Atomic Spectroscopy - A Primer
Atomic spectroscopy (also termed atomic spectrometry) is one of the oldest and most well established of the analytical methods. Its roots can be traced to very early observations in which the presence of specific salts in a chemical sample imparted characteristic colors to a luminous flame. Atomic spectrometry is perhaps the most prominent and widely used of the family of methods employed for elemental analysis.
All methods for elemental analysis, including atomic spectrometry, exploit quantized transitions characteristic of each individual element. Conveniently, atomic spectrometry utilizes quantized transitions in the conveniently accessible ultraviolet, visible, and near-infrared regions of the electromagnetic spectrum. As a result, the spectroscopic instrumentation is relatively inexpensive, readily available, and easy-to-use. Of course, it is valance electronic transitions that occur in these spectral regions. For narrow-band, characteristic spectra to be generated then requires each atom to be isolated from all others, so the transitions are not perturbed by neighboring atoms or by bonding effects. If this requirement is not met, the resulting spectra are representative more of molecules or molecular fragments than of atoms themselves.
Accordingly, the underlying requirement for all atomic methods of analysis is that a sample be decomposed to the greatest extent possible into its constituent atoms. Ideally, this atomization step should be quantitative; there should be no residual bonding in the gas-phase atomic cloud. Anything less than complete atomization will understandably yield lower sensitivities in any atomic method. Even more importantly, changes in the fraction of atomization from sample to sample or from sample to standard will cause errors in calibration. Thus, it is less important that complete atomization be achieved than that the fractional atomization be extremely consistent. To the extent that this condition is not met, interelement (matrix) interferences in atomic spectrometry can be extremely troublesome.
It is not surprising, then, that a consistent theme in the history of atomic spectrometric analysis is a search for improved methods of atomization for samples in solid, solution, or gaseous form. It is therefore appropriate that we consider such systems in some detail here.
Systems for Atom Formation in Atomic Spectroscopy
Schemes for the atomization of various samples have followed several traditional paths, outlined in Figure 1. The most common first step is to dissolve the sample if it is not already in solution form. Although this step is an inconvenient and often time-consuming one, it also offers a number of important benefits. First, after a sample is dissolved, the principal constituent in the sample solution is the solvent. Consequently, most sample solutions look more or less the same and similarly resemble standard solutions that are prepared. Secondly, samples in solution form are relatively easy to handle and lend themselves readily to automation. Third, sample solutions permit relatively simple and straightforward background correction, simply by use of a solvent or reagent blank. Lastly, other constituents can readily be added to sample and standard solutions to simplify such procedures as standard additions (spiking) and internal standardization.
(Click on image to enlarge)
Once a sample solution is prepared, it can be converted to free atoms in several alternative ways. One is to pipette a small aliquot of the sample solution into an electrothermal atomizer (ETA) such as a graphite furnace. This method corresponds to the right-hand path in Fig. 1. In the graphite furnace, the solvent is first evaporated from the sample at a moderate temperature. The furnace temperature is then raised so organic material is ashed, and the temperature next increased rapidly to the point where the sample is vaporized and ultimately atomized. Not surprisingly, the furnace temperature and the sample composition are extremely important in ensuring the efficiency and consistency of this atomization.
A second path to the atomization of solution samples is through a nebulization (spraying) process. In simple terms, nebulization serves to increase the surface area of the solution sample, so solvent evaporation (desolvation) can proceed more rapidly and so the resulting dried solute particles can be volatilized better. This scheme, which is probably the most common in analytical atomic spectrometry, is employed in flame atomic absorption, flame emission, and plasma emission spectrometry and is shown in the left-hand path in Fig. 1. Once formed, droplets in the nebulized spray are sent into a high-temperature environment such as a chemical flame or flowing rare-gas plasma. There, desolvation and solute-particle vaporization take place, and the resulting vapor converted more or less efficiently into free atoms. Indeed, the environment in these discharges is often hot enough that many of the atoms that are formed wind up as positively charged ions. Also, the environment in these atomization sources is commonly energetic enough to yield strong emission from either the freed atoms or their ionic counterparts. Correspondingly, this general scheme lends itself well to a number of different detection approaches that will be discussed shortly: atomic emission spectrometry, atomic absorption spectrometry, atomic fluorescence spectrometry, and atomic mass spectrometry.
For some elements, it is possible to employ a more straightforward means of generating free atoms. In the case of mercury, for example, free atoms can be formed simply by the chemical reduction of inorganic mercury in solution to the free atomic form. The neutral mercury atoms can then be driven from solution merely by passing an appropriate carrier gas (for example, air or argon) through it. The liberated atoms, present in a relatively cool environment, can then be measured alternatively by such techniques as atomic absorption or atomic fluorescence spectrometry. This approach is found in the center route in Fig. 1.
