Anal. Chem. 199466, 462R-499R
Plasma Emission Spectrometry? Diane Beauchemin,’ J. C. Yves Le Blanc, Gregory R. Peters,* and Anthony T. Persaud Department of Chemistry, Queen’s University, Kingston, Ontario, Canada K7L 3N6 Review Contents Introduction Conferences Books and Reviews Sample Pretreatment Digestion Extraction/Separation Other Methods Sample Introduction Liquid Sample Introduction Solid Sample Introduction Gaseous Sample Introduction Atomization and Excitation Sources Spark Discharge Direct Current Plasma (DCP) Microwave-Induced Plasma (MIP) Inductively Coupled Plasma (ICP) Glow Discharge (GD) Hollow Cathode Discharge (HCD) Laser-Induced Plasma (LIP) Detection Systems Data Acquisition, Processing, and Simulation Calibration Kalman Filtering Other Multivariate Techniques Line Selection Plasma Modeling
462R
465R
468R
485R
J. C. Yves Le Blanc received his B.Sc.
493R 494R
A. INTRODUCTION Since our last and first fundamental review ( A I ) following the impressive work of the P. N. Keliher group (A2), a few changes occurred in our group. Jane M. Craig decided to concentrate on completing her Ph.D. thesis (on mixed-gas plasmas and segmented flow injection for inductively coupled plasma mass spectrometry (ICP-MS)) that she successfully defended on January 27,1994. Anthony Persaud (who started in September 1992) took her place. On the other hand, Gregory R. Peters remained part of the team even though he completed his Ph.D. on March 3, 1994 (on novel sample introduction strategies for inductively coupled plasma mass spectrometry). (In case the reader is wondering why so many details, it so happens that these two new Ph.D. graduates were the first ones supervised by D.B.!) We heard relatively few comments about our last review. Fortunately, they were all positive, so we decided to carry on. As previously, the reader is referred to the review, which is also in this issue, entitled Atomic Absorption, Atomic Emission and Flame Emission Spectrometry by K. W. Jackson for emission spectrometry using flames or electrothermal atomizers. Examples of hybrid techniques combining a graphite ‘This review is dedicated to the memory of Dr. Sc. Klaus Dittrich, who unexpectedly died of a heart attack on December 2, 1993. Present address: Fenwick Laboratories Ltd., 5595 Fenwick St., Suite 200, Halifax, NS B3H 4M2.Canada.
*
462R
Dlane Beauchemlnis Associate Professor of Chemistry at Queen’s University. She received a B.Sc. in Chemistry from the Universit6 de Montr6al and stayed on to obtain a Ph.D. in Analytical Chemistry at the end of 1984 (under the supervision of Joseph Hubert). She then worked as a Research Associate in the Chemistry Division (now the Institute for Environmental Chemistry)of the National Research Council of Canada, in collaboration with Jim W. Mclaren. She moved at Queen’s in July 1988 and was promoted in July 1993 (and granted tenure as of July 1, 1994). She is actively involved in extending the capabilities of inductively coupled plasma mass spectrometry (ICPMS) by taking several approaches, either individually or in combination. These include flow injection techniques (on-line pretreatment, on-line preconcentration,hydride generation, etc.), mixed-gas plasmas, coupling to chromatography (both gaseous and liquid), and slurry nebulization. The techniques are used for both environmental and medical applications.
AnalyticalChemistry, Vol. 66, No. 12, June 15, 1994
degree in chemistry from the Universit6 de Montr6al in 1989. He then completed a M.Sc. degree on electrospray mass spectrometry (1992) at Queen’s University, under the joint supervision of Professor Beauchemin and Dr. K. W. Michael Siu of the Institute for Environmental Chemistry (NationalResearchCouncilof Canada). He is currently working toward his Ph.D. degree at the same university, and his major interest is liquid sample introduction systems for both ICP and mass spectrometry instruments. Gregory R. Peters is currently employed as a postdoctoral fellow with Fenwick Laboratories in Halifax, NS,having recentty c* completed his Ph.D. at Queen’s University under the supervision of Professor Beauchemin. He received his B.Sc. (Honours Chemistry, Minor Nuclear Science) degree from Simon Fraser University (British Columbia, Canada) in 1989. During the course of his undergraduate studies, he was employed by numerous govemment and industrial laboratories as part of the cooperative education program at SFU. As / B a Ph.D. student, Greg’s interests focused (Ir on altemate sample introduction strategies ration for ICP-MS (including gas chromatograph preevaporation). He is continuing his research on sample introduction methods for ICP-MS and ICP-AES at Fenwick. Anthony T. Persaud is finishing his M.Sc. on geological slurry analysis by ICP-MS at Queen’s University under the joint supervision of Professor Diane Beauchemin, Professor Heather E. Jamieson, and Professor Robert J. C. McLean. His B.Sc. in geology and chemistry was completed at the University of Toronto in 1992.
