j. Cosmet. Sci.

, 50, 231-248 (July/August1999)

Influenceof matrix compounds on the analysisof oxidalive hair dyesby HPLC

URSULA VINCENT, GUY BORDIN, and

ADELA R.RODRIGUEZ, European Commission, Joint Research

Centre, Institute for Rerence Materialsand Measurements, Retieseweg, B-2440 Geel,Belgium.

Accepted for publication June30, 1999.

Synopsis

Oxidativehair dyeingproducts consist of a mixtureof a broadspectrum of organic compounds including the hair dye'sso-called activecompounds and the matrix-formingcompounds. Forty-seven dye intermediatescommonly usedin cosmetic formulations havethusbeenanalyzed by RP-HPLC, and their chromatographic characteristics havebeenrecorded. Since the matrix compounds couldinterfere with the quantitative analysis of the activecompounds, eighteen matrix compounds commonly usedin cosmetic formulations have been testedfor their influenceon dye intermediatedetermination.Since some of them do affect the chromatographic behavior of the dye intermediates, an isolationprocedure, for separating matrix componentsfrom the dye-forming compounds, consisting of a liquid-liquid extraction by n-heptane, hasbeenset up. In mostcases, this procedure is effective for the extraction of the matrix products. Moreover, the dye intermediates are not extracted by n-heptane, and their chromatographic behavior is not alteredby the extraction procedure. In summary, this studyhasshown that a reference methodfor the analysis of oxidative hair dyesshouldincludea compulsory extraction stepbeforesubmission of samples to RP-HPLC.

INTRODUCTION

Among the differentmethods of changing the colorof humanhair, oxidativehair dyeing playsan important role. Formulations consist of a wide rangeof organiccompounds of two distinct types,i.e., the hair dye intermediates and the matrix-formingcompounds. Someof the hair dyes, which are availablefor use in hair dye formulations,have a toxicological or sensitizing potentialandareprohibitedor restricted in concentration by the 6th Amendmentof the European Union CouncilDirective 93/35/EEC.

To identify and quantify substances used in hair dye formulations, with the aim of implementing the EuropeanUnion CosmeticDirective, there is need for a reliable analyticalmethod. Initially, a high-performance liquid chromatography method was developed that enables the identification of a broadrangeof dye-forming compounds in standard solutions (1,2). The next step for the setupof a reliableanalyticalmethod, besides additionaloptimizationsteps,involves the investigation of possible effects of
matrix components.
231

232

JOURNAL OF COSMETIC SCIENCE

Due to the different and specificfunctionsof matrix compounds in hair dye formulations, the chemicalcomposition of matrix compounds covers a broadrange,including surfactants, pH adjusters,consistency providers,antioxidants,emulsifiers,perfumes, perfumesolubilizers, and preservatives. Eachof theseadditivesplaysa clearrole in the formulation. For instance, preservatives suchasmethylparaben or DEDM-hydantoin and
antioxidants such as BHT, sodium sulfite, and L-ascorbic acid sodium salt have to be

addedto keep the productstable.

Literature on the influence of the various matrix products on the chromatographic separation of oxidativehair dyesis very rare, althoughit is of tremendous importance. Someauthorssimply do not consider their influenceon the analysis (3-5), while others try to reducematrix interference by optimizing the measurement wavelength (6).

From the point of view of the chromatographic separation of the dyes, two main
situations can occur:

a given matrix component showsa specificretentive behaviorand can be detected

with the UV detector.

the matrix component formsa complexwith a dye intermediatethat will changethe retentive behaviorof the dye intermediate. Furthermore,it must be noted that somematrix components canalsointerferewith the column itself, i.e., adsorbonto the stationaryreversed phase,thereforechangingthe separation properties of the oxidativehair dyes. Therefore,the aims of the work describedin this article were to investigatepossible matrix influences on the chromatographic separation of the dye-formingcompounds and to setup an effective methodof separating the matrix components from the dye-forming compounds. This methodshouldnot, of course, affectthe chromatographic behavior of the dyes.Due to the high numberof dyesand matrix compounds, a selection of products had to be made accordingto lists of frequently used dye intermediates and matrix productswith their concentrations in formulations provided by COLIPA (Comit de LiaisonEuropen de l'Industrie de la Parfumerie,de Produits Cosmtiqueset de Toilette; seeAppendix).The forty-seven dyesand eighteenmatrix components selected for theseexperiments arerepresentative of four groupsof dye-forming compounds, classified according to their chemical characteristics, and of five classes of additivesusedasmatrix productsin the formulations. The influenceof matrix components on the dye determination has beenperformedon selected dyesfrom eachgroup.

EXPERIMENTAL

The chromatographic procedure has been described in detail in a previousarticle (1).

INSTRUMENTATION

All chromatographic separations were carried out using the following equipment:a two-pistonHPLC pump with a low-pressure ternarygradient systemmodule (System 325 from Kontron Instruments S.P.A., Milan, Italy), an autosampler 360 with a loopof 20 ml (Kontron InstrumentsS.P.A.), a diodearraydetector440 (Kontron Instruments S.P.A.), and a vacuum degassing system,DegasysDG 1300 (Unifiows, Japan). The

MATRIX

COMPOUNDS

AND

OXIDATIVE

HAIR

DYES

233

column temperature waskept constant by means of the thermostat of an electrochemical detector (Decade, AntecLeyden, Leiden,The Netherlands). Data processing wasdone with the Data System 450-MT2/DAD series (KontronInstruments S.P.A.).The column wasa Merck Lichrospher RP 60 Select B, 250 x 4 mm, 5-1am particlesize.In some cases, a UV-Vis spectrophotometer (Lambda 7 fromPerkinElmer)wasused additionally
with the aboveequipment.
CHEMICALS

