Plasma Spectrometry and Molecular Information* Plenary Lecture Journal of Analytical 1 Atomic OLIVIER F. X. DONARD AND RYSZARD LOBINSKI Laboratoire de Photophysique et Photochimie Molkculaire CNRS Universitk de Bordeaux I F-33405 Talence France Trends in species-selective analysis by GC and HPLC with plasma source element selective detection are discussed and illustrated with examples of applications. The potential of plasmas for the acquisition of auxiliary molecular information and conditions for its reliability are highlighted. The need for interaction of plasma techniques and classical organic MS techniques is emphasized. Keywords Plasma spectrometry; gas chromatography; high- performance liquid chromatography; speciation Metals exist in the environment or in biological systems as part of larger molecular structures.' Each and every such structure is the result of a series of biochemical processes defined by nature.They carry messages that are vital for the evaluation of the bioavailability of the metal and hence of its toxicity or essentiality and for the understanding of mechan- isms controlling biological life and regulatory pathways. The question of how to access this molecular information (speci- ation) has become one of the most pertinent and challenging issues to modern analytical chemistry during the past decade because of its impact on environmental chemistry eco- and clinical toxicology and to food and energy Classical approaches to speciation analysis have included the use of an intrinsically species-selective technique such as an electrochemical technique or use of soft-fragmentation mass or tandem mass spectrometry.Although these techniques can work well for standard solutions they usually fail for sub- ngg-' concentrations of an analyte in the presence of a complex matrix. As the metal of interest forms only a tiny part of the whole species detection focused on the polyatomic fragment usually lacks selectivity and suffers from high back- ground noise and isobaric matrix interferences. The two key issues for a successful analysis (sensitivity and selectivity) cannot be properly addressed other than by targeting the metal itself. Plasmas especially the ICP have become established in the field of inorganic analytical chemistry as a successful way of overcoming chemical interferences matrix signal suppression and the poor excitation efficiency associated with a chemical combustion flame." There is however an intrinsic contradic- tion between the plasma and the molecular information required.Plasma spectrometry needs a cloud of individual atoms (or atomic ions) in order to provide qualitative and quantitative information on trace metals present in the matrix being studied. As a consequence the structure of the compound has to be destroyed and the molecular information is lost during the determination stage. The paradox is that the sensitivity attainable makes plasma techniques the only ones which at the moment are capable of answering the demands of environmental and clinical trace metal species-selective analysis. *Presented at the 1996 Winter Conference on Plasma Spectrochemistry Fort Lauderdale FL USA January 8-1 3 1996. Spectrometry Information on trace-metal speciation can be addressed at three levels which is schematically illustrated in Fig.1. They include ( i ) time-resolution of analyte species prior to their entering the plasma; ( i i ) empirical formula determination; and (iii) controlled fragmentation of polyatomic ions. The isotopic or wavelength snapshot contributes to the confirmation of the identity of the target element. Most attention has been given to on-line chromatographic or electrophoretic separation of analyte species prior to the plasma. This has resulted in a new generation of analytical techniques (termed hyphenated tandem coupled and hybrid) which are able to discriminate between the different forms of a metal or metall~id."