Chem 223: Experiment 8

Spectrophotometry, Beer's Law, and Precision

(This experiment is completely based on the lab. Spectrophotometry, Beer's Law, and
Precision. Chem. 223- UIUC)

SAFETY FIRST! Remember to bring AND WEAR your goggles. "But we aren't doing any
reactions this week, and the chemicals are largely innocuous." "They're hot and steam up."
"They're uncomfortable." "I forgot." "I look like I came from Mars." We've heard 'em all.
Doesn't matter. You'll be in a lab. Get to the point that it's a reflex:

BRING AND WEAR YOUR SAFETY GOGGLES.

Simple way to tell if you've been careful this week: if you don't have blue fingers, you did
fine. (Methylene blue is used to stain tissue during surgery, so it's safe, just ugly.)

Purpose

Analytical Chemistry is the science that identifies the components of a portion of the physical
world, quantifies them, and characterizes their interactions. Doing this requires:

• A problem whose solution requires knowledge of chemical composition or dynamics.

• One or more means of measurement
• An understanding of the chemistry and physics that effects the measurement
• A means to characterize the quality of the measurement (both precision and accuracy)
• A theoretical or practical way to use the results of the measurement to solve a problem.

Our purpose here is to demonstrate a common means of measurement (spectrophotometry), to

show examples where the measurement tool works, to show examples where the measurement
tool fails, and to show examples where the chemistry interferes with the measurement. All
this will be done under circumstances where the operator's manual dexterity is of minimal or
no influence on the results.

In later labs, your skill will contribute to the precision of the experiment. Many past students
have felt that imprecision was some mark of shame or that "skill is proportional to 1/standard
deviation." There are times when this is true, but in this experiment you will see differences in
uncertainty caused by chemistry, sample, and instrument. What you should learn is:

• Non-zero uncertainty is a fundamental property of nature.

• Many factors contribute to uncertainty.
• Understanding the sources of uncertainty can lead to their minimization.
• Learning analytical chemistry means learning the "tools of the trade" PLUS developing "lab
technique" PLUS learning how to integrate many pieces of information into the solution of a
single problem.
Important Advice

Bring a thumb drive to lab. Save ALL your data files both to the hard disk (during the
experiment, immediately after obtaining raw data) and to your thumb drive BEFORE leaving
lab. The computers are NOT on the network; you can NOT get at your data (without which
you can't write your report) unless you remember to save your data!

Background
One of the hardest parts of doing analytical chemistry is that we must deal with the world as a
SYSTEM (many things interacting and happening at once), yet we have been taught to deal
with learning and living as COMPONENTS (single activities, single calculations, individual
controls or pieces of apparatus).

Think of driving a motor bike. You don't think about each spark plug firing or turning the
wheel 2.78º to the right or pushing on the throttle to get an additional 10-3 Newtons to
accelerate. You Just Drive. You are taking a systems approach, with a lot of activity implicit
in what you do. But when you learned to drive, you DID pay attention to pressure on the gas
or the angle at which you steered. A mechanic pays careful attention to the details of e.g.
timing. Similarly, when you're doing error propagation or taking a spectrum, you're dealing
with a component. The big picture is that there's some problem involving a chemical system
you're trying to solve. For this lab, the components connect like this:

We'll discuss the theory of the components here, you'll make measurements and do
computations, and then in the questions at the end, we'll tie it all together in a systematic way.
 Fig. 1: UV-1601 PC Photomultiplier Spectrophotometer

I. Beer's Law

See Harris, Seventh Edition, Chapter 18, or similar material in any other quantitative
analysis text written since the Second World War).

Suppose we have light of a wavelength λ (that's a Greek lambda; you can find it as the lower-
case l in the symbol font on nearly any computer) such that a molecule can absorb the light. If
a homogeneous solution of the analyte is illuminated by a source containing light of
wavelength λ, only a fraction of the light will be transmitted through the solution i.e.
illuminating the solution with light of intensity I0 (Watts cm-2) results in intensity I<I0 being
transmitted. Only the transmitted light can reach the detector.

Without worrying about the electronics and optics involved, we can say that we MEASURE
Transmittance T = I/I0. If there are no molecules to absorb the light, T = 1. If the solution
contains some absorbing molecules, 0<T<1.

