Atomic Spectrometry Update- Advances in Atomic Absorption and Fluorescence Spectrometry and Related Techniques I Atomic I Spectrometry Update i STEVE J. HILL* Department of Environmental Sciences University of Plymouth Plymouth Devon UK PL4 8AA JOHN B . DAWSON Department of Instrumentation and Analytical Science UMIST P.O. Box 88 Manchester UK M60 1 QD W. JOHN PRICE Ellenmoor East Budleigh Budleigh Salterton Devon UK EX9 7DQ IAN L. SHUTTLER Bodenseewerk Perkin-Elmer GmbH Postfach 101 761 0-88647 Uberlingen Germany JULIAN F . TYSON Department of Chemistry University of Massachusetts Box 34510 Amherst M A 01003-451 0 USA SUMMARY OF CONTENTS 1. 1.1. 1.1.1. 1.1.2. 1.1.3. 1.1.3.1. 1.1.3.2. 1.1.3.3. 1.1.4. 1.1.5. 1.1.6. 1.2. 1.2.1. 1.2.1.1. 1.2.1.2. 1.2.2. 1.2.2.1. 1.2.2.2. 1.2.2.3.1.2.2.4. 1.2.3. 1.2.4. 1.2.4.1. 1.2.4.2. 1.2.4.3. 1.2.4.4. 1.2.5. 1.3. 1.3.1. 1.3.1.1. 1.3.1.2. 1.3.2. 1.3.3. 1.4. 1.4.1. 1.4.2. 1.4.3. 1.4.4. Atomic Absorption Spectrometry Flame Atomization Fundamental studies of flames Flames for AAS Sample introduction Atom trapping techniques Nebulization/vaporization Flow injection Interference studies Sample pre- treatmen t Chromatographic detection Electrothermal Atomization Atomizer design and surface modification Graphite atomizers Metal atomizers and metallic coatings Sample introduction Slurry sampling Gas sampling Coupled techniques and preconcentration Electrodeposition Fundamental processes Interferences Spectral interferences Chemical modifiers-general Chemical modifiers-palladium Other chemical modifiers Developments in technique Chemical Vapour Generation Hydride generation General studies of fundamentals techniques and instrumentation Determination of individual elements Mercury by cold vapour generation Volatile organometallic compound generaion and metal vapour separation Spectrometers Light sources Continuum source and simultaneous multi-element AAS Background correction Detectors 1.5.1.5.1. 1.5.2. 1.5.3. 2. 2.1. 2.2. 3. 3.1. 3.2. Instrument Control and Data Processing Instrument control Data processing Chemometrics Atomic Fluorescence Spectrometry Discharge-excited Atomic Fluorescence Laser-excited Atomic Fluorescence Laser-based Spectroscopy Laser Ablation and Excitation Laser Enhanced Ionization This review follows on from last year’s (J. Anal. At. Spectrom.1995 10 199R) and describes the developments in atomic absorption and atomic fluorescence spectrometry since that time. Included in this review are fundamental processes and instrumentation in the areas of atomic absorption and atomic fluorescence spectrometry together with advances in related techniques such as atomic magneto-optical rotation spectrometry and laser-enhanced ionization. The review of ‘Atomic Emission Spectrometry’ may be found in J. Anal. At. Spectrom. 1996 11 (4). Once again this year a number of changes have been implemented to aid presentation. The most important of these is the inclusion of Tables to present information in Section 1.2 Electrothermal Atomization. It is hoped that this will facilitate easier and quicker access to material covering a specific area.The sub-headings in Section 1 have also been modified to reflect the material presented in the review although in general the format is the same as in previous years. The full references names and addresses of authors can be readily found from the Atomic Spectrometry Updates References in the relevant issue of J. Anal. At. Spectrum. However as an additional service to readers an abbreviated form of each reference quoted (except for those of Conference Abstracts) is given at the end of the review. Comments as to these changes and possible improvements for future reviews are welcomed by the review coordinator. 1. ATOMIC ABSORPTION SPECTROMETRY 1.1. Flame Atomization * Review Coordinator to whom correspondence should be AS a mature analytical technique developments in FAAS are principally directed at solving specific analytical problems and 281 R addressed.Journal of Analytical Atomic Spectrometry August 1996 Vol. 11 (281 R-325R)therefore are presented in the ASU Review appropriate to the sample matrix. There has not been a readily identifiable development of a new generally applicable technique. Studies of the fundamental parameters of atoms and flames continue to be made. Examples of specific analyses will be presented in this review when they are either representative of a general class or contain a novel feature of more general applicability. This introductory section will contain small items of interest in FAAS which otherwise may be overlooked by the reader if recorded elsewhere in the text.Several books and reviews on analytical atomic spectrometry have appeared during the past year (95/3467 95/3512 95/3513 95/3967 95/4526). The most comprehensive review was that appearing in Volume 10 of this journal (95/3732). Reviews on the determination of individual elements by AAS include A1 (95/3676) and T1 (95/2868). The sensitivity of a determination by AAS is affected by the oscillator strength of the atomic transition utilized for the analysis. Hannaford (95/2214) reported theoretical and exper- imental studies of the role of oscillator strength in AAS. The results were used to revise estimates of the theoretical character- istic masses of Ag Al Au Cay Mo Na Ti and V in ETAAS. Data on the determination of 25 elements by FAAS has been gathered together by Zehr and Van Kuren (95/1032) and presented as a single graph based on LODs and RSDs to give analysts a rapid guide to the potential analytical performance of a specified element.In the early days of AAS spectrophoto- meters used for the measurement of molecular spectra were adapted for AA measurement. In contrast the present-day widespread availability of AAS spectrometers has encouraged the use of AA instruments for the study of molecular spectra (95/3668)! Indirect methods based on FAAS have been developed either to improve sensitivity or to determine analytes that otherwise cannot be measured by FAAS. Recent examples of such methods include B determined via the Cu content of the complex Cu( hen),^' -B-phenylhydroxyacetic acid after extrac- tion of the complex with isobutyl methyl ketone (MIBK) (95/3785). The method was applied to the determination of B in cast iron.Copper was also the determined element in the indirect determination of iodide (95/2861). Iodide was reacted with the NCP-Cur‘ complex in the presence of thiourea. The complex NCP-Cur-I- was formed and extracted with chlorob- enzene for determination of its Cu content by FAAS. Nitrophenols were determined indirectly at the 100 ppb level by extraction of ionic associates of bipyridylcopper or phen- anthrolinylcopper into nitrobenzene or MIBK followed by direct nebulization and determination of Cu by FAAS (95/3842). Aluminium in HNO solution has been determined by its displacement of Pb from a solution in benzene of N-nitrophenylhydroxylamine lead salt (96/358).Lead was determined in the aqueous phase. The LOD for A1 was -0.005 pg ml-’ with recoveries of 94-100%. Indirect methods are more frequently applied to the determination of non- atomic analytes. Sulfate was precipitated with barium chromate to liberate an equivalent amount of chromic acid. Excess barium chromate was removed with lime prior to FAAS determination of Cr (95/1213). The total acid in soda water has been determined by adding a suspension of CaCO to the soda water centrifugation of the mixture and measurement of Ca in the supernatant solution by FAAS (95/3436). The acids studied were acetic citric fumaric lactic malic succinic and tartaric. Copper was again the determinand in an indirect method for the measurement of tannins in tea (95/3627).The tannins were first extracted with MIBK-butanol; after separa- tion a C U ~ ( P O ~ ) ~ reagent was added to the organic phase and following centrifugation unconsumed Cu was determined in the supernatant. Recoveries were 94-104% with an RSD of 3.1%. The results agreed with those of the official method for tannin determination. 1.1.1. Fundamental studies ofjames The use of flames in analytical atomic spectrometry has necessitated the study of the properties of flames themselves in order to maximize their analytical usefulness. At the present time there is widespread interest in flames but most studies relate to non-analytical uses. Some indication of the range of those studies will be given in this section for more compre- hensive information the literature on combustion should be consulted.Temperature is often the most important characteristic of a flame when considering its suitability for FAAS. An approxi- mate estimate of the temperature of a flame can be made by measurement of the intensity of its IR emission. However as several components including C02 contribute to the total radiation calculation of the precise temperature is not possible (95/1171). The temperature distribution within a flame has a significant bearing on the zone of the flame most suitable for analytical use. Two-line thermometry employing CO molecular absorption of radiation from a tunable dye laser was used in a study of a low pressure stoichiometric pre-mixed CH,-O2 flame (95/3889). Tomographic analysis was used to correct the underestimation of the observed temperature.The correction was a non-linear function of the height above the burner surface. Measurements of the temperature of an axisymmetric flame using a Talbot interferometer with circular gratings agreed with those by thermocouple to within +0.2% (95/1238). In addition to temperature the composition of the flame gases plays a significant part in the generation of free analyte atoms. The formation of soot in a pre-mixed CH,-O2 laminar flame at atmospheric pressure was studied by D’Anna D’Alessio and Minutolo (95/3878). Particles ( M 3-4 nm) form- ing early in the flame did not absorb or fluoresce in the middle waveband but were the precursors of soot particles. It was suggested that the ‘building bricks’ of these particles were aromatic compounds with few condensed rings (< 3) connected by aliphatic and oxygen bonds. Particle inception was con- trolled by the internal arrangement of these polymers leading to structures with more condensed aromatic rings and more compact three-dimensional shape thus forming the first soot nuclei.The temporal and spatial distribution of soot in heptane and decane fuelled flames was determined by laser-induced incandescence (95/2259) using a 10 Hz Nd YAG laser at 1064nm. The laser intensity in the flame was > 1 x lo7 W cmV2 with pulse lengths of 8-10 ns. The distribution of water vapour in flames has been mapped by means of 2-photon excitation of the water molecule using a tunable excimer laser at 248 nm (95/1237) with detection at 400-500nm. The detection limit for H20 with a single laser shot was estimated to be 0.2% H,O.The detection of free radical emission e.g. OH in the spectral range 200-400 nm has been improved by incorporation of stray light rejection measures into the experimental procedure (95/3066). These measures include filters to exclude from the spectrometer radiation outside the required spectral region elimination of the zero-order spectrum in the spectrometer and optical baffles to ensure that the spectrometer mirrors and gratings are not overfilled. Laser-induced fluorescence is frequently used in studies of the behaviour of Nu atoms and OH radicals in flames. A high repetition rate (82 MHz) picosecond laser was used to measure the quenching rate coefficients for Na and OH in an air-CH flame (95/2257 95/2258).Calibration based on the AAS of Na in the flame predicted a detection limit of 2 x cm- for OH radicals. Hydroxyl radicals in an atmos- pheric pressure air-C2H flame were detected via laser enhanced ionization of In (95/2213,95/3882). This information was used to determine OH rotational temperature with high resolution. Anomalously high temperatures were detected at the edge of the flame. The decay ofjuorescence of Nu seeded 282R Journal of Analytical Atomic Spectrometry August 1996 Vol. 11into a laminar H,-02-Ar diffusion flame has been measured by means of picosecond time-resolved laser induced fluores- cence (95/2261). This was achieved by sampling the fluores- cence signal at the rate of lo9 samples s-l for ~ 4 0 0 p s . The absolute concentration of atomic oxygen in a H,-air flame has been measured using intracavity laser spectroscopy based on a dye laser pumped by an argon-ion laser (95/3766) 95/4432).Observations were made at 630.03 and 336.38 nm with an equivalent optical path of 10 km. The LOD for 0 was 6 x lo1 atoms ~ m - ~ . The distribution of CN in a turbulent air-CH flame was mapped by planar single shot laser induced fluorescence (95/2232). The CN zone appeared to be located outside the inner cone and to be z 1 mm thick. 1.1.2. Flames for AAS Optimization of flame conditions is an essential step in achiev- ing sensitive accurate analysis by FAAS. The specijication of a nebulizer-burner system designed to be a ‘universal flame atomizer’ has been given as follows firstly to be capable of analysing a wide range of samples; secondly to atomize all elements determinable by AAS; thirdly to be chemically inert to solvents used in AAS; fourthly to aspirate solutions with high dissolved solids content without blockage; fifthly to exhibit good sensitivity; and sixthly to be cost effective.The development of an atomizer designed to meet these criteria has been reported (95/2534). Selection of appropriate oxidant andfuel gases their flow rates and analytical zone within the flame are options open to most users of FAAS instruments. Over the past 40 years much has been written on this subject and even today the subtleties of these adjustments provide opportunities fo further research publications. Zheng and Meng reviewed the problems of determining Cr by FAAS.As a means of optimizing flame temperature a system was devel- oped whereby 0 could be added to the air flow of an air-C,H flame (95/3639). Nitrous oxide when used as the oxidant in a C,H,-based flame facilitates the atomization of refractory elements. This flame has been studied in depth for the determi- nation of Nb (95/1861). Lateral and vertical profiles of Nb-free atoms were measured. When the flame was shielded with Ar or N2 flatter lateral profiles were observed. A possible atomiz- ation mechanism based on vaporization from condensed phase Nb oxide species inhibited by non-volatile carbide formation in reducing flames was suggested. The solvent and other reagents associated with sample prep- aration can have a significant bearing on the atom population in the flame.Samples are most commonly aspirated into the atomizing flame in aqueous media frequently containing HCl. Kantor and Ernyei (95/2225 95/2226) have studied the effect of adding C1 to the jame gases uia the introduction of CCI into the flame. It was inferred that only 19% of the total C1 introduced as CC1 was converted into HC1 and C1 species. From the signal depression caused by the introduction of halocarbon into air-C,H flame it was possible to elucidate the formation of the monochlorides of Sn and Mo under strongly reducing flame conditions as well as the formation of the dichlorides of Ba Ca Cr Fe Mg Mn and Sr under slightly and strongly reducing conditions. It was suggested that a strongly acidic solution of 10% (m/v) HC1 would be equivalent to the introduction of 25 ml min-’ of CCI4 vapour into the flame and could cause significant depression of an AA signal.Organic solvents are used to increase sensitivity in FAAS either by improving nebulizer-atomizer efficiency or as a consequence of a solvent extraction preconcentration process. They also modify the atmosphere within the flame. Compared with aqueous solutions pure MIBK water-ethanol-acetone mixture ,and water-ethanol-MIBK were found to produce significant signal enhancement (95/38 16). Ethanol was pre- ferred to MIBK as the solvent for the determination of Ru in organic solutions as it gave greater sensitivity lower detection limits a more pleasant smell less contamination of the spray chamber and burner and the burner position was less critical (95/1839).lsoamyl acetate (mainly 3-methylbutyl acetate) was used to extract chlorosilanes from anhydrous magnesium chloride for direct determination of the chlorosilanes by FAAS (95/4406). The RSD for replicate analyses was ~ 3 % . The determination of metals in organic solvents is complicated by the difficulty of preparing appropriate standard solutions. A method has been proposed for the coulometric preparation of solutions of Co Cu Ni V and Zn in an acetone-ethanol- o-xylene mixture (95/2799). The metal was the working elec- trode with an auxiliary Pt electrode. The cell was filled with an acetone-ethanol mixture to which the supporting electrolyte (0.1 mol 1-1 NaClO or 0.1 mol 1-l LiCl) was added. After electrolysis the solution was diluted with o-xylene.These reference solutions were suitable for the FAAS analyses of bitumen and petroleum products dissolved in the solvent mixture. 1.1.3. Sample introduction No novel means of sample introduction have caught the eye of this reviewer during the past year. Published information generally describes developments or modifications of existing practice. Much of such work relates to the needs of specific analyses and thus will be presented in ASU reviews concerned with application of FAAS. Selected techniques that may have a wider application than the one for which they were specifi- cally designed will be presented here. 1.1.3.1. Atom trapping techniques. The most common form of this technique employs a tube with an aperture often a slot to admit the vaporized sample mounted above the burner of an FAAS instrument and on its optical axis.The sample is nebulized into the flame. There are two modes ofoperation of atom trapping systems. In the most commonly used mode the vaporized sample flows continuously through the atom trapping tube. The gain in sensitivity over a conventional long path burner atomizer is of the order of 3- to 5-fold for easily atomized elements. This improvement is attributable to the increased residence time of the atom in the optical path. The alternative approach depends upon first collecting the analyte atoms on the interior wall of the tube by condensation of the atom vapour over a period of time; this is then followed by rapid release of the atoms when the tube temperature is raised. In the latter mode the gain in sensitivity can be of the order of 50-fold.The slotted tube atomizer not only improves sensitivity but it has also been reported to improve precision by factors of from 2- to &fold (95/3384). The tube is usually made of silica although studies have shown that graphite can also be used successfully for that purpose (95/2388). Atom trapping systems have been combined with chromato- graphic separation of the analyte. Usually the atomizer is a quartz T-tube heated by a flame or electrical furnace into which the sample is introduced via the lower branch of the T. The performance of a coupled gas chromatograph-quartz tube atomizer AA system was studied by Johansson Baxter and Frech (95/4685) with a view to optimizing the S/N ratio. They found the system to be complex particularly the interaction between the demands of the GC separation and atomization in the quartz tube.Changes in the radius of the quartz tube were found to be more important than changes in its length. A GC-quartz T-tube atomizer system based on common heater modules was evaluated for the determination of organolead and organotin compounds (95/3847). The detection limits achieved were z 1.5 pg for Et,Pb (as Pb) and x 3.5 pg for Bu4Sn (as Sn). When the separation of compounds is by HPLC the eluate requires nebulization prior to transfer to the atomization tube. An interface has been developed which Journal of Analytical Atomic Spectrometry August 1996 Vol. 11 283Rincorporates into the lower branch of the T a pyrolysis chamber for the combustion of the aerosol prior to entering the heated silica optical tube for AAS measurement (95/2267). The LODs for some As and Se ionic species were in the range 0.6-2 ng. Continuous flow slotted tube atom traps (STAT) have been successfully applied to a number of analyses including Pb in water where an LOD of 1.9 ng ml-I was achieved (95/3384) Se in rice garlic bulb and tea (95/3499) and Sb in copper- based alloys where interferences were minimized by coating the internal surface of the quartz tube with copper (95/2501).The sensitivity of the determination of Sn in copper based alloys by STAT was increased 100-fold by generating the hydride of Sn with NaBH and feeding this gas into the furnace in place of an Sn-containing aerosol (95/2502). As noted earlier the interruptedjlow STAT is more sensitive by virtue of its concentration step which is achieved by initially reducing the tube temperature to collect the analyte.The simplest approach is to modify the composition of the gas mixture of the flame heating the sample aerosol and furnace tube. When determining Ag Huang Qian and Liu (95/2862) used a C2H flow of 0.2 1 min-l during the 1 min trapping phase and 0.41 min-’ for the release of Ag. Compared with conventional AAS sensitivity was improved 62-fold. An alternative approach uses water as cooling to reduce the tube temperature during the trapping phase. Such a water-cooled atom trap has been constructed and evaluated (95/C4191 95/4686). Sensitivites for Cd and Pb were 0.36 and 3.0 ng ml-’ respectively. 