L-Ascorbic acid sodium salt (NaAsc), 3,4-diaminobenzoic acid (3,4-daba), p-

aminophenol (4-ap),m-aminophenol (3-ap),0-aminophenol (2-ap),2,4-diaminophenol (2,4-dap), 2,4-diaminophenol HC1 (2,4-daph), 4-amino-m-cresol (4-a-3-mp),2-aminop-cresol(2-a-4-mp), 6-amino-m-cresol (2-a-5-mp), 5-amino-0-cresol (4-a-2-ht), 2-amino-5-nitrophenol (2-a-5-np),2-amino-4-nitrophenol (2-a-4-np),N,N-diethyl-maminophenol (3-deap), p-phenylenediamine (1,4-pd),m-phenylenediamine (1,3-pd),0phenylenediamine (1,2-pd), 2-nitro-p-phenylenediamine (2-n-l,4-pd), 4-nitro-0phenylenediamine (4-n-l,2-pd),2-chloro-p-phenylenediamine sulfate (2-cl-l,4-pds), Nphenyl-p-phenylenediamine (4-adp), 4,4'-diaminodiphenylamine sulfate (4,4'-dadps),
resorcinol (res),4-chlororesorcinol (chlres), 4-hexylresorcinol (hres), p-anisidine (1,4-ad), 2,6-diaminopyridine (2,6-dap),2-amino-3-hydroxypyridine (2-a-3-hp),2-methylresorcinol (2,6-dht), toluene-2,4-diamine (2,4-dat), toluene-3,4-diamine (3,4-dat), toluene2,5-diamine sulfate (2,5-dats), 1-naphtol (1-nap),2-naphtol (2-nap),1,6-naphtalenediol (1,6-dhnap),2,3-naphtalenediol (2,3-dhnap),2,7-naphtalenediol (2,7-dhnap),phloroglucinol (phlg), pyrogallol (pg), hydroquinone (hq), pyrocatechol (pc), p-

methylaminophenol sulfate(met), 2-methoxy-p-phenylenediamine sulfate(2,5-das), 3-tert-butyl-p-hydroxyanisole (3-tb-4-ha),4-chloroaniline (4-cla),3-methyl-l-phenyl-2pyrazoline-5-one (3-m-l-p-2-p-5-o)were obtained fromFluka. p-Phenylenediamine sulfate(1,4-pds), andm-phenylenediamine sulfate (1,3-pds) werekindlyprovided by "Les
ColorantsWackherr," Saint-Ouen l'Aum6ne, France. Sodium tetraboratedecahydrate

(p.a.),acetic acid95% (suprapure), ammonia 25% (suprapure), hydrochloric acid(0.1

M), oleicacid(OA) (p.a.),n-heptane (p.a.),isopropanol (p.a.),sodium sulfite(p.a.)(SS), andpolyvinylpyrrolidon (PVP) wereobtained fromMerck.BHT (butylated hydroxytoluene), methylparaben (MP), andlaurylsulfate (LS)wereobtained fromSigma. Lauric

diethanolamide (LDA), TEA-dodecylbenzenesulfonate (TDS), and Syntopon 8 D1 (ethoxylated octylphenol-EOP) wereprovided by Witco S.A. Triethanolamine (TEA)
was obtainedfrom Mobi-Lab bvba, Zutendaal,Belgium. DEDM-hydantoin (DMDM),

dimethicone copolyol (DC), nonoxynol-12 (NOL), and polyquaternium-11 (PQ) were kindly providedby "Keuringsdienst van Waren," Enschede, The Netherlands. n-

Nonylamine 98%(NNO) was obtained from Janssen Chimica, andOranex HT (ORA) wasprovided by Spinnrad GmbH, Gelsenkirchen, Germany. Cetrimonium chloride (CC)andmethanol (HPLCquality) were obtained fromFluka. Pure water (18.2Ml/cm
quality)usedfor the preparation of solutions wasobtained from a MilliQ Plus 185
system(Millipore, Molsheim, France).
PREPARATION OF REAGENTS

Solvents and chromatographic mobile phase

Buffersolution pH 8 (Soerensen buffer):440 ml hydrochloric acid(0.1 N) and 2 g/1

234

JOURNAL OF COSMETIC SCIENCE

L-ascorbic acid sodiumsalt (NaAsc)as an antioxidantagentwere addedto a 560-ml

sodium teraboratesolution. The solvents were mixtures of MeOH and Soerensen buffer, pH 8 in variousproportions.

Mobile phase (aqueous phase B): a 0.05 M aceticacid solutionwasadjusted to a pH of 5.9 with a 10% ammonia solution andfilteredthrougha 0.45-lm filter. When not in use,the eluentwasstoredat a temperature of 4C to preventmicrobiological growth.
Samples. The matrix productsand the dye intermediates used in the experiments are given in Tables I and II, respectively. All dye intermediates and matrix component concentrations were selected accordingto COLIPA data.

Stocksolutions of dye intermediates at a concentration of 0.1 g/100 g wereprepared in mixturesof various proportions of MeOH and Soerensen buffer(containing NaAscasan antioxidant agent), ranging from40% to 90% Soerensen buffer.Sample solutions of dye intermediates at a concentration of 0.025 g/100 g wereprepared by dilution from the stocksolutions in mixturesof Soerensen bufferand MeOH rangingfrom 30% to 60%
Soerensen buffer.