The negative base 10 logarithm of transmittance is called absorbance, A = - log10T. A is a

computed quantity, not a directly measured quantity. Beer's Law says that, absent
complications (some of which we'll discuss),

A = εbC (1)
where A = absorbance
ε = molar absorptivity in liter mole-1 cm-1
b = path length over which absorbance is observed (typically 1 cm in a square cuvette,
but it can range from 1 nm for a molecular monolayer to light-years in astronomy)
C = concentration of the analyte in mole/L

This is a deceptively simple result. If one knows the molar absorptivity and path length, an
absorbance measurement directly gives concentration. Experimentally, if one has a standard
solution of known concentration Cs, then measuring its absorbance As should give the product
εb. The problem is that absorbances are additive. If there is any hint of absorbance from a
background substance, then A =(εbC)analyte + Abackground. This behavior is commonly referred to
as a linear working curve, since a plot of A vs C is just a straight line, with slope εb and
intercept Abackground. Such working curves are the most common relationships used in doing
concentration measurement in chemical analysis. Almost always, one measures BOTH slope
AND intercept. Without measuring that Abackground = 0, don't assume it or you're likely to get
yourself into a bind. The reason we're doing spectrophotometry as the first lab is to get used to
the idea that there are such things as working curves. Note that to find the slope and intercept,
you need a set of standards. This sets up a problem: how do you know the composition of the
standards and the source of absorbance in the blank? We'll answer those questions later in the
course!

II.Deviations from Beer's Law

There are many situations where there appear to be deviations from Beer's Law. In most
cases, it's not Beer's Law that's the problem – it's that one or more assumptions that went into
Beer's Law don't apply.
For example, if there are many substances that absorb at the same wavelength,

A    bC (2)
i i i
where the index i refers to different species in solution. For example, what happens if
substances can form new compounds or complexes at high concentrations? Suppose a
compound M can pair with another molecule of M to form M2.

M+M ⇔ M2
We have no way to know the molar absorptivity of M2. What we do know is that there has to
be conservation of matter, so if CM is the total concentration of M prepared initially,
CM = [M] + 2[M2]. We also know from the law of mass action that

(3)

In this lab, we use the well-known example of methylene blue which, in water, forms such
dimer adducts. Since we have two unknowns, the concentrations of the monomeric and
dimeric species, we will need measurements at two wavelengths to figure out the molar
absorptivities and the equilibrium constant. At low concentrations, practically no dimer will
form. Furthermore, we can make solutions in which we know CM, the total concentration of
methylene blue moieties in solution (some of which will be present as monomer and some as
dimer).

(4)

Note carefully that the molar absorptivity of the dimer, M2, at wavelength λ is different than
that of the monomer, because it is a different molecule. Substituting the equilibrium constant
into Equation 4, we obtain

(5)
What we know experimentally is CM (that's what we can make up with a balance and a
volumetric flask – NOT [M]). So to express the dependence of absorbance on an
experimentally-known variable, we use CM = [M] + 2[M2] = [M] + 2K[M]2

This is a quadratic equation in the variable [M], which has the solution:

(6)

Substituting this into Equation 5 yields

(7)

That formula should bother you. If you go back to the original expression relating CM and
[M], setting K=0 gives (unsurprisingly) [M] = CM. But in the full formula, CM shows up under
a square root sign! How do we bail out of this problem? Using the nearly universal solution to
messy math problems: turn the expression into a Taylor series and look for limiting cases.

(Recall "h.o.t." doesn't mean what you say after biting into habanero peppers, but rather
"higher order terms"). Demonstration: 1.11/2 ≈ 1.04880, but 1 + 0.1/2 = 1.05, which isn't too
far off. So if K is small,

(If you think this is reasonable, you're happy with the primary mathematical trick used in
physical chemistry!)

An equation such as Equation 7 is true, but awkward. Dealing with it now is getting ahead of
ourselves; it will be the middle of the course before you have enough experience with
equilibria, Excel, and data reduction to finish processing the data. For right now, it's enough to
know that this is one example of how Beer's Law can work, but a working curve can appear to
be nonlinear.