For other elements, a chemical reaction will yield not free atoms but rather other volatile species that can be dissociated at moderate temperatures into free atoms. The most common example here is the use of a chemical reduction to form a stable hydride of such elements as tin, antimony, and arsenic. As in the case of mercury atoms, these volatile hydrides can be driven from solution by bubbles of an appropriate gas, which will carry them into a moderate-temperature flame or furnace for atomization. Once formed, those atoms can then be measured by atomic absorption, atomic fluorescence, or certain other spectrometric methods.
Naturally, it would be desirable in many cases to be able to analyze solid samples directly. Even more beneficial in special cases would be the possibility of measuring sample concentrations in a solid on a three-dimensional spatial basis. Some methods employed for sample volatilization and atomization have been aimed at exactly that goal and are also indicated in Fig. 1.
Of such methods, the best established employs a glow discharge, dc arc, or high-voltage spark. In a glow discharge, ordinarily operated at pressures in the range of 1 torr, the sample surface is bombarded by energetic rare-gas ions, usually of argon. Through this steady bombardment, the surface is eroded layer-by-layer, a process called "sputtering". Thus, atoms freed from the surface as a function of time are taken from successively deeper layers within the sample. Recording the spectrometric signal as a function of time then permits a "depth profile" of the sample to be obtained. In early work, glow-discharge devices were applicable only to conductive solid samples because of the need to impart a negative charge to the sample surface, in order to attract argon-ions to it. More recent developments, however, have shown it possible to utilize radiofrequency-sustained discharges for the same purpose; such discharges permit the depth-resolved analysis of nonconductive samples as well.
In recent years, intense laser beams also have been exploited for depth resolution. Unlike the glow discharge, however, a laser beam can be focused to discrete, extremely small spots on the sample surface, so that not only depth resolved but also laterally resolved information can be obtained. In its most straightforward version, the laser is employed not as an atomization source but rather as an ablation-based sampling device. The sample material ablated by a laser can be fed into any of several sources for further atomization. The most attractive combinations have been between laser ablation and either inductively coupled plasma-atomic emission spectrometry and inductively coupled plasma mass spectrometry.
Some sources employed in atomic spectrometry are not as good at generating atoms as they are at exciting or ionizing them. The microwave-induced plasma (MIP) typifies such sources. In fact, many MIP systems tolerate only a small loading of solvent vapor before they are visibly perturbed. The MIP therefore finds an important role in atomic spectrometry in the dissociation of gas-phase samples and in the production of atomic emission and mass spectra. It is therefore particularly attractive as a source for detection of a gas chromatographic effluents, of gas-phase species generated by a chemical-based hydride-formation apparatus, or for atoms volatilized from an auxiliary source such as a carbon furnace.
Detection Methods in Atomic Spectroscopy
Once atoms are in the gas phase, they can be probed by any of several spectrometric techniques, including atomic emission spectrometry (AES), atomic absorption spectrometry (AAS), atomic fluorescence spectrometry (AFS), atomic mass spectrometry (AMS), coherent forward scattering spectrometry (CFS), photothermal deflection spectrometry, atomic magneto-optic rotation spectrometry, and several others. Of these alternatives, diagrammed in Fig. 2, the most common have become AAS, AES, and AMS.
AAS dates from the earliest observations by Fraunhofer of dark lines in the Sun?s spectrum and, as an analytical method, from the pioneering work by Walsh and by Alkemade in 1955. It has become a workhorse in atomic spectrometric analysis, in large part because of the simplicity of the instrumentation that it requires. To perform atomic absorption requires only a primary light source, usually a hollow-cathode lamp, an atomization cell such as any of those described in the preceding paragraphs, a spectral isolation device such as a monochromator, an appropriate photodetector, and associated electronics. Working curves are usually constructed in the form of a Beer?s-law plot familiar from molecular spectrophotometry. However, for Beer?s law to be followed closely (that is, for linear calibration curves to be obtained), the band of detected radiation must be narrower than the atomic absorption line that is being measured. Because atomic absorption lines are extraordinarily narrow, the primary light source (hollow cathode lamp) must emit a band of light that is at least as narrow. It is for this reason that the hollow cathode lamp was chosen. Significantly, this choice also simplifies the measurement, since spectral lines emitted by the hollow cathode lamp are typical of the element being measured and are naturally locked onto the narrow atomic absorption lines of interest. Thus, spectral mismatch is almost impossible and spectral interferences are relatively uncommon because of the narrowness of the spectral lines that are being probed.
Unfortunately, it is not particularly convenient to measure more than one element at a time by means of AAS. There is a necessary straight-line geometry among the primary light source (hollow cathode lamp), the atom cell (flame or furnace) and the detection equipment. This alignment makes it difficult to incorporate more than a single source into the system; because each hollow cathode lamp emits efficiently the spectrum of only one, two, or three elements at a time, measuring additional elements requires substituting a new hollow cathode lamp. Although recent advances have been made in continuum-source atomic absorption, even those arrangements are somewhat limited: continuum sources that extend well into the important ultraviolet region of the spectrum are not widely available. Moreover, when furnace atomization is employed, it is often necessary to employ a different temperature program for each element.