1.1.3.2. Nebulization-vaporization.Slurry nebulization is now an established though not widely used analytical pro- cedure. Mauri et al. (95/3462) applied the technique to the speciation of calcium compounds by FAAS. The speciation of Ca’ calcium hydroxide and calcium phosphate was achieved via the determination of total Ca and absorbance values at burner heights of 5 and 7mm with an air-C2H2 flame and heights of 10 and 13 mm with a N,0-C2H2 flame. The different behaviour of Ca in slurries and in aqueous solutions was discussed. When the sample is a liquid that is immiscible with water emulsijication may be used to produce a sample suitable for nebulizing into a flame. This approach was used for the determination of cobalt napthenate in gasolene (95/943) and ferrocene in gasolene (95/2837).Desolvation has from time to time been found to be advantageous in FAAS; in the case of MIP atomization AAS it is almost essential. Details of a desolvation system incorporating nebulization heated tubing a water-cooled condenser and a concentrated sulfuric acid desiccator have been published (95/1028). The effects of operating parameters were investigated and the desolvation efficiencies of the different stages assessed. An over-all efficiency of 99% was achieved. Nebulization under normal operating conditions is a very inefficient process with well over 90% of the sample passing to waste. Microsample injection (pulse or discrete nebulization) leads to more efficient sample utilization; further by incorpor- ating first-order derivative signal processing discrimination against background noise can be improved.When this tech- nique was applied to the determination of Cu in whole blood a 2- to 4-fold improvement in signal over normal FAAS was obtained (95/3861). The RSD of measurements was 4.5%. Simple pulse sampling has been used for the determination of Z n in serum (95/3872) and Pb in botanical samples after pretreatment and solvent extraction with APDC-CHC1 (95/3985). The sensitivity of another old-established but little used transient response technique vaporization from an electrical/flame heated platinum loop has been increased by adsorbing metal-APDC complexes onto activated carbon (95/C4215). The method was applied to the determination of Cd Pb and Zn in dried apricots. Recoveries were up to 90% with 100mg of activated carbon in 20ml of APDC solution at pH 6.Lead was preconcentrated with APDC-polystyrene on a modified platinum wire matrix for direct FAAS determi- nation (96/156). The detection limit was 0.65 ng ml-’. A thermospray interface combines the generation of a fine aerosol with its practically instantaneous desolvation. Such an interface between an HPLC column and an FAAS instrument has been used for studies of Cd speciation in human body fluids urine and perspiration (95/1041). Compared with a conventional nebulizer interface the concentration sensitivity was enhanced 25-fold. High pressure nebulization produces finer aerosols and hence more efficient sample utilization. An HPLC pump was used to feed the sample solution to a special fine nebulization nozzle (95/4392).By inserting an HPLC column between the pump and the nozzle on-line separation and preconcentration were possible. Supercritical Jluids have been used to dissolve transport and nebulize lipophilic analytes in FAAS determination. In one such application the analyte metal in aqueous medium was complexed in situ with tetrabutylammonium dibutyldithiocar- bamate and the derivate mobilized into supercritical CO (95/3413). The superheated extractor eluate was nebulized into the upper region of an 02-H2 flame prior to transport to the absorption tube. Detection limits in the pg range were obtained for As Cd Cu Mn Pb Se and Zn. Similar detection limits were obtained when a 5 yl aqueous sample was injected uia a heated T-junction into a flowing stream of supercritical C02 at 1500-2000 psi (95/C3024).An emulsion was formed which when rapidly expanded through a fused silica capillary or stainless steel restrictor produced a very fine aerosol. Droplet sizes were dependent on the concentration of analyte dissolved in the aqueous solution and typically half the mass was contained in aqueous droplets in the 0.7-1.0ym range. The washout time for the system was <5 s. 1.1.3.3. Flow injection. There are signs in the material reviewed for this year’s update that FI is being used routinely as a sample introduction method. Several papers describing work in which FI has been used are clearly no more than applications papers i.e. they report on a new or improved method for the determination of certain analyte elements in particular matrices.Among such papers are reports of the determination of Fe in caustic soda (95/4421) of Cu and Fe in gelatin (95/4246) of K in sodium antimonate (95/1957) of Co Cu Fe and Ni in water (95/2350) and of Ca Fe K Mg and Na in fluorescent screen glass (95/2514). There has also been the usual crop of overview presentations at conferences (95/C803 95/C800 95/C764 95/C4200). Clearly there is still interest in and considerable research activity related to develop- ments in both the techniques of FI and applications. Several overview articles have appeared concerning (a) the performance of the FI-FAAS combination (Ma et al. Mikrochim. Acta 1995 119 95) (b) the use of time-based or volume-based sample introduction devices (95/3515) and (c) the history of and general trends in FI-AAS (Fang et al.J. Anal. At. Spectrom. 1996 11 1). This latter article contains the bibliography (now numbered) of Fang’s book (95/4469) from which it can be readily seen that the book covers the contributions from some 645 original papers. The book is devoted to FI-AAS does not cover any other atomic spec- trometry and was written according to the author’s preface ‘mainly with the practising AA spectroscopist in mind with the aim of providing the reader with both a general reference source book and a laboratory manual for up-grading AA spectrometric methods in various fields of application’. The coverage of the literature stops in late 1994. The book contains substantial sections devoted to electrothermal atomization and to chemical vapour generation (these two topics account for 284 R Journal of Analytical Atomic Spectrometry August 1996 Vol.11about 20% of the text). The book contains much valuable commentary from the author and is highly recommended. A number of applications have specifically taken advantage of the greater tolerance to high dissolved solids imparted by the use of FI. For example (95/C3747) Cd Cu and Pb have been determined in foodstuffs with high sodium contents such as fish and instant soup by the simple expedient of injecting a sample volume of 100 pl. A similar approach has been used for the determination of Pb in paint (95/C763) Pb in copper (95/1941) Ag and Cd in rocks (95/1941) and for the solution of a potentially sticky problem the determination of Ca K Mg Mn Na and Zn in honey (96/145).The injection of slurry samples was used in the determination of several elements in cements of varying silica content (95/2849). For the determination of minor components slurries (0.05-0.25% m/v) were prepared in 30% glycerol-2% nitric acid and 2% hydrofluoric acid solution of which 50-2OOpl were injected. For the determination of major components the slurry concentration was 0.1 9'0 m/v in a suspension medium also containing 0.25% solid lanthanum nitrate. The carrier composition was changed from water to a 5% glycerol and 1 YO lanthanum nitrate solution. A variable volume single well stirred tank functioned as an on-line dilution device and in both cases aqueous standards were used. Several reports of novel flow injection hardware have appeared.Details of the air transport system with hydro- dynamic injection mentioned in a conference presentation (see J. Anal. At. Spectrom. 1995 10 199R) have been published (95/4564). The apparatus was used in a method for the determination of Cu in serum in which 80 pl sample volumes were injected (in triplicate) into an aqueous carrier flowing at 5.2ml min-'. To conserve sample transport from the auto- sampler to the injector was by an air carrier. A total of 800 p1 was needed for the three injections. A bovine serum RM was accurately analysed. A microsampling device based on the 'controlled dispersion analysis' concept of Rocks Sherwood and Riley (Analyst 1985 110 493) has been constructed and used for the determination of Cu Fe Ni and Sr in fluorite.Again sample consumption was reduced (to about a third that of a conventional FI system) as there was no injection device as such. A moveable probe was transferred between sample and carrier reservoirs. A sample volume of between 50 and 600 pl was aspirated into the manifold by control of the pump. There have been no reports apparently of the use of on-line microwave assisted digestions with FAAS though the use of this method of heating in chemical vapour generation pro- cedures (see section 1.3.) has featured in a number of publications. Further developments in systems involving on-line dialysis have been described by van Staden (95/4687). The system was applied to the determination of Cu over a wide range of concentrations in industrial effluent. The dialysers also pro- vided on-line filtration removing particles which otherwise would have blocked the nebulizer.For a 75 pl injection volume a throughput of 120 h-' was obtained at a precision of 0.3% RSD. Dialysis was also used (95/4014) to dilute soil extracts for the determination of K and Na by FAES and of Ca and Mg by FAAS. Analytes were speciated as either 'soluble' (extracted by shaking the sample with water) or 'extractable' (shaking with a pH 7 ammonium acetate solution). The working ranges were between approximately 1.0 and 300 ppm for Ca K and Na and between 0.2 and 20 ppm for Mg. The possible use of tubular Donnan dialysis described some time ago (Koropchak and Allen Anal. Chem. 1989 61 1410) has been discussed (95/C423 1 ). In addition to manifolds containing dialysers a number of dilution manifolds have been described.A procedure based on the slow introduction of sample solution into the injection valve by a peristaltic pump followed by delivery to the spectrometer at a rate compatible with the requirement of the nebulizer achieved dilutions of up to 2000-fold (95/305). The sampling frequency was between 40 and 120 h-' and it was shown that a linear response for Cu over three orders of magnitude (10-10000 ppm) could be obtained. The perform- ance was extended (96/301) to dilution factors of 20000 with precisions of between 1 and 3% RSD by using an optical device to monitor the position of the pump rollers at the beginning of the metering stage. Volumes as low as 20 nl were introduced at pump head rotation speeds of 0.56 rpm with pump tubing of 0.38 mm internal diameter.Even larger dilution factors were obtained by Xu and Fang (95/3809) for a manifold in which two sample zones injected sequentially overlapped to give a measurable minimum between the two peaks. For volumes of 1.4 pl the dispersion coefficient at the valley between the peaks was 27000 when the injections were 18 s apart. The procedure was applied to the determination of Mg at concentrations of several g 1-' in brines used for the coagulation of colloids in bean curd production. As has been the case for several years now a major theme in the published work in FI-FAAS concerns the use of solid- phase extraction (SPE ) procedures for preconcentration and matrix removal. The boundaries between SPE and on-line precipitation are somewhat blurred in the case of the formation of insoluble analyte derivatives.In some instances these are retained on packed-bed reactors (as in the so-called sorbent extraction procedure) and in others on the interior wall of the FI tubing configured as a knotted tubular reactor. Although chelating functionalities offer the best possibilities in terms of selectivity there is little interest in the use of Chelex-100 because of the dimensional instabilities associated with the pH changes which occur when loading and eluting the analyte. A comparison of the relative performance of quinolin-8-ol (Q80L) immobilized on controlled pore glass (CPG) and Muromac-A-1 (an iminodiacetate functionality immobilized on a highly cross-linked styrene divinylbenzene co-polymer) for the possible speciation of Cr described in a conference presen- tation last year (see J.Anal. At. Spectrom. 1995 10 199R) has now been published (95/1499). Neither of these materials is apparently suitable for the direct preconcentration of Crv' but both will retain Cr"'. The optimum pH values for retention being 9 and 4 for the Q80L and Muromac materials respect- ively. LODs of a few ppb for a sample volume of 18 ml were obtained. The retained ions were eluted with a mixture of hydrochloric acid (2 mol 1-') and nitric acid (2 mol 1-l). As Muromac had better tolerances towards the interferences from calcium and iron it was used for the determination of Cr"' in coastal and estuarine waters at concentrations of 100 ppb. By reducing CrV' with hydroxyammonium chloride sequential speciation was achieved.The same determination has been performed by a procedure in which derivatized cellulose mate- rials (95/889) were used. A manifold with two columns in series was used. In the first column Cellex T (cellulose derivat- ized with quaternary amine groups) retained CrV1 while the second column containing Cellex P (cellulose with phosphonic acid groups) retained Cr"'. Loading was at pH 2.5 and Cr"' was eluted with 1 moll-' hydrochloric acid and the CrV' with 1 mol 1-' sodium hydroxide solution. The eluted CrV' by-passed the Cellex P column. Detection limits of around 1 ppb were obtained. Further work from the University of Strathclyde featured immobilized Q80L on CPG in the determination of T1 in saline samples (95/2331).The method was optimized in terms of the nature of the buffer solution to be used and a pH 10 solution of ammonium acetate and ammonium chloride was selected. Better results were obtained for T1"' than for TI' and so an off-line oxidation procedure with bromine was devised. The excess bromine was removed with phenol. For a sample Journal of Analytical Atomic Spectrometry August 1996 Vol. 1 I 285Rvolume of 18 ml the LOD was 3 ppb. The method was applied to the determination of T1 in river and sea-waters and in potassium-enriched table salt. Chinese workers have been active in devising procedures for the determination of Au. A chelating rhodanine jibre was used (95/2368) in a procedure which involved passing the sample solution in 20% aqua regia through the column at 8 ml min-.' for 40 s with elution with thiourea solution at 2.6 ml min-l.The enrichment factor of 25 allowed the determination of ALL at concentrations down to 0.02 g t-'. A device described as the YIL-110 (95/4378) incorporated a special3bre microcolumfil for the detection of Au down to 70 ppt in solution. Yet a third. group of workers reported (95/4428) that after preconcen- tration on polyurethanefoam at a flow rate of 16 ml min-' an enrichment factor of 50 was obtained. Some preliminary results for the performance of poly-~- cysteine as an extractant have been presented (95/C3019) and information concerning the binding characteristics for Cd Cu and Zn have been published (Autry and Holcombe Analyst 1995 120 2643). The logarithms of the binding constants for Cd and Zn were measured by a spectrophotometric titration procedure to be 8.0 & 0.5 and 9.5 & 0.8 respectively.A chelating material for the determination of A1 has been synthesized (95/3528) by the immobilization of Chromotrope 2B on an AG 1-X8 ion exchange resin. For a loading time of 1Omin at 1 ml min-' a detection limit of 10 ppb was obtained. A sorbent extraction procedure (95/3511) for the speciation ofFe based on the selective complexation of Fe" with ferrozine has been developed. This complex was retained from a 5 ml sample solution buffered at pH 4.5 on a C18 column (15 x 2 mm) whereas any Fe"' in the sample solution passed through the column and gave a broad signal. The retained Fe" complex was eluted with 200 p1 of methanol.The working ranges for the two species were 5-50 ppb and 250-10 000 ppb for Fe" and Fe'" respectively. The same material has also been used (95/2825) in a procedure for the determination of Mo in sea-water. The analyte was retained as the QSOL complex from 2.5 ml of a pH 3 solution. Elution was with ammoniacal methanol at pH9. An N20-C2H2 flame was used and an LOD of 5 ppb and a thoughput of 30 h-' were obtained. A variety of possible chelating agents were studied by Lancaster and co-workers (95/1891) for use in sorbent extraction with CI8. In their procedure the chelating agent was first loaded onto the column followed by the passage of the analyte solution. The relative merits of QSOL DDC 4-( 2-pyridylazo)resorcinol 1-( 2-pyridylazo)-2-naphthol and dithizone were evaluated for the determination of Cu and Pb.It was found that the best performance was obtained with Q80L and DDC for which LODs down to a few ppb were obtained with enhancement factors of between 50 and 100. A macro-porous resin coated with di(methylhepty1)methyl phos- phonate has been used (95/3681) for the determination of Au. The retained analyte was eluted with a thiourea-hydrochloric acid mixture. Simplex optimization was used to select the operating parameters. Recoveries from synthetic salt mixtures of between 93 and 110% at a concentration of 20 ppb were obtained at a throughput of 48 h-'. For the determination of Pb (95/2508) a column material was prepared by coating silica gel with the chloroform extract of the ion pair formed from Aliquat 336 (methyltricaprylammon- ium chloride) and nitroso-R-salt.The samples buffered at pH 5.5 were passed through the column at 1 ml min-'. The retained Pb was eluted with 150 pl of 0.1 mol 1-' hydrochloric acid. Serious interferences from nickel and fluoride were observed but at the concentrations encountered in natural waters were not significant in the determination of Pb in this matrix. Activated carbon has been used (95/3451) in a procedure for the determination of Pb in skimmed milk powder and some mineral RMs by retention of the APDC or dithizone complex from pH 1 solution. For 2 min loading at 3 ml min-' LODs of around 10-15 ppb were obtained. Alumina has been used to retain Pt as the chloro complex (95/1840) in a procedure based on retention from pH 2.5 hydrochloric acid solution and elution with 25 pl of 2 moll-' ammonia solution.The manifold design was such that unretained sample components passed into the nebulizer and the eluted analyte was passed through the column rather than being back-flushed. It is well known that this design is inferior to the alternative manifold in which the column is located in the loop of the injection valve but it is not clear why the authors selected this design. The procedure was applied to the determination of Pt in spiked tap water. A strong cation exchange material has been used (95/3629) to retain Sr in a procedure for the measurement of this element in natural mineral water. A water carrier stream was used with 2 mol 1-' hydrochloric acid as the eluent. The detection limit was 15 ppt.Natural waters have also been analysed (95/4482) by a procedure in which Cd Co Cr Cu Fe Mn Ni Pb and Zn were preconcentrated by retention on a poly (aminophos- phonic acid) column after adjustment of the pH to 5. Elution was with 100 p1 of 3 mol 1-l hydrochloric acid solution. An average preconcentration factor of 19 was obtained LODs were 0.5-5 ppb and a water SRM was analysed accurately. The procedure reported last year in a conference presen- tation (see J. Anal. At. Spectrom. 