-'~ The increasing need for species-selective analytical information and the emerging market for instrumentation capable of delivering this information are making speciation analysis one of the fields capturing most interest in modern analytical chemistry.The purpose of the present paper is to discuss briefly the status of plasma techniques in trace element species-selective analysis of environmental and biological materials and to indicate the most important trends in development. DISCUSSION Species-selective Analysis Using Plasmas In species-selective analysis the plasma is used only at the final determination stage. Therefore the signal obtained carries information pertinent to the element alone but not for the whole compound. The key to a successful analysis is the quality of the separation of the species prior to their entering the plasma.Identification is based on a parameter that is external to the plasma; this is usually retention time in the case of Signal f intensity I Molecular m h ratio Empirical formula \ \ Elemental m/z ratio or emission wavelength Fig. 1 Pathways to molecular information in spectrochemical analysis Journal of Analytical Atomic Spectrometry September 1996 Vol. 11 (871 -876) 871chromatographic methods or electrophoretic mobility in the case of capillary electrophoresis. Simultaneous multiwavelength (multimass) monitoring enables the determination of several species containing differ- ent metals (metalloids) during one chromatographic run.In addition it renders possible the acquisition of more information on the eluted species in terms of the presence of other analyte lines (confirmation of elemental identity) and the mutual ratio of the constituent elements in the detected compounds (deter- mination of empirical formula). Gas chromatography with plasma detection The attractive features of capillary GC such as the high resolution and the availability of plasma source detectors with absolute detection limits at the low femtogram level have made GC with plasma source detection the technique of choice for many applications especially those related to anthropog- enic environmental contaminants. This has been supported by the commercial availability of a GC-MIP-AES system which can be easily tuned to the desired ~avelength.l~-~* The principal disadvantage of GC is the usual need for conversion of analytes into thermally stable and volatile species (derivatization).This limits the field of application to a restric- ted number of compounds and has largely contributed to GC-based techniques reaching a stage of maturity in relation to speciation analysis. The novel aspects of work now appear- ing usually concern newer designs of the interface between the GC and ICP-MS instruments on-line preconcentration prior to GC and developments at the sample preparation stage. The basic advantage that ICP-MS offers over MIP-AES is its higher tolerance to water vapour and carbon dioxide. This is particularly critical in the determination of gaseous trace contaminants in air following their cryogenic preconcentration. Cryotrapping in liquid nitrogen is very efficient but it is not selective for the analyte compounds; the thermodesorption process often does not discriminate between the organometallic analytes and matrix interferents (COz H20).Purge and trap thermal desorption ICP-MS for the determination of volatile organometallics answers the demand for a simple robust and automated system for the analysis of the gaseous air fraction for organometallic compounds. A chromatogram is shown in Fig. 2 of urban air sampled on a busy car park indicating the presence of tetraalkyllead and volatile mercury compounds. Liquid chromatography with plasma detection HPLC has usually been used in this context but a series of low-pressure columns for the speciation of redox forms and some organometallics has recently been marketed.21 Me,,Pb (2.62 ng mS) A 175 000 i 150 000 3 125000 I/ Etpb (i.5 ng m-3) II I \ s 2 loo 0 75 50 25 M%Et,Pb 000- - Pb 0 1 2 3 4 5 6 Time/min Fig.2 Chromatogram obtained for urban air (Bordeaux area) by cryotrapping (- 196 "C)-thermal desorption GC-ICP-MS; mass spec- trometer; SCIEX ELAN 5000 (Perkin-Elmer); and 0.1 m3 of air ~ I Serum 168 126 2 84 42 5 .- E v) c c .- - CT) I I I I 4.277 5.348 1.069 2.139 3.208 0' Retention time/min Fig.3 Chromatogram of a human serum sample obtained with reversed-phase HPLC-ICP-MS by monitoring m/z = 82 (selenium). Mass spectrometer SCIEX ELAN 5000 (Perkin-Elmer); column PRP-1 (Hamilton); sample diluted five times with water; injection volume 100~~1; mobile phase 2% methanol in water lOmmoll-' C,H,,S03Na at pH 4.5; flow rate 1 ml min-'.The retention time of the signal corresponds to selenocystine The need for desolvation of the mobile phase makes ICP the only source able to cope directly with this problem. ICP-MS is more popular than ICP-AES because of the much higher sensitivity and has been comprehensively evaluated as an element-selective LC detector by the group of Caruso.16 The major drawbacks of ICP-MS are the vulnerability to salt and organic solvent content of the LC mobile phase and polyatomic ion interferences which hamper ICP-MS analysis for transition metals (m/z<80). Axial ICP-AES can offer a cost-effective alternative to the light transition