Food Safety Agilent

Solutions that meet your demands for
food safety testing
Excellent choices for food applications
Table of Contents
The table of contents on the right has been linked to the individual sections in this compendium. Click on the text to jump to a specific section. If you prefer to search for words or phrases used in any of the enclosed applications notes, click on the Search button below to open Acrobats Search window.
Solution Segments
Pesticides and Residues
Pesticides and Herbicides Antibacterial Drug Residues
Contaminants
Acrylamides Food Packaging PCBs Toxins Trace Metals
Search
Natural Compounds and Additives

Additives Amino Acids Carbohydrates Fats and Oils Flavors and Fragrances Miscellaneous Proteins Vitamins
Bioanalysis
Applications
Productivity Tools
Applications
Index
Pesticides & Residues Pesticides and Herbicides
> Return to Table of Contents > Search entire document
Replacing Multiple 50-Minute GC and GC-MS/SIM Analyses with One 15-Minute Full-Scan GC-MS Analysis for Nontargeted Pesticides Screening and >10x Productivity Gain Application
Food Safety
Author
Chin-Kai Meng and Mike Szelewski Agilent Technologies 2850 Centerville Road Wilmington, DE 19808
Introduction
To have a plentiful food supply, most fruits and vegetables are treated with pesticides (insecticides, fungicides, herbicides, etc.) to protect primarily against insects, molds, and weeds. Therefore, in order to ensure food safety, the food supply is frequently monitored for pesticide residues. Nowadays, the pesticide monitoring is expanding beyond food, for example, to botanical dietary supplements. The analytical challenge to monitor (identify and quantify) trace multiresidues requires an effective and universal extraction and analysis method for maximum productivity and efficiency. Up to now, the trade-off in analysis has been between sensitivity and confirmation. Element-selective gas chromatograph (GC) detectors, such as the flame photometric (FPD), electron capture (ECD), electrolytic conductivity (ELCD), and halogen selective (XSD) detectors, provide excellent selectivity and sensitivity; however, they lack the capability to identify. On the other hand, mass spectrometry (MS) is capable of identifying an analyte by fullscan library match or multiple target and qualifier ion ratios from selected ion monitoring (SIM). However, MS sometimes lacks the selectivity to
Abstract
Pesticide analysis of fruits and vegetables requires finding trace-level residues in complex matrices. Up to now, the typical trade-off is between sensitivity and confirmation. Therefore, multiple injections are needed for screening and confirmation using gas chromatography with mass spectrometry (GC-MS) or GC-MS in combination with GC with element-selective detection. With the recent introduction of hardware and software tools, for example, capillary flow three-way splitter, trace ion detection, and deconvolution, a 15-minute fast analysis can match the results obtained from three injections of approximately 50 minutes each. A table comparing the results from the Food and Drug Administration/Center for Food Safety and Applied Nutrition (FDA/CFSAN) procedure using the traditional multi-instrument approach and the new Agilent single injection approach shows that Agilent's fast analysis is capable of finding all the target analytes in less than one-tenth of the current FDA/CFSAN total analysis time.
find target analytes in a complex matrix full of interferences and chemical background. Analyte spectra are sometimes overwhelmed by similiar ions contributed from the coextractives in the matrix that prevent the analyte of interest from being identified or confirmed. The compromise and the typical approach are to use selective GC detector(s) to flag potential target analytes and use MS SIM for confirmation. For instance, many laboratories screen food samples for semivolatile pesticides using the ECD or ELCD (or XSD) for organohalogen, FPD or pulsed FPD (PFPD) for organophosphorus, and NPD for nitrogen-containing targets [1 5]. Any found targets are further confirmed by GC-MS/SIM. In addition, other procedures have used GC-MS/SIM entirely for the screening of pesticides in foods [6 8]. In most of these procedures, multiple injections are needed to identify hundreds of compounds at the detection limit in the low parts-per-billion (ppb) levels. To improve the efficiency and increase the productivity of screening for all of these pesticides, the challenge is to reduce the GC-MS or the combination of GC and GC-MS analysis times. There are several hundred pesticides typically used in the world, and each country has its own pesticide tolerance levels for different agricultural commodities. This presents another analytical challenge in multiresidue monitoring: to develop a nontargeted procedure to identify pesticides at trace levels in different food matrices. These challenges are met by the recent introduction of hardware and software tools, including GCMS, capillary flow three-way splitter, trace ion detection, and deconvolution reporting software (DRS). The splitter allows multiple GC as well as MS signals to be acquired from a single injection for productivity gains (from three injections down to one). Trace ion detection minimizes noise on the signal and DRS separates target analyte ions from matrix background ions. Several sample extracts were analyzed by the current Food and Drug Administration/Center for Food Safety and Applied Nutrition (FDA/CFSAN ) multiple injection process and this new Agilent pesticide system. With DRS, the demand for chromatographic resolution is minimum; therefore, the Agilent system was running the analysis at a 3x faster speed (one-third the analysis time) to further increase productivity. A table comparing the
results from the current FDA/CFSAN multi-instrument approach and the new Agilent single-injection approach shows that not only is Agilents fast analysis capable of finding all the target analytes, but it is also accomplished in just one-tenth of the current FDA/CFSAN total analysis time.
Experimental
Sample Preparation Sample extracts of fresh produce were prepared by FDA based on modifications of the QuEChERS protocols [9, 10]: Homogenize 1 to 2 kg of sample 15 g sample + 15 mL 1% AcOH/ACN, homogenized Add 6 g MgSO4 and 2.5 g NaOAc, shake vigorously for 1 minute, and centrifuge Transfer ~15 mL + 0.5 g C-18 + 1.2 g MgSO4, shake, and centrifuge for 5 min at 3,000 rpm Transfer ~12 mL + 0.4 g PSA + 0.2 g GCB + 1.2 g MgSO4, vortex Add 4 mL toluene, shake, and centrifuge Transfer 6 to 8 mL, evaporate and bring to volume with toluene, add I.S. Add MgSO4, vortex and centrifuge, transfer to ALS vials GC and GC-MS analysis Sample preparation of dried ginseng powder is similar to that used for fresh produce, but smaller sample sizes (2 g) were used [11]. Capillary Flow Three-Way Splitter One of the capillary flow devices is a three-way splitter, which consists of two half plates bonded together (diffusion bonding) to form a plate with the etched flow channels inside. The splitter is only 6.5 cm tall and 3 cm wide and is mounted on the side of the oven wall (see Figure 1). The low thermal mass minimizes cold spots and peak broadening. All capillary flow devices use metal column ferrules, have extremely low dead volumes, are inert, and do not leak, even after many oven cycles.
To MSD
To FPD To ECD
Aux EPC in
Column in
different inlets and outlets. Aux pressure can be either an inlet (for the splitter flow restrictors connected to different detectors) or an outlet (for the analytical column). A graphical user interface makes the configuration easy to set up. Once all the columns and restrictors are configured, the backflush can be executed easily. Backflush Traditional bakeout step for removing late eluters could be very time consuming, or even as long as the analysis time depending on the matrix. Backflush is a simple technique to remove high boilers from the column faster and at a lower column temperature to cut down analysis time and increase column lifetime. Capillary flow devices (in this case, a three-way splitter) also provide backflush [13, 14] capability. Backflush is a term used for the reversal of flow through a column such that sample components in the column are forced back out the inlet end of the column. By reversing column flow immediately after the last compound of interest has eluted, the long bake-out time for highly retained components can be eliminated. Therefore, the column bleed and ghost peaks are minimized, the column will last longer, and the MS ion source will require less frequent cleaning. The split vent trap may require replacement more frequently than usual. Figures 2 and 3 are two screen shots from the MSD ChemStation software, providing a summary of the backflush operation. In Figure 2, the column and three restrictor dimensions and respective detectors are shown (the setup came from the column configuration section). For MSD, the user can choose the vacuum pump installed on the system. This information will be used to calculate if the backflush is within the system flow limits. By clicking on the Evaluate button, the screen shown in Figure 3 appears, listing the maximum flow for each detector and the void volumes for a certain backflush time. In this example, Aux pressure is at 60 psi, inlet is at 1 psi, and oven is at 280 C. The backflushing flow is shown to be 8.66 mL/min, and the void time is shown to be 0.16 min. Therefore, backflushing for 2.5 minutes will send 15.6 void volumes through the column. This is useful for developing the backflush method. Figures 2 and 3 simplify the setup and development of a backflush method.
Etched flow channel inside two diffusion bonded plates Capillary tube connection via metal ferrule
Figure 1.
Flow diagram of the three-way splitter. The picture shows the splitter mounted on the oven wall.
The three-way splitter enhances productivity by splitting column effluent proportionally to multiple detectors: MSD, dual flame photometric detector (DFPD) and micro-electron capture detector (ECD). Therefore, two GC detector signals can be acquired together with the MS data (both SIM and scan signals if desired) from one injection [12]. The exit end of the analytical column is installed into one of the four ports on the splitter using a metal ferrule. The other three ports are connected to three detectors via restrictors (deactivated capillary tubing) of varying diameter and length to set the split ratio among the three detectors. Restrictors are sized for 1:1:0.1 split ratio in favor of MSD and DFPD (ECD has 1/10 of the flow to MSD), with similar hold-up times. The splitter uses auxiliary (Aux) electronic pneumatics control (EPC) for constant pressure makeup flow. The makeup gas (Aux pressure 6) at the splitter is fixed at 3.8 psi to maintain the split ratio throughout the run. This multisignal configuration provides full-scan data for library searching, SIM data for trace analysis, DFPD (phosphorus or sulfur mode), and ECD data for excellent selectivity and sensitivity from complex matrices. The trade-off is the decrease of analyte concentration in any detector due to the flow splitting and the additional makeup gas from the splitter. An analyte would have similar retention times in all three detectors. Therefore, the GC data can be used in two ways: first, to confirm the presence of target analytes found by the MSD deconvolution reporting software (DRS), and second, to highlight potential target compounds to be further confirmed by MSD. With the new 7890A GC software, up to six columns/ restrictors can be configured/assigned to
Figure 2.
Backflush setup in ChemStation.
Figure 3.
Automated backflush calculations in ChemStation.
Deconvolution
Another useful feature in Figure 3 is the warning, shown as highlighted yellow cells. In this example, setting the backflush pressure to 60 psi sends more than the allowable flow (60 mL/min) to the FPD. Therefore, the backflush pressure setting and the actual flow value to FPD are shown in yellow as warnings. Although the system will accept the setup, the high flow may cause consequences in the analysis, for example, flameout. Trace Ion Detection Trace ion detection [15] is a filtering algorithm to smooth peaks. This filtering is an advanced form of averaging used to remove the noise riding on the signal. The implications from TID are typically a slight loss in peak height and some peak broadening. The default setting in ChemStation for TID is off. It should be turned on for any analysis that uses deconvolution and has more than six sampling points across a peak. TID provides better signal-to-noise ratios and helps deconvolution to confirm target compounds as shown in the Results section. Deconvolution In GC/MS, deconvolution is a mathematical technique that separates overlapping mass spectra into deconvoluted spectra of the individual components. Figure 4 is a simplified illustration of this process. Here, the total ion chromatogram (TIC) and apex spectrum are shown on the left. In a complex matrix, a peak may be composed of multiple overlapping components and matrix background ions; therefore, the apex spectrum is actually a composite of these constituents. A mass spectral library search would give a poor match, at best, and certainly would not identify all of the individual components that make up the composite spectrum. The deconvolution process groups ions whose individual abundances rise and fall together within the spectrum. The deconvolution process first corrects for the spectral skew that is inherent in quadrupole mass spectra and determines a more accurate apex retention time of each chromatographic peak. As illustrated in Figure 4, deconvolution produces a cleaned spectrum for each overlapping component. These individual spectra can be library searched with a high expectation for a good match. Deconvolution significantly reduces chromatographic resolution requirements, allowing much shorter analysis times.
TIC and spectrum

TIC
Deconvoluted peaks and spectra

Component 1 extracted spectrum
Library search each component to identify
Figure 4.
Deconvolution process of three overlapped peaks.
Agilent Deconvolution Reporting Software (DRS) utilizes the AMDIS deconvolution program from the National Institute of Standards and Technology (NIST), originally developed for trace chemical weapons detection in complex samples [16]. DRS presents the analyst with three distinct levels of compound identification: (1) ChemStation, based on retention time and four ion agreement; (2) AMDIS, based on cleaned spectra full spectral matching and expected retention time window as a qualifier; and (3) NIST05 search using a >163,000compound library [17, 18]. In this application, both the ChemStation quantitation database and the AMDIS library have the same 926 entries. These entries include pesticides, numerous metabolites, endocrine disruptors, important PCBs and PAHs, certain dyes, synthetic musk compounds, and several organophosphorus fire retardants [18]. The AMDIS software, shipped with the NIST05 Library CD-ROM, is also capable of deconvoluting selected ion monitoring (SIM) data [19], while previous AMDIS revisions were not. Testing has shown that proper compound identification requires four ions per compound. All Agilent DRS databases are retention time locked and have both full-scan and SIM libraries for AMDIS. Instrument Method The system used for this study consists of an Agilent 7890A GC with split/splitless inlet, a threeway splitter, ECD, DFPD, and 5975 MSD. For a detailed description of SIM/scan and the splitter system configuration, please refer to the experimental section of reference [12]. See Table 1 for hardware detail and settings.
Table 1. Gas Chromatograph, Mass Spectrometer, and Three-Way Splitter Operating Parameters GC Injector Syringe size Injection volume Solvent A wash Solvent B wash Sample wash Sample pump Plunger speed Inlet Mode Inlet temperature Pressure Agilent Technologies 7890A with 240V fast oven option Agilent Technologies 7683 10 L 1 L 1 (pre), 3 (post) 1 (pre), 3 (post) 0 4 Fast EPC split/splitless Splitless 250 C ~24.4 psi (chlorpyrifos methyl RT locked to 5.531 min, 3x speed) constant pressure mode 50.0 mL/min 2 min 3 mL/min Switched Off Helium Helix double taper liner, deactivated, p/n 5188-5398 C /min 75 9 24 13.96 min 1.0 min 2.5 min 280 C Agilent Technologies HP-5MS, p/n 19091S-431 15.0 m 0.25 mm 0.25 m Constant pressure RT locked to chlorpyrifos methyl at 5.531 min, 3x analysis speed 3.5 mL/min Aux pressure 6 3.8 psi (Aux EPC pressure to three-way splitter), helium gas 280 C 2.5 min 1 psi 60 psi (column outlet) ECD 300 C 60.0 mL/min Nitrogen 20 Hz Dual FPD 250 C Final (C) 70 150 200 280 Hold (min) 0.67 0 0 3.33 Hydrogen flow Air flow Const Col + Makeup Make gas type Lit offset Data rate Transfer line AUX Thermal 1 AUX Pressure 6 Gas type Initial pressure Backflush pressure MSD Tune file Mode Solvent delay EM voltage Low mass High mass Threshold Samples Scans/sec Quad temp Source temp Three-way splitter Pressure Split ratio MSD restrictor DPFD restrictor ECD restrictor Flow to MSD Flow to DFPD Flow to ECD Makeup (Aux 6) Software GC/MSD ChemStation MS Libraries Agilent part number G1701EA (version E.01.00 or higher) NIST05a mass spectral library (Agilent part number G1033A) Agilent RTL Pesticide and Endocrine Disruptor Libraries (926 entries) in Agilent and AMDIS formats (part number G1672AA) Automated Mass Spectral Deconvolution and Identification Software (AMDIS_32 version 2.65 Build 116.66) NIST MS Search (version 2.0d or greater) (comes with NIST'05a mass spectral library Agilent part number G1033A) Agilent part number G1716AA (version A.03.00 or higher) 75.0 mL/min 100.0 mL/min 60.0 mL/min Nitrogen 2.00 20 Hz 250 C MSD transfer line, 280 C Three-way splitter Helium 3.8 psi 60 psi Agilent Technologies 5975C MSD Atune.u Scan 1.50 min Atune voltage 50 amu 550 amu 0 2 2.91 150 C 230 C Agilent 7890A Option 890, installed during factory assembly 3.8 psi (Aux pressure 6 setting) 1:1:0.1 MSD:DFPD:ECD 1.444 m 0.18-mm id deactivated fused silica tubing, p/n 160-2615-10 0.532 m 0.18-mm id deactivated fused silica tubing, p/n 160-2615-10 0.507 m 0.10-mm id deactivated fused silica tubing, p/n 160-2635-1 3.43 mL/min (at 70 C), 1.53 mL/min (at 280 C) 3.43 mL/min (at 70 C), 1.53 mL/min (at 280 C) 0.343 mL/min (at 70 C), 0.153 mL/min (at 280 C) 3.19 mL/min (at 70 C), 1.52 mL/min (at 280 C)
Purge flow Purge time Septum purge flow Septum purge mode Gas saver Gas type Liner Oven Oven ramp Initial Ramp 1 Ramp 2 Ramp 3 Runtime Oven equilib time Post-run time Post-run temperature Column Length Diameter Film thickness Mode
Nominal initial flow Outlet Outlet pressure Backflush (post-run) Oven Time Inlet Aux pressure 6 Front detector Temperature Const col + makeup Make gas type Data rate Back detector Temperature 6
Deconvolution software Library searching software
Deconvolution reporting software
Results and Discussion

Backflush Example Blank runs, made after separate milk analyses with different backflush (BF) times, are shown in Figure 5. The top TIC is a blank run after a milk extract analysis stopped at 42 minutes and the system backflushed for 1 minute. The next TIC is a blank run after another milk extract analysis stopped at 42 minutes and backflushed for 2 minutes and so on for the other five TICs. It is interesting to confirm graphically that the latest eluters disappeared from the TIC earliest in backflushing.
100000 50000 0 100000 50000 0 100000 50000 0 100000 50000 0 100000 50000 0 100000 50000 0 100000 50000 0
Trace Ion Detection Figure 6 compares the signals when TID is on and off. Visually, it is obvious that TID smoothes the noise riding on top of the signal. When TID was on, Atrazine was successfully identified by AMDIS. When TID was off, Atrazine was not found by AMDIS and resulted in a false negative. Figure 7 compares TID on and off for two different analytical conditions of the same ginseng extract. On the right, the fast (3x) analysis was a 1-L splitless injection with TID on. The analyte Diazinon was found by AMDIS with a peak width less than 5 seconds. On the left side, the normal (1x)
BF for 1 min
BF for 2 min
BF for 3 min
BF for 4 min
Note: late eluters were backflushed out first
BF for 5 min
BF for 6 min
BF for 7 min
10.00 20.00 30.00 40.00 50.00
Column is clean
60.00 70.00 80.00
Figure 5.
Differences in blank runs as the result of seven different backflush times. Atrazine found by AMDIS
TID on
TID off
Atrazine not found by AMDIS Figure 6. The power of deconvolution with TID for atrazine, from AMDIS. 7
analysis was a 5-L cold splitless injection using a PTV with TID off. The 9-second-wide Diazinon was not found by AMDIS, also a false negative. Both examples show that TID is a very useful feature for trace target analysis. Benefits of TID: Improves the signal-to-noise ratio AMDIS is more thorough in identifying components, resulting in fewer false positives Improves library match quality Improves area repeatability, resulting in more reliable quantitation DRS and Splitter Figures 8 and 9 are DRS reports for ginseng and peach extracts with pesticides highlighted (for a detailed explanation of the report, please refer to references [17] and [20]). Figures 10 and 11 show simultaneously collected GC and MS signals (RT locked) for the corresponding ginseng and peach extracts from a three-way splitter. The presence of the GC peaks from the ECD and FPD (P) helps confirm the targets reported by DRS. Each run is finished at 15 minutes using the 3x speed and a 240V oven. With deconvolution, less peak resolution is required for compound identification. A 4minute backflush is added after the run to make sure that the column is clean to maintain the next runs locked RTs for all peaks.
PTV, 1x speed, 5 L injection No TID, Diazinon not found (false negative)
Deconvolution Figures 12 through 15 show the results from AMDIS. There are three spectra for each target compound found by AMDIS. The top window shows the spectrum (scan) from the TIC. This is the only spectrum that would be available for library searching without deconvolution obviously quite useless. The middle window shows the deconvoluted spectrum and the bottom window is the target compounds spectrum in the library. The compound confirmation can be done easily and with confidence by visually comparing the bottom two spectra. The power of deconvolution is appreciated while comparing the top two spectra (the raw scan and the spectrum hidden in the raw scan). It is easy to further confirm the hits found by deconvolution. In Figure 9, four pesticides found by AMDIS in the peach extract have a match factor of about 80 or lower. The four pesticides are Cabaryl, Captan, Propiconazole, and Fenbuconazole. A SIM method of these compounds was set up to analyze the peach extract. By selecting the proper AMDIS library (full-scan or SIM), DRS can process full-scan as well as SIM data files [19]. Figure 16 is the DRS report of the peach SIM analysis. The high match factor (99 or higher) and the small RT difference of all targets found by AMDIS confirm the presence of all compounds.
Splitless, 3x speed, 1 L injection with TID, Diazinon found
Figure 7.
The power of deconvolution with TID for diazinon.
Figure 8.
DRS report for ginseng with pesticides highlighted.
Figure 9.
DRS report for peach with pesticides highlighted.
Ginseng
1e+07 8000000 6000000 4000000 2000000 2.00 3.00 4.00 3.524 5.00 6.00 7.00 8.00 8.699 9.00 10.00 9.413 11.00 12.00 13.00
TIC
9e+07 8e+07 7e+07 6e+07 5e+07 4e+07 3e+07 2e+07 1e+07
Tetrachloro-m-xylene (ISTD) Chlorthaldimethyl

4.501 6.248 2.00 1.660 3.00 4.00 3.665 5.00 6.00 7.005 7.00 8.00
8.290
7.783
8.744 8.992 9.093 9.059 9.265 9.528 9.646 9.784
ECD Azoxystrobin
9.00
10.00
11.00
12.00
13.00
6000000 5000000 4000000 3000000 2000000 1000000
Tributylphosphate (ISTD)
9.413
Diazinon
4.831 9.313 7.641
FPD (P)
2.00
3.00
4.00
5.00
6.00
7.00
8.00
9.00
10.00
11.00
12.00
13.00
Figure 10. Simultaneous display of MSD and GC selective detector signals for ginseng.
2200000 1800000 1400000 1000000 600000 200000 2.00 3.5e+07 3e+07 2.5e+07 2e+07 1.5e+07 1e+07 5000000 2.00 5000000 4000000 3000000 2000000 1000000 2.00 3.00 4.00 9.358 4.773 5.00 6.00 7.00 8.00 9.00 10.00 9.826 9.910 2.703 4.696 3.932 3.00 4.00 3.641 1.838 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00
Peach TIC
11.00
12.00
13.00
Tetrachloro-m-xylene (ISTD)
3.519
Carbaryl
5.640 5.00
7.093 6.680
Endosulfan(alpha) 7.556 Phosmet 9.538 Captan

9.081 8.00 9.00 9.523 10.559 10.00 11.00
ECD
12.382 12.00 13.00
6.00
7.00
Tributylphosphate (ISTD)
Phosmet FPD (P)
11.066 11.00 12.00 13.00
Figure 11. Simultaneous display of MSD and GC selective detector signals for peach. 10
Scan at 12.299 min
Deconvoluted/extracted spectrum
Library spectrum Azoxystrobin
Figure 12. Raw (dirty) spectrum, deconvoluted (clean) spectrum, and library spectrum of azoxystrobin found in ginseng, from AMDIS.
11
Scan at 5.615 min
Library spectrum Carbaryl
Figure 13. Raw (dirty) spectrum, deconvoluted (clean) spectrum, and library spectrum of cabaryl found in peach, from AMDIS.
12
Scan at 10.776 min
Library spectrum Fenbuconazole
Figure 14 Raw (dirty) spectrum, deconvoluted (clean) spectrum, and library spectrum of fenbuconazole found in peach, from AMDIS.
13
Scan at 8.934 min
Library spectrum Endosulfan sulfate
Figure 15 Raw (dirty) spectrum, deconvoluted (clean) spectrum, and library spectrum of endosulfan sulfate found in tomato, from AMDIS.
14
Figure 16. DRS report from the SIM analysis of peach. Please refer to reference 18 for the explanation of the fictitious CAS number assigned to Propiconazole-II (999048032).
Comparison of Incurred Samples The current approach at FDA/CFSAN is to find a wide suite of organohalogen and organophosphorus pesticide residues. This requires four injections (GC-MS/SIM and GC-ELCD for organohalogen and GC-MS/SIM and GC-FPD for organophosphorus screening) of approximately 50-minutes runtime each (total runtime = 200 minutes). Table 2 shows that FDA found several target compounds in three extracts as well as quantitation
Table 2.
results from both GC and MS. In comparison, using the new tools (splitter, TID, and deconvolution) found as many target compounds and a few more in just one short (15-minute) full-scan analysis. The three-way splitter was used to get selective GC signals (ECD and FPD) for confirmation purposes. Due to column effluent splitting to three detectors (1:1:0.1), the MSD is getting less than half of the amount injected. FDA/CFSAN GC and GC/MS/SIM analyses for organohalogen monitor-
Comparison of the Agilent Pesticide System Results with the FDA Results Agilent DRS (full scan/TID) FDA (FPD, ELCD, SIM) Diazinon (FPD, SIM) GC-FPD or ELCD 25 3 ppb GC-MS/SIM 25 2 ppb
Ginseng
Diazinon Chlorthal-dimethyl Azoxystrobin Carbaryl Captan Endosulfan (alpha) Phosmet Propiconazole I and II Fenbuconazole Chlorothalonil Endosulfan (alpha) Endosulfan (beta) Endosulfan sulfate 1 15-min injection (splitter) found these
Peach
Phosmet (FPD, SIM)
320 37
230 23
Tomato
Chlorothalonil (ELCD, SIM) Endosulfan (alpha) (ELCD, SIM) Endosulfan (beta) (ELCD, SIM) Endosulfan sulfate (ELCD, SIM) 2 50-min injections found these
205 10 16 2 34 4 14 2
153 47 26 4 47 5 21 6
FDA quant results
15
ing found endosulfan sulfate at 14/21 ppb (pg/L) in tomato. Agilent MSD/DRS also found this compound in full-scan mode with less than half the amount reported by FDA/CFSAN available at the MSD due to the split. Several other target compounds that did not contain any halogens or the organophosphorus moeity at the low ppb concentrations were also identified by DRS, such as carbaryl (C12H11NO2) in peach and azoxystrobin (C22H17N3O5) in ginseng. The two FDA/CFSAN procedures for organohalogen and organophosphorus pesticides never would have been able to detect these additional nitrogen-containing pesticides. This shows that deconvolution of data acquired with TID is capable of identifying compounds below 10 pg on column in full-scan mode.
Department of Food and Agriculture, Fresenius J. Anal. Chem, 1991, 339, 376383 2. J. Cook, M. P. Beckett, B. Reliford, W. Hammack, and M. Engel, Multiresidue Analysis of Pesticides in Fresh Fruits and Vegetables Using Procedures Developed by the Florida Department of Agriculture and Consumer Services, J. AOAC Int, 1999, 82, 14191435 3. G. E. Mercer, Determination of 112 Halogenated Pesticides Using Gas Chromatography/ Mass Spectrometry with Selected Ion Monitoring, J. AOAC Int, 2005, 88, 14521462 4. G. E. Mercer and J. A. Hurlbut, A Multiresidue Pesticide Monitoring Procedure Using Gas Chromatography/Mass Spectrometry and Selected Ion Monitoring for the Determination of Pesticides Containing Nitrogen, Sulfur, and/or Oxygen in Fruits and Vegetables, J. AOAC Int, 2004, 87, 12241236 5. USDA Pesticide Data Program Analytical Methods: http://www.ams.usda.gov/science/pdp/ Methods.htm 6. J. Fillion, R. Hindle, M. Lacroix, and J. Selwyn, Multiresidue Determination of Pesticides in Fruit and Vegetables by Gas ChromatographyMass-Selective Detection and Liquid Chromatography with Fluorescence Detection, J AOAC Int, 1995, 78, 12521266 7. S. Nemoto, K. Sasaki, S. Eto, L. Saito, H. Sakai, T. Takahashi, Y. Tonogai, T. Nagayama, S. Hori, Y. Maekawa, and M. Toyoda, Multiresidue Determination of 110 Pesticides in Agricultural Products by GC/MS(SIM), J. Food Hyg. Soc, Japan, 2000, 41, 233241 8. G. F. Pang, C. L. Fan, Y. M. Liu, Y. Z. Cao, J. J. Zhang, X. M. Li, Z. Y. Li, Y. P. Wu, and T. T. Guo, Determination of Residues of 446 Pesticides in Fruits and Vegetables by Three-Cartridge SolidPhase Extraction Gas Chromatography-Mass Spectrometry and Liquid ChromatographyTandem Mass Spectrometry, J. AOAC Int, 2006, 89, 740771 9. M. Anastassiades, S. J. Lehotay, D. Stajnbaher, and F. J. Schenck, Fast and Easy Multiresidue Method Employing Acetonitrile Extraction/Partitioning and Dispersive Solid-Phase Extrac-
Conclusions
The trade-off in trace-level pesticide residue analysis is sensitivity versus confirmation. Therefore, the common practice is to use element-selective GC detectors to screen the extracts and use MS/SIM to confirm hits found by GCs. This can take as many as four injections to have a complete residue analysis from a sample extract. Recent introduction of hardware and software tools, which include the capillary flow three-way splitter, trace ion detection, and deconvolution reporting software, can increase productivity dramatically. With deconvolution the demand for chromatographic resolution is lowered; therefore, the Agilent system can run the analysis at a 3x faster speed to further increase productivity. A single-injection approach even at the 3x fast speed can replace the three-injection approach. A table comparing the results from the current FDA/CFSAN multi-instrument approach and the new Agilent single-injection approach shows that not only is Agilents fast analysis capable of finding all the target analytes, but it can also do it in just one-tenth of the current FDA/CFSAN total analysis time.
References
1. S. M. Lee, M. L. Papathakis, H. C. Feng, G. F. Hunter, and J. E. Carr, Multipesticide Residue Method for Fruits and Vegetables: California
16
tion for the Determination of Pesticide Residues in Produce, 2003, J. AOAC Int, 86:412431 10.S. J. Lehotay, K. Matovsk, and A.R. Lightfield, Use of Buffering and Other Means to Improve Results of Problematic Pesticides in a Fast and Easy Method for Residue Analysis of Fruits and Vegetables, 2005, J. AOAC Int, 88:615629 11.J. W. Wong, M. K. Hennessy, D. G. Hayward, A. J. Krynitsky, I. Cassias, and F. J. Schenck, Analysis of Organophosphorus Pesticides in Dried Ground Ginseng Root by Capillary Gas Chromatography-Mass Spectrometry and -Flame Photometric Detection, J. Agric. Food Chem, 2007, 55, 11171128 12.Chin-Kai Meng and Bruce Quimby, Identifying Pesticides with Full Scan, SIM, ECD, and FPD from a Single Injection, Agilent Technologies publication, 5989-3299, July 2005 13.Chin-Kai Meng, Improving Productivity and Extending Column Life with Backflush, Agilent Technologies publication, 5989-6018EN, December 2006 14.Matthew Klee, Simplified Backflush Using Agilent 6890 GC Post Run Command, Agilent Technologies publication, 5989-5111EN, June 2006 15.Randy Roushall and Harry Prest, The 5975C Series MSDs: Method Optimization and Trace Ion Detection, Agilent Technologies publication, 5989-6425EN, March 2007 16.http://chemdata.nist.gov/mass-spc/amdis/ explain.html
17. Philip L. Wylie, Michael J. Szelewski, Chin-Kai Meng, and Christopher P. Sandy, Comprehensive Pesticide Screening by GC/MSD Using Deconvolution Reporting Software, Agilent Technologies publication, 5989-1157EN, May 2004 18.Philip L. Wylie, Screening for 926 Pesticides and Endocrine Disruptors by GC/MS with Deconvolution Reporting Software and a New Pesticide Library, Agilent Technologies publication, 5989-5076EN, April 2006 19.Mike Szelewski and Chin-Kai Meng, New Features of Deconvolution Reporting Software Revision A.02, Agilent Technologies publication, 5989-4159EN, November 2005 20.Bruce Quimby and Mike Szelewski, Screening for Hazardous Chemicals in Homeland Security and Environmental Samples Using a GC/MS/ ECD/FPD with a 731 Compound DRS Database, Agilent Technologies publication, 5989-4834EN, February 2006
Acknowledgements
The authors would like to thank Jon Wong, FDA/CFSAN, for the sample extracts and results used in this study as well as valuable feedback on this application.
For More Information

For more information on our products and services, visit our Web site at www.agilent.com/chem.
17
www.agilent.com/chem
Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2007 Printed in the USA December 18, 2007 5989-7670EN
Reducing Analysis Time Using GC/MSD and Deconvolution Reporting Software
Application
Food and Flavors
Authors
Mike Grady, Steve Morrison, and Bob Deets Campbell Soup Company Campbell Place Camden, NJ 08103 USA Mike Szelewski Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808 USA
Abstract
Analyzing a complex matrix can be accomplished using multiple specific detectors, but a significant time savings is realized using GC/MSD. Using an available Retention Time Locked database with Deconvolution Reporting Software (DRS) adds a second expert opinion. GC/MSD with DRS provides the fastest methodology to the fewest number of false positives/negatives with the greatest confidence in the results.
dual flame photometric for phosphorus and sulfur (DFPD). Most of the analytes, depending on their molecular formula, could be run on two of these specific detectors for confirmation. However, that would take twice the time. A second choice is to run each sample on unlike columns of differing polarities to the same type detector. Both of these approaches, using ECD, NPD, and/or DFPD, involve more sample handling/tracking, usually twice the number of runs and subsequent data interpretation, and still only have retention time as an identifier. Some methodologies require final confirmation by GC/MSD. Also, it is very difficult to analyze for hundreds of compounds using a nonMS method due to peak overlaps. Campbell Soup Company has been moving from GC-specific detector analyses to GC/MSD analyses for the above-mentioned reasons. A significant time savings can be realized using GC/MSD. It is imperative, however, that data quality be maintained while increasing productivity. Agilent Technologies recently introduced Deconvolution Reporting Software (DRS) for use with a GC/MSD system [1]. DRS automatically combines the results from the MSD ChemStation Quantitation with AMDIS Deconvolution and NIST Search into an easy-to-read report. Using an available Retention Time Locked database reduces methods development for DRS and speeds data comparison among labs. Positives found by normal quantitation are confirmed, and the analyst is directed to further target compound hits to verify. The analyst
Introduction
Analyzing a complex matrix can be accomplished using multiple specific detectors such as electron capture (ECD), nitrogen-phosphorus (NPD), and
need not spend the time to review hundreds of possible compounds, which could take hours per sample. GC/MSD with DRS provides the fastest methodology to the fewest number of false positives/negatives with the greatest confidence in the results. It typically takes less than 2 minutes to process a DRS Report. The purpose of this application note is to show the use of GC/MSD with DRS on complex samples in Campbells Research Lab. It is also intended to show a time savings while maintaining data quality.
The HP-5MS column was used by Agilent to develop the original method and is run in constant pressure mode. Constant pressure methods can be precisely scaled, or sped up, for faster analyses. The retention times of 927 analytes have been recorded on this column. The system is Retention Time Locked to methyl chlorpyrifos at 16.596 min. The primary benefit of RTL for a food laboratory is the ability to maintain retention times after clipping or changing the column. Other benefits include: 1) constant quantitation database and integration events times; 2) switching group times remain constant for laboratories performing SIM analyses; 3) multi-site laboratories can easily compare data; 4) commercially available RTL databases can be used. Additional information is available at www.agilent.com/chem, with application notes detailing the numerous benefits of RTL. The injector parameters are modified from the default Fast Injection to Variable. This allows matching the injector parameters to the PTV-SV requirements. Most importantly, the injection speed is slowed to accommodate the evaporation of the solvent during injection. The 5973N MSD had been upgraded to an inert source and Performance Electronics. The inert source has shown improved response and less degradation for active compounds. Performance electronics minimize noise at faster Scan sampling rates and allow shorter dwell times and more ions/group for SIM acquisitions. A sampling rate of 2 and a scan range of 35-500 yields 3.12 scans/sec. This results in at least 10 data points across the earliest narrow peaks that are 0.055 min wide. This is a good number of points for both quantitation and deconvolution while minimizing noise.
Experimental
Instrument Operating Parameters The instrument operating parameters are listed in Table 1. These conditions may have to be optimized for use in another laboratory. A Programmable Temperature Vaporizing inlet (PTV) is used in the solvent vent mode (SV). The sample is injected at or below the boiling point of the solvent, in this case 50 C. Solvent is evaporated through the split vent line with helium at 200 mL/min for 0.3 min. At 0.5 min, the PTV is rapidly heated to 320 C, transferring the analytes, with minimum solvent, onto the column. The column is held at the initial temperature of 70 C during this process. The PTV-SV allows larger volumes of sample to be injected, 10 L for this study, versus the typical 1 L for this column. The PTV inlet liner, 5183-2037, is multi-baffled and deactivated. It does not contain glass wool, which could contribute to active compound degradation. This liner has sufficient capacity to accommodate the 10 L injection volume
Table1. GC Inlet Mode
Gas Chromatograph and Mass Spectrometer Conditions Agilent Technologies 6890N EPC PTV Solvent vent C/min 600 50 On 50 C 10.00 min (On) On 17.98 psi (On) 0.30 min 200.0 mL/min 0.0 psi 400.0 mL/min 1.00 min 403.9 mL/min On 20.0 mL/min 2.00 min Helium Agilent PTV Liner part# 5183-2037 120 V C/min 25 3 8 41.87 min 0.5 min 325 C Agilent Technologies HP 5 MS, part# 19091S-433 30.0 m 0.25 mm 0.25 m Constant Pressure 17.98 psi 1.9 mL/min Front MSD Vacuum Next C 70 150 200 280 Hold min 2.00 0.00 0.00 10.00 Next C 50 320 200 Hold min 0.50 3.00 0.00
RTL Front injector Sample washes Sample pumps Injection volume Syringe size PreInj Solv A washes PreInj Solv B washes PostInj Solv A washes PostInj Solv B washes Viscosity delay Plunger speed Injection speed Draw speed Dispense speed PreInjection dwell PostInjection dwell MSD Upgrades Solvent delay Low mass High mass Threshold Sampling Scans/sec Quad temperature Source temperature Transfer line temp Tune Type Calibration Standards
System Retention Time Locked to methyl chlorpyrifos at 16.596 min
Temperature ramp Initial Ramp 1 Ramp 2 Cryo Cryo use temperature Cryo timeout Cryo fault Pressure Vent time Vent flow Vent pressure Purge flow Purge time Total flow Gas saver Saver flow Saver time Gas type PTV liner Oven Oven ramp Initial Ramp 1 Ramp 2 Ramp 3 Total run time Equilibration time Oven max temperature Column Length Diameter Film thickness Mode Pressure Nominal initial flow Inlet Outlet Outlet pressure
0 3 10.00 L 25.0 L 0 1 2 2 1s Variable 50.00 L/min 600.00 L/min 1000.00 L/min 0.00 min 0.00 min Agilent Technologies 5973N Inert source and Performance Electronics 4.00 min 35 amu 500 amu 50 2 3.12 150 C 230 C 280 C Autotune
Prepared from certified reference standards available from ChemServe and Crescent Chemical Company. All standards were corrected for purity.
Extraction Procedure An appropriate amount of commodity is weighed, typically 10-15 grams. Surrogates and, if necessary, fortification standards (spike) are added. The commodity is extracted with 1% acetic acid in acetonitrile, centrifuged [2], and passed through an SPE cartridge [3]. Analytes are eluted from the cartridge using acetonitrile/toluene. A 1-gram volume equivalent is taken from the eluant and internal standard(s) is added. The extract is brought to near dryness and solvent exchanged into ethyl acetate for GC/MSD analysis.
trace quantities (< 0.02 ppm) to 0.48 ppm. In most cases, the trace quantities were not found by the GC/MSD standard quantitation using 3 qualifier ion identification. As a verification of the methodology, a second sample of strawberries and tomatoes were each fortified with six analytes at the 0.1 ppm level, together with the Aldrin SS/IS. The analyses results for these spiked samples are shown in Table 3. There are excellent recoveries for most analytes. Responses for two analytes in tomato, atrazine, and permethrin were higher than expected and could be due to matrix enhancement during injection. Time constraints did not allow for further investigation. Duplicate results are also shown in Table 3. The chlorothalonil in tomato at 0.08 ppm compares favorably with the 0.09 ppm found in the first sample. The same is true for captan at 0.47 ppm in the duplicate versus 0.48 ppm in the original sample. The GC/MSD system was calibrated for more than 50 compounds. Using DRS, the analyst can get a verification of the presence of those 50 compounds together with an automated expert second opinion of other compounds that may be present. This second opinion is in two distinct parts. First, the deconvolution of the complex TIC with subsequent matching of clean spectra to a database is provided. Second, the matching of these clean spectra to an independent database, in this case the NIST05a library of > 163,000 compounds.
Celery Gr pepper Strawberry Tomato
Results
Six commodities - apples, lettuce, carrots, celery, green peppers, strawberries, and tomatoes - were purchased at local supermarkets. They were extracted and analyzed using the GC/MSD conditions described earlier. Aldrin was added to each during the extraction process and acts as both a surrogate standard (SS) and an internal standard (IS). The datafiles were processed using Agilents Deconvolution Reporting Software. DRS automatically combines the results from the MSD ChemStation Quantitation with AMDIS Deconvolution and NIST Search into an easy-to-read report. The results are shown in Table 2. Each of the samples showed at least one residue, ranging from
Table 2. Market Basket Commodity Results Commodity Apple Lettuce Compound Table R.T. Methamidopho Acephate Diphenylamine Chlorothalonil Carbaryl Metalaxyl Malathion Isodrin Thiabendazole Captan Phosmet Permethrin Cyfluthrin Cypermethrin Fenvalerate Addional DRSonly compounds
* First R.T. of multiple isomers.
Carrot
5.655 7.690 10.516 14.784 16.806 17.337 18.800 20.031 20.939 21.227 28.504 31.369* 32.218* 32.690* 34.271*
t t 0.02 0.02
0.04 0.23 t t t t 0.09
0.02 0.48 0.03 t t t t 8 0 1 0.03 t
13
10
t = Trace quantity
Table 3.
Spiked and Duplicate Commodity Results, ppm Spikes Strawberry 0.15 0.11 0.11 0.13 0.09 0.14 0.15 Tomato 0.25 0.11 0.15 0.19 0.13 0.26 0.25 Duplicates 0.08
Compound Table R.T. Atrazine 13.159 Lindane 13.461 Carbaryl 16.806 Linuron 18.187 Parathion 19.275 Permethrin I 31.369 Permethrin II 31.550 Chlorthalonil Captan 14.784 21.227
0.47
A portion of the DRS Report for celery is shown in Figure 1. Chlorothalonil was found at 14.823 minutes at 0.02 ppm. AMDIS verified chlorothalonil with a 97 match eluting only 2.7 seconds from its expected RTL time. NIST search further verified chlorothalonil with a 93 match, as the third hit out of the top 100 hits. Aldrin and permethrin are similarly verified with permethrin below the normal reporting limit. Malathion was found by AMDIS and verified by NIST. It was not found by ChemStation because one of three qualifier ions was out of range. AMDIS mitigates this problem because deconvolved spectra are cleaned of interferences and full spectrum matching is used. It would be nearly impossible to identify the analytes of interest in the presence of > 650 individual components in the celery without deconvolution. The TIC for celery is shown in Figure 2.
For all of the commodities tested, DRS verified the presence of all the calibrated peaks found by the ChemStation. Most laboratories only calibrate a fixed number of compounds, say 50-100, as it is not practical to calibrate for all 927. At the same time, these laboratories are interested in identifying other compounds that may be present in samples. Numerous uncalibrated compounds were identified using the 927 compound DRS database. The number of these additional compounds is shown on the last line in Table 2, excluding phthalates, cresols, and sulfur. When these are important to the laboratory, the GC/MSD system can, of course, be calibrated using additional standards. If an estimate of the amount is needed, or if a standard is unavailable, an average response factor can be used. The DRS database provides an average response factor for all 927 compounds. In contrast to the above methodology is the use of multiple element specific detectors such as ECD, NPD, and DFPD. Most of the analytes could be run on two of these specific detectors, but that would take twice the time. Also, it is very difficult to analyze for 927 compounds using a non-MS method due to peak overlaps. A second choice is to run each sample on unlike columns of differing polarities to the same type detector. Both of these approaches, using ECD, NPD, and/or DFPD, involve more sample handling/tracking, usually twice the number of runs, and still only have retention time as an identifier.
MSD Deconvolution Report Sample Name: celery Data File: C:\msdchem\1\Data\sjm03.D Date/Time: 02:30 PM Friday, Jan 19 2007 The NIST library was searched for the components that were found in the AMDIS target library.
R.T. 14.823 18.5992 18.7799 31.3134 Figure 1. CAS # 1897456 309002 121755 52645531 Compound Name Chlorothalonil Aldrin Malathion Permethrin I Agilent ChemStation Amount (ppm) 0.02 1 0.03 Match 97 99 59 71 AMDIS R.T. Diff Sec 2.7 4.3 -1.2 -3.3 NIST Reverse Match Hit Num. 91 93 47 53 3 1 1 5
DRS report for celery.
Aldrin
Chlorothalonil Permethrin Malathion
10
12
14
16
18
20
22
24
26
28
30
32
Figure 2.
Market basket celery total ion chromatogram.
Conclusions
Analyzing a complex matrix can be accomplished using multiple specific detectors, but a significant time savings is realized using GC/MSD. Many more analytes can be determined simultaneously using mass spectra for confirmation. Using an available Retention Time Locked database reduces methods development and speeds data comparison among labs. Deconvolution Reporting Software adds a second expert opinion. Positives found by normal quantitation are confirmed, and the analyst is directed to further target compound hits to verify. GC/MSD with DRS provides the fastest methodology to the fewest number of false positives/negatives with the greatest confidence in the results.
2. Fast and Easy Multiresidue Method Employing Acetonitrile Extraction/Partitioning and Dispersive Solid-Phase Extraction for the Determination of Pesticide Residues in Produce, M. Anastassiades, S. J. Lehotay, D. Stajnbaher, F. J. Schenck. (2003) J.AOAC Int. 86, 412-431. 3. Analytical Methods for Residual Compositional Substances of Agricultural Chemicals, Feed Additives, and Veterinary Drugs in Food, Japan Department of Food Safety, Ministry of Health, Labour and Welfare.

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2007 Printed in the USA May 14, 2007 5989-6677EN
References
1. Comprehensive Pesticide Screening by GC/MSD using Deconvolution Reporting Software, Philip L. Wylie, Michael J. Szelewski, and Chin-Kai Meng, Agilent Technologies Pub # 5989-1157EN.
Rapid Analysis of Herbicides by Rapid Resolution LC with Online Trace Enrichment Application
Environmental
Author
Michael Woodman Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA
Abstract
Environmental and Food Safety agencies are constantly updating methods to improve detection limits and to resolve interfering compounds. One particular method, EPA 555, is used for the analysis of chlorinated phenoxy acid herbicides in drinking water. A mandated trace enrichment step significantly impacts the ease of use and reliability of the method. The method uses 5-m analysis columns and online trace enrichment. The variation here uses small ZORBAX 3.5- and 1.8-m RRHT columns and an autoSPE (Solid Phase Extraction) cartridge with an automated switching valve mounted in the column compartment. Combined with sample introduction via direct injection to the autoSPE cartridge, instead of the loading pump specified in the EPA method, we dramatically reduce the overall analysis time and virtually eliminate the potential of sample cross-contamination.
disposable SPE cartridges, which require an elution step into a vial prior to analysis, online SPE assures 100% sample transfer to the analysis column and dramatically increases sensitivity by increasing the analyte mass delivered to the column. In EPA Method 555, 20 mL of drinking water is loaded through a pump to an SPE cartridge mounted on a high-pressure switching valve on the HPLC system. Because few, if any, autosamplers can inject this large volume, the sample must be pumped onto the cartridge. Contamination of the loading pump with prior samples is always a concern, and adequate flushing and blank runs become an important part of the overall method procedure. To reduce the sampling volume sufficient for available automatic preparative samplers, without losing sensitivity in the method, it is necessary to reduce the analysis column size while preserving resolving power. Ancillary benefits of using smaller columns generally include reduced analysis time and solvent consumption, and greater compatibility with ionization sources in mass spectrometers. If the ratios of their column length to particle size are equal, columns are considered to have equal resolving power. Significant reductions in column volume can be made by reducing the length and/or internal diameter of the column. In the latter case, the flow rate would normally be reduced as in Equation 1.
Introduction
Trace analyte detection in relatively clean matrices is an excellent application for online SPE procedures. Compared to manually loading samples with
Flowcol. 1
Diam.column2 Diam.column1
= Flowcol. 2
(eq. 1)
little improvement is observed. In the subsequent examples, we will see the results associated with the calculations discussed above. Experimental Conditions See figure 1 for configuration.
System Agilent 1200 Series Rapid Resolution LC consisting of: G1379B micro degasser G1312B binary pump SL G1312A binary pump with solvent selection valve option, or G1354A quaternary pump G1367C HiP ALS autosampler SL, and G2258A Dual Loop Prep autosampler 5 ml G1316B Thermostatted column compartment SL with 6- or 10-port 2-position switching valve G1315C UV/VIS diode array detector (DAD) SL, flow cell as indicated in individual chromatograms ChemStation 32-bit version B.02.01 Columns Agilent ZORBAX SB-C18, 4.6 250 mm, 5 m Agilent ZORBAX SB-C18, 3.0 150 mm, 3.5 m Agilent ZORBAX SB-C18, 2.1 80 mm, 1.8 m Agilent ZORBAX SB-Aq, 4.6 12.5 mm, 5 m Mobile phase conditions Organic solvent: Aqueous solvent: Gradient conditions Gradient slope: 7.8 or 2.3% per column volume, as indicated. See individual chromatograms for flow rate and time Acetonitrile 25 mm phosphoric acid in Milli-Q water
The combined effect of reduced length and diameter contributes to a reduction in solvent consumption. We normally scale the injection mass to the size of the column and a proportional injection volume would be calculated from the ratio of the void volumes of the two columns, multiplied by the injection volume on the original column, as in Equation 2 below.
Volumecolumn2 Volumecolumn1
Inj. vol.col. 1
= Inj. vol.col. 2
(eq. 2)
Short columns packed with small particle sizes are typically operated at high linear velocities. The increase in elution speed will decrease absolute peak width and may require the user to adjust data acquisition rates and reduce signal filtering parameters. This will ensure that the chromatographic separation is accurately recorded in the acquisition data file. For gradient elution separations, where the mobile phase composition increases through the initial part of the analysis until the analytes of interest have been eluted from the column, successful method conversion to smaller columns requires that the gradient slope be preserved. We can express the gradient slope as in Equation 3.
% Gradient slope =
(End% Start%) #Column volumes
Sample
(eq. 3)
Note that the use of % change per column volume rather than % change per minute frees the user to control gradient slope by altering gradient time and/or gradient flow rate. A large value for gradient slope yields very fast gradients with minimal resolution, while lower gradient slopes produce higher resolution at the expense of increased solvent consumption and somewhat reduced sensitivity. Longer analysis time may also result unless the gradient slope is reduced by increasing the flow rate, within acceptable operating pressure ranges, rather than by increasing the gradient time. Resolution increases with shallow gradients because the effective capacity factor, k*, is increased. Much like in isocratic separations, where the capacity term is called k', a higher value directly increases resolution. The effect is quite dramatic up to a k value of about 510, after which
2
EPA 555 Group A chlorinated phenoxy acid herbicides (picloram, chloramben, dicamba, bentazon, 2,4-D, dichlorprop, 2,4,5-TP, acifluorfen), 100 g/mL in methanol or diluted to 20 ng/L (20 ppb) in reagent water acidified with 25 mm phosphoric acid.
Results
The separation was initially performed via direct injection of concentrated standard on a 4.6 250 mm, 5-m ZORBAX SB-C18 column, thermostatted to 25 C, using conditions referenced in US EPA method 555 (Figure 2). The described trace enrichment procedure using pump A as the loading pump was performed (Figure 3). The method was then scaled in flow and time for exact translation to a 3.0 150 mm 3.5-m column (Figure 4) using 5-mL trace enrichment injection. Finally, a 2.1 80 mm 1.8-m configuration (50-mm plus 30-mm columns in series) is used to demonstrate the feasibility of this separation under conditions for trace enrichment requiring less than 1.5-mL injection. (Figure 5)
Load/Wash position
Elute/Analyze position
2 Position/6 Port valve
Figure 1.
Trace enrichment autoSPE scheme.
Figure 1 shows the schematic placement of modules and columns in the system. The A pump is the loading pump in case of volumes exceeding the 5-mL capacity of the G2258A Dual Loop Autosampler, thus pump A uses one line for sample and a second line for the aqueous eluent, 25 mm phosphoric acid. If direct injection from the autosampler is used, pump A is delivering 25 mm phosphoric acid. If the A pump is fitted with a degasser, the sampling line should bypass the degasser module to minimize contamination with sample solutions. To conduct sampling through the A pump, the valve should be in position B while the new sample is flushed through the A pump. Then switch the valve to the A position and load the required 20 mL sample volume. The analysis
begins when the valve is returned to the B position, at which time the sampling line on the A pump would be flushed with reagent water or the next sample, as appropriate. Figures 2 and 3 show the standard separation by direct injection and pumped trace enrichment, respectively. With column regeneration steps, this results in a total analysis time of 60 minutes. Translation of the gradient to the 3.0- 150-mm column requires a reduction in flow rate, due to the smaller diameter, and a reduction in gradient time because of the shorter column length. The resulting analysis is reduced from 60 to 36 minutes and solvent consumption is proportionately reduced from 60 mL to 15.5 mL.
100
80
12.679 - Picloram
14.436 - Chloramben
mAU 120
Conditions EPA Method 555 with ZORBAX SB-C18 columns and fast DAD detector ZORBAX SB-C18 4.6 mm 250 mm, 5 m Column temp: 25 C Gradient: 25 mM H3PO4, ACN, 10% to 90% ACN in 30 min Gradient slope: 7.8% ACN/column volume Analysis flow rate: 1 mL/min Group A compounds: 1 L of 100 g/mL Total analysis time: 60 min Detection: UV 230 nm, 10-mm 13-L flow cell, filter 2 seconds (default)
18.895 - Bentazon 17.845 - Dicamba 24.678 - Acifluorfen
60
21.104 - Dichlorprop
19.480 - 2,4-D
40
13.303
0 10 12 14 16 18 20 22 24 26 28 min
Figure 2.
Gradient separation of herbicides on 4.6 mm 250 mm, 5 m ZORBAX SB-C18.
17.653
20.238
29.609
20
23.095 - 2,4,5-TP
14.380 - Chloramben
12.557 - Picloram
mAU 350 300 250 200 150 100
18.871 - Bentazon
19.414 - 2,4-D
21.063 - Dichlorprop
17.779 - Dicamba
23.050 - 2,4,5-TP
24.667 - Acifluorfen
13.194
17.607
18.439
0 10 12 14 16 18 20 22 24 26 28 min
Conditions EPA Method 555 with ZORBAX SB-C18 columns and fast DAD detector ZORBAX SB-C18 4.6 mm 250 mm, 5 m Column temp: 25 C Gradient: 25-mM H3PO4, ACN, 10% to 90% ACN in 30 min Gradient slope: 7.8% ACN/column volume Analysis flow rate: 1 mL/min Group A compounds: 20 mL of 20 g/L trace enrichment Total analysis time: 60 min Detection: UV 230 nm, 10-mm 13-L flow cell, filter 2 seconds (default) Figure 3. Trace enrichment (20 mL) of 20-ppb solution on 4.6 250 mm 5-m ZORBAX SB-C18.
4.094
120 100 80 60 40 20 0 2
5.118
mAU
Conditions: EPA Method 555 with ZORBAX SB-C18 columns and fast DAD detector ZORBAX SB-C18 3 mm 150 mm, 3.5 m Column temp: 25 C Gradient: 25 mm H3PO4, ACN, 10% to 90% ACN in 18 min Gradient slope: 7.8% ACN/column volume Analysis flow rate: 0.43 mL/min Group A compounds: 5 mL of 20 g/L (20 ppb) Detection: UV 230 nm, 3-mm 2-L flow cell, filter 0.2 seconds Total analysis time: 36 min
7.870
7.073
8.221
9.211
21.720
10.352
27.714
50
4.462
6.642
10
11.555
12
14
14.041 min
Figure 4.
Trace enrichment (5 mL)of 20-ppb solution on 3.0 150 mm, 3.5-m ZORBAX SB-C18.
The last peak in Figure 4 is missing due to a valve timing error that was not detected until sometime after the lab work was completed. Peak 8 was not eluted from the trace enrichment column before
4
the valve switched offline for regeneration and equilibration. Note the baseline shift that occurs after peak 7, not seen in other autoSPE examples.
29.595 29.786
1.532
700 600 500 400 300 200
1.672
mAU
Conditions EPA Method 555 with ZORBAX SB-C18 columns and fast DAD detector ZORBAX SB-C18, 2.1 mm 80 mm (50mm + 30mm in series), 1.8 m Column temp: 50 C Gradient: 25-mM H3PO4/ACN, 10% to 90% ACN in 2.7 min 7.8% ACN/column volume Analysis flow rate: 0.72 mL/min Detection: UV 230 nm, 3-mm 2-L flow cell, filter 0.2 seconds Sample: Aged 1 L 100 g/mL Total analysis time: 6 min
2.048 2.723 2.626 2.830
2.161 2.206
2.385
1.643
1.121
0 1.25 1.5 1.75 2 2.25 2.5 2.75 3 3.25 min
Figure 5.
High-speed gradient separation of herbicides on 2.1 80 mm, 1.8-m ZORBAX SB-C18.
In Figure 5 we see the combination of highest speed and resolution, using the full capability of the 1200 Rapid Resolution LC. Operating pressure was, at the maximum point, about 520 bar. We maintain comparable resolution to the original 4.6 250 mm, 5-m method, a 60-minute run time, with an analysis time of only 6 minutes.
Conclusion
As is the case for many existing methods, it is both possible and practical to modernize this method to improve throughput and overall performance. Here we see the potential for a 10-fold increase in analysis speed and elimination of the loading pump scheme found in the original method. Approximately 1.3 mL of sample solution, injected to the autoSPE setup using the 2.1 80 mm configuration, is all that is needed to replace the 20-mL injection previously loaded through the pump. This approach can greatly improve productivity and ensure minimal sample cross-contamination.

2.310
3.025
100
2.577
Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2007 Printed in the USA March 30, 2007 5989-5176EN
Improving Productivity and Extending Column Life with Backflush
Application Brief
Chin-Kai Meng
All Industries
A previous application note [1] has shown that multiple GC signals and MS signals can be acquired from a single sample injection. When a 3-way splitter is connected to the end of a column, column effluent can be directed proportionally to two GC detectors as well as the MSD. This multi-signal configuration provides full-scan data for library searching, SIM data for quantitation, and element selective detector data for excellent selectivity and sensitivity from complex matrices. The system used in this study consists of a 7683ALS, a 7890A GC with split/splitless inlet, 3-way splitter, ECD, dual flame photometric detector (DFPD), and a 5975C MSD. Figure 1 shows four chromatograms from a single injection of a milk extract. The synchronous SIM/scan feature of the 5975C MSD provides data useful for both screening (full scan data) and quantitation (SIM data). DFPD provides both P and S signals without the need to switch light filters. Noticeably in the full scan TIC in Figure 1, a significant number of matrix peaks were observed after 32 minutes. It is not uncommon to add a bake-out oven ramp to clean the column after analyzing complex samples. The bake-out period is used to quickly push the late eluters out of the column to be ready for the next injection. Therefore, it is common to use a higher oven temperature than required for the analysis and an extended bake-out period at the end of a normal
Full scan TIC
Highlights
Backflush a simple technique to remove high boilers from the column faster and at a lower column temperature to cut down analysis time and increase column lifetime. The milk extract example shows that a 7-minute 280 C backflush cleaned the column as well as a 33-minute 320 C bake-out. The cycle time was reduced by more than 30%. Using backflush, excess column bleed and heavy residues will not be introduced into the MSD, thus reducing ion source contamination.
SIM
ECD
DFPD(P)
10
15
20
25
30
35
40
Figure 1.
Four chromatograms collected simultaneously from a single injection of a milk extract.
over program to clean out the column, which adds to the cycle time and shortens the column lifetime. Adding the bake-out period to the milk extract analysis, additional matrix peaks were observed even up to 72 minutes, while target compounds already eluted before 42 minutes. This means that 30 minutes were lost in productivity for each injection. Backflush [2] is a simple technique to drastically decrease the cycle time by reversing the column flow to push the late eluters out of the inlet end of the column. Late eluters stay near the front of the column until the oven temperature is high enough to move them through the column. When the column flow is reversed before the late eluters start to move down the column, these late eluters will take less time and at a lower oven temperature to exit the inlet end of the column. There are many benefits in using backflush: Cycle time is reduced (no bake-out period, cooling down from a lower oven temperature) Column bleed is reduced (no high-temperature bake-out needed), resulting longer column life Ghost peaks are eliminated (no high boilers carryover into subsequent runs) Contamination that goes into the detector is minimized, which is especially valuable for the MSD (less ion source cleaning) of 320 C. A column effluent splitter or QuickSwap is required to do the backflush.
References
1. Chin-Kai Meng and Bruce Quimby, Identifying Pesticides with Full Scan, SIM, ECD, and FPD from a Single Injection, Agilent Application Note, 5989-3299EN, July 2005. 2. Matthew Klee, Simplified Backflush Using Agilent 6890 GC Post Run Command, Agilent Application Note, 5989-5111EN, June 2006.
Acknowledgement
Milk extract is courtesy of Dr. Steven Lehotay from USDA Agricultural Research Service in Wyndmoor, Pennsylvania, USA.
Figure 2 shows three total ion chromatograms from the Agilent 7890A GC/ 5975C MSD. The top chromatogram is a milk extract analysis with all the target compounds eluted before 42 minutes (over program goes to 280 C). However, an additional 33-minute bake-out period at 320 C was needed to move the high boilers out of the column. This bake-out period was almost as long as the required time to elute all target compounds. The middle chromatogram is the same milk extract analysis stopped at 42 minutes with a 7-minute backflush post-run at 280 C added to the analysis. The bottom chromatogram is a blank run after the backflushing was completed. The blank run shows that the column was very clean after backflushing. The example shows that a 7-minute backflush cleaned the column as well as a 33-minute bake-out. The milk extract example in Figure 2 illustrates the backflush technique in reducing cycle time and column bleed. The cycle time was reduced by more than 30% and the column was kept at 280 C, without going to the bake-out temperature

It took an additional 33 min and heating the column to 320 C to remove these high boilers
Run stopped at 42 min and backflushed at 280 C for 7 mins Blank run after backflushing showing the column was clean
5 10 15 20 25 30 35 40 45 50 55 60 65 70 min
Figure 2.
Three total ion chromatograms comparing the results with and without backflush.
Using RTL and 3-Way Splitter to Identify Unknown in Strawberry Extract Application Brief
Chin-Kai Meng
Food Safety and Environmental
Fruit and vegetable extracts are usually very complex to analyze. It is common to use the very selective GC detectors, for example NPD, ECD, and FPD, to look for trace pesticide residues in the extracts. Mass spectrometry is most often used to confirm the hits from GC detectors. A previous application note [1] describes a GC/MS system with a three-way splitter added to the end of the column. The column effluent could be split three ways to two GC detectors and the MSD. The splitter system is therefore capable of providing up to four signals (two GC signals, SIM, and full-scan chromatograms) from a single injection. The combination of element selective detectors, SIM/scan, and Deconvolution Reporting Software (DRS) makes a very powerful pesticide analysis system [2]. The trade-off is the decrease of analyte concentration in any detector due to the flow splitting at the end of the column. The system used for this study consists of an Agilent 7890A GC with split/splitless inlet, a three-way splitter, ECD, dual flame photometric detector (DFPD), and 5975C MSD. Figure 1 shows chromatograms from 2 separate injections (each injection provides two GC signals) of the same strawberry extract without any hardware or filter changes. All of the target compounds were found and confirmed by DRS, GC, and MS signals except the unknown peak at about 41 minutes. The peak shows responses from ECD, DFPD(S) and DFPD(P). However, no peak was observed in the MS full-scan signal. This makes it difficult to confirm the unknown peak using the full-scan TIC. Since the analysis was retention time locked, it is therefore possible to find potential matches by examining the RTL pesticide database (part number G1672AA). The unknown compound, containing electron-capturing atoms (for example, Cl or O), P, and S atoms, would have a target retention time inside the
Highlights
Splitter+an inert, easy-to-use capillary flow technology that splits column effluent to multiple detectors (for example, MSD, DFPD, and ECD). The splitter configuration provides a comprehensive screening and quantitative system. By combing RTL, element-selective detector chromatograms, and the RTL pesticide database, a trace level pesticide residue was identified without the full-scan mass spectrum.
TIC
ECD
DFPD (S)
DFPD (P)
5.00
10.00
15.00
20.00
25.00
30.00
35.00
40.00
Figure 1.
Unknown compound detected by GC signals not found in strawberry extract TIC.
Table 1.
Compound List Extracted from the RTLPest3.tab File CAS 117337196 191242 3383968 60044260 83794 Mol form C15H15CIFN3O3S2 C22H12 C16H20O6P2S3 C12H4Br6 C23H22O6 Mol wt 403.9 276.3 466.5 627.6 394.4 R.T. 39.10 39.13 40.74 40.93 41.70 Target Ion 403 276 466 308 192 Q1 56 277 125 468 191 Q2 405 138 93 148 394 Q3 232 275 109 154 177
Name Fluthiacet-methyl Benzo[g,h,i]perylene Temephos PBB 169 hexabrombiphenyl Rotenone
41 \ 0.5-minute window (that is, 40.5 to 41.5 min) in the database, if it is included in the database. Table 1 is a portion of the RTLPest3.tab file1 opened in Microsoft Excel. The compound temephos at locked retention time 40.74 min meets all the criteria for the unknown peak. To further confirm peak identity, extracted ion chromatograms (EICs) of the four major ions of temephos were plotted. Figure 2 shows EICs of target ion and three qualifiers (ions 466, 125, 93, and 109 from Table 1) of temephos. Although the ion intensities were weak, the noticeable presence of all four ions at 40.9 min helped to confirm that the unknown peak was temephos.
1. The RTLPest3.tab file is created in the C:\Database directory while executing the Tools\List Screen Database command (in MSD Enhanced Data Analysis software) and selecting the RTLPest3.scd from the C:\Database directory.
200 100 0
Ion 466
200 100 0
Ion 125
200 100 0
Ion 93
200 100 0
Ion 109
36
37
38
39
40
41
42
43
44
Figure 2.
EICs of target ion 466 (temephos) and three qualifier ions.
References
1. Chin-Kai Meng and Bruce Quimby, Identifying Pesticides with Full Scan, SIM, ECD, and FPD from a Single Injection, Application Note, 5989-3299, July 2005. 2. Mike Szelewski and Bruce Quimby, New Tools for Rapid Pesticide Analysis in High Matrix Samples, Application Note, 5989-1716, October 2004.
Acknowledgement
Strawberry extract is courtesy of Dr. Steven Lehotay from USDA Agricultural Research Service in Wyndmoor, Pennsylvania, USA.

Copyright 2006 Agilent Technologies All Rights Reserved. Reproduction, adaptation, or translation without prior written permission is prohibited, except as allowed under the copyright laws. Printed in the USA December 13, 2006 5989-6007EN
Automated Screening of 600 Pesticides in Food by LC/TOF MS Using a Molecular-Feature Database Search Application
Food Safety
Authors
E. Michael Thurman and Imma Ferrer Pesticide Residue Research Group University of Almera 04120 Almera, Spain Jerry A. Zweigenbaum Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA
the compounds to <0.5 mg/kg for 95% of the compounds. Strengths of the MF algorithm include rapid screening of hundreds of compounds at sensitive levels in minutes compared to a manual approach of hours to days.
Introduction
Contamination of foodstuffs with pesticides always means a risk to the consumer. Fetuses, infants, and children, a particularly sensitive group, are currently protected by a limit of 0.01 mg/kg in baby food by European Union legislation [1]. Maximum residue limits (MRLs) should be set at the lowest possible level with the aim of setting products on the markets without any measurable residues. However, this is often difficult because of the use of fungicides for transport and storage of fruit and vegetables, as well as the use of insecticides for crop protection. The analysis of pesticides in baby food has involved both GC/MS and LC/MS methods. Classically, GC/MS has been used to look for volatile pesticides, especially the chlorinated insecticides and herbicides, and LC/MS has been used to monitor the more polar insecticides, herbicides, and fungicides. GC/MS methods rely on both full scan and selected ion monitoring. Recently the use of reverse search methods for GC/MS has made it possible to search large NIST pesticide libraries in minutes [2] and has made screening quite simple for pesticides amenable to GC/MS methods.
Abstract
Searching a database using a molecular feature (MF) algorithm was developed for the screening of 600 pesticides and degradates in extracts of food by liquid chromatography time-of-flight mass spectrometry in positive ion mode with full-scan accurate mass spectra. The database search works by compiling the accurate mass of the ions detected and identified as real compound chromatographic peaks without ion extraction and compares them with the monoisotopic exact masses of the compounds in the database. The screening criteria consisted of 5 ppm accurate mass window, 0.2 minute retention time window, and a minimum area count of 1,000 counts (signal-to-noise ratio of ~10:1). The limit of detection and retention time was determined for 100 of the 600 compounds and varied from <0.01 mg/kg for 34% of
Unfortunately, similar reverse-search methods have not been available for LC/MS for two reasons. First, the single quadrupole and triple quadrupole methods do not operate in full scan mode for pesticide screening because of a lack of sensitivity. Second, standardization and reproducibility of CID fragmentation energy for broad usage has been difficult, making common spectral libraries unavailable. Recently, the use of LC/TOF MS has shown that it has the capability and sensitivity to obtain full scan spectra with accurate masses of less than 1 ppm [3] suitable for database searching [4].
to 100% B in 30 minutes at a flow rate of 0.6 mL/min Agilent 6210 LC/MS TOF with dual spray electrospray source Positive ESI, Capillary 4000 V Nebulizer 40 psig, drying gas 9 L/min, gas temp 300 C Fragmentor 190 V, skimmer 60 V, Oct DC1 37.5 V, OCT RF V 250 V Reference Masses: mass range (m/z) 121.0509 and 922.0098, resolution: 9500 500 at m/z 922.0098, m/z 50 to 1,000, Reference A Sprayer 2 is constant flow rate during the run
Experimental
Vegetable and Fruit Extraction (QuEChERS) QuEChERS is the acronym for the extractionmethod, which stands for quick, easy, cheap, effective, rugged, and safe. It is a method that is widely receiving acceptance for rapid extraction of pesticides in food [5, 6]. Weigh 15 g of a previously homogenized sample into a 40-mL Teflon centrifuge tube. Add 15 mL of acetonitrile (containing 1% acetic acid), add 6 g of anhydrous MgSO4 and 2.5 g of NaAc63H2O (sodium acetate trihydrate) and shake the sample vigorously for 1 min by using a vortex mixer at maximum speed or hand shaking. Centrifuge for 3 minutes at 3,700 rpm. Take 5 mL of supernatant into a 15-mL tube, add 250 mg of PSA adsorbent and 750 mg of MgSO4, and vortex and shake for 20 s. Then centrifuge again for 3 min at 3700 rpm. Transfer 1.0 mL into LC/MS vial. The 1.0 mL supernatant is then evaporated to dryness and brought back up in 8/92% methanol/water for LC/MSD TOF and ion trap analysis. Direct analysis of the fruit and vegetable extracts were analyzed by injecting 50 L. Nonfortified samples were analyzed directly at this same point by LC/MS TOF. LC/MS TOF Methods LC pumps: Agilent 1100 binary pumps, injection volume 50 L with standard Agilent 1100 ALS Column: ZORBAX Eclipse XDB 4.6 150 mm C-8, 5-micron (p/n 993967-906) Mobile Phase A = 0.1% formic acid in water, and B = acetonitrile, gradient began with 5 minutes isocratic at 10% B followed by a linear gradient

Molecular Feature Database Search Theoretical monoisotopic exact masses of compounds based on their molecular formula were calculated using an Excel spreadsheet and put into csv (comma separated values) format for use by the TOF software of the Agilent LC/TOF MS system for 600 pesticides known to ionize by positive ion electrospray. The csv file is created in the TOF software by an Excel spreadsheet tool called Formula DB Generator. The csv file is then searched automatically by the LC/TOF MS instrument at the completion of the sample run and a report generated on compounds that were found in the database. Search criteria include ppm mass tolerance (5 ppm), retention time window (0.2 minutes) if available, and minimum peak height count, which is called the compound threshold (1,000 counts or a signal-to-noise ratio of ~10:1 or 0.06% relative volume), and adducts and neutral fragment losses. The search routine is called a molecular feature extractor and is software recently available on the Agilent LC/TOF MS (November 2005). The molecular feature extractor finds all ions in an LC/MS TOF data file representing ions of real compounds in the sample analyzed. Noise and other extraneous ions are excluded. The resulting list of ions are then searched against the csv database using the chosen criteria and ions found are then tabulated from the full scan spectrum and checked for accuracy and retention time against the database. The molecular feature approach is more suited to large libraries because of the ease of operation and the quickness for which the search is done. Thus,
each ion of interest in the database is not extracted from the sample file as in a reverse search. This procedure requires much more time with LC/TOF MS data files collected in profile mode. From this point, confirmation is carried out manually by checking the positive screens for retention time match and fragment ions (if present). The sample may be reanalyzed at a higher fragmentor voltage to check for fragment ions and confirmed by authentic standard analysis. Limit of Detection The limit of detection (LOD) was determined in several matrices, including spiked food samples and solvent extracts for 100 compounds (Table 1). These compounds consist of the major classes of pesticides that are commonly used in the United States and Europe for treating crops of fruits and vegetables. The limit of detection was based on an accurate mass of less than 3 ppm and the appearance of the correct accurate mass of A+1 and A+2 isotopic signatures. The LOD was determined at various levels for pesticide work, including the EU regulation of 0.01 mg/kg for baby food, and 0.05, 0.1, and 0.5 mg/kg for various food levels, depending on pesticide and crop type. The LOD was equal to or less than 0.01 ppm for 33 compounds and less than or equal to 0.05 for 60 compounds (60%). The 0.05 mg/kg LOD is also a critical one for food monitoring of banned substances or controlled compounds. The LOD for 95% of the compounds was equal to or less than 0.2 mg/kg for food. Only six compounds were found
to be insensitive with LODs of 0.5 mg/kg for food. The insensitive compounds were promecarb and aldicarb, which are two carbamate pesticides that fragment easily in the electrospray source and give low abundance for the MH+ ion. Likewise, malathion oxon and dimethoate are two organophosphate pesticides that fragment easily. Thus, these compounds could be more sensitively detected by the use of the more abundant fragment ion rather than the MH+. For example, Figure 1 shows the mass spectrum of dimethoate. The MH+ ion is not the major fragment ion of the spectrum, and in fact has an intensity about three to four times less than the m/z 124.9819 ion. Furthermore, it must be taken into consideration that the LOD is affected by the matrix for two reasons. One is suppression of ionization and the second is interfering ions of nearly identical mass. Suppression of ionization has been tested in our previous work for some of these pesticides in food [7], including pepper, broccoli and tomato, melon, orange, and lemon, so we have experience on which matrices are most difficult. For example, Figure 2 shows the matrix chromatogram for pesticide-free pepper, which gives a complicated chromatogram. The MF database search identified ~3,000 compounds of which none were pesticides. These peaks contained signal-to-noise ratio of 10:1 or greater and present an extremely difficult matrix for which to find ions, especially at trace levels. Spiked food extracts of these difficult matrices were used to establish LOD for the compounds shown in Table 1. An example printout is shown in Figure 3. The report contains formula, compound,
SH P H3CO OCH3 O C5H13NO3PS2+ Exact mass: 230.0069 S H N CH3
Figure 1.
Dimethoate mass spectrum showing low intensity of MH+ ion and importance of using characteristic fragment ions to lower LODs on some compounds of low intensity, specifically, m/z 124.9819 ion.
Figure 2.
Blank pepper sample showing complexity of the sample ~3000 accurate mass peaks were detected in this sample at signal to noise of 10:1 or greater.
accurate mass of the neutral molecule, error in mDa and error in ppm, retention time error in minutes, and a description (specifically, fungicide). The mass spectrum of the MH+ and isotope signature of the compound are also shown, which is useful for a quick check and partial confirmation of the formula, especially since most pesticides show an interesting A+2 ion from a halogen or sulfur atom. A+2 Ions and Empirical-Formula Confirmation Because the MF database search is a screening program, a formula is not unequivocally confirmed at this point with a mass accuracy window of 5-ppm. The molecular formula (not molecular identification) may be confirmed by the A+2 isotopic signature of the compound. For example, 70% of the 600-compound database contains either S, Cl, or
Br, which give the isotopic cluster of the A+2 ions. The accurate mass of the isotope along with the intensity profile can then be used as a first level confirmation of the empirical formula. Thus, this adds a great deal of confidence to the screening data of the accurate mass but of course does not yet meet the standards for identification of the molecule, a topic discussed later. For example, let us examine the isotopic signature of imazalil in a pear extract (Figure 4). The measured mass of the MH+ was 297.0564 and the chlorine 37 isotope was 299.0533. Thus the difference in mass is 1.997 mass units, which is the mass defect of a chlorine 37 atom relative to the chlorine 35 atom that has been replaced. Furthermore, the intensity of the A+2 peak is about 2/3 of the A peak, which is consistent with two chlorine atoms in the molecule. This is further established by the presence of an A+4 peak at 301.0503. Thus, these
Limits of Detection for Pesticides in Food Samples with Retention Time and Accurate Mass Accurate mass Compound Ret time Formula molecule Atrazine 21.1 C8H14N5Cl 215.0938 Azoxystrobin 24.0 C22H17N3O5 403.1168 Benalaxyl 26.8 C20H23NO3 325.1678 Buprofezin 27.2 C16H23N3OS 305.1562 Cyanazine 22.0 C9H13N6Cl 240.0890 Diazinon 27.6 C12H21N2O3PS 304.1010 Difenconazole 26.4 C19H17Cl2N3O3 405.0647 Isomer 26.6 C19H17Cl2N3O3 405.0647 Difenoxuron 21.3 C16H18N2O3 286.1317 Dimethomorph 22.2 C21H22NO4Cl 387.1237 Fenamiphos 23.9 C13H22NO3PS 303.1058 Imazalil 18.0 C14H14N2OCl2 296.0483 Imazapyr 20.0 C13H15N3O3 261.1113 Imazaquin 20.0 C17H17N3O3 311.1270 Irgarol 21.2 C11H19N5S 253.1361 Irgarol metabolite 17.0 C8H15N5S 213.1048 Isoproturon 21.3 C12H18N2O 206.1419 Mebendazole 18.2 C16H13N3O3 295.0957 Metolachlor 25.6 C15H22NO2Cl 283.1339 Metribuzin 15.0 C8H14N4OS 214.0888 Nicosulfuron 17.0 C15H18N6O6S 410.1009 Prochloraz 23.0 C15H16Cl3N3O2 375.0308 Prometon 16.6 C10H19N5O 225.1590 Prometryn 19.0 C10H19N5S 241.1361 Propazine 23.0 C9H16N5Cl 229.1094 Propiconazole 25.9 C15H17CI2N3O2 341.0698 Isomer 26.1 C15H17CI2N3O2 341.0698 Simazine 18.8 C7H12N5Cl 201.0781 Spinosyn A 20.9 C41H65NO10 731.4608 Spinosyn D 21.9 C42H67NO10 745.4765 Spiroxamine 19.6 C18H35NO2 297.2668 Isomer 19.7 C18H35NO2 297.2668 Terbuthylazine 23.4 C9H16N5Cl 229.1094 Terbutyrn 20.4 C10H19N5S 241.1361 Triflumazole 25.9 C15H15ClF3N3O 345.0856 Acetamiprid Acetochlor Alachlor Bensultap Bromuconazole Carbaryl Carbendazim Carbofuran Cartap Chlorfenvinphos Cyproconazole Cyromazine Deethylatrazine Deisopropylatrazine Dichlorvos Dimethenamide Dimethoate Diuron 16.3 23.0 23.0 21.4 23.8 21.3 6.2 20.4 3.1 26.5 23.4 2.9 15.3 12.1 20.0 24.0 16.3 21.0 C10H11N4Cl C14H20NO2Cl C14H2ONO2Cl C17H21NO4S4 C13H12N3OCl2Br C12H11NO2 C9H9N3O2 C12H15NO3 C7H15N3O2S2 C12H14Cl3O4P C15H18N3OCl C6H10N6 C6H10N5Cl C5H8N5Cl C4H7Cl2O4P C12H18NO2SCl C5H12NO3PS2 C9H10N2OCl2 222.0672 269.1183 269.1183 431.0353 374.9541 201.0790 191.0695 221.1052 237.0606 357.9695 291.1138 166.0967 187.0625 173.0468 219.9459 275.0747 228.9996 232.0170
Table 1.
LOD food mg/kg 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01
Table 1.
Limits of Detection for Pesticides in Food with Retention Time and Accurate Mass (Continued) Ret time 21.8 15.7 15.7 19.2 22.5 16.9 21.2 23.5 12.1 18.7 11.9 19.1 28.6 24.0 25.6 29.0 3.7 17.7 4.5 18.5 6.0 20.6 20.4 25.0 11.5 28.6 14.0 24.1 17.4 24.8 27.3 17.0 27.5 11.4 18.5 23.8 15.0 25.0 29.2 18.8 27.2 14.6 25.4 25.0 27.6 24.4 28.2 23.0 3.2 Formula C11H15NO2S C9H12N2O C9H10N5O2Cl C13H18N2O2 C10H19O6PS2 C10H19O6PS2 C15H21NO4 C11H15NO2S C5H10N2O2S C9H11ClN2O C11H15CIN4O2 C14H18N2O4 C11H15BrCIO3PS C12H17NO2 C11H14NOCl C14H21NOS C10H7N3S C10H9N4SCl C5H11NS3 C7H14N2O2S C7H14N2O3S C11H13NO4 C10H13N2OCl C14H13N3O2SF4 C8H15N5O C17H8N2O3Cl2F8 C10H10N4O C6H11N2O4PS3 C11H15NO4S C9H17NOS C10H14NO5PS C9H9NOCl2 C13H9CI3N2O C7H14N2O4S C9H13N2O2Br C13H12N3OCI2Br C11H23NOS C14H9N2O2ClF2 C21H11N2O3ClF6 C7H5N2O3Cl2F C16H8N2O3Cl2F6 C11H10N2OCl2 C13H13N3O3Cl2 C13H19N3O4 C14H6N2O2Cl2F4 C9H8CI3NO2S C7H7CI3NO3PS C23H30O4 C5H13NO6S4 Accurate mass molecule 225.0823 164.0950 255.0523 234.1368 330.0361 330.0361 279.1471 225.0823 162.0463 198.0560 270.0884 278.1267 371.9351 205.1341 211.0764 251.1344 201.0361 252.0236 181.0054 190.0776 206.0725 223.0845 212.0716 363.0665 197.1277 509.9784 202.0855 301.9619 257.0722 187.1031 291.0330 217.0061 313.9780 222.0674 260.0160 374.9541 217.1500 310.0321 488.0362 253.9661 459.9816 256.0170 329.0334 281.1376 379.9742 298.9341 320.8950 370.2144 310.9626 LOD food mg/kg 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.5 0.5 0.5 0.5
Compound Ethiofencarb Fenuron Imidacloprid Lenacil Malathion 1 Malathion 2 Metalaxyl Methiocarb Methomyl Monuron Nitenpyram Oxadixyl Profenfos Promecarb Propachlor Prosulfocarb Thiabendazole Thiacloprid Thiocyclam Aldicarb Aldicarb sulfoxide Bendiocarb Chlorotoluron Flufenacet Hydroxyatrazine Lufenuron Metamytron Methidathion Methiocarb sulfone Molinate Parathion ethyl Propanil Triclocarban Aldicarb sulfone Bromacil Bromuconazole Butylate Diflubenzuron Flufenoxuron Fluroxypyr Hexaflumuron Imazalil degradate Iprodione Pendimethalin Teflubenzuron Captan Chloropyrifos methyl Spiromesifen Thiosultap
Figure 3.
Example of a report from the molecular feature database search.
data are excellent for the confirmation of the molecular formula for imazalil (however not for the identification of imazalil). The mass accuracy was 0.8 mDa off or 2.7 ppm error. Thus, this is an example of how one confirms the database screening of a pesticide molecular formula. The database report prints the isotopic signature and accurate masses to help for the manual screening of the database hits. When the formula does not contain an A+2 atom (that is, consists of only C, H, N, and O), then formula confirmation is not readily possible by this method. The reason being that only an A+1 peak is present and this peak is dominated by the carbon 13 signal of the molecule, which in itself is not enough for formula confirmation. Thus, for these pesticides (about 30% of the database) more data are required (specifically, retention time or fragment ion). Screening of Fruits and Vegetables Table 2 shows the results of screening six fruit and vegetable samples from a nearby grocery store (apple, pear, tomato, potato, pepper, and cucumber) and one commercial brand of olive oil for the 600 pesticides in the MF database search. The MF database search found from 617 to 2,681 accurate mass peaks in the sample chromatograms. The least complicated sample matrix was the tomato with 617 peaks, and the apple was the most complex sample with 2,681 peaks. The sensitivity of the MF database search was set at a signal-to-noise ratio of 10:1. The quantity of peaks found approximately doubles with decreasing the signal-to-noise
ratio from 20:1 to 10:1. The value of 10:1 is chosen in order to obtain good values for the isotopic signature of the compound, which is the A+1 and A+2 isotope signatures with the maximum instrument sensitivity. The accuracy window of the MF database search is set at 5 ppm in order to be well within the mass accuracy of the LC/TOF MS system, which typically operates at less than 3 ppm and often at 1 to 2 ppm or approximately 0.3 mDa [3]. The number of pesticides found in the 5 ppm mass window of these samples varied from 8 to 41 compounds. The only criterion to be included in this match was that the MH+ ion was within 5 ppm of the database value. Thus, as an example, the pepper sample, which had 2,402 peaks, had only 41 of these peaks that met the 5-ppm accuracy window (Table 2). Of these 41 peaks only three formulae were confirmed based on the correct isotope signature and retention time match (for compounds not containing an A+2 isotope), which was checked not only in the printout of the automated database match, but also by manual confirmation of the data file. The confirmation of pesticides varied from no detections in the potato sample, one pesticide in olive oil, three pesticides in pepper and tomato, and five pesticides in the cucumber and apple. The most common compound found in the fruit and vegetable samples was imazalil, which is a postharvest fungicide used for transport and storage of fruits and vegetables before their sale. Other compounds included organophosphate insecticides, such as diazinon, phosmet, and malathion and the
7
Cl
+1.997
H2C
A+2
N NH
Cl
1 37Cl
+1.997
1 35Cl A+4
Figure 4.
Isotope cluster of the m/z 297 ion and imazalil structure in a pear extract.
oxon of malathion, which is a pesticide degradate. The insect growth regulator buprofezin was found in a tomato sample, as was thiophanate methyl and carbendazim, both fungicides. The software originally used for this work did not address saturated peaks and compounds at high concentrations such as imazalil in the pear sample and buprofezin in the tomato sample were incorrectly identified. The latest software handles saturated peaks and correctly identified these compounds. The accuracy of all confirmed samples had an absolute-value average of 0.3 mDa or 1.2 ppm and a standard deviation of 0.25 mDa and 1.0 ppm, respectively (Table 2 ). The absolute-value average for retention time match was 0.07 minutes and standard deviation of 0.09 minutes. Thus, the windows chosen for the database are chosen with enough margin of error to find 99% of the compounds based on two standard deviations of the mean for mass accuracy and retention time.
Screening of Baby-Food Samples Of the 100 tested compounds, 33 met the baby-food screening level of 0.01 mg/kg. The only compound detected by the database search was a trace level of imazalil in the puree of pear, banana, and orange. Furthermore, examination of the label shows that lemon juice was also used in the baby food, which was a possible source for the imazalil. The, concentration though was approximately 0.0005 mg/kg, which is considerably less than the levels for baby-food safety of 0.01 mg/kg. Imazalil is easily screened because it contains the characteristic A+2 chlorine signature with two chlorine atoms. The error in identification was 0.3 mD or 1 ppm. None of the other compounds of the database was detected. Confirmation was not possible on this sample because of the low signal of the fragment ion and the low concentration of the suspected imazalil (0.0005 mg/kg). The sample was screened as safe, though, based on the health limit of 0.01 mg/kg for baby food. Approximately 10 different baby-food samples have been
Table 2.
Screened Pesticides in Food Samples Using the MF Database Peaks screened 2681 Pesticide matches < 5 ppm 12 Pesticides confirmed LC/TOF MS Imazalil Imazalil degradate Iprodione Fluqinconazole Difenoconazole Terbuthylazine (Deisopropylatrazine) Imazalil Diazinon Buprofezin Buprofezin (Benzthiazuron) Carbendazim Thiophanate methyl Thiabendazole Malathion isomer 1 Malathion isomer 2 Malathion oxon Imazalil Imazalil Carbendazim Imazalil degradate Phosmet None 0.2 0.04 0.22 0.25 0.1 0.21 0.51 0.31 1 0.1 0.7 0.8 0.3 1.1 2 1 0.01 0.05 0.03 0.08 0.05 0.05 0.03 0.04 0 0 Error mDa 1 0.22 0.05 0.74 0.06 0.3 0.11 1 Error ppm 3.9 0.7 0.1 1.8 0.2 1 0.3 3.3 Retention time error (min.) 0.08 0.12 FALSE negative 0 FALSE positive 0
Sample Apple
Olive oil Pepper
1678 2402
10 41
0.09 0.07 0.12 0.1
0 0
0 1
Tomato
617
0 1
Cucumber
1619
17
Pear
1209
14
Potato
1150
11
Mean Standard deviation
0.30 0.25
1.20 1.00
screened, including a variety of brand-name vegetables and fruits and, fortunately, no positive detections have been found for the pesticides in the MF database search with the exception of imazalil shown above. The baby food samples represent the most difficult samples to screen because of the low LODs required. Weakness and Strengths of MF Database Search The only weakness of the database is the loss of mass accuracy because of interferences in the matrix. This problem is easily solved by the addition of a second ion (fragment ion) or a sodium or ammonium adduct ion for added confidence from matrix interference. The use of accurate mass LC/TOF MS combined with database searching is a powerful example of a new wave of monitoring technology for identification of pesticides in food and water. The use of
classical fragmentation libraries with comparison of fragmentation patterns is not needed in LC/TOF MS. Rather the use of molecular formula and the calculated accurate mass, especially when combined with one or two fragment ions of accurate mass, gives the identity of compounds without the worry of the intensity of the fragment ions and how this may vary from instrument to instrument and matrix to matrix. For example, the Molecular Hunter Software works in conjunction with the Molecular Feature Database using the .mhd files to link ions in groups according to their exact retention time (matching within 0.005 minutes). Therefore, it is possible to find and differentiate the fragment ions of a pesticide, such as imazalil, from the background ions of the matrix. Thus, it is the view of the authors that a large problem in LC/MS libraries is on the verge of being solved with the use of accurate mass
databases for pesticide screening in food with a molecular feature algorithmic approach using the MH+ ion and a major fragment ion.

References
1. PAN Europe position on the European Commission Proposal for a Regulation of the European Parliament and of the Council on maximum residue levels of pesticides in products of plant and animal origin COM(2003) 117 final, 2003/0052 (COD). 2. Phillip L. Wylie, Michael J. Szelewski, Chin-Kai Meng, and Christopher P. Sandy, Comprehensive Pesticide Screening by GC/MSD using Deconvolution Reporting Software, Agilent Technologies, publication 5989-1157EN. 3. Imma Ferrer, E. M. Thurman, 2005, Measuring the mass of an electron by LC/TOF MS: A study of twin ions, Analytical Chemistry, 77, 33943400. 4. E. Michael Thurman, Imma Ferrer, 2005, Identification of unknown pesticides in food using both LC/MSD TOF and ion trap MSn, Agilent Technologies, publication 5989-1924EN. 5. M. Anastassiades, S. J. Lehotay, D. Stajnbaher, and F. J. Schenck, Fast and Easy Multiresidue Method Employing Acetonitrile Extraction/Partitioning and Dispersive Solid-Phase Extraction for the Determination of Pesticide Residues in Produce, (2003) Journal of AOAC International, 86:412431. 6. S.J. Lehotay, K. Matovsk, A.R. Lightfield, Use of Buffering and Other Means to Improve Results of Problematic Pesticides in a Fast and Easy Method for Residue Analysis of Fruits and Vegetables, (2005) Journal of AOAC International, 88:615629. 7. Imma Ferrer, Juan F. Garcia-Reyes, M. Mezcua, E. M. Thurman, A. R. Fernandez-Alba, 2005, Multiresidue pesticide analysis in fruits and vegetables by liquid chromatography time-offlight mass spectrometry, J. Chromatography A, 1082, 8190.
Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2006 Printed in the USA October 26, 2006 5989-5496EN
Multiresidue Analysis of 100 Pesticides in Food Samples by LC/Triple Quadrupole Mass Spectrometry Application
Food Safety
Authors
Imma Ferrer and E. Michael Thurman Pesticide Residue Research Group University of Almeria Almeria, Spain Yanyan Fang, Paul Zavitsanos, and Jerry A. Zweigenbaum Agilent Technologies, Inc. USA
Introduction
In recent years, the established regulations regarding the maximum residue limits (MRLs) in commodities have become more and more stringent. The European Union (EU) has set new directives for pesticides at low levels in vegetables in order to meet health concerns. For fruits and vegetables intended for production of baby food, an MRL of 10 g/kg is applicable for all pesticides, and compounds without a stated regulation also have the lowest MRLs at 10 g/kg. The low MRLs have encouraged the development of more sensitive analytical methods to meet the requirements in complex samples. In this sense, liquidchromatography tandem-mass spectrometry (LC-MS-MS) with triple quadrupole in multiple reaction monitoring (MRM) mode has become, so far, the most widely used technique for the monitoring and quantitation of pesticides in food, as reported extensively in the literature. On the other hand, high-resolving power mass spectrometric techniques, such as time-of-flight mass spectrometry (TOF-MS), have been applied recently for screening purposes as well. Nevertheless, the simplicity of methodologies using triple quadrupole as a detection technique, together with the low limits of detection achieved and the MS/MS capability make this technique a valuable tool for routine
Abstract
An analytical methodology for confirming the presence of a group of 100 pesticides in vegetable and fruit samples was developed using the Agilent G6410AA Triple Quadrupole Mass Spectrometer (QQQ). One transition per parent compound was monitored in a single chromatographic run containing two time segments. The sensitivity obtained meets the maximum residue levels (MRLs) established by the European Union regulation for food monitoring programs. The analytical performance of the method was evaluated for different types of fruit and vegetables + orange, tomato, and green pepper + showing little or no matrix effects. Linearity of response over two orders of magnitude was demonstrated (r > 0.99). This study is a valuable indicator of the potential of the QQQ for routine quantitative multiresidue analysis of pesticides in vegetables and fruits.
monitoring programs established in regulatory official laboratories. The easiness of use is sometimes an essential for these types of regulatory agencies, which lack the high-skilled personnel required for more sophisticated techniques such as TOF-MS. Triple quadrupole technology is not new in the sense that it needs to be validated for monitoring purposes and its basis is already wellestablished for routine analysis. Our study in this report is one of the first of its kind to examine the new Agilent Triple Quad for the analysis of pesticides in fruit and vegetables. This topic was chosen because of the relevance of these compounds and their significant use on food commodities. The sensitivity of the QQQ easily meets the levels required by the regulations on pesticides in food.
LC/MS/MS Instrumentation
LC Conditions Column: Agilent ZORBAX Eclipse XDB C-8, 4.6 mm 150 mm, 5 m, (p/n 993967-906). 25 C A = 0.1% formic acid in water B= Acetonitrile 0.6 mL/min 10% B at 0 min 10% B at 5 min 100% B at 30 min 1-5 L
Column temperature: Mobile phase: Flow-rate: Gradient:
Injection volumes: MS Conditions Mode:
Experimental
Sample preparation Pesticide analytical standards were purchased from Dr. Ehrenstorfer (Ausburg, Germany). Individual pesticide stock solutions (around 1,000 g/mL) were prepared in pure acetonitrile or methanol, depending on the solubility of each individual compound, and stored at 18 C. From these mother solutions, working standard solutions were prepared by dilution with acetonitrile and water. Vegetable samples were obtained from the local markets. Blank vegetable and fruit extracts were used to prepare the matrix-matched standards for validation purposes. In this way, two types of vegetables and one fruit (green peppers, tomatoes, and oranges) were extracted using the QuEChERS method already described in a previous application [1]. The vegetable extracts were spiked with the mix of standards at different concentrations (ranging from 2 to 100 g/kg) and subsequently analyzed by LC/MS/MS.
Positive ESI using the Agilent G6410AA Triple Quadrupole Mass Spectrometer Nebulizer: 40 psig Drying gas flow: 9 L/min V capillary: 4000 V Drying gas temperature: 350 C Q1 resolution: Unit Q2 resolution: Unit Fragmentor voltage: 70 V Collision energy: 525 V MRM: 1 transition for every compound as shown in Table 1 Dwell time: 15 msec

Optimization of LC/MS/MS conditions A preliminary study of the optimal MRM transitions for every compound was carried out by injecting groups of analytes (around 10 analytes in one chromatographic run) at a concentration level of 10 g/mL. Various collision energies (5, 10, 15, 20, and 25 V) were applied to the compounds under study. The optimum energies were those that gave the best sensitivity for the main fragment ion and, as a general rule, left about 10% of parent compound in the spectra, and they were selected as optimum ones. Only one fragment ion was chosen as the most abundant product ion for every target compound. Results are shown in Table 1.
Table 1.
Analytical Conditions and Limits of Detection (LOD) for Each of the Compounds Tested Retention time (min) 2.7 2.7 3 4.5 6.4 6.6 7.9 10.8 11 11.2 11.5 11.9 12.5 13.9 14.5 14.8 14.8 15.4 15.5 15.7 16 16.4 16.9 17 17.2 17.2 17.5 17.8 17.9 17.9 18 18 18.4 18.5 18.5 18.6 18.7 18.9 19.4 19.5 19.5 19.6 19.7 20 20.1 20.3 20.3 20.4 20.4 20.4 20.5 Protonated molecule [M+H]+ 167 312 150 182 207 192 202 223 271 198 163 174 262 203 165 188 256 230 223 226 214 258 411 253 297 296 213 312 279 255 202 199 235 241 166 298 221 215 213 242 242 222 224 732 202 254 216 280 287 207 432 Product ion (m/z) 125 232 105 137 89 160 175 148 225 156 88 132 234 175 72 146 209 199 126 184 158 122 182 126 159 264 89 284 219 209 132 72 153 214 109 144 109 187 72 200 186 165 167 142 145 198 174 248 123 72 290 Collision energy 20 10 15 10 5 15 25 5 10 15 5 15 15 15 15 15 10 5 15 20 15 5 15 15 15 20 10 20 10 10 15 10 10 10 5 15 15 15 15 20 15 10 5 5 5 15 15 10 15 15 15 LOD (pg) 10 90 10 8 9 5 10 50 7 3 4 18 8 8 2 4 7 7 6 4 0.8 6 6 3 7 2 10 15 10 120 5 2 20 70 2 10 10 5 3 2 1 2 2 12 2 0.1 0.3 5 5 1 6 3
Compound name Segment 1 Cyromazine Thiosultap Cartap Thiocyclam Aldicarb sulfoxide Carbendazim Thiabendazole Aldicarb sulfone Nitenpyram Hydroxyatrazine Methomyl Deisopropylatrazine Imazapyr Metamitron Fenuron Deethylatrazine Imidacloprid Dimethoate Acetamiprid Prometon Irgarol metabolite Methiocarb sulfone Nicosulfuron Thiacloprid Imazalil Mebendazole Aldicarb Imazaquin Oxadixyl Fluroxypyr Simazine Monuron Lenacil Cyanazine Metolcarb Spiroxamine Dichlorvos Metribuzin Chlorotoluron Prometryn Terbutryn Carbofuran Bendiocarb Segment 2 Spinosad A Carbaryl Irgarol 1051 Atrazine Metalaxyl Difenoxuron Isoproturon Bensultap
Table 1.
Analytical Conditions and Limits of Detection (LOD) for Each of the Compounds Tested (Continued) Retention time (min) 20.5 20.7 20.7 21.3 21.6 21.7 21.9 22.2 22.5 22.6 22.7 22.8 23 23 23.2 23.2 23.2 23.3 23.3 23.7 23.7 24.1 24.6 24.7 24.8 24.9 24.9 24.9 25 25.1 25.2 25.3 25.4 25.5 25.8 26.2 26.4 26.5 26.7 26.8 26.9 27.1 27.6 27.9 28 28.5 28.7 29.2 29.7 Protonated molecule [M+H]+ 233 746 226 388 212 388 376 218 292 226 230 376 304 303 404 318 300 276 208 376 188 311 330 342 331 342 284 346 270 270 364 406 406 359 326 292 315 461 306 305 381 322 373 511 252 489 218 282 336 Product ion (m/z) 72 558 107 301 170 301 308 162 70 169 174 159 217 145 372 160 264 244 151 159 126 158 245 159 127 159 252 278 238 224 194 251 251 155 294 236 162 158 201 169 158 212 303 158 91 158 57 212 236 Collision energy 15 5 5 20 10 20 10 15 10 5 15 20 15 5 10 5 10 10 10 20 10 10 10 20 5 20 10 10 10 10 5 20 20 10 5 10 15 10 10 15 15 15 10 10 15 10 10 5 15 LOD (pg) 5 100 5 11 1 8 6 10 6 15 0.3 6 0.7 5 0.4 2 50 1 5 6 5 9 8 5 5 5 2 7 8 8 5 4 4 8 5 9 8 7 1 1 22 15 7 10 2 6 2 5 30
Compound name Diuron Spinosad D Ethiofencarb Dimethomorph isomer 1 Propachlor Dimethomorph isomer 2 Prochloraz Propanil Cyproconazole Methiocarb Terbutylazine Bromuconazole isomer 1 Fenamiphos Methidathion Azoxystrobin Phosmet Captan Dimethenamide Promecarb Bromuconazole isomer 2 Molinate Diflubenzuron Iprodione Propiconazole isomer 1 Malathion Propiconazole isomer 2 Metolachlor Triflumizole Alachlor Acetochlor Flufenacet Difenoconazole isomer 1 Difenoconazole isomer 2 Chlorfenvinphos Benalaxyl Parathion ethyl Triclocarban Hexaflumuron Buprofezin Diazinon Teflubenzuron Chlorpyrifos methyl Profenofos Lufenuron Prosulfocarb Flufenoxuron Butylate Pendimethalin Trifluralin
The MRM transitions were included in the method with a dwell time of 15 msec, and two different time segments were recorded in the chromatographic run (each one of them containing about half of the pesticides studied). Figure 1 shows the chromatogram corresponding to 100 pg on column for all the compounds studied. Extracted ion chromatograms are overlaid for each one of the target analytes according to their respective protonated molecule and product ion MRM transition. Linearity and Limits of Detection Linearity was evaluated by analyzing the standards solutions at five different concentration levels in the range 2 to 100 pg on column. As an
example, the calibration curve generated for atrazine is shown in Figure 2. As it can be observed in this figure, the linearity of the analytical response across the studied range is excellent, with a correlation coefficient of 0.998. Similar results were obtained for the rest of the compounds analyzed. The limits of detection (LOD) were estimated from the injection of standard solutions at concentration levels corresponding to a signal-to-noise ratio of about 3. The results obtained are included in Table 1 as well. The best limits of detection were obtained for the triazines (from 100 fg to 2 pg on column) and the highest limits of detection were for fluoroxypyr and spinosad D (above 100 pg).
x105 1.6 1.5 1.4 1.3 1.2 1.1 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
+ MRM (282.0 212.0) mix100_100 pg_5May.d
12
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 Abundance vs. acquisition time (min)
Figure 1.
Product ion chromatograms of a mix of 100 pesticides (concentration: 100 pg on column).
500000 400000 Area 300000 200000 100000 0 0
y = 3828.6x % 1790.3 R2 = 0.9988
20
40
60 Amount injected (pg)
80
100
120
Figure 2.
Calibration curve for atrazine using a linear fit with no weighting and no origin treatment.
Application to Vegetable Matrices To confirm the suitability of the method for analysis of real samples, matrix-matched standards were analyzed in three different matrices + green pepper, tomato, and orange + at two different concentration levels (10 and 100 g/kg). Figure 3 shows the analysis of a green pepper spiked with the pesticide mix at 10 g/kg (10 pg on column). As it can be observed in two of the MS/MS extracted product ion chromatograms, for dimethoate and azoxystrobin, compounds can be easily identified in these complex matrices due to the selectivity of the MRM transitions, thus fulfilling the regulation limits imposed by the EU directives. In general, the LOD obtained meet the requirements regarding the MRLs imposed by the existing European regulations.
Reference
1. Imma Ferrer and E. Michael Thurman, Determination of Fungicides in Fruits and Vegetables by Time-of-Flight and Ion Trap LC/MS (2005) Agilent Technologies, publication 5989-2209EN www.agilent.com/chem.

x104 1 1.8 1.4 1 0.6 0.2 0 x103 1 3 2 1 0 x103 5 4 3 2 1 0 1
+ TIC MRM (** **) mix100_10pg_green pepper.d
12
+ MRM (230.0 199.0) mix100_10pg_green pepper.d
12
(a)
+ MRM (404.0 372.0) mix100_10pg_green pepper.d
12
(b)
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 Abundance vs. acquisition time (min)
Figure 3.
MRM chromatogram of a spiked green pepper sample at 10 g/kg. Product ion chromatograms for (a) dimethoate and (b) azoxystrobin.
Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2006 Printed in the USA August 9, 2006 5989-5469EN
Improving the Effectiveness of Method Translation for Fast and High Resolution Separations Application
Author
Abstract
The increased availability of sub-2-micron (STM) columns and increased demand for methods friendly to mass spectrometers has led to strong trend toward conversion of existing HPLC methods to smaller diameter and smaller particle size columns. While the conversion is a simple mathematical exercise requiring the scaling flow rates, gradient times and injection volumes, many users observe less than perfect results. Here we look closely at the problem and propose calculations that improve the speed and/or resolution in a more predictable and beneficial way.
Simplistically, a column of 250-mm length and containing 5-m particles can be replaced by a 150-mm length column packed with 3-m particles. If the ratio of length to particle size is equal, the two columns are considered to have equal resolving power. Solvent consumption is reduced by L1/L2, here about 1.6-fold reduction in solvent usage per analysis. If an equal mass of analyte can then be successfully injected, the sensitivity should also increase by 1.6-fold due to reduced dilution of the peak as it travels through a smaller column of equal efficiency. LC/MS (Liquid Chromatography/Mass Spectrometry) ionization sources, especially the electrospray ionization mode, have demonstrated greater sensitivity at lower flow rates than typically used in normal LC/UV (UltraViolet UV/VIS optical detection) methods, so it may also be advantageous to reduce the internal diameter of a column to allow timely analysis at lower flow rates. The relationship of flow rate between different column diameters is shown in Equation 1.
Flowcol. 1 Diam.column2 Diam.column1
2
Introduction
Methods developed on older columns packed with large 5- or 10-m particles are often good candidates for modernization by replacing these columns with smaller dimension columns packed with smaller particle sizes. The potential benefits include reduced analysis time and solvent consumption, improved sensitivity and greater compatibility with mass spectrometer ionization sources.
= Flowcol. 2
(eq. 1)
The combined effect of reduced length and diameter contributes to a reduction in solvent consumption and, again assuming the same analyte mass can be injected on the smaller column, a proportional increase in peak response. We normally scale the injection mass to the size of the column,
though, and a proportional injection volume would be calculated from the ratio of the void volumes of the two columns, multiplied by the injection volume on the original column.
Inj. vol.col. 1 Volumecolumn2 Volumecolumn1 = Inj. vol.col. 2 (eq. 2)
For isocratic separations, the above conditions will normally result in a successful conversion of the method with little or no change in overall resolution. If one wishes to improve the outcome of the method conversion, though, there are several other parameters that should be considered. The first of these parameters is the column efficiency relative to flow rate, or more correctly efficiency to linear velocity, as commonly defined by van Deemter [1] and others, and the second is the often overlooked effect of extracolumn dispersion on the observed or empirical efficiency of the column. Van Deemter observed and mathematically expressed the relationship of column efficiency to a variety of parameters, but we are most interested here in his observations that there is an optimum linear velocity for any given particle size, in a wellpacked HPLC column, and that the optimum linear velocity increases as the particle size decreases. Graphically, this is often represented in van Deemter plots as shown in Figure 1, a modified version of the original plot [2]. In Figure 1 we observe that the linear velocity at which 5-m materials are most efficient, under the conditions used by the authors, is about 1 mm/sec. For 3.5-m materials the optimum linear velocity is about 1.7 mm/sec and has a less distinct opti0.02
mum value, suggesting that 3.5-m materials would give a more consistent column efficiency over a wider flow range. For the 1.8-m materials, the minimum plate height, or maximum efficiency, is a broad range beginning at about 2 mm/sec and continuing past the range of the presented data. The practical application of this information is that a reduction in particle size, as discussed earlier, can often be further optimized by increasing the linear velocity which results in a further reduction in analysis time. This increase in elution speed will decrease absolute peak width and may require the user to increase data acquisition rates and reduce signal filtering parameters to ensure that the chromatographic separation is accurately recorded in the acquisition data file. The second important consideration is the often overlooked effect of extracolumn dispersion on the observed or empirical efficiency of the column. As column volume is reduced, peak elution volumes are proportionately reduced. If smaller particle sizes are also employed there is a further reduction in the expected peak volume. The liquid chromatograph, and particularly the areas where the analytes will traverse, is a collection of various connecting capillaries and fittings which will cause a measurable amount of bandspreading. From the injector to the detector flow cell, the cumulative dispersion that occurs degrades the column performance and results in observed efficiencies that can be far below the values that would be estimated by purely theoretical means. It is fairly typical to see a measured dispersion of 20 to 100 L in an HPLC system. This has a disproportionate effect on the smallest columns and smallest particle sizes, both of which are expected to yield the smallest
Plate height (mm)
0.015
0.01
0.005
5.0 m SB-C18 3.5 m SB-C18 1.8 m SB-C18
Lin. vel. 4.6 mm 3 mm 2.1 mm 1 mm
mm/sec mL/min mL/min mL/min mL/min
1 0.7 0.3 0.14 0.033
2 1.4 0.6 0.29 0.066
3 2.1 0.9 0.44 0.1
4 2.8 1.2 0.58 0.133
5 3.5 1.5 0.73 0.166
Figure 1.
van Deemter plot with various flow rates and particle sizes.
possible peak volumes. Care must be taken by the user to minimize the extracolumn volume and to reduce, where practical, the number of connecting fittings and the volume of injection valves and detector flow cells. For gradient elution separations, where the mobile phase composition increases through the initial part of the analysis until the analytes of interest have been eluted from the column, successful method conversion to smaller columns requires that the gradient slope be preserved. While many publications have referred to gradient slope in terms of % change per minute, it is more useful to express it as % change per column volume. In this way, the change in column volume during method conversion can be used to accurately render the new gradient condition. If we think of each line of a gradient table as a segment, we can express the gradient by the following equation:
Experimental Conditions
System Agilent 1200 Series Rapid Resolution LC consisting of: G1379B micro degasser G1312B binary pump SL G1367C autosampler SL, with thermostatic temperature control G1316B Thermostatted column compartment SL G1315C UV/VIS diode array detector SL, flow cell as indicated in individual chromatograms ChemStation 32-bit version B.02.01 Columns Agilent ZORBAX SB-C18, 4.6 mm 250 mm, 5 m Agilent ZORBAX SB-C18, 3.0 mm 150 mm, 3.5 m Mobile phase conditions Organic solvent: Aqueous solvent: Gradient Conditions Gradient slope: 7.8% or 2.3% per column volume, as indicated. See individual chromatograms for flow rate and time Acetonitrile 25 mm phosphoric acid in Milli-Q water
% Gradient slope =
(eq. 3)
Sample Standard mixture of chlorinated phenoxy acid herbicides, 100 g/mL in methanol
Note that the use of % change per column volume rather than % change per minute frees the user to control gradient slope by altering gradient time and/or gradient flow rate. A large value for gradient slope yields very fast gradients with minimal resolution, while lower gradient slopes produce higher resolution at the expense of increased solvent consumption and somewhat reduced sensitivity. Longer analysis time may also result unless the gradient slope is reduced by increasing the flow rate, within acceptable operating pressure ranges, rather than by increasing the gradient time. Resolution increases with shallow gradients because the effective capacity factor, k*, is increased. Much like in isocratic separations, where the capacity term is called k', a higher value directly increases resolution. The effect is quite dramatic up to a k value of about 5 to 10, after which little improvement is observed. In the subsequent examples, we will see the results associated with the calculations discussed above.
Results
The separation was initially performed on a standard 4.6 250 mm, 5-m ZORBAX SB-C18 column thermostatted to 25 C (Figure 2) using conditions referenced in US EPA Method 555. The method was then scaled in flow and time for exact translation to a 3.0 150 mm, 3.5-m column (Figure 3). Solvent consumption is reduced from 60 mL to 15.5 mL per analysis. The separation was then re-optimized for faster separation with the identical slope, 7.8%, by increasing the flow rate from 0.43 to 1.42 mL/min, and proportionately reducing the gradient time (Figure 4). Finally, increased resolution is demonstrated by keeping the original times used in Figure 3 with the increased flow rate (Figure 5). This yields a gradient with identical time but a reduced slope of 2.3%. The increased resolution of peaks 4 and 5 is readily apparent. The conditions in Figure 4, 7.8% slope at increased linear velocity on 3.0 150 mm, 3.5-m material, yield a separation with comparable resolution to the original 4.6 250 mm method, but with only a 12-minute total analysis time. This is excellent for
12.557
mAU 350 300 250 200 150 100 50 0
14.380
17.779
18.871
19.414
21.063
23.050
24.667
13.194
12.5
15
17.5
17.607
20
22.5
25
27.5
Conditions EPA Method 555 with ZORBAX SB-C18 columns and fast DAD detector ZORBAX SB-C18 4.6 mm 250 mm, 5 m Column temp: 25 C Gradient: 10% to 90% ACN vs. 25 mM H3PO4 Gradient slope: 7.8% ACN/column volume Analysis flow rate: 1 mL/min Group A Compounds Total analysis time: Detection: Figure 2. 60 min UV 230 nm, 10-mm 13-L flow cell, filter 2 seconds (default)
Gradient separation of herbicides on 4.6 250 mm 5-m ZORBAX SB-C18.
800 700 600 500 400 300 200
9.990
mAU
Conditions: EPA Method 555 with ZORBAX SB-C18 columns and fast DAD detector ZORBAX SB-C18 3.0 mm 150 mm, 3.5 m Column temp: 25 C Gradient: 25 mm H3PO4/ACN, 0% to 90% ACN in 18 minutes Gradient slope: 7.8% ACN/column volume Analysis flow rate: 0.43 mL/min Detection: UV 230 nm, 3-mm 2-L flow cell, filter 0.2 seconds Total analysis time: 36 min.
12.061 12.831 16.314
8.781
14.106
13.046
15.317
15.786
17.081
0 8 10 12 14 16 18 min
Figure 3.
Gradient separation of herbicides on 3.0 150 mm, 3.5-m ZORBAX SB-C18.
13.854
100
18.348
9.120
29.595 min
400
300
2.674
3.011
mAU
Conditions EPA Method 555 with ZORBAX SB-C18 columns and fast DAD detector ZORBAX SB-C18, 3.0 mm 150 mm, 3.5 m Column temp: 25 C Gradient: 25 mM H3PO4/ACN, 10% to 90% ACN in 5.4 min. Gradient slope: 7.8% ACN/column volume Analysis flow rate: 1.42 mL/min Detection: UV 230 nm, 3-mm 2-L flow cell, filter 0.2 seconds Total analysis time: 12 min.
3.620
3.850 3.919
4.240
200
2.780
0 2 2.5 3 3.5 4 4.5 5 5.5 min
Figure 4.
High speed gradient separation of herbicides on 3.0 150 mm, 3.5-m ZORBAX SB-C18.
3.964
mAU 400 350 300 250
Conditions EPA Method 555 with ZORBAX SB-C18 columns and fast DAD detector ZORBAX SB-C18, 3.0 mm 150 mm, 3.5 m Temp: 25 C Gradient: 25 mM H3PO4/ACN, 10% to 90% ACN in 18 min. Gradient slope: 2.3% ACN/column volume Analysis flow rate: 1.42 mL/min Detection: UV 230 nm, 3-mm 2-L flow cell, filter 0.2 seconds Total analysis time: 36 min.
6.793 7.583 11.291 10.257 11.465 8.905
150 100
0 4 6 8 10 12 min
Figure 5.
Reduced slope gradient separation of herbicides on 3.0 150 mm, 3.5-m ZORBAX SB-C18.
12.692
4.093 4.253
50
7.912
10.056
200
4.933
4.743
100
4.611
4.914
high throughput screening and quantitation of a large number of samples. Figure 5, with the gradient slope reduced to 2.3%, results in a high-resolution separation with a calculated R value of 3.3 vs. the standard 3.0 150 mm separation value of 1.9, for the critical pair seen in Figure 5 at 7.5 to 8 minutes. In Table 1 the column has been replaced with a low dead volume connecting union in a system fitted with 0.12-mm id capillary tubing at all points of sample contact. A 1-L injection of dilute actone
Table 1. Volumetric Measurements of Various Flow Cells Elution volume (L) 11 14 Half height width (L) 5 6 5 Sigma width (L) 12 18
Conclusion
Careful analysis of the existing gradient conditions, coupled with an awareness of the need to accurately calculate new flow and gradient conditions can lead to an easy and reliable conversion of existing methods to new faster or higher resolution conditions. In addition, awareness of extracolumn dispersion, especially with small and high resolution columns, will ensure good column efficiency which is critical to a successful translation of the method.
References
1. J. J. van Deemter, F. J. Zuiderweg, A. Klinkenberg, Chemical Engineering Science 1956, 5, 271289 2. The Influence of Sub-Two Micron Particles on HPLC Performance, Agilent Technologies, application note 5989-9251EN, May 2003
Flow cell New SL 2 L 3 mm Micro 6 mm 1.7 L (n = 2) Semi-micro 6 mm 5 L (n = 2) Standard 10 mm 13 L New SL 10 mm 13 L
13
6.5
18.5
26 27
11 11
26 25

is made to determine the bandspreading contribution of the system, with various flow cells. Multiple flow cells were tested, and the average result reported, where possible. The elution volume summarizes the total volume of all tubing in the system. While the absolute volume from the 2-L to the 13-L flow cells is 11 L, we observe an increase of 15 to 16 L because of the larger diameter inlet tubing integral to the larger volume flow cells.
Determination of 44 Pesticides in Foodstuffs by LC/MS/MS Application
Food Safety
Authors
Masahiko Takino Agilent Technologies, Inc. Hachioji, Tokyo Japan Toshitsugu Tanaka Kobe Institute of Health Department of Food Chemistry Chuo-ku, Kobe Japan
Abstract
A sensitive and selective analytical method for the determination of 44 pesticide residues in several foodstuffs using the Agilent G6410AA Triple Quadrupole Mass Spectrometer (QQQ) was developed. This method use two different sample preparation methods followed by LC/MS/MS (liquid chromatography/tandem mass spectrometry). The limits of detection for all pesticides were less than 10 ng/mL in foodstuff. The sensitivity of QQQ easily met the maximum residue limits (MRLs) of all investigated pesticides in Japan Food Hygiene Law.
safety. In recent years, the established regulations regarding MRLs in commodities have been more and more stringent. In Japan, the positive list system was introduced this year, and MRLs have been set for over 500 pesticides in all foodstuffs. This new system sets different MRLs for each pesticide within each food group. Typically, the MRLs range from 0.01 to 3 g/g depending on the commodities and pesticides. The low MRLs fostered the development of more sensitive analytical methods to meet the requirements of complex samples. In this sense, liquid chromatography/tandem mass spectrometry (LC/MS/MS) with QQQ in multiple reaction monitoring (MRM) mode has become so far the most widely used techniques for the quantitation of polar pesticides in food. MRM mode provides for more specific detection in a complex matrix such as food. In this work, 44 pesticides (Tables 1 and 2) are analyzed in two separate runs with sample analytical conditions. The sensitivity requirements set by the positive lost system for these pesticides are easily met.
Experimental
Chemicals The acetonitrile was of LC/MS grade from Wako Pure Chemical Ind (Japan). Toluene, acetone, nhexane, formic acid, sodium chloride, and anhydrous sodium sulfate were of analytical grade from Wako Pure Chemical. All SPE cartridges were purchased from Spelco Japan (Japan). Pesticide standards were obtained from Hayashi Pure Chemical (Japan).
Introduction
Pesticides are widely used in agricultural practices. The main application can be classified in production and post-harvest treatment of agricultural commodities for transport purposes. In this sense, production agriculture comprises the main category of use of pesticides subject to control requirements and, therefore, maximum residue levels (MRLs) have been fixed to assess food
Sample Preparation
Extraction
LC/MS/MS Instrument The LC/MS/MS system used in this work consists of an Agilent 1100-series vacuum degasser, binary pump, well-plate autosampler, thermostatted column compartment, and the Agilent G6410 Triple Quadrupole Mass Spectrometer with an electrospray ionization source (ESI). The objective of the method development was to obtain a fast and sensitive analysis for quantifying pesticides in fruits and vegetables. For chromatographic resolution and sensitivity, different solvents and columns were optimized. It was found that a simple solvent system using water, acetonitrile, formic acid, formic acetate, and a 1.8-m particle size C18 column would work very well.
LC Conditions Instrument: Column: Column temp: Mobile phase: Agilent 1100 HPLC ZORBAX Extend C18, 100 mm 2.1 mm, 1.8 m (p/n 728700-902) 40 C A = 0.1% formic acid +5 mM ammonium formate in water B= Acetonitrile 10% B at 0 min, 80% B at 30 min 0.2 mL/min 5 L Agilent 6410 QQQ Positive ESI 10 L/min 50 psig 350 C 4000 V m/z 100 to 550 Variable 100 V Shown in Tables 1 and 2 Shown in Tables 1 and 2
Vegetable and fruit samples were obtained from the local markets. A sample of 10 to 500 g was chopped in a food processor to obtain thoroughly mixed homogenates. A 20-g portion of sample homogenate was weighed in a 200-mL PTFE centrifuge tube. Then 50 mL of acetonitrile was added and blended in a Polytoron. The extract was then filtered by applying vacuum. The filtrate was collected and the residue was re-extracted with 20 mL of acetonitrile. The filtrates were combined in a 100-mL volumetric flask and made up to volume with acetonitrile. A 20-mL portion of the extract was transferred into a PTFE centrifuge tube, and 10 g of NaCl and 20 mL of 0.5 M phosphate buffer (pH 7.0) were added to the extract followed by shaking for 5 min. Five grams of anhydrous Na2SO4 were added to the acetonitrile layer obtained after salting out. After removing anhydrous Na2SO4, the extract was evaporated to dryness by rotary evaporator (water bath temperature did not exceed 40 C). The residue was dissolved in 2 mL of acetonitrile-toluene (3:1).
Cleanup
Group 1 - The extract was loaded into a GCB/amino propyl SPE cartridge (500 ng/500 mg) preconditioned with 10 mL of acetonitrile-toluene (3:1). The 20 mL of acetonitrile-toluene (3:1) was further added to the SPE cartridge. All eluate was collected and evaporated by rotary evaporator. The residue was dissolved in 4 mL of methanol. Group 2 - The extract was loaded into a silica gel SPE cartridge (500 mg) preconditioned with 10 mL each of methanol, acetone, and n-hexane (10 mL of methanol, 10 mL of acetone, and 10 mL of n-hexane, total volume is 30 mL). The 10 mL of acetone-triethylamine-n-hexane (20:0.5:80) was further added to the SPE cartridge. All eluate was discarded. The 20 mL of acetone-methanol (1:1) was applied and the eluate was collected and evaporated by rotary evaporator. The residue was dissolved in 4 mL of methanol. Standard Preparation Stock solutions of individual pesticides were prepared in methanol at 1 g/mL. Serial dilutions using methanol produced a range of standard mixture solutions at 0.001 g/mL to 1 g/mL. The blank matrix residues were fortified with a mixture of pesticides studied at 10 ng/g.
2
Gradient: Flow rate: Injection vol: MS Conditions Instrument: Source: Drying gas flow: Nebulizer: Drying gas temp: Vcap: Scan: Fragmentor: MRM ions: Collision energy:
LC/MS/MS Method Quantitative analysis was carried out using MRM mode with time program. The parameters of MRM transition are shown in Tables 1 and 2.
Table 1. No 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33
Data Acquisition Parameters of MRM Transitions of Each Pesticide in Group 1 Pesticides Thiabendazole Thiamethoxam Clothianidin Chloridazon Imidacloprid Dimethirimol Oxycarboxine Thiacloprid Azamethiophos Ferimzone(E) Ferimzone(Z) Phenmedipham Azinphos-methyl Simeconazole Isoxaflutol Pyriftalid Tridemorph Methoxyfenozide Chromafenozide Fenoxycarb Naproanilide Butafenacil Cyazofamide Anilofos Pyrazolate Benzofenap Cyflufenamid Indoxacarb Clomeprop Cloquincet-mexyl Furathiocarb Lactofen Tralkoxydim RT (min) 5.018 6.16 7.83 8.19 8.39 8.8 11.02 11.03 12.87 13.21 13.7 17.77 17.9 18.5 18.7 18.7 19.21 20.06 20.57 20.63 21.27 21.55 21.7 22.5 23.5 24 24.3 24.37 24.78 24.8 25.7 26.3 26.7 Molecular weight 201 291 249 221 255 209 267 252 324 254 254 317 318 293 359 318 297 312 394 301 291 491 324 367 438 430 412 527 372 335 365 478 329 Precursor ion (m/z) 202 292 250 222 256 210 268 253 325 255 255 318 132 294 360 319 298 313 175 302 292 492 325 368 439 431 413 528 373 336 383 479 330 Product ion (m/z) 175 211 169 104 209 171 175 126 183 124 132 136 77 70 251 139 130 149 141 88 171 331 108 199 173 105 241 150 299 238 195 344 284 Collision energy(V) 20 5 10 10 20 20 10 20 10 20 20 20 15 15 15 20 15 20 20 15 10 20 10 10 15 20 20 15 5 15 15 15 10
Table 2. No 1 2 3 4 5 6 7 8 9 10 11
Data Acquisition Parameter of MRM Transitions of Each Pesticide in Group 2 Pesticides Flumetsulam Thidiazuron Imazaquin Thifensulfuron-methyl Florasulam Forchlorfenuron-methyl Clorasulam-methyl Diclosulam Fomesafen Triflusulfuron-methyl Haloxyfop RT (min) 9.96 11.95 12.25 12.89 13.75 14.63 16.41 16.83 18.27 19.29 19.67 Molecular weight 325 220 311 387 359 247 429 405 438 492 361 Precursor ion (m/z) 326 221 312 388 360 248 430 406 456 493 362 Product ion (m/z) 129 102 267 167 129 129 398 161 344 264 316 Collision energy (V) 20 10 20 10 20 10 10 20 10 15 15

Optimization of MRM Transitions Determination of the optimal MRM transitions for each pesticide was carried out using full scan mode followed by product ion scan mode using two pesticide standard mixtures at 1 g/mL. TICs of these standard mixtures in full scan mode and product ion scan mode are shown in Figures 1 and 2. The mass spectrum of each pesticide by full scan mode exhibited protonated molecular ions; [M+H]+ as the base peak ion except azinphos-methyl, furathiocarb, and fomesafen, which exhibited fragment ion and ammonium adduct ion [M+NH4]+. These ions were selected as precursor ions for MRM mode. It was possible to generate individual product ion MS/MS spectrum of each pesticide by using multiple acquisition and time programming mode. As shown in Tables 1 and 2, 10 time segments for 33 pesticides in group 1 and 7 time segments for 11 pesticides in group 2 were used for MRM mode.
Total ion chromatograms of pesticide standard mixture corresponding to the minimum MRL value for pesticides (10 ng/mL) are shown in Figure 3. These show excellent signal-to-noise (S/N) ratios for all pesticides. The limit of detection (LOD) for each pesticide was determined using an S/N ratio of 3 with an MRM chromatogram of each pesticide at 1 ng/mL (see Table 3). To evaluate the linearity of the calibration curves, various concentrations of pesticide standard solutions ranging from 0.001 ng/mL to 1 ng/mL were analyzed. As shown in Table 3, the linearity was very good for all pesticides with correlation coefficients (r2) greater than 0.998 The matrix effect of this method was investigated by using orange, apple, potato, and cabbage extracts spiked with pesticide standards at 10 ng/mL. Typical MRM chromatograms of orange extract are shown in Figures 4 and 5. The other chromatograms of apple, potato, and cabbage extract are shown in Figure 6. There was not additional peak from sample matrix in all food when compared with the pesticide standard mixture. These results indicate that MRM mode has very high selectivity.
x107 1.2
(A)
6 18 19, 20 24 4 2 3 11 9 15,16 13 21 17 22 23
29,30
7,8 0.8 1
10
12
31 27, 28 26 25 32 33
0.4
14
1 x106 6.5
9 6
11
13
15
17
19
21
23
25
27
29
31
Abundance versus acquisition time (min)
(B)
9 10 12 19, 20 18 21 22 23 21 23 33 25 27, 28 26 25 32 27 29 31 29, 30 24 31 7, 8
4.5
2.5 1 0.5 1 3 5
5 34
11
13
14
15, 16 17
11
13
15
17
19
Figure 1.
TIC of 33 pesticides standard in full scan mode (A) and product ion scan mode (B) at 1 g/mL.
x107 4.0
(A)
6 3.0 2.0 1.0 1 2 4 5 7 8 9
10
11
10
11
12 3
13
14
15
16
17
18
19
20
21
22
23
Abundance versus acquisition time (min) x106 3.0
(B)
4 5 6 10
2.0
1.0
7 8 9
11
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Figure 2.
TIC of 11 pesticides standard in full scan mode (A) and product ion scan mode (B) at 1 g/mL.
Furthermore, the change on the peak intensity of each pesticide by sample matrix was calculated by comparing with the peak intensity of pesticide standards. As these results show in Table 4, the relative intensity of each pesticide ranged from 91 to 116%. Thus, matrix effect such as ion suppression may be insignificant and it was possible to use external standards instead of matrix matched standards. The repeatability of each pesticide in orange extract is also shown in Table 4, and the RSD of each pesticide was in the range from 1.7 to 5.9%.
7,8 5000 4000 3000 2000 1 1000 2 4 3 5 7 6 10 11 9 11 13 15 17 14 13 15,16 17 19 22 23 24 21 26
(A)
12 9 18 19,20,21 25 29,30
28 27 25
31 32 33 27 29
23
Abundance versus acquisition time (min) 6 16000 12000 1 8000 4000 23 7 9 1 3 5 7 9 11 13 15 17 19 11 21 23 25 27 29 8
(B)
5 4 10
Figure 3.
TIC of 33 pesticide standards (A) and 11 pesticides standard (B) at 10 ng/mL in MRM mode.
Table 3.
Linearity and LOD of 44 Pesticide Standard Solutions r2 0.9999 0.9992 0.9999 0.9993 0.9995 0.9989 0.9993 0.9991 0.9988 0.9993 0.9995 0.9993 0.9997 0.9992 0.9991 0.9988 0.9991 0.9996 0.9994 0.9992 0.9989 0.9969 0.9977 LOD (ng/mL) <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 0.34 0.53 <0.1 <0.1 <0.1 <0.1 <0.1 1.21 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 No 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 Pesticides Methoxyfenozide Chromafenozide Fenoxycarb Naproanilide Butafenacil Cyazofamide Anilofos Pyrazolate Benzofenap Cyflufenamid Clomeprop Indoxacarb Quinclorac-methyl Furathiocarb Lactofen Tralkoxydim r2 0.9993 0.9992 0.9988 0.9993 0.9994 0.9987 0.9991 0.9990 0.9982 0.9993 0.9993 0.9991 0.9988 0.9987 0.9987 0.9992 LOD (ng/mL) 0.55 0.49 <0.1 <0.1 <0.1 0.43 <0.1 0.51 0.49 0.43 0.61 1.04 0.63 <0.1 1.10 0.52
No Pesticides Group 1 1 Thiabendazole 2 Thiamethoxam 3 Clothianidin 4 Chloridazon 5 Imidacloprid 6 Dimethirimol 7 Oxycarboxine 8 Thiacloprid 9 Azamethiophos 10 Ferimzone(E) 11 Ferimzone(Z) 12 Phenmedipham 13 Azinphos-methyl 14 Simeconazole 15 Isoxaflutol 16 Pyriftalid 17 Tridemorph Group 2 1 Flumetsulam 2 Thidiazuron 3 Imazaquin 4 Thifensulfuron-methyl 5 Florasulam 6 Forchlorfenuron-methtyl
7 8 9 10 11
Clorasulam-methyl Diclosulam Fomesafen Triflusulfuron-methyl Haloxyfop
0.9987 0.9989 0.9989 0.9992 0.9995
<0.1 <0.1 0.32 <0.1 0.19
25000 15000
1
m/z = 202>>>175
700 300
13
m/z = 132>>>77
550 350
2
m/z = 292>>>211
900 500
14
m/z = 294>>>70
300 200
3
m/z = 250>>>169
2400 1200
15
m/z = 360>>>251
10 11 12 13 14 15 16 17 18 19 20
700 300
4
m/z = 222>>>104
250 150 50
16
m/z = 319>>>139
300 100
24
5 m/z = 256>>>209
17
m/z = 298>>>130
12
800 400
2500
18
m/z = 313>>>149
m/z = 210>>>171
1500
10
11
12
13 900
7
2500 1500
19
m/z = 175>>>141
m/z = 268>>>175
500
4000 2000
8
m/z = 253>>>126
900 600
20
m/z = 302>>>88
3000
9
m/z = 325>>>183
7 8 9 10 11 12 13 14 15 16
1100 500
21
m/z = 292>>>171
1000
300 200 100 400 200
10
m/z = 255>>>124
500 300
22
m/z = 492>>>331
11
m/z = 255>>>132
300 200
23
m/z = 325>>>108
3000 1000
12
m/z = 318>>>136
10 11 12 13 14 15 16 17 Retention time (min) 18 19 20
2200 1000
24
m/z = 368>>199
15 16 17 18 19 20 21 Retention time (min) 22 23 24
Figure 4.
MRM of 33 pesticides in orange extract spiked at 10 ng/mL. (Continued)
400 200
25
m/z = 439>>>173
1000 400
31
m/z = 383>>>195
200 80
26
m/z = 431>>>105
120 60
32
m/z = 479>>>344
180 100
27
m/z = 413>>>241
15 16 17 18 19 20 21 22 23 24
300 100
33
m/z = 330>>>284
21 22 23 24 25 26 Retention time (min) 27 28 29
800 400
28
m/z = 373>>>299
45 25
29
m/z = 528>>>150
3500 1500
30
m/z = 336>>>238
21 22 23 24 25 26 Retention time (min) 27 28 29
Figure 4.
MRM of 33 pesticides in orange extract spiked at 10 ng/mL.
6000 2000
1
m/z = 326>>>129
6000 4000
7
m/z = 430>>>398
7500 2500
2
m/z = 221>>>102
8000 4000
8
m/z = 406>>>161
7500 2500
3
m/z = 312>>>267
800 400
9
m/z = 456>>>344
20000 10000
4
m/z = 388>>>167
20000 10000
10
m/z = 493>>>264
11
7500 2500
5
m/z = 360>>>129
2000 1000
m/z = 362>>>316
16 17 18 19 Retention time (min) 20
6
10000 5000 1
m/z = 248>>>129
2 3 4 5 6 7 8 9 10 11 12 13 14 15
Retention time (min)
Figure 5. 8
MRM of 11 pesticides in orange extract spiked at 10 ng/mL.
7, 8 5000
12 9 13 15,16 10 11 14 17 22 24 29, 30 18 19, 20, 21 23 25 28 26 27 31 32 33
(A)
5 1 2 6 34
3000
1000
5000
(B)
3000
1000
5000
(C)
3000
1000 1 3 5 7 9 11 13 15 17 Retention time (min) 6 4 1 8 2 3 5000 7 9 11 19 21 23 25 27 29
15000 10000
(D)
10
15000 10000 5000
(E)
15000
(F)
10000
5000
9 11 13 Retention time (min)
15
17
19
Figure 6.
TIC of spiked at 10 ng/mL.
Table 4.
Relative Intensity of Each Pesticide in Sample Extracts Relative intensity(%) Orange* Cabbage 105 (3.2) 101 103 (2.1) 98 106 (2.9) 101 105 (3.3) 102 (1.7) 103 (4.6) 106 (3.7) 104 (3.1) 93 (4.6) 116 (4.1) 96 (5.3) 90 (2.1) 104 (4.4) 102 (2.7) 97 (4.1) 92 (3.1) 96 (2.8) 97 (3.4) 99 (2.1) 91 (4.3) 102 (2.6) 93 (3.5) 102 (2.7) 103 (4.7) 108 (5.2) 108 (3.4) 109 (2.6) 105 (4.2) 105 (4.1) 102 (1.8) 100 (3.7) 101 (3.3) 97 (2.6) 104 (4.8) 105 (3.1) 106 (2.9) 99 (3.1) 101 (4.4) 94 (3.9) 95 (3.3) 99 (5.9) 97 (4.1) 108 (4.8) 106 97 107 102 104 90 109 99 103 102 104 103 99 102 96 105 97 114 92 105 101 111 110 105 107 104 104 109 111 110 101 100 112 106 103 104 102 101 111 114 Apple 116 104 109 101 102 103 104 106 94 102 100 104 106 108 104 104 103 100 102 98 104 87 103 103 98 105 100 106 104 105 105 102 156 102 100 116 103 100 97 96 95 104 110 Potato 107 105 112 109 104 108 106 108 84 112 104 110 110 103 93 97 101 111 101 103 114 95 107 97 108 101 111 104 105 101 112 117 104 113 101 113 109 108 142 107 109 108 124
No Pesticides Group 1 1 Thiabendazole 2 Thiamethoxam 3 Clothianidin 4 5 6 7 8 9 Chloridazon Imidacloprid Dimethirimol Oxycarboxine Thiacloprid Azamethiophos
10, 11 Ferimzone(E,Z) 12 Phenmedipham 13 Azinphos-methyl 14 15 16 17 18 19 20 21 22 23 24 25 26 27 29 28 30 31 32 33 Simeconazole Isoxaflutol Pyriftalid Methoxyfenozide Chromafenozide Tridemorph Fenoxycarb Naproanilide Butafenacil Cyazofamide Anilofos Pyrazolate Benzofenap Cyflufenamid Indoxacarb Clomeprop Cuinclorac-methyl Furathiocarb Lactofen Tralkoxydim
Group 2 1 Flumetsulam 2 Thidiazuron 3 Imazaquin 4 5 6 7 8 9 10 11 Thifensulfuron-methyl Florasulam Forchlorfenuron-methyl Clorasulam-methyl Diclosulam Fomesafen Triflusulfuron-methyl Haloxyfop
*( ): RSD,% calculated based on five replicates within one day
10
Conclusions
The multiresidue method by LC/MS/MS described here was suitable for the determination of 44 pesticides in a variety of food samples due to its high sensitivity and high selectivity. Another advantage of this method is that ion suppression was not observed for all food samples studied. Thus, it may eliminate the need for matrix-matched standards, which make analysis more tedious for samples from different origins. For more details concerning this application, please contact masahiko_takino@agilent.com

11
Screening for 926 Pesticides and Endocrine Disruptors by GC/MS with Deconvolution Reporting Software and a New Pesticide Library Application Note
Food and Environmental
Authors
Philip L. Wylie Agilent Technologies, Inc. 2850 Centerville Rd. Wilmington, DE 19808-1610 USA
endocrine disrupters in about two minutes per sample. Deconvolution helps identify pesticides that are buried in the chromatogram by co-extracted materials. The new database was compared to the smaller one for the DRS analysis of 17 surface water samples. With the new database, DRS found 99 pesticides, metabolites, fire retardants, and related contaminants that were not contained in the original RTL Pesticide and Endocrine Disruptor Library.
Abstract
An updated and greatly expanded collection of mass spectral libraries has been introduced, replacing Agilents RTL Pesticide Library and DRS pesticide solution. The new library contains 926 pesticides, endocrine disruptors, and related compounds 359 more than the original library. Included are all compounds specified for GC/MS analysis in the new Japanese Positive List regulations. All compounds have locked retention times that can be accurately reproduced using an Agilent GC/MS system with the ChemStation's Retention Time Locking software. The new Database can be used as a standard GC/MS library for compound identification or with Agilent's Screener software for identifications based upon retention time and mass spectral matching. The greatest benefit accrues when these libraries are used with Agilents new version of Deconvolution Reporting Software (part number G1716AA version A.03.00). This solution allows one to screen GC/MS files for all 926 pesticides and
Introduction
Several years ago Agilent Technologies introduced Retention Time Locking (RTL) for gas chromatography (GC) and GC with mass spectral detection (GC/MS). RTL software makes it possible to reproduce retention times from run-to-run on any Agilent GC or GC/MS, in any laboratory in the world, so long as the same nominal method and GC column are used (1). Since any laboratory can reproduce retention times generated in another, it is possible to create mass spectral libraries that contain locked retention times. By locking their method to the published database, users can screen GC/MS files for all of the librarys compounds. Hits are required to have the correct retention time as well as the correct spectrum, which eliminates many false positives and gives more confidence in compound identifications (2).
More recently, Agilent introduced Deconvolution Reporting Software (DRS) that incorporates mass spectral deconvolution with conventional library searching and quantification. DRS results from a marriage of three different GC/MS software packages: 1) The Agilent GC/MS ChemStation, 2) The National Institute of Standards and Technology (NIST) Mass Spectral Search Program with the NIST 05 MS Library, and 3) The Automated Mass Spectral Deconvolution and Identification System (AMDIS) software, also from NIST. The original DRS software was intended to be a comprehensive solution for pesticide analysis and, therefore, included the mass spectra (in AMDIS format) and locked retention times for 567 pesticides and suspected endocrine disrupters (3). Recently, Agilent introduced an updated and greatly expanded Pesticide and Endocrine Disruptor Database (part number G1672AA) that now contains 926 entries. This represents the addition of 359 new compounds to the original library. At the same time, Agilent introduced a new version of the DRS software (part number G1716AA version A.03.00) that can be used with any Agilent-provided or user-developed DRS library. Pesticide and Endocrine Disruptor Database Contents The G1672AA Pesticide and Endocrine Disruptor Database contains virtually all GC-able pesticides, including those introduced very recently. In addition, the database includes numerous metabolites, more endocrine disruptors, important PCBs and PAHs, certain dyes (for example, Sudan Red), synthetic musk compounds, and several organophosphorus fire retardants. This new database includes: A conventional mass spectral library for use with Agilent GC/MS ChemStations
A screener database for use with Agilents powerful screener software that is integrated into the GC/MS ChemStation Locked Retention Times for all 926 compounds that any Agilent 5975 or 5973 GC/MS user can reproduce in their laboratory Files for use with Agilents G1716AA (A.03.00) Deconvolution Reporting Software An e-method that can be loaded into Agilents G1701DA (version D.02.00 SP1 or higher) with instrument parameters for acquiring GC/MS files and analyzing the data with DRS. These parameters are listed in Table 1. Example files Application notes On November 29, 2005, the Japanese Government published a Positive List system for the regulation of pesticides, feed additives, and veterinary drugs. Maximum Residue Limits (MRL) have been set for 758 chemicals while 65 others have been exempted from regulation. Fifteen substances must have no detectable residues. Other agricultural chemicals not mentioned have a uniform MRL of 0.01 ppm (4). This new regulation is scheduled to take effect on May 29, 2006. Of the pesticides in the Japanese Positive List, 265 are to be analyzed by GC/MS. The new G1672AA Pesticide library contains mass spectra and locked retention times for all of these compounds. Thus, a laboratory could screen for all 265 positive list compounds and several hundred more pesticides in just 13 minutes after the GC/MS run.
Experimental
Table 1 lists the instrumentation, software, and analytical parameters used by Agilent for pesticide analysis. Depending upon the desired injection volume, a PTV inlet or split/splitless inlet can be used.
Table 1.
Instrumentation and Conditions of Analysis Agilent 6890N Agilent 7683 Injector and AutoSampler Agilent PTV operated in the solvent vent mode or Split/Splitless Agilent 30 m 0.25 mm 0.25 m HP-5MSi (part number 19091S-433i) Helium in the constant pressure mode Chlorpyrifos-methyl locked to 16.596 min (nominal column head pressure = 17.1 psi) 70 C (2 min), 25 C/min to 150 C (0 min), 3 C /min to 200 C (0 min), 8 C /min to 280 C (1015 min) Temp program: 40 C (0.25 min), 1600 C/min to 250 C (2 min); Vent time: 0.2 min; Vent flow: 200 mL/min; Vent pressure: 0.0 psi; Purge flow: 60.0 mL/min; Purge time: 2.00 min 15 L (using a 50-L syringe) Agilent 5975 inert Atune.u Scan (or SIM with SIM DRS library) 50550 u 230, 150, and 280 C, respectively 4.00 min Autotune voltage
Gas Chromatograph Automatic Sampler Inlet Column Carrier gas Retention time locking Oven temperature program PTV inlet parameters Injection volume Mass Selective Detector Tune file Mode Scan range Source, quad, transfer line temperatures Solvent delay Multiplier voltage Software GC/MSD ChemStation Deconvolution Reporting Software Library searching software Deconvolution software MS Libraries
Agilent part number G1701DA (version D02.00 sp1 or higher) Agilent part number G1716AA (version A.03.00) Deconvolution Reporting Software NIST MS Search (version 2.0d or greater) (comes with NIST '05 mass spectral library Agilent part number G1033A) Automated Mass Spectral Deconvolution and Identification Software (AMDIS_32 version 2.62 or greater; comes with NIST '05 mass spectral library Agilent part number G1033A) NIST 05 mass spectral library (Agilent part number G1033A) Agilent RTL Pesticide and Endocrine Disruptor Libraries in Agilent and NIST formats (part number G1672AA)

DRS, which has been described in preceding papers (3,5,6), can be summarized as follows: Three separate, but complimentary, data analysis steps are combined into the DRS. First, the GC/MS ChemStation software performs a normal quantitative analysis for target pesticides using a target ion and up to three qualifiers. An amount is reported for all calibrated compounds that are detected. For other compounds in the database, an estimate of their concentration can be reported based upon an average pesticide response factor
that is supplied with the DRS software. The DRS then sends the data file to AMDIS, which deconvolutes the spectra and searches the Agilent RTL Pesticide Library using the deconvoluted full spectra. A filter can be set in AMDIS, which requires the analytes retention time to fall within a userspecified time window. Because RTL is used to reproduce the RTL database retention times with high precision, this window can be quite small typically 1020 seconds. Finally, the deconvoluted spectra for all of the targets found by AMDIS are searched against the 147,000-compound NIST mass spectral library for confirmation; for this step, there is no retention time requirement.
This approach was rapidly adopted by many laboratories because of its ability to identify pesticides in complex chromatograms containing high levels of co-extracted interferences. Indeed, the solution proved to be so useful that users began to create their own DRS libraries (7). Therefore, the DRS was unbundled from the pesticide database so that it could be used with any agilent-provided or user-created database. The original 567-compound RTL Pesticide Library (G1049A) included pesticides, a few metabolites, and most of the GC-amenable endocrine disruptors that were known at the time. The new version of the library includes many more pesticides, endocrine disruptors, and metabolites. This update also contains important compounds from other classes of contaminants that have been found in food and water supplies. Included are eighteen polychlorinated biphenyls (PCBs), four polybrominated biphenyls (PBBs), several polynuclear aromatic hydrocarbons (PAHs), several organophosphorus fire retardants, three important toxaphene congeners, and three Sudan dyes.
Advantages of Deconvolution Figure 1 shows a screen from AMDIS that illustrates the power of this deconvolution software. The white trace in Figure 1A is the total ion chromatogram while the other three are extracted ions of a deconvoluted peak (a component in AMDIS terminology). Note that the TIC and extracted ions are not scaled to each other and this component is actually obscured by co-eluting compounds. Figure 1B juxtaposes the deconvoluted component spectrum (white) with the complete undeconvoluted spectrum (black). Clearly, this component is buried under co-eluting peaks that would ordinarily obscure the analyte. Figure 1C shows that the deconvoluted peak (white spectrum) is a good library match for norflurazon (black spectrum). The locked retention time for norflurazon in the RTL Pesticide Database is 26.933 min, which is just 2.3 seconds away from its observed RT in this chromatogram. Confidence in peak identifications is greatly enhanced by the combination of spectral deconvolution and locked retention time filtering.
Figure 1.
AMDIS screen showing the identification of norflurazon. A) The total ion and extracted ion chromatograms where norflurazon elutes. B) The deconvoluted component spectrum (white) juxtaposed with the spectrum at 26.972 min (black). C) The deconvoluted component matched to the library spectrum of norflurazon.
Surface Water Analysis - Revisiting an Earlier Study In an earlier study, a comparison was made between Agilents DRS and conventional pesticide analysis (3). The California Department of Food and Agriculture (CDFA) provided data files for 17 surface water extracts that had been analyzed in their laboratory. Since the GC/MS chromatograms were locked to the Agilent pesticide method, it was possible to analyze these data files using DRS without having to re-run the samples. The original DRS analysis was made using the 567-compound RTL Pesticide Database. For comparison, these same data files were re-analyzed using the new 926-compound RTL Pesticide Database. The chromatogram (Figure 2) and the DRS report (Figure 3) from one of these samples are shown below.
Excluding phthalates, seven new compounds (shown with bold type in Figure 3) were identified using the 926-compound database: 4-chlorophenyl isocyanate (a phenylurea herbicide metabolite); 3,4-dichlorophenyl isocyanate (diuron metabolite); tris(2-chloroethyl) phosphate (a fire retardant); caffeine (a stimulant); Cyprodinil (a fungicide); desmethyl-norflurazon (a metabolite of norflurazon, an herbicide); and tris(2-butoxyethyl) phosphate (a fire retardant). Although caffeine is not generally considered to be dangerous, it is included in the database because it has been found frequently in sewage effluent and in numerous waterways together with a various pharmaceuticals and pesticides (8).
2800000 2600000 2400000 2200000 2000000 Abundance 1800000 1600000 1400000 1200000 1000000 800000 600000 400000 200000 5.00 10.00 15.00 20.00 Time 25.00 30.00 35.00 40.00 TIC: E02-557.d\data.ms
Figure 2.
Chromatogram of a surface water extract that was analyzed by DRS using the new RTL Pesticide and Endocrine Disrupter Database. The results of this analysis are shown in Figure 3.
MSD Deconvolution Report Sample Name: E02-557 Data File: C:\MSDChem\1\DATA\CDFA surface water data\E02-557.d Date/Time: 11:24 AM Tuesday, Apr 4 2006 The NIST library was searched for the components that were found in the AMDIS target library. Agilent ChemStation amount (ng) AMDIS match 62 84 99 84 93 67 63 RT Diff (sec.) 3.2 1.8 3.1 2.0 2.1 1.7 7.7 62 1.29 85 98 86 96 83 88 79 95 80 90 97 90 99 90 69 2.2 2.6 2.6 3.0 0.7 1.4 1.0 1.3 1.6 3.2 0.4 1.5 0.4 0.7 0.1 70 87 87 94 89 75 98 65 4.5 1.5 0.5 0.8 3.3 0.3 1.9 71 10 1 69 79 94 83 83 90 1 2 1 1 1 1 3 84 92 88 90 74 86 78 83 74 88 90 84 94 87 2 1 2 1 1 2 1 1 1 4 1 1 1 1 1 NIST reverse match 48 86 95 85 89 84 Hit number 1 2 1 1 2 2
RT 4.4689 4.4689 4.8840 6.3879 6.8357 7.6988 7.9342 8.1112 8.1112 8.941 9.7903 10.0019 10.7109 10.9684 11.6491 12.9326 13.4309 13.7478 15.4048 15.9474 16.5988 17.3653 18.4213 18.9214 20.5633 20.5633 26.4247 26.9700 26.9992 27.3984 28.0127 29.6537 33.9298 33.9298 13.739 Figure 3.
Cas # 106445 0000 104121 102363 759944 95761 131113 25013165 0000 29878317 134623 84662 119619 126738 1582098 122349 115968 1517222 58082 84695 5598130 7287196 84742 51218452 121552612 76470252 23576241 27314132 85687 51235042 78513 117817 84764 0000
Compound name 4-Methylphenol 3-Carbobenzyloxy-4-ketoproline 4-Chlorophenyl isocyanate Diuron Metabolite [3,4-Dichlorophenyl isocyanate] EPTC 3,4-Dichloroaniline Dimethylphthalate Butylated hydroxyanisole 7-Methoxy-2,2,4,8-tetramethyltricyclo [5.3.1.0(4,11)]undecane Tolyltriazole [1H-Benzotriazole, 4-meth-] N,N-Diethyl-m-toluamide Diethyl phthalate Benzophenone Tributyl phosphate Trifluralin Simazine Tris(2-chloroethyl) phosphate Phenanthrene-d10 Caffeine Diisobutyl phthalate Chlorpyrifos Methyl Prometryn Di-n-butylphthalate Metolachlor Cyprodinil 9,9-Dimethoxy-9-sila-9, 10-dihydroanthracene Norflurazon, DesmethylNorflurazon Butyl benzyl phthalate Hexazinone Tris(2-butoxyethyl) phosphate Bis(2-ethylhexyl)phthalate Di-n-nonyl phthalate Phthalic acid, 3,4-dichlorophenyl propyl ester Phenanthrene-d10
DRS report from the analysis of a surface water sample. The compounds shown in bold type were found by the new RTL Pesticide Database but not the original one because these compounds were not included.
For this sample, the ChemStation identified only tolyltriazole at 8.941 min, but AMDIS did not confirm this assignment, nor could it be confirmed manually. Butylated hydroxyanisole was tentatively identified by AMDIS with a low match value, but the retention time is off by 7.7 seconds which is considerably more than most other hits. This compound is not in the NIST library so it could not be confirmed. The ChemStation method used for this analysis required that all three qualifier ions fall within 20% (relative) which is a rigorous requirement for such a complex sample. This explains why so few compounds were found by the ChemStation. Cyprodinil (20.563 min) was identified by AMDIS but the NIST library search failed to confirm its presence. The next line shows that the best NIST library match is an anthracene derivative that is nothing like cyprodinil. This result was obtained when AMDIS was configured to use uncertain peaks as shown in Figure 4. When this feature is
turned off in DRS Compound Identification Configuration, the best NIST library hit for this spectrum is, indeed, cyprodinil. When a compound's identity is ambiguous, as with cyprodinil, it may be useful to perform the DRS search both ways and compare the results. In the comparison described earlier (3), DRS was able to identify all 37 pesticides found by the CDFA chemist. However, DRS completed the task for all 17 samples in about 20 minutes compared to ~8 hours for the manual procedure (Table 2). Moreover, DRS identified one false positive in the CDFA report and found 34 additional pesticides and related compounds. Using the new 926-compound Database, it took 32 minutes to analyze all of the samples and DRS was able to find an additional 99 pesticides, metabolites, fire retardants, and related compounds (Table 2).
Figure 4.
DRS configuration screen for the method called Tri_Pest. When the box labeled Use Uncertain Peaks is checked, AMDIS will use uncertain peaks for library searches. When unchecked, AMDIS ignores uncertain mass spectral peaks. Sometimes, this can affect the quality of a library match.
Table 2.
Comparison of the Results Obtained by Screening 17 Surface Water Extracts Using Traditional Methods (CDFA) and Using DRS With Two Different Databases the G1049A With 567 Compounds and the G1672AA With 926 Entries Agilent DRS (Original G1049A database) Same 37 +34 more 0 20 minutes Agilent DRS (G1672 AA database) Same 37 +99 more 0 32 min
CDFA Targets found (not counting ISTD) False positives Processing time 37 1 ~8 hrs (ChemStation only)
Handling Stereoisomers Many pesticides have multiple stereoisomers with virtually identical mass spectra. For example, cyfluthrin has four diastereomers arising from its three chiral centers. It is very difficult and sometimes impossible to determine the elution order of these isomers and most analysts report them as a sum of the isomer amounts. Agilents G1049A RTL Pesticide database arbitrarily assigned each isomer a Roman numeral with I for the earliest eluting isomer, II for the next, and so on. The same Chemical Abstracts Service number (CAS #) was assigned to all of the isomers. Generally, it was a CAS # for the compound with unstated stereochemistry. This caused some incompatibility with AMDIS as explained below. AMDIS software differentiates among compounds using a chemical identification number. The easiest and most consistent approach is to use each compound's CAS #. The default setting for AMDIS is to allow each CAS # to be used only once when analyzing a GC/MS data file. While this seems logical, it requires that each database entry have a different CAS #. It is possible to allow multiple hits per compound by checking the box in AMDIS found in the drop down menu under Analyze/ Settings/Identif. However, this allows multiple peaks to be assigned the same compound name.
In the new RTL Pesticide Database (G1672AA), the Roman numeral designations remain and the first isomer in the series is given its genuine CAS #. Subsequent isomers in the series are given unique, but fictitious CAS #s generated by Agilent. The compound's real CAS # appears in braces after the compound name. For example, the cyfluthrin isomers are entered into the database as shown in Table 3.
Table 3.
RT 32.218 32.359 32.477 32.536
Method for Listing Compounds with Multiple Stereoisomers in the New G1672AA RTL Pesticide Database Compound name* CAS #** Cyfluthrin I Cyfluthrin II {CAS # 68359-37-5} Cyfluthrin III {CAS # 68359-37-5} Cyfluthrin IV {CAS # 68359-37-5} 68359-37-5 999028-03-4 999029-03-7 999030-03-4
* In a series, the earliest eluting isomer is identified with I and is assigned its legitimate CAS #. Subsequent isomers are assigned unique, but fictitious CAS #s (see footnote **). Their actual CAS # is put in braces behind the compound name. **Cyfluthrin I has been given it's genuine CAS #. Cyfluthrin II-IV have been given unique numbers that can be distinguished from actual CAS numbers because they all have six digits before the first hyphen (9 total) and all begin with the series 999.
Figure 5 shows how permethrin was identified in a spinach sample using both databases with AMDIS configured to allow one hit per compound. Using the older 567-compound database (G1049A) only one permethrin isomer was identified because its CAS # could be used only once. With the new format used in the 926-compound RTL Pesticide Database (G1672AA), both isomers of permethrin were identified. Not surprisingly, the NIST library search found no hits with the same fictitious CAS # assigned to permethrin II. So, the software printed the best match on the following line. This compound, a cyclopropanecarboxylic acid derivative, is a permethrin isomer. So long as the NIST library search is turned on in DRS, it will always print another line after reporting a compound with a fictitious CAS #. Note that these fictitious CAS #s always contain 9 digits and begin with 999. A)
Agilent ChemStation amount (ng) AMDIS match 88 RT Diff (sec.) 3.9 NIST reverse match 91 Hit number 3
RT 31.6158
Cas # 52645531
Compound name Permethrin II
B)
Agilent ChemStation amount (ng) AMDIS match 78 65 RT Diff (sec.) 2.6 3.5 95 1 NIST reverse match 81 Hit number 3
RT 31.4127 31.6088 31.6088
Cas # 52645531 999046036 51877748
Compound name Permethrin I Permethrin II {CAS # 52645-53-1} Cyclopropanecarboxylic acid, 3-(2,2-dichlorovinyl)-2,2-dimethyl-, (3-phenoxyphenyl)methyl ester, (1R-trans)-
Figure 5.
A) A single isomer of permethrin was identified by DRS using the G1049A 567-compound database when AMDIS was not allowed to use multiple hits per compound. B) Two permethrin isomers are identified by DRS with the G1672AA 926-compound database under the same circumstances.
Conclusions
The new G1672AA RTL Pesticide and Endocrine Disruptor library contains substantially more target analytes than its predecessor. With the addition of 359 new compounds, it is the most comprehensive library of its type available today. Many new pesticides, metabolites, and endocrine disruptors were added along with important PCBs, PBBs, PAHs, synthetic musk compounds, Sudan dyes, and organophosphorus fire retardants. The database contains all of the analytes specified for GC/MS analysis in the new Japanese Positive List regulations. When combined with the complete DRS solution, one can screen GC/MS data files for all 926 compounds in about two minutes per sample. This is the fastest, most comprehensive, most accurate, and least tedious method for screening food and environmental samples for these compounds.
5. C. P. Sandy, A Blind Study of Pesticide Residues in Spiked and Unspiked Fruit Extracts using Deconvolution Reporting Software, Agilent Technologies, publication 5989-1564EN, www.agilent.com/chem 6. C. Lesueur and M. Gartner, Routine Identification and Quantification of Pesticide Multiesidues in Fruit and Vegetable Samples with Full Scan, SIM, and Deconvolution Reporting Software, 2005 Ernhrung/Nutrition, 29 (11) 466471 7. X. Ping, C.-K. Meng, and M. Szelewski, Building Agilent GC/MSD Deconvolution Reporting Libraries for any Application, Agilent Technologies, publication 5989-2249EN, www.agilent.com/chem 8. Large-scale studies of the occurrence and distribution of new contaminants in the environment Reconnaissance studies, USGS, Contaminant Occurrence Studies, http://toxics.usgs.gov/topics/reconnaissance_ studies.html
References
1. V. Giarocco, B. Quimby, and M. Klee, Retention Time Locking: Concepts and Applications, Agilent Technologies, publication 5966-2469E, www.agilent.com/chem 2. H. Prest, P. Wylie, K. Weiner, and D. Agnew, Efficient Screening for Pesticides and Endocrine Disrupters Using the 6890/5973 GC/MSD System, Agilent Technologies, publication 5968-4884E, www.agilent.com/chem 3. P. L. Wylie, M. J. Szelewski, C.-K. Meng, C. P. Sandy, Comprehensive Pesticide Screening by GC/MSD using Deconvolution Reporting Software, Agilent Technologies, publication 5989-1157EN, www.agilent.com/chem 4. Introduction of the Positive List System for Agricultural Chemical Residues in Foods Department of Food Safety, Ministry of Health, Labour and Welfare http://www.mhlw.go.jp/english/topics/foodsafety/positivelist060228/introduction.html

Acknowledgments
The author wishes to thank Dr. G. Kempe of the Landesuntersuchungsanstalt Sachsen, Institut, Chemnitz, Germany for his help in acquiring much of the data for this library update. The author also thanks Dr. Mark Lee and Mr. Steve Siegel of the California Department of Food and Agriculture for providing the surface water extract data files.
10
Appendix A
Lists of Compounds in Databases

1,2,4-Trichlorobenzene 1,2-Dibromo-3-chloropropane 1,3,5-Tribromobenzene 1,3-Dichlorbenzene 17a-Ethynylestradiol 1-naphthalenol 2-(1-naphthyl)acetamide 2-(2-Butoxyethoxy)ethyl thiocyanate 2-(Octylthio)ethanol 2,3,4,5-Tertrachloronitrobenzene 2,3,4,5-Tetrachlorophenol 2,3,4,6-Tetrachlorophenol 2,3,5,6-Tetrachlorophenol 2,3,5,6-Tetrachloro-p-terphenyl 2,3,5-Trichlorophenol 2,3,5-Trimethacarb 2,3,6-Trichloroanisole 2,3,7,8-Tetrachlorodibenzofuran 2,3,7,8-Tetrachlorodibenzo-p-dioxin 2,4,5,6-Tetrachloro-m-xylene 2,4,5-T methyl ester 2,4,5-Trichloroaniline 2,4,5-Trichlorophenol 2,4,5-Trichloro-p-terphenyl 2,4,5-Trimethylaniline 2,4,6-Tribromoanisole 2,4,6-Tribromophenol 2,4,6-Trichloroanisole 2,4,6-Trichlorophenol 2,4-D methyl ester 2,4-D sec-butyl ester 2,4-DB methyl ester 2,4'-Dichlorobenzophenone (2,4'-Dicofol decomposition product) 2,4-Dichlorophenol 2,4-Dichlorophenyl benzenesulfonate 2,4-Dimethylaniline 2,4-Dimethylphenol 2,6-Dichlorobenzamide 2,6-Dichlorobenzonitrile 2,6-Dimethylaniline 2-[3-Chlorophenoxy]propionamide 2-Chlorophenol 2-Ethyl-1,3-hexanediol 2-ethyl-6-methylaniline 2-Hydroxyestradiol 2-Methyl-4,6-dinitrophenol 2-Methylphenol 2-Nitrophenol 2-Phenoxypropionic acid 3,4,5-Trimethacarb 3,4-Dichloroaniline 3,5-Dichloroaniline 3-Aminophenol 3-Chloro-4-fluoroaniline 3-Chloro-4-methoxyaniline 3-Chloroaniline 3-Hydroxycarbofuran 3-Indolylacetonitrile 3-Trifluormethylaniline 4,4'-Dichlorobenzophenone 4,4'-Oxydianiline 4,6-Dinitro-o-cresol (DNOC) 4-Aminodiphenyl 4-Bromoaniline 4-Chloro-2-methylaniline 4-Chloro-3-methylphenol 4-Chloroaniline 4-Chlorophenyl isocyanate 4-Isopropylaniline 4-Methylphenol 4-Nitrophenol 4-Nonylphenol 5,7-Dihydroxy-4'-methoxyisoflavone 9,10-Anthraquinone Acenaphthene Acenaphthylene Acephate Acequinocyl acetamiprid Acetochlor Acifluorfen methyl ester Aclonifen Acrinathrin Alachlor Aldrin Allidochlor Ametryn Amidithion Aminocarb Amitraz Amitraz metabolite [Methanimidamide, N(2,4-dimethylphenyl)-N'-methyl-] Ancymidol Anilazine Aniline Anilofos Anthracene Aramite I Aramite II {CAS # 140-57-8} Atraton Atrazine Atrazine-desethyl Azaconazole Azamethiphos Azibenzolar-S-methyl Azinphos-ethyl Azinphos-methyl Aziprotryn metabolite [2-Amino4-isopropylamino-6-methylthio1,3,5-triazine] Aziprotryne Azobenzene Azoxybenzene Azoxystrobin Barban Beflubutamid Benalaxyl Benazolin-ethyl Bendiocarb Benfluralin 11
Benfuracarb Benfuresate Benodanil Benoxacor Bentazone Bentazone methyl derivative Benthiocarb Benzene, 1,3-bis(bromomethyl)Benzenesulfonamide Benzidine Benzo(a)anthracene Benzo(a)pyrene Benzo[b]fluoranthene Benzo[g,h,i]perylene Benzo[k]fluoranthene Benzophenone Benzoximate metabolite Benzoylprop ethyl Benzyl benzoate b-Estradiol BHC alpha isomer BHC beta isomer BHC delta isomer BHC epsilon isomer Bifenazate metabolite (5-Phenyl-o-anisidine) Bifenox Bifenthrin Binapacryl Bioallethrin Bioallethrin S-cyclopentenyl isomer Bioresmethrin Biphenyl Bis(2,3,3,3-tetrachloropropyl) ether Bis(2-butoxyethyl) phthalate Bis(2-ethylhexyl)phthalate Bisphenol A Bitertanol I Bitertanol II {CAS # 55179-31-2} Boscalid (Nicobifen) Bromacil Bromfenvinphos-(E) Bromfenvinphos-(Z) Bromobutide Bromocyclen Bromophos 12
Bromophos-ethyl Bromopropylate Bromoxynil Bromoxynil octanoic acid ester Bromuconazole I Bromuconazole II {CAS # 116255-48-2} Bufencarb Bupirimate Buprofezin Butachlor Butafenacil Butamifos Butoxycarboxim Butralin Butyl benzyl phthalate Butylate Butylated hydroxyanisole Cadusafos Cafenstrole Caffeine Captafol Captan Carbaryl Carbetamide Carbofuran Carbofuran-3-keto Carbofuran-7-phenol Carbophenothion Carbosulfan Carboxin Carfentrazone-ethyl Carpropamid Carvone Cashmeran Cekafix Celestolide Chinomethionat Chloramben methyl ester Chloranocryl Chlorbenside Chlorbenside sulfone Chlorbicyclen Chlorbromuron Chlorbufam Chlordecone Chlordene, trans-
Chlordimeform Chlorethoxyfos Chlorfenapyr Chlorfenethol Chlorfenprop-methyl Chlorfenson Chlorfenvinphos Chlorfenvinphos, cisChlorfenvinphos, transChlorflurecol-methyl ester Chlormefos Chlornitrofen Chlorobenzilate Chloroneb Chloropropylate Chlorothalonil Chlorotoluron Chlorpropham Chlorpyrifos Chlorpyrifos Methyl Chlorthal-dimethyl Chlorthiamid Chlorthion Chlorthiophos Chlorthiophos sulfone Chlorthiophos sulfoxide Chlozolinate Chrysene Cinerin I Cinerin II Cinidon-ethyl cis-Chlordane Clodinafop-propargyl Clomazone Cloquintocet-mexyl Coumaphos Crimidine Crotoxyphos Crufomate Cyanazine Cyanofenphos Cyanophos Cyclafuramid Cycloate Cyclopentadecanone Cycluron
Cyflufenamid Cyfluthrin I Cyfluthrin II {CAS # 68359-37-5} Cyfluthrin III {CAS # 68359-37-5} Cyfluthrin IV {CAS # 68359-37-5} Cyhalofop-butyl Cyhalothrin I (lambda) Cyhalothrin (Gamma) Cymiazole Cymoxanil Cypermethrin I Cypermethrin II {CAS # 52315-07-8} Cypermethrin III {CAS # 52315-07-8} Cypermethrin IV {CAS # 52315-07-8} Cyphenothrin cisCyphenothrin trans- {CAS # 39515-40-7} Cyprazine Cyproconazole Cyprodinil Cyprofuram Cyromazine d-(cis-trans)-Phenothrin-I d-(cis-trans)-Phenothrin-II {CAS # 260002-80-2} Dazomet DDMU [1-Chloro-2,2-bis(4'-chlorophenyl)] Decachlorobiphenyl Deltamethrin Demephion Demeton-S Demeton-S-methylsulfon Desbromo-bromobutide Desmedipham Desmetryn Dialifos Di-allate I Di-allate II {CAS # 2303-16-4} Diamyl phthalate Diazinon Diazinon-oxon Dibenz[a,h]anthracene Dicamba Dicamba methyl ester Dicapthon Dichlofenthion Dichlofluanid
Dichlofluanid metabolite (DMSA) Dichlone Dichlormid Dichlorophen Dichlorprop Dichlorprop methyl ester Dichlorvos Diclobutrazol Diclocymet I Diclocymet II {CAS # 139920-32-4} Diclofop methyl Dicloran Dicrotophos Dicyclohexyl phthalate Dicyclopentadiene Dieldrin Diethatyl ethyl Diethofencarb Diethyl dithiobis(thionoformate) (EXD) Diethyl phthalate Diethylene glycol Diethylstilbestrol Difenoconazol I Difenoconazol II {CAS # 119446-68-3} Difenoxuron Diflufenican Diisobutyl phthalate Dimefox Dimepiperate Dimethachlor Dimethametryn Dimethenamid Dimethipin Dimethoate Dimethomorph-(E) Dimethomorph-(Z) {CAS # 110488-70-5} Dimethylphthalate Dimethylvinphos(z) Dimetilan Dimoxystrobin Di-n-butylphthalate Di-n-hexyl phthalate Diniconazole Dinitramine Di-n-nonyl phthalate Dinobuton
Dinocap I Dinocap II {CAS # 39300-45-3} Dinocap III {CAS # 39300-45-3} Dinocap IV {CAS # 39300-45-3} Di-n-octyl phthalate Dinoseb Dinoseb acetate Dinoseb methyl ether Dinoterb Dinoterb acetate Di-n-propyl phthalate Diofenolan I Diofenolan II {CAS # 63837-33-2} Dioxabenzofos Dioxacarb Dioxathion Diphacinone Diphenamid Diphenyl phthalate Diphenylamine Dipropetryn Dipropyl isocinchomeronate Disulfoton Disulfoton sulfone Ditalimfos Dithiopyr Diuron Diuron Metabolite [3,4-Dichlorophenyl isocyanate] Dodemorph I Dodemorph II {CAS # 1593-77-7} Drazoxolon Edifenphos Empenthrin I Empenthrin II {CAS # 54406-48-3} Empenthrin III {CAS # 54406-48-3} Empenthrin IV {CAS # 54406-48-3} Empenthrin V {CAS # 54406-48-3} Endosulfan (alpha isomer) Endosulfan (beta isomer) Endosulfan ether Endosulfan lactone Endosulfan sulfate Endrin Endrin aldehyde Endrin ketone 13
EPN Epoxiconazole EPTC Erbon Esfenvalerate Esprocarb Etaconazole Ethalfluralin Ethidimuron Ethiofencarb Ethiolate Ethion Ethofenprox Ethofumesate Ethofumesate, 2-Keto Ethoprophos Ethoxyfen-ethyl Ethoxyquin Ethylenethiourea Etoxazole Etridiazole Etridiazole, deschloro- (5-ethoxy3-dichloromethyl-1,2,4-thiadiazole) Etrimfos Eugenol Exaltolide [15-Pentadecanolide] Famoxadon Famphur Fenamidone Fenamiphos sulfoxide Fenamiphos-sulfone Fenarimol Fenazaflor Fenazaflor metabolite Fenazaquin Fenbuconazole Fenchlorazole-ethyl Fenchlorphos Fenchlorphos-oxon Fenclorim Fenfuram Fenhexamid Fenitrothion Fenitrothion-oxon Fenobucarb Fenoprop 14
Fenoprop methyl ester Fenothiocarb Fenoxanil Fenoxaprop-ethyl Fenoxycarb Fenpiclonil Fenpropathrin Fenpropidin Fenson Fensulfothion Fensulfothion-oxon Fensulfothion-oxon -sulfone fensulfothion-sulfone Fenthion Fenthion sulfoxide Fenthion-sulfone Fenuron Fenvalerate I Fenvalerate II {CAS # 51630-58-1} Fepropimorph Fipronil Fipronil, desulfinylFipronil-sulfide Fipronil-sulfone Flamprop-isopropyl Flamprop-methyl Fluacrypyrim Fluazifop-p-butyl Fluazinam Fluazolate Flubenzimine Fluchloralin Flucythrinate I Flucythrinate II {CAS # 70124-77-5} Fludioxonil Flufenacet Flumetralin Flumiclorac-pentyl Flumioxazin Fluometuron Fluoranthene Fluorene Fluorodifen Fluoroglycofen-ethyl Fluoroimide Fluotrimazole
Fluoxastrobin cisFluquinconazole Flurenol-butyl ester Flurenol-methylester Fluridone Flurochloridone I Flurochloridone II {CAS # 61213-25-0} Flurochloridone, deschloroFluroxypyr-1-methylheptyl ester Flurprimidol Flurtamone Flusilazole Fluthiacet-methyl Flutolanil Flutriafol Fluvalinate-tau-I Fluvalinate-tau-II {CAS # 102851-06-9} Folpet Fonofos Formothion Fosthiazate I Fosthiazate II {CAS # 98886-44-3} Fuberidazole Furalaxyl Furathiocarb Furilazole Furmecyclox Halfenprox Haloxyfop-methyl Heptachlor Heptachlor epoxide isomer A Heptachlor exo-epoxide isomer B Heptenophos Hexabromobenzene Hexachlorobenzene Hexachlorophene Hexaconazole Hexazinone Hexestrol Hydroprene Imazalil Imazamethabenz-methyl I Imazamethabenz-methyl II {CAS # 81405-85-8} Imibenconazole Imibenconazole-desbenzyl
Indeno[1,2,3-cd]pyrene Indoxacarb and Dioxacarb decomposition product [Phenol, 2-(1,3-dioxolan-2-yl)-] Ioxynil Ioxynil octanoate Ipconazole Iprobenfos Iprodione Iprovalicarb I Iprovalicarb II {CAS # 140923-25-7} Irgarol Isazophos Isobenzan Isobornyl thiocyanoacetate Isocarbamide Isocarbophos Isodrin Isofenphos Isofenphos-oxon Isomethiozin Isoprocarb Isopropalin Isoprothiolane Isoproturon Isoxaben Isoxadifen-ethyl Isoxaflutole Isoxathion Jasmolin I Jasmolin II Jodfenphos Kinoprene Kresoxim-methyl Lactofen Lenacil Leptophos Leptophos oxon Lindane Linuron Malathion Malathion-o-analog MCPA methyl ester MCPA-butoxyethyl ester MCPB methyl ester m-Cresol Mecarbam
Mecoprop methyl ester Mefenacet Mefenpyr-diethyl Mefluidide Menazon Mepanipyrim Mephosfolan Mepronil Metalaxyl Metamitron Metasystox thiol Metazachlor Metconazole I Metconazole II {CAS # 125116-23-6} Methabenzthiazuron [decomposition product] Methacrifos Methamidophos Methfuroxam Methidathion Methiocarb Methiocarb sulfone Methiocarb sulfoxide Methomyl Methoprene I Methoprene II {CAS # 40596-69-8} Methoprotryne Methoxychlor Methoxychlor olefin Methyl (2-naphthoxy)acetate Methyl paraoxon Methyl parathion Methyl-1-naphthalene acetate Methyldymron Metobromuron Metolachlor Metolcarb Metominostrobin (E) Metominostrobin (Z) {CAS # 133408-50-1} Metrafenone Metribuzin Mevinphos Mirex Molinate Monalide
Monocrotophos Monolinuron Musk amberette Musk Ketone Musk Moskene Musk Tibetene (Moschustibeten) Musk xylene Myclobutanil N,N-Diethyl-m-toluamide N-1-Naphthylacetamide Naled Naphthalene Naphthalic anhydride Naproanilide Napropamide Nicotine Nitralin Nitrapyrin Nitrofen Nitrothal-isopropyl N-Methyl-N-1-naphthyl acetamide Nonachlor, cisNonachlor, transNorflurazon Norflurazon, desmethylNuarimol o,p'-DDD o,p'-DDE o,p'-DDT Octachlorostyrene o-Dianisidine o-Dichlorobenzene Ofurace Omethoate o-Phenylphenol Orbencarb ortho-Aminoazotoluene Oryzalin Oxabetrinil Oxadiazon Oxadixyl Oxamyl Oxycarboxin Oxychlordane Oxydemeton-methyl Oxyfluorfen 15
p,p'-DDD p,p'-DDE p,p'-DDM [bis(4-chlorophenyl)methane] p,p'-DDT p,p'-Dibromobenzophenone p,p'-Dicofol Paclobutrazol Paraoxon Parathion PBB 52 Tetrabrombiphenyl PBB 101 PBB 15 PBB 169 Hexabrombiphenyl PCB 101 PCB 105 PCB 110 PCB 118 PCB 126 PCB 127 PCB 131 PCB 136 PCB 138 PCB 153 PCB 169 PCB 170 PCB 180 PCB 30 PCB 31 PCB 49 PCB 77 PCB 81 p-Dichlorobenzene Pebulate Penconazole Pendimethalin Pentachloroaniline Pentachloroanisole Pentachlorobenzene Pentachloronitrobenzene Pentachlorophenol Pentanochlor Permethrin I Permethrin II {CAS # 52645-53-1} Perthane Phantolide Phenamiphos 16
Phenanthrene Phenanthrene-d10 Phenkapton Phenol Phenothiazine Phenothrin I Phenothrin II Phenoxyacetic acid Phenthoate Phorate Phorate sulfone Phorate sulfoxide Phorate-oxon Phosalone Phosfolan Phosmet Phosphamidon I Phosphamidon II {CAS # 13171-21-6} Phthalide Phthalimide Picloram methyl ester Picolinafen Picoxystrobin Pindone Piperalin Piperonyl butoxide Piperophos Pirimicarb Pirimiphos-ethyl Pirimiphos-methyl Plifenat p-Nitrotoluene Potasan Prallethrin, cisPrallethrin, trans- {CAS # 23031-36-9} Pretilachlor Probenazole Prochloraz Procymidone Prodiamine Profenofos Profenofos metabolite (4-Bromo2-chlorophenol) Profluralin Prohydrojasmon I Prohydrojasmon II {CAS # 158474-72-7}
Promecarb Promecarb artifact [5-isopropyl3-methylphenol] Prometon Prometryn Propachlor Propamocarb Propanil Propaphos Propargite Propargite metabolite [Cyclohexanol, 2-(4-tert-butylphenoxy)] Propazine Propetamphos Propham Propiconazole-I Propiconazole-II {CAS # 60207-90-1} Propisochlor Propoxur Propyzamide Prosulfocarb Prothioconazole-desthio Prothiofos Prothoate Pyracarbolid Pyraclofos Pyraflufen-ethyl Pyrazon Pyrazophos Pyrazoxyfen Pyrene Pyrethrin I Pyrethrin II Pyributicarb Pyridaben Pyridaphenthion Pyridate Pyridinitril Pyrifenox I Pyrifenox II {CAS # 88283-41-4} Pyriftalid Pyrimethanil Pyrimidifen Pyriminobac-methyl (E) Pyriminobac-methyl (Z) {CAS # 136191-64-5}
Pyriproxyfen Pyroquilon Quinalphos Quinoclamine Quinoxyfen Quintozene metabolite (pentachlorophenyl methyl sulfide) Quizalofop-ethyl Rabenzazole Resmethrin Resmethrine I Resmethrine II {CAS # 10453-86-8} Rotenone S,S,S-Tributylphosphorotrithioate Schradan Sebuthylazine Sebuthylazine-desethyl Secbumeton Silafluofen Silthiopham Simazine Simeconazole Simetryn Spirodiclofen Spiromesifen Spiroxamine I Spiroxamine II {CAS # 118134-30-8} Spiroxamine metabolite (4-tert-butylcyclohexanone) Sudan I Sudan II Sudan Red Sulfallate Sulfanilamide Sulfentrazone Sulfotep Sulfur (S8) Sulprofos Swep Tamoxifen TCMTB Tebuconazole Tebufenpyrad Tebupirimifos Tebutam Tebuthiuron
Tecnazene Tefluthrin, cisTemephos Terbacil Terbucarb Terbufos Terbufos-oxon-sulfone Terbufos-sulfone Terbumeton Terbuthylazine Terbuthylazine-desethyl Terbutryne Tetrachlorvinphos Tetraconazole Tetradifon Tetraethylpyrophosphate (TEPP) Tetrahydrophthalimide, cis-1,2,3,6Tetramethrin I Tetramethrin II {CAS # 7696-12-0} Tetrapropyl thiodiphosphate Tetrasul Thenylchlor Theobromine Thiabendazole Thiazopyr Thifluzamide Thiofanox Thiometon Thionazin Thymol Tiocarbazil I Tiocarbazil II {CAS # 36756-79-3} Tolclofos-methyl Tolfenpyrad Tolylfluanid Tolylfluanid metabolite (DMST) Tolyltriazole [1H-Benzotriazole, 4-methyl-] Tolyltriazole [1H-Benzotriazole, 5-methyl-] Tonalide Toxaphene Parlar 26 Toxaphene Parlar 50 Toxaphene Parlar 62 trans-Chlordane Transfluthrin Traseolide Triadimefon
Triadimenol Tri-allate Triamiphos Triapenthenol Triazamate Triazophos Tributyl phosphate Tributyl phosphorotrithioite Trichlamide Trichlorfon Trichloronate Triclopyr methyl ester Triclosan Triclosan-methyl Tricresylphosphate, metaTricresylphosphate, orthoTricresylphosphate, para Tricyclazole Tridemorph, 4-tridecylTridiphane Trietazine Triethylphosphate Trifenmorph Trifloxystrobin Triflumizole Trifluralin Triphenyl phosphate Tris(2-butoxyethyl) phosphate Tris(2-chloroethyl) phosphate Tris(2-ethylhexyl) posphate Triticonazole Tryclopyrbutoxyethyl Tycor (SMY 1500) Uniconizole-P Vamidothion Vernolate Vinclozolin XMC (3,4-Dimethylphenyl N-methylcarbama XMC (3,5-Dimethylphenyl N-methylcarbama Zoxamide Zoxamide decomposition product
17
Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2006 Printed in the USA April 18, 2006 5989-5076EN
Screening for Hazardous Chemicals in Homeland Security and Environmental Samples Using a GC/MS/ECD/FPD with a 731 Compound DRS Database Application Note
Homeland Security, Environmental
Authors
Bruce Quimby Mike Szelewski Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808 UASA
interpretation of the MS data, especially in samples with high matrix contamination. The combination of selective GC detectors, SIM/Scan, and deconvolution makes a very powerful hazardous chemical analysis system that shows significant progress toward the above goals.
Introduction
In recent years, there has been increasing concern over the release of hazardous chemicals through either accidental or intentional acts. Both the homeland security and environmental communities recognize the need for preparing analytical laboratories that can respond quickly to such incidents. The terms toxic industrial chemicals/toxic industrial materials (TIC/TIM) are used in homeland security to refer to hazardous chemicals, while the environmental community uses different terminology like hazardous materials. In either case, the challenge is to develop laboratory methods with the capability of identifying any hazardous chemical(s) involved in an incident and to be able to measure its concentration in collected samples. There are several significant challenges to face when developing methods for this analysis. The methods must able to: Rapidly and accurately identify the specific toxic agents involved Measure concentration correctly at high levels of agent at the epicenter (high dynamic range) Measure concentration correctly at low levels of agent at perimeters and during decontamination (low detection limits)
Abstract
Response to homeland security or environmental incidents involving hazardous chemicals requires first, the rapid and accurate identification of the chemical agent(s) involved and second, the quantitative measurement of that agent in large numbers of samples to aid in managing the response. Given the unknown nature of the analytes and the complexity of matrices that could be encountered, developing analytical methods for this analysis is challenging. The approach described in this work uses a gas chromatography/mass spectrometry (GC/MS) system with a micro-fluidic splitter added to the end of the column. The splitter divides the column effluent between the MS and either a dual-wavelength flame photometric detector (DFPD) or a micro electron-capture detector (ECD) and a single-wavelength FPD. This approach allows the simultaneous collection of MS and two channels of selective GC detector data from a single injection. This multisignal configuration provides: full-scan MS data for library searching, selective ion monitoring (SIM) data for trace analysis, ECD and FPD data for excellent selectivity and sensitivity in complex matrices. The systems use retention time locking (RTL) to produce retention times (RTs) that precisely match those in a 731 compound database of hazardous chemicals. Deconvolution Reporting Software (DRS) is used to provide fast and accurate
Be highly selective over matrix interferences (wood smoke, fuels, burning tires, etc.) to minimize both false positives and false negatives Indentify as many toxic agents as possible Handle large numbers of samples It is clear that there is no single analytical technique that can be used for detecting all possible hazardous chemicals. However, one technique that is widely applicable for the identification and measurement of broad classes of hazardous chemicals is GC/MS. GC/MS is widely used in laboratories worldwide for the analysis of thousands of different chemicals. GC/MS methods are typically developed to analyze between 10 and 100 individual compounds. A target compound is deemed to be present if the target ion and two or three qualifier ions, with specific abundance ratios, fall within a defined RT window. The identity of the target may be further confirmed by comparison of the scan at the apex of the peak with a library reference spectrum. Matrix interferences are usually minimized by optimizing a combination of the sample preparation, GC, and MS parameters. Since most methods only deal with at most a few matrix types, the ions chosen for identification purposes can be selected such that they are minimized in the matrix. With the limited number of targets addressed by the method, recalibration of response factors, RTs, and qualifier ion abundance ratios can be accomplished with the injection of a few calibration mixtures. General screening methods for very large numbers of targets in widely varying and complex matrices offer a new set of challenges for the method developer. When screening for hundreds of targets, several factors must be addressed: Use of sample preparation to reduce matrix interferences is now significantly limited because rigorous cleanup steps may unintentionally remove targets. This reduced level of cleanup can result in significantly higher levels of matrix interferences to contend with. Recalibration of response factors, RTs, and qualifier abundance ratios is difficult or impossible because of the large number of targets. The methods may be deployed in laboratories without access to standards for all of the targets.
The time required for data review of hundreds of targets in complex matrices can become unmanageably large. Even with a very large database of targets, it is possible that hazardous chemicals not in the target list could be present in a sample. Recently, several techniques have become available to help address the above set of challenges. RTL produces RTs that precisely match from instrument-to-instrument and to those in a database [1]. This eliminates the need for recalibration of the individual RTs and timed events. The introduction of reliable and inert microfluidic splitters allows for the simultaneous collection of mass spectral data and, for example, phosphorus, sulfur, and/or electron capture data [2]. The selective detector chromatograms can highlight suspect compounds even if they are not in the MS target list. They can also offer an alternative means for quantitation of target analytes. The introduction of the synchronous SIM/Scan feature allows for the simultaneous acquisition of both full scan and SIM data from the same injection [2, 3]. The scan data can be used for screening the full list of targets in the database while the SIM data looks for a high priority subset of compounds down to very low levels. One of the most significant tools developed for dealing with complex matrices is Agilents Deconvolution Reporting Software (DRS) [4]. It uses advanced computational techniques to extract the spectra of targets from those of overlapped interference peaks. It then compares the extracted spectrum with a library to determine if the target is present. Any hits are confirmed by searching against the main NIST MS reference library. This process is automated and provides significant time savings in data interpretation. Since it deals with the entire spectrum instead of just four ions, DRS can often correctly identify a target in the presence of interferences where the typical approach would fail. The use of DRS substantially reduces the number of both false positives and false negatives. This application note describes the combination of the above techniques with a database of 731 hazardous chemicals, the Agilent Hazardous Chemical DBL (HCD), to be used for screening purposes. The compounds were chosen because of their significance in environmental or food safety analysis. The reasoning is that if the materials are manufactured in significant quantities and are toxic, they would be likely to appear in an
environmental method. The pesticides are included because many exhibit toxicity. The list is comprised of: Chlorinated Dioxins and Furans: EPA 8280A, 10 compounds Polychlorinated biphenyls: EPA 8082, 19 compounds Volatiles: EPA 502/524, 60 compounds Semivolatiles: EPA 8270C Appendix IX, 140 compounds Pesticides: Agilent RTL Pesticide Database (adapted), 567 compounds Total: 796 compounds, with 65 compounds in two groups, or 731 individual compounds The names of all the compounds in the database are listed in Appendix A at the end of this note. The above list by no means contains all of the hazardous chemicals that could be encountered. However, it does screen for a large number of known hazards and with the addition of selective detection can highlight other nontarget compounds that may be of interest.
and where the fuel components are not of interest. In the examples shown below the database with hydrocarbons removed was used, since fuels were used as prototype matrices.
System Configuration
The system configurations used are shown in Figure 1A and 1B.
A
Auto-sampler Phosphorus FPD
3-Way effluent splitter with makeup
AUX EPC 3.8 psig
ECD
Column 6890N GC
5975 Inert MSD
The chromatographic conditions chosen for development of the database are general in nature and are compatible with the analysis of other types of compounds beyond those in the table. For example, laboratories with access to calibration standards for chemical warfare agents (CWA) can add CWA data to the tables and screen for them as well. The RTs for compounds in the database were collected with the column outlet pressure at 3.8 psig using a microfluidic splitter. This was done to assure that the RTs observed during sample analysis would closely match those in the database when a microfluidic splitter or QuickSwap is used. The chromatographic conditions for the database were chosen to be compatible with the method translation technique. Constant pressure mode was used in the GC inlet so that method translation can be used to precisely time scale the methods for faster operation [5]. Provided with the Agilent Hazardous Chemicals DBL are the files to run the analysis precisely threefold (3X) and sevenfold (7X) faster than the primary database (1X). Also, each of the three-speed variations of the database are provided in two forms: one with the entire set of 731 compounds and one with the 36 aromatic hydrocarbons removed. The latter is provided for use with samples known to contain fuels
Auto-sampler Dual Flame Photometric Detector Sulfur Phosphorus AUX EPC 3.8 psig 3-Way effluent splitter with makeup
Column 6890N GC
5975 Inert MSD + Performance electronics
Figure 1.
System configurations. A). GC/MS/ECD/FPD system used for 1X and 3X screening analyses. B). GC/MS/DFPD system used for 7X screening analyses.
Key components are: Fast Oven The primary 1X method only requires the 120V oven. With the 6890N 240V oven (option 002), the screening analysis method can be run precisely three times faster (14.33 min) using a 15-m HP-5MS column. If the 240V GC is further equipped with SP1 2310-0236 (puts MSD interface in back of oven under rear injection port) and
3
using the G2646-60500 oven-insert accessory, the speed can be increased to seven times faster (6.14 min) with a 5-m HP-5MS column. Note that use of the oven insert prevents use of the front inlet and detector positions. Only one detector is available for splitting. The DFPD is a good choice for this configuration, as it uses only one detector position but generates two signals. ECD The 6890N Option 231 is a ECD. The signal from the electron capture detector (ECD) is collected, stored, and processed by the MS ChemStation simultaneously with the MS data. ECDs are selective in nature and exhibit very sensitive response to halogenated compounds, with detection limits below 1 pg for polyhalogenates. They also respond to several other functional groups like nitro compounds. They do, however, also respond to some fairly low-priority compounds, like phthalate esters. The ECD data can be used in several ways. Nontarget halogenated or nitro compounds are highlighted. The presence of an electrophore at the RT of an identified compound can be used to support confirmation of identity. The response on the ECD can be used for quantitative analysis, but only after calibration with a standard, as the response factors are compound dependent and can vary significantly with compound class. Single FPD The 6890N Option 240 is a single FPD. It is used to selectively detect either sulfur or phosphorus. The detector is usually run in the phosphorus mode to highlight such compounds as organophosphorus pesticides and nerve agents. In the phosphorus mode, the detector is highly selective (>106) with a very low (~0.050 pg) detection limits for phosphorus. The ability of the FPD to uncover nontarget organophosphorus compounds like new pesticides or designer nerve agents is especially helpful. The presence of phosphorus at the RT of an identified compound can be used to support confirmation of identity. Because the response per unit weight of phosphorus is relatively consistent from compound to compound, the FPD can be used for semi-quantitative analysis in situations where no calibration standard is available for an identified analyte. Dual FPD The 6890N Option 241 is a DFPD with two optical detection channels that measures sulfur and phosphorus simultaneously. The DFPD sulfur response is also selective (>104) and sensitive (detection limits <10 pg) , although not as much as phosphorus.
4
The sulfur signal is also quadratic with respect to the amount of sulfur injected. It is often used to detect sulfur-mustard agents and for confirmation of sulfur-containing pesticides. The response per unit weight of sulfur is relatively consistent from compound to compound, but varies more than that of the phosphorus signal. Microfluidic Splitter The 6890N Option 890 (3-way splitter) or Option 889 (2-way splitter) uses diffusion-bonded plate technology combined with metal column ferrules to make an inert, easy-to-use, leak free, hightemperature column-effluent splitter. The splitter uses Auxiliary EPC for constant pressure makeup (6890N Option 301). The Auxiliary EPC makeup can be pressure programmed at the end of the run to higher pressure, while at the same time the inlet pressure is lowered to near ambient. This causes the flow in the column to reverse direction, backflushing heavy materials out the split vent of the inlet. Backflushing can greatly reduce analysis times for samples that contain high-boiling matrix components [6]. The Aux EPC also allows column changing and maintainance without venting the MSD. When the column fitting is removed from the splitter, helium from the makeup supply purges the fitting, preventing air from entering the MSD. If the column is attached to the splitter but removed from the inlet, helium flows backwards through the column and out the inlet end. Inlet maintainance or column headtrimming can be done without cooling and venting the MSD to prevent sucking air into a hot source. MSD System The 5975 inert MSD with performance turbo (G3243A) or 5973N inert MSD with performance electronics and performance turbo (G2579A), EI (electron impact ionization mode) MSD is used. These configurations provide faster full scan rates while maintaining sensitivity. The scan rates are compatible with the narrower peaks generated by fast chromatography. The performance turbo pump is required to handle the higher flows associated with the screening method. Synchronous SIM/Scan The D.02.00 (or higher) revision of the Agilent MSD ChemStation is used because it supplies the synchronous SIM/Scan feature. SIM/Scan operates by collecting SIM data every other cycle and scan data on alternate cycles throughout the entire chromatogram. The signal-to-noise performance of the collected SIM and scan data is virtually identical to that obtained with SIM-only and scan-only
methods. As with conventional SIM methods, not all 731 targets can be monitored in a single run due to the required time separation between SIM groups. In general, the acquisition of SIM data is set up to collect high-priority targets at very low levels. Examples would be the chlorinated dioxins and CWAs. DRS Software (G1716AA) Spectral deconvolution of the MS data enables identification of analytes in the presence of overlapped matrix peaks [4]. This significantly reduces chromatographic resolution requirements, which allows detection of targets in higher levels of matrix or can be used with fast chromatography to shorten analysis times. DRS uses the AMDIS deconvolution program from NIST, originally developed for trace chemical-weapons detection in complex samples. DRS presents the analyst with three distinct levels of compound identification: ChemStation, based on RT and four-ion agreement AMDIS, based on cleaned spectra full-ion matching and locked RT NIST05 search using a 163000 compound library Hazardous Chemical DBL (G1671AA) This supplies the mass spectral library, method, and DRS files for the 731 compound-screening method.
volatility from gases to large polynuclear aromatic hydrocarbons (PAHs). Splitless injections are usually incompatible with the lowest boiling volatiles due to problems with the solvent. For low matrix samples where semivolatiles are of interest, splitless injections can be used. For ambient headspace analysis [7], the conditions are listed separately at the bottom of Table 1. The liner used for ambient headspace was 1-mm id straight through (no glass wool) and Siltek coated (Restek, part number 20973-214.5). The auto injector parameters are critical in ambient headspace and are listed in Table 1. The volatiles samples run by ambient headspace were prepared as described in Reference 7. While the targets in the table cover a very broad range of boiling points, it is usually not practical to screen for all of them in one run. This is because an analysis for semivolatile compounds would be done with a solvent that would occlude the lowest boiling volatiles in the table. Conversely, a method for injecting the lowest boiling compounds would usually not be suitable for the highest boiling. The MSD solvent delays listed in Table 1 are based on isooctane as the solvent in a semivolatiles analysis. If a lower boiling solvent is used, it may be possible to reduce these delays accordingly. Some of the target compounds were found to have sufficiently high boiling points to require higher inlet and detector temperatures. These were the higher molecular weight PAHs, the polychlorinated dioxins, and the polychlorinated furans. For these compounds the inlet temperature, MS source, and transfer line were also raised to 300 C. Without this increase in temperature, the compounds would exhibit tailing and in some cases reduction in signal. The trade-off with temperature is that the performance of some thermally labile compounds is degraded at the higher temperatures. The MSD data acquisition sampling rates listed in Table 1 are for scan mode only. For volatiles analysis, the scan rate is increased one step. It is also increased one step when SIM/Scan is used. In SIM/Scan mode the SIM dwell time was set to 40 milliseconds for each ion monitored. The microfluidic splitter parameters are chosen to provide the desired flow ratio between detectors while meeting the flow requirements of the detectors used. A primary consideration is to make sure that the flow to the MSD does not exceed ~4 mL/min while collecting analyte data. It was also desired to split the effluent equally between the DFPD and MSD in the 2-way split configuration. In the 3-way configuration, the split to the ECD was reduced
5
Instrument Operating Parameters

The instrument operating parameters used (unless noted otherwise) are listed in Table 1. These are starting conditions and may have to be optimized. The split/splitless injection port was used for all work described here. It was chosen for its flexibility, allowing splitless injections for clean samples and split injections for dirty or high-concentration samples. It is also compatible with column backflushing. For all cases (except ambient headspace), the inlet liner used was the 4-mm id Siltek Cyclosplitter (Restek, part number 20706-214.1). This inlet liner was found to be of low activity, as it does not contain glass wool. Proper mixing for split injections is done by the internal liner geometry. Except as noted, split injections with a split ratio of 10:1 were used. For high matrix samples, this roughly matches the amount of matrix injected with the column capacity. If excess amounts of matrix are injected, the RTs of targets can shift. Split injection is also the easiest and most reliable way of screening samples for analytes ranging in
Table 1.
Gas Chromatograph and Mass Spectrometer Conditions Original 1X Method 3X Method 7X Method
GC Agilent Technologies 6890N 7683 Autoinjector and Tray Inlet Mode Injection type Injection volume (uL) Inlet temp ( C) Pressure, nominal (psig) RT Locking compound RT Locking time (min) Split ratio Gas saver Gas type Oven Voltage (VAC) Initial oven temp (C) Initial oven hold (min) Ramp rate (C/min) Final temp (C) Final hold (min) Total run time (min) Equilibration time (min) Column Type Agilent part number Length (m) Diameter (mm) Film thickness (um) Outlet pressure (AUX EPC, psig) FPD or DFPD Type Temperature (C) Hydrogen flow (mL/min) Air flow (mL/min) Mode: Constant makeup flow Nitrogen makeup flow (mL/min) Data rate (Hz) ECD Temperature (C) Nitrogen makeup flow (mL/min) Mode: Constant makeup flow Data rate (Hz) AUX EPC Pressure Pressure (psig) Gas type EPC Split/Splitless Constant pressure Split 1.0 250 31.17 Tripropyl phosphate 12.874 10:1 Off Helium EPC Split/Splitless Constant pressure Split 1.0 250 23.96 Tripropyl phosphate 4.291 10:1 Off Helium EPC Split/Splitless Constant pressure Split 1.0 250 8.84 Tripropyl phosphate 1.839 10:1 Off Helium
120 or 240 40 2 10 300 15 43.00 0.5
240 40 0.667 30 300 5 14.33 0.5
240 (and pillow) 40 0.286 70 300 2.143 6.14 0.5
HP 5-MS inert 19091S-433i 30 0.25 0.25 3.8
HP 5-MS 19091S-431 15 0.25 0.25 3.8
HP 5-MS Custom 5 0.25 0.25 3.8
Single, Phosphorus 250 75 100 60 5
Single, Phosphorus 250 75 100 60 10
Dual, S and P 250 75 100 60 10
30 0 60 5
300 60 10
N/A N/A N/A
3.8 Helium
3.8 Helium
3.8 Helium
Table 1.
Gas Chromatograph and Mass Spectrometer Conditions (Continued)
MSD Agilent Technologies Tune file Mode Solvent delay (min) EM voltage Low mass (amu) High mass (amu) Threshold Sampling Scans/s Quad temp (C) Source temp (C) Transfer line temp (C) Splitter Type 6890N option number Flow ratio [Deactivated fused silica tubing] MSD restrictor length (m) MSD restrictor id (mm) FPD/DFPD restrictor length (m) FPD/DFPD restrictor id (mm) ECD restrictor length (m) ECD restrictor id (mm) Ambient Headspace Inlet Mode Injection type Inlet temp ( C) Pressure, nominal (psig) RT locking compound RT locking time (min) Split ratio Gas saver Gas type Autoinjector Sample washes Sample pumps Injection volume (L) Syringe size (L) PreInj Solvent A washes PreInj Solvent B washes PostInj Solvent A washes PostInj Solvent B washes Viscosity delay (s) Plunger speed Pre-injection dwell (min) Post-injection dwell (min) Sampling depth (mm) [critical!]
5975 inert MSD Atune.U Scan 2.20 Atune voltage 35 565 0 1 5.23 150 230 280
5975 inert MSD Atune.U Scan 0.82 Atune voltage 35 565 0 1 5.23 150 230 280
5973 inert with Performance Electronics Atune.U Scan 0.40 Atune voltage 35 565 0 0 9.46 150 230 280
3 way 890 1:1:0.1 MSD:FPD:ECD 1.44 0.18 0.53 0.18 0.51 0.10
3 way 890 1:1:0.1 MSD:FPD:ECD 1.44 0.18 0.53 0.18 0.51 0.10
2 way 889 1:1 MSD:DFPD 1.44 0.18 0.53 0.18 N/A N/A
EPC Split/Splitless Constant pressure Split 200 31.17 Tripropyl phosphate 12.874 1:1 Off Helium
0 3 50 100 0 0 1 3 5 Fast 0 0 20
to 1/10th that going to the MSD and FPD because of the extreme sensitivity of the detector. The lengths and diameters of the detector restrictors were calculated using the spreadsheet calculator included with the splitter. The peak recognition windows used in the Agilent ChemStation were set to 0.2 min and in AMDIS to 12 s. these values were found to be sufficiently wide enough to compensate for some RT drift yet narrow enough to minimize the number of false positives. The minimum match factors setting in AMDIS was set to 45. This value seemed to give the least number of false positives and false negatives.
relatively non-polar volatiles in water. It is convenient for labs that need to screen samples for volatiles but do not have a dedicated headspace sampler. The conversion from liquid sampling to ambient headspace simply requires changing the inlet liner and the autosampler syringe. Figure 2 shows the chromatograms from a run using the system in Figure 1A. A mixture of 14 halogenated volatiles was spiked into water at 2 ppm. Fifty microliters of the approximately 1 mL of headspace in the vial was injected. With the exception of peaks 3 and 4, which coelute, the compounds are well separated. The ECD chromatogram is inverted for comparison with the MS total ion chromatogram from the full-scan data. All of the volatiles respond on the ECD, although the response to compounds 1, 2, and 8 is significantly lower than for the rest of the compounds. In general with an ECD the response to a compound increases dramatically with the number of halogens in the molecule. Since none of the compounds contain phosphorus, there is no response on the FPD. Figure 3 shows the DRS report for the sample. For each compound identified, the RT, Chemical Abstracts number (CAS#), and compound name are listed. A line is generated in the report if a compound is found by either the Agilent ChemStation, AMDIS, or both.
Results
Volatiles To evaluate the HCD method for volatiles analysis, headspace injection was chosen. Headspace injections are usually done with an automated heated sampler specifically designed for the purpose. Ambient headspace [7] is a variant of the technique that uses a gastight syringe in the liquid autosampler and injects the headspace from a 2-mL vial. It is unheated, and is thus limited to compounds that are volatile at room temperature. Ambient headspace works well for the analysis of
3,4 7 2 1 5 6 8 10 9 11 12
TIC: 2ppmMIX 3_Only_simscan.D\DATA.MS
14 13
TIC
ECD
FPD
2 4 6 8 10 12
Peak identities 1) 1,2-Dichloroethane 2) 1,1-Dichloropropylene 3) 1,2-Dichloropropane 4) Trichloroethylene
5) 6) 7) 8) 9)
cis-1,3-Dichloropropylene trans-1,3-Dichloropropylene 1,1,2-Trichloroethane 1,3-Dichloropropane 1,2-Dibromoethane
10) 11) 12) 13) 14)
1,1,1,2-Tetrachloroethane 1,1,2,2-Tetrachloroethane 1,2,3-Trichloropropane 3-Chloro-1,2-dibromopropane Hexachlorobutadiene
Figure 2. 8
Ambient headspace analysis of volatile organics in water, spiked at 2 ppm per component.
Figure 3.
DRS report for the analysis in Figure 2.
The report shows that a compound has been determined as present by the Agilent ChemStation if a value appears in the Agilent ChemStation Amount column. This means the identification criteria set in the DATA ANALYSIS section of the method have been met. Typically the criteria are that the target ion is present and all three qualifier ions are present in ratios that fall within the percent uncertainty values for that compound. The Agilent ChemStation Amount listed is a very rough approximation of the amount of the compound, in nanograms, reaching the MS. This is based on the response factor originally observed when the HCD table data was collected. Since valid quantitation requires recent recalibration of response factors on the specific instrument used for analysis, the numbers in this column should never be used to report concentrations of identified analytes. The error in these values can easiliy be a factor of 10 or higher. The purpose of the listed values is to give an approximate amount that can be used to guide standard preparation for quantitative calibration of the compound, if needed. The match value listed under the AMDIS column is the degree to which the extracted (deconvolved) spectrum of the peak at that RT matched the spectrum in the HCD AMDIS target library. The higher this number, the better the spectra agree. The
column R.T. Diff sec. lists the difference in seconds between the observed RT and that in the AMDIS target library. The lower this number, the better the RTs agree. The NIST column lists the reverse-match quality of the extracted spectrum compared with the NIST05 main library spectrum with the same CAS#. The entry Hit Num. is the number of the hit in the NIST search results that has the same CAS# as the identified compound. The higher the reverse-match value and the lower the hit number, the better the extracted spectrum matches with NIST05. The NIST column serves as a second opinion on the identity of the extracted spectrum. The analysis in Figure 2 is of course an easy one, but serves to demonstrate how the system works. All 14 spiked compounds were found by both the Agilent ChemStation and AMDIS. The certainty of identification is very high because: The target ion and three qualifier ions are present in appropriate ratios and at the appropriate time as determined by the Agilent ChemStation The deconvolved spectrum and the RT at which it appears closely matches the data in the AMDIS target library.
9
The extracted spectrum of the identified compound also matches the spectrum with the same CAS # in the NIST05 library. The compounds all have a significant response on the ECD, as expected from their halogen content. To challenge the system in a more realistic way, the effect of matrix and dilution of the analytes was studied. Additional samples were prepared that contained: the same 2-ppm mixture of analytes plus 100 ppm of pump gasoline; 100 ppb of analytes only; and 100 ppb of analytes plus 100 ppm of pump gasoline. Figure 4 shows the chromatograms from the 100 ppb of analytes with 100 ppm of gasoline. The complexity of the TIC chromatogram illustrates the severe matrix challenge presented by the thousand-fold excess of gasoline. In the ECD chromatogram, interference peaks are now apparent. However, with the exception of peaks 1, 2, 8, and 12, all of the analytes peaks are still visible above the matrix interferences.
Table 2 summarizes the results from the matrix and dilution experiments. In the sample that was 2 ppm of analytes with 100 ppm of gasoline, the Agilent ChemStation (column labeled Quant) found all but two of the compounds. Those two compounds had qualifier ions out of range due to interferences from the matrix. AMDIS successfully found all 14 compounds. Also, with the exception of compound 8, all of the analytes were clearly visible above the matrix responses on the ECD chromatogram. In the sample that contained 100 ppb of analytes but without gasoline, quant found 7 of the 14 analytes. Using full-scan data, the signal to noise ratio for most of the analytes at the 100-ppb level is very low. This results in difficulties with finding the qualifier ions in ratios that fall within the specified uncertainty range in the quant calibration table. AMDIS found 11 of the 14 compounds. Peak 3 was not found due to a severe overlap with the coeluting peak number 4. Peaks 9 and 13 were missed by AMDIS because the signal to noise ratio was too low.
TIC
8 1 2 5 6 7 11
12
ECD
9 10 13 14
3,4
FPD
2 4 6 8 10 12
Peak identities 1) 1,2-Dichloroethane 2) 1,1-Dichloropropylene 3) 1,2-Dichloropropane 4) Trichloroethylene
5) 6) 7) 8) 9)
cis-1,3-Dichloropropylene trans-1,3-Dichloropropylene 1,1,2-Trichloroethane 1,3-Dichloropropane 1,2-Dibromoethane
10) 11) 12) 13) 14)
1,1,1,2-Tetrachloroethane 1,1,2,2-Tetrachloroethane 1,2,3-Trichloropropane 3-Chloro-1,2-dibromopropane Hexachlorobutadiene
Figure 4.
Ambient headspace analysis of volatile organics in water. Analytes at 100 ppb plus pump gasoline at 100 ppm.
10
Table 2.
Effect of Matrix and Concentration on DRS Results 2 ppm STD only 2 ppm STD with 100 ppm gasoline Quant AMDIS (ng) (match) 2.47 7.34 5.59 7.71 4.81 93 98 64 97 98 84 3.05 3.50 96 97 95 5.32 2.41 1.85 2.40 3.54 12 99 98 98 90 89 14 0.65 7 0.07 0.12 0.37 0.21 0.40 0.21 100 ppb STD only Quant (ng) AMDIS (match) 73 89 Overlap 90 88 53 72 66 S/N 89 48 79 S/N 75 10 0.36 7 0.31 0.19 0.14 0.22 0.30 0.23 100 ppb STD with 100 ppm gasoline Quant AMDIS (ng) (match) 65 85 Overlap 82 74 Overlap Overlap 46 66 88 53 75 59 52 11
RT (min) 1.491 1.536 1.793 1.863 2.317 2.658 2.735 2.938 3.250 4.003 5.151 5.283 8.208
Compound 1,2-Dichloroethane 1,1-Dichloropropylene 1,2-Dichloropropane Trichloroethylene cis-1,3-Dichloropropylene trans-1,3-Dichloropropylene 1,1,2-Trichloroethane 1,3-Dichloropropane 1,2-Dibromoethane 1,1,1,2-Tetrachloroethane 1,1,2,2-Tetrachloroethane 1,2,3-Trichloropropane 3-Chloro-1,2-dibromopropane
Peak Number 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Quant (ng) 2.27 7.60 4.92 7.58 4.39 3.30 2.82 3.39 2.60 5.15 2.38 1.89 1.62 16.46 14
AMDIS (match) 97 100 95 99 98 97 99 98 91 100 99 98 93 94 14
10.435 Hexachlorobutadiene Total Found
With 100 ppm of gasoline added to the 100-ppb sample, quant again found 7 of the 14 compounds and AMDIS again found 11 of the 14. Curiously, in both cases some of the compounds missed in the absence of matrix were now found. It is possible that the presence of matrix enhances the concentration of some of the analytes in the headspace. The compounds missed in quant were again the result of low signal to noise and/or interference. In AMDIS the three missed peaks were due to severe interferences from the gasoline. As indicated above, the ECD response from 10 of the 14 compounds was still visible above the peaks due to interferences. SIM/Scan The quant data in Table 2 was generated using full scan mode. Peak 13 was missed in quant due to low signal to noise ratio. SIM/Scan mode can be
used to collect SIM data simultaneously with the scan data. The 100 ppb plus 100-ppm gasoline sample was run in SIM/Scan mode with SIM groups for each of the 14 analytes. Figure 5 compares the target and qualifier extracted ion chromatograms in both modes with the ECD response for peak 13. The signal-to-noise (peak to peak) for the target ion increases from 34 in full scan mode to 433 in SIM mode. The peaks lost in quant due to low signal-to-noise were all recovered in SIM mode. This example demonstrates the power of SIM/Scan when looking for high-priority targets at low levels. If necessary, the ECD could also be used for quantitation, as it has a high signal to noise ratio and is free from interference.
11
Ion 157 Scan Ion 75 Scan
s/n (pk-pk) = 34
Ion 155 Scan
Ion 39 Scan s/n (pk-pk) = 433
Ion 157 SIM Ion 75 SIM Ion 155 SIM
Ion 39 SIM
ECD
s/n (pk-pk) = 122
8.0
8.1
8.2
8.3
8.4
Figure 5.
Target and qualifier extracted ion chromatograms for peak 13 (3-Chloro-1,2-dibromopropane) in Figure 4. SIM, scan, and ECD data collected simultaneously.
AMDIS Figure 6 illustrates the ability of AMDIS to clean the interference ions from the spectrum of an analyte. The raw spectrum at the top of Figure 6 was taken at the apex of peak 13 in the 100 ppb plus 100-ppm gasoline sample. When searched against the NIST05 library using the NIST search program, the actual compound (3-Chloro-1,2-dibromopropane) was the 70th hit in the search results. Using manual subtraction of nearby spectra in the Agilent ChemStation data analysis program improved the quality of the spectrum so that it was now the second hit when searched in NIST. This is a tedious process, however, when dealing with a large number of analytes. The spectrum as deconvolved by AMDIS is shown in Figure 6 above the
NIST05 library spectrum. When this spectrum is searched, it is the first hit in the results. The automated deconvolution provided by AMDIS saves an enormous amount of time in the data review process. Fast Methods When a retention time locked database is constructed, the RTs are (or at least should be) collected under the highest resolution conditions expected for the application. If the database is collected under constant pressure mode, method translation can then be used to adjust the speed of the method to meet the needs of different situations.
12
100
155 44 75 55 81 97 117 132 188
50
Raw spectrum
211 227 246 261 280 300
0 157 75 38 0 30 58 60 95 90 115 120 132 150 157 75 50 39 49 0 49 50 100 30 50 39 75 70 90 110 130 157 150 170 190 61 85 93 93 105 136 119 129
100
50
Manual subtraction
188 180
211 210
230
246 240
261 270
282
300 300
100
Deconvolved spectrum
188 187 199
NIST 05 library
Figure 6.
Comparison of raw, manually subtracted, AMDIS deconvoluted, and NIST05 reference spectra for peak 13 (3-Chloro-1,2-dibromopropane) in Figure 4.
The 3X method uses RTs in its database that are simply the RTs from the 1X method divided by exactly 3. The 7X method likewise uses RTs that are 1/7 of those in the original database. The quality of RTs matching between the two new faster methods and the new divided databases is demonstrated in Figure 7. Three different mixtures containing 13 chlorinated hydrocarbons and 36 pesticides were run with the two methods. The RTs were compared to those in the two new databases. The graph at the top of Figure 7 plots the database RT on the x-axis versus the difference of the measured RT from the database on the y axis. If the RT matching were perfect, the plot would be a straight horizontal line at zero height on the y axis. The maximum deviation from the table values for the 3X method was 0.047 min. The plot
indicates that a peak recognition window of 0.1 min should be sufficient. The maximum deviation in the 7X plot at the bottom of Figure 8 is +0.032 min indicating that the same peak recognition window could be used here as well. In general the RTs in scaled methods agree very well with the predicted RTs. The conditions for the two higher-speed methods were chosen to increase speed while maintaining the same column capacity. The capacity is important for both the dynamic range of quantitative measurements and for minimizing analyte RT shifts in samples with high levels of matrix. In gas chromatography, the well-known triangle of speed, resolution, and capacity dictates that if the capacity is to be maintained and the speed is to be increased, then the resolution will decrease.
13
0.100
3X Measured vs. Table
0.050 Difference (min)
0.000
_0.050
_0.100 0.0 2.0 4.0 6.0 Table RT (min) 8.0 10.0 12.0
0.100
7X Measured vs. Table
Difference (min)
0.050
0.000 _0.050 _0.100 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 Table RT (min)
Figure 7.
Difference between scaled HCD table and experimental retention times for 50 compound test set. Y axis is table value minus experimental, X axis is table RT. Top plot is 3X, bottom is 7X.
Figure 8 shows three sets of chromatograms using the HCD database at three different speeds. The sample consists of nine organophosphorus pesticides (identified in the caption to Figure 8) at 50 ppm and a matrix consisting of an equal volume mixture of gasoline, kerosene, and diesel fuel spiked at 50,000 ppm total mixture. The 1X and 3X data were collected on the three-way splitter instrument and the 7X was collected on the DFPD instrument. All nine compounds also contained sulfur as can be seen in the DFPD sulfur chromatogram at the bottom of Figure 8. Note that the sulfur tails somewhat compared to the phosphorus.
14
TIC: OP_50K_GKD1.D\DATA.MS
TIC
1 2 3 4 5 6 7 8 9
1X
P (FPD) ECD
4 8 12 16 20 24 28
TIC: OP_50K_GKD1.D\DATA.MS
TIC P (FPD) ECD

1 2 3 4 5 6 7 8 9
3X
TIC: 50_OP_50K_GKD.D\DATA.MS
TIC
7X
P (FPD) S (FPD)
1 2 3 4
Peak identities
1) 2) 3) 4)
O,O,O-triethyl phosphorothioate Thionazin Sulfotepp Phorate
5) 6) 7) 8) 9)
Dimethoate Disulfoton Methyl parathion Parathion Famphur
Figure 8.
Comparison of 1X, 3X, and 7X chromatograms. 1X and 3X were run on GC/MS/ECD/FPD system, 7X on GC/MS/DFPD.
15
Figure 9 expands the RT region of the phosphorous chromatogram containing peaks 3, 4, and 5 from Figure 8. The decrease in resolution with increasing speed is clearly evident. If only the standard target and three qualifier ion approach is used, the loss in resolution causes a significant problems. With the 1X method, all nine of the analytes are identified and eight false positives are reported. With the 3X method, all analytes are again found but now with 25 false positives. With the significantly decreased resolution of the 7X method, only seven of the nine analytes are identified and 48 false positives are reported.
3
The situation is much different when using the approach described here. Even in the worst situation, the 7X method, AMDIS finds all nine analytes with high-quality matches and only three false positives. The DRS report for the 7X analysis is shown in Table 3. To simplify the table, the 48 false positives that only appear in the quant column are not shown. The analyte compounds are shown in bold. All show close RT and high-quality spectral matches to both the AMDIS target library and to the NIST05 library.
1X
16.50
17.75
3X
5.50
5.916
7X
2.365
2.544
Peak identities
3) Sulfotepp 4) Phorate 5) Dimethoate Figure 9. Comparison of FPD phosphorus chromatograms from 1X, 3X, and 7X runs in Figure 8.
16
Table 3.
DRS Report for 7x Analysis of 50 ppm Pesticides In 50,000 ppm Gasoline/Kerosine/Diesel Matrix Agilent ChemStation amount (ng) 13.92 AMDIS match 71 69 46 64 55 91 88 90 84 92 92 91 93 RT Diff (sec.) 9.5 0.7 7.6 0.6 2.3 0.5 0.5 0.6 0.7 0.6 0.6 0.7 0.8 NIST reverse match 74 71 74 80 85 85 83 85 85 88 82 85 85 Hit number 50 1 1 3 1 1 1 1 1 1 1 1 1
RT 0.973 1.380 1.520 1.520 2.113 2.138 2.138 2.275 2.417 2.427 2.485 2.619 2.748 2.901 3.360
Cas # 98862 126681 94597 52417502 132649 90437 2131411 297972 3689245 298022 60515 298044 298000 56382 52857
Compound name Acetophenone O,O,O-triethyl phosphorothioate Safrole Benzeneacetaldehyde, ,2,5-trimethylDibenzofuran o-Phenylphenol Naphthalene, 1,4,5-trimethylThionazin Sulfotepp Phorate Dimethoate Disulfoton Methyl parathion Parathion (ethyl) Famphur (48 quant-only hits not shown)
0.35
89.2 23.31 27.34 22.7 25.12
The peak at 0.973 minutes is a reasonable spectral match to acetophenone, but the large time difference and being the 50th hit in the NIST search results suggests that this is not the compound. The peak at 1.520 min is a poor spectral match with a large time difference. The absence of a NIST reverse search and hit entry means that the listed compound was not in the top 100 hits in the NIST search. The next compound listed at 1.520 min is the top entry from the NIST search. It is quite clear that safrole is not present. The peak at 2.113 min, dibenzofuran, was not one of the analytes added to the sample. However, it probably is present in the diesel fuel matrix. Its presence is supported by both reasonably good spectral matches and close time matching with a database. The last extraneous peak at 2.138 min is also questionable. The time match is somewhat poor and the NIST reverse search suggests the identification is not correct.
All nine analytes are detected with the FPD on both the phosphorus and sulfur chromatograms. All analytes except peak 1 are detected selectively on the ECD as well. These results suggest that while the loss of resolution in going to 7X is unacceptable when using only conventional screening approaches, with the method discussed here, it is a viable option. By using the DRS report combined with the selective detector data, the number of false positives and false negatives are significantly reduced. For those situations where speed is a critical factor, for example in response to homeland security incidents, the fastest method may be the one of choice. For many laboratories, the 3X method would be an attractive choice. It has higher resolution than the 7X and higher speed than the 1X and still allows the use of two GC detectors in parallel with the MSD. It also only requires a 240V oven, not the repositioning of the MSD to the back position.
17
Conclusions
The systems described here offer several advantages when screening samples for the presence of hazardous chemicals. The advantages derive from a combination of techniques that result in both faster and more accurate screening results. Retention time locked target database of 731 hazardous chemicals for screening with MS Microfluidic splitter - using selective detection simultaneous with MS data for added confirmation, finding non-target suspect compounds, and alternate quantitation SIM/Scan - Acquire SIM data on high priority targets simultaneously with scan data. Saves time by eliminating need to run samples in both modes. DRS - automated deconvolution dramatically increases accuracy of target identification, even in the most challenging matrices. The reduction of data interpretation from hours to minutes is especially useful for response to hazardous chemical incidents. Fast chromatography using shorter columns, faster ovens, and backflushing to greatly reduce run times. This combination of techniques offers a viable solution to the hazardous chemicals challenge.
3. Chin-Kai Meng, Improving Productivity with Synchronous SIM/Scan, Agilent Technologies, publication 5989-3108EN www.agilent.com/chem 4. Philip Wylie, Michael Szelewski, Chin-Kai Meng, Christopher Sandy, Comprehensive Pesticide Screening by GC/MSD using Deconvolution Reporting Software, Agilent Technologies, publication 5989-1157EN www.agilent.com/chem 5. B. D. Quimby, L. M. Blumberg, M. S. Klee, and P. L. Wylie, Precise Time-Scaling of Gas Chromatographic Methods Using Method Translation and Retention Time Locking, Agilent Technologies, publication 5967-5820E www.agilent.com/chem 6. Michael J. Szelewski and Bruce Quimby, New Tools for Rapid Pesticide Analysis in High Matrix Samples, Agilent Technologies, publication 5989-1716EN www.agilent.com/chem 7. Michael J. Szelewski and Bruce D. Quimby, Ambient Headspace GC and GC-MSD Analysis of Nonpolar Volatiles in Water, Agilent Technologies, publication 5968-9455E www.agilent.com/chem

References
1. Retention time locking (RTL), Vince Giarrocco, Bruce Quimby, and Matthew Klee, Retention Time Locking: Concepts and Applications, Agilent Technologies, publication 5966-2469E www.agilent.com/chem 2. Chin Kai-Meng and Bruce Quimby, Identifying Pesticides with Full Scan, SIM, ECD, and FPD from a Single Injection, Agilent Technologies, publication 5989-3299EN www.agilent.com/chem
18
Appendix A
Lists of Compounds in Databases

Volatiles: EPA 502/524, 60 compounds
1,1,1,2-Tetrachloroethane 1,1,1-Trichloroethane 1,1,2,2-Tetrachloroethane 1,1,2-Trichloroethane 1,1-Dichloroethane 1,1-Dichloroethylene 1,1-Dichloropropylene 1,2,3-trichlorobenzene 1,2,3-Trichloropropane 1,2,4-trichlorobenzene 1,2,4-trimethylbenzene 1,2-Dibromoethane 1,2-dichlorobenzene 1,2-Dichloroethane 1,2-Dichloropropane 1,3,5-trimethylbenzene 1,3-dichlorobenzene 1,3-Dichloropropane 1,4-dichlorobenzene 2,2-Dichloropropane 2-chlorotoluene 3-Chloro-1,2-dibromopropane 4-chlorotoluene Benzene Bromobenzene Bromochloromethane Bromodichloromethane Bromoform Bromomethane Carbon Tetrachloride Chlorobenzene Chlorodibromomethane Chloroethane Chloroform Chloromethane cis-1,2-Dichloroethylene cis-1,3-Dichloropropylene Dibromomethane Dichlorodifluoromethane Ethylbenzene Hexachlorobutadiene Isopropylbenzene Methylene Chloride m-xylene Naphthalene n-butylbenzene n-propylbenzene o-Xylene p-isopropyltoluene p-xylene Styrene tert-butylbenzene Tetrachloroethylene Toluene trans-1,2-Dichloroethylene trans-1,3-Dichloropropylene Trichloroethylene Trichlorofluoromethane Vinyl chloride 4,4'-DDD 4,4'-DDE 4,4'-DDT 4-aminobiphenyl 4-bromophenyl phenyl ether 4-chloro-3-methylphenol 4-chloroaniline 4-chlorophenyl phenyl ether 4-nitroaniline 4-nitrophenol 4-nitroquinoline-1-oxide 5-nitro-o-toluidine 7,12-dimethylbenz[a]anthracene a,a-dimethylphenethylamine Acenaphthene Acenaphthylene Acetone Acetophenone Aldrin Alpha-BHC (alpha-HCH) Aniline Anthracene Aramite (total) Benz[a]anthracene Benzene Benzo[a]pyrene Benzo[b]fluoranthene Benzo[ghi]perylene Benzo[k]fluoranthene Benzyl alcohol Beta-BHC (beta-HCH) Bis(2-chloroethoxy)methane Bis(2-chloroethyl) ether Bis(2-chloroisopropyl) ether Bis(2-ethylhexyl)phthalate Butyl benzyl phthalate Chlorobenzilate Chrysene Delta-BHC (delta-HCH) Diallate (total) Dibenz[a,h]anthracene Dibenzofuran Dieldrin Diethyl phthalate Dimethoate Dimethyl phthalate Di-n-butyl phthalate
Semivolatiles: EPA 8270C Appendix IX, 140 compounds

1,2,4,5-tetrachlorobenzene 1,2,4-trichlorobenzene 1,2-dichlorobenzene 1,3,5-trinitrobenzene 1,3-dichlorobenzene 1,4-dichlorobenzene 1,4-naphthoquinone 1-naphthylamine 2,3,4,6-tetrachlorophenol 2,4,5-trichlorophenol 2,4,6-trichlorophenol 2,4-dichlorophenol 2,4-dimethylphenol 2,4-dinitrophenol 2,4-dinitrotoluene 2,6-dichlorophenol 2,6-dinitrotoluene 2-acetylaminofluorene 2-chloronaphthalene 2-chlorophenol 2-methyl-4,6-dinitrophenol 2-methylnaphthalene 2-naphthylamine 2-nitroaniline 2-nitrophenol 2-picoline 3,3'-dichlorobenzidine 3,3'-dimethylbenzidine 3-methylcholanthrene 3-nitroaniline
19
Di-n-octyl phthalate Dinoseb Diphenylamine Disulfoton Endosulfan I Endosulfan II Endosulfan sulfate Endrin Endrin aldehyde Ethyl methanesulfonate Famphur Fluoranthene Fluorene Gamma-BHC (lindane) Heptachlor Heptachlor epoxide -isomer B Hexachlorobenzene Hexachlorobutadiene Hexachlorocyclopentadiene Hexachloroethane Hexachlorophene Hexachloropropene Indeno[1,2,3-cd]pyrene Isodrin Isophorone Isosafrole Kepone m-cresol (3-methylphenol) m-dinitrobenzene Methapyrilene Methoxychlor Methyl methanesulfonate Methyl parathion Naphthalene Nitrobenzene N-nitrosodiethylamine N-nitrosodimethylamine N-nitrosodi-n-butylamine N-nitrosodi-n-propylamine N-nitrosodiphenylamine N-nitrosomethylethylamine N-nitrosomorpholine (4-nitrosomorpholine) N-nitrosopiperidine (1-nitrosopiperidine) N-nitrosopyrrolidine (1-nitrosopyrrolidine) O,O,O-triethyl phosphorothioate o-cresol (2-methylphenol) o-toluidine p-(dimethylamino)azobenzene Parathion (ethyl) p-cresol (4-methylphenol) Pentachlorobenzene Pentachloroethane Pentachloronitrobenzene Pentachlorophenol Phenacetin 20
Phenanthrene Phenol Phorate p-phenylenediamine Pronamide Pyrene Pyridine Safrole Sulfotepp Thionazin
Chlorinated Dioxins and Furans: EPA 8282, 19 compounds

2,3,7,8-Tetrachlorodibenzo-p-dioxin 1,2,3,7,8-Pentachlorodibenzo-p-dioxin 1,2,3,4,7,8-Hexachlorodibenzo-p-dioxin 1,2,3,4,6,7,8-Heptachlorodibenzo-p-dioxin Octachlorodibenzo-p-dioxin 2,3,7,8-Tetrachlorodibenzofuran 1,2,3,7,8-Pentachlorodibenzofuran 1,2,3,4,7,8-Hexachlorodibenzofuran 1,2,3,4,6,7,8-Heptachlorodibenzofuran Octachlorodibenzofuran
Polychlorinatedbiphenyls: EPA 8082, 19 compounds

2-chlorobiphenyl 2,3-dichlorobiphenyl 2,2',5-trichlorobiphenyl 2,4',5-trichlorobiphenyl 2,2',5,5'-tetrachlorobiphenyl 2,2',3,5'-tetrachlorobiphenyl 2,3',4,4'-tetrachlorobiphenyl 2,2',4,5,5'-pentachlorobiphenyl 2,2',3,4,5'-pentachlorobiphenyl 2,3,3',4',6-pentachlorobiphenyl 2,2',3,5,5',6-hexachlorobiphenyl 2,2',4,4',5,5'-hexachlorobiphenyl 2,2',3,4,5,5'-hexachlorobiphenyl 2,2',3,4,4',5'-hexachlorobiphenyl 2,2',3,4',5,5',6-heptachlorobiphenyl 2,2',3,4,4',5',6-heptachlorobiphenyl 2,2',3,4,4',5,5'-heptachlorobiphenyl 2,2',3,3',4,4',5-heptachlorobiphenyl 2,2',3,3',4,4',5,5',6-nonachlorobiphenyl
Pesticides: Agilent RTL pesticide database (adapted), 567 compounds

1,2,4-Trichlorobenzene 1,2-Dibromo-3-chloropropane 17a-Ethynylestradiol 2-(1-naphthyl)acetamide 2-(2-Butoxyethoxy)ethyl thiocyanate 2-(Octylthio)ethanol
2,3,4,5-Tetrachlorophenol 2,3,4,6-Tetrachlorophenol 2,3,5,6-Tetrachlorophenol 2,3,5-Trichlorophenol 2,3,5-Trimethacarb 2,3,5-Trimethylphenyl methyl carbamate (Trimethacarb) 2,3,7,8-Tetrachlorodibenzofuran 2,3,7,8-Tetrachlorodibenzo-p-dioxin 2,4,5-T methyl ester 2,4,5-Trichlorophenol 2,4,6-Trichlorophenol 2,4-D methyl ester 2,4-D sec-butyl ester 2,4-DB methyl ester 2,4-Dichlorophenol 2,4-Dichlorophenyl benzenesulfonate 2,4-Dimethylaniline 2,6-Dichlorobenzonitrile 2,6-Dimethylaniline 2-[3-Chlorophenoxy]propionamide 2-Ethyl-1,3-hexanediol 2-Hydroxyestradiol 2-Methylphenol 2-Phenoxypropionic acid 3,4,5-Trimethacarb 3,4-Dichloroaniline 3,5-Dichloroaniline 3-Chloroaniline 3-Hydroxycarbofuran 4,4'-Dichlorobenzophenone 4,6-Dinitro-o-cresol (DNOC) 4-Chloroaniline 4-Methylphenol 5,7-Dihydroxy-4'-methoxyisoflavone 9,10-Anthraquinone Acephate Acetochlor Acifluorfen methyl ester Alachlor Aldrin Allidochlor Ametryn Amidithion Aminocarb Amitraz Ancymidol Anilazine Aniline Atraton Atrazine Azaconazole Azamethiphos Azinphos-ethyl Azinphos-methyl Aziprotryne Azobenzene Barban
Benalaxyl Benazolin-ethyl Bendiocarb Benfluralin Benfuresate Benodanil Bentazone Bentazone methyl derivative Benthiocarb Benzo(a)pyrene Benzophenone Benzoylprop ethyl b-Estradiol BHC alpha isomer BHC beta isomer BHC delta isomer Bifenox Bifenthrin Binapacryl Bioallethrin Bioallethrin S-cyclopentenyl isomer Bioresmethrin Biphenyl Bis(2-ethylhexyl)phthalate Bisphenol A Bitertanol I Bitertanol II Bromacil Bromobutide Bromocyclen Bromophos Bromophos-ethyl Bromopropylate Bromoxynil Bromoxynil octanoic acid ester Buprofezin Butachlor Butamifos Butoxycarboxim Butralin Butyl benzyl phthalate Butylate Butylated hydroxyanisole Captafol Captan Carbaryl Carbetamide Carbofuran Carbofuran-3-keto Carbophenothion Carbosulfan Carboxin Chinomethionat Chloramben methyl ester Chloranocryl Chlorbenside Chlorbromuron
Chlorbufam Chlordecone Chlordimeform Chlorfenethol Chlorfenprop-methyl Chlorfenson Chlorfenvinphos Chlorflurecol-methyl ester Chlormefos Chlornitrofen Chlorobenzilate Chloroneb Chloropropylate Chlorothalonil Chlorotoluron Chlorpropham Chlorpyrifos Chlorpyrifos Methyl Chlorthal-dimethyl Chlorthiamid Chlorthion Chlorthiophos Chlorthiophos sulfone Chlorthiophos sulfoxide Chlozolinate cis-Chlordane Clomazone Coumaphos Crimidine Crotoxyphos Crufomate Cyanazine Cyanofenphos Cyanophos Cycloate Cycluron Cyfluthrin I Cyfluthrin II Cyfluthrin III Cyfluthrin IV Cyhalothrin I (lambda) Cymoxanil Cypermethrin I Cypermethrin II Cypermethrin III Cypermethrin IV Cyprazine Cyprofuram Cyromazine d-(cis-trans)-Phenothrin-I d-(cis-trans)-Phenothrin-II Dazomet Decachlorobiphenyl Deltamethrin Demephion
Demeton-S Demeton-S-methylsulfon Desbromo-bromobutide Desmedipham Desmetryn Dialifos Di-allate I Di-allate II Diamyl phthalate Diazinon Dibrom (naled) Dicamba Dicamba methyl ester Dicapthon Dichlofenthion Dichlofluanid Dichlone Dichlormid Dichlorophen Dichlorprop Dichlorprop methyl ester Dichlorvos Diclobutrazol Diclofop methyl Dicloran Dicrotophos Dicyclohexyl phthalate Dicyclopentadiene Dieldrin Diethatyl ethyl Diethofencarb Diethyl dithiobis(thionoformate) (EXD) Diethyl phthalate Diethylene glycol Diethylstilbestrol Difenoconazol I Difenoconazol II Diflufenican Dimefox Dimethachlor Dimethametryn Dimethipin Dimethoate Dimethylphthalate Dimethylvinphos(z) Dimetilan Di-n-butylphthalate Diniconazole Dinitramine Dinobuton Dinocap I Dinocap II Dinocap III Dinocap IV
21
Dinoseb Dinoseb acetate Dinoseb methyl ether Dinoterb Dinoterb acetate Di-n-propyl phthalate Dioxacarb Dioxathion Dioxydemeton-S-methyl Diphacinone Diphenamid Diphenylamine Dipropetryn Disulfoton Ditalimfos Dithiopyr Diuron Dodemorph I Dodemorph II Drazoxolon Edifenphos Endosulfan (alpha isomer) Endosulfan (beta isomer) Endosulfan ether Endosulfan lactone Endosulfan sulfate Endrin Endrin aldehyde Endrin ketone EPN Epoxiconazole EPTC Erbon Esfenvalerate Esprocarb Etaconazole Ethalfluralin Ethiofencarb Ethiolate Ethion Ethofumesate Ethoprophos Ethoxyquin Ethylenethiourea Etridiazole Etrimfos Famphur Fenarimol Fenazaflor Fenbuconazole Fenchlorphos Fenfuram Fenitrothion Fenobucarb Fenoprop Fenoprop methyl ester Fenoxycarb 22
Fenpropathrin Fenson Fensulfothion Fenthion Fenthion sulfoxide Fenuron Fenvalerate I Fenvalerate II Fepropimorph Flamprop-isopropyl Flamprop-methyl Fluazifop-p-butyl Flubenzimine Fluchloralin Flucythrinate I Flucythrinate II Flumetralin Fluometuron Fluorodifen Fluotrimazole Flurenol-butyl ester Flurenol-methylester Fluridone Flurochloridone I Flurochloridone II Fluroxypyr-1-methylheptyl ester Flusilazole Flutolanil Flutriafol Fluvalinate-tau-I Fluvalinate-tau-II Folpet Fonofos Formothion Fuberidazole Furalaxyl Furathiocarb Furmecyclox Heptachlor Heptachlor epoxide Heptachlor exo-epoxide isomer B Heptenophos Hexabromobenzene Hexachlorobenzene Hexachlorophene Hexaconazole Hexazinone Hexestrol Imazalil Ioxynil Iprobenfos Iprodione Isazophos Isobenzan Isobornyl thiocyanoacetate Isocarbamide Isodrin
Isofenphos Isomethiozin Isoprocarb Isopropalin Isoprothiolane Isoproturon Isoxaben Isoxathion Jodfenphos Kinoprene Lenacil Leptophos Leptophos oxon Lindane Linuron Malathion Malathion-o-analog MCPA methyl ester MCPB methyl ester m-Cresol Mecarbam Mecoprop methyl ester Mefenacet Mefluidide Menazon Mephosfolan Mepronil Metalaxyl Metamitron Metasystox thiol Metazachlor Methacrifos Methamidophos Methfuroxam Methidathion Methiocarb Methiocarb sulfone Methiocarb sulfoxide Methomyl Methoprene I Methoprene II Methoprotryne Methoxychlor Methyl paraoxon Methyl parathion Methyl-1-naphthalene acetate Methyldymron Metobromuron Metolachlor Metolcarb Metribuzin Mevinphos Mirex Molinate Monalide Monocrotophos Monolinuron
Myclobutanil N,N-Diethyl-m-toluamide N-1-Naphthylacetamide Naphthalic anhydride Napropamide Nicotine Nitralin Nitrapyrin Nitrofen Nitrothal-isopropyl N-Methyl-N-1-naphthyl acetamide Norflurazon Nuarimol o,p'-DDD o,p'-DDE o,p'-DDT Octachlorostyrene o-Dichlorobenzene Omethoate o-Phenylphenol Oryzalin Oxabetrinil Oxadiazon Oxadixyl Oxamyl Oxycarboxin Oxychlordane Oxydemeton-methyl Oxyfluorfen p,p'-DDD p,p'-DDE p,p'-DDT Paclobutrazol Paraoxon Parathion p-Dichlorobenzene Pebulate Penconazole Pendimethalin Pentachloroaniline Pentachloroanisole Pentachlorobenzene Pentachloronitrobenzene Pentachlorophenol Pentanochlor Permethrin I Permethrin II Perthane Phenamiphos Phenkapton Phenoxyacetic acid Phenthoate Phorate Phosalone Phosfolan Phosmet Phosphamidon I
Phosphamidon II Phthalide Picloram methyl ester Pindone Piperalin Piperonyl butoxide Piperophos Pirimicarb Pirimiphos-ethyl Pirimiphos-methyl Plifenat p-Nitrotoluene Pretilachlor Probenazole Prochloraz Procymidone Profenofos Profluralin Promecarb Prometon Prometryn Propachlor Propamocarb Propanil Propargite Propazine Propetamphos Propham Propiconazole-I Propiconazole-II Propoxur Propyzamide Prothiofos Prothoate Pyracarbolid Pyrazon Pyrazophos Pyrazoxyfen Pyributicarb Pyridaben Pyridaphenthion Pyridate Pyridinitril Pyrifenox I Pyrifenox II Pyrimethanil Pyroquilon Quinalphos Quinoclamine Quizalofop-ethyl Resmethrin S,S,S-Tributylphosphorotrithioate Sebuthylazine Secbumeton Simazine Simetryn Sulfotep
Sulfur (S8) Sulprofos Swep Tamoxifen TCMTB Tebuconazole Tebutam Tecnazene Temephos Terbacil Terbucarb Terbufos Terbumeton Terbuthylazine Terbutryne Tetrachlorvinphos Tetradifon Tetraethylpyrophosphate (TEPP) Tetramethrin I Tetramethrin II Tetrapropyl thiodiphosphate Tetrasul Thenylchlor Thiabendazole Thiofanox Thiometon Thionazin Tiocarbazil I Tiocarbazil II Tolclofos-methyl Tolylfluanid trans-Chlordane Triadimefon Triadimenol Tri-allate Triamiphos Triazophos Tributyl phosphate Tributyl phosphorotrithioite Trichlorfon Trichloronate Triclopyr methyl ester Tricyclazole Tridiphane Trietazine Triflumizole Trifluralin Tryclopyrbutoxyethyl Tycor (SMY 1500) Uniconizole-P Vamidothion Vernolate Vinclozolin
23
Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2006 Printed in the USA February 9, 2006 5989-4834EN
Analysis of Post-Harvest Fungicides and Their Metabolites in Citrus Fruits and Juices by Time-of-Flight and Ion Trap LC/MS Application
Food Safety
Authors
E. Michael Thurman and Imma Ferrer Pesticide Residue Research Group University of Almera, 04120 Almera, Spain Michael Woodman and Jerry Zweigenbaum* Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA
*Agilent contact
Introduction
Thiabendazole and imazalil are widely used for post-harvest treatment of citrus fruits (Figure 1) in order to preserve the fruit during the transport process, which may take from several days to several weeks. The fungicide treatment is of the fruit skin only. The maximum residue limits (MRLs) for imazalil and thiabendazole in Europe and the United States require that these compounds be monitored before consumption of the fruit [1, 2]. Since these compounds are used ubiquitously, they are frequently detected in imported citrus fruits [3]. Thus, many commercial samples must be inspected to ensure food safety. Additionally, imazalil is easily metabolized to 1-(2,4dichlorophenyl)-2-(1H-imidazole-1-yl)-1-ethanol, which is sometimes detected in citrus fruits [4]. In the United States, the sum of imazalil and its metabolite is regulated, so a method for the parent compound should also include its metabolite [5]. While papers have reported analytical LC/MS procedures for imazalil and thiabendazole, less is known about the metabolite of imazalil, and there are, as yet, no published reports using LC/MS accurate-mass analysis [6]. This application note describes a rapid and simultaneous method using electrospray liquid chromatography/time-of-flightmass spectrometry and liquid chromatography/ion trap mass spectrometry (LC/TOFMS and LC/ITMS) for the analysis of these two important postharvest fungicides in citrus. An earlier application note reported on the analysis of carbendazim and thiabendazole in fruits and vegetables and may be of interest as well [7].
Abstract
This application note applies both time-of-flight and ion trap liquid chromatography/mass spectrometry (LC/TOFMS and LC/ITMS) to determine two important post-harvest fungicides (imazalil and thiabendazole) and a metabolite of imazalil in oranges and in their juice. Included is both a detailed rapid procedure for sample preparation of citrus fruits and a solid-phase extraction for the isolation of these fungicides and metabolite from juice. Identification of these fungicides and metabolite is carried out using both accurate mass for elemental composition and fragmentation pathways with LC/ITMS to help elucidate pathways of degradation. The results of the analysis of actual samples from the marketplace for fruit and juice are included.
1 ppm
0 ppm
5. Shake the sample vigorously for 1 minute by using a Vortex mixer at maximum speed or hand shaking. 6. Centrifuge for 3 minutes at 3700 rpm.
Treated
Untreated
7. Add 5 mL of supernate into a 15-mL tube. 8. Add 250 mg of PSA adsorbent and 750 mg of MgSO4. 9. Vortex and shake for 20 s.
Cl O H2C N NH
10. Centrifuge again for 3 minutes at 3700 rpm.

HO Cl
11. Transfer 1.0 mL into a LC/MS vial. 12. Evaporate the 1.0 mL supernate to dryness and bring back up in 8%/92% methanol/water for LC/TOFMS analysis and ion trap analysis. LC/TOFMS analysis of fruit and vegetable extracts is performed by injecting 50 L. Nonfortified samples are analyzed directly at this same point by either LC/TOFMS or LC/ITMS. Fruit-Juice Extraction
Cl
m/z 297
NH
Cl
m/z 257
Imazalil
H2 N N N
Imazalil metabolite
m/z 202
Thiabendazole Figure 1. Structure of imazalil, thiabendazole, and an imazalil metabolite, which are post-harvest fungicides used on citrus (oranges and lemons), showing the [M+H]+ ion. Top panel compares treated and untreated oranges.
Solid-phase extraction (SPE) is used for the extraction of post-harvest fungicides from fruit juices and sodas based on juices. The cartridge of choice is the Accubond C-18 SPE cartridge (part number 188-1356). 1. Prepare the 5-mL SPE cartridge by taking 5 mL of methanol to wet the cartridge followed by 5 mL of deionized water to remove the methanol for good recovery of the fungicides. 2. Allow juice to stand, equilibrate to room temperature and degas prior to SPE, especially if bubbles are present. 3. Centrifuge and filter (0.2-micron TFPE filter to remove remaining solids) 5 mL of juice prior to dilution with 5 mL of deionized water. 4. Pass the juice through the cartridge and rinse with 2 mL of deionized water. 5. Elute with 5 mL of methanol. 6. Evaporate the methanol to 200 L under nitrogen. 7. Dilute 1:3 with deionized water before injection into the LC/TOFMS or LC/ITMS.
Experimental Methods
Fruit Extraction: QuEChERS QuEChERS (quick, easy, cheap, effective, rugged, and safe) is a method that is receiving wide acceptance for rapid extraction of pesticides in food [8]. 1. Weigh 15 g of a previously homogenized sample into a 40-mL Teflon centrifuge tube. 2. Add 15 mL of acetonitrile (containing 1% acetic acid). 3. Add 6 g of anhydrous MgSO4. 4. Add 2.5 g of NaAc3H2O (sodium acetate trihydrate).
LC/TOFMS Method
LC Pumps Column Mobile Phases Gradient Instrument Nebulizer Drying gas Voltages Reference masses Resolution Mass range Flow rate Agilent 1100 binary pumps, injection volume 50 L with standard Agilent 1100 ALS ZORBAX Eclipse XDB-C8, 4.6 mm 150 mm, 5 m, product number 993967-906 A = acetonitrile and B = 0.1% formic acid in water 5 minutes isocratic at 10% A, followed by a linear gradient to 100% A in 30 minutes at a flow rate of 0.6 mL/min Agilent LC/MSD TOF time-of-flight mass spectrometer with dual electrospray source, positive ESI, capillary 4000 V 40 psig 9 L/min, 300 C Fragmentor 190 V, skimmer 60 V, Oct DC1 37.5 V, OCT RF V 250 V m/z 121.0509 and 922.0098 9500500 at m/z 922.0098 501000 m/z Reference A Sprayer 2 is at a constant flow rate during the run
LC/ITMS Method
Chromatography LC Pumps Column Mobile phases Gradient Instrument Detection Identical to LC/MSD TOF methods for direct comparison of peaks Agilent 1100, injection volume 50 L ZORBAX Eclipse XDB-C8, 4.6 mm 150 mm, 5 m, product number 993967-906 A = acetonitrile and B = 0.1% formic acid in water 5 minutes isocratic at 10% A, followed by a linear gradient to 100% A in 30 minutes at a flow rate of 0.6 mL/min Agilent 1100 Series LC/MSD Trap. Positive ESI, Capillary 3200 V, Nebulizer 40 psig, drying gas 9 L/min, gas temp 300 C. Capillary exit 70 V, Skimmer 60 V Trap accumulation, 50,000 counts with maximum accumulation time of 200 ms

Imazalil in Oranges Figure 1 shows two oranges, one with a treatment of the post-harvest fungicides (imazalil and thiabendazole) and the other orange without treatment. Note the intense growth of mold when no treatment of post-harvest fungicides is used. The purpose of the post-harvest fungicides is to prevent serious mold formation during transport and sale of citrus fruits. This treatment is directed at the orange peel itself, rather than the entire fruit; therefore, it poses a lesser risk to the consumer. Figure 2 shows the LC separation and chromatogram of the orange-peel extract that contains two large peaks, which were suspected to be the post-harvest fungicides, imazalil and thiabendazole. They were identified using the following
technique. The accurate mass spectrum of each peak was taken and A+2 isotopes were examined (Figures 2 and 3). For example, the peak at 17.9 minutes was suspected to be a chlorinecontaining species based on the mass spectrum (Figure 3). In fact, the presence and number of chlorine atoms in the compound can be easily attained taking into account both the relative intensity of the 37Cl/35Cl signals and the accurate mass differences between the two masses. As can be seen in Figure 3, the accurate mass of the m/z 297 peak was 297.0556 with a 37Cl isotope signal of 299.0527, with a relative intensity of about two-thirds of the main peak. The mass difference between both signals is 1.9971, which is very near to the exact mass difference between 35Cl (34.9689) and 37Cl (36.9659) (1.9970). This evidence combined with peak area shows that the unknown compound unequivocally contains chlorine atoms.
3
3.9e7 3.8e7 3.6e7 3.4e7 3.2e7 3.0e7 2.8e7 2.6e7 2.4e7 Intensity, cps 2.2e7 2.0e7 1.8e7 1.6e7 1.4e7 1.2e7 1.0e7 8.0e6 6.0e6 4.0e6 2.0e6 0.0 2 4 6 8 10 12 14 2.1 2.4 2.7 7.6 16.7
17.9
Imazalil
Thiabendazole Imazalil degradate

14.9 18.2 21.0 20.4 21.5 24.8 26.0 26.4 22.7 23.9 25.4
16 18 Time, min
20
22
24
26
28
30
Figure 2.
LC/MSD TOF total ion chromatogram (TIC) of orange peel, note the two large peaks in the chromatogram are imazalil and thiabendazole.
Furthermore, the relative abundance of the isotopic signal indicates that the chlorine isotope is present with two atoms (the natural distribution of 35 Cl: 37Cl is nearly 3:1 (75.77% 35Cl; 24.23 % 37Cl), and two chlorine atoms give the characteristic pattern shown in Figure 3. This fact makes much easier the assignment of an elemental composition for the suspected species. Using the calculator tool of the TOF software, we set, as a stringent criterion, the number of chlorine atoms to two. Using an accuracy error threshold of 3 ppm (the maximum error of the LC/MSD TOF), only one formula
was found (C14H15N2OCl2, 0.3 ppm error). This formula was searched against the The Merck Index database. The use of the Merck Index is explained in an application note [9]. It has to be taken into consideration that the additional hydrogen atom present in the measured ions due to the positive ionization mode must be subtracted from the elemental compositions provided by the calculator tool before entering the formula into the database. The Merck Index gave a unique match with the formula for imazalil.
297.0556
A (Cl35)
Imazalil
A+2 (Cl35 and Cl37)

299.0527
Intensity
Accurate mass Accurate mass spectrum spectrum
299.0501
A+4 (Cl37 and Cl37)

301.0498
288
290
292
294
296
298
300 302 m/z, amu
304
306
308
310
312
314
Figure 3.
Imazalil detection in orange peel measured accurate mass, m/z = 297.0556 theoretical mass = 297.0556, accuracy 0.3 ppm, formula is C14H15N2OCl2.
Likewise the mass spectrum was taken for the peak at 7.6 minutes in Figure 2 and the spectrum is shown in Figure 4, m/z 202.0434. The relative abundance of the isotopic signal indicates that sulfur may be present with one atom. The relative intensity of the A+2 peak is 4.2% with a mass depletion of 1.996 daltons when going from 32S (31.9721) to 34S (33.9679). The mass spectrum shown in Figure 4 has a mass difference of 1.997 daltons and a peak height of 4.7%, which is
quite close to the expected values. Thus, these data indicate the presence of 1 sulfur atom. Entering a mass of m/z 202.0434 gives the unique formula of C10H8N3S, which is the formula for a protonated ion of thiabendazole, based on the Merck Index search. The next step is to examine the fragmentation of the molecules using the LC/MS ion trap for further structural confirmation.
1.9e5 1.8e5 1.7e5 1.6e5 1.5e5 1.4e5 Intensity, counts 1.3e5 1.2e5 1.1e5 1.0e5 9.0e4 8.0e4 7.0e4 6.0e4 5.0e4 4.0e4 3.0e4 2.0e4 1.0e4 186 188 190 192 194 196 198 200 202 m/z, amu 203.0458 204.0406 204 206 208 210 212 214 216 202.0434
Figure 4.
LC/MSD TOF mass spectrum of thiabendazole.
Fragmentation of Imazalil The next step in the discovery process was to search for characteristic fragment ions of the pesticide to confirm (or refute) its identity as imazalil. The fragmentation pathway of the large peak at 17.9 minutes was determined with LC/ITMS at MS3. The pathway is examined to see if reasonable structures could be drawn for the imazalil fragments based on the detection by LC/TOFMS. Figure 5 shows the structure for the proposed ions from the fragmentation of the unknown peak. The m/z 255 ion results from the loss of 42. MS/MS of the m/z 255 ion gives the m/z 187 and 159 ions, whose proposed structures are shown in Figure 5.
The m/z 159 ion is proposed as a doubly chlorinated tropylium ion. All the structures are consistent with the proposed identification as imazalil. Next is to examine the accurate mass spectrum from the LC/TOFMS. Here we found two fragment ions (m/z 255 and 159) with a relative abundance of respectively, 5% and 10% of that of the protonated molecule. The accurate mass of fragment 1 was 255.0084 (Table 1), and a 37Cl signal m/z of 257.0055 (data not shown). From both the relative intensity of these signals and the difference between the two masses, it can be deduced that the two chlorine atoms present in imazalil remain in this fragment.
5 6 4 2 0 100 200 300
297.0
Cl O H2C N
400 500 m/z 600 700 800 900
Cl Neutral loss of 42 H2C HO
NH
Cl
NH
Cl
m/z 255 at MS2
5000 4000 Intensity 3000 2000 1000 109.0 0 100 200 200.8 158.7
254.8
Imazalil, EE ion
EE ion fragment
MS/MS (297 255, 201,159)
Neutral loss of 68
NH
Cl
300 400 500 m/z 600 700 800 900
Cl _CO of 28
+
1.0 0.8 Intensity x104 0.6 0.4 0.2 0.0 100 200 300 400 500 m/z 600 700 800 900 158.7
m/z 159 at MS4
Double Cl trophilium Cl
Cl
m/z 187 at MS3
MS3 (255 187,159)

186.7
EE ion fragment
EE ion fragment
82.0
219.8
Figure 5.
Ion trap MS2 and MS3 to show fragmentation pathway of imazalil.
As can be seen in Table 1, the accurate mass of this fragment gave also three possible elemental compositions in the calculator tool. The first formula (C11H9N2OCl2), (0.8 ppm error) fitted with the proposed structure, involving a loss of C3H6 (accurate mass loss of 42.0469 versus 42.0465 u) in relation with the proposed parent species. Moreover, the accurate mass spectrum (relative intensity and mass differences) evidenced the presence of two chlorine atoms on fragment 2. The measured mass (m/z 158.9762) gave a unique elemental composition (C7H5Cl2), that corresponds to the formation of a doubly chlorinated tropylium fragment-ion. At this point, we combined the fragmentation information from the LC/MSD Trap, as shown in Figure 5 for the proposed structure of
imazalil, with the accurate mass information, that all show a perfect match for the detection of imazalil. These two lines of evidence (LC/TOFMS and LC/ITMS) provide fragment ions and information to confirm the identity of the proposed species based on fragmentation of the parent structure as imazalil, without the use of standards! Likewise, thiabendazole was examined by ion trap and gives a m/z 175 fragment ion, that fits with the structure of thiabendazole. The data are not shown here but were shown in a previous application note for thiabendazole [9]. We verified these determinations by standards for final confirmation.
Table 1.
LC/MSD TOF Fragment Ions of Imazalil Elemental compositions C14H15N2OCl2 C3H15N8O4Cl2 C11H19N2OSCl2 m/z calculated 297.0556 297.0588 297.0590 255.0086 255.0073 255.0060 158.9763 Error, ppm <0.1 8.3 9.0 0.8 4.3 9.4 0.6 Comments Imazalil Fragment 1 Fragment 2
m/z experimental 297.0556
255.0084
C11H9N2OCl2 C9H7N5Cl2 C8H11NO4Cl2
158.9762
C7H5Cl2
Imazalil Metabolite In the orange extract, we found a peak (at a retention time of 14.9 minutes, see Figure 2) with an ion with the same isotopic pattern as imazalil (two chlorine atoms). Taking into account that it had the same number of chlorine atoms, and it also appeared before imazalil, we considered that it could be related to imazalil, perhaps a metabolite. The accurate mass of the ion was 257.0245 with a 37 Cl signal of 259.0216 (see Figure 6). It gave a unique elemental composition in the calculator tool: C11H10N2OCl2 (0.8 ppm error, also see Table 2). For confirmation purposes, we searched for additional fragments but we did not find any characteristic fragment ion of that compound by TOF LC/TOFMS. The obtained elemental composition was then searched against two databases (The Merck Index and ChemIndex) with no positive
results. Then, we search the elemental composition against the Sigma-Aldrich commercial electronic catalogue, and we found a unique match: 1-(2,4dichloro-phenyl)-2-imidazol-1-yl-ethanol. The structure of this compound, shown in Figure 1, is compatible with the degradation of imazalil (see Figure 5 the m/z 255 ion). This suggests that the degradation product is the neutral species corresponding to the degradation of imazalil at the same site that cleaves to yield the m/z 255 fragment. A literature search on imazalils degradation products produced data and reports that agreed with our results (Imazalil-metabolite, R14832) [4]. In fact, in the United States, the sum of imazalil and imazalil metabolite is regulated, so the survey of residual imazalil metabolite is also required [4]. Finally, we confirmed the identity of the degradate by standard matching.
Table 2.
Accurate Mass and Elemental Composition of Imazalil, Imazalil Degradate, and Thiabendazole in Orange Peel Extracted with Methanol Selected ion [M+H]+ [M+H]+ [M+H]+ m/z experimental 297.0556 257.0245 202.0434 m/z calculated 297.0556 257.0243 202.0434 Error mDa +0.1 +0.2 +0.2
Compound Imazalil Imazalil Degradate Thiabendazole
Formula C14H14Cl2N2O C11H10Cl2N2O C10H7N3S
ppm 0.3 0.8 <0.1
2.4e5 2.3e5 2.2e5 2.1e5 2.0e5 1.9e5 1.8e5 1.7e5 1.6e5 1.5e5 Intensity, counts 1.4e5 1.3e5 1.2e5 1.1e5 1.0e5 9.0e4 8.0e4 7.0e4 6.0e4 5.0e4 4.0e4 3.0e4 2.0e4 1.0e4 0.0 235 240 245 250
257.0245
C11H11N2OCl2
HO N N Cl
1 (2,4 dichloro-phenyl)-2-imidazol-1-YL ethanol (Imazalil degradate)
[M+H]+ Protonated imazalil degradate Cl
259.0216
261.1115
255
260 m/z, amu
265
270
275
280
285
Figure 6.
LC/MSD TOF spectrum of imazalil degradate.
Identification of Imazalil in Orange Juice Figures 7 and 8 show the LC/TOFMS chromatogram (see peak at 18.0 minutes) and identification of imazalil (m/z 297.0556) in a juice product made from oranges. Several samples were analyzed and only this sample contained imazalil. No identifications of thiabendazole were observed. It is assumed that the production of juice usually does not involve the squeezing of the peel. However, in some types of juice machines the entire orange is squeezed without removing the peel. In this case, the opportunity for including fungicides in the juice may occur. Note that the LC/TOFMS analysis gave good results in spite of the complex matrix and complex chromatogram of the juice extract.
3.7e7 3.6e7 3.4e7 3.2e7 3.0e7 2.8e7 2.6e7 2.4e7 Intensity, cps 2.2e7 2.0e7 1.8e7 1.6e7 1.4e7 1.2e7 1.0e7 8.0e6 6.0e6 4.0e6 2.0e6 0.0 2 2.1 3.0
3.8
25.4 3.4 16.7 12.2 17.5 24.3 23.4 6.2 11.5 4.9 14.4 14.8 15.3 13.8 21.1 23.1 18.0 19.5 20.3 21.5 22.1 28.5 27.9 26.1
24.7
10
12
14
16 Time, min
18
20
22
24
26
28
30
Figure 7.
LC/MSD TOF TIC of orange juice extract.
9.2e4 9.0e4 8.5e4 8.0e4 7.5e4 7.0e4 6.5e4 6.0e4 Intensity, counts 5.5e4 5.0e4 4.5e4 4.0e4 3.5e4 3.0e4 2.5e4 2.0e4 1.5e4 1.0e4 5000.0 0.0 260 265 270 275 280 285 290
297.0556
299.0528
298.0585
295 m/z, amu
300
305
310
315
320
325
330
Figure 8.
LC/MSD TOF mass spectrum of imazalil in orange juice made from peel.
10
Conclusions
LC/TOFMS analysis is a powerful tool for identification of fungicides in fruits and juices and their metabolites and is a new tool for environmental food chemistry. Elemental composition and structural fragmentation of standards and fragment ions are possible, especially with ion trap MS/MS confirmation. The identification of metabolites is possible, in the case of imazalil, without standards, when both LC/TOFMS and LC/ITMS are used. These concentrations are equal to or lower than the EU directives for controlled fungicides in fruits and juices.

