Quantitative Applications - Atomic Absorption Spectroscopy

Quantitative Applications

Atomic absorption using either flame or electrothermal atomization is widely used for the analysis of trace metals in a variety of sample matrices. Using the atomic ab- sorption analysis for zinc as an example, procedures have been developed for its de- termination in samples as diverse as water and wastewater, air, blood, urine, muscle tissue, hair, milk, breakfast cereals, shampoos, alloys, industrial plating baths, gaso- line, oil, sediments, and rocks.

Developing a quantitative atomic absorption method requires several consider- ations, including choosing a method of atomization, selecting the wavelength and slit width, preparing the sample for analysis, minimizing spectral and chemical in- terferences, and selecting a method of standardization. Each of these topics is con- sidered in this section.

Flame Versus Electrothermal Atomization 

The choice of atomization method is determined primarily by the analyte’s concentration in the samples being analyzed. Because of its greater sensitivity, detection limits for most elements are significantly lower when using electrothermal atomization (Table 10.10). A better precision when using flame atomization makes it the method of choice when the analyte’s concentration is significantly greater than the detection limit for flame atomization. In addition, flame atomization is subject to fewer interferences, allows for a greater throughput of samples, and requires less expertise from the operator. Electrother- mal atomization is the method of choice when the analyte’s concentration is lower than the detection limit for flame atomization. Electrothermal atomization is also useful when the volume of sample is limited.

Selecting the Wavelength and Slit Width 

The source for atomic absorption is a hollow cathode lamp consisting of a cathode and anode enclosed within a glass tube filled with a low pressure of Ne or Ar (Figure 10.41). When a potential is applied across the electrodes, the filler gas is ionized. The positively charged ions collide with the negatively charged cathode, dislodging, or “sputtering,” atoms from the cathode’s surface. Some of the sputtered atoms are in the excited state and emit ra- diation characteristic of the metal from which the cathode was manufactured. By fashioning the cathode from the metallic analyte, a hollow cathode lamp provides emission lines that correspond to the analyte’s absorption spectrum.

The sensitivity of an atomic absorption line is often described by its characteris- tic concentration, which is the concentration of analyte giving an absorbance of 0.00436 (corresponding to a percent transmittance of 99%). For example, Table 10.11 shows a list of wavelengths and characteristic concentrations for copper.

Usually the wavelength providing the best sensitivity is used, although a less sensitive wavelength may be more appropriate for a high concentration of analyte. A less sensitive wavelength also may be appropriate when significant interferences occur at the most sensitive wavelength. For example, atomizing a sample produces atoms of not only the analyte, but also of other components present in the sam- ple’s matrix. The presence of other atoms in the flame does not result in an inter- ference unless the absorbance lines for the analyte and the potential interferant are within approximately 0.01 nm. When this is a problem, an interference may be avoided by selecting another wavelength at which the analyte, but not the interfer- ant, absorbs.

The emission spectrum from a hollow cathode lamp includes, besides emission lines for the analyte, additional emission lines for impurities present in the metallic cathode and the filler gas. These additional lines serve as a potential source of stray radiation that may lead to an instrumental deviation from Beer’s law. Normally the monochromator’s slit width is set as wide as possible, improving the throughput of radiation, while being narrow enough to eliminate this source of stray radiation.

Preparing the Sample 

Flame and electrothermal atomization require that the sample be in a liquid or solution form. Samples in solid form are prepared for analysis by dissolving in an appropriate solvent. When the sample is not soluble, it may be digested, either on a hot plate or by microwave, using HNO₃, H₂SO₄, or HClO₄. Alternatively, the analyte may be extracted via a Soxhlet extraction. Liquid samples may be analyzed directly or may be diluted or extracted if the matrix is in- compatible with the method of atomization. Serum samples, for instance, may be difficult to aspirate when using flame atomization and may produce unacceptably high background absorbances when using electrothermal atomization. A liquid–liquid extraction using an organic solvent containing a chelating agent is frequently used to concentrate analytes. Dilute solutions of Cd2+, Co2+, Cu2+, Fe3+, Pb2+, Ni2+, and Zn2+, for example, can be concentrated by extracting with a solu- tion of ammonium pyrrolidine dithiocarbamate in methyl isobutyl ketone.

Minimizing Spectral Interference 

A spectral interference occurs when an analyte’s absorption line overlaps with an interferant’s absorption line or band. As noted pre- viously, the overlap of two atomic absorption lines is seldom a problem. On the other hand, a molecule’s broad absorption band or the scattering of source radia- tion is a potentially serious spectral interference.

An important question to consider when using a flame as an atomization source, is how to correct for the absorption of radiation by the flame. The prod- ucts of combustion consist of molecular species that may exhibit broad-band ab- sorption, as well as particulate material that may scatter radiation from the source. If this spectral interference is not corrected, then the intensity of the transmitted radiation decreases. The result is an apparent increase in the sam- ple’s absorbance. Fortunately, absorption and scattering of radiation by the flame are corrected by analyzing a blank.