metals. In addition the tolerance of ICP-AES to the salt and organics load is higher than that of ICP-MS because of the higher plasma powers used and the absence of the sampler-skimmer interface.Further development of a high-resolution ICP-MS detector is still limited by the high cost of the instrumentation. The present challenge in HPLC-ICP-MS coupling is how to improve the sensitivity of detection for metal-containing compounds; conventional nebulization can hardly be con- sidered suitable because the losses in the nebulizer are too large which thus degrades the sensitivity. This technique is sufficient for some applications however as demonstrated in Figs 3 and 4. A high-pressure nebulizer (HPN) with a desolv- ation chamber is more s ~ i t a b l e . ~ ~ ' ~ Even more efficient .- t I I I Retention time/s Fig.4 Chromatographic profile of copper in a commercial mixture of metallothioneins (Sigma Rabbit liver No. M7641). 1 Copper bound to metallothionein fraction; and 2 Cu2+. Column Progel-TSK PWXL (4.0 cm x 6.0 mm); 30 ng ml-' of thionein (as Cu); mobile phase 30 mmol 1-' TRIS-HC1 pH 7.1 0.7 ml min-'; sample 50 pl injected in 22 mmol 1-' 2-mercaptoethanol in mobile phase. Mass spectrometer SCIEX ELAN 6000 872 Journal of Analytical Atomic Spectrometry September 1996 Vol. 11(although at the expense of the amount of sample handled) are the microconcentric (MCN) and direct injection nebuliz- ers (DIN). The MCN is a fixed concentric nebulizer that fits easily onto a conventional spray chamber. It operates at low sample flow rates (0.03-0.3 ml min-l) using normal aerosol carrier gas pressures (80 psi; 1 psi = 6894.76 Pa) and performs well as the interface between low pressure chromatography and ICP-MS.Shown in Fig. 5 is an example of the speciation of toxic arsenic forms in canned crab meat leachate. The DIN. is a microconcentric nebulizer with no spray chamber; it nebulizes the liquid sample directly into the central channel of the ICP torch. Compared with conventional nebuliz- ers it offers 100% sample introduction efficiency at flow rates of 10-100 pl min-'. The low dead volume (<2 pl) and the absence of a spray chamber minimize post-column peak broad- ening and memory effects. The DIN provides absolute detec- tion limits that are superior by 1-2 orders of magnitude to those obtained with conventional nebulizers. Concentration detection limits were however not Another advan- tage is that the sampler and skimmer orifices do not clog to any great extent with carbon.24 A wider range of applications of microbore HPLC-DIN-ICP-MS coupling in clinical chem- istry are clearly imminent.Reliability of Molecular Information The reliability of the information obtained is based on the reliability of the atomic information obtained and on the quality and reproducibility of the separation. Atomic information in plasma spectrometry That the signal measured definitely corresponds to the expected element being measured is crucial. Although the selectivity of plasma source spectrometers used as chromatographic detec- tors is superior to flame ionization conductivity or UV detec- tors a few concerns do arise.In AE detection the confirmation of elemental identity requires the ability to carry out a wavelength scan at the peak 1 2 0 200 400 600 800 1000 Retention time/s Fig. 5 Speciation of arsenic in commercial canned crab meat leachate by low-pressure anion-exchange chromatography-MCN-ICP MS. Mass spectrometer; HP 4500 (Hewlett-Packard); column ANX3206 (CETAC Omaha USA); sample in OSmmoll-' malonic acid; load time 55 s. Elution program 0-300 s 0.9 mmoll-' malonic; 300-400 s 5 mmol 1-' malonic acid pH 8; above 400 s 50 mmol 1-' malonic acid pH 8. Mobile phase flow rate 30 pl min-'. 1 As"' (ca. 5 ng ml-'); and 2 As" (ca. 0.5 ng ml-') apex. In the Hewlett-Packard AE detector a 40nm wide emission spectrum can be taken at the peak apex and checked for the presence of the analyte characteristic atomic emission lines and their ratio (fingerprint).20 In MS validation is obtained by the isotopic pattern.Problems can occur for monoisotopic elements e.g. arsenic which is known to suffer from an overlap by the ArCl' ion. High-resolution MS is the expensive alternative. The potential advantages of simpler background spectra with an He plasma remain largely unful- filled. In fact NO' and 02+ problems are even more severe because of the lower gas flow rates and easier entrainment of air in many He plasmas. Also the lower temperature can in some cases prevent dissociation of polyatomic ions. The most reliable proof of identity would be obtained by splitting the effluent to