JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL. 2 537 State of the Art of Glow Discharge Lamp Spectrometry* Plenary Lecture Jose A. C. Broekaert lnstitut fur Spektrochemie und angewandte Spektroskopie, Postfach 778, 4600 Oortmund 1, FRG The development of glow discharge sources for atomic spectroscopy is traced. The electrical characteristics of the analytically relevant sources, sample volatilisation and the predominant excitation and ionisation processes are discussed. The possibilities of using the Grimm-type glow discharge lamp with a flat cathode for the spectral emission analysis of compact metallic samples, for non-conducting powders, after mixing with metal powder and making into pellets and for in-depth profiling are described. New types of lamps and the features of glow discharges with hollow cathodes are evaluated.The use of glow discharges as atom reservoirs for atomic absorption and fluorescence spectrometry and recent advances in glow discharge mass spectrometry are also covered. Keywords: Glow discharge; atomic absorption spectrometry; atomic emission spectrometry; atomic fluorescence spectrometry; mass spectrometry Different types of glow discharges have been used for a long time as radiation sources in atomic spectroscopy. Not only because of their abilities to excite atomic spectra but also because of the volatilisation processes involved, they have been found to be interesting alternatives to arc and spark sources operated at atmospheric pressure. At 10-1000 Pa, the filler gas pressure range for the analytical sources being discussed, the electrical characteristic of the gas discharge (Fig.1)’ commences with a Townsend discharge. There is only a small amount of ion and free electron production and a transition range where, owing to the increased energy exchange by collisions, the current even increases at decreasing discharge voltages. In the glow discharge region, the current increases at constant current density, as here the cathode surface covered by the discharge grows with the discharge current. Once the discharge covers the whole cathode (restricted glow discharge), the current can only increase with increasing current density which requires an increase in the discharge voltage. In this “abnormal” region, charged particles acquire high energies by passing through the high electric fields produced by the high discharge voltage.t .g Q, gk! ‘v, I- I - 1 - - _ _ - - - - I I I 1 I 10-9 10-7 10-5 10-3 10-1 10 CurrentiA - Fig. 1. Characteristic of a self-sustaining gas discharge [V = f(log i)]: Vb = breakdown voltage; V, = normal cathode fall of potential; and Vd = arc voltage (similar to reference 1) * Presented at the 1987 Winter Conference on Plasma and Laser Spectrochemistry, Lyon, France, 12th-16th January, 1987. This particularly applies for the energy-rich cathode region. Accordingly, filler gas species impact on the cathode surface and may cause “sputtering” of the cathode material. When the discharge current increases further, the current density becomes so high that intensive heating of the cathode through bombardment with filler gas species may cause thermal evaporation.Then, the availability of high number densities of analyte perturbs the high fields by which the characteristic becomes normal, i.e., the current then increases at decreasing discharge voltage as is the situation for a d.c. arc. Sample volatilisation in glow discharges is mainly due to cathodic sputtering (Fig. 2 ) . Indeed, because of the acquisi- tion of high energies in the cathode region, positive ions, through the transition of mechanic momentum, may release cathode material on impact. When the sample to be analysed is brought into the cathode region or the sample itself acts as the cathode, the analyte volatilises by cathodic sputtering. In the analytically important glow discharges, hard-sphere colli- sions mainly occur.As is known from sputtering experiments with an ion source under high vacuum, the sputtering yield S then is given by K m.M s=-.