In Class Problems: Atomic Spectroscopy

I use a lecture format to introduce general background information and aspects of flame and furnace atomization that are discussed in sections 6.1-6.2B of the text before asking them the following question.

1. What are the relative advantages and disadvantages of using a flame or furnace as an atomization source?

I prompt them by asking them to think back on the types of things we have usually considered with analytical methods to assess their utility. Eventually we generate a list of things like sensitivity, reproducibility, matrix effects and sample size.

It is obvious to them that graphite furnace requires less sample size. They are often at a loss to address the issue of sensitivity. I might ask some leading questions:

Do both methods use all of the sample that has been introduced?

Compare how long the atoms spend in the light path in the two techniques?

Is there a difference between a method that integrates the signal versus one that measures a steady state value?

They are often at a loss to address the question of reproducibility and matrix effects, and we discuss these as a group.

Before examining the next question, I give a brief lecture on the use of the cold vapor method for mercury and the generation of volatile hydrides for elements such as arsenic and selenium.

I then give a lecture on the use of an inductively coupled plasma for atomization (section 6.2D of the text) as well a very brief coverage of the use of arc and spark devices.

2. If you were to run an analysis using an atomic absorption spectrophotometer, you would note that a separate source lamp called a hollow cathode lamp is needed for each individual element that you wish to measure. For example, a lead lamp emits the specific lines of light that are absorbed by lead. Why is the cathode designed with a hollow configuration?

Before giving them this question, I give a brief overview of the design and operation of a hollow cathode lamp.

Groups can usually suggest the two reasons why it the lamp uses a hollow cathode. If not, I might ask the following:

What problem do you see if you had a flat cathode (Note, I draw a picture on the board)?

What ultimately happens to the gas phase atoms that are sputtered off the surface and what would you prefer happen with them?

They are never able to answer question 3 below without a series of prompting questions that I include in the problem set I provide to them. I introduce question 3 and 4 below at the same time.

3. Needing a different lamp for each element is expensive and not as simple as using a continuum source with a monochromator. Why is it apparently not feasible to use a broadband continuum source with a monochromator when performing atomic absorption spectroscopy?
4. One thing you might consider is whether continuum lamps have enough power in the part of the electromagnetic spectrum absorbed by elements. In what part of the electromagnetic spectrum do most atoms absorb (or emit) light?

The groups usually have no problem determining that this is in the UV/Vis portion of the spectrum.

5. Do powerful enough continuum sources exist in this region of the electromagnetic spectrum?

I remind them of UV/Vis and fluorescence spectroscopy and ask whether those methods used continuum sources and whether we were concerned about their power. They realize that powerful enough continuum sources do exist so that is not the reason for using hollow cathode lamps.

6. A more helpful thing to consider is the width of an atomic line. What are the two major contributions to the broadening of atomic lines? (Hint: We went over these earlier in the course).

They can usually come up with collisional and Doppler broadening, although some have to go back through their notes to find these.

7. When these contributions to line broadening are considered, the width of an atomic line is observed to be in the range of 0.002-0.005 nm. Using information about the width of an atomic line, explain why a continuum source will not be suitable for measuring atomic absorption.

They are initially stumped by this question, so I ask them the following:

How did we select a monochromatic wavelength from a continuum source when recording absorbance with a UV/Vis spectrophotometer?

They can relate this back to the use of a grating and slit.

What was the typical bandpass from such a device?

They usually realize that is was about 1 nm. I then draw the output of a continuum source on the board – y-axis being intensity and x-axis the wavelength showing a width of about 1 nm – and ask them to copy this into their notebook and superimpose on it the atomic line that would be absorbed. This is usually sufficient to get them to realize that the atomic absorbance will only remove a small fraction of the power from such a source, hence the reason for using a hollow cathode lamp.

What is the problem with reducing the slit width of the monochromator to get a narrower line?

From prior material in the course, they realize that the power will now be too low and the noise too high.

8. Why does the hollow cathode lamp have a low pressure instead of a high pressure of argon filler gas?

If they are stumped with this, I ask them the following:

What would happen to gas phase atoms sputtered from the hollow cathode if there was more argon in the lamp?

They can usually determine that more collisions would occur.

Based on prior things we learned in the course with other spectroscopic methods, what will these collisions do to the emission from the sputtered atoms?