One of the problems encountered in AAS is the generation of a broad-band spectral background. This background arises from absorption caused by residual molecules or molecular fragments in the atom source and by scattering from smoke or other airborne particulate matter. Because a hollow-cathode lamp is utilized as a light source, this molecular absorption or scattering cannot be distinguished from atomic absorption, since it simply attenuates the beam. If left uncorrected, this "non-specific absorption" then makes element concentrations appear to be higher than they really are.
Not surprisingly, non-specific absorption is more of a problem when furnace atomizers are employed than when a chemical flame is utilized. In a furnace, a dense smoke cloud often arises during atomization of the sample, whereas in a chemical flame the particulate matter is more thoroughly volatilized. Nevertheless, it is considered to be prudent to employ some form of background correction with either a furnace or flame atomizer.
Several alternative schemes for background correction have been developed for AAS. The most common of these schemes, termed "continuum-source background correction", is particularly ingenious and was the first of such methods to be developed. It exploits the spectral difference between narrow-band atomic absorption and broadband molecular absorption or scattering. However, two sources are required, a hollow cathode lamp and an auxiliary continuum source, usually a deuterium arc lamp. Still, it remains the most widely used method.
Another technique for background correction in AAS utilizes the Zeeman effect. The Zeeman effect is the splitting of atomic lines that occurs when the atoms are present in a magnetic field. Accompanying this spectral splitting is a change in polarization of the spectral lines that are generated. By use of appropriate instrumentation, it is possible to distinguish atomic absorption from the background features. Although no auxiliary source is required, a magnetic field and polarizers usually are.
Another approach employed for background correction in atomic absorption employs a pulsed hollow cathode and is similar in concept to the continuum-source procedure. In essence, the same hollow cathode lamp is used as both a narrow-band and broad-band source. The broad-band spectrum arises from spectral-line broadening that occurs at unusually high hollow-cathode operating currents.
In contrast to AAS, atomic emission spectrometry is inherently a multielement procedure. In a high-temperature flame or plasma (such as the inductively coupled plasma or ICP), atoms are not only formed extremely efficiently, but also they generate intense atomic emission. The emission is isotropic and also occurs from all elements at the same time. It is therefore possible to perform simultaneous multielement determinations simply by means of a multichannel detection system. Multichannel devices are routinely being introduced that employ two-dimensional spectral dispersion coupled with two-dimensional arrays of detector elements of unprecedented sensitivity and low noise.
Like AES, atomic mass spectrometry is inherently a multielement approach. Moreover, AMS can provide information about isotopes present in a sample. Although a small spectral shift occurs in atomic emission spectra because of the "isotope effect", for most elements in the periodic table that shift is too small to be detected with a conventional spectrometer. In contrast, AMS can readily distinguish one isotope from another; one of the important applications of AMS has therefore become the measurement of isotope ratios in various kinds of samples.
AMS offers also the benefit of extraordinarily low background levels. The result is that typical detection limits in AMS are lower than those in AES by a factor of 100 or so. Recent manufacturers? literature suggests that it is now routinely possible to measure solution concentrations in inductively coupled plasma-mass spectrometry at the level of one part per quadrillion or so.
For the ultimate in sensitivity, however, none of the above-mentioned techniques can better atomic fluorescence spectrometry. Indeed, AFS studies have demonstrated the detection of single atoms of a desired element. The reason for this extraordinary sensitivity is quite straightforward: in AFS, each atom can be detected many times, since it can be excited over and over again. Each time the atom emits a photon, it can be re-excited by an incident beam of photons, probably from a laser. In contrast, detection in atomic mass spectrometry is usually destructive, so there is only a single chance to observe each atomic ion. Although AES also offers the opportunity to collect several photons from each atom, the background levels are usually high enough that it is difficult to distinguish the few photons an atom emits from others that a detector registers.
From this brief narrative, it should be clear that atomic spectrometric analysis is not only an extremely important area of chemical analysis, but also one that continues to evolve rapidly. Over the past decade, the nature of instrumentation for atomic spectrometry has changed dramatically. Benchtop units are now available that measure virtually the entire atomic emission spectrum of a complex chemical sample at once. Similarly, alternative kinds of mass spectrometric instrumentation (quadrupole mass filters, sector-field instruments, time-of-flight mass spectrometers, ion traps, and Fourier transform mass spectrometers) have been applied to AMS. These alternative approaches enable samples to be analyzed with unprecedented speed, accuracy, and precision. They permit isotope ratios to be determined on samples having elemental concentrations below 1 ppt and at a speed that would have been considered impossible only a few years ago. Because of this rapid development, the reader should attempt to keep abreast of ongoing developments.