both ICP-AES and ICP-MS instru- ments which are considered orthogonal techniques.However apart from the fact that few laboratories have both techniques coupled to chromatography the detection limits of ICP-AES may not be sufficient for some analytes and also some analyte elements may not be determinable by ICP-MS. Compound identijication In the simplest case the identification of analyte compounds is based on the retention times of their corresponding standards. This can be influenced by the column matrix and separation conditions but within the same laboratory an acceptable reliability can be achieved. Standards however are not gener- ally available with the exception of the case of a small number of anthropogenic pollutants but not necessarily products of their degradation as well. Some compounds can be custom- synthesized and the use of MS (electron impact chemical ionization or electrospray) to confirm their structural identity is then mandatory.In practice there is often a need for the prediction of retention times by interpolation or extrapolation of those obtained for the standards available. In GC the use of the Kovat's retention index is a valid choice. Actual retention times of the mixed compounds can be confirmed using transal- kylation mixtures. Once identified the compound should be synthesized and its actual retention time then confirmed. Quantification of the mixed compounds is made by interp- olation of the signal intensity using the standards available. Discrimination of the response with the boiling-point of the analyte must be taken into account.In metal-containing pro- tein analyses a role similar to that of the Kovats index is played by the use of relative molecular mass markers in size- exclusion chromatography. Attribution of the relative molecu- lar masses to the proteins was performed after calibration of the column with for example myoglobin (17.8 kDa) thyroglo- bulin (669 kDa) alcohol dehydrogenase (150 kDa) albumin (67 kDa) which act as the relative molecular mass standards.25 The reliability of identification based on the retention time is highly dependent on the quality of separation. The better the separation the lower the probability that an unknown and unaccounted for compound elutes at the retention time of a particular analyte compound. An example is given in Fig. 6 that illustrates the decreasing possibility of interference when moving from packed to capillary columns.The chromatogram obtained using a capillary column shows between the peaks of tetrabutylin (SnBu,) and tripentylmethyltin (SnMePe,) a well separated small peak (dipentyldipropyltin SnPr,Pe,) due to a standard impurity. This peak was not shown by packed column GC analysis. Whereas the presence of an unknown compound is unlikely to be a serious source of mis-identification in the case of analysis for anthropogenic organometallic pollutants it does become a serious problem when the complexity of the species (relative molecular mass) increases and a variety of similar Journal of Analytical Atomic Spectrometry September 1996 VoZ. 11 873molar ratios of the molecular structure of the compound which is usually not the case as has been demonstrated by Huang et a1.27 Another problem is the need for an undisturbed compound band in the plasma at a given moment and the need for relatively sensitive detection of all the compound- forming elements at compromise plasma operating conditions.Complexity of the matrix leads to a high background especially on the carbon and hydrogen channels which affects the precision and often does not allow for element quantification. Under optimum conditions typical precision values vary between 2 and 5% which works well for the calculation of the formula of compounds with elemental ratios equal to whole numbers but fails for higher hydrocarbons e.g. C2oH44 where a 5% error can easily change the resulting calculated formula.The principal area of applications is low relative molecular 0 1 3 5 7 9 1 1 - 3 5 1 9 1 1 mass halogen- and sulfur-containing compounds. In general empirical formulae can be determined only for standard solu- tions as has been demonstrated for tin.28 The method fails for 5 1 l 1 1 1 Retention time/min Fig.6 Effect of resolution of the separation on the accuracy of AAS; and (b) capillary GC-AES (HP-1 column). 