References
1. EU Food Directives 2002, 91/414/EEC. 2. Food Quality Protection Act, 1998 (FQPA). 3. M. Fernandez, R. Rodriguez, Y. Pico, J. Manes, (2001) J. Chromatogr. A 912, 301310. 4. N. Yoshioka, Y. Akiyama, K. Teranishi, (2004) J. Chromatogr. A, 1022, 145150. 5. http://www.epa.gov./pesticides/registration/ imazalil/ImazProductChem.pdf. 6. E. M. Thurman, Imma Ferrer, Jerry A. Zweigenbaum, Juan F. Garca-Reyes, Michael Woodman, and Amadeo R. Fernndez-Alba, (2005) Discovering Metabolites of Post-Harvest Fungicides in Citrus with LC/MS TOF and Ion Trap MS/MS, J. Chromatogr., in press. 7. Determination of Fungicides in Fruits and Vegetables by LC/MS/ESI/TOF and LC/MS Ion Trap, Agilent Technologies, publication 5989-2209EN www.agilent.com/chem 8. M. Anastassiades, S.J. Lehotay, D. Stajnbaher, and F.J. Schenck, (2003) J. AOAC Internat., 86, 412431. 9. Identification of Unknown Pesticides in Food Using Both LC/MSD TOF and Ion Trap MSn, Agilent Technologies, publication 5989-1924EN www.agilent.com/chem
11
Acknowledgments We acknowledge Amadeo Fernandez-Alba and Juan-Francisco Garcia-Reyes, Department of Hydrogeology and Analytical Chemistry, University of Almera, for help, ideas, and analytical support. Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. ZORBAX is a registered trademark of E.I. duPont de Nemours Co. Inc. Agilent Technologies, Inc. 2005 Printed in the USA August 24, 2005 5989-2728EN
Analysis of Terbuthylazine in Olive Oil by LC/Ion Trap MS and LC/Time-of-Flight MS Application
Food safety
Authors
Imma Ferrer and E. Michael Thurman University of Almera, 04120 Almera, Spain
Introduction
Olive oil is one of the most-used food products in Mediterranean countries. The positive effects of olive oil on health have prompted a demand for this product world-wide. Virgin olive oil is obtained from the fruit of the olive tree (Olea Europaea) exclusively by mechanical and physical processes without any further treatment, which may alter the olive oil quality. The most extensively applied agrochemicals in olive plantations of Mediterranean countries (Greece, Spain, and Italy) are herbicides and insecticides. Although herbicides are mainly applied to soils, some residues can persist to the harvest stage, thus contaminating the olives picked up from soil. This can result in the presence of trace amounts of these pesticides in olive oil. Consequently, both the European Union and the Codex Alimentarius Commission of the Food and Agriculture Organization (FAO) of the United Nations have established maximum pesticide residue levels in olives and olive oil. Currently, various olive oil pesticide residue regulatory programs are being carried out to update and to establish new and more stringent regulations concerning the maximum residue levels in these commodities [1]. This fact has fostered the development of more powerful analytical tools in order to provide enough sensitivity and selectivity to meet these requirements in food samples such as edible oils, which have a complex matrix due to the high fat content of the extracts obtained after the sample treatment step.
Abstract
This application note describes the use of liquid chromatography/ion trap mass spectrometry (LC/ITMS) and liquid chromatography/time-of-flight mass spectrometry (LC/TOFMS) for the identification and quantitation of terbuthylazine in olive oil samples. The method includes a sample treatment step based on a preliminary liquidliquid extraction, followed by matrix solid-phase dispersion (MSPD) using an aminopropyl-bonded silica as a sorbent material. A final clean-up step is performed with Florisil using acetonitrile as an eluting solvent. The analysis by ion trap was achieved in MS/MS mode, monitoring the characteristic fragment ion at m/z 174. The identification by LC/TOFMS was accomplished with the accurate mass (and the subsequently generated empirical formula) of the protonated molecule ([M+H]+ m/z 230), along with the accurate mass of the fragment ion and the characteristic chlorine isotope cluster present in terbuthylazine. The accuracy typically obtained was routinely better than 2 ppm. The method sensitivity, linearity, precision, accuracy, matrix effects, and limit of detection were also studied; they supported the potential of LC/TOFMS and LC/ITMS for the routine quantitative analyses of terbuthylazine in olive oil.
Many multiresidue procedures employing different clean-up techniques and a variety of detection methods are reported on the determination of pesticide residues in olive oil. The most commonly used methodology is based on GC after a comprehensive clean-up step, in most cases based on liquid-liquid partitioning or gel permeation chromatography to separate the low molecular mass pesticides from the higher molecular mass fat constituents of the oil, such as triglycerides [2, 3]. The preparation of oil samples for the determination of pesticides by GC requires the complete removal of the high-molecular-mass fat from the sample to maintain the chromatographic system in working order. Recently, a multiresidue method for the determination of triazines and organophosphorous pesticides using MSPD followed by GC/MS and ion trap MS techniques was reported [4]. In this work, we further explore the capabilities of LC/ITMS and LC/TOFMS techniques for the identification of terbuthylazine, one of the most common pesticides found in olive oil.
- Another 10 mL of acetonitrile saturated with petroleum ether was added to the petroleum ether extract. The mixture was shaken again for 3 minutes and the acetonitrile phase was collected and added to the previous one. - Finally, a 7-mL aliquot of the acetonitrile extract was transferred to a 10-mL glass test tube. The extract was then carefully evaporated down to a final volume of about 2 mL. This remaining extract was transferred to a glass mortar to be subjected to matrix solid-phase dispersion. 2. The MSPD: - The extract obtained in the liquid-liquid extraction step was homogenized with 2 g of aminopropyl-bonded sorbent (Bondesil-NH2, 40-m particle size, Varian Inc.) until a fine powder was obtained. - The mixture was transferred to a commercially available minicolumn containing 2 g of Florisil (12-mL Bond-Elut Varian minicolumn, Varian Inc.). This minicolumn was connected to a vacuum system for solid phase extraction adjusting the flow to 3 mL/min. - An elution step was carried out with 2 5 mL of acetonitrile. The final extract was evaporated until near dryness, then dissolved with 1:1 acetonitrile:water. - Prior to LC/ITMS and LC/TOFMS analysis, the extract was filtered through a 0.45-m PTFE filter (Millex FG, Millipore, Milford, MA, USA). Using the MSPD method, recoveries for terbuthylazine from olive oil samples were higher than 96% with a 6% relative standard deviation (RSD) (n = 5).
Olive Oil Extraction MSPD was used for the extraction of terbuthylazine from olive oil after a preliminary liquidliquid extraction with organic solvents. 1. The liquid-liquid extraction: - An aliquot of approximately 5 g (ca. 5.5 mL) of olive oil sample was mixed with 15 mL of petroleum ether, saturated with acetonitrile, in a 100-mL separatory funnel, in which a two-step liquid-liquid extraction was performed. - A solution of 25 mL of acetonitrile, saturated with petroleum ether, was added to the olive oil mix obtained previously. The funnel was shaken vigorously for 3 minutes, and the acetonitrile phase was separated from the petroleum ether phase.
Agilent 1100 Series LC/MSD TOF Methods

LC conditions LC Pumps were Agilent 1100 binary pumps Injection volume: 50 L with standard Agilent 1100 ALS Column: ZORBAX Eclipse XDB-C8, 4.6 mm 150 mm, 5 m, p/n 993967-906 Mobile phases: A = acetonitrile and B = 0.1% formic acid in water Gradient: 5 minutes isocratic at 10% A, followed by a linear gradient to 100% A in 25 minutes at a flow rate of 0.6 mL/min MS conditions LC/MSD TOF Source: Capillary: Nebulizer: Drying gas: Gas temp: Fragmentor: Skimmer: Oct DC1: Oct RF V: Reference masses: Resolution: Mass range: Reference A sprayer 2:
Positive ESI (electrospray ionization) 4000 V 40 psig 9 L/min 300 C 190 V 60 V 37.5 V 250 V m/z 121.0509 and 922.0098 9500 500 @ m/z 922.0098 501000 m/z Constant flow during the run
Agilent 1100 Series LC/MSD Trap Methods LC/MSD Trap Methods identical to LC/MSD TOF for direct comparison of peaks. LC Pumps: HP 1100 Inject vol: 50 L Column: ZORBAX Eclipse XDB-C-8, 4.6 mm 150 mm, 5 m, p/n 993967-906 Mobile phases: A = acetonitrile (ACN) B = 0.1% formic acid in water Gradient: 10% A, isocratic, for 5 minutes followed by linear gradient to 100% A in 25 min at a flow rate of 0.6 mL/min LC/MSD Trap Positive ESI: Nebulizer: Drying gas: Capillary exit: Skimmer: Trap accumulation: Capillary 3200 V 40 psig 9 L/min, gas temp 300 C 70 V 60 V 50,000 counts with maximum accumulation time of 200 ms

Identification of Terbuthylazine by LC/ITMS and LC/TOFMS Olive oil is one of the most difficult food matrices due to the presence of numerous interferences that show up in full-scan mode. For this reason, the ion trap method was optimized in MS/MS mode by isolating the precursor ion at m/z 230, which corresponds to [M+H]+. An isolation mass window of m/z 2 and optimized fragmentation amplitude was used in order to enhance the signal-to-noise (S/N) ratio. An amplitude voltage of 0.8 V was used to
obtain good fragmentation for terbuthylazine, which fragments by losing the terbutyl group (56) forming the m/z 174 fragment ion. Figure 1 shows the analysis of an olive oil sample (spiking level 0.025 mg/kg) by ion trap MS/MS. The extracted ion chromatogram (EIC) for m/z 174, the main fragment of terbuthylazine, as well as its mass spectrum is shown. The fragmentation occurs at the CN bond via the cleavage of the terbutyl group. This represents a characteristic fragmentation for this analyte, allowing for the structural confirmation of terbuthylazine in a relative complex matrix such as olive oil.
1.25
1.00 Intensity 105
0.75
0.50
0.25
0.00 0 5 10 15 20 25 Time [min]
MSPAC25.D: EIC 174 All 1.50
173.8
1.25 Cl Intensity 105 1.00 0.75 0.50 0.25 0.00 100 N N H N N N H _56
+MS2(230.0), 23.7 min #556
210.8
200 300 400 500 600 700 800 900 m/z
Figure 1.
Ion Trap MS/MS chromatogram corresponding to the analysis of a spiked olive oil sample with terbuthylazine (0.025 mg/kg). The EIC for m/z 174 and its corresponding spectrum are shown.
LC/TOFMS analyses were optimized in terms of fragmentation. The in-source collisionally induced dissociation (CID) fragmentation is greatly enhanced at high fragmentor voltages. This provides highly valuable structural information since the accurate mass of the characteristic fragment ion can be used along with that of the molecular ion for confirmation purposes. The relative abundances for both the main fragment and the protonated molecule of terbuthylazine are summarized in Table 1 at three different voltages: 160 V (low),
Table 1. m/z ion 230 [M+H]+ 174 [M+HC4H8]+ Effect of the Fragmentor Voltage on CID Fragmentation for LC/TOFMS Relative abundance 160 V 190 V 230 V 100 5 100 30 20 100
190 V (medium), and 230 V (high). In order to obtain sufficient sensitivity for quantitative purposes (using the protonated molecule) and additional qualitative spectral information provided by the fragment ions generated by in-source fragmentation, a value of 190 V was chosen for further analyses. The accurate masses for both the main fragment and the protonated molecule of terbuthylazine in a spiked olive oil matrix are shown in Table 2.
Table 2.
LC/TOFMS Accurate Mass Measurements for Terbuthylazine and its Main Fragment Ion in Olive Oil Matrix-Matched Standard (Fragmentor Voltage 190 V). Spiking Level: 0.025 mg/Kg Ion [M+H]+ Theoretical mass 230.1167 232.1137 174.0541 176.0511 Measured mass 230.1168 232.1134 174.0542 176.0511 Error mDa 0.1 0.3 0.1 0.05 ppm 0.4 1.5 0.6 0.3
Elemental composition C9H16N535Cl C9H16N537Cl C5H9N535Cl C5H9N537Cl
[M+HC4H8]+
Along with the accurate mass of the protonated molecule and the information provided by the fragments obtained with an optimized in-source fragmentation, terbuthylazine presents another feature which enables its identification; the presence of a chlorine atom. The accurate mass pattern of the 37 Cl isotope signal confirms that the peak unequivocally contains only one chlorine atom (Figure 2). Therefore, the accurate mass of each ion as well as the presence of the chlorine signal, together with the characteristic retention time represent sufficient information to unequivocally identify and confirm this specie in such complex matrices.
230.1168 9.50e4 8.50e4 7.50e4 6.50e4 5.50e4 4.50e4 3.50e4 2.50e4 1.50e4 5000.00
Intensity, counts
174.0542 176.0511
232.1134
60 80 100 120 140 160 180 200 220 240 260 280 m/z, amu
Figure 2.
LC/TOFMS total ion chromatogram (TIC) corresponding to the analysis of a spiked olive oil sample with terbuthylazine (0.025 mg/kg). The EIC for m/z 230 and its corresponding spectrum are also shown.
Analytical Performance The analytical performance of the proposed methods was studied in order to explore the ability to carry out quantitative analyses of terbuthylazine in these complex matrices with a high content of fat. The calibration was carried out using matrixmatched standards prepared by the extraction method based on MSPD described in the experimental section. Linearity was evaluated by analyzing solutions of matrix-matched standards at seven different concentration levels in the range 0.0050.5 mg/kg. Using ion trap the quantitation was performed in MS/MS mode of the m/z 230 ion. Using LC/TOFMS the quantitation was carried out using the peak area from the EIC of the protonated molecule with a mass window of 0.1 Da. As an example, Figure 3 shows the linear calibration curve obtained by LC/TOFMS for terbuthylazine in an olive oil matrix compared to the curve obtained in pure solvent.
The limits of detection (LOD) were estimated from the injection of matrix-matched standard solutions with concentration levels giving a S/N ratio of about 3. The results obtained are shown in Table 3. The LOD obtained are remarkable since they are far below the maximum residue level regulations established for these pesticides. In this sense, LC/TOFMS analyses benefits of the use of narrow mass windows for quantitation purposes, which results in enhanced S/N ratio, thus providing lower detection limits. This fact illustrates the analytical potential of the proposed method based on MSPD and LC/TOFMS for the analyses of pesticides in complex matrices with high content of fat.
40
Pure solvent
y = 70.697x + 0.5666 R2 = 0.9977
30
Spiked oil solvent

y = 60.442x + 0.7386 R2 = 0.9977
Area 10 E+06
20
10
0 0.0 0.1 0.2 0.3 Concentration (mg/kg) 0.4 0.5 0.6
Figure 3.
Calibration plot obtained from spiked olive oil samples versus solvent based samples by LC/TOFMS.
Table 3.
Analytical Parameters for the Analysis of Terbuthylazine in Olive Oil Samples by Ion trap MS/MS and LC/TOFMS Concentration Linearity LOD RSD (%) Method range (mg/kg) (R2) (g/kg) n = 5 LC/ITMS LC/TOFMS 0.0050.5 0.0050.5 0.991 0.995 0.2 1 5.5 2
overcome using the isotopic chlorine signal or a characteristic fragment not affected by other interfering species.
Conclusion
LC/ITMS and LC/TOFMS can be used for the identification and quantitation of terbuthylazine in olive oil samples after performing several preliminary sample treatments. The analysis by ion trap was achieved in MS/MS mode, monitoring the characteristic fragment ion at m/z 174. The identification by LC/TOFMS was accomplished with the accurate mass (and the subsequently generated empirical formula) of the protonated molecule ([M+H]+ m/z 230), along with the accurate mass of the fragment ion and the characteristic chlorine isotope cluster present in terbuthylazine. The accuracy typically obtained was routinely better than 2 ppm. The method sensitivity, linearity, precision, accuracy, matrix effects, and LOD were also studied and they supported the potential of LC/TOFMS and LC/ITMS for the routine quantitative analyses of terbuthylazine in olive oil.
Analysis of Olive Oil Samples To evaluate the effectiveness of the proposed method, it was applied to the analysis of olive oil samples. An example is shown in Figure 4 where traces of terbuthylazine were found in an olive oil extract; the EIC for m/z 230 is also shown. This is a real example illustrating the usefulness of the routine accurate mass measurement capabilities of LC/TOFMS, especially when analyzing traces of pesticides in complex samples such as olive oil. In this sense, the selectivity of LC/TOFMS relies on the resolving power of the instrument on the m/z axis, enabling discrimination between the target species and an isobaric interference within 0.05 Da of the mass difference (using 230 m/z, as example). In the case of terbuthylazine, it is easily
Figure 4.
Upper: LC/TOFMS TIC of a "positive" for terbuthylazine in an olive oil sample. Lower: EIC of terbuthylazine using a 20 mDa mass window.
References
1. Proposal for a Regulation of the European Parliament and of the Council on maximum residue levels of pesticides in products of plant and animal origin, COM/2003/0117/final COD2003/0052*/; www.europa.eu.int/pol/food/indexes.htm 2. Ch. Lentza-Rizos, E. J. Avramides, and F. Cherasco, (2001) J. Chromatogr. A 912: 135142. 3. J. J. Vreuls, R. J .J. Swen, V. P. Goudriaan, M. A. T. Kerkhoff, G. A. Jongenotter, and U. A. Th. Brinkman, (1996) J. Chromatogr. A 750: 275286. 4. C. Ferrer, M. J. Gmez, J. F. Garca-Reyes, I. Ferrer, E. M. Thurman, and A. R. FernndezAlba, (2005) J. Chromatogr. A 1069: 183194.
Acknowledgments
We acknowledge Amadeo Fernandez-Alba, JuanFrancisco Garca-Reyes and Carmen Ferrer from the Department of Hydrogeology and Analytical Chemistry, University of Almera, for help, ideas, and analytical support. Agilent contact: Jerry Zweigenbaum

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice.
Agilent Technologies, Inc. 2005 Printed in the USA August 23, 2005 5989-3573EN
Identifying Pesticides with Full Scan, SIM, ECD, and FPD from a Single Injection Application
Food Safety, Environmental
Authors
Chin-Kai Meng and Bruce Quimby Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA
Abstract
In this application note, a gas chromatography/mass spectrometry (GC/MS) system capable of providing up to four signals from a single injection is described. When a three-way micro-fluidic splitter is added to the end of the column, two additional signals from GC detectors can be acquired together with the MS data from a single injection. This multi-signal configuration provides: full-scan data for library searching, selective ion monitoring (SIM) data for trace analysis, micro-electron capture detector and flame photometric detector data for excellent selectivity and sensitivity from complex matrices. A combination of element selective detectors, SIM/Scan, and deconvolution reporting software makes a very powerful pesticide analysis system. Examples for trace-level compound quantitation/confirmation or for screening are discussed.
Chromatograph (LC) depending on the nature of the compounds. For GC amenable compounds, the traditional detectors are NPD (Nitrogen Phosphorus Detector), ECD (micro-Electron Capture Detector), and FPD (Flame Photometric Detector) for their excellent sensitivity and selectivity. However, even with dual-column confirmation analysis, these GC detectors cannot be used to verify the identity of the compounds with high confidence. Full scan mass spectral data and library searching are typically used for final compound verification. However, full-scan analysis has a worse (higher) detection limit (DL) compared to selective detectors on a GC. To improve the DL, the technique selective ion monitoring (SIM) is often used. With SIM, the MS monitors only a few characteristic ions for each target compound within the retention time (RT) range that the target elutes from the column. By monitoring only a few specific ions, the signal-to-noise ratio (S/N) improves significantly. The ions monitored are time programmed in groups corresponding to the RTs of the targets. SIM analyses with closely eluting targets require precise alignment of chromatographic RTs with the time programming of SIM groups. The Retention Time Locking (RTL) technique can be applied to eliminate the need to adjust SIM group time-windows after column maintenance or replacement. In this application note, a GC/MS system capable of providing up to four signals from a single injection is described. The benefits of the multi-signal detection include: Confirmatory information Full-scan data for library search capability
Introduction
Many laboratories in the world are analyzing pesticide residue levels in both foods and the environment to protect human health. The process usually involves homogenizing the sample, extracting the pesticides, and analyzing the target compounds with a Gas Chromatograph (GC) or a Liquid
Maximum sensitivity SIM data enables trace analysis Excellent selectivity ECD and FPD detect trace-level hetero-compounds from complex matrices
Experimental
A recent technical note describes Synchronous SIM/Scan, which takes advantages of the Performance Electronics in the 5975 inert MSD to get both SIM and full-scan signals in a single run without sacrificing performance [1]. The SIM method can be easily developed automatically using the ChemStations AutoSIM tool [2]. By simply selecting a checkbox in the method, the SIM and fullscan data can be acquired together. The trade-off is giving up some cycles per second but gaining an additional signal (full-scan data or SIM data) for the whole analysis. With properly chosen acquisition parameters, for example, increasing the scan speed, the decrease of cycles per second is usually not significant and does not affect peak quantitation or the quality of results (for example, S/N).
Figure 2.
A close-up view of the micro-fluidic three-way splitter in the 6890 GC oven.
EPC Injection MSD (SIM/Scan) Split vent ECD Splitter FPD (P)
At the end of the column, effluent flow is split three ways according to the length and diameter of the capillary tubing (restrictor) used.
Figure 1.
A schematic of the multi-signal configuration. Note: the EPC flow adds to the column flow into the splitter.
The size of the micro-fluidic plate is 1.25 inches (3.2 cm) wide and 2.5 inches (6.4 cm) tall. The device was designed to eliminate the common problems of large thermal mass, excess dead volume, and leaky connection due to oven temperature cycling etc. The splitter's flow paths and connection points are laid out and etched onto a thin, stainless steel plate using photolithography and chem-milling technologies. The plate is diffusion bonded, mounted with column connectors, and surface deactivated, resulting in an integrated and compact micro-fluidic splitter. Metal ferrules are used at the connectors that are leak-free after temperature cycling and will not absorb solvents or sample matrix, improving sensitivity for trace analysis applications. Deactivated capillary tubing between the splitter and each detector was used as a flow restrictor. Aux EPC pressure and the restrictor dimensions were determined using a spreadsheet-like calculator program to achieve the proper split ratio among all detectors. The three-way splitter can easily turn into a two-way splitter when a connector is capped. Other advantages of a splitter include backflushing [3] and quick-swapping. The Aux EPC flow can be run-time programmed to a higher pressure, while at the same time the inlet pressure is lowered to near ambient. This causes the column flow to reverse direction, back-flushing the less volatile materials out of the split vent of the inlet. The Aux EPC on the splitter also allows column changing and inlet maintenance without cooling and venting the MSD. The splitters flow paths and connection points were designed in such a way
Besides the SIM/Scan data, the ChemStation software can simultaneously acquire up to two additional GC detector signals, for example, FPD (in phosphorus- or sulfur- mode) and NPD (nitrogenphosphorus detector) signals or both P- and S- signals from a dual-wavelength FPD (DFPD). See Figure 1. Figure 1 is a schematic for multi-signal detection. At the end of the column, a three-way micro-fluidic splitter was used to split the column effluent to different detectors [3]. For this study, an FPD and a ECD were installed. Notice on the figure that an Auxiliary Electronic Pneumatics Control (Aux EPC) gas channel was connected to the splitter to maintain the pressure at the end of the column so that the split ratios/flows are kept constant throughout a run. Figure 2 shows a close-up view of the microfluidic splitter installed in the GC oven.
2
that when the column fitting is removed, the helium gas from the Aux EPC purges the fitting, preventing air from entering the splitter/MSD. See Table 1 for hardware details and settings.
Table 1. GC Inlet
Gas Chromatograph, Mass Spectrometer, and Three-Way Splitter Operating Parameters Agilent Technologies 6890 EPC Split/Splitless Splitless, 1.0 L injected (7683 ALS) 280 C ~27 psi (chlorpyrifos methyl RT locked to 16.596 min) 50.0 mL/min 0.75 min 55.3 mL/min Off Helium Siltek Cyclosplitter, 4-mm id, Restek p/n 20706-214.1
Mode Inlet temp Pressure Purge flow Purge time Total flow Gas saver Gas type Inlet liner Oven Oven ramp Initial Ramp 1 Ramp 2 Ramp 3 Total run time Equilibration time Oven max temp Column Length Diameter Film thickness Mode Nominal initial flow Outlet Outlet pressure
C/min 25 3 8
Final (C) 70 150 200 280
Hold (min) 2.00 0.00 0.00 15
46.87 min (last standard elutes around 35 min) 0.5 min 325 C Agilent Technologies HP 5-ms, p/n 19091S-433 30.0 m 0.25 mm 0.25 m Constant pressure 2.5 mL/min Unspecified 3.8 psi (Aux EPC pressure to splitter)
Front detector (FPD) Phosphorus mode Hydrogen flow: Oxidizer flow: Temperature: Oxidizer gas type: Mode: Makeup flow: Makeup gas type: Lit offset: Data rate: 75.0 mL/min 100.0 mL/min 250 C Air Constant makeup flow 60.0 mL/min Nitrogen 2.00 5 Hz Sulfur mode Hydrogen flow: Oxidizer flow: 50.0 mL/min 60.0 mL/min
Table 1.
Gas Chromatograph, Mass Spectrometer, and Three-Way Splitter Operating Parameters (Continued) 300 C Constant makeup flow 60.0 mL/min Nitrogen 5 Hz MSD Transfer line heater 280 C Helium 3.80 psi 0.00 min (this value will follow oven ramp) Agilent Technologies 5975 inert MSD Atune.U Scan 3.00 min Atune voltage 45 amu 555 amu 100 2 4 2.89 150 C 230 C Agilent 6890N Option 890, when installed on the GC during factory assembly 10:10:1 MSD:FPD:ECD 1.444 m 0.18-mm id Deactivated fused silica tubing 0.532 m 0.18-mm id Deactivated fused silica tubing 0.507 m 0.10-mm id Deactivated fused silica tubing 1.53 mL/min 1.53 mL/min 0.153 mL/min 1.38 mL/min
Back detector (ECD) Temperature: Mode: Makeup flow: Makeup gas type: Date rate: Thermal AUX 2 Use: Initial temp: Pressure AUX 5 Gas type: Initial pressure: Initial time: MSD Tune file Mode Solvent delay EM voltage Low mass High mass Threshold Sampling A/D Samples Scans/s Quad temp Source temp Three-way splitter Split ratio MSD restrictor FPD restrictor ECD restrictor Flow to MSD (at 280 C) Flow to FPD (at 280 C) Flow to ECD (at 280 C) Makeup flow (at 280 C)
Software Used in this Application Note GC/MSD ChemStation G1701DA Deconvolution Reporting Software (DRS) G1716AA NIST Library G1033A AMDIS (included for free with the NIST library CD)

Figure 3 shows four signals that were simultaneously acquired from a single injection of a pesticide mixture. Due to the high sensitivity of the ECD, the split ratios for the three detectors was set to MSD:FPD:ECD = 10:10:1. This split ratio distributes the sample of a 1-L splitless injection of a 1-ppm (1000 pg/L) sample to the different detectors as labeled in Figure 3.
Scan: ~ 480 pg
SIM: ~ 480 pg
ECD: ~ 48 pg
FPD (P): ~ 480 pg
5.00
10.00
15.00
20.00
25.00
30.00
35.00
40.00
45.00
Figure 3.
Signals acquired simultaneously from a 1-L splitless injection of 1-ppm standard. The split ratios were MSD:FPD:ECD = 10:10:1.
Figure 4 shows the signals when the pesticide standard was diluted 100-fold in a produce matrix. The total ion chromatogram (TIC) from full scan was not shown due to the lack of sensitivity. The FPD(P) and ECD were able to detect all the pesticides spiked in this extract. For trace-level target compound analysis, the SIM signal can be used for quantitation and the GC signals used for further confirmation.
1800 1700 1500 1300 1100 900 700 500 300 5.00 10.00 15.00 20.00 25.00 30.00 35.00
SIM
Produce extract 2688, spiked
700000 600000 500000 400000 300000 5.00 1240000 1200000 1160000 1120000 1080000 1040000 1000000 5.00
FPD(P)
10.00
15.00
20.00
25.00
30.00
35.00
ECD
10.00
15.00
20.00
25.00
30.00
35.00
Figure 4.
Data of a produce extract spiked at 10 ppb. FPD and ECD were able to detect the respective standards spiked into the extract.
Another application for this multi-signal system is for screening. In screening, no target list is available for the analysis; therefore, SIM acquisition or MS/MS is not possible. Figure 5 shows three signals (no SIM) from a produce extract.
800000 700000 600000 500000 400000 300000 200000 100000 5.00 50000 45000 40000 35000 30000 25000 20000 15000 10000 5.00 110000 100000 90000 80000 70000 60000 50000 40000 30000 5.00 10.00 15.00 20.00 25.00 30.00 35.00 10.00 15.00 20.00 25.00 30.00 35.00 10.00 15.00 20.00 25.00 30.00 35.00
Produce extract 13-10927
SCAN
40.00
45.00
FPD (P)
40.00 45.00
ECD
40.00
45.00
Figure 5.
Full-scan, FPD(P), and ECD data for extract 13-10927.
The Deconvolution Reporting Software (DRS) [3, 4] found several pesticides in the TIC as shown in Figure 6.
Figure 6.
Report for extract 13-10927 generated from DRS.
The possible pesticides in the sample were benzophenone, chlorpyrifos methyl, and thiabendazole. Propoxur and metamitron were not confirmed by both AMDIS and NIST; therefore, they were most likely false positives. Due to the complexity of the sample matrix and other interferences, it is sometimes difficult to get a high library match factor from peaks in the TIC, even after background subtraction. Therefore, element selective detectors would be very useful in providing the supporting information for compound confirmation. The multi-signal system was retention time locked, therefore, from the RT and the aligned peaks from the FPD(P) and the ECD responses, chlorpyrifos methyl (C7H7Cl3NO3PS) was confirmed.
It usually takes less than 3 minutes to turn off the FPD photomultiplier, swap the P-filter with the S-filter, and turn the photomultiplier back on. After the swap, adjust the detector gas flows to optimize the response in either P- or S- mode. A new injection of the same extract was made in FPD(S) mode. The FPD(S) result is shown with previously acquired signals in Figure 7. Two major peaks were seen on the FPD(S) chromatogram. From the peak RTs, they supported the presence of chlorpyrifos methyl and thiabendazole (C10H7N3S) respectively. Note that the full-scan TIC barely showed a peak for either compound, which made it impossible for traditional data analysis to identify both compounds. The FPD(S) mode is very selective, but it is not as sensitive as the FPD(P) mode. Although the ECD is very sensitive, it is not as selective as the FPD. A combination of GC detectors, SIM/Scan, and DRS makes a very powerful pesticide analysis system.
Conclusion
The Synchronous SIM/Scan provides users with library searchable full-scan spectra as well as trace level SIM data in a single analysis. When a threeway micro-fluidic splitter is added to the end of the column, two additional signals from element selective detectors can be acquired together with the MS data from a single injection. This configuration makes it very attractive for the analysis of trace-level pesticide residues in foods or environmental samples. This multi-signal configuration provides: full-scan data for library searching, SIM data for trace analysis, ECD and FPD data for excellent selectivity and sensitivity from complex matrices. In this application note, examples of ECD signal and FPD signal (P- or S- mode) were acquired together with the SIM/Scan data from a single injection for trace-level compound quantitation/confirmation, or for screening.
800000 700000 600000 500000 400000 300000 200000 100000 5.00 23000 21000 19000 17000 15000 13000 5.00 50000 45000 40000 35000 30000 25000 20000 15000 10000 5.00 10.00 15.00 20.00 25.00 30.00 10.00 15.00 20.00 25.00 30.00 10.00 15.00 20.00 25.00 30.00
Produce extract 13-10927
SCAN
35.00 40.00 45.00
Thiabendazole
FPD (S) (new injection)
Chlorpyrifos methyl
35.00
40.00
45.00
FPD (P)
35.00 40.00 45.00
Figure 7.
Full-scan, FPD(S), and FPD(P) data for extract 13-10927.
References
1. Chin-Kai Meng, Improving Productivity with Synchronous SIM/Scan, Agilent Technologies, publication 5989-3108, www.agilent.com/chem 2. Harry Prest and David W. Peterson, New Approaches to the Development of GC/MS Selected Ion Monitoring Acquisition and Quantitation Methods, Agilent Technologies, publication 5988-4188, www.agilent.com/chem 3. Michael J. Szelewski and Bruce Quimby, New Tools for Rapid Pesticide Analysis in High Matrix Samples Agilent Technologies, publication 5989-1716, www.agilent.com/chem 4. Philip L. Wylie, Michael J. Szelewski, Chin-Kai Meng, and Christopher P. Sandy, Comprehensive Pesticide Screening by GC/MSD Using Deconvolution Reporting Software Agilent Technologies, publication 5989-1157, www.agilent.com/chem

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2005 Printed in the USA July 6, 2005 5989-3299EN
Determination of Fungicides in Fruits and Vegetables by Time-of Flight and Ion Trap LC/MS Application
Food Safety
Author
Imma Ferrer and E. Michael Thurman Pesticide Residue Research Group University of Almera, 04120 Almera, Spain
Introduction
The importance of the identification and quantitation of fungicides in fruits and vegetables was described [1, 2]. Increasingly, LC/MS is being used for the analysis of many polar fungicides of the European Union (EU) Directives and is rapidly becoming an accepted technique in pesticide residue analysis for regulatory monitoring. In particular LC/MS works well on the various families of fungicides that are polar and labile. Thus, often LC/MS methods are preferred over the older gas chromatography/mass spectrometry (GC/MS) methods. For example, the use of MS and MS/MS for pesticides (including fungicides) in food was reviewed by Pico et al. [3], and there are currently about 100 published papers dealing with the analysis of pesticides in food. Of these, there are approximately 25 papers that deal with fungicides by LC/MS, the majority published in the early 2000s. Thus, LC/MS is an emerging technology for the analysis of food products. Interestingly though, none of the reported papers in this most recent review deal with the use of LC/TOFMS and accurate mass analysis for fungicides in food. Thus, there is an important need for research studies and methods development on the analysis of pesticides in food by accurate mass using LC/ TOFMS [1, 2]. Our study in this report is one of the first of its kind to examine the new Agilent LC/MSD TOF time-of-flight mass spectrometer for the analysis of fungicides in food. This topic was chosen because of the relevance of these fungicides and their significant use on apples, oranges, lemons, melons, tomatoes, broccoli, and peppers.
Abstract
This application examines the feasibility of the new instrumentation of electrospray and liquid chromatography/ time-of-flight mass spectrometry (LC/TOFMS) in conjunction with liquid chromatography/ ion trap mass spectrometry (LC/ITMS) to analyze five major fungicides in fruit extracts (apple, lemon, melon, and orange) and in salad vegetables (tomato, broccoli, and pepper). Included are the LC/TOFMS and LC/ITMS MS/MS spectra of five important fungicides (carbendazim, thiabendazole, azoxystrobin, dimethomorph, and triflumizole), the TOF empirical formula and MS/MS fragmentation and diagnostic ion(s) for these fungicides in the matrices of various important fruits and vegetables. A detailed rapid procedure for sample preparation and extraction of these fungicides from the fruit and vegetables is given. Spectral quality at the limit of detection (LOD), linearity, and quantitation of the fungicides in pure solvent and in the fruit and vegetable extracts using TOF and ion trap are provided. Last are the results of the analysis of real samples from the marketplace for these fungicides in fruits and vegetables: apples, oranges, lemons, and melons.
Furthermore, LC/ITMS will be demonstrated as a companion technique to LC/TOFMS for the analysis of fungicides in fruits and vegetables. The companion use of LC/ITMS was described in more detail [2]. The molecular structures of the common fungicides of this study (carbendazim, thiabendazole, azoxystrobin, dimethomorph, and triflumizole) are shown in Figure 1.
2. Take 5 mL of the supernatant into a 15-mL tube, add 250 mg of PSA adsorbent and 750 mg of MgSO4, and vortex for 20 s. Then microcentrifuge again for 3 min at 3700 rpm. Transfer 1.0 mL into an autosampler vial. Then evaporate the fruit and vegetable residues to dryness and redissolve in 8/92% methanol/ water. Standards were prepared by fortifying with a mixture of carbendazim, thiabendazole, azoxystrobin, dimethomorph, and triflumizole at concentrations ranging from 0.01 to 0.5 mg/kg for analysis using the Agilent LC/MSD TOF and the Agilent 1100 Series LC/MSD Trap. Nonfortified samples were analyzed directly at this same point by either the LC/MSD TOF or the LC/MSD Trap by injecting 50 L. The extraction method is summarized in Figure 2.
O H N NH N Carbendazim [M+H]+ = m/z 192 OCH3
H N
S N
Thiabendazole [M+H]+ = m/z 202
N O CN Azoxystrobin [M+H]+
N O H3CO = m/z 404 O OCH3 Cl F
Sample 1 kg of fruit or vegetable. F
H3CO
OCH3 O N O H3C O N
N Cl Dimethomorph [M+H]+ = m/z 388 Triflumizole [M+H]+ = m/z 346
Combine 15 g of sample, 15 mL of acetonitrile (1% acetic acid), extract and shake for 60 s with some Vortex to mix. Add 6 g of MgSO4 and 2.5 g of NaAc3H2O and shake. Centrifuge for 3 min at 3700 rpm.
Put 5 mL of sample into a 15-mL tube, add 250 mg of PSA adsorbent and 750 mg of MgSO4, and vortex and shake for 20 s. Then microcentrifuge for 180 s at 3700 rpm.
Figure 1.
Structures of major fungicides used on fruits and vegetables, and the corresponding [M+H]+ ion. Figure 2.
Transfer 1.0 mL into LC/MS vials
Fruit and Vegetable Extraction (QuEChERS) QuEChERS is the acronym for the method of extraction, which stands for quick, easy, cheap, effective, rugged, and safe [4]. It is a method that is receiving wide acceptance for rapid extraction of pesticides in food. 1. Weigh 15 g of a previously homogenized sample into a 40-mL Teflon centrifuge tube. Add 15 mL of acetonitrile (containing 1% acetic acid), 6 g of anhydrous MgSO4 and 2.5 g of NaAc3H2O (sodium acetate trihydrate) and shake the sample vigorously for 1 min by using a Vortex mixer at maximum speed or by hand shaking. Afterwards, centrifuge for 3 min at 3700 rpm.
2
QuEChERS Extraction method for fungicides in fruits and vegetables.
Agilent LC/MSD TOF Methods LC Pumps were Agilent 1100 binary pumps, injection volume 50 L with standard Agilent 1100 ALS. Column: ZORBAX Eclipse XDB-C8, 4.6 mm 150 mm, 5 m, part number 993967-906 Mobile Phase A=acetonitrile and B=0.1% formic acid in water, gradient was 10% A for 5 min, then to 100% A in 25 min at a flow rate of 0.6 mL/min Agilent LC/MSD TOF with electrospray source Positive ESI, Capillary 4000 V
Nebulizer 40 psig, drying gas 9 L/min, gas temp 300 C Fragmentor 190 V, Skimmer 60 V, Oct DC1 37.5 V, OCT RF V 250 V Reference Masses: m/z 121.0509 and 922.0098, resolution: 9500500 @ m/z 922.0098, Reference A Sprayer 2 is constant flow rate during the run Agilent 1100 Series LC/MSD Trap Methods Chromatographic methods were identical to Agilent LC/MSD TOF for direct comparison of peaks LC Pumps were Agilent 1100, injection volume 50 L Column: ZORBAX Eclipse XDB-C8, 4.6 mm 150 mm, 5 m, part number 993967-906 Mobile Phase A=acetonitrile and B=0.1% formic acid in water, gradient was 10% A for 5 min, then to 100% A in 25 min at a flow rate of 0.6 mL/min. Agilent 1100 Series LC/MSD Trap Positive electrospray ionization (ESI), Capillary 3200 V Nebulizer 40 psig, drying gas 9 L/min, gas temp 300 C Capillary exit 70 V Trap accumulation, 50,000 counts
Intensity, cps
17.6 5.0e7 4.0e7 3.0e7 2.0e7 1.0e7 0.0 2 1.6e6 1.4e6 1.2e6 1.0e6 8.0e5 6.0e5 4.0e5 2.0e5 0.0 2 4 4 6 8 2.1 3.6 4.1 13.4 17.4 15.1 18.0 20.8 22.7 24.0 25.1
(a) 28.9
10 12 14 16 18 20 22 24 26 28 30 Time, min 4 (b)
Intensity, cps
3 5
10 12 14 16 18 20 22 24 26 28 30 Time, min
Figure 3.
(a) TIC of a lemon standard solution (0.125 ppm); (b) EICs of carbendazim (1), thiabendazole (2), dimethomorph (3), azoxystrobin (4), and triflumizole (5)
of the compounds are shown in Figure 3b. The window of extraction was clean, which is commonly a feature of accurate mass extraction of ions, where the width of the window of extraction may be narrowed to (0.02 amu) or 100 ppm. In Figure 3, the window of extraction was 0.1 amu. At concentrations as low as 0.05 mg/kg, the extracted ions (m/z 192, 202, 346, 388, and 404) still yielded clean chromatograms testifying to the importance of accurate mass and its ability to give clean extracted ion chromatograms (EICs) with a narrow mass window. This accurate mass window is reflected in the determinations of the accurate mass of the protonated molecule of each of the fungicides. Table 1 shows the elemental composition, the exact mass, and errors, in mDa and ppm, for each of the fungicides at the 0.50 mg/kg concentration. Mass accuracy was always better than 2 ppm in all the fruit and vegetable matrices, except for dimethomorph.

Fragmentation and Mass Accuracy Figure 3 shows the LC separation of five fungicides: carbendazim, thiabendazole, azoxystrobin, dimethomorph, and triflumizole. The sample contained the fungicides at 0.125 mg/kg (ppm) in a fortified lemon extract. The extracted ions for each
Table 1.
Accurate Mass and Elemental Composition of Five Fungicides at 0.50 mg/kg in Lemon Matrix Error mDa -0.05 -0.34 +1.1 0.20 -0.35 Error ppm 0.27 1.7 2.5 0.50 1.0 3
Compound Carbendazim Thiabendazole Dimethomorph Azoxystrobin Triflumizole
Formula C9H9N3O2 C10H7N3S C21H22NO4Cl C22H17N3O5 C15H15N3OF3Cl
Selected ion [M + H]+ [M + H]

+
m/zexperimental 192.0767 202.0430 388.1321 404.1243 346.0925
m/zcalculated 192.07675 202.04334 388.13101 404.1242 346.09285
[M + H]+ [M + H]+ [M + H]+
These results show that the use of continuous calibration is effective for accurate mass across an order of magnitude concentration range in complex fruit and vegetable matrices. Figures 4a4e show the fragmentation pathway of each of the five fungicides, based on accurate mass spectra with pure standards in a methanol solution. The fragmentation ions of the LC/MSD TOF were verified with the LC/MSD Trap spectra at MS2 with identical chromatographic conditions. Beginning with carbendazim, the MS2 shows that the protonated molecule loses methanol (32 Da) to give the m/z 160 fragment ion (Figure 4a). This fragment ion of m/z 160 is also seen in the LC/MSD TOF. The fragmentation pathway for thiabendazole is shown in Figure 4b with the protonated molecule, m/z 202, fragmenting to give the m/z 175, with the neutral loss of 27 Da or HCN. The fragmentation pathway for azoxystrobin is shown in Figure 4c with the protonated molecule, m/z 404, fragmenting to m/z 372 (loss of methanol of 32 Da) at MS2, and to m/z 342 at MS3. The fragmentation pathway for dimethomorph is shown in Figure 4d with the protonated molecule, m/z 388, fragmenting to give the m/z 301 at MS2, then m/z 165 at MS3. The fragmentation pathway for triflumizole, m/z 346, is to m/z 278 at MS2, then to m/z 250 at MS3 (Figure 4e). The combination of the LC/MSD Trap and LC/MSD TOF is extremely valuable for interpretation of spectra in that both instruments work well for these compounds and give complementary information on the structure and identity.
MSD Ion Trap of carbendazim

+H+ NH O
Fragment ion at MS2

+H+ O N
N H
OCH3
Carbendazim, [M+H]+ = m/z 192
Fragment ion, m/z 160
192.0766
4.7e4 4.0e4
N N O
H2 N N
O
Int ensity, counts
N H
NH
OCH3
3.0e4
2.0e4
160.0502
1.0e4
193.0796
0.0 145 155 165 175 m/z, amu 185 195 205 215
Figure 4a. Structure and fragmentation pathways for carbendazim using the LC/MSD TOF and LC/MSD Trap. The spectrum is from the LC/MSD TOF.
MSD Ion Trap of thiabendazole

H N +H+ S
Fragment ion at MS2

+H+ N S
N Thiabendazole, [M+H]+
N N = m/z 202 Fragment ion, m/z 175
1.00e5 9.00e4 8.00e4 Intensity, counts 7.00e4 6.00e4
202.0433
S 5.00e4 4.00e4 3.00e4 202.2072 2.00e4 203.0459 1.00e4 200.0 202.0 204.0404 204.0 206.0 m/z, amu 208.0 210.0 N H
Figure 4b. Structure and fragmentation pathways for thiabendazole using the LC/MSD TOF and LC/MSD Trap. The spectrum is from the LC/MSD TOF. 4
N O
N Cl O H3CO +H+ OCH3 O
Ion Trap MS/MS of dimethomorph
Fragment ion at MS2

H3CO OCH3
Ion Trap MS/MS of azoxystrobin, [M+H]+ = m/z 404
CN
Azoxystrobin, m/z 404
+H+ N H3CO H3CO O Dimethomorph, [M+H+]+ = m/z 388 O
N O +H+
Fragment ion, m/z 372, at MS2

CN
Cl Fragment ion, m/z 301
Fragment ion, m/z 372 H3CO O
H3CO
Fragment ion at MS3 Fragment ion, m/z 342, at MS3

CN N O N O +H+ O 8.0e4 7.0e4 Intensity, counts Intensity, counts 4.0e5 1 ppm accuracy 3.0e5 O 2.0e5 CN H3CO 1.0e5 372.0982 OH CN H3CO OH 405.1275 426.1061 427.1092 420 430 N N O 1 ppm accuracy N O N O OCH3 2.0e4 1.0e4 404.1246 6.0e4 5.0e4 4.0e4 3.0e4 389.1341 Cl 390.1280 O O 388.1310 O NH
O +H+ Fragment ion, m/z 165 MS3
CH3 O CH3
391.1317 386.0 387.0 388.0 389.0 390.0 391.0 392.0 393.0 394.0 m/z, amu
373.1008 370 380 390
406.1301 400 410 m/z, amu
Figure 4d. Structure and fragmentation pathways for dimethomorph using the LC/MSD TOF and LC/MSD Trap. The spectrum is from the LC/MSD TOF.
Figure 4c. Structure and fragmentation pathways for azoxystrobin using the LC/MSD TOF and LC/MSD Trap. The spectrum is from the LC/MSD TOF.
Ion Trap MS/MS of triflumizole
Fragment ion at MS2
+H+ O+ O
N N N N
CF3
CF3
Cl
Triflumizole [M+H]+ = m/z 346
Cl
Cl
Fragment ion at MS3

N O + CF3 Fragment ion, m/z 250 at MS3
broccoli, and pepper). Results showed the similarity among matrices with the LC/MSD TOF and LC/MSD Trap, given the variability in fruit and vegetable matrices. For example, Figures 5a5b show the standard curves for two representative fungicides, carbendazim and azoxystrobin. These figures demonstrate that there was little or no matrix suppression in the LC/MSD TOF system for the fruit extracts up to 0.5 ppm of parent compound for carbendazim, thiabendazole, and azoxystrobin. However, triflumizole and dimethomorph showed both enhanced signal and suppressed signal (data not shown here). The enhancement occurred for the fruit extracts of melon, orange, and lemon. The vegetable extracts of pepper showed no enhancement and some suppression for broccoli was observed. One possible explanation for these results for triflumizole has to do with the late retention time of the analyte (26 min), which means that the mobile phase is approximately 80% acetonitrile. The ESI signal is susceptible to matrix effects at these high concentrations of organic solvent [5], which indicates the importance of using matrix matched standards for unknown analysis of food extracts. Thus, best accuracy for reporting concentrations was with the standard curve made up in matrix for each fruit or vegetable. This was the procedure that was used for unknown analysis in real fruit and vegetable samples.
(a)
F
Cl 278.0558 2.5e5
Carbendazim
30 F Area 10 E+06
O+ H3C N CF3 O N
2.0e5 Intensity, counts
20
1.5e5
10
Solvent Pepper Lemon Orange Broccoli Melon
1.0e5
Cl 280.0529 279.0586 281.0556 280 290 300 310 320 m/z, amu 330
N 346.0930
0.1
0.2
5.0e4 1.0e4
0.3 0.4 0.5 Concentration (mg/kg)
0.6
347.0958 348.0912 340 350 Area 10 E+06
50
Azoxystrobin
(b)
40
Figure 4e. Structure and fragmentation pathways for triflumizole using the LC/MSD TOF and LC/MSD Trap. The spectrum is from the LC/MSD TOF.
30 Solvent Pepper Broccoli Lemon Orange Melon
20
Linearity and Detection Limits

Calibration curves were established for the five fungicides using both the LC/MSD TOF and LC/MSD Trap over the analyte concentration range of interest, which was from 0.01 mg/kg to 0.5 mg/kg (ppm) in solvent and in each of the fruit and vegetable matrices (orange, lemon, melon,
6
10
0 0 0.1 0.2 0.3 0.4 0.5 Concentration (mg/kg) 0.6
Figure 5. Overlay of standard curves for the LC/MSD TOF analysis of carbendazim (5a) and azoxystrobin (5b) in five matrices.
Furthermore, Figures 5ab also show that the standard curves were linear across the range of concentration tested with correlation coefficients of 0.990 to 0.999 for all matrices tested and for all four of the fungicides (Table 2). Similar results for standard curves were seen with the LC/MSD Trap, with correlation coefficients typically of 0.9900.001.
Table 2. Correlation Coefficients for Standard Curves for Various Pesticides in Representative Fruit Matrices Orange 0.998 0.999 0.990 0.996 Melon 0.997 0.999 0.990 0.999 Lemon 0.999 0.999 0.997 0.996
For example, carbendazim is regulated at 0.10 mg/kg in tomatoes, thiabendazole is 0.50 mg/kg for tomatoes, while azoxystrobin is not regulated in tomatoes in Spain; therefore, the maximum residue limit is automatically 0.01 mg/kg. The LC/MSD TOF is capable of these LODs for each of the fungicides in the various fruit and vegetable matrices. The quantitation for real fruit extracts from actual grocery store samples are shown in Table 5 for LC/MSD TOF analysis of fungicides with LC/MSD Trap confirmation. Thus, these data show that the LC/MSD TOF is capable of accurate mass measurements of the fungicides in fruit and vegetable extracts with LODs needed for their monitoring in the EU.
Compound Carbendazim Thiabendazole Azoxystrobin Trifluomizole
Table 3 shows the limits of detection (LODs) of the fungicides for various matrices. Typically, the values of Table 3 vary from 0.001 ppm to 0.010 ppm. The European standard for pesticides with no regulatory standard is 0.010 ppm. Thus, the LC/MSD TOF is sufficiently sensitive to detect these compounds in all matrices. A similar result was achieved for the LC/MSD Trap for LODs (Table 4) using single MS. The accurate mass capabilities of the LC/MSD TOF appear capable of seeing through the complexity of the fruit and vegetable matrices without interferences to the LODs. These LOD results of the various fungicides in fruits and vegetables are capable of meeting the regulation limits of Spain and at those of the EU.
Table 3.
Table 4.
LOD in mg/kg for Fungicides by LC/MSD Trap in Three Matrices in Full-Scan Mode Orange 0.005 0.005 0.005 0.002 0.002 Melon 0.005 0.010 0.003 0.005 0.002 Lemon 0.010 0.005 0.010 0.010 0.005
Compound Carbendazim Thiabendazole Dimethomorph Azoxystrobin Triflumizole
Table 5.
Concentration (in mg/kg) of Fungicides by LC/MSD TOF and Confirmed by LC/MSD Trap in Samples of Fruits from Grocery Store Produce Orange 0.03 <LOD 0 0 0 Melon <LOD 0 0 0 0 Apple 0.1 0 0 0 0 Lemon 0.5 0.1 0 0 0
Compound LOD in mg/kg for Fungicides by LC/MSD TOF in Three Carbendazim Matrices Thiabendazole Compound Orange Melon Lemon Dimethomorph Carbendazim 0.004 0.005 0.008 Azoxystrobin Thiabendazole 0.005 0.010 0.005 Triflumizole Dimethomorph 0.005 0.002 0.008 Azoxystrobin Triflumizole 0.001 0.001 0.001 0.001 0.001 0.001
Conclusions
LC/MSD TOF analysis is a powerful tool for identification of fungicides in fruits and vegetables and is a new tool for environmental food chemistry. Quantitation is easily possible over 2 orders of magnitude with accuracy better than 3 ppm, typically less than 2 ppm, and in this work was an amazing 1 ppm in ESI+ ion for most compounds! Elemental composition of fungicides and fragment ions are possible with LC/MSD TOF, and further confirmed with LC/MSD Trap. LODs of the five fungicides in fruit and vegetables were from 0.001 to 0.010 g/g. These concentrations are equal to or better than the EU directives for controlled fungicides in fruits and vegetables.
References
1. Imma Ferrer and E. Michael Thurman, (2004) "Determination of chloronicotinyl insecticides in salad vegetables by LC/MSD TOF and LC/MSD ion trap", Agilent Technologies, publication 5989-1842EN www.agilent.com/chem 2. E. Michael Thurman and Imma Ferrer, (2005) "Identification of unknown pesticides in food using both LC/MSD TOF and Ion Trap MSn", Agilent Technologies, publication 5989-1924EN www.agilent.com/chem 3. Y. Pico, C. Blasco, and G. Font, (2004) Mass Spectrometry Reviews, 23, 45-85. 4. M. Anastassiades, S.J. Lehotay, D. Stajnbaher, and F.J. Schenck, (2003) Journal of AOAC International, 86: 412-431. 5. E.M. Thurman, I. Ferrer, and A.R. FernndezAlba, Chapter 8: LC/MS I. Basic Principles and Technical Aspects of LC/MS for Pesticide Analysis, in Chromatographic-Mass Spectrometric Food Analysis for Trace Determination of Pesticide Residues, Ed. A.R. Fernndez-Alba, Elsevier, Amsterdam, 2005.
Acknowledgements
We acknowledge Amadeo Fernndez-Alba, Department of Hydrogeology and Analytical Chemistry, University of Almera for analytical support.

For more information on our products and services, visit our Web site at www.agilent.com/chem. Agilent Contact: Jerry Zweigenbaum
ZORBAX is a registered trademark of E.I. duPont de Nemours Co. Inc. Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2005 Printed in the USA March 15, 2005 5989-2209EN
Identification of Unknown Pesticides in Food Using Both LC/MSD TOF and Ion Trap MSn Application
Food Safety
Author
E. Michael Thurman and Imma Ferrer Pesticide Residue Research Group University of Almera 04120 Almera Spain
searching of ions. The procedure here consists of multiple steps. First is the initial detection of a possible unknown pesticide in actual market-place vegetable extracts (tomato or another salad vegetable) using either the LC/MSD TOF or the LC/MSD ion trap. Second is the application of the "Dual Approach." This consists of accurate mass identification of an empirical formula or molecular mass (usually containing an A+2 isotope signature, such as from S or Cl), followed by MS/MS ion trap confirmation of the chlorinated and/or sulfur-containing pesticides and their major fragment ions and chlorine or sulfur signatures. Third is the database analysis of the empirical formula, which is carried out with either the Merck Index CD, containing 10,000 compounds, or the ChemINDEX CD, containing 77,000 compounds, and a laptop computer. In the examples given, the unknowns are identified as the organophosphate insecticide, malathion, in cucumbers; and the insect growth regulator, buprofezin in tomato extracts. The fourth and final step involves confirmation with authentic standards.
Agilent Contact: Jerry Zweigenbaum 2850 Centerville Road Wilmington, DE 19808-1610
Abstract
Traditionally, the screening of unknown pesticides in food was accomplished by gas chromatography/mass spectrometry (GC/MS) methods using conventional library searching routines. However, many of the new polar and thermally labile pesticides are more readily and easily analyzed using liquid chromatography/mass spectrometry (LC/MS) methods, even though no searchable libraries currently exist (with the exception of some limited user libraries). There is, therefore, a need for LC/MS methods to detect true pesticide unknowns. This application develops an identification scheme for unknown pesticides using a combination of liquid chromatography/mass selective detector (LC/MSD) with time-of-flight (TOF) and ion trap (MSn) options, and searching for the empirical formula using the accurate mass and the ChemINDEX or Merck Index databases. The approach is different from the conventional library
Introduction
The identification and quantitation of insecticides in vegetables is of great importance to individuals and health organizations around the world. The European Union (EU), has set new Directives for pesticides at low levels in vegetables in order to
meet these health concerns. For example, new laws such as the European Directive 91/414/EEC [1], or the Food Quality Protection Act (FQPA) [2] in the US have increased the standards for human health, workers, and environmental protection. They also require re-registration for older pesticides. Furthermore, the review programs have withdrawn authorizations for many of the crop protection products currently on the market, 177 compounds in the US and 320 in Europe. Moreover, it was announced in Europe that a total of 110 products will be withdrawn in the near future [1]. The revision of the review program in Europe dictates that over 450 existing active substances will be taken off the market by 2008 with about 400 pesticides remaining in use. Next, the quality standards within the new regulations include the re-assessment of the maximum residue limits (MRLs) for vegetables, and EU directives are setting different MRLs for each pesticide within each food group. Typically, the MRLs are lower than the previous ones. Furthermore, the new Directive leads to different MRLs for each EU country, which are still being decided. The EU Directive states that individual country MRLs will be maintained in the new program. Finally, banned compounds have the lowest MRLs, which is set now at 0.01 mg/kg (ppm). With the planned program to remove so many compounds from the market, it is important, even necessary, that screening for unknown pesticides may be done by LC/MS on vegetable extracts. Because it is not always possible to know which banned substances may be used, it is of vital importance to environmental food monitoring that there be a system to give fast and accurate screening of unknown substances in food and food products. Thus, there is an important need for research studies and methods development on the analysis of unknown pesticides in food by new LC/MS methods, such as the combination of accurate mass using LC/MSD TOF, and MS/MS using LC/MSD ion trap. Our study in this report is one of the first of its kind to examine the new Agilent LC/MSD TOF combined with LC/MSD ion trap, and the use of commercial databases, such as the Merck Index and the ChemINDEX database to identify unknown pesticides in food. Several advantages of the combination of LC/MSD TOF and ion trap are that accurate mass and empirical formulas may be combined with the MS/MS spectra or even MSn for spectral information. The identification of unknowns using the
LC/MSD TOF and ion trap consists of six steps, which are outlined below. 1. Analyze the vegetable extract with ion trap in full-scan mode looking for important unknown peaks or alternatively, using in-source CID on LC/MSD TOF. 2. Next check for A+2 isotope patterns to see if Cl, Br, or S is present. 3. Search Merck Index or ChemINDEX for possible unknowns using molecular mass (nominal mass weight), A+2 isotopes, and molecular formula. 4. Proceed to ion trap MS/MS with proposed ideas for structure and do MS2 or MS3. Identify ion fragments, accurate mass fragments, and possible structures. 5. Combine with LC/MSD TOF data of accurate fragment ions and protonated molecule to make tentative identification. 6. Obtain and analyze appropriate standard for final confirmation. Given are two detailed examples of this process using store-purchased salad vegetables, cucumbers and tomatoes, both of which contained "unknown white powders" that were subsequently identified by the above process for various "unknown pesticides."
Standard Vegetable Extraction [3] 1. Tomato, lettuce, pepper, and cucumber, in 2-kilogram portions, were obtained from the market place and homogenized by chopping. 2. Then 15 g of the homogenized vegetable was accurately weighed and placed into a 200-mL PTFE centrifuge tube. 3. Ethyl acetate (45 mL) and 6.5 M NaOH (1 mL) were added and blended in a high-speed blender (Polytron) for 30 s at 21,000 rpm [4]. After this time, 13 g of anhydrous Na2SO4 was added and the extraction repeated for 30 s. 4. The combined extracts were then filtered through a thin layer of 20 g of anhydrous Na2SO4. The solid was washed with 50 mL of ethyl acetate and the combined extracts were evaporated to dryness on a vacuum rotary evaporator using a water bath at 45 5 C.
5. The dried residue was dissolved by sonication in 15 mL of methanol. The extracts, which contained 1 g of sample per mL, were filtered through 0.2-m PTFE filters (Millex FG, Millipore) prior to LC/MS analysis. Rapid Vegetable Extraction 1. Select tomato or cucumber containing white powder from a commercial market place. 2. Carefully wash the vegetable skin three times with methanol to remove the white powder, recovering a total of 2 to 5 mL of wash, (depending on the size of the vegetable) into a 150-mL Pyrex beaker. 3. After mixing, transfer the methanol wash to a 5-mL syringe and filter through a Millex-FH PTFE filter and aliquot 0.3 mL. 4. Dilute with 0.6 mL of deionized water. 5. Analyze by either LC/MSD TOF or LC/MSD ion trap. LC/MSD TOF Methods Instrument: Agilent model LC/MSD TOF with electrospray source LC pumps were Agilent 1100 binary pumps, injection volume 50 L with standard Agilent ALS Column: ZORBAX Eclipse XDB-C8, 4.6 mm 150 mm 5 m, part number 993967-906 Mobile phases: A: ACN, B: 0.1% formic acid in water Gradient: 15% A to 100% A over 30 minutes, at 0.6 mL/min Positive ESI, capillary: 4000 V Nebulizer 40 psig, drying gas 9 L/min, gas temp 300 C Fragmentor: 190 V, skimmer: 60 V Oct DC1: 37.5 V, OCT RF V: 250 V Reference masses: m/z 121.0509 and 922.0098, resolution: 9500 500 @ m/z 922.0098 Reference A Sprayer 2 set to constant flow rate during the run
LC/MSD Ion Trap Methods Chromatographic conditions identical to those using LC/MSD TOF for direct comparison of peaks. Instrument: LC/MSD Ion Trap Classic LC pumps were Agilent 1100 binary pumps, injection 50 L with standard Agilent ALS Column: ZORBAX Eclipse XDB-C8, 4.6 mm 150 mm 5 m, part number 993967-906 Mobile phases: A: ACN, B: 0.1% formic acid in water Gradient: 15% A to 100% A over 30 minutes, flow rate: 0.6 mL/min Positive ESI, Capillary 3000 V Nebulizer 40 psig, drying gas 9 L/min, gas temp 350 C Skimmer 1: 20 V, Skimmer 2: 10 V, Capillary exit offset: 50 V, Capillary exit: 70 V Octopole 3 V, Oct RF 100 V, Octopole : 2 V, Lens 1: -3 V, Trap drive: 50, Lens 2: -50 V Target: 50,000, max. accumulation time: 200 ms, scan m/z 50 to 800, averages: 5 Rolling average: on, MS/MS m/z 2.0 isolation width, amplitude: 1.2 V, fragmentation cutoff m/z 83

Cucumber Extract Figure 1 shows the ion trap total ion chromatogram (TIC) for the rapid extract of the white powder present on a store-purchased cucumber. The largest peak in the chromatogram, deemed the Star*TIC, has a spectrum shown in Figure 1 (bottom panel) with ions at m/z 331 and 353. The spacing of m/z 22 indicates a sodium adduct. Because the analysis was carried out in positive ion, the [M+H]+ is the m/z 331, and the [M+Na]+ is the m/z 353. Furthermore, note that the m/z 331 ion also contains an A+2 ion at m/z 333, which is 8% of the m/z 331 ion. The two important A+2 ions to consider in unknown ID are chlorine and sulfur (occasionally bromine). Sulfur-34 is present in natural abundance of 4.8%; thus, the 8% peak is most likely due to two sulfur-34 atoms present in the unknown at m/z 331.
Intens. 106 2.5 2.0 1.5 1.0 0.5 0.0 0 5 TIC EIC 301 Intens. 106 0.8 0.6 0.4 0.2 0.0 98.9 170.0 100 200 285.1 300 400 500 Possible fragment ions 353.1 575.5 600 682.8 709.5 700 m/z 10 15 20 EIC 331 m/z 331 25 m/z 301 Time [min]
Intens. 106 2.0 1.5 1.0 0.5 0.0 20.5 Intens. 3000 2000 1000 0 Fragment ions 126.9 172.9 240.1 98.7 100 200 300 284.9 21.0 21.5 22.0 22.5 23.0 23.5 24.0 24.5 Time [min] TIC
MS/MS of 331 No fragmentation of 353 verifying it is a Na adduct The m/z 285, 240, and 99 ions are present 400 500 600 700 m/z
Figure 2.
Ion-Trap MS/MS and corresponding product mass spectrum for m/z 331.
TIC of cucumber skin Star*TIC
The next step is the search of the Merck Index for the molecular weight and sulfur atom content. Figure 3 shows an example search sheet on a laptop computer. Because the ion trap gives the protonated molecule (m/z 331) the search is done by subtracting a proton to give the molecular weight of 330. The molecular weight is then searched across one mass unit to 331. The molecular formula has only the S2, which is a requirement now of the database search.
331.1
1. [M+H]+ = m/z 331 2. m/z 353 = [M+Na]+ 3. S-34 ~8% = 2 S atoms
Figure 1.
Upper: cucumber skin ion-trap TIC; lower: extracted ion chromatogram (EIC) for m/z 331. The lower panel is the spectrum of the Star*TIC peak.
Figure 2 shows the LC/MS/MS of m/z 331. The ions present include: m/z 285, 240, 127, and 99. These ions will be used later for structural identification after a database search on the Merck Index finds a possible structure. These ions will then be compared with that structure.
Figure 3.
Database search for MW 330-331.
Figures 4a and 4b show the two results of the Merck Index database search. The compounds found were penicillin O and malathion, with molecular weights of 330.43 and 330.36, respectively.
Results of library search for MWt = 330-331 Molecular Formula = S2*
Monograph Number: 7170 Title: Penicillin O CAS Registry Number: 87-09-2

H2C S N H N O COOH S CH3 CH3 O H H
CAS Name: (2S,5R,6R)-3,3-Dimethyl-7-oxo-6-[[(2-propenylthio)acetyl]amino]-4-thia-1-azabicyclo[3.2.0]heptane-2carboxylic acid Additional Names: [(allylthio)methyl] penicillin; allylmercaptomethylpenicillin; allylmercaptomethylpenicillinic acid; penicillin AT Molecular Formula: C13H18N2O4S2
Molecular Weight: 330.43. Percent Composition: C 47.25%, H 5.49%, N 8.48%, O 19.37%, S 19.41% Literature References: Antibiotic produced by Penicillium chrysogenum. Biosynthesis of salts: Behrens et al., J. Biol. Chem. 175, 793 (1948); Rhodehamel, Behrens et al., US 2528175 and US 2623876 (1950, 1952, both to Lilly); Ford et al., Antibiot. & Chemother. 3, 1149 (1953); Ford, US 2647894 (1953 to Upjohn); Paleckov, Slechta, C.A. 50, 17309g (1956).
Derivative Type: 2-Chloroprocaine salt monohydrate Additional Names: Chloroprocaine penicillin O; penicillin O 2-chloroprocaine Trademarks: Depo-Cer-O-Cillin Chloroprocaine (Upjohn) Molecular Formula: C H ClN O S .H O 26 37 4 6 2 2 Molecular Weight: 619.20. Percent Composition: C 50.43%, H 6.35%, Cl 5.73%, N 9.05%, O 18.09%, S 10.36% Properties: Slender needles from hot water, mp 79-81. Practically insol in cold water. Stable in dry form at room temp. Aq suspensions are stable at room temp for 1 week, at refrigerator temps for 3 weeks. Calculated activity: 949 units/mg. Solubilities: Weiss et al., Antibiot. & Chemother. 7, 374 (1957). Melting point: mp 79-81
Figure 4a. First result for database search for MW 330-331.
Monograph Number: 5723 Title: Malathion CAS Registry Number: 121-75-5 CAS Name: [(Dimethoxyphosphinothioyl)thio]butanedioic acid diethyl ester
H3CO H3CO S P S
O O
CH3 CH3
Additional Names: diethyl mercaptosuccinate S-ester with O,O-dimethyl phosphorothioate; S-(1,2dicarbethoxyethyl) O,O-dimethyldithiophosphate; insecticide no. 4049; carbofos; malathon (obsolete); mercaptothion; phosphothion Manufacturers' Codes: ENT-17034 Trademarks: Cythion; Derbac-M (International); Malamar 50; Malaspray; Organoderm (Mundipharma); Prioderm (Coates & Cooper); Suleo-M (International) Molecular Formula: C10H19O6PS2
Molecular Weight: 330.36. Percent Composition: C 36.36%, H 5.80%, O 29.06%, P 9.38%, S 19.41% Literature References: Cholinesterase inhibitor. Prepn: Johnson et al., J. Econ. Entomol. 45, 279 (1952); Cassaday, US 2578652 (1951 to Am. Cyanamid). Purification: Usui, US 2962521 (1960 to Sumitomo). Treatment of Pediculus humanus (head lice) infestation: D. Taplin et al., J. Am. Med. Assoc. 247, 3103 (1982). Toxicity data: T. B. Gaines, Toxicol. Appl. Pharmacol. 14, 515 (1969). Review of distribution, transport and fate in the environment: M. S. Mulla et al., Residue Rev. 81, 1-159 (1981); of carcinogenic risk: IARC Monographs 30, 103-129 (1983). Properties: Deep brown to yellow liq, mp 2.9. bp 156-157. Characteristic odor. d 25 1.23. n 25 1.4985. Vapor pressure at 30: 4 10-5 mm Hg. Slightly sol in 4 D and alkylated aromatic hydrocarbons and vegetable oils. water (145 ppm). Misc with many organic solvents 0.7 including alcohols, esters, ketones, ethers, aromatic Limited soly in certain paraffin hydrocarbons. Petroleum ether is sol to about 35% in malathion. Hydrolyzed at pH >7.0 or <5.0. Stable in an aq soln buffered to pH 5.26. LD in female, male rats (mg/kg): 1000, 1375 orally (Gaines). 50 Melting point: mp 2.9 Boiling point: bp 156-157 0.7 Index of refraction: n 25 1.4985 D Density: d 25 1.23 4 Toxicity data: LD in female, male rats (mg/kg): 1000, 1375 orally (Gaines) 50 CAUTION: Potential symptoms of overexposure are miosis, aching eyes, blurred vision and lacrimation; eye and skin irritation; salivation; anorexia, nausea, vomiting, abdominal cramps, diarrhea, giddiness, confusion and ataxia; rhinorrhea, headache; tightening of chest, wheezing and laryngeal spasms. See NIOSH Pocket Guide to Chemical Hazards (DHHS/NIOSH 97-140, 1997) p 188. Use: Insecticide. Therap-Cat: Pediculicide. Therap-Cat-Vet: Ectoparasiticide.
Figure 4b. Second result for database search for MW 330-331. 5
The fragmentation ions shown in Figure 5 fit the ions from the MS/MS experiment of the m/z 331 (Figure 2). Therefore, malathion is a good possibility of an identification from the Database search of the Merck Index.
Intensity, cps OH S H3CO H3CO P S O C10H20O6PS2 Exact Mass: 331.0433 Error = 0.8 ppm
+
2.1 1.00e7 0.7 8.00e6 2.8 6.00e6 4.00e6 2.00e6 4.0 18.6 2.4 3.5 Star*TIC 27.5 22.5 28.6 24.9 26.0 27.8 30.5 21.1 23.3
O O
CH3 CH3 S H3CO H3CO P S Na
O O O O CH3 CH3
0.00 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 Time, min
Figure 6a. LC/MS TOF analysis of a cucumber extract.
-46
HO
C10H19NaO6PS2+ Exact Mass: 353.0253 Error = 1 ppm
H3CO H3CO
S P S O
When this formula is entered into the Merck Index database the only match is malathion (Figure 7).
OH
H3CO H3CO
S P S
O C8H14O5PS2+ Exact Mass: 285.0015 Error = 0.8 ppm HO
O C7H14O4PS2+ Exact Mass: 257.0066 Error = 0 ppm O
O C6H7O3+ Exact Mass: 127.0390 Error = 0 ppm
O H
C4H3O3+ Exact Mass: 99.0077 Error = 3.0 ppm
Thus, the accurate mass gives the same formula match that is consistent with the identification by ion trap. Furthermore, Figure 6b also shows the accurate mass ions for the possible fragmentation of the unknown at masses of m/z 353.0256, 285.0017, 127.0389, and 99.0081. These accurate mass ions are the same ones measured with the ion trap in MS/MS mode of the m/z 331 at nominal mass (with the exception of the m/z 353 adduct). The masses of these ions match the formulas shown in Figure 5 including the sodium adduct of the m/z 331. Thus, the probability of this being the correct identification is quite high. The final step is the identification by standard matching. This was done by both ion trap MS/MS and by TOF; both gave this compound as the correct identification. Malathion is a banned substance for cucumbers so this identification represents an important finding. Based on the identification point scheme for unknowns, this compound receives 2.0 points for the protonated molecule, 2.0 points for the m/z 285 ion, and 1.5 points for the MS/MS transition from m/z 331 to 285 for a total of 5.5 points. Four identification points are required for banned substances in food (see reference by Stolker et al. [5] in book by Ferrer and Thurman [6]). The complementary nature of the two instruments is also shown with the next unknown example.
Figure 5.
Fragmentation possibilities for malathion.
The next step is obtaining the accurate mass with the LC/MSD TOF. Figure 6a shows the TIC and the nearly identical retention times between the two instruments for the unknown Star*TIC because of using the same HPLC column. Figure 6b shows the accurate mass for the m/z 331.0435 and only one formula match of C10H19O6PS2.
OH 1.6e5 1.5e5 1.4e5 1.3e5 1.2e5 1.1e5 1.0e5 Intensity, counts 9.0e4 8.0e4 7.0e4 6.0e4 5.0e4 4.0e4 3.0e4 2.0e4 1.0e4 333.0404 O 127.0389 O H3CO H3CO S P S O OH 257.0066 353.0256 99.0081 HO H3CO O HO S P S O H3CO H3CO O S P S O Na O O O CH3 CH3 331.0435
Accurate Mass of 331 m/z 331.0435 Formula matches...C10H19O6PS2
S H3CO H3CO P S O
O O
CH3 CH3
285.0017
O H H3CO
60
80
100
120
140
160
180
200
220
240 m/z, amu
260
280
300
320
340
360
380
400
Figure 6b. LC/MS TOF spectrum of peak at 22.5 min (m/z 331).
Monograph Number: 5723 Title: Malathion CAS Registry Number: 121-75-5 CAS Name: [(Dimethoxyphosphinothioyl)thio]butanedioic acid diethyl ester
H3CO H3CO S P S
O O
CH3 CH3
Additional Names: diethyl mercaptosuccinate S-ester with O,O-dimethyl phosphorothioate; S-(1,2dicarbethoxyethyl) O,O-dimethyldithiophosphate; insecticide no. 4049; carbofos; malathon (obsolete); mercaptothion; phosphothion Manufacturers' Codes: ENT-17034 Trademarks: Cythion; Derbac-M (International); Malamar 50; Malaspray; Organoderm (Mundipharma); Prioderm (Coates & Cooper); Suleo-M (International) Molecular Formula: C10H19O6PS2
Molecular Weight: 330.36. Percent Composition: C 36.36%, H 5.80%, O 29.06%, P 9.38%, S 19.41% Literature References: Cholinesterase inhibitor. Prepn: Johnson et al., J. Econ. Entomol. 45, 279 (1952); Cassaday, US 2578652 (1951 to Am. Cyanamid). Purification: Usui, US 2962521 (1960 to Sumitomo). Treatment of Pediculus humanus (head lice) infestation: D. Taplin et al., J. Am. Med. Assoc. 247, 3103 (1982). Toxicity data: T. B. Gaines, Toxicol. Appl. Pharmacol. 14, 515 (1969). Review of distribution, transport and fate in the environment: M. S. Mulla et al., Residue Rev. 81, 1-159 (1981); of carcinogenic risk: IARC Monographs 30, 103-129 (1983). Properties: Deep brown to yellow liq, mp 2.9. bp 156-157. Characteristic odor. d 25 1.23. n 25 1.4985. Vapor pressure at 30: 4 10-5 mm Hg. Slightly sol in 4 D and alkylated aromatic hydrocarbons and vegetable oils. water (145 ppm). Misc with many organic solvents 0.7 including alcohols, esters, ketones, ethers, aromatic Limited soly in certain paraffin hydrocarbons. Petroleum ether is sol to about 35% in malathion. Hydrolyzed at pH >7.0 or <5.0. Stable in an aq soln buffered to pH 5.26. LD in female, male rats (mg/kg): 1000, 1375 orally (Gaines). 50 Melting point: mp 2.9 Boiling point: bp 156-157 0.7 Index of refraction: n 25 1.4985 D Density: d 25 1.23 4 Toxicity data: LD in female, male rats (mg/kg): 1000, 1375 orally (Gaines) 50 CAUTION: Potential symptoms of overexposure are miosis, aching eyes, blurred vision and lacrimation; eye and skin irritation; salivation; anorexia, nausea, vomiting, abdominal cramps, diarrhea, giddiness, confusion and ataxia; rhinorrhea, headache; tightening of chest, wheezing and laryngeal spasms. See NIOSH Pocket Guide to Chemical Hazards (DHHS/NIOSH 97-140, 1997) p 188. Use: Insecticide. Therap-Cat: Pediculicide. Therap-Cat-Vet: Ectoparasiticide.
Figure 7.
Database search for C10H19O6PS2. 7
Tomato Extract In this next example, a similar protocol is followed except that this time we use the LC/MSD TOF first to obtain an in-source CID spectrum. Figure 8a shows the Star*TIC for the rapid extraction of the tomato white powder.
23.9 3.0e7 2.5e7 2.0e7 Intensity, cps 14.7 1.5e7 1.0e7 2.2 5.0e6 0.0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Time, min 8.3 14.5 11.3 18.5 24.9 3.2 2.0 26.0 26.7 30.4
The accurate mass is m/z 306.1642, which results in two possible formulae after examination of the A+2 isotope pattern, which shows one sulfur atom in the molecule. Furthermore, the A+1 has an area relative to the m/z 306 of 17%, which indicates 15 carbon atoms. See Figure 8b.
4.0e5 306.1642 m/z 306.1642 C16H23N3OS C14H28N O2PS 2.4 ppm -3.0 ppm
Star*TIC
3.0e5 Intensity, counts
2.0e5
1.0e5 306.6200 0.0 280 290 300
307.1664 A+117%~C15 308.1624 A+25%S1 310 m/z, amu 320 330
Figure 8b. LC/MS TOF spectrum of peak at 23.9 min (m/z 306).
Figure 8a. LC/MS TOF analysis of a tomato extract.
The two formulae were both then searched in the Merck Index and only one formula gave a database hit. That compound was C16H23N3OS, the insect growth regulator buprofezin. Buprofezin is used extensively on white flies according to the Merck Index (Figure 9). Thus, this compound was a good candidate for a positive identification.
Monograph Number: 1486 Title: Buprofezin

N N O CH3 H3C S N C(CH3)3
CAS Registry Number: 69327-76-0
CAS Name: 2-[(1,1-Dimethylethyl)imino]tetrahydro-3-(1-methylethyl)-5-phenyl-4H-1,3,5-thiadiazin-4-one Additional Names: 2-tert-butylimino-3-isopropyl-5-phenylperhydro-1,3,5-thiadiazin-4-one Manufacturers' Codes: NNI-750; NNK-758; NN-29285; PP-618 Trademarks: Applaud (Nihon Nohyaku) Molecular Formula: C16H23N3OS
Molecular Weight: 305.45. Percent Composition: C 62.92%, H 7.59%, N 13.76%, O 5.24%, S 10.50% Literature References: Insect growth regulator which inhibits chitin synthesis. Prepn: Z. Grnecker et al., DE 2824126; K. Ikeda et al., US 4159328 (1978, 1979 both to Nihon Nohyaku); H. Kanno, Pure Appl. Chem. 59, 1027 (1987). Mode of action study: T. Asai et al., Appl. Entomol. Zool. 20, 111 (1985). Control of whiteflies and scale insects: I. Ishaaya et al., Meded. Fac. Landbouwwet., Univ. Gent 54, 1003 (1989). GC-MS determn in clementine citrus: P. Cabras et al., J. Agr. Food Chem. 46, 4255 (1998). Persistence in olives and olive oil: idem et al., Food Addit. Contam. 17, 855 (2000). Review of physical properties, activity and field trials: H. Kanno et al., Proc. Br. Crop Prot. Conf. - Pests Dis. 1981, 59-66. Properties: Crystals from isopropyl alcohol, mp 106.1. Soly at 25 (g/l): acetone 240, chloroform 520, ethanol 80, toluene 320; water 0.9 mg/l. Vapor pressure at 25: 9.4 10-6 mmHg. LD in mice, rats (mg/kg): 10000, 8740 orally. LC (48 hr) in carp: 2-10 mg/l (Kanno, 1981). 50 50 Melting point: mp 106.1 Toxicity data: LD Use: Insecticide. 50 in mice, rats (mg/kg): 10000, 8740 orally; LC 50 (48 hr) in carp: 2-10 mg/l (Kanno, 1981)
Results of Library SearchC14H28NO2PSNo hit. Molecular Formula = C16H23N3OS

Figure 9. Database search for C14H28NO2PS and C16H23N3OS. 8
Note in the spectrum the presence of the ion at m/z 201.1059 (Figure 10). LC/MSD ion trap MS/MS of the m/z 306 ion gave the m/z 201 and the further MS3 yielded the m/z 116 ion.
Intens. 107 1.0 306.1656 0.5 0.0 23.00 Intens. 6 10 3 2 1 115.9 1.0e5 201.1059 116 0.0 100 200 300 400 Formula = C9H17N2OS 500 600 m/z, amu 700 800 900 1000 0 MS3 gives 116
4.0e5
23.50
24.00
Intensity, counts
3.0e5
24.50 25.00 Time [min]
200.9 MS/MS gives 201
2.0e5
Furthermore, the expected mass for the 201.1056 fragment ion matched the value from the LC/MSD TOF (201.1059) quite closely (0.0003 u), which gave a high certainty for identification. After obtaining the buprofezin standard, the final data show a perfect match, which further shows the ability of the LC/MSD TOF and LC/MSD ion trap to identify unknowns. Buprofezin is allowed for tomatoes; therefore, it is not a banned substance. The quantitative tolerance for buprofezin may then be measured by the standard extraction and analysis for vegetable acceptance to European Union standards [1].
100 200 300 400 500 600 700 m/z
Conclusions
LC/MSD TOF and LC/MSD Trap are complementary and powerful tools for identification of pesticides in vegetables and represent a new approach for environmental food chemistry using LC/MS. The combination of these two tools, the twin mass spectrometer(s), with a pesticide database, Merck Index or ChemINDEX, works often for identification of unknown pesticides. The use of identification of fragment ions with MS2 and MS3 is also powerful when combined with accurate mass of LC/MSD TOF CID spectra. Combining TOF and Trap for unknown ID works [4, 7]! Future work should include chromatographic data to help in the identification of unknowns and to provide a simple pesticide library generated through calculation of empirical formula and isotope ratios.
Figure 10. LC/MS TOF spectrum and LC/MS/MS Ion Trap spectrum for m/z 306.
It was possible to draw reasonable chemical structures for the m/z 201 and 116 fragment ions (Figure 11).
Buprofezin m/z 306 S N N O H3C CH3 N C(CH3)3
MS2
S MS3 N N O H3C 201 m/z fragment ion C9H17N2OS Exact mass: 201.1056 CH3 C(CH3)3
H N m/z 116 C(CH3)3
Conclusion: Buprofezin Standard confirmationyes.
Figure 11. Reasonable structures of fragment ions consistent with the structure of buprofezin.
References
1. EU Food Directives, 2002, 91/414/EEC 2. Food Quality Protection Act, 1998 (FQPA) 3. Aguera, A.; Lopez, S.; Fernandez-Alba, A.R.; Contreras, M.; Crespo, J.; Piedra, W., 2004, J. Chromatogr. A., 1045: 125-135. 4. Imma Ferrer and E. Michael Thurman, "Determination of Chloronicotinyl Insecticides in Salad Vegetables by LC/MS/ESI/TOF and LC/MS Ion Trap", Agilent Technologies publication 5989-1842EN, www.agilent.com/chem 5. A.A.M Stolker, E. Dijkman, W. Niesing, E.A. Hogendoorn, 2003, "Identification of residues by LC/MS/MS" In: Imma Ferrer and E. Michael Thurman, editors Liquid Chromatography/Mass Spectrometry/Mass Spectrometry and Time-ofFlight Mass Spectrometry for the analysis of emerging contaminants, (2003) American Chemical Society Symposium Volume 850. 6. Imma Ferrer and E.M. Thurman, (2003), Liquid Chromatography/Mass Spectrometry/Mass Spectrometry and Time-of-Flight Mass Spectrometry for the analysis of emerging contaminants, American Chemical Society Symposium, 850. 7. E. Michael Thurman, Imma Ferrer, A. R. Fernandez-Alba, (2005), "Matching Unknown Empirical Formulas to Chemical Structure Using LC/MS TOF Accurate Mass and Database Searching: Example of Unknown Pesticides on Tomato Skins" Journal of Chromatography, Special Issue on Mass Spectrometry, In press.
Acknowledgments Amadeo Fernandez-Alba, Department of Hydrogeology and Analytical Chemistry, University of Almera for analytical support. ChemOffice with Chemfinder, the Merck Index, and ChemIndex are products of CambridgeSoft. ZORBAX is a registered trademark of E.I. duPont de Nemours Co. Inc. Millex is a registered trademark of the Millipore Corp. Merck Index is a registered trademark of Merck. Pyrex is a registered trademark of Corning Corporation. Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2005 Printed in the USA February 18, 2005 5989-1924EN
New Tools for Rapid Pesticide Analysis in High Matrix Samples Application
Food Analysis
Authors
Mike Szelewski and Bruce Quimby Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA
Tools like Method Translation [1] have made it straightforward to reduce analysis time by a known factor and maintain the exact relative elution order of the analytes. The use of appropriate shorter and smaller diameter columns can maintain the same resolution while achieving a much shorter analysis time. One application area where this approach has met difficulty, however, is the gas chromatography/ mass spectrometry (GC/MS) analysis of pesticides in complex matrices like food. This application requires that speed-up schemes maintain column capacity in order to handle the large matrix peaks and achieve low detection limits for analytes. Since chromatography is governed by the triangle of speed, resolution, and capacity, resolution must be sacrificed to increase speed at the same capacity. The problem is that chromatographic resolution is also needed to confirm the identity of any target analytes detected in the presence of interferences from the sample. In this note, the reduction in chromatographic resolution in faster analysis is more than adequately compensated for by use of spectral deconvolution [2] and simultaneous element-selective detection for the confirmation step. The system consists of a GC/MS with a dualwavelength flame photometric detector (DFPD) for the simultaneous collection of phosphorus, sulfur, and mass spectral data. The GC column effluent is split between the two detectors in the ratio of 2:1 in favor of the mass selective detector (MSD). The system is retentiontime locked to the Agilent pesticide library [1] scaled to threefold faster times, which contains the
Abstract
Recent developments in GC/MS hardware and software make it possible to analyze samples with high levels of matrix contamination much faster than ever before. New tools such as mass spectral deconvolution, reliable and inert effluent splitters, and column backflushing capabilities can be combined to produce large time savings. By accelerating the chromatographic run, post-run bakeout, and data interpretation steps, analysis times can be shortened by at least three-fold versus conventional methods. These tools are especially useful in analyses with high levels of matrix background, such as the inspection of the food supply for contaminants. In addition to monitoring for pesticide residues, the threat of terrorism has recently raised concerns over deliberate contamination of food with other toxic materials. This article describes a GC/MS system for the rapid screening of foodstuffs for chemical contaminants with a special emphasis on pesticides, organophosphorus, and organosulfur compounds.
Introduction
Techniques for decreasing the analysis time of gas chromatography (GC) methods have been developed in recent years.
retention times and spectra for 567 pesticides used worldwide. Samples are analyzed with MS in fullscan electron-impact ionization (EI) mode. The combination of precise retention times, elemental, and mass spectral data is used to screen for specific target compounds. The flame photometric detector (FPD) data also highlights any non-target, P- or S-containing compounds for identification by MS. The MS data is screened using the standard quantitation software based on retention time (RT), ion ratios, and spectral cross correlation. The MS data is also processed using spectral deconvolution software, which greatly reduces spectral interferences from the matrix. The deconvoluted spectra are then searched against a table of targets. Any hits are confirmed by searching against the main NIST library. This process is automated by the Agilent Deconvolution Reporting Software (DRS), also providing significant time savings in data interpretation.
The system described here uses column backflushing, a technique used to save large amounts of time with complex samples. Backflushing is done with the splitter hardware. This technique removes heavy residues from the column much faster and at lower temperatures than the conventional bakeout step at the end of the run. This reduces MS source contamination by preventing the higher levels of column bleed and heavy matrix components from entering the MSD. It also increases the column lifetime. The approach used thus reduces analysis time in three major ways: shortening the chromatographic run time; automating data interpretation; and reducing bakeout time. Other notable advantages are the ability to change columns and/or inlet liners without venting the MSD, and a reduced need for MS source cleaning. System Configuration The system configuration used is shown in Figure 1. Key components are:
Auto-sampler Dual Flame Photometric Detector

Sulfur Phosphorus
Effluent Splitter with Makeup
AUX EPC
0.814 m 0.18 mm id
Column
1.1 m 0.18 mm id
6890N GC
15 m 0.25 mm id 0.25 um HP-5MS
5973 Inert MSD
Figure 1.
System configuration.
Key Components Fast Oven With the 6890N 220V oven (option 002), the pesticide analysis method can be run precisely 3 times faster (14 min) using a 15 m HP-MS column. If the 220V GC is further equipped with SP1 2310-0236 (puts MSD interface in back of oven under rear injection port) and the G2646-60500 oven insert accessory (reduces oven volume twofold), the speed can be increased to 4.8 times faster (9 min). The cool-down time of the oven is also reduced. Dual FPD 6890N Option 241 is a single flame photometric detector with two optical detection channels, one for sulfur and one for phosphorus. The signals from the DFPD are collected, stored, and processed by the MS ChemStation simultaneously with the MS data. The FPD data can be used in several ways. Nontarget organophosphorus compounds like new pesticides or designer nerve agents are highlighted. The presence of an element at the retention time of an identified compound can be used to support confirmation of identity. The response on the FPD can be used for quantitative or semi-quantitative analysis, especially for situations where no calibration standard is available for an identified analyte. Microfluidic Splitter 6890N Option 889 uses diffusion bonded plate technology combined with metal column ferrules to make an inert, easy-to-use, leakfree, high-temperature column effluent splitter. The splitter uses Auxillary (Aux) electronic pneumatics control (EPC) for constant pressure makeup (6890N Option 301). The Aux EPC makeup can be pressure programmed at the end of the run to higher pressure, while at the same time the inlet pressure is lowered to near ambient. This causes the flow in the column to reverse direction, backflushing heavy materials out the split vent of the inlet. The Aux EPC also allows column changing and maintainance without venting the MSD. When the column fitting is removed from the splitter, helium from the makeup supply purges the fitting, preventing air from entering the MSD. If the column is attached to the splitter but removed from the inlet, helium flows backwards through the column and out the inlet end. Inlet maintainance or column headtrimming can be done without cooling and venting the MSD and air is not introduced into a hot source. MSD System The 5973N Inert with Performance Electronics and performance turbo (G2579A) EI MSD is used. This configuration provides faster full scan rates while maintaining sensitivity. The scan rates are compatible with the narrower peaks
generated by fast chromatography. The performance turbo pump is required to handle the higher flows associated with fast chromatography and backflushing. DRS Software (G1716AA) Spectral deconvolution of the MS data allows identification of analytes in the presence of overlapped matrix peaks. This significantly reduces chromatographic resolution requirements, allowing much shorter analysis times. DRS utilizes the AMDIS deconvolution program from NIST, originally developed for trace chemical weapons detection in complex samples. DRS presents the analyst with three distinct levels of compound identification: (1) ChemStation, based on retention time and four ion agreement; (2) AMDIS, based on cleaned spectra full ion matching and locked retention time; (3) NIST02 search using a >147,000 compound library. Instrument Operating Parameters The recommended instrument operating parameters are listed in Table 1. These are starting conditions and may have to be optimized. Split injection was used to match the amount of matrix to the column capacity. Citrus oils cause retention shifts if excess sample is injected. Splitless injection could be used for samples with significantly less matrix. The inlet liner was found to be of low activity, as it does not contain glass wool. Proper mixing for split injections is done by the internal liner geometry. The 6890 220V oven was needed for the ramps described in Tables 1 and 2. This oven program is necessary for the precise 3 speed increase of the RTLocked pesticide database. The 15-m HP-5ms column has the same phase ratio as the 30 m column traditionally used for the 1 method. This shorter column allows a flow rate for a 3 precisely scaled faster method. The outlet is listed as unspecified because the column connects to the splitter. The splitter pressure is operated at a constant 3.8 psig using an auxillary EPC module. The 5973 inert Performance Electronics data acquisition sampling rate was set to 1, which is faster than the typical setting of 2. Signal-to-noise is improved over previous systems at faster sampling rates. More data points allows for easier integration and better deconvolution to compensate for the loss in resolution using a shorter column. The microfluidic splitter parameters are chosen to provide the desired split ratio between detectors
3
while meeting the flow requirements of the detectors used. A primary consideration with the current system is to make sure that the flow to the MSD does not exceed ~4 mL/min while collecting analyte data. It was also desired to split the effluent 2:1 in favor of the MSD. These parameters were entered into the spreadsheet calculator (included with the splitter), which calculated the lengths and diameters of the detector restrictors
Table 1. GC Inlet Mode Inlet temp Pressure Split ratio Split flow Total flow Gas saver Gas type Inlet Liner Oven Oven ramp Initial Ramp 1 Ramp 2 Ramp 3 Ramp 4 Total run time Total run time Equilibration time Oven max temp Column Length Diameter Film thickness Mode Inlet Outlet Outlet pressure Back Detector (FPD) Temperature Hydrogen flow Oxidizer flow Oxidizer gas type Mode Makeup flow Makeup gas type Flame Lit offset Photo multiplier Gas Chromatograph and Mass Spectrometer Operating Parameters Agilent Technologies 6890 EPC Split/Splitless Split, 1.0 L injected 250 C 23.84 psi 10:1 44.1 mL/min 48.1 mL/min Off Helium Siltek Cyclosplitter, 4 mm id, Restek part number 20706-214.1 220V C/min 75 9 24 50 Next C 70 150 200 280 320 Hold min 0.67 0.00 0.00 3.33 (end of pesticide ramp) 50.0 (end of oil elution)
13.96 min to elute pesticides 64.76 min to elute heavy components from citrus oils 0.5 min 325 C Agilent Technologies HP-5MS, p/n 19091S-431 15.0 m 0.25 mm 0.25 m Constant Pressure = 23.84 psi Front Unspecified 3.8 psi (aux pressure to splitter) 250 C 75.0 mL/min 100.0 mL/min Air Constant makeup flow 60.0 mL/min Nitrogen On 5.00 On
Table 1.
Gas Chromatograph and Mass Spectrometer Operating Parameters (Continued) 5 Hz Back detector On 0.0 (Off) 0 Off 0 Signal 2 Data rate: Type: Save data: Zero: Range: Fast Peaks: Attenuation: 5 Hz Front detector On 0.0 (Off) 0 Off 0
Signal 1 Data rate Type Save data Zero Range Fast Peaks Attenuation AUX Pressure 5 Description Gas type Initial pressure Initial time MSD Tune file Mode Solvent delay EM voltage Low mass High mass Threshold Sampling Scans/sec Quad temp Source temp Transfer line temp Splitter Split ratio MSD restrictor DFPD restrictor
Helium 3.80 psi 0.00 min (this value will follow oven ramp) Agilent Technologies 5973 Inert Performance Electronics Atune.U Scan 1.00 min Atune voltage 45 amu 450 amu 0 1 6.68 150 C 230 C 280 C Agilent 6890N Option 889 2:1 MSD:DFPD 1.1 m 0.18 mm id deactivated fused silica tubing 0.81 m 0.18 mm id deactivated fused silica tubing
Backflush Instrument Operating Parameters Instrument operating conditions for backflushing are shown in Table 2. These are starting parameters and will have to be optimized depending on sample matrix. Conditions listed here are only those that are different from the Table 1 conditions. Instead of baking the column at 320 C for 50 min, the heavy matrix material is backflushed through the column inlet, out the split vent. This is accomplished in 5 min at 300 C, saving 45 min of run time, preserving column life, and shortening cool down time. The column inlet pressure is reduced to 1.1 psi by using the ramped pressure feature of the EPC. At the end of the backflush time, it is ramped back to initial conditions.
Simultaneous with ramping the inlet pressure down to 1.1 psig, the Aux EPC splitter pressure is ramped up to 23 psig. This increase in pressure at the column outlet, along with the decrease in inlet pressure, backflushes the column. The backflush pressure for the Aux EPC is set to limit the flow to the MSD to < 8 mL/min. This is calculated using the Agilent Flow Calculation software program, also supplied with the splitter kit. The calculator is used to find the pressure which gives the desired flow through the MSD splitter restrictor (1.1 m 0.18 mm id) at the backflushing temperature 300 C. The result was 23.6 psig which would produce a flow of 8 mL/min, so 23 psig was used. The backflush time is determined empirically and depends on the sample matrix. The process is to try a backflush run followed by a blank run with the conventional long-hold to see if the heavy materials
are completely removed. If not, the process is repeated with a longer backflush time. As a very rough guide, start with a backflush time of five void times for the backwards flow. Using the Agilent Flow Calculation software with the inlet pressure at 23 psig, the outlet pressure at 1.1 psig, and the temperature at 300 C, the hold-up time (void time) is 0.423 min. The rough guide says that the column should be backflushed for 2.12 min. This works for removing most heavies, but 5 min is required in this case to remove them all. At the end of the backflush time, the Aux EPC is ramped back to initial conditions. The MSD is time-programmed to collect data over the time range of 1 to 13.96 min. All of the pesticides elute during this time range.
Table 2.
Backflush Gas Chromatograph and Mass Spectrometer Operating Parameters. Use Table 1 Conditions Except for These Differences C/min 75 9 24 50 Next C 70 150 200 280 300 Hold min 0.67 0.00 0.00 3.33 (end of pesticide ramp) 5.40 (end of oil backflush)
Backflush Oven Conditions Oven ramp Initial Ramp 1 Ramp 2 Ramp 3 Ramp 4 Total run time Total run time
13.96 min to elute pesticides 19.76 min to elute heavy components from citrus oils
Backflush Column Conditions Mode Ramped Pressure Press ramp psi/min Initial Ramp 1 99 Ramp 2 99 Backflush AUX 5 Conditions Press ramp psi/min Initial Ramp 1 99 Ramp 2 99 Backflush MSD Conditions Timed MS detector entries Time (min) 13.96
Next psi 23.84 1.1 23.84 Next psi 3.8 23.0 3.8
Hold min 13.96 (end of pesticide ramp) 5.57 (backflush time) 0.00 Hold min 13.96 (end of pesticide ramp) 5.61 (backflush time) 0.00
State (MS on/off) Off
Results
A mixture of 25 pesticides was run at 1, 3, and 4.8 speeds. The Total Ion Chromatograms (TICs) are shown in Figure 2. There is some loss in resolution, but the loss is limited because the shorter columns are run at flows closer to the optimum. The resolution losses are mitigated by using the faster scanning capabilities of the performance electronics together with DRS.
1 30 m
5.00
10.00
15.00
20.00
25.00
30.00
35.00
40.00
3 15 m
2.00
3.00
4.00
5.00
6.00
7.00
8.00
9.00
10.00
11.00
12.00
13.00
4.8 10 m
1.00
1.50
2.00
2.50
3.00
3.50
4.00
4.50
5.00
5.50
6.00
6.50
7.00
7.50
8.00
8.50
Figure 2.
TICs of pesticide test mix at three different scaled speeds. All columns have the same phase ratio.
A neat lemon oil was analyzed using the 3 speed conditions. The TIC is shown in Figure 3 with the DFPD phosphorus (P) and sulfur (S) data channels. The ChemStation software can simultaneously acquire two signals of GC data with the MSD data. The pesticides elute within the 14 min window shown, but the matrix continues to elute for an additional 50 min at 320 C.
TIC
7.441 min
FPD (P)
7.257 min
FPD (S)
10
12
14
Figure 3.
MS and DFPD data from lemon oil analyzed with 3x method.
The P and S chromatograms indicate the possible presence of numerous pesticides. The largest P peak, 7.441 min, also contains S. A PBM reverse Library search identified the peak as Methidathion (C6H11N2O4PS3) with a match quality of 45. It was also identified using a target ion and three qualifier ions. The raw apex spectrum of the P peak at 7.257 min is shown in the top of Figure 4. No match was found for a P-containing compound in the top 20 library search results. It was also not identified by the ChemStation target and qualifier ion criteria due to out-of-range ratios.
69 81 93 55 131
Raw spectrum at 7.257 min

109 121 137 159 171 187
180
206
200
227 241
220 240
256 270 283 296

260 280 300
314
329
320
40
60
80
100
120
140
160
100
58 86 76 65
131 97 159
Deconvoluted spectrum
296
171 198 226 252
329
0
86 97
50 80 110
296
329
159 131
140 170
100
Library spectrum of Mecarbam

200 230 260 290 320
Figure 4.
Top - Raw mass spectrum of peak at 7.257 min in lemon oil. Bottom - Deconvoluted spectrum of 7.257 min peak compared to NIST02 library spectrum of Mecarbam.
The DRS report is shown in Figure 5. The peak at 7.257 min is clearly identified as Mecarbam by the DRS software. The deconvoluted spectrum has a match factor of 72, compared to both the Agilent Pesticide AMDIS database and the NIST02 library. Additionally, the DRS report shows the elution time to be only 0.5 s from expected. Further confirmation is the presence of P with S barely visible. The deconvoluted spectrum for the peak at 7.257 min is shown at the bottom of Figure 4, together with the NIST library spectrum of Mecarbam. The DRS report displays the amount for compounds found by the normal ChemStation quant process. The compounds must be properly
calibrated to have a meaningful value. In this lemon oil, the ChemStation found four compounds. The AMDIS portion of DRS found two of the same compounds and an additional five compounds missed by the ChemStation. The NIST02 library search confirmed all of the compounds found by AMDIS using the NIST02 >147,000 compound library. The DRS results show good match quality at the locked retention times for seven target compounds. No single software package can produce this same confidence level in compound identification. An experienced analyst would take 14 hours to process this complex sample manually with each of the three software packages used by DRS. DRS produced this report is less than two minutes.
Figure 5.
DRS Report for lemon oil.
10
Backflushing Citrus oils contain significant amounts of material that elute after the last pesticide. This requires a 150-min hold at 320 C to elute all of the heavy material with a 1 method. The total run time for the 1 method is therefore 195 min, as shown in Table 3.
Table 3. Method Run Time Comparison 30 m 1 min 42 195 n/a 15 m 3 min 14 65 20 10 m 4.8 min 8.75 40.6 12.5
The 3 method requires a 50-min hold at 320 C, as shown at the top in Figure 6, resulting in a 65-min run time. With backflushing, all of this heavy material is removed in 5 min at 300 C, as shown in the bottom of Figure 6. This is a 9-fold reduction in analysis time compared to the 1 method. 4.8x Method Using the 220V oven, SP1 2310-0236, and oven insert accessory, the method can be scaled to 4.8 faster, as shown in Table 3. There is a practical limit to the amount of matrix that can be tolerated with the reduced resolution using a 10-m column. However, for matrices less complex than a citrus oil, the 4.8 method can be successfully used to save even more time. The Performance Electronics allows running at faster scan speeds while maintaining signal/noise ratio. Sufficient points across a peak are maintained even with faster chromatography.
Column Speed Run time Pesticides No backflush matrix With backflush matrix
Normal bakeout: 65 min run
Backflush time range
Backflush: 19 min run
10
20
30
40
50
60
Figure 6.
Top - 3 lemon oil analysis with 50 min bakeout at 320 C. Bottom - 3 lemon oil analysis with 5 min backflush at 300 C.
11
Conclusions
New tools are available for the analysis of pesticides in complex matrices. These tools can be combined to significantly reduce analysis and data processing time. Fast oven - programming rates needed for faster runs Shorter column - 3 faster runs with sufficient resolution Microfluidic splitter - confidence in results using selective detection simultaneous with MS data Backflush - additional 3 reduction in run time with lower column temperatures for extended lifetime Performance Electronics - maintain signal/noise at faster sampling rates DRS - three levels of target analyte identification in less than two minutes Using the above tools, the run time for analysis of lemon oil was reduced from 195 minutes to 20 minutes (nine-fold). DRS reduced the data analysis from hours to minutes.
References
1. B. D. Quimby, L. M. Blumberg, M. S. Klee, and P. L. Wylie Precise Time-Scaling of Gas Chromatographic Methods Using Method Translation and Retention Time Locking Agilent Technologies publication 5967-5820E www.agilent.com/chem 2. P. Wylie, M. Szelewski, and C. K. Meng Comprehensive Pesticide Screening by GC/MSD using Deconvolution Reporting Software Agilent Technologies publication 5989-1157EN www.agilent.com/chem

A Blind Study of Pesticide Residues in Spiked and Unspiked Fruit Extracts Using Deconvolution Reporting Software Application
Food
Author
Christopher P. Sandy Agilent Technologies, Inc. Block A, CSC Eskdale Road Winnersh Triangle Estate Wokingham, Berk RG41 5DZ United Kingdom
Abstract
The Agilent Technologies mass selective detector (MSD) coupled with deconvolution reporting software (DRS) provides additional powerful data processing capabilities to the MSD ChemStation software. Reviewing full scan gas chromatography/mass spectrometry data for the confirmation of pesticide residues can be a labor-intensive and time-consuming process requiring great skill and concentration by an experienced analyst. The DRS is able to process a complex food extract total ion chromatogram in about 1 minute, whereas an experienced analyst may take more than 30 minutes to achieve the same quality result. Extensive data shown in this report supports the high confidence level that an analyst can have in results rapidly produced by the DRS.
ion ratios. It is sometimes very difficult to confirm target compounds from high matrix background because the matrix affects the ion ratios of the target compounds or complicates the spectrum with additional ions. To be certain of the results, background subtraction and manual integration are often practiced. It is, therefore, a timeconsuming process to confirm target compounds in a dirty matrix. It can take an experienced analyst 15 to 30 minutes to review/confirm one data file. Two powerful gas chromatography/mass spectrometry (GC/MS) techniques - Retention Time Locking (RTL) and deconvolution were combined to create a quantitation and screening tool that can identify 567 pesticides and endocrine disrupters from a single run in 12 minutes. The Agilent Technologies GC/MSD-DRS provides the additional functionality to the MSD ChemStation.
Experimental
DRS Overview A detailed overview of the DRS is given in an application note 5989-1157EN [1], available for download at www.agilent.com/chem. The operating principles of the DRS appear in Figure 1.
Introduction
Typical mass spectral pesticide residue analysis requires finding target ions and meeting qualifier
2.5E7 2E7 1.5E7 1E7 .5E7 5 10 15 20 25 30 35 40
Total ion chromatogram (TIC)
Targets are identified by comparison to locked retention times (RTs) and three qualifying ion ratios, quantified using target ion area versus internal standard (ISTD) calibration table Quant results
AMDIS 32 deconvolutes component spectra and searches target MS database, locked RT used as a qualifier
Deconvoluted target spectra confirmed by AMDIS searched against NIST02 MS database
Confirmed AMDIS hits
Confirmed NIST02 hits
Combined quantitative and qualitative HTML Summary report
Figure 1.
Schematic diagram summarizing the GC/MS DRS.
The quantitation capabilities of the MSD ChemStation are combined with the deconvolution power of the industry standard AMDIS program from NIST. AMDIS is able to separate spectra of interest from dirty matrix spectra present in samples analyzed for pesticides. A third level of confidence is obtained by sending the deconvoluted spectra for library searches of the NIST02 145,000 compound library. A comprehensive report is produced in about 1 minute.
Samples Six samples of fruit extracts, supplied in 90/10 iso-octane/toluene solvent were received for analysis by GC/MS. The samples were prepared by an accredited food pesticide laboratory based in Scandinavia. Three of the samples were spiked with a number of pesticides at varying concentration levels. Although the range of concentrations of the pesticides in each sample was given, neither the actual number of pesticides spiked into each control sample nor the identities were supplied. Details of the samples appear in Table 1. The other three samples were real, unspiked extracts.
Table 1. Sample number 1 2 3 4 5 6
Sample Details for Blind Study Matrix extracted Orange Lettuce Apple Grapes Orange Apple Number of pesticides 2040 2040 2040 24 24 24 Concn range (mg/Kg) 0.020.20 0.020.20 0.010.20 0.11.0 0.25.0 0.052.0 Comments Control sample - spiked Control sample - spiked Control sample - spiked Real sample Real sample Real sample
Instrumentation
The samples were analyzed by full-scan GC/MS using the analytical conditions given in Table 2. Data processing and reporting were performed using the default settings provided with the DRS.
Table 2.
RTL GC/MS Analysis Conditions for Fruit Extract Samples Agilent 6890N 30 m x 0.25 mm id x 0.25 m HP-5MS (p/n 19091S-433) Helium 1.9 mL/min at 70 C 18 psig, constant pressure mode Method RTLocked to methyl chlorpyrifos at 16.593 min PTV, septumless head 90 C (0.3 min) - 1720 C/min - 250 C 0.2 min 30 mL/min 0 psig 60 mL/min 1.0 min 50 L 15 L Empty multibaffle 70(2)-25-150(0)-3-200(0)-8-280(10)
Gas chromatograph Column Carrier gas Flow rate Head pressure Injector type Injector temperature (C), hold time (min), and ramp rate (C/min) Vent time Vent flow Vent pressure Purge flow Purge time Syringe volume Injection volume Liner Oven program: temperature (C), hold time (min), and ramp rate (C/min) MSD MS interface MS source MS quad Detection mode EM voltage
Agilent 5973 inert 280 C 230 C 150 C EI, Scan 40550 amu ATUNE value
Results
The results for the three spiked extracts appear in Table 3 - note that the details of which pesticides were added to the spiked samples were not supplied until after the results were shown to the customer. Those pesticides confirmed by the DRS, are shown lightly shaded. The analytes, shown darkly shaded, are not present in the Agilent RTL Pesticides database. Analyte entries left unshaded were not confirmed.
Table 3.
MSD-DRS Results for Three Spiked Fruit Extract Samples Sample 2: Control-lettuce, spiked Added mg/kg 0.10 0.10 0.10 0.10 0.20 0.10 0.10 0.04 0.02 0.10 0.10 0.10 0.04 0.10 0.04 0.10 0.10 0.10 0.10 0.10 0.10 0.10 0.10 0.04 0.04 0.10 0.10 0.10 0.10 0.10 0.04 0.04 Pesticide Diphenylamine HCB Lindane (HCH-gamma) Diazinon Chlortalonil Vinclozolin Carbaryl Metalaxyl Pirimiphos-methyl Malathion Chlorpyrifos Cyprodinil Penconazole Captan Folpet** Procymidone Endosulfan-a pp-DDE Bupirimate Endosulfan-b Aclonifen Ethion Triazophos Endosulfan-sulfate Iprodione Bromopropylate Methoxychlor Phosalone Lambda-Cyhalothrin Permethrin Cypermethrin Fenvalerate Deltamethrin Added mg/kg 0.10 0.02 0.04 0.04 0.04 0.04 0.20 0.10 0.10 0.10 0.10 0.04 0.04 0.10 0.10 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.10 0.10 0.10 0.04 0.10 0.10 0.04 0.10 Sample 3: Control- apple, spiked Pesticide Mevinphos Trichlorfon Heptenophos Tecnazene HCH alpha HCH beta Dichloran Pyrimethanil Etrimphos Ethiofencarb Metribuzin Toclophos methyl Linuron Aldrin Diethofencarb Trichloronate Triadimenol Disulfoton sulfoxide Disulfoton sulfone Fluazinam Chlorbenzilate Oxadixyl Benalaxyl Dicofol Fenazaquin Pyrazophos Acrinathrin Bitertanol Cyfluthrin beta Alpha cypermethrin Added mg/kg 0.05 0.05 0.02 0.01 0.01 0.02 0.05 0.02 0.02 0.10 0.05 0.01 0.05 0.02 0.02 0.02 0.05 0.20 0.02 0.05 0.05 0.05 0.05 0.05 0.02 0.05 0.02 0.05 0.05 0.05
Sample 1: Control-orange, spiked Pesticide Methamidofos* Dichlorvos* Acephate* Omethoate Propachlor Chlorprofam Monocrotophos Dimethoate Quintozene Parathion-methyl Dichlofluanid Fenpropimorph Triadimefon Thiabendazole Tolylfluanid Mecarbam Methidation Vamidothion Imazalil Myclobutanil Kresoxim methyl Tebuconazole Phosmet Fenpropathrin Tetradifon Azinphos-methyl Fenarimol Azinpfos-ethyl Prochloraz Flucythrinate Esfenvalerate Azoxystrobin
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 * See Discussion item 1. ** See Discussion item 2.
The results for the three real extracts appear in Table 4. Those pesticides confirmed by the DRS are shown lightly shaded. The darkly-shaded analytes are not present in the Agilent RTL Pesticides database. Analyte entries left unshaded were not confirmed. Analytes with an associated concentration were confirmed as present by the customer using NPD/ECD. Lightly-shaded analytes without a concentration label were detected and confirmed by the DRS, but not by the customer.
Table 4. MSD-DRS Results for Three Real Fruit Extract Samples Sample 4: Grapes 0.68 mg/Kg Captan 0.21 mg/Kg Cyprodinil 0.27 mg/Kg Fludioxinil Diphenylamine Sample 5: Orange 2.5 mg/Kg Imazalil 0.25 mg/Kg Medidathion 3.0 mg/Kg Thiabendazole Sample 6: Apple 0.86 mg/Kg Diphenylamine 0.05 mg/Kg Chlorpyrifos 0.79 mg/Kg Thiabendazole Dimethoate Ethoxyquin Methyl parathion Endosulfan sulfate Propargite
2. Control - Lettuce spiked extract This control sample was spiked with 33 pesticides at levels ranging between 0.02 and 0.20 mg/kg. Twenty-nine pesticides were detected and confirmed by the DRS software, three were not reported since they are not present in the Agilent RTL Pesticide database and one was not detected. The one undetected analyte, (Folpet, marked with two asterisks in Table 3), was detected and confirmed if a higher sensitivity setting was used in the AMDIS deconvolution program. 3. Control - Apple spiked extract This control sample was spiked with 30 pesticides at levels ranging between 0.01 and 0.20 mg/kg. Twenty-two pesticides were detected and confirmed by the DRS software, six were not reported since they are not present in the Agilent RTL Pesticide database and two were not detected. Overall, of the 95 spiked analytes in the three control samples, 93% of the pesticides present in the Agilent RTL Pesticide database were detected and confirmed by full-scan library searching of the deconvoluted mass spectra. 4. Real Grape extract The customer had detected and confirmed three pesticide residues in the Grape extract sample Captan, Cyprodinil, and Fludioxinil. Of these three analytes, Captan was confirmed by the DRS and Cyprodinil and Fludioxinil are not entries in the Agilent RTL Pesticide database. However, DRS also confirmed an additional pesticide residue Diphenylamine, which was not reported by the customer. 5. Real Orange extract The customer had detected and confirmed three pesticide residues in the Orange extract sample Imazilil, Methidathion, and Thiabendazole. All three of these pesticides were confirmed by the DRS software and no other analytes were confirmed.
Discussion
1. Control - Orange spiked extract This control sample was spiked with 32 pesticides at levels ranging between 0.02 and 0.10 mg/kg. Twenty-six pesticides were detected and confirmed by the DRS software, two were not reported since they are not present in the Agilent RTL Pesticide database and four were not detected. The spiking was done to the raw matrix, not to a matrix extract. For the polar pesticides (methamidofos and acephate), the recovery was in the 20%30% range as confirmed by NPD/ECD. Therefore, that explains why these pesticides were not detected by DRS.
6. Real Apple extract The customer had detected and confirmed three pesticide residues in the Apple extract sample Diphenylamine, Chlorpyriphos, and Thiabendazole. All three of these pesticides were confirmed by the DRS. In addition, the DRS also confirmed the presence of five additional pesticide residues Dimethoate, Ethoxyquin, Methyl Parathion, Endosulfan Sulfate, and Progargite. These five pesticides had not been reported by the customer. The extensive data shown in this report, run under totally blind conditions, shows the high degree of confidence that an analyst can have in the results produced by the DRS in minutes.
Reference
1. Philip L. Wylie, Michael J. Szelewski, Chin-Kai Meng, and Christopher P. Sandy, Comprehensive Pesticide Screening by GC/MSD Using Deconvolution Reporting Software, Agilent Technologies, publication 5989-1157EN, www.agilent.com/chem
Conclusions
The Agilent Technologies MSD-DRS provides additional powerful data processing capabilities to the MSD ChemStation software. Reviewing full scan GC/MS data for the confirmation of pesticide residues can be a labor-intensive and time consuming process requiring great skill and concentration by an experienced analyst. The DRS is able to process a complex food extract TIC in the order of 1 minute, whereas an experienced analyst may take more than 30 minutes to achieve the same quality result. The DRS software was proven to report the lowest number of false positives and false negatives in the shortest time period. In scan mode, the detection limit is not as low as in selected ion monitoring (SIM) mode; however, any prior knowledge of the target analytes (retention times or characteristic ions) is not required for the DRS.

Analysis of Organochlorine and Pyrethroid Pesticides with Agilent 6820 Gas Chromatograph/Micro-Electron Capture Detector Application
Environmental and Food Analysis
Author
Chuanhong Tu Agilent Technologies Co., Ltd. (Shanghai) 412 YingLun Road Waigaoqiao Free Trade Zone Shanghai, 200131 P. R. C.
To address these problems, the ECD, developed by Agilent Technologies, uses a smaller flow cell. The ECD is optimized for capillary columns and designed for improved sensitivity. It was successfully used with an Agilent 6890 series GC with better detector sensitivity and a wider linear range. In this note, the Agilent 6820 GC with ECD was used to determine organochlorine and pyrethroid pesticides following the Chinese National Standard Method GB/T 5009.146-2003 [2]. System sensitivity and limits of detection (LOD) were examined for organochlorine and pyrethroid compounds.
Abstract
The Agilent 6820 gas chromatograph (GC) with microelectron capture detector (ECD) was used to analyze organochlorine and pyrethroid compounds. All compounds demonstrated good linearity among a wide concentration range. The sensitivity provided by ECD was much better than the requirements of the routine pesticide residue analysis.
Experimental
All experiments were performed on an Agilent 6820 GC with split/splitless inlet and ECD. Single-tapered deactivated liner (p/n 5183-4696) and Agilent green septa (p/n 5183-4759) were used. Cerity Networked Data System (NDS) software was used for instrument control, signal acquisition, and data processing. Samples were manually introduced into the GC with a 10-L micro-syringe (p/n 5182-3428). Experimental conditions are listed in Table 1. All organochlorine and pyrethroid compounds were diluted with hexane.
Introduction
The electron capture detector (ECD) is a type of detector with high sensitivity and selectivity for halogenated compounds. However, there are some drawbacks in ECD design. The ECD is inherently nonlinear, with a limited linear range. Due to the narrow linear range, sample concentration or dilution, and re-analysis have to be employed, resulting in lower productivity. In addition, in traditional ECD design, a large flow cell is necessary to be compatible with both packed and capillary columns, leading to lower detector sensitivity [1].
Table 1. Instrumental Parameters Instrument Agilent 6820 GC Software Inlet Injection volume Column Carrier Oven Cerity NDS Chemical for QA/QC Split/Splitless; 250 C; splitless mode; purge time: 0.75 min 2 L HP-1, 30 m 0.32 mm 0.25 m (p/n 19091Z-413) Nitrogen, 6.0 psi, constant head pressure mode, 1.2 mL/min (60 C) 60 C (1 min), 30 C/min to 180 C, 5 C/min to 250 C (5 min), 3 C/min to 280 C (10 min) ECD; 330 C; make-up: nitrogen, 60 mL/min
Results
ECD Sensitivity A chromatogram of organochlorine and pyrethroid pesticides on an HP-1 column is shown in Figure 1. The concentrations, for four benzene hydrochlorides (BHCs), heptachlor, aldrin, heptachlor epoxide, and six pyrethroids are 10 ppb; for p,p'-DDE, dieldrin, endrin, endosulfan I, and endosulfan II, 20 ppb; and for p,p'-DDD, endrin aldehyde, endosulfan sulfate, and p,p'-DDT, 60 ppb. Except for endrin and endosulfan II, the other 20 compounds were fully separated. Among pyrethroid pesticides, two permethrin, four cypermethrin and two fenvalerate isomers were separated.
Detector
Hz 1400
Peak identification
1200 1. 2. 3. 4. 5. 6. 7. 8. 3 600 4 5 400 2 6 7 -BHC -BHC -BHC -BHC Heptachlor Aldrin Heptachlor epoxide Endosulfan I 9. 10. 11. 12. 13. 14. 15. 16. p,p'-DDE Dieldrin Endrin Endosulfan II p,p-DDD Endrin aldehyde Endosulan sulfate p,p'-DDT 8 10 12 11 14 13 15 16
1000
800
200
10 Hz 240
12
14
16
18
20
min
18 17. 18. 19. 20. 21. 22. 17 Fenpropathrin Cyhalothrin Permethrin Cypermethrin Fenvalerate Deltamethrin
220
200
180 19 160
20
21 22
140
120 25 27.5 30 32.5 35 min
Figure 1. 2
Chromatogram of organochlorine and pyrethroid pesticides on HP-1 column.
The signal-to-noise ratios are larger than 20 for organochlorine pesticides at concentrations of 16 ppb. For 5 ppb pyrethroids, the signal-to noise ratios are larger than 10. They can be easily quantitated. Therefore, the ECD provides more than enough sensitivity to meet the requirements of quantitative analysis of pesticides residues. Linear Range and Response Factors The calibration curves of -BHC and permethrin, typical of organochlorine and pyrethroid pesticides, are shown in Figures 2 and 3, respectively. The linear range and response factors (RFs) are listed in Table 2. The RFs are the ratios of compound concentrations to peak areas. The relative standard deviations (RSDs) of RFs are less than 20%, better than the precision requirements for response factors in the contract laboratory program of the USEPA (the United States Environmental Protection Agency). The linear correlation coefficients for all compounds are better than 0.995.
pA 60000
40000
-BHC y = 119.98x + 354.05 R2 = 0.9988
20000
0 0 100 200 Concentration, ppb 300 400
Figure 2.
pA 4000
Calibration curve of -BHC, a typical organochlorine pesticide.
3000
Permethrin y = 9.0737x + 97.257 R2 = 0.9986
2000
1000
0 0 100 200 Concentration, ppb 300 400
Figure 3.
Calibration curve of permethrin, a typical pyrethroid.
Table 2. Compound -BHC -BHC -BHC -BHC Heptachlor Aldrin Heptachlor epoxide Endosulfan I p,p'-DDE Dieldrin Endrin Endosulfan II p,p'-DDD p,p'-DDT Endrin aldehyde Endosulfan sulfate Fenpropathrin Cyhalothrin Permethrin Cypermethrin Fenvalerate Deltamethrin Linearity Results for Organochlorine and Pyrethroid Pesticides Average RF 0.0064 0.0159 0.0078 0.0087 0.0097 0.0078 0.0100 0.0109 0.0088 0.0123 0.0152 0.0121 0.0379 0.0175 0.0139 0.0140 0.0500 0.0225 0.1007 0.0809 0.0728 0.0461 %RSD of RF 7.5 18.1 9.4 8.8 13.2 8.0 11.7 11.7 11.0 13.0 16.3 15.8 8.7 15.8 10.6 10.0 18.0 8.6 10.2 7.2 6.9 15.0 Linear range (ppb) 1400 1400 1400 1400 1400 1400 1400 2800 2800 2800 2800 2800 62400 62400 62400 62400 1400 1400 10400 10400 10400 10400 R2 0.9983 0.9984 0.9988 0.9986 0.9987 0.9982 0.9983 0.9983 0.9948 0.9981 0.9991 0.9987 0.9983 0.9972 0.9989 0.9988 0.9951 0.9982 0.9986 0.9991 0.9990 0.9996
Conclusion
The Agilent 6820 GC/ ECD system shows good sensitivity and wide linear range for organochlorine and pyrethroid pesticides, and are much better than routine pesticide residue analysis requirements.
References
1. Channel, I., and Chang, I. L., Analysis of Organochlorine Pesticides and PCB Congeners with the Agilent 6890 Micro-ECD, Agilent Technologies, Publication (23) 5965-8556E www.agilent.com/chem 2. China National Standard Method GB/T 5009. 146-2003, Multiresidue analytical methods for organochlorine and pyrethroid pesticides for plant-originated food, August, 2003

Analysis of Organophosphorus Pesticides with Agilent 6820 Gas Chromatograph/ Nitrogen Phosphorus Detector Application
Author
Chuanhong Tu Agilent Technologies Co., Ltd. (Shanghai) 412 YingLun Road Waigaoqiao Free Trade Zone Shanghai, 200131 P. R. C.
Abstract
Fifteen of the most common and important organophosphorus pesticides (OPs) were studied using the Agilent 6820 gas chromatograph (GC) equipped with the new nitrogen-phosphorus detector (NPD). The 6820 GC-NPD demonstrated good linearity for concentrations of pesticides range from 1 to 500 ng/mL (R2 >0.999) for most compounds. All OPs produced high signal-to-noise ratios for splitless injections at 10 ng/mL (ppb) concentrations with the NPD detector. Instrumental limits of detection for most OPs studied were at low or sub-ppb concentrations. This suggests the 6820 GC with an NPD is well suited for OP pesticide residue determinations in foods, water, or other samples.
food safety and the identities and residual concentrations of pesticides due to their stability, inappropriate or illegal usage. In China, organophosphorus pesticides (OPs) account for 70% of the total amount of the pesticides used [1]. Maximum residue levels (MRLs) have been set up for 27 OPs in different kinds of food, and analytical methods have been developed for their analysis [2]. The China National standard method GB/T 5009.1452003 is a method for the determination of 16 organophosphorus and 4 carbamate pesticides by GC-NPDs [3]. In this application note, the Agilent 6820 GC equipped with an NPD is employed to determine 15 of the most common OPs of concern.
Experimental
The experiments were carried out on an Agilent 6820 GC with split/splitless inlet and NPD. De-activated liners for splitless injection (p/n 5183-4696) were used to improve the inertness of the system; the septa were Agilent green septa (p/n 5183-4759). Cerity Networked Data System (NDS) for Chemical QA/QC software was used for instrument control, data collection, and data processing. The sample was introduced manually with a 10-L syringe (p/n 5182-3428). All target compounds were dissolved in acetone. The experimental conditions are listed in Table 1. All flows were set using the Veriflow-500 digital flowmeter (p/n HVF-500-2).
Introduction
Synthetic organic pesticides are widely used in modern agriculture to protect crops and improve production. The green revolutions of many countries are obtained through the application of these compounds. There are increasing concerns over
Table 1. Software Inlet
Instrumental Parameters Cerity NDS for chemical QA/QC Split/Splitless inlet 250 C Splitless 1 L 0.75 min HP-5ms, 30 m 0.32 mm 0.25 m (p/n 19091S-413) He, head pressure: 12 psi, 2.5 mL/min at 60 C. 60 C for 1 min, to 200 C at 10 C/min, to 250 C at 5 C/min, 5 min hold. NPD at 325 C with white rubidium bead (p/n G1534-60570) H2: 3 mL/min; Air: 60 mL/min; makeup N2: 10 mL/min
Inlet temperature Injection mode Injection volume Purge time Column Carrier gas Oven temperature Detector Detector gases
by 1-L injections of standards at 1, 10, 50, 100, and 500 ng/mL concentrations were linear with R2 >0.999 for most compounds. The calibration curve for dichlorvos, a typical OP, is shown in Figure 2.
Results and Discussions

The separation of 15 OPs is illustrated in Figure 1. The compounds, except for chlorpyrifos and parathion, were well separated with the HP-5ms column. Peak identifications are listed in Table 2. Calibration curves constructed from data obtained
Table 2. Retention Times of Target Pesticides Peak number Compound Retention time 1 Methamidophos 8.26 2 Dichlorvos 8.63 3 Acephate 11.21 4 Monocrotophos 14.44 5 Phorate 14.60 6 Dimethoate 15.00 7 Parathion-methyl 17.03 8 Fenitrothion 17.75 9 Malathion 18.04 10 Fenthion 18.27 11 Parathion 18.34 12 Chlorpyrifos 18.34 13 Methidathion 19.97 14 Ethion 22.52 15 Triazophos 22.92
pA 600 11, 12
400 2 1 200 3 4 5
6 7 10 8 9 13 15 14
0 10 15 20 min
Figure 1.
Chromatogram of 15 OPs at 1-ppm using the NPD.
pA 800
600
Dichlorvos y = 1.3849x + 3.3834 R2 = 0.9999
400
200
0 0 100 200 300 Concentration, ppb 400 500 600
Figure 2.
Calibration curve for dichlorvos, a typical OP.
Approximate Instrumental Limits of Detection The chromatogram for 10-ppb pesticides using the NPD is shown in Figure 3. All compounds are easily quantitated. Acephate, with the lowest response factor, provided around a 30 ratio of signal to noise. In fact, most compounds at 1 ppb show good peaks except methamidophos, acephate, and monocrotophos. The limits of detection (LODs) are much lower than the maximum residue levels (MRLs) for the OPs.
11, 12 10 9
50
pA
52
14
7 6 5
13 15
48
2
46
44
3 1
42
40
7.5
10
12.5
15
17.5
20
22.5
min
Figure 3.
Chromatogram of 10-ppb OPs using the Agilent 6820 GC/NPD.
Sample Solvent Precautions In some sample extraction procedures, organic solvents, containing elements with high electronegativity, (such as methylene chloride), may be used. However, these solvents may cause baseline shifts, change the detector sensitivity, shorten lifetime of the rubidium bead, or even make the detector unusable. This is true for all types and vendors of NPD [4]. If the final solution for injection consists of methylene chloride or chloroform, it is best to change the solvent to acetone or hexane. visit our Web site at www.agilent.com/chem.
Conclusion
The Agilent 6820 GC with NPD can be used for the sensitive and selective determination of OPs. The NPD detector provides good linearity for most of these compounds in the 1500-ppb range.
References
1. Xingui Sun et al, Survey of organophosphorus pesticide residues in vegetables and fruits in Beijing, (2003) Chinese Journal of Food Hygiene, 15, No. 6, 536-538. 2. Jiming Ye et al, Introduction of maximum residues limits in China, (2000) Pesticide Science and Management. 3. China National Standard Method GB/T 5009.145-2003, Determination of organophosphorus and carbamate pesticides multiresidues in vegetable foods. 4. Kai Meng, Yeugeny Kaplun, Rich White, The New Features of the HP 6890 NitrogenPhosphorus Detector to Deal with Hostile Solvents, Agilent Technologies, publication (23) 5963-6808E. www.agilent.com/chem

For more information on our products and services,
Analysis of Carbamate Pesticides Using Agilent 6820 Gas Chromatograph/Nitrogen Phosphorus Detector Application
Authors
Weijun Yao, Chuanhong Tu Agilent Technologies Co., Ltd. (Shanghai) 412 YingLun Road Waigaoqiao Free Trade Zone Shanghai, 200131 P. R. C.
because of the excellent separation capability of the capillary column. In this note, the Agilent 6820 GC/NPD is used to determine seven common carbamates including metolcarb, isoprocarb, baycarb, propoxur, carbofuran, primicarb, and carbaryl.
Experimental
All experiments were performed on an Agilent 6820 GC with NPD, deactivated splitless inlet liner (p/n 5183-4696), and Agilent advanced green septa (p/n 5183-4759). Agilent Cerity Networked Data System (NDS) is used for instrument control and data acquisition. Instrumental parameters are listed in Table 1. The standard pesticides were dissolved in acetone solution. Samples were injected manually using a 10-L syringe (p/n 5182-3428).
Table 1. Instrumental Parameters Instrument Agilent 6820 GC Software Cerity NDS for chemical QA/QC Inlet Split/Splitless inlet Inlet temperature 250 C Injection mode Splitless Injection volume 1 L Purge time 0.75 min Column HP-5ms, 30 m 0.32 mm 0.25 m (p/n 19091S-413) Carrier gas Nitrogen, constant pressure of 5 psi, 1.0 mL/min (50 C) Oven temp 50 C (1 min) Ramp 1: 20 C/min to 100 C Ramp 2: 5 C/min to 150 C (5 min) Ramp 3: 10 C/min to 200 C (10 min) Detector NPD, 325 C, white bead (p/n G1534-60570) H2, 3 mL/min Air, 60 mL/min Detector gas Makeup, N2, 10 mL/min
Abstract
The Agilent 6820 gas chromatograph/nitrogen phosphorus detector (GC/NPD) was employed to determine carbamate pesticides, following the China National Standard Method GB/T 5009.104-2003. Seven common carbamates were fully separated using an HP-5ms column, 30 m 0.32 mm 0.25 m. Each compound showed good linearity in the 101000-ppb range, with a detection limit lower than 10 ppb, which is 101000 times lower than the maximum allowable residue level.
Introduction
The nitrogen phosphorus detector (NPD) is both sensitive and selective for organic compounds containing nitrogen and phosphorus, and is widely used in the trace analysis of organonitrogen or organophosphorus pesticides. The carbamates are types of organic synthetic pesticides, widely used to prevent plant diseases. The application of these pesticides may leave residues in agricultural products that may pose potential human health risks via the food chain. The Chinese government publishes maximum residue limits for the pesticides carbofuran, primicarb, and carbaryl in food. GB/T 5009.104-2003 (originally GB14877-1994) is a method used to determine six carbamate residues based on GC/NPD using a packed column [1]. In recent years, capillary-column methods have gradually replaced packed-column methods
Results and Discussions

All compounds were separated with a HP-5ms column. The resolution of baycarb and propoxur is 1.0, although their structures are very similar. The chromatogram of these pesticides is shown in Figure 1. The peak areas are linear with concentrations for all carbamates in the range of 101000 ppb. Typical calibration curves for carbofuran and primicarb are shown in Figure 2.
pA 90
80
70
Peak identification 5. Carbofuran 1. Metolcarb 6. Primicarb 2. Isoprocarb 7. Carbaryl 3. Baycarb 4. Propoxur 1
60
50
3 4 2
5 7
40
30
20 16 18 20 22 24 26 min
Figure 1.
Chromatogram of 1-ppm carbamates using NPD detector.
pA 140
120
100
Primicarb y = 0.1277x + 0.5557 R2 = 0.9992
80
60
40
Carbofuran y = 0.0706x + 0.7961 R2 = 0.9969
20
0 0 200 400 600 Concentration, ppb 800 1000 1200
Figure 2.
Calibration curves for primicarb and carbofuran.
The chromatogram for the 10-ppb carbamate standard is shown in Figure 3. The signal to noise ratios are between 612 for the 10-ppb carbamates. The limits of detection are 46 ppb; these are 10 to 1000 times lower than the maximum allowable residue limits stipulated in China National Standard (GB14928.2-94, GB14928.7-94, and GB14971-94), which is 50-ppb primicarb and 5000-ppb carbaryl in cereals [2, 3, 4]. The NPD can easily meet these requirements for the routine analysis of carbamate pesticide residues.
pA
6
30.6
30.5
30.4
4 1 3 2
30.3
30.2
30.1
30
16
18
20
22
24
26
min
Figure 3.
Chromatogram of 10-ppb carbamates standard with 1-L injection.
Conclusion
The Agilent 6820 GC with NPD detector can be used for the sensitive and selective measurement of carbamate pesticide residues in food. The detection limit is much lower than the maximum residue limits. Good linearity was obtained in the range of 101000 ppb.
References
1. GB/T 5009.104-2003, Determination of carbamate pesticide residues in plant-originated foods. 2. GB14928.2-94 Maximun residue limits of primicarb in food. 3 GB14928.7-94 Maximun residue limits of carbofuran in rice grains. 4. GB14971-94 Maximun residue limits of carbaryl in food.
3

Comprehensive Pesticide Screening by GC/MSD using Deconvolution Reporting Software Application
Food Safety
Author
Philip L. Wylie, Michael J. Szelewski, and Chin-Kai Meng Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA Christopher P. Sandy Agilent Technologies, Inc. Block A, CSC Eskdale Road Winnersh Triangle Estate Wokingham, Berk RG41 5DZ United Kingdom
pesticides in food and environmental samples to track their distribution in the environment and to ensure a safe food supply. Current analytical methods target only a subset of the possible compounds. Whether for food or environmental samples, analyses are often complicated by the presence of co-extracted natural products. Food or tissue extracts can be exceedingly complex matrices that require several stages of sample cleanup prior to analysis [3]. Even then, it can be difficult to detect trace levels of contaminants in the presence of the remaining matrix. For efficiency, multiresidue methods (MRMs) must be used to analyze for most pesticides. Traditionally, these methods have relied upon gas chromatography (GC) with a constellation of element-selective detectors to locate pesticides in the midst of a variable matrix [4, 5, 6]. GC with mass spectral detection (GC/MS) has been widely used for confirmation of hits. Liquid chromatography (LC) has been used for those compounds that are not amenable to GC [2]. Today, more and more pesticide laboratories are relying upon LC with mass spectral detection (LC/MS) and GC/MS as their primary analytical tools [7, 8]. Still, most MRMs are target compound methods that look for a small subset of the possible pesticides. Any compound not on the target list is likely to be missed by these MRMs. Using the techniques of retention time locking (RTL) [9, 10, 11] and spectral deconvolution [12], a method has been developed to screen for 567 pesticides and suspected endocrine disrupters in a single GC/MS analysis. Spectral deconvolution
Introduction
According to The Pesticide Manual, more than 700 pesticides are currently approved for use around the world [1]. About 600 more were used in the past, but are either banned or no longer marketed. In spite of their discontinuance, some of these still persist in the environment where they may bioaccumulate in the flora and fauna. Many pesticides or their degradation products can be found at trace levels in food and beverages; in soil, water, and air; in aquatic and terrestrial flora and fauna; and in human blood, adipose tissue, and breast milk. The World Health Organization has classified pesticides into five groups based upon their acute toxicity to humans [2]. The categories range from Acutely Hazardous to those that are Unlikely to Present Acute Hazard in Normal Use. Certain pesticides are classified as persistent organic pollutants (POPs), carcinogens, teratogens, or endocrine disrupters. It is now common to analyze for
helps to identify pesticides even when they are buried under co-eluting matrix compounds. RTL helps to eliminate false positives and gives greater confidence in the results. Users can easily add compounds to the method if they wish.
Samples Vegetable extracts were obtained from Dr. Mark Lee and Stephen Siegel at The California Department of Food and Agriculture (CDFA; Sacramento, CA USA) and from Dr. J.G.J. Mol at TNO Nutrition and Food Research (Zeist, The Netherlands). Seventeen data files from the GC/MS analysis of surface water samples were also contributed by CDFA and were processed in this laboratory using the Deconvolution Reporting Software (DRS). GC/MS data files (locked to the Agilent Pesticide Method) for 17 crop extracts were supplied by NRM Laboratories, Berkshire, UK.
Experimental
Table 1 lists the instrumentation, software, and analytical parameters used by Agilent for pesticide analysis. Depending upon the desired injection volume, a PTV inlet or split/splitless inlet can be used.
Table 1.
Instrumentation and Conditions of Analysis Agilent 6890N Agilent 7683 Agilent PTV operated in the solvent vent mode Agilent 30 m 0.25 mm 0.25 m HP-5MS (p/n 19091S-433) Helium in the constant pressure mode Chlorpyrifos-methyl locked to 16.596 min (nominal column head pressure = 17.1 psi) 70 C (2 min), 25 C/min to 150 C (0 min), 3 C /min to 200 C (0 min), 8 C /min to 280 C (1015 min) Temp program: 40 C (0.25 min), 1600 C/min to 250 C (2 min); Vent time: 0.2 min; Vent flow: 200 mL/min; Vent pressure: 0.0 psi; Purge flow: 60.0 mL/min; Purge time: 2.00 min 15 L (using a 50-L syringe) Agilent 5973 inert 50550 amu 230, 150, and 280 C, respectively 4.00 min Autotune voltage
Gas chromatograph Automatic sampler Inlet Column Carrier gas RTL Oven temperature program PTV inlet parameters Injection volume Mass Selective Detector (MSD) Scan range Source, quad, transfer line temperatures Solvent delay Multiplier voltage Software GC/MSD ChemStation Deconvolution Reporting Software (DRS) Library searching software Deconvolution software MS Libraries
Agilent p/n G1701DA (Version D01.00 sp1) Agilent p/n G1716AA NIST MS Search (version 2.0) (included with NIST '02 mass spectral library, Agilent p/n G1033A) Automated Mass Spectral Deconvolution and Identification Software (AMDIS) (included with NIST '02 mass spectral library, Agilent p/n G1033A) NIST '02 mass spectral library (Agilent p/n G1033A); Agilent RTL Pesticide Library (p/n G1049A)

RTL and RTL Databases RTL is a technique developed by Agilent that allows users to match analyte retention times (RTs) on any Agilent 6890 GC, in any laboratory in the world, so long as the same nominal GC method and capillary column are used [13]. Using RTL, Agilent has developed several retention-timelocked databases for GC and GC/MS that include the locked retention time, compound name, CAS number, molecular formula, molecular weight, and mass spectrum (GC/MS databases only) for each entry [14]. The Agilent RTL Pesticide Library contains this information for almost all GC-amenable pesticides, as well as several endocrine disrupters - 567 compounds in all. For use with the DRS discussed below, this library was converted into the NIST format [15]. Separate Automated Mass Spectral Deconvolution and Identification Software (AMDIS) libraries for the RTs and compound information were created from the original RTL Pesticide Library. Users can easily augment these libraries with newer pesticides or other compounds of interest [15].
Basics of Deconvolution In GC/MS, deconvolution is a mathematical technique that separates overlapping mass spectra into cleaned spectra of the individual components. Figure 1 is a simplified illustration of this process. Here, the total ion chromatogram (TIC) and apex spectrum are shown. As is often the case, the peak is composed of multiple overlapping components and the apex spectrum is actually a composite of these constituents. A mass spectral library search would give a poor match, at best, and certainly would not identify all of the individual components that make up the composite spectrum. The deconvolution process finds ions whose individual abundances rise and fall together within the spectrum. In this case, it first corrects for the spectral skew that is inherent in quadrupole mass spectra and determines a more accurate apex RT of each chromatographic peak. As illustrated in Figure 1, deconvolution produces clean spectra for each overlapping component. These individual spectra can be library searched with a high expectation for a good match. The AMDIS that is incorporated into the Agilent DRS is supplied by the National Institute of Science and Technology (NIST) [12].
TIC and spectrum
Deconvoluted peaks and spectra
TIC Component 1 Component 2 Component 3 Matrix
Deconvolution
Interference
Target
Figure 1.
An illustration of mass spectral deconvolution process. 3
DRS Agilent's DRS results from the combination of three different GC/MS software packages: 1) the Agilent GC/MS ChemStation, 2) the NIST Mass Spectral Search Program with the NIST '02 MS Library, and 3) the AMDIS software, also from NIST. Included in the DRS, are mass spectral and locked RT libraries for 567 pesticides and suspected endocrine disrupters. Three separate, but complimentary, data analysis steps are combined into the DRS. First, the GC/MS ChemStation software performs a normal quantitative analysis for target pesticides using a target ion and up to three qualifiers. An amount is reported for all calibrated compounds that are detected. For other compounds in the database, an estimate of their concentration can be reported based upon an average pesticide response factor
(RF) that is supplied with the DRS software. The DRS then sends the data file to AMDIS, which deconvolutes the spectra and searches the Agilent RTL Pesticide Library (in AMDIS format) using the deconvoluted full spectra. A filter can be set in AMDIS, which requires the analyte's RT to fall within a user-specified time window. Because RTL is used to reproduce the RTL database RTs with high precision, this window can be quite small typically 20 seconds or less. Finally, the deconvoluted spectra for all of the targets found by AMDIS are searched against the 147,000-compound NIST mass spectral library for confirmation; for this step, there is no RT requirement. Once the appropriate method is loaded, the DRS report can be generated with a single mouse click as shown in Figure 2. The software can run automatically after each analysis or at a later time on a single file or a batch of files.
Figure 2.
ChemStation pull down menu showing options for running the DRS on single or multiple files.
Pesticides in an Herbal Mix Figure 3 shows a TIC from the extract of an herbal mix. Figure 4 shows the MSD Deconvolution Report for this sample, which is produced in html format so it can easily be emailed or copied into a spreadsheet. This sample was chosen because herbs are among the most difficult vegetable products to analyze. Their extracts contain a large number of natural products that interfere with pesticide analysis.
450000 400000 350000 Abundance 300000 250000 200000 150000 100000 50000 5.00 10.00 15.00
TIC: MOL_4A.D
20.00
25.00 Time
30.00
35.00
40.00
45.00
Figure 3.
TIC of an herbal mix.
Figure 4.
MSD Deconvolution Report generated for the herbal mix extract shown in Figure 3.
The DRS report in Figure 4 lists the RT, CAS number, and compound name for each hit. Phenanthrene-d10, listed at the bottom of the report, is the internal standard (ISTD) used by the ChemStation to estimate the quantity of each compound that it found. Since an average pesticide response factor was used for all 567 target compounds, the amounts listed in column 4 are only estimates. Experience has shown that most estimates reported using an average pesticide response factor fall within a factor of 10 of their actual values. True quantitation requires calibration with pesticide standards in the normal way, but this is not practical for all of the pesticides in the database. A laboratory would normally generate calibration curves for their target set of pesticides and use the average RF for the remaining compounds in the database. In this way, when a new compound is detected, the lab can immediately get a rough estimate of its concentration and decide if it should be added to the calibration list. Column 5 in the report shows the match factor obtained through AMDIS deconvolution and RTL Pesticide Library searching using the deconvoluted full spectra. In this case, several more targets were identified by AMDIS than were found by the ChemStation software (for example, Prometon and p,p-DDE), which is typical for complex samples. When locked RTs are available, it is a significant advantage to set a RT requirement in the AMDIS software. In this case, hits that did not fall within 10 seconds of the database RT were eliminated. Column 6 shows the RT difference (in seconds) between the compound's library RT and its actual value in the chromatogram. Figure 4 shows that the software identified two phthalates (suspected endocrine disrupters) in addition to the pesticides. Phthalates are ubiquitous in the environment and are extremely difficult to remove from the background. In this case, no attempt was made to determine if the phthalates were actually extracted from the sample or were introduced in the laboratory. The last two columns in the DRS report show the results from searching all of the AMDIS hits against the NIST 147,000-compound mass spectral library. When the NIST library search finds a compound in the top 100 matches (a user-settable value) that agrees with the AMDIS results, its match factor is listed in column seven. The hit number is shown in the last column, with 1 being the best match (highest match factor) in the NIST database. Occasionally, the NIST library search does not find the AMDIS hit among the top
100 spectral matches. In this case, the next line in the report shows the best library match for that spectrum. This is evident for fluvalinate-tau-I (Figure 4), which eluted at 34.779 min. The next line shows the best NIST library match for that spectrum - fluvalinate. In this case, no compound with the same CAS number as fluvalinate-tau-I is contained in the NIST mass spectral library. In fact, fluvalinate-tau-I is the D isomer, while fluvalinate is the DL isomer mixture. Blind Comparison Between DRS and Traditional Data Review Many comparisons have shown that the DRS is much better than conventional methods at identifying target compounds in complex samples, such as food and environmental extracts. Two such studies are described here. In the first case, 17 unspiked crop samples were analyzed by NRM Laboratories in Berkshire, UK using Agilent's RT-locked pesticide method. The data files, but not their list of pesticide hits, were sent to Agilent for analysis using the new DRS. Table 2 shows a comparison of the results from the two laboratories. Using manual data review, NRM identified 28 pesticides in the 17 samples, four of which were below their lowest calibration level. Using the same data files, the DRS identified 33 pesticides. Agilent's automated method did not identify azoxystrobin in the spring onion sample because it is not included in the RTL pesticide library. While it can be found in the NIST library, it has a molecular ion at 403 amu and method used at NRM only scanned to 400 amu. The DRS method confirmed all four pesticides that were below the NRM calibration range and found five more (terbacil, pyrimethanil, methiocarb, pyridaben, and propamocarb) that were not included in their method. The agreement between the manual and automated methods was excellent. However, the DRS looks for many more pesticides and was able to find several that were missed by the manual method. In addition, manual data review took a chemist about 7 hours for the 17 samples while the DRS finished the task in 50 minutes of unattended computer time.
Table 2.
A Comparison of the Pesticides Found in 17 Unspiked Crop Samples Using Conventional Data Review and Agilent's DRS. Pesticides that Were Found by Only One Method Are Underlined Agilent DRS results* Propyzamide Chlorthal-dimethyl p,p'-DDE Terbacil Pirimicarb Chlorthal-dimethyl Propyzamide Pyrimethanil Pirimicarb Metalaxyl Iprodione Not in DRS library Methiocarb Iprodione Procymidone Pyridaben Propamocarb Procymidone Buprofezin Endosulfan sulfate Iprodione Chlorthal-dimethyl Iprodione Dimethoate Metalaxyl Procymidone Terbuconazole Omethoate Procymidone lamda-Cyhalothrin Chloropropham Pirimicarb Pirimicarb NRM Manual Analysis** Propyzamide Chlorthal-dimethyl p,p'-DDE Not found*** Pirimicarb Chlorthal-dimethyl Propyzamide Not found*** Pirimicarb Metalaxyl Iprodione Azoxystrobin Not found*** Iprodione Procymidone Not found*** Not found*** Procymidone Buprofezin Endosulfan sulfate Iprodione Chlorthal-dimethyl Iprodione Dimethoate Metalaxyl Procymidone Terbuconazole Omethoate Procymidone lamda-Cyhalothrin Chloropropham Pirimicarb Pirimicarb
Sample Coriander
Rosemary
Spring Onion
Chives Cherry Tomato Courgette Aubergine
Flat Leaf Parsley Lambs Lettuce Cos Lettuce
Fine Endive Red Potato Fine Endive

* **
Pesticides found by re-analyzing NRM datafiles using Agilent's DRS software. Pesticides found by NRM using target compound analysis and manual verification.
*** This compound was not in the NRM target compound list.

This compound is not included in the Agilent RTL Pesticide Library or the DRS software. Found by the Agilent ChemStation but not found by AMDIS or NIST library searching after deconvolution. After careful review of this hit, omethoate was judged not to be in the sample. Compound was detected but was below the calibration range.
Analysis of Surface Water Samples: In another study, the CDFA analyzed 17 surface water extracts for pesticides. TICs for two typical samples are shown in Figure 5. The CDFA used RTL and RTL database searching but without the benefit of spectral deconvolution. The same data files were then analyzed using the DRS for comparison. Table 3 shows the results from the CDFA manual analysis of the 17 samples compared to results using the DRS. The CDFA found 38 pesticide hits in the 17 samples, some of which were for the same pesticide in multiple samples. It took a skilled analyst about 8 hours to review the results, eliminate false positives, and verify all of the hits. The DRS found 37 of the compounds seen by the CDFA and identified one CDFA hit as a false positive. In addition, 34 more pesticides were
found for a total of 71 hits in the 17 samples. The process was fully automated and took about 20 minutes of unattended computer time to process all of the data files.
Table 3. A Comparison of Results from the Analysis of 17 Surface Water Samples by GC/MS. The CDFA Used RTL and RTL Database Searching, but No Deconvolution. Agilent's DRS Was Used to Analyze the Same Data Files CDFA Number of pesticide hits Number of false positives Time required for analysis 37 1 ~ 8 hours DRS Same 37 + 34 additional 0 20 minutes
5.00
10.00
15.00
20.00
25.00
30.00
35.00
40.00
5.00
10.00
15.00
20.00
25.00
30.00
35.00
40.00
Figure 5.
TICs of typical surface water extracts provided by the CDFA.
Conclusions
Agilent's new DRS solution for pesticide analysis offers laboratories a number of real benefits. Ease of use: This software solution is very simple to use and takes no more skill than is needed to operate the 6890N/5973 inert GC/MS system. There is no need for the user to learn about the intricacies of deconvolution or to master a new software package. Automation: The deconvolution report can be generated automatically after each run or a batch of samples can be processed all at once. Time savings: Data review is reduced from hours to minutes. Quality: It produces results with the fewest false positives and false negatives. Reproducibility: Results are not dependent upon the skill or experience of the operator. Accuracy: Comparisons such as those discussed in this application note show that the DRS finds pesticides with greater accuracy than manual methods of data analysis. It is particularly useful for relatively complex samples where co-eluting matrix components might obscure traces of target pesticides. Comprehensive: This method screens for almost all GC-amenable pesticides as well as several suspected endocrine disrupters in a single GC/MS run. With 567 compounds in the method, it is the most comprehensive pesticidescreening tool available. Users can add more compounds to the method as needed. Produces quantitative, semi-quantitative, and qualitative results: All calibrated compounds can be quantified. The concentrations of any other compounds can be estimated using an average pesticide response factor provided with the software. Use of the DRS is not limited to pesticide analysis. Other target compound mass spectral libraries can be converted into the AMDIS format and used with this software. For example, one could use existing libraries for forensic drugs, flavors and fragrances, organic pollutants, etc. Users can even generate their own libraries and use them with the DRS. While not required, it is a big advantage to have an RTL library with locked RTs for each entry, as this will give the fewest false positives.
Acknowledgements
The authors wish to thank Dr. Mark Lee and Stephen Siegel of the California Department of Food and Agriculture, Dr. J.G.J. Mol of TNO Research, The Netherlands, and the management and staff at NRM Laboratories, UK, for their contribution of samples and data.
References
1. C.D.S Tomlin, editor. The Pesticide Manual, 13th edition, British Crop Protection Council, Surry, UK (2003). 2. http://www.who.int/pcs/docs/ Classif_Pestic_2000-02.pdf 3. J. Cook, M.P. Beckett, B. Reliford, W. Hammock, and M. Engel (1999) J. AOAC Int., 82 (6), 14191435. 4. M.A. Luke, J.E. Froberg, and H.T. Masumoto, (1975) J. Assoc. Off. Anal. Chem. 58, 1020-1026. 5. M. Luke, J. Froberg, G. Doose, and H. Masumoto, (1981) J. Assoc. Off. Anal. Chem. 64, 1187-1195. 6. B. McMahon and N. Hardin (1994) Pesticide Analytical Manual, Vol. 1, 3rd Ed., U.S. Food and Drug Administration, Washington, DC. 7. J. Fillion, R. Hindle, M. Lacroux, and J. Selwyn (1995) J. AOAC Int. 78, 1252-1266. 8. J. Fillion, F. Sauv, and J. Selwyn (2000) J AOAC Int., 83, 698-712. 9. P. Wylie and B. Quimby, A Method Used to Screen for 567 Pesticides and Suspected Endocrine Disrupters, Agilent Technologies, publication 5967-5860E www.agilent.com/chem. 10.H. Prest, P. Wylie, K. Weiner, and D. Agnew, Efficient Screening for Pesticides and Endocrine Disrupters Using the 6890/5973 GC/MSD System, Agilent Technologies, publication 5968-4884E www.agilent.com/chem. 11.K. Weiner and H. Prest, Retention Time Locking: Creating Custom Retention Time Locked Screener Libraries, Agilent Technologies, publication 5968-8657E www.agilent.com/chem. 12.National Institute of Standards and Technology, AMDIS Literature and Downloads website: http://www.amdis.net/What_is_AMDIS/ AMDIS_Literature_and_Downloads/ amdis_literature_and_downloads.html.
13.V. Giarocco, B. Quimby and M. Klee, Retention Time Locking: Concepts and Applications, Agilent Technologies, publication 5966-2469E www.agilent.com/chem. 14.http://www.chem.agilent.com/cag/servsup/ usersoft/main.html#RTL. 15.M. Szelewski and C.K. Meng, Building and Editing RTL Screener Databases and Libraries, Agilent Technologies, publication 5989-0916EN www.agilent.com/chem.

Improving the Analysis of Organotin Compounds Using Retention Time Locked Methods and Retention Time Databases Application
Environmental
Authors
Frank David Research Institute for Chromatography Pres. Kennedypark 20, B-8500 Kortrijk, Belgium Pat Sandra University of Gent Krijgslaan 281 S4, B-9000 Gent, Belgium Philip L. Wylie Agilent Technologies 2850 Centerville Road Wilmington, DE 19808-1610 USA
Introduction
For many years, organometal speciation has been an important topic in environmental analysis, primarily due to increasing awareness of the toxicological effects of many organometal compounds. Within the class of organometalics, organotin compounds are probably the most widely spread in the environment due to their use as additives in polymers and in antifouling paints. Organotin compounds degrade in the environment into more polar metabolites [1]. Tributyltin, one of the most frequently used organotin additives (as tributyltinchloride or tributyltinoxide), for instance, degrades into dibutyltin and monobutyltin species. Consequently, a large diversity of organotin compounds can be detected in various environmental samples [2]. More recently, organotin contamination of diapers and printed T-shirts was reported and numerous analyses were performed on different consumer products, including all types of absorbent hygiene products. Different methods were used for the extraction and analysis of organotin compounds in environmental, food, and consumer product matrices. Since the organotin compounds with less than four alkyl groups are very polar, they cannot be analyzed directly by GC and must be derivatized into tetraalkyltin compounds prior to analysis. Initially, most methods were based on extraction with
Abstract
The analysis of organotin compounds is becoming increasingly important in both environmental analysis and in food and consumer product analysis. This application note describes a retention time locked (RTL) gas chromatography/mass spectrometry (GC/MS) method for the analysis of derivatized organotin compounds. Three retention time locked libraries are made available, corresponding to three different derivatization methods. The retention time databases allow easy peak location and identification of the target solutes based on mass spectra and retention times.
tropolone (a complexing agent) and n-hexane, followed by Grignard derivatization and determination with GC-flame photometric detection (FPD) [39]. Recently, in-situ ethylation with sodium tetraethylborate (NaBEt4) [1013] has largely replaced Grignard derivatization. At the same time, mass selective detectors (MSD) and atomic emission detectors (AED) have replaced the FPD as the preferred GC detector for organotin compounds [11,13]. A few years ago, solid phase micro extraction (SPME) in combination with capillary gas chromatography-inductively coupled plasma mass spectrometry (CGC-ICP-MS) was used for the determination of volatile and semi-volatile organometal compounds, resulting in excellent sensitivity and selectivity [14,15]. SPME was performed in the headspace or directly in the aqueous sample using a 100 mm polydimethylsiloxane (PDMS) coated fiber. Using NaBEt4, organotin compounds could be derivatized in-situ and simultaneously extracted into the PDMS phase. More recently, stir bar sorptive extraction (SBSE) using a magnetic stir bar coated with a 0.51 mm PDMS layer was developed [16]. After extraction, the solutes were thermally desorbed online to GC/MS, GC-AED or GC-ICP-MS. SBSE in combination with CGC-ICP-MS was applied for the determination of organotins in environmental samples after in-situ derivatization with NaBEt4, resulting in unsurpassed sensitivity with detection limits reaching the ppq (pg/L) level [17]. For standard applications such as the determination of organotin compounds in sediments, or soils, and in extracts or leachates of consumer products, these extremely high sensitivities are not required. For these applications, sufficient sensitivity is obtained using mass spectrometric detection. In comparison to AED or ICP-MS, where specific tin-chromatograms are obtained, the chromatograms obtained by mass spectroscopy are far more complex, even when using the selected ion monitoring (SIM) mode. Several ions per solute need to be monitored, and the derivatized sample extracts often contain many co-extracted solutes
or by-products of the derivatization reaction. Therefore, data interpretation is more demanding requiring the use of extracted ion chromatograms, retention time matching, and calculation of the relative abundances of target and qualifier ions. In this respect, the use of retention time locked methods offers several advantages. If a selected ion method is used, the switching times between groups of monitored ions are fixed and do not need to be adjusted after column maintenance or column change, since the retention times of all solutes can be relocked. Moreover, quantification databases do not need to be updated for variations in retention times. Finally, a retention time locked database can be used, allowing easy peak allocation. Solute detection and confirmation are far more reliable using the results screener option [18,19], which combines the power of spectral matching with locked retention time matching. In this application note, a GC/MS method is described for the analysis of organotin compounds in environmental, food, or consumer product extracts. Since derivatization by Grignard reaction and derivatization using NaBEt4 are both easy and convenient, three types of derivatives are considered: methyl-derivatives using methylmagnesium bromide, pentyl- derivatives using pentylmagnesium bromide (both Grignard reagents), and ethylderivatives using NaBEt4. The most important organotin compounds are listed in Table 1 together with typical ions for the mass spectra of all three derivatives. Tin has several isotopes and the mass spectra are characterized by typical isotope clusters. The relative abundances of the tin isotopes are Sn-116 (14.24%), Sn-117 (7.57%), Sn-118 (24.01%), Sn-119 (8.59%), Sn-120 (32.97%), Sn-122 (4.71%), and Sn-124 (5.98%). For the organotin compounds listed in Table 1, mass spectral libraries and retention-time-locked screener libraries were created for all three types of derivatives. After selecting the appropriate derivitization method, a library and screener database can be selected, allowing fast data interpretation. Sample extraction and clean-up are beyond the scope of this application note.
Experimental
Samples The organotin compounds listed in Table 1 were purchased from Dr Ehrenstorfer, Augsburg, Germany (http://www.analytical-standards.com). For analysis, the standards were dissolved in methanol at a 1000 ppm (1mg/mL) concentration. These solutions were further diluted, depending on the derivatization method used. For creation of the databases, approximately 10 g of compound was derivatized, resulting in a final concentration of 10 ppm. Derivatization method 1: The sample extract is concentrated to 1 mL in an apolar solvent (typically hexane) in a reaction tube. To this solution, 0.5 mL methylmagnesiumbromide Grignard reagent (1.4 M in 75/25 toluene/THF, SigmaAldrich cat no 28,223-5) is added. The solution is vortexed for 10 s and allowed to stand at room temperature for 15 min. This procedure should be performed in a fume hood, since toxic vapors evolving from the reaction and the solvents are
flammable. The reaction is stopped and the excess reagent is removed by adding 2 mL of a saturated ammoniumchloride solution in water or 2 mL 0.25 mol/L aqueous sulphuric acid. The mixture is vortexed for 10 s and the two phases are allowed to separate. The clear upper layer (apolar hexane phase) is transferred to an autosampler vial for analysis. The resulting organotin compounds are the methyl-derivatives. Derivatization method 2: The sample extract is concentrated to 1 mL in an apolar solvent (typically hexane) in a reaction tube. To this solution, 0.5 mL pentylmagnesiumbromide Grignard reagent (2 M in diethylether, Sigma-Aldrich cat no 29,099-8) is added. The remaining steps in this procedure are identical to those used in derivitization method 1. The resulting organotin compounds are the pentyl-derivatives. Derivatization method 3: The sample extract is concentrated to 1 mL in a polar solvent (typically ethanol) in a reaction tube. To this solution, 1 mL acetate buffer (82 g/L sodium acetate in water, adjusted to pH 4.5 with acetic acid) and 50 L
Table 1:
Organotin Compounds and Characteristic Ions for the Three Derivatization Products Abbreviation Derivatization 1 Methylmagnesium bromide Methyl193, 191, 165, 163 207, 205, 179, 177 179, 177, 221, 219 249, 247, 207, 205 165, 163, 151, 149 151, 149, 207, 205 193, 191, 249, 247 291, 289, 179, 177 227, 225, 223, 197 289, 287, 285, 197 351, 349, 347, 197 351, 349, 347, 197 301, 299, 219, 217 165, 163, 263, 261 263, 261, 151, 149 Derivatization 2 Pentylmagnesium bromide Pentyl179, 177, 249, 247 207, 205, 179, 177 277, 275, 165, 163 249, 247, 207, 205 319, 317, 193, 191 319, 317, 179, 177 305, 303, 179, 177 291, 289, 179, 177 339, 337, 197, 195 345, 343, 197, 195 351, 349, 347, 197 351, 349, 347, 197 357, 355, 205, 203 375, 373, 193, 191 417, 415, 375, 373 Derivatization 1 Sodium tetraethylborate Ethyl207, 205, 179, 177 207, 205, 179, 177 235, 2331, 249, 247 249, 247, 207, 205 235, 233, 179, 177 263, 261, 207, 205 291, 289, 207, 205 291, 289, 179, 177 255, 253, 197, 195 303, 301, 197, 195 351, 349, 347, 197 351, 349, 347, 197 315, 313, 233, 231 291, 289, 179, 177 375, 373, 263, 261 3
Organotin solute Reagent Derivatives Triethyltin Tetraethyltin Tripropyltin Tetrapropyltin Monobutyltin Dibutyltin Tributyltin Tetrabutyltin Monophenyltin Diphenyltin Triphenyltin Tetraphenyltin Tricyclohexyltin (Cyhexatin) Monooctyltin Dioctyltin
TET TeET TPT TePT MBT DBT TBT TeBT MPhT DPhT TPhT TePhT TCT MOT DOT
derivatization reagent are added. The derivatization reagent is prepared by dissolving 2 g NaBEt4 (Sigma-Aldrich cat no 48,148-3) in 10 mL ethanol. This solution should be freshly prepared. The sample is shaken and allowed to react for 30 min. After addition of 5 mL water, the derivatized compounds are extracted in 1 mL hexane. The mixture is vortexed for 10 s and the two phases are allowed to separate. The clear upper layer (apolar hexane phase) is transferred to an autosampler vial for analysis. The resulting organotin compounds are the ethyl-derivatives. These derivatization methods can be adapted to the type of sample analyzed. For example, derivatization method 3 is often applied to aqueous samples directly, combining in-situ derivatization and simultaneous extraction. This method is also used for sediment samples. Typically 1 g sample (dry weight) is extracted with 10 mL acetate buffer, 7 mL methanol and 10 mL hexane. Four mL of a 5% NaBEt4 solution is added while stirring. The derivatized organotin compounds are simultaneously extracted into the hexane layer.
Analytical Conditions
All analyses were performed on an Agilent 68905973N GC-MSD system. Automated splitless injection was performed using an Agilent 7683 automatic liquid sampler. The instrumental configuration and analytical conditions are summarized in Table 2. The retention time of tetrabutyltin (used as the locking standard) was locked at 16.000 min. To duplicate this method, the initial column head pressure can be set to the pressures indicated in Table 2 (nominal pressure). Then the retention time locking (RTL) calibration runs can be performed automatically (at 20%, 10%, +10% and +20% of the nominal pressure) [18]. The retention time versus head pressure curve is then calculated and stored in the method. Agilents RTL software uses this curve to set the column head pressure so that retention time of the locking standard (tetrabutyltin) is 16.000 min.
Table 2.
Instrumentation and Conditions of Analysis
Instrumentation Chromatographic system Inlet Detector Automatic sampler Liner Column Experimental conditions Inlet temperature Injection volume Injection mode Carrier gas Head pressure Oven temperature Transfer line temperature Detector 280 C 1 L Splitless, purge time: 1 min, purge flow: 50 mL/min. Helium Tetrabutyltin is retention time locked at 16.000 min (pressure around 45 kPa at 50 C, 34 cm/s at 50 C) 50 C, 1 min, 10 C/min to 300 C, 4 min. 300 C Scan (40550 amu), threshold 100, MS quad 150 C, MS source 230 C. Solvent delay: 4 min SIM mode: 50 ms dwell time per ion, ions listed in Table 3 Agilent 6890 GC Split/Splitless Agilent 5973 N MSD Agilent 7683 Splitless liner (part number 5062-3587) 30 m 0.25 mm id 0.25 m HP-5MS (Agilent part number 19091S-433)
derivatization. With this derivatization, the elution sequence of the butyltin compounds is MBT A typical chromatogram, for an organotin standard (10 C atoms) < DBT (12 C atoms) < TBT (14 C atoms) < TeBT (16 C atoms). The spectrum obtained for mixture, derivatized using method 3 (ethylderivatives with NaBEt4), is shown in Figure 1. The tributyltin (as tributylethyltin) is shown in Figure 2. The typical ion clusters, resulting from the different compounds elute according to their boiling point, tin isotopes, are clearly detected. and the elution sequence can be predicted by calculating the total number of carbon atoms after
3000000
TCT
2800000 2600000 2400000 2200000 2000000 1800000 Abundance 1600000 1400000 1200000 1000000 800000 600000 400000 200000 0 10.00 11.00 12.00 13.00 14.00 15.00 16.00 17.00 18.00 19.00 20.00 21.00 22.00 23.00 Time
TBT
TeBT
DOT MBT TPT TePT DBT MOT
Figure 1.
GC/MS chromatogram for the analysis of an organotin standard mixture after derivatization with NaBEt4 (ethyl-derivatives).
207
130000 120000 110000 100000 90000 Abundance 80000 70000 60000 50000 40000 30000 20000 10000 0 40 60 41 57 81 80 103 100 120 140 160 180 200 220 m/z 240 260 280 313 300 334 320 340 360 389 406 380 400 121 235 151 177 263 291
Figure 2.
Mass spectrum of tributyltin after derivatization with NaBEt4 (ethyl-derivative). 5
The analysis of a coastal sediment sample is shown in Figure 3. In this case, derivatization method 2 (Grignard reaction with pentylmagnesium bromide) was applied and a complex chromatogram was obtained. Using the extracted ion chromatogram at m/e 179 the butyltin compounds were easily detected (Figure 4). Tetrabutyltin, eluting at 16.000 min, was added as internal standard. In this case, pentyl- derivatives are analyzed. Therefore the elution order is reversed since the derivatization adds a C5-group for every free valency. The elution sequence is now TeBT (16 C atoms = unchanged) < TBT (17 C atoms) <DBT (18 C atoms) < MBT (19 C atoms). The mass spectrum obtained for the pentyl derivative of tributyltin is shown in Figure 5.
9500000 9000000 8500000 8000000 7500000 7000000 6500000 6000000 Abundance 5500000 5000000 4500000 4000000 3500000 3000000 2500000 2000000 1500000 1000000 500000 0 15.50 16.00 16.50 17.00 17.50 Time 18.00 18.50 19.00 19.50
Figure 3.
GC/MS chromatogram for the analysis of a coastal sediment sample after derivatization with pentylmagnesium bromide (pentyl-derivatives).
16.01
650000 600000 550000 500000 450000 400000 Abundance 350000 300000 250000 200000 150000 100000
TeBT
17.00
TBT
DBT
17.94
MBT
50000 0 15.50 16.00 16.50 17.00 17.50 Time 18.00 18.50 19.00 19.50
17.72
19.00
Figure 4.
Extracted ion chromatogram showing the presence of butyltin compounds in the coastal sediment sample extract(shown in Figure 3) after derivatization with pentylmagnesium bromide (pentyl-derivatives).
179 280000 260000 240000 220000 200000 235 180000 160000 Abundance 140000 120000 100000 80000 60000 40000 20000 0 305
121
41 69 147 99 207
254 283 327
355
405
452
477
50
100
150
200
250
m/z
300
350
400
450
Figure 5.
Mass spectrum of tributyltin after derivatization with pentylmagnesium bromide (pentyl-derivative).
Using the Agilent results screener and the appropriate screener library, the files can be screened for the presence of all compounds listed in the screener database. Figure 6 shows a typical result, with the identification of pentyltricyclohexyltin at 24.908 min. The target ions for this compound are extracted and overlaid in the top window. For easy comparison, the apex mass spectrum is displayed. Though not shown in Figure 6, the Agilent RTL Screener Software can display the library and apex spectra together for easy spectral comparison. In addition, the relative abundances of the
target ion and qualifiers are measured and compared to the library data. What distinguishes the Agilent screener methods from conventional GC/MS techniques is the comparison of a peak's locked retention time to values stored with the RTL database. In this case, the locked retention time of pentyltributyltin is within 0.002 min (0.12 s) of the database value. The Agilent results screener compares locked retention times and spectral information for fast peak allocation and more reliable identification.
320000 280000 240000 Abundance 200000 160000 120000 80000 40000 0 24.00 24.20 24.40 24.60 24.80 25.00 Time 25.20 25.40 25.60 25.80
24.91
Ion 205 Ion 203 Ion 357 Ion 287
205 200000 180000 160000 287 140000 Abundance 120000 100000 80000 121 60000 40000 20000 147 0 50 100 150 200 250 177 231 253 309 300 m/z 331 350 378 405 429 451 479 503 529 400 450 500 55 81 357
(13) Pentyltricyclohexyltin 24.908 min (-0.002) response 4581312 Ion 205.00 203.00 357.00 287.00 Exp % 100 82.90 77.20 66.70 Act % 100 83.87 80.36 68.98
Figure 6.
Screener result for the detection of tricyclohexyl tin in a sample extract after derivatization with pentylmagnesium bromide (pentyl-derivative).
For added specificity and sensitivity, SIM methods were developed for all three alkyl derivatives of the target tin compounds. Table 3 lists the SIM ions and target compounds in each group. Note that the start time for each SIM group is also listed. Normally, this timing could not be published with confidence, because of retention time differences
between instruments. However, RTL allows analysts to duplicate locked methods directly and reproduce all analyte retention times within a few thousandths of a minute. Thus, it is possible to apply this method directly, including the SIM group timing, after locking tetrabutyltin to the method-specified retention time of 16.000 minutes.
Table 3.
SIM Groups and Timing for Methyl, Pentyl, and Ethyl Derivatives of the Target Tin Compounds Listed in Table 1. The GC/MS Method Shown in Table 2 was used with the Retention Time of Tetrabutyltin Locked to 16.000 Minutes. Start time (min)
Solutes
Ions
Derivatization 1 1 2 3 4 5 6 7 8 Derivatization 2 1 2 3 4 5 6 7 8 Derivatization 3 1 2 3 4 5 6 7 8 5.00 8.50 11.40 13.00 14.80 17.00 22.00 25.00 TeET (=TET) MBT, TPT TePT, DBT MPhT, TBT MOT, TeBT DPhT, DOT TPhT, TCT TePhT 207, 205, 179, 177 235, 233, 179, 177, 249, 247 249, 247, 207, 205, 263, 261, 207, 205 255, 253, 197, 195, 291, 289 291, 289, 179, 177, 291, 289 303, 301, 197, 195, 375, 373, 263, 261 351, 349, 347, 197, 315, 313, 233, 231 351, 349, 347, 197 5.00 9.00 13.50 16.50 20.00 22.80 24.00 26.00 TeET TET, TePT TPT, TeBT TBT, DBT, MBT MPhT, MOT, DPhT DPhT, TCT DOT, TPhT TePhT 207, 205, 179, 177 179, 177, 249, 247, 207, 205 277, 275, 165, 163, 291, 289, 179, 177 305, 303, 179, 177, 319, 317, 193, 191 339, 337, 197, 195, 375, 373, 193, 191 345, 343, 275, 273, 357, 355, 205, 203 417, 415, 375, 373, 351, 349, 347, 197 351, 349, 347, 197 5.00 6.50 8.00 10.50 12.50 16.40 21.00 25.00 TET, MBT TeET MPhT, DBT, TPT MOT, TePT TBT, TeBT DPhT, DOT TCT, TPhT TePhT 193, 191, 165, 163, 151, 149 207, 205, 179, 177 227, 225, 223, 151, 149, 207, 205, 179, 177, 221, 219 165, 163, 263, 261, 249, 247, 207, 205 193, 191, 249, 247, 291, 289, 179, 177 289, 287, 285, 197, 263, 261, 151, 149 301, 299, 219, 217, 351, 349, 347, 197 351, 349, 347, 197
Conclusion
A GC/MS method is presented for the analysis of organotin compounds in extracts of environmental, food, or consumer product samples. Three different derivatization methods are described. For each derivatization method, mass spectral and retention time-locked screener databases were created. By itself, RTL is a valuable tool for maintaining GC and GC/MS methods and for comparing results among different laboratories. It also allows analysts to duplicate methods exactly, including SIM group timing and peak timing in quantitative methods. When combining RTL with locked mass spectral database searching, peak identifications become far more convenient and reliable. While many compounds can have similar spectra, they usually do not have similar spectra and identical retention times. Agilents ability to reproduce retention times for a given method on any 6890 GC makes it possible to differentiate closely-related compounds and to screen for large numbers of analytes in a matter of seconds. This rapid GC/MS screening technique is now available for a wide variety of important tin compounds. The three organotin databases are available for free from the Life Sciences and Chemical Analysis portion of the Agilent web site (www.agilent.com).
8. M. Nagase and K. Hasabe, (1993) Anal. Sci., 9, 517. 9. H. H. Van de Broek, G. B. M. Hermes, and C. E. Goewie, (1988) Analyst, 113, 1237. 10. N. Flsvik and E. M. Brevik, (1999) J. High Resolut. Chromatogr., 22, 177. 11. M. Ceulemans, C. Witte, R. Lobinski, and F. C. Adams, (1994) Appl. Organomet. Chem., 8, 451. 12. Y. Morcillo and C. Porte, (1998) Trends Anal. Chem., 17, 109. 13. J. S. Lobinska, M. Ceulemans, R. Lobinski, and F. C. Adams, (1993) Anal. Chim. Acta, 278, 99. 14. L. Moens, T. De Smaele, R. Dams, P. van den Broek, and P. Sandra, (1997) Anal. Chem., 69, 1604. 15. J. Vercauteren, A. De Meester, T. De Smaele, F. Vanhaecke, L. Moens, R. Dams, and P. Sandra, (2000) J. Anal. At. Spectrom., 15, 651. 16. E. Baltussen, P. Sandra, F. David, and C. Cramers, (1999) J. Microcolumn Sep., 11, 737. 17. J. Vercauteren, C. Prz, C. Devos, P. Sandra, F. Vanhaecke and L. Moens, (2001) Anal. Chem., 73, 1509. 18. V. Giarrocco, B. Quimby and M. Klee, Retention Time Locking: Concepts and Applications, Agilent Technologies, publication 5966-2469E www.agilent.com/chem 19. K.R. Weiner and H.F. Prest, Retention Time Locking: Creating Custom Retention Time Locked Screener Libraries, Agilent Technologies, publication 5968-8657E www.agilent.com/chem
References
1. P. J. Graig, (1986) Organometallics in the environment; Principles and Reactions, 133. 2. Harino H., M. Fukushima, Y. Yamamoto, S. Kawai. N. Miyazaki, (1998) Archives of Toxicology, 35, 558. 3. A. M. Caricchia, S. Chiavarini, C. Cremisini, R. Morabito, and R. Scerbo, (1994) Anal. Chim. Acta, 286, 329. 4. K. Fent and J. Hunn, (1991) J. Environ. Sci. Technol., 25, 956. 5. J. L. Gomez-Ariza, E. Morales, and M. Ruiz-Benitez, (1992) Analyst, 117, 641. 6. H. Harino, M. Fukushima, and M. Tanaka, (1992) Anal. Chim. Acta, 264, 91. 7. M. D. Mller, (1987) Anal. Chem., 59, 617.

For more information on our products and services, visit our web site at www.agilent.com/chem.
Validated Method for the Determination of Phenyl Urea and Triazine Herbicides in Potable and Groundwater by LC/MS Using Selective Ion Monitoring Application
Environmental
Authors
Neil Cullum Anglian Water Services Laboratories Huntingdon UK Paul Stephens Agilent Technologies, Ltd. Bracknell UK
(a triazinone herbicide). Phenyl urea herbicides, such as isoproturon, are widely used for weed control in crops. The triazine herbicides, such as atrazine, are extensively used general weed control agents. Both phenyl urea and triazine herbicides are found in environmental samples and are detected in drinking water. The Prescribed Concentration or Value (PCV) for an individual pesticide in drinking water in the UK, as defined by the Water Supply (Water Quality) Regulations, is set at 0.1 g/L. Ideally, the method of analysis should be capable of detecting 10% to 20% of the PCV, that is 0.01 to 0.02 g/L. Standard methods for the determination of these herbicides in water matrices involve either liquidliquid extraction or, more recently, solid phase extraction (SPE) followed by high performance liquid chromatography (HPLC) with ultra violet (UV)/diode array detection for both classes of herbicide, but more typically gas chromatograph/ nitrogen phosphorus detector (GC/NPD) or gas chromatography/mass spectrometry (GC/MS) detection for the triazines. The phenyl urea herbicides are not suited to GC analysis as they are thermally labile. This application note details the extraction method (SPE) and analysis using liquid chromatography/mass spectrometry (LC/MS) as a combined suite, using a relatively small volume of sample.
Abstract
This application note describes a validated liquid chromatography/mass spectrometry method for phenyl urea and triazine herbicides in potable and groundwater using an atmospheric pressure electrospray ionisation source in both positive and negative ion mode. The performance requirements set by the Drinking Water Inspectorate for standard deviation, bias, and total error are all met for each of the 16 herbicides studied.
Introduction
This application note describes the analysis of a single analytical suite composed of five phenyl urea herbicides including eight triazine herbicides, carbetamide (a carbamate herbicide), chloridazon (a pyridazinone herbicide), and metamitron
Experimental
All analyses were performed using the Agilent 1100 series LC/MS quadrupole coupled to an Agilent 1100 series LC system consisting of a binary pump, autosampler, thermostated column compartment, and vacuum degasser. The system also had a diode array detector in-line before the mass spectrometer, as a trouble shooting tool. The quadrupole mass spectrometer was operated with an atmospheric pressure electrospray ionisation (API-ES) source in positive and negative ion modes.
MS Conditions
Ionisation mode: Drying gas flow: Nebulizer pressure: Drying gas temperature: Vcap voltage: Positive/Negative API-ES 13.0 L/min 40 psig 350 C 3000 V (positive), 2500 V (negative)
LC Conditions
Column: 40 C Flow rate; Injection volume: Mobile phase: 0.5 mL/min 25 L A = 0.001% formic acid in water B = Methanol Initial 2.0 min 15.0 min 15.1 min 22.0 min A 90% 90% 30% 90% 90% B 10% 10% 70% 10% 10% Zorbax Eclipse XDB-C8 50 mm long x 2.1 mm id, 3.5 m particles,
The selected ion monitoring (SIM) ions and fragmentor voltages listed in the SIM table parameters of Table 1 were all optimised using Flow Injection Analysis (FIA). Ten mg/L standard solutions of each herbicide were injected using scan mode 150 400 amu and the fragmentor voltage was ramped from 70 to 150 V in steps of 5 V.
Sample Preparation
SPE was performed using automated equipment and Baker SDB1 200 mg, 3 mL cartridges. The cartridges were conditioned with 5 mL of ethyl acetate, followed by 5 mL methanol, followed by 2 mL HPLC grade water. Fifty mL of sample was diluted to 200 mL using de-ionised water. One hundred mL of the diluted solution was pumped through the conditioned cartridge at 10 mL/min. Cartridges were dried for 25 minutes by forcing air through, followed by elution using ethyl acetate 1 1.5 mL and 1 1 mL. The final extracts were evaporated to dryness using a heated block set at 45 C
Gradient program:
Table 1.
SIM Table parameters, positive ion mode Compound Metamitron Chloridazon Monuron Simazine Carbetamide Cyanazine Isoproturon Chlortoluron/ Isoproturon-d6 Atrazine Propazine Terbuthylazine Trietazine Prometryn Terbutryn Time 2.50 7.50 Group 1 2 SIM ions, Quantitation, Qualification (q) 203.0, 204.0q 222.0, 224.0q 199.0, 201.0q 202.0, 204.0q 237.0, 238.0q 241.0, 243.0q 207.0, 208.0q 213.0, 215.0q 216.0, 218.0q 12.40 4 230.0, 232q Fragmentor voltage 70 100 115 70 95 130 140 130 135 130
Compound # 1 2 3 5 4 6 9 7 8 12 13 15 14 16
10.25
242.0, 243.0q
130
SIM Table parameters, negative ion mode 10 11 q Qualifier ion Diuron Linuron 2.50 1 231.0, 233.0q 247.0, 249.0q 130 115
and a gentle stream of air. The residue was redissolved in 250 L 90:10 HPLC grade water:methanol. The final make-up solution contains an internal standard at a concentration of 0.1 g/L. The internal standard used is isoproturon-d6. Isoproturon-d6 is also present in the calibration standards at the above concentration for all levels of the calibration.
450000 400000 350000 300000
MSD1 TIC, MS File (T:\AGILEN~2\922259~1\DATA\PT1\SAMPLE03.D) API-ES, Pos, SIM, Frag: 70 (TT)
15 + 16 14
9 250000 200000 150000 100000 50000 0 4 6 8 10 12 14 16 18 min 1 3 2 4 5+6 7 12.902 10 18000 16000 14000 12000 10000 8000 11 6000 4000 2000 0 4 6 8 10 12 14 16 18 min
MSD2 TIC, MS File (T:\AGILEN~2\922259~1\DATA\SAMPLE03.D) API-ES, Neg, SIM, Frag: 130 (TT)
8 12 13
Figure 1. Chromatogram for the low level standard in positive ion mode.
Figure 2. Chromatogram for the low level standard in negative ion mode.
IS 70000 60000 50000 40000 30000 20000 10000 0 4 6 8 10 12 14 16 18 min 7

MSD1 213, EIC=212.7:213.7 (T:\AGILEN~2\922259~1\DATA\PT1\SAMPLE05.D) API-ES, Pos, SIM, Frag: 70 (TT)
Figure 3. Extracted ion chromatogram from a spiked tap water sample containing chlortoluron at 0.10 g/L. The isoproturon-d6 internal standard (IS) is also shown. 3
Results
A typical chromatogram for the low level standard (each compound at 0.1 g/L) appears in Figures 1 and 2 in both positive and negative ion modes respectively. Figure 3 shows an extracted ion chromatogram from a spiked tap water sample containing chlortoluron (0.10 g/L) and the isoproturon-d6 internal standard. Validation of the method was done on 11 batches of samples. Standards were spiked at three levels: 0.01, 0.10, and 0.40 g/L. The borehole raw water was spiked at two levels: 0.01 and 0.10 g/L. The potable tap water (which was from a surface water source) was also spiked at two levels: 0.01 and 0.10 g/L . All samples were analysed in duplicate in each batch in a random order. See Table 2. The limit of detection (LOD) for each herbicide was calculated from the within-batch standard deviation of the standard spiked at 0.01 g/L. Recovery for both the groundwater and potable water samples was calculated from the 0.10 g/L spike after the subtraction of 0.01 g/L, hence the recovery value is based on 0.09 g/L. Calibration curves were produced using three calibration levels at 0.1, 0.3, and 0.5 g/L. The calibration curves were all produced using a quadratic fit and forced through the origin and referenced against an internal standard. Typical correlation values are 0.9997 or better for all the herbicides in the suite.
Table 2. Results Groundwater sample Compound Atrazine Carbetamide Chloridazon Chlortoluron Cyanazine Diuron Isoproturon Linuron Metamitron Monuron Prometryn Propazine Simazine Terbuthylazine Terbutryn Trietazine Recovery % 83.4 92.0 94.8 89.5 95.1 92.7 93.0 89.4 102.4 96.6 81.8 85.8 91.0 78.1 81.9 79.1 RSD % 5.1 7.5 5.8 3.9 4.3 4.6 2.6 4.9 2.8 4.6 4.6 5.8 4.7 8.3 4.4 4.7
Discussion
All 16 compounds could be analysed in positive ionisation mode, but diuron and linuron were analized in negative ionisation mode because some interferences were observed around the 0.01 g/L level. No interferences were observed for these compounds in negative ionisation mode and good limits of detection were obtained. The recovery values obtained for all compounds in both water matrices tested are all in agreement. It appears that there is no suppression of ionisation, which has not been the case for other methods [1] where suppression is noted by a reduction in the recovery of individual compounds in some matrices. Complete separation of all 16 compounds was not achieved by this method. Compounds 5 and 6 coeluted as well as compounds 15 and 16. However, all compounds were still quantitatively analysed as they differ in molecular weight. For instance, since compound 5 (simazine) has a molecular weight of 201 amu ( [M + H]+ = 202 m/z) and compound 6 (cyanazine) has a molecular weight of 240 amu ( [M + H]+ = 241 m/z), they were distinguished by selected ion mass spectroscopy. The same is applicable for compounds 15 and 16. Three compounds 12, 13, and 15 are isomers and all have a molecular weight of 229, but all three compounds show good chromatographic separation.
Potable water sample Recovery % 83.3 88.5 94.0 89.5 96.3 94.3 93.2 91.0 102.8 97.3 82.6 85.6 91.1 80.2 83.6 78.7 RSD % 5.6 7.8 4.6 4.6 4.7 4.5 3.3 4.6 3.1 4.8 4.4 6.2 4.8 6.8 4.6 4.8 LOD g/L 0.00146 0.00544 0.00293 0.00219 0.00352 0.00348 0.00209 0.00330 0.00257 0.00221 0.00208 0.00218 0.00179 0.00397 0.00268 0.00179
Figure 4 shows an extracted ion chromatogram, at 230 m/z ( [M + H]+ ) for compounds 12, 13, and 15. Two other compounds, 14 and 16, are also isomers with a molecular weight of 241 amu ( [M + H]+ = 242 m/z). Compounds 14 and 16 are also separated by the chromatography.
13.979 13.105 13.391 14 14.775 4 6 8 10 12
MSD1 230, EIC=229.7:230.7 (T:\AGILEN~2\922259~1\DATA\SAMPLE01.D) API-ES, Pos, SIM, Frag: 70 (TT)
800000 700000 600000 500000 400000 300000 200000 100000 0 16 18 min
Figure 4. Extracted positive ion chromatogram, 230 m/z, for compounds 12, 13, and 15.
Conclusion
The data shows that the method presented is capable of quantitative analysis for the 16 herbicides in a single analytical suite. The performance requirements set by the Drinking Water Inspectorate (DWI) for standard deviation, bias, and total error are all met for each individual herbicide. Although spiked recovery targets of 90%110% are not achieved in all cases, this can be compensated for by the application of recovery factors calculated from the performance data.
References
1. N. Cullum, P. Stephens and S.Evans, Determination of Acidic Herbicides in Groundwater and Potable Water by LC/MSD Using Selective Ion Monitoring. Agilent Technologies, publication 5988-5882EN www.agilent.com/chem

Analyzing Phenyl Ureas and Carbamate Pesticides Using ESI-, APPI-, and APCI-LC/MSD Application
Environmental and Food
Author
Chin-Kai Meng Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA
Abstract
Carbamates (a class of highly effective insecticides) and phenyl ureas (a class of herbicides) were successfully analyzed by liquid chromatography/mass spectrometry using electrospray ionization (ESI), atmospheric pressure chemical ionization (APCI), and atmospheric pressure photo ionization (APPI) sources. APPI and APCI exhibited lower background signals and fewer background peaks than ESI. Phenyl ureas, in general, exhibited better signal-to-noise ratios (S/N) from APCI or APPI than from ESI. However, it was just the opposite for carbamates. The S/N was better from ESI than from APCI or APPI. A small amount of post-column acetone was needed in the APPI source as a photoionizable dopant for charge transfer to the analyte.
Gas chromatography/mass spectrometry (GC/MS) systems have traditionally been used for analyzing pesticides in environmental samples. Due to the thermally labile nature of these compounds, liquid chromatography has been the method of choice for the separation of these pesticides. Many of the pesticides within the same class exhibit similar ultraviolet (UV) spectra. With the development of atmospheric pressure ionization (API) techniques, liquid chromatography/mass spectrometry (LC/MS) systems have become the preferred method for the analysis of ground- and surface-water contaminants. The LC/MS provides better sensitivity and specificity than typical diode array detectors (DAD), even if the analytes are not fully resolved from neighboring eluants. Three different API sources were used in this study for comparison: electrospray ionization (ESI), chemical ionization (APCI) [1], and photo ionization (APPI) [2]. In ESI, the ionization process happens before the solvent evaporation process. It is, therefore, a more universal ionization for polar compounds. In APCI and APPI, the analyte is not ionized until after solvent evaporation. APCI involves a charge transfer (proton or electron) between ionized reagent gas and the analyte. APPI requires that the analytes and/or a dopant be photoionized by absorbing photons from the 10.6 eV krypton light. The dopant then transfers the charge to the analyte, which could be thought of as photon-induced chemical ionization.
Introduction
Carbamates, a class of highly effective insecticides, are widely used worldwide to protect crops from pests. Phenyl ureas, a class of herbicides, are highly potent chemicals for agricultural weed control. All are potential endocrine disrupters found in ground- and surface-water samples.
Experimental
A mixture of carbamates and urea pesticides at 10 ppm (ng/L) in acetonitrile was purchased from AccuStandard (New Haven, CT). A series of dilutions in acetonitrile were made for the linearity studies. The compounds and their structures are shown in Figures 1 and 2, respectively. All experiments were performed on two G1946D LC/MSD systems equipped with ESI on one and APCI/APPI on the other. The LC/MS conditions are listed in Table 1.
O H3C H3C N H3C Cl O Cl H3C NH O HN O O
O NH CH3 H 3C NH
O O O CH3 CH3
Aminocarb
Barban
Carbaryl
Carbofuran
NH O Cl
CH3 CH3 H3C S
CH3 O O NH CH3 H3C S N CH3 O
O NH CH3
H 3C H3C N CH3
O HN
CH3
H3C
H3C
Chlorpropham
CH3 H3C N H3C S CH3 N O O HN O
Methiocarb
Methomyl
Mexacarbate
H3C NH O O CH3 O CH3 O
CH3 O NH CH3
Cl Cl
O O NH
CH3
Oxamyl
Propham
Propoxur
Swep
Figure 1.
The 12 carbamates used in this study.
Cl H3C O N CH3 NH Cl NH
O N CH3 F F NH F
O N CH3 Cl Cl
NH N
O O CH3
H3C
CH3
CH3
Diuron
Fenuron
Flurometuron
Linuron
CH3 NH Cl O N CH3
Cl Cl NH
O N CH3 CH3 CH3 NH
O NH
Monuron
Neburon
Siduron
Figure 2.
The seven phenyl ureas used in this study.
Table 1.
LC/MS Conditions ESI APCI APPI
Column Column temperature Column flow rate Solvent A Solvent B
Zorbax Eclipse XDB-C8, 4.6 x 50 mm, 3.5 m (p/n 935967-906) 30 C 1 mL/min H2O, 0.1% acetic acid or as specified Acetonitrile, 0.1% acetic acid or as specified n/a B: 10% at 0 min, 30% at 4 min, 80% at 10 min, 80% at 16 min 2 L 12 L/min 350 C 60 V 3500 V 60 psi n/a 0.1 0.15 Off 150800 Positive 30 C 1 mL/min H2O Acetonitrile 30 C 1 mL/min H2O Methanol
Post column Solvent gradient
n/a
40 L/min acetone added as dopant
B: 30% at 0 min, 40% at 4 min, 70% at 13 min, 80% at 16 min or as specified in the figure 2 L 11 L/min 275 C 110 V 4500 V 35 psi 225 C 0.1 0.15 Off 115500 Positive 2 L 11 L/min 275 C 110V 4500V 35 psi 225 C 0.1 0.15 Off 115500 Positive
Injection volume Drying gas flow Drying gas temperature Fragmentor voltage Vcap Nebulizer pressure Vaporizer Step size Peak width (min) Time filter Scan (m/z) Polarity

Figure 3 shows the effect of adding a post-column dopant (acetone) in APPI to significantly improve the responses of the analytes. Acetone at a flow rate of 40 L/min was introduced to the column effluent before entering the nebulizer. A gain of 1000 in peak height was seen for some of the analytes. The difference in peak-height gain among the analytes may be due to their structural difference and the charge transfer efficiency with the dopant.
140000 120000 100000 80000 60000 40000 20000 0 2 4 6 8 10 12 14 min
APPI without post-column acetone
2250000 2000000 1750000 1500000 1250000 1000000 750000 500000 250000 0 2 4 6 8 10 12 14 min
APPI with post-column acetone
Figure 3.
TICs show the effect of adding post-column acetone to enhance analyte signal in APPI. Methanol was used as solvent B for both TICs.
The LC mobile phase in APPI can interfere with ionization if the solvent has more affinity for the proton than the analyte. Figure 4 shows that acetonitrile can be a problem for baseline stability and it is always best to also try methanol in APPI. Typical background spectra between 15 and 15.5 minutes from the three sources appear in
Figure 5. Because ESI is a more universal ionization source for polar compounds and ionic modifiers are used in ESI, the baseline has a lot more peaks and the noise is usually 5 to 15 times higher than APCI and APPI.
1200000 1000000 800000 600000 400000 200000 2 4 6 8 10
Mobile A: H2O Mobile B: ACN with dopant
min
1000000 800000 600000 400000 200000 0 2 4 6 8 10
Mobile A: H2O Mobile B: MeOH with dopant
min
B: 10% at 0 min, 30% at 3 min, 80% at 8 min, 80% at 12 min
Figure 4.
TICs show that methanol is a better mobile phase than acetonitrile in APPI. Acetone was used as dopant.
6
116.1 117.1 158.9 183.0 200 210.0 233.1 158.1 256.2 170.2 179.2 299.2 338.1 201.1 400 363.0 385.1 219.2 233.2 250 279.1 298.3
Figure 5.
149.1
128.1 138.2 150
136.1
150 205.1
200 188.1
171.2
200 233.2
220.2
337.1
426.1
250 250.1
260.2 259.2
259.2
485.0 522.5 275.1 550.5 288.3
280.2 288.3 279.2
300
300
600
313.2
ESI
APPI
APCI
Max: 3155
Max: 1091
Max: 20907
633.1
350
350
Baseline noise of the three sources are compared. ESI exhibits higher noise and more peaks in the baseline.
391.3
794.6
The spectra of 20 ng Diuron on column from the three sources (ESI, APCI, and APPI) appear in Figures 6, 7, and 8. The peak at mass 233 is the [M+H]+ peak. Peaks at masses 233, 235, and 237 match the isotope peak-intensity pattern for two chlorine atoms, which is additional information for compound confirmation. Although ESI gives better response compared to the other two sources, it has lower signal-to-noise ratio (S/N).
4.777
6.637
2500000
Diuron
ESI
9.134 8.458 9.201 9.453 10.427
2000000
1500000 2.266 2.340 2.230 2.687 5.095
9.607
1000000
10.676 10.538
11.445 11.624
4.667
500000
0 2 4 6 8 10 12 14 min
60 40
Diuron
235.0
80
233.0
100
10.237
Max: 439851
210.1
20 0 160 180 200
220
222.1
237.0
240
m/z
0.05% formic acid in both solvents A and B
Figure 6.
The TIC of all target compounds from ESI. The bottom half is the ESI spectrum of 20 ng Diuron on column.
13.180
300000 5.476 250000 200000 150000 1.017 1.208 1.910 3.954 100000 0.449 50000 0 2 4 6 8 10 12 14 5.977 6.148 6.346 6.580 8.115
2.300
APPI
Diuron
11.041 11.292 11.667 13.453
9.040
10.012 10.426
7.225
9.550
min
60 40 20 0
160 180
Diuron
235.1 200 220 237.1 240
80
233.1
100
Max: 66405
m/z
Figure 7.
The TIC of all target compounds from APPI. The bottom half is the APPI spectrum of 20 ng Diuron on column. APCI
400000 5.310 300000 200000 5.573 7.124 8.356 100000 0 2 4 6 8 10 12 14 min 1.008 1.170 6.224 6.471 8.009
2.202
Diuron
10.957 13.401
10.381 10.569
11.196 11.569
8.942
100 233.1 80 60 40 20 0 160 180 200 220 240 m/z 237.1 235.1
Diuron
Max: 85332
Figure 8.
The TIC of all target compounds from APCI. The bottom half is the APCI spectrum of 20 ng Diuron on column.
15.027 15.386 15.551
Figure 9 shows the spectra of Siduron (protonated ion at m/z 233) from all three sources. Protonated dimer (m/z 465) and sodiated dimer (m/z 487) of Siduron are fairly significant peaks in the ESI spectrum. However, only negligible amounts of protonated dimers (no sodiated dimers) were observed in the APPI and APCI spectra. This is mainly due to the different sequence of desolvation and ionization steps. In ESI, ionization happens before desolvation. In APPI and APCI, ionization comes after desolvation.
233.2
7 6 5 4 3 2 1 0
APPI
Max: 111072 234.2

250
300
350
400
450
500 m/z
10 8 6 4 2 0
233.2
APCI
Max: 171264 234.2

250
300
350
400
450
500 m/z
233.1
100 80
ESI
Max: 1.62867e+006 465.2
[M+H]+
60 40
234.1
466.3
[2M+H]+
255.1
487.2
[2M+Na]+
20 0
250
300
350
400
450
488.3
500 m/z
Figure 9.
Spectra of Siduron from APPI, APCI, and ESI. Significant dimer peaks are observed in ESI.
Figure 10 compares the S/N of some of the target compounds analyzed by the three sources. The figure shows that carbamates, in general, give better S/N from ESI than from APCI or APPI. However, it is just the opposite for phenyl ureas, for example, the S/N is better from APCI or APPI than from ESI. It is also worth noting that APPI shows consistently higher S/N than APCI for this study.
A calibration based on the Methomyl [M+H]+ (m/z 163) is linear over the concentration range of 202000 pg on column. Figure 11 shows the ESI calibration curve with the linear correlation coefficient of 0.9996. Four of the Methomyl peaks that were used for the linearity calibration appear in Figure 12.
ESI APPI APCI
Carbofuran
Mexacarbate
Diuron
Monuron
Carbamate Figure 10. The S/N of four analytes from the three ionization modes.
Phenyl urea
Figure 11. The calibration curve of Methomyl, by ESI in SIM mode (mass 163), shows good linearity over the concentration range of 202000 pg on column (2-L injection).1
1
The calibration curve was generated using the PE Sciex AnalystTM software.
10
20 pg on column
100 pg
200 pg
1000 pg
Figure 12. The AnalystTM software shows four of the Methomyl peaks (163 amu) that were used for generating the calibration curve in Figure 11.
11
Three different modifiers (formic acid, ammonium acetate, and acetic acid) were used in the ESI mode to compare their effectiveness. The results in Figure 13 show that, in general, the peak intensities are comparable among three modifiers. However, the elution order of the peak Mexacarbate (as well as Aminocarb) was different. Therefore, a single selected-ion-monitoring (SIM) method should not be used for different modifiers.
1000000 800000 600000 400000 200000 0 2 1000000 800000 600000 400000 200000 0 2
Mexacarbate
0.1% Formic acid
10
12
min
Mexacarbate
10 mM NH4Ac
10
12
min
1000000 800000 600000 400000 200000 0 2
Mexacarbate
0.1% Acidic acid
10
12
min
Figure 13. Different modifiers in ESI resulted in comparable peak intensities but changed the peak elution order.
12
Conclusions
Carbamates and phenyl ureas were successfully analyzed using ESI, APCI, and APPI sources. APPI and APCI have lower background signals and fewer background peaks than ESI. The results show that carbamates, in general, give better S/N from ESI than from APCI or APPI. However, it is just the opposite for phenyl ureas, for example, the S/N is better from APCI or APPI than from ESI. A small amount of post-column acetone was needed in the APPI source as a dopant. The dopant is photoionizable and is used to transfer charge to the analyte. The typical ESI quantitation limit for these compounds is 20 pg on column using SIM. Some carbamates and phenyl ureas (for example, Propoxur, Carbofuran, and Siduron) exhibited protonated and sodiated dimers in the full scan ESI spectra.
References
1. Basics of LC/MS, Agilent Technologies Primer, Publication 5988-2045EN, February 15, 2001. 2. PhotoMate Atmospheric Pressure Photoionization Source from Syagen Technology, Agilent Technologies Brochure, Publication 5988-3130EN, May 25, 2001.

13
Analysis of the Active Compound in an Agricultural Fungicide Formulation by Liquid Chromatography Application
Agricultural, Speciality Chemical, Environmental, Ag Chem
Author
James D. McCurry Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA
Abstract
Liquid chromatography was shown to be an excellent tool for the routine analysis of the active ingredient in an agricultural fungicide formulation. No extensive sample preparation or cleanup was required to remove the active ingredient from the rest of the sample matrix. The active ingredient was easily separated and detected using conventional reverse-phase conditions with a UV/VIS detector.
This work was done on a commercially available fungicide formulation. The active ingredient in this product is 6.5 % (wt) of N,N-[1,4-piperazinediylbis (2,2,2-trichloroethylidene)] bisformamide. This is also known as triforine (CAS registry number 26644-46-2) and the structure is shown in Figure 1. The inactive ingredients in this formulation are listed as cyclohexanone, N-methyl pyrrolidone and Atlox 3406-F. The Atlox 3406-F is an agricultural dispersant that contains ionic and nonionic surfactants and mixed aromatic solvents. Electrospray ionization liquid chromatography/mass spectrometry (LC/MS) analysis has shown that the triforine can be easily separated and identified in the formulation [1].
Cl Cl * Cl NH O
Introduction
Agricultural chemical formulations usually contain an active ingredient and several inert components, such as surfactants, that are designed to enhance the efficacy of the product. Gas chromatographic analysis of these formulations cannot be performed due to the polarity or thermal instability of the active ingredient as well as the high molecular weight and polarity of the surfactants. Therefore, liquid chromatography offers the best solution for the routine analysis of the active ingredients in an agricultural formulation.
O HN Cl Cl N * Cl
*Optically active
Figure 1.
Chemical structure of triforine, the active ingredient in some commercial fungicide formulations.
Experimental
A 10% (v/v) solution of the fungicide formulation was made in acetonitrile. This solution was run on the Agilent 1100 Series LC System. This system included a vacuum degasser, a binary pump, an autoinjector, a thermostated column compartment, and a diode array UV/VIS detector. LC instrument conditions for this analysis are shown in Table 1.
Table 1. LC Analysis Conditions

Figure 2 shows the chromatogram of the fungicide formulation. The active ingredient in the formulation, triforine, elutes as two chromatographic peaks between 7.5 minutes and 7.8 minutes. The presence of two triforine peaks is due to the stereochemistry of the structure. Figure 3 shows the four triforine stereoisomers. These four configurations can be grouped into two pairs of mirror images that are diastereoisomers. The S,R and R,S configurations are mirror images that are superimposable, resulting in a meso compound that exhibits no optical activity or differences in physical properties. Therefore, because the S,R and R,S configurations are identical, they will elute as one chromatographic peak. The second pair of mirror images are the R,R and S,S configurations. These are not superimposable and are, therefore, enatiomers that will have different optical activity, but identical physical properties. Conventional reverse-phase liquid chromatography cannot separate these enantiomers, and they will co-elute as a single peak. However, these enantiomers can be separated from the meso compound by reversephase LC. This is why there are two triforine peaks, one for the meso compound and one for the enatiomers. Without pure standards of the stereoisomers, it is not possible to determine which configurations can be attributed to the observed chromatographic peaks.
Liquid chromatograph conditions Column: Mobile phase A: Mobile phase B: Mobile phase gradient: Flow rate: Injection volume: Column temperature: Detector: Signal wavelength: Signal bandwidth: Reference wavelength: Reference bandwidth: Zorbax XDB-C8,150 4.6 mm, 5 m (p/n 993967-906) 0.1% Formic acid in water Acetonitrile 30% B at 0 min, 50% B at 7 min, 95% B at 10 min 1.0 mL/min 1 mL 30 C Diode array 254 nm 10 nm 500 nm 40 nm
7.554 min
mAU
7.757 min
40
30
20
254 nm
10
0 0 2 4 6 8 10 12 14 min
Figure 2.
LC of an agricultural fungicide formulation containing the active ingredient triforine. The two peaks at 7.554 min and 7.757 min were shown to be optical isomers of triforine.
Mirror
Mirror
O HN H
(S)
O NH HN H Cl3C
(R)
O NH
CCl3
Cl3C
(R)
(S)
CCl3
N H
(R)
N CCl3 Cl3C
(S)
N H H
(R)
N CCl3 Cl3C
(S)
NH O (S,R)
HN O (R,S) O
NH
HN O
(R,R)
(S,S)
Meso structure no optical activity
Enantiomers optically active
Figure 3.
The four triforine stereoisomers arising from the two chiral carbons in the structure. These two pairs of mirror images account for the two triforine peaks observed in the chromatogram.
Conclusions
Liquid chromatography was shown to be an excellent tool for the routine analysis of the active ingredient in an agricultural fungicide formulation. No extensive sample preparation or cleanup was required to remove the active ingredient from the rest of the sample matrix. The active ingredient was easily separated and detected using conventional reverse-phase conditions with a UV/VIS detector.
References
1. McCurry, J.D., and Zavitsanos, P., Analysis of Components, Contaminants, and Impurities in Fungicide Chemical Formulations by GC/MS and LC/MS, Agilent Technologies Application Note, Publication 5988-6085EN, April 2002.

Herbicides
Examples of Bonded-Phase Selectivity Differences
Application
Environmental
Robert Ricker
1. Bentazon 2 .Tebuthiuron 3. Simazine 4 Atrazine 5. Prometron 6. Diuron 7. Propazine 8. Propanil 9. Prometryne 10. Metolachlor
Highlights
Useful differences in band spacing can often be obtained by varying column type, keeping mobile-phase conditions constant.
For the same mobile phase, stationaryphase strength (retentivity) is in the order C8>Phenyl>CN for these sterically protected bonded phases.
Long-term stability and reproducibility are provided by ZORBAX StableBond columns.

Conditions: ZORBAX StableBond Columns (4.6 x 150 mm) (Agilent P/N: above) Mobile Phase: 35% ACN, 65% H20, 1mL/min
A wide range of pesticides can be separated with excellent peak efficiency and peak shape.
Robert Ricker is an application chemist based at Agilent Technologies, Wilmington, Delaware.
For more information on our products and services, visit our website at: www.agilent.com/chem
Copyright 2002 Agilent Technologies, Inc. All Rights Reserved. Reproduction, adaptation or translation without prior written permission is prohibited, except as allowed under the copyright laws. Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Printed in the USA April 25, 2002 5988-6340EN
Pesticides
Analysis of Pesticides in Drinking Water
Application
Environmental
Robert Ricker
1. Desisopropylatrazine 2. Metamitron 3. Fenuron 4. Chloridazon 5. Desethylatrazine 6. Metoxuron 7. Carbetamid 8. Bromacil 9. Hexazinon 10. Simazine 11. Metribuzin 12. Desethylterbutylazine 13. Carbutilat 14. Methabenzthiazuron 15. Chlortoluron 16. Atrazine 17. Monolinuron 18. Diuron 19. Isoproturon 20. Metobromuron 21. Metazachlor 22. Buturon 23. Propazine 24. Dimefuron 25. Terbuthylazine 26. Linuron 27. Chlorbromuron 28. Chloroxuron
Courtesy of Dr. rer.nat. Claus Schlett, Gelsenwasser AG
Highlights
The 3mm-diameter ZORBAX LowVolume Columns offer significant advantages over standard 4.6 mm i.d. columns: - A 2-fold increase in detection sensitivity -- less sample required - A 50% solvent savings -- and reduced solvent-disposal costs
ZORBAX SB-C18 has a sterically protected bonded phase that permits reliable results run after run.
28 pesticides are separated with good resolution and peak shape in a single run using simple mobile phases.
Conditions: ZORBAX SB-C18 (3.0 x 250 mm) (Agilent P/N: 880975-302) Mobile Phase:! A=2mM Sodium Acetate (pH 6.5) with 5% ACN B=100% Acetonitrile!(ACN) Gradient Elution:!!2min, 10% B; 10 to 45% B in 70 min. Injection volume 25l, 0.35 mL/min, 40C, Detect. UV (245 nm)
SUMMARY A variety of pesticides have had extensive use in many countries around the world over the last twenty years. These chemicals are currently present in surface water in very low concentrations, and need to be analyzed. High-Performance Liquid Chromatography with diode-array detection is an excellent tool for analysis of these compounds. TECHNICAL DETAILS Drinking-water regulations have been developed in many locations that set limits for maximum allowable levels of pesticides. A reliable method of analysis is required to monitor these levels, preferably in a single run. HPLC using diode-array detection after solid-phase extraction can meet this need. Generally, substances can be detected in concentrations less than 0.1 mg/L (i.e., the maximum level set in the drinking-water regulation of Germany). ACKNOWLEDGMENT Agilent Technologies, Inc. wishes to thank Dr. rer.nat. Claus Schlett, Gelsenwasser AG who developed the method and provided this chromatogram.
High-Speed Separation of Pesticide Mixture
Application
Environmental
Robert Ricker
Columns with particle sizes under 5 m and columns with dimensions up to ten-times smaller than standard analytical sized columns are ideal for high-speed analyses. In this application, smaller particles and column dimensions are used to resolve six pesticides in less than two minutes. Extra-column volume is a crucial factor in chromatographic performance when using low volume columns. In this example, no modification to a modern instrument (plumbing or flow cell) is necessary.
Operating Conditions:
HPLC System: Column: Mobile Phase: Detection: Flow: Temperature: Agilent 1100 with quaternary pump ZORBAX Eclipse XDB-C18 Rapid-Resolution (3.5m) Cartridge-Column, 4.6 x 30 mm Agilent Part No. 931975-932 MeOH: water (60:40) UV 254 nm with standard flow cell (13 L) 2 mL/ min. ambient
Highlights
Reducing column length and particle size simultaneously can: - Reduce analysis time - Maintain resolution - Reduce solvent use
mAU 500 400 300 200 100 0 0
ZORBAX Eclipse XDB-C18

1. Tebuthiuron 2. Atrazine 3. Diuron 4. Propazine 5. Propranil 6. Prometryne
Higher flow rates are possible due to decreased pressure when using ZORBAX Rapid Resolution cartridge columns.
6 4
0.5
1.5
2.5
3.5
min
Analysis of Components, Contaminants, and Impurities in Fungicide Formulations by GC/MS and LC/MS Application
Agriculture, Specialty Chemical, Environmental, Ag Chem
Author
James D. McCurry Paul Zavitsanos Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA
Abstract
A commercially available fungicide formulation was analyzed by both gas chromatography/mass spectrometry (GC/MS) and electrospray ionization liquid chromatography/mass spectrometry (ESI-LC/MS). The GC/MS analysis provided a detailed look at the volatile components in the formulation, but did not yield any results for the active ingredient, triforine. The ESI-LC/MS provided information on the stereoisomers of triforine as well as the nonvolatile surfactants and contaminants in the formulation. This paper demonstrates the complementary nature of these two analytical techniques when trying to fully characterize a complex chemical formulation containing a broad range of components.
unexpected impurities and breakdown products that can affect product quality. However GC/MS can only provide meaningful information for compounds that are volatile, nonionic, thermally stable, and have relatively low molecular weight. Liquid chromatography is much better suited to analyzing compounds that are nonvolatile, ionic, polar, thermally labile, or have high molecular weight. This includes about 80% of all known organic compounds [1]. When coupled with a modern atmospheric pressure ionization (API) mass spectrometer, LC/MS offers a complementary tool to GC/MS in the chemical diagnostic laboratory. Commercial pest control formulations contain one or more active compounds along with a recipe of ingredients that can play an important role in the products efficacy. These inactive ingredients are often a combination of solvents and surfactants that allow for easy application and dispersal of the active ingredient onto the target substrate. For this work, an over-the-counter fungicide formulation was purchased at a local home products store. The active ingredient in this product is 6.5 % (wt) of N,N-[1,4-piperazinediylbis(2,2,2-trichloroethylidene)] bisformamide. This is also known as triforine (CAS registry number 26644-46-2), and the structure is shown in Figure 1. The inactive ingredients in this formulation are listed as cyclohexanone, N-methyl pyrrolidone, and Atlox 3406-F. The Atlox 3406-F is an agricultural dispersant that contains ionic and nonionic surfactants and mixed aromatic solvents.
Introduction
Gas chromatography/mass spectrometry (GC/MS) is an indispensable tool for solving complex problems in the chemical industry. This fast and powerful technique yields detailed information about the expected compounds in the mixture along with any
Cl Cl * Cl NH O
Table I.
GC/MS Analysis Conditions
Gas chromatograph conditions Column: Carrier gas: Flow rate: Inlet: 30 m 0.25 mm HP5-MS, 0.25 m (p/n 19091S-433) Helium at 13.00 psi 1.6 mL/min., constant flow mode Cool on-column at 50 C, oven track mode
N O HN Cl Cl Figure 1. N * Cl
*Optically active
Oven temperature program: 50 C for 3 min 10 C/min to 275 C 275 C for 4 min MS Transfer line: Injection volume: Electron multiplier: Solvent delay: Scan range: Scan threshold: A/D Samples: Scan rate 280 C 1 L 1400 V 3 min 30 to 800 m/z 50 counts 2 1.95 scans/s
Chemical structure of triforine, the active ingredient in some commercial fungicides. The nominal molecular weight is 432, and the structure contains two optically active carbons.
Mass spectrometer conditions
A complete analysis of this formulation requires GC/MS to separate and identify the volatile components and LC/MS for the surfactants and polar components. Analysis of the active ingredient, triforine, presents a separate challenge. References for triforine analysis cite gas chromatography as the method of choice when analyzing environmental residues [2]. However, the melting point is reported to be 155 oC with decomposition, indicating that gas chromatography may only be possible with on-column injection.
Electrospray Ionization Liquid Chromatography/Mass Spectrometry (ESI-LC/MS) The same fungicide sample was run on the Agilent 1100 Series LC/MSD. This system included a vacuum degasser, a binary pump, an autoinjector, a thermostatted column compartment, and the LC/MSD SL quadrupole mass spectrometer. LC/MS instrument conditions for this analysis are shown in Table 2.
Experimental
Gas Chromatography/Mass Spectrometry (GC/MS) A 1% (v/v) solution of the triforine formulation was made in acetonitrile and the GC/MS analysis was performed with an Agilent 5973 GC/MS system. The components in this system were a 6890N gas chromatograph, a 7683 autoinjector, and a 5973 mass spectrometer. A cool-on-column inlet in the Agilent 6890 GC was used to avoid decomposition of the triforine. Instrument conditions for the GC/MS analysis are listed in Table 1.

Gas Chromatography/Mass Spectrometry (GC/MS) The complex nature of this fungicide formulation is revealed when one looks at the GC/MS data. Figure 2 shows the total ion chromatogram (TIC) of the fungicide sample. The volatile components
Table 2.
LC/MS Analysis Conditions
Liquid chromatograph conditions Column: Mobile phase A: Mobile phase B: Mobile phase gradient: Flow rate: Column temperature: Injection volume: Source: Drying gas flow: Nebulizer: Drying gas temperature: Vcap: Stepsize: Peak width: Time filter: Scan range Fragmentor 150 4.6 mm Zorbax XDB-C8, 5 m (p/n 993967-906) 0.1% Formic acid in water 0.1% Formic acid in acetonitrile 30% B at 0 min; 50% B at 7 min; 95% B at 10 min 1.0 mL/min 30 C 1 L Electrospray 12 L/min 40 psig 350 C 3500 V (positive) and 3000 V (negative) 0.1 amu 0.1 min On 120 to 1200 m/z Fixed at 60 V
in the formulation are easily identified from the mass spectral data. The major solvents, cyclohexanone and N-methyl-2-pyrrolidone, dominate the chromatogram while smaller amounts of C9 aromatics, C10 aromatics, and substituted napthalenes are easily separated and identified. There were no peaks in the TIC whose spectra matched the triforine reference spectra from the Wiley mass spectral library. An extracted ion profile using the triforine base peak of 203 m/z did not produce any chromatographic peak indicating the presence of triforine. From this data, it appears that the triforine did not elute from the column into the mass spectrometer. However, a spectral average of the large hump between 18 and 20 minutes shows an isotope pattern indicating one chlorine atom (Figure 3A). Since no chlorinecontaining species other than triforine are components in the formulation, the presence of chlorine and the broad peak shape indicates triforine decomposition in the gas chromatograph. The peak at approximately 20-minute retention time also has a mass spectrum containing an isotope pattern indicating the presence of two chlorine atoms in the structure (Figure 3B). This peak could be a decomposition product or a contaminant in the formulation.
Mass spectrometer conditions
5 2 4 1 2 3 4 5 6 7 1-hexanol Cyclohexanone C9-aromatics N-methyl-2-pyrrolidone C10-aromatics Naphthalenes Triforine decomposition
6 3
10
12
14
16
18
20
Figure 2.
GC/MS TIC showing the complex volatile components in the commercial fungicide formulation.
145 158 187 152
215
140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 320 m/z
288
144
140 150
157
160 170 180
189
190 200 210
225
220 230 240 250
259
260 270 280 290 300 310 320 m/z
Figure 3.
(A) Average mass spectrum of broad hump between 18 and 20 minutes of TIC. Isotope patterns of the peaks at m/z 145, 158 and 187 indicate the presence of one chlorine atom. (B) Mass spectrum of the peak at 20 minutes shows the presence of two chlorine atoms in the structure.
Electrospray Ionization Liquid Chromatography/Mass Spectrometry (ESI-LC/MS) The positive ion ESI-LC/MS chromatogram is shown in Figure 4. Several major peaks are observed along with several minor components. The spectra of the three peaks eluting between 0 and 2 minutes are shown in Figure 5. Since electrospray is a soft ionization technique, these spectra do not exhibit the detailed fragmentation
needed to interpret structures for these three compounds. However, peak number 2 does have an isotopic pattern indicating the presence of two chlorine atoms in the structure. This compound could be a contaminant related to triforine production or a triforine decomposition product. Figure 6 shows the spectra of the three peaks between 10.5 and 13 minutes. These compounds are the various surfactants that make up the agricultural dispersant used in the formulation.
3 8
2 6
10
12
Figure 4.
TIC from positive ion ESI-LC/MS of fungicide formulation.
167
122
Peak 1
126 146 185 208
200 m/z
120
140
160
180
260 262
Peak 2
264
250 275
300
325
350
375
m/z
199
Peak 3
200
180 190 200
210
220
m/z
Figure 5.
Electrospray spectra from LC/MS peaks 1, 2, and 3. The spectra from peak 2 shows an isotope pattern indicating two chlorine atoms in this structure. This compound may be a contaminant in the formulation from the active ingredient triforine.
Peak 6
600
700
800
900
m/z
Peak 7
400
600
800
1000
m/z
Peak 8
400
500
600
700
800
m/z
Figure 6.
Electrospray mass spectra of LC/MS peaks 6, 7, and 8 from Figure 4. These compounds are the surfactants used in the formulation.
The spectra of LC/MS peaks 4 and 5 (Figure 7) are identical and correspond to the active ingredient, triforine. The protonated molecular ion is observed at m/z 433 along a sodium adduct at m/z 455. The multiplets for m/z 433 to 439 and m/z 455 to 461 exhibit an isotopic pattern consistent with six chlorine atoms. The ion at m/z 388 is due to a rearrangement and subsequent loss of a formamide group from the protonated molecular ion (m/z 433). This is also confirmed by the isotopic pattern indicating six chlorine atoms (m/z 388 to 396). The presence of two triforine peaks in Figure 4 can be explained by the stereochemistry of the structure. Triforine contains two optically active carbons that give rise to four stereoisomers. Figure 8 shows the four configurations that can be grouped into two pairs of mirror images that are diastereomers. The S,R and R,S configurations are mirror
images that are superimposable, resulting in a meso compound that exhibits no optical activity or differences in physical properties. Therefore, because the S,R and R,S configurations are identical, they will elute as one chromatographic peak. The second pair of mirror images are the R,R and S,S configurations. These are not superimposable and are, therefore, enatiomers that will exhibit different optical activity, but identical physical properties. Conventional reverse-phase liquid chromatography cannot separate these enantiomers, and they will co-elute as a single peak. However, these enantiomers are not mirror images of the meso compound and can be chromatographically separated from the meso compound. This is why there are two triforine peaks, one for the meso compound and one for the enatiomers. Without pure standards of the stereoisomers, it is not possible to determine which configurations can be attributed to the observed chromatographic peaks.
(M+H)+ -43
Peak 4
7.554 min
(M+H)+ (M+Na)+
390 392
Peak 5
7.757 min
388
394
433 435 437 439
400
420
440
457 459 461

460 m/z
Figure 7.
Electrospray mass spectra of peaks 4 and 5 from Figure 4. Both spectra show a protonated molecular ion at m/z 433 representing the active ingredient triforine. There is also a sodium adduct (m/z 455) of triforine observed for both peaks. A rearrangement and loss of a formamide group from the protonated molecular ions give rise to the multiplet at m/z 388 to 396. Mirror Mirror
O HN H
(S)
396
O NH HN H Cl3C
(R)
O NH
CCl3
Cl3C
(R)
(S)
CCl3
N H
(R)
N CCl3 Cl3C
(S)
N H H
(R)
N CCl3 Cl3C
(S)
NH O (S,R)
HN O (R,S) O
NH
HN O
(R,R)
(S,S)
Meso structure no optical activity Figure 8.
Enantiomers optically active
The four triforine stereoisomers arising from the two chiral carbons in the structure. These two pairs of mirror images account for the two triforine peaks observed in the chromatogram (Figure 4). 7
The fungicide formulation was also run by ESI-LC/MS in the negative ion mode. The results of this analysis are shown in Figure 9. The negative ion mass spectra for these two peaks are shown in Figure 10. For both triforine peaks, the most stable negative ion species is the chloride adduct (m/z 467). However, the spectra for the first peak contains a
7.554
deprotonated molecular ion (m/z 431) and a formate adduct (m/z 477) that is not observed in the spectra of the later eluting peak. This selective adduct formation is likely related to the stereochemstry of the triforine, but again, without pure standards, the correct configurations cannot be assigned to the chromatographic peaks.
7.757
10
12
Figure 9.
TIC from negative ion ESI-LC/MS of fungicide formulation.

_ (M+Cl)
469 471
RT = 7.554 min
(M+HCO2)
479 _
(M+TFA)
_ 545 547
467
473
477
483
(M_H)
(M+NH4CO2)
494 494 494 494
431 433 435 437
RT = 7.757 min
425
450
475
500
525
550
551 m/z
Figure 10. Negative ion electrospray mass spectra of the two triforine peaks. The spectra from peak at 7.554 minutes shows a deprotonated molecular ion (m/z 431) and a formate adduct (m/z 477) that is not seen in the later eluting peak (7.757 minutes).
549
481
Conclusions
This paper demonstrates the complimentary nature of GC/MS and LC/MS when trying to characterize a formulation that is composed of many different chemical species. The volatile compounds in the formulation can be easily separated and identified by GC/MS. In this case, polar solvents such as cyclohexanone and N-methyl-2-pyrrolidone were the major components while 1-hexanone, C9 aromatics, C10 aromatics, and substituted naphthalenes were present as minor components or contaminants. However, GC/MS did not yield any information on the active fungicidal ingredient, triforine, a hexachlorinated compound. This was most likely due to thermal decomposition during GC/MS analysis. Evidence for this was seen in a broad chromatographic hump containing chlorine-containing constituents. The nonvolatile components in this fungicide were quickly analyzed by ESI-LC/MS. This analysis yielded information on several polar contaminants, some containing chlorine, which may be by-products of triforine production or triforine breakdown products. Also observed were several surfactants that are used in agricultural products as dispersants. The LC/MS analysis did yield significant information on the triforine active ingredient, showing a distribution of stereoisomers in the formulation.
References
1. Willoughby, R., Sheehan, E, and Mitrovich, S. A., Global View of LC/MS: How to Solve your Most Challenging Problems, p. 80, Global View Publishing, 1998. 2. The Pesticide Manual 9th Edition, Worthing, C.R. and Hance R.J. eds., p. 853, The British Crop Protection Council, 1991.

Rapid Analysis of CLP Pesticides Using High-Temperature DB-35ms and DB-XLB Columns Application
Environmental
Author
Cameron George Agilent Technologies, Inc. 91 Blue Ravine Road Folsom, CA 95630-4714 USA
thermal stability needed to achieve optimum resolution and sensitivity in the shortest possible time. These needs are realized with Agilent Technologies J&W Brand DB-35ms (primary) and DB-XLB (confirmation) columns. The excellent selectivity of high phenyl content phases for chlorinated pesticides is well documented. However, these phases typically suffer from poor thermal stability resulting in high bleed and excessively long analysis times. DB-35ms uses arylene-phase technology to provide improved thermal stability through a stiffening of the polymer backbone. The result is increased sensitivity and an upper temperature limit of 360C. The column bleed contribution to background noise is reduced, giving a much improved signal-tonoise ratio and increased usable sensitivity compared to standard 35%-phenyl phases. The high thermal limit translates into shorter analysis times, increased column lifetime and the ability to periodically bake the column at a high temperature to remove semivolatile contaminants. DB-XLB uses a proprietary second-generation arylene technology giving it the same 360C upper temperature limit and the lowest bleed of any phase available.
Abstract
Arylene-phase column pairs (primary and confirmation) permit high oven temperature for rapid analysis of CLP chlorinated pesticides. The columns are also suitable for phenoxy acids, haloacetic acids, polychlorinated biphenyls and Environmental Protection Agency Method 508.1 pesticides.
Introduction
Accurate identification and confirmation of trace level chlorinated pesticides are difficult tasks facing environmental laboratories. The chromatographic system, including the analytical columns, must be optimized. The gas chromatography (GC) columns must possess the selectivity, inertness and
Experimental
The columns and related inlet parts are described in Table 1.
Table 1.
Columns and Related Parts

id (mm) 0.32 0.32 0.53 Film (m) 0.25 0.50 Length (m) 30 30 5 Part Number 123-3832 123-1236 5181-3398 160-2535-5
Phase/Description DB-35ms DB-XLB Quartz deactivated splitter Deactivated fused silica guard column
This is a small sampling of the many DB columns and dimensions available.

Simple window diagramming identified the exact film thickness necessary to allow DB-XLB to give optimum confirmation power when run using the primary column temperature program. Figure 1 shows the optimized primary and confirmation chromatograms for the DB-35ms/DB-XLB column pair. Because these columns are designed for enhanced thermal resistance, it is not necessary to bake them excessively upon installation to reduce bleed to acceptable levels. A simple 1 to 2 hour conditioning period is typically more than adequate. Conditioning columns overnight is a common requirement with cyanopropyl- and trifluoropropyl-containing CLP pesticide columns. This practice can result in increased column activity and decreased column life time, but is not required with DB-35ms/DB-XLB. In short, you are ready sooner after column installation. Environmental laboratories are also interested in other gas chromatograpy/electron capture detector (GC/ECD) methods with the same dual column
pair used for the chlorinated pesticides. These methods include phenoxy acid herbicides (EPA Method 8151A), haloacetic acids (EPA Method 552.2), PCBs (EPA Method 8082) and EPA Method 508.1 pesticides. This column pair provides baseline resolution of all 8151A analytes on both columns in just over 16 minutes. In addition, DB-35ms and DB-XLB provide the best confirmation and the fewest coelutions of any dual column pair commercially available for 508.1 pesticides.
Conclusions
The DB-35ms and DB-XLB column pair perform rapid, high-resolution separations of CLP pesticides. The high temperature limit and low column bleed make these columns attractive for analyses of similar semivolatile sample mixtures.

Column:
DB-35ms 30 m 0.32 mm id, 0.25 m P/N: 123-3832 DB-XLB 30 m 0.32 mm id, 0.50 m P/N: 123-1236
Carrier: Oven:
Helium at 45 cm/sec (EPC in constant flow mode) 110 C for 0.5 min 110-320 C at 15 /min 320 C for 2 min Splitless, 250 C 30 sec purge activation time 50 pg per component
Column:
Injector:
Detector: Micro ECD, 350 C Nitrogen makeup gas (column + makeup flow = 30 mL/min constant flow)
2 3 7 12 8 1 5 9 10 11 13
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22.
Tetrachloro-m-xylene (SS) -BHC -BHC -BHC Heptachlor -BHC Aldrin Heptachlor epoxide -Chlordane -Chlordane Endosulfan I 4,4-DDE Dieldrin Endrin 4,4-DDD Endosulfan II 4,4-DDT Endrin aldehyde Endosulfan sulfate Methoxychlor Endrin ketone Decachlorobiphenyl (SS)
SS-Surrogate Standard
15 16 21 18 19 22
DB-35ms
14
17
20
C784
10 Time (min)
12
14
16
7 8 9 10
12 13 16 15
Complete resolution and confirmation of 22 CLP Pesticides in under 16 minutes!
11 21 18 17 19 22
6 5 14
DB-XLB
20
C783
10 Time (min)
12
14
16
Figure 1. CLP Pesticides 3
Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2001 Printed in the USA December 18 2001 5988-4973EN
A Complete Solution for Chlorinated Pesticides and Herbicides Using DB-35ms and DB-XLB Columns Application
Environmental
Author
Cameron George Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA
confirmation in less than 16 minutes. However, there is always the desire for a faster analysis. This note reports the result of changing to hydrogen carrier gas and increasing the oven ramp rate.
Experimental
Table 1 describes the columns and related inlet parts.
Table 1. Columns and Related Parts
id (mm) 0.32 0.32 0.53 Film (m) 0.25 0.50 Length (m) 30 30 5 Part number 123-3832 123-1236 5181-3398 160-2535-5
Abstract
DB-35ms (primary) and DB-XLB (confirmation) columns, used with inert inlet components and hydrogen carrier gas, perform CLP pesticide analyses in less than 15 minutes total cycle time. Phenoxy acids can be analyzed with the same configuration.
Phase/Description DB-35ms DB-XLB Quartz deactivated splitter Deactivated fused silica guard column
Introduction
Obtaining high quality data in the shortest possible time is a desire of all analytical testing laboratories. To achieve this goal, all aspects of the chromatographic system must be optimized. The GC columns must possess the selectivity, inertness and thermal stability needed to achieve optimum resolution and sensitivity in the least amount of time. For the analysis of CLP pesticides and phenoxy acid herbicides, these needs are met with DB-35ms (primary) and DB-XLB (confirmation) columns. In another Application Note[1], DB-35ms and DB-XLB show excellent selectivity and inertness for CLP pesticides, achieving one hundred percent
This is a small sampling of the many DB columns and dimensions available.

To reduce analysis time without a significant loss in resolution, the carrier gas was changed to hydrogen. Using hydrogen carrier gas with a linear velocity of 65 cm/sec, and increasing the oven ramp rate from 15 C/min to 25 C/min, reduces analysis time to less than 10 minutes. Considering a typical cool-down time of 4 to 5 minutes, the total instrument cycle-time is now less than 15 minutes.

Figure 1 shows the excellent resolution and confirmation available for CLP pesticides using DB-35ms/DB-XLB with hydrogen carrier gas and a properly scaled temperature program.
Baseline resolution for 22 CLP pesticides

7 2 3 12 10 13 8 9 11 56 4
DB-35ms
14 1516 18 19 17 21
22
20
Columns: DB-35ms 30 m 0.32 mm I.D., 0.25 m Part No.: 123-3832 DB-XLB 30 m 0.32 mm I.D., 0.50 m Part No.: 123-1236 Carrier: Hydrogen at 65 cm/sec (EPC in constant flow mode) Oven: 110 C for 0.5 min 110-320 C at 25 C/min 320 C for 2 min Injector: Splitless, 250 C 30 sec purge activation time 50 pg per component Detector: ECD, 350 C Nitrogen makeup gas (column + makeup flow = 30 mL/min constant flow)
10
C794
7 Time (min)
100% confirmation in under 10 minutes

12 2 7
DB-XLB
3 65 4
11 8 10 9
13 14 1516 18 19 17 20 21
22
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22.
10
Tetrachlorom-xylene (SS) -BHC -BHC -BHC Heptachlor -BHC Aldrin Heptachlor epoxide -Chlordane -Chlordane Endosulfan I 4,4-DDE Dieldrin Endrin 4,4-DDD Endosulfan II 4,4-DDT Endrin aldehyde Endosulfan sulfate Methoxychlor Endrin ketone Decachlorobiphenyl
C795
7 Time (min)
Figure 1 Analysis of CLP pesticides using DB-35ms and DB-XLB columns.
DB-XLB and DB-35ms have flexibility for a range of GC/ECD methods, a result of their ideal selectivity, inertness and the robustness of low bleed phases. Phenoxy acid herbicides (EPA Method 8151A) are nicely resolved with these columns. All twenty common herbicides are resolved in slightly
over 16 minutes, as shown in Figure 2. The analysis can be optimized for faster analysis. To obtain chromatograms and analysis conditions for additional GC/ECD methods, go to Agilent's Technical Support at www.agilent.com/chem
Columns: DB-35ms 30m 0.32 mm I.D., 0.25 m Part No.: 123-3832 DB-XLB 30m 0.32 mm I.D., 0.50 m Part No.: 123-1236 Carrier: Helium at 45 cm/sec (EPC in constant flow mode) Oven: 50 C for 0.5 min 50-100 C at 25 C/min 100-320 C at 12 C/min 320 C for 2 min
Injector:
Splitless, 250 C 30 sec purge activation time 50 pg per component Detector: ECD, 350 C Nitrogen makeup gas (column + makeup flow = 30 mL/min constant flow) 1. Dalapon 2. 3,5-Dichlorobenzoic acid 3. 4-Nitrophenol 4. Methyl-2,4-dichlorophenyl-acetate (SS) 5. Dicamba 6. MCPP
7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.
MCPA 4,4, Dibromooctafluorobi-phenyl (IS) Dichloroprop 2,4-D Pentachlorophenol 2,4,5-T,P 2,4,5-T Chloramben Dinoseb 2,4-DB Bentazone DCPA Picloram Acifluorofen
11
DB-35ms
8 18 13 14 15 9 6 7 10 16 17 19 20
12 5 1 2 3 4
C784-2
10 Time (min)
12
14
16
11
DB-XLB
5 1 2 3 4 6 7 9 10
12
18 19 15 16 17
13 14
C785
10 Time (min)
12
14
16
Figure 2.
EPA 8151A phenoxy acid herbicides.
Conclusions
DB-XLB and DB-35ms columns, when used with inert inlet components, hydrogen carrier gas and an appropriate carrier velocity, yield these benefits: Short analysis times for better productivity Excellent thermal stability with 360 C upper temperature limit Confirmation for CLP pesticides and phenoxy acid herbicide
Reference
1. Rapid Analysis of CLP Pesticides Using HighTemperature DB-35ms and DB-XLB Columns, Application Note 5988-4973EN, Nov 26, 2001.

Routine Analysis of Trace Level Carbamate Pesticides in Food Using LC/MSD
Application
Food, Environmental
Author
Masahiko Takino Agilent Technologies, Inc. 3-3-11, Niitaka, Yodogawa-ku, Osaka, 532-0033 Japan
Abstract
System linearity, sensitivity, and repeatability were evaluated on nine carbamates using positive electrospray ionization. Instrument detection limits were determined to be about 3 to 10 ppb for all the analytes in broccoli extract. The results demonstrate that the liquid chromatography with mass selective detection method is a viable technique to replace the nonspecific post-column derivatization process using fluorescence detection.
including N-methyl carbamates, are also suspected endocrine disrupters. The most often used method for carbamate determination is the post-column derivatization technique using fluorescence detection. Drawbacks of the fluorescence method include false positives due to lack of specificity, required additional hardware and plumbing, and insufficient limits of detection. False positives were confirmed in samples having a positive carbaryl identification by a liquid chromatograph (LC) with a fluorescence detector [1]. The goal of this study is to develop a routine liquid chromatograph/mass selective detector (LC/MSD) assay to detect carbamates in foods at low ppb levels without the derivatization step.
Experimental
A mixture of nine analytes (see Figure 1) in methanol was used to evaluate system linearity, sensitivity, and repeatability. A spiked broccoli extract was used to study the matrix effect on system performances. The LC/MSD system used was the Agilent 1100 MSD SL Quad system.
Introduction
Pesticides in food are a significant route to human exposure. Besides toxicity, many of the pesticides,
O (CH 3)2NC C N S CH3 O
O C NHCH3
CH3 S
O C(CH3 )2 CH N OCNCH3
OCONHCH3 CH2SC 2H5
Oxamyl
Aldicarb
Ethiofencarb
OCONHCH3
OCONHCH3 O CH3 CH3

CH3
OCONHCH3
CH3 O S
CH3 CH 3
CH3 S CH3
Methiocabsulfoxide
Bendiocarb
Methiocarb
OCONHCH3
O (CH 3) N
2
CH3
CO CH3
N N CH3
N( CH3 )2
C 2H 5
CH
OCONHCH3
CH3 S O
CH3 CH3 O
Methiocabsulfone
Figure 1. Structures of the analytes used in this study.
Pirimicarb
Fenobcarb
Table 1.
LC and MS Conditions Inertsil ODS3 (150 mm 2.1 mm 5 m) Zorbax Eclips XDB C18 (150 mm 2.1 mm 5 m) A: MeCN B: 10 mM CH3COONH4 in H2O 20% A/B to 100% A in 30 min 0.2 mL/min 40 C 10 L Positive electrospray 100 to 500 Base peak 10 L/min at 350 C 50 psi 60 V 7

TIC and SIM Figure 2 shows the total ion chromatogram (TIC) of all nine analytes at 1 ppm. The base peaks for the analytes were either MH+ or MNH4+, except Aldicarb, which was a fragment from the breakage of the N-O bond. Table 2, later in this note, lists the base peaks. Figures 3 and 4 are the selected ion monitoring (SIM) chromatograms of all nine carbamate standards at 0.2 ppb. The signal-to-noise (S/N) ratios of these peaks range from 7 to 56.
LC Conditions Columns
Mobile phase
Flow rate Column temp Sample volume MS Conditions Ionization Scan range SIM ion Drying gas Nebulizer gas Fragmentor EM gain
160000
1 Oxamyl
140000
6 Pirimicarb 7 Ethiofencarb 8 Fenobcarb 9 Methiocarb

5
8,9
2 Methiocarbsulfoxide 3 Methiocarbsulfone
1
120000
4 Aldicarb 5 Bendiocarb
2
100000
80000
4
60000
40000
20000 0 0 2.5 5 7.5 10 12.5 15 17.5 20
Retention time/min
Figure 2.
TIC of carbamate standards at 1 ppm.
1
700 600 500 400 0 2.5
Oxamy
m/z = 237
5 7.5 10 12.5 15 17.5 20
1200 1000 800 600 0
2 Methiocarbsulfoxide
m/z = 242
2.5 5 7.5 10 12.5 15 17.5 20
900 700 500 0
3 Methiocarbsulfone
m/z = 275
2.5 5 7.5 10 12.5 15 17.5 20
4 Aldicarb
2600 2200 1800 0 2.5 5 7.5 10 Retention time/min 12.5 15 17.5 20
m/z = 116
Figure 3.
SIM chromatograms of four carbamate standards at 0.2 ppb. The base peak is labeled in each chromatogram.
1500 1000 500 0 0 2.5 5 7.5 10 12.5
5 Bendiocarb
m/z = 224
15 17.5 20
6000 4000 2000 0 0 2.5 5 7.5 10 12.5
6 Pirimicarb
m/z = 239
15 17.5 20
9 Methiocarb
1500 1000 500 0 0 1500 1000 500 0 0 2.5 5 7.5 10 12.5 15 17.5 20 Retention time/min 2.5 5 7.5 10 12.5 15 17.5 20
7 Ethiofencarb
m/z = 226
8 Fenobcarb
m/z = 208
Figure 4.
SIM chromatograms of five carbamate standards at 0.2 ppb. The base peak is labeled in each chromatogram.
Ionization Electrospray is a soft ionization technique that produces a large number of molecular-related ions. These ions are typically protonated molecular ions [M+H]+. In this study, both [M+H]+ and [M+NH4]+ are dominant adduct ions. The fragmentor voltage is applied to the exit of the capillary and affects the transmission and fragmentation of sample ions by the in-source collision-induced dissociation in this region. By changing the voltage of the fragmentor, various degrees of fragmentation may be achieved. With a low voltage, there is little fragmentation; with higher voltages, a molecularrelated ion is fragmented to a larger degree.
Figure 5 has five SIM chromatograms collected at different fragmentor voltages ranging from 40 V to 120 V. Using the abundance of carbamates at 40 V as reference, Figure 6 shows the changes in relative abundances of carbamates at four fragmentor voltages. For compounds that do not fragment easily, ion transmission improves at the higher fragmentor voltage because the fragmentor voltage gives these ions a push that helps them travel through the relatively high pressure region between the exit of the capillary and the skimmer. For stable compounds, base peak abundance increases with fragmentor voltage up to about 100 V. Setting the fragmentor to either 60 or 80 V will have the best results.
Fragmentor voltage
3000000 2000000 1000000 0 0 3000000 2000000 1000000 0 0 3000000 2000000 1000000 0 0 3000000 2000000 1000000 0 0 3000000 2000000 1000000 0 0 2.5 5 7.5 10 12.5 15 17.5 20 2.5 5 7.5 10 12.5 15 17.5 20 2.5 5 7.5 10 12.5 15 17.5 20 2.5 5 7.5 10 12.5 15 17.5 20 2.5 5 7.5 10 12.5 15 17.5 20
40 V
1 2 3 4
8,9
60 V
80 V
100 V
120 V
Retention time/min Figure 5 SIM chromatograms of carbamate standards at different fragmentor voltages.
140
120
100
80
1 Oxamyl
60
6 Pirimicarb 7 Ethiofencarb 8 Fenobcarb 9 Methiocarb
2 Methiocarbsulfoxide 3 Methiocarbsulfone 4 Aldicarb 5 Bendiocarb
40
20
0 40 60 80 Fragmentor voltage (V) 100 120
Figure 6
Relative abundance of carbamate standards at different fragmentor voltages. Abundances at 40 V were used as the reference values. 5
Repeatability and Linearity Figure 7 shows five repeat injections of the carbamate standards at 5 ppb. The %RSD for each analyte ranges from 1.4% to 13% (Methiocarbsulfone). The 13% variation reflects the relatively weak response of the peak at this low concentration.
Figure 8 shows some of the calibration curves of carbamate standards. The responses are linear for all the analytes over a very wide concentration range. For the range from 0.2 to 1000 ppb, the correlation coefficients of all analytes are all greater than 0.998 (see Table 2).
20000 10000 0 2.5 20000 10000 0 2.5 20000 10000 0 2.5 20000 10000 0 2.5 20000 10000 0 2.5 5 7.5 10 12.5 5 7.5 10 12.5 5 7.5 10 12.5 5 7.5 10 12.5 5 7.5 10 12.5
5 (2.2%) 1 (2.5%) 2 (4.1%) 3 (13%) 4 (1.4%)
6 (1.9%) 7 (2.1%)
8 (1.7%) 9 (1.7%)
15
17.5
20
15
17.5
20
15
17.5
20
15
17.5
20
15
17.5
20
Retention time/min
Figure 7.
Repeatability of five repeat injections of the carbamate standards at 5 ppb. The percentage number in each parenthesis is the %RSD value for the peak.
Fenobcarb, 0.2-1000 ppb correlation coefficient = 0.998

7
800000 6000000 600000
160000 120000 80000 40000
4 5 6 8 7 9
0 50 ppb 4000000
8 9 11 10 12
5 ppb Correlation: 0.99903
400000 200000 0
2 3 5 4 6
0 500 ppb
Correlation: 0.99938
2000000
0 0
0.2-10 ppb
2-100 ppb
20-1000 ppb
Methiocarbsulfone, 0.2-1000 ppb correlation coefficient = 0.999

40000
300000
4
3000000
30000 200000 20000
5 8 9 11 10
Correlation: 0.99836 100000
2000000
2 3 54
6 0 500 ppb
10000
0 0
12
5 ppb
8
0 0
6
1000000
9
50 ppb
0.2-10 ppb
2-100 ppb
20-1000 ppb
Figure 8.
Calibration curves of carbamate standards.
Column Size Figure 9 is a comparison of 10-ppb chromatograms on a 2.1-mm column and a 4.6-mm column. The column length and particle size are the same. The responses from the 4.6-mm column are about 40% of the responses from the 2.1-mm column even though the flow rate is five times higher on the 4.6-mm column. Matrix Effect and LOD In many cases, the sample matrix interferes with the analyte responses. A SIM method collects less background signal. See Figure 10 for the SIM
chromatograms of 10-ppb carbamates in broccoli extract. The injection was 5 L on a 2.1-mm column. The percentage number in each parenthesis is the relative response of the analyte from the broccoli extract to the peak intensity from the analyte standard in methanol. Most of the analytes showed responses very similar (98 to 115%) to the standards, except Methiocarbsulfoxide, which had 71% of the response of the standard in methanol. The Limit of Detection (LOD) is defined as the sample concentration that produces a signal-to-noise (S/N) ratio of three in the broccoli extract. The LOD values of the nine carbamates are listed in Table 2.
6
35000 30000 25000 20000 15000 10000 5000 0 2.5 30000 25000 20000 15000 5 7.5 10 12.5 15 17.5 20
2.1-mm column (Flow rate: 0.2 mL/min)

5 2 1 3 7 4
8 9
4.6-mm column (Flow rate: 1.0 mL/min)

5 (36%)
6 (44.8%)
7 (30.8%)
8 (41%) 9 (34.6%)
2 (41.5%)
10000 5000 0 2.5 5 7.5 10 12.5 Retention time/min 15 17.5 20
3 (31.7%) 1 (29.4%)
4 (36.6%)
Figure 9.
Comparison of 10-ppb chromatograms on a 2.1-mm column and a 4.6-mm column. The percentage number in each parenthesis is the relative response of the analyte from the 4.6-mm column to the peak intensity from the 2.1-mm column.
2000 1000 0 2.5 3000 2000 1000 0 2.5 3000 2000 1000 0 2.5 4000 2000 0 2.5
1 (98%)
8000 4000 0
5 (98%)
7.5
10
12.5
15
17.5
20 20000 10000 0
2.5
7.5
10
12.5
15
17.5
20
2 (71%)
6 (101%)
7.5
10
12.5
15
17.5
20 10000 5000 0
2.5
7.5
10
12.5
15
17.5
20
3 (109%)
9 (121%) 7 (98%)
7.5
10
12.5
15
17.5
20 8000 4000 0
2.5
7.5
10
12.5
15
17.5
20
4 (115%)
8 (101%)
7.5 10 12.5 Retention time/min
15
17.5
20
2.5
7.5
10 12.5 15 Retention time/min
17.5
20
Figure 10. SIM chromatograms of 10-ppb carbamates in broccoli extract.
Table 2. # 1 2 3 4 5 6 7 8 9
Carbamates with the Base Peak, Correlation Coefficient and LOD in Broccoli Extract Base peak 237 [M+NH4]+ 242 [M+H]+ 275 [M+NH4]
+
Compound Oxamyl Methiocarbsulfoxide Methiocarbsulfone Aldicarb Bendiocarb Pirimicarb Ethiofencarb Fenobcarb Methiocarb
Correlation coefficient (0.21000 ppb) 0.999 0.999 0.999 0.999 0.999 0.999 0.999 0.998 0.998
LOD (S/N = 3 in broccoli extract) 3 ppb 10 10 2 1 3 2 1 2
116 (fragment) 224 [M+H]+ 239 [M+H]

+
226 [M+H]+ 208 [M+H]+ 226 [M+H]

+
Conclusions
Several carbamates can be analyzed routinely using LC/MSD with low ppb detection limits and excellent linearity in broccoli extract. The results demonstrate that the LC/MSD method is a viable technique to replace the non-specific post-column derivatization process using fluorescence detection.
References
1. Detection of Low Levels of Carbaryl in Food Using Agilent LC/MSD Trap System, Agilent Application Note, 5980-0332, April, 2000.

For more information on our products and services, visit our Web site at: www.agilent.com/chem.
Identification and Quantitation of Pesticides in the Parts-per-Trillion Range Using Retention Time Locking and GC/MS Application
Environmental, Food
Author
C. Kai Meng Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA
Abstract
The typical pesticide quantitation limit for a mass spectrometer in the Scan mode is in the sub-ppm range. By using a selected ion monitoring method, a lab can lower the target compound quantitation limit to the low parts-per-billion (pg/L) range using a retention time locked gas chromatography/mass spectrometry method. By adding large volume injection capability to the method, target compounds at parts-per-trillion can be quantified. A specially developed 567-compound retention time locking pesticide mass spectral library can automatically screen an acquired samples data file for all 567 compounds in seconds. The library can also be applied for rapid screening of samples acquired in selected ion monitoring method. Using the compound library information, a selected ion monitoring method for 80 target compounds was created in less than 2 hours without running any analyses.
detection limits and are easy to operate, the data do not provide sufficient information to confirm a compounds presence with confidence. Due to the universal nature of mass spectrometric detection, a mass spectrometer (MS) provides additional information and increased confidence in the assignment of compound identity. With recent advances in GC/MS hardware and software and the decrease in cost of ownership, more and more laboratories are routinely analyzing pesticide residue samples with MS detection. To match the GC/ESD detection limits and/or to eliminate sample concentration steps, a user must lower the MS detection limit by 2 to 3 orders of magnitude. This application note, discusses the following approaches. Run the MS in single ion monitoring (SIM) mode Make large volume injections (LVIs) Use higher electron multiplier voltage (EMV) For compound identification, a specially developed 567-compound retention time locking (RTL) [1] pesticide library could perform the entire 567-compound screening in seconds using Scan data. A subset of the library could be screened in seconds from SIM data.
Introduction
Most pesticides are typically analyzed on a gas chromatograph (GC) with element-selective detectors (ESDs). Although these ESDs provide low ppb
Experimental
A pesticide standard mixture was used to compare the lowest detection limits of splitless injection and LVI under Scan and SIM modes.
System Configuration for Screening and Quantitation: 6890 GC with a programmable temperature vaporizer (PTV) [2,3] inlet 5973 Mass Selective Detector (MSD) 7683 Automatic Liquid Sampler (ALS) tray and autoinjector HP-5MS capillary column (30 m 0.25 mm 0.25 m), P/N 19091S-433 G1701BA version B.00.00 MSD ChemStation software or higher G1049A MSD RTL Pesticide Database/Library
Table 3. MS Method Parameters Solvent delay 3 min Tune file Atune.u Transfer line 280 C MS Quad 150 C MS source 230 C Threshold 150 Sample # 2 Scan range 35 to 500 amu (in Scan mode) Forty (40) SIM groups (in SIM mode)
Table 4. Pesticide Screening Parameters for the SIM Method Extraction window Qualifier mode Qualifier % Zero qualifiers Subtraction mode Screen database 0.100 minute Absolute 30 Included Average start/stop Rtlpest.SCD
Table 1. GC Method Parameters Oven Inlet 70(2)/25/150(0)/3/200(0)/8/280(10) = 41.87 min PTV
Inlet pressure 17.30 psi (locked to methyl chlorpyrifos at 16.593 min), constant pressure mode

RTL [1] was used to: 1. Expedite data comparison in overlay format
Table 2. Injection Parameters Injection mode Injection volume (syringe) Injection speed Inlet temp Solvent vent 25 L (50-L syringe, P/N 5183-0318) Inject @ 100 L/min Draw @ 300 L/min Dispense @ 4500 L/min 40(0.35)/600/320 (3)/50/200 (Hold until end) Vent time = 0.29 min Vent flow = 150 mL/min Vent pressure = 0.00 psi 60 mL/min @ 2 min Deactivated, Multi Baffled (P/N 5183-2037) Liquid CO2 60 mL/min @ 2 min Deactivated, Multi Baffled (P/N 5183-2037) None Splitless 1 L (10-L syringe, P/N 9301-0713) Fast
2. Achieve lower target compound detection limit 3. Allow rapid pesticide screening using the RTL pesticide database/library 4. Help to differentiate isomers by their retention time (RT) differences 5. Eliminate the tedious SIM method RT updating process after column maintenance
280 C
6. Simplify the editing of the SIM ion groups A mixture from the California Department of Food and Agriculture (CDFA) of 80 pesticides at 5000 pg/L each was used as the stock solution for this study. The mixture contained carbamate, organochlorine, organophosphorus, and organonitrogen pesticides. Figure 1 is an offset overlay of three total ion chromatograms (TIC) with 50, 100, and 500 pg of each of the pesticides injected. These TICs were obtained in the Scan mode from 1-L spiltless injections. For many of these pesticides the quantitation limit in the Scan mode is about 500 pg on column.
Vent
Purge Liner
Inlet cooling
240000
50 pg on column
220000
180000
100 pg on column
140000
100000
60000
500 pg on column
20000
5.00
10.00
15.00
20.00
25.00
30.00
35.00
40.00
Figure 1.
Total ion chromatograms from 1-L splitless injections of 80 pesticides with 50, 100, and 500 pg of each compound injected.
SIM Mode To lower the detection limit, a SIM method was created. Instead of the traditional way of making a SIM method, a user can use the information in the RTL Pesticide Database to build a SIM method
without running an analysis. Here are the steps for editing SIM ion group parameters: 1. List the MSD RTL Pesticide Database from the ChemStation (Figure 2 is a partial listing) and paste the complete listing into a spreadsheet.
Figure 2.
A partial listing of the pesticide screener database. The listing includes the compound number, compound name, target ion, expected retention time, and three qualifier ions. 3
2. In the spreadsheet, delete the rows of the compounds not needed in the method. 3. Separate target compounds into groups (see the added Group # column on Figure 3) using these criteria: One to three compounds in each group, and The RTs of the adjacent compounds in adjacent groups are at least 0.2 minute apart. For example, compounds 42 and 51 are more than 0.2 minute apart, so they are in different groups. Compounds 51 and 55 are less than 0.2 minute apart, so they are in the same group. 4. Use the average RT of the adjacent compounds in adjacent groups as the SIM group RT (see the added Group RT column on Figure 3). For example, the average retention time of compound 42 (7.91 min, in group 2) and compound 51 (8.78 min, in group 3) is 8.35 minute which is used as the starting retention time of group 3. When all the group numbers and respective starting retention times are determined, make a hardcopy of the spreadsheet for easy entry into the MS SIM/Scan Parameters in the next step. 5. Enter the target ion and qualifier ion(s) (Q1, Q2, and/or Q3) of all compounds into the respective ChemStation SIM group (Figure 4). Notice that all the information for building the SIM groups came from Figure 3.
# 24 35 42 51 55 76 82 92 98 102 103 104 111 113 117 120 122 124 129 Compound Name 2,6-Dichlorobenzonitrile Mevinphos Propham o-Phenylphenol Pentachlorobenzene Propoxur Diphenylamine Chlorpropham Ethalfluralin Bendiocarb Trifluralin Benfluralin Phorate BHC alpha isomer Hexachlorobenzene Dicloran Demeton-S Dimethoate Simazine MSD_RT 6.75 7.60 7.91 8.78 8.95 10.35 10.52 11.05 11.28 11.54 11.64 11.73 11.96 12.09 12.38 12.56 12.63 12.68 12.91 T 171 127 93 170 250 110 169 127 276 151 306 292 75 181 284 206 88 87 201 Q1 173 192 179 169 252 152 168 213 316 126 264 264 121 219 286 176 60 93 186 Group # Group RT 1 3.00 2 3 4 5 6 7.75 8.35 9.60 10.76 11.41
The number of qualifier ions used in a SIM method depends on the number of analytes of interest. For a method monitoring 20 to 30 compounds, all three qualifier ions should be used in the SIM method. As the list of target compounds grows, fewer qualifier ions should be used in the method to maintain a reasonable and comparable ion dwell time and sampling rate. In general, 10 scans (cycles) per peak are recommended for quantitation purposes. For example, if an analyte peak is 6 seconds wide, about 1.7 cycles per second should be maintained for that SIM ion group. Once the number of cycles per second is determined, the dwell time of the ions can be varied to meet that. As the dwell time is entered for each ion, the ChemStation automatically shows the number of cycles per second. In Figure 4, Group 6 has 3.03 cycles per second.
Figure 4.
A screen capture of the MSD ChemStation showing the MS and SIM parameters. The SIM parameters (group ID, group retention time, and ions) were all derived from Figure 3.
Figure 3.
A spreadsheet of target compounds separated into different SIM groups with RTs of the adjacent compounds in adjacent groups at least 0.2 minute apart. The starting retention time of each group was determined by calculating the average RT of the adjacent compounds in adjacent groups.
Figure 5 shows two chromatograms obtained from 1-L splitless injections at 50 pg/L using both Scan and SIM modes. The Scan mode has significantly higher baseline noise than the SIM mode. Some of the compounds, especially the late eluters, were not detected in the Scan mode. When the Scan method was changed to a SIM method at this concentration, the signal-to-noise ratio (S/N) increased by a factor of 100. It is worth pointing out that a SIM method does not record background ions from the sample matrix, therefore minimizing the baseline noise and improving the S/N.
21000 19000 17000 15000 13000 11000 9000 7000 5000 3000 1000 5.00 10.00 15.00 20.00 25.00 30.00 Injection volume: 1 L Concentration: 50 pg/L PTV mode: Splitless
Scan
SIM
35.00 40.00
Figure 5.
Chromatograms of 1-L splitless injections at 50 pg/L from Scan and SIM modes.
In a SIM method, the retention times of the ion groups normally need updating after column maintenance. By using RTL, a user can not only eliminate the tedious RT updating process [4] but also decrease the detection limit. With reproducible known RTs of target compounds, the start and end time of each ion group can be determined optimally. By narrowing the time windows of an ion group to monitor only one or two compounds at a time, the MS can monitor fewer ions in each window, allowing more sampling time for the target ions. Ideally, a SIM method will have the maximum number of ion groups and the minimum number of ions in each group. In this way, each ion group can get more scans per unit time resulting in better peak shape and more accurate quantitation.
LVIs To decrease the detection limit further, a user can put more sample on column using the LVI technique. The typical solvent-vent approach is to inject the sample slowly into a PTV inlet at a temperature just below the solvent boiling point and let solvent evaporate before ramping up the inlet temperature to move the compounds onto the capillary column. Figure 6 compares a 1-L splitless injection with a 25-L solvent-vent injection. Both injections resulted in 50 pg per compound on column. Note that the solvent-vent ion chromatogram is plotted upside down for ease of comparison with the splitless ion chromatogram. It is obvious from the figure that the two techniques provide very similar results. This demonstrates that the solvent-vent technique is a viable approach for sample introduction.
7000 6000 5000 4000 3000 2000 1000 0 -1000 -2000 -3000 -4000 -5000 6.00 8.00
SIM
Injection volume: 1 L Concentration: 50 pg/L PTV mode: Splitless
Injection volume: 25 L Concentration: 2 pg/L PTV mode: Solvent vent
10.00
12.00
14.00
16.00
18.00
20.00
22.00
24.00
26.00
28.00
30.00
32.00
Figure 6.
SIM results of 50 pg on column using either a 1-L splitless or a 25-L solvent-vent injection.
Higher EMV It is known that the signal increases with higher EMV on the MS. In Figure 7, the upper signal, after 10-fold magnification, is a 25-L LVI of 0.5 pg/L at tune voltage. The bottom signal is the same injection with the electron multiplier set to tune +400 V. Adding 400 V to the EMV increases
the signal by 10X, which makes the integration more accurate. However, the baseline noise also increases by 10X, so the S/N stays the same. Although increasing the EMV does help to bring small peaks over the detection threshold, it shortens the life of the multiplier. In general, the EMV should be kept at the tune voltage.
40000 35000 30000 25000 20000 15000 10000 5000 0 -5000 -10000 -15000 -20000 -25000 -30000 -35000 6.00 8.00 10.00 12.00 14.00 16.00 18.00 20.00 22.00 24.00 26.00 28.00 30.00 32.00
Tune voltage
(Magnified 10X)
Injection volume: 25 L Concentration: 0.5 pg/L
Tune voltage + 400 V
Figure 7. SIM results of 12.5 pg on column using either EMV at Tune voltage or Tune +400 V.
LVIs in Combination with SIM Methods Combining LVI and SIM, Figures 8 and 9 show quantifiable peaks of three compounds at as low as 5 pg on column. In Figure 8, ion chromatograms of endosulfan sulfate and p,p-DDT at 0.2 and 500 pg/L are shown. The top chromatogram was from a 25-L solvent-vent SIM method and the bottom chromatogram was from a 1-L splitless Scan method. By using LVI and SIM, it is interesting to see that similar S/N ratios were achieved
even with a 2500-fold decrease (from 500 to 0.2 pg/L) in sample concentration. By increasing the injection volume to 100 L, samples at concentration as low as 0.05 pg/L can also be quantified as shown in Figure 9. The top portion shows the chlorthal-dimethyl extracted ion chromatograms (EIC) of mass 299 and 301 from a 100-L full Scan run. The bottom portion shows the same ions from a 100-L SIM run. The SIM method shows better peak shape and lower baseline noise.
280 250 220 190 160
Injection volume: Concentration:
25 L 0.2 pg/L (5 pg on column)
SIM
30000 24000 18000 12000 6000 26.30
Injection volume: Concentration:
1 L 500 pg/L (500 pg on column)
p,p-DDT
Scan
Endosulfan sulfate
26.40
26.50
26.60
26.70
26.80
26.90
27.00
27.10
27.20
27.30
Figure 8.
Ion chromatograms of endosulfan sulfate and p,p-DDT at 0.2 and 500 pg/L. The top chromatogram was from a 25-L solvent-vent SIM method and the bottom chromatogram was from a 1-L splitless Scan method.
1400 1200 1000 800 600 400 200 0 1400 1200 1000 800 600 400 200 0 18.6 18.8 19.0 19.2 19.4 19.6 19.8 20.0 20.2 20.4
Ion 301
Scan
Concentration: 0.05 pg/L
Ion 299
1000 800 600 400 200 0 1000 800 600 400 200 0 18.6 19.0 19.2
SIM
Ion 301
Concentration: 0.05 pg/L
Ion 299 *SIM group start time
*
19.4 19.6
*
19.8 20.0 20.2
Figure 9.
Ion chromatograms of 100-L chlorthal-dimethyl injected at 0.05 pg/L. The top portion was from a full Scan run and the bottom portion was from a SIM run.
Target Compound Screening Combing RTL and the G1049A MSD RTL Pesticide Database/Library, a user can screen for 567 pesticides and suspected endocrine disrupters from any Scan run [5]. A user can screen a subset of the library with improved sensitivity using a SIM method. The MSD ChemStation can generate a
567-compound screening report automatically in less than 30 seconds. Figure 10 is a report of the 0.5 pg/L sample (25 L injected in SIM mode) that lists the probable hits (marked with an x) and possible hits (marked with a ?). All target compounds at this 12.5 pg on column level were found by the software.
Figure 10. Typical report from the GC/MS pesticide screener showing probable "hits" (marked with an x) and possible hits (marked with a ?). Other information includes the library retention time followed by the RT difference in this chromatogram, the target ion, its abundance, out of range qualifier(s), and a cross correlation value with the library spectrum.
Conclusions
Using the information (compound names, retention times, and ion masses) in the RTL pesticide database, a SIM method of 80 target compounds can be created in less than 2 hours without running any analyses. The examples show that both LVI and SIM are effective techniques to decrease the quantitation limit of target compounds from sub-ppm to ppt. Any lab can decrease the quantitation limit by a factor of 100 without any hardware modification. Lowering the quantitation limit from 500 pg down to 5 pg on column can be done using a SIM method and RTL. By adding LVI to the system, target compounds in femtogram/L can be quantified.
Acknowledgement
The author would like to acknowledge Alex Chung and Mark Lee at CDFA for providing the pesticide mixture used in this study.
References
1. Vince Giarrocco, Bruce Quimby, and Matthew S. Klee, Retention Time Locking: Concepts and Applications, Agilent Technologies Application note, 5966-2469E, printed March 2000, www.agilent.com/chem 2. Bill Wilson, Philip L. Wylie, and Matthew S. Klee, Large Volume Injection for Gas Chromatography Using a PTV Inlet, Agilent Technologies Application note, 5965-7770E, printed February 2000, www.agilent.com/chem
3. Philip L. Wylie, Trace Level Pesticide Analysis by GC/MS Using Large-Volume Injection, Agilent Technologies Application note, 5966-1214E, printed April 2000, www.agilent.com/chem 4. Prest, H. and P. Cormia, Retention Time Locking: Advantages in GC/MS SIM Analysis, Agilent Technologies Application Note, 5968-3797E, printed December 1999, www.agilent.com/chem 5. Harry Prest, Philip L. Wylie, Ken Weiner, and Doug Agnew, Efficient Screening for Pesticides and Endocrine Disrupters Using the 6890/5973 GC/MSD System, Agilent Technologies Application note, 5968-4884E, printed December 1999, www.agilent.com/chem

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2001 Printed in the USA November 14, 2001 5988-4392EN
GC/MS Approaches to the Analysis of Monochloropropanediol Application
Foods and Flavors
Author
Harry Prest Agilent Technologies, Inc. 1601 California Avenue Palo Alto, California 94301-1111 USA
monochloropropanediol byproduct was classified by the European Union's Scientific Committee for Food as a suspected carcinogen [1]. Although efforts were made to reduce the presence of 3-MCPD, continuing concerns about its presence lead to regulation of the allowable concentration. Recently the Association of Official Analytical Chemists (AOAC) has published a method for the extraction, separation and identification of 3-MCPD in foods and ingredients using gas chromatography with mass spectrometric detection [2]. In brief, a homogenized sample is mixed with a salt solution, then mixed with an Extrelut refill pack before being added to chromatographic column. The 3-MCPD is eluted with diethyl ether and a portion is derivatized with heptafluorobutyrylimidazole. Quantitation with GC-MS using electron impact ionization provides detection limits less than 0.01 mg/kg (which is equivalent to about 10 pg/L in the final extract at injection). This brief examines approaches to 3-MCPD as the heptafluorobutyryl-derivative described in the AOAC method using the Agilent 5973N MSD.
Abstract
The suspected carcinogen 3-chloro-1,2-propanediol (3-MCPD) is found in hydrolyzed vegetable protein, a widely used flavoring. Gas chromatography with mass spectrometric detection is a standard AOAC method. Electron impact ionization permits subpicogram measurement. Electron capture negative ionization is more selective and probably better suited to actual samples, with sensitivity of a few picograms in scan mode and less than 1 picogram in the selected ion mode.
Introduction
Hydrolyzed vegetable protein (HVP) is a widely used flavoring found in soups, sauces, and some meat products, etc. HVP is traded internationally both as solid and liquid depending upon the intended application. During the acid hydrolysis process of vegetable proteins, the hydrochloric acid agent reacts with triglycerides (Equation 1) to produce 3-chloro-1,2-propanediol (3-MCPD). This
O O R O O O O R" R' HCl OH OH OH HCl Cl OH OH
Experimental
3-MCPD liquid (Sigma Scientific, St. Louis, MO.) was diluted in dichloromethane (VWR Scientific, San Francisco, CA). An aliquot was added to a reaction vial containing 1 mL isooctane and derivatized with heptafluorobutyrylimidazole (Pierce, Rockford, IL) at 70 C for 30 minutes according to the procedure outlined in the AOAC method [2].
Equation 1. The acid hydrolysis of vegetable matter from triglycerides to glycerol, and then to 3-MCPD.

Derivatizing 3-MCPD with heptafluorobutyrylimidazole replaces the hydrogens on the diol groups with ester linkages to a perfluorinated propyl side chain. The molecular formula of the derivative is C3H5O2Cl(COC3F7)2, and Figure 1 shows the molecular structure. Figure 2 shows the electron impact (EI) ionization mass spectrum of the heptafluorobutyrylimidazole derivative of 3-MCDP from 60 to 510 m/z.
F
197 m/z
275 m/z
F F F F F O 169 m/z F
F O F F O 453 m/z 289 m/z -HCl 253 m/z Cl
Figure 1. The molecular structure and a suggested fragmentation pattern for heptafluorobutyrylimidazole derivative of 3-MCPD in electron impact ionization.
Abundance 100 90 80 70 60 50 40 30 20 10 0 100 150 200 250 300 350 400 450 500 m/z
169
69 197
253 289 100 453
Figure 2. Electron impact ionization mass spectrum of the 3-MCPD heptafluorobutyryl derivative at 70 eV for the 60 to 510 m/z mass range. The molecular ion [M]+ would be expected at 502 m/z.
Electron impact ionization produces a mass spectrum that lacks a molecular ion and has a base peak at m/z 169 from [C3F7]+ fragments. As seen from the fragmentation diagram of Figure 1, the ions at 169 m/z and 197 m/z contain no structural relevance to 3-MCPD and consequently can not be used to indicate 3-MCPD. This suggests the 453, 289, 275, and 253 m/z fragments, which contain 3-MCPD structure, be used for detection. In the AOAC collaborative study, several laboratories had difficultly detecting the 453 m/z fragment. This is not a problem for the 5973N due to the high-energy dynode arrangement and high transmission quadrupole which provide good signal for high molecular weight fragments. In fact, work with the standard showed good signal-to-noise for the 453 m/z ion in the scan mode even at only a few picograms injected. Selected ion monitoring using the four ions suggests detection at subpicogram levels is possible. In actual samples, matrix interferences may emerge and contribute to noise. However, the derivatization technique suggests applying electron capture negative ionization (ECNI) mass
spectrometry which provides more selective ionization than electron impact. Figure 3 shows the ECNI mass spectrum of the 3-MCPD derivative under standard conditions with methane buffer gas (tat is, source 150 C, methane at 2 mL/min). Unlike the EI results, the molecular ion (502 m/z) is detected, although at low relative intensity. Unfortunately like EI, the base peak at 213 m/z and next most abundant peak at 194 m/z, due to [OCOC3F7]+ and [OCOC2F7]+ fragments respectively, also contain no structural relationship to 3-MCPD. This leaves the 502, 482 and 446 m/z ions as good candidates for 3-MCPD detection and quantitation. Analysis of standards suggests that it would be possible to detect a few picograms in the scanning mode and less than one picogram in selected ion mode (SIM). Further optimization of ECNI is possible, such as a lower source temperature to take advantage of the low boiling point of the derivative, which may improve the spectrum and detection limits. The real advantage of ECNI is expected to be in typical food samples where the greater selectivity of ECNI will demonstrate a strong suppression of chemical noise and enhance method detection limits.
Abundance 100 90 80 70 60 50 40 30 20 10 0 160 180 200 220 240 260 280 300 320 340 360 380 400 420 440 460 480 500 m/z-->
213
194 482 179 233 446 502
Figure 3. Electron capture negative ion chemical ionization mass spectrum of derivatized 3-MCPD with methane buffer gas from 150 to 510 m/z. Note the presence of the molecular anion [M]- not seen in EI.
Acknowledgments
The author is grateful to Phil Wylie and Norman Low for their contributions to the manuscript.
References
1. Reports of the Scientific Committee for Food, Food Sciences and Techniques, 36th Series, European Commission, Luxembourg, Belgium. Brereton, P., et al., Determination of 3-chloro1,2-propanediol in foods and food ingredients by gas chromatography with mass spectrometric detection: Collaborative study. Journal of the AOAC International, 2001. 84(2): p. 455-465.
2.