Stock solutionsof matrix components were preparedin MeOH- or isopropanolSoerensen buffermixtures of different proportions. Sample solutions of the matrixcomponentsat variousconcentrations (Table I) were preparedby dilution from the stock solutions in MeOH- or isopropanol-Soerensen buffermixtures (abbreviations aregiven
in the Chemicals section).
Table I

Selected Matrix Compounds, Functions, Concentrations, and MeOH Proportions Used

MeOH

Concentration

proportion
(%)

Compound
Oleic acid (OA)

Function

(g/100 g)

Surfactant, cleansing agent

5
3

100
100

Unknown Oranex HT (ORA) TEA-dodecylbenzenesulfonate Surfactant, deansingagent

0.5 0.5 2
0.25 6

90
90
75

(TDS) Cetrimonium chloride (CC) Lauric diethanolamide (LDA)

BHT

Surfactant, emulsifying agent Surfactant, deansingagent

Antioxidant Perfume solubilizer

70 70

Syntopon 8 D1 (EOP)
Nonoxynol-12 (NOL) n-Nonylamine 98% (NNO) Sodium lauryl sulfate(LS) Polyquaternium-11 (PQ)

Surfactant, emulsifying agent

Unknown

3
25

70

Surfactant, cleansing agent,denaturant Antistaticagent,film former,hair fixative

Preservative

3 2
0.05

60 60 60 60
60 40 40
0

Methylparaben (MP) Polyvinylpyrrolidon (PVP)

Sodium sulfite (SS) Triethanolamine (TEA)

Antistaticagent,film former,hair fixative Antioxidant,reducingagent pH adjuster

Preservative

2 2 1.5
0.05

DEDM-Hydantoin (DEDM) Dimethiconecopolyol(DC)

L-ascorbic acid sodium salt

Emollient, hair conditioning agent Antioxidant,pH adjuster

2.5 **

(NaAsc)

* DC wasprepared in an isopropanol-Soerensen buffermixturewith 67% isopropanol. ** The Soerensen buffer contains NaAsc 2 g/l.

MATRIX

COMPOUNDS

AND

OXIDATIVE

HAIR

DYES

235

Table

II

RetentionTimes for Forty-Seven Dye Intermediates, Each of Which Was Measuredas a Single ComponentSolution(n = 3)
Retention time (rain)

Dye

Mean

RSD (%)

3,4-Diaminobenzoicacid

2.75

1.54

2,4-Diaminophenol 2,4-Diaminophenol HC1 p-Phenylenediamine sulfate Phloroglucinol p-Phenylenediamine Pyrogallol 2-Amino-3-hydroxypyridine Hydroquinone p-Aminophenol m-Phenylenediamine sulfate m-Phenylenediamine 2,6-Diaminopyridine m-Aminophenol
Toluene-2,5-diamine sulfate Resorcinol 4-Amino-m-cresol

3.95 3.98 5.21 5.23 5.60 5.90 5.91 6.02 6.13 7.00 7.15 7.91 8.04
8.50 9.23 9.70

0.17 0.18 0.54 12.56 9.34 3.59 0.84 6.92 3.46 3.23 6.42 3.93 3.43
7.73 4.67 1.82

0-Phenylenediamine 2-Methoxy-p-phenylenediamine sulfate 0-Aminophenol Pyrocatechol p-Methylaminophenol sulfate 2-Methylresorcinol 2-Chloro-p-phenylenediamine sulfate Toluene-2,4-diamine

9.86 9.94 10.51 11.23 11.88 11.89 12.33 12.78

3.87 1.35 5.98 18.06 2.02 6.84 2.01 8.85

2-Nitro-p-p henylenediamine 4-Nitro-0-phenylenediamine 2-Amino-4-nitrophenol

5-Amino-o-cresol

13.14 15.10 15.50

15.66

2.20 2.53 3.01

4.11

Toluene-3,4-diamine
6-Amino-m-cresol

17.53
17.95

7.50
1.34

2-Amino-p-cresol p-Anisidine 2-Amino5-nitrophenol

4-Chlororesorcinol

18.07 18.47 18.66

20.44

2.90 4.21 3.14

3.70

4,4'-Diaminodiphenylamine sulfate 1,6-Naphtalenediol

2,7-Naphtalenediol 3-Methyl-1-phenyl-2-pyrazoline-5-one
4-Chloroaniline

20.83 25.23
25.81 26.01
26.82

0.75 2.24
1.59 6.39
2.50

2,3-Naphtalenediol N- Phenyl-p-phenylenediamine 2-Naphtol N, N-Diethyl-m-aminophenol 1-Naphtol 3-tert-butyl-4hydroxyanisole 4-Hexylresorcinol

27.88 29.56 29.71 30.00 30.01 31.85 32.26

2.38 3.64 0.93 0.02 0.42 0.58 0.55

236

JOURNAL OF COSMETIC SCIENCE

Samplesolutions of dye intermediates at a concentration of 0.025 g/100 g in matrix media (at the differentconcentrations given above)wereprepared by dilution from the respective stocksolutions at variousproportions of MeOH and Soerensen. It hasto be notedthat all samples contained NaAscasan antioxidantagentaddedto the Soerensen buffer to ensurethe stability of the dye samples.
PROCEDURES

Reversed-phase HPLC conditions. A non-linearMeOH (A)/aqueous phase (B) gradientwas

usedasfollows: 0-25% A for 19 rain, 25-80% A for 10 rain, 80% A for 5 rain, 80-95% A for 5 rain, 95% A for 10 rain, and 95-0% A for 3 min. The total flow was ! ml/min.