1995 10 199R) for the determination of Cd involving retention of the DDC complex on the interior surface o f a PTFE knotted tubular reactor has now been published (95/3401). The material retained from about 4ml of sample solution at pH 9.2 was dissolved in IBMK to give an enhancement factor of 66 and a concentration efficiency (defined as the product of the enhancement factor and the sample throughput in samples min-') of 6 1 min-'.The LOD was 0.1 ppb. Interferences from copper and iron were overcome by the addition of thiourea and ascorbic acid-1,lO-phenanthroline. The method was applied to the determination of Cd in rice and hair. The same laboratory (the Flow Injection Analytical Research Centre of the Institute of Applied Ecology Academica Sinica in Shenyang) has also published procedures based on this chemistry for the determi- nation of Ag in geological materials (95/334) Cu in rice drinking water and sea-water (95/2111 95/C1993). In the first procedure (95/334) Ag was collected by co-precipitation with iron(11) and DDC with dissolution in IBMK.The enhancement factor was 26 the concentration efficiency was 28 min-' and the detection limit was 0.5 ppb. Interestingly enough the interference from large amounts of iron in the sample was masked by the addition of 1,lO-phenanthroline. In the second procedure (95/21 l l ) the Cu was retained as the DDC complex and eluted with IBMK. An enhancement factor of 120 a concentration efficiency of 92 min-' and a detection limit of 0.5 ppb were obtained. The interference from iron was removed by the addition of ascorbic acid. A somewhat similar procedure was reported (95/4511) by Danish workers for the determi- nation of Cd. The analyte was precipitated with iron(II)-APDC retained on knitted tubular reactor and eluted with a discrete volume of IBMK (rather than a continuous stream) which was injected into a nitric acid carrier for transport to the spectrometer.The LOD was 74 ppt but no information about the concentration efficiency was provided. Spanish workers (Mir et al. At. Spectrosc. 1995 16 127) have determined Co in steel by precipitation as the acetylacetonate complex reten- tion on a sintered glass filter (pore size 10-16 pm) located in the external loop of a 6-port rotary injection valve and elution with 80 p1 of IBMK. An enhancement factor of 70 was obtained from a 500-p1 injection volume with a working range of 8-500 ppb. It was claimed that Co in steels at concentrations of between 100 and 500 ng was determined accurately. Further work on the indirect determination of anions (see 286 R Journal of Analytical Atomic Spectrometry August 19'26 Vol.1 1J . Anal. At. Spectrom. 1995 10 199R) by FI precipitation and dissolution of metal salts has been described by Esmadi and co-workers. Dichromate oxalate or carbonate were determined uia their silver salts. The calibrations covered the range 0.01-0.22mmol l-' with LODs between 1 and 8 pmol 1-'. The procedure has been extended (Esmadi et al. J. Flow Inject. Anal. 1995 12 35) to allow the determination of each anion in the presence of chloride (which is also determined). Following dissolution of the oxyanion salt in acid the chloride is dissolved in ammonia. Hungarian workers (95/4424) have used the dissolution of inorganic salts (copper sulfide or silver halide) by complexing ligands as the basis of indirect pro- cedures for the determination of the ligands.It appears that there have been no reports of the use of FI liquid-liquid extraction procedures with FAAS during this review period. In the broad category of FI and AAS are numerous papers in which experimental work is described where an FI procedure is used for chemical vapour generation. These studies are considered under section 1.3. unless the atomizer was a graphite furnace. In this case the work is surveyed in section 1.2.2.3. 1.1.4. Interference studies The identification and elimination of interference effects is one of the most difficult steps when applying FAAS to an analyt- ical problem. Optical interference may take the form of either background emission and absorption or spectral line overlap. Physical and chemical interferences can occur at any stage between the raw sample and the final absorbing atom.In the case of interfering background absorption by macroelements (e.g. Ca K Mg Na etc.) when determining trace levels of heavy metals (Cd Co Ni and Pb) in biological samples (soils plants food products) interference was minimized by extracting the metals from the aqueous matrix solution (pH 9) with dithizone in CCl (95/1170). The effect of spectral overlap of the Mn 403.307 nm line by the 403.298 nm line of Ga has been corrected by means of a partial least squares model (95/2310). It was suggested that the method could be developed for the simultaneous determination of several elements using a multi-element hollow cathode lamp. A computational technique for the elimination of interfering eflects in FAAS has been proposed by Wu et al.(95/4519) and applied to the determination of Ca (95/4409). A polynomial was used to describe the interferent effects occurring between analyte-interferent pairs e.g. Ca-A1 Ca-PO:- and the coefficients for 10 pairs tabulated. In addition to the interaction of sample components the reagents used in processing may themselves cause interference. The mechanisms of enhancement and suppression effects of surfactants in FAAS were studied by Wang and Wan (95/3667) using Mo as the analyte ion. Surfactant micelles were found to react with analytes and co-existing ions to produce signal suppression. Incomplete combustion also contributed to the suppression. A review of the analytical application of organized media interfaces in atomic spectrometry has been presented by Sanz-Medel (95/C4235).The benefits of surfactants in the generation of volatile species with NaBH and in LC separa- tions were noted. Interference effects are usually studied in the context of specific analyte element-matrix combinations; however a number of reports relating to more general situations have been published and will now be presented. In the determination of Ag by FAAS ethanolamine (EA) was used to eliminate the interfering effects of some organic and inorganic compounds (95/4345). The releasing action of EA was attributed to decomposition of an Ag-EA complex and to the presence of excess EA which generated reducing fragments in the flame. The interference efects of 40 single and mixed inorganic ions as well as 24 organic substances on the Al absorption spectra in an N20-C2H flame were investigated (95/3395).The mechanisms of the interferences were explained in terms of physical spectroscopic and chemical influences. In the determi- nation of Cd by FAAS potassium cyanide was used to eliminate the interference effects of some cation and organic compounds (95/3814). The interferences of acids were overcome by prepar- ing both standards and samples in 0.02 mol 1-' solutions of the acid. The determination of Ca by FAAS was found to be less subject to interference than when FAES was employed (96/111). The sensitivity of the latter was however the greater. Ascorbic acid citric acid and EDTA are widely used as inter- ference reducing agents. These reagents were used to reduce the interfering effects of Mg Sr and ammonium at various concentration levels on the determination of Pb (95/1894).Interferences in the speciation analysis by GC-AAS of organolead and organotin compounds in various samples were studied by Szpunar-Lobinska et al. (95/3808). Particular atten- tion was given to the effects of matrix co-extraction and reagent purity. Their influence on LOD was discussed in terms of baseline noise blank value artefacts and signal suppression. Loss of column resolution was noted during the analysis of some matrices. The FAAS determination of Mo in aqueous solutions of tetrathiomolybdate (TTM) and paramolybdate (PM) species using an N,O-C2H flame was studied by Voulgaropoulos et al. (95/3374). In the case of TTM there was an 85% loss of sensitivity which was reversed when TTM was acid digested and neutralized with 0.14 mol 1-' ammonia solution.Interfering effects notably Cu in the determination of Se were eliminated by the addition of n-butylamine with thiourea to sample and standard solutions in an acidic medium (95/3654). The mechanisms of the elimination were investigated by a continuous inhibition-release titration method to examine the stoichiometric ratios between the analyte and interferents. In a series of papers Belloni and co-workers (95/1923,95/1924 95/1925) examined interferences in the determination of I/. By careful optimization of the air-C,H flame conditions and use of 20% PrOH solutions sensitivities and freedom from inter- ference comparable with those of the N,O-C,H flame were obtained. The method was applied to the determination of V in steels.When samples are presented to the FAAS system as emulsions slurries or homogenates calibration can be a major problem. By selection of an appropriate medium for dispersion of the sample calibration can be greatly simplified. This approach was successfully followed for the determination of Z n in rat liver homogenate (95/3699). Water HC1 and tris- acetate buffer (pH 8.7) were tested as dispersing media. Hydrochloric acid (1.0 mol-') was found to be the most satisfactory medium while simple standard solutions prepared in acid or water were adequate. 1.1.5. Sample pre-treatment The material covered in this section follows the pattern estab- lished in previous years namely that a paper is only discussed here if there is some element of novelty in the sample pre- treatment and there is some possibility of the procedure being implemented in an on-line format.Most of the procedures surveyed here are therefore separation procedures in which the analyte and potentially interfering matrix are separated. In addition the procedure may produce an increased concen- tration of the analyte species in the solution presented to the spectrometer. Solid phase extraction procedures feature prominently in this section. A variety of materials were evaluated (95/1012) for their suitability for use in the determination of A1 in waters and renal dialysate solutions. The best performance ( 500-fold Journal of Analytical Atomic Spectrometry August 1996 Vol.1 1 287Renrichment) was obtained from silica gel or Amberlite IRA 400 modified with Chrome Azurol S. Amberlite XAD-1180 was used (95/3482) in a procedure for the determination of Bi Cd Cu and Pb in high purity zinc. The sample solution in 0.1 moll- sulfuric acid also contained 0.05 moll -' potassium thiocyanate and 0.4 mol I-' sodium iodide). Detection limits of around 0.2ppm were obtained. Further studies on the synthesis of chelating resins have been reported by workers at the Indian Institute of Technology (95/1228 95/3102). In the first study (95/1228) Amberlite XAD-2 was modijied by the immobilization of Alizarin red-S and in the second (95/3102) by derivatization with salicylic acid. The resins were used for the determination of Cd Ni Pb Zn and Pb Zn respectively.The salicylate material had a higher sorption capacity for both Pb and Zn leading to preconcentration factors in the methods used of up to 180. The materials were used in the determination of metals in well-water. The interferences from a number of commonly encountered anions were investigated. The resins showed slightly different behaviours in that the Alizarin red-S material showed susceptibility to interference from nitrate whereas the salicylate material was susceptible to phosphate. Chinese work- ers (95/4405) prepared thiosemicarbazide cellulose in a two step synthesis and then used the material in a procedure for the determination of Au. A number of sorption extraction procedures have been described. Thallium (95/3682) was retained on CI8 silica as an ion pair [ tetrachlorothallate(rr1) and a cationic surfactant] prior to elution with 96% ethanol.It was found that a more sensitive determination was obtained by emission from the N,0-C2H2 flame in comparison with absorption in this flame. A CI6 cartridge has been modijied with Q80L by Russian workers (95/2377) for the preconcentration of Cd and Pb prior to elution with 0.1 moll-' EDTA solution. Silica gel modified (95/4532) by suspension in n-propanol containing sodium diethyldithiocarbamate was used to collect Cd Pb and Zn prior to elution with EDTA solution buffered with ammonia- ammonium chloride. A number of reports have described the retention of metal chelates on activated carbon (95/C4226 95/3039 95/3372 95/4028 and 96/313). The procedure has been applied to the determination of Mo in silicates (95/3372) in which the analyte was adsorbed onto the surface of 60mg of activated carbon as the APDC complex the determination of Ni (APDC was also used) in urine (95/C4226) the determi- nation of Cd and Pb (95/3095) in vegetables (Q80L or cupferron was used as the ligand) the determination of V (95/4028) in biological materials and the determination of Co (96/3 13) in vegetables.The procedure was similar in all cases. After stirring for up to 1 h the carbon was separated by filtration and dried. The metals were redissolved in nitric acid and then measured at enrichment factors of up to 100. In the determination of Cr"' and CrV' (95/4688) a froth Jlotation procedure with oleic acid as surfactant was used to separate the butylamine complex of Cr"'.Butylamine was also found to act as a releasing agent eliminating the interference effect of a number of inorganic and organic species. A similar procedure was used by the same workers (96/311) for the determination of Pb. The precipitated chromate was separated by froth flotation and dissolved in dilute nitric acid. The procedure was applied to the measurement of Pb in non- ferrous alloys. There are still reports of liquid-liquid extraction based on APDC and MIBK appearing (95/C1991) although it is difficult to see what is novel. A procedure (95/3683) for the determi- nation of As Bi Ga In Sb and Sn based on the extraction of the iodides into MIBK incorporated a back-extraction with a benzene-nitric acid-water mixture and a procedure (95/3874) for the determination of Cu Fe Mn and Zn in water was based on the extraction of the APDC complexes into MIBK- xylene followed by back-extraction into concentrated nitric acid dilution with water (1 + 9) and nebulization. Chinese workers determined Cd and Pb in barite (95/4522) by extrac- tion of the ethyl xanthates into MIBK and back extraction into hydrochloric acid solution. Some further applications of the liquid membrane procedure described last year (see J.Anal. At. Spectrom. 1995 10 199R) have appeared in the Chinese literature (95/2834 95/4418). In this method the analyte elements were extracted into the non- aqueous phase of an emulsion consisting of isooctyl phosphon- ate and liquid paraffin stabilized with a surfactant poly( bis- succinimide). After extraction and standing the aqueous phase was discarded and the remaining phase de-emulsified and nebulized.The procedure was applied to the determination of Co Cu and Ni (95/2834) in lithium salts. A number of precipitation procedures have been described. Copper was determined (95/1951) in sulfide ores by a method involving precipitation with 2-amino-3-hydroxypyridine. For the determination of Cd in tobacco (95/3621) the analyte was collected by co-precipitation with strontium carbonate separ- ated by centrifugation and re-dissolved in a small volume of hydrochloric acid. Chromium was determined in fish bone meal (95/4032) by coprecipitation of Cr"' by iron(II1) hydroxide allowing CrV' to be determined in the supernatant. The Cr"' was determined following redissolution of the hydroxide pre- cipitate.Reductive precipitation was used to separate the matrix (96/368) prior to the determination of a number of trace elements in tellurium oxide or bismuth oxide. After dissolution of the sample in hydrochloric acid the matrix element was precipitated by the addition of hydrazine (as the hydrochloride sulfate or monohydrate). 1.1.6. Chromatographic detection There have been a number of overviews of the role of chromato- graphic separation in speciation studies (95/ C3750 96/C224 95/3692) with particular reference to application in biological and environmental materials. A high pressure nebulizer was used (95/3494) to couple HPLC and FAAS for the determi- nation of Mg in solutions containing aluminium which would interfere if introduced at the same time as the analyte.The procedure was also used to determine Ca in saturated sodium chloride solutions. Various complexes of Cr"' and CrV' were determined (95/1884) by a procedure in which the anionic species were retained and separated by a fast protein liquid chromatography anion-exchange column of Mono Q HR 5/5. Cationic species and kinetically labile negatively charged species such as Cr"'-citrate passed through the column. A FI system was used to interface the chromatographic separation with the spectrometer and the method was applied to the determination of Crvl Cr"'-EDTA and Cr"'-oxalate in cab- bage xylem at ppb concentrations. Size-exclusion chromatog- raphy interfaced with a variety of detectors including a microatomizer device consisting of an oxygen-hydrogen flame and thermospray (see Anal.Chem. 1992 64 3197) has been used for the characterization of metal binding proteins extracted from rabbit liver freshwater mussels and osprey blood (95/3534 and 95/3603). These metallothionein and metal- lothionein-like proteins have a high proportion of cysteine residues capable of binding a variety of metals. It was suggested that the redox state of these residues could be probed by the addition of cadmium in the presence or absence of a reducing agent; however some problems were encountered in the appli- cation of this procedure to the molluscan tissue owing to the presence of 2-mercaptoethanol in the extract indicating that the metallothionein-like proteins may be partially oxidized in uivo and/or during isolation under anoxic conditions.It should be noted that many of the uses of AAS as a chromatographic detection mode are classified under the 288R Journal of Analytical Atomic Spectrometry August 1996 Vol. 11sections dealing with various aspects of chemical vapour generation (see section 1.3.). 