Why else might working curves be non-linear? Two reasons are other chemical equilibria
involving the measured species, e.g. a protonoation/deprotonation reaction or instrumental
effects. An example of the latter is spectrometer resolution, which can be a problem. If the
absorption band is narrow compared to the spectrometer resolution, even a very concentrated
solution of the analyte won't absorb all the light before reaching the detector. Here's a sketch.
In the case of a very narrow spectral profile, it doesn't matter how high the molar absorptivity
is; almost all the light will reach the detector. It's just like putting a single piece of black
thread in front of a bay window – the black thread doesn't transmit any light, but the room
isn't darkened. If we make a Beer's Law plot for the ideal situation (wide absorbance spectrum
compared to the resolution of the instrument) we get the straight line in the following plot,
while the narrow spectrum above (line width less than 2 nm, while the instrument resolution
is 2 nm) gives the dashed line. At low concentrations, the slope of the Beer's Law plot for the
narrow line is 0.84 that of the broad line plot. At high concentration, the slope has fallen to
0.78. If you use a ruler on the plot, you can see that the shallower-sloped line is curved.

III. Error Propagation

See Harris, Chapter 3, Section 5 and Appendix C for general error propagation formulas.
Alternatively, see Appendix II of this lab manual. The error illustrated at the end of section II
above is an example of determinate error. It results from a systematic problem with the
measurement itself, namely, inadequate spectrometer resolution. It could be completely
eliminated by the expedient of acquiring the spectrum with another spectrometer with smaller
spectral resolution. There is another kind of error, indeterminate error, which is random, and
can never be completely eliminated from your measurements. This section addresses the
question: If I know the uncertainties in all but one quantity involved in an experiment, what
is the uncertainty in the remaining quantity? Error/fluctuation/distribution/variance/variability
of a quantity on successive measurements are fundamental characteristics of nature and useful
characteristics we can study.

Suppose that we are interested in a measured value Z, which is a function of other variables,
Z = f(A, B, …) then

(8)

where δZ represents the uncertainty in Z and similarly for A and B. Since all random errors, if
averaged, have a zero mean, just averaging this equation would reduce to 0=0, which is not
very helpful. So instead, we square both sides, THEN average (because the square of real
numbers is positive or 0, never negative).
(9)

The function's variance, δZ2, is either positive or zero provided the δAδB averages to 0, as it
usually does. We can average the variance and take the square root to obtain the root mean
square error, <δZ2>1/2, or the
relative root mean square error, < δZ2>1/2/Z, where the angle brackets, < >, denote that an
average has been taken. These are more commonly known, respectively, as the standard
deviation and the relative standard deviation.

IV. Sources of Random Error in the Spectrophotometer

A potential source of systematic error is small differences of baseline between different

spectra. In order to minimize baseline errors it is preferable that neither sample nor reference
cell should be moved between measurements of spectra. In practice this means either using a
flow-cell or a fibre-optic probe or building a titration cell for a particular spectrophotometer.
If measurements are to be made in alkaline solutions then the necessity of excluding
atmospheric CO2 indicates that a closed titration system must be used.

Baseline error is also affected by whether the spectrophotometer is a single- or double-beam

device. Instruments based on diode-array detectors are usually single-beam devices, so the
background has to be measured on the same cell as the sample spectra.

The type of random error depends principally on the nature of the detector. Older instruments
used a photomultiplier detector. The error of this type of detector increases with the intensity
of the light falling on it. In Hyperquad there is a module for determining an absorbance error
function which is based on the use of repeated scans of a standard spectrum.

In the instance shown below the data were obtained with an instrument that has a
photomultiplier detector, using a holmium glass filter as sample. There is a general increase in
error as absorbance increases but the trend is irregular because of correlation of errors
between absorbance values. They appear to belong to different sets according to whether
absorbance is increasing or decreasing.
Modern instruments use semiconductor detectors such as a diode array or charge-coupled
detector (CCD). The error associated with these detectors tends to be constant and
independent of wavelength. Here is an example from a diode-array spectrometer, also using
the holmium filter. The error is virtually constant.