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance or use of this material. Information, descriptions and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2001 Printed in the U.S.A. November 8, 2001 5988-4287EN
SPE of Chlorinated Pesticides in Water
Application Brief
Condition: Add 5 mL acetone to the cartridge. Apply vacuum and discard the eluant. Repeat with 5 mL methanol then 5 mL water. Do not allow the sorbent to go dry at any point during this step. Load: Attach a sample reservoir to the top of the cartridge. Add 0.2 mL methanol to 20 mL of the water sample.1 Mix, then add to the cartridge. Apply the vacuum and discard the eluant. The flow rate should be no greater than 10 mL/min.2 Rinse: Add 3 mL water to the cartridge. Apply vacuum and discard the eluant. Leave the vacuum on for 30 seconds after all of the water has passed through the cartridge. Centrifuge the cartridge at 10001500 rpm for 5 minutes.3
Elution: Place a collection tube beneath the cartridge. Add 3 mL of acetone to the cartridge. Apply vacuum and collect the eluant. Concentrate to dryness under a stream of dry nitrogen4 (transfer to a small vial may be necessary). Dissolve the residue in 200 L acetone. Inject 2L into the GC.
1
Equipment
AccuBOND II ODS (C18) 6 mL/500 mg cartridge (P/N 188-1356) 25 mL sample reservoir (P/N 700-4007) coupling fitting (P/N 5185-5794) vacuum manifold (P/N 5185-5754, 10-port) (P/N 5185-5765, 20-port)
Volume up to 100 mL may be used. Adjust the amount of added methanol so that the final concentration is 1%. Using slower flow rates will result in slightly better recovery values. Centrifuging removes additional water which aids in sample concentration. The use of heat to aid in sample concentration may result in reduced recovery values.
Reagents
water (HPLC grade) methanol (pesticide grade) acetone (pesticide grade)
Compound
Trifluralin Chloroneb Propachlor Hexachlorobenzene (HCB) -BHC -BHC (lindane) -BHC Heptachlor Chlorothalonil -BHC Alachlor Aldrin DCPA Heptachlor epoxide o,p'-DDE Endosulfan 1 p,p'-DDE Dieldrin o,p'-DDD Chlorobenzilate Endrin p,p'-DDD o,p'-DDT Endosulfan II p,p'-DDT Endrin aldehyde Endosulfan sulfate Methoxychlor
Water spiked at 50 ppb n=4
k
5.52 5.63 6.65 7.10 7.43 8.16 8.26 8.75 8.76 8.93 9.03 9.38 9.80 10.86 11.49 11.55 12.15 12.52 13.12 13.35 13.85 14.17 14.25 14.48 15.27 15.64 16.07 18.50
x
90 90 87 69 73 49 83 62 86 97 83 67 66 72 66 81 88 86 68 86 97 80 64 91 79 85 90 92
Std. Dev.
12 6 7 13 3 18 4 7 2 4 5 5 3 7 11 9 5 11 9 8 12 8 14 7 6 3 7 11
Pond Water Extract Containing Chlorinated Pesticides

DB-608 30 m x 0.53 mm I.D. P/N: 125-1730 Film Thickness: 0.83 m Carrier: Helium at 40 cm/sec (measured at 140C) Oven: 140C for 2 min 140-240C at 10/min 240C for 5 min 240-265C at 5/min 265C for 5 min Injector: Megabore direct, 250C Detector: ECD, 325C Nitrogen makeup gas at 30 mL/min Column: 1. Hexachlorobenzene 2. a-BHC 3. g-BHC 4. b-BHC 5. Heptachlor 6. d-BHC 7. Aldrin 8. Heptachlor epoxide 9. Endosulfan 1 10. p,p'-DDE 11. Dieldrin 12. Endrin 13. p,p'-DDD 14. Endosulfan II 15. p,p'-DDT 16. Endrin aldehyde 17. Endosulfan sulfate 18. Methoxychlor
Pond Water Extract Containing Chlorinated Pesticides

Column: DB-608 30 m x 0.53 mm I.D. P/N: 125-1730 Film Thickness: 0.83 m Carrier: Helium at 40 cm/sec (measured at 140C) Oven: 140C for 2 min 140-240C at 10/min 240C for 5 min 240-265C at 5/min 265C for 5 min Injector: Megabore direct, 250C Detector: ECD, 325C Nitrogen makeup gas at 30 mL/min 1. Trifluralin 2. Chloroneb 3. Propachlor 4. Hexachlorobenzene 5. Chlorothalonil 6. Alachlor 7. DCPA 8. o,p'-DDE 9. o,p'-DDD 10. Chlorobenzilate 11. o,p'-DDT
k = partition ratio (a measure of retention) x = % recovery To order or for technical assistance, contact your local authorized Agilent distributor, or visit our website at www.agilent.com/chem.
Agilent Technologies, Inc. 2001 Information, descriptions and specifications in this publication are subject to change without notice. All rights reserved. Reproduction, adaptation, or translation without prior written permission is prohibited, except as allowed under copyright laws. Printed in the U.S.A. October 25, 2001 5988-4057EN
The Analysis of Organophosphate Pesticides by LC/MS Application
LC-MS
Authors
Paul Zavitsanos Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA Paul Yang MOE Ontario Canada Lorna Grey MOE Ontario Canada
Additionally, LC-MS provides unequivocal identification of each pesticide, even if the pesticide was not completely resolved from neighboring eluants. Traditional UV detection cannot provide the required specificity because many of the pesticides within the same class exhibit similar UV spectra.
Sample case
A mixture of organophosphate pesticides and an internal standard were analyzed using an Agilent 1100 LC/MS with an ESI source (Table 1).
Table 1. Mixture of Organophosphate Pesticides
Abstract
Organophosphate pesticides were readily analyzed using liquid chromatography-mass spectrometry with electrospray ion source. Sensitivity and selectivity were significantly better than using a diode-array UV detector.
Elution order 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Compound Mevinphos isomer 1 Dimethoate Mevinphos isomer 2 Dichlorvos Azinphos methyl Parathion methyl Malathion Diazinon Triphenyl orthophosphate* Parathion ethyl Phorate Reldan Ronnel Terbuphos Dursban Ethion Temephos
[M+H]+ 225 230 225 221 318 264 331 305 327 292 261 322 321 289 350 385 467
Concentration g/mL 0.2 0.5 0.5 0.5 0.05 0.2 0.5 0.2 1.0 0.1 0.1 0.5 0.1 0.2 0.1 0.2 0.1
Overview
Liquid chromatography-mass spectrometry (LC-MS) is rapidly becoming a routine technique for efficient trace analysis of polar pesticides in various types of samples. In comparison to existing methodologies, such as gas chromatography-mass spectrometry (GC-MS) and ultraviolet (UV) detection, LC-MS considerably simplifies cleanup procedures, reducing both time of analysis and method development time.1 LC-MS with an electrospray ion (ESI) source avoids the thermal degradation of labile pesticides encountered with GC and eliminates the need for preliminary derivatization to increase compound volatility.
* Internal standard
Method summary
Column 2.1 mm id 5 cm long, filled with
Results
Simultaneous UV (220 nm) and MS detector outputs are compared in Figure 1. The MS plot is a composite of all the individual extracted ion chromatograms. Each was obtained at the [M+H]+ value given in Table 1, and are separated and stacked in Figure 2 for easy comparison.
3.5 m particles, C18 chemistry

20 mM ammonium acetate vs. acetonitrile
mobile phase gradient 5 % to 95 % acetonitrile in 4 minutes Hold 2 minutes

Splitless 400 L/min flow 3 L injection volume Scan data 120 to 600 m/z
SIM data as per Table 1. 95 msec dwell/ion in two groups
5 0 -5 -10 -15 -20 -25 0 2 4 6 8 Normalized UV Plot 10 12 14 min.
7000000 6000000 5000000 4000000 3000000 2000000 1000000 0 2 4 6 8 10 12 14 min.
MS selected ion mode (SIM), overlayed and normalized
Figure 1. Comparison of UV and MS chromatograms.
17 16
m/z 467 385
15 350 14 289 13 321 12 322 11 261 10 292 9 327 8 305 7 331 6 264 5 348 4 221 2 230 1
2
3 225
4 6 Minutes 8 10
Figure 2. Stacked normalized extracted ion chromatograms for compounds 1 through 17.
Figures 3 through 6 show the resulting normalized mass ion spectra for each compound included in Table 1.
Mevinphos isomer 1
193.0
[M+H]+
225.0
Dimethoate
[M+H]+
232.0
230.0
226.0
[M+H]+
192.9
222.9
Dichlorvos
[M+H]+
220.9
237.9
100
150
200
224.8
250
[M+NH4]+
239.9
242.9
[M+NH4]+
Mevinphos isomer 2
225.0
242.1
283.0
300
m/z
Figure 3. Stacked normalized ion mass spectra for compounds 1 through 4.
Diazinon
Malathion
Azinphos methyl
Parathion methyl
100
118.9 133.1 150.0 160.0
132.0
150
200
235.4
250

[M+H]+ 265.1 285.0 [M+H]+ 305.1 291.4 Background 263.9 260.9 290.1 291.5 290.4 [M+H]+ [M+H]+ 332.0 348.1 331.0 340.1 318.8 317.9
300
306.1
m/z
Triphenyl orthophosphate
[M+H]+
327.0
Parathion ethyl
[M+H]+
292.1
235.1
[M+H]+
Phorate
263.0
261.0
293.0
304.9
[M+H]+
Reldan
325.8
100 150 200 250 300
321.9 323.9
324.9
328.0
m/z
[M+H]+ 324.9 101.9
322.9 232.9 253.3 244.8 186.9 198.9 234.9 [M+H]+

150 200 250
Ronnel
320.8
Terbuphos
289.0
100
291.0
98.0
300
324.0
339.1
m/z
350.9
Dursban
[M+H]+
353.9
351.9
[M+H]+
Ethion
Temephos
[M+NH4]+ 386.9 467.0
400 425 450
325
350
375
[M+H]+
475
485.0
484.0
m/z
Figure 6. Stacked normalized ion mass spectra for compounds 13 through 17. 7
www.agilent.com Conclusions
When determining organophosphate pesticides using LC-MS with an ESI source:
All the tested organophosphate pesticides
ionized well and gave definite [M+H]+ ions

Sensitivity and selectivity are significantly
better than using diode-array UV detector

Overall chromatography and analysis is simple
and straightforward
Positive identification and quantification are
performed using integrated software For more information, contact your local Agilent sales representative, or visit www.agilent.com.
Reference
1. Elbert Hogendoorn and Peit van Zoonen, Journal of Chromatography A, 892 (2000) 435-453.
Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Copyright 2001 Agilent Technologies, Inc. Printed in the USA September 13, 2001 5988-3774EN
A New Approach to the Analysis of Phthalate Esters by GC/MS Application

Gas Chromatography/Mass Spectrometry March 2001
Authors
Cameron George Agilent Technologies, Inc. 91 Blue Ravine Road Folsom, California 95630-4714 USA Harry Prest Agilent Technologies, Inc. 1601 California Avenue Palo Alto, California 94304-1111 USA
Introduction
The widespread use and manufacture of plastics have made the phthalate esters one of the most ubiquitous classes of compounds in our everyday environment. These plasticizers increase polymer flexibility due to their function as intermolecular lubricants. Because they are additives and not reagents, they are not chemically bound in the polymer and are available to leach from the matrix. Phthalates are also components of cosmetics, detergents, building products (flooring, sheeting, films), lubricating oils, PCB substitutes, carriers in pesticide formulations and solvents. Consequently, the potential for human exposure is very high. Toxicological studies have linked some of these compounds to liver and kidney damage, and to possible testicular or reproductive-tract birth defect problems, characterizing them as endocrine disruptors. Scientists at the U.S. Centers for Disease Control have, for the first time, documented human exposure to phthalates by determinations of the monoester metabolites in human urine [1]. Their work leads to the conclusion that phthalate exposure is both higher and more common than previously suspected. Of particular concern were the significantly higher concentrations of the dibutyl phthalate metabolite in urine of women of childbearing age (20-40 years) than in other portions of the population. The presence of phthalate esters in polyvinyl chloride (PVC) toys has generated the most
Abstract
A new instrumental method for the determination of 29 phthalate esters, including six recently banned from baby toys by the European Union, using positive chemical ionization and retention-time locking is described. Positive chemical ionization provides a high degree of selective ionization for the phthalates, primarily producing spectra in which the protonated molecule (M+1) is the base peak. This provides easy discrimination among the phthalates on the basis of their molecular weight, while retentiontime locking increases confidence in the identification of the various isomers. In this approach, both pure compounds and technical mixtures are considered. Although this work focuses on the more commonly used 1,2-substituted esters, the 1,3-isomers and 1,4-isomers are also characterized. The combination of positive chemical ionization and retention-time locking makes the method rugged, durable and applicable to a wide variety of matrices.
controversy. While regulators in Greece have completely banned soft PVC toys, Austria, Denmark, Finland, France, Germany, Norway and Sweden have unilaterally banned phthalates in PVC toys for children under three years old. In December of 1999, the European Union (EU), concerned with a serious and immediate risk to children, placed an emergency ban on six of the phthalate esters in soft PVC toys and childcare products meant to be placed in the mouths of children under the age of three [2]. None of the six banned phthalates may exceed 0.1% by weight. These heightened concerns suggest the need for an improved method of detecting and characterizing phthalate esters which is applicable to a wide variety of matrices. This application note describes such an analytical method.
O R O O R' O
O R
Phthalate Structure and Mass Spectra

The three primary structures of phthalates are shown in Figure 1. Although there are three possible positions for the ester linkages, the most commonly used phthalates are based on the 1,2-benzenedicarboxylic acid structure (top). There are an infinite number of possible alkyl side chains, (R) and an infinite number of combinations of the side groups (R and R'). For example, the diisononyl phthalate consists of an array of compounds due to the isomeric branched-chain alkyl groups on both side chains. For phthalate esters with saturated alkyl side chains (without oxygen), the most intense peak in the electron impact (EI) ionization mass spectrum at 70 eV is always at m/z 149 due to the rapid formation and stability of the ion shown in Figure 2. (The only exception is R=R'=CH3 where the base peak is at m/z 163).
O R' O
O R
R' O O
Figure 1. Phthalic ester (top) or the 1,2-benzenedicarboxylic acid ester, isophthalic ester (middle) or the 1,3-benzenedicarboxylic acid ester, and terephthalic ester (bottom) or the 1,4-benzenedicarboxylic acid ester. R and R' represent alkyl side chains which may be branched and contain oxygen.
OH
Figure 2. The most abundant ion in the mass spectra of the phthalate esters with saturated alkyl side chains; m/z 149. The exception is for dimethyl phthalate where both R and R' are CH3 and so the H on the oxygen is replaced by CH3 and consequently m/z 163 becomes the base peak.
Invariably, the molecular ion is very weak or altogether absent; other fragments that provide information on the phthalate identity are also of very low abundance. As an example, consider the EI mass spectrum of dibutyl phthalate, one of the six banned by the EU, and bis(4-methyl-2-pentyl) phthalate in Figure 3. Identifying fragments have relative intensities of less than 10%. Gas chromatography provides some separation of the phthalates, but with the array of possible isomers and essentially a single identifying ion (i.e., m/z 149), distinguishing the individual phthalates of concern is difficult. More confident identification of the phthalates is possible using chemical ionization mass spectrometry in conjunction with retention-time locking (RTL).
1000000 900000 800000 700000 Abundance 600000 500000 400000 300000 200000 100000 0 40 60 80 100 120
149 O O O O
57
76
104
121
140
167
160
187
180
205
200
223 250
220 240 260
278
280
296 313
300 320
341
340
149 800000
O
700000 600000 Abundance 500000 400000 300000 200000 100000 55 0 40 60 76 80 104 100 121 120 140 167 160 189 180 m/z 208 200 233 220 240 260 251 278 280 306 300 334 320 340
O O O
Figure 3. Electron impact ionization mass spectra of di-n-butyl phthalate (upper panel) and bis(4-methyl-2-pentyl) phthalate (lower panel) from m/z 50 to 350 at 70 eV. Notice the lack of intense fragments and molecular ions. The molecular weights are 278 and 334 g/mole, respectively.
Retention-time locking allows compound retention times achieved on any one Agilent 6890 gas chromatograph (GC) to be replicated to within a few seconds on any other Agilent 6890 gas chromatograph (GC) applying the same GC method [3-5]. RTL is a powerful approach to compound identification. RTL allows the creation of compound acquisition methods and quantitation databases that can be reproduced in any laboratory, anywhere, because a compound can have a universally fixed and reproducible retention time. It is important that RTL be applied in conjunction with the appropriate detection scheme and sample reparation methods. Chemical ionization provides a more selective form of ionization than electron impact [6]. By judicious choice of the reagent gases, the degree of compound fragmentation can be controlled to a certain extent. In positive chemical ionization, methane reagent gas usually provides more fragmentation than gases of higher proton affinity such as ammonia. Less fragmentation would be helpful in identifying the phthalates. Instead of all phthalates generating a single, similar ion, positive ionization can provide phthalate ester molecular weights.
Injection Parameters Injection Mode Injection Port Temperature Pulse Pressure & Time Purge Flow & Time Gas Saver Flow & Time Oven Parameters Temperature Program 50.00C/min 15.00C/min Oven Equilibrium Time MSD Transfer Line Temp Column Parameters GC column (122-5532) Initial Flow & Mode Detector & Outlet Pressure Mass Spectrometer Parameters Tune Parameters Electron Multiplier Voltage Solvent Delay Scan Parameters Quadrupole Temperature Source Temperature Ammonia Gas Flow (MFC setting) Miscellaneous Parts Septa Liner GC column ferrule MSD interface ferrule
Pulsed Splitless 300C 25.0 psi 20.0 mL/min 20.0 mL/min
1.00 min 3.00 min 3.00 min
80C 200C 350C 0.25 min 325C
1.00 min 0.00 min 2.00 min
DB-5MS 30 m; 0.25 mm i.d.; 0.25 m film 1.2 mL/min Constant Flow MSD Vacuum
PCI Autotune (NH3) Autotune + 400V 4.00 min 194 - 550 m/z 150C 250C 0.5 mL/min (10%)
Experimental
Phthalate esters were obtained from Ultra Scientific (North Kingstown, RI), AccuStandard (New Haven, CT), and ChemServices (West Chester, PA) as neat compounds and mixtures. Dilutions were made in isooctane (Burdick and Jackson Grade, VWR Scientific). The configuration and operating parameters of the Agilent 6890Plus GC (standard 120V or faster ramping 220V), 7683 Automatic Liquid Sampler and 5973N MSD with CI option used for acquiring the data are given in the following tables. PCI reagent gas purities were 99.99% or higher.
5182-0739 5062-3587 5181-3323 5082-3508
BTO septa (400C) Deactivated 4 mm i.d. single taper 250 m Vespel 0.4 mm i.d. graphitized Vespel

As expected using methane as the reagent gas, the PCI mass spectra of the phthalates show ions corresponding to the protonated molecule [M+H]+ and adducts [M+C2H5]+ and [M+C3H5]+. Because of the relatively vigorous fragmentation produced by methane, the spectra of the dialkyl phthalate esters still resemble that produced in EI. In most cases, the fragment at m/z 149 is the base peak, however ions at m/z M+1, M+29 and M+41 are relatively intense with [M+H]+ from 10% to 30% (Figure 5). The dialkyl phthalate spectra also show a fragment corresponding to loss of one of the alkyloxy side chains to produce an ion shown in Figure 4. This is the most intense fragment for the dimethyl and diethyl phthalates and for the dibutyl and dipentyl (diamyl) phthalates, about 75% of the 149 base peak. As the length of ester alkyl chain increases, the intensity of this fragment decreases. (Apparently, in the dialkyl isophthalates, loss of the alkyl side chain not accompanied by the oxygen may be a preferred route.) Although positive chemical ionization with methane provides more information than EI on phthalate identity, the methane reagent is still rather unselective in ionization and will produce more chemical noise in the background, complicating identification in complex matrices.
OR
O
Figure 4. One of the most intense fragments in the methane PCI spectra of the phthalate esters is formed by loss of one of the alkyloxy side groups.
Applying ammonia as the reagent gas in PCI to reduce chemical noise and enhance identification of the phthalates is a more useful approach. The relatively gentle ionization produces protonation of the dialkyl phthalates, with m/z M+1 the base peak in their spectra. When combined with retentiontime locking, identification of phthalates becomes further simplified. Compare the spectra of the di-n-butyl phthalate acquired using methane versus ammonia as the reagent gas (Figure 5). The protonated molecule is the single dominant peak in the ammonia PCI mass spectrum of the di-n-butyl phthalate, and the adduct at m/z 296 ([M+NH4]+) is relatively small.
149
280000 260000 240000 220000 200000 180000 Abundance 160000 140000 120000 100000 80000 60000 40000 20000 0 80
O O O 205 177 O 205
177 123 84 105

100 120
279 189 223

200 220 240
135
140
163
160
251 263
260 280
307
300
319 333
320
180
1300000 1200000 1100000 1000000 900000 800000 Abundance 700000 600000 500000 400000 300000 200000 100000 0 80 100 120 140 160 180 200 m/z 220 240 260
279
94
108
122
136 148
166 177
194 205
222
240 252
280
296
300 320
Figure 5. PCI methane (upper panel) and ammonia (lower panel) mass spectra of di-n-butyl phthalate. The PCI methane mass spectrum shows substantial fragmentation but relative to the EI spectrum in Figure 3, high abundance for the higher m/z ions such as the protonated molecule at m/z 279. The ion at m/z 205 is generated by loss of an oxybutyl fragment; a process described in Figure 4. The PCI-ammonia mass spectrum consists almost completely of the protonated molecule.
This implies an easy method for identification. Whereas the EI spectra of the phthalates most frequently result in a base peak at m/z 149, the dialkyl phthalate PCI-ammonia spectra have base peaks at m/z = M+1. All dialkyl phthalates molecular formulas can be expressed as C8H6O4(CH2)y (CH2)x. These phthalates have (nominal) molecular masses given by M = 166 + (x + y)14, or M = 166 + (w)28, where x and y are the side chain lengths, and the second formula applies to symmetrical side chains (i.e., x = y = w). For example, di-n-butyl phthalate has x = y = 4, and therefore a (nominal) molecular mass of 278 which produces m/z 279 as the base
peak. Interestingly, the PCI-ammonia spectra of the dialkyl isophthalates and terephthalates appear to have base peaks at m/z M+18 due to [M+NH4]+. Because of the greater steric access to the ester linkages, adduct formation may be preferred. Table 1 gives the phthalate names, CAS numbers, molecular formula, nominal molecular mass, base peak in the PCI-ammonia spectrum and the RTL elution times. These retention times are "locked" relative to diphenyl phthalate, which has been chosen as the RTL locking compound and locked to elute at 9.450 min. Notice that the branched chain isomers elute prior to their straight chain forms on this column phase.
Table 1. Phthalate compound names, Chemical Abstracts Services numbers (CAS), molecular weights (M. Wt.), molecular formulas, nominal base peak in the PCI-ammonia spectrum and retention time (RT) in minutes. Retention times are locked relative to diphenyl phthalate (9.450 min). Retention time ranges are given for the isoalkyl phthalate technical mixtures. Phthalates banned by the EU are indicated by an asterix*. Benzyl benzoate is included since it is used as a surrogate in U.S. Environmental Protection Agency Method 8061. Name dimethyl phthalate dimethyl isophthalate diethyl phthalate diethyl terephthalate benzyl benzoate diisobutyl phthalate di-n-butyl phthalate* bis(2-methoxyethyl) phthalate diamyl phthalate bis(2-ethoxyethyl) phthalate butyl benzyl phthalate* diphenyl phthalate diphenyl isophthalate dicyclohexyl phthalate bis(4-methyl-2-pentyl) phthalate diisohexyl phthalates dihexyl phthalate dibenzyl phthalate hexyl-2-ethylhexyl phthalate bis(2-n-butoxyethyl) phthalate bis(2-ethylhexyl) phthalate* di-n-octyl phthalate* dioctyl isophthalate diisononyl phthalates* dinonyl phthalate diisodecyl phthalates* didecyl phthalate diundecyl phthalate didodecyl phthalate ditridecyl phthalate CAS 131-11-3 1459-93-4 84-66-2 636-09-9 120-51-4 84-69-5 84-74-2 117-82-8 131-18-0 605-54-9 85-68-7 84-62-8 744-45-6 84-61-7 146-50-9 146-50-9 84-75-3 523-31-9 75673-16-4 117-83-9 117-81-7 117-84-0 137-89-3 28553-12-0 84-76-4 26761-40-0 84-77-5 3648-20-2 2432-90-8 119-06-2 M. Wt. 194 194 222 222 212 278 278 282 306 310 312 318 318 330 334 334 334 346 362 366 390 390 390 418 418 446 446 474 502 530 Molecular Formula C8H4O4(CH3)2 C8H4O4(CH3)2 C8H4O4(C2H5)2 C8H4O4(C2H5)2 C14H12O2 C8H4O4(C4H9)2 C8H4O4(C4H9)2 C8H4O4(C2H4OCH3)2 C8H4O4(C5H11)2 C8H4O4(C2H4OC2H5)2 C8H4O4(C4H9)(CH2C6H5) C8H4O4(C6H5)2 C8H4O4(C6H5)2 C8H4O4(C6H11)2 C8H4O4(CH3C5H10)2 C8H4O4(C6H13)2 C8H4O4(C6H13)2 C8H4O4(CH2C6H5)2 C8H4O4(C2H5C6H12)(C6H13) C8H4O4(C2H4OC4H9)2 C8H4O4(C2H5C6H12)2 C8H4O4(C8H17)2 C8H4O4(C8H17)2 C8H4O4(CH3C8H17)2 C8H4O4(C9H19)2 C8H4O4(CH3C9H18)2 C8H4O4(C10H21)2 C8H4O4(C11H23)2 C8H4O4(C12H25)2 C8H4O4(C13H27)2 Base Peak 195 212 223 240 230 279 279 283 307 311 313 319 319 331 335 335 335 347 363 367 391 391 408 419 419 447 447 475 503 531 RT (min) 4.32 4.54 4.81 5.06 5.62 5.95 6.40 6.57 6.94 7.13 8.42 9.45 10.30 9.32 6.93 7.55 - 8.28 8.34 10.51 8.84 8.98 9.32 10.28 10.84 9.40 - 11.10 11.19 10.16 - 11.86 12.05 12.87 13.65 12.01 - 13.81
Technical formulations of the isoalkyl phthalates tended to contain substantial amounts of the straight chain isomer, which may convolute quantitation as well as peaks that may be construed as originating from nonequivalent side chains i.e., x y in equation 1). These impurities can be detected as M14 around the mass of the nominal isomer. For example, technical grade diisononyl phthalate contains compounds that generate ions at m/z 391 (minor), 405, 433, and 447 in addition to the nominal diisononyl phthalate compound at m/z 419. The gentle ionization of ammonia reagent gas, the elution times and the study of the
spectra of other pure isomers, such as the dinonyl phthalate, suggest that these fragments are not formed by the PCI process but are due to these different alkyl side chain impurities (Figure 6). To demonstrate the utility of the PCI-ammonia compared to conventional EI analysis, consider the chromatograms presented in Figure 7. The EI spectra of the phthalates produce m/z 149 as the base peak for all the phthalates present; distinguishing ions are minor constituents (<10% relative intensity), making identification complicated. However, by examining the appropriate PCI-ammonia ions, the various phthalates are easily distinguished.
m/z 419
Abundance
m/z 433
m/z 405
m/z 447
9.40
9.60
9.80
10.00
10.20
10.40 Time
10.60
10.80
11.00
11.20
11.40
Figure 6. PCI-Ammonia extracted ion chromatogram of technical diisononyl phthalate. The diisononyl appears as the major component at m/z 419 while ions at m/z 405, 433, and 447 indicate alkyl chains shorter by one CH2 unit and longer by one and two CH2 units respectively.
EI TIC
Abundance 9.50
10.00
10.50
11.00
11.50
12.00
12.50
13.00
13.50
14.00
Abundance
m/z 149
9.50
10.00
10.50
11.00
11.50 di-n-nonyl
12.00
12.50
13.00
13.50
14.00
di-n-undecyl di-n-dodecyl
di-n-decyl
diisodecyl Abundance
diisononyl
ditridecyl
9.50
10.00
10.50
11.00
11.50
12.00 Time
12.50
13.00
13.50
14.00
Figure 7. Chromatograms of dinonyl, diisononyl, didecyl, diisodecyl, diundecyl, didodecyl, ditridecyl phthalate esters in EI (upper panel), EI as an extracted ion chromatogram at m/z 149 (middle panel), and PCI-extracted ion chromatogram with ions selected for the individual phthalate classes as given in Table 1. The EI information is insufficient to identify coeluting phthalates. For example, the dinonyl and diundecyl phthalates are "buried under" the signals from the isodecyl and ditridecyl phthalates.
Conclusions
Applying GC - electron impact (EI) mass spectrometry to the determination of phthalates requires full chromatographic separation. The EI spectra of the phthalates are distinguished only by ions of very low intensity. In EI, the phthalates produce a single common ion (m/z 149) as the most intense spectral peak, regardless of the alkyl side chain substitution. Applying tandem mass spectrometry (i.e., EI/MS/MS) gains nothing, because there is a common parent ion, and therefore any daughter ions would also be non-unique. However, the combination of positive chemical ionization with retention-time locking allows even complex mixtures of phthalates to be characterized. Ammonia reagent gas produces the protonated molecule as the base peak, which immediately allows the phthalates to be distinguished on the basis of their substitution. PCI is also an advantage in complex matrices, where the non selective ionization of EI produces a high chemical background. This method should therefore be suitable for use in phthalate determinations in environmental media, plastics, cosmetics and many other matrices. Locking the retention time enhances confidence in the characterization of the various phthalate isomers on the basis of their definitive retention time. This is especially helpful for determinations using selected ion monitoring (SIM), since SIM groups need not be edited after column maintenance [4]. The data in Table 1 facilitate the development of a SIM method. The extension of the method to phthalates which elute at higher temperatures ( >350C) is also easily accomplished.
References
1. Blount, B.C., et al., Levels of seven urinary metabolites in a human reference population. Environmental Health Perspectives, 2000. 108 (10): p. 979-982. 2. Official Journal of the European Communities, Decision 198/815/EC. 1999, European Commission; European Union Scientific Committee on Toxicology, Ecotoxicology, and the Environment. 3. Giarrocco, V., B. Quimby, and M. Klee, Retention Time Locking: Concepts and Applications. 1998, Publication number (23) 5966-2469E, Agilent Technologies. 4. Prest, H. and P. Cormia, Retention Time Locking: Advantages in GC/MS SIM Analysis. 1999, Publication number (23) 5967-3797E, Agilent Technologies. 5. Prest, H. and K. Weiner, Retention Time Locking: Creating Custom Retention Time Locked Screener Libraries. 1999, Publication number (23) 5968-8657E, Agilent Technologies. 6. Harrison, A.G., Chemical Ionization Mass Spectrometry. Second ed. 1992: CRC Press.
For more information

For more information on our products and services, you can visit our site on the World Wide Web at: http://www.agilent.com/chem.
Agilent Technologies shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Copyright 2001 Agilent Technologies All rights reserved. Reproduction and adaptation are prohibited. Printed in the U.S.A. March 7, 2001 5988-2244EN
Development of an LC/MS Method for the Analysis of Rodenticides Application Note

Environmental
Friedrich Mandel and Juergen Wendt Agilent Technologies Raffaele Vistocco and Christian Bachmann Chromatography Department, A.P.P.A. Bolzano (Italy)
Introduction
Rodents such as rats and mice must be controlled because they destroy food supplies and serve as vector hosts for human diseases such as hantavirus. Although individual animals or small groups can be removed by trapping, rodenticides are frequently used in rodent control. Most rodenticides are also toxic to humans and domesticated animals such as dogs. In this work, an LC/MS method was developed to monitor several coumarin rodenticides in complex matrices. Anticoagulant poisons work by interfering with the blood clotting mechanism. After repeated ingestion of relatively small doses, a lethal dose accumulates and causes death due to internal bleeding. Anticoagulant poisoning can also cause spontaneous bleeding from the nose, gums and the gastrointestinal and urinary tracts.
Development of an LC/MS Method for the Analysis of Rodenticides
Agilent Technologies
Experimental
All experiments were done on an Agilent 1100 Series LC/MSD system that was comprised of a binary pump, vacuum degasser, autosampler, thermostatted column compartment with column-switching valve, diode-array detector, and an LC/MSD. The LC/MSD was used with either the electrospray ionization (ESI) or atmospheric pressure chemical ionization (APCI) source. Complete system control and data evaluation was done on the Agilent ChemStation for LC/MS. Reagent grade chemicals and HPLC grade solvents were used for preparing mobile phases and standards. The compounds used for this study are all classified as coumarin rodenticides due to the common coumarin moiety in the structures (Figure 1). Method development included assessing ionization technique (ESI and APCI), ionization polarity and fragmentor voltage to determine the best conditions for the three analytes. In order to test the performance of the method, spiked sausage extract and dog stomach extract were analyzed. The samples were spiked at either 0.5 or 5 ppm, then subjected to a general extraction and sample cleanup.1 The spiked samples were extracted in acetone for four hours. The acetone was then evaporated to a volume of approximately 50 ml and extracted with 100 ml of methylene chloride. The acetone layer was dried with anhydrous sodium sulfate, filtered, and evaporated to dryness. After redissolving in 2 ml of acetone, the extract was applied to a column containing 15 g silica gel and 1 g charcoal. After elution with methylene chloride/benzene/acetone (10/2/2 v/v), the eluate was evaporated to dryness and redissolved in 5 ml of methanol. The sample extracts were then analyzed by ESI-LC/MS.
O O CH OH H2C OH O C CH3 O O O
OH
Difenacoum C31H24O3 mass 444.173

O O
Coumatetralyl C19H16O3 mass 292.11

Figure 1. Coumarin rodenticides.
Coumafuryl C17H14O5 mass 298.084

Both coumafuryl and coumatetralyl showed poor response in positive mode ESI compared to negative mode ESI (Figure 2). The positive mode response for these compounds was better in APCI than ESI, but the negative mode APCI response for coumatetralyl was weak compared to ESI. Based on these results, ESI negative mode was selected for the optimized method.
10000 0
Positive m/z 299
ANALYSIS METHOD Chromatographic Conditions Column: 150 2.1 mm Zorbax XDB-C18, 5 m (p/n 993700-902) Mobile phase: A = 2 mM ammonium acetate in water B = methanol Gradient: start with 30% B at 2 min 50% B at 4 min 100% B Flow rate: 0.4 ml/min from 0 to 2 min, then 0.5 ml/min Column temp: 50C Injection vol: 2 l Diode-array detector: signal: 280, 16 nm; reference: 550, 100 nm
Negative m/z 297 40000 0 Positive m/z 293 1000 0 10000 0 Positive m/z 445 20000 0 20000 0 0 1 2 3 4 5 6 7 min Negative m/z 443 Negative m/z 291
coumafuryl
coumatetralyl
difenacoum
ESI-MS Conditions Source: Drying gas flow: Nebulizer: Drying gas temp: Vcap: Stepsize: Peakwidth: Time filter: Scan: SIM ions (negative mode):
Fragmentor:
ESI 10 l/min 40 psig 350C 2500 V (positive and negative) 0.1 0.2 min On m/z 150500 296, 297, 298 Coumafuryl 290, 291, 292 Coumatetralyl 442, 443, 444 Difenacoum variable 180 V (50275) 100 V (280) 120 V (400)
Figure 2. Comparison of positive and negative ionization modes for electrospray LC/MS.
Typically, positive ions will fragment more easily than negative ions, so the fragmentor voltage was optimized for both modes. For example, difenacoum showed extensive fragmentation at 160 V in positive ion mode but no fragmentation at the same voltage in negative ion mode. Optimized fragmentor voltages were used for each ion in the final method. For samples extracted from a complex matrix, the deprotonated molecule as well as the ions 1 m/z above and below can be monitored to check for coeluting artifacts. In the final method, this strategy was employed so a set of three ions was used for each analyte. The ions for each analyte were time-programmed as a SIM ion group in order to maximize the dwell time and thus maintain maximum sensitivity. Figure 3 shows the results for the
spiked sausage extract. For this analysis, data was collected in scan mode because the coumafuryl and difenacoum were present in sufficient concentration. The mass spectra from the spiked analytes show good quality at the 5 ppm level. The dog stomach extract was spiked at a lower level (0.5 ppm) so SIM analysis was done for this sample (Figure 4). Only the spiked analyte, coumafuryl, shows a significant signal. This work demonstrates that anticoagulant rodenticides can be detected by APCI and ESI in both positive and negative ion modes. Negative mode ESI-LC/MS was found to be the best choice for the three target compounds. This LC/MS method was capable of detecting the rodenticides in complex matrices.
coumafuryl
250000 150000 50000 m/z 297
coumafuryl
60000 m/z 297
2.833
2.882
40000 443.1 20000 0
coumafuryl
Max: 153123 311.2 250000 150000 265.1 343.2 50000 m/z 291
60000 40000 20000 1 2 3 4 5 minutes 6 200 300 400 m/z 0
m/z 291
difenacoum
difenacoum
297.0 60000 Max: 60723 5.374 40000 20000 0 211.1 0 1 2 3 4 5 6 min m/z 443
250000 150000 50000
m/z 443
3 4 5 minutes
200 300 400 m/z
Figure 4. Extracted ion chromatograms from the analysis of dog stomach extract spiked with 0.5 ppm of coumafuryl. Data was collected in negative ion using SIM mode.
Figure 3. Extracted ion chromatograms (left) and mass spectra (right) from the analysis of sausage extract spiked with 5 ppm of coumafuryl and difenacoum. Data was collected in negative ion using scan mode.
References
1. Faucannet, V., Pouliquen, H., and Pinault, L., Journal of Analytical Toxicolocy, 27, 1997.
For more information on our products and services, you can visit our site on the World Wide Web at: http://www.agilent.com/chem. Agilent Technologies shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance or use of this material. Information, descriptions and specifications in this publication are subject to change without notice. Copyright 2000 Agilent Technologies All rights reserved. Reproduction and adaptation is prohibited. Printed in the USA June 2000 (23) 5968-0561E
Authors
Friedrich Mandel and Juergen Wendt are application chemists at Agilent Technologies in Waldbronn, Germany. Raffaele Vistocco and Christian Bachmann are in the Chromatography Department of A.P.P.A. Bolzano in Italy.
Fast Screening of Pesticides and Endocrine Disrupters Using the Agilent 6890/5973N GC/MSD System, Part II
Application
Gas Chromatography May 2000
for 567 of the most common pesticides and endocrine disrupters of concern worldwide. A GC/MSD method was developed based on the standard 42-min method1 to screen for all 567 of the most common analytes. A specific combination of column stationary phase, carrier gas flow rate, and oven temperature programming is required to lock all the compounds to an expected retention timetable2. Compound identification based only on spectral searching alone is difficult when analyzing extracts containing significant sample matrix content because of overlapping peaks and noisy baselines. The new screening tool, integrated within Agilent's ChemStation for MSD, searches for all 567 compounds. It first checks and integrates four characteristic ions within the expected time window and then prints a report showing "hits" and "possible hits" (ratios of characteristic ions that do not match the expected values in the library within specified limits). In Part I of the MSD fast screening application brief 3 , a 10 m 0.1 mm 0.1 m Agilent HP-5 column was used to increase analysis speed up to fourfold. In this application brief, a 15 m 0.25 mm 0.25 m Agilent HP-5MS column was used. The faster methods were scaled exactly as predicted by using a combination of Agilent's method translation (MTL) and RTL software. Because scaling was exact, these faster methods can be used with precisely-scaled pesticide libraries, making the screening process even more powerful and adaptable to individual needs.
Authors C. Kai Meng Agilent Technologies, Inc. 2850 Centerville Road Wilmington, Delaware 19808-1610 USA Michael J. Szelewski Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA
Abstract Agilent Technologies' new, fast GC/MSD method can significantly speed up the screening of pesticides. Agilent's GC Method Translation software (available free from the Agilent Technologies Web site, http://www.chem.agilent.com/cag/ servsup/usersoft/main.html#mxlator) was used in developing the new method based on the standard 42-min method. A 15 m 0.25 mm 0.25 mm Agilent HP-5MS column was used to increase analysis speed up to fourfold. The time savings were implemented in increments (down to 10.5 minutes) to verify the predictability of scaling and the effect of scaling on the signal-to-noise ratio. Key Words
RTL, pesticide, environmental, screening, fast GC, method translation, 5973, 6890, MTL
Introduction
Analysts want faster analyses to improve laboratory productivity. Often, when speeding up GC methods, an analyst will trade resolution for increased analysis speed. This loss of resolution can complicate peak identification, even with a mass selective detector (MSD). Agilent Technologies has developed new techniques to solve the peak identification problem based on Agilent's retention time locking (RTL) and a new mass spectral library that contains the locked retention times and characteristic ions
Experimental
The GC method translation software tool was used to find operating conditions for the faster methods. Figure 1 is a screen capture of MTL software data entry showing the original conditions and the new chromatographic conditions for a fourfold speed gain. The column flow rate, which is helpful to avoid exceeding MSD pumping capacity 4, also is found in the table. In this study, a turbo pump was used, which could handle the 3.8 mL/min carrier flow. The program also determined the required column head pressure and corresponding oven ramp. The Agilent 6890 GC fast oven option (220/240V in the U.S.) was required for the faster oven ramp used in this study. General chromatographic conditions are listed in table 1. The standard used was a mixture of 26 pesticides at 10 ppm. A 15 m 0.25 mm 0.25 m Agilent HP-5MS column (part number 19091S-431) was used. The head pressure determined by the method translation software (18 psi) was used as the starting point for retention time locking. The column head pressure required to lock retention times of the compounds to the library (the original retention time divided by 4) was determined using the automated RTL process integrated within the Agilent ChemStation for MSD.
Figure 1. Screen capture showing the method translation (MTL) software data entry used in a 4X speed gain translation.
This process (first translate the method then lock the retention times) was repeated for the 2.5X time reductions. Figure 2 shows the results of the shortened analysis times. The three chromatograms look extremely similar, except that the time axis is scaled proportionally. Because MTL followed by RTL scales methods very precisely, scaled screening libraries for corresponding time reductions can be obtained by dividing the retention times in the library by the speed gain (which does not have to be an integer). Using the same injection method (1-L splitless), the peak heights of the faster runs were twice those from the original
Table 1 Chromatographic Conditions
Speed GC Column Injection mode Column head pressure Column flow (mL/min) Inlet control mode Carrier gas Injector Temp. Oven Temp. Ramp 1 Ramp 2 Ramp 3 Oven equilibration Injection volume Liner MS Conditions (Turbo pump) Solvent delay Tune file Low mass High mass Threshold Sampling Scans/sec Quad Temp. Source Temp. Transfer line Temp. Acquisition mode
Onefold 110 V 30 m 0.25 mm 0.25 mm HP-5MS (P/N 19091S-433) Splitless 18.0 psi 1.9 Constant pressure Helium 250 C 70 (2 min) 25 C/min 150 (0 min) 3 C/min 200 (0 min) 8 C/min 280 (10 min) 2 min 1 mL 5183-4647
Two and a half fold 220/240 V 15 m 0.25 mm 0.25 mm HP-5MS (P/N 19091S-431) Splitless 5.74 psi 1.49 Constant pressure Helium 250 C 70 (0.8 min) 62.5 150 (0 min) 7.5 200 (0 min) 20 280 (4 min) 2 min 1 mL 5183-4647
Fourfold
18.0 psi 3.8
70 (0.5 min) 100 150 (0 min) 12 200 (0 min) 32 280 (2.5 min)
3 min Atune.u 35 amu 500 amu 150 2 3.15 150 C 230 C 280 C Scan (EI)
1.44 min Atune.u 35 amu 450 amu 250 2 3.50 150 C 230 C 280 C Scan (EI)
0.9 min
1 6.54
analysis. A faster oven ramp and the shorter column made the peaks narrower and higher, so an improvement in the signal-to-noise ratio is realized with the faster methods.
Abundance 7000000 6000000 5000000 4000000
TIC: RTLDEMO.D
1X
Conclusion
The highly accurate and reproducible pressure and temperature control of the Agilent 6890 GC allows precise scaling of the standard 42-min GC/MSD pesticide method. Run time was shortened to 10.5 minutes using a fast oven ramp rate and a 15-meter, 250-micron column. The combination of MTL and RTL facilitated scaling and yielded exact scaling. RTL libraries can be scaled accurately to correspond to the faster analyses.
3000000 2000000 1000000 0 Time 5.00 10.00 15.00 20.00 25.00 30.00 35.00 40.00
Agilent HP-5MS, 30 m 0.25 mm 0.25 m

Abundance 18000000 16000000
TIC: 15MMIX1A.D
References 1. B. D. Quimby, L.M. Blumberg, M. S. Klee, and P. L. Wylie, "Precise Time-Scaling of Gas Chromatographic Methods Using Method Translation and Retention Time Locking," Application Note 228-401, Agilent publication number 5967-5820E, May 1998. 2. H. Prest, P. L. Wylie, K. Weiner, and D. Agnew, "Efficient Screening for Pesticides and Endocrine Disrupters Using the HP 6890/ 5973 GC/MSD System," Agilent publication number 5968-4884E, April 1999.
3. C. K. Meng and M. Szelewski, "Fast Screening of Pesticide and Endocrine Disrupters Using the Agilent 6890/5973N GC/MSD System", Agilent publication number 5968-9220, January 2000. 4. H. Prest, "GC Column Selection and Pumping Considerations for Electron and Chemical Ionization MSD operation," Agilent publication number 5968-7958E, November 1999.
14000000 12000000 10000000 8000000 6000000 4000000 2000000 Time 0 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 11.00 12.00 13.00 14.00 15.00 16.00
2.5X
Abundance 18000000 TIC: 4X-MIX1A.D 16000000 14000000 12000000 10000000 8000000 6000000 4000000 2000000 0 Time 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00
4X
Agilent HP-5MS, 15 m 0.25 mm 0.25 m

Figure 2. The TICs of the 2.5X and 4X speedups. The standard analysis (1X) was 42 minutes long.
Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Copyright 2000 Agilent Technologies, Inc. Printed in the USA 5/00 5980-1057E
Trace Level Pesticide Analysis by GC/MS Using Large-Volume Injection
Application
Gas Chromatography September 1997
Author
Philip L. Wylie Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA
were injected into a single PTV liner. This application note includes recommendations for doing LVI using the PTV/6890/5973 GC/MSD system.
Introduction
More than 700 pesticides are registered for use in the world1 , and many more continue to persist in the environment, even though they are no longer being applied. For the protection of human health and the environment, pesticide residues are routinely monitored in food, water, soil, and tissue samples. "Acceptable" residue limits have been set for various foods and environmental samples by agencies such as the United States Environmental Protection Agency (U.S. EPA), the Codex Alimentarius Commission2 , and many other governmental organizations around the world. A great many methods have been developed to screen for pesticides in food3-7 and the environment8-10 to ensure that risks associated with pesticide use are minimized.
Abstract
Large-volume injection (LVI) using the Agilent programmable temperature vaporizing (PTV) inlet can improve gas chromatography system detection limits by one to two orders of magnitude over standard methods that call for 1- or 2- L injections. An Agilent 6890 Series gas chromatograph (GC), configured with a PTV inlet, a 6890 Series automatic liquid sampler (ALS), and an Agilent 5973 mass selective detector (MSD), was used for the analysis of pesticides in standards and several food extracts. By making 100- L injections, several pesticides could be identified by scanning gas chromatography/mass spectrometry (GC/MS) at the 100 ppt (100 ng/L) level. The PTV inlet tolerated dirty food extracts very well; more than 1,500 L of such samples
Recently, concern has increased that certain pesticides and other synthetic chemicals may be acting as pseudo hormones which disrupt the normal function of the endocrine system in wildlife and humans. Birth defects, behavioral changes, breast cancer, lowered sperm counts, and reduced intelligence are among the many disorders that have been blamed on these "endocrine disrupting" compounds, though much research must be done to verify these assertions. In 1996, Colborn, Domanoski, and Myers11 brought these issues into the public spotlight with the publication of their book Our Stolen Future. Recently, the United States Congress passed legislation calling for increased testing of suspected endocrine disrupters and monitoring their levels in food12 and water13 supplies. Because the endocrine system can be exquisitely sensitive to extremely low hormone concentrations, there is a need to measure concentrations of suspected endocrine disrupters (many of which are pesticides) at very low levels. Initiatives such as the Pesticide Data Program, developed by the United States Department of Agriculture14 , seek to
determine the lowest measurable pesticide levels in various foods to develop a total exposure model. Clearly, there is pressure to push pesticide detection limits to even lower levels than are routinely achieved today. Most residue measurements are made by gas chromatography using a variety of element-selective or mass spectral detectors (GC/MS). Therefore, to achieve lower detection limits, it is necessary to improve the detection limits of these GC methods. In GC, there are primarily four ways to improve method detection limits: 1) increase the concentration of analytes in a sample, usually by reducing the volume of an extract; 2) increase the sensitivity of the detector; 3) increase the selectivity of the detector to reduce chemical background "noise" or 4) increase the volume of sample injected. Because GC/ MS can be highly selective and extremely sensitive, it is often the method of choice for pesticide analysis and/or confirmation. However, for the reasons discussed above, there are occasions when even greater sensitivity is required. This application note describes a method for increasing GC/MS system detection limits by making large-volume injections (LVI) using Agilent's new programmable temperature vaporizing (PTV) inlet. Because this LVI technique is detector-independent, it is applicable to other GC configurations that may be used for pesticide residue analysis.
I, p,p'-DDE, propargite, iprodione, methoxychlor, and fenvalerate (mix of isomers I and II) to individual 20mL vials and diluting with 10.0 mL of acetone. Permethrin was obtained as a mixture of permethrin I and permethrin II comprising 32 percent and 27 percent of the sample, respectively, so 16.95 mg of this mixture was diluted with 10 mL of acetone giving a solution in which the combined permethrins represented 1 mg/mL. A stock mixture was prepared by adding 4 mL of the permethrin and fenvalerate solutions and 1 mL of each of the other stock solutions to a 100-mL volumetric flask and diluting to volume with acetone. The resultant solution contained 40 ng/ L each of the combined permethrin and fen-
valerate isomers and 10 ng/ L each of the other 12. This sample was diluted further with acetone to prepare standards that were analyzed by LVI. All these pesticides were obtained in neat form from Chem Service (West Chester, PA USA).
Extracts
Fruit and vegetable extracts were obtained from the Florida Department of Agriculture and Consumer Services (Tallahassee, FL USA). Commodities were extracted using a version of the Luke procedure15-17 that gave a final sample representing 1.75 g of the commodity per mL of extract.
Table 1.
Instrumentation and Conditions Used for Pesticide Samples

6890 Series GC 6890 Series ALS 5973 Series MSD PTV with CO2 cooling HP Vectra XU 6/200 G1701AA Version A.03.00 running Microsoft Windows 95 30 m x 0.25 mm x 0.25 m Agilent HP-5MS
GC/MS System Gas chromatograph Automatic liquid sampler Mass spectral detector Programmable temperature vaporizing inlet Computer for data acquisition and analysis Software Column Instrumental Conditions GC Parameters Carrier gas Inlet liner Syringe size Injection volume Injection delay Inlet temperature program Vent flow Purge flow to split vent Column head pressure Oven temperature program MSD Parameters Acquisition mode Temperatures
Experimental
Pesticide Standard Solution
Stock solutions of 14 pesticides were prepared at 1 mg/mL by adding 10 mg each of trifluralin, hexachlorobenzene, pentachloronitrobenzene, dichloran, chlorothalonil, chlorpyrifosmethyl, chlorpyrifos, endosulfan
Helium Prototype deactivated borosilicate with fritted glass on interior walls (part no. 5183-2041) 50 L 100 L (Inject 10 L 10 times) 12 sec 40 C (4.2 min), 200 C/min to 320 C (2 min) 400 mL/min Vent pressure 0.0 psi for 4.00 min 50.0 mL/min at 6.50 min 0 psi (4 min) then 17.3 psi (constant pressure) 50 C (6.13 min), 30 C/min to 150 C (2 min), 3 C/min to 205 C (0 min), 10 C/min to 250 C (20 min) Scan (35-550 amu) Transfer line = 280 C, MS quad = 150 C, MS source = 230 C
Instrumentation
Table 1 lists the instrumentation and chromatographic conditions used for LVI and GC/MS analysis of pesticide samples.
manual injections are usually impractical and good precision may be hard to achieve. The 6890 Series ALS is designed to make one or more injections of up to 25 L into the PTV inlet. After the desired number of injections has been made, the inlet is heated and the chromatography begins. Though the system controls allow up to 99 injections, a reasonable upper limit is about 10, making 250 L the typical injection volume limit for this system. For even larger injections, the controlled speed injector19 should be used. For all of the analyses described below, 100 L were injected by making 10 sequential injections of 10 L each.
How the PTV Works in the Solvent Vent Mode

Figure 1 shows a diagram of the PTV inlet. For large-volume injections, three steps are required. These are: 1) injection and solvent elimination; 2) splitless sample transfer to the GC column; and 3) chromatographic separation and, if desired, a simultaneous inlet bake-out step. The steps are described more completely below.
Brief PTV Tutorial

Before focusing on the PTV/GC/ MS analysis of pesticides, it is important to understand how the PTV inlet operates in the solvent vent mode for large-volume injections.
The PTV Inlet

The PTV inlet has the same basic functions as the split/ splitless inlet except that it is temperature programmable from -60 C (using CO2 cooling) or -160 C (using liquid N2 cooling) to 450 C at rates up to 720 C/min. However, the PTV's design has been optimized for its main uses-LVI and cold split/splitless injection. Although hot split and splitless injections may be made with or without a pressure pulse, care must be taken not to exceed the small internal volume of the PTV inlet. In practice, it is best to choose the Agilent split/splitless inlet for hot injections and the PTV inlet for LVI and cold split/ splitless techniques. Most GC pesticide methods call for injecting 1-2 L; splitless injection is used because it is compatible with dirty extracts of food, soil, or water. Pulsed splitless injection allows one to make injections of up to 5 L using standard equipment18. Enormous gains in system sensitivity can be realized by using the PTV inlet in the "solvent vent" mode, which is compatible with injections of 5-1,000 L. These large injections may be made manually or automatically using either a standard 6890 Series ALS in the multiple injection mode or by using a controlled speed injector available from Gerstel19. Because the injection process may take several minutes,
Injection and Solvent Elimination (Step 1)

During injection, the column head pressure is set to 0 psi to eliminate or, in the case of GC/MS, reduce the flow through the column. When mass spectral detection is used, there is still
Septumless Sampling Head Carrier Gas Line Coolant Liner Seal Heating Coil
Split/Splitless Solenoid Proportional Valve
Glass Inlet Liner
Capillary Column
Figure 1. The PTV inlet shown with the septumless head. The inlet is also available with a septum head that may be equipped with a standard septum or a Merlin Microseal. (Figure reproduced with permission of Gerstel GMBH.)
some flow because the column outlet is under vacuum. At the same time, a steady stream of carrier gas passes through the inlet and out through the split vent. This flow is typically between 100 and 500 mL/min. The sample is injected into the cool liner where it remains as a liquid, dispersed over the liner walls or any packing material that may be in the liner. The steady flow of carrier gas through the liner causes the solvent (and any volatile fraction of the sample) to evaporate and be swept with the carrier gas out through the split vent. This is analogous to "blowing down" a sample with a stream of inert gas, except that this takes place inside the PTV inlet. When most of the solvent has evaporated, the next injection is made and the evaporation process repeats, accumulating more sample in the inlet. To recover an analyte completely, its boiling point should be at least 100 C greater than that of the solvent; most pesticides fall into this category.
Normal
The timing of these multiple injections can be important. If the sample is introduced too rapidly, the liner may become flooded and liquid will be forced out through the split vent. Chromatographically, this shows up as reduced area counts for all analytes (see figure 2A). If there is too much time between injections, all of the solvent may evaporate and more of the volatile analyte fraction may be lost too. This results in poor recovery of volatiles but 100 percent recovery of the less volatile compounds (see figure 2B). Set-points such as inlet temperature, vent flow, and injection delay times can affect recovery of volatiles. Note that for 100 percent recovery, an analyte should have a boiling point at least 100 C greater than the solvent. One can adjust the delay between injections by entering the desired value in the ChemStation software. Some experimentation is usually necessary when setting this delay for a new method. It will be dependent upon such factors as the solvent type, injection volume, vent flow, and inlet temperature.
a splitless injection, except that instead of flash vaporization, the sample is transferred as the inlet temperature is programmed up. For the most gentle treatment of labile analytes, slow ramp rates may be used. This allows analytes to be flushed into the column at the minimum temperature needed for volatilization. When sample decomposition is not a problem, the inlet may be heated as fast as 720 C/min.
Chromatographic Separation (Step 3)

During sample transfer, the oven temperature is usually held between 30 C below and 20 C above the solvent's atmospheric boiling point, depending on whether the solvent effect is needed to focus the more volatile fraction of the analytes. Again, some experimentation is necessary to optimize peak shapes. After the sample has been transferred in step 2, the oven temperature is programmed up and chromatography begins. After the inlet has reached its maximum temperature and sufficient time has elapsed to transfer the sample to the column, a purge flow of 30-50 mL/min is restored to the split vent. If desired, one can set a very large split flow for a few minutes and bake out the inlet at a higher temperature to remove nonvolatile impurities. To conserve carrier gas, gas saver should be turned on at the end of this bake-out step.
Splitless Sample Transfer to the GC Column (Step 2)

Inlet Flooding
Once the desired number of injections has been made, the column head pressure is restored and the vent flow is tur ned off. At this point, the inlet temperature is programmed up to a value that is sufficient to transfer all of the desired analytes to the GC column. This step is similar to
A Sample is injected too rapidly
Normal
Volatiles Lost
B Solvent evaporates completely between injections
Figure 2. Chromatograms A and B illustrate the result of poor timing of multiple injections.
Entering PTV Inlet Parameters into the Agilent ChemStation

When preparing the PTV portion of a GC method, one should first decide on the sample size and how many injections are required. In this work, ten 10- L injections were made for a total of 100 L. When entering parameters into the ChemStation screen, the Injector icon is first selected (figure 3) under the "GC edit parameters" menu. Next, the Configure button is pressed to enter the syringe size and enable multiple injections. From the main injector screen, the injection volume (10 L) and number of injections are entered10 . For this work, a 12-second delay was chosen between injections to allow for solvent evaporation. The estimated total injection time is listed on the Inlets screen (figure 4). This is helpful when setting the inlet and oven parameters. First, the vent flow rate (400 mL/min for these analyses) is chosen, which sets the vent pressure to 0 psi until the injection sequence is done and solvent from the last injection has largely evaporated (4.00 min in figure 4). This is done by entering these values in the following fields: Vent Flow 400 mL/min Vent pressure 0.0 psi until 4.00 min Next, the purge flow and elapsed time are set by entering values in the following field: Purge Flow to Split Vent 50.0 mL/min @ 6.50 min Note that as an aid in setting up the method, the "estimated total injection time" is shown just above the previous data entry fields.
In this example, the normal column head pressure was restored and the vent flow was turned off at 4.00 min. This prepares the inlet for the splitless transfer of the sample to the column. The vent flow remained off until it was set to 50 mL/min at 6.5 min. Thus, there is a 2.5-min period for inlet temperature
programming and splitless sample transfer to the column. In this example, the inlet was held at 40 o C for 4.2 min, enough time to make 10 injections, turn off the purge flow, and restore the column head pressure; the PTV was then programmed to 320 o C at 200 o C/min (figure 4).
Figure 3. The injector screen from Agilent GC and GC/MS ChemStation software showing the setpoints available for multiple injections. To configure the sampler for multiple injections, set the syringe size, and choose slow injection, click on the Configure button.
Figure 4. The inlets screen from Agilent GC and GC/MS ChemStation software showing the setpoints available for operation of the PTV inlet in the solvent vent mode.
Although not done for these analyses, the inlet could be baked out by setting the "purge flow to split vent" to a large value (perhaps 500 mL/min) at the end of the splitless time (6.50 min) and at the same time, program the inlet to a higher temperature. After the bake-out period, the inlet temperature is programmed downward and gas saver is turned on. Normally, the GC oven is held at its starting temperature until the splitless injection is complete (6.50 min in this case) at which time oven temperature programming is begun. For this work, the oven temperature program was begun at 6.13 min so that the pesticide retention times would match a retention time data base that was in use. Figure 5 diagrams the PTV and GC oven setpoints used for this work.
PTV Inlet Liner Considerations

The correct liner choice is critical to the success of any pesticide analysis by PTV injection. The liner must be thoroughly deactivated or many labile pesticides may decompose or adsorb in the inlet. In general, any liner containing glass wool will be unsatisfactory for the analysis of labile pesticides, whether or not the glass wool is deactivated. At this time, two PTV liners are suggested for pesticide analysis: Part no. 5183-2037 is a deactivated, open multibaffled liner with no internal packing that may be used for single or multiple injections of 5 L or less. This liner gives very good recovery for pesticides, even extremely difficult ones such as acephate and methamidophos.
Part no. 5183-2041 is a deactivated liner with an internal coating of sintered glass to give it more surface area and is, therefore, suitable for single or multiple 25- L injections. This liner gives better than 70 percent recovery for most pesticides, although tests have shown that acephate and methamidophos cannot be analyzed using this liner, and that recoveries of guthion are often less than 50 percent. A prototype version of this liner was used for all of the work described in this application note.
Multiple injections
12 sec injection delay
PTV purge flow 400 mL/min 4.00 min 0 psi 4.00 min 40 C 4.20 min 50 C 6.13 min
Column head pressure PTV temperature Oven temperature
Figure 5. Illustration of the GC and sampler setpoints used for 100- L injections of pesticide samples. Note that normally, the GC oven hold period would have been at least 6.5 min for this method. A value of 6.13 min pesticide retention times to a data base.
}
n 0 mL/min 280 C 200 C/min 50 mL/min 6.50 min 30 C/min
6

When compared to a typical 2-L splitless injection, 100- L PTV injections can often result in a 50-fold improvement in system detection limits. Selective detectors such as the MSD can help the analyst to realize the full measure of this sensitivity improvement by excluding background that may be introduced from solvent impurities, vial cap extract, and indigenous compounds coextracted with the analytes. In this application, it was possible to see most of the pesticides in the 14-component mixture at 100 ppt in the scan mode (400 ppt for the isomer mixes of permethrin and fenvalerate). Figure 6 shows extracted ion chromatograms for trifluralin and hexachlorobenzene (HCB) at 100 ppt. Library searching gave a match quality of 93 for the HCB peak. Fenvalerate isomers I and II were found in the solution in a ratio of about 78:22. Figure 7 shows extracted ion chromatograms for fenvalerate I at a concentration of 311 ppt.
Trifluralin (100 ppt) m/z 306
m/z 264
A Extracted ion current chromatograms of trifluralin
Hexachlorobenzene (100 ppt) Extracted ions 284, 286, and 282 Match quality = 93
B Extracted ion current chromatogram of HCB with its mass spectrum and library match
Figure 6. Scanning GC/MS results for a pesticide standard containing Trifluralin and Hexachlorobenzene at 100 ppt. (Ten 10- L injections were made using the PTV inlet.)
Fenvalerate I (311 ppt) m/z 167
m/z 125
m/z 225
Figure 7. Extracted ion current chromatograms of Fenvalerate I at a concentration of 311 ppt in a pesticide standard. (Ten 10- L injections were made using the PTV inlet.)
Analysis of a bell pepper extract revealed several pesticide residues. As seen in figure 8, chlorpyrifos and the endosulfans were easily detected. The Florida Department of Agriculture determined the concentration of chlorpyrifos, alpha-endosulfan, betaendosulfan, and endosulfansulfate to be 0.210, 0.011, 0.018, and 0.013 ppm, respectively. It is important to note that these compounds could be detected with very high selectivity by extracting high mass ions that are characteristic of these pesticides but not of the matrix. Using LVI, there is ample signal from these less abundant ions for good quantitation. With normal injection volumes, selectivity may have to be compromised and the most abundant ions extracted in a pesticide spectrum to gain sensitivity. Phosmet, captan, and propoxur were all easily detected in a pear sample. The total ion current chromatogram (TIC) is shown in figure 9 along with spectrum obtained for captan juxtaposed with the library spectrum. Figure 10 shows the propoxur peak along with 2,4,6-tribromoanisole and 2,4,6-tribromophenol, two other compounds that were surprising to find in a pear sample. Though the origin of these brominated compounds is not known, a recent paper by Hoffmann and Sponholz 20 suggests that tribromophenol is used to treat storage palettes for the prevention of fire and mold growth, and that the anisole is formed from the phenol microbiologically. Perhaps these pears were shipped in containers that had been similarly treated.
Figure 8. GC/MS Analysis of a bell pepper extract. (Ten 10- L injections were made using the PTV inlet.) Using LVI, there was sufficient signal to use high mass ions with smaller abundances to achieve greater selectivity.
Figure 9. TIC of a pear extract resulting from a 100- L Injection (10 x 10 L). Captan was easily detected, and its spectrum gave a library match quality of 96.
Figure 10. TIC of a pear extract resulting from a 100- L Injection (10 x 10 L). Propoxur and two brominated phenolics were easily identified.
A single sintered glass coated liner of the type described above (part no. 5183-2041) was used for about ten 50- and ten 100- L injections (ca. 1,500 L total) of vegetable extracts before it was replaced. All of the extracts were rather dirty, and an inlet bake-out step was not used. Although the liner looked somewhat discolored for about 2 cm where injections were made, it still performed well at the time it was replaced.
Conclusion
Using the PTV inlet in the solvent vent mode, it is relatively simple to increase system detection limits by one or two orders of magnitude. When combined with the Agilent 6890 Series automatic liquid sampler,
multiple injections of up to 25 L each into the inlet can be made, allowing the solvent to vent while pesticides and other less volatile analytes accumulate. After the desired sample volume has been introduced (typically 5-250 L), the solvent vent is closed and the sample is transferred to the column in a temperature-programmed splitless injection. By making 100- L injections into a PTV-equipped Agilent 6890 Series GC coupled to the Agilent 5973 MSD, it was possible to see several pesticides at the 100 ng/L level (100 ppt) in the scan mode. With such low detection limits, less abundant ions can be used to identify and quantitate pesticides at low ppb levels, thereby gaining in selectivity as well.
When performing LVI, there are several parameters to adjust and some method development time is usually required. However, the method described herein worked well and can be duplicated for the PTV/GC/MS analysis of pesticides in food.
Acknowledgment
The author wishes to thank Ms. Joanne Cook of the Florida Department of Agriculture and Consumer Services for supplying the food extracts used in these experiments and Dr. Bill Wilson (Agilent Technologies) for supplying liner deactivation test results.
References
1. Tomlin, Clive, ed (1994), The Pesticide Manual, Tenth Edition, British Crop Protection Council, Surry, UK. 2. Miller, R. W., This is Codex Alimentarius, Secretariat of the Joint FAO/WHO Food Standards Programme, Food and Agriculture Organization of the United Nations, Rome. 3. McMahon, B. M. and Hardin, N. F. eds. (1994), Pesticide Analytical Manual, Vol I, Third Edition, U.S. Food and Drug Administration, Washington, DC. 4. Lee, S. M., Papathakis, M. L., Feng, H.-M. C., Hunter, G. G., and Carr, J. E. (1991), Fresenius' A. Anal Chem 339, 376-383. 5 Fillion, J., Hindle, R., Lacroix, M., and Selwyn, J. (1995), J AOAC Int 78, 1252-1266.
7. Luke, M. A., Froberg, J. E., Doose, G. M., Masumoto, H. T. (1981), J Assoc Off Anal Chem 64, 1187-1195. 8. Stan, H. J., ed. (1995), Analysis of Pesticides in Ground and Surface Water II, SpringerVerlag, Berlin, Germany. 9. Wagner, R. E., Kotas, W., and Yogis, G. A., eds. (1994), Guide to Environmental Analytical Methods, 2nd Edition Genium, Schenectady, NY. 10. U.S. Environmental Protection Agency, Test Methods for Evaluating Solid Waste, SW-846, Draft Method 8085: Pesticides by GC/AED. 11. Colborn, T., Dumanoski, D., and Myers, J. P. (1996), Our Stolen Future, Penguin, New York, NY. 12. Food Quality Protection Act of 1996, Public Law 104-170, Congressional Record pp. H8127-H8141. 13. Safe Drinking Water Act Amendments of 1996, Public Law 104-182, Congressional Record pp. H9680-H9700.
14. Pesticide Data Program Annual Summary Calendar Year 1994, U.S. Department of Agriculture, Agricultural Marketing Service, Washington, DC. 15. Luke, M. A., Froberg, J. E., Doose, G. M., and Masumoto, H. T. (1981), J Assoc Off Anal Chem 64, 1187-1195. 16. Luke, M. A., and Doose, G. M. (1983), Bull Environ Contamin Toxicol 30, 110-116. 17. Sawyer, L. D. (1985), J Assoc Off Anal Chem 68, 64-71. 18. Wylie, P. L., Phillips, R. J., Klein, K. J., Thompson, M. Q., and Hermann, B. W. (1991), J High Resol Chromatog 14, 649-655. 19. The controlled speed injector is available from Gerstel US, 1510 Caton Center Dr., Baltimore, MD 21227 USA. 20. Hoffmann, A. and Sponholz, W. (March 1997), American Laboratory, 22-23.
6. Working Group on Development and Improvement of Residueanalytical Methods (1996), Analytical Methods for Pesticide Residues in Food-stuffs, General Inspectorate for Health Protection, Ministry of Public Health, Welfare & Sport, The Netherlands.
Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Microsoft is a U.S. registered trademark and Windows is a U.S. trademark of Microsoft Corporation. HP is a registered trademark of Hewlett-Packard Company. Copyright 2000 Agilent Technologies, Inc. Printed in the USA 4/2000 5966-1214E
Fast Dual-Column GC/ECD Analysis of Chlorinated PesticidesEPA Methods 608 and 8080
Application Note 228-305
Author
Imogene L. Chang, Ph.D. Agilent Technologies Wilmington, DE 19808-1610
Experimental
EPA Method 608 and 8080 targeted pesticides were separated using 30 m x 0.53 mm x 1.0 m HP-35 and HP PAS-1701 columns (part no. 19095G123 and 19094U-023, respectively). Analyses were performed on an HP 5890 Series II GC with EPC, dual split/splitless inlets, and dual ECDs. An Agilent 7673 automatic liquid sampler was used to process the simultaneous splitless injections. A deactivated single-tapered glass liner with a small plug of glass wool (part no. 5181-3316) and a Merlin Table 1. Experimental Conditions
Instrument Requirement Gas Chromatograph Injection Ports Column Detector Sample Introduction Data Collection Experimental Conditions Injection Carrier gas
Microseal septum (part no. 51818816) were used with each split/ splitless inlet. Instrumentation and GC conditions are listed in Table 1. A test mix containing 18 pesticides (50 ppb per component) and two surrogates was prepared from the dilution of certified standard mixes with pesticide-grade hexane (Burdick & Jackson). Pesticides in the test mix are listed in Table 2.
Abstract
Dual-column analysis with HP-35 and HP PAS-1701 columns was used to analyze chlorinated pesticides targeted in EPA Methods 608 and 8080 for wastewater and solid wastes. GC parameters were optimized using the Agilent 5890 Series II gas chromatograph (GC) with electronic pressure control (EPC), a dual injector, and a dual electron capture detector (ECD) system. The analysis of 18 pesticides was completed in 12 minutes.
Agilent Technologies 5890 Series II with EPC Dual split/splitless inlets HP-35, 30 m x 0.53 mm x 1.0 m (Part no. 19095G-123) HP PAS-1701, 30 m x 0.53 mm x 1.0 m (Part no. 19095S-123) Dual ECD 7673 automatic sampler with dual injectors 3365 ChemStation and HP Vectra 486/33T PC
Introduction
Currently, many testing laboratories use dual-column/dual-ECD GC systems to analyze the chlorinated pesticides specified in EPA Methods 608 and 80801,2. For this application, EPC was used with an HP-35 column (35% phenyl, 65% methyl polysiloxane phase) as the primary column and the HP PAS-1701 column for confirmation. The unique selectivity of the HP-35 column for this set of chlorinated pesticides permitted focus on the optimization of oven temperature for the HP PAS-1701 column. Individual EPC ports for each injector permitted individual regulation of column flow for both the HP-35 and the HP PAS-1701.
Splitless 1 l, purge delay, 0.75 min, inlet temperature of 250C (A) HP-35, pressure program: 8.6 psi (1 min) at 0.5 psi/min to 12 psi and at 3.0 psi/min to 25 psi (0 min) (B) HP-1701, helium, 10 ml/min constant flow 160C (1 min) to 280C at 10C/min and to 300C (2 min) at 25C/min ECD (300C), 120 ml/min N2 makeup, 6 ml/min anode purge
Oven Detector

In a dual-column/dual-ECD system, samples introduced in a single injection can be split between two columns using a Y-connector and detected by different ECDs. However, when using a Y-connector without EPC, the split sample flow to each column cannot be optimized, and equal and consistent sample splits cannot be presumed. The only variable that can be optimized, in dual-column ECD analysis using a Y-connector is the oven temperature program, which can be optimally balanced for the two dissimilar columns. Using dual-column GC/ECD without EPC, it would typically require 45 to 60 minutes to obtain baseline separations for EPA Method 608 and 8080 targeted pesticides (see Figure 1). A typical run from an environmental testing laboratory for a test mix containing 18 targeted pesticides and two surrogates is shown in Figure 1. A
Table 2. Chlorinated Pesticides.

Peak No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Pesticides Tatrachloro-m-xylene (SS1) alpha-BHC Lindane beta-BHC Heptachlor delta-BHC Aldrin Heptachlor epoxide Endosulfan I 4,4-DDE Dieldrin Endrin 4,4-DDD Endosulfan II 4,4-DDT Endrin aldehyde Endosulfan sulfate Methoxychlor Endrin ketone Decachlorobiphenyl (SS2)
Yconnector was used to split samples for both columns, DB-608 and DB-1701, and good baseline separations were obtained for most analytes. This dual-column run was completed in 45 to 53 minutes using the following oven temperature program: 150C (1 minute) to 260C (18.34 minute) at 3C/minute, then to 275C (5 minutes) at 25C/minute. Clearly this oven temperature program was optimized to separate critical pairs, such as DDE/dieldrin, DDD/endosulfan II, endosulfan sulfate/mehtoxychlor, and methosychlor/endrin ketone for both columns. Figure 2 shows chromatograms of the same pesticide test mix using the HP-35 and HP PAS-1701 columns and EPC. The oven program, 160C (1 minute) to 280C at 10C/minute and to 300C (2 minutes) at 25C/minute, was optimized to separate the critical pairs, endosulfan
DB-608
1 2 3 4 5 6 7 18 8 9 10 11 12 13 15 16 17 19 20
14
12
16
20
24
28
32
36
40
44
48
52
56 min
DB-1701
1 2 3 5 7 6 8 9 10 11 12 14 13 32 15 16 17 18 20 19 4
12
16
20
24
28
36
40
44
48
52
56 min
Figure 1. Typical chromatograms of a pesticides standard mix using DB-608 and DB-1701 columns under GC conditions used in environmental testing laboratories. (See Table 2 for peak identification.)
HP-35 30 m x 0.53 mm x 1m
1 3 20 2 6 4 5 7 8 9 10 11 14 16 17 13 12 15 19 18
00 2 1 4 6 8 10 20 2 3 5 4 7 6 8 10 11 9 14 13 16 18 17 15 0 0 2 4 6 8 10 12 min 12 min
HP PAS-1701 30 m x 0.53 mm x 1m
12
19
Figure 2. Chromatograms of a pesticides standard mix using HP-35 and HP PAS-1701 columns under the GC conditions listed in Table 1. (See Table 2 for peak identification.) II/DDT and methoxychlor/endosulfan sulfate, for the HP PAS-1701 column. In this run, EPC provided a constant 10 ml/minute helium flow to the HP PAS-1701 column throughout the entire run. For the HP-35 column, the following pressure program was used: 8.6 psi (hold 1 minute) at 0.5 psi/minute to 12 psi and at 3.0 psi/minute to 25 psi (hold for constant flow for the remaineder of the run). This pressure program actually provided a 10 ml/minute constant flow to elute most of the pesticides and an increased flow (up to 20 ml/minute) near the end of the run to elute the last analyte, surrogate decachlorobiphenyl and other high-boiling materials from the column. GC parameters optimized for dualcolumn/dual-injector/dual-ECD analysis of chlorinated pesticides reduced analysis time to less than
Copyright , 1995, 2000 Agilent Technologies Printed in USA 04/00 Printed on recycled paper Publication Number 5963-6734E
12 minutes. In addition to speed, all EPA Methods 608 and 8080 targeted pesticides and surrogates were well resolved with good sharp peaks for accurate quantitation.
Acknowledgement
The author wishes to thank Ms. Joann Faulkner and her colleagues at the Pace Laboratory in Petaluma, California, for providing chromatograms and pesticide standards.
Conclusion
The use of EPC permitted individual column flow control to each ECD. The unique selectivity of the HP-35 column for chlorinated pesticides permitted focus on the optimization of oven temperature for the HP PAS1701 column. Run time was 11.5 minutes with good baseline separations for all 20 target pesticides and surrogates. The result was a reduction in sample turnaround time from 54 to 11.5 minutes for a 400% increase in productivity. This is more than a twofold improvement in productivity when compared with conventional methods currently used at many environmental testing laboratories with DB-608 and DB-1701 columns.
References
1. USEPA SW-846 Test methods for Evaluating Solid Wastes, Methods 8000 and 8080, September 1986. 2. USEPA Methods for Organic Chemical Analysis of Municipal and Industrial Wastewater, Method 608, 1982. 3. I. L. Chang, The Analysis of Chlorinated Pesticides and PCBs Using the HP-608 Capillary Column, Agilent Application Note 228-236, Publication No. 5091-7567E.
Detection of Low Levels of Carbaryl in Food Using Agilent LC/MSD Trap System Application Note
Environmental
Jeff Keever, Research Triangle Park Ken Imatani and Paul Goodley, Agilent Technologies
Introduction
Pesticides in food and beverages can be a significant route to human exposure. Extracts of composite food samples typically contain many compounds, which produce interfering compounds at the retention time of interest. This note describes the application of ion trap LC/MS/MS to determine pesticide and herbicide contamination in food samples. The goal of this method was to develop a confirmaThe specific determination of the carbamate pesticide carbaryl in composite food extracts that represent typical meals have previously been investigated using a fluorescence technique.13 tion assay to detect carbaryl in foods at low ng/g levels. Absolute confirmation, identification and quantification of the low levels of carbaryl can be acheived with the LC/MSD Trap primarily due to the high sensitivity and specificity of MS/MS with the ion trap. However, the fluorescence method has drawbacks related to: (1) time consuming post-column derivatization, (2) false positives due to lack of specificity, (3) long sample preparation, (4) insufficient limits of detection, and (5) the lack of confirmatory information from the analytical results.
Detection of Low Levels of Carbaryl in Food Using Agilent LC/MSD Trap System
Experimental
Pesticides in food and beverages are considered a significant route for human exposure. To better assess exposure from dietary intake, whole food samples were collected over a 24-hour period, homogenized using a 5-gallon blender, and then analyzed as a composite sample. Aliquots of the homogenate were extracted using a nine step scheme: (1) acid precipitation using zinc acetate, (2) drying with potassium oxalate and (3) sodium sulfate, (4) soxhlet extraction, (5) solvent exchange, (6) liquid-liquid extraction followed by (7) GPC cleanup with a (8) final solvent exchange and (9) concentration into acetonitrile. Table 1 lists the HPLC method used for the determination of carbaryl in food extracts.
Table 1. LC Conditions. LC column: Mobile phase: 4.6 X 250 mm Zorbax XDB-C8 A = 95% H20: 5% ACN, 0.1M NH4Ac, 0.1% CH3COOH B = 40% H20: 60% ACN, 0.1 NH4Ac, 0.1% CH3COOH
the detection of non-target compounds in the food extract. The selectivity offered by MS/MS enabled the detection of the carbaryl product ion at m/z 145 (Figure 2) generated from the [M+H]+ ion of carbaryl. The subsequent quantitative analysis was performed using this product ion at m/z 145. The limit of detection in this food matrix was found to be 1 ng/g and the limit of quantitation (LOQ) used for the determination in food was 10 ng/g, exhibiting a signal-to-noise of 40:1 in the composite food matrix ( Figure 3). A calibration based on the m/z 145 product ion was linear (linear correlation coefficient of 0.994) over the concentration range of 11000 ng/g (Figure 4). When the Limit of Quantitation of the carbaryl for both techniques were compared, the LC/MS/MS was shown to be 12 times more sensitive than the HPLC fluoresence technique. The LOQ for LC/MS/MS was found to be 110 ng/g while the LOQ for the fluoresence was 120 ng/g.
1.50
Gradient program: 20% B for 10 minutes; ramp up to 100% B over 30 minutes; hold for a further 11 minutes at 100% B Flowrate: 1 mL/min
1.25
Intensity 10 8
1.00
LC/MS/MS was performed using an Agilent LC/MSD Trap mass spectrometer in the MS/MS full scan mode. Carbaryl was detected using positive ion electrospray, in which MS/MS was used to provide additional fragmentation information needed to confirm the presence of carbaryl.
0.75
0.50

Extracts of composite food samples typically contain many compounds (Figure 1) which produce interfering compounds at the retention time of interest. This LC/MS/MS positive ion electrospray method for carbaryl proved 2030 times more sensitive compared to data acquired from a scanning instrument, such as a single quadrupole, enabling
0.25
0.00 0 10 20 30 40
Time
Figure 1. HPLC/ion trap MS full scan determination of carbaryl (retention time of carbaryl shown by arrow) in a composite food sample.
145 m/z
100
145
+H O O C N H CH3
145 m/z
8000
S/N = 40:1
80
Carbaryl
Abundance
6000 60
Intensity
225 250 275
+ O
4000
40
20
2000
[M + H] + 202
0 100 125 150 175 200 0 0 10 20 30 40
m/z
Figure 2. HPLC/MS/MS spectrum of carbaryl from Figure 1. The product ion spectrum was generated by mass selecting the [M+H]+ ion at m/z 202 for collision induced decomposition using an end cap fragmentation voltage of 0.9V for 20 ms.
Time
Figure 3. Ion trace of the carbaryl product ion at m/z 145 at the limit of quantitation LOQ (10 ng/g) in food.
r 2 = .994 Response 107

0.8 0.6 0.4 0.2 0.0 0 1 2 3 4 5 6 7 8 9 10
The LC/MS/MS ion trap method confirmed that there were false positives in samples having a positive carbaryl identification by HPLC fluorescence (5 g/g of carbaryl). The coeluting interference that also underwent post column derivatization, was a different molecular weight (m/z 213 vs 201 for carbaryl) and showed a different product ion spectrum (Figure 5), compared to carbaryl (Figure 2). Based on the MS/MS spectrum of the interference, a possible structure could be assigned as shown in Figure 5.
Concentration (ng/g)[100]
Figure 4. Calibration curve based on the m/z 145 product ion for carbaryl over the concentration range of 11000 ng/g.
References
O C2 H4 N O
+
214
NH C2 H5
100
N H 86 H 145 O mw 213 C O
80
1. Forest, D. L., U. S. Environmental Protection Agency, Method 531.1. Measurement of N-Methylcarbamoyloximes and N-Methylcarbamates in Water by Direct Aqueous Injection HPLC with Post Column Derivatization, Revision 1.0 (1985). 2. Engels, T., National Pesticide Survey Method 5, Revision 2.0 (1987).
Abundance
60
+ N H H
3. Graves, R. L., Method 531.1, Revision 3.0 (1989).

O N H C2 H4 N O+
86
40
Acknowledgements
This work was supported by EPA contract No. 68-C5-0011.
20
145
Author
0 60 80 100 120 140 160 180 200 220
m/z
Figure 5. HPLC/MS/MS spectrum of the coeluting interference that produced a false positives in the HPLC fluorescence method.
Dr. Jeff Keever is a research associate at Research Triangle Institute, Research Triangle Park, NC. Ken Imatani is a product manager and Paul Goodley is a senior applications chemist at Agilent Technologies, Palo Alto, CA.
Conclusions
This application has shown that the analysis of food samples for pesticides/herbicides using ion trap MS/MS detection can be accomplished successfully. This method allows for the detection, confirmation and quantitation of the carbamate pesticide carbaryl in extracts of whole food samples down to 1 ng/g levels.
For more information on our products and services, you can visit our site on the World Wide Web at: http://www.agilent.com/chem. Agilent Technologies shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance or use of this material. Information, descriptions and specifications in this publication are subject to change without notice. Copyright 2000 Agilent Technologies All rights reserved. Reproduction and adaptation is prohibited. Printed in the USA April 2000 (23) 5980-0332E
A Method Used to Screen for 567 Pesticides and Suspected Endocrine Disrupters
Application
Authors
Philip L. Wylie and Bruce D. Quimby Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA
Abstract
A gas chromatographic (GC) method has been developed that can be used to screen for 567 pesticides and suspected endocrine disrupters. In principle, it can be used to screen for any GC-amenable pesticide, metabolite, or endocrine disrupter. The method relies on a technique called retention time locking (RTL). RTL is a procedure that allows the chromatographer to reproduce analyte retention times independent of GC system, column length, or detector so long as columns with the same stationary phase, nominal phase ratio, and diameter are used. Because RTL increases retention time precision and predictability, raw retention times can be used as a more reliable indicator of compound identity. The chromatographer first locks the GC method so that all retention times match those listed in a 567-compound pesticide and
endocrine disrupter retention time table. After analyzing a sample by GC with atomic emission detection (GC-AED), the analyst enters a peaks retention time and known elemental content (presence or absence of heteroatoms) into a dialog box. If elementselective detectors are used, detector response can be entered in addition to or in place of GC-AED data. The software then searches the pesticide table for those compounds that elute at the correct retention time and have the right elemental content or detector response. Most often, the software finds just one compound that meets these criteria, and rarely does it find more than three. Confirmation is performed by GC with mass spectral detection (GC-MS) or by calculation of elemental ratios using GC-AED data. With retention time locking, pesticides have the same retention time on all GC systems; this makes GC-MS confirmation much easier because the analytes retention time is already known.
Introduction
The Pesticide Manual1 lists 759 compounds and biological agents that are used currently as active ingredients in various pesticide formulations. Many compounds, though no longer used, still persist in the environment. For the protection of human health and the environment, acceptable limits in food and water have been set by governmental bureaus such as the United States Environmental Protection Agency (USEPA) and the Codex Alimentarius Commission.2 Numerous methods have been developed to screen for pesticide contamination in food37 and the environment810 to ensure that these standards are met. Certain pesticides and other synthetic chemicals have been suspected of behaving as pseudo hormones, disrupting normal functions of the endocrine system in wildlife and humans. Maladies such as birth defects, behavioral changes, breast cancer, lowered sperm counts, and reduced intelligence have been blamed on exposure to endocrine disrupters.11 The 1996 publication of Our Stolen Future, a book by Colborn,
Key Words
Pesticides, endocrine disrupters, gas chromatography, retention time locking, RTL
Dumanoski, and Myers,11 brought these concerns to the attention of the public. Recently passed legislation in the U.S. calls for more testing of suspected endocrine disrupters and monitoring of them in food12 and water13 supplies. To facilitate more research into the endocrine disrupter issue, methods are needed to detect suspected compounds at trace levels. Because so many pesticides are in use, it is usually impractical to screen for large numbers of them individually and, therefore, multiresidue methods are preferred. Most laboratories that analyze for pesticides in food or the environment screen for only a few dozen compounds because it is often very difficult to screen for more. Recently however, methods have been developed using gas chromatography with mass spectral detection (GC-MS), that can screen for more than 2005 or even 3006 pesticide residues. Still, there is no universal method to analyze for all GC-amenable pesticides. While GC-MS methods are gaining in popularity, there are still some limitations. When methods employ selected ion monitoring (SIM) or tandem mass spectrometry (MS-MS), method development is more tedious and any shift in GC retention times requires that individual analyte retention time windows be shifted accordingly. These methods are only capable of detecting compounds on the target list; there are still hundreds of pesticides, metabolites, and suspected endocrine disrupters that could be missed. On the other hand, methods based on scanning GC-MS alone may require more sample cleanup to avoid interferences from co-extracted indigenous compounds. Typically, these methods do not screen for many pesticide metabo-
lites, endocrine disrupters, or other environmental contaminants. A method that could be used to screen for endocrine disrupters and almost all of the volatile pesticides and metabolites would offer a better means of monitoring the food supply and the environment. This paper describes a universal method that, in principle, could be used to screen for any pesticide, metabolite, or endocrine disrupter that can elute from a gas chromatograph. The screening procedure relies on a new gas chromatographic technique called retention time locking (RTL)1416 with database searching based on retention time and elemental content or detector response. This technique is used to narrow an analytes identity to a few possibilities. Confirmation is performed by GC-MS or by calculation of a compounds elemental ratio using GC with atomic emission detection (GC-AED).
intended for analysis by halogenselective detectors were also subjected to floracil SPE.
Pesticide Retention Time Table

The table containing GC and GC-MS retention times for 567 pesticides, metabolites, and suspected endocrine disrupters was obtained from Agilent Technologies, Wilmington, DE, USA (G2081AA).
Instrumentation
Table 1 lists the instrumentation and chromatographic conditions used for GC-AED screening and GC-MS confirmation.
Software for Method Translation

Software for use in translating the normal GC method to one that runs three times faster was obtained from Agilent Technologies,Wilmington, DE, USA.17
Experimental
Standards and Extracts
Pesticide standards used to develop the retention time table were obtained from Chem Service (West Chester, PA, USA), Promochem Ltd (Welwyn Garden City, Hertfordshire, England), Dr. Ehrenstorfer (Augsburg, Germany), Hayashi Pure Chemical Industries, Ltd (Osaka, Japan), Wako Pure Chemical Industries, Ltd (Osaka, Japan), and GL Sciences Inc (Tokyo, Japan). Fruit and vegetable extracts were obtained from the Florida Department of Agriculture and Consumer Services (Tallahassee, FL, USA). Samples were extracted with acetonitrile followed by solid-phase extraction (SPE) using a C-18 cartridge. Extracts

Retention Time Locking
Key to the development of this method is a new concept in gas chromatography called retention time locking (RTL).1416 Agilent RTL software allows the chromatographer to match analyte retention times from run to run, independent of the GC system, detector, or manufacturing variations in column dimensions. The only requirement is that the columns used have the same stationary phase and the same nominal diameter and phase ratio. For example, with RTL it is possible to match analyte retention times on a GC-AED and a GC-MS even though the MS operates under vacuum and the AED operates at 1.5 psi above ambient pressure. The
procedure also compensates for differences in GC column length resulting from variations in manufacturing or from column cutting required during routine maintenance. RTL is accomplished by adjusting the GC column head pressure until a given analyte, such as an internal standard, has the desired retention time. When this is done, all other analytes in the chromatogram will have the correct retention times as well. Software has been developed that can be used to determine the column head pressure that will lock the retention times correctly after one or two scouting runs. With RTL, it is possible to measure pesticide retention times using a given GC method, and then reproduce those retention times in subsequent runs on the same or different instruments. With this increased retention time precision and predictability, retention times become a far more useful indicator of analyte identity. For many years, relative retention times3,6 or retention indices18,19 have been used to identify compounds. These techniques were developed to compensate for the fact that retention times were not predictable from day to day, column to column, or instrument to instrument. With the increased retention time precision of the Agilent 6890 GC and RTL, it seemed that raw retention times could be used for compound identification instead of retention indices. The chromatographer could simply scan a table of pesticide retention times, eliminating all possibilities but those with close elution times under the same locked GC conditions.
Table 1.
Instrumentation and Conditions of Analysis

6890 6890 Series automatic sampler G2350A atomic emission detector G2360AA GC-AED software running on Microsoft Windows 3.11 30 m 0.25 mm 0.25 m HP-5MS (part no. 19091S-433) Split/splitless, 250 C or 260 C Single-tapered deactivated (part no. 5181-3316) with 2-cm deactivated glass wool plug centered ~3 cm from the top 35 L splitless when running method at 3 speed; 23 L splitless at 1 speed 87.5 psi constant pressure for method at 3 speed; 27.6 psi constant pressure for 1 speed 60 psi (2.01 min), 10 psi/min to 27.9 psi (hold) 70 C (2 min), 25 C/min to 150 C (0 min), 3 C/min to 200 C (0 min), 8 C/min to 280 C (10 min) 290 C 320 C Group 1: Cl 479, Br 478 Group 2: C 193, S 181, N 174 Group 3: P 178 Group 4: F 690 (optional) 6890 6890 Series automatic sampler 5973 MSD G1701AA Version A.03.00 running on Microsoft Windows 95 30 m 0.25 mm 0.25 m HP-5MS (part no. 19091S-433) Split/splitless, 250 C Single-tapered deactivated with small amount of glass wool at the bottom (part no. 5062-3587) 2 L 15.5 psi (constant pressure) Same as GC-AED Scan (35550 amu) 200 rel 3.20 min 150 2.86 Transfer line = 280 C, MS quad = 150 C, MS source = 230 C
Agilent GC-AED System Gas chromatograph Automatic sampler Atomic emission detector Software Column GC inlet Inlet liner Injection volumes Inlet pressure (splitless)* Inlet pressure program (pulsed splitless)* Oven temperature program AED transfer line temperature AED cavity temperature AED elements and wavelengths (nm)
Computer for data acquisition and analysis HP Vectra XM Series 4 5/150
Agilent GC-MS System Gas chromatograph Automatic sampler Mass selective detector Software Column Inlet Inlet liner Injection volume Inlet pressure* Oven temperature program MSD parameters Acquisition mode EM voltage Solvent delay Threshold Scans/sec Temperatures
Computer for data acquisition and analysis HP Vectra XU 6/200
*The column head pressures shown are typical values. Exact values were determined as part of the retention time locking procedure.
Pesticides almost always contain heteroatoms and often have several in a single molecule. The most frequently encountered heteroatoms are O, P, S, N, Cl, Br, and F. GC with atomic emission detection (GC-AED) has been shown to be a useful tool for pesticide screening because it is selective for all of the elements found in these compounds.2022 Thus, GC-AED screening provides valuable information about the elemental content of an unknown molecule. By including this elemental information along with the retention time, it should be possible to narrow pesticide hits to just a few possibilities. To implement this screening procedure, a table of pesticide and endocrine disrupters retention times had to be created using a suitable method under locked conditions.
tested to see if it could still be locked with a retention gap installed. The column chosen for the method was a 30 m 0.25 mm 0.25 m HP-5MS because the same column could be used with any GC-detector combination. In particular, this column was chosen for its low bleed at high temperatures and because its optimum column flow is compatible with GC-MS. The 5% phenyl methyl silicone phase in this column has been widely used for pesticides. Method translation software17,25,26 can be used to increase the speed of a method while retaining the same relative retention times. This can be done by translating the method to a column having the same phase ratio but a smaller id or by increasing the flow rate and oven temperature program while using the same column. The final goal was to design a method that could run at three times the normal speed on the 30-m 0.25-mm 0.25- m HP-5MS column or be translated to a 100- m id column. After several weeks of method development, the GC oven temperature program shown in figure 1a was chosen because it met all of the development criteria. Chlorpyrifos-methyl (C7H7Cl3NO3PS) was chosen as the locking standard. It is an ideal choice because chlorpyrifos-methyl elutes near the middle of the chromatogram (16.596 minutes), has good peak shape, and can be seen by most element-selective detectors. Because GC-AED requires three runs to generate element-selective chromatograms for C, Br, Cl, N, S, and P, the method was translated to run three times faster using software for method translation.17,25,26 The faster oven temperature program used by this method requires 6890 GC systems that are configured for fast oven tem-
perature ramping. The method translation software can be used to speed up the method by any desired factor; even 120-V 6890 GCs can run the method two times faster. However, the original method must be used for GC-MS because of the restriction in flow rates into the MSD. Figure 1b lists the threefold (3 faster GC method.
Pesticide Retention Time Table

Once developed, this method was employed to create a table of locked retention times for the 567 pesticides, metabolites, and suspected endocrine disrupters. Increasing international food trade requires the analysis of pesticides that may be used in the supplying country but not in the recipient country. The goal was to create a table that included pesticides used around the world so pesticide standards were obtained from sources in Europe, Japan, and the USA. A list of suspected endocrine disrupters was compiled from various lists published on the World Wide Web.2731 Many of these compounds are, in fact, pesticides. Most of the GC-amenable endocrine disrupters were analyzed and their retention times appear in the table. However, the 209 polychlorinated biphenyl congeners were not included because their inclusion might actually complicate the identification of organochlorine pesticides. Standards, diluted to 10 ppm in acetone, were first analyzed by GC-MS using the oven temperature program shown in figure 1a and instrumental conditions listed in table 1. Compound identities were verified by matching their spectra to library entries,32 by comparison with a published spectral compendium,33 or by matching spectra to a list of charac-
GC Method for Pesticide Screening

First, a GC method was needed that could elute hundreds of pesticides and endocrine disrupters in a reasonable time with adequate separation. However, the goal was not to separate every possible analyte in a single GC run. Because the intention was to build a table of locked retention times using this method, it had to reproduce these retention times under a variety of conditions. For example, the method needed to accommodate a variety of injection techniques including splitless, pulsed splitless,23,24 cold splitless using a PTV inlet, and oncolumn injection which is occasionally used for the more labile pesticides. The method also needed to perform well with samples dissolved in common solvents such as acetone and methylene chloride. Because a retention gap (or guard column) is sometimes added to protect the analytical column, the method had to be
teristic ions.6 When reference spectral information was not available, the pesticides were verified by spectral interpretation. Samples were then analyzed on two different 6890 GC-FID instruments under the same locked conditions (chlorpyrifosmethyl retention time = 16.596 minutes). The GC-MS retention time and the average of the two GC-FID retention times were tabulated for each compound along with its molecular formula, molecular weight, and CAS number. In addition to these fields, there are four user-definable columns in table 2 that can be used to add such things as mass spectral information, internal catalog numbers, or comments. Table 2 lists a small portion of the database. It must be noted that all retention time values were created using constant column head pressure. This is because GC-MS retention times are very close to those obtained with other detectors when constant pressure is used. In this mode, GC-MS and GC-FID retention times match within 0.1 minute except for three compounds that elute at the very end of the chromatogram. Even in this case, the differences are no more than 0.2 minute. The discrepancy between GC-MS and GC-FID retention times is larger in the constant flow mode.
280 C 10 min 200 C 0 min 150 C 0 min 70 C 2 min b) 3 Speed 150 C 0 min 70 C 0.67 min 75 C/min 9 C/min 25 C/min 200 C 0 min 3 C/min 280 C 3.3 min 24 C/min 8 C/min
a) Normal Speed
Figure 1. a) GC oven temperature program for the Agilent pesticide method at normal speed. When using this method, chlorpyrifos-methyl must be locked to 16.596 minutes. This method is used by GC-MS and can be used by any other GC system. b) GC oven temperature program for the Agilent pesticide method translated to run three times faster. This method may be used with 6890 GCs configured with any detector except an MSD so long as the GC is configured for fast oven temperature ramping. Chlorpyrifos-methyl must be locked to 5.532 minutes.
Table 2.
Small Portion of the Pesticide and Endocrine Disrupter Retention Time Table That Contains 567 Entries. The retention times shown here are for the pesticide method run at normal speed as shown in figure 1a. Chlorpyrifos-methyl was locked to 16.596 minutes ( 0.015 minute for the collection of the tabulated retention time values. The table includes four additional columns for user-defined information.
Name Acetochlor Fuberidazole Methyl parathion Chlorpyrifos methyl Vinclozolin Plifenat Terbucarb Chloranocryl Heptachlor Carbaryl CAS No. 34256-82-1 3878-19-1 298-00-0 5598-13-0 50471-44-8 21757-82-4 001918-11-2 2164-09-2 76-44-8 63-25-2 Molecular Formula C:14,H:20,Cl:1,N:1,O:2, C:12,H:8,N:2,O:1, C:8,H:10,N:1,O:5,P:1,S:1, C:7,H:7,Cl:3,N:1,O:3,P:1,S:1, C:12,H:9,Cl:2,N:1,O:3, C:10,H:7,Cl:5,O:2, C:17,H:27,N:1,O:2, C:10,H:9,Cl:2,N:1,O:1, C:12,H:15,N:1,O:4, C:10,H:5,Cl:7, C:12,H:11,N:1,O:2, MW 269.77 196.21 263.20 322.53 286.11 336.43 277.41 230.09 237.26 373.32 201.22 MSD RT 16.542 16.549 16.594 16.593 16.630 16.641 16.686 16.736 16.741 16.796 16.806
FID RT 16.542 16.549 16.583 16.596 16.637 16.650 16.689 16.730 16.752 16.773 16.800
Pesticide Screening Method

Figure 2 diagrams the pesticide screening method. First, RTL was used to match GC-AED and GC-MS analyte retention times to those listed in the pesticide table. Software for RTL1416 was used to determine the
3-Hydroxycarbofuran 16655-82-6
column head pressure needed to produce a retention time of 16.596 minutes for chlorpyrifos-methyl. When analyzing samples by GC-AED, the method was usually run at 3 speed and chlorpyrifos-methyl was locked to 5.532 minutes. Figure 3 shows the RTL software screen that is used to develop the retention time calibration. To accomplish this for the pesticide method, one should install the 30 m 0.25 mm 0.25 m HP-5MS column (part no. 19091S-433) and set the column head pressure to one of the appropriate nominal values as shown below, making sure to use the constant pressure mode. 26 psi for atmospheric pressure detectors run at normal speed (eg, NPD, FPD) 16 psi for GC-MSD operated at normal speed 27.5 psi for GC-AED operated at normal speed 88 psi for GC-AED operated at 3 speed
Run GC/AED element-selective chromatograms Use retention time locking so GC/AED, GC/MS, and database have same RTs GC/MSD confirmation Confirmation using GC/AED element ratioing Possible compounds Second column confirmation Done -pesticide identified
Perform pesticide database search based on RT and elemental content
Figure 2. Diagram of the screening method that uses retention time locking and retention time table searching to identify pesticides and suspected endocrine disrupters.
To prepare a calibration table similar to the one shown in figure 3, the chromatographer must make five analyses of chlorpyrifos-methyl at the following column head pressures: the nominal pressure, the nominal pressure + 20%, the nominal pressure + 10%, the nominal pressure 10%, and the nominal pressure 20%. Because of the first run affect, it is usually wise to make one or two blank runs before performing the five calibration runs. The five pressures and the chlorpyrifos-methyl retention times are entered into the table provided by the RTL software. This calibration table stays with the method and can be used to lock, or re-lock, the GC
Figure 3. RTL software screen showing typical retention time locking calibration data for the pesticide method run at normal speed using a GC detector that operates at atmospheric pressure.
method as long as that method is used. That is, the five calibration runs only need to be made once for a given method. The software screen for locking the GC method is shown in figure 4. To lock the method, one enters the retention time of chlorpyrifos-methyl and clicks on the Calc new pressure button. The RTL software calculates the pressure needed to lock the chlorpyrifos-methyl peak at the desired retention time. By clicking on the Update current 6890 Method button, this value is entered automatically into the method. One can use Agilents software for method translation17 to convert the method to other speeds (eg, 1.9 ) and determine the nominal column head pressure required. If this is done, the pesticide table must be exported to a spreadsheet program where the analyte retention times can be divided by the appropriate factor (1.9 in this case). This new table can then be imported back into the ChemStation for use with the new method. After locking the method to the table, GC-AED element-selective chromatograms were obtained for C, Cl, Br, N, S, P, and sometimes F. From the GC-AED chromatograms, it was usually possible to determine which heteroatoms were present or absent in the suspected pesticide peak. RTL software was then used to search the database by retention time and elemental content. Figure 5 shows the RTL software screen used for retention time table searching. One can enter the elements known to be present or not present in the GC-AED peak of interest. Up to six other element-selective detectors can be configured for use in the search algorithm. When the presence or absence of a heteroatom is uncertain,
Figure 4. RTL software screen used to calculate the column head pressure needed to lock or re-lock a method. In this case, the chlorpyrifos-methyl retention time was 16.581 minutes and the pressure needed to re-lock the method was calculated to be 26.33 psi. By clicking on the Update current 6890 Method, button, the new pressure is entered automatically into the GC method.
Figure 5. RTL software screen used to search a retention time table on the basis of retention time and known elemental content. In this case, the software will search the Agilent pesticide table at 16.638 0.1 minutes for compounds that contain N, P, and S but do not contain Br or Cl. If element-selective detectors (such as the NPD) are used, this information can be provided to the search routine. Up to six different element-selective detectors can be configured as shown for NPD, FPD (P), FPD (S), and ELCD. nothing is added to the search routine for that element. One must choose a search time window wide enough to include the correct analyte, but narrow enough to eliminate as many extraneous hits as possible. Experience has shown that the normal speed method requires a search window of 0.2 to 0.3 minute. The 3 speed method can use a search window of 0.1 minute. If the heteroatom content is known for a peak, retention time table searching
with these search windows most often finds just one pesticide and rarely finds more than three possibilities. Confirmation is usually done by GC-MS under locked conditions so that all GC-MS retention times match the values listed in the pesticide retention time table. This was found to be of enormous benefit. Prior to GC-MS confirmation, the analyst already knows which pesticides to look for and their expected retention times. Alternatively, when there is adequate signal to quantitate the analyte in multiple AED element-selective chromatograms, it is often possible to confirm a pesticides identity simply by calculating its heteroatom ratio. GC-AED software for element ratioing facilitates this procedure.
easily confirmed at the expected retention time. In addition, the pesticides trichlorophenol, chlorothalonil, propoxur, and prochloraz were identified. Searching the Cl peak at about 6 minutes gave no pesticide hits. However, GC-MS suggested the presence of a trichloronaphthalene isomer at the corresponding retention time in the GC-MS chromatogram (about 12 minutes because the GC-MS was operated at normal speed). Though not a pesticide, trichloronaphthalene is considered to be a hazardous compound that should not be in food. The same green onion sample was then analyzed by the newer model GC-AED system (6890/ G2350A) at 3 speed (figure 8). Several more pesticides were identified by searching the pesticide/ endocrine disrupter table using a 0.1-minute retention time window. Table 3 lists the pesticide hits that were obtained for each retention time search using the available GC-AED data. Sulfur was not included in any of the searches
because onion extracts have such a high sulfur background. Confirmation by GC-MS was much easier because the GC-MS retention time for each pesticide hit was printed out with the RT search report. Thus, the retention times and probable identities of each pesticide were already known before the GC-MS analysis was run. As is shown in figure 7 for folpet, one can simply extract the ions characteristic for each pesticide hit and look in the extracted ion chromatogram at the expected retention time.
Quantitative Analysis
The Agilent pesticide screening method is a qualitative tool to identify any of the 567 pesticides and endocrine disrupters listed in the retention time table. This, of course, is the first step in any pesticide screening method. Quantitative analysis can be performed in one of two ways.
Analysis of a Green Onion Extract

Numerous samples of fruit and vegetable extracts have been analyzed using this methodology. The results for a green onion extract illustrate the versatility and potential of this method. Green onion extracts are usually very dirty and contain a large number of co-extracted sulfur compounds that can obscure sulfur-containing pesticides. The onion chromatograms shown in figure 6 were run under locked conditions at 2 speed in Tallahassee, Florida, by the Department of Food and Agriculture using a 5890 SERIES II/5921A GC-AED system. Retention time searching indicated that folpet was present in the sample, but it could not be confirmed at the time. The same sample was sent to the Agilent Technologies Little Falls Site in Wilmington, DE, where it was analyzed by scanning GC-MS using an 6890/5973 system. As shown in figure 7, folpet was
40 2,4,5-Trichlorophenol 30 20 10 35 25 Prochloraz 15 4 6 8 10 12 14 16 18 Propoxur Chlorothalonil Trichloronaphthalene Folpet
Cl
N
20
Figure 6. Cl- and N-selective chromatograms of a green onion extract from an 5890/5921A GC-AED system. The analysis was performed at 2 speed under locked conditions in Tallahassee, Florida, by the Department of Agriculture and Consumer Services. In addition to folpet, trichlorophenol, propoxur, and prochloraz were identified by retention time table searching and confirmed by GC-MS at their expected retention times. There were no hits for the Cl peak at about 6 minutes, which was identified by GC-MS as a trichloronaphthalene isomer.
The traditional method is to inject standards into the GC, GC-AED, or GC-MS system to determine response factors from which quantitative results are calculated by the ChemStation software. However, because the GC-AED elemental response is almost independent of molecular structure, compound-independent calibration (CIC) can be used to quantitate all of the pesticides and endocrine disrupters that are found. For example, one could spike chlorpyrifos-methyl (C7H7Cl3NO3PS) at a known concentration into each pesticide extract and obtain elementspecific calibration curves for Cl, N, P, and S. These curves could then be used to calibrate for any other compound containing one or more of these elements. Because the GC-AED is quite stable, external standard CIC often works just as well. The GC-AED software facilitates CIC. Unfortunately, this procedure determines the amount of a compound that reaches the AED and does not compensate for losses due to decomposition or adsorption in the inlet or column.
Green Onion
Folpet
4.00
8.00
12.00
16.00
20.00
24.00
28.00
M/Z 260, 294, 297 Folpet (21.637 min) Pesticide table RT = 21.594 min
21.20
21.60
22.00
22.40
22.80
23.20
Figure 7. Confirmation of folpet in a green onion extract. The tabulated GC-MS retention time is 21.594 minutes, and folpet was detected in this sample at 21.637 minutes by simply extracting its characteristic ions.
N 174
6 1
1. Dichlorvos 2. 2,4,5-Trichlorophenol 3. Propoxur 4. Trichloronaphthalene 5. Chlorothalonil 6. Chlorpyrifos-methyl 7. Folpet 8. Mirex 9. Prochloraz
Conclusions
Most screening procedures in use today are capable of finding only a fraction of the pesticides that are registered around the world. This new method has the capability of screening for virtually any volatile pesticide, metabolite, or endocrine disrupter. Although confirmation is usually required, GC-MS analysis is made much easier and more reliable because the pesticides retention time and probable identity are already known.
P 178
4 5
Cl 479
7 8 9 12 14
10
Figure 8. Element-selective chromatograms obtained for the same green onion extract shown in figure 6. These chromatograms were obtained at 3 speed using an 6890/G2350A GC-AED system.
While GC-AED is an ideal tool for element-selective pesticide screening,2022 many laboratories rely on a combination of other selective detectors. It is still possible to apply this method if each GC system runs the Agilent pesticide method under the same locked conditions. Any combination of GC-AED and/or elementselective detector response data can be entered into the RTL searching software. When combined with RTL and retention time searching, GC-AED and GC-MS provide the most comprehensive and reliable screening method available for pesticides, metabolites, and suspected endocrine disrupters. Unlike most target compound methods in use today, this procedure has a good chance of finding and identifying unexpected or unknown pesticides, even in complex food extracts. RTL software makes it easy to add more compounds to the method, simply by determining their retention times under the same locked conditions. Retention time locking with database searching could easily be applied to similar types of analyses. For example, one might use the procedure to identify polychlorinated biphenyls, polynuclear aromatics, drugs of abuse, or flavor and fragrance compounds.
Table 3.
Green onion pesticide "hits" obtained by searching the 567-compound pesticide/endocrine disrupter RT table using a 0.1-minute RT window and elementselective GC-AED data. Compounds confirmed by GC-MS are shown for comparison. The GC/MS (figure 7) and GC-AED (figure 8) chromatograms were obtained at normal and 3X speeds, respectively. Sulfur peaks were not used to narrow the search because of the high background of sulfur-containing compounds in the onion extract.
RT Search Hits Dichlorvos 2,4,6-Trichlorophenol 2,4,5-Trichlorophenol Fenobucarb Propoxur 4,6-Dinitro-o-cresol No pesticide hits Terbacil Chlorothalonil Chlorpyrifos-methyl Folpet Chlorbenside Mirex Prochloraz Confirmed by GC-MS Dichlorvos 2,4,5-Trichlorophenol Propoxur
GC-AED RT 1.933 2.281 3.440
3.854 4.955 5.538 7.232 9.965 10.588
Trichloronaphthalene isomer Chlorothalonil Chlorpyrifos-methyl Folpet Mirex Prochloraz
beta testing the method; Matthew Klee and Leonid Blumberg for many useful discussions; and James Green and Takeshi Otsuka for their help in developing the retention time table.
5. J. Fillion, R. Hindle, M. Lacroix, and J. Selwyn, J AOAC Int 78, 12521266, 1995. 6. Working Group on Development and Improvement of Residue-analytical Methods, Analytical Methods for Pesticide Residues in Foodstuffs, General Inspectorate for Health Protection, Ministry of Public Health, Welfare and Sport, The Netherlands, 1996. 7. M. A. Luke, J. E. Froberg, G. M. Doose, and H. T. Masumoto, J Assoc Off Anal Chem 64, 11871195, 1981. 8. H.-J. Stan (Ed.), Analysis of Pesticides in Ground and Surface Water II, Springer-Verlag, Berlin, Germany, 1995. 9. R. E. Wagner, W. Kotas, and G. A. Yogis (Eds.), Guide to Environmental Analytical Methods, 2nd Edition Genium, Schenectady, NY, 1994. 10. U.S. Environmental Protection Agency, Test Methods for Evaluating Solid Waste, SW-846, Draft Method 8085: Pesticides by GCAED.
References
1. C. Tomlin (Ed.), The Pesticide Manual, Eleventh Edition, British Crop Protection Council, Farnham, Surry, UK, 1997. 2. R. W. Miller, This is Codex Alimentarius, Secretariat of the Joint FAO/WHO Food Standards Programme, Food and Agriculture Organization of the United Nations, Rome. 3. B. M. McMahon and N. F. Hardin (Eds.), Pesticide Analytical Manual, Vol. I, Third Edition, U.S. Food and Drug Administration, Washington, D.C., 1994. 4. S. M. Lee, M. L. Papathakis, H.-M. C. Feng, G. G. Hunter, and J. E. Carr, Fresenius J Anal Chem 339, 376383, 1991.
Acknowledgments
The authors wish to thank the following people for their contributions to the development of this method: Joanne Cook and Marc Engel (Florida Department of Agriculture and Consumer Services) for supplying a green onion chromatogram, for contributing fruit and vegetable extracts, and for
10
11. T. Colborn, D. Dumanoski, and J. P. Myers, Our Stolen Future Penguin, New York, NY, 1996. 12. Food Quality Protection Act of 1996, Public Law 104-170, Congressional Record, pp. H8127H8141. 13. Safe Drinking Water Act Amendments of 1996, Public Law 104-182, Congressional Record pp. H9680H9700. 14. V. Giarrocco, B. Quimby, and M. Klee, Retention Time Locking: Concepts and Applications, Application Note 228-392, Publication (23) 5966-2469E, December 1997. 15. P. L. Wylie and B. D. Quimby, Abstracts of the 1st European Pesticide Residue Workshop, Alkmaar, The Netherlands,| Paper # O-023, June 1012, 1996. 16. M. Klee and T. Sullivan, Environ. Testing and Analysis, pp. 1617 and 4445, March/April 1998. 17. Hewlett Packard Company, Software for Method Translation, Available on the World Wide Web at: http://www.chem. agilent.com/cag/servsup/ usersoft/main.html. 18. M. L. Lee, F. J. Yang, and K. D. Bartle, Open Tubular Column Gas Chromatography, Theory and Practice, Wiley, New York, NY, pp. 217218, 1984. 19. C. Bicchi, A. DAmato, and A. Binello, J High Resol Chromatogr 19, 8084, 1996. 20. S. M. Lee and P. L. Wylie, J Agric Food Chem 39, 21922199, 1991. 21. P. L. Wylie and R. Oguchi, J Chrom 517, 131142, 1990. 22. N. L. Olson, R. Carrell, R. K. Cummings, and R. Rieck, LC-GC 12, 142154, 1994. 23. P. L. Wylie and K. Uchiyama, J AOAC Int 79, 571577, 1996.
24. P. L. Wylie, K. L. Klein, M. Q. Thompson, and B. W. Hermann, J High Resol Chromatog 15, 571577, 1992. 25. W. D. Snyder and L. M. Blumberg, Proceedings of the Fourteenth International Symposium on Capillary Chromatography, Baltimore, MD, pp. 2838, May 2529, 1992. 26. B. D. Quimby, L. M. Blumberg, M. S. Klee, and P. L. Wylie, Precise Time-Scaling of Gas Chromatographic Methods Using Method Translation and Retention Time Locking, Application Note 228401, Publication 5967-5820E, April 1998. 27. http://www.nijs.go.jp/hse/ environ/edsubs/substancesnew. html. 28. http://www.wwfcanada.org/ hormone-disruptors. 29. http://www.epa.gov/opptintr/ opptendo/index.htm. 30. http://www.osf-facts.org/ basics.chemlist.html. 31. http://www.mst.dk/liste.htm. 32. Hewlett-Packard Company. Mass Spectral Libraries: HP Pest library, Wiley 275 library, NBS75K Library, Palo Alto, CA. 33. Pesticide Reference Spectra, Vol. 14, Spectral Service Gmbh, Kln, Germany and Riedel-de Han AG, Seelze, Germany, 1992.
11
Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Microsoft and Windows NT are U.S. registered trademarks. HP is a registered trademark of Hewlett-Packard Company. Copyright 2000 Agilent Technologies, Inc. Printed in the USA 3/2000 5967-5860E
The Analysis of Chlorinated Pesticides and PCBs Using the HP-608 Capillary Column
Application Note 228-236
Authors
Imogene L. Chang, PhD Winfred J. Sanders, PhD
Abstract
Chlorinated pesticides and PCBs targeted in EPA Methods 608, 8080, 8081, and CLP pesticides for wastewater and solid wastes are analyzed under optimum conditions at a constant flow of 2.4 ml/min. The merits of splitless and on-column injection techniques using the Agilent 5890 Series II GC with electronic pressure control (EPC) are compared. Key Words: chlorinated pesticides, PCBs, on-column injection, splitless injection, HP-608 capillary column, EPA 608, EPA 8080/8081, CLP pesticides, electronic pressure control.
These EPA Methods allow laboratories to substitute columns of their choice provided that performance data such as chromatographic resolution, analyte breakdown, and MDLs (minimum detectable levels) are equal to or better than those provided with the EPA methods. The HP-608 is a wide bore (530 m-id) capillary column specially designed for the analysis of organic pesticides. GC/ECD separations of chlorinated pesticides and PCBs were done using the HP-608 column with both on-column and splitless inlet sample introductions. In both cases, the HP-608 provided superior chromatographic resolution, excellent reproducibility, and minimal analyte breakdown for the analysis of pesticides and PCBs. Table 1. Experimental Conditions
Instrument Requirements Gas chromatograph: Injection ports: Column: Detector: Sample introduction: Data collection: Experimental Conditions Column: Carrier gas: Oven:
Experimental
A 30 m x 530 m x 0.5 m HP-608 column (part no. 19095S-023) was used under constant carrier gas flow using the 5890 Series II GC with EPC equipped with a split/splitless inlet and a cool on-column inlet. Equipment included the 7673 automatic sampler with tray and the electron capture detector (ECD). Samples were introduced in both the on-column and splitless modes. The MerlinTM Microseal septum (part no. 5181-8816) was used in the split/splitless inlet to replace the conventional inlet septum. A deactivated tapered glass liner (part no. 51813316) was used for all splitless injection runs. GC conditions were controlled using the HP 3365
Introduction
Chlorinated pesticides and PCBs have been banned in the U.S. for several years. However, because of their persistence in the environment, EPA methods 8080/8081 and CLP pesticides target 16 to 20 chlorinated organic pesticides in the evaluation of solid waste. This includes pesticides, their degradation products, technical grades of chlordane, toxaphene, and PCBs in solid waste.1,2 EPA Method 608 targets similar pesticides in industrial and wastewater discharges.3 EPA Methods 608 and 8080 prescribe packed-column analysis, whereas Methods 8081 and CLP pesticides prescribe capillary column analysis.
Agilent 5890 Series II with EPC Split/splitless inlet with temperature and pressure programmable features On-column inlet with temperature and pressure programmable features HP-608, 30 m x 530 m x 0.5 m (Part number 19095S-023) ECD 7673 splitless fast injection On-column injection 3365 ChemStation and HP Vectra 486/133T HP-608, 30 m x 530 m x 0.5 m (Part number 19095S-023) He, 20 cm/sec, 2.2 psi at 80C with EPC under constant flow of 2.4 ml/min First ramp: 80C (hold 1 min) to 190C at 30C/min Second ramp: 190C to 280C (hold 1 min) at 6C/min Third ramp: 280C to 300C (hold 2 min) at 20C/min Splitless: 1 l, inlet temperature of 250C On-column: 1 l oven track for inlet temperature program ECD (330C), 65 ml/min N2 makeup, 6 ml/min anode purge Pesticides and PCB standard solutions in isooctane
Injection: Detector: Sample:
ChemStation. Data was managed with a HP Vectra PC (486/33T). Instrument parameters and experimental conditions are listed in Table 1. Pesticide solutions containing 16 to 22 components were prepared from the dilution of certified standards (part no. 8500-5873 and 8500-5876, mixes A and B: level 2) with isooctane (pesticide residue grade from Burdick & Jackson). Pesticide standards (part no. 5062-3589), including four vials of 16 EPA-608 pesticides and two vials of two component inlet check solutions (endrin/DDT concentrations are 50 ppb/100 ppb), were used without further dilution. These pesticide compounds are listed in Table 2.
Figure 1.
1.6e5 1.4e5 1.2e5 1.0e5 8.0e4 6.0e4 4.0e4 2.0e4 0 6 1.6e5 1.4e5 1.2e5 1.0e5 8.0e4 6.0e4 4.0e4 2.0e4 0 6
Chromatograms of the 16 chlorinated pesticides under optimum GC conditions, 100 pg of each pesticide injected. Peak identification in Table 2.
1 2 5 6 4 3 7 10 11 12 15 14 17 18 13 16
Figure 1A. Splitless Injection
a 8 1 10 12 14 16 18 20 min
Figure 1B. On-Column Injection

2 5 6 4 7 11 10 12 13 3 16
15 14 17 18
a 8 10 12 14 16 18 20 min
Table 2. Chlorinated Pesticides

Peak No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 SS1 SS2 EPA-608 alpha-BHC Lindane beta-BHC Heptachlor delta-BHC Aldrin Heptachlor epoxide Compound Name EPA-8080/8081 alpha-BHC Lindane beta-BHC Heptachlor delta-BHC Aldrin Heptachlor epoxide Chlordane-gamma Endosulfan I 4,4-DDE Dieldrin Endrin 4,4-DDD Endosulfan II 4,4-DDT Endrin aldehyde Endosulfan sulfate Methoxychlor EPA-CLP Pesticides alpha-BHC Lindane beta-BHC Heptachlor delta-BHC Aldrin Heptachlor epoxide Chlordane-gamma Chlordane-alpha Endosulfan I 4,4-DDE Dieldrin Endrin 4,4-DDD Endosulfan II 4,4-DDT Endrin aldehyde Endosulfan sulfate Methoxychlor Endrin ketone Tetrachloro-m-xylene Decachlorobiphenyl

Splitless Analysis
Figure 1A shows the analysis of a standard solution containing the 16 EPA-608 targeted pesticides at a constant column flow of 2.4 ml/minute. One microliter of sample (100 pg of each component) was introduced in splitless mode at 250C under the conditions4 listed in Table 1. All 16 components were well resolved in sharp symmetric peaks, and the analysis was completed in less than 17 minutes. The 30-m HP-608 (530 m id) column possesses sufficient efficiency to completely resolve the complex pesticides mix, including chlorinated compounds with similar or isomeric structures. The absence of coeluting peaks on the HP-608 column permitted fast and accurate identification and quantitation.
Endosulfan I 4,4-DDE Dieldrin Endrin 4,4-DDD Endosulfan II 4,4-DDT Endrin aldehyde Endosulfan sulfate a-Degradation product
Low-Temperature On-Column Analysis

Figure 1B shows the same pesticides standard mix using the cool on-column injection technique. On-column injection of 1 l of sample at 80C resulted in little sample degradation, minimal byproducts, and good sensitivity (see Table 3). Common to both Figures 1A and 1B is the absence of tailing peaks, including the endrin aldehyde peak (peak 17), indicating the HP-608 column surface is very inert.
Table 3. Reproducibility of Pesticide Analysis

Retention Times, min Pesticides alpha-BHC Lindane beta-BHC Heptachlor delta-BHC Aldrin Endosulfan I 4,4-DDE Dieldrin Endrin 4,4-DDD Endosulfan II 4,4-DDT Endrin aldehyde Endosulfan sulfate a, Degradation product alpha-BHC Lindane beta-BHC Heptachlor delta-BHC Aldrin Endosulfan I 4,4-DDE Dieldrin Endrin 4,4-DDD Endosulfan II 4,4-DDT Endrin aldehyde Endosulfan sulfate a, Degradation product Mean 8.423 9.225 9.352 9.984 10.181 10.760 13.036 13.623 13.838 14.814 15.135 15.311 15.975 16.208 16.570 18.690 Std Dev 0.004 0.004 0.004 0.004 0.005 0.004 0.003 0.004 0.004 0.004 0.004 0.004 0.004 0.003 0.004 0.003 0.003 % RSD 0.047 0.046 0.046 0.042 0.044 0.039 0.028 0.031 0.026 0.027 0.025 0.024 0.025 0.021 0.022 0.021 0.017 Mean 431643 393514 208287 310294 390027 359246 359586 321622 341930 336042 268560 254389 297580 259369 205588 281397 3416 A. On-column injection (100 pg each component) 7497 6496 3428 5430 7428 6996 5740 5478 7070 4832 5298 3017 4326 3881 1876 4143 97 1.74 1.65 1.65 1.75 1.90 1.95 1.60 1.70 2.07 1.44 1.97 1.19 1.45 1.50 0.91 1.47 2.83 Area Counts Std Dev % RSD
Heptachlor epoxide 12.385
Reproducibility
Reproducibility for the analysis of chlorinated pesticides using HP-608 columns with the HP GC/ECD system was excellent (see Table 3). The RSD (relative standard deviation) in absolute area counts for all 16 EPA targeted pesticides was less than 2% for on-column runs (two sets of six replicate injections). Similarly, the peak area counts reproducibility for all splitless injection runs (three sets of six replicate injections) was in the 1% to 2% RSD range using the same standard sample. The standard deviation of retention times was within 0.0030.005 minutes and 0.002 minutes for on-column and splitless runs, respectively. In comparison, the standard deviation of retention times for EPA Method 8081 analysis (Table 10, reference 1) using wide-bore capillary columns ranged from 0.007 minutes to 0.013 minutes for the same set of pesticides. This clearly demonstrates that chromatographic reproducibility obtained using the HP-608 capillary column is better than that obtained using the capillary columns stipulated in EPA Method 8081.
B. Splitless injection (100 pg each component) 8.351 9.146 9.273 9.898 10.097 10.671 12.938 13.527 13.735 14.710 15.034 15.207 15.874 16.103 16.467 18.584 0.002 0.002 0.002 0.002 0.001 0.002 0.001 0.002 0.001 0.002 0.002 0.002 0.002 0.002 0.002 0.002 0.002 0.020 0.020 0.018 0.018 0.013 0.015 0.011 0.014 0.011 0.014 0.013 0.013 0.015 0.012 0.010 0.013 0.012 376446 317405 165105 207924 301779 308689 289985 253489 313249 209054 160235 168113 228810 168810 148655 190284 21513 7222 6592 3129 4637 6113 6422 6216 5496 6102 3925 3104 3094 4868 2129 3687 3003 1747 1.92 2.08 1.90 2.23 2.03 2.08 2.14 2.17 1.95 1.88 1.94 1.84 2.13 1.26 2.48 1.58 8.12
Heptachlor epoxide 12.289
Comparison of Sample Introduction Techniques

For all on-column injection runs, degradation was negligible due to the low initial column temperature (80C) and the direct introduction of a liquid sample plug into an inert column. As a result, inlet-related sample discrimination, alteration, and degradation were eliminated, while the advantages of solvent focusing and stationary phase focusing were maximized. Routine analysis of the inlet check solution (specified by the EPA methods) showed that the average degradation was less than 3% for endrin and 1% for DDT. As demonstrated by the clean baseline in Figure 1A, little sample degradation occurred at an inlet temperature of 250C. However, a small endrin ketone peak (RT of 18.69 minutes) appeared on the chromatograms from the GC runs with both on-column and splitless injection shown in Figures 1A and 1B. A closer look (Table 3), shows that the area counts for endrin ketone (peak a, a byproduct of endrin degradation) measured 5 times larger in the splitless runs than for the on-column runs (average absolute area counts of 3,400 versus 21,000). The GC runs of the inlet check standard (after 200 repeated splitless injections), showed a 7% endrin degradation and 10% DDT degradation. These values were well below the EPA requirement of 15% degradation for both endrin and DDT. Use of the MerlinTM Microseal5 and the deactivated glass liner also contributed directly to the low degradation rate in the splitless mode. The Microseal is designed to provide a good inlet seal without using a conventional septum. By eliminating the introduction of particulates into the inlet liner from conventional septum, useful life for the inlet liner is extended, down time (to change a liner and a conventional septum) is reduced, and laboratory throughput is increased. The use of splitless injection technique may also prevent interference from extraneous and high boiling
Figure 2.
2.2e4 2.0e4
Chromatograms of isooctane under optimum GC conditions, 1 l injected. (b,k=solvent contaminants)
Figure 2A. Splitless Injection

1.8e4 1.6e4 1.4e4 1.2e4 1.0e4 8000 6000 4000 6 2.2e4 2.0e4 8 10 12 14 16 18 20 min b b
Figure 2B. On-Column Injection

1.8e4 1.6e4 1.4e4 1.2e4 1.0e4 k 8000 6000 4000 6 8 10 12 14 16 18 20 min b b b b
materials in dirty samples. This is demonstrated in Figures 2A and 2B. Figure 2 shows the analysis of isooctane solvent (pesticide-residue grade) using both splitless (Figure 2A) and on-column injection (Figure 2B). The late-eluting peak (peak k) , at 16.69 minutes retention time in the on-column run, does not appear in the chromato-gram of the splitless run (Figure 2A). This peak, possibly a high boiling contaminant in isooctane, appears again in Figure 3B. Figures 3A and 3B show analyses of a 10-ppb pesticide standard using splitless injection and on-column injection, respectively. The peak (peak k) eluting just before endosulfan sulfate
(peak 18) may cause a higher value for the determination of trace endosulfan sulfate in the sample. Both area counts and peak heights for the splitless runs were smaller than those for the on-column injection runs (see Table 3). For example, the average counts of lindane from the splitless runs were approximately 80% of those from the on-column injections (Table 3). Therefore, oncolumn injection is a good choice for clean samples and trace analyses demanding high sensitivity and low detection limits (large area counts).
Analysis of PCBs and EPA Methods 8080, 8081, and CLP Pesticides
For wastewater and solid waste samples, the EPA recommends splitless injection for the determination of pesticides and PCBs. Using splitless injection under optimum 5890 Series II GC conditions, all 17 pesticides targeted by EPA Method 8080B are resolved as shown in Figure 4. Among the 20 components targeted by EPA Methods 8081 and CLP pesticides, all but alpha-chlordane and endosulfan I (they are partially separated) are well resolved by the HP-608 column (Figure 5). Since the HP-608 column can effectively separate the complex mix of these pesticides, it is a good column choice for the determination of PCBs and multiple-peak response pesticides such as chlordane and toxaphene. Figure 6 shows a comparison of chromatograms for technical grade chlordane and toxaphene, while Figure 7 is a comparison of chromatograms for seven PCBs, all analyzed under the same GC conditions using the HP-608 capillary column.
Figure 3.
2.2e4 2.0e4
Chromatograms of dilute pesticides mix under optimum GC conditions; 10 pg of each pesticide injected. (Peak ID, see Table 2)
Figure 3A. 1 Splitless Injection

1.8e4 1 1.6e4 1.4e4 1.2e4 3 1.0e4 8000 a 6000 4000 6 2.2e4 1.0e4 1.2e4 1.6e4 1.4e4 1.2e4 1.0e4 8000 6000 4000 6 8 10 12 14 16 18 20 min k a 3 8 1 2 5 4 6 7 10 11 14 12 13 15 16 18 17 10 12 14 16 18 20 min 4 2 5 6 7 10 11 12 13 15 14 16 18 17
Figure 3B. 1 On-Column Injection
Figure 4.
Chromatograms of the EPA-Method 8080 pesticides under optimum GC conditions. Splitless injection of 100200 pg per component. (Peak ID, see Table 2)
1 2 5 6 7 10 4 3 13 12 14 15 18 16 17 19
1.2e5
11
a 0 6 8 10 12 14 16 18 20 min
Conclusion
Under optimal conditions, the HP-608 column separates 16 EPA-608 pesticides in 17 minutes and 20 EPA-CLP pesticides (and EPA-8081 pesticides) in 19 minutes (22 minutes including the surrogate, decachlorobiphenyl). Both splitless and on-column injections yield little sample degradation and provide excellent reproducibility of retention times and area responses. On-column injection is more suitable for clean samples and trace analysis, while splitless injection is better used for wastewater and waste samples.
Figure 5.
Chromatogram of pesticides targeted in EPA-method 8081 and CLP pesticides under optimum GC conditions. Splitless injection of 50100 pg per component. (Peak ID, see Table 2)
SS1 11 19 SS2
1.2e5
1.0e5 1 8.0e4 12 10 5 6.0e4 4 4.0e4 3 6 7 9 8 13 15 20 14 16 17 18
2.0e4
0 5 10 15 20 min
Figure 6.
Chromatogram of technical grade toxaphene and chlordane under optimum GC conditions. Splitless injection of 1 l 2.5 ppm mix
2.0e5
Chlordane
1.5e5
1.0e5
6.0e5
0 5 1.4e5 1.2e5 1.0e5 8.0e5 6.0e5 4.0e5 2.0e5 0 5 10 15 20 25 min Toxaphene 10 15 20 25 min
Figure 7.
1.4e5
A comparison of seven PCBs under optimum GC conditions. Splitless injection of 1 l 2.5 ppm each
1.6e5 Aroclor 1016 Aroclor 1248
0 6 2.0e5 Aroclor 1221 8 10 12 14 16 18 20 min
0 6 1.6e5 Aroclor 1254 8 10 12 14 16 18 20 min
0 6 70000 Aroclor 1232 8 10 12 14 16 18 20 min
0 6 1.6e5 Aroclor 1260 8 10 12 14 16 18 20 min
0 6 1.6e5 Aroclor 1242 8 10 12 14 16 18 20 min
0 6 8 10 12 14 16 18 20 min
0 6 8 10 12 14 16 18 20 min
Acknowledgment
The authors wish to thank Dr. D. Pautler for his many helpful discussions.
References
1. EPA Method 8080B and 8081, Test Methods for Evaluating Solid Waste, SW-846, Revision 2, Nov. 1990. 2. U.S. EPA Contract Laboratory Program Statement of Work for Organics Analysis Document Number OLM01.0, 1990. 3. EPA Method 608, Methods for Organic Chemical Analysis of Municipal and Industrial Wastewater, PB82-201798, 1982. 4. I. L. Chang and W. J. Sanders, Method Development for EPA608 Analysis Using a HP-608 Capillary Column, HewlettPackard Application Note 1993 (in preparation). 5. Introducing the Merlin MicrosealTM Septum, Pub. No. 5091-3197EUS, 1991.
www.agilent.com
Copyright 1993, 2000 Agilent Technologies Printed in USA 3/00 5091-7567E
Fast Screening of Pesticide and Endocrine Disrupters Using the Agilent 6890/5973N GC/MSD System, Part I
Application
Gas Chromatography January 2000 retention times and characteristic ions for 567 of the most common pesticides and endocrine disrupters of concern worldwide. A GC/MSD method was developed based on the standard 42-min method1 to screen for all 567 of the most common analytes. A specific combination of column stationary phase, carrier gas flow rate, and oven temperature programming is required to lock all the compounds to an expected retention timetable2. Compound identification based only on spectral searching alone is difficult when analyzing extracts containing significant sample matrix content because of overlapping peaks and noisy baselines. The new screening tool, integrated within Agilents ChemStation for MSD, searches for all 567 compounds by first checking and integrating four characteristic ions within the expected time window, and second by printing out a report showing hits and possible hits (ratios of characteristic ions that do not match the expected values in the library within specified limits). In one application, the analysis time of the standard pesticide method was reduced by one half, two-thirds, and three-fourths. The faster methods were scaled exactly as predicted by using a combination of Agilents method translation (MTL) and RTL software. Because scaling was exact, these faster methods can be used with precisely-scaled pesticide libraries, making the screening process even more powerful and adaptable to individual needs.
Authors C. Kai Meng Agilent Technologies 2850 Centerville Road Wilmington, Delaware 19808-1610 USA Michael Szelewski Agilent Technologies 2850 Centerville Road Wilmington, Delaware 19808-1610 USA
Abstract Agilent Technologies new, fast GC/MSD method can significantly speed up the screening of pesticides. Agilents GC method translation software (available free from the Agilent Technologies Web site, http://www. chem. agilent.com/cag/ servsup/usersoft/main.html#mxlator) was used in developing the new method based on the standard 42-min method. A 10 m x 0.1 mm x 0.1 m HP-5 column was used to increase analysis speed up to fourfold. The time savings were implemented in increments (down to 10.5 minutes) to verify the predictability of scaling and the effect of scaling on the signal-tonoise ratio. Key Words RTL, pesticide, environmental, screening, fast GC, method translation, 5973, 6890, MTL Introduction Analysts want faster analyses to improve laboratory productivity. Often, when speeding up GC methods, an analyst will trade resolution for increased analysis speed. This loss of resolution can complicate peak identification, even with a mass selective detector (MSD). Agilent Technologies has developed new techniques to solve the peak identification problem based on Agilents retention time locking (RTL) software and a new mass spectral library that contains the locked
Experimental The GC method translation software tool was used to find operating conditions for the faster methods. Figure 1 is a screen capture of MTL software data entry showing the original conditions and the new chromatographic conditions for a twofold speed gain. The column flow rate, which is helpful to avoid exceeding MSD pumping capacity3, is also found in the table. A 16:1 split ratio was suggested in the table as a proportional scaling from the original column to the smaller i.d. column with corresponding lower capacity. The program also determined the required column head pressure and corresponding oven ramp. The Agilent 6890 GC fast oven option (220/240V in the U.S.) was required for the faster oven ramp used in this study.
Figure 1. Screen capture showing the method translation (MTL) software data entry used in a twofold speed gain translation.
General chromatographic conditions are listed in table 1. The standard used was a mixture of 26 pesticides at 10 ppm. A 10 m x 0.1 mm x 0.1 m HP-5 column (part number 19091J141) was used. The head pressure determined by the method translation software (30.72 psi) was used as the starting point for retention time locking. The column head pressure required to lock retention times of the compounds to the library (the original retention time divided by 2) was determined using the automated RTL process integrated within the Agilent ChemStation for MSD. This process (first translate the method then lock the retention times) was repeated for the threefold and fourfold time reductions.
Table 1. Chromatographic Conditions Speed GC Column Injection mode Column head pressure Column flow (mL/min) Inlet control mode Carrier gas Injector temperature Oven temperature Ramp 1 Ramp 2 Ramp 3 Oven equilibration Injection volume Liner MS Conditions Solvent delay Tune file Low mass High mass Threshold Sampling Scans/sec Quad temperature Source temperature Transfer line temperature Acquisition mode Onefold (1X) 110 V 30 m x 0.25 mm x 0.25 m HP5-MS (P/N 19091S-433) Splitless 18.0 psi 1.5 Constant pressure Helium 250 C 70 (2 min) 25 C/min 150 (0 min) 3 C/min 200 (0 min) 8 C/min 280 (10 min) 2 min 1 L 5183-4647 Twofold (2X) Threefold (3X) 220/240 V 10 m x 0.1 mm x 0.1 m HP-5 (P/N 19091J-141) 16:1 split 36.55 psi 63.17 psi 0.4 0.8 Constant pressure Helium 250 C 70 (1 min) 70 (0.67 min) 50 75 150 (0 min) 150 (0 min) 6 9 200 (0 min) 200 (0 min) 16 24 280 (5 min) 280 (3.33 min) 2 min 1 L 5183-4647 Fourfold (4X)
90.0 psi 1.5
70 (0.5 min) 100 150 (0 min) 12 200 (0 min) 32 280 (2.5 min)
3 min Atune.u 35 amu 500 amu 150 2 3.15 150 C 230 C 280 C Scan (EI)
1.8 min Atune.u 35 amu 450 amu 250 2 3.50 150 C 230 C 280 C Scan (EI)
1.2 min
0.9 min
1 6.54
1 6.54
Figure 2 shows the results of the shortened analysis times. The three chromatograms look extremely similar, except that the time axis is scaled proportionally. Because MTL followed by RTL scales methods very precisely, scaled screening libraries for corresponding time reductions can be obtained by dividing the retention times in the library by the speed gain (which does not have to be an integer). The peak heights from all the methods are very similar. Although the sample was split 16:1 for the smaller column, the small column i.d. and faster oven ramp combination made the peaks narrower and higher, so there was minimal loss in the signal to noise ratio. Conclusion The highly accurate and reproducible pressure and temperature control of the Agilent 6890 GC allows precise scaling of the standard 42-min GC/MSD pesticide method. Run time was shortened to 10.5 minutes using a fast oven ramp rate and a 10-meter 100-micron column. The combination of MTL and RTL facilitated scaling and yielded exact scaling. RTL libraries can accurately be scaled to correspond to the faster analyses. References 1. B. D. Quimby, L.M. Blumberg, M. S. Klee, and P. L. Wylie, Precise TimeScaling of Gas Chromatographic Methods Using Method Translation and Retention Time Locking, Application Note 228-401, Agilent publication number 5967-5820E, May 1998. 2. H. Prest, P. L. Wylie, K. Weiner, and D. Agnew, Efficient Screening for Pesticides and Endocrine Disrupters Using the HP 6890/ 5973 GC/MSD System, Agilent publication number 5968-4884E, April 1999. 3. H. Prest, GC Column Selection and Pumping Considerations for Electron and Chemical Ionization MSD operation, Agilent publication Number 5968-7958E, November 1999.
Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Copyright 2000 Agilent Technologies Printed in USA 2/00 5968-9220E
Abundance 2,400,000 2,200,000 2,000,000 1,800,000 1,600,000 1,400,000 1,200,000 1,000,000 800,000 600,000 400,000 200,000 Time 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 20.00
2X Speed
18 min
Abundance 6,000,000 5,500,000 5,000,000 4,500,000 4,000,000 3,500,000 3,000,000 2,500,000 2,000,000 1,500,000 1,000,000 500,000 Time 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 11.00 12.00 13.00
3X Speed
Abundance 6,000,000 5,500,000 5,000,000 4,500,000 4,000,000 3,500,000 3,000,000 2,500,000 2,000,000 1,500,000 1,000,000 500,000 Time 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00
12 min
4X Speed
9.00
10.00
9 min
Figure 2. Three TICs of the 2X, 3X, and 4X speedups. The standard analysis (1X) was 42 minutes long. The two vertical lines on the figure are used as references to show the similarity of the TICs.
Analysis of Glyphosate in Water with Postcolumn Derivatization using HPLC

Rainer Schuster Food
Abstract The HPLC method presented here was used for the direct analysis of glyphosate in water with postcolumn derivatization. HPLC method performance Limit of detection 1ppb Repeatability of RT over 10 runs <0.8 % of areas over 10 runs <2.2 %
Conditions
Column 150 4 mm cation exchange, K+ form from Pickering, 8 m Mobile phase A = 5 mM KH2PO4, pH = 2.0, B = 5 mM KOH Flow rate 0.4 ml/min Gradient at 15 min 0% B; at 17 min 100% B Column compartment 55 C Injection vol 50 l standard Fluorescence detector Excitation wavelength: 230 nm or 330 nm, Emission wavelength: 425 nm Slit width excitation: 2 mm (25 nm) Slit width emission 1: 4 mm (50 nm) Slit width emission 2: 4 mm (50 nm) Photomultiplier gain: 2 Cut-off filter: 370 nm Lamp: 55 Hz (always on) Response time: 4 s Derivatization reagent pump flow rate for hydrolization agent: 0.3 ml/min (OCl-) flow rate for derivatization agent: 0.3 ml/min (OPA). Sample preparation None
Norm. 7.5 7
Glyphosate AMPA
6.5 6 5.5 5 2.5 5 7.5 10 12.5 Time [min] 15 17.5 20
Figure 1 EIC traces from amine standards
Innovating the HP Way
Equipment
Agilent 1100 Series degasser quaternary pump autosampler Pickering post-column derivatization system fluorescence detector, Agilent ChemStation + software
Rainer Schuster is application chemist at Agilent Technologies, Waldbronn, Germany. For more information on our products and services, visit our worldwide website at http://www.agilent.com/chem
Copyright 1997 Agilent Technologies Released 09/97 Publication Number 5966-0743E
Analysis of Pesticides in Salad Samples and Spices using HPLC

Abstract The following compound classes of pesticides have been analyzed: triazines, phenylureaherbicides, methabenzthiazuron, diquat, paraquat, and mercaptobenzothiazol. Carbamates and glyphosate also have been analyzed but with different equipment. In most countries, growing concern about the residues of pesticides in food products is evident. Therefore, regulations limiting the concentration of pesticides in foodstuffs have been introduced to protect consumers from contaminated food products. Several methods are used to control these limits. HPLC is recommended for the analysis of low volatile compounds and for compounds that are unstable when heated.
Conditions
mAU 120
80
3 different salad samples
Vinclozolin
Folpet 40 0 10 15
20 25 Time [min] 30
Carbendazim*
35
40
Figure 1 Analysis of pesticide residues in three different salad samples * Carbendazim has a low recovery rate of only approximately 40 %
Column 100 3 mm Hypersil BDS, 3 m Mobile phase water/ACN (95:5) Gradient at 10 min 25% ACN at 26 min 42% ACN; at 34 min 60% ACN Flushing time 10 min at 100% ACN Post time 6 min Flow rate 0.5 ml/min Oven temp 42 C Injection vol 310 l Detector UV-DAD detection wavelengths 214/15 nm, 230/20 nm, and 245/20 nm reference wavelength 400/80 nm Sample preparation Salad was homogenized and then extracted with liquid/liquid extraction. The extract was cleaned with gel permeation chromatography using cyclohexane/ethyl acetate. Spices were prepared according to Specht 22 with gel permeation chromatography.
Sample preparation Sample preparation and enrichment depend strongly on the matrix. Drinking water samples, for example, must be extracted using solidphase extraction, whereas vegetables are extracted with liquid/liquid extraction after homogenization, followed by additional cleaning and sample enrichment. Chromatographic conditions The HPLC method presented here was used for the analysis of pesticides in salad samples and spices.1, 2 HPLC method performance Limit of detection 0.01 g/l Repeatability of RT over 10 runs <0.2 % of areas over 10 runs <1 %
Equipment
Agilent 1100 Series degasser quaternary pump autosampler thermostatted column compartment diode array detector, Agilent ChemStation + software
mAU 100
80 60
Vinclozolin Procymidon Nitro compounds Chlorpyripho-ethyl Paprika (Spain)
40 20 0 10 20 30 Time [min] 40
Paprika (Turkey)
50
Figure 2 Analysis of pesticide residues in two paprika samples
Pesticides & Residues Antibacterial Drug Residues
Detection, Confirmation, and Quantification of Chloramphenicol in Honey, Shrimp and Chicken Using the Agilent 6410 LC/MS Triple Quadrupole
Application
Food Safety
Authors
Yanyan Fang Agilent Technologies (Shanghai), Co. 412 Yin Lun Road 200131 China Jerry Zweigenbaum Agilent Technologies, Inc. 2850 Centerville Road Wilimington DE 19808 USA Zhuwei Wang Agilent Technologies (China) Beijing
meet the EU identification points needed for confirmation. This study is a valuable indicator of the ability of the QQQ for routine quantitative trace analysis of chloramphenicol in honey, shrimp, and chicken.
Introduction
Chloramphenicol (CAP) is a broad-spectrum antibiotic. It was concluded that human exposure to CAP can cause aplastic anemia [1]. Chloramphenicol and other bacterial inhibitors have arguably been the biggest issue facing international seafood trade over the past year. Because chloramphenicol has displayed significant toxicological effects on humans, it has been banned from foods in the European community, Japan, and the United States at levels of 0.3 ppb. LC/MS has been demonstrated for this analysis by the U.S. Food and Drug Administration[2-4] and others[5]. In this application, the Agilent G6410AA QQQ is used. This method employs negative ion mode with electrospray ionization. An internal standard (IS), CAP-d5, is added at the beginning of the extraction. The use of this IS self-corrects for any extraction variability from sample to sample and response variability caused by the matrix. With the use of this IS, 50 parts per trillion (ppt) CAP levels can be reliably quantified. A solid phase extraction (SPE) procedure is used along with a mobile phase of only methanol and water without salt buffers, which should help minimize MS maintenance.
Abstract
This method was developed using the Agilent G6410AA Triple Quadrupole Mass Spectrometer (QQQ) for chloramphenicol in honey, shrimp, and chicken. The sensitivity obtained exceeds the minimum required performance level (MRPL) established by the European Union regulation for food monitoring programs. Using a deuterated internal standard and one simple sample solid phase extraction (SPE) procedure can provide a limit of detection at 10 ppt in sample matrix. The analytical performance of the method was evaluated for three different matrixes and the results show little or no matrix effects. Linearity of response over 2 orders of magnitude was demonstrated (r > 0.99). In addition, good reproducibility of the two required product ion ratios was obtained to
Experimental
Reagents and Materials Agilent AccuBond SPE ENV PS DVB Cartridges (P/N 188-3060) Ethyl acetate from Burdick and Jackson (Morristown, NJ) Methanol HPLC-Grade from Burdick and Jackson Water (18 MW) from Milli-Q Synthesis System Chloramphenicol (CAP) from Aldrich Chemical Co. (Milwaulkee, WI) Deuterated (d5) CAP internal standard from Cambridge Isotope Laboratories (CIL, Andover, MA, U.S.) Syringe filter (0.2 m, PTFE) from Agilent (P/N 5185-5843) Overview of Method Internal Standard Preparation 1. A 100-g/mL (100 ppm) stock standard CAP-d5 solution in methanol (MeOH) is purchased from Cambridge Isotope Laboratories, Inc. (Lot SCCE-005) 2. A 1:100 dilution in MeOH of the stock standard gives an intermediate standard concentration of 1 g/mL (1 ppm) or 1000 ng/mL CAP-d5 3. A 1:100 dilution in MeOH gives a diluent solution (This diluent solution is used to prepare the samples) concentration of 10 ppb. 4. Every 1-g sample is fortified with 25 L of CAP-d5 diluent solution for a 0.25 ppb IS (internal standard) concentration Standard Solution Preparation 1. A 100-g/mL stock standard CAP solution in methanol (MeOH) is prepared by weighing 5.0 mg CAP std into 50 mL methanol. 2. A 1:100 dilution with methanol of the stock standard gives an intermediate standard concentration of 1 g/mL (1 ppm) or 1000 ng/mL CAP 3. Add 25 L CAP-d5 diluent solution into each vial. 4. Prepare standard solutions in these vials: 1 ppb, 0.2 ppb, 0.1 ppb, 0.02 ppb, and 0.01 ppb, with IS at 0.25 ppb level.
Sample Preparation All SPE cartridges are conditioned with 2 mL of water before use. 1. Honey, 1 g of sample is diluted to 5 mL with water and 25 L 10 ppb IS is added. The solution is loaded onto the SPE cartridge and allowed to stand for 5 min. Elution is performed with 10 mL ethyl acetate. The eluate is collected and the solvent is evaporated under a nitrogen stream at 40 C. The residue is redissolved in 1 mL methanol and put in an ultrasonic bath for 1 min. The solution is filtered, using a syringe filter, before injection. No additional clean-up of the sample solution is performed. 2. Shrimp, 1 g of shrimp is defrosted and mixed in a blender. To the 1 g of the mixed shrimp, 3 mL of water and 25 L 10 ppb IS is added. The portion is centrifuged for 5 min (8,000 rpm). The supernatant is loaded on the cartridge and allowed to stand for 5 min. Elution is performed with 5 mL ethyl acetate. The eluate is collected and the solvent evaporated under a nitrogen stream at 40 C. The residue is redissolved in 1 mL methanol and put in an ultrasonic bath for 1 min; the solution is filtered before injection. 3. Chicken, 1 g of chicken is defrosted and mixed in a blender. To the 1 g of the mixed chicken, 3 mL of water and 25 L 10 ppb IS is added. The portion is centrifuged for 5 min (8,000 rpm). The supernatant is loaded on the cartridge and allowed to stand for 5 min. Elution is performed with 5 mL ethyl acetate. The eluate is collected and the solvent evaporated under a nitrogen stream at 40 C. The residue is redissolved in 1 mL methanol and put in an ultrasonic bath for 1 min.; the solution is filtered before injection. LC/MS conditions The LC system was the Agilent 1200-SL binary pump with the ALS-SL autosampler. The MS was an Agilent 6410 LC/MS triple quadrupole mass spectrometer. See Table 1 for conditions.
Table1. HPLC Column
LC/MS Conditions ZORBAX SB-C18, 2.1 50 mm, 1.8 m (p/n 827700-902) 0.4 mL/min A: water B: methanol 0-5 min, 30~70% B 5-6 min, 70~100% B 8 min, 100% B 4 min 45 C 5 L ESI Negative 350 C 10 L/min 45 psi 3500 V 100 V 10 V for m/z 257(qualifier ion) 15 V for m/z 152 (quantitation ion)
Table 2.
Structure and Fragment Ions of CAP and CAP-d5 (* indicates deuterated positions for the CAP-d5 IS)
Chloramphenicol structure
OH
Flow rate Mobile phase Gradient
*
O2N
* * *
m/z 257 152 262 157
H N O
CI CI
OH
Post time Temperature Injection MS Source Settings Source Ion polarity Drying gas temperature Drying gas flow rate Nebulizer Vcap Fragmentor Collision energy
CAP CAP-d5
Qualifier ion Quantitiation ion
Optimization of MS Condition Figure 1 shows the results of varying the Vcap. For this analyte there was little effect from varying this parameter. Only a slight increase in signal is observed at 3,500 V, and this voltage was used. The fragmentor was varied from 90 V to 160 V. Above 120 V, fragment ions are observed and the precursor ion signal drops significantly. At 160 V on the fragmentor almost no m/z 321 is observed. This results show that 100 V on the fragmentor provided the highest precursor ion signal. Finally, using a product ion scan of the precursor, m/z 321, the collision energy (CE) was varied from 2 V, 5 V, 8 V, 10 V, 15 V, 18 V to 40 V.
Response 4.00E+06 3.00E+06 2.00E+06 1.00E+06 0.00E+00 4500 4000 3500 3000 2500 Vcap

Spectral Quality and Sensitivity of Standard Table 2 lists the structure of the CAP and the fragment ions used for quantitation and confirmation as described by the identification point system.[6] To obtain the most sensitivity, only two or three parameters need to be optimized on this instrument. They are the fragmentor, to provide highest transmission of the precursor ion, the collision energy, to maximize signal for the quantitation and qualifier ion, and possibly the Vcap (electrospray voltage), to maximize the number of ions generated.
Figure 1.
Plot of Vcap voltage vs. response of precursor ion at m/z 321.
Comparison of extracted ion chromatograms of the quantitation and qualifier ions showed that response maximized at 10 V for m/z 257 and at 15 V for m/z 152. The product ion spectra for these two collision energy experiments are shown in Figure 2 and Figure 3. As shown in Table 3, the same CE were used for the deuterated internal standard.
104 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 100 120 140 160 180 200 220 240 260 Abundance vs. Mass-to-Charge (m/z) 280 300 320 340
321.0 152.1
257.1
194.0
176.0 249.0 127.0 164.3 206.9 219.1 237.0
Figure 2.
x104 1.1 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
Product ion spectrum of m/z 321 at 10 V collision energy.

152.1
176.0
194.1 121.0
120
257.1 207.1 237.2 248.7

220 240 260 280 300
105.8
100
136.2
140
158.9 160 180 200
320.8
320 340
Abundance vs. Mass-to-Charge (m/z)
Figure 3.
Product ion spectrum of m/z 321 at 15 V collision energy.
Table 3.
MRM Mode Parameters Transition 321257 321152 Dwell time (ms) 200 200 200 200 Fragmentor Voltage (V) 100 100 100 100 Collision) Energy (V) 10 15 10 15 MS2 resolution Unit Unit Unit Unit
Compound CAP
CAP-d5
326262 326157
Repeatability Using honey matrix spiked at 0.1 ppb level as an example, the repeatability was tested by running the extract 15 times. Table 4 shows the area of the qualifier and quantitation ions in both the analyte and the IS. On average the areas of each ion vary about 8% and the ratios 5%, well within the 20% required for ratios 50% and above. Masshunter quantitation software tabulates these results and gives a graphic representation as shown in Figure 4.
Table 4.
Integrated Areas of the Quantitation Ion and Qualifier Ion and Their Associated Internal Standard Ion Chloramphenicol Quantitative ion (321152) 350 346 346 313 301 313 320 326 317 290 300 281 303 290 261 8.11% Qualifier ion (321257) 165 157 5 164 154 168 160 145 141 135 138 136 143 140 131 8.30% Ratio 47.1 45.2 44.6 52.3 49.5 53.6 50.1 44.5 44.5 46.6 46.2 48.4 47.3 48.3 50.3 5.91% d5-chloramphenicol Quantitative ion (326157) 262 258 259 267 261 253 228 225 241 226 253 240 220 214 217 7.67% Qualifier ion (326262) 121 114 118 127 121 124 111 113 117 107 90 90 101 107 101 9.99% Ratio 50.4 55.3 49.4 47.6 46.4 49.0 48.6 50.4 48.6 47.1 45.7 47.6 45.9 49.8 46.6 4.83%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 RSD
- MRM (321.0 -> 152.0) x102 Abundance 5 4 3 2 1 0 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8
321 -> 152.0 , 257.0
1.849
Abundance
x102 5 4 3
Ratio = 46.8
2 1 0 1 1.2 1.4 1.6 1.8 2 2.2
2.4
2.6
2.8
Acquisition Time (min) - MRM (326.0 -> 157.0) x101 Abundance 7 6 5 4 3 2 1 0 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 326 -> 157.0 , 262.0
Acquisition Time (min)
1.818
Abundance
x101 7 6 5 4
Ratio = 50.3
3 2 1 0 1 1.2 1.4 1.6 1.8 2
2.2
2.4
2.6
2.8
Figure 4.
Panels A and B show the CAP and IS peak for the quantitation transition. Panels C and D are the graphic representation of quantitation ion and qualifier ion ratio as shown by MassHunter software.
Linearity The linearity of the method was determined for CAP in solvent and each of the matrices. This was done from 10 ppt to 1 ppb, well below the minimum required performance level (MRPL) and above that concentration. Figures 5 through 8 show the graphic representation of those results. Each was well above an r2 value of 0.99.
Relative Responses 8 7 6 5 4 3 2 1 0 -0.2 0.2 0.6 1 1.4 1.8 2.2 2.6 3 3.4 3.8 4.2 Relative Concentration
Relative Responses 6 5 4 3 2 1 0 -0.2 0.2 0.6 1 1.4 1.8 2.2 2.6 3 3.4 3.8 4.2 Relative Concentration
y = 1.9565 * x + 0.0669 R^2 = 0.99967712
y = 1.3871 * x + 0.3325 R^2 = 0.99854947
Figure 5. 6
Linearity of CAP in solvent from 10 ppt to 1 ppb.
Figure 6.
Linearity of CAP in honey from 10 ppt to 1 ppb.
Relative Responses 8 7 6 5 4 3 2 1 0 -0.2 0.2
Relative Responses 7 6 5 4 3 2 1 0
y = 1.7891 * x + 0.2085 R^2 = 0.99890473
y = 1.7027 * x + 0.0150 R^2 = 0.99985348
0.6
1.4
1.8
2.2
2.6
3 3.4 3.8 4.2 Relative Concentration
-0.2
0.2
0.6
1.4
1.8
2.2
2.6
3 3.4 3.8 4.2 Relative Concentration
Figure 7.
Linearity of CAP in shrimp from 10 ppt to 1 ppb.
Figure 8.
Linearity of CAP in chicken from 10 ppt to 1 ppb.
Sensitivity The sensitivity of CAP standard in solvent is observed at 10 ppt with an injection volume of 5 L. The MRM chromatogram is shown in Figure 9. Although this demonstrates the sensitivity of the instrument, it is also important to determine the sensitivity in real sample matrix. This is shown in Figure 10 with a spike concentration of CAP at 10 ppt with a 5-L injection. Not only is the analyte detectable, but the ratio of the qualifier ion is within the specified tolerance so confirmation can be obtained.
x10 2 2.0 1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
- TIC MRM (****) CAP 10 ppt
Recovery Recovery was determined by spiking CAP into three samples of matrix and extracting using the specified SPE. Table 5 shows both the repeatability of extraction and analysis and the mean recovery. Using the internal standard spiked before extraction, recovery is automatically compensated. Thus accuracy of the quantification is very good using this methodology. The recovery results show the overall effectiveness of the method.
1.724
2.6
2.8
3.2
3.4
3.6
3.8
Abundance vs. Acquisition Time (min)
Figure 9. Table 5.
MRM chromatogram of 10 ppt CAP in solvent with injection volume of 5 L.
Recovery of CAP at 0.1 ppb Where Three Sample Aliquots of Each Matrix Were Spiked and Determined Honey Shrimp Chicken (n=3) (n=3) (n=3) RSD (%) 6.29 3.93 3.29 Recovery (%) 89.5 85.4 86.4
Conclusions
The method described herein for the analysis of CAP in three important matrices has been shown to be highly effective and meet the criteria for
quantitation and confirmation well below the required 0.3 ppb MRPL. Optimization of the method was simple, as few parameters in the mass spectrometer need adjustment. In addition, the requirements for a validated method have been shown. These include sensitivity, repeatability, linearity, and recovery. The Agilent 6410 LC/MS triple quadrupole instrument has been shown to be a highly effective instrument for the analysis of chloramphenicol.
10 1 1.0 0.9 0.8 0.7 Abundance Abundance 0.6
- MRM (321 152.0) CAP in Honey.d
10 1 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 _ 0.1
321 152.0, 257.0
1.751
Ratio=57.8
0.5 0.4
0.3 0.2 0.1 0.0 _ 0.1 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 Acquisition Time (min)
1.2
1.4 1.6 1.8 2 2.2 2.4 2.6 Acquisition Time (min)
Figure 10. MRM chromatogram of quantitation ion and ratio of qualifier ion for 10 ppt CAP in honey.
References
1. Chemical Safety Information from Intergovernmental Organisations (IPCS-INCHEM), Web page http://www.inchem.org/documents/ jecfa/jecmono/v33je03.htm 2. S. Turnipseed, et al. (2002) Confirmation of Multiple Phenicol Residues in Honey by Electrospray LC/MS, Laboratory Information Bulletin (4281) U.S. Food and Drug Administration. 3. A. Pfenning, et al. (2002) Confirmation of Multiple Phenicol Residues in Shrimp by Electrospray LC/MS, Laboratory Information Bulletin (4284) Food and Drug Administration. 4. B. K. Neuhaus, et al. (2002) LC/MS/MS Analysis of Chloramphenicol in Shrimp, Laboratory Information Bulletin (4290) Food & Drug Administration 5. P. Mottier, V. Parisod, E. Gremaud, P. A. Guy, and R. H. Stadler. Determination of the antibiotic chloramphenicol in meat and seafood products by liquid chromatography - electrospray ionization tandem mass spectrometry. Journal of Chromatography A 2003, 994, (1-2), 75-84. 6. Commission, E., 2002/657/EC: Commission Decision of 12 August 2002 implementing Council Directive 96/23/EC concerning the performance of analytical methods and the interpretation of results (Text with EEA relevance) (notified under document number C(2002) (3044) 2002.

The information contained in this publication is intended for research use only and is not to be followed as a diagnostic procedure. Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2007 Printed in the USA August 21, 2007 5989-5975EN
Analysis of Nitrofuran Metabolites in Tilapia Using Agilent 6410 Triple Quadrupole Application
Food Safety
Authors
Feng Liang, Peibin Hu, and Ping Li Agilent Technologies, Inc. Beijing, China
analytical detection of the metabolites and not the parent compounds are required in samples of animal origin. The criteria for detection and confirmation of veterinary drugs in animal and animal products established by the European Union (EU) [2] has been accepted in much of the world. This criteria mandates a separation technique combined with a spectrometric technique. For banned substances such as the nitrofurans, no maximum residue limit (MRL) could be set. Therefore a minimum required performance level (MRPL) was set at 1 g/kg for each metabolite [3]. Only LC/MS could meet these criteria, and very good methods have been reported [46]. However, the most widely accepted methodology employs triple quadrupole tandem mass spectrometers. This is the first report showing analysis of these metabolites using the new Agilent triple quadrupole LC/MS system.
Abstract
The metabolites of nitrofuran antibiotics banned in meat and meat products are analyzed by LC/MS/MS with the new Agilent 6410 triple quadrupole. The method is shown to be highly sensitive, to 0.01 ppb (10 ppt), for each of the four analytes. Calibration from 0.1 ppb to 10 ppb is presented with all criteria for confirmation as set by the European Union decisions for analytical method performance. Extracts of tilapia are used to show the performance of the LC/MS/MS method for aquaculture samples.
Introduction
Nitrofurans are inexpensive antibiotics used for Gram positive and Gram negative bacteria. They have been used to treat gastrointestinal and dermatological infections in farm animals and fish. In addition, they have been used to treat bacteria in bees. Because both parent compounds and their metabolites are suspect carcinogens, they have been banned around the world. The Rapid Alert System for Food and Feed Annual report for 2005 [1] shows that these compounds continue to be detected in food samples and remains a major concern for food safety. The four compoundsfurazolidone, furaltadone, nitrofurantoin, and nitrofurazonehave been found to metabolize rapidly, and the metabolites bind to muscle tissue. Thus the
Experimental
Chemicals Derivatized standards of nitrofuran metabolites and all chemicals for sample preparation were received from a food manufacturing company. Acetonitrile was HPLC grade from Merck (Darmstadt, Germany). Formic acid was reagent grade from Merck (Darmstadt, Germany). Sample Preparation The accepted procedure for sample preparation was followed. To 2 g of tilapia was added 15 mL 0.125 M HCl and the mixture homogenized.
To this solution, a 50-L solution of 2-nitrobenzaldehyde (50 mM in DMSO) was added and shaken. The solution was then incubated at 37 C for 16 hours. This was followed by neutralization to pH ~7 with NaOH and K2HPO4. The neutral derivatized sample was then extracted with ethyl acetate, concentrated to dryness, and reconstituted in 100 L of initial LC mobile phase. Standards of the four metabolites were spiked into 0.125 M HCl, derivatized, and extracted for calibration using the same procedure as was used for the samples of tilapia. LC/MS/MS Method
LC Conditions Instrument: Column: Column temp.: Mobile phase: Gradient: Agilent 1100 LC C18, 2.1 mm 150 mm, 3 m 40 C A = 0.1% formic acid in water B = acetonitrile 22% B at 0 min 99% B at 6 min 99% B at 9 min 0.3 mL/min 50 nL
Quantitation Quantitative analysis was done with the first transition listed in the MRM parameter table. The second transition was used as a qualifier ion for confirmation as per the confirmation criteria. Quantitative results were performed with the new MassHunter quantitative analysis software.

The instrument sensitivity is an important performance parameter for this analysis when considering the derivatization and extraction needed to meet the required detection limit of 1 ppb for each metabolite, aminohydantoin (AH), 3-amino-5-morpholinomethyl-2-oxazolidinone (AMOZ), 3-amino2-oxazolidinone (AOZ), and semicarbazide (SC). To demonstrate this performance, a standard of each 2-nitrobenzaldehyde (2-NBA) derivatized metabolite is shown at 0.01 ppb (10 ppt) in Figure 1. The structure of each derivatized metabolite is given and each is shown with a signal-to-noise ratio of greater than 3:1. Another indicator of performance is linearity. Calibration curves from this concentration (10 ppt) to 10 ppb are displayed in Figure 2 showing the linearity for each compound. Treatment of fish with nitrofurans is a continual problem for food safety and import into EU member countries. To demonstrate the capability of the Agilent triple quadrupole LC/MS, tilapia samples were spiked with the four metabolites, hydrolyzed, derivatized, and extracted. An analysis of a tilapia extract at 500 ppt is shown in Figure 3 and demonstrates the signal obtained at half the MRPL. In addition to meeting the sensitiv-
Flow rate: Injection volume: MS Conditions Instrument: Ionization mode: Drying gas flow: Nebulizer: Drying gas temp.: Vcap:
Agilent 6410 LC/MS Triple Quadrupole Positive ESI 10 L/min 35 psig 350 C 5000 V
MRM Mode Parameters Dwell time (ms) 60 60 60 60 60 60 60 60 Fragmentor voltage (V) 100 100 100 100 100 100 100 100 Collision energy (V) 5 5 5 5 5 5 5 5 MS2 resolution Unit Unit Unit Unit Unit Unit Unit Unit
Compound AMOZ
Transition 335.1 & 291.4 335.1 & 262.4
SC
209.1 & 192.3 209.1 & 166.3
AH
249.1 & 134.2 249.1 & 104.2
AOZ
236.0 & 134.1 236.0 & 104.1
ity requirement, the analysis must also meet the confirmation criteria, including both chromatographic retention time match with the standards and measuring a qualifying ion with a relative intensity ratio within a specified tolerance of the quantitation ion. This tolerance is set by the ratio obtained when analyzing standards and increasing as that ratio decreases. This tolerance ranges from 20% for ions with relative ratio intensities above 0.5 and to 50% for ratios below 0.1. Table 1 shows tilapia samples spiked with the metabolites, derivatized and extracted. The spikes
were used as the calibrants, so the final concentration is obtained from the curve. The table is produced as the batch using the MassHunter software results with outliers highlighted in blue (low) and red (high). The table shows that in the blank a peak is found within the tolerance set for the retention time of AMOZ but the qualifier ratio is low. For AOZ and SC, retention times for suspect peaks are below the specified retention in the same blank. For AH, the 0.5 ppb spike, the qualifier ion ratio is outside the 35% tolerance limit set for this ion (again low).
x10 2 0.9 0.7 0.5
+ MRM (335.1 291.4) 10ppt_1.d Smooth (1)
12
23
34 O N O O N
+ _
O O O N N N+ O
_
x10 1 + MRM (209.1 192.3) 10ppt_1.d Smooth (1) 7 6 5 x10 1 6.0 + MRM (249.1 134.2) 10ppt_1.d Smooth (1) 5.6 5.2 4.8
12
23
34
N NH
O NH2
12
23
34
O O N
+
N N O O
_
x10 + MRM (236.0 134.3) 10ppt_1.d Smooth (1) 1.2 1.0 D 0.8 0.6
2
12
23
34
N N
O NH
O 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2
Abundance vs. acquisition time (min)
Figure 1.
The MRM quant ion chromatogram for each derivatized metabole at 10 ppt of A) 2-NBA AMOZ, B) 2-NBA SC, C) 2-NBA AH, and D) 2-NBA-AOZ
x105 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0
y = 49247.4270 * - 1850.8024 R 2 = 0.99989213
Response
R 2 > 0.999
Response
AMOZ
x105 1.2 1.1 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 _0.1
y = 11978.9741 * - 1213.1580 R 2 = 0.99852795
SEM
R 2 > 0.998
0 x104 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 _0.5
3 4 5 6 7 Concentration (ng/mL)
10
10
y = 5697.5004 * - 565.5495 R 2 = 0.99879989
Response
R 2 > 0.998
Response
AHD
x105 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0
y = 45077.4032 * + 4420.7065 R 2 = 0.99523855
AOZ
R 2 > 0.995
10
10
Figure 2.
Calibration curves of nitrofuran metabolites linear range from 10 ppt to 10 ppb.
x10 3
+ TIC MRM (** **) 0619_500ppt_spike_MRM_50uL_4seg_4.d Smooth (1)
12
23
34
AOZ AMOZ
1.0
SC AH
Figure 3.
Spiked tilapia sample extract at 500 ppt each metabolite.
Table 1.
Analysis of Talapia Spikes Self-Calibrated. Note Qualifier Ratios and Retention Times Reported.
AH
Name Nitrofuran Blank Nitrofuran 0.5 ppb Nitrofuran 1 ppb Nitrofuran 3 ppb Nitrofuran 5 ppb RT 8.66 8.66 8.67 8.66 Final Conc. 0.00 0.54 1.07 3.42 5.71
Qualifier
Ratio 6.07 15.83 13.78 14.24 RT 2.72 2.62 2.64 2.65 2.66
AMOZ
Final Conc. 0.02 0.51 0.92 3.41 4.66
Qualifier
Ratio 8.70 17.44 20.42 18.67 19.28 RT 8.71 9.19 9.19 9.20 9.19
AOZ
Final Conc. 0.00 0.37 1.06 3.18 4.89
Qualifier
Ratio 7.99 10.19 9.51 8.61 RT 7.94 8.44 8.44 8.44 8.44
SC
Final Conc. 0.00 0.50 1.05 2.89 5.05
Qualifier
Ratio 65.32 70.85 67.47 69.06
Conclusions
This work shows the high performance of the new Agilent 6410 LC/MS triple quadrupole system for the sensitive analysis of the nitrofuran metabolites in fish samples. The system readily meets the performance requirements and provides advanced quantitation software for calculating and reporting all confirmation parameters specified by the European Commission decision.
4. B. Wst, C. Sauber, and H. J. A. van Rhijn, Quantitation of Nitrofuran Metabolites in Shrimp and Poultry by LC/MS/MS Using the Agilent LC/MSD Trap XCT, Agilent Technologies publication, 5989-0738EN: March 25, 2004. 5. M. Takino, Determination of the Metabolites of Antibacterial Drugs in Chicken Tissue by Liquid Chromatography Electrospray Ion ization Mass Spectrometry (LC-ESI-MS), Agilent Technologies publication, 5988-8903EN: March 19, 2003. 6. F. Mandel, B. Wst, and C. Sauber, High Resolution Quantitative Analysis of Nitrofuran Derivatives in Poultry and Shrimp Using a New oa-ESI-TOF, Agilent Technologies publication, 5989-1302EN: July 9, 2004.
References
1. The Rapid Alert System for Food and Feed (RASFF) Annual Report 2005, p 29, http://ec.europa.eu/food/food/rapidalert/ index_en.htm. 2. E. Commission, 2002/657/EC: Commission Decision of 12 August 2002 implementing Council Directive 96/23/EC concerning the performance of analytical methods and the interpretation of results (Text with EEA relevance) (notified under document number C [2002] 3044) 2002. 3. E. Commission, 2003/181/EC: Commission Decision of 13 March 2003 amending Decision 2002/657/EC as regards the setting of minimum required performance limits (MRPLs) for certain residues in food of animal origin (Text with EEA relevance) (notified under document number C[2003] 764) 2003.

Determining Malachite Green and Leucomalachite Green in Food by LC/MS/MS Application
Food Safety
Authors
Feng Liang, Peibin Hu, and Ping Li Agilent Technologies Beijing, China
have already banned MG in fishery. Due to its low cost and antifungal effectiveness, MG is still being used illegally as indicated in the European Rapid Alert System for Food and Feed.2 HPLC with UV detection has been used to analyze MG and LMG. Figure 1 shows the structure of the two compounds. Loss of conjugation by reduction changes the chromaphore of LGM significantly. To obtain the sum of both, the method employs postcolumn oxidation with PbO2 to convert LMG to MG, thus providing a sum of both comounds.3 Most recently, LC/MS has been used to both meet the EU confirmation criteria and provide quantitative results for both compounds without the need for post-column oxidation. In this application, a simple and sensitive method for simultaneously determining MG and LMG is presented.4, 5 The LC/MS/MS methods LOQ is 0.01 g/Kg, which easily meets the import requirement set by Japan or the EU.6
Abstract
This application note demonstrates a complete method to rapidly and precisely determine residue levels of malachite green and leucomalachite green in fish with the new Agilent 6410 LC/MS triple quadrupole system. Using positive mode electrospray ionization (ESI+) and multiple reaction monitoring (MRM), qualification and quantification were accomplished without the traditional tedious PbO2 oxidation process. The LC/MS/MS methods LOQ is 0.01 g/Kg, which easily meets the import requirement of 2 g/Kg set by Japan and the EU.
Introduction
Malachite green (MG) is a metallic-looking crystal. It dissolves in water easily as a blue-green solution. It is a toxic chemical primarily used as a dye and has been found very effective in treating parasites, fungal infections, and bacterial infections in fish and fish eggs.1 On uptake, MG is rapidly reduced into leucomalachite green (LMG) and deposited in the fatty tissue of the fish with little MG remaining. MG can cause significant health risk for humans who eat contaminated fish. For example, it can cause liver tumor formation and is suspected of carcinogenesis.1 The United States, Japan, China, the European Union, and many other countries
Experimental
Reagents MG LMG Sigma-Aldrich, CAS 569-64-2, USA Dr. Ehreastorfer's lab, D-86199, 99% pure, Augsburg, Germany CAS 75-05-8; Burdick & Jackson; Morristown, New Jersey, USA Merck, Germany CAS 631-61-8, Acros Organics, Morris Plains, New Jersey, USA
Acetonitrile
Acetic acid Ammonium acetate
H C N C N
N +
Cl
Malachite green
Figure 1.
Leucomalachite green
Molecular structure of malachite green and leucomalachite green.
Calibration Solutions A stock standard solution of MG and LMG in acetonitrile was prepared at 100 g/mL and stored at %18 oC, avoiding light. The stock solution was diluted in 50:50 acetonitrile:water to make the calibration solutions+10, 50, 100, 500, 1000, 5000, and 10,000 fg/L. Sample Preparation To 5 g tilapia tissue was added 1 mL (0.25 mg/mL) hydroxylamine, 2 mL 1 M toluene sulfonic acid, 2 mL of 0.1 M ammonium acetate buffer (pH 4.5), and 40 mL acetonitrile. The mixture was then homogenized for 2 min. The supernatant was decanted, and to the precipitate was added 20 mL acetonitrile. This was filtered and added to the supernatant. To the combined acetonitrile extracts, 35 mL water and 30 mL methylene chloride were added. The solution was shaken and the methylene chloride layer collected. A second extract of 20 mL methylene chloride was made, and this layer added to the first extract. The methylene chloride was taken to dryness with a gentle stream of nitrogen and the extract reconstituted in 100 L of acetonitrile
Instrumentation
LC Column Column temp. Mobile phase 1100 LC C18, 2.1 x 150 mm, 5 m 40 oC A % 10 mmol/L ammonium acetate (adjust to pH 4.5 with acetic acid) B % acetonitrile 0.3 mL/min Time %B 0 30 1 50 2 95 8 95 8.01 30 13 30 10 L
Column flow Gradient
Injection vol. MS
Agilent 6410 LC/MS Triple Quadrupole Ionization ESI(+) Capillary 4000 V Nebulizer P. 35 psi Drying gas 11 L/min Gas temp. 350 oC Skimmer 15 V OctDc1 (Skim2) 45 V Oct RF 500 V Q1 resolution Unit Q3 resolution Unit Collision gas Nitrogen The MRM parameters are listed in Table 1.
Table 1. Time 0 7
MRM Method Parameters Compound MG LMG Precursor 329.3 329.3 331.3 331.3 Product 313.3 208.2 316.3 239.2 Dwell (ms) 40 40 40 40 Fragmentor (V) 100 100 100 100 Collision Energy (V) 40 40 30 30

To obtain the most sensitive results, optimization of certain fragmentor voltages is important. Figure 2 shows the EICs of both target compounds at fragmentor values of 70 V, 90 V, and 100 V. The results show that the three different fragmentor values have little effect on the intensity of [M+H]+ ions. Thus, 100 V was chosen for this study. In addition, an optimal collision energy for the MS/MS must be set. Figure 3 shows the MS/MS spectra from three different collisional voltages,
(a) 20 V, (b) 30 V, and (c) 40 V. Due to their structural differences, the voltage required for optimum fragmentation of each compound is different. For MG, the optimum fragmentation was observed at 40 V. The ion m/z 313 was due to the neutral loss of methane. The ion at m/z 208 was due to the neutral loss of N,N-dimethylaniline. For LMG, the optimum fragmentation was observed at 30 V. The ion at m/z 316 was due to the loss of a methyl radical. The ion at m/z 239 resulted from a subsequent loss of a benzene radical or, more likely, the rearrangement and neutral loss of toluene.
x107 + EIC(329.4, 331.4 m/z) Scan optimizing FRG70_3.d 5 3 1

+ EIC(329.4, 331.4 m/z) Scan optimizing FRG90_4.d
70 V
x107 5 3 1 x107 5 3 1
90 V
+ EIC(329.4, 331.4 m/z) Scan optimizing FRG100_5.d
100 V
0 1 2 3 4 5 6 7 8 9 10 11 12
Figure 2.
EICs of malachite green and leucomalachite green at fragmentor values of 70 V, 90 V, and 100 V.
329.3 x105 1.2 1.0 0.8 0.6 0.4 0.2 0.0 193.1 208.3 236.9 268.4 284.3 313.4 331.8
+ Product Ion (5.499-5.633 min, 17 scans) (329.3, 331.4 **) optimizing MS2_FRG100_CE20_2.d
x105 + Product Ion (8.349-8.466 min, 15 scans) (331.4, 329.3 **) optimizing MS2_FRG100_CE20_2.d 239.8 2.0 1.6 1.2 0.8 0.4 0.0 40 60 80 100 120.8 134.5 120 140 165.6 160 195.8 180 209.8 223.9 200 220 240 272.8 286.6 301.8 260 280 300 316.7
320
340
Abundance vs. mass-to-charge (m/z)
Figure 3a. MS/MS spectra of MG and LMG at collisional voltage of 20 V.

x104 7 6 5 4 3 2 1 0 134.3 165.1 192.8 208.2 217.4 237.2 251.4 270.3 285.3298.9 313.4
329.3
239.8
x105 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 40 60 80 100 120
315.8 134.5 140 209.8 223.9 165.8 180.6 194.7 160 180 200 220 240 Abundance vs. mass-to-charge (m/z) 260 272.7 286.8 280 301.9 300 320 331.8 340
Figure 3b. MS/MS spectra of MG and LMG at collisional voltage of 30 V.
x104 4.0 3.0 2.0 1.0 0.0
313.3
208.2 165.3 241.4 284.2 299.4 329.4
91.5
117.9 134.1
192.1
221.0
270.3
x105 2.5 2.0 1.5 1.0 0.5 0.0 40 60 80 91.6 100 120.6 134.5 120 140 165.8 180.7 194.7 208.7 223.8
239.8
315.8 257.9 272.7 286.7 260 280 300 320 340
160 180 200 220 240 Abundance vs. mass-to-charge (m/z)
Figure 3c MS/MS spectra of MG and LMG at collisional voltage of 40 V.
Figure 4 shows the calibration curves for both MG (4a) and LMG (4b). Calibration solution concentrations were from 10 to 10,000 fg/L. The linear calibration range is 100 to 100,000 fg on column for both compounds. The R2 for both compounds was > 0.999 (origin ignored and no weighting). To demonstrate the sensitivity of the instrument,
x105 2.6 2.4 2.2 2.0 1.8 1.6 Response 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 0 1 2 3 4 5 6 7 Concentration (ng/mL)
Figure 5 shows MS/MS spectra of a blank sample extract (5a) and sample extract spiked with 10 ppt of each compound (5b). A sample of tilapia spiked at 100 ppt MG and LMG before extraction was made to demonstrate method performance. The MRM results after extraction and cleanup are shown in Figure 6. The recover-
Malachite green - 7 levels, 7 levels used, 14 points, 14 points used, 0 QCs y = 23363.3374 * - 1766.9951 R 2 = 0.99946103
R 2 = 0.9995
10
11
12
Figure 4a. Calibration curve of malachite green, linear range: 10 ppt to 10 ppb.
x10 6 1.0 0.9 0.8 0.7 Response 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0 1 2 3 4 5 6 7 Concentration (ng/mL) 8 9 10 11 12
Leucomalachite green - 7 Levels, 7 Levels Used, 14 Points, 14 Points Used, 0 QCs y = 93199.4712 * - 7543.3588 R 2 = 0.99942595
R 2 = 0.9994
Figure 4b. Calibration curve of leucomalachite green, linear range: 10 ppt to 10 ppb.
8.433 x101 2.8 2.4 2.0 1.6 1.2 0.8 0.4 0.0 0 1 2 3 4 5 6 7 Abundance vs. acquisition time (min) 8 9 10 11 12
+ MRM MRM (331.3 239.2) malachite green_200606121.d
12
Figure 5a. MG and LMG MRM of a blank sample.

x102 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 0 1 2 3 4 5 6 7 Abundance vs. acquisition time (min) 8 9 10 11 12 5.440
+ MRM MRM (329.3 313.3) malac hite green_200606121.d
12
ppt spiked sample.

Figure 5b. MG and LMG MRM of a 10-ppt spiked sample.
x102 3.2 2.8 2.4 2.0 1.6 1.2 0.8 0.4 0.0 0
+ MRM MRM (331.3 239.2) Spike_100 ppt_1.d
12
8.315
5 6 7 Abundance vs. acquisition time (min)
10
11
12
Figure 6. MRM result of talapia extract spiked with 100-ppt MG and LMG.
ies for MG were 48% and 23% for LMG. A mixture of MG and LMG at 100 fg/L in 50:50 acetonitrile: ammonium acetate was used for the repeatability study for instrument performance. The RSD from eight injections for MG was 3.52% (S/N > 20). The RSD from eight injections for LMG was 2.25% (S/N > 40).
4. M. D. Hernando, M. Mezcua, J. M. SuarezBarcena, and A. R. Fernandez-Alba, Liquid chromatography with time-of-flight mass spectrometry for simultaneous determination of chemotherapeutant residues in salmon. Analytica Chimica Acta 2006, 562, (2), 176%184. 5. K.-C. Lee, J.-L. Wu, and Z. Cai, Determination of malachite green and leucomalachite green in edible goldfish muscle by liquid chromatography-ion trap mass spectrometry. Journal of Chromatography B 2006, In Press, Corrected Proof. 6. 2004/25/EC: Commission Decision of 22 December 2003 amending Decision 2002/657/EC as regards the setting of minimum required performance limits (MRPLs) for certain residues in food of animal origin (Text with EEA relevance) (notified under document number C [2003] 4961). 2003.
Conclusions
This application note demonstrates a complete method to rapidly and precisely determine residue levels of malachite green and leuco-malachite green in fish. Using positive mode electrospray ionization (ESI+) and multiple reaction monitoring (MRM) technique, the LC/MS/MS method shows detection limit of 10 ppt, which easily meets the import requirement set by Japan or EU.
References
1. S. Srivastava, R. Sinha, and D. Roy, Toxicological effects of malachite green. Aquatic Toxicology 2004, 66, (3), 319%329. 2. The Rapid Alert System for Food and Feed (RASFF) Annual Report 2005. 2005, 29. 3. C. A. Hajee and N. Haagsma, Simultaneous determination of malachite green and its metabolite leucomalachite green in eel plasma using post-column oxidation. Journal of Chromatography B Biomed Appl. 1995, 669, (2), 219%227.
Acknowledgement
The authors would like to thank Dr. Yanqin Liu for providing the standard solutions and sample extracts.

Quantitation of Nitrofuran Metabolites in Shrimp and Poultry by LC/MS/MS Using the Agilent LC/MSD Trap XCT Application
Food
Authors
Bernhard Wst, Christian Sauber Agilent Technologies GmbH, Hewlett-Packardstr. 8, 76337 Waldbronn Germany Hans (J.) A. van Rhijn RIKILT -Institute of Food Safety Bornsesteeg 45,6700 AE Wageningen the Netherlands
shrimp, and poultry. In 1995 the four drugs Furazolidone, Furaltadone, Nitrofurantoin, and Nitrofurazone were banned by the European Union (EU) for their usage in food-producing animals. Due to their rapid metabolism nitrofuran parent substances are not suitable for monitoring and typically their metabolites are analyzed. A liquid chromatography /mass spectrometry/mass spectrometry (LC/MS/MS) method was developed for the sensitive, qualitative, and quantitative analysis of four derivatized nitrofuran metabolites found in shrimp and poultry. See Figure 1.
Abstract
An LC/MS/MS method was developed for the qualitative and quantitative measurement of nitrofuran metabolites in chicken and shrimp using the Agilent LC/MSD XCT Ion Trap. The limit of quantitation for all four nitrofurans investigated easily met the specified European Union Minimum Required Performance Level of 1 g/kg and ranged from 0.125 g/kg to 0.5 g/kg.
Experimental
All liquid chromatography/mass spectrometry (LC/MS) experiments were performed using an Agilent 1100 Series LC system coupled to a mass selective detector (MSD) Ion Trap XCT mass spectrometer. The Ion Trap was operated with an orthogonal electrospray source in positive ion mode using multiple reaction monitoring (MRM). A gradient method was used for chromatography. Deuterated NBA-AMOZ was used as the internal standard (ISTD) for NBA-AMOZ and deuterated NBA-AOZ was used as the ISTD for the other metabolites.
Introduction
Nitrofuran antibiotics are widely used for the treatment of infectious diseases in cattle, pigs,
O O
O O O N N O H2N N O O N
O O N N O
N+
Furazolidone
O O
_
AOZ
O O N N
NBA-AOZ
N+
O N N O O N
H2N
N N
O N
O N
Furaltadone
_
AMOZ
O O N O
NBA-AMOZ
O H N O
O O
O O N N O NH H2N N O
N+
NH
Nitrofurantoin
O O
_
AHD
O O N O H N NH2
NBA-AHD
N+
O N
H N NH2
H2N
H N NH2
Nitrofuranzone
SEM
NBA-SEM
Drugs Figure 1.
Metabolites
Derivatized metabolites
Chemical structures of nitrofurans, nitrofuran metabolites, and corresponding derivatives.
Sample Preparation One (1) g of homogenized tissue from shrimp or poultry was mixed with 5 mL of a 0.2 M hydrochloric acid and 50 L of 2-nitrobenzaldehyde (2-NBA, 100 mM in methanol) and incubated overnight at 37 C. This is a protocol from the State Institute for Quality Control of Agricultural Products (RIKILT, The Netherlands). Using this protocol, tissue-bound residues of the nitrofurans with an intact side chain are released through acidic hydrolysis of the imine bond. The free amino groups of the corresponding metabolites are derivatized with 2-NBA to form an aromatic imine bond.
The sample was then neutralized with 500 L of a 0.3 M Na3PO4 solution in water and adjusted to pH7 with 2 M NaOH solution. After the addition of 4-mL ethyl acetate, the sample was centrifuged and the organic layer was transferred to a clean tube. The sample was further extracted with a 4-mL aliquot of ethyl acetate, centrifuged, and the organic layer added to the first extract. After evaporation to near dryness, the sample was reconstituted in 500-L solvent (50-mL acetonitrile, 80-mL water and 0.5-mL acetic acid) for subsequent LC/MS analysis. Calibration standards were made by spiking known concentrations of all four underivatized analytes into shrimp and poultry prior to sample preparation.
Chemicals
Methanol Hydrochloric acid, HCL 32% 2-Nitrobenzaldehyde (2-NBA), C7H5NO3 Tri-sodiumphosphate-dodecahydrate, Na3PO4 12(H2O) Sodium hydroxide, NaOH Ethyl acetate, CH3COOC2H5 Acetic acid, CH3COOH Acetonitrile, CH3CN Methanol-d, 99.5% D (Biosolve, 13683502) (Merck, 100319) (Sigma, N6001) (Merck, 106578) (Merck,106498) (Biosolve 05402602) (Merck 10063) (Biosolve, 01203502) (Aldrich, 15.193-9)
LC/MS/MS Method Details

HPLC: Flow rate: Column: Mobile phases: Agilent 1100 0.4 mL/min Zorbax XDB-C18, 2.1 mm 150 mm, 3.5 m A: Water + 0.1% acetic acid B: Acetonitrile + 0.1% acetic acid Gradient: 014 min: 10% A - 45% A; 1416 min: 45% A - 90% A 50 L out of 500 L
Injection: MS 1100 LCMSD XCT Ion Trap Ionization mode: Nebulizer pressure: Drying gas flow: Drying gas temperature: Skimmer: Capillary exit: Trap drive: ICC: MRM mode 4 Segments: Segment 1: Segment 2: Segment 3: Segment 4:
Positive ESI 45 psi 12 L/min 350 C 20 V 55 V 55 On
02.2 min, 2.27.4 min, 7.410.6 min, 10.613.9 min Divert valve: to waste, no spectra MS/MS of m/z 335, Isolation width: 2.0, Cut-off: 140; MS/MS of m/z 340, Isolation width: 2.0, Cut-off: 140; MS/MS of m/z 209, Isolation width: 2.0, Cut-off: 120; MS/MS of m/z 249, Isolation width: 2.0, Cut-off: 100; MS/MS of m/z 236, Isolation width: 2.0, Cut-off: 100; MS/MS of m/z 240, Isolation width: 2.0, Cut-off: 100;
Amplitude: 1.16 Amplitude: 1.16 Amplitude: 1.28 Amplitude: 1.25 Amplitude: 1.25 Amplitude: 1.25
Quantitation NBA-AMOZ: NBA-dAMOZ: NBA-SEM: NBA-AHD: NBA-AOZ: NBA-dAOZ: Maximum accumulation time: Smart target: Scan: EIC of 261 + 291 (MS/MS of 335), EIC of 266 + 296 (MS/MS of 340), EIC of 166 + 192 (MS/MS of 209), EIC of 134 (MS/MS of 249), EIC of 134 (MS/MS of 236), EIC of 134 (MS/MS of 240), 150 ms 100.000 100350 Ret. Time: 4.5 min Ret. Time: 4.5 min Ret. Time: 9.9 min Ret. Time: 10.0 min Ret. Time: 10.8 min Ret. Time: 10.8 min

Very low limits of detection (LOD) are required for nitrofuran metabolites and the derivatization method increased the ionization efficiency, as well as improving the chromatographic separation. A liquid-liquid extraction procedure was used which resulted in a relatively high concentration factor to further improve LOD. The ion trap mass spectrometer was operated in MRM mode. In this mode, only precursor ions are chosen and full-scan MS/MS-spectra of the corresponding analytes are acquired. These full scanMS/MS spectra are then used for identification by comparing them with MS/MS-spectra stored in a library. No further qualifier ion has to be monitored. See Figure 2.
Intensity x104 3 2 1 0 x105 0.8 0.4 0.0 x104 2.0 1.0 0.0 x104 2.0 1.0 0.0 6000 4000 2000 0 x104 1.2 0.8 0.4 0.0 2 4 6 8 10 12 Time [min]
Intensity
2A
NBA-AMOZ
x104 3 2 1 0
x104 6 4 2 0
2B
NBA-AMOZ
192.9
290.9
261.9
317.8 333.9
NBA-d5AMOZ
NBA-d5AMOZ
264.9
295.9
x104 3.0
NBA-AOZ
2.0 1.0 0.0 x104 2.0

119.9
164.7 133.8
217.7 235.8 206.7
NBA-AOZ
148.8
171.8
190.7
NBA-d4AOZ
133.8
1.0 0.0 8000 6000 4000 2000 0 8000 6000 4000 2000 0 100 125
122.9 158.9 176.8 148.8 166.7 239.9 204.8 222.8
NBA-d4AOZ
133.8 120.8
138.7
162.7 230.7 170.7 188.7 200.7212.8 248.8
NBA-AHD
144.8
NBA-AHD
165.7
NBA-SEM
NBA-SEM
135.8 148.7 190.8 207.8 217.8 228.7
150
175
200
225
250
275
300
325
m/z
Figure 2.
Representative chromatograms (2A) and MS/MS spectra (2B) for all analytes plus ISTDs (1 g/kg).
Quantitation is performed by selecting one or more product ions to create extracted ion chromatograms for each analyte and ISTD. The product ions used for quantitation were selected for best signal-tonoise (S/N) ratio post-acquisition. See Figures 3 and 4.
4000
3000
Intensity
2000
1000
0 2.5
3.0
3.5
4.0
4.5 5.0 Time [min]
5.5
6.0
6.5
7.0
Figure 3.
Limit of quantitation (LOQ) for NBA-AMOZ, 0.125 g/kg in shrimp matrix.
1.0
1.0
0.8
0.8
ISTD Response x106
Relative response
0.6
0.6
0.4
0.4
0.3
0.3
0.0 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25 2.50 Relative concentratiion
0.0
Figure 4.
Calibration curve for NBA-AMOZ, 0.125 g/kg 2 g/kg poultry matrix, three replicates.
NBA-AMOZ and NBA-SEM were quantified using the sum of two product ions, while NBA-AHD and NBA-AOZ were quantified using one product ion. The European Union has set a Minimum Required Performance Level (MRPL) of 1 g/kg for nitrofuran metabolites. These detection limits are easily reached using this method with LOQs ranging from 0.125 g/kg for NBA-AMOZ to 0.25 g/kg for NBA-AOZ and NBA-SEM, and 0.5 g/kg for NBA-ADH. See Figure 5.
Intensity x104 4 3 2 1 0
NBA-AMOZ
x104 6 4 2 0 x104 1.50 1.00 0.50 0.00 x104 1.2 0.8 0.4 0.0 6000 5000 4000 3000 2000 1000 0 x104 4 3 2 1 0 2 4
NBA-d5 AMOZ
NBA-AOZ
NBA-d4 AOZ
NBA-AHD
NBA-SEM
10
12
Time [min]
Figure 5.
Representative chromatogram of a positive shrimp sample at a level of 0.25 g/kg.
Linearity of the method was evaluated up to twice the MRPL (2 g/kg) and showed a linear weighted regression (1/) with coefficients of correlation of 0.99 or better. Intraday relative standard deviations (RSDs) were below 10% for all analytes at all concentrations. See Table 1.
Table 1. Method Reproducibility and Accuracy for the Four Target Derivatized Metabolites NBA-AOZ SD % 2.29 2.52 3.35 3.01 2.13 n=6 Accuracy % (average) 98.94 101.76 100.58 101.90 96.82 NBA-AMOZ Accuracy % SD % (average) 4.34 101.10 5.87 95.55 6.21 105.19 5.46 99.11 5.77 99.05 NBA-AHD SD % 6.05 4.59 6.53 6.97 Accuracy % (average) 98.31 103.87 100.22 97.60
NBA-SEM Standard Accuracy % (g/kg) SD % (average) 0.125 3.77 98.81 0.25 2.26 102.72 0.5 3.40 100.72 1 3.56 96.62 2 2.94 101.12 All calibration curves linear weighted 1/x
Conclusions
An LC/MS/MS method was developed for the qualitative and quantitative measurement of nitrofuran metabolites in chicken and shrimp using the Agilent XCT Ion Trap. The LOQ for all four nitrofurans investigated easily met the specified EU MRPL of 1 g/kg and ranged from 0.125 g/kg to 0.5 g/kg.

The Analysis of Fluoroquinolones in Beef Kidney Using HPLC Electrospray Mass Spectrometry Application
Food
Authors
Ralph Hindle Access Analytical Labs #3, 2616 - 16 Street N.E. Calgary, Alberta T2E 7J8 Canada Agilent contact: Chin-Kai Meng Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA
Introduction
Fluoroquinolones are synthetic antibacterial compounds derived from nalidixic acid, and are useful to treat animal infections that are resistant to other antibacterial agents. They have a broad spectrum of activity, acting against both gram-positive and gram-negative bacteria. The maximum residue limit (MRL) for enrofloxacin (as the sum of enrofloxacin and ciprofloxacin) was entered into Annex 1 of Council Regulation (EEC) No. 2377/90 for kidney at 200 g/kg in bovine and ovine species, and 300 g/kg for porcine, poultry, and rabbits. For all other food producing species, the MRL is 200 g/kg in kidney [1]. There are a number of methods describing the analysis of fluoroquinolones in various tissues, with HPLC coupled with fluorescence and mass spectrometric detection being very popular. Most methods involve extraction into acidic or basic organic solvents, followed by some type of cleanup, most notably solid phase extraction (SPE). The Canadian Food Inspection Agency extracts animal tissue with acidic ethanol, followed by strong cation exchange SPE cleanup, and HPLC fluorescence analysis [2]. Chen and Schneider [3] described a screening method for enrofloxacin in chicken, where extracts were detected by fluorescence without cleanup, following extraction and centrifugation.
Abstract
A fast and simple screening method was validated for the analysis of three fluoroquinolone antibiotics in beef kidney. Samples were extracted with acidified methanol, centrifuged, diluted with water, and filtered. The diluted extract was analyzed directly by HPLC mass spectrometry using electrospray ionization in positive ion mode. Using an internal standard, mean recoveries were 73%96% at spiking levels of 33 g/kg (ppb), with statistically derived detection limits of 819 g/kg. This is below the European Union maximum residue limit of 200 g/kg for enrofloxacin and ciprofloxacin in bovine kidney. The method is evaluated relative to the requirements of the European Commission Decision 2002/657/EC for use as a confirmatory method.
European Community Commission Decision 2002/657/EC allows the use of HPLC coupled with fluorescence detection [4] for substances in Group B of Annex I to Directive 96/23/EC. Quinolones and other veterinary drugs fall into Group B, where three identification points are required for confirmation by Selected Ion Monitoring (SIM) using mass spectrometry (MS). With low resolution HPLC/MS, one point can be earned for each ion detected, provided that the ion ratios meet relative intensity criteria. Additional requirements of Directive 2002/657/EC, based on spiking levels of 33 g/kg carried out in this study, are as follows: The internal standard (IS) shall be added to the test portion at the beginning of the extraction procedure. In order to allow the use of data corrected for mean recovery, the range of recoveries allowed are 20% to +10%. The reproducibility of coefficient variation (CV) (%) is expected to be about one-half to twothirds of the 100 g/kg CV, which is 23%, at a concentration of half the permitted limit. For liquid chromatography/mass spectrometry (LC/MS) procedures, the minimum acceptable retention time (RT) for the analyte under examination is twice the RT corresponding to the void volume of the column. The ratio of the chromatographic RT of the analyte to that of the IS, that is, the relative RT of the analyte, shall correspond to that of the calibration solution at a tolerance of 2.5% for LC. The molecular ion shall preferably be one of the selected diagnostic ions. The maximum permitted tolerances for relative ion intensities shall meet the criteria in the Annex, (in this case, either 25% or 30%), as reproduced in Table 6.
Acidified methanol was prepared by adding 100 L of 98% formic acid to 100 mL of methanol. Acidified deionized water was prepared by adding 100 L of 98% formic acid to 100 mL deionized water. Ultra-Turrax T25 homogenizer, 50-mL polypropylene centrifuge tubes, and 13-mm polyvinylidene fluoride (PVDF) syringe filters (0.2 m), were purchased from VWR Scientific. All fluoroquinolones, including the IS, were provided as a gift from the Canadian Food Inspection Agency, Calgary, Alberta, Canada, as stock solutions of 100 ng/L (ppm) in 1% acetic acid in methanol. Solutions were stored at 4 C. Standard solutions at different concentrations were prepared for spiking by dilution with acidified methanol solution. The analytes ciprofloxacin, enrofloxacin, and sarafloxacin were chosen as targets since these compounds are included in the Canadian Food Inspection Agencys proficiency check samples. The spiking standard for these compounds (1 ng/L) was prepared by diluting 100 L of each the stock solutions to a 10-mL volumetric flask, and made to volume with acidified deionized water. A separate IS solution at 1 ng/L was prepared the same way, except that it only contained norfloxacin and danofloxacin. Sample Preparation 1. For beef kidney, 3 g samples were weighed directly into 50-mL polypropylene centrifuge tubes. 2. For spiked samples, 100 L of the 1-ng/L (100 ng) spiking solution was added, resulting in fortification levels of 33 g/kg. Samples were allowed to stand for 1 hour before subsequent extraction. 3. For the sample blank, 100 L of acidified methanol solution was added. 4. For all spiked samples, 100 L of the 1-ng/L (100 ng) IS solution was added just prior to extraction. Norfloxacin was included in this solution at the same level, to be used as an alternate IS, if required due to potential interferences for danofloxacin. 5. The samples were homogenized for 2 min with 15 mL of acidified methanol using the Ultra-Turrax homogenizer. 6. The samples were then centrifuged for 10 min, and the supernatant decanted into a clean test tube.
Experimental
Chemicals and Materials HPLC-grade methanol and acetonitrile were purchased from Caledon Labs (Georgetown, Ontario). Formic acid, min. 98%, was purchased from EM Science. Acidified methanol solution: 30% methanol in pH 3 deionized water (100 L of formic acid per 100 mL of water).
7. The extract was diluted with acidified deionized water 1 in 4 (250 L of extract + 750 L of water), filtered through a 0.2-m PVDF filter into an autosampler vial, and analyzed directly by LC/MS. By adding an accurately known amount of IS to the initial sample before extraction, there is no need to measure the final volume of the extracts, nor the aliquot to be diluted. The IS calculations, performed by the ChemStation, measure the relative amounts of the analytes and IS. This corrects for any concentration or dilution effects in the samples. Standard Preparation A 5-point calibration curve was used for the determination of each of the three target compounds, and a 1-point curve was used for norfloxacin, the alternate IS. Table 1 gives the volumes of the IS and target solutions added (1 ng/L each) to each of five test tubes. The standards were prepared by adding 250 L of the blank extract and 750 L of acidified deionized water to the tubes containing the analytes, after which the solutions were filtered through 0.2-m PVDF filters. The final solution of each standard contained 5 ng of IS per mL of diluted extract, or 5 pg/L. With 50 L injected, this results in 250 pg injected. The amount of target analyte in each of the five solutions varies to produce the calibration curves, as shown in Table 1. The correlation coefficient (R ) for the target analytes ranged from 0.9987 to 0.9992, as shown in Table 4. Preparation of the standards in this fashion will compensate for any ion suppression or enhancement that may occur, due to the presence of co-eluting material at the MS source, which may not otherwise occur if pure solvents alone are used.
Table 1. Preparation of Analytical Standards (50-L Injections into LC/MSD) Target volume added (L) 1 2 5 10 20 Target amount injected (pg) 50 100 250 500 1,000
2
LC/MS Conditions The HPLC system was made up of an Agilent Technologies 1100 series solvent degasser, binary pump, autosampler, column oven, diode array detector (DAD), and quadrupole mass selective detector (MSD) (Table 2).
Table 2. HPLC Column Zorbax Eclipse XDB-C8, 150 mm 4.6 mm, 5 m (P/N 993967-906) 0.1% Formic acid in water 0.1% Formic acid in acetonitrile t0 = 20% B t1 = 20% B t8 = 90% B t15 = 90% B Post time = 2.0 min 0.4 mL/min 50 L 30 C LC/MSD Conditions
Solvent A Solvent B Gradient
Flow rate Injection volume Column temp MSD Source Ion dwell time Fragmentation Drying gas flow Nebulizer pressure
Electrospray Ionization (ESI) (positive ion mode) 14 ions at 40 ms each Varies by ion, see Table 3 12 L/min 30 psi
Drying gas temperature 350 C Capillary voltage 4000 V
Table 3.
Fragmentor Voltages for Acquired Ions in SIM (single acquisition group) Ion 320 302 276 332 314 288 358 340 360 342 316 386 368 342 Fragmentor (V) 120 200 200 120 200 200 120 220 120 220 220 120 220 220 3
Compound Norfloxacin (IS)
Ciprofloxacin
Standard 1 2 3 4 5
IS Volume added (L) 5 5 5 5 5
IS Amount injected (pg) 250 250 250 250 250
Danofloxacin (IS) Enrofloxacin
Sarafloxacin
All ions were included in a single acquisition group, which started at injection (time = 0). An alternative approach would be to set the group start time to a value around half a minute before the elution of the first compound, as this will keep the eluant stream diverted to waste as long as possible. This will reduce the amount of co-extracted material being introduced into the source, reducing contamination. Another alternative is to add an additional timeprogrammed acquisition group to the method, and only include the ions for compounds eluting within the group times. This will take on more significance as the overall number of compounds in a method increases, and with three ions per compound required for identity confirmation. Fragmentor voltages were chosen that maximized the response for each selected ion. For each fluoroquinolone, a value of 120 V produced only the protonated parent ion, while higher voltages were required to induce fragmentation to confirmatory ions. The ions monitored corresponded to the neutral losses of water and carbon dioxide in each case. Note that although mass 342 is acquired for both enrofloxacin and sarafloxacin, it is only added to the MSD acquisition table once.
Chromatography All compounds eluted between 5 and 9 minutes, however the total run time was set to 15 minutes with 90% organic solvent to allow co-extractives to elute from the column. Otherwise, their eventual elution could interfere with subsequent injections. This is more of a potential problem when methods with abbreviated cleanups, such as dilution-only, are used. The following figures compare the blank beef kidney sample to a sample fortified at 33 g/kg. In each case, the selected ions are the protonated forms of the parent ion, as well as the protonated ions resulting from the loss of H2O (M-18) and CO2 (M-44). The qualifier ion for danofloxacin, the compound used as the IS for this study, is mass 340. The matrix causes an interference at mass 340. The interference is shown as a small peak in the beef kidney blank as shown in Figure 1. Since a diagnostic qualifier ion is not required for the IS calculations, it had no impact on the results. It does, however, indicate that there is elution of co-extractive material in the samples, and that without further cleanup, ion suppression may result from its presence. All standards were prepared in blank beef kidney extract in order to compensate for these potential effects.
Norfloxacin (IS) 320 = [M + H]+ 302 = [M - H2O + H]+ 276 = [M - CO2 + H]+
12000 8000 4000 0 12000 8000 4000 0 12000 8000 4000 0 5.5
5.530 - NOR 320
5.531- NORq1
302
5.532 - NORq2
6 6.5 7 7.5 8 8.5 9
276
min
Beef kidney blank
12000 8000 4000 0 12000 8000 4000 0 12000 8000 4000 0
5.484 - NOR 320
5.485 - NORq1
302
5.482 - NORq2
5.5 6 6.5 7 7.5 8 8.5 9
276
min
Figure 1.
Comparative extracted ion chromatograms for fluoroquinolones spiked into beef kidney.
Ciprofloxacin spike 332 = [M + H]+ 314 = [M H2O + H]+ 288 = [M CO2 + H]+
6000 4000 2000 0 6000 4000 2000 0 6000 4000 2000 0 5.5
5.854 - CIPRO 332
5.850 - CIPROq1
314
5.846 - CIPROq2
6 6.5 7 7.5 8 8.5 9
288
min
Beef kidney blank
6000 4000 2000 0 6000 4000 2000 0 6000 4000 2000 0 5.5 6 6.5 7 7.5 8 8.5 9
332
314
288
min
Danofloxacin (IS) 358 = [M + H]+ 340 = [M - H2O + H]+
8000 6000 4000 2000 0 8000 6000 4000 2000 0 5.5 6
6.009 - DANO 358 6.012 - DANOq1 340

6.5 7 7.5 8 8.5 9 min
Beef kidney blank
8000 6000 4000 2000 0 8000 6000 4000 2000 0 5.5 6
358
6.04 - DANOq1
6.5 7 7.5 8 8.5 9
340
min
Figure 1.
Comparative extracted ion chromatograms for fluoroquinolones spiked into beef kidney (Continued).
Enrofloxacin spike 360 = [M + H]+ 342 = [M H2O + H]+ 316 = [M CO2 + H]+
6000 4000 2000 0 6000 4000 2000 0 6000 4000 2000 0 5.5 6 6.5
6.910 - ENRO 360
6.914 - ENROq1
342
6.917 - ENROq2
7 7.5 8 8.5 9
316
min
Beef kidney blank
6000 4000 2000 0 6000 4000 2000 0 6000 4000 2000 0 5.5 6 6.5 7 7.5 8 8.5 9
360
342
316
min
8000 4000
8.604 - SARA 386 8.601- SARAq1 368 8.594 - SARAq2 342

5.5 6 6.5 7 7.5 8 8.5 9 min
Sarafloxacin spike 386 = [M + H]+ 368 = [M H2O + H]+ 342 = [M CO2 + H]+
0 8000 2000 0 8000 4000 0
8000 4000
386
Beef kidney blank
0 8000 4000 0 8000 4000 0
368
342
Figure 1.
Comparative extracted ion chromatograms for fluoroquinolones spiked into beef kidney (Continued).
Sarafloxacin elutes from the column in the same region as a number of other co-extractives, making identification and quantitation more difficult. However, as shown in Table 7, the qualifier ions still meet the identification criteria for relative responses of the qualifiers, and so further cleanup of the samples may not be necessary. The effect of these co-extractives will also be reduced at higher incurred residue levels, closer to those permitted by the European Union MRL. Recoveries In order to allow results to be corrected for recoveries, where the determined incurred levels are divided by the percent recovered from certified reference materials or spiked samples, Table 2 of the Annex requires that the recoveries for analytes at levels greater than 10 g/kg be within the range of 80% to 110%. Table 4 shows that recoveries for ciprofloxacin and enrofloxacin meet this requirement, with 96.3% and 86.0%, respectively. However, sarafloxacin fails the requirement, with only 72.6%
mean recovery. With a CV of only 8% for this compound, it looks as though the method may still produce acceptable results for screening purposes, but some additional work may be required to produce higher recoveries. Since the work presented here involves spiked samples only, recoverycorrection calculations do not apply. Norfloxacin was added along with danofloxacin as an additional IS. However, examination of the blank beef kidney used in this study shows norfloxacin to be present as an incurred residue, at a concentration approximately one half of the spiking level. Assuming a linear response through the origin, this would mean that norfloxacin was detected at approximately 1520 g/kg, which is about 10% of the permitted level for enrofloxacin in bovine kidney. Recoveries for norfloxacin are included in Table 4, even though they were calculated with a single point calibration, and not corrected for incurred residues. However, there is some compensation for this since the standards used for calibration were prepared by addition of the targets to the blank extracts.
Table 4.
Recoveries of Fluoroquinolones from Beef Kidney Amount recovered (ng) Ciprofloxacin Enrofloxacin 96.3 84.9 94.0 85.6 89.6 83.8 95.4 86.2 93.8 85.4 109.3 87.9 101.3 83.3 90.8 91.3 100.0 100.0 96.3 86.0 6.3 2.5 19.0 7.6 63.4 25.4 6.6 3.0 96.3 86.0 0.9987 0.9992 3.00 3.00
Description Kidney spike 1 Kidney spike 2 Kidney spike 3 Kidney spike 4 Kidney spike 5 Kidney spike 6 Kidney spike 7 Kidney spike 8 Amount spiked (ng) Mean SD (Precision) ng MDL (SD t-stat) ng LOQ (SD 10) ng CV (SD/Mean) % Accuracy (%) Linearity (R2) t-stat (N = 8)
Norfloxacin 111.8 93.1 88.0 98.9 82.2 143.0 102.6 110.6 100.0 103.8 18.9 56.7 189.1 18.2 103.8 0.9895 3.00
Sarafloxacin 68.5 64.1 77.6 75.2 82.1 72.9 73.0 67.7 100.0 72.6 5.8 17.4 58.1 8.0 72.6 0.9987 3.00
Compound Identification For chromatographic separation, Section 2.3.3.1 of the Annex to 2002/657/EC requires that the minimum acceptable RT for the analyte under investigation be at least twice the RT corresponding to the void volume of the column (k'=1). The first compound to elute under these conditions is norfloxacin, with a k' of 2.6, therefore this condition is easily met. The second condition is that the ratio of the RT of the analyte to that of the IS, that is the relative RT, shall correspond to that of the calibration solution at a tolerance of 2.5% for LC. Table 5 shows the RT times of each analyte in the spiked samples, compared to those of the standards, and that they are well within the allowable tolerance.
Table 5. Relative RTs of Analytes in Samples, Compared to Standards Average RRT in standards (N = 15) 0.922 0.975 1.150 1.439 CV (%) RRT in standards (N = 15) 0.12% 0.05% 0.16% 0.47% RRT in samples, relative to standards (N = 8) 99.8%100.1% 99.9%100.1% 99.8%100.2% 99.5%100.3%
Compound Norfloxacin Ciprofloxacin Enrofloxacin Sarafloxacin
Compound Confirmation Section 2.3.3.2 of the Annex to 2002/657/EC gives the maximum permitted tolerances for relative ion intensities, which is reproduced in Table 6.
Table 6.
Maximum Permitted Tolerances for Relative Ion Intensities Using a Range of Mass Spectrometric Techniques GC/MS(EI) (relative) 10% 15% 20% 50% GC/MS(CI), GC/MSn, LC/MS, LC/MSn (relative) 20% 25% 30% 50%
Relative intensity (% of base peak) >50% >20% to 50% >10% to 20% 10%
Note MSn equals MS/MS if n = 2
Table 7 shows the relative intensities for each of the qualifier ions for the three target compounds, as well as norfloxacin and danofloxacin (one ion). As expected, norfloxacin meets the criteria in each of the eight spiked samples, even though it had incurred residues. The presence of additional norfloxacin should not negatively affect this qualitative aspect of performance, and it does not. Danofloxacin, however, showed an interference for the single qualifier ion monitored, and so the relative amount of this signal would be expected to vary to a larger degree, depending upon the exact amount of blank extract used in preparing the
sample dilutions and standards. As previously mentioned, the standards are prepared by accurately measuring the relative amounts of target and IS compounds into a tube or vial, followed by addition of blank kidney extract and water. The exact proportions of extract and water do not have to be known, since the IS calculations uses amount and response ratios, rather than absolute amount and response, in determining concentrations in unknowns. An accurate measurement of extract and water volumes can, however, reduce interference variability.
Table 7.
Relative Intensities of Qualifier Ions for Fluoroquinolones in Beef Kidney, Compared to Permitted Tolerances Relative intensities (%) of qualifier ions Ciprofloxacin Danofloxacin Enrofloxacin Q1 = 314 Q2 = 288 Q1 = 340 Q1 = 342 Q2 = 316 47 17 64 44 28 48 17 58 46 24 46 20 63 44 28 46 17 60 43 30 45 21 65 44 26 39 17 72 45 29 41 18 65 46 29 42 19 62 39 24 44 20 86 43 26 3 2 22 2 2 25 30 20 25 25 33 14 69 32 19
Sample Spike 1 Spike 2 Spike 3 Spike 4 Spike 5 Spike 6 Spike 7 Spike 8 Average for Stds Std Dev for Stds Tolerance(Table 7) Lower Allowable (calculated) Upper Allowable (calculated)
Norfloxacin Q1 = 302 Q2 = 276 49 15 45 15 48 16 42 15 49 17 50 19 49 17 47 17 49 20 2 1 25 30 37 14
Sarafloxacin Q1 = 368 Q2 = 342 50 15 41 17 45 14 43 13 45 11 43 12 46 13 46 12 47 15 6 1 25 30 35 11
62
26
55
25
103
53
32
59
20
Conclusion
A fast and sensitive single quadrupole LC/ESI/MS method was validated for the detection of three fluoroquinolone antibiotics (ciprofloxacin, enrofloxacin, and sarafloxacin) in beef kidney. The detection limits ranged from 8 to 19 g/kg (ppb), with direct analysis of sample extracts after dilution with water. All qualitative requirements were met with respect to the Annex to EU Directive 2002/657/EC for spiked samples, and recoveries of two of the three compounds met the quantitative requirements. Recovery of sarafloxacin was slightly lower than the level required to allow correction for recoveries in reported results.
References
1. The European Agency for the Evaluation of Medicinal Products, Veterinary Medicines and Inspections, EMEA/MRL/820/02-FINAL, January 2002. 2. Determination of Fluoroquinolones in Bovine, Porcine and Avian Tissues by Liquid Chromatography with Fluorescence Detection, FQL-SP04, Canadian Food Inspection Agency, Saskatoon, Saskatchewan, Canada; 2001/03. 3. Chen, G., Schneider, M. J., (2003) A Rapid Spectrofluorometric Screening Method for Enrofloxacin in Chicken Muscle. J. Agric. Food Chem., 51(11), 3249-3253. 4. Annex of Commission Decision 2002/657/EC, Commission Decision of 12 August 2002, implementing Council Directive 96/23/EC concerning the performance of analytical methods and the interpretation of results, Official Journal of the European Communities, 17.8.2002, L 221/8-36, Table 5, Footnote 4.

A Validated Atmospheric Pressure Chemical Ionization Method for Analyzing Sulfonamides in Pork Muscle Application
Food
Author
Ralph Hindle Access Analytical Labs #3, 2616 - 16 Street N.E. Calgary, AB. T2E 7J8 Canada Agilent contact: Chin-Kai Meng Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA
Introduction
Meat, edible organs, animal feed, and animal waste may contain antibiotics, growth hormones, and other chemicals that can enter the food supply. These compounds are added to maintain animal health, to increase animal growth rate, and to reduce stress. Human exposure can result from eating contaminated meat, or contacting runoff and leaching from manure and compost. Health specialists warn that there may be reduced options for effectively treating disease with antibiotics, such as penicillin and sulfa drugs, since antibioticresistant strains of bacteria may develop from the low-level exposure. Sulfonamides are broad-spectrum antimicrobials used in both humans and animals. The maximum residue limit (MRL) in Canada for sulfonamides in meat is 100 ppb (ng/g), and 10 ppb in milk, while the MRL in the European Union is 100 ppb for both of these matrices. The Canadian Food Inspection Agency method for sulfonamides in meat tissue calls for extraction in ethyl acetate, partitioning with glycine buffer, followed by a pH-adjusted back extraction into methylene chloride [1]. Extracts are evaporated, reconstituted, then separated by thin layer chromatography (TLC), derivatized, and quantitated by densitometry. Alberta Agriculture has improved the quantitative and qualitative aspects by using liquid chromatography/mass spectrometry (LC/MS) with atmospheric pressure chemical ionization (APCI) for the final analysis [2]. There are a number of
Abstract
This application note presents a simple method for the analysis of sulfonamide antibiotics in pork muscle. Samples were extracted with acidified methanol, centrifuged, and a portion of the extract was diluted with water. This dilution was analyzed directly by HPLC mass spectrometry using chemical ionization, with all compounds eluting in less than 5 minutes. Using an internal standard, recoveries for seven sulfonamides ranged from 84%118% at a spiking level of 50 ppb (ng/g). The statistically derived detection limit was 1025 ppb. A comparison was made to the cleaned extracts using solid phase extraction, as well as a comparison of mass selective detector settings for both screening (maximum sensitivity) and confirmation (greater fragmentation). The enhanced sensitivity of the Agilent quadrupole mass selective detector allows this dilution cleanup technique to be used in labs where high throughput is required.
extraction steps in the Alberta method, and a faster method would greatly benefit laboratories monitoring the food supply for residues. The goal of this method was to reliably quantitate the sulfa drugs at one-half of the regulatory limit or lower, with minimal sample preparation, and a maximum injection cycle time of 10 minutes. Maximum sensitivity is generally obtained by forming as many parent ions [M+H]+ as possible and minimizing fragmentation. Due to the operational complexity of triple quadrupole instruments, it is also desirable to confirm positive findings on a single quadrupole. This could be achieved by using collision induced dissociation (CID) to enhance fragment ions characteristic of the compounds.
Sample Preparation 1. For pork muscle, 3 g samples were weighed directly into 50-mL polypropylene centrifuge tubes. 2. The samples were homogenized for 3 minutes with 10 mL acidified methanol using the UltraTurrax homogenizer. 3. The samples were then centrifuged for 10 minutes, and the supernatant decanted into a clean test tube. 4. The samples were then re-extracted with a further 10 mL acidified methanol, and centrifuged again. 5. The supernatants were combined, and 1 mL IS (2 mg) was added to the combined extract. 6. The extract was diluted with de-ionized water 1 in 4 (250 L extract + 750 L water), filtered through a 0.2 m PVDF filter into an autosampler vial, and analyzed directly by LC/MS. By adding an accurately known amount of IS to the combined extracts, there is no need to measure the final volume of the extract. The IS calculations performed by the ChemStation measure the relative amounts of the analytes and IS. This corrects for any concentration or dilution effects in the samples. Sample extracts were also taken through SPE cleanup cartridges in order to compare with the dilution-only extracts. The 60-mg Oasis HLB cartridges are prewashed by eluting 1.5 mL acidified methanol, followed by 1.5 mL de-ionized water. The 1 mL extract was diluted to 10 mL with de-ionized water, eluted through the cartridges, and the eluant was discarded. The sulfa drugs were then eluted with 1.5 mL acidified methanol. This eluant was evaporated to near dryness under nitrogen. Samples were reconstituted in 1 mL of 25% methanol in water, filtered, and analyzed by LC/MS. A further comparison was done by evaporating 1 mL methanol extract to near dryness, and reconstituting it in 1 mL of 25% methanol in water without the SPE cleanup. This gave the sample extract the proper solvent composition for HPLC analysis, but without the dilution step to negatively affect the detection limits (DL) of the compounds.
Experimental
Chemicals and Materials All sulfonamide standards were purchased from Sigma Aldrich Canada, with a minimum purity of 99%. Stock solutions were prepared at 2 mg/mL in acetone, with the exceptions of sulfadiazine and the sodium salt of sulfaquinoxaline. Three mL of 0.2N NaOH was added in order to completely dissolve these compounds. Standard solutions at different concentrations were prepared for spiking and quantitation by diluting with de-ionized water. Internal standard (IS): sulfachloropyridazine (SCPD) at 2 mg/mL in de-ionized water. HPLC-grade methanol and acetonitrile were purchased from Caledon Labs (Georgetown, Ontario). Formic acid (min. 98%), was purchased from EM Science. Acidified methanol was prepared by adding about 100 L of 98% formic acid to 100-mL methanol. Ultra-Turrax T8 homogenizer with 8-mm diameter dispersing element, 50-mL polypropylene centrifuge tubes, and 13-mm polyvinylidene fluoride (PVDF) syringe filters (0.2 m), were purchased from VWR Scientific. Oasis HLB (3 cc, 60 mg) solid phase extraction (SPE) cartridges were purchased from Waters.
LC/MS Conditions
The LC/MS system was made up of Agilent Technologies 1100 Series solvent degasser, binary pump, autosampler, column oven, diode array detector, and quadrupole mass selective detector (MSD) (Table 1).
Compound Identification and Confirmation

In general, the goal of a monitoring method for target analytes is to separate the compounds from potential interferences and maximize sensitivity on the instrument. Using mass spectrometry (MS), maximum sensitivity is achieved by the production of a single ion, for example, the protonated parent
ion [M+H]+ in LC electrospray ionization (ESI) or APCI in target ion mode. However, once a positive is detected, a confirmation must be made as to whether the suspect peak is actually the target analyte, or simply a co-eluting compound that produces the same ion. There are a number of ways to perform the confirmation: re-extract the sample with a different solvent system; further clean up the sample to a higher final concentration, to allow detection of additional confirmation ions or analysis in scan mode; derivatize and analyze by gas chromatography/mass spectrometry (GC/MS); or re-analyze the extract on a triple quadrupole LC/MS/MS. All of these techniques are useful, but the drawback is the additional time and expense involved, especially with LC/MS/MS.
Table 1. HPLC Column Solvent A Solvent B Gradient
LC/MSD Conditions Zorbax Eclipse XDB-C8, 150 mm 4.6 mm, 5 m (p/n 993967-906) 0.1% Formic acid in water 0.1% Formic acid in acetonitrile t0 = 20% B t1 = 20% B t3 = 90% B t6.5 = 90% B Post time = 1.5 min 1.0 mL/min 50 L 30 C APCI (positive ion mode) 8 Ions at 63 ms each 70 V 6.0 L/min 60 psi 350 C 400 C 3000 V 4 A
Flow rate Injection volume Column temp MSD Source Ion dwell time Fragmentor Drying gas Nebulizer pressure Drying gas temperature Vaporizer temperature Capillary voltage Corona current
The Agilent 1100 MSD has the capability of acquiring up to four separate MS signals during the same run, where each signal can be made up of a number of selected ions (SIM) or a full scan spectrum. For example, Signal 1, with a low fragmentor voltage to maximize parent ion response, can include each of the [M+H]+ ions in the target list, while Signal 2, at higher fragmentor voltages can acquire the confirmatory fragment ions. For analytes expected at higher concentrations, Signal 1 could acquire in SIM mode for quantitation, while Signal 2 could be set for scan mode for identification. Figure 1 demonstrates the former example, with the Fragmentor set to 70 V for Signal 1 (MSD1), and 200 V for Signal 2 (MSD2).
MSD1 250, EIC=249.7:250.7 (SULFAMSD\SSSCI01.D) APCI, Pos, SIM, Frag: 70, "Quantitation"
8000 4000 0 2.5 3.0
3.365 - SPY
Simultaneous 2-signal aquisition; Fragmentor at 70 V or 200 V
3.5
4.0
4.5
5.0
5.5
min
MSD1 285, EIC=284.7:285.7 (SULFAMSD\SSSCI01.D) APCI, Pos, SIM, Frag: 70, "Quantitation"
8000 4000 0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 min
4.344 - SCPD (IS)
MSD2 108, EIC=107.7:108.7 (SULFAMSD\SSSCI01.D) APCI, Pos, SIM, Frag: 200, "Confirmation"
3000 2000 1000 2.5 3.0 3.5 4.0 4.5 5.0 5.5 min
MSD2 156, EIC=155.7:156.7 (SULFAMSD\SSSCI01.D) APCI, Pos, SIM, Frag: 200, "Confirmation"
700 500 300 2.5 3.0 3.5 4.0 4.5 5.0 5.5 min
Figure 1.
Dual MSD acquisition signals (Masses 108 and 156 are class-specific fragments for sulfonamides).
Table 2 shows the mass spectra for the sulfonamides using various fragmentor voltages. Masses 108 and 156 are class-specific fragments for sulfonamides (H2N+=[C6H4]=O and H2N+=[C6H4]=SO2, respectively), and, as such, are very useful diagnostic ions, when acquired along with the protonated molecular ion.
Table 2.
Compound
APCI Spectra of Sulfonamides, Using Various Fragmentor Voltages

Fragmentor 70 V
*MSD1 SPC, time=2.859:3.513 of SFCISCAN\SULFA006.D APCI, Pos, Scan, Frag: 70
Fragmentor 160 V
Fragmentor 200 V
156.1
256.1
100
101.1 108.2
0
100 200 300
m/z
0 100 200 300
107.5
m/z
0 100 150 200 250
256.1 m/z
20
20
157.1
257.1
20
251.1
251.1
252.1
158.1
252.1
0
100 200
300 m/z
0 100 200
185.1
20
20 300 m/z 0
100
150
200
m/z
250.1
108.1
156.1
120.1
101.2
251.1
20 0
108.1
184.1
251.1
20 m/z 0
20 m/z 0 100
157.1
183.4
40
40
40
107.4
Sulfapyridine (SPY) C11H11N3SO2 MW = 249 RT = 3.33 min
80 60
250.1
80 60
80 60 156.1
184.1
100
Max: 35037
100
Max: 16213
100
Max: 2943
100
150
200
250
100
200
300
150
200
m/z
265.1
265.1
156.1 172.1 184.1 199.1
101.2 117.1
110.2
250.1
266.2
20 0
20
300 m/z
155.3 172.1
20
300 m/z
100
200
100
200
109.4 120.1
266.1
265.1
Sulfamerazine (SMR) C11H12N4SO2 MW = 264 RT = 3.78 min
80 60 40
80 60 40
80 60 40
108.2
110.2
100
Max: 20181
100
Max: 9216
100
Max: 4083
100
200
300
m/z
250.1
251.1
20
158.0
185.1
40
107.5 109.1 120.1
40
40
107.5 108.2
156.1
Sulfadiazine (SDZ) C10H10N4SO2 MW = 250 RT = 3.09 min
60
60
156.1
80
80
80 60
108.1
100
Max: 66929
100
Max: 23650
100
107.5 120.1
40
257.1
40
40
156.1
Sulfathiazole (STZ) C9H9N3S2O2 MW = 255 RT = 3.05 min
80 60
80 60
256.1
Max: 151737
100
Max: 69275
100 80 101.1 60
108.1
Max: 32290
Max: 10375
Table 2.
Compound
APCI Spectra of Sulfonamides, Using Various Fragmentor Voltages (Continued)

Fragmentor 70 V
Fragmentor 160 V
Fragmentor 200 V
100
Max: 57189
279.1
100 80 60 40
Max: 35356
279.1
100 80 60
108.1 123.5 124.2
Max: 11707
279.1 156.1 186.1 213.2 280.1 300 m/z 200
280.1
124.2
20 0
100 200
20
m/z
186.1
280.1
Sulfamethazine (SMZ) 60 C12H14N4SO2 MW = 278 40 RT = 4.06 min
80
40 20
m/z
300
100
200
300
100
100
285.0
Max: 15461
100
156.0
Max: 6222
285.0
111.2 123.3 124.2 135.3 138.1
40 20
m/z
163.1 179.1
107.4 120.1 130.1
40 20 0
40 20
265.1 284.3 287.0
101.2 115.2
100
200
300
m/z
100
200
300
184.2
Sulfachloropyridazine (SCPD); istd C10H9N4SO2Cl MW = 284 RT = 4.31 min
80 60
80 60
287.1
80 60
108.1
100
Max: 3906
156.0
100
150
200
250
m/z
301.1
301.1
100 80
80 60
302.1 302.1
80 60 40 20
m/z
40 20 0
100 146.1
40
108.2 129.1
107.5 120.1 129.1 145.3
235.1
20
m/z
200
300
100
200
300
288.4 302.1
Sulfaquinoxaline (SQ) C14H12N4SO2 MW = 300 RT = 4.51 min
156.1
301.1
60
156.1
Max: 148192
100
Max: 63906
100
108.2
Max: 16120
100
200
300
m/z
311.1
311.1
156.1
100
Max: 467058
100 80 60
156.1
Max: 261189
100 80 60
311.1 218.1 245.2
Max: 80490
312.1
312.1
156.1
20 0
100
20
m/z
20
m/z
200
300
100
200
300
157.1
312.1
Sulfadimethoxine (SDMX) C14H12N4SO2 MW = 310 RT = 4.54 min
80 60 40
108.1
40
40
155.3
100
200
300
m/z
Chromatography
While complete separation of target compounds is not always necessary when using mass spectral detection, it is, however, essential when common ions are present. For example, the protonated molecular ion of SPY is 250 mass units. Due to the naturally-occurring C13 isotope, ions 251 coexist with the parent ions 250. Separating SPY from SDZ (m/z = 251) was, therefore, important when trying to optimize the chromatographic conditions, and was achieved as shown in Figure 2. While this results in a slightly longer chromatographic run than would otherwise be necessary, there is more consistent integration of the peaks during data
analysis; the chromatogram is easier to interpret; and the amount of SDZ is not underestimated due to co-elution of SPY in the standard mix. A recently published application shows four sulfonamides were analyzed with an injection cycle time of 1.1 minutes, using a 2-position 10-port valve, two analytical columns in parallel, and a second binary pump [3]. Since most labs do not have such high sample volume requirements, the method described in this application note was developed using more conventional techniques, without the additional hardware costs. Conditions were set up to provide good chromatographic separation in a relatively short time of 6 minutes (total cycle time was 10 minutes).
MSD1 256, EIC=255.7:256.7 (SULFAMSD\SSS_CI5.D) APCI, Pos, SIM, Frag: 70
3.051 - STZ
1
2
min
3.084 - SDZ
1
2
min
3.332 - SPY
1
2
min
3.783 - SMR
1
2
min
4.056 - SMZ
4 5 min
4.314 - SCPD (IS)

4 5 min
4.514 - SQ
4 5 min
4.537 - SDMX
4 5 min
Figure 2.
Sulfonamide standard mix, 500 pg each (SIM).
Sample Cleanup
The total ion chromatograms (TIC) in Figure 3 show that there is considerable matrix background from the samples. A simple solvent exchange was performed, where 1 mL of extract was evaporated under nitrogen, and reconstituted in 25% methanol in water. One of the problems with solvent exchange only is the amount of matrix material that is injected onto the HPLC column. Peak shape can be negatively affected by overloading, and eventually the performance of the column will deteriorate. All of this matrix material is also introduced into the MSD. Frequent cleaning and maintenance may be required for the MSD, further reducing productivity.
In order to develop a high-throughput method, keep the number of required steps to a minimum. The Agilent liquid chromatography/mass selective detector (LC/MSD) has enough sensitivity to allow simple dilution of the extracts with water to act as a cleanup technique. This eliminates the need for costly SPE cartridges and analyst time to further prepare the samples. Minimal sample handling can also improve recoveries, since losses are possible at each step. The third chromatogram in Figure 3 shows how the use of SPE cleanup techniques can remove the majority of co-extracted materials, allowing for a more concentrated final extract and ultimately lower DLs. This also results in a simpler chromatogram for integration and interpretation.
MSD1 TIC, MS File (SF030817\SULFA018.D) APCI, Pos, SIM, Frag: 70 200000 150000 100000 50000 0 1 2 3 4 5 6 min
Spiked pork extract - solvent exchange to 25% MeOH in water
Spiked pork extract - Diluted 1 in 4 with water
Spiked pork extract - Oasis HLB cleanup
Figure 3.
TIC comparisons of various cleanup techniques.
However, where the goal of a method is to screen large numbers of samples to find potential violations of MRLs, a simple dilution technique may be preferred. Dilution could offer enough cleanup for good chromatographic separation, while remaining concentrated enough to meet DL requirements. The second chromatogram in Figure 3 shows a much improved baseline. Figures 4 through 6 show the same analyses with all the target ions in SIM mode.
MSD1 256, EIC=255.7:256.7 (SF030817\SULFA018.D) APCI, Pos, SIM, Frag: 70
3 .0 5 4 - S T Z 3.0 3.5 4.0 4.5 5.0 5.5 min
3.094 - SDZ
3.0 3.5 4.0 4.5 5.0 5.5 min
3.340 - SPY
3.0 3.5 4.0 4.5 5.0 5.5 min
3.794 - SMR
3.0 3.5 4.0 4.5 5.0 5.5 min
3.180
3.0 3.5 4.0
4.083 - SMZ
4.5
4.760
5.0 5.5 min
4.345 - SCPD (IS)

3.0 3.5 4.0 4.5 5.0 5.5 min
4.549 - SQ
3.0 3.5 4.0 4.5 5.0 5.5 min
4.572 - SDMX
3.0 3.5 4.0 4.5 5.0 5.5 min
Figure 4.
Solvent exchange only (SIM)
3.044 - STZ
3.0 3.5 4.0 4.5 5.0 5.5 min
3.078 - SDZ
3.0 3.5 4.0 4.5 5.0 5.5 min
3.324 - SPY
3.0 3.5 4.0 4.5 5.0 5.5 min
3.782 - SMR
3.0 3.5 4.0 4.5 5.0 5.5 min
3.139
3.0 3.5 4.0
4.056 - SMZ
4.5 5.0 5.5 min
4.316 - SCPD (IS)

3.0 3.5 4.0 4.5 5.0 5.5 min
4.517 - SQ
3.0 3.5 4.0 4.5 5.0 5.5 min
4.541 - SDMX
3.0
3.5
4.0
4.5
5.0
5.5
min
Figure 5.
Diluted 1 in 4 with water.
10
3.042 - STZ
3.0 3.5 4.0 4.5 5.0 5.5 min
3.078 - SDZ
3.0 3.5 4.0 4.5 5.0 5.5 min
3.322 - SPY
3.0 3.5 4.0 4.5 5.0 5.5 min
3.777 - SMR
3.0 3.5 4.0 4.5 5.0 5.5 min
3.166
3.0 3.5 4.0
4.077 - SMZ
4.5 5.0 5.5 min
4.343 - SCPD (IS)

3.0 3.5 4.0 4.5 5.0 5.5 min
4.546 - SQ
3.0 3.5 4.0 4.5 5.0 5.5 min
4.569 - SDMX
3.0 3.5 4.0 4.5 5.0 5.5 min
Figure 6.
After HLB cleanup (SIM).

The recoveries obtained for seven samples spiked at a level of 50 ppb (150 ng of each sulfonamide in 3 g sample) appear in the following tables. The spiking solutions were added before homogenization and allowed to stand for at least 30 minutes before extraction. SMR (sulfamerazine) was added separately at 300 ng per sample before homogenization, and could be used as a surrogate. Results in Table 3 were obtained by simply diluting the extracts 4-fold with water (recovery 84%118%), while results in Table 4 are from extracts taken through SPE cleanup (recovery 79%104%).
In both cases, a five-point IS calibration with SCPD was used, with 20 to 200 pg of each target compound injected, plus 2,000 pg SCPD. The five standards were injected both before and after the set of seven spikes, and the curves were created by using the average responses of the two sets of standards. Peak height was used to measure response, as there was less variability compared to peak area, due to the noticeable tailing of these compounds. The linearity results (R2) are tabulated in Tables 3 and 4.
11
Table 3.
Recoveries of Sulfonamides by Diluting Extracts 1 in 4 with Water Amount recovered (ng) STZ 167 168 160 158 151 147 144 150 156 9 29 94 6 104 0.9997 3.14 SDZ 172 197 183 189 169 161 72 150 178 13 40 126 7 118 0.9996 3.14 SPY 164 68 158 167 154 144 141 150 157 11 34 108 7 104 0.9997 3.14 SMR 317 343 315 336 295 322 272 300 314 24 77 245 8 105 0.9972 3.14 SMZ 151 164 157 156 169 143 151 150 156 9 28 88 6 104 0.9996 3.14 SCPD(IS) 2,000 2,000 2,000 2,000 2,000 2,000 2,000 2,000 2,000 100 1.0000 3.14 SQ 148 169 133 138 133 120 124 150 138 17 53 167 12 92 0.9984 3.14 SDMX 130 137 121 129 129 112 125 150 126 8 26 82 7 84 0.9992 3.14
Description Pork spike 1 Pork spike 2 Pork spike 3 Pork spike 4 Pork spike 5 Pork spike 6 Pork spike 7 Amount spiked (ng) Mean SD (Precision) MDL (SD t-stat) ng LOQ (SD 10) ng RSD (SD 100/Mean) Accuracy (%) Linearity (R ) t-stat (N=7)
2
Table 4.
Recoveries of Sulfonamides Using Oasis HLB Cleanup Cartridges Amount recovered (ng) STZ 161 154 149 145 151 136 148 150 149 8 24 76 5 99 0.9994 3.14 SDZ 157 156 158 152 162 147 161 150 156 5 17 53 3 104 0.9994 3.14 SPY 132 132 124 122 127 127 128 150 127 4 11 36 3 85 0.9997 3.14 SMR 273 293 267 279 294 274 275 300 279 10 33 104 4 93 0.9979 3.14 SMZ 149 157 155 144 149 136 155 150 149 7 23 73 5 100 0.9998 3.14 SCPD(IS) 2,000 2,000 2,000 2,000 2,000 2,000 2,000 2,000 2,000 100 1.0000 3.14 SQ 139 153 132 119 127 116 124 150 130 13 40 128 10 87 0.9989 3.14 SDMX 126 131 113 111 121 108 116 150 118 8 26 82 7 79 0.9989 3.14
Description Pork spike 1 Pork spike 2 Pork spike 3 Pork spike 4 Pork spike 5 Pork spike 6 Pork spike 7 Amount spiked (ng) Mean SD (Precision) MDL (SD t-stat) ng LOQ (SD 10) ng RSD (SD 100/Mean) Accuracy (%) Linearity (R ) t-stat (N=7)
2
12
Table 5 summarizes the comparison of recoveries when diluted with water versus using Oasis HLB cartridge cleanup. Generally there is a greater difference in recoveries for the early eluting compounds, as one might expect. Since the samples are loaded onto the cartridge with a mostly aqueous phase (10% methanol in water), the water-soluble matrix components would tend to pass through the cartridge to waste. Because these early eluting compounds were removed prior to injection on the HPLC column, the chromatograms are cleaner with more reproducible chromatography, as shown by the smaller standard deviations in recoveries. The results from the HLB cleanup exhibited smaller standard deviations and lower minimum detection levels (MDLs).
when SPE cleanup is used. Instrumental conditions allow injection cycle-time of 10 minutes using typical columns and conditions for most labs.
References
1. TLC-Densitometric Procedure for Sulfonamide Residues in Animal Tissue, SUL-SP08, Canadian Food Inspection Agency, Saskatoon, Saskatchewan, Canada; 2001/04. 2. Sulfonamides in Tissue by LC/MS, Alberta Agriculture, Edmonton, Alberta, Canada, Standard Operating Procedure TX-0278-01. 3. Mark Stahl, High-throughput analysis with the Agilent 1100 Series high-throughput LC/MS system, Agilent Technologies, publication 5988-9638EN. www.agilent.com/chem
Conclusion
A fast and sensitive single quadrupole LC/APCI/MS method was developed and validated for detection of sulfonamide residues in pork. The DL ranged from 10 to 25 ng/g of tissue when analyzed by simple dilution of the extracts, and 4 to 13 ng/g

Table 5.
Comparison of Recoveries Obtained by Dilution vs Oasis HLB Cleanup STZ 104 9.4 29 99 7.6 24 SDZ 118 12.6 40 104 5.3 17 SPY 104 10.8 34 85 3.6 11 SMR 105 24.5 77 93 10.4 33 SMZ 104 8.8 28 100 7.3 23 SCPD(IS) 100 100 SQ 92 16.7 53 87 12.8 40 SDMX 84 8.2 26 79 8.2 26
Description Accuracy % (1 in 4 dilution) SD (Precision) MDL (ng) Accuracy % (HLB cleanup) SD (Precision) MDL (ng)
13
Detection, Confirmation, and Quantification of Chloramphenicol in Honey and Shrimp at Regulatory Levels Using Quadrupole and Ion Trap LC/MS Application
Foods, Environmental
Authors
Gerd Vanhoenacker, Frank David, Pat Sandra Research Institute for Chromatography Kennedypark 20, B-8500 Kortrijk, Belgium Non-Author Contact Jerry Zweigenbaum Agilent Technologies, Inc. 2850 Centerville Road, Wilmington DE 19808-1610 USA e-mail: Jerry_Zweigenbaum@Agilent.com
Abstract
Methodology capable of meeting regulatory requirements has been developed for the determination of chloramphenicol in honey and shrimp. Samples of the two foodstuffs are extracted with Isolute HN-M cartridges and analyzed with both the Agilent 1100 LC/MSD Trap (SL) and the Agilent 1100 LC/MSD (SL) quadrupole with negative mode electrospray ionization. Using deuterated internal standard and one simple sample extraction procedure, both instruments provide a limit of detection at or below 0.1 ppb in both shrimp and honey. Detection limits are lower using the ion trap for shrimp because of less matrix interference. The Agilent 1100 LC/MSD gives quantitative results and the Agilent 1100 LC/MSD Trap gives full spectrum confirmation.
and shrimp. Because it has displayed significant toxicological effects on humans, it has been banned from foods in the European community and the United States at levels greater than 0.1 ppb. Analytical methods used to determine this limit must achieve both the required sensitivity and maintain sufficient selectively. LC/MS has been demonstrated by the US Food and Drug Administration for these analysis [1-3]. In addition, the Commission of European Communities has issued guidelines stipulating that for mass spectral detection, a molecular ion (or quasimolecular ion) and at least two fragment ions are needed for positive confirmation [4]. For quantitative analysis the Agilent 1100 LC/MSD provides excellent results and can give some confirmation information. The Agilent 1100 LC/MSD Trap gives excellent full spectrum confirmation at the regulated concentration.
Experimental
Reagents and Materials ISOLUTE HM-N cartridges from IST (Hengoed, UK, Part-nr. 800-1300-FM) Ethyl acetate from Vel (Merck Eurolab, Leuven, Belgium) Methanol HPLC-grade from Merck (LiChrosolv, Darmstadt, Germany) Deuterated (d5) CAP internal standard from Cambridge Isotope Laboratories (CIL, Andover, MA, USA) Syringe filters (0.2 m, PTFE) from Alltech Associates Inc. (Lokeren, Belgium)
Introduction
Chloramphenicol is a broad range antibiotic that has found its way into foodstuffs such as honey
Sample Preparation For honey, 5 g of sample is diluted to 20 mL with water and 5 L of 1 ng/L internal standard (IS) is added. The solution is loaded on the cartridge and allowed to stand for 5 minutes. Elution is performed with 50 mL ethyl acetate. The eluate is collected and the solvent is evaporated under a nitrogen stream at 40 C. The residue is redissolved in 1 mL water/methanol (9/1, v/v) and put in an ultrasonic bath for 1 minute. The solution is filtered, using a syringe filter, before injection. No additional clean-up of the sample solution is performed. For shrimp, a portion of at least 10 g of frozen shrimp is defrosted and mixed in a blender. To 10 g of the mixed shrimp, 30 mL of water and 10 L of
1 ng/L IS is added. This portion is centrifuged for 10 minutes (2000 rpm). A 20-mL portion of the supernatant is loaded on the cartridge and allowed to stand for 5 minutes. Elution is performed with 50 mL ethyl acetate. The eluate is collected and the solvent evaporated under a nitrogen stream at 40 C. The residue is redissolved in 1 mL water/methanol (9/1, v/v) and put in an ultrasonic bath for 1 minute. The solution is filtered before injection. LC/MS Conditions The LC/MS systems were the Agilent 1100 LC/MSD quadrupole mass spectrometer and the Agilent 1100 LC/MSD Trap. Both were equipped with Agilent 1100 binary pumps and 1100 well plate autosamplers. See Table 1.
Table 1. HPLC Column Flow-rate
LC/MS Conditions
Eclipse XDB C18, 4.6 mm 150 mm, 5 m (p/n 993967.902) 0.9 mL/min 10 mM ammonium acetate in water (solvent A) Methanol/acetonitrile 1/9 (solvent B) both from Merck (LiChrosolv, Darmstadt, Germany) 01 min 18 min 88.5 min 8.512 min Post time 30% B 30%70% B 70%100% B 100% B 4 min at 30% B
Mobile phase
Gradient
Injection Injection solvent Column temperature MSD source settings Source Ion polarity Drying gas temperature Drying gas flow-rate Nebulizer pressure Vcap Quadrupole MSD MSD acquisition on Fragmentor SIM settings
100 L with needle wash (methanol) Water/methanol (9/1 v/v) for both standards and samples 30 C ESI Negative 340 C 11 L/min 50 psig 3500 V Between 3 and 7.5 min 160 V m/z 257, 321, 323 (CAP) m/z 262, 326, 328 (CAP-d5)
Table 1. Trap MSD
LC/MS Conditions (continued)
MSD acquisition on Target mass (SPS) Trap parameters Max. accumulation time ICC target Scan range Averaging
Between 3 and 7.5 min 323 m/z
300 ms 30,000 160340 2
Fragmentation parameters (MS/MS) Smart Frag Isolation mass Isolation width Fragmentation amplitude Fragmentation cutoff On, 30%200% (default) m/z 325.0 10.0 m/z 1.0 V m/z 88

Spectral Quality and Sensitivity of Standards For analysis with the quadrupole LC/MSD, selected ion monitoring (SIM) was used to obtain the required sensitivity. Table 2 shows the structure, fragment ions and identity of CAP and CAP-d5. Figure 1 shows the analysis of a standard mixture containing 2.5 pg/L CAP and 5 pg/L CAP-d5. By applying a fragmentor voltage of 160 V, fragment ions at m/z 257 and 262 are detected for confirmation purposes. Lowering the fragmentor voltage to optimize for the m/z 321 and m/z 326 and monitoring those ion alone would obtain greater sensitivity. However, the confirmation of the fragment ions would be lost. For screening analysis without confirmation this would be acceptable and provide a much lower limit of detection (LOD).
Table 2
Structure and Fragment Ions and Identity of CAP and CAP-d5 (* Indicates Deuterated Positions for the CAP-d5 IS)
Chloramphenicol structure OH * * O N+ * O* * H N Cl OH O Cl
m/z CAP 257 249 194 176 262 254 199 180
Identity [M-H-HCOCl] [M-H-2HCl] [M-H-NH2CoCl2H] [M-H-NH2CoCl2H-H2O] [M-H-HCOCl] [M-H2HCl] [M-H-NH2COCl2H] [M-H-NH2COCl2H-HDO]
CAP-d5
7000
CAP (2.5 pg/L)
321 m/z
4000
323 m/z
257 m/z
1000
14000
CAP-d5 (5 pg/L)
326 m/z
8000
328 m/z
2000
262 m/z
Time (min)
Figure 1.
Analysis of a standard solution containing 2.5 ppb of CAP and 5 ppb of CAP-d5 (IS) on the quadrupole MSD. The extracted ion chromatogram for the corresponding ions are shown.
Using the LC/MSD Trap in MS/MS mode both the needed sensitivity (through reduction in chemical noise) and selectivity (for confirmation) is obtained. The compound shows a clear and reproducible fragmentation pattern. An example of the analysis of the standard mixture together with the corresponding MS/MS spectra is shown in Figure 2. Optimizing the fragmentation energy [turning off Smart Frag] and fragmentation cutoff in the ion trap will increase sensitivity even further than shown here. Using an isolation width of 10 m/z allows inclusion of the chlorine isotopes in the resulting full scan mass spectra of the analyte and the Cl35 isotope of the internal standard. Contact Agilent for more details on these and other ion trap settings.
CAP: EIC 176, 194, 249, 257 m/z (2.5 pg/L)

1000
2000
CAP-d5: EIC 180, 199, 254, 262 (5 pg/L)
1000
0 4
Time (min)
2.5 pg/L CAP
257
194
5 pg/L d5-CAP6 262 249
176
160 240
0.2 pg/L CAP (LOD)
257 180
160
199
240
254
320 m/z
176
194 249
160
240
320
m/z
Figure 2.
Analysis of a standard solution containing 2.5 pg/L CAP and 5 pg/L CAP-d5 (IS) on the LC/MSD Trap together with the corresponding MS/MS spectra and the MS/MS spectrum resulting from an analysis of a standard solution containing 0.2 pg/L CAP.
Method Performance Standard solutions of CAP containing 5 pg/L of CAP-d5 were injected six consecutive times to test repeatability of injection on the mass selective detector (MSD) quadrupole instrument. This was done at two concentration levels. Each time, the response of CAP relative to CAP-d5 was recorded. For a solution containing 0.5 pg/L CAP the relative standard deviations (RSDs) on the relative response were 5.05%. This 0.5-pg/L level would correspond to a sample containing approximately 0.1 ppb CAP with the five-fold concentration step. When a solution containing 5 pg/L CAP was analyzed, RSDs on the relative response were 1.28% for the quadrupole.
A calibration line was constructed by injecting standard solutions of CAP with a concentration of 0 to 25 pg/L with 5 pg/L of the IS added to each solution. One injection was performed per concentration. The quadrupole showed a linear response for CAP in this concentration range. Calibration curves and correlation coefficients are shown in Figure 3. The LOD with this method was determined to be ca. 0.2 pg/L in a standard solution for both mass spectrometers. With the 100-L injection used, this corresponds with 20 pg on-column.
8.E+05
CAP (without IS)
R2 = 0.9994
CAP/CAP d5 (with IS) R2 = 1
4 6.E+05 3 4.E+05 2 2.E+05
0.E+00 0 5 10 15 20 25 30 Concentration CAP (ppb)
0 0 5 10 15 20 25 30 Concentration CAP (ppb)
Figure 3.
Calibration graphs for standard solutions of CAP on the quadrupole with and without CAP-d5 (IS).
Extraction Recovery and Repeatability of Extraction The extraction procedure was evaluated on repeatability and linearity with the quadrupole instrument. Blank honey was spiked with 1 ppb CAP and 1 ppb CAP-d5. The extraction procedure was carried out six times and the recovery was calculated. The recovery for CAP varied from 85.31% to 94.94% and the mean recovery was 90.60%. The RSD on the recovery was 4.34% for CAP and 3.39% when the IS was taken into account. An analysis of blank honey spiked only with the IS is shown in Figure 4 run on both instruments. With the quadrupole, LC/MSD matrix interferences are present but chromatographically separated from the CAP signal. The ion trap results show that no matrix interference is present in the isolation window from m/z 318 to 328. The data suggest that other endogenous compounds in honey produce fragments at the same m/z as CAP. This supports an even lower detection limit for this matrix if a screening analysis were conducted with a lower fragmentor voltage monitoring only the m/z 321.
60000
TIC
30000
15000
EIC 257, 321, 323 m/z (CAP)
QUADRUPOLE
7500
12000
EIC 262, 326, 328 m/z (CAP-d5)

6000
0 4 5 Time (min) 15000 6 7
EIC 176, 194, 249, 257 m/z (CAP)
0 15000
TRAP
EIC 180, 199, 254, 262 m/z (CAP-d5)
0 4 5 Time (min) 6 7
Figure 4.
Analysis of a blank honey sample containing 1 ppb CAP-d5.
A calibration curve was constructed with blank honey samples spiked with 0, 0.1, 0.2, 0.5, 1.0, and 2.0 ppb CAP. The samples also contained 1 ppb of the IS. The correlation coefficients were 0.9997 and 0.9998 without and with correction with the IS, respectively. The slope for the calibration curve constructed with these extracts for CAP with correction with the IS was 0.1822. This is in good agreement with the slope obtained with the standard solutions, which is 0.1758 (see Figure 3).
Spectra on the trap were similar for standard solutions and real samples. An example of an MS/MS spectrum of an extract of a honey sample spiked with 0.5 ppb CAP and 1 ppb CAP-d5 is shown in Figure 5. Since the analyte and the IS coelute, a mixed spectrum is obtained. This could be avoided by using a smaller isolation width and the multiple reaction monitoring (MRM) function of the ion trap. Note that the chlorine isotope for Cl35Cl37 is not observed for the deuterated internal standard because its precursor ion is at the edge of the isolation width and thus not trapped.
262
257 194 176 180

160
249 199
240
254
320 m/z
Figure 5.
Ion trap MS/MS spectrum from analysis of a honey sample spiked with 0.5 ppb CAP and 1 ppb CAP-d5.
Analysis of Honey The extraction procedure and LC/MS methods were applied to the analysis of honey samples that were known to contain CAP. Sample results obtained with the quadrupole and trap MSD were compared (Figure 6).
80000 40000 0 14000
TIC
QUADRUPOLE
EIC 257, 321, 323 m/z (CAP)

8000 2000 15000
EIC 262, 326, 328 m/z (CAP-d5)

7500 0 4 5
Time (min)
EIC 176, 194, 249, 257 m/z (CAP)

15000
TRAP
0 15000
EIC 180, 199, 254, 262 m/z (CAP-d5)
0 4 5
Time (min)
Figure 6.
Analysis of a honey sample containing 0.5 ppb CAP and 1 ppb CAP-d5.
The LOD for the honey samples varies between detectors. For the quadrupole, it is found to be 0.5 pg/L in the analytical solution. This corresponds with 50 pg on-column. Taking into account the sample preparation with a five-fold concentration, samples containing 0.1 ppb CAP can be detected. It is obvious that the sample matrix interferes with the sensitivity (Figures 4 and 6). Due to the increased selectivity using MS/MS in the trap, the LOD with this MS is similar for honey samples as for the standard solutions and is ca. 0.2 pg/L in the analytical solution. This is equivalent to 0.04 ppb CAP in the sample because of the five-fold concentration step.
Analysis of Shrimp The same sample preparation method was applied to the analysis of shrimp. The total volume of shrimp and water added was about 40 mL. Taking 20 mL of the 10 g shrimp aliquot for the Isolute sample preparation and reconstituting the dried extract in 1 mL produced a five-fold concentration as with the honey. This sample preparation shows less matrix interference with the analysis compared to honey samples. An example of an analysis of shrimp is shown in Figure 7. Due to the reduced matrix effect, the LOD with the quadrupole is lowered to nearly the same level as for the trap (0.05 ppb in the sample with the five-fold concentration). A concentration of 0.35 ppb was recovered in the shrimp sample by both the quadrupole and the trap MSD. Extraction recovery was approximately 85%.
30000
TIC
15000
7000
QUADRUPOLE
EIC 257, 321, 323 m/z (CAP)

4000
16000
EIC 262, 326, 328 m/z (CAP-d5)

8000
Time (min)
10000
EIC 176, 194, 249, 257 m/z (CAP)
TRAP
0 10000
EIC 180, 199, 254, 262 m/z (CAP-d5)
0 4 5
Time (min)
Figure 7.
Analysis of a shrimp containing 0.35 ppb CAP and 1 ppb CAP-d5.
Conclusion
Honey and shrimp samples were successfully analyzed for CAP with both the quadrupole and trap MSD. A simple liquid-liquid extraction procedure using ISOLUTE HM-N cartridges was found to perform excellently in view of recovery and repeatability. The LC method used a standard 4.6-mm id column and produced the required sensitivity on both instruments. The LC/MSD quadrupole instrument produced excellent linearity and demonstrated its quantitative ability. The LC/MSD Trap showed the needed sensitivity with excellent full scan capability below the regulated limit in both sample matrices. The use of a broad isolation window for full scan spectra using the ion trap produced more transition ions than required for confirmation.
References
1 S. Turnipseed, et al. (2002) Confirmation of Multiple Phenicol Residues in Honey by Electrospray LC/MS, Laboratory Information Bulletin (4281) U.S. Food and Drug Administration. 2 A. Pfenning, et al. (2002) Confirmation of Multiple Phenicol Residues in Shrimp by Electrospray LC/MS, Laboratory Information Bulletin (4284) Food and Drug Administration. 3 B. K. Neuhaus, et al. (2002) LC/MS/MS Analysis of Chloramphenicol in Shrimp, Laboratory Information Bulletin (4290) Food & Drug Administration. 4 D. Byrne, (2002) Performance Criteria, other Requirements and Procedures for Analytical Methods. Official Journal of European Communities L221, 1417.

For more information on our products and services visit our web site at www.agilent.com/chem
Determination of Chloramphenicol in Fish Meat by Liquid Chromatograph-Atmospheric Pressure Photo Ionization-Mass Spectrometry (LC-APPI-MS) Application
Foods, Environmental
Author
Masahiko Takino and Shigeki Daishima Yokogawa Analytical Systems Inc 2-11-13, Nakacho, Musashino Tokyo, 180-8453 Japan
Abstract
A liquid chromatography-atmospheric pressure photoionization-mass spectrometry method was developed for the determination of chloramphenicol antibiotics in fish meats. For the optimization of APPI, several ion source parameters were examined. Using the optimized parameters, simple mass spectra and a strong signal cor responding to [M-H] was observed. The samples were extracted with ethylacetate and evaporated to dryness followed by a clean-up step using liquid-liquid distribution by acetonitrile and n-hexane. Mean recoveries of chloramphenicol from young yellowtail meat and flatfish meat spiked at 0.12 ng/g were 89.3%102.5% and 87.4%94.8%, respectively. The limit of detection (signal-to-noise = 3) of the young yellowtail meat and the flatfish meat were 0.27 and 0.10 ng/g.
therapy can result in a well-understood and irreversible type bone marrow depression called aplasia or hypolasia. This, in turn, can lead to aplastic anemia and although uncommon, it is often fatal. Because of these health concerns, a joint Food and Agriculture Organization/World Health Organization (FAO/WHO) Expert Committee on Food Additives has proclaimed that CAP residues in the human food supply are unacceptable [1]. The use of CAP in food products has been banned in EU and U.S.A. However, CAPs broad-spectrum activity, ready availability, and low cost attract its use by some third world countries. Admittedly, whenever CAP is accessible, indiscriminate and illegal use potentially exists. In fact, the presence of CAP has been detected in shrimp imported from China and Vietnam that was intended for human consumption. Liquid chromatography/mass spectrometry (LC/MS) methods are very useful in analyzing CAP in food because of the high selectivity and sensitivity of MS detection [2-7]. Atmospheric pressure ionization (API) interfaces, represented by atmospheric pressure chemical ionization (APCI) and electrospray ionization (ESI), are commonly used in LC/MS. Atmospheric pressure photoionization (APPI) is a new ionization technique for LC/MS [8, 9]. The APPI source is based on a high-fluence gas discharge lamp that generates vacuum-ultraviolet (VUV) photons of 10 and 10.6 eV energy. The energy of this discharge lamp is normally greater
Introduction
Chloramphenicol (CAP) is a broad-spectrum antibiotic, that exhibits activity against a variety of aerobic and anaerobic microorganisms. Its action works through interference with or inhibition of protein synthesis. However, weeks or months of CAP
than a first ionization potential (IP) of an analyte because many organic compounds have IPs in the range of 710 eV. On the other hand, the IPs of the most common LC solvents, which are used as a mobile phase, have higher values (water, IP = 12.6 eV; methanol, IP = 10.8 eV; acetonitrile, IP = 12.2 eV). This provides ionization of many analytes with lower IPs without interference from the mobile phase. To our knowledge, APPI has not yet been applied to residual analysis in food. This application note describes how parameters affect the ionization efficiency of APPI for the analysis of CAP. In addition, the suitability of LC/MS and liquid-liquid extraction using the APPI technique is evaluated for the determination of CAP in fish meat.
step was repeated with another 1 mL of n-hexane. Finally, the acetonitrile phase was evaporated to dryness under a stream of dry nitrogen using a heating block at 50 C, redissolved in 5 mL of a 10% acetonitrile in 10 mM ammonium acetate water solution, and filtered through a 0.22 m nylon centrifuge filter. The samples were spiked with 0.1100 ng/mL of CAP after the homogenation step to generate a calibration by LC/APPI-MS selected ion monitoring (SIM). LC/MS An Agilent 1100 series LC, consisting of a vacuum solvent degassing unit, a binary high-pressure gradient pump, a standard automatic sample injector, and a column thermostat, was used for the separation. An 1100 series diode array detector (DAD) was connected in line with an 1100 MSD for detection and confirmation. See Table 1. The separation was performed on a 150 3 mm id column packed with 5 m Zorbax Eclipse XDB C18 (Agilent Technologies, Palo Alto, USA). A 15-min linear solvent gradient was used for elution with the mobile phase. Quantitative analysis was carried out using SIM of m/z 321 with a dwell time of 500 msec. The following six parameters were optimized using
Table 1. LC: Column: Solvent A: Solvent B: Dopant: Gradient: Column temp: Sample volume: Flow rate: MS: Ionization: Scan range: SIM ion: Drying gas: Nebulizer gas: Fragmentor: Capillary: Vaporizer temp: Instrument Parameters 1100 series LC Zorbax Eclipse XDB C18 (150 mm 3 mm, 5 m) Water with 10 mM ammonium acetate Methanol Acetone at 0.05 mL/min 90/10 A/B 15 min to 70/30 A/B 40 C 20 L 0.5 mL/min 1100 MSD, SL APPI (Negative) m/z 100400 for optimization m/z 321; (M-H)
_
Experimental
Chemicals and Solvents CAP was purchased from Sigma-Aldrich Japan (Tokyo, Japan). The purity of this compound was greater than 99%. Stock solutions at 1 mg/mL were prepared in methanol, stored in the dark at 4 C, and diluted to the desired concentrations prior to use. Ammonium acetate, pesticide-grade ethyl acetate, anhydrous sodium sulfate, acetonitrile, HPLC-grade methanol and n-hexane were obtained from Wako Chemical (Osaka, Japan). Water was purified with a Milli-Q system (Millipore, Tokyo, Japan). A nylon-type 0.22 m centrifuge filter was obtained from Toyo Soda (Tokyo, Japan). Sample Preparation The samples analyzed (young yellowtail and flatfish) were obtained from a local market. To a centrifuge tube, 5 g fish meat and 5 g anhydrous sodium sulfate were weighed and 10 mL ethyl acetate was added. The mixture was homogenized for 20 s with an Ultra-Turrax TP 18/10 (Janke & Kunkel KG, Staufen, Germany). After centrifugation for 5 min at 6000 rpm, the supernatant was removed and transferred to a round flask. The extraction step was repeated twice, each with 10 mL ethyl acetate. The combined ethyl acetate extract was then evaporated in a rotary evaporator at 40 C under vacuum. One mL acetonitrile and 1 mL n-hexane was added to the residue, transferred into a graduated glass stopper reagent bottle, and shaken. The n-hexane phase was discarded. The
Nitrogen, 7 L/min at 350 C Nitrogen, 50 psi 120 V 3500 V 350 C
the analytical column with CAP at 100 ng/mL: the voltages for in-source-fragmentation (the fragmentor voltage), the capillary voltage (Vcap), the drying gas flow rate, the nebulizer pressure, the mobile phase composition, and the mobile phase flow rate. The ion lens voltages in the MS were automatically optimized using a Calibrant Delivery System and the AutoTune program. Negative ion mass spectra were acquired over the scan range m/z 100400 using a step size of 0.1 amu and a scan rate of 2 s per scan for the optimization of fragmentor voltage.
was found that modification of drying gas flow rate and nebulizer gas pressure did not drastically improve the sensitivity of CAP. In addition, the fragmentor voltage was included in optimization because of its compound dependence and its significant effect on the mass spectral response. Effect of Capillary Voltage The capillary voltage is applied to the inlet of the capillary and influences the transmission efficiency of the ions through the capillary sampling orifice. To establish the optimum capillary voltage, this parameter was varied from 1000 to 4000 V. As shown in Figure 1, 1500 V was found optimum. A tremendous effect of this parameter on the intensity of CAP was observed in the case where acetone was not used as the dopant. On the other hand, when acetone was introduced into the APPI source as the dopant, the maximum intensity of the ion was found at 3500 V. The intensity found at 3500 V with the dopant was higher than the maximum intensity without the dopant. Based on the above results, the capillary voltage was set at 3500 V with acetone.

Optimization of the APPI Parameters To optimize the APPI conditions, parameters that influence the ionization efficiency were investigated. The drying gas flow, the nebulizer gas pressure, the vaporizer temperature, the capillary voltage, and the mobile phase composition were evaluated under the chromatographic conditions mentioned in the Experimental section by SIM mode using the m/z 321 ion as the target ion. It
3000000
2500000
2000000 Peak intensity
1500000
Without acetone as the dopant With acetone as the dopant

1000000
500000
0 1000
1500
2000
2500 Capillary voltage [V]
3000
3500
4000
Figure 1.
The effect of the capillary voltage on the peak intensity of CAP concentration : 1 ng/mL. For the other conditions, see Experimental section.
Effect of Vaporizer Temperature In APPI, the vaporizer temperature plays a key role for the complete evaporation of CAP because ionization occurs in the vapor state like APCI. Thus, in the case of using linear gradient elution, this temperature must be kept sufficiently high so that the change of mobile phase composition does not influence the ion intensity of CAP. Under high temperature, however, the risk of thermal degradation occurs. In this study, the vaporizer temperature was modified between 250 and 450 C to optimize the intensity and the S/N ratio. The highest temperature for a maximum intensity and S/N ratio of CAP was observed at 350 C. The intensity of CAP decreased as the vaporizer temperature was increased over 400 C. In addition, intense fragmentation was observed in the mass spectrum at 400 C. Therefore, the decrease in intensity above 400 C seems to be a result of the thermal degradation. Based on the above results, the vaporizer temperature was set at 350 C.
Optimization of Fragmentor Voltage The fragmentor voltage is applied to the exit of the capillary and affects the transmission and fragmentation of sample ions between the exit of the capillary and the skimmer at relatively high pressure (3 torr). In general, the higher the fragmentor voltage (which helps the transfer of ions), the more fragmentation will occur. To establish the optimum fragmentor voltage for the analysis of CAP, the intensity of this compound versus the fragmentor voltage was studied in the range from 80 to 200 V. As shown in Figure 2, the optimum fragmentor voltage was found at 120 V, whereas at higher values a significant intensity reduction was observed. Further, the best S/N ratio was also observed at 120 V. The mass spectra of CAP at optimal and higher fragmentor voltages are shown in Figure 3. The deprotonated molecule (M-H) was the predominant ion at 120 V, and this included isotopic ions (m/z 321, Cl35 Cl35; m/z 323, Cl35 Cl37; m/z 325, Cl37 Cl37) because CAP includes two
2500000
2000000
1500000 Peak intensity 1000000 500000
0 80 100 120 140 Fragmentor voltage [V] 160 180 200
Figure. 2. The effect of the fragmentor voltage on the peak intensity of CAP concentration : 1 ng/mL. For the other conditions, see Experimental section.
chlorines. A higher fragmentor voltage (180 V) generated structurally relevant fragment ions. The m/z 152 fragment ion gives the greatest intensity and might be produced by the cleavage of the carbon-carbon bond on the alkyl branch as shown in Figure 3. Other fragment ions are observed at m/z 121 and 257. The m/z 121 may be the nitrophenyl fragment. The m/z 257 fragment might be explained by a charge migration hydrogen shift with a concerted loss of HCl and CO. These observed fragment ions in the APPI source corresponded with the fragment ions in an ESI source and an APCI source. Based on the above results, the fragmentor voltage was set to 120 V.
Optimization of the Chromatographic Conditions The separation of CAP from sample matrix peaks was optimized using acetonitrile, methanol, water, and ammonium acetate. The combination of methanol and ammonium acetate was found optimum for the separation of CAP. When methanol was replaced with acetonitrile, a significant signal intensity and S/N decrease was observed. This result indicates that methanol may be a source of electrons for the hydrogen abstraction from CAP. Therefore, methanol and 10 mM ammonium acetate was used as the mobile phase in this study. The flow rate was set at 0.5 mL/min considering the size of the used column.
321
600000
(A). Low fragmentor voltage (120 V)
400000
200000
100000 100 150 200 250 300 350 400
(B). High fragmentor voltage (180 V)

250000 152 200000
321
121 O2N
H CH OH
NH
CO
CHCl2
150000
CHCH2OH
100000
152
257 121
50000
0 100 150 200 250 300 350 400
Figure 3.
The mass spectra of CAP at two different fragmentor voltages.
Linearity, Detection Limit and Precision of LC/APPI-MS System The analytical performance characteristics of the optimized LC/APPI-MS were first determined on standard solutions of CAP in pure solvent. See Figure 4. In order to achieve optimum sensitivity, all experiments were carried out under SIM mode using the mass corresponding to the [M-H] ions for CAP. To test the linearity of the calibration curves, various concentrations of CAP ranging from 0.1 to 100 ng/mL were analyzed. The calibration curves of APPI showed good linearity with correlation coefficients (r2) = 0.9998. The repeatability of APPI for a standard solution was calculated on the basis of five replicates at 0.5 ng/mL in the same day. The limit of detection (LOD) was calculated by using a S/N ratio of 3 at 0.1 ng/mL. The SIM chromatogram of CAP with APPI is shown in Figure 4 (the S/N ratio of this chromatogram was 4.2); LOD and RSD were 0.07 ng/mL and 2.1%.
900
700
500
300
10 Retention time [min]
15
20
Figure 4.
SIM chromatogram of CAP in pure solvent at 0.1 ng/mL with APPI.
APPI Method Evaluation To evaluate recoveries, the proposed method was applied to the analysis of spiked CAP-free samples of young yellowtail and flatfish meat. Eighteen samples of two different fish were each spiked with CAP and each sample was spiked at three levels. The spiking levels ranged from 0.1 to 2 ng/g. Typical chromatograms from the fish meat extracts spiked at 1 ng/g and 0.1 ng/g are shown in Figure 5.
20000 15000 10000 5000 0 16000 12000 8000
(A)
10
15
20
(B)
0 12000 8000 4000
10
15
20
(C)
0 10000
10
15
20
(D)
6000
2000 0 5 10 Retention time [min] 15 20
Figure 5.
SIM chromatograms of A) Young yellowtail meat, B) Spiked young yellowtail meat at 1 ng/g CAP, C) a flatfish meat, and D) a spiked flatfish meat at 0.1 ng/g CAP.
Data from 18 spiked samples led to recoveries and RSD are summarized in Table 2.
Table 2
Recovery of CAP for Spiked Fish Meat Recovery [RSD (%)]* Young yellowtail Flatfish 89.3 5.1 102.5 4.9 96.1 4.3 87.4 6.1 94.8 6.7 91.8 4.9
Spiking levels (ng/g) 0.1 0.5 2.0
*Three spiked samples at the same amount were analyzed.
Mean recoveries ranged from 87.4% to 102.5% with RSD of 4.3% to 6.7%. The LODs of CAP in fish meats were determined by the signal corresponding to three times the background noise on SIM chromatogram of spiked sample at 1 ng/g and 0.1 ng/g and shown in Table 3. The intraday precision (repeatability) was estimated by injecting the same spiked fish meat extract at 0.1 ng/g five times during a working day. The interday precision (reproducibility) was evaluated by analyzing the same sample over 5 working days. The repeatability and reproducibility for CAP in fish meats were 4.8%, 9.4% and 2.1%, 7.3%, respectively. These results indicate that this LC/APPI-MS method is suitable for the analysis of residues of CAP in fish meats.
Table 3.
LODs, Repeatability, and Reproducibility of CAP in Standard Solution Using APPI LODs* Repeatability** Reproducibility*** Fish meats (ng/g) (RSD, %) (RSD, %) Young yellowtail 0.27 4.8 9.4 Flatfish 0.10 2.1 7.3
*Detection limit is LOD defined as S/N = 3 at 0.1 ng/mL. **Repeatability was calculated on the basis of five replicates at 0.5 ng/mL within 1 day. ***Reproducibility was calculated by analyzing one fish meat spiked at 0.1 ng mL1 per day during 5 days.
Conclusion
APPI is an ideal ionization technique because of high sensitivity and high selectivity for the determination of CAP in fish meats. An important advantage of using APPI for CAP content of fish meats is that sample matrix did not significantly affect ion intensity of CAP. The data presented here demonstrate that this method is convenient for routine analysis of CAP residues in fish meats at trace levels, as excellent recoveries and precision for different samples were obtained.
References
1. A. H. Allen J. AOAC Int. 1985, 68, 990. 2. T. L. Li; Y.J. Chung-Wang; Y. C. Shih J. Food Science 2001, 67, 21. 3. B. Roudaut J. Liq. Chrom. & Rel. Technol. 1994, 19, 1097. 4. C. N. Kenyon; A. Melera; F. Mrmi J. Anal. Toxicol. 1981, 5, 216. 5. V. Hormazabal; M. Yndestad J. Liq. Chrom. & Rel. Technol. 2001, 24, 2477. 6. K. Richard; V. Kruft; H. Sommer LaborParaxis 2000, 24, 91. 7. D. G. Kennedy; R. J. McCracken; A. Cannavan; S. A. Hewitt J. Chromatogr. A. 1998, 812, 77. 8. D. B. Robb; T. R. Covey; A. P. Bruins Anal. Chem. 2000, 72, 3653. 9. J. A. Syage; M. D. Evans; K. A. Hanold Amer. Lab. 2000, 32, 24.