furnace and a plasma include furnace atomization plasma emission spectrometry (FAPES), furnace atomization nonthermal excitation spectrometry (FANES), and laser-induced plasma in an electrothermal atomizer. Similarly, the reader is referred to the review Atomic Mass Spectrometry by David 0003-2700/94/0366-0462$14.00/0
0 1994 American Chemical Society
W. Koppenaal (also in this issue) for developments pertaining strictly to ICP-MS. The present review will cover the plasma emission/fluorescence spectrometry developments which occurred during 1992 and 1993 (as well as some late 1991 ones which did not get covered in the last review). Again, we have made a point of only reporting on papers that we had thoroughly read, so the review will by no means provide an all-inclusive bibliography. In these days of recession, most libraries do not hold as many journals as they used to do. This is also true for those to which we have access. We therefore found it more difficult to get a hold of interesting papers. In order to facilitate our task for the next review, we would be very grateful if authors would send us reprints/ photocopies of their pertinent work as they become available. Abstracts are fine for an overview but they are not without mistakes. For instance, D.B. wrote a review entitled A Comparison of ICP Atomic Spectrometry Techniques ( A 3 ) which was described by Chemical Abstracts as being authored by N. J. Miller-Ihli, the Contributing Editor who invited D.B. to write this paper! Even if plasma emission spectrometry is no longer regarded as a “hot” area in analytical chemistry, research is nonetheless continuing because a better understanding of the underlying measurement principles is still needed, together with effective methods and improved instrumentation (A4). (Actually, this last publication is recommended reading to anybody who does not know what analytical chemistry is!) This review will therefore report significant (or potentially significant) developments on sample pretreatment, followed by those on sample introduction, plasma sources, detection systems, and, finally, data processing. We have tried to keep our coverage as selective and critical as in the last review. Easier and faster ultratrace analysis was really the goal of most of the developments which occurred during 1992 and 1993. This is not surprising, given the ever-increasing needs for ultratrace analysis of environmental and high tech materials. (By the way, as truly pointed out by Ortner ( A 5 ) , “ultratrace analysis” still awaits a clear definition.) Analytical laboratories have to perform accurate and reproducible analyses of a varying nature (in terms of analytes, matrix, etc.) at a low cost in order to remain competitive (A6). Selecting cost-effective instrumentation and then developing adequate sample preparation procedures are major concerns of these laboratories. For instance, Ortner (AS) said that although ICP-AES could not match most of the detection limits of ICP-MS, it could still be quite appropriate for a number of applications, especially following a matrix-separation method to reduce spectroscopic interferences. Several examples of the latter are included in section B. Gill ( A 7 ) discussed the dilemma of the AAS/ICP-AES choice, appropriately pointing out that numerous factors must be taken into consideration, aside from the cost of the instrument itself. These include the number of elements and their identity, the number and types of samples, and the detection limit required. Obviously, each of these techniques has advantages compared to the other. This is why both of them are well established; Le., ICP-AES has not replaced AAS. Nonetheless, the stringent cost-effective requirement has forced instrument manufacturers to build highly reliable user-friendly instruments which do not require highly trained personnel and can,
more and more, be operated unattended. Some examples of these developments are included in this review. The comments which were made on the previous review almost unanimously concerned two paragraphs of the introduction, Le., those which dealt with two forms of dishonesty: not appropriately quoting previous work or reporting the same results as new results more than once (there is another appalling exampleof the latter in section D). It appeared that these are the pet peeves of many people and not only those of the principal author. The first one in particular touched a very sensitive string, which D.B. would like to “vibrate” a little more. In the previous review, the onus was put on the authors and the reviewers to ensure that appropriate references