Betweenthe injections, the columnwasequilibrated by a 25-ml mobile phase. Each analysis wasrepeated five times.The columntemperature waskept at 48C. The data acquisition wascarriedout at two or threeselected wavelengths: 220 nm, 235 nm, and 290 nm, in parallel with the spectra acquisition.
Isolation of the matrix components from thefinal analytesolution. First, experiments using anion exchange, solid phaseextraction,liquid cation exchange, and liquid-liquid extraction were carried out. The resultsled to the selectionand the optimization of a liquid-liquid extraction procedure of the matrix components from the sample solution by n-heptane. Two milliliters of samplesolution(single-matrixproduct solutionor mixture of dye intermediates and a matrix compound, at the concentrations given in the Experimental section) were treatedwith n-heptane.Dependingon the matrix product,the extraction involvedoneto threesteps. In the first step,the extraction wasperformed on the 2-ml sample using20 ml n-heptane, the two phases wereseparated, andthe resulting aqueous phase1 wasthereafter submittedto HPLC or analyzed by UV-Vis spectrophotometry. For additionalextractionsteps,20 ml n-heptanewere addedto the resultingaqueous phase n-l, separation of the two phases wasperformed, and the resultingaqueous phase n wassubmittedto HPLC or analyzed by UV-Vis spectrophotometry.

RESULTS

AND

DISCUSSION

A systematic studywascarried out on eachof the eighteen selected matrix compounds and forty-seven selected hair dyes(TablesI and II). First, their individual retentive behaviorwas determinedand their UV spectrumrecorded, using the DAD or the UV-Vis spectrophotometer. Second, the efficiency of the liquid-liquid extraction protocolon the selected matrix compounds waschecked. When a matrix compound could not be extracted from the sample solution, its influence on the retentive behavior of the
dye intermediates was investigated.
CHROMATOGRAPHIC STUDY

Single solutions of dye-forming compounds. Forty-seven dye-forming compounds weresubjected to chromatography and their retentiontimes and UV spectra were recorded.
Resultsare presented in Table II.

It comes out that 91% (43 dyeintermediates) of the dyeintermediates tested havetheir

MATRIX

COMPOUNDS

AND

OXIDATIVE

HAIR

DYES

237

retention time between5 min and 30 min. Someof the tested dyes have very close retentiontimes--4-n-l,2-pd, 2-a-4-np, and 4-a-2-ht for instance, their retentiontimes being 15.10 min, 15.50 min, and 15.66 rain, respectively. However,sincetheir respective UV spectra showremarkable differences (8), their identification canstill be carried out easily.

In orderto test the repeatability, five injections of eachsamplewere carriedout. The

relative standard deviation (RSD) for the retention time was between 0.2% and 1.1%,

while that of the peakcorrected areas (definedasthe peakareadividedby the retention time), recordedat 290 nm, was between 1.1% and 6.0% (9). The chromatographic behaviorshowedgood repeatability. The injections were repeated on three differentcolumns in orderto test the reproducibility (Table II). For thirty-six hair dye intermediates, the relativestandard deviation (RSD) on the retentiontime waslessthan 5%; for nine dye intermediates, this RSD was between6% and 10% and more than 10% for only two dyes.The reproducibility betweencolumnswas considered satisfactory.

Single solutions of matrixcompounds. Eight differentcompounds were subjected to chromatographic separation, andtheirretention timesandUV spectra wererecorded. Results arepresented in Table III. It appears that someof the testedcompounds havevery close
retention times--BHT, OA, and ORA, for instance.However, with their respective UV

spectra alsoshowing remarkable differences (Figure 1), their identification canstill be carried out easily.It must alsobe notedthat mostof the matrix compounds, with the exceptionof DMDM, PVP and MP, have retention times greater than 30 minutes, meaningthat confusing dye intermediates with matrix products shouldnot occursince the retention timesof the hair dyes generally rangefrom 5 to 30 minutes(1). Concerning thefourdyeintermediates forwhichretention timesareoutof therange of 5 to 30 min, the discrimination from the matrix compounds canalsobe easily madeaccording to both
retention time and UV spectrum.

In orderto test the repeatability, five injectionsof eachsamplewere carriedout. The

relative standard deviation (RSD) for the retention time was between 0.02% and 0.7%,

while that of the peakcorrected areas wasbetween 0.3% and 3.7%, except for DMDM. The chromatographic behavior of the matrix compounds showed goodrepeatability.
Table III

RetentionTimes and Corrected Peak Areasfor Eight Matrix Compounds, Each in a Single
Solution (n = 5) Retention time (min)

Corrected peakarea(AU)

Matrix compound
DMDM PVP MP NNO EOP

Mean
6.80 7.10 26.87

RSD (%)
0.21 0.24 0.06 0.70 0.08

Mean
4.3 1.1 1.0 5.2 5.8

RSD (%)
12.4 0.2 1.8 3.4 0.3

k of measurement (rim)
220 220 290 235 290

33.80
35.85

OA
BHT ORA

37.97
38.11 38.35

0.30
0.02 0.09

20.6
2.5 12.2

2.9
3.7 0.5

235
290 235

AU: arbitrary unit.

238

JOURNAL OF COSMETIC SCIENCE

Butylated Hydroxytoluene

1000-500

-s22|
200

I
300
400

Wavelength (nm)
Figure 1. UV spectra of three matrix products.

Sample solutions containing an additional matrixcompound and three dyeintermediates. From the forty-seven dyeintermediates tested, fourwerechosen for furtherinvestigation based on how representative they wereof their differentclasses of hair dyesaswell ason their
use in formulations.