-. E . . . . . . ( 1 ) ( m + W 1 h(E) = - . . . JI: R2no where K is a constant, rn the mass of the incident particles, M the mass of the static particles, E the energy of the incident particles, no the number of target atoms per unit volume and Ro the distance of closest approach under hard-sphere conditions.2 Accordingly, the influence of various parameters on the ablation rates in analytical glow discharges, as discussed later, can be investigated. However, as several ionic and neutral species at pressures of 10-1000 Pa coexist with very different energies and collide with the filler gas, it is impossible to derive quantitative conclusions from such formulae.Selective \ Mv Photoionisation and electron emission Cathodic sputtering Thermal evaporation Fig. 2. Analyte volatilisation in atomic spectrometry538 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL. 2 volatilisation, which occurs when the sample material enters the discharge plasma by thermal evaporation, is typical of arc discharges at atmospheric pressure but is less so for the glow discharges being discussed here. Excitation and ionisation in glow discharges are mainly due to electron impact, but other processes also contribute.3 1. Electron impact (mainly with high energy electrons) : Excitation e + Ar+ Ar* + e e + M+ M* + e Ionisation e + Ar -+ Ar+ or AT+* 2. Collisions of the second kind: Excitation Ar* + M-+ M* + Ar Ar*+M+-+M+*+Ar Ar* + Ar+ Ar+ + Ar* Ionisation Arm + M + M+ * + Ar (also Arm) (analyte excitation by electron impact) (for cathodic sputtering, emission of secondary electrons) (analyte excitation by collisions of the second kind) (analyte excitation by collisions of the second kind) (charge transfer, sputtering by neutrals) excitation by Penning ionisation) (analyte 3.Recombination (mainly with slow electrons): M++e+M* (radiative recombina- tion) Glow discharge plasmas are known not to be in local thermal equilibrium (LTE). Indeed, electron temperatures, as measured with probes, are very high (>lo000 K). The electrons, because of their small mass, may acquire high velocities at the high field gradients. Excitation temperatures, as derived from spectroscopic measurements, are lower and differ from the thermometric lines used, This also relates to deviations from the Boltzmann distribution.The existence of temperatures of ca. 5000 and 10000 K in a hollow-cathode discharge plasma, for instance, can be explained by the non-Maxwellian velocity distribution of the electrons as a result of the high electric field and by the existence of two groups of electrons, namely a low-energy group involved in recombination and a high-energy group responsible for the excitation of high energy levels. Hollow cathode Glow discharge lamp FANES Fig. 3. spectrometry Discharges under reduced pressure for atomic emission Kinetic gas temperatures in glow discharges are low. They may be measured from the Doppler broadening of spectral lines or approximated by the rotation - vibration tempera- tures.For a glow discharge with a flat cathode, gas tempera- tures of <lo00 K resulted from Doppler width measure- ments.4 For a hollow-cathode glow discharge, values from rotation - vibration spectra5 are of the same order of magni- tude (800-1500 K). Glow Discharge Atomic Emission Spectrometry Discharges under reduced pressure, from the point of view of sample volatilisation and excitation, have unique features as radiation sources for atomic emission spectrometry. When the sample is cooled, material volatilisation is solely based on cathodic sputtering, which excludes selective volatilisation and related interferences. However, it is also possible to adjust the cathode temperature carefully by regulation of the discharge current and/or external heating.The latter provides ideal conditions for the selective volatilisation of elements or species from a complex matrix. Owing to the non-LTE character of the discharge, high line to background intensity ratios are obtained for the analyte lines. The spectra hence consist mainly of the most sensitive lines, which are narrow because of the low Doppler and pressure broadening and do not suffer significantly from interferences by molecular bands. Indeed, the discharge gas is nearly always a noble gas which, moreover, because of its high ionisation energy, also permits an efficient excitation of high-energy terms. Three main types of discharges under reduced pressure (Fig. 3) are used as radiation sources for atomic emission spectrometry and are discussed below.