Determination of the Metabolites of Nitrofuran Antibacterial Drugs in Chicken Tissue by Liquid Chromatograph-Electrospray Ionization-Mass Spectrometry (LC-ESI-MS) Application
Food, Environmental
Author
Masahiko Takino Yokogawa Analytical Systems Inc 2-11-13, Nakacho, Musashino, Tokyo, 180-8453 Japan
additives to prevent bacterial enteritis by Escherichia coli and Salmonella in cattle, fish, swine, and poultry. The occurrence of furazolidone residue in edible tissue is a major human health concern. Effective June 1995, these drugs were banned from use in food animal production in the European Union (EU) because of concerns about their carcinogenicity and mutagenicity (Commission Regulation 1442/95). Nitrofuran antibacterial drugs are characterized by their rapid metabolism, with in vivo half-lives of less than a few hours. Therefore, the detection of parent drugs in animal tissue is not practical. Studies using radioactive-labeled furazolidone have shown that protein-bound metabolites are formed in tissues [1-3]. The tissue-bound metabolites are detectable for several weeks after administration. Hence, the analysis of nitrofuran drugs is based on the detection of the tissue-bound metabolites of the parent drugs. These tissue-bound metabolites are very small molecules which are not UV absorbing, and they elute too quickly out of a column. To induce UV absorption in the molecule and to be reasonably retained on a column, they are derivatized. It is possible to release these metabolites from the proteins under moderately acidic conditions and derivatize the metabolites with 2-nitrobenzaldehyde (2-NBA) to produce 2-NBA-derivatives for liquid chromatography (LC), UV detection, and mass spectrometry (MS) confirmation. The goal of this study is to develop a routine analytical method to simultaneously detect the target nitrofuran metabolites. Because no maximum residue limit (MRL) has
Abstract
A liquid chromatography-electrospray ionization-mass spectrometry (LC-ESI-MS) method was developed for the simultaneous determination of the metabolites of four nitrofuran antibacterial drugs in chicken tissues: furazolidone, furaltadone, nitrofurazone, and nitrofurantoin. Sample clean-up and analyte enrichment were performed by liquid-liquid extraction with ethyl acetate followed by solvent washing, hydrolysis of the protein-bound drug metabolites, and derivatization with 2-nitrobenzaldehyde (2-NBA). ESI parameters were optimized, and the chromatographic separation of all metabolites was examined. Each metabolite produced a simple mass spectrum containing a strong signal corresponding to [M+H]+. Metabolite calibration curves, in the 0.25 to 1 ng/mL range, exhibited correlation coefficients greater than 0.999. The limit of detection (LOD) for each analyte ranged from 0.02 to 0.06 ng/mL.
Introduction
The four drugs shown in Figure 1, furazolidone, furaltadone, nitrofurazone, and nitrofurantoin, belong to the group of nitrofuran antibacterial drugs. These drugs have been widely used as feed
Drugs
O N N2O O N O H2N
Metabolites
O N O
Furazolidone
3-amino-2-oxazolidinone (AOZ)
O N N2O O N O H2N N
O O
N
Furaltadone
5-morpholinomethyl-3-amino-oxazolidinone (AMOZ)
O N N2O O N H
Nitrofurazone
O H2N NH2 N H NH2
Semicarbazide (SEM)
O N N2O O N NH H2N N
O NH
O
Nitrofurantoin
O
1-aminohydantoin (AHD)
Figure 1.
Structure of the nitrofuran antibacterial drugs and their metabolites .
been set by any regulatory agency, the goal of the analytical method was to estimate the lowest possible detection limit.
Experimental
Chemicals and Solvents Three metabolites: 3-amino-2-oxazolidinone (AOZ), semicarbazide (SEM), and 1-aminohydantoin (AHD) were purchased from Sigma Aldrich Japan (Tokyo, Japan). The purity of these compounds was greater than 99%. The 2-NBA derivatives of these metabolites were prepared by the Livestock
Department in Thailand (Palm Thani, Thailand) using the procedure described by Leitner [4]. Stock solutions of these three 2-NBA derivatives were prepared in methanol at 1000 ng/mL and stored in the dark at 4 C. The stock solution was diluted to the desired concentration just prior to its use for the optimization of ESI parameters. Acetonitrile, ethyl acetate, formic acid, and dimethyl sulfoxide (DMSO) were supplied by Wako Chemical (Osaka, Japan). Hydrochloric acid and 2-NBA were purchased from Tokyo Kasei Kogyo (Tokyo, Japan). Water was purified with a Milli-Q system (Millipore, Tokyo, Japan).
Sample Preparation Sample preparation procedures included solvent wash and acid extraction by homogenization and derivatization with 2-NBA. Chicken muscle and liver were prepared by the Livestock Department in Thailand. Calibration curves for the four nitrofuran metabolites (from the Livestock Department) were constructed in the range 0.25 to 1.0 ng/mL. The derivatization and sample preparation procedures used by the Livestock Department are the following: 1. The four metabolite solutions in water, 12.5, 25.0, 37.5, and 50 L at 100 ng/mL, were transferred to separate 40 mL glass vials with screw caps. 2. A solution of 10 mL HCl (125 mM in water) and 200 L 2-NBA (50 mM in DMSO) were added to each vial. 3. The reaction mixtures were kept in a water bath at 37 C for 16 hours. 4. The solutions were cooled to room temperature. 5. The pH was adjusted to about 7.4 by adding 0.1 M aqueous KHPO4 or 0.8 M aqueous NaOH. 6. A 5-mL measure of ethyl acetate was added to each reaction mixture, and shaken for 2 min. 7. Each ethyl acetate phase was transferred to a separate glass vial and evaporated under a stream of nitrogen. 8. Finally, each residue was reconstituted in 5 mL of 1:1 methanol:water (V/V). The calibration curve was based on the metabolite concentration in clean solvent and derivatization using 2-NBA. Previous studies done by the Livestock Department showed that recovery of all metabolites from chicken extracts was above 80%. Therefore, the amounts of metabolite in chicken extract can be calculated by comparing the responses of 2-NBA derivatives from the samples against the calibration curve.
Instrument and Experimental Conditions An Agilent 1100 series LC, with a solvent degassing unit, a binary high-pressure gradient pump, an automatic sample injector, and a column thermostat, was used for separation. An 1100 series diode array detector (DAD) was connected in line with an 1100 MSD for detection and confirmation. The column and MS conditions are described in Table 1.
Table 1. LC: Column: Solvent A: Solvent B: Gradient: Column temp Sample volume Flow rate: Instrument Parameters Agilent 1100 series Inertsil ODS3, 150 mm 2.1 mm, 5 m (GL Science, Tokyo, Japan) Acetonitrile Aqueous 0.5% formic acid 20/80 A/B to 70/30 A/B in 20 min 20 C 30 L 200 L/min
MS: Ionization: Scan range: SIM ion: Drying gas: Nebulizer gas: Fragmentor: Vcap
Agilent 1100 MSD, SL ESI (Positive) 100500 m/z for optimization Base peak for quantitation Nitrogen, 10 L/min at 350 C Nitrogen, 50 psi 120 or 140 V 2000 V
Quantitative analysis was carried out using selective ion monitoring (SIM) of the base peak ions according to the program shown in Table 2. To confirm the presence of the target analytes in chicken extract, the sodium adduct ions (qualifier ions) of all target analytes were also monitored.
Table 2. Group 1 2 3
SIM Program Time window min 06 612.5 12.514 Analyte(s) 2-NBA-AMOZ 2-NBA-SEM and 2-NBA-AHD 2-NBA-AOZ Target ion 335 209 and 249 236 Qualifier ion 357 231 and 271 258 Dwell time msec 500 250 and 250 500 Fragmentor voltage, V 140 120 and 140 140
System Optimization Positive ion mass spectra were acquired over the scan range m/z 100500 using a step size of 0.1 amu and a scan rate of 2 seconds per scan for the optimization of fragmentor voltage. Ion lens voltages in the MS were automatically optimized using a Calibrant Delivery System and the AutoTune program. Using the analytical column and three 2-NBA derivatives (AOZ, SEM, and AHD) at 100 ng/mL, instrument performance was optimized by adjusting the four major ESI parameters: the capillary voltage, fragmentor voltage, the nebulizer gas pressure, and the drying gas flow rate. However, significant variation in the intensity of analytes was not observed when the drying gas flow rate and nebulizer gas pressure were varied from 4 L/min to 13 L/min and 20 psi to 60 psi, respectively. Capillary and fragmentor voltages applied to the inlet and exit end of the capillary affected the ion
transmission significantly. Fragmentor voltage also affected the fragmentation of sample ions. In general, higher fragmentor voltage helps the transmission of ions through the relatively high-pressure region between the exit of the capillary and the entrance of the skimmer. High fragmentor voltage can cause fragmentation to occur which provides structural information of the ion. For compounds that do not fragment easily, higher fragmentor voltage often results in better ion transmission. Optimal fragmentor voltage is compound dependent. Evaluation of the fragmentor voltages for the three 2-NBA-metabolites was done under the same chromatographic conditions as the analysis. Mass spectra of three 2-NBA-metabolites are shown in Figure 2. Each mass spectrum exhibited [M+H]+ as the base peak. Adducts ions [M+NH4]+ and [M+Na]+ were observed at lower fragmentor voltage (120 V) and some fragment ions (m/z=178 and 192) were observed at higher fragmentor voltage (180 V). Interestingly, the [M+NH4]+ ion was not observed at
Fragmentor = 120 V
800000
Fragmentor = 180 V
209(M+H)+ 192
200
600000
231(M+Na)+
400000 200000 0
400000 200000 100000 100 500000 400000 300000 200000 100000 0 100 150 150
200
250
300
350 400000
100
150
166
231(M+Na)+
250
600000
209(M+H)+
2-NBA-SEM
2-NBA-SEM
300
350
249(M+H)+ 266(M+NH4)+
249(M+H) 178
200 250
271(M+Na)+
2-NBA-AHD
300000 200000 100000 0
2-NBA-AHD
200
250
300
350
100
150
271(M+Na)+
300
350
236(M+H)+
40000
258(M+Na)+
40000 20000
20000
0 100 150 200 250 300 350
0 100 150 200 250 300 350
Figure 2.
Mass spectra of 2-NBA-SEM, 2-NBA-AHD, and 2-NBA-AOZ from two ESI fragmentor voltages.
192
258(M+Na)+
2-NBA-AOZ
253(M+NH4)+
60000
2-NBA-AOZ
236(M+H)+
180 V fragmentor voltage due to its stability. As seen in Figure 3, in order to ensure the best sensitivity, the fragmentor voltage for 2-NBA-SEM was set to 120 V and that of 2-NBA-AHD and 2-NBA-AOZ was set to 140 V for the analysis. Although 2-NBA-AMOZ was not examined, fragmentor voltage of this compound was set to 140 V because of its structural similarity to 2-NBA-AOZ. For the capillary voltage varied between 1500 and 4500 V, the optimal voltage was found to be 2000 V for all three metabolites.
Linearity, Detection Limits, and Precision In order to achieve optimal sensitivity, all quantitation experiments were carried out under SIM conditions, and the [M+H]+ ions were monitored for all 2-NBA-metabolites. To evaluate the linearity of the calibration curves, various metabolite solutions ranging from 0.25 ng/mL to 1 ng/mL were derivatized and then analyzed. As shown in Table 3, the linearity was very good for all 2-NBA-metabolites with correlation coefficients (r2) greater than 0.999.
1000000
800000
700000
600000 Peak intensity
500000
400000
2-NBA-SEM
300000
2-NBA-AOZ 2-NBA-AHD
200000
100000
0 100 120 140 Fragmentor voltage/V 160 180 200
Figure 3.
Effect of fragmentor voltage on peak intensity. Mobile phase, 20% acetonitrile/80% water 0.1% formic acid; Analyte concentration, 100 ng/mL.
The LOD for all 2-NBA-metabolites was estimated by extrapolating to a signal-to-noise ratio (S/N) of 3 using the signal from the standard solution at 0.25 ng/mL. These SIM chromatograms are shown in Figure 4. The LODs of the metabolites were in the range of 0.02 ng/mL to 0.06 ng/mL. These LODs were lower than those of the LC/MS/MS method developed by Leitner [4]. The intraday
instrument precision (repeatability) was determined by injecting aqueous standard solutions containing all of the 2-NBA-metabolites at 0.5 ng/mL five times during a working day. The interday instrument precision (reproducibility) was evaluated by analyzing the same sample three times over 3 working days. The precision for all analytes ranged from 3.1% to 8.2%, as seen in Table 3.
6000 6000 5500
2-NBA-AMOZ
5000
5500
2-NBA-AHD
5000 4500 4500 4000 4 8000
m/z = 357
6 8 10 12 14 60000 4 6 8
m/z = 271
10 12 14
7000
2-NBA-SEM
6000
56000
2-NBA-AOZ
52000 5000
m/z = 258
4000
m/z = 231
4 6 8 10 12 14
48000 4 6 8 10 12 14
Retention time(min)
Retention time(min)
Figure 4.
SIM chromatograms of aqueous 2-NBA nitrofuran metabolites solution at 0.25 ng/mL.
Table 3.
Linearity, LOD, and Instrument Precision of Metabolites in Aqueous Solutions Instrument precision (%RSD) Repeatability** Reproducibility*** 5.0 7.3 4.7 8.1 4.9 7.9 3.1 8.2
Metabolites AMOZ SEM AHD AOZ
r2 0.9999 0.9998 0.9989 0.9997
LOD* (ng/mL) 0.04 0.02 0.06 0.06
*Detection limit is LOD defined as S/N = 3 for standard solution at 0.25 ng/mL **Repeatability was calculated based on five replicates at 0.5 ng/mL within 1 day ***Reproducibility was calculated based on once per day for 3 days at 0.5 ng/mL
Evaluation of Chromatographic Separation Several reverse-phase columns were evaluated for HPLC performance. In terms of minimizing the inherent matrix suppression effects on the ESI process, Inertsil ODS3 column provided the best separation between analytes and the majority of the matrix components with the given mobile phase. Further, the linear solvent gradient gave the best compromise between short analysis time and sufficient matrix and analytes separation. Figure 5 shows individual SIM chromatograms for the four metabolite derivatives in spiked chicken muscle at 0.2 ng/g. No interference peaks were observed for 2-NBA-AMOZ and 2-NBA-AOZ, but it was difficult to separate 2-NBA-SEM and 2-NBA-AHD from the interfering matrix peaks. However, these peaks could still be identified by comparison with the blank sample, and the analyte amounts could then be calculated.
25000 15000 5000 0 30000 2
2-NBA-AMOZ
m/z = 335
4 6 8 10 12 14 16
2-NBA-SEM
20000 10000 0 0 40000 30000 20000 10000 0 0 50000 40000 30000 20000 10000 0 0 2 4 6 8 10 12 14 16 2 4 6 8 10 12 14 16 2 4 6 8 10 12
m/z = 209
14 16
2-NBA-AHD
m/z = 249
2-NBA-AOZ
m/z = 236
Retention time(min)
Figure 5.
SIM chromatograms of a spiked chicken muscle tissue sample containing 0.2 ng/g of each of the four 2-NBA nitrofuran metabolites.
Application of the Method to Chicken Liver Samples It has been reported that AOZ concentrations in liver tissue are several times higher than in muscle tissue [1,2]. This indicates that detection of nitrofuran metabolites in liver would be possible over an even longer period of time. Since the nature of the liver matrix is considered to be different from muscle and more difficult to separate target compounds from the interfering matrix, the developed LC/MS muscle method was also tested for the applicability to liver matrix. Figure 6 shows individual SIM chromatograms of the four metabolite derivatives in spiked chicken liver tissue. AMOZ, SEM, and AOZ derivatives were identified unambiguously and quantified down to 0.2 ppb. However, the AHD derivative overlapped with the matrix component and was difficult to quantify.
35000 25000 15000 5000 0 30000 20000 10000 0 0 25000 20000 15000 10000 5000 0 0 70000 50000 30000 10000 0
2-NBA-AMOZ (0.84 ppb)
m/z = 335
2 4 6 8 10 12 14 16
2-NBA-SEM (1.21 ppb)

m/z = 209
2 4 6 8 10 12 14 16
2-NBA-AHD (0.33 ppb)

m/z = 249
2 4 6 8 10 12 14 16
2-NBA-AOZ (0.72 ppb)
m/z = 236
12 14 16
10
Retention time (min)
Figure 6.
SIM chromatograms of a spiked chicken liver tissue sample containing 0.2 ng/g of each of the four 2-NBA nitrofuran metabolites.
Conclusion
The development of a routine and sensitive LC/MS method allows for the simultaneous detection of four nitrofuran metabolite derivatives. The detection limit of each analyte ranges from 0.05 to 0.2 ng/g in chicken muscle and liver tissues.
References
1. L.A.P. Hoogenboom, M. van Kammen, M.C.J. Berghmans, J.H. Koeman and H.A. Kuiper, Food Chem. Toxicol. 1991; 28: 321. 2. L.A.P. Hoogenboom, M.C.J. Berghmans, T.H.G. Polman, R. Paker and I.C. Shaw, Food Addit. Contam. 1992; 9: 623. 3. D.W. Gottschall and R. Wang, J. Agric. Food. Chem. 1995; 43: 2520. 4. A. Leitner, P. Zollner, W. Lindner, J. Chromatogr. A 2001; 939: 49.
Acknowledgement
The author would like to acknowledge the Livestock Department staff in Thailand for providing the four 2-NBA-metabolite standards and chicken muscle and liver extracts.

HPLC Analysis of Antibacterial Drugs with Penicillin-Like Structure Application
Drug Development
Udo Huber, Adebayo O. Onigbinde
Penicillins can be isolated from the culture medium of certain fungi-producing natural penicillin, such as Penicillium notatumor and P. chrysogenum. Other penicillins can be synthesized semisynthetically or by precursor-indicated biosynthesis. Total synthesis would not be economical. Penicillin inhibits the polymerization of murin, which is responsible for the stability of the bacteria's cell wall. Because many antibacterials are toxic, various countries regulate the level of antibacterial residues in agricultural, veterinary, dairy, and meat-based food products. Figure 1 shows the HPLC separation of four common antibacterial drugs with pencillin-like structure (amoxicillin, ampicillin, penicillin G, and penicillin V) on an SB-C18 reversed phase column.
Highlights
There is excellent resolution of penicillin analogs without ion pairing agent. There is rapid resolution of the penicillin analogs on the SB-C18 column at low pH and buffer concentration. Penicillins are eluted from the column with good and narrow peak shape. Extreme stability of sterically protected SB-C18 bonded phases allows for excellent separation at low pH. The SB-C18 column provides excellent peak shape and selectivity for antibacterial drugs. The HPLC method shows an easy but reliable and precise analysis of the antibacterial drugs. The values for limit of detection (LOD), precision of retention time (RT) and area show the good performance of the HPLC analysis.
1000 Absorbance (mAU) 800 600 400 200 0 0 2 4 6 Time (min) 8 10 12
1 2 3 4 4
Amoxicillin Ampicillin Penicillin G Penicillin V
Figure 1. Separation of four penicillin analogs.
Equipment: Agilent 1100 Series HPLC; UV Detector: Variable wavelength detector, 204 nm, standard cell; Column: Zorbax SB-C18, 3. 5 m, 4.6 75 mm (part number 866953-902); Mobile phase: A = 0.025 M KH2PO4 in water (pH = 3), B = acetonitrile; Injection volume: 5 L; Temp: 40 C; Flow rate: 1.0 mL/min; Gradient: at 5% B to 60% in 10 min; Column wash: 60% B to 5% B in 2 min; Stop time: 12 min; Post time: 5 min
Table 1. HPLC Method Performance of Antibacterial Drugs with Penicillin-Like Structure LOD for S/N = 2 (mg/L)* 1.0 1.0 1.0 1.0 Precision of RT (RSD of 10 runs) (100 mg/L)* 0.32 0.32 0.32 0.25 Precision of area (RSD of 10 runs) (100 mg/L)* 0.54 0.55 0.49 0.48
Compound Ampicillin Amoxicillin Penicillin G Penicillin V

*Injection volume: 5 L
Udo Huber is an application chemist based at Agilent Technologies, Waldbronn, Germany. Adebayo Onigbinde is an applications chemist based at Agilent Technologies, Wilmington, Delaware, USA.

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2002 Printed in the USA September 10, 2002 5988-7629EN
HPLC Separation of Antibacterial Drugs with Tetracycline Structure Application
Drug Development
Udo Huber, Adebayo O. Onigbinde
Tetracyclines occur naturally in some streptomyces species. Besides being used in human and veterinary medicine, they are fed as nutritional antibiotics in pig and poultry farming. Because of their long half-life and resistance, there is a high restriction on their usage in some European countries, such as Germany. Figure 1 shows the HPLC separation of three common tetracycline analogs on a Zorbax SB-C18 reversed phase column. This application demonstrates separation without ion pairing and the use of an alternative mobile phase to TFA in separating antibacterial drugs.
2
120 100 Absorbance (mAU) 80 60 40 20 0 0 2 4 Time (min) 6 8 10
Highlights
The SB-C18 column provides excellent peak shape and selectivity for basic antibacterial drugs. The SB-C18 column shows excellent stability at low pH. The SB-C18 column shows excellent and rapid resolution of antibiotics at low pH and buffer concentration. The HPLC method shows an easy but reliable and precise analysis of the antibacterial drugs. The values for limit of detection (LOD), precision of retention time (RT), and area show the good performance of the HPLC analysis.
1 3
1 Minocycline 2 Tetracycline 3 Doxycycline
Figure 1. Separation of three antibacterial drugs with tetracycline structure.
Equipment: Agilent 1100 Series HPLC; UV Detector: Variable wavelength detector, 350 nm, standard cell; Column: Zorbax SB-C18, 3. 5 m, 4.6 75 mm (part number 866953-902); Mobile phase: A = 0.025 M KH2PO4 in water (pH = 3), B = acetonitrile; Injection volume: 5 L; Temp: 25 C; Flow rate: 1.0 mL/min; Gradient: at 5% B to 60% B in 10 min; Column wash: 60% B to 5% B in 2 min
Table 1. HPLC Method Performance of Antibacterial Drugs with Tetracycline Structure LOD for S/N = 2 (mg/L)* 0.1 0.1 0.1 Precision of RT (RSD of 10 runs) (100 mg/L)* 0.06 0.05 0.04 Precision of area (RSD of 10 runs) (100 mg/L)* 0.14 0.13 0.21
Compound Minocycline Tetracycline Doxycycline

*Injection volume: 5 L
Udo Huber is an application chemist based at Agilent Technologies, Waldbronn, Germany. Adebayo Onigbinde is an applications chemist based at Agilent Technologies, Wilmington, Delaware, USA.

Analysis of Residual Synthetic Antibacterials in Meat by HPLC Application
Food
Hiroki Kumagai, Adebayo Onigbinde
Many domestic cattle receive various antibacterials in their feed for the prevention and control of disease caused by fungi and bacteria. Residues of antibacterials are found in food made from the meat of these animals. Since many antibiotics are toxic, many countries regulate acceptable residue levels of compounds allowable in agricultural and animal products. Many alkyl-C18 columns tail with basic compounds and have a shorter life time at low pH. Purospher column separated basic antibacterials with good resolution, peak shape,and efficiency.
Absorbance [mAU] 25 20 15 10 5 0 0 Absorbance [mAU] 25 20 15 10 5 0 0 5
Highlights
Separation of 10 antibacterials in meat at low pH Excellent and rapid resolution of antibacterials at low sample concentration Elution of antibacterials from the column with good peak shape and narrow peak width Separation of low level amounts of a wide range of pharmaceutical compounds with differing structures in a single analysis by Purospher column
3 8 9 10
1 2 3 4 5 Time (min) SMr PYM TCP SDD FZD 6 7 8 9 10 SMMX DFZ SDMX SQX OXA
at 224 nm
1
10 15
4 5
67
20
25
30
35
at 360 nm
DFZ FZD
10
15
20
25
30
35
Time (min)
Figure 1. Chromatogram of standard solution, 2 g/mL each analyte.

Absorbance [mAU] 4 3 2 1 0 -1 0 Absorbance [mAU] 4 3 2 1 0 -1 0 5 10
at 224 nm SDD
15
20
25
30
35
Time (min)
Analyzed Compounds Sulfamerazine (SMR) Sulfadimidine (SDD) Sulfamonomethoxine (SMMX) Sulfadimethoxine (SDMX Sulfaquinoxaline (SQX) Pyrimethamine (PYM) Thiamphenicol (TPC) Furazolidone (FZD) Difurazone (DFZ) Oxolinic acid (OXA) Sample: Extracts from bovine muscle Sample preparation: According to the official procedure of the Japanese food sanitation law.
at 360 nm
10
15
20
25
30
35
Time (min)
Figure 1. Chromatogram of extract of bovine muscle.
Instrument: Agilent 1100 Series HPLC; Column: 250 mm 4 mm id, RP-18 Purospher, 5 m, Part no. 79925PU-584; Mobile phase: A = 0.7 % Phosphoric acid, B = CH3CN; Gradient: 0.0 min 5% B; 10.0 min 65% B; 40.0 min 65% B; 45.0 min 65% B; Post Time 7.0 min 5% B; Flow rate: 1.0 mL/min; Temperature: 40 C; Injection volume: 20 L; Diode array detector: A338/10 nm, reference wavelength off; B264/8 nm, reference wavelength off; C360/8 nm, reference wavelength off
Hiroki Kumagai is the PHS Support manager based at Yokogawa Analytical Systems Inc., Tokyo, Japan. Adebayo Onigbinde is an application chemist based at Agilent Technologies, Wilmington, Delaware, USA.

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2002 Printed in the USA June 18, 2002 5988-7135EN
Analysis of tetracyclines by HPLC
Abstract Tetracyclines are used worldwide as oral or parenteral medication in the form of additives in animal feed. In food-producing animals, these drugs exhibit a high degree of activity toward a wide range of bacteria.1, 2 Sample preparation After homogenization or mincing and addition of mineral acids to dissociate tetracyclines from proteins, the samples were extracted using liquid/liquid extraction followed by degreasing and/or deproteinization, purification, and concentration.3 Chromatographic conditions The HPLC method presented here for the analysis of meat is based on reversed-phase chromatography and UV-visible diode-array detection. UV spectra were evaluated as an additional identification Conditions tool.
Absorbance [mAU] 1 2 3 4 5 6 7 Oxytetracycline Tetracycline epi-Tetracycline Demeclocycline epi- Demeclocycline Chlortetracycline Doxycycline
6 5 4 3 2 1 0 -1 2 4 6 1
2 4 6
Column: 100 4 mm Hypersil BDS, 3 m Mobile phase: A = water, pH = 2.1 with sulfuric acid B = ACN Gradient: start with 15 % B at 10min 60% B Flow rate: 0.5 ml/min Column compartment: 25 C Detector: UV-DAD detection wavelength 355 nm/20 nm, reference wavelength 600/100 nm
5 10 ng each
Sample preparation
1. 1 g sample was mixed with citric acid (100 mg). 2. add 1 ml nitric acid (30 %) or 0.1 m oxalic acid 3. add 4 ml methanol 5 min in the ultrasonic bath 4. add water up to 10 ml total volume 5. centrifuge 6. inject
1 ng each 8 10 12 14 Time [min]
Figure 1 Analysis of tetracyclines by HPLC
Area Area = 2.59596374*Amt -0.0821516 25 20 15 10 5 0 0 5 Amount [ng/ul] 3 2 Correlation: 0.99996 Rel. Res%(3): 2.461 1
Equipment
Agilent 1100 Series vacuum degasser quaternary pump autosampler thermostatted column compartment diode array detector, Agilent ChemStation + software
HPLC method performance

Limit of detection for UV-DAD 100 ppb Repeatability of RT over 10 runs <0.2 % of areas over 10 runs <2 %
Figure 2 Linearity for oxytetracycline 1-10 ng

Absorbance [mAU] 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 200 250 300 350 400 Wavelength [nm] 1 ng Tetracycline Library Tetracycline
References
1. H. Malisch et al., Determination of residues of chemotherapeutic and antiparasitic drugs in food stuffs of anomaly origin with HPLC and UV-Vis diodearray detection J. Liq. Chromatogr., 1988, 11 (13), 28012827.14. 2. M.H. Thomas, J. Assoc. Off. Anal.; 1989, 72 (4) 564. 3. Farrington et. al., Food Additives and Contaminants, 1991, Vol. 8, No. 1, 55-64.
Figure 3 Analysis of tetracyclines at 100 ppb by HPLC

Contaminants Acrylamides
Gas Chromatography/Mass Spectrometry Approaches to the Analysis of Acrylamide in Foods Application
Food Safety
Author
Bernhard Rothweiler Agilent Technologies Deutschland GmbH Hewlett-Packard Strasse 8 76337 Waldbronn Germany Eberhardt Kuhn Agilent Technologies, Inc. 91 Blue Ravine Road Folsom, CA USA Harry Prest Agilent Technologies, Inc. 5301 Stevens Creek Blvd. Santa Clara, CA USA
Introduction
The discovery announced in April 2002 by scientists at Swedens National Food Administration of acrylamide (2-propenamide) in fried and baked foods at levels many times that allowed in water suggested a much higher exposure than previously estimated [1-3]. Acrylamide (Figure 1), a known neurotoxin, is considered a probable human carcinogen. The World Health Organization considers 0.5 g/L the maximum level for acrylamide in water. However, foods such as french fries, baked potato chips, crisp breads, and other common cooked foods, were found to contain acrylamide between 100 and 1000 g/kg. Acrylamide was not found in the raw foodstuffs and cooking by boiling produced no detectable levels. Recent work has suggested that acrylamide forms via the Maillard reaction, which occurs when amino acids and sugars (for example, asparagine and sucrose) are heated together [4]. The concern over these relatively high concentrations has led to studies of the occurrence of acrylamide in a wide variety of foods.
H H H O H2N
Abstract
Discovery of acrylamide in cooked foods has required an examination of foods for potential exposure. A classic approach employs extracting acrylamide from the food with water and converting the acrylamide to brominated derivatives. These derivatives are described here in terms of their spectra and response in electron impact and positive chemical ionization. Additionally, a more direct and simple approach involving extraction and direct injection and analysis of acrylamide by positive chemical ionization is described. This screening approach is rapid, robust, and provides low detection limits.
Figure 1
Acrylamide (2-propenamide), CH2=CHCONH2, 71.08 g/mole, CAS number 79-06-1.
Acrylamide Analytical Methodologies A wide variety of instrumental approaches have been applied to acrylamide. Recent methods using
liquid chromatography with tandem mass spectrometry (MS-MS) detection have proved useful to approximately 50 g/kg (ppb) or better using the 72 to 55 m/z transition (for example, [5]). This approach has appeared attractive in providing a simple sample preparation strategy. Gas chromatographic methods using MS detection with electron impact (EI) ionization typically suffer from the relatively small size of the molecule and therefore use derivatization. This application note presents alternative gas chromatography/ mass spectrometry (GC/MS) approaches aimed at more rapid screening, as well as the conventional, definitive quantitation via derivatization. These methods are rapid and relatively simple approaches to acrylamide analysis.
71
Rapid Screening via GC/MS-SIM with Positive Chemical Ionization

EI ionization mass spectrum for acrylamide (Figure 2) reveals very low mass ions; 71, 55, 44 m/z. Although there is good intensity at sub-ng levels, the ions are subject to interferences in food samples. The positive chemical ionization (PCI) spectrum achieved with ammonia provides more selective ionization and is of greater utility than EI in food matrices, Figure 3. Ammonia PCI results in two ions; 72 m/z, the protonated molecule, [M+H]+, and 89 m/z due to the adduct, [M+NH4]+. PCI provides good selectivity and sensitivity for acrylamidepicogram amounts can be detected.
44 55
40 35 40 45
51 59 50 55 60 65 65 70 76 75 80 80 85 85 89 90 93 95 98
103 100 105
108
112 110
117 115
Figure 2
The EI ionization spectrum of acrylamide (40120 amu).
89
72
63 55 60 70 78 80 90
100 108 100 114 110 122 120 128 130 138 140 146 152 150 158 165 160 171 170 179 180 189 190 198
Figure 3
The PCI spectrum of acrylamide with ammonia reagent gas (60200 amu).
Figure 4 shows a calibration curve from 100 pg to 10 ng collected under the method cited below in the section on Instrumental Parameters.
Table 1.
GC/MS Instrumental Method Parameters for PCI Screening of Native Acrylamide Agilent p/n 5062-3587 Single-taper with glass wool 220 C Pulsed splitless 30.0 psi 1.20 min 50.0 mL/min 1.20 min 54.7 mL/min Off 260 C 0.20 min 60 C 1.00 min Time 10.00 min 25.17 min Agilent 19091X-136 HP-INNOWax 260 C 60.0 m 250.00 m 0.25 m Helium Constant flow 2.0 mL/min MSD Vacuum 7.00 min PCI Ammonia at 24% (1.2 mL/min) PCI Autotune + 400 V 250 C 150 C High Dwell (ms) 60 60 60 60
Inlet parameters Liner:
Figure 4
PCI-ammonia SIM calibration curve from 100-picograms to 10 ng (R2 = 1.00).
Temperature: Mode: Pulse pressure: Pulse time: Purge flow: Purge time: Total flow: Gas saver: Oven parameters Oven maximum: Oven equilibrium time: Initial temperature: Initial time: Ramp Temperature 12 C/min Run time: 230 C
Screening Sample Preparation The enhanced specificity obtained through PCI can be used for rapid screening using a very simple and rapid sample cleanup. A food sample is homogenized and pulverized, and 0.4-g subsample is transferred to a centrifuge tube. The sample is extracted with 1 mL of methanol:water (9:1 v/v) solution for 10 minutes using an ultrasonic cleaner. Prior to sonication, 1 g of labeled 13 C3-acrylamide is added to the 1-mL solution. After sonication, the sample is centrifuged for about 5 minutes at 8000 rpm. The upper layer is decanted and transferred to a vial for injection and analysis by GC/MS-PCI conditions with selected ion monitoring (SIM). See Table 1 for method parameters for PCI screening of native acrylamide.
Column parameters Capillary column Maximum temperature: Nominal length: Nominal diameter: Nominal film thickness: Carrier: Mode: Outlet and pressure: MSD Parameters Solvent delay Tuning: EM Setting: Source temperature: Quad temperature: SIM Parameters Resolution: Group ions 72.0 75.0 89.1 92.1
Screening Method Results and Discussion Figure 5 shows the extracted ion chromatograms for a sample of white bread. The baseline shows very little disturbance near the acrylamide analyte due to the selective nature of the PCI with ammonia. The extracted concentration is approximately 34 ng/mL or 85 ng acrylamide per gram white bread. Since acrylamide is formed when amino acids and sugars are heated together, it is logical to suspect the possibility of acrylamide formation in the inlet during injection. To test this possibility, the white bread extract was spiked with 100 ng of acrylamide and reanalyzed. The results calculated 135 ng/mL and suggest that either the relatively low temperature and short duration in the liner due to pressure pulsing mitigate acrylamide formation for this sample or acrylamide formed in the inlet is highly reproducible. This may not be the case in all extracts or under all similar conditions.
Ion 89 Ion 92
12.40
12.60
12.80
13.00
13.20
13.40
13.60
13.80
14.00
14.20
14.40
Figure 5.
Extracted ion chromatograms for acrylamide (84 ng/g) in sample of white bread.
GC/MS Approaches to Acrylamide Involving Derivatives

Another approach to extraction from foods uses water, in situ derivatization, and liquid-liquid extraction [6, 7]. In this approach acrylamide in a homogenized sample is extracted with (hot) water, 1 g : 10 mL. A strong brominated agent is added and allowed to react. This reaction converts acrylamide to the 2,3-dibromopropionamide. Excess brominating reagent is removed by addition of sodium thiosulfate and the solution centrifuged and/or filtered. The 2,3-dibromopropionamide is extracted by partitioning into ethyl acetate. An option is to further treat this derivative to form a more stable analyte, the 2-bromopropenamide. The overall chemistry is given in Equation 1. Methacrylamide, CH2=CH(CH3)CONH2, is frequently used as a recovery surrogate so its behavior is also reported here.
H H H O H2N
KBr, HBr, Br2
Instrumental conditions for the dibromopropionamide and bromopropenamide are cited in Tables 2 and 3. All data was collected using 2-L injections.
Table 2. GC/MS Instrumental Method Parameters for Dibromopropionamide (Dibromo-Derivative of Acrylamide) in EI and PCI with Methane and Ammonia Agilent p/n5181-3315 double-taper 250 C Pulsed splitless 30.0 psi 1.20 min 50.0 mL/min 1.20 min 54.7 mL/min On 325 C 0.50 min 50 C 1.00 min Time 2.00 min 8.56 min
Inlet parameters Liner: Temperature: Mode: Pulse pressure: Pulse time: Purge flow: Purge time: Total flow: Gas saver: Oven parameters Oven maximum: Oven equilibrium time: Initial temperature: Initial time: Ramp 45 C/min Run time: Temperature 300 C
Br H H
O H H2N Br H
Br
O H2N H
2,3-dibromopropionamide C3H5Br2NO mol. wt. 230.9 g/mole
2-bromopropenamide C3H4BrNO mol. wt. 150.0 g/mole
Equation 1 Experimental Acrylamide and methacrylamide were obtained as neat standards (Sigma-Aldrich Corp) and dissolved in HPLC grade methanol. Labeled acrylamide, 1,2,3-13C3-acrylamide, was obtained at 1 mg/mL methanol (Cambridge Isotope Laboratories, Andover, MA). The brominating reagent solution was made according to the literature [6] with reagent grade KBr, HBr, and bromine water (VWR, San Francisco,CA). Sodium thiosulfate was obtained as a 1-Normal solution (VWR, San Francisco,CA). Derivatization also followed the literature [6] with addition of 1 mL of brominating reagent to solutions containing acrylamide; over-night derivatization, neutralization by 1-drop 1N sodium thiosulfate and extraction by 1-mL ethyl acetate (pesticide grade, VWR). The dibromo-derivatives were directly injected. The mono-bromo-derivatives were generated by addition of triethylamine.
Column parameters Capillary column Maximum temperature: Nominal length: Nominal diameter: Nominal film thickness: Carrier: Mode: Outlet and pressure: MSD Parameters for EI and PCI Solvent delay EI Parameters EI Tuning: EM Setting: Source temperature: Quad temperature: EI SIM parameters Resolution:
Agilent 122-3832 DB-35 ms 340 C 30 m 250 m 0.25 m Helium Constant flow 1.2 mL/min MSD Vacuum 5.00 min Autotune Autotune + 400 V 230 C 150 C Low
(Continued)
Table 2.
GC/MS Instrumental Method Parameters for Dibromopropionamide (Dibromo-Derivative of Acrylamide) in EI and PCI with Methane and Ammonia (Continued) Dwell (ms) Acrylamide analyte 10 ms 10 ms 10 ms Internal standard 10 ms 10 ms 10 ms Ancillary surrogate 10 ms 10 ms 10 ms 10 ms PCI Autotune PCI Autotune + 400 V 250 C 150 C MFC 20% (1.0 mL/min) Low Dwell (ms) Acrylamide analyte 10 ms 10 ms 10 ms 10 ms Internal standard 10 ms 10 ms Ancillary surrogate 10 ms 10 ms MFC 20% (1.0 mL/min) Low Dwell (ms) Acrylamide analyte 10 ms 10 ms 10 ms Internal standard 10 ms 10 ms 10 ms Ancillary surrogate 10 ms 10 ms 10 ms
Table 3.
GC/MS Instrumental Method Parameters for 2-bromopropenamide (Monobromo-Derivative of Acrylamide) in EI Agilent p/n 5062-3587 Single-taper with glass wool 250 C Pulsed splitless 30.0 psi 1.20 min 50.0 mL/min 1.20 min 54.7 mL/min Off 325 C 0.50 min 50 C 1.00 min Agilent 122-5533 DB-5MS Time 0.00 min 1.50 min 9.66 min 350 C 30.0 m 250 m 1.00 m Helium Constant flow MSD Vacuum 5.00 min Autotune Autotune + 400V 230 C 150 C Low Dwell (ms) Native acrylamide 20 ms 20 ms 20 ms Internal standard 20 ms 20 ms Ancillary surrogate 10 ms 10 ms 10 ms 10 ms
Group ions 2,3-dibromopropionamide 149.9 151.9 106.0 13 C3-2,3-dibromopropionamide 152.9 154.9 109.9 2,3-dibromo-2-methylpropionamide 120.0 122.0 164.0 166.0 PCI Parameters PCI Tuning: EM Setting: Source temperature: Quad temperature: PCI SIM Parameters Methane reagent gas: Resolution: Group ions 2,3-dibromopropionamide 231.9 233.9 149.9 151.9 13 C3-2,3-dibromopropionamide 234.9 236.9 2,3-dibromo-2-methylpropionamide 245.9 247.9 Ammonia reagent gas: Resolution: Group ions 2,3-dibromopropionamide 248.9 246.9 250.9 13 C3-2,3-dibromopropionamide 251.9 249.9 253.9 2,3-dibromo-2-methylpropionamide 262.9 260.9 264.9
Inlet parameters Liner:
Temperature: Mode: Pulse pressure: Pulse time: Purge flow: Purge time: Total flow: Gas saver: Oven parameters Oven maximum: Oven equilibrium time: Initial temperature: Initial time: Column parameters Capillary column Ramp 25 C/min 45 C/min Run time: Maximum temperature: Nominal length: Nominal diameter: Nominal film thickness: Carrier: Mode: 1.2 mL/min Outlet and pressure: MSD Parameters for EI and PCI Solvent delay EI Parameters EI Tuning: EM setting: Source temperature: Quad temperature: EI SIM Parameters Resolution: Group ions 2-bromopropenamide 148.9 150.9 105.9 13 C3-2-bromopropenamide 151.95 153.95 2,3-dibromo-2-methylpropionamide 120.0 122.0 164.0 166.0 Temperature 140 C 300 C

EI Ionization Figures 6 and 7 show the EI mass spectrum of the 2,3-dibromopropionamide and the 2-bromopropenamide, respectively. Note the similar spectra for the two brominated acrylamide derivatives. In EI, the 2,3-dibromopropionamide loses bromide to generate the C3H5ONBr ion that shows the isotopic abundance expected from a monobrominated species. The addition of the triethylamine (base) leads to loss of HBr in solution, generating the monobrominated species C3H4ONBr which contains one less hydrogen than the dibromo-derivative and appears as the molecular ion in EI. The spectra share a common C2H3Br ion that accounts for the fragments at 105.9 and 107.9 m/z. Note that use of the 13C3-acrylamide as an internal standard prohibits use of the 107.9 ion in acrylamide quantitation due this C2H3Br fragment. The dibromoderivative shows greater response than the monobrominated compound and lacks the 149 fragment which is subject to interferences from phthalates which are ubiquitous in solvents and food
packaging. Both compounds demonstrate good linearity over the range of 10 to 500 pg/L in EI-SIM as shown in Figures 8 and 9, but better EI detection and elution at a higher oven temperature makes the dibromo-derivative more attractive than the monobromo-derivative. However, it has long been known that the 2,3-dibromopropionamide breaks down in the injection port to form the 2-bromopropenamide. The fraction converted is a function of the injection port activity hence the use of the double-tapered liner for the dibromopropionamide analysis as opposed to the single-tapered liner with wool for the bromopropenamide. Use of the 13C-labeled surrogate is necessary to correct for the degradation of the dibromo-derivative but the methacrylamide surrogate may correct fairly well for recoveries of the mono-brominated acrylamide. Because of this and citations of its use in the literature, the EI spectrum for the brominated methacrylamide is shown in Figure 10 and ions are presented in the acquisition method tables. As the 2,3-dibromo-2-methylpropionamide, this surrogate elutes just prior to the 2,3-dibromopropioamide and much later than the 2-brompropenamide on the GC programs cited.
100 90 80 70 60 50 40 30 20 10 0
50 60 70 80 90 100 110 120 130 140
150 152
106
55
70
150
160
170
180
190
200
210
220
230
240
Figure 6.
EI ionization spectrum of 2,3-dibromopropionamide.
100 90 80 70 60 50 40 30 20 10 0
50 60
70
106
149 151
53
135
70
80
90
100
110
120
130
140
150
160
170
180
190
200
210
220
230
240
Figure 7.
EI ionization spectrum of 2-bromopropenamide.
Figure 8.
Calibration Curve plot for 2,3-dibromopropionamide from 10 to 500 pg/L (R2 = 0.998).
Figure 9.
Calibration Curve plot for 2-bromopropenamide from 10 to 500 pg/L (R2 = 0.999).
100 90 80 70 60 50 40
120
85
30 20
164
69
10 0
50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240
Figure 10. EI ionization spectrum of the alternative, methacrylamide surrogate, 2,3-dibromo-2-methylpropionamide.
PCI The 2,3-dibromopropionamide spectra in PCI with methane and ammonia reagent gas are shown in Figures 11 and 12. In methane, the highest mass fragment is due to [M+H]+ and in ammonia, [M+NH4]+. Response with methane is higher than with ammonia and would make a good choice in acrylamide quantitation in samples, if background for that particular food are not an issue. Calibration is similar to that in EI between 10 and 500 pg/L for both methane and ammonia (R2 >0.998). It is important that the lower mass fragments that occur in methane and ammonia PCI, m/z 72 and 89, respectively, are not used in SIM quantitation. These intense fragments apparently originate through elimination of Br2 and do not coincide with the cited ions.
232
150 72
60
80
100
120
140
160
180
200
220
240
260
280
Figure 11. The PCI spectrum of 2,3-dibromopropionamide with methane reagent gas (60300 amu).
89
249
72
60
80
100
120
140
160
180
200
220
240
260
280
Figure 12. The PCI spectrum of 2,3-dibromopropionamide with ammonia reagent gas (60300 amu).
10
Similar to the situation in EI, PCI response of the 2,3-dibromopropionamide exceeds that of the 2-bromopropenamide under either reagent gas. Spectra for this analyte using methane and ammonia are presented in Figures 13 and 14. Highest mass fragments for 2-bromopropenamide also are due to [M+H]+ in methane and in ammonia, [M+NH4]+. For completeness, the spectra are also included for the brominated methacrylamide surrogate, Figure 15 and 16.
100 90 80 70 60 50 40 30 20 55 10 0 50 60
72
150
70
80
90
100
110
120
130
140
150
160
170
180
190
200
210
220
230
240
Figure. 13 The PCI spectrum of 2-bromopropenamide with methane reagent gas (50250 amu).
100 90 80 70 60 50 40 30 63 20 10 0 60 70 80 72
89
167 102
90
100
110
120
130
140
150
160
170
180
190
Figure 14. The PCI spectrum of 2-bromopropenamide with ammonia reagent gas (60200 amu).
11
246 86
164
121
60
80
100
120
140
160
180
200
220
240
260
280
Figure 15. The PCI spectrum of 2,3-dibromo-2-methylpropionamide (the methacrylamide derivative) with methane reagent gas (60300 amu).
103
263
86
60
80
100
120
140
160
180
200
220
240
260
280
Figure 16. The PCI spectrum of 2,3-dibromo-2-methylpropionamide with ammonia reagent gas.
12
Conclusions
Since acrylamide was found in a wide range of foodstuffs, a variety of approaches were presented here. The rapid screening approach for native acrylamide using PCI provides a direct and simple method for sensitive detection and quantitation. For approaches using the brominated derivatives, the dibromopropionamide shows superior opportunities for detection and quantitation relative to the 2-bromopropenamide. If, for a particular food product, there are problems in EI, PCI will provide a worthwhile approach for exploration. Methane reagent gas provides about twice the response of ammonia. The degradation of the dibromopropionamide can and must be accounted for by an appropriate labeled internal standard. The methacrylamide surrogate also may be useful for recovery calculations. Data collected on potato chips, and not presented here, suggests this is the case.
3. Hileman, B., Acrylamide found in cooked foods. (2002) Chemical & Engineering News, 80(19): p. 33. 4. Yarnell, A., Acrylamide mystery solved. (2002) Chemical & Engineering News, 80(40): p. 7. 5. Rosen, J. and K. E. Hellenas, Analysis of acrylamide in cooked foods by liquid chromatography tandem mass spectrometry. (200) Analyst (Cambridge, UK), 127(7): p. 880-882. 6. Castle, L., M. J. Campos, and J. Gilbert, Determination of acrylamide monomer in hydroponically grown tomato fruits by capillary gas chromatography/mass spectrometry. (1991) J Sci Food Agric, 54(4): p. 549-555. 7. Castle, L., Determination of acrylamide monomer in mushrooms grown on polyacrylamide gel. (1993) J Agric Food Chem, 41(8): p. 1261-1263.
For More Information References

1. Swedish researchers report acrylamide found in starchy foods, (2002) Chemical & Engineering News. p. 38. 2. Hileman, B., Acrylamide worries experts, (2002) Chemical & Engineering News, 80(27): p. 9. For more information on our products and services, visit our Web site at www.agilent.com/chem.
13
Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2004 Printed in the USA January 15, 2004 5989-0602EN
Contaminants Food Packaging
LC-TOF-MS As a Tool to Support Can Coating/Food Interaction Studies
Application
Food Safety
Authors
M. Driffield, E. L. Bradley, and L. Castle Central Science Laboratory Sand Hutton York, YO41 1LZ UK J. Wagner and B. Wedzicha Proctor Department of Food Science University of Leeds Leeds, LS2 9JT UK J. Zweigenbaum Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808 USA
Food ingredients such as fat or water can cause some coatings to swell, which may enhance any migration, particularly if the food is heat processed in its packaging. Migration can also depend on other factors: contact time and temperature, the type and thickness of the coating, and the molecular mass and size of the migrating species. Studies, in particular migration modeling, of interactions between the can coating and the foodstuff are important in understanding, and eventually reducing, migration of compounds from the can coating into the foodstuffs. In previous applications we have described the analysis of can coatings based on epoxy resins [1] and polyester resins [2] and how the accurate mass information for the parent compounds and fragment ions greatly increased the confidence in the identification of unknown compounds. In this application we describe how liquid chromatography/time-of-flight mass spectrometry (LC/TOF-MS) was used as a tool to support studies on can coating/food interactions
Abstract
This application illustrates how time-of-flight mass spectrometry has been used in the studies of interactions at the can coating/food interface of internally coated metal cans intended for use in the food industry. Previously unconfirmed migrants were confidently identified using accurate mass information provided.
Experimental
Coated Panels The epoxyphenolic (EPH) lacquer was applied to metal panels and cured in an oven at 200 C for 10 min. Small test specimens with an area of 9 cm2 were cut from the panel and folded to a concertina shape ready for migration studies. Exposure to Sunflower Oil The folded test specimens of coated panel were submerged in sunflower oil in a pressurised vial at 121 C in a silicon oil bath for 10, 20, 30, 40, 50, 60, 70, and 120 min. After exposure, the coated
Introduction
The internal surface of metal cans used to pack foodstuffs is often coated to form a barrier between the food and the metal of the can. The coating formulation may contain various components, such as resins, cross-linking agents, catalysts, lubricants, wetting agents, and solvents. The potential exists for these ingredients, or by-products of reactions between them, to migrate from the can coating into foods.
metal test specimens were wiped to remove the oil and submerged in acetonitrile (25 mL) overnight. A portion of the acetonitrile (1 mL) was passed through a 0.2-mm PTFE filter. Hydrochlorination Concentrated hydrochloric acid (100 L) was added to a portion (500 L) of the concentrated (10 times) EPH acetonitrile extract. This was allowed to react at 60 C for 18 h in a sealed vial and made up to 1 mL with acetonitrile. Liquid Chromatography-Fluorescence Detection (LC-FLD)
Instrument: Mobile phases: Agilent 1200 Series LC and G1321A FLD A = 0.1% acetic acid in water B = acetonitrile Gradient: At t = 0, B = 35%; t = 5 min, B = 50%; t = 10 min, B = 50%; t = 20 min, B = 100%; t = 25 min, B = 100% Flow rate: 1.0 mL/min Column: Agilent ZORBAX Eclipse XDB, 100 mm 2.1 mm, 3.5-m particle size (part number 961753.902) Injection: 20 L Excitation wavelength: 275 nm Emission wavelength: 305 nm Gain: 2^10
acetonitrile extract was evaporated to dryness under nitrogen. Acetic anhydride (25 L) and pyridine (25 L) were added and allowed to react for 15 min. The excess reagents were removed by evaporation under nitrogen and the residue was redissolved in acetonitrile (500 L). Another portion (5 mL) of EPH acetonitrile extract was evaporated under nitrogen and treated with acetic anhydride as described above. After removal of the excess reagents by evaporation, trifluoroacetic acid (TFAA, 100 L) was added. This was allowed to react for 15 min and the excess reagent removed by evaporation under nitrogen. The dry residue was redissolved in acetonitrile (500 L).

During migration tests using panels coated in a generic EPH coating, two co-eluting peaks were studied by LC-FLD. The identity of a pair of peaks at a similar retention time has been previously reported in the literature as two isomers of cyclo-diBADGE [4]. The structure of cyclo-di-BADGE is shown in Figure 1. The EPH coated panels were exposed to sunflower oil at 121 C for increasing periods of time and the co-eluting peaks were seen to behave differently, as shown in Figure 2. The profile of the two peaks, 2-1 and 2-2, was seen to change: as the length of exposure increased, the relative peak height of peak 2-1 decreased compared to that of peak 2-2, suggesting that this compound was migrating at a greater
H 3C CH 3
LC-FLD-TOF-MS
Instrument: Agilent 1200 Series LC, G1321A FLD and TOF positive electrospray Mobile phases: A = 0.1% acetic acid in water B = acetonitrile Gradient: At t = 0 min, B = 35%; t = 5 min, B = 50%; t = 20 min, B = 50%, t = 30 min, B = 100%; t = 40 min, B = 100% Flow rate: 0.3 mL/min Column: Agilent ZORBAX Eclipse XDB, 100 mm 2.1 mm, 3.5-m particle size (part number 961753.902) Injection: 5 L Excitation wavelength: 275 nm Emission wavelength: 305 nm Gain: 2^10 Nebulizer pressure: 30 psi Capillary: 4000 V Gas temperature: 325 C Drying gas: 10 L/min Fragmentor: 150 V
O HO O
O OH O
H 3C
CH 3
Acylation
Figure 1. Structure of cyclo-di-BADGE, C36H40O6.
A method reported by Biedermann and Grob was adapted for use [3]. A portion (5 mL) of the EPH
rate. After 120 min, the two peaks had approximately equal peak heights. The identification power of TOF-MS shown in previous applications [1-2] was used to investigate the EPH extracts with the aim of confidently identifying the co-eluting peaks and explaining the migration behavior seen. The flow rate of 1.0 mL/min used in the LC-FLD was deemed too high to be used on the LC-TOF-MS (although this ESI source can accept a flow rate of 1.0 mL/min), so the LC conditions were adapted. The LC-TOF-MS apparatus had a FLD in series. Figure 3 shows the FLD chromatogram of a concentrated acetonitrile EPH extract. The slower gradient means that the peaks are now eluting later (22.4 min and 23.15 min) and there are now three peaks, because the extra time on the column has allowed greater
interactions with the stationary phase, which in turn has allowed greater resolution between the peaks. This becomes clearer when looking at the TOF-MS data for this chromatogram (Figure 4). There are three peaks: 4-1 at 22.13 min, 4-2 at 22.61 min, and 4-3 at 23.45 min. Peaks 4-2 and 4-3 have the same mass spectra and molecular formula (C36H40O5), as shown in Figure 5 and Table 1. Peak 4-1 has a different molecular formula (C25H34O5). From this data the molecular formula of peak 4-1 was proposed as C25H34O5 with an identity of BADGE.BuOH. This identity was deduced based on methods reported in an earlier application [1]. The identity was confirmed by the addition of HCl to the extract. As expected, the BADGE.BuOH peak disappeared and was replaced by a peak with a
1.6E+05
0 min 10 min
1.2E+05
2-1
2-2
20 min 30 min
Intensity (mV)
40 min
8.0E+04
50 min 60 min 70 min
4.0E+04
120 min
0.0E+00 15.5 15.7 15.9 Retention time (min) 16.1 16.3
Figure 2.
LC-FLD chromatogram obtained after exposing the EPH can coating to sunflower oil for different lengths of time.
65.0
22.40
23.15
21.87
60.0
55.0 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0 13.0 14.0 15.0 16.0 17.0 18.0 19.0 20.0 21.0 22.0 23.0 24.0 Time, min
Figure 3.
LC-FLD chromatogram of the concentrated EPH acetonitrile extract.

22.13
a)
1.5e5
4-1
1.0e5
5.0e4
0.0 20.0 20.2 20.4 20.6 20.8 21.0 21.2 21.4 21.6 21.8 22.0 22.2 22.4 22.6 22.8 23.0 23.2 23.4 23.6 23.8 24.0 24.2 24.4 6.0e4 5.0e4 4.0e4 3.0e4 2.0e4 1.0e4 0.0 19.8 20.0 20.2 20.4 20.6 20.8 21.0 21.2 21.4 21.6 21.8 22.0 22.2 22.4 22.6 22.8 23.0 23.2 23.4 23.6 23.8 24.0 24.2 24.4 Time, min Time, min
b)
4-2
22.61
23.45
4-3
Figure 4. 4
Extracted ion chromatograms for the three peaks of interest: a) m/z 436.98 - 437.48 and b) m/z 591.03 - 591.53.
2.2e4 2.0e4 1.8e4 1.6e4

Intensity, counts
591.2733
a)
607.2459 586.3161
1.4e4 1.2e4 1.0e4 8000.0 6000.0 4000.0 2000.0 0.0 100 150 200 250 300 350 400 m/z, amu 450 437.2297 6.0e4 5.6e4 5.2e4 4.8e4 4.4e4 4.0e4 3.6e4 3.2e4 2.8e4 2.4e4 2.0e4 1.6e4 1.2e4 8000.0 4000.0 0.0 100 500 550 600 650
b)
432.2762
Intensity, counts
457.1966 473.1746
150
200
250
300
350 400 m/z, amu
450
500
550
600
650
Figure 5.
Mass spectrum of peak a) 4-1, b) 4-2, and 4-3.
Table 1. Peak 4-1 4-2 4-3
Molecular Formula Database Information for Peaks 4-1, 4-2 and 4-3 Mass 432.2746 437.2300 586.3163 591.2722 586.3161 591.2725 Molecular formula predicted C25H38NO5 C25H34O5Na C36H44NO6 C36H40O6Na C36H44NO6 C36H40O6Na Theoretical mass 432.2744 437.2298 586.3163 591.2717 586.3163 591.2717 Mass error (PPM) 0.12 0.35 0.026 0.83 0.37 1.3 Molecular adduct M+NH4 M+Na M+NH4 M+Na M+NH4 M+Na
molecular formula consistent with that of BADGE.BuOH.HCl, as the HCl adds across the remaining epoxide ring. Figure 6 shows the relevant extracted ion chromatograms. Figure 7 shows
the mass spectrum of BADGE.BuOH.HCl, with excellent correlation between experimental and theoretical chlorine isotope patterns.
1.8e5 1.6e5 1.4e5

Intensity, counts
22.13
a)
BADGE.BuOH peak
1.2e5 1.0e5 8.0e4 6.0e4 4.0e4 2.0e4 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0
9.46 10.69 18.23
9.0 10.0 11.0 12.0 13.0 14.0 15.0 16.0 17.0 18.0 19.0 20.0 21.0 22.0 23.0 24.0 Time, min
8.5e4 7.5e4 6.5e4

Intensity, counts
5.5e4 4.5e4 3.5e4 2.5e4 1.5e4 5000.0 1.0 4.8e4 4.4e4 4.0e4 3.6e4 2.0 3.0 4.0 5.0 6.0 7.0 8.0
No BADGE.BuOH.HCl peak
9.0 10.0 11.0 12.0 13.0 14.0 15.0 16.0 17.0 18.0 19.0 20.0 21.0 22.0 23.0 24.0 Time, min
3.86
b)
No BADGE.BuOH peak
11.94
Intensity, counts
3.2e4 2.8e4 2.4e4 2.0e4 1.6e4 1.2e4 8000.0 4000.0 0.0 1.0 1.4e5 1.3e5 1.2e5 1.1e5 9.5e4 8.5e4 7.5e4 6.5e4 5.5e4 4.5e4 3.5e4 2.5e4 1.5e4 5000.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0
6.83
9.41 11.58
13.51 17.46
9.0 10.0 11.0 12.0 13.0 14.0 15.0 16.0 17.0 18.0 19.0 20.0 21.0 22.0 23.0 24.0 Time, min
9.07
BADGE.BuOH.HCl peak
20.08
Intensity, counts
0.74
11.94
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
9.0 10.0 11.0 12.0 13.0 14.0 15.0 16.0 17.0 18.0 19.0 20.0 21.0 22.0 23.0 24.0 Time, min
Figure 6.
Extracted ion chromatograms for BADGE.BuOH (m/z 437 438) and BADGE.BuOH.HCl (m/z 451 452) for a) untreated EPH extract and b) EPH extract treated with HCl.
468.2500 5.2e4 4.8e4 4.4e4 4.0e4 3.6e4 3.2e4

Intensity, counts
M+NH 4 +
2.8e4 2.4e4 2.0e4 1.6e4 1.2e4 8000.0 4000.0 0.0 400 410 420 430 440 450 460 470 480 453.2254
M+K+ M+H+
451.2251 470.2502 469.2558 488.2191
M+Na+
473.2071 490.2162
490
500
510
520
530
540
550
m/z, amu
Figure 7.
Mass spectrum of the peak at 20.1 min, corresponding to BADGE.BuOH.HCl.
Peaks 4-2 and 4-3 had the same mass spectra and were proposed to be C36H40O6. As well as corresponding to cyclo-di-BADGE as suggested above, this could also correspond to BADGE.BPA, a linear BADGE derivative with the same molecular formula (see Figure 8).
H 3C CH 3
H 3C
CH 3
O O
O OH
OH
Figure 8.
Structure of BADGE.BPA, C36H40O6.
The addition of HCl did not change the two co-eluting peaks, which suggests that they are in fact due to cyclo-di-BADGE and not BADGE.BPA, as any peaks due to BADGE.BPA would be expected to disappear as the HCl will add across the epoxide ring. This conclusion was tested further; one of the differences between the proposed structures is that cyclo-di-BADGE has two hydroxide groups but BADGE.BPA has additional epoxide functionality as well as two hydroxide groups. It has been reported that acetic anhydride and TFAA react differentially with hydroxide and epoxide groups [3]. Acetic anhydride causes acylation of free hydroxide groups while further addition of TFAA causes
acylation across the epoxide ring (see Figure 9). LC-TOF-MS analysis of the EPH extract after treatment with acetic anhydride showed the presence of a doubly acylated compound, either compound 9-1 or 9-2 (C40H44O8). After further addition of TFAA the absence of a peak corresponding to compound 9-3 suggests that the two co-eluting peaks are in fact both due to cyclo-di-BADGE. The reason for two peaks is proposed to be because of the presence of stereoisomers, with the two hydroxide groups being cis- or trans- to each other, depending upon the side from which the phenol group attacks the epoxide ring during formation [3].
a)
H 3C CH 3
O O O CH 3
H 3C
CH 3
O O O CF 3
O OH O
O OH O
H 3C
O O O CH 3 O
O O O O H 3C
F 3C
No further reaction
H 3C
CH 3
H 3C
CH 3
Cyclo-di-BADGE b)
H 3C O CH 3 O OH O H 3C CH 3
H 3C O O O CH 3
9-1
H 3C
CH 3 O O O O
H 3C
CH 3
OH
O H 3C O
9-2
CH 3
BADGE.BPA
O F 3C O O CF 3
O F 3C F 3C O O O O
H 3C
CH 3 O O O O CH 3
H 3C
CH 3
O H 3C O
9-3
Figure 9.
Acylation reactions for a) cyclo-di-BADGE and b) BADGE.BPA.
Conclusions
During migration studies at Leeds University, two co-eluting peaks were seen to behave differently during exposure to sunflower oil at different temperatures. There were in fact three co-eluting peaks and these were identified using LC-TOF-MS. The first peak was identified as BADGE.BuOH, and the second and third peaks were confirmed as cis- and trans- isomers of cyclo-di-BADGE. It is suggested that the differences seen in the original migration studies were due to the differences in structure between BADGE.BuOH and cyclo-di-BADGE, with the BADGE.BuOH migrating faster than cyclo-diBADGE into the simulants, and the two stereoisomers of cyclo-di-BADGE migrating at the same rate as each other, which would account for the changes in peak height.
Acknowledgements
This work was carried out as a part of a Defra LINK project: FQS45 New technologies and chemistries for food can coatings. Funding by Defra and matching funds in kind from Valspar Corporation, Impress Group, and H. J. Heinz are gratefully acknowledged. The contents of this application are the responsibility of the authors alone and should not be taken to represent the views of the supporting organizations.

References
1. M. Driffield, E. L. Bradley, L. Castle, and J. Zweigenbaum, Identification of Unknown Reaction By-Products and Contaminants in Epoxyphenolic-Based Food Can Coatings by LC-TOF-MS (2006) Agilent Technologies publication 5989-5898EN 2. M. Driffield, E. L. Bradley, L. Castle, and J. Zweigenbaum, Identification of Unknown Polyester Oligomers in New Polyester Food Can Coatings by LC-TOF-MS Using Molecular Feature Extraction and Database Searching, (2007) Agilent Technologies publication 5989-7393EN 3. M. Biedermann and K. Grob, Food Contamination from Epoxy Resins and Organosols Used as Can Coatings: Analysis by Gradient NPLC, Food Additives & Contaminants, (1998) 15, 5, 609-618. 4. A. Schaefer and T. J. Simat, Migration from Can Coatings: Part 3. Synthesis, Identification and Quantification of Migrating Epoxy-Based Substances Below 1000 Da, Food Additives & Contaminants, (2004) 21, 4, 390-405.
Identification of Unknown Reaction By-Products and Contaminants in Epoxyphenolic-Based Food Can Coatings by LC/TOF-MS
Application
Food Safety
Authors
M. Driffield, E. L. Bradley, and L. Castle Central Science Laboratory Sand Hutton York, YO41 1LZ UK J. Zweigenbaum Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808 USA
foods. Thus existing and especially new coatings need to be evaluated for their safety for contact with food and beverages. We will illustrate this evaluation using the example of epoxyphenolic can coatings based on bisphenol A epoxy resins. These are cured by stoving with phenolic resins to produce a three-dimensional crosslinked network to provide the chemical and pack resistance required for food and beverage cans. The epoxy monomer bisphenol A diglycidyl ether (BADGE, see Figure 1) participates in these polymerization reactions via its reactive epoxide groups. However, it can also undergo addition from attacking nucleophiles such as water or solvents to give lower molecular weight products that might migrate into the packed food [13]. These potential migrants need to be identified.
H 3C CH 3
Abstract
This application illustrates how time-of-flight mass spectrometry can be used in the safety evaluation of new and existing can coatings used in the food industry. The accurate mass provides information for the parent compound and fragment ions greatly increase the confidence in the identification process.
O O
Introduction
The internal surface of metal cans used to pack foodstuffs is often coated to form a barrier between the food and the metal of the can. The coating formulation may contain various components such as resins, crosslinking agents, catalysts, lubricants, wetting agents, and solvents. The potential exists for these ingredients, or by-products of reactions between them, to migrate from the can coating into
O Figure 1. Chemical structure of BADGE, C21H24O4.
The accurate mass measurements provided by time-of-flight (TOF) mass spectrometry (MS) for unknown compounds makes this identification process possible without the need for authentic standards of every possible minor impurity and reaction by-product.
Experimental
Sample Extraction A metal panel (250 cm2) coated with an epoxyphenolic lacquer and stoved under industrial conditions was cut into pieces (approximately 1 cm2) and extracted by immersion in acetonitrile (100 mL). After 18 hours the extract was evaporated to a small volume (1 mL). LC Conditions
Instrument: Mobile phases: Gradient: Agilent LC 1200 SL A: water B: acetonitrile 20% B to 50% B over 25 min, hold 20 min, 100% B at 60 min, hold 10 min, return to 20% B over 10 min 0.2 mL/min Agilent ZORBAX Eclipse XDB, 100 mm 2.1 mm, 3.5-m particles Part number 961753.902 5 L
MS Conditions
Instrument: Nebulizer press.: Capillary: Gas temp.: Drying gas: Agilent 6210 LC/MS TOF in positive ion ESI mode 30 psi 4000 V 300 oC 7 L/min

TOF-MS parameters were optimized using solvent standards of BADGE, as mainly BADGE derivatives were expected to be extracted from the coating [1]. A fragmentor value of 150 V was used first, to cause no fragmentation, and so molecular ion adducts were seen. Figure 2 shows the TIC for the acetonitrile extract of the epoxyphenolic coating. There are many unknown peaks, and the one at 27.2 min was chosen for this example.
Flow rate: Column:
Injection:
7.0
6.0
5.0
Intensity x 10 7, cps
4.0
3.0
2.0
1.0
0.0
10
15
20
25
30
35
40 45 Time, min
50
55
60
65
70
75
80
Figure 2.
Total ion chromatogram of the acetonitrile extract of the epoxyphenolic coating.
Figure 3 shows the mass spectrum of the peak at 27.2 min. The differences in masses between the ions suggest that these are due to the protonated, ammoniated, sodiated, and potassiated molecule. No ammonia, sodium, or potassium was added to the mobile phase, and it is likely that these adducts arose due to contamination from other work carried out on the instrument, or were present in the solvents used in the mobile phase. The formula calculator was used to propose identities for the peak, using the accurate mass determined for [M+NH4]+, as it was the most intense.
Only one possible empirical formula was provided limiting the elements to C, H, O, and only one N within the 5 ppm mass error limit used. For the experimentally derived mass 494.3118, the formula C27H44O7N was proposed (theoretical mass 494.3112, 1.15 ppm error). As it is proposed that this is the ammoniated adduct (subtract NH4), this gives a formula of C27H40O7 for the unknown peak. Furthermore, it is suspected that this peak is a BADGE derivative (subtract C21H24O4 from the formula) and this suggests that the unknown peak is BADGE + C6H16O3.
[M+NH4] + 494.3118
10.0
[M+Na] + 499.2674
Intensity x 10 4, counts
7.50
Max. 1.2e5 counts
5.00
2.50
[M+H] + 477.2842
[M+K] + 515.2399
0.00 480 490 500 m/z, amu 510 520
Figure 3.
Mass spectrum of the unknown peak at 27.2 min (fragmentor = 150 V).
Fragmentation experiments were carried out to aid the identification process. A fragmentor value of 275 V dissociated the ammoniated molecular adduct into fragment ions, see Figure 4. The accurate masses of the fragment ions were put into the formula calculator and the structures of the ions were theorized from the proposed empirical formulae. The fragment ions confirmed the presence of the BADGE unit (m/z 341.1727), that one of the epoxide rings had reacted with water (fragment ion at m/z 209.1149), and the other had reacted with butoxyethanol (BuOEtOH, C6H16O3) (fragment ion at m/z 309.2036), a solvent used in the manufacturing process of the coating formulation. Figure 5 shows the structure of BADGE.H2O.BuOEtOH.
Using the same approach, the identity of virtually all of the peaks in Figure 2 was established and different can coating chemistries have been studied.
Conclusions
Solvent extracts of epoxyphenolic can coatings have been analyzed by LC/TOF-MS to identify potential migrants into food and beverages. Accurate mass data of the parent compound and the fragment ions allows confident assignment of previously unknown peaks. Using the LC/TOF-MS has helped the testing of existing can coatings and guided the development of new coating chemistries.
Max. 1.2e5 counts

[M+NH4] + 494.3118 [M+Na] + 499.2674
10.0 Intensity x 10 4, counts
Fragmentor = 175 V
7.50
5.00
2.50
0.00
Fragmentor = 275 V
H 3C CH 3 CH 3 H 3C + + O O OH H 3C CH 3 + H 2C OH O
Max. 1.2e6 counts [M+Na] + 499.2657

C 4H 9
10.0 Intensity x 10 4, counts

+ CH 2 HO OH
309.2036
7.50
HO
209.1149 135.0782
5.00
341.1727
2.50
0.00 100
200
300 m/z, amu
400
500
Figure 4.
TOF-MS of the unknown peak at 27.2 min.
H 3C
CH 3 O
HO OH
O OH
C 4H 9
Figure 5.
Structure of the identified compound: BADGE.H2O.BuOEtOH.
References
1. A. Schaefer and T. J. Simat (2004) Food Additives and Contaminants, 21, 4, 390405. 2. N. Leepipatpiboon, O. Sae-Khow, and S. Jayanta (2005) Journal of Chromatography A, 1073, 12, 331-339. 3. O. Pardo, V. Yusa, N. Leon, and A. Pastor (2006) Journal of Chromatography A, 1107, 12, 7078.
Acknowledgements
This work was carried out as part of a Defra LINK project: New technologies and chemistries for food can coatings, Project number: FQS45.

The information contained in this publication is intended for research use only and is not to be followed as a diagnostic procedure. Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2006 Printed in the USA December 21, 2006 5989-5898EN
Trace Level Hydrocarbon Impurities in Ethylene and Propylene
Application
Gas Chromatography March 1997
Authors
Vince Giarrocco and Roger Firor Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA
Abstract
An Agilent 6890 Series gas chromatography system was used to determine trace (low ppm) levels of hydrocarbon impurities in high-purity ethylene and propylene. The gas chromatograph (GC) was equipped with a heated gas sample valve, split/splitless inlet, and flame ionization detector (FID). An Agilent HP-PLOT Al2O3 column was used for separation of the trace hydrocarbons. Impurity levels below 10 ppm (mole) were easily detected in both ethylene and propylene.
Typically, these low molecular weight monomers are of very high purity (99.9+ percent). However, hydrocarbons, sulfur compounds, and other impurities in feed streams can cause such problems as reduced catalyst lifetime and changes to product quality. Process yields can also be adversely affected. Many impurities have been identified as potential contaminants (1,2). Recently, ASTM has proposed several procedures to determine trace hydrocarbon impurities in both ethylene and propylene (3). These methods, currently in the investigation stage, use alumina porous layer open tubular (PLOT) columns. This application note describes the suggested Agilent configuration for such methods and illustrates resulting separations of both quantitative calibration blends and actual process samples. These proposed methods should be valuable in meeting commercial specifications.
ionization detector (FID). All gas flows and pressures within the GC were controlled electronically. Gas sample injections were made using an automated sample valve placed in the 6890 valve oven (80 C). The gas sample valve was interfaced to the capillary inlet using an aluminum tube (1/8-in.) that jacketed the stainless steel transfer line (option 860). The inlet was fitted with a split/splitless liner (part no. 19241-60540). All injections were made in the split mode. A 50-m 0.53-mm, HP-PLOT Al2O3 M column was used for separation. For ethylene analysis only, a 30-m 0.53-mm, 5-m HP-1 column was placed directly behind the HP-PLOT column. The two columns were joined using a glass press-fit connector. The Agilent ChemStation was used to control the 6890 Series GC and to provide data acquisition and peak integration. The ChemStation was operated at a data acquisition rate of 10 Hz.
Introduction
High-purity ethylene and propylene are commonly used as feedstocks for production of polyethylene, polypropylene, and other chemicals.
Experimental
All experiments were performed on an 6890 Series gas chromatograph (GC) equipped with a split/splitless inlet and capillary optimized flame
Standards for retention time and response factor calculation were obtained from DCG Partnership (Houston, Texas, USA 77061). Samples used for this work were obtained from commercial sources. Table 1 lists the entire set of equipment and conditions.
Table 1.
Instrument Configuration and Operating Conditions

Description 6890 Series GC Split/splitless inlet Capillary optimized FID 6-port gas sample valve and automation Valve oven Valve to inlet interface 50-m x 0.53-mm HP-PLOT Al2O3 M (part no. 19095P-M25) 30-m x 0.53-mm, 5- m HP-1 (part no. 19091Z-236), used for ethylene analysis only Agilent ChemStation 200 C 250 C 10/1 to 50/1 depending on sample 30 mL/min hydrogen, 350 mL/min air, nitrogen make-up (25 mL/min column + makeup) Ethylene: 35 C (2 min), 4 C/min to 190 C Propylene: 40 C (2 min), 4 C/min to 190 C Ethylene: 0.5 mL Propylene: 0.25 mL Ethylene: 6 mL/min constant flow (10 psi) Propylene: 3.5 mL/min constant flow (4 psi) 80 C
Item Gas Chromatograph G1540A Option 112 Option 211 Option 701 Option 751 Option 860 Column

Ethylene
The configuration used for ethylene analysis is found in figure 1. The HP-PLOT Al2O3 column was used for hydrocarbon separation. The use of HP-PLOT Al2O3 columns for light hydrocarbon analyses has been previously described (4). These columns exhibit excellent separation characteristics for the C1 through C5isomers. The proposed method for ethylene specifies the use of a second nonpolar column placed after the HP-PLOT alumina column to improve the separation of impurity peaks eluting on the tail of ethylene. This nonpolar column gains importance for trace level analysis, where higher concentrations of ethylene (99.9 percent and higher) exhibit increased tailing. No attempt was made to compare separations without the nonpolar Agilent HP-1 column.
Data Acquisition G2070AA Operating Parameters Injection port temperature Detector temperature Split ratio FID conditions Temperature program Injection volume Column flow Valve temperature
Split Inlet 10/1 to 40/1 Helium 4 5 6 1 3 2 30-m x 0.53-mm, 5- m HP-1
FID 0.5-cc Loop Sample In/Out 50-m x 0.53-mm HP-PLOT Al 2 O 3M
Figure 1. Valve drawing for impurities in ethylene.
Figure 2 shows the chromatogram of an ethylene calibration blend containing most of the major hydrocarbon impurities. This sample was analyzed at a split ratio of 10/1. The concentration of most components (except for ethane) ranges from 8 to 12 ppm (mole). For this analysis, baseline separation is achieved for all the impurities except for propane. Total analysis time is approximately 30 minutes. Because this separation is more than adequate, analysis time can be reduced by increasing the temperature program rate. Based upon conditions used for this analysis, most components can be detected at the 1-ppm level. Chromatographic results for two process ethylene samples are given in figures 3 and 4. The sample presented in figure 3 contains only methane, ethane, and propylene as impurities. Less than 1-ppm methane was detected. The ethylene sample in figure 4 shows a high concentration of methane, with trace amounts of ethane, propane, and propylene.
1. 2. 3. 4. 5.
Methane (10 ppm) Ethane (219 ppm) Ethylene Propane (12 ppm) Propylene (9 ppm),
6. Isobutane (10 ppm) 7. n-Butane (10 ppm) 8. Propadiene (10 ppm) 9. Acetylene (10 ppm) 10. t-2-Butene (8 ppm)
11. 1-Butene (8 ppm) 12. Isobutylene (9 ppm) 13. c-2-Butene (9 ppm) 14. 1,3-Butadiene (10 ppm) 15. Methylacetylene (9 ppm)
Figure 2. Chromatogram of ethylene calibration blend, split ratio 10/1.
1. 2. 3. 4.
Methane (<1 ppm) Ethane (5 ppm) Ethylene Propylene (113 ppm)
Figure 3. Chromatogram of process ethylene sample, split ratio 20/1.
Propylene
The configuration used for propylene analysis is illustrated in figure 5. This configuration is essentially the same as for ethylene, but without the HP-1 column. The sample volume was reduced to 0.25 mL. Propylene was sampled in the gas state.
1. 2. 3. 4. 5. Methane (4969 ppm) Ethane (18 ppm) Ethylene Propane (2 ppm) Propylene (5 ppm)
Figure 4. Chromatogram of process ethylene sample.
Split Inlet 10/1 to 40/1 Helium 4 5 6 1 3 2 FID
0.25-cc Loop Sample In/Out
50-m x 0.53-mm HP-PLOT Al 2 O 3 PLOT M
Figure 5. Valve drawing for impurities in propylene.
A chromatogram representing the trace hydrocarbon impurities in propylene is shown in figure 6. This sample was analyzed at a split ratio of 20/1. The concentration of most impurities range from 8 to 20 ppm. Ethylene is present at a higher concentration level. Most of the impurities in the sample are well separated using the conditions described in table 1. Cyclopropane elutes just before propylene and is baseline separated under these conditions. Several of the C4 hydrocarbons elute on the tail of the high-purity propylene. This affects the lower limit of detection for these peaks, compared to those components that are baseline separated. The remainder of the C4 and C5 impurities are well separated.
1. 2. 3. 4. 5. 6. 7. 8. 9.
Methane Ethane (10 ppm) Ethylene (50 ppm) Propane Cyclopropane (10 ppm) Propylene Isobutane (10 ppm) n-Butane (7 ppm) Propadiene
10. Acetylene 11. t-2-Butene (10 ppm) 12. 1-Butene 13. neo-Pentane 14. Isobutylene (9 ppm) 15. Isopentane 16. c-2-Butene (9 ppm) 17. n-Pentane (10 ppm) 18. 1,3-Butadiene (9 ppm)
Figure 6. Chromatogram of propylene calibration standard.
For comparison, figure 7 shows the analysis of a second calibration blend containing a higher level of impurities (50 to 1000 ppm). Figure 8 presents the chromatographic results for a high-purity propylene process sample. This sample contains only ethane and propane impurities.
1. 2. 3. 4. 5. 6. 7. 8. 9.
Methane Ethane Ethylene Propane (988 ppm) Cyclopropane (100 ppm) Propylene Isobutane (129 ppm) n-Butane Propadiene (62 ppm)
10. Acetylene (48 ppm) 11. t-2-Butene 12. 1-Butene 13. Isobutylene 15. Isopentane 16. c-2-Butene 17. 1,3-Butadiene 18. Methylacetylene (100 ppm)
Summary
This application note describes two methods for analyzing trace hydrocarbon impurities in ethylene and propylene. These methods use a gas sample valve with split injection, an Agilent HP-PLOT Al2O3 and HP-1 (for ethylene only) column, and an FID. Impurities below the 10-ppm mole level can be easily quantitated using these methods. For some impurities, especially those that are well separated from the large ethylene or propylene peaks, detection limits were estimated to be about 1 ppm.
Figure 7. Chromatogram of propylene calibration blend containing higher levels of impurities.
1. Ethane (82 ppm) 2. Propane (4358 ppm) 3. Propylene
Figure 8. Chromatogram of process propylene sample.
References
1. ASTM Method D 5325, Standard Guide for the Analysis of Ethylene Product, Annual Book of Standards, Volume 5, ASTM, 100 Bar Harbor Drive, West Conshohocken, PA 19428 USA. 2. ASTM Method D 5273, Standard Guide for the Analysis of Propylene Concentrates, Annual Book of Standards, Volume 5, ASTM, 100 Bar Harbor Drive, West Conshohocken, PA 19428 USA. 3. Proposed methods for hydrocarbon impurities in ethylene and propylene by gas chromatography are being investigated under ASTM committee D-2, subcommittee D. 4. Optimized Determination of C1C6 Impurities in Propylene and Ethylene Using HP-PLOT/Al2O3 Columns, Agilent Technologies, Inc. Publication (43) 5062-8417E, March 1994.
Contaminants PCBs
Direct Injection of Fish Oil for the GC-ECD Analysis of PCBs: Results Using a Deans Switch With Backflushing Application
Environmental and Pharmaceutical
Author
Philip L. Wylie Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808 USA
Abstract
A Deans switch, employing Agilent's Capillary Flow Technology, was configured on an Agilent 7890A gas chromatograph (GC) equipped with dual electron capture detectors (ECDs). A method was developed for the analysis of fish oil for polychlorinated biphenyl (PCB) contamination. The Deans switch was used to heart cut 7 indicator PCBs (IUPAC congeners 28, 52, 101, 118, 138, 153, and 180) from the primary DB-XLB column onto a DB200 column for further separation. Fish oil from a supplement capsule was simply diluted 1:10 in isooctane and injected directly. In a separate experiment, the fish oil was analyzed by GC with a flame ionization detector (GC/FID) without backflushing. From these analyses, it was estimated that about two-thirds of the fish oil components would remain on the column after the 17.4-minute GC/ECD run. To prevent carryover, contamination, and retention time shifts, the Deans switch was used to backflush the primary column at the end of each run. Evidence shows that backflushing removed the fish oil residue, which otherwise would quickly degrade the chromatography.
beneficial health affects. In addition to eating fish, many people take fish oil as a supplement to their daily diet. However, fish, especially those high on the aquatic food chain, can bioaccumulate fat-soluble pollutants. Among these are polychlorinated dibenzodioxins (PCDDs), polychlorinated dibenzofurans (PCDFs), and polychlorinated biphenyls (PCBs). Therefore, fish oil used in supplements undergoes a variety of analyses, including tests for halogenated pollutants. One of the quality assurance tests is to analyze fish oil for PCB contamination. This is complicated by the fact that fish oil is a very complex mixture containing high-boiling fatty acids and triglycerides of fatty acids; chain lengths are mostly between 14 and 22 carbons. They also contain varying amounts of phospholipids, glycerol ethers, wax esters, and fatty alcohols. PCB analysis is complex by itself, with 209 possible congeners. Of these, 140 to 150 have been observed in commercial PCB mixtures called Aroclors. PCB analysis usually focuses on the 12 planar, dioxin-like PCBs and/or on seven indicator PCBs (IUPAC Numbers 28, 52, 101, 118, 138, 153, and 180). To obtain sufficient sensitivity and selectivity for these compounds, analysts have traditionally employed very expensive techniques such as highresolution mass spectrometry (HR/MS) or HR/MS/MS. Analysis of the fish oil generally follows a series of extraction and cleanup steps. This paper focuses on the analysis of the seven indicator PCBs in fish oil using an Agilent 7890A GC configured with a Deans switch, two columns of differing selectivity, and dual electron capture
Introduction
Fish oils contain high levels of eicosapentanoic acid (EPA) and docosahexanoic acid (DHA), omega-3 fatty acids that are thought to have
detectors (ECDs). Fish oil from a commercially available supplement was simply diluted 10:1 in isooctane and injected into the GC. No cleanup steps were employed.
Instrumental Conditions for Analysis Inlet Split/splitless at 330 oC Oven temperature program 80 oC (1 min), 50 oC/min to 200 oC (0 min), 10 oC/min to 290 oC (5 min) Detectors Dual ECD at 340 oC N2 at 60 mL/min H2 at 41.040 psig (constant pressure mode) H2 at 20.610 psig (constant pressure mode) ECD make-up gas Inlet pressure
Experimental
The fish oil supplement was obtained from a local grocery store. According to the bottles label, each gelatin capsule contains 1.0 g of fish oil of which 180 mg is EPA and 120 mg is DHA. Oil was removed from a capsule and diluted with isooctane (pesticide grade from Sigma-Aldrich, Milwaukee, WI, USA) to make a 10% solution. This solution was spiked with various Aroclors (Supelco, Bellefonte, PA, USA) or with individual PCB congeners (AccuStandard, New Haven, CT, USA). Table 1 lists the instrumentation and experimental conditions for the analysis.
Table 1. Instrumentation and Experimental Conditions
PCM pressure to Deans switch
Post-Run Backflush Conditions Post-run duration 2.4 min Inlet pressure PCM pressure Oven temperature during backflush H2 at 0 psig H2 at 80 psig 290 oC for 2.4 min

Without backflushing, the high-boiling components of fish oil can be retained by the GC column, causing severe carryover problems from one run to the next. After a few injections, so much of the fish oil residue builds up on the column that it causes PCB retention times to shift by a minute or more. Such dramatic retention time shifts would prevent the use of the Deans switch, where heart cuts are just a few seconds wide. Deans SwitchHeart Cutting The Deans switch is one of Agilents new devices that employ Capillary Flow Technology. These devices have extremely low dead volumes, are inert, and do not leak, even with large cycles in oven temperature. Columns are easy to install into the Deans switch, which is mounted on the side of the oven wall (Figure 1).
Instrumentation and Software Gas chromatograph Agilent 7890A Automatic sampler Agilent 7683B Series injector and tray Primary column Primary column connections Secondary column Secondary column connections Restrictor Restrictor connections Inlet liner Auxiliary pressure control device Deans switch calculator software J&W 30-m 0.18-mm 0.18-m DB-XLB (P/N 121-1232) Split/splitless inlet to Deans switch J&W 30-m 0.25-mm 0.50-m DB-200 (P/N 122-2033) Deans switch to back ECD 76.8-cm 0.100-mm deactivated fused silica tubing Deans switch to front ECD Agilent deactivated single taper with glass wool (P/N 5062-3587) Agilent 7890A Pneumatic Control Module (PCM) Option # 309 Agilent Technologies Deans Switch Calculator (Rev. A.01.01)
Software for data acquisition Agilent GC ChemStation and analysis (Rev. B.03.01)
Restrictor to back detector
Primary column DB-XLB
Pneumatic connections to solenoid valve
Secondary column DB-200
Figure 1.
Photograph of the Deans switch installed on the side of the 7890A GC oven. The column and restrictor connections are indicated by an * in Figure 2a.
As shown in Figure 2a, the 30-m 0.18-mm 0.18-m DB-XLB column is connected between the split/splitless inlet and the Deans switch. A short length of deactivated fused silica tubing (76.8 cm 0.100 mm) connects the Deans switch to the front ECD. The secondary column (30-m 0.25-mm 0.5-m DB-200) was chosen because it is more polar than the DB-XLB column and has a different selectivity for PCBs. It has an upper temperature limit of 300 C, which is high enough to elute the PCBs of interest. Figure 2a shows the Deans switch in the normal mode with the solenoid valve in the off position.
In this mode, the effluent from the primary DB-XLB column is directed through the restrictor to the front ECD. When the solenoid valve is switched, the effluent is directed through the secondary DB-200 column to the back ECD (Figure 2b). The retention times for the seven indicator PCBs were initially determined with the valve in the off position. Using the timed events table in the ChemStation, the valve was switched to on just before each PCB peak and off immediately after. This produced seven heart cuts that were directed through the DB-200 column to the back ECD.
Front ECD
*
Restrictor Solenoid valve (off)
S/S Inlet
41.040 psig
*
DB-XLB
PCM
20.610 psig
Back ECD
DB-200
Figure 2a. Deans switch in the no cut position. The effluent from the DB-XLB column goes directly to the front ECD through the short restrictor. The intersections marked with an * are column and restrictor connections to the Deans switch plate (Figure 1).
Front ECD
Restrictor
Solenoid valve (on)
S/S Inlet
41.040 psig
PCM
20.610 psig
DB-XLB
Back ECD
DB-200 Figure 2b. Deans switch in the cut position. The effluent from the DB-XLB column goes to the DB-200 column and then to the back ECD.
In some Deans switch systems, the second column is placed in a separate GC oven or cryogenic cooling is used to trap the heart cut components at the head of the second column. In this case, both columns were mounted inside of the 7890A oven and cooling was not used to focus compounds at the head of the DB-200 column. Figure 3a shows the chromatogram for a fish oil sample spiked with Aroclor 1260. PCBs 28, 52, 101,
units 120000
118, 138, 153, and 180 were cut out of the primary chromatogram (Figure 3b) and sent to the second column (Figure 3c). The purpose of the DB-200 column is to resolve the target PCBs from other PCBs and matrix components that co-elute with them on the DB-XLB column. Six of the 7 PCBs appear to be well resolved on the DB-200 column. PCB 118 is only partially resolved by this method.
100000
80000
60000
40000
20000
0 8 10 12 14 16 min
Figure 3a. GC/ECD chromatogram of Aroclor 1260 spiked into fish oil. This is the effluent from the primary DB-XLB column with seven heart cuts.
units
700
600
500
400
300
200
28
100 8
52
101
118 153
138
180
10
12
14
16
min
Figure 3b. Enlargement of the chromatogram in Figure 3a showing where heart cuts were made for the seven target PCBs. 5
units
153 138
80000 70000 60000 50000 40000 30000 20000 10000
180 101
28
0 8
52
10 12
118
14
16
min
Figure 3c. GC/ECD chromatogram from the DB-200 column. The peaks in this chromatogram were heart cut from the DB-XLB column. Except for congener 118, the target PCBs were separated from co-eluting interferences by the DB-200 column.
Deans SwitchBackflushing Data collection with the Deans switch system ended at 17.4 min with the oven at 290 C. While it was assumed that a lot of the fish oil components remained on the column at this point, it was impossible to tell because the ECD responds poorly to these compounds. The fish oil does contribute some small peaks (both positive and negative) to the chromatogram, but it is impossible to see the full contribution of the fish oil. So a sample of the fish oil was analyzed on an identical DB-XLB column using a flame ionization detector (FID) with no Deans switch installed. The temperature was held at 290 C for an extra 25 minutes to determine if high boiling compounds were still eluting.
Figure 4 shows that a great deal of the fish oil continued to elute after 17.4 minutes (arrow in figure). When a blank run was made with a final oven temperature of 310 C, much more of the fish oil eluted from the column (Figure 4, middle chromatogram). A second blank run (Figure 4, top chromatogram) showed that fish oil components were still eluting from the column. In actuality, only about a third of the fish oil comes off the column under the Deans switch conditions. This is why other fish oil methods begin with a solvent extraction followed by solid phase extraction for sample cleanup.
Response 9000000 8500000 8000000 7500000 7000000 6500000 6000000 5500000 5000000 4500000 4000000 3500000 3000000 2500000 2000000 1500000 1000000 500000 0 5.00 10.00 15.00 20.00 25.00 30.00 35.00 40.00 Time
2nd Bakeout
1st Bakeout
1 L 10% fish oil splitless injection
Figure 4.
GC/FID chromatogram from a 1 L splitless injection of 10% fish oil using a 30-m 0.18-mm 0.18 m DB-XLB column. The arrow indicates where the GC/ECD method ends and the post-run backflush begins. In this case, there was no backflushing so the oven was held at 290 C for an extra 25 min. The run was repeated two more times without injection but with the oven held at 310 C for 30 minutes at the end of the run. Residue from the fish oil injection continued to elute, even during a second bakeout step.
The 7890A has been designed to make column backflushing a routine process. It has been shown empirically that backflushing should continue for about five times the holdup time. In this case the column was held at 290 C during the post run backflush. At the same time, the inlet pressure was dropped to 0 psig while the PCM pressure was increased to 80 psig. Using Agilents GC Pressure/ Flow Calculator software, the H2 flow rate backwards through the column was 3.81 mL/min and the holdup time was 0.466 min. Backflushing was, therefore, continued for 2.4 minutes, which is slightly more than five times the calculated holdup value. Figure 5 shows the Deans switch in the backflush mode.
Front ECD
Restrictor
Solenoid valve (off)
S/S Inlet
0 psig
PCM
80.000 psig
DB-XLB
Back ECD
DB-200 Figure 5. Deans switch in the backflush mode. The inlet pressure is dropped to 0 (or 1) psig while the PCM pressure is raised to 80 psig. This causes the carrier gas to flow backwards through the DB-XLB column. The reverse flow sweeps highboiling fish oil components off the head of the column and out the split vent.
As mentioned earlier, just a few injections of fish oil can cause dramatic shifts in PCB retention times. Backflushing forces the remaining fish oil components backwards through the primary column and out through the split vent. This prevents fish oil buildup on the column, thus eliminating carryover and retention time shifts. Figure 6a compares the first and last chromatograms in a six-run sequence. One-L splitless injections were made of 10% fish oil spiked with Aroclor 1260. This sequence was run after many previous injections of fish oil using this method, and it is clear that the retention times did not shift. Figure 6b shows the seven PCBs that were heart cut from the two analyses shown in Figure 6a. Figure 6b shows no differences in the first and last heart cut chromatograms, providing further proof that there were not even subtle shifts in the PCB retention times. Each heart cut was just 4 to 5 seconds wide, so very small RT shifts in the first column would dramatically alter the results in the second.
units
1500
1000
500
-500
-1000
-1500
10
11
12
13
14
15
16
min
Figure 6a. First (top) and sixth (inverted) injections of 10% fish oil spiked with Aroclor 1260. Seven Deans switch cuts were made from this DB-XLB column in order to isolate PCBs 28, 52, 101, 118, 138, 153, and 180. The DB-XLB column was backflushed after each run, preventing build-up of fish oil residue. The comparison shows that there was no shift in retention times caused by fish oil accumulation.
101
units
180 153 138
6000
4000
118 28 52
2000
-2000
118
-4000
-6000
-8000 9 10 11 12 13 14 15 min
Figure 6b. Chromatogram of the seven PCB congeners and interferences that were cut from the DB-XLB column to the DB-200. The first chromatogram (top) and sixth (inverted) are identical, providing further proof of retention time stability. Any retention time shift on the primary column would severely alter the appearance of the secondary chromatogram.
Conclusions
This paper demonstrates that it is possible to analyze PCBs in fish oil without performing laborious sample cleanup prior to GC injection. A Deans switch was used to cut seven target PCBs (28, 52, 101, 118, 138, 153, and 180) from a DB-XLB column for further separation on a DB-200 column. This produced nearly baseline separation of the target PCBs. Only congener 118 was not well separated from co-eluting PCBs. Further refinement of the oven temperature program would be needed to isolate this congener. It has been estimated that about two-thirds of the fish oil remained on the primary GC column at the end of the run. By setting the Deans switch to the backflush mode for just 2.4 minutes at the end of each run, this material was swept backwards through the column and out the split vent. There was no evidence for retention time shifts or carryover from run to run.

The information contained in this publication is intended for research use only and is not to be followed as a diagnostic procedure. Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2007 Printed in the USA January 5, 2007 5989-6095EN
Analysis of Organochlorine Pesticides and PCB Congeners with the Agilent 6890 Micro-ECD
Application
Gas Chromatography June 1997
Authors
Isabelle Chanel Agilent Technologies, Inc. European Marketing Group 76337 Waldbron Germany Imogene Chang Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA
Introduction
The electron capture detector (ECD) is the detector of choice in many Contract Laboratory Programs (CLP)1 and EPA methods for pesticide analysis because of its sensitivity and selectivity for halogenated compounds. However, there are drawbacks to the ECD design. The ECD is inherently nonlinear2, with a limited linear range. The limited linear range means that dilution and reanalysis are frequently required for samples that are outside the calibration range. Also, the typical ECD is designed to be compatible with both packed and capillary columns. This results in a flow cell that is larger than that required for capillary columns alone, which reduces detector sensitivity.
To address these problems, a new ECD was developed for the 6890 Series gas chromatograph (GC). The 6890 Micro-ECD has a smaller flow cell optimized for capillary columns and was redesigned to improve the linear operating range. This application note examines the linearity, reproducibility, and limit of detection of the new ECD with mixtures of polychlorobiphenyl (PCB) congeners and organochlorine pesticides (OCPs).
Abstract
A new electron capture detector (ECD) for the Agilent 6890 Series gas chromatograph (GC) was used to analyze polychlorinated biphenyl congeners and organochlorine pesticides. The linearity of the 6890 Micro-ECD in the calibration range of 2 to 400 ppb was evaluated. The micro-ECD easily meets the linearity requirements of U.S. EPA contract laboratory programs for pesticides. Its limit of detection for these compounds goes down to less than 50 ppt. The micro-ECD also exhibits good reproducibility.
Experimental
All experiments were performed on an 6890 Series GC with electronic pneumatics control (EPC) and the 6890 Micro-ECD. Table 1 shows the experimental conditions for PCB congeners and OCPs.
Table 1.
Experimental Conditions for PCB Congener and OCP Analysis.

PCB Congener Analysis 80 C (2 min); 30 C/min to 200 C; 10 C/min to 320 C (5 min). Split/splitless; 300 C Helium, 16.8 psi (80 C); 1.3-mL/min constant flow Agilent 7673, 10- L syringe, 1- L splitless injection 30-m, 250- m id, 0.25- m film HP-5MS (part no. 19091S-433) 330 C; makeup gas: nitrogen, constant column and makeup flow OCP Analysis 80 C (2 min); 25 C/min to 190 C; 5 C/min to 280 C; 25 C/min to 300 C (2 min). Split/splitless; 250 C Helium, 23.9 psi (80 C); 2.2-mL/min constant flow 7673, 10- L syringe, 1- L splitless injection 30-m, 250- m id, 0.25- m film HP-5MS (part no. 19091S-433) 330 C; makeup gas: nitrogen, constant column and makeup flow
System Conditions Oven
Inlet Carrier Sampler Column
Key Words
Organochlorine pesticides, PCB congeners, 6890 GC, micro-ECD; pesticide analysis, ECD.
Detector
The solutions were prepared by making appropriate dilutions of a stock solution with isooctane. For PCB congeners, the stock solution was an EPA PCB congener calibration check solution (from Ultra Scientific Company, part number RPC-EPA-1). For OCPs, the solution was an OCP calibration check solution (part number 8500-5876).

Linearity and Response Factors
A series of dilutions of the PCB mixture from 2 ppb to 200 ppb and of the OCP mixture from 2 ppb to 400 ppb was injected into the 6890 Micro-ECD system. The linearity was determined by calculating the correlation coefficient from the resulting calibration curve. Figures 1 and 2 present typical chromatograms of OCPs and PCBs at 20 or 40 ppb and 50 ppb, respectively. Figure 3 is a calibration curve of decachlorobiphenyl, typical of other PCB congeners. Figure 4 shows the calibration curve of 4, 4 DDE, typical of OCPs. The correlation coefficient, Figure 1. Typical chromatogram of OCPs at 20 or 40 ppb. See table 1 for conditions. See table 5 for peak identification.
Figure 2. Typical chromatogram of PCB congeners at 50 ppb. See table 1 for conditions. See table 4 for peak identification.
Figure 3. Typical linearity of PCB congener analysis: decachlorobiphenyl from 2-200 ppb.
Figure 4. Typical linearity of OCP analysis: 4,4 DDE from 4 to 400 ppb.
average response factors, and percent relative standard deviation (%RSD) for the response factors for each analyte are shown in tables 2 and 3. All correlation coefficients were at least 0.9996. In these experiments, the 6890 Micro-ECD is linear over this range. The typical range required by CLP methods is 5-80 ppb1, so the 6890 Micro-ECD exceeds the range by almost twofold. In addition, the CLP method requires the percent RSD of the response factors for most components to be less than 20 percent for a three-point calibration curve (5 to 80 ppb). As shown in tables 2 and 3, the percent RSD of the response factors ranged from 0.55 percent to 12. 5 percent for the PCB congeners and from 2.8 percent to 10 percent for the OCPs over a concentration range of two orders of magnitude (2 to 400 ppb). Furthermore, the average response factor of each analyte was so consistent and reproducible that the internal standard technique can be used to quantitate all OCPs and PCB congeners.
Table 2.
PCB Congener Analysis: Linearity of the 6890 Micro-ECD 2 ppb to 200 ppb. See table 1 for conditions.
Peak
Name
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
2,4-Dichlorobiphenyl 2,2,5-Trichlorobiphenyl 2,4,4-Trichlorobiphenyl 2,2,5,5-Tetrachlorobiphenyl 2,2,3,5-Tetrachlorobiphenyl 2,3,4,4-Tetrachlorobiphenyl 2,2,4,5,5-Pentachlorobiphenyl 3,3,4,4-Tetrachlorobiphenyl 2,3,4,4,5-Pentachlorobiphenyl 2,2,4,4,5,5-Hexachlorobiphenyl 2,3,3,4,4-Pentachlorobiphenyl 2,2,3,4,4,5-Hexachlorobiphenyl 3,3,4,4,5-Pentachlorobiphenyl 2,2,3,4,5,5,6-Heptachlorobiphenyl 2,2,3,3,4,4-Hexachlorobiphenyl 2,2,3,4,4,5,5-Heptachlorobiphenyl 2,2,3,3,4,4,5-Heptachlorobiphenyl 2,2,3,3,4,4,5,6-Octachlorobiphenyl 2,2,3,3,4,4,5,5,6-Nonachlorobiphenyl Decachlorobiphenyl
Average Response Factor 2e-2 2e-2 8.5e-3 1.3e-2 1e-2 8e-3 9e-3 1.2e-2 8e-3 8e-3 6e-3 6.5e-3 9e-3 8e-3 5.6e-3 5.8e-3 5.8e-3 6e-3 8e-3 1e-2
%RSD of Response Factor 12.5 11.1 7.5 10.2 9.4 6.7 8.8 12.6 5.5 8.1 1.9 3.8 6.5 5.7 1.8 1.0 0.57 0.78 3.1 9.5
Correlation (%) 99.97 99.97 99.99 99.97 99.98 99.99 99.98 99.97 99.99 99.98 99.99 99.99 99.99 99.99 99.99 99.99 99.99 99.99 99.96 99.98
Table 3.
OCP Analysis: Linearity of the 6890 Micro-ECD 2 or 4 ppb to 200 or 400 ppb. See table 1 for conditions.
Name Average Response Factor 4.2e-3 1.1e-2 6.4e-3 4.7e-3 4.7e-3 6.6e-3 5e-3 5e-3 2.9e-3 4.5e-3 5.1e-3 4.7e-3 3.7e-3 % RSD of Response Factor 5.3 7.1 4.7 9.5 5.4 6.6 4.3 2.8 4.4 5.9 5.3 9.0 9.9 Correlation (%) 99.97 99.99 99.99 99.97 99.99 99.99 99.98 99.99 99.98 99.94 99.97 99.89 99.96
Peak
1 2 3 4 5 6 7 8 9 10 11 12 13
2,4,5,6-Tetra-m-xylene beta-BHC delta-BHC Aldrin Heptachlor epoxide gamma-Chlordane alpha-Chlordane 4,4' DDE Endosulfan II Endrin aldehyde Endosulfan sulfate Endrin ketone Decachlorobiphenyl
Reproducibility
The reproducibility of the 6890 MicroECD was established by analyzing each mixture using identical conditions five times. Each analyte in the PCB congener mixture was injected at a concentration of 50 ppb, and the analytes in the OCP mixture were 20 or 40 ppb. The results are shown in tables 4 and 5. The highest %RSD for any analyte is 3.69 percent for aldrin, which is well below the CLP maximum allowable RSD of 15 percent.1
Table 4.
PCB Congener Analysis: Reproducibility of the 6890 Micro-ECD 50 ppb; N=5. See table 1 for conditions.
Name 2,4-Dichlorobiphenyl 2,2,5-Trichlorobiphenyl 2,4,4-Trichlorobiphenyl 2,2,5,5-Tetrachlorobiphenyl 2,2,3,5-Tetrachlorobiphenyl 2,3,4,4-Tetrachlorobiphenyl 2,2,4,5,5-Pentachlorobiphenyl 3,3,4,4-Tetrachlorobiphenyl 2,3,4,4,5-Pentachlorobiphenyl 2,2,4,4,5,5-Hexachlorobiphenyl 2,3,3,4,4-Pentachlorobiphenyl 2,2,3,4,4,5-Hexachlorobiphenyl 3,3,4,4,5-Pentachlorobiphenyl 2,2,3,4,5,5,6-Heptachlorobiphenyl 2,2,3,3,4,4-Hexachlorobiphenyl 2,2,3,4,4,5,5-Heptachlorobiphenyl 2,2,3,3,4,4,5-Heptachlorobiphenyl 2,2,3,3,4,4,5,6-Octachlorobiphenyl 2,2,3,3,4,4,5,5,6-Nonachlorobiphenyl Decachlorobiphenyl Average Area 2229 2547 5687 3721 4941 5943 5089 3822 6203 6189 8375 7538 5092 6224 8921 8527 8625 8338 6097 4622 RSD (%) 1.26 1.29 1.41 1.43 1.46 1.40 1.47 1.72 1.62 1.44 1.68 1.56 2.02 1.69 1.67 1.82 1.91 2.13 2.55 2.85
Peak 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Table 5.
OCP Analysis: Reproducibility of the 6890 Micro-ECD; N=5. See table 1 for conditions.
Name 2,4,5,6-Tetra-m-xylene beta-BHC delta-BHC Aldrin Heptachlor epoxide gamma-Chlordane alpha-Chlordane 4,4' DDE Endosulfan II Endrin aldehyde Endosulfan sulfate Endrin ketone Decachlorobiphenyl Concentration (ppb) 20 20 20 20 20 20 20 40 40 40 40 40 40 Average Area 4785 1802 3251 402 4316 2958 4219 4103 7176 4719 4040 4386 5369 RSD (%) 0.7 0.81 1.50 3.69 1.58 1.23 1.06 1.76 1.27 0.85 3.04 2.52 0.85
Peak 1 2 3 4 5 6 7 8 9 10 11 12 13
Detection Limit
To establish the lower limit of detection for the 6890 Micro-ECD with PCBs and OCPs, 1-L injections were made at gradually decreasing concentrations. Figures 5 and 6 show chromatograms with analyte concentrations of 50 to 100 ppt. All the analyte peaks for both the PCB congener and OCP mixtures are still easy to quantitate, and in fact smaller concentrations can be reliably analyzed. Aldrin, which has the lowest response of the OCPs, still exhibits an adequate signal-to-noise ratio at the 50 ppt level under these analysis conditions.
Figure 5. PCB congener mixture at 50 ppt each. See table 1 for conditions. See table 4 for peak identification.
Conclusion
The Agilent 6890 Micro-ECD response was linear over the concentration range of 2 to 200 ppb, produced reproducible results, and exhibited excellent sensitivity for mixtures of PCB congeners and OCPs.
References
1. U.S. EPA Contract Laboratory Program, Statement of Work for Organic Analysis, OLM03.1, August 1994. Figure 6. OCP Mixture at 50 to 100 ppt. See table 1 for conditions. See table 5 for peak identification.
2. R. Buffington and M.K. Wilson, eds., Detectors for Gas Chromatography, Agilent Technologies, Inc., Part Number 5958-9433, 1987.
Contaminants Toxins
Determination of Aflatoxins in Food by LC/MS/MS
Application
Food Safety
Authors
Masahiko Takino Agilent Technologies 9-1 Takakura-Cho Hachiouji-Shi, Tokyo Japan Toshitsugu Tanaka Kobe Institute of Health Department of Food Chemistry 4-6 Minatojima-nakamachi Chuo-ku, Kobe Japan
Abstract
A sensitive and selective analytical method for the determination of aflatoxin G1, G2, B1, and B2 residues in cereals using the Agilent G6410AA LC/MS Triple Quadrupole Mass Spectrometer was developed. This method uses simple sample preparation methods followed by LC/MS/ MS. The limits of detection for all aflatoxins were less than 1 ng/mL in cereals.
disease from AF exposure and to advance worldwide surveillance of food. Analysis of AFs in food products is routinely performed by thin-layer chromatography (TLC) and liquid chromatography (LC) with fluorescence detection (FD) in combination with both precolumn derivatization and postcolumn derivatization. The LC/FD technology is often used due to the high selectivity and sensitivity of these methods. Furthermore, hyphenated techniques such as LC coupled to mass spectrometric (MS) detection have been developed and applied in residual analysis of foods. The high selectivity and sensitivity of MS detection methods associated with the resolution of LC provide decisive advantages to perform qualitative as well as quantitative analysis of a wide range of molecules at trace levels. Several papers describing different kinds of MS methods for the analysis of AFs have been published [2-4.]
Experimental
Sample Preparation The samples analyzed (peanuts, corn, nutmeg, and red pepper) were obtained from a local market and did not include any AFs. The extraction and cleanup steps for AFs were carried out according to validated methods reported by Tanaka [5]. Briefly, 20 g fine ground sample was poured into a 200-mL Erlenmeyer flask, followed by adding 40 mL acetonitrile-water (9:1, v/v) for corn and cereals. After shaking for 30 min, the mixed solution was centrifuged for 5 min at 1,650 g. The supernatant obtained was filtered through a glass microfiber GF/B grade filter (Whatman Interna-
Introduction
Aflatoxins (AFs) belong to a closely related group of secondary fungal metabolites. These mycotoxins are severely toxic metabolites produced mainly by Aspergillus flavus and A. parasiticus, and exposure to them can cause cancer in humans and livestock [1]. Based on epidemiological evidence, AFs have been classified as human liver carcinogens by the World Health Organization and by the U.S. Environmental Protection Agency. Thus, accurate determination of AFs is required to avoid human
tional Ltd, Maidstone, UK). A 5-mL portion of the filtrate was applied to a MultiSep number 228 cartridge column for the cleanup. After passing through at a flow rate of 1 mL/min, 2 mL of the first eluate was collected. The eluate was evaporated to dryness at 40 C under a gentle stream of nitrogen. The residue was reconstituted in 1 mL methanol-water (4:6 v/v) containing 10 mM ammonium acetate. Standard Preparation Each of the standard reagents, aflatoxin G2 (AFG2), aflatoxin G1 (AFG1), aflatoxin B2 (AFB2) and aflatoxin B1 (AFB1), was dissolved in acetonitrile at 1 mg/mL and was stored at 4 C in the dark until use. To prepare the working standard for LC/MS analysis, each AF stock solution was equally pipetted and transferred to a vial, and it was then diluted with the mobile phase. The final concentration of each AF was 1 ng/mL. Chemicals The standards AFG2, AFG1, AFB2, and AFB1 were obtained from Sigma Aldrich Japan (Tokyo, Japan). The purity of these compounds was greater than 99%. Ammonium acetate, toluene, HPLC-grade acetonitrile, and HPLC-grade methanol were obtained from Wako Chemical (Osaka, Japan). Water was purified in-house with a Milli-Q system (Millipore, Tokyo, Japan). The cartridge column of MultiSep number 228 was purchased from Showa Denko (Kanagawa, Japan). LC/MS Instrument The LC/MS/MS system used in this work consists of an Agilent 1200 Series vacuum degasser, binary
pump, well-plate autosampler, thermostatted column compartment, the Agilent G6410 Triple Quadrupole Mass Spectrometer with an electrospray ionization (ESI) source. The objective of the method development was to obtain a fast and sensitive analysis for quantifying AFs in foods. For chromatographic resolution and sensitivity, different solvents and columns were optimized. It was found that a simple solvent system using water, methanol, ammonium acetate, and a 1.8-m particle size C18 column worked very well.
LC Conditions Instrument: Column: Column temp: Mobile phase:
Flow rate: Injection volume: MS Conditions Instrument: Source: Drying gas flow: Nebulizer: Drying gas temp: Vcap: Scan: Fragmentor: MRM ions: Collision energy:
Agilent 1200 HPLC ZORBAX Extend C18, 100 mm 2.1 mm, 1.8 m (p/n 728700-902) 40 C A = 10 mM ammonium acetate in water B= Methanol 40% A/60% B 0.2 mL/min 5 L Agilent 6410 LC /MS Triple Quadrupole Positive ESI 10 L/min 50 psig 350 C 4000 V m/z 100 550 Variable 100 V Shown in Table 1 Shown in Table 1
LC/MS/MS Method Quantitative analysis was carried out using MRM mode. The parameters for MRM transitions are shown in Table 1.
Table 1.
Data Acquisition Parameters of MRM Transitions for Each Aflatoxin RT (min) 5.21 6.61 8.44 10.89 Molecular weight 330 328 314 312 Precursor ion (m/z) 331 329 315 313 Product ion (m/z) 245 243 259 241 Collision energy (V) 30 30 30 30
No 1 2 3 4
Mycotoxins Aflatoxin G2 Aflatoxin G1 Aflatoxin B2 Aflatoxin B1

Optimization of MRM Transitions Determination of the optimal MRM transitions for each aflatoxin was carried out using single-MS fullscan mode followed by product ion scan mode using taflatoxin standard mixtures at 1 ug/mL. Mass spectra of these standard mixtures in full scan mode and product ion scan mode are shown in Figures 1 and 2. The mass spectrum of each aflatoxin by full-scan mode exhibited the protonated molecule [M+H]+ as the base peak ion. These ions were selected as precursor ions for MRM
mode. The optimum collision voltage is compound dependent. To establish the optimum collision voltage, this parameter was varied from 5 to 40 V using a step size of 5V. As shown in Figure 2, a distinct optimum in the intensity of the product ion of each AF was observed at 30 V. The product ions that indicated the highest intensity were m/z 245 (AFG2), 243 (AFG1), 259 (AFB2), and 241 (AFB1), respectively. On the basis of the above results, the collision voltage was set to 30 V. Table 1 shows the parameters of MRM mode of each aflatoxin.
10 5 2.4 2.0 Abundance
331(M+H) +
10 5 2.4 2.0 Abundance 1.6 1.2 0.8
315(M+H) +
(A)
(C)
1.6 1.2 0.8 353 0.4 313
0.4
287
337
10 5 1.8 1.4 Abundance
329(M+H) +
(B)
Abundance
10 5 1.8 1.4 1.0 0.6
313(M+H) +
(D)
1.0 0.6 0.2 100 140 180 220 m/z
243
311
351
0.2 340 100 140 180 220 m/z
241
260
285
335
300 340
260
300
Figure 1.
Mass spectra of four aflatoxins standard in single-MS full-scan mode at 1 g/mL (A): aflatoxin G2, (B): aflatoxin G1, (C): aflatoxin B2, and (D): aflatoxin B1.
10 3 4 Abundance
189
(A)
245
141
2
217 285 331(M+H) +
10 3 6 5.5 5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 10 3 9 8
(C)
259
Abundance
115 287 203 315(M+H) +
10 3 7
200
(B)
128 215 171 243
Abundance
(D)
128 185 213
241
Abundance
7 6 5 4 3
157.1
269 313(M+H) +
1 120 160 200 240
283
280
329(M+H) +
320
2 1 0
m/z
120
160
200
m/z
240
280
320
Figure 2.
Mass spectra of four aflatoxins standard in product ion scan mode at 1g/mL (A): aflatoxin G2, (B): aflatoxin G1, (C): aflatoxin B2, and (D): aflatoxin B1.
The MRM chromatogram of each aflatoxin at 0.1 ng/mL is shown in Figure 3. These show excellent signal-to-noise (S/N) ratios for all aflatoxins. The limit of detection (LOD) for each aflatoxin was determined using an S/N ratio of 3 with this MRM chromatogram and is shown in Table 2. To evaluate the linearity of the calibration curves, various concentrations of aflatoxin standard solutions ranging from 0.1 ng/mL to 100 ng/mL were analyzed. As shown in Figure 4 and Table 2, the linearity was very good for all aflatoxins with correlation coefficients (r2) greater than 0.999.
5000 4000 Abundance
1. Aflatoxin G 2
5000
3. Aflatoxin B2
4000 Abundance
3000 2000 1000
3000
2000
1000
6000 8000
2. Aflatoxin G1
Abundance 4000
4. Aflatoxin B1
Abundance
6000
4000
2000 2000
6 7 8 5 Retention time (min)
10
11
10
11
Figure 3.
MRM chromatograms of four aflatoxin standards at 0.1 ng/mL in MRM mode.
80000 40000
1. Aflatoxin G 2
60000
3. Aflatoxin B2
30000 Abundance Abundance 40000
20000
10000
20000
100000
50
100 90000
50
100
80000
2. Aflatoxin G1
Abundance
70000 50000 30000 10000
4. Aflatoxin B1
Abundance
60000
40000
20000
50 Conc. (ng/mL)
100
50 Conc. (ng/mL)
100
Figure 4. Calibration curves of four aflatoxins ranged from 0.1 ng/mL to 100 ng/mL. 5
Table 2. No 1 2 3 4
Linearity and LODs of Four Aflatoxins Mycotoxins Aflatoxin G2 Aflatoxin G1 Aflatoxin B2 Aflatoxin B1 r2 0.9999 0.9992 0.9999 0.9993 LOD (ng/mL) 0.025 0.020 0.025 0.020
The matrix effect of this method was investigated by using cereal and corn extracts spiked with mycotoxin standards at 0.2 ng/mL. Typical MRM chromatograms of cereal and corn extract are shown in Figures 5 and 6, respectively. There were no additional peaks from sample matrix in either food when compared with the mycotoxin standard mixture. These results indicate that MRM mode has very high selectivity.
10000 8000 Abundance
Aflatoxin G2
10000 8000 Abundance 6000 4000 2000
Aflatoxin B2
6000 4000 2000
16000
Aflatoxin G1
Abundance Abundance 12000
12000 10000 8000 6000 4000 2000
Aflatoxin B1
8000
4000
10
11
10
11
Figure 5.
MRM of four aflatoxins in cereal extract spiked at 0.2 ng/g.
10000 Abundance 8000 6000 4000 2000
Aflatoxin G2
10000 8000 6000 4000 2000
Aflatoxin B2
16000
Abundance
Aflatoxin G1
Abundance Abundance 12000
12000 10000 8000 6000 4000 2000
Aflatoxin B1
8000
4000
10
11
10
11
Figure 6.
MRM of four aflatoxins in corn extract spiked at 0.2 ng/g.
Furthermore, the change on the peak intensity of each aflatoxin by the sample matrix was investigated by comparison with the peak intensity of aflatoxin standards. As these results show in Table 3, the relative intensity of each pesticide ranged from 88 to 96%. Thus, matrix effects such as ion suppression may be insignificant and it is possible to use external standards instead of matrix-matched standards.
Table 3. No 1 2 3 4 Relative Intensity of Each Aflatoxin in Sample Extracts Mycotoxins Relative intensity (%) Cereal Corn Aflatoxin G2 88 91 Aflatoxin G1 92 94 Aflatoxin B2 93 96 Aflatoxin B1 97 95
References
1. K. K. Sinha and D. Bhatnagar, 1998, Mycotoxins in Agriculture and Food Safety, 1998 (New York: Marcel Dekker) 2. W. J. Hurst, R. A. Martin, and C. H. Vestal, 1991, The Use of HPLC/Thermospray MS for the Confirmation of Aflatoxons in Peanuts, J. Liq. Chromatogr., 14, 2541-2540. 3. A. Cappiello, G. Famiglini, and B. Tirillini, 1995, Determination of Aflatoxins in Peanut Meal by LC/MS with a Particle Beam Interface, Chromatographia, 40, 411-416. 4. M. Vahl and K. Jorgensen, 1998, Determination of Aflatoxins in Food Using LC/MS/MS, Z Lebensm Unters Forsch A., 206, 243-245. 5. T. Tanaka, A. Yoneda, Y. Sugiura, S. Inoue, M. Takino, A. Tanaka, A. Shinoda, H. Suzuki, H. Akiyama, and M. Toyoda, 2002, An Application of Liquid Chromatography and Mass Spectrometry for Determination of Aflatoxins, Mycotoxins, 52, 107-113.
Conclusions
The multi-aflatoxin method by LC/MS/MS described here was suitable for the determination of four aflatoxins in cereal and corn extract due to its high sensitivity and high selectivity. Another advantage of this method is that ion suppression was not observed for all food samples studied. Thus, it may eliminate the need for matrixmatched standards, which makes analysis more tedious for samples from different origins.

For more information on our products and services, visit our Web site at www.agilent.com/chem. For more details concerning this application, please contact masahiko_takino@agilent.com.
Rapid Analysis of Crude Fungal Extracts for Secondary Metabolites by LC/TOF-MS A New Approach to Fungal Characterization Application
Food
Authors
enyuva Hamide Z. S Ankara Test and Analysis Laboratory Scientific and Technological Research Council of Turkey Ankara 06330 Turkey
John Gilbert Central Science Laboratory Sand Hutton York YO41 1LZ UK
Abstract
A novel approach to studying the production of secondary metabolites by fungi using LC/TOF-MS has been developed. Fungi grown on culture media are solvent-extracted and directly analyzed by LC/TOF-MS. Searching against a database of 465 secondary metabolites, mycotoxins and other compounds of interest can be readily identified. The methodology was validated by spiking culture media with 20 mycotoxin standards and identifying these toxins in the crude solvent extracts. Subsequently, using seven different fungi from culture collections, after culturing for 7 to 14 days in three different media, anticipated metabolites were readily identified.
Traditionally, this fungal characterization has been based on classical mycology, involving culturing the fungi on different media and then classifying depending on morphological and growth behavior characteristics. However, such classification can be time-consuming and is somewhat subjective, being dependant on the skill and experience of the mycologist. Additionally, such typing of fungi only provides anecdotal evidence about actual profiles of secondary metabolites, as it is based on previously observed secondary metabolism of particular fungal species. This empirical approach is further confounded by the fact that fungi of the same species can be both toxigenic and nontoxigenic; that is, some readily produce mycotoxins, but some otherwise indistinguishable fungi of the same species are genetically incapable of toxin production. Classification of fungal species alone therefore provides no real insight into mycotoxin production. In the past, direct analysis of fungal culture media for the presence of mycotoxins has of necessity involved target analysis with the inevitable assumption as to which toxins should be sought. However, LC/TOF-MS offers new possibilities for studying the behavior of fungi with regard to toxin production. Providing that efficient extraction from the medium of toxins with widely differing polarity can be demonstrated, the specificity of TOF-MS means that any further sample clean-up is not necessary. Furthermore, targeted analysis is also unnecessary as the instrument can provide accurate mass measurement of molecular ions of any components detected in an LC run, and these can be identified by searching a database of exact masses of relevant secondary metabolites.
Introduction
From a food safety perspective there is a need to characterize molds (fungi) isolated from agricultural products, as these may represent a potential source of mycotoxin contamination in food.
In this note we describe suitable conditions for extraction of secondary metabolites from cultured fungi and LC/TOF-MS conditions for subsequent analysis. The methodology has been validated by spiking aflatoxins, ochratoxin A, trichothecenes, zearalenone, and fumonisins into various growth media, and demonstrating good recovery from the media at low levels and subsequent identification by searching against a database of 465 secondary metabolites. The methodology has been applied to one Penicillium species and six Aspergillus species, which were obtained from a culture collection, and their secondary metabolites have been compared with the anticipated toxin profiles.
Table 1.
LC/MS-TOF Operational Conditions in ESI+ Ion Mode 3000 V 40 psig 10 L/min 300 C 150 V 60 V 250 V 37.5 V 1001000 121.050873; 922.009798
Parameter Capillary voltage Nebulizer pressure Drying gas Gas temperature Fragmentor voltage Skimmer voltage OCT* RF OCT* DC Mass range (m/z) Reference masses *Octapole
Experimental
All analytical work was performed using an Agilent 6210 TOF-MS coupled to an Agilent 1200 Series HPLC. The separation of mycotoxins and other fungal metabolites was also carried out using an HPLC system (consisting of vacuum degasser, autosampler with thermostat, binary pump, and DAD system) equipped with a reversed-phase C18 column (ZORBAX Eclipse XDB 100 2.1 mm, 1.8 m). The TOF-MS was equipped with a dualnebulizer electrospray source, allowing continuous introduction of reference mass compounds. The instrument was scanned from m/z 100 to 1,000 for all samples at a scan rate of 1 cycle/sec in 9,429 transient/scan. This mass range enabled the inclusion of two reference mass compounds, which produced ions at m/z 121.0508 and 922.0097. The injected sample volume was 5 L. The HPLC analysis used a mobile phase of acetonitrile and 2 mM ammonium acetate in an aqueous solution of 1% formic acid at a flow rate of 0.3 mL/min. The gradient elution started with 15% acetonitrile and reached 100% acetonitrile in 20 min. The column was washed with 100% acetonitrile for 5 min. and equilibrated for 5 min between chromatographic runs. UV spectra were obtained using diode array detection scanning every 0.4 sec from 200 to 700 nm with a resolution of 4 nm. The optimum TOF-MS conditions are given in Table 1. The data recorded were processed with Analyst-QS software with accurate mass application. The database of 465 mycotoxins and other fungal metabolites was created in Excel from reference sources [1,2], which were easily adapted to use in a search capacity using Agilent software.
Fungal Extraction Well-characterized isolates of A. paraciticus (NRRL 2999), were obtained from the USDA culture collection and isolates of A. flavus, (200198), A. ochraceus (200700), A. oryzae (200828), A. niger (200807), A. fumigatus (200418), and P. citrinum (501862) were obtained from the TBITAK Mamara Research Center culture collection. Fungi were inoculated onto malt extract agar (MEA), potato dextrose agar (PDA), and yeast extract sucrose agar (YES) in petri dishes. After allowing the fungi to grow for 7 to 14 days at 25 C, typical prolific growth of fungal colonies was observed on the surface of the media. Samples of fungal hyphae, together with underlying culture media, were taken by vertically cutting two 6-mm diameter plugs using a cork borer. The plugs were transferred to 5-mL disposable screw-cap bottles. Extraction conditions were modified from previous published methods [3,4]. One of the plugs was extracted twice with 2 mL ethyl acetate with 1% formic acid and then 2 mL isopropanol. The second plug was extracted twice with 2 mL ethyl acetate with 1% formic acid and then 2 mL acetonitrile, followed by 1 min vortexing and 30 min total ultrasonication. The extracts were filtered and evaporated gently under a nitrogen stream. The residues in both cases were dissolved in 1 mL methanol, ultrasonicated for 10 min and passed through a 0.2-L disposable filter prior to HPLC analysis.

Optimization of LC/TOF-MS Conditions The most important instrumental parameters, which were capillary voltage, nebulizer pressure, drying gas, gas temperature, and skimmer voltage, were initially optimized by autotune to achieve
maximum sensitivity. However, the fragmentor voltage also needed to be optimized to provide maximum structural information, which sometimes required a compromise. Optimization was carried out by varying the fragmentor voltage in the range of 55 to 250 V without changing any other conditions. The fragmentor voltage that provided minimum fragmentation was found at 150 V. To validate the whole procedure, 20 commercially available standards (aflatoxins B1, B2, G1, and G2; aflatoxin M1; ochratoxin A; zearalenone; 4-deoxynivalenol; 3-acetyldeoxynivalenol; 15-acetyldeoxynivalenol; diacetoxyscirpenol; fusarenone X; neosolaniol; fumonisins B1, B2, and B3; nivalenol; HT-2 toxin; T2 toxin; and kojic acid) and internal standard (benzophenone) were mixed together. Using positive electrospray, the accurate masses of protonated molecule ions, retention times, and UV spectra were obtained in each case. Construction of Database of Accurate Masses of Fungal Metabolites An Excel spreadsheet was constructed containing the exact mass data for each of the 465 mycotoxins and fungal metabolites, together with their empirical formulas [1,2]. Theoretical monoisotopic exact masses of the compounds were calculated based on their molecular formula using an Excel spreadsheet (called Formula DB Generator and provided with the Agilent TOF) and put into csv (comma-separated values) format for use by the Agilent TOF automated data analysis software. The csv file is searched automatically by the LC/TOFMS instrument at the completion of the sample run and a report is generated on compounds that were found in the database. The creation of the data analysis method is done using a data analysis editor. The editor allows selection of adducts (for example, in positive ion H+, NH4+, Na+, etc.) and neutral losses to be searched automatically, as well as mass accuracy and retention time tolerances, report options, and other search and detection criteria. Retention times are not required but if they are known add a degree of confidence to the identification. We use samples of various growth media that had been spiked with the standard mixture of 20 mycotoxin standards to determine retention times. The standards were injected 10 times to establish the repeatability of those retention times. The criteria used for identification were a fit for the accurate mass of the M+1 ion to a mass tolerance of 5 ppm, a retention time match to 0.2 min (if standards available), a minimum peak height
count, which is called the compound threshold of 1,000 counts (or a signal-to-noise ratio of ~10:1 or 0.06% relative volume), and, if present, good correspondence (to 5 ppm) with predicted adducts and neutral fragment losses. Method Validation by Spiking and Analysis Based on the above detection criteria, all 20 standards were correctly identified when spiked at 25 to 100 g/kg into growth media, and analyzed as described above. Utilization of the Method to Determine Metabolite Production from Well-Characterized Fungi Rather than simply looking at theoretical situations with spiked growth media, the above technique was applied to the real situation of well-characterized fungi being cultured on various media. One Penicillim species and six Aspergillus species were grown on three different media. Using the simple solvent extraction described above, the crude extracts were directly analyzed by LC/TOFMS. By way of illustration, Figure 1 shows the total ion chromatogram for an A. flavus extract indicating about 20 components detected. The peak eluting at 8.9 min on database searching was found to have an accurate mass of m/z 313.0712. Based on the M+H+ ion this corresponded to aflatoxin B1 with a 0.2 ppm mass match as compared to the database exact mass for aflatoxin B1. The software uses a molecular feature (MFE) algorithm that removes all ions that are not real peaks and groups the real ions into molecular features. Those molecular features can be characterized by their relationship with each other and adducts, dimers, trimers, etc., and their isotopes (depicted as +1, +2, etc.) are deduced. The molecular features and accurate mass measurement of these species for the peak at a retention time of 8.9 min identified as aflatoxin B1 are summarized in Table 2. Selecting a molecular feature, the software will calculate possible empirical formulas and score the isotopes for the fit to the proposed formula; this is also shown in Table 2. The formula with the score of 100 is that of aflatoxin B1. This formula then can be automatically translated to a Web connection search with NIST, ChemIndex, and Medline. The search results in NIST indicated the formula and structure of aflatoxin B1 as illustrated in Figure 2. In addition to the identification of aflatoxin B1 as a secondary metabolite from A. flavus, this fungi was also found to produce aflatoxin B2, aflatoxin B3, and aflatoxin G1, which are consistent with
3
TIC of A. flavus
m/z 313.0712 C17H13O6
(M+H)+ 8.903 min 313.0712
Figure 1.
Analysis of an extract from A. flavus by LC/TOF-MS illustrating: (a) Total ion chromatogram (TIC) with * peak corresponding to aflatoxin B1, (b) Extracted ion chromatogram from m/z 313.058 to 313.093 for aflatoxin B1, (c) Full-scan spectrum showing accurate mass with 0.2 ppm error for M+1 ion for aflatoxin B1.
Table 2.
Typical Clusters Seen in ESI+ LC/MS-TOF on the Peak Retention Time of 8.90 min, m/z 313.0706
MFE
Feature #27 (RT = 8.903) Species RT M M+H M+H+1 M+H+2 M+H+3 M+Na M+Na+1 M+Na+2 2M+H 2M+Na 2M+Na+1 2M+Na+2 2M+Na+3 2M+Na+4 8.903 8.903 8.903 8.902 8.906 8.904 8.904 8.916 8.906 8.902 8.902 8.900 8.899 8.897 m/z 313.0706 314.0744 315.0766 316.0795 335.0529 336.0563 337.0593 625.1357 647.1164 648.1202 649.1228 650.1256 651.1290 Mass 312.0633 312.0633 Abundance 5541933 4035186 622147 74349 7943 86580 15898 1848 741 226965 65639 14941 2677 257 Width 0.088 0.09 0.088 0.09 0.085 0.094 0.097 0.091 0.049 0.061 0.063 0.065 0.058 0.048
312.0637
312.0642 312.0636
MFE
Compositions chemical formula C17H12O6 C18H8N4O2 C14H4N10 C9H16N2O8S C13H8N6O4 dm(Da) 0.0001 0.0014 0.0013 0.0006 0.0026 dm(ppm) 0.2 4.5 4.2 1.9 8.4 dm(ppm) 0.2 4.5 4.2 1.9 8.4 DBE 12 17 18 3 13 Score 100 77 68 58 55
National Institude of St andards and Technology
Standard Reference Data Program
Data Gateway
Online Databasis
Chemistry WebBook
Aflatoxin B1
Formula: C17H12O6 Molecular weight: 312.27 IUPAC International Chemical Identifer: o InChI=1/C17H12O6/c1-20-10-6-11-14(8-4-5-21-17(8)22-11)15-13(10)7-2-3-9(18)12(7) 16(19)23-15/h4-6,8,17H,2-3H2,1H3
O O O O O
CAS Registry Number: 1162-65-8

O
Chemical structure:
Figure 2. Database search result for emprical formula using NIST (Medline and ChemIndex results were the same but are not given here). Note molecular weights should not be searched in these databases as they are often the average molecular weight and not the monoisotopic weight. 5
known metabolic behavior. In Table 3 the screening results from the database search with a 5 ppm tolerance are shown with the accurate masses of some other peaks, which corresponded to known compounds. Kojic acid and methoxysterigmatocystin, which are a good match, are both wellknown fungal metabolites that might be expected to be found from A. flavus. A good match was also found for cinnamic acid, which is not known as a metabolite. When this new approach was applied in a preliminary study of a total of seven different fungi obtained from culture collections and grown on
three different media, the results shown in Table 4 were obtained. In most cases the predicted metabolites were found, which gives good confidence in the methodology. Some of these initial results showed that predicted mycotoxins were not detected and unexpected metabolites were found. The possibility of a misidentified culture exists or that metabolites not previously reported were detected. While this demonstrates the power of the approach, these results do need to be followed up with more in-depth study. Future Prospects The use of accurate mass LC/TOF-MS combined
Table 3.
Results of Automated Mycotoxin Database Search for A. flavus Extract (Extraction compound list is sorted in ascending order of retention time within 5 ppm error. Benzophenone was used as an internal standard.)
Mass Value = 142.03 Formula C6H604 Mass Value = 148.05 Formula C9H802 Mass Value = 328.06 Formula C17H12O7 Mass Value = 354.07 Formula C19H1407 Mass Value =312.06 Formula C17H1206 Mass Value =312.06 Formula C17H14O6 Mass Value = 338.08 Formula C19H1406 Mass Value = 182.07 Formula C13H100
Compound Kojic acid Compound Cinnamic acid Compound Aflatoxin G1 Compound 5-Methoxysterigmatocystin Compound Aflatoxin B1 Compound Aflatoxin B2 Compound Methylsterigmatocystin Compound Benzophenone
Mass 142.03 Mass 148.05 Mass 328.06 Mass 354.07 Mass 312.06 Mass 314.08 Mass 338.08 Mass 182.07
Error (mDa) 0.10 Error (mDa) 0.08 Error (mDa) 1.01 Error (mDa) 0.99 Error (mDa) 0.05 Error (mDa) 0.06 Error (mDa) 0.16 Error (mDa) 0.73
*Error (ppm) 0.7 Error (ppm) 0.5 Error (ppm) 1.4 *Error (ppm) 2.8 *Error (ppm) 0.2 *Error (ppm) 0.2 *Error (ppm) 0.5 *Error (ppm) 4.0
Ret. Time Error Ret. Time Error Ret. Time Error 0.05 Ret. Time Error Ret. Time Error 0.11 Ret. Time Error 0.06 Ret. Time Error Ret. Time Error 0.15
Table 4.
A Comparison of Detected and Predicted Metabolites from Culture Collection Fungi Grown in MEA, YES, and PDA Medium Fungi
Metabolites AFB1 AFB2 AFB3 AFG1 AFG2 KA MST 5-MST OTA RO-A FU-B MA AA Nig Ter Cit
P. citrinum
A. flavus ` ` ` `
A. paraciticus ` `
A. niger
A. fumigatus
A. oryzae
A. ochraseus
` `
` ` `
` ` ` ` ` ` ` ` ` `
` - metabolites detected by LC/TOF-MS; Key: AFB1 KA MST 5-MST OTA RO-A Aflatoxin B1 etc. Kojic acid Methylsterigmatocystin 5-methoxysterigmatocystin Ochratoxin A Roquefortine A (isofumigaclavine A)
- metabolites predicted to be present
FU-B MA AA Nig Ter Cit
Fumigaclavine B Malformin (peptides) Aspergillic acid Nigragillin Terrein Citrinin
with database searching is a powerful example of a new, versatile identification technique that can be used in targeted analysis. In the area of fungal metabolites, the potential to screen fungi for a range of metabolites for which dedicated methods are not available has been demonstrated. This approach offers new possibilities for fungal typing based on metabolite production and rapid screening of agricultural products for mycotoxins of food safety interest. Where previously unknown metabolites are detected, although LC/TOF-MS can provide some insight, further work with a hybrid quadrupole time-of-flight LC/MS system (LC/QTOF-MS) will be required for structural elucidation.
Conclusions
A simple and rapid method has been developed using LC/TOF-MS to determine the profile of secondary metabolites produced by fungi under various culture conditions. The approach has been validated by spiking representative metabolites into solid cultures and demonstrating good recovery and identification by searching accurate masses against a metabolite database. Results for a range of well-characterized fungi from a culture collection showed that the anticipated toxins could be readily detected.
References
1. R. J. Cole and R. H. Cox. (1981) Handbook of Toxic Fungal Metabolites, Academic Press Inc. (New York). 2. J. C. Frisvad and U. Thrane (1987) Standardized High-Performance Liquid Chromatography of 182 Mycotoxins and Other Fungal Metabolites Based on Alkylphenone Retention Indices and UV-VIS Spectra (diode array detection), Journal of Chromatography A, 404:195214. 3. J. Smedsgaard (1997) Micro-Scale Extraction Procedure for Standardized Screening of Fungal Metabolite Production in Cultures, Journal of Chromatography A, 760: 264270. 4. J. Stroka, M. Petz, U. Joerissen, and E. Anklam, (1999) Investigation of Various Extractants for the Analysis of Aflatoxin B1 in Different Food and Feed Matrices, Food Additives & Contaminants, 16: 331338.

Rapid Determination of Hydroxymethylfurfural in Foods Using Liquid Chromatography-Mass Spectrometry Application
Food Industry
Authors
Vural Gkmen Food Engineering Department Hacettepe University, Beytepe Campus 06800 Beytepe Ankara, Turkey Hamide Z. Senyuva Ankara Test and Analysis Laboratory Scientific and Technical Research Council of Turkey Ankara 06330, Turkey
catalyzed dehydration of hexoses. Formation of HMF in foods is especially dependent on temperature and pH [1]. In recent years, the presence of HMF in foods has raised toxicological concerns: the compound and its similar derivatives were shown to have cytotoxic, genotoxic, and tumoral effects. However, further studies suggest that HMF does not pose a serious health risk, but the subject is still a matter of debate. Several HPLC techniques were reported for the determination of HMF in various foods. These techniques use UV detection because of the strong absorption of furfurals at approximately 280 to 285 nm. However, many compounds naturally present or formed in foods during processing may also absorb at this wavelength. Poor chromatographic resolution of these compounds may adversely affect the quantification of HMF during UV detection. A rapid and reliable liquid chromatography/mass spectrometry (LC/MS) method was developed for the determination of HMF in foods. The method entailed aqueous extraction of HMF, solid-phase extraction (SPE) cleanup and analysis by LC/MS. The separation was performed on a narrow-bore column to shorten the chromatographic run.
Abstract
Hydroxymethylfurfural is a product of food deterioration and is still under investigation for possible toxic effects. It can also be used to monitor food quality. A sensitive and selective LC/MS method for monitoring this compound is presented. The method can quantitatively determine hydroxymethylfurfural in food with a detection limit of 0.005 g/g. Sample preparation and analytical conditions are given.
Introduction
Hydroxymethylfurfural (HMF) is recognized as an indicator of quality deterioration in a wide range of foods. It is formed as an intermediate in the Maillard reaction and is also formed during acid-
Experimental
LC/MS experiments were performed using an Agilent 1100 series HPLC system consisting of a binary pump, an autosampler, and a temperaturecontrolled column oven, coupled to an Agilent 1100 MS detector equipped with atmospheric pressure chemical ionization (APCI) interface. Data acquisition was performed in selected ion monitoring (SIM) mode using the interface parameters: drying gas (N2, 100 psig) flow of 4 L/min, nebulizer pressure of 60 psig, drying gas temperaLC/MS Flow rate: Gradient: Mobile phase: Injection: MS conditions Ionization mode: Nebulizer pressure: Drying gas flow: Drying gas temperature: Vaporizer temperature: Skimmer: Capillary voltage: Fragmentor voltage: Dwell time: Positive APCI 60 psi 4 L/min 325 C 425 C 20 V 4kV 55 eV 439 ms 0.2 mL/min ZORBAX Bonus RP, 100 mm 2.1 mm, 3,5 m 0.01 mM acetic acid in 0.2% aqueous solution of formic acid 20 L out of 1000 L
II solutions were prepared by dissolving 15 g of potassium hexacyanoferrate and 30 g of zinc sulfate in 100 mL of water, respectively. A total of 100 L Carrez I and 100 L Carrez II solutions were added to the sample and the volume completed to 10 mL with 0.2 mM acetic acid. HMF was extracted by mixing the tube for 3 min using a vortex mixer. It was then centrifuged for 10 min at 5,000 rpm at 0 C. The clear supernatant was further cleaned up by using Oasis HLB SPE cartridge. Prior to use, the SPE cartridge was conditioned by passing 1 mL of methanol and equilibrated by passing 1 mL of water at a flow rate of approximately two drops per second using a plastic 2-mL syringe. The excess water was removed from the cartridge by passing 2 mL of air. One milliliter of aqueous extract was eluted through the preconditioned cartridge at a flow rate of approximately one drop per second using a plastic syringe and the eluate was discarded. The cartridge was washed by passing 0.5 mL of water. Then the cartridge was dried under a gentle stream of nitrogen. HMF was eluted from the cartridge by passing 0.5 mL of diethyl ether at a flow rate of approximately one drop per second using a plastic 2-mL syringe. The eluate was collected in a conical bottom glass test tube placed in a water bath at 40 C (Zymark Turbo Vap LV Evaporator) and evaporated to dryness under nitrogen at 3 psig. The remaining residue was immediately redissolved in 1 mL of water by mixing in a vortex mixer for 1 min. Twenty microliters of this test solution was injected onto the HPLC system.

Positive APCI-MS analysis of HMF showed both the precursor [M+1] ion and the compound-specific ion [C6H5O2] due to loss of water from the protonated molecule. See Figure 1. These characteristic
127.1 125.0 128.1
100
tures of 325 C, vaporizer temperature of 425 C, capillary voltage of 4 kV, corona current of 4 A, fragmentor voltage of 55 eV, and dwell time of 439 ms. Ions monitored for HMF were m/z 109 and m/z 127. The quantification was performed based on the signal response of the ion having m/z of 109. The chromatographic separations were performed on a ZORBAX Bonus RP Narrow Bore column (2.1 mm 100 mm, 3.5 m) using the isocratic mixture of 0.01 mM acetic acid in 0.2% aqueous solution of formic acid at a flow rate of 0.2 mL/min at 40 C. Method Sample Preparation Finely ground sample (1 g) was weighed into a 10-mL glass centrifuge tube with cap. Carrez I and
2
80
60
40
0 80 100 120
110.0
20
73.2
109.1
m/z
Figure 1.
Mass spectrum for HMF obtained with positive APCI.
ions having m/z of 127 and 109 were used to monitor HMF in SIM mode. The ratio of these ions (response of ion 127/response of ion 109 = 1.12) was used to confirm the purity of HMF peak. The signal response was linear over a concentration range of 0.05 to 2.0 g/mL for both ions with correlation coefficients of higher than 0.99. On the basis of a signal-to-noise ratio of 3, the limit of detection (LOD) was determined to be 0.005 g/mL and 0.006 g/mL for ions having m/z 127 and m/z 109, respectively. LC/MS with APCI was found to be a powerful tool that allowed us to determine HMF sensitively and precisely. The chromatographic separation of HMF was performed on a ZORBAX Bonus RP narrow-bore column. The solution of 0.01 mM acetic acid in 0.2% aqueous solution of formic acid was used as
the mobile phase at a flow rate of 0.2 mL/min to increase the ionization yield during MS detection with an adequate separation of HMF in the column from interfering matrix co-extractives. Under these conditions, HMF eluted at 5.087 min with good retention time reproducibility (5.09 0.04 min, n = 10). See Figure 2. The capacity factor (k') was determined to be 2.33 for HMF based on the holdup time of 1.55 min. Usual approach for the extraction of free furfurals from solid food matrices entails extraction with water followed by clarification using Carrez I and II reagents. Direct LC/MS analysis of aqueous extract showed the presence of interfering compounds. Oasis HLB cartridge packed with a macroporous copolymer of the lipophilic divinylbenzene and the hydrophilic N-vinylpyrrolidone was, there-
600 550 500 450 400 350 300 1 700 650 600 550 500 450 400 350 300 250 1 2 3 4 5 2 3 4 5
m/z 109
min
m/z 127
O O
OH
HMF
min
Figure 2.
Extracted ion chromatogram (EIC) of an HMF standard (HMF concentration is 100 ng/mL).
fore, used to clean the extract prior to LC analysis. The clear aqueous extract was passed through a preconditioned cartridge. HMF present in the extract strongly interacted with the sorbent material while much of the co-extractives did not. HMF retained in the cartridge was then eluted with diethyl ether. It was determined that 0.5 mL of diethyl ether was sufficient to recover HMF from the cartridge completely. SPE cleanup brought significant improvement for the detection of HMF using MS in SIM mode. Total ion chromatogram indicated the presence of three
major peaks in the sample. HMF peak was identified by comparing both retention time and mass spectral data. The ratio of characteristic ions having m/z 127 and m/z 109 also confirmed the purity of HMF peak. The compound-specific ion [C6H5O2] having m/z of 109 was found to be more selective than the parent compound ion. So, the quantification of HMF was performed using the signal response recorded for this ion. The accuracy of the method was verified by analyzing spiked cereal-based baby foods. The recovery of HMF was determined by analyzing each of the
900 800 700 600 500 400 300 2.5 5 7.5 10 12.5 min
m/z 109
1600 1400 1200 1000 800 600 400 200 2.5 5 7.5 10 12.5
m/z 127
O O
OH
HMF
min
Figure 3.
EIC of a fruited yogurt sample (HMF concentration is 0.2 g/g).
5000
m/z 109
4000
3000
2000
1000
2.5
7.5
10
12.5
min
7000 6000 5000 4000 3000 2000 1000
m/z 127
O O
OH
HMF
2.5
7.5
10
12.5
min
Figure 4.
EIC of a crisp bread sample (HMF concentration is 17.5 g/g).
spiked samples four times for spiking levels ranging from 0.25 to 5.0 g/g. The mean percentage recoveries exceeded 90% for all levels. The method is capable of low concentrations, but also high concentrations of HMF in foods precisely and accurately. Figure 3 illustrates the EICs of a fruited yogurt sample having 0.2 g/g of HMF. It is difficult to measure such a low concentration of HMF using LC coupled to UV detection. Figure 4 illustrates the EICs of a crisp bread sample having 17.5 g/g of HMF.
Conclusion
The growing attention of the scientific community with regard to the potentially toxic effects of HMF requires new efforts to be made to establish new
rapid, reliable, and sensitive methods to determine HMF in real matrices. Previous methods usually dealt with the food items where HMF concentrations are comparatively higher and use extraction procedures that usually do not avoid potential interfering compounds prior to LC analysis. Presence of interferences may be problematic, particularly during the UV detection after LC separation when low concentrations of HMF are being measured in baby foods. The method described in this application combines 1) a rapid separation of HMF from the matrix co-extractives in a narrow-bore column, 2) an efficient cleanup of the extract using SPE, and 3) a selective detection of HMF using MS in a single analytical method.
Reference
1. V. Gkmen, H. Z. Senyuva, Improved method for the determination of hydroxymethylfurfural in baby foods using liquid chromatography-mass spectrometry, Journal of Agricultural and Food Chemistry, 2006, 54, 28452849.

Separation of Aflatoxins by HPLC Application
Environmental, Food Safety
Authors
Coral Barbas Facultad de CC Experimentales y de la Salud Universidad San Pablo-CEU Urbanizacin Monteprncipe Boadilla del Monte, 28668 Madrid Spain Andre Dams* Agilent Technologies, Inc. Amstelveen The Netherlands Ronald E. Majors Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA
*Present address: Andre Dams Analytical Consultancy, Amstelveen The Netherlands
abnormalities), and carcinogenic. Unfortunately, A. flavus is a common mold found in tropical and subtropical countries and has been found to cause aflatoxin contamination. This contamination is a result of inadequate storage conditions for certain agricultural commodities such as peanuts, cereal seeds, dried fruit, and a wide range of tree nuts such as pistachio, pecans, walnuts, almonds, and herbal seeds such as red and black pepper, cloves, and cinnamon. Because of their toxic and carcinogenic nature, there is a very low minimum detectable quantity (MDQ) for aflatoxin contamination in food. Chemical Nature Although 18 different aflatoxins have been identified, the four most prevalent aflatoxins are B1, B2, G1, and G2, whose chemical structures are shown in Figure 1. Aflatoxin B1 is one of the most potent and abundant environmental mutagens and carcinogens known. Aflatoxins are quite stable in many foods and are fairly resistant to degradation. Collectively, the aflatoxins are chemical derivatives of difurancoumarin. Pure aflatoxin B1 is a pale-white to yellow crystalline, odorless solid. Aflatoxins are freely soluble in moderately polar solvents such as chloroform, methanol, and dimethyl sulfoxide, and dissolve in water to the extent of 1020 mg/L. In methanol, they have fairly strong extinction coefficients (around 10,000) at 265 nm and 360362 nm. They fluoresce under UV radiation, and fluorescence detection is often used for trace analysis in HPLC. Since there are differences in fluorescence yields between B1 and B2, and between G1 and G2, it can be useful to run both UV and fluorescence detectors in series [1]. Aflatoxins have no polar functional groups, and can be separated by virtue of their hydrophobicity.
Abstract
Four target aflatoxins (B1, B2, G1, and G2) were separated by HPLC using an isocratic ternary mixture of water, methanol and acetonitrile, and detected using UV 365 nm. A baseline separation was achieved in less than 5.5 min.
Introduction
Aflatoxins are mycotoxins that are produced by various Aspergillus flavus molds. Not only are these compounds extremely toxic, but they are also mutagenic, teratogenic (causing fetal
O O O O O CH3 O O O O O CH3 O O
Experimental Conditions Chemicals: HPLC Conditions Column: Mobile phase: Flow rate: Temperature: Detector: Injection volume: The aflatoxins were purchased from Sigma Aldrich (Madrid, Spain). ZORBAX Eclipse XDB-C18, 4.6 mm 150 mm, 3.5 m Water/MeOH/ACN; 50/40/10 (V/V/V) 0.8 mL/min Ambient UV 365 nm 10 L (0.044 mg/mL)
Aflatoxin B1
O O O O O CH3 O O O O
Aflatoxin B2
O O O O
CH3

All four aflatoxins were separated using an isocratic ternary mixture of water, methanol, and acetonitrile. A baseline separation was achieved in less than 5.5 min. Although UV detection was shown here, in some cases, lower levels of detection may be obtained for B2 and G2 using fluorescence (ex = 365 nm, em = 455 nm) detection. Mass spectroscopic detection has also been used [1].
Aflatoxin G1
Aflatoxin G2
Figure 1.
Chemical structures of target aflatoxins.
HPLC Methodology While thin-layer chromatography was frequently used in the past, HPLC has been used in recent years because of its ease of operation and better quantitation. Most HPLC methods published to date have used reversed-phase HPLC on C18 bonded phases [14], where the aflatoxins are separated by their hydrophobicity. Most published separations have been performed on 5-m columns of 25-cm in length. The use of smaller particle size packings in shorter columns with faster separation times are now in vogue. These columns show that the same separations can be achieved in less time than on the longer columns with similar resolution. In the present study, we desired to show that using such a column can provide improved results compared to the older methods. See Figure 2.
300 250 200 mAU 150 100 50 0 1 2 3 4 5 6 Minutes 7 8 9 10
References
1. R. Schuster, G. Marx, and M. Rothaupt, Analysis of Mycotoxins by HPLC with Automated Spectroscopic Confirmation by Spectral Library, Agilent Technologies, publication 5091-8692E, www.agilent.com/chem 2. A. Gratzfeld-Heusgen, HPLC Analysis of Aflatoxins in Pistachio Nuts, Agilent Technologies, publication 5966-0632E, www.agilent.com/chem 3. I. Ferreira, E. Mendes, and M. Oliveira, (2004) J. Liquid Chromatog., 27 (2), 325334. 4. E. Chiaavaro, C. Dall'Asta, G. Galaverna, A. Biancardi, and E. Gambarelli, (2001) J. Chromatogr. A, 937 (12), 3140.
G2 G1 B2 B1

Figure 2.
Reversed-phase separation of target aflatoxins using ZORBAX XDB-C18 Rapid Resolution column.
Identification and Isolation of DSP-Toxins Using a Combined LC/MS-System for Analytical and Semipreparative Work Application
Food Safety
Authors
Norbert Helle TeLA GmbH Bremerhaven, Germany Sebastian Lippemeier BlueBioTech Microalgen Biotechnologie Ellerbek, Germany Jrgen Wendt Agilent Technologies Sales and Support GmbH Waldbronn, Germany
mussels, clams and oysters. The symptoms include abdominal pain, vomiting, nausea, headache, diarrhea, chills, and fever. DSP toxins can be classified in three groups: the okadaic acid (OA) group involving OA and the dinophysistoxins (DTXs), the pectenotoxin group (PTXs) and the yessotoxin group (YTXs). Inside the OA group, OA and DTX-1 are the main toxins responsible for DSP outbreaks. The outbreaks led to the establishment of control programs for marine biotoxins in many countries. In Germany residues of DSP toxins in mussels are controlled at present under the regulation of the Fischhygiene-Verordnung of 8th June 2000. This Order requires the testing of shellfish for the presence of toxins by means of animal tests (mouse bioassays) or by chemical analytical procedures [1, 2]. Liquid chromatography with fluorescence detection is an established technique, but it requires the derivatization of the not naturally fluorescent DSP toxins. Using LC/MS coupled with electrospray ionization (ESI) more sensitive and selective results are attainable with less sample preparation. The greatest problem regarding the analytical methods for monitoring DSP toxins is the availability of pure reference material. The DSP toxins OA and DTX-1 can be isolated from crude extracts of Prorocentrum lima algae (see Figure 1) using mass-based fraction collection in semipreparative mode. The present work describes the configuration, setup, and operation of a combined LC/MS system for analytical and semipreparative work.
Abstract
The configuration and operation of a combined liquid chromatography/mass spectrometry (LC/MS) system to identify and isolate DSP-toxins is described. In the analytical mode, okadaic acid (OA) and dinophysistoxin-1 (DTX-1) are more selectively and sensitively monitored when compared to LC with fluorescence detection. With less sample preparation, the detection limits are decreased by a factor of 35, depending on the matrix. In semipreparative mode, OA and DTX-1 could be isolated from crude extracts of Prorocentrum lima algae using mass-based fraction collection with a purity >98%. Due to this method, reference standards of DSP toxins are now commercially available.
Introduction
Diarrheic shellfish poisoning (DSP) is a gastrointestinal syndrome that occurs in humans after the consumption of bivalve mollusks such as scallops,
Figure 1.
Prorocentrum lima algae under the microscope.
Experimental
The DSP toxins shown in Figure 2 were analyzed in this work. The analyses were conducted in two modes: Analytical and Semipreparative.
CH3 OH H O HO CH3 H O O OR1 H CH3 O O H H OH H CH3 R2 O OH CH2 H3C O
OA: R1 = H, R2 = H DTX-1: R1 = H, R2 = CH3
C44H68O13 C45H70O13
Chemical and physical properties: Polyether structure, Carboxylic acid Lipophilic, and no chromophore
Figure 2.
DSP toxins.
LC/MS Method Details - Analytical LC Conditions Instrument: Column: Mobile phase: Gradient: Agilent 1100 HPLC (Quaternary pump) 150 3.0 mm ZORBAX SB-C18, 5 m A Water (0.1% Formic acid) B Methanol 20% B at 0 min 20% B at 5 min 80% B at 20 min 28 min; 4 min 0.6 mL/min 10 L Agilent LC/MSD Positive/Negative switching ESI 12 L/min 60 psig 350 C 3000 V (positive and negative)
Stop time: Post time: Flow rate: Injection vol: MS Conditions Instrument: Source: Drying gas flow rate Nebulizer: Drying gas temp: Vcap: 2
LC/MS Method Details - Semipreparative LC Conditions Instrument 1: Column: Mobile phase: Gradient: Agilent 1100 HPLC (Quaternary pump) 50 9.4 mm ZORBAX SB-C18, 5 m A Water (0.1% Formic acid) B Methanol 20% B at 0 min 20% B at 5 min 80% B at 20 min 28 min 4 min 7.0 mL/min 100 L (250 L using Multiple Draw Mode) Agilent 1100 HPLC (Isocratic pump) for makeup flow 0.8 mL/min (50% H2O + 50% MeOH + 0.1% Formic acid) Split ratio 271:1 Agilent LC/MSD Negative ESI 12 L/min 60 psig 350 C 3000 V (positive) Use method target mass; Adducts: (MH)
Stop time: Post time: Flow rate: Injection vol: Instrument 2: Flow rate: Active splitter: MS Conditions Instrument: Source: Drying gas flow: Nebulizer: Drying gas temp: Vcap: MSD Fraction Collection Setup FC Mode:

Analytical Work In the analytical mode of the LC/MS system (Figure 3) the DSP toxins were monitored using ESI with positive/negative mode switching. The positive ion mode is four times more sensitive than the negative ion mode (Figure 4). Mass spectra for OA and DTX-1 show a sodiated molecular ion instead of a protonated molecular ion, and characteristic fragment ions [M+H nH2O]+, where n = 14, formed by a sequential loss of water. In negative ion mode only the [MH] ion is detected. LC/MS provided a more selective and sensitive method for monitoring DSP toxins in comparison to LC with fluorescence detection (Figure 5), by a factor of 35.
Analytical column
Fluorescence detector
LC/MSD
2-position/6-port valve
Splitter off
Figure 3.
System diagram (analytical work).
OA positive
100 80
769.4 751.4
Max: 611673
60 40 20 0 650 30 25
OA positive
827.4 752.4 770.4 828.4 733.4 753.4 771.4 829.4
787.4
675
700
725
750
775
800
825
850
m/z
Max: 177930
OA negative
20 15 10 5 0 650 675 700 725 750 775 800 825 850 m/z
805.4
OA negative
804.4
17.0
17.2
17.4
17.6
Figure 4.
LC/MS analysis of OA.
1467 g/kg Okadaic acid
1467 g/kg Okadaic acid
mV
mV
HPLC/Fluorescence
LC-MS, API pos.
10
20
30
40
min
10
14
18 min
Figure 5.
Comparative analysis of DSP toxins in shellfish.
Semipreparative Work The reference standards could be obtained by switching the system to semipreparative mode (Figure 6). The valve is switched to position 2.
The main flow now goes to the semipreparative column and then through the splitter to fraction collector (AS). The make-up flow goes through the splitter where it picks up some of the compound from the main flow and goes to the MS-detector (MSD).
2-position/6-port valve Semi-prep column
Splitter on Fraction collector Figure 6. System diagram (semipreparative work).
Make-up pump
Besides OA and DTX-1, a new OA-toxin with similar mass spectral properties could be isolated from crude extracts of Prorocentrum lima algae using mass-based fraction collection (Figure 7). The mass-based fraction collection of a methanolic extract of Prorocentrum lima algae results in three fractions: OA, DTX-1 and an unknown toxin. From MSn experiments it can be determined that the molecular structure of the unknown toxin must be very similar to those of OA and DTX-1.
Norm. 200000
MSD1 TIC, MS File (E:\VIENNA\311003\002-0301.D) API-ES, Neg, SIM, Frag: 160
DTX-1
175000
OA
150000
125000
100000
1-Vial 11
75000
50000
Unknown DSP
1-Vial 10 1-Vial 12
25000
0 18 19 20 21 22 23 24 min
Figure 7.
Mass-based fraction collection of DSP toxins.
Because of the low concentration, the target compounds had to be collected from multiple injections of the same sample, a process usually referred to as pooling (Figure 8). Reanalysis of the collected fractions gave results for purity >98%. This method is robust (Figure 9) and has now resulted in making reference standards of DSP toxins commercially available.
Fraction collector module
Pooling lines
Up to 10 bottles
1 3 2 1
Figure 8.
Pooling.
Multiple fraction collection scheme
19
20
21
22
23
Figure 9.
Robustness of the method - overlay of 10 mass-based fraction collection runs. 7
Conclusions
Configuration and operation of a combined LC/MS system to identify and isolate DSP toxins is described. In the analytical mode, OA and DTX-1 were monitored more selectively and sensitively than by using LC with fluorescence detection. With less sample preparation, the detection limits could be decreased by a factor of 35, depending on the matrix. In the semipreparative mode OA and DTX-1 could be isolated from crude extracts of Prorocentrum lima algae using mass-based fraction collection with a purity >98%. Due to this method reference standards of DSP toxins are now commercially available.
References
1. M.A. Quilliam, A. Gago-Martinez, and J.A. Rodriguez-Vasquez, Improved method for preparation and use of 9-anthryldiazomethane for derivatization of hydroxycarbolic acids Application to diarrhetic shellfish poisoning toxins, (1988), Journal of Chromatography A, 807, 229239. 2. A.G. Bauder; A.D. Cembella, V.M. Bricelj, and M.A. Quilliam, Uptake and fate of diarrhetic shellfish poisoning from the dinoflagellate Prorocentrum lima in the bay scallop Argopecten irradians", (2001), Marine Ecol. Progr. Ser., 213, 3952.

For more details concerning this note, please contact Juergen_Wendt @Agilent.com
Identification and characterization of new ergot alkaloids
Application
Abstract
Ergotamine
267 331 358 252
O
Mark Stahl Edgar Ngele
314
- H2O -
296 - CH2 -CH2 CO

N O CH2
CH2-C 6H5 -CH 206
CH3 OH O NH O N CH 3 N
245
223 - CH2 =N-CH2 180 208
322
NH
260
Ergometrine
252 103
O
267 75
NH
60
C H2 O H CH CH3 C
A crude extract of ergot fungus was analyzed with LC/MS for identification and characterization of its ergot alkaloid content. Alkaloids posessing a lysergic acid structure were identified with a nano LC/MS ion trap system and purified using mass based fraction collection. The purified alkaloids were then structurally characterized with MSn experiments. Several well-known ertgot alkaloids were identified and several so far unknowns were determined and structurally characterized using MSn.
223
N
Introduction
The ergot fungus is a 4-cm long and 3-mm wide black ascomycet, living parasitic on grasses and crops. It produces several different alkaloids which in former times lead to many mass poisonings. The latest poisoning occurred 1951 in France. It resulted from contaminated flour and caused 300 deaths. To ensure that this does not happen again today, crops are routinely tested.1,2 For this reason, as well as for medical purposes it is important to identify and characterize potential toxins as thoroughly as possible. They are mainly derivatives of the lysergic acid which can be classified into two different types: ergometrine type, where the carboxy group of the lysergic acid is esterified with an amino alcohol, and peptide type, where the carboxy group of the lysergic acid is esterified with a tricyclic peptide. Ergot alkaloids are used for medical purposes to start uterus contraction and the detachment of the placenta. Furthermore, they are used in the treatment of migraine and heart rhythm disturbances. These effects are caused by the alkaloids acting as partial agonist or antagonist (depending on the single alkaoid and the single receptor) on the a-adreno, the dopamine and the serotonine receptor. These pharmacological targets are also responsible for the symptoms occurring in case of intoxication: mydriasis, hallucinations, diarrhea, spasms, paralysis, loss of extremities and finally exitus. For treatment of intoxication the antidote diazepam and emergency medical treatment is used.
- CH2 =N-CH2 180 208 NH
CH3 C
Dehydroergotamine
265 331
NH
314
CH 3 OH O
296 - CH2 -CH2 CO

N N O CH 2
- CH 2 -C C6 H5
206
358
250 O
245
221
O N CH 3
- CH2 =N-CH2
178 206
NH
320
260
Hydroxyergotamine
332 267 347 OH 374 252 O
NH CH2 O - H2O
314
- H2O
296 -CH -C C H 206
OH N N CH2 O
-CH 2 -CH 2 CO 245
223 -CH2 =N =N-CH CH2 208
O N CH3
180
338
NH
260
Ergoval
145 223
267 116 252

O
NH CH
CH3 CH
101
CH H COOH
43
-CH CH 2=N-CH CH 2 180 208

NH
N CH3
325
MSn analysis of ergotamine (A)
MSn analysis of ergometrine (B)

Intens . 267.3 208.1 223.1 308.2 283.2
Intens. x10 4 5 4 3 2 1 0 100 223.2 207.9 297.2 180.1 200 300 400 500 268.2
564.2
6000
4000 320.2 MS 2 of ergotamine (582 Da) 536.3 600 700 m/z 0 100 200 300 400 500 600 700 m/z 197.3 2000 252.8 MS 2 of ergometrine (326 Da)
Intens. 2000 1500
297.1 268.1 320.2 223.1
Intens.
292.2
MS 3 582 564
1500 1000
235.1
MS 3 582 320
Intens 120 100 80 180.1
1000 500 0 100 200 300 400 500 600 700 m/z 536.3
500 0
223.1 197.1
100 200 300
400 500
600 700 m/z
60 40
197.3 252.8 208.1 223.1 100 200 300 400
MS 3 326 267
. Intens
223.1
Intens.
208.1
20
5000 4000 3000 2000 1000 0 251.1
MS 3 582 268
1500
MS 3 582 223
0 500 600 700 m/z
1000 500 180.1
0 100 200 300 400 500 600 700 m/z 100 200 300 400 500 600 700 m/z
Figure 1 MSn spectra of ergotamine (A) and ergometrine (B)
Materials and methods

Sample: The crude extract of ergot fungus was purchased from Firma Dr. Hetterich (Fuerth, Germany), 0.22 m filtrated and directly injected. Analysis: In a first analysis using the Agilent 1100 Series nano LC system directly coupled to an Agilent ion trap mass spectrometer, substances showing the lysergic acid structure (m/z 223 or m/z 208) (structures front page and figure 1) were identified by
TIC x107 1.25 1.00 0.75 0.50 0.25 0.00
60
120
180 Time [min]
240
300
360
Figure 2 Total ion chromatogram of the separation of the crude ergot extract
MSn analysis of dehydroergotamine (C)

Intens. x10 4 5 4 3 2 1 0 100 200 300 400 Intens 500 600 700 m/z
205.1 264.8 290.2 562.2 249.1 318.2
MSn analysis of hydroxyergotamine (D)

Intens. . x10 4 3 580.2
MS2 of dehydroergotamine 580 Da) (580
318.2
249.1
MS2 of hydroxyergotamine (598 Da)

562.3 400 500 600 700 m/z
0 100
222.1 267.3 336. 336.2 200 300
Intens . 6000
249.1 297.2 3 4 334.2 318.2
4000 221.1 290.2 2000 Intens . 249.2 300 Intens . x10 4 1.0 249.1
.
MS 3 598 580
208.1 267.3 221.1 0 100 200 300 400 500
562.3 600 Intens . 1500 700 m/z 233.9
2000
221.8
1500
1000 500
MS 3 580 318
200 205.1 100
MS 3
0.8 0.6 0.4 0.2
580
249
MS 3 598 318
1000 500
178.1
MS 3 598 249
208.0 0 100 200 300 400 500 600 700 m/z 180 200 220 240 260 m/z
0 200 400 600 800 1000 m/z
0 100 200 300 400 500 600 700 m/z
0.0
Figure 3 MSn spectra of dehydroergotamine (C) and hydroxyergotamine (D)
LC/MS/MS experiments. To confirm the structure of already known alkaloids LC/MSn experiments were carried out. To achieve thorough characterization of the unknown alkaloids they were purified using an Agilent 1100 Series purification system in mass-based fraction collection mode. The purified alkaloids were then structurally characterized with flow injection MSn experiments thus allowing to optimize fragmentation parameters for each ion individually.
Results and discussion

The analytical chromatogram of the crude ergot extract is shown in figure 2. It clearly demonstrates the large amount of different substances that are present in the sample. Using the technique described above the two well known alkaloids ergotamine and ergometrine were identified (figure 1)2, 3, 4. Additionally, three so far unknowns were found. They were purified and structural investigations were done as described
before. The comparison of their MSn spectra to those of the known derivatives allowed to characterize the derivatives and led to the structural proposals shown in figures 3 and 4. Dehydroergotamine is an oxidized derivative of ergotamine whereas in hydroxyergotamine the amino acid alanine has been replaced by serine. In ergoval the lysergic acid is esterified with the amino acid valine. As this amino acid is also used in other alkaoids of the tricyclic peptide type 4 it may either
represent a pre-state or a reduction product of other forms, but these could not be found within the extract used. Thus it is more likely that the valine is the final product.
MSn analysis of ergoval (E)

Intens. 1500 208.0
1000
MS 2 of ergoval ( (368 Da)

322.2 311.3 300
Conclusion
We demonstrated that nano-LCMS/MS screening allows to identify analytes posessing a certain structure out of a crude mixture. Their purification followed by massbased fraction collection allows their structural characterization with MSn. Apart from two wellknown ergot alkaloids three so far unknowns have been identified and structurally characterized by MSn experiments. Based on these spectra their structures could be proposed.
223.0 500 197.0 0 50 100 Intens. 80 60 40 20 0 50 100 150 200 250 300 350 m/z 150 200 250 350 m/z 251.0 269.1
353.1
MS 3 368 322
223.0
191.9 207.9
280.1 251.2 311.6 322.1
Figure 4 MSn spectra of ergoval (E)
References
1. Methode zur Bestimmung von Mutterkornalkaloiden in Lebensmitteln, Klug, C.; Baltes, W.; Krnert, W.; Weber, R., Z Lebensm Unters Forsch, 186, 108 113, 1988. 2. Analysis of ergot alkaloids in endophyte infected tall fescue by liquid chromatography electrospray ionisation mass spectrometry, Shelby, R. A.; Olsovska, J.; Havlicek, V.; Flieger, M.; J. Agric. Food Chem., 45, 1997. 3. Untersuchungen zur Massenspektrometrie von Mutterkornalkaloiden, Voigt, D.; Johne, S.; Grger. D., Pharmazie, 29, H. 10 11, 697 700, 1974. 4. Mass spectrometric amino acid structure determination in ergopeptides, 4. Halada, P.; Sedmera, P.; Havlicek, V.; Jegorov, A.; Cvak, L.; Ryska, M., Eur. Mass Spectrometry, 4, 385 392, 1998. Mark Stahl and Edgar Ngele are Application Chemists at Agilent Technologies GmbH, Waldbronn, Germany. www.agilent.com/chem/1100
The information in this publication is subject to change without notice. Copyright 2003 Agilent Technologies All Rights Reserved. Reproduction, adaptation or translation without prior written permission is prohibited, except as allowed under the copyright laws. Published November 1, 2003 Publication number 5989-0261EN
0
Analysis of Mycotoxins by HPLC with Automated Confirmation by Spectral Library Application Note
Food Analysis
Rainer Schuster Gerhard Marx Michael Rothaupt This note describes the sample preparation, chromatographic (FLD) for aflatoxins, ochratoxin and zearalenones or mass control and by retention time tagging.
separation and detection of four different types of mycotoxins in food samples. Deciding which approach to adopt for analyzing these depends on the sample matrix and the type of fungus it has been contaminated with. Various professional organizations have proposed a variety of sample preparation methodsthose for aflatoxins, ochratoxin A, patulin and zearaleneone are described here. All HPLC separations have been performed on reversed phase material (normal phase chromatography; a diol column can be used for patulin) and monitored with UV-visible absorbance diode-array detection (DAD) and fluorescence detection
spectrometry (MS) for aflatoxins. Most compounds have been identified and confirmed by UV-visible absorbance spectral library search, purity
Introduction
Mycotoxins are highly toxic compounds produced by fungi. These toxins can contaminate foodstuffs when storage conditions are favorable to fungal growth. Mycotoxin nomenclature very often results from the fungi where the substance was first detected, for example aflatoxins in Aspergillus flavus strains, ochratoxin in Aspergillus ochraceus, patulin in Penicillium and Aspergillus, zearalenone in Fusarium. Most of these mycotoxins have been identified after cases of poisoning in livestock or the population at large. In 1969 more than 100,000 turkeys died of an unknown condition (so-called Turkey X) that was finally traced to peanutsa component of their feed contaminated with Aspergillus flavus. During the wartime winter of 1940 in the USSR many people died after eating grain poorly stored and highly contaminated with Fusarium toxins such as zearalenone and fusarin C. A similar case occurred in 1965 in South Africa with ochratoxins found in cereals accumulated unmetabolized in animal kidneys.
Aflatoxins are known to be mutagens, teratogens (causing fetal abnormalities) and carcinogens (particularly in cancer of the liver or kidneys). Ochratoxins cause nephropathies in pigs, are teratogenic, and carcinogenic particularly in the liver and kidneys. Zearalenone shows estrogenic effects in sows and poultry, and affects the liver and kidneys. Patuline is a powerful mutagenic and cytotoxic compound. The intake of these mycotoxins over a long period at very low concentrations may be highly dangerous, yet difficult to combat since the small quantities are difficult to trace. Currently most mycotoxins are still assayed using thin-layer chromatography (TLC), which permits effective compound separation and characterization. Such assay may be performed with satisfactory sensitivity when the compounds to be detected are fluorescenta fluorodensitometer reads the plates quantitatively and objectively and has become indispensable to the control laboratory. However due to its higher separation power and shorter analysis times, use of HPLC has expanded rapidly in recent times. Reversed phase chromatography separates mycotoxins of widely different polarity ranges. The diversity of detection systems (diode array, fluorimetry, and even mass spectrometry) permit identification, confirmation and accurate assay of a great variety of these compounds. In addition HPLC is well suited to existing safety regulations and automation in laboratories dealing with a large number of samples.
The complexity in composition of processed foodstuffs makes a fixed routine necessary for analysis: 1. sampling protocol that ensures representative data from any one batch 2. extraction of mycotoxins, using mostly chloroform, acetone, or methanol 3. purification of the extract with clean-up methods 4. concentration of the extract 5. qualitative detection and assay of the mycotoxins. In this paper we describe the analysis of 4 different types of mycotoxins. First we describe the chemical nature and occurrence of these toxins.
Experimental
Compound class Matrix Nuts, spices, animal food, milk, dairy products
Sample preparation Extraction 35LMBG
Chromatographic conditions Hypersil ODS 100 x 2.1-mm id, 3 m particles HP 799160D-352 Water-methanol-ace tonitrile 63:26:11 as isocratic mixture* Flow 0.3 ml/min at 25C
Table 1 gives a short overview of analysis conditions for the four different mycotoxins aflatoxins, ochratoxin A, zearalenone and patuline.
Aflatoxins G2, G2, B2, B1, M2, M1
Ochratoxin A
Cereals, flour, figs
Extraction 35LMBG Acidify with HCI. Extract with toluene. Clean up SiO2. Elute toluene-CH3COOH 9: 1
Lichrospher 100 RP18 125 x 4-mm id, 5 m particles HP 799250D-564-3
Water with 2 % acetic acid/acetonitrile, 1 : 1* Flow 1 ml/min at 40C Zearalenone Cereals Extract with toluene. Sep-pak clean up. Elute toluene-acetone 95: 5. Hypersil ODS 100 x 2.1-mm id, 3 m particles HP 799160D-352 Water-methanol-ace tonitrile 5:4:1 as isocratic mixture* Flow 0 : 45 ml/min. at 45C Patuline Apple products Clean-up on Extrelut Silica gel clean up Elute toluene-ethyl acetate 3: 1. a)Superspher RP 18 125 x 4-mm id, 4 m particles HP 799250D-464, Water-acetonitrile, 95 % to 5 % gradient Flow 0.6 ml/min at 40 C b) Lichrospher Diol 125 x 4-mm id, 5 m particles Hexane-isopropanl 95:5 as isocratic mixture Flow 0.6 ml/min * 100 % B is recommended for cleaning the column. Table 1. Sample preparation and chromatographic conditions for mycotoxins in foodstuffs
Compound types: Aflatoxinsare chemical derivatives of difurancoumarin (figure 1). Although a number of different aflatoxin metabolites are known, interest is usually focused on the four main aflatoxins B1, B2, G1, G2 and the so-called milk toxin M 1. Aflatoxin B1 is in the majority of cases the most abundant toxin, the most toxic and the most potent carcinogen. Maximum levels for B1 are usually given for the individual compound, 2 ppb in Germany, 5 ppb in France and 1 ppb in Switzerland, for example. United States legislation regulates the aflatoxin content of a contaminated product as the sum of B1-plus-B2-plus-G1-plus-G2, which may not exceed 20 ppb. Aflatoxins are most often analyzed in nuts, for example peanuts and pistachios, cereals, figs, bread, meat, eggs, butter, milk, margarine, juices, cottonseed products, and cocoa beans. Considering the complexity of these matrices, sample preparation is the most important step for reliable results. Since fungal growthand therefore contamination by aflatoxinis not homogeneous, normal sampling gives mediocre results. The US Department of Agriculture (USDA) tackled the problem by defining a sampling protocol for peanuts which involves as many as eight assays on four samples of more than 1 kg. Sampling is not specified in European countries. At best, a few hundred grams are taken to determine the mean level of aflatoxins in a batch as large as of a couple of thousand tons of grain.
Figure 1. Structure of aflatoxins and their maximum permitted concentrations (given for Germany1)
OchratoxinsThe mycotoxin ochratoxin A can be produced by different fungi including Aspergillus and Penicillium. Of the ochratoxins A, B, and C, the latter two so far have not been found in naturally contaminated products. Beside nephrotoxicity, ochratoxin A has hepatoxic, teratogenic and carcinogenic properties in the kidneys. Ochratoxin A was found in various foodstuffs. Analysis of more that 900 plant samples show a contamination rate of about 13 %, mostly in barley, oats and wheat. The concentrations of ochratoxin A found varied from 0.1 to 200 g/kg.2 A review of results from various countries, covering around 7000 samples, reported that contamination was about 14 %.3 Ochratoxin A is the primary agent in so-called
mycotoxic porcine nephropathy (MPN) a disease prevalent in pigs. The toxicity is a third that of the toxicity of aflatoxin B1 in rats. The main human intake is assumed to be through the consumption of pork and wholemeal products.
Figure 2. Structure of three common ochratoxins
Figure 3. Chemical structures of zearalenone, aand b-zearalenol and a -zearalanol (Zeranol)
Figure 4. Chemical structure of patulin
Zearalenonean estrogenic-efficient mycotoxin produced by Fusarium, occurs mainly in a variety of natural products such as corn and other grains. Whereas the acute toxicity of zearalenone is low, its intake is linked to various possible estrogenic disease effects in children.4 After carcinogenity was determined in rodents, a recent risk assessment resulted in an estimated safe intake of not more than 0.05 g for each kilogram body weight per day for humans.12 Use of contaminated animal feed means that these compounds are present in cow's milk as zearalenone and the diastereomer metabolites and b-zearalenol. Another zearalenone derivative is the synthetic a -zearalanol (also known as Zeranol) which is used in some countries for fattening cattle. The recommended limit for zearalanol is 10 g kg for liver and
2 g/kg for other meats.7 Their use as anabolic agents is prohibited in the European Community. investigations of food and animal feedstuffs have shown zearalenone concentrations between 0.001 and 2.0 mg/kg.5 The highest levels, 1700 mg/kg, were found in silo-corn.6
Patulinunstable in cereals, mainly occurs in fruit, especially apples, is a metabolite of several fungi of Penicillium and Aspergillus. Most of the survey work has been done on apple juices and apple-based products.
Aflatoxin sample preparation
Extraction of aflatoxins with chloroform
For sample preparation different methods are described in the CB method (AOAC and EEC) for extraction of aflatoxins literature.8,9,18 Lipid should be 50 g of finely ground sample are mixed with 25 g of diatomaceous earth and moistened with eliminated after analyte extraction 25 ml of water. This mixture is carefully shaken, diluted in 250 ml of chloroform and shaken vigorously for 30 minutes on a vibration shaker. A 50 ml portion of chloroform extract is if lipid content exceeds 5 %. collected for purification and assay. The addition of water facilitates chloroform penetration Solvents in which the aflatoxins into substrates derived from plants, while the diatomaceous earth retains various are insoluble are hexane, substances like pigments. petroleum ether, pentane and Method 35 Lebensmittel und Bedaffspeyenstands Gesetz (LMBG)8 for extraction of isooctane, and are used in Soxhlet aflatoxins apparatus (6h), shaking or column 20 g of finely ground sample are mixed with 20 g of silica gel, particle size 2045 m (for example celit 545, Serva, Germany). This mixture is diluted with 200 ml of chloroform and clean up. The contaminants branch (CB) extraction method is 20 ml of water and vigorously shaken for 30 minutes on a vibration shaker. After filtration, 100 ml are evaporated close to dryness on a rotary evaporator (temperature 40 C). based on a chloroform-water 9 mixture, a method adopted by the Method 35 LMBG for extraction of M1 in milk and milk powder 50 ml acetone and 5 g sodium chloride and 1 ml 1 N H3PO4 are added to 50 g milk, or 10 9 Association of Official Analytical milk powder homogenized in 40 g water, and shaken for 10 minutes. After addition of 100 ml Chemists (AOAC) 20.029 for dichloromethane and a further 10 minutes shaking, 25 g of silica gel, particle size 2045 m determination of aflatoxins in (for example celit 545, Serva, Germany) is added and shaken again. The dichloromethane/ ground nuts and recommended by acetone phase is filtered and 100 ml of the filtrate (equivalent to 33.33 g milk or 6.66 g milk powder) is evaporated to dryness at 40 C in a rotary evaporator. the European Economic Community (EEC) for B1 in simple animal nutrition foodstuffs. Purification of aflatoxins Whatever extraction method is used the resulting extract still contains, besides the aflatoxins, various impurities (lipids, pigments and so on) requiring an extra clean-up step. Apart from purification by precipitation or by liquid-liquid partition, the most commonly used technique is column adsorption chromatography.
Method according to AOAC/ EEC and 35 LMBG regulations8 A glass column (400 x 30mm) is filled in succession with 5 g sodium sulfate, 10 g of silica gel (63200 m, dried at 105 C for lie), and 15 g anhydrous sodium sulfate topped up with some cotton-wool. The extract to be cleaned is added on top and eluted with 1520 ml chloroform. Then the column is washed with 150 ml hexane and 150 ml diethylether to remove lipids and other interfering compounds from the aflatoxins. The aflatoxins are eluted with 150 ml of a chloroform/methanol (97: 3) mixture. The eluate is dried down and redissolved in a suitable solvent for assay by HPLC (methanol).
Extraction of aflatoxins from milk, AOAC 26.139

To 25 ml of milk, 10 drops of NH4OH are added, swirled and diluted by 70 ml acetonitrile. The mixture is shaken for 1 minute and centrifuged for 5 minutes at 1000 rpm. The aqueous alkaline acetonitrile supernatant is transferred and evaporated on a rotary evaporator at 45 C. The residue is acidified with 15 drops (about 500 l) of HCI to pH 1.3 and partitioned into methylene chloride on the liquid-liquid extraction column, ChemElutTM (Analytichem, United Kingdom). The dichloromethane is evaporated off by rotary evaporation. After cooling, the residue is redissolved in 1.5 ml of dichloromethane, evaporates and redissolves in 500 l hexane-dichloromethane (1: 9). Clean-up is performed on a BondElut NH2 cartridge (Analytichem, United Kingdom) conditioned with hexane-dichloromethane (1: 9).100 l of the milk extract is transferred to the column; fats are removed with 230 l hexane-dichloromethane (1: 9) while zearalenone is eluted with 1 ml of methanol. After evaporation of methanol the residue is dissolved in 500 l of mobile phase.
Extraction of ochratoxins with toluene

According to 35 LMBG Method 15-00-1, AOAC 26.10026.125 30 ml of 2 M HCI in 50 ml of 0.4 M magnesium chloride solution is added to 20 g of ground and mixed sample. After homogenization, 100 ml toluene is added and shaken vigorously for 60 minutes. The suspension is separated by centrifuge and 50 ml of the toluene supernatant is passed through a preconditioned Sep Pak silica gel column. The column is washed with two 10-ml aliquots of hexane, 10 ml of toluene/acetone (95: 5) and 5 ml of toluene. Ochratoxin A is eluted with two aliquots of 15 ml toluene/acetic acid (9: 1) and dried down at 40 C. The residue is redissolved in 1 ml of mobile phase and filtered.
Ochratoxin sample preparation Methylester derivatives of the ochratoxins can also be analyzed. From the extracted sample, 500 l is evaporated to dryness and redissolved with 1 ml of dichloromethane and 2 ml of 14 % boron trifluoride in methanol. The solution is heated for 15 minutes at 5060 C and after cooling diluted in 30 ml of distilled water and extracted with three 10-ml aliquots of dichloromethane. The organic phase is filtered through sodium sulfate, dried down and dissolved in 500 l of mobile phase.
Zearalenone sample preparation Thanks to Eppley s technique, zearalenone, aflatoxins and ochratoxin can be simultaneously extracted on Sep-Pak silica cartridges. The sample is added as a toluene extract, washed with toluene and zearalenone is eluted with 10 ml of toluene-acetone (95: 5) mixture.
13
Patulin sample preparation Two approaches are documented.1 Fruit juices can be cleaned on an Extrelut cartridge followed by analyte extraction on a silica gel column with toluene-ethyl acetate (3: 1) before HPLC assay.14 The analyte can be extracted into ethyl acetate, followed by partition extraction into 1.4 % Na2CO3 solution and back into ethyl acetate. After evaporation of the ethyl acetate at 40 C, the residue is dissolved in methanol-ethyl acetate (9: 1) if it is to be analyzed on a reversed-phase column packing material or in hexane-isopropanol if a diol-phase column packing material is used (details of suitable columns are given in table 1).18
Chromatographic separations, peak confirmation and quantification We used a Hewlett-Packard HP 1090 Series M liquid chromatograph with DR 5 binary solvent-delivery system, variable-volume auto-injector, temperature-controlled column compartment and solvent-preheating device. Mobile phase methanol and acetonitrile were of HPLC reagent quality (Baker, Gross-Gerau, Germany). A diode-array UV-Visible absorbance detector was used together with HPLC3D ChemStation software to automatically quantify the mycotoxins and identify them using spectral libraries. Fluorescent species were detected using an HP 1046A programmable flourescence detector (FLD) under the control of the HPLC3D ChemStation, using lex 265 nm, lem 455 nm for aflatoxins, lex 247 nm, lem 480 nm for ochratoxin A and lex 236 nm, lem 464 nm for zearalenone. Aflatoxins were also determined using mass spectrometry on an HP 5989 MS Engine equipped with negative ion detection and Thermospray options. The electron filament capability was used to provide higher sensitivity. LC eluant passed through a capillary tube and was simultaneously heated to approaching the boiling point. The resulting liquid-vapor was injected into the mass spectrometer where it was ionized and analyzed. The mass spectrometer was controlled and the data were analyzed by the HP 59940A MS ChemStation (HP-UX series).
Results and discussion Aflatoxin assay by HPLC-DAD and HPLC-FLD Thin layer chromatography can be replaced by reversed phase HPLC, improving accuracy, and dramatically speeding up the time required to assay, for example B1 takes three hours by TLC, M1 four hours. Figure 6 shows a separation of the common aflatoxins M2 (5 ng), M1 (10 ng), G2 (1.5 ng), G1 (5 ng), B2 (1.5 ng), B1 (5 ng) on a reversed phase column (refer to table 1 for conditions).
Due to the extreme differences in fluorescence yields for B2 and B1 respectively G2 and G1 (B2 FLD yield is about 60 times higher than B1) it can be useful to run both detectors in series. Diode-array detection in addition gives us the UV-visible absorbance spectra dimension for further identification of the aflatoxins. As an alternative to the isocratic run with subsequent 100 % B wash, a gradient analysis from 35 % B (methanol-acetonitrile, 26: 11) to 55 % B in 10 min and 100 % B in 14 min (at 35 C) might be used. Peaks become much sharper than under isocratic conditions, with higher signal-to-noise, and less polar compounds in the food extract are eluted in this run.
Flow rate Mobile phase
0.30 ml/min Isocratic water methanolacetonitrile (63: 26: 11) mixture
Detection: Fluorescence Diode-array lex 365 m, lem 455 nm 365 nm
Figure 5. Analysis of the common aflatoxins by fluorescence and diode-array detection
Figure 6. Structure of aflatoxins after hydrolysis with trifluoracetic acid (TFA), from B1 to B2a
If fluorescence is used alone it might be desirable to improve the B1 and G1 fluorescence yield by hydrating the double bond of the furanic ring (figure 6) with trifluoroacetic acid (TFA) to form the corresponding hemiacetals B2a and G2a. This approach can also be used as a confirmation tool for B1 and G1. A separation of the four aflatoxins and the hemiacetals B2a and G2a is shown in figure 7.17
Figure 7. Analysis of the aflatoxins G2, G1, B2, B1 and G2a and B2a (hemiacetals), with conditions as for figure 1
Sensitivity of the B1 can also be improved by formation of an iodine derivative10 or by modification of the flow cell to a cell filled with fine silica particles.11
Aflatoxin assay by LC-MS For highest sensitivity and selectivity we have investigated the use of mass spectrometry. The aflatoxin standard (not including M2 and M1) was diluted fivefold and 1 l was injected resulting in concentrations of 1 ppm for G1 and B1, and 300 ppb for G2 and B2 (figure 8). A further dilution of 1: 10 and the 1-l injection is shown in figure 9. Detection limits for G1, B2 and B1 are less than 50 ppb for this separation. An example of thermospray application is shown in figure 11.
HPLC Stationary phase Mobile phase Flow Gradient MS Tune parameters Source temperature SIM parameters Dwell time Electron multiplier Thermospray stem temperature Manual tune on 367 adduct ion for polypropylene glycol 250C Quadropole temperature 120 C SIM ions 312B1, 314B2, 328G1, 330G2, 286 and 284 fragments of G1 respectively G2 600 msec 2500 V On Mode negative 95 C Filament ON Hypersil ODS 100 x 2.1 mm, 3 l 799160D-352 Water-methanol-acetonitrile 163: 26: 11 ) 0.3 ml 32 %60 % B in 10 min
Figure 8. Analysis of G2, G1, B2 and B1 using HPLC and thermospray mass spectrometry
Figure 9. G1, B2, and B1 in the low picogram range at 1-l injection volume
Figure 10. Total ion monitoring spectrum of aflatoxin B1
Figure 11. Extract of pistachio nut according to 35 LMBG with 1-l injection (see also figure 12 with 2-l injection volume)
10
Automatic operation for routine applications Considering the toxicity of aflatoxins, most countries keep permitted concentrations low, for examples see table 2.
Aflatoxin B1 Germany 2 ppb B1 20 ppb B1 animal feed diet and baby food France Switzerland 5 ppb B1 1 ppb B1
Total aflatoxin 4 ppb {B1, B2, G1, G2} 50 ppt {B1 , B2, G1, G2}
Milk toxin 50 ppt M1 milk
5 ppb {B1, B2, G1, G2}
50 ppt M1 in milk 20 ppt M1 baby food 250 ppt M1 cheese
2 ppb B1 corn, cereals USA, FDA WHO, FAO 5ppb B1
10 ppt {B1 , B2, G1, G2} diet and baby food 20 ppb {B1, B2 , G1 , G2 } 10 ppb {B1, B2, G1, G2} 500 ppt M1 milk
Table 2. Limits for Aflatoxins in different countries
Figure 12. The analysis of a 2-l injection of pistachio-nut extract (35 LMBG) detected by FLD and DAD and UV-Spectrum
Sample preparation in the following applications was performed according to 35 LMBG. With diode-array detection, retention time and spectral information can both be incorporated automatically in the report. We created a library of standard mycotoxin spectra tagged with their HPLC retention times. After each run, peak spectra were automatically compared with library spectra, and their purity checked by overlaying several spectra taken in each peak. The customized
report prints all this information retention times, chromatogram, library and calibration table, amounts, library search match and purity match factor. The method is fully automatic. Data acquisition and data evaluation including quantification and qualitative identification are performed in one run. Figure 12 shows a 2-l injection of pistachio-nut extract detected using FLD and DAD. B1 is present in less than 1.0 ng (absolute) corresponding to 11 ppb (20 g of material extracted in 500 l
11
methanol of which 2 l was injected corresponds to a multiplication factor of 12500 for the ppb value). This is close to the detection limit for fluorescence, while for UV-visible absorbance much lower values are detectable. To determine lower concentrations by FLD, larger injection volumes are needed. Fluorescence has the advantage of high selectivity- and therefore no matrix effectswhereas the diode-array detector shows much higher sensitivity for B1 and G1 and can be used for additional confirmation using the automatic library search program. Figure 13 shows the printout of such a search. A report header contains all the important information, such as data file name, the library used, search threshold values, peak purity, while the quantitative report contains the corresponding retention times (from library, calibration table and chromatogram), purity and library match factors and names of the identified compounds.
Report from automatic Library search

***** REPORT ***** Operator Name: Date & Time: Data File Name: Integration File Name: Calibration File Name: Quantitation method: Sample Info: Misc. Info: Method File Name: Library File Name: Reference Spectrum: Time window from: Dilution Factor: 12.5 Name B1 Amount [g/kg] 11.48 11.48 Figure 13. Printout of the report of the analysis in figure 12 RS 10 Apr 92 4:06 pm FOOD:AFLA-PIST DATA:AFLA1.I FOOD:AFLA-DAD.Q ESTD calibrated by Area response PISTATIO NUTS according to $35 LMBG (reduced to 500 l MeOH) 20g extr. CH2CL2clean SilicagHEXANE, ETHER, CHC13 AFLA-DAD.M FOOD:AFLATOX.L 8.32 min 5.0 % to: 5.0 % Sample Amount: 0.0 Peak-Ret. [min] A 7.956 Call-Ret. [min] 7.18 Wavelength from: 250 to: 400 nm Library Threshold: 959 Peak Purity Threshold: 950 Smooth Factor: 5 Resp.Fact.uncal.peaks: None Lib.-Ret [min] 8.18 Purity Match factor 967 987 6.4 Library Res. Vial/Inj.No.: 100/1
A milk sample spiked with 600 ng/l of aflatoxin M1 was prepared according to 35 LMBG. 2 l were injected and detected with FLD and DAD (figure 14). The sub-nanogram amounts of M1 (0.6 ng) could be detected by DAD and a spectral search for confirmation was performed.
Figure 14. Analysis of spiked milk sample (600 ng/l) using sample preparation 36 LMBG
12
Ochratoxin A assay by HPLC Separation was achieved on a reversed phase column (LiChrospher 100 RP 18 125 x 4-mm id, 5 m particles) with water /2 % acetic acid / acetonitrile (1: 1) and detected at lex 247 nm, lem 480 nm with a fluorescence detector, 20-l injection volume and 40C column temperature. This analysis was confirmed with a derivatization of the mycotoxin to the methyl ester (figure 16). The analysis also works well in more complicated matrices, for example, figs (figure 17).
Figure 15. Analysis of ochratoxin A in wheat flour with the corresponding standard
Figure 16. Separation of ochratoxin A and the ochratoxin A methylester derivative (1 ng absolute)
Figure 17. Analysis of a fig extract where ochratoxin A has been derivatized and overlaid with the corresponding methyl ester standard
13
Zearalenone assay by HPLC-DAD and HPLC-FLD Separation was achieved on an Hypersil ODS narrow-bore column (100 x 2.1-mm id, 5-m particles) using a 50 parts water, 40 parts acetonitrile, 10 parts methanol isocratic mobile phase mixture. DAD detection wavelength was 236 nm with 20 nm bandwidth, fluoresecence detection was at lex 236 nm, lem 464 nm. Figure 18 shows a standard composed of 5 ng a-zearalenol, 2 ng b-zearalenol and 8 ng zearalenone. We recommend the DAD for sensitivity and spectral confirmation, while higher selectivity is given by fluorescence. Patulin assay by HPLC-DAD A major problem for the analysis of patulin in apple products (juice, pies and so on) is the high content of 5-hydroxy methyl furfural (HMF) a compound that elutes close to patulin and absorbs light in the ultraviolet region also. Separation of both HMF and patulin was achieved on a silicagel column and also on a diol column (a reversed phase column based on silica gel with two hydroxyl endings). Figure 18 shows a separation of the compounds in an apple juice sample on a diol column (conditions given in table 1). Patulin was detected at 270 nm with subsequent identification using spectral library search. With improved reverse phase column materials, separation can be done on Spherisorb RP 18, 5-m particles using an acetonitrile gradient from 5 % to 100 % at 40 C.16
Figure 18. Analysis of zearalenone and its metabolites with FLD and DAD detection
Figure 19. Resolution of patulin and 5-hydro methyl furfural (HMF) in apple juice overlaid with a 10-ng standard of patulin
14
Conclusion We have been able to show that with suitable sample preparation four classes of mycotoxins can be successfully quantified at nanogram levels in a variety of solid and liquid foodstuffs. Considering sample complexities, a variety of approaches are possible and we have discussed these at length. HPLC separations were performed on reversed phase materials. Derivatization and subsequent fluorescence detection can improve selectivity for aflatoxins and ochratoxin A, and serve as an additional confirmatory analysis. An alternative to confirmation by FLDUV-visible spectral libraries acquired on a diode-array detector can be incorporated in the analytical run and automated, generating a single comprehensive report. For most of the mycotoxins, fluorescence detection was used for high sensitivity. Mass spectrometry was able to lower detection limits to the low picogram range for aflatoxins, including confirmation via molecular mass.
References 1 "Verordnung uber Hochstmengen an Aflatoxinen in Lebensmitteln" vom 30.11.1976 (E GB1. I S. 3313) i.d.F. vom 06.11.1990 (BGB1. I S. 2443). 2 Bauer et al., Tierrtzl. Praxis 1987, 15, 3336. 3 Dwivedi et al., WPSA Journal 1986, 42(1), 3247. 4 Rodrigues et al., J. Pediatr. 1986, 107, 393397. 5 Enders, C., Dissertation, University of Munich, 1984. 6 Shotwell et al., Cer. Chem, 1975, 52, 687697. 7 Joint FAD/WHO Expert Committee on Food Additives, Report 763, 1988. 8 Amtliche Sammlung von Untersuchungsverfahren nach 35 Lebensmittel und Bedarfsgegenstndegesetz (LMBG) Methode 00.00-2 Methode 1,01.00-14/15. 9 Amtliche Sammlung von Untersuchungsverfahren nach 35 Lebensmittel und Bedarfsgegenstndegesetz (LMBG) Methode L 15-00-1.
10 N. D. Davis, U. L. Diener, J.A.O.A.C. 1980, 63, 107. 11 M. Blanc, Industries Alimentaires et Agricoles, 1980, 893. 12 T. Kuiper-Goodman, P. M. Scott, H. Watanabe, Toxicol. Pharmacol. 1987, 7, 255306. 13 R. Eppley, J.A.O.A.C., 1968, 51, 174. 14 Official methods of analysis for patulin, Mitt.-Geb.Lebensmittelunters.-Hyg., 1984, 75, 506513. 15 International Union of Pure and Applied Chemistry, Pure Appl. Chem. 1988, 60, 871876. 16 D. Ehlers, Lebensmittelchem. Gerichtl.-Chem. 1986, 40, 24. 17 Gregory Manley, J.A. O.A. C., 1981, 64, 144-151. 18 Official methods of analysis of the AOAC, 5th Edition, volumes 1 and 2, BSU publishers, USA, 1993, ISBN 09355 84 420.
Acknowledgments Martin Greiner, mass spectrometry specialist at Hewlett-Packard's European Customer Support Center, performed the MS work on aflatoxins.
15
Rainer Schuster is a senior application chemist based at Hewlett-Packard's Waldbronn Analytical Division in Germany. Dr Gerhard Marx is head of the analysis laboratory at the Landesuntersuchungsamt (State Analytical Services Agency) in Karlsruhe, Germany. Michael Rothanpt is a marketing development engineer based at Hewlett-Packard's European Marketing Center also in Germany. Correspondence should be addressed to one of the HP authors.
Hewlett-Packard shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Copyright 1993 Hewlett-Packard Company 8/93 12-5091-8692
16
GC/MS Approaches to the Analysis of Monochloropropanediol Application
Foods and Flavors
Author
Harry Prest Agilent Technologies, Inc. 1601 California Avenue Palo Alto, California 94301-1111 USA
monochloropropanediol byproduct was classified by the European Union's Scientific Committee for Food as a suspected carcinogen [1]. Although efforts were made to reduce the presence of 3-MCPD, continuing concerns about its presence lead to regulation of the allowable concentration. Recently the Association of Official Analytical Chemists (AOAC) has published a method for the extraction, separation and identification of 3-MCPD in foods and ingredients using gas chromatography with mass spectrometric detection [2]. In brief, a homogenized sample is mixed with a salt solution, then mixed with an Extrelut refill pack before being added to chromatographic column. The 3-MCPD is eluted with diethyl ether and a portion is derivatized with heptafluorobutyrylimidazole. Quantitation with GC-MS using electron impact ionization provides detection limits less than 0.01 mg/kg (which is equivalent to about 10 pg/L in the final extract at injection). This brief examines approaches to 3-MCPD as the heptafluorobutyryl-derivative described in the AOAC method using the Agilent 5973N MSD.
Abstract
The suspected carcinogen 3-chloro-1,2-propanediol (3-MCPD) is found in hydrolyzed vegetable protein, a widely used flavoring. Gas chromatography with mass spectrometric detection is a standard AOAC method. Electron impact ionization permits subpicogram measurement. Electron capture negative ionization is more selective and probably better suited to actual samples, with sensitivity of a few picograms in scan mode and less than 1 picogram in the selected ion mode.
Introduction
Hydrolyzed vegetable protein (HVP) is a widely used flavoring found in soups, sauces, and some meat products, etc. HVP is traded internationally both as solid and liquid depending upon the intended application. During the acid hydrolysis process of vegetable proteins, the hydrochloric acid agent reacts with triglycerides (Equation 1) to produce 3-chloro-1,2-propanediol (3-MCPD). This
O O R O O O O R" R' HCl OH OH OH HCl Cl OH OH
Experimental
3-MCPD liquid (Sigma Scientific, St. Louis, MO.) was diluted in dichloromethane (VWR Scientific, San Francisco, CA). An aliquot was added to a reaction vial containing 1 mL isooctane and derivatized with heptafluorobutyrylimidazole (Pierce, Rockford, IL) at 70 C for 30 minutes according to the procedure outlined in the AOAC method [2].
Equation 1. The acid hydrolysis of vegetable matter from triglycerides to glycerol, and then to 3-MCPD.

Derivatizing 3-MCPD with heptafluorobutyrylimidazole replaces the hydrogens on the diol groups with ester linkages to a perfluorinated propyl side chain. The molecular formula of the derivative is C3H5O2Cl(COC3F7)2, and Figure 1 shows the molecular structure. Figure 2 shows the electron impact (EI) ionization mass spectrum of the heptafluorobutyrylimidazole derivative of 3-MCDP from 60 to 510 m/z.
F
197 m/z
275 m/z
F F F F F O 169 m/z F
F O F F O 453 m/z 289 m/z -HCl 253 m/z Cl
Figure 1. The molecular structure and a suggested fragmentation pattern for heptafluorobutyrylimidazole derivative of 3-MCPD in electron impact ionization.
Abundance 100 90 80 70 60 50 40 30 20 10 0 100 150 200 250 300 350 400 450 500 m/z
169
69 197
253 289 100 453
Figure 2. Electron impact ionization mass spectrum of the 3-MCPD heptafluorobutyryl derivative at 70 eV for the 60 to 510 m/z mass range. The molecular ion [M]+ would be expected at 502 m/z.
Electron impact ionization produces a mass spectrum that lacks a molecular ion and has a base peak at m/z 169 from [C3F7]+ fragments. As seen from the fragmentation diagram of Figure 1, the ions at 169 m/z and 197 m/z contain no structural relevance to 3-MCPD and consequently can not be used to indicate 3-MCPD. This suggests the 453, 289, 275, and 253 m/z fragments, which contain 3-MCPD structure, be used for detection. In the AOAC collaborative study, several laboratories had difficultly detecting the 453 m/z fragment. This is not a problem for the 5973N due to the high-energy dynode arrangement and high transmission quadrupole which provide good signal for high molecular weight fragments. In fact, work with the standard showed good signal-to-noise for the 453 m/z ion in the scan mode even at only a few picograms injected. Selected ion monitoring using the four ions suggests detection at subpicogram levels is possible. In actual samples, matrix interferences may emerge and contribute to noise. However, the derivatization technique suggests applying electron capture negative ionization (ECNI) mass
spectrometry which provides more selective ionization than electron impact. Figure 3 shows the ECNI mass spectrum of the 3-MCPD derivative under standard conditions with methane buffer gas (tat is, source 150 C, methane at 2 mL/min). Unlike the EI results, the molecular ion (502 m/z) is detected, although at low relative intensity. Unfortunately like EI, the base peak at 213 m/z and next most abundant peak at 194 m/z, due to [OCOC3F7]+ and [OCOC2F7]+ fragments respectively, also contain no structural relationship to 3-MCPD. This leaves the 502, 482 and 446 m/z ions as good candidates for 3-MCPD detection and quantitation. Analysis of standards suggests that it would be possible to detect a few picograms in the scanning mode and less than one picogram in selected ion mode (SIM). Further optimization of ECNI is possible, such as a lower source temperature to take advantage of the low boiling point of the derivative, which may improve the spectrum and detection limits. The real advantage of ECNI is expected to be in typical food samples where the greater selectivity of ECNI will demonstrate a strong suppression of chemical noise and enhance method detection limits.
Abundance 100 90 80 70 60 50 40 30 20 10 0 160 180 200 220 240 260 280 300 320 340 360 380 400 420 440 460 480 500 m/z-->
213
194 482 179 233 446 502
Figure 3. Electron capture negative ion chemical ionization mass spectrum of derivatized 3-MCPD with methane buffer gas from 150 to 510 m/z. Note the presence of the molecular anion [M]- not seen in EI.
Acknowledgments
The author is grateful to Phil Wylie and Norman Low for their contributions to the manuscript.
References
1. Reports of the Scientific Committee for Food, Food Sciences and Techniques, 36th Series, European Commission, Luxembourg, Belgium. Brereton, P., et al., Determination of 3-chloro1,2-propanediol in foods and food ingredients by gas chromatography with mass spectrometric detection: Collaborative study. Journal of the AOAC International, 2001. 84(2): p. 455-465.
2.

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance or use of this material. Information, descriptions and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2001 Printed in the U.S.A. November 8, 2001 5988-4287EN
Analysis of Fumonisin Mycotoxins by LC/MS Application Brief

Agilent 1100 Series LC/MSD Foods, Environmental Friedrich Mandel
Introduction
The fungus Fusarium, which is known to infest corn and corn products, produces a group of mycotoxins called fumonisins. The toxicities of the most abundant fumonisins, B1-3, have been extensively studied, and a variety of species-specific toxicities have been published. These compounds may be carcinogenic to humans. Fumonisins are characterized by a 19-carbon aminopolyhydroxyalkyl chain that is diesterified with propane-1-2, 3- tricarboxylic acid. Analogues B1-3 show a difference in the number and position of the hydroxyl groups (Figure 1). Fumonisins B2 and B3 have the same molecular weight.
Most analytical methods exclude the detection of one or more of the known fumonisins. Traditional HPLC analysis requires the derivatization of the amino group. In this paper, we show that the Agilent 1100 Series LC/MSD can detect fumonisins without derivatization.

The fumonisin analogues were analyzed in scan mode at a high concentration (25 ng) to determine the molecular ion and confirming fragments. The initial conditions showed the molecular ion [M+H]+, but no significant fragment ions. Collision induced dissociation (CID) was used to generate more fragments for structural confirmation. Fumonisin B2 and B3, indistinguishable by their spectra, were easily separated chromatographically (Figure 2).
Experimental
The system comprised of an Agilent 1100 Series binary pump, vacuum degasser, autosampler, thermostated column compartment, diode-array detector (DAD), and LC/MSD. The LC/MSD used electrospray ionization (ESI). Complete system control and data evaluation were done on the Agilent ChemStation for the LC/MSD.
Figure 1. Structure of fumonisins.
5.2 27
200000 100000 0
Fumonisin B
200000 100000 0
Fumonisin B
200000 100000 0 200 300
Fumonisin B
400
500
600
700
800 m/z
Chromatographic Conditions Column: 150 x 2.1 mm Zorbax Eclipse XDB, C18, 5 m Mobile phase: A = 5 mM ammonium acetate in water, pH 3 B = acetonitrile Gradient: Start with 33% B at 8 min 60% B at 9 min 33% B Flow rate: 250 l/min Injection vol: 5 l Column temp: 40C Diode-array detector: Signal 220, 4 nm; reference 550,100 nm MS Conditions Source: Ion mode: Vcap: Nebulizer: Drying gas flow: Drying gas temp: Scan range: Step size: Peak width: Time filter: Fragmentor: ESI Positive 4000 V 30 psig 10 l/min 350C 120-820 amu 0.1 0.15 min On Variable 230 V (100-680) 100 V (680-800)
5.6 07 5.6 07
Figure 2. Mass spectra for fumonisin analogues.
The total ion chromatogram (TIC) shows very good sensitivity at 25 ng (Figure 3). To further improve sensitivity, the standards were run in the selected ion monitoring (SIM) mode.
Fumonisin B
Abundance
Fumonisin B
Fumonisin B
096.7
182.3
800000
472.6
400000
152.1
min
Retention Time (min)
Figure 3. Chromatographic separation of fumonisin analogues at 25 ng.
Figure 4 shows the extracted ion chromatograms for 250 pg of fumonisins in a corn extract. The mass spectra showing the molecular and fragment ions provide highconfidence identification and quantification.
1000000
m/z 723
732.3
Fumonisin B
0 250000
m/z 335
822.3
0 350000
m/z 707 Fumonisin B

386.7 842.6
Fumonisin B
3
0 120000
m/z 337
576.7
Chromatographic Conditions Column: 150 x 2.1 mm Zorbax Eclipse XDB, C18, 5 m Mobile phase: A = 5 mM ammonium acetate in water, pH 3 B = acetonitrile Gradient: Start with 33% B at 8 min 60% B at 9 min 33% B Flow rate: 250 l/min Injection vol: 5 l Column temp: 40C Diode-array detector: Signal 220, 4 nm; Reference 550,100 nm
min
Figure 4. SIM of molecular and fragment ions for fumonisins in spiked corn extract.
MS Conditions Source: Ion mode: Vcap: Nebulizer: Drying gas flow: Drying gas temp: SIM ions: Step size: Peak width: Time filter: Fragmentor:
142.6
ESI Positive 4000 V 30 psig 10 l/min 350C at 0 min, 334.4, 352.4, 370.4, 722.5 at 5 min, 336.4, 354.4, 706.5 0.1 0.15 min On Variable 230 V (334.5, 352.4, 370.4) 100 V (706.5, 722.5)
Conclusion
The Agilent 1100 Series LC/MSD is capable of detecting fumonisins at low levels without derivatization. Mass spectrometry allows specific and sensitive detection in complex matrices such as corn extract. Friedrich Mandel is an application chemist at Agilent Technologies, Inc.
Agilent Technologies shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance or use of this material. Information, descriptions and specifications in this publication are subject to change without notice. Copyright 1998 Agilent Technologies, Inc. All rights reserved. Reproduction and adaptation is prohibited. Printed in the U.S.A. April 2000 (23) 5968-2124E
Analysis of Poisoned Food by Capillary Electrophoresis
Tomoyoshi Soga and Maria Serwe Food
Abstract In cases of poisoning, analytical tools are needed to determine the identity of the toxins quickly and accurately. This enables healthcare professionals to administer appropriate treatment as quickly as possible and helps police to find those responsible. A rapid determination of anionic toxins in adulterated foods and beverages is possible using capillary electrophoresis (CE) with indirect UV detection. Cyanide, arsenite, arsenate, selenate, azide and other anions can be detected within 15 minutes, requiring only minimal sample preparation.
Conditions
1 Cyanide 5 Sulfate 9 Arsenate 13 Arsenite 2 Chloride 6 Azide 10 Phosphate 14 Lactate 3 Nitrite 7 Carbonate 11 Glutamate 4 Nitrate 8 Fluoride 12 Acetate Absorbance 8 [mAU] 3 5 10 11 12 25 12 4 6 7 9 13 14 A 20 B 15 C 10 5 0 -5 5 6 7 9 8 Time [min] 10 11 E D
Injection 6 s @ 50 mbar Capillary fused silica capillary total length 112.5 cm effective length 104 cm internal diameter 50 m Buffer Agilent Basic Anion Buffer Voltage -30 kV Temperature 30 C Detection signal 350/20 nm reference 275/10 nm
Figure 1 Analysis of cyanide and arsenite in food. A = anion standard (50 ppm each), B= Oolong tea (1:100 diluted with H2O), C= Oolong tea as in B, spiked with 100 ppm NaCN, D=curry (1:100 diluted with H2O, filtered through 0.22 m filter), E=curry as in D, spiked with 100 ppm NaAsO2
Experimental Anion analysis was performed using the Agilent Capillary Electrophoresis system equipped with diode-array detection and computer control via Agilent ChemStation. The analysis is based on the Agilent Forensic Anion Analysis Kit (part number 5064-8208). Prior to first use, a new capillary was flushed with run buffer for 15 minutes (at 1 bar). Between the analyses the capillary was flushed 2 minutes from the OutHome vial into waste, then 2 minutes from the InHome vial into waste. This procedure avoids baseline fluctuations as a result of buffer depletion. Buffer vials were replaced after 10 runs when using 2 ml vials, after 5 runs, when using 1 ml vials. Sample preparation consisted simply of dilution with water, or dilution and additional filtration through a 0.22 m filter, as indicated in figure 1. Results Figure 1 shows the analysis of food spiked with cyanide and arsenite. Depending on the results of this quick analysis, the sample can then undergo a more detailed analysis. The assay was linear over the range 10100 ppm with r2 > 0.999. The method detection limit was 510 ppm. For the analysis of curry, the repeatability for arsenite (n = 6) was 0.06 % RSD for migration time and 2.7 % RSD for peak area. For cyanide in Oolong tea the respective values were 0.13 % RSD for migration time (n = 10) and 4 % for peak area (sample diluted in 0.01 N NaOH). Other toxic anions that can be determined are arsenate, azide and selenate (which migrates between azide and carbonate). Compared to ion chromatography (IC), the advantages of CE for this type of analysis are the shorter analysis time and the minimal sample preparation needed for samples with a complex matrix (e.g. curry). Additionally, the analysis of azide and arsenate together with cyanide and arsenite is not possible in one run with IC.
Equipment
Agilent Capillary Electrophoresis system Agilent ChemStation Agilent Forensic Anion Analysis Kit
Tomoyoshi Soga is an application chemist at Yokogawa Analytical Systems Inc., Tokyo, Japan. Maria Serwe is an application chemist at Agilent Technologies, Waldbronn, Germany. For more information on our products and services, visit our worldwide website at http://www.agilent.com/chem
HPLC Analysis of Aflatoxines in Pistachio Nuts using HPLC
Angelika Gratzfeld-Heusgen Food
Abstract The following mycotoxins have been analyzed: aflatoxins G2, G1, B2, B1, M2, and M1, ochratoxin A, zearalenone, and patuline. Mycotoxins are highly toxic compounds produced by fungi. They can contaminate food products when storage conditions are favorable to fungal growth. These toxins are of relatively high molecular weight and contain one or more oxygenated alicyclic rings. The analysis of individual mycotoxins and their metabolites is difficult because more than 100 such compounds are known, and any individual toxin is likely to be present in minute concentration in a highly complex organic matrix. Most mycotoxins are assayed with thin-layer chromatography (TLC). However, the higher separation power and shorter analysis time of HPLC has resulted in the increased use of this method. The required detection in the low parts per billion (ppb) range 4,1, 2, 3 can be performed using suitable sample enrichment and sensitive detection. Sample preparation Samples were prepared according to official methods.2 Different sample preparation and HPLC separation conditions must be used for the different classes of compounds. The table on the next page gives an overview of the conditions for the analysis of mycotoxins in foodstuffs.
mAU 20 15 10 5 0 2 4 Time [min] 6 8 10 FLD:
em ex
Chromatographic conditions The HPLC method presented here for the analysis of mycotoxins in nuts, spices, animal feed, milk, cereals, flour, figs, and apples is based on reversed-phase chromatography, multisignal UVvisible diode-array detection, and fluorescence detection. UV spectra were evaluated as an additional identification tool.
365 nm 455 nm
DAD: 365 nm
Figure 1 Analysis of aflatoxins with UV and fluorescence detection
HPLC method performance Limit of detection 15 g/kg Repeatability of RT over 10 runs <0.12 % of areas over 10 runs <1.5 %
Column class Aflatoxins G2, G1, B2, B1, M2, M1 Matrix nuts, spices, animal feed, milk, dairy products
References
1. R. Schuster, G. Marx, G, M. Rothaupt, Analysis of mycotoxins by HPLC with automated confirmation by spectral library, Agilent Application Note 5091-8692, 1993. 2. Lebensmittel- und Bedarfsgegenstndegesetz, Paragraph 35, Germany. 3. Official Methods of Analysis, Food Compositions; Additives, Natural Contaminants, 15th ed; AOAC: Arlington, VA, 1990, Vol. 2.; AOAC Official Method 980.20: aflatoxins in cotton seed products; AOAC Official Method 986.16: Aflatoxins M1, M2 in fluid milk; AOAC Official Method 985.18: -zearalenol. 4. Farrington et. al., Food Additives and Contaminants, 1991, Vol. 8, No. 1, 55-64. 5. Official Methods of Analysis; Horwitz, W., Ed.; 14th ed.; AOAC: Arlington, VA, 1984; secs 12.01812.021.
Linearity of UV-visible DAD 1500 ng of fluorescence 30 pg to 2 ng

Chromatographic conditions
Sample preparation
extraction according to Hypersil ODS, 100 2.1 mm id, Para. 35, LMBG* 4, 5 3-m particles water/methanol/ACN (63:26:11) as isocratic mixture (100% B is recommended for cleaning the column flow rate: 0.3 ml/min at 25 C DAD: 365/20 nm Fluorescence detector (FLD): excitation wavelength 365 nm, emission wavelength 455 nm extraction according to Para. 35, LMBG acidify with HCl extract with toluene SiO2 cleanup elute toluene/acetic acid (9:1) extract with toluene Sep-pak cleanup elute toluene/acetone (95:5) AOAC 985.18:3 -zearalenol and zearalenone in corn cleanup on Extrelut silica gel cleanup elute toluene/ ethylacetate (3:1) Lichrospher 100 RP18, 125 4 mm id, 5-m particles water with 2 % acetic acid/ACN (1:1)* flow rate: 1ml/min at 40 C FLD: excitation wavelength 347 nm, emission wavelength 480 nm Hypersil ODS, 100 2.1 mm id, 3-m particles water/methanol/ACN (5:4:1) as isocratic mixture* flow rate: 0.45 ml/min at 45 C DAD: 236/20 nm FLD: excitation wavelength 236 nm, emission wavelength 464 nm Superspher RP18, 125 4 mm id, 4-m particles water 5 %95 % ACN flow rate: 0.6 ml/min at 40 C DAD: 270/20 nm or Lichrospher diol, 125 4 mm id, 5-m particles hexane/isopropanol (95:5) as isocratic mixture flow rate: 0.6 ml/min at 30 C DAD: 270/20 nm
Ochratoxin A
cereals, flour, figs
Zearalenone
cereals
Patuline
apple products
Equipment
Agilent 1100 Series degasser isocratic pump autosampler thermostatted column compartment diode array detector, fluorescence detector Agilent ChemStation + software
Angelika Gratzfeld-Heusgen is application chemist at Agilent Technologies, Waldbronn, Germany.
mAU 5
Pistachio nut
4 3 2 1 2 4 Time [min] 6 8 FLD DAD
Figure 2 Analysis of aflatoxins in pistachio nuts with UV and fluorescence detection
Contaminants Trace Metals
Multi-Element Determination of Heavy Metals in Dietary Supplements Using Collision/Reaction Cell ICP-MS Application
Foods
Authors
Emma Peachey, Ruth Hearn, and Selvarani Elahi LGC, Queens Road Teddington, Middlesex, TW11 0LY UK
well-balanced meals, our diets are at an increased risk of becoming nutrient deficient. Dietary supplements, sold in capsule, tablet, or liquid form offer a simple and convenient way to supplement our diets and reduce the risk of nutrient deficiency. In the UK, most products described as dietary supplements are regulated as foods and are subject to the general provisions of The Food Safety Act 1990 and secondary legislation on safety and labeling (Food Labeling Regulations 1996 [as amended] and the Food Supplements Directive 2002/46/EC). Past investigations have shown that some dietary supplements can contain elevated concentrations of metals and other elements. While some metals, such as selenium, iron, copper, chromium, and zinc [1], are essential at low concentrations, others, such as arsenic, cadmium, lead, and mercury, are toxic [2]. In the UK, arsenic is the only element for which there is legislation on permitted levels in foods. Arsenic is regulated under the Arsenic in Food (as amended) Regulations 1959, which states a limit of 1 mg/kg in general food [3]. Cadmium, lead, mercury, and tin in specific foods are regulated under Commission Regulation 1881/ 2006 [4]. It is important that up-to-date information on the levels of metals and other elements in dietary supplements consumed in the UK is obtained, in order to assess whether there is any risk to consumers. The purpose of this study was to provide the Food Standards Agency (FSA) with up-to-date and accurate information on the concentration of a suite of metals contained within dietary supplements consumed in the UK.
Abstract
Determination of 11 metals (including arsenic, chromium, mercury, iron, copper, nickel, zinc, selenium, lead, cadmium, and thallium) in UK-consumed dietary supplements was carried out using ICP-MS. The instrument used was the Agilent 7500ce, which is equipped with a collision/ reaction cell (Octopole Reaction System), and was operated in no-gas, helium, and hydrogen modes, all acquired within a single method. Samples were microwave digested with nitric acid/hydrogen peroxide/hydrofluoric acid and quantified using external calibration. The method was assessed by the analysis of two certified reference materials (LGC7160 and SRM1577b), and recoveries were 100 15% of the certified value for all elements.
Introduction
Heavy metals are natural components of the earths crust and are widely used in agricultural, construction, manufacturing, and food/material processing industries. As trace elements, some heavy metals are metabolically essential to humans at low levels, but at higher concentrations they can become toxic. Heavy metal toxicity can result from high ambient air concentrations near emission sources, drinkingwater contamination, or intake via the food chain. Conversely, as our lifestyles have become increasingly busy, leaving less time to prepare and eat
To do this, microwave acid digestion was used for sample preparation, followed by multi-element determination by collision/reaction cell inductively coupled plasma mass spectrometer (CRC-ICP-MS) using He and H2 collision gases to remove spectral interferences. The suite of metals included arsenic (As), chromium (Cr), iron (Fe), copper (Cu), nickel (Ni), zinc (Zn), selenium (Se), lead (Pb), mercury (Hg), cadmium (Cd), and thallium (Tl). Most elements were measured in standard mode (no gas) since multi-isotope ICP-MS data obtained for the food samples suggested the absence of significant polyatomic interferences. However, the ICP-MS detection of three of these elements, two essential (Fe and Se) and one toxic (As), was found to be strongly hampered by polyatomic ions. The purpose of this application is to demonstrate the ability of the Agilent 7500ce using the Octopole Reaction System (ORS) to eliminate these interferences (Table 1), providing accurate determination of these three elements in food supplements.
USA) containing 0.73 0.06 mg/kg Se and 184 15 mg/kg Fe and crab paste LGC7160 (LGC, Teddington, UK) containing 11 1 mg/kg As. Instrumentation Sample digestion was undertaken in a Mars 5 microwave (CEM, Buckingham, UK). Elemental measurements were performed using an Agilent 7500ce CRC-ICP-MS operating in hydrogen mode for Se and Fe, and helium mode for As to remove spectral interferences (Table 1). All other elements were measured in standard (no gas) mode within the same method. Typical operating conditions are illustrated in Table 1. The Integrated Sample Introduction System (ISIS) was used with a pump speed set at 0.1 rps during the analysis and washout in order to minimize the amount of matrix onto the interface and optimize sample throughput.
Table 1. Instrumental Conditions for 7500ce ORS Collision/Reaction Cell Mode ORS Cell Mode He As
40 40
Parameter
Experimental
Elements measured
H2 Se, Fe
38
Samples Two hundred different dietary supplements (either tablet, capsule, liquid, or powder form) commercially available in the UK were sourced. The average weight of each tablet/capsule was determined using an electronic balance. Sample Preparation Tablets were crushed with a pestle and mortar. Crushed tablets, liquids, and powders were subsampled after thorough mixing. Oil capsules were digested whole. Approximately 0.7 g of sample was accurately weighed and microwave digested with 7 + 3 + 0.2 mL of nitric acid + hydrogen peroxide + hydrofluoric acid, respectively. The microwave program consisted of heating the samples to 180 C over 20 min and holding for a further 10 min. Once cool, the digests were made up to 100 g using deionized water, and the resultant solutions were subjected to element determination by ICP-MS. Approximately 10% of the samples were digested in duplicate. A blank and QC material were included in each digestion run, containing a maximum of 12 samples. QC Materials Two certified reference materials were analyzed to assess the accuracy of the methodology. These were bovine liver SRM 1577b (NIST, Gaithersburg,
Spectral interferences removed by ORS gas
Ar35Cl+ on 75As+ Ca35Cl+ on 75As+
Ar40Ar+ on 78Se+ Ar37Cl+ on 77Se+ 40 Ca37Cl+ on 77Se+ 40 Ar16O+ on 56Fe+ 40 Ca16O+ on 56Fe+
40
RF power (W) Carrier gas (L/min) Make-up gas (L/min) Nebulizer Spray chamber Interface cones Cell gas Cell gas flow rate (mL/min) Points per peak Repetitions Integration time per mass (sec) He 2.5
1520 0.9 0.26 Glass concentric, MicroMist Quartz cooled to 2 C Ni H2 2.2 3 10 0.3
Measurement Five-point external calibrations with standards traceable to the National Institute of Standards and Technology (NIST, Gaithersburg, USA) were used to quantify the elements in the digests. Rhodium (Rh) was used as an internal standard and added on-line (1:1 with samples). The internal standard solution also contained 4% propanol to compensate for enhancement of As and Se signal from any residual carbon in the samples. Water
Research Councils (WRC) Aquacheck solutions (used in a proficiency testing scheme) with known concentrations of the analytes (660.3 ng/g Fe, 12.6 ng/g As, and 13.27 ng/g Se) were also analyzed as independent checks on the accuracy and precision of each ICP-MS run.
This makes it difficult for the FSA to accurately assess the dietary intake of these elements from such supplements based on label claim alone and demonstrates the necessity of this survey. The limits of detection (LOD) and quantification (LOQ) of the described procedure, calculated according to International Union of Pure and Applied Chemistry (IUPAC) guidelines [6], are presented in Table 3. For 25% of the dietary supplements tested the Se concentration was found to be less than the LOD.
Table 3. Limits of Detection and Quantification of Se, Fe, and As Se Concentration (mg/kg)1 Fe 0.072 0.240 As 0.006 0.022

All sample results were checked against legislative limits and/or limits agreed with the FSA. The action limits used were 1 mg/kg arsenic, lead, and cadmium and 0.5 mg/kg for mercury. Samples that exceeded these limits were reanalyzed to confirm the results. As illustrated in Table 2, five samples were confirmed to contain As concentrations greater than the 1 mg/kg limit recommended in The Arsenic in Food (as amended) Regulations 1959 [3]. A further two samples were found to contain As concentrations between 0.75 mg/kg and 1 mg/kg. It should be noted that several of these supplements were derived from marine animals, and the form of the As in such materials is likely to be nontoxic arsenobetaine. The majority of supplements (> 75%) were found to contain < 0.1 mg/kg As and Se, and > 20 mg/kg Fe.
Table 2. Samples with Arsenic Concentrations Above Recommended Limit As in sample Product Form (mg/kg) Product 1 Product 2 Product 3 Product 4 Product 5 Tablet Capsule Capsule Capsule Capsule 3.3 0.7 2.5 0.6 20.5 4.8
LOD LOQ
1
0.009 0.029
Values shown are based on the average weight of dietary supplement tablet/ liquid/capsule digested (0.67 g).
Recovery results for the QC materials and Aquacheck solutions were very good, with results for all elements falling within 100 15% (n = > 9) of the certified/expected value. The recoveries obtained for the reference materials analyzed are illustrated in Figure 1. A number of sample and blank solutions were also spiked with Se, Fe, and As prior to microwave digestion. The resultant recoveries fell within 100 10% (n = > 5) of the expected value. As a check on the repeatability of
120
100
7.3 1.7
Recovery of certified value (%)
1.5 0.4
80
Legislative limit = 1 mg/kg [3] The uncertainty quoted is the expanded uncertainty calculated using a coverage factor of 2, which gives a level of confidence of approximately 95%. The uncertainty was calculated based upon the principles of the Eurachem Guide [5].
60
40
None of the samples was found to contain an elevated concentration of Cd, and only one sample was close to the limit for Hg. Ten samples were confirmed to contain concentrations of Pb above 1 mg/kg. Results in mg/kg were calculated back to mg/tablet in order to allow comparison with label claims. For a number of supplements, differences were found between the measured concentrations of Se, Fe, Zn, Cu, Cr, or Ni and the values stated on the label.
20
0 Se Fe Element NIST bovine liver LGC crab paste As
Figure 1.
Recovery of Se, Fe, and As from the CRMs analyzed.
the method, approximately 10% of the samples were also independently digested and analyzed in duplicate. For these, coefficient of variation values of < 5% were found for all three elements measured in collision/reaction mode on the 7500ce ICP-MS. 3. The Arsenic in Food Regulations 1959 (S.I. [1959] No. 831), as amended by The Arsenic in Food (Amendment) Regulations 1960 (S.I. [1960] No. 2261) and The Arsenic in Food (Amendment) Regulations 1973 (S.I. [1973] No. 1052). The Stationery Office. 4. Commission Regulation (EC) No 1881/2006 of 19 December 2006 setting maximum levels for certain contaminants in foodstuffs. 5. Eurachem, Quantifying uncertainty in analytical measurement. Laboratory of the Government Chemist. London 1995 (ISBN 0 94892608). 6. A Statistical Overview of Standard (IUPAC and ACS) and New Procedures for Determining the Limits of Detection and Quantification: Application to Voltammetric and Stripping Techniques, Pure & Appl. Chem. Vol 69, No. 2, pp 297328, 1998 7. Food Standard Agencys Food Surveillance Information sheet No 85/05 December 2005, Survey of Metals and Other Elements in Dietary Supplements.
Conclusions
Microwave digestion followed by analysis by ICPORS-MS has been shown to be a simple, reliable method for the multi-element determination of trace metals in nutritional supplements and foodstuffs. A number of supplements were found to contain Se, Fe, Zn, Cu, Cr, and Ni at concentrations that deviated from the label claim. Five samples for As and 10 samples for Pb were found to contain elevated concentrations above the recommended 1 mg/kg limit. The data generated in this survey has provided the FSA with up-to-date concentrations of these metals in a range of dietary supplements. The results have enabled the risk of metal toxicity from the consumption of dietary supplements to be assessed and published in a Food Standard Agencys Food Surveillance Information sheet [7].
References
1. World Health Organization, Trace Elements in Human Nutrition and Health, Geneva, 1996 2. The Ministry of Agriculture, Fisheries and Food (1982), Survey of Arsenic in Food. Food Surveillance Paper No. 8, Published by HMSO
Acknowledgements
The authors would like to thank Malcolm Burn and Kam Lee from the Food Chemistry team at LGC for performing the sample preparation/digestion, and Sheila Merson, Linda Evans, and Dave Curtis from the Specialised Techniques team for conducting the elemental determinations.

Measurement of Trace Elements in Malt Spirit Beverages (Whisky) by 7500cx ICP-MS Application
Food
Author
Glenn Woods Agilent Technologies UK Ltd. Lakeside Business Park Cheadle Royal, Cheshire, SK8 3GR UK
of trace elements can also significantly affect the taste of the whisky. Consequently, there is a requirement to measure elemental concentrations in the final product. While ICP-MS offers high sensitivity and excellent detection limits for many elements, interferences on key elements arising from the alcohol content and required sample preparation can be problematic. The 7500cx features the Octopole Reaction System (ORS) collision/reaction cell, which removes matrix-based polyatomic interferences using a single set of cell conditions (helium mode). For the analysis of spirits, the major interferences resulting from the sample would be carbon-based (for example, 40Ar12C on 52Cr). Many elements are much more stable in a chloride matrix than simple acidification using nitric acid; for this reason, hydrochloric acid (HCl) was added to the samples. New interferences are created by the addition of HCl (for example, 35Cl16O on 51V; 40Ar35Cl on 75As, etc.) but they are removed by the ORS in helium mode. An optional cell gas line is available for the 7500cx, enabling operation in hydrogen (H2) reaction mode, which allows for the measurement of selenium at ultratrace levels. Since several of the solutions contained less than 40 ng/L Se (some significantly lower than this) in solution after dilution, H2 reaction mode was also used during this study.
Abstract
A method for the measurement of trace elements in malt spirits (whisky) is described with reference to six different samples. An Agilent 7500cx ICP-MS featuring the Octopole Reaction System (ORS) was used for the analysis. The 7500cx ensures simple operation as a single method and a single set of conditions can be used to remove interferences regardless of their source. Excellent spike recoveries were obtained (between 97 and 107%) following a simple dilution of the samples. A 5-hour stability test yielded excellent precision (< 2%) for almost all elements. The study shows that the 7500cx can be used for the routine measurement of trace metals in beverages.
Introduction
The measurement of trace elements in alcoholic beverages is required from a quality control standpoint and also to ensure that the final product complies with any regulatory requirements. Metal content can originate from the raw ingredients, such as water or grain, as well as during processing, for example, from fermentation or distillation equipment. An example would be high arsenic concentration from distillation vessels manufactured from poor-quality copper. The levels
Experimental
Sample Preparation & Instrumental Conditions
Four Scottish whiskies (Highland, Speyside, Islay, and a blend), one Irish whisky, one U.S. bourbon
as well as a further Scottish whisky and a U.S. bourbon that had been stored in lead crystal decanters were analyzed. The samples were prepared by simple 5x dilution using 1% HNO3 and 0.5% HCl (v/v). Using an acid mix significantly improves the stability of many elements, particularly Hg and Sn, compared to the use of nitric acid alone. Standards were prepared from 1,000 ppm stock single-element solutions to produce final mixed-element calibration solutions. In order to compensate for sample transport effects and solvent evaporation rates, the alcohol content of the standards was matched to that of the samples by adding 8% ethanol to all standard solutions (equivalent to 5x dilution of the original samples, which contained 40% v/v alcohol). This also compensates for ionization enhancement effects for As and Se in the presence of high carbon concentrations. Gold (400 g/L) was also added to the standards and samples in order to further improve the stability of Hg. Table 1 lists the instrumental conditions used for the analysis; sample uptake rate was approximately 150 L/min and sampling was facilitated by the Agilent ASX-520 autosampler. The solution pump program was optimized using the preemptive rinse function in the ChemStation software in addition to a multichemistry rinse regime [1]. The 7500cx was operated under standard conditions, and internal standards (Ge, Rh, and Ir) were added automatically on line by the systems peristaltic pump. No special precautions were necessary for these sample types.
Table 1. RF power Sampling depth Carrier gas flow Makeup gas flow Spray chamber temperature Helium cell gas flow Hydrogen cell gas flow Agilent 7500cx Operating Conditions 1550 W 8 mm 0.68 L/min 0.33 L/min 15 C 5.5 mL/min 4.0 mL/min
As each collision causes energy loss, the interfering species lose more energy than the analyte and are subsequently filtered from the mass spectrum by discriminating between the two different energies (called energy discrimination). As this process takes place regardless of the analyte-interference combination, a single set of conditions can be used for all analytes. Selenium was measured in hydrogen mode as the concentration of this element in the diluted sample was at low ppt levels. Although selenium can be measured in helium mode, hydrogen mode removes the Ar-based interference with greater efficiency, improving the detection limit for this element, and is the better option for low-ppt concentrations. Some isotopes were determined in both helium mode and no-gas mode to provide comparative data on cell performance. For routine analysis this would not be necessary, of course. All cell modes were acquired within a single acquisition and sample pass.

Table 2 summarizes the detection limits (DLs), background equivalent concentrations (BECs), and calibration regression for the isotopes studied in the different cell modes (default mode is highlighted in bold typeface). For those elements that suffer from interferences in this carbon and chloride matrix, BECs and DLs are severely compromised when operating the instrument in no-gas mode (that is, conventional ICP-MS). This can be clearly observed in the data for chromium: 52Cr BEC without cell gas is 526 g/L, and in helium mode is 0.07 g/L. The interference is effectively reduced to background contamination levels as the BECs for both Cr isotopes are very similar. Improvements can also be observed for V, Fe and 65 Cu (63Cu does not suffer from interferences in this relatively simple matrix), all of which were acquired in helium mode. Figures 1 to 3 display the calibration profiles for selected interfered elements with and without cell gas applied; Figures 4 and 5 illustrate the calibration profiles for Be (low mass, difficult to ionize) and Hg (high mass, difficult to ionize, low-abundance isotope). The line does not pass through the origin in the calibrations for those elements that suffer from an interference and this offset can be seen clearly. Be and Hg are also presented to demonstrate the excellent sensitivity for these difficult-to-ionize elements. In order to obtain low detection limits, it is essential to maximize the ion-
Data Acquisition Data was acquired operating the ORS in helium [He], hydrogen [H2], and no-gas modes. Helium mode is the default mode of operation of the 7500cx. The inert He cell gas conditions remove interferences based on their ionic cross-section rather than relying on a reactive gas. As almost all interferences in ICP-MS are polyatomic in nature, they possess a greater cross-section than the monatomic analyte at the same mass and therefore undergo a greater number of collisions in the cell.
ization efficiency of the plasma. This is done through optimization of the sample introduction system (low solution and gas flow rates and widebore injector torch) and plasma generator design (27.12 MHz, solid-state fixed frequency, and highefficiency digital drive). All of these factors combine to increase the effective central channel temperature, improving ionization efficiency. This is allied to an ion lens system designed to improve low-mass ion transmission efficiency, further improving the DL of this important and relatively difficult element. The elements Ge, Rh, and Ir were used as internal standards and were added on line. Table 3 displays the quantitative data for all samples, including a spike recovery for the Islay whisky. Data are displayed in the preferred cell gas mode (usually helium). Although some elements were calibrated under gas and no-gas conditions, only the most appropriate cell mode is displayed to
simplify the data set. Taking Cr as an example, the data for both isotopes did not match in no-gas mode and were significantly higher than the data obtained in helium mode due to the intensity of the C- and Cl-based interferences. The helium mode data for both Cr isotopes produced comparable results, which is a good indication of the accuracy of the data. The two samples that had been stored in lead crystal decanters have obviously higher Pb concentration in comparison to the other samples. The mean Pb concentration in the noncrystal samples was about 1.3 g/L, while the Pb content of the samples stored in the crystal decanters was almost 10x higher. As a comparison, the UK maximum permissible Pb concentration in drinking water is 25 g/L (at the tap), which means that the Pb concentration is within this guideline; however, in 2013 this level is due to be reduced to 10 g/L, meaning products stored in crystal would fail to meet drinking water quality standards.
Table 2. Element Be V V Cr Cr Cr Cr Mn Mn Fe Fe Co Co Ni Ni Cu Cu Cu Cu Zn Zn As As Se Se Cd Cd Sn Sb Ba Hg Pb U
Limit of Detection, Background Equivalent Concentrations, and Regression Coefficients for the Studied Isotopes (Data are presented as ng/L [ppt] and are corrected for dilution.) Mass 9 51 51 52 52 53 53 55 55 56 56 59 59 60 60 63 63 65 65 66 66 75 75 78 78 111 111 118 121 137 201 208 238 Mode No gas He No gas He No gas He No gas He No gas He No gas He No gas He No gas He No gas He No gas He No gas He No gas H2 No gas He No gas No gas No gas No gas No gas No gas No gas r 1 1 0.9999 1 0.985 1 0.9997 1 1 0.9999 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 DL 0.5 26.8 495 50.4 48600 38.6 2020 7.8 17.5 17.8 1160 0.5 5.6 13 15.2 10.4 16.6 18.1 58.6 33.9 22 2.0 46.1 3.6 135 5.7 3.3 7.8 7.1 2.8 1.7 2.1 0.1 BEC 0.3 12.2 5220 73.7 526000 71.3 52100 20.8 30.4 406 58300 3.7 6.7 38.7 73.6 41.5 59.6 36.9 230 119 171 3.8 382 13 1390 5.3 6.1 50.5 30.8 5.3 10.7 10.2 0.2 3
Figure1A. Chromium calibration [He mode]. Note BEC of 0.0504 g/L.
Figure 1B. Chromium calibration [no-gas mode]. Note BEC of 526 g/L.
Figure 2A. Vanadium calibration [He mode]. Note BEC of 0.0122 g/L.
Figure 2B. Vanadium calibration [no-gas mode]. Note BEC of 4.57 g/L.
Figure 3A. Selenium calibration [H2 mode]. Note BEC of 0.013 g/L.
Figure 3B. Selenium calibration [no-gas mode]. Note BEC of 1.4 g/L.
Figure 4.
Beryllium calibration [no-gas mode]. Note detection limit of 0.303 g/L.
Figure 5.
Mercury calibration [no-gas mode]. Note detection limit of 0.0107 g/L. 5
Table 3.
Quantitative Data Obtained in the Preferred Cell Mode for the Spirit Samples with Spike Recovery for the Islay Sample (Recoveries were generally excellent. All results presented as dilution-corrected g/L.)
Bourbon Whisky decanter 0.048 0.14 4.331 4.114 30.54 67.03 0.368 1.992 367.6 355.9 21.9 0.424 0.293 0.028 15.12 0.188 2.41 0.009 11.15 0.049 Islay spike 13.09 25 30.95 30.81 38.19 232.2 12.83 25.91 579.8 568.6 137.5 25.72 26.54 12.55 41.3 24.87 25.71 0.252 25.33 24.38 Spike qty 12.5 25 25 25 25 125 12.5 25 125 125 125 25 25 12.5 25 25 25 0.25 25 25 % recovery 104.4 98.6 107.5 107.4 98.7 106.0 101.2 97.9 100.3 101.6 103.3 101.8 105.4 100.3 98.5 97.9 97.1 97.0 97.7 97.3 Highland 9 51 52 53 55 56 59 60 63 65 66 75 78 111 118 121 137 201 208 238 No gas He He He He He He He He He He He H2 He No gas No gas No gas No gas No gas No gas 0.140 1.564 27.39 26.15 54.51 1125 1.097 14.02 542.9 525.5 21.02 0.503 0.458 0.036 9.18 0.817 3.282 0.013 1.13 0.295 Speyside 0.052 0.443 14.05 13.7 31.76 191.2 0.376 3.586 370.8 359.2 18.54 0.427 0.357 0.024 14.82 0.514 3.05 0.011 0.898 0.049 Islay 0.037 0.344 4.064 3.955 13.52 99.76 0.180 1.442 454.4 441.6 8.414 0.272 0.190 0.012 16.68 0.397 1.426 0.010 0.903 0.051 Blend 0.008 0.073 12.62 12.57 12.22 583.8 0.130 5.065 258 251.4 14.18 0.256 0.073 0.010 5.161 0.308 2.001 0.011 1.902 0.026 Irish 0.035 0.431 22.71 22.61 26.95 250.5 0.336 3.078 38.45 37.4 8.149 0.164 0.045 0.024 2.245 0.311 3.37 0.010 1.21 0.060 Bourbon 0.015 6.321 4.077 3.661 9.753 131.4 0.172 2.274 22.2 21.43 13.69 2.192 0.497 0.036 1.681 0.765 3.303 0.018 0.912 0.104 decanter 0.042 1.693 31.87 31.52 90.4 1114 0.323 12.88 445.5 430.8 68.27 0.434 0.069 0.193 0.239 0.316 1.396 0.008 12.59 0.028
Sample Be V Cr Cr Mn Fe Co Ni Cu Cu Zn As Se Cd Sn Sb Ba Hg Pb U
The benefit of operating the instrument in helium mode can clearly be observed for those isotopes suffering from interferences. As helium is a totally inert gas, no side reactions or new product interferences are formed this lends itself to full mass acquisition allowing interference-free qualitative or semiquantitative analysis. The samples were prepared in an identical way as above (although a separate preparation on a different day) and Table 4 displays the semiquantitative data obtained for the samples analyzed under identical helium cell conditions as with the previous data set. The full mass spectrum (Figure 6) is from the crystal-stored bourbon sample. The graphic includes an inset, zoomed-in region to demonstrate the excellent isotopic fit for those elements suffering most from interferences. The fit for Cr is particularly important as all three isotopes (50, 52, and
53) demonstrate good agreement with the expected natural ratio in this carbon-based matrix (50Cr suffers interferences from 38Ar12C, 13C37Cl, 36 Ar14N, and 35Cl15N; 52Cr has interferences from 36 Ar16O, 40Ar12C, 35Cl17O, and 37Cl15N; 53Cr has interferences from 40Ar13C, 37Cl16O, 35Cl18O, and 35 Cl17O1H). Several other interferences are also possible, but each is polyatomic in nature and so is removed by the same process and using a single set of helium mode conditions. To demonstrate instrument stability (Figure 7), 54 separate measurements were made of a spiked Highland malt whisky sample; total measurement time was 5 hours 18 minutes. Stability for the majority of elements was < 2% RSD over the run, indicating applicability of the method to routine analysis.
Table 4.
Semiquantitative Data for Spirit Samples Using Helium Mode (Data are presented as g/L [ppb] unless indicated and are corrected for dilution.) Highland Speyside
0.2 N/D 44 55000 ppm 2100 48 1.8 1400 9.7 190 440 ppm 150 49 1.6 0.54 14 32 210 0.47 3.5 380 18 0.45 0.3 0.48 150 1.4 0.84 0.03 0.019 0.0076 0.31 0.02 N/D 0.027 0.047 18 0.3 0.38 0.42 0.15 3.2 0.07 0.24 0.028 0.12 N/D 0.0083 0.04 0.005 0.0096 0.0057 0.081 0.003 0.013 0.00096 0.0032 N/D 0.065 0.0049 N/D 0.026 0.019 0.084 0.95 0.0088 0.045
Islay
0.049 N/D 42 55000 ppm 1600 23 1.8 1300 6.2 120 420 ppm 100 5.5 0.77 0.11 5.1 13 100 0.16 1.8 450 9 0.27 0.21 N/D 140 0.92 0.22 0.024 0.12 0.005 0.37 N/D N/D 0.0034 0.047 20 0.27 0.38 0.65 0.026 1.3 0.063 0.17 0.019 0.045 0.01 0.0055 0.048 0.0025 0.024 0.0057 0.081 0.004 0.0043 N/D 0.0032 0.0019 0.11 0.012 0.01 0.0044 0.029 00.064 0.94 0.012 0.044
Blend
0.084 N/D 49 55000 ppm 1600 30 2.3 1300 11 240 420 ppm 130 10 0.51 0.6 13 14 630 0.18 5.7 260 15 0.39 0.2 N/D 130 1 0.36 0.0084 0.064 0.0051 0.41 0.021 0.0075 0.0035 0.024 6 0.3 0.19 0.41 0.0092 2.2 0.086 0.11 0.02 N/D 0.033 0.011 0.016 N/D 0.029 0.0011 N/D 0.001 0.0089 N/D 0.0033 N/D 0.13 0.0051 0.0051 0.018 0.029 0.04 2 0.0079 0.023
Irish
0.16 N/D 62 54000 ppm 1000 63 2.4 1300 86 250 420 ppm 320 37 1.5 0.6 23 27 260 0.36 3.4 39 8.9 0.66 0.19 N/D 120 2.2 1.4 0.1 0.22 0.091 0.33 N/D N/D 0.014 0.024 3.7 0.39 0.19 0.5 0.063 3.8 0.28 0.61 0.063 0.3 0.098 0.016 0.11 0.012 0.083 0.012 0.046 0.0061 0.013 N/D 0.019 0.021 0.07 0.0025 0.005 0.013 0.059 0.041 1.3 0.009 0.073
Bourbon
0.16 N/D 69 76000 ppm 12000 120 1.9 2000 200 520 430 ppm 470 18 1.7 6.5 5.1 11 150 0.24 2.4 22 14 0.52 2 N/D 150 8.1 1.7 0.034 0.093 0.014 1.5 N/D 0.025 0.0039 0.1 2.2 0.8 N/D 0.89 0.24 3.8 0.15 0.24 0.02 0.14 0.037 0.0063 0.056 0.0029 0.028 0.0013 0.0075 0.0046 0.02 N/D N/D 0.0011 0.31 0.0058 0.011 0.0051 0.022 0.039 0.93 0.027 0.094
Bourbon decanter
0.34 N/D 75 55000 ppm 540 45 2.5 1500 33 210 410 ppm 460 26 0.99 1.2 27 90 1100 0.4 13 440 71 0.3 0.4 N/D 150 6.1 0.57 0.0056 0.069 0.0077 0.76 0.01 N/D 0.021 0.14 0.6 0.3 N/D 0.45 0.045 1.3 0.035 0.04 0.0044 0.022 0.032 0.0056 0.016 N/D 0.0049 0.0046 0.0032 0.001 0.0044 0.00097 0.0066 0.0019 0.07 0.0076 0.005 0.017 0.039 0.038 12 0.0056 0.032
Whisky decanter
0.26 N/D 51 57000 ppm 2100 48 1.6 1400 18 250 430 ppm 160 22 0.8 0.21 5.4 32 76 0.4 2.3 380 25 0.51 0.42 0.47 150 1.5 0.78 0.035 0.0097 0.005 0.13 0.02 N/D 0.01 0.071 17 0.21 0.18 0.47 0.15 2.4 0.087 0.36 0.024 0.14 0.032 N/D 0.024 0.0037 0.029 0.0034 0.078 0.001 0.017 0.00096 0.0032 0.00098 0.077 N/D 0.015 0.017 0.058 0.052 10 0.0067 0.05
7 9 11 12 23 24 27 29 31 34 35 39 43 47 51 52 55 56 59 60 63 66 69 75 78 79 85 88 89 90 93 95 101 105 107 111 118 121 125 127 133 137 139 140 141 146 147 153 157 159 163 165 166 169 172 175 178 181 182 185 189 195 202 205 208 232 238
Li Be B C Na Mg Al Si P S Cl K Ca Ti V Cr Mn Fe Co Ni Cu Zn Ga As Se Br Rb Sr Y Zr Nb Mo Ru Pd Ag Cd Sn Sb Te I Cs Ba La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Hf Ta W Re Os Pt Hg Tl Pb Th U
0.24 0.15 46 53000 ppm 1400 61 3.9 1300 13 100 450 ppm 150 60 0.51 0.8 26 54 1200 1.1 13 530 22 0.65 0.44 N/D 160 0.9 1.3 0.046 0.18 0.0024 0.7 0.01 0.0072 0.017 N/D 10 0.32 N/D 0.46 0.052 4.6 0.16 0.47 0.042 0.21 0.074 0.0027 0.071 0.0024 0.071 0.0045 0.1 0.0059 0.055 0.0018 0.0096 0.0038 0.11 0.012 0.0049 0.021 0.057 0.1 1.2 0.02 0.26
Figure 6a. Full scan (in He mode) of lead-crystal stored bourbon sample. The "major" peaks are indicated, including spectral fit for higher intensity peaks.
Figure 6b. Zoomed spectrum for those elements suffering from interferences in this matrix. Note good spectral fit, particularly for Cr (suffers from ArO, ArC, and ClO interferences).
318 min stability spiked whisky
Concentration (ppb) 900 800 700 600 500 400 300 200 100 0 Be V Cr Cr Mn Fe Co Ni Cu Cu Zn As Se Cd Sn Sb Ba Pb U 0 10 20 Sample repeat 30 40 50
Figure 7.
Stability for a spiked Highland malt sample (dilution corrected) taken over 5 hours 18 minutes (54 repeat measurements). Measurement precision was < 2% for almost all elements.
Conclusions
The analysis of high percentage alcoholic beverages using the 7500cx ICP-MS is routine after a simple acidification/dilution. The use of the ORS in the appropriate gas mode efficiently removes the plasma-based and matrix-based interferences, improving detection limits and reliability of the analysis with a simple set of conditions. The use of helium mode also allows interference-free semiquantitative analysis, permitting greater elemental coverage and rapid screening.

References
1. Achieving Optimum Throughput in ICP-MS Analysis of Environmental Samples with the Agilent 7500ce ICP-MS, Agilent ICP-MS Journal 27, page 4; May 28, 2006, 5989-5132EN
Determination of Organic and Inorganic Selenium Species Using HPLC-ICP-MS
Application
Environmental
Authors
Mat Bueno, Florence Pannier, and Martine Potin-Gautier Laboratoire de Chimie Analytique Bio Inorganique et Environnement Universit de Pau et des Pays de l'Adour, 64000 Pau France Jrme Darrouzes Agilent Technologies France
Abstract
A methodology based on coupling isocratic high-performance liquid chromatography (HPLC) and inductively coupled plasma mass spectrometry (ICP-MS) with optimized collision/reaction cell conditions has been developed for the simultaneous analysis of organic and inorganic selenium species in natural water samples. Selenium concentrations found in total and speciation analysis of a number of water samples showed good agreement. Because HPLC-ICP-MS coupling is easily automated, the method can be considered robust and applicable to the routine monitoring of selenium species in environmental and nutritional samples.
chemical forms in which an element can exist, that is, in the determination of its speciation. Indeed, knowledge of total concentrations of elements is not sufficient to assess their effects on human health or the environment. Among the elements of concern, there is a growing interest in selenium. Selenium is a very important element from an ecotoxicological point of view due to the narrow concentration range between its essential and toxic effects. Selenium compounds are distributed throughout the environment as a result of human activities (industrial and agricultural uses) and natural processes (weathering of minerals, erosion of soils, and volcanic activity). In waters, concentrations can vary from 2 ng/L to 1,900 g/L depending on the system [1]. The natural cycle of selenium shows its existence in four oxidation states (-II, selenide; 0, elemental selenium; +IV, selenite; and +VI, selenate) and in a variety of inorganic and organic compounds. The organically bound Se(-II) compounds include seleno-amino acids and volatile forms (dimethylselenide and dimethyldiselenide), which are less toxic relative to other species and result from various detoxification pathways. The toxic dose of selenium as a function of its chemical form is shown in Table 1.
Introduction
In the last 20 years, there has been increasing interest in the determination of the different
Table 1. Compound
Selected Selenium Compounds and Their Toxicity Formula (CH3)2Se H2Se (CH3)3Se+ [HO2CCH(NH2)CH2Se]2 CH3Se(CH2)2CH(NH2)CO2H SeO SeO
2 3 2 4
Lethal doseLD-50* 1600 mg/kg (Int.) 0.02 mg/L (Resp.) 49 mg/kg (Int.) 35.8 mg/kg (Or.) 4.3 mg/kg (Int.) 3.5 mg/kg (Int.) 5.8 mg/kg (Int.)
Ref. [2] [3] [3] [4] [3] [5] [5]
Dimethylselenide (II) Hydrogen selenide (II) Trimethylselenonium (II) Selenocystine (I) Selenomethionine (II) Selenite (+IV) Selenate (+VI)
*Lethal doses obtained on mice or rats by intraperitoneal (Int.), oral (Or.), or respiratory (Resp.) absorption.
A number of analytical procedures exist for the determination of selenium and its various species in samples from different environmental sources. Existing methods can be divided in three groups, depending on selenium concentration: Total selenium Selenite species Species including inorganic and organic forms of selenium Various redox reactions are often used to determine selenite species. However, the series of required reagents and pretreatment steps increases the possibility of element loss and contamination. Speciation results can also be distorted as back-oxidation of selenite to selenate may occur during sample pretreatment. Moreover, selenite and selenate are distinguished by two separate analyses, which is not the case for individual organic selenium species that remain unidentified. Hence, methods able to separate and quantify different selenium species simultaneously, in a single analysis, are preferred and are becoming more widespread. In this application, the coupling of high-performance liquid chromatography (HPLC) with inductively coupled plasma mass spectrometry (ICP-MS) is presented for selenium speciation analysis with emphasis on its application to natural water samples. Instrumentation A 7500ce ICP-MS from Agilent Technologies (Tokyo, Japan), equipped with an Octopole Reaction System (ORS) cell, was used for this study; see Table 2 for operating parameters. The sample introduction system consisted of a concentric nebulizer (Meinhard Associates, California, USA) and a Scott double-pass spray chamber cooled to 2 C. Nickel sampler and skimmer cones were used.
2
Table 2.
Instrumental Parameters for Agilent 7500ce ORS ICP-MS Value 1590 W 15.0 L/min 0.86 L/min 11.1 L/min 2 C 400 ms 77 to 82 100 ms
Parameter RF power Ar plasma gas flow Ar auxiliary gas flow Ar nebulizer gas flow Spray chamber temperature Integration time per isotope for speciation analysis m/z ratio monitored Integration time per isotope for elemental analysis
Chromatographic separation was carried out using the Agilent 1100 Series HPLC pump, equipped with an autosampler and variable volume sample loop. The analytical column was a Hamilton PRPX-100, 10 m particle size, 25 cm length 4.1 mm internal diameter (id). The chromatographic separation of selenocystine (SeCyst), selenomethionine (SeMet), selenite (SeIV), and selenate (SeVI) was adapted from Ge et. al. [6] and performed using a 5 mmol/L ammonium citrate buffer with pH adjusted to 5.2. Injection volume was fixed at 100 L. Methanol (2% v/v) was added to the mobile phase to improve sensitivity [7]. The mobile phase was delivered at 1 mL/min isocratically. The HPLC-ICP-MS interface consisted simply of polyetheretherketone (PEEK) tubing. Polyatomic Interference Removal ICP-MS is the detector of choice for trace element analysis due to its high sensitivity and selectivity. It is also one of the most often used detection systems for total and speciation analyses of selenium. Nevertheless, selenium detection limits obtained with a conventional ICP-MS (quadrupole filter without collision/reaction cell system) are not sufficient when dealing with selenium determinations in natural waters. Difficulties in Se determination by ICP-MS are mainly due to its high first ioniza-
tion potential (9.75 eV) compared to argon (15.75 eV) and, as a consequence, its low ionization in an Ar plasma (around 33% [8]). Secondly, argon polyatomic interferences, especially 40Ar40Ar+ and 40 Ar38Ar+ dimers, prevent selenium determination from its most abundant isotopes 80Se (49.6% abundance) and 78Se (23.8% abundance). Hence, the less interfered and less abundant 82Se isotope (9.2% abundance) is generally monitored. The problem of argon-based polyatomic interferences can be solved with the use of ICP-MS systems equipped with a collision/reaction cell (CRC). A 10- to 20fold improvement in total Se and speciation analysis detection limits was observed using the ORS cell of the Agilent 7500ce. Speciation analysis detection limits are below 15 ng/L based on monitoring 80Se (see Table 3). Better detection limits were achieved for 80Se compared to 78Se because the 7500ce was optimized on 80Se.
Table 3. Optimization of ORS Operating Conditions Agilent 7500ce 5.5 mL/min H2 0.5 mL/min He*
78
The use of CRC technology allows efficient removal of argon-based interferences, resulting in improved ICP-MS detection power for selenium by permitting monitoring of its most abundant isotope, 80Se. However, such improvements are mitigated, in some cases, by reaction cell induced interferences. Indeed, hydrogen, or impurities contained in gases, can cause hydride formation from elements such as bromine, selenium, or arsenic [9-11]. Therefore, in samples containing bromine, as in the case of natural waters, there would be an interference on 80 Se and 82Se from bromine hydride. As a result, the 78Se signal should be monitored to avoid misinterpretation of the results and alleviate the need for correction equations. Selenium concentrations determined in different mineral and spring waters, under the ICP-MS operating conditions described in Table 3, are summarized in Table 4. Results for certified simulated rain water (TM-Rain 95 from National Water Research Institute, [Ontario, Canada]) are also given. Total Se was established by measuring the 78 Se isotope without correction equations.
Instrument Cell gases Elemental Analysis
Experimental
80
Se
Se
Detection limit (ng/L) Repeatability (%) HPLC Coupling Detection limit (ng/L) Repeatability (%)
6 2
78 Se 1430 2
4 2
80
Se 715 2
Figure 1 shows a chromatogram of 1 g(Se)/L per species standard obtained using HPLC-ICP-MS. The method was then applied to the mineral and spring water samples previously analyzed for their total selenium content. The results of selenium species concentrations are summarized in Table 4, together with the total selenium data.
*Addition of He is optional. Similar detection limits should be achievable without He. 18000 16000 14000 12000 10000 8000 6000 4000 2000 0
SeIV
m/z 80 m/z 78 m/z 82
Counts
SeVI SeCyst SeMet
200
400
600 Time (s)
800
1000
1200
Figure 1.
Chromatogram of standard, 1 g(Se)/L per species; 100 L injected, Hamilton PRP X-100 column, citrate buffer pH 5.2 and 2% methanol as mobile phase. 3
Table 4.
Selenium Concentrations Determined in Different Natural Waters [units: ng(Se)/L] Elemental Analysis HPLC Coupling Se SeIV SeVI
78
Natural Water TM-Rain 95 A B C D E
78
Se
Se SeIV
80
SeVI < DL 72 6 143 4 267 13 492 5 1920 20
622 19* 67 1 142 24 240 20 467 17 1890 160
629 7 < DL < DL < DL < DL 55 2
< DL 69 2 140 9 232 13 475 4 1840 30
615 8 < DL < DL < DL < DL 57 6
*Certified value 740 290 ng(Se)/L
Concentrations found in total and speciation analyses are in complete agreement, showing the suitability of the method when applied to natural water samples. Although the bromine hydride interference on m/z 80 is present, it is separated chromatographically without overlapping with the selenium species. The chromatogram of water sample C (Figure 2) shows bromine elutes after the selenate peak.
200 180 160 140 Counts (m/z 78,80) 120
Selenate, commonly found in oxygenated waters, was determined in commercial waters A-D. Selenite was identified in TM-Rain 95 water, which is only certified for its total selenium content. Only water E, a noncommercial ground water, contained both inorganic selenite and selenate species (see Figure 3).
450
m/z 81 (81Br) m/z 79 (79Br) m/z 80 (80 Se + 79BrH) m/z 78 (78 Se)
400 350 300 250 Counts (m/z 79,81)
100 200 80 60 40 20 0 0 200 400 600 Time (s) 800 1000 150 100 50 0 1200
Figure 2.
Chromatogram of natural water C showing reaction cell induced interference from bromine hydride elutes after the selenate peak.
14000
12000
m/z 80
10000
m/z 78
Counts
8000
6000
4000
2000
0 0 200 400 Time (s) 600 800
Figure 3.
Chromatogram of natural water E, the only sample to contain both inorganic species. First peak is SeIV, second peak is SeVI.
Conclusions
Interest in selenium speciation has grown in recent years due to its characteristics as both an essential and toxic element. However, the complete speciation of selenium, including organic and inorganic forms, is still a major challenge. This is particularly true when exploring selenium speciation in natural waters due to the low levels of Se present. A hyphenated technique consisting of isocratic HPLC coupled to ICP-MS with optimized collision/reaction cell conditions allows for a quick and precise simultaneous analysis of organic and inorganic selenium species. Moreover, as HPLCICP-MS coupling is easily automated, it can be considered a robust routine method to monitor selenium species levels in environmental and nutritional samples.
4. Y. Sayato, T. Hasegawa, S. Taniguchi, H. Maeda, K. Ozaki, I. Narama, K. Nakamuro, Eisei Kagaku 39 (1993) 289. 5. World Health Organization (W.H.O.) (1987) Environmental health criteria 58 : selenium. 6 H. Ge, X. J. Cai, J. F. Tyson, P. C. Uden, E. R. Denoyer, and E. Block, Anal. Commun. 33 (1996) 279. 7. E. H. Larsen and S. Strup, J. Anal. At. Spectrom 9 (1994) 1099. 8. A. R. Date and A. R. Gray (Eds), Applications of Inductively Coupled Plasma Mass Spectrometry, Blackie & Son, 1989. 9. L. Hinojosa-Reyes, J. M. Marchante-Gayon, J. L. Garcia-Alonso, and A. Sanz-Medel, J. Anal. At. Spectrom, 18 (2003) 11. 10.D. Wallschlager and J. London, J. Anal. At. Spectrom, 19 (2004) 1119 11.J. Darrouzes, M. Bueno, G. Lespes, and M. Potin-Gautier, J. Anal. At. Spectrom, 20 (2005) 88.
References
1. J. E. Conde and M. Sanz Alaejos, Chem. Rev. 97 (1997) 1979. 2. M. A. Al Bayati, O. G. Raabe, and S. V. Teague, J. Toxicol. Environ. Health 37 (1992) 549. 3. C. G. Wilber, Clin. Toxicol. 17 (1980) 171.

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2007 Printed in the USA July 25, 2007 5989-7073EN
Unmatched Removal of Spectral Interferences in ICP-MS Using the Agilent Octopole Reaction System with Helium Collision Mode Application
Metals Analysis
Authors
Ed McCurdy, Glenn Woods and Don Potter Agilent Technologies Lakeside Business Park Cheadle Royal, Cheshire UK
(plasma-based), such as 40Ar, 40Ar16O, and 40Ar38Ar, and those derived from sample matrix components (matrix-based), such as 35Cl16O, and 32S34S. Plasmabased polyatomic ions are both predictable and reasonably constant, regardless of sample matrix, whereas matrix-based polyatomic ions are less predictable and vary with sample matrix components and their relative concentrations. Recent advances in CRC technology have led to dramatic improvements in the analysis of interfered elements which previously proved difficult or impossible to measure at required levels in certain sample matrices. In a CRC ICP-MS, the cell is typically pressurized with a reactive gas that reacts with the interference (referred to as reaction mode). Attenuation of the interfering species occurs by one of several different processes depending on the gas and the interference. However, in practice, reaction mode-only CRCs limit the system to the removal of single interfering ions from single analytes [18], using highly reactive gases and specific measurement conditions. Some instruments use simpler or less reactive cell gas such as H2, but its use is limited mainly to plasmabased interferences, as it reacts slowly or not at all with matrix-based interferences which are much more difficult to remove. Helium (He) Collision Mode The development of the Agilent Octopole Reaction System (ORS) introduced a new and much more powerful mode of CRC operation He collision mode which uses an inert collision gas to remove all polyatomic species based on their size rather
Abstract
Many routine laboratories have adopted ICP-MS as their primary technique for metals analysis due to its simple operation as a multi-element analyzer. However, despite its higher performance for the targeted removal of specific interferences, collision/reaction cell (CRC) ICP-MS remains relatively understudied in terms of its multielement capability. This work demonstrates that the Agilent 7500ce ICP-MS can be operated with a single set of He cell gas conditions, to provide effective interference removal for a range of elements in a challenging and complex sample matrix.
Introduction
ICP-MS is an immensely powerful multi-element analytical technique, but it does suffer from some well-documented spectral interferences, which can be especially problematic when complex and variable samples are analyzed. Most interferences in ICP-MS arise due to an overlap from a molecular (or polyatomic) ion at the same nominal mass as the analyte of interest. Commonly reported interferences can be broadly divided into two groups: those derived from the plasma and aqueous solution
than their relative reactivity with a reaction gas. Since all polyatomics are larger than analyte ions of the same mass, their larger cross-section means that they suffer more collisions with the cell gas and so lose more energy as they progress through the pressurized region. On arrival at the cell exit, the large cross section polyatomic species all have distinctly lower ion energy (due to collisions with the He cell gas) than the analyte ions and so can be prevented from leaving the cell using a stopping voltage, allowing only the analytes to pass through to the analyzer. This separation process is known as kinetic energy discrimination (KED), and this simple yet extremely effective approach offers a number of significant analytical advantages over reaction mode. Advantages of He Collision Mode: In contrast with a reactive cell gas, He is inert so does not react with the sample matrix - no new interferences are formed in the cell As He is inert, it does not react with and cause signal loss for analyte or internal standard ions ALL interferences (plasma-based AND matrixbased) are removed or attenuated so multielement screening or semiquant analysis can be combined with effective interference removal Since He collision mode is not interferencespecific, multiple interferences can be removed from the same analyte (or different analytes) simultaneously [9, 10] No prior knowledge of the sample matrix is required, and no method development is required, in contrast to the extensive, analyteand matrix-specific method development which is required for any reactive mode of interference removal [11] He collision mode can be applied to every sample, every matrix, and the same setup (gas flow rate) is used for every application No cell voltages to set up or optimize NO interference correction equations are used Why Cant Other CRC-ICP-MS Use He Collision Mode? To work properly, He collision mode requires efficient analyte/interference separation by KED, which requires two conditions to be met: first, the energy of all the ions entering the cell must be very tightly controlled. Agilents unique ShieldTorch
interface insures a very narrow ion energy spread of 1 eV: its physically grounded shield plate provides better control of initial ion energy than electrically grounded plasma designs (such as balanced, center-tapped or interlaced coils). Second, in the cell, polyatomic species must experience a sufficiently high number of collisions to differentiate them from the analyte ions at the cell exit. In the Agilent ORS this is achieved by the use of an octopole ion guide the only implementation of an octopole cell in ICP-MS. There are two key benefits to the use of an octopole cell: Octopoles have a small internal diameter. As a result, the cell entrance and exit apertures are small so the cell operates at relatively higher pressure compared to quadrupole or hexapole cells which increases ion/gas collisions. Octopoles also have better focusing efficiency than hexapole and quadrupole ion guides. The ion beam is tightly focused, which insures good ion transmission and high sensitivity at its higher cell operating pressure. Only the Agilent ORS combines the ShieldTorch interface with an octopole cell and so only the Agilent ORS can effectively use He collision mode. Testing He Collision Mode a Worst Case Scenario A synthetic sample matrix was prepared to give rise to multiple interferences across a range of common analytes and test the ability of He collision mode to remove all overlapping polyatomic species. A standard solution was prepared, containing 1% HNO3, 1% HCl and 1% H2SO4 (all UpA UltraPure Reagents, Romil, Cambridge, UK), 1% Butan-1-ol (SpS Super Purity, Romil, Cambridge, UK) and 100 mg/L (ppm) each of Na and Ca (both prepared from 10,000 mg/L Spex CertiPrep Assurance single element standards), to simulate a very complex natural sample matrix. Table 1 summarizes the potential polyatomic species in this sample matrix, illustrating that practically every element in the mid-mass region (from 50 to 80 amu) suffers from multiple interferences. This makes the accurate determination of these elements in complex sample matrices extremely challenging for conventional ICP-MS, as the complex nature of the multiple interferences means mathematical corrections will be unreliable. This also illustrates why reactive cell gases are unsuitable for the multielement analysis of complex samples; no single reaction gas can be effective for a range of
polyatomic ions, each of which will have different reactivity with any given reactive cell gas. However, every interference shown in Table 1 is a polyatomic ion and can therefore be attenuated effectively using a single set of He collision mode conditions. Two sets of spectra were acquired to show the ability of the He collision mode to remove multiple interferences; one in no-gas mode and the second with He added to the cell. No data correction or background subtraction was applied. Finally, a 5-ppb multi-element spike was added to
Table 1. Isotope
51 52 53 54 55 56 57 58 59 60 61 63 64 65 66 67 68 69 70 71 72 73 74 75 77 78 80
the matrix and spectra acquired to confirm the recovery of all analytes and check for correct isotopic fit. Instrumentation An Agilent 7500ce ICP-MS was optimized using the typical tuning conditions for high and variable sample matrices (plasma conditions optimized as usual for ~0.8% CeO/Ce). No attempt was made to optimize any parameter for the targeted removal of any specific interference. 5.5 mL/min He gas (only) was added to the cell for the collision mode measurements. Comparison of Spectra The background spectrum obtained in no-gas mode is shown in Figure 1a, together with the same spectrum (same mass range and intensity scale) under He collision mode conditions, in Figure 1b. From Figure 1a, it is clear that the normal background components of the argon plasma gas and aqueous sample solution (Ar, O, H), together with the additional components of the synthetic sample matrix (HNO3, HCl, H2SO4, butanol, Ca and Na), lead to the formation of several high intensity background peaks in the no-gas mode spectrum, notably 40 Ar16O+ and 40Ar2+ from the plasma, but also 40 Ar12C+, 32S2+, 35Cl16O+, etc, from the matrix. These high intensity background peaks show why several interfered elements (56Fe, 78Se and 80Se, 52Cr in a carbon matrix, 64Zn in a sulfur matrix) have traditionally been considered as difficult elements for ICP-MS. When helium is added to the cell (He collision mode conditions) all of these high intensity background peaks are removed from the spectrum, (Figure 1b same sample, same intensity scale as Figure 1a) demonstrating the effectiveness and the universal applicability of He collision mode. Figures 2a and 2b are the same two spectra as in Figure 1, but with the vertical scale expanded 100x. Many more, lower intensity, matrix-derived polyatomic species are now observed. These interferences, though present at lower levels than the plasma-based polyatomic ions, have the potential to cause more serious errors in routine sample analysis, as their presence and intensity is dependent on matrix composition, which, in routine laboratories, may be variable and unknown. At this expanded scale, it is clear that the use of He collision mode has reduced the background
Principal Polyatomic Interferences from an Aqueous Matrix Containing N, S, Cl, C, Na, and Ca Principal interfering species
35 36 36 40 37 40 40 40 40 44 44 40
V Cr Cr Fe Mn Fe Fe Ni Co Ni Ni Cu Zn Cu Zn Zn Zn Ga Zn Ga Ge Ge Ge As Se Se Se
Cl16O, 37Cl14N Ar16O, 40Ar12C, 35Cl16OH, 37Cl14NH Ar16OH, 40Ar13C, 37Cl16O, 35Cl18O, 40Ar12CH Ar14N, 40Ca14N Cl18O, 23Na32S Ar16O, 40Ca16O Ar16OH, 40Ca16OH Ar18O, 40Ca18O, 23Na35Cl Ar18OH, 43Ca16O Ca16O, 23Na37Cl Ca16OH, 38Ar23Na, 23Na37ClH Ar23Na, 12C16O35Cl, 12C14N37Cl S O2, 32S2, 36Ar12C16O, 38Ar12C14N, 48Ca16O S O2H, 32S2H, 14N16O35Cl, 48Ca16OH S O2, 32S34S, 33S2, 48Ca18O S SH, 33S2H, 48Ca18OH, 14N16O37Cl, 16O235Cl S O2, 34S2 S O2H, 34S2H, 16O237Cl S O2, 35Cl2 S O2H Ar32S, 35Cl37Cl, 40Ar16O2 Ar33S, 35Cl37ClH, 40Ar16O2H Ar34S, 37Cl2 Ar34SH, 40Ar35Cl,40Ca35Cl Ar37Cl, 40Ca37Cl Ar38Ar Ar2, 40Ca2, 40Ar40Ca
32 16 32 16 34 16 32 34 32 18 32 18 34 18 34 18 40 40 40 40 40 40 40
species to very low levels, including the high intensity plasma-based species ArO+ and Ar2+. The only peaks clearly visible in He collision mode (Figure 2b) on this scale are Fe and Zn (the peak template confirms the Zn isotopic pattern at m/z 64, 66, and 68), due to trace level contamination present in the matrix components. By contrast, in no-gas mode (Figure 2a), almost every isotope of every element in this mass region has an overlap from at least one matrix-derived polyatomic interference.
a
2.0E7 [2] Spectrum No.1 [ 110.528 sec]:001SMPL.D / Tune #2 [CPS] [Linear]
Ar2, Ca2, CaAr, S2O, SO3
ArO, CaO ArC, SO

1.0E7
ArOH, CaOH
S2, SO2
ClO ClO
50 52
ArN
56 58
CaO
60
ArNa CaOH
62 64
S2, SO2
66 m/z 68 70
ArS
72
ArCl, CaCl Ar 2
74 76
Ar2
78 80 82 84
54
b
50 52 54 56 58 60 62 64 66
68 m/z
70
72
74
76
78
80
82
84
Figure 1.
High intensity interfering polyatomic ions from complex matrix sample (see text for composition) in (a) no-gas mode and (b) He collision gas mode, on same intensity scale (2.0E7).
[2] Spectrum No.1
[ 110.528 sec]:001SMPL.D / Tune #2 [CPS] [Linear]
a
2.0E5
ClO
ArC, SO
ClO
ArO, CaO
ArOH, CaOH
S2O, SO3, ArCO, ArCN
ArNa2H, SO2H
ArS, Cl2
Ar2, Ca2, ArCa, S2O, SO3
ArNa ArN CaO, NaCl CaOH, ArNa NaClH ArCl, CaCl, ArSH
Ar2
1.0E5
CaO, NaCl
Br, Ar2H S2, SO2
CaO, NaS
CaO ArS, SO2 ClO2 ArS, Cl2

56 58 60 62 64 66 m/z 68 70 72
ArCl ArS, Cl2 ArS

74 76
Br, Ar2 Ar H 2
50
52
54
78
80
82
84
b
50 52 54
Fe
56 58 60 62
Zn
64
Zn
66 m/z
Zn
68
Zn
70 72 74 76 78 80 82 84
Figure 2.
Low intensity interfering polyatomic ions from complex matrix sample in (a) no-gas mode and (b) He collision gas mode on same intensity scale (2.0E5), which is expanded 100x compared to Figure 1.
Measurement of Analytes in the Presence of the Sample Matrix Having demonstrated the effective reduction of both plasma-based and matrix-based polyatomic ions using a single set of He collision mode cell conditions (Figures 1b and 2b), a second sample was analyzed. This time the sample consisted of the same multi-component matrix, but was spiked with a 5-ppb multi-element standard. Data was acquired in He collision mode to ensure that the same cell conditions used for interference removal also gave sufficient analyte sensitivity to permit the measurement of the previously interfered trace elements in this mass range. The spike consisted of 5 ppb each of V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Ge, As and Se, all of which had at least one analytically useful isotope which suffered a polyatomic overlap in no-gas mode in this matrix. Spectra obtained in He collision mode for the blank (unspiked) matrix and the spiked matrix are
compared in Figures 3a and 3b respectively. Note that these spectra are shown on an intensity scale that is a further 4x lower than that used for Figures 2a and 2b, allowing the presence of the contaminant elements (Fe, Ni, Cu, Zn) to be confirmed from their isotopic templates (Figure 3b). The spectrum shown in Figure 3a clearly illustrates the capability of He collision mode to perform multi-element measurements at the low ppb level in this most complex and challenging sample matrix. Good isotopic fit is shown for every analyte. The only residual interferences observed were the plasma-based species ArOH and Ar2 at mass 57 and 80 respectively. The Ar2 signal at mass 80 is equivalent to ~5 g/L Se. However, the polyatomic interferences on the other Se isotopes at m/z 77, 78, and 82 were removed completely, allowing Se determination at any of these isotopes (76Se would also be available, but is overlapped by 76Ge which was in the spike mix).
a
5.0E4 [1] Spectrum No.1 [ 300.207 sec]:002SMPL.D / Tune #1 [CPS] [Linear]
As Se, Ar2
Inset: Intensity scale expanded x10 compared to main spectra
Se, Ge Br Br Se
78 80 82 84
Co Cu Se
2.5E4
Se
Fe
Ni
b
76
Ar2 Cr Cu V Mn Cr
50 52
Br
76 78 80
Br
82 84
Ni
Zn Zn Zn Zn
64 66 m/z 68
Cr Fe
54 56
Fe
58 60
Ni
Ni
62
Zn, Ge
70
Ge Ge
72
Ge
As Ar2 Br
76 78 80
Br
82 84
74
Fe
b
Ni
50 52 54 56 58 60 62
Zn Cu
64
Zn Cu
66 m/z
Zn
68
Zn
70 72 74 76 78
Br
Ar2
80
Br
82 84
Figure 3.
Complex matrix sample in He collision mode, (a) spiked at 5 ppb with V, Cr, Fe, Mn, Ni, Co, Cu, Zn, Ge, As, and Se and (b) unspiked. Intensity scale is 5.0E4 (5.0E3 for inset spectra).
Conclusions
The ability to remove ALL polyatomic interferences under a single set of conditions means that He mode is effectively universal being suitable for any isotope of any element in any sample matrix. The use of He collision mode provides a unique new mode of operation, in which ALL the isotopes of each analyte become accessible. This, in turn, means that major isotopes that could not previously be used due to interferences (for example: 52Cr in a carbon matrix, 56Fe in any aqueous sample, 63Cu in a sodium matrix, and 64Zn in a sulfate matrix) - now become available. This is a great advantage to the analyst since, if desired, results can be verified by measuring many elements at both the preferred isotope AND at a second,
qualifier isotope. Since both isotopes are free from polyatomic interference when measured using He collision mode, the use of two independent measurements gives a valuable confirmation of the reported result. A further benefit of this powerful mode of analysis concerns sample preparation. In normal (non-CRC) ICP-MS, the choice of dilution media was limited mostly to nitric acid. Hydrochloric and sulfuric acid could not be used because of the problems of chloride or sulfur-based matrix interferences. Analysts can now choose the most appropriate digestion technique for the sample, secure in the knowledge that any new polyatomic interferences will be removed under the existing, standard He mode conditions.
The use of He collision mode on the 7500ce was demonstrated to provide effective removal of all polyatomic interferences under a single set of conditions, thereby enabling accurate multi-element analysis in complex and unknown samples. The use of an inert cell gas insures that there is no loss of analyte signal by reaction and that no new interfering species are generated, in contrast to the use of a reactive cell gas. Since no analytes are lost by reaction and no new interferences are formed, uninterfered elements (and internal standards) can be measured under the same conditions as potentially interfered elements, and the use of a single set of cell conditions for all analytes allows multi-element analysis of transient signals (such as those derived from chromatography or laser ablation sample introduction), as well as semiquantitative screening analysis. He collision mode is suitable for all analytes that suffer from polyatomic ion interferences and the cell conditions do not need to be set up specifically for each analyte, so the same cell conditions can be applied to new analyte suites, without requiring method development. Furthermore, since the He mode conditions are not set up specifically for the removal of individual interferences, identical cell conditions can be used for highly variable or completely unknown sample matrices, which greatly simplifies operation in a routine laboratory. The ORS enables ICP-MS to be used for the trace multielement measurement of the most complex, real world sample matrices with no method development and with complete confidence.
5. H-T. Liu and S-J. Jiang, (2003), Anal. Bioanal. Chem., 375, 306. 6. D. R. Bandura, S. D. Tanner, V. I. Baranov, G. K. Koyanagi, V. V. Lavrov and D. K. Bohme, in Plasma Source Mass Spectrometry: The New Millennium, eds. G. Holland and S. D. Tanner, The Royal Society of Chemistry, Cambridge, 2001, p. 130. 7. C. C. Chery, K. DeCremer, R. Cornelis, F. Vanhaecke and L. Moens, (2003), J. Anal. Atom. Spectrom., 18, 1113. 8. F. Vanhaecke, L. Balcaen, I. Deconinck, I. De Schrijver, C. M. Almeida and L. Moens, (2003), J. Anal. Atom. Spectrom., 18, 1060. 9. N. Yamada, J. Takahashi and K. Sakata, (2002), J. Anal. Atom. Spectrom., 17, 1213. 10.E. McCurdy and G. Woods, (2004), J. Anal. Atom. Spectrom., 19, 607. 11. J. W. Olesik and D. R. Jones, (2006), J. Anal. Atom. Spectrom., 21, 141.