Three sample solutions wereprepared, eachcontaining a mixture of threeselected dye intermediates (1,4-pd or 4-ap, res, and 2-n-l,4-pd), the intrinsicmatrix compound
NaAsc asan antioxidant,and anothermajor matrix compound, BHT, DMDM, or EOP, respectively. BHT is an importantantioxidant, DMDM a commonly usedpreservative,

and EOP a perfumesolubilizer. The retentiontime of BHT and EOP is morethan 30 min, while DMDM is eluted within about 7 min. Three controlsolutions, mixturesof the respective dyeswithout the major matrix compound, were alsoprepared. All the components were at the concentrations used in formulations according to COLIPA, and five injections of eachtype of solutionwere made.For eachseries of
measurementsof control and mixture, a new column was used, as it has been shown in

a preliminary studythat after50 injections, the efficiency of the columnis dramatically decreased (data not reported). Table IV givesthe statistical resultsobtained for the retentiontimesand the corrected peakareas for eachof thesecompounds.
Comparison of the data betweenmixture ! (BHT) and control 1, mixture 2 (DMDM) and control2, andfinally mixture 3 (EOP) and control3 leadsto several observations. First of all, as shownby the results,DMDM or EOP do not interferewith the dye intermediates or with the column in a way that would changethe chromatographic

MATRIX

COMPOUNDS

AND

OXIDATIVE

HAIR

DYES

239

Table

IV

RetentionTimes and Corrected PeakAreasfor Three Matrix Compounds and Four Dye Intermediates in
D,e-Matrix Mixtures and in Control Solutions (n = 5)

Retentiontime (min)
Column Mixture Compound Mean RSD (%)

Corrected peak area(AU)

Mean RSD (%)

4-ap
res

5.59
9.31

0.85
1.14

11.5
0.3

7.2
3.1

2-n-l,4-pd
NaAsc

13.75
2.18

0.53
0.38

9.1
54.7

21
7.8

1
Control 1

BHT
4-ap
res

38.07
5.57
9.55

0.55
0.40
0.82

2.5
10.6
0.3

19.0
5.9
5.0

2-n-l,4-pd
NaAsc

13.89
2.18

0.50
0.21

10.0
56.5

6.5
8.0

1,4-pd
res

5.36
8.32

0.39
0.34

14.6
0.5

1.6
2.9

2-n-l,4-pd
NaAsc

12.91
2.35

0.55
0.60

14.4
70.3

1.1
2.3

2
..........

DMDM

6.72

0.42

5.4

15.1

1,4-pd
res
Control 2

5.33
8.28

0.19
0.53

15.0
0.5

2.0
1.3

2-n-l,4-pd
NaAsc

13.28
2.35

3.39
0

14.4
80.7

3.8
1.5

4-ap
res

5.99
9.53

0.39
0.79

22.1
0.6

1.7
1.5

2-n-l,4-pd
NaAsc

14.09
2.27

0.39
0.25

20.4
35.7

1.6
5.5

EOP

36.05

0.71

6.4

0.6

4-ap
res
Control 3

6.32
10.26

0.88
1.99

22.7
0.6

2.0
5.9

2-n-l,4-pd
NaAsc

14.84
2.24

1.42
0.24

20.3
53.2

1.6
6.3

behavior(retention time, corrected peak area)of the dye intermediates. Furthermore, they do not alter the repeatability of the separation. For instance, when DMDM is added to the sample, the RSD on the retentiontime of the various hair dyesvariesfrom 0.3%

240

JOURNAL OF COSMETIC SCIENCE

to 0.5% and that of the corrected peakareas stays belowthe limit of 5%. For DMDM, this RSD value is of the order of what was obtained when DMDM was analyzed separately (seeTable III). Nevertheless, DMDM has a retention time greater than 5 minutes and could (potentially) be confusedwith a dye intermediate (such as pphenylenediamine, for instance). However,DMDM, which shows two chrornatographic peaksat 220 nm with exactlythe sameUV spectrum,doesnot showany absorbance signalat any wavelength greaterthan 250 nm. Thus, a selective detectionof the dye intermediates can be cacriedout at a wavelengthgreater than 250 nm.
For EOP, TablesIII and IV showthat EOP hasa retention time much greaterthan 30 minutes,and thuscannotbe confused with anydye intermediate. Furthermore, the peak retentiontimesandthe peakareas of the dyeintermediates arenot significantly changed by the matrix compound regardingeither the meanvaluesor the RSD.

Nevertheless, sinceEOP doesnot interferewith the hair dye separation in theseconditions,an additionalparameter, i.e., the concentration, hasbeenchecked. A broadrange
of concentrations of EOP were tested around the concentration used in formulations.

Five RP-HPLC analyses were carriedout for eachconcentration of EOP in a three-dye intermediatesolutioncontainingNaAsc asan antioxidant.Figure 2 shows the evolution of the mean corrected areaof the peaksobtainedfor the three dye intermediates with increasingconcentrationof EOP in the sample. The graphs clearly show that the corrected areasdo not vary significantly(p -- 0.05) when EOP has beenaddedto the samplesolutionover a concentration rangeof 2 g/100 g-10 g/100 g. The statistical analysis of the results (datanot reported) shows goodrepeatability of the analysis for each
selected concentration of EOP. The RSD is less than 5% for the retention time and the

corrected peak area. Concerning BHT, the results showthat the presence of this matrix compound doesnot

25t

:0
10

't

-,

--

T
,,a.

..

2-n-l,4-pd

*T 4-ap

10

Concentrationof EOP (g/100g)

Figure 2. Influence of the concentration of EOP on the corrected peakareas of 4-ap, res,and 2-n-l,4-pd. (Seetext for experimental data and abbreviations.)