1.2. Electrothermal Atomization As one would expect in a mature analytical technique there have been few major developments during the time period covered by this Update. However in a number of conference presentations various key workers in the field of ETAAS have considered the future of the field and reviewed its capabilities in comparison with competitive techniques such as ICP-MS. Most of these had provocative titles! Sturgeon (96/C228) discussed the use of a graphite electro- thermal atomizer and its role in atomic spectroscopy.An extra- ordinary diversity of applications of the graphite atomizer were shown encompassing the fields of physics thermochem- istry spectroscopy and analytical chemistry. Although many may think that ETAAS is declining in use as a tool for ultratrace analysis owing to the availability of competitive techniques such as ICP-MS the attractive features of ETAAS such as high atomization/vaporization efficiency comparative long atomic vapour residence times controllable chemical and thermal environment and its ability to handle high dissolved solids will serve to ensure its place in analytical atomic spectrometry for many years to come. Recent developments in the design and application of ETAAS atomizers were presented by Frech (96/C220).Currently a platform equipped trans- versely heated graphite atomizer (THGA) most closely approaches an ideal system. There is still room for improve- ment such as the addition of end-caps to the tubes that restrict the open ends and as a result atomic vapour containment within the atomizer can be increased improving sensitivity without resorting to longer tubes which is technically difficult. The present status of ETAAS in comparison with other AS techniques was presented by Hoenig (95/C3737) who con- sidered that the use of simultaneous multi-element ETAAS has removed one of the major disadvantages of ETAAS in that sample throughput can be considerably improved. In addition the ability of ETAAS to handle high dissolved solids and its advantages for biological fluids and slurries will preserve a role for ETAAS against techniques such as ICP-MS which are more dependent on the matrix composition of the sample.In comparing ICP-MS and ETAAS McLaren et al. (95/C848) considered them both to be complementary and although ICP-MS is the National Research Council of Canada's main analytical technique for the production of CRMs partly owing to the ability to perform rapid multi-element isotope dilution analysis ETAAS is also a very important analytical technique and they do not rely solely on ICP-MS. Continuing the theme of comparing ICP-MS and ETAAS Gregoire and Gregoire (95/C760) asked the question 'to atomize or not to atomize?' and whether ETV-ICP-MS will ever supersede ETAAS as the combination of ETV with ICP-MS gives one the benefits of both techniques.They considered that the limited number of ETV-ICP-MS applications in the literature is due to the lack of published fundamental studies on such things as analyte transport efficiencies the use of physical carriers and chemical modifiers and analyte vaporization mechanisms. This would certainly seem an area ripe for further work and complemen- tary to that which has already been performed for many years with respect to ETAAS. Finally Welz and Sperling (95/C8466) asked with a certain degree of humour 'ETAAS-is it still worth the trouble?' However they showed that although ETAAS is a mature technique it has a number of unique properties. The world- wide inter-laboratory stability and reproducibility of the characteristic mass (mo) within 10-20% is a feature that cannot be approached by any other spectrometric technique.The ability to tolerate high dissolved solids such as undiluted sea- water blood or urine and use of a transversely heated atomizer with a combined palladium and magnesium nitrate chemical modifier has produced a widely applicable robust analytical technique. Combined with the ability to perform simultaneous multi-element determinations this has eliminated one of the disadvantages of ETAAS. For further improvements the coupling of FI automatically and on-line with the atomizer allows matrix separation and preconcentration taking routine achievable detection limits down to the low ng I-' range.A few reviews were published during this period. Li and Matsumoto (95/1917) produced a Japanese review with 70 references which discussed recent developments in methods sample introduction techniques and atomizer design. Slavin (95/3966) reviewed the use of ETAAS for the determination of trace elements in biological specimens. 1.2.1. Atomizer design and surface modiJication 1.2.1.1. Graphite atomizers. For some time it has been clear that the surface within a graphite electrothermal atomizer plays an important role however to date very little information has been collected regarding the surface changes that take place during the repeated heating and cooling cycles to which a graphite atomizer is subjected. Habicht et al. (95/4732) applied atomic force microscopy to observe the nanotopographical changes on graphite tube surfaces in an ETAAS atomizer.These revealed severe topographical modifications on both the micrometre and nanometre scale after only a few atomization cycles for the determination of Cr in 0.1 mol 1-' nitric acid with an atomization temperature of 2700 "C. Initially pro- trusion diameters varied between 5 and 7 pm. With an increase in the number of firings the relatively smooth surface of these protrusions became increasingly rough and the diameter reduced to between 300 and 600 nm. During the first atomiz- ation cycles the protrusion size distribution was found to become more uniform which is in agreement with the observed increase in reproducibility of ETAAS measurements after 10-20 atomizer cycles. The homogenization of the graphite surface was interpreted as the initial phase of a secondary coating.There is no doubt that this work will lead to a greater understanding of the role the graphite surface plays in ETAAS and further publications from Professor Ortner's group are awaited with interest. During the period of this Update a number of workers have compared the performance of conventional longitudinally heated Massmann-type atomizers with transversely heated graphite ato- mizers. While there is no argument that an isothermal atomizer will produce the best over-all performance there is considerable discussion as to which designs and operational conditions result in performance that comes close to an isothermal atomizer. In recent studies it has been shown that the static wall temperature of a graphite tube and also the gas-phase temperature measured by the two-line method are poor indi- cators of the internal gas-phase temperature.In an extensive study Welz et al. (95/C657) discussed the differences in spatial and temporal gas phase temperature distributions within both of these different atomizer designs obtained with coherent anti- Raman stokes (CARS) thermometry measurements. Transversely heated atomizers show a uniform temperature distribution within & 50 "C under steady-state conditions whereas during rapid heating the temperature at the tube end is as much as 900 "C higher than over the platform depending upon the start temperature. Longitudinally heated atomizers show a pronounced gas-phase temperature gradient largely independent of the experimental conditions.This results in condensation of the analyte and matrix species at cooler parts of the tube and in the gas phase leading to increased atomic peak tailing memory effects spectral and non-spectral inter- ferences. The memory effects and peak-shape tailing could be Journal of Analytical Atomic Spectrometry August 1996 Vol. 1 1 289Rreduced by introducing a moderate internal gas flow however this is at the expense of reduced sensitivity. They found that a longitudinally heated atomizer cannot approach the perform- ance of a transversely heated one under any circumstances. From a practical approach Granadillo and Romero (95/C640) compared the use of a Massmann-type atomizer with continuum source background correction (HGA) with a transversely heated atomizer and longitudinal Zeeman-effect background correction for the determination of Cr in biological and environmental samples.They found that a high pyrolysis temperature 1650 "C could be used which removed all matrix constituents and the need for background correction and chemical modification. Wall atomization was preferred to the integrated platform of the THGA design and the LODs (3s) were 0.03 and 0.2 pg I-' respectively. The accuracy of the method was verified by the analysis of seven RMs and recover- ies were in the range 94-103% with within- and between- batch %RSDs 3.1 for HGA and 3.7 for THGA. These data seem to contradict the findings of other workers in fact the LOD for the THGA system in this work is a factor of two worse and the HGA 3-fold better than what the manufacturer claims.In contrast Hinds (95/C647) compared the use of HGA and THGA atomizers for the determination of Ti and other elements in a gold matrix. For most elements both atomizers provided adequate results however the HGA atom- izer was very insensitive for elements such as Pd and Pt. Titanium could only be determined with the THGA system because of the reduction in analyte memory effects. For both atomizers analyte sensitivity decreased as the the amount of precious metal matrix increased owing to the effect of con- densed metal particles forming in the gas phase. Bradley and Leung (95/936) compared HGA with platform atomization and THGA with integrated platform atomizers for the determi- nation of A1 in plasma and urine.Both systems gave compar- able results with good agreement between them with a correlation coefficient of 0.995. The advantages of the THGA system were that it allowed a lower atomization temperature and that the analyte memory effects and analysis time were significantly reduced. The use of integrated contact cuvettes discussed last year (see J. Anal. At. Spectrom. 1995 10 199R) has now been published (95/4372). A transversely heated graphite atomizer with continuum source backround correc- tion was employed for the determination of Cr and Se in serum and Pb in urine using three different types of graphite tube design. A platform-free tube a tube with fork platform and a ring-shaped platform were used. The tubes without a platform and with a ring-shaped platform were suitable for automated analysis in that there was no need to change the tube but the fork-shaped platform showed low sensitivity for Cr.Compared with the platform-free tube the tube with the ring-shaped platform showed an improved S/N. The design of this atomizer and the ring-shaped platform and tube are outlined in two patents (95/1907 95/1908) respectively. In a novel application of the double atomizer for ETAAS discussed last year (see J. Anal. At. Spectrom. 1995 10 199R) Smith and Harnly (95/C642) were able to differentiate between inorganic and porphyrin bound Fe. An inlet hole was drilled in the bottom of the integrated contact cuvette (ICC) atomizer to admit gases from the graphite cup placed below. The sample was placed in the cup and heated to vaporize the sample.The sample vapours then passed into the heated upper atomizer where atomization took place. This double atomizer was located within a modified vacuum cross which permitted operation at low pressures (10 mTorr; 1 Torr = 133.322 Pa) and high pressures (10 atm; 1 atm= 101.325 Pa). The combination of slow heating of the cup and atomization under reduced pressure allowed separation and quantitation of the different types of Fe. There is as yet no perfect electrothermal atomizer and workers continue to investigate and propose modifications to all the available commercial designs. Even with a relatively recent design such as the transversely heated electrothermal atomizer a new design of graphite tube incorporating end- caps was discussed in various presentations during the last few years (see J. Anal.At. Spectrom. 1994 9 213R and J. Anal. At. Spectrom. 1995 10 199R) and has now been published (95/1019). Hadgu and Frech (95/1019) compared end-caps with apertures of 2 3 and 4 mm in THGA graphite tubes. The presence of the end-caps reduced diffusional losses of analyte and matrix through the tube ends and enhanced gas-phase temperatures. The increase in sensitivity with end-caps with an aperture of 3 mm and a thickness of 1.5-2mm were established for 13 elements and found to be in the range of 1.7-2.5-fold. As the end-caps do not reduce spectral through- put S/N are improved provided that the analyte atomization times are comparable. The background absorbances for sodium sulfate and sodium chloride obtained with end-capped tubes were found to be similar to those obtained with standard tubes and hence even in the presence of background signals the S/N are improved with the end-capped tubes.The design of tubes equipped with end-caps is described in a patent (95/1909). Cimadevilla et al. (95/1485) continue to investigate and compare wall platform and probe atomization this time turning their attention to the direct determination of As Cd and Pb in sea-water. The best analytical performance was obtained using total pyrolytic graphite platforms inserted into pyrolytic graphite coated electrographite tubes with peak absorbance measurement and the addition of 10 pg of palladium nitrate as chemical modifier. For As the chemical modifier was added with the sample but for Cd and Pb the chemical modifier was dried first and pre-reduced on the graphite surface before addition of the sample.Sea-water was injected directly for As; for Pb a 1+ 1 dilution was preferred and for Cd a 1+3 dilution with 25% m/v ammonium nitrate. LODs (3s) with respect to undiluted sea-water were found to be 1.7 1.0 and 1.3 pg 1-' for As Cd and Pb respectively. The methods were applicable to the monitoring of contaminated sea-water but for the assessment of 'normal' levels preconcentration tech- niques would have to be employed. Chinese workers (95/3638) examined the use of a probe atomizer for the determination of In in human hair. The In content of human hair was found to be 12-158 ng g-' (90 samples). No details were given in the abstract of the sample preparation procedure.Deng and Gao (95/1024) found that lanthanum nitrate heated on a graphite probe produced blisters on the graphite surface. The formation of these was ascribed to the formation of La(OH) produced from La,O and water vapour on the graphite probe surface. Katskov et a!. (95/C648) discussed the further benefits of a graphite electrothermal atomizer containing a graphite filter. A V-shaped pyrolytic graphite coated electrographite boat inserted into a pyrolytic graphite coated electrographite tube was used by Zheng and Wang (95/1042) for the standardless determination of T1 though there is no indication of the advantages of such a boat over a conventional platform. Hocquellet (96/3 12) applied an arch-shaped graphite vault coated with tantalum carbide inserted into a Massmann-type atomizer for the determination of Cd in microwave digestions of plant and animal tissues.The initial atomization took place at 1650°C followed by cooling at 20°C and then reatomiz- ation from the vault at 1500 "C. The claimed advantages are that the method was less susceptible to spectral and chemical interferences than platform atomization. 1.2.1.2. Metal atomizers and metallic coatings. The work produced during the period covered by this update is summar- ized in Table 1. A steady stream of work concerning metal atomizers and metallic coatings was produced though there 290 R Journal of Analytical Atomic Spectrometry August 1996 Vol. 11Table 1 METAL ATOMIZERS AND SURFACE MODIFICATIONS ~ Atomizer material/metallic coating Element Matrix Au-coated Hg Hair urine soil and leaves Comparison of electrographite Ba pyrolytic graphite coated electrographite and W-coated electrographite Ir-coated tube Ir-coated tube Ir-coated tube Mo-coated graphite tube Mo-coated graphite tube Mo-coated graphite tube Mo-tube atomizer Pd- and Rh-coated tubes W metal atomizer W-atomizer WETA 90 W-coated graphite tube W coil W coil Citrus leaves Various (5) - Various (9) - - Fluoride Ga In T1 - Ni Serum Au Biological As Se Au Be Sn Pb River and waste water Drinking and waste water Serum Whole blood Various (6) Ground and tap water Sample treatment/comments Hg vapour trapped within a gold-coated graphite atomizer.Atomization temperature of 500-600 "C. Optimum characteristic mass and lifetime with pyrolytic graphite coated tube without a platform and Ar purge gas method of analyte additions applied. Tungstate coated tube performance poor owing to discontinuous coating.Use of a solid chemical modifier (Ir) deposited onto surface by means of cathodic sputtering; homogeneity and distribution of Ir examined. See ref. 9514689. Homogeneous layer of Ir deposited on inner graphite surface by cathodic sputtering. Performance of permanent modifier assessed for Cd Mn Pb Se and V. Integrated absorbance remained constant for 700 firings but atomization of Mn and V unfavourably delayed in Ir-coated tubes. See ref. 95lC2271. W-carbide coated graphite platform for hydride trapping studies. Ir-Zr-carbide was the preferred combination. AIF molecular absorption spectrometry used to determine F.Large enhancement in sensitivity with Mo-coated tube. A Pt hollow cathode lamp used as light source. Examination of evaporation behaviour and interference effects. Molybdenum participated in atomization process changed mechanism and improved atomization efficiency. Chemical modifiers of ammonium vanadate and lanthanum atomization at 2700 "C LOD ca. 40 pg 1- ' recoveries 95-1 11%. See ref. 961108. Samples decomposed with 14 mol I-' HN03 and 30% v/v H,O evaporated to dryness and dissolved in thiourea Ar-H purge gas LOD 1.3 pg %RSD (n= 10) 2.9% for 250 pg Au. From same group as ref. 9512623. Various types of graphite tube electroplated with either Pd or Rh. Found to reduce losses of As and Se and reduce influence of oxide and carbide formation (e.g.Si). Pretreated tubes found to have lifetimes of up to 160 firings with an Rh coating. Au"' preconcentrated as chloro anionic complex on chitin. Chitin suspension (10 pl) taken for analysis LOD 0.1 pg l-' recoveries 98-103%. Use of H2 to protect atomizer Al(N03) chemical modifier peak-absorbance measurements LOD (3s) 0.16 pg for 20 p1 sample volume calibration by method of analyte additions. Sample pre-digestion with HN0,-H,O chemical modifiers of ammonia and molybdate recoveries of 96-101 YO. See ref. 9513394. Prototype instrument aim was to develop a portable AA instrument for screening purposes. Sensitivity in peak-absorbance mode better than integrated absorbance D background correction RSD in peak absorbance mode ca. 3% for an aqueous Pb concentration of 100 pg I-'.Ar-H (90+ 10) mixed gas. Parameters studied current applied to coil; temperature and duration of three heating steps (dry pyrolysis and vaporization); purge gas flow rate; coil alignment in radiation beam; and sample volume. LODs (3s) ranged from 0.6 (Cd) to 23 (Pb) pg 1-' for a 10 pl sample volume. Iridium 2 pg deposited on a Zr- or Unenclosed W-coil atomizer operated in Reference 951966 9511675 95fC227 1 9514689 96/28 1 9513807 951945 9513394 9513452 95/550 961401 9511486 961 108 95/C43 1 1 9514356 Journal of Analytical Atomic Spectrometry August 1996 Vol. I I 291 RTable 1 (continued) Atomizer material/metallic coating Element Matrix Sample treatmentlcomments W-tube atomizer Zr-coated platform Zr-coated platform Zr-coated tube Zr coating A1 Biological Use of H to protect atomizer LOD 52 ng 1-' severe interferences overcome by method of analyte additions.From same group as ref. 9513452. HF-HN03-HC10 residue dissolved in HCl and extracted with IBMK 20 pl of organic phase taken with addition of Pd and Ni as mixed chemical modifier. LODs 0.21 and 0.31 ng for Se and Te respectively. Sn Tap water Use of a mixed palladium and magnesium nitrate modifier and a pyrolytic graphite coated tube preferred as reduced problems with matrix effects. Interferences due to calcium and improved precision. Chemical modifier of Ni and Zr interference effects due to the major component iron and 8 minor components were studied. Zr( NO& under reduced pressure then heated to 120°C for 2 h. Se Te Geological Samples (1 g) digested with B Iron- and nickel-based alloys Sb Sn Not given Graphite tubes immersed in 6% m/v Reference 9512623 9513503 95/21 14 9511877 961390 seems to be little new material.Many of these are from the same research groups and generally are the determination of a different element or the same element in a different matrix with the particular metallic atomizer or coating that happens to be the choosen atomizer for the particular group. Little of the work with these atomizers appears to show a distinct trend or the near universal applicability that one sees with a graphite electrothermal atomizer. Two areas show some interesting developments. The group led by Parsons and Slavin (95/C4311) at the New York State Department of Health have taken the tungsten-coil atomizer and applied it to the development of a low cost portable system for the determination of Pb in blood.It could be that concentrating on applying this atomization system to only one particular application may find a niche for this atomizer. As discussed in previous reviews (see J. Anal. At. Spectrom. 