They can usually determine that it will broaden the lamp output, and based on the prior discussion, realize that this creates the same problem as a continuum source with a monochromator.

9. Flame noise (either emission from the flame or changes in the flame background as a sample is introduced) presents a significant interference in atomic methods. Can you design a feature that could be incorporated into an atomic absorption spectrophotometer to account for flame noise?

I usually set this up that they work at an instrument company selling AAs. A customer has called and identified this problem and asks if they can provide a solution to it. I am usually pleasantly surprised at how some of the groups can hone in on key things to consider with this problem. Some propose measuring the flame noise before introducing sample, but then I draw on the board a representation of flame noise as some positive signal above zero and indicate that its magnitude may change up or down as a sample different from distilled water is added. Some realize that they need a way to measure the flame signal in the absence of the hollow cathode lamp and someone usually proposes the idea of somehow blocking the source beam. We have talked about the use of pulsed sources in fluorescence spectroscopy so some suggest that we could do this as well. We then discuss the possibility of using a chopper.

10. Can you design to feature that could be incorporated into an atomic absorption spectrophotometer than can be used to account for both scattered light and light absorbed by molecular species?

I usually draw Figure 6.13 on the board showing an atomic absorption line superimposed over a molecular absorption background. The groups can usually realize that they ought to be able to do some sort of background correction by measuring the absorbance near and at the atomic line and then subtracting the two. Some even ask whether it would be possible to have a second line, and I use this to mention the possibility of using an instrument that exploits the Zeeman Effect but put off a full discussion of that.

Thinking back to our discussion of sources for atomic absorption, can you propose a method that would allow you to measure molecular absorption while not having any significant contribution from atomic absorption?

With this prompt, the groups can usually reflect back on the reason why the use of a continuum lamp was a problem and a hollow cathode lamp was used instead.

Can a continuum source be used to measure molecular absorption?

Based on our unit on UV/Vis spectroscopy, they immediately answer this. We can then discuss the concept of background correction using a deuterium lamp. I remind them of the reason why hollow cathode lamps have a low pressure of Argon gas – which usually prompts them to ask whether it would somehow be possible to pulse the pressure in the lamp – which then serves as a way to introduce the Smith-Hieftje method.

11. Metal complexes with low volatility are often difficult to analyze when performing atomic absorption measurements because the atomization efficiency is reduced to unacceptably low levels. Can you devise a strategy or strategies for eliminating the problem of a non-volatile metal complex?

I indicate that there are three different strategies that they might be able to come up with and let them start discussing the question. If they are completely stumped, I remind them that earlier we had discussed a specialized technique for analyzing arsenic. The groups can usually come up with the ideas of using a protecting agent.

If needed, I might ask whether they can propose a way to eliminate the effect of an undesirable ligand that forms a non-volatile complex.

With this they can come up with the idea of adding a releasing agent. Groups also often suggest raising the temperature of the flame as another possibility.

12. Can you devise a strategy to overcome unwanted ionization of the analyte?

We have talked about how the high amount of argon ions in an ICP helps suppress ionization, so they have that as background. Also, since the discussion on non-volatile complexes promoted the idea of adding things to a sample, members within groups often suggest the idea of adding something that easily ionizes.

13. Devise a general method that can be used to account for the presence of unknown matrix effects.

I remind them of what we mean by a matrix effect, and that matrix effects can either enhance the signal or suppress the signal of the analyte. I also remind them that when we use a standard curve, the standards are usually prepared in distilled water, which might have a substantially different matrix than the sample. Therefore, we somehow need a way to determine whether the matrix is enhancing or suppressing the signal of the analyte.

With this background, usually someone comes up with the idea of adding extra analyte to the sample. When they do, I pose the following question:

How should the amount of additional increment compare to the initial concentration in the sample?

They usually realize that this should be relatively small so that it does not swamp out the matrix.

I then draw a set of axes on the board and show the response for the initial sample and indicate where we will plot the response for samples with additional increments of analyte. I then pose the following:

Draw the curves that would result for two samples with identical concentrations but one has a matrix that enhances the signals and the other a matrix that suppresses the signal.

Groups can usually determine the correct form of the curve and we can then discuss any remaining aspects about standard addition and how one can extrapolate the curve back to get the concentration of the species in the sample.