MATRIX

COMPOUNDS

AND

OXIDATIVE

HAIR

DYES

241

affecteither the retentiontimes of the dyesor their RSD, nor the mean valuesof the corrected peak area.Nevertheless, it is importantto note that the repeatability of the measurements for one component (2-n-l,4-pd) seems to be stronglyaffectedby the presence of BHT, with the RSD increasing from 6.4% to 21%. BHT itself hasa high RSD for its peakcorrected area. Regardingtheseresults,it appears, quite clearly,that although the matrix compound EOP doesnot interferewith the dye intermediates or with the column in a way that would affect the chromatographic behaviorof the dye intermediates, it would be of advantage to remove BHT andDMDM from the sample solution before proceeding with the separation in orderto obtain an accurate separation of the dye intermediates.
OPTIMIZATION OF THE LIQUID-LIQUID EXTRACTION METHOD

Sinceseveralmatrix compounds are presentin a real formulation,it is of immense interestto determinethe effectof the isolationmethod on eachof them. The optimizationof the extraction protocol leads to a one-step (LDA, TDS, NNO, ORA), two-step (OA, BHT, DMDM, NaAsc), or three-step(CC, DC) extractionby n-heptane,while EOP, NOL, LS, SS,TEA, PQ, MP, andPVP arenot extracted. Two differentapproaches were considered for testing the extractionmethod. BHT and EOP were submitted to extractionin a mixture containingthe dye intermediates as well, whereas in a second approach, the extraction wasperformed on singlesolutions of matrix compounds.
Extraction of sample solutions containing a matrix compound and three dyeintermediates. The extraction protocolhasto be efficientto extractthe matrix compounds, of course, but the target dyesshouldnot be extracted.A one- to three-stepextractionprotocolwas thus appliedto sample solutions containing threeselected dye intermediates (4-ap, res,and 2-n-l,4-pd), the antioxidant NaAsc,and a major matrix product(BHT or EOP). The aqueous phases obtained weresubmittedto HPLC, and the extraction yield wascalculated according to the followingequation:

yield (%) =100 - k, (correcte------- corrected ;re--efor-' area after extraction extractio---- X100 ) (1)
Table V givesthe extraction yield for eachcomponent in the solution,and the successive chromatograms obtainedfor the solutioncontainingBHT before(a) and after (b) the extractionprocedure are shownin Figure 3. First of all, it is clearly seenthat the extractionprocedure affectsneither the chromatographic characteristics of the dye inTable V

Statistical ExtractionYield of the Components of a Dye-EOP and a Dye-BHT Mixture

Compound One-stepextractionyield (%) Two-stepextractionyield (%)

4-ap
Resorcinol

0
0

0
0

2-n- 1,4-pd
NaAsc
EOP --

0
19

0
50
0

BHT

89

100 (n.p.)

n.p.: no peak.

242

JOURNAL OF COSMETIC SCIENCE

550,

250__

--

-60

Retentiontime (rain)

550

I 250---

-60

I
0.0 15.0 30.0 49.0

Retentiontime (rain)
Figure3. Successive chromatograms of a dyeintermediate mixture with BHT added to thesample before andafterextraction. a: Before extraction. b: Aftera two-step extraction. Peaks: (1) NaAsc (intrinsic matrix, antioxidant); (2) 4-ap;(3) res;(4) 2-n-l,4-pd; and (5) BHT (matrix,antioxidant). Peaks 2, 3, and 4 are unchanged between steps a and b. (Seetext for experimental dataandabbreviations.)

termediates (retention times, corrected peakareas of peaks 2, 3, and4) northe repeatability of the analysis. Then, after a two-step extraction by n-heptane, 50% of the antioxidant NaAscis extracted while 100% of BHT (peak 5) is extracted from the sample. However, it appears that EOP is not at all extracted by n-heptane. Additional experimental workhasshown thatEOPwas100%extractable by othertested organic solvents (trichloromethane, dichloromethane, or diethylicether,for instance) but that theyalsoleadto the partialextraction of the phenolic dyes (datanot reported). This extraction procedure by n-heptane is therefore veryeffective for the separation of

MATRIX

COMPOUNDS

AND

OXIDATIVE

HAIR

DYES

243

BHT from the sampleand partially effectivein the caseof NaAsc, beforethe HPLC analysis. Moreover,it hasto be stressed that with this extraction procedure by n-heptane, the non-extractability of the dye intermediates is ensured (9).

Extraction ofsingle-compound matrixsolutions. Due to the factthat the dyeintermediates are not extractedby n-heptane,the extractionprotocolwas then performedon singlecompound matrix solutions. The extraction protocolwas appliedto 16 individual solutions of matrix compounds, namely OA, NNO, ORA, DMDM, BHT, PVP, MP, LDA, TDS, DC, NOL, SS, LS, TEA, PQ, and CC (seetheir function in Table I). The resultingaqueous phases obtainedwere submittedto HPLC or analyzedby UV-Vis
spectrophotometry when necessary.

The first sevencompounds (OA, NNO, ORA, DMDM, BHT, PVP, and MP) show chromatographic peaksfor which the chromatographic characteristics aregiven in Table III. For theseproducts, the extraction yield wascalculated according to equation1. For the lastnine compounds, i.e., LDA, TDS, DC, NOL, SS,LS,TEA, PQ, andCC, the final analysis could not be performedby RP-HPLC for several reasons: because LDA, NOL, andTDS aresurfactants, the submission of these compounds to RP-HPLC leadsto their adsorption in the system,thus provokingfurther contamination, and the compounds CC, LS, SS, TEA, DC and PQ do not showany chromatographic peak when submitted to RP-HPLC. As an alternative,a UV-Vis spectrophotometer was then usedto test the extraction procedure on thesenine compounds, by recording the spectra obtainedbefore and after extraction.In orderto determineif the UV-Vis spectrophotometer was sensitive enough for the extraction measurements, BHT, which has been shown to be

extracted by n-heptane (seeabove), wasusedfor comparing results obtained usingboth

procedures.