1994 9,213R and J. Anal. At. Spectrorn. l995,10,199R) the tungsten coil atomizer does suffer from some severe interference effects in spite of its attractions as a simple low cost unit. It will be interesting to see how this develops further. Of particular note is the interest being shown in developing permanent metallic coatings as in situ chemical modifiers (95/550 95/4689 95/C2271,96/281). This work has its origins in the development of interest in hydride trapping within a graphite electrothermal atomizer.The realisation that an iridium metallic coating can last for several 100 atomizer cycles while trapping the hydrides has stimulated interest in not only procedures as to how the best coating of iridium can be achieved and the distribution and characterization of the coating on the surface (95/C2271) but as to how reliable the iridium coating is when a sample solution is pipetted on top as opposed to the trapping of a gas (95/4689). Currently all the proposed procedures whether electrodeposition (95/550) or sputtering (95/4689) can only treat one tube at a time. This does make the procedures time consuming in spite of the improvement in analysis times as a liquid chemical modifier does not have to be added and dried. This work also crosses the topic groups in this review and asks the question as to whether this work should be discussed under metallic coatings or chemical modifiers.For the present update the material is covered in both sections. 1.2.2. Sample introduction 1.2.2.1. Slurry sampling. Slurry sampling direct into an elec- trothermal atomizer continues to be a rich source of publi- cations and presentations. It would be fair to say that most of the fundamental parameters and requirements of this sampling technique have been established and virtually all the work examined in this review period mainly concerns applications. Atuya (95/4369) reviewed (107 references) the available studies on slurry/solid sampling for ETAAS and discussed the advan- tages and disadvantages. For solid sampling disadvantages are the introduction of the sample and the preparation of a calibration however slurry sampling overcomes these.It is always a pleasure to read of workers who present data on 100s of samples rather than a few RMs. Such studies clearly demonstrate that the technique is viable. Epstein et al. (95/C575) applied slurry sampling to assess the Pb contami- nation in 400 soil and paint samples taken from 53 parks and playgrounds. Soil Pb concentrations were found to vary between 0 and 400 pg g-' and paint Pb concentrations from 0 to 8% m/m. The slurry method had the advantage of being rapid and cost effective while retaining adequate accuracy and precision to allow conclusions to be drawn from the data. In a series of presentations Miller-Ihli (95/4723 95/C727 95/C3029) discussed the conclusions to be drawn from an international collaborative study designed to evaluate the current level of solid sampling capabilities.Samples were sent to 28 laboratories and data received back from 18 16 of whom reported slurry sampling results. In order to establish prelimi- nary performance the collaborators had to provide accurate results within f 10% of NIST SRM 1643c Acidified Water for the mean certified values for Pb and Cr. The average perform- ance for the determination of Pb in NIST SRM 2704 Buffalo River Sediment was 103% recovery based on the mean certified reference value. For NRCC PACS-1 a marine estuarine sedi- ment supplied as an unknown the recovery was 84% again based on the mean certified value.For Cr similar results were obtained with recoveries of 96 and 78% for the same materials respectively. The second phase of this study included samples of coal fly ash marine material glass botanical materials and a pre-made slurry. Results from phase 2 indicated the potential problems that can be encountered with solid/slurry sampling. Important findings were the usefulness of secondary wave- lengths and oxygen ashing during the pyrolysis step and that a representative sub-sample must be taken. Possible problems leading to inaccurate results being reported by the collabor- ators included the use of mini-internal gas flows chemical modifiers low atomization temperatures short atomization times and expulsion losses and indicate how important it is that ETAAS conditions for solid/slurry sampling be optimized.292R Journal of Analytical Atomic Spectrometry August 1996 Vol. 1 1However it has to be said that these last findings would most probably be found by any collaborative study involving ETAAS. As stated above much of the current work involving solid/slurry sampling is application related and there are few fundamental studies. The work covered during the period of this review are shown in Table2. One group Vinas et al. (951956 951982 951900 9512528 95/3840 95/4597 961103) from the University of Murcia Spain have published a series of papers on the application of slurry sampling for a variety of analytes and matrices. It would appear that the slurry sampling and preparation procedure was identical in nearly all cases.1.2.2.2. Gas sampling. Although an electrothermal atomizer would not appear to be the ideal sampler for gaseous samples there is steady interest in this application. Sneddon’s group have discussed a variety of approaches and have now published the work (95/3369 95/C2900) discussed in the last review (see J. Anal. At. Spectrom. 1995 10 199R). The electrothermal atomizer is used as an impaction collector that allows near real time analysis. In conjunction with a multi-element AA spectrometer this allows the determination of several elements in aerosols. Zhang et al. (95/4550) modified a graphite probe for use as a filter collection device for the determination of In in atmospheric particulate matter. Air was sampled at 270 1 min-’ for 1 h.The filter was attached to the probe control system of an electrothermal atomizer and In determined. An LOD of 21.5 pg was found and the method was applied to the determination of In in NIST SRM 1648 Urban Particulate Matter. Sadly no details of the calibration procedure used are available. Baaske and Telgheder (96/292) have continued working on the challenging task of determining trace amounts of heavy metals in gaseous hydrogen chloride. The original work on the determination of Fe (see J . Anal. At. Spectrom. 1995 10 199R) has now been extended to include Cr Ni and Mn. Using a probe atomizer modified to insert a quartz capillary into the atomizer a sample volume of 0.43 ml was delivered into the atomizer held at 1100°C with no internal gas flow. The sample volume was controlled by the length of time a magnetic valve was open.The hydrogen chloride was removed by flushing for 7s followed by atomization at 2300°C. Calibration was achieved by standard additions of the penta- carbonyl compound of the analyte element of interest in argon as there are no certified gaseous standards of elements in hydrogen chloride. Detection limits were found to be 1.1 39 0.4 and 8.7 pg at 25 “C for Cr Fe Mn and Ni respectively. The reproducibility of the integrated absorbance signals varied between 5.3 and 13% (n=10). 1.2.2.3. Coupled techniques and preconcentration. The coup- ling ofFI techniques with ETAAS continues to generate a large volume of work. What is interesting to note is that the workers active in this area are expanding and are no longer solely research-based but increasingly application orientated. There has been a great deal of progress in this area and it would appear that ‘in-atomizer trapping ’ primarily for elements amen- able to hydride generation has rapidly become a routine technique especially for those who require very low detection limits and where the sample volume is not a limiting factor. Willie and Sturgeon (95/C798) discussed the ‘in-atomizer trap- ping’ of hydrides for the determination of As Sb Se and Sn in environmental samples.They converted what were pre- viously batch methods into on-line FI procedures using a commercial software package developed for ‘in-atomizer trap- ping’. The benefits of this approach were rapid simple and sensitive methods with additional benefit of reduced sample consumption.Sturgeon et al. (95/C728) showed that ‘in- atomizer trapping’ coupled with HG permits significant enhancement in relative detection power over conventional methods for the determination As Bi Sb Se and Sn. Trapping the hydrides on the graphite atomizer surface effects a clean rapid separation and concentration. The graphite atomizer can then function as the vaporizer-atomizer in AAS or as a vaporizer in ETV-ICP-MS. An area that has received a great deal of attention is the trapping reagent used on the graphite surface. While trapping for some elements can be achieved on the graphite surface alone the use of a metallic trapping reagent does improve the trapping efficiency. Previously palladium had to be added prior to each trapping sequence but the realization that elements such as iridium could be used as ‘permanent’ trapping reagents has stimulated a great deal of work.Tsalev et al. (961281) investigated the use of 2 pg of iridium deposited on a carbide-coated platform pre-treated with 1.2- 1.3 pmol of zir- conium or tungsten as a permanent trapping reagent/chemical modifier for the trapping of hydride-forming and volatile elements As Bi Cd Pb Sb Se Sn Te and T1. The best characteristic masses for integrated absorbance measurements were achieved with iridium-zirconium-treated platforms. While these workers did not use ‘in-atomizer trapping’ but conven- tional solution sampling for the experimental work the con- clusions drawn do show the potential of these noble elemental combinations as permanent trapping reagents.The determination of Ge appears to be the most popular element for the use of ‘in-atomizer trapping’ during the period of this review especially with Chinese workers. Haug and Yiping (96/291) systematically examined the use of a variety of permanent trapping reagents for germane such as the carbide forming elements zirconium niobium tantalum and tungsten and noble metals iridium and palladium. Effective trapping of germane was possible on zirconium coated platforms at trap- ping temperatures of 600-800 “C. Good signal stability was observed for more than 400 complete trapping-atomization cycles with an RSD of 3% for 2.8 ng of Ge. The iridium-coated tubes allowed trapping of germane at lower temperatures 400-500 “C but the signals observed were smaller and with reduced stability compared with zirconium.An LOD (3s) of 18 pg was found for a 1 ml sample volume and the method was applied to the determination of Ge in sediment geological and low-alloy steel RMs. Ni and He (95/4690) determined Ge in environmental samples in dilute perchloric acid solutions (used to digest the samples) by continuous flow HG and ‘in- atomizer trapping’ with palladium as the trapping reagent. It is interesting that these workers examined the use of zirconium- treated tubes for trapping germane but were not successful hence the use of palladium even though this had to be added before each trapping cycle. A detailed study of both of these reports does not provide any reasons for the different findings.Ma et al. (95/1940) applied a laboratory made FI-ETAAS system to the determination of Ge in soil and Chinese herbal drugs. The generation of germane was performed in 20% m/v Na3P04 at 5 ml min-’ in a 40 cm reaction tube followed by 0.03 mol 1-l of H3P0 at 3.2ml min-’ in a 120cm long reactor before reacting with 0.5% m/v potassium borohydride. The generated hydride was transferred into an automatic sampler coated with palladium chloride and potassium borohy- dride then to the AAS instrument. Further details are not available but the HG chemistry appears interesting. Other Chinese workers (95/2253) successfully determined Ge with FI and ‘in-atomizer trapping’ with ETAAS using palladium as the trapping reagent and chemical modifier. Willie (95/4365) described an automated FI and ‘in atomizer trapping’ procedure for the determination of Pb using ethyl- ation.The Pb was converted into the volatile PbEt by reaction with NaBEt and collected in the graphite atomizer at 400 “C. An absolute LOD of 12 pg was found which corresponded to a relative LOD of 2 ng 1-l using a 5 ml sample loop. For the Journal of Analytical Atomic Spectrometry August 1996 Vol. 11 293RTable 2 SLURRY AND SOLID SAMPLING Type of sampling Slurry Element - Matrix Algae Slurry activated carbon impregnated Ag Au Sediments with dithizone Slurry Slurry Slurry Slurry of ion-exchange resin Slurry prepared by grinding and ultrasonic mixing in a bath Slurry from dried and ground material Ga TI Geological Various (6) Silica gel Various (9) Silicon nitride powder Cr Water Al Cr Vegetables c u Biscuits bread Slurry prepared from ashed residue A1 Chewing gum Slurry prepared on ground dried Cd Pb Vegetables of chewing gum to achieve small particle size samples Slurry prepared on ground dried samples Slurry samples pre-ashed at 400 "C for 1 h Cd Pb Various (5) Cereal-based products biscuits bread Chewing gum and sweets Sample treatment/comments Use of algae to separate and preconcentrate trace elements study of background interference which could be eliminated by selection of appropriate pyrolysis time and temperature signal recording and heating mode. Sediment digest extracted with 25 mg of activated carbon impregnated with dithizone.Slurry of activated carbon suspended in 5 ml H,O and 10 yl taken for analysis. LODs 0.05 and 0.4 yg 1-1 for Ag and Au respectively.solution of 1 mgml-' Ni 1 mgml-' V and 0.15% m/v agar-agar. Mixture stirred for 5-10 min 20 pl aliquots taken. Absolute LODs 24 and 28 pg for Ga and T1 respectively. Determination of Al Cd Cu Fe Pb and Ti. Influence of silica gel on background assessed. Advantages of method were simplicity and speed. Method of analyte additions used. Use of NH4H,P04-Mg(N03) as chemical modifier for Cr Cu Fe and Mn. Use of the slurry technique reduced blanks and allowed an 8-100-fold improvement in LODs for elements such as Na K Mg and Zn. See also 95 4594. Speciation of CrIll CAI preconcentration with a factor of 50-100-fold onto ion- exchange resins and slurried resin taken for direct analysis. LODs 0.01-1.0 pg 1-'. Samples 10-500 mg plus 2 ml EtOH+ 1 ml 30% v/v H,O + 0.25 ml conc.HNO mixed and diluted to 25 ml. Ultrasonic mixing and stirring 20 pl aliquots taken for ETAAS measurements. Chemical modifier claimed to be H20 as reduced carbonaceous residue during pre- treatment. Results agreed with wet digestion procedures. See also refs. 951900 951982 9512528 9513840 961103. Samples 50-200 mg mixed with 25 ml of 20% v/v EtOH containing 0.5% v/v HN03 mixed by ultrasonic agitation and stirring. 20 pl aliquots taken for analysis by 'fast furnace' ETAAS programs. Results for RMs (2) and 15 samples compared with independent AAS method of analysis were in good agreement. See also refs. 951900 9512528 9513840 9514597 961103. Aliquots of 10 ~1 slurry taken. See also refs. 951900 951982 9513840 9514597 9512528.Sample 0.05-0.5 g suspended in 10 ml Samples 10-500 mg + 5 ml EtOH + 2 ml 30% v/v H20z + 25 mg NH4H,P04 diluted to 25 ml with water. Mixed by ultrasonic agitation and stirring. 20 p1 aliquots taken for analysis by 'fast furnace' ETAAS programs with platform atomization. Results in good agreement with acid digestion method and slurry sampling based on a partial ashing of the samples. See also refs. 951900 951982 95/956 9513840 9514597 961103. Slurries prepared in EtOH and H,Oz mixture NH4H2P0 as chemical modifier for Pb. Both Pd and ammoniacal solution of CU" used for Cd. LODs 0.5 and 8 ng g-' for Cd and Pb respectively RSDs in the range 2.5-14%. See also refs. 95/900 951982 9512528 9513840 9514597 961103. ground residue with 2 ml EtOH + 1 ml 30% H,Oz + 25 mg NH4H2P04 diluted to 25 ml.Ultrasonic agitation and stirring used to produce suspension. 