were made. This still holds but one more “player” must take part in this game for it to be thoroughly fair: the Editor of the journal where the work is submitted for publication. Indeed, while acting as a reviewer, D.B. noticed on a few occasions that papers appeared with essentially no revisions even though she had specifically requested several changes. It is very discouraging to spend many hours thoroughly reviewing a paper if the Editor does not ensure that the necessary changes are made. Editors are bound to be scrupulously objective; Le., they should not allow any interference, such as the potential changes in readership and/or the number and type of authors. During this review, as in the last one, we noticed that detection limits were often reported without mention of the associated number of replicates or integration time. Yet, the standard deviation, on which detection limits are based, depends on the number of measurements (n), especially if n is small. To better illustrate how widespread this “omission” is, all the tables quoting detection limits also include n when known. Most authors followed IUPAC’s recommendation (A8). This definition is very convenient for those working with solutions but not for those dealing with solids where true blanks are often unavailable. In contrast, the definition proposed by Boumans (see refs A4-AS of ref A l , as well as ref F12 of this review) is more universal since it is based on the background emission rather than on that of a blank. Another disturbing trend also became evident during this review: many authors do not include a complete description of their instrumental setup. Too often, important details are missing. This is also illustrated by several examples in the text. Conferences. The second pet peeve mentioned above can certainly be extended to conferences. Too many people proceed from one conference to another to make essentially the same presentation@) (plenary presentations are excluded since they are usually expected to be of a reviewing nature) on the basis that part of the audience will be different. What about that part of the audience who already heard it? The least that the speaker could do is to briefly summarize what was presented at the previous conference(s) and carry on with new material. In this way, both those who never heard the preceding talk(s) and those who did benefit. This problem is actually a “secondary effect” of there being too many conferences. They end up being very close to one another (or even overlapping!), which makes it very difficult to have new material to present to many of them (assuming that the speaker is very rich to afford all these trips or that he/she is invited, with part or all of his/her expenses covered). For an up-to-date list of all Analytical Chemism, Vol. 66, No. 12, June 15, 1994
463R
Table 1. Books Related to Plasma Eml8slon/Fluorrrcence Spectrometry
title
authors and/or editors
publisher
Flow Injection Separation and Preconcentration Z. Fang Element-Specific Chromatographic Detection by P. C. Uden, Ed. Atomic Emission Spectroscopy; ACS Symposium Series 479 L. H. J. Lajunen Spectrochemical Analysis by Atomic Absorption and Emission A. Montaser and Inductively Coupled Plasmas in Analytical D. W. Golightly Atomic Spectrometry, 2nd ed. J. Sneddon, Ed. Advances in Atomic Spectrometry, Vol. 1 M. Valcarcel and Non-chromatographic Continuous Separation Techniques M. D. Luque de Castro I. S. Krull, Ed. Trace Metal Analysis and Speciation; Journal of Chromatography Library 47 K. A. Smith, Ed. Soil Analysis. Modern Instrumental Techniques, 2nd ed. J. C. Van Loon and Determination of the Precious Metals: R. R. Barefoot Selected Instrumental Methods E. R. Malinowski Factor Analysis in Chemistry, 2nd ed. Principles and Practices of Spectroscopic H. Mark Calibration
plasma-related conferences, the reader is referred to the ICP Information Newsletter, edited by Ramon M. Barnes. This publication also regularly contains abstracts of papers presented at numerous conferences. The principal author has given a lot of thought to the problem and has come up with three possible solutions. (1) Each “faulty” speaker acquires the discipline to only present new material (unless it is specifically a review), which also means that he/she learns to say “non when he/she is invited but does not have anything new to say. From the number of repetitions currently occurring, this option, although the simplest, is obviously difficult to implement. An alternative would be for the author to indicate in the abstract where the same presentation was made so that those who already heard it can attend