What sources of random error could there possibly be from a high-class diode array
spectrometer such as the HP8452A? Here are some possibilities.
• Digitization noise. Going from the raw diode charge signal to a number involves a device
called an analog-to-digital converter. It has only 216 = 65536 different numbers it can
record. Thus, there's a limit to how finely we can measure raw intensity.
• Dark noise. Even when no light hits the photodiodes, there's an uncertainty in the signal
level.
• Dark current. Some charge leaks off the diode as a function of time.
• Readout pattern noise. Because of the way the diode array integrated circuit is made,
alternate diodes have different readout characteristics.
• Light source flicker. Fluctuations in the light source.
• Shot noise. For N discrete, random events such as detecting a photon, the uncertainty is
N1/2.
• Sample cell positioning error. The cell never seats exactly the same way twice, so the path
length
 and refraction give slightly shifted values each time the cuvette is reseated..
• Wavelength drift from temperature changes in the spectrometer.
• Density changes in the sample due to temperature changes.
There are additional factors, but this is a sufficient start. Notice that none of these say
anything about changes in the sample! Any random variation in the sample adds to the noise.

Before You Come to Lab

Read the ENTIRE lab. BRING YOUR SAFETY GLASSES. Wearing safety glasses in lab will be
strictly enforced. Bring a floppy disk or thumb drive so you can copy your data.

Procedure
I. Using the UV 1601 PC Shimadzu Spectrophotometer

I.A. Turning On and Warming Up the Instrument.

When you come to lab, the computer and instrument will likely be turned on. If not, push the
power switch on the lower left side of the Shimadzu UV-1601 PC spectrophotometer to turn
on the instrument. Turn the spectrometer on BEFORE turning the computer on. Turn the
spectrometer OFF at the end of the lab. It takes approximately 20 minutes for the
spectrometers to warm up to the point where their output is stable.

I.B. Using the Software.

The computers are not on the network – just turn them on, when you get to the sign-on screen,
leave the user as "Administrator" and the password blank. On the Windows desktop, double –
click the UVProbe icon, followed by OK button. The UV Probe software will load and the
spectrum window will appear.

The TAs will provide additional instruction from here. For a preview, see the tutorial on the
software menus hot-linked in the on-line lab book

I.C. Handling the Glassware

Think of the walls of the cuvette as windows. If you set 100% T when the windows are clean,
what happens if the windows become dirty? Less light gets through,. Fingerprints can't be
differentiated from anything else that causes absorbance. It is CRITICAL that cuvettes be
scrupulously clean. For quartz cuvettes, one typically wipes the cuvette with a lintless wipe
(Kimwipes® are a sufficiently common brand that "wiping with a Kimwipe" has become the
pseudogeneric that "make a Xerox copy" has become for photocopying) just before inserting a
cuvette into the spectrophotometer. However, for plastic cuvettes such as you will use, wiping
tends to scratch the outside surface, making things worse than if you hadn't done anything at
all. So when you remove a cuvette from the box,

• Only touch the top at first.

 • Locate the frosted sides – henceforth only touch those frosted sides or the very top, far
from where light will go through the cuvette.
• Load the cuvette with solution, buffer, or water, being careful not to overfill.
• Insert the cuvette into the spectrometer with the cuvette clamp lever in the UP position.
Make sure the light goes through the clear sides, not the frosted sides, of the cuvette.
• Lower the lever to lock the cuvette in place.

NEVER mouth pipette. EVER. Suction to fill the pipette must ALWAYS be provided by a
mechanical device. That way, if solution squirts out the top of the pipette, you may
contaminate the pipette bulb but you won't hurt yourself. DO NOT THROW PIPETTES
INTO THE TRASH CANS. There are separate receptacles for glass such as disposable
pipettes. Plastic cuvettes can go in the trash cans after they've been rinsed.

II.Precision

II.A. Cuvette Transparency.

While the plastic cuvettes you'll be using aren't terribly expensive, they don't transmit light at
wavelengths shorter than 300 nm. Shorter wavelengths can be observed using quartz cuvettes.
These can run over $100 apiece and should be treated accordingly. To see how transparent an
empty cuvette is,
• Blank the instrument with only air in the optical path. That is, make sure there's
nothing in the cuvette holder. Put the lever DOWN so that the optical path is
unobstructed. Take a blank spectrum.
• Raise the lever, put the clean, dry cuvette into the holder, push it into place, lock it by
lowering the lever.
• Take a spectrum. In your lab report, describe what you see. You should report apparent
absorbance vs. wavelength and describe how much light gets through the cuvette in
the UV, visible, and near IR regions of the spectrum. You can tell by eye that the
cuvette is clear in the visible.