20 pl aliquots taken with 'fast furnace' ETAAS programs. LODs 2 10 8,6 and 75 ng g-' for Pb Zn Fe Cu and Cr respectively. See also refs. 951956 951982 9512528 9513840 9514597 961103. Slurry prepared by mixing 5-200 mg of Reference 9512684 9 514 506 9 513 501 9513660 9513767 95lC2898 9514597 951982 961103 9512528 951956 95/900 294R Journal of Analytical Atomic Spectrometry August 1996 Vol. 11Table 2 (continued) Type of sampling Slurry seafood freeze dried plants dried and ground Element Co Ni Slurry 0.5% m/v prepared by bottle and zirconia bead method particle size < 250 pm Slurry prepared by bottle and zirconia bead method and magnetic agitation As As Slurry prepared by bottle and zirconia bead method particle size <20 pm Slurry 10% m/v automated preparation via FI Slurry automatic preparation with an ultrasonic probe Slurry automatic preparation with an ultrasonic probe Slurry manual preparation by mixing Slurry particle size < 50 pm Solid Solid Solid Solid Solid A1 Fe Ni Various (6) Se As Cu Pb Cu Pb Fe Matrix Food samples Lyophilized mussel Marine sediments Marine sediments Yoghurt suspensions Edible olis.fats Mineral waters biological Biological Tuna fish spinach cabbage and sediments RMs Hair Geological Silicon nitride and silicon carbide Pasta food Sample treatmentlcomments Samples 50-150 mg plus 5 ml of a solution containing 0.25% v/v Triton X-100 10% v/v H20 and 1% v/v HNO mixed by ultrasonic agitation and stirring.20 yl aliquots taken for ETAAS measurements. LODs 36 and 42 ng g-' for Co and Ni respectively. See also refs. 951900 951956 951982 9512528 9514597 961103. Chemical modifier of Pd-Mg(NO,) used 20 1 slurry aliquots taken. LOD 1.3 pg g-rin mussel. Different chemical modifiers examined LaC1,-HNO mixture and Pd-Mg( mixture for total As. LaC1,-HNO produced double peaks and not recommended. LOD with Pd-Mg(NO,) was 44 pg kg-'. Use of different thickening agents Triton X-100 Viscalex HV30 and gl cerol were studied. Palladium 15 mg I-[ used as chemical modifier found to stabilize Hg up to 400°C for aqueous solutions and 200 "C for slurries. Atomization temperature 1450°C. LOD 70 pg kg-l method of analyte additions used. Use of FI to avoid need for sample weighing dilution and homogenization and addition of aqueous standards automatically.particulate Fe and Ni with sample tray heated to 60 "C. Method gave better precision than manual method. Preconcentration of analyte elements onto Ployorgs YIIM sorbent introduction of sorbent slurry 20mg in 1 ml water to ETAAS by automated agitation. LODs of the order of 0.01-0.4 pg 1 - l . Sample 0.6 plus 5 ml glycerin + 2.5 ml 1 mg ml- Cu 1.5 ml Mg(NO,),+0.5 ml of 50% v/v HNO finally diluted to 25 ml. 25 pl aliquots taken for ETAAS measurements. Method suitable for determination of >0.1 pg g-' Se. Slurry sampling and measurements found to be simpler and faster than digestion method. Samples ground 50-100 mg taken in 1 ml of 6% v/v HC1-t 2% v/v HNO + 4% v/v H202 plus 0.1 ml of 0.5% v/v Triton X-100 and slurries prepared by manual shaking 10-20 pl aliquots taken for ETAAS measurements. Maximum pre- treatment temperature of 220 "C with atomization at 900°C with 1 s ramp.Found better not to use a Pd chemical modifier. minimum sample size required for an RM. See also ref. 951952 9512839. Axial distribution of As in individual 5 mm segments of hair. Use of mixed Pd-Mg(NO,) chemical modifier and aqueous calibration standards. Results agreed well with those obtained by NAA. Application of continuous laser vaporization coupled with ETAAS. Use of a cup-type graphite atomizer. Sample size > 20 ym mixed with graphite powder 0.5-3 mg taken for analysis. Calibration based on aqueous solutions. Double peaks seen for Cu first assigned to Cu vaporizing from Si second due to Cu vaporizing from SIC.Results agreed well with ETAAS measurements after digestion. wavelengths atomization conditions and sample mass were optimized for the direct determination of Fe. A sample size of at least 2 mg was required and inserted into the tube through an enlarged sampling hole. Results compared well with a total digestion FAAS method. Fully automated determination of k Use of solid sampling for evaluating the Fundamental parameters such as Reference 9513840 9513371 9511500 9511838 95/55 1 9512222 9512824 9514373 951948 961290 9513659 951100 951892 9513096 9512 1 15 Journal of Analytical Atomic Spectrometry August 1996 Vol. 11 295RTable 2 (continued) Type of sampling Element Solid Ga Solid Solid Solid Solid Solid Solid samples preconcentrated onto sorbent Solid Solid Solid Solid Solid Solid Solid and slurry manual preparation Pb Pb Pb Se Matnx - Biological Muscle tissue Copper alloy Biological V Lignite Various (6) Natural waters Various Various (3) Various (4) Various (4) Various Various Various (6) Molybdenum- based materials RMs Biological RMs metals and biologicals Biological RMs and industry dust Powdered Mo metal and molybd'enum silicide Biological Rm:j Sample treatmentjcomments Preconcentration onto an anion-exchange resin with Tiron.Single bead of resin inserted into cup atomizer. LOD 1.5 ng ml-'. Tolerance to sulfate nitrate and chloride interferences was enhanced lo- 100- and 1000-fold respectively. Use of miniature cup atomizer method of analyte additions. Samples pre-ashed at 500°C for 30 min.See also ref. 9511131. bovine muscle tissue can be caused by cysts produced by Cysticercus bovis the larval stage of Tuenia suginutu. Solid sampling ETAAS was well suited to the determination of endogenously contaminated muscle which would be missed by sample decomposition methods. Spatial distribution of Pb in copper alloy determined by reduced pressure ETAAS. Three Pb peaks observed at 0.1 Torr (1 Torr = 133.322 Pa) first peak assigned to release from surface second peak from grains and third from Pb in the bulk material. chemical modifier of Pd and Cu in 0.3% v/v Triton-X-100 was applied. Calibration against aqueous solutions. Method was applied to the determination of Se in heart muscle. Lignite mixed with graphite powder and sampled heated at 600°C to remove A1 interference atomization at 2660 "C for 30 s.50-100 mg of sobent. Sorbent filtered off and mixed with graphite powder (1 + 1) and placed in a ring-chamber tube for ETAAS measurements. LODs 2-5 ng 1-' for Ag and Cd 10-30 ng 1-' for Bi In Pb and T1. Study of separation of analyte and matrix from metallic Mo powder using radio tracers under different ETA AS conditions. See also ref. 9514594. Calibration procedure by extrapolation to zero matrix. Peak absorbance signal divided by analyte or sample mass (for standard and sample respectively) plotted as function of analyte or sample mass and extraploated to zero mass. This produced a signal not influenced by deviations from linearity and free from matrix effects. See also ref. 9513659 9512839.Use of pre-ashing stage as a preconcentration step in solid sampling. See also ref. 9511079. purposes. L'vov platforms employed and peak height measurements. Calibration curves established by an iterative procedure. See also refs. 9513659 951952. participants for the determination of Cd Cu Hg and Pb. Most results satisfactory except for the determination of Cu in bovine liver and Cd and Cu in industry dust. Powdered samples analysed using either a cup atomizer (CRA-90) or boatjplatform (HGA-72). LODs improved by 1-2 orders of magnitude compared with solution methods of analysis for Cu K Mg Mn Na and Zn. See also 9513767 9512119. Use of older AA instrumentation (1980s) with D2 background correction. Solid sampling produced poor %RSDs (13-15%) but good accuracy and precision were obtained with slurry sampling (3-13%).Non-homogeneous Pb distributions in Use of cup-in-tube method. A mixed Water samples 100 ml treated with Use of solid sampling for certification Results of a collaborative study with 11 Reference 9513381 9 51 1079 9511 125 9511898 951916 951957 9513375 95/2119 951952 9511131 9 5/28 3 9 9513858 9514594 9511039 296 R Journal of Analytical Atomic Spectrometry August 1996 Vol. 1 1Table 2 (continued) Type of sampling Element Matrix Sample treatment/comments Reference Solid sample ground. Cd Biological sediments Sample mixed with chemical modifier of 95/3863 0.5% Ca(NO,) 0.5% La(NO,) 0.5% NH,NO and 50% EtOH then dried and applied onto the graphite platform with a laboratory-made device. Air ashing applied at > 500 "C for 40 s to remove matrix.Method of analyte additions applied and recoveries of 93-105% found. determination of ionic alkyllead compounds Erber et al. (95/3790) applied FI coupled with 'in atomizer trapping' and a THGA atomizer. The LOD was found to be approximately 7 ng 1-' but can be lowered to < 1 ng 1-l if larger sample volumes are taken. A number of workers applied HG and 'in-atomizer trapping' for the determination of various hydride forming elements in a variety of matrices. For the determination of ultra-trace amounts of Sn Tao and Fang (95/4452) used an FI-HG technique coupled with an on-line ion-exchange separation. The sample was prepared in 2 moll- ' HCl before being passed through a micro-column 3cmx3 mm id packed with a strongly basic anion exchanger.The Sn was retained as its chlorostannate complex and eluted with water into the HG system. The graphite atomizer was coated with palladium and the trapping took place at 300 "C. Up to 30 samples h-' could be processed with an LOD (3s) of 0.01 pg 1-' using a 10.7 ml sample volume. Walcerz et al. (95/3846) determined As and Sb by continuous flow HG and 'in-atomizer trapping' on a palladium-coated graphite atomizer. Trapping took place at 200°C with an argon flow rate of 50ml min-' followed by atomization at 2100 "C for Sb and 2400 "C for As. The method was applied to the determination of As and Sb in a rye grass RM and results agreed well with the certified values. Hanna et al. (96/78) used a commercial system and software to determine Se in microwave digested nutritional supplement formulas.With a 500 pl sample volume the LOD (3s) was 0.6 pg 1-' and sensitivity was increased by over an order of magnitude compared with conventional electrothermal atomiz- ation. Peng and Yao (95/3382 95/3637) eliminated the inter- ferences from nickel and iron in the determination of Se (95/3382) and Te (95/3637) in high-temperature nickel-base alloys by using 'in-atomizer trapping' with a palladium trap- ping reagent in the graphite atomizer. The HG solution was 10% m/v potassium borohydride in 6% m/v sodium hydroxide. Sundin et al. (95/C762,95/C2907) considered the determination of Se in urine. A hydrobromic acid-bromine reagent was used to decompose the organoselenium species in urine and quanti- tative recoveries of Se" SeV' selenopurine selenomethionine and selenocystine were obtained from the urine matrix.The method has the advantage of avoiding the need for perchloric acid but further experimental details were not given in the abstract. Ni et al. (95/C1994) examined the use of silver-coated graphite atomizers for the 'in-atomizer trapping' of Se. The advantage of a silver trapping reagent is that a lower atomization temperature is required 1800 "C compared with >2000 "C when a palladium trapping reagent is used. However the main problem with using elements such as silver or palladium is that after every atomization cycle they are vaporized from the atomizer and the coating has to be renewed. Those workers investigating the use of iridium find that several 100 firings can be achieved before the effectiveness of the iridium coating decreases and it would seem that this approach has much to recommend it.The late Klaus Dittrich and co-workers (95/2223) compared the determination of Hg with furnace atomic non-thermal excitation spectrometry (FANES) either by direct solution injection or by cold vapour generation and trapping within Journal of the FANES atomizer. A coating of iridium was found to form a permanent coating and with this technique an LOD of 0.9 ng 1-' was obtained in comparison with 22ng I-' for direct injection. By reducing the concentration of the sodium borohy- dride reagent to 0.002% m/v it was possible to determine Hg in the presence of high excess concentrations of other hydride forming elements without any depression of the Hg emission intensity.Chinese workers (95/966) used a gold-coated graphite atomizer for the 'in-atomizer trapping' determination of Hg in human hair urine soil and peach leaves though the method of analyte additions was used indicating that the cold vapour generation and trapping procedure was not free of interferences. For the more complex procedure of on-line preconcentration and elution directly into a graphite atomizer Fang and Tao (96/C204) reviewed the recent developments in F I on-line preconcentration systems. They discussed the use of both solid- phase extraction preconcentration and liquid-liquid extraction. To date liquid-liquid extraction is not as popular as solid- phase extraction but these workers illustrated that with modern commercially available equipment and correct design of solvent delivery systems improved phase separators and manifold designs there is no reason why liquid-liquid extrac- tion coupled with ETAAS should not become more widely used.This reviewer (ILS) has his doubts as solid-phase extrac- tion systems are inherently simpler and easier to manipulate. Dong and Fang (95/1188) determined Cd in whole blood by an on-line preconcentration procedure with dithizone copre- cipitation. The Cd was coprecipitated with dithizone in a weakly acidic aqueous solution and the precipitate collected within a knotted reactor and micro-membrane filter. The precipitate was dissolved in 60 pl of IBMK and transferred into the graphite atomizer. An enrichment factor of 11 was obtained with a 20s sampling time at a sample flow rate of 2 ml min-'.The LOD (3s) was 0.003 pg 1-' in the sample digest. From the same group led by Professor Fang Chen et al. (95/3728) used a similar procedure to determine Mo in water human hair and high-purity reagents. The Mo was coprecipitated from dilute HC1 solution with iron@) pyrrolidin-1-yldithioformate. The LOD (3s) was 0.04 pg 1-' for a 30 s loading time. Burguera et al. (95/2390 95/2741 95/3516) described a number of automated on-line procedures coupling FI with ETAAS. In one publication (95/2390) blood was sampled directly from the patients arm via a time-based solenoid injected into the FI stream passed through a microwave oven and then to a solenoid injector where 20 pl of the digested sample were pipetted onto a platform in a graphite atomizer.The claimed advantage of this in vivo sampling and on-line digestion procedure is lower sample handling which reduces contamination and complete automation. This reviewer (ILS) wonders whether any patient would wish to remain linked to such a system given the amount of sample required 2 ml and the length of time the whole procedure takes approximately 5min per measurement cycle. These same workers turned their attention to an inanimate sample namely soap (95/2741) and applied a similar experimental approach to determine titanium dioxide in soap. An automated on-line procedure for the dissolution and microwave digestion of the soap allowed Analytical Atomic Spectrometry August 1996 Vol. 11 297 Rthe rapid on-line determination of Ti02.The values obtained with this procedure agreed well with those from more conven- tional techniques. In the final publication from this group (95/3516) the sulfur anion interferences on the determination of Mn in geothermal fluids were eliminated by an on-line FI column separation procedure. The interfering species were retained on a micro-column filled with 25 mg of Dowex 1-X8 resin and the sample eluate; free of interfering sulfur species stored in a sampling loop and then pumped to a solenoid sampling device where 20 pl of this solution were dispensed into the graphite atomizer. The LOD (3s) was 0.2 pg 1-l and results agreed well with those obtained by batch chelation and extraction procedures. For the determination of Mo in sea-water Yin and Liu (95/C1992) applied on automated FI on-line preconcentration procedure coupled with ETAAS.The Mo was extracted as the dithiocarbamate complex and sorbed onto a cl8 column which was then extracted into a small volume of ethanol and delivered into the atomizer by a low flow of air in what appears to be similar to the procedure originally proposed by Welz et al. (see J. Anal. At. Spectrom. 1994 9 213R). These workers also claimed that if a small internal gas flow of 50 ml min-' was used during the atomization step then the memory effects from Mo can be eliminated. While this is not new it should be pointed out that the internal gas flow during atomization will also lower the magnitude of the signals obtained. For the determination of Cd and Pb at theng 1-' level Abollino et al.(95/C4228) applied on-line FI preconcen- tration on an anion-exchange micro-column after com- plexation with Tiron. The metals were eluted from the column with 0.1 mol 1-l HC1. Interference studies showed that the presence of a saline matrix non-ionic surfactants and an excess of calcium and magnesium did not affect the analyte recovery. The method was applied to the determination of Cd and Pb in tap water and lake water from Antarctica. Workers from the University of Lund Sweden and the University of Botswana (95/C2020 95/C2018) used 150 p1 micro-columns containing quinolin-8-01 immobilized on controlled pore glass as field samplers. These field samplers were then connected to an on-line FI system coupled with ETAAS to extract the precon- centrated analytes with 1 rnol I-' HNO,.Although in the procedure described the eluate is collected into an autosampler cup there is no reason why this could not be automated. The advantages of using these micro-columns as field samplers is that the sampling can be done with minimum contamination and the collected metals are protected during transport. It is easier to transport small tubes than the corresponding number of water bottles. Rayson and Baker (95/C2999) argued that the current procedures for preconcentrating analytes prior to introduction into the atomizer such as column preconcen- tration followed by elution into a much reduced volume or adsorption of the analyte onto a particulate material a slurry of which can then be introduced directly into the atomizer involve a significant increase in the complexity of sample introduction systems for ETAAS.These workers proposed the use of acid-insoluble humic acid as the analyte binding sub- strate and the subsequent introduction of a basic metal-bound humic acid solution into the electrothermal atomizer. While this approach appears to be simple it would presumably be a manually operated batch system and the advantages of the on-line column preconcentration procedures are that they can be automated with equipment and software that is currently commercially available. Bendicho (95/3459) evaluated an automated thermospray interface for coupling HPLC with ETAAS for the determination of As species by ion-pair chromatography on a DeltaPak c18 column with water-methanol buffered with 2 mmol 1-l ammonium citrate as the mobile phase.The thermospray interface consisted of a fused silica capillary 0.1 mm id inserted into a stainless-steel tube 0.5 mm id. The effluent from the HPLC column was introduced through the thermospray vapo- rizer and injected for 2 s into the graphite atomizer. For As"' AsV and dimethylarsinate the detection limits were 5 8 and 15 pg 1-' respectively. Sadly there is no further information about the column flow rate. The thermospray interface was operated in real-time mode as the time between the measure- ments was considerably reduced as the chromatographic eluent was dried on impact with the curved pre-heated platform at 150 "C. A total thermal program time of 40 s was feasible 15 s for the atomizer cycle plus 25 s for cool down.Astruc et al. (95/1882) and Potin-Gautier et al. (95/1219) from the same research group have determined butyltin compounds and selenoamino acids respectively by on-line coupled HPLC with ETAAS. Although mention is made of a laboratory-made interface no details were given in the abstract. 1.2.2.4. Electrodeposition. A small amount of work on elec- trodeposition continues. Matousek and Powell (95/4744) have extended the in situ procedures that they have described earlier (see J. Anal. At. Spectrom. 1995 10 199R). Previous work with a 0.5 mol 1-' sodium chloride medium using a platinum anode attached to the PTFE sampling capillary of the auto- sampler showed that although quantitative separation of ana- lyte was achievable in short times (120 s) from 50 p1 sample volumes by the use of high deposition potentials the atomizer lifetime was severely limited by the exfoliation of the pyrolytic graphite coating.A maximum of 50-70 firings could be achieved. The exfoliation was thought to be due to the penetration of Na' or H+ into the pyrolytic graphite coating followed by reduction and evolution of H,. These workers concluded that the surface of the atomizer required a more protective coating prior to electrodeposi tion. The current work (95/4744) described the use of electrodeposited palladium as a renewable atomizer coating and as a chemical modifier. The sequence of steps to achieve a complete measurement appear to be as follows. Clean the tube by adding 30-60 p1 of 3% v/v nitric acid electrolyse at 3.2 V (30-35 mA) for 60 s remove residual solution and any released analyte is then removed by an atomization cycle.Pre-deposition of palladium 40-50 pl of 10 pg ml-' of palladium in 0.5% v/v nitric acid is pipetted into the tube and electrolysed for 20 s at 4-5 V (25-50 mA) and then the solution is withdrawn. This coats the tube surface with approximately 0.25 pg of palladium. The tube is 'rinsed' by pipetting 50-60 PI of 3% v/v nitric acid into the furnace and allowing it to remain in the atomizer for 5 s before being removed by the autosampler. Next the analyte solution 25-30 p1 (in this series of experiments either 0.5 mol 1-' of sodium chloride or 1% v/v nitric acid containing 20 pg 1-' of Pb) was electrolysed at 4-5 V (20-40 mA) for 60 s and then the residual analyte solution removed.For washing the deposited analyte and dilution of residual salts 50-60 p1 of water were added and then withdrawn if necessary to release deposit entrained sodium chloride then 40-50 pl of nitric acid were electrolysed and withdrawn followed by another 'wash- ing' step and finally an atomization cycle was performed. While there have been previous publications describing this procedure this is the first that provides a detailed step-by-step listing of the procedures so that one can follow exactly what the workers proposed. It is fair to say that this is not the easiest nor quickest procedure that has been described to determine Pb in 0.5 mol 1 - l sodium chloride solution and the degree of complexity both in parameters that have to be defined and coordination and sequencing of the autosampler do not appear trivial.Chinese workers (96/154) described an electrochemical preconcentration procedure for the determi- nation of Pb in polluted soil. The Pb is deposited on a coiled tungsten wire modified with a thin film of Nafion which is then placed inside a graphite electrothermal atomizer. 