other talks without fear of missing important information. (2) The conference organizers ensure the originality of the material to be presented by only accepting papers which are also submitted to the journal/book where the proceedings of the conference are to be published, Le., a paper (which will be reviewed) must be written for every presentation and proceedings must be published for every conference. (3) Some concertation is made between the various organizers and the number of conferences is reduced. A North American example of a possible reduction would be to combine the Federation of Analytical Chemistry and Spectroscopy Societies (FACSS) meeting (or at least the spectroscopy part of it) and the Canadian Spectroscopy Conference, which are usually held two months apart from one another. The resulting meeting could be called the North American Spectroscopy Conference and be held alternately in Canada and the United States. At the moment, many people attend both meetings and see many repetitions. It is also more and more difficult to find sponsors for invited speakers. Having one conference instead of two would not only facilitate this latter process, but it may also reduce traveling expenses (leaving money for more research leading to more high-quality presentations!). Admittedly, option 2 is the most contentious: although it would definitely eliminate “repetitions”, it would also decrease the number of papers presented altogether as many people present work in progress which is not quite ready for publication. Would this really be a bad thing, however? Some 464R
9
Analytical Chemistty, Vol. 66, No. 12, June 15, 1994
year
related technique
book review A9 A10-Al2
VCH American Chemical Society
1993 FIA 1992 various plasma-AES
The Royal Society of Chemistry VCH
1992 ICP-AES, DCP-AES, A13-Al6 MIP-AES 1992 ICP-AES, ICP-AFS A17, A18
JAI Press The Royal Society of Chemistry Elsevier
1992 ICP-AES, FIA 1991 FIA, HG
A19 A20-A23 A24-A27
Marcel Dekker
1991 MIP-AES, ICP-AES, DCP-AES 1991 ICP-AES, FIA, IC
Wiley
1991 ICP-AES, IC, FIA
A3 1
John Wiley & Sons John Wiley & Sons
199 1 data processing 1991 data processing
A32, A33 A34
A28-A30
people say that the purpose of conferences is to learn about what is going on long before publications appear, if ever; if this is true, maybe the purpose should be rethought. It is not in the interest of anybody to only make a conference presentation since this greatly reduces the likelihood of receiving appropriate credit. However, many people do so as a result of outright laziness. This is supported by a strange conversation that D.B. had at the 1994 Winter Conference on Plasma Spectrochemistry. In answer to her question on why no publication was going to follow the poster presentation, an author simply replied that a poster could be quickly put together whereas writing a paper was too time-consuming because of the mandatory literature search! How can any author make a presentation without even knowing if it is original? It therefore appears that the second pet peeve is indeed very closely related to the first one. From personal experience, D.B. knows that her presentations were a lot more satisfying (in terms of clarity and “density” of the content) when they had already been written up for publication. The writing process (aside from allowing the author to check the originality of his/her work by the necessary literature review which must be carried out in the introduction) forces the author to organize his/her ideas so as to appropriately explain and/ or defend them. There is therefore a very high likelihood that the quality of the presentations would greatly increase if option 2 was implemented, which should certainly make each conference worth the trip! Having proceedings for each conference would also allow people who cannot attend them to still benefit from them. Books and Reviews. Table 1 includes information on several books which have appeared since the last review with references to detailed book reviews written by experts. Some apparent outliers do not specifically talkof a plasma technique but were nonetheless included because of their direct applicability to plasma emission spectrometry. The books listed in Table 1 span a range of readership, from being well suited as an undergraduate textbook (e.g., Spectrochemical Analysis by Atomic Absorption and Emission) to covering one topic in great depth (e.g., Inductively Coupled Plasmas in Analytical AtomicSpectrometry). The latter could have also been titled “Everything that you always wanted to know about the ICP”. It is probably the most important contribution to the field.