(GD-AES) Glow Discharge Lamp In 1968 the first practical glow discharge lamp for spectro- chemical analysis with a flat cathode was described by Grimm.6 In his lamp (Fig. 4) the sample, which must be electrically conducting, is taken as the cathode and the discharge is restricted to the sample surface. This is achieved by keeping the distance between the anode tube on one side and the cathode block and the sample on the other below the mean pathway of the free electrons, and by evacuating the interspace with a supplementary vacuum pump. The analyte is volatilised uniquely by cathodic sputtering (typical penetra- tion rate 3 pm min-1) and is excited in the negative glow of the discharge mainly by electron impact.The radiation is measured by end-on observation, As indicated by the black area in the right-hand part of Fig. 4, the material is partly deposited on the anode tube, which normally has a diameter of between 6 and 10 mm, and the remainder is sucked away into the vacuum system. The lamp is normally operated in argon at a pressure of 10-300 Pa. The vacuum supply includes a dual-vacuum pump and a needle valve for admitting the filler gas (see, for example, references 7 and 8). In commercial instrumentation interlocks and automatic rinsings are pro- vided. To pump 2 Cathode body). , Sample 0.2 mm thick ' 1 min- PTFE sheath TO pump I CatLode Fig. 4. Sample volatilisation and excitation in a Grimm-type glow discharge lamp539 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL.2 In the glow discharge lamp, the material volatilises uniquely by cathodic sputtering. The influences of the sample composi- tion and the discharge parameters on the ablation rate have been studied extensively7.9 and they can be explained by the impulse theory. Indeed, this clearly indicates that the ablation rate will increase in the sequence: helium < neon < argon < krypton (for aluminium samples), and C < Al < Fe < Cu < Zn (in argon gas). In most instances argon is used as the filler gas, as the heavier noble gases are expensive and their ionisation energies are lower. When initiating the glow discharge and bringing the operation voltage and current to their nominal values of about 700 V and 80-120 mA, respectively, a certain burning time is required before the sputtering and excitation conditions become stable.This time, for example, for steel samples, may be as short as 20 s. It may be reduced by high-energy pre-burning. When equilibrium conditions are reached, the composition of the ablated material becomes equal to that of the sample, which is a pre-requisite of low matrix interferences and for the exclusion of structural effects. As shown by the burning spot records (Fig. 5 ) , no melting of the surface and accordingly no selective volatilisation occur in the glow discharge in contrast to spark erosion. As shown in many publications (see, for example, refer- ences 10-14) very different types of alloys (e.g., pure aluminium and different A1 alloys*s) can be analysed with GD-AES using the same calibration function.This demon- strates the potential interest in glow discharge lamp emission spectrometry by users who have to analyse a wide variety of samples. As shown by results for steel,lh.'7 the detection limits of GD-AES are in the range 1-10 yg g-1. Results are often higher than for spark atomic emission spectrometry. However, as will be discussed later, progress with respect to this point is still being made. Apart from compact metallic samples, non-conducting powders can also be analysed with the glow discharge lamp. They can be mixed with a metal powder and the mixture can be briquetted into pellets. By using appropriate techniques, vacuum-tight and mechanically stable pellets can be obtained even with small amounts of sample (Fig.6).*,18 As is known from the analysis of slags, for instance, analyses can be performed by calibrating with synthetic samples. The method is of topical interest because, in contrast to X-ray fluorescence spectrometry, light elements such as beryllium and boron, which may be important for the quality of new ceramic materials, can also be determined. It has been shown that the graininess of the material plays an important role in the analytical performance. 19 A major feature of the glow discharge lamp lies in the easily controllable layer by layer ablation. The penetration rate depends on the voltage, the current and the gas pressure and is material specific.20-22 By integrating the analytical signals with a small time window, information on the in-depth variation of the elemental composition is obtained.2" The penetration rate of the glow discharge lamp operated at about 100 Pa of argon and 90 W (i = 50-70 mA and V = ca.800 V) is of the order of 3.5 ym min-1. As discussed by Quentmeier and Laqua,22 RSDs of the intensities at a concentration level of 10 mg g-1 for an integration time of 1 s are below 0.1. Accordingly, for many applications integration times down to 0.1 s are still tolerable and the resolution of in-depth profiling by GD-AES is of the order of 5 nm. The technique has been proven to be useful for the control of the thickness