Why does the spectrometer indicate that less than 100% of the light gets through in this
wavelength region?

• Squirt some distilled water into the cuvette (if any slops over the side, you'll have to
get another cuvette and go back to step 2) You only need to have enough to make the
top of the liquid be above the observation axis hole). Take another spectrum.

What has changed? Why?

• Usually, one takes a blank spectrum with buffer or solvent in the cuvette. With water
still in the cuvette, re-blank the instrument.

Explain what is going on in the UV region of the spectrum. Describe physically what is
happening and explain the implications for making measurements on substances with
absorbance only in the wavelength range 200-300 nm.

II. B. Reproducibility of Absorbance Measurements

 • Remeasure 100%T to see how reproducible this setting is.
• Put blank 30% methanol in a cuvette and measure absorbance for this combination vs.
water only. Does the methanol absorb? (note: methanol is poisonous. Do not drink,
avoid breathing fumes)
• Reblank system using the blank solution and cuvette.
• Take the cuvette out and put it back in.

Is there a measureable change? Why or why not?

There is an ambiguity in that last step. To find out if the change is due to moving the cuvette
or the precision of the absorbance measurement itself:
• Leave the cuvette in place and measure the absorbance of the buffer 15 times at 500
nm. Compute the mean and standard deviation of these measurements. In a perfect
world, you would get A = 0.0000 with a standard deviation of 0.0000. You won't.
• Make another 15 measurements, but this time lift the lever, remove the cuvette, put it
back into the holder, lower the lever, and make the absorbance measurement. Again,
compute mean and standard deviation.

IN YOUR LAB REPORT: are the mean readings with and without physically removing and
replacing the cuvette statistically different? See Harris, 7th edition, section 4-3 on Student's t.
The question here is identical in character to Case 3, P. 62. Are the standard deviations
different? Does removing and reseating the cuvette reduce experimental precision? From what
you have seen here about the precision of measuring A=0 and what you saw for wavelengths
less than 300 nm for the empty cuvette, suggest the range of absorbances one can reliably
measure with these spectrophotometers.

III.Working Curve for Crystal Violet

III.A. Make the working curve.

You are provided with standard solutions of crystal violet, otherwise known as Gentian
Violet, in DI H2O. Concentrations are given in terms of the mass of crystal violet per mL
solvent in parts per million, ppm, i.e. μg/mL. We define concentration on a weight-per-
volume basis so we don't need to assume the density of the solution is 1 g/mL (in fact, we
don't care about density at all).

In any order you like (as long as you carefully record the concentration to which each
spectrum corresponds), obtain and store spectra for the following solutions:
Blank – 3 replicate measurements
1.70 μg/mL = 1.70 ppm crystal violet
3.40 ppm crystal violet
5.70 ppm crystal violet
7.60 ppm crystal violet
8.60 ppm crystal violet

III.B. Baseline correction

 - Open the spectrophotometer sample- chamber cove. Make sure that neither the
sample beam nor the reference beam is obstructed, and that there is no sample
in the sample compartment.
- Add 2 mL of the solvent (methanol 30%) to a quart cuvette and place the
cuvette in the sample compartment. (Caution: Quartz cuvettes are very
expensive! Only touch the frosted, not the clear surfaces. When preparing a
solution for analysis, be careful not to contaminate the sample nor the
equipment with acetone, which has a large absorbance maximum at 268 nm).
Close the sample- chamber cover.
- Click the connect button, followed by the baseline button. The baseline
parameter window will appear. Enter the appropriate wavelength limits. Then
click the OK button to acquire a background spectrum.