The reader will therefore not be totally surprised to find out that, within our group, we affectjonately refer to this book as the "ICP bible"! However, in all fairness, another book may also be an invaluable asset to the ICP user (we have not found any book review and have not seen it yet): the CRCHandbook of Inductively Coupled Plasma Atomic Emission Spectroscopy, by Varma (published by CRC Press, Inc.), which, according to the advertizement, covers everything from the preparation of sample and standard solutions to the optimization and maintenance as well as the selection of a system and includes instrumental parameters for the determination of numerous elements. Jarvis and Jarvis wrote two overlapping reviews on plasma spectrometry for geochemical exploration and the earth sciences (A35, ,436). Instrumentation (Le., sample introduction system, plasma source, and spectrometer), sample preparation, calibration techniques, sources of interferences, etc., are discussed, and ICP-AES is compared to other techniques such as that based on emission from a direct current plasma (DCP-AES) and ICP-MS. We also know of the existenceof a bookon microwave-induced plasmas (MIPS), which was published by Elsevier in 1992, Microwave Excited Plasmas. Plasma Technology 4, edited by Moisan and Pelletier. Jinno edited Hyphenated Techniques in Supercritical Fluid Chromatography and Extraction, which is volume 53 of the Journal of Chromatography Library Series published by Elsevier Science Publishers B. V. in 1992 and includes one chapter (Chapter 9, by Jinno) on the coupling of SFC to ICP-AES and one (Chapter 10, by Luffer and Novotny) on coupling SFC to MIP-AES. Similarly, Boenig edited volume 4 of Advances in Low-temperature Plasma Chemistry, Technology, Applications, which was published in 1992 by Technomic Publishing Co. Inc. and includes a discussion by Reszke and Parosa (pp 1-7) of new microwave cavities and a report by Ismail, Atwee, and Ed-Nadi on the use of a laser to enhance emission in MIPS (pp 9-24).
B. SAMPLE PRETREATMENT Digestion. Kingston and Walter ( B I ) reviewed the status of microwave digestion compared to conventional dissolution techniques, especially for environmental analysis by several atomic spectrometry techniques. They demonstrated that the closed-vessel microwave approach is faster, more efficient (Le., a smaller number of reagents and in smaller quantities are usually required), and achieves more completedissolution than open-vessel acid dissolution. To further facilitate the task of the operator, some microwave digestion systems now come equipped with a notebook computer which already contains a selection of methods for a variety of matrices (B2). These systems are also safer since they include internal probes for continuous monitoring of temperature and pressure, rupture disks being only used as a backup. Zehr (B3) discussed a systematic approach to microwave digestion for elemental analysis. The author stressed the importance of minimizing not only sample size (being careful about homogeneity) but also particle size. Valuable summaries of the conditions which may induce loss of elements (through the formation of volatile species) and the properties of useful dissolution reagents (and mixtures of them) are provided, as well as a list of developed methods for a variety of matrices. TatPr et al. ( B 4 ) showed that, in addition to grain size, other
physicochemical properties (e.g., phase state and specific surface area of Alz03) affected the dissolution time and efficiency of microwave digestion. For instance, samples composed of only a-Alz03 required 363 min for complete dissolution at 170 OC and 207 kPa, whereas 80-90% of other A1203 samples were dissolved within 70 min. Their results also showed that even a relatively small percentage of the a-A1203 substantially increased the dissolution time (2 h for 17% of a-AlzO3) when compared to samples free of a-AlzO3 ( 5 min). Using standard reference materials (SRMs), they also showed that the element recovery (e&, Fe) often matched the matrix dissolution efficiency unless the element was located on the surface (e.g., Ca). The dependence of digestion conditions on phase state was turned into an advantage by Jianzhong et al., who developed an apparatus for the phase analysis of