as well as for the study of the composition of coatings on technical surfaces. Compara- tive studies of several surface analysis techniques showed that the in-depth resolution of GD-AES is lower than in secondary ion mass spectrometry and Auger electron spectroscopy, the latter being a real "surface" technique.Therefore, GD-AES has a high power of detection and a high multi-element capacity and it just may solve problems where low in-depth resolution is required.23 Applications to the quality control of coatings on steel24 and aluminium23 have been described. As in all known surface techniques, with GD-AES it is still difficult to derive quantitative information from in-depth resolved intensity profiles. This applies particularly to infor- mation from layers near to the surface where the excitation conditions in the glow discharge are often perturbed by gases released from the surface during the initiation of the disc- harge.For deeper layers, it has previously been shown22 that it is possible to quantify elements by referring to the elemental line intensities from the bulk. This idea has been pursued by Bengston ,25 who introduced model sputtering constants and discharge parameters derived from measurements with the pure metals. For galvanisation layers his glow discharge values agreed well with those obtained by electropolishing. Despite the fact that the problem of quantification has not been satisfactorily resolved, GD-AES is already used for solving many practical problems, particularly in metallurgy. 0.2 rnrn Copper + 'sample Cobper Fig. 6. Preparation of pellets from electrically non-conducting powders for glow discharge lamp spectrometry (according to refer- ence 18) Fig- 5.medium voltage spark ( b ) , respectively; sample, aluminium Burning spots obtained with a glow discharge ( a ) and aJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL. 2 540 In addition to the application areas mentioned above there is still considerable development in GD-AES. As described in papers by Ko26 and Ferreira et al., 27 a lamp in which the anode is further away from the sample and where a ceramic restrictor defines the burning spot is advantageous. Indeed, the field across the sample is then more uniform, which results in a flatter burning spot and, accordingly, in a better in-depth resolution. It was found that the analytical performance of this lamp in terms of precision and freedom from matrix interfer- ences is even better than that of the original Grimm-type lamp.Special efforts are being made to improve the power of detection of GD-AES. This is required by the demands for elemental characterisation of metals in the sub-yg g-1 range in technological applications. Some success could be obtained by increasing the ablation rate. This is possible to a certain extent with the magnetic field lamp where a strong Co - Sm magnet is used to focus the sputtering beam on the sample and at the same time to increase the length of the negative glow. As shown for stee1,28,2’ both the ablation rate and the power of detection with the magnetic lamp could be improved. It was also found that it was no longer necessary to use a second vacuum line to restrict the discharge to the sample surface.Therefore, the rinsing of the lamp became difficult for ferromagnetic samples. A further method of improving the power of detection lies in the use of a cross excitation. Indeed, as shown by Ferreira and Human,30 the greater part of the sputtered analyte in a conventional glow discharge lamp is present as ground-state neutral atoms. The use of a d.c.-boosted discharge has been propagated in the work of Lowe and other workers.31-33 It has been shown that the net intensities are considerably increased by the boosting. This helps to overcome detector noise and counting error limitations, which for the low radiative output in glow discharge work often limit the power of detection. Similar experience has been reported for h.f. boosted lamps.34 Leis et al.35.36 showed with a microwave-boosted lamp that not only a gain in net intensities but also a gain in signal to background ratio could be obtained. The detection limits for steel samples could be improved by a factor of up to five as compared with those of a conventional lamp while preserving the same analytical precision and a linear dynamic range of more than three orders of magnitude.Hollow-cathode Glow Discharge Lamps and FANES Glow discharges with hollow cathodes have long been important as radiation sources for emission spectrochemical analysis. It has been shown by