III.C. Method preparation

* Spectrum method:
- Select the File > Open menu item. Them choose the desires method ( e.g.
crystal_violet.smd) and click the Open button. The spectrum Method window
will appear
- Set the following parameters in the measurement subwindow
a. Wavelenght Range: Choose the appropriate wavelength limits.
b. Scan speed: medium
c. Sampling Interval: 1.0
d. Scan mode: Single.
- Click on the Instrument parameters tab. Then choose the Absorbance
Measuring mode. Click the OK button to send the method parameters to the
instrument.
 Photometric method ( see detail in appendix)
- Select the File > Open menu item. Them choose the photometric method and
click the Open button. The photometric Method window will appear
- Set the parameters in the measurement subwindow.

III.D. Sample preparation

- Open the sample- chamber cover and remove the cuvette. Empty the cuvette.
Then use a solution of the sample dissolved in the solvent to rinse the cuvette
twice. Use about ½ mL for each rinse.
- Once the cuvette is rinsed, add appropriately 2 mL of the sample solution and
place the cuvette in the sample compartment. Close the sample-chamber cover.

IV. Data Acquisition

IV. A. For recording spectrum:
- Click the overlay tab the place the spectrum in Overlay mode. Click the minimum
absorbance value on the Y- axis and set it to 0. then click the maximum absorbance
value on the y- axis and set it to 2. (Note: If the acquired spectrum do not display
properly on the graph, right- click the mouse and choose the auto Scale function to
adjust the graph’s coordinate.)
Click the Start button to begin the scan. Watch the spectrum trace in the graph pane to
see that the absorbance values do not exceed the upper limit during the acquisition.
The new Data Set window will appear the scan is complete. Enter an appropriate File
name.
-Click the Print button to print the spectrum
- open the sample- chamber cover, remove the sample from the instrument and close
the sample- chamber cover.
- Clean the cuvette thoroughly (without acetone).

IVB. For calibration method

- Click to read the standard and view a calibration curve

- Click to read unknown samples

-Clean the cuvette thoroughly (without acetone). and return it to the instructor

IN YOUR REPORT: Identify the wavelength of maximum absorbance. Use this wavelength
to plot a working curve (absorbance as a function of concentration). Find the slope and
intercept of the working curve using least-squares regression procedure given Section VI.
Find the standard deviation of the slope and standard deviation of the intercept. What's going
on at wavelengths shorter than 300 nm? What if you chose a wavelength other than 588 nm
for your working curve? Would it affect accuracy? Precision?
What if you averaged absorbance over a range of wavelengths – would that affect accuracy?
Precision?

You will notice "funny" regions in the spectrum near 656 nm and near 486 nm. Recall that the
light source in the spectrometer is a deuterium lamp. Atomic hydrogen (whether H, D, or T)
emits and absorbs light in the "Balmer series" that includes Hα (656 nm) and Hβ (486 nm). For
details, see the discussion of the Bohr atom in your general chemistry text.

From your spectrum, find the absorbance at the wavelength you used for making your
working curve and, from the equation you fitted to your working curve data, compute the
concentration and uncertainty in the concentration of the unknown.

V. Chemical and Spectral Interference

• Recheck 100% T
• Take a fresh spectrum of crystal violet
• Take a spectrum of crystal violet containing 12.5 ppm Bromophenol Blue

IN YOUR REPORT explain what the presence of Bromophenol Blue does to the crystal violet
spectrum and what effect this would have on a quantitative analysis. Estimate the smallest
concentration of crystal violet which could be reliably determined in the presence of 12.5 ppm
Bromophenol Blue.

IMPORTANT INSIGHT: methods for determination only work if there are no interferences.
All the sample preparation and separation are for the purpose of providing a clean sample,
free of interferences, so that accurate determination is possible.

VI. Methylene Blue

Not only can methylene blue form dimers, but also trimers. In non-aqueous solvents, there's
an anhydrous form that has a different spectrum than the aquated species. We've chosen a
solvent where the anhydrous form doesn't appear, 30% methanol. In this section of the lab,
you'll see how the shape of a spectrum changes with concentration.
• Blank the spectrophotometer again using 30% methanol
• Record spectra for all the concentrations of methylene blue in water/30% methanol
provided. You will have a set of about 15 spectra.
• Make working curves from the data at 648 nm, 656 nm, and 664 nm

Methylene Blue

IN YOUR REPORT: Is the curvature you see due to the instrument? How do you know? If it's
not due to the instrument, what might cause the curvature? If you wanted to use the
spectrometer to assay methylene blue, what wavelength would you choose and why? What's
the lowest concentration at which you get an absorbance that is unambiguously greater than
that of just 30% methanol/water? That's your detection limit. What's the highest concentration
you can measure? The difference between the high and low concentrations is the WORKING
RANGE of the method. The RATIO of the highest usefully measurable concentration to the
detection limit is the DYNAMIC RANGE. Report this number.