Spectral interferences also occur when components of the sample’s matrix react in the flame to form molecular species, such as oxides and hydroxides. Ab- sorption and scattering due to components in the sample matrix other than the analyte constitute the sample’s background and may present a significant prob- lem, particularly at wavelengths below 300 nm, at which the scattering of radia- tion becomes more important. If the composition of the sample’s matrix is known, then standards can be prepared with an identical matrix. In this case the background absorption is the same for both the samples and standards. Alterna- tively, if the background is due to a known matrix component, then that compo- nent can be added in excess to all samples and standards so that the contribution of the naturally occurring interferant is insignificant. Finally, many interferences due to the sample’s matrix can be eliminated by adjusting the flame’s composi- tion. For example, by switching to a higher temperature flame it may be possible to prevent the formation of interfering oxides and hydroxides.

When the identity of the matrix interference is unknown, or when it is impossi- ble to adjust the flame to eliminate the interference, then other means must be used to compensate for the background interference. Several methods have been devel- oped to compensate for matrix interferences, and most atomic absorption spec- trophotometers include one or more of these methods.

One of the most common methods for background correction is the use of a continuum source, such as a D₂ lamp. Since the D₂ lamp is a continuum source, the absorbance of its radiation by the analyte’s narrow absorption line is negligible. Any absorbance of radiation from the D₂ lamp, therefore, is due to the background. Absorbance of radiation from the hollow cathode lamp, how- ever, is due to both the analyte and the background. Subtracting the absorbance for the D₂ lamp from that for the hollow cathode lamp gives an absorbance that has been corrected for the background interference. Although this method of background correction may be quite effective, it assumes that the background absorbance is constant over the range of wavelengths passed by the monochro- mator. When this is untrue, subtracting the two absorbances may under- or over-correct for the background.

Other methods of background correction have been developed, including Zee- man effect background correction and Smith–Hieftje background correction, both of which are included in some commercially available atomic absorption spec- trophotometers.

Minimizing Chemical Interferences 

The quantitative analysis of some elements is complicated by chemical interferences occurring during atomization. The two most common chemical interferences are the formation of nonvolatile com- pounds containing the analyte and ionization of the analyte. One example of a chemical interference due to the formation of nonvolatile compounds is ob- served when PO₄3– or Al3+ is added to solutions of Ca2+. In one study, for exam- ple, adding 100 ppm Al3+ to a solution of 5 ppm Ca2+ decreased the calcium ion’s absorbance from 0.50 to 0.14, whereas adding 500 ppm PO₄3– to a similar solution of Ca2+ decreased the absorbance from 0.50 to 0.38.21 These interfer- ences were attributed to the formation of refractory particles of Ca₃(PO₄)₂ and an Al–Ca–O oxide.

The formation of nonvolatile compounds often can be minimized by increas- ing the temperature of the flame, either by changing the fuel-to-oxidant ratio or by switching to a different combination of fuel and oxidant. Another approach is to add a releasing agent or protecting agent to solutions containing the analyte. A releasing agent is a species whose reaction with the interferent is more favorable than that of the analyte. Adding Sr2+ or La3+ to solutions of Ca2+, for example, minimizes the effect of PO₄3– and Al3+ by reacting in place of the analyte. Thus, adding 2000 ppm SrCl₂ to the Ca2+/PO₄3– and Ca2+/Al3+ mixtures discussed in the preceding paragraph gave absorbances for each of 0.48, whereas a solution of 2000 ppm SrCl₂ and Ca2+ alone gave an absorbance of 0.49. Protecting agents react with the analyte to form a stable volatile complex. Adding 1% w/w EDTA to the Ca2+/PO₄3– solution discussed in the preceding paragraph gave an absorbance of 0.52, compared with an absorbance of 0.55 for just the Ca2+ and EDTA. On the other hand, EDTA does not serve as a protecting agent for solutions of Ca2+ and Al3+.

Ionization interferences occur when thermal energy from the flame or elec- trothermal atomizer is sufficient to ionize the analyte

             10.28

where M is the analyte in atomic form, and M+ is the cation of the analyte formed by ionization. Since the absorption spectra for M and M+ are different, the position of the equilibrium in reaction 10.28 affects absorbance at wavelengths where M ab- sorbs. If another species is present that ionizes more easily than M, then the equilib- rium in reaction 10.28 shifts to the left. Variations in the concentration of easily ionized species, therefore, may have a significant effect on a sample’s absorbance, resulting in a determinate error. The effect of ionization can be minimized by adding a high concentration of an ionization suppressor, which is simply another species that ionizes more easily than the analyte. If the concentration of the ioniza- tion suppressor is sufficient, then the increased concentration of electrons in the flame pushes reaction 10.28 to the left, preventing the analyte’s ionization. Potas- sium and cesium are frequently used as ionization suppressors because of their low ionization energy.

Standardizing the Method 

Because Beer’s law also applies to atomic absorp- tion, we might expect atomic absorption calibration curves to be linear. In prac- tice, however, most atomic absorption calibration curves are nonlinear, or linear for only a limited range of concentrations. Nonlinearity in atomic absorption is a consequence of instrumental limitations, including stray radiation from the hol- low cathode lamp and a nonconstant molar absorptivity due to the narrow width of the absorption line. Accurate quantitative work, therefore, often requires a suitable means for computing the calibration curve from a set of standards. Non- linear calibration curves may be fit using quadratic and cubic equations, al- though neither works well over a broad range of concentrations.

When possible, a quantitative analysis is best conducted using external stan- dards. Unfortunately, matrix interferences are a frequent problem, particularly when using electrothermal atomization. For this reason the method of standard additions is often used. One limitation to this method of standardization, however, is the re- quirement that there be a linear relationship between absorbance and concentration.