1 Me2Pe2Sn; 2 Pr,Sn; 3 Pr,PeSn; 4 Bu,Sn; 5 MePe,Sn; 6 Bu,PeSn; 7 Bu2Pe2Sn; 8 BuPe,Sn; and 9 Pr,Sn. Me methyl; Pr propyl; Pe pentyl; Bu butyl unknown compounds that are present in extracts of real determination of carbon and hydrogen owing to the very high backgrounds Present and low concentrations of the species of interest. Although precision cannot match that of classical microanalysis the measurements can be carried out on-line and considerably lower amounts of sample are required for analysis (pico- or nanograms instead of milligrams).analysis. (a) Packed (3y0 OV-l On Chromosorb)-quartz furnace samples. The reason is the virtual impossibility of an accurate structures could exist (metal-binding protein analysis). For this reason there is a need for orthogonal (independent) separation techniques for the same applications. The most suitable choice seems to be the parallel use of GC and HPLC (for low relative molecular mass compounds) and that of HPLC and capillary Alternative Ion Sources for Species-selective MS Analysis zone electrophoresis-(CZE) for high relative molecular mass compounds. An alternative approach is the confirmation of the molecular identity by MS or IR spectroscopy.However some incompati- bility in terms of the detection limits and matrix interference can occur. The advances in the sensitivity of electron impact MS for GC detection and electrospray MS for LC and CZE detection are likely to create valuable auxiliary tools for species-selective analysis. Since both the ICP and the MIP usually destroy any molecular structure and focus on the element itself the quest for other sources is on-going. The glow discharge has been reported to ionize a sample to a sufficient degree for MS detection at ultratrace levels and also to be capable of providing structural information at these levels.29 An alternative is the modification of atmospheric pressure ionization (API) techniques which have become a landmark in the field of LC detectors in organic analysis.These tech- Quantitative information Regarding the quantitative information required it is possible that a species will respond differently in the plasma depending on the organic part of the molecule. Internal standard cali- bration could thus turn out to be inappropriate. Use of a calibration curve or the method of standard additions requires the use of standards which are rarely available. A method to solve this problem is on-line conversion of the analyte species into forms which would respond in the same way. An example is the use of post-column HG for the determination of arsenic and selenium species. Harsh UV radiation thermochemical or microwave-assisted techniques are required in order to convert all of the species containing the element in question into the same form and thus to assure a species-independent response in the plasma.26 Empirical Formula Determination by Plasma Spectrometry The atomic information obtained as consequence of the plasma-induced radiation (or ion current) can be readily translated into molecular information in terms of the empirical formula of the compound.For this the mutual concentration ratios for all elements are necessary which in practice means the need for the simultaneous determination of all the elements that form the compound in question. The need for the quantification of a non-metal (carbon hydrogen sulfur and halogens) limits the choice of detectors to those operating with an He plasma and in particular to MIP-AES.The key to success is the independence of elemental niques produce gas-phase analyte ions directly from solution by using electrospray or ion-spray (a pneumatically assisted version of electrospray) interfaces. The problem is that under the operating conditions usual in commercial instruments the ionization energy is not sufficient to achieve efficient fragmen- tation down to the elemental ion. Recently a combination of modifications to the curtain gas the declustering potential and the geometry of the vacuum interface region of the mass spectrometer was reported to enhance the fragmentation effi~iency.~' In such a system ions from the spray chamber are sampled into the MS system through a dry N2 curtain gas which assists in the declustering of ion-solvent aggregates.The degree of adjustability of the potential difference between the orifice and skimmer (referred to as the declustering potential) allows control of the declustering reactions in the atmosphere through to the vacuum region of the mass spectrometer. As a function of this potential the system is able to operate either as an elemental analyser or as a molecular detector. A spectrum of selenomethionine obtained in the molecular and elemental mode taken at the apex of the chromatographic peak is illustrated in Fig. 7 . In Fig. 7(a) it is shown that in the molecular mode there is a noisy background and poorer sensitivity than in elemental mode [Fig. 7 ( b ) ] . The modification of the instrument allowing for its running in the elemental mode still needs to be optimized as shown by the residue of the molecular peak at m/z 196.The availability of information on the relative molecular mass of analyte species and further (by placing a second MS) fragmentation information are likely to make electrospray (ion spray) atmospheric pressure ioniz- ation (AP1)-MS a serious competitor especially for elements that suffer serious interferences in ICP-MS (e.g. As Se and Cr). 