Evaluation of the extraction yieldusingRP-HPLC.Figure4 shows the successive chromatograms obtainedbeforeand after extraction of four compounds (OA, NNO, ORA, DMDM). Table VI gives the extractionyield obtained. It must be noted that the efficiency of the extraction protocol variesslightly,depending on the matrix product. OA needs a two-stepextraction procedure to be fully extracted (Figure4, la-c), while the peaks of NNO (Figure4, 2a,b)andORA (Figure4, 3a,b)disappear aftera one-step
extractionprocedure. Moreover,evena two-stepprocedure doesnot lead to a complete

extraction of DMDM (Figure4, 4a-c) and NaAsc.Finally, PVP and MP are not at all
extracted.

The extraction protocol hastherefore beenshown to be efficientfor isolating OA, NNO, and ORA from the sample.Concerning DMDM, the extractionyield is about 80% and MP is not extracted. Nevertheless, by analogywith EOP (Figure 2), a similar study involving the concentration as an additional parameterleads to the conclusion that DMDM andMP do not interfere with the dyeintermediates or with the columnin a way that would affectthe chromatographic behavior of the dye intermediates (resultsnot shown).Concerning PVP, special attention shouldbe given to real samples containing this matrixproduct, asit formsa film on the columnsurface duringthe first injection, leadingto a modification of the retentiontimesof the dye intermediates. The determinationof the dyeintermediates hastherefore to be carriedout mainly according to their
spectrum(resultsnot shown).

Evaluation of the extraction yield usingUV-Vis spectrophotometry. The extractionyields

244

JOURNAL OF COSMETIC SCIENCE

1300

OA

la

16u

NNO I

2a

900 -

600 -

3oo-

-75

I
0.0

I
30.0

I
45.0 57.0

-25

15.0

0.0

15.0

30.0

44.0

Retention time (min)

1300

Retention time (min)

160

113
9oo_

213

.1oo
E

600-

,.

300-

-75 oo

I
15.0

I
30.0

I
45.0 57.0

-25
o.o

I
15.o

I
30 o 44.0

Retention time (min)

1300

Retention time (rain)

lc

900-

g600-

" 300-

o.o

15.o

30.0

45.0

57 o

Retention time (rain)

Figure 4. Successive chromatograms of four matrix productsamples beforeand after extraction.1: OA; 2: NNO; 3: ORA; 4: DMDM. a: Beforeextraction.b: After a one-step extraction.c: After a two-step extraction.(Seetext for experimental data and abbreviations.)

obtained are reportedin Table VI. The completeextractionof BHT in a two-step extraction procedure wasconfirmed, and the extraction protocolhasalsobeenshown to be efficientfor isolatingLDA, TDS, CC, and DC from the sample.Nevertheless, once

MATRIX

COMPOUNDS

AND

OXIDATIVE

HAIR

DYES

245

1300

ORA

3_a

DMDM

4a

6OO300_

-75
0.0

-50

I
0.0

I
15 0

I
22.5 30 0

150

30.0

45.0

57 0

7.5

Retention time (min)

1300

Retention time (mln)

3b

4b

9OO-

600-

300-

-75

I
15.0

-50

I
0.0

I
150

I
22.5 30.0

0.0

30.0

45 0

57.0

7.5

Retentiontime (min)
100

Retention time (min)

4_qc

50
0.0

7.5

15.0

22.5

30.0

Retentiontime (min) Figure 4. Continued

more,it hasto be noted that the efficiency of the extraction protocol varies slightly, depending on the matrixproduct. LDA andTDS needa one-step extraction procedure to befullyextracted, whilethecomplete extraction of CC andDC involves a three-step
extraction procedure. Moreover, NOL, LS, SS,TEA, and PQ are not at all extracted.

246

JOURNAL OF COSMETIC SCIENCE

Table VI

StatisticalExtractionYield of EighteenMatrix Compounds

Evaluation

using

Compound
OA NNO ORA
DMDM

Extractionyield (%)
100 100 100
80

Extractionstep number

H
P
L

(n.p.) (n.p.) (n.p.)

NaAsc
EOP
MP PVP

50
0
0 0

BHT LDA
U V / V
I

100 100

(n.p.)

TDS
CC DC
LS SS

100
100 100
0 0

NOL

PQ
TEA

0
0

n.p.: no peak. UV/VIS: UV-Vis spectrophotometry.

CONCLUSION

The chromatographic methodpreviously set up for the determination of hair dye intermediates hasbeenshown to be extremely effective for the analysis of forty-seven hair dye intermediates commonly usedin cosmetic formulations. Using that method,a data base of the retention time andof the spectrum hasbeenbuilt up for the forty-seven dyes, which allowsthe determination of the dyesand their discrimination. In mostcases, the retentiontimes of the hair dyesrangefrom 5 to 30 minutes.Thus problems with confusing dye intermediates with matrix products, whichgenerallyhaveretentiontimes of lessthan 5 minutesor greaterthan 30 minutes,shouldnot be encountered. Once more, the UV spectrumconstitutes a powerfulcriterion of discrimination when the retention time of a particular dye or a particular matrix compoundis outsidethe
expectedrange.