Mandelstam and Nedler’7 that, in particular, owing to the high residence times of the analyte in the excitation zones which results from the cathode geometries, they are the most sensitive emission techniques.This fact is reinforced by the departures from LTE, by which the line to background ratios increase as compared with those of LTE sources.3* Both demountable cooled and hot hollow- cathode lamps have been used for spectrochemical analysis. In cooled hollow cathodes, the sample is placed as a disc at the bottom of the cathode or the hollow cathode is machined from the sample. Volatilisation takes place by cathodic sputtering only. In contrast with the glow discharge lamp with a plane cathode, however, no equilibrium in sample ablation is attained. Similarly, with the flat cathode the technique can be used for the analysis of metals and of electrically non- conducting powders. For the latter, Caroli39 showed that detection limits are in the pg g-1 range. Microanalyses with liquid sample aliquots can be performed by transferring them into a metal or graphite cathode and exciting the dry solution residues .40 The hot hollow cathode has continued to be of interest for emission spectrochemical analysis.Here, the sample volatilisation is due mainly to thermal evaporation. This has been shown by measurements on brass chips which have been excited for short periods.41 As the temperature of the cathode can be controlled very efficiently by selecting the proper discharge conditions, volatile elements can be evaporated reproducibly from refractory matrices. Accordingly, the hot hollow cathode is still being used for the analysis of high- temperature alloys, which can be used in the form of chips in graphite cathodes.Indeed, on the one hand these alloys are often difficult to dissolve which could hamper their analyses by ICP-AES but on the other hand the emitted matrix spectra are very line rich. However, in the hot hollow cathode volatile elements such as arsenic, bismuth, lead and selenium can be volatilised selectively from the matrix, hence spectral inter- ferences are circumvented. Owing to the high excitation temperatures in the hollow cathode these elements, which have high excitation energies, can be effectively excited and detection limits down to the sub-yg g-1 level are 0btained.42-~~ Hot hollow cathodes can also be used for applications such as the direct determination of toxic elements in airborne dust collected in graphite electrodes.45 The hot hollow cathode is considered to be the emission spectrochemical radiation source with the highest absolute power of detection.46 Indeed, absolute detection limits are in the picogram range in many instances, but only in the absence of a matrix.An effective separation of volatilisation and excitation is realised in furnace atomic non-thermal excitation spec- trometry (FANES), as developed by Falk et al. (see, for example, references 47 and 48). Here the samples are transferred into a graphite furnace, similar to those used in graphite furnace atomic absorption (GFAAS) work. The furnace is operated under reduced pressure and the analyte that is released by heating is excited in the negative glow between the furnace and an additional electrode. For dry sample aliquots, detection limits for a series of elements are in the picogram range.In contrast to GFAAS, simultaneous multi-element determinations are possible and calibration graphs are linear over several decades of concentration. The technique has been shown to improve considerably the power of detection as compared with furnace emission spectrometry at atmospheric pressure.49 Applications in the analysis of biological fluids have been described”) and direct solids sampling of pulverised biological substances has also recently been shown to be possible.51 Glow Discharges as Atom Reservoirs As the material volatilised in a glow discharge is to a large extent present as a vapour cloud of free atoms, glow discharges are suitable atom reservoirs for atomic absorption and atomic fluorescence work.By using a Grimm-type glow discharge as an atomiser, direct analyses of compact metallic samples can be performed by atomic absorption spectrometry (see, for example, refer- ence 52). By optimising the discharge conditions at the highest sample ablation rates, high powers of detection can be obtained. Further, the use of a reference signal has been shown to enable high precision analyses. Atomic fluorescence can also be performed on the vapour cloud. The fact that decay of the excited species by