874 Journal of Analytical Atomic Spectrometry September 1996 Vol. 116x105 5x105 4x105 3x105 T~ 2x10~ 2 1x10~ G 5 100 120 140 160 180 200 A 'z 122.0 145.8 168.0 ':to 2.0X1O6 - i 90 120 150 180 m/z Fig. 7 Snapshot on the selenomethionine peak in an HPLC chroma- togram by ion spray MS. Sample selenomethionine (50 pg ml-' in H,O); Mass spectrometer modified API-300 (SCIEX); column LichroCART 250-4 (Merck); mobile phase 50% methanol 1 ml min-' Split 1 200; post-column acidification with HC1 to pH 3.4.(a) Molecular mode (orifice voltage 40 V) and (b) elemental mode (orifice voltage 400 V) Towards Analysis of Real-world Materials Very few studies have provided data on 'real-world' materials with reasonably well-substantiated analytical accuracy. Whereas few problems exist with water speciation analysis of sediments and biomaterials requires isolation of analytes from the matrix without changing their chemical form. As the organometallic species in sediments are not built into the crystal lattice leaching is the approach of choice. They can however be incorporated in tissues of a living organism so biological materials must be solubilized.Alkaline hydrolysis with tetramethylammonium hydroxide (TMAH) is faster and cheaper than enzymatic hydrolysis with a mixture of lipase and protease. Until recently the procedures reported for sediments have not only been extremely time consuming (they take from 1 h to 2 d ) but also are usually inefficient in terms of recovery of analytes and are unreliable. As shown by Zhang et ~ l . ~ ' only three out of ten sample preparation methods described in the literature for the analysis of sediments were able to recover more than 90% of tributyltin (TBT) whereas none were able to recover monobutyltin (MBT) in a non-erratic and reproduc- ible manner. High scattering of the results due to leaching problems prevented certification of MBT by the Community Bureau of Reference (BCR).32 However it would seem that a lot of these problems can be solved by the recently proposed application of low-power focused microwave-assisted technol- ~ g y .~ ~ In contrast to the common high-temperature and press- ure acid attack a focused low-power microwave field was demonstrated to be successful for quantitative extraction of organometallic species from the matrix without destroying the carbon-metal bonds.34 Also in the case of the species-selective analysis of biological materials the impact of recent developments in microwave technology for solvent extraction leaching and enhancement of extraction kinetics with the simultaneous preservation of the organometallic moiety and thus structural information is a milestone regarding the sample preparation step.Microwave assisted TMAH hydrolysis allows dissolution of a biotissue within a few minutes without degradation of the a n a l y t e ~ . ~ ~ Studies however have been restricted to small organotin and 8o 1 2 5 I 6 Time/min Fig. 8 Capillary GC-AES chromatogram of a fish tissue (NIES 11) extract obtained after simultaneous acetic acid hydrolysis ethylation with NaBEt and extraction into nonane in a microwave field (3 min). Column DB-210. 1 BuEt,Sn; 2 Pr,EtSn; 3 Bu,Et,Sn; 4 Bu,EtSn; and 5 Bu4Sn organomercury compounds. The real challenge lies in the analysis of high relative molecular mass compounds often with unknown structures which require not only sophisticated and up-to-date separation procedures but also their efficient isolation from the matrix.Hitherto sample preparation pro- cedures have been based on leaching with water or a neutral buffer followed by (ultra) centrifugation with recoveries of 50-80%. While the identity of the leached compounds often raises some questions virtually nothing is known of the identity of the compounds remaining in the solid phase. Another challenge is the integration of sample decompo- sition derivatization of polar analytes and their extraction into a non-polar solvent. Hydrolysis with acetic acid carried out in a low-power focused microwave field in the presence of sodium tetraethylborate and nonane was recently shown to shorten radically