The influence of eighteen matrix products commonly usedin hair dyeformulations has alsobeeninvestigated in relationto the determination of hair dye intermediates. Nine of the eighteenmatrix compound showretentivebehavior and UV-detectable peaks. Their individualretentiontimes and UV spectra havetherefore alsobeenrecorded and includedin the database with the aim of providingascomplete a pictureaspossible of the separation of hair dyesin complexmatrix media.The nine matrix compounds left overcannotbe identifiedunderthesechromatographic conditions, but their UV spectra havealsobeenrecorded in orderto complete the data base. However,on the whole,the presence of some matrix compounds seems to affectthe final accuracy of the quantitative measurements of some hair dyes (caseof 2-n-l,4-pd), leading to the conclusion that a separation of the matrix compounds from the dye

MATRIX

COMPOUNDS

AND OXIDATIVE

HAIR

DYES

247

solutions is desirable. A liquid-liquid extraction procedure by n-heptane hastherefore been setup. Thismethod wastested onall theeighteen selected matrixproducts andwas 100% effective for eightproducts (LDA, TDS, CC, DC, NNO, ORA, OA, BHT) in a one-to three-step extraction procedure. DMDM andNaAscwereonlypartiallyextracted (80% and 50%, respectively), while EOP, MP, PVP, LS, SS,NOL, PQ and TEA were
not extracted.Nevertheless, the non-interference of these compounds with the dye intermediates or with the column,hasbeenproved,exceptin the case of PVP.

A three-step liquid-liquid extraction procedure by n-heptane, whichis not toxicto the environment, is therefore suggested asa compulsory stepfor the separation of matrix products from the dye-containing sample before analysis by reversed-phase HPLC to avoid all problems of potential interference between thedyes andthematrixcompounds or betweenthe matrix compounds and the columnin a way that would affect the chromatographic determination of thedyes. Thisprocedure will nowbeapplied to actual
commercial formulations.

As a generalconclusion, it has beenshownthat after applyinga matrix extraction procedure usingn-heptane, the RP-HPLC method,combined with diodearraydetection,is verypromising in terms of beingan efficient candidate reference method for the
identification and quantitationof oxidativehair dyes.
NOTE

A spectra' database of eighteen matrixproducts hasbeenrecorded throughthe diode array detector andtheUV spectrophotometer aswellasa spectra' database offorty-seven dye intermediates (8). For further informationabout the database, email contact:
vincent@irmm.jrc.be
ACKNOWLEDGMENTS

We would like to thank COLIPA for providingraw materialsand information.We would alsolike to thank Dr. E. Pel for performinginitial studies on separation procedures.This work was carriedout in the frame of support to DG XXlV "Consumer Policy"and DG III "Industry" of the European Commission.
APPENDIX

List of SomeFrequently Used Matrix Products Matrix component

Ammonia
Ascorbic acid L-ascorbic acid sodium salt BHT

Usualconcentration (g/100 g)
6.00-9.00
0.20 0.20 0.25

Diethyleneglycolmonoether
Citric acid Diethanolamine

5.00
0.30 2.00

Triethanolamine

0.10-1.50

TEA-dodecylbenzenesulfonate DEDM-hydantoin

0.50 0.10

248

JOURNAL OF COSMETIC SCIENCE

Appendix (continued)

Syntopon 8 D1
Ethoxylatednonylphenol Ethyl acetate Hydroxyethylcellulose Isopropanol
Lauric diethanolamide

6.00
3.00 2.00 2.40 3.00-15.00
1.50-2.00

Methyl paraben Propyleneglycol Sodiumlauryl sulfate

Sulfated castor oil

0.05 5.00-9.00 2.95

4.00

Polyvinylpyrrolidon
Oleic acid

2.00
5.00

n-Nonylamine
Oranex HT
Cetrimonium chloride

25.00
3.00
0.50

REFERENCES

(1) E. Pel, G. Bordin,and A. R. Rodrfguez, HPLC candidate reference methodfor oxidative hair dye analysis. I. Separation and stability testing,J. Liq. Chrom. Re/. TechnoL, 21(6), 883-901 (1998). (2) E. Pel,Application ofHighPerformane LiquidChromatography (HPLC) totheStudy ofOxidative Hair Dyes (Internal annualreport, GE/R/ACH/03/95, 1995), pp. 1-19. (3) D. H. LiemandJ. Rooselaar, HPLC of oxidation haircolours, Mitt. Gebiete Lebensm. Hyg.,72, 164-176
(1981).

(4) N. Goetz,J. Mavro, L. Bouleau, and A. de Labbey,"Analysis of OxidativeHair DyesUsing High Performance Liquid Chromatography," in Cosmetic Analysis: Selective Methods and Techniques, P. Bord, Ed. (MarcelDekker, New York, 1985), pp. 245-287. (5) C. Gennaro,P. L. Bertolo,and E. Marengo,Determinationof aromatic aminesat tracelevelsby ion interaction reagentreversed-phase high-performance liquid chromatography--Analysis of hair dyes and otherwater-soluble dyes, J. Chromatogr., 518, 149-156 (1990). (6) V. Andrisano, R. Gotti, A.M. di Pietra,and V. Cavrini,Analysis of basic hair dyesby HPLC with on-line post-column photochemical derivatisation, Chromatographia, 39, 138-145 (1994). (7) C. Genova,A. Zatta, L. Deiana, F. Montesion,F. Buosi, and G. Gazzaniga,"EthoxylatedMonoBranched PrimaryAlcohols in Hair Dyes,"in Cosmetics 1997 ConjUrerice Proceedings: Modern Challenges to theCosmetic Formulation, 1997, pp. 359-377. (8) U. Vincent,Practitwl Aspects oftheDetermination of Oxidative Hair Dyes byRP-HPLC--Development ofa Data Base (Annualreport,GE/R/ACH/04/99, 1999), pp. 1-24. (9) U. Vincent,Evaluation of High Performance LiquidChromatography (HPLC) in Relation to theStudy of Oxidative Hair Dyes (Internalannualreport,GE/R/ACH/13/97, 1997) pp. 1-26.