References
1. G. K. Koyanagi, V. I. Baranov, S. D. Tanner and D. K. Bohme, (2000), J. Anal. Atom. Spectrom., 15, 1207. 2. P. R. D. Mason, K. Kaspers and M. J. van Bergen, (1999), J. Anal. Atom. Spectrom., 14, 1067. 3. J. M. Marchante Gayon, I. Feldmann, C. Thomas and N. Jakubowski, (2000), J. Anal. Atom. Spectrom., 16, 457. 4. E. H. Larsen, J. Sloth, M. Hansen and S. Moesgaard, (2003), J. Anal. Atom. Spectrom., 18, 310.
Ion Chromatography (IC) ICP-MS for Chromium Speciation in Natural Samples Application
Environmental
Authors
Tetsushi Sakai Agilent Technologies Musashino Center Building 1-19-18 Naka-cho Musashino-shi Tokyo 180-0006 Japan Ed McCurdy Agilent Technologies Lakeside, Cheadle Royal Business Park Stockport United Kingdom Steve Wilbur Agilent Technologies, Inc. 3380 146th PI SE Suite 300 Bellevue, WA 98007 USA
spectrometry), to give a simple and reliable method for the separation and measurement of Cr(III) and Cr(VI), and so provide an accurate indication of the toxicity of the Cr level in a sample. This method has the merit of being applicable to high matrix samples, such as hard drinking water, due to the optimization of the sample preparation method and the chromatography. Also, the ICP-MS method provides excellent signal to noise, as a result of the removal of potentially interfering background species in the reaction cell, allowing the accurate determination of toxicologically useful levels of Cr(VI), at concentrations below 0.1 g/L.
Introduction
The measurement of chromium toxicity is a requirement across a wide range of sample types, including drinking water, foodstuffs, and clinical samples (the latter used primarily to assess occupational exposure). However, it is the hexavalent form of Cr - Cr(VI) that is the toxic form, while the trivalent form - Cr(III) is an essential element for human nutrition. Methods to establish the potential toxicity of Cr must therefore determine the concentration of Cr(VI), rather than simply total Cr. Two common approaches are used to address the issue: First, if the total Cr level measured is below the toxic level for Cr(VI), then it is reasonable to state that the Cr level will not be toxic, even if all of the Cr is present as Cr(VI). However, this approach can lead to a large number of false positives if samples contain a high concentration of Cr(III), so a more accurate approach is to separate and measure the Cr(VI) itself or, ideally, separate and measure both forms of Cr, giving an indication of the level of total Cr AND the level of toxic
Abstract
Trace measurements of the element chromium (Cr) are of interest in a wide range of applications and matrices. In the environment, Cr exists in two different oxidation states, the trivalent Cr(III) cation and hexavalent Cr(VI) anion. In mammals, Cr(III) is an essential element involved in the regulation of glucose; however, the element in its hexavalent form demonstrates mutagenic and carcinogenic effects at relatively low levels. Because of this duality, total Cr measurements do not provide sufficient information to establish potential toxicity. In order to assess the potential toxicity of the Cr level in a sample, it is the Cr(VI) concentration that must be measured, rather than the total Cr concentration. A new method was developed to couple Ion Chromatography to Octopole Reaction Cell ICP-MS (inductively coupled plasma mass
Cr(VI), from a single analysis. Separating and detecting individual forms or species of elements is usually a straightforward analytical challenge, but Cr is an unusual case in this respect. This is because the common forms of Cr in natural samples such as water are chromate (CrO42) for Cr(VI) and chromic ion (Cr3+) for Cr(III). Chromate is an anion and the chromic ion is cationic, so a single ion exchange method will not work for both forms under the same conditions. A further problem is that Cr(III) is the most stable oxidation state in samples such as water, whereas Cr(VI) ions are strong oxidizing agents and are readily reduced to Cr(III) in the presence of acid or organic matter. Consequently great care must be taken during sample collection, storage and preparation, to ensure that the Cr species distribution present in the original sample is maintained up to the point of analysis.
at room temperature. Note that a relatively high concentration of EDTA is required for this method to be applicable to natural water samples, since other ions, such as Ca and Mg, which are commonly present at 10s or 100s mg/L in hard drinking water for example, would compete with the Cr(III) to form EDTA complexes, leading to low and matrix dependent Cr(III) recovery. The combination of the separation of the Cr species using ion chromatography (IC), together with analysis of the separated species using ICP-MS, offers an ideal analytical method, as it permits the individual Cr species to be separated using a simple, low cost IC configuration. ICP-MS detection also allows the separated Cr species to be measured at extremely low concentrations, providing accurate assessment of exposure levels, even for natural or background Cr concentrations. ICP-MS has excellent sensitivity and so is a good detector for many trace elements. The introduction of collision/reaction cells (CRCs) for ICP-MS allows Cr to be measured even more accurately and with better sensitivity, using the main isotope at mass 52, with removal of the primary matrix-based interferences ArC and ClOH. The sample preparation method, column type and chromatographic conditions used for Cr speciation are shown in Table 1. Note that, in addition to the stabilization of the samples with Na EDTA, EDTA was also added to the mobile phase, to stabilize the Cr(III) complex during separation. In addition, it was found that the use of pH 7 was essential for species stabilization and
Experimental
The method described in this application note used an optimized sample stabilization method, in which the samples were incubated at 40 C with EDTA, which forms a complex with the Cr(III), allowing a single chromatographic method to be used to separate the Cr(III)EDTA complex and the Cr(VI). The reaction to form the Cr(III)EDTA complex is dependent on the incubation time and temperature, with complete conversion occurring after less than 1 hour at 60 C or 3 hours at 40 C. Complete conversion did not occur even after 7 hours incubation
Table 1. Cr column Mobile phase Flow rate Column temperature Injection volume Sample preparation Reaction temperature Incubation time EDTA concentration 40 C 3h Chromatographic Conditions for Cr Speciation
Agilent part number G3268A, 30 mm 4.6-mm id 5 mM EDTA (2Na), pH 7 adjust by NaOH 1.2 mL/min Ambient 50~500 L
5~15 mM pH 7 adjust by NaOH
optimum chromatographic separation. The IC configuration used for the work presented in this note is illustrated in Figure 1. Note that the nonmetal IC pump (Metrohm 818 IC Pump was used to deliver the mobile phase, but the sample loop was filled and switched using the optional Integrated Sample Introduction System (ISIS) of the Agilent 7500ce ICP-MS. While this configuration maintains the high precision and relatively high pressure of the IC pump, it provides a much simpler and lowercost alternative to a complete IC or HPLC system,
Column (G3268A) Metrohm 818 IC Pump Valve 1 ISIS ICP-MS Nebulizer
Mobile phase 0.05~1.0 mL
Drain
Autosampler (ASX500)
Sample
Pump 1 ISIS
Integrated Sample Introduction System (ISIS)
Figure 1.
IC-ICP-MS configuration used for Cr speciation.
since only the IC pump module is required in addition to the ICP-MS system.

Under the conditions described above, with ICP-MS detection using the Agilent 7500ce in H2 cell gas mode to remove the ArC and ClOH interferences on Cr at mass 52, detection limits (DLs) of <20 ng/L were obtained for the individual Cr species, as shown in Table 2. Many international regulations for hexavalent Cr specify a maximum allowable concentration of 1 g/L, with a required DL of one-tenth of this level (100 ng/L), and even the small sample volume injection of 100 L easily meets these
Table 2. Inject/L 50 100 250 500
DLs for Cr Species by IC-ICP-MS RT/min Cr(III) Cr(VI) 0.79 0.79 0.85 0.97 2.09 2.09 2.21 2.39 Peak height/counts Cr(III) Cr(VI) 8548 13688 33967 44870 4261 7173 20830 37502 Peak area/counts Cr(III) Cr(VI) 1082295 1704312 4939876 10268086 914804 1525147 4546219 9398651 DL (ng/L)* Cr(III) Cr(VI) 69.5 43.4 17.5 13.2 139.4 82.8 28.5 15.8
*Detection limits calculated as three times the peak-to-peak signal-to-noise as measured on standard chromatograms.
requirements. However, increasing the injection volume to 500 L allowed the DLs to be reduced to 13.2 ng/L for Cr(III) and 15.8 ng/L for Cr(VI). For a simple standard solution, these conditions give an excellent signal to noise for both Cr species, as illustrated in Figure 2. This chromatogram shows the separation of the two Cr species each at a concentration of 0.1 g/L (ppb), using an injection
9000 8000 Abundance/ counts 7000 6000 5000 4000 3000 2000 0.00 1.00 2.00 3.00 4.00 5.00 Retention Time / min
Cr(III) 0.1 g/L Cr(VI) 0.1 g/L
Figure 3.
Calibration for Cr(III) - Standard concentrations 0.05, 0.1, 0.5 and 1.0 g/L.
Figure 2.
Separation and detection of Cr(III) and Cr(VI) at a concentration of 0.1 g/L each species.
volume of 500 L. Clearly the peaks are easily detected above the background and the baseline separation of the two species in a total time of about 3 minutes is also illustrated. Using a series of synthetic standard solutions at low concentrations, a calibration was created for each of the two Cr species. Quantification was based on
Figure 4.
Calibration for Cr(VI) - Standard concentrations 0.05, 0.1, 0.5 and 1.0 g/L.
peak area. The calibrations obtained for Cr(III) and Cr(VI) are illustrated in Figures 3 and 4, respectively, each showing excellent sensitivity and linearity. In addition to sensitivity, species stability, chromatographic separation and calibration linearity, for the method to be suitable for routine analysis, it is essential that it provides acceptable long-term stability. In chromatographic analysis, stability is governed by two factors, RT stability and peak area stability. The data in Table 3 illustrates both of these parameters and indicates that the stability of the method is certainly acceptable for routine operation.
Of more interest for the routine analysis of chromium species (or specifically hexavalent Cr) in natural water samples is the maintenance of this excellent sensitivity, stability and chromatographic separation in samples that contain a high concentration of other ions. In order to test the suitability of the method for these real-world sample types, the method was applied to the determination of both Cr species in both spiked and unspiked mineral water samples. The first sample evaluated was a leading French mineral water referred to in this study as mineral water A. Figure 5 shows the chromatogram of the two Cr species in the unspiked and spiked samples of mineral water A. The major element composition of the water is also shown in the table inset in Figure 5, indicating the typical drinking water com-
Routine Analysis
Table 3.
Stability of RT and Peak Area for Multiple 500 L Injections of 0.5 g/L Each Cr Species RT/min Cr(III)-EDTA 0.969 0.969 0.969 0.952 0.969 0.969 0.985 0.969 0.969 0.969 0.97 0.008 0.80 Peak height/counts Cr(III)-EDTA Cr(VI) 23514 18437 22642 18784 22832 18615 24104 19944 22797 19203 23830 19328 23971 19479 23393 19675 23070 20097 23826 19896 23398 19346 534.45 581.88 2.28 3.01 Peak area/counts Cr(III)-EDTA Cr(VI) 5331427 4621752 5280683 4758462 5220349 4742259 5470760 4800723 5287094 4726640 5498172 4760285 5481984 4824934 5474510 4883193 5355106 4892160 5428247 4886400 5382833 4789681 100413.18 85782.42 1.87 1.79
Number 1 2 3 4 5 6 7 8 9 10 Avg STD RSD%
Cr(VI) 2.338 2.338 2.338 2.338 2.372 2.405 2.338 2.338 2.355 2.372 2.38 0.014 0.57
9000 8000 7000 6000 5000 4000 3000 2000 0.00 1.00 2.00 3.00 RT/min
Water A + 0.1 g/L each Cr(III),Cr(VI) Cr(III): 0.105 g/L Cr(VI): 0.153 g/L
Abundance/Counts
Original Water A Cr(III): 0.005 g/L Cr(VI): 0.055 g/L
Na Ca Mg K
7.3 mg/L 91.0 mg/L 19.9 mg/L 4.9 mg/L
4.00
5.00
6.00
Figure 5.
Major element composition (mg/L) and chromatogram for spiked and unspiked mineral water A.
position of about 100 mg/L Ca and between 5 mg/L and 20 mg/L of the other major elements K, Mg and Na. The spike recovery data for mineral water A is shown in Table 4, indicating the very low level at which the Cr species were quantified (0.005 g/L
Table 4. Spike Recovery Data for 0.1 g/L Spikes of Cr(III) and Cr(VI) in Mineral Water A Found (g/L) Spike added 0.10 0.10
Element Cr(III) Cr(VI)
Original mineral water A 0.005 0.055
Spike found 0.105 0.150
Recovery (%) 100.0 95.0
and 0.055 g/L for Cr(III) and Cr(VI), respectively), and the excellent spike recovery accuracy for the low concentration spikes in this sample - better than 5% error in both cases. The second mineral water sample analyzed was another French mineral water, referred to as mineral water B, which has among the highest levels of calcium and sulfates of any commonly available mineral water (over 450 mg/L Ca and more than 1000 mg/L sulfates). As for the mineral water A sample, mineral water B was analyzed with and
14000 13000 12000 11000 Abundance/Counts 10000 9000 8000 7000 6000 5000 4000 3000 2000 0.00 1.00 2.00 3.00 RT/min 4.00 5.00 6.00 Original Water B Cr(III): 0.05 g/L Cr(VI): 0.24 g/L Water B + 1 g/L each Cr(III),Cr(VI) Cr(III): 1.10 g/L Cr(VI): 1.27 g/L Na Ca Mg K 9.1 mg/L 486.0 mg/L 84.0 mg/L 3.2 mg/L
Figure 6.
Major element composition (mg/L) and chromatogram for spiked and unspiked mineral water B.
without a spike of the two Cr species and the spike recovery was assessed. The results for the measured chromatograms are shown in Figure 6, while the spike recovery data are shown in Table 5. As shown for the mineral sample A, the major element composition of the mineral water is shown as an inset in the chromatogram, illustrating the very high mineral levels in mineral water B. Despite these high major element levels, the optimized sample preparation and chromatographic method gave good chromatographic separation and excellent spike recovery results for both Cr species. A higher spike level was used for mineral water B, due to the higher baseline (unspiked) concentration for the Cr species in this sample. The ability to recover low concentration spikes for both Cr species in such a high matrix sample indicates the effectiveness of the optimized method for sample stabilization, which ensures that a high enough concentration of EDTA is available for complete complexation of the Cr(III) species, even in the presence of a high level of competing ions. Furthermore the accurate recovery of low concentration spikes of both species indicates that potential problems of species interconversion (reduction of Cr(VI) to Cr(III)) was avoided through the selection of an appropriate pH for the samples and the
mobile phase, together with the use of EDTA in the mobile phase as well as for sample stabilization. See Table 5.
Conclusions
A new method for the stabilization and analysis of Cr(III) and Cr(IV) in natural, high matrix water samples was developed and optimized with DLs in the region of 0.05 g/L for 100-L injections, or 0.015 g/L for larger, 500-L injections. Reliable and stable separation of the Cr(III) and Cr(VI) species was achieved in a method taking approximately 3 minutes per sample and the separation and accurate quantification of the two species could be maintained even in the presence of a high concentration of competing ions, such as >500 mg/L mineral elements in the highly mineralized water. Accurate and interference-free determination of Cr at the low concentrations (0.1 g/L) required by international regulations was made possible by the simple and consistent operation of the Agilent 7500ce in reaction mode, using H2 as a cell gas. This mode of operation does not preclude the simultaneous analysis of other analytes of interest, such as As, in contrast to the use of highly reactive cell gases such as CH4 or NH3.
Table 5.
Spike Recovery Data for 1.0 g/L Spikes of Cr(III) and Cr(VI) in a Highly Mineralized Water (B) Found (g/L) Spike added 1.0 1.0
Element Cr(III) Cr(VI)
Original mineral water B 0.05 0.24
Spike found 1.10 1.27
Recovery (%) 105 102

A Comparison of the Relative Cost and Productivity of Traditional Metals Analysis Techniques Versus ICP-MS in High Throughput Commercial Laboratories Application
Author
Steve Wilbur Agilent Technologies, Inc. 3380 146th PI SE Suite 300 Bellevue, WA 98007 USA
by inductively coupled plasma optical emission spectroscopy (ICP-OES), which is less sensitive but capable of simultaneous multi-element analysis. As the need for sub-ppb detection limits extends to more elements in more samples, ICP-OES becomes less useful and the reliance on GFAA increases. However, GFAA, while sensitive, is slow, expensive to operate, and has limited dynamic range. Because GFAA is much slower than ICP-OES, many routine labs have a dedicated GFAA instrument for each analyte that is required to be measured by GFAA - multiple GFAAs working with one ICP-OES. Furthermore, the analysis of mercury will add the need for a third technique, either cold vapor AA or atomic fluorescence. However, in the interest of simplicity, a separate mercury analyzer was not considered in the examples used. Each of these techniques may require separate sample handling and preparation, as well as separate analysis, data processing and archival, significantly increasing the cost per sample. The subject of this application note is to evaluate the productivity and cost effectiveness of ICP-MS as a routine, highly sensitive, multi-element technique where a single ICP-MS instrument has the potential to replace an ICP-OES, multiple GFAAs, and a mercury analyzer for most routine elemental analyses. The analytical applicability of ICP-MS to many types of samples is already well established. More recently, the introduction of the Octopole Reaction System on the 7500 Series ICP-MS instruments from Agilent has removed the final performance barriers that have prevented ICP-MS being proposed as a complete replacement for GFAA and ICP-OES.
Abstract
A financial model was developed to help the metals laboratory using graphite furnace atomic absorption and inductively coupled plasma optical emission spectroscopy calculate the potential savings by switching to inductively coupled plasma mass spectrometry. Results based on several typical laboratory examples are presented.
Introduction
The past 5 years have seen significant growth in the use of inductively coupled plasma mass spectrometry (ICP-MS) for the analysis of trace metals in many applications in the environmental, semiconductor, geological, and health sciences industries. This growth is driven by three factors. First is the need for increasingly lower limits of detection for many metals in many applications. Second is the significantly improved performance, reliability, and ease of use of modern ICP-MS instruments. And third is economics. Traditionally, most elemental analysis has been performed by either atomic absorption (AA) or optical emission spectroscopy (OES). Generally, the ultratrace (sub-ppb) elements were measured by graphite furnace atomic absorption (GFAA), a highly sensitive single-element technique. The trace and minor (ppb to ppm) elements were measured
Methods
To facilitate this study, a spreadsheet-based sample cost comparison model was developed in Excel. This tool allows the user to provide detailed parameters related to numbers and types of samples, as well as associated costs of sample preparation, instrumentation, and analysis. Output is simply cost of analysis per sample. Also reported are the total time required for sample analysis per month, the number of analysts required, and the number of instruments. The model compares the results for GFAA, ICP-OES, and ICP-MS. While it will allow almost any values to be entered for most parameters, the results presented here are based on values obtained from several commercial laboratories doing these analyses. No model can exactly predict the results for all situations and still be simple enough to be useful. Therefore, in the interest of simplicity, a number of assumptions were made in the design of the model and in the example data entered. We feel that the assumptions are realistic and do not impart significant bias on the results. The tool is easy to use and can allow a laboratory to quickly and simply evaluate the cost effectiveness of the three techniques based on laboratory-specific information.
An instrument operator can keep a modern, automated GFAA, ICP-OES, or ICP-MS running for two shifts (16 hours) per day. When analysis times exceed 16 hours per day for any technique, additional instrumentation and operators will be required. Instruments are added in increments of one; operators are added in fractions since it is assumed that they can be shared with other tasks in the laboratory and cost calculations are based only on the portion of time the operator spends on the specific analysis. GFAA is a single element technique. Instruments with multiple lamps still perform a single analysis at a time. Typical analysis time is 90 seconds per element and each element requires two replicate analyses (burns). ICP-OES and ICP-MS are multi-element techniques and the number of elements does not significantly effect the analysis time. This is not strictly true, but the assumption is reasonable for the sake of simplicity. GFAA will use pressurized argon and the consumption is 40 hours of use per cylinder ($100). GFAA graphite tubes and platforms cost $50 per set and last for 100 burns. ICP-MS and ICP-OES will use liquid argon and the typical consumption is 3 weeks of use per dewar ($250). ICP-MS detectors last typically for 3 years and the cost per year is amortized based on 3-year lifetime.
Assumptions
GFAA system costs US$30K ICP-OES system costs US$100K ICP-MS system costs US$180K Cost of funds (finance) is 6% General facilities costs, such as laboratory space, utilities etc., are ignored since they are difficult to estimate and do not significantly affect the results in most cases.
Results
Several typical laboratory scenarios were evaluated by varying the current instrument complement of the laboratory, and by varying the current and anticipated number of samples to be analyzed per month. Also examined was the effect of the number of elements that must be analyzed by GFAA (in the case of laboratories without ICP-MS) to meet required DLs.
Scenario 1 Laboratory currently has one GFAA plus one ICP-OES, which are paid for. ICP-MS must be purchased and amortized over 3 years. See Table 1.
Table 1. Scenario 1 GFAA elements 8 8 8 # GFAA required 1 2 9 Cost/sample GFAA + ICP-OES $41 $33 $31 # ICP-MS required 1 1 2 Cost/ sample ICP-MS $30 $15 $9 Savings/ month $4,536 $18,196 $112,968
Samples/ month 400 1000 5000
Scenario 2 Laboratory currently has two GFAA plus one ICP-OES, which are paid for. ICP-MS must be purchased and amortized over 3 years. See Table 2.
Scenario 3 Laboratory currently has no instrumentation and must decide on purchasing GFAA plus ICP-OES versus ICP-MS. See Table 3.
Scenario 4 Comparison of costs per sample as a function of number of GFAA elements. (All instruments must be purchased.) See Table 4.
Table 4. Scenario 4 GFAA elements 2 4 8 10 # GFAA required 1 1 2 3 Cost/sample GFAA + ICP-OES $24 $28 $38 $42 # ICP-MS required 1 1 1 1 Cost/ sample ICP-MS $14 $14 $14 $14 Savings/ month $9,601 $12,751 $22,151 $27,490
Samples/ month 1000 1000 1000 1000
Discussion
In all cases, even when the laboratory already owns two graphite furnaces and one ICP-OES (a common configuration) and must purchase the ICP-MS, the cost per sample is lower for ICP-MS. This is mainly due to the high cost of consumables for GFAA plus the fact that GFAA and ICP-OES requires two separate sample prep steps. Additionally, as the number of samples increases from a conservative number of 400 per month to 1000 and 5000 per month, the differential becomes much greater. This is caused by rapidly increasing labor costs for GFAA, as well as the much higher sample capacity of ICP-MS, lower consumables costs, and requirements for only a single sample prep.
Return on Investment for ICP-MS

A simple return on investment (ROI) can be calculated from the above tables. In this case, the cost per month of the new ICP-MS system is approximately US $5500.00 (assuming purchase price of US$180K financed for 3 years at 6%). Figure 1 shows the payback times for a laboratory that already owns two GFAAs and one ICP-OES as a function of the sample load. The y-axis represents the accumulated monthly savings of using ICP-MS versus GFAA + ICP-OES for three different sample loads compared to the unpaid balance on the ICP-MS instrument. As can be seen, the accumulated savings of ICP-MS is equal to the payoff amount after just 4 months when analyzing 2000 samples per month. Even when analyzing as few as 400 samples per month, the accumulated savings is sufficient to pay off the ICP-MS instrument in around 20 months. In this case, eight furnace elements are assumed. Other assumptions are as above.
1800000 1600000 1400000 1200000 Balance ($) 1000000
Unpaid balance
800000 600000 400000 200000 0 1 200000 4 7 10 13 16 19 22 25 28 31 34
400 samples per month 1000 samples per month ~8 months ~20 months ~4 months 2000 samples per month
Payment period (months)
Figure 1.
Cumulative return on investment of ICP-MS purchase for three sample levels plotted against the monthly unpaid balance on the ICP-MS. In this case, it is assumed that the accumulated revenue will be used to pay off the loan when the balance equals the residual loan amount. At that point, the net monthly revenue is increased by the loan amount. In this example, laboratories running 2000 samples per month will be able to pay off the ICP-MS in about 4 months, 1000 sample laboratories in about 8 months, and 400 sample laboratories in about 20 months. At the end of 36 months (the original loan period), net revenue exceeds $200K for the 400 sample lab, $750K for the 1000 sample lab, and $1.7 million for the 2000 sample lab.
Conclusions
For almost any metals laboratory, analyzing at least 100 samples per week (400 per month) and using a combination of GFAA and ICP-OES for the analysis, converting to ICP-MS will save money. Depending on the number of samples, the payback for the ICP-MS can be as short as a few months. The cost advantages are not reduced significantly, even if the laboratory already owns its GFAA and ICP-OES instruments. They are also not significantly affected by the number of GFAA elements. As Scenario 4 shows, for the laboratory analyzing at least 1000 samples per month with only two elements by GFAA, the cost savings of switching to ICP-MS is approximately $10,000 per month. Add to this the increased confidence in results obtained by ICP-MS, the ability to analyze all analyte elements at GFAA (or better) DLs, and the robustness and simplicity of operation of modern ICP-MS instruments, and the choice becomes simple. The productivity of ICP-MS in a highvolume laboratory can quickly pay off the purchase price and increase laboratory profitability significantly.

Performance Characteristics of the Agilent 7500ce - The ORS Advantage for High Matrix Analysis
Part 1 of a 3 part series on Environmental Analysis
Application
Environmental Analysis
Authors
Steve Wilbur and Emmett Soffey Agilent Technologies, Inc. 3380 146th Pl SE Suite 300 Bellevue, WA 98007 USA Ed McCurdy Agilent Technologies Lakeside, Cheadle Royal Business Park, Stockport, Cheshire, SK8 3GR UK
Part Three is an application note covering the analysis of various high matrix environmental samples using the Agilent 7500ce ICP-MS [2]. This application note details the advances in ion optics and octopole reaction system (ORS) design that were incorporated into the 7500ce. These advances came about as a result of extensive testing and development of its predecessor (Agilent 7500c) with difficult, high-matrix samples. The design goals of the 7500ce were: Develop an ICP-MS system specifically to meet the needs of analytical laboratories to analyze unknown, variable, high-matrix samples, which are currently depending on inductively coupled plasma-optical emission spectroscopy (ICP-OES), graphite furnace atomic absorption (GFAA), and hydride and cold vapor techniques, in addition to ICP-MS. Maintain the simple, effective interference removal characteristics of the ORS - successfully introduced in the 7500c. Improve the overall sensitivity to allow ultratrace analysis of mercury (Hg) and other low level elements, which were previously difficult in some very high matrix sample types. These goals were achieved through enhancements in the sample introduction system, interface, ion optic, and ORS regions of the instrument. In common with all the other models in the 7500 Series, the 7500ce uses highly efficient 27 MHz plasma coupled to a low-flow nebulizer and cooled-spray chamber to minimize plasma and interface matrix effects. This approach has been
Abstract
The Agilent 7500ce ICP-MS was specifically designed and optimized for the analysis of trace metals in high matrix samples including environmental, clinical, geological, and others. The 7500ce uses enhanced Octopole Reaction System (ORS) technology for improved sensitivity and robustness over previous generation inductivity coupled plasma mass spectrometry (ICP-MS) instruments. This application note outlines the theory of interference removal using the ORS, the design enhancements employed, and the typical performance of the Agilent 7500ce.
Introduction
This application note is Part One of a three part series on environmental analysis using the Agilent 7500ce ICP-MS system. Part Two is an application note demonstrating the ability of the Agilent 7500ce ICP-MS system to measure trace elements in drinking water, at substantially below regulated levels, under challenging real-world conditions [1].
successfully used in all Agilent ICP-MS instruments since the 4500 Series in 1994, but recent enhancements with the development of a new digitally driven, all solid-state RF generator have further increased plasma robustness. This serves to reduce metal oxide interferences, as evidenced by a very low CeO+/Ce+ ratio of <1.5% (<0.8% in He cell gas mode). Following the successful strategy of the 7500c, all ion lenses with the exception of the octopole are outside the high vacuum region and can be serviced without venting the mass analyzer. This design greatly reduces downtime for routine system maintenance. The 7500ce maintains a linear, axial flow of ions from the sampler and skimmer cones through a pair of on-axis extraction lenses, enhancing ion transmission and reducing the effects of matrix accumulation on the extraction lenses. Borrowed from the successful 4500 and 7500a systems, the 7500ce uses a simplified Omega lens to eliminate photons and neutrals from the ion beam before entering the octopole. Unlike older photon stop designs, the Omega lens eliminates photons and neutrals while maintaining high ion transmission, particularly at low masses. After the Omega lens, ions enter the octopole
reaction cell, which is now located on-axis to the quadrupole and detector, further enhancing ion transmission. The redesigned ion lens and ORS provide improved ion transmission without compromising the tight control of ion energy, which is essential for efficient interference removal by energy discrimination (ED). Figure 1 compares the 7500c and 7500ce configurations, highlighting the simplification in the ion trajectory that has led to the improved performance specifications of the 7500ce. Enhancements in software designed specifically for routine high matrix analyses add additional capability and ease of use. These include the introduction of Virtual Internal Standardization (VIS) which allows the user to interpolate between internal standard (ISTD) response factors to create a VIS at a mass where no appropriate ISTD exists. Intelligent calibration resloping can automatically fine-tune a calibration curve, if needed, during a long sequence of high matrix samples, without the time consuming recalibration. This can be accomplished in the same process as monitoring a required continuing calibration verification (CCV).
7500c configuration
MFC MFC
H2 He
Detector
Octopole reaction cell
7500ce configuration
On-axis octopole
MFC MFC
H2 He
Detector
Octopole reaction cell
High transmission ion optic with Omega lens outside high vacuum region
Figure 1. 2
A comparison of the ion optic and octopole configurations between the Agilent 7500c and 7500ce ICP-MS systems.
Enhanced ORS Like its predecessor the 7500c, the 7500ce uses collision/reaction cell (CRC) technology in the form of the ORS to remove polyatomic interferences. The use of CRC technology to reduce interferences in ICP-MS is well-documented [3]. However, until now, there have been compromises associated with the use of some designs of CRC ICP-MS for trace level, multi-element analysis in unknown or variable matrices. These compromises included poor sensitivity for low-mass analytes, poor stability, and the necessity for matrix matching of samples and standards to avoid unexpected new interferences caused by complex, sequential reaction chemistry in the cell. As a result, some CRC systems allow only the analysis of a small number of analytes under a specific set of conditions for a single sample matrix. Numerous publications [3, 4, 5] have discussed the mechanisms for polyatomic interference removal using CRC technology including: Collisionally induced dissociation (CID) Chemical reaction Charge transfer Atom transfer Kinetic energy discrimination (KED) Mechanism 1, CID, does not occur to any great degree with the relatively light gases typically used in the collision cell because the combined kinetic energy of the collision does not generally exceed the bond energy of the polyatomic species. In most CRC ICP-MS systems, chemical reaction mechanisms including charge transfer and atom transfer
are the predominant mechanisms [4, 5]. However, in order to provide sufficient reduction of interferences, the reaction must be highly favored, which can require the use of very reactive gases for many interferences. Such gases can also react with analyte ions, so reducing sensitivity and compromising multi-element analysis, or form reaction by-products that can interfere with other analytes [4]. In this case, reaction cell conditions must be matched to a specific analyte/matrix combination and cannot be used simultaneously for multiple analytes in variable matrices. Mechanism 3, KED, relies on the fact that at the exit of the collision cell polyatomic species will possess lower kinetic energy than atomic ions at the same mass-tocharge ratio [3, 4]. This is due to the fact that collision cross sections of polyatomic ions are larger than for atomic ions, so that polyatomic species suffer more collisions with the cell gas, thus losing more of their initial energy. A bias voltage at the cell exit is then used to reject the low-energy polyatomic species, while allowing the high-energy atomic ions to enter the quadrupole for analysis and detection Three Modes of Operation - One Set of Conditions Table 1 lists the typical instrument conditions used for high-matrix analysis for the 7500ce. Instrument parameters are essentially the same for all three modes of operation1. This is because no complex procedures are required to remove newly created interferences or to avoid the reactive loss of analyte in any ORS mode.
1
Slightly higher bias voltages are required in the octopole and quadrupole to maintain ion velocity in a pressurized collision cell compared with a nonpressurized cell. Other parameters, with the exception of the cell gas flow, are identical in all modes of operation.
Table 1. Instrument Parameters for Robust Plasma Conditions Used with the 7500ce Instrument parameter Normal mode Hydrogen mode Helium mode RF Power 1500 W <Same <Same Sample depth 8 mm <Same <Same Carrier gas 0.85 L/min <Same <Same Makeup gas 0.2 L/min <Same <Same Spray chamber temp 2 C <Same <Same Extract 1 0V <Same <Same Extract 2 160 V <Same <Same Omega bias 24 V <Same <Same Omega lens 0.6 V <Same <Same Cell entrance 30 V <Same <Same QP focus 3V 11 V <Same as H2 Cell exit 30 V 44 V <Same as H2 Octopole bias 7 V 18 V <Same as H2 QP bias 3.5 V 14.5 V <Same as H2 Cell gas flow 0 3.0 mL/min H2 4.5 mL/min He
The Hydrogen Reaction Mode In hydrogen reaction mode, the ORS is pressurized using a small flow of pure hydrogen at 15 mL min1. Simple reactions with hydrogen remove argonbased polyatomics according to the following examples. Charge (e) transfer: Ar + H2 H + Ar (interference on Ca )
+ + 2 40 +
Note that some of these reaction processes lead to new polyatomic ion species, principally hydrides of the original interference. However, these new, cell-formed species all have low energy and are removed from the ion beam using the same bias voltage, as was discussed above, under interference removal by KED. Reaction mechanisms can be highly efficient, as evidenced by the calibration curves in Figure 2 for 78 Se, 40Ca and 56Fe, which show that all the normal background species are reduced significantly under a single set of cell conditions. Figure 3 illustrates the reduction in background from Ar+ at m/z = 40 as hydrogen flow in the cell is increased, yielding a 109 reduction in background. Since the reaction chemistry is specific to argon polyatomics, no signal is lost due to reaction of the analyte with hydrogen, as could occur with other more reactive gases. However, due to the specificity of reaction mode, there are numerous examples where it is not useful. For example, in samples where the matrix composition is unknown, or there are multiple polyatomic interferences at a single m/z, it is not possible to use reaction mode effectively. In this case, a more generic method of interference removal is needed.
Proton transfer: Ar2+ + H2 Ar2H+ + H (interferences on 78Se+ and 80Se+) ArO + H2 ArOH + H (interference on Fe)
+ + 56
Atom transfer: ArO+ + H2 H2O+ + Ar (interference on 56Fe) In all cases the Ar interference is removed from mass 40, 56, 78, and 80. Since Ca, Fe, and Se do not react with H2, there is no loss of analyte signal. Ca+ + H2 No reaction Se+ + H2 No reaction Fe+ + H2 No reaction
Figure 2.
Calibration plots for 78Se, 40Ca, and 56Fe under hydrogen reaction conditions.
1.00E+10 1.00E+09 1.00E+08 Mass 40 signal (cps) 1.00E+07 1.00E+06 1.00E+05 1.00E+04 1.00E+03 1.00E+02 1.00E+01 1.00E+00 1.00E_01 0.3 Efficient removal of 40Ar+ background, by reaction with H2 cell gas. Ar+ background is reduced from 2.5E9 cps to ~1 cps - greater than 9 orders of magnitude interference reduction 0.7 1.1 1.5 1.9 2.3 2.7 3.1 3.5 3.9 4.3 4.7 5.1 5.5 5.9 Standard (cps) Blank (cps) BEC (ppb)
1.00E+10 Background equivalent concentration (ppb) 1.00E+09 1.00E+08 1.00E+07 1.00E+06 1.00E+05 1.00E+04 1.00E+03 1.00E+02 1.00E+01 1.00E+00 1.00E_01
H2 cell gas flow
Figure 3.
Reduction in background on mass 40 for calcium using hydrogen reaction mode. In this case, as the hydrogen flow is increased to about 5 mL/min, the background at mass 40 decreased from approximately 2.5 billion cps to about 1 cps, a >109 reduction.
The Helium Collision Mode Helium collision mode can reduce or eliminate polyatomic interferences by one of two mechanisms; either CID or KED. Both are nonreactive mechanisms and so they do not form any new polyatomic ionic species that must be managed. CID can occur when the collision energy between the polyatomic ion and the collision gas (typically He) is sufficient to break the polyatomic bond. The result is two (usually atomic) fragments at lower mass, one of which will retain the charge of the original ion. A few common polyatomic
interferences are bound weakly enough for this to occur. They include NaAr+, which can interfere with the measurement of 63Cu in high sodium samples and ArO+ which interferes with iron. However, when ion energies are properly controlled, KED is the more useful of the two techniques. Kinetic energy discrimination depends on the fact that polyatomic ions are always larger in collisional cross section than monatomic ions (Figure 4), and as a result undergo more collisions and so lose more energy when traversing a pressurized collision cell. Figures 5 and 6 depict the KED process.
400
m/z 65
350 300 Approx ionic radius (pm) 250 200 150 100 50 0 Complex species
NOCl KED set at this cut-off point _ Excludes ALL interfering species, but transmits the Cu SO2
CaOH S2 ArMg
Cu
Figure 4.
Comparison of approximate ionic radii (in picometers) for copper and several polyatomic species.
Figure 5 shows the greater loss of energy of the polyatomic ion relative to the atomic ion, in this case ArCl+ relative to As+. However, for KED to be effective, it must remove the polyatomic ion effectively while not significantly reducing the response of the atomic ion. This means there must be minimal overlap in ion kinetic energies between the polyatomic and atomic ions at the exit of the octopole. For this to be the case, the energy spread of incoming ions must be less than the energy difference between analyte and polyatomic interference at the octopole exit.
As
Ar Cl
Collision
He
As
Electrical potential (Q-pole)
Ar Cl
Reaction cell Electrical potential (Octopole)
Figure 5.
KED. Polyatomic species have a larger collision cross section, and so experience more energy dissipating collisions and exit the cell with lower kinetic energy. A small stopping potential between the exit of the octopole and the entrance of the quadrupole keeps the polyatomic ions from entering the quadrupole and being detected.
Only the Agilent ORS can accomplish this, as a result of the use of the ShieldTorch system, which minimizes plasma potential and eliminates secondary ionization in the interface, which would otherwise cause broadening of the ion energy spread. It is also essential to avoid band-broadening collisions induced by high extraction voltages in the high-pressure region immediately behind the skimmer cone. On the 7500ce, this is accomplished by using soft-extraction, (extract 1 operates at 0 to +5 V), as a result of which, the mean ion energy is maintained at less than 2 eV with an ion energy spread of about 0.5 eV, ideal for the KED of plasma-source polyatomic interferences. A simplified schematic representation of ion kinetic energy and energy distribution for a typical ICP-MS system and from an Agilent 7500 ORS ICP-MS is shown in Figure 6.
Kinetic energy of atomic analytes and polyatomic interferences in the plasma is the same, only the energy spread varies
Kinetic energy of polyatomic ions is reduced relative to atomic ions through collisions in the cell
Typical ICP-MS
Agilent ORS system

Only those ions whose kinetic energy exceeds the blocking potential are allowed to pass into the quadrupole for analysis
Polyatomic ion kinetic energy
Kinetic energy barrier = b eV

Atomic ion kinetic energy
Figure 6.
Simplified schematic representation of ion kinetic energy and energy distribution of a typical ICP-MS system (upper), and from an Agilent 7500 ORS ICP-MS (lower).
After multiple collisions in the CRC, in both cases the average kinetic energy of the larger polyatomic ions is decreased relative to the smaller atomic ions by b eV. If a kinetic energy barrier is applied to the ion beam at the exit to the collision cell that is equivalent to the kinetic energy difference, b, (indicated by dashed line), then any ions whose kinetic energy is lower than the barrier will be blocked. If the energy distribution of ions is larger than the difference in average energies of the ions, only partial rejection of polyatomic ions occurs accompanied by a loss of atomic ions. While ED was described on other designs of ICP-MS systems, these systems do not have the tight control of ion energy provided by the ShieldTorch System and so the ED is only effective for reducing the very low energy polyatomics formed within the cell, typically as a result of sequential reaction chemistry, which is characteristic of the use of a highly reactive cell gas, such as NH3. Table 2 shows the reactants and products for a number of polyatomic interferences using both hydrogen and ammonia reaction mode. As can be seen, in the presence of common matrix components such as carbon and sulfur, the use of NH3 can create multiple new interferences, which must be removed. Avoiding the use of NH3 eliminates the possibility of creating new, cell-formed, polyatomic cluster ions in the first place.
Reaction of hydrogen with plasma-based polyatomics such as Ar+ is highly favored and results in elimination of the interferent. Neither hydrogen nor NH3 are effective at removing the interference from ClO+ on vanadium. Additionally, the use of ammonia can lead to reaction with other common matrix elements such as carbon and sulfur, creating new interferences such as HCN+ on aluminum and NHSH+ on titanium. Using the Agilent system with an inert cell gas and KED would eliminate the interference from ClO+ on vanadium AND, for example, ArC+ on Cr, without producing any new interferences. An excellent test of the efficiency of interference removal can be seen in low-level calibration plots. When interferences are present, the response curve will be offset in the y direction by the magnitude of the interference, increasing the background equivalent concentration (BEC) and the detection limit (DL). When the interference is removed, the calibration curve intersects the y-axis at a point much nearer to zero with a correspondingly lower BEC and DL. Figure 7 depicts sub-ppb calibration curves for chromium and vanadium in 1% each methanol, HCl, and HNO3, with and without the use of helium collision mode. Since KED does not depend on chemical reaction, it is independent of matrix concentration as well as composition.
Table 2. Comparison of Reaction Products for Several Possible Reactants Involving Hydrogen and Ammonia Reactants Ar+ H2 Ar
+ 2 +
Interfered analytes Ca+ Se Fe V+ V

+ + + +
Products H2+ ArH

+ +
Ar Ar, H None None
H2 H2 H2 NH3
+
ArO ClO C+
H2O , ArOH None None
ClO+
+
HCN
H2 NH3
Al
HCNH (28) NH3+ (17) HCN+ (27) HCNH+ (28)
H C H2 H
S+
NH3
NHSH+ (48)
7.7 ppb
0.09 ppb
1.8 ppb
0.05 ppb
Figure 7.
Calibration plots of 52chromium and 51vanadium in 1% nitric, 1% hydrochloric, 1% methanol showing the contribution of interferences from ArC+ and ClO+ in normal mode on the left and after removal of interferences by the ORS using He on the right.
Figure 8 depicts the effect of increasing HCl on the measured concentration of a 5-ppb solution of vanadium in both helium and normal (no gas) mode. Increasing the HCl from 0% to 1% causes an 80% increase in measured vanadium concentration in the no-gas mode. There is no increase in the V concentration reported for the variable sample matrix, when measured in He mode.
5-ppb Vanadium vs. HCl concentration
10.000 9.000 8.000 Reported value for V 7.000 6.000 5.000 4.000 3.000 2.000 1.000 0.000 5 ppb V 5 ppb V + 0.1% HCl 5 ppb V + 0.5% HCl 5 ppb V + 1% HCl
V 51 He V 51 Norm
Figure 8.
Effects of increasing HCl concentration on vanadium response in both normal and helium modes. 9
The Normal Mode Normal mode uses the octopole as an ion guide only with no additional gas added. In this mode, the Agilent 7500ce ICP-MS operates as a conventional (non-collision cell) instrument. Because an octopole is a highly efficient ion guide as compared with a lower-order multipole, such as a hexapole or quadrupole, there are no compromises in ion transmission efficiency and the addition of a collision gas to promote collisional focusing in normal mode is not required. Because of this, the Agilent
Table 3.
7500ce exhibits exceptional sensitivity for uninterfered low mass elements, such as lithium, beryllium, and boron. Typically, normal mode is used only for these elements, though it is also acceptable for other elements that do not require interference removal such as lead, mercury, thallium, and uranium. Examination of Table 3 will show that the DLs for the interference-free heavy metals are essentially the same in all three modes, giving the user the flexibility to select as appropriate.
Three Sigma Detection Limits (ppt). All Isotopes 1 s Total Integration Time Except Ca and Hg Which Were 3 s Total Integration Time Element Li Li Be B Na Mg Al P K Ca Ca Ca Sc Ti Ti V Cr Cr Mn Fe Fe Co Ni Cu Cu Zn Ga Ga Ge Ge As Se Se Rb Sr Y Zr STD Mode (No gas) DL (3-Sigma) BEC 4.99 17.09 1.67 14.36 0.19 0.11 5.88 47.26 3.36 148.40 0.27 0.72 3.05 50.70 418.27 12521.95 1347.50 47564.25 460.04 8520.74 2932.48 50407.443 7.95 183.06 4.07 40.86 11.49 57.33 0.40 2.52 5.53 212.670 7.98 52.87 1.69 25.24 1443.70 55093.34 444.36 23132.21 0.21 2.15 26.523 672.20 1.32 37.48 2.70 47.46 2.85 9.01 0.30 3.03 1.22 7.40 2.60 53.69 3.72 32.74 23.24 660.78 48.10 6351.29 26.92 251.29 0.27 1.06 0.19 0.84 0.10 0.26 0.07 0.09 H2 Mode (5 mL/min) DL (3-Sigma) BEC 161.491 95.83 23.755 26.31 6.932 2.62 83.182 128.03 62.682 313.64 2.570 1.75 8.079 5.51 29.532 118.74 2.936 7.13 129.640 121.68 102.104 121.94 6.446 19.41 32.197 19.88 17.535 10.46 1.309 0.73 19.919 68.63 28.504 82.65 1.362 4.10 5.034 20.21 66.614 261.56 0.816 0.26 71.224 742.27 20.271 20.14 27.845 27.76 1.612 1.83 0.273 0.19 0.125 0.15 1.448 1.14 5.556 2.74 14.130 14.00 2.396 2.556 15.225 56.51 0.349 0.57 0.072 0.04 0.054 0.04 0.709 0.18 He Mode (4 mL/min) DL (3-Sigma) BEC 331.32 142.58 9.38 20.55 7.812 2.61 48.21 107.28 37.65 299.38 3.37 1.41 37.56 53.52 1903.62 3800.52 2838.73 27943.17 191.85 742.92 48.01 352.38 1.34 6.46 4.98 3.52 5.69 1.70 0.42 0.19 3.10 22.70 8.70 21.60 4.25 12.23 53.99 451.26 71.06 215.90 0.38 0.30 41.70 491.67 6.37 68.35 7.52 59.76 1.84 2.18 0.82 1.04 1.80 2.22 3.32 7.55 7.30 6.04 10.72 65.72 48.93 195.45 20.63 116.01 0.72 0.34 0.38 0.13 0.16 0.04 0.15 0.06
Mass 6 7 9 11 23 24 27 31 39 40 43 44 45 47 49 51 52 53 55 56 57 59 60 63 65 66 69 71 72 73 75 78 82 85 88 89 90 10
Table 3.
Three Sigma Detection Limits (ppt). All Isotopes 1 s Total Integration Time, Except Ca and Hg Which Were 3 s Total Integration Time (Continued) Element Nb Mo Ru Rh Pd Ag Cd In Cd Sn Sb Te Te I Cs Ba La Ce Pr Nd Sm Eu Gd Tb Dy Dy Ho Er Tm Yb Lu Hf Ta W W Re Ir Pt Au Hg Hg Hg Tl Pb Pb Pb Bi Th U STD Mode (No gas) DL (3-Sigma) BEC 0.12 0.20 0.32 0.67 0.60 1.40 0.08 0.11 0.30 0.27 0.23 0.33 0.56 0.83 0.07 0.11 0.33 0.41 0.24 0.43 0.11 0.08 1.96 0.94 1.12 1.64 2.02 21.73 0.09 0.04 0.22 0.20 0.17 1.94 0.223 2.65 0.11 0.25 0.39 0.44 0.22 0.17 0.02 0.03 0.17 0.14 0.03 0.02 0.18 0.15 0.15 0.08 0.04 0.00 0.15 0.06 0.02 0.02 0.11 0.09 0.04 0.020 0.13 0.08 0.04 0.06 0.32 0.35 5.07 1.07 0.12 0.07 0.09 0.10 0.14 0.17 0.22 0.11 0.82 2.00 1.11 2.54 0.86 1.84 0.20 0.24 0.33 0.84 0.51 0.94 0.47 0.712 0.05 0.04 0.04 0.04 0.05 0.04 H2 Mode (5 mL/min) DL (3-Sigma) BEC 0.68 0.53 14.53 4.33 19.27 5.74 4.24 1.75 10.11 6.39 1.42 1.61 0.32 0.20 0.05 0.03 0.40 0.47 0.55 0.54 0.21 0.010 2.05 1.29 2.08 1.67 3.57 22.30 0.06 0.04 0.38 0.16 2.49 2.46 2.18 3.21 0.12 0.28 0.43 0.50 0.11 0.04 0.04 0.02 0.15 0.05 0.01 0.01 0.05 0.05 0.08 0.04 0.02 0.01 0.05 0.02 0.02 0.01 0.07 0.02 0.02 0.01 0.06 0.06 0.06 0.049 1.39 0.5 0.87 0.42 0.07 0.05 0.25 0.08 1.94 0.52 1.76 0.43 1.04 1.78 2.07 2.29 0.75 1.77 0.13 0.22 0.28 0.64 0.25 0.69 0.40 0.55 0.03 0.02 0.03 0.01 0.04 0.01 He Mode (4 mL/min) DL (3-Sigma) BEC 0.15 0.05 0.31 0.1 0.28 0.09 0.05 0.0 0.33 0.15 0.28 0.4 0.86 0.54 0.08 0.05 0.34 0.23 0.91 0.73 0.46 0.25 9.57 4.12 7.33 4.27 7.71 20.41 0.15 0.06 0.9 0.38 0.66 2.14 0.47 2.88 0.12 0.31 0.70 0.73 0.58 0.24 0.11 0.03 0.35 0.22 0.055 0.03 0.23 0.17 0.23 0.16 0.05 0.02 0.17 0.09 0.03 0.03 0.30 0.18 0.06 0.03 0.32 0.15 0.11 0.08 0.56 0.33 0.43 0.47 0.12 0.08 0.33 0.15 0.22 0.18 0.18 0.07 1.15 2.18 1.56 2.58 0.59 1.91 0.35 0.30 0.34 0.73 0.95 1.05 0.53 0.755 0.06 0.05 0.06 0.05 0.05 0.044 11
Mass 93 95 101 103 105 107 111 115 116 118 121 125 126 127 133 137 139 140 141 146 147 153 157 159 161 163 165 166 169 172 175 178 181 182 183 185 193 195 197 200 201 202 205 206 207 208 209 232 238
Conclusions
The Agilent 7500ce ICP-MS has achieved its design goals of providing sensitive, robust, interferencefree analysis of difficult, high-matrix samples. With five times the sensitivity of its predecessor, nine operating orders of dynamic range and increased matrix tolerance, it is capable of replacing both GFAA and ICP-OES instruments in addition to older generation ICP-MS systems. The 7500ce is unique in offering a single solution for multielemental analysis of complex and variable, high matrix samples, while allowing the operator the freedom to use simple and consistent sets of instrument conditions for almost all elements in almost any matrix.
3. E. McCurdy and G. Woods, The Application of collision/reaction cell inductively coupled plasma mass spectrometry to multi-element analysis in variable sample matrices, using He as a non-reactive cell gas (2004) JAAS 19, (3). 4. S. D. Tanner, V. I. Baranov, and D. R. Bandura, (2002) Spectrochimica Acta Part B, 57, 1361. 5. Using automated collision cell ICP-MS with rapid in-sample switching to achieve ultimate performance in environmental analysis Thermo Electron Corporation Application Note AN_E0640, (2003).

References
1. Real World Analysis of Trace Metals in Drinking Water using the Agilent 7500ce ICP-MS with Enhanced ORS Technology Agilent Technologies publication 5989-0870EN www.agilent.com/chem 2. Analysis of High Matrix Environmental Samples with the Agilent 7500ce ICP-MS with Enhanced ORS Technology Agilent Technologies publication 5989-0915EN www.agilent.com/chem
Determination of Mercury in Microwave Digests of Foodstuffs by ICP-MS Application
Food Safety
Author
John Entwisle LGC Queens Road Teddington Middlesex TW11 0LY, UK
Abstract
The quantitative determination of mercury in foodstuffs is presented using a 7500i ICP-MS. Microwave digests were prepared and then analyzed by ICP-MS. To avoid memory effects often experienced with mercury, gold was added offline to all standards/samples and wash solutions to act as a cleansing agent. The instrumental setup used a second vacuum pump, the integrated sample introduction system in the high sample throughput mode, and a microflow concentric nebulizer. This allowed the robust and rapid determination of mercury in the digests at the ppt range. Excellent agreement with the certified value was obtained for two certified reference materials and stability of the system was demonstrated over a 36-hour analytical run.
Introduction
The determination of sub-ppb concentrations of mercury has always been of special importance in the field of trace metal analysis. Even at trace levels, mercury is toxic and causes neurological damage, particularly in fetuses and young children. Anthropogenic sources of mercury in the environ-
ment include coal-fired power stations and chlor-alkali works. In the aquatic environment, bacteria convert elemental mercury Hg(0) to methylmercury which is accumulated and passed up the food chain. It has been reported that some whale meat contains 5000 times the Japanese legal limit of 0.4 g/g. In addition, fish and shellfish are significant contributors to the human diet. Today mercury pollution is a global problem and extensive monitoring of foodstuffs is required. Therefore fast efficient and robust methods are needed. Mercury however is recognized as a problem element. It is known to adsorb onto the walls of storage containers and volatilize as mercury vapor. Additionally, its high first ionization potential and numerous isotopes have limited its sensitivity in ICP-MS analysis. ICP-MS allows the rapid determination of ultratrace levels of metals in food digests, however, extensive washout times have been required to reduce carryover for mercury analysis. Other workers have tried the addition of a number of chemical agents in the past. One of the most effective washout agents is gold chloride. To avoid memory effects and ensure stability, gold chloride (at the 5-ppm level) was added offline to all samples/standards and wash solutions. Extensive washout times were reduced by using by the integrated sample introduction system (ISIS). With the use of the high throughput pump, a large flush volume can be pumped through in a much shorter time. By summing the responses for multiple isotopes (199, 200, 201, and 202) and with the help of a second interface vacuum pump and a micro-flow concentric nebulizer, detection levels of between 10 and 30 ppt were routinely achieved for the digests.
Procedure
Microwave Digestion Varying aliquots of each sample (generally between 0.2 and 0.6 g, depending on the moisture content of the sample) were weighed to the nearest 0.01 g into the digestion vessels. Wet oxidation was induced using concentrated, ultra-high purity nitric acid (10 mL, from Romil LTD, Cambridge, UK) with the addition of a 0.2 mL of concentrated hydrochloric acid (Romil LTD, Cambridge, UK). Oxidation was carried out in heavy-duty vessels (HDV) using a high-pressure microwave digestion oven (Mars 5 from CEM). Temperature control was used as opposed to pressure control. Samples were ramped to 180 C over 20 minutes and held at 180 C for 10 minutes before cooling to below 50 C before venting the vessel. Both pressure and temperature were monitored by direct measurement throughout the digestion to ensure that samples attained the critical temperature of 180 C, at which food components, such as fat, are digested. The sample digests were then made up to 100 g using ultrahigh purity water (18 mega ohms, from Elga Maxima). The resultant solution was used for determination. Operating and Acquisition Parameters Ten milliliter portions of the sample digests were accurately pipetted into sample tubes, and using a micropipette, 20 L of a 1000-ppm gold chloride solution (Romil LTD, Cambridge, UK) was added. This gives a final gold concentration of 5 ppm in solution. Fifty milliliters each, of blank and four standard solutions covering the range, were prepared from a 100 g/g stock mercury solution (from SPEX CertiPrep Assurance, Metuchen, New Jersey, USA). Ten percent wt/wt nitric acid containing 5 ppm of gold was used as the wash solution for the autosampler and nebulizer. Gold is thought to have its effect by acting as an oxidizing agent ensuring that mercury stays in an ionized form in solution. Gold was added at elevated levels to ensure that any residual amounts of organic compounds in the digests would not reduce Au(III) to elemental gold and render it ineffective. A 250-ppb Thalium standard was added online as an internal standard (ISTD), using the ISIS system. There was an online dilution factor of 1:20. Gold chloride was also added to the standard solutions at 5 ppm.
Instrument Conditions
Plasma gas flow rate Carrier gas flow rate Make-up flow RF Power Nebulizer Spray chamber Spray chamber temperature ICP Torch injector Sample tubing Internal standard tubing Instrument Peri pump Sample/Skimmer cones Rotary pumps Autosampler 16 L/min 0.85 L/min 0.14 L/min 1400 Watts Agilent micro-flow 100 L Glass double pass Cooled to 2 C 2.4 mm 0.89 mm id 0.19 mm id 0.1 rps Nickel 2 AX500
Acquisition Parameters
Mass 199201 202 205 Element Hg Hg Tl Integration/Point 3.5 3.5 0.05 3 43.79 s 3 131 s Time (s)/Mass 10.5 10.5 0.15
Number of points per mass: Acquisition time: Number of repetitions: Total acquisition time:
Peristaltic Pump Program Memory effects arise when the analyte signal is enhanced due to contributions from previous high concentration sample. This is due to adsorption/desorption of mercury in the sample introduction system. As a result, the analyst has to program long washout times. With the use of ISIS this wash-out time can be reduced. ISIS Peristaltic Pump Program
Analysis Speed : 0.10 rps Before acquisition Uptake speed Uptake time Stabilization time (undiluted) After acquisition (probe rinse) Rinse speed Rinse time (sample) Rinse time (standard) After acquisition (rinse) Rinse vial Uptake speed Uptake time (undiluted) Stabilization time 0.80 rps 32 s 150 s 0.80 rps 8s 8s 1 0.8 rps 32 s 85 s
12
10
Hg/Tl ratio (Tl 12 ng/g)
R2 = 0.998
Aquacheck 217 independent QC standard recovery 106.9% 8% (n = 10, K = 2)

0 0 1000 2000 3000 4000 Hg concentration pg/g 5000 6000
Figure 1. Calibrations over 36 hours continuous operation.

0.140 0.120 Concentration Hg mg/kg 0.100 0.080 0.060 0.040 0.020 0.000 0 2 4 6 8 10 Determination number 12 14 16
Figure 2. NIST 1547 Peach leaves (bottom trace) and LGC 7160 Crab paste (top trace) were analyzed 14 times over a period of 1 month. Each point represents a separate digest. LGC 7160 certified value 0.096 = 0.108 mg/kg, NIST 1547 certified value 0.031 = 0.007 mg/kg.
Tee joint
Pump 1 1
Pump 2
ISIS Autosampler AX500
ISTD Agilent 7500i Waste
The sample is split at the first Tee joint. A constant quantity of sample is introduced to the second Tee joint (by pump 2) where it is mixed with ISTD before going to the nebulizer. When going to the next sample the ISIS pump (pump 1) turns at high speed to shorten the sample and washout transfer times.
Rotary pumps
Figure 3. Schematic of ISIS in high sample throughput mode. 3
Results
Five calibration plots over a 36-hour period are shown in Figure 1 demonstrating excellent stability of the system. As can be seen, excellent linearity was achieved over an extended calibration range demonstrating that memory effects had been effectively eliminated. Between the calibrations over 120 various microwave foodstuff digests were analyzed.
Conclusion
This procedure proved robust, as in excess of 500 samples were analyzed in runs lasting in excess of 36 hours without loss of sensitivity. The 7500i ICP-MS has shown to be a robust and sensitive tool for the analysis of foodstuff for mercury. The system proved stable over an extended time period. The use of the ISIS enables the operator to take advantage of the high sample throughput possible with this technique. The use of Au(III), in particular, shortens washout time considerably, reducing the possibility of carry-over. The sensitivity of the 7500i ICP-MS gives method detection limits at low ppt levels in the foodstuff digests. The extended calibration range reduces the need for dilutions and reruns.
Quality Control
An Aquacheck proficiency testing material solution from the Water Research Council (1010 ppt) was analyzed 10 times during the run and a mean recovery of 106.9% (8%) was achieved (Figure 1). This material was analyzed throughout a run of 120 various foodstuffs samples and acted as a quality control for the quantitation. Two certified reference materials (CRMs) were analyzed throughout a survey of 500 samples. These acted as quality control materials for the microwave digestion as well as the quantitation. The CRMs used were a crab pasteMetals LGC 7160, 0.096 mg/kgand peach leavesNIST 1547, 0.031 mg/kg. Each CRM was analyzed 14 times on different runs during a survey of more that 500 samples of various foodstuffs over a 1-month period. The results can be seen in Figure 2.

Fast and Accurate Determination of Arsenobetaine (AsB) in Fish Tissues Using HPLC-ICP-MS Application
Foods
Author
Raimund Wahlen LGC Limited, Queens Road, Teddington, Middlesex, TW11 0LY, United Kingdom
Introduction
The element Arsenic (As) has long been thought of as poisonous and highly toxic. However, it has since been shown that the toxicity of As is largely dependent on the form or species the arsenic is in. Arsenic is ubiquitous in the environment due to natural and anthropogenic sources, and the relative contribution of these factors is estimated as roughly 60% and 40% respectively. In the environment, As behaves in similar ways to the Group V elements nitrogen (N) and phosphorus (P). As a result of these similarities, arsenic gets taken into the biochemical pathways of N and P. This results in the formation of compounds such as arsenobetaine (AsB) in fish and arseno-sugars, which are found in marine algae. The toxicity of the inorganic As-species (such as arsenite, As(III) and arsenate, As(V)) is far greater than the organic forms, such as monomethylarsonic and dimethylarsinic acid (MMAA and DMA) and AsB. The International Agency for Research on Cancer (IARC) has classified inorganic arsenic as a human carcinogen, whereas AsB, the predominant form of As in most marine organisms [1], is considered nontoxic to humans. Although AsB is the major form of As in many marine organisms, it is not present in all fish species [2]; therefore, an evaluation of the proportion of AsB to the total As determined can give a useful and rapid estimate of the toxicological significance of a sample. In order to determine the toxicity of seafood, the determination of the total As alone is of limited value, and the different species of As have to be extracted, separated, and determined. Fast, reliable, and practical methods are therefore required that can provide speciation information for the screening of large sample batches.
Abstract
A high performance liquid chromatography-inductively coupled plasma mass spectrometry method was developed for the fast and accurate analysis of arsenobetaine in fish samples extracted by accelerated solvent extraction. The combined extraction and analysis approach was validated using certified reference materials for arsenobetaine in fish and during a European intercomparison exercise with a blind sample. Up to six species of arsenic can be separated and quantified in the extracts within a 10-minisocratic elution. The method was optimized so as to minimize time-consuming sample preparation steps and to allow for automated extraction and analysis of large sample batches. A comparison of standard addition and external calibration showed no significant difference in the results obtained, which indicates that the liquid chromatography-inductively coupled plasma mass spectrometry method is not influenced by severe matrix effects. The extraction procedure could process up to 24 samples in an automated manner while the robustness of the developed high performance liquid chromatographyinductively coupled plasma mass spectrometry approach is highlighted by the capability to run more than 50 injections per sequence which equates to a total run-time of more than 12 hours. The method can therefore be used to rapidly and accurately assess the proportion of nontoxic arsenobetaine in fish samples with high total arsenic content during toxicological screening studies.
Aims and Objectives

The aim of this study was to develop a semiautomated analytical method for the extraction and determination of As-species in fish tissues. Requirements for high sample throughput analysis were the automation of the extraction procedure as well as a fully automated separation and detection method capable of analyzing large sample batches (up to 50 injections per run) during overnight runs. In order to streamline the analytical procedure, an attempt was made to develop a method with a minimal number of sample preparation steps. It was intended that the method should be established using calibration by external calibration curves, rather than the lengthy alternative of standard additions. The use of an isocratic liquid chromatography (LC) elution can be favorable in terms of time-efficiency during the liquid chromatographyinductively coupled plasma mass spectrometry (LC-ICP-MS) analysis because it negates the need for column re-equilibration between injections.
Table 1.
Extraction Conditions Used for ASE Dionex ASE 200 2 min 5 min 5 2 min 100 C 1500 psi Methanol
Instrument Preheat Heat Extraction steps Temperature Pressure Solvent
HPLC-ICP-MS Methodology
The HPLC-ICP-MS instrumentation consisted of an Agilent Technologies 1100 HPLC system coupled to an Agilent Technologies 7500i ICP-MS fitted with a second roughing pump, which enhances sensitivity by increasing ion transmission across the interface. The HPLC system comprised a quaternary pump module, a vacuum degasser, a temperature controlled autosampler, and column compartment. The ICP-MS instrument was tuned for sensitivity, reduced oxides, and doubly charged species prior to connection to the liquid chromatograph by performing a standard instrument tune using a 10 ng/g solution of Li, Y, Ce, and Th in 1% HNO3. The pulse to analog (P/A) factor was adjusted on a daily basis using a solution containing ~50 ng/g Li, Mg, Mn, Cu, As, Gd, Y, Cd, Pb, and Ba. After this optimization, a 50 ng/g solution of As in 1% HNO3 was used to specifically optimize the sensitivity for arsenic. The ICP-MS nebulizer was then connected to the HPLC-column using a length of PEEK tubing (yellow, 1/16-inch od, 0.007-inch id). See Table 2 for the ICP-MS conditions used.
Table 2. RF Power RF Matching Sampling depth Carrier gas flow Make up gas flow Optional gas Spray-chamber temperature Cones Isotopes monitored ICP-MS Conditions Used for HPLC-ICP-MS Determination of As-Species 14301550 W 1.891.92 V 4.04.8 mm 0.890.93 L/ min 0.100.14 L/ min Oxygen at 5% 0 C Platinum
75
Calibration Standards
The following standards were obtained from Fluka (Sigma-Aldrich, Gillingham, UK): di-sodium hydrogen arsenate heptahydrate (AsHNa2O4.7H2O) 98.5%, sodium (meta)arsenite (AsNaO2) 99.0%, and cacodylic acid (dimethylarsinic acid, DMA, C2H2AsO2) 99.0%; monomethylarsonic acid disodium salt (MMAA, CH3AsNa2O3) >98% was obtained from Argus Chemicals (Vernio, Italy). Arsenobetaine (AsB, C5H11AsO2) was obtained from BCR (Brussels, Belgium) as a solution of AsB in water at 1031 6 (95% C.I.) mg/kg (BCR 626).
Extraction
Accelerated solvent extraction (ASE) has been used previously for As-speciation [3, 4] and was chosen as the sample preparation method because it allows for the automated extraction and online filtration of up to 24 samples. In addition, the extraction solution is collected in glass vials, which negates further sample preparation steps such as filtration or centrifuging. The samples were extracted using a Dionex ASE 200 accelerated solvent extractor. Sample sizes from 0.10.3 g were weighed accurately into 11-mL stainless steel extraction cells fitted with filter papers and PTFE liners. The extraction program was set up as shown in Table 1.
As Rh 77 Se (40Ar37Cl) to monitor Cl interferences 53 Cr (40Ar13C) to monitor C interferences

103
Other parameters
Injector diameter: 2.4 mm Nebulizer 100 L/min PFA, Two interface pumps used
In order to develop a rapid chromatographic separation of the main As-species in fish tissues, an anion exchange column (Hamilton PRP X-100) was chosen in combination with an isocratic elution profile. Several mobile phases were tested and the best separation of AsB and As(III) as well as DMA and MMAA was achieved within 10 min using 2.2-mM NH4HCO3/2.5-mM tartaric acid at pH 8.2 delivered at 1 mL/min isocratic flow. This evaluation was carried out initially using matrix-free calibration standards containing the species of interest and refined using an oyster tissue extract that contained arsenocholine (AsC), two arsenosugars (As-sug. B and As-sug. D), TMAs+ and several unknown species in addition [5]. The injection volume for samples and standards was 50 L. In order to enhance the ionization of the As-species [6, 7], methanol was added to the mobile phase at concentrations ranging from 0.5% to 5% v/v. At concentrations above 1%, the chromatographic separation degraded significantly to the degree that base-line resolution between AsB and As(III) was no longer achieved. However, the addition of 1% MeOH to the mobile phase resulted in a significant improvement in the sensitivity (34-fold increase in peak height) for all analytes. A chromatogram for a 5-ng/g mixed calibration standard with the final chromatography conditions is shown in Figure 1.
Variations in Signal Response for Different As-Species

The chromatogram shows that the four species analyzed here have very different response factors with this method, even when made up to contain the same concentration of As in solution. This is further illustrated by the calibration curves and their respective slopes, as shown in Figure 2. Such differences in the analyte signal intensity were reported previously in the literature [7] and appear to be due to a combination of the ICP-MS hardware used and the plasma conditions, which are in turn affected by the mobile phase composition. This points to possible differences in the nebulization, transport and/or ionization of different species by such methods. In order to determine whether this effect could be attributed to the coupling of the ICP-MS with a liquid chromatograph, aqueous standards of AsB and As(III) were made up to equivalent concentrations as As and analyzed by direct aspiration without chromatography. This indicated that the signal response of AsB was ~10%15% higher compared to the inorganic As standard and, therefore, the difference in signal response does not appear to be related to the coupling with a liquid chromatograph.
14000
AsB
12000
10000 Abundance (counts)
8000
DMA As(III)
6000
4000
MMAA
2000
0 2.00 4.00 6.00 8.00 Time (min) 10.00 12.00 14.00
Figure 1.
Chromatography A: 2.2-mM NH4HCO3 2.5-mM tartaric acid, 1% MeOH, pH 8.2, Hamilton PRP X-100 column. Concentration of standard ~ 5 ng/g as As.
3000000
2500000
AsB y = 51320x R2 = 0.9996 DMA y = 42505x R2 = 1
2000000 Peak area
1500000
1000000
MMAA y = 28061x R2 = 1 As(III) y = 20235x R2 = 0.9965

0 10 20 30 ng/g As 40 50 60
500000
Figure 2. Calibration curves for AsB, DMA, MMAA and As(III) over a range of 050 ng/g as As.
In order to increase the signal intensity for species such as As(III) and MMAA by the approach described here, additional MeOH was added via a T-piece post-column so as not to impact on the chromatographic resolution. Although the relative volume of MeOH could be increased by 50%70% in this way without deteriorating plasma stability, the relative signal responses of the four species were not influenced significantly. Because the relative signal response was stable on a day-to-day basis, no further attempts were made to equalize the signal responses. The instrumental detection limit for AsB by this method was 0.04 ng/g as As. The linearity obtained, as indicated by the correlation coefficient of the calibration line, was 0.9991.000 over a calibration range of 0700 ng/g as As.
Plasma Disturbance Due to Elution of MeOH

During the analysis of fish samples, which had been extracted under the ASE conditions highlighted in Table 1, a disturbance of the plasma was observed between ~2.3 to 4.3 min after injection. This affected all of the isotopes monitored and the effect on 75As and 103Rh is highlighted in Figure 3. As can be seen from the chromatogram, the effect on these two isotopes is nonlinear. The 103Rh signal decreases significantly during this time, whereas the shoulder on the tailing side of the AsB peak indicates an increase in the 75As signal.
40000 36000
AsB
32000 28000
103 Abundance (counts)
Rh
24000 20000 16000 12000 8000 4000 0 2.00 4.00 6.00 8.00
Time (min) 75
DMA
10.00
As
12.00
14.00
Figure 3.
Signals for 103Rh and 75As in an undiluted fish extract. Notice the increase in the 75As signal at the tailing side of the major peak (AsB) coinciding with the decrease in the 103Rh signal.
The observed fluctuation in the signal intensities for the different isotopes coincides with the elution of the organic methanol fraction of the fish extracts from the analytical column. This effect could be reduced slightly by lowering the temperature of the spray-chamber from 5 C to 0 C, but the effect was not completely eliminated. During the injection of undiluted sample extracts, the volume of methanol that passes through the column and into the ICP-MS is ~10%. It has already been discussed that the addition of MeOH enhances the 75As signal by increasing the ionization efficiency of this analyte; this effect is observed on a small scale here. Although there is no detectable As(III) in this fish material, the accurate quantitation of this compound (compared to aqueous calibration standards) could obviously lead to an overestimation if the signal of this analyte is enhanced due to the simultaneous elution of MeOH from the column. In this case, a standard addition calibration would represent a more accurate approach for quantitation. However, the spiking of each sample extract at different levels, which is necessary for this type of calibration, would make such an approach less suitable for a high sample-throughput application. In addition, the accurate integration of AsB is influenced by the signal increase on the tailing side of the peak.
In order to eliminate the effect of these signal variations on the accurate quantitation of the As-species in the methanolic extracts, the methanol fraction could either be reduced by evaporation or dilution with water. Dilution was chosen as the preferred option over evaporation in order to avoid possible analyte losses and because of time-efficiency. Whereas evaporation would either involve passing an inert gas over the solution or using rotary evaporation equipment, gravimetric dilutions were easily and quickly achieved by pipetting an aliquot of the extract into a sealed HPLC autosampler vial, weighing, and then adding the appropriate amount of water. In order to observe the effect of different dilution factors on the observed plasma disturbance, a fish extract was diluted 10-, 5-, and 2-fold in water and also injected undiluted. The effects of the different dilutions on the 103Rh signal are shown by the chromatograms in Figure 4. As demonstrated in Figure 4, a 10-fold dilution is sufficient to eliminate the plasma disturbance sufficiently; therefore, all extracts were diluted 1:10 in water prior to injection.
Abundance (counts) 30000 a) Undiluted
10000 2.00 30000 b) 2-fold 4.00 6.00 8.00 10.00 12.00 14.00
10000 2.00 30000 c) 5-fold 10000 2.00 30000 d) 10-fold 10000 2.00 4.00 6.00 8.00 Time (min) 10.00 12.00 14.00 4.00 6.00 8.00 10.00 12.00 14.00 4.00 6.00 8.00 10.00 12.00 14.00
Figure 4.
Signal of the internal standard 103Rh, for fish sample extracts a) undiluted and diluted b) 2-fold, c) 5-fold and d) 10-fold.
Comparison of External Calibration and Standard Addition for the Quantitation of AsB in Fish Tissues
Due to the fact that arsenic is mono-isotopic, isotope dilution analysis cannot be used for the highaccuracy quantitation of this compound by LC-ICP-MS. In such circumstances, calibration by standard additions is often used in order to achieve matrix matching of standards and samples. It is also a useful technique in chromatographic applications where the possibility of retention time (RT) shifts of analytes due to matrix components exists. This can result in misidentification, and thus erroneous results. However, standard addition calibration can be very time- consuming because several aliquots of the sample require spiking with different levels of a calibration standard, and at least three levels of standard addition are needed for accurate quantitation of the same sample. External calibration by non-matrix matched standards can be used for applications where the difference in the matrix between samples and standards does not influence the accuracy of the result to a significant extent. Standard addition calibration and non-matrix matched external calibration were compared for AsB in two certified reference materials (DORM-2, Dogfish muscle, NRC Canada and BCR 627, Tuna Fish, BCR EU) in order to assess whether the
6
calibration technique used significantly influenced the accuracy or precision of the analytical result. The results showed that there was no significant difference in the mean results determined by the different calibration techniques with this method. The mean results for repeat analysis of both materials showed that the difference in the DORM-2 material was less than 1.4% and less than 4.5% for the BCR 627 material. When taking into account the standard deviations (SD) associated with the mean result obtained by each calibration technique, there was no statistically significant difference between the AsB results obtained by either approach in either of the fish tissue certified reference materials (CRMs).
Results of CRM Analysis

In order to test the accuracy of the developed ASE extraction and HPLC-ICP-MS method, a variety of certified and candidate reference materials of marine origin were extracted and analyzed. The samples included the certified fish reference materials DORM-2 and BCR 627, as well as an oyster tissue material (BCR 710)*, which is pending certification.
* The "MULSPOT" project has been financed by the SM&T Program (EU) (Contract SMT4-CT98-2232) and coordinated by ENEA (IT). The Project is at the certification stage and the material is not yet available on the market.
Table 3.
Data Obtained for AsB in Two CRMs and a Candidate Reference Material Measured value 16.3 0.9 (1 SD) 3.69 0.21 (1 SD) 31.8 1.1 (1 SD) Certified value 16.4 1.1 (95% C.I.) 3.90 0.22 (95% C.I.) 32.7 5.1 (1 SD)
Expressed as mg/kg As unless otherwise stated DORM-2 (Dogfish muscle) BCR 627 (Tuna fish) BCR 710 (Oyster tissue) (Concentration as species)
The data shown for this material is based on the consensus mean of the final certification round after the removal of statistical outliers.
Subsamples of the different materials (n = 46) were extracted, diluted in water, and analyzed as described above. The data for AsB determined in these samples is shown in Table 3. A chromatogram of the tuna fish material BCR 627 is shown in Figure 5. The chromatogram indicates that the major species in this sample is AsB with two minor species, which were also extracted and detected. One peak was identified as DMA, and the peak labelled P1 is most likely to be AsC from RT matching. The data in Table 3 shows that the combined ASE/HPLC-ICP-MS methodology is capable of delivering accurate and reproducible results for AsB in these matrices. In addition, the extraction of other minor species, such as DMA and AsC, was achieved in the fish tissues; up to six species were extracted and separated in the oyster material, although none of these (apart from DMA) were quantified during this study. This DMA data for BCR 710 (730 30 ng/g DMA) showed a good agreement with the consensus mean value of the certification round (820 200 ng/g DMA).
Evaluation of Method Performance During a CRM Feasibility Study

The method performance was assessed in comparison to a number of European expert laboratories during the SEAS feasibility study organized by the The University of Plymouth Enterprise Limited and sponsored by the European Union (BCR, EU). A fish material was prepared for this intercomparison by the University of Plymouth and distributed to participating laboratories. Participants were asked to determine AsB in a fish material from two different bottles using a methodology of their choice and making their determinations, at least, in duplicate on separate days. The developed As-speciation method was used to extract and analyze the fish samples provided. A total of 12 subsamples from the two bottles were
The SEAS feasibility study was co-ordinated by The University of Plymouth Enterprise Limited (Plymouth, UK) under the EC contract: G6RD CT2001 00473 "SEAS" with the title: 'Feasibility Studies for Speciated CRMs For Arsenic in Chicken, Rice, Fish and Soil and Selenium in Yeast and Cereal'.
7000
6000
AsB
Abundance (counts)
5000
4000
3000
2000
1000
P1
0 2.00 4.00
DMA
6.00
8.00 Time (min)
10.00
12.00
14.00
Figure 5.
Chromatogram of a tuna fish extract (BCR 627) enlarged to show the detection of minor species in this material. 7
extracted and analyzed on 3 different days. The data were combined to provide the value labelled LGC in Figure 6 below. The error bars indicate the SD of the mean of individual results. The mean of all result (excluding a statistical outlier) together with 1 SD above and below the mean is indicated by the solid and dashed horizontal lines, respectively. The data provided by the combined ASE extraction and developed LC-ICP-MS methodology (94.92 3.95 mg/kg AsB) is in very good agreement with the mean result of all labs (95.72 7.79 mg/kg AsB, n = 11). The precision achieved was also satisfactory at 4.2% (RSD) for 12 subsamples from different bottles analyzed on 3 separate days. The performance of the method in this international intercomparison is highlighted by the good agreement with data provided by several European expert laboratories with longstanding expertise in As-speciation analysis. It should also be noted that the intercomparison was carried out with a blind sample of unknown concentration, rather than based on the analysis of a CRM with known certified values.
Conclusions
A robust and practical method has been developed based on accelerated solvent extraction and HPLC-ICP-MS analysis for the fast and accurate determination of AsB in fish samples. The benefits of the methods include automated extraction of up to 24 samples, minimal sample preparation steps (dilution only) after extraction, and rapid and automated analysis by HPLC-ICP-MS. The separation of four to six species of toxicological interest is achieved within 10 min using an isocratic elution. This increases the sample throughput by negating the column equilibration period needed with most gradient elution profiles. The method was validated using commercially available CRMs and during a European intercomparison study with a fish sample of unknown concentration. The performance of the method was very satisfactory in terms of both accuracy and precision compared to several other expert laboratories. This method can be used to rapidly determine the nontoxic proportion (AsB) in fish samples with high total As content and could therefore be used to determine whether a particular sample poses a toxicological risk in the food chain.
120
AsB in fish
110
AsB measured (mg\kg)
100 90 80 70 60 50 40 1 2 3 4 5 6 LGC 7 8 9 10 11
Solid line: Mean of means: 95.72 mg/kg Dashed lines - 1 SD: 7.79 mg/kg
Laboratory
Figure 6.
Comparison of data submitted by 12 participants for the determination of AsB in fish during the SEAS feasibility study. The error bars associated with the individual data points represent 1 SD of analysis of separate subsamples.
Acknowledgements
The work described in this application note was supported under contract with the Department of Trade and Industry (UK) as part of the National Measurement System Valid Analytical Measurement (VAM) program.
References
1. S. Branch, L. Ebdon, and P. ONeill (1994) J. Anal. At. Sprectrom., 9, 33-37. 2. J. S. Edmonds, Y. Shibata, K. A. Francesconi, R.J. Rippingale, and M. Morita (1997) Appl. Organomet. Chem., 11, 281. 3. P. A. Gallagher, S. Murray, X. Wei, C. A. Schwegel, and J. T. Creed (2002) J. Anal. At. Spectrom., 17, 581-586. 4. J. W. McKiernan, J. T. Creed, C. A. Brockhoff, J. A. Caruso, and R. M. Lorenzana (1999) J. Anal. At. Spectrom., 14, 607-613. 5. S. McSheehy, P. Pohl, R. Lobinski, and J. Szpunar, (2001) Analyst, 126, 1055-1062. 6. E. H. Larsen and S. Strup (1994) J. Anal. At. Sprectrom., 9, 1099-1105. 7. U. Kohlmeyer, J. Kuballa, and E. Jantzen (2002) Rapid Commun. Mass Spectrom,. 16, 965-974.

A Comparison of GC-ICP-MS and HPLC-ICP-MS for the Analysis of Organotin Compounds Application
Environmental
Author
Raimund Wahlen LGC Limited, Queens Road Teddington, Middlesex TW11 0LY, United Kingdom
Abstract
An inductively coupled plasma mass spectrometer (ICP-MS) was used as a detector for gas chromatography (GC) and high performance liquid chromatography (HPLC) analysis of organotin compounds. ICP-MS is a highly sensitive detector with detection limits in the pg-ng range, as well as enabling calibration by isotope dilution mass spectrometry (IDMS). Calibrating using isotopically labeled organotin species reduces measurement uncertainties and leads to greater precision compared to external calibration methods. This application note details the relative merits of the two techniques for the analysis of organotin compounds.
Introduction
The toxic effects of organotin compounds in the environment have been well documented [1] and have led to extensive research into analytical methodologies for their determination in a variety of matrices. The widespread use of organotin compounds has resulted in their detection in most marine and fresh-water sediments as well as in open-ocean waters [2]. In recent years, the focus of research in organotin analysis has begun to include matrices with human health implications, such as seafood [3], manufactured products (PVC pipes used for drinking water distribution [4]), and human blood samples [5].
Organotin analysis has traditionally been performed by chromatographic separation (gas chromatography (GC) or high performance liquid chromatography (HPLC)) coupled to a variety of detectors. GC separations enable the analysis of many different groups of organotin compounds (for example, butyl-, phenyl-, octyl-, and propyl) in a single analysis after derivatization [6]. However, derivatization is time-consuming and yields may vary between species and in terms of efficiency depending on matrix components. GC-ICP-MS has the potential to facilitate simultaneous multielemental speciation analysis, because species of Se [7], Pb [8], Hg [9], and Sn [10] have volatile forms and could be analyzed in a single analysis. Organotin separations by HPLC offer the advantage that derivatization is not required, which eliminates a potential source of uncertainty in the final result and can reduce analysis time significantly. However, the range of compounds that can be analyzed in a single run are limited compared to GC. The use of ICP-MS as a detector enables calibration by isotope dilution mass spectrometry as well as providing very low limits of detection (pg-ng range). In conjunction with isotopically labeled organotin species, this approach offers many advantages from an analytical point of view including reduced measurement uncertainties and greater precision compared to external calibration methods.
Experimental
Reagents and Standards Acetonitrile (UpSTM ultra-purity solvent grade) was obtained from Romil (Cambridge, UK). Glacial
acetic acid (TraceSelect) and anhydrous sodium acetate (Microselect 99.5% NT) were obtained from Fluka (Gillingham, Dorset, UK). Triethylamine, methanol and hexane were used as HPLC grade. Deionized water was obtained from a water purification unit at >18MW (Elga, Marlow, UK). Sodium tetra-ethylborate (NaBEt4) was obtained from Aldrich (Gillingham, Dorset, UK). Tributyltinchloride (TBTCl), Dibutyltinchloride (DBTCl2), Triphenyltinchloride (TPhTCl) and Diphenyltinchloride (DPhTCl2) were obtained from Aldrich and purified according to the procedure described by Sutton et al [11]. The 117Sn isotopically enriched TBTCl was synthesized according to the procedure described in the same paper. Monobutyltinchloride (MBTCl3) and Tetrabutyltinchloride (TeBTCl) were obtained from Aldrich, and Dioctyltin (DOT), Tripropyltin (TPrT), and Tetrapropyltin (TePrT) were obtained from Alfa Aesar (Johnson Matthey, Karlsruhe, Germany). Instrumentation Accelerated solvent extraction was carried out using a Dionex ASE 200 system. An Agilent 7500i ICP-MS was used for time-resolved analysis of 120 Sn, 118Sn, and 117Sn. The ShieldTorch system was used, and a second roughing pump was added in-line to increase sensitivity. An Agilent Technologies (Palo Alto, California, USA) 1100 HPLC system was used for HPLC separations. All stainless steel parts of the HPLC system that come into contact with the sample were replaced by polyether ether ketone (PEEK) components. A 100-cm length piece of PEEK tubing was used to connect the analytical column to the 100-L min-1
PFA MicroFlow nebulizer of the ICP-MS. Optimization of the ICP-MS conditions was achieved prior to HPLC analysis by adjusting the torch position and tuning for reduced oxide and doubly charged ion formation with a standard tuning solution containing 10 ng g-1 of 7Li, 89Y, 140Ce, and 205Tl in 2% HNO3. After this preliminary optimization, the HPLC system was coupled to the nebulizer and a final optimization was carried out using 103Rh added to the HPLC mobile phase. To reduce the solvent loading on the plasma, the double-pass spray-chamber was Peltier-cooled to -5 C. Oxygen (0.1 L min-1) was mixed into the make-up gas and added post-nebulization to convert organic carbon to CO2 in the plasma and avoid a carbon build-up on the cones. The final optimization was important because the nebulizer gas and make-up gas flows had to be adjusted to ensure plasma stability with the organic mobile phase conditions. HPLC separations were performed using a C-18 ACE column (3-m particle size, 2.1 mm 15 cm) with a mobile phase of 65: 23: 12: 0.05 % v/v/v/v acetonitrile/ water/ acetic acid/TEA. The flow rate was 0.2 mL min-1, and 20 mL of sample blends and mass-bias blends were injected. See Table 1. GC separations were performed on an Agilent 6890 GC. The Agilent G3158A GC interface [12] was used to couple the GC to the ICP-MS. The GC method was used as described by Rajendran et al [6]. The analytical column was connected to a length of deactivated fused silica, which was inserted along the ICP transfer line and injector. After installation of the interface, the torch position and the ion lenses were tuned using a 100-ppm xenon in oxygen mixture, which was added to the ICP-MS carrier gas at 5% volume via a T-piece. The isotope monitored for this adjustment was 131Xe.
Table 1.
ICP-MS Parameters Used HPLC-ICP-MS Platinum 14.5-14.9 L min-1 0.65-0.75 L min-1 0.15-0.25 L min 1350-1550 W 4-7 mm 300 ms
120 117 -1
Interface cones Plasma gas flow Carrier gas flow Make-up gas flow RF power Sampling depth Integration time per mass Isotopes monitored
GC-ICP-MS Platinum 14.5-14.9 L min-1 0.80-0.85 L min-1 Not used 1100-1200 W 6.5-7.5 mm 100 ms
120 118
Sn Sn 103 Rh ICP torch injector diameter: 1.5 mm Peltier cooled spray chamber at -5 C 5% O2 added post-nebulization ShieldTorch fitted
Sn Sn 117 Sn 5% N2 or O2 added to enhance sensitivity ShieldTorch fitted
Other parameters
Extraction of Organotin Compounds The ASE extraction cells were fitted with PTFE liners and filter papers and filled with dispersing agent. The sediment and the isotopically enriched spike were added and left to equilibrate overnight. Each cell was extracted using five 5-minute cycles at 100 C and 1500 psi after a 2-minute preheat and 5-minute heat cycle. 0.5 M sodium acetate/ 1.0 M acetic acid in methanol was used as the extraction solvent [13]. A calibrated solution (mass-bias blend) was prepared by adding the appropriate amounts of both 120Sn TBTCl and 117Sn TBTCl into an ASE cell filled and extracting under the same conditions as the samples. Digestion blanks were prepared by extracting ASE cells filled with hydromatrix and PTFE liners. After the extraction, each cell was flushed for 100 seconds with 60% of the volume and purged with N2. Prior to analysis, the extracts were diluted two- to fivefold in ultrapure water for HPLC-ICP-MS analysis. For GC-ICP-MS analysis, 5 mL of sample-, blank-, and mass-bias blend solutions were derivatized with 1 mL of 5% NaBEt4 and shaken for 10 minutes with 2 mL of hexane. An aliquot of the hexane fraction was then injected for analysis. Isotope Dilution Mass Spectrometry (IDMS) Methodology The method used for IDMS consisted of analyzing a blend of the sample together with a mass-bias calibration blend. Each sample blend was injected four times and bracketed by injections of the mass-bias calibration blend. The mass-bias calibration blend was prepared to match the concentration and isotope amount ratio in the sample by mixing the same amount of spike added to the sample with a primary standard of the analyte of interest [14], [15]. The estimation of the standard uncertainties for the measured isotope amount ratios was different to the one described in [14] as they were calculated as peak area ratios and not spectral measurement intensities. The chromatographic peaks were integrated manually using the RTE integrator of the Agilent ICP-MS chromatographic software. The mass fraction obtained from the measurement of each sample blend injection was then calculated according to: m m wX = wZ Y Zc mX mYc
RY RB RB RBc RBc
RBc RBc RZ
RBc RZ RY RBc
RB RBc RBc RZ RY wX wZ mY mX mZc mYc
Measured isotope amount ratio of sample blend (X+Y) Measured isotope amount ratio of calibration blend (Bc=Z+Y) Gravimetric value of the isotope amount ratio of calibration blend (Bc=Z+Y) Isotope amount ratio of Primary standard Z (IUPAC value) Isotope amount ratio of spike Y (value from certificate) Mass fraction of Sn in sample X obtained from the measurement of one aliquot Mass fraction of Sn in primary standard Z Mass of spike Y added to the sample X to prepare the blend B (=X+Y) Mass of sample X added to the spike Y to prepare the blend B (=X+Y) Mass of primary standard solution Z added to the spike Y to make calibration blend Bc (=Y+ Z) Mass of spike Y added to the spike Y primary standard solution Z to make calibration blend Bc (=Y+ Z)
The representative isotopic composition of Sn taken from IUPAC was used to calculate the isotope amount ratios of the primary standard. For the spike TBTCl, the isotopic composition was obtained from the certificate supplied with the 117 Sn enriched material from AEA Technology plc (UK). For the measured isotope amount ratio of the calibration blend (RBc), the average of the two ratios measured before and after each sample blend isotope amount ratio (RB) were taken. A mass fraction was calculated for each sample blend injection and the average of the bracketing mass-bias calibration blend injections. The average of the four mass fractions was then reported as the mass fraction obtained for the blend analyzed. The final mass fraction was recalculated back to the original sample and corrected for moisture content.

General Comparison Analysis of mixed organotin standard solutions showed that the GC method could separate a greater number (10-12) of compounds in a single run compared to HPLC-ICP-MS (5-6). The injection-to-injection time was ~40% shorter for HPLC-ICP-MS, due to the temperature profile used for GC separations. Because of the cost of the derivatizing agent, the reagent cost per sample is approximately double for GC sample preparation. Sensitivity Enhancement of GC-ICP-MS by Using Additional Gases Figure 1 and Table 2 illustrate the effect of adding different additional gases on the signal response
for a range of organotin compounds. Adding 5% O2 results in an increase in the measured peak area ranging from 9-fold (DBT and MPhT) to 12-fold (MBT). The addition of N2 results in a further increase compared to analysis without addition of an optional gas. Response factors range from 105 (DBT and TPhT) to 136 for MBT and 150 for TeBT. This translates to a reduction of the method detection limit (3s) for TBT from 0.4 ng mL-1 (no gas) to 0.03 ng mL-1 (with 5% O2 added) to 0.006 ng mL-1 (with 5% N2 added). The table below summarizes detection limits based on analysis of a calibration standard for MBT, DBT, and TBT.
Detection limits (ng mL-1 as Sn) by GC-ICP-MS MBT DBT TBT No gas added 0.7 0.5 0.4 5% N2 added 0.01 0.008 0.006
TeBT
650000 600000 550000 500000 450000 400000 350000 300000 250000 200000 150000 100000 50000 0 1.00 2.00 3.00 4.00 5.00 Time 6.00 7.00 8.00 9.00 10.00
TBT
DBT
DPhT
TPhT
MBT
a)
MPhT
b) c)
Figure 1.
Sensitivity increase on a 20 ng mL-1 mixed standard by using a) no additional gas, b) 5% O2, and c) 5% N2.
Table 2.
Effect of Different Additional Gases on Sensitivity of Organotin Compounds by GC-ICP-MS Retention time (min) 5.57 6.38 6.84 7.02 7.54 8.46 9.81 a) No gas added (peak area) 2274 3247 2026 3490 3717 3181 4287 b) 5% O2 added (peak area) 27029 29238 18173 33132 34225 29665 41119 Response factor compared to a) 12 9 9 10 9 9 10 c) 5% N2 added (peak area) 309702 340436 215182 399868 558916 338057 450803 Response factor compared to a) 136 105 106 115 150 106 105 Response factor compared to b) 12 12 12 12 16 11 11
Compound MBT DBT MPhT TBT TeBT DPhT TPhT
Comparison of HPLC-ICP-MS and GC-ICP-MS for Analysis of TBT in Sediment Table 3 shows the comparative data obtained by analysis of the same sediment extracts by both methodologies. There is no statistically significant difference between the two data sets. This confirms that the chromatographic separation and the different sample pretreatment (dilution/derivatization) used has no influence on the analytical result obtained. The chromatography for both methods appears in Figure 2 and Figure 3. The isotope amount ratio measurement precision, measured for 15 injections over a 6-8 hour period, is good for both methods (1.6% for HPLC-ICP-MS and 1.7% for GC-ICP-MS). The uncertainty estimates provided by HPLC-ICP-MS tend to be larger than for GC
separations. This is a result of broader peaks (50-60s by HPLC, compared to 4-6s by GC) and greater baseline noise. Detection limits for sediment analysis are estimated by peak height measurements (3s) as 3 pg TBT as Sn for HPLC-ICP-MS and 0.03 pg TBT as Sn for GC-ICP-MS with 5% O2 addition. This demonstrates the superior sensitivity of GC-ICP-MS even without sample preconcentration. The accuracy of the analytical procedure was evaluated by measuring extractions of the certified reference sediment PACS-2 (NRC, Canada). The mean mass fraction obtained by the HPLC-ICP-MS analysis of four extracts was 864 35 ng g-1 TBT as Sn compared to a certified value of 980 130 ng g-1 TBT as Sn.
Table 3.
TBT Data for Sediment Extracts HPLC-ICP-MS (ng/g as Sn) n=4 827 805 845 826 87 Standard uncertainty k = 1 (ng/g as Sn) 19 38 9 22 GC-ICP-MS (ng/g as Sn) n=4 853 846 838 846 39 Standard uncertainty k = 1 (ng/g as Sn) 12 13 8 11
Sample 1 2 3 Mean Expanded uncertainty (k = 2)
24000 22000 20000 18000 16000 14000 12000 10000 8000 6000 4000 2000
0 1.00
Ion 120.00 (119.70 to 120.70): msd1.ms
DBT
Ion 117.00 (116.70 to 117.70): msd1.ms
TBT
Abundance
TPhT
2.00
3.00
4.00
5.00
6.00
Time
7.00
8.00
9.00
10.00
11.00
Figure 2.
HPLC-ICP-MS chromatogram.
80000 70000 60000 Abundance 50000 40000 30000 20000 10000 0 1.00 2.00 3.00 4.00 5.00 Time 6.00
Ion 120.00 (119.70 to 120.70): msd1.ms
EtBu3Sn (TBT) Et3Bu2Sn (DBT)
Et3BuSn (MBT)
EtBu3Sn (TPhT)
7.00 8.00 9.00 10.00
80000 70000 60000 Abundance 50000 40000 30000 20000 10000 0 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00
Ion 117.00 (116.70 to 117.70): msd1.ms
117Sn
enriched TBT
spike
Figure 3.
GC-ICP-MS chromatogram.
Conclusions
Both HPLC-ICP-MS and GC-ICP-MS offer advantages for organotin speciation analysis. While there is no statistical difference in the results obtained, HPLC-ICP-MS can be used for cheaper and faster determinations of large sample batches, while the superior sensitivity and the greater number of analytes separated make GC-ICP-MS an ideal tool for monitoring studies at the ultratrace level.
9.
J. P. Snell, I. I. Stewart, R. E. Sturgeon, and W. J. Frech, (2000) J. Anal. At. Spectrom., 15 (12): 15401545.
10. J. R. Encinar, P. R. Gonzalez, J. I. Garcia Alonso, and A. Sanz-Medel, (2002) Anal. Chem., 74, 270281. 11. P. G. Sutton, C. F. Harrington, B. Fairman, E. H. Evans, L. Ebdon, and T. Catterick, (2000) Applied Organometallic Chemistry 14, 110. 12. Agilent Technical Note GC-ICP-MS Interface publication 5988-3071EN. 13. C. G. Arnold, M. Berg, S. R. Mller, U. Dommann, and R. P. Schwarzenbach, (1998) Anal. Chem., 70, 30943101. 14. T. Catterick, B. Fairman, and C. J. Harrington, (1998) J. Anal. At. Spectrom. 13, 1109. 15. Guidelines for achieving high accuracy in isotope dilution mass spectrometry, edited by M. Sargent, C. Harrington, and T. Harte RSC London, 2002.
References
1. 2. S. Nicklin and M. W. Robson, (1988) Applied Organometallic Chemistry, 2, 487508. H. Tao, R. B. Rajendran, C. R. Quetel, T. Nakazato, M. Tominaga, and A. Miyazaki, (1999) Anal. Chem., 71, 42084215. J. C. Keithly, R. D. Cardwell, and D. G. Henderson, (1999) Hum. Ecol. Risk. Assess., 5, No. 2, 337354. A. Sadiki, and D. T. Williams, (1996) Chemosphere, 32, 12, 23892398. S. Takahashi, H. Mukai, S. Tanabe, K. Sakayama, T. Miyazaki, and H. Masuno, (1999) Environmental Pollution, 106, 213218. R. B. Rajendran, H. Tao, T. Nakazato, and A. Miyazaki, (2000) Analyst, 125, 17571763. J. L. Gomez-Ariza, J. A. Pozas, I. Giraldez, and E. J. Morales, (1998) J. Chromatogr. A., 823 (12): 259277. I. A. Leal-Granadillo, J. I. Garcia-Alonso, and A. Sanz-Medel, (2000) Anal-Chim-Acta., 423 (1): 2129.
3.
4. 5.
Acknowledgments
The work described in this application note was supported under contract with the Department of Trade and Industry (UK) as part of the National Measurement System Valid Analytical Measurement (VAM) program.
6. 7.

8.
Discover the Full Capabilities of ICP-MS in Food Safety
Extending the Capabilities of the Laboratory Measuring Traces in Milk Powder
Reaching New Limits
Arsenite, Glycerol-ribose Sulfonate -ribose Phosphate -ribose Arsenate Sulfate -ribose
http://www.agilent.com/chem
Measurement of Macro and Trace Elements in Plant Digests Using the 7500c ICP-MS System Application
Food
Author
Kazuo Yamanaka Agilent Technologies Nakacho 1-15-5 Musashino-shi Tokyo 180-8543 Japan Fred Fryer Agilent Technologies 12/2 Eden Park Drive, North Ryde NSW 2113 Australia
grains such as wheat and rice is resulting in a nutritionally poor diet. Micronutrient [1] malnutrition is an identified problem that has coincided with the rapid adoption of modern cereal cropping systems. Profitable and sustainable agriculture depends on the understanding of the nutrients required and available for plant growth, as well as the nutrients for a balanced human diet. World food production will need to double over the next 30 years to keep pace with increasing demands from both industrialized and rapidly developing countries. As well as the need to increase production, there will be an increase in demand for higher quality and healthier food products as developing countries become more affluent. Taken from the Commonwealth Scientific and Industrial Research Organisation (CSIRO) website: http://www.csiro.gov.au (select: Agribusiness/Field Crops/Field Crops & Australia) Human dietary micronutrients are required by humans in very small amounts. They include at least 14 trace elements (As, B, Cr, Cu, F, I, Fe, Mn, Mo, Ni, Se, Si, V, Zn) as well as 13 vitamins (thiamin, riboflavin, niacin, pantothenic acid, biotin, folic acid, vitamins B6, B12, C, A, D, E, K) The recommended daily intake of the micronutrient trace elements is of the order of: mg per day for B, Cu, F, Fe, Mn, Zn g per day for As, Cr, I, Mo, Ni, Se, Si, V
Abstract
Inductively coupled plasma mass spectrometry is a powerful tool for the investigation of many materials. The Agilent 7500c with Octopole Reaction System was used to analyze major, minor and trace elements in two standard reference plant materials. The data obtained using the 7500c is compared to the certificate reference values and to results that were generated using inductively coupled plasma optical emission spectroscopy. Results for all elements obtained using the 7500c agree with the certified values.
Introduction
The reliable measurement of trace elements in food is becoming more important as information is revealing that over-dependence on processed
Accurate determination of these trace elements in food materials is useful in ensuring that dietary intake is providing adequate levels of micronutrient elements. Due to the very low concentrations that must be measured and, in many cases, the high and variable sample matrix in which the measurements must be made, this analysis has proved challenging for elemental analysis instrumentation. Traditionally, a combination of techniques was required for a complete analysis of the plant digesttypically Graphite Furnace Atomic Absorption Spectroscopy (GFAAS), Hydride-Atomic Absorption Spectroscopy (HG-AAS) and Inductively Coupled Plasma Optical Emission Spectroscopy (ICP-OES). Such is the performance and elemental coverage of modern inductively coupled plasma mass spectrometry (ICP-MS) instrumentation, in many cases (metals analysis in drinking water, for example) a single ICP-MS has replaced all of the above mentioned techniques, enabling all analytes to be determined in a single measurement. The analysis of plant and food digests for nutritional studies is more challenging. In ICP-MS, isobaric interferences arise from the argon used to sustain the plasma and from the reagents used for sample preparation. Table 1 summarizes some well-known interfering species. In biological sample analysis, there are well-documented interferences for ICP-MS that can bias the measurement of Fe, Cr, V, As and Se at trace levels, with the result that ICP-MS has not yet been widely adopted by the foods industry.
Table 1. Examples of Potential Interferences in Biological/Clinical Matrices Element Cr V Fe Cu As Se Mass 52; 53 51 56 63 75 77; 78; 80 Molecular interference 40 Ar12C, 36Ar16O, 35Cl16O1H; 37Cl16O 35 16 Cl O 40 Ar16O 40 Ar23Na 40 Ar35Cl 40 Ar37Cl; 40Ar38Ar; 40Ar40Ar;
digestion with hydrogen peroxide and nitric acid. This digestion media does not generate additional interferences for ICP-MS and is a complete digest. However, for high sample numbers, the traditional hot plate digest offers higher sample throughput than closed vessel microwave digestion [2]. Recently, the advent of collision/reaction cells has improved the detection capability of quadrupole ICP-MS (ICP-QMS) by removing spectral interferences on analytes such as Fe, Cr, V, As and Se. The Agilent 7500c ICP-MS features an Octopole Reaction System (ORS) for highly efficient removal of multiple interferences arising from complex sample matrices. The ORS removes interferences by either reacting a gas with the interference or by preventing the interfering species from entering the analyzer stage using a process called energy discrimination. The 7500c exhibits highly efficient interference removal. The Ar2 overlap on Se at mass 80 is virtually eliminated, reducing the background equivalent concentration from 100s of ppb to <10 ppt. Moreover, the 7500c was designed specifically to handle complex matrices such as plant and food digests. The key to the successful multi-element determination of trace elements in complex samples is a combination of matrix tolerance and efficient interference removal. Matrix tolerance is mainly determined by the plasma efficiency, which must be optimized to ensure efficient sample decomposition, and is monitored by the CeO/Ce ratio. An efficient plasma minimizes the formation of plasmaand matrix-based interferences, while maximizing the conversion of analyte atoms into ions. The importance of matrix tolerance of any ICP-MS system should not be underestimated, as this leads to improved analytical accuracy, better tolerance to matrix changes and reduced requirements to carry out routine maintenance of the vacuum, ion lens and pump components. All of these aspects contribute to the usability of the analytical instrument, as routine maintenance contributes far more to the down-time of a modern, reliable ICP-MS instrument than hardware breakdowns. The unique capability of the Agilent 7500 Series lies in the mode of operation of the plasma source, which decomposes sample matrices five to 10 times more efficiently than is typical for other ICP-MS instruments.
One obvious way to remove interferences is to eliminate the source of the interfering species. Traditionally plant materials are digested on a hot plate using a mixture of nitric and perchloric acids. Chloride-based mass spectral interferences are introduced by this method. An alternative sample preparation method is available using microwave
The 7500c was designed specifically to handle complex, high matrix samples. A robust 27.12-MHz plasma, low sample uptake rate, cooled spray chamber and proven small orifice interface protect the ORS from contamination by undissociated sample matrix. A novel ion optic, mounted outside the high vacuum region for easy access, further protects the reaction cell, which features an octopole for optimum ion transmission. The octopole is mounted off-axis to minimize random background levels. A schematic of the 7500c is shown in Figure 1. Some of the important instrument parameters that contribute to good matrix decomposition are: The standard low sample flow rate (100 to 400 L/min) and Peltier-cooled spray chamber reduce the sample and water vapor loading on the plasma, which leads to a hotter plasma central channel. The 7500 Series uses a high efficiency, solid state 27.12-MHz plasma RF generator, ensuring good energy transfer into the plasma central channel. The unique wide internal diameter plasma torch design ensures that the sample aerosol is resident in the plasma for sufficient time to ensure complete matrix decomposition, leading to exceptionally good matrix decomposition (low CeO/Ce ratio).
The optimized interface design, which uses the smallest skimmer cone orifice of any commercial ICP-MS instrument, ensures that minimal sample matrix is passed into the high-vacuum part of the instrument, dramatically reducing the requirement for routine maintenance of the interface cones, the ion lenses and the collision cell. In summary, as the complexity of the sample matrix increases, the benefit of minimized interference levels becomes more significant. Because modern analytical laboratories rarely have the luxury of pre-analyzing samples to identify the matrix, it is impractical to rely on matrix matching of the samples or data correction using complicated interference equations.
Sample Preparation and Analysis

About 800 mg of sample was accurately weighed and carefully heated with 10 mL nitric acid (70%), followed by gentle heating with the addition of 8 mL perchloric acid (70%) until colorless. After cooling, 30 mL water was added and heating resumed for 10 min. Finally, the solutions were cooled, then made to 100 mL volume with water.
Reaction gas inlet Quadrupole Octopole reaction cell
Octopole Off-axis lens
Figure 1: Schematic diagram of the Agilent 7500c Octopole Reaction System.
The instrument was tuned and optimized as detailed in Table 2. Calibrations were performed using external standards prepared from 1000 ppm single element stock, made up as appropriate with 2% nitric acid.
Table 2. Agilent 7500c Operating Conditions Plasma RF power Sample depth Carrier gas flow Spray chamber temperature Sample flow rate Nebulizer Interface 1500 W 9.5 mm from load coil 1.1 L/min 2 C 240 L/min Agilent microflow (PFA) Nickel sample and skimmer cones
Table 4. NIST 1573a (Tomato Leaves, Blank Corrected)

Name 43 Ca 39 K 52 Cr 53 Cr 54 Fe 56 Fe 63 Cu 65 Cu 75 As 78 Se 111 Cd Certified (mg/kg) 5.05% 2.70% 1.99 1.99 368 368 4.7 4.7 0.112 0.054 1.52 ICPOES (mg/kg) 5.00% 2.72% 1.7, 1.8 1.7, 1.8 342, 347 342, 347 2.49, 2.40 2.49, 2.40 5.7, 6.6 0.1, 0.8 5.5, 5.9 7500c (mg/kg) 5.08% 2.62% 1.60 1.63 368 368 4.43 4.47 0.175 0.061 1.32
Table 5. NIST 1570a (Spinach, Blank Corrected)
The external calibrations were run in the same analytical sequence as the samples. Sample concentration was calculated using the internal standard method. Table 3 summarizes the element and relevant internal standard information.
Name 39 K 43 Ca 52 Cr 53 Cr 54 Fe 56 Fe 63 Cu 65 Cu 75 As 78 Se 111 Cd 54 Fe 56 Fe 63 Cu 65 Cu 75 As 78 Se 111 Cd
Certified (mg/kg) 2.90% 1.53% 12.20 12.20 0.07 0.12 2.89 12.20 12.20 0.07 0.12 2.89
Table 3. Reaction Gases and Internal Standards Used Measured element Potassium Calcium Chromium Iron Copper Zinc Arsenic Selenium Cadmium Reaction gas Helium Helium Helium Helium Helium Helium Helium Hydrogen Hydrogen Internal standard Scandium Scandium Gallium Gallium Cobalt Cobalt Yttrium Indium (115) Indium (115)

The practical effect of the 7500cs unique combination of matrix tolerance and interference removal is that complex and variable samples can be measured with a simple quantification procedure using external standard calibration and internal standard correction for all masses. As and Se were accurately quantified at sub-ppb levels, even in a matrix containing 8% perchloric acid. Tables 4 and 5 summarize the results obtained in a blind analysis of plant digests using the 7500c, comparing the results with both the certified values and data obtained from analysis by ICP-OES.
Reference 2: (mg/kg) 2.63% 1.32% 252 252 11.6 11.6 252 252 11.6 11.6 -
7500c (mg/kg) 2.56% 1.39% 1.24 1.29 248 250 10.48 10.51 0.062 0.09 2.33 248 250 10.48 10.51 0.062 0.09 2.33
Measurements of Cr, Fe and Cu were made on two separate isotopes for each element. Because molecular interferences will, in many cases, only affect one of the analyte isotopes, the presence of an interference can cause a large discrepancy between results for different isotopes of the same element. An example of this is the measurement of Cu in a high Na matrix, where 40Ar23 Na gives an overlap on 63Cu, but no interference on 65Cu. As the results indicate, the 7500c obtained excellent agreement for all the pairs of isotopes, highlighting the capabilities of the ORS in reducing interfering molecular species that, until now, have prevented the accurate trace analysis of transition metals in complex matrices by ICP-QMS.
Values for major and trace element concentrations agreed both with the expected value and the results obtained from ICP-OES. In the cases where the trace values for some elements were below the detection limit of the ICP-OES, the 7500c returned results in excellent agreement with the certified value. This data illustrates the wide dynamic range of the system and demonstrates its advantages as a replacement for traditional techniques such as ICP-OES. The quantitative analysis of the NIST SRM samples also demonstrates that both the 7500c and the operating conditions are robust and tolerant of the changing matrix composition found in plant digests.
Conclusions
The trace analysis of plant digests is an application that can be suitably addressed by the 7500c. Advances in technology now allow the determination of multiple elements in complex sample matrices, with efficient interference removal and, in the case of the 7500c, with the excellent matrix tolerance for which the 7500 Series is renowned. Accurate quantification of As and Se at low and even sub-ppb levels in plant digests is possible, even where high concentrations of perchloric acid have been added during the sample preparation stage.
Acknowledgement
The ICP-OES measurements and NIST sample preparation were performed at the University of Queensland, School of Land and Food Sciences, Australia.
References
1. Ross M. Welch; USDA-ARS, US plant and soil and nutrition laboratory, Towner Road, Ithaca, NY 14853, USA, Micronutrients, Agriculture and Nutrition: Linkages for Improved Health and Wellbeing 2. Da-Hai Sun; Waters, J. K.; Mawhinney, T. P. Determination of thirteen common elements in food samples by inductively coupled plasma atomic emission spectrometry: comparison of five digestion methods.; Journal of AOAC International 2000, 83 (5) 1218-1224

5
Natural Compounds & Additives
Additives
The Separation of Seven Synthetic/Artificial Food Colors on Agilent HC(2)/TC(2) Reversed Phase Columns Application
Food and Flavors
Author
Rongjie Fu Agilent Technologies, Inc. 412 Ying Lun Road Pu Dong, Shanghai 200131 China
Abstract
Many synthetic or artificial colorants (Red #33, Sunset yellow, etc.) are used in food and beverages to improve product appearance. These compounds can be easily separated by reversed-phase liquid chromatography. [1] A new C18 column was used to separate seven food colorants, using a gradient method with a phosphate buffer-acetonitrile mobile phase. This method is suitable for many samples and is applied here to the analysis of these colorants in beverages.
possible harmful effects of these compounds, including allergic reactions and hyperactivity in children. Artificial food colorings are restricted in some countries because of these possible effects. The quantity of these compounds in food quality control is becoming more important, as is the need to prove that they meet international food quality control standards. The structures of seven colorants, mostly azo dyes, are shown in Figure 1. These compounds are used in beverages to give them a colorful, attractive appearance. In this application, we focused on developing a method for separating seven colorants on the Agilent TC-C18(2) column and then applying this method to the analysis of these food colors in beverage samples. This new column is packed with 5-m high-purity (B type) silica with a surface area of 290 m2/g and bonded with C18, achieving a carbon load of 12%. This is a moderate carbon load for a column with this surface area and is therefore more hydrophilic than columns with a higher carbon load. The special treatment of the surface of the silica and improved end-capping give this column excellent performance, especially for the separation of polar and basic compounds.
Introduction
Synthetic or artificial food colorings are almost all water soluble, making them ideal for analysis by HPLC with reversed phase columns. These compounds are generally safe, but there are some
O O OH NaO S O O N N N N NaO O S ONa O NaO S O N N HO
ONa S O
1. Tartrazine
2. Amaranth
O ONa
HO O O NaO S O N N OH O O S ONa S O ONa N S O O S NaO O O N
NaO
3. Ponceau 4R
4. Sunset yellow
O OH N N O NaO S O O S S S ONa NH2 NaO OCH3 O S O H3C O NaO I O I O N N I HO ONa I
5. Red 33
6. Allura red
ONa
7. Erythrosine
Figure 1.
Structures of colorants.
HPLC conditions
Instrument Column Mobile phase Gradient Flow rate Temperature Injection volume Agilent 1200SL with DAD Agilent TC-C18(2), 4.6 mm 150 mm, 5 m A: 20 mM phosphate buffer, pH 7.0 B: Methanol 015 min, 10% B 90% B 1 mL/min 30 C 5 L
Detector wavelength 254 nm
Seven Food Color Standards Separated on Agilent TC-C18(2) The compounds shown in Figure 1 are polar, water-soluble compounds with sulfonic acid groups on them, increasing their solubility in
water. In fact, many of these compounds are used in the salt form in food. A C18 column was selected to separate these compounds. Two new C18 columns are available from Agilent for routine or QC/QA-type applications. These columns are the Agilent HC-C18(2) column with a carbon load of 17% and the TC-C18(2) column with a carbon load of 12%. The difference in carbon load was designed to provide column choices optimized for the analyses of very polar and very nonpolar analytes. Polar analytes typically require high aqueous mobile phases to be effectively retained. For the artificial food colors, a gradient method starting at low organic in the mobile phase was developed. Starting at very high aqueous mobile phases, the TC-C18(2) column with a carbon load of 12% was the better column for this method development, and can be used for mixtures of polar compounds or for samples of polar and nonpolar compounds.
Figure 2 shows all seven colorants separated on the 4.6 mm 150 mm, 5 m Agilent TC-C18(2) column. Using pH 7.0 phosphate buffer-methanol mobile phase, all the compounds are symmetrical with a U.S. Pharmacopeia (USP) tailing factor close to 1.0 and baseline resolution with good reproducibility. Because of their strong acidity, the peak shape and resolution of the compounds are dramatically influenced by the pH selected for the mobile phase. The separation was repeated with mobile phases at three different pH values. The
chromatograms for this are shown in Figure 3. The chromatograms clearly show that the peak shape and resolution are influenced by pH; especially the first three components, which are the most acidic and polar. At low pH, pH 2.2, poor resolution was seen between peaks 1 and 2 and peaks 4 and 5; however, pH had less influence on resolution for the later eluting compounds, such as erythrosine (peak 7). Erythrosine is the most hydrophobic of these compounds. The overall elution order of these compounds was tied to their hydrophobic properties, as expected in reversed phase LC.
Norm. 140 120 100 80 60 40 20 0 2 4
2 4 3 5 6 7 1 2 3 4 5 6 7 Tartrazine Amaranth Ponceau 4R Sunset yellow Red 33 Allura red Erythrosine
10
12
14
16
min
Figure 2.
mAU 120 100 80 60 40 20 0
Chromatogram of colorants standards on the Agilent TC-C18(2) column.
Tf=1.24 Rs4,5=1.3 Tf=1.21 Tf=1.98
Rs4,5=1.7
pH2.2
2 mAU 140 120 100 80 60 40 20 0 2 mAU 160 140 120 100 80 60 40 20 0 2
10
12
14
16
min
Tf=1.11
Tf=1.14 Tf=1.14
Rs4,5=1.7
pH4.5
10
12
14
16
min
Tf=1.04
Tf=1.05 Tf=1.03
Rs4,5=2.3
pH7.0
10
12
14
16
min
Figure 3.
Chromatogram of colorants standards in different pH mobile phase on the Agilent TC-C18(2) column. 3
Lot-to-Lot Reproducibility Lot-to-lot reproducibility of columns is important for rugged HPLC methods and routine monitoring. Figure 4 shows the reproducibility test used to evaluate the TC-C18(2) column in this method. Three different batches of columns were used in the test under the same HPLC conditions. The chromatograms show almost no retention time drift and no peak shape variations occurring from column lot-to-lot, which demonstrates the highquality manufacturing and packing techniques used for these columns. Analysis of Artificial Food Colors in Beverages Synthetic or artificial food colorants exist in many common beverages and food, such as fruit-flavored
drinks and sodas, cakes, and candy. Some of the colorants in two different beverages were separated in this application. Two popular beverages were selected, an orange flavored soda and a drink powder. For the sample preparation, orange flavored soda was degassed, filtered through a 0.45-m filter, and then analyzed. The drink powder was dissolved in water, filtered through a 0.45-m filter, and injected for analysis. The chromatograms from these samples are shown in Figure 5. Three colorants were found and completely separated in the orange flavored soda sample. Based on the standard, the first known peak could be resolved from the unknown peak before it. Two target colorants were found and adequately resolved in the beverage powder sample for effective quantitative analysis.
mAU 250
200
150
Lot 1
100
Lot 2
50
Lot 3
0 0 2 4 6 8 10 12 14 16 min
Figure 4.
Lot-to-lot reproducibility(TC-C18[2], 4.6 mm 150 mm, 5 m).
mAU 250
200
150
Standards
100
1 3
50
Orange flavored soda
1 4
Orange beverage powder

10 12 14 16 min
0 0 2 4 6 8
Figure 5.
Chromatogram of two samples.
In the beverage powder, there were lower levels of these food colors added. For the low-concentration samples, a large-volume injection is needed to meet the detection and quantitation requirements. Two chromatograms of large-volume 50-L injections are shown in Figure 6. The separation was still
mAU 350 300 250 200 150 100 50 0 2 mAU 1000 800 600 400 200 0 2 4 6 4 6
quite good, with symmetrical peaks and little band broadening. For the sample orange flavored soda, peak 1 was completely separated from the unknown compound. All these demonstrate the high capacity of the high surface area TC-C18(2) column.
Separated with unknown compound
Orange flavored soda 50-L injection
10
12
14
min
Orange beverage powder 50-L injection
10
12
14
min
Figure 6.
Chromatogram of 50-L injection.
Selectivity Comparison with HC-C18(2) As mentioned above, the carbon load of HC-C18(2) is higher than TC-C18(2), which leads to strong retention of nonpolar compounds but not quite so much retention of polar compounds. Figure 7 shows that the HC-C18(2) column retains the more polar food colors slightly less than the TC-C18(2) column does. The change in carbon load also causes a change in selectivity of peaks 4 and 5 between the two columns, which is also shown in the chromatograms. Figure 8 shows chromatograms of orange flavored soda both on TC-C18(2) and HC-C18(2) columns. Peak 1 and the potential interfering compound on the TC-C18(2) column
reverse order on the HC-C18(2) column. The possible interfering compound elutes before the target compound 1 on the TC-C18(2) column and has excellent resolution. When the large-volume injection was made, baseline separation could still be achieved; when the elution order was reversed eluted, the two compounds were not baseline separated until the mobile phase composition was adjusted. Because the mobile phase is a low organic (only 10%) starting gradient for this method, we chose the TC-C18(2) column with a lower carbon load for this application. Its important to note that the different selectivity of these two columns provides more column selection opportunities for method development.
Norm. 140 120 100 80 60 40 20 0 2 Norm. 160 140 120 100 80 60 40 20 0 2 4 4
2 4 3 5 6 7
TC-C18(2)
10
12
14
16
min
HC-C18(2)
10
12
14
16
min
Figure 7.
Chromatogram of colorants standards on TC-C18(2) and HC-C18(2).
mAU
1
80
Interferential compound
60
HC-C18(2)
40
20
TC-C18(2)
0 2 4 6 8 10 12 14 16 min
Figure 8.
Different selectivity of TC-C18(2) and HC-C18(2).
Conclusions
Using a simple gradient method, many common synthetic colorants can be separated on the Agilent TC-C18(2) column. This method allows baseline separation for many artificial food colors and provides symmetrical peaks and good reproducibility. The Agilent TC-C18(2) column can be used for daily QC analysis of synthetic colorants in foods and beverages with reliable results.
Reference
1. Fang Yanyan, Agilent Technologies, publication 5989-3639CHCN. www.agilent.com/chem

A Rapid and Sensitive Analysis Method for Sudan Reds in Curry and Chili Powder Using LC/MS/MS Application
Food Safety
Authors
Yanyan Fang Agilent Technologies, Inc. 412 Ying Lun Road Pu Dong, Shanghai 200131 China Jeffry Keever Agilent Technologies, Inc. 200 Regency Forest Drive, Suite 330 Cary, NC 27518-8695 USA Dustin Yaworsky and Jerry Zweigenbaum Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA
Introduction
The group of color additives known as Sudan dyes consists of a number of red colorants, for example, Sudan I through IV. This group, together with other dyes, such as Para Red, are synthetically produced azo dyes. Because their degradation products are considered to be carcinogens and teratogens, the EU and the U.S. do not permit the use of these colorants as food additives. However, in some countries, these dyes are still occasionally used to intensify the color of bell pepper and chili powders. The red dyes Sudan I, II, III, and IV are oil-soluble azo dyes used legally in the leather and fabric industries. The International Agency for Research on Cancer (IARC), a part of the World Health Organization, has assessed the Sudan dyes as Group 3 genotoxic carcinogens. [1] The industrial dye Para Red is chemically similar to Sudan I and is also a dye not permitted for use in food. The EU issued Decision 2003/460/EC requiring as a condition of import that all hot chili and hot chili products be tested for Sudan I. [2] The Decision was amended in January of 2004 (2004/92/EC) to include Sudan II, III, and IV. [3] This requirement remains in effect. The highly selective and sensitive triple quadrupole (QQQ) is used to meet the testing requirement.
Abstract
A method was developed using the Agilent G6410 Triple Quadrupole LC/MS for Sudan Reds in different matrices, including curry powder, red pepper powder, and chili sources. The analytical performance of the method was evaluated for four different matrices and the results show little or no matrix effects. Linearity of response over three orders of magnitude was demonstrated (r > 0.99). In addition, good reproducibility of the two required product ion ratios was obtained to meet the EU identification points needed for confirmation.
Experimental
Reagents and Materials Ethyl acetate from Burdick and Jackson ( Morristown, NJ) Methanol HPLC-grade from Burdick and Jackson Water at 18 M from a Milli-Q Synthesis System by Millipore (Billerica, MA) Syringe filter (0.2 m, PTFE) from Agilent, p/n 5185-5843 Samples 1. Curry powder (House of Spices [India] Inc.) 2. Red pepper powder (Korean Farm, Inc.) 3. Red pepper powder (Oriental Mascot, Summit Import Corp.) 4. Worcestershire sauce (Lea & Perrins) Overview of Method Standard Preparation 1. Standard solution preparation: 10 mg Para Red, Sudan Red I, Sudan Red II, Sudan Red III, and Sudan Red IV were dissolved into 100.0 mL acetontrile/water (90:10) to a final concentration of 100 ppm. 2. Diluents solution preparation: using acetontrile to obtain concentrations at 1 pmm, 0.1 ppm, 10 ppb, 5 ppb, 2 ppb, and 100 ppt. Sample Preparation 1. Curry powder was spiked with 10 L of standard solution into 1 g of powder, diluted with 5 mL ethyl acetate, and allowed to ultrasonicate for 10 min. The solution was filtered with a 0.22-m syringe filter and evaporated to dryness under a nitrogen stream at 40 C (used below). The residue was reconstituted in 1 mL acetonitrile. This was filtered again before injection. No additional clean up of the sample solution was performed. This gave a spiked sample at a concentration at of 1 ppm. This spiking procedure was repeated to obtain spiked samples at levels of 100, 10, 1, and 0.1 ppb.
2. The two red pepper chili powders: The procedure used to prepare the curry powder was also used to prepare the spiked samples. 3. Worcestershire sauce was spiked with 10 L of standard into 1 mL of sauce. The sauce, at about pH 3, was diluted with 5 mL ethyl acetate and allowed to ultrasonicate for 10 min. The solution was centrifuged for 2 min (8,000 rpm) and the top solvent layer (ethyl acetate) was transferred to a clean tube and evaporated under a nitrogen stream at 40 C. The residue was redissolved in 1 mL acetonitrile and put in an ultrasonic bath for 1 min. This solution was filtered before injection. LC/MS/MS Conditions All analyses were performed with the Agilent 6410 Triple Quadrupole LC/MS equipped with an electrospray ionization source operated in positive mode. The HPLC was the Agilent 1200 LC system equipped with binary pump and well plate autosampler. See Table 1 for the conditions.
Table 1. HPLC Column Flow rate
LC/MS Conditions ZORBAX XDB C18 2.1 mm 100 mm, 1.8 m Agilent p/n: 928700-906 0.4 mL/min A: 0.1% formic acid in water B: 0.1% formic acid in acetonitrile 0.14 min, 70~98% B 45 min 98% B 57 min 70% 10 min including re-equilibration 40 C 5 L
Mobile phases Gradient
Total run time Temperature Injection
MS Source Settings Source ESI Ion polarity Drying gas temp. Drying gas flow rate Nebulizer Vcap MRM parameters Positive 350 C 10 L/min 45 psi 3500 V As shown in Table 2
The MS analysis was divided into five segments each containing one of the analytes. The appropriate fragmentor and collision energy for the analyte eluting in that segment was contained therein. Segment 1 started at 0 min, segment 2 at 2.0 min, segment 3 at 2.8 min, segment 4 at 3.6 min, and segment 5 at 4.2 min.

Each of the dyes was analyzed by LC/MS/MS in product ion scan and examined for spectral quality. Appropriate transition ions for quantitative analysis and confirmation were selected. The structure of each of the compounds is shown below. The azo bond cleaves under the collisioninduced dissociation conditions and produces unique fragment ions.
Optimization of MS Condition
Optimization usually consists of simply finding the fragmentor voltage that maximizes the abundance
of the precursor ion and the collision energy that maximizes the abundance of the transition ion. Optimization of the fragmentor was done by stepping through the voltages and recording intensity as shown for Sudan Red I in Figure 1. The other compounds were optimized by the same process. Figure 2 shows the product ion MS/MS spectra for Sudan Red I at two collision energies (CE). It can be seen from the spectra that the transition m/z 156 gives a maximum intensity at 15 V and m/z 93 at 25 V. Typically the most abundant ion is used for quantitation to maximize precision and accuracy at the lowest levels and the less abundant ion is used as the qualifier for confirmation. However, for the highest selectivity, unique transition ions are sought. In the example of Sudan Red I the transition m/z 93 is the aniline radical cation produced by cleavage of the azo bond and hydrogen transposition. [4] This also produces the transition m/z 156, both unique. Measuring CE vs intensity for the two most abundant ions of all the compounds resulted in the voltages shown in Table 2 along with the other MS settings.
Dye Structures Para red

OH
N
Sudan Red I
HO
N N
NO2
Sudan Red II
H3C CH3 HO N N
Sudan Red III

N N N N
HO
Sudan Red IV
CH3 N N
CH3 HO N N
250000
Sudan Red 1 fragmentor Sudan Red 1 area
200000
150000
100000
50000
0 80 100 120 140 160 180 200
Figure 1.
Optimization of fragmentor voltage for Sudan Red I.
+ Product Ion (2.165 min) (249.0 **) 15v_10.d 10 3 9 8 7 6 N N HO 5 4 3 2 1 0 65.2 50 60 70 80.1 80 90 101.1 128.0 168.1 179.2 191.3 204.2 221.1 249.1 156.0 93.1
CE = 15V
232.1
m/z = 156
100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 Abundance vs. Mass-to-Charge (m/z)
10 4 1.1 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
+ Product Ion (2.165 min) (249.0 **) 25v_01.d 93.0
CE = 25V m/z = 93
128.0
156.0 66.2 76.9 50 60 70 80 90 101.1 115.0 Abundance vs. Mass-to-Charge (m/z) 204.0 220.7
232.1 248.2
100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250
Figure 2.
MS/MS spectra of Sudan Red I at collision energies (CE) 15 and 25 V.
Table 2. Analyte Para Red
MRM MS/MS Parameters Transition 294156 294128 249156 24993 277156 277121 35377 353197 38191 381225 Dwell (msec) 200 200 200 200 200 200 200 200 200 200 Fragmentor voltage 120 V 120 V 120 V 120 V 100 V 100 V 120 V 120 V 120 V 120 V Collision energy 15 V 30 V 15 V 25 V 10 V 20 V 25 V 20 V 30 V 15 V MS2 Res. Unit Unit Unit Unit Unit Unit Unit Unit Unit Unit Gain 1 1 1 1 1 1 1 1 1 1
Sudan Red I Sudan Red II Sudan Red III Sudan Red IV
Chromatography and Sensitivity

Figure 3 shows the chromatographic separation achieved for the five analytes allowing the segmentation of the MS/MS conditions. The sensitivity of
the methodology is shown with the spike of the chili powder at 1 ppb. This is shown with a blank chili powder (Figure 4) and spike where each transition of the quantifier and qualifier ion for all five analytes is displayed (Figure 5).
10 3 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0
+ BPC MRM (** **) WorklistData3-r002.d
12
23
3 34
45
2 4 1 5
1. 2. 3. 4. 5.
Para Red Sudan Red I Sudan Red II Sudan Red III Sudan Red IV
0.0 0.4 0.8
1.2
1.6
2.4 2.8 3.2 3.6 4 4.4 4.8 5.2 5.6 6 6.4 6.8 Abundance vs. acquisition time (min)
7.2
7.6
8.4 8.8 9.2 9.6
10
Figure 3.
MRM chromatogram of Sudan Reds and Para Red.
+ TIC MRM (****) WorklistData1-r002.d 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 0.0 0.4 0.8 1.2 1.6
12
23
3 4
45
2.4 2.8 3.2 3.6
4 4.4 4.8 5.2 5.6 6 6.4 6.8 7.2 Counts vs. acquisition time (min)
7.6
8.4 8.8 9.2 9.6
10
Figure 4.
Blank chili pepper extract showing no response for any analyte.
10 2 1.0 0.8 0.6 0.4 0.2 0 10 1 1.0 0.8 0.6 0.4 0.2 0 10 1 1.0 0.8 0.6 0.4 0.2 0
+ TIC MRM (** **) WorklistData1-r010.d
12
2 3
3 4
4 5
+ MRM (294.00000[z=0] 156.00000) WorklistData1-r010.d
12
2 3
3 4
4 5
12
23
3 4
4 5
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 10 1 + MRM (249.00000[z=0] 93.00000) WorklistData1-r010.d 4 1 12 23 3 4 4 5 3 2 1 0 10 1 1.5 1.0 0.5 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 10 1 + MRM (277.00000[z=0] 121.00000) WorklistData1-r010.d 6 1 12 23 3 4 4 5 5 4 3 2 1 0 10 1 + MRM (277.00000[z=0] 156.00000) WorklistData1-r010.d 2.5 2.0 1.5 1.0 0.5 0
1
5.5
6.5
7.5
8.5
9.5
10
5
12
23
3 4
4 5
5.5
6.5
7.5
8.5
9.5
10
5
12
23
34
45
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 10 1 + MRM (353.00000[z=0] 77.00000) WorklistData1-r010.d 12 23 34 45 2.0 1 1.5 1.0 0.5 0
5.5
6.5
7.5
8.5
9.5
10
5
12 23 34 45 8 1 6 4 2 0 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 10 1 + MRM (381.00000[z=0] 91.00000) WorklistData1-r010.d 12 23 34 45 2.0 1 1.5 1.0 0.5 0 10 1 + MRM (381.00000[z=0] 225.00000) WorklistData1-r010.d
5.5
6.5
7.5
8.5
9.5
10
5
1.5 1.0 0.5 0
12
23
34
45
0.5
1.5
2.5
3.5 4 4.5 5 5.5 6 6.5 Counts vs. acquisition time (min)
7.5
8.5
9.5
10
Figure 5. 6
Chili pepper spiked at 1 ppb. Total MRM chromatogram on top and then each transition, quantitation and qualifier of the respective five analytes.
Confirmation by Ion Ratios

The EU Decision 2002/657/EC set confirmation criteria to include two MS/MS transition ions and their relative ratios (to the most abundant of the two). [5] The criteria for the tolerances of those ratios were given by the relative intensity. This is shown in Table 3 and demonstrates that the higher the relative ratio, the tighter the tolerances for acceptance. We have measured these ratios for each of the analytes in both solvent and matrix and found the matrix has little effect on the measured ratios. Using Sudan Red I as the example, the ratios of the two ions are given in Table 4. In addition, the repeatability of the method was tested using a 1 ppb spike in curry powder. The results of 20 repeats shown in Table 5 demonstrate that the method can perform well within the accepted tolerance of 25%. All of the samples used for the validation study met the relevant identification criteria.
Table 5.
Repeatability of Sudan Red I Ion Ratios at 1 ppb in Curry Powder Quantitation ion (249-93) 130 141 145 152 135 146 134 139 137 147 143 156 147 144 148 152 143 140 141 142 143.1 4.49 Qualifier ion (249-156) 58 65 71 62 66 66 64 69 71 64 61 65 64 64 65 63 68 66 62 62 64.8 4.87 Ratio 45.0 46.0 48.9 41.0 49.1 45.3 47.7 49.5 51.8 43.9 42.4 41.7 43.5 44.7 44.1 41.2 47.7 46.7 44.2 43.8 45.4 6.55
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Average STD (%)
Table 3.
Maximum Permitted Tolerances for Relative Ion Intensities Using LC-MSn Tolerance for LC-MSn (%) 20 25 30 50
Relative intensity between two transition ions (% of most intensive ion) > 50 > 2050 > 1020 10
Linearity, Sensitivity, and Recovery

Pearsons Correlation Coeffcient (R2) was used as a measure of the standard curve linearity where an R2 value of at least 0.99 was deemed acceptable. The linearity of the dyes in each of the food matrices was at least 0.99 except where the 1 ppm standard showed saturation, probably of the electrospray current. When that concentration is omitted, all calibrations in all matrices have an R2 > 0.99. An example of this is shown in Figure 6, where red chili pepper is spiked with Sudan Red I. The detection limit of each of the dyes was estimated by analyzing them in a standard at 0.2 ppb. The total MRM chromatogram is shown in Figure 7. Because each of the signals has little noise, it is difficult to determine with accuracy an s/n of 3:1.
Table 4.
Ratio of Quantitation vs Qualifier Transition Ion in Tested Matrices at 100 ppb Spike of Sudan Red I Ratio 46.8 48.5 46.5 46.6 47.7
Matrix Solvent Curry powder (Madras) Red pepper powder (Korean Farm) Red pepper powder (Oriental Mascot) Worcestershire sauce (Lea & Perrins)
Increasing the gain on the electron multiplier would have given both greater signal and noise and made this determination more accurate. It is reasonable to estimate that the limit of detection (LOD) is close to this concentration.
Finally, recovery of the sample preparation procedure was examined for the Worcestershire sauce since this was a liquid-liquid extraction. This was done with a replicate of three for each level of 100 and 1 ppb. These results are given in Table 6 and show reasonable recoveries.
1.8 1.7 1.6 1.5 1.4 1.3 1.2 1.1 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 _0.1
Sudan Red 1 to 4 Levels, 4 Levels Used, 12 Points, 12 Points Used, 0 QCs
y = 1701.9406 * + 767.3814 R^2 = 0.99989049
Responses (10 5)
_5
10
15
20
25
30
35
40
45
50
55
60
65
70
75
80
85
90
95
100 105
Concentration (ng/mlL)
Figure 6.
Standard curve of 1 ppb 100 ppb Sudan Red I in chili powder.
10 1 4.8 1 4.4 4.0 3.6 3.2 2.8 2.4 2.0 1.6 1.2 0.8 0.4 0.0
+ TIC MRM (** **) STD_LOD_04.d
12
23
3 4
4 5
0.5
1.5
2.5
3.5
4.5
5.5
6.5
7.5
8.5
9.5
10
Counts vs. acquisition time (min)
Figure 7.
MRM chromatogram of each of the dyes at 0.2 ppb in solvent, representing a response close to the detection limit.
Table 6.
Recovery of the Five Dyes in the Worcestershire Sauce Para Red (%) Sudan Red I (%) 71.0 66.2 Sudan Red II (%) 63.4 50.1 Sudan Red III (%) 62.1 54.1 Sudan Red IV (%) 70.2 51.2
100 ppb recovery (n = 3) 1ppb recovery (n = 3)
57.8 54.5
Conclusions
The method described herein for the analysis of four Sudan Red compounds and Para Red in four different matrices has been shown to be highly effective in meeting the criteria for quantitation and confirmation. Optimization of the method was simple, because few parameters in the mass spectrometer need adjustment. In addition, basic requirements for a validated method have been shown. These include sensitivity, repeatability, linearity, and recovery. This shows the Agilent 6410 Triple Quadrupole LC/MS system to be a highly effective instrument for the analysis of Sudan Reds and other azo dyes in food spices.
4. F. Calbiani, M. Careri, L. Elviri, A. Mangia, I. Zagnoni, Accurate Mass Measurements for the Confirmation of Sudan Azo Dyes in Hot Chili Products by Capillary Liquid Chromatography-Electrospray Tandem Quadrupole Orthogonal-Acceleration Time of Flight Mass Spectrometry. Journal of Chromatography A 2004, 1058, (12), 127135. 5. 2002/657/EC: Commission Decision of 12 August 2002 implementing Council Directive 96/23/EC concerning the performance of analytical methods and the interpretation of results (notified under document number C[2002] 3044), OJ L221/8 (17.8.2002) In 2002; Vol. L221, p 8.
References
1. IARC Summaries and Evaluations; Sudan I; International Agency for Research on Cancer: 1975. 2. 2003/460/EC: Commission Decision of 20 June 2003 on emergency measures regarding hot chili and hot chili products (notified under document number C[2003] 1970) OJ L154/114 (21.6.2003). 3. 2004/92/EC: Commission Decision of 21 January 2004 on emergency measures regarding chili and chili products (notified under document number C[2004] 68) OJ L27/52 (30.1.2004).

Analysis of Suspected Flavor and Fragrance Allergens in Cosmetics Using the 7890A GC and Capillary Column Backflush Application
Food
Authors
Frank David Research Institute for Chromatography Pres. Kennedypark 26 8500 Kortrijk Belgium Matthew S. Klee Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA
quantitative methods are needed to monitor these target solutes in various types of cosmetic products. Depending on the sample matrix and solute concentrations, different sample preparation methods are developed and applied [2]. For the determination of allergens in cosmetic products, one of the major problems is related to the presence of less volatile or nonvolatile constituents, such as detergents (nonionic or ionic), waxes, lipids, etc. These constituents will contaminate the analytical system if the samples are introduced without selective sample preparation. Selective extraction or selective sample introduction is not easy, however, since the target compounds cover a broad volatility range (from limonene to benzyl benzoate) and polarity range (from relatively polar benzyl alcohol to apolar benzyl benzoate). The method of choice should therefore give ppm sensitivity on one hand, and avoid discrimination of the target solutes based on relative volatility or polarity, on the other hand. Moreover, for routine analysis in a quality control environment, sample preparation should be minimized and direct injection of a nonselective solution or extract is preferred. Recently, liquid sample introduction with selective retention of nonvolatiles in a packed PTV liner in combination with automated liner exchange was developed and validated [3]. This approach, however, requires a dedicated sampler. In this application, an alternative method is proposed using a standard split/splitless inlet and Capillary Flow Technology. A QuickSwap device is used at the end of the column (coupled to the mass spectrometer transfer line), thereby allowing
Abstract
Flavor and fragrance allergens are determined in cosmetics using GC-MS. After simple sample preparation by nonselective extraction/dilution, extracts were injected and analyzed under fast screening conditions and locked retention times. After elution of the target solutes, the low-volatility matrix constituents, such as detergents, were effectively removed using capillary column backflush. Column and detector contamination were thereby strongly reduced and sample throughput was significantly increased.
Introduction
According to EU directive 2003/15/EC [1], 27 fragrance compounds in cosmetic products should be labeled if their concentrations exceed 100 ppm (mg/kg) for wash-off products, such as shower gels or soaps, or 10 ppm for leave-on products, such as perfumes or creams. Therefore, qualitative and
column outlet pressure to be controlled with auxiliary electronic pneumatic control (EPC). By lowering the inlet pressure and raising the outlet pressure after the last peak of interest has eluted from the column, sample components remaining in the column are forced back out the head of the column into the split inlet and are subsequently trapped on the split vent trap. The analysis is performed by GC-MS under retention-time locked conditions. The reference method, using a 30 m 0.25 mm id 0.25 m HP-5MS column operated under helium [2], was translated to a fast screening method for maximum throughput, using a 15-m column and hydrogen as the carrier gas. The analysis time needed for the separation of the target solutes was thereby reduced from 24 to 8 minutes (3X speedup). The low-volatility sample matrix constituents are backflushed from the column, avoiding column and detector contamination, baseline shifts, and ghost peaks due to carryover into subsequent runs.
GC-MS Conditions Column Carrier gas and pressure Column outlet and pressure Inlet Oven temperature program MSD setpoints
Sample Preparation
Samples are diluted to the 5% level (50 mg/mL) in an appropriate solvent (typically, acetone or dichloromethane is used). The sample is placed in an ultrasonic bath for 15 minutes to completely dissolve the target solutes in the solvent. After extraction and dissolution, the sample can be centrifuged and the supernatant transferred to an autosampler vial. In this application, data were obtained on a shampoo sample containing fragrance compounds and nonionic detergents.
GC conditions
All analyses were performed on an Agilent 7890A GC-5975 MSD combination. Injection was done using a 7683 ALS. The GC-MS conditions can be summarized as follows:
15 m x 0.25 mm id x 0.25 m HP-5MS Hydrogen QuickSwap Split/splitless in split mode Fast analysis (3X speedup*) Transfer line temperature Source temperature Quad temperature
Agilent P/N 19091-431 11.050 psi constant pressure 4 psi via auxiliary EPC 250 C, split ratio = 50:1 50 C (0.33 min) 240 C at 24C/min 250 C 300 C 150 C EMV +0V P/N G3185-60063 40350 amu, samples = 21 1.5 min 8.0 min
Tune QuickSwap restrictor Detection MSD events
Autotune 17 cm x 110 m id (4 psi) MS in scan mode Solvent delay Detector OFF (during backflush)
* Under these conditions, alpha isomethyl ionone elutes at 5.17 min, corresponding to a speed gain factor of 3 in comparison to a previously published retention time locking (RTL) method [2].
Backflush conditions (initiated at 8 min) Inlet pressure Auxiliary pressure Backflush time Backflush temperature 2 psi 70 psi 2.75 min 240 C
Results
First, the shampoo extract was analyzed in a typical modewithout applying backflush and programming the oven to 320 C to ensure that late eluters were eluted. In Figure 1, the overlay of the total ion chromatograms of 10 consecutive runs is shown. Excellent retention time and peak area repeatability is obtained in the first part of the chromatogram. In this sample, some allergens could be detected, including limonene (peak 1), linalool (2), eugenol (3), lilial (4), hexyl cinnamaldehyde (5), benzyl benzoate (6), and benzyl salicylate (7). After 8 minutes, no target solutes elute, but peaks corresponding to nonionic detergents are detected. Even using
a bakeout at 320 C, these compounds are not completely removed from the column. This can be seen from the appearance of ghost peaks (for instance, one at 11.7 minutes indicated by an arrow). This peak and others due to carryover increase regularly with added sample injections, clearly indicating that not all low-volatility sample material elutes from the column. Also, an increasing baseline is clearly observed after 10 minutes. It should be noted that from this 14-minute run, only the first 8 minutes are in fact needed for the necessary separation and quantitation of the target allergens. The remaining time represents the common practice of trying to removing highly retained sample components from the column by baking the column out. As demonstrated here, this is not so easily accomplished.
Abundance
5
8000000
7
7000000
6000000
Surfactants
5000000
4
4000000
320 C
6
3000000
2000000
1000000
2.00
3.00
4.00
5.00
6.00
7.00 Time
8.00
9.00
10.00
11.00
12.00
13.00
Figure 1.
Overlay of 10 consecutive analyses of shampoo extract (oven temperature programmed to 320 C, no backflush).
After this sequence of 10 sample runs, two blank runs were made. The chromatograms are given in Figure 2. Some contaminant peaks (probably extractables from repeated penetration of sample vial septum) elute around 6 to 8 minutes and are constant in both blank runs. The large peaks, eluting after 10 minutes, clearly show that high molecular weight materials were building up in the column and that these compounds were not removed, even by programming to 320 C. In a subsequent experiment, another six consecutive runs of the shampoo extract were made. For each analysis, the run was stopped at 8 minutes after the retention time of the most highly retained target allergen. After the sample runs, two blanks were run: one with the same temperature program as the samples, ending at 240 C (8 minutes), and another in which the temperature program continued to 320 C. The chromatograms of the sixth sample analysis, the first blank (stopped at 8 minutes) and the second blank (run to 320 C) are overlaid in Figure 3.
Some ghost peaks appear within the 8-minute analysis time window, even in the first blank. From the second blank run to 320 C, it is clear that lowvolatility solutes were accumulating in the column from each injected sample. Accumulation of sample material such as that shown in this example quickly leads to column deterioration and greatly reduces the ability to detect and quantify minor sample components. By following the typical approach of attempting to remove late-eluting sample components (cleaning off the column) at high temperature, not only is the column prone to premature degradation due to oxidation and cleavage of stationary phase polymer, but the contamination is moved from the column into the mass spectrometer source, degrading its performance and requiring more frequent cleaning. Next, a backflush method was set up and 10 new sample runs were made, followed by a blank run. The chromatograms of the sample analyses are shown in Figure 4.
Abundance 1000000 900000 800000 700000 600000 500000 400000 300000 200000 100000 0 2.00 3.00 4.00 5.00 6.00 7.00 Time 8.00 9.00 10.00 11.00 12.00 13.00
First blank run Second blank run
Retained contamination from samples
Septa extract peaks are constant for both runs
Figure 2.
Two consecutive blank runs after analysis of shampoo extract.
Abundance 9000000
2
8000000
7000000
6000000
5000000
7 Second blank, extended temperature
4000000
3000000
First blank following 6 sample analyses
2000000
1000000
Fast analysis of allergens (no backflush)

2.00 3.00 4.00 5.00 6.00 7.00 Time 8.00 9.00 10.00 11.00 12.00 13.00
Figure 3.
Overlay of sixth analysis of shampoo extract with run stopped at 240 C (bottom), first blank run to 240 C (middle), and second blank run to 320 C (top).
Abundance
8000000
2 5
7000000
6000000
5000000
7 4
4000000
3000000
6 1 3
2000000
1000000
0 2.00 2.50 3.00 3.50 4.00 4.50 5.00 Time 5.50 6.00 6.50 7.00 7.50 8.00
Figure 4.
Overlay of 10 consecutive analyses of shampoo extract (oven temperature programmed to 240 C, with backflush). 5
From Figure 4, it is clear again that excellent retention time and peak area repeatability were obtained with no evidence of carryover: no emerging ghost peaks; no increasing baseline. In Figure 5, the tenth run is overlaid with a blank that was run immediately following it. In the blank run, only contaminant peaks coming from the solvent vial septum are observed. The detergent peaks were efficiently and effectively removed from the column. The total analysis time was reduced from 13.6 min (programmed to 320 C, with a 2-minute hold) to
11 minutes (programmed to 240 C, with 2.75-minute backflush). Moreover, all low-volatile material was removed from the column, which was not the case with the longer run without backflush. An added bonus was that the oven cooldown and equilibration times were reduced because of the lower final oven temperature. Retention time peak area repeatability was determined for each of the seven identified allergens and is listed in Table 1. The standard deviation on the retention times is better than 0.002 minute (RSD < 0.03%). Also, excellent values are obtained for peak area repeatability.
Abundance 450000 400000 350000 300000 250000 200000 150000 100000 50000 0 2.00 2.50
Last of 10 runs with backflush
First blank following 10 sample runs
3.00
3.50
4.00
4.50
5.00 Time
5.50
6.00
6.50
7.00
7.50
8.00
Figure 5.
Overlay of 10 analyses of shampoo extract (oven programmed to 240 C) with backflush (top) and subsequent blank run (bottom).
Table 1.
Seven Identified Allergens RT min 2.3771 2.8372 4.4671 5.4312 6.5514 6.6467 7.1405 RT SD min 0.0005 0.0004 0.0003 0.0015 0.0016 0.0000 0.0013 0.0008 RT RSD % 0.020 0.015% 0.007 0.028 0.022 0.000 0.018 0.015 Area RSD % 1.80 1.60 1.60 1.53 2.00 2.00 2.98 1.95
Limonene Linalool Eugenol Lilial Hexyl cinnamaldehyde Benzyl benzoate Benzyl salicylate Average
Conclusions
For the determination of flavor and fragrance allergens in cosmetics, direct sample injection in a split/splitless inlet can be used. In comparison to a previously presented retention time locked method, the analysis time was reduced by a factor of three using a shorter column and hydrogen as carrier gas in combination with 5975 MSD. Contamination of the column and detector was minimized using the backflush method with the 7890A GC. A 20% reduction of the run time is obtained, with faster oven recycle times. Ghost peaks from previous injections were eliminated. Excellent retention time repeatability and peak area repeatability were obtained. Since the analysis of flavor and fragrance compounds is also performed on columns with a polar stationary phase and limited maximum operating temperature, for example, polyethylene glycol columns (MAOT 250 C), the capillary column backflush technique using Capillary Flow Technology with the 7890A GC is a very interesting tool to remove highly retained sample components at moderate temperatures.
References
1. Directive 2003/15/EC, Official Journal of the European Union, L 66/26, 11.3.2003. 2. F. David, C. Devos, and P. Sandra, LC. GC Europe, 19, 602-616, November 2005. 3. F. David, C. Devos, D. Joulain, A. Chaintreau, and P. Sandra, J. Sep. Science, 29, 1587-1594 (2006).

Rapid Screening and Analysis of Components in Nonalcoholic Drinks Application
Food
Author
Abstract
Soft drinks, juices, and prepared teas are popular drinks that are carefully formulated to provide unique flavor and reliable product stability. Many end-product facilities have the means to monitor critical ingredients, such as sugars, flavors, colorants, and preservatives, but may find their analyses are not fast enough to keep pace during periods of high seasonal demand. Here we take a classical approach to the analysis of additives and sweeteners and convert to a new method that improves sample throughput and resolution while also reducing solvent consumption.
10-m particle size columns may be candidates for modernization by replacing these columns with smaller-dimension columns packed with smaller particle sizes. The greatest opportunities in beverage QC labs exists with the small molecule separations of additives, preservatives, flavorants, and sweeteners. In methods development and research labs, compatibility with mass spectrometer ionization sources has also been a driving force in reducing column size to support low flow-rate operation. To reduce analysis time without sacrificing resolution, we can reduce column length and packing particle size simultaneously. The balancing effect is remarkably simple if reasonable optimization of the HPLC system, with respect to gradient delay volume and extracolumn dispersion, is practical. For example, a 250-mm long column with 5-m particles can be replaced by a 150-mm long column packed with 3-m particles. If the ratio of length to particle size is equal, in other words, resolution is preserved. Solvent consumption is proportionately reduced, and, if an equal mass of analyte can be successfully injected, the detector response should also increase due to reduced dilution of the peak as it travels through a smaller column of equal efficiency. Liquid chromatography/mass spectrometry (LC/MS) ionization sources, especially the electrospray ionization mode, have demonstrated greater sensitivity at lower flow rates than typically used in normal ultra violet UV/VIS optical detection (LC/UV) methods. In MS detection mode, it is
Introduction
Quality control (QC) laboratories are constantly challenged to meet higher sample loads with fewer instruments, reduced staff, and without sacrificing data quality in the process. While some labs have successfully adopted greater automation in sample preparation and data handling, they still find that opportunities exist for increasing their chromatographic throughput. Methods developed on 5- or
advantageous to reduce the column volume via reduced internal diameter. Lower flow rates will be required, proportional to the cross-sectional area of the inside diameter of the columns. The relationship of flow rate between different column diameters is shown in Equation 1.
2
Flowcol. 1
Diam.column2 Diam.column1
= Flowcol. 2
(eq. 1)
The second important consideration is the often overlooked effect of extracolumn dispersion on the observed or empirical efficiency of the column. As column volume is reduced, peak elution volumes are proportionately reduced. If smaller particle sizes are also employed there is a further reduction in the expected peak volume. Care must be taken by the user to minimize the extracolumn volume and to reduce, where practical, the number of connecting fittings and the volume of injection valves and detector flow cells. For gradient elution separations, where the mobile phase composition increases through the initial part of the analysis until the analytes of interest have been eluted from the column, successful method conversion to smaller columns requires that the gradient slope be preserved. It is most useful to express the gradient slope as % change per column volume. In this way, the change in column volume during method conversion can be used to accurately render the new gradient condition, regardless of flow rate or column diameter and length. We can express the gradient as shown in Equation 3.
We normally scale the injection mass to the size of the column, and a proportional injection volume would be calculated from the ratio of the void volumes, as shown in Equation 2.
Volumecolumn2 Volumecolumn1
Inj. vol.col. 1
= Inj.Vol.col. 2
(eq. 2)
For isocratic separations, the above conditions will normally result in a successful conversion of the method with little or no change in overall resolution. There are several other parameters that should be considered. The first of these parameters is the column efficiency relative to flow rate, more correctly efficiency to linear velocity, as commonly defined by van Deemter [1] and others, and the second is the often overlooked effect of extracolumn dispersion on the observed or empirical efficiency of the column. Van Deemter observed and mathematically expressed the relationship of column efficiency to a variety of parameters, but we are most interested here in his observations that there is an optimum linear velocity for any given particle size, in a wellpacked HPLC column, and that the optimum linear velocity increases as the particle size decreases. Graphically, it is typically represented as van Deemter plots [2] of plate height versus linear velocity. Generally speaking, a reduction in particle size leads to higher efficiency, at higher linear velocities, and the optimum range of highest efficiency expands to offer a wide flow range where good performance can be expected.
% Gradient slope =
(eq. 3)
A large value for gradient slope yields very fast gradients with minimal resolution, while lower gradient slopes produce higher resolution at the expense of increased solvent consumption and somewhat reduced sensitivity. Lower gradient slopes are formed by increasing the gradient time or flow rate, or a combination, within an acceptable range of analysis time and operating pressure. Resolution increases with shallow gradients because the effective capacity factor, k*, is increased. Much like in isocratic separations, where the capacity term is called k', a higher capacity value directly increases resolution. The effect is quite dramatic up to a k value of about 510, after which little improvement is observed. In the subsequent examples, we will see the results associated with the calculations discussed above.
System
Agilent 1200 Series Rapid Resolution LC, consisting of: G1379B micro degasser G1312B binary pump SL G1367C HiP ALS autosampler SL, with thermostatic temperature control G1316B thermostatted column compartment SL G1315C UV/VIS diode array detector SL, flow cell as indicated in individual chromatograms ChemStation 32-bit version B.02.01
Results
The separation in Figure 1 was performed on a 4.6 250 mm, 5-m ZORBAX SB-C18, column thermostatted to 30 oC using buffered phosphate conditions. At this pH, benzoate and sorbate were adequately resolved on the SB-C18 material with ACN organic modifier. With the formic acid modifier, nominal pH 2.5, poor resolution of these two components was observed, though selectivity and peak shape of other components was comparable. Additionally, on-column degradation of benzaldehyde was observed, presumably to benzoate, at the lower pH. This has been observed in other experiments (data not shown) and is particularly evident as column temperature is increased. Preparing ammonium formate buffer at about the same pH as the phosphate would likely solve both problems and would be MS compatible. The method was then scaled in flow and time for exact translation to 3.0 100 mm, 3.5-m and 1.8-m columns (data not shown). The gradient was recalculated for 3.0 50 mm, 1.8-m material (Figure 2). Subsequent gradient calculations yielded higher speed examples (Figures 3 and 4) without resolution loss.
Columns
Agilent ZORBAX SB-C18, 4.6 mm 250 mm, 5 m Agilent ZORBAX SB-C18, 3.0 mm 50 mm, 1.8 m
Mobile phase conditions

Organic solvent: Aqueous solvent: Acetonitrile (ACN) or ACN containing 0.1% formic acid 20 mm phosphoric acid in Milli-Q water, pH 3.65 with ammonium hydroxide, or Milli-Q water containing 0.1% formic acid
Gradient conditions
Gradient slope: 2.8% per column volume See individual chromatograms for flow rate and gradient time.
Samples
1. Standard mixture of sodium saccharin, caffeine, aspartame, vanillin, benzoic acid, sorbic acid, benzaldehyde, all 50 g/mL in 1/1 methanol/water 2. Various soft drinks, decarbonated where applicable Conditions ZORBAX SB-C18 4.6 mm 250 mm, 5 m Column temp: 30 C Gradient: 20 mM H3PO4 pH 3.65 with ammonium hydroxide, 10% to 50% ACN in 25 min Gradient slope: 2.8% ACN/column volume Analysis flow rate: 1.41 mL/min Sample: Standards 50 g/mL each in methanol/water 1/1, 15-L injection Total analysis time: 37.5 min Detection: UV 230 nm, 10-mm 13-L flow cell, filter 2 seconds (default) (Datafile SDADDS000006.D)
9.955 - Vanillin
600 4.389 - Saccharin
500
8.104 - Aspartame
400
6.701 - Caffeine
mAU
300
200
100 1.901 5.641
0 0 2 4 6 8 10 12 14 16 18 min
Figure 1
Gradient separation of soft drink additives on 4.6 250 mm, 5-m ZORBAX SB-C18. 3
12.053 - Benzoate 12.499 - Sorbate
15.512 - Benzaldehyde
mAU 200 150 100 50 0 -50 0.5
2.697 - Aspartame
3.455 - Benzoate
1.254 - unk
1.5
2.5
3.5
3.681 - Sorbate
4.270 - Benzaldehyde min 1.320 - Benzaldehyde 1.4 min
1.832 - Caffeine
Conditions ZORBAX SB-C18 3.0 mm 50 mm, 1.8 m Column temp: 45 C Gradient: 20 mM H3PO4 pH 3.65 with ammonium hydroxide, 10% to 50% ACN in 5 min Gradient slope: 2.8% ACN/column volume Analysis flow rate: 0.6 mL/min Sample: Standards 50 g/mL each in methanol/water 1/1, 2.5-L injection Total analysis time: 9 min Detection: UV 210 nm, 6-mm 5-L flow cell with 0.12-mm id inlet heat exchanger, filter 0.2 seconds (Datafile SDADDS_3MM21004.D) Figure 2. Gradient separation of soft drink additives on 3.0 50 mm, 1.8-m ZORBAX SB-C18.
0.297 - Saccharin 0.624 - Caffeine
0.793 - Saccharin
mAU 175
0.849 - Aspartame
150 125 100 0.452 - unk. 75 50 25 0 -25 0 0.2 0.4 0.6 0.8
0.933 - Vanillin
2.952 - Vanillin
1.071 - Benzoate 1
1.136 - Sorbate 1.2
Conditions ZORBAX SB-C18, 3.0 mm 50 mm, 1.8 m Column temp: 45 C Gradient: 20 mM H3PO4 pH 3.65 with ammonium hydroxide, 10% to 50% ACN in 1.5 min Gradient slope: 2.8% ACN/column volume Analysis flow rate: 2.0 mL/min Sample: Standards 50 g/mL each in methanol/water 1/1, 2.5-L injection Total analysis time: 3 min Detection: UV 210 nm, 6-mm 5-L flow cell with 0.12-mm id inlet heat exchanger, filter 0.2 seconds (Datafile SDADDS_3MM31029.D) Figure 3. High-speed gradient separation of soft drink additives on 3.0 50 mm, 1.8-m ZORBAX SB-C18.
0.849 - Aspartame
1.071 - Benzoate
mAU 150 100 50 0 0 0.121 mAU 400 300 200 100 0 0 0.2 0.227 - Acesulfame K* 0.2
0.452 - unk.
1.136 - Sorbate
1.320 - Benzaldehyde
0.297 - Saccharin
0.624 - Caffeine
0.933 - Vanillin
0.4
0.6 0.628
0.8
1.2
1.4 min
0.935
0.4
0.6
0.8
1.2
1.322
1.4 min
mAU 300 200 100 0 0 0.2
0.625
0.850
0.935
1.070
0.4
0.6
0.8
1.2
1.321
1.4 min
Conditions See Figure 3 for chromatographic conditions *Acesulfame K tentative identification by reference to other samples under similar conditions and product label claim Sample: Upper panel: mixed standard Center panel: black cherry vanilla cola, 20-oz polymeric bottle, decarbonated, 2.5-L injection Lower panel: diet black cherry vanilla cola, 20-oz polymeric bottle, decarbonated, 2.5-L injection (Datafiles SDADDS_3MM31029.D, SDADDS_3MM31043.D, SDADDS_3MM31044.D) Figure 4. High-speed gradient separation of soft drink additives on 3.0 50 mm, 1.8-m ZORBAX SB-C18.
The conditions in Figures 3 and 4, using a 2.8% slope at increased linear velocity on 3.0 50 mm, 1.8-m material, yield a separation with better resolution than the original 4.6 250 mm method. Because the absolute plate count is lower for the 1.8-m column (12,000 vs. 24,000 predicted for the 250-mm 5-m column, based on typical calculations), the resolution increase is presumably related to the increased operating temperature, which lowers both solvent viscosity and nonspecific column-analyte interactions, and an improvement in solvent linear velocity relative to the optimum for 1.8-m materials. With only a 3-minute total analysis time, this is an excellent procedure for high-throughput screening and quantitation of a large number of samples.
Conclusion
Careful translation of gradient slope and the use of optimum linear velocity with sub-2-micron particles can enable users to take advantage of small format columns that yield fast analysis times without compromising resolution. Optimization of extracolumn volume helps minimize resolution losses that unduly degrade column performance, assuring the maximum resolution possible while improving analysis time substantially. We have demonstrated this potential with a complex mixture of typical beverage additives and encourage users to contact us for guidance on this application and when other method translation opportunities are identified.
1.484
References
1. J. J. van Deemter, F. J. Zuiderweg, A. Klinkenberg, A. Chemical Engineering Science 1956, 5, 271289 2. The Influence of Sub-Two Micron Particles on HPLC Performance, Agilent Technologies, application note 5989-9251EN, May 2003

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2006 Printe in the USA August 3, 2006 5989-5178EN
Superior Peak Shape of Xanthines and Metabolites Separated by Eclipse Plus C18 Application Brief
John W. Henderson Jr., William J. Long
Food and Beverage
A new, improved HPLC column technology is shown to provide better peak shape and efficiency. ZORBAX Eclipse Plus columns combine improved ZORBAX Rx-Sil, a high-purity, type B silica that is engineered for the best peak symmetry, with optimized bonding (C18, C8) reagents and bonding processes. The result is superior peak shape and efficiency.
Highlights
Advanced HPLC technology incorporated into Eclipse Plus columns delivers superior chromatographic performance. Zorbax Eclipse Plus C18 columns separate xanthines with similar selectivity to other C18 columns, however with superior peak shape. Sharper peaks yield higher sensitivity. Efficiency (N) of Eclipse Plus C18 is over 10% higher than any competitor for this analysis. Tailing factor (TF) of Eclipse Plus C18 is closer to ideal (1.00) for every peak in this analysis compared to the competition.
Xanthines are naturally occurring alkaloids, synthesized by certain species of plants. Caffeine is the most widely known, found in coffee beans, tea leaves, and cacao beans. Others found in food and beverages include theophylline and theobromine. Below is an isocratic analysis of xanthines and metabolites on Eclipse Plus and four other C18 columns. While all provide satisfactory peak shape and resolution, to the trained eye, one chromatogram exhibits exceptional chromatographic performance, Eclipse Plus C18 (PN 959993-902). For every peak in this analysis, Eclipse Plus C18 consistently has sharper peaks (higher plates (N) and better peak shape (TF) ).
2
mAU 80 Eclipse Plus 40 0 0 mAU 80 Competitor G 40 0 0 mAU 80 Competitor S 40 0

0 mAU 80 Competitor L 40 0 0 mAU Competitor X 80 40 0 0
3 4 5 6 6 8 min
min
Figure 1 analyte elution order: 1. 2. 3. 4. 5. 6. 1-Methylxanthine Theobromine Theophylline -hydroxyethyltheophylline 3-Propylxanthine Caffeine
8 min
8 min
min
A: 25 mM Na phosphate pH 7.0 B: MeCN (90:10) Flow: 1 mL/min Temperature 40 C Columns 4.6 150 mm 5 m
Figure 1.
Separation of xanthines on various C18 columns.
Tables 1 and 2 show the difference in efficiency and peak shape between Eclipse Plus and other C18 columns based on the chromatograms in Figure 1. Clearly, Eclipse Plus C18 has higher efficiency and better peak shape than any of the competitors. Many factors influencing efficiency and peak shape include: extra column volume, detector data sampling rate, and mobile phase conditions. These factors can be ignored in this example however, because the chromatographic system and method were the same for each analysis. Therefore, the difference in performance is entirely from the column. Factors attributed to the column that influence efficiency include: silanol interactions with basic analytes such as xanthines (see Figure 2 for chemical structures), column voids inherent in manufacture or from dissolution of silica from mobile phase, and particle size and distribution (column loading). Eclipse Plus superior peak shape over the others is based on advances in HPLC technology that reduce or eliminate these factors.
Table 1: Theoretical Plates (N) of Xanthines by Various C18 Columns
Eclipse Plus
1-Methylxanthine Theobromine Theophylline -hydroxyethyltheophylline 3-Propylxanthine Caffeine 14900 15100 15100 15000 14500 15600
References
High Throughput Separation of Xanthines by Rapid Resolution HPLC, Agilent publication 5989-4857EN New ZORBAX Eclipse Plus LC Columns, Agilent publication 5989-4934EN
G
13100 13300 13800 13300 13500 13600
S
10700 10700 10400 10100 10400 10600
L
12600 12600 13400 13200 13300 14000
X
15900 13100 12600 12400 12700 12800
John W. Henderson Jr. and William J. Long are application chemists based at Agilent Technologies, Wilmington, Delaware.

Table 2: Tailing Factor (USP 5%) of Xanthines by Various C18 Columns
Eclipse Plus
1-Methylxanthine Theobromine Theophylline -hydroxyethyltheophylline 3-Propylxanthine Caffeine O N N N CH3 N CH3 1.11 1.08 1.05 1.02 1.01 1.00 O N N N H N H H3C N N CH3
G
0.94 0.95 0.93 0.92 0.91 0.91
S
1.18 1.19 1.15 1.13 1.11 1.10 O N N
L
1.13 1.14 1.12 1.10 1.09 1.09
X
1.15 1.18 1.17 1.15 1.14 1.14 O N N CH3 H
CH3
CH3 N N
Theophylline (1,3-dimethylxanthine)
O H3C N N CH3 H N NH O
Basic xanthine structure
Caffeine (1,3,7-trimethyxanthine)
Theobromine (3,7-dimethylxanthine)
Copyright 2006 Agilent Technologies All Rights Reserved. Reproduction, adaptation or translation without prior written permission is prohibited, except as allowed under the copyright laws. Printed in the USA June 20, 2006 5989-5237EN
O H3C N N H N N
CH3
O H 3C N N H H N N
1,3-dimethyluric acid (theophylline metabolite)
1,7-dimethylxanthine (caffeine metabolite)
1-methylxathine (theophylline metabolite)
Figure 2.
Structures of xanthines and metabolites in this study.
High Throughput Separation of Xanthines by Rapid Resolution HPLC Application Note
Biochemistry, Food and Beverage, Pharmaceutical
Author
John W. Henderson and Ronald E. Majors Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA
described to analyze theobromine, theophylline and caffeine in liquid refreshments (tea, chocolate syrup, and cocoa).
Introduction
Xanthines are a group of alkaloids that are commonly used for their effects as mild stimulants and as bronchodilators, notably in treating the symptoms of asthma. The most common xanthine is caffeine and it is found in foods such as coffee beans, tea, kola nuts, and in small amounts in cacao beans. Surprisingly, chocolate is a weak stimulant due to its content of theobromine, theophylline, and caffeine which are all methylxanthines. The chemical structures of these xanthines and some of their metabolites are depicted in Figure 1.
O N H O H3C O CH3
Caffeine (1,3,7-trimethylxanthine)
Abstract
Xanthines were found to be optimally separated by reversed-phase HPLC on a C18 column. By a reduction in column length and particle size, the separation time for a mixture of the xanthines investigated was reduced from 8 minutes to 1.5 minutes without a major loss in resolution. A simple isocratic HPLC method was used and
H N N N CH3
Theophylline (1,3-dimethylxanthine)
O N CH3 O
H N N
CH3 N N
H3C N N
O N N CH3 H O
N H
Basic xanthine structure
Theobromine (3,7-dimethylxanthine)
O H3C O N N CH3
H N N H O
O H3C O N H
CH3 N N
O H3C O N H
H N N
1,3-dimethyluric acid (theophylline metabolite)
1,7-dimethylxanthine (caffeine metabolite)
1-methylxanthine (theophylline metabolite)
Figure 1.
Structures of selected xanthines and metabolites used in this study.
The xanthines are absorbed in the body almost 100% and they appear in the blood in a few minutes after ingestion. Xanthines stimulate the central nervous system, can affect the circulatory system, and relax muscles in the bronchi. Caffeine is well known for its effect on reducing drowsiness and fatigue and improving alertness. These common xanthines are metabolized to a variety of compounds that may have physiological effects on the human body. The xanthines are most often separated by reversed-phase HPLC (RP-HPLC) on a C18 column [13]. Although ion pair chromatography has been used for xanthine separation [2], RP-HPLC with buffered water and acetonitrile requires a much simpler mobile phase system. This application note will show how different stationary phases may impart different selectivities for xanthine separations and will also investigate the effect of particle size and column length on the separation speed. Finally, it will show an application of a method for the analysis of caffeine and theobromine in chocolate-based drinks.
Selection of the Stationary Phase for the Separation of Xanthines HPLC allows for the resolution of peaks of interest in the shortest possible time. Selection of the appropriate stationary phase is an important step in method development. Initially, several different stationary phases were tried in order to choose an appropriate one for these investigations. Figure 2 shows the separation of the xanthine components in the test mixture using ZORBAX StableBond phases [cyano (CN), phenyl, C18] and a polar embedded stationary phase (ZORBAX Bonus RP). Although the particle size was different for the Bonus column, our objective here was to choose the phase with the best selectivity and the shortest retention time. Under the conditions employed, the CN column gave least retention (by virtue of its shorter alkyl chain length), but the column failed to resolve two of the test xanthines. The SB-C18 column gave the best overall separation in the shortest time and thus became the stationary phase of choice.
mAU 40 20 0 0.5 mAU 30 20 10 0 0.5 mAU 100 50 0 0.5 mAU 40 20 0 0.5
1 2
ZORBAX StableBond (SB)-C18 1.8 m

4
1 1
1.5
2.5
min
SB-Phenyl 1.8 m*
2 3 4
1 1,2
1.5
2.5
min
SB-CN 1.8 m
1 1 2,3
1.5 4
2.5
min
ZORBAX Bonus RP 100 mm 3.5 m
1 Columns:
1.5
2 SB-C18, 4.6 50 mm, 1.8 m SB-phenyl, 4.6 50 mm, 1.8 m SB-CN, 4.6 50 mm, 1.8 m Bonus-RP, 4.6 100 mm, 3.5 m A = 0.2% FA B = AcCN w 0.2% FA 98% A 2% B (v/v) 1.5 mL/min 2, 4, 6 L, respectively DAD, 254 nm 3 L, 2-mm flow path
2.5
min
Compounds: 1 1-methylxanthine 2 1,3-dimethyluric acid 3 3,7-dimethylxanthine 4 1,7-dimethylxanthine * Preproduction batch
Mobile phase: Isocratic composition: Flow rate: Injection volume: Detector: Flowcell:
Figure 2.
ZORBAX stationary phase selectivity comparisons for xanthines.
The chromatographic conditions chosen for subsequent experiments appear below:
The Effect of Particle Size and Column Length on the Separation of Xanthines Recent trends in HPLC have pointed to the use of shorter columns with smaller particles. The end result is a faster separation with the same or similar resolution. Figure 3 depicts the isocratic separation of the xanthine test mixture on three different columns (250 mm, 100 mm, and 50 mm) packed with three different particle sizes of ZORBAX StableBond C18 (5 m, 3.5 m, 1.8 m, respectively). As the column length decreases, one would expect to see shorter retention times, proportional to the decrease in length. Indeed Figure 3 clearly shows a decrease in overall separation time from 8 minutes to 1.5 minutes.
LC Conditions Column: Mobile phases: Isocratic composition: Flow rate: Detection: Flowcell: ZORBAX SB-C18 (various lengths and particle diameters shown on chromatograms), A= 0.2% Formic acid (FA) B=Acetonitrile with 0.2% FA 98% A 2% B (v/v) 1.5 mL/min; Injection volumes are shown on chromatograms DAD, 254 nm 3 L, 2-mm flow path
mAU 60 40 20 0 1 mAU 60 40 20 0 1 mAU 60 40 20 0 1
4.6 250 mm, 5 m
R2,3 = 9.30 R3,4 = 3.71 R4,5 = 4.31
8 min.
min
4.6 100 mm, 3.5 m R2,3 = 8.53 R3,4 = 3.37 R4,5 = 3.69
2 3 4 5 6 7
3 min.
min
4.6 50 mm, 1.8 m R2,3 = 9.30 R3,4 = 3.61 R4,5 = 3.87

2 3 Column: Mobile phase: Isocratic composition: Flow rate: Injection volume: Detector: Flowcell: 4 5 6 7
1.5 min.
min
Solutes: 1 1-methylxanthine 2 1,3-dimethyluric acid 3 3,7-dimethylxanthine 4 1,7-dimethylxanthine
ZORBAX SB-C18 (see chromatogram for dimensions and particle size) A = 0.2% FA B = AcCN w 0.2% FA 98% A 2% B (v/v) 1.5 mL/min 2, 4, 6 L, respectively DAD, 254 nm 3 L, 2-mm flow path
Figure 3.
Column scalability: change in column configuration to increase speed while maintaining resolution.
On the other hand, one would also expect to see a reduction in column efficiency. However, by reducing the particle size, the overall efficiency and resolution is nearly the same. The calculated resolution for all pairs of xanthines is shown on Figure 3. This is the rapid resolution concept where a combination of shorter columns and smaller particles led to equivalent separations at greatly reduced separation time. Since the flow rate is the same, in this case 1.5-mL/min, the solvent use is decreased proportional to column length resulting in an overall cost reduction. Another advantage when converting to shorter columns is that the peaks are narrower. Thus, if the same sample mass is injected the resulting increase in peak height provides greater sensitivity. In Figure 3, the sample volume was reduced proportional to column length to keep peaks nearly the same peak height. Of course, as one decreases the particle size of a column, the column backpressure increases with the inverse square of the particle diameter. Thus, if the same column length was used, the pressure at the same flow rate (or more correctly the linear velocity), the pressure would go up by a factor of 2 for a 3.5-m particle versus a 5.0-m particle and a factor of almost 8 for a 1.8-m particle. However, with the increase in plate count for the smaller
mAU 60 40 20 0 0.2 mAU 60 40 20 0 0.2 0.4 0.6 0.8 1 0.4 0.6 0.8 1
particles, columns can be shortened and the actual pressure increase is more nominal as can be seen in Table 1. Agilent's engineered particle size distribution helps to keep the pressure lower than what one would anticipate for a 1.8-m column.
Table 1. Particle diameter, um 5 3.5 1.8 Pressure as a Function of Particle Diameter and Column Length* Column length, mm 250 100 50 Pressure, bar 181 155 264 Pressure increase (relative to 5.0) 1.0 0.86 1.46
* Conditions of Figure 3
In order to demonstrate that a change in the particle size of the column packing has a minimal effect on selectivity, the isocratic separation of the xanthine test mix as a function of particle size at constant column length was investigated. Figure 4 shows a minimal variation in retention but a significant decrease in peak width in going from the 5-m column to the 1.8-m column. In other words, the column showed more efficiency and subsequent better resolution for the 1.8-um column but the selectivity was mostly unaffected. The HPLC conditions are shown on the chromatogram in Figure 4.
2,1 = 1.36
5 m
3,2 = 1.27 4,3 = 1.79

1.2 1.4 1.6 1.8 min
3.5 m
2,3 = 1.40 3,2 = 1.29 2,3 = 1.78

1.2 1.4 1.6 1.8 min
mAU 60 40 20 0
1.8 m
2,3 = 1.36 3,2 = 1.27 4,3 = 1.79
0.2
0.4
0.6
0.8 Column: Mobile phase:
1.2
1.4
1.6
1.8
min
Solutes: 1 1-methylxanthine 2 1,3-dimethyluric acid 3 3,7-dimethylxanthine 4 1,7-dimethylxanthine
Isocratic composition: Flow rate: Injection volume: Flowcell:
ZORBAX SB-C18 A = 0.2% FA B = 0.2% FA in MeCN 98% A 2% B (v/v) 1.5 mL/min 2 L 2 L, 3-mm flow path
Figure 4. 4
Column selectivity as a function of particle size.
Analysis of Xanthines in Liquid Refreshments The three most common xanthines are caffeine, theophylline, and theobromine. These xanthines may be present in a variety of drinks, either as part of the flavoring or added to enhance taste or increase alertness. We developed a simple isocratic method to analyze for them in chocolate drink and tea. Using the same chromatographic conditions described earlier; an excellent separation of a standard xanthine mixture was achieved. See Figure 5.
Next, three different liquid refreshments-hot cocoa, chocolate syrup, and black tea (bag) were prepared using directions on the container, but using sonication for mixing. After preparation, all solutions were centrifuged and then the aqueous portion was filtered through a 0.45-micron filter to remove any particulates that may foul the HPLC column. In particular, the hot cocoa gave a distinctive fat layer, but only the aqueous layer was sampled for analysis. For each sample, a 3-L injection of the aqueous extract was made. See results in Figure 6.
Norm. 60 50 40 30 20
Theobromine
Sample: 0.1 mg/mL of each analyte, 3-L injection
Theophylline
Caffeine
10 0 0.5 1 1.5 2 2.5 3 3.5 4 min
Figure 5.
mAU 250
Separation of xanthine standards.
Black tea bag

1 2
200
150
Chocolate syrup
2 1
100
50
Hot cocoa mix

2
0 0 1 2 3 4 5 min
Peak 1: Theobromine Peak 2: Caffeine (callouts show expanded absorbance range X10)
Figure 6.
Analysis of liquid refreshments for xanthines.
From the raw areas we were able to do a semiquantitative analysis (single-point calibration) of the three xanthines in the drinks. As can be seen from Table 2, in the chocolate drinks, relatively large levels of theobromine were observed but smaller amounts of caffeine while for the tea sample, caffeine was in an excess. The results of Table 2 were based on a weight/weight basis and not on total milligrams in the drink solution itself. These results are within the concentrations expected based on the manufacturers approximations. No theophylline was observed in any of the drinks.
Table 2. Determination of Xanthines in Liquid Refreshments Theobromine (%, wt/wt) 0.15 0.13 0.056 Caffeine (%, wt/wt) 0.011 0.011 0.17
Beverage Hot chocolate Chocolate syrup Tea
Conclusions
Xanthines were found to be optimally separated by reversed-phase HPLC on a C18 column. By a reduction in column length and particle size, the separation time for a mixture of the xanthines investigated was reduced from 8 minutes to 1.5 minutes without a major loss in resolution. A simple isocratic HPLC method was used to analyze theobromine, theophylline and caffeine in liquid refreshments (tea, chocolate syrup, and cocoa).
References
1. U. Huber, Analysis of Antiasthmatic Drugs by HPLC, Agilent Technologies, publication 5988-2523EN, www.agilent.com/chem. 2. R. Ricker, High Speed Separation of Analgesics, Agilent Technologies, publication 5988-6414EN, www.agilent.com/chem. 3. Q. Wang, Analysis of Xanthines in Serum, Agilent Technologies, publication 5988-2523EN, www.agilent.com/chem.
Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material.

Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2006 Printed in the USA March 6, 2006 5989-4857EN
Using TOF for Screening and Quantitation of Sudan Red Colorants in Food Application
Food
Authors
Yanyan Fang Agilent Technologies, Inc. 412 Ying Lun Road Pu Dong, Shanghai 200131 China Michael Zumwalt Agilent Technologies, Inc. 9780 S Meridian Blvd., MS 1-1E Englewood, Colorado 80112-5910 USA
Regulations Relating to Food Colorants (R.1008) of the Foodstuffs, Cosmetics and Disinfectant Act 54 of 1972. Furthermore, the European Commission extended the scope of the ban on Sudan Red after it was revealed that related chemicals were also being found in chili products adulterated with Sudan Red. These chemical varieties are classified as Sudan Red I, II, III, and IV. All are considered to be carcinogenic. There is now an emergency measure in force dictating that chili and chili products, including curry powder, can only enter an EU country provided it has proof that these illegal chemical dyes are not present. These rulings extend tight measures that have been in place since June 2003, when France alerted the European Commission that traces of Sudan Red I were found in chili and chili powder imports from India. The inclusion of curry powder, which is found extensively in European food products, could lead to a surge in product recalls for the food industry. In the UK alone, the food industry has recalled for destruction more than 160 products from the supermarket shelves since July 2003 as part of the enforcement of the new measures. Many countries including Canada, South Africa, France, Germany, and China have also discovered traces of Sudan I in their cuisines [1]. Figure 1 shows the structures of the banned dyes.
Abstract
The Sudan Reds are four commercial dyes, which could increase the risk of cancer if consumed in foods over long periods, or in large quantities. Now, they are banned in most countries. This application note presents the capability of the Agilent Time of Flight Mass Spectrometer (LC/MSD TOF) to be used as a screening, confirmation and quantitation tool-within-one analytical run of 5 min.
Introduction
Sudan Red is a red dye, traditionally used for coloring solvents, oils, waxes, petrol, and shoe and floor polishes. However, it is prohibited for use by the
Sudan Red I
HO N N
Sudan Red II
HO H3C CH3 N N
Sudan Red III

HO N N N N
Sudan Red IV
CH3 CH3 N N N N HO
Figure 1.
Structures of Sudan Reds.
Some methods exist for Sudan Red analysis. In method GB 19681-2005, the separation is complimented with a solid phase extraction, which simplifies the analysis and concentrates the sample [2]. Accurate mass measurement greatly increases the confidence of identification because it inherently limits the possible number of candidate compounds. That is, the better the precision and accuracy of the mass measurement, the fewer the number of compounds theoretically possible for a given accurate mass. Even without complicated sample preparation, good results can be obtained. This application note provides an overview of the power of the Agilent TOF mass spectrometer for the screening, confirmation, and quantitation analysis of Sudan Reds. The TOF mass spectrometer provides accurate mass determination (<3 ppm) with good linearity, proving its use as an excellent tool for selectivity, specificity, and sensitivity for detection of analytes in different source matrices. The method used here is not intended to represent one that will determine the lowest possible level of any one particular analyte, or class of analytes, but rather is a procedure that could be expanded to cover a wider range of components used in screening analyses.
Experimental
Instrument Agilent 1100 Series LC/MSD TOF with Agilent 1100 binary pump and well plate autosampler
LC Conditions Column Mobile phases Gradient Post time Flow rate MS Conditions Ionization Gas temp Drying gas ESI, Positive 350 C 8 L/min ZORBAX XDB C18, 2.1 mm 50 mm, 1.8 m Agilent p/n: 922700-902 A: H2O with 5 mM NH4OAc B: Acetonitrile 03 min 95% B to 98% B 35 min 98% B 3 min 0.5 mL/min
Nebulizer pressure 45 psi Capillary V (+ve) 4000 V
Reference Mass Introduction with LC-TOF The Agilent TOF MS uses a reference mass for the generation of reliable and high-level accurate mass spectra. The electro-spray ion source for the TOF is a unique dual-spray assembly that allows the simultaneous and constant introduction of reference mass components consisting of ~2-M purine (m/z 121.050873 and HP-0921 (m/z 922.009798) calibrant compounds.
Results
Shown in Figure 2 is the total ion chromatogram (TIC) and the corresponding MS spectra for Sudan Red I, Sudan Red II, Sudan Red III, and Sudan Red IV. Note that all the compounds can be separated well within 5 min for a high-throughput analysis, which corresponds to at least half the time seen in most other methods.
Figure 2.
TIC showing the four peaks of the Sudan Red dyes along with the corresponding mass spectum of each peak.
Sudan Red I
Sudan Red II
Sudan Red III
Sudan Red IV
Figure 2.
TIC showing the four peaks of the Sudan Red dyes along with the corresponding mass spectum of each peak.
The ability to closely match the expected mass and the observed mass provides the analyst with a higher level of confidence in the assignment given to a chromatographic peak. In the screening for compounds in matrices such as food, this additional confidence is of great importance. This capability also allows the possibility of using this technique as a screening tool for a wide range of components. To demonstrate the use of high mass accuracy in confirming the presence of compounds like the Sudan Red dyes, the mass of Sudan Red I, from Figure 3, which is 249.1022 is entered into the
calculator of the TOF software to determine the empirical formula. (See Table 1.) So, given the accurate mass of 249.1022, mass accuracy of 3 ppm, and assumption that the compound only contains carbon (C), hydrogen (H), nitrogen (N), and Oxygen (O), only two empirical formulas are determined. However, the second match with only one carbon does not make any sense for the type of compounds being examined here and can be discarded so that only C16H13N2O is possible, which just happens to be the expected formula for the pseudo-molecular ion, (M+H)+, of Sudan Red I.
Table 1.
Theoretical Accurate Mass, Observed Mass and Mass Error Monoisotopic Retention Observed Compound mass time Adduct mass Sudan Red I Sudan Red II Sudan Red III Sudan Red IV 248.0950 276.1263 352.1318 380.4482 0.66 1.01 1.30 2.24 [M+H]+ [M+H]+ [M+H]+ [M+H]+ 249.1022 277.1340 353.1393 381.1709
Adduct accurate mass 249.1022 277.1335 353.1390 381.1709
Mass (ppm) 0.1589 0.1434 0.6993 0.2306
Figure 3.
Empirical formula calculator of TOF software determines two possible formulas given the observed mass of 249.1022 and known mass accuracy of <3 ppm.
A greater than two-fold increase in sensitivity for many components is seen with the narrowing of the extracted ion chromatogram (EIC) window. To demonstrate this, Figure 4 shows the reduction in noise that is observed with the extraction of a smaller mass range for Sudan Red I. As a result, the ability of TOF-MS to accurately determine the presence of the components at low level may assist with investigations into reporting the illegal use of the synthetic dyes, and prove to be a critical factor in confirmation when dealing with complex matrixes.
1-amu window (4000 ppm)
1000-ppm window
10-ppm window
Figure 4.
Effect of extracted ion range on noise-Sudan Red I in a real sample of Chinese Source spice oil.
It can be seen from Figure 4 that the ability to narrow the mass extraction window greatly reduces the noise for a given mass, and with the retention time information, can provide a high level of confidence in the assignment of a component.
Furthermore, the TOF can use in-source collisioninduced dissociation (CID) to generate fragment ions for confirmation as well. These additional ions can be seen in both the unextracted standards and the compounds in the food matrix. Such spectra obtained with the unextracted standards are shown in Figure 5.
Sudan Red I at different voltages
100 V
320 V
Sudan Red II at different voltages
100 V
320 V
Figure 5.
Sudan Red I, II, III, and IV at different in-source CID voltages. 7
Sudan Red III at different voltages
100 V
320 V
Sudan Red IV at different voltages
100 V
320 V
Figure 5.
Sudan Red I, II, III, and IV at different in-source CID voltages (continued).
TOF Linearity The components analyzed by TOF were tested for linearity as part of this study. Figure 6 shows the linearity of these four compounds. The regression values, r2, of over 0.99 were seen for each of these compounds.
Sudan Red I (r2 = 0.9996)
Sudan Red II (r2 = 0.9990)
Sudan Red III (r2 = 0.9995)
Sudan Red IV (r2 = 0.9992)
Figure 6.
Linearity of the four Sudan Red dyes.
LOD and LOQ Table 2 summarizes the limits of detection (LODs) and limits of quantitation (LOQs) for each of the components analyzed by this method. As shown in these results, this method is designed to provide a broad screening tool for the analysis of banned colorants.
Table 2. LOD and LOQ for Components by LC/MSD TOF LOD (ppb) (s/n = 3) 3.6 2.4 17.6 20.4 LOQ (ppb) (s/n = 20) 24.4 18.7 126.4 132.2
Component Sudan Red I Sudan Red II Sudan Red III Sudan Red IV
An example of the determination for LOD and LOQ is shown for the Sudan Red I dye in Figures 7 and 8, respectively. The upper chromatogram of each figure is the TIC, while the lower chromatogram corresponds to an EIC of m/z 249.1022 with a window of 10 ppm. The LOD is calculated based on a signal-to-noise (S/N) value of 3. Figure 7 corresponds to an on-column injection of 4.4 ppb and has a S/N value of 3.7. Therefore, the LOD is calculated as 4.4 ppb * (3/3.7) = 3.6 ppb, which is listed in Table 2. Figure 8 corresponds to an on-column injection of 33.7 ppb. The LOQ is calculated based on a S/N value of 20. Therefore, the LOQ is calculated as 33.7 ppb * (20/27.6) = 24.4 ppb, which is also shown in Table 2. Similar reasoning is used for the other dyes.
TIC
10-ppm window
Figure 7.
LOD of the Sudan Red I.
10
TIC
10-ppm window
Figure 8.
LOQ of the Sudan Red I.
Analysis of Sample The instrument and software provides the user with the ability to create a database for all components they wish to automatically screen for. The minimum requirement for this database is the empirical formula and name for the component of interest, although the inclusion of a retention time will assist with confidence in the confirmation and reduce analysis time. Sample Preparation The sample analyzed in this work is Indian curry powder, but it should be noted that similar results were found in the analysis of Indian, Chinese, and
South Korea chili powders. About 5 g of the Indian chili powder is weighed to within 10 mg in a ground Erlenmeyer flask with 100 mL of acetonitrile added using a graduated test tube. Sonicate for 1 hour and filter on filter paper in a ground Erlenmeyer flask. This method is identical to the one recommended by the EU. Screening and Recovery Results The prepared sample was spiked with 0.4 g/mL Sudan Red solution. The TIC for this sample is shown in Figure 9. However, also shown is the EIC of the sum of the four m/z values for each of the Sudan Red dyes.
11
Figure 9.
Real sample spectrum (upper); EIC of the 4-Sudan Red colorants (lower).
12
Table 3 illustrates the excellent quantitation accuracy of the spiked levels as compared to what is expected from the calibration curves. The levels of 0.4 and 0.2 ppm are below the 0.5-ppm cutoffs regulated by the EU but are well quantitated at the 0.4-ppm level with accuracies all close to 100%. At both levels the %RSDs of peak areas are still below 10%.
Table 3. Formula C16H12N2O Species [M+H]+ Formula C18H16N2O Species [M+H]+ Formula C22H16N4O Species [M+H]+ Formula C24H20N4O Species [M+H]+ Method Accuracy and Reproducibility for the Four Target Colorants Spiked in Indian Chili Compound name Sudan Red I Abundance (cts) 2.5354 e5 Compound name Sudan Red II Abundance (cts) 2.9543E5 Compound name Sudan Red III Abundance (cts) 8.5544E4 Compound name Sudan Red IV Abundance (cts) 7.9608E4 Mass 248.09496 Ion mass 249.10224 Mass 276.12626 Ion mass 277.13354 Mass 352.13176 Ion mass 353.13904 Mass 380.44821 Ion mass 381.17095 Peak RT (min) 0.65 Measured mass 249.10249 Peak RT (min) 1.01 Measured mass 277.13372 Peak RT (min) 1.30 Measured mass 353.13931 Peak RT (min) 2.24 Measured mass 381.17098 Peak area 6.6932E5 Error (mDa) 0.26 Peak area 8.8136E5 Error (mDa) 0.18 Peak area 3.2776E5 Error (mDa) 0.27 Peak area 2.7679E5 Error (mDa) 0.03 CAS number 842-07-9 Error (ppm) 1.0 CAS number 3118-97-6 Error (ppm) 0.7 CAS number 85-86-9 Error (ppm) 0.7646 CAS number 85-83-6 Error (ppm) 0.0787
Reproducibility
Standard (ppm) 0.2 0.4 0.8 1.6 2 Sudan Red I RSD Accuracy (%) (%, avg) 6.04 6.98 4.61 5.17 6.12 97.31 101.95 104.75 102.53 96.77 Sudan Red II RSD Accuracy (%) (%, avg) 5.76 5.64 6.12 5.99 4.74 97.61 100.27 103.78 105.92 94.72 Sudan Red III RSD Accuracy (%) (%, avg) 7.11 8.04 8.37 7.56 7.31 82.3 103.12 96.23 92.79 101.27 Sudan Red IV RSD Accuracy (%) (%, avg) 8.87 8.26 8.43 7.73 6.20 87.4 104.32 102.72 95.78 100.79
13
Conclusions
An overwhelming advantage of using TOF MS for the trace-level detection of any component is the confirmation that is provided through accurate mass. The data shown here demonstrates the ability of LC/MSD TOF to confirm with accurate mass measurement, and quantitate with selective, narrow mass window. High-mass resolution and accuracy (in every spectrum) provides the selectivity that reduces chemical noise for quantitation and confirmation. Also, this method can easily include new components such as other banned or limited use dyes. Furthermore, this application shows the capability of the Agilent LC/MSD TOF as a single tool for both screening and confirmatory analysis, with quantitative information, often at levels below those currently analyzed for today. Using the TOF can save more than 50% the time required as compared to many other methods used today. It provides a strong tool for the large quantity samples for screening.
References
1. ASTA White paper on Sudan and Related Dyes, a publication of the American Spice Trade Association (http://www.astaspice.org/pubs/ ASTAWhitePaper.cfm), copyright 2005 2. Solid-phase extraction procedure adapted from People's Republic of China approved national standards method of analysis for sudan dyes, 2005, Standardization Administration of China (http://www.sac.gov.cn/english/home.asp), Standard Number GB/T 19681-2995.

Separation of Paraben Preservatives by Reversed-Phase HPLC Application
Foods, Beverages, and Cosmetics
Authors
Coral Barbas and Javier Ruprez Facultad de CC Experimentales y de la Salud, Universidad San Pablo-CEU Urbanizacin Monteprncipe Boadilla del Monte, 28668 Madrid Spain Andre Dams* Agilent Technologies, Inc. Amstelveen The Netherlands Ronald E. Majors Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA
*Present address: Andre Dams Analytical Consultancy, Amstelveen The Netherlands
Synthetic methyl, ethyl, and propyl parabens were developed from benzoic acid and were considered effective and economical since they were inexpensive to use as both a cosmetic and food grade preservative. It is estimated that 99% of all cosmetic and body care products contain some form of paraben preservatives. Methyl and propyl parabens are generally recognized as safe (GRAS) substances. Recently, however, this preservative system has come into question as these substances were found in cancerous tissues, especially breast tissue. A study by the Journal of Pharmaceutical Science revealed that after receiving multiple doses of a gentamicin formula containing paraben preservatives, six infants found traces of up to 82.6% of the parabens in urine samples. Researchers of the Department of Biology and Biochemistry of Brunel University in the United Kingdom found that the greatest concern regarding parabens focuses on their estrogen-mimicking ability in laboratory animals. In addition, 2-phenoxyethanol (2PX), a chemical substance also used as a preservative in several vaccines, is sometimes used in conjunction with parabens. Paraben mixtures have the advantages of being broad-spectrum, leading to reduced inventory levels and cost savings. It is easier to handle one liquid in reasonable quantities rather than several small quantities of powders or liquids. Phenonip, a product of Clariant Ltd, Horsforth, Leeds, United Kingdom, is a mixture of parabens in 2PX solution. This product is probably the best known of the paraben mixtures and is often copied. Analysis of these substances at formulation and trace levels in foods and cosmetics is of great interest. HPLC is an ideal method for their separation and analysis.
Abstract
Paraben preservatives are shown to be readily and quickly analyzed using reversed-phase HPLC with a ZORBAX Eclipse XDB-C18 Rapid Resolution column.
Introduction
Preservatives are a class of chemical agents that are commonly used to prevent the growth of bacteria in foods, beverages, and cosmetics. The paraben preservatives (4-hydroxybenzoic acid esters) are among the most widely used. These preservatives were developed in the 1930s to stabilize creams.
Chemical Characteristics The structures of 2PX and the parabens are depicted in Table 1. Due to their phenyl ring, these compounds are UV-detectable at extremely low concentrations. Since they have no ionic functional groups, they are considered lipophilic. Due to this lipophilicity, some accumulation in fatty tissues of the body would be expected. Parabens are slightly soluble in water, with the solubility decreasing as
Table 1.
the ester chain length increases. For example, methyl paraben dissolves at the 0.25% (w/w) level at 20 C while butyl paraben is soluble at the 0.02% (w/w) level. Most of the parabens are freely soluble in alcohol, acetone, ether and a number of other organic solvents. With such solubility properties, reversed-phase chromatography (RPC) is an ideal separation technique. Many reversed-phase separations of parabens are published in the chromatography literature [14].
Structures and Concentrations of Preservative Compounds
O CH2
2PX: 2-Phenoxyethanol (1.4 mg/mL)
CH2 OH
O C OCH 3
MEP: Methylparaben (0.30 mg/mL)
HO O C OC 2 H5
ETP: Ethylparaben (0.07 mg/mL)
HO O CO
PRP: Propylparaben (0.04 mg/mL)
CH2 HO O CO CH2 HO O CO C 4 H9
CH2
CH3
CH3 CH CH3
IBP:
Isobutylparaben (0.04 mg/mL)
BTP:
Butylparaben (0.08 mg/mL)
HO
Chromatographic Conditions Column: Mobile phase: Gradient: ZORBAX Eclipse XDB-C18 Rapid Resolution, 4.6 mm 150 mm, 3.5 m Solvent A: Water Solvent B: Methanol Time 0 5 6 16 17 20 % MeOH 38 38 60 60 62 38
Conclusion
Paraben preservatives are readily and quickly analyzed using reversed-phase HPLC with a ZORBAX Eclipse XDB-C18 Rapid Resolution column, 4.6 mm 150 mm, 3.5 m.
References
1. Robert Ricker, High-Speed Separation of Parabens, Agilent Technologies, publication 5988-6356EN www agilent.com/chem. 2. M. Borremans, J. van-Loco, P. Roos, and L. Goeyens, (2004) Chromatographia., 59(12), 4753. 3. E. Marengo, M.C. Gennaro, and V. Gianotti, (2001) J. Chromatogr. Sci.; 39(8), 339-344. 4. J. E. Koundourellis, E. T. Malliou, and T. A. Broussali, (2000) J. Pharm. Biomed. Anal., 23(23), 469475.
Flow rate: Temperature: Detector: Injection volume:
0.8 mL/min 40 C UV 254 nm 5 L

The separation of the parabens and 2PX contained in a paraben product mix is depicted in Figure 1.
800

2
600
mAU
3
400 200 0 0.0 2.5 5.0 1 2 3 4 5 6 MEP 2PX ETP PRP IBP BTP
6 5
7.5 10.0 12.5 15.0 17.5 20.0 22.5 25.0 27.5 30.0 Minutes
Figure 1.
Separation of preservatives by reversed-phase HPLC.
Sample preparation merely involved dilution of the sample mix with methanol. All components were separated to baseline. On other columns, the separation of MEP and 2PX and IBP and BTP is usually quite difficult, especially in such a short analysis time (16 min). The method is reproducible with good separation efficiency.
Aspartame: Metabolites and Applications
Application
Food Analysis
Robert Ricker
Aspartame is of interest to the food industry as an artificial sweetener in diet drinks and light foodstuffs. It is also used as a sugar substitute and does not have the bitter aftertaste of saccharin and cyclamates. Speedy analysis is of particular interest to quality control labs which perform routine testing for aspartame and its metabolites in dietary food products.
2 1 Aspartame and Its Metabolites 1. Phenylalanine 2. 5-benzyl-3,6-dioxo-2-piperazineacetic acid 3. Aspartic acid-phenylalanine dipeptide 4. Aspartame 3 4
Highlights
Rapid analysis (while maintaining resolution of aspartame and its metabolites) is achieved with shorter columns (75!mm) and smaller particlesize (3.5!m) packings.
1 2 3 Diet Coke 1. Caffeine 2. Aspartame 3. Unknown
Sample preparation is minimal for liquid samples which can be diluted and injected directly onto the column. Solid samples are extracted with water and then injected onto the column.
Equal Sweetener 1. Aspartame
Conditions: ZORBAX SB-C18 (4.6 x 75 mm, 3.5 m) (Agilent P/N: 866953-902) Mobile Phase:!!85:15, 0.1% TFA:ACN 1 mL/min, 35C, Detect. UV (210 nm)
FD&C Colors
Application
Food Analysis
Robert Ricker
FD&C colorants are commonly used in foods and cosmetics. Propylparaben is a preservative found in the mixture of food colorings used for this analysis. Green food coloring is comprised of yellow # 5 and blue # 1. A combination of ion-pairing and gradient HPLC is used because of the wide chemical differences of compounds. Note the molecular formulas.
HPLC System: Column: Mobile Phase: Gradient: Detection: Flow: Temperature: Agilent 1100 with quaternary pump ZORBAX Eclipse XDB-C18 Rapid-Resolution (3.5 m), 4.6 x 50 mm Agilent Part No. 935967-902 A: 0.1% TFA, pH to 4.4 with TEA, B: MeOH 17 to 100% B / 4 min UV 254 nm 1 mL/ min. ambient
Highlights
ZORBAX Eclipse-XDB is designed for extended column life at intermediate pH.
1. Yellow # 5 2. Red # 40 3. Blue # 1 4. Propylparaben 5. Red # 3
C16H9N4Na3O9S2 C18H14N2Na2O8S2 C37H34N2Na2O9S3 C10H12O3 C20H4I4Na2O5
MW= 534 MW= 496 MW= 760 MW= 180 MW= 878
mAu 800 600 400 200 0 0 1
1 ZORBAX Eclipse XDB-C18 2
3 4 5
2 3 4 5 min
High-Efficiency and High-Speed Separation of Natural Anthocyanins

Application
Food Analysis
Robert Ricker
Anthocyanins are natural pigments responsible for the brilliant red and blue (and purple) colors found in many fruits and flowers. The colors of blueberries are the result of many different anthocyanins being present in the fruit. Qualitative and quantitative analysis of anthocyanins can be used to distinguish between different cultivars of blueberry plants and determine their quality. Therefore the chromatographic separation of anthocyanins is of increasing importance to the agricultural and wine industries. Recent interest in medicinal use of anthocyanins, as antioxidants/anticancer agents, has also stimulated interest in their chromatographic separation.
4.6 x 250 mm, 5m Time Percent B 0 23% 35 26% 97 60%
Highlights Traditionally, a low-pH mobile phase (containing formic acid) in these types of separations has caused degradation of the column and change in the separation (Goiffon, J.-P., Brun, M., and Bourrier, M.-J., J. Chromatogr. 537 101-121, 1991). ZORBAX StableBond SB-C18 columns provide the chromatographer with long-term stability for reverse-phase separations requiring very low pH.
4.6 x 150 mm, 3.5m
Time Percent B 0 23% 21 26% 58.2 60%
4.6 x 75 mm, 3.5m
Time Percent B 0 23% 10.5 26% 29.1 60%
In these experiments, phosphoric acid has replaced the traditional formic acid in the mobile phase (Gao, L. and Mazza, G., J. Liq. Chromatogr., 18(2) 245-259, 1995), resulting in a superior separation of >25 anthocyanins on the ZORBAX SB-C18 column.
Conditions: ZORBAX SB-C18 (4.6 x 250 mm, 5 m; 4.6 x 150 mm, 3.5 m; 4.6 x 75 mm; 3.5 m) (Agilent P/N: 880975-902, 863953-902, 866953-902, respectively) Mobile Phase: A:3% Phosphoric acid B: 100% MeOH, Gradient: (see individual chromatogram) Inj.: 20 L, 1 mL/min, 30C, Detect. UV(525 nm), Sample preparation: (see back)
As a result of reproducible products and smaller chromatographic particles (3.5m) of narow size-distribution, the analyst can systematically change column configuration to save both time and solvent.
Method of Preparing Blueberry Extracts* Begin my mixing: 10 g. Blueberries 10mL Solvent (70:28:2, MeOH:H2O, Formic acid) Blend for 10 minutes on ice. Filter through glass wool in a 10 mL syringe. Allow filtrate to sit for 1 hour. Filter through a 0.2 m filter. Inject 50 L immediately for HPLC analysis. *Method of Extraction and blueberry extracts were kindly provided by Drs. Willy Kalt and Jane McDonald, Agriculture and Agri-Food Canada, Kentville, Nova Scotia, Canada
application
LC / M S
N a t u ra l Fo o d C o l o ra n t s
Analysis of Natural Food Colorants by Electrospray and Atmospheric Pressure Chemical Ionization LC/MS Introduction
Many kinds of natural colors are used in beverages, jellies, and candies. In many countries, food regulations have been recently revised to cover natural colorants to the same degree as synthetic ones. Accordingly, it has become necessary to develop reliable analytical methods for various natural colorants in food. In this study, LC/MS methods using electrospray ionization (ESI) or atmospheric pressure chemical ionization (APCI) were developed to identify major pigments in four natural colorants: red cabbage, paprika, Monascus, and lac. compartment, vacuum degasser, autosampler, and LC/MSD. The LC/MSD used either an ESI or APCI source. Complete system control and data handling were done on the Agilent ChemStation for LC/MS. Operating conditions were optimized for each sample.

Red cabbage colorant. Figure 1 shows the structure of seven major pigments of red cabbage. The pigments share the basic cyanidin 3-diglucoside structure with differing R1 and R2 groups. Figure 2 shows the total ion chromatogram (TIC) and extracted ion chromatograms (EIC) of red cabbage pigments. Although every major pigment can be chromatographically separated using 10% formic acid in the mobile phase, the high acid concentration reduces sensitivity. Therefore, 1% formic acid was used in this study. The EICs show the separation of the pigments based on their main ion (base peak).
Experimental
Paprika and Monascus colorants were dissolved in acetone; the other colorants were dissolved in deionized water. Eachcolorant was filtered through a 0.2-m filter. A 10-l portion was injected into the system, which consisted of an Agilent 1100 Series binary pump, thermostatted column
Figure 1. The structure of major pigments in red cabbage colorant.
Figure 2. Total and extracted ion chromatograms of red cabbage colorant.
LC conditions Column: Mobile phase: Gradient: Flow rate: Column temp: Injection vol: MS Conditions Source: Ion mode: Vcap Voltage: Nebulizer: Drying gas flow: Drying gas temp: Corona: Vaporizer temp: Scan range: Step size: Peak width: Time filter: Fragmentor:
250 x 2.1 mm Inertsil ODS3, 5 m A = 1% formic acid B = acetonitrile Start with 5% B At 30 min 50% B 0.2 ml/min 40 C 10 l ESI positive 4000 V 50 psig 10 l/min 350 C 4 A 350 C 100-1200 amu 0.1 0.15 min on 200 V
Figure 3 shows the mass spectra of the seven major pigments in red cabbage colorant. For these pigments, the singly charged molecular ion is observed rather than the more typical [M+H]+ ion, because the cyanidin group already has a positive charge on an oxygen. In-source collision-induced dissociation (CID) can be used to generate fragment ions to provide structural confirmation. Using CID, mass spectra of these pigments show common fragments corresponding to the loss of a glucose, as well as cyanidin (m/z 287) and cyanidin 3glucoside (m/z 449) ions. Monascus colorant
Figure 3. Mass spectra of major pigments in red cabbage colorant.
Monascus contains six major pigments; their structures are shown in Figure 4. Four pigments were identified from the mass spectra of major peaks in the TIC. See Figure 5.
Figure 4. The major pigments of Monascus colorant.
LC conditions Column: Mobile phase: Gradient: Flow rate: Column temp: Injection vol: MS Conditions Source: Ion mode: Vcap voltage: Nebulizer: Drying gas flow: Drying gas temp: Corona: Vaporizer temp : Scan range: Step size: Peak width: Time filter: Fragmentor:
250 x 2.1 mm Inertsil ODS3, 5 m A = 1% formic acid B = acetonitrile Start with 50% B At 10 min 90% B 0.2 ml/min 40 C 10 l ESI positive 4000 V 50 psig 10 l/min 350 C 4 A 350 C 100-1200 amu 0.1 0.15 min on 100 V
Figure 5. The total ion chromatogram of Monascus colorant.
Figure 6. Mass spectra of major pigments in Monascus colorant.
Three major peaks with base peaks at m/z 439, 467, and 495 were not identified. Figure 6 shows the mass spectra of the identified pigments. Protonated molecular ions [M+H]+ were observed for the four identified pigments. Paprika color Capsanthin and the mono- and diesters of capsanthin with fatty acids are known as the major pigments in paprika colorant. See Figure 7. Two monoesters and five diesters of capsanthin were identified in the paprika colorant analyzed in this study. See Figure 8.
The protonated molecular ions [M+H]+ were observed for every major pigment. See Figure 9. However, with the exception of capsanthin monoeicosanoate, the intensity of these ions was very low. Except for capsanthin monoeicosanoate, the pigments show fragment ions resulting from the loss of one or two fatty acid fragments. A common fragment ion was observed at m/z 567 in the mass spectra of these pigments.
Figure 7. The structure of major pigments of paprika colorant.
Figure 8. The total ion chromatogram of paprika colorant.
LC conditions Column: Mobile phase: Gradient: Flow rate: Column Temp: Injection vol: MS Conditions Source: Ion mode: Vcap voltage: Nebulizer: Drying gas flow: Drying gas temp: Corona: Vaporizer temp: Scan range: Step size: Peak width: Time filter:
250 x 2.1 mm Inertsil ODS3, 5 m A = acetone B = methanol Start with 10% B At 10 min 90% B 0.2 ml/min 40 C 10 l APCI Positive 4000 V 50 psig 5 l/min 350 C 4 A 350 C 100-1200 amu 0.1 0.15 min On
Lac colorant Figure 10 shows the structure of the major pigments in lac colorant. Laccaic acids A, B, C are known as the major pigments in lac colorant. These compounds have the same basic anthraquinone structure but with different R groups. Three major peaks were detected in the TIC. See Figure 11. Although laccaic acids A, B, and C were identified, A and B could not be separated. Figure 12 shows the mass spectra of two peaks, laccaic acid C and a combination of laccaic acids A and B. The deprotonated molecular ions were observed at m/z 495, 536, and 538. Fragment ions resulting from the loss of carbon dioxide were observed at m/z 451, 492, and 494.
Figure 9. Mass spectra of major pigments in paprika colorant.
Figure 10. The strucuture of major pigments of lac colorant.
Figure 11. The total ion chromatogram of lac colorant.
LC conditions Column: Mobile phase: Flow rate: Column Temp: Injection vol: MS Conditions Source: Ion mode: Vcap voltage: Nebulizer: Drying gas flow: Drying gas temp: Scan range: Step size: Peak width: Time filter: Fragmentor:
250 x 2.1 mm Inertsil ODS3, 5 m 30% acetonitrile in 5 mM dibutylamine, isocratic 0.2 ml/min 40 C 10 l ESI Negative 4000 V 50 psig 10 l/min 350 C 100-1200 amu 0.1 0.15 min On 100 V
Figure 12. Mass spectra of the major pigments in lac colorant.
N a t u ra l Fo o d C o l o ra n t s Conclusion
Four commercial natural colorants were analyzed using ESI and APCI-LC/MS. The MS data provided molecular weight information and some structural information for the major pigments. Masahiko Takino is an applications chemist at Yokogawa Analytical Systems, Inc.
Windows NT is a U.S. registered trademark of Microsoft Corporation. Windows is a U.S. registered trademark of Microsoft Corporation. Agilent Technologies shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance or use of this material. Information, descriptions and specifications in this publication are subject to change without notice. Copyright 1998 Agilent Technologies, Inc. All rights reserved. Reproduction and adaptation is prohibited. Printed in the U.S.A. April 2000 (23) 5968-2979E
Analysis of Acidulants in White Wine using HPLC
Angelika Gratzfeld-Huesgen Food
Abstract Sorbic acid and citric acids are commonly used as acidulants1 and/or as preservatives. Acetic, propionic, succinic, adipic, lactic, fumaric, malic, tartaric, and phosphoric acids can serve as acidulants as well. Acidulants are used for various purposes in modern food processing. For example, citric acid adds a fresh, acidic flavor, whereas succinic acid gives food a more salty, bitter taste. In addition to rendering foods more palatable and stimulating, acidulants act as flavoring agents to intensify certain tastes and mask undesirable aftertastes buffering agents to control the pH during food processing and of the finished products preservatives to prevent growth of microorganisms synergists to antioxidants to prevent rancidity and browning viscosity modifiers in baked goods melting modifiers in cheese spreads and hard candy meat curing agents to enhance color and flavor
mAU 400 300 200

? ? 1
1 2 3 4 5 6
? 6 7 8
Oxalic acid Citric acid Tartaric acid Malic acid Sulfur-trioxide Succinic acid
7 8 9 10 11 12
Lactic acid Glycerol DEG Acetic acid Methanol Ethanol White wine
Conditions
Column 300 7.8 mm BioRad HPX 87-H, 9 m Mobile phase 0.0035M H2SO4 isocratic Flow rate 0.6 ml/min Column compartment 65 C Injection vol 10 l Detector UV-VWD detection wavelength 192 nm or 210 nm Sample preparation Filtration
100
2
3 4
? 9 10 11 12
0
0 5 10 15 Time [min] 20 25
Standard
Figure 1 Analysis of acidulants in white wine
Sample preparation Sample preparation depends strongly on the matrix to be analyzed, but in general steam distillation and solid-phase extraction techniques can be used. Chromatographic conditions High-performance liquid chromatography (HPLC) with UV-visible diode-array detection (UV-DAD) has been applied in the analysis of citric acid in wine and in a vodka mixed drink. Retention time and spectral data were used as identification tools. HPLC method performance Limit of detection 100ng injected amount, S/N = 2 equivalent to 2 ppm with 50 l injected volume Repeatability of RT over 10 runs <0.1% areas over 10 runs <3 %
Citric acid
Conditions
Sample preparation filtration Column 300 7.8 mm BioRad HPX 87-H, 9 m Mobile phase 0.007M H2SO4 isocratic Flow rate 0.6 ml/min Column compartment 65 C Injection vol 10 l Detector UV-DAD
Equipment
Agilent 1100 Series degasser isocratic pump autosampler thermostatted column compartment diode array detector, variable wavelength detector or refractive index Agilent ChemStation + software
20
mAU 100
Citric acid
Glucose Fructose
20
Sample spectrum overlaid with library spectrum match 994
0 190 Wavelength [nm] 276
Ethanol 0 0 5 10 Time [min] 15
Figure 2 Analysis of citric acid in vodka References 1. Official Methods of Analysis, Food Compositions; Additives, Natural Contaminants, 15th ed; AOAC: Arlington, VA, 1990, Vol. 2.; Official Method AOAC 986.13: quinic, malic, citric acid in cranberry juice cocktail and apple juice.
Angelika Gratzfeld-Huesgen is application chemist at Agilent Technologies, Waldbronn, Germany. For more information on our products and services, visit our worldwide website at http://www.agilent.com/chem
Analysis of Antioxidants in Chewing Gum using HPLC
Abstract The following compounds are used as antioxidants in food products:1 Natural antioxidants: vitamin C vitamin E Synthetic antioxidants: BHT butylated hydroxytoluene BHA butylated hydroxyanisole TBHQ mono-tert-butylhy-droquinone THBP 2,4,5-trihydroxybuty-rophenone PG propyl gallate OG octyl gallate DG dodecyl gallate Ionox-100 4-hydroxymethyl-2,6-di(tert-butyl) phenol NDGA nordihydroguaiaretic acid TDPA 3,3'-thiodipropionic acid ACP ascorbyl-palmitate
mAU 1500 1 2 3 4 5 6 7 8 3 4 56 10 Vitamin C PG THBP TBHQ BHA 4-hydroxy BHT Chewing gum extract ACP 7 12 8 Standard
Conditions
Column 100 4 mm BDS, 3 m Mobile phase A = water + 0.2 ml H2SO4, pH = 2.54 B = ACN Gradient start with 10% B; at 3 min 60% B at 4 min 80% B; at 11 min 90% B Flow rate 0.5 ml/min Post time 4 min Column compartment 30 C Injection vol 5 l Detector UV-DAD detection wavelength 260/40 nm, reference wavelength 600/100 nm Sample preparation Ultrasonic liquid extraction with acetonitrile (ACN)
1000 2 1 0 2 4
500
6 8 Time [min]
Figure 1 Analysis of antioxidants in chewing gum
Antioxidants may be naturally present in food, or they may be formed by processes such as smoking. Examples of natural antioxidants include tocopherols (vitamin E) and acsorbic acid (vitamin C). A second category of antioxidants comprises the wholly synthetic antioxidants. When these antioxidants are added to foodstuffs, they retard the onset of rancidity by preventing the oxidative degradation of lipids. In most countries where antioxidants are permitted either singly or as combinations in foodstuffs, maximum levels for these compounds have been set.2 Sample preparation Sample preparation depends strongly on the matrix to be analyzed. For samples low in fat, liquid extraction with ultrasonic bath stimulation can be used. For samples with more complex matrices, solid-phase extraction, liquid/liquid extraction, or steam distillation may be necessary. Chromatographic conditions HPLC and UV-visible diode-array detection have been applied in the analysis of antioxidants in chewing gum. Spectral information and retention times were used for identification. HPLC method performance Limit of detection 0.12 ng (injected amount), S/N = 2 Repeatability of RT over 10 runs <0.2 % areas over 10 runs <1 % References 1. Official Methods of Analysis, Food Compositions; Additives, Natural Contaminants, 15th ed; AOAC: Arlington, VA, 1990, Vol. 2.; AOAC Official Method 983.15: Antioxidants in oils and fats. 2. M. Rothaupt, Food Analysis, Introduction and Applications, Agilent Technologies Publication Number 5963-2317E, 1994.
Equipment
Analysis of aspartame using HPLC
Abstract The following compounds are used as artificial sweeteners in food products: acesulfam aspartame saccharin1 Nowadays, low-calorie sweeteners are widely used in foods and soft drinks. Investigations of the toxicity of these compounds have raised questions as to whether they are safe to consume. As a result, their concentration in foods and beverages is regulated through legislation in order to prevent excessive intake. Sample preparation Sample preparation depends strongly on the matrix to be analyzed. For sample low in fat, liquid extraction at low pH with ultrasonic bath stimulation can be used. For samples with more complex matrices, solid-phase extraction, liquid/liquid extraction, or steam distillation may be necessary.
mAU
scaled
Aspartame spectra
Conditions
Derivatization agent o-phthalaldehyde (OPA) mercapto-propionic acid (MPA) Column 100 2.1 mm Hypersil ODS, 5m Mobile phase A = 0.01 mM sodium acetate B = methanol Gradient start with 5% B; at 5 min 25% B at 10 min 35% B; at 13 min 55% B at 18 min 80% B; at 20 min 95% B Flow rate 0.35 ml/min Post time 5 min Column compartment 40 C Injection vol 1 l Injector program for online derivatization 1. Draw 5.0 l from vial 3 (borate buffer) 2. Draw 0.0 l from vial 0 (water) 3. Draw 1.0 l from vial 1 (OPA/MPA) 4. Draw 0.0 l from vial 0 (water) 5. Draw 1.0 l from sample 6. Mix 7 l (6 cycles) 7. Inject Detectors UV-DAD detection wavelength 338/20 nm or fluorescence: excitation wavelength 230 nm, emission wavelength 445 nm
60 50 40 30 20 10 0 0 2 4 6 Time [min] Apartame
original
derivatized
250
350 300 Wavelength [nm]
400
10
Figure 1 Determination of the quality of vanillin extract
Chromatographic conditions The HPLC method presented here for the analysis of aspartame is based on automated on-column derivatization and reversed-phase chromatography. UV spectra were evaluated as an additional identification tool.2 HPLC method performance Limit of detection for fluorescence 200 pg (injected amount), S/N = 2 for DAD 1 ng (injected amount), S/N = 2 Repeatability of RT over 10 runs <0.1 % of areas over 10 runs <5 % References 1. Official Methods of Analysis, Food Compositions; Additives, Natural Contaminants, 15th ed; AOAC: Arlington, VA, 1990, Vol. 2.; Official Method AOAC 979.08: Benzoate, caffeine, saccharin in soda beverages. 2. A.M. Di Pietra et al., HPLC analysis of aspartame and saccharin in pharmaceutical and dietary formulations; Chromatographia, 1990, 30, 215219.
Equipment
Agilent 1100 Series vacuum degasser quaternary pump autosampler thermostatted column compartment diode array detector, fluorescence detector Agilent ChemStation + software
Analysis of Preservatives in White Wine and Salad Dressing using HPLC

Abstract The following compounds are used as preservatives in food products: benzoic acid sorbic acid propionic acid methyl-, ethyl-, and propylesters of p-hydroxy benzoic acid (PHB-methyl, PHB-ethyl, and PHB-propyl, respectively).1 Preservatives inhibit microbial growth in foods and beverages. Various compound classes of preservatives are used, depending on the food product and the expected microorganism. PHBs are the most common preservatives in food products. In fruit juices, in addition to sulfur dioxide, sorbic and benzoic acid are used as preservatives, either individually or as a mixture.
Conditions
Absorbance (scaled) Spectral library 50 library match 999 30 10 sample 200 Wavelength [nm] 320
mAU 60 50
40 30 20 10 0
Standard White wine Salad dressing
Column 125 4 mm Hypersil BDS, 5 m Mobile phase A = water + 0.2 ml H2SO4, pH = 2.3 B = ACN Gradient start with 10% B at 3 min 60% B; at 4 min 80% B at 6 min 90% B; at 7 min 10% B Flow rate 2 ml/min Post time 1 min Column compartment 40 C Injection vol 2 l Detector UV-DADdetection wavelength 260/40 nm Sample preparation Carrez clearing and filtration for the salad dressing. None for white wine.
Sorbic acid PHB-methyl
Benzoic acid
PHB-propyl
PHB-ethyl
BHA
Time [min]
BHT
Sample preparation Sample preparation depends strongly on the matrix to be analyzed. For samples low in fat, liquid extraction with ultrasonic bath stimulation can be used. For samples with more complex matrices, solid-phase extraction, liquid/liquid extraction, or steam distillation may be necessary. Chromatographic conditions HPLC and UV-visible diode-array detection have been applied in the analysis of preservatives in white wine and salad dressing. Spectral information and retention times were used for identification. HPLC method performance Limit of detection 10 ppm, S/N = 2 Repeatability of RT over 10 runs <0.1 % areas over 10 runs <3 % References 1. Official Methods of Analysis, Food Compositions; Additives, Natural Contaminants, 15th ed; AOAC: Arlington, VA, 1990, Vol. 2.; AOAC Official Method 979.08: Benzoate, caffeine, saccharine in carbonated beverages.
Equipment
Sensitive Analysis of Synthetic Colors using HPLC and Diode- Array Detection at 190950 nm
Application Note
Food and Flavors Pharmaceutical Chemical Processing Angelika Gratzfeld-Hsgen Rainer Schuster
In this study synthetic dyes were analyzed using ion-pair, reversed phase chromatography on a special base-deactivated HPLC column. This separation mechanism was chosen to reduce tailing effects of highly polar colors. Detection was performed using a new design of diode-array detector based on two lamps a deuterium lamp and a tungsten lamp. This ensured highest light output at 190 950 nm, which resulted in lowest detection limits over the entire wavelength range. It was possible to analyze blue, black or green colors in the low ng range at their absorption maxima of 600 700 nm. Complete spectra, feasible from 190 950 nm, facilitated identification with an automated library search.
Introduction
Synthetic colors are widely used in the food, pharmaceutical and chemical industries. The regulation of colors and the need for quality control relating to traces of starting products and byproducts have forced the development of analytical methods. Nowadays various HPLC methods are in use, based either in ion-pair reversed phase chromatography or ion-exchange chromatography. UV absorption is the detection method of choice. The UV absorption maxima of colors are very characteristic, starting with maxima at about 400 nm for yellow colors, 500 nm for red colors and at 600 700 nm for green, blue and black colors. To analyze all different colors at maximum sensitivity and selectivity, the light output from the detector lamp should be high across the complete wavelength range. For black, green and blue colors, which show absorption maxima at or above 600 nm, this is not possible with conventional UV-Visible detectors based on a one-lamp design. For example, deuterium lamps have their maximum output in the UV range, whereas the visible range shows low light output. A further analytical problem is the tailing-free separation of dyes. Figure 1 shows that some colors are of more polar nature, which sometimes causes problems even on reversed phases depending on the strength of polarity tailing. This results in worse detection limits and integration problems.
In the following study we evaluated the influence of: special deactivated columns on the separation of colors using different mobile phases, a newly designed diode-array detection system on limit of detection for colors which absorb at above 500 nm, and low noise and complete spectra on identification of sample compounds using automated library search in the low mAU range.
Experimental
For the experiments the Agilent 1100 Series HPLC system was used. The system comprised a low-pressure quaternary gradient pump with vacuum degasser, autosampler, Peltier-regulated column compartment, and diodearray detector with a wavelength range of 190 950nm. System control and data evaluation was done through an Agilent ChemStation for HPLC.
NHCO CH 3 OH
Brilliant black
NaO3 S
N N
N N SO3 Na NaO 3 S SO 3Na
Quaternary pump
(C H3 )2 N
HO
SO3Na
Brilliant acid green

(CH 3) 2 N +
SO 3
Figure 1 Chemical structure of brilliant black and brilliant green BS
Agilent 1100 Series System Configuration
Solvent cabinet Vacuum degasser Agilent ChemStation for HPLC
Autosampler Column compartment Diode-array detector
Figure 2 HPLC system used

Separation of synthetic colors During method development a special base-deactivated column and four different mobile phases were evaluated. Two mobile phases were simple buffer systems and the other two phases contained ion-pairing reagents (see figure 3). A 125 x 3 mm Hypersil BDS, 3-m column (Agilent part number 79926BD-363) was used. The use of 3-mm id. and 3-m material allowed optimum flow
rates below 1 ml/min. This saved purchasing disposal costs of solvents. Mobile phases C and D showed tailing for compounds with sulfonic groups. Using ionpairing chromatography (see figure 3, examples A and B), the separation of colors with different functional groups and different chemical structure was achieved with minimum or no tailing for all different color classes. As a conclusion, we recommend using mobile phases from A.
Chromatographic conditions For the application examples that follow we used mobile phases from A. Column : Mobile phase A: 125 3 mm Hypersil BDS, 3 m (HP part number 79926BD-363) A = 0.01 M NaH2PO4 + 0.001 M Tetrabutylammoniumdihydrogenphosphate pH 4.2 B = acetonitrile (ACN) start with 15 %, in 10 min to 40 %, in 14 min to 90 %, until 19 min at 90 %, in 20 min to 15 % ACN A = 0.01 M NaH2PO4 + 0.001 M Tetrabutyl ammoniumhydrogen sulfate pH 4.8 B = acetonitrile (ACN) start with 15 %, in 10 min to 40 %, in 14 min to 90 %, until 19 min at 90 %, in 20 min to 15 % ACN A = 0.01 M ammoniumacetate pH = 4.9, B = ACN start with 7 %, in 10 min to 40 %, in 14 min to 90 %, until 19 min at 90 %, in 20 min to 7 % ACN A = 0.01 M NaH2P04, pH = 4.3, B = acetronitrile (ACN) start with 5 %, in 10 min to 60 %, in 14 min to 90 %, until 19 min at 90 % in 20 min to 5 % ACN 20 min 4 min 0.8 ml/min o 40 C 1 l signal A: 254 nm/50 nm (for optimization of separation) B: 350 nm/20nm, C: 465 nm/30 nm, D: 600 nm/40 nm E: 750/40
Gradient for A
Mobile phase B
mAU 1000
3
3 2 2 2 5
3 none
Gradient for B
800 600 400

2 200 1 1 4 4 3 5 5 41
5 6 4 1 6 6 6
Mobile phase C
0 2 4 6 8 10 12 14
A B C D
Gradient for C:
Time [min]
Mobile phase D Gradient for D
Figure 3 Separation of synthetic colors on BDS column using different mobile phases
Peak 1 2 3 4 5 6
Compound name E102 Tartrazine E123 Amaranth E124 Ponceau 4R E110 Sunset yellow E125 Scarlet red E127 Erythrosine
Number of sulfonic group 2 3 3 2 none
Stop time Post time Flow rate Col. temp. Inject. vol. Detector
Table 1 Content of sulfonic groups for colors shown in figure 3
Detection of synthetic colors using diode-array UV-visible absorption detector

The diode-array detector used here was equipped with two lamps, a deuterium lamp and a tungsten lamp. This ensured highest light output from 190 to 950 nm and therefore lowest detection limits over the entire wavelength range. The use of 1024 diodes and a programmable slit ensured highest spectral resolution. This gave the following advantages for the analysis of colors: acquisition of five signals simulataneously, highest sensitivity and selectivity even for blue, green and black colors with absorption maxima above 500 nm, complete spectral data up to 950 nm, and optimization of signal to noise ratio using different slit width without the need to exchange optical slits mechanically. Figure 4 shows the complete spectra of a yellow, red and two blue colors. For each of the analyzed colors characteristic absorption maxima were obtained. The yellow color tartrazine had its maxima at around 400 nm, the red color amaranth absorbed best at about 500 nm, the blue color patent blue had its maxima around 600 nm whereas the darker blue color brilliant blue showed its maxima at 740 nm. This clearly demonstrated that several signals had to be acquired for optimum sensitivity and selectivity for all colors.
The spectra of pure compounds can be stored in spectral libraries and used for later identification of colors in food, paints or pharmaceutical preparations. That sensitivity is of utmost importance is shown in figure 8. The quality of inks is often determined by the content of traces of other colors, which may be unwanted by-products from production process or which are added to influence the nuance of a color. Therefore the quantitation and identification of trace compounds is as important as the determination of main compounds.
Norm 40 35 30 25 20 15 10 5 0 300
Tartrazine
Amaranth
Patent blue
Brilliant blue
400
500
600 Wavelength [nm]
700
800
Figure 4 Spectra of different colors
As already demonstrated colors have very characteristic spectra which can be used to identify peaks not only by retention times but also by spectral data. The detector used here allowed identification using spectral data, see figure 5, even in the low mAU absorption range. The detection limit for the evaluated colors, measured at their specific absorption maximum, was in the low ng range. The repeatability of the HPLC method used here was measured using the standard mixture of figure 3. The relative standard deviation for retention times measured over 10 runs was below 0.2 %. For the areas the relative standard deviation was below 1 %. The linearity was evaluated using blue ink color measured at 600 nm. Linearity was given from the low ng up to the low g range.
mAU 120
100 80
Normalized to 10 mAU 10 8 6 4 Match 998.9
74.5ng = 120 mAU
60 40 20
2 0 300 500 700 nm
7.45ng = 10mAU 0 0 2 4 6 8 10 Time [min] 12 14 16 18
Figure 5 Identification of spectra in the low mAU absorption range for a blue ink color through overlay of trace level sample spectrum with library spectrum
EC E102 E103 E104 E105 E110
Name Tartrazine Chryosine Quinoline yellow Yellow Sunset yellow Sunset yellow Azorubine Amaranth FD&C Ponceau 4R Scarlet red Ponceau 6R Erythrosine Patent blue V Indigo carmine Acid brilliant green Black PN Ponceau SX
Food, Drug & Cosmetics FD&C Yellow No. 5
CI 19140 14270
FD&C Yellow No. 10
47005 13015
FCF FD&C Yellow No. 6 FD&C Orange No. 2 Carmoisine Red No. 2 Ponceau 4R Scarlet red
15985 15980 14720 16185 16255 14815 16290
Application Examples Food colors

Colors are vital constituents of foods and probably the first characteristic perceived by the human senses. Today synthetic dyes have widely replaced natural colors. Table 2 lists some of the most frequently used food colors.
E111 E122 E123 E124 E125 E126 E127 E131 E132 E142 E151
FD&C Red No. 3 FD&C Violet No. 1 FD&C Blue No. 2 FD&C Green
45430 42051 73015 No. 3 28440
FD&C Red No. 4
Table 2 List of commonly-used colorings in EC and US classifications with color index numbers (Cl
These dyes are used to supplement and enhance the natural colors destroyed during processing or storage, and substantially increase the appeal and acceptability of foodstuffs where no natural colors exist, for example, soft drinks or ice cream. The usage of synthetic colors is well regulated worldwide, but the regulations differ from one country to the next. To ensure compliance with regulatory requirements, the used colors have to be identified and qualified according to national directives. As an example of the determination of colors in foodstuffs we analyzed the synthetic colors used for a green carbonated drink a woodruff-ade. The sample was injected directly and the compounds were identified using a library search (figure 6). The green color was produced by a mixture of quinoline yellow and patent blue. The yellow color quinoline yellow split into four peaks showing an absorption maximum at 410 nm. The blue color patent blue had its maxima at 600 nm.
mAU
12 10 8 6 4 2 4 65 n m/ 30 n m
Chinolin yellow
Patent blue
6 0 0n m / 4 0n m
8 Time [min]
10
12
14
Figure 6 Chromatogram from the analysis of woodruff lemondade
mAU
20
llow
15
Tartrazine Amarath
10
Erythrosine 465 /3 0nm 5 0 -5 -10 2 4 6 Time [min] 8 10 3 5 0 /2 0 n m 600/40 nm
Colors for pharmaceutical preparations

The pharmaceutical industry uses colors, for example, for tablets, capsules and syrups. Here the intention is not only to improve the optical appearance but also to give more safety to the consumer. For example, different colors may help to avoid errors for a patient who has to take several medicaments.
Figure 7 Chromatogram from the analysis of a red capsule
Figure 7 shows the analysis of tablet capsules which were dissolved in water, filtered and injected directly. The red coloring comprised two red colors erythrosine and amaranth and a yellow color tartrazine.
Ink colors
The quality of ink colors relating to color brightness, stability against light and reproducibility of the same colors nuance for years, is determined by the accurate composition of different dyes in different concentrations. Here the right concentration of color traces is as important as the concentration of the main color compounds. In figure 8 the chromatograms of a blue and a black ink color are overlaid showing that both inks do not only differ in one main compound but also in some trace compounds. The blue color compound in the black ink is producing the main difference between black and blue.
Black Ink mAU blue 80 80 mAU SB37 Blue Ink BVL3
SO = solvent orange SR = solvent red BVL = basic violet BR = basic red SB = solvent blue
60
SO25 40 SB39 SR49 SR39 BR1 465 nm 0 600 nm
40 SB38 20 4 8 12 Time [min] 16 SB38
465 nm 600 nm
12 16 Time [min]
Figure 8 Comparison between a black and a blue ink
Conclusion
In this study we have demonstrated that the separation and detection of synthetic colors can be improved. Ion-pairing reversed phase chromatography on a special deactivated column allowed the separation of dyes of different polarity with no or only slight tailing. The UV-visible detection, especially from 400 to 950 nm, gained sensitivity by increasing the light output with a tungsten lamp in addition to a deuterium lamp.
Angelika Gratzfeld-Hsgen and Rainer Schuster are application chemists based at Agilent Technologies, Waldbronn Analytical Division, Germany.
For the latest information and services visit our world wide web site: http://www.agilent.com/chem
Copyright 1995 Agilent Technologies All Rights Reserved. Reproduction, adaptation or translation without prior written permission is prohibited, except as allowed under the copyright laws. Publication Number 5964-3559E
Analysis of Synthetic Dyes in Food Samples by Capillary Zone Electrophoresis

Application Note
Food Analysis Rainer Schuster Angelika Gratzfeld-Hsgen
Synthetic food dyes were separated by capillary zone electrophoresis using an alkaline phosphate buffer as background electrolyte. The precision had a relative standard deviation of less than 0.5 % for the run-to-run migration times and 2 % for the peak areas with buffer replenishment after each run. The detection limit for the individual dyes was about 1 ng using a 50-m internal diameter Agilent extended path length capillary. Compounds were detected at different wavelengths 215, 410, 520 and 598 nmand the identities of the individual dyes were confirmed using peak-purity routines and a UV-Visible spectral library.
Introduction
Color is a vital constituent of foods and probably the first characteristic perceived by the human senses. Almost all foods from raw agricultural commodities to finished productshave an associated color. Further, many tests have shown that color can organoleptically dominate the flavor of a food. Colors have been added to foods since ancient times and include chlorophylls, carotinoids, flavonoids and anthocyans extracted from different plants. Today synthetic dyes have widely replaced natural colors. These dyes are used to supplement and enhance the natural colors destroyed during processing or storage, and substantially increase the appeal and acceptability of foodstuffs where no natural colors exist, for example, soft drinks or ice cream.
But, synthetic dyes are also used to mask decay, to redye food, to mask aging effects or to disguise poor products. Colors permitted for food use can be divided in three categories: 1. synthetic dyesdescribed in this study, 2. natural colors (for example, caramel or beetroot) and naturally identical colors (canthaxanthine), and 3. inorganic pigments. Synthetic colors in food products are predominantly azo and triarylmethane dyes. These are mostly acidic or anionic dyes containing carboxylic acid, sulfonic acid or hydroxy groups, which form negatively-charged colored ions at basic pH ranges. Capillary zone electrophoresis (CZE) is an ideal tool because it can separate all the different functional groups in one analysis within a short run time. CZE
separates compounds based on charge and size. The range of color shades covered by the azo group includes red, orange, yellow, brown and blue-black. These dyes are prepared by coupling diazotized sulfanilic acid to a phenolic sulfonic acid moiety that often contains unwanted byproducts from corresponding impurities. Triarylmethane colors are distinguished by their brilliance of color and high tinc- torial strength, but have poor light stability. The United States (US) and the European Community (EC) are the two major geographical areas where color regulations are enforced. The lists of admitted colors are updated continuously because of suspected carcinogenicity. For example: amaranth (E123, FD&C Red No. 2)one of the most widelyused red colorwas delisted in the US in 1970, indigo carmine (E132, FD&C Blue No. 2) was delisted in 1980, tartrazine (E102, FD&C Yellow No. 5) was subjected to rigorous tests, and in 1990 the use of erythrosine (E127, FD&C Red No. 3) was discontinued. Figure 1 shows the chemical structures of some common synthetic dyes.
Azo Dyes
Triphenylmethane Dyes
OH NaO 3S N N
NaO 3S HO
SO3 Na SO3
NaO3S
Amaranth E123 (FD&C Red No. 2)
(C2 H5 )N HO NaO3S N N NaO3S SO3Na

Sunset yellow FCF E110 (FD & C Yellow No. 6) Patent blue E131
N+ (C 2 H5)
O SO2Na
C C C=C HN HN
Sulfonated Indigo
Indigo carmine E132 (FD & C Blue No. 2)
Figure 1 Chemical structures of some common synthetic dyes
In the EC colors are regulated by council directives on coloring matters. Of relevant interest are the proposals from 1985 and 1992 in foodstuff directives for international harmonization of colorant regulations. Table 1 shows some of the most widelyused synthetic dyes with their EC name (E-number), the US name (FD&C) and the color index number (CI). To ensure compliance with regulatory requirements, analytical methods are required to determine nature and concentration of a colorant in a food product. Paper chromatography,
thin-layer chromatography, and more recently high-performance liquid chromatography (HPLC) have helped the analyst in the examination of synthetic colors.1,2,3 Conventional HPLC separations use reversed-phase ion-pair chromatography with silica ODS or RP-2 material and a quaternary ammonium salt as the counterion.4,5 Several workers have adopted this approach for different applications using different sample-preparation techniques.6,7,8,9 For example, in the
wool-fiber method the sample is extracted in a tartaric acid solution and a wool fiber is used to absorb the colors of interest.8 After washing with water and methanol the colors are washed off with a diluted ammonia solution, evaporated to dryness, and dissolved in water or ethanol. The polyamide method involves extracting the sample with acetic acid and polyamide powder, whereby a chromatographic column is filled with polyamide and washed with water and methanol.9 The colors are extracted with an ammonia solution, evaporated to dryness, and dissolved in water or ethanol. Both wool-fiber and polyamide extraction are official methods in Germany and described in detail in 35 LMBG.10 In the US individual methods for colors in food are described in the handbook of official methods of analysis of AOAC.11 Ion pair extraction with tetrabutyl-ammonium phosphate and back extraction using perchlorate has been reported by Puttemans et al.12
EC E102 E103 E104 E105 E110 E111 E122 E123 E124 E125 E126 E127 E131 E132 E142 E151
Name Tartrazine Chryosine Quinoline yellow Yellow Sunset yellow FCF Sunset yellow Azorubine Amaranth Ponceau 4R Scarlet red Ponceau 6R Erythrosine Patent blue V Indigo carmine Acid brilliant green Black PN Ponceau SX
Food, Drug & Cosmetics FD&C Yellow No. 5
CI 19140 14270
FD&C Yellow No. 10
47005 13015
FD&C Yellow No. 6 FD&C Orange No. 2 Carmoisine FD&C Red No. 2 Ponceau 4R Scarlet red
15985 15980 14720 16185 16255 14815 16290
FD&C Red No. 3 FD&C Violet No. 1 FD&C Blue No. 2 FD&C Green No. 3
45430 42051 73015
28440 FD&C Red No. 4 14700
Table 1 List of commonly-used colorings in EC and US classifications with color index numbers (CI)
Experimental
Separations were performed using an Agilent CE system with a builtin diode-array detector and Agilent CE ChemStation software. Separations were achieved with fused-silica 50-m id capillaries (64.5 cm total length, 56 cm effective length) with an extended path length or bubble cell at the detector end. All separations were performed at 30C using a 10-mM sodium phosphate with 5-mM sodium hydrogen carbonate buffer at pH 10.5. Capillaries were preconditioned by flushing with 1M sodium hydroxide for 3 minutes followed by running buffer for 10 minutes. Samples were introduced hydrodynamically with 100 and 200 mbars followed by a 200 mbars buffer plug. The samples were analyzed with an applied voltage of 30 kV and
detected at 215/50 nm, 520/60 nm, 410/60 nm and 598/4 nm. After each run the inlet and outlet vials were replenished and the column was rinsed with the separation buffer for 1 minute. A voltage ramp from 0 to 30 kV within 0.5 minutes was performed to avoid possible thermal expansion and loss of sample. Details of the separation conditions are listed alongside Figure 2 which shows the separation of common synthetic dyes: patent blue E131, acid brilliant green E142, erythrosine E127, indigo carmine E132, chryosine E103, sunset yellow FCF E110, sunset yellow E111, scarlet red E125, quinoline yellow E104, azorubine E122, ponceau 4R E124, amaranth E123, black PN E151, tartrazine E102 and ponceau 6R E126. The E numbers correspond to the EC regulations for additives in food.
mAU 40
30
3
14 6 12 13 7 15
16
18 598 nm 17 520 nm
20
10
11 410 nm
0 4 2 4 6 8
9 10 215 nm
-10
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
E 131 E 142 E 127 E 132 E 105 Carminic acid E 103 E 110 E 111 E 125 E 104 E 122 E 124 Naphthol yellow E 123 E 151 E 102 E 126
Buffer
Electric field Capillary Injection Temperature Detection Signal
10-mM sodium borate with 5-mM sodium hydrogen carbonate at pH 10.5 465 V/cm I = 56 cm, L = 64.5 cm, id = 50 m 100 mbars 30 C 215/50 nm 520/60 nm red 410/60 nm yellow 598/4 nm blue off
Reference
8 Time [min]
10
12
Figure 2 Electropherogram of a standard sample

Classical methods still used for determination of synthetic dyes are thin-layer chromatography (TLC), paper chromatography, and HPLC together with diode-array detection (DAD) system. In HPLC separation is based on ion-exchange or reversed phase using ion-pairing reagents separation modes which require time (ion exchange) or chromatographic skill (ion pairing). Further HPLC can only analyze some of the existing colors. The use of CZE with a diode-array detector enabled us to separate most of the synthetic dyeslegal and illegal onesin one run and detect them at their individual wavelength: 410 nm for yellow, 520 nm for red, and 598 nm for blue.
A wavelength of 215 nm was used as a pilot wavelength for spectral acquisition and as a universal monitoring wavelength (for example, sweeteners and preservatives). The spectral information of the compound of interest was compared with the spectra in a librarycompounds were analyzed not only by migration time but also by spectral comparisonand a peak-purity check was performed by overlaying spectra taken in the peak. With the new software generation just one step was required to combine all these capabilities (migration time, spectral identification and peak purity) to produce quantitative reports based on three-dimensional data.
Table 2 shows a report based on the analysis of Figure 2 with the corresponding library search and relative standard deviation data (%RSD). Depending on the functional groupsmostly sulfonic acid groups as shown in Figure 1a pH in the range 811 was chosen for method development to separate colors as anions by CZE. Different buffers, such as borate, CAPS (3-cyclohexylalmino-1propane sulfonic acid) and phospate were tested at different pH values and concentrations. The optimum buffer was found to be 10 mM phospate at pH 10.5.
Library search mode: Automatic library search Signal 1: DAD1 A, Sig=215,50 Ref=off Meas. Library CalTbl %RSD (10 runs) 100 mbs 200 mbs 60-260ng/l 6-26ng/l Mt Area Mt Area -----|------|------|---|-------|------|-|------|------------------------------------------4.71 4.79 4.79 1 91.069 - 1 999 erythrosine E127 0.2 3.3 0.08 5.0 6.45 6.71 6.71 1 84.340 - 1 975 sunset yellow E110 0.6 2.4 0.05 2.9 6.93 7.19 7.19 1 138.535 - 1 998 scarlet red E125 0.5 2.7 0.06 3.3 7.71 8.00 8.00 1 69.244 - 1 1000 ponceau 4R E124 0.6 2.7 0.1 1.9 8.15 8.54 8.54 1 31.708 - 1 1000 amaranth E123 0.8 3.1 0.07 3.2 10.23 10.81 10.81 1 84.758 - 1 1000 tartrazine E102 1.2 3.7 0.1 5.9 Time Timet [min] [min] Time Sig [min] Amount Library ppm/100m Match Name
Table 2 Report and standard deviation on migration time and area for some color compounds
mAU
30 25 20 15 10 5 215 nm 410 nm 598 nm 2 4 6 8 10 12
Time [min]
Figure 3 Electropherogram of a carbonated drink containing colors and artificial sweeteners
However, the reliability data showed a clear trend to shorter migration times depending on the pH changes of the buffer even when replenishment of the two vials was used. This was mainly due to the instable pH of this buffer system. According to literature data we found at this pH sodium hydrogen carbonate was a stable alternativea buffer system which did not give a ideal separation of the colors.
Finally a mix of 10 mM phosphate and 5 mM sodium hydrogen carbonate adjusted to pH 10.5 with sodium hydroxide gave optimum conditions for separation of all colors, see figure 2. The buffer was filtered through a 0.45-m filter and its pH value had to be checked when used over several days. The migration order of the two colors azorubine E122 and quinoline yellow E104 was reversed using different pH values (for example, 10.510.2) whereas migration times of other compounds were more or less stable.
Application
The method was applied to different matrices: carbonated beverages, pasta, glac cherries, taramasalata extract and tablet capsules. The method was also tested to control quality of the colors themselves, for example, quinoline yellow. The carbonated drinka woodruff-adewas injected directly and the compounds were identified using a library search, see figure 3. The mint impression (green color) was produced by a mix of quinoline yellow E104 monitored at 410 nm with patent blue E131 monitored at 598 nm. Other compounds such as the sweeteners labeled on the bottle could also be quantified using a library search (acesulfam and saccharine). Figure 4 shows the electropherogram of a pasta extract containing quinoline yellow E104 with impurities (see figure 9), and sunset yellow FCF E110 to simulate the use of eggs. Both yellow compounds were selectively detected at 410 nm. Figure 5 shows the analysis of glac cherries containing ponceau 4R E124 which gives the intensive red color monitored at 520 nm.
mAU 30 25 20 15 10 5 410nm 0 0 2 4 6 Time [min] 8 10 12
215nm
Figure 4 Electropherogram of pasta extract using wool-fiber sample preparation method
mAU 50 40 Ponceau 4R
30 20 215 nm 10 520 nm 0 2 4 6 Time [min] 8 10 12
Figure 5 Electropherogram of glac cherries colored with E124
mAU 25 PHB-methyl 20 Amaranth 15 PHB-ethyl Erythrosine
10
Tartrazine 215 nm
5 520 nm 0 0 2 4 6 8 Time [min] 10 12 410 nm 14
Figure 6 shows the analysis of tablet capsules which were dissolved in water and injected directly. The red coloring of the capsule comprised two red colors erythrosine E127 and amaranth E123 (monitored at 520 nm), and a yellow color tartrazine E102 (monitored at 410 nm). The preservatives PHB-propyl, -ethyl and -methyl (215 nm) could be identified using a library search. It was found that under these conditions some of the sweeteners and preservatives could also be determined by this method.13 In addition to classical food applications this method could be applied to the quality control of the individual colors where the use of intermediates, containing impurities and coupling reactions, results in the formation of unwanted products.
Meas Library CalTbl Sig Amount Library Name Mt Mt Mt ppm/100m Match [min] [min] [min] bar*s Factor ------|-------|------|---|--------|-------|--------------4.99 4.79 4.79 1 11.100 998 erythrosin E127 8.89 8.54 8.54 1 21.133 980 amaranth E123 11.43 10.81 10.81 1 20.978 971 tartrazine E102 Uncalibrated compounds: Meas Library CalTbl Sig Amount Library Name Mt Mt Mt ppm/100m Match [min] [min] [min] bar*s Factor ------|-------|------|---|--------|-------|--------------4.14 4.13 1 983 PHB-propyl 4.29 4.25 1 1000 PHB-ethyl 4.61 4.41 1 1000 PHB-methyl
Figure 6 Electropherogram of a tablet capsule and corresponding report based on a library search
For example, figure 7 shows the analysis of quinoline yellow E104 which contained seven impurities. After spectral analysis two different types of yellow were identifiedtype I with a wavelength maximum at 422 nm and type II with a maximum at 387 nm. All other compounds could be associated with either of these two spectral types. This was also observed in all samples containing quinoline yellow E104, see figure 3. Similar impurities were found in the analysis of quinoline yellow FCF E110 (FD&C Yellow No. 6), tartrazine E102 (FD&C Yellow No. 5) and FD&C Red No. 40 (allura red is not used in Europe).11 In recent years concern has been expressed about the safety of certain synthetic dyes and this has prompted increased consideration for the use of natural colorants in food samples. Natural and naturally-identical colors are usually mixtures of several colored as well as noncolored compounds. Figure 8 shows the analysis of an aqueous beetroot extract with two red compounds (520 nm).
mAU 40 30 25 20 15 10 5 0 -5 2 4 6 8 Time [min] 10 12 I II II Quinoline yellow E104 I 80 40 0 Type II Type I
II
300 400 500 Wavelength [nm] 410 nm 350 nm 215 nm
Figure 7 Electropherogram of quinoline yellow E104
mAU 14 12 10 8 6 4 2 0 -2 0 2 4 6 Time [min] 8 10 12 215 nm 520 nm Ponceau 4R E124
Figure 8 Electropherogram of an aqueous beetroot extract
Area 100 80 60 40 20 5 3
Erythrosine E127 FD&C Red 3 1
Correlation Formula m b x y
0.99920 y = mx + b 6.1747e-1 1.06058 amount area
50
100 Amount [ppm/100 mbars]
150
Figure 9 Linearity curve for erythrosine E127
Reproducibility and Linearity

The migration-time precision calulated as %RSD for all compounds was better than 0.5 % and the peak-area precision was between 24 %. The calculation was based on ten runs with amounts of 60260 ng injected at 100 mbars and 626 ng injected at 200 mbars, see table 2. To achieve such excellent reproducibility, buffer replenishment after each run was necessary. The method was nearly linear over a range from low nanogram amounts to about 0.5 g with variable injection volumes, see figure 9.
Conclusion
We have shown that CZE is well suited for control of synthetic dyes in food samples and for some sweeteners and preservatives. The combination of this separation method with library searches and peak-purity checks enabled separation, identification and identity confirmation of the analytes in a single run. Diodearray detection enabled selective monitoring of individual color group at their appropriate wavelength. The detection limit for most of the compounds was found to be in the low nanogram range.
References 1 D. Pearson, J. Assoc. Public Anal., 11 (1973) 127 2 R. A. Hoodless, et.al., J.Chromatogr., 54 (1971) 393 3 N. P. Boley, et. al., Analyst (London), 105 (1980) 589 4 J. F. Lawrence, et.al., J.Chromatogr., 210 (1981) 168 5 M. L. Puttemans, et.al., J. Assoc. Off. Anal. Chem., 66 (1983) 670 6 W. J. Hurst, et. al. J. Assoc. Off. Anal. Chem., 64 (1981) 1411 7 M. H. Salagoity-Auguste, P. Sudraud, A. Bertrand, Sci. Aliments., 3 (1983) 127 8 J. Maslowsaka, K. Marszal, Dtsch. Lebnesm.-Rdsch., 77 (1981) 275 9 F. Kirchnawy, et al. Ernhrung, 9 (1985) 388
10 F. Kirchnawy, et. al. Ernhrung, 9 (1985) 38810 Amtliche Sammlung von Untersuchungsverfahren nach 35 Lebensmittel und Bedarfsgegenstndegesetz (LMBG) Methode L 08.0012 11 Official methods of analysis of the AOAC, 15th Edition, 1990, volume 2 12 M. Puttemans, et.al. Anal. Chim. Acta, 113 (1980) 307 13 R. Schuster, et.al. CZE Analysis of Artificial Sweeteners and Preservatives in Drinks Agilent Technologies Application Note, 1994, publication number 5963-1122E
Rainer Schuster and Angelika Gratzfeld-Hsgen are application chemists at Agilent Technologies, Waldbronn, Analytical Division, Germany.
CZE Analysis of Artificial Sweeteners and Preservatives in Drinks

Application Note
Food Analysis Rainer Schuster, Angelika Gratzfeld-Hsgen
Capillary electrophoresis can reduce the complexity associated with the analysis of sweeteners in drinks. Doing away with the need for a multitude of derivatization chemistries and their separations, the method described here has been successfully applied to both beverages and tablet formulations. A single run on an 50-m id Agilent Extended Path Length capillary at 192 nm, with simultaneous UV-visible absorbance spectral library and peak purity routines, detects and confirms most compounds in the low nanogram range. With buffer replenishment every five injections, repeatability is better than 0.15 % for migration times and approximately 2 % for areas.
Introduction
Sweetening agents can be classified as belonging to one of two main groups: caloric, or nutritive, and noncaloric or non-nutritive compounds. Nutritive sweeteners are carbohydrates (or their derivatives such as glucose, fructose and maltose) or products hydrolyzed from starch. Non-nutritive sweeteners do not belong to any particular chemical group. Synthetic sweeteners are steadily increasing in importance with increased public awareness of diabetes and its special dietary requirements, and more consumers becoming concerned about obesity and dental caries. The most frequently used synthetic sweeteners are: saccharin, cyclamate, aspartame
Saccharin Saccharin Aspartame Dulcin ODS
and acesulfame, see figure 1. To date, artificial sweeteners (table 1) have been determined by HPLC with reversed phase chromatography using different buffer systems, ion pairing reagents and specific derivatization procedures (aspartame with o-phthalaldehyde [OPA]; cyclamate with 4-fluoro-7nitrobenzofurazone [NBDF]). Derivatization overcomes detection limitations for these compounds in the low UV range. A number of methods have been published 1, 2 for simultaneous determination of aspartame and saccharin. Conditions are shown in table 1.1
Acesulfame-K Cyclamate NBDF-Derivative RP-18
Hermann and coworkers reported on a method for the detection of aspartame, cyclamate, dulcin, and saccharin using an ion-pair HPLC separation with indirect photometric detection. Toxicological data has led to the use of some artificial sweeteners being controlled, for example cyclamate is banned in the United States, the United Kingdom and Japan. Aspartame is metabolized to aspartic acid, methanol, and phenylalanine a substance critical to persons who suffer from phenylketonuria (PKU). Reliable means of obtaining analytical data are required for food samples containing these compounds. However, such varied methods with their differing derivatization protocols make the analysis of artificial sweeteners time consuming and labor intensive. An alternative to HPLC is capillary zone electrophoresis (CZE). All compounds can be separated sufficiently in one run.
Column Mobile phase
RP-18 0.05 mM H2 PO4 acetonitrile, 9:1
ODS
0.01 mM KH2 PO4 85:15 216 nm
0.01 mM TBAHS H2O, acetonitrile, 55:45
pH 3.5 -acetonitrile, methanol, 9:1 227 nm 490 nm Fluorescence ex485nm, em 530 nm
Detector
230/260 nm
Experimental
CZE separations were performed using the Agilent CE system with a built-in diode-array detector and Agilent CE ChemStation (DOS Series) software. Separations were achieved with fused-silica 50-m id capillaries (64.5 cm total length, 56 cm effective length) with an extended path length or bubble cell at the detector end. All separations were performed at 25C using a 20-mM sodium tetraborate buffer at pH 9.4. New capillaries were preconditioned by flushing with 1M sodium hydroxide for 3 minutes followed by running buffer for 10 minutes.
Table 1 HPLC conditions for determination of sweeteners
CH3
Acesulfame
O SO2 N H O
O H
Saccharine
SO2
H N
SO3H
Cyclamate
COOCH3 CH2 CH H NH CO CH NH2
Aspartame
CH2COOH
Figure 1 Chemical structure of the common artificial sweeteners
mAU 80 60 40
4 2 3 5 6 7 8
20 0 3.8
11 10 12
4.8
5.8
6.8
7.8
min
Figure 2 Electropherogram of a standard sample Meas. Library CalTbl Time Time Time Sig Amount Purity Library Name %RSD- (n=5) [min] [min] [min] [ng/l] Factor # Match MT AREA ------ |----- |------ |- |------- |------ |- |------ |----------------------------------------------------------4.21 4.10 4.20 1 177.884 1000 1 998 phenylalanine 0.02 9.4 4.78 4.67 4.78 1 87.063 1000 1 999 aspartame 0.006 1.1 5.00 4.90 4.99 1 62.506 1000 1 999 PHB-propyl 0.141 6.0 5.00 5.10 5.12 1 57.134 1000 1 998 ? PHB-ethyl 5.13 5.10 5.12 1 109.938 1000 1 999 PHB-ethyl 0.03 2.8 5.37 5.30 5.36 1 45.242 1000 1 999 PHB-methyl 0.03 1.4 5.95 5.85 5.94 1 50.572 1000 1 998 dehydroacetic acid 0.02 0.22 6.04 6.00 6.03 1 245.684 - 1 621 x cyclamate 0.03 7.0 6.29 6.20 6.29 2 69.561 1000 1 998 sorbic acid 0.026 2.3 6.91 6.90 6.89 1 45.588 1000 1 998 benzoic acid 0.03 1.9 6.99 6.86 6.96 1 157.647 1000 1 955 aspartic acid 0.045 0.9 7.22 7.20 7.20 1 53.055 1000 1 1000 saccharine 0.035 1.4 7.88 7.80 7.87 2 136.327 969 1 923 x acesulfame 0.04 1.7 Table 2 Report of figure 2 for artificial sweeteners and preservatives
Buffer 20mM borate pH 9.4 E 465 V/cm Effective 56 cm capillary length Total capillary 64.5 cm length id 50 m .s Injection 100 mbar Temperature 25C Detection Signal 192/2 nm Reference 450/100 nm Key 1 phenylalanine 7 cyclamate 2 aspartame 8 sorbic acid 3 PHB propyl 9 benzoic acid 4 PHB ethyl 10 aspartic acid 5 PHB methyl 11 saccharine 6 Dehydroacetic acid 12 acesulfame
be achieved by overlaying spectra taken in the peak. The systems software performs all three actions (migration time report, libraray search and peak purity) in one step, producing quantitative reports based on three-dimensional data. Table 2 shows a report based on the analysis of figure 2 with the corresponding library search and peak purity data. Concentrations injected were in the range 50200 ppm. The x flag in the report shows that acesulfame might contain an impurity: its library match factor of 923 and purity factor of 969 are lower than could be expected for a pure peak. Although flagged, cyclamate concentration is too low, even at 192 nm, to make any conclusive judgements. A spectral overlay of spectra taken over the peak migrating at 7.88 (all spectra in peak) show a rather interesting aspect, see figure 3. The overlay reveals an isosbestic point (245 nm) acesulfame exists in a tautomeric equilibrium in that buffer, stable at this pH value.
Samples were introduced hydrodynamically in 2 s at 50 mbar and analyzed with an applied voltage of 30 kV and detected at 191 nm (2-nm bandwidth). After each run the column was rinsed with the separation buffer for 2 minutes. Detailed separation conditions are listed alongside figure 2, the separation of common artificial sweeteners: aspartame (and its decomposition products phenylalanine and aspartic acid, cyclamate, saccharine and acesulfame together with the normally occuring preservatives PHB-esters (propyl-, ethyl-, and methyl), sorbic acid and benzoic acid.