quenching, because of the reduced pressure, is low is a distinct advantage. The non-dispersive resonance fluorescence spectrometer de- scribed by Human et a1.53 makes use of one glow discharge as the primary radiation source and a second glow discharge, for which the sample is used as the cathode, as the atom reservoir.For a series of applications spectral resolution of the analytical signal has been shown not to be required.54 Multi-elemental determinations can be performed by using a multi-element primary source and a spectrometer. Laser induced fluor- escence with a glow discharge as the atom reservoir has alsoJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER been described.55.56 In experiments with dry solution residues on a flat cathode, detection limits in the picogram range are obtained (see, for example, reference 57). Glow discharges are also of potential interest as atom reservoirs for laser enhanced ionisation techniques. Glow Discharges as Ion Sources Apart from their use as sources for optical atomic spec- trometry, glow discharges also have been recognised as being powerful ion sources for inorganic mass spectrometry. Owing to the material volatilisation by sputtering, selective volatilisa- tion as compared with classical spark source or vacuum arc ion sources is low, which is a pre-requisite for reducing the matrix interferences.As steady-state conditions in sample ablation can be obtained the use of sequential systems, such as low-priced quadrupole systems, became possible. A glow discharge mass spectrometry system thus consists of the glow discharge source, which is operated at a pressure in the range of one to several hundred Pa, a pump-backed skimmer unit to extract the ions at a suitable location and to bridge the pressure difference between the source and the mass spec- trometer and the ion optics and the mass spectrometer.Systems using discharges with one or two pin electrodes but also hollow-cathode plumes have been described by Harrison et al.57 Detection limits in the sub-pg g-1 range and a better analytical precision and accuracy than with conventional spark source systems have been reported .58 Apart from quadrupole mass filters, double focusing mass spectrometers providing higher spectral resolution have been used. It is also possible to use a modified Grimm glow discharge with a flat cathode as an ion source,59 by which, in addition to bulk analyses, in-depth profiling is possible. It has been shown that by suitable selection of the ion sampling location the intensities of the analyte and the cluster signals and accordingly the interfer- ences in the mass spectrum can be greatly influenced.Glow discharge mass spectrometry is now commercially available and may be expected to become of considerable interest for trace analyses of solid samples. Again, the technique may be applied to metals and electrically non-conducting powders, as discussed for GD-AES. Dry residues can also be analysed. The possibility of using isotopic dilution enables high ana- lytical precision and accuracy to be realised. The highest selectivity can be obtained by resonant ionisa- tion of the atoms released in the glow discharge with the aid of laser radiation. Resonance ionisation mass spectrometry thus enables it to suppress considerably the signals produced by the matrix.6” As an analytical method, however, it is still in the development stage. Conclusion Glow discharge lamp spectrometry has been shown to be a powerful method in analytical atomic spectrometry.In emission spectrometry it is an interesting alternative to spark excitation especially for in-depth analyses. Its powers of detection are still being improved. From this point of view, its use as an ion source for mass spectrometry could be an important development and, when properly optimised, high precision and accuracy as well as the capability of getting in-depth resolved information can then be realised. This work was supported by the Ministerium fur Wissenschaft und Forschung des Landes Nordrhein-Westfalen and by the Bundesministerium fur Forschung und Technologie. References 1 .2. Penning, F. M . , “Electrical Discharges in Gases,” Philips Technical Library, Eindhoven, 1957, p. 41. Kaminsky, M., “Atomic and Ionic Impact Phenomena on Metal Surfaces,” Springer-Verlag, Berlin, 1965, p. 227. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20, 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 1987, VOL. 2 541 Flugge, S., Editor, “Encyclopedia of Physics,” Volume XXII, Springer-Verlag, Berlin, 1956. Ferreira, N. P., Human, H. G . C., and Butler, L. R. P., Spectrochim. Acta, Part B, 1980, 35, 287. Broekaert, J. A . C., Bull. SOC. Chim. Belg., 1977, 86, 895. Grimm, W., Spectrochim. Acta, Part B, 1968, 23, 443. Dogan, M., Laqua, K., and Massmann, H., Spectrochim. Acta, Part B , 1971, 26, 631.El Alfy, S . , Laqua, K., and Massmann, H., Fresenius Z . Anal. Chem., 1973, 263, 1. Boumans, P. W. J . M., Anal. Chem., 1972, 44, 1219. Dogan, M., Laqua, K., and Massmann, H., Spectrochim. Acta, Part B, 1972, 27, 65. Klockenkamper, R., Laqua, K., and Dogan, M., Spectrochim. Acta, Part B, 1980, 34, 527. Kruger, R. A., Butler, L. R. P., Liebenberg, C. J . , and Bohmer, R. G . , Analyst, 1977, 102, 949. Jager, H., Anal. Chim. Acta, 1972, 60, 303. Ferreira, N. P., and Butler, L. R. P., Analyst, 1978, 103, 607. KO, J. B., and Laqua, K., XVIII Colloquium Spectroscopicum Internationale, Grenoble, Abstracts 11, 1975, p. 543. Wagatsuma, K., and Hirokawa, K., Anal. Chem., 1984, 56, 908. Rademacher, H . W., and de Swardt, M. C., Spectrochim.Acta, Part B, 1975, 30, 353. El Alfy, S . , PhD Dissertation, Dortmund, 1978. Mai, H., and Scholze, H., Spectrochim. Acta, Part B, 1986,41, 797. Berneron, R., Spectrochim. Acta, Part B., 1978, 33, 665. Waitlevertch, M. E., and Hurwitz, J . K., Appl. Spectrosc., 1976, 30, 510. Quentmeier, A . , and Laqua, K., in Koch, K. H., and Massmann, H . , Editors, “13. Specktrometertagung,” W. de Gruyter, Berlin, 1981, p. 37. Quentmeier, A., Bubert, H., Garten, R. P. H., Heinen, H. J., Puderbach, H., and Storp, S., Mikrochim. Acta, Suppl., 1985, 11, 89. Koch, K. H., Kretschmer, M., and Grunenberg, D., Mikro- chim. Acta, 1983, 2, 225. Bengston, A., Specrrochirn. Acta, Part B, 1985, 40, 631. KO, J . B., Spectrochim. Acta, Part B, 1984, 39, 1405. Ferreira, N. P., Strauss, J.A., and Human, H. G. C., Spectrochim. Acta, Part B , 1983, 38, 899. Kruger, R . A., Bombelka, R. M., and Laqua, K., Spectrochim. Acta, Part B, 1980, 35, 581. Kruger, R. A., Bombelka, R. M., and Laqua, K., Spectrochim. Acta, Part B, 1980, 35, 589. Ferreira, N. P., and Human, H. G. C., Spectrochim. Acra, Part B, 1981,36, 1981. Lowe, R. M., Spectrochim. Acta, Part B, 1978, 31, 257. Lomdahl, G . S . , McPherson, R . , and Sullivan, J. V., Anal. Chim. Acta, 1983, 148, 171. Gough, D. S . , and Sullivan, J. V., Analyst, 1978, 103, 887. Walters, P. E . , and Human, H . G . C., Spectrochim. Acta, Part B , 1983,36, 585. Leis, F., Broekaert, J. A. C., and Laqua, K . , XXIV Col- loquium Spectroscopicum Internationale, Garmisch-Partenkir- chen, Book of Abstracts, Volume 4, 1985, p.640. Leis, F., Broekaert, J . A. C., and Laqua, K., Spectrochim. Acta, Part B, in the press. Mandelstam, S. L., and Nedler, V. V., Spectrochim. Acta, 1961, 17, 885. Falk, H., Spectrochim. Acta, Part B, 1977, 33, 437. Caroli, S . , Progr. Anal. At. Spectrosc., 1983, 6, 253. Harrison, W. W., and Prakash, N. J., Anal. Chim. Acta, 1970, 49, 151. Broekaert, J . A. C . , Spectrochim. Acta, Part B, 1979, 34, 11. Thornton, K., Analyst, 1969, 94, 958. Thelin, B., Appl. Spectrosc., 1981, 35, 302. Berglund, B., and Thelin, B., Analyst, 1982, 107, 867. Broekaert, J. A. C., Bull. SOC. Chim. Belg., 1976, 85, 755. Zil’bershtein, Kh.I., “Spectrochemical Analysis of Pure Sub- stances,’’ Adam Hilger, Bristol, 1977. Falk, H., Hoffmann, E., and Ludke, Ch., Spectrochim. Acta, Part B, 1981, 36, 767. Falk, H . , Hoffman, E., and Ludke, Ch., Spectrochim. Acta, Part B, 1984, 39,283. Falk, H., Hoffmann, E., Ludke, Ch., Ottaway, J. M., and Giri, S. K., Analyst, 1983, 108, 1459.542 50. 51. 52. 53. 54. 55. 56. JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL. 2 Falk, H., Hoffmann, E., Ludke, Ch., Ottaway, J. M., and Littlejohn, D., Analyst, 1986, 111, 285. Falk, H., Hoffmann, E., Ludke, Ch., and Schmidt, K. P., Spectrochim. Acta, Part B, 1986, 41, 853. McDonald, D. C., Anal. Chem., 1977, 49, 1337. Human, H. G. C., Ferreira, N. P., Kruger, R. A., and Butler, L. R . P., Analyst, 1978, 103, 469. Bubert, H., Spectrochim. Acta, Part B, 1984,39, 1337. Smith, B. W., Omenetto, N., and Winefordner, J. D . , Spectrochim. Acta, Part B, 1984, 39, 1389. Patel, B. M., and Winefordner, J. D., Spectrochim. Acta, Part B, 1986,41,469. 57. Harrison, W. W. , Hess, K. R. , Marcus, R. K., and King, F. L., Anal. Chem., 1986, 58, 341A. 58. Bruhn, C. G., Bentz, B. L., and Harrison, W. W., Anal. Chem., 1979,51, 673. 59. Jakubovsky, N., Stuwer, D., and Tolg, G., Znt. J. Mass Spectrom. Ion Proc., 1986, 71, 183. 60. Hess, K. R., and Harrison, W. W., Anal. Chem., 1986, 58, 1696. Paper 5715 Received January 12th, 1987