high-temperature superconductors (B5).Selective dissolution was accomplished by stirring ground sample in 0.05% HC1 in a quartz container as dropwise addition of concentrated HCl was made while heating at 2 OC/min. A sampling peristaltic pump was used to retrieve 2-mL aliquots of the digesting solution at discrete time intervals. An internal standard was then added and the solution was partly neutralized (to pH 3) before analysis by ICP-AES. A factorial design can be a very effective tool for parameter optimization by allowing one to extract useful information from a limited number of experiments. Mohd et al. (B6)used a fractional factorial design to optimize microwave digestion of biological samples prior to DCP-AES analysis. A linear equation relating the emission intensity of an element to nine terms was proposed to represent the response surface. A total of 21 experiments was required to determine the parametric coefficients of the various terms, which were used to look at the dependence of the response on individual parameters or the combination of interdependent parameters. The results clearly showed that aqua regia did not achieve quantitative recovery of the analytes. A study (B7)clearly demonstrated a new area of application for ICP-AES: the determination of the diffusion coefficient of Cu in A1 and Ag. Thin films of Cu were electroplated on high-purity (99.999%) Ag and deposited by vacuum evaporation on high-purity (99.9995%) Al. Determination of the diffusion coefficient was made by determining the Cu content of layers of constant thickness that were removed by chemical dissolution using mineral acids. The thickness of the layer removed was calculated using the surface and the density of the metal, as well as its weight loss. Table 2 summarizes selected dissolution procedures which were developed for a variety of matrices. Mateo and Sabat6 (B26) get the prize for developing the least expensive microwave digestion system! They simply modified slightly a domestic microwave oven by coating its interior with a lubricant spray of Teflon particles and drilling a 1.5-cm hole in its side to let a Teflon tubing out. This tubing connects the center of a Pyrex container in which are placed open Teflon bottles of sample/acid mixture to a water pump whosevacuum is sufficient to continuously evacuate the acid fumes generated. Since the dead space in the container is minimized using fine sand, it need not be vacuum-tight to facilitate the evacuation of acid fumes. As many as 25 samples can be processed simultaneously as long as the Teflon coating and a microwave Analytical Chemishy, Vol. 66, No. 12, June 15, 1994
405R
m
a
9 m
" I
a
5 a
2 a
0 " s ep,
2
E
B
0
2cr:
VI
m
5cr:
B
'cl
B
VI
TI
e
-' e,
> m'
m
d
466R
r;i
AnalytlcalChemistry, Vol. 66,No. 12, June 15, 1994
a
0 E
8 .r=
B
CI
za
6
.-x
c)
ea
J= M
.E
II
zm
d
a
-0
9
d
5
3
e E
6 U
E'
N
AnalyticaIChemistry, Vol. 66, No. 12, June 15, 1994
467R
sink (Le., a beaker of water) is maintained inside the oven. Totland et al. performed an extensive comparison of LiB02 fusion, HF/HC104 open acid digestion, and closed-vessel microwave digestion for the analysis of geological materials by ICP-AES (and ICP-MS) (B28). The results demonstrated that no single dissolution technique was effective for the determination of all elements. For instance a fusion was best for the determination of Cr, Hf, and Zr as it was most effective in dissolving refractory minerals, but it also led to the loss of volatile elements. Whereas acid digestion with or without microwave assistance was better for all trace elements (volatile or not), the microwave approach providing lower blanks and requiring less time than the open digestion. Borsier (B29) described a fully automated sample dissolution unit for the multielemental analysis of geological materials by ICP-AES. The only manual operations involve grinding samples to 90% of Sod2- recovered with 430 pL of 2 M NH40H; on-line standard additions; agreement with certified value
C1-48
C1-49
0 Ratio of peak height over the steady-state signal observed from direct aspiration. b Ratio of the detection limit by direct aspiration over that by preconcentration or FIA. CRatio of peak area over the FIA peak observed without preconcentration. 1,5-Bis(di-2pyridy1)methylene thiocarbonohydrazide.