Cyclamate and aspartame lack chromophores and require detection wavelengths in the low UVrange below 200 nm, that part of the spectrum where certain ion pairing reagents also absorb. Monitoring with the CE systems built-in diode-array detector permits detection at 192 nm and simultaneous acquisition of spectra. This spectral information compared to spectra in a library stored on the ChemStation can confirm that the response is indeed from the compound of interest and not from interfering matrix compounds. Peak purity analysis can
neutral Norm. 8.2 7.2 6.2 5.2 4.2 3.2 2.2 1.2 0.2 220 240 260 280 300 320 340 360 380nm * O CH3 O SO2 N CH3 OHH at pH 9.4 OO SO2 N basic
Buffer 20 mM borate pH 9.4 E 465 V/cm Effective 56 cm capillary length Total capillary 64.5 cm length id 50 m .s Injection 100 mbar Temperature 25 C Detection Signal Reference
192/2 nm 450/100 nm
Figure 3 Overlay of spectra taken from acesulfame
mAU 160 140 120 100 80 60 40 20 0 2
438 ppm aspartame
63 ppm benzoic acid
diet cola carbonated drink standard 4 6 8 min
192/2 nm 450/100 nm
Figure 4 Electropherogram of a diet cola containing aspartame and benzoic acid and a carbonated drink containing benzoic acid
mAU 35 30 25
mAU 120 74ppm benzoic 100 80 60 40 20 0 2 4 6 8 acid
library match 999
20 15 10 5 0
220 260 300 340 380 nm
carbonated drink standard min
192/2 nm 450/100 nm
Figure 5 Electropherogram of a carbonated drink with benzoic acid and spectral overlay
Application
The method has been applied to different matrices: beverages, such as diet cola and coffee, and tablets. All compounds have been identified with library search see figure 4.
figure 7. Buffer replenishment after five injections is necessary for highest reproducibility. CZE is known to have linearity characteristics half that of HPLC, nevertheless the equipment used here is linear up to 600 mAU.4
Reproducibility
The repeatability for all compounds was better than 0.15 % for retention time and between 1 and 7 % (9 % for phenylalanine due to an impurity) for peak area. The calculation was based on five runs with injected amounds of 50 to 250 ng absolute, see table 2 and
mAU 140 120 100 80 60 40 20 0 -20 4 5 6 7 Caffeic acids Cyclamate
Saccharine sweetener tablet
65ppm coffee sweetened with two tablets 8 9 min
192/2 nm 450/100 nm
Figure 6 Electropherogram of analysis of a sweetener tablet overlaid with the electropherogram of directly injected coffee sweetened with two such tablets mAU 100 80 4 60 40 20 0 2 4 6 8 min 2 3 5 6 7 8 1 Buffer 20 mM borate pH 9.4 E 465 V/cm Effective 56 cm capillary length Total capillary 64.5 cm length id 50 m .s Injection 100 mbar Temperature 25 C Detection Signal Reference
9 10
192/2 nm 450/100 nm
Figure 7 Overlay of five artificial sweetener standards
Key 1 phenylalanine 2 aspartame 3 PHB propyl 4 PHB ethyl 5 PHB methyl
6 7 8 9 10
cyclamide sorbic acid benzoic acid saccharine acesulfame
Conclusion
We have been able to show that capillary electrophoresis and UVvisible absorbance spectral library search is well suited for controlling food samples for artificial sweeteners and preservatives. Due to the transparency of the borate buffer detection could be done at 192 nm. Detection limit for most of the compounds is in the low nanogram range. Peak purity control for identification and confirmation of the compounds is possible in the same single run.
References
1 U. Zache, H. Gruending, Z. Lebensm. Unters. Forsch. 1987 184, 503. 2 J. F. Lawrence, C. F. Harbonneau, J. Assoc. Offic. Anal. Chem. 1988, 71, 934. 3 A. Hermann, E. Damawandi, and M. J. Wagmann, Chromatogr. 1983, 280, 85. 4 D. Heiger, P. Kaltenbach, H.-J. P. Sievert, Diode-array detection in capillary electrophoresis manuscript, submitted to Electrophoresis in April 1994, in publication.
Rainer Schuster and Angelika Gratzfeld-Hsgen are HPLC application chemists based at Agilent Technologies Waldbronn Analytical Division, Germany.
Amino Acids
High-Speed Amino Acid Analysis (AAA) on 1.8 m Reversed-Phase (RP) Columns Application
Pharmaceuticals and Foods
Authors
Cliff Woodward and John W. Henderson Jr. Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808 USA Todd Wielgos Baxter Healthcare Corp. Round Lake, IL 60073 USA
Introduction
The detection and quantitation of amino acids has been a large part of protein and food analysis since Moore and Stein developed the first analyzer in 1951 utilizing ion exchange chromatography to separate underivatized amino acids (AAs) followed by post-column derivatization with ninhydrin and detection in the visible region [5]. They were able to develop a completely automated analyzer, introduced commercially by Beckman in 1958. This was a revolution in the analysis of proteins and strongly contributed to the winning of the 1972 Nobel Prize in Chemistry for their work on ribonuclease. A single analysis had previously taken weeks to complete but was now accomplished in less than a day. Subsequent work reduced the analysis time to approximately 110 minutes on the last revision of the Beckman Amino Acid Analyzer produced, but the sensitivity was still a problem using ninhydrin. Post-column reaction with o-phthalaldehyde (OPA) and mercaptoethanol offered some promise, but only primary AAs were detected (see Figure 1a). In 1971 the first precolumn derivatization of AAs with OPA was published [6]. The derivatives were rather unstable but the process could be automated. Sensitivity was enhanced, particularly when fluorescence detection (FLD) was used. While the OPA is nonfluorescent, the derivatives are highly fluorescent; however, the secondary AAs are still not detected. Another derivatizing agent, 9-fluorenylmethyl chloroformate (FMOC), formed stable derivatives and derivatized both pri-
Abstract
Amino acid analysis (AAA) is commonly used in proteomics and food quality testing as a tool to discover the precise amino acid (AA) makeup of samples. In order to do this in a timely fashion a short turnaround time is needed; to do it with limited sample amounts, a highly sensitive method is needed; and to maximize laboratory productivity, an automated method is desirable. This paper describes a new methodology using recently developed columns with an engineered particle size of 1.8 m. These patented particles are specifically designed to create less backpressure than other sub-two-micron particles on the market; therefore, they can be used on 400 bar HPLC systems as well as higher pressure limit instruments. We take advantage of the OPA/FMOC chemistry first introduced by Hewlett-Packard/Agilent in 1988 and improve on the precision, run time, column longevity, and ruggedness of the previous methods [1%4].
mary and secondary AAs but was itself strongly UV absorbing and highly fluorescent (see Figure 1b) [7]. An extraction step or reaction with very hydrophobic amine at the end was needed to remove the excess FMOC and its reaction byproducts [8]. This latter methodology was also automated as a precolumn method [9] but had problems, in commercial versions, of having higher relative standard deviations (RSDs) than systems using no back extractions. By combining these two chemistries in a sequential way, in 1986 Hewlett-Packard/Agilent was able to fully automate the derivatization, chromatography, detection, and reporting of AAs from protein hydrolysates without back extraction [1]. A commercial analyzer was developed and sold by 1988 that had a total turnaround time of 36 minutes with femtomole sensitivity. This was definitely a step in the right direction for biotech companies and university researchers, who were sample limited, particularly in the early R&D phases of drug discovery.
a)
A few chemical innovations were needed in order to make this commercially viable: OPA derivatives were made more stable by changing the source of incorporated thiol to 3-mercaptopropionic acid (MPA); all primary AAs were reacted first with OPA to quantitatively remove them from further reaction; then FMOC was introduced to react only with the secondary AAs. Since FMOC, its derivatives, and reaction byproducts were all more hydrophobic than any of the OPA derivatives, they did not interfere with any primary AA detection. Because there were only a few FMOC derivatives, it was possible, using a simple two-segment gradient, to adequately separate those AAs from the reaction byproducts and FMOC [1]. This is summarized in Figures 1 and 2. The names corresponding to the peak numbers are given in Table 1.
OPA
O H H O
SR
+RNH 2
+ RSH Room Temperature
NR
Nonfluorescent, absorbs at 230 nm Does not absorb at 338 nm
Fluorescence: Ex 230 nm, Em 450 nm DAD: 338, 10 nm; Ref. 390, 20 nm Absorbs at 230 nm and 338 nm
b)
FMOC
RRNH + or RNH 2
O Cl O
Fluorescent Absorbs at 266 nm and Fluoresces at 305 nm
HCl
O O
Room Temperature
NRR or NHR
Fluorescence: Ex 266 nm, Em 305 nm DAD: 262, 16 nm; Ref. 324,8 nm
Figure 1.
o-Phthalaldehyde (OPA) and 9-Fluorenylmethyl chloroformate (FMOC) reactions with amines.
*
*Degradation peaks from reagents
12 3 5 1 4 2 7 6 8
FMOC and by-products

22 15
20 10 11 13 14 16 17 18 19 21
* *
23
Primary AAs (OPA) 1 2 3 4 5
Secondary AAs (FMOC) 6 min
Figure 2.
Amino acid analysis on Rapid Resolution HT Eclipse Plus C18, 4.6 x 50 mm, 1.8-m column: DAD 125 pMoles on column.
Table 1.
Peak # 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Names and Order of Elution for OPA and FMOC Derivatives of Amino Acids AA Derivative AA name abbreviation type Aspartic Acid Glutamic Acid Asparagine Serine Glutamine Histidine Glycine Threonine Arginine Alanine Tyrosine Cystine Valine Methionine Norvaline Tryptophan Phenylalanine Isoleucine Leucine Lysine Hydroxyproline Sarcosine Proline ASP GLU ASN SER GLN HIS GLY THR ARG ALA TYR CYS-CYS VAL MET NVA TRP PHE ILE LEU LYS HYP SAR PRO OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA FMOC FMOC FMOC
Experimental
Instrument The results obtained in this application were all performed on an Agilent 1200SL HPLC system consisting of the following components: G1312B, binary pump SL G1379B, micro degasser G1367C, high-performance well plate autosampler (WPS) configured with 54 x 2 mL sample tray in front and 15 x 6 mL tray in back G1316B, thermostatted column compartment SL (TCC) with low dispersion kit installed G1315C, diode array detector SL (DAD) with semimicro flow cell G1321A, fluorescence detector (FLD) All tubing used throughout the instrument is 0.12 mm id (0.005 inch). It is likely that older instruments, such as 1100 binary systems, and possibly 1100 and 1200 quaternary pump-based systems can run this analysis on 4.6 mm id columns, though this has not been tested. The use of 3.0 mm id and 2.1 mm id is not recommended on those systems without extensive testing due to the larger delay volume encountered using those systems; this may cause anomalies in some of the earlier eluting peaks due to dwell volume effects.
3
Internal Standard
Instrument Configuration Pump Parameters: mixer and pulse dampener bypassed with 100 mm long tubing (0.12 mm id), bypass programmed (at 0.1 min after inject command; see WPS parameters below), compressibility settings used: A = 35, B = 80 Flow: 0.42 mL/min for 2.1 mm id; 0.85 mL/min for 3.0 mm id; 2.00 mL/min for 4.6 mm id
Gradient Timetable: Time (min) 0.0 1.0 7.0 7.1 8.4 8.6 Stop time 8.7 %B 2.0 2.0 57.0 100.0 100.0 2.0
FLD: PW 0.01 min; Stop time 7 min (adjust as needed), never use this detector in series before another due to fragility of flow cell Ex 230 nm; Em 450 nm; Filter 295 nm (Default filter) Timetable Signal: 0.00 min Ex 230 nm, Em 450 nm; PMT Gain 9 (as needed) 5.53 min Ex 266 nm, Em 305 nm; PMT Gain 9 (as needed; wave length transition times as for DAD + approximately 0.03 min) TCC: used with low dispersion kit installed; T = 40 C for column side, 30 C for exit side. Low dispersion kit used on both sides. WPS: Default volume set to 0.5 l, default speed used throughout injector program is 200 l/min Injector program 1) Draw 2.5 l from Borate vial (Agilent P/N 5061-3339) 2) Draw 0.5 l from Sample vial 3) Mix 3 l in washport 5X 4) Wait 0.2 min 5) Draw 0.5 l from OPA vial (Agilent P/N 5061-3335) 6) Mix 3.5 l in washport 6X 7) Draw 0.4 l from FMOC vial (Agilent P/N 5061-3337) 8) Mix in 3.9 l in washport 10X 9) Draw 32 l from Injection Diluent vial 10) Mix 20 l in washport 8X 11) Inject 12) Wait 0.10 min 13) Valve bypass Amino Acid Mix for Calibration Curves For the construction of calibration tables and curves, 17 amino acids, plus the 4 extended amino acids, are combined at various concentrations with fixed amounts of internal standards. The internal standards (ISs) (norvaline and sarcosine) are part of the supplemental amino acid kit (P/N 50622478). The remaining amino acids in this kit (glutamine [GLN], asparagine [ASN], tryptophan [TRP], and hydroxyproline [HYP]) form the extended amino acids (EAA). To make the appropriate solutions, refer to Tables 2 and 3 for low- and high-sensitivity standards, respectively.
DAD: PW 0.01 min; slit 4 nM; Stop time 7 min (adjust as needed); Cell volume = 5 l, 6 mm flow path (Agilent P/N G1315-60025) Signal A: 338, 10 nm; Ref 390, 20 nm Signal B: 262, 16 nm; Ref 324, 8 nm Signal C: 338, 10 nm; Ref 390, 20 nm Signal D: 230, 16 nm; Ref 360, 100 nmM Timetable Signal C): 0.00 min 338, 10 nm; Ref 390, 20 nm; 5.53 min 262, 16 nm; Ref 324, 8 nm (adjust as needed; 3.0 mm id transition at approximately 5.45 min, 4.6 mm id transition at approximately 5.4 min)
Table 2. Preparation of Low-Sensitivity Amino Acid Standard Solutions (Prepare the three low-sensitivity standards by mixing together stock solutions in the volumes shown.) Concentration of final AA solutions (pMoles/L) 900 225 90 Take 5 mL 18 nMoles EAA Dilute with 0.1 N HCl Diluted EAA mix Take 5 mL diluted EAA mix Add 10 nMoles ISs solution EAA-ISs mix 5 mL 5 mL 5 mL 5 mL 10 mL 5 mL 15 mL 20 mL 5 mL 5 mL 10 mL 5 mL 45 mL 50 mL 5 mL 5 mL 10 mL
Take 100 L EAA-ISs mix 100 L Add 1000 pMoles AA standard 900 L Add 250 pMoles AA standard Add 100 pMoles AA standard Final AA solution with EAA and 500 pMoles/L ISs 1 mL
100 L 100 L 900 L 900 L 1 mL 1 mL
Table 3.
Preparation of High-Sensitivity Amino Acid Standard Solutions (Prepare the three high-sensitivity standards by mixing together stock solutions in the volumes shown.) Concentration of final AA solutions (pMoles/L) 90 22.5 9 5 mL 5 mL 5 mL 5 mL 10 mL 5 mL 15 mL 20 mL 5 mL 5 mL 10 mL 5 mL 45 mL 50 mL 5 mL 5 mL 10 mL
HPLC Columns ZORBAX Rapid Resolution HT Eclipse Plus C18, 2.1 x 50 mm, 1.8 m, P/N 959741-902 ZORBAX Rapid Resolution HT Eclipse Plus C18, 3.0 x 50 mm, 1.8 m, P/N 959941-302 ZORBAX Rapid Resolution HT Eclipse Plus C18, 4.6 x 50 mm, 1.8 m, P/N 959941-902 Mobile Phase and Injection Diluent Mobile phase A: 10 mM Na2HPO4: 10 mM Na2B4O7, pH 8.2: 0.5 mM NaN3 (5.6 gm anhydrous Na2HPO4 + 15.2 gm Na2B4O7 10H2O in 4 L water + 32 mg NaN3). Adjust to approximately pH 9 with 6 mL concentrated HCl and then small drops until pH 8.2. Be cautious with strong acids. Filter through 0.45 m regenerated cellulose membranes (Agilent P/N 3150-0576). Stable for approximately 1.5 weeks at room temperature. Mobile phase B: Acetonitrile methanol: water (45:45:10 by volume). All mobile-phase solvents must be HPLC grade. Injection Diluent: 100 ml mobile phase A + 1,500 L concentrated H3PO4 in a 100-mL bottle. Keep at 4 C. Dispense into 6-mL vials before use. Be cautious with strong acids. See Ordering Information for descriptions and part numbers of all supplies available for AAA. Derivatization Reagents Borate Buffer: Agilent P/N 5061-3339 Solution is 0.4 N in water, pH 10.2. Keep refrigerated (4 C). Dispense as necessary. FMOC Reagent: Agilent P/N 5061-3337 Pipette 200-L aliquots of the 1-mL FMOC reagent into conical inserts, cap immediately, and refrigerate (4 C); solution is useable for 7 to 10 days, maximum, after dispensing if kept at 4 C; useable for 1 day if kept at room temperature. OPA Reagent: Agilent P/N 5061-3335 Pipette 200-L aliquots of the 1-mL OPA reagent into conical inserts, cap immediately, and refrigerate (4 C); solution is useable for 7 to 10 days, maximum, after dispensing, if kept at 4 C; useable for 1 day if kept at room temperature. Water: Deionized water, HPLC Grade See Ordering Information for descriptions and part numbers of all supplies available for AAA.
Take 5 mL 1.8 nMoles EAA Dilute with 0.1 N HCl Diluted EAA mix Take 5 mL diluted EAA mix Add 1.0 nMoles ISs solution EAA-ISs mix
Take 100 L EAA-ISs mix 100 L Add 100 pMoles AA standard 900 L Add 25 pMoles AA standard Add 10 pMoles AA standard Final AA solution with EAA and 50 pMoles/L ISs 1 mL
100 L 100 L 900 L 900 L 1 mL 1 mL
Amino Acid Standards (10 pMoles/L to 1 nMoles/L): Divide each 1-mL ampoule of standards (P/N 5061-3330 through 5061-3334) into 100-L portions in conical vial inserts, cap, and refrigerate aliquots at 4 C. Calibration curves may be made using from two to five standards, depending on experimental need. Extended Amino Acid (EAA) Stock Solution: This solution is made using four of the six amino acids in the supplemental amino acid kit (P/N 5062-2478). For use with low-sensitivity standards (Table 2), make a 25-mL solution containing 18 nMoles/L of glutamine, asparagine, tryptophan, and 4-hydroxy-proline in deionized water. Sonicate the solution until dissolved. Store the solution at 4 C. For use with high-sensitivity standards (Table 3), make a 1.8 nMoles/L solution by diluting 5 mL of the 18 nMoles/L standard with 45 mL deionized H2O. ISs Stock Solution: These solutions are made using two of the six amino acids in the supplemental amino acid kit (P/N 5062-2478). For use with low-sensitivity standards (Table 2), make a 25-mL solution containing 10 nMoles/L of norvaline and sarcosine in deionized water. Sonicate the solution until dissolved. Store in refrigerator (4 C). For use with high-sensitivity standards (Table 3), make a 1 nMoles/L solution by diluting 5 mL of the 10 nMoles/L standard with 45 mL deionized H2O. Store at 4 C.
Sample Preparation The various bottled beer samples analyzed in this paper were obtained locally. The only preparation needed was degassing, which was accomplished quite simply by sonication. Caution must be used because sonication causes instantaneous, explosive degassing of carbonated beverages. It is recommended that the beer be poured into a wide-mouth beaker of greater volume than the beer, to allow for the large amount of foam generated.
50-mm length. The only differences in each method are the flow rate, so that a constant column flow velocity is maintained, and detector wavelength switching time, so that secondary AAs can be included on the same chromatogram. We have also run this separation on Rapid Resolution HT Eclipse Plus C8 columns and on Rapid Resolution HT Eclipse XDB-C18 and -C8 columns, packed with 1.8-m particles and the same variety of dimensions (data not shown). All columns gave very similar results to those obtained on the Rapid Resolution HT Eclipse Plus C18 (data not shown). The comparative separations obtained on different diameter Rapid Resolution HT Eclipse Plus C18 columns are shown in Figures 3 and 4 for DAD and FLD, respectively. As may be seen, there are only slight shifts in retention times for all chromatograms no matter what diameter is used. The shifts are caused by the slight changes in delay times among the three different diameter columns owing to the fact that the delay volume is constant; thus, the delay times are less for higher flow rates.
15 22 20 17 18 19 21 23

Separation Options In this application we demonstrate that the new method allows AAA to be done on a wider variety of columns than in previous versions. We report the results run on Rapid Resolution HT Eclipse Plus C18 columns, all of which were made with 1.8-m particles in 4.6-, 3.0-, or 2.1-mm id and
mAU 50 1 2
125 pMoles on column

3 4 5 6 78 9 10 11 12 13
14
16
40
4.6 mm id
30
3.0 mm id
20
2.1 mm id
10
0 0 1 2 3 4 5 6 7 min
Figure 3.
Scalability of AAA on various Rapid Resolution HT Eclipse Plus C18, 50 mm long, 1.8-m columns, DAD.
LU 800
50 pMoles on column
15 22
4.6 mm id
700 1 3 2 4 7 8 56
10
11
13
14
16
17
18
19 21 20 23
600
500
3.0 mm id
400
300
200
2.1 mm id
100
0 0 1 2 3 4 5 6 7 min
Figure 4.
Scalability of AAA on various Rapid Resolution HT Eclipse Plus C18, 50 mm long, 1.8-m columns, FLD. Table 4. Absolute Peak Area Reproducibility on Rapid Resolution HT Eclipse Plus C18, 2.1 x 50 mm, 1.8-m Columns Using an Agilent 1200SL HPLC System, no ISs
Reproducibility and Linearity The reproducibility of the secondary AAs in previous versions of this analysis was always less than that of the primaries [1%4]. With the current methodology we have improved the secondary AAs reproducibility to equality with the primary AAs. This data is shown in Tables 4 and 5 for ZORBAX Eclipse Plus-C18, 2.1-mm id columns, without ISs and in Table 6 for 4.6-mm id columns with ISs. The absolute peak area reproducibilities in Table 4 are superior. The average RSD is below 2% and only two are above 3% for all the AAs. Note that the linearities shown in Tables 5 and 6 are also excellent for all the AAs, being very close to 1.00 for all with a small edge to the linearities obtained with ISs. The results without ISs may be seen graphically in Figures 5 and 6 for Eclipse Plus-C18, 3.0-mm id columns. The chromatographic reproducibility of low-level AAA on the Eclipse Plus-C18, 2.1-mm id column may be seen in Figure 7 for 5.0 pMoles on column. The staggered overlay allows the retention time reproducibility and peak sizes to be easily compared. This is raw data with no corrections and no ISs.
Detector reproducibility 50 pMoles FLD; 125 pMoles DAD; raw data areas FLD RSD AA name Abbreviation (n = 10) Aspartic acid ASP 1.1 Glutamic acid GLU 0.8 Asparagine ASN 1 Serine SER 0.9 Glutamine GLN* 2.9* Histidine HIS 0.8 Glycine GLY 1 Threonine THR 1.2 Arginine ARG 1 Alanine ALA 0.9 Tyrosine TYR 1 Cystine CYS-CYS NA Valine VAL 1.5 Methionine MET 2.7 Norvaline NVA 3.3 Tryptophan TRP 1.3 Phenylalanine PHE 1 Isoleucine ILE 0.9 Leucine LEU 2.1 Lysine LYS 4.4 Hydroxyproline HPA 1.8 Sarcosine SAR 1.9 Proline PRO 3.0 Average RSD = 1.7 Asparagine is somewhat unstable in solution * Glutamine is very unstable in solution
DAD RSD (n = 12) 0.8 2.3 1.6 1 2.1* 2.1 1.4 0.9 1.9 .3 1 1.1 0.9 0.8 2.3 1.1 0.9 1.5 0.7 0.8 1.7 2.3 2.2 1.4
Table 5.
Linearity of AAA on Rapid Resolution HT Eclipse Plus C18, 2.1 x 50 mm, 1.8-m Columns Using an Agilent 1200SL HPLC System, no ISs Using raw data areas DAD FLD coefficients coefficients of linearity (r2) of linearity (r2) 0.9949 0.9976 0.9928 0.9933 0.9966* 0.9951 0.9931 0.9942 0.9960 0.9965 0.9936 0.9818 0.9952 0.9944 0.9989 0.9923 0.9925 0.9934 0.9944 0.9857 0.9980 0.9975 0.9981 0.9992 0.9995 1.0000 0.9997 0.9937* 0.9996 0.9996 0.9996 0.9998 0.9996 0.9997 NA 0.9998 0.9997 0.9991 0.9987 0.9996 0.9993 0.9994 0.9957 0.9924 0.9960 0.9969
Table 6.
Linearity of AAA on Rapid Resolution HT Eclipse Plus C18, 4.6 x 50 mm, 1.8-m Columns Using an Agilent 1200SL HPLC System with ISs DAD coefficients of linearity (r2) 0.99995 0.99916 1.00000 0.99982 0.99996* 0.99978 0.99982 0.99996 0.99993 0.99984 0.99991 0.99992 0.99992 0.99992 IS+ 0.99997 0.99988 0.99973 0.99986 0.99956 0.99991 IS+ 0.99998 FLD coefficients of linearity (r2) 0.99887 0.99879 0.99581 0.99880 0.99549* 0.99993 0.99681 0.99941 0.99946 0.99924 0.99946 NA 0.99886 0.99996 IS+ 0.99687 0.99943 0.99910 0.99932 0.99979 0.99751 IS+ 0.99979
Amino acid ASP GLU ASN SER GLN* HIS GLY THR ARG ALA TYR CYS-CYS VAL MET NVA TRP PHE ILE LEU LYS HYP SAR PRO
Amino acid ASP GLU ASN SER GLN* HIS GLY THR ARG ALA TYR CYS-CYS VAL MET NVA TRP PHE ILE LEU LYS HYP SAR PRO
Asparagine is somewhat unstable in solution. * Glutamine is very unstable in solution.
Asparagine is somewhat unstable in solution. * Glutamine is very unstable in solution. + IS at 250 pMoles on column for DAD and 25 pMoles on column for FLD
Extended Amino Acids (EAAs) and Internal Standards (ISs) There are several amino acids that are destroyed during acid hydrolysis of proteins (asparagine, glutamine and tryptophan) and one that is rarely seen anywhere but in structural proteins (hydroxyproline). Additionally, many analysts prefer to use ISs in their sample in an effort to improve accuracy and reproducibility. It is necessary to choose the ISs carefully in order to obtain the best results. There are two common ways these are used: 1. Addition to the sample before any sample manipulations at all (i.e., before acid hydrolysis). This corrects for all variations, providing, of course, that the ISs are stable to hydrolysis and have the same derivatization reactivity as the other components in the sample. In order to properly use this mode the standards should also undergo the hydrolysis process; but this
8
usually leads to more variation due to the different stabilities of various amino acids. An example is serine, which is destroyed progressively, dependent on hydrolysis time. This brings us to the most useful way of running with ISs. 2. Addition just before analysis, which corrects for volumetric pipetting errors and autosampler variations and also assumes that the reactivities of ISs and sample components are the same. This is the method used in this application. For the OPA/FMOC analysis demonstrated here, there are two ISs available: norvaline for primary amino acids and sarcosine for secondary amino acids. These AAs may be used for both methods of using ISs (unpublished data); but method 2 is definitely the preferred way. When the four amino acids listed above (EAAs) are included in the standards using the Agilent Amino Acids Supplement
1200
1000 ASP GLU 800 SER HIS Area Counts GLY 600 THR ARG ALA 400
200
0 0 20 40 60 pMoles on Column 80 100 120 140
Figure 5.
Amino acid linearities by FLD on Rapid Resolution HT Eclipse Plus C18, 3.0 x 50 mm, 1.8 m.
1400
1200
1000
TYR VAL MET PHE ILE LEU
800 Area Counts
600
LYS PRO
400
200
0 0 20 40 60 pMoles on Column 80 100 120 140
Figure 6.
Amino acid linearities by FLD on Rapid Resolution HT Eclipse Plus C18, 3.0 x 50 mm, 1.8 m.
kit (P/N 5062-2478), all are well separated in this new version of AAA. This has been shown in Figures 3 and 4 already on a Rapid Resolution HT Eclipse Plus C18, 2.1, 3.0, or 4.6 x 50 mm, 1.8-m columns. In these two figures the ISs, norvaline and sarcosine, are added at 100 pMoles on column for FLD chromatograms and at 250 pMoles on column for DAD simply to demonstrate where they appear in the chromatogram. The original process for the addition of the EAAs and ISs has been described previously [4]. This procedure is used to
LU 16 0 1 14 0 12 0 10 0 80 60 40 20 0 0 1 2 3 4 2 4 8 6 7 9 10 11
produce an IS-based calibration table for the quantitation needed for real-world samples. The process is described in the Experimental section and is shown in Tables 2 and 3. Real samples The result of a comparison of a variety of lager beers is shown in Figure 8. Each beer has been spiked with ISs at a level of 500 pMoles/L each so that a comparison of the relative AA content
13 18 19
14
17
23 20
7 min
Figure 7.
Reproducible AAA of 5.0 pMoles on column using a Rapid Resolution HT Eclipse Plus C18, 2.1 x 50 mm, 1.8 m, FLD, no ISs.
Column: Rapid Resolution HT Eclipse Plus C18, 4.6 mm 500 mm, 1.8 m
mAU 140 120 100 80 60 American beer brewed to German purity laws 40 1 20 0 0 1 2 3 4 5 6 min 2 900 pMoles/L AAA std 6 7 8 + 500 pMoles/L ISs 3 4 5 9 10 11 13 14 15 17 18 19 16 21 Typical American beer English beer brewed for U.S. market
German beer brewed for U.S. market
12 20
22 23
Figure 8. 10
Comparison of a variety of beers available in the USA using the 1200SL, DAD and ISs.
could be calculated. This comparison shows that the relative AA content of these beers is not consistent with many peoples perception+that typical American beers have lower AA content than do traditional German beers. The results indicate that at least some German and English beer manufacturers now brew their beers sold in the U.S. for American taste. It may be that the micro-brew from the U.S. is more consistent with a German beer brewed for German distribution, since it is brewed in compliance with German purity laws and there-
fore, more representative of traditional German brewing. The results are tabulated in Table 7. Acid hydrolysis and subsequent AAA are commonly used in the analysis of proteins. Cell culture media are also monitored, both in the laboratory and in the new Process Analytical Technology (PAT) initiative now being implemented throughout biotechnology and pharmaceutical companies. AAA on these subjects will be presented in subsequent papers.
Table 7.
Comparison of the Amino Acid Content of Various Beers Typical American beer Moles /mL 1.06 1.07 1.19 0.45 0.67 0.68 1.54 0.30 0.48 4.28 1.36 0.08 1.92 0.17 IS 0.58 1.33 0.54 1.02 0.30 0.20 IS 8.65 American beer brewed to German purity laws Moles /mL 0.73 0.93 0.98 0.40 0.76 1.00 2.00 0.34 0.94 5.14 2.86 0.10 3.35 0.27 IS 1.07 2.21 0.97 1.71 0.32 1.74 IS 34.84 German beer brewed for U.S. market Moles /mL 0.26 0.80 0.14 0.12 0.38 1.07 1.61 0.74 1.29 4.56 1.65 0.08 1.93 0.11 IS 0.80 1.23 0.34 0.55 0.08 1.17 IS 15.88 English beer brewed for U.S. market Moles /mL 1.15 1.98 0.54 0.26 0.33 0.89 1.57 0.20 1.83 4.60 1.77 0.05 2.83 0.16 IS 0.67 2.07 0.94 1.75 0.90 1.26 IS 10.29
Amino acid ASP GLU ASN SER GLN HIS GLY THR ARG ALA TYR CYS-CYS VAL MET NVA TRP PHE ILE LEU LYS HYP SAR PRO
Total =
27.9
62.7
34.8
36.0
11
Conclusions
All aspects of the AAA have been improved by this work. The use of 1.8-m sized particle columns has enabled the cycle time of the analysis to be more than cut in half, from approximately 35 minutes to approximately 13.5 minutes. The peak shapes of the early eluting aspartic acid and glutamic acid have been improved. The reproducibility of the secondary AAs has been improved, as has that of the slowest AA to react, lysine. The linearity is excellent (nearly 1); and the average peak area reproducibility is below 2%. Use of the new Rapid Resolution HT Eclipse Plus C18, 1.8-m columns of any diameter from 2.1 to 4.6 mm has been demonstrated. The new mobile phase and injector conditions have made it very simple to convert from 2.1- to 3.0- or 4.6-mm id columns by simply changing flow rate. It is also possible to use Rapid Resolution HT Eclipse Plus C8 or Rapid Resolution HT Eclipse XDB-C18 or -C8 columns as well (data not shown).
References
1. Rainer Schuster and Alex Apfel, HewlettPackard App. Note, 5954-6257 (1986) 2. Rainer Schuster, J. Chromatogr., 431, 271-284 (1988) 3. Herbert Godel, Petra Seitz, and Martin Verhoef, LC-GC International, 5(2), 44-49 (1992) 4. John W. Henderson, Robert D. Ricker, Brian A. Bidlingmeyer, and Cliff Woodward, Agilent App. Note, 5980-1193E (2000) 5. S. Moore and W.H. Stein, J. Biol. Chem., 192, 663 (1951) 6. M. Roth, Anal Chem, 43, 880-882 (1971) 7. S.B. Einarsson, B. Josefsson, and S. Lagerkvist, J. Chromatogr., 282, 609-618 (1983) 8. I. Betnr and P. Fldi, LC-GC International, 2(3), 44-53 (1989) 9. B. Gustavsson and I. Betnr, J. Chromatogr., 507, 67-77 (1990)

12
Ordering Information
ZORBAX Rapid Resolution HT Eclipse Plus C18 HPLC Columns
Description Eclipse Plus C18 Eclipse Plus C18 Eclipse Plus C18 Derivatization Reagents Description
Size (mm) 4.6 x 50 mm 3.0 x 50 mm 2.1 x 50 mm
Particle Size (m) 1.8 m 1.8 m 1.8 m
Agilent Part No. 959941-902 959941-302 959741-902
Agilent Part No. 5061-3339 5061-3337 5061-3335 5062-2479
Borate Buffer: 0.4 M in water, pH 10.2, 100 mL FMOC Reagent, 2.5 mg/mL in ACN, 10 x 1 mL ampoules OPA Reagent, 10 mg/mL in 0.4 M borate buffer and 3-mercaptoproprionic acid, 6 x 1 mL ampoules DTDPA Reagent for analysis of cysteine, 5 g Mobile Phase and Injection Diluent Components Description Na2HPO4, Sodium Phosphate, Dibasic, Anhydrous Na2B4O7.10H2O, Sodium Tetraborate Decahydrate NaN3, Sodium Azide H3PO4, ortho Phosphoric Acid Vials Description 100-L conical insert with polymer feet, 100/pk Amber, wide-opening, write-on, screw-cap vial, 2 mL, 100/pk Blue polypropylene cap, PTFE/silicone septum, 100/pk Clear glass screw-cap vial, 6 mL, 16 mm cap size, 100/pk Screw caps, 16 mm, 100/pk PTFE/silicone septa, 16 mm, 100/pk Standards Description Amino Acid Standards in 0.1 M HCl, 10 x 1 mL ampoules 1 nMoles/L 250 pMoles/L 100 pMoles/L 25 pMoles/L 10 pMoles/L Supplemental Amino Acids: Nva, Sar, Asn, GIn, Trp, Hyp, 1 g each
Manufacturer Sigma Sigma Sigma Sigma
Manufacturer Part No. 71639 S 9640 S 2002 79617
Agilent Part No. 5181-1270 5182-0716 5182-0721 9301-1377 9301-1379 9301-1378
Agilent Part No. 5061-3330 5061-3331 5061-3332 5061-3333 5061-3334 5062-2478
13
Measurement of Macro and Trace Elements in Plant Digests Using the 7500c ICP-MS System Application
Food
Author
Kazuo Yamanaka Agilent Technologies Nakacho 1-15-5 Musashino-shi Tokyo 180-8543 Japan Fred Fryer Agilent Technologies 12/2 Eden Park Drive, North Ryde NSW 2113 Australia
grains such as wheat and rice is resulting in a nutritionally poor diet. Micronutrient [1] malnutrition is an identified problem that has coincided with the rapid adoption of modern cereal cropping systems. Profitable and sustainable agriculture depends on the understanding of the nutrients required and available for plant growth, as well as the nutrients for a balanced human diet. World food production will need to double over the next 30 years to keep pace with increasing demands from both industrialized and rapidly developing countries. As well as the need to increase production, there will be an increase in demand for higher quality and healthier food products as developing countries become more affluent. Taken from the Commonwealth Scientific and Industrial Research Organisation (CSIRO) website: http://www.csiro.gov.au (select: Agribusiness/Field Crops/Field Crops & Australia) Human dietary micronutrients are required by humans in very small amounts. They include at least 14 trace elements (As, B, Cr, Cu, F, I, Fe, Mn, Mo, Ni, Se, Si, V, Zn) as well as 13 vitamins (thiamin, riboflavin, niacin, pantothenic acid, biotin, folic acid, vitamins B6, B12, C, A, D, E, K) The recommended daily intake of the micronutrient trace elements is of the order of: mg per day for B, Cu, F, Fe, Mn, Zn g per day for As, Cr, I, Mo, Ni, Se, Si, V
Abstract
Inductively coupled plasma mass spectrometry is a powerful tool for the investigation of many materials. The Agilent 7500c with Octopole Reaction System was used to analyze major, minor and trace elements in two standard reference plant materials. The data obtained using the 7500c is compared to the certificate reference values and to results that were generated using inductively coupled plasma optical emission spectroscopy. Results for all elements obtained using the 7500c agree with the certified values.
Introduction
The reliable measurement of trace elements in food is becoming more important as information is revealing that over-dependence on processed
Accurate determination of these trace elements in food materials is useful in ensuring that dietary intake is providing adequate levels of micronutrient elements. Due to the very low concentrations that must be measured and, in many cases, the high and variable sample matrix in which the measurements must be made, this analysis has proved challenging for elemental analysis instrumentation. Traditionally, a combination of techniques was required for a complete analysis of the plant digesttypically Graphite Furnace Atomic Absorption Spectroscopy (GFAAS), Hydride-Atomic Absorption Spectroscopy (HG-AAS) and Inductively Coupled Plasma Optical Emission Spectroscopy (ICP-OES). Such is the performance and elemental coverage of modern inductively coupled plasma mass spectrometry (ICP-MS) instrumentation, in many cases (metals analysis in drinking water, for example) a single ICP-MS has replaced all of the above mentioned techniques, enabling all analytes to be determined in a single measurement. The analysis of plant and food digests for nutritional studies is more challenging. In ICP-MS, isobaric interferences arise from the argon used to sustain the plasma and from the reagents used for sample preparation. Table 1 summarizes some well-known interfering species. In biological sample analysis, there are well-documented interferences for ICP-MS that can bias the measurement of Fe, Cr, V, As and Se at trace levels, with the result that ICP-MS has not yet been widely adopted by the foods industry.
Table 1. Examples of Potential Interferences in Biological/Clinical Matrices Element Cr V Fe Cu As Se Mass 52; 53 51 56 63 75 77; 78; 80 Molecular interference 40 Ar12C, 36Ar16O, 35Cl16O1H; 37Cl16O 35 16 Cl O 40 Ar16O 40 Ar23Na 40 Ar35Cl 40 Ar37Cl; 40Ar38Ar; 40Ar40Ar;
digestion with hydrogen peroxide and nitric acid. This digestion media does not generate additional interferences for ICP-MS and is a complete digest. However, for high sample numbers, the traditional hot plate digest offers higher sample throughput than closed vessel microwave digestion [2]. Recently, the advent of collision/reaction cells has improved the detection capability of quadrupole ICP-MS (ICP-QMS) by removing spectral interferences on analytes such as Fe, Cr, V, As and Se. The Agilent 7500c ICP-MS features an Octopole Reaction System (ORS) for highly efficient removal of multiple interferences arising from complex sample matrices. The ORS removes interferences by either reacting a gas with the interference or by preventing the interfering species from entering the analyzer stage using a process called energy discrimination. The 7500c exhibits highly efficient interference removal. The Ar2 overlap on Se at mass 80 is virtually eliminated, reducing the background equivalent concentration from 100s of ppb to <10 ppt. Moreover, the 7500c was designed specifically to handle complex matrices such as plant and food digests. The key to the successful multi-element determination of trace elements in complex samples is a combination of matrix tolerance and efficient interference removal. Matrix tolerance is mainly determined by the plasma efficiency, which must be optimized to ensure efficient sample decomposition, and is monitored by the CeO/Ce ratio. An efficient plasma minimizes the formation of plasmaand matrix-based interferences, while maximizing the conversion of analyte atoms into ions. The importance of matrix tolerance of any ICP-MS system should not be underestimated, as this leads to improved analytical accuracy, better tolerance to matrix changes and reduced requirements to carry out routine maintenance of the vacuum, ion lens and pump components. All of these aspects contribute to the usability of the analytical instrument, as routine maintenance contributes far more to the down-time of a modern, reliable ICP-MS instrument than hardware breakdowns. The unique capability of the Agilent 7500 Series lies in the mode of operation of the plasma source, which decomposes sample matrices five to 10 times more efficiently than is typical for other ICP-MS instruments.
One obvious way to remove interferences is to eliminate the source of the interfering species. Traditionally plant materials are digested on a hot plate using a mixture of nitric and perchloric acids. Chloride-based mass spectral interferences are introduced by this method. An alternative sample preparation method is available using microwave
The 7500c was designed specifically to handle complex, high matrix samples. A robust 27.12-MHz plasma, low sample uptake rate, cooled spray chamber and proven small orifice interface protect the ORS from contamination by undissociated sample matrix. A novel ion optic, mounted outside the high vacuum region for easy access, further protects the reaction cell, which features an octopole for optimum ion transmission. The octopole is mounted off-axis to minimize random background levels. A schematic of the 7500c is shown in Figure 1. Some of the important instrument parameters that contribute to good matrix decomposition are: The standard low sample flow rate (100 to 400 L/min) and Peltier-cooled spray chamber reduce the sample and water vapor loading on the plasma, which leads to a hotter plasma central channel. The 7500 Series uses a high efficiency, solid state 27.12-MHz plasma RF generator, ensuring good energy transfer into the plasma central channel. The unique wide internal diameter plasma torch design ensures that the sample aerosol is resident in the plasma for sufficient time to ensure complete matrix decomposition, leading to exceptionally good matrix decomposition (low CeO/Ce ratio).
The optimized interface design, which uses the smallest skimmer cone orifice of any commercial ICP-MS instrument, ensures that minimal sample matrix is passed into the high-vacuum part of the instrument, dramatically reducing the requirement for routine maintenance of the interface cones, the ion lenses and the collision cell. In summary, as the complexity of the sample matrix increases, the benefit of minimized interference levels becomes more significant. Because modern analytical laboratories rarely have the luxury of pre-analyzing samples to identify the matrix, it is impractical to rely on matrix matching of the samples or data correction using complicated interference equations.
Sample Preparation and Analysis

About 800 mg of sample was accurately weighed and carefully heated with 10 mL nitric acid (70%), followed by gentle heating with the addition of 8 mL perchloric acid (70%) until colorless. After cooling, 30 mL water was added and heating resumed for 10 min. Finally, the solutions were cooled, then made to 100 mL volume with water.
Reaction gas inlet Quadrupole Octopole reaction cell
Octopole Off-axis lens
Figure 1: Schematic diagram of the Agilent 7500c Octopole Reaction System.
The instrument was tuned and optimized as detailed in Table 2. Calibrations were performed using external standards prepared from 1000 ppm single element stock, made up as appropriate with 2% nitric acid.
Table 2. Agilent 7500c Operating Conditions Plasma RF power Sample depth Carrier gas flow Spray chamber temperature Sample flow rate Nebulizer Interface 1500 W 9.5 mm from load coil 1.1 L/min 2 C 240 L/min Agilent microflow (PFA) Nickel sample and skimmer cones
Table 4. NIST 1573a (Tomato Leaves, Blank Corrected)

Name 43 Ca 39 K 52 Cr 53 Cr 54 Fe 56 Fe 63 Cu 65 Cu 75 As 78 Se 111 Cd Certified (mg/kg) 5.05% 2.70% 1.99 1.99 368 368 4.7 4.7 0.112 0.054 1.52 ICPOES (mg/kg) 5.00% 2.72% 1.7, 1.8 1.7, 1.8 342, 347 342, 347 2.49, 2.40 2.49, 2.40 5.7, 6.6 0.1, 0.8 5.5, 5.9 7500c (mg/kg) 5.08% 2.62% 1.60 1.63 368 368 4.43 4.47 0.175 0.061 1.32
Table 5. NIST 1570a (Spinach, Blank Corrected)
The external calibrations were run in the same analytical sequence as the samples. Sample concentration was calculated using the internal standard method. Table 3 summarizes the element and relevant internal standard information.
Name 39 K 43 Ca 52 Cr 53 Cr 54 Fe 56 Fe 63 Cu 65 Cu 75 As 78 Se 111 Cd 54 Fe 56 Fe 63 Cu 65 Cu 75 As 78 Se 111 Cd
Certified (mg/kg) 2.90% 1.53% 12.20 12.20 0.07 0.12 2.89 12.20 12.20 0.07 0.12 2.89
Table 3. Reaction Gases and Internal Standards Used Measured element Potassium Calcium Chromium Iron Copper Zinc Arsenic Selenium Cadmium Reaction gas Helium Helium Helium Helium Helium Helium Helium Hydrogen Hydrogen Internal standard Scandium Scandium Gallium Gallium Cobalt Cobalt Yttrium Indium (115) Indium (115)

The practical effect of the 7500cs unique combination of matrix tolerance and interference removal is that complex and variable samples can be measured with a simple quantification procedure using external standard calibration and internal standard correction for all masses. As and Se were accurately quantified at sub-ppb levels, even in a matrix containing 8% perchloric acid. Tables 4 and 5 summarize the results obtained in a blind analysis of plant digests using the 7500c, comparing the results with both the certified values and data obtained from analysis by ICP-OES.
Reference 2: (mg/kg) 2.63% 1.32% 252 252 11.6 11.6 252 252 11.6 11.6 -
7500c (mg/kg) 2.56% 1.39% 1.24 1.29 248 250 10.48 10.51 0.062 0.09 2.33 248 250 10.48 10.51 0.062 0.09 2.33
Measurements of Cr, Fe and Cu were made on two separate isotopes for each element. Because molecular interferences will, in many cases, only affect one of the analyte isotopes, the presence of an interference can cause a large discrepancy between results for different isotopes of the same element. An example of this is the measurement of Cu in a high Na matrix, where 40Ar23 Na gives an overlap on 63Cu, but no interference on 65Cu. As the results indicate, the 7500c obtained excellent agreement for all the pairs of isotopes, highlighting the capabilities of the ORS in reducing interfering molecular species that, until now, have prevented the accurate trace analysis of transition metals in complex matrices by ICP-QMS.
Values for major and trace element concentrations agreed both with the expected value and the results obtained from ICP-OES. In the cases where the trace values for some elements were below the detection limit of the ICP-OES, the 7500c returned results in excellent agreement with the certified value. This data illustrates the wide dynamic range of the system and demonstrates its advantages as a replacement for traditional techniques such as ICP-OES. The quantitative analysis of the NIST SRM samples also demonstrates that both the 7500c and the operating conditions are robust and tolerant of the changing matrix composition found in plant digests.
Conclusions
The trace analysis of plant digests is an application that can be suitably addressed by the 7500c. Advances in technology now allow the determination of multiple elements in complex sample matrices, with efficient interference removal and, in the case of the 7500c, with the excellent matrix tolerance for which the 7500 Series is renowned. Accurate quantification of As and Se at low and even sub-ppb levels in plant digests is possible, even where high concentrations of perchloric acid have been added during the sample preparation stage.
Acknowledgement
The ICP-OES measurements and NIST sample preparation were performed at the University of Queensland, School of Land and Food Sciences, Australia.
References
1. Ross M. Welch; USDA-ARS, US plant and soil and nutrition laboratory, Towner Road, Ithaca, NY 14853, USA, Micronutrients, Agriculture and Nutrition: Linkages for Improved Health and Wellbeing 2. Da-Hai Sun; Waters, J. K.; Mawhinney, T. P. Determination of thirteen common elements in food samples by inductively coupled plasma atomic emission spectrometry: comparison of five digestion methods.; Journal of AOAC International 2000, 83 (5) 1218-1224

5
Rapid Screening of Amino Acids in Food by CE-ESI-MS

Tomoyoshi Soga and Maria Serwe Food Industry Abstract In the food industry, amino acids are measured to correlate flavor trends, monitor fermentation and assess the quality of the final product. Although amino acids are most commonly analyzed by HPLC with pre- or post column derivatization, capillary electrophoresis coupled to electrospraymass spectrometry (CE-ESI-MS) is ideally suited for the rapid screening of free amino acids at low mg/L levels in complex matrices.
Conditions
Sample soy sauce, diluted 1:100 and ultrafiltrated Injection 3 sec @ 50 mbar Capillary bare fused silica L= 100 cm, 50 m id Buffer 1 M formic acid Voltage 30 kV Temperature 20C Preconditioning 4 min flush with run buffer at 1 bar Sheath liquid 5 mM ammonium acetate in 50 % aqueous methanol, 10 L/min Nebulizing gas nitrogen, 10 psi Drying gas nitrogen, 10 L/min, 300 C Acquisition positive mode, Vcap -4 kV, fragmentor 70 V Scan range 50350 m/z
m/z 76, GLY m/z 90, ALA m/z 106, SER m/z 116, PRO m/z 118, VAL m/z 120, THR m/z 132, ILE/LEU m/z 133, ASN m/z 134, ASP m/z 147, LYS m/z 148, GLU m/z 150, MET m/z 156, HIS m/z 166, PHE m/z 175, ARG m/z 182, TYR m/z 205, TRP m/z 241, CYSCYS 7.5 10 12.5 15 Time [min] 17.5
Figure 1 Amino acid analysis in soy sauce (selected ion traces).
Experimental CE-ESI-MS analysis was performed using the Agilent Capillary Electrophoresis system with CE-MS capillary cassette coupled to the Agilent 1100 LC/MSD equipped with electrospray source and orthogonal sprayer for CE-MS. Agilent ChemStation software was used for instrument control. For method parameters see figure 1. A new capillary was flushed with run buffer for 20 minutes. A 17-amino acid standard (2.5 mmol/L except Cys at 1.25 mmol/L in 0.1 N HCl, not containing Gln) was obtained from Pierce (Rockford, IL, USA). Asp and Trp were purchased from Wako (Osaka, Japan) and prepared at a concentration of 2.5 mmol/L in 0.1 N NaOH. Amino acid standards were mixed and diluted with deionized water to a final concentration of 250 mol/L (Cys 125 mol/L). Sample preparation was simple and consisted of dilution with deionized water and ultrafiltration through a 30 kDa cutoff filter to remove peptides and proteins. Results An optimized method for the rapid analysis of amino acids by CE-ESI-MS has been developed. Using an electrolyte with a pH value below the analyte's isoelectric point (< 2.77) allows the simultaneous determination of basic and acidic amino acids as positive molecular ions. In this study an ion at m/z 122 for Cys could not be observed. However, an ion at m/z 241 for Cys-Cys indicated oxidation by exposure to air. When analyzing a standard composed of 19 amino acids, the assay was linear from 10500 mol/L. Detection limits ranged from 0.3 - 1.1 mol/L for basic amino acids to 611 mol/L for acidic amino acids. Migration time reproducibility was < 1.2 % RSD (n = 8), RSD for peak area was 2.04.7 % (except for Met, which gave decreasing peak areas due to oxidation). The analysis of amino acids in food is important. Their monitoring can help track fermentation metabolites and is used to correlate flavor trends. Figure 1 shows the analysis of soy sauce by CE-ESI-MS. In addition to amino acids, soy sauce contains a great number of organic compounds with amino-functional groups. Using methods with derivatization can cause very complex separations. Without derivatization, as shown here, well-defined extracted ion traces can be obtained, allowing detection of amino acids in less than 13 minutes. Migration time reproducibility was < 0.4 % RSD (n = 5). Area reproducibility was 1.16.0 % RSD except for Met. This method was further applied to the analysis of amino acids in beer and sake.
Equipment
Agilent Capillary Electrophoresis system Agilent CE-MS Adapter Kit Agilent 1100 Series LC/MSD module with API Electrospray Source Agilent CE-ESI-MS Sprayer Kit Agilent ChemStation and CE-MS add-on software
References
1 Soga, T., Heiger, D., J. Anal. Chem.2000, 72, 1236-1241.
Tomoyoshi Soga is an application chemist at Yokogawa Analytical Systems, Inc. Tokyo, Japan. Maria Serwe is an application chemist at Agilent Technologies, Waldbronn, Germany.
For more information on our products and services, visit our worldwide website at http://www.agilent.com/chem
Analysis of Amino Acids in Beer using HPLC with Online Derivatization

Abstract
Conditions
Both primary and secondary amino acids were analyzed in Column 200 2.1 mm Hypersil ODS, 5 m Mobile phase A = 0.03M sodium acetate one run. pH = 7.2 + 0.5% THF B = 0.1M sodium acetate/ ACN (1:4) The amino acid composition of proteins can be used to Gradient determine the origin of meat products and thus to detect at 0 min 0% B at 0.45 ml/min flow rate adulteration of foodstuffs. Detection of potentially toxic at 9 min 30% B amino acids is also possible through such analysis. at 11 min 50% B at 0.8 ml/min flow rate Through the use of chiral stationary phases as column at 13 min 50% B material, D and L forms of amino acids can be separated at 14 min 100% B at 0.45 ml/min flow rate and quantified. at 14.1 min at 0.45 ml/min flow rate HPLC in combination with automated online derivatization at 14.2 min at 0.8 ml/min flow rate is now a well-accepted method for detecting amino acids at 17.9 min at 0.8 ml/min flow rate at 18.0 min at 0.45 ml/min flow rate owing to its short analysis time and relatively simple at 18 min 100% B; at 19 min 0% B sample preparation. Post time 4 min Flow rate 0.45 ml/min Column compartment 40 C mAU Injection vol 1 l standard 70 Detector UV -DAD 338 nm and 266 nm 60 Fluorescence 50 Excitation wavelength: 230 nm WL Emission wavelength: 450 nm at 11.5 min 40 switch Excitation wavelength: 266 nm 30 Emission wavelength: 310 nm 20 Slit width excitation: 2 mm (25 nm) 10 Slit width emission 1: 4 mm (50 nm) Slit width emission 2: 4 mm (50 nm) 0 10 4 8 12 2 6 Time [min] Photomultiplier gain: 12 Cut-off filter: 280 nm Figure 1 Lamp: 55 Hz (always on) Analysis of amino acids in beer after online derivatization Response time: 4 s
ARG ILE TYR PHE VAL PRO ALA GLY LEU GLU HIS CIT MET LYS GLN SER ASN ASP
Sample preparation Hydrolyzation with HCl or enzymatic hydrolysis is used to break protein bonds. Chromatographic conditions The HPLC method presented here was used in the analysis of secondary and primary amino acids in beer with precolumn derivatization and fluorescence detection.1 HPLC method performance Limit of detection 5 pmol Repeatability of RT over 6 runs areas over 6 runs Linearity 1 pmol to 4 nmol References 1. R. Schuster, Determination of amino acids in biological, pharmaceutical, plant and food samples by automated precolumn derivatization and HPLC, J. Chromatogr., 1988, 431, 271284. <1 % <5 %
Conditions
Injector program for online derivatization 1. Draw 3.0 l from vial 2 (borate buffer) 2. Draw 1.0 l from vial 0 (OPA reagent) 3. Draw 0.0 l from vial 100 (water) 4. Draw 1.0 l from sample 5. Draw 0.0 l from vial 100 (water) 6. Mix 7.0 l (6 cycles) 7. Draw 1.0 from vial 1 FMOC reagent 8. Draw 0.0 l from vial 100 (water) 9. Mix 8.0 l (3 cycles) 10. Inject
Equipment
Agilent 1100 Series vacuum degasser quaternary pump autosampler thermostatted column compartment diode array detector fluorescence detector Agilent ChemStation + software
Analysis of Vanillin Extract Quality using HPLC
Abstract The following compounds are examples of flavoring agents used in food products: lupulon and humulon (hop bittering compounds) vanillin naringenin and hesperidin (bittering compounds) Three major classes of compounds are used as flavoring agents: essential oils, bitter compounds, and pungency compounds. Although the resolution afforded by gas chromatography (GC) for the separation of flavor compounds remains unsurpassed, HPLC is the method of choice if the compound to be analyzed is low volatile or thermally unstable. Sample preparation Turbid samples require filtration, whereas solid samples must be extracted with ethanol. Afterfiltration, the solution can be injected directly into the HPLC instrument.
Conditions
Column 100 4 mm Hypersil BDS, 3 m Mobile phase A = water + 0.15ml H2SO4 (conc.), pH = 2.3 B = ACN Gradient start with 10% B at 3 min 40% B; at 4 min 40% B at 6 min 80% B; at 7 min 90% B Flow rate 0.8 ml/min Post time 3 min Column compartment 30 C Injection vol 5 l Detector UV-DADdetection wavelength 280/80 nm, reference wavelength 360/100 nm Sample preparation Injection without further preparation
Norm. 400 300 200 100 0 0
4-hydroxy benzoic acid
Vanillin 4-hydroxybenzaldehyde Ethylvanillin Coumarin
Vanillin alcohol
Standard Vanillin extract
4 Time [min]
Chromatographic conditions The HPLC method presented here for the analysis of vanillin is based on reversed-phase chromatography. UV spectra were evaluated as an additional identification tool.1 HPLC method performance Limit of detection 0.25 ng (injected amount) S/N = 2 Repeatability of RT over 10 runs <0.2 % of areas over 10 runs <1 %
mAU 60
Conditions
Conditions as above, except Column 100 2.1 mm Hypersil ODS, 5 m Mobile phase A = water + 5 mM NaH2PO4 B = methanol Gradient at 10 min 70% B Flow rate 0.4 ml/min
Equipment
Vanillin
40
30 20 10
Scaled
50
Gallic acid
Syringaaldehyde
50 40 30 20 10 0
Vanillin
Match 991
Salicylaldehyde
217 Wavelength [nm] 400
0 0 2 4 Time [min] 6
8
Standard Cognac
10
Figure 2 Analysis of vanillin in cognac. Identification of vanillin through spectra comparison References 1. Herrmann, A, et al.;,Rapid control of vanilla-containing products using HPLC; J. Chromatogr., 1982, 246, 313316.
Carbohydrates
Effect of Mobile-Phase Strength in Separations of Mono- and Disaccharides

Application
Food Analysis
Robert Ricker
Sugars are important component of many foodstuffs derived from plant and animal sources Sugars are also of great clinical importance. They have traditionally separated by normal- phase on NH2 bonded phases. Effect of mobile-phase solvent strength in these separations is demonstrated for various mono- and disaccharides.
Highlights
Good resolution of various mono- and disaccharides is obtained using a ZORBAX NH2 column.
1. Fructose 2. Glucose 3. Saccharose 4. Palatinose 5. Trehalulose 6. Isomaltose
Retention of sugars increases with increasing organic content in the mobile phase.
Conditions: ZORBAX NH2 (4.6 x 250 mm) (Agilent P/N: 880952-708) Mobile Phase:!!ACN : H2O, as indicated " mL/min, Detect. = Refractive Index
High-Performance Carbohydrate Analysis
Application
Food Analysis
Robert Ricker
Analysis of mono and disaccharides (simple sugars) in foods is required when their presence is greater than 1%, according to the Food and Drug Administration's new policy. Furthermore, sugar content must be included on the food label. When sugar analyses are required, HPLC provides speed and ease of use. In fact, the Association of Official Analytical Chemists (AOAC) has specified HPLC and a propylamine column as an official method for several matrices (e.g., pre-sweetened cereals). The high performance ZORBAX NH2 columns (5m) are an optimal choice for these applications as well as for analysis of some high-fructose corn syrups which are added to foods instead of cane sugar. The Agilent ZORBAX NH2 column provides higher resolution than many so-called 10m, "sugar analysis specified" columns.
Highlights
Good resolution of various mono- and disaccharides is obtained using a ZORBAX NH2 column.
1. Fructose 2. Glucose 3. Saccharose 4. Palatinose 5. Trehalulose 6. Isomaltose
Retention of sugars increases with increasing organic content in the mobile phase.
ZORBAX NH2 (4.6 x 250 mm) (Agilent Part No. 880952-708) Mobile Phase: ACN : H2O, as indicated 1 mL/min, Detect. = Refractive Index
Application of Liquid Chromatography/ Mass Spectrometry to the Analysis of Sugars and Sugar-Alcohol Application
Food
Author
Hiroki Kumaguai
Liquid chromatography/mass spectrometry (LC/MS) methods using electrospray ionization (ESI) also require pre- or post-column derivatization to obtain a high level of sensitivity. LC/MS method, described here, using atmospheric pressure chemical ionization (APCI) does not require pre- or post-column derivatization to attain high sensitivity. However, this method does require CHCl3 and CH2Cl2 to be added in the post-column stage. Methods using derivatization are the most sensitive, with a minimum detectability of several to tens of picograms (pg). Although the APCI method, employing post-column addition, has a lower sensitivity (several hundred picograms), it is an easy and versatile method of analysis, with superior ionization repeatability, generated by the [M+Cl] ion.
Abstract
Sugars and sugar alcohols were successfully analyzed by liquid chromatography/mass spectrometry using atmospheric pressure chemical ionization in negative ion mode.
Background
Sugars, quantitatively the largest organic compound group on earth, are widely distributed among both flora and fauna. Higher classes of vegetation and algae contain large quantities of sugars, and the shells of arthropods, represented by crabs and shrimp, are made of chitin, which are polysaccharides. Although sugars represent a huge biomass; they also exist in very small amounts within individual living organisms. Various kinds of sugars and compound sugars are involved in bodily functions, and as sources of energy. Sugars are used as raw materials within the textile, food processing and pharmaceutical industries. Sugars have been analyzed by various methods: Gas chromatorgraphy/mass spectrometry (GC/MS) methods require preliminary derivatization to increase sugar's volatility. High-performance liquid chromatography (HPLC) methods have limitations, including detector sensitivity. Commonly, one of two different detectors is used. - differential refraction detector - ultraviolet (UV)/fluorescence detector, following application of pre- or post-column derivatization; without derivatization, sugars are not UV detectable
Method
Instrument: HP 11001 LC/MS with APCI Negative ion mode Drying gas: N2 13 L/min and 350 C Nebulizer: N2 40 psi Fragmentor: 20 V Corona current: 30 A Mass range: 100500 m/z Mode: SIM (m/z) negative Sorbitol 217, 219 Glucose, Fructose 215, 217 Xylitol 187, 189 Sucrose 377, 379 LC Conditions: Mobile phase: CH3CN/H2O (75/25) at 0.2 mL/min Oven temperature: 40 C Injection volume: 10 L Post-column addition: CH3CN/CHCl3 (50/50) at 0.2 mL/min Column: Asahipak NH2-50 2D, 2.0 mm id 150 mm long
1
Now available as the Agilent 1100 LC/MS from Agilent Technologies, Inc.
Sample Analysis
Selected sugars and sugar alcohols in some standards and common beverages were measured using this
LC/MS method with APCI in negative ion mode. The three figures below illustrate both the sensitivity and applicability of this method.
Adundance 200000
180000
Sorbitol
160000
140000
Glucose
120000
100000
80000
60000
40000 1 Adundance 200000 2 3 4 5 6 7 8 min
Xylitol
180000
160000
140000
120000
Fructose Glucose
100000
Sucrose
80000
60000
40000 1 2 3 4 5 6 7 8 min
Figure 1. Stacked SIM chromatograms of sugar standards at 1 g/mL each.
Adundance
120000 100000 80000 60000 40000 20000

0
Coffee
Sucrose
9 min
Adundance
300000 250000 200000 150000 100000 50000

0
Tea
Sucrose
9 min
Adundance 140000
Orange Juice
120000 100000 80000 60000 40000 20000
Fructose
Glucose
9 min
Figure 2. Stacked SIM chromatograms for typical sugars in three common beverages.
www.agilent.com
Adundance 800000 700000 600000 500000 400000 300000 200000 100000
TIC
10
12 min
Adundance 180000 160000 140000 120000 100000 80000 60000 40000 0 2 4 6 8 10 12 min
SIM
Glucose
Figure 3. Stacked TIC and SIM chromatograms for Sake (Japanese rice wine).
Conclusion
Underivatized sugars and sugar alcohols can be successfully analyzed by LC/MS using atmospheric pressure chemical ionization (APCI) in negative ion mode. Hiroki Kumagai is an application chemist at Agilent/Yokogawa Analytical Systems, Tokyo, Japan.
Agilent Technologies shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Copyright 2001 Agilent Technologies, Inc. www.agilent.com/chem Printed in the USA October 4, 2001 5988-4236EN
Process control of starch
Application
Heinz Goetz and Peter Kilz
Starch is a polysaccharide composed of two components with different molecular weightabout 2030 % amylose and about 8070 % amylopectin. Starch is a very important part of the diet for both humans and animals. For example, it is available in potatoes, flour and bread. Further areas of use are in the production of paste, adhesive, glucose, dextrin and beer. Figure 1 shows an overlay of three chromatograms obtained under the same conditions. The chromatogram of the
Starch Original Degradation I Degradation II nRIU 6000 5000 4000 3000 2000 1000 0 10 11 12 13 14 15 Time [min] 16 17 18 I Mn * 1000 910 Mw * 4700 6100
Conditions
Sample preparation Sample was dissolved in 1 mL eluent at 20 C (concentration 0.1 % w/w). Dextran standards from Polymer Standards Services (PSS) were used for narrow standard calibration. Column PSS Suprema 100 + 1000 in series, 10 m, 300 8 mm Mobile phase 0.1 M sodium nitrate Flow rate 1 ml/min Column compartment temperature 25 C Injection volume 100 l Detector refractive index detector
* not calculated because of steric exclusion
Original starch Starch after incorrect processing
Glucose II
Figure 1 Overlay of chromatograms of original starch and starch after incorrect processing
original starch shows a big peak of the polysaccharide eluting at a retention time of approximately 11 minutes. This peak almost disappears in the chromatograms of the degraded starch (chromatograms I and II). New peaks from degradation products elute between 12 and 17 minutes, especially from the degradation end product glucose. Degradation of the original starch is also shown by the molecular weight data (table in figure 1). Both the number (Mn) and the weight average molecular weight (Mw) of the degraded products are significantly below 10000 dalton while starch in it's original state gives data of 300000 to 2000000 dalton. The GPC-SEC method described here shows an easy but reliable analysis for process and quality control of starch.
Equipment
Agilent 1100 Series GPC-SEC system consisting of vacuum degasser for efficient degassing of the mobile phase isocratic pump with large solvent cabinet autosampler with single valve design thermostatted column compartment for precise column temperatures refractive index detector with automatic recycle valve ChemStation Plus with GPCSEC data analysis software
Columns supplier: Polymer Standards Service, Mainz, Germany
Heinz Goetz is an application chemist at Agilent Technologies, Waldbronn, Germany. Peter Kilz is Managing Director at Polymer Standards Service, Mainz, Germany For more information on our products and services, visit our worldwide website at http://www.agilent.com/chem
Copyright 2000 Agilent Technologies Released 09/2000 Publication Number 5988-0117EN
Rapid Monitoring of Carbohydrates in Food with Capillary Electrophoresis

Tomoyoshi Soga and Maria Serwe Pharmaceutical Abstract Processed food often contains additives such as sweeteners or preservatives. For regulatory compliance the concentration of these substances needs to be monitored. This is important to ensure the consumer's protection and also has economic implications. In Japan, for example, the tax rate of imported juice depends on the sucrose content. Non-derivatized carbohydrates can rapidly be determined using capillary electrophoresis with indirect UV-detection. The method's simplicity and stability make it suitable for the application in routine labs. A Carbohydrate standard (1000 ppm each) B Orange juice (diluted with H2O 1:20, filtered through 0.22-m filter) C Yogurt (diluted with H2O 1:20, filtered through 30 kDa cut-off filter) 1 Fructose 2 Glucose 3 Maltose 4 Lactose Absorbance 5 Sucrose [mAU] 30 25 20 15 10 5 0 -5 10 11 12 13 14 Time [min] 15 16 Figure 1 Analysis of sugars in orange juice and yogurt C B A 1 2 3 4 5
Conditions
Sample see figure Injection 6 s @ 50 mbar Capillary bare fused silica capillary total length 80.5 cm effective length 72 cm internal diameter 50 m Buffer Agilent Basic Anion Buffer Voltage -25 kV Temperature 15 C Detection signal 350/20 nm reference 275/10 nm
Experimental Carbohydrate analysis was performed using the Agilent Capillary Electrophoresis system equipped with DAD detection and computer control via Agilent ChemStation. The method is based on the use of the Agilent Basic Anion Buffer (part number 5064-8209) together with a standard bare fused silica capillary (G1600-62211). Prior to first use, a new capillary was flushed with run buffer for 15 minutes (at 1 bar). Between analyses the capillary was flushed 4 minutes from an extra buffer vial into waste. Sample preparation was simple and consisted, even in the case of yogurt, of only dilution and ultrafiltration for protein removal. Results Figure 1 shows the analysis of orange juice and yogurt in comparison with a carbohydrate standard. The standard assay was linear over the range 5010 000 ppm with r2 > 0.999. Method detection limits were 715 ppm. Sucrose levels were 43 g/l in orange juice and 109 g/l in yogurt. Fructose and glucose in orange juice were determined as 23 g/l and 25 g/l, respectively. The yogurt contained 26 g/l lactose. Repeatability was < 0.16 % RSD for migration times and < 2.3 % RSD for peak areas (n = 6). High performance liquid chromatography (HPLC) with refractive index detection is routinely used for the analysis of carbohydrates in food, yielding low ppm detection limits when modern equipment is used. However, capillary electrophoresis with indirect UV-detection, providing the same sensitivity at short overall run times, can be a worthy complement. This is especially true for complex sample matrices, which present challenges for the classical stationary phases used for carbohydrate analysis by HPLC. This makes the method applicable to a wide range of food with complex matrices. In Japan, the method presented here is routinely applied as a quick screening for carbohydrates in fruit juices, whiskey or rice plant sap.
Equipment
Agilent Capillary Electrophoresis system Agilent ChemStation Agilent Basic Anion Buffer
Tomoyoshi Soga is an application chemist at Yokogawa Analytical Systems Inc., Tokyo, Japan. Maria Serwe is an application chemist at Agilent Technologies, Waldbronn, Germany. For more information on our products and services, visit our worldwide website at http://www.agilent.com/chem
Analysis of sugars in foods and beverages by HPLC with pulsed amperometric detector
Hiroki Kumagi and Ikuko Tajima Food
Abstract Sugars are generally analyzed by HPLC with refractive index detection. UV or fluorescence detection are also frequently used, coupled with pre or post derivatization for high sensitive analysis. However, the methods mentioned above have some limitations, such as the difficulty of simultaneously analyzing various sugars, or the tedious procedure of derivatization. The pulsed amperometric detector (PAD) is a common method for sugar analysis. The coupling of the PAD and post-column pHchanging offers a simple, easy and highly sensitive way of detecting sugars. This brief demonstrates the analysis of various sugars in candy and juice using PAD with postcolumn pHchanging. The sugars are separated by both ligand exchange and size exclusion mode. Analyzed Compounds Sucrose, glucose, galactose, mannose, fructose, mannitole, and sorbitol. Sample Candy and apple juice.
Conditions
Column 300 x 7.8 mm Excelpak CHA-L31 (This column is same as SPELCOGEL C-611) Mobile phase 0.1 mM NaOH, 0.75 ml/min Post column reagent 400 mM NaOH, 0.6 ml/min Column temp 80 C Injection vol 20 l Detector Agilent 1049 Electrochemical detctor Mode; Pulse mode Working electrode; gold Applied Potential; Pot1 = 0.65 V, Pot2 = -0.95 V, Pot3 = 0.15 V Sample preparation Candy was dissolved in water and filtrated Juice was diluted and filtrated.
Absorbance [mAU] 7 6 5 4 3 2 1 0 0 2.5 5 7.5 10 12.5 15 1 3 6 4 5 7 2
1 2 3 4 5 6 7
Sucrose Glucose Galactose Mannose Fructose Mannitole Sorbitol
17.5 Time [min]
Figure 1 Chromatogram of standard solution, 50 mgl/l each
Equipment
Absorbance [mAU] 1 Glucose 2 Galactose 3 Fructose
17.5 15 12.5 10 7.5 5 2.5 0 0 2.5 5 7.5 10 12.5 15 2 3 1
Agilent 1100 Series 2 isocratic pumps with vacuum degasser autosampler thermostatted column compartment programmable electro chemical detector Agilent ChemStation + software
17.5 Time [min]
Figure 2 Chromatogram of sugars in candy
Absorbance [mAU]
60 50 40 30 20 10 0 0 2.5 5 7.5 1
2 3
1 2 3 4
Sucrose Glucose Fructose Sorbitol
Method Performance Limit of detection 0.23~0.95 mg/L (S/N = 3) RSD of peak area 1.0~4.0 % RSD of retention time 0.1~0.2 %
10
12.5
15
17.5 Time [min]
Figure 3 Chromatogram of sugars in apple juice
Hiroki Kumagi and Ikuko Tajima are application engineers at Yokogawa Analytical Systems Inc., Tokyo, Japan. For more information on our products and services, visit our worldwide website at http://www.agilent.com/chem
Analysis of Carbohydrates in Lemonade using HPLC
Abstract The following carbohydrates have been analyzed: glucose, galactose, raffinose, fructose, mannitol, sorbitol, lactose, maltose, cellobiose, and sucrose. Food carbohydrates are characterized by a wide range of chemical reactivity and molecular size. Because carbohydrates do not possess chromophores or fluorophores, they cannot be detected with UV-visible or fluorescence techniques. Nowadays, however, refractive index detection can be used to detect concentrations in the low parts per million (ppm) range and above, whereas electrochemical detection is used in the analysis of sugars in the low parts per billion (ppb) range. Sample preparation Degassed drinks can be injected directly after filtration. More complex samples require more extensive treatment, such as fat extraction and deproteination. Sample cleanup to remove less polar impurities can be done through solid-phase extraction on C18 columns.
Norm 800 600
Conditions
Citric acid? Lactose Glucose Raffinose Galactose Standard
400 200 5 Time [min] 10 Fructose 15
Lemonade
Column 300 7.8 mm Bio-Rad HPXP, 9 m Mobile phase water Column compartment 80 C Flow rate 0.7 ml/min Detector refractive index Sample preparation Samples were directly injected.
Figure 1 Analysis of carbohydrates in lemonade
Chromatographic conditions The HPLC method presented here was used to analyze mono-, di-, and trisaccharides as well as sugar alcohols. HPLC method performance Limit of detection <80 ng with S/N = 2 Repeatability of RT over 10 runs <0.05 % areas over 10 ru ns 2 %
Glucose
Equipment
Agilent 1100 Series degasser isocratic pump autosampler thermostatted column compartment refractive index detector Agilent ChemStation + software
180 Raffinose 160 140 120 100 80 Cellbiose 5
Lactose
Norm
Galactose
Fructose Standard Corn extract Sucrose 15 Standard 20
Maltose 10
Time [min]
Figure 2 Analysis of carbohydrates in corn extract References 1. Official Methods of Analysis, Food Compositions; Additives, Natural Contaminants, 15th ed; AOAC: Arlington, VA, 1990, Vol. 2; AOAC Official Method 980.13: Fructose, glucose, lactose, maltose, sucrose in milk chocolate; AOAC Official Method 982.14: Glucose, fructose, sucrose, and maltose in presweetened cereals; AOAC Official Method 977.20: Separation of sugars in honey; AOAC Official Method 979.23: Saccharides (major) in corn syrup; AOAC Official Method 983.22: Saccharides (minor) in corn syrup; AOAC Official Method 984.14: Sugars in licorice extracts.
Fats and Oils
Analysis of Triglycerides by Liquid Chromatography/Mass Spectrometry Application
Food
Author
Hiroki Kumagai
Table 1. Chemical Structure of Triglycerides CH2 CH CH2 OCO OCO OCO R1 R2 R3 R1=R2=R3 C11 C13 C15 C18 C18:1 C18:2
Abstract
Triglycerides were readily analyzed using liquid chromatography/mass spectrometry with atmospheric pressure chemical ionization in positive ion mode.
Triglyceride Trilaurin Trimyristin Tripalmitin Tristearin Triolein Trilinolein
Background
Oil is a primary component of many foods, such as cooking oils and dairy products. Consequently, it is desirable to conduct compositional analyses of oil-based fatty acid components in food. Such analyses may use either gas chromatography (GC) or high-performance liquid chromatography (HPLC), but the GC methods have problems with the complexity of sample preprocessing, which requires saponification. With HPLC a direct analysis is possible; however, these methods are low in sensitivity because the compounds of interest do not absorb ultraviolet and separation is barely adequate. In this study, six triglycerides (Table 1) with identical fatty acid compositions were analyzed using liquid chromatography/mass spectrometry (LC/MS) and atmospheric pressure chemical ionization (APCI) as the ionization mode.
Method
Instrument: Agilent 1100 LC/MS with APCI in positive mode Mass range: 100 to 1000 m/z Drying gas: N2 4 L/min, 350 C Nebulizer: N2 50 psi Fragmentor: 160 V Vaporizer: 400 C LC Conditions: Mobile phase: (CH3)2CO/H2O (98/2) Flow rate: 1.0 mL/min Oven temperature: 40 C Injection volume: 15 L Column: Develosil ODS-UG3, 4.6 mm id 75 mm long
Results
The following figures show total ion chromatogram (TIC) and selected ion mode (SIM) chromatograms, and mass spectra for the selected triglyceride standards.
Abundance 40000
2 1. 2. 3. 4. 5. 6. Trilaurin Trilinolein Trimyristin Triolein Tripalmitin Tristearin
35000
4
30000
25000
20000
15000
10000
5000
1.00
2.00
3.00
4.00
5.00
6.00
min
Figure 1. TIC of triglyceride standards, each at 0.2 ppm.
439.5
90
T il i Trilaurin
90
Trilinolein
599.5
879.8 (MH)+
90
495.5
Trimyristin
50
50
183
150 250 350 450 550 650 750 850 150
50
305
250 350 450 550 650 750 850
211
150 250 350 450 550 650 750 850
Triolein
90
603.5
Tripalmitin
90
551.5
90
Tristearin
607.5
50
50
50
881
150 250 350 450 550 650 750 850 150 250 350 450 550 650 750 850 150 250 350 450 550 650 750 850
Figure 2. Mass spectra of individual triglyceride standards at 0.2 ppm. The tallest peak in each case represents the [MH-HOOCR]+ ion created after the loss of a fatty acid group from the protonated psuedomolecular ion. With the exception of Trilinolein, the protonated psuedomolecular ion was not observed.
5000 4000 3000 2000 1000 0.40
Trilaurin 439.5 m/z
5000 4000 3000 2000 1000 0
Trilinolein 599.5 m/z
0.80
1.20
1.60
2.00 2.40
2.80
0.40 0.80 1.20
1.60
2.00 2.40
2.80
7000 6000 5000 4000 3000 2000 1000 0 0.40
Trimyristin 495.5 m/z
0.80
1.20
1.60
2.00
2.40
2.80
8000 6000
Triolein 603.5 m/z
14000 10000 6000
Tripalmitin 551.5 m/z
20000 15000 10000 5000
Tristearin 607.5 m/z
4000 2000 2000 3.00 4.00 5.00 6.00 3.00 4.00 5.00 6.00
3.00
4.00
5.00
6.00
Figure 3. SIM chromatograms of individual triglyceride standards at 1 ppb.
The protonated molecular ions were rarely observedtrilinolein in the scan mode is the exceptionand the base peaks consisted of fragmented ions from which fatty acid had been detached. Sensitivity was favorable, extending down to 0.2 ppm in TIC mode. In the SIM mode, measurements at 1 ppb were possible by selecting the base peak in the mass spectrum of each composition of monitored ions. With this technique, it is possible to measure fat-soluble substances such as triglycerides with a high degree of sensitivity.
Conclusions
LC/MS, with APCI in positive ion mode, readily detected selected triglyceride standards with high sensitivity. The analytical peaks were the positvely charged fragmented ions from which fatty acid had been detached. SIM mode yielded sensitivity to 1 ppb.
Hiroki Kumagai is an application chemist at Agilent/Yokogawa Analytical Systems, Tokyo, Japan.
www.agilent.com
Agilent Technologies shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Copyright 2001 Agilent Technologies, Inc. www.agilent.com/chem Printed in the USA August 22, 2006 5988-4235EN
Column Selection for the Analysis of Fatty Acid Methyl Esters Application
Food Analysis
Authors
Frank David Research Institute for Chromatography President Kennedy Park 20 B-8500 Kortrijk, Belgium Pat Sandra University of Gent Krijgslaan 281 S4, B-9000 Gent Belgium Allen K. Vickers Agilent Technologies, Inc. 91 Blue Ravine Road Folsom, CA 95630-4714 USA
In this application note, three different stationary phases are compared for the separation of FAMEs. Polyethylene glycol columns gave good separation for the less complex samples, but they did not separate cis-trans isomers. A medium polar cyanopropyl column (DB23) provided excellent separation for complex FAME mixtures and also achieved some cis-trans separation. For more detailed cis-trans separation, the highly polar HP-88 cyanopropyl column is preferred.
Introduction
The analysis of FAMEs is used for the characterization of the lipid fraction in foods, and is one of the most important analyses for food. Lipids mainly consist of triglycerides, being esters of one glycerol molecule and three fatty acid molecules. Most edible fats and oils contain mainly fatty acids ranging from lauric acid (dodecanoic acid) to arachidic acid (eicosanoic acid). Besides the linear saturated fatty acids, branched fatty acids, monounsaturated, di-unsaturated, and higher unsaturated fatty acids can also occur. An overview of the most important fatty acids and their abbreviations appears in Table 1.
Abstract
The analysis of fatty acid methyl esters (FAMEs), derived from food, is a very important food characterization procedure. These esters are normally analyzed on columns coated with polar stationary phases, such as polyethylene glycols or cyanopropyl silicones, allowing separation of fatty acids according to their carbon number, the degree of unsaturation, the cis-trans configuration, and the location of the double bonds.
Table 1.
Fatty Acids, Common Names, and Abbreviations Common Name Butyric acid Caproic acid Lauric acid Myristic acid Palmitic acid Palmitoleic acid Stearic acid Oleic acid Elaidic acid Linoleic acid Linolelaidic acid -Linolenic acid -Linolenic acid Arachidic acid Abbreviation C4:0 C10:0 C12:0 C14:0 C16:0 C16:1 C18:0 C18:1- cis (n9) C18:1- trans (n9) C18:2 - cis (n6) C18:2 - trans (n6) C18:3 (n3) C18:3 (n6) C20:0 C20:1 (n9) C20:2 (n6) C20:3 (n3) Dihomogammalinolenic acid Arachidonic acid EPA Behenic acid Erucic acid DHA Lignoceric acid Nervonic acid C20:3 (n6) C20:4 (n6) C20:5 (n3) C22:0 C22:1 (n9) C22:4 (n6) C22:6 (n3) C24:0 C24:1 (n9)
Fatty acid Butanoic acid Decanoic acid Dodecanoic acid Tetradecanoic acid Hexadecanoic acid Hexadecenoic acid Octadecanoic acid cis-9-Octadecenoic acid trans-9-Octadecenoic acid all cis-9,12-Octadecadienoic acid all trans-9,12-Octadecadienoic acid all cis-9,12,15-Octadecatrienoic acid all cis-6,9,12-Octadecatrienoic acid Eicosanoic acid cis-11-Eicosenoic acid all cis-11,14-Eicosadienoic acid all cis-11,14,17-Eicosatrienoic acid all cis-8,11,14-Eicosatrienoic acid all cis-5,8,11,14-Eicosatetraenic acid all cis 5,8,11,14,17-Eicosapentenoic acid Docosanoic acid cis-13-Docosenoic acid all cis-7,10,13,16-Docosatetraenoic acid all cis 4,7,10,13,16,19-Docosahexenoic acid Tetracosanoic acid cis-15-tetracosenoic acid
For the characterization of the lipid fraction, the triglycerides are hydrolyzed (saponified) into glycerol and free fatty acids. Although the free fatty acids can be analyzed directly on polar stationary phases (such as an HP-FFAP column), more robust and reproducible chromatographic data are obtained if the fatty acids are derivatized to the methyl esters. For the derivatization, including hydrolysis and methylation, different methods are available [1]. These methods are easy to use and do not require expensive reagents or equipment. A typical sample preparation method is described in the sample preparation section. After preparation of the FAMEs, they are separated according to the carbon number (number of carbon atoms in the fatty acid chain, excluding the methyl ester carbon) and the degree of unsaturation. Moreover, the position of the double bond(s)
and the geometric configuration (cis/trans) are also important parameters and their determination adds additional information to the characterization of the lipid fraction in food. In this application note, three stationary phases are compared for the separation of FAMEs. The first method uses DB-Wax, a polyethylene glycol column, where FAMEs from C4 (butyric acid) to C24 (lignoceric acid) can be separated according to carbon number and degree of unsaturation. On these columns, no separation of cis- and transisomers is obtained, and for complex mixtures, such as fish oils, some FAMEs are difficult to separate. However, the separation of FAMEs on polyethylene glycol columns is widely used and are applied to the characterization of classical samples, such as vegetable oils from corn, maize, olive, and soybean. Moreover, animal fats can also be
analyzed. One important application is the analysis of butyric acid in milk fat. The concentration of butyric acid in milk is an important indicator of milk quality, and its analysis is therefore very important in milk, dairy, and chocolate products. For the analysis of complex samples, such as fish oils, additional resolution of the FAMEs is needed, and is obtained using a capillary column coated with a cyanopropyl-stationary phase, such as a DB-23. On this column, highly unsaturated fatty acids, such as all cis 5, 8, 11, 14, 17-eicosapentenoic acid methyl ester (EPA, C20:5 3) and all cis 4,7,10,13,16,19-docosahexenoic acid methyl ester (DHA, C22:6 3) are separated from other FAMEs. This analysis is very important in the framework of recent interest in omega-3 fatty acid determination. On the cyanopropyl column, separation of the cis- and trans-isomers is also possible. Due to stronger interaction of the cis-isomer with the cyano-dipole, the trans-isomers elute before the cis-isomers. In this way, the determination of trans-fatty acids is also performed, however, the polarity of the stationary phase is not sufficient to fully separate complex cis-trans mixtures. For the separation of a complex FAME mixture containing a relatively large amount of trans-fatty acids, a highly polar HP-88 column is preferred. On this highly polar column excellent separation between different cis- and trans-isomers is obtained, however, some higher molecular weight fatty acids are more difficult to separate. An overview of columns and their application area is summarized in Figure 1.
Experimental
Samples Reference standards of FAMEs can be obtained from different sources as solutions or as neat compounds. For analysis, the standards are typically dissolved in hexane at a 0.01%0.1% (w/v) concentration. For column check-out, a 37-component mixture (Supelco #18919) was used. The mixture is available as a 100-mg neat mixture, containing C4C24 FAMEs (2%4% relative concentration). The whole sample was diluted in 10-mL hexane (final concentration = 0.20.4 mg/mL per FAME) before use. Oil and fat samples can be prepared using different methods [15]. Sample Preparation Method [5] Weigh 100-mg sample in a 20-mL test tube (with screw cap) or reaction vial. Dissolve the sample in 10-mL hexane. Add 100-L 2 N potassium hydroxide in methanol (11.2 g in 100 mL). Close the tube or vial and vortex for 30 s. Centrifuge. Transfer the clear supernatent into a 2-mL autosampler vial. Analytical Conditions The analyses were performed on an Agilent 6890 GC equipped with a flame ionization detector (FID). Automated split injection was performed using an Agilent 7683 autosampler. The instrumental configuration and analytical conditions are summarized in Table 2 (DB-Wax column), Table 3 (DB-23 column) and Table 4 (HP-88 column).
FAME Analysis
DB-Wax No cis-trans Some overlap "Simple" mixtures Milk fat
DB-23 Complex mixtures Partial cis-trans separation Omega 3 (EPA, DHA) analysis Fish oils
HP-88 Complex mixtures Excellent cis-trans separation Olive oil Refined (hydrogenated) oils
Figure 1.
Overview of column selection for FAMEs analysis.
Table 2.
DB-Wax Method 1 Agilent 6890 GC Split/Splitless FID or Agilent 5973 MSD Agilent 7683 Split liner (p/n 5183-4647) 30 m 0.25 mm ID, 0.25 m DB-Wax (J&W 122-7032) 250 C 1 L 1/50 Hydrogen 53 kPa constant pressure (36 cm/s at 50 C) 50 C, 1 min, 25 C/min to 200 C, 3 C/min to 230 C, 18 min. 280 C Hydrogen: 40 mL/min; Air: 450 mL/min; Helium make-up gas: 30 mL/min.
Instrumentation Chromatographic system: Inlet Detector Automatic Sampler Liner Column Experimental Conditions GC-FID Inlet temperature Injection volume Split ratio Carrier gas Head pressure Oven temperature Detector temperature Detector gases Table 3. DB-23 Method 2
Instrumentation Chromatographic system: Inlet Detector Automatic Sampler Liner Column Experimental Conditions GC-FID Inlet temperature Injection volume Split ratio Carrier gas Head pressure Oven temperature Detector temperature Detector gases Table 4. HP-88 Methods 3A and 3B
Agilent 6890 GC Split/Splitless FID or Agilent 5973 MSD Agilent 7683 Split liner (p/n 5183-4647) 60 m 0.25 mm ID, 0.15 m DB-23 (J&W 122-2361) 250 C 1 L 1/50 Helium 230 kPa constant pressure (33 cm/s at 50 C) 50 C, 1 min, 25 C/min to 175 C, 4 C/min to 230 C, 5 min. 280 C Hydrogen: 40 mL/min; Air: 450 mL/min; Helium make-up gas: 30 mL/min.
Instrumentation Chromatographic system: Inlet Detector Automatic Sampler Liner Column A Column B Experimental Conditions GC-FID Inlet temperature Injection volume Split ratio Carrier gas A Carrier gas B Head pressure Oven temperature A Oven temperature B Detector temperature Detector gases
Agilent 6890 GC Split/Splitless FID or Agilent 5973 MSD Agilent 7683 Split liner (p/n 5183-4647) 100 m 0.25 mm ID, 0.2 m HP-88 (J&W 112-88A7) 60 m 0.25 mm ID, 0.2 m HP-88 (J&W 122-8867) 250 C 1 L 1/50 Hydrogen Helium 2 mL/min constant flow 120 C, 1 min, 10 C/min to 175 C, 10 min, 5 C/min to 210 C, 5 min 5 C/min to 230 C, 5 min 175 C, 10 min, 3 C/min, 220 C, 5 min 280 C Hydrogen: 40 mL/min; Air: 450 mL/min; Helium make-up gas: 30 mL/min.
Results
A typical chromatogram for the analysis of the 37-compound FAMEs reference standard, obtained on the DB-Wax column is shown in Figure 2.
C10:0 C8:0 C12:0
Norm. 100
C6:0
C18:1 c+t
90
C14:0
C16:0
DB-wax
C11:0
C18:0
70
C4:0
80
C14:1 C15:0 C15:1
C18:2 c+t C18:3 n6 C18:3 n3
C13:0
60
50
C20:0 C20:1 C20:2 C20: 3n6 + C21:0 C20:3 n3 C20:4 C20:5 C22:0 C22:1
C16:1 C17:0 C17:1
30
20 0 5 10 15 Time (min) 20 25 30
Figure 2.
GC-FID analysis of 37-component FAMEs standard mixture on a 30 m 0.25 mm ID, 0.25 m DB-Wax column using Method 1. (See Table 2).
A good separation is obtained, except for the following compounds: cis- and trans-C18:1 coelute at 14.38 min, cis- and trans-C18:2 coelute at 15.13 min, C20:3 n6 and C21:0 coelute at 19.44 min, and C22:6 and C24:1 coelute at 30.73 min. However, this separation is sufficient for some classical oil and fat characterization methods. Butyric acid elutes at 4.28 min and can be determined in milk fat using this method. This is demonstrated in Figure 3, showing the analysis of a certified reference sample of milk fat (CRM 164, [6]).
C22:2 C23:0
C24:0
40
C22:6 + C24:1
C14:0
C16:0
Norm. 100
DB-wax
90 80 70 60 50 40 30 20
C4:0
C6:0
C10:0
C12:0
C18:0 C18:2 c+t

14
C8:0
C14:1
C15:0
C16:1
C18:1 c+t
10 Time (min)
C15:1
12
16
Figure 3.
GC-FID analysis of milk fat (CRM 164) fatty acids on a 30 m 0.25 mm ID, 0.25 m DB-Wax column using Method 1, Table 2.
The separation of the 37-compound FAME standard mixture on the 60 m 0.25 mm ID, 0.15 m DB-23 column is shown in Figure 4.
Norm. 100
C10:0
C12:0
C14:0
C16:0
C8:0
90
DB-23
C18:0 C18:1, cis
80
70
C6:0
C18:1, trans
C11:0
C14:1 C15:0 C15:1
C20:2 C21:0 C20:3 n6 C20:3 n3 C20:4 n6
C18:2, trans C18:2, cis C18:3 n6 C18:3 n3
C22:0
60
C13:0
C20:0
C18:3 n3
C17:0
C16:1
C20:1n9
C20:5 C22:1
50
C17:0 C17:1
C4:0
C22:2 C23:0
C24:0
20 22
30
20 4 6 8 10 12 14 Time (min) 16 18
Figure 4.
GC-FID analysis of FAMEs standard mixture on a 60 m 0.25 mm ID, 0.15 m DB-23 column using Method 2, Table 3.
C22:6
40
C24:1
Using these conditions, all compounds in the standard mixture are well resolved. Important is the separation of the cis/trans isomers and the separation of EPA (19.15 min) and DHA (22.38 min) components. This method is very useful for the analysis of fatty acid in complex mixtures, and especially for the determination of omega-3 fatty acids (such as EPA and DHA). An example of the separation obtained for a mixture of polyunsaturated fatty acids from a marine source is given in Figure 5. EPA and DHA can easily be detected and quantified.
C20:5 (EPA)
DB-23
C22:6 (DHA) C24:0
20
C14:0
100
90
80
C16:0
60
C18:2, cis
70
C18:1, cis
50
C18:1, trans
C17:0
C18:3 n3
C20:1n9
Norm.
C16:1
C18:2, trans
C22:1
C21:0 C20:2 C20:3 n6 C20:4 n6 C20:3 n3 C22:0
40
20 8 10 12 14 16 18 22 24
Figure 5.
GC-FID analysis of unsaturated fatty acid mixture from marine origin on a 60 m 0.25 mm ID, 0.15 m DB-23 column using Method 2, Table 3.
The separation of the 37-compound mixture on the highly polar HP-88 column is shown in Figure 6.
C12:0
C18:0
C20:0
30
C14:1 C15:0 C15:1
C18:3 n6
C23:0 C22:2
C17:1
C24:1
Norm.
C12:0
45
C14:0
C16:0
50
C10:0
40 35
C220 + C 20:3 n6
HP-88
C18:1 cis
C18:0
C11:0
C13:0
C20:0
30 25
C8:0
C20:3 n3 C22:1 C20:4 C23:0 C22:2 C20:5
C14:1 C15:0 C15:1
C24:0 C24:1
30
C18:1 trans
C18:2 trans C18:2 cis
C6:0
15 10 5 5 10 15
C4:0
Time (min)
20
25
Figure 6.
GC-FID analysis of 37-component FAMEs standard mixture on a 100 m 0.25 mm ID, 0.2 m HP-88 column using Method 3A, Table 4A.
Again a quite good separation is obtained, except for the separation of C22:0 and C20:3 (n-6) that coelute at 24.7 min. Using this column, however, the separation of cis- and trans-isomers is excellent. This is illustrated by the separation of a standard mixture containing five C18:1 isomers. The cis- and trans- positional isomers are well separated, as shown in Figure 7.
Norm. 90
c6
c11
80
HP-88
c9
70
60
18:1 t6 t11
50
40
30
20
10 15 16 17 Time (min) 18 19 20
Figure 7. 8
GC-FID analysis of C18:1 isomers on a 100 m 0.25 mm ID, 0.2 m HP-88 column using Method 3A, Table 4A.
C22:6
20
C18:3 n6 C18:3 n3 C20:1 C21:0 C20:2
C16:1 C17:0 C17:1
Equally good separation is obtained for four C18:2 isomers (trans-trans, cis-trans, trans-cis and cis-cis), as shown in Figure 8.
C18:2 tt
Norm. 90
HP-88
80
70
60
50
ct
tc
40
30
cc
20
10 15 16 17 18 19 Time (min) 20 21 22
Figure 8.
GC-FID analysis of C18:2 isomers on a 100 m 0.25 mm ID, 0.2 m HP-88 column using Method 3A, Table 4.
A comparison between a DB-23 and a HP-88 column was made for the separation of a highly hydrogenated oil. Due to the hydrogenation process, all possible positional and geometrical (cis-trans) isomers are formed. The sample was analyzed on a DB-23 column and an HP-88 column respectively, both isothermally at 180 C. the chromatograms are compared in Figure 9 (details of C18:1 elution window). Although Figure 4 shows baseline separation of trans-C18:1(n9) and cis-C18:1 (n9) in the 37-component standard, Figure 9 demonstrates that for real samples containing several C18:1 isomers, the cis-trans separation with the HP-88 is a preferred column choice.
pA 100 80 60 40 20 0 14.5 pA 35 30 25 20 15 10 5 15.5 16.0 16.5 17.0 Time (min) 17.5 18.0 18.5 15.0 15.5 Time (min) 16.0 16.5
DB-23
18:1 trans 18:0
18:2
17.0
min
HP-88
18:1 cis
18:0
18:1 trans
Figure 9.
Comparison of the separation of C18:1 isomers from a hydrogenated oil obtained on a 60 m x 0.25 mm ID, 0.15 m DB-23 column (upper window) and on a 100 m 0.25 mm ID, 0.2 m HP-88 column. Both analyses were performed at 180 C isothermally.
The application of the HP-88 is demonstrated by the analysis of a partially hydrogenated rapeseed oil. The separation of the trans fatty acids is clearly illustrated in Figure 10. The valley between transand cis-isomers can easily be determined. Also the other trans-isomers (C18:2 and C18:3) can be detected.
10
Norm. 28
26
HP-88
24
22
1. trans-C18:1 2. trans-C18:2 3. trans-C18:3
20
18
1
16
14
2
12
10 5.0 7.5 10.0 12.5 15.0 Time (min) 17.5 20.0 22.5 25.0
Figure 10. GC-FID analysis of FAMEs from a partially hydrogenated rapeseed oil on a 100 m 0.25 mm ID, 0.2 m HP-88 column using Method 3A. (See Table 4).
The same column can also be used for quality control of olive oil according to EC regulation 2568/91 [5]. Using the method described in Table 4 (Column A Method A), the analysis time for the 100-m column is approximately 35 min using hydrogen as carrier gas. For the QC analysis of olive oil, a 60-m column can also be used (Table 4 column B). Using helium as carrier gas and a different temperature program (that is, oven temperature B, Table 4), an excellent separation is obtained in less than 20 min, as shown in Figure 11. The obtained separation fully conforms to the EC regulation [5].
11
Norm. 14 13 12 11 10 9 8 7 6 C18:1 trans 5 0 2 4 6 8 10 12 14 16 18 20 C18:3 C16:0 C18:0 C18:1 C18:2
HP-88
C16:1
C20:0
C20:1
Figure 11. GC-FID analysis of olive oil FAMEs on a 60 m 0.25 mm ID, 0.2 m HP-88 column using Method 3B, Table 4.
Conclusions
Three types of stationary phases can be used for the analysis of FAMEs. 1. A DB-Wax column, is useful for the analysis of classical edible oils and fats, including the determination of butyric acid in milk fat. Using this column, however, no separation of cis-trans isomers is obtained. 2. A medium polar DB-23 cyanopropyl column is excellent for the analysis of complex FAME mixtures, including fish oils, allowing the determination of omega 3 fatty acids such as EPA and DHA. Partial cis-trans separation is obtained. 3. For the most demanding separation of cis-trans separation, an HP-88 column is recommended. This column is also the column of choice for olive oil QC analysis.
3. IUPAC, Standard methods for Analysis of Oils, Fats and Derivatives, Blackwell Scientific Publications, IUPAC Method 2.301. 4. International Standard ISO 15304, version 2003-05-15 (ISO 15304/2002) (can be ordered from www.iso.org) 5. EU regulation 2568/91, Official Journal EU, document L248, 5/9/1991. 6. EC, Bureau of Reference (BCR), Brussels, Belgium (see catalog at www.irmm.jrc.be)

References
1. W. W. Christie, Gas Chromatography and Lipids, A Practical Guide, (1989), The Oily Press, Ayr, Scotland (ISBN 0-9514171-O-X). 2. AOAC Official Methods of Analysis (2000), method Ce 266.
The HPLC Preparative Scale-Up of Soybean Phospholipids Application
Food and Flavors
Authors
Cliff Woodward and Ron Majors Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA
development of an analytical normal phase HPLC separation of the most common phospholipids found in crude soy lecithin extracts and the possibility of performing a linear scale-up to preparative columns filled with the same packing.
Properties of Phospholipids
From a chromatographic viewpoint, phospholipids are in a class of polar lipids and are amphipathic. They contain an ionic portion due to the phosphate moiety and they may contain additional hydrophilic or ionic portions due to the presence of choline, serine, ethanolamine, myo-inositol and/ or glycerol functionality. All phospholipids contain a hydrophobic portion due to the presence of the fatty acid moieties and their complexity results from the two fatty acids on the 1,2-diacylglycerol portion. These fatty acid moieties may be the same or different. The main phospholipids present in soybeans along with their abbreviations are shown in Table 1. In a typical soybean extract, about half of the extractables are phospholipids with neutral lipids, mainly triglycerides, and carbohydrates making up the remainder. Since phospholipids are generally very complex and difficult to prepare in a pure form, most commercially prepared mixtures contain varying amounts of these extractables. Phospholipids generally have low water solubility and can be dissolved in nonpolar solvents. Since their ultraviolet (UV) absorbance arises only from the double bond in the fatty acid portion, they provide low response with UV detectors. Refractive index-, evaporative light scattering- or mass spectrometric-detectors are, therefore, more frequently used.
Abstract
Classes of major soybean phospholipids can be separated from a crude mixture, under both analytical and preparative conditions, using a matched set of columns, normal phase conditions, and a small amount of volatile salt added to the mobile phase. At higher concentrations of phosphatidylcholine, however, two distinct peaks, whose areas changed with the mass injected, were observed. These two peaks were determined by mass spectrometry to be the same or, at least, very similar.
Introduction
Phospholipids have an important role in animal and plant systems since they are commonly found in cellular and intracellular membranes where they fulfill the role of primary structural components. Phospholipids are also widely used in the food, cosmetic and industrial manufacturing industries and account for 80% of worldwide plant oil production. The structure of a typical phospholipid is depicted in Figure 1. Phospholipids are the major components of soybean lipids and are often desired in a pure form as standards for chromatography and other studies. Preparative amounts are therefore needed and HPLC is often the first choice for this purpose. This application note describes both the
Glycerol C O CH2 O C O CH O Fatty acid O

_
Phosphatidylcholine (PC)
Choline CH3 CH3
CH2 O P O CH2 CH2 N+ CH3 O Phosphate
Hydrophobic tail
Polar head group
Figure 1. Table 1.
Structure of typical phospholipid. Important Phospholipids in Commercial Plant Oil Extracts Abbreviation PE PS PI PC
Phospholipid Phosphatidylethanolamine Phosphatidylserine Phosphatidylinositol Phosphatidylcholine
UV [1]. In their work, they used a ternary gradient system. However, since our preparative chromatograph employed a binary pumping system, we had to adapt our gradient conditions for both the analytical and preparative chromatography accordingly.
Chromatographic Separation of Phospholipids

Because of their chemical properties, phospholipids may be separated by both normal phase and reversed-phase chromatography. For preparative scale-up from analytical dimensions, the use of silica normal phase chromatography has advantages when compared to reverse phase chromatography (RPC): 1) The mobile phases are more volatile; thus, collected fractions can easily be recovered. 2) The viscosity of the common normal phase solvents is low, allowing higher flow rates to be employed, thereby improving throughput. 3) Using the adsorption mechanism, class separations of major phospholipids are easier to perform compared to RPC. 4) In normal phase work, since phospholipids absorb poorly in the UV, detection is better accomplished using the evaporative light scattering detector (ELSD) that allows gradient systems to be employed and provides excellent sensitivity in organic mobile phases. Row and Kang demonstrated successful separation of phospholipids using normal phase chromatography but they operated at 208- or 210-nm in the
Evaporative Light Scattering as a Detection Principle

The ELSD detects compounds based somewhat on a compound's mass, and not on its spectroscopic characteristics. It is nearly universal and it responds to all compounds as long as they are less volatile than the mobile phase. As seen in Figure 2, the chromatographic flow stream enters the heated nebulizer and is nebulized using a stream of dry nitrogen; the resultant mist containing the analyte is sequentially directed through the heated evaporator where is loses the solvent and becomes a particle, which is then directed through the polychromatic light beam. The particles cause photon scattering as they traverse the path of the light beam. The signal generated is proportional to the analyte particle size present. The solvent is evaporated and does not interfere with the analyte measurement. The presence or absence of chromophores, fluorophores, or electroactive groups has no impact on the signal. The ELSD can be optimized for analyte sensitivity and baseline stability by adjustment and optimization of evaporation and nebulization temperatures. For preparative applications, the upper flow rate capability of the ELSD is 4 mL/min, and thus requires post-column splitting of the chromatographic effluent.
Detection chamber
Light source
Heated nebulizer Heated evaporating chamber
Syphon
Figure 2.
Schematic of ESA Chromachem ELSD.
Scaling Up from Analytical to Preparative Chromatography

In scaling up to a preparative separation, it is recommended that initial method development use an analytical-sized column. Since flow rates and sample injection sizes are lower than for prep, solvent costs are lowered and less sample is used, which is important when sample mass is limited or for samples that are relatively expensive. Once the separation is optimized using an analytical (or scalar) column, linear scale-up to a preparative column of the same materials is easily achieved. Initially, we adjusted our gradient to approximate the one used by Row and Kang [1]. Using 2.5% methanol (MeOH), 2.5% isopropanol (IPA), and 95% hexane as solvent A, and using 40% IPA and 60% MeOH as solvent B, we achieved a successful separation of the four standard phospholipids but found that our retention times often varied from run to run. In earlier normal phase work [2], Christie found that the addition of a small amount of ionic material to the most polar solvent, in his case, isopropanol-water, gave better resolution and
extended the column life. However, several of his suggested ionic additives were nonvolatile. Since the ELSD requires the use of a volatile mobile phase, we chose a low concentration of ammonium acetate (10-mM) but added it to both mobile phase constituents. Higher concentrations could be used to give improved retention behavior, but at >10-mM levels there is danger of salt precipitation, and should therefore be avoided. Using this hexane-methanol-isopropanol-ammonium acetate (NH4Ac) gradient system allowed us to fine tune the separation, especially relative to isocratic methods of the past. The final set of chromatographic conditions are depicted in Table 2. Figure 3 shows the optimized separation on a silica-gel scalar column as a function of the injected amount of phospholipid standards: PE, PS, PI, and PC. The small amount of NH4Ac, a volatile salt, gave much better retention behavior for PC, a zwitterion with a quaternary amine functionality (see Figure 1). The salt addition also assured minimal retention variation with load at low injection mass. Surprisingly, at higher injection masses, we observed two peaks for PC.
Table 2.
Final Set of Experimental Conditions Mobile phase: Flow rate: Agilent Prep SIL Scalar Column, 4.6 150 mm, 10 m A = 95:2.5:2.5% hexane-isopropanolmethanol B = 40:60% isopropanol-methanol Both solvents contain 10-mM ammonium acetate 1.5 mL/min Time, min 0 20 20.2 25 25.1 %B 0 18.7 100 100 0 Gradient: Temperature: Detection Detector: Temp. nebulizer: Temp. evaporator: Filter: Gas pressure: Attenuation: Sample Dissolved in: Analytical: Prep: CHCl3:Hexane 2:1 5 mg/mL 25 mg/mL Same as above 31.9 mL/min Same as above Ambient Chromachem ELSD, (ESA, Cambridge, MA) 45 C 65 C High 23 psi 1
Chromatography Analytical runs: Column: Mobile phase:
Flow rate: Gradient:
Temperature: Preparative runs: Column:
Ambient Agilent Prep SIL Column, 21.2- 150 mm, 10 m

PE PS PI
mV 2000 1750
PC
1500 1250
Z2
1000 750 500
Z1 500 g 200 g 100 g 50 g 25 g
250 0 5 10 15 20 25
30
35
min
Figure 3.
Agilent prep 4.6 150 mm, 10 m Mixture of soy phospholipid standards. - Loadability.
Study of PC Elution Behavior

As stated above, at higher injection masses, in Figure 3, we observed two peaks for PC, one eluting at a 16.1 minute (PC-Z1) and the second 21.8 minutes (PC-Z2). As sample mass increased (above 75 g on the 4.6 150-mm column), the latter peak increased while the former decreased but the total area increased as expected. As can be
seen in Figure 4, with increasing sample mass, the peak areas for all of the phospholipids, including PC, increased. Typical of the ELSD, response was nonlinear, especially at the lowest concentrations (less than 50 g); however, since the double peak phenomenon occurred with PC only, we suspected that it had something to do with the zwitterion functionality. We, therefore, investigated this aspect more thoroughly.
80000 70000
PE
60000
PI
50000
PS
Area counts 40000 30000 20000 10000 0 -10000 0 100 200 300 Micrograms injected 400 500 600
PC (Total) PC-Z2 PC-Z1
Figure 4.
Area counts for ELSD versus micrograms injected (analytical column, 4.6 150 mm).
Using liquid chromatography/mass spectrometry (LC/MS), we attempted to determine the structure of the two peaks observed with PC. Since in normal phase chromatography ionization could be a problem, we first investigated MS ionization approaches. MS sensitivity is related to the formation of stable ions in the source. Using electrospray ionization, a minimal signal was observed; so, we resorted to atmospheric pressure photometric ionization (APPI), a more useful ionization technique for neutral organic compounds. Since PC
is zwitterionic, its natural ionic character allowed it to be detected in either positive or negative APPI. To enhance sensitivity, acetone and toluene dopants were added post-column as a 50:50 mixture. From other studies [3] it is known that soybean PC consists of multiple phospholipid entities with varying fatty acid content. We were able to assign certain m/z values to individual PCs on the basis of the fatty acid portion of the 1,2diacylglycerol functionality. Figure 5 shows the raw, unsmoothed mass spectral data of the two PC peaks (PC-Z1 and PC-Z2).
600000
500000
400000
C18:2, C18:2 PC (dilinoleoyl) MW 781 C16:0, C18:2 PC (palmitoyl/lineoleoyl) MW 757 C18:1, C18:2 PC (oleoyl/linoleoyl) MW 783 C18:2, C18:3 PC (lineoleoyl/linolenoyl) MW 779
300000
PC-Z1 PC-Z2
200000
100000
0 16 18 20 22 24 min
Figure 5.
Overlay of the most common phosphatidylcholine (lecithin) (PC) species present in crude soy lecithin Extracted ions from LC/MS (unsmoothed data). 5
Clearly, one can determine that there was partial resolution by the type of fatty acid moiety present. Note that the same species are present in both peaks PC-Z1 and PC-Z2. Furthermore, as can be seen in Table 3, if one explores the peak height ratios of the individual PCs within the PC-Z1 and PC-Z2 umbrellas, they are similar enough to deduce that the peaks are the same. In addition, based on fatty acid ratio studies of the individual PCs in purified PC as determined by gas chromatography [3], for a first approximation, the fatty acid composition of the PCs depicted in Figure 5 are similar. This observation also led us to believe that these two PC peaks were the same compound (Table 3). Our speculation is that the two peaks are a result of the competitive silanol binding between the positively charged ammonium ion in the mobile phase modifier and the positively charged portion of PC. At low PC concentrations, the ammonium binds more to the silica silanol groups and the concentration of unbound PC-Z1 is insufficient to displace it. As the concentration of PC increases, however, some of the PC adsorbs to the silanol sites in competition with the ammonium ion. This strongly held PC (identified as PC-Z2) elutes only at very high concentrations of IPAMeOH. As larger amounts of crude soy lecithin are
injected, the available PC increases so that more binds to the surface, resulting in an increase in its elution peak, PC-Z2. If one estimates the total amount of available silanol groups (assuming 8-moles/m2 for a totally hydroxylated surface) on the surface of the packing inside of the column, the number of moles of PC potentially bound to the surface at saturation (around 600 moles) is consistent with the calculated level of binding sites available.
Preparative Studies of Crude Soy Lecithin

Figure 6 depicts the preparative scale-up of the same phospholipid standard to a 21.2-mm 150-mm column packed with the same material as the scalar column. The flow rate was increased to 31.88 mL/min to achieve the same linear velocity and separation time as under the analytical conditions. The same relative amounts injected onto the preparative column allowed as much as 20 mg of total phospholipid before the ELSD detector showed signs of overload as noted in the upper chromatogram. Note that the same PC-Z1 and PC-Z2 peaks can be observed in the chromatogram of Figure 6.
Table 3.
Table of Relative Extracted Ion Currents Weight,% typical soybean* 47.5 30.0 10.5 6.1 Relative weight.% 50 31 13 6 Relative height PC-Z1 39 31 28 3 Relative height PC-Z2 45 24 21 9
Fatty acid composition C18:2/ C18:2 PC (dilinoleoyl) C16:0/ C18:2 PC (palmitoyl/ lineoleoyl) C18:1/ C18:2 PC (oleoyl/ linoleoyl) C18:2/ C18:3 PC (lineoleoyl/ linolenoyl)
*B. Herslof, U. Olsson, and P. Tingvall in I. Hanin and G. Pepeu,Eds. (1990). Phospholipids: Biochemical, Pharmaceutical, and Analytical Considerations, Plenum Press, New York, pp. 295298.
mV 2000
PE
PS PI PC
1750 1500 1250 1000 750
Z1
Z2
20 mg
500
10 mg 4 mg 2 mg 1 mg
250 0 0 5 10 15 20 25
30
35
min
Figure 6.
Agilent prep 21.2 150 mm, 10-m Mixture of soy phospholipid standards. - Loadability.
Finally, in Figure 7, we investigated the scale-up of a 20-mg injection of a crude soy lecithin sample which, by overlaying the standards run under the same conditions, revealed the presence of three phospholipids, including PE, PI, and PC. No PS is seen in this crude extract.
Norm.
Standard run conditions

1200
1000
Crude soy PC
800
Pure soy PC
600
Pure soy PI
400
Pure soy PS
200
Pure soy PE
0 0 5 10 15 20 25 min
Figure 7.
Agilent prep SIL, 21.2 150 mm,10 m Crude soy lecithin (PC) and pure individual standards.
Conclusion
These experiments show that the classes of major soybean phospholipids can be successfully separated from a crude mixture, under both analytical and preparative conditions, using a matching set of columns, normal phase conditions, and a small amount of volatile salt added to the mobile phase to cut down on possible electrostatic interactions. At higher concentrations of PC, however, two distinct peaks, whose areas change with mass injected, are observed. By mass spectrometric measurements of the eluted peaks, these two peaks were determined to be the same or, at least, very similar. It is suspected that PC is strongly binding to the silica silanol group in competition with the ammonium acetate, a mobile phase modifier, and is eluting at high gradient solvent polarity.
Acknowledgement
We would like to thank ESA for the loan of the Chromachem ELSD detector for the duration of this study.
References
1. Kyung Ho Row and Duk Hui Kang, (2003) Amer. Biotech. Lab., 21(9) 4043. 2. W. M., Christie, (1986) J. Chromatogr. 361, 396399. 3. B. Herslof, U. Olsson, and P. Tingvall in I. Hanin and G. Pepeu, Eds. (1990) Phospholipids: Biochemical, Pharmaceutical, and Analytical Considerations, Plenum Press, New York, pp. 295298.

Improving the Analysis of Fatty Acid Methyl Esters Using Retention Time Locked Methods and Retention Time Databases
Application
Food
Authors
Frank David Research Institute for Chromatography Pres. Kennedypark 20, B-8500 Kortrijk, Belgium Pat Sandra University of Gent Krijgslaan 281 S4, B-9000 Gent, Belgium Philip L. Wylie Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA
acid methyl esters is performed on a DB-Wax column using gas chromatography/flame ionization detector or gas chromatography/mass spectrometry. In the second method, a DB-23 column is used. Retention time and mass spectral libraries are available for both methods. Retention time locking allows easy peak identification, easy exchange of data between instruments (gas chromatography/flame ionization detector, gas chromatography/mass spectrometry, different labs), and avoids the need to modify the retention times in calibration tables after column maintenance or column change.
Introduction
The analysis of fatty acid methyl esters (FAMEs) is used for the characterization of the lipid fraction in foods and is one of the most important applications in food analysis. Lipids mainly consist of triglycerides, which are esters of one glycerol molecule and three fatty acids. Most edible fats and oils are composed largely of 12- to 20-carbon fatty acids [lauric acid (dodecanoic acid) to arachidic acid (eicosanoic acid)]. Besides linear saturated fatty acids, branched, mono-unsaturated, di-unsaturated, and higher unsaturated fatty acids can occur. An overview of the most important fatty acids and their common abbreviations appears in Table 1.
Abstract
The analysis of fatty acid methyl esters is a very important application in food analysis. Fatty acid methyl esters are normally analyzed on columns coated with polar stationary phases such as polyethylene glycols or cyanopropyl silicones, allowing separation of fatty acids according to carbon number and according to the degree of unsaturation. Two retention time locked methods are presented in this application note. In the first method, the analysis of fatty
Table 1. Fatty acid
Fatty Acids, Common Names and Abbreviation Used Common Name Butyric acid Caproic acid Lauric acid Myristic acid Palmitic acid Palmitoleic acid Stearic acid Oleic acid Elaidic acid Linoleic acid Linolelaidic acid alpha-Linolenic acid gamma-Linolenic acid Arachidic acid Simplified Abbreviation1 4:0 10:0 12:0 14:0 16:0 16:1 n-7 18:0 18:1 n-9 t18:1 n-9 18:2 n-6 t18:2 n-6 18:3 n-3 18:3 n-6 20:0 20:1 n-9 20:2 n-6 20:3 n-3 20:3 n-6 20:4 n-6 20:5 n-3 22:0 22:1 n-9 22:4 n-6 22:6 n-3 24:0 24:1 n-9 Abbreviation specifying cis and trans bonds1 4:0 10:0 12:0 14:0 16:0 9c-16:1 18:0 9c-18:1 9t-18:1 9c12c-18:2 9t12t-18:2 9c12c15c-18:3 6c9c12c-18:3 20:0 11c-20:1 11c14c-20:2 11c14c17c-20:3 8c11c14c-20:3 5c8c11c14c-20:4 5c8c11c14c17c-20:5 22:0 13c-22:1 7c10c13c16c-22:4 4c7c10c13c16c19c-22:6 24:0 15c-24:1
Butanoic acid Decanoic acid Dodecanoic acid Tetradecanoic acid Hexadecanoic acid Hexadecenoic acid Octadecanoic acid cis-9-Octadecenoic acid trans-9-Octadecenoic acid all cis-9,12-Octadecadienoic acid ll trans-9,12-Octadecadienoic acid all cis-9,12,15-Octadecatrienoic acid all cis -6,9,12-Octadecatrienoic acid Eicosanoic acid cis-11-Eicosenoic acid all cis -11,14-Eicosadienoic acid all cis -11,14,17-Eicosatrienoic acid all cis -8,11,14-Eicosatrienoic acid all cis -5,8,11,14-Eicosatetraenoic acid all cis 5,8,11,14,17-Eicosapentaenoic acid Docosanoic acid cis -13-Docosenoic acid all cis -7,10,13,16-Docosatetraenoic acid all cis 4,7,10,13,16,19-Docosahexaenoic acid Tetracosanoic acid cis -15-tetracosenoic acid
1
Dihomogammalinolenic acid Arachidonic acid EPA Behenic acid Erucic acid DHA Lignoceric acid Nervonic acid
Several different versions of fatty acid nomenclature and structural abbreviation have been used in the past. For discussions of past and currently-accepted nomenclature, the following web sites are recommended: http://www.ajcn.org/misc/lipid.shtml http://www.chem.qmul.ac.uk/iupac/lipid/ http://www.aocs.org/member/division/analytic/fanames.htm http://www.cyberlipid.org/index.htm
For the characterization of the lipid fraction, the triglycerides are hydrolyzed (saponified) into glycerol and free fatty acids. Although the free fatty acids can be analyzed directly on polar stationary phases (such as a FFAP column), more robust and reproducible chromatographic data are obtained if the fatty acids are derivatized to the FAMEs. Several methods are available for derivatization, which requires hydrolysis of the glycerides and methylation of the resulting fatty acids [1]. These easy-to-use methods do not require expensive reagents or equipment. Two useful methods are described in the Experimental section. After preparation of the FAMEs, the FAMEs are separated according to carbon number (number of carbon atoms in the fatty acid chain, not including the methyl ester carbon) and according to the degree of unsaturation. Moreover, the position of the double bond(s) and the geometric configuration (cis/trans) are also important parameters and
2
their determination adds additional information to the characterization of the lipid fraction in food. In this application note, two methods are described for the GC analysis of FAMEs. The method choice depends both on sample complexity and the degree of fatty acid characterization that is required (Figure 1). Method 1 uses a DB-Wax column that separates FAMEs from C4 (butyric acid) to C24 (lignoceric acid) according to carbon number and unsaturation. On this column, no separation of cis and trans isomers is obtained, and for complex mixtures (such as fish oils), some FAMEs are not resolved. However, the separation of FAMEs on polyethylene glycol columns is widely used and can be applied to the characterization of classical samples such as vegetable oils (corn oil, maize oil, olive oil, soybean oil, and so on). For certain applications, animal fats can also be analyzed using the Agilent DB-Wax column. An important application, for instance, is the analysis of butyric
acid in milk fat. The concentration of butyric acid in milk is an important indicator of its quality. This determination is very important in milk and dairy analysis and in the analysis of chocolate products. All these applications can be performed on the Agilent DB-Wax column using method 1.
FAME Analysis
times in the calibration table remain unchanged even after column maintenance or column change (after re-locking the method). Moreover, gas chromatography/flame ionization detection (GC/FID) and gas chromatography/mass spectrometry (GC/MS) methods can be scaled [2, 3], so that retention times obtained on the two different systems match very closely. For GC/MS, a custom spectral library was created so that data files can be screened using spectra together with their locked retention times [4].
Experimental
DB-Wax DB-23
Samples Reference standards of FAMEs can be obtained from different sources as solutions or as neat compounds. For analysis, the standards are typically dissolved in hexane at a 0.01 to 0.1 % (w/v) concentration. For method and instrument check-out, a 37-component mixture (Supelco number 18919) was used. The mixture was purchased as a 100-mg neat mixture, containing C4 to C24 FAMEs (2 to 4% relative concentration). The whole sample was diluted in 10 mL of hexane (final concentration = 0.2 to 0.4 mg/mL per FAME). Oil and fat samples can be prepared using any one of several different methods [1]. The following sample preparation methods were tested. Sample preparation method 1 [5] : Weigh 100-mg sample in a 20-mL test tube (with screw cap) or a reaction vial. Dissolve the sample in 10 mL of hexane. Add 100 L of 2N potassium hydroxide in methanol (11.2 g in 100 mL). Close the tube or vial, and vortex for 30 seconds. Centrifuge. Transfer the clear supernatant to a 2-mL autosampler vial. Sample preparation method 2 [6]: Weigh 50-mg sample in a 20-mL test tube (with screw cap) or a reaction vial. Add 2 mL of 2N sodium hydroxide in methanol (8 g in 100 mL). Close the tube or vial, and heat at 80 C for 1 hour. Allow to cool. Add 2 mL of a 25% borontrifluoride solution in methanol (Sigma-Aldrich, cat. no. 13,482-1, 50% solution in methanol, to be diluted to 25% in methanol). Close the tube or vial, and heat again for 1 hour at 80 C. Allow to cool. Add 5 mL of water and 5 mL of hexane. Shake well. Allow the phases to separate (or centrifuge). Transfer the clear supernatant to a 2-mL autosampler vial. Both methods performed equally well.
Edible oils Milk fat No cis/trans Some overlap
Complex mixtures EPA and DHA analysis cis/trans separation
Figure 1.
Overview of method selection for FAME analysis.
For the analysis of more complex samples, such as fish oils, additional resolution of FAMEs is obtained using a capillary column coated with a cyanopropyl stationary phase (method 2). On an Agilent DB-23, highly unsaturated fatty acids, such as all cis 5,8,11,14,17-eicosapentenoic acid methyl ester (EPA, 20:5 n-3) and all cis 4,7,10,13,16, 19-Docosahexenoic acid methyl ester (DHA, 22:6 n-3), are separated from other FAMEs. Separation of cis and trans isomers is also possible on the cyanopropyl column due to the stronger interaction of cis isomers with the cyano-dipole. This causes the trans isomers to elute before the cis isomers. An increasingly important food analysis is the determination of trans fatty acids, which can be performed using the DB-23 column with the conditions described in method 2. Both methods are retention time locked using methyl stearate as the locking compound. Retention time locking (RTL) allows the analyst to obtain virtually identical retention times on any GC, independent of the inlet, injection technique, or detector used [2, 3, 4]. Because RTL reproduces retention times so accurately, FAME identification can be done based on absolute retention times. It is unnecessary to have all of the FAME standards available because peak identification is possible using Agilent's published retention time database. An additional benefit of RTL is that retention
Analytical Conditions The analyses were performed on an Agilent 6890 GC equipped with an FID or on an Agilent 6890/5973 GC/MSD system. Automated split injection was performed using an Agilent 7683 automatic sampler. The instrumental configuration and analytical conditions are summarized in Table 2 (DB-Wax column) and Table 3 (DB-23 column). For both methods, methyl stearate was used as the
locking standard. The retention time for methyl stearate was locked at 14.000 min. When duplicating this method, the column head pressure can be set to the pressures indicated in Tables 2 and 3 (nominal pressure). Then the RTL calibration runs can be performed (at -20%, -10%, +10% and +20% of the nominal pressure)[4]. The retention time versus head pressure curve is then automatically calculated and stored in the method.
Table 2.
DB-Wax Method
Instrumentation Chromatographic system Inlet Detector Automatic sampler Liner Column Experimental conditions GC/FID Inlet temperature Injection volume Split ratio Carrier gas Head pressure 250 C 1 L 1/50 Hydrogen Methyl stearate is retention time locked to 14.000 min Constant Pressure mode (pressure approximately 53 kPa at 50 C, 36 cm/s at 50 C) 50 C, 1 min, 25 C/min to 200 C, 3 C/min to 230 C, 18 min 280 C Hydrogen: Air: Helium make-up gas: 40 mL/min 450 mL/min 30 mL/min Agilent 6890 GC Split/splitless FID or Agilent 5973 MSD Agilent 7683 Split liner (part no. 5183-4647) 30-m 0.25-mm id 0.25-m DB-Wax (part no. 122-7032)
Oven temperature Detector temperature Detector gases
Experimental conditions GC/MS Inlet temperature Injection volume Split ratio Carrier gas Head pressure 250 C 1 L 1/50 Helium Methyl stearate is retention time locked to 14.000 min Constant Pressure mode (pressure approximately 55 kPa at 50 C, 36 cm/s at 50 C) 50 C, 1 min, 25 C/min to 200 C, 3 C/min to 230 C, 18 min 280 C Scan (40 to 500 amu), threshold 100 MS quad 150 C MS source 230 C Solvent delay: 2 min
Oven temperature Detector temperature MSD Parameters
Table 3.
DB-23 Method
Instrumentation Chromatographic system Inlet Detector Automatic sampler Liner Column Experimental conditions GC/FID Inlet temperature Injection volume Split ratio Carrier gas Head pressure 250 C 1 L 1/50 Helium Methyl stearate is retention time locked to 14.000 min Constant Pressure mode (pressure approximately 230 kPa at 50 C, 33 cm/s at 50 C) 50 C, 1 min, 25 C/min to 175 C, 4 C/min to 230 C, 5 min 280 C Hydrogen: Air: Helium make-up gas: 40 mL/min 450 mL/min 30 mL/min Agilent 6890 GC Split/splitless FID or Agilent 5973 MSD Agilent 7683 Split liner (part no. 5183-4647) 60-m 0.25-mm id 0.15-m DB-23 (part no. 122-2361)
Oven temperature Detector temperature Detector gases
Experimental conditions GC/MS Inlet temperature Injection volume Split ratio Carrier gas Head pressure 250 C 1 L 1/50 Helium Methyl stearate is retention time locked to 14.000 min Constant Pressure mode (pressure approximately 180 kPa at 50 C, 33 cm/s at 50 C) 50 C, 1 min, 25 C/min to 175 C,4 C/min to 235 C, 5 min 250 C Scan (40 to 500 amu), threshold 100 MS quad 150 C, MS source 230 C Solvent delay: 3.5 min
Oven temperature Transfer line MSD parameters

A typical chromatogram for the analysis of the FAME reference standard (obtained on the DB-Wax column using method 1) is shown in Figure 2. A good separation is obtained, except for the following compounds: cis and trans 18:1 co-elute at 14.38 min, cis and trans 18:2 co-elute at 15.13 min, 20:3 n-6 and 21:0 co-elute at 19.44 min, and 22:6 and 24:1 co-elute at 30.73 min. However, this separation is sufficient for most classical oil and fat characterization. Butyric acid elutes at 2.85 min and can be determined in milk fat using the same
method. Because the GC/MS and GC/FID instruments were locked, virtually identical chromatograms were obtained on both systems. A comparison between the GC/FID and GC/MS profiles is shown in Figure 3. Although different outlet pressures (ambient versus vacuum) and different carrier gases (hydrogen in GC/FID and helium for GC/MS) were used, the correspondence between the two chromatograms was very good. Peaks detected in the GC/FID trace can easily be located in the GC/MS trace and identification is possible based on retention times alone or in combination with mass spectra.
5
10:0 12:0
Norm.
8:0
100
16:0
FID1 A, (FAMINNO\INJ00002.D) 6:0 18:1 c+t 18:0 13:0 18:2 c+t 20:0 20:1 20:2 20:3 n-6 + 21:0 20:3 n-3 20:4 20:5 22:0 22:1 18:3 n-6 18:3 n-3 15:1 16:1 17:0 17:1
90
80
70
4:0
60
11:0
50
14:1 15:0
14:0
30
20 0 5 10 15 20 25 30 minute
Figure 2.
GC/FID analysis of FAMEs standard mixture on a 30-m 0.25-mm id 0.25-m DB-Wax column (part no. 122-7032) using method 1 (see Table 2).
Abundance 2.2e + 07 2e + 07 1.8e + 07 1.6e + 07 1.4e + 07 1.2e + 07 1e + 07 8000000 6000000 4000000 2000000 0 -2000000 -4000000 -6000000 -8000000 -10000000 -12000000
Signal: INJ00002.D\FID1A.CH (*)
22:2 23:0
GC/FID
GC/MS
TIC: WAXTEST.D (*)
6.00
8.00
10.00
12.00 Time
14.00
16.00
18.00
24:0
40
20.00
Figure 3.
Comparison of GC/FID and GC/MS chromatograms obtained on a 30-m 0.25-mm id 0.25-m DB-Wax column (part no. 122-7032) using method 1 (see Table 2).
22:6 + 24:1
The retention time locked method on the DB-Wax column was applied to the analysis of two certified reference samples (CRM 162, a soy-maize blend oil and CRM 164, a milk fat)[7]. Both samples were prepared according to sample preparation method 2. The resulting chromatograms are shown in Figure 4 (soy-maize oil blend) and in Figure 5 (milk fat). The peaks were automatically identified using the RTL FAMEs retention time database. The quantitative results are summarized in Tables 4 and 5.
Norm. 100 90 80 70 60
Figure 4 shows a classical fatty acid profile normally obtained for edible oils. Figure 5 shows the fatty acid profile typical for milk fat. Butyric acid elutes at 2.85 min and is easily detected. Very good reproducibility is obtained. The standard deviation of the relative areas is smaller than 1% in all cases. Also the correspondence between the measured fatty acid composition and the certified values is good.
18:1 c+t
18:0
40 30
18:3 n-6
18:3 n-3
50
16:0
18:2 c+t
FID1 A, (A:\0712-007.D)
20:0
16
14:0
16:1
20
10
12
17:0
14
20:1
18
minute
Figure 4.
GC/FID analysis of soy-maize oil (CRM 162) FAMEs on a 30-m 0.25-mm id 0.25-m DB-Wax column (part no. 122-7032) using method 1 (see Table 2).
16:0
100 90
70 60 50 40 30 20
4:0
6:0
10:0
12:0
18:0 18:2 c+t

14
80
8:0
16:1
18:1 c+t
14:0
Norm.
FID1 A, (A:\0712-009.D)
10
15:1
12
17:0
16
18:3 n-3
14:1
15:0
minute
Figure 5.
GC/FID analysis of milk fat (CRM 164) FAMEs on a 30-m 0.25-mm id 0.25-m DB-Wax column (part no. 122-7032) using method 1 (see Table 2). 7
Table 4:
Quantitative Data Obtained Using RTL Method 1 for CRM 162 Measured concentration (g/100 g) 10.607 2.917 24.461 57.051 4.286 0.368 0.246 Certified concentration (g/100 g) [7] 10.65 2.87 24.14 56.66 4.68 (0.3) (*) (0.2) (*)
Fatty acid 16:0 18:0 18:1 18:2 18:3 20:0 20:1
Standard deviation 0.003 0.005 0.013 0.017 0.017 0.003 0.003
Uncertainty (**) 0.17 0.07 0.28 0.54 0.21
(*) Indicative values, not certified (**) Uncertainty is taken as half-width of the 95% confidence interval of the certified mean value.
Table 5: Fatty acid 4:0 6:0 8:0 10:0 12:0 14:0 16:0 18:0 18:1 18:2 () 18:3
Quantitative Data Obtained Using RTL Method 1 for CRM 164 Measured concentration (g/100 g) 3.522 2.318 1.420 3.010 3.907 11.383 27.693 10.882 24.832 2.844 0.604 Standard deviation 0.012 0.003 0.002 0.006 0.008 0.014 0.005 0.009 0.009 0.012 0.005 Certified concentration (g/100 g) [7] 3.49 2.36 1.36 2.89 4.03 10.79 26.91 10.51 24.82 2.68 0.51
Uncertainty (*) 0.06 0.19 0.10 0.12 0.10 0.35 0.84 0.40 0.61 0.40 0.04
(*) Uncertainty is taken as half-width of the 95% confidence interval of the certified mean value.
The separation for the 37-component standard mixture on the 60-m 0.25-mm id 0.15-m DB-23 column using method 2 is shown in Figure 6. Using these conditions, all compounds in the standard mixture are resolved. Important is the separation of the cis/trans isomers and the separation of EPA (20:5, 19.15 min) and DHA (22:6, 22.38 min). This method is very useful for the analysis of fatty acid profiles in complex mixtures. An example of the separation obtained for a mixture of polyunsaturated FAMEs from a marine source appears in Figure 7. The identifications shown on the chromatogram were done by using a classical calibration table with a 5% retention time window for identification and quantitation (the default setting). Using this setting, the last three peaks (at
22.122, 22.262, and 22.414 min) were identified as 24:0, 24:1, and 22:6, respectively. The result from the peak identification using the locked retention time database appears in Table 6. RTL library searching shows that the initial identification (using the calibration table) was wrong. The correct identification was: 22.122 min = unknown (no fatty acid methyl ester), 22.262 min = 24:1, and 22.414 min = 22:6. These identifications could be confirmed easily by GC/MS. The profiles obtained by GC/FID and GC/MS for the same sample of marine FAMEs appears in Figure 8. An excellent retention time correlation was obtained.
Norm. 100
16:0
FID1 A, (RTL\60DB2303.D)
90
70
18:0 18:1, cis
80
6:0
8:0
18:1, trans
10:0
12:0
14:1
11:0
20:2 21:0 20:3 n-6 20:4 n-6 20:3 n-3 22:0 20:5 22:1 22:2 23:0
60
13:0
14:0 15:0 15:1
18:2, trans 18:2, cis 18:3 n-3 18:3 n-6
20:0
16:1 17:1 17:1
50
20:1 n-9
4:0
24:0 24:1
22 24 minute
40
30
20 4 6 8 10 12 14 16 18 20 minute
Figure 6.
GC/FID analysis of FAMEs standard mixture on a 60-m 0.25-mm id 0.15-m DB-23 column (part no. 122-2361) using method 2 (see Table 3).
FID1 A, (RTL\PUFA0001.D) 20:1 n-9 14:0 16:1 18:1, cis

Norm. 100
20:5
90
80
60
16:0
18:1, trans
18:2, cis
70
50
18:3 n-6
18:2, trans
21:0 20:2 20:3 n-6 20:4 n-6 20:3 n-3 22:0
40
20 8 10 12 14 16 18 20 22
Figure 7.
GC/FID analysis of unsaturated FAME mixture of marine origin on a 60-m 0.25-mm id 0.15-m DB-23 column (part no. 122-2361) using method 2 (see Table 3). The peak at 22.122 min was initially identified as 24:0 using a normal calibration table with the default retention time window. However, RTL library searching proved that this compound was, instead, an unknown.
12:0
18:0
20:0
30
14:1 15:0 15:1
18:3 n-3
17:1
23:0 22:2
24:0 24:1
17:0
22:1
22:6
22:6
Table 6.
GC/FID Peak Identifications Using the FAMEs Retention Time Database D:\HPCHEM\2\DATA\RTL\PUFA0001.D PUFA 1 2/15/02 2:40:50 PM 2/14/02 5:04:32 PM PUFA 1 VH RTLDB23.M Seq. Line : Vial: Inj: 1 01 1
Data file: Sample name: Instrument 2: Injection date: Sample name: Acq. operator: Acq. method: Results of RT table search
Search results for 22.122 0.100 minutes Contains elements: {No restriction} Does not contain elements: {No restriction} RTT file searched: D:\HPCHEM\RTL\FAMDB23.RTT No matches found. Search results for 22.262 0.100 minutes Contains elements: {No restriction} Does not contain elements: {No restriction} RTT file searched: D:\HPCHEM\RTL\FAMDB23.RTT RT 22.254 Compound Nervonic acid methyl ester Code 24:1
Search results for 22.414 0.100 minutes Contains elements: {No restriction} Does not contain elements: {No restriction} RTT file searched: D:\HPCHEM\RTL\FAMDB23.RTT RT 22.382 Compound cis-4, 7, 10, 13, 16, 19-docosahexaenoic acid methyl ester Code 22:6
10
Abundance
6000000 5000000 4000000 3000000 2000000 1000000 0 -1000000 -2000000 -3000000 -4000000
Signal: PUFA0001.D\FID1A.CH (*)
GC/FID
GC/MS
TIC: TEST4.D (*)
-5000000
9.00
10.00
11.00
12.00
13.00
14.00
15.00
16.00 Time
17.00
18.00
19.00
20.00
21.00
22.00
23.00
Figure 8.
Comparison of GC/FID and GC/MS chromatograms obtained for the unsaturated FAME mixture of marine origin on a 60-m 0.25-mm id 0.15-m DB-23 column (part no. 122-2361) using method 2 (see Table 3).
Conclusions
Two methods are described for the analysis of FAMEs. Method 1, using an Agilent DB-Wax column, is useful for the analysis of classical edible oils and fats, including the determination of butyric acid in milk fat. Method 2, applying an Agilent DB-23 cyanopropyl column, can be used for the analysis of more complex samples, including fish oils and hydrogenated fats, for the determination of EPA and DHA and for cis/trans determination. Using retention time locking, GC/FID and GC/MS retention times can be closely matched for easy correlation of chromatograms between the instruments. RTL database searching makes peak identification more accurate.
Translation and Retention Time Locking, Agilent Application Note, Publication 5967-5820E, May 1998. 3. V. Giarrocco, B. Quimby, and M. Klee, Retention Time Locking: Concepts and Applications, Agilent Application Note, Publication 5966-2469E, December 1997. 4. K.R. Weiner and H.F. Prest, Retention Time Locking: Creating Custom Retention Time Locked Screener Libraries, Agilent Application Note, Publication 5968-8657E, December 1999. 5. IUPAC Workshop 2/87. 6. AOAC Official Methods of Analysis (1990), method 969.33. 7. European Community Bureau of Reference (BCR), Brussels, Belgium (see catalog at www.irmm.jrc.be).
References
1. W.W. Christie, Gas Chromatography and Lipids, A Practical Guide, 1989, The Oily Press, Ayr, Scotland (ISBN 0-9514171-O-X). 2. B.D. Quimby, L.M. Blumberg, M.S. Klee, and P.L. Wylie, Precise Time-Scaling of Gas Chromatographic Methods using Method

11
High-Speed Separation of Parabens
Application
Food Analysis
Robert Ricker
Parabens, or para-hydroxy benzoic acid alkyl esters are popular preservatives used in the cosmetic and food industry to battle microbe degradation. Analysis time of less than one minute is accomplished with a Rapid Resolution SB-C18, 4.6 x 30 mm column. Columns with particle sizes under 5 m and dimensions up to ten-times smaller than traditional analytical-size columns are ideal for high-speed methods. If desired, increasing temperature cuts analysis time further. Extra-column volume is a crucial factor in chromatographic performance. Here, no modification to the modern instrument is necessary for optimal resolution.
HPLC System: Column: mm Mobile Phase: Detection: Flow: Inj. Volume: Temperature: Agilent 1100 with quaternary pump ZORBAX StableBond-C18 Rapid-Resolution (3.5 m) Cartridge-Column, 4.6 x 30 Agilent Part No. 833975-902 0.1% H3PO4: ACN, (50:50) UV 254 nm with standard flow cell (13 L) 2 mL/ min. 1 L top: ambient, bottom: 60C
Highlights
Elevated operating temperature is effective in reducing run time.
ZORBAX SB-C18
mAU 200 100 0 mAU 200 100
1 2 3 4
1. Methylparaben 2. Ethylparaben 3. Propylparaben 4. Butylparaben
ZORBAX StableBond SB-C18 can operate at higher temperatures and lower pH than other commercial reversed-phase columns.
Ambient
60C
0 0 0.2 0.4 0.6 0.8 1 min
The Analysis of Triglycerides in Edible Oils by APCI LC/MS

Doug McIntyre
Application Note
Food
Abstract
Triglycerides are the major components in edible oils. This note describes a method for analyzing intact triglycerides by LC/MS and the type of information available.
known as well as for determining other characteristics of the oil such as product purity, suitability for deep-frying, etc. Traditionally, they have been analyzed by hydrolyzing the triglycerides to yield the fatty acids, which are then derivatized and analyzed as methyl esters by GC/MS.1 This offers only indirect analysis of the original triglycerides. Triglycerides can be chromatographed intact by HPLC but their lack of a chromophore
unless conjugated makes detection difficult.2 In addition, in order to identify each individual triglyceride in an oil, a larger number of standards than is practical would be required and the analysis time would be quite lengthy. LC/MS analysis offered the ease of an LC separation plus the specificity of mass detection. Due to the non-polar nature of the molecules, atmospheric pressure chemical ionization (APCI) was employed.
Chromatographic Conditions Column: 200 2.1 mm Hypersil MOS, 5 m Mobile phase: A = 60:40 water isopropanol + 25 mM ammonium formate B = 10:10:80 water:isopropanol:n-butanol + 25 mM ammonium formate Gradient: Start with 30% B at 1.5 min 30% B at 25 min 60% B at 28 min 100% B Flow rate: 0.25 ml/min Column temp 50C Injection vol: 1.5 l MS Conditions Source: Ion mode: Vcap: Nebulizer: Drying gas flow: Drying gas temp: Corona Vaporizer: Scan range: Stepsize: Peakwidth: Time filter: Fragmentor APCI Positive 4000 V 50 psig 4 l/min 325 C 4 A 300 C 3001100 m/z 0.1 m/z 0.3 min On 80 V
Introduction
Triglycerides are found in both plant oils and animal fats. Their characterization is important for nutritional reasons, where the amount of unsaturation must be
Abundance 100000 90000 80000 70000 60000 50000 40000 30000 20000 10000
Trilinolein (C18:2,[cis,cis]-9,12)
Triolein (C18:1,[cis]-9)
Trielaidin (C18:1,[trans]-9) Tristearin (C18:0)
16.00
17.00
18.00
19.00
20.00
21.00
22.00
23.00
min
Figure 1. TIC showing separation obtained for four C18 triglyceride standards. Note that cis and trans isomers are well resolved.
Experimental
The system was comprised of an 1100 Series binary pump, vacuum degasser, autosampler, thermostatted column compartment, and LC/MSD. The LC/MSD was used with the atmospheric pressure chemical ionization (APCI) source. Complete system control and data evaluation was done on the ChemStation for LC/MS. Earlier work to demonstrate feasibility was done on the 1090 HPLC and 5989B MS Engine equipped with the 5987A Electrospray Accessory and the G1075A Atmospheric Pressure Chemical Ionization source. A variety of vegetable oils as well as animal fat products such as butter were purchased at local supermarkets. All products were dissolved in isopropanol at a concentration of 20 l of oil per 10 ml of IPA or 20 mg of fat per 10 ml IPA. Triglyceride standards were purchased from Sigma and
439
8000 6000 4000 2000
prepared by dissolving 10 mg of standard in 5 ml chloroform then further diluted to the desired concentration with IPA.

Separation and Ionization In order to have the non-polar triglycerides elute in a reasonable mobile phase composition; a Hypersil MOS (C8) column was used. Even this column required using isopropanol and n-butanol to elute the larger triglycerides. The final gradient was able to resolve cis and trans isomers of C18:1 triglyceride standards as shown in Figure 1. It was found that the addition of ammonium formate to the mobile phase allowed the formation of an ammonium adduct [M+18]+ which was more stable than the protonated [M+1]+ species formed without the ammonium formate. The effect of adding ammonium
formate is shown in Figure 2. Additionally, the vaporizer had to be lowered to 300C to minimize fragmentation and enhance the molecular signal. If structural information was desired, in-source collision induced dissociation (CID) was employed by raising the fragmentor from 80 volts to 150 volts. Qualitative Spectral Information Under these analytical conditions, each triglyceride shows predominantly an M+18 adduct ion and few other major fragments. Since triglycerides in plants are made up predominantly of fatty acids with even numbers of carbons (C12, C14, etc) and either 0, 1 or 2 double bonds per acid, much can be determined from this single adduct ion. A spreadsheet was created to determine the possible combinations of double bonds and carbon
No ammonium formate
383
300 350
411
467 495
450 500 550 600
639 656
650
699
400
656.6
8000 6000 4000 2000
25 mM ammonium formate
439 383 411

300 350 400 450
474 496
500 550 600 650
695
Figure 2. Effect of ammonium formate on the spectrum of trilaurin from coconut oil.
chain length. This was determined from the basic formula: M = 218.03 + 28.03*m + 26.02*n Where M is the mass of the observed adduct ion, m is the number of ethylene (CH2CH2) groups and n is the number of ethenyl (CH=CH) groups. Since
only integer solutions are allowed, the number of possible answers is typically very short. The choice is reduced even further if you assume fatty acids of the same length and unsaturation as the most likely solution, followed by only small differences in the fatty acids. For example, an
unsaturated butyric acid with two oleic acids is not a likely solution. Figure 3 shows a typical TIC trace for coconut oil with the masses of the base ion annotated over each peak. Inserting these masses into the spreadsheet produces the results shown in Table 1. Further confirmation, if needed, can be
700000 600000 500000 400000 300000 200000 100000 0 8 10 572.5 600.6 628.6
656.6
684.7 712.7 740.7
12
14
16
768.7 768.7 18
20
min
Figure 3. TIC showing mass of the base peak of the spectrum from peak apex. Peaks differ by m/z 28 corresponding to difference in the number of (CH2CH2) groups on sidechains.
Table 1. Structure calculations for the major peaks in coconut oil.
Observed Ion 516.50 544.55 572.55 600.55 628.55 656.70 684.75 712.75 740.80 768.75 796.75
Calc. MW 498.52 526.52 554.52 582.52 610.52 638.67 666.72 694.72 722.77 750.72 778.72
Backbone Removed 280.42 308.42 336.42 364.44 392.44 420.59 448.64 476.64 504.69 532.64 560.64
m 10 11 12 13 14 15 16 17 18 19 20
n 0 0 0 0 0 0 0 0 0 0 0
Possible Sidechains C8,C8,C10 C8,C10,C10 or C8,C8,C12 3 X C10 C10,C10,C12 C10,C12,C12 or C10,C10,C14 3 X C12 C12,C12,C14 C12,C14,C14 or C12,C12,C16 3 X C14 C14,C14,C16 C14,C16,C16 or C14,C14,C18
# Double Bonds 0 0 0 0 0 0 0 0 0 0 0
obtained using in-source CID. This is shown in Figure 4. Most oils do not produce as simple a TIC trace as coconut oil. Figure 5 shows a trace for two grades of olive oil. Some differences are visible in the TIC but not much useful information is apparent. However, by looking at
extracted ion chromatograms (EICs), much more can be determined. Figure 6 shows the results from extracting the signals for m/z 908.8, 906.8, 904.8, 902.8, 900.8 and 888.8. These correspond to triglycerides made up from 3 C18 fatty acids with 0, 1, 2, 3, 4 and 5 double bonds respectively. Not
surprisingly, the dominant ion is at m/z 902.8, which corresponds to the ammonium adduct ion for triolein (3 C18:1). Oleic acid is the major fatty acid component in olive oil. Using EICs in this way, different oils can be compared to see how they differ in degree of unsaturation.
684.65 Fragmentor = 80 volts

8000 6000 4000 2000
439 467
350 400 450
502
500 550 600 650 700
684.65
8000 6000 4000 2000
Fragmentor = 150 volts 467 439 loss of C14 fatty acid 411
350 400 450
loss of C12 fatty acid
495
500 550 600 650 700
Figure 4. The effect of raising the Fragmentor voltage on a mixed triglyceride in coconut oil. The fragment masses and ratios are consistent with two C12 fatty acids and one C14.
Lite Olive Oil
400000
300000
Spanish Extra Virgin Olive Oil
200000
100000
0 16 17 18 19 20 21 22 min
Figure 5. The TIC of extra lite and spanish virgin olive oils. The data indicates lite oil is more unsaturated or has shorter fatty acids. Further examination of EICs indicate more unsaturation.
400,000
m/z 00 double bonds m/z908.8 908.8- double bonds m/z 11 double bond m/z906.8 906.8- double bonds m/z904.8 904.8- double bonds m/z 22 double bonds m/z902.8 902.8- double bonds m/z 33 double bonds m/z900.8 900.8- double bonds m/z 44 double bonds m/z 898.8 - 5 double bonds
m/z 898.8 5 double bonds
300,000
200,000
100,000
22
23
24
25
26
min
Figure 6. Extracted ion chromatograms (EICs) for olive oil corresponding to the 3 C18 series. Triolein is the major component.
Conclusion
LC/MS using APCI as an ionization process provides a simple way to analyze oils and fats for triglycerides. Information can be obtained directly rather than indirectly on the type of fatty acids including the degree of unsaturation. This information can be used for determining nutritional value and other physical properties of the oil or fat. It also has the potential of determining the amount of processing the oil has undergone (extra virgin versus pure). Little or no sample cleanup is required, allowing for a rapid analysis.
References
1. M. M. Mossoba and D. Firestone, Food Testing and Analysis, 1996, 2 (2), 2432. 2. R. Schuster, Multicomponent Analyses of Fats and oils using Diode-Array Detection, Hewlett-Packard Application Note, 1987, publication number 12-5954-6269.
Doug McIntyre is an applications chemist at Agilent Technologies
Agilent Technologies shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance or use of this material. Information, descriptions and specifications in this publication are subject to change without notice. Copyright 1998 Agilent Technologies, Inc. All rights reserved. Reproduction and adaptation is prohibited. Printed in the U.S.A. May, 2000 (23) 5968-0878E
Analysis of Triglycerides in Olive Oil and Rape Oil using HPLC

Abstract Unsaturated triglycerides in olive oil have very characteristic patterns. Other fats and oils are also complex mixtures of triglycerides but with different patterns. Sample preparation Triglycerides can be extracted from homogenized samples with petrol ether. Fats and oils can be dissolved in tetrahydrofurane. Chromatographic conditions The presented HPLC method was used to analyze the unsaturated triglycerides, LnLnLn, LLL, and OOO.1
mV 200 180 160 140 120 100
Conditions
Column 200 2.1 mm Hypersil MOS, 5 m Mobile phase acetone/ACN (30:70) Flow rate 0.5 ml/min Column compartment 30 C Injection vol 2 l Detector refractive index Sample preparation Samples were dissolved in tetrahydrofurane.
80
60 40 2 4 Time [min] 6 8
Rape oil Olive oil Standard
Figure 1 Analysis of the triglyceride pattern of olive and rape oil
HPLC method performance Limit of detection for ECD 50 g/l with S/N = 2 Repeatability of RT over 10 runs <0.3 % areas over 10 runs 5 % References 1. Determination of triglycerides in vegetable oils, EC Regulation No. L248, 28ff.
Equipment
Agilent 1100 Series degasser isocratic pump autosampler thermostatted column compartment refractive index detector Agilent ChemStation + software
Analysis of Unsaturated Triglycerides using HPLC
Abstract The HPLC method presented here was used to analyze triglycerides, hydroperoxides, sterols, and vitamins with UV-visible diode-array detection (UV-DAD). Spectra were evaluated in order to trace hydroperoxides and to differentiate saturated from unsaturated triglycerides. Unsaturated triglycerides in olive oil have a very distinctive pattern. Other fats and oils are also complex mixtures of triglycerides but exhibit an entirely different pattern. Adulteration with foreign fats and the use of refined triglycerides in olive oil also can be detected through triglyceride analysis. Sample preparation Triglycerides can be extracted from homogenized samples with petrol ether. Fats and oils can be dissolved in tetrahydrofuran.1
Conditions
Column 200 2.1 mm Hypersil MOS, 5 m Mobile phase A = water B = ACN/methyl-tert.butylether (9:1) Gradient at 0 min 87% B; at 25 min 100% B Post time 4 min Flow rate 0.8 ml/min Column compartment 60 C Injection vol 1 l standard UV absorbance 200 nm and 215 nm to detect triglycerides 240 nm to detect hydroperoxides 280 nm to detect tocopherols and decomposed triglycerides (fatty acids with three conjugated double bonds) Sample preparation Samples were dissolved in tetrahydrofuran (THF).
140 120 100 80

60 40
* Hydroperoxides
*
*
* * *
215 nm
20 0 5
240 nm 10 15 Time [min] 20

25
Figure 1 Triglyceride pattern of aged sunflower oil. The increased response at 240 nm indicates hydroperoxides
HPLC method performance Limit of detection for saturated triglycerides >10 g for unsaturated triglycerides fatty acids with 1 double bond >150 ng fatty acids with 2 double bonds >25 ng fatty acids with 3 double bonds <10 ng Repeatability of RT over 10 runs <0.7 % areas over 10 runs <6 %
Olive oil mAU 20 15 10 5 0 13.0 Time [min] 215 nm 280 nm 23.0 mAU 20 Good quality 15 10 5 0 13.0 Time [min] 215 nm 280 nm 23.0
Equipment
Poor quality
Figure 2 Analysis of olive oil. The response at 280 nm indicates a conjugated double bond and therefore poor oil quality
References 1. Determination of triglycerides in vegetable oils, EC Regulation No. L248, 28ff.
Analysis of Hydrolized Fatty Acids in Dietary Fat using HPLC
Abstract Saturated and unsaturated fatty acids from C4 through C22 have been analyzed. Fatty acids are the primary components of oils and fats and form a distinctive pattern in each of these compounds. For example, butter and margarines can be differentiated by the percentage of butyric acid in the triglycerides. To determine the fatty acid pattern of a fat or oil, free fatty acids first are obtained through hydrolysis. Derivatization is then performed to introduce a chromophore, which enables analysis of the fatty acids using HPLC and UV-visible detection. Sample preparation The triglycerides were hydrolyzed using hot methanol and KOH, followed by derivatization. Chromatographic conditions The HPLC method presented here was used in the analysis of the fatty acid pattern of dietary fat. The method involves hydrolysis with hot KOH/methanol and online derivatization with bromophenacyl bromide.
mAU C18-3 C18-2 C18-1 1400 1000 600 C16 200 15 20 Time [min] 25 Standard Dietary fat C18 C20 C14 C22 Standard
Conditions
Column 200 2.1 mm, MOS, 5 m Mobile phase A = water (70 %) B = ACN + 1% THF (30%) Gradient at 5 min 30% B; at 15 min 70% B at 17 min 70% B; at 25 min 98% B Flow rate 0.3 ml/min Column compartment 50 C Detector variable wavelength, 258 nm Derivatization 60 mg/ml bromophenacyl bromide was dissolved in ACN
30
Figure 1 Analysis of a dietary fat triglyceride pattern. Overlay of one sample and two standard chromatograms
HPLC method performance Limit of detection 200 pg injected amount, S/N = 2 Repeatability of RT over 10 runs <0.1 % areas over 10 runs 5 %
Norm 40 30 20 10 0 20 22 24 26 Time [min] 28 30 32 DAD VWD
Conditions
Injector program for online derivatization 1. Draw 2.0 l from vial 2 (ACN) 2. Draw 1.0 l from air 3. Draw 1.0 l from vial 3 (derivatization agent) 4. Draw 0.0 l from vial 4 (wash bottle) (ACN/THF, 50:50) 5. Draw 1.0 l from sample 6. Draw 0.0 l from vial 4 (wash bottle) 7. Draw 1.0 l from vial 3 (derivatization agent) 8. Draw 0.0 l from vial 4 (wash bottle) 9. Draw 1.0 l from vial 5 (acetonitrile + 5 % TEA) 10. Draw 0.0 l from vial 4 (wash bottle) 11. Mix 9 l in air, 30 l/min speed, 10 times 12. Wait 2.0 min 13. Inject Sample preparation 0.215 g fat was hydrolyzed with 500 l MEOH/ KOH at 80 C for 40 min in a thermomixer. After cooling 1.5 ml ACN/THF (1:1) was added, and the mixture was shaken for 5 min. The mixture was then filtered through a 0.45-m Minisart RNML from Satorius.
Figure 2 Trace analysis of triglycerides with a diode-array and a variable wavelength detector in series
Equipment
Agilent 1100 Series vacuum degasser quaternary pump autosampler thermostatted column compartment variable wavelength detector Agilent ChemStation + software
Flavors and Fragrances
Comprehensive GC System Based on Flow Modulation for the 7890A GC
Application Brief
Introduction
A hardware solution is available on the 7890A for the practice of comprehensive GC. The system uses a capillary flow modulator controlled by the 7890A GC. The system is offered with factory checkout using an FID detector. Other detectors, preferably those operating at 50 Hz or greater, can be used. Comprehensive two-dimensional (2D) GC, or GCxGC, is a powerful technique that can be used to separate very complex mixtures, such as those found in the hydrocarbon processing, environmental, and food/fragrance industries. The method uses two columns, typically of very different polarities, installed in series with a modulator in between. The second column is much shorter than the first column to effect a fast separation. The entire assembly is located inside the GC oven. The modulator performs three functions: 1. It collects effluent from the first column for a fraction of the time equal to peak width. For example, if a peak from column one is six seconds wide, the modulator will accumulate material every two to three seconds, thereby dividing the peak from the first column into two or three cuts. 2. It focuses the material collected from each cut into a very narrow band through flow compression. 3. It introduces the bands sequentially onto the second column, resulting in additional separation for each band injected onto the second column.
ALS Conventional GC Column 1 Power to separate in second column In reality, some correlation in retention characteristics will be present between the two columns.
Modulator
Column 2 Peak capacity is the maximum number of equally resolved peaks that can be obtained in a given time, usually the entire run. The classic argument made is that GCxGC increases peak capacity over what is possible with a one-dimensional separation.
Fast GC
Detector
Comprehensive 2D GC uses a primary column (conventional separation), a flow modulator, a second column (very fast separation), and a fast detector. This technique provides a second dimension of information that can increase the peak resolution and capacity. A number of different modulator designs have been described in the literature, most relying on thermal cycling to focus the bands from the first column and release them into the second column. Some disadvantages to this approach are: Large usage of expensive cryogenic gases leading to a high cost of analysis Complexity of the hardware Longer analysis times Agilents proprietary Capillary Flow Technology and fourth-generation Electronic Pneumatics Control (EPC) enable the use of a differential flow modulator to conduct comprehensive 2D-GC without the use of cryogenic gases or complex hardware.
The key to operation is the flow differential (typically 20 to 1) between the second and first columns, respectively. This compresses and focuses the analytes present in any given modulation inject pulse into the second column. Precise timing of the modulator is made possible by installing a driver board in the Aux det 2 detector slot of the 7890A mainframe. The Capillary Flow Technology modulator uses a deactivated, stainless steel structure with all flow splitters and the collector channel incorporated internally in the device. It has low thermal mass so it can track the oven temperature very closely, and its GC oven location allows precise temperature control without lag during programmed runs. All external connections are made using Agilents Ultimate Union technology for leak-free operation and extremely small, well-swept volumes. A micro three-way solenoid valve, installed on the side of the gas chromatograph, connects to a pneumatics control module (PCM) to accurately and precisely control the flows through the modulator. The figures below illustrate the modulator. A three-way solenoid valve receives a controlled supply of hydrogen gas from a PCM. The periodic switching of this three-way valve drives the modulator. The precisely timed and synchronized switching between the collect and inject states directs discrete sample pulses continuously to the second column for additional fast separation throughout the chromatographic run. Both columns are run in constant flow mode. For optimal performance, injection size and split ratio should be carefully adjusted to avoid overloading, which can lead to excessive peak tailing.
Split/splitless inlet Column 1 (25-30M)
The primary column effluent enters the modulators top tee connection and flows into the collection channel. The analytes from this column enter one end of the collection channel. Hydrogen flow from the PCM/three-way micro valve exits the modulator at the bottom tee and is sent to the second column.
Split/splitless inlet Column 1 (25-30M) Inject flow direction
H2
Modulation valve
Collection channel
0.8 mL/min
Column 2 (5M)
FID Flow Modulator

Collection channel is quickly injected into second column in about 0.1 second
approx. 21 mL/min
Flow rates and flow directions during the transfer or inject portion of the modulation cycle Inject or flush state (above): Hydrogen gas flow from the three-way solenoid valve is directed to the top tee. A high flow of typically 20 mL/min for about 0.1 second rapidly flushes the collection channel, transferring material in a very narrow band onto the second column where any analytes collected in the channel undergo rapid separation.
What is required: Agilent 7890A GC with firmware version A.04.06 or higher FID with 200 Hz data collection rate or other fast detector Split/splitless inlet Capillary Flow Technology modulator option or accessory Capillary Flow Technology modulator checkout kit Pneumatics control module (PCM) Agilent GC ChemStation B.03.02 or other data collection and analysis system that can control the flow modulator cycle 30-m 0.25-mm 0.25-m DB-5ms column (included with option or accessory) 5-m 0.25-mm 0.15-m INNOWax column (included with option or accessory) 2D data analysis software, GC Image recommended (not provided by Agilent) Internal column nuts and SilTite ferrules
H2
Modulation valve Collect flow direction
Collection channel
0.8 mL/min
Column 2 (5M)
FID
Flow Modulator
approx. 21 mL/min
Flow rates and flow directions during the load or collect portion of the modulation cycle Load or collect state (above): At the beginning of this state, the collection channel is filled with hydrogen gas from a previous injection cycle flush.
2
Ordering Information
Description 7890A GC with Capillary Flow Technology Modulator (requires checkout kit) 7890A GC with 200 Hz FID 7890A GC with split/splitless inlet Capillary Flow Technology modulator checkout kit PCM for 7890A GC SilTite metal ferrules, 1/16-in 0.4-mm id, 10/pk, includes 2 column nuts Agilent 32-bit ChemStation for 1 GC Agilent 32-bit ChemStation Bundle for 1 GC includes: G2070BA 32-bit ChemStation software Computer with monitor and Windows operating system Printer 2D GC software Recommend GC Image software, which can be purchased from Zoex Corporation Part number G3440A Option 887 or accessory G3486A G3440A Option 211 or accessory G3462A G3440A Option 112 or accessory G3452A G3487A G3440A Option 309 or accessory G3471A 5184-3569 G2070BA G1875BA
www.zoex.com
Application Examples
Several applications are shown. Note that primary column lengths have been chosen to give optimal results. While the 30M column that is shipped with the system is an excellent choice for a wide range of applications, other lengths can be used to optimize a given separation. Various columns have been used in these examples to illustrate some of
the possibilities. The GC Image software package was used for processing the ChemStation data. 1. B20 biodiesel based on soy FAMES. Section of the 2D image showing the C16 and C18 FAMES is shown. Column 1: 60 m 0.25 mm 0.10 m DB-5ms Column 2: 5 m 0.25 mm 0.15 m INNOWax Modulation: 1.40 s load, 0.10 s inject
C18:1
C16
C18:2 C18:0 C18:3
2. Complete 2D image of a sample of heavy gasoline. Each series of substituted 1-ring aromatics is well separated, making hydrocarbon class grouping possible. Column 1: 60 m 0.25 mm 0.10 m DB-5ms Column 2: 5 m 0.25 mm 0.15 m INNOWax Modulation: 1.40 s load, 0.10 s inject
3. Lime oil 2D image. Column 1: 15 m 0.25 m 0.25 m DB-5ms Column 2: 5 m 0.25 mm 0.15 m DB-17HT Modulation: 1.40 s load, 0.10 s inject
Limonene
C20 reference
Thermal vs. Flow Modulation

Since competitors offer only systems based on thermal modulation, the following table summarizes the key points about the respective approaches of thermal vs. flow modulation.
Thermal modulation Cryo-focusing provides potentially narrower peaks in second dimension Lower flows Can be used with highvacuum detectors (TOF) Large consumption of cryogen Complex hardware design, set-up, and maintenance Long chromatographic runs required for best performance System price (estimate) $60 to $70K
Differential Flow modulation Peak widths comparable to thermal. Usually no more than 20% wider. Many users want to sum regions of peaks where peak width is not as critical MSD can be used with a splitter over limited scan range No cryogen required Simple, reliable Capillary Flow Technology based hardware; small thermal foot print Run times comparable to a 1D separation Agilent system approximately $60K (list)

Rapid Analysis of Food and Fragrances Using High-Efficiency Capillary GC Columns Application
Food, Flavors and Fragrances
Authors
Mark Sinnott and Simon Jones Agilent Technologies, Inc. 91 Blue Raving Road Folsom, CA 95630-4714 USA
Abstract
An analysis of various essential oils/flavors was performed using both polar and nonpolar high-efficiency 0.18-mm id GC columns. Agilents GC method translation software was used to translate existing 0.25-mm id column methods to 0.18-mm id columns. The ease with which this software can be used allowed for simple method development. Elution times were compared between the standard 0.25-mm id columns and the highefficiency 0.18-mm id columns. The benefits of using hydrogen carrier gas for shorter analysis time will also be illustrated.
Efficiency is often related to the number of theoretical plates (n) that a column has and is expressed as plates per meter. It follows that the longer the column, the more plates you have, and thus the more efficient the column. One way to measure column efficiency is to calculate the height equivalent of a theoretical plate (HETP = h). Following Equation 1, the lower the value of h is, the greater the value of n and therefore the efficiency. The shorter the plate is, the larger the number of plates that can be stacked in a given length of column. By reducing the column id, the plate height is reduced, which results in more plates per meter (see Equation 2). A more efficient, smaller id column can be used to obtain the same number of plates in a shorter length of column. The shorter the column, the less time the analytes take to travel that length of column, which equates to shorter analysis times without the loss of efficiency or resolution. L n Equation 1: Height Equivalent to a Theoretical Plate h= h = height equivalent to a theoretical plate L = length of the column in millimeters n = number of theoretical plates Equation 2: Height Equivalent to a Theoretical Plate in Relation to Column Diameter
Introduction
There are many misconceptions about what it means to perform fast gas chromatography (GC) and what the term fast GC implies. Fast GC is often associated with the use of hydrogen as a carrier gas, and although this is certainly a good approach, it is not always necessary in order to shorten the analysis time. A second misconception is that changing column dimension results in time-consuming method development. Utilizing 0.18 mm id high-efficiency GC columns can greatly reduce the analysis time. When coupled with Agilents GC method translation software, the time spent on method development can be greatly minimized.
hmin = r
(11 k 2 + 6 k +1) 3 (1 + k)2
hmin = height equivalent to a theoretical plate r = radius of column k = capacity factor (partition coefficient) of an analyte
The high-efficiency GC columns are designed to maintain the same phase ratio as the more commonly used 0.25-mm id columns, making for easy method translation, as will be illustrated. Phase ratio is a unitless measure of the relationship between the column radius and the stationary phase thickness. If this calculated number changes when changing from one dimension column to another, there is a change in the retention of a particular solute (k). Equation 3 illustrates that even though a shorter column means a shorter time that the analyte takes to elute from the column, k will remain constant because the unretained compound will also elute more quickly. Equation 3: Partition Ratio t t k = r o to k = partition coefficient of an analyte tr = retention time of analyte to = retention time of an unretained compound During the chromatographic process, the resulting chromatogram and its associated resolution are the product of the thermodynamics of the system. If the dimensions of the column are changed, then the thermodynamics of the system also change. A new temperature program must be developed to match the new column dimensions. This is why most analysts avoid trying to go faster; the time and energy that goes into developing a new method just isnt worth it. One solution to this problem is to utilize the GC method translation software that is available online at the Agilent Web site http://www.chem.agilent.com/cag/servsup/ usersoft/files/GCTS.htm (see Figure 1). This free software takes the guesswork out of developing a new temperature program. This assumes that the same column phase type and same phase ratio are being used between the two methods. It is not imperative to use the same phase ratio; however, if the phase ratio is not maintained, the elution order should be confirmed. An additional option for faster analysis is to utilize a more efficient carrier gas. When changing carrier gas types from one to another, the method translation software takes into account the efficiencies of the four most commonly used carrier gases (argon, nitrogen, helium, and hydrogen) and adjusts the method parameters accordingly. In this application the benefits of using highefficiency columns to shorten run times will be illustrated. Two commonly used columns for food/fragrance analysis are the DB-1 and DB-WAX. A comparative analysis will be performed between a more commonly used dimension (30 m 0.25 mm 0.25 m) and that of the high-efficiency column
2
dimension (20 m 0.18 mm 0.18 m). In addition, the ability of the GC method translator to minimize time spent performing method development will be demonstrated. The benefits of using hydrogen as a carrier gas in conjunction with the high-efficiency columns will also be addressed.
Figure 1.
Method translation input screen.
Experimental
All analyses were performed using an Agilent 6890 GC with a 5973 MSD, equipped with a split/splitless inlet. The analytical conditions are summarized in Table 1 (DB-1 columns) and Table 2 (DB-WAX columns). Original method parameters were not optimized for each compound, but rather developed to accommodate a wide range of essential oils and fragrances. Method parameters used for the high-efficiency columns were translated directly from the Agilent method translation software. Spearmint and ylang-ylang samples were prepared by dilution of neat oils with acetone at roughly 40:1.
Table 1. Method A Column Carrier Injector Oven Method B Column Carrier Oven Method C Column Carrier Oven Method Conditions for DB-1 Columns 30 m 0.25 mm 0.25 m DB-1 p/n 122-1032 Helium 25 cm/sec measured at 40 C 250 C, Split 40:1, 1-L injection 40 C hold 1 min 5 C/min to 290 C 20 m 0.18 mm 0.18 m DB-1 p/n 121-1022 Helium 26 cm/sec measured at 40 C 40 C hold 0.64 min 4.67 C/min to 290 C 20 m 0.18 mm 0.18 m DB-1 p/n 121-1022 Hydrogen 47 cm/sec measured at 40 C 40 C hold 0.38 min 13 C/min to 290 C hold 13.09 min
Table 2. Method A Column Carrier Injector Oven Method B Column Carrier Oven Method C Column Carrier Oven
Method Conditions for DB-WAX columns 30 m 0.25 mm 0.25 m DB-WAX p/n 122-7032 Helium 25.4 cm/sec measured at 45 C 250 C, Split 30:1, 1-L injection 45 C hold 2 min 3 C/min to 250 C hold 34 min 20 m 0.18 mm 0.18 m DB-WAX p/n 121-7022 Helium 26.3 cm/sec measured at 45 C 45 C hold 1.28 min 4.68 C/min to 250 C hold 21.81 min 20 m 0.18 mm 0.18 m DB-WAX p/n 121-7022 Hydrogen 44.3 cm/sec measured at 45 C 45 C hold 0.77 min 7.79 C/min to 250 C hold 13.09 min
compound for spearmint is viridiflorol, which decreased from 27.41 minutes to 17.73 minutes, as illustrated in Figures 1 and 2. This represents a speed gain of approximately 35%. Resolution between three close-eluting compounds remained nearly identical as is illustrated in Table 4. Elution time of the last compound for ylang-ylang oil tested on the DB-WAX column, benzyl salicylate, decreased from 63.47 to 41.07 minutes. This is illustrated in Figures 3 and 4. This represents a speed gain of approximately 35%. Resolution between two pairs of compounds remained essentially unchanged (see Table 5). Using hydrogen as a carrier gas in conjunction with the high-efficiency columns resulted in additional speed gains. Due to its small molecular size, hydrogen can be used at higher velocities without loss of efficiency. These additional benefits are illustrated in Figures 5 and 6 as well as in Tables 4 and 5. The overall speed gain from the original method was found to be 61% for both the DB-1 and DB-WAX methods. The method translation software allowed for essentially plug-and-play method development. The results that were obtained were used without modification to the values provided by the translation software.

Typical chromatograms are presented here for all three GC methods on both the DB-1 and DB-WAX columns. Peak identities can be found in Tables 3A and 3B. Significant speed gain is achieved by simply switching to the high-efficiency column and continuing to use helium as the carrier, without a loss of resolution. Spearmint was tested on the DB-1 column. The elution time for the last-eluting
Table 3A. Component List for DB-1 Chromatograms
Compound List for Spearmint Oil Chromatogram 1 2 3 4 5 6 7 8 9 10 11 -Pinene Sabinene -Pinene 3-Octanol Myrcene -Terpinene -Cymene 1,8-Cineol Limonene cis-Ocimene trans-Ocimene 12 13 14 15 16 17 18 19 20 21 22 -Terpinene trans-Sabinene hydrate Terpinolene Linalool 3-Octyl acetate Isomenthone Terpinen-4-ol Dihydrocarvone trans-Carveol l-Carvone trans-Dihydrocarveol acetate 23 24 25 26 27 28 29 30 31 cis-Carvyl acetate cis-Jasmone -Bourbonene -Bourbonene -Caryophylene -Copaene trans--Farnesene Germacrene-d Viridiflorol
Table 3B. Component List for DB-WAX Chromatograms Compound List for Ylang-Ylang Oil Chromatogram 1 2 3 4 5 6 -Pinene Methyl--cresol -Copaene -Gurjunene Linalool -Caryophyllene 7 8 9 10 11 12 Methyl benzoate -Caryophyllene Germacrene-d Benzyl acetate Farnescene -Cadinene 13 14 15 16 17 18 Geranial acetate trans-Cinnamyl acetate -Bisbolene Farnesyl acetate Benzyl benzoate Benzyl salycilate
Table 4.
Resolution of Closely Eluting Compounds by Column ID and Carrier Gas DB-1 ID and Carrier Gas Type 0.25 mm Helium 1.52 1.61 0.18 mm Helium 1.59 1.73 0.18 mm Hydrogen 1.56 1.86
Compound(s) Sabinene -Pinene -Terpinene p-Cymene Speed gain
N/A
35%
61%
Table 5.
Resolution of Closely Eluting Compounds by Column ID and Carrier Gas DB-WAX ID and Carrier Gas Type
Compound(s) -Farnesene -Cadinene -Cadinene Geranial acetate Speed gain
0.25 mm Helium 2.16 1.67
0.18 mm Helium 2.14 1.66
0.18 mm Hydrogen 2.13 1.64
N/A
35%
61%
21
25 8 23 19 13 22 27 29 30 12 6 7 10 11 16 14 15 14 17 16 Ti 24 26 28 18 ( i ) 20 22 24 26
4 1 2 3
18
20
31
10
12
28
Figure 1.
Spearmint oil sample on a DB-1, 30 m x 0.25 mm x 0.25 m column and He carrier. (See Table 1, Method A, for experimental parameters.)
21
25 5 8 23 19 4 3 2 67 10 13 18 22 27 29 30
12 11
16 14 15 17
20
24
26 28
31
11
13
15
17
Figure 2.
Spearmint oil sample on DB-1, 20 m x 0.18 mm x 0.18 m column, He carrier. (See Table 1, Method B, for experimental parameters.)
9 10
6 5 17
11 2 7 13 18 4 8 1 3 12 14 15 16
12
17
22
27
32
37 Time (min)
42
47
52
57
62
67
Figure 3.
Ylang-ylang oil sample on a DB-WAX, 30 m x 0.25 mm x 0.25 m column and He carrier. (See Table 2, Method A, for experimental parameters.)
10
6 5
17 2 11 4 7 13 8 3 1 12 15 16 18 14
10
12
14
16
18
20
22
24
26
28
30
32
34
36
38
40
42
44
46
48
50
Figure 4.
Ylang-ylang oil on DB-WAX, 20 m x 0.18 mm x 0.18 m column, He carrier. (See Table 2, Method B, for experimental parameters.)
21
8 5
25
13 1 3 4 2 6 3 7 12 10 11 5 16 15 17
19 18
22
23
27
30 14 20 26 24 7 Time (min) 9 28 29 31 11
Figure 5.
Spearmint oil sample on DB-1, 20 m x 0.18 mm x 0.18 m with H2 carrier. (See Table 1, Method C, for method parameters.)
10
4 7 8 3
11 13 12 14 15
17
18 16
10
12
14
16 Time (min)
18
20
22
24
26
28
30
Figure 6.
Ylang-ylang oil sample on a DB-WAX, 20 m x 0.18 mm x 0.18 m column with H2 carrier. (See Table 2, Method C, for method parameters.)
Conclusions
The use of high-efficiency columns has many benefits, as illustrated here. Shorter analysis times were achieved without significant loss of resolution. Time spent in method development was kept to a minimum through the use of the GC method translation software and the fact that the high-efficiency columns were phase ratio matched. While there are additional benefits for using hydrogen as the carrier gas, significant speed gain can be realized by simply using the high-efficiency columns while maintaining helium as the carrier. *Although this application only depicts two oils, several additional flavors/fragrances were analyzed. Please contact Agilent Technologies Application Support for additional information.
Reference
D. Rood, The Practical Guide to the Care, Maintenance and Troubleshooting of Capillary Gas Chromatographic Systems, Huthig, Heidelberg, 1991

For more information on our products and services, visit our Web site at www.agilent.com/chem. For more information about the high-efficiency GC columns, visit our Web site at http://www.chem. agilent.com/scripts/PDS.asp?1Page=60005.
The High-Resolution Reversed-Phase HPLC Separation of Licorice Root Extracts Using Long Rapid Resolution HT 1.8-m Columns Application
Food Additive, Natural Products, and Pharmaceuticals
Author
Bernard Permar and Ronald E. Majors Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA
COOH
COOH O
Abstract
High-resolution reversed-phase HPLC analytical studies using licorice, a licorice hydrolysis product, and commercial licorice samples, showed that resolution and throughput using a ZORBAX 1.8-m column greatly exceeded that obtained using the conventional 5.0-m column.
COOH O O
Figure 1.
Structure of glycyrrhizinic acid.
Introduction
Licorice is derived from the root of the Glycyrrhiza glabra plant, a 4- to 5-foot woody shrub that grows in Europe, the Middle East and Western Asia. The root of the plant is known to contain about 4% glycyrrhizin, the potassium or calcium salt of glycyrrhizinic acid. The latter is a glycoside of a pentacyclic triterpine carboxylic acid (18--glycyrrhetic acid) with two molecules of glucuronic acid (Figure 1). Glycyrrhizin is about 50 times sweeter than sucrose (cane sugar) but at high dosage is known to have toxicity. Upon hydrolysis, the glycoside loses its sweet taste and is converted to the aglycone glycyrrhetinic acid plus two molecules of glucuronic acid.
Extractions from any of the many species of this plant will yield a complex mixture containing more than 100 compounds. Several of these compounds are used as additives in candy as sweeteners, in cough syrup as flavoring agents, and in drugs to mask a bitter taste, or for their therapeutic qualities, mainly in Traditional Chinese Medicines (TCMs). The medicinal properties of licorice have been known for several centuries in China, as well as India, Egypt, Greece, and Rome. Uses included cough suppressant, laxative, and treatments for gastric ulcer, early Addison disease, and liver disease. Most recently, glycyrrhizin has been shown to have antiviral activity against DNA and RNA viruses (influenza A and B, HIV, VZV, Hepatitis B and C) [1]. Licorice has also been used in topical cosmetic applications.
The abundance of certain compounds of interest will vary greatly according to the species of the plant, the time of harvest, and the method of extraction. Thus, analytical methods to follow the active ingredients are required. Gradient elution reversed-phase HPLC has been found to be an effective method for separating some of the important compounds in licorice [2]. This application note compares the traditional HPLC methodology and the newer Rapid Resolution high-throughput (RRHT) columns. We will apply these HPLC techniques to investigate the differences between two commercially available licorice root extracts.
Standards:
Purchased from Sigma Aldrich
(G) 0.1-mg glycyrrhizic acid ammonium salt, ~75 %, dissolved in 0.5-mL mobile phase B, then brought to 1.0 mL by adding 0.5-mL mobile phase A (GA) 0.1-mg 18-beta-glycyrrhetinic acid, 97%, dissolved in 0.5-mL mobile phase B, then brought to 1.0 mL by adding 0.5-mL mobile phase A Samples: Licorice root extract A (HERB FARM brand) Licorice root extract B (Newark Natural Foods) Both extracts should be vortexed, then filtered (0.2 micron) prior to injection. Injection volume: 5 L of extract
Experimental
Two reversed-phase (RP) columns were used in this study: Conventional ZORBAX StableBond (SB)-C18, 4.6 mm 250 mm, 5 m ZORBAX SB-C18 RRHT, 4.6 mm 150 mm, 1.8 m The smaller particle size of the RRHT column allows use of a shorter column to achieve the same resolution as the longer conventional column, and also allows more rapid separations.
Results
The most important compound found in a typical licorice extract is G and to a lesser extent, its hydrolysis product, GA. These substances can be purchased commercially. Although some of the other components of licorice have been identified and are available commercially, they are quite expensive. Since our main objective was to demonstrate the advantage of using shorter, highresolution HPLC columns, we used only two standards (G and GA) to develop the initial method. Figure 2a shows the gradient separation of G and GA on the conventional column (ZORBAX SB-C18, 4.6 mm 250 mm, 5 m) using gradient elution. Since the licorice extract to be examined was quite complex, isocratic conditions were not usable to separate all of the components. The G being more polar by virtue of the additional sugar moieties eluted first while the GA came off the column much later. Using this gradient, the GA eluted in just under 42 minutes. By switching to the shorter ZORBAX SB-C18 RRHT column (4.6 150 mm, 1.8 m), the separation was virtually the same, as can be seen in Figure 2b. However, the separation time was now just over 25 min, a savings of about 40% in time.
HPLC conditions Instrument: Detector: Mobile phase: Agilent 1200 Series Rapid Resolution System Multiple wavelength detector (MWD), 254 nm/100 BW, 450 nm reference A = 1% Acetic acid in water B = 1% Acetic acid in acetonitrile
Gradient conditions for ZORBAX SB-C18 columns: Conventional: RRHT: Flow: Temperature: 4.6 mm 250 mm, 5 m 5% to 100% B in 50 minutes 4.6 mm 150 mm, 1.8 m 5% to 100% B in 30 minutes 1.0 mL/min Ambient
mAU 80 60 40 20 0 0
ZORBAX SB-C18, 4.6- 250-mm, 5 m 2a

G 22.9 min GA 41.4 min
10
20
30
40
50
min
mAU 160 120 80 40 0 0
ZORBAX SB-C18 RRHT, 4.6- 150-mm, 1.8 m 2b

10
15
20
25
30
min
Figure 2.
Gradient reversed-phase separation of G and GA on the test 5.0- and 1.8-m columns.
To investigate the use of these columns for the separation of actual licorice root extracts, Figures 3 and 4 depict the complex chromatograms observed by injection of filtered extracts, identified in the Experimental section. Figure 3a shows the complex chromatogram obtained using the 5-m 250-mm column. The cut-away shows the small amount of GA that was present in the extract. Since GA is a hydrolysis product of G, it should be at a much lower concentration in a licorice extract, unless the extract was treated to enhance the concentration of the hydrolyzed product. Based on the area count, the GA concentration was less than 0.5% of the G concentration in extract A. Although the actual peaks were not counted, the calculated peak capacity (3) for the 5-m column was determined to be 290 (resolution: 1.0). Running the same extract A on the 1.8-m 150-mm column, one can see the finer structured (that is, higher resolution) chromatogram that results
(Figure 3b). The calculated peak capacity for this higher efficiency column was determined to be 442, over 50% higher than by using the longer 5-m column. Thus, it would be easier to determine minor components on this shorter rapid resolution column. The peaks per unit time (Resolution = 1.0) was calculated to be 17.7 peaks/min for the 1.8-m column versus 7.1 for the 5-m column. Figures 4a and 4b show similar runs using extract B. This extract was even more complex than extract A which is borne out by comparing the high resolution chromatograms of Figure 3b versus Figure 4b. Again, the calculated peaks per minute for the 1.8-m column greatly exceeded that of the 5-m column (17 versus 7.5 respectively). Based on the peak area counts for GA, it was roughly 1% of the concentration of G in extract B.
(A)
mAU 400 300 200 100 0 0 10 20 mAU 50 30 40 50 min
G 22.8 min
Peak capacity (@ Rs=1.0) = 290
GA 41.4 min
Analysis of licorice root extract A using a ZORBAX SB-C18, 4.6 250 mm, 5-m column
40 30 20 10 0 -10 -20 25 30 35 40 45 50 min
GA
(B)
mAU 450 400 350 300 250 200 150 100 50 0 0 5 10 15 mAU 70 20 25 30 min
Analysis of licorice root extract A using a ZORBAX SB-C18 RRHT 4.6 150 mm, 1.8-m column For conditions: see Experimental
60 50 40 30 15.896 20 10 0
GA
16
18
20
22
24
26
28
30
32
34 min
Figure 3a and 3b.
The gradient reversed-phase separation of licorice root extract A on the 5.0-m column (A) and on the 1.8-m column (B).
(A)
mAU 400
G 22.8 min
300 200 100 0 0 10 20 mAU 30
Peak capacity (@ Rs 1.0) = 306
GA 41.40 min
40
50
min
Analysis of licorice root extract B using a ZORBAX SB-C18 4.6 250-mm column, 5 m
70 22.016 60 50 40 30 20 10 0 25 30 35 40 45 50 min
GA
(B)
mAU 450 400 350 300 250 200 150 100 50 0 0 5 10 15 mAU 90 20 25 30
m i n
G 14.6 min
GA 25.5 min
Analysis of licorice root extract B using a ZORBAX SB-C18 RRHT 4.6 150-mm column
80 70 60 50 40 30 20 16 18 20 22 24
GA
26
28
30
32
34 min
Figures 4a and 4b. Gradient reversed-phase separation of licorice root extract B on the 5.0-m column (A) and on the 1.8-m column (B). 5
Conclusions
No attempt was made to perform quantitative analysis on the components of the licorice extracts. From our studies, it was obvious that resolution and throughput using the 1.8-m column greatly exceeded that obtained using the 5.0-m conventional column. As more complex samples of natural products are encountered and researchers require more detailed component analyses, the use of high resolution, small particle columns should grow. In the investigation of licorice root, other natural products, and TCMs, it is necessary to have gradient capability and sensitive detection.
References
1. S. Fanali, Z. Aturki, G. DOrazio, M. A. Raggi, M. G. Quaglia, C. Sabbioni, and A. Rocco, (2005) J. Sep. Sci., 28, 982986. 2. I. Kitagawa, (2002) Pure Appl. Chem. 74 (7), 11891198.

Complete Separation and Quantitation of Fusel Oils by Capillary GC Application
Authors
Eberhardt R. Kuhn and Allen K. Vickers Agilent Technologies, Inc. Life Sciences and Chemical Analysis 91 Blue Ravine Road Folsom, CA 95630 Susan E. Ebeler, Dawn M. Ahlgren, and John H. Thorngate University of California, Davis Department of Enology and Viticulture
amyl alcohol (2-methyl-1-butanol). They are formed through transamination of carbohydrates by amino acids:
Glucose -Keto-Acids Decarboxylation and Reduction Alcohols
Amino acid Leucine Isoleucine Valine Threonine
-Keto acid -Isocaproate -Ketoisovalerate -Ketobutyrate
Fusel alcohol 3-Methyl-1-butanol
-Keto- -methyl valerate 2-Methyl-1-butanol 2-Methyl-1-butanol 1-Propanol
2-Phenylalanine -Phenyl-2-ketopropionate 2-Phenylethanol
Abstract
Fusel oils are of great importance in the alcoholic beverages industry, since they affect the flavor and aroma of the beverage. Thus, their accurate quantitation is essential in assuring consistent quality of alcoholic beverages. Traditionally, packed GC columns were used for this analysis. Capillary columns do not offer the specific stationary phases that were available, and necessary, for packed columns. This application note describes the successful separation of all fusel oils, including other compounds typically found in alcoholic beverages, in a single run on a common stationary phase. In particular, baseline separation of methanol/acetaldehyde and isoamyl/active amyl alcohol was achieved. Examples include standards as well as real samples of fermented and distilled spirits, with quantitative data provided for a number of spirits. Fusel oils are important flavor constituents in alcoholic beverages. As a group they contribute fusel/diesel-like character. Individual aromas range from banana (isoamyl acetate) to rose-like (phenylethanol). At high levels they are considered undesirable; however, low to moderate levels contribute to the complexity of the beverage. Analysis of fusel oils is used to monitor distillation processes, malfunctions in distillation apparatuses, as well as confirming fermentation substrate authenticity. Thus, their accurate quantitation is essential in assuring consistent quality of alcoholic beverages. Separation of all fusel oils and the low boiling components on a single capillary column can be problematic. In particular, baseline separation of isoamyl alcohol (3-methyl-1-butanol) and active amyl alcohol (2-methyl-1-butanol) present some challenges. Traditionally, packed GC columns were used for this analysis. Capillary columns do not offer the plethora of different stationary phases that were available, and necessary, for packed columns to accomplish specific separations. In general, isoamyl- and active amyl alcohol will not
Introduction
Fusel oils are small alcohols that typically include 1-propanol, 1-butanol, isobutanol, as well as isoamyl alcohol (3-methyl-1-butanol) and active
separate on polar columns typically used for separation of alcohols (Figure 1). By contrast, their separation is easily achieved on a DB-MTBE (Figure 2), one of the least polar columns with respect to polar analytes. Unfortunately, resolution of other analytes typically found in alcoholic beverages, such as methanol and acetaldehyde, behaves in just the opposite manner (Figures 1 and 2). It would then stand to reason that a compromise of the two, that is, a mid-polarity column, should separate all 4 compounds. Complete baseline resolution of all fusel oils in about 13 minutes (Figure 3) was achieved with a DB-624.
SPME Fibers
Polyacrylate 85um (Supelco, Inc.) Carbowax-Divinylbenzene 65 um (Supelco, Inc.)
Sampling Conditions Beverage samples were diluted to 20% ethanol with deionized water prior to sampling. A 10 mL aliquot of standard or diluted sample was placed into a 20 mL headspace vial and 100 mL of IS was added prior to closing with a Teflon lined crimp top seal. The SPME fiber was inserted into the headspace and allowed to equilibrate at 25 C for 30 min. The fiber was then inserted into the GC inlet and desorbed at 250 C for 5 min. Statistical Analyses All samples and standards were analyzed a minimum of two times. Means, standard deviations, and relative standard deviations (%RSD) were calculated where appropriate. Linear calibration curves were prepared and used for quantitation of unknown samples.
Materials and Methods

Samples All alcoholic beverage samples were purchased from commercial sources. Standards All standards were purchased commercially and were of the highest grade available. A list of analytes is given in Table 1. Standard solutions containing 0, 10, 50, 100, 250, and 500 ppm (vol/vol) of each analyte were prepared in 20% (vol/vol) aqueous ethanol. Internal Standards Two internal standards were evaluated, 2-propanol and 3-pentanol. Separate solution of each internal standard were prepared by diluting 25 mL of the neat IS component to a final volume of 250 mL with absolute ethanol. GC Conditions
GC: Agilent 6890 Gas Chromatograph; ChemStation Software Autosampler: Column: Carrier gas: Oven: Gerstel Model MPS2 DB 624 60 m 0.25 mm id 1.4 m Helium at 35 cm/s at 40 C (1.9 mL/min) 40 C for 5 min; 10 C/min to 250 C Injector: 250 C Splitless; Split vent open at 5.00 min at 17.7 mL/min Detector: Agilent 5973 MSD; Interface 280 C

SPME Fiber Choice Two different fibers were evaluated for their response to the analytes monitored. Peak areas for early eluting analytes (methanol, acetaldehyde, ethanol, 2-propanol) were approximately two times greater using the carbowax-divinylbenzene fiber compared to the polyacrylate fiber. Response of the later eluting analytes (amyl alcohols, 1-hexanol, phenylethanol) was slightly higher using the polyacrylate fiber compared to the Carbowax fiber. The Carbowax fiber was used for all quantitative analyses. Internal Standards Both 2-propanol and 3-pentanol were evaluated for use as internal standards. The retention time for 2-propanol was close to that of ethanol. 3-Pentanol was well resolved from all other analytes (Table 1; Figure 3). Standard curves calculated using peak area ratios for both internal standards gave similar results in terms of linearity and reproducibility. 3-Pentanol was chosen as the internal standard for quantitation in this study.
Standard Curves
Table 1. Standard Curve Equations. Analyte Concentration vs. Peak Area Ratio for Analyte and Internal Standard Retention time (min) 4.6 4.9 7.3 9.7 10.7 11.9 13.1 14.0 15.2 15.3 18.5 24.5 10500 10500 10500 10500 Y = 0.0094(x) + 0.0607 Y = 0.013(x) + 0.2127 Y = 0.029(x) + 0.7569 Y = 0.0497(x) + 1.5034 0.9994 0.9928 0.9907 0.9918 Range (ppm) 10500 10500 10500 10500 10500 10500 10500 Equation Y = 6x 105(x) + 0.0048 Y = 0.0001(x) 0.0033 Y = 0.0007(x) 0.0006 Y = 0.0013(x) + 0.05 Y = 0.0049(x) + 0.0412 Y = 0.0037(x) + 0.0728 Y = 0.0038(x) + 0.0184 r2 0.9954 0.9938 0.9994 0.9637 0.9983 0.9716 0.9772
Analyte Acetaldehyde Methanol Acetone 1-Propanol Ethylacetate 2-Methyl-1-propanol 1-Butanol 3-Pentanol (IS) 3-Methyl-1-butanol 2-Methyl-1-butanol 1-Hexanol Phenyethanol
All standards prepared in 20% (vol/vol) ethanol.
Precision
Table 2. Analyte Acetaldehyde Methanol Acetone 1-Propanol Ethylacetate 2-methyl-1-propanol 1-Butanol 3-Methyl-1-butanol 2-Methyl-1-butanol 1-Hexanol Phenylethanol Relative Sstandard Deviation (%) for Replicate Analyses (n>3) at 5 Concentrations on Different Days 500 ppm 9.5 6.9 8.2 13.5 5.3 4.6 2.4 7.1 4.6 7.2 6.7 250 ppm 15.9 13.2 15.4 11.7 12.4 3.0 4.0 3.0 4.5 8.3 10.9 100 ppm 7.3 20.9 8.4 15.7 3.3 9.6 8.8 7.8 8.4 13.4 11.0 50 ppm 14.6 4.6 15.1 12.2 11.7 6.6 12.5 10.3 8.1 7.0 9.6 10 ppm 22.0 18.6 11.9 22.3 10.2 10.8 17.6 11.2 12.1 10.9 15.9
Note that %RSD is slightly greater at the low concentrations, particularly for the low molecular weight analytes. Volatilization from the standards between days and during analysis can occur and further work is needed to minimize this variation at low concentrations.
Analysis of Alcoholic Beverages Chromatograms of selected alcoholic beverages are presented in Figures 46. Quantitative results for individual analytes are given in Table 3. Analyte concentrations are typical of those reported for these beverages (Nykanen, 1986).
Table 3. Analyte
Fusel Alcohol, Methanol, Acetaldehyde, Acetone, and Ethylacetate Concentrations (ppm) in Commercial Beverages* Brandy A (80 proof) 14.7 <10 <10 54.4 78.2 142.8 <10 <10 571.9 153.3 <10 Brandy B (80 proof) 58.9 <10 <10 74.5 44.4 192.4 <10 <10 764.8 215.9 <10 Vodka (80 proof) <10 <10 <10 <10 <10 <10 <10 <10 <10 <10 <10 Gin (80 proof) 85.6 <10 <10 <10 <10 <10 <10 <10 <10 <10 <10 Scotch (80 proof) <10 <10 <10 107.9 74.3 261.4 <10 <10 87.2 40.9 <10
Acetaldehyde Methanol Acetone 1-Propanol Ethylacetate 2-methyl-1-propanol 1-Butanol 3-Methyl-1-butanol 2-Methyl-1-butanol 1-Hexanol Phenylethanol
*Concentrations reported in ppm (vol/vol).
Summary
Resolution of all analytes of interest was achieved using a mid-polarity stationary phase (DB-624). Further optimization of column length and film thickness provided baseline resolution in under 30 minutes for all components quantified. Use of SPME headspace sampling combined with either FID or MSD detection was easy, rapid (~30 min per sample) and readily automated. Volatility of the low molecular weight analytes requires careful sample preparation and temperature control to ensure reproducible results between days. This technique holds promise for the routine analysis of alcoholic beverages in order to monitor distillation processes, malfunctions in still operations, and fermentation substrate authenticity.
References
L. Nykanen, Am. J. Enol. Vitic., 37, 84 (1986).
Column: P/N: Carrier: Oven: Injector: Detector:
DB-WAX 30 m 0.25 mm id, 0.25 m 122-7032 Hydrogen at 45 cm/s 30 C for 5 min 3080 C at 1/min Split 1:50, 1.5 L FID, Nitrogen makeup gas at 30 mL/min
8 1 2 3 4 5 6 7 8 Acetaldehyde Ethyl acetate Methanol Ethanol 1-Propanol 3-Pentanol 2-Methyl-1-butanol 3-Methyl-1-butanol 7
2 1
CO57
14 min
Figure 1.
Scotch whiskey.
Column: Inlet: FID: Carrier: Oven:
DB-MTBE 30 m .45 mm id 2.55 m 250 C, split 300 C H2, 50 cm/s 40 C for 5 min. 10 C/min to 250 C
1 2 3 4
Acetaldehyde Methanol 3-Methyl-butanol (isoamyl alcohol) 2-Methyl-butanol (active amyl alcohol)
1, 2
4 3
C1101
4 Time (min)
10
Figure 2.
Fusel oil simple standard.
DB-624 60 m .25 mm id 1.4 m 250 C, split 300 C H2, 50 cm/s 40 C for 5 min. 10 C/min to 250 C
1 2 3 4 5 6 7
Acetaldehyde Methanol Ethanol Acetone 1-Propanol Ethyl acetate Isobutanol
8 9 10 11 12 13
1-Butanol 3-Pentanol (IS) 3-Methyl-butanol (isoamyl alcohol) 2-Methyl-butanol (active amyl alcohol) Hexanol Phenylethanol
1 2
Rs=1.8
7 9 8 11 10 6 C1098 0 5 10 Time (min) 15 12 13
20
Figure 3.
Fusel oil standard.
DB-624 60 m .25 mm id 1.4 m 250 C, split 300 C H2, 50 cm/s 40 C for 5 min. 10 C/min to 250 C 3
1 2 3 4 5 6 7
8 9 10 11 12 13 14
1-Butanol 3-Pentanol (IS) 3-Methyl-butanol (isoamyl alcohol) 2-Methyl-butanol (active amyl alcohol) Hexanol Phenylethanol Acetic acid
10 1
14 2 5 6 13 7 4 11 12
C1098
10 Time (min)
15
20
25
Figure 4.
Sherry sample.
DB-624 60 m .25 mm id 1.4 m 250 C, split 300 C H2, 50 cm/s 40 C for 5 min. 10 C/min to 250 C
1 2 3 4 5 6 7
8 9 10 11 12 13
9 10
5 12 4
11 6 7
12
13
C1098
10 Time (min)
15
20
25
Figure 5.
Brandy sample (SPME).
DB-624 60 m .25 mm id 1.4 m 250 C, split 300 C H2, 50 cm/s 40 C for 5 min. 10 C/min to 250 C 3
1 2 3 4 5 6 7
8 9 10 11 12 13
12
4 5
10 6 7
11
12
13
C1104 0
10 Time (min)
15
20
Figure 6.
Vodka sample (SPME).

Analysis of Essential Oil Compounds Using Retention Time Locked Methods and Retention Time Databases Application
Food and Flavors
Author
Frank David Research Institute for Chromatography Pres. Kennedypark 20, B-8500 Kortrijk Belgium Francis Scanlan Quest International Naarden, The Netherlands Pat Sandra Laboratory of Organic Chemistry University of Gent Krijgslaan 281, B-9000 Gent Belgium Michael Szelewski Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA
and two retention time databases were created. Flavor compounds in food extracts or essential oil constituents can be automatically searched based on retention times and/or mass spectra. Finally, transformation of existing retention index libraries into locked retention time databases is discussed.
Introduction
Capillary gas chromatography (GC) has been for many years the method of choice for the analysis of essential oils [1]. The constituents of essential oils are identified using a combination of different GC techniques, including GC with flame ionization detection (FID) and determination of retention indices, GC with olfactometric detection (sniffing), GC in combination with mass spectrometry (GC/MS) and GC with element-selective detection (flame photometric detection, nitrogen phosphorous detection, atomic emission detection, etc). Although GC/MS is probably the most powerful technique, and extended mass spectral libraries are available, it does not allow complete identification. Essential oils are complex mixtures of monoterpenes, monoterpenoids, sesquiterpenes, sesquiterpenoids, diterpenes, and diterpenoids. No single capillary column can resolve all possible compounds, and spectral data are not always conclusive because isomers give similar spectra. In flavor and fragrance quality control, retention indices are still frequently used as a complementary technique to GC/MS. Several libraries are
Abstract
Two retention time locked methods for the analysis of essential oil compounds are described. The first method is a gas chromatography/flame ionization detector method using a 50 m 320 m id 1.05 m HP-5 column. The second method is a gas chromatography/mass spectrometry method using a 30 m 250 m id 0.25 m HP-5MS column. Retention times of approximately 400 essential compounds were measured using both methods,
available with retention indices for many flavor and fragrance compounds [24]. Retention indices are less dependent on operational parameters than absolute retention times, but they still depend significantly on the column type (stationary phase and supplier), on the temperature program, and to a lesser extent, on the carrier gas velocity. Therefore, it is sometimes difficult to reproduce published retention indices in different laboratories. Moreover, most companies in the flavor and fragrance industry are still using in-house methods based on historical choices of columns and conditions. Finally, the use of retention indices (with n-alkanes as reference compounds) is not possible with element-selective detectors. A peak detected at a certain retention time using a sulphur selective detector, for instance, might be difficult to locate in an FID or MS chromatogram. Recent developments in GC have led to the ability of locking and matching retention times for a given application [5-6]. Using retention time locking, it is no longer necessary to calculate the retention index, but the absolute retention time can be used as an identification tool. Of course, retention times are still dependent on operating conditions, but small differences in carrier gas velocity and column length are compensated by re-locking the GC method by adjustment of the column headpressure. After re-locking, elution temperatures of the solutes are also constant. Moreover, retention time locking can also be used in combination with different detectors, and exact scaling of GC/FID, GC/sniffing, GC/MS, and GC/AED chromatograms is possible [6]. Retention time locking and retention time databases are, therefore, excellent tools in essential oil and in flavor QA/QC analysis. In this paper, two retention time locked methods are presented. For each method, a retention time database is available containing approximately 400 flavor compounds and essential oil constituents. The first method is a GC/FID method. A long, thick film column is used in combination with a slow temperature program. These conditions are frequently used in QA/QC analysis in the flavor and essential oil industry because a high sample capacity is combined with a high resolving power,
resulting in a detailed picture of the samples. The second method is a GC/MS method. For this method, a 30 m 0.25 mm id 0.25 m HP-5MS column was selected because this is the most frequently used column in GC/MS analysis. While the resolution and sample capacity are lower on this column, the GC/MS analysis mainly focuses on identification of analytes. For this method, a combined retention time and mass spectral library Screener Database is available. Because a lot of retention data are already available as retention indices, it was also evaluated if these data could be transferred into absolute retention times that match with locked retention times. It was shown that retention indices from existing retention index libraries can be recalculated as absolute retention times that match with experimental data.
Experimental
GC/FID analyses were performed on an Agilent 6980 gas chromatograph equipped with a split/splitless inlet. Separation was done on a 50 m 0.32 mm id 1.05 m HP-5 column (=72) (Agilent part number 19091J-215). The analytical conditions are summarized in Table 1. Helium at approximately 85 kPa (12.33 psi) constant pressure was used as carrier gas. The inlet pressure was adjusted to give a retention time of 70.000 min for n-pentadecane. This is done by retention time locking (RTL), using five runs at different pressures (respectively 70, 80, 90, 100, and 110 kPa), and plotting the retention time of n-pentadecane as a function of the inlet pressure [5]. From this curve, the inlet pressure can be calculated to adjust the retention time of n-pentadecane to exactly 70.000 min. The analytical conditions in this GC/FID method are typical conditions used in quality control of essential oils. The column choice and the slow temperature program offer high resolution and a detailed sample profiling. The column also offers high sample capacity, which is also important in essential oil profiling, because important trace constituents can be present and elute close to major constituents. On columns with restricted sample capacity, overloading and, consequently, peak leading is frequently observed for the main constituents.
Table 1. Column Injection Carrier RTL
GC/FID Conditions 50 m 0.32 mm id 1.05 m HP-5 (=72) (Agilent part number 19091J-215) Split, split ratio 25:1, 250 C, 1 L injection volume Helium (approximately 85 kPa) (12.33 psi), constant pressure The inlet pressure is adjusted to give a retention time of 70.000 min for n-pentadecane 50 C to 280 C at 2 C/min (110 min analysis time) FID, 300 C
Table 2. Column Injection Carrier RTL
GC/MS Conditions 30 m 0.25 mm id 0.25 m HP-5MS (=72) (Agilent part number 19091S-433) Split, split ratio 25:1, 250 C, 1 L injection volume Helium (approximately 65 kPa) (9.43 psi), constant pressure The inlet pressure is adjusted to give a retention time of 27.500 min for n-pentadecane 60 C to 240 C at 3 C/min (60 min analysis time) MS in scan mode (40_400 amu); solvent delay: 2 min; transfer line: 300 C
Oven program Detection
Oven program Detection
The second method is a GC/MS method. GC/MS analyses were performed on an Agilent 6980 gas chromatograph equipped with a split/splitless inlet in combination with an Agilent 5973N MSD. Separation was done on a 30 m 0.25 mm id 0.25 m HP-5MS column (=250) (Agilent part number 19091S-433). The analytical conditions are summarized in Table 2. Helium at approximately 65 kPa (9.43 psi) constant pressure was used as carrier gas. The inlet pressure was adjusted to give a retention time of 27.500 min for n-pentadecane. This is done by retention time locking, using five runs at different pressures, and plotting the retention time of n-pentadecane as a function of the inlet pressure. This is automatically done by starting the acquire RTL calibration runs command in the GC/MS instrument control. From this curve the inlet pressure can be calculated to adjust the retention time of n-pentadecane to exactly 27.500 min. These analytical conditions can be used to screen essential oils using GC/MS. Essential oil constituents can be identified based on the mass spectral data, and on retention times, using a screener library. The operational conditions are identical to the conditions used by Adams [4]. Spectra and retention data published in this reference are also valid for this method. Test mixtures of flavor compounds and n-alkanes were prepared from pure chemicals at 0.1% concentration in carbon tetrachloride or chloroform. Essential oil mixtures are diluted in carbon tetrachloride or chloroform at a 5% level (50 mg/mL).

GC/FID Method The described GC/FID method is used for quality control of essential oil mixtures. The long, thick film column results in high resolution and high sample capacity. Traces of important compounds can be detected besides the main constituents. A typical separation obtained by this method appears in Figure 1, showing the analysis of a Spanish orange oil. The chromatogram shows a detailed picture of the main compounds and of minor constituents. This type of analysis, giving both qualitative and quantitative information, is used for quality control of essential oils. This GC/FID method was locked to n-pentadecane (tR = 70.000 min). Under these locked conditions, n-decane elutes at 31.640 min and n-eicosane at 99.557 min. Using these conditions, a retention time locked database was created containing approximately 400 compounds that are important in quality control of essential oil mixtures. Using this database and the GC ChemStation RTL option, peaks in the GC chromatogram can be identified based on a retention time search in a given retention time window (for instance 0.2 min). For the 10 main peaks of the Spanish orange oil, the results of such a retention time search are given in Table 3. It is clear, that in some cases, several compounds elute in the 0.4-min window and further identification is needed. However, this tool already allows an excellent profiling of samples.
7 4
2 1
5 6 8
10 3
10
20
30
40
50
60
70
min
Figure 1. GC/FID chromatogram of Spanish orange oil. (Conditions: Table 1, peak identification: Table 3)
Table 3
Identification of main compounds in Spanish orange oil using a retention time database and a combined mass spectral and retention time identification. GC/FID* tR (min) 26.793 30.042 tR identification -pinene 1-octen-3-ol 3-(methylthio)-1-propanol sabinene hexanoic acid -pinene 6-methyl-5-hepten-2-one 2-octanone myrcene furfuryl acetate octanal trans-2-hexenoic acid -3-carene limonene benzylalcohol ocimene GC/MS tR (min) 5.172 6.181 MS + tR identification -pinene sabinene
Peak number 1 2
30.539
6.282
-pinene
31.053
6.658
myrcene
5 6 7
31.987 33.190 35.001
6.987 7.267 8.130
octanal -3-carene limonene
Table 3 Peak number 8
Continued GC/FID* tR (min) 40.162 GC/MS tR (min) 10.391
tR identification n-undecane cis-3-hexenylpropionate -hexalactone 1-methyl-2,3-cyclohexadione linalool methyl benzoate dihydrocarveol methyl salicylate estragole n-decanal octylacetate anisylproprionate valencene piperonyl acetate
MS + tR identification linalool
48.728
14.750
n-decanal
10
71.366
27.134
valencene
* For the GC/FID retention time identification, a 0.4-min window was used (0.2 min)
Another example of the GC/FID method appears in Figure 2, showing the analysis of an Indonesian nutmeg oil. Again a very detailed chromatogram is obtained. Using the retention time locked database, most constituents are identified. The small peak eluting at 67.468 min is, for instance, identified as isoeugenol. This is an important flavor compound.
67.468
10
20
30
40
50
60
70
80
90
min
Figure 2.
GC/FID chromatogram of Indonesian nutmeg oil. (Conditions: Table 1)
GC/MS Method For confirmation of solute identities and for the identification of unknown peaks, the essential oils are analyzed by GC/MS. The described method uses a standard column and a faster temperature program. These conditions are similar to the method published by Adams [4]. The chromatogram obtained for the Spanish orange oil is given in Figure 3. A similar separation is obtained as in Figure 1, but the resolution is lower due to the lower column plate number. Moreover, some peak overloading can be observed. Due to the fact that a different temperature program is used, the compounds also elute at different temperatures (method not translated) and, therefore, also some
differences in relative elution profiles are observed (see, for instance, relative elution of peaks 2, 3, and 4). Nevertheless, this method can be used for detailed identification of the essential oil constituents. For this GC/MS method, a mass spectral library was created containing approximately 400 compounds, together with the retention times. With this library, identification is possible based on mass spectra AND on retention time. The identification of the 10 main compounds (same as labelled in Figure 1) using the Results Screener is included in Table 3. The combination of mass spectral and retention time data allows complete identification. It is also important to notice that the correct compound was, in all cases, already selected in the GC/FID retention time window.
9 5 8 10 6 3
10
12
14
16
18
20
22
24
26
28
min
Figure 3.
GC/MS chromatogram of Spanish orange oil. (Conditions: Table 2, peak identification: Table 3)
The Indonesian nutmeg oil was also analyzed by this GC/MS method. The chromatogram is given in Figure 4. A classical library search using the standard NIST mass spectral library of the peak at 25.304 min gave isoeugenol as the first hit (probability 96%) and eugenol as the second hit (probability 94%). The spectra of both compounds are very similar (Figure 5). Using the Results Screener, the compound was identified unequivocally as isoeugenol (retention time plus mass spectral match). This example clearly demonstrates the power of combined retention time and mass spectral search.
25.304
10
15
20
25
30
35
40
45
min
Figure 4.
GC/MS chromatogram of Indonesian nutmeg oil. (Conditions: Table 2)
164
Abundance 9000 8000 7000 6000 5000 4000
273 Isoeugenol
149
3000 2000 1000 0
77 55 32
30
91
103 121
131 138
120 130 140 150 160 170
39
40 50 60
65
70 80 90 100 m/z
110
110
Abundance 9000 8000 7000 6000 5000 4000 3000 2000 1000 0 30 40
164
72243 Eugenol
149 77 55 39 27 46
50 60
103 91 65 121
131
138
70 80 90 100 m/z 110 120 130 140 150 160 170
Figure 5.
Comparison of mass spectra of eugenol and isoeugenol.
Transformation of Retention Indices Further it was evaluated if published retention indices could be transferred into retention times and if these calculated retention times match with experimental data. A total of 34 test solutes were analyzed. The compounds are listed in Table 4, together with the FEMA code, the retention index in an existing database (RI) [7], and the measured retention time (tR exp) under retention time locked conditions. From the retention index, the absolute retention time was calculated using the retention times of n-alkanes as reference compounds. The calculated values are also listed in Table 4 (tR calc).
Thus, these retention times are not the original retention times used for the retention index calculation, but calculated values. In the last column, the difference between calculated and experimental retention times are also given. From these data, it is clear that the calculated and experimental retention times match very well (within 0.2 min). This means that retention times can be calculated from the retention indices present in an existing database using the locked retention times for n-alkanes if the column dimensions and the temperature program are the same. This is also valid for the GC/MS method as shown in the following example. Isoeugenol is present in the database of
Table 4.
FEMA names, FEMA codes, retention indices, calculated retention times, experimental retention times, and retention time differences for test solutes. FEMA name Acetal Amyl alcohol Hexyl alcohol Anisole Ethyl acetoacetate Heptyl alcohol Octanal Methyl 3-(methylthio)propionate Benzyl alcohol Isoamyl butyrate Octyl alcohol Acetophenone Benzylformate Benzyl acetate Allylheptanoate Decanal Benzyl propionate 1-Decanol Anisyl alcohol Isobornyl acetate Benzyl isobutyrate Undecanal Triacetin Benzyl butyrate Acetanisole gamma-Nonalactone Anisyl acetate Allyl cyclohexylpropionate Lauryl alcohol Isoamyl octanoate Isoamyl phenylacetate Ethyl-methylphenylglycidate Ethyl-3-phenylglycidate gamma-Undecalactone FEMA code 2002 2056 2567 2097 2415 2548 2797 2720 2137 2060 2800 2009 2145 2135 2031 2362 2150 2365 2099 2160 2141 3092 2007 2140 2005 2781 2098 2026 2617 2080 2081 2444 2454 3091 RI 725.1 760.0 865.8 923.0 944.4 968.4 1003.8 1026.8 1036.9 1055.5 1069.7 1072.4 1081.2 1169.1 1180.0 1207.4 1266.2 1272.1 1295.0 1301.7 1305.0 1310.6 1344.4 1354.5 1369.0 1373.4 1426.0 1435.0 1475.0 1487.0 1503.0 1517.0 1529.0 1589.0 tR calc (min) 11.836 13.844 21.068 25.529 27.300 29.285 32.218 34.140 34.984 36.539 37.726 37.952 38.688 45.848 46.729 48.917 53.444 53.899 55.662 56.171 56.411 56.819 59.279 60.014 61.070 61.390 65.116 65.735 68.489 69.315 70.405 71.317 72.099 76.007 tR exp (min) 11.836 13.833 20.941 25.403 27.319 29.180 32.215 34.121 35.011 36.524 37.663 38.119 38.851 45,915 46.664 48.970 53.378 53.904 55.799 56.252 56.452 56.798 59.192 60.049 60.960 61.273 65.203 65.747 68.531 69.230 70.387 71.447 72.063 75.955 tr diff (min) 0.000 -0.011 -0.127 -0.126 0.019 -0.105 -0.003 -0.019 0.027 -0.015 -0.063 0.167 0.163 0.067 -0.065 0.053 -0.066 0.005 0.137 0.081 0.041 -0.021 -0.087 0.035 -0.110 -0.117 0.087 0.012 0.042 -0.085 -0.018 0.130 -0.036 -0.052
Compound 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
Adams [4] with a retention index of 1447. This retention index can be transferred into an absolute retention time using the following formula: 100)] t ( [[RI (Z 100 _t
RZ + 1
RZ
)] + t
RZ
= tRX
whereby: RI = retention index (from existing data base), Z = carbon number of preceding n-alkane, tRZ+1 and tRZ = retention times of following and preceding n-alkanes (in RTL method) and tRX = retention time of solute in RTL method
For isoeugenol, with an RI = 1447, and the preceding tetradecane (Z=14) eluting at 23.26 min (tRZ), and the following n-pentadecane (Z+1) eluting at 27.50 min (tRZ+1) using the retention time locked GC/MS method, the calculated retention time is 25.25 min. This corresponds well (within 0.2 min) with the measured retention time (25.30 min). Using these calculations, compounds can also be added to the RTL databases.
Conclusion
Two methods were developed for the analysis of essential oils. The first method is used for quality control analysis. The method is locked using n-pentadecane as locking standard. A retention time locked database, containing approximately
9
400 compounds was created. This database can be used to identify constituents based on their absolute retention time under the locked conditions. The locked method also guarantees retention time stability in function of time, between columns and between instruments. Secondly, a GC/MS method was developed. This method can be used for identification of essential oil constituents. Identification is done based on the combination of retention time and mass spectral matching. Finally, it is shown that retention indices for flavor compounds measured under specific operational conditions can be transferred into locked retention times using the locked retention times of n-alkanes. Thus, existing retention index databases can be translated into locked retention time databases. Moreover, a retention time locked database is not restricted to the use of one (FID) detector, but compounds detected by any GC detector can be searched if the locked method is used.
References
1. R. Tabacchi and J. Garnero in P. Sandra and C. Bicchi (Eds) , Capillary Gas Chromatography in Essential Oil Analysis, Huthig, Heidelberg, 1987, pp 1-11. 2. W. Jennings and T. Shibamoto, Qualitative Analysis of flavor and Fragrance Volatiles by Glass Capillary Gas Chromatography, 1980, Academic Press, New York. 3. T. Shibamoto in P. Sandra and C. Bicchi (Eds) , Capillary Gas Chromatography in Essential Oil Analysis, Huthig, Heidelberg, 1987, pp 259-274. 4. R.P. Adams, Identification of Essential Oil Components by Gas Chromatography - Mass Spectroscopy, 1995, Allured Publishing Corporation, IL, USA. 5. V. Giarrocco, B.D. Quimby and M.S. Klee, Agilent Application Note 228-392, publication (23) 5966-2469E, December 1997. 6. B.D. Quimby, L.M. Blumberg, M.S. Klee and P.L. Wylie, Agilent Application Note 228-401, publication (23) 5967-5820E, May 1998. 7. Quest International, Naarden, the Netherlands.

Epigallocatechin 3-O-Gallate Extract from Green Tea
Application
Food Analysis
Robert Ricker
Epigallocatechin 3-O-gallate (EGCG) belongs to a class of compounds called flavonoids and is further subclassified as a flavanol, due to the level of oxidation in its chemical structure. This compound has been recognized as a cancer-preventive compound due to its ability to inhibit urokinase, an enzyme which has been associated with excelerated tumor cell growth. Due to the interest in holistic-type, preventative medicine approaches in today's society, a method was developed for a series of catechins and an actual extract of green tea to serve as a general interest application.
Highlights
Catechin Mixture
20
Green Tea Extract

2
2 1. Epigallocatechin 2. Epicatechin 3. Epigallocatechin gallate 50 4. Catechol 5. Epicatechin gallate
Good peak shape and resolution are maintained for a group of catechins on Agilent ZORBAX SB-C8 at low pH.
15
3
10
unknown
5
Sterically protected bonded phases, like SB-C8, offer extended column lifetime even with TFA-containing (low-pH) mobile phases.
25
1 3 4 Good retention of the catechins allows adequate separation from other UV-absorbing compounds in the actual tea extract.
15
0 0
10
15
5 Minutes
10
ZORBAX SB-C8 (4.6 x 150 mm; 3.5 m) (Agilent Part No. 863953-906) Mobile Phase: 75% 0.1% Trifluoroacetic acid: 25% Methanol Inj. Vol. 5 L, 1 mL/min, 40C Det. UV (280 nm)
Flavoring Agents
Application
Food Analysis
Robert Ricker
Several flavoring agents can be found in Cool Mint Listerine, a popular mouthwash. Carvone is oil of caraway, anethole is a constituent of anise and fennel oils, which has a licorice flavor and methyl salicylate (wintergreen oil). This sample also contains the antiseptic, thymol and the preservative, benzoic acid.
Highlights
Rapid analysis of complex mixtures while maintaining resolution can be achieved with shorter length (50 mm) columns.
2
Cool Mint Listerine Sample 1. Unknown 2. Benzoic Acid 3. Methyl Salicylate 4. Carvone 5. Unknown 6. Thymol 7. Anethole
Sample preparation is minimal for liquid samples which can be diluted and injected directly onto the column.
This method can be applied directly to LC/MS analysis.

7
Sterically protected bonded phases provide superior lifetime at low pH.

6 5
Conditions: Column: Agilent ZORBAX SB-Phenyl, Narrow-Bore LC/MS, 2.1 x 50 mm (5m), Agilent PN 860975-912 Mobile Phase: 0.3% TFA : ACN, 65:35 UV: 254 nm; Flow: 0.3 mL / min.; Ambient
Aromatics
Application
Food Analysis
Robert Ricker
Aromatics are used in both the flavor and fragrance industries. Only a few of the many compounds that can be classified as aromatic were separated for this application note. The odors of the various aromatics in this study are as follows: acetophenone has an orange-blossom-like odor; cinnamaldehyde is found in some cinnamon oils and has a characteristic cinnamon odor; eugenol has an odor of cloves and is also used as a dental analgesic; ethyl cinnamate has a fruity and balsamic odor reminiscent of cinnamon. Cymene occurs in a number of essential oils, but no odor is described in the literature.
Highlights
Neutral compounds have excellent peak shape on Agilent ZORBAX Eclipse XDB-Phenyl columns.
1 Aromatic Sample 1. Acetophenone 2. Cinnamaldehyde 3. Eugenol 4. Cinnamaldehyde Impurity 5. Ethyl Cinnamate 6. p-Cymene
Efficiency can be increased with the use of smaller size (3.5 m vs. 5.0 m) particles. Plate counts of 13,000 18,000 were achieved for the five compounds in this application.
2 5 3
Conditions: Column: ZORBAX Eclipse XDB-Phenyl, 4.6 x 150 mm (3.5m) Agilent P/N: 963967-912 Mobile Phase: H2O : MeOH, 40:60 UV: 254 nm; Flow: 1.0 mL / min.; 35C
Analysis of Flavonoids in Plant Extracts by CE-MS

Martin Greiner and Gordon Ross Food Abstract Complex extracts of plant compounds often need a very effective separation mechanism for a clear compound identification. MS detection can provide a specific identification of the analyte in such complex and difficult matrices. Here we show the analysis of flavonoids from extracts of oranges, illustrating the specificity and sensitivity of an Electrospray Mass Spectrometer (ESIMS) which can add a new dimension to capillary electrophoresis separations. Hesperidine, a flavone glucoside, is present in the peel of green oranges while naringin, a rhamnoglucoside flavanone, is a bitter component of the orange peel. If not controlled properly, the high squeeze pressure applied to oranges during production of orange juice yields a high content of these flavonoids which gives it a bitter taste. Monitoring the presence of these in orange juice provides a way of monitoring the quality of the product. CE was the separation technique of choice to allow a fast and efficient separation and additionally to reduce sample preparation to centrifugation and filtering of the juice. Abundance 4000 3500 3000 2500 2000 1500 1000 500 0
Conditions
Hesperetin m/z 325 Sample spiked orange juice Injection 1500 mbar s e c Capillary 75 cm (22 cm UV) x 50 m id Buffer 5 mM borate pH 9.3 Voltage 30 kV Temperature 25 C Sheath liquid 4 l/min, 50 % MeOH/H20, 1 % formic acid Nebulizing gas 20 psi Drying gas 6.0 l/min Acquisition positive SIM 325 m/z 581 m/z
Naringin m/z 581
5 4 3 Time [min]
Figure 1 Analysis of hesperetin and naringin
Experimental All experiments were performed using a Agilent Capillary Electrophoresis system equipped with DAD detection and controlled via Agilent CE ChemStation software. The Agilent capillaries for CEMS were 50 m id with a total length of 75 cm. These capillaries have a UV detection window at 22 cm. The buffer used was 50 mM borate, pH 9.3 diluted tenfold by Agilent CE-grade water (all Agilent Technologies, Waldbronn, Germany). Sheath flow (4 l/min) was delivered by an Agilent 1100 isocratic HPLC pump running at 400 l/min split 1:100 by an Agilent splitting device. Sheath liquid was 1 % formic acid in 50 % methanol for positive ion detection. Figure 1 shows hesperetin (a glycon of the hesperidine), found at m/z value 325 (the Na adduct mass). Naringin was found under these conditions as the protonated adduct at m/z 581. Besides the qualitative aspects of identifying compounds by mass spectra CE-MS can also provide quantitative information. The calibration curve shown in figure 2 resulted from a naringin standard diluted in a range from 200 to 2 g/ml. The lowest amount used here still gave a reasonable signal larger than 5:1 s/n. Figure 3 shows the MS trace of 2 g/ml naringin. Area 1600000
6 3 4 12 5
Equipment
Agilent Capillary Electrophoresis system Agilent CE-MS Adapter Kit Agilent 1100 Series LC/MSD module with API Electrospray Source Agilent CE-ESI-MS Sprayer Kit Agilent ChemStation +CE-MS Add-on software
Figure 2 Linearity of naringin detection with MS
800000 0
Correlation: 0.99884 100 Amount [g/ml] 200 Figure 3 Analysis of 2 g/ml naringin
20
Area 3500 3000 2500 2000 1500 1000 1 2 3 6 5 4 Time [min] 7 8 9
Analysis of Bitter Compounds, in Orange Juice using HPLC
Abstract The following compounds are examples of flavoring agents used in food products: lupulon and humulon (hop bittering compounds) vanillin naringenin and hesperidin (bittering compounds) Three major classes of compounds are used as flavoring agents: essential oils, bitter compounds, and pungency compounds. Although the resolution afforded by gas chromatography (GC) for the separation of flavor compounds remains unsurpassed, HPLC is the method of choice if the compound to be analyzed is low volatile or thermally unstable. Sample preparation for bitter compounds in orange juice1
Conditions
The samples were prepared according to Carrez 1 and 2. This method uses potassium ferrocyanide and zinc sulfate for protein precipitation. Column 125 4 mm Hypersil BDS, 5 m Mobile phase A = water + 0.15ml/l H2SO4 (conc.), pH = 2.4 B = ACN Gradient start with 20% B; at 3 min 20% B at 5 min 90% B; at 6 min 20% B Flow rate 2 ml/min Post time 1 min Column compartment 40 C Injection vol 1 l Detector UV-DAD detection wavelength 260/80 nm, reference wavelength 380/80 nm Sample preparation The orange juice was prepared according to Carrez 1 and 2
mAU 20 15 10 5 0 -5 0.5 1
Time [min] Naringenin
Hesperidin
Standard
Orange juice
1.5
2.5
Figure 1 Analysis of bitter ccmpounds in orange juices
Chromatographic conditions The HPLC method presented here for the analysis of hesperidin and naringenin is based on reversed-phase chromatography. UV spectra were evaluated as an additional identification tool. HPLC method performance Limit of detection 1 ng (injected amount), for DAD S/N = 2 Repeatability of RT over 10 runs <0.2 % of areas over 10 runs <1 % References 1. Official Methods of Analysis; Horwitz, W., Ed.; 14th ed.; AOAC: Arlington, VA, 1984; secs 12.01812.021.
Equipment
Miscellaneous
Scale-Up of Anthocyanin Separations and Re-Analysis of Collected Fractions on an Agilent Prep-C18 Column Application
Biochemical
Cliff Woodward
Anthocyanins, potent anti-oxidants, are now recognized as health important components of many foods. Blueberries, of all natural foods, contain the highest concentrations of these interesting compounds. The purification and identification of various anthocyanins is an important step in understanding which components are most beneficial to human health. Anthocyanins are highly retained on C18 columns; in addition, they require high concentrations of organic solvents for extraction. In attempting to separate large quantities of some of the components for use as standards or for further structural workups, the analyst may encounter solubility issues, which can limit the loadability. The factors affecting loadability of anthocyanins were explored in previous work [1]. The first step in preparative purification is to prove that scale-up works. Once that is known, preparative chromatography is simple. Fractions can be collected and rechromatographed to demonstrate purity. Figure 1 shows the scalability of the Agilent Prep-C18 column. The resolution of the analytical column is, of course, better than the prep column.
Highlights
The Agilent Prep-C18 columns demonstrate excellent scalability, allowing method development on analytical scale columns The Agilent Prep-C18 enables high resolution and high purity fractionation The Purification software of the Agilent ChemStation combined with the Agilent Prep-C18 makes complex purifications easy
An Agilent Prep-C18 column is shown above with an Agilent 1100 HPLC system.
Agilent Prep-C18, 21.2 250 mm, 10 m (p/n 410910-102) 1A
10
20
30
40
50
60
70
80 min
Agilent Prep-C18, 4.6 250 mm, 5 m (p/n 440905-902) 1B
10
20
30
40
50
60
70
80 min
1A Agilent Prep-C18, 21.2 250 mm, 10 Temperature: Ambient DAD wavelength: 525 nm Injection: 2000 L Sample: Blueberry extract, 46.1 mg/mL total dissolved solids (~5 mg/mL anthocyanins) Flow: 21.2 mL/min 1B Agilent Prep-C18, 4.6 250 mm, 5 Temperature: Ambient DAD wavelength: 525 nm Injection: 100 L Sample: Blueberry extract, 46.1 mg/mL total dissolved solids (~5 mg/mL anthocyanins) Flow: 1.0 mL/min Figure 1. Scalability of Agilent Prep columns.
Mobile phase A = 0.1% TFA in water B = 0.1% TFA in methanol Gradient timetable Time (min) 0.00 35.00 85.00 % Solvent B 23.0 26.0 53.5
By using the ChemStation fraction collection software, fractions are obtained that are significantly purer than the separation would seem to allow. See Figure 2. Fraction 1 is >99% pure Delphinidin3galactoside and Fraction 2 is >97% pure Delphinidin3glucoside. This purity was obtained using the threshold and slope settings of the software to cut the fractions appropriately. Identities of the fractions were verified by liquid chromatography/mass spectrometry (LC/MS) (data not shown). The conditions for the Agilent Prep-C18 columns below are the same as in Figure 1.
Slope setting Agilent Prep-C18 21.2 250 mm, 10 Threshold setting Actual peaks collected are between the dashed vertical lines
10
15
20
25
30
35
40
min
Agilent Prep-C18 4.6 250 mm, 5 Fraction 1 rechromatographed
Delphinidin-3-galactoside
10
15
20
25
30
35
40
min
Agilent Prep-C18 4.6 250 mm, 5 Fraction 2 rechromatographed

5 10 15 20
Delphinidin-3-glucoside
25
30
35
40
min
Figure 2.
Fraction collection and rechromatography to demonstrate purity.
References
1. "Scalability and Volume Loadability for Highly Retentive Compounds Anthocyanins", (2004) Separation Times, 17, 1.

For more information on our products and services, visit our Web site at www.agilent.com/chem. Search Agilent Prep. The author, Cliff Woodward is an Application Biochemist based at Agilent Technologies, Wilmington, Delaware.
Using Lead Isotope Ratios to Distinguish between Samples of the Traditional Chinese Medicine Dan-shen
Geological
Authors
Dengyun Chen Agilent Technologies Co., Ltd China Chaoyong Yang, Zhiyong Huang & Xiaoru Wang Analytical Sciences Laboratory Xiamen University Xiamen 361005 China
Most Traditional Chinese Medicines (TCMs) are a mixture of several different herbs that undergo special treatment to make them useful. The amounts of the effective components in the plants are influenced by the soil, water, weather conditions etc in the area where they are grown. Experienced practitioners of traditional Chinese healing know that the quality of herb medicines is strongly related to their source of origin. It is therefore important to know the specific location of a TCM herb component so ensure its effectiveness. Many techniques have been applied to characterize various herb medicines and to correlate them with their place of origin. High Performance Liquid Chromatography, Mass Spectrometry, Nuclear Magnetic Resonance, Infrared Spectroscopy and X-ray Fluorescence have all been used to produce spectra that can be used to fingerprint the TCM habitat. Lead isotope ratio measurements provide analytical information relating to the source of lead contamination in naturally occurring samples. Concentration measurements cannot provide this information. Studies of the isotopic composition of lead are therefore commonly used in environmental science as well as geological and anthropological studies. Among all the naturally occurring lead isotopes, only 204Pb is non-radiogenic, whereas, 206 Pb, 207Pb and 208Pb are the daughter products from the radioactive decay of 238U and 235U and 232Th respectively. As a consequence, small Pb isotope abundance variations occur in nature and the isotopic composition of lead in the environment is dependent on local ore deposits. If lead is present in the soil, a plant will take up small amounts and subsequent isotope ratio studies might provide a unique
Abstract
Quadrupole Inductively Coupled Plasma Mass Spectrometry (ICP-MS) was used to determine lead (Pb) isotope ratios in Dan-shen, a type of herb used in Traditional Chinese Medicines (TCM), and in water and soil samples all taken from the same geographical location. The precision obtained for the 208Pb/206Pb ratio, the 207Pb/206Pb ratio and the 204Pb/206Pb ratio values was considerably lower than 0.5% demonstrating the applicability of the technique for Pb isotope ratio studies. The results show that it is possible to distinguish Dan-shen samples originating from different geographical areas using Pb isotope ratio measurements. As the medicinal effectiveness of a TCM is highly dependant on the source of origin of its herb components, it is useful to have a reliable, routine means of fingerprinting the components grown in different habitats.
Introduction
means of differentiating between different plant sources of origin. Of course, local lead levels may become mixed with external sources of contamination (e.g. automobile exhaust) that vary with time depending on the anthropogenic activity. These mixing processes can be quantified as long as each lead source shows a different lead isotope abundance. The work described in this application note was undertaken to investigate if lead isotope ratios can be used to fingerprint the Dan-shen herb grown in different habitats. While isotope ratio studies have traditionally been undertaken with a multicollector mass spectrometer such as Thermal Ionization Mass Spectrometry (TIMS), the technique is relatively slow and the instrumentation is expensive. These experiments were carried using quadrupole Inductively Coupled Plasma Mass Spectrometry (ICPMS) that provides a fast, convenient and precise method to determine isotope ratios. The Agilent 7500 Series ICP-MS is an excellent tool for routine isotope ratio measurements. It is easy to operate with its fully automated optimization system (Auto-tune function) and its user-friendly software. Because of these advantages, Agilent ICP-MS systems have been widely used for isotope ratio studies for environmental monitoring applications and nuclear research projects, e.g., U isotope ratios in human urine [2], in the cooling water of nuclear plants etc., the certified values of some certified reference materials such as rock samples [3], biological samples [3,4] by isotope dilution. In Encinar et al.s work[4,5], an older model Agilent ICP-MS (HP-4500) was used, and the performance was as good as the more expensive multi-collector ICP-MS and double focusing sector field ICP-MS especially for long term stability and isotope dilution applications.
SRM981 (National Institute of Standards & Technology, USA) - for mass bias correction. HBr, HCl, HF, and HNO3, (GR grade, after subboiling purification) - for sample digestion and dilution. Anion exchange resin (DOWEX 18, 200-400 mesh, USA), - for separating Pb from the matrix in the samples. Ultra Pure Water, supplied by Milli-Q water system (18.2 MOhms) Sample preparation Dry samples and filtered water samples were digested using a HNO3/HF acid mixture in a microwave oven then heated to dryness in a fume cupboard to remove the HF. 1.0 mL HBr (0.5N) was used to redissolve the residue. Samples were then passed through a DOWEX 18 column to remove the matrix. The Pb containing eluant was dried and diluted with 3% HNO3. In order to avoid memory effects and to keep the ICP-MS detector working in pulse counting mode only, the concentrations were controlled between 1 and 80 ug/L. Optimization of the ICP-MS operation parameters Since the sample volumes collected after the column pretreatment step were limited, most of them were less than 1 mL, the Agilent PFA micro flow nebulizer (100 uL type) was selected for sample introduction. Self-aspiration mode was used and the carrier gas flow rate was controlled to maintain the sample introduction flow rate at about 50 uL/min. The 1 mL sample was sufficient for 2 or 3 individual measurements so that repeat analyses could be undertaken in case of error during the determination. Instrument sensitivity and the other parameters, such as oxide level and doubly charged ions, were optimized automatically using the AutoTune feature of ChemStation on a 10 ppb tuning solution containing Li, Y, Ce, Tl. The carrier gas was then set to 0.4 L/min to obtain a sample uptake rate of about 50 uL/min. Finally, the make-up gas was adjusted to produce the best sensitivities and lowest interferences. Instrument operating parameters are given in Table 1.
Experimental
Instrumentation An Agilent 7500i ICP-MS, (Agilent Technologies, Palo Alto, CA, USA), fitted with a PFA micro flow nebulizer (100 uL/min) in self-aspiration mode, a quartz spray chamber and a one-piece quartz torch was used throughout the study. A MK-II microwave sample digestion system (Shang Hai Xin Ke Factory, China) was used for sample digestion. Reagents The following reagents were used during the course of the study:
Data Acquisition Parameters: Table 1. ICP-MS operating conditions Isotope Ratio Analysis mode, 1000 scan/sec, minimum dwell time 100 us. The integration times were 20sec for 204 Pb, 10sec for 206Pb and 207Pb, 5 sec for 208Pb. Each data point is the average of 3 repetitions.
Plasma Conditions: RF power: RF Matching: Sample Depth: Torch-H: Torch-V: Carrier Gas: Make-up Gas: Peripump1: Spray Chamber Temperature: 1250 W 1.7V 7 mm 0 mm 0.3 mm 0.4 L/min 0.9 mL/min 0.1 rps 2.0 degC
Results
Before beginning the analysis, the instrument was allowed 30 minutes following plasma ignition to reach thermal stability. A 20ppb solution of NIST SRM 981 isotope ratio standard was measured after every 5 unknowns. This data provided the 7500i short-term and long-term stability information and provided a means of making the small mass bias corrections required. The instrument demonstrated excellent precision. See Tables 2 and 3. Note that the instrument consistently delivered %RSDs at less than 0.1% over the short and long term.
Table 2: Short-term Stability of Pb Isotope Ratio Determinations of 20 ppb SRM981 using the Agilent 7500i ICP-MS (20min, 5 measurements) File Name Acq Date Acq Time 204/Total 206/Total 207/Total 208/Total 204/206 207/206 208/206 PB-206A.D Jun 9 2001 3:45 PM 0.01427 0.2413 0.2211 0.5233 0.05913 0.9163 2.169 PB-206B.D Jun 9 2001 3:50 PM 0.01426 0.2414 0.2208 0.5235 0.05904 0.9146 2.168 PB-206C.D Jun 9 2001 3:55 PM 0.01426 0.2414 0.2209 0.5235 0.05907 0.9149 2.169 PB-206D.D Jun 9 2001 4:00 PM 0.01426 0.2415 0.2209 0.5233 0.05906 0.9146 2.167 PB-206E.D Jun 9 2001 4:06 PM 0.01425 0.2416 0.2209 0.5232 0.05897 0.9144 2.165 RSD(%)
0.05 0.05 0.05 0.03 0.10 0.08 0.08
Table 3. Long-term Stability (10 hours) of Pb Isotope Ratio Determinations of 20 ppb SRM981 using the Agilent 7500i ICP-MS File Name PB20-A.D PB20-B.D PB20-C.D PB20-D.D PB20_1A.D Pb-202A.D PB_202B.D PB-203A.D PB-204A.D PB-204B.D PB_205A.D PB_205B.D RSD(%) Acq Date Jun 9 2001 Jun 9 2001 Jun 9 2001 Jun 9 2001 Jun 9 2001 Jun 9 2001 Jun 9 2001 Jun 9 2001 Jun 9 2001 Jun 9 2001 Jun 9 2001 Jun 9 2001 Acq Time 6:14 AM 6:26 AM 6:33 AM 6:40 AM 7:59 AM 9:12 AM 9:18 AM 11:17 AM 12:52 PM 12:59 PM 2:27 PM 2:32 PM 206/Total 0.2411 0.2412 0.2412 0.2411 0.2411 0.2410 0.2414 0.2413 0.2412 0.2410 0.2412 0.2415 0.06 207/Total 0.2209 0.2209 0.2209 0.2210 0.2211 0.2211 0.2210 0.2211 0.2212 0.2214 0.2210 0.2210 0.07 208/Total 0.5235 0.5236 0.5235 0.5236 0.5236 0.5236 0.5234 0.5233 0.5233 0.5234 0.5236 0.5232 0.03 207/206 0.9162 0.9159 0.9158 0.9167 0.9172 0.9173 0.9153 0.9161 0.9168 0.9185 0.9163 0.9153 0.10 208/206 2.171 2.171 2.170 2.172 2.172 2.172 2.168 2.168 2.169 2.172 2.171 2.167 0.08
The excellent results from the NIST quality assurance sample suggest that the data from the unknown samples is also very reliable and can be reported with a high degree of confidence. In this study, three types of samples were measured: Four samples of surface water taken from different areas in Zhong-Jiang, Si-Chuan Province Soil samples came from Er-Mei Mountain (2), Zhong-Jiang, Si-Chuan Province (5), Jiang-Xi Province (1)
Dan Shen plant samples were from Zhong-Jiang, Si-Chuan Province (3), Er-Mei Mountain (2), Bei-Jing (1), Jiang-Xi Province (1), Tai Mountain, Shan-Dong Province (1), Shang-Luo, Shan-Xi Province (1), Xin-Jiang Province (1), He-Nan Province (1)
All the samples were taken from different areas even within the same province. The measured results and the standard deviations are listed in Table 4 a-c.
Table 4a: Results of Pb Isotope Ratio Measurements of Dan-shen Plant Samples Sample Name Jiang-Xi Province Shan-Luo, Shan-Xi Province Bei-Jing Shan Dong Province Xin-Jiang Province He-Nan Province ErMei Mountain, D1 ErMei Mountain, D2 Zhong-Jiang, D1 Zhong-Jiang, D2 Zhong-Jiang, D3 Pb / Pb sd(n=3) 0.8485 0.0005 0.8502 0.0014 0.8674 0.0019 0.8570 0.0018 0.8495 0.0007 0.8576 0.0007 0.8467 0.0003 0.8469 0.0004 0.8529 0.0022 0.8524 0.0012 0.8537 0.0003
207 206
Pb / Pb sd(n=3) 2.079 0.002 2.100 0.002 2.159 0.003 2.107 0.002 2.095 0.001 2.115 0.002 2.093 0.001 2.095 0.004 2.105 0.004 2.105 0.002 2.103 0.002
208
206
Table 4b: Results of Pb Isotope Ratio Measurements of Soil Samples

207 206 208 206
Sample Name Jiang-Xi Province ErMei Mountain, T1 ErMei Mountain, T2 Zhong-Jiang, T1 Zhong-Jiang, T2 Zhong-Jiang, T3 Zhong-Jiang, T4 Zhong-Jiang, T5
Pb / Pb sd(n=3) 0.846 3 0.0017 0.8520 0.0016 0.8523 0.0007 0.8395 0.0007 0.8417 0.0004 0.8411 0.0015 0.8400 0.0007 0.8391 0.0014
Pb / Pb sd(n=3) 2.090 0.001 2.101 0.004 2.103 0.001 2.084 0.001 2.089 0.001 2.084 0.002 2.090 0.001 2.083 0.002
Table 4c: Results of Pb Isotope Ratio Measurements of Water Samples

207 206 208 206
Sample Name Zhong-Jiang, w1 Zhong-Jiang, w2 Zhong-Jiang, w3 Zhong-Jiang, w4
Pb / Pb sd(n=3) 0.8702 0.0015 0.8735 0.0005 0.8675 0.0002 0.8693 0.0014
Pb / Pb sd(n=3) 2.140 0.003 2.137 0.002 2.136 0.002 2.138 0.003
Discussion
Since Pb, Pb and Pb are the daughter products from the radioactive decay of 238U and 235U and 232Th respectively, the Pb isotope abundance varies in nature. In most of the samples, the 208Pb/206Pb ratio values are between 1.95 and 2.15, the 207Pb/206Pb ratio values are between 0.78 and 0.86 and the 204Pb/206Pb ratio values are between 0.05 and 0.06[1]. In order to distinguish between the different types of samples by their lead isotope ratio, the measurement precision is extremely important, both over the short and long term. The RSD
206 207 208
values should be at most 0.5% in order to make a clear differentiation between the samples. As indicated in Table 4a-c, the SD values for the determinations are considerably better (lower) than 0.5%, so it is possible to use lead isotope ratio measurements to distinguish between the different samples.
2. 17 2. 16 2. 15 2. 14 208Pb/206Pb 2. 13 2. 12 2. 11 Shan- Xi 2. 1 2. 09 2. 08 2. 07 0. 845 Er - M ei Xin-Jiang Zhong- Ji ang Shan-Dong Bei-Jing
He-Nan
Jiang-Xi
0. 85
0. 855 207Pb/ 206Pb
0. 86
0. 865
0. 87
Figure 1: Pb Isotope Ratio Distribution of the Herb Medicine, Dan-Shen, from Different Sources As a means of looking for variations/similarities in the measurements, the 208Pb/206Pb ratio values were plotted on the y-axis vs the 207Pb/206Pb ratio values on the x-axis. Figure 1 illustrates the data from the analyses of the Danshen plant and shows very interesting results. Obviously, the Pb isotope ratio values show their own special pattern for the Dan-shen samples from different sources. For example, Dan-shen samples grown in the Zhong-Jiang area have similar Pb isotope ratio values, although the samples were taken from different sampling areas. When compared to the Pb isotope values of the Dan-shen samples from the Er-Mei Mountain region (the two samples also have similar Pb isotope ratio values), the difference is distinct. In conclusion, it is possible to distinguish the Dan-shen samples from different sources.
2. 11 Zhong- Ji ang D an- Shen 2. 105
2. 1 208Pb/206Pb Er - M ei Soi l 2. 095
2. 09
Er - M ei D an- Shen
Zhong- Ji ang Soi l 2. 085
2. 08 0. 836
0. 838
0. 84
0. 842
0. 844
0. 846 207Pb/ 206Pb
0. 848
0. 85
0. 852
0. 854
0. 856
Figure 2: Pb Isotope Ratio Distribution of Soil and Dan-Shen Samples taken from Different Sources When the soil samples from different sources are considered, it is found that the isotope values of the soil samples from different sources also have their own special pattern, as shown in Figure2. But, interestingly, the isotope values of the soils are different to the Dan-shen samples from the same place. This suggests that soil isnt the only source of lead within the plant. When the water samples are considered, it is obvious that the Pb isotope ratio values are a combination of those found in the soil and in the water, as shown in Figure 3.
2. 15
2. 14
2. 13
W at er
208Pb/206Pb
2. 12
2. 11 D an- Shen 2. 1
2. 09 Soi l 2. 08
2. 07 0. 835
0. 84
0. 845
0. 85
0. 855
0. 86
0. 865
0. 87
0. 875
0. 88
207Pb/ 206Pb
Figure 3: Pb Isotope Ratio Distribution of the Soil, Dan-Shen and Water Samples Taken from similar Sampling Area (Zhong-Jiang, Si-Chuan Province)
Conclusions
The Agilent PFA micro flow nebulizer is a good choice for isotope ratio analysis since it has very good nebulization efficiency and stability, especially when the sample amount is small. Self-aspiration mode avoids small amounts of pulsation from the peristaltic pump that could affect the precision of the isotope ratio analysis. As this data suggests, the Agilent 7500 Series ICP-MS is well suited for routine isotope ratio analysis. Further improvements in precision may be obtained by modifying the method used in this study slightly. For instance, Tl may be added to the samples as an internal standard for lead isotope ratio analysis[5]. In addition, optimization of the dead-time correction, the stand-by mass selection, may also improve the performance of Agilent 7500 ICP-MS so that the theoretical minimum %RSD of 0.03 can be obtained. Future work at Xiamen University will look in to these potential improvements. This preliminary research indicates that it is possible to distinguish Dan-shen samples from different areas using
Pb isotope ratio measurements. Continuing these studies by running more types of samples will allow a large database to be set-up, and a chemometric model to be built to provide a convenient way to distinguish herb medicines from different sources.
References
1 2 3 4 5 Viczian M, Lasztity A, Barnes RM, J. Anal. At. Spectrom., 1990(5), 293-300 Don Potter, Glenn Carey, Agilent Technologies, Application Note 228-317 Yang Chaoyong, Dissertation for Master of Science, Xiamen University, 2001 Encinar JR, Alonso JIG, Sanz-Medel A, Main S, Turner PJ, J. Anal. At. Spectrom., 2001(16), 322-326 Encinar JR, Alonso JIG, Sanz-Medel A, Main S, Turner PJ, J. Anal. At. Spectrom., 2001(16), 315-321
Information, descriptions and specifications in this publication are subject to change without notice. Visit our website at http:/www.agilent.com/chem/icpms Copyright 2002 Agilent Technologies, Inc. Printed 07/2002 Publication number: 5988-7450
Plant Hormones Rapid Gradient Elution Separation

Application
Agrichemical
Robert D. Ricker
Plant hormones play an essential role in the growth and development of plant species. Rapid analysis of these compounds by HPLC may be used for general research into metabolism and the study of plant disease mechanisms. HPLC may also be used in studies of plant control, whether that is stimulation of growth for the purpose of larger and sturdier plants, or for stopping growth, as in plant pesticides (herbicides). The wide variety and structure of these molecules makes them a candidate for gradient HPLC, to reduce runtime and increase throughput.
1 - Kinetin 2 - N-6-Benzyl Adenine 3 - 3-Indole Acetic Acid 4 - 1-Naphthyl Acetamide 5 - 3-Indole Proprionic Acid 6 - o-Chlorophenoxy Acetic Acid 7 - p-Chlorophenoxy Acetic Acid 8 - 3-Indole Butyric Acid 9 - 1-Naphthyl Acetic Acid 10 - o-Chlorophenoxy Propionic Acid 11 - 3,4,5-trichlorophenoxy Acetic Acid 12 - 3,4,5-Trichlorophenoxy Propionic Acid
Highlights
These sterically protected bonded phases provide long column life and good peak shape at low pH (0.1% TFA)
Use of temperature (60C) with short column length (75mm) permits operation with reasonable back pressure at higher flow rates (e.g. 3.0mL/min).
Agilent ZORBAX 3.5 m particles can be used in columns to provide an excellent balance in speed of separation and resolution.
0.5
1.5
2.5
3.5
Conditions: LC: Hewlett Packard 1050 Column: ZORBAX SB-C8, 4.6 x 75 mm (Agilent Part No. 866953-906) Mobile Phase: A Solvent: Water with 0.1% TFA B Solvent: Acetonitrile with 0.1% TFA Gradient: 22% B to 50% B in 1.5 min. UV: 254 nm; Flow: 3.0 mL / min.; 60C
application
Structural Determination of Ginsenosides Using MSn Analysis

Linda L. Lopez
MSn analysis in an ion trap mass spectrometer permits multiple isolation and fragmentation stages, ensuring that product ions in each stage are specifically related to the precursor ion from that particular stage. This type of stepwise fragmentation can be quite advantageous because it allows product ion origins to be unambiguously assigned, making MS/MS spectra simpler to interpret and permitting individual fragmentation pathways to be followed. This note demonstrates the power of MSn analysis for the structural determination of ginsenosides from a ginseng root extract.
Introduction
Ginseng root, a traditional Chinese herbal remedy, contains more than a dozen biologically active saponins called ginsenosides. This class of natural products is believed to play an important role in the treatment and prevention of a number of diseases including atherosclerosis, arthritis, asthma, diabetes, stroke, multiple sclerosis, and endotoxin liver injury.1 3 Ginsenosides are among a growing class of herbal and vitamin products know as nutraceuticals, that is, food products that have pharmacological benefits to human health because of their therapeutic properties. With an estimated 15 million patients at risk of potentially adverse drug-herb interactions,4 there is renewed interest in the isolation and characterization of these compounds. Ginsenosides are structurally described as glycosides consisting of an aglycone moiety, which is typically a triterpenoid or steroid, and one or more covalently linked sugar monomers. Since most ginsenosides contain multiple oligosaccharide chains at different positions in the molecule, structural elucidation of these compounds can be quite complicated. Tandem mass spectrometric methods have been developed for the characterization of ginsenosides contained in ginseng extracts.5 However, MS/MS experiments carried out on a triple quadrupole mass spectrometer using a collision cell typically generate complex product ion spectra that are often difficult to interpret. This is because first-stage product ions tend to undergo further collisions with the background gas to yield second and third generation fragments that cannot be easily distinguished from first-stage MS/MS product ions.
Experimental
All experiments were done using an Agilent 1100 Series LC/MSD Trap system composed of a binary pump, vacuum degasser, autosampler, and thermostatted column compartment with column-switching valve. The system was operated with the electrospray ionization (ESI) source in the positive ion mode. Reagent grade chemicals and HPLC grade solvents were used in preparing mobile phases and standards.

Figures 1ac show the full scan MS, MS/MS and MS3 spectra from direct infusion of the Rb1 ginsenoside standard, along with proposed origins of the observed product ions. The mass spectrum of Rb1 (Figure 1a) shows predominantly the intact [M+Na]+ pseudomolecular ion at m/z 1131.7, with little or no decomposition of the labile ginsenoside adduct under typical ESI interface conditions using a drying gas temperature of 350C. This is in contrast to previous studies in which a room temperature API interface was required to observe an intact molecular ion,5 and emphasizes the gentle nature of the orthogonal spray ion source on the LC/MSD Trap.
ANALYSIS METHOD:
100
HO OH OH
1131.7
OO OH O O OH OH OH
Direct infusion Flow rate: 5 mL/min MS Conditions Source: Drying gas flow: Nebulizer: Drying gas temperature: Skim 1: Cap exit offset: Averages: ICC: Max accu time: Target: Ion mode: ESI 6 l/min 10 psig 300C 25.0 V 50.0 V 5V On 200 ms 40000 Positive
80
OH
% Relative Abundance
[M + Na] m/z 1131.70 60

HO OH OO
40
HO
HO OH HO OH OO
20
0 200 400 600 800 1000 1200
m/z
Figure 1a. Full scan mass spectrum of ginsenoside Rb1 standard.
MS/MS of the m/z 1131.7 sodium adduct shown in Figure 1b yields a product ion at m/z 789.6 corresponding to cleavage of a single glycosidic bond. For the case of the ginsenoside Rb1, which contains two isomeric oligosaccharide chains, cleavage at either glycosidic linkage would result in isobaric fragment ions at m/z 789.6. Subsequent isolation and fragmentation of m/z 789.6 (Figure 1c) yields two products: (1) a more abundant ion corresponding to loss of the oligosaccharide chain (Glc2Glc) at m/z 365.1 and (2) loss of a deoxyhexose sugar at m/z 627.5. The stepwise fragmentation observed in the ion trap provides specific information on molecular structure; however, in this case it is not possible to establish the initial site of glycosidic bond cleavage in the molecule. For use as a reference in the interpretation of ginsenoside Rb1 fragmentation spectra, several analogs of ginsenoside Rb1 were analyzed by ion trap LC/MSn. In this manner an MS/MS comparative method was devised based on the premise that ginsenoside Rb1 would be expected to retain substructures of the related ginsenosides. For example, Figure 2 shows the
total ion chromatogram generated from online LC/MS3 analysis, as well as extracted ion chromatograms of the pseudomolecular ions of each chromatographic peak generated from a water extract of American ginseng root. Figures 3a and 3b show the online product ion spectra of the ginsenoside Rb2 which contains two different oligosaccharide chains. Isolation and fragmentation of the pseudomolecular ion at m/z 1101.6 yields a single ion at m/z 789.7 resulting from cleavage of the Glc6Ara(p) dissacharide linkage. Comparison of the product ion spectra of ginsenosides Rb1 with Rb2 shows isobaric product ions (m/z 789.7), thus supporting the proposed fragmentation pattern for ginsenoside Rb1 shown in Figure 1. Furthermore, m/z 789.7 was a major fragment ion observed in the product ion spectra of all the ginsenosides analyzed, providing a substructural template that supports the fragmentation patterns proposed for Rb1 and Rb2. Subsequent isolation of the m/z 789.7 precursor generates a full scan MS3 spectrum, which is shown in Figure 3b along with proposed origins of the observed product ions.
ANALYSIS METHOD:
100
HO
OO OH OH OH
789.7
O OH OH OH OH O
MS/MS of 1131.7
Direct infusion Flow rate: 5 mL/min MS Conditions Source: Drying gas flow: Nebulizer: Drying gas temperature: Skim 1: Cap exit offset: Averages: ICC: Max accu time: Target: Ion mode: ESI 6 l/min 10 psig 300C 25.0 V 50.0 V 5V On 200 ms 40000 Positive
80
m/z 789.7 [M + Na] m/z 1131.70

+
60
HO O OH HO HO OH HO O
m/z 365.2
40
OH
20
365.2
0 200 400 600 800 1000
1131.7 1200
m/z
Figure 1b. Product ion spectrum generated from infusion of ginsenoside Rb1 standard by MS/MS analysis of m/z 1131.7 ([M+Na] +).
365.1 100
1131.7
OH
789.7
ANALYSIS METHOD: Direct infusion Flow rate: 5 mL/min
80
Na+ m/z 627.5
m/z 365.1
HO O O
60
HO
OH OH O OH O
40
OH OH
MS Conditions Source: Drying gas flow: Nebulizer: Drying gas temperature: Skim 1: Cap exit offset: Averages: ICC: Max accu time: Target: Ion mode:
ESI 6 l/min 10 psig 300C 25.0 V 50.0 V 5V On 200 ms 40000 Positive
20
627.5
0 200 300 400 500 600 700 800
m/z
Figure 1c. Product ion spectrum generated from infusion of ginsenoside Rb1 standard by MS3 analysis of m/z 789.7 (from m/z 1131.7).

Figure 2. Total ion chromatogram (TIC) and extracted ginsenoside pseudomolecular ion chromatograms generated from LC/MSn analysis of ginseng root extract.
100 80 60 40 20 0 100 80 60 40 20 0
m/z 823
6.29 Rg1
10.36 Rf
m/z 1131
11.13 Rb1
100 80 60 40 20 0 100 80 60 40 20 0 100 80 60 40 20 0
m/z 969 6.67 Re
12.27 Rd 12.75
10.36
m/z 1101
11.37 Rc
11.76 Rb2
TIC
11.13
11.37 6.29 2 4 6 6.67 10.36 8 10 12 12.27 12.75 14
Time (min)
ANALYSIS METHOD:
100
789.7
O HO
O OH O OH OH O
MS/MS of 1101.6
Chromatographic Conditions Column: 2 50 mm Zorbax SB-C18, 3.5 m (p/n 863954-302) A = 0.01% acetic acid in water; B = acetonitrile start with 20% B at 1 min 20% B at 11 min 95% B 0.3 ml/min 10 l ESI 7 l/min 30 psig 300C 25.0 V 50.0 V 5V On 200 ms 40000 Positive
HO OH
m/z 789.7
OH
80
Mobile phase:
+
[M + Na] m/z 1101.60 60

HO O OH HO HO OO OH HO OH O
Gradient:
Flow rate: Injection volume: MS Conditions Source: Drying gas flow: Nebulizer: Drying gas temperature: Skim 1: Cap exit offset: Averages: ICC: Max accu time: Target: Ion mode:
40
20
0 500 600 700 800 900 1000 1100 1200
m/z
Figure 3a. Product ion spectrum generated from ginseng root extract by LC/MSn analysis of m/z 1101.6 ([M+Na] +).
ANALYSIS METHOD:
100
365.2
1101.6
OH
789.7
Chromatographic Conditions Column: 2 50 mm Zorbax SB-C18, 3.5 m (p/n 863954-302) A = 0.01% acetic acid in water; B = acetonitrile start with 20% B at 1 min 20% B at 11 min 95% B 0.3 ml/min 10 l ESI 7 l/min 30 psig 300C 25.0 V 50.0 V 5V On 200 ms 40000 Positive
80 m/z 365.2
HO
Na m/z 627.6
Mobile phase:
Gradient:
O OH OH O O OH O
60
HO
Flow rate: Injection volume: MS Conditions Source: Drying gas flow: Nebulizer: Drying gas temperature: Skim 1: Cap exit offset: Averages: ICC: Max accu time: Target: Ion mode:
40
OH OH
20
627.6
0 200 400 600 800 1000 1200
m/z
Figure 3b. Product ion spectrum generated from ginseng root extract by LC/MSn analysis of m/z 789.7 (from m/z 1101.6).
Structural Determination of Ginsenosides Using MSn Analysis Conclusions

MS analysis using an ion trap mass spectrometer specifically selects the desired precursor ion and dissociates it to produce a specific fragmentation pattern in individual stages. As a result, it is a powerful analytical tool for deducing molecular structure. Electrospray ionization provides a soft ionization technique for generating predominantly intact molecular or pseudomolecular ions with little or no structurally relevant fragment ions in the mass spectra. MS/MS fragmentation in an ion trap mass spectrometer is useful because the product ions generated are derived only from the original molecular ion and are not the result of any additional fragmentation, as can be the case with collision induced dissociation (CID) in a collision cell. Additional MS stages tend to show stepwise fragmentations in which all or most of the ion current is localized in a single product ion, greatly facilitating interpretation of the spectra.
n
References
1. Kim Y. C., Kim S. R., Markelonis G. J., Oh T. H., J. Neurosci. Res. 53, 1998, 42632. 2. Kim H. S., Hong Y. T., Jang C. G., J. Pharm. Pharmacol. 50, 1998, 55560. 3. Yokozawa T.; Liu Z. W.; Dong E., Nephron. 78, 1998, 2016. 4. Smolinske S. C., J. Am. Med. Womens Assoc. 54, 1999, 1912. 5. Wang X., Sakuma, T., Asafu-Adjaye, E., Shiu, G.K., Anal. Chem. 71, 1999, 157984.
Author
Linda L. Lopez is an applications chemist at Agilent Technologies in Palo Alto, CA.
Agilent Technologies shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance or use of this material. Information, descriptions and specifications in this publication are subject to change without notice. Copyright 2000 Agilent Technologies All rights reserved. Reproduction and adaptation is prohibited. Printed in the U.S.A. January 2000 (23) 5968-8869E
Simultaneous analysis of inorganic anions, organic acids, amino acids and carbohydrates using the Agilent Basic Anion Buffer
Application Note
Food Tomoyoshi Soga
Absorbance [mAU] 12.5 10 7.5 5 2 2.5 0 -2.5 -5 10 5 4 1 3 6
1 2 3 4 5 6 7 8
Br Cl NO2 NO3 SO4 Oxalate F Citrate
9 10 11 12 13 14 15 16
HPO4 Malate Asp Glu Acetate Tyr Gly Lactate
17 18 19 20 21 22 23 24
Ala Ser Thr Pro Val Met His Leu+Ile
25 26 27 28 29 30 31 32 33
Phe 34 Lys 35 Trp 36 NGNA 37 NANA 38 Ribose 39 Mannose 40 Xylose 41 Glucosamine 42 31
Glucose Galactosamine Galactose Fucose Sucrose Mannitol Sorbitol Xylitol Inositol
7 19 9 8 11 10 12 13 17
23 21 22 24 20 28 26 25 27
30 32 29 33 35 36 34 37 38
40 39
41 42
15 16 18
14 15 20 25 30 35
Time [min]
Introduction
This application note describes extended applications of the Agilent Basic Anion Buffer. It facilitates the analysis of many anions including inorganic anions, organic acids, amino acids and carbohydrates. The method described here is useful for the screening analysis of anions in food and beverage samples.
To separate these anions simultaneously, a highly alkaline pH condition is used to confer a negative charge not only on inorganic and organic anions but also on amino acids and carbohydrates, and promote their migration towards the anode. Electroosmotic flow (EOF) is reversed in the direction of the anode by adding quaternary ammonium salt to the electrolyte. This is necessary to have both anion migration and
EOF in the same direction (towards the anode) and ensure anion migration past the detector. In this method, indirect UV detection is employed to visualize anions which have little or no chromophore. The Agilent Basic Anion Buffer is pre-made with the pH already adjusted, therefore no further preparation is required. Detailed method parameters and some typical electropherograms are shown.
Necessary supplies
The following parts are necessary for the simultaneous analysis of anions: The following Agilent parts should be ordered separately when used with the Agilent CE system:
Component Fused silica capillary (id = 50 m, l=104 cm L=112.5 cm) CE buffer vials 2 ml (glass*) * For this buffer, only 50-m id straight capillaries are usable. Quantity Part No. 1 pk G1600-64211
Component Agilent Basic Anion Buffer* Product literaure
Quantity Part No. 50 ml 1 5064-8209 5968-7715E
100/pk
5182-9697
CE sample vials, 100 l 1000/pk 9301-0978 (polypropylene) CE caps (polyurethane) 100/pk CE water 500 ml 5181-1512 5062-8578
* It is recommended to use glass vials rather than polypropylene vials.
Procedures
Buffer preparation The Basic Anion Buffer is premade and ready to use. Do not leave the solution uncapped since the buffer is a highly alkaline solution and will readily absorb gaseous carbonate from the atmosphere. This will cause its pH to drop immediately. Make certain the bottle is capped after use. Do not store opened bottles which are a third or less full. These should be discarded because gaseous carbonate in the bottle can be adsorbed and will reduce the pH of the residual buffer. The buffer should be stored at room temperature (not less than 20 C, since some buffer components may crystallize at lower temperatures). Buffer and waste vials Prepare 3 vials (one flushing vial and two home vials). When using 2-ml glass vials (PN 5182-9697), fill each vial with 1.4 ml of the buffer. Also prepare a waste vial (filled with 300-ml CE water or deionized water). Since buffers used for indirect UV detection have limited buffering capacity, the buffer should be replaced every 8 runs when using 2-ml glass vials (PN 5182-9697). For this application, do not use the replenishment system for buffer replacement. The level of dissolved carbonate may increase due to pressurization of the buffer bottle, causing a pH drop. Also, crystallization of the buffer might block the tubes. Standard Preparation Individual stock solutions of inorganic and organic anions should be prepared from their sodium salts or free acids. Since carbohydrates are unstable, they should be prepared shortly before use. For amino acids, individual stock solution of Tyr, Cys-Cys, Asp, Trp, Leu, Ile and Phe should be prepared at a concentration of 10 g/l in 0.1 M NaOH. Other amino acids should be prepared in 0.01 M HCl. The working mixture standard should be prepared by diluting stock solutions with deionized water. If a commercially available amino acid standard mixture in 0.1 M HCl is used, Arg may not be detected since the pKa value of Arg is 10.76. Arg is positively charged in 0.1 M HCl and migrates
toward the opposite direction of the detector. In this case, the standard solution should be diluted with 0.01 M NaOH. The concentration of the standard mixture should be in the range of 50 to 1000 mg/l to obtain good peak shape and sensitivity. Sample Preparation For actual samples, dilution with deionized water is necessary in order to reduce the conductivity of the samples, for example, 1:50 dilution for soy sauce. If the sample contains proteins and the migration times increase from run to run, removal of the proteins is recommended by using centrifugal filtering through a 30-kDa cutoff filter.
Capillary Only 50-m id straight capillaries are suitable for this method. Baseline noise is markedly increased if a 75-m id capillary is used due to the high UV absorptivity of the buffer. Neither bubble cell capillaries nor the High Sensitivity Detection Cell should be used. A 50-m id capillary (L=112.5 cm, l=104 cm) is recommended. Capillary conditioning Avoid capillary conditioning with sodium hydroxide since this degrades the performance of this application. Prior to first use, a new capillary should be flushed only with the run buffer for 15 minutes.
Between analyses it is recommended that the capillary be flushed for 4 minutes with buffer from the flushing vial. Capillary storage If the capillary is removed from the instrument it should be washed for 10 minutes with deionized water and then flushed with air for 10 minutes. When the capillary is to be reinstalled, it is necessary to flush with the run buffer for at least 15 minutes.
Method summary
The following method can be used to separate most inorganic anions, organic acids, amino acids and carbohydrates simultaneously. Below are the general analytical
Capillary Injection Applied voltage: Capillary temperature Detection wavelength Preconditioning
conditions. The method as it should be entered into the Agilent ChemStation is shown on the following page.
Fused silica id = 50 m, l=104 cm, L=112.5 cm (G1600-64211) 1. Pressure: 50 mbar for 6 seconds from sample vial 2. Post-injection of buffer from InHome vial, 50 mbar for 4 s - 30 kV 15 C Signal 350/20 nm, reference 230/10 nm Buffer flush for 4 min at 1 bar prior to each run
Programming the method

HPCE mode: CE Home values: Lift Offset Cassette Temperature Inlet Home Vial Outlet Home Vial Electric: Electric Polarity 4 15.00 C 10 11 Voltage Current
On Negative 30.0 kV 150.0 mA
Negative polarity is used since EOF is reversed. A current limit is not necessary but may be used to prevent excessive current generation in case the wrong vial is used. System Limit 2 A
Place the buffer vials at position 10 and 11. Vial locations are exemplary only.
Power Low Current Limit Store Data: Collect voltage Yes Collect currentYes
Replenishment and Preconditioning: Serial processing Replenishment Entries: No Replenishment used
Do not use replenishment. The level of dissolved carbonate might increase due to pressurization of the buffer bottle causing a pH drop.
Collect power Collect pressure Collect temperature Time entries: Stop time
It is recommended to store the current for every analysis. Current in this method shows from -30 to -40 A. No No Yes
Preconditioning Entries: Function Parameter 1 Flush 4.00 min, I: 9, O:1
Place flushing vial at 9 and waste vial at 1. Remember to monitor the waste vial volume for overflow.
Post time Time Table is empty. Diode array detector Settings: Stop Time Post Time Response Time 1.3 sec
40.00 min Adjust as needed when running actual sample. Off
Postcondition Entries: No Postconditioning used Injection Table Entries: Function Parameter 1 PRESSURE 50.0 mbar, 6.0 sec, I: InjectVial, O:OutHomeVial
May be increased or decreased depending on sample concentration.
Peak width > Prerun Autobalance Postrun Autobalance Spectrum: Store Signals: Store Signal, Bw A: Yes 350/20
as HPCE: 40.00 min Off This is recommended to reduce baseline noise. 0.1 min On Off
2 PRESSURE 50.0 mbar, 4.0 sec, I: InHomeVial, O:OutHomeVial
The post injection plug helps to minimize sample loss upon application of voltage. A voltage ramp is used for the same purpose.
None
Reference Reference, Bw [nm] 230/10

Simultaneous analysis of inorganic anions, organic acids, amino acids and carbohydrates Figure 1 shows a typical electropherogram of a 43-component anion sample including seven inorganic anions, five organic acids, 16 amino acids and 15 carbohydrates using the standard method. If the results are not similar to these please refer to the section Troubleshooting in this note. In this method most inorganic anions and carbohydrates can be separated. However, migration times of several organic acids such as tartarate, succinate, malate and a-ketoglutarate are close. This separation can be improved by using the Organic Acids Analysis Kit from Agilent Technologies (PN 5063-6510). With respect to amino acids, Leu and Ile cannot be resolved. Although Arg is not observed in figure 1, Arg can be detected at approximately 35 minutes if the sample is dissolved in an alkaline solution. Since Tyr and Trp have UV absorbance at 230 nm, they are recorded as negative peaks. Ser migrates just before a system peak. Electrophoretic mobilities at 20 C The effective mobilities of 82 compounds including nine inorganic anions, 23 organic acids, 18 amino acids and 32 carbohydrates were determined by this method at 20 C and are listed in table 1. If the mobilities of the compounds of interest are close, they are difficult to separate. In this case investigating a change of temperature or pH is recommended.
Buffer
1 2 3 4 5 6 7 8 Br Cl NO2 NO3 SO4 Oxalate F Citrate 9 10 11 12 13 14 15 16 HPO4 Malate Asp Glu Acetate Tyr Gly Lactate 17 18 19 20 21 22 23 24 Ala Ser Thr Pro Val Met His Leu+Ile 25 26 27 28 29 30 31 32 33 Phe 34 Lys 35 Trp 36 NGNA 37 NANA 38 Ribose 39 Mannose 40 Xylose 41 Glucosamine 42 31 30 32 28 26 25 27 29 33 35 36 34 37 38 40 39 41 42 Glucose Galactosamine Galactose Fucose Sucrose Mannitol Sorbitol Xylitol Inositol
Sample Capillary Injection Temperature Voltage Detection
Absorbance [mAU] 12.5 10 7.5 5 2 2.5 0 -2.5 -5 10 5 4 1 3 6
7 19 9 8 11 10 12 13 17
23 21 22 24 20
Agilent Basic Anion Buffer (PN 5064-8209) Cl 110 mg/l, carbohydrates 200 mg/l each, others 50 mg/l each fused silica, I=104 cm, L=112.5 cm, id= 50 m 300mbars 15 C -30 kV signal 350/20 nm, reference 230/10 nm
15 16 18
14 15 20 25 30 35
Time [min]
Figure 1 Analysis of 43-component anion standard mixture
Compound
Mobility Compound Mobility Compound Mobility Compound Mobility ( X10-4 cm2/Vs) ( X10-4 cm2/Vs) ( X10-4 cm2/Vs) ( X10-4 cm2/Vs)
-7.181 -6.983 -6.648 -6.442 -6.140 -5.784 -5.409 -5.093 -4.990 -4.911 -4.775 -4.760 -4.677 -4.584 -4.565 -4.520 -4.513 -4.418 -4.196 -4.084 -3.934 Acetate Pyruvate CysCys Glycolate Tyr Gly n-Propionate Lactate Borate Ala Ser n-Butyrate Levulinate Mannuronic Pyroglutamate n-Pentanoate Thr Glucuronic Pro Val Met -3.589 -3.540 -3.514 -3.495 -3.493 -3.260 -3.111 -3.041 -2.963 -2.858 -2.795 -2.781 -2.729 -2.674 -2.631 -2.578 -2.542 -2.497 -2.450 -2.444 -2.389 n-Hexanoate Galacturonic His Leu Ile Phe n-Heptanoate Gluconate Lys Trp NGNA n-Octanoate NANA ManNAc Ribose Fructose GlcNAc Mannose Xylose GalNAc Rhamnose -2.385 -2.337 -2.310 -2.300 -2.300 -2.220 -2.149 -2.060 -2.026 -1.910 -1.719 -1.707 -1.675 -1.221 -1.037 -0.983 -0.975 -0.966 -0.935 -0.919 -0.904 Glucosamine Mannosamine Lactose Arabinose Glucose Maltose Galactosamine Lactulose Galactose Fucose Sucrose Raffinose Mannitol Trehalose Sorbitol Galactitol Xylitol Erythritol Inositol -0.846 -0.832 -0.774 -0.764 -0.761 -0.731 -0.725 -0.676 -0.659 -0.485 -0.291 -0.284 -0.134 -0.121 -0.118 -0.104 -0.086 -0.076 -0.064
Bromide Chloride Nitrite Nitrate Sulfate Oxalate Ascorbate Malonate Fluoride Formate Citrate Pyrophosphate Phosphate Tartarate Succinate Malate a-Ketoglutarate Asp Glutarate Glu Adipate
Table 1 Electrophoretic mobilities of anions at 20 C
Applications
Soy sauce analysis Figure 2 shows the analysis of a soy sauce. The sample was diluted 1:50 with CE water. Centrifugal filtering through a 30 kDa cutoff filter was applied to remove proteins and peptides. A well-defined electropherogram was obtained without interference from other matrix compounds. Satisfactory reproducibilities were obtained for all compounds with RSD values (n=5) for migration times better than 0.3 % and for peak areas between 0.6 and 5.4 %.
Buffer Absorbance [mAU] 35 30 25 20 15 10 5 0 10 15 20 Time [min] 25 30 2 45 7 8 3 9 1 6 14 15 13 16 10 11 12 17 18 19 20 1 2 3 4 5 Cl 6 Citrate 7 Succinate 8 Malate 9 Asp 10 Glu Acetate Gly Lactate Ala 11 12 13 14 15 Ser Thr Pro Val Met 16 17 18 19 20 Leu+Ile Phe Lys Glucose Galactose
Agilent Basic Anion Buffer (PN 5064-8209) Sample soy sauce, 1:50 diluted with water, ultrafiltration with 30-kDa cutoff filter Preconditioning 4 min with run buffer Capillary fused silica, I=104 cm, L=112.5 cm, id= 50 m Injection 300 mbars Temperature 15 C Voltage -30 kV Detection signal 350/20 nm, reference 230/10 nm
Figure 2 Analysis of soy sauce
Pineapple analysis This method was applied to the analysis of organic acids and carbohydrates in pineapple. In the agriculture industry, technical experts are trying to develop a new crossbreed of fruits. Since the content of organic acids and carAbsorbance [mAU] 30 25 20 15 10 5 0 4 6 8 Citrate Malate
bohydrates determines the taste of the pineapple juice, their analysis can help to characterize the product. If citrate and malate concentrations are high, the taste tends to be sour. If the carbohydrate concentration is high, the taste is
sweet. Determination of these compounds is traditionally performed using two HPLC methods. However, the method described here enables the simultaneous analysis of both organic acids and carbohydrates in a much shorter time (less than 18 min) and in a single run. In order to reduce the analysis time, a shorter length capillary was used for this sample. Squeezed pineapple juice was diluted 50-fold with CE water prior to injection. Figure 3 shows a result of the analysis.
Buffer Agilent Basic Anion Buffer (PN 5064-8209) pineapple, 1:50 diluted with water, fused silica, I=72 cm, L=80.5 cm, id= 50 m 300 mbars 20 C -25 kV (reversed polarity) signal 350/20 nm, reference 275/10 nm
Sucrose
Fructose Glucose
Sample Capillary Injection Temperature Voltage Detection
10 12 Time [min]
14
16
Figure 3 Analysis of pineapple
Troubleshooting
Problem Poor resolution or broad/split peaks Possible Cause Buffer is old Capillary damaged Sample overloaded Capillary too short Buffer absorbed carbonate Sample not injected Capillary damaged Wrong buffer used Wrong setting of power supply polarity Detection wavelength incorrect Noisy baseline Wrong setting of response time Capillary window not adjusted Capillary window dirty Lamp is old Buffer pH higher than 12.3 Buffer overused Capillary broken Capillary not filled with buffer Solution Use new buffer Replace capillary Dilute sample Replace capillary Use new buffer Verify no air bubble trapped in bottom of sample vial Verify inlet capillary set correctly Replace capillary Verify buffer Check that polarity is negative Verify signal: 350/20, reference: 230/10 nm Verify response time 1.3 sec DAD Examine capillary window Examine and clean with lint-free paper/MeOH Replace lamp Verify buffer pH Replace buffer Replace capillary Increase flush time
Peak leading No signal
Poor reproducibility Unstable current
References 1. Soga, T. and Ross, G. A. Capillary Electrophoretic Determination of Inorganic and Organic Anions using 2,6-pyridinedicarboxylic Acid: Effect of Electrolyte's Complexing Ability, J. Chromatogr. A, 1997, 767, 223-230, Agilent publication number 5965-8067E. 2. Soga, T. and Heiger, D. N. Simultaneous Determination of Monosaccharides in Glycoproteins by Capillary Electrophoresis, Anal. Biochem., 1998, 261, 73-78, Agilent publication number 5968-0772E. 3. Soga, T. and Ross, G. A. Simultaneous Determination of Inorganic Anions, Organic Acids, Amino Acids and Carbohydrates by Capillary Electrophoresis J. Chromatogr. , 1999, 837, 231-239, Agilent publication number 5968-4470E. Tomoyoshi Soga is application chemist at Yokogawa Analytical Systems, Japan. Agilent Technolgies recognizes his efforts in the development of this work.
For more information on our products and services, visit our website at: http://www.agilent.com/chem
Copyright 1999 Agilent Technologies All Rights Reserved. Reproduction, adaptation or translation without prior written permission is prohibited, except as allowed under the copyright laws. Published in Germany 01/2000 Publication Number 5968-7715E
Analysis of catechins in tea by HPLC with electrochemical detector

Kuniko Koizumi and Hiroki Kumagai Food
Abstract Catechin (C), epicatechin (EC), epigallocatechin (EGC) and epigallocatechingallate (EGCG) collectively called catechins, have many phenolic hydroxyl groups in the molecule. It is known that catechins have various physiological benefits such as anti-oxidation, anti-bacterium, and anti-tumor effects, as well as the suppression of the increase of cholesterol concentration in blood. Catechins also effect the flavor of foods such as tea, wine, and beer. This brief demonstrates the analysis of catechins in tea using the Agilent 1100 Series modules and systems for LC with an electrochemical detector. Analyzed Compounds (+)-catechin, (-)-epicatechin (-)-epigallocatechin (-)-epigallocatechingallate Sample Green tea, tea and woolong tea
Absorbance [mA] 400 300 200 100 0 2 4 6 8 10 12 14 Time [min] EGC C EGCG EC
Conditions
Column 250 x 4.0 mm Hypersil BDS (Agilent Parts No: 79926 OB-584) Mobile phase Methanol: (2g NaNO3 + 0.05 g H2SO4 in 1l Water) = 22:78 Column temp 25 C Injection vol 20 l Detector Agilent 1049 Electrochemical detector Mode Amperometry Working electrode; glassy carbon. Applied Potential:0.850 V Sample preparation Green tea, tea and woolong tea were diluted with water and filtrated.
Figure 1 Chromatogram of standard solution, 2 mg/L each
Method Performance Limit of detection 0.262.00 ng (S/N = 3) RSD of peak area 0.731.14 % RSD of retention time 0.090.14 %
Equipment
Agilent 1100 Series binary pump with vacuum degasser autosampler thermostatted column compartment programmable electro chemical detector diode array detector Agilent ChemStation + software
Absorbance [nA] 120 100 80 60 40 20 0 Absorbance 2.5 [nA] 80 60 40 20 0 Absorbance 2.5 [nA] 140 120 100 80 60 40 20 0 2.5
EGC
Green tea
EGCG EC C 5 7.5 10 12.5 15 17.5 Time [min]
Woolong tea
EGC C 5 7.5 EGCG EC 10 12.5 15 17.5 Time [min]
Tea
Kuniko Koizumi and Hiroki Kumagai are application engineers based at Yokogawa Analytical Systems Inc., Tokyo, Japan.
EGC C 5 7.5
EGCG 10
EC 12.5 15 17.5 Time [min]
Figure 2 Chromatogram of catechins in various teas
For more information on our products and services, visit our worldwide website at http://www.agilent.com/chem
Normal Phase Analysis of Tocopherols in Margarine using HPLC

Abstract Tocopherols cannot be separated completely using reversed-phase chromatography. However, normal-phase chromatography can separate isocratically all eight tocopherols (T) and tocotrienols (T3) naturally occurring in fats, oils, and other foodstuffs. Fluorescence detection is recommended for the analysis of total lipid extraction because UV absorbance detection is not selective enough to prevent detection of coeluting peaks. Chromatographic conditions The HPLC method presented here was used in the analysis of margarine.
mAU 20 15 10 5 0 1 2
-tocopherol
Conditions
Column 100 2.1 mm Hypersil SI 100, 5 m Mobile phase hexane + 2 % isopropanol Stop time 8 min Flow rate 0.3 ml/min Column compartment 25 C Injection vol 0.5 l Detector UV-DAD 295/80 nm Fluorescence excitation wavelength 295 nm, emission wavelength 330 nm Sample preparation 20 g sample dissolved in 15 ml hexane
-tocopherol -tocopherol -tocopherol FLD DAD 3 4 Time [min] 5 6 7
Figure 1 Analysis of tocopherols on normal phase using UV and fluorescence detection
HPLC method performance Limit of detection for diode-array 1020 ng, S/N = 2 Limit of detection for fluorescence 0.52 ng S/N = 2
Equipment
Agilent 1100 Series vacuum degasser isocratic pump autosampler thermostatted column compartment electrochemical detector Agilent ChemStation + software
Repeatability of RT over 10 runs <2 % areas over 10 runs <2 %
%F 90 70
-tocopherol
-tocopherol
-tocopherol 50 30 10 1 2 3 4 Time [min] 11.2 % 77.3 % 9.5 %
-tocopherol Standard
1.9 %
Margarine
Figure 2 Analysis of tocopherol concentration in margarine fat extract with fluorescence detection
Analysis of Selected Anions with HPLC and Electrochemical Detection

Application Note
Food
Angelika Gratzfeld-Huesgen and Rainer Schuster

Figure 1 shows how certain anions can be selectively detected in the presence of several other anions. In total, 16 anions were injected. Four that can be oxidized: iodide, nitrite, bromide, thiocyanide, and twelve that cannot be oxidized: fluoride, chloride, azide, nitrate, phosphite, sulfate, molybdate, hydrogencarbonate, perchlorite, phosphate, selenate and perchlorate.
Traditionally, conductivity is the standard detection method for ion analysis. During the last 10 years, however, there has been increasing interest in electrochemical detection (ECD) for the analysis of anions. This is due to the high sensitivity and selectivity of this type of detection.
Chromatographic conditions
1000 900 800 700 600 500 400 300 200 0 2 4 6 Time [min] 8 10 12 SCN 118 ng 14 Br 77 ng NO2 67 ng I152 ng
Column: Mobile phase:
Flow rate: Inj. volume.: Potential: Mode: Electrode:
200 x 4 mm, Spherisorb ODS2, 5 m water with 5.2 g/l K2HPO4 + 3.6 g/l KH2PO4 + 3 g/l TBAHSO4*/ACN=85:15 1 ml/min 0.1 l 1V amperometry glassy carbon
mV
* Tetrabutylammoniumdihydrogensulphate
Figure 1 Analysis of standard
In order to demonstrate the selectivity and sensitivity, iodide in table salt was analyzed. There is currently a great deal of interest in the determination of iodide in various foodstuffs, because some data indicates that the consumption of iodide has increased, leading to a potential rise in thyroid disorders. Table salt is almost 100% sodium chloride and
180 170 I-
with conductivity detection a large chloride peak may overlap with the iodide. Figure 2 shows the selective detection of iodide in a large excess of cloride. Detection potential is of great importance for selectivity and sensitivity and must be determined for each set of analysis parameters.
The Agilent 1049A ECD in combination with an auto-sampler provides a time-saving way to find the optimum potential for several compounds through its autoincrement mode. The potential is automatically increased within a series of runs and the best detection potential can be selected from the chromatogram plots.
Chromatographic conditions Column: 200 x 4 mm, Spherisorb ODS2, 5 m water with 5.2 g/l K2HPO4 + 3.6 g/l KH2PO4 + 3 g/l TBAHSO4*/ACN=85:15 1 ml/min 0.1 l amperometry glassy carbon
Standard (880 pg) 160 150
Mobile phase:
mV
140 130 120 110 100 0 2 4
I-
Sample (313 pg iodide or 875 g/kg salt)
Flow rate: Inj. volume.: Potential:1 V Mode: Electrode:
* Tetrabutylammoniumdihydrogensulphate
8 Time [min]
10
12
14
Figure 2 Analysis of iodide in table salt, dissolved in water

1.2 V 90 80 70 60 Br 154 pg NO 2 134 pg mV 110.0 I304 pg SCN 236 pg
An example is given in figure 3. The optimum potential was 1.1 V. At lower potentials the response was insufficient, at higher potentials the detector drift was unacceptable. The minimum detectable level for iodide was about 40 g iodide per kilogram table salt. The relative standard deviation for peak heights over 75 runs was 3% for 300 pg of iodide and an injection volume of 10 l.
Angelika Gratzfeld-Huesgen and Rainer Schuster are applications chemists based at Agilent Technologies, Waldbronn, Germany Copyright 1991 Agilent Technologies All Rights Reserved. Reproduction, adaptation or translation without prior written permission is prohibited, except as allowed under the copyright laws. Publication Number 5091-1815E
mV
50 40 30 20 10 0 0
SCN enlarged
108.5 13 14 15 16 1.1 V 1.0 V 0.9 V 20
10 Time [min]
15
Figure 3 Optimizing detection potential using the 'auto-increment' mode
Proteins
Characterization of Transgenic Soybean Seedlines by Protein Expression with the Agilent 2100 Bioanalyzer Application
Food
Authors
Debra Dempsey E.I. DuPont de Nemours DuPont Experimental Station Wilmington, DE 19880 USA Mark Jensen Agilent Technologies 2850 Centerville Road Wilmington, DE 19808-1610 USA
importance. Characterizing the expression of these proteins in various soybean seed lines is also essential in expanding the range of soy protein applications in food. The relative levels of these two proteins have been shown to significantly impact the nutrition, taste, and texture of food products derived from soy protein extracts [1]. For this reason, soybean lines that preferentially express the 11S or 7S proteins continue to be an active target in the efforts to improve seed quality. Both conglycinin and glycinin are complex aggregates made of smaller protein subunits. -conglycinin is a 7S protein with a trimeric structure and is composed of 53, 70, and 76 kDa units. Glycinin is an 11S hexameric protein consisting of six monomer units, where each monomer is made up of 40 and 20 kDa subunits [2]. Given the sizes of protein subunits, it is relatively straightforward to characterize the levels of these two proteins by electrophoresis. The Agilent 2100 Bioanalyzer, an automated microfluidic electrophoresis platform, is well suited for the analysis of proteins in this size range. The Protein 200 LabChip has a size range of 14-200 kDa. Samples of soy protein isolate can be loaded, separated, and analyzed for relative protein composition in less than 45 minutes. In this application we describe the use of the Agilent 2100 Bioanalyzer in the analysis of soybeans, to determine the level of expression of 7S and 11S proteins.
Abstract
This application note describes how the Agilent Technologies 2100 Bioanalyzer can be used to analyze protein extracts from transgenic seedlines. Accuracy and precision in the determination of protein size and concentration was sufficiently good to allow for the characterization of experimental seedlines based solely on expressed protein profiles.
Introduction
-conglycinin(7S) and glycininin(11S) are the primary seed storage proteins in soybean, comprising about 70% of the total storage proteins. Because these proteins make up such a large portion of soya protein, they are of critical economic
Experimental
Extraction Protocol Grind the seed into a fine powder. Place a 30 mg sample in an Eppendorf tube and add 1000 L of extraction buffer (50 mM Tris [pH 7.5], 10 mM 2-mercapto-ethanol, 0.1% SDS). Agitate the mixture on a rotary shaker for 30 minutes and then centrifuge at 15,000 g for 10 minutes. Remove the supernatant and introduce the extract directly into the Protein 200 LabChip to begin the assay protocol. If the extracts contain an excessive amount of oil, the supernatant may be removed and further diluted prior to beginning the Protein 200 protocol.
Methodology To determine if a transgenic line of soybeans preferentially expressed the -conglycinin or glycinin protein groups, seed extracts were compared to extracts made from wild-type seed lines that strongly expressed either the 7S or the 11S group. Twenty extracts from the unknown seedline were prepared as described above. The protein profiles were determined by separating the proteins in the Agilent 2100 Bioanalyzer. Because of the size range of the proteins, the Protein 200 Plus chip (14-200 kDa) was used for the separation. The concentration of the individual proteins was determined by a comparison with an internal concentration marker (myosin). The ratio of 7S/11S proteins was then calculated from those values and the ratio was compared to the ratio of wild-type seedlines that preferentially expressed either 7S or 11S protein groups. Figure 1 shows the electropherogram for a control seedline, a 7S wild-type, an 11S wild type, and a representative sample from the unknown transgenic seedline. A simulated gel image of the same traces is shown in Figure 2.
Glycinin subunits (11S) -Conglycinin subunits (7S) Control
7S
11S
Unknown
20 Time (seconds)
40
Figure 1.
Electropherograms of soya protein extracts.
The ratio of 7S to 11S for the 20 sample extracts was calculated by integrating the individual components comprising the 7S and 11S groups and then determining the summed areas of the two groups. The levels of extracted protein and the 7S/11S ratios for the control 7S, 11S, and unknown groups are summarized in Table 1. The ratios determined for the high 11S and high 7S seedlines indicate the range of expected 7S/11S ratios should fall between 0.04-3.4. The ratios determined for both the controls and unknown extracts both fall within this range. All 20 unknown samples showed a higher 7S/11S ratio than the control. Average ratio values for 20 unknown extracts and 5 control extracts was 0.72 and 0.39, respectively. Measurement precision was excellent for both sample sets.
Table 1.
Summary of Extracted Protein Levels and 7S/11S Protein Ratios Extracted protein level g/mL 14,000 5,200 14,000 13,000 7S/11S Ratio 0.39 0.004 (n=5) 3.4 0.04 0.72 0.1 (n=20)
SeedLine Control 7S 11S Unknowns
Conclusions
This application note describes the use of the Agilent 2100 Bioanalyzer and the Protein 200 LabChip Kit for evaluating the relative expression of -conglycinin and glycininin in unknown seedlines. In the 20 protein extracts taken from the unknown seedline, the average ratio was 0.72 0.1. This ratio lays in range that is characteristic of high 7S expression seedlines. Given the precision of the ratio determination, it is clearly apparent that the assignment of this unknown seedline to the high 7S group is statistically significant. This conclusion is further supported by a comparison to a normal control seedline where the ratio of 7S/11S is 0.39 0.004. The Agilent 2100 Bioanalyzer together with the Protein 200 LabChip Kit are quick and efficient tools for the determination of relative protein expression. The resulting protein expression profiles are in turn a highly effective means for the characterization of new transgenic seedlines.
A B C D
Seedline for protein extraction A Control B 7S cultivar C 11S cultivar D Unknown transgenic
References:
1. Kitamura, K., (1995) "Genetic Improvement of Nutritional and Food Processing Quality in Soybean," Jpn. Agric. Res. Quart., 29(1), 1-8. 2. Yaklich, R. W., (2001) "-Conglycinin and Glycinin in High-Protein Soybean Seeds," J. Agric. Food Chem. 49, 729-735.