If there is curvature in a working curve that can not be attributed to instrumental behavior,
then either the methylene blue is reacting or complexing with the solvent or the methylene
blue is reacting with itself to form dimers or higher complexes. For now, you should conclude
which circumstances lead to ideal methylene blue behavior (Beer's Law is followed and the
working curve is straight) and the circumstances which lead to Beer's Law not being followed
(working curve is non-linear).

Questions to Answer at the End of Your Report

• For some of the experiments, you set 0 absorbance/100% transmittance using the same
cuvette in which you made the sample measurement. In others, you set the baseline
using a different cuvette. Does this make a difference in accuracy? In precision?
• In every case, we told you (or hinted strongly) what wavelength to use. What
characteristic did you notice that led us to choose those wavelengths?
• When there was an interference because another compound absorbed at the chosen
wavelength, how would the computed concentration for an unknown change if you
didn't account for the interference, i.e., if you made a working curve with the
interference absent, but unknown to you the interference was present in the unknown, so
simply used the original working curve? Would you overestimate, underestimate, or
correctly estimate the concentration of the unknown? Explain.
 • How can you tell if a supposedly linear working curve really is linear? Use your own
data and appropriate curve fitting to provide examples for your discussion.

Things we didn't talk about but you might worry about in the future

• You didn't make up any of the solutions. You'll learn to do this in the next lab or two.
• You didn't prepare the samples starting from solids, so you didn't weigh anything, nor
did you have to deal with complete dissolution, filtering, pH adjustment, or any other
wet chemistry.
• The operation of the spectrometer wasn't explained very much. You'll learn about this in
lecture.
• Noise sources in the spectrometer weren't discussed in detail. The best textbook on
spectrometer noise is J. D. Ingle and S. R. Crouch, Spectrochemical Analysis. There are
newer journal articles (see particularly the work of Joel Tellinghuisen), but Ingle and
Crouch is the gold standard.
• We ignored systematic errors that can occur if the spectrometer's resolution is poor
compared to the width of the spectral features being observed. In the Real World, you
can't do that, or the working curve may be highly nonlinear.
• We assumed that light could only be transmitted or absorbed. It can also be emitted,
reflected, or scattered, and each of these can be used either to characterize samples or
can interfere with sample characterization. Imagine trying to do an absorbance
measurement in fog – the fog would act as a scatterer, not an absorber (since water is
transparent in the visible, but you know you can't see through a fog).
• When is spectrophotometry useful for solving problems, and when is some other tool
preferable? When can it be used alone, and when must it be combined with other
techniques? Until we've studied more techniques, these questions can't be answered.

Additional References
 Beer's Law: Harris, Seventh Edition, Chapter 18, or similar material in any other
quantitative analysis text written since the Second World War)
 Harris 7th Ed., Chapter 4, Sections 7-9 for the derivation of regression formulas
 Harris 7th Ed., Chapter 3, Sections 4-5 and Appendix C for general error propagation
formulas
 E. B. Sandell, Colorimetric Determination of Traces of Metals, 3rd Ed., Interscience
(New York, 1959).
 J. O. Rawlings, Applied Regression Analysis, Wadsworth and Brooks/Cole (Belmont,
CA, 1988).
 Z. Zhao and E. R. Malinowski, "Determination of the Hydration of Methylene Blue
Aggregates and Their
 Dissociation Constants Using Visible Spectroscopy," Appl. Spectrosc. 53, 1567-1574
(1999).
 B. D.McCullough and Berry Wilson, "On the Accuracy of Statistical Procedures in
Microsoft Excel 97,"
 Computational Statistics and Data Analysis, 31, 27-37 (1999).