the sample preparation time for GC determi- nation of organotin compounds in biological materials.36 After a 3 min reaction time ethyl derivatives of mono- di- and tributyltin and triphenyltin were quantitatively found in the supernatant organic phase that was injected onto a capillary GC column (Fig.8). The use of a slightly polar column instead of the commonly recommended non-polar phases allowed a clean-up step to be avoided and thus to shorten the overall procedure. CONCLUSIONS The rapidly growing interest to plasma spectroscopists in speciation of elements has been followed by a rapid develop- ment of more sophisticated analytical techniques. Plasma (both ICP and MIP) techniques have an already established position in the field of species-selective analysis. They are considered to be the best detection option for modern separation tech- niques.The molecular information accessible is however based on the retention time of the metal-containing compounds which implies the availability of standards. Simultaneous AE detection permits the acquisition of data on carbon hydrogen sulfur and the halogens of the same time as that for a metal and metalloid and thus offers the possibility of structure identification. Plasma MS techniques are evolving towards the possibility of controlling the fragmentation of a species of interest so that at least some molecular information is retained. On the other hand recent advances in organic LC-MS make imminent the advent of a molecule smashing device to obtain the bare metal. Where will these approaches meet? Journal of Analytical Atomic Spectrometry September 1996 Vol.1 1 875REFERENCES 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Metals and their Compounds in the Environment ed. Merian E. VCH Weinheim 1991. Trace Element Speciation Analytical Methods and Problems ed. Batley G. E. CRC Press Boca Raton 1987. Metal Speciation Theory Analysis and Application eds. Kramer J. R. and Allen H. E. Lewis Chelsea 1988. Environmental Analysis Using Chromatography Interfaced with Atomic Spectroscopy eds. Harrison R. M. and Rapsomanikis S. Horwood Chichester 1989. Trace Metal Analysis and Speciation ed. Krull I. K. Elsevier Amsterdam 1991. Chau Y. K. Zhang S. and Maguire R. J. Analyst 1992 117 1161. Donard 0. F. X. and Martin F. M. Trends Anal. Chem. 1992 11 17. Caroli S. Microchem. J. 1995 51 64.tobinski R. in 1995 Colloquium on Analytical Atomic Spectrometry ed. Welz B. Perkin-Elmer Uberlingen 1996. Inductively Coupled Plasmas in Atomic Spectrometry eds. Montaser A. and Golightly D. W. VCH New York 2nd edn. 1992. Ebdon L. Hill S. and Ward R. W. Analyst 1986 111 1 113. Ebdon L. Hill S. and Ward R. W. Analyst 1987 112 1. Cappon C. J. LC-GC 1987 5,400. Element-specijic Chromatographic Detection by Atomic Emission Spectroscopy ed. Uden P. ACS Washington DC 1992. Hill S. J. Bloxham M. J. and Worsfold P. J. J. Anal. At. Spectrom. 1993 8 499. Vela N. P. Olson L. K. and Caruso J. A. Anal. Chem. 1993 65 585A. tobinski R. Analusis 1994 22 37. Szpunar J. Witte C. tobinski R. and Adams F. C. Fresenius’ J. Anal. Chem. 1995 351 351. 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 Lobinski R. and Marczenko Z.Spectrochemical Trace Analysis for Metals and Metalloids Elsevier Amsterdam 1996. tobinski R. and Adams F. C. Trends Anal. Chem. 1993,12,41. Gjerde D. T. Wiederin D. R. Smith F. G. and Mattson B. M. J. Chromatogr. 1993 640 13. Berndt H. Fresenius’ Z . Anal. Chem. 1988 331 321. Berndt H. and Schaldach G. J. Anal. At. Spectrom. 1994 9 39. Shum S. C. K. and Houk R. S. Anal. Chem. 1993,65,2972. Nagase M. Kondo H. and Hasebe K. Analyst 1995 120 1923. Rubio R. Alberti J. Padro A. and Rauret G. Trends Anal. Chem. 1995 14 274. Huang Y.-R. Ou Q.-Y. and Yu W.-L. J. Anal. At. Spectrom. 1990 5 115. tobinski R. Dirkx W. M. R. Ceulemans M. and Adams F. C. Anal. Chem. 1992 64 159. Caruso J. ICP In$ Newsl. 1996 21 60. Corr J. J. and Anacleto J. F. Anal. Chem. 1996 68 2155. Zhang S. Chau Y. K. Li W. C. and Chau A. S. Y. Appl. Organomet. Chem. 1991 5 431. Quevauviller Ph. Astruc M. Ebdon L. Desauziers V. Sarradin P. M. Astruc A. Kramer G. N. and Griepink B. Appl. Organomet. Chem. 1994 8 629. Donard 0. F. X. Lalere B. Martin F. and tobinski R. Anal. Chem. 1995,67,4250. Szpunar J. Schmitt V. Donard 0. F. X. and Lobinski R. Trends Anal. Chem. 1996 15 181. Szpunar J. Schmitt V. O. tobinski R. and Monod J.-L. J. Anal. At. Spectrom. 1996 11 193. Rodriguez Pereiro I. and Schmitt V. O. personal communication. Paper 6/00819D Received February 5 1996 Accepted May 8 1996 876 Journal of Analytical Atomic Spectrometry September 1996 Vol. 1 1