Figure 2. Gel simulation of electropherograms for soya protein extracts.
Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Eppendorf is a registered trademark of Eppendorf AG. Agilent Technologies, Inc. 2003 Printed in the USA May 9, 2003 5988-9441EN
Effect of Temperature on Separation of a Wheat Protein

Application
Food Analysis
Robert Ricker
Highlights
Wide-pore ZORBAX 300 SB-C8 demonstrates high peak capacity for complex samples of plant protein.
Temperature can be effectively used to increase resolution in reversedphase separations of complex samples.
Conditions: ZORBAX 300 SB-C8, 4.6 x 150 mm (Agilent P/N: 883995-906) 1 ml/min., 15 l of Neepawa Wheat Extract Abs. 210 nm 23-48% B in 120 min. A=0.1% TFA/water, B=0.1% TFA/ACN
Stationary-Phase Selectivity Comparison:

High-Molecular-Weight, Alkylated, Wheat Glutenin, Protein Subunits Application
Food Analysis
Robert Ricker
Highlights
Improved separations are achieved by taking advantage of different bondedphase chemistries.
Sterically protected SB bonded phases are hydrolytically stable, especially at low pH, and can be relied upon for stable, reproducible separations.
Conditions: ZORBAX 300SB-C8, 300SB-CN (4.6 x 150 mm) Agilent P/N: (883995-906) (883995-905) Gradient: 60 min., Mobile Phase (as above) with 0.1% TFA, 1.0 mL/min., 50C
Vitamins
Quantitative Analysis of Water-Soluble B-Vitamins in Cereal Using Rapid Resolution LC/MS/MS Application
Food Analysis
Authors
Sheher Mohsin Agilent Technologies, Inc. Schaumburg, Il USA Michael Zumwalt Agilent Technologies, Inc. 9780 S. Meridian Blvd. Englewood, CO 80112-5910 USA Indarpal Singh ConAgra Foods, Inc. Omaha, NE 68102 USA
ence of ion pair reagents such as heptafluorobutyric acid in the mobile phase has been shown to improve the separation and retention of these compounds. However, the drawback of using such ion pair reagents is the high background levels that are generated inside the mass spectrometer. Therefore, we have developed a rapid and sensitive method with ammonium formate in the mobile phase solvent using a column with a bonded phase designed to retain hydrophilic compounds. The Agilent 1200 Series liquid chromatography (LC) system used in this work was designed to take advantage of sub-2-micron particle columns for rapid, high-resolution separations. Included in the LC design were decreased delay volume, increased pressure range, and increased column temperature. This LC system was coupled to the Agilent 6410 Triple Quadrupole Mass Spectrometer (QQQ) by way of the G1948B electrospray ionization source. Target compound separation was achieved on a ZORBAX AQ 1.8-micron column using a water and methanol gradient with ammonium formate. Typical LC/MS methods for water-soluble vitamins have shown analysis times as high as 30 minutes with heptafluorabutyric acid ion pairing reagent in the mobile solvent. We have developed a rapid and sensitive method for the LC/MS/MS analysis of water-soluble vitamins by employing a high-efficiency 1.8-micron column in a low-dispersion, 600-bar LC/MS configuration that allowed screening and quantitation with a run-to-run cycle time as low as 10 minutes. Linearity of the mass spec-
Abstract
An Agilent 6410 Triple Quadrupole Mass Spectrometer (QQQ) is used to analyze several water-soluble B-vitamin compounds in breakfast cereal. A simple gradient elution is carried out on a Rapid Resolution High Throughput SB-Aq column (particle size 1.8 m). All compounds elute in less than 7.5 minutes, and with the exception of pyridoxine, good linearity over more than three orders of magnitude, from 0.5 to 500 ppb, is demonstrated with good peak area reproducibility at the 0.5 ppb level, which is the lowest level of quantitation considered.
Introduction
Water-soluble vitamins are very polar and have poor retention on reverse-phase columns. The pres-
trometer response was observed over three orders of magnitude with limits of quantitation in the 0.5 pg/L range for all of the analytes except for pyridoxine. In the case of pyridoxine, good sensitivity was demonstrated, but good linearity was limited to just under three orders of magnitude. Calibration curves and chromatograms for the vitamins between 0.5 and 500 pg/L were generated
O O H 3C H 3C H 2N O H 3C H 3C HN O CH 3 HO NH 2
for all compounds with the exception of pyridoxine, which was between 0.5 and 250 pg/L. The structures of the B vitamins are shown below.
NH 2 O H 3C CH 3 N+ N+
Cyanocobalamin C 63H88CoN14O14P
O NH 2 N
O OH
Nicotinic acid C 6H 5NO2
H 2N
O N+ N+ Co 4
_
N CH 3 CH 3 H 2N N+ N CH 3 O CH 3 H 2N N O N
NH 2
Nicotinamide C 6 H 6 N 2O
H N
N N
HN
Folic acid C19H19N7O6

O
NH OH OH
O O P HO O
OH H N _ O HO N HO N N OH OH
H 3C N
NH 2 N+
CI S
Pyridoxine C 8H11NO3
OH
CH 3
H 3C
CH 3
Riboflavin C17H20N4O6
Pantothenic acid C9H17NO5

OH HO O NH O H 3C CH 3 OH
H 3C HO
Thiamine C12H17CIN4OS
OH
OH
Figure 1.
Structures of B vitamins analyzed in this work.
Experimental
Sample Preparation A standard mix containing all eight compounds in methanol was provided by ConAgra Foods and diluted in 90% water/10% methanol with 20 mM ammonium formate and 0.1% formic acid to the following concentrations: 500, 250, 100, 50, 5, and 0.5 pg/L. These dilutions were used for the quantitation of unknown samples. One B-vitamin-fortified sample was also provided using the following sample preparation procedure: 1. Grind and homogenize breakfast cereal in blender 2. Weigh 1 gram of homogenized sample into a 50-mL vial 3. Add 25 mL of 0.1M HCl and heat in water bath at 100 C for 20 min. This solubilizes the vitamins.
4. Cool to ambient temperature 5. Adjust volume to 1 L with deionized water 6. Filter with 0.45-m glass microfiber membrane. It should be noted that the provided fortified sample was created for testing the sensitivity of the instrument for customer demonstration purposes. A typical unfortified sample extraction consists of 1 g homogenized sample treated with enzymatic digestion, to release naturally occurring vitamins from their conjugated forms, and volume adjusted to 10 mL, which is 1/100th the volume used in the fortified sample analyzed in this work. At the higher concentration, salts and other matrix contributions are seen to cause interference in the analysis of some vitamins. As a result, further dilution may be used to accommodate the matrix effect in these samples.
Table 1. Segment 1
MRM Mode Parameters Compound Thiamine Pantothenic acid Pyridoxine Nicotinic acid Nicotinamide Cyanocobalamin Folic acid Riboflavin Transition 265.2 > 122.0 220.2 > 90.0 170.1 > 152.1 124.1 > 80.0 123.1 > 80.0 678.6 > 146.7 442.2 > 295.1 377.2 > 243.1 Fragmentor (V) 85 110 100 100 100 130 120 110 Collision Energy (V) 10 13 10 27 25 35 10 25
LC/MS Method Details

LC Conditions Agilent 1100 Series binary pump, degasser, wellplate sampler, and thermostatted column compartment Column: Column temperature: Mobile phase: Flow rate: Injection volume: Gradient: Agilent ZORBAX RRHT SB-Aq, 3.0 mm 100 mm, 1.8 m (PN: 828975-314) 35 C A = 20 mM ammonium formate and 0.1% formic acid in water B = 20 mM ammonium formate and 0.1% formic acid in methanol 0.5 mL/min 10 L Time (min) %B 0.0 10 8.0 55 Stop time: 10 min 8.1 10 75:25 methanol/water (flush port 20 seconds) Positive ESI using the Agilent G1948B ionization source 30 psig 10 L/min 350 C 1850 V Q1 = low res; Q2 = low res 200 msec 3
Needle wash: MS Conditions Mode: Nebulizer: Drying gas flow: Drying gas temperture: Vcap: Resolution (FWHM): Dwell time for all MRM transitions
The precursor ion mass for cyanocobalomin (m/z 678.6) is about half of the expected value in which the empirical formula for this compound is C63H88CoN14O14P, as denoted in Figure 1. It is believed that the structure of this compound is not very stable and breaks apart during the ionization process.

The calibration curves for all eight compounds are shown in Figures 2A through 2H. Only for the compound pyridoxine is the 500 ppb level needed to be removed for good linearity. No internal standard is included.
Thiamine 0.5 500 ppb (5 5000 pg on-column) R2 > 0.99
7 replicate injections
Figure 2A. Linearity of thiamine over three orders of magnitude.
Pantothenic acid 0.5 500 ppb (5 5000 pg on-column) R2 > 0.99
Figure 2B. Linearity of pantothenic acid over three orders of magnitude.
Pyridoxine 0.5 250 ppb (5 2500 pg on-column) R2 > 0.99
Figure 2C. Linearity of pyridoxine acid over nearly three orders of magnitude. 5
Nicotinic acid 0.5 500 ppb (5 5000 pg on-column) R2 > 0.99
Figure 2D. Linearity of nicotinic acid over three orders of magnitude.
Nicotinamide 0.5 500 ppb (5 5000 pg on-column) R2 > 0.99
Figure 2E. Linearity of nicotinamide acid over three orders of magnitude. 6
Cyanocobalamin 0.5 500 ppb (5 5000 pg on-column) R2 > 0.99
Figure 2F. Linearity of cyanocobalamin over three orders of magnitude.
Folic acid 0.5 500 ppb (5 5000 pg on-column) R2 > 0.99
Figure 2G. Linearity of folic acid over nearly three orders of magnitude. 7
Riboflavin 0.5 500 ppb (5 5000 pg on-column) R2 > 0.99
Figure 2H. Linearity of riboflavin acid over three orders of magnitude.
All line fits to data are carried out as linear with the origin ignored and a 1/x weighting. An example of reproducibility at the 0.5 ppb level for pyridoxine is shown in Figure 3. The peak area %RSD values for all compounds at the 0.5 ppb level are given in Table 2.
Table 2.
Peak Area Reproducibility for Each Compound at the Lowest Level of Quantitation, Which Is 0.5 ppb Peak area %RSD 5.6 11.6 2.6 12.9 10.1 19.4 13.4 17.1
500 ppt Pyridoxine Area %RSD = 2.6
Compound Thiamine Pantothenic acid Pyridoxine Nicotinic acid Nicotinamide Cyanocobalamin Folic acid Riboflavin
Figure 3.
Peak area reproducibility of pyridoxine at 500 ppt level, 6 injections.
A fortified cereal extract is also analyzed and quantitated using the diluted standard mix already mentioned. An example of the batch results using the MassHunter Quantitative Analysis is shown in Figure 4. The concentration of nicotinamide present in the sample is calculated to be 111.6 pg/L. The concentrations calculated for all compounds in the fortified extract are given in Table 3. The corresponding chromatographic elution of the eight compounds detected in the fortified extract is shown in Figure 5.
Table 3.
Calculated Concentrations for Each Compound in Fortified Cereal Extract Calculated concentration (pg/L) 24.0 1.2 15.5 43.5 111.6 0.4* 2.6 8.6
Compound Thiamine Pantothenic acid Pyridoxine Nicotinic acid Nicotinamide Cyanocobalamin Folic acid Riboflavin
* Outside quantitation limits.
Figure 4.
Quantitation batch results for nicotinamide in sample. Concentration calculated to be 111.6 pg/L (highlighted). 9
Thiamine (24.0 pg/L)
Pantothenic acid (1.2 pg/L)
Pyridoxine (15.5 pg/L)
Nicotinic acid (43.5 pg/L)
Nicotinamide (111.6 pg/L)
Cyanocobalamin (0.4 pg/L)
Folic acid (2.6 pg/L)
Riboflavin (8.6 pg/L)
Figure 5.
Chromatogram of compounds in fortified extract and calculated concentrations.
10
Conclusions
The water-soluble B vitamins are successfully analyzed using LC/MS/MS. Good linearity with R2 > 0.99 is demonstrated over three orders of magnitude for all compounds, with reproducibility as low as 2.5 %RSD at the lowest level of quantitation for pyridoxine. An extracted fortified sample is successfully analyzed with only the cyanocobalamin concentration falling below the quantitation limit.

For more information on our products and services, visit our Web site at www.agilent.com/chem. For more details concerning this application, please contact Michael Zumwalt at Agilent Technologies, Inc.
11
Direct Analysis of Folic Acid in Digestive Juices by LC/TOF-MS Application
Food Safety
Authors
M. H. Blokland, P. W. Zoontjes, S. S. Sterk, and L. A. van Ginkel, National Institute of Public Health and the Environment (RIVM) Laboratory for Food and Residue Analysis Bilthoven The Netherlands
consumption of enriched food products. In the latter case, it is important to know the fraction of the total intake that is accessible in the gastric tract. To study the accessibility, an in vitro digestion model can be used. Due to the complex nature of gastrointestinal juices (for example, saliva, stomach juice, and chyme), analysis of folic acid in these juices generally requires an extensive cleanup procedure. Traditionally, cleanup procedures consist of immuno affinity chromatography or solid phase extraction. Using the high-resolution capabilities and unique accurate mass measurement of LC/TOF-MS, these cleanup steps can be eliminated and direct detection of folic acid in gastrointestinal juices is possible. By doing so, the risk of degradation of folic acid during cleanup is minimized and highthroughput analysis becomes possible. This application describes the direct measurement of folic acid in gastrointestinal juices.
Jerry Zweigenbaum Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE USA
Abstract
By use of the high resolving power and accurate mass measurement of the LC/TOF-MS it is possible to measure folic acid without cleanup directly in gastrointestinal juices. The samples are obtained from an in vitro digestion model. The method is validated for measuring uptake in this type of model.
Experimental
Cleanup Procedure Saliva, stomach juice, and chyme were prepared as described elsewhere.[1] Sample preparation is straightforward; an aliquot of the juices is taken and directly injected in the LC/TOF-MS system.
Introduction
Health benefits associated with the intake of folic acid are well known. Increasing the daily intake can be achieved through medication or through the
LC/MS Method Details

LC Conditions Instrument: Column: Column temp: Mobile phase: Agilent LC 1200 SL ZORBAX SB-C8, 100 mm x 2.1 mm, 1.8 m (p/n 828700-906) 80 C A: 1% formic acid in water B: 26/60/14 v/v acetonitrile/water/methanol 5% B at 0 min 5% B at 4 min 40% B at 8 min 40% B at 8.1 min 5% B at 9 min 0.5 mL/min 10 L Agilent 6210 LC/TOF-MS Positive ESI 9 L/min 60 psig 350 C 4000 V 150 V 60 V m/z 110%940, 12419 transients/scan, 1.5 scan/sec 100 mg/L solutions of purine 2 ml/L and HP-921 1 ml/L in acetonitrile infused into second sprayer at constant rate m/z 121.020873 and 922.009798

Calibration Standards The structure of folic acid is shown in Figure 1. Because sensitivity is not a concern in this analysis conditions were only optimized for accurate mass measurement and linearity of response within the concentrations expected for the samples. In Figure 2 a typical chromatogram is shown of a standard of 1 g/mL (expressed as concentration in matrix). The S/N for this concentration is typically greater then 500. Calibration curves are produced from seven calibration points in the range between 1 g/mL to 100 g/mL. It is demonstrated that the LC/TOF-MS has the necessary linear dynamic range.
Gradient:
Flow rate: Injection volume: MS Conditions Instrument: Source: Drying gas flow: Nebulizer: Drying gas temp: Vcap: Fragmentor: Skimmer: Scan mode: Reference mass solution: Reference mass:
O N H2N N H N N O NH NH O
Molecular Formula: Monoisotopic Mass: C19H19N7O6 441.139683 Da
Figure 1.
HO HO Structure of folic acid, its exact molecular weight, and its empirical formula.
2.6e4 2.4e4 2.2e4 2.0e4 1.8e4 1.6e4 Intensity, cps 1.4e4 1.2e4 1.0e4 8000.0 6000.0 4000.0 2000.0 0.0 XIC of +TOF MS: 442.142 to 442.151 amu from 06072001.wiff
6.47
4.0
4.2
4.4
4.6
4.8
5.0
5.2
5.4
5.6
5.8
6.0 6.2 Time, min
6.4
6.6
6.8
7.0
7.2
7.4
7.6
7.8
8.0
Figure 2.
Extracted ion chromatogram of a standard of folic acid at 1 3g/mL (ion extraction window of 10 ppm of exact mass of M+H ion).
Response 25000000
Calibration curve folic acid
20000000
y = 415215x - 69578 R2 = 0.9924

15000000
10000000
5000000
0 0 10 20 30 40 Concentration folic acid (g/mL) 50 60
Figure 3.
Calibration curve of folic acid standards from 1 to 100 3g/mL.
Validation of the Method To obtain reliable validation results it is important to obtain clean chromatograms without matrix interference. Figures 4 to 6 show examples of total ion chromatograms, extracted ion chromatograms (EIC), and spectra of spiked gastrointestinal juices (1 g/mL). In all samples mass accuracy was better then 3 ppm. EIC of each were extracted with a mass window of 442.1425 to 442.1515 ( 10 ppm) to obtain chromatograms with no matrix interference.
2.8e6 2.4e6 Intensity, cps 2.0e6 1.6e6 1.2e6 8.0e5 4.0e5 0.0 4.0 2.4e4 2.0e4 Intensity, cps 1.6e4 1.2e4 8000.0 4000.0 0.0 4.0 4.2 4.4 4.6 4.8 5.0 5.2 5.4 5.6 5.8 6.0 6.2 Time, min 6.4 4.2 4.4 4.6 4.8 5.0 5.2 5.4 5.6 5.8 6.0 6.2 Time, min 6.4
TIC of +TOF MS: from 2006062111.wiff
Validation of the Sample Data Validation of the method was done by spiking saliva, gastric juice, and chyme at the following concentrations: 1, 2.5, 5, 10, 25, 50, 75, and 100 g/ml. The results would indicate whether there were ion-suppression effects in these matrices. Figure 7 shows the response for each of these matrices. The data show that chyme produces the least effects, whereas saliva and gastric juice produce slightly higher ion suppression effects.
4.88
5.05
6.08
6.38
6.64
7.45
7.83
Max. 5.8e6 cps
6.6 6.63
6.8
7.0
7.2
7.4
7.6
7.8
8.0
XIC of +TOF MS: 442.142 to 442.151 amu from 2006062111.wiff
6.6
6.8
7.0
7.2
7.4
7.6
7.8
8.0
+TOF MS: 6.615 to 6.661 min from 2006062111.wiff Agilent, subtracted (6.764 to 6.902 ...
442.1472
2.4e4 Intensity, counts 2.0e4 1.6e4 1.2e4 8000.0 4000.0 0.0
295.0950
C
148.0561 130.0502
120 140 160 180 200 220 240 260 280 300 320 m/z, amu 340 360 380 400 420 440 460 480 500
221.5781
Max. 2.5e4 counts
Figure 4.
TIC (A), EIC (B), and spectra (C) for folic acid in saliva spiked at 1g/mL.
3.2e7 2.8e7 2.4e7 2.0e7 1.6e7 1.2e7 8.0e6 4.0e6 0.0 4.0 4.0e4 3.6e4 3.2e4 2.8e4 2.4e4 2.0e4 1.6e4 1.2e4 8000.0 4000.0 0.0 4.0
Max. 3.6e7 cps.

7.35
7.64
8.00 7.87
Intensity, cps
A
6.05 4.26 5.07
6.26 6.52
6.81
7.03
4.2
4.4
4.6
4.8
5.0
5.2
5.4
5.6
5.8 6.0 6.2 Time, min
6.4
6.6
6.8 6.61
7.0
7.2
7.4
7.6
7.8
8.0
Intensity, cps
4.2
4.4
4.6
4.8
5.0
5.2
5.4
5.6
5.8 6.0 6.2 Time, min
6.4
6.6
6.8
7.0
7.2
7.4
7.6
7.8
8.0
442.1473
1.8e4 +TOF MS: 6.594 to 6.639 min from 2006062129.wiff Agilent, subtracted (6.697 to 6.823 ... 1.6e4 1.4e4 295.0950 1.2e4 245.1864 1.0e4 8000.0 148.0554 187.1446 221.5782 336.1990 445.9498 6000.0 261.6512 366.2044 333.8643 4000.0 130.0451 409.2084 2000.0 0.0
Max. 1.9e4 counts.

540.2673
Intensity, counts
470.2495 475.4894 509.0044 450 500 550 594.2597
150
200
250
300
350 m/z, amu
400
600
Figure 5.
1.4e7 1.2e7 Intensity, cps 1.0e7 8.0e6 6.0e6 4.0e6 2.0e6 0.0 4.0
TIC (A), EIC (B), and spectra (C) for folic acid in gastric juice spiked at 1 g/mL.
Max. 5.8e7 cps.
7.77 6.98 7.31 7.97
A
6.07
4.2
4.4
4.6
4.8
5.0
5.2
5.4
5.6
5.8 6.0 6.2 Time, min
6.4
6.6 6.54
6.8
7.0
7.2
7.4
7.6
7.8
8.0
Intensity, cps
7.0e4 6.0e4 5.0e4 4.0e4 3.0e4 2.0e4 1.0e4 0.0 4.0
B
7.78
4.2
4.4
4.6
4.8
5.0
5.2
5.4
5.6
5.8 6.0 6.2 Time, min
6.4
6.6
6.8
7.0
7.2
7.4
7.6
7.8
8.0
429.2354
Max. 6.6e4 counts.
6.0e4 Intensity, counts 5.0e4 4.0e4 3.0e4 2.0e4 1.0e4 0.0
+TOF MS: 6.525 to 6.560 min from 2006062156.wiff Agilent, subtracted (6.617 to 6.732 ...
C
295.0950 148.0553 221.5787 290.1621 375.2249 417.2346 413.6461
442.1474 551.1909 531.2793
120 140 160 180 200 220 240 260 280 300 320 340 360 380 400 420 440 460 480 500 520 540 560 m/z, amu
Figure 6.
TIC (A), EIC (B), and spectra (C) for folic acid in chyme spiked at 1g/mL. 5
Response in different gastrointestinal juices

35000000
30000000
Saliva Gastric juice Chyme
25000000
20000000 Response 15000000 10000000 5000000
0 0 20 40 60 Concentration (g/mL) 80 100 120
Figure 7.
Calibration curve of folic acid spiked in different juices.
To determine the accuracy and covariance of the measurement, gastric juice and chyme were spiked (n = 4) at 50 g/mL and analyzed on four separate days, with a one-week interval. These results are shown in Table 1. The Limit of Detection (LOD) was determined by preparing a standard curve in saliva, gastric juice, and bile. From this calibration curve the Y-inter-
cept and slope were calculated. The LOD is the corresponding concentration at the Y-intercept plus three times the standard deviation of the Y-intercept. The results of these determinations are given in Table 2.
Table 1.
Accuracy (%) and Covariance (%) Week 1 Accuracy (%) Week 2 Accuracy (%) 99.3 102.5 Week 3 Accuracy (%) 82.9 91.4 Week 4 Accuracy (%) 91.7 93.3
CV (%) 1.2 2.1
CV (%) 4.2 3.7
CV (%) 6.0 4.2
Gastric Chyme
91.9 99.9
Table 2. LOD
Determination of the Limit of Detection (LOD) Saliva (g/mL) 0.4 Stomach (g/mL) 0.7 Chyme (g/mL) 0.6
Conclusions
The aim of this work was to develop a method with a minimal cleanup to detect folic acid in gastrointestinal juices. By using the high resolving power of the Agilent LC-TOF, it was possible to measure directly in these juices without any cleanup. The measurements proved to be reproducible with a high accuracy in complex matrices. The response was linear for the different digestive fluids tested with little matrix effects, and the method was validated within the concentrations required for folic acid uptake studies.
Reference
1. C. H. Versantvoort, A. G. Oomen, E. Van de Kamp, C. J. Rompelberg, and A. J. Sips. Applicability of an in vitro digestion model in assessing the bioaccessibility of mycotoxins from food. Food Chem Toxicol, 2005 Jan, 43(1).

The information contained in this publication is intended for research use only and is not to be followed as a diagnostic procedure. Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies, Inc. 2006 Printed in the USA December 27, 2006 5989-5869EN
Improved and Simplified Liquid Chromatography/Atmospheric Pressure Chemical Ionization Mass Spectrometry Method for the Analysis of Underivatized Free Amino Acids in Various Foods Application
Food
Authors
Sreyya zcan Food Engineering Department Hacettepe University 06532 Beytepe, Ankara Turkey Hamide Z. Senyuva Ankara Test and Analysis Laboratory Scientific and Technical Research Council of Turkey Ankara 06330 Turkey
threonine considered essential, which means that humans need to get a certain amount of them in their diet to function properly. The determination of amino acids is of great importance to the food industry because of nutritional labeling requirements. There are several approaches to amino acids analysis based on the pre- or post-column derivatization techniques, which have demonstrated good results but have drawbacks of very long analysis time or poor repeatability. Most of the other methods used for amino acid analysis involve a chromatographic separation following or preceding the derivatization of amino acids with UV absorptive or fluorescent functional group detection. Although these methods are often faster than the classical ionexchange method, which involves quantification after ninhydrin and OPA derivatization and offers lower detection limits, these methods are still very time consuming, taking 20 to 75 min, are not repeatable, and are difficult to undertake. The objectives of the present study were 1) to develop an analytical method for the simple, rapid, accurate, and repeatable determination of 22 free amino acids in various foods using an improved LC/APCI-MS method, 2) to develop an easy aqueous extraction of amino acids without clean-up, and 3) to perform screening and both qualitative and quantitative analysis on a narrow-bore column with a short run time.
Abstract
An improved analytical method that offers rapid, accurate determination and identification of 22 amino acids in a variety of matrixes is reported. The amino acids were extracted from the matrixes using acidified water. Simultaneous determination of 22 underivatized amino acids was carried out by liquid chromatographymass spectrometry (LC/MS). A narrow-bore column allowed rapid screening and quantitative analysis by LC/atmospheric pressure chemical ionization (APCI) MS in positive ion mode.
Introduction
There are a total of 22 amino acids classified by their functional group, or their R group. Of those 22 amino acids, eight are histidine, lysine, phenylalanine, methionine, leucine, isoleucine, valine, and
Method
LC Conditions Column: Flow rate: Mobile phase: Injection: MS Conditions Ionization mode: Nebulizer pressure: Drying gas flow: Drying gas temperature: Vaporizer temperature: Skimmer: Capillary voltage: Fragmentor voltage: Dwell time: ZORBAX Bonus-RP, narrow-bore (100 mm 2.1 mm, 3.5 m) 0.2 mL/min 0.01 mM acetic acid in 0.2% aqueous solution of formic acid 20 L out of 1000 L Positive APCI 55 psi 4 L/min 320 C 425 C 20 V 3 kV 55 V 27 ms
mixing in a vortex mixer for 2 min, the mixture was centrifuged at 5,000 rpm for 10 min at 5 C. The clear supernatant was quantitatively transferred into a vial, avoiding the top oil layer if present. The supernatant was filtered through 0.45-m nylon syringe filter prior to LC/MS analysis.
Table 1. Characteristic Fragments of 22 Amino Acids and Ions Used in SIM Mode for Quantification Fragment ions m/z 90, 73 175, 129 133, 116, 87, 74 134, 116, 88 122, 105, 87, 73 241, 122 148, 130, 102 147, 130, 101 76, 59 156, 110, 96, 73, 59 132, 86 132, 86 147, 130, 84 150, 133, 104 166, 149, 120 116, 70 106, 88, 60 120, 102, 74 182, 165, 136, 123 205, 188, 130 118, 72 Selected ion m/z 90 175 133 134 122 241 148 147 76 156 132 132 147 150 166 116 106 120 182 188 118
ID Alanine (ala) Arginine (arg) Asparagine (asn) Aspartic acid (asp) Cysteine (cys) Cystine(cys-cys) Glutamic acid (glu) Glutamine (gln) Glycine (gly) Histidine (his) Hydroxyproline (hyp) Leucine-isoleucine (leu-ile) Lysine (lys) Methionine (met) Phenylalanine (phe) Proline (pro) Serine (ser) Threonine (thr) Tyrosine (tyr) Tryptophan (trp) Valine (val)
Experimental
LC/MS experiments were performed using an Agilent 1100 Series HPLC system consisting of a binary pump, an autosampler, and a temperaturecontrolled column oven, coupled to an Agilent 1100 MS detector equipped with an atmospheric pressure chemical ionization (APCI) interface. Data acquisition was performed in SIM mode using the interface parameters: drying gas (N2, 100 psig) flow of 4 L/min, nebulizer pressure of 55 psig, drying gas temperature of 320 C, vaporizer temperature of 425 C, capillary voltage of 3 kV, corona current of 8 A, fragmentor voltage of 55 V, and dwell time of 27 msec. Ions monitored for 22 underivatized amino acids are given in Table1. The chromatographic separations were performed on a ZORBAX Bonus-RP, narrow-bore (100 mm 2.1 mm, 3.5 m) using the isocratic mixture of 0.01 mM acetic acid in a 0.2% aqueous solution of formic acid at a flow rate of 0.2 mL/min at 400 C. Sample preparation According to the sample matrix, the sample was ground (by using a blender, mesh size 2 mm) or mixed (by Ultra Turrax). Dry samples like baby foods, breakfast cereal, and cookies were milled, and soft samples, like tomato, and liquid samples, like juices, were mixed with an Ultra Turrax blender. The pH of each homogenized sample was measured before sample preparation. Subsamples of the homogenate were stored at 20 C in high-density polyethylene bottles with plastic screw-cap lids. Finely ground or homogenized sample (1 g) was weighed (fresh weight [FW]) into a 10 mL glass centrifuge tube with cap. After
2

Optimization of MS fragmentation and chromatographic conditions Fragmentation patterns were reproducible in both modes, and quantification could be performed on single m/z fragments. In-source collision induced dissociation at 55, 70, and 100 V was applied for all amino acids, and optimum conditions were found to be 55 V and positive polarity. All the most abundant fragmentations were present, with optimization of the MS parameters. Under the positive APCI conditions applied here, SIM-MS analysis showed excellent identification and quantification conditions within the 7.5 min total run time. Specificity MS detection has the advantage of providing structural information about the eluted compounds. Moreover, resolution of co-eluting compounds can be achieved by selecting different m/z ions for monitoring. In this study, a simple, rapid, and economic sample preparation method was used.
During the extraction with a 0.2 mM acetic acid solution, colloids soluble in water (starch) were simultaneously precipitated by centrifugation. Fat was separated by cold centrifugation, if present in the food sample. Figure 1 shows the MS signals for specific m/z ions of 22 amino acids in green peas using a single injection.
35000 30000 25000 20000 15000 10000 5000 0 8000 1600 1400 1200 1000 800 600 400 200 5 80000 min 12000 10000 8000 6000 4000 2000 0 min 2000 1750 1500 1250 1000 750 500 250 min 5000
Lysine Glutamine
6000 4000 2000 0
Alanine
Cystine
5 1400 1200 1000 800 600 400 200
min
min
Histidine
60000 40000 20000 0
Threonine
Valine
5 20000 15000 10000 5000 0 5 1000 800 600 400 200 5 10000 8000 6000 4000 2000 0 5 4000 3000 2000 1000 0 5
min 14000 12000 10000 8000 6000 4000 2000 0 min 900 800 700 600 500 400 300 200 min 7000 6000 5000 4000 3000 2000 1000 0 min 10000
min
Arginine
Leucine-isoleucine
Methionine
Hydroxyproline
5
min
Glysine
Cysteine
4000 3000 2000 1000 0
Tyrosine
min 6000 5000 4000 3000 2000 1000 0 min 1400 1200 1000 800 600 400 200
min
Serine
Glutamic acid
Phenylalanine
min
Aspartic acid
8000 6000 4000 2000 0 min
Proline
Tryptophan
min
min
Figure 1.
Determination of underivatized free amino acids using LC/APCI-MS. LC conditions: column, ZORBAX Bonus RP (2.1 mm 100 mm, 3.5 m); flow rate, 0.2 mL min1; mobile phase composition, 0.01 mM acetic acid in a 0.2% aqueous solution of formic acid; and MS signals for specific m/z ions of 22 amino acids in green pea.
Food samples In this study, 22 food samples were analyzed for their amino acid content. The food groups are baby foods, vegetables, fruits, juices, nuts, wine, beer, milk, chicken, and honey. The mean concentrations of amino acids in each food are given in Table 2.
Table 2. The Results of 22 Free Amino Acids in Various Food, as mg/100 g FW
Ala 3.59 10.07 69.84 < LOD < LOD 2.03 1.53 3.95 15.66 3.38 6.54 8.58 4.00 7.02 4.49 12.23 41.06 12.98 6.56 11.39 25.15 Leu-ile 0.30 0.96 26.84 1.38 < LOD 0.32 0.23 0.37 1.37 < LOD 1.54 1.80 < LOD 2.04 0.23 2.06 6.37 3.24 1.81 7.21 2.75 Arg 1.25 3.94 37.26 1.81 1.40 1.79 1.62 8.30 1.99 1.64 2.18 0.60 0.88 16.50 1.35 < LOD 2.38 1.28 2.71 4.14 7.58 Lys < LOD < LOD 17.41 < LOD < LOD < LOD < LOD 1.76 < LOD < LOD 1.21 1.39 1.17 1.43 1.28 4.80 4.35 2.06 1.64 4.93 3.64 Asn 1.07 11.02 112.10 45.27 3.80 12.80 24.10 8.08 13.58 78.57 10.50 5.32 2.98 8.99 0.94 4.16 31.36 2.21 0.71 57.51 14.45 Met 0.20 0.63 4.44 < LOD < LOD < LOD 4.52 4.72 0.53 < LOD 0.89 0.64 0.30 1.80 0.25 < LOD 1.66 < LOD 0.52 1.93 1.43 Asp 0.50 36.86 30.22 17.01 3.56 10.58 7.30 15.15 17.33 21.52 16.08 5.54 1.15 4.00 0.55 8.43 70.26 13.05 7.42 44.92 27.40 Phe 0.26 3.25 13.52 0.44 < LOD < LOD 0.14 0.99 0.94 0.45 0.81 1.88 0.15 0.61 0.25 11.18 14.31 2.41 1.09 6.08 381 Cys 0.17 1.31 6.96 0.60 0.30 0.98 0.38 0.70 0.27 0.84 0.50 0.74 < LOD 0.36 0.66 < LOD 43.19 < LOD < LOD 0.73 1.36 Pro 0.78 6.30 21.50 17.92 < LOD 4.48 6.52 30.55 8.52 12.89 8.94 14.05 0.84 11.61 30.50 81.32 32.30 5.38 5.50 33.65 55.98 Cys-Cys 9.77 1.52 3.62 13.58 16.04 12.52 10.95 9.99 16.53 12.26 16.06 17.60 5.33 16.12 0.92 203.22 14.82 6.02 4.90 4.50 5.82 Ser 0.40 3.38 55.52 3.86 0.57 0.78 1.29 6.07 3.01 2.48 2.80 4.30 < LOD 3.14 0.36 6.31 20.48 3.84 1.44 7.82 3.80 Glu 0.37 85.94 15.78 8.19 2.18 4.82 7.11 5.10 17.23 3.38 8.34 7.98 1.89 9.06 0.73 7.93 57.18 21.01 46.80 3100 35.64 Thr 1.66 2.70 0.03 1.23 < LOD 0.33 0.53 1.13 1.94 1.17 1.75 5.33 0.32 3.12 0.31 2.65 3.63 2.38 1.12 3.34 3.71 Gln 1.06 49.64 57.49 86.33 23.46 39.89 52.45 3.36 42.35 10.05 41.50 4.17 1.54 4.85 0.42 18.90 3.89 0.51 3.55 6.31 39.28 Tyr 0.50 2.05 15.45 0.68 < LOD < LOD 0.25 0.99 1.76 0.33 1.80 3.90 0.27 2.33 2.37 8.63 13.58 3.69 1.71 4.07 6.00 Gly 0.52 2.03 7.92 1.14 < LOD < LOD < LOD 1.38 < LOD 1.44 1.26 1.52 1.48 1.38 1.69 2.93 3.43 2.57 1.72 3.76 2.88 Trp 0.12 0.74 3.31 0.48 < LOD < LOD < LOD 0.58 0.64 < LOD 0.64 0.75 < LOD 0.65 1.86 < LOD 7.66 2.16 1.75 2.28 1.51 His 0.30 0.34 3.27 < LOD < LOD 0.13 1.01 1.11 0.17 0.14 0.11 1.46 0.26 0.75 0.46 1.26 3.27 < LOD 0.38 0.34 < LOD Val 1.21 2.43 38.60 5.25 < LOD 1.70 2.62 3.64 4.62 2.18 5.10 5.03 3.84 5.40 0.88 6.34 8.61 4.08 3.01 10.33 5.73 Hyp 1.42 0.73 1.17 < LOD < LOD 0.43 0.32 0.53 0.44 0.35 0.43 0.52 8.59 0.36 0.24 0.76 < LOD < LOD 0.74 0.46 7.26
Food Types Baby food Tomato Pea Pear Apple paste Apple juice Sour cherry juice Orange juice Pomegranate juice Peach juice White grape juice Red grape juice Beer Milk Wine Honey Green coffee Hazelnut Walnut Almond Pistachio Food Types Baby food Tomato Pea Pear Apple paste Apple juice Sour cherry juice Orange juice Pomegranate juice Peach juice White grape juice Red grape juice Beer Milk Wine Honey Green coffee Hazelnut Walnut Almond Pistachio
There was a great variation in the amino acid levels in each food. The variations in amino acid concentrations might be due to the differences in the composition of formulas and/or the differences in processing conditions. Also the protein composition can give rise to slight differences in the free amino acid profiles of foods. The pH of the samples was measured before extraction and found to range from 2 to 3 for juices, beer, and wine; about pH 4 for tomato; pH 6 for peas; and pH 7 for milk. It is very well known that the measurement of amino acids is very important in various areas, such as: 1) The determination of the geographical sources of honey from the ratios between concentrations of amino acids 2) Nutritional analysis as the amino acid supply in an infants first month of life must be sufficient in quantity and quality to fulfill the needs of this critical period. 3) Some of the important free amino acids, such as asparagine, glutamine, aspartic acid, and glutamic acid, are important precursors for acrylamide in various products.
Reference
S. zcan, H. Senyuva, Improved and simplified liquid chromatography/atmospheric pressure chemical ionization mass spectrometry method for the analysis of underivatized free amino acids in various foods, J. Chromatogr. A (2006), Doi:10.1016/j.chroma.2006.09.039

Conclusion
A simple, reliable, and rapid LC/MS method for the determination of 22 free amino acids in foods was developed in this study. The method was found to be applicable for a wide variety of foods. Our goal was to obtain a single-run LC/APCI-MS analysis of underivatized amino acids exhibiting sensitivity in positive mode, and this was improved by the use of further acidic LC/MS mobile phase to achieve a very short run time, ca. 7.5 min. Sample preparation and the subsequent chromatographic run took < 25 min to complete. Previous studies, which used conventional LC columns, long extractions with clean-up, and subsequent derivatization were expected to take one day to analyze five samples, whereas with this new, simplified, faster method, a batch of 20 samples can be completed in the same time.
Separation of Vitamin D2 from D3 and Other Fat-Soluble Vitamins
Application
Food Analysis
Robert D. Ricker
Analysis of fat-soluble vitamins is of widespread importance in clinical diagnosis, general research, and food product analysis. Separation of vitamins D2 and D3 is particularly difficult. Change in retention and resolution of these D vitamins and others is shown for both a classic, non-endcapped Agilent ZORBAX ODS column and its endcapped counterpart. Endcapping of bonded phases is performed in an effort to minimize secondary interaction of polar sample components with the column packing.
ZORBAX ODS Classic, 4.6 x 250mm Agilent Part No. 884950-543 (non-endcapped)
1 2 4 5 8
Highlights
6
Absorbance
The non-endcapped ZORBAX ODS column outperforms the endcapped ODS in obtaining resolution of vitamins D2 and D3.
ZORBAX ODS, 4.6 x 150mm Agilent Part No. 883952-702 (endcapped)

4 5 1 2 8 3 6 7
1. all-Trans Retinol 2. all-Trans Retinol Acetate 3. (+)-d-Tocopherol 4. Vitamin D2 5. Vitamin D3 6. ()-a-Tocopherol (Vitamin E) 7. (+)-a-Tocopherol Acetate 8. Vitamin K1
Endcapped and non-endcapped versions of the ZORBAX ODS column are both able to separate a wide variety of fat-soluble vitamins; although, with differing selectivities (see peak pairs 3-4 and 5-6).
Conditions: LC: Hewlett Packard HP 1050 Mobile Phase: 75% ACN : 25% MeOH Det.: UV: 325 nm for 4 min., 280 nm after Flow: 1.0 mL/min.; 40C Inj Vol: 5 L (10 g / L)
High-Speed Separation of Water-Soluble Vitamins with Ion-Pairing Chromatography

Application
Food Analysis
Robert Ricker
The rapid separation and detection of vitamins is becoming increasingly important to both food and pharmaceutical industries, as well as basic research. Water-soluble vitamins can be easily separated using reverse-phase ion-pairing chromatography. Four B vitamins, folic acid and vitamin C were analyzed in less than 5 minutes using a 75!mm SB-C8 column.
Highlights Short run times and excellent peak shape of water-soluble vitamins are achieved using a low-pH mobile phase and a short (75!mm) ZORBAX SB-C8 column with small-particle packing (3.5!m).
3 2
". Vitamin C 2. B3 - Niacin 3. B6 - Pyridoxine 4. Folic Acid 5. B2 - Riboflavin 6. B" - Thiamine
ZORBAX StableBond technology offers reproducibility and column stability even at low pH (0."% phosphoric acid) and with ion-pairing agents. 6 " 4 5
min
Conditions: ZORBAX SB-C8 (4.6 x 75 mm, 3.5m) (Agilent P/N: 866953-906) Mobile Phase:!"0mM Hexane Sulfonate with 0."% Phosphoric Acid:MeOH (74:26) Injection volume "0 L, " mL/min, Ambient, Detect: UV(245 nm)
Separation of Water-Soluble Vitamins Using the USP 23 Method

Application
Food Analysis
Robert Ricker
Recently, USP-recommended methods for vitamin analysis have become increasingly important. Analysis and quantitation of water-soluble vitamins using USPrecommended chromatographic methods can become difficult when column performance is questionable. Routine analyses performed by analysts in the pharmaceutical industry require the use of reproducible and stable columns.
Highlights
StableBond columns provide analysts with long-term stability and reliability for the demands of USP-generated methods.
USP Standard Preparation
3 4
Analysis of Vitamin Formulation 1. B3 - Niacin 2. B6 - Pyridoxine 3. B2 - Riboflavin 4. B1 - Thiamine E. Excipient 3
Lot-to-lot reproducibility of StableBond column packings assures the chromatographer of consistency in USP-generated methods.
Details of the USP Method appear on page!2"68 of! USP!23.
2.5
7.5
10
12.5
15
17.5
min
Conditions: ZORBAX SB-C18 (L1 packing) (4.6 x 250 mm) (Agilent P/N: 880975-902) Mobile Phase:!7.2 mM Hexane Sulfonate:MeOH:Acetic Acid (73:27:1) (ratio to 101%) Injection: 20 L, 1 mL/min, 30C, Detect. UV(280 nm)
Separation of Water-Soluble Vitamins Using Reversed-Phase Chromatography

Application
Food Analysis
Robert Ricker
The analysis and quantitation of water-soluble vitamins has recently become an area of interest in the pharmaceutical industry. An alternative to routine isocratic separation, the use of a reversed-phase gradient (without ion pairing), allows analysis of an 8component sample containing B!vitamins, pantothenic acid, folic acid, and vitamin C. A ZORBAX SB-C8 column was used for the analysis.
Highlights
High-speed gradient analysis of 8 watersoluble vitamins is achieved using a short (150!mm) ZORBAX SB-C8 column and a low-pH mobile phase.
". B" - Thiamine 2. Vitamin C 3. B3 - Niacin 4. B6 - Pyridoxine 5. Pantothenic Acid 6. Folic Acid 7. B"2- Cyanocobalamin 8. B2 - Riboflavin
Gradient reversed-phase separations produce good peak shape for vitamin B12 and pantothenic acid in addition to typical water-soluble vitamins.
5 6 4 7
StableBond packings offer reproducibility and stability at low pH.
8 " 3
2.5
10
12.5
min
Conditions: ZORBAX SB-C8 (4.6 x 150 mm) (Agilent P/N: 883975-906) Mobile Phase:!A=50mM Sodium Phosphate, pH 2.5:MeOH (90:10) B=50mM Sodium Phosphate, pH 2.5:MeOH (10:90); Gradient 0-70% B/18 min. Injection: 10!L, 1 mL/min, Ambient, Detect. UV(245 nm)
Fat-Soluble Vitamins on ZORBAX XDB-C8

Application
Food Analysis
Robert Ricker
The fat-soluble vitamins are comprised of vitamins A,D, E and K. Margarine, milk products, breakfast cereals and infant formulas are fortified with vitamins A, D and E; Vitamin K is fortified in infant formulas, only. Assays are needed to conform with nutrient-labeling regulations and to study changes in vitamin content due to storage, packaging, and processing. Normal-phase chromatography has been the chromatographic mode of choice. This application demonstrates a reversed-phase separation using highorganic concentration (>90% methanol) on a ZORBAX XDB-C8 column. By using reversed-phase chromatography for water-soluble and fat-soluble vitamins, labs requiring both assays can now minimize changeover time. The extradense bonding and double endcapping of ZORBAX Eclipse packings (e.g.,!XDB- C8) shield ionized silanols that may interact with analytes during high-organic, reversedphase separations.
Highlights
Methods for normal-phase chromatography of fat-soluble vitamins can be replaced by highorganic, reversed-phase separation.
Operation at high temperature results in reduced elution times, peak widths, and run-times.
". Retinol 2. Retinol Acetate 3. Vitamin D3 4. -Tocopherol 5. -Tocopherol 6. Tocopherol Acetate 7. Retinol Palmitate
Conditions: ZORBAX XDB-C8 (4.6 x 150mm) (Agilent P/N: 993967-906) Mobile Phase:!(5:95) water : MeOH Injection 5L, 1 mL/min, Detect. UV (280 nm)
Separation of Retinal (Vitamin A) Isomers:

Comparison of Mobile-Phase Composition
Application
Food Analysis
Robert Ricker
HPLC separation of vitamin A isomers (retinal isomers) is important in the analysis of various food stuffs, metabolism of visual pigments, liver function, and the corpus luteum. HPLC has traditionally been carried out using 1,4-dioxane. Efforts here were to find a more friendly solvent and a column that provides good chromatographic performance.
Highlights
Various solvent combinations can be used in conjunction with ZORBAX Sil to provide adequate resolution of Vitamin A isomers.
A= Retinal Isomers 11,13 -di-cis13 -cis9,11,13 -tri-cis9,13 -di-cis11 -cis7,11 -di-cis9 -cis7,9 -di-cis7,13 -di-cisAT all trans retinal B-F= Retinol Isomers
ZORBAX Sil silica provides good peak shape for these retinoids.
Courtesy of Dr. G. Nll Physiologisches Inst. -- Justus Liebig Uni. Giessen Conditions: ZORBAX Sil (4.6 x 250 mm) (Agilent P/N: 880952-701) Injection volume 20L, 25C
Rapid Analysis of Water-Soluble Vitamins with Extraction from Cat Food Application
Food and Flavors

Robert D. Ricker* Thomas Schellenberger**
The qualitative and quantitative analysis of vitamins is of increasing importance in both the medical and food-related sciences. Such analyses may be carried out for vitamins found in a variety of matrices, including tablets, feed and foodstuffs, pharmaceuticals and cosmetics, and beverages. The analysis of water-soluble vitamins in pet foods is important for ensuring that products have met internal (corporate) as well as government quality standards. The following work describes the analysis of water-soluble vitamins in cat food but is applicable to analysis of these vitamins from many sources. The goal was to develop a high-throughput analysis of essential water-soluble vitamins in cat food.
Highlights
Six water-soluble vitamins can be analyzed and sensitively detected within their background matrix after extraction from cat food, within 20 minutes. Agilent Zorbax 3.5 m packing materials in shorter columns (75 mm were used here) provide rapid, high-resolution analysis. Shorter columns allow high flow rates that result in rapid isocratic or gradient analysis.
20
Nicotinic acid Pantothenic acid
Riboflavin B2
Mobile Phase A = 1L H2O, 2.5 g hexane sulfonic acid, 2.5 mL HOAc, 4 g NaH2PO4, 50 L triethylamine B = 600 mL H2O, 2.5 g hexane sulfonic acid, 2.5 mL HOAc, 50 L triethylamine, 400 mL ACN Gradient Timetable
6 8 10 12 min
15
10
Folic acid
B6
5 UV = 200nm
4
Thiamin B1
0 0 2 4
Conditions Instrument Column Flow Temperature Injection volume Detection
Agilent 1100 3D HPLC System Agilent Zorbax SB-C18, 4.6 75 mm, 3.5 m particle size, part no: 866953-902 1.9 mL/min Ambient 20 L UV at 275 nm plus spectra
Time (min) 0.00 3.60 6.00 13.50 14.40 16.50 16.80 20.00
% Solvent B 0.0 0.0 20.0 55.0 90.0 90.0 0.0 0.0
Sample Preparation
The total content of one can (125 g) of cat food is homogenized using a kitchen mixer or by mashing. Of the total sample, 10.0 g are transferred into a 250 mL brown glass flask. Approximately 120 mL saturated Titriplex-V solution (Titriplex-V dissolved in MeOH/Water, HPLC-grade) is added and the sample is stirred for exactly 5 minutes in a water bath at 90C. The measuring flask is filled nearly to the mark with saturated Titriplex-V solution and cooled to room temperature. After filling to the mark with saturated Titriplex-V solution, the sample is mixed and filtered using a syringe-filter (for example, hydrophilic Teflon membrane, pore size 0.45 m). An aliquot is placed in the Agilent 1100 autosampler using appropriate vials and 20 L are injected.
Water-Soluble Vitamins Detected 1. Nicotinic acid 2. Pantothenic acid 3. Folic acid 4. Pyridoxine B6 5. Riboflavin B2 6. Thiamine B1

* Robert Ricker is an application chemist based at Agilent Technologies, Wilmington, Delaware. ** Thomas Schellenberger is a manager of the analytical laboratory at the fine chemical division at BASF AG, Ludwigshafen, Germany.
Analysis of Fat-Soluble Vitamins by HPLC
Udo Huber Pharmaceutical
Vitamins are compounds, which are necessary to maintain a healthy and properly functioning human organism. A few milligrams a day are enough to regulate the utilization of the nutrients, such as carbohydrates, fats, proteins and minerals. Usually vitamins are supplied in food. Figure 1 shows the separation of the three fat-soluble vitamins retinol (A), cholecalciferol (D3) and a-tocopherol (E) using gradient analysis on a reversed phase column and UV detection. The autosampler temperature was set to 4 C to avoid decomposition of the samples.
Conditions
Absorbance [mAU] 800 600 400 200 0 0 2.5 5 7.5 10 12.5 15 Time [min] 17.5 20 1 2 1 Retinol (A) 2 Cholecalciferol (D3) 3 -Tocopherol (E) 3 Column 4.6 x 75 mm Zorbax Eclipse XDB-C18, 3.5 m Mobile phase A = water, B = methanol Flow rate 1.0 ml/min Gradient at 0 min 90 % B at 15 min 100 % B at 20 min 100 % B Column wash at 21 min 90 % B UV detector variable wavelength detector 210 nm, standard cell Column compartment temperature 20 C Stop time 21 min Post time 5 min Injection volume 5 l
Figure 1 Analysis of three fat-soluble vitamins
HPLC Performance Compound LOD for S/N=2 (mg/l)* 4.0 2.5 2.0 Precision of RT (RSD of 10 runs) (1000 mg/l)* 0.10 0.04 0.04 Precision of Area (RSD of 10 runs) (1000 mg/l)* 0.14 0.16 0.20
Equipment
Agilent 1100 Series Quaternary pump (includes vacuum degasser) Thermostatted autosampler Thermostatted column compartment Variable wavelength detector, standard flow cell 10-mm path length, 13-l cell volume Alternative: Binary pump Vacuum degasser Diode array detector standard flow cell 10-mm path length, 13-l cell volume Agilent ChemStation + 3D software Columns Zorbax Eclipse XDB, 3. 5 m, 4.6 x 75 mm (Agilent part number 7995118-344) Recommended: Guard cartridges Zorbax Eclipse XDB-C18, 5 m, 4 x 4 mm (Agilent part number 7995118-504, 10/pk) Note: Since the method was specifically developed on the Agilent 1100 Series system you might not be able to reproduce this analysis on an older system or even on a new system with lower performance. To avoid sample decomposition it is necessary to use a cooled autosampler, for example, the Agilent 1100 Series thermostatted autosampler.
Retinol Cholecalciferol a-Tocopherol
* Injection volume: 5 l The performance of the HPLC method is shown in the table above. The HPLC method presented here shows an easy but reliable and precise analysis of the fat-soluble vitamins retinol (A), cholecalciferol (D3) and a-tocopherol (E). The values for LOD, precision of RT and precision of area show the good performance of the analysis.
HPLC Analysis of Vitamins in Tablets using HPLC
Abstract Fat-soluble vitamins, such as vitamins E, D, and A, and water-soluble vitamins, such as vitamins C, B6, B2, B1, and B12, have been analyzed. Vitamins are biologically active compounds that act as controlling agents for an organisms normal health and growth. The level of vitamins in food may be as low as a few micrograms per 100 g. Vitamins often are accompanied by an excess of compounds with similar chemical properties. Thus not only quantification but also identification is mandatory for the detection of vitamins in food. Vitamins generally are labile compounds that should not exposed to high temperatures, light, or oxygen. HPLC separates and detects these compounds at room temperature and blocks oxygen and light.1 Through the use of spectral information, UV-visible diode-array detection yields qualitative as well as quantitative data. Another highly sensitive and selective HPLC method for detecting vitamins is electrochemical detection.
Folic acid, diCa
Conditions
Column 100 4 mm Hypersil BDS, 3 m Mobile phase A= water with pH = 2.1 (H2SO4) = 99% B = ACN + 10% A = 1 % Gradient at 3.5 min 1% B; at 11 min 25% B at 19 min 90% B Post time 6 min Flow rate 0.5 ml/min Column compartment 30 C Injection vol 25 l Detector UV-DAD detection wavelength 220/30 nm, reference wavelength 400/100 nm Sample preparation Filtration
1000 500 0 Citric acid 0 2 4
B12 Riboflavin Biotin 12
1500
Pantothenic acid
Norm
B1
B6
Standard Vitamin tablet
Saccharin 10
8 6 Time [min]
Figure 1 Analysis of water-soluble vitamins in a vitamin tablet
Sample preparation Different food matrices require different extraction procedures.1 For simple matrices, such as vitamin tablets, water-soluble vitamins can be extracted with water in an ultrasonic bath after homogenization of the food sample. Chromatographic conditions The HPLC method presented here was used to analysis vitamins in a vitamin drink. HPLC method performance Limit of detection <500 pg (injected amount), S/N = 2 Repeatability of RT over 10 runs <0.2 % areas over 10 runs <2 %
Norm
800 400 0 Riboflavin
Equipment
Agilent 1100 Series vacuum degasser quaternary pump autosampler thermostatted column compartment diode array detector Agilent ChemStation + software
Norm
Folic acid 400 200
250
350 Norm 1000

600 200
450
0 550 nm
250
350
450
550 nm
Vitamin B 1,B 6, B 12
250
350
450
550 nm
Figure 2 Analysis of carbohydrates in corn extract References 1. L.M. Nollet, Food Analysis by HPLC, New York, 1992.
Analysis of Selected Vitamins with HPLC and Electrochemical Detection

Application Note
Food and Clinical
Angelika Gratzfeld-Hsgen, Rainer Schuster, Wolfgang Haecker
Several oxidizable vitamins (vitamin A, B6, C, D3 and E) were analyzed using HPLC and electrochemical detection. Excellent sensitivity was achieved with all vitamins. The minimum detectable level for vitamin Apalmitate was about 80 pg, for -carotene about 50 pg, for vitamin B6 about 30 pg, for vitamin C below 1 pg, for vitamin D3 about 70 pg and for vitamin E about 30 pg. The optimum potentials were evaluated using an automated increment for the potentials. The reproducibility was tested using vitamin B6 and was found to be 3% RSD of peak height for 100 runs. Due to the high selectivity of electrochemical detection, it was possible to keep sample preparation very simple.
Introduction
In the last decade, liquid chromatography (LC) with electrochemical detection (ECD) has been used more and more for the determination of electroactive compounds at trace levels and in complex matrices. For the analysis of electroactive vitamins in food, pharmaceutical preparations, tissue and body fluids, the combination of HPLC and ECD provides high sensitivity and excellent selectivity 1,2,7,8,9,11. Electroactive vitamins can be determined in the low pg range and moreover, fewer matrix effects are encountered. This enables sample preparation and enrichment to be simplified and reduced 11.
Despite the high sensitivity and selectivity of electrochemical detection for vitamin analysis, it is not widely used. It has a reputation for low response stability due to the fact that the electroactive surface of the detector cell is in direct contact with the mobile phase and all the compounds and products of the oxidative and reductive reaction. This can result in contamination of the working electrode and an unstable response. The response instability can be minimized by using modern electrochemical detectors with electrochemical self-cleaning routines 19. This note describes a method for analyzing oxidizable vitamins with high sensitivity and reproducibility using a modern electrochemical detector.
Experimental
We used Agilent 1049A electrochemical detector with HP 1050 Series isocratic pump and autosampler. The chromatographic conditions for each experiment are listed with the appropriate figures.
Results and discussion Analysis of vitamin A (retinol), provitamin A (-carotene) in fruit juice
Vitamin A is one of the fat-soluble vitamins and is of importance for normal growth and development of the human body. Malabsorbtion results in disease of the liver 1 and an insufficiency results in extreme dryness of the skin and nightblindness 12. Sources of vitamin A are liver, fish oil, milk, butter and the yolk of eggs. A further source is provitamin A (-carotene), which is present in many plants and is converted in the human body to retinol (vitamin A1). The analysis of vitamin A is of interest for quality control of food and pharmaceutical products and for diagnosis and therapeutic reasons in clinical routine testing and medical research.
Measurement of this vitamin is nowadays based on HPLC methods 10 using UV-Visible, fluorescence or electrochemical detection. The analysis of vitamin A and/or -carotene with electrochemical detection has several advantages over UV-Visible or fluorescence detection. UV detection shows a lack of sensitivity and selectivity 1, 2, 4, whereas fluorescence detection normally offers sufficient sensitivity but sometimes (depending on the matrix) insufficient selectivity 1,3. Using an electrochemical detector, the minimum detectable level for vitamin A-palmitate was about 80 pg and for -carotene about 50 pg with a signal-to-noise ratio of 2, (figure 1).
In figure 1 the analysis of vitamin A-palmitate and -carotene is shown. 0.5 l of a standard mixture was injected and the minimum detectable level was calculated. In order to measure at optimum sensitivity or at the highest possible selectivity, it is necessary to evaluate the optimum detection potential for each compound of interest. This can be done using built-in software routines (auto-increment mode). The sample is injected as often as specified between various selected potentials. The voltage difference between these potentials can also be determined by the user.
Vitamin A-palmitate
CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH2OR
-carotene
Column
Mobile phase
Flow rate Oven temperature Injection volume Detector parameters Operation mode Potential Range Reference electrode Response time
125 x 4 mm Lichrospher RP18, 5 (Part No. 799250-564) Methanol + 5 g/l lithiumperchlorate + 1 g/l acetic acid 1.5 ml/min 30 C 0.5 l
mV 119.0 118.8 118.6 118.4 118.2 118.0 117.8 117.6 117.4 117.2 117.0 116.8
CH3 CH3
3HC
3HC
CH3
CH3
Amperometry 1V 0.5 A AgCl/KCl 8 s.
Vitamin A-palmitate 208pg
-carotene 146pg
10 Time [min]
15
20
Figure 1 Sensitivity of electrochemical detection with vitamin A and -carotene
Figure 2 shows the result of the automated optimization procedure. At 1 V the optimum signal-to-noise ratio was obtained. Further increase of the oxidative potential did not increase the signal height but produced more noise and drift, and so no better signal-to-noise ratio could be achieved. The optimization procedure also showed how easily -carotene can be oxidized. Oxidation started at
0.5 V. This can be of advantage when high selectivity is more important than high sensitivity. For the analysis of vitamins in food, the main problem is often not only sensitivity but also selectivity 13-17. The high selectivity of electrochemical detection means that problems due to the matrix or coeluting peaks can be avoided without time-consuming
sample preparation. For example, -carotene in fruit juice can be analyzed by direct injection into an HPLC instrument, figure 3. Figure 3 shows that, with 1 l injection volume and no sample preparation, -carotene in fruit juice was easily determined. The concentration of -carotene was found to be 2.7 mg/100 ml fruit juice.
Scaled 11 10
Column 125 x 4 mm Lichrospher RP18, 5 (Part No. 799250-564) Methanol + 5 g/l lithiumperchlorate 1.5 ml/min 30 C 3 l
Vitamin A-palmitate 1.248 ng
-carotene
0.876 ng
9 8 7 6 5 4 3 2 1 0 6 8 10 12 Time [min] 14 16 18
1.1 V 1.0 V 0.9 V 0.8 V 0.7 V 0.6 V 0.5 V 0.4 V
Mobile phase + 1 g/l acetic acid Flow rate Oven temperature Injection volume
Detector parameters Auto-increment parameters Start potential 0.4 V End potential 1.1 V Increment 0.1 V
Figure 2 Optimization of detection potentials
Scaled
Column 125 x 4 mm Lichrospher RP18, 5 (Part No. 799250-564) Methanol + 5 g/l lithiumperchlorate + 1 g/l acetic acid 1.5 ml/min 30 C 1 l
800 700 600 500 400 300 200
Vitamin A-palmitate 420 ng

-carotene
290 ng
Mobile phase
Flow rate Oven temperature Injection volume Detector parameters Operation mode Potential Range Reference electrode Response time
Standard
Amperometry 1V 0.5 A AgCl/KCl 8 s.
100 0 5 10
2.7 mg/100 ml fruit juice 27 ng 15 Time [min] 20 25
Figure 3 Analysis of -carotene in fruit juice
Analysis of vitamin B6 (pyridoxine hydrochloride) in a pharmaceutical vitamin preparation

Vitamin B6 is a water-soluble vitamin and consists of three interconvertible forms (pyridoxine hydrochloride, pyridoxamine dihydrochloride and pyridoxal hydrochloride). Vitamin B6 is of great importance for the nervous system 5, 6. In food, it is present in liver, bananas, wheat and vegetables. Whereas other methods show lack of sensitivity and selectivity, e.g. the minimum detectable level with HPLC and fluorescence detection is in the low ng range 5, a combination of HPLC and electrochemical detection enables the analysis of vitamin B6 in the pg
range (30 pg with a signal-to-noise ratio of 2), figure 4. Figure 4 shows, that the determination of vitamin B6 is possible even when vitamin C is present in large quantities. The sample preparation was kept very simple: 10 l of the vitamin preparation was diluted with 1 ml distilled water and 0.5 l of the liquid was injected. The vitamin concentration was 804.16 mg/l vitamin preparation. In order to evaluate the reproducibility of this type of analysis, 5 l (450 pg) of the standard sample were injected. The reproducibility was about r.s.d. = 3% for peak heights and about 5% for area counts over 100 runs.
Column
Mobile phase
Flow rate Oven temperature Injection volume
125 x 4 mm Lichrospher RP18, 5 (Part No. 799250-564) Water + 0.02M KH2PO4 + 0.03M tetrabutylammoniumhydrogensulfate + 0.03M heptanesulfonic acid + 2% acetonitrile 0.8 ml/min 30 C Standard 1l Sample 0.5 l
mV
Vitamin C
240 220 200 180 160 140 120 0 1
3HC
Vitamin preparation with 804.21ug/l Vitamin B6
HO
CH2OH
CH2OH
Vitamin B6, 402 pg *
Detector parameters Working electrode Operation mode Potential Range Reference electrode Response time
Glassy carbon Amperometry 1.2 V 0.5 A AgCl/KCl 1s
Vitamin B6, 90 pg * 2 3 Time [min] 4 5
Standard
Figure 4 Analysis of vitamin B6 in a vitamin preparation for chicken feed
Analysis of vitamin C (ascorbic acid) in fruit juice

Vitamin C belongs to the watersoluble vitamins and is of vital importance for mammalian health. Lack of this vitamin is responsible for scorbutus (scurvy) and in former times many people (specially sailors) died because of this deficiency disease. Vitamin C is present in many fruits and vegetables 12. The analysis of vitamin C in food is mainly performed for quality control purposes. Determination of this vitamin in tissue and body fluids is mainly carried out to increase knowledge of its physiological role in mammalians 7. There are several methods available for the analysis of vitamin C, but most of them show lack of sensitivity and/or selectivity. Tedious sample preparation steps are needed to remove interfering compounds 7,11,18. HPLC, in combination with UV-Visible
detection for example, shows insufficient sensitivity for the analysis of vitamin C in tissue and body fluids 7,11. The minimum detectable level for this method is about 4 ng. Using HPLC and electrochemical detection, the minimum detectable level for vitamin C is below 1 pg. (Due to the low stability of vitamin C, the determination of the minimum detectable level must be done directly after dilution of freshly prepared stock solutions.) The analysis of vitamin C in food such as fruit and fruit juices, is mainly done to control the freshness and proper storage of these products. Figure 5 shows the chromatogram of a fruit juice which contained 43.16 mg vitamin C in 100 ml fruit juice. The fruit juice was prefiltered, 10 times diluted and 1 l of the diluted sample was injected. The standard solution contained 1 mg vitamin C in 100 ml water.
Column
Mobile phase tetrabutylammonium-
100 x 4.6 mm ODS Hypersil, 5 (Part No. 79916OD-554) Water + 5.4 g/l Na acetate-trihydrate + 3g/l hydrogensulfate + 3 ml KCl solution (3.3 m), pH 5 with acetic acid. 0.8 ml/min 30 C 1 l
Scaled 500 400 300 200 100 0

HO OH CH2OH H-C-OH O
43.16 ng Ascorbic acid = 43.16 mg/100 ml fruit juice
Flow rate Oven temperature Injection volume Detector parameters Working electrode Operation mode Potential Range Reference electrode
10 ng/ul Standard
Glassy carbon Amperometry 0.6 V 0.5 A in-situ reference electrode AgCl/KCl in mobile phase 8s
Time [min]
Response time
Figure 5 Analysis of vitamin C (ascorbic acid) in fruit juice
Analysis of vitamin D3 (colcalciferol) in a multi-vitamin preparation

Vitamin D3 belongs to the fatsoluble vitamins and is of great importance for the metabolism of bone. Lack of this vitamin results in bone deformation from rachitis (rickets). The determination of vitamin D3 in pharmaceutical preparations, tissue and body fluids plays an important role in quality assurance procedures and in the understanding of bone metabolism 8. The methods in use are generally timeconsuming, involving several sample preparation steps and some of them are not sufficiently sensitive or selective 8. The combination of HPLC and electrochemical detection offers excellent sensitivity. The minimum detectable level is about 70 pg with a signal-to-noise ratio of 2.
Figure 6 shows the analysis of vitamin D3 in a highly viscous liquid multivitamin solution, which is used in poultry farms. Due to the excellent selectivity of electrochemical detections, sample preparation could be kept very simple. 40 l of the vitamin liquid was shaken with 1 ml of cyclohexane for 1 min and 5 l of the supernatant liquid were injected. The vitamin D3 concentration in this preparation was found to be 1,73 mg/l.
mV
Vitamin preparation 1.73 mg/ml Vitamin D3
3HC
CH 3 CH 3
140
CH 3
Column RP18, 5
H
125 x 4 mm Lichrospher (Part No. 799250-564) Methanol + 5 g/l lithiumperchlorate + 1 g/l acetic acid 1 ml/min 30 C 5 l
130 Vitamin D3, 346.97 pg *
H CH2
Mobile phase
HO
Flow rate Oven temperature Injection volume Detector parameters Working electrode Operation mode Potential Range Reference electrode Response time
120
Standard 136.7pg 110 0 5 10 15
Time [min]
Figure 6 Analysis of vitamin D3 in a pharmaceutical multivitamin preparation
Analysis of vitamin E (-tocopherol) in vitamin E capsules

Vitamin E belongs to the fatsoluble vitamins. In mammalians it fulfills the function of a lipidsoluble antioxidant and is necessary for the maintenance and protection of neuronal, muscular and reproductive tissue 9. The analysis of vitamin E is of interest in the food and cosmetic industry, see for example, the analysis of this vitamin in butter or milkpowder 10 or in beauty creams, where it is used as an antioxidant 12 . The analysis of vitamin E in tissue has become more and more important, because there is a possibility that vitamins A and E may act as cancer-chemopreventive agents 4.
Measurements of plasma vitamin A and E provides an indication of the proper absorbtion of these two compounds. Malabsorbtion of these vitamins occurs in cystic fibrosis and in cholestative liver diseases 1. The analysis of tocopherols with HPLC and electrochemical detection also offers more sensitivity and selectivity compared with HPLC and UV-Visible or fluorescence detection, figure 7 9. A standard solution of 100 pg in 1 l (figure 7) was injected and the minimum detectable limit was calculated. It was found to be about 30 pg for a signal-to-noise ratio of 2.
mV
3HC
CH3 O
CH3
CH 3 CH 3
118.5
Column 125 x 4 mm Lichrospher RP18, 5 (Part No. 799250-564) Methanol + 5 g/l lithiumperchlorate + 1 g/l acetic acid 1 ml/min 30 C 1 l
HO
CH3 CH3
118.0
Mobile phase
CH 3
-Tocopherol 100 pg
117.5
117.0
116.5 0 2 4 Time [min] 6 8 10
Figure 7 Sensitivity of tocopherol analysis with electrochemical detection
In figure 8, the analysis of -tocopherol and traces of other tocopherol isomers in a pharmaceutical preparation is shown. For sample preparation the capsule was cut and extracted with 5 ml of methanol in an ultrasonic bath for 10 minutes. 3 l of the supernatant liquid was injected. The concentration of a-tocopherol was found to be 270 mg/capsule.
The chromatogram shows that not only -tocopherol, but also traces of other tocopherol isomers, such as beta, delta and/or gamma tocopherol, were present. Under the chromatographic conditions used, gamma and delta tocopherol could not be separated.
mV 260
mV 124 123 122 121 120 119 118 117 116 115
-tocopherol -tocopherol
270 mg/capsule
Column
Mobile phase
125 x 4 mm Lichrospher RP18, 5 (Part No. 799250-564) Methanol + 5 g/l lithiumperchlorate + 1 g/l acetic acid 1 ml/min 30 C 3 l
240 220 200 180 160
/ -tocopherol ? -tocopherol
140 120 0
3 Time [min]
3 Time [min]
Figure 8 Analysis of tocopherols in vitamin E capsules
References
1 May-Lynn Huang, Gilbert J. Burckart, Raman Venkataramanan, Sensitive high performance liquid chromatographic analysis of plasma vitamin A using amperometric and ultraviolet detection, J.Chromatogr., 1986, 380, 331-338 2 Johanna K.Lang, Lester Packer, Quantitative determination of vitamin E and oxidized and reduced coenzyme Q by highperformance liquid chromatography with in-line ultraviolet and electrochemical detection, J.Chromatogr., 1987, 385, 109-117 3 A.T.Rhys Williams, Simultaneous determination of serum vitamin A and E by liquid chromatography with fluorescence detection, J.Chromatogr., 1985, 198-201 4 David W.Nierenberg, Deborah C.Lester, Determination of Vitamins A and E in serum and plasma using simplified clarification method and high performance liquid chromatography, J.Chromatogr., 1985, 345, 275-284 5 Leslie A. Morrison, Judy A. Driskell, Quantities of B6 vitamers in human milk by high performance liquid chromatography, J.Chromatogr., 1985, 337, 249-258
6 Weiying Hou, Huamin Ji, Erkang Wang, Liquid chromatography of vitamin B6 with electrochemical detection using a carbon fibre electrode, Anal.Chim.Acta, 1990, 230, 207-211 7 Philip W. Washko, William O. Hartzell, Mark Levine, Ascorbic acid analysis using high performance liquid chromatography with coulometric electrochemical detection, Anal.Biochem., 1989, 181, 276-282 8 A. Sanchez Perez, M. Delgado Zamarreno, J. Hernandez Mendez, R.M. Sanchez Rodriguez, Flowinjection analysis of vitamin D3 and 25-hydroxy-vitamin D3 by amperometric detection, Anal.Chim.Acta, 1989, 225, 247-251 9 Michael E. Murphy, James P.Kehrer, Simultaneous measurement of tocopherols and tocopheryl quinones in tissue fractions using high performance liquid chromatography with redox-cycling electrochemical detection, J.Chromatogr., 1987, 421, 71-82 10 W. Kneifel, F. Ulberth, U. WinklerMacheiner, Simultaneous determination of retinol and tocopherol in butter and milk powder with high performance liquid chromatography, (original in German) Z. Deutsche LebensmittelRundschau, 1987, 83, 137-139
11 Karel Stulik, Vera Pacakova, Electroanalytical measurements in flowing liquids, Prague 1987, Ellis Horwood Ltd series in analytical chemistry, New York 12 Roempps Chemie Lexikon, 1979, 8. Aufl., Franckhsche Verlagshandlung, Stuttgart 13 David W. Nierenberg, Serum and plasma B-carotene levels measured with an improved method of high performance liquid chromatography, J.Chromatogr., 1985, 339, 273-284 14 R. Ohmacht, Gy. Toth, G. Voigt, Separation of serum carotenoids and vitamin A on Chromsil-Amino and Cyanophases by bi-directional gradient elution technique, J.Chromatogr., 1987, 395, 609-612 15 R.J. Bushway, Separation of carotenoids in fruits and vegetables by high performance liquid chromatography , J.Liq.Chromatogr., 1985, 8(8), 1527-1547 16 Yen-Ping C. Hsieh, Marcus Karel, Rapid extraction and determination of a- and B-carotenes in foods, J.Chromatogr., 1983, 259, 515-518
17 Adriana Z. Mercadante, Delia B. Rodriguez-Amaya, Comparison of normal-phase and reversed-phase gravity-flow column methods for provitamin A, Chromatographia, 1989, 28, 249-252 18 Antal Bognar, Determination of vitamin C with high performance liquid chromatography, (original in German) Z. Deutsche Lebensmittel-Rundschau, 1988, 84, 73-76 19 A.G. Huesgen, R. Schuster Selective HPLC analysis of phenols in river water, Agilent Technologies Application Note ,1990, 04/90, Pub. no. 5952-1548
Angelika Gratzfeld-Hsgen, Rainer Schuster and Wolfgang Hcker are based at Agilent Technologies, Waldbronn, Germany.
The information contained in this note is subject to change without notice.
Bioanalysis Applications
Rapid wheat varietal identification using the Agilent 2100 bioanalyzer and automated pattern-matching
Application Note
Dr Dhan Bhandari
Abstract
Agilent Equipment: 2100 Bioanalyzer Protein 230 Kit Application Area: Food analysis
Accurate identification of wheat varieties is of paramount importance to the milling industry in many countries. This Application Note describes how the Agilent 2100 bioanalyzer and the Protein 230 assay can be used in conjunction with a third-party software, to analyze wheat proteins for varietal identification.
Introduction
Authentication of varieties is important for the cereal industry for maintenance and testing of grain quality to meet market requirements. The charge-based separation of proteins by the acidPAGE (polyacrylamide gel electrophoresis) technique is widely used for wheat varietal identification. However, this requires highly skilled operators to prepare, run and scan the gels and interpret band patterns. Also, there are safety concerns regarding the toxicity of unpolymerised acrylamide. While acid-PAGE is effective in analytical laboratories, the routine method can take up to two days. This is too slow for use at mill intake, which requires assessment of the wheat shipment during the period of delivery typically under one hour. In this Application Note we demonstrate the use of the Agilent 2100 assay, with bioanalyzer and Protein 230 the Nonlinear Dynamics Phoretix 1D Advanced (TotalLab TL120 DM) computerized pattern-recognition software to provide a better alternative to acid-PAGE. The aim of this study was to develop a robust, automated method for rapid identification of wheat varieties. Methods Total wheat proteins (including glutenins) were extracted from individual grains in 0.4 mL of 2M urea, 15 % glycerol, 0.1 M DTT and 0.1 M Tris/HCl, pH 8.8, using an ultra-sonic water bath for 15 minutes. Extracts were centrifuged at 11,000 g for 5 minutes and treated with the Protein 230 assay reagents in according to the
-, -, Gliadins
HMW-Glutenins -Gliadins, L MW-Glutenins
Lower Marker
Upper Marker
Figure 1 Typical wheat protein separation by the Protein 230 assay.
Figure 2 Bioanalyzer gel-like image of MW standards and 10 different wheat varieties.
assay protocol. The samples were separated on the Agilent 2100 bioanalyzer, with each analysis of 10 samples plus ladder taking less than 25 minutes. Replicates of 34 wheat cultivars representing the 2004/5 UK Recommended List
varieties were analyzed. The electropheromram profiles were processed using the Phoretix 1D Advanced and 1D Database (Nonlinear Dynamics) software for pattern-matching purposes.
Results The Protein 230 assay produced well-resolved protein profiles, suitable for varietal discrimination (figures 1 & 2). The Phoretix software was able to compare the electropherogram profiles. Figure 3 shows an example of a dendrogram where all the replicates of three different varieties are correctly grouped. A prototype wheat library was developed by selecting the most representative varietal profile. Results showed that 90 % of test samples could be identified within the top three matches. Work is in progress to optimize the performance of this library. The practicality of the method and the robustness of the system is bourne out by the fact that the system is now in routine use in UK commercial mill intake laboratories.
Conclusions
Our study has demonstrated that using the Agilent 2100 bioanalyzer with the Phoretix system offers a standardized, objective method for rapid varietal discrimination. The ease of use and total analysis time of less than 50 minutes makes it most suitable for mill intake use. The optimized system will enable millers to make more confident decisions in accepting grain consignments, and could become widely adopted as an effective policing tool within the grain industry. A number of UK mills have purchased the combined systems for screening wheat deliveries at intake.
Figure 3 Dendrogram illustrating pattern-matching of replicate profiles of 3 wheat varieties.
Acknowledgements CCFRA is grateful to the National Association of British and Irish Millers (nabim) for sponsoring this study. Dr Dhan Bhandari is Senior Scientist at Campden & Chorleywood Food Research Association (CCFRA), Chipping Campden, Gloucestershire, GL55 6LD, UK.
www.agilent.com/chem/2100
2008 Agilent Technologies Inc. Published February 1, 2008 Publication Number 5989-7735EN
Use of the Agilent 2100 Bioanalyzer for Basmati Rice Authenticity Testing
Application
Food
Author
Steve Garrett and Marie-Anne Clarke Molecular Biology Group Dept. Chemistry & Biochemistry Campden & Chorleywood Food Research Association Chipping Campden Gloucestershire GL55 6LD UK
to be supplying materials or products that are incorrectly labeled due to substitution or contamination. One food area which has been under the spotlight in recent years is the supply of basmati rice to the UK from India and Pakistan. In Europe, Commission Regulation 1549/04 grants a lower import tax on nine basmati varieties: Basmati 370, Dehradun (Type 3), Basmati 217, Taraori, Ranbir Basmati, Kernel, Basmati 386, Pusa Basmati, and Super Basmati. Other basmati rice varieties approved by India, Pakistan, and the UK include Basmati 198, Basmati 385, Haryana Basmati, Kasturi, Mahi Suganda, and Punjab Basmati. In a Code of Practice developed by Indian, Pakistani, and UK industry and enforcement organizations that came into effect for products packed and labeled after January 2006, the level of non-basmati rice in a basmati rice product must not exceed 7% (see www.riceassociation.org.uk). In order to check the supply of basmati rice, a DNA variety testing method using PCR amplification of eight rice microsatellite sequences has been developed for UK compliance (www.foodstandards. gov.uk/multimedia/pdfs/fsis4704basmati.pdf). During 2003, the UK Food Standards Agency carried out a surveillance exercise on basmati rice products using this method and revealed that 74% of them contained > 7% non-basmati varieties. This study evaluates the use of the Agilent 2100 bioanalyzer to differentiate approved and nonapproved varieties using three primer sets and to estimate the level of non-basmati using reference rice admixtures.
Abstract:
Ensuring integrity of raw food materials, ingredients, and products is both a product quality and regulatory compliance concern. Food suppliers and manufacturers may suffer economic and legal damages if proven to be supplying incorrectly labeled products. For example, EU Commission 1549/04 grants lower import tax on nine basmati rice varieties. A quick and cost-effective analytical method utilizing the Agilent 2100 bioanalyzer and DNA 1000 assay is shown as an alternative method to establish the authenticity of basmati rice products and to estimate the level of some varieties of non-basmati rice in ground rice products.
Introduction
The integrity of raw food materials, ingredients, and products must be maintained to ensure that they meet appropriate quality and legislative requirements. Food ingredient suppliers, manufacturers, and retailers can face legal action if proven
Experimental
Method All reference basmati rice samples were obtained from the UK Food Standards Agency via the University of Wales (Bangor, UK ). Other samples were from in-house basmati rice sample collections. PCR amplification was performed using either a PE9600 or PE2400 PCR machine (Applied Biosystems). Extraction DNA was extracted from ground rice grains using Qiagens DNeasy Plant Mini Kit or Promegas Maxwell 16 automated DNA extractor PCR DNA extracts were diluted 1 to 1 in sterile distilled water (SDW) to produce template DNA prior to use in PCRs. Amplification was performed in 25-L PCRs containing 1x Amplitaq Gold PCR buffer (Applied Biosystems), 60 nM of each primer, 200 nM dNTPs, 3 mM MgCl2, 0.05 U/L of Ampli-Taq Gold (Applied Biosystems), and 2 L of template DNA.
Marker RM201 RM212 RM339 Marker RM201 RM212 RM339 Forward Primer CTC gTT TAT TAC CTA CAg TAC C CCA CTT TCA gCT ACT ACC Ag GTA ATC gAT gCT gTg ggA Ag Reverse Primer CTA CCT CCT TTC TAg ACC gAT A CAC CCA TTT gTC TCT CAT TAT g gAg TCA TgT gAT AgC CgA TAT g
Table 1.
Analysis of Authenticated Basmati Rice Varieties Using Three Microsatellite Primer Sets List of microsatellite amplification product sizes obtained with primer set* RM201 RM212 RM339
Rice variety
Varieties listed in Commission Regulation 1549/04 Basmati 370 Dehra Dun (Type 3) Basmati 217 Ranbir Taraori Basmati 386 Kernel Pusa Super 162 (162) 162 162 162 162 162 162 162 162 134 (139) 134 (139) 134 134 134 (139) 134 (138) 134 134 (139) 134 (140) 200 (193) 200 (195) 200 200 200 (195) 200 (195) 200 200 (194) 204 (196)
Other varieties approved as basmati by UK Food Standards Agency Basmati 198 Basmati 385 Kasturi Haryana Basmati Mahi Sugandha Punjab Basmati 162 162 162 162 176 162 152 152 132 152 152 152 200 200 166 166 166 200
Non-approved varieties Basmati 2000 Shaheen Basmati Sherbati Mugad Sugandha Pak 386 Superfine Pusa Sugandha 162 162 178 (176) 178 178 178 162 162 152 152 130 (135) 132 130 (135) 132 132 134 204 200 166 (167) 166 166 (167) 166 178 200
Amplification profiles (95 C for 15 minutes [denaturation]; 50 cycles of: 95 C for 1 minute, 60 C for 1 minute [amplification]; 72 C for 10 minutes [final extension]) were used in all PCR reactions. DNA amplification was confirmed by separating PCR products using the Agilent 2100 bioanalyzer. Capillary Gel Electrophoresis on 2100 Bioanalyzer Reagents were prepared following manufacturers instructions. Batches (~500 L) of gel matrix (used to fill LabChip capillaries) were prepared as required or at 4 weekly intervals. All reagents were stored at 4 C and allowed to reach room temperature for 1 hour before use. PCR products (1 L)
Yamini
*These are from the FSA method developed by the University of Wales, Bangor. Actual size of fragments determined by the bioanalyzer using a DNA 500 chip kit are shown in brackets. The variation in bioanalyzer-determined fragment sizes can be about 5%. Shaded cells show how the varieties can be grouped using the three primer sets with analysis performed on the bioanalyzer.
were loaded directly onto prepared Series I DNA 500/DNA 1000 or Series II DNA 1000 labchips. All analysis was performed on the 2100 bioanalyzer, as per the manufacturers instructions.

The different-sized PCR products generated when these three primer sets are applied to the different varieties are easily resolved on the bioanalyzer using the DNA 500 or DNA 1000 chip. RM 212 primers produce a 139 bp product with varieties listed in Commission Regulation 1549/04 and a 154 bp product with FSA-approved varieties apart
from Kasturi and a few non-basmati varieties. The other primer sets enable separation of these varieties apart from the Yamini variety, which produces fragments similarly sized to the ECapproved varieties for all three primer sets. This variety is also difficult to distinguish from approved varieties using the standard method. The other FSA-approved basmati varieties could be distinguished from the EC-approved varieties but not from Basmati 2000 or Basmati 386 using these microsatellite primer sets. Use of further microsatellites that give PCR products that can be separated on the bioanalyzer will give improved differentiation of non-basmati rice varieties.
a) Primer set RM201
RM201 PCR with samples containing 0% to 40% (w/w) Pak 386 (nonbasmati) in basmati
0%
10%
20%
40%
b) Primer set RM339
RM339 PCR with samples containing 0% to 40% (w/w) Pak 386 (non-basmati) in basmati
0%
10%
20%
40%
? sample
Figure 1.
Analysis of non-basmati/basmati rice admixtures using two microsatellite primer sets. Estimation of percentage nonbasmati variety (Pak 386) in an unknown sample.
Table 2.
Experimental Summary Primer RM201 mean ratio 156 bp/170 bp PCR fragments (n = 3) 9.6 3.5 2.7 0.6 0.7 3.1 Primer RM339 mean ratio 199 bp/169 bp PCR fragments (n = 2) 0 3.1 1.1 0.4 0.3 2.1
% Pak 386 in basmati rice 0 10 20 40 50 Unknown sample estimated to contain between 1020% Pak 386
Rice Agilent Data
Basmati/non-basmati ratio
RM201 y = _0.0745x + 4.1482 R 2 = 0.9275

2
RM201 RM339
RM339 y = _0.0632x + 3.1246 R 2 = 0.804

0 0 10 20 30 % Basmati 40 50 60
Figure 2.
Standard curves based on the PCR product concentration ratios.
Both the RM201 and RM339 primer sets (Figure 1 and Table 2) could be used to produced separate and measurable PCR products with the PAK 386 basmati rice admixtures. The RM201 primers set gave a 170 bp product generated from the basmati rice and a 156 bp product from the Pak 386 variety, whereas RM339 primers gave 199 bp and 169 bp products. RM281 also gave other larger PCR products, which seemed to correlate with the presence of Pak 386 in the admixture. Standard curves (Figure 2) were produced using the PCR product concentration ratios. An unknown admixture sample was also analyzed and estimated to contain between 10 and 20% non-basmati rice in basmati. Results show that the bioanalyzer can be used to estimate the level of non-basmati rice (Pak 386) in a basmati rice which produces different sized microsatellite PCR products from the Pak 386.
estimate the level of some varieties of non-basmati rice. It should be feasible to develop further primer sets that would allow the bioanalyzer to be used to identify individual varieties. Simple manual and automated DNA extraction followed by fast PCR and post-PCR analysis on the bioanalyzer would allow rapid screening of rice materials prior to export from India and Pakistan and also allow enforcement bodies to efficiently test for microsatellite markers from non-approved basmati rice varieties in imported products.
Acknowledgements
Thanks to Mark Woolfe at UK Food Standards Agency and John Gorham and Katherine Steele at The University of Wales, Bangor, for the supply of primer set details and authentic basmati rice materials.
Conclusions
The Agilent 2100 bioanalyzer can be used as a quick and cost-effective alternative to establish the authenticity of basmati rice products and to

Agilent in Life Sciences
> Genomics
> Proteomics > Drug Discovery > Development > QA/QC
Agilent 2100 bioanalyzer

Application compendium

Application compendium
Agilent Technologies is a leading provider of life science and chemical analysis solutions. We offer systems for the acquisition and interpretation of genetic and chemical information from sample handling, to analysis, data management and reporting. Let us provide you with the right solution for your success.
Whatever your life science needs, Agilent Technologies has the solution. We are committed to providing you with superior technology designed for maximum productivity and cost efficiency.
The Agilent 2100 bioanalyzer is a unique analysis tool capable of handling nucleic acids, proteins and cells on one platform. When combined with any one of our kits, you will discover how, lab-on-a-chip technology can revolutionize your laboratory. One of the many benefits the Agilent 2100 bioanalyzer has over conventional bioanalytical methods is the elimination of time consuming procedures you enjoy standardized handling and interpretation of data. And as our portfolio of microfluidics continues to expand, you will benefit from the wide range of applications to which the technology can be used. Lab-on-a-Chip technology simplifies the process of data gathering and analysis down to three quick and easy steps:
Load sample, run analysis, view data.

The Agilent 2100 bioanalyzer utilizes micro-fabricated chips with up to 12-wells requiring minimum sample consumption, in the low l-range. Prepackaged reagents included with the chip kits help to speed up the entire process.
The Agilent 2100 bioanalyzer with lab-on-a-chip technology will increase the efficiency of your analysis and the productivity of your day.
With it's single, compact system architecture the Agilent 2100 bioanalyzer integrates sample handling, separation, detection and data analysis, all in the name of speed. Eliminate potential mistakes that can occur while interpreting and storing data. The Agilent 2100 bioanalyzer automatically incorporates steps some researchers might otherwise ignore in the interest of time.
The Agilent 2100 bioanalyzer helps to optimize PCR reactions for gene expression, sequencing, cloning and typing. When used in conjunction with the DNA kits, it provides higher sensitivity, improved sizing accuracy and automated, reproducible quantitation, compared with regular slab gel electrophoresis, which is crucial for RT-PCR and any type of multiplex PCR. RNA 6000 Pico kit: catch RNA degradation with sample amounts as low as 200 pg of total RNA and automatically detect ribosomal RNA contamination in mRNA. The RNA 6000 Nano kit is the industry standard for sample QC in the context of microarray analysis. With the introduction of RIN (RNA Integrity Number), each RNA electropherogram receives a RIN assigned by the software. The RIN unambiguously assesses RNA quality in terms of degradation. DNA and RNA LabChip kits enable you to check the quality of probes and targets in your microarray gene expression analysis. Agilent also provides the full solution for gene expression analysis with its high performance microarray scanner and the suite of off-the-shelf microarrays. The Protein 200 Plus LabChip kit is a fast and reliable assay capable of quantifying and sizing a multitude of different protein samples. Used with the Agilent 2100 bioanalyzer it can analyze ten, 4 l samples in less than 30 minutes. Agilent offers an add-on pressure cartridge, cell fluorescence software and Cell LabChip kit for multiple types of cell assay applications. Combined with the Agilent 2100 bioanalyzer, this makes performing simple flow cytometric analyses a reality, even for the smallest lab. Make your Agilent 2100 bioanalyzer system compliant! The Agilent 2100 bioanalyzer security pack software ensures full 21 CFR part 11 compliance of your system. Along with IQ and OQ/PV services offered for all assays of our LabChip kits, your Agilent 2100 bioanalyzer system will be compliant in no time.
Contents
Page
I. Cell fluorescence analysis

Protein expression monitoring Cell surface antibody staining - CD4 in CCRF CEM T-cells Cell surface antibody staining - CD3 in T-cell leukemia CD3 expression in T-cell leukemia via on-chip staining Intracellular glucocorticoide receptor (GR) antibody staining in H4 hepatocytes Analyzing a limited number of cells Baculovirus titre determination Upregulated gene expression in primary cells Transfection efficiency monitoring Green fluorescent protein in CHO cells On-chip staining of GFP expression for optimizing transfection conditions with different DNA:lipid ratios Verification of stable transfected cell clones by on-chip antibody staining Transfection of primary cells Apoptosis detection Detection of phosphatidylserine on the cell surface via Annexin V binding Intracellular Caspase-3 antibody staining assay Fast Annexin protocol for time course of apoptosis induction via anti-FAS antibody Apoptosis detection in primary cells Gene silencing in cell culture siRNA transfection optimization Monitoring of gene silencing experiments
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19-20 21 22 23
Contents
Page
II. DNA analysis

Restriction digest analysis Sizing range exemplified by the separation of Adenovirus 2/Dra I Detection of single base mutations PCR product analysis Separation of 3 different mixtures of PCR products Determination of PCR product impurity Multiplex PCR analysis of bacteria in chicken Multiplex PCR with 19 products Gene expression analysis mRNA expression study by comparative multiplex PCR Standardized end-point RT-PCR Co-amplification of GAPDH and hsp72 Co-amplification of GAPDH and hsp72 - response curves Competitive PCR Food analysis Development of meat specific assays Fish species identification by RFLP GMO detection Development of a multiplex assay for soya DNA stability during food processing GMO detection by nested multiplex PCR Oncology Tumor cell detection from carcinoma patient blood SNP analysis in cancer related P16 gene K-ras gene SNP detection METH-2 downregulation in lung carcinomas Label-free analysis of microsatellite instability in carcinoma Clinical research Genotyping of H. pylori Duplications and deletions in genomic DNA Forensic testing Optimization of PCR on mtDNA Pitfalls in mtDNA sequencing
24 25 26-27 28 29-30 31 32 33 34 35 36 37 38-39 40 41 42 43 44 45 46 47 48 49 50 51 52
Contents
Page
III. RNA analysis

Analysis of total RNA RNA integrity Standardization of RNA Quality Control Reproducibility of quantitation Genomic DNA contamination Low amounts of total RNA Detection of low levels of RNA RNA integrity with the RNA 6000 Pico kit RNA quality after staining and microdissection Analysis of minimum RNA amounts Genomic DNA in low concentrated RNA extracts Low RNA amounts from kidney sections Analysis of mRNA RNA integrity Ribosomal RNA contamination in mRNA samples Analysis of Cy5-labeled samples Analysis of cRNA with and without dye in gel matrix Optimization of labeling reactions cRNA fragmentation Analysis of T7-RNA transcripts Size estimation
53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69
Contents
Page
IV. Protein analysis

Protein expression Analysis of cell lysates - protein induction Protein purification: Comparison between lysate and flow through Analysis of protein purification GFP Streptag fusion protein purification Analysis of column capacity Analysis of column fractions to optimize conditions His-tag protein purification with Ni++ ZipTips Enzymatic removal of His Tags from recombinant proteins Complementing RP-HPLC protein purification Antibody Analysis Analysis of antibodies under reducing and non-reducing conditions Quantitation of the half-antibody content in IgG4 preparations Comparison of SDS-PAGE, CGE and Agilent 2100 bioanalyzer for humanized monoclonal antibody analysis Absolute quantitation of IgG Quality control of stressed antibodies Separation of bispecific antibodies chains Food analysis Bovine milk analysis Protein pattern of different transgenic seedlines Protein - others Absolute protein quantitation Glycoprotein sizing Protein quality control prior to MS-analysis Depletion of high abundant proteins from blood samples Increased sensitivity by desalting protein samples
70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93
V. Literature
Drug Discovery/Drug Development
Forensics/Homeland Security
Drug Manufacturing QA/QC
I. Cell fluorescence analysis

Agriculture/Food
Pharmaceuticals
Proteomics
Genomics
Protein expression monitoring Cell surface antibody staining - CD4 in CCRF CEM T-cells Cell surface antibody staining - CD3 in T-cell leukemia CD3 expression in T-cell leukemia via on-chip staining Intracellular glucocorticoide receptor (GR) antibody staining in H4 hepatocytes Analyzing a limited number of cells Baculovirus titre determination Upregulated gene expression in primary cells Transfection efficiency monitoring Green fluorescent protein in CHO cells On-chip staining of GFP expression for optimizing transfection conditions with different DNA:lipid ratios Verification of stable transfected cell clones by on-chip antibody staining Transfection of primary cells Apoptosis detection Detection of phosphatidylserine on the cell surface via Annexin V binding Intracellular Caspase-3 antibody staining assay Fast Annexin protocol for time course of apoptosis induction via anti-FAS antibody Apoptosis detection in primary cells Gene silencing in cell culture siRNA transfection optimization Monitoring of gene silencing experiments
Protein expression monitoring

Cell surface antibody staining - CD4 in CCRF CEM T-cells
Flow cytometer (10,000 Events) Agilent 2100 bioanalyzer (500 Events)
Kit: Cell fluorescence kit Assay: Antibody staining assay Application: CCRF-CEM cells were stained with hCD4-APC labeled antibodies and calcein live dye. 65% of all CCRF-CEM live cells (yellow curve) are expressing CD4 protein which is good in comparison to conventional flow cytometer results.
Corresponding application note: 5988-4322EN

Cell surface antibody staining - CD3 in T-cell leukemia
Averaged data per instrument
2100-1 60.9 34.4 17.3 8.9 5.1 0.8
2100-2 67.8 36.7 17.6 9.4 4.4 0.6
Mean % CD3+ cells 2100-3 2100-4 66.6 65.0 36.7 34.3 18.7 17.2 9.9 8.3 5.3 4.9 0.3 0.3
Flow cyt. 60.9 29.8 13.8 6.5 3.2 0.0
Kit: Cell fluorescence kit Assay: Antibody staining assay Application: Jurkat (T-cell leukemia) cells were stained with calcein alone or with calcein and APC-labeled anti-CD3 antibody. To mimic different subpopulation sizes, mixtures of both populations were prepared at various ratios. Samples were analyzed with 4 Agilent 2100 bioanalyzer instruments on 5 chips and compared to a flow cytometer reference instrument. Interestingly, small subpopulations (like 10 - 20%) could be analyzed with good accuracy and reproducibility.

CD3 expression in T-cell leukemia via on-chip staining
A. On-chip Agilent 2100 bioanalyzer
B. Conventional flow cytometry
Kit: Cell fluorescence kit Assay: Antibody staining assay Application: Jurkat cells were stained on-chip with anti hCD3-APC prediluted 1:5.5 in cell buffer and Calcein (1:50 in cell buffer). After an incubation time of 25 minutes in the chip, samples were measured in the Agilent 2100 bioanalyzer. The faster and easier on-chip staining procedure has the advantage here of reducing cell consumption 17 fold and antibody reagent costs 80 fold. A) Overlay of representative histograms of calcein and antibody treated cells. B) Comparison between on-chip staining data and data obtained by measuring cells stained by conventional staining on a flow cytometer. Corresponding application note: 5988-7111EN

Intracellular glucocorticoide receptor (GR) antibody staining in H4 hepatocytes
Chip histogram overlay from 700 cells/sample
Correlation of chip vs. flow cytometer results
Kit: Cell fluorescence kit Assay: Generic assay Application: H4 hepatocytes cells were stained with SYTO16 DNA dye alone or with SYTO16 and GR primary antibody. After washing, both cell preparations were stained with APC-labeled secondary antibody. Mixtures of both populations were prepared at various ratios. The insert in the left picture shows the overlay of all six cell samples in the blue reference color. The black histogram represents data from the control sample, no GR detected. All other 5 samples have significant staining above marked fluorescence intensity in the red. Good chip to chip reproducibility and comparison to flow cytometer is demonstrated.

Analyzing a limited number of cells
Cells 20,000 10,000 5,000 2,500 1,250 625
Live-CD3+ 83.7% 85.6% 87.7% 84.0% 89.8% 90.0%
STD(n=4) 3.5% 4.1% 4.2% 3.0% 6.5% 9.3%
Kit: Cell fluorescence kit Assay: On-chip antibody staining assay Application: The direct comparison of different input numbers of cells (down to 625 cells in 10 l) for the on-chip staining protocol reveals that even with a much lower number than the recommended 20000 cells/10 l for the standard protocol reliable and meaningful results can be achieved with good reproducibility. The data shown were generated with CD3-positive Jurkat cells stained with an anti-CD3 antibody for the CD3 protein and counterstained with the live cell stain Calcein AM. Similar results were obtained with primary human dermal fibroblasts (PHDF) indicating the usefulness of this method for scarce specimen. The lack of sensitivity, automation and convenient quantitation found with other methods can be circumvented easily by using the Agilent 2100 bioanalyzer. Corresponding application note: 5989-0746EN
10

Baculovirus titre determination
Fluorescent Light Transmitted Light
104 103 102 Red Fluorescence 101
Infected Sf21 cells
100 10-1
10-2
10-1
100 101 102 Blue Fluorescence
103
Kit: Cell fluorescence kit Assay: GFP Assay Application: A fast and convenient method exists for the calculation of baculovirus titre for expression systems facilitating insect cells. Using GFP-linked co-expression plasmids, the Agilent 2100 bioanalyzer and the flow cytometry set allows the calculation of the viral titre for six samples in approximately 90 minutes. It is superior to traditional plaque assays in terms of labor time, automation and user-to-user variability.

Upregulated gene expression in primary cells
A
B Kit: Cell fluorescence kit Assay: On-chip antibody staining assay Application: Flow cytometric analysis of primary cells can present a challenge for researchers due to limited availability and life span of primary cells. A dose-respondent upregulation of protein expression in primary cells using only a minimum number of cells in a fast on-chip-staining approach is shown here. Activation of peripheral blood lymphocytes by phorbol-12-myristate-13-acetate (PMA) leads to increased expression of the T cell receptor CD3 (Figure A, mean from 3 experiments). For HUVECs (human umbilical vein endothelia cells) the induction of E-selectin (CD62E) expression upon IL-1 treatment is shown (Figure B, white bars) in comparison to results from a conventional flow cytometer (white bars). Corresponding application note: 5989-2718EN
11
12
Transfection efficiency monitoring

Green fluorescent protein in CHO cells
Mock transfected cells
GFP transfected cells
Kit: Cell fluorescence kit Assay: GFP assay Application: Chinese hamster ovary (CHO-K1) cells were transfected with EGFP DNA by a lipofection method. The upper panel shows the control mock transfection; here cells don't express GFP. Examples for data evaluation in dotplot view and histogram view are shown in comparison to the microscopy view. For analysis on the Agilent 2100 bioanalyzer, cells were stained with a red dye for live cells (reference stain). The transfection efficiency of 56% can be easily determined with the Agilent 2100 bioanalyzer. Corresponding application note: 5988-4320EN

On-chip staining of GFP expression for optimizing transfection conditions with different DNA:lipid ratios
Control
1:2
1:4
1:6
1:8
1:10
Kit: Cell fluorescence kit Assay: On-chip GFP assay Application: Chinese hamster ovary (CHO-K1) cells were transfected with EGFP DNA by alipofection method. Optimization of transfection conditions were done on one chip. Several DNA:lipofectamine ratios were tried. A ratio of 1:8 gave the best transfection efficiency. All cells were reference stained with a red live dye. On-chip staining was applied, minimizing the staining time, reagent usage and cell consumption.
13
14

Verification of stable transfected cell clones by on-chip antibody staining
Hek 293 control cells
CD 80 stable clone
Kit: Cell fluorescence kit Assay: On-chip antibody staining assay Application: Verification of CD80 protein expression in stable transfected Hek 293 cells with the Agilent 2100 bioanalyzer. Control (left dot plot) and CD80 transfected cells (right) are stained on-chip with blue calcein live dye and anti-CD80-CyChrome antibody. Red region marks CD80 protein expressing 293 cells within live cell population - confirming expression in the CD80 stable clone Hek 293 cells.

Transfection of primary cells
Control recovery 100%
DNA:LP ratio 1:2 recovery 64%
Kit: Cell fluorescence kit Assay: GFP assay Application: Monitoring the transfection efficiency in primary cells requires low cell consumption, high reproducibility of results, a fast on-chip staining procedure and ease-of-use all provided by the Agilent 2100 bioanalyzer. The transfection efficiency using a GFP-coding plasmid (pEGFP-C2) at varying plasmid:lipofectamine ratios (DNA:LP ratio) obtained with human umbilical vein endothelial cells (HUVEC) is measured in this optimization series. Images from a fluorescence microscope (A) and dot plots (B), as well as histograms (C) of control- and GFP-transfected cells are shown. Using increasing ratios, better transfection efficiency was achieved, whereas the toxicity of LP caused decreased recovery of living cells. Such data facilitates optimizing transfection conditions. Corresponding application note: 5988-8154EN
15
16
Apoptosis detection
Detection of phosphatidylserine on the cell surface via Annexin V binding
Agilent 2100 bioanalyzer histogram: blue channel (Calcein) Agilent 2100 bioanalyzer histogram: red channel (Annexin-Cy5) Agilent 2100 bioanalyzer dot plot: events from both channels
16h treated sample Subpopulation of all live cells which are apoptotic Kit: Cell Fluorescence kit Assay: Apoptosis assay Application: Apoptosis (programmed cell death) in Jurkat cells was induced with camptothecin. Cells treated for 16 hours and untreated cells were stained with calcein and Annexin-Cy5. Annexin-V binds to phosphatidylserine - a membrane lipid which is kept to the inner leaflet of the cell membrane of intact cells. Exposure of phopshatidylserine on the outer leaflet is an early indicator of apoptotic processes. Annexin-V binding is made detectable by Cy5 staining of the Annexin-V via a biotin-streptavidin interaction. Calcein staining of cells is used as a live control to distinguish living and apoptotic cells from dead cells. Calcein enters the cell via the membrane as a non-fluorescent ester. The ester is cleaved inside the cell which results in fluorescence. The histograms on the left show the number and intensity value of all events which generated a signal in the blue channel, corresponding to calcein-stained cells. The histograms on the right shows all events which generated a signal in the red channel, corresponding to Annexin-V binding to apoptotic cells. While the control shows only low intensity values (background noise), the treated sample shows high intensity values (within the red markers) corresponding to apoptotic cells. The dot plot of the treated sample nicely shows the subpopulation of all live cells which are apoptotic. Corresponding application note: 5988-4319EN
Apoptosis detection
Intracellular Caspase-3 antibody staining assay
Kit: Cell fluorescence kit Assay: Generic assay Application: Induction of apoptosis in Jurkat cells was done with anti-FAS antibody treatment. Intracellular staining with specific antibodies against 'active' Caspase-3 were performed. Reference staining was done with SYTO16 DNA dye. Good chip to chip reproducibility and good comparison to conventional flow cytometer results were obtained.
17
18
Apoptosis detection
Fast Annexin protocol for time course of apoptosis induction via anti-FAS antibody
Kit: Cell fluorescence kit Assay: Apoptosis assay Application: Apoptosis (programmed cell death) in Jurkat cells was induced with anti-FAS antibody. Cells treated for 0,1,2,3,4 and 6 hours were stained with calcein and Annexin-Cy5. Annexin-V binds to phosphatidylserine - a membrane lipid which is kept to the inner leaflet of the cell membrane of intact cells. Exposure of phopshatidylserine on the outer leaflet is an early indicator of apoptotic processes. Annexin V binding is detectable by Cy5 staining of the Annexin-V via a biotin-streptavidin interaction. Calcein staining of cells is used as a live control to distinguish living and apoptotic cells from dead cells. Calcein enters the cell via the membrane as non-fluorescent ester. The ester is cleaved inside the cell which results in fluorescence and indicates apoptosis.
The histograms on page 18 (A) show the number and intensity value of all events which generated a signal in the blue channel, corresponding to calcein-stained cells. The histograms on the right show all events which generated a signal in the red channel, corresponding to Annexin-V binding to apoptotic cells. While the control shows only low intensity values (background noise), the treated sample shows high intensity values (within the red markers) corresponding to apoptotic cells. (B) Time course of the induction of apoptosis by anti-FAS antibody in Jurkat cells. Apoptosis is detectable in a significant amount of cells after 2 hours. Following a treatment of 4 hours, approximately 95% of the cells are apoptotic. Corresponding application note: 5988-4319EN
19
20
Apoptosis detection
Apoptosis detection in primary cells
Kit: Cell fluorescence kit Assay: On-chip antibody staining assay Application: The Agilent 2100 bioanalyzer has been used to study induced apoptosis by monitoring annexin V-binding in primary human endothelial cells (HUVEC, not shown) and human dermal fibroblasts (NHDF, shown). A simple and fast assay protocol was used on cells left untreated or treated for 5 hours with different concentrations of staurosporine, which induces apoptosis. See row A for dot blots and B for histograms at different concentrations. Evaluation of the same samples on a conventional flow cytometer (row C) yielded similar results.
Gene silencing in cell culture

siRNA transfection optimization
A
Transfection Viability (TV) Viability in Transfected cells (ViT) Transfection Efficiency (TE)
Kit: Cell fluorescence kit Assay: On-chip antibody staining assay Application: In gene silencing experiments (HeLa cells) we found that increasing amounts of transfection reagent (TransMessengerTM)to a constant amount of siRNA leads to a plateau of transfection viability (panel A). Transfection viability reflects the product of the viability of the transfected cells and the transfection efficiency. With a constant siRNA/transfection reagent ratio of 1:4 and increasing total amounts of introduced siRNA (panel B) the viability of transfected cells decreases at a certain point although the transfection efficiency increases. Thus, there are experimental conditions where the number of living and transfected cells are at a maximum. The Agilent 2100 bioanalyzer features on-chip staining and leads to excellent results with a minimal consumption of cells and reagents. Corresponding application note: 5988-9872EN
21
22
Gene silencing in cell culture

Monitoring of gene silencing experiments
Kit: Cell fluorescence kit Assay: GFP Assay Application: After co-transfection of a GFP plasmid and Cy5-labeled siRNA (GFP-specific), GFP expression and viability of cells were detected. The course of GFP expression in control (GFP only) and siRNA/GFP transfected cells was measured on the Agilent 2100 bioanalyzer. Accurate results were obtained fast and in an automated manner. They easily allow the efficiency and reliability of a given protocol and transfection reagents to be judged. Thus, such an experiment provides efficient monitoring and optimization of any gene silencing experiment.
II. DNA analysis

Agriculture/Food
Pharmaceuticals
Proteomics
Genomics
Restriction digest analysis Sizing range exemplified by the separation of Adenovirus 2/Dra I Detection of single base mutations (I) Detection of single base mutations (II) PCR product analysis Separation of 3 different mixtures of PCR products Determination of PCR product impurity Multiplex PCR analysis of bacteria in chicken Multiplex PCR with 19 products Gene expression analysis mRNA expression study by comparative multiplex PCR Standardized end-point RT-PCR Co-amplification of GAPDH and hsp72 Co-amplification of GAPDH and hsp72 - response curves Competitive PCR Food analysis Development of meat specific assays (I) Development of meat specific assays (II) Fish species identification by RFLP GMO detection Development of a multiplex assay for soya DNA stability during food processing GMO detection by nested multiplex PCR Oncology Tumor cell detection from carcinoma patient blood SNP analysis in cancer related P16 gene K-ras gene SNP detection METH-2 downregulation in lung carcinomas Label-free analysis of microsatellite instability in carcinoma Diagnostic research Genotyping of H. pylori Duplications and deletions in genomic DNA Forensic testing Optimization of PCR on mtDNA Pitfalls in mtDNA sequencing
23
24
Restriction digest analysis

Sizing range exemplified by the separation of Adenovirus 2/Dra I
Kit: DNA 12000 kit Assay: DNA 12000 assay Application: Restriction digest analysis of Adenovirus 2/Dra I. For restriction fragment analysis the large linear dynamic range of the lab-on-a-chip approach is very advantageous. Analyzing samples with large and short fragments on slab gels can be difficult because of bands running off the gel and insufficient staining (or over-staining) of bands.

Detection of single base mutations (I)
Amplify exons 7 and 8 (resulting products: 618 bp fragment and 200 bp fragment) Digest with Hpa II In each example one of the restriction sites can be deleted by a point mutation
Analyze using Agilent 2100 bioanalyzer and 4-20 acrylamide gel
Kit: DNA 7500 kit Assay: DNA 7500 assay Application: Mutation detection by RFLP highlights the use of the Agilent 2100 bioanalyzer. Two different regions of the p53 gene were amplified with specific primers and digested with Hpa II, which cuts in a location that is prone to mutations. In the presence of a point mutation, the enzyme Hpa II does not cleave the DNA, leaving larger fragments that can be revealed by gel electrophoresis or by analysis with the DNA 7500 LabChip kit (see next page).
Corresponding application note: data not published
25
26

Detection of single base mutations (II)
Kit: DNA 7500 kit Assay: DNA 7500 assay Application: Analysis on the chip showed an identical pattern of digest fragments as seen on the slab gel for the wildtype and Exon 7 & 8 PCR products. Comparison of the calculated sizes of the bands shows 1-2% variance with the LabChip assay, which allows fast and accurate detection of point mutations.
PCR product analysis

Separation of 3 different mixtures of PCR products
Kit: DNA 500 kit Assay: DNA 500 assay Application: Overlay of three different electropherograms, which are mixtures of PCR samples ranging from 25 to 500 base pairs in size. The two closest eluting bands (50 bp and 53 bp) are partially separated and identified by the software as two separate peaks. The DNA 500 assay achieves a resolution of five base pairs from 25 to 100 base pairs and a 5% resolution from 100 to 500 base pairs where the sizing error is less than 10% over the entire size range.
27
28

Determination of PCR product impurity
Kit: DNA 7500 kit Assay: DNA 7500 assay Application: Comparison between the analysis of two PCR reactions (300 and 3000 bp products) using the DNA 7500 LabChip kit vs. an agarose gel. Two different concentrations are shown side by side for each PCR reaction (undiluted and 1:4 dilution). The Agilent 2100 bioanalyzer shows superior performance in locating impurities over a broader concentration range than the gel. The 300 bp fragment appears to be uncontaminated in both the gel and on the Agilent 2100 bioanalyzer. The 3000 bp fragment shows few impurities on the gel, which become invisible at the 1:4 dilution. These impurities can easily be detected with the Agilent 2100 bioanalyzer. Corresponding application note: 5968-7496EN

Determination of PCR product impurity
Sample
c (DNA) Main peak All peaks (300bp) 41.4 ng/l 40.7 ng/l 9.6 ng/l 9.6 ng/l
Pure PCR 1:4 dilution
Sample
c (DNA) Main peak All peaks (3000bp) 61.9 ng/l 40.7 ng/l 14.8 ng/l 9.8 ng/l
Pure PCR 1:4 dilution
Kit: DNA 7500 kit Assay: DNA 7500 assay Application: The quantitative data generated by the Agilent 2100 bioanalyzer indicate the amount of impurity or non-specific products in the PCR reactions from the previous page. Even in the 300 bp fragment a small impurity can be detected, while the 3000 bp fragment shows more than 50% impurities.
29
30

Multiplex PCR analysis of bacteria in chicken
Upper marker
E.coli C.jejuni Salmonella C.coli
Lower marker Data kindly provided by GenPoint, NL
Kit: DNA 500 kit Assay: DNA 500 assay Application: Multiplex PCR with four primer pairs, each one specific for a certain DNA sequence from one of the 4 bacteria to be tested for. Total DNA was extracted from chicken and subjected to PCR. The gel-like image shows traces from different chicken samples with bands showing up when an amplicon could be detected. The electropherogram is one example where bacterial DNA from two species of the Campylobacter genus could be detected.

Multiplex PCR with 19 products
Data kindly provided by Qiagen, Germany Kit: DNA 7500 kit Assay: DNA 7500 assay Application: Many molecular applications include PCR multiplexing as shown above with a PCR that yields 19 products. Applications are genotyping of transgenic organisms, detection of pathogens or GMs and microsatellite genotyping (e.g. short tandem repeat (STR) and variable number tandem repeat (VNTR) analyses). The sample shows optimization of PCR conditions (Mg2+ concentration) performed to ensure annealing of the multiple primers under identical conditions. Visualization and evaluation of the results can be performed efficiently with the Agilent 2100 bioanalyzer because of the high resolution, the accurate sizing, quantitation and extended linear range. Corresponding application note: 5988-9342EN
31
32
Gene expression analysis

mRNA expression study by comparative multiplex PCR
Data kindly provided by the Roy Castle Centre Kit: DNA 1000 kit Assay: DNA 1000 assay Application: Two genes were co-amplified in this study. A tumor specific gene (upper band) along with a housekeeping gene (lower band). The upregulation of the tumor gene is visualized via analysis on the Agilent 2100 bioanalyzer. Building the ratio of the concentration values obtained from the Agilent 2100 bioanalyzer, numerical values are obtained that are normalized with regard to the RT-PCR amplification efficiency. This way tumor tissue can be distinguished from normal tissue more unambiguously. Corresponding application note: data not published

Standardized end-point RT-PCR
10
11
12
Data kindly provided by the Medical College of Ohio
Kit: DNA 7500 kit Assay: DNA 7500 assay Application: Complementary DNA from bronchial epithelial cells (BEC) was analyzed by a Standardized RT-PCR (StaRT) for the expression of 15 different genes. This analysis can be performed at the end-point of PCR without the need for real-time measurement at each cycle of PCR. Three methods for evaluation of representative results were compared (see above). The coefficient of variance (CV) from at least 3 measurements was calculated. The direct comparison of the reproducibility for agarose gel analysis (A, CV = 0.50) and the ABI Prism310 Genetic Analyzer (C, CV = 0.39) with the Agilent 2100 bioanalyzer (B, CV = 0.29) reveals that the Agilent 2100 bioanalyzer is superior. It is a reliable and valuable tool in quantitative gene expression analysis. Corresponding application note: 5988-3674 EN
33
34

Co-amplification of GAPDH and hsp72
Data kindly provided by Dr. Eric Gottwald, Forschungszentrum Karlsruhe, Germany
Kit: DNA 1000 kit Assay: DNA 1000 assay Application: Gel-like image and electropherograms showing the results of separate amplifications and co-amplifications of GAPDH and hsp72 in unstimulated HepG2 cells. Primers for GAPDH yield a PCR product of 443 bp (lane 1), primers for hsp72 yield PCR products of 384 and 650 bp (lane 2 and 3). Lane 4 and 5 show the results of the co-amplification reactions. Due to the competitiveness of the reaction, very little hsp72 products could be detected in lane 4 (insert) and no product was detected in lane 5 (lane 6 = negative control). The broad linear dynamic range of the analysis allows detection of weak bands next to strong bands and helped in the determination of gene expression in this case.

Co-amplification of GAPDH and hsp72 - response curves
Kit: DNA 1000 kit Assay: DNA 1000 assay Application: The optimized PCR conditions were used to monitor the response of a stimulus to hsp. Gene expression was monitored by comparing the RT-PCR amplification of a housekeeping gene with the co-amplification of hsp. In the current case, the highest gene expression was measured after about 10 minutes. As a comparison, the same set of samples was analyzed using the DNA 500 kit. Virtually identical results are obtained with both kits, demonstrating thatlab-on-a-chip technology can serve as a standardized approach to gel electrophoresis.
35
36

Competitive PCR
Kit: DNA 1000 kit Assay: DNA 1000 assay Application: Two genes were reverse transcribed and co-amplified in one reaction tube. The PCR products were analyzed using the DNA 1000 LabChip kit. Primers for hsp72 were present from the beginning of the PCR reactions, while primers for GAPDH were added after various cycle numbers ranging from 20 to 40 cycles (primer dropping method). This allowed optimization of this competitive PCR reaction. The left graph displays the dynamic range (arrow) in the gel like view, whereas the right graph indicates conditions with greatest sensitivity (red line). Corresponding application note: 5988-4556EN
Food analysis
Development of meat specific assays (I)
Data kindly provided by CCFRA, UK
Kit: DNA 500 kit Assay: DNA 500 assay Application: For detection of individual species in processed food, PCR assays with specific sets of primers can be developed. Example: turkey specific primers do not amplify any other meat species, including beef, chicken, lamb, or pork (see lane 5 and respective electropherogram).
37
38
Food analysis
Development of meat specific assays (II)
Kit: DNA 500 kit Assay: DNA 500 assay Application: For detection of individual component types in processed food, PCR assays with specific sets of primers can be developed. Example: Primers that amplify any type of meat, but do not amplify other food constituents, including soya, maize, wheat or fish.
Food analysis
Fish species identification by RFLP
Common name (UK) Atlantic Cod Pacific Cod Coley (Saithe) Haddock European Hake South African Hake European Plaice Whiting Alaskan (Walleye) Pollock Hoki Atlantic Salmon Red / Sockeye Salmon Pink / Humpback Salmon Chinook Salmon Coho / Silver Salmon Keta / Chum Salmon Cut-throat Trout Dolly Varden Cherry Salmon
Latin name Gadus morhua Gadus macrocephalus Pollachius virens Melanogrammus aeglefinus Merluccius merluccius Merluccius paradoxus Pleuronectes platessa Merlangus merlangus Theragra chalcogramma Macruronus novaezelandiae Salmo salar Oncorhynchus nerka Oncorhynchus gorbuscha Oncorhynchus tschawytscha Oncorhynchus kisutch Oncorhynchus keta Oncorhynchus clarki clarki Salvelinus malma malma Oncorhynchus masou masou
Restriction enzyme: DdeI
Restriction enzyme: HaeIII
Restriction enzyme: NlaIII Kit: DNA 500 kit Assay: DNA 500 assay Application: Identification of white fish and salmon species in the processed state presents a challenge. However, evaluation of PCR-RFLP profiles (PCR-restriction fragment length polymorphism) of a 464 bp region from the cytochrom b gene cut separately with three restriction enzymes facilitated the differentiation of 19 commercially important species. Analysis of the restriction digests was performed with the Agilent 2100 bioanalyzer. This approach was successfully tested in an interlaboratory study. Corresponding application note: 5989-2982EN
39
40
GMO detection
Development of a multiplex assay for soya
Kit: DNA 500 kit Assay: DNA 500 assay Application: Multiplex assay for genetically modified (GM) soya. The aim was to develop a model assay that could be used to assess the quality of DNA extracted from heat-processed soya flour samples, in particular, to investigate differences in PCR amplification between small DNA targets. A single multiplex PCR assay was developed that enabled three GM soya targets and one control to be analyzed in a single reaction mix. Primer concentration was optimized in order to obtain four PCR products resolved by gel electrophoresis which corresponded in size to the soya lectin gene target of 80 bp, and the EPSPS (5-enolpyruvyl-shikamate- 3-phosphate synthase) gene targets of 117 bp, 150 bp and 202 bp respectively. These latter targets are only found in Roundup Ready GM soya. Figure A: Peaks produced by the four PCR products when analyzed with the Agilent 2100 bioanalyzer and DNA 500 LabChip kit. Figure B: Analysis of certified reference materials containing known amounts of GM soya. Corresponding application note: 5988-4070EN
GMO detection
DNA stability during food processing
Time at 100C and pH 3.3 (min) 80 bp 0 3 6 9 12 15 18 21 100 74 57 36 67 48 0 0
Amount of PCR product* 118 bp 100 77 58 23 33 27 0 0 150 bp 100 73 21 24 47 16 0 0 202 bp 100 67 6 15 21 0 0 0
* % product determined relative to the amount at 0 minutes Data kindly provided by CCFRA, UK
Kit: DNA 500 kit Assay: DNA 500 assay Application: The multiplex PCR assay was applied to soya flour samples containing approx. 1.3 % GM soya and boiled at either pH 3.3, 4.3 or 6.7 for up to 21 minutes. For accurate determination of the quantity of each PCR product, the samples were applied to the DNA 500 LabChip. The concentration of each PCR product was calculated using the Agilent 2100 bioanalyzer software. At pH 3.3 where an effect of heating time was observed, the amount of each PCR product at each time point was compared to the amount of each product at 0 minutes (Table 2). At pH 3.3, the relative amount of the 80 bp product was reduced to 48 % after 15 minutes and no product was detected at 18 or 21 minutes. After 15 minutes, the relative amounts of products of 118 bp and 150 bp were reduced to 27 % and 16 % respectively and the 202 bp product was not detected. None of the products were detected after 18 or 21 minutes.
41
42
GMO detection
GMO detection by nested multiplex PCR
1 2 3 4 5 6 7
Low molecular weight marker Primer-dimer in negative control Soy lectin - 118 bp 35S GMO - 153 bp Internal control - 217 bp Corn zein - 278 bp High molecular weight marker
Kit: DNA 1000 kit Assay: DNA 1000 assay Application: GMO detection by multiplex PCR is widely used for soy and corn. Often sequences from the transgene and species specific controls or internal standard are co-amplified by endpoint PCR in a screening procedure. Multiple products can be analyzed with the Agilent 2100 bioanalyzer at high resolution and sensitivity. Quantification and comparison of product amounts may already lead to qualification of a positive screening result prior to analysis by expensive quantitative real time PCR. Corresponding application note: 5989-0124EN
Oncology
Tumor cell detection from carcinoma patient blood
Data kindly provided by AdnaGen
Kit: DNA 500 kit Assay: DNA 500 assay Application: A combined method of specific tumor cell enrichment and a high sensitivity tumor cell detection by multiplex PCR allows analysis of several tumor marker genes. The method is so sensitive that it allows the detection of only a few tumor cells per 5 ml EDTA-blood. The Agilent 2100 bioanalyzer provides the performance to detect the PCR products with high sensitivity and automated result flagging. This method offers new possibilities for monitoring and prognosis in routine diagnosis, and may facilitate an appropriate selection of patients for adjuvant therapy. Corresponding application note: 5988-9341EN
43
44
Oncology
SNP analysis in cancer related P16 gene
Data kindly provided by SAIC-Frederick
Kit: DNA 1000 kit Assay: DNA 1000 assay Application: Mutations in the exon 3 region of P16 gene are closely related to human cancer. A PCR yields 198 bp fragments with single, expected bands or additional, multiple bands in the Agilent 2100 bioanalyzer analysis. These observations correspond perfectly to genotyping sequencing data of normal and mutant tissues. The pattern of bands is visible due to slower mobility of the heteroduplex formed by heterozygote mutant of the samples. The method provides fast and reliable acquisition of genetic diagnostic data from cancer patients, also on single nucleotide polymorphisms (SNP). Corresponding application note: 5989-0487EN
Oncology
K-ras gene SNP detection
het ero zyg ous neg con tro l
Data kindly provided by SAIC-Frederick
- PCR product - Restriction
Kit: DNA 1000 kit Assay: DNA 1000 assay Application: Mutations in the K-ras gene coding 12 region can lead to cancer in different human tissues. A dedicated combination of PCR and specific restrictions (BstNI digest) reveals the underlying single nucletide polymorphisms (SNPs). The integral element within this test is the rapid and precise analysis of short amplicons (135 bp, see PCR-product lanes above) and fragments (106 bp, visible in lanes labeled with restriction) with the lab-on-a-chip technique. The test was used to ultimately determine a cancer patients eligibility for a clinical trial for a peptide vaccine. Corresponding application note: 5989-0487EN
Wi ld t ype
Mu tan t1
Mu tan t2
45
46
Oncology
METH-2 downregulation in lung carcinomas
Data kindly provided by Roy Castle Lung Cancer Research Programme, University of Liverpool, UK Kit: DNA 1000 kit Assay: DNA 1000 assay Application: Microarray analysis reveals under- or over-representations of transcripts. Screening of several cell lines for independent validation of such observations can be done with different techniques such as comparative multiplex PCR. This application shows the downregulation of a characteristic antiangiogenetic factor (METH-2) for a series of patient samples. Expression in normal tissue and tissue from the non small lung carcinomas is compared. Results from the array experiments were confirmed on a broad basis. Fast and convenient analysis with the Agilent 2100 bioanalyzer with given quantitation capability fit perfectly in such analytical workflow. Corresponding application note: 5989-3514EN
Oncology
Label-free analysis of microsatellite instability in carcinoma
Kit: DNA 1000 kit Assay: DNA 1000 assay Application: Microsatellite instability (MSI) is caused by a failure of the DNA mismatch repair system and occurs frequently in various types of cancer. Given that conventional techniques used for MSI detection, for example, polyacrylamide gel electrophoresis (PAGE) or capillary electrophoresis, turned out to be laborious or expensive, this study aimed to develop a simple and efficient procedure of MSI detection. Detection of MSI could be demonstrated by microsatellite loci-associated, well defined deviations in the electropherogram profiles of tumor and non-tumor material and confirmed the classification of the MSI cases performed by conventional technology (95% concordance rate). Whereas the results of the MSI detection were comparable to conventional techniques, the on-chip electrophoresis on the Agilent 2100 bioanalyzer was superior in terms of speed, usability and data management. Corresponding application note: 5989-2626EN
47
48
Clinical research
Genotyping of H. pylori
A B
Data kindly provided by Institute for Pathology, Cologne Kit: DNA 1000 kit Assay: DNA 1000 assay Application: Different allelic variants are associated with different stages of H. pylori virulence. Multiplex PCR on five alleles with products in the range of 102 to 301 bp were used to analyze DNA from paraffin embedded tissues. Agarose gel (A) yields only limited distinctiveness, whereas gel-like images (B) and electropherograms (C) show good resolution and superior reproducibility allowing convenient analysis of all desired products in parallel (D). An extended spectrum of prognostic or therapeutic relevant information is now routinely accessible for simultaneous analysis. Corresponding application note: 5989-0078EN
Clinical research
Duplications and deletions in genomic DNA
No mutation
Duplication in exon 4
Duplication in exon 12
Deletion in exon 45
Data kindly provided by Center for Human and Clinical Genetics, Leiden
Kit: DNA 500 kit Assay: DNA 500 assay (in expert software) Application: Multiplex amplifiable probe hybridization (MAPH) and multiplex ligation-dependent amplification (MLPA) are high throughput techniques for the detection of reordered genomic segments. These methods include hybridization of amplifiable probes with either stringent washing or ligation events prior to amplification. Exact and reproducible sizing and quantitation of multiple products are important prerequisites which are delivered by the Agilent 2100 bioanalyzer and lead to quick and simple analysis of genetically related diseases. Corresponding application note: 5989-0192EN
49
50
Forensic testing
Optimization of PCR on mtDNA
Amplified areas in human mtDNA
Kit: DNA 500 kit Assay: DNA 500 assay Application: Human mitochondrial DNA (mtDNA) is amplifiable even from small or badly degraded samples, even if genomic DNA is not available. Lanes B and C show homogenous PCR products which can subsequently be sequenced for identification. However, careful optimization of PCR parameters, like pH, Mg2+ concentration or polymerase amount is necessary and shown in detail in this application note. For example, a high Taq concentration increased the yield but also increased the level of byproducts. PCR for samples in lane D (impurities) and lane A (C-heteroplasmy) need to be improved. The Agilent 2100 bioanalyzer provides a rapid quantitative analysis over the broad size and concentration range needed for optimization and QC. It has proven to be an indispensable tool for forensic labs. Corresponding application note: 5989-3107EN
Forensic testing
Pitfalls in mtDNA sequencing
Kit: DNA 500 kit Assay: DNA 500 assay Application: Analysis of the non-coding sequence of human mitochondrial DNA (mtDNA) is performed for the purpose of identification in forensics. PCR amplification of limited or degraded mtDNA is done prior to sequencing. Quantitation and quality control of these PCR products (10-100 ng/ml, homogenous fragment in the range of 200-500 bp) was performed. Difficult PCR templates may cause G-stutters or other unintended byproducts of higher or lower mass (left). This may lead to indistinct sequence readings (right). Therefore, e.g. FBI guidelines enforce a 10% impurity level at the most. Fulfillment of this prerequisite can be satisfactorily verified with the Agilent 2100 bioanalyzer. Corresponding application note: 5989-0985EN
51
52
III. RNA analysis

Agriculture/Food
Pharmaceuticals
Proteomics
Genomics
Analysis of total RNA RNA integrity Standardization of RNA Quality Control Reproducibility of quantitation Genomic DNA contamination Low amounts of total RNA Detection of low levels of RNA RNA integrity with the RNA 6000 Pico kit RNA quality after staining and microdissection Analysis of minimum RNA amounts Genomic DNA in low concentrated RNA extracts Low RNA amounts from kidney sections Analysis of mRNA RNA integrity Ribosomal RNA contamination in mRNA samples Analysis of Cy5-labeled samples Analysis of cRNA with and without dye in gel matrix Optimization of labeling reactions cRNA fragmentation Analysis of T7-RNA transcripts Size estimation
Analysis of total RNA

RNA integrity
High quality total RNA
Partially degraded products

Agilent 2100 bioanalyzer: electropherogram Agilent 2100 bioanalyzer: single lane gel-like image
Kit: RNA 6000 Nano kit Assay: Eukaryote total RNA Nano assay Application: Analysis of total RNA integrity - a typical first QC step during cDNA or cRNA sample prep for microarrays. In Figure A the upper electropherogram and gel-like image show the analysis of high quality total RNA with the 18S and 28S subunit as two distinct bands. Figure B shows the analysis of a partially degraded total RNA sample. Many degradation products appear between the two ribosomal bands and below the 18S band. With the help of the Agilent 2100 bioanalyzer and the RNA 6000 Nano kit the important sample QC step prior to an expensive microarray experiment can be easily and quickly achieved. Corresponding application note: 5968-7493EN
53
54

Standardization of RNA Quality Control
Kit: RNA 6000 Nano kit Assay: Eukaryote total RNA Nano assay Application: The RNA integrity number (RIN) is calculated by a dedicated software algorithm (expert 2100 software, starting with Rev 02.01) to assess the quality of RNA preparations. The RIN tool is a major step in the standardization of user-independent RNA evaluation and delivers more meaningful information than simple ratio calculations for ribosomal RNA peaks. It is not influenced by instrument, sample integration and most important, concentration variability, thereby facilitating the comparison of samples and avoiding cost-intensive experiments with low quality RNA preparations. The RIN algorithm is based on a large collection of RNA data of various tissues and qualities. Furthermore, anomalies like genomic DNA contaminations are indicated with weighted error messages (critical/non-critical) to achieve a maximum of reliability. Corresponding application note: 5989-1165EN

Reproducibility of quantitation
Reproducibility for 12 consecutive runs
Kit: RNA 6000 Nano kit Assay: Eukaryote total RNA Nano assay Application: Alongside the quality control of RNA samples, measurement of RNA concentration is important for (bio-)chemical reactions, such as labeling reactions in the context of microarray experiments. With the RNA 6000 Nano kit good reproducibility can be achieved (here 6% CV), which is little affected by sample contaminants, such as phenol.
Corresponding application note: 5988-7650 EN
55
56

Genomic DNA contamination
Kit: RNA 6000 Nano kit Assay: Eukaryote total RNA Nano assay Application: Gel representation of a chip run with total RNA samples (mouse brain) spiked with varying amounts of herring sperm genomic DNA before and after treatment with RNase. The left panel shows the intact RNA with broad bands in the low MW region stemming from the genomic DNA. After the RNase digest (right panel) only the DNA bands remain, ranging in intensity according to the amount of DNA spiked into the sample.
Low amounts of total RNA

Detection of low levels of RNA
Analysis of mouse brain RNA at two different concentrations
Kit: RNA 6000 Pico kit Assay: Eukaryote total RNA Pico assay Application: The RNA 6000 Pico kit is complementary to the RNA 6000 Nano kit and is suitable for all applications where the amount of RNA (or cDNA) is limited, e.g. for biopsy samples, samples from microdissection experiments, QC of cDNA made from total RNA, microarray samples, etc. Here Agilent 2100 bioanalyzer results obtained from mouse brain RNA (Ambion) at 200 and 1000 pg/l are shown. By analysis in repetitions the reproducibility of quality control is demonstrated. Detection of 200 pg total RNA could be achieved without problems. Corresponding application note: data not published
57
58

RNA integrity with the RNA 6000 Pico kit
Intact RNA
Degraded RNA
Kit: RNA 6000 Pico kit Assay: Eukaryote total RNA Pico assay Application: Detection of RNA degradation with the RNA 6000 Pico kit. Sample: mouse liver total RNA (Ambion) concentration: 1 ng. Degradation was accomplished by adding a low amount of RNase. In Figure A the upper electropherogram and gel-like image show the analysis of high quality total RNA with the 18S and 28S subunit as two distinct bands. Figure A shows the analysis of a partially degraded total RNA sample. Many degradation products appear between the two ribosomal bands and below the 18S band. Corresponding application note: data not published

RNA quality after staining and microdissection
Laser micro dissection
Check and optimize RNA quality and yield Kit: RNA 6000 Pico kit Assay: Eukaryote Total RNA Pico assay Application: RNA derived from laser-microdissected tissue isolated by the PALMMicroBeam system was shown to be of high quality by convenient analysis with the RNA 6000 Pico assay. RNA-purification kits from different manufacturers and various common staining procedures have been tested and yielded 130-700pg/l RNA from 1000 cells with different quality (see above). The RNA 6000 Pico kit was well suited to show differences in RNA quality and yield and, therefore, is an ideal tool to optimize and adapt experimental conditions to individual tissue. The experiments were accompanied by a more laborious real time PCR that revealed similar results. Due to its unprecedented sensitivity, the RNA 6000 Pico assay is an indispensable tool for quality control in the context of microdissection experiments, ensuring successful gene expression profiling experiments. Corresponding application note: 5988-9128EN
59
60

Analysis of minimum RNA amounts
Kit: RNA 6000 Pico kit Assay: Eukaryote Total RNA Pico assay Application: The challenge of analysis of minimal amounts of RNA from e.g. laser micro dissections calls for detailed knowledge of extraction conditions. Some commonly used RNA isolation kits and buffer components were assessed in detail. The majority of the kits had no negative effect on the performance of the analysis, whereas, some kits include buffers which lead to shifted, missing and diminished RNA-peaks. In figure A, RNA isolated after microdissection shows lack of the 28S-peak due to high salt concentration introduced during the isolation process. In figure B, a standard RNA was diluted in water and subsequently extracted with a commercially available RNA extraction kit. The original samples (red) and the eluates after extraction are shown. These data show the importance of evaluating the individual method used for RNA extraction to exclude misleading conclusions. Corresponding application note: 5989-0712EN

Genomic DNA in low concentrated RNA extracts
Concentration 0.14 ng/l
Concentration 0.5 ng/l
Kit: RNA 6000 Pico kit Assay: Eukaryote Total RNA Pico assay Application: Laser capture microdissection enables collection of cells from small tissue areas. A low RNA yield is in the nature of the extraction method from such a specimen that usually complicates quality assessment a fact that can be circumvented by taking advantage of the Agilent 2100 bioanalyzer capabilities. A comparative study using mouse kidney cryosections showed that on-column DNase digestion is indispensable to obtain a reasonable result for integrity and yield (figure A). Experiments with omitted on-column DNA digestion confirmed that the peak visible in the inter-region consists of genomic DNA which caused overestimation of extracted RNA amounts (figure B). Corresponding application note: 5989-0991EN
61
62

Low RNA amounts from kidney sections
Renal medulla RNA
(1200 pg/l) before DNAse treatment
650 pg/l after DNAse I treatment
Kit: RNA 6000 Pico kit Assay: Eukaryote Total RNA Pico assay Application: High sensitivity quality control of RNA samples using the RNA 6000 Pico LabChip kit are demonstrated for microdissected samples (0.1 mm3). DNAse I digestion revealed that DNA contamination was present in the sample. Removal of DNA revealed total RNA with a low degree of degradation. Under ideal conditions, the RNA Pico assay has a linear response curve and, therefore, allows estimation of RNA concentrations.
Analysis of mRNA
RNA integrity
Highly enriched Poly (A)+ RNA
Progressive degradation of Poly (A)+ RNA
Kit: RNA 6000 Nano kit Assay: mRNA Nano assay Application: Progressive degradation of Poly (A)+ RNA. Poly (A)+ RNA (60 ng/L) from cultured Jurkat cells was incubated for 15 minutes at room temperature with very dilute RNase A (1 x 10-6 and 2 x 10-6 mg/mL, respectively). A progressive shift towards shorter fragment sizes can be observed. Even with a mild degradation, the absence of very long transcripts can be noticed.
63
64
Analysis of mRNA
Ribosomal RNA contamination in mRNA samples
Kit: RNA 6000 Nano kit Assay: mRNA Nano assay Application: Ribosomal contamination in mRNA samples. During the isolation of mRNA, varying amounts of ribosomal RNA can remain in a sample. Since the purity of mRNA is of importance for a number of downstream applications, samples should be checked on the Agilent 2100 bioanalyzer. This slide shows the analysis of 6 commercially available RNA samples from different suppliers. Analysis on the Agilent 2100 bioanalyzer reveals large differences in the purity of the mRNA samples. Corresponding application note: 5968-7495EN
Analysis of Cy5 labeled samples

Analysis of cRNA with and without dye in gel matrix
Assay conditions: - Cy5 labeled nucleic acids - no intercalating dye used - 5nM Cy5 dCTP added to gel matrix and sample buffer for focusing Lanes: 1. Unlabeled cRNA 2. Cy5 labeled cRNA 3. Cy5 labeled cRNA
Assay conditions: as above, but intercalating dye included in gel matrix Lanes: 1. RNA transcript ladder 2. Unlabeled cRNA 3. Cy5 labeled cRNA 4. Cy5 labeled cRNA
Kit: RNA 6000 Nano kit Assay: mRNA Nano and Cy5 labeled nucleic acids Nano assay Application: Analysis of Cy5 labeled and non-labeled cRNA samples. Cy5-labeled samples show the combined signals of the fluorescent label and the RNA signal created by the fluorescence of the RNA 6000 dye. If the RNA 6000 dye is omitted from the gel matrix, only the signal created by Cy5 is detected, allowing the determination of dye incorporation after a labeling reaction. Please note that for Cy3 labeled samples the intactness of the sample can be verified, but the dye incorporation can not be checked. Corresponding application note: 5980-0321EN
65
66

Optimization of labeling reactions
Kit: RNA 6000 Nano kit Assay: Cy5 Labeled Nucleic Acids Nano assay Application: An experiment was designed to check the influence of Cy5 dCTP concentration on labeling efficiency. Lane 2 represents the negative control (primer ommitted from the reaction mixture), while lane 3 shows the analysis of a reaction with a 6-fold increased Cy5 dCTP concentration. A look at the electropherograms reveals that the high Cy5 dCTP concentration not only gave a high peak of unincorporated Cy5, but also the labeling efficiency for longer fragments was very low. This approach allows the optimization of labeling reactions. Corresponding application note: 5980-0321EN

cRNA fragmentation
1 Cy5-1 2 Cy5-1 3 Cy5-1 4 Cy5-1 5 Cy5-1
40 min 30 min 20 min 10 min no fragmentation
Kit: RNA 6000 Nano kit Assay: Eukaryote total RNA Nano assay Application: The RNA 6000 Nano LabChip kit can be used to monitor completion of a cRNA fragmentation reaction. In this example, the profile of a Cy5 labeled cRNA sample was monitored at different time points during a fragmentation reaction. It can be seen that after 10 minutes most of the fragments are in the desired size range. After 20 minutes, no further shift of fragmentation can be observed indicating completion of the fragmentation reaction.
67
68
Analysis of T7 RNA transcripts

Size estimation
Kit: RNA 6000 Nano kit Assay: Eukaryote total RNA Nano assay Application: A number of RNA transcripts, ranging from 350 to 1400 nt in size, were analyzed on the RNA 6000 Nano LabChip kit. Although the assay runs under native conditions and the transcripts exhibit a certain degree of secondary structure, a good size estimation can be achieved.
IV. Protein analysis

Agriculture/Food
Pharmaceuticals
Proteomics
Genomics
Protein expression Analysis of cell lysates - protein induction Protein purification: Comparison between lysate and flow through Analysis of protein purification GFP Streptag fusion protein purification Analysis of column capacity Analysis of column fractions to optimize conditions His-tag protein purification with Ni++ ZipTips Enzymatic removal of His-tags from recombinant proteins Complementing RP-HPLC protein purification Antibody analysis Analysis of antibodies under reducing and non-reducing conditions Quantitation of the half-antibody content in IgG4 preparations Comparison of SDS-PAGE, CGE and Agilent 2100 bioanalyzer for humanized monoclonal antibody analysis Absolute quantitation of IgG Quality control of stressed antibodies Separation of bispecific antibodies chains Food analysis Bovine milk analysis Protein pattern of different transgenic seedlines Protein - others Absolute protein quantitation Glycoprotein sizing Protein quality control prior to MS-analysis Depletion of high abundant proteins from blood samples Increased sensitivity by desalting protein samples
69
70
Protein expression
Analysis of cell lysates - protein induction
-Galactosidase
Kit: Protein 200 Plus kit Assay: Protein 200 Plus assay Application: Two cell lysates, induced and non-induced were compared to verify the induction of protein expression. The overlay feature of the bioanalyzer software allows quick sample comparison. The blue electropherogram trace shows the cell lysate highly expressing -galactosidase (128 kDa).
Protein purification
Comparison between lysate and flow through
38 kDa
77 kDa
load onto the column
column flow through
Kit: Protein 200 Plus kit Assay: Protein 200 Plus assay Application: Cells were lysed using the Pierce B-Per kit and then loaded onto an affinity column. The protein of interest, a 38 kDa protein, should bind to the column and not show up in the flow through. By overlaying the 2 electropherograms from both samples, the lysate and the flow through, it is visible that the protein of interest has bound to the column as expected. In addition, a 77 kDa protein has bound to the column, which could be attributed to unspecific binding or the binding of a dimer. Corresponding application note: data not published
71
72
Analysis of protein purification
Courtesy of P. Sebastian and S.R. Schmidt GPC-Biotech AG, Martinsried, Germany
Kit: Protein 200 Plus kit Assay: Protein 200 Plus assay Application: A 18 kDa protein was purified using affinitiy chromatography. The starting material and the column fraction were analyzed with the protein assay. The protein of interest was determined to be 99% pure and the concentration in the column fraction was 167 ng/l. The protein assay allows protein purity and concentration to be determined in one step, in addition it calculates protein size for reconfirmation.
GFP Streptag fusion protein purification
Kit: Protein 200 Plus kit Assay: Protein 200 Plus assay Application: This example shows the analysis of various steps during the purification workflow of a GFP Streptag fusion protein (28 kDa). The protein was expressed in E.coli and purified via affinity chromatography with Strep Tactin Poros as the column matrix. The protein assay allows each purification step from the cell lysis to the elution of the purified protein to be monitored and optimized.
73
74
Analysis of column capacity
column load conc.: 150 g/ml flow through conc.: 60 g/ml
Kit: Protein 200 Plus kit Assay: Protein 200 Plus assay Application: The binding of a recombinant antibody Fab fragment to a Sepharose column with immobilized Protein G was analyzed to determine the column capacity and prevent column overloading. The protein assay allows this purification step to be monitored and quickly optimized.
Analysis of column fractions to optimize conditions
column 1-eluate 33 ng/l
column 2-eluate 44 ng/l
column 2-eluate not detected
Kit: Protein 200 Plus kit Assay: Protein 200 Plus assay Application: Different column conditions were tested to optimize the purification conditions for a 30 kDa protein. The column fractions were analyzed for protein purity and concentration to identify the optimal conditions providing a highly purified protein in a good yield. Using the protein assay it was possible to determine the optimum purification conditions in a short time frame.
75
76
His-tag protein purification using Ni++ZipTips
E.coli cell lysate
purified protein, 25 kDa eluted with acetic acid
Kit: Protein 200 Plus kit Assay: Protein 200 Plus assay Application: ZipTips loaded with a Ni2+-resin (in development by Millipore) were used to purify a His-tagged protein expressed in E.coli. Both the cell lysate and the purified protein were analyzed with the Agilent 2100 bioanalyzer to demonstrate the performance of the tips. The purification with the tips takes approximately 5 minutes, usually followed by the analysis of the samples with SDS-PAGE analysis which takes a further 2 hours. The SDS-PAGE analysis was substituted by the much faster Protein 200 Plus assay run on the Agilent 2100 bioanalyzer. Corresponding application note: data not published
Enzymatic removal of His-tags from recombinant proteins
Kit: Protein 200 Plus kit Assay: Protein 200 Plus assay Application: For some applications, it might be necessary to remove the His-tag after the protein purification because of its effects on enzymatic activity or protein structure. Here the TAGZyme system (Qiagen) was used to remove the N-terminal His-tag from two different proteins, a GFP variant and a recombinant Interleukin 1. Samples were taken at different time points to study the kinetics of the enzymatic cleavage. The dipeptide cleavage can be detected by a size shift on the gel-like images and the electropherograms. The fast analysis with the bioanalyzer allows multiple kinetic studies to done in one day instead of waiting until the next day for the results from SDS-PAGE analysis. Poster presented at ABRF Conference, March 2002 by F. Schfer, K. Steinert, C. Feckler, J.Drees, and J.Ribbe, QIAGEN GmbH, Hilden, Germany Corresponding application note: 5988-8144 EN
77
78
Complementing RP-HPLC protein purification
Kit: Protein 200 Plus kit Assay: Protein 200 Plus assay Application: Protein purification and characterization was carried out facilitating an Agilent 1100 Series purification system for reverse phase HPLC assisted by the Agilent 2100 bioanalyzer. The final polishing of a 56 kDa protein by RP HPLC from a pre-purified sample (starting material, right: red electropherogram and gel) and the analysis of three HPLCfractions containing the major components are shown (fractions 1-3). No impurity is visible by RP HPLC reanalysis (left chromatogram, fraction 2) of the fraction containing the target protein. However, because the Agilent 2100 bioanalyzer is an orthogonal technique compared to reverse phase HPLC a 20 kDa protein could be found as an impurity (see insert). The reverse phase HPLC purification leads to a purity of only 76% for the protein of interest and the Agilent 2100 bioanalyzer reveals the necessity of further purification. Corresponding application note: 5988-8630EN
Antibody analysis
Analysis of antibodies under reducing and non-reducing conditions
Kit: Protein 200 Plus kit Assay: Protein 200 Plus assay Application: The protein kit allows analysis of both reduced and non-reduced antibodies on the same chip. This is not possible using SDS-PAGE, as the reducing agent will diffuse within the gel and will also reduce other samples. Under non-reducing conditions, it is expected to detect the intact antibody around 160 kDa. Here the single light and heavy chains and half-antibodies are also visible. Under reducing conditions this is all completely reverted to single light and heavy chains, due to the reduction of the disulfide bonds.
79
80
Antibody analysis
Quantitation of the half-antibody content in IgG4 preparations
Kit: Protein 200 Plus kit Assay: Protein 200 Plus assay Application: In the given host cell line for antibody production usually up to 30% of IgG4 is secreted as half molecule (half antibody). The half-molecule has only a single disulfide bond between the heavy and light chains, the inter-heavy chain disulfide bonds are absent. The protein assay allows the half-antibody content in IgG4 preparations to be determined automatically. In addition, the sizing provided by the Agilent 2100 bioanalyzer compares very well to the theoretical size and is superior to SDS-PAGE in terms of accuracy and reproducibility. Poster presented at WCBP Conference, January 27-30, 2002 by E. Vasilyeva, H. Fajardo, P. Bove, F. Brown and M. Kretschmer. BIOGEN, Cambridge, MA , USA Corresponding application note: data not published
Antibody analysis
Comparison of SDS-PAGE, CGE and Agilent 2100 bioanalyzer for humanized monoclonal antibody analysis
Kit: Protein 200 Plus kit Assay: Protein 200 Plus assay Application: The analysis of a humanized monoclonal antibody under reducing condition was compared using 3 different techniques, the Agilent 2100 bioanalyzer, 4-20% SDS-PAGE, stained with Coomassie, and capillary gel electrophoresis. All 3 techniques result in a similar separation pattern showing the light and the heavy chain of the antibody. In addition, the determined sizes of the light and heavy chain were comparable for all 3 techniques and compared well to the molecular weights determined by MALDI-TOF (light chain: 23762 Da, heavy chain: 51003 Da). However, the Agilent 2100 bioanalyzer provides significant time saving compared to the other techniques. Poster presented at WCBP Conference, January 2002 by S.H. Bowen, M. Chan, P. McGeehan, J. Smith, L. Inderdass, R. Strouse, M. Schenerman MedImmune Inc., Gaithersburg, MD, USA Corresponding application note: data not published
81
82
Antibody analysis
Absolute quantitation of IgG
Kit: Protein 200 Plus kit Assay: Protein 200 Plus assay Application: The calibration feature of the software allows determination of the absolute antibody concentration in comparison to user defined standards with known concentration, accurate determination of IgG concentrations and carrying out batch comparison during antibody QA/QC.
Corresponding application notes: 5988-4021EN and 5988-6576EN
Antibody analysis
Quality control of stressed antibodies
Kit: Protein 200 Plus kit Assay: Protein 200 Plus assay Application: A quality control step in pharmaceutical QA/QC departments is to trigger typical degradation and aggregation patterns for a specific antibody. The given samples from heat stress stability studies show expected protein byproducts after aging at elevated temperatures. The content of heavy and light chain, representing the intact antibody, is reduced by 5% or 13% within 1 month or respectively 12 weeks. Excellent reproducibility in the range from 0.6 to 1.7% CV for this quantification was achieved in a validation study with three different users and two bioanalyzer instruments over several days. Corresponding application note: 5988-9648EN
83
84
Antibody analysis
Separation of bispecific antibodies chains
Kit: Protein 200 Plus kit Assay: Protein 200 Plus assay Application: In general, antibodies are biopharmaceuticals of great interest. Especially bispecific antibodies often require high resolution to allow analysis of both sets of chains (Agilent 2100 bioanalyzer: A, gel like view, resolved heavy chains; C electropherogram). A labor intensive SDS-PAGE could not resolve the heavy chains (B, marked by an arrow) in the given sample. In contrast, the Agilent 2100 bioanalyzer is a superior tool for antibody quality control since it is a convenient, fast and easy to standardize method which additionally enables quantitative analysis. Corresponding application note: 5988-9651EN
Food analysis
Bovine milk analysis
Kit: Protein 50 kit Assay: Protein 50 assay Application: The Protein 50 LabChip is suitable for analysis of diary products such as milk. The kit delivers an excellent reproducibility, as shown in this example (A, bovine milk diluted 1:10). Here the main protein fractions can be identified running the individual purified proteins for comparison (not shown). The overlay of the electropherograms from two separate runs under reducing conditions demonstrates the high reproducibility of the assay. Furthermore, milk from different animals could be distinguished based on their protein pattern (B) which facilitates a fast incoming inspection in routine labs. Corresponding application note: data not published
85
86
Food analysis
Protein pattern of different transgenic seedlines
SeedLine
Extracted protein level g/ml 14,000 5,200 14,000 13,000
7S/11S Ratio
Control 7S 11S Unknows
0.390.004(n=5) 3.4 0.04 0.720.1(n=20)
Kit: Protein 200 Plus kit Assay: Protein 200 Plus assay Application: Determination of protein size and concentration with sufficient accuracy and precision allows the highly efficient characterization of transgenic seed lines. Expressed protein was available after grinding, extraction of seeds and dilution with buffer. Electropherograms were evaluated by integration of regions specific for 7S or 11S seed storage proteins. The elevated ratio of 7S/11S for the analyzed unknown line shows significant changes in the expression profile in comparison with the control. Corresponding application note: 5988-9441EN
Protein - others
Absolute protein quantitation
Kit: Protein 200 Plus kit Assay: Protein 200 Plus assay Application: A comparative analysis of different techniques used for absolute protein quantitation was performed analyzing 3 different proteins (CA, BSA, OV) in 4 different concentrations (40 - 1250 ug/ml). The same samples were quantitated using the Agilent 2100 bioanalyzer, two commonly used total protein quantitation assays, Lowry and Bradford, and SDS-PAGE, stained with Coomassie. The relative standard deviation (CV) and the error compared to the target concentration were determined. A comparison shows that the CV and error for the Agilent 2100 bioanalyzer are better than for the SDS-PAGE by a factor of 2. This data demonstrates that the Agilent 2100 bioanalyzer is a viable alternative for protein quantitation. It allows the quantitation of individual proteins and simultanous determination of protein purity and size. Corresponding application notes: 5988-4021EN and 5988-6576EN
87
88
Protein - others
Glycoprotein sizing
100.0
80.0
SDS-PAGE / untreated SDS-PAGE / deglycol. Bioanalyser / untreated Bioanalyser / deglycol. Calculated / deglycol.
60.0
40.0
20.0
0.0 Rnase B IgG Light Chain IgG Heavy Chain Ovomucoid Ovalbumin alpha 1-acid glycoprotein Transferrin
Kit: Protein 200 Plus kit Assay: Protein 200 Plus assay Application: Due to large carbohydrate moieties glycosylated proteins can differ in amount of incorporated SDS and shape of the protein/SDS-complex from non-glycosyslated proteins. This may lead to different migration times in SDS-PAGE, as well as in the Protein 200 Plus assay run on the Agilent 2100 bioanalyzer. The data compare deglycosylation of a mixture of three proteins (electropherogram on the left) with a commercial N-Glycosidase F Deglycosylation kit. Sizing experiments comparing glycosylated and non glycosylated states for additional proteins are compared and summarized on the right. Such an approach avoids misinterpretation of sizing due to glycosylation and allows detection of a posttranslational modification of unknown proteins. Corresponding application note: 5989-0332EN
Protein - others
Protein quality control prior to MS-analysis
Kit: Protein 200 Plus kit Assay: Protein 200 Plus assay Application: By applying soft ionization methods like MALDI or ESI mass spectroscopy (MS), mass information from proteins up to 300kDa can be obtained. However, proper sample preparation is an important precondition. Concentration, purity and assumed size are valuable ex ante information usually given by biochemists to MS-analysis services. Two different examples for proteins analyzed by an LC/MS-method (right panel) with good results for sample A and discrepancies for sample B are shown. An analysis of the samples with the Protein 200 Plus assay (left panel) showed an impure protein preparation for sample B. Here, two major peaks at higher masses (66 and 132kDa) potentially caused by aggregates were encountered. The protein of interest (30kDa) yielded high noise background in the MS. A quality check of the sample with the Protein 200 Plus assay, therefore, may avoid an unproductive MS analysis or data evaluation. Corresponding application note: 5989-0771EN
89
90
Protein - others
Depletion of high abundant proteins from blood samples
Kit: Protein 200 Plus kit Assay: Protein 200 Plus assay Application: Depletion of high abundant proteins in human blood plasma was facilitated by the Agilent Multiple Affinity Removal System. Unprocessed serum, the flow through (i.e. the immunodepleted serum) and the bound proteins after specific elution were analyzed on the Agilent 2100 bioanalyzer. Equivalent amounts were analyzed and resulting electropherograms (left) and gel like images (right) show in comparison the completeness of separation. The Agilent 2100 bioanalyzer, in combination with the Protein 200 Plus LabChip kit, proves to be an excellent method for evaluation of serum processed with albumin removal kits. The system offers a rapid and accurate method to detect proteins both quantitatively and qualitatively. Corresponding application note: data not published
Protein - others
Increased sensitivity by desalting protein samples
Kit: Protein 200 Plus kit Assay: Protein 200 Plus assay Application: Protein purification steps, such as ion exchange chromatography, often implicate high salt concentrations. Nevertheless, quantitation of these samples is effective since the upper marker serves as internal protein standard and is subjected to the same conditions. However, under high salt conditions lower amounts of protein are injected into the microfluidic channels for analysis. Therefore, the sensitivity can be increased by usage of convenient desalting spin columns. Comparably higher and homogeneous peak intensities are obtained while the recovery after such treatment is good.
91
92
Literature
Microfluidics application notes from Agilent Technologies

To download an application note visit the library section at: www.agilent.com/chem/labonachip Description DNA Publication number
5968-7496EN 5968-7501EN 5980-0549EN 5988-1086EN 5988-3041EN 5988-3674EN 5988-4069EN 5988-4070EN 5988-4847EN 5988-9341EN 5988-9342EN 5989-0078EN 5989-0124EN 5989-0192EN 5989-0487EN 5989-0985EN 5989-0991EN 5989-1870EN 5989-2487EN 5989-2626EN 5989-2982EN 5989-3107EN 5980-0321EN 5989-3514EN
Quantitative analysis of PCR fragments with DNA 7500 LabChip kit High precision restriction fragment sizing with DNA 12000 LabChip kit Comparing the Agilent 2100 bioanalyzer performance to traditional DNA analysis Agilent 2100 bioanalyzer replaces gel electrophoresis in prostate cancer research High resolution DNA analysis with the DNA 500 and DNA 1000 LabChip kits Quantitative end-point RT-PCR gene expression using DNA 7500 LabChip kit Development of meat speciation assays using the Agilent 2100 bioanalyzer Analysis of genetically modified soya using the Agilent 2100 bioanalyzer Detecting genetically modified organisms with the Agilent 2100 bioanalyzer Sensitive detection of tumor cells in peripheral blood of carcinoma patients by a reverse transcription PCR method Highly efficient multiplex PCR using novel reaction chemistry Microfluidic analysis of multiplex PCR products for the genotyping of Helicobacter pylori Nested multiplex PCR for the determination of DNA from genetically modified corn and soy beans using the Agilent 2100 bioanalyzer Rapid detection of genomic duplications and deletions using the Agilent 2100 bioanalyzer Mutation detection for the K-ras and P16 genes Use of the Agilent 2100 bioanalyzer and the DNA 500 LabChip in the analysis of PCR amplified mitochondrial DNA Assessing genomic DNA contaminations of total RNA isolated from kidney cells obtained by Laser Capture Microdissection using the Agilent RNA 6000 Pico assay Integrating high-throughput, on-chip electrophoresis analysis into PCR diagnostics projects DNA QC for Oligonucleotide Array CGH (aCGH) with the Agilent 2100 bioanalyzer Label-free analysis of microsatellite instability in colorectal carcinoma by on-chip electrophoresis Determination of PCR-RFLP Profiles for Fish Species Using the Agilent 2100 bioanalyzer Using the Agilent 2100 bioanalyzer to Optimize the PCR Amplification of Mitochondrial DNA Sequences Analysis of Cy5-labeled cRNAs and cDNAs using the Agilent 2100 bioanalyzer and the RNA 6000 LabChip kit Measuring the METH-2 promoter hypermethylation and transcript dwonregulation in non-small cell lung carcinomas with the Agilent 2100 bioanalyzer
RNA
Analysis of total RNA using the RNA 6000 LabChip kit Analysis of messenger RNA using the RNA 6000 LabChip kit Analysis of Cy5-labeled cRNAs and cDNAa using the RNA 6000 LabChip kit Quantitation comparison of total RNA using the Agilent 2100 bioanalyzer, ribogreen analysis and UV spectrometry Characterization of RNA quality using the RNA 6000 LabChip kit Comparing performance of the Agilent 2100 bioanalyzer to traditional RNA analysis The total RNA story Interpreting mRNA electropherograms Optimizing cRNA fragmentation for microarray experiments using Agilent 2100 bioanalyzer High sensitivity quality control of RNA samples using the RNA 6000 Pico LabChip kit Quality assurance of RNA derived from laser microdissected tissue samples obtained by the PALM MicroBeam system using the RNA 6000 Pico LabChip kit Advancing the quality control methodology to assess isolated total RNA and generated fragmented cRNA siRNA transfection optimization with the Agilent 2100 bioanalyzer Successful analysis of low RNA concentrations with the Agilent 2100 bioanalyzer and the RNA 6000 Pico Stringent RNA quality control using the Agilent 2100 bioanalyzer Gene Expression Profiling of Esophageal Cancer Using Laser Capture Microdissected Samples RNA Integrity Number (RIN) - Standardization of RNA Quality Control Confirming gene silencing mechanism by pGFP/GFP22 - siRNA co-transfection Sensitive detection of tumor cells in peripheral blood of carcinoma patients by a reverse transcription PCR method Assessing genomic DNA contaminations of total RNA isolated from kidney cells obtained by Laser Capture Microdissection using the Agilent RNA 6000 Pico assay High-Purity RNA isolation from a wide range of plant species and tissue types Isolation of high purity total cellular RNA from muscle tissues
5968-7493EN 5968-7495EN 5980-0321EN 5988-7650EN 5980-0472EN 5980-2206EN 5988-2281EN 5988-3001EN 5988-3119EN 5988-8554EN 5988-9128EN 5988-9861EN 5988-9872EN 5989-0712EN 5989-1086EN 5989-1088EN 5989-1165EN 5989-0103EN 5988-9341EN 5989-0991EN 5989-2271EN 5989-2312EN
Cells
Apoptosis detection by annexin V and active caspase-3 with the Agilent 2100 bioanalyzer Detection of cell surface proteins with the Agilent 2100 bioanalyzer by on-chip antibody staining Monitoring transfection efficiency in cells using an on-chip staining protocol A fast protocol for apoptosis detection by Annexin V with the Agilent 2100 bioanalyzer Cell fluorescence assays on the Agilent 2100 bioanalyzer - general use Monitoring transfection efficienency by green fluorescence protein (GFP) detection Detection of antibody-stained intracellular protein targets with the Agilent 2100 bioanalyzer Measuring multiple apoptosis parameters with the Agilent 2100 bioanalyzer Flow cytometric analysis of human primary cells using the Agilent 2100 bioanalyzer and on-chip staining Confirming gene silencing mechanism by pGFP/GFP22 - siRNA co-transfection Flow cytometric analysis of a limited number of cells using the Agilent 2100 bioanalyzer A new method for the calculation of baculovirus titre using the Agilent 2100 bioanalyzer and the flow cytometry kit Cytometric analysis of Upregulated Functional Gene Expression in Primary Cells by On-Chip Staining Detection of Apoptosis in Primary Cells by Annexin V Binding Using the Agilent 2100 bioanalyzer siRNA transfection optimization with the Agilent 2100 bioanalyzer
5988-4319EN 5988-7111EN 5988-7296EN 5988-7297EN 5988-4323EN 5988-4320EN 5988-4322EN 5988-8028EN 5988-8154EN 5989-0103EN 5989-0746EN 5989-1644EN 5989-2718EN 5989-2934EN 5988-9872EN
93
94
Proteins
Protein sizing and analysis using the Protein 200 LabChip kit Differences and similarities between Protein 200 Assay and SDS-PAGE (technical note) Comparison of different protein quantitation methods Using the Agilent 2100 bioanalyzer for analysis of His-tag removal from recombinant proteins Absolute quantitation with the Protein 200 LabChip kit Optimization of protein purification using the Agilent 2100 bioanalyzer Comparison of different methods for purification analysis of a green fluorescent strep-tag fusion protein Fast analysis of proteins between 5-50 kDA using the Agilent 2100 bioanalyzer and Protein 50 assay Using the Agilent 2100 bioanalyzer for analysis of His-tag removal from recombinant proteins Protein purification and characterization using the 1100 Series purification system and the Agilent 2100 bioanalyzer Characterization of Transgenic Soybean Seedlines by Protein Expression with the Agilent 2100 bioanalyzer Quality control of antibodies using the Agilent 2100 bioanalyzer and the Protein 200 Plus assay Analysis of bispecific antibodies using the Agilent 2100 bioanalyzer and the Protein 200 Plus assay Evaluation of albumin removal using the Agilent 2100 bioanalyzer Increased sensitivity by desalting protein samples prior to analysis on the Agilent 2100 bioanalyzer Glycoprotein sizing on the Agilent 2100 bioanalyzer Using the Agilent 2100 bioanalyzer for quality control of protein samples prior to MS-analysis Quality Control for the Agilent 2100 Bioanalyzer Protein 200 Plus LabChip Kits (technical note)
5988-0975EN 5988-3160EN 5988-6576EN 5988-8144EN 5988-4021EN 5988-4022EN 5988-5025EN 5988-8322EN 5988-8144EN 5988-8630EN 5988-9441EN 5988-9648EN 5988-9651EN 5988-9911EN 5989-0228EN 5989-0332EN 5989-0771EN 5989-3336EN
Scientific publications
The Agilent 2100 bioanalyzer is used in virtually every kind of scientific laboratory. A list of existing scientific publications with more than 2000 citations from well-known and high- impact journals is continuously updated and can be viewed at: http://www.agilent.com/chem/2100publications.

Ordering
Visit our online store at www.agilent.com/chem/store for the Agilent 2100 bioanalyzer products, accessories and services. Instruments Agilent 2100 bioanalyzer Agilent 2100 electrophoresis bioanalyzer Agilent 2100 bioanalyzer desktop system Agilent 2100 bioanalyzer laptop system Cell solutions Cell Fluorescence Checkout kit Cell Fluorescence LabChip kit DNA solutions DNA 500 LabChip kit DNA 1000 LabChip kit DNA 7500 LabChip kit DNA 12000 LabChip kit RNA solutions RNA 6000 Nano LabChip kit RNA 6000 Pico LabChip kit Protein solutions Protein 200 Plus LabChip kit Protein 50 LabChip kit Software Agilent 2100 expert software Agilent 2100 bioanalyzer security pack Services DNA/RNA assays operational service Protein assays operational service Cell assays operational service Compliance services Agilent 2100 bioanalyzer system IQ Agilent 2100 bioanalyzer system OQ/PV Accessories
Notes
Notes
97
98
Find out more:

U.S. and Canada
1-800-227-9770 agilent_inquiries@agilent.com
Europe
info_agilent@agilent.com
Asia Pacific
adinquiry_aplsca@agilent.com
To download an application note visit: www.agilent.com/chem/labonachip For online ordering visit our online store: http://www.agilent.com/chem/store For more information about lab-on-a-chip products visit: www.agilent.com/chem/labonachip For more information about microarray products visit: www.agilent.com/chem/dna
Copyright 2007 Agilent Technologies Printed in Germany, June 1, 2007 Publication Number: 5989-3542EN
One platform, endless possibilities

The Agilent 2100 Bioanalyzer is an indispensable tool for food chemists and biologists.
Now you can have a multi-purpose platform that streamlines your workflows from product development to QA/QC of bulk materials and finished food products. The Agilent 2100 Bioanalyzer can handle all your needs, whether you want to measure protein content in seeds, bulk material or finished product; measure DNA or RNA for molecular detection of genetically modified organisms, allergenic species or pathogens; or count cells.
Uniquely, the 2100 Bioanalyzer performs both electrophoretic separation and flow cytometric analysis of cell fluorescence parameters. It is rapidly replacing gel electrophoresis for DNA fragment analysis and SDS-PAGE analysis of protein samples.
Versatile, fast and mess-free

The Agilent 2100 Bioanalyzer is the industrys only platform that can cover your entire workflow with a single compact system. The first commercial, analytical instrument based on lab-on-a-chip technology, the 2100 Bioanalyzer has proven to be an excellent alternative to antibody-based, labor-intensive gel electrophoresis. This technology replaces subjective, time-consuming techniques associated with agarose or SDS-PAGE slab gels with fast, automated, high quality digital data.
Answers within minutes

The 2100 Bioanalyzer provides you with a convenient and productive way to gather and store experimental and routine test data. Automation and standardization of different processes on a chip give you high quality digital data fast, increasing lab productivity. You get answers within 30 minutes.
Advantages of miniaturization
Miniaturization of analytical instrumentation has a number of advantages over conventional techniques: Data precision and reproducibility Short analysis times Minimal sample consumption Improved automation Integration of complex workflows
Pre-packaged LabChip kits, which include sample-specific reagents and chips, let you analyze specific sample classes. A variety of kits for RNA, DNA, protein and cell assays are currently available to meet your needs.
Why you need to test-drive the Agilent 2100 Bioanalyzer

Advantages of the Agilent 2100 Bioanalyzer
Faster than gelsdigital data for up to 12 samples within 30 minutes Reproducible and complete digital data Compliance Service options allow you to work within regulated environments (particularly QA/QC labs) 2100 Expert Software for easy digital handling and storage of all bioanalyzer data Multiplex detection capabilities
Advantages of Agilents lab-on-achip technology

Minimal sample consumption and fast results Improved assay accuracy and precision Digital data for convenient archiving and storage Various data display options Ease-of-use with simplified sample-to-sample comparison Minimum exposure to hazardous materials
Digital data
The 2100 Bioanalyzer provides fully digital data easily shared with colleagues worldwide. The fully functioning data analysis software is available free at http://www.chem.agilent.com/ cag/wad/registration/2100expert.asp
1 Fast and easy operation

Add sample
2 Automation
Start chip run
3 Digital data in 30 minutes

Watch real-time data display
Ready-to-use reagent kits Quick-start instructions Chip preparation in less than 5 minutes Minimal use of hazardous chemicals and waste disposal Sample volumes in the L-range
Start analysis at the press of a button Predefined protocols System uses internal standards to calculate results
Automated data analysis Digital data can be filed in a database or shared No user-dependent data interpretation
Food safety applications

Nuts and allergens
The 2100 Bioanalyzer is more cost effective and reduces labor compared to other test methods. Recent regulatory and policy changes in the United States (US), European Union (EU), Japan, Canada and Australia require labeling of multiple allergens in food products. EU regulations, for instance, identify eight nut species that must be declared on the ingredient list. Currently available allergen content tests for food products and materials often have difficulty in complex matrices and usually detect only a single allergen per test. In contrast, the 2100 Bioanalyzer coupled with molecular detection techniques (PCR) can cost-effectively screen for multiple allergens using a single test. The electropherograms shown in Figure 1 demonstrate multiplex detection of three nut speciesBrazil nut (Bertholletia excelsa), pistachio nut (pistacia vera) and macadamia nut (macadamia integrifolia)in a single run. Furthermore, the bioanalyzer, which uses only 1 l of reaction mixture to load onto the LabChip, can reduce PCR volumes by up to 75 percent compared to standard gel-electrophoresis, which requires 10 to 15 l, saving reagent costs.
ia
am
hi
ac
ad
il
az
st
ac
Br
Pi
Positive
Figure 1: Cost-effective screening for multiplex PCR reaction product for nut allergen species in food using a single test. The 2100 Bioanalyzers gel-like image shows positive test results for pistachio nut (lane 4), Brazil nut (lane 5) and macadamia nut (lane 6), as well as for a mixture of these three nuts (lane 7) in wheat. Source: Campden and Chorleywood Food Research Association Group, Gloucestershire, UK.
Ad
ixt
ur

Genetically modified organisms
The number of genetically modified organisms (GMO) and regulatory requirements for testing and labeling continue to increase. Many countries now require GMO content labelingor ban GMO materials altogether. Antibody tests to detect GMO proteins often require individual tests for each suspected GMO organism/event. Test accuracy can be problematic in processed materials where detected proteins may become denatured or damaged. The 2100 Bioanalyzer, in conjunction with an appropriate PCR kit, offers multiplex screening to detect and identify multiple GMO events in processed foods, improving test accuracy and reducing the number of tests, which in turn reduce costs. Figure 2 shows multiplex detection of GMO in corn meal and soya powder.
4% NuSieve 3:1 Plus Gel. 150 V for 60 min.
35S GMO
Gel-like 2100 Bioanalyzer image.
Figure 2. Comparison of slab gel electrophoresis (A) with Agilent 2100 Bioanalyzer gel-like image (B) in detecting multiple GMO events in corn meal and soybean powder. Samples: L) molecular weight ladder (26, 34, 67, 89, 110, 147, 190, 242, 353, 404, 489 and 501 bp); 1) 0% MON810 corn; 2) 0.1% MON810 corn; 3) 0.5% MON810 corn; 4) 1.0% MON810 corn; 5) 2.0% MON810 corn; 6) 5.0% MON810 corn; 7) commercial corn meal; 8) Allin positive control0.5% Bt 176 maize and 0.5% Roundup Ready soybean; and 9) negative control (de-ionized water). Band identification: 118 bp soy lectin; 153 bp35S GMO; 217 bpPCR reaction internal control; 278 bpcorn zein.
Meat and fish

The multiplex detection capability of the 2100 Bioanalyzer can also determine food product authenticity, its source and homogeneity. For example, meat authenticity is an important factor for economic as well as religious and cultural reasons. Identifying authenticity and homogeneity of the meat source is difficult for processed meats in which proteins may be damaged by heat, rendering antibody-based tests inaccurate. Using molecular detection methods (PCR), the 2100 Bioanalyzer successfully determines meat species and sample homogeneity in one test, reducing the number of tests required and thereby reducing labor and testing costs. Figure 3 shows the bioanalyzers ability to detect turkey, lamb or pork in a variety of meat samples while discriminating for other various matrices such as fish or grain.
Figure 3. The 2100 Bioanalyzer can detect multiple targets at one time, including turkey, lamb or pork in a variety of meat samples. The dark band (lane 6) indicates a positive molecular test (PCR) for turkey. Negative tests (lanes 1 through 4) demonstrate test discrimination for other meat products (beef, chicken, lamb and pork).
Figure 4. Agilent 2100 Bioanalyzer gel-like image showing PCR test designed to discriminate meat. Positive results are evident in lanes 1 through 5 for various meat products. Lanes 6 through 12 show negative results for grain and fish. Source: Campden and Chorleywood Food Research Association Group, Gloucestershire, UK.

Dairy: Milk protein
In addition to DNA analysis for molecular detection applications, the 2100 Bioanalyzer can detect and quantify proteins to determine protein source, quality and content. Figure 5 demonstrates determination of protein content to check the authenticity of the milk source (such as goat, sheep or cow) as well as determine the protein content and quality for inprocess or finished product QA/QC.
[FU] 300 250 200 150 100 50 0 25 30 35 40 45 50 (Sec) Lower marker System peaks Upper marker Lactoglobulin Casein Casein
Figure 5. Milk protein analysis. The 2100 Bioanalyzer protein assay provides rapid analysis of milk proteins. Shown here are two overlaid electropherograms of bovine milk (dil. 1:10; peaks identified by separately analyzing individual protein standards).
[FU] 600 500 400 300 200 100 0 25 30 35 40 45 50 (Sec) Cow Buffalo Sheep
Figure 6. Milk authenticity. The 2100 Bioanalyzer protein assay also provides a rapid means to determine product authenticity quickly. Shown here are three overlaid electropherograms of milk samples (dil. 1:10) from different sources. In this example, milk samples from cow, buffalo and sheep are differentiated quickly based on relative milk protein content.
Wheat
Because the 2100 Bioanalyzer can electrophoretically separate complex protein samples, providing both protein molecular weight and quantity, you can use pattern matching of the resulting electropherogram to determine product authenticity as an alternative (or complement to) molecular detection methods. For example, verifying seed variety and quality is important economically because high quality and desirable seeds usually command a premium price. For many milled materials such as wheat flour, protein quality and quantity determine the suitability of the flour for the intended finished product. In a single test, the 2100 Bioanalyzer can determine both the genetic identity of various wheat varieties (DNA analysis) and protein content for either variety identification or product quality control. Figure 7 shows wheat variety identification based on protein pattern.
Pattern matching can identify the particular strain of wheat, in this case Halberd.
Figure 7. Protein patterns obtained by the 2100 Bioanalyzers protein chip identify various kinds of wheat. This figure covers extracts (1% SDS + 1% DTT) of six wheat varieties: Arrino, Tatiara, Rosella, Yanac, Sunbri and Sunsoft. The elution profiles from size-based capillary electrophoresis appear on the left, corresponding to the simulated gel patterns at right. Source: S. Uthayakumaran, I.L. Batey, C.W. Wrigley, Value Added Wheat CRC and Food Science Australia, North Ryde, Australia.
Compact, robust and simple to use

To learn more about the Agilent 2100 Bioanalyzer and additional food applications, go to: http://www.agilent.com/chem/labonachip http://www.agilent.com/chem/foods and click Bioanalysis and Applications You can also call 1-800-227-9770 (in the U.S. and Canada) or contact your local Agilent representative or Agilent Authorized Distributor.
This information is subject to change without notice. Agilent Technologies, Inc. 2007 Printed in U.S.A. March 6, 2007 5989-6306EN
Determination of PCR-RFLP Profiles for Fish Species Using the Agilent 2100 Bioanalyzer Application
Food Safety
Authors
Steve Garrett and John Dooley Molecular Biology Group, Dept. of Chemistry and Biochemistry Campden & Chorleywoood Food Research Association, Chipping Campden Gloucestershire, GL55 6LD UK
There is, therefore, a need to have reliable and simple species identification methods to support enforcement and compliance with labelling legislation (EC Council Regulation No. 104/2000 and EC Commission Regulation No. 2065/2001). Methods of fish species identification based on morphological characteristics are suited to whole fish; however, fish species identification becomes more problematic once it is processed. Protein profiling is used for fish identification; however, this method requires the analysis of species reference materials alongside the samples and is less reliable when applied to processed food products as the proteins become denatured. DNA based methods offer an alternative approach to species identification as DNA remains detectable in all but the most heavily processed samples. Direct sequencing is the most definitive method of identification; however, it cannot easily be applied to samples suspected or known to contain more than one species. Alternative techniques, using polymerase chain reaction (PCR), were developed to identify fish species based on DNA fingerprint patterns. Methods used include RAPDs (random amplified polymorphic DNA), SSCP (single strand conformation polymorphism) and PCR-RFLPs. A PCR-RFLP technique, which involved digesting an amplified 464bp region of the cytochrome b gene with restriction enzymes to generate DNA profiles, was previously developed for the identification of salmon species [1, 2].
Abstract
This application note shows how the Agilent 2100 Bioanalyzer was used in polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) fragment analysis for fish species identification. A 464bp cytochrome-b target sequence, found in all vertebrate fish, was first amplified and then digested with restriction enzymes. The fragments were resolved on the DNA 500 LabChip, allowing simple comparison with authentic species profiles due to the accuracy of fragment size determination. Use of the Bioanalyzer offered significant benefits over traditional gel electrophoresis and DNA staining techniques for PCR-RFLP fragment analysis.
Introduction
The diversity of fish products available to consumers has increased significantly in recent years. Products can range from premium grade fish steaks to low cost fish fingers. As fish are caught, processed, and distributed by a global network of operators, there is a need to ensure the authenticity and the origin of fish used in the products.
The aim of this work was to improve the method for identification of salmon and other species by replacing conventional gel electrophoresis and staining steps with the Agilent 2100 Bioanalyzer. The generation of species-specific PCR-RFLP profiles on the 2100 Bioanalyzer combined with accurate sizing and quantification of individual DNA fragments, offered significant advantage over gelbased approaches in terms of ease-of-use, speed, and accuracy of identification. Materials and Methods All chemicals used for this work, unless otherwise stated, were supplied by Sigma-Aldrich and were of molecular biology grade or equivalent. PCR primers were supplied by MWG-Biotech UK Ltd. PCR-RFLP profiles were generated using a DNA500 LabChip and the Agilent 2100 Bioanalyzer. AmpliTaq Gold DNA polymerase from Applied Biosystems was used in all PCR reactions. Restriction enzymes were obtained from New England Biolabs and used per the manufacturers instructions. Fish Samples Commercially important salmon and white fish species samples were obtained from appropriate fishery research laboratories in the UK, Canada, Alaska, New Zealand, and Japan. Five individuals were used to minimize the effects of polymorphic variation within the population. Additional samples of each fish species were obtained from local UK fishmongers and retailers. DNA Extraction DNA extraction was performed using a modification of the CTAB method. Samples (2 g wet weight) were suspended in 5 mL of CTAB buffer (2% CTAB [hexadecyltrimethylammonium bromide], 100-mM Tris-HCl, 20-mM EDTA, 1.4-M NaCl, pH 8.0) and 40 L of Proteinase K solution (20 mg/mL) was added. Samples were mixed thoroughly and then incubated overnight at 60 C. After incubation, 1 mL of supernatant was transferred to a 2.0-mL Eppendorf tube, cooled to room temperature (RT), and centrifuged at 13,000g for 10 minutes. The clear supernatant was recovered and an equal volume of chloroform was added. The solution was vortexed and then centrifuged at 16,000g for 15 minutes before the upper aqueous layer was transferred to a clean 1.5-mL Eppendorf tube. An equal volume of isopropanol was added and the DNA precipitated at RT for 30 minutes. DNA was pelleted by centrifugation at 16,000g for 15 minutes,
washed in 70% ethanol and air dried for 30 minutes at RT. The DNA pellet was resuspended in 100 L of sterile distilled water (SDW) and purified using Promegas Wizard Purification Resin per the manufacturers protocol. DNA extracts were recovered in 50 L of 1TE (10-mM Tris-HCl, pH 7.4, 1-mM EDTA, pH 8.0) buffer. Final DNA concentrations (ng/L) were determined using a GeneQuant pro DNA calculator (Pharmacia). DNA Amplification PCR products (464bp target from the cytochrome b gene) were produced by amplification of DNA extracts (50 ng) in 20-L reactions containing 1 AmpliTaq Gold PCR buffer (Applied Biosystems), 300 nM of each primer (L14735: 5'-AAA AAC CAC CGT TGT TAT TCA ACT A-3 and H15149: 5'-GCI CCT CAR AAT GAY ATT TGT CCT CA-3'), 200-nM dNTPs, 5-mM MgCl2 and 0.05-U/L of AmpliTaq Gold (Applied Biosystems). Amplification profiles (94 C for 5 minutes [denaturation]; 40 cycles of: 94 C for 40 s, 50 C for 80 s, 72 C for 80 s [amplification]; 72 C for 7 minutes [final extension] were applied using a PE9600 PCR machine (Applied Biosystems). Unpurified PCR products (1 L) were applied to the 2100 Bioanalyzer to confirm amplification. PCR-RFLP Profiling Unpurified PCR product (2.5 L) was digested for 3 or more hours with two to five units of enzyme in a total volume of 5 L. Reactions were terminated by incubation at 65 C for 10 minutes. Digested PCR products (5 L) were mixed with 5 L of 20-mM EDTA, to achieve a final concentration of 10-mM EDTA, prior to loading on to DNA500 LabChips. Aliquots (1 L) of the reaction mix were analyzed on the 2100 Bioanalyzer, per manufacturers instructions.
Results
Evaluation of PCR-RFLP Profiles Generated on the 2100 Bioanalyzer for Species Identification An initial evaluation of the PCR-RFLP method was performed using salmonid species. Following cleavage of the amplified DNA fragment with restriction enzymes, species-specific PCRRFLP patterns were resolved on the 2100 Bioanalyzer. An example of a PCR-RFLP pattern is given in Figure 1, which shows PCR-RFLP profiles for salmon and trout samples generated with enzymes DdeI and HaeIII.
DdeI
S1 L 1 2 3 S2 4 5 T 6 7 S1 8
HaeIII
S2 9 10 11 T 12
Observed and expected fragment sizes for a selection of five enzymes and four salmonids are shown in Table 1. As can be seen, patterns were similar to those reported previously [1, 2]. Expected DNA fragments of greater than about 25bp were readily detected; however, some smaller fragments were not detected because they could not be distinguished from the lower 15bp size marker or were outside the detection range (25bp500bp) of the LabChip. Small DNA fragments (25bp100bp), which were not reported previously, were observed in some digests [2]. This highlights the improved band resolution of this method in comparison to gel electrophoretic methods. This improved resolution is also highlighted by profiles generated with DdeI, where all four species have an extra fragment that is about 9bp smaller than the expected larger fragment. This is due to an extra DdeI site situated in primer H15149 Table 1 shows that O. gorbuscha and O. mykiss profiles generated with NlaIII have only two fragments when three are expected. This is believed to be due to the comigration of the two larger fragments. An analysis of the sequence of the O. mykiss 464bp amplicon indicated that cleavage of the amplicon with NlaIII should produce fragments of 192bp, 180bp and 91bp. These respectively equate to the 210bp, 190bp and 100bp fragments reported by Russell et al. (2000) and appear in Table 1. From the sequence data there is
600bp
15bp
Figure 1.
PCR-RFLP patterns obtained from salmon and trout with enzymes DdeI and HaeIII. PCR-RFLP patterns obtained when amplified DNA from two salmon (S1, S2) and one trout (T) samples were digested with DdeI (lanes 16) or HaeIII (lanes 712). A 15bp600bp ladder (L) is shown. All wells contain 15bp and 600bp size markers. DNA fragments of note are indicated (arrows).
Table 1.
Expected and Observed PCR-RFLP Fragment Sizes Obtained with Five Restriction Enzymes and Four Salmonid Species Expected* (E) and observed (O) fragment sizes for each enzyme (bp) DdeI** Bsp1286I HaeIII NlaIII E O E O E O E O 360, 130 353, 346, 114 360, 130 349, 343, 112 350, 130 321, 312, 110 360, 130 348, 339, 111 300, 200 289, 172 U/C*** 464 300, 200 287, 172 300, 200 279, 174 350, 130 320, 102, 35 or 421 U/C 421 350, 130 318, 98, 35 350, 130 313, 100, 33 310, 180 272, 160 210, 190, 100 181, 92 U/C 438 210, 190, 100 185, 92
Species O. nerka (Red salmon) O. gorbuscha (Pink salmon) S. salar (Atlantic salmon) O. mykiss (Rainbow trout)
Sau3AI 390, 120 340, 115 390, 120 338, 115 410, 110 370, 88 U/C 451
*Sizes as reported by Russell et al. (2000). **Extra fragments in observed DdeI profiles are due to restriction site introduced by primer H15149. ***U/C Uncut with enzyme.
no evidence that the smaller 180bp fragment contains a higher proportion of heavier A or G bases or the larger 192bp fragment a higher proportion of lighter C or T bases, which could cause their respective molecular weights to converge. The calculated molecular weight difference (3277 Daltons) between the two fragments is equivalent to the difference in the number of bases. This makes it unlikely that comigration is due to molecular weight similarities between the two fragments. The comigration of these two fragments as a single fragment is consistently observed and does not detract from the identification of these species. Overall, the profiles generated by the 2100 Bioanalyzer matched those expected or previously reported [1, 2], which supports the use of this approach for the identification of fish species. Further work was performed using Atlantic salmon and trout and a smaller number of enzymes to confirm the application of this approach. Experimental Repeatability In order to determine the experimental repeatability (LabChip-to-LabChip variability) of the complete assay, duplicate PCRs were produced from two salmon and one trout sample. Amplified fragments were cleaved with DdeI and digests stored at 4 C until required. PCR-RFLP patterns were separated on four occasions using different DNA500 LabChips primed with two different batches of gel matrix, A and B. Two LabChips were run using a freshly prepared gel matrix (matrix A) while a third LabChip was run when gel matrix A was 1-week old. The fourth LabChip was run on
Table 2.
the same date as the third LabChip but using a second, fresh batch of gel matrix (matrix B). Overall variation (encompassing variation due to LabChip-to-LabChip, PCR and gel matrix) appears in Table 2, which shows the results of analysis with the four LabChips following digestion with enzyme DdeI. Results are the mean fragment sizes observed on each LabChip from two PCR replicates of each species. Absolute fragment size variations within a single LabChip, that is, for PCR replicates of the same sample or between the two salmon samples, were less than 2bp and were only observed between the larger (>300bp) fragments. The overall absolute size variation for each fragment, which included variation due to different LabChips, PCRs and gel matrices, was slightly greater; the biggest variation was 6bp for the 320bp fragment in salmon (321bp to 327bp). Fish Species Identification Using sequence data generated from 10 different white fish species, theoretical PCR-RFLP profiles were generated from a range of restriction enzymes. Closer examination of these theoretical profiles indicated that only three enzymes would be needed to differentiate all the white fish species. Experimental profiles were generated to confirm that species identification was possible using just these three enzymes. Results of this analysis are shown in Figure 2, which shows that 9 of the 10 species could be identified based on profiles generated with only one enzyme. The remaining two species were differentiated using the other two enzymes.
PCR-RFLP Fragment Sizes Obtained Following Separation of DNA Cleaved with DdeI on Four Different DNA500 LabChips Observed fragment sizes on each LabChip (bp) LabChip 1 LabChip 2 LabChip 3 (Fresh matrix A) (Fresh matrix A) (Week old matrix A) 111 314 323 111 338 347 111 316 325 111 341 349 110 314 323 110 338 347 LabChip 4 (Fresh matrix B) 111 317 326 111 343 352
Expected band size (bp) Salmon analysis* 117 312 320 Trout analysis 116 339 348
Mean 111 315 324 111 340 349
%RSD 0.34 0.43 0.51 0.45 0.72 0.61
*An expected 27bp fragment from salmon is not detected by the 2100 Bioanalyzer.
A - Ddel
Figure 2.
C - Nlalll
Co d Ha ell l Sa m ple Co d Sa m ple Dd el
B - Haelll
Co d Sa m ple 17 Sa Ha m ell ple l 18 Sa Ha m ell ple l 2 Ha Sa ell m l ple 14 Sa Ha m ell ple l 5 Ha Sa ell m l ple 8 Ha Sa ell m l ple 9 H Sa ae m lll ple 19 Sa Ha m ell ple l 4 Ha ell l Dd el
PCR-RFLP profiles from the 10 white fish species used in this study. Profiles were generated using enzymes DdeI (A), HaeIII (B) or NlaIII (C). Each sample number indicates a different fish species.
17 Sa Nl m all ple l 18 Sa Nl m all ple l 2 Nl Sa all m l ple 14 Sa Nl m all ple l 5 Nl Sa all m l ple 8 Nl Sa all m l ple 9 Nl Sa all m l ple 19 Sa Nl m all ple l 4 Nl all l
17 Sa Dd m el ple 18 Sa Dd m el ple 2 Dd Sa el m ple 14 Sa Dd m el ple 5 Dd Sa el m ple 8 Dd Sa el m ple 9 Dd Sa el m ple 19 Sa Dd m el ple 4 Dd el
Analysis of a further nine salmon species was performed using these three enzymes only (data not shown). Again it was found that the salmonids could be differentiated using the three enzymes. A list of all the species studied is found in Table 3.
Table 3. Fish Species that Could be Differentiated Using PCR-RFLP with Enzymes Dde I, Hae III and NlaIII Latin name Gadus morhua Gadus macrocephalus Pollachius virens Melanogrammus aeglefinus Merluccius merluccius Merluccius paradoxus Pleuronectes platessa Merlangus merlangus Theragra chalcogramma Macruronus novaezelandiae Salmo salar Oncorhynchus nerka Oncorhynchus gorbuscha Oncorhynchus tschawytscha Oncorhynchus kisutch Oncorhynchus keta Oncorhynchus clarki clarki Salvelinus malma malma Oncorhynchus masou masou
Common name (UK) Atlantic cod Pacific cod Coley (Saithe) Haddock European hake South African hake European plaice Whiting Alaskan(Walleye) Pollock Hoki Atlantic salmon Red / Sockeye salmon Pink / Humpback salmon Chinook salmon Coho / Silver salmon Keta / Chum salmon Cut-throat trout Dolly Varden Cherry salmon
thin (<2 mm) acrylamide gels, to resolve the PCRRFLP patterns. This makes handling and staining difficult and requires the use of large equipment and volumes of solution. All this makes these methods potentially hazardous and time consuming and can sometimes produce variable results. This type of detection is, therefore, not suited to use in enforcement and quality control laboratories where a rapid, robust detection method is required. As an alternative, the 2100 Bioanalyzer incorporates conventional capillary electrophoresis (CE) technology into an easy-to-use chip-based format, which enables accurate sizing and quantification of individual DNA fragments. Coupled with the small (2 cm2) size of the LabChip, this gives the system a significant advantage over conventional gel-based approaches in terms of ease-of-use, speed, and safety. This makes the 2100 Bioanalyzer ideally suited to the analysis of multiple small DNA fragments such as those found in PCR-RFLP profiles. Using the 2100 Bioanalyzer it was possible to generate PCR-RFLP profiles that resembled those previously published for salmon species. However, generating PCR-RFLPs on the 2100 Bioanalyzer produced profiles with improved fragment resolution and detection, especially of smaller fragments that were not detected using conventional gel electrophoresis. PCR-RFLP profiles generated on the 2100 Bioanalyzer were also more consistently produced compared to gel electrophoresis. Consequently, profiles based on three enzymes DdeI, NlaIII and HaeIII were developed for 19 commercially important species. Further studies should increase the range of species in the database, making it an important tool for the study of the authenticity of fish products. Complete details of the development of the assays, application to fish species identification in products, and an interlaboratory study were recently published [3, 4, 5].
Discussion
To identify species present in a sample when no prior knowledge of the sample is available requires a universally applicable method. PCR-RFLP profiling of a common region of the vertebrate cytochrome b gene, which is present in all fish species, enabling comparison with profiles in a database is one such universal approach. The conventional PCR-RFLP fragment analysis involves gel electrophoresis, on large (over 30 cm2),
Acknowledgements
This work formed part of the UKs Food Standards Agency project Q01069 and the authors wish to acknowledge the FSA for their financial support.
References
1. G. L. Hold, V. J. Russell, S. E. Pryde, H. Rehbein, J. Quinteiro, M. Rey-Mendez, et al., (2001), Validation of a PCR-RFLP based method for the identification of salmon species in food products. European Food Research and Technology, 212, 385389. 2. V. J. Russell, G. L. Hold, S. E. Pryde, H. Rehbein, J. Quinteiro, M. Rey-Mendez, et al., (2000). Use of restriction fragment length polymorphism to distinguish between salmon species. Journal of Agricultural and Food Chemistry, 48 (6), 21842188. 3. J. J. Dooley, H. D. Sage, H. M. Brown, and S. D. Garrett, Improved fish species identification by use of lab-on-a-chip technology. (2005) Food Control 16 (7), 601607. 4. J. J. Dooley, M. A. Clark, H. D. Sage, H. M. Brown, and S. D. Garrett, (2005). Fish species identification using PCR-RFLP analysis and lab-on-a-chip capillary electrophoresis: application to detect white fish species in food products and an Interlaboratory study. Journal of Agricultural and Food Chemistry, 53 (9), 33483357. 5. J. J. Dooley, M. A. Clark, H. D. Sage, H. M. Brown, and S. D. Garrett, (2004) Application of a chipbased capillary electrophoresis system to enable simple PCR-RFLP identification of fish species. FSA Final Report, Q01069

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. AmpliTaq is a US registered trademark of Roche Molecular Systems, Inc. LabChip is a US registered trademark of Caliper Technologies, Inc. Wizard is a US registered trademark of Promega Corporation. Agilent Technologies, Inc. 2005 Printed in the USA July 29, 2005 5989-2982EN
Nested Multiplex Polymerase Chain Reaction for the Determination of DNA From Genetically Modified Corn and Soy Beans Using the Agilent 2100 Bioanalyzer Application
Agriculture
Author
Mark Jensen Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA
detected levels. In spite of these issues, the number of available transgenic events has continued to grow. At the current time there are 14 transgenic varieties of corn and soy beans that have been deregulated by the Animal and Plant Health Inspection Service of the United States Department of Agriculture (USDA). Although enzyme immunoassay is an efficient means for detecting transgenic proteins in raw products, only DNA analysis has proved to be effective for the entire range of sample matrices from raw materials to highly processed foods. The polymerase chain reaction (PCR) has been widely accepted as a method for the detection of DNA from genetically modified ingredients such as soya or corn [1, 2]. PCR detection of GMOs can be done either as a screening test using endpoint PCR or as a quantitative test using real-time fluorescence detection of the PCR product. Quantitation by real-time PCR is an expensive analysis requiring assay calibration for each sample lot and multiple replicates of each unknown sample. Typical service charges for a single analysis are between $150$300. Since each transgenic event must be evaluated individually, the cost of rigorously testing an unknown food sample for all the possible current transgenic events is cost prohibitive. These cost constraints make it necessary to screen samples for DNA components that are present in
Abstract
This application note describes how the Agilent Technologies 2100 bioanalyzer and the DNA 500 LabChip can be used to detect polymerase chain reaction products corresponding to genetically modified elements and endogenous sequences in corn and soy beans. The DNA extraction protocol used in the preparation of polymerase chain reaction samples was characterized using the Protein 200 Plus LabChip.
Introduction
Six years after the introduction of genetically modified organisms (GMO), consumer concerns about the presence of such modified organisms in food remains an ongoing issue. There has been a continuous debate surrounding issues of how food products that contain GMO ingredients should be regulated and labeled. This debate has been further complicated by disagreements over how GMOs should be detected and the significance of the
most GMOs prior to any quantitative analysis. This type of screening analysis can be carried out with a commercially available PCR test kit, such as the Biosmart Allin 1.0 GMO Screening System from Promega (Madison, Wisconsin). This kit provides a protocol and reagents for a nested multiplex PCR assay for the detection of DNA from modified organisms containing the 35S promoter. This genetic element is derived from the cauliflower mosaic virus and is found in most transgenic crops. In addition to the 35S promoter, the multiplex PCR reaction also detects sequences for soya (lectin), corn (zein) and an internal positive control. These multiplex PCR products are all easily resolved and detected using the DNA 500 LabChip and the Agilent 2100 bioanalyzer.
5. Two volumes of the CTAB precipitation solution (0.014 M CTAB, 0.04 M NaCl ) were added to aqueous extract and mixed. 6. The mixture was incubated at 25 C for 30 min and then centrifuged for 10 min at 15,000 g. The supernatant was discarded. 7. The precipitate was redissolved by adding 250 L of the 1.2 M NaCl and incubating 37 C for 10 min. The mixture was centrifuged for 5 min at 15,000 g and the supernatant was transferred to a new tube. 8. One volume of isopropanol was added to the mixture and then the mixture was stored at 4 C for at least 30 min. 9. The solution was centrifuged for 10 min at 15,000 g and then the supernatant was discarded. 10. The pellet was washed with 500 L of cold 70% ethanol, the mixture was centrifuged for 5 min at 15,000 g, and then the ethanol solution was discarded. 11. After drying the tube, the DNA pellet was redissolved in 25 L of Tris buffer (0.01 M Tris, 0.001 M EDTA pH 8.0). The soya reference standards were used in the evaluation of the DNA extraction because soya contains a high level of protein that can be accurately tracked through each step of the extraction procedure. Although the soya protein is not problematic for the PCR assay, it serves as a useful indicator of overall DNA purity. Protein levels in the final DNA extracts were determined by a protein measurement using the Protein 200 Plus LabChip and the Agilent 2100 bioanalyzer. Absorbance measurements were made with the ND-1000 spectrophotometer from NanoDrop Technologies, Inc. (Wilmington, Delaware). PCR Protocol Samples used in the GMO analysis consisted of Institute for Reference Materials and Measurements (IRMM) Certified Reference GMO Soy and GMO Corn, as well as commercial corn meal and soya powder. The PCR analysis was carried out using the Biosmart Allin 1.0 GMO Screening System from Promega (Madison, Wisconsin). A modified PCR cycling protocol was developed for use with the PTC 200 Peltier Thermal Cycler from MJ Research, Inc. (Waltham, Massachusetts). The PCR protocol is described below.
Experimental
DNA Extraction Three DNA extraction protocols were evaluated prior to the PCR analysis. These protocols were the DNeasy Plant Mini Kit from Qiagen N. V. (Venlo, Netherlands), the Wizard Genomic DNA Purification Kit from Promega (Madison, Wisconsin) and a cetyltrimethylammonium bromide (CTAB) precipitation procedure that was developed in-house. Detailed descriptions of the DNeasy and Wizard Genomic kit protocols can be found in the technical manuals that accompany these products. The CTAB protocol is described below. CTAB Extraction protocol: 1. A 50-mg sample was added to a 1.0-mL aliquot of CTAB extraction buffer (0.055 M CTAB (cetyltrimethylammonium bromide), 0.1 M Tris, 1.4 M NaCl, 0.2 M disodium EDTA pH 8.0), heated at 65 C for 15 min and then placed on ice. 2. The extract mixture was centrifuged at 15,000 g for 10 min to remove particulates. 3. The supernatant was removed and stored in a new microfuge tube at 4 C. (Samples with high starch levels, for example corn, were treated with 0.2 L of -amylase and incubated at 37 C for 30 min.) 4. The recovered supernatant was combined with an equal volume of chloroform mixed for 30 s and then centrifuged for 10 min at 15,000 g. The upper (aqueous) phase was transferred to a new tube.
Multiplex PCR Protocol 1. Dilute the DNA sample to an initial concentration of 550 ng/L. (This corresponds to an optical density, (OD), at 260 nm of 0.11.0 for a 1-cm cell.) 2. For the first stage of the multiplex PCR combine the following per reaction: 17-L 2X Qiagen Multiplex PCR Master Mix 27-L Allin Mix 1 5-L Internal control 1-L Extracted DNA 3. Vortex the solution. 4. Perform first stage of PCR amplification using the following cycling program. Temperature Step Time (C) 1 15 min 95 2 15 s 95 3 60 s 55 4 30 s 72 5 Repeat steps 24 an additional 39 times 6 3 min 72 5. For the second stage of the Multiplex PCR combine the following per reaction 24.5-L 2X Qiagen Multiplex PCR Master Mix 24.5-L Allin Mix 2 1.0-L Product from first stage PCR 6. Perform second stage of PCR amplification using the following cycling program. Temperature Step Time (C) 1 15 min 95 2 15 s 95 3 60 s 57 4 30 s 72 5 Repeat steps 24 an additional 39 times 6 3 min 72 Analysis of PCR Products The PCR products were analyzed using the DNA 500 LabChip and the Agilent 2100 bioanalyzer. For comparison, the same PCR products were also characterized by gel electrophoresis in a 4% NuSieve 3:1 Plus Gel. The electrophoretic separation conditions were 150 V for 60 min in a 1X Tris-borate-EDTA (TBE) buffer.
Results
DNA Extraction DNA extraction protocols were evaluated for both yield and purity of DNA. Protein levels in the DNA extracts were determined using the Protein 200 Plus LabChip [3]. The results of this comparison are summarized in Table 1.
Table 1. Method CTAB DNeasy Wizard Genomic Summary of DNA Extraction Yield (%) 0.008 0.004 0.008 DNA/Protein w/w 2.7 2.5 0.3 A260/A280 1.82 1.89 1.62
The CTAB and Wizard Genomic protocols showed the highest levels of DNA recovery with final DNA yields of 0.008% of the initial sample weight. The DNA/protein ratio of 0.3 found in the Wizard DNA extract indicated that this extract still contained a high level of protein. The corresponding ratio for CTAB was 2.7, indicating that the protein content of the CTAB extract was only 1/9 of the Wizard protocol. The A260/A280 ratio is also a useful indicator of DNA purity, with high purity DNA having an absorbance ratio of 1.81.9. A comparison of the A260/A280 ratios for the CTAB and Wizard extracts, 1.82 and 1.62 respectively, confirmed the higher DNA purity in the CTAB process. The DNeasy extraction process produced DNA of comparable quality to the CTAB process, but at a significantly lower yield. Since the CTAB extraction protocol gave the highest yield of high purity DNA, this protocol was used to prepare the DNA extracts used in the PCR analysis. Analysis of PCR Products The Biosmart Allin 1.0 GMO Screening System is a multiplex PCR kit that is capable of generating four PCR products. These PCR products include the following: Base pairs (bp) Product 118 Soy lectin gene 150 35S Promoter 217 Internal control (corn zein sequence added to the PCR mix) 278 Corn zein gene
All four of these products can be found in either the positive or the negative control reactions using the Biosmart GMO Screening System. The DNA 500 LabChip has ample separation to resolve all of these PCR products. This resolution is illustrated in Figure 1, which shows a composite of the bioanalyzer electropherograms for the positive and negative controls.
6 1 2 3 4 5 6
Fluorescence
Low molecular weight marker Primer-dimer in negative control Soy lectin - 118 bp 35S GMO - 153 bp Internal control - 217 bp Corn zein - 278 bp High molecular weight marker
7 4 1 3 5
Migration time
Figure 1. Composite of bioanalyzer electropherograms for positive and negative controls.
The PCR analysis was carried out on corn and soya sample sets. The corn sample set consisted of IRMM Certified Reference MON810TM corn standards at the following levels: 0% MON810, 0.1% MON810, 0.5% MON810, 1.0% MON810, 2.0% MON810, and 5.0% MON810, and commercial corn meal. The soya sample set was made of IRMM Certified Reference Roundup Ready soya samples at the following levels: 0% Roundup Ready, 0.1% Roundup Ready, 0.5% Roundup Ready, 1.0% Roundup Ready, 2.0% Roundup Ready, and 5.0% Roundup Ready and commercial soya powder. Both sample sets contained a positive control with 0.5% each Bt176 corn and Roundup Ready soya, and deionized water as a negative control.
Figure 2 shows the electrophoretic gel separation and a gel-like bioanalyzer image for all of the samples listed above. The PCR products for the corn samples are shown in Figures 2A and 2C. Figures 2B and 2D show the soya results. A B
Figure 2A and 2B. 4% NuSieve 3:1 Plus Gel, 150 V for 60 min.
C
500 400
D
500 400
300 250 200 150 100
300 250 200 150 100
Figure 2C and 2D. Gel-like bioanalyzer image.

Figure 2. 2A (gel) and 2C (bioanalyzer) L) Molecular weight ladder: 501, 489, 404, 353, 242, 190, 147, 110, 89, 67, 34, 34, and 26 bps 1) 0% MON810 corn, 2) 0.1% MON810 corn, 3) 0.5% MON810 corn, 4) 1.0% MON810 corn, 5) 2.0% MON810 corn, 6) 5.0% MON810 corn, 7) Commercial corn meal, 8) Allin positive control (0.5% Bt 176 maize and 0.5% Roundup Ready soybean), and 9) Negative control (deionized water). 2B (gel) and 2D (bioanalyzer) L) Molecular weight ladder, 1) 0% Roundup Ready soya, 2) 0.1% Roundup Ready soya, 3) 0.5% Roundup Ready soya, 4) 1.0% Roundup Ready soya, 5) 2.0% Roundup Ready soya, 6) 5.0% Roundup Ready soya, 7) Commercial soya powder, 8) Allin positive control (0.5% Bt 176 maize and 0.5% Roundup Ready soybean), and 9) Negative control (deionized water).
Comparison of Gel and Bioanalyzer Response The bioanalyzer DNA 500 LabChip clearly shows superior resolution and uniformity of band location compared to the 4% NuSieve gels. The enhanced reproducibility of band location in the gel-like bioanalyzer image is readily apparent in a visual comparison of Figures 2C and 2D to Figures 2A and 2B. Both the 4% gel and the bioanalyzer have sufficient sensitivity to visualize the 35S PCR product at the minimum corn and soy reference standard levels of 0.1% GMO.
A comparison of the initial 35S level in the sample to the amount of 35S PCR product shows that the 35S GMO PCR band increases as the GMO content increases. The PCR response in both corn and soy samples appears to saturate before the maximum GMO standard level of 5% is reached. This effect can be clearly demonstrated using the bioanalyzers ability to measure PCR product concentrations. Tables 2 and 3 show the concentrations of the corn, soy, and 35S PCR products. For the corn samples, the PCR product response saturates at 1.0% GMO. In the soy samples, the saturation occurs at 0.5% GMO.
Table 2.
GMO Corn Response - Corn and 35S PCR Products 278 bp-Corn amplicon (ng/L) 2.5 3.4 6.2 3.5 5.2 4.7 6 153 bp-35S amplicon (ng/L) 0.03 1.2 3.8 4.3 4.4 4.7 5.2 Ratio of 35S/ corn amplicons 0.01 0.4 0.6 1.2 0.9 1.0 0.9
Sample 0% MON810 Corn 0.1% MON810 Corn 0.5% MON810 Corn 1.0% MON810 Corn 2.0% MON810 Corn 5.0% MON810 Corn Corn meal
Table 3.
GMO Soya Response - Soya and 35S PCR Products 118 bp-Soya amplicon (ng/L) 5.1 5.4 4.9 5.3 5.1 4.8 4 4.8 153 bp-35S amplicon (ng/L) 0.1 0.8 3.4 4.0 4.4 5.0 0.9 2.2 Ratio of 35S/ soya amplicons 0.02 0.15 0.7 0.8 0.9 1.0 0.2 0.5
Sample 0% Roundup Ready soy 0.1% Roundup Ready soy 0.5% Roundup Ready soy 1.0% Roundup Ready soy 2.0% Roundup Ready soy 5.0% Roundup Ready soy Soya powder Positive control
Although concentrations of PCR products can be determined quite accurately, some care must be exercised in the interpretation of these quantitative results. Under conditions where a correlation can be seen between the GMO content and the concentration of GMO amplicon, for example, GMO <0.5%, a rough estimate of GMO concentration can be made. However, such an estimate is only reliable if the same PCR master mix and thermocycler are used for all the amplification reactions. In addition, sample and calibration standard matrices must also be highly similar. For the Biosmart GMO Screening System, quantitative conclusions cannot be made at concentrations >0.5% GMO because the PCR response is saturated. Once the PCR product concentrations have reached this level, small differences in amplicon concentrations are not useful in making quantitative estimates. An example of this can be seen in Table 3. The 0.5% GMO soy reference has a 35S/soy amplicon ratio of 0.7. In the positive control, where the sample has 0.5% GMO soy and 0.5% GMO corn, the corresponding ratio is only 0.5. Since the positive control contains at least twice as much 35S GMO marker as the 0.5% soy reference, this ratio would be expected to be greater than 0.7. It is difficult to explain why the ratio is lower in the positive control. This behavior could be the result of the complex reaction kinetics in a multiplex nested PCR assay or may simply reflect the imprecision in endpoint PCR amplicon concentrations. PCR Product Composition The internal control in the Biosmart Allin 1.0 GMO Screening System is a 217 bp corn sequence that uses the same primer sequences as the corn PCR product. According to the manufacturer, when high levels of corn DNA are present, competition for primer may result in the loss of the internal control PCR product. An examination of the corn PCR products in Figure 2C shows this effect in that the 217 bp is either absent or visible only at trace levels. However, in the soy PCR assay in Figure 2D, the 217 bp fragment can easily be seen in all the samples except the positive control that contains corn DNA. In the corn sample set, a weak band can be seen at around 120 bp. This suggests that during the DNA isolation step, trace amounts of soya DNA were introduced into the corn samples. Likewise, the
presence of a band at 280 bp in some of the soy samples indicates low levels of corn DNA were present in several of the soy samples. It is not surprising that trace levels of cross contamination should be apparent in a nested PCR assay. Since all of the samples undergo a net 80 cycles of amplification, even a few copies of soy or corn DNA will be sufficient to produce a detectable PCR product. In PCR assays using a large number of amplification cycles, it is not uncommon for amplicons of similar sequence to cross hybridize. In the case of the Biosmart Allin 1.0 GMO Screening System, both the corn PCR product and the internal control share a region of common DNA sequence. When these two amplicons cross-hybridize, the resulting product has both single-stranded and double-stranded regions. These structures, known as heteroduplexes, have substantially lower mobility than a corresponding double-stranded structure. The relative mobility shift depends on such factors as gel composition, ionic strength, and gel temperature [4]. In Figure 2, PCR products that are larger than the 278 bp corn amplicon are observed. These bands occur at about 320 bps in the 4% gel and at 400 and 500 bps in the bioanalyzer electropherogram. Cross hybridization of the corn and the internal control amplicons is probably responsible for these products.
Conclusion
This application note described the use of the Agilent 2100 bioanalyzer with the Protein 200 Plus and DNA 500 LabChip Kits in the evaluation of sample preparation and the analysis of multiplex PCR products. The Protein 200 Plus was used to determine which DNA extraction procedure was most effective in removing residual protein. The DNA 500 LabChip was used to characterize the Biosmart Allin 1.0 GMO Screening System, a nested multiplex PCR assay for the genetically modified corn and soy beans. Resolution and sensitivity in these assays was sufficient to identify all of the targeted multiplex PCR amplicons and to differentiate these targets from PCR artifacts. Sensitivity of the assay was sufficient to detect GMO content even at the minimum GMO standard level of 0.1% in both corn and soy reference standards.
References
1. S. Vollenhofer, K. Burg, J. Schmidt and H. Hroath, Genetically Modified Organisms in Food - Screening and Specific Detection by Polymerase Chain Reaction (1999) J. Agric. Food Chem. 47, 5038-5043. 2. S. Garret, O. Arun and J. Dooley, Analysis of Genetically Modified Soya Using the Agilent 2100 bioanalyzer, Agilent Technologies, publication 5988-4070EN www.agilent.com/chem. 3. D. Dempsey, and M. Jensen, Characterization of Transgenic Soybean Seedlines by Protein Expression with the Agilent 2100 bioanalyzer, Agilent Technologies, publication 5988-9441EN, www.agilent.com/chem. 4. M. A. Jensen, and N. Straus, Effect of PCR Conditions on the Formation of Heteroduplex and Single-stranded DNA Products in the Amplification of Bacterial Ribosomal DNA Spacer Regions, (1993) PCR Methods and Applications 3, 186.

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. MON810TM is a trademark of the Monsanto Company. Roundup Ready is a registered trademark of the Monsanto Company. Agilent Technologies, Inc. 2003 Printed in the USA October 23, 2003 5989-0124EN
Characterization of Transgenic Soybean Seedlines by Protein Expression with the Agilent 2100 Bioanalyzer Application
Food
Authors
Debra Dempsey E.I. DuPont de Nemours DuPont Experimental Station Wilmington, DE 19880 USA Mark Jensen Agilent Technologies 2850 Centerville Road Wilmington, DE 19808-1610 USA
importance. Characterizing the expression of these proteins in various soybean seed lines is also essential in expanding the range of soy protein applications in food. The relative levels of these two proteins have been shown to significantly impact the nutrition, taste, and texture of food products derived from soy protein extracts [1]. For this reason, soybean lines that preferentially express the 11S or 7S proteins continue to be an active target in the efforts to improve seed quality. Both conglycinin and glycinin are complex aggregates made of smaller protein subunits. -conglycinin is a 7S protein with a trimeric structure and is composed of 53, 70, and 76 kDa units. Glycinin is an 11S hexameric protein consisting of six monomer units, where each monomer is made up of 40 and 20 kDa subunits [2]. Given the sizes of protein subunits, it is relatively straightforward to characterize the levels of these two proteins by electrophoresis. The Agilent 2100 Bioanalyzer, an automated microfluidic electrophoresis platform, is well suited for the analysis of proteins in this size range. The Protein 200 LabChip has a size range of 14-200 kDa. Samples of soy protein isolate can be loaded, separated, and analyzed for relative protein composition in less than 45 minutes. In this application we describe the use of the Agilent 2100 Bioanalyzer in the analysis of soybeans, to determine the level of expression of 7S and 11S proteins.
Abstract
This application note describes how the Agilent Technologies 2100 Bioanalyzer can be used to analyze protein extracts from transgenic seedlines. Accuracy and precision in the determination of protein size and concentration was sufficiently good to allow for the characterization of experimental seedlines based solely on expressed protein profiles.
Introduction
-conglycinin(7S) and glycininin(11S) are the primary seed storage proteins in soybean, comprising about 70% of the total storage proteins. Because these proteins make up such a large portion of soya protein, they are of critical economic
Experimental
Extraction Protocol Grind the seed into a fine powder. Place a 30 mg sample in an Eppendorf tube and add 1000 L of extraction buffer (50 mM Tris [pH 7.5], 10 mM 2-mercapto-ethanol, 0.1% SDS). Agitate the mixture on a rotary shaker for 30 minutes and then centrifuge at 15,000 g for 10 minutes. Remove the supernatant and introduce the extract directly into the Protein 200 LabChip to begin the assay protocol. If the extracts contain an excessive amount of oil, the supernatant may be removed and further diluted prior to beginning the Protein 200 protocol.
Methodology To determine if a transgenic line of soybeans preferentially expressed the -conglycinin or glycinin protein groups, seed extracts were compared to extracts made from wild-type seed lines that strongly expressed either the 7S or the 11S group. Twenty extracts from the unknown seedline were prepared as described above. The protein profiles were determined by separating the proteins in the Agilent 2100 Bioanalyzer. Because of the size range of the proteins, the Protein 200 Plus chip (14-200 kDa) was used for the separation. The concentration of the individual proteins was determined by a comparison with an internal concentration marker (myosin). The ratio of 7S/11S proteins was then calculated from those values and the ratio was compared to the ratio of wild-type seedlines that preferentially expressed either 7S or 11S protein groups. Figure 1 shows the electropherogram for a control seedline, a 7S wild-type, an 11S wild type, and a representative sample from the unknown transgenic seedline. A simulated gel image of the same traces is shown in Figure 2.
Glycinin subunits (11S) -Conglycinin subunits (7S) Control
7S
11S
Unknown
20 Time (seconds)
40
Figure 1.
Electropherograms of soya protein extracts.
The ratio of 7S to 11S for the 20 sample extracts was calculated by integrating the individual components comprising the 7S and 11S groups and then determining the summed areas of the two groups. The levels of extracted protein and the 7S/11S ratios for the control 7S, 11S, and unknown groups are summarized in Table 1. The ratios determined for the high 11S and high 7S seedlines indicate the range of expected 7S/11S ratios should fall between 0.04-3.4. The ratios determined for both the controls and unknown extracts both fall within this range. All 20 unknown samples showed a higher 7S/11S ratio than the control. Average ratio values for 20 unknown extracts and 5 control extracts was 0.72 and 0.39, respectively. Measurement precision was excellent for both sample sets.
Table 1.
Summary of Extracted Protein Levels and 7S/11S Protein Ratios Extracted protein level g/mL 14,000 5,200 14,000 13,000 7S/11S Ratio 0.39 0.004 (n=5) 3.4 0.04 0.72 0.1 (n=20)
SeedLine Control 7S 11S Unknowns
Conclusions
This application note describes the use of the Agilent 2100 Bioanalyzer and the Protein 200 LabChip Kit for evaluating the relative expression of -conglycinin and glycininin in unknown seedlines. In the 20 protein extracts taken from the unknown seedline, the average ratio was 0.72 0.1. This ratio lays in range that is characteristic of high 7S expression seedlines. Given the precision of the ratio determination, it is clearly apparent that the assignment of this unknown seedline to the high 7S group is statistically significant. This conclusion is further supported by a comparison to a normal control seedline where the ratio of 7S/11S is 0.39 0.004. The Agilent 2100 Bioanalyzer together with the Protein 200 LabChip Kit are quick and efficient tools for the determination of relative protein expression. The resulting protein expression profiles are in turn a highly effective means for the characterization of new transgenic seedlines.
A B C D
Seedline for protein extraction A Control B 7S cultivar C 11S cultivar D Unknown transgenic
References:
1. Kitamura, K., (1995) "Genetic Improvement of Nutritional and Food Processing Quality in Soybean," Jpn. Agric. Res. Quart., 29(1), 1-8. 2. Yaklich, R. W., (2001) "-Conglycinin and Glycinin in High-Protein Soybean Seeds," J. Agric. Food Chem. 49, 729-735.

Figure 2. Gel simulation of electropherograms for soya protein extracts.
Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Eppendorf is a registered trademark of Eppendorf AG. Agilent Technologies, Inc. 2003 Printed in the USA May 9, 2003 5988-9441EN
Analysis of genetically modified soya using the Agilent 2100 bioanalyzer
Application
Steve Garrett zge Arun John Dooley
Abstract
In this Application Note we describe how the Agilent 2100 bioanalyzer was used to analyze multiple PCR products from Roundup Ready soya DNA. The multiplex assay assessed the effects of heat and low pH on subsequent amplification of genetically modified DNA and estimated levels of Roundup Ready soya within a sample.
Introduction
Genetically modified organisms (GMOs) and derived food ingredients are regulated throughout the European Union (EU). Legislation requires appropriate labeling of products containing GM DNA. Whilst DNA methods based on the polymerase chain reaction (PCR) are suitable for monitoring known GMOs in raw materials and processed foods, there is concern that the analytical methods are less reliable for quantification purposes. Particular concern is for processed foods, where soya and maize ingredients, which are most likely to contain GMOs, are only a minor component of the finished product. Food processing has a significant effect on the quality of DNA. Physical and chemical factors such as shear forces, heat treatment, nuclease activity and low pH will lead to degradation of the DNA. Soya is a common component of a wide range of foods, used as flour, protein isolate or concentrate. Soybeans are usually defatted by pressing and/or solvent extraction. In both processes, the soybeans are heated to 6080 C and the resulting protein meal can then be concentrated by extraction using weak acid (pH 4.5). These processes combined with further processing during product manufacture significantly reduce the quality of the soya DNA in the final product. This fragmentation
of DNA reduces the probability of PCR detection particularly if the fragment sizes are smaller than the DNA sequence that is amplified by the primers. Studies indicate that small sequences of DNA remain detectable following all but the most extreme processing conditions. In routine screening analysis for GMOs the use of small targets (<200 bp) is common. However, if there is differential degradation in these small targets (some sequences will be more susceptible to degradation than others) quantifying GMOs in processed foods using amplification of two similar but slightly differently sized targets may affect the accuracy of the results. The aim of this project was to study the effect of food processing on PCR-based amplification and quantification of GM DNA. The approach was to develop a simple model assay system to observe differences in detection when using small targets in Roundup Ready (RR) soya heat treated at low pH. The Agilent 2100 bioanalyzer was used in post-PCR analysis to measure the concentration and number of differently sized PCR products.
Results
Assay development The aim was to develop a model assay that could be used to assess the quality of DNA extracted from heat-processed soya flour samples, in particular, to investigate differences in PCR amplification between small DNA targets. A single multiplex PCR assay was developed that enabled four GM soya targets to be analyzed in a single reaction mix. Primer concentration was optimized in order to obtain four PCR products resolved by gel electrophoresis which corresponded in size to the soya lectin gene target of 80 bp, and the EPSPS (5-enolpyruvylshikamate-3-phosphate synthase) gene targets of 117 bp, 150 bp and 202 bp respectively. These latter targets are only found in Roundup Ready GM soya (Monsanto).
Although gel-based analysis enables sizing of PCR products, it cannot be used to provide accurate information on the quantity of a PCR product. Therefore, postPCR analysis was performed using the Agilent 2100 bioanalyzer, which can accurately size and quantify PCR products. Initially, this was carried out using the DNA 7500 LabChip kit. Four peaks were observed, corresponding to the PCR products within the lectin and EPSPS genes. However, the 80 bp peak from the lectin gene was not completely resolved from the alignment marker so quantification was not possible. Subsequently, post-PCR analysis was performed using the DNA 500 LabChip when it was made available and resolution of all four peaks was observed (figure 1).
Fluorescence
117 bp
80
60
150 bp 202 bp alignment marker 25 bp 80 bp
40
20
0 30 40 50
Time [sec] Figure 1 Multiplex assay for GM soya. Peaks produced by the four PCR products when analyzed with the Agilent 2100 bioanalyzer and DNA 500 LabChip kit.
60
70
80
The multiplex PCR assay was applied to DNA from RR soya flour reference materials (figure 2). The results show that there is an increase in PCR product concentration of the 117 bp, 150 bp and 202 bp products and little change in the concentration in the 80 bp product. This increase corresponds to the increase in RR content of the soya flour. No
increase in the product from the lectin gene was expected, as it is common to both the GM- and nonGM soya. It should therefore be possible to estimate levels of GM soya in an unknown material by applying the multiplex assay and comparing the ratio of lectin product to the other products in the sample with ratios produced from certified reference materials (CRMs). The assays would have to
Fluorescence 5%
117 bp
2%
150 bp
202 bp
80 bp 0.5 % 0.1 %
Migration Figure 2 Peaks produced by the Agilent 2100 bioanalyzer using the multiplex assay on CRMs containing different levels of GM soya
Sample Extract 1a Extract 1b Extract 2a Extract 2b 5% CRM 2% CRM 1% CRM
Ratio 117/80 1.50 1.45 1.40 1.54 2.10 0.74 0.41
Ratio 150/80 0.68 0.68 0.46 0.65 1.00 0.21 0.18
Ratio 202/80 1.04 1.04 0.71 1.10 0.94 0.68 0.43
Table 1 Analysis of a soya flour
be performed using a limited number of PCR cycles in order to perform the end-point detection during linear stages of amplification. The reference materials and the unknown samples would also have to be similar in nature. Analysis of a soya flour containing GM soya The multiplex assay was applied in duplicate to two DNA extracts prepared from a soya flour sample which had given a positive result when screened for a common GM promoter sequence (CaMV 35S promoter). The assay was also applied to 1 %, 2 %, and a 5 % CRMs. Ratios of each EPSPS product compared to the lectin product were calculated (table 1). The same extracts were analyzed using a real-time PCR method for quantitative determination of GM soya. Results from the real-time analysis indicated that the sample contained approximately 5 % GM soya while the multiplex assay gave ratios indicating that the sample contained between 2 % and 5 % GM soya.
The effect of heating time and pH on detection and quantification of GMDNA The multiplex PCR assay was applied to soya flour samples containing approximately 1.3 % GM soya and boiled at either pH 3.3, 4.3 or 6.7 for up to 21 minutes. For accurate determination of the quantity of each PCR product, the samples were applied to the DNA 500 LabChip. The concentration of each PCR product was calculated using the Agilent 2100 bioanalyzer software. At pH 3.3 where an effect of heating time was observed, the amount of each PCR product at each time point was compared to the amount of each product at 0 minutes (table 2). At pH 3.3, the relative amount of the 80 bp product was reduced to 48 % after 15 minutes and no product was detected at 18 or 21 minutes. After 15 minutes, the relative amounts products of 118 bp and 150 bp were reduced to 27 % and 16 % respectively and the 202 bp product was not detected. None of the products were detected after 18 or 21 minutes.
Time at 100 C and pH 3.3 (min) 80 bp 100 74 57 36 67 48 0 0
Amount of PCR product* 118 bp 100 77 58 23 33 27 0 0 150 bp 100 73 21 24 47 16 0 0 202 bp 100 67 6 15 21 0 0 0
0 3 6 9 12 15 18 21
* % product determined relative to the amount at 0 minutes Table 2 The effect of heating time on RR flour held at pH 3.3, determined using the multiplex PCR method.
To eliminate any variation due to amount of DNA in each PCR reaction, the ratio of the lectin 80 bp product to each of the other three products was determined for all experiments (table 3), that is, normalized with respect to the 80 bp product. The ratios of each would be expected to remain constant if no degradation of the tar-
get DNA occurred or if the degree of degradation between the 80 bp target and the other targets was comparable. At pH 3.3 the ratios tended to increase with increasing heating time. This suggests that at low pH there were differences in the detectability of the three EPSPS targets compared to the smaller lectin target, with the
Time at 100 C (min) pH 3.3 0 3 6 9 12 15 18 0 3 6 9 12 15 18 0 3 6 9 12 15 18 21
Ratio lectin 80bp/ RR-117bp 1.8 1.8 1.7 3 3.6 3.8 NP 2 2.2 1.3 1.3 1.5 1.8 1.9 1.8 1.7 1.2 1.5 1 1.2 1.4 1.6
Ratio lectin 80bp/ RR-150bp 3 3 7.8 5 4.4 9 NP 4.4 2.9 2 2.2 2.6 3.7 3.9 4.2 3.9 2.3 2.4 1.9 2.1 2 2.4
Ratio lectin 80bp/ RR-202bp 1.9 2.1 17.5 5 6 NP NP 1.9 1.8 1.9 2.3 2.6 2.7 3 1.7 1.6 1.5 1.7 1.3 1.3 1.4 1.4
pH 4.3
pH 6.7
NP= no PCR products observed Table 3 The effect of heating time on RR flour held at pH 3.3, 4.3 and 6.7, determined using the multiplex PCR method
80 bp target being degraded at a slower rate compared with the other targets. At pH 4.3, the 80/118 bp and 80/145 bp ratios decreased during the first 39 minutes of heating, then increased returning to the their original value, whereas the 80/202 bp ratio increased with heating time. Similar trends were observed at pH 6.7 except for the 80/202 bp ratio where little change occurred. However, further analyses are required to replicate these observations and focus around the pH where an effect is observed. These initial results indicate that the different targets used in PCR are not detected equally in these experiments. Other studies show similar results. In 1998 Hpfer et al.1 demonstrated that PCR detection of GM maize in polenta could be influenced by pH during thermal treatment of the product. They showed that detection of a 1,914 bp segment of the cry1A(b) gene was not possible after boiling at neutral pH for 30 minutes, whereas a 211 bp fragment was detected after boiling for 105 minutes. At pH 2-3, the larger segment was not detected after boiling for 5 minutes and the smaller fragment was not detected after 15 minutes. As a result of such observations, it is common practice to use small target sequences in screening methods for GMOs. However, the work reported here suggests that at low pH, degradation of DNA results in
differences in detection of very small target sequences. This may not be important for qualitative analysis, however, it is likely to have significance for the accuracy of quantitative analysis of processed foods with low levels of GM-DNA, when two target sequences are analyzed simultaneously as in real-time PCR.
Conclusion
The RR multiplex assay was used to quantify the amount of GM soya in a soya flour and assess the effects of pH and heat on the detection of GM soya DNA. A key component of the assay is the Agilent 2100 bioanalyzer which is used to accurately quantify the four PCR products simultaneously. This user-friendly instrument replaces gel based analysis and offers enormous potential for the routine screening of raw materials for levels of genetically modified organisms.
References
1. Hpfer C, Hotzel H., Sachse K., Engelk E.H. Detection of the genetic modification in heat treated products of Btmaize by polymerase Zeitschrift fr Lebensmittel Untersuchung und Forschung A., 206, 203-201, 1998
Steve Garett and John Dooley are scientists in the Molecular Biology Group, Dept. of Chemistry and Biochemistry, CCFRA Technology Ltd., Chipping Campden, Glouchester, UK. zge Arun is a scientist at the Tubitak Marmara Research Center, Food Science and Technology, Research Institute, Kocaeli, Turkey.
www.agilent.com/chem/labonachip
LabChip and the LabChip Logo are registered trademarks of Caliper Technologies Corp. in the U.S. and other countries. Copyright 2001 Agilent Technologies All Rights Reserved. Reproduction, adaptation or translation without prior written permission is prohibited, except as allowed under the copyright laws. Published November 1, 2001 Publication Number 5988-4070EN
Detecting genetically modified organisms with the Agilent 2100 bioanalyzer
Application
Anne-Cline Minvielle, Ph.D. Hafida Oujati Sverine Dameron

Genolife
Abstract
Labeling of food containing more than 1% of genetically modified organisms (GMOs) has been obligatory in Europe since January 2000. To guarantee transparency and labeling, methods to distinguish between transgenic food and their traditional counterparts must be available. Genolife developed a method to detect Ready RoundUp soy (RRS) and a multiplex PCR to detect five corn transgenes (Bt176, Bt11, MON810, T25 and GA21). The Agilent 2100 bioanalyzer and DNA 500 LabChip kit provided a simple, high throughput and standardized way to analyze multiplex PCR products. The methods developed by Genolife allow detecting 0.01% of RRS in food ingredients on the one hand and 10 copies of transgene MON810, GA21 and Bt11, and 100 copies of transgene Bt176 and T25 on the other.
Introduction
The rapid development of biotechnology has launched products and ingredients derived from genetically modified organisms (GMOs) into the food market. The general public, however, has shown anxiety about this new technology. Information and transparency regarding these products are essential in order to become accepted by the consumers. In Europe, labeling of GMOs is regulated by the Novel Food directives 258/97/EEC1 and 1139/98/EEC2. More recently, the threshold regulation 49/2000/EEC3 has been approved, specifying that foodstuffs are subject to labeling. When the proportion of an individual food component is higher than 1 % manufacturers must label their products. Moreover, the presence of GMOs must be adventitious and therefore, food manufacturers must be able to supply evidence that they have taken appropriate steps to avoid using GMOs. A key factor to guarantee transparency and labeling is the availability of methods to distinguish between transgenic food and their traditional counterparts, not only in raw materials but also in food products. Several analytical methods using polymerase chain reaction (PCR) technology have been developed to qualitatively detect the presence of a modified sequence of nucleic acid in transgenic food.4,5 But these analytical methods detect only one modified sequence of one genetically modified organism. It would be advantageous to detect more than one sequence
per genetically modified organism (one endogenous gene and several transgenic markers) or to screen several GMOs in one analysis. We therefore developed a method to detect RoundUp Ready soy (RRS) from Monsanto in one part, and another method to detect one endogenous maize gene and five genetically modified maize genes four are authorized in Europe (Bt176, Bt11, MON810 and T25) and one is non-authorized (GA21). The Agilent 2100 bioanalyzer and DNA 500 LabChip kit provided a simple, rapid and standardized alternative to analyze multiplex PCR products.

RRS detection in food ingredients DNA was extracted from commercial transgenic soybean reference standards (Fluka) and different food samples (lecithin, soybean proteins, soybean flour) with spe-
cific protocol developed by Genolife. For each sample, four PCR reactions were done one for amplification of an endogenous gene (ACC1, 115 bp) to check the quality of the extracted DNA, and three PCR reactions for specific RR soybean sequences (T1: 167 bp, T2: 141 bp and T3: 189 bp). After amplification, PCR products were mixed and 1 l of each mixed PCR was analyzed on the Agilent 2100 bioanalyzer using the DNA 500 LabChip kit, which allows analysis of DNA fragments ranging in size from 25 to 500 bp. Twelve samples were analyzed simultaneously and the 2100 bioanalyzer produced raw data and analysis in multiple formats. It displayed a simulated gel view an electropherogram. A data table labels each of the peaks and furnishes information about the size and concentration for each fragment. Results are shown in figure 1. Non transgenic soy gives
Fluorescence
2 4 6
Migration Time
Figure 1 RR soybean detection A) Gel view: 1-soya protein, 2- lecithin, 3-soybean flour, 4-RR soy 1 %, 5-RR soy 0.1 %, 6-non transgenic soy, 7-PCR blank. B) Overlay of the electrophoretic traces of lanes 2, 4 and 6.
one band corresponding to the endogenous gene (115 bp) while RR soy 1 % and RR soy 0.1 % show four bands corresponding to the endogenous gene (115 bp) and the three specific RR soybean sequences (141, 167 and 189 bp). The Agilent 2100 bioanalyzer software compares unknown samples with commercial transgenic soybean reference standards. Soybean proteins and lecithin are transgenic (four bands at 115, 141, 167 and 189 bp) and soybean flour is not transgenic (only one band at 115 bp corresponding to endogenous gene). The 2100 bioanalyzer performs quantification using an internal standard (marker) added to each sample before loading, and the software calculates the DNA concentration in each band. Soybean protein contains less than 0.1 % and lecithin more than 1 % RR soy, compared to concentration DNA in transgenic band obtained with commercial transgenic soybean reference standards (table 1). The detection limit of this RR soy PCR is 0.01 %.
GMO maize detection in food ingredients by multiplex PCR In Europe, four GMOs maize were authorized two insect-resistant corn species from Novartis (Bt11 and Bt176), one insect-resistant corn from Monsanto (MON810) and one glufosinate-tolerant corn developed by Agrevo (T25). We developed a PCR multiplex to detect these four corn lines and one endogenous gene to check the integrity of the extracted DNA. We also added a couple of primers to detect a glyphosate-tolerant corn GA21 produced by Monsanto. This glyphosate-tolerant corn is authorized in the USA and can be exported in Europe with authorized corn. The PCR multiplex amplified a 152-bp fragment for endogenous gene, a 343-bp fragment for Bt176 corn, a 149-bp frag-
ment for T25, a 199-bp fragment for MON810, a 110-bp fragment for Bt11 and a 270-bp fragment for GA21. Results are presented in figure 2. The endogenous gene was amplified in all lanes except PCR blank. Only Bt176 fragment (343 bp) was obtained when only Bt176 corn was present in PCR tube (lane 4). Specificity was checked for each corn (lane 5: Bt11, lane 6: T25, lane 7: GA21 and lane 8: MON810). Lanes 1 to 3 presented PCR products when all corn lines were analyzed together. This multiplex PCR allows detection of five corn lines present at 0.2 % each (lane 3). The detection limit of this multiplex PCR is 10 copies for transgene MON810, GA21 and Bt11 and 100 copies of transgene Bt176 and T25.
125 100 Fluorescence 75 50 25 0 30
RR1 % RR 0.1 % 141 bp 167 bp 189 bp 1 0.1 1 0.24 0.56
Soya Lecithin protein 0.12 0.48 1.3 0.34 1.3
Table 1 DNA concentration (ng/l) of the band corresponding to the transgenic markers
1*
2 34 5 6 7 50 70 Time [s]
8* 90
Figure 2 Multiplex PCR to detect GMO corn A) 1-Bt176 5 %, Bt11 2 %, T25 5 %, GA21 5 %, MON810 5%, 2-Bt176 2 %, Bt11 2 %, T25 2 %, GA21 2 %, MON810 2 %, 3-Bt176 0.2 %, Bt11 0.2%, T25 0.2 %, GA21 0.3 %, MON810 0.3%, 4-Bt176 2 %, 5-Bt11 2 %, 6-T25 2 %, 7-GA21 2 %, 8-MON810 2 %, 9-non GMO corn, 10-PCR blanc B) Electrophoregram of lane 2 (multiplex PCR with mix of 2 % corn). The peaks are: Bt11 corn (2), T25 corn (3), endogenous gene (4), MON810 corn (5), GA21 corn (6) and BT176 corn (7).
Conclusion
We developed detection methods for GMO soy and corn in food ingredients. The Agilent 2100 bioanalyzer performed the analysis of multiplex PCR product and allowed a semi quantification of GMO content. Two detection methods were presented one for RR soy and one for GMO corn detection. The sensibility of the soya detection method is 0.01 % and 10 to 100 copies of transgene for corn mulitplex detection.
3. Commission Regulation (EC) No 49/2000 of 10 January 2000 amending Council Regulation (EC) No 1139/98 concerning the compulsory indication on the labeling of certain foodstuffs produced from genetically modified organisms of particulars other than those provided for in Directive 79/112/EEC. Official Journal L 006, 11/01/2000 p.0013-0014. 4. Vatilingom M., Pijnenburg H., Gendre F., Brignon P. (1999) RealTime Quantitative PCR Detection of Genetically Modified Maximizer Maize and Roundup Ready Soybean in some representative Foods. J. Agric. Food Chem., 47, 5261-5266. 5. Zimmermann A., Hemmer W., Liniger M., Lthy J., Pauli U. (1998) A sensitive Detection Method for Genetically Modified MaisGard Corn using a Nested PCR-system. Lebensm.Wiss.u.-Technol., 31, 664-667. Anne-Cline Minvielle, Ph.D, Hafida Oujati and Sverine Dameron are Research Scientists at Genolife, Clermont-Limagne, France.
References
1. Regulation (EC) No 258/97 of the European Parliament and the Council of 27 January 1997 concerning novel foods and novel food ingredients. Official Journal L 043, 14/02/1997 p.0001-0007. 2. Council Regulation (EC) No 1139/98 of 26 May 1998 concerning the compulsory indication of the labeling of certain foodstuffs produced from genetically modified organisms of particulars other than those provided for in Directive 79/112/EEC. Official Journal L 159, 03/06/1998 p.00040007.
LabChip and the LabChip Logo are registered trademarks of Caliper Technologies Corp. in the U.S. and other countries.
Copyright 2001 Agilent Technologies All Rights Reserved. Reproduction, adaptation or translation without prior written permission is prohibited, except as allowed under the copyright laws. Published November 1, 2001 Publication Number 5988-4847EN
Development of meat speciation assays using the Agilent 2100 bioanalyzer
Application
John Dooley Steve Garrett
Abstract
The use of real-time PCR assays for quantitative PCR is becoming more frequent. During the development of such assays it is necessary to match both the PCR primers and the fluorescent probe (used for detection) in a single reaction. The probe production costs are high compared to the primer production costs. It is, therefore, useful to know that the newly designed PCR primers are functioning in an expected manner before the cost of probe production is incurred. This Application Note describes the use of the Agilent 2100 bioanalyzer with the DNA 500 LabChip kit to confirm that primer sets are suitable before the probe is finally produced.
Introduction
Campden & Chorleywood Food Research Association (CCFRA) is interested in the development of PCR based methods for food authenticity, particularly in relation to the detection and quantification of meat species in meat products. The assays must be applicable to processed foods and, therefore, use small DNA targets as the extracted DNA is often degraded. One approach is to develop real-time methods based on the ABI Prism 7700 Sequence Detection System, known as TaqMan. The method is suitable for amplicon detection in the range 60150 base pairs. It is common during the TaqMan assay development stage to find several suitable probes, each with several different primer sets. Although, theoretically, these assays should all work to consistent levels, practically there are variations between them and some assays are unlikely to work at all. Therefore, it is advisable to confirm that the primers designed
are specific (produce only a single PCR product with no primerdimerization) and will work under TaqMan conditions (high MgCl2 concentration and strict cycling parameters). These conditions are fundamental to the accurate quantification of samples. Although confirmation of primers can be performed using traditional agarose gel methods, the Agilent 2100 bioanalyzer has several advantages over the agarose methods including speed of analysis, accurate quantification of PCR yield and sizing of products. In addition, safety is improved as there is a reduced risk from handling DNA staining dyes such as ethidium bromide. Using the DNA 500 LabChip kit allows accurate sizing of small PCR products. This is advantageous for real-time PCR where the amplicon size required is small (less than 150 bp). We describe the use of the Agilent 2100 bioanalyzer to assist in the development of assays suitable for sensitive detection of one meat species in another.
Materials and Methods

Design of PCR assay TaqMan PCR assays (primers and probes) were designed using the Primer Express software (Applied Biosystems, Warrington, Cheshire, UK). Primers (forward and reverse) were designed to amplify single genomic DNA targets from pig, cow, sheep, turkey and chicken of less than 150 bp, in accordance with TaqMan design restraints. Primers were produced by MWG-Biotech, UK. Performance of PCR reaction PCR was performed in 25-l volumes using 300 nM of each primer, 5 mM MgCl2 and 100 ng of template DNA. A TaqMan-based amplification protocol (30 cycles of a two-step reaction consisting of 95 C for 15 seconds and 60 C for one minute) was applied to the reactions. PCR was finished with a final 10-minute step at 72 C.

To perform absolute quantification it was necessary to design two assay types, a species-specific assay and a total meat assay that would be suitable for all meat species.
Species-specific assay development Figure 1 shows examples of results. Figure 1A was obtained following amplification of different mammal or poultry species with a specific turkey assay. A single band was obtained with turkey (lane 5) only, i.e. there was no amplification with chicken, pork,
A
chicken
Ladder
lamb
beef
pork
turkey
Fluorescence 40 30 20 10 0 30 1* 2 40 50 3* 90 100
60
70
80
Time [seconds]
DNA 500 LabChip preparation Chips were primed according to Agilents instructions, provided with the chips. Samples (1l) of each PCR reaction were loaded onto the DNA 500 LabChip following Agilent protocols and the chips were loaded into the Agilent 2100 bioanalyzer. The analysis of the DNA products was performed using the DNA 500 protocol of the accompanying software.
chicken
Ladder
turkey
chicken
turkey
blank
Figure 1 Species specific PCR amplification of meat samples. A) With turkey specific primers only turkey samples are amplified. The electrophoretic trace confirms the high purity of the fragment. B) Different sets of primers can be designed that allow the species specific PCR amplification of lamb (lanes 1-6) or pork (lanes 7-12).
blank
lamb
beef
pork
beef
lamb
pork
beef and lamb. This band was not seen in any other poultry or mammal species. Figure 1B shows results of amplification with a set of specific mammal primers. A fragment can only be seen in the correct species, with no amplification detectable in any other species. Similar results were obtained for all five species under investigation. These results suggested that these primer sets would be suitable for individual species detection on the TaqMan. The appropriate probes were produced and the assays optimized for TaqMan usage. Species-specific amplification was observed on the TaqMan system. Total meat assay development Figure 2A shows the results of designing a total meat assay. The aim was to develop an assay that would amplify all meat species with the same degree of efficiency. The assay was also designed to show no amplification with nonmeat species, including fish. Figure 2A shows that a single band of equal intensity was observed in all five species, whether of mammal or poultry origin. Figure 2B shows the overlay of the electrophoretic traces for these PCR products. The yields for all species were similar (mean 5.43 0.64 ng/l suggesting that this assay would be suitable for developing TaqMan-based, absolute quantification protocols. No amplification was observed in non-meat samples tested, including maize, soya, wheat and fish (figure 2A lanes 6-12). Figure 3 shows results from the complete TaqMan assays where the primers and probes were combined. Poor clarity (smudging) of the bands is possibly due to the use of dUTP in the TaqMan assay as opposed to the
chicken
Ladder
fish 51
lamb turkey
beef
pork
maize
soya
wheat
fish 52
fish 53
fish 54
B
Fluorescence
30 beef pork chicken lamb turkey
20
10
0 30
1* 40
2 50 60 70 80 Time [seconds]
3* 90 100
Figure 2 Development of a meat specific assay. A) A set of primers can be designed that amplifies specifically all meat samples but does not amplify grain or fish. B) The overlay of the electrophoretic traces reveals uniform amplification levels for different meat species.
300F1:50R1 300F1:300R1 300F1:900R1
900F1:300R1
900F1:900R1
300F1:300R1
50F1:300R1
50F1:900R1
900F1:50R1
Ladder
50F1:50R1
chick
chick
lamb
beef
Ladder
water
Figure 3 PCR products using TaqMan probes. Results from assy optimization test using pork DNA. Primers (forward [F] or reverse [R]) were used at 50, 300 or 900 nM concentration. Results show that at least 300 nM of F or R primer is required for amplification. 300 nM of each primer was found to be optimal for this amplification.
Figure 4 Non-specific amplification using non-optimal sets of primers.
beef
lamb
use of dTTP, which was used in the conventional PCR reactions. An example of a primer set, although designed to be specific, is not specific in practice, is shown in figure 4. This assay was designed to amplify a single target in all meat species. As can be seen the number, size and yield of PCR fragments varies between the species. This primer pair was, therefore, inappropriate for use but having used the Agilent 2100 bioanalyzer to check the primers before purchasing the probe saved a considerable expense.
Conclusion
We believe the Agilent 2100 bioanalyzer provides a quick, visual method to confirm primer specificity and suitability for use in TaqMan assays. Although it would be possible to perform similar checks using SYBR Green DNA stains in the TaqMan machine itself, it is not possible to determine if the observed fluorescent change is due to primerdimer formation or from the target of interest. The Agilent 2100 bioanalyzer allows confirmation of this and also confirmation that only a single target of expected size is being amplified. The ability of the Agilent 2100 bioanalyzer to quantify PCR yields is useful especially if assays being designed are required for quantitative or semi-quantitative determination, or as in our case to design a single assay suitable for detecting multiple species.
The authors are scientists in the Molecular Biology Group, Dept. of Chemistry and Biochemistry, Campden & Chorleywood Food Research Association (CCFRA), Chipping Campden, Gloucestershire, UK. The authors wish to acknowledge the support of the Food Standards Agency for this work.
LabChip and the LabChip Logo are registered trademarks of Caliper Technologies Corp. in the U.S. and other countries.
Copyright 2001 Agilent Technologies All Rights Reserved. Reproduction, adaptation or translation without prior written permission is prohibited, except as allowed under the copyright laws. Published November 1, 2001 Publication Number 5988-4069EN
Productivity Tools Applications
Low-Pressure Retention Time Locking with the 7890A GC
Application
HPI
Authors
Russell Kinghorn and Courtney Milner BST, 41 Greenaway Street Bulleen, VIC 3105 Australia Matthew S. Klee Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808 USA
Most RTL-based methods were established on GC/MS systems, where a typical head pressure for a 30 m 0.25 m id column is > 10 psi. With setability to two decimal places, four-digit pressure setpoints for such columns (for example, 11.54 psig) result in excellent inter-instrument and intrainstrument RTL precision. Amongst the many optimization tasks in GC method development is deciding on the best column dimension to select. This can realistically only be determined with full knowledge of the sample characteristics and analysis goals (that is, components of interest, complexity, detection limits required, and the matrix of the sample). Larger diameter columns have the advantages of ruggedness and sample capacity over smaller dimension columns. The larger the diameter of the column, the less pressure is needed to establish optimal flows. However, for the most precise RTL, one needs the ability to set pressure very precisely. With two-decimal-place precision at low pressures (for example, 1.28 psig), locking a system to a target retention time is less precise for column dimensions, such as 0.32-m columns. The 7890 GCs fifth generation EPC provides excellent low-pressure control, and with third-decimalplace control of the pressure, providing the precision demanded for RTL at low pressures. This application explores the suitability of the 7890 for RTL at low pressures with a 320-m column and uses a translation of the method described in Agilent Technologies publication 5989-6569EN, Reliable transfer of existing Agilent 6890/5973GC/MSD methods to the new 7890/5975 GC/MSD.
Abstract
Retention time locking was introduced over a decade ago with the Agilent 6890 gas chromatograph. The next generation of GC from Agilent the 7890A has enhanced and extended the functionality previously available, including a new electronic pneumatic control system capable of pressure control to the third decimal place. This application demonstrates the ability of the EPC system to be used for retention time locking at low pressures, in this case with a 320-m column on a 5975 GC-MS.
Introduction
The introduction of retention time locking (RTL) with the 6890 GC gave users a new way of improving productivity by eliminating the need to constantly update compound retention time data whenever a column was trimmed or replaced. It also allowed the same methods to be run on multiple systems with the same retention times. The introduction of eMethods further enhanced the portability of these methods.
Experimental
The method used for this example was translated using the Agilent Method Translator software to convert to a 250-m id column method to a 320-m method. A series of standards was run from 10 ppb to 5 ppm to illustrate the performance of the method, followed by trimming approximately 45 cm off the column and relocking the method.
Table 1. Column Carrier gas RTL peak Split/splitless inlet Oven Sample MSD Method Conditions HP-5 25 m 0.32 m 0.52 m p/n 19091J-112 Helium, constant pressure mode nominal 3.761 psi Pyrene at 11.112 min 300 C, pulsed splitless 7 psi for 0.3 in, 50 mL/min purge at 0.75 min 55 C (1.1 min) to 320 C (6.5 min) at 22.88 C/min 1-L injection, PAH 0.01 to 5 ppm concentration range Scan 45400 u Samples = 22 Autotune EM + 200V Source = 230 C Quad = 150 C Transfer line = 280 C
peaks of interest, and relocking must be effective or peak identification will fail on the different column. Samples were injected in triplicate to measure the retention time precision of the locked method at different concentrations from 0.01 to 5 ppm. Table 2 shows the performance metrics for the retention time and also the correlation co-efficient for the analysis. The initial retention times in Table 2 were achieved at a head pressure of 3.761 psig. This setpoint was determined from the retention time calibration and relocking process to be appropriate to achieve the target retention time of 11.112 min for the locking compound pyrene. Table 3 shows the calibration data from the RTL runs.
Table 3. Run RTLOCK1.D RTLOCK2.D RTLOCK3.D RTLOCK4.D RTLOCK5.D Maximum deviation Correlation co-efficient RTL Data Pressure (psi)* 3.01 3.38 3.76 4.14 4.51 Retention Deviation time (mins) (seconds) 11.212 6.018 11.166 3.264 11.112 0.000 11.070 2.508 11.020 5.508 6.018 seconds 0.999

The test sample with 16 polynuclear aromatic hydrocarbons (PAHs) was chosen for this example as it covers a wide range of physical properties and provides several challenging separations. This requires the retention time precision to be as tight as possible to ensure correct identification of the
* Even though pressure setpoints used for RTL calibration need only be to be to two decimal places, the ability to precisely set locking pressures based on the calibration curve requires setability to the third decimal place.
Figure 1 shows the calibration curves corresponding to the linearity metrics summarized in Table 2.
Table 2.
Retention Time Precision of Low-Pressure RTL Average retention time RSD (%) 0.055 0.045 0.044 0.048 0.089 0.044 0.031 0.090 0.094 0.046 0.045 0.069 0.035 0.051 0.066 0.089 Calibration linearity (r2) 0.997 0.997 0.997 0.997 0.997 0.996 0.996 0.996 0.996 0.997 0.995 0.995 0.996 0.996 0.995 0.997 RT before relocking 5.809 7.676 7.893 8.486 9.621 9.671 11.041 11.308 12.719 12.774 14.168 14.201 14.686 16.890 16.903 17.508 Relocked RT 5.876 7.746 7.959 8.556 9.692 9.746 11.116 11.383 12.799 12.849 14.277 14.306 14.807 17.074 17.082 17.709 RT when relocked 0.008 0.007 0.009 0.006 0.010 0.006 0.003 0.005 0.009 0.006 0.008 0.008 0.008 0.014 0.016 0.016
Naphthalene Acenapthylene Acenaphthene Fluorene Phenanthrene Anthracene Fluoranthene Pyrene Chrysene Benz[a]anthracene Benz[b]fluoranthene Benz[k]fluoranthene Benz[a]pyrene Indeno[1,2,3-cd]pyrene Dibenz[a,h]anthracene Benzo[ghi]perylene 2
5.884 7.753 7.968 8.562 9.702 9.752 11.119 11.388 12.808 12.855 14.285 14.314 14.815 17.088 17.098 17.725
4500000 4000000 3500000 3000000

Response
2500000 2000000 1500000 1000000 500000 0

0 1000 2000 3000 Concentration ppb 4000 5000 6000
Naphthalene Acenapthylene Acenaphthene Fluorene Phenanthrene Anthracene Fluoranthene Pyrene Chrysene Benz[a]anthracene Benz[b]fluoranthene Benz[k]fluoranthene Benz[a]pyrene Indeno[1,2,3-cd]pyrene Dibenz[a,h]anthracene Benzo[ghi]perylene
Figure 1. GCMS calibration curves for 16 PAHs using 320-m id column with low-head pressure.
Figure 2 presents overlayed chromatograms for the replicate sample injections, showing the excellent precision of the replicate injections summarized in Table 2.
Abundance 3000000 2600000 2200000 1800000 1400000 1000000 600000 200000 9.60 9.80 10.00 10.20 10.40 10.60 Time 10.80 11.00 11.20 11.40 TIC: 1802024.D\data.ms TIC: 1801023.D\data.ms TIC: 1803025.D\data.ms
Abundance 3800000 3400000 3000000 2600000 2200000 1800000 1400000 1000000 600000 200000 4.00 6.00 8.00 10.00 Time 12.00 14.00 16.00 18.00 TIC: 1802024.D\data.ms TIC: 1801023.D\data.ms TIC: 1803025.D\data.ms
Figure 2.
Overlay of three replicate injections of 2 ppm standard prior to column maintenance and relocking, including a zoomed area of the four peaks from phenanthrene to pyrene. 3
A 45-cm length of column was removed from the column to simulate typical maintenance that may be performed on a column during routine use. The method was then relocked and the sample re-run to check the efficacy of relocking at low pressure. The relocked method had a resulting pressure of 3.168 psig. The change in locking pressure can, over time, provide guidance as to the extent at which column trimming can be undertaken with the need for full re-locking of the method.
Figure 3 shows an overlay of the before and after column trimming and the extent of the retention time change. Figure 4 presents an overlay of chromatograms, one of the originals and one after column trimming and relocking. The last two columns of Table 2 compare the relocked retention times of target compounds to the originals.
Abundance 1e+07 9000000 8000000 7000000 6000000 5000000 4000000 3000000 2000000 1000000 0
TIC: RELOCK1.D\data.ms TIC: 1902027.D\data.ms
9.40 9.60 9.80 10.00 10.20 10.40 10.60 10.80 11.00 11.20 11.40 Time TIC: RELOCK1.D\data.ms
Abundance 1e+07 9000000 8000000 7000000 6000000 5000000 4000000 3000000 2000000 1000000 0 4.00
TIC: 1902027.D\data.ms
6.00
8.00
10.00 Time
12.00
14.00
16.00
18.00
Figure 3.
Overlay of injection before and after column maintenance showing the extent of retention time variation, including a zoomed area of the four peaks from phenanthrene to pyrene
Abundance 1e+07 9000000 8000000 7000000 6000000 5000000 4000000 3000000 2000000 1000000 0 9.60 9.80 10.00 10.20 10.40 10.60 Time 10.80 11.00 11.20 11.40 TIC: 1901026.D\data.ms TIC: RELOCKCHECKA.D\data.ms
Abundance 1e+07 9000000 8000000 7000000 6000000 5000000 4000000 3000000 2000000 1000000 0 4.00
TIC: 1901026.D\data.ms TIC: RELOCKCHECKA.D\data.ms
6.00
8.00
10.00 Time
12.00
14.00
16.00
18.00
Figure 4.
Overlay of injections before and after column maintenance and relocking, including a zoomed area of the four peaks from phenanthrene to pyrene.
Conclusions
This application demonstrates the ability of the 7890 GC system to perform RTL at low pressures (sub 5 psi), such as those experienced when using a 320-m column in a GC-MS system. The average retention time variation before and after column maintenance for a 16-PAH mixture is less than 0.5 sec, providing high confidence in peak assignments, even with critical separations.