An unusual approach to HPLC/MIP-AES was described by Mason et al. (CI-36), who used a moving band interface, typically designed for HPLC/MS, to sequentially dry and then vaporize the sample prior to entry into a He Beenakkertype MIP. Absolute detection limits of 20-100 ng of C1 were obtained for 9-chlorofluorene, p-chlorobiphenyl, 4-chlorobenzophenone, and a,a’-dichloro-o-xylene. However, the response decreased as the boiling point of the chlorinecontaining analyte decreased, suggesting that part of the analyte might have been prematurely lost before entering the vaporization region of the interface. On- Line Preconcentration and Flow Injection Analysis (FIA). Some of the separation/preconcentration techniques described in section B could not be directly coupled to ICPAES because the optimum elution flow rate was higher than the optimum uptake rate of the nebulizer. Israel et al. (CI-37)used a T-junction between the column and the peristaltic pump controlling solution delivery to the nebulizer, so that both could operate at their respective optimum flow rate. Excess effluent was collected in a trap and, after mixing, could later be aspirated by the nebulizer. This setup allowed the determination of the transient elution peak as well as the steady-state signal for the excess effluent. It was successfully applied to the determination of trace elements in ultrapure and analytical-reagent grade KNO3 and KOAc. This interface could certainly be adapted to other separation techniques that operate at flow rates incompatible with a nebulizer. Tyson (CI-38)reviewed the basic concepts and theory of flow injection techniques. Although atomic spectrometry techniques are not necessarily ideal detectors for FI, their combination nonetheless offers many advantages. Table 4 provides some examples of such FI systems. Some on-line preconcentration systems where no FI (i.e., injection of a plug of sample into a flow of carrier) is carried out are also included. Kasthurikrishnan and Koropchak (CI-47)used FI Donnan dialysis prior to ICP-AES determination. They carried out preconcentrationof tracemetal cations on the basis that cations from a relatively large sample volume will migrate through a permeable membrane into a small volume of concentrated electrolyte solution if a high ionic strength gradient exists
across the membrane. The great simplicity of this method makes it very attractive. However, its application to real samples, including those containing large concentrations of alkalis and alkaline earths has yet to be investigated. On the other hand, an efficient computer-controlled FIA system was developed by dos Reis et al. (CI-50)for on-line dilution or performing standard additions. The latter is carried out by first loading a standard into a coil and then delivering and mixing it with theseparately injected sample through a merging zone manifold. The size of the standard and sample loops as well as their carrier flow rates (which are independently controlled) determined the number of standard additions possible. As many as 11 standard additions could be performed within 5 min. The method was successfully applied to the analysis of plant digests SRMs. Gluodenis and Tyson (CI-51)developed an on-line microwave digestion system. The sample was injected in the form of a slurry and pumped, along with concentrated HNO3, into a glass reactor within a microwave oven. The flow was then stopped to allow total digestion of the sample with as small a residual dissolved carbon content as possible. Results for the determination of Fe in cocoa powder by ICP-AES (with internal standardization using Sc) agreed with those obtained by batch microwave digestion, hot plate, or dry ashing. Solid Sample Introduction. Solid samples may be directly analyzed by a number of methods. Their major advantage is reduced contamination, since nodissolution and little sample preparation is generally required. However, techniques of improving sample analysis by reducing matrix effects that can be important must bedevised. These may includechemical modification, standard calibration with matrix-matched solutions or solids, and internal standardization. Selected examples of successful applications are summarized in Table 5 . Ablation Techniques. Moenke-Blankenburg (C2-9)wrote an extensive review of laser ablation (LA) techniques for ICPAES and ICP-MS. It discusses laser conditions according to the type of laser, as well as transport processes, including the design of the ablation cell and the transfer line to the ICP. Calibration methods are also covered with tables demonstrating numerous applications. Analytlcal Chemistry, Vol. 66,No. 12, June 15, 1994
477R
Spark ablation (SA) was used with ICP-AES detection by Coedo et al. (C2-IO) to analyze ferrochromium alloys, which were diluted by remelting with iron in a 12:28 FeCr/Fe ratio. Low voltage was initially used for best plasma stability (400 V, 600 Hz), but reproducibility was found to relate to sample hardness, which in turn was related to carbon content. Better results were achieved with 600 V at 600 Hz, with reproducibilityindependent ofcarboncontent. Acarbondioxidestream was used to keep the sample cool, reducing the effect of sample heating with sparking and hence improving sampling reproducibility. Good accuracy was found in the analysis of standard reference materials, but polishing was required after a small number of sparks. Uebbing et al. (C2-11 ) coupled LA to a MIP. The ablation cell was placed at 1 mm from the microwave cavity to allow ablation of the sample directly into the microwave cavity for maximum analyte line intensity. A low pressure of Ar buffer gas was used to entrain ablated material into the MIP where analyte emission was monitored by a gateable photodiodearray spectrometer. They indeed demonstrated that, under these conditions, LA-MIP-OES required time-resolved detection since strong perturbations of the plasma by a laserproduced vapor plume only occurred within the first 50 ps of a laser shot. Furthermore, the low gas temperature of the MIP (