298R Journal of Analytical Atomic Spectrometry August 1996 Vol. 11The work of Beinrohr and co-workers (95/C4232) continues in investigating the use of an electrochemical flow-through cell for preconcentration prior to elution into a graphite atomizer. The current work (95/C4232) described the application of such a device to the determination of Hg. Unlike the previous work on Pt (see J. Anal. At. Spectrom. 1995 10 199R) whereby the cell was then placed into the atomizer here the Hg is precon- centrated onto crushed glassy carbon with a particle size of 10-100 pm with a negative potential.After preconcentration the trapped Hg is eluted at a positive potential with a small volume of dilute nitric acid and delivered directly into the graphite atomizer. 1.2.3. Fundamental processes The atomization of Ba was investigated by Monteiro and Curtius (95/1675) from the wall of electrographite pyrolytic graphite coated electrographite and tungsten-coated graphite tubes using either argon or nitrogen as the sheath gas. The sensitivity appearance times pyrolysis and atomization tem- perature curves tube lifetimes and SEMs were obtained for all of these combinations. The best characteristic mass and longest lifetime were obtained in a pyrolytic graphite coated tube with argon as the purge gas.Nitrogen was found to increase the appearance time and decrease the sensitivity owing to the formation of gaseous barium monocyanide. The appearance temperature approximately 2200 "C suggests that Ba is atom- ized through the reduction of Ba by carbon. Complete atomiz- ation was not obtained with any of the conditions as evidenced by the absence of a plateau on the atomization curve up to 2650 "C. The SEM micrographs explained the poor perform- ance of the tungsten coated tubes the tungsten coverage was discontinuous and melting of the tungsten phase decreased the coverage and led to destruction of the pyrolytic graphite coating. The analysis of real samples was only possible using the method of analyte additions.Fonseca et al. (95/1017) investigated the mechanisms of Cr vaporization from a graphite platform with an ETAAS-MS instrument. Aqueous samples of chromium nitrate were deposited on the graphite platform dried under N2 and atomized at 2740K and the species evolved with respect to temperature monitored. Free metal atoms were monitored by AA in uacuo which enabled comparisons to be made with the MS data. Absorption profiles were obtained by conventional AAS measurements under atmospheric pressure conditions. The data indicated that two dominant mechanisms contribute to the analytical signal the decomposition of Cr3C produced by carbon reduction of Cr203 and the thermal desorption of adsorbed Cr. Gaseous Cr oxides were identified at approxi- mately 1500K by AA-MS these species probably undergo re-deposition at active sites in the tube under atmospheric pressure and contribute to the formation of adsorbed Cr on the surface.Gaseous Cr carbides were not detected by MS. Thermal pre-treatment of deposited chromium nitrates at 1700 K which led to a shift in the position of the absorption maxima with time was seen as due to a gradual reduction in the contribution of a mechanism associated with the decomposition of Cr203. The atomization eficiencies for Cr Bi Ga Mn and Pb in a longitudinally heated graphite atomizer (HGA-400) operated under STPF conditions were calculated by Yang and Ni (95/3519) with and without the use of palladium as a chemical modifier. These workers found that the proportionality factor p which relates the number of absorbing atoms to the measured absorbance was found to be almost constant for these elements over the temperature range studied.This indicated that the assumptions made in the calculations for the rate constants for atom production and atom removal were acceptable. In some cases the use of palladium produced higher atomization efficiencies. For Bi and Pb similar experimental and calculated atomization efficiencies were found and in the presence of palladium the atomization efficiencies were temperature inde- pendent. For Cr Ga and Mn the experimental atomization efficiencies were lower than the calculated atomization efficiencies and whether or not palladium was used the atomiz- ation efficiencies were temperature dependent.In general the atomization efficiencies agreed with those found in previous studies though it is difficult to draw relevant conclusions from this publication. Other Chinese workers (96/363) derived a formula for evaluating atomization efficiencies based on the relationship between the AA signal and the mass of analyte. Majidi and Xu (95/C653) investigated the atomization of Ga In and T1 with time and wavelength resolved molecular absorption spectroscopy. A laser-induced plasma was used as a continuum source to obtain the molecular absorption of the transient species in the graphite atomizer. During the atomiz- ation of Ga Ga,O Ga2 and GaCl were identified as intermedi- ate species. Both atomizer wall reduction and thermal decomposition of these species were suggested as pathways leading to the formation of free Ga atoms.Similar pathways were suggested for the atomization of In. The indium dimer In was found to play a major role in free In atom formation. For the atomization of T1 there was difficulty in identifying the intermediates although the evidence suggests rejecting the formation of thallium carbides. Chinese workers (95/945) examined the vaporization of Ga In and Tl from a molyb- denum-coated graphite tube. The results indicated that these elements formed stable sintered materials such as Ga,O InO and T1,O with MOO,. Japanese workers (95/3772) also exam- ined the atomization mechanism of Ga. These workers found that the electrothermal atomizer signal for Ga from a pyrolytic graphite coated electrographite tube showed an unresolved double peak.The first peak was interpreted as being due to the thermal decomposition of Ga20 produced by carbon catalysed decomposition of Ga,O and the second peak owing to a surface reduction of Ga,O to gaseous Ga atoms. In the presence of an organic matrix the first signal was enhanced owing to the occlusion of Ga20 in the porous residue of the pyrolysed matrix. These same workers (95/548) investigated the atomization of Au from organic matrices. The atomic signal for Au was altered by the pyrolysis of a volatile organic matrix (methanol propan-1-01 and pentane- 1,5-diol) and also a non-volatile matrix (ascorbic acid glucose and sucrose) with a low-temperature shift of the signal and development of a shoulder or second peak on the tail of the original peak.The workers claimed that pyrolysis of the organic matrix produces two types of carbon residue active carbon and a thermally stable carbon residue. The early shift ocurrs as a result of the formation of smaller microdroplets of Au by adsorption of the analyte on the active carbon. The latter shift and peak is due to the formation of larger sized microdroplets by admission diffusion or transport of the Au into the inside of the thermally stable carbon residue. These conclusions were supported by kinetic investigations. Aller (95/1814) considered the vaporiz- ation mechanism for Au with and without the presence of vanadyl chloride. Both wall and platform atomization show similar results for the orders of reaction and the absorbance profiles were found to be sensitive to the mass of vanadium present showing larger integrated absorbances when the vanadium mass increased.Various workers from Chakrabarti's group (95/C608 95/C654 95/C656 95/4739) used ETV-ICP-MS to investigate the atomization mechanisms of U (95/C608 95/4739) W (95/C654) and B (95/C656). Electrothermal atomization stud- ies from a graphite surface indicated that U atoms were formed at temperatures < 2400 "C. Using ETV-ICP-MS an appear- ance temperature of 1100 "C was obtained indicating that some U vaporizes as uranium oxide. Although uranium car- Journal of Analytical Atomic Spectrometry August 1996 Vol. 11 299Rbides form at temperatures above 2000 "C the ETV-ICP-MS studies showed that they do not vaporize until 2600 "C.In the temperature range between 2200 and 2600 "C U atoms within a graphite atomizer are most probably formed by thermal dissociation of uranium oxide whereas at higher temperatures U atoms are formed via thermal dissociation of uranium carbide. Chemical modifiers such as 0.2% v/v HF and 0.1% v/v CHF in the argon gas were ineffective in preventing the formation of uranium carbide at 2700 "C. The origin of the U signal suppression by sodium chloride was investigated. At temperatures above 2000 "C signal suppression could be caused by the accelerated rate of formation of carbide species while at temperatures below 2000 "C the presence of sodium chloride could cause intercalation of the U in the graphite layers resulting in partial retention of U during the vaporization step.Vaporizing the U from a tungsten surface using tungsten foil inserted into the graphite tube prevented the formation of uranium carbide and eliminated the ETV-ICP-MS signal suppression caused by the sodium chloride matrix. The vapor- ization of W (95/C654) by ETV-ICP-MS in the range from 800 to 2700 "C showed that two distinct processes were taking place. An early peak appeared at a temperature as low as 850 "C and was attributed to tungsten oxide vaporization. At temperatures above 2500 "C a second peak appeared and was considered to be due to the vaporization of tungsten carbide from the surface. Results showed that sodium chloride and sodium fluoride chemical modifiers were ineffective in pre- venting the formation of tungsten carbide. In investigating the atomization mechanism of B (95/C656) the poor sensitivity of ETAAS for this element was shown to be due to the formation of molecular boron oxide species prior to atomization.Both ETAAS and ETV-ICP-MS were used to investigate the tem- poral atomization and vaporization of B from a graphite surface. Additional information regarding the spatial and tem- poral distribution of the atomic and molecular species during atomization was provided by digital imaging using a CCD camera. Kantor et al. (95/2127) applied a combined electrothermal vaporization and flame atomization technique to investigate the vaporization mechanisms of calcium nitrate calcium chlor- ide gallium nitrate gallium chloride and gallium metal. Gallium was found to evaporate from a nitric acid solution residue as Ga,O(g) at temperatures of between 900 and 1000 "C when introduced as analyte and as both Ga,O(g) and Ga(g) between 1100 and 2300 "C when present as a matrix.Calcium as the analyte vaporizes rapidly above 1520 "C producing atomic vapour and permitting partial separation from Ga during the pyrolysis stage. The advantage of this approach appears to be that it is quicker to generate volatiliz- ation signals with this technique than generating conventional pyrolysis curves with ETAAS and it can produce adequate information about permissible pyrolysis temperatures for sev- eral substances. However these workers state that it is still important to determine the ETAAS pyrolysis curve at least in the narrow temperature range which is judged to be critical based on the volatilization signal.Redfield et al. (95/C609) investigated the atomization mechan- isms of Pb in a two-step atomizer based on Frech's design whereby the tube atomizer can be heated to a constant temperature before the sample is heated. All gaseous species are then vaporized in the hot tube and molecular species are thermally decomposed. The relative amounts of Pb released as the oxide were determined with and without the presence of aluminium or magnesium present as 1+1 mixtures on a mass basis with Pb. The presence of magnesium results in the near total loss of the Pb signal observed from atomization with no magnesium present though sadly there were no further discussions given in the abstract as to the conclusions to be drawn from these observations.Two basic mechanisms have been proposed to account for the appearance of metal oxide species after the drying of metal nitrates on the graphite surface at temperatures well below the expected vaporization temperature. The first is a physical expulsion mechanism that involves the direct production of solid metal oxide components and the second a coherent gasification mechanism that involves a gas phase metal oxide intermediate. Holcombe et al. (95/C650) carried out gas phase MS studies in an effort to determine the origin of such signals from nitrates of Ag Cd Cu and Pb. Their results showed many differences in the decomposition trends for the different metal nitrates and they concluded that one mechanism may not accurately describe the decomposition of all metal nitrates.Deng and co-workers (95/1024,95/3508,95/3792) continued to examine the atomization mechanisms of a variety of elements from graphite probe surfaces using a variety of surface tech- niques such as SEM X-ray diffraction X-ray photoelectron spectrometry and Auger electron spectroscopy. As part of an investigation into the role of blister formation on atomization mechanisms these techniques were applied to examine the formation of blisters on graphite or tantalum probes during atomization of large amounts of lanthanum nitrate (95/1024). The formation of such blisters has been described previously by other workers and the interpretation was that these were blisters of carbon. These workers found that blisters formed on both the graphite and tantalum probe surfaces.They concluded that carbon does not play a role in the formation of these blisters and that they are due to lanthanum oxide absorbing water vapour at 1900K and hence enlarging its volume. The ability to absorb water vapour decreases rapidly with increasing temperature and the amount of lanthanum hydroxide formed is too small to account for the blisters at 2300 K. Blisters were not observed on the graphite probe at temperatures higher than 2300 K and this was considered to be due to the formation of lanthanum carbides. Lattice defects caused by interaction between the lanthanum and tantalum species on the tantalum probe were observed and it was thought that these were responsible for the formation of blisters on the tantalum probe. For the atomization of Eu(NO,) from a graphite probe (95/3508) it was concluded that the nitrate was first decomposed to Eu203 then to EuO before vaporiz- ation into the gaseous phase and thermal decomposition during atomization.Europium carbide was formed at 1600 and 1920K. A similar mechanism was proposed for the atomization of Sm (95/3792). In considering the efects of oxygen within a graphite atomizer the group at the University of Karlsruhe have extended their previous studies in which uncoated electrographite tubes were used to now include pyrolytic graphite coated electrographite tubes and pyrolytic graphite platforms. Muller-Vogt et al. (95/4695) discussed the chemical processes occurring within these atomizers when the oxide forming elements Bi Pb and Ti were determined.Pyrolysis curves were established for all the different atomizer types with flowing argon and gas-stop conditions during atomization. These workers found that the reactions occurring for T1 and Bi are reduction reactions via volatile suboxides to the elements whereas Pb is directly reduced to the metal. The volatile suboxides are responsible for material losses during the pyrolysis step. Under gas-stop conditions the losses can be reduced by further reduction of the suboxides by additional collisions with the graphite surface. Separate atomization from the atomizer wall and from the platform after pyrolysis gave an indication of the location and kinetics of the different reaction steps. The oxygen conditioning of the uncoated atomizers and the uncoated platforms resulted in the thermal stabilization of the elements.This was considcrcd to be due to the intercalation of the metals into the graphite layers. This does not occur on the pyrolytic graphite coated surfaces owing to the absence of edge sites hence there is 300R Journal of Analytical Atomic Spectrometry August 1996 Vol. 11no direct reduction of T1,0 to TI and no stabilization by intercalation of the T1 atoms. The same behaviour was observed for Bi. For Pb the PbO is reduced directly to Pb. This work appears to agree with the findings of Eloi et al. (Anal. Chem. 1995 67 335) which was discussed last year (see J. Anal. At. Spectrom. 1995 10 199R). The work concerning the mechanisms that control the atomiz- ation of Si from solid gold samples and discussed previously (see J.Anal. At. Spectrom. 1994 9 213R) now has been extended and published by Hinds and co-workers (95/537 95/3727 95/C652). Initial results (95/537) showed that Si is released only after the gold matrix melts and migration of the Si to the surface is assisted by the formation of convection cells in the melt. This convection is driven by thermal gradients in the sample. Diffusion alone either in the solid or the melt was too slow to account for the relatively short time for total Si release before a significant portion of the gold matrix was depleted. They showed that the liquid lines of the binary phase diagrams associated with a gold solvent can be used to predict the existence of analyte carry-over in the gold. Consequently these diagrams offer a means of identifying a priori candiate elements that could be determined by solid sampling ETAAS in gold via aqueous solution calibration.Brown et al. (95/3727) investigated the mechanisms of Si atomization from aqueous solutions and solid gold samples both in vacuo and in real time by MS to explain the correlations between free Si evol- ution from aqueous and from solid gold samples. The results from monitoring the gaseous Si species that evolved from these two types of samples indicated that the atomization mechan- isms differed substantially. Free Si and SiO vaporized from both sample types but SiC,(g) and Si,(g) were unique to the aqueous and to the solid gold samples respectively. These observations implied that atomization occurs via dissociative adsorption processes during heating of the aqueous samples.For the solid gold samples the Si impurities are carried by convection to the surface of the molten gold where they vaporize. These workers speculated that the correlations between free Si formation at atmospheric pressure are because of gas-phase collisions. These collisions can enhance dis- sociation of the Si2(g) that forms only during heating of the solid samples. Torsi et al. (95/1217) applied a model ofpure diflusion and verified it based on experimental data obtained from a labora- tory-made atomization system which allows all analyte atoms from the sample to be present in the optical beam simul- taneously (see J. Anal. At. Spectrom. 