Precise Time-Scaling of Gas Chromatographic Methods Using Method Translation and Retention Time Locking Application
Authors
B. D. Quimby, L. M. Blumberg, M. S. Klee, and P. L. Wylie Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA
Key Words
Pesticides, GC, GC-AED, retention time locking, RTL, method translation, scalable RT libraries
the observed element content of the peak. The combination of time and element content narrows rapidly the possible compounds that could have produced the heteroatom response to a few pesticides. The element-selective detection is done with either gas chromatography-atomic emission detection (GC-AED), which can screen for all the individual elements found in pesticides, or with a combination of other selective detectors like the electron capture detector (ECD), the nitrogen-phosphorus detector (NPD), the flame photometric detector (FPD), or the electrolytic conductivity detector (ELCD). The GC-AED technique can also be used to calculate element ratios and to quantitate unknown peaks that are detected because of its equimolar element response factors. The measured element ratios can be used to further distinguish between possible identities of detected heteroatomic compounds, often resulting in a single entry as the likely identity of a given peak. With compound-independent calibration, the amount of the unknown can be calculated using element response factors generated with a different standard compound.
Introduction
Interest in the analysis of pesticide residues has been increasing recently, in part due to the discovery that some of these compounds act as endocrine disrupters. Agilent Technologies has responded to the need for rapid, accurate, and comprehensive screening analysis for pesticides by developing a method to screen for 567 pesticides and suspected endocrine disrupters. The method uses element- selective detection and a retention time locked library of retention times to find and identify pesticides in a sample.1 In the method, sample extracts are run with element-selective detection using a prescribed set of chromatographic conditions and with the column retention time locked to the retention times in a table. If any peaks containing heteroatoms are observed, the section of the table corresponding to a small time window around the observed peak is searched. The time search results are further sorted using
Abstract
Complete development of a gas chromatographic method often involves a significant amount of effort. Once a method is completed, retention time locking (RTL) can be used to implement the method and to obtain the same retention times on multiple systems. This application note describes how to use method translation combined with RTL to implement precise time-scaled versions of a method on multiple instrument types. This allows the original method to be re-used with minimal effort, while optimizing the method for a given sample type or instrument setup. In this way, the utility of the original method is extended greatly, increasing the payback on the investment in its development and optimizing its use for specific analyses. In this note, the Agilent RTL Pesticide Library method is used as an example. The steps involved in precise time-scaling of the method to different speeds, detectors, and columns are presented.
Once the element-selective screen is completed, samples that contain any suspect compounds are run on a GC with mass spectral detection (GC-MS) system that is retention time locked to the pesticide method, thus having the same retention times as the element-selective detectors. Using the possible identities generated from the element screen, the GC-MS data is evaluated to decide which (if any) of the possible identities for suspect peaks is correct. The confirmation process is simplified greatly because the element screen usually yields only a few possibilities and because the retention time in the GC-MS run is accurately known. In practice, extracted ion chromatograms for characteristic ions of each possible compound are used to determine the identity of suspect compounds. This screening method minimizes false negatives, even in dirty samples, by using element-selectivity and time in the initial screen. With elementselective detection, all compounds containing chlorine, phosphorus, nitrogen, etc. are detected. Even if a detected heteroatomic compound is not in the table, its presence is known, and it can be marked for further GC-MS evaluation. By using GCMS for confirmation, false positives are also minimized. The RTL Pesticide Library method is a good example of a method in which a substantial investment of time and material has been made. As with many methods intended for use in multiple laboratories, it would be desirable to be able to scale the method for use in different situations of sample type and instrument setup. Because the method relies on the measured retention times of 567 compounds, it would be impractical to remeasure all the retention times
whenever the method is modified, for example, to increase its speed. Method translation24 is a calculation technique developed at Agilent Technologies that allows a capillary column GC method to be translated to different chromatographic conditions. The technique calculates the required changes in inlet pressure and oven temperature ramp rates and hold times required to maintain peak elution order identical to that of a reference method. In this way, the speed of an analysis can be scaled predictably to accommodate the needs of a specific sample or instrument type. The inlet pressure calculated for the new version of a method by the method translation software is based on the assumed or nominal dimensions of the column. As such, the calculated inlet pressure will provide a close, but not exact, match to the desired scaled retention times. To match precisely the retention times of the scaled method to the desired scale factor, the new method must be retention time locked. Retention time locking3 (RTL) is a technique developed by Agilent Technologies whereby the inlet pressure required to match retention times precisely is calculated from a calibration curve of inlet pressure versus retention time. Using method translation followed by RTL allows a method to be scaled by a precisely known factor. Once the chromatography has been scaled, a retention time table, such as the RTL Pesticide Library, can then be scaled by the same factor, resulting in a new library whose retention times match those of the scaled method precisely. The steps required to scale the method are: 1. Determine the desired scale factor for the new method.
2. Use the method translation software4 to calculate the inlet pressure and oven temperature adjustments to obtain the desired scaling of the method. The scale factor is the speed gain value reported in the method translation software. Make sure that the new method parameters are consistent with the hardware capabilities of where the new method will be used. 3. Perform the RTL calibration runs for the new method. Alternatively, the method translation software can be used to calculate the RTL calibration points for the new method using those from the original method. 4. Retention time lock the new method using the locking reference standard from the original method. The new method should be locked to the original reference standard retention time divided by the scale factor. 5. Export the retention time table as a text file using the EXPORT function in the RTL SEARCH menu of the RTL ChemStation software. 6. Divide the retention times in the table by the scale factor in a spreadsheet program like Microsoft Excel. 7. Re-import the new, scaled table. 8. Run a representative test mixture to validate the scaled method. Several examples of scaling the HP RTL Pesticide Library are presented below.
Experimental
All data were collected on Agilent 6890 Series GC systems. All systems were equipped with: Electronic pneumatics control (EPC)
Split/splitless inlet Automatic liquid sampler

Locking GC-MS with Other GC Detectors
When using selective GC detectors in conjunction with GC-MS, one problem that is encountered is knowing the relationship between retention times on the selective detector and that of the GC-MS. In GC-MS, the outlet pressure of the column is vacuum, while with most other GC detectors, the outlet pressure of the column is at or near atmospheric pressure. This difference in outlet pressures results in large differences in retention time between GC with MS detection and GC with other detectors. Comparison of GC-FID, a general detector, with GC-MS is reasonably straightforward, because the total ion chromatogram (TIC) of the GC-MS system has similar response to the FID. Retention times on the GC-MS system corresponding to those on the GC-FID can be determined by looking for similar patterns of response. With selective detectors, this is much more difficult because the response patterns from selective detectors usually do not resemble the TIC. For this reason, matching the retention times of selective detectors precisely with the GC-MS system simplifies data analysis greatly. In this first example of scaling the RTL Pesticide Library, the method will be scaled from the GC-AED method to the GC-MS method. In this case, the desired scale factor is exactly 1, that is, the GC-MS retention times are desired to be exactly the same as those of the GC-AED. The first step is to use the method translation software to determine the GC conditions to use for GC-MS.
The GC-AED system also included an Agilent G2350A atomic emission detector with GC-AED ChemStation software (rev B.00.00) for Microsoft Windows NT. The GC-micro-ECD system was controlled by Agilent GC ChemStation software (rev A.05.04). Both the GC-AED and the GC-micro-ECD ChemStations contained RTL software for GC ChemStation (G2080AA) and the Retention Time Locking Pesticide Library for GC ChemStation (G2081AA). The GC-MS system (G1723A) used consisted of an 6890 Series GC equipped with an Agilent 5973 mass selective detector (MSD). The process for retention time locking the GC-MS system is described in reference 2. All systems except the micro-ECD instrument used 30 m 0.25 mm id 0.25 m HP-5MS columns (part no. 19091S-433). The Agilent micro-ECD instrument used 10 m 0.1 mm id 0.1 m HP-5 column (part no. 19091J-141). RTL measurements were made with a solution of dichlorvos, methyl chlorpyrifos, and mirex, each at 10-ppm concentration in acetone. All injections were 1- L splitless, except for the micro-ECD experiments, which were 1- L split 100:1. In all methods, inlets were operated at 250 C and detectors at 300 C. Method translation requires inlets to be run in constant pressure mode to obtain precise scaling of retention times. Thus, all methods discussed in the note were run in this mode.
Figure 1 shows the method translation software. The original method conditions for the GC-AED pesticide method are entered in the column labeled Original Method. The column dimensions, carrier gas type, inlet pressure, outlet pressure, ambient pressure, and oven temperature program are entered here. Note that the inlet pressure is in psi (gauge), while the outlet pressure and ambient pressure are psi (absolute). The original method here is being used on a GC-AED system, so the outlet pressure is entered as atmospheric pressure plus 1.5 psi, the operating pressure of the GC-AED. The Criterion parameter is set to None, which allows the user to select a specific value of speed gain by adjusting the value of hold-up time for the translated method (see figure 1). In the column labeled Translated Method, the parameters of column dimensions, carrier gas type, outlet pressure, and ambient pressure for the GC-MS method are entered. Note that the inlet pressure and oven program are not entered; they are calculated by the program. To set the speed gain to a desired value, take the calculated value of hold-up time in the first column (0.996060 minute) and divide it by the scale factor. Because in this case the desired scale factor (speed gain) is 1, the same hold-up time for both the GC-AED and the GC-MS methods is required. Clicking the radio button next to the hold-up time in the Translated Method column will do this automatically. The method translation indicates that to obtain the same retention times on the GC-MS system as on the GC-AED, use all the same method parameters
except inlet pressure. Instead of using 27.6 psi as is used on the GC-AED, method translation calculates that 17.93 psi on the GC-MS system will result in matching retention times. As mentioned above, this inlet pressure is calculated on the assumed dimensions of the column in the GC-MS system. To get the retention times to match precisely, RTL3 is used. To retention time lock the GC-MS method to the GC-AED method in this example, it is necessary to construct an RTL calibration file for the GC-MS system. Construction of this file only needs to be done once. All subsequent users of the GC-MS method will then be able to use this calibration file for a similarly configured GC-MS instrument. The RTL calibration file is constructed by running five calibration runs of the target compound, in this case methyl chlorpyrifos, at five different inlet pressures. The runs are made at conditions identical to the nominal method except that four of the runs are made at different pressures. The pressures used are typically: Target pressure 20% Target pressure 10% Target pressure (nominal method pressure) Target pressure + 10% Target pressure + 20%
Figure 1. Method translation software showing scaling HP RTL Pesticide Method from GCAED conditions to GC-MS with a scale factor of 1. the nominal method pressure, and the retention time is observed. The pressure and resulting retention time are then entered into the (Re)Lock New Column menu item of the RTL software to calculate the correct pressure for obtaining locked retention times. Normally, the RTL calibration for a new method is determined by actually making the five calibration runs. In the current example, methyl chlorpyrifos would be run at: 17.93 psi 20% = 14.34 psi 17.93 psi 10% = 16.14 psi 17.93 psi (nominal method pressure) 17.93 psi + 10% = 19.72 psi 17.93 psi + 20% = 21.56 psi calibration data, method translation can be used to calculate the new RTL calibration points. This is useful when you want to try a scaled method rapidly and save the time required in making the five runs. (Note: For methods that will be used extensively, the five-runs approach may provide a somewhat better calibration. It is recommended that for these methods, the standard calibration be performed.) To calculate the five RTL calibration pairs of pressure and retention time for the GC-MS method from those of the GC-AED method: Take the inlet pressure used for each original GC-AED RTL calibration run, and enter it into the method translation software for the inlet pressure of the original method. Make sure the hold-up times are locked, giving a speed gain of 1.
The retention time of the target compound is determined for each run. The resulting set of five pressures and corresponding retention times is then entered in the RTL calibration dialog box for the method and saved with the method. To lock the method on the GC-MS setup, the target compound is run at
However, because the new GC-MS method is scaled from an existing GC-AED method that already has RTL
The inlet pressure calculated in the Translated Method column will now change to a new value, corresponding to the pressure that would be obtained if the calibration run were made on a GC-MS system. This pressure is used with the retention time obtained for the corresponding GC-AED calibration run as a calibration point for the GC-MS method.
When all five points have been calculated in this way, they are entered into the RTL calibration dialog box for the GC-MS method and saved with the method. Table 1 lists the original RTL calibration pressures and times with the calculated pressures and times for the GC-MS method. To test the accuracy of using a predicted RTL calibration file for GC-MS, a real calibration set was measured on the GC-MS system. The data is shown in the first two columns of table 2. (Note: The calibration points are spaced ~ 5% apart in pressure instead of the typical 10%.) A GC-MS RTL calibration file was constructed with these measured points. For each point, the locking pressure required to lock the method was calculated and is shown in column 3 of table 2. The locking pressure is the pressure determined by the RTL software that would make methyl chlorpyifos have a retention time of 16.596 minutes. This is determined by entering the pressure and retention time for each point into the (Re)Lock New Column menu item of the RTL software. If the calibration is done correctly, the locking pressures determined from each point should be very similar, as they are in column 3 of table 2.
Column 4 of table 2 shows the locking pressures for the same set of runs but determined using the GC-MS RTL calibration points calculated using method translation. The calculated data provide locking pressures that agree well with those based on measured data. The range in locking pressures pressure is only from 17.72 to 17.75 psi. This range of 0.03 psi corresponds to only about a 0.006-minute range in the retention time of methyl chlorpyrifos. Figure 2 shows the locked chromatograms from a threecomponent mixture run on GC-AED and GC-MS systems. As can be seen, the retention times are well matched between the two methods.
The RTL Pesticide Library contains the retention times of the 567 pesticides measured with GC-FID. The values measured with the FID would be the same observed with any detector that is operated at or near atmospheric pressure. Because retention time matching is critical in this application, the retention times for all the compounds in the table were also measured on the GC-MS system after scaling as described here. Figure 3 is a plot of the difference between the retention times measured on the GC-FID and the GC-MS systems. The plot shows the retention times match well within 0.1 minute out to 30 minutes. A few compounds at the end deviate outside this window, with one compound 0.2-minute different. The
Table 1.
RTL Calibration Points from Original GC-AED Method and Calculated Points for GC-MS
GC-MS RTL Calibration Calculated Pressure (psi) 24.27 21.18 17.934 14.654 11.449 Calculated Ret Time (min) 15.346 15.919 16.578 17.338 18.242
GC-AED RTL Calibration Pressure (psi) 33.1 30.4 27.6 24.8 22.1 Ret Time (min) 15.346 15.919 16.578 17.338 18.242
Table 2.
Comparison of Locking Pressures Calculated Using Measured and Predicted GC-MS RTL Calibration Data
Locking Pressures Using Measured RTL Cal Points Pressure (psi) 17.73 17.72 17.72 17.74 17.72 Using Calculated RTL Cal Points Pressure (psi) 17.75 17.73 17.72 17.74 17.74
GC-MS Locking Runs Measured GC-MS RTL Cal Points Pressure ( psi) 20 19 18 17 16 Ret Time (min) 16.127 16.326 16.536 16.760 16.988
deviation is clearly largest in the isothermal hold region, which starts at 31.87 minutes. This effect is seen with GC-MS, but not with scaling to other atmospheric pressure detectors. While the cause is not yet clearly understood, it appears related to the vacuum outlet pressure of the GC-MS column. Although this level of matching is very good, the table includes both the GC-FID and GC-MS retention times so that smaller time windows can be used in searching unknowns.
Gaining Speed in the Same Instrument Setup

In the analysis of pesticide residues in food, there are usually only a few compounds encountered in any one sample. Because the screening method uses selective detectors, it makes sense to consider trading speed for chromatographic resolution. Selective detectors respond to only those compounds containing a specific heteroatom(s), and the chromatography only needs to resolve those compounds from each other, not from every other compound in
1
the matrix. This approach can save a significant amount of analysis time. In this example of scaling the RTL Pesticide Library, the method will be increased in speed at the expense of chromatographic resolution. The first consideration is by what factor to increase the speed. The method translation software is useful for determining this. A candidate speed gain, in this example threefold, is entered into the method translation software. The resulting inlet pressure and oven temperature ramp rates are then inspected to see if the instrument on
Locking GC-AED with Other GC Detectors

When the method translation step is done to scale the GC-AED method to other atmospheric pressure detectors, the only different parameter to enter is the outlet pressure. The outlet pressure for the GC-AED method is 16.2 psi and that for the others is 14.696 psi. The method translation calculates that the nominal GC-AED inlet pressure of 27.6 psi would be changed to 26.29 psi for the other atmospheric detectors. This difference (<5%) is so small that it can be neglected, because corrections in this range are compensated easily by the retention time locking step. Thus, the method conditions and RTL calibration points used with GC-AED are interchangeable with FID, NPD, ECD, FPD, and other atmospheric detector methods. Note that this would not always be the case. If for example, a method is being scaled that uses a very low inlet pressure, the 1.5-psi difference in outlet pressure could become significant. It is best to check the method with method translation and see if the inlet pressure will change by >10%. If it does, it would be advisable to collect (or translate) a new RTL calibration centered around the translated nominal inlet pressure.
GC-MS
2
10
15
20
25
30
35
GC-AED
10
15
20
25
30
35
40 min
Figure 2. GC-AED chlorine and GC-MS TIC chromatograms of three-component locking mixture. Peak identifications: 1. dichlorvos, 2. methyl chlorpyrifos, 3. mirex.
0.300 0.200 Difference (mm) 0.100 0.000 0 -0.100 -0.200 -0.300 Retention Time (min) 5 10 15 20 25 30 35 40 45
Figure 3. Difference plot of GC-MS and GC-FID retention times in RTL Pesticide Library.
which the new method will be run is compatible with those parameters. Figure 4 shows the method translation software with the data entered for a speed gain of 3. Note that columns for Original Method and Translated Method are set up as in the previous example with two exceptions. Because the scaling is from GC-AED to GC-AED, the outlet pressure in both columns is entered as 16.2 psi. The second and most significant difference is the holdup time. The desired speed gain is 3. To set the speed gain, the calculated value of hold-up time in the first column (0.996060 minute) is divided by exactly 3. This value (0.33202 minute) is entered for the hold-up time in the second column. This will force the speed gain to exactly 3. The inlet pressure and oven temperature ramp for the new threefold speed method are now calculated. The calculated inlet pressure is 87.862 psi, which is compatible with the EPC module on the current system (maximum 100 psi). Note that the helium source supplying the GC must be capable of reaching 100 psi of helium. An optional 150-psi EPC module is available for the HP 6890 GC to provide additional inlet pressure, if necessary. The oven temperature program calculated for the new method has the first ramp listed as 75 C/min. This ramp rate is compatible with the 240-V oven option on the current instrument but would not work with a 120-V oven, which is limited to about 50 C/min in this temperature range. With a 120-V oven, the speed gain would be limited to about 2. The next step is to calculate the RTL calibration points from the original
GC-AED method. This is done by the same process as shown in the GC-MS scaling above. In this case, when one of the original method RTL calibration pressures is entered, the resulting holdup time must be divided by 3 and entered for the holdup time in the Translated Method column. This will force the speed gain back to 3. The resulting inlet pressure is then paired with the retention time of the corresponding original GC-AED calibration run, but divided by 3 as a calibration point for the new method. Table 3 shows the RTL calibration points from the original GC-AED method and calculated points for the threefold speed gain (3 ) method. When the calibration data is entered into the RTL calibration dialog box, the target time for methyl chlorpyrifos is entered as 5.532 minutes, which is 16.596 minutes divided by 3.
Table 4 compares the locking pressures determined with measured and with calculated RTL calibration points. As in the above GC-MS example, the range of the locking pressures from the calculated data is only 0.11 psi (87.88 to 87.99), which corresponds to ~ 0.003 minute. Figure 5 compares the chromatograms of the RTL locking mixture from both the original and the 3 scaled methods. Note that while the chromatographic resolution is reduced, the speed is increased by a factor of 3. Figure 6 shows a plot of the difference between the RTL Pesticide Library retention times, divided by 3, and those of the 3 method. The data were taken with a 36-component subset of the library. The plot shows the retention times match well within 0.05 minute for all compounds, even
Figure 4. Method translation software showing scaling RTL Pesticide method scaled to threefold faster method.
those in the 3.3-minute hold time at the end of the run.
Table 3.
RTL Calibration Points from Original GC-AED Method and Calculated Points for Threefold Speed Gain (3 ) Method
3x GC-AED RTL Calibration Calculated Pressure (psi) 106.21 97.23 87.86 78.44 69.31 Calculated Ret Time (min) 5.115 5.306 5.526 5.779 6.081
GC-AED RTL Calibration
Gaining Speed with a Small-Bore Column

In the previous example, speed was gained at the expense of resolution. In this example, speed will be gained while maintaining most of the resolution but sacrificing capacity. This is done by scaling the original method to a 0.1-mm id column. In scaling to columns of a different diameter, there are two important considerations that must be obeyed to obtain precise matching to a library or reference method. The first is that the stationary phase composition must be the same as that used in the original method. The second is that the phase ratio of the column being scaled to must be the same as that of the reference method. Columns of the same phase ratio have the same ratio of inner diameter to film thickness. Because the reference method was developed on a column with 0.25 mm id 0.25 m film thickness, scaling to a 0.1-mm id column will require a 0.1- m film thickness. A 10-m column of these dimensions was chosen for this example. The micro-ECD for the 6890 GC is extremely sensitive, with detection limits in the low femtogram range for polyhalogenated pesticides. These detection limits are so low that it is reasonable to consider using split mode for a rapid screening method. Using split mode with a split ratio of 100 still gives a detection limits in the range of a few picograms. The split is also more compatible with the relatively low capacity of the column.
Pressure (psi) 33.1 30.4 27.6 24.8 22.1
Ret Time (min) 15.346 15.919 16.578 17.338 18.242
Table 4.
Comparison of Locking Pressures Calculated Using Measured and Predicted 3 GC-AED RTL Calibration Data
Locking Pressures Using Measured RTL Cal Points Pressure (psi) 87.99 87.94 87.99 87.99 87.97 Using Calculated RTL Cal Points Pressure (psi) 87.99 87.95 87.99 87.96 87.88
3x GC-AED Locking Runs Measured 3x GC-AED RTL Cal Points Pressure ( psi) 97 92 87 82 77 Ret Time (min) 5.319 5.433 5.557 5.689 5.832
3 1
GC-AED (1x)
2
10
15
20
25
30
35
40 min
GC-AED (3x)
10
12
min
Figure 5. Chlorine chromatograms from original and 3x GC-AED methods of three-component locking mixture. Peak identifications: 1. dichlorvos, 2. methyl chlorpyrifos, 3. mirex.
Difference (min)
Figure 7 shows the method translation from the GC-AED method to the 0.1-mm id column with a scale factor of 3. A speed gain of 3 was again chosen based on oven and inlet limitations as described above. The same scaling process as used above is followed. The RTL calibration points for the new 3 0.1-mm micro-ECD method were both calculated with method translation and measured. Table 5 shows the calculated values. When the locking pressures from the measured and calculated values were examined, the calculated values provided much poorer predictions of locking pressure than expected. The pressure required to actually lock the column was confirmed to be 65.95 psi, as predicted by the measured RTL calibration data. Method translation had predicted the inlet pressure would be 58.514 psi for an assumed 10-m column length. Because the actual locking pressure was noticeably higher, this suggests that the actual column length was longer and/or the column diameter was smaller and/or the film thickness larger than the assumed values. As an experiment, it was assumed that the problem was in the assumed length of the column used in calculating the RTL calibration points. The column length entry for the 0.1-mm column was iteratively adjusted until the calculated inlet pressure matched the actual locking pressure, 65.95 psi. This resulted in a calculated column length of 10.5622 m. A new set of calculated RTL calibration points were calculated using 10.5622 m as the length of the 0.1-mm column. The results are shown in table 6.
0.2 0.15 0.1 0.05 0 0 -0.05 -0.1 -0.15 -0.2 Retention Time (min) 2 4 6 8 10 12 14
Figure 6. Difference plot of RTL Pesticide Library (GC-FID) retention times divided by 3 minus 3 GC-AED retention times for 36-compound subset of the library.
Figure 7. Method translation software showing scaling RTL pesticide method scaled to a threefold faster method on a 10-m 0.1-mm id column.
Table 7 shows a comparison of locking pressures calculated using measured and predicted 3 0.1-mm id micro-ECD calibration data. The range of locking pressures from the measured data (66.03 to 65.93) only corresponds to a spread in retention times of about 0.004 minute. However, with the data calculated based on a 10-m assumed length, the spread (66.38 to 63.18) is much larger and would correspond to a time range of 0.14 minute. The locking pressures calculated using the 10.5622 value are much more consistent with the measured values. The range in retention times would be ~ 0.03 minute if all the calculated points are used, and if the first value in column 5 is ignored, the range drops to ~ 0.005 minute. The fact that the agreement in locking pressures is much improved by using 10.56 m instead of 10 m suggests that length is probably the largest contributor to the discrepancy. These results should reinforce the recommendation that if a method is to be used extensively, it is prudent to obtain measured RTL calibration data. It should be noted, however, that even with the RTL calibration from the 10-m assumed length, the worst consequence would be that the RT locking step would need to be repeated an extra time to get a more precise match. Figure 8 compares the chromatograms of the RTL locking mixture from both the original and the 3 0.1-mm id micro-ECD methods.
Table 5.
RTL Calibration Points from Original GC-AED Method and Calculated Points for 3 0.1-mm id Micro-ECD Method Assuming 10-m Column Length
3x Micro-ECD RTL Calibration Calculated Pressure (psi) 71.03 64.90 58.51 52.11 45.91 Calculated Ret Time (min) 5.115 5.306 5.526 5.779 6.081
Table 6.
RTL Calibration Points from Original GC-AED Method and Calculated Points for 3 0.1-mm id Micro-ECD Method Assuming 10.5622-m Column Length
3x Micro-ECD RTL Calibration Calculated Pressure (psi) 80.03 73.13 65.95 58.74 51.75 Calculated Ret Time (min) 5.115 5.306 5.526 5.779 6.081
Table 7.
Comparison of Locking Pressures Calculated Using Measured and Predicted 3 0.1-mm id Micro-ECD Calibration Data
Locking Pressures Using Measured Using 10-m Calculated Using 10.56-m Calculated RTL Cal Points RTL Cal Points RTL Cal Points Pressure (psi) 65.95 66.03 65.95 65.93 66.00 Pressure (psi) 66.38 65.77 65.12 64.36 63.18 Pressure (psi) 65.30 65.85 65.96 65.95 65.90
3x Micro-ECD Locking Runs Measured 3x Micro-ECD RTL Cal Points Pressure (psi) 48.81 52.66 58.51 64.36 70.22 Ret Time (min) 6.323 6.041 5.797 5.585 5.396
10
Note that while the most of the chromatographic resolution is preserved, the speed is increased by a factor of 3. After being locked, the three peaks in the 3 micro-ECD method had retention times of 1.924, 5.533, and 9.963 minutes, respectively. These values are very close to the RTL Pesticide Library retention times for the three compounds divided by 3: 1.932, 5.532, and 9.949. The fact that the largest difference between the scaled table and the 3 micro-ECD method is only 0.014 minute again demonstrates the precision of retention time matching achievable with the scaling technique described here.
References
1. P. L. Wylie and B. D. Quimby, A Method Used to Screen for 567 Pesticides and Suspected Endocrine Disrupters, Hewlett-Packard Company, Application Note 228-402, Publication 5967-5860E, April 1998. 2. M. Klee and V. Giarrocco, Predictable Translation of Capillary GC Methods for Fast GC, Hewlett-Packard Company, Application Note 228-373, Publication 5965-7673E, March 1997. 3. V. Giarrocco, B. D. Quimby, and M. S. Klee,Retention Time Locking: Concepts and Applications, Hewlett-Packard Company, Application Note 228-392, Publication 5966-2469E, December 1997. 4. Capillary Column Method Translator, user contributed software, free download from: www.hp.com/go/mts.
Conclusions
Using method translation combined with retention time locking provides a means of extending the usefulness of existing capillary GC methods. The ability to precisely scale a method to meet the needs of different samples and instrument types greatly reduces the effort required to re-use methods, thus saving time and money.
3 1
GC-AED (1x)
2
10
15
20
25
30
35
40 min
GC-micro-ECD (3x)
10
12
min
Figure 8. Chlorine chromatogram from 1 GC-AED method (top) and 3 micro-ECD method (bottom) of three-component locking mixture. Peak identifications: 1. dichlorvos, 2. methyl chlorpyrifos, 3. mirex.
11
Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Microsoft and Windows NT are U.S. registered trademarks. Copyright 2006 Agilent Technologies, Inc. Printed in the USA 1/2006 5967-5820E
The 5973N inert MSD: Using Higher Ion Source Temperatures Application
Authors
Harry Prest and Charles Thomson Agilent Technologies, Inc. 5301 Stevens Creek Boulevard Santa Clara, CA 95052-8059
Tuning
Figures 1 and 2 show the results for autotuning the Inert Source at the standard 230 C ion source temperature and the 300 C temperature limit of the new source (quadrupole temperature 200 C). The higher temperature for the source produces a perfluorotributylamine (PFTBA) spectrum that shows lower abundances of the higher mass fragments, which is not entirely unexpected. The m/z 219 fragment has dropped to an abundance comparable to the m/z 69 ion and the ion at m/z 502 has dropped about 50%. This is to be expected as the internal energy of the calibrating gas has increased. Note, however, that the isotopic ratios are maintained. The user should also expect to see a higher background in the higher temperature tunes. A portion of the background will be due to ions associated with column bleed. Bleed, which usually condenses in the source, now is volatized and will appear as an increase in background and baseline.
Abstract
The new 5973N inert MSD and ChemStation software (G1701DA) offers the capability of operating the ion source at higher temperatures. This feature, combined with the improved inertness of the source, can provide the user with improvements in analysis, if exploited coherently. This application note provides advice and examples of how to explore the utility of ion source temperature.
Introduction
The default ion source temperature of 230 C is commonly applied in electron impact (EI) ionization on the 5973 MSD platforms. The new Inert Source when used with the new revision of the ChemStation software (rev. DA) allows ion source temperature to be set to a maximum of 300 C. As with all advances, there are advantages and disadvantages in operating at higher source temperatures. This note will address several general aspects in EI operation.
100 90
219
69
80 70 60 50 40 30 20 10 Mass 69.00 219.00 502.00 Abundance 382336 461504 51720 Relative abundance 100.00 120.71 13.53 Iso mass 70.00 220.00 503.00 Iso abundance 4302 19976 5073 Iso ratio 1.13 4.33 9.81
502
50
100
150
200
250
300
350
400
450
500
550
600
650
700
Figure 1.
Autotune results for an ion source temperature of 230 C.
100 90 80 70 60 50 40 30 20 10
69 219
Mass 69.00 219.00 502.00
Abundance 425024 395392 24688
Relative abundance 100.00 93.03 5.81
Iso mass 70.00 220.00 503.00
Iso abundance 4657 17000 2563
Iso ratio 1.10 4.30 10.38
502
50 100 150 200 250 300 350 400 450 500 550 600 650 700
Figure 2.
Autotune results for an ion source temperature of 300 C.
Implications for Analytical Applications

Although the tuning compound showed a spectral change that favored more fragmentation, and all compounds could be expected to be influenced similarly, there are some advantages that can occur for less fragile compounds, especially those that have higher boiling points and are late eluting in GC. Analysis of the class of compounds known as persistent organic pollutants (POPs) is likely to benefit from higher source temperatures. To illustrate the aspects that need to be examined, consider the six polychlorinated biphenyls (PCBs) acquired in full-scan and presented in Figure 3. The
overlaid reconstructed total-ion-current chromatograms (RTICCs) suggest that the higher source temperature increases the total response for the later eluting PCBs but produces little enhancement for the early eluters. This could be due to more fragmentation and may not necessarily be useful if the increase in the RTIC is due to lower mass fragments since these lower mass ions are usually compromised by interferences. A calculation of the signal/noise (S/N) for the RTICCs shows that while there is an increase in signal at the source higher temperature, there is also an increase in the background noise and the result is a lower S/N ratio for the higher source temperature.
110 100 90 Relative abundance 80 70 60 50 40 30 20 10 0 6.00 6.10 6.20 6.30 6.40 6.50 6.60 6.70 6.80 Time 6.90 7.00 7.10 7.20 7.30 7.40 7.50
300C 230C
Figure 3.
Overlaid RTICC of six PCBs acquired in full-scan (50505 amu) at source temperatures of 230 C and 300 C. From left to right, or earlier to later, in the chromatogram, the PCBs consist of a Cl3-Biphenyl, Cl4-B, Cl5-B, Cl6-B, another Cl6-B and a Cl7-B.
Figure 4 shows the same analytes acquired in selected-ion-monitoring mode (SIM) using three ions for each component (M, M+2 or M2, and M70). The same trend appears with an enhancement apparent in signal for the later eluting PCBs but little increase for the earlier PCBs. Now, however, the RTIC for the SIM acquisition does show a higher S/N ratio for these later PCBs. As opposed to the full-scan acquisition, the SIM mode acquisition at higher source temperature does increase signal for the ions of interest and, because there was no increase in background, a useful S/N increase was obtained. As always, the guiding principle that an increase in signal is only useful if
it exceeds the concomitant increase in background holds. This is clearly illustrated by the third PCB, the pentachlorobiphenyl (Cl5B). Figure 5 shows the behavior of the signal and background for the two source temperatures for one of the pentachlorobiphenyl confirming ions. The higher source temperature raises the signal and the background for this ion of interest over the lower temperature but fortunately signal increases faster than background. In this case, the background is due to column bleed components and is unavoidable but fortunately not very intense. This may or may not be the case in sample analysis.
120 110 100 90 80 Abundance 70 60 50 40 30 20 10 0 5.90 6.00 6.10 6.20 6.30 6.40 6.50 6.60 6.70 Time 6.80 6.90 7.00 7.10 7.20 7.30 7.40 7.50
300C 230C
Figure 4.
Overlaid RTICC of six PCBs acquired in SIM at source temperatures of 230 C and 300 C. From left to right, or earlier to later, in the chromatogram the PCBs consist of a Cl3-Biphenyl, Cl4-B, Cl5-B, Cl6-B, another Cl6-B and a Cl7-B.
6000 5500 5000 4500 4000 Abundance 3500 3000 2500 2000 1500 1000 500 0 6.40 6.45 6.50 6.55 6.60 Time 6.65 6.70 6.75 6.80
300C 230C
Figure 5.
Overlaid extracted ion-current chromatograms of one ion (M-70) for the pentachlorobiphenyl acquired in SIM at source temperatures of 230 C and 300 C.
The detection limits for many late eluting, highboiling compounds that will improve by implementing higher source temperatures (for example, PAHs, terphenyls, etc.). As an illustration of the enhancement for very high-boiling compounds, consider the 6-ring benzenoid hydrocarbon (PAH), coronene (CAS 191-07-1). This compound is difficult to determine due to low response and poor chromatography, although it is present in many sediment samples. Figure 6 shows overlaid RICCs for acquisitions of coronene at 230 C and 300 C. Although the peak area is the same, the enhanced Gaussian peak shape achieved at 300 C improves detection.
1800000 1600000 1400000 1200000 Abundance 1000000 800000 600000 400000 200000 0 9.38 9.40 9.42 9.44 9.46 9.48 9.50 9.52 9.54 Time 9.56 9.58 9.60 9.62 9.64 9.66 9.68
300C 230C
Figure 6.
Overlaid extracted ion-current chromatograms of one ion (m/z 300) for coronene acquired in full scan at source temperatures of 230 C, and 300 C.
Source "Bakeout"
There may be considerable temptation to use the higher source temperature for source cleaning by baking. In other words, when the user notices a higher background in the source or a reduction in response, the ill-conceived approach of baking the source clean may come to mind. The result will be that garbage coating the source will be volatized further into the analyzer; the other lenses will get dirtier, as will the multiplier, etc. Baking is not a substitute for mechanical cleaning of the source. However, baking a source after a cleaning is a good approach and a macro that provides this option is given in Table 1. After a source has been cleaned, and the MS system pumped down and checked to be leak free, this macro can be implemented either
manually or in a sequence. (Note that the temperature limits in the tune file need to be altered to 300 and 200 for source and quadrupole, respectively). Manually the bakeout is called from the command line in TOP by macro "bake.mac" <enter> bake 2 <enter> The 2 calls for a 2 hour bakeout, and which can be set to anytime the user requires. Copy the lines in Table 1 into Notepad and save the file as BAKE.MAC in the MSDCHEM\MSEXE directory. The ! indicates a comment (line) which is not executed. Note that the temperature limits, which reside in the tune file, must be edited to allow the higher settings.
Table 1.
ChemStation Macro for Baking the Source and Quadrupole After Source Maintenance
name Bake ! this macro sets the source and quad temps to their maximum and holds for a set period parameter hours def 6 ! default setting is 6 hours -this is customizable msinsctl "mstemp QUAD, , , 200" ! sets the quad temperature to bake at 200C synchronize msinsctl "mstemp SOURCE, , , 300" ! sets the source temperature to bake at 300C synchronize SLEEP hours*60*60 ! bakes for set period msinsctl "mstemp QUAD, , , 150" ! sets the quad temperature to operating temp at 150C synchronize msinsctl "mstemp SOURCE, , , 230" ! sets the source temperature to operating temp at 230C synchronize return
Usually a source cleaning is executed at the end of the working day, and the system pumped down overnight for operation the next day. In this case, a pumpdown sequence is useful. After the system is confirmed to be leak-tight, this sequence is loaded and executed which bakes the source and quad overnight, then executes an Autotune, and then makes a few injections of a checkout standard to confirm system performance. In this way, the analyst returns the next day to review data about the system prior to beginning new analyses. An example of this is given in Figure 7.
Figure 7.
Pumpdown sequence table using source bakeout.
Line 1 Loads the Bake macro. Line 2 sets the bake time to 10 hours. After the bake, (Line 3) an autotune is executed. Lines 4 and 5 run the system performance method, CHECKOUT.M, on the system checkout standard. Note: after the system has been cleaned and leak-checked, the CHECKOUT.M method should be loaded, THEN this sequence should be run!
Conclusions
The increased source temperature limit available on the 5973N inert MSD can provide improved detection limits for common, late-eluting, recalcitrant compounds such as the POPs when properly applied. A requirement, that must be explored, is that the higher source temperatures do not increase compound fragmentation or reduce the intensity of the (useful) higher mass ions. These improvements are most likely to be realized in SIM acquisitions where the increased background that must result from higher source temperatures is not as likely to affect the signal. This application note also describes a programmed bake-out of the source and quadrupole that can be automatically implemented after source cleaning. This bake-out provides a rapid lowering of the airwater background and can be used within the sequence table as part of the instrument performance checkout.

Combined EI and CI Using a Single Source Technical Overview
Chris Sandy Agilent Technologies
Introduction
The Agilent 5973x gas chromatograph/mass selective detectors (GC/MSDs) come with sources optimized for electron ionization (EI) and chemical ionization (CI). However, there are occasions where another ionization mode is desired without changing sources. This note demonstrates the capability of acquiring high-quality EI spectra with the CI source.
Data was acquired in positive CI (PCI) and EI modes. Figure 1 shows the CI and EI total ion chromatograms using the CI source. The major and minor peaks are easily comparable in the two chromatograms. Figure 2 shows the CI spectrum for Hexadecanolide (MW = 254) with the expected adduct ions for methane. Note the relatively large response for the 255 ion. As expected, there is little fragmentation due to the soft ionization.
Data Acquisition
An Agilent 5973 inert MSD with a CI source was set up for the experiments. The following process was used to tune the MS: 1. Perform the CI autotune at the normal methane reagent gas flow rate (typically at a mass flow controller (MFC) setting of 20%). 2. Reduce the CI flow to 2%. 3. Set the emission current to 250 a. 4. In Manual Tune, ramp the repeller from 05 volts for the mass 69 ion. 5. Set the repeller voltage to the maximum value. 6. Turn off the CI gas. 7. Save tune file. 8. Associate tune file with method.
Figure 1.
PCI and EI total ion chromatograms using the CI source.
Figure 2.
PCI and EI spectra for Hexadeconolide.
Figure 3.
Acquired EI spectrum compared to the NIST02 library reference spectrum.
The EI data in Figure 3 shows much more fragmentation useful for compound identification. The response for 255 is relatively small. Using the NIST02 library, the EI reference spectra for Hexadecanolide (Oxacyclohelptadecan-2-one) was retrieved with a 98% quality match.
Summary
This data demonstrates the Agilent 5973 inert GC/MSDs ability to acquire high quality EI spectra using the CI source. The EI spectra can be searched against standard libraries for identification while the CI spectra provide molecular weight information. The ability to acquire both types of data without changing sources results in increased productivity.

For more information on our products and services, visit our Web site at: www.agilent.com/chem
The author, Chris Sandy, is a GC MS Applications Specialist for Agilent Technologies in the UK.
Optimizing the Agilent Technologies 6890 Series GC for High Performance MS Analysis Technical Overview
Author
Linda Doherty Agilent Technologies, Inc. Life Sciences and Chemical Analysis 5301 Stevens Creek Blvd Santa Clara, California 95051 USA
The Carrier Gas Line The GC carrier gas should be at least 99.999% helium. Lower grades of helium are available and may contain impurities that can damage the GC column (for example, oxygen) and contribute to the chemical noise background. Even with a high purity gas there may be trace water, oxygen, and hydrocarbons. Putting a trap in the carrier line will eliminate these contaminants (see Figure 1). The mass spectrometer (MS) gas purifier trap is recommended and shipped with all new mass selective detectors (MSDs). It must be installed diagonally or vertically for optimum performance. Do not install it horizontally. High-purity helium also increases the effective lifetime of traps. Precleaned, refrigeration grade 1/8-inch copper tubing should be used with high quality carrier gas. Other tubing can be cleaned by running solvents (methanol, ethyl acetate, hexane) through it in a water aspirator vacuum setup. The use of chlorinated solvents is not recommended due to
Two-Stage regulator Main supply on/off valve Agilent 6890 Series GC Main gas supply On/Off valve Gas purifier
Abstract
Trace level GC/MS analysis requires a system that is performing at its best. Without a properly optimized GC, the mass spectrometer may not give the sensitivity expected. In other words, when more of the sample gets from the injection port to the ion source, the more likely a detector will produce a signal. Also, if the chemical noise from the GC is too high, the signal-to-noise ratio and ability to detect small analyte concentrations will be reduced. This note is a "how-to" guide for improving the GC performance. This will, in many instances, improve the overall performance of a GC/MS system. This guide is specific for the Agilent Technologies 6890 Series GC used with the Agilent Technologies 5973 MSD.
Supplies
In order to improve the system performance, some supplies should be on hand. These may not give significant improvements by themselves, but when installed together, they will give the best results. Many of the referenced supplies have changed over the past few years and will continue to be improved. It is important to stay informed and purchase the most recently updated versions of the consumables. Refer to www.agilent.com/chem for the latest comsumables available.
Figure 1.
General gas plumbing assembly. The MS gas purifier must be installed diagonally or vertically.
possible long term contamination of flow lines and controllers. Another commonly used cleaning technique is to heat the copper tubing with a Bunsen burner, propane torch, or heat gun, while helium is flowing through the tubing. This is done after connecting the tubing to the helium supply but before connecting it to the GC. This process bakes off all the volatile contaminants. Take proper safety precautions while heating the tubing. Laboratory manifold systems, especially when new, tend to have hydrocarbon contaminants. Purging the new lines, before connecting the clean tubing to analytical instruments, is essential. The supplies needed for the carrier gas line are: 99.999% He: Gas supplier Clean copper tubing (50 ft): p/n 5180-4196 Mass Spectrometer gas purifier: p/n RMSH-2 Bracket to mount the gas purifier: p/n UMC-5-2 Splitless Inlet Consumables The capillary inlet (Figure 2) has many consumable parts that should be kept on hand. Many of these consumables, such as liners (5), come in a variety of designs (Appendix A). The proper liner to use is dependent on the application. For trace level analysis, the single tapered, deactivated liner is recommended. The Viton O-ring (4) holding the liner in place should be replaced periodically to reduce the chance of leaking. The seal and the washer in the bottom of the injection port (10, 11) should be replaced whenever the reducing nut (12) is removed. The recommended seal is gold plated to reduce metal catalyzed thermal degradation of analytes. Septa (2) should be replaced quite frequently: for example, every 100 injections. The low bleed, precored, red septa should be used. Keeping a beaker of septa in an oven at 250 C at all times will eliminate the need to condition the septum once it is in place. A Merlin MicrosealTM is highly recommended over a conventional septum nut and septum (1, 2). The Microseal eliminates the need for septa and lasts for tens of thousands of injections without leaking. It is most appropriate with the Automatic Liquid Sampler (ALS) injection tower and only works with untapered, blunt tip syringe needles. The Microseal has been improved this past year. Older seals have a maximum pressure rating of 30 psi. The new seals are rated to 100 psi. The new seal and nut are recommended. The gold nuts (stamped 303C) are not compatible with the new seals. A gray nut, stamped 221B, should be installed.
1. Septum retainer nut 2. Septum 3. Insert assembly 4. O-ring 6. Split vent line 5. Liner
7. Insulation 8. Capillary inlet body 9. Retaining nut 10. Inlet base seal 11. Washer 12. Reducing nut
13. Insulation
14. Insulation cup
15. Ferrule 16. Column nut
Figure 2.
Capillary inlet assembly.
Electronic pneumatic control (EPC) is an integral part of the Agilent 6890 Series GC. The EPC version of the Agilent 6890 is required when using an Agilent 5973 MSD. The manual version of the Agilent 6890 will not work with the Agilent 5973 MSD. Electronic control gives the best repeatability in retention time and area counts. Using the electronic pulsed splitless mode allows for complete transfer of larger volume injections (up to 5 L) onto the column. Even larger volume injections (>5 L) may result in more inlet maintenance, especially with dirty samples. The Agilent 6890 does have an inlet that accommodates injections up to 250 L. It is called the programmable temperature vaporizer (PTV). It works with the ALS to deliver large volume injections (LVI). The inlet works by venting the solvent before analysis. The analyte is trapped and concentrated. It is then delivered to the column as a single plug.
The list of splitless inlet consumables is: Molded Septa (11 mm, red, 25/pk): p/n 5181-3383 Or High Pressure Merlin Microseal starter kit: p/n 5182-3442 Merlin Microseal septum: p/n 5182-3444 Merlin Microseal septum nut: p/n 5182-3445 Liner, single taper, deactivated, no glass wool: p/n 5181-3316 Viton O-ring (12/pk): p/n 5180-4182 Gold plated seal: p/n 18740-20885 Washer (to go with seal): p/n 5061-5869 10-L Blunt needle syringe: p/n 9301-0713 Column Consumables The optimal choice of column is once again dependent on the application. For trace-level, highsensitivity applications, a column with a thin film and low bleed is best. A 30 m, 0.25-mm id, 0.25-m film, 5% phenyl95% methyl silicone column is a versatile column used for many applications. Special low-bleed MS columns cost more but give better results. The proper column nut and ferrule combination are critical for a leaktight seal. Newer column nuts may not be compatible with all ferrules. The proper ferrule will be dependent on column outer diameter (od) and is specified below. The ferrule should only be slightly larger than the column od. The use of 100% graphite ferrules is not recommended, as they are easily over-tightened causing graphite to extrude into the injection port. This will be apparent when disassembling the injection port. If there are pieces of graphite in the bottom of the injection port, the ferrule(s) was/were overtightened. The presence of graphite in a hot injection port can cause thermally labile compounds to degrade. It can also affect the chromatography and cause tailing. Thus, 10% graphite, 90% Vespel ferrules are highly recommended. Vespel ferrules will shrink as they are heated. Conditioning them for 4 hours in a 250 C oven will preshrink them before use. Alternatively, the column nuts (Figure 2, number 16) can be retightened after the column oven cycles a few times.
The column, column nut, and ferrules supplies include: 30-m column, 0.25-mm id, 0.25 m, low bleed (HP-5MS): p/n 19091S-433 Column nuts (wrench-tighten only) 2/pk: p/n 5181-8830 Ferrules for 0.2-mm id columns, 10/pk: p/n 5062-3516 Ferrules for 0.25-mm id columns, 10/pk: p/n 5181-3323 Ferrules for 0.32-mm id columns, 10/pk: p/n 5062-3514 Ceramic scoring wafer (column cutter), 4/pk: p/n 5181-8836 A sharp column-cutting tool is needed for making clean cuts. The ceramic scoring wafers or sapphire square edge pens are desirable. The diamond point pens are harder to use. Ceramic scoring wafers are extremely sharp. They should be used with care. An X-ACTO or Swiss Army knife is not a column cutting tool. Use a 10 magnifier to assure that the cut is clean and no column shards are lodged inside the column. Interfacing the Column to the MS The column is connected to the MS through an interface that is sealed with a column nut and ferrule. The specific ferrule used depends on the column diameter. Never use a 100% graphite ferrule. Similar to the injection port, pieces of graphite may extrude into the interface and contaminate the MS. The ferrules required are 15% graphite, 85% Vespel. The column nut listed is brass; stainless steel nuts should never be substituted. Stainless nuts may damage the threads on the interface. Damaged threads cause air leaks and replacement of the entire interface. The MS interface supplies include: Brass column nut: p/n 05988-20066 Ferrules for 0.2 and 0.25-mm id columns, 10/pk: p/n 5062-3508 Ferrules for 0.32-mm id columns, 10/pk: p/n 5062-3506
Installation of Consumables
This section assumes that you are going to perform preventative maintenance (PM) on your Agilent 6890 Series GC. If this is a new gas chromatography/mass selective detector (GC/MSD) system, many of these steps will be completed by a Customer Engineer during installation. Before beginning the PM, please read this section carefully. The necessary manuals will be referenced frequently. These manuals can be downloaded from our Web site at www.agilent.com/chem. The manuals necessary include: Agilent 6890N GC User Information Manual: G1530-90210 Agilent 6890 Series GC Service Manual: G1530-90220 Agilent 5973 Series MSD Hardware Installation Manual: G2589-90006 Agilent 5973 inert MSD Hardware Manual: G2589-90071 Agilent 5973 Series MSD Site Preparation Guide: G2589-90070 With all of the consumable supplies previously mentioned at hand, a proper PM can be completed. To begin a PM, it is necessary to cool the GC zones (oven, inlet, MS interface). Vent the Agilent 6890 Series MSD. Please refer to the MS hardware manuals for venting instructions. Installation of Gas Supplies Follow the directions in the GC site preparation/installation manual and install the gas line supplies. Take care in making the Swagelok connections. The trap can break if too much stress is placed on the connection during tightening. Leak check all connections with a helium leak detector. (No Snoop please!) Make sure to purge all lines with helium before connecting them to the GC. Installation of Capillary Inlet Supplies Before handling any of the injection port supplies, wash hands and/or wear lint-free gloves. Oils on the hands will be transferred to these parts and become background in the system, requiring extra bakeout time. Following the instructions in the GC operating manual, remove the septum nut, septum, and liner. Discard the septum, liner, and liner O-ring. Open the oven door, loosen the 1/16-inch column nut, and remove the column and nut.
Remove the insulation cup and any necessary insulation (Figure 2, number 14) to provide access to the reducing nut (Figure 2, number 12). If the lower insulation cup was not in place, find it, because this piece improves the inlet temperature profile. With a 1/2-inch wrench, remove the reducing nut (Figure 2, number 12). Due to heat cycling of the GC, the reducing nut will be very tight. Remove the seal and the washer (Figure 2, number 10, 11) and discard. Place a new washer in the reducing nut and a new seal (flat side with groove up). Handtighten the reducing nut back into place within the retaining nut and then wrench-tighten until very tight. Replace the insulation cup. Insert a new liner and O-ring. The single taper liners are installed with the taper down, toward the column end of the inlet. Hand-tighten the insert assembly (Figure 2, number 3). Add the Merlin Microseal or proper preconditioned septum and septum nut. The molded septum is installed with the hole up. Follow the directions supplied with the Merlin Microseal to insure proper installation. If the green septum nut is used, wrench-tighten the weldment and septum nut with the septum nut wrench until the C-ring lifts off the top of the green septum nut. At this point, the inlet should be leak checked. Follow the directions in the GC maintenance and troubleshooting section of the manual. Column Installation Working with fused silica columns may be dangerous. Wear proper eye protection. Inspect the column for damage or breakage. Unweave 0.51 coil of the column from its basket to make it easier to install. Push a septum onto the inlet end of the column about 10 cm. Put the column nut and appropriate ferrule on the column. Cut 510 cm off the inlet end of the column. Check the cut with a 10 magnifier; the cut should be straight, not jagged, with no column shards within the column. If the cut is jagged or shards are inside, try again. After a clean cut is obtained, mark the proper column position with the septum (Figure 3). The septum will hold the column nut and ferrule in place. Place the column on the column hanger. Insert the column nut into the inlet reducing nut and finger-tighten. Wrench-tighten the column nut. The column should be stationary in the ferrule. Carefully slide the septum down away from the nut without disturbing the column positioning. The septum can be left in place if desired. Using the GC keypad or the MS ChemStation software, input the
Table 1.
Head Pressures and Calculated Flowrates for a Splitless Inlet at an Oven Temperature of 25 C with the Outlet Pressure set to Vacuum Length (meters) 12 25 50 30 30 50 Head pressure (psi) 6.0 15.0 28.0 6.2 3.4 5.5 Linear velocity (cm/s) 57 39 28 36 50 34 Column flow (mL/min) 1.0 1.0 1.0 1.0 2.0 1.5
Column id (mm) 0.20 0.20 0.20 0.25 0.32 0.32
Column hanger position
Marking the column position Capillary column Capillary inlet
2 cm
4-6 mm Inlet reducing nut Ferrule Column nut
the column nut in case it loosened. Check once again for column flow. Remove the end of the column from the beaker and close the oven door. Condition the column by slowly (5 C/min) ramping it to its maximum operating temperature. Leave it at that temperature for least 2 hours; overnight is preferable. The maximum operating temperature for an HP-5MS column is 325 C. Cool the oven to ambient and insert the interface nut and ferrule onto the column. Properly cut off 510 cm of the column. Properly place the column into the interface by following the directions in the MS hardware manual for the Agilent 5973 and for the Agilent 6890 series MSD. Hand-tighten the interface nut and then wrench-tighten the nut. The nut should be tightened only until you hear two squeaks. This is a firm seal. Pump down the detector as directed by the MSD manual (for the Agilent 5973 and for the Agilent 6890 Series MSD). Keep the oven at ambient temperature until the source is hot. Check the interface connection after the interface is heated. The interface nut may need additional tightening.
Septum
Tips for Better Method Performance

Numerous splitless parameters need to be optimized for the best splitless injection. Those parameters include: Injection port temperature Column flow
Septum
Capillary column
Figure 3.
Proper installation of capillary columns in a capillary inlet.
Liner design Solvent Sample volume Sample volatility Splitless valve time Injection speed
Column Dimensions, set Outlet Pressure to Vacuum and Column Flow between 1 and 1.5 mL/min helium (Table 1). The Split Vent Flow should be approximately 50 mL/min. These parameters are accessed through the inlet and column screens. Place the detector end of the column into a beaker of water and check for bubbles to show helium flow. Heat the injection port. When the injection port reaches the setpoint temperature, retighten
The proper inlet temperature is needed to evaporate high boiling point compounds without thermally degrading other compounds. Normally, the inlet temperature is a compromise between these two factors. A good starting point is 250 C. Liner design is one of the most difficult choices simply because of the variety of liners available. The features that are most important in a liner are the volume, whether it is deactivated or not, and whether or not it contains deactivated glass wool. As a general choice for high sensitivity work, a 4-mm single tapered, deactivated liner with no glass wool is recommended. For large volume injections (2 L) and for the highest repeatability (especially with small volume injections of 0.5 L [1]), deactivated glass wool is necessary. For dirty samples, deactivated glass wool helps to keep the nonvolatiles from getting to the column, but too much deactivated glass wool can greatly decrease sensitivity. Often, the most appropriate liner must be determined through experimentation. Please note: Removing and/or breaking deactivated glass wool creates active sites. Splitless valve timing is critical. The ON time (splitless mode) should to be long enough to assure that all of the injected sample reaches the column. A textbook splitless injection has the liner volume swept at least two times (during the ON time). A 4-mm liner has an approximate volume of 1 mL. With a GC/MS flow rate of 1 mL/min, a 2-min splitless injection would be necessary. However, this long splitless time has not been common. There are two reasons for this: Conventional (manually controlled) capillary inlets were pressure regulated (constant pressure, regardless of oven temperature) and not flow regulated (changing pressure with oven temperature), so a higher-than-optimal flow was set initially so that the flow did not go to zero at high oven temperatures. Thus, a typical splitless or OFF time has been between 0.5 to 1.5 min. Liner volumes smaller than the textbook examples have typically been used. Since a 2-mm liner (250 L volume) was more commonly employed, the splitless time was proportionally shorter. Finally, with the programmable control afforded by EPC, flows can be reliably pulsed during the
injection process. With pulsed splitless injections, flows during the splitless time can be 2-6 mL/min, resulting in splitless times less than 2 min for a 4-mm id liner. The pulsed splitless injection mode on the Agilent 6890 is recommended for GC/MS work. After the injection pulse, the system returns to analytical flow rates of 14 mL/min He. The highest flow allowable depends on the MSD. Refer to the appropriate MS hardware manual for your detectors limit. Unless all analytes have high boiling points, the initial oven temperature should be set to take advantage of the solvent effect. The solvent effect focuses the analytes on the head of the column. The oven temperature should typically be set to 10 C below the boiling point of the solvent used (Table 2). On the other hand, the MS interface temperature should be hot enough to avoid loss of analytes on cold spots. The interface should be set to the maximum oven temperature for the analysis or 10 C15 C higher if the upper temperature limit for the column is not exceeded. The default method temperature is 280 C; the interface temperature should be optimized as part of method development. Finally, the rate of auto-injection of a sample has been studied for splitless injections. It has been found that fast injections, such as with the ALS, tend to give the most repeatable and non-discriminating results.
Table 2. Solvent Diethyl ether n-Pentane Methylene chloride Carbon disulfide Acetone Chloroform Methanol n-Hexane Ethyl acetate Acetonitrile n-Heptane Isooctane Toluene Boling and Initial Oven Temperatures for Common Solvents Boiling point (C) 36 36 40 46 56 61 65 69 77 82 98 99 111 Initial oven temperature (C) 10 to 25 10 to 25 10 to 35 10 to 35 25 to 45 25 to 50 35 to 55 40 to 60 45 to 65 50 to 70 70 to 90 70 to 90 80 to 100
Using Pulsed Splitless Injections

Pulsed splitless injections are the best way to do splitless injections. EPC of the splitless inlet allows for high flow rates initially, followed by more typical GC/MS flow rates. A pulsed splitless injection transfers more of the sample onto the column and allows for increased injection volumes up to 5 L. When the injected volume is flash vaporized, the required expansion volume for the solvent is greatly increased (solvent choice also effects expansion volume). The use of higher initial inlet pressures reduces the volume (P1V1 = P2V2) so the entire injected volume can move to the column. The higher pressure also decreases the likelihood that highly volatile compounds will escape out the top of the injection port through the septum purge vent (Figures 4 and 5). In the case of thermally labile compounds, the faster they leave the hot injection port the less likely they are to degrade [2, 3]. The flow rate is then reduced to the value desired for the chromatographic separation. This flow is held constant by increasing the pressure as the oven temperature increases. Figure 6 is a graphical representation of the pulsed splitless technique with constant flow. Pulsed splitless injections should always be used when sensitivity and/or repeatability are critical. Refer to the GC operating manual for how to set up a pulsed splitless injection. Electronic pressure and flow control of carrier gases not only help with larger volumes, they also help to decrease run times and maintain stable MS sensitivity by keeping the carrier gas flow constant. These lead to shorter analyses, higher sensitivity, and higher reproducibility. The benefits for the analyst are more chromatographic analyses per shift, better data, and higher revenues per instrument.
Septum purge vent
Septum
Flow Carrier gas in
Inlet purge (off) Low inlet pressure/Low column flow

= More volatile compound = Less volatile compound
Liner
Column
Figure 4.
A low initial inlet pressure causes loss of volatile compounds.
Septum
Septum purge vent
Flow Carrier gas in
Inlet purge (off) High inlet pressure/High column flow

= More volatile compound = Less volatile compound
Liner
Column
Figure 5.
With the correct inlet pressure there is no loss of volatile compounds.
30 psi 1.5 min 320 C
400 C
300 C
98 psi/min
16 psi 200 C 180 C 130 C 5.4 psi 90 C 2.0 min 100 C
30 m 0.25-mm id column; vacuum compensation on; constant flow rate, 0.8 mL/min
10 15 20 25 30
Injection
Inlet pressure ( Oven temperature ( ) )
Run time
Figure 6.
Pulsed splitless injection technique employing constant flow with EPC. This technique allows larger injection volumes and inhibits the loss of volatile compounds out of the septum purge vent.
Summary
Following the instructions in this guide will improve analytical results with your GC/MSD system. Contamination interferring with the determination of your analytes will be minimized and sample transfer optimized, both improving sensitivity. A final note: the choice of vials and septa will affect your results. Screw cap vials are not recommended for GC/MS analyses. Application note Effects of Vial Septa Used in GC/ECD Analysis of Trace Organics Agilent Technologies publication 5091-8980E www.agilent.com/chem will help you make the right choice.
References
1. Analyzing aromatics in reformulated gasoline by GC/MS, Agilent Technologies, publication (23) 5964-0116E. 2. Philip L. Wylie et al. Improving the Analysis of Pesticides by Optimizing Splitless Injections. 1995 Pittsburgh Conference paper number 347. 3. Philip L. Wylie and Katsura Uchiyama. Improved Gas Chromatogaphic Analysis of Organophoshorous Pesticides with Pulsed Splitless Injection. (1996) JAOAC 79: 571-577.
Appendix A
Capillary Inlet System Liners
Configuration ID volume 4 mm (0.8-mm end) 900 L 4 mm (0.8-mm end) 900 L 4 mm (0.8-mm end) 800 L 4 mm 990 L 4 mm with cup 4 mm with cup 2 mm 250 L 2 mm 250 L 1.5 mm 140 L Glass wool * packing** YES**
Applications Fast Slow & injection manual (7673 ALS) injection Splitless Splitless
Linear seals Fluorocarbon (Max 350 C) 5180-4182 (12/pk) 12 5180-4182 (12/pk) 12 5180-4182 (12/pk) 12
Graphite (350 C and above) 5180-4173 (12/pk) 48 5180-4173 (12/pk) 48 5180-4173 (12/pk) 48 5180-4168 (12/pk) 48 5180-4168 (12/pk) 48 5180-4168 (12/pk) 48 5180-4173 (12/pk) 48 5180-4173 (12/pk) 48 5180-4173 (12/pk) 48
Liner Single-Taper liner
Part no. 5062-3587
Price 34
Glass type Borosilicate
Deactivated YES
Split C
Split C
Single-Taper liner
5181-3316
31
Borosilicate
YES
NO
Double-Taper liner
5181-3315
34
Borosilicate
YES
NO
Capillary liner
19251-60540
26
Borosilicate
NO
YES** YES + column packing NO
5180-4182 (12/pk) 12 5180-4182 (12/pk) 12 5180-4182 (12/pk) 12 5180-4182 (12/pk) 12 5180-4182 (12/pk) 12 5180-4182 (12/pk) 12
Split liner
18740-60840
46
Borosilicate
NO
Split liner
18740-80190
41
Borosilicate Quartz (Purity 8) Quartz (Purity 8)
NO
Splitless liner
18740-80220
27
NO
NO
A,B
Splitless liner
5181-8818
30
YES
NO
A,B
Direct liner
18740-80200
17
Borosilicate
NO
NO
A,B
General recommendation See notes at left regarding use
* Quality discounts available; please inquire. ** Silanized glass wool, 10 gm, (pesticide grade) (Agilent p/n 5181-3317).
A. Can be used without the glass wool in some applications, but liners with glass wool are generally recommended for best reproducibility in fast injections. B. Recommended only for small (< 0.5 L) volumes, depending on solvent and conditions. C. The glass wool packing as supplied may not provide adequate mixing for good precision in split injections. Liners can be packed with silanized glass wool positioned as in the straight 4-mm capillary liner, part number 19251-60540. D. Taper liners are recommended for best performance in this application, particularly with labile samples and wide-boilingrange mixtures. E. Not recommended for use with EPC.
9

For more information on our products and services, visit our Web site at: www.agilent.com/chem
Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. MicrosealTM is a trademark of the Merlin Instrument Company. SwagelokTM is a registered trademark of the Swagelok Complany. SnoopTM is a registered trademark of the Nupro Company Vespel is a registered trademark of E. I. du Pont de Nemours Co. (Inc.) X-Acto is a registered trademark of the Hunt Corporation. Agilent Technologies, Inc. 2003 Printed in the USA November 14, 2003 5988-9944EN
HPLC for Food Analysis

A Primer
Copyright Agilent Technologies Company, 1996-2001. All rights reserved. Reproduction, adaption, or translation without prior written permission is prohibited, except as allowed under the copyright laws. Printed in Germany September 01, 2001 Publication Number 5988-3294EN
HPLC for Food Analysis

A Primer
The fundamentals of an alternative approach to solving tomorrows measurement challenges
Angelika Gratzfeld-Hsgen and Rainer Schuster
Acknowledgements We would like to thank Christine Miller and John Jaskowiak for their contributions to this primer. Mrs. Miller is an application chemist with Agilent Technologies and is responsible for the material contained in chapter 5. Mr. Jaskowiak, who wrote chapter 7, is a product manager for liquid chromatography products at Agilent Technologies.
Copyright Agilent Technologies Company 1996-2001. All rights reserved. Reproduction, adaption, or translation without prior written permission is prohibited, except as allowed under the copyright laws. Printed in Germany, September 1, 2001. Publication Number 5988-3294EN
Preface
Modern agriculture and food processing often involve the use of chemicals. Some of these chemicals and their functions are listed below: Fertilizers: increase production of agricultural plants Pesticides: protect crops against weeds and pests Antibiotics: prevent bacteria growth in animals during breeding Hormones: accelerate animal growth Colorants: increase acceptability and appeal of food Preservatives and antioxidants: extend product life Natural and artificial sweeteners and flavors: improve the taste of food Natural and synthetic vitamins: increase the nutritive value of food Carbohydrates: act as food binders Such chemicals improve productivity and thus increase competitiveness and profit margins. However, if the amounts consumed exceed certain limits, some of these chemicals may prove harmful to humans. Most countries therefore have established official tolerance levels for chemical additives, residues and contaminants in food products. These regulations must be monitored carefully to ensure that the additives do not exceed the prescribed levels. To ensure compliance with these regulatory requirements, analytical methods have been developed to determine the nature and concentration of chemicals in food products. Monitoring of foodstuffs includes a check of both the raw materials and the end product. To protect consumers, public control agencies also analyze selected food samples. High-performance liquid chromatography (HPLC) is used increasingly in the analysis of food samples to separate and detect additives and contaminants. This method breaks down complex mixtures into individual compounds, which in turn are identified and quantified by suitable detectors
III
and data handling systems. Because separation and detection occur at or slightly above ambient temperature, this method is ideally suited for compounds of limited thermal stability. The ability to inject large sample amounts (up to 12 ml per injection) makes HPLC a very sensitive analysis technique. HPLC and the nondestructive detection techniques also enable the collection of fractions for further analysis. In addition, modern sample preparation techniques such as solid-phase extraction and supercritical fluid extraction (SFE) permit high-sensitivity HPLC analysis in the ppt (parts per trillion) range. The different detection techniques enable not only highly sensitive but also highly selective analysis of compounds.
Hydrophilic
HPLC
Amino acids Volatile carboxylic acids Aldehydes Ketones Sulfonamides Nitriles Glycols Synthetic Glyphosate food dyes Enzymes PG, OG, DG phenols Aflatoxins Antibiotics
Inorganic ions Sugar alcohols Sugars
Polarity
Nitrosamine TMS derivative of sugars Epoxides
BHT, BHA, THBQ antioxidants Alcohol
Fatty acids
Flavonoids Anabolica Natural food dyes
Organophosphorous pesticides PCB Essential oils
PAHs
Aromatic amines
Fat soluble vitamins Triglycerides Phospho-lipids
Polymer monomers Fatty acid methylester
C2/C6 hydrocarbons Hydrophobic
Aromatic esters
GC Volatile
Volatility
HPLC
Nonvolatile
Figure 1 Match of analyte characteristics to carrier medium
IV
Its selective detectors, together with its ability to connect a mass spectrometer (MS) for peak identification, make gas chromatography (GC) the most popular chromatographic method. HPLC separates and detects at ambient temperatures. For this reason, agencies such as the U.S. Food and Drug Administration (FDA) have adopted and recommended HPLC for the analysis of thermally labile, nonvolatile, highly polar compounds. Capillary electrophoresis (CE) is a relatively new but rapidly growing separation technique. It is not yet used in the routine analysis of food, however. Originally CE was applied primarily in the analysis of biological macromolecules, but it also has been used to separate amino acids, chiral drugs, vitamins, pesticides, inorganic ions, organic acids, dyes, and surfactants.1, 2, 3 Part 1 is a catalog of analyses of compounds in foods. Each section features individual chromatograms and suggests appropriate HPLC equipment. In addition, we list chromatographic parameters as well as the performance characteristics that you can expect using the methods shown. In part 2 we examine sample preparation and explain the principles behind the operation of each part of an HPLC systemsampling systems, pumps, and detectorsas well as instrument control and data evaluation stations. In the last of 11 chapters, we discuss the performance criteria for HPLC, which are critical for obtaining reliable and accurate results. Part 3 contains a bibliography and an index.
Contents
Part One The HPLC Approach
Chapter 1 Analytical examples of food additives Acidulants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Antioxidants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Preservatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Artificial sweeteners . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Colorants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Flavors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Vanillin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Bitter compounds: hesperidin and naringenin . . . . . . . 14 Chapter 2 Analytical examples of residues and contaminants Residues of chemotherapeutics and antiparasitic drugs . . Tetracyclines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fumonisins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mycotoxins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bisphenol A diglydidyl-ether (BADGE) . . . . . . . . . . . . . . . . Pesticides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Carbamates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Glyphosate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chapter 3 Analytical examples of natural components Inorganic anions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Lipids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Triglycerides and hydroperoxides in oils . . . . . . . . . . . Triglycerides in olive oil . . . . . . . . . . . . . . . . . . . . . . . . . Fatty acids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Carbohydrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Vitamins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Water-soluble vitamins . . . . . . . . . . . . . . . . . . . . . . . . . . Fat-soluble vitamins . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis of tocopherols on normal-phase column . . . . Biogenic amines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Amino acids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Peptides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
VI
16 18 19 21 24 26 28 29
32 35 35 37 38 40 42 42 45 46 48 50 52
Part Two The Equipment Basics
Chapter 4 Separation in the liquid phase Separation mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . Reversed-phase materials . . . . . . . . . . . . . . . . . . . . . . . . Ion-exchange materials . . . . . . . . . . . . . . . . . . . . . . . . . . Size-exclusion gels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Adsorption media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The advent of narrow-bore columns . . . . . . . . . . . . . . . . . . Influence of column temperature on separation . . . . . Chapter 5 Sample preparation Sample preparation steps . . . . . . . . . . . . . . . . . . . . . . . . . . . Automation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Solids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ultrasonic bath liquid extraction . . . . . . . . . . . . . . . . . . Steam distillation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Supercritical fluid extraction . . . . . . . . . . . . . . . . . . . . . Liquids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Liquid-liquid extraction . . . . . . . . . . . . . . . . . . . . . . . . . Solid-phase extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . Gel permeation chromatography . . . . . . . . . . . . . . . . . Guard columns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chapter 6 Injection techniques Characteristics of a good sample introduction device . . . Manual injectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Automated injectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Autosampler with sample pretreatment capabilities . . . . Derivatization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chapter 7 Mobile phase pumps and degassers Characteristics of a modern HPLC pump . . . . . . . . . . . . . . Flow ranges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gradient elution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gradient formation at high pressure . . . . . . . . . . . . . . . Gradient formation at low pressure . . . . . . . . . . . . . . .
58 58 58 59 59 59 60
62 62 63 63 64 64 65 65 65 66 67
70 71 72 72 73
76 76 76 77 77
VII
Pump designs for gradient operation . . . . . . . . . . . . . . . . . Low-pressure gradient Agilent 1100 Series pump . . . . High-pressure gradient Agilent 1100 Series pump . . . . Degassing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Helium degassing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Vacuum degassing . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
78 78 80 82 83 84
Chapter 8 Detectors Analytical parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 Limit of detection and limit of quantification . . . . . . . 87 Selectivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 Linearity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 Qualitative information . . . . . . . . . . . . . . . . . . . . . . . . . . 88 UV detectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 Diode array detectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 Three dimensions of data . . . . . . . . . . . . . . . . . . . . . . . . 91 Fluorescence detectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 Cut-off filter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 Signal/spectral mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 Online spectral measurements and multi signal acquisition . . . . . . . . . . . . . . . . . . . . . . . . . . 96 Multisignal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 Electrochemical detectors . . . . . . . . . . . . . . . . . . . . . . . . . . 98 Electrode materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 Flow cell aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 Automation features . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 Mass spectrometers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 API interfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 Refractive index detectors . . . . . . . . . . . . . . . . . . . . . . . . . 104
VIII
Chapter 9 Derivatization chemistries Addition of UV-visible chromophores . . . . . . . . . . . . . . . . 108 Addition of a fluorescent tag . . . . . . . . . . . . . . . . . . . . . . . 109 Precolumn or postcolumn? . . . . . . . . . . . . . . . . . . . . . . . . . 109 Automatic derivatization . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 Chapter 10 Data collection and evaluation techniques Strip chart recorders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 Integrators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 Personal computers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 Local area networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 Networked data systems . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 Chapter 11 Factors that determine performance in HPLC Limit of detection and limit of quantification . . . . . . . . . 121 Accuracy and precision . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122 Qualitative information . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
Part Three References and Index
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
IX
The HPLC Approach
Part One
A demonstration of liquid chromatographic separations in food analysis
Chapter 1
Analytical examples of food additives
1
Acidulants
Sorbic acid and citric acids are commonly used as acidulants4 and/or as preservatives. Acetic, propionic, succinic, adipic, lactic, fumaric, malic, tartaric, and phosphoric acids can serve as acidulants as well. Acidulants are used for various purposes in modern food processing. For example, citric acid adds a fresh, acidic flavor, whereas succinic acid gives food a more salty, bitter taste. In addition to rendering foods more palatable and stimulating, acidulants act as flavoring agents to intensify certain tastes and mask undesirable aftertastes buffering agents to control the pH during food processing and of the finished products preservatives to prevent growth of microorganisms synergists to antioxidants to prevent rancidity and browning viscosity modifiers in baked goods melting modifiers in cheese spreads and hard candy meat curing agents to enhance color and flavor Sample preparation Sample preparation depends strongly on the matrix to be analyzed, but in general steam distillation and solid-phase extraction techniques can be used. Chromatographic conditions High-performance liquid chromatography (HPLC) with UV-visible diode-array detection (UV-DAD) has been applied in the analysis of citric acid in wine and in a vodka mixed drink. Retention time and spectral data were used as identification tools.
Water Isocratic Autopump + sampler vacuum degasser
Column compartment
Detector (VWD, DAD or refractive index)
Control and data evaluation
Sample preparation Column Mobile phase isocratic Flow rate Injection volume Detector
filtration 300 x 7.8 mm BioRad HPX 87-H, 9 m 0.0035 M H2SO4
mAU 400 300

1
0.6 ml/min 10 l UV-VWD detection wavelength 192 nm or 210 nm
Column compartment 65 C
200
?
? ?
1 2 3 4 5 6
? 6 7 8
Oxalic acid Citric acid Tartaric acid Malic acid Sulfur-trioxide Succinic acid
7 8 9 10 11 12
Lactic acid Glycerol DEG Acetic acid Methanol Ethanol White wine
100
2
3 4
? 9 10 11 12
0
0 5 10 15 Time [min] 20 25
Standard
Figure 2 Analysis of acidulants in white wine
Conditions as above except Mobile phase 0.007 M H2SO4 isocratic Detector UV-DAD
mAU 100
Citric acid
Citric acid
Glucose Fructose
20
Sample spectrum overlaid with library spectrum match 994
HPLC method performance Limit of detection 100 ng injected amount, S/N = 2 equivalent to 2 ppm with 50 l injected volume Repeatability of RT over 10 runs < 0.1 % areas over 10 runs < 3 %
190
Wavelength [nm] 276
Ethanol 0 0 5 10 Time [min] 15

20
Figure 3 Analysis of citric acid in vodka
4. Official Methods of Analysis, Food Compositions; Additives, Natural Contaminants, 15th ed; AOAC: Arlington, VA, 1990, Vol. 2.; Official Method AOAC 986.13: quinic, malic, citric acid in cranberry juice cocktail and apple juice.
1
Antioxidants
The following compounds are used as antioxidants in food products:4 Natural antioxidants: vitamin C vitamin E Synthetic antioxidants: BHT BHA TBHQ THBP PG OG DG Ionox-100 NDGA TDPA ACP butylated hydroxytoluene butylated hydroxyanisole mono-tert-butylhydroquinone 2,4,5-trihydroxybutyrophenone propyl gallate octyl gallate dodecyl gallate 4-hydroxymethyl-2,6-di(tert-butyl)phenol nordihydroguaiaretic acid 3,3'-thiodipropionic acid ascorbyl-palmitate
Antioxidants may be naturally present in food, or they may be formed by processes such as smoking. Examples of natural antioxidants include tocopherols (vitamin E) and acsorbic acid (vitamin C). A second category of antioxidants comprises the wholly synthetic antioxidants. When these antioxidants are added to foodstuffs, they retard the onset of rancidity by preventing the oxidative degradation of lipids. In most countries where antioxidants are permitted either singly or as combinations in foodstuffs, maximum levels for these compounds have been set. Sample preparation Sample preparation depends strongly on the matrix to be analyzed. For samples low in fat, liquid extraction with ultrasonic bath stimulation can be used. For samples with more complex matrices, solid-phase extraction, liquid/liquid extraction, or steam distillation may be necessary.
Chromatographic conditions HPLC and UV-visible diode-array detection have been applied in the analysis of antioxidants in chewing gum. Spectral information and retention times were used for identification.
ultrasonic liquid extraction with acetonitrile (ACN) Column 1 100 x 4 mm BDS, 3 m Mobile phase A = water + 0.2 ml H2SO4, pH = 2.54 B = ACN Gradient start with 10 % B at 3 min 60 % B at 4 min 80 % B at 11 min 90 % B Flow rate 0.5 ml/min Post time 4 min Column compartment 30 C Injection volume 5 l Detector UV-DAD detection wavelength 260/40 nm, reference wavelength 600/100 nm Sample preparation
mAU 1500
1000 2 1 0 2 4 3 4
500
1 2 3 4 5 6 7 8 56
Vitamin C PG THBP TBHQ BHA 4-hydroxy BHT Chewing gum extract ACP 7 10 12 8 Standard
6 8 Time [min]
HPLC method performance Limit of detection Repeatability of RT over 10 runs areas over 10 runs 0.12 ng (injected amount), S/N = 2 < 0.2 % <1%
Quaternary pump + vacuum degasser
Control and data evaluation Autosampler Column compartment Diodearray detector
Water
Acetonitrile
4. Official Methods of Analysis, Food Compositions; Additives, Natural Contaminants, 15th ed; AOAC: Arlington, VA, 1990, Vol. 2.; AOAC Official Method 983.15: Antioxidants in oils and fats.
1
Preservatives
The following compounds are used as preservatives in food products: benzoic acid sorbic acid propionic acid methyl-, ethyl-, and propylesters of p-hydroxy benzoic acid (PHB-methyl, PHB-ethyl, and PHB-propyl, respectively)4 Preservatives inhibit microbial growth in foods and beverages. Various compound classes of preservatives are used, depending on the food product and the expected microorganism. PHBs are the most common preservatives in food products. In fruit juices, in addition to sulfur dioxide, sorbic and benzoic acid are used as preservatives, either individually or as a mixture. Sample preparation Sample preparation depends strongly on the matrix to be analyzed. For samples low in fat, liquid extraction with ultrasonic bath stimulation can be used. For samples with more complex matrices, solid-phase extraction, liquid/liquid extraction, or steam distillation may be necessary.
Control and data evaluation Quaternary pump + vacuum degasser Autosampler Column compartment Diodearray detector
Water
Acetonitrile
Chromatographic conditions HPLC and UV-visible diode-array detection have been applied in the analysis of preservatives in white wine and salad dressing. Spectral information and retention times were used for identification.
Carrez clearing and filtration for the salad dressing. None for white wine. Column 125 x 4 mm Hypersil BDS, 5 m Mobile phase A = water + 0.2 ml H2SO4, pH = 2.3 B = ACN Gradient start with 10 % B at 3 min 60 % B at 4 min 80 % B at 6 min 90 % B at 7 min 10 % B Flow rate 2 ml/min Post time 1 min Column compartment 40 C Injection volume 2 l Detector UV-DAD detection wavelength 260/40 nm Sample preparation
Absorbance (scaled) Spectral library 50 library match 999 30 10 sample 200 Wavelength [nm] 320
mAU 60 50
Sorbic acid PHB-methyl
Benzoic acid
40 30 20 10 0
PHB-propyl
PHB-ethyl
BHA
BHT
3 4
Standard White wine Salad dressing
Time [min]
HPLC method performance Limit of detection Repeatability of RT over 10 runs areas over 10 runs 10 ppm, S/N = 2 < 0.1 % <3%
Figure 5 Analysis of preservatives in white wine and salad dressing
4. Official Methods of Analysis, Food Compositions; Additives, Natural Contaminants, 15th ed; AOAC: Arlington, VA, 1990, Vol. 2.; AOAC Official Method 979.08: Benzoate, caffeine, saccharine in carbonated beverages.
1
Artificial sweeteners
The following compounds are used as artificial sweeteners in food products: acesulfam aspartame saccharin4 Nowadays, low-calorie sweeteners are widely used in foods and soft drinks. Investigations of the toxicity of these compounds have raised questions as to whether they are safe to consume. As a result, their concentration in foods and beverages is regulated through legislation in order to prevent excessive intake. Sample preparation Sample preparation depends strongly on the matrix to be analyzed. For sample low in fat, liquid extraction at low pH with ultrasonic bath stimulation can be used. For samples with more complex matrices, solid-phase extraction, liquid/liquid extraction, or steam distillation may be necessary.
Fluorescence detector Quaternary pump + vacuum degasser
Autosampler
Column compartment
Diodearray detector
Water Methanol
Chromatographic conditions The HPLC method presented here for the analysis of aspartame is based on automated on-column derivatization and reversed-phase chromatography. UV spectra were evaluated as an additional identification tool.5
Derivatization agent o-phthalaldehyde (OPA) mercapto-propionic acid (MPA) Column 100 x 2.1 mm Hypersil ODS, 5 m Mobile phase A = 0.01 mM sodium acetate B = methanol Gradient start with 5 % B at 5 min 25 % B at 10 min 35 % B at 13 min 55 % B at 18 min 80 % B at 20 min 95 % B Flow rate 0.35 ml/min Post time 5 min Column compartment 40 C Injection volume 1 l Injector program for online derivatization 1. Draw 5.0 l from vial 3 (borate buffer) 2. Draw 0.0 l from vial 0 (water) 3. Draw 1.0 l from vial 1 (OPA/MPA) 4. Draw 0.0 l from vial 0 (water) 5. Draw 1.0 l from sample 6. Mix 7 l (6 cycles) 7. Inject Detectors UV-DAD: detection wavelength 338/20 nm or fluorescence: excitation wavelength 230 nm, emission wavelength 445 nm
mAU
scaled
Aspartame spectra
60 50 40 30 20 10 0 0 2 4 6 Time [min] Aspartame
original
derivatized
250
400
10
Figure 6 Chromatogram and spectra of derivatized and non derivatized aspartame

HPLC method performance Limit of detection for fluorescence for DAD 200 pg (injected amount), S/N = 2 1 ng (injected amount), S/N = 2
Repeatability of RT over 10 runs < 0.1 % of areas over 10 runs < 5 %
5. A.M. Di Pietra et al., HPLC analysis of aspartame and saccharin in pharmaceutical and dietary formulations; Chromatographia, 1990, 30, 215219. 4. Official Methods of Analysis, Food Compositions; Additives, Natural Contaminants, 15th ed; AOAC: Arlington, VA, 1990, Vol. 2.; Official Method AOAC 979.08: Benzoate, caffeine, saccharin in soda beverages.
1
Colorants
We have selected the food color E104 Quinolin yellow and E131 Patent blue as application examples. Synthetic colors are widely used in the food processing, pharmaceutical, and chemical industries for the following purposes:4 to mask decay to redye food to mask the effects of aging The regulation of colors and the need for quality control requirements for traces of starting product and by-products have forced the development of analytical methods. Nowadays, HPLC methods used are based on either ion-pairing reversed-phase or ion-exchange chromatography. UV absorption is the preferred detection method. The UV absorption maxima of colors are highly characteristic. Maxima start at approximately 400 nm for yellow colors, 500 nm for red colors, and 600700 nm for green, blue, and black colors. For the analysis of all colors at maximum sensitivity and selectivity, the light output from the detector lamp should be high for the entire wavelength range. However, this analysis is not possible with conventional UV-visible detectors based on a one-lamp design. Therefore, we have chosen a dual-lamp design based on one deuterium and one tungsten lamp. This design ensures high light output for the entire wavelength range. Sample preparation Whereas turbid samples require filtration, solid samples must be treated with 0.1 % ammonia in a 50 % ethanol and water mixture, followed by centrifugation. Extraction is then performed using the so-called wool-fiber method. After desorption of the colors and filtration, the solution can be injected directly into the HPLC instrument.
Water Acetonitrile Quaternary AutoColumn pump + sampler compartvacuum ment degasser
Diodearray detector
10
Chromatographic conditions The HPLC method presented here for the analysis of dyes is based on ion-pairing reversed-phase chromatography. UV spectra were evaluated as an additional identification tool.6
injection without further preparation Column 125 x 3 m Hypersil BDS, 3 mm Mobile phase A = 0.01 M NaH2PO4+ 0.001 M tetrabutylammoniumdihydrogenphosphate, pH = 4.2 B = ACN Gradient start with 15 % in 10 min to 40 % in 14 min to 90 % until 19 min at 90 % in 20 min to 15 % ACN Stop time 20 min Post time 4 min Flow rate 0.8 ml/min Column compartment 40 C Injection volume 1 l Detector UV-DAD signal A: 254/50 nm (for optimization of separation) signal B: 350/20 nm signal C: 465/30 nm signal D: 600/40 nm Sample preparation
mAU 12 10 8 6 4 2 0 2
Woodruff lemonade
Chinolin yellow
Patent blue
465 nm/30 nm
600 nm/40 nm 4 6 8 Time [min] 10 12 14
Norm Tartrazine yellow 40 30 20 10 0 300
Spectra of different colors Amaranth Patent blue red
Brilliant blue
400
700
800
Figure 7 Analysis of synthetic colors in lemonade. Overlay of spectra of yellow, red, blue and black colors
HPLC method performance Limit of detection for UV-DAD Repeatability of RT over 10 runs of areas over 10 runs 2 ng (injected amount) S/N = 2 < 0.2 % <3%
4. Official Methods of Analysis, Food Compositions; Additives, Natural Contaminants, 15th ed; AOAC: Arlington, VA, 1990, Vol. 2.; Official Method AOAC 981.13: Cresidine sulfonic acid in FD&C Red No. 40; Official Method AOAC 982.28: Intermediates and reaction by-products in FD&Y Yellow No. 5; Official Method AOAC 977.23: 44 (Diazoamino) dibenzene sulfonic acid (DAADBSA) in FD&C Yellow No. 6; Official Method AOAC 980.24: Sulfanilic acid in FD&C Yellow No. 6. 6. A.G. Huesgen, R.Schuster, Sensitive analysis of synthetic colors using HPLC and diode-array detection at 190950 nm, Agilent Application Note 5964-3559E, 1995.
11
1
Flavors
The following compounds are examples of flavoring agents used in food products: lupulon and humulon (hop bittering compounds) vanillin naringenin and hesperidin (bittering compounds) Three major classes of compounds are used as flavoring agents: essential oils, bitter compounds, and pungency compounds. Although the resolution afforded by gas chromatography (GC) for the separation of flavor compounds remains unsurpassed, HPLC is the method of choice if the compound to be analyzed is low volatile or thermally unstable.
Vanillin Sample preparation

Turbid samples require filtration, whereas solid samples must be extracted with ethanol. After filtration, the solution can be injected directly into the HPLC instrument.
Water
Acetonitrile
12
injection without further preparation Column 100 x 4 mm Hypersil BDS, 3 m Mobile phase A = water + 0.15 ml H2SO4 (conc.), pH = 2.3 B = ACN Gradient start with 10 % B at 3 min 40 % B at 4 min 40 % B at 6 min 80 % B at 7 min 90 % B Flow rate 0.8 ml/min Post time 3 min Column compartment 30 C Injection volume 5 l Detector UV-DAD detection wavelength 280/80 nm, reference wavelength 360/100 nm
Sample preparation
Chromatographic conditions The HPLC method presented here for the analysis of vanillin is based on reversed-phase chromatography. UV spectra were evaluated as an additional identification tool.7
Norm. 400 300 200 100 0 0 1 2 3 4 Time [min] 5 6 Vanillin extract 7 Ethylvanillin Vanillin 4-hydroxybenzaldehyde Coumarin
4-hydroxy benzoic acid
Vanillin alcohol
Standard
Conditions as above, except Column 100 x 2.1 mm Hypersil ODS, 5 m Mobile phase A = water + 5 mM NaH2PO4 B = methanol Gradient at 10 min 70 % B Flow rate 0.4 ml/min
mAU 60
Gallic acid
Syringaaldehyde
50 40
30 20
Vanillin
50 40 30 20 10 0
Vanillin
Match 991
Salicylaldehyde
HPLC method performance Limit of detection 0.25 ng (injected amount) S/N = 2
10
0 0 2 4 Time [min] 6
8
Standard Cognac
10
Repeatability of RT over 10 runs < 0.2 % of areas over 10 runs < 1 %
Figure 9 Analysis of vanillin in cognac. Identification of vanillin through spectra comparison 7. Herrmann, A, et al.;,Rapid control of vanilla-containing products using HPLC; J. Chromatogr., 1982, 246, 313316.
13
1
Bitter compounds: hesperidin and naringenin
Sample preparation for bitter compounds in orange juice8 The samples were prepared according to Carrez 1 and 2. This method uses potassium ferrocyanide and zinc sulfate for protein precipitation. Chromatographic conditions The HPLC method presented here for the analysis of hesperidin and naringenin is based on reversed-phase chromatography. UV spectra were evaluated as an additional identification tool.
The orange juice was prepared according to Carrez 1 and 2. Column 125 x 4 mm Hypersil BDS, 5 m Mobile phase A = water + 0.15 ml/l H2SO4 (conc.), pH = 2.4 B = ACN Gradient start with 20 % B at 3 min 20 % B at 5 min 90 % B at 6 min 20 % B Flow rate 2 ml/min Post time 1 min Column compartment 40 C Injection volume 1 l Detector UV-DAD detection wavelength 260/80 nm, reference wavelength 380/80 nm Sample preparation
mAU 20 15 10 5 0 -5 0.5 1
Time [min] Naringenin
Hesperidin
Standard
Orange juice
1.5
2.5
Figure 10 Analysis of bitter compounds in orange juice
HPLC method performance Limit of detection for DAD Repeatability of RT over 10 runs of areas over 10 runs 1 ng (injected amount), S/N = 2 < 0.2 % < 1 %.
8. Official Methods of Analysis; Horwitz, W., Ed.; 14th ed.; AOAC: Arlington, VA, 1984; secs 12.01812.021.
14
Chapter 2
Analytical examples of residues and contaminants
2
Residues of chemotherapeutics and antiparasitic drugs
In addition to several other drugs, nitrofurans and sulfonamides such as sulfapyridine, N-acetyl metabolite, ethopabat, chloramphenicol, meticlorpindol, metronidazol, ipronidazol, furazolidone, and nicarbazin are frequently fed to domestic cattle. Modern intensive animal breeding demands permanent suppression of diseases caused by viruses, bacteria, protozoa, and/or fungi. A number of chemotherapeutics are available for the prevention and control of these diseases. After application, residues of these drugs can be found in foods of animal origin such as milk, eggs, and meat. These chemotherapeutics can cause resistancy of bacteria. Because of the toxic nature of chemotherapeutics, for example, choramphenical, government agencies in many countries, including the United States, Germany, and Japan, have set tolerance levels for residues of these drugs. Simple and reliable analysis methods are necessary in order to detect and quantify residues of chemotherapeutic and antiparasitic drugs in food products. Malisch et al. have developed an HPLC method to determine 11 of these compounds.9,10 The internal standard (ISTD) comprises benzothiazuron and pyrazon. Sample preparation After homogenization or mincing and pH adjustment, the samples were extracted using liquid/liquid extraction followed by degreasing, purification, and concentration.
Water
Acetonitrile
16
Sample preparation
Column Mobile phase
Gradient
Flow rate Injection volume Detector
Sample preparation was done according to reference9 250 x 4.6 mm Spherisorb ODS-2, 5 m A = sodium acetate buffer, 0.02 M, pH = 4.8 B = ACN/water (60:40) start with 8 % B at 5 min 8 % B at 7 min 20 % B at 14 min 23 % B at 16 min 33 % B at 19 min 40 % B at 21 min 50 % B at 26 min 60 % B at 30 min 80 % B at 33 min 90 % B at 43 min 90 % B at 55 min 8 % B 1.5 ml/min 20 l UV-DAD detection wavelengths 275/80 nm, 315/80 nm, and 360/80 nm, reference wavelength 500/100 nm
Chromatographic conditions The HPLC method presented here for the analysis of residues of drugs in eggs, milk, and meat is based on reversed-phase chromatography and multisignal UV-visible diode-array detection (UV-DAD). UV spectra were evaluated as an additional identification tool.
Sulfapyridine t R= 12.2 min match 997 Pyrazon t R= 9 min match 998
Scaled
40 0 250 300 350 Wavelength [nm] offset 400
Scaled
80
80
40 0 300 350 250 Wavelength [nm] 9 offset 400
mAU 20 10 0 1 2 10
3 4
5 8 6,7 20
10 11 Egg sample Standard 30
Time [min]
HPLC method performance Limit of detection 0.0010.05 mg/kg Repeatability of RT over 10 runs < 0.12 % of areas over 10 runs < 1.5 %
1 2 3 4
metronidazol meticlorpindol sulfapyridine furazolidone
5 6 7 8
pyrazon ipronidazol chloramphenicol N-acetyl metabolite of 3
9 3-ethopabat 10 benzothiazuron 11 nicarbazin
Figure 11 Analysis of residues in an egg sample. Identification through spectra comparison
9. H. Malisch, et al.,Determination of residues of chemotherapeutic and antiparasitic drugs in food stuffs of anomaly origin with HPLC and UV-Vis diode-array detection, J. Liq. Chromatogr., 1988, 11 (13), 28012827.14. 10. EC Guideline 86/428 EWG 1985.
17
2
Tetracyclines
Tetracyclines are used worldwide as oral or parenteral medication in the form of additives in animal feed. In food-producing animals, these drugs exhibit a high degree of activity toward a wide range of bacteria.9, 11 Sample preparation After homogenization or mincing and addition of mineral acids to dissociate tetracyclines from proteins, the samples were extracted using liquid/liquid extraction followed by degreasing and/or deproteinization, purification, and concentration.12 Chromatographic conditions The HPLC method presented here for the analysis of meat is based on reversed-phase chromatography and UV-visible diode-array detection. UV spectra were evaluated as an additional identification tool.
Sample preparation
1 g sample was mixed with citric acid (100 mg). add 1 ml nitric acid (30 %) or 0.1 m oxalic acid add 4 ml methanol 5 min in the ultrasonic bath add water up to 10 ml total volume centrifuge inject Column 100 4 mm Hypersil BDS, 3 m Mobile phase A = water, pH = 2.1 with sulfuric acid B = ACN Gradient start with 15 % B at 10 min 60 % B Flow rate: 0.5 ml/min Column compartment 25 C Detector UV-DAD detection wavelength 355 nm/20 nm, reference wavelength 600/100 nm
Oxytetracycline 1.8 ng 6 5 4 3 2 1 0 2 4 Time [min] 6 370 ppb
3 2 1
Oxytetracycline
Library match 980
Pork muscle Blank 8
HPLC method performance Limit of detection for UV-DAD 100 ppb Repeatability of RT over 10 runs < 0.2 % of areas over 10 runs < 2 %
Figure 12 Trace analysis of tetracycline residues in meat. Identication of oxytetracycline through spectra comparison 9. H. Malisch et al., Determination of residues of chemotherapeutic and antiparasitic drugs in food stuffs of anomaly origin with HPLC and UV-Vis diode-array detection J. Liq. Chromatogr., 1988, 11 (13), 28012827.14. 11. M.H. Thomas, J. Assoc. Off. Anal.; 1989 , 72 (4) 564. 12. Farrington et. al., Food Additives and Contaminants, 1991, Vol. 8, No. 1, 55-64.
18
Fumonisins
Fumonisins are characterized by a 19-carbon aminopolyhydroxyalkyl chain which is diesterified with propane1,2,3-tricarboxylic acid. Analogues B 1-3 in figure 13 show a difference only in the number and position of the hydroxyl groups present on the molecule. Fragmentation experiments using collision induced dissociation (CID) show no difference between fumonisins B2 and B3. Consequently, it was necessary to separate these compounds chromatographically for quantitative analysis. However, in crude corn extracts the CID-fragment ions provide important confirmatory information. In order to obtain spectra of the fragment ions as well as the pseudomolecular ions in a single scan, operating at maximum sensitivity, the fragmentor voltage was set to 230 V while scanning from 150 amu to 680 amu and then to 100 V when scanning from 690 amu to 800 amu. Sample preparation Extraction according to 35, LMBG.13 Chromatographic conditions The Agilent 1100 Series LC/MSD proved to be capable of detecting and quantifying fumonisins at 250 picograms per component regardless of their chemical structure and without the need for derivatization during the sample preparation procedure. The Agilent 1100 Series LC/MSD provided optimum sensitivity in the selected ion monitoring mode. Even when operating in scan mode (150 amu to 800 amu), the Agilent 1100 Series LC/MSD still provided sensitivity more than a factor of 10 better than reported for a fluorescence detector.
19
334.4
100000 0
Zorbax Eclipse XDB-C18, 2.1 mm x 150 mm, 5 m Mobile phase A 5 mM ammonium acetate pH3 Mobile phase B acetonitrile Gradient 0 min 33% B 8 min 60% B 9 min 33% B Flow rate 250 l/min Injection vollume 5 l Column compartment 40C Ionization mode Nebulizer pressure Dryng gas temp. Drying gas flow Vcap. Fragmentor Scan range API-ES positive or APCI negative 30 psig 350C 6 l/min 4000 volts 100 volts m/z 120 820
200000 163.1 170.1 220.1 100000 0 200 250000 336.4 354.5 150000 50000 200 300 300 400 500 336.2 354.4 376.6
FB3
600 FB2
700 706.5
512.0
553.5
400
500
600
700
Figure 13 Mass spectra of Fumonisins B 1,2,3 when the fragmentor is ramped from 230 to 100V
3.237 MS EIC m/z 723
100000 60000 20000
FB1
3.228
250000 150000 50000
FB1
MS EIC m/z 335
100000 60000 20000 1
MS EIC m/z 337 6.241
6.248
300000 200000 100000
FB3 MS EIC m/z 707
7.683
FB2
FB3 7.675
FB2 9.862 8 Time [min]
Figure 14 Identification of different Fumonisin species in corn extract by retention time with further confirmation through fragment ion
13. Lebensmittel- und Bedarfsgegenstndegesetz, Paragraph 35, Germany.
20
707.3 708.7 728.5 m/z
728.5
704.7 728.5 750.5 769.5 m/z
706.5 707.5
LC/MS conditions Colum
370.5
411.6
723.5 500 600 700 m/z
352.4
200000
200
300
400
392.5
FB1
722.5
Mycotoxins
The following mycotoxins have been analyzed: aflatoxins G2, G1, B2, B1, M2, and M1; ochratoxin A; zearalenone; and patuline. Mycotoxins are highly toxic compounds produced by fungi. They can contaminate food products when storage conditions are favorable to fungal growth. These toxins are of relatively high molecular weight and contain one or more oxygenated alicyclic rings. The analysis of individual mycotoxins and their metabolites is difficult because more than 100 such compounds are known, and any individual toxin is likely to be present in minute concentration in a highly complex organic matrix. Most mycotoxins are assayed with thin-layer chromatography (TLC). However, the higher separation power and shorter analysis time of HPLC has resulted in the increased use of this method. The required detection in the low parts per billion (ppb) range 4,13 can be performed using suitable sample enrichment and sensitive detection. Sample preparation Samples were prepared according to official methods.13 Different sample preparation and HPLC separation conditions must be used for the different classes of compounds. The table on the next page gives an overview of the conditions for the analysis of mycotoxins in foodstuffs. Chromatographic conditions The HPLC method presented here for the analysis of mycotoxins in nuts, spices, animal feed, milk, cereals, flour, figs, and apples is based on reversed-phase chromatography, multisignal UV-visible diode-array detection, and fluorescence detection. UV spectra were evaluated as an additional identification tool.
21
2
Column class Matrix Sample preparation Chromatographic conditions Hypersil ODS, 100 2.1 mm id, 3-m particles water/methanol/ACN (63:26:11) as isocratic mixture** flow rate: 0.3 ml/min at 25 C DAD: 365/20 nm Fluorescence detector (FLD): excitation wavelength 365 nm, emission wavelength 455 nm Lichrospher 100 RP18, 125 4 mm id, 5-m particles water with 2 % acetic acid/ACN (1:1)* flow rate: 1 ml/min at 40 C FLD: excitation wavelength 347 nm, emission wavelength 480 nm Aflatoxins nuts, extraction G2 , G1 , B2 , B1 , spices, according to Para. M2 , M1 animal 35, LMBG*8,12 feed, milk, dairy products
Ochratoxin A
cereals, flour, figs
extraction according to Para. 35, LMBG acidify with HCl extract with toluene SiO2 cleanup elute toluene/acetic acid (9:1)
Zearalenone
cereals
extract with toluene Sep-pak cleanup
Hypersil ODS, 100 2.1 mm id, 3 m particles
water/methanol/ACN (5:4:1) isocratic mixture* elute toluene/acetone (95:5) flow rate: 0.45 ml/min at 45 C DAD: 236/20 nm AOAC 985.18:4 -zearalenol and FLD: excitation wavelength 236 nm, zearalenone in emission wavelength 464 nm corn Patuline apple products cleanup on Extrelut Superspher RP18, 125 4 mm id, 4-m particles silica gel cleanup water 5 %95 % ACN elute toluene/ flow rate: 0.6 ml/min at 40 C ethylacetate (3:1) DAD: 270/20 nm or Lichrospher diol, 125 4 mm id, 5-m particles hexane/isopropanol (95:5) as isocratic mixture flow rate: 0.6 ml/min at 30 C DAD: 270/20 nm
* Lebensmittel- und Bedarfsgegenstndegesetz, Germany ** 100 % B is recommended for cleaning the column
22
HPLC method performance Limit of detection Repeatability of RT over 10 runs of areas over 10 runs Linearity of UV-visible DAD of fluorescence 15 g/kg < 0.12 % < 1.5 % 1500 ng 30 pg to 2 ng
mAU 20 15 10 5 0 2 4 Time [min] 6 8 10

1
M2 5 ng
FLD:
em ex
365 nm 455 nm
DAD: 365 nm
Figure 15 Analysis of aflatoxins with UV and fluorescence detection
mAU 5 4 3 2 1
2 4 Time [min]
6
Pistachio nut
G1 5 ng
B1 5 ng
FLD DAD
8
Figure 16 Analysis of aflatoxins in pistachio nuts with UV and fluorescence detection
Water Methanol Quaternary AutoColumn pump + sampler compartvacuum ment degasser
Diodearray detector
13. Lebensmittel- und Bedarfsgegenstndegesetz, Paragraph 35, Germany. 4. Official Methods of Analysis, Food Compositions; Additives, Natural Contaminants, 15th ed; AOAC: Arlington, VA, 1990, Vol. 2.; AOAC Official Method 980.20: aflatoxins in cotton seed products; AOAC Official Method 986.16: Aflatoxins M1 , M2 in fluid milk; AOAC Official Method 985.18: -zearalenol.
23
2
Bisphenol A diglycidyl-ether (BADGE)
Bisphenol A diglycidyl-ether (BADGE) is present in the three most common coatings (epoxy lacquer, organosol lacquer and polyester lacquer) used to protect the inside surfaces of cans used for food packaging. In canned foods containing a high proportion of fat, BADGE tends to migrate into the fatty phase where it remains stable, whereas in water it is hydrolyzed. BADGE was originally determined to be mutagenic during in vitro tests but a later re-assessment, using in vivo tests, led to a different conclusion. While further tests are being performed, a maximum concentration of 1 mg BADGE per kg of food has been agreed. Sample preparation Extracted with water/alcohol 50/50 or n-heptane at reflux temperature for six hours. Chromatographic conditions A fast separation was developed by using the enhanced specificity provided by the Agilent 1100 Series LC/MSD in CID (collision induced dissociation) mode allowing the detection of BADGE via the molecular ion combined with confirmation using the most abundant fragment ion.
24
Zorbax Eclipse XDB-C8, 2.1 mm x 50 mm, 5 Mobile phase A 5 mM ammonium acetate in water, pH3 Mobile phase B acetonitrile Gradient 0 min 25% B 5 min 50 % B Flow rate 300 l/min Injection volume 1 l Column compartment 40 C Detector UV-DAD 210 nm/6 nm, ref. 360/60 nm 254 nm/6 nm, ref. 360/60 nm
7.992
LC/MS conditions Colum
mAU 2 0 -2 -4 -6 -8 -10 -12 500000
8.341
UV-Vis 230 nm
12.112 12.429 13.773 14.095 15.059 15.918
MS EIC m/z 358 300000 100000 0 2.5 5 7.5 10 12.5 Time [min] 15 17.5 20
Figure 17 Extract from tuna 0.2 ppm, 1 l injected
-10
500000 300000 100000 0 2.5
MS EIC m/z 358
7.5
10 12.5 Time [min]
10.870
15
16.509
Ionization mode Nebulizer pressure Dryng gas temp. Vcap. Fragmentor Scan range Scan speed
API-ES positive 50 psig 350 C 3500 volts 70 volts m/z 250 400 2 s/scan
10.815
0.574 0.769
5 0 -5
16.329 16.861 17.302 17.922
10
5.929
10.185 10.494
15.949
mAU
UV-Vis 230 nm
13.332 13.461 13.874 14.367 15.100 15.343
17.5
Figure 18 Extract from sardine 20 ppm, 1 l injected
20.712
20
25
2
Pesticides
The following compound classes of pesticides have been analyzed: triazines, phenylurea-herbicides, methabenzthiazuron, diquat, paraquat, and mercaptobenzothiazol. Carbamates and glyphosate also have been analyzed but with different equipment. In most countries, growing concern about the residues of pesticides in food products is evident. Therefore, regulations limiting the concentration of pesticides in foodstuffs have been introduced to protect consumers from contaminated food products. Several methods are used to control these limits. HPLC is recommended for the analysis of low volatile compounds and for compounds that are unstable when heated. Sample preparation Sample preparation and enrichment depend strongly on the matrix. Drinking water samples, for example, must be extracted using solid-phase extraction, whereas vegetables are extracted with liquid/liquid extraction after homogenization, followed by additional cleaning and sample enrichment.
Water
Acetronitrile
14. Specht, W. Organochlor- und Organophosphor-Verbindungen sowie stickstoffhaltige sowie andere Pflanzenschutzmittel, DFG-Methodensammlung, 1982, 19.
26
Chromatographic conditions The HPLC method presented here was used for the analysis of pesticides in salad samples and spices.
Sample preparation
Column Mobile phase Gradient
Flushing time Post time Flow rate Oven temperature Injection volume Detector
Salad was homogenized and then extracted with liquid/liquid extraction. The extract was cleaned with gel permeation chromatography using cyclohexane/ethyl acetate. Spices were prepared according to Specht14 with gel permeation chromatography. 100 3 mm Hypersil BDS, 3 m water/ACN (95:5) at 10 min 25 % ACN at 26 min 42 % ACN at 34 min 60 % ACN 10 min at 100 % ACN 6 min 0.5 ml/min 42 C 310 l UV-DAD detection wavelengths 214/15 nm, 230/20 nm, and 245/20 nm reference wavelength 400/80 nm
mAU 120
80
3 different salad samples
Vinclozolin
Folpet 40 0 10 15
20 25 Time [min] 30
Carbendazim*
35
40
Figure 19 Analysis of pesticide residues in three different salad samples * Carbendazim has a low recovery rate of only approximately 40 %
mAU 100
80 60
Vinclozolin Procymidon Nitro compounds Chlorpyripho-ethyl Paprika (Spain)
40
HPLC method performance Limit of detection 0.01 g/l Repeatability of RT over 10 runs < 0.2 % of areas over 10 runs < 1 %
20 0 10 20 30 Time [min] 40
Paprika (Turkey)
50
Figure 20 Analysis of pesticide residues in two paprika samples
27
2
Carbamates
Chromatographic conditions The HPLC method presented here was used for the direct analysis of carbamates in water with postcolumn derivatization.15 Fruits and vegetables must be extracted at neutral pH with water prior to HPLC analysis.
Sample preparation Column none 250 x 4 mm C18 phase from Pickering, 5 m Mobile phase water/methanol (MeOH, 88:12) Gradient at 2 min 12 % MeOH at 42 min 66 % MeOH at 46 min 66 % MeOH at 46.1 min 100 % MeOH at 49 min 100 % MeOH Flow rate 0.8 ml/min Column compartment 37 C Injection volume 10 l standard Fluorescence detector Excitation wavelength: 230 nm or 330 nm Emission wavelength: 425 nm Photomultiplier gain: 12 Response time: 4 s Derivatization reagent pump flow rate for hydrolization agent: 0.3 ml/min (NaOH) flow rate for derivatization agent: 0.3 ml/min (OPA)
%F 6 3 1 4.5 4 19 15 17 7 9 14
5.5
5
Sample A
20
4
3.5 10 %F 5.5 5 5 15 20 25 Time [min] 30 35 40 45
Sample B
10 8 11 18 21
4.5
4 3.5 10 15 20
23 22
12
13
16
25
30 Time [min]
35
40
45
Sample A 1 butocarboxim sulfoxide 2 aldicarb sulfoxide 3 butoncarboxim sulfone 4 aldicarb sulfone 6 methomyl 7 ethiofencarb sulfoxide
9 ethiofencarb sulfone 14 butocarboxim 15 aldicarb 17 propoxur 19 carbaryl 20 ethiofencarb
Sample B 5 oxamyl 8 thiofanox sulfoxide 10 thiofanox sulfone 11 3-hydroxycarbofuran 12 methiocarb sulfoxide 13 methiocarb sulfone
16 18 21 22 23
3-ketocarbofuran carbofuran 1-naphthol thiofanox methiocarb
Figure 21 Analysis of two different carbamate standards

HPLC method performance Limit of detection 100 ppt, S/N = 2 Repeatability of RT over 10 runs < 0.1 % of areas over 10 runs < 0.55 %
Water Methanol
Control and data evaluation Quaternary pump + vacuum degasser Autosampler Pickering post-column derivatization system Fluorescence detector
15. A new approach to lower limits of detection and easy spectral analysis Agilent Primer 5968-9346E, 2000
28
Glyphosate
Chromatographic conditions The HPLC method presented here was used for the direct analysis of glyphosate in water with postcolumn derivatization.16
none 150 x 4 mm cation exchange, K + form from Pickering, 8 m Mobile phase A = 5 mM KH2PO4 , pH = 2.0 B = 5 mM KOH Flow rate 0.4 ml/min Gradient at 15 min 0 % B at 17 min 100 % B Column compartment 55 C Injection volume 50 l standard Fluorescence detector Excitation wavelength: 230 nm or 330 nm Emission wavelength: 425 nm Photomultiplier gain: 12 Response time: 4 s Derivatization reagent pump flow rate for hydrolization agent: 0.3 ml/min (OCl*) flow rate for derivatization agent: 0.3 ml/min (OPA)
Sample preparation Column
Norm. 7.5 7
Glyphosate AMPA
6.5 6 5.5 5 2.5 5 7.5 10 12.5 Time [min] 15 17.5 20
Figure 22 Analysis of glyphosate standard
HPLC method performance Limit of detection 500 ppt Repeatability of RT over 10 runs < 0.8 % of areas over 10 runs < 2.2 %
Quaternary pump + vacuum degasser Pickering post-column derivatization system Fluorescence detector
Control and data evaluation Autosampler
Water
KOH
16. R. Schuster, A comparison of pre- and post-column sample treatment for the analysis of glyphosate, Agilent Application Note 5091-3621E , 1992.
29
30
Chapter 3
Analytical examples of natural components
3
Inorganic anions
Anions containing halogen, nitrogen, and sulfur are used as additives in food industries. For example, nitrites act as preservatives in smoked sausage. Nowadays, dedicated instrumentation such as special columns and electroconductivity detectors are used in the analysis of inorganic anions. Because specialized equipment has a very limited application range, a method was developed for analyzing anionsusing reversed-phase chromatography and indirect UV detection. Another, more selective and sensitive approach for the analysis of selected anions is electrochemical detection. Sample preparation Excepting filtration, sample preparation normally is unnecessary if the sample is aqueous. Other matrices can be extracted with hot water, followed by filtration.
Control and data evaluation Isocratic pump + vacuum degasser
AutoAutosampler sampler
Column compartment
Water/ACN
Variable wavelength detector
32
Chromatographic conditions The HPLC method presented here was used for the analysis of anions in drinking water.
Sample preparation Column filtration HP-IC (modifiers for the mobile phase are included) water/acetonitrile (ACN) (86:14), adjusted to pH = 8.6 with carbonate-free NaOH 1.5 ml/min 40 C 25 l UV-VWD detection wavelength 266 nm
mAU 100 80 60 40 20 0 -20
F-
HCO3 Cl - = 15 ppm Cl - NO2
Mobile phase
Drinking water
NO3 = 0.9 ppm Br NO3
SO2 4 = 40 ppm H2PO4

2
Flow rate Oven temperature Injection volume Detector
SO4
2-
Standard
6 Time [min]
10
HPLC method performance Limit of detection for UV-VWD 0.11 ppb with S/N = 2 and 25 l injected volume Repeatability of RT over 10 runs < 0.8 % areas over 10 runs < 1 %
Figure 23 Analysis of anions in drinking water with indirect UV-detection
33
3
Chromographic conditions for electrochemical detection The HPLC method presented here was used for the analysis of iodide in table salt.17
mV . 180 160 140 120 100 2 4 6 8 Time [min] 10 12 14 I-
IStandard
Table salt was dissolved in water. Column 200 x 4 mm Sperisorb ODS2, 5 m Mobile phase water with 5.2 g/l K2HPO4 + 3 g/l tetrabutylammoniumdihydrogenphosphat/ACN (85:15) Flow rate 1 ml/min Oven temperature ambient 24 C Injection volume 0.1 l Detector electrochemical (ECD) Electrode: glassy carbon, Working potential: 1 V Operation mode: amperometry
Sample preparation
Table salt
Figure 24 Analysis of iodide in table salt
HPLC method performance Limit of detection for ECD 40 g/l Repeatability of RT over 10 runs < 0.1 % areas over 10 runs 3 % Linearity min 50 pg to 150 ng
Column compartment
Electrochemical detector
Water
17. A.G. Huesgen, R. Schuster, Analysis of selected anions with HPLC and electrochemical detection, Agilent Application Note 5091-1815E, 1991.
34
Lipids
Both saturated and unsaturated triglycerides have been analyzed. Fats and oils are complex mixtures of triglycerides, sterols, and vitamins. The composition of triglycerides is of great interest in food processing and dietary control. Owing to the low stability of triglycerides containing unsaturated fatty acids, reactions with light and oxygen form hydroperoxides, which strongly influence the taste and quality of fats and oils. Adulteration with foreign fats and the use of triglycerides that have been modified by a hardening process also can be detected through triglyceride analysis. The HPLC method presented here was used to analyze triglycerides, hydroperoxides, sterols, and vitamins with UV-visible diode-array detection (UV-DAD). Spectra were evaluated in order to trace hydroperoxides and to differentiate saturated from unsaturated triglycerides. Unsaturated triglycerides in olive oil have a very distinctive pattern. Other fats and oils are also complex mixtures of triglycerides but exhibit an entirely different pattern. Adulteration with foreign fats and the use of refined triglycerides in olive oil also can be detected through triglyceride analysis. Sample preparation Triglycerides can be extracted from homogenized samples with petrol ether. Fats and oils can be dissolved in tetrahydrofuran.17
Triglycerides and hydroperoxides in oils
Water
Acetonitrile
35
3
Absorbance [scaled] 120 100 80 000 *
60 40 * Hydroperoxides
S00
* *
PLL
00L
140
LLL
Sample preparation
Samples were dissolved in tetrahydrofuran (THF). Column 200 x 2.1 mm Hypersil MOS, 5 m Mobile phase A = water B = ACN/methyltert.butylether (9:1) Gradient at 0 min 87 % B at 25 min 100 % B Post time 4 min Flow rate 0.8 ml/min Column compartment 60 C Injection volume 1 l standard UV absorbance 200 nm and 215 nm to detect triglycerides 240 nm to detect hydroperoxides 280 nm to detect tocopherols and decomposed triglycerides (fatty acids with three conjugated double bonds)
20 0 5
215 nm
240 nm 10 15 Time [min] 20

25
Figure 25 Triglyceride pattern of aged sunflower oil. The increased response at 240 nm indicates hydroperoxides
5 0 13.0
LLL
HPLC method performance Limit of detection for saturated triglycerides > 10 g for unsaturated triglycerides fatty acids with 1 double bond >150 ng fatty acids with 2 double bonds > 25 ng fatty acids with 3 double bonds < 10 ng Repeatability of RT over 10 runs < 0.7 % areas over 10 runs <6%
Olive oil LL0 00L 000 LL0 mAU 20 15 10 215 nm 280 nm 23.0 mAU 20 Good quality 15 S00 10 5 0 13.0 Time [min] 00L 000 Poor quality LL0 S00 215 nm 280 nm 23.0
Time [min]
Figure 26 Analysis of olive oil. The response at 280 nm indicates a conjugated double bond and therefore poor oil quality
36
Triglycerides in olive oil
Unsaturated triglycerides in olive oil have very characteristic patterns. Other fats and oils are also complex mixtures of triglycerides but with different patterns. Sample preparation information Triglycerides can be extracted from homogenized samples with petrol ether. Fats and oils can be dissolved in tetrahydrofurane. Chromatographic conditions The presented HPLC method was used to analyze the unsaturated triglycerides, LnLnLn, LLL, and OOO.18
Samples were dissolved in tetrahydrofurane. Column 200 2.1 mm Hypersil MOS, 5 m Mobile phase acetone/ACN (30:70) Flow rate 0.5 ml/min Column compartment 30 C Injection volume 2 l Detector refractive index
Sample preparation
mV 200 180 160 120 100 LnLnLn 000 LLL 140
80
HPLC method performance Limit of detection for ECD 50 g/l with S/N = 2 Repeatability of RT over 10 runs < 0.3 % areas over 10 runs 5 %
Rape oil Olive oil Standard 2 4 Time [min] 6 8
60 40
Figure 27 Analysis of the triglyceride pattern of olive and rape oil
Column compartment
Refractive index detector
Acetronitrile
18. Determination of triglycerides in vegetable oils, EC Regulation No. L248, 28ff.
37
3
Fatty acids
Saturated and unsaturated fatty acids from C4 through C22 have been analyzed. Fatty acids are the primary components of oils and fats and form a distinctive pattern in each of these compounds. For example, butter and margarines can be differentiated by the percentage of butyric acid in the triglycerides. To determine the fatty acid pattern of a fat or oil, free fatty acids first are obtained through hydrolysis. Derivatization is then performed to introduce a chromophore, which enables analysis of the fatty acids using HPLC and UV-visible detection. Sample preparation The triglycerides were hydrolyzed using hot methanol and KOH, followed by derivatization. Chromatographic conditions The HPLC method presented here was used in the analysis of the fatty acid pattern of dietary fat. The method involves hydrolysis with hot KOH/methanol and online derivatization with bromophenacyl bromide.
Control and data evaluation Quaternary pump + vacuum degasser Acetonitrile Autosampler Column compartment Variable wavelength detector
Water
38
Sample preparation 0.215 g fat was hydrolyzed with 500 l MEOH/ KOH at 80 C for 40 min in a thermomixer. After cooling 1.5 ml ACN/THF (1:1) was added, and the mixture was shaken for 5 min. The mixture was then filtered through a 0.45-m Minisart RNML from Satorius. Column 200 x 2.1 mm, MOS, 5 m Mobile phase A = water (70 %) B = (ACN + 1 % THF) (30 %) Gradient at 5 min 30 % B at 15 min 70 % B at 17 min 70 % B at 25 min 98 % B Flow rate 0.3 ml/min Column compartment 50 C Detector variable wavelength, 258 nm Derivatization 60 mg/ml bromophenacyl bromide was dissolved in ACN. Injector program for online derivatization 1. Draw 2.0 l from vial 2 (ACN) 2. Draw 1.0 l from air 3. Draw 1.0 l from vial 3 (derivatization agent) 4. Draw 0.0 l from vial 4 (wash bottle) (ACN/THF, 50:50) 5. Draw 1.0 l from sample 6. Draw 0.0 l from vial 4 (wash bottle) 7. Draw 1.0 l from vial 3 (derivatization agent) 8. Draw 0.0 l from vial 4 (wash bottle) 9. Draw 1.0 l from vial 5 (acetonitrile + 5 % TEA) 10. Draw 0.0 l from vial 4 (wash bottle) 11. Mix 9 l in air, 30 l/min speed, 10 times 12. Wait 2.0 min 13. Inject
mAU C18-3 1000 600 C16 200 15 20 Time [min] 25 Standard Dietary fat C18 C20 30 C14 C22 Standard C18-2 C18-1 1400
Figure 28 Analysis of a dietary fat triglyceride pattern. Overlay of one sample and two standard chromatograms
Norm 40 30 20 10 0 20 22 24 26 Time [min] 28 30 DAD C10, 9.9 ng
C12, 4.0 ng
C14, 3.0 ng
C16, 6.7 ng
C18, 4.5 ng
C20, 5.2 ng
VWD
Figure 29 Trace analysis of triglycerides with a diode-array and a variable wavelength detector in series
HPLC method performance Limit of detection 200 pg injected amount, S/N = 2 Repeatability of RT over 10 runs < 0.1 % areas over 10 runs 5 %
C22, 3.3 ng 32
39
3
Carbohydrates
The following carbohydrates have been analyzed: glucose, galactose, raffinose, fructose, mannitol, sorbitol, lactose, maltose, cellobiose, and sucrose. Food carbohydrates are characterized by a wide range of chemical reactivity and molecular size. Because carbohydrates do not possess chromophores or fluorophores, they cannot be detected with UV-visible or fluorescence techniques. Nowadays, however, refractive index detection can be used to detect concentrations in the low parts per million (ppm) range and above, whereas electrochemical detection is used in the analysis of sugars in the low parts per billion (ppb) range. Sample preparation Degassed drinks can be injected directly after filtration. More complex samples require more extensive treatment, such as fat extraction and deproteination. Sample cleanup to remove less polar impurities can be done through solid-phase extraction on C18 columns.
Column compartment
Refractive index detector
Water
40
Chromatographic conditions The HPLC method presented here was used to analyze mono-, di-, and trisaccharides as well as sugar alcohols.
Sample preparation Samples were directly injected. Column 300 x 7.8 mm Bio-Rad HPXP, 9 m Mobile phase water Column compartment 80 C Flow rate 0.7 ml/min Detector refractive index
Norm 800 600
Citric acid?
Lactose Glucose
Raffinose
Galactose Standard
400
HPLC method performance Limit of detection < 10 ng with S/N = 2 Repeatability of RT over 10 runs < 0.05 % areas over 10 runs 2 %
Lemonade
200 5 Time [min] 10 Fructose 15
Figure 30 Analysis of carbohydrates in lemonade

Lactose
Norm 180
Raffinose
Glucose Galactose
140 120 100 80
Fructose
160
Standard Cellbiose 5 Maltose 10

Time [min]
Corn extract Sucrose 15 Standard 20
Figure 31 Analysis of carbohydrates in corn extract 4. Official Methods of Analysis, Food Compositions; Additives, Natuaral Contaminants, 15th ed; AOAC: Arlington, VA, 1990, Vol. 2; AOAC Official Method 980.13: Fructose, glucose, lactose, maltose, sucrose in milk chocolate; AOAC Official Method 982.14: Glucose, fructose, sucrose, and maltose in presweetened cereals; AOAC Official Method 977.20: Separation of sugars in honey; AOAC Official Method 979.23: Saccharides (major) in corn syrup; AOAC Official Method 983.22: Saccharides (minor) in corn syrup; AOAC Official Method 984.14: Sugars in licorice extracts.
41
3
Vitamins
Fat-soluble vitamins, such as vitamins E, D, and A, and water-soluble vitamins, such as vitamins C, B6, B2, B1, and B12, have been analyzed. Vitamins are biologically active compounds that act as controlling agents for an organisms normal health and growth. The level of vitamins in food may be as low as a few micrograms per 100 g. Vitamins often are accompanied by an excess of compounds with similar chemical properties. Thus not only quantification but also identification is mandatory for the detection of vitamins in food. Vitamins generally are labile compounds that should not exposed to high temperatures, light, or oxygen. HPLC separates and detects these compounds at room temperature and blocks oxygen and light.19 Through the use of spectral information, UV-visible diode-array detection yields qualitative as well as quantitative data. Another highly sensitive and selective HPLC method for detecting vitamins is electrochemical detection. Sample preparation Different food matrices require different extraction procedures.19 For simple matrices, such as vitamin tablets, water-soluble vitamins can be extracted with water in an ultrasonic bath after homogenization of the food sample.
Water-soluble vitamins
Water
Acetonitrile
42
Chromatographic conditions for UV detection The HPLC method presented here was used to analysis vitamins in a vitamin drink.
filtration 100 x 4 mm Hypersil BDS, 3 m Mobile phase A= water with pH = 2.1 (H2SO4) = 99 % B = ACN 1 % Gradient at 3.5 min 1 % B at 11 min 25 % B at 19 min 90 % B Post time 6 min Flow rate 0.5 ml/min Column compartment 30 C Injection volume 25 l Detector UV-DAD detection wavelength 220/30 nm, reference wavelength 400/100 nm
1500 1000 500 0 0 2
Pantothenic acid
Norm
B1 Vitamin C
B6
Folic acid, d Riboflavin 5'phos B12 Riboflavin Biotin 12

Folic acid 350 450 550 nm
Sample preparation Column
Standard Vitamin tablet
Citric acid 4
Saccharin 10
8 6 Time [min]
Figure 32 Analysis of water-soluble vitamins in a vitamin tablet
HPLC method performance Limit of detection < 500 pg (injected amount), S/N = 2 Repeatability of RT over 10 runs < 0.2 % areas over 10 runs < 2 %
Norm
800 400 Riboflavin
Norm
400 200 0
250
350 Norm 1000

600 200
450
0 550 nm
250
550 nm
Vitamin B 1,B 6, B 12
250
350
450
Figure 33 Spectra of water-soluble vitamins
19. L.M. Nollet, Food Analysis by HPLC, New York, 1992.
43
3
Chromatographic conditions for electrochemical detection The HPLC method presented here was used in the analysis of vitamins in animal feed.20
mV 240 220 200 180 160 140 120 0 1 2 Time [min] 3 4 Vitamin B 6 Standard 5 6 Vitamin B 6
Vitamin C
Vitamin preparation was diluted with water 1:100 Column 125 x 4 mm, Lichrospher RP 18, 5 m Mobile phase water + 0.02 M KH2PO4 + 0.03 M tetrabutylammoniumhydrogensulfat + 0.03 M heptanesulfonic acid + 2 % ACN Stop time 15 min Flow rate 0.8 ml/min Column compartment 30 C Injection volume 1 l standard 0.5 l sample Detector electrochemical Working electrode: glassy carbon Operation mode: amperometry Working potential: 1.2 V Range: 0.5 A Reference electrode: AgCl/KCl Response time: 1s
Sample preparation
Figure 34 Analysis of vitamin B6 in a vitamin preparation
HPLC method performance Limit of detection 30 pg (injected amount) S/N = 2 Repeatability of RT over 10 runs < 0.5 % areas over 10 runs < 5 % Linearity 30 pg to 1 ng
Isocratic pump + vacuum degasser
Column compartment
Water
20. A.G. Huesgen, R. Schuster, Analysis of selected vitamins with HPLC and electrochemical detection, Agilent Application Note 5091-3194E , 1992.
44
Fat-soluble vitamins
Sample preparation Different food matrices require different extraction procedures. These procedures include alkaline hydrolysis, enzymatic hydrolysis, alcoholysis, direct solvent extraction, and supercritical fluid extraction of the total lipid content.
100 x 2.1 mm Hypersil MOS, 5 m Mobile phase A = water B = ACN (70 %) Gradient at 15 min 90 % B at 16 min 95 % B Post time 3 min Flow rate 0.5 ml/min Column compartment 40 C Injection volume 25 l Detector UV-DAD detection wavelengths 230/30 nm, 400/100 nm; reference wavelengths 280/40 nm, 550/100 nm
Column
Chromatographic conditions for UV detection The HPLC method presented here was used in the analysis of a vitamin standard.
mAU 700 600 500 400 300 200 100 0 2 4 6 8 Time [min] 10 -tocopherol -and -tocopherol Standards Vitamin D3 -tocopherol
HPLC method performance Limit of detection Repeatability of RT over 10 runs areas over 10 runs 1 ppb with S/N = 2 < 0.82 % < 2.2 %
12
14
Figure 35 Analysis of fat-soluble vitamins with UV detection
Diodearray detector
45
3
Chromatographic conditions for electrochemical detection The HPLC method presented here was used in the analysis of a vitamin standard.20
Column Mobile phase 125 x 4 mm Lichrospher RP18, 5 m methanol + 5 g/l lithiumperchlorate + 1 g/l acetic acid 20 min 1 ml/min 30 C 1 l standard electrochemical glassy carbon amperometry 0.9 V 0.5 A AgCl/KCl 8s
mV 118.5 118.0 117.5 117.0 116.5

0
Standard
Stop time Flow rate Oven temperature Injection volume Detector Working electrode: Operation mode: Working potential: Range: Reference electrode: Response time:
-tocopherol
4 Time [min]
10
Figure 36 Analysis of a fat-soluble vitamin with electrochemical detection
HPLC method performance Limit of detection 80 pg (injected amount), S/N = 2 Repeatability of RT over 10 runs < 0.5 % areas over 10 runs < 5 % Linearity 30 pg to 1 ng
Column compartment
Water
Analysis of tocopherols on normal-phase column
Tocopherols cannot be separated completely using reversed-phase chromatography. However, normal-phase chromatography can separate isocratically all eight tocopherols (T) and tocotrienols (T3 ) naturally occurring in fats, oils, and other foodstuffs. Fluorescence detection is recommended for the analysis of total lipid extraction because UV absorbance detection is not selective enough to prevent detection of coeluting peaks.
46
Chromatographic conditions for analysis of tocopherols on normal-phase column The HPLC method presented here was used in the analysis of margarine.
Sample preparation 20 g sample dissolved in 15 ml hexane Column 100 x 2.1 mm Hypersil SI 100, 5 m Mobile phase hexane + 2 % isopropanol Stop time 8 min Flow rate 0.3 ml/min Column compartment 25 C Injection volume 0.5 l Detector UV-DAD 295/80 nm Fluorescence excitation wavelength 295 nm, emission wavelength 330 nm
mAU 20 15 10 5 0 1 2
-tocopherol
-tocopherol -tocopherol -tocopherol FLD DAD 3 4 Time [min] 5 6 7
Figure 37 Analysis of tocopherols on normal phase using UV and fluorescence detection
HPLC method performance Limit of detection 1020 ng, S/N = 2 for diode-array Limit of detection 0.52 ng S/N = 2 for fluorescence Repeatability of RT over 10 runs <2% areas over 10 runs < 2 %
%F 90 70 50
-tocopherol
-tocopherol
-tocopherol -tocopherol 77.3 % 11.2 % 9.5 % Standard
30 10
1.9 %
Margarine
Hexane
Isocratic pump + vacuum degasser Autosampler Column compartment Diodearray detector
4 Time [min]
Figure 38 Analysis of tocopherol concentration in margarine fat extract with fluorescence detection
47
3
Biogenic amines
The following amines were analyzed: ammonia, amylamine, 1-butylamine, 1,4-diaminobutane, 1,5-diaminopentane, diethylamine, ethanolamine, ethylamine, hexylamine, histamine, isobutylamine, isopropylamine, methylamine, 3-methylbutylamine, morpholine, phenethylamine, propylamine, pyrrolidine, and tryptamine. Free amines are present in various food products and beverages, including fish, cheese, wine, and beer. High concentrations of specific amines can have toxic properties. As a result, several countries have set maximum tolerance levels for these compounds in foodstuffs. HPLC is now preferred for the analysis of amines in food matrices because of its shorter analysis time and relatively simple sample preparation. Sample preparation Amines can be extracted from different matrixes using liquid/liquid extraction or solid-phase extraction followed by derivatization.
Control and data evaluation Quaternary pump + vacuum degasser Autosampler Column compartment Variable wavelength detector
Water
Acetonitrile
48
Chromatographic conditions for UV detection The HPLC method presented here was used to analyze amines in wine.21
Sample preparation 25 ml wine was decolored with polyvinylpyrrolidoine. After filtration, the amines (5 ml sample, pH = 10.5) were derivatized with 2 ml dansyl chloride solution (1 %). The reaction solution was cleaned with solid-phase extraction using C18 cartridges (500 mg). After elution with 2 ml ACN, the solution was concentrated to 100 l. Column 250 x 4.6 mm Spherisorb ODS2, 5 m Mobile phase A = water + 5 % ACN = 75 % B = ACN (25 %) Gradient at 5 min 45 % B at 30 min 45 % B at 50 min 60 % B at 55 min 80 % B at 60 min 80 % B Stop time 60 min Post time 4 min Flow rate 1 ml/min Column compartment 60 C Detector UV-VWD 250 nm
mAU 8.0e4
4 7 8 10
14 5 16 15 13 11 12 18
Standard
6.0e
20 17 19 60
15 16 17 18 19 20 amylamine 1,4-diaminobutane 1,5-diaminopentane hexylamine histamine heptylamine (internal standard)
4.0e4 2 2.0e4 20
1 2 3 4 5 6 7 ethanolamine ammonia methylamine ethylamine morpholine i-propylamine propylamine
Time [min] 40
8 9 10 11 12 13 14 pyrrolidine i-butylamine 1-butylamine tryptamine diethylamine phenethylamine 3-methylbutylamine
Figure 39 Analysis of amine standard with UV detection after derivatization
HPLC method performance Recovery rate > 85 % Limit of detection 50150 g/l Method repeatability for 5 red wine analyses < 5 % Linearity 500 g/l to 20 mg/l
21. O. Busto, et al., Solid phase extraction applied to the determination of
biogenic amines in wines by HPLC, Chromatographia, 1994, 38(9/10), 571578.
49
3
Amino acids
Both primary and secondary amino acids were analyzed in one run. The amino acid composition of proteins can be used to determine the origin of meat products and thus to detect adulteration of foodstuffs. Detection of potentially toxic amino acids is also possible through such analysis. Through the use of chiral stationary phases as column material, D and L forms of amino acids can be separated and quantified. HPLC in combination with automated online derivatization is now a well-accepted method for detecting amino acids owing to its short analysis time and relatively simple sample preparation. Sample preparation Hydrolyzation with HCl or enzymatic hydrolysis is used to break protein bonds. Chromatographic conditions The HPLC method presented here was used in the analysis of secondary and primary amino acids in beer with precolumn derivatization and fluorescence detection.22, 23
Fluorescence detector Quaternary pump + vacuum degasser
Autosampler
Column compartment
Diodearray detector
Water
Acetonitrile
50
Sample preparation Column Mobile phase
filtration 200 x 2.1 mm Hypersil ODS, 5 m A = 0.03 M sodium acetate pH = 7.2 + 0.5% THF B = 0.1 M sodium acetate/ ACN (1:4)
ARG
ILE
mAU ALA 70 GLY 60 50 40 30 20 ASP 10 0 GLU
TYR
PHE
VAL
PRO
ABA
LEU
Gradient at 0 min 0 % B at 0.45 ml/min flow rate at 9 min 30 % B at 11 min 50 % B at 0.8 ml/min flow rate at 13 min 50 % B at 14 min 100 % B at 0.45 ml/min flow rate at 14.1 min at 0.45 ml/min flow rate at 14.2 min at 0.8 ml/min flow rate at 17.9 min at 0.8 ml/min flow rate at 18 .0 min at 0.45ml/min flow rate at 18 min 100 % B at 19 min 0 % B Post time 4 min Flow rate 0.45 ml/min Column compartment 40 C Injection volume 1 l standard Detector UV -DAD 338 nm and 266 nm Fluorescence Excitation wavelength: 230 nm Emission wavelength: 450 nm at 11.5 min Excitation wavelength: 266 nm Emission wavelength: 310 nm Photomultiplier gain: 12 Response time: 4s
WL switch CIT MET
HIS GLN SER ASN 2 4
LYS 10
8 Time [min]
12
Figure 40 Analysis of amino acids in beer after online derivatization

Injector program for online derivatization 1. Draw 3.0 l from vial 2 (borate buffer) 2. Draw 1.0 l from vial 0 (OPA reagent) 3. Draw 0.0 l from vial 100 (water) 4. Draw 1.0 l from sample 5. Draw 0.0 l from vial 100 (water) 6. Mix 7.0 l (6 cycles) 7. Draw 1.0 from vial 1 FMOC reagent 8. Draw 0.0 l from vial 100 (water) 9. Mix 8.0 l (3 cycles) 10. Inject
HPLC method performance Limit of detection DAD < 5 pmol FLD < 100 fmol Repeatability of RT over 6 runs <1% areas over 6 runs <5% Linearity DAD 1 pmol to 4 nmol
22. Sensitive and reliable amino acid analysis in protein hydrolysates using the Agilent 1100 Series, Agilent Technical Note 5968-5658E, 1999 23. R. Schuster, Determination of amino acids in biological, pharmaceutical, plant and food samples by automated precolumn derivatization and HPLC, J. Chromatogr., 1988, 431, 271284.
51
3
Peptides
Peptide mapping of phytochrome from dark grown oat seedlings using capillary liquid chromatography The analyzed phytochrome is a photoreceptor protein that controls light-dependent morphogenesis in plants. For example, potato clod forms pale long sprouts if it germinates in a dark cellar. However, if this process takes place in the light, a normal plant with green leaves grows and photosynthesis occurs. Phytochrome proteins are present in very low concentrations in potato clod, and sample volume and concentration of these proteins is rather low following sample preparation. In this case, columns or capillaries with a small internal diameter are preferred because sensitivity increases with decreasing internal diameter of the column. The use of capillaries with an internal diameter of 100300 m enables flow rates as low as 0.54.0 l/min, which reduces solvent consumption. Such flow rates are well-suited to liquid chromatographymass spectroscopy (LC/MS) electrospray ionization. In our experience, the appropriate conversion of standard HPLC equipment to a capillary HPLC system is cost-effective and yields the highest performance for running capillary columns. For conversion, a flow stream-split device, a 35-nL capillary flow cell for the detector, and capillary connections between system modules are required. System delay volume should be as low as possible. To meet the demands of such a system, the Agilent 1100 Series binary pump, which has inherently low delay volume, was selected as a pumping system. The flow splitter, the capillary flow cell for the detector, and the column were purchased from LC Packings in Amsterdam.24 With this design, a standard flow rate (for example, 100 or 50 l/min) can be set for the pump. This flow then can be reduced by calibrated splitters between 0.5 and 4 l/min, for example. This flow rate is optimal for capillary columns with an internal diameter of 300 m.
52
Chromatographic conditions Capillary HPLC with UV and MS detection has been used in the analysis of phytochrome protein from dark grown oat seedlings. Figures 41, 42 and 43 show the UV and total ion chromatogram together with two mass spectra of selected fragments. The Agilent 1100 Series LC system was used without mixer. All tubings were as short as possible, with an internal diameter of 75120 m id. Sample preparation The extracted protein was reduced and alkylated prior to digestion with trypsin.
tryptic digest of phytochrome from oat seedlings, 7 pmol/l Capillary column 300 m x 25 cm, C18 Mobile phase A = 0.025 % TFA in water B = 0.02 % TFA in ACN Gradient 0.35 % B/min Flow rate 100 l/min split to 4 l/min Column compartment 25 C Injection volume 2.5 l Detector UV-VWD wavelength 206 nm with a 35-nl, 8-mm flow cell
Sample
mAU 120 100 80 60 40 20 40 60

Time [min]
80
100
HPLC method performance Limit of detection 1 pmol Repeatability of RT over 10 runs < 0.7 % areas over 6 runs <1%
Figure 41 Capillary LC-MS of a phytochrome tryptic digest (17.5 pmol)UV trace
Control and data evaluation Binary pump + vacuum degasser Flow split device Autosampler Column compartment Mass spectrometer or VWD detector
Water Acetonitrile
53
3
MS data was used for further evaluation. Some of the tryptic mass fragments of the phytochrome are signed. As an example, figure 42 shows two mass spectra.
Voltages Scan Threshold Sampling Stepsize Drying gas Nebulizer Vcyl -5500, Vend -3500, Vcap -4000, CapEx 150 4001800 m/z 150 1 0.15 amu nitrogen, 150 C gas nitrogen, < 20 psi
T15 160000 140000 T12 T60-61
Abundance
120000 100000 80000 60000 40000 20000 40
T8 T92
T46
T42 T14
T58
The Agilent 5989B MS engine was equipped with an Iris Hexapole Ion Guide
50
60
70 80 Time [min]
90
100
110
Figure 42 Capillary LC-MS of a phytochrome tryptic digest (17.5 pmol)total ion chromatography (TIC)
415.4 10000 Abundance 8000 6000 4000 2000 0 450 550 650 750 850 m/z Time [min] 829.7 T12 (MW = 828.5) Abundance 6000 5000 4000 3000 2000 1000
796.6 T58 (MW = 2387.2)
1194.7
500 700 900 1100 1300 m/z Time [min]
Figure 43 Mass spectra of T12 and T58
24. Capillary Liquid Chromatography with the Agilent 1100 Series Modules and Systems for HPLC, Agilent Technical Note 5965-1351E , 1996.
54
55
The Equipment Basics
Part Two
An overview of the hardware and the software components needed for successful HPLC, and an introduction to the analytical techniques that have become routine in food analysis
Chapter 4
Separation in the liquid phase
4
Liquid chromatography offers a wide variety of separation modes and mobile phases for optimizing your separation system.
Separation mechanisms
Stationary phases can be classified according to the mechanism by which they separate molecules: partition phases adsorption phases ion-exchange phases size-exclusion phases Nowadays the most popular column material is reversed phase, in which separation is achieved through partition and through adsorption by unprotected silanol groups. In reversed-phase chromatography, the stationary phase is nonpolar (or less polar than in the mobile phase) and the analytes are retained until eluted with a sufficiently polar solvent or solvent mixture (in the case of a mobile-phase gradient).
Reversed-phase materials
Reversed-phase materials have wide application and a long lifetime. Moreover, these media have good batch-to-batch reproducibility, low equilibration times, high mechanical stability, and predictable elution times and elution order. Reversed-phase chromatography is frequently used in food analysis, as shown in part one of this primer. Compared with reversed-phase media, ion-exchange materials have a shorter lifetime, are less mechanically stable, and take longer to equilibrate. These columns have limited application in food analysis and are used primarily for inorganic cations and anions or for glyphosate.
Ion-exchange materials
58
Size-exclusion gels
Size-exclusion chromatography is used for sample cleanup and fractionation and is described in more detail in chapter 5 (Sample preparation). Adsorption chromatography is used for sampling and cleanup. For example, flavonoids from plant material can be cleaned, fractionated, and enriched on alumina. Other examples are given in chapter 5.
Adsorption media
The advent of narrow-bore columns
%F 140 120 100 80 60 40 20 0 5
250 x 2.1 mm id column
Discussions of HPLC methods often revolve around the internal diameter (id), or bore, of the column to be used. Standard-bore columns have an internal diameter of approximately 4 or 5 mm, whereas narrow-bore columns have an internal diameter of approximately 2 mm. When packed with the same materials as the standard-bore column, the narrow-bore column can achieve the same resolving power with less solvent because the analytes can be eluted at a lower flow rate (< 0.5 ml/min) than the 23 ml/min required for standard-bore columns. In addition, narrow-bore columns are 46 times more sensitive using the injection volume required for a standard-bore column (see figure 44). Narrow-bore columns nonetheless place higher demands on the equipment used than standard-bore columns. First, the HPLC pump must yield these low flow rates in a way that is both reproducible and precise. Second, all capillary connections, that is, from injector to column and from column to detector, must be kept to a minimum. Third, because column frits block more often, guard columns are recommended. An HPLC system designed for narrow-bore columns (low dead volume and high-performance pumping system) can achieve solvent economies of more than 60 % as well as improve detection limits with the same injection volume. Moreover, under the same conditions, a standardbore column may have higher resolving power.
250 x 4.6 mm id column
10 15 Time [min]
20
Figure 44 Effect of bore dimensions on separation
59
4
Influence of column temperature on separation Many separations depend not only on the column material and mobile phases but also on the column temperature. In such cases, column temperature stability is the dominating factor for the elution order. A thermostatted column compartment using Peltier control with good ambient temperature rejection ensures stable chromatographic conditions. Periodic fluctuations in room temperature during 24-hour use influence these conditions. Figure 45 shows the advantage of Peltier control over conventional air cooling.
Retention time Agilent 1100 Series thermostatted column compartment
71.20 71.00 70.80 70.60 70.40 1 2 3 4 5 Run number 6 7 8 9 10 Conventional column oven Day Night Day
Figure 45 Comparison of Peltier and conventional cooling as demonstrated using retention time fluctuations of a peptide peak over a sequence of 10 consecutive tryptic peptide maps
In brief
Reversed-phase stationary phases are the most popular LC media for the resolution of food mixtures. The use of narrow-bore columns can result in gains in sensitivity and reduced solvent consumption. For example, these columns have been applied successfully in the analysis of aflatoxins and fatty acids.
60
Chapter 5
Sample preparation
5
The isolation of analytes from other matrix constituents is often a prerequisite for successful food analysis. The broad selection of cleanup and enrichment techniques takes into account the many matrices and compound classes under study.
Sample preparation steps
Sample preparation for HPLC can be broken down into the following main steps: 1. Sampling Collection Storage 2. Cleanup/enrichment offline Homogenization, centrifugation, precipitation, hydrolyzation, liquid/liquid extraction, solid-phase extraction, ultrasonic bath liquid extraction, supercritical fluid extraction, concentration 3. Cleanup/enrichment online Guard columns Online solid-phase extraction Gel-permeation chromatography (GPC) 4. Chemical derivatization Precolumn, online, or offline (see also discussion of postcolumn derivatization, chapter 9)
Automation
Manual extraction, cleaning, and concentration of the sample prior to transfer to the HPLC instrument is time-consuming and can drain resources. Sample preparation therefore should be automated where possible. Nowadays the sample can be fractionated and/or derivatized automatically.
62
Supercritical fluid extraction (SFE) systems and automated solid-phase extraction equipment also have been interfaced directly to liquid chromatographs. Equipment used to automate preparation of HPLC samples includes: ValvesValves are used to switch to guard columns and online solid-phase extraction techniques.25 Switching valves are common in HPLC, and some instruments even have built-in column compartment valves. With a six-port valve, for example, the eluant stream can be switched from one column to another to cut out a peak. This peak is then analyzed on the second column. SFE interfacesThis technique is rather new, and online systems are under development.26 An offline procedure has been used successfully in the analysis of vitamin K in infant formula.27 Precolumn derivatizationThis well-accepted and commercially available technique28 has been applied in the analysis of amino acids in beer (see page 50 ff.). Reagents also can be used postcolumn (see page 28). Automated solid-phase extractionThis relatively new technique is used to analyze bittering compounds in beer.29
Solids
Solid samples, for example chocolate or meat, should be homogenized before such techniques as steam distillation, SFE, or ultrasonic-stimulated liquid extraction are applied.30 Ultrasonic bath liquid extraction Ultrasonic bath liquid extraction is a very simple extraction method. Selectivity is achieved through the use of appropriate solvents. Antioxidants and preservatives can be extracted with this technique if the matrix is low in fat.
63
Uses relatively small quantities of organic solvents, thereby reducing costs and facilitating disposal. Extraction times are in minutes rather than hours.
Samples high in fat cannot be extracted.
Steam distillation
Steam distillation is only used to extract volatile compounds from solid homogenized matrices. For example, biphenyl and 2-phenylphenol pesticides can be extracted from citrus fruits with this technique.31
Enables selective extraction of volatile compounds.
Extraction times are long and offline. Narrow range of use.
Supercritical fluid extraction
Until now, supercritical fluid extraction (SFE) was rarely used in food analysis. However, the input of modern SFE instruments can be automated with sampling devices. This method is used primarily for GC,26, 32 although LC coupling also has been performed with SFE.33
Uses small quantities of organic solvents, thereby reducing costs and facilitating disposal. Extraction times are in minutes rather than hours. Can be automated.
Weak solvating power limits range of analytes. Ultrapure fluids for trace analysis are not always available.
64
Liquids
Liquid-liquid extraction
Liquid-liquid extraction, on- and offline solid-phase extraction, and GPC are used in the analysis of liquid samples or extracts from solid samples. Liquid-liquid extraction is the most common extraction method. It requires an appropriate solvent and a separating funnel, or a continuous or counter-current distribution apparatus
Simple, with highly selective modifiers (pH, salts, or ion-pairing reagents).
Requires large amounts of toxic solvents, can emulsify, and is difficult to automate.
Solid-phase extraction
Suitable for cleaning clear liquids such as filtered beverages, solid-phase extraction (SFE) is simple in principle. The sample is first sucked through a preconditioned cartridge or disk filled with adsorbents. The solid then traps the compounds of interest, which can be extracted later with a small amount of an organic eluant. A variety of materials provides a choice of selectivities for use as a fractioning tool. Two or more separate cartridges filled with specific adsorption materials can trap individual fractions of the sample. SPE is one of the fastest-growing sample preparation and cleanup techniques.34 Attempts have been made to automate both the procedure and its interface with the chromatograph. Systems based on robotics and valves are available. Pumping a certain volume of water sample through one or more precolumns filled with extraction materials will extract and concentrate the compounds of interest. After desorption with a suitable solvent, the analytes can be introduced into a liquid or gas chromatograph for identification and quantification. The precolumns are
65
5
exchanged automatically between analyses to prevent clogging and memory effects.35 So far this system has been used only to extract pesticides and polynuclear aromatics in river water. A different online solid-phase extraction system has been used to extract and analyze iso-a-acids in beer.36, 40
Uses small amounts of organic solvents, can run several samples at once, and can be automated.
Differing batch-to-batch efficiencies can reduce reproducibility. Risk of irreversible adsorption. Degradation by surface catalysis can occur.
Gel permeation chromatography
Also known as size-exclusion chromatography, gel permeation chromatography (GPC) has become a standard technique for isolating compounds of low molecular weight from samples that contain compounds of high molecular weight, such as oil or fats. The separation is based on differences in size, with higher molecular weight compounds retained less than smaller ones. GPC has been used successfully to separate vitamins A, D, and E from glycerides in infant formula and clean-up of pesticides in spices (see chapter 2, page 22 ff).37
Highly reproducible, good automation possibilities.
Large amounts of solvents needed, separation efficiency may differ from batch to batch.
66
Guard columns
A guard column is connected in front of the analytical column to prevent contamination of the analytical column by the matrix. Either the guard column can be included in analytical column design, or both columns can be interconnected by a valve that, when switched, transfers fractions from the precolumn to the analytical column. The latter technique is more flexible and can be used for sample cleanup and enrichment. Alternatively, a backflush valve can be used to enrich the sample on a precolumn. Reversing the direction of flow transfers compounds concentrated from the precolumn to the analytical column.
Highly reproducible, good automation possibilities.
More complex, and more expensive if a valve is used.
In brief
Many food analyses are governed by officially recognized methods, which often include details on sample preparation. Recent trends toward automated sample preparation increase precision by eliminating operator variances. Should you adopt a newly developed sample preparation technique, however, please be aware that the method must comply with existing good laboratory practice (GLP) regulations and with accreditation standards.38
67
68
Chapter 6
Injection techniques
6
After the sample has been prepared for introduction onto the LC column, analysis can begin. Judgements based on analyte concentration require a reliable quantity of sample volume. The process of introducing the sample onto the column with precision syringes can be automated for increased throughput.39, 40
Characteristics of a good sample introduction device

Normal
The main requirements for any sampling device are good precision of injection volumes, low memory effects (carry-over of material from one injection to another), and the ability to draw viscous samples and inject variable volumes. Modern sampling systems can further increase productivity with features such as online precolumn derivatization for selective detection, heating and cooling for improved stability, and microsampling of material in low supply. Some analyses may require corrosive solvents or mobile-phase additives such as 0.1 N HCl or 60 % formic acid. Some vendors supply devices of corrosion-resistant titanium to solve this problem. Injection systems often are based on a six-port valve, which is put through several steps for each injection, as illustrated in figure 46. In the first step, denoted here as load, the sample is either aspired by a vacuum (in automated systems) or expressed by a syringe plunger (in manual systems) into a sample loop, where it rests until the valve is switched to inject. This second step connects the pump and the mobile phase with the column. The contents of the sample loop then move into the solvent flow path and onto the analytical column. Because all parts of the system are constantly flushed during analysis, the remnants of a previous injection are removed before the next injection occurs.
Load
Inject
Figure 46 A typical 6-port injection valve
70
The quality of the separation on the column depends on the quality of the injectiona short, sharp injection increases the likelihood of short, sharp peaks. The use of a minimum number of fittings between the injector and the column reduces the diffusion of the contents of the sample loop into the mobile-phase fronts in front of and behind the column. So-called low dead volume fittings with the minimum required internal capacity are available. These fittings have no dead ends or unnecessary spaces where solvent and sample can mix.
Manual injectors
Simple manual injectors remain popular in some laboratories because they are inexpensive and because their operation requires little previous experience, which is important if the equipment is used infrequently. With a precision syringe, the operator can fill the sample loop at atmospheric pressure by injecting the contents into the injection port. A switch of the rotor attached to the valve realigns the valve ports to the inject position. Solvent from the pump then flushes the contents of the sample loop onto the column. Continual flushing during the run keeps the injection port and valve clear of remnants of previous samples.
Inexpensive.
No automation and no provision for online derivatization. Syringe must be cleaned manually, offline.
71
6
Automated injectors
Automated injectors contain a mechanically driven version of the same six-port valve found in manual injectors. Pneumatic or electrical actuators control the valve as it switches between steps in the injection cycle. A metering device can handle injection volumes of 0.11500 l. With sample loops of larger capacity, such a device can inject up to 5 ml. Vials designed to hold microliter volumes can be used to inject as little as 1 l of a 5-l sample. Even the way the needle enters the vial can be controlled with computer software: deep down to aspire from the denser of two layers, or a shallow dip into the supernatant phase. With this technique, even viscous samples can be analyzed if the right equipment is used. The time required to extract the syringe plunger is simply protracted, permitting meniscus motion of higher reproducibility.
Highly reproducible. Can be fully automated. Flow maintained over all parts in contact with sample, preventing inaccuracies from intermingling between runs.
Equipment can be costly.
Autosampler with sample pretreatment capabilities
Autosamplers can provide online precolumn derivatization, dilute small volumes of sample, and add internal standards. You may need to protect unstable species by keeping the sample chilled with a cooling device connected to a flow of refrigerated water or, more conveniently, by Peltier elements. Alternatively, you might want to induce reactions using heat applied within the injection device of the autosampler. Commercially available autosamplers offer all these features.
72
Derivatization
As discussed later in chapter 8, derivatization may be required if the analytes lack chromophores and if detection is not sensitive enough. In this process, a chromophore group is added using a derivatization reagent. Derivatization can occur either in front of or behind the analytical column and is used to improve sensitivity and/or selectivity. Precolumn derivatization is preferable because it requires no additional reagent pump and because reagents can be apportioned to each sample rather than pumped through continuously. Automated precolumn derivatization yields excellent precision. Moreover, it can handle volumes in the microliter range, which is especially important when sample volume is limited. The principles involved are illustrated in figure 47.
Sample Reagent
Reagent Metering device Sampling unit
6-port valve
From pump
To column To waste
Figure 47 Automated precolumn derivatization
The robotic arm of the autosampler transports, in turn, a sample vial and several reaction vials under the injection needle. The needle is extended by a length of capillary at the point at which the derivatization reaction takes place.
73
6
The injector draws distinct plugs of sample and derivatization reagent into the capillary. The back-and-forth movement of the plunger mixes the plugs. With the right software, the autoinjector can be paused for a specified length of time to allow the reaction to proceed to completion. If the reaction requires several reagents, the autosampler must be programmable, that is, it must be able to draw sequentially from different reagent vials into one capillary. In this complex sample manipulation, the needle must be cleaned between vials, for example by dipping into wash vials of distilled water.
In brief
Automated sampling systems offer significant advantages over manual injectors, the most important of which is higher reproducibility of the injection volume. Sample throughput also can be increased dramatically. Modern autosamplers are designed for online sample preparation and derivatization. For food analysis, an automated injection system is the best choice.
74
Chapter 7
Mobile phase pumps and degassers
7
The pump is the most critical piece of equipment for successful HPLC. Performance depends strongly on the flow behavior of the solvent mixture used as mobile phasevarying solvent flow rates result in varying retention times and areas. Conclusions from a calibration run for peak identification or quantification depend on reproducible data. In this chapter we discuss multiple aspects of pump operation, including solvent pretreatment and its effect on performance.
Characteristics of a modern HPLC pump
A modern HPLC pump must have pulse-free flow, high precision of the flow rates set, a wide flow rate range, and low dead volume. In addition, it must exhibit control of a maximum operating pressure and of at least two solvent sources for mobile-phase gradients, as well as precision and accuracy in mixing composition for these gradients. We discuss two gradient pump types: that constructed for flow rates between 0.2 and 10 ml/min (low-pressure gradient formation), and that designed for flow rates between 0.05 and 5 ml/min (high-pressure gradient formation). In separating the multiple constituents of a typical food sample, HPLC column selectivity with a particular mobile phase is not sufficient to resolve every peak. Changing the eluant strength over the course of the elution by mixing increasing proportions of a second or third solvent in the flow path above the column improves peak resolution in two
Flow ranges
Gradient elution
76
ways. First, resolution is improved without extending the elution period, which prevents long retention times (peaks that have been retained on the column for a longer period of time tend to broaden and flatten through diffusion, lowering the S/N and therefore detection levels). Second, gradient elution sharpens peak widths and shortens run time, enabling more samples to be analyzed within a given time frame. The solvents that form the gradient in front of the column can be mixed either after the pump has applied high pressure or before, at low pressure. Gradient formation at high pressure If mixing takes place after pressure has been applied, a high-pressure gradient system results (this is most often achieved by combining the output of two isocratic pumps, each dedicated to one solvent).
Ability to form sharp gradient profiles and to change solvents rapidly (100% A to 100% B), without degassing, for standard applications.
Expensive. An additional mixer for lowest mixing noise at flow rates below 200 l is needed for mobile-phase compositions.
Gradient formation at low pressure
At low pressure, mixing of the gradient solvents occurs early in the flow path before the pump applies pressure, as in the two examples below.
Less expensive than gradient elution. Can mix more than two channels. Low mixing noise without a dedicated mixer.
Degassing is necessary for highest reproducibility.
77
7
Pump designs for gradient operation
Low-pressure gradient Agilent 1100 Series pump In food analysis, pump performance is critical. In the examples, we describe a low-pressure gradient system and a high-pressure gradient system, both of which perform according to food analytical requirements. The former has a single dual-piston mechanism for low-pressure gradient formation, whereas the latter has a double dual-piston mechanism for high-pressure gradient formation. After passing the online vacuum degasser, the mobile phase enters the first pump chamber through an electronically activated inlet valve (see figure 48). Active valves resolve the problem of contaminated or sticky ball valves by making the pump easy to prime. Output from the first piston chamber flows through a second valve and through a low-volume pulse dampener (with pressure transducer) into a second piston chamber. Output from the second chamber flows onto the sampling unit and column. The pistons in the pump chambers are motor driven and operate with a fixed-phase
0.08% 0.09% 0.09% 0.10% 0.11% 0.15% 0.07% 0.08% 0.08% 0.10% 0.08% 0.09% 0.09%
Vacuum chamber
From solvent bottles Damper Purge valve Outlet valve Proportioning valve Inlet valve To sampling unit and column
To waste
Figure 48 Low pressure gradient pump
mAU 50 40 30 20 10 0 5 10
0.08% 0.08%
Time [min]
15
20
25
Figure 49 Retention time precision (% RSD) of 10 injections of a polycyclic aromatic hydrocarbon (PNA) standard sample
78
difference of 180, so that as one delivers mobile phase, the other is refilling. The volume displaced in each stroke can be reduced to optimize flow and composition precision at low flow rates. With solvent compressibility, compensation, and a low-volume pulse dampener, pulse ripple is minimal, resulting in highly reproducible data for retention times and areas (see figure 49). A wide flow range of up to 10 ml/min and a delay volume of 8001100 l support narrow-bore, standard-bore, and semipreparative applications. Four solvents can be degassed simultaneously with high efficiency.
mAU mAU 80 60 40 20 0 0 5 10 15 Time [min] 20 0 5 10 15 20
80 60 40 20 0
Time [min]
Figure 50 Results of a step-gradient composition (07%) of a high-pressure pump (left) and of a low-pressure pump (right)
Performance of low-pressure pump design Flow precision < 0.3 % (typically < 0.15 %) based on retention times of 0.5 and 2.5 ml/min Flow range 0.29.999 ml/min Delay volume ca. 8001100 l Pressure pulse < 2 % amplitude (typically < 1 %), 1 ml/min propanol, at all pressures Composition 0.2 % SD precision at 0.2 and 1 ml/min
In this design, gradients are formed by a high-speed proportioning valve that can mix up to four solvents on the low-pressure side. The valve is synchronized with piston movement and mixes the solvents during the intake stroke of the pump. The solvents enter at the bottom of each chamber and flow up between the piston and the chamber wall, creating turbulences. Compared with conventional multisolvent pumps with fixed stroke volumes, pumps with variable stroke volumes generate highly precise gradients, even at low flow rates (see figure 50).
79
7
High-pressure gradient Agilent 1100 Series pump The Agilent 1100 Series high-pressure gradient pump is based on a double dual-piston mechanism in which two pumps are connected in series in one housing. This configuration takes up minimal bench space and enables very short internal and external capillary connections. Both pistons of both individual pumps are servocontrolled in order to meet chromatographic requirements in gradient formation (see figure 51). Three factors ensure gradients with high precision at low flow rates: a delay volume as low as 180480 l internal volume (without mixer), maximum composition stability and retention time precision, and a flow range typically beginning at 50 l/min. The same tracer gradient used to determine composition precision and accuracy also was used to determine the ripple of the binary pump (see figures 50 and 52). The delay volume was measured by running a tracer gradient. Large delay volumes reduce the sharpness of the gradient and therefore the selectivity of an analysis. They also increase the run-time cycle, especially at low flow rates.
Damper To sampling unit and column Purge valve Mixer
Performance of high-pressure pump design Flow precision < 0.3 % Flow range 0.055 ml/min Delay volume 180480 l (600900 l with mixer Pressure pulse < 2 % amplitude (typically, 1 %), 1 ml/min isopropanol, at all pressure > 1MPa Composition precision < 0.2 % at 0.1 and 1.0 ml/min
Outlet valve Inlet valve
Outlet valve Inlet valve
Figure 51 Schematics of the high-pressure gradient Agilent 1100 Series pump
80
When working at the lowest detection limits, it is important to use a mixer to reduce mixing noise, especially at 210220 nm and with mobile phases containing solvents such as tetrahydrofuran (THF). Peptide mapping on 1-mm columns places stringent demands on the pump because small changes in solvent composition can result in sizeable changes in retention times. Under gradient conditions at a flow rate of 50 l/min, the solvent delivery system must deliver precisely 1 l/min per channel. A smooth baseline and nondistorted gradient profiles depend on good mixing and a low delay volume. Figure 53 shows six repetitive runs of a tryptic digest of myoglobin with a retention time precision of 0.070.5% RSD.
mAU binary pump without mixer 380 l with mixer 850 l
300 200 100 0 3 4
at 5 min start of gradient
quaternary pump 950 l 5 6 Time [min] 7 8 9 10
Figure 52 Delay volume of high- and low-pressure gradient pumps
81
7
mAU 300 250 200 150 100 50 0 20 40 60 80 Time [min] 100 120 0.53% 0.15% 0.08% 0.06% 0.04% 0.02% 0.04% 0.04% 0.38%
0.07%
Figure 53 Overlay of six repetitive runs of a tryptic digest of myoglobin in RSD of RT is as low as 0.070.5 %
Degassing
Degassing removes dissolved gases from the mobile phase before they are pumped over the column. This process prevents the formation of bubbles in the flow path and eliminates volumetric displacement and gradient mixing, which can hinder performance. Instable flow causes retention on the column and may increase noise and drift on some flow-sensitive detectors. Most solvents can partially dissolve gases such as oxygen and thereby harm detectors. Detrimental effects include additional noise and drift in UV detectors, quenching effects in fluorescence detectors, and high background noise from the reduction of dissolved oxygen in electrochemical detectors used in reduction mode (in oxidation mode, the effect is less dramatic).
82
Fluorescence 12 10 8 6 4 2 10 11 1 2 3 4 12 13 14 6 Signal heights for selected PNAs 5
The oxygen effect is most apparent in the analysis of polycyclic aromatic hydrocarbons (PNAs) with fluorescence detection, as shown in figure 50. The less oxygen present in the mobile phase, the less quenching occurs and the more sensitive the analysis. In general, one of three degassing techniques is used: on- or offline vacuum degassing, offline ultrasonic degassing, or online helium degassing. Online degassing is preferable since no solvent preparation is required and the gas concentration is held at a constant, minimal level over a long period of time. Online helium and online vacuum degassing are the most popular methods.
Time [min] Agilent on-line degassing Helium degassing No degassing
Figure 54 The loss of response due to quenching can be recovered with either helium or vacuum degassing.
Helium degassing
In helium degassing, gas is constantly bubbled through the mobile-phase reservoir. This process saturates the solvent and forces other gases to pass into the headspace above.
Requires only a simple regulator. Several channels can be purged simultaneously without additional dead volume.
Expensive. Evaporation of the more volatile components can change composition over time. Oxygen is better purged by vacuum degassing.
83
7
Vacuum degassing In vacuum degassing, the solvent is passed through a membranous tube made of a special polymer that is permeable to gas but not to liquids under vacuum. The pressure differences between the inside and outside of the membrane cause continuous degassing of the solvent. New online degassers with low internal volume (< 1 ml) allow fast changeover of mobile phases.
Less expensive to use and maintain than helium degassing. The composition of premixed solvents is unaffected, and removal of oxygen is highly efficient. Several channels can be degassed simultaneously.
Increases dead volume and may result in ghost peaks, depending on the type of tubing and type of solvent used.
In brief
The choice of pump depends on both elution mode (isocratic or gradient) and column diameter (narrow bore or standard bore). Although an isocratic system often is sufficient, gradient systems are more flexible. Moreover, their short analysis times make gradient systems ideal for complex samples, sharp peaks, resolution of multiple species, and automatic system cleansing with additional online solvent channel. Agilent 1100 Series pumps are best suited for flow ranges from 0.05 ml/min up to 10 ml/min and can therefore be used with columns that have an inner diameter of 1 mm to 8 mm. Although many officially recognized methods are based on standard columns and flow rates, the trend is toward narrow-bore columns. These consume less solvent, which also reduces waste disposal, thus lowering operating costs.
84
Chapter 8
Detectors
8
Most detectors currently used in HPLC also can be applied in the analysis of food analytes. Each technique has its advantages and disadvantages. For example, diode array UV-absorbance detectors and mass spectrometers provide additional spectral confirmation, but this factor must be weighed against cost per analysis when deciding whether to use a detector routinely. The ability to use UV spectra to confirm the presence of certain food analytes and their metabolites and derivatives makes UV absorbance the most popular detection technique. However, for analytical problems requiring high sensitivity and selectivity, fluorescence detection is the method of choice. Although electrochemical detectors are also highly sensitive and selective, they are rarely used in food analysis. Conductivity detectors, on the other hand, are well-suited for the sensitive and selective analysis of cations and anions, and thermal energy detectors are used for high-sensitivity determination of nitrosamines down to 10 parts per trillion (ppt). Refractive index (RI) detectors are appropriate only if the above-mentioned detectors are not applicable or if the concentration of analytes is high, or both.
86
Analytical parameters
The most important parameters for food analysis are: limit of detection (LOD) and limit of quantification (LOQ) linearity selectivity qualitative information
Limit of detection and limit of quantification
The LOD and LOQ of an analytical system depend on the noise and drift of the detection equipment. Absolute detector LOD can be determined by injecting a sample directly into the detector. It is often expressed as minimum detectable level, which is sometimes defined as equal to the noise level. However, the LOD depends not only on the detector but may also depend on the oxygen content of the mobile phase, the injection system, peak broadening on the column, and temperature differences among system components. Taking these factors into account, the LOD is defined as 2 to 3 times the noise level. The LOQ is defined as 10 to 20 times the noise level. A UV detection system can be used to measure quantitative amounts down to 500 pg per injection. The LOD can be as low as 100 pg for food compounds such as antioxidants if detection wavelengths have been optimized to match the extinction coefficients of as many compounds as possible. Fluorescence and electrochemical detectors operate in the very low picogram range. The LOD of a mass spectrometer connected to HPLC equipment depends on the type of interface used. Instruments with electrospray interfaces can detect down to the picogram range. Refractive index detectors normally are appropriate above 500 ng. We define the selectivity of a detection system as the ability to select only those compounds of interest in a complex matrix using specific compound properties. A detector is selective if it does not respond to coeluting compounds that
Selectivity
87
8
could interfere with analyte quantification. A UV absorbance detector can be made selective by setting an appropriate wavelength with a narrow bandwidth for the compound of interest. However, the selectivity of detectors based on such a universal feature is low compared with the selectivity of detectors based on fluorescence and electrochemistry. Response characteristics are very selective, shown by a limited number of compounds. Mass spectrometers can be applied selectively or universally (in total scan mode), depending on the analysis to be performed. RI detectors are universal by definition. Linearity Detector response can be expressed both as dynamic range and as linear dynamic range. Dynamic range is the ratio of the maximum and the minimum concentration over which the measured property (absorbance, current, and so on) can be recorded. However, in practice, linear dynamic range the range of solute concentration over which detector response is linearis more commonly used. Plotting the response of injections of different analyte concentration against their concentrations should give a straight line over part of the concentration range. Response often is linear for only one tenth of the full dynamic range. UV detectors are linear over a range of a maximum of five orders of magnitude, whereas fluorescence and electrochemical detectors are linear over a range of two orders of magnitude. Mass spectrometers are usually linear over three orders of magnitude, and RI detectors are linear over a maximum of four orders of magnitude. A classical identification tool in chromatography is the mass spectrogram, which is recorded by a mass spectrometer. Its appeal in HPLC, however, is limited owing to the cost of interfacing the mass spectrometer equipment. If the spectra of the analytes differ markedly, UV absorbance spectra can be used for identification using diode array technology. Fluorescence and electrochemical detectors can be used only to identify compounds based on their retention times.
Qualitative information
88
UV detectors
Figure 55 shows the optical path of a conventional variable wavelength detector. Polychromatic light from a deuterium lamp is focused onto the entrance slit of a monochromator using spherical and planar mirrors. The monochromator selectively transmits a narrow band of light to the exit slit. The light beam from the exit slit passes through the flow cell and is partially absorbed by the solution in the flow cell. The absorbance of the sample is determined by measuring the intensity of the light reaching the photodiode without the sample (a blank reference) and comparing it with the intensity of light reaching the detector after passing through the sample.
Deuterium lamp
Cut-off filter Holmium oxide filter
Slit Lens Mirror 1 Sample diode Grating
Flow cell Beam splitter Mirror 2
Reference diode
Figure 55 Conventional variable wavelength detector
Most variable wavelength detectors split off part of the light to a second photodiode on the reference side. The reference beam and the reference photodiode are used to compensate for light fluctuations from the lamp. For optimum sensitivity,
89
8
the UV detector can be programmed for each peak within a chromatographic run, which changes the wavelength automatically. The variable wavelength detector is designed to record absorbance at a single point in the spectrum at any given point in time. However, in practice, different wavelengths often must be measured simultaneously, for example when two compounds cannot be separated chromatographically but have different absorbance maxima. If the entire spectrum of a compound is to be measured, the solvent flow must be stopped in order for a variable wavelength detector to scan the entire range, since scanning takes longer than elution.
Sensitive; can be tuned to the wavelength maxima of individual peaks. Some instruments are equipped with scanning mechanisms with stopped-flow operation.
Single-wavelength measurement is not always sufficient. Without spectra, peaks cannot be identified.
Diode array detectors
Figure 52 shows a schematic diagram of a photodiode array detector (DAD). An achromatic lens system focuses polyTungsten lamp Holium oxide filter Achromatic lens
950 nm 190 nm
Deuterium lamp
1024-element diode array Standard flow cell
Programmable slit
Figure 56 Diode array detector optics
90
0.6 0.7 mAU 0.4 0.2 0
chromatic light from the deuterium and tungsten lamps into the flow cell. The light then disperses on the surface of a diffraction grating and falls on the photodiode array. The range varies from instrument to instrument. The detector shown here is used to measure wavelengths from 190 to 950 nm using the twin-lamp design.
280 nm
240
260
Figure 57 High-resolution spectrum for benzene in the low absorbance range
Conventional Signal acquisition Spectra acquisition 1
DAD 8
stop flow
on-line
In our example, the array consists of 1024 diodes, each of which measures a different narrow-band spectrum. Measuring the variation in light intensity over the entire wavelength range yields an absorption spectrum. The bandwidth of light detected by a diode depends on the width of the entrance slit. In our example, this width can be programmed to selected values from 1 to 16 nm. If very high sensitivity is required, the slit is opened to 16 nm for maximum light throughput. If maximum spectral resolution is needed, the slit is narrowed to 1 nm. At this setting, the fine structure of benzene can be detected, even at 0.7 mAU full-scale (mAUFS; see figure 57). Because the relative positions of the sample and the diffraction grating are reversed compared with a conventional instrument, this configuration is often referred to as reversed optics. The most significant differences between a conventional UV absorbance detector and a DAD are listed at left. DADs connected to appropriate data evaluation units help optimize wavelengths for different compounds over the course of the run. Maxima can be seen easily using three-dimensional plots of data, or as absorbance intensity plotted over time at different wavelengths, that is, as an isoabsorbance plot (see figure 58). Figure 55 illustrates the optimization result for antibiotics. The ability to acquire and store spectra permits the creation of electronic spectral libraries, which can be used to identify sample compounds during method development.
Three dimensions of data
91
Figure 58 Isoabsorbance plot
Multisignal detection yields optimum sensitivity over a wide spectral range. However, the spectral axis in figure 58 shows that no single wavelength can detect all antibiotics at highest sensitivity.
mAU 9
100
10
100
meticlorpindolmetronidazol nicarbazine
5 80 60 1
40
Absorbance (scaled)
2 4
6,7
260 300 340 380 Wavelength [nm]
275 nm
11
315 nm 20 360 nm
Figure 59 Multisignal detection of antibiotic drugs
1 2 3 4 5 6 7 8 9 10 11 40
Metronidazol Meticlorpinol Sulfapyridine Furazolidon Pyrazon Ipronidazol Chloramphenicol N-Acetylsufapyridine Ethopabat Benzothiazuron Nicarbazin
10
20
Time [min] 30
92
Figure 60 Peak purity analysis
In light of the complexity of most food samples, the ability to check peak purity can reduce quantification errors. In the most popular form of peak purity analysis, several spectra acquired during peak elution are compared. Normalized and overlaid, these spectra can be evaluated with the naked eye, or the computer can produce a comparison. Figure 60 shows a peak purity analysis of antibiotics. If a spectral library has been established during method development, it can be used to confirm peak identity. Analyte spectra can be compared with those stored in the library, either interactively or automatically, after each run.
93
8
Figure 61 shows both the quantitative and qualitative results of the analysis. Part one of this primer contains several applications of UV absorbance DAD detection.
18 14 3
9 8 10
Peak Purity Check and Identification
Part 1: General information
10 6
2
1
2
7 4
6
11
* * * * *
Operator Name: 0/ 1 (s0B Date & Time: Data File Name: Integration File Name: Calibration File Name: Quantitation method: Sample Info: Misc. Info: BERWANGER
R E P O R T
* * * * *
(s1B Vial/Inj.No.:
10 Match > 950

1 2 3 4 5 6 ?-*Metronidazole ?-*Meticlorpindol Sulfapyridine Furazolidone Pyrazon ?-*Ipronidazole
20
30
10 Sep 86 9:17 am LH:LETAA00A DATA:DEFAULT.I DATA:ANTI.Q ESTD DOTIERUNGSVERSUCHE
calibrated by
Area response
7 Chloramphenicol 8 N-Acetylsulfapyridine 9 Ethopabate 10 Benzothiazuron 11 *Nicarbazin
Method File Name: Library File Name: Reference Spectrum: Time window from: Dilution Factor: 1.0
ANTIBI.M DATA:ANTIBI.L Apex 6.0 % to: 2.0 % Sample Amount: 0.0
Wavelength from: 230 to: 400 nm Library Threshhold: 950 Peak Purity Threshold: 950 Smooth Factor: 7 Resp.Fact.uncal.peaks: None
Part 2: Quantitation, peak purity check and peak identification

Name Amount [ng/l] Peak-Ret. Cal.-Ret. Lib.-Ret Purity Library Res. [min] [min] [min] Matchfactor
________________________________________________________________________________
Sulfapyridine Furazolidone Pyrazon 10.31 4.54 13.72 A 12.183 A 16.096 A 19.024 12.143 16.024 18.987 12.159 16.028 19.000 999 992 1000 1000 984 1000 0.9 1.3 1.7
N-Acetylsulfapyidine
14.66 13.40 12.80
A 23.307 A 23.874 A 24.047

A 32.733
23.282 23.840 24.024

32.722
23.282 23.848 24.029

32.709
976 911 998

336
1000 996 1000

984
1.1 2.3 0.7

1.2
Figure 61 Quantitative and qualitative results for the analysis of antibiotic drugs
Ethopabat Benzthiazuron
Nicarbazin
*up
*up
3.00
========
72.41
Enables maximum peak purity and identity, measurement of multiple wavelengths, acquisition of absorbance spectra, and spectral library searches.
DADs are best suited for universal rather than sensitive analysis (for which electrochemical or fluorescence detection is more appropriate).
94
Fluorescence detectors
Xenon flash lamp Lens
Emission monochromator
Fluorescence is a specific type of luminescence that is created when certain molecules emit energy previously absorbed during a period of illumination. Luminescence detectors have higher selectivity than, for example, UV detectors because not all molecules that absorb light also emit it. Fluorescence detectors are more sensitive than absorbance detectors owing to lower background noise. Most fluorescence detectors are configured such that fluorescent light is recorded at an angle (often at a right angle) to the incident light beam. This geometry reduces the likelihood that stray incident light will interfere as a background signal and ensures maximum S/N for sensitive detection levels.
Photomultiplier
Mirror
Lens
Excitation monochromator Sample
Photodiode
Figure 62 Schematics of a fluorescence detector
The new optical design of the Agilent 1100 Series fluorescence detector is illustrated in figure 62. A Xenon flash lamp is used to offer the highest light intensities for excitation in the UV range. The flash lamp ignites only for microseconds to provide light energy. Each flash causes fluorescence in the flow cell and generates an individual data point for the chromatogram. Since the lamp is not powered on during most of the detector operating time, it offers a lifetime of several thousand hours. No warmup time is needed to get a stable baseline. A holographic grating is used as a monochromator to disperse the polychromatic light of the Xenon lamp. The desired wavelength is then focused on the flow cell for optimum excitation. To minimize stray light from the excitation side of the detector, the optics are configured such that the emitted light is recorded at a 90 degree angle to the incident light beam. Another holographic grating is used as the emission monochromator. Both monochromators have optimized light throughput in the visible range. A photomultiplier tube is the optimum choice to measure the low light intensity of the emitted fluorescence light. Since flash lamps have inherent fluctuations with respect to flash-to-flash intensity, a reference system based on a
95
8
photodiode measures the intensity of the excitation and triggers a compensation of the detector signal. Cut-off filter Since the vast majority of emission maxima are above 280 nm, a cut-off filter (not shown) prevents stray light below this wavelength to enter the light path to the emission monochromator. The fixed cut-off filter and bandwidth (20 nm) avoid the hardware checks and documentation work that is involved with an instrument that has exchangeable filters and slits. The excitation and emission monochromators can switch between signal and spectral mode. In signal mode they are moved to specific positions that encode for the desired wavelengths, as with a traditional detector. This mode offers the lowest limits of detection since all data points are generated at a single excitation and emission wave length. A scan of both the excitation and the emission spectra can be helpful in method development. However, only detectors with motor-driven gratings on both sides can perform such a scan. Some of these detectors also can transfer this data to a data evaluation computer and store spectra in data files. Once the optimum excitation and emission wavelength has been determined using scanned spectra, detectors with grating optics can be programmed to switch between these wavelengths during the run. The spectral mode is used to obtain multi-signal or spectral information. The ignition of the flash lamp is synchronized with the rotation of the gratings, either the excitation or emission monochromator. The motor technology for the gratings is a long-life design as commonly used in high-speed PC disk drive hardware. Whenever the grating has reached the correct position during a revolution, the Xenon lamp is ignited to send a flash. The flash duration is below two microseconds while the revolution of the grating takes less than 14 milliseconds. Because of the rotating monochromators, the loss in sensitivity in the spectral
Signal/spectral mode
Online spectral measurements and multisignal acquisition
96
mode is much lower compared to conventional dual-wavelength detection with UV detectors. Multisignal PNA analysis, for example, can be performed with simultaneous multi wavelength detection instead of wavelengthswitching. With four different wavelengths for emission, all 15 PNAs can be monitored (figure 63).
1 2 3 4 5 6 7 5 6 Naphthalene Acenaphthene Fluorene Phenanthrene Anthracene Fluoranthene Pyrene 8 9 10 11 12 13 14 15 10 11 12 2 7 3 8 9 4 13 14 5 15 Benz(a)anthracene Chrysene Benzo(b)fluoranthene Benzo(k)fluoranthene Benz(a)pyrene Dibenzo(a,h)anthracene Benzo(g,h,i)perylene Indeno(1,2,3-cd)pyrene
1 excitation WL at 260 nm 4 emission WL at 350, 420, 440 and 500 nm
LU 180 160 140 120 100 80 60 40 20 0

Ex=260, Em=500 Ex=260, Em=440 Ex=260, Em=420 Ex=260, Em=350 1 2 3 1 4 Ex=275, Em=350, TT Reference chromatogram with switching events
10
15
20
25 Time [min]
Figure 63 Simultaneous multi wavelength detection for PNA-analysis The upper trace was received with traditional wavelength switching. 1 Ex/Em = 260/420 nm 2 Ex/Em = 270/440 nm 3 Ex/Em = 260/420 nm 4 Ex/Em = 290/430 nm 5 Ex/Em = 250/550 nm
97
Highly specific. Flash lamps eliminate drawback of baseline drift from heat transfer. Fluorescent tagging improves detection limits.
Fluorescence spectra are not commonly used to confirm peak identity.
Electrochemical detectors
Reference Reference electrode electrode Counter electrode Working Working electrode electrode Cell Cell
Electrochemical detection techniques are based on the electrical charge transfer that occurs when electrons are given up by a molecule during oxidation or absorbed by a molecule during reduction. This oxidation or reduction takes place on the surface of a so-called working electrode. Whether a compound is reduced or oxidized and the speed of the reaction depend on the potential difference between the working electrode and the solution containing the compounds. From the activation energies and redox potentials expressed by the Nernst equation, reaction speed can be determined. The resulting current is proportional to the number of reactions occurring at the electrode, which in turn is an indicator of the concentration of the compound of interest at the surface. In the detection process, three electrodes are used: the working electrode, in which the reaction takes place; the counter electrode, which applies the potential difference between mobile phase and the working electrode; and the reference electrode, which compensates for any change in eluant conductivity (see figure 64). The reference electrode readings feed back to the counter electrode in order to keep the potential difference constant during peak elution as current flows through the working electrode.
+ +
Figure 64 Three-electrode electrochemical detector
98
Current
E2
E1/2 Optimum potential E1 0.4 0.8 0.6 Potential [V] 1.0
Figure 65 Current-voltage relationship
Detector response results from amplification of the electron flow and its subsequent conversion to a signal. Extremely low currents representing analyte quantities in the picogram range and below can be measured with todays advanced electronics. Although electrochemical detection can detect only those substances that can be electrolyzed, this limitation is actually an advantage when applied to complicated food matrices because it improves selectivity. To determine the optimum working electrode potential, the relationship between detector response (current) and potential applied (voltage) must be plotted for each compound as a current-voltage (CV) curve, as shown in figure 65. At a potential less than E1, oxidation cannot occur because the supply of energy is insufficient. Increasing the potential to E1/2 will electrolyze 50 % of all molecules at the surface of the electrode. Maximum response requires a potential just above E2. This potential is known as the limiting current because any further increase in voltage will limit detection by raising noise. Several materials are used in working electrodes, the most common of which is glassy carbon. These materials also include gold (for sugars and alcohols), platinum (for chlorite, sulfite, hydrazine, and hydrogen peroxide), silver (for halogens), copper (for amino acids), mercury (in reductive mode for thiosulfate), and combined mercury-gold (in reductive mode for nitrogenous organic compounds). Numerous cell designs have been described in the literature. The majority can be classified as one of three principal types: thin-layer design, wall-jet design, and porous flow-through design (see figure 66). The porous flow-through cell design differs significantly from the other two in that coulometric detection ensures 100 % reaction yield on the surface of the electrode. The other designs allow an efficiency of only 110 % by amperometric detection. However, amperometric detection is usually the more sensitive technique and is preferred over coulometric
Electrode materials
Flow cell aspects
99
8
detection. electrochemical detectors can employ 1-l flow cells and are well-suited to narrow-bore HPLC.
Thin layer
Reference electrode Auxilliary electrode
Wall jet
Reference electrode Auxilliary electrode
Porous flow-through
Reference Auxilliary electrode electrode
Working electrode Working electrode Working electrode
Figure 66 Thin-layer design, wall-jet design, and porous flow-through design
Automation features
oxidative cleaning (+1.3 V) Potential [V] working potential (1.2 V) reductive cleaning (-0.1 V)
Figure 67 Cleaning of working electrode
Until recently, the electrochemical technique was considered difficult to apply and not stable enough for routine analysis. However, recent improvements have made the use of these detectors routine, for example in the analysis of catecholamines in clinical research and routine testing laboratories. When applied between runs or even during peak elution (for example in sugar analysis using gold electrodes), self-cleaning routines based on pulsed amperometry improve stability (see figure 67). Although an optimum potential for a mixture of compounds can be determined by evaluating the voltamograms for each compound, these optimizing steps can be automated using certain electrochemical detectors in so-called auto-increment mode. The HPLC equipment runs a series of injections over a range of increasing potentials (defined by start and end potentials and increment parameter), as shown in figure 68. A drift sensor helps ensure that a specified threshold is maintained before the next analysis begins (see figure 69).
3-Nitrophenol
p-Chloro m-cresol
1.4 1.3 V 1.2 1.1 1.0 0.9 0.8 4 6
8 Time [ms]
10
12
Figure 68 Autoincrement mode
100
Current
Falling current detector not ready
Current steady detector triggers next injection Threshold set by drift trigger parameter
Should the electrode surface of the flow cell become severely contaminated, as is likely for food matrixes, the cell must be disassembled and the electrode removed and cleaned in a strong acid or other suitable cleaning agent. Modern detectors are designed for ease of access and disassembly. Part one of this primer contains several applications of electrochemical detection.
Time [min]
Baseline
Figure 69 Drift trigger
Mass spectrometers
Electrospray ThermoParticle spray beam GC/MS Polarity/solubility in water
Figure 70 Suitability of MS interfaces
The identification of complex samples presents a problem for LC analysis. Coeluting compounds generally can be identified using UV absorbance detection with diode array technology, but this method may not be specific enough where spectra differences are low. Detection techniques such as fluorescence may offer higher specificity than UV detection, but if many different compounds are to be analyzed, these techniques also may not yield desired results. With mass spectroscopy (MS), several different analyte classes in a wide variety of sample types can be identified with greater certainty. Although GC/MS is a well-established technique for food analysis, LC/MS is only now emerging as a useful tool in this area. A GC-based analysis is appropriate only for those food compounds that are volatile and thermally stable (see figure 70). However, many compounds are nonvolatile, extremely polar, or thermally labile. Such compounds often can be separated successfully with LC, and the development of improved interfaces has made LC/MS more popular.
Molecular weight
101
8
HPLC High pressure liquid phase separation Produces large quantities of volatilized solvent (1001000 ml/min gas)* No mass range limitation MS High vacuum required Typical MS vacuum systems designed for low ml/min gas load Depends on masss/charge and mass range of analyzer Prefers volatile buffers
Can use inorganic buffers
* About 1000-fold increase going from liquid to gas phase with typical LC solvent
Mass spectrometers nevertheless are more easily interfaced with GC equipment than with LC equipment. The table at left lists the different operational conditions of LC and MS. Early efforts to interface LC with MS used direct liquid injection and moving-belt interfaces, but these methods proved ineffective and unreliable. In the 1980s, thermospray and particle beam interfaces improved both the range of applicability and the reliability of LC/MS. However, low sensitivity, the narrow mass and polarity range of analytes, and frequent maintenance requirements limited the effectiveness of these interfaces. More recently, two atmospheric pressure ionization (API) interfaceselectrospray and atmospheric pressure chemical ionization (APCI)have replaced almost completely thermospray and particle beam techniques. These interfaces have a broad range of analyte molecular weights and polarities, high sensitivity, improved usability, and reduced maintenance needs. Selection of the appropriate LC/MS interface for an application depends on factors such as the polarity, molecular weight, and thermal lability of the analyte. In electrospray, effluent is directed through a nebulizing needle into a high-voltage field where charged droplets are formed (see figure 71). The charged droplets are then dried
HPLC inlet Nebulizer Skimmers Octopole
API interfaces
Capillary
+ +++ + + + + + + + + + +
Fragmentation zone (CID)
Lenses
Quadrupole
Figure 71 API-electrospray LC/MS interface
102
Abundance
8 1 2 3 4 5 6 7 8
4 5 3 12
Aldicarb sulfoxide Aldicarb sulfone Methomyl 3-hydroxycarbofuran Aldicarb Carbofuran Carbaryl Methiocarb
and, as they shrink, analyte ions are desorbed. The ions are transported to the mass analyzer through a series of vacuum stages and ion-focusing elements. Electrospray ionization can produce multiply charged ions of macromolecular analytes such as proteins and peptides. Because mass analyzers separate ions based on mass-tocharge ratio (m/z), lower-cost mass spectrometers with mass ranges of several thousand m/z can be used to analyze compounds in excess of 150,000 daltons. The primary use of electrospray has been the analysis of compounds of higher molecular weight. However, this technique also has been applied successfully to small polar molecules. Fig. 72 shows a separation of carbamate pesticides using electrospray.
HPLC inlet Nebulizer Skimmers Capillary
+ + + + +
10 Time [min] 20
Figure 72 Carbamate analysis
Octopole
Corona needle Fragmentation zone (CID) Lenses Quadrupole
Figure 73 APCI LC/MS interface
APCI also can be used to analyze moderate polarity analytes. As in electrospray, APCI ionization occurs at atmospheric pressure via a chemical ionization process (see figure73).
103
8
Abundance
100000 80000 60000 40000 20000 m/z
<- [M + NH ] +
4
603 400
639 800
987
1000
600
Figure 74 Mass spectrum of the fatty acid triolein (C18:1, [cis]-9) molecular weight = 884.781 molecular formula = C57H104O6
APCI requires some compound volatility and is less suitable for highly thermally labile compounds. Figure 74 shows a typical triglyceride mass spectrum. Both the degree of unsaturation and the length of the fatty acid side chains can be determined from the [M + NH4] + ion, which corresponds to mass M + 18. In-source CID experiments also can be helpful in determining the fatty acid composition of chromatographic peaks. Full-scan methods allow easy identification at the low nanogram level. If more precise quantitation is required, selected ion mode (SIM) can be used to obtain detection limits at the low picogram level.
Polar and semipolar compounds up to 150,000 daltons can be analyzed. Highly sensitive. Strong molecular ions. Fragments, depending on in-source CID parameters.
Data analysis for complex heterogeneous mixtures of multiply charged analytes is not straightforward. Matrix can interfere with the ionization process.
Refractive index detectors
Refractive index (RI) detection is based on the difference in RI between the solution in the sample cell and the pure mobile-phase solution in the reference cell. Because the composition of the eluents must remain fixed throughout the analysis, this detector is not suitable for gradient analysis. Four main types of RI detectors are available: deflection according to Snells law, reflection according to Fresnels law, interference, and Christiansen effect. The first, which uses the dual-cell design, is by far the most popular. However, the nearly designed Agilent 1100 Series refractive index detector allows detection limits to the low ng range. Because RI detectors lack sensitivity and exhibit a tendency to drift owing to temperature changes, they are used primarily in the analysis of carbohydrates and nonaromatic acids.
Universal detector.
Low sensitivity, no gradient operation.
104
In brief
The following table reviews the detection techniques discussed in this chapteryour decision ideally should reflect a balance between desired results and financial resources.
Detector UV variable wavelength UV-DAD Sensitivity + Selectivity Advantages Low cost, universal acids Peak purity confirmation Applications Organic acids, fatty after derivatization, inorganic anions Antioxidants, preservatives, flavors, colorants, antiparasitic drugs, mycotoxins, pesticides, vitamins, amines after derivatization Artificial sweeteners, mycotoxins, vitamins, carbamates, glyphosate Vitamins, inorganic anions Carbamates, lipids Pesticides, proteins Carbohydrates, nonaromatic acids
Fluorescence
++
High sensitivity
Electrochemical Mass spectrometer scan Mass spectrometer SIM RI
++ ++ -
+ ++ ++ -
High sensitivity Identity, structure High selectivity Universal
105
106
Chapter 9
Derivatization chemistries
9
When analyte concentrations are particularly low, sample handling equipment for chemical derivatization can enhance the sensitivity and selectivity of results. As discussed in chapter 6, such equipment is available both pre- and postcolumn. In this chapter, we detail the chemistries that can be applied to food compounds and list the detection techniques for which they are best suited.
Addition of UV-visible chromophores
Labeling compounds with reagents that enable UV absorption is one of the most popular derivatization techniques. The reagent should be selected such that the absorption maximum of the reaction product exhibits not only improved sensitivity but also good selectivity. This combination reduces matrix effects resulting from the reagent, from by-products, or from the original matrix. The following table lists common compounds and reactions. In part one of this primer we give examples of compound derivatization, including that of fatty acids and amino acids.
Target compound Alcohols Oxidizable sulfur compounds Fatty acids Aldehydes and ketones Primary amines -OH SO32-COOH -CO-COOH, =C=O, and -CHO -NH2 Reagent phenylisocyanate 2,2-dithiobis (5-nitro-pyridine) p-bromophenacyl bromide 2-naphthacyl bromide 2,4-dinitrophenyl hydrazine -phthalaldehyde (OPA) 9-fluorenylmethyl chloroformate (FMOC) 250 nm 320 nm 258 nm 250 nm 365 nm 340 nm 256 nm
Primary and NHR secondary amines
108
Addition of a fluorescent tag
Fluorescence is a highly sensitive and selective detection technique. Adding fluorescent properties to the molecule of interest is of particular benefit in food analysis, in which components must be detected at very low concentrations. The following table lists common fluorescent tags. In part one of this primer we give examples for carbamates41 and glyphosate.42
Target compound Alcohols Primary amines -OH -NH2 Tagging reagent phenylisocyanate o-phthalaldehyde (OPA) 9-fluorenylmethyl, chloroformate (FMOC) ex 230 nm, em 315 nm ex 230 nm, l em 455 nm
Primary and NHR secondary amines
ex 230 nm, l em 315 nm
Precolumn or postcolumn?
Pickering system
Precolumn techniques can be run either offline or online, but postcolumn techniques should be run online for maximum accuracy. In postcolumn derivatization, reagents can be added only through supplementary equipment (see figure 75) such as pumps. Mixing and heating devices also may be required. Increasing the dead volume behind the column in this way will result in peak broadening. Although this broadening may have no effect on standard-bore columns with flow rates above 1 ml/min, postcolumn derivatization is not suitable for narrow-bore HPLC.
Figure 75 Pickering postcolumn derivatization equipment for the analysis of carbamates
109
9
Automatic derivatization Both pre- and postcolumn derivatization techniques can be automated with modern HPLC equipment. The single-step mechanical functions of an autoinjector or autosampler can be programmed prior to analysis and stored in an injector program (see left). These functions include aspiration of the sample and of the derivatization agent, and mixing. Precolumn derivatization is fully compatible with narrow-bore HPLC and can result in fivefold improvements in S/N, with much lower solvent consumption than that from standard-bore methods. The analysis of fatty acids in part one of this primer illustrates this principle.
1 2 3 4 5 6 7 8 9
Draw Draw Draw Draw Draw Draw Mix Draw Inject
1.0 l 0 l 1.0 l 0 l 1.0 l 0 l 8 cycles 1.0 l
from vial 12 from vial 0 from vial 8 from vial 0 from sample from vial 0 from vial 12
Derivatization improves detectability of trace species. It can be automated and integrated online within the analysis. Many chemistries have been developed for routine use both pre- and postcolumn.
Additional investment in equipment.
In brief
Derivatization offers enhanced analytical response, which is of benefit in food analysis. Chemical modifications can be automated either before or after separation of the compounds under study. In precolumn derivatization, autoinjectors with sample pretreatment capabilities (see chapter 6) are used, whereas in postcolumn derivatization, additional reagent pumps are plumbed to the chromatograph upstream of the detector. The latter approach adds dead volume and therefore is not suitable for the narrow-bore column technique described in chapter 4.
110
Chapter 10
Data collection and evaluation techniques
10
Regardless which detection system you choose for your laboratory, the analytical data generated by the instrument must be evaluated. Various computing equipment is available for this task. The costs depend on the reporting requirements and on the degree of automation required.
Depending on individual requirements, increasingly complex techniques are available to evaluate chromatographic data: at the simplest level are strip chart recorders, followed by integrators, personal computerbased software packages and, finally, the more advanced networked data systems, commonly referred to as NDS. Although official methods published by the U.S. Environmental Protection Agency (EPA) and by Germanys Deutsche Industrienorm (DIN) provide detailed information about calculation procedures and results, they give no recommendations for equipment.
Strip chart recorders
Strip chart recorders traditionally have been used in connection with instruments that record values over a period of time. The recorder traces the measurement response on scaled paper to yield a rudimentary result. In the age of electronic data transfer, such physical records have been largely surpassed by data handling equipment preprogrammed to make decisions, for example to reject peaks that lie outside a certain time window.
Inexpensive.
No record of retention times, no quantitative results on-line, no automatic baseline reset between runs, no electronic storage.
112
Integrators
Integrators offer several advantages over strip chart recorders and consequently are becoming the minimum standard for data evaluation. Integrators provide a full-scale chromatographic plot and multiple report formats. Area percent, normalization, and external and internal standard calculations are basic features of almost all modern integrators. Annotated reports list amounts, retention times, calculation type (peak areas or heights), and integration parameters as well as the date and time of measurement. Advanced features may provide for automated drawing of the baselines during postrun replotting and for the plotting of calibration curves showing detector response. For unattended analyses in which several runs are performed in series, integrators normally are equipped with a remote control connected to the autosampler in the system. Most models can also store raw data for replotting or reintegration at a later date. Some instruments have computer programming capabilities and can perform more advanced customized statistical calculations using the BASIC programming language, for example. Multichannel integrators are available for some analytical methods requiring two or more detection signals.
Inexpensive. Facilitates reporting of retention times, quantitative results, and automatic baseline resets.
No instrument control or report customization.
113
10
Personal computers
In recent years personal computers (PCs) have become increasingly popular as data analysis tools in analytical laboratories. PCs offer more flexibility and better data storage capabilities than traditional storage methods. Moreover, on-line functions such as word processing, spreadsheet analyses, and database operations can be performed simultaneously (see figure 76). Through computer networks, laboratory instruments can be interconnected to enable the central archival of data and the sharing of printer resources. Client/server-based software extends these capabilities by distributing the processing across multiple processing units and by minimizing the time spent validating software. With PCs, all aspects of the HPLC system can be accessed using a single keyboard and mouse. Parameters for all modules, including pump, detector, and autosampler, can be entered in the software program, saved to disk, and printed for documentation. Some HPLC software programs include diagnostic test procedures, instrument calibration procedures, and extensive instrument logbooks, all of which can facilitate the validation processes of various regulatory agencies. Such complementary functions, although not
Figure 76 Cross sample reports regression analyses, trend charts and other calculations consolidate sample data, enhancing the overall productivity and efficiency of the laboratory
114
directly related to the control of the equipment, are more easily built into a software program than into the equipment itself. In fact, many GLP/GMP features are added to every new version of the software programs sold with HPLC equipment (see figure 77). For example, in some chromatography software, the raw data files can store more than just signal data. A binary check-sum protected file stores instrument parameters (system pressure, temperature, flow, and solvent percent) as well as all aspects of the analytical method, including integration events, calibration settings, and a date-stamped logbook of events as they occurred during the run. Additionally, with spectral libraries, compounds can be identified not only on the basis of their elution profile but also according to their spectral characteristics. Such procedures can be fully automated to reduce analysis time and user interaction.
Figure 77 Maintenance and diagnosis screen
115
10
A single PC running the appropriate chromatography software can process data from several detectors simultaneously. This feature is particularly useful in analyses in which sensitivity and selectivity must be optimized to different matrices and concentrations. For example, in the analysis of polynuclear aromatic hydrocarbons, UV absorbance and fluorescence detection are applied in series. The PC displays graphically the chromatographic signals and spectra, enabling detailed interpretation of the data. Software purity algorithms can be used to help determine peak homo-geneity, even for coeluting peaks. Flexible software programs can report data in both standard and customized formats. For example, some chromatography software can be programmed to yield results on peak purity and identification by spectra or, for more complex analyses, to generate system suitability reports. Any computer-generated report can be printed or stored electronically for inclusion in other documents. PCs are well-suited for the modification of calibration tables and for the reanalysis of integration events and data. The software must record such recalculation procedures so that the analysis can be traced to a particular set of parameters in accordance with GLP/GMP principles. A computer can automate entire sequences of unattended analyses in which chromatographic conditions differ from run to run. Steps to shut down the HPLC equipment also can be programmed if the software includes features for turning off the pump, thermostatted column compartment, and detector lamp after completion of the sequence. If the HPLC equipment malfunctions, the software reacts to protect the instrumentation, prevent loss of solvents, and avoid unnecessary lamp illumination time. A good software application should be able to turn off the pump, thermostatted column compartment, and detector lamp in the event of a leak or a faulty injection. System suitability tests also can be incorporated in a sequence. When performed on a regular
116
basis, such tests can validate assumptions about performance of the analytical system and help verify results.
Enables control of multiple instruments. Additional software can be used for many other tasks. Provides for better data storage and archival.
Requires more bench space for peripherals such as printers or plotters.
Local area networks
Shared printing peripherals
A laboratory running food analyses frequently requires multiple instruments from multiple instrument vendors for sample analysis. Although the integrators and PC systems described above can evaluate data at analytical instrument stations throughout the laboratory, this data must be collected centrallyover a network, for examplein order to generate a single report for multiple analytical techniques. Local area networks (LANs) offer several advantages in addition to shared data processing (see figure 78). Centralized printing saves bench space and reduces equipment expenses, and centralized file security through a single computerthe serveraccelerates data backup. Standard network software and hardware cannot handle data files from diverse analytical instrument vendors. The analytical software therefore should have file conversion utilities based on the Analytical Instrument Association ANDI file format (*.cdf).
Figure 78 A laboratory LAN connecting instruments and collating analytical results
Integrates multiple techniques and instruments from multiple vendors. Saves bench space and computer processing resources. Access to network utilities such as e-mail.
Data processing features may not match those of dedicated data analysis software applications.
117
10
Networked data systems
The Agilent ChemStation remote access and data storage modules combine isolated islands of data into a powerful client/server networked information system. Each Agilent ChemStation becomes a network client. It is possible to oversee and control all laboratory operations securely and easily from any computer on the network. The progress of each analysis is monitored to ensure the quality of the results the first time the sample is analyzed. Appropriate action can be taken with the access remote capability from wherever you happen to be if the performance looks suspect. Laboratory data is automatically stored on one centralized and secure server system.
In brief
Which data handling technique is most effective and economical for your laboratory depends on several factors: the size of the laboratory the role of the laboratory in the organization industrial testing, public safety testing, and so on the demands on sample throughput the range of analytes under study For laboratories with few instruments and low sample throughput, integrator systems normally suffice, although a PC may be more appropriate for automated operation of multiple HPLC instruments. A client/server networked data system helps consolidate documentation and validation processes for multiple techniques and instruments from multiple vendors.
118
Chapter 11
Factors that determine performance in HPLC
11
The analysis of food samples places high demands on HPLC equipment, notably in the areas of performance, stability, and reliability. Modern evaluation software enables you to determine the suitability of a particular piece of HPLC equipment for analysis. The factors that influence the outcome of a measurement thus can be identified before results are published to confirm assumptions made during analysis or to draw attention to erroneous data.
In this chapter we focus on those instrument-related parameters that strongly influence the limit of detection (LOD) and the limit of quantification (LOQ). We also discuss the accuracy, precision, and qualitative information that an HPLC system can provide. Some vendors address the performance of specific instrumentation in technical notes.43 Such notes include detailed performance test procedures and results for individual modules as well as for complete HPLC systems.
120
Limit of detection and limit of quantification
The principle determinant of the LOD in HPLC is the response of the detector to the compound of interest. The response factor thus depends primarily on the choice of detection technique. However, regardless of the quality of the detector, the LOD or LOQ remains a function of peak height. This height can sink if the peak is allowed to disperse within the surrounding liquid in the flow path. All parts of the flow path in front of the detector therefore must be designed to limit broadening and flattening of the response. A minimum of narrow capillaries between injector and column and from column to detector helps keep dead volume low. With low injection volumes, separation efficiency of the column can be utilized to the maximum, thereby improving peak height. In other words, the lower the column volume, the lower the peak volume eluted. Other factors that influence peak dispersion include pump performance, degassing efficiency, capacity factor (k), and column particle size. Any improvements can be registered by calculating the S/N of the analyte. Indeed, the noise of the detector should be tested regularly in this way to ensure that performance is maintained. Dead volume of the complete injection system can be determined by first injecting a tracer mobile-phase additive into the flow path with the column disconnected and then recording the time this additive takes to reach the detector at a particular flow rate. The flow cell volume of the detector should be as low as possible, whereas its pathlength should be as long as possible, according to Beers law. Maximizing analyte response is not sufficient to ensure good results, however, since the level of background noise from the detector can counter any gains made. In particular, the performance of the pump in combination with certain solvents can increase detector noise level, as described in chapter 7. Degassing is necessary in order to avoid gas
121
11
bubbles, which can cause noise or spikes, or oxygen quenching in fluorescence. High k values result from higher elution volume or from longer retention time. These values are accompanied by broader peak width and smaller peak height, that is, peaks with longer retention times have poorer S/N. The use of different columns, different mobile phases, and different flow rates can improve S/N. Packing material also directly influences peak dispersion; for example, smaller-sized particles reduce peak dispersion.
Accuracy and precision
Accuracy is the degree of agreement between test results and true values. It is influenced by the analytical method, the extraction procedure used, and the choice of column or detector. Prior to the adoption of any HPLC method for routine use, the degree of agreement with an established reference method should be determined, or a control run should be performed with a known quantity of spiked sample matrix. In practice, however, the degree of agreement will never reach 100 %. This mismatch can be corrected by calibration with standards of known concentration and, based on these results, by calculating the accurate results from an unknown sample. Inclusion of an external or internal standard calibration procedure ensures accuracy in food analysis. The precision of a method is the degree of agreement among individual test results when an analysis is applied repeatedly to multiple samplings. Precision is measured by injecting a series of standards and then calculating the relative standard deviation of retention times and areas or peak heights. Precision may be measured at three levels: repeatability, intermediate precision, and reproducibility. Repeatability is associated with an analysis performed in one laboratory by one operator using a single piece of equipment over a relatively short time period. Intermediate precision is
122
the long-term variability of the measurement process for a method performed within one laboratory but on different days. Reproducibility applies to an analysis performed in more than one laboratory. Any HPLC method used in food analysis should be tested for both repeatability and reproducibility. The precision of a method is strongly influenced by the performance of the HPLC instrumentation. Repeatability of flow rates, gradient formation, and injection volumes can affect precision, as can response stability of the detector, aging of the column, and temperature stability of the column oven. The equipment should be inspected on a regular basis using the test methods recommended by the supplier to ensure reliability, high performance, and good analytical results.
Qualitative information
HPLC analytes can be identified on the basis of their retention times and either their UV-visible or mass spectra. Compounds, on the other hand, are identified primarily according to the degree of agreement between retention times recorded using calibration standards and those obtained from the sample. Unfortunately, co-eluting peaks can falsify results obtained with samples containing unknowns, especially for food matrices such as meat, vegetables, or beverages. In such cases, samples often can be identified using UV-visible spectral information. A diode array detection system enables online acquisition, and a number of software packages offer automatic evaluation, for example for the analysis of polynuclear aromatic hydrocarbons (PNAs) and pesticides.41
123
References and Index
Part Three
References
01. D.N. Heiger, High Performance Capillary ElectrophesisAn Introduction, Agilent Primer 5968-9936E, 2000. 02. CD-ROM CE Partner, Agilent publication 5968-9893E 03. CD-ROM CE Guidebook, Agilent publication 5968-9892E 04. Official Methods of Analysis, Food Compositions; Additives, Natural Contaminants, 15th ed; AOAC: Arlington, VA, 1990, Vol. 2. 05. A.M. Di Pietra, et al., HPLC analysis of aspartame and saccharin in pharmaceutical and dietary formulations, Chromatographia, 1990, 30, 215219. 06. A.G. Huesgen, R. Schuster, Sensitive analysis of synthetic colors using HPLC and diode-array detection at 190950 nm, Agilent Application Note 5964-3559E, 1995. 7. A. Herrmann, et al., Rapid control of vanilla-containing products using HPLC, J. Chromatogr., 1982, 246, 313316. 08. Official Methods of Analysis; W. Horwitz, Ed.; 14th ed.; AOAC: Arlington, VA, 1984; secs 12.018 12.021. 09. H. Malisch, et al., Determination of residues of chemotherapeutic and antiparasitic drugs in food stuffs of anomaly origin with HPLC and UV-Vis diode-array detection, J. Liq. Chromatogr., 1988, 11 (13), 28012827. 10. EC Guideline 86/428 EWG 1985. 11. M.H. Thomas, J. Assoc. Off. Anal.; 1989, 72 (4) 564. 12. Farrington et. al., Food Additives and Contaminants, 1991, Vol. 8, No. 1, 55-64. 13. Lebensmittel- und Bedarfsgegenstndegesetz, Paragraph 35, Germany.
14. W. Specht, Organochlor- und Organophosphor-Verbindungen sowie stickstoffhaltige sowie andere Pflanzenschutzmittel, DFG-Methoden sammlung, 1982, 19. 15. A new approach to lower limits of detectionand easy spectral analysis, Agilent Primer 5968-9346E, 2000. 16. R. Schuster, A comparison of preand post-column sample treatment for the analysis of glyphosate, Agilent Application Note 5091-3621E, 1992. 17. A.G. Huesgen, R. Schuster, Analysis of selected anions with HPLC and electrochemical detection, Agilent Application Note 5091-1815E, 1991. 18. Determination of triglycerides in vegetable oils, EC Regulation No. L248, 28ff. 19. L.M. Nollet, Food Analysis by HPLC New York, 1992. 20. A.G. Huesgen, R. Schuster, Analysis of selected vitamins with HPLC and electrochemical detection, Agilent Application Note 5091-3194E, 1992. 21. O. Busto, et al. Solid phase extraction applied to the determination of biogenic amines in wines by HPLC, Chromatographia, 1994, 38(9/10), 571578. 22. Sensitive and reliable amino acid analysis in protein hydrolysates using the Agilent 1100 Series, Agilent Technical Note, 5968-5658E, 2000 23. R. Schuster, Determination of amino acids in biological, pharmaceutical, plant and food samples by automated precolumn derivatisation and HPLC, J. Chromatogr., 1988, 431, 271284. 24. Capillary Liquid Chromatography with the Agilent 1100 Series Modules and Systems for HPLC, Agilent Technical Note 5965-1351E, 1996.
126
25. R. W. Frei and K. Zech, Selective sample handling and detection in HPLC, J. Chromatogr. 1988, 39A. 26. D. R. Gere et al., Bridging the automation gap between sample preparation and analysis: an overview of SFE, GC, GC/MSD and HPLC applied to several types of environmental samples, J. Chromatogr. Sci., 1993, July. 27. M.A. Schneidermann, et al., J. Assoc. Off. Anal. Chem., 1988, 71, 815. 28. R.Schuster, A comparison of pre- and postolumn sample treatment for the analysis of glyphosate, Agilent Application Note 5091-3621E, 1992. 29. M. Verzele et al., J. Am. Soc. Brew. Chem., 1981, 39, 67. 30. W.M. Stephen, Clean-up techniques for pesticides in fatty foods, Anal. Chim. Acta, 1990, 236, 7782. 31. J.E. Farrow, et al., Analyst 102, 752 32. H. Schulenberg-Schell et al., Poster presentation at the 3rd International Capillary Chromatography Conference, Riva del Garda, 1993. 33. S. K. Poole et al., Sample preparation for chromatographic separations: an overview, Anal. Chim. Acta, 1990, 236, 342. 34. R. E. Majors, Sample preparation perspectives: Automation of solid phase extraction, LC-GC Int. 1993, 6/6. 35. E. R. Brouwer et al., Determination of polar pollutants in river water using an on-line liquid chromatographic preconcentration system, Chromatographia, 1991, 32, 445. 36. I. McMurrough, et al., J. Am. Soc. Brew. Chem., 1988. 37. K. K. Unger, Handbuch fr Anfnger und Praktiker, 1989, Git Verlag, Germany.
38. W.O. Landen Jr., J. Assoc. Off. Anal. Chem., 1985, 68, 183. 39. L. Huber, Good laboratory practice for HPLC, CE and UV-Visible spectroscopy, Agilent Primer, 5968-6193E, 2000 40. R. L. Grob, M. A. Kaiser, Environmental problem solving using gas and liquid chromatography, J. Chromatogr. ,1982, 21. 41. A. G. Huesgen et al., Polynuclear aromatic hydrocarbons by HPLC, Agilent Application Note, 5091-7260E, 1992. 42. R. Schuster, A comparison of preand postolumn sample treatment for the analysis of glyphosate, Agilent Application Note, 5091-3621E, 1992. 43. H. Godel, Performance characteristics of the HP 1100 Series modules and systems for HPLC, Agilent Technical Note, 5965-1352E, 1996.
127
128
Index
129
Numerics
1,4-diaminobutan, 48 1 5-diaminopentane, 48 1-butylamine, 48 1-naphthol, 28 2,2'-dithiobis (5-nitro-pyridine), 108 2,4-dinitrophenyl hydrazine, 108 2-naphthacyl bromide, 108 2-phenylphenol, 58 3-hydroxycarbofuran, 28 3-ketocarbofuran, 28 3-methylbutylamine, 48 9-fluorenylmethyl chloroformate (FMOC), 108
A
absorption spectrum, 91 accreditation standards, 59 accuracy, 120, 122 acesulfam, 8 acetic acid, 2 acids 3,3'-thiodipropionic, 4 acetic, 2 adipic, 2 amino, 50 ascorbic, 4 benzoic, 5 butyric, 38 citric, 2, 3, 43 fatty, 35, 38, 60,108 folic, 43 fumaric, 2 lactic, 2 malic, 2 mercapto-propionic (MPA), 9 nordihydroguaiaretic, 4 oxalic, 3 panthothenic, 43 phosphoric, 2 propionic, 2, 5 sorbic, 2, 6 succinic, 2 tartaric, 2 acidulants, 2, 3 additives, III adipic acid, 2
adsorption chromatography, 59 adulteration, 35 aflatoxins, 21, 22, 23, 60 alcohols, 108 alcoholysis, 45 aldehydes, 108 aldicarb, 28 aldicarb sulfone, 28 aldicarb sulfoxide, 28 alducarb, 28 alkaline hydrolysis, 45 amines, 48 primary, 108 secondary, 108 amino acids, V, 50, 99 ammonia, 48 AMPA, 29 amperometric detection, 98 amylamine, 48 animal feed, 18, 21, 22, 44 anions, 33 inorganic, 32 antibiotics, III antioxidants, III, 4, 63 apples, 21, 22 artificial sweeteners, III, 8 ascorbic acid, 4 aspartame, 8 atmospheric pressure chemical ionization (APCI), 102 - 104 autoincrement mode, 100 automated injector, 72 autosampler, 72, 109
bitter compounds, 12, 14 bromophenacyl bromide, 38 butocarboxim, 28 butocarboxim sulfone, 28 butocarboxim sulfoxide, 28 butter, 38 butyric acid, 38
C
calibration curves, 113 settings, 115 tables, 116 capacity factor, 121 capillary electrophoresis, V capillary liquid chromatography, 52 carbamates, 26,103,109 carbaryl, 28 carbendazim, 27 carbofuran, 28 carbohydrates, III, 40, 41 Carrez, 7, 14 cell design (electrochemical) porous flow-through, 99, 100 thin-layer, 99, 100 wall-jet, 98, 100 cellobiose, 39 cereals, 19, 20 cheese, 48 chemical residues, 16 chemotherapeutics, 16 chewing gum, 5 chiral drug, V chloramphenicol, 15 chlorite, 99 chlorpyripho-ethyl, 27 chromophore, 38, 108 citric acid, 2, 3, 43 cleanup, 54 client/server-based software, 114 cognac, 13 collision induced dissociation (CID), 19 colorants, III, 10 column guard, 59, 67 narrow-bore, 59 standard-bore, 59 temperature, 60
B
backflash valve, 67 bacteria, 15 BASIC programming language, 112 beer, 48, 50 Beer's law, 121 benzoic acid, 6 benzothiazuron, 16 BHA butylated hydroxyanisole, 4 BHT butylated hydroxytoluene, 4 biogenic amines, 48 biotin, 43 biphenyl, 84 bisphenol A (BADGE), 24
130
compressibility, 79 computer networks, 114 computing equipment, 112 conductivity detector, 86 copper, 99 corn, 41 coulometric detection, 98 counter electrode, 98 cross sample reports, 114
direct solvent extraction, 45 drift, 87 drift trigger, 101 drinking water, 26, 33 dual-lamp design, 10 dual-piston mechanism, 78 dyes, V dynamic range, 88
D
dairy products, 22 dansyl chloride, 49 data evaluation, 112 generation, 112 storage, 114 dead volume, 59, 71, 121 DEG, 3 degassing, 82 helium, 83 ultrasonic, 83 vacuum, 83, 84 derivatization, 73 chemical, 62, 108 postcolumn, 109, 110 precolumn, 109, 110 detection amperometric, 98 coulometric, 99 detector, 86-105 conductivity, 86 diode array, 86, 90, 105 electrochemical, 86, 87, 98, 105 electroconductivity, 32 fluorescence, 87, 95, 105 mass spectrometer, 86, 88, 101, 105 refractive index, 86, 87, 104 response, 88 thermal energy, 86 UV, 89, 90 variable wavelength, 89, 90, 105 deuterium lamp, 10, 90, 92 DG dodecyl gallate, 4 diagnostic test, 114 diethylamine, 48 diode array detector, 87, 91 diquat, 26
E
eggs, 16,17 electrochemical detector, 86, 87, 98, 105 electroconductivity detector, 32 electrospray ionization, 102, 103 elution order, 60 emission, 96, 97 emission grating, 95 enzymatic hydrolysis, 45 essential oils, 12 ethanol, 3 ethanolamine, 48 ethiofencarb, 28 ethiofencarb sulfone, 28 ethiofencarb sulfoxide, 28 ethopabat, 16 ethylamine, 48 excitation, 96, 97 excitation grating, 95 extinction coefficients, 87 extraction liquid-liquid, 65 solid-phase, 63, 65 supercritical fluid, 64
fluorescence detection, 109 fluorescence detector, 87, 95, 105 fluorescent tag, 109 folic acid, 43 folpet, 27 Food and Drug Administration (FDA), V food colors, 10 fragmentation, 19 fructose, 40 fruit juices, 6 fruits, 28 fumaric acid, 2 fumonisins, 19 fungi, 15, 21 furazolidone, 15
G
galactose, 40 gel permeation chromatography (GPC), 27, 66 glassy carbon, 99 GLP/GMP principles, 114 glucose, 40 glycerol, 3 glyphosate, 26, 29 gold, 99 good laboratory practice (GLP), 67 gradient elution, 76 formation, 77 high pressure, 80 low-pressure, 78 guard column, 59, 67
F
fats, 35, 37, 38 fatty acids, 35, 38, 60, 108 fertilizers, III figs, 21, 22 fish, 48 flavors, III, 12 flour, 21, 30 flow precision, 79 ranges, 76 rates, 76
H
halogens, 99 hesperidin, 12, 14 hexylamine, 48 histamine, 48 hormones, III humulon, 12 hydrazine, 99 hydrogen peroxide, 99 hydrolysis, 38 hydroperoxides, 35, 36
131
I
indirect UV-detection, 33 injection volumes, 70 injector automated, 72 manual, 71 program, 110 inorganic anions, 32 inorganic ions, V instrument calibration, 114 logbooks, 114 parameters, 115 integration events, 115 parameters, 113 integrators, 112 , 113 interface (MS) moving belt, 102 particle beam, 102 thermospray, 102 intermediate precision, 122 iodide, 34 ion-exchange chromatography, 10 ion-exchange phases, 58 ionox-100 4-hydroxymethyl-2,6-di(tert-butyl) phenoI, 4 ion-pairing reversed-phasechromatography, 10, 11 ipronidazol, 16 isoabsorbance plot, 91, 92 isobutylamine, 48 isopropylamine, 48
limit of detection (LOG), 87, 120, 121 limit of quantification (LOQ), 87, 120, 121 linearity, 87, 88 lipids, 35 liquid-liquid extraction, 65 local area network (LAN), 117 long-term variability, 123 luminescence, 95 lupulon, 12
N
N-acetyl metabolite, 15 naringenin, 12, 14 narrow-bore column, 59 natural sweeteners, III Nernst equation, 98 networked data systems, (NDS), 112, 118 nicarbazin, 16 nitrites, 32 nitro compounds, 27 nitrofurans, 16 NOGA nordihydroguaiaretic acid, 4 noise, 87, 122 normalization, 113 normal-phase column, 46, 47 nuts, 21, 22
M
malic acid, 2 maltose, 40 mannitol, 40 manual injector, 71 margarine, 38, 47 mass spectra, 53, 54 mass spectrometer, V, 86, 88, 101, 105 meat, 16, 63 memory effect, 70 mercaptobunzothiazol, 26 mercapto-propionic acid, 9 mercury, 99 mercury-gold, 99 methabenzthiazuron, 26 methanol, 3 methiocarb, 28 methiocarb sulfone, 28 methiocarb sulfoxide, 28 methomyl, 28 methylamine, 48 meticlorpindol, 16 metronidazol, 16 microbial growth, 6 microorganism, 6 microsampling, 70 milk, 16, 21, 22 mixing noise, 81 molecular weight, 66 monochromator, 89 morpholine, 48 moving-belt interface (MS), 102 multichannel integrators, 113 multisignal, 97 multisignal detection, 92 mycotoxins, 21
O
oat seedlings, 52, 53 ochratoxin A, 21, 22 OG octyl gallate, 4 oils, 35 - 38 one-lamp design, 10 online spectral measurements, 96 o-phthalaldehyde (OPA), 9, 108 orange juice, 14 organic acids, V oxalic acid, 3 oxamyl, 28 oxidizable sulfur compounds, 108 oxytetracycline, 18
P
pantothenic acid, 43 paprika, 27 paraquat, 26 partition phases, 58 Patent blue, 10 patuline, 21, 22 p-bromophenacyl bromide, 108 peak co-eluting, 123 dispersion, 121, 122 elution, 93 identity, 93, 94 purity, 93, 94
K
ketones, 108
L
lactic acid, 2 lactose, 42 LC/MS, 52, 101, 102 LC/MSD, 19, 24 lemonade, 41 light intensity, 91
132
Peltier control, 60 peptides, 52 performance test, 122 personal computers, 114 pesticides, III, V, 26, 58 petrol ether, 35, 37 PG propyl gallate, 4 PHB-ethyl, 6 PHB-methyl, 6 PHB-propyl, 6 phenethylamine, 48 phenylisocyanate, 108 phenylurea-herbicides, 26 phosphoric acid, 2 photodiode, 89 array, 91 photomultiplier tube, 97 photoreceptor protein, 52 phytochrome proteins, 52 pistachio nuts, 23 platinum, 99 polycyclic aromatic hydrocarbons, 96 pork muscle, 18 postcolumn derivatization, 28, 29, 109, 110 potassium ferrocyanide, 14 precision, 120, 122 precolumn derivatization, 109, 110 precolumns, 65 preservatives, III, 6, 7, 63 procymidon, 27 propionic acid, 2, 6 propoxur, 28 propylamine, 48 protein precipitation, 14 proteins, 18 protozoa, 16 pulse ripple, 79 pump high-pressure gradient, 80 low-pressure gradient, 78 pumps, 76-84 pungency compounds, 12 pyrazon, 15 pyrrolidine, 48
R
raffinose, 40 redox potentials, 98 reference electrode, 98 refractive index detector, 86, 87, 104 regression analysis, 114 reintegration, 113 repeatability, 122 reproducibility, 122 residues, 16, 26 reversed optics, 91 reversed phase, 58 riboflavin 5' phosphate, 43
S
saccharin, 8, 43 salad, 27 salad dressing, 7 sample cleanup, 59 preparation, 62 pretreatment, 72, 110 volume, 72 sampling device, 70 scanning, 96 selected ion mode (SIM), 105 selectivity, 87 separation, 58 sequences, 116 silver, 99 size-exclusion chromatography, 66 slit, 91 smoked sausage, 32 soft drinks, 8 solid-phase extraction, 65 sorbic acid, 2, 6 sorbitol, 39 spectral libraries, 115 resolution, 91 spices, 21, 22, 27 spikes, 122 spreadsheet, 114 standard external, 113 internal, 113 standard-bore column, 59
steam distillation, 64 sterols, 35 strip chart recorders, 112 succinic acid, 2 sucrose, 40 suitability reports, 110 sulfapyridine, 16 sulfite, 99 sulfonamides, 16 sulfur dioxide, 6 supercritical fluid extraction (SFE), IV, 63, 64 surfactants, V sweeteners, 8 switching valves, 63 system suitability, 116
T
table salt, 34 tartaric acid, 2 TBHQ mono-tert-butylhydroquinone, 4 TDPA 3,3'-thiodipropionic acid, 4 tetracyclines, 18 tetrahydrofurane, 35 THBP 2,4,5-trihydroxybutyrophenone, 4 thermal energy detector, 86 thermal stability, IV thin-layer chromatography (TLC), 21 thiofanox, 28 thiofanox sulfone, 28 thiofanox sulfoxide, 28 thiosulfate, 99 tocopherols, 4, 45 - 47 tocotrienols, 46 tolerance levels, III total ion chromatography, 53, 54 toxicity, 8 toxins, 19 tracer, 80, 121 trend charts, 114 triazines, 26 triglycerides, 35 - 39 trypsin, 53 tryptamine, 48 tungsten lamp, 10, 91
Q
qualitative information, 87, 88, 120, 123 quenching effects, 82 Quinolin yellow, 10
133
U
ultrasonic bath liquid extraction, 63 UV absorbance, 86 UV detector, 89, 90
V
validation processes, 113 vanillin, 12, 13 variable volumes, 70 variable wavelength detector, 89, 90 vegetables, 26, 28 vinclozolin, 27 viruses, 15 viscous samples, 60 vitamins, V, 4, 35, 42, 44, 66 drink, 43 fat-soluble, 42, 46 natural, III standard, 46 synthetic, III tablets, 42, 43 water-soluble, 42, 43 vodka, 2
W
wavelength switching, 96 wine, 2, 7, 48 wool-fiber method, 10 working electrode, 98
X
xenon flash lamp, 95
Z
zearalenone, 21, 22 zinc sulfate, 14
134
Retention Time Locking: Concepts and Applications
Application
Gas Chromatography December 1997
Authors
Vince Giarrocco Bruce Quimby Matthew Klee Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA
Key Words
Retention time locking, method validation, styrene analysis, ASTM D 5135, capillary gas chromatography, laboratory productivity
Introduction
Retention time is the fundamental qualitative measurement of chromatography. Most peak identification is performed by comparing the retention time of the unknown peak with that of a standard. It is much easier to identify peaks and validate methods if there is no variation in the retention time of each analyte. However, shifts in retention time occur frequently. Routine maintenance procedures such as column trimming alter retention times. In a multi-instrument laboratory running duplicate methods, the retention times for each instrument will differ
Abstract
The concepts and applications of retention time locking (RTL) are described. RTL simplifies the process of transferring methods from chromatographic instrument to chromatographic instrument, column to column, and detector to detector. The analysis of impurities in styrene according to ASTM D 5135 is used to demonstrate the efficacy of the approach. Using RTL, the retention times matched within an average of 0.16% (0.020.03 minute) in constant pressure modes.
from each other, even when run under nominally identical conditions. These differences in retention times mean that each instrument must have a separate calibration and integration event table, making it time-consuming to transfer methods from one instrument to another. Differences in retention time also complicate comparison of data between instruments and over time. Retention time locking (RTL) is the ability to very closely match chromatographic retention times in any Agilent 6890 gas chromatograph (GC) system to those in another 6890 GC system with the same nominal column. There are several subtle effects that combine to cause retention time differences between similarly configured GC systems. Columns of the same part number can vary slightly in length, diameter, and film thickness.
GC pneumatics can have small variations in the actual inlet pressure applied at a given setpoint. The actual temperature of the GC oven also has minute but real deviations from the indicated value. The sum of these and other effects result in the observed retention time differences between similarly configured GC systems. The pneumatics and oven temperature control of the 6890 GC have advanced the state of the art in GC hardware accuracy and precision. Agilents advances in fused silica capillary column technology have resulted in highly reproducible column-to-column retention characteristics. With these advances, retention time precision for a given peak in a single GC setup is usually better than 0.01 minute. However, even with these advances in columns and instrument hardware, the sum of the effects mentioned above can cause retention time differences between identically configured GC systems of as much as 0.4 minute. It would be impractical to control all of the instrument and column variables to a degree where retention time differences between similarly configured GC systems are removed. There is, however, a means of greatly reducing these differences. By making an adjustment in the inlet pressure, the retention times on a given GC setup can be closely matched to those of a similarly configured GC system. RTL is based on this principle. The process of RTL is to determine what adjustment in inlet pressure is necessary to achieve the desired match in retention times. Agilent RTL software (G2080AA), which integrates into the Agilent GC ChemStation (version A.05.02 or later), provides the tool required to determine the correct inlet pressure quickly and simply.
There are several advantages gained by using RTL in the laboratory. Peak identification becomes easier and more reliable. It is easier to compare data both between instruments and over time. Comparison of data when using different detectors for analyte identification is simplified. Transferring methods from instrument to instrument or laboratory to laboratory is easier because calibration time windows normally will not require readjustment. Validation of system performance is easier. With locked GC methods, the development and use of retention time data bases for unknown identification is much more straightforward. To maintain a locked method, RTL should be performed whenever: The column is changed or trimmed The method is installed on a new instrument A detector of different outlet pressure is used System performance is validated Troubleshooting chromatographic problems
Systems where the predicted locking pressure falls outside the range of the current calibration
A specific solute (usually one found in the normal method calibration standard) must be chosen and then used for both developing the locking calibration and locking all future systems. The solute, or target peak, should be easily identifiable, symmetrical, and should elute in the most critical part of the chromatogram. Solutes that are very polar or subject to degradation should be avoided. Once the target solute has been chosen and all other chromatographic parameters of the method have been determined, five calibration runs are performed. The runs are made at conditions identical to the nominal method except that four of the runs are made at different pressures. The pressures used are typically: Target pressure 20% Target pressure 10% Target pressure (nominal method pressure) Target pressure + 10% Target pressure + 20%
To lock a given method for the firsttime or for the reasons below, one must first develop a retention time versus pressure (RT vs. P) calibration. Even when using columns with the same part number (same id, stationary phase type, phase ratio, and same nominal length), separate/different locking calibration curves are needed when using: Systems with different column outlet pressures (FID/atmospheric, MSD/vacuum, AED/ elevated) Columns differing from the nominal length by more than 15% (e.g., due to trimming)
The retention time of the target compound is determined for each run. The resulting five pairs of inlet pressures and corresponding retention times are entered into the ChemStation software to generate an RTL calibration file. Figure 1 shows the dialog box used to enter the calibration data. After the data is entered, a plot is displayed, as shown in figure 2. The maximum departure of the fitted curve from the data is given for both time and pressure. If the fit is acceptable, the retention time versus pressure calibration is stored and becomes part of the GC
method. This calibration need only be generated once. Subsequent users of the method can use this calibration when running the method on a similar instrument setup, regardless of location. To relock a system or lock a new one: 1. Set up the method conditions and run a standard containing the target compound. 2. Enter the actual retention time of the target compound into the (Re)Lock current method dialog box (see figure 3). 3. Update the 6890 method with the new calculated pressure, and save the method. 4. Validate the retention time lock by injecting the standard at the new pressure, and compare the retention time obtained to the desired retention time. 5. Repeat steps 2 to 4, if necessary. Figure 1. Dialog box used for entering retention time locking calibration data
A Note on Constant Flow versus Constant Pressure Modes of EPC Operation

Many GC chromatographers prefer to use the constant flow mode of EPC operation. In this mode, inlet pressure increases automatically to maintain constant outlet flow rate as the oven temperature increases during the run. Constant flow mode reduces run time and ensures that flow-sensitive detectors see a constant column effluent flow. The constant pressure mode of EPC operation is also popular. In this mode, the pressure remains constant during the run (outlet flow will decrease as temperature increases). For those wishing to reduce run time in constant pressure mode, a higher pressure can be chosen. For
Figure 2. Plot of calibration data as displayed by RTL software
Figure 3. Dialog box used to calculate locking pressure and update the 6890 method
flow-sensitive detectors, one can set constant column flow + makeup via the 6890 keyboard or ChemStation. In this mode, the makeup flow is increased as the column flow decreases to keep the sum of the two constant. The underlying theory of RTL predicts that constant pressure mode of EPC provides the closest matching of retention times. If one desires to compare data from systems with very different configurations, such as GC/FID to GC/MSD, it is best to use constant pressure mode. As can be seen from the styrene analysis data herein, retention time matching between systems of the same configuration (GC/FID, in this case) is still quite good in the constant flow mode. This application note shows the use of RTL to lock retention times between multiple chromatographic instruments, columns, and detector types and demonstrates RTL in both constant flow and constant pressure modes.
Temperature program: 80 C (9 min), 5 C/min to 150 C
The inlet pressures/flows used are indicated with each chromatogram. A third 6890 Series GC was also used. This system was equipped with an Agilent 5973 mass selective detector (MSD) and was used for peak identification. The GC-MSD chromatographic parameters used were the same as the GC systems noted above except for the inlet pressures as indicated.
The sample was then run at four other pressures to collect the five data pairs for RTL calibration. Because this method was run in constant flow mode, the pressures entered into the RTL software were the initial pressures. The =-methylstyrene peak (peak 10) was chosen as the target compound. The calibration data are shown in figure 1. The method conditions and RTL calibration were then moved to GC system 2, a different GC and column. The sample was run at the original method inlet pressure of 18.2 psi. The chromatogram obtained using this scouting run is overlaid on the original chromatogram in figure 5. The retention times shifted about 0.3 minute on the second GC. This is a typical result obtained when trying to replicate an analysis on a second instrument or with a second column. The retention time of =-methylstyrene was entered into the RTL software

GC-FID to GC-FID Locking
Figure 4 shows the original chromatogram (GC system 1) obtained from running a styrene sample under the conditions specified in ASTM D 5135.1 Many of the typical impurities found in styrene are found here. The phenylacetylene peak represents about 60 ppm. The peaks are identified in table 1.
pA 28 26 24 2
Experimental
Two 6890 Series GC systems were used. Each system was equipped with: Electronic pneumatics control (EPC) Split/splitless inlet (250 C, He carrier gas, split 80:1) Automatic liquid sampler
10
6 22 20 18 16 5 11 7 14 12 12 1 8 13
GC ChemStation (version A.05.02) Flame ionization detector (FID) 60 m 0.32 mm, 0.5 mm HP-INNOWax column (part no. 19091N-216)
2.5
7.5
10
12.5
15
17.5
20
22.5
min
Figure 4. Styrene sample run on GC system 1 at 18.2 psi initial pressure, constant flow mode
dialog box on GC system 2, as shown in figure 3. The RTL software indicated the initial pressure should be modified from 18.2 psi to 18.96 psi. The new initial pressure was entered into the method and saved. Figure 6 compares the chromatograms obtained from the original run and after locking retention times using the =-methylstyrene. Table 2 compares the retention times before and after using this approach. The retention times are now closely matched.
Table 1.
Peak # 1 2 3 4 5 6 7
Peak Identities for Figure 4

Name Nonaromatics Ethylbenzene p-Xylene m-Xylene i-Propylbenzene o-Xylene n-Propylbenzene Peak # 8 9 10 11 12 13 Name p/m-Ethyltoluene Styrene =-Methylstyrene Phenylacetylene >-Methylstyrene Benzaldehyde
pA 27.5 25.6
Ethylbenzene
=-Methylstyrene
10.318 min 22.5 20 10.658 min
17.778 min 18.099 min
GC-FID to GC-MSD Locking

17.5
A second experiment was conducted to lock the original method from GC system 1 to the GC-MSD. This is useful for identification of unknown impurities that show up in the FID chromatogram. For example, there is a shoulder evident on the front side of the phenylacetylene peak in figure 4. It would simplify locating the impurity in the GC-MSD data if the retention times closely matched that of the GC-FID. Because constant pressure mode is preferred when comparing data from FID and MSD systems, constant pressure mode was chosen, and the styrene sample was re-run on GC system 1 at 18.2 psi for reference. The next step was to determine the chromatographic conditions to be used on the GC-MSD. The Agilent method translation software tool was used to calculate the conditions necessary to have the peaks elute in the identical order on the two systems.2,3 Because the retention times need to match, the dead time and temperature program used for running the GC-MSD must be the same as the GC
15 12.5 10
Original (GC system 1, column 1)
Scouting (GC system 2, column 2) 5 7.5 10 12.5 15 17.5 20 22.5 min
Figure 5. Comparison of original chromatogram on GC system 1 with GC system 2 before retention time locking
pA 27.5 25.6 22.5 20 17.5 15 12.5 10
Ethylbenzene 10.318 min vs. 10.298 min
=-Methylstyrene
17.778 min vs. 17.776 min
Original (GC system 1, column 1)
Locked (GC system 2, column 2) 5 7.5 10 12.5 15 17.5 20 22.5 min
Figure 6. Comparison of original chromatogram on GC system 1 with GC System 2 after retention time locking
method. The pressure used, however, will be different due to the difference in column outlet pressure. The GC-MSD inlet pressure is calculated using the none mode of the method translation software (figure 7). In this mode, the holdup time between the two columns was forced to be identical to the GC-FID. This gives a speed gain of 1. The pressure calculated for use on the GC-MSD was 8.44 psi. Note that this calculated pressure is only the nominal pressure required to get similar retention times, not the exact locking pressure. A different RTL calibration is required for GC-MSD because the outlet pressure is vacuum, and that of the FID is atmospheric pressure. Five runs were made on the GC-MSD system bracketing the 8.44 psi nominal method pressure. Because the GC-MSD used in this study was not equipped with RTL software, a dummy method was created in GC system 1 and the GC-MSD RTL calibration data was entered into it. A scouting run of the Styrene sample was made on the GC-MSD, and the =-methylstyrene retention time was used for locking. The locking inlet pressure calculated with the dummy method was 7.9 psi and was entered into the GC-MSD. Figure 8 shows the resulting matched chromatograms from the GC-FID and GC-MSD. As seen in table 3, the retention times are now closely matched within 0.02 minute. Figure 9 shows the MSD first choice of library search result of the impurity that created the shoulder on the front side of the Phenylacetylene peak. RTL ensured that this shoulder remained separated on the MSD system and eluted at the same time
Table 2.
GC-FID Retention Times Before and After Locking for Styrene Impurities (Constant Flow Conditions). Chromatograms Shown in Figures 4, 5, and 6.
Original Run GC 1/Column 1 18.2 psi 10.318 10.616 10.858 11.985 12.533 13..360 17.778 18.806 20.248 24.097 GC2GC1 Before RTL 0.340 0.333 0.337 0.359 0.345 0.364 0.321 0.275 0.310 0.279 0.326 Scouting Run GC 2/Column 2 18.2 psi 10.658 10.949 11.195 12.344 12.878 13.724 18.099 19.081 20.558 24.376 GC2GC1 After RTL 0.020 0.026 0.022 +0.005 0.012 0.016 0.002 0.040 0.006 0.069 0.028 Locking Run GC 2/Column 2 19.0 psi 10.298 10.590 10.836 11.990 12.521 13.376 17.776 18.766 20.242 24.028
Component Ethylbenzene p-Xylene m-Xylene i-Propylbenzene o-Xylene n-Propylbenzene =-Methylstyrene* Phenylacetylene >-Methylstyrene Benzaldehyde Average , * Used in locking calculation
Figure 7. Method translation software provides scaled conditions for GC systems with different configurations
for easy comparison to the FID results.
Conclusions
Retention time locking facilitates replicating results from instrument to
instrument, from column to column, and from detector to detector by locking retention times. The retention times of a styrene sample analyzed according to ASTM D 5135 matched to within 0.06 minute after locking.
References
1. ASTM D 5135-95, Analyses of Styrene by Capillary Gas Chromatography, Annual Book of Standards, Volume 06.04, ASTM, 100 Bar Harbor Drive, West Conshohocken, PA 19428 USA. 2. M. Klee and V. Giarrocco, Predictable Translation of Capillary GC Methods for Fast GC Agilent Technologies, Inc., Application Note 228-373, Publication 5965-7673E, March 1997. 3. GC Pressure/Flow Calculator for Windows, Version 2.0 and Method Translation Tool Version 2.0. Available at http://www. chem.agilent.com/servsup/ usersoft/main.html.
GC-FID
GC-MSD, TIC
1.0
3.0
5.0
7.0
9.0
11.0 13.0
15.0
17.0
19.0
21.0
23.0
25.0 min
Figure 8. Comparison of chromatogram on GC system 1 with GC-MSD system after retention time locking, Constant Pressure Mode Table 3. GC-FID vs. GC-MSD, Method Translated then LockedRetention Times (Constant Pressure Conditions)
GC-FID Original 18.2 psi 10.315 10.620 10.869 12.038 12.613 13.492 18.276 19.406 21.008 25.475 GC-MSD 7.9 psi 10.338 10.642 10.890 12.053 12.630 13.508 18.267 19.389 20.987 25.415 Average
* Used in locking calculation
Component Ethylbenzene p-Xylene m-Xylene i-Propylbenzene o-Xylene n-Propylbenzene a-Methylstyrene* Phenylacetylene b-Methylstyrene Benzaldehyde
RT Difference min 0.023 0.022 0.021 0.015 0.017 0.016 0.009 0.017 0.011 0.060 0.021
Phenylacetylene
1-Ethenyl-3-methyl-benzene
17.6
18.0
18.4
18.8
19.2
19.6
20.0
20.4
20.8
min
Figure 9. GC-MSD identification of impurity in shoulder of phenylacetylene peak
Large Volume Injection for Gas Chromatography Using a PTV Inlet
Application
Gas Chromatography March 1997
Authors
Bill Wilson, Philip L. Wylie, and Matthew S. Klee Agilent Technologies, Inc. 2850 Centerville Road Wilmington, DE 19808-1610 USA
Introduction
The demand for lower detection limits is important in environmental, pharmaceutical, food analysis, and other gas chromatography (GC) applications. This demand has driven instrument manufacturers to provide more sensitive instrumentation and procedures, which has prompted regulators to reduce allowable limits, and so on in a never-ending cycle. Improvements in sample handling, sample injection techniques, and detectors have all contributed to the ability to measure compounds at decreasing levels. Concentrating samples is an established approach for increasing method sensitivity. For many environmental analysis, this involves extraction followed by solvent evaporation, which generates large volumes of waste solvent and increases sample preparation time significantly. Recent advances, such as supercritical fluid extration (SFE), solid-phase extraction (SPE), solid-phase microextraction (SPME), and pressurized fluid extraction, are making inroads on liquid/liquid extractions, but these still involve additional sample preparation time.
Abstract
Reduced sample detection limits is a continuing goal in gas chromatography. Large sample injection volume is one possible approach. New inlets and injection techniques supporting large volume injection (LVI) have been developed in recent years. This paper discusses the uses and limitations of LVI, describes LVI with a programmable temperature vaporizer (PTV) inlet, and reviews the results of LVI analysis of pesticides and straight-chain hydrocarbons.
Detection limits can be reduced by lowering system background and interferences. This can be done through sample cleanup, such as florisil column chromatography, SPE, and SPME, and by using selective detectors that do not respond to the background. How-ever, the cleanup requires time, and the need for reduced limits has surpassed the sensitivity of even the best detectors. Large volume injection (LVI) is another approach to lower detection limits. The typical injection volume for capillary column analysis is 0.5 to 2 L. Agilent 6890 Series and 5890 gas chromatographs (GCs) allow approximately two times the normal injection volume (up to 5 L, depending on the solvent) using "pulsed" splitless injection. Injecting still larger volumes with standard techniques can lead to contamination of the system, irreproducible results, and loss of sample. With the new LVI technique, good chromatography can be obtained with injection volumes of 5 to 500 L or more. Table 1 summarizes several common ways to lower detection limits.
Key Words
Large volume injection, LVI, programmed temperature vaporizer, PTV, gas chromatography, GC, pesticide analysis, hydrocarbon analysis
Table 1. Approaches to Lowering Detection Limits

Concentrate the Sample Extraction followed by evaporation of solvent Solid-phase extraction (SPE) Solid-phase micro extraction (SPME) Headspace sampling Purge and trap (P&T) sampling Large volume injection (LVI) Transfer More Sample into the Column Cool on-column (COC) injection Splitless injection Large volume injection (LVI) Use More Sensitive or Selective Detectors Electron capture detector (ECD) Electrolytic conductivity detector (ELCD) Atomic emission detector (AED) Selected-ion monitoring mass spectrometry (SIM-MS) Nitrogen-phosphorus detector (NPD) Flame photometric detector (FPD) Decrease System Noise Selective detectors Sample cleanup to reduce interferences Use headspace or purge and trap sampling Decrease column bleed
In LVI, a large volume of sample is injected. The bulk of the solvent is evaporated before transfer of the sample to the analytical column and the start of the analytical sepa-ration. There are two primary techniques to eliminate solvent: PTV and cool oncolumn injection with solvent vapor exit (COC-SVE). COC-SVE is most appropriate for clean samples with volatile (early-eluting) compo-nents such as extracts of drinking water. The COC-SVE technique is discussed elsewhere.1
The Agilent 6890 Series GC uses a standard automatic sampler with syringe sizes up to 50 L. A 50- L syringe can inject up to 25 L. Multiple injections can be used with the PTV inlet when even larger volumes are required. With the 6890 GC system, delay between injections can be controlled as well as the number of injections or the total injection volume. Injection parameters are set through the Agilent ChemStation. A problem with multiple injections is the increased number of punctures of the GC inlet and vial septa. This reduces septum life and increases the possibility of contamination of sample and inlet. Using a "septumless head" for the inlet can eliminate the inlet problems. Figure 2A shows septum cap extract that contaminated the sample after the vial was pierced 40 times during several multiple-injection experiments. 100% Teflon septa minimize sample contamina-tion such as this, but once punctured, Teflon septa do not reseal. The PTV inlet can be considered a temperature-programmable split/ splitless inlet with the same basic configuration. While it can be used hot for split and splitless applications, this is not recommended because the volume of vaporized solvent may exceed the low internal volume of the PTV inlet. PTV is ideal for cold split or splitless applications,
avoiding most of the problems associated with hot inlets such as sample discrimination, liner overload, and sample decomposition. For large volume injections, the PTV is used in a "solvent vent" or "solvent elimination" mode. Sample is introduced into the inlet with the inlet temperature near the boiling point of the solvent and with a relatively high split ratio. The solvent (and low-boiling solutes) is vented while the higher boiling solutes (more than about 100C above the solvent boiling point) remain and are concentrated in the inlet. After a preset time, the split vent is closed and the inlet temperature increased to transfer the solutes and any residual solvent to the column for separation. Because the sample is evaporated from the inlet, nonvolatile sample components and degradation products remain behind in the inlet, minimizing column contamination. There is evidence that inlet contamination in PTVs influences subsequent injections less than in hot inlets. If contamination becomes an issue, the inlet liner is easily changed. Thus, PTV is a better choice for dirty samples than cool-on-column and split/ splitless inlet.
PTV
LVI with PTV is ideal for trace analysis of later eluting solutes (boiling points approximately 100C higher than the solvent) and for dirty samples. Typical injection volumes for solvent elimination PTV are 25 to 100 L. Injection volumes up to 1 mL have been demonstrated.2 The large sample volumes are injected by manual injection, by multiple sample injections from a standard automatic sampler, or by LVI with a variablespeed injector. An automatic injector is recommended for maximum reproducibility. A standard automatic sampler making repeat injections is more cost effective than purchasing a variable-speed sampler and requires less solvent for syringe cleaning.
For a PTV inlet to work well, inlet temperature must be programmed independently from the column oven. The 6890 provides this function. For example, the inlet can be heated with the split flow off to transfer the sample to the column before the oven temperature program begins. After sample transfer, the inlet can be heated further to bake off contaminants with a high split flow to minimize inlet contamination. LVI by PTV is not a good choice when the target compounds include highly volatile species because these lowboiling compounds are vented along with the solvent. The lowest boiling target compound should boil at least 100C above the solvent to have a reasonable chance of success. Table 2 lists the advantages and disadvantages of LVI by solvent elimination PTV.
Table 2. Advantages and Disadvantages of LVI by Solvent Elimination PTV

Advantages Most flexible LVI technique Good for late-eluting compounds such as pesticides, PAHs, etc Inlet protects column so one can use dirty samples Disadvantages Loss of volatile sample components Possibility of sample decomposition(although less than with split/splitless) More difficult to use than conventional inlets like split/splitless
Table 3. Software, Hardware, and Firmware Versions that Support LVI-PTV

Item ChemStation software G1513A injector G1512A ALS 6890 GC Software/Firmware A.04.02 or higher A.09.10 A.01.08 A.02.01 or higher
Experimental
A 6890 GC with electronic pneumatics control (EPC) was used. A G1916A automatic liquid sampler (ALS) with a G1513A controller performed sample injection. An Agilent ChemStation (version A.04.02) controlled the instruments and acquired and processed data. Table 3 lists the hardware and software revisions that support multiple injections, postdwell time, and the slow plunger mode required for LVI-PTV. Experimental conditions for the GC methods are given with the chromatograms.

Multiple Injections
Multiple injections are a straight-forward and reliable way to introduce large sample volumes into the inlet. Figure 1 shows the linearity obtained from two sets of multiple injections using a standard G1513 ALS. Depending on solvent type and injection volume, liquid sample may run down the liner and enter the column. If this occurs, the column may overload with solvent causing catastrophic peak splitting and possible damage to the stationary phase. To minimize this possibility, a packed liner should be used with multiple injections of more than 5 L. In addition to preventing solvent from flowing into the column, glass wool or other packing provides a surface to retain a film of solvent which, in turn, helps retain early-eluting compounds. Figure 2A shows loss of analytes eluting before C18. Figure 2B, with glass wool packing in the liner, shows complete recovery down to C14.
Area Count 12,000
Packed Liner
10,000 Response Factor RSD C23: 2.1% C22: 2.6% 6,000
8,000
4,000
2,000
0 0
50
100
150
200
250
Total Injection Volume, L
Figure 1. Linearity of multiple injections

pA 2500
Inlet = 40 C at injection Open baffle liner 50 mL/min purge for 2.5 min 30 m x 0.32 mm x 0.25 m HP-5
C18
2000
1500 C16 Septum extract 1000 C14 (vial pierced 40 times)
500 C12 C10 0 0 pA 2.5 5 7.5 10 12.5 Minutes 15 17.5 20 22.5
B
6000
C14 C16 C18
Glass wool packed liner 10 x 25 L injections
5000 4000 3000 2000 C12 1000 C10 0 0 2.5 5 7.5 10 12.5 Minutes
15
17.5
20
22.5
Figure 2. Comparison of packed and open baffle liners using PTV sample: C10-C44 in Hexane
PTV
Figure 3 is a chromatogram of an LVI of pesticides using a PTV inlet. By injecting 25 L divided into five injections, good response is obtained from a 0.01-ppm mixture COC-SVE is usually the preferred technique for samples with low-boiling components, because volatile compounds evaporate with the solvent (as shown in Figure 2A) with PTV. However, for volatile samples that are too dirty for COC-SVE sampling, the PTV inlet can be cryocooled below ambient temperature resulting in much better recovery of the low boilers. Figure 4 compares the recovery of C10 using PTV withthe inlet temperature at 40C and -10C during the injection step. There is 100% recovery of C10 with cryocooling. Much less solvent was eliminated at the lower temperature, which helped retain the early-eluting compounds. It is important to determine carefully the best injection parameters for LVI methods. Figure 5 shows the improvement in peak shape obtained for a sample containing pesticides when the initial PTV temperature, vent flow, and injection delay are optimized. Inlet temperature and vent flow influence the speed and extent of solvent removal. The chromatogram in figure 5B indicates that insufficient solvent was vented, thereby increasing the chance of carry-over,degrading peak shape, andincreasing ghost peaks.
1. Methamidophos 2. Acephate 3. Dimethoate 4. Diazinon
5 X 5 L Injections 5. Chlorothalonil 6. Chlorpyrifos-Methyl 7. Chlorpyrifos 8. Thiabendazole 9. Imazalil 10. Ethion 11. Phosmet 12. Azinphos-methyl
4 5 6 1 2 3
8 9
10
11 12
0 5 6 7 8 Minutes 9 10 11 12
Figure 3. LVI with solvent elimination PTV Pesticides: 0.01 ppm; PTV conditions: vent flow, 300 mL/min;vent pressure: 0 until 1 min; purge flow: 50 mL/min at 3.50 min; gassaver on at 4.70 min; PTV initial temperature: 20 C; PTV initial time:1.1 min; PTV rate: 700 C/min; PTV final temperature: 300 C; injection delay: 0.00 min; column: 30 m x 0.25 mm x 0.25 m HP-5MS
pA 3000 2000 C20 C16 1000 0 Hexane Solvent C14 C10 C12
PTV = 40 C
C23 100 % Recovery
0
pA 3000 2000 1000 0
6 C10
8 C14
10
12
C23 C16
PTV = 10 C
C12 C20
Conclusion
GC using LVI is useful for lowering detection limits. It has broad applicability for applications that require more sensitivity than can be obtained with standard injection volumes. PTV inlets are appropriate for LVI of lateeluting samples, for dirty samples, and for any cold split/splitless injection. Total automation of temperature, pressure, and flow simplifies
6 Minutes
10
12
Figure 4. PTV with cryocooled inlet
PTV use. LVI with solvent elimination PTV requires careful method development for maximum accuracy and reproducibility.
UsingCOC-SVE," Agilent Technologies, Application Note 228-377, Publication Number (23)5965-7923E, March 1997. 2. J. Staniewski and J. Rijks, HRC, 16(1993)182.
References
1. B. Wilson, et al., "Large Volume Injection for Gas Chromatography
Abundance 9e+06
9e+05 Initial PTV Temp = 20 C Vent Flow = 300 mL/min
7e+06
5e+05
5e+06
2e+05 7.60 8.00 8.40 8.70
3e+06
A
500000 0 5 1.2e+07 6e+07 9e+06 Initial PTV Temp = 35 C Vent Flow = 100 mL/min 6 7 8 9 10 11 12
4e+07
5e+06 2e+06
B
2e+07 8.20
8.60
9.00
5000000 0 5 6 7 Minutes 8 9 10 11 12
Figure 5. Effect of PTV setpoints on peak shape sample: pesticides at 1.0 ppm (peaks identified in figure 3); column: 30 m x 0.25 mm x 0.25 mm HP-5MS
Index
Pesticides and Residues Applications
Pesticides and Residues: Pesticides and Herbicides Replacing Multiple 50-min GC and GC-MS/SIM Analyses with a 15-min Full-Scan GC-MS Analysis for Nontargeted Pesticides Screening Reducing Analysis Time Using GC/MSD and DRS Analysis of Herbicides by RRLC with Online Trace Enrichment Improving Productivity and Extending Column Life with Backflush Using RTL and 3-Way Splitter to Identify Unknown in Strawberry Extract Automated Screening of 600 Pesticides in Food by LC/TOF MS Multiresidue Analysis of 100 Pesticides in Food Samples by LC/Triple Quadrupole MS Improving the Effectiveness of Method Translation for Fast and High Resolution Separations Determination of 44 Pesticides in Foodstuffs by LC/MS/MS Screening for 926 Pesticides and Endocrine Disruptors by GC/MS with DRS and a Pesticide Library Screening for Hazardous Chemicals Using a GC/MS/ECD/FPD with a DRS Database Analysis of Fungicides and Their Metabolites in Citrus Fruits and Juices by TOF and Trap LC/MS Analysis of Terbuthylazine in Olive Oil by TOF and Trap LC/MS Identifying Pesticides with Full Scan, SIM, uECD, and FPD from a Single Injection Determination of Fungicides in Fruits and Vegetables by TOF and Trap LC/MS Identification of Unknown Pesticides in Food Using TOF and Trap LC/MS New Tools for Rapid Pesticide Analysis in High Matrix Samples Study of Pesticide Residues in Spiked and Unspiked Fruit Extracts Using DRS Analysis of Organochlorine and Pyrethroid Pesticides with 6820 GC/Micro-ECD Analysis of Organophosphorus Pesticides with 6820 GC/NPD Analysis of Carbamate Pesticides Using 6820 GC/NPD Comprehensive Pesticide Screening by GC/MSD Using DRS Improving the Analysis of Organotin Compounds Using RTL Methods and Databases Validated Method for Phenyl Urea and Triazine Herbicides in Water by LC/MS Using SIM Analyzing Phenyl Ureas and Carbamate Pesticides Using ESI-, APPI-, and APCI-LC/MSD Analysis of the Active Compound in a Fungicide Formulation by LC Herbicides Examples of Bonded Phase Selectivity Differences Pesticides - Analysis of Pesticides in Drinking Water High-Speed Separation of Pesticide Mixture Analysis of Components, Contaminants, and Impurities in Fungicides by GC/MS and LC/MS Analysis of CLP Pesticides Using High-Temperature DB-35ms and DB-XLB Columns Complete Solution for Chlorinated Pesticides and Herbicides Using DB-35ms and DB-XLB Columns Analysis of Trace Level Carbamate Pesticides in Food Using LC/MSD Identification and Quantitation of Pesticides in the Parts-per-Trillion Range Using RTL and GC/MS GC/MS Approaches to the Analysis of Monochloropropanediol SPE of Chlorinated Pesticides in Water Analysis of Organophosphate Pesticides by LC/MS A New Approach to the Analysis of Phthalate Esters by GC/MS Development of an LC/MS Method for the Analysis of Rodenticides Screening of Pesticides and Endocrine Disrupters Using 6890/5973N GC/MSD, Part II
Index
Pesticides and Residues: Pesticides and Herbicides (continued) Trace Level Pesticide Analysis by GC/MS Using LVI Fast Dual-Column GC/ECD Analysis of Chlorinated Pesticides -- EPA Methods 606/8080 Detection of Low Levels of Carbaryl in Food Using LC/MSD Ion Trap System Screening Method for 567 Pesticides and Suspected Endocrine Disrupters Analysis of Chlorinated Pesticides and PCBs Using HP-608 Capillary Column Fast Screening of Pesticides and Endocrine Disrupters Using 6890/5973N GC/MSD, Part I HPLC Analysis of Glyphosate in Water with Postcolumn Derivatization Analysis of Pesticides in Salad Samples and Spices Using HPLC Pesticides and Residues: Antibacterial Drug Residues Detection, Confirmation, and Quantification of Chloramphenicol in Honey, Shrimp and Chicken Using 6410 QQQ LC/MS Analysis of Nitrofuran Metabolites in Tilapia Using 6410 QQQ LC/MS Determining Malachite Green and Leucomalachite Green in Food by LC/MS/MS Quantitation of Nitrofuran Metabolites in Shrimp and Poultry Using LC/MSD Trap XCT Analysis of Fluoroquinolones in Beef Kidney Using Electrospray LC/MS Validated APCI Method for Analyzing Sulfonamides in Pork Muscle Analysis of Chloramphenicol in Honey and Shrimp at Regulatory Levels Using Quadrupole and Ion Trap LC/MS Determination of Chloramphenicol in Fish Meat by LC-APPI-MS Determination of the Metabolites of Nitrofuran Antibacterial Drugs in Chicken Tissue by LC-ESI-MS HPLC Analysis of Antibacterial Drugs with Penicillin-Like Structure HPLC Separation of Antibacterial Drugs with Tetracycline Structure Analysis of Residual Synthetic Antibacterials in Meat by HPLC Analysis of Tetracyclines by HPLC
Contaminants Applications
Contaminants: Acrylamides GC-MS Approaches to the Analysis of Acrylamide in Foods Contaminants: Food Packaging LC-TOF-MS As a Tool to Support Can Coating/Food Interaction Studies Identification of Unknown Reaction By-Products and Contaminants in Can Coatings by LC/TOF-MS Trace Level Hydrocarbon Impurities in Ethylene and Propylene Contaminants: PCBs Direct Injection of Fish Oil for the GC-ECD Analysis of PCBs Using a Deans Switch with Backflushing Analysis of Organochlorine Pesticides and PCB Congeners with the 6890 Micro-ECD Contaminants: Toxins Determination of Aflatoxins in Food by LC/MS/MS Rapid Analysis of Crude Fungal Extracts by LC/TOF-MS Determination of Hydroxymethylfurfural in Foods using LC/MS Separation of Aflatoxins by HPLC Identification and Isolation of DSP-Toxins Using a Combined LC/MS System for Analytical and Semipreparative Work
Index
Contaminants: Toxins (continued) Identification and Characterization of New Ergot Alkaloids Analysis of Mycotoxins by HPLC with Automated Confirmation by Spectral Library GC/MS Approaches to the Analysis of Monochloropropanediol Analysis of Fumonisin Mycotoxins by LC/MS Analysis of Poisoned Food by Capillary Electrophoresis HPLC Analysis of Aflatoxines in Pistachio Nuts Using HPLC Contaminants: Trace Metals Determination of Heavy Metals in Dietary Supplements Using CRC ICP-MS Measurement of Trace Elements in Whiskey by 7500cx ICP-MS Determination of Organic and Inorganic Se Species Using HPLC-ICP-MS Unmatched Spectral Interference Removal in ICP-MS Using ORS with He Collision Mode Ion Chromatography ICP-MS for Cr Speciation in Natural Samples Comparison of the Relative Cost and Productivity of Traditional Metals Analysis versus ICP-MS in High Throughput Labs Performance Characteristics of the Agilent 7500ce - The ORS Advantage Determination of Hg in Microwave Digests of Foodstuffs by ICP-MS Determination of AsB in Fish Tissues Using HPLC-ICP-MS Comparison of GC-ICP-MS and HPLC-ICP-MS for Organotin Analysis Discover the Full Capabilities of ICP-MS in Food Safety Measurement of Macro and Trace Elements in Plant Digests Using the 7500c ICP-MS System
Natural Compounds and Additives Applications

Natural Compounds and Additives: Additives Separation of Artificial Food Colors on HC(2)/TC(2) RP Columns Analysis for Sudan Reds in Curry and Chili Powder Using LC/MS/MS Analysis of Suspected Allergens in Cosmetics Using the 7890A GC and Backflush Rapid Screening and Analysis of Components in Nonalcoholic Drinks Superior Peak Shape of Xanthines and Metabolites Separated by Eclipse Plus C18 High Throughput Separation of Xanthines Using TOF for Screening and Quantitation of Sudan Red Colorants in Foods Separation of Paraben Preservatives by Reversed-Phase HPLC Aspartame: Metabolites and Applications FD&C Colors High-Efficiency High-Speed Separation of Natural Anthocyanins Analysis of Natural Food Colorants By Electrospray and APCI LC/MS Analysis of Acidulants in White Wine Using HPLC Analysis of Antioxidants in Chewing Gum Using HPLC Analysis of Aspartame Using HPLC Analysis of Preservatives in White Wine and Salad Dressing Using HPLC Analysis of Synthetic Colors using HPLC and Diode- Array Detectionat 190950 nm Analysis of Synthetic Dyes in Food by CZE CZE Analysis of Artificial Sweeteners and Preservatives in Drinks
Index
Natural Compounds and Additives: Amino Acids High-Speed Amino Acid Analysis on 1.8 m RP Columns Measurement of Macro and Trace Elements in Plant Digests Using the 7500c ICP-MS Rapid Screening of Amino Acids in Food by CE-ESI-MS Analysis of Amino Acids in Beer using HPLC with Online Derivatization Analysis of Vanillin Extract Quality Using HPLC Natural Compounds and Additives: Carbohydrates Effect of Mobile-Phase Strength in Separations of Mono- and Disaccharides High Performance Carbohydrate Analysis Application of LC/MS to Analysis of Sugars and Sugar-Alcohol Process Control of Starch Rapid Monitoring of Carbohydrates in Food with Capillary Electrophoresis Analysis of Sugars in Foods and Beverages by HPLC with Pulsed Amperometric Detector Analysis of Carbohydrates in Lemonade Using HPLC Natural Compounds and Additives: Fats and Oils Analysis of Triglycerides by LC/MS Column Selection for the Analysis of FAMEs The HPLC Preparative Scale-Up of Soybean Phospholipids Improved Analysis of FAMEs Using RTL Methods and Databases High-Speed Separation of Parabens The Analysis of Triglycerides in Edible Oils by APCI LC/MS Analysis of Triglycerides in Olive Oil and Rape Oil Using HPLC Analysis of Unsaturated Triglycerides Using HPLC Analysis of Hydrolized Fatty Acids in Dietary Fat Using HPLC Natural Compounds and Additives: Flavors and Fragrances Comprehensive GC System Based on Flow Modulation for the 7890A GC Analysis of Food and Fragrances Using High-Efficiency Capillary GC Columns High-Resolution RP HPLC Separation of Licorice Root Extracts Complete Separation and Quantitation of Fusel Oils by Capillary GC Analysis of Essential Oil Compounds Using RTL Methods and Databases Epigallocatechin 3-O-Gallate Extract from Green Tea Flavoring Agents Aromatics Analysis of Flavonoids in Plant Extracts by CE-MS Analysis of Bitter Compounds in Orange Juice Using HPLC Natural Compounds and Additives: Miscellaneous Scale-up of Anthocyanin Separations and Re-Analysis of Collected Fractions on a Prep-C18 Column Using Pb Isotope Ratios to Distinguish between Samples of Dan-shen Plant Hormones - Rapid Gradient Elution Separation
Index
Natural Compounds and Additives: Miscellaneous (continued) Structural Determination of Ginsenosides Using MSn Analysis Analysis of Inorganic Anions, Organic Acids, Amino Acids and Carbohydrates Analysis of Catechins in Tea by HPLC with Electrochemical Detector Normal Phase Analysis of Tocopherols in Margarine Using HPLC Development of a RP HPLC Separation of Acids in Coffee Analysis of Selected Anions with HPLC and Electrochemical Detection Natural Compounds and Additives: Proteins Characterization of Transgenic Soybean Seedlines by Protein Expression with the 2100 Bioanalyzer Effect of Temperature on Separation of a Wheat Protein Stationary-Phase Selectivity Comparison: High-Molecular-Weight, Alkylated, Wheat Glutenin, Protein Subunits Natural Compounds and Additives: Vitamins Quantitative Analysis of Water-Soluble B-Vitamins in Cereal Using RRLC/MS/MS Direct Analysis of Folic Acid in Digestive Juices by LC/TOF-MS Improved and Simplified LC/APCI MS Method for the Analysis of Underivatized Free Amino Acids in Foods Separation of D2 from D3 and other Fat-Soluble Vitamins Separation of Water-Soluble Vitamins with Ion-Pairing Chromatography Separation of Water-Soluble Vitamins Using the USP 23 Method Separation of Water-Soluble Vitamins Using RP HPLC Fat Soluble Vitamins on ZORBAX XDB-C8 Separation of Retinal Isomers: Comparison of Mobile-Phase Composition Rapid Analysis of Water-Soluble Vitamins with Extraction from Cat Food Analysis of Fat Soluble Vitamins Using HPLC HPLC Analysis of Vitamins in Tablets Using HPLC Analysis of Selected Vitamins with HPLC and Electrochemical Detection
Bioanalysis Applications
Bioanalysis: Applications Rapid Wheat Varietal Identification Using the 2100 Bioanalyzer Use of the 2100 Bionalyzer for Basmati Rice Authenticity Testing Agilent 2100 Bioanalyzer Application Compendium One Platform, Endless Possibilities Determination of PCR-RFLP Profiles for Fish Species Using the 2100 Bioanalyzer Nested Multiplex PCR Determination of DNA from GM Corn and Soy Beans Using the 2100 Bioanalyzer Characterization of Transgenic Soybean Seedlines by Protein Expression with the 2100 Bioanalyzer Analysis of Genetically Modified Soya Using the 2100 Bioanalyzer Detecting GMOs with the 2100 Bioanalyzer Development of Meat Speciation Assays Using the 2100 Bioanalyzer
Productivity Tools Applications

Productivity: Applications Low-Pressure RTL with the 7890A GC Precise Time-Scaling of GC Methods Using Method Translation and RTL
Index
Productivity: Applications (continued) The 5973N-Inert MSD: Utilizing Higher Ion Source Temperatures Combined EI and CI Using a Single Source Optimizing the 6890 Series GC for High Performance MS Analysis HPLC Primer for Food Analysis Retention Time Locking: Concepts and Applications Large Volume Injection for GC Using a PTV Inlet

Food Safety Agilent

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Food Safety Agilent

Uploaded by

Copyright:

Available Formats

Solutions that meet your demands for

food safety testing

Excellent choices for food applications

Natural Compounds and Additives

Pesticides & Residues Pesticides and Herbicides

> Return to Table of Contents > Search entire document

Backflush setup in ChemStation.

Automated backflush calculations in ChemStation.

TIC and spectrum

Deconvoluted peaks and spectra

Component 2 extracted spectrum

Component 3 extracted spectrum

Library search each component to identify

Deconvolution process of three overlapped peaks.

Deconvolution software Library searching software

Deconvolution reporting software

Results and Discussion

Note: late eluters were backflushed out first

Splitless, 3x speed, 1 L injection with TID, Diazinon found

The power of deconvolution with TID for diazinon.

DRS report for ginseng with pesticides highlighted.

DRS report for peach with pesticides highlighted.

9e+07 8e+07 7e+07 6e+07 5e+07 4e+07 3e+07 2e+07 1e+07

Tetrachloro-m-xylene (ISTD) Chlorthaldimethyl

8.744 8.992 9.093 9.059 9.265 9.528 9.646 9.784

6000000 5000000 4000000 3000000 2000000 1000000

Endosulfan(alpha) 7.556 Phosmet 9.538 Captan

12.382 12.00 13.00

Phosmet FPD (P)

11.066 11.00 12.00 13.00

Scan at 12.299 min

Library spectrum Azoxystrobin

Scan at 5.615 min

Library spectrum Carbaryl

Scan at 10.776 min

Library spectrum Fenbuconazole

Scan at 8.934 min

Library spectrum Endosulfan sulfate

Phosmet (FPD, SIM)

FDA quant results

For More Information

Reducing Analysis Time Using GC/MSD and Deconvolution Reporting Software

Table1. GC Inlet Mode

System Retention Time Locked to methyl chlorpyrifos at 16.596 min

0.04 0.23 t t t t 0.09

0.02 0.48 0.03 t t t t 8 0 1 0.03 t

DRS report for celery.

Chlorothalonil Permethrin Malathion

Market basket celery total ion chromatogram.

For More Information

(End% Start%) #Column volumes

2 Position/6 Port valve

Trace enrichment autoSPE scheme.

Gradient separation of herbicides on 4.6 mm 250 mm, 5 m ZORBAX SB-C18.

mAU 350 300 250 200 150 100

700 600 500 400 300 200

0 1.25 1.5 1.75 2 2.25 2.5 2.75 3 3.25 min

High-speed gradient separation of herbicides on 2.1 80 mm, 1.8-m ZORBAX SB-C18.

For More Information

Improving Productivity and Extending Column Life with Backflush

Four chromatograms collected simultaneously from a single injection of a milk extract.

For More Information

Food Safety and Environmental

+ TIC MRM ( ) mix100_10pg_green pepper.d