1993 8 197R). The validation was obtained through a ratio of experimental par- ameters derived from each absorbance uersus time curve.The elements used as analytes were Cd and Hg which satisfied the models assumptions satisfactorily however data obtained for Pb did not agree with the theory. From measurements at different temperatures the spectroscopic constant K does not change with temperature over a broad range. Diffusion coefficient values in argon calculated from measurements at different temperatures with this atomization system were found to be at variance with accepted values. This was found to be due to a wrong assignment of the length of the atomizer graphite tube Torsi (95/4731) has extended this work to consider the spectroscopic constant K as a more useful parameter to compare different atomizers. This constant in the Beer-Lambert law applied to AA measurements with line sources can be theoretically derived.Hence this value can be compared with the experimental data. In comparing the avail- able literature data for K a large difference was found between those obtained with metal atomizers and the values obtained theoretically when highly volatile elements are considered. An even bigger difference is found between the values for graphite atomizers. Torsi proposed that if this difference is real then a loss mechanism until now undetected in graphite atomizers must be found. In addition the use of the K value can be applied to atomization mechanism studies if the atomization efficiency is very low and this can be checked by assuming that the theoretical values of K are correct then the possibility for errors is greatly increased.Yan and co-workers (95/3392 95/4397 95/4733) have con- tinued to apply the method originally proposed by Yan et al. (see J. Anal. At. Spectrom. 1994 9 213R or Spectrochim. Acta Part B 1993 48 605) to determine simultaneously the kinetic order and activation energy. This method was applied to study the atomization of Au (95/4397) from a pyrolytic graphite surface. The kinetic order and activation energy were found to be 0.31 f0.03 and 293 & 8 kJ mol-l respectively. No analyte mass dependence of the measured kinetic order and activation energy for Au was found over the range 0.25-2ng. It would appear that workers in the same group (95/4733) have modified the Yan et al. method to allow it to be used for calculating these kinetic parameters for atom formation under isothermal conditions. A tungsten wire probe was used to achieve constant temperature atomization conditions.By means of this model the kinetic parameters for the atomization of Bi Ge Mn and Pb at constant temperature were found to be 0.67k0.01 and 302f8 kJ mo1-l for Bi 1.10f0.08 and 109f2 kJ rno1-l for Ge 0.97k0.03 and 372+5 kJ mol-I for Mn and 0.46+0.01 and 159f2 kJ mol-' for Pb respectively. However it was stated that the method is only applicable to elements for which condensation and re-atomization do not contribute signifi- cantly to the shape of the atomic absorption uersus time signal. In those cases where the analyte species vaporizes from the graphite surface followed by condensation on the graphite wall and re-atomization the absorption signal will be a composite of these processes and this will affect the kinetic parameters derived with the method. Yan and Ni (95/3392) have prepared a critical review (in Chinese) of thermodynamic and kinetic studies of atom formation in ETAAS.There are now several alternative methods available for calculating kinetic parameters from ETAAS data and it is always a pleasure when someone else compares the alternatives and saves the writer of this Update from having to do it. Fonesca et al. (95/2215) compared three of the current methods for the determination of these parameters. The method pro- posed by McNally and Holcombe (Anal. Chem. 1987,59,1015) uses Arrhenius-type plots to extract activation energy values and an approximation of the order of release is obtained by alignment of the absorption maxima at increasing analyte concentrations.In the method proposed by Rojas and Olivares (Spectrochim. Acta Part B 1992 47 387) plots are prepared for different orders of release with the correct order yielding a longer linear region from whose slope the activation energy is calculated. The method of Yan et al. (Spectrochim. Acta Part B 1993 48 605) uses a single absorption profile for the calculations activation energy and order of release are obtained from the slope and intercept respectively of the graph pre- pared. All three methods assume linear heating rates constant activation energies and isothermal atomization conditions. The three methods were tested with the same sets of experimental data for Au Cu and Ni obtained with a spatially isothermal atomizer.In general the three methods produced comparable results for the activation energies for the three elements. The simple Arrhenius plots appeared to provide an easier way to estimate the activation energy since they did not require the intensive mathematical treatment involved in the methods proposed by Yan et al. and Rojas and Olivares. In addition the last two methods require much better S/N to produce acceptable plots since the data transformation accentuates any small fluctuations in the original data. With the method of Rojas and Olivares establishing the activation energy and order of release often becomes subjective because of the requirement of selecting a value for the order that produces Journal of Analytical Atomic Spectrometry August 1996 Vol.11 301 Rthe 'most extensive linear range' on data sets that are all non- linear. As different activation energy values are obtained with different orders of release then an error in choosing the correct value for the order also generates an error in determining the activation energy. With the method of McNally and Holcombe the order of release is only a semiquantitative idea of the value while the other two methods produce quantitative results. In addition the methods of Yan et al. and Rojas and Olivares only require a single absorption profile for the calculations while that of McNally and Holcombe requires absorption signals for a range of different concentrations. However it is important to realize that the activation energy and order of release can vary depending upon the analyte mass. Russian workers (96/412) considered the temperature depen- dence of the atomization efJiciency in a graphite electrothermal atomizer.In general the thermodynamic simulation of the processes occurring in a graphite electrothermal atomizer adequately describes the relation between the temperature and the degree of atomization. Chinese workers (95/3630 95/3631) have continued their investigations into the spectral line profiles with Zeeman-efect background correction AAS. The profiles of 37 lines were calculated theoretically and plotted. The overlapping relation- ship of AE AA and Zeeman-effect AA calculated profiles for the 37 lines were established and Doppler and Lorentz broad- ening and the hyperfine structure of the lines were taken into account.These workers also compared the theoretical and experimental values for the Zeeman ratio R and found that the discrepancy between calculation and experiment was less than 10% for 33 lines examined. 1.2.4. Interferences Katskov et al. (95/C655 95/4736) have continued with their studies on the graphite filter atomizer. Although discussed in the last Update under section 1.2.1 Atomizer design and surface modiJcation (see J. Anal. At. Spectrom. 1995 10 199R) in the present Update it is discussed under interferences as this is the area investigated in these recent publications. These workers have examined the theoretical and experimental reasons for the elimination of interferences in the graphite filter atomizer.Observations of stepped or multiple background absorption signals for NaCl NaI KCl KBr KI CaC1 and MgS04 characteristic for each matrix and type of graphite used allowed a hypothesis to be proposed about the interaction of molecules with graphite which display signs of formation of stable stoichiometric compounds. These effects provided differences in the diffusion rates for atomic and molecular vapour that could be used for the development of methods for high-temperature separation of these materials and with the correct design of electrothermal atomizer for a reduction in spectral and chemical interferences. The atomization of A1 is still an important topic as Tang et al. (95/3726) made clear. While ETAAS is acknowledged to be the method of choice for the determination of A1 in biological materials there still appear to be numerous problems with this determination when one considers the wide range of different chemical modifiers and acid media proposed.Tang et al. (95/3726) investigated the effects of several acids such as HNO HCI and H2S04 and tried to resolve some of the confusing and conflicting reports in the literature. The aim of their work was to identify the best chemical modifier and acid medium for the determination of A1 in serum bone and other biological samples by ETAAS. Based on atomization from a solid pyrolytic graphite platform and with Zeeman-effect back- ground correction no interferences were observed for any of the acids investigated. The expected characteristic mass for A1 was unaffected by chemical modifiers such as magnesium nitrate palladium nitrate calcium nitrate and ammonium dihydrogenphosphate. Addition of the calcium magnesium and palladium nitrates produced a sharper absorbance peak and the A1 signal was delayed in time.Either calcium or magnesium nitrate can stabilize A1 during pyrolysis allowing a very high (>17OO"C) pyrolysis temperature to be used. However multiple atomization peaks were seen when mag- nesium nitrate was used. These multiple peaks became more troublesome as the tube aged. Calcium nitrate was rec- ommended by these workers as a better chemical modifier for samples such as bone in which calcium is a large matrix component. A serious interference from chloride salts varied with chloride concentration pH and pyrolysis temperature.This 'suppressive' interference from chloride was overcome by using calcium nitrate as a modifier in 1% v/v nitric acid and a pyrolysis temperature in excess of 1400 "C. It would seem that much of the earlier confusion stems from the use of peak absorbance measurements; however in the present instance it was found that it was essential to use integrated absorbance measurements which provided consistent results. Zhang et al. (95/2393) determined trace amounts of A1 in high-purity tin dissolved in HCl. From bulk experiments withmg amounts of A1 and Sn and XRD measurements the following mechanism was proposed. On drying AlCl and SnC1 were converted into Al(OH) and Sn(OH)Cl respect- ively. During pyrolysis when the Al(OH) was converted into A1,0 no loss of A1 was found but a large amount of Sn(0H)Cl evaporated and a small amount of Sn(0H)Cl converted into SnO which did not interfere with the A1 measurement.It is interesting to note that these workers did not find it necessary to use a chemical modifier and in addition found that with a tin matrix that a pyrolysis temperature of 1900 "C was possible with no apparent loss of Al. They clearly showed that the pyrolysis step which involved heating from 1000 to 1900°C with a ramp of 15s was critical to ensure sublimation of the Sn(0H)Cl. 1.2.4.1. Spectral interferences. Sadagov (96/414) compared three methods of background correction; Zeeman-effect Smith- Hieftje and continuum source for the determination of Ga emission line 287.424nm in the presence of Fe absorption line 287.417nm.It was found that the absolute value of the error of the correction by the Zeeman-effect method was only 20% of that of the other methods. Although the determination of Ga in the presence of Fe is known to be difficult even with Zeeman-effect background correction as there is a known interference owing to the very close proximity of the two lines this work shows that this type of background correction produces less of an error than the other two. Similarly Wassall (95/883) compared the advantages and disadvantages of con- tinuum source and Zeeman-effect background correction methods though no conclusions were presented in the abstract. Zong et al. (95/2220) discussed the background correction problems for the determination of Pb in bone digests in the presence of phosphate as a chemical modifier. Phosphate in various chemical forms is widely used as a chemical modifier for the determination of Pb.In this study the workers used both a Massman-type electrothermal atomizer equipped with transverse Zeeman-effect background correction and a side- heated electrothermal atomizer with longitudinal Zeeman- effect background correction. They used phosphate as a chemi- cal modifier but they found that at the 217.0 239.4 261.4 283.3 and 364.0nm lines in matrices such as bone and other biological materials that could contain relatively high concen- trations of Ca Mg or Na an intense background absorption signal in the atomizer resulted. This caused an overcorrection problem at these Pb lines in the longitudinal Zeeman-effect based system.In a transverse Zeeman-effect system only the 217.0 and 239.4nm lines were affected. The background absorption was thought to be caused by molecular absorption 302 R Journal of Analytical Atomic Spectrometry August 19516 Vol. 11of PO which is formed from the dissociation of phosphate in the presence of metals such as calcium or magnesium. These workers speculated that the PO molecular bands are split in the magnetic field thereby producing a Zeeman-effect over- correction error at the principal analytical wavelength for Pb at 283.3 nm. Tahvonen and Kumpulainen (95/540) found a similar spectral interference at the 217.0nm Pb line with transverse Zeeman-effect background correction when ammonium phosphate was employed as the chemical modifier for the determination of Pb in digested food samples.Other workers have also observed spectral interferences with Zeeman- effect background correction. Epstein et al. (95'2221 ) deter- mined As Sb and Te in lead alloys and high-purity lead. They observed a spectral interference by lead absorption lines resulting from excited-state transitions on both primary As lines at 193.696 and 197.197 nm. Analytical bias was caused by the interference at 197.197 nm but was eliminated by using temperature programming and temporal resolution. These workers concluded with a very valid comment that their study emphasizes the importance of never assuming the absence of spectral interferences even when working in a presumably 'simple' matrix using a technique (ETAAS) that is generally free of such interferences.While during each Update period a couple of papers are reported that consider spectral inter- ferences with Zeeman-effect background correction it has to be noted that for most practical purposes the performance of Zeeman-effect background correction far exceeds that of other methods of background correction but one should always beware of the fact that there is no perfect method of background correction and every method has its limitations. Yuan (95/3620) investigated the interfering mechanisms of perchloric acid and sodium chloride on the atomization of In using molecular absorption spectrometry. The results showed that the suppression effect was primarily due to the formation and loss of InCl after the pyrolysis step. The interference was due to a gas-phase reaction.The application of a palladium and ascorbic acid chemical modifier reduced the molecular absorbance due to InCl and MgCl and no suppression effect was observed. Akman and Doner (95/2121) have continued their investi- gations into interference efSects using dual cauity platforms. In this latest study the interference mechanisms of sodium chlor- ide on Co and Zn were examined. Formation of highly volatile zinc chloride contributes mainly to the losses of Zn especially during high pyrolysis temperatures. The losses of both Co and Zn could be partly attributed to the occlusion of the analytes in microcrystals of sodium chloride and some of these are expelled from the atomizer without decomposing during the atomization step. 1.2.4.2.Chemical modifiers-general. The numbers of pub- lished papers and conference presentations concerning chemi- cal modifiers continues to expand each year. Many of these are more application orientated and hence are discussed in detail in the relevant application based Update. However it is appreciated that those concerned with fundamental ETA research require an overview of the various chemical modifiers that have been examined during the last Update period. In order to reduce the amount of text a reader has to wade through all the relevant work concerning chemical modifiers is listed in an abbreviated form in Table 3. In the text below attention is drawn to work which in the opinion of the Update authors merits more attention or discussion.The use of various chemical modifiers in routine ETAAS work is now widely accepted. However while there are undoubted advantages to using chemical modifiers the disad- vantage is that they have to be added with the sample prior to each measurement cycle. The addition of another reagent can increase the risks of contamination and it also reduces the total volume of sample which can be taken and in many cases increases the over-all time of the complete measurement sequence. Some workers are now turning their attention to the possibility of permanent' chemical modifiers by coating the inside surface of the platform or tube with a suitable metallic coating. This represents a new approach to chemical modifi- cation; however the history of the development of ETAAS is littered with reports of various coatings that have been tried many of these only having a specific application and the disadvantage of the difficulty of producing a reproducible coating in tube after tube.Rademeyer et al. (95/4689,95/C2271) investigated the suitability of a pure metal layer deposited on the inner surface of a graphite atomizer by cathodic sputtering of iridium in a low-pressure argon discharge. The analytical performance of the chemical modifier was assessed by the determination of Cd Mn Pb Se and V. When sufficient iridium 10 mg was deposited a homogeneous layer covering the entire tube was formed that changed only slightly during the entire lifetime of the graphite tube. Integrated absorbance measurements were found to remain constant for more than 700 atomization cycles.The volatile elements Cd Pb and Se were stabilized in the absence of matrix to 800 1200 and 1400 "C respectively and the analytical performance of this permanent iridium coating was equivalent to that of the recommended chemical modifiers for these elements. Unfortunately for the elements Mn and V the rate of atomiz- ation was unfavourably delayed in the iridium-coated tubes and for Mn this was thought to be due to the fact that a stable compound is formed between Mn and iridium. Other workers also investigated using iridium but in conjunction with a zirconium or tungsten carbide coated platform as a permanent chemical modifier. In contrast to Rademeyer et al. (95/4689) Tsalev et al. (96/281) applied a carbide coating by repeatedly pipetting aqueous solutions of either tungsten or zirconium onto the platform and running a complex sequence of different temperature/time programs followed by an iridium solution onto the carbide coating with a total time for the complete coating sequence of approximately 45 min.The iridium-zir- conium carbide coating appeared to be the best for elements such as As Cd Bi Pb Sb Se Sn Te and T1 allowing more than 800 atomizer cycles to be performed providing the pyrolysis atomization and clean-out temperatures were kept below 1400 2050-2100 and 21 50-2200 "C respectively. With the iridium-tungsten carbide coating higher atomization temperatures were required but these were either too near or greater than the temperatures at which iridium is vaporized.Both of these approaches are novel although it remains to be seen how readily they will be adopted or investigated by other groups. The approach proposed by Tsalev et al. (96/281) does have the advantage that it can be performed in any laboratory with current instrumen tation. Workers from Hitachi (95/1880 95/3645 96/353) continue their investigations into the action of various chemical modifiers. In a previous study (see J. Anal. At. Spectrom. 1995 10 199R) these workers found an unexpected increase in the atomic vapour temperature for Se in the presence of a palladium chemical modifier. This work has now been extended to the determination of As in urine (95/1880). Again the two-line method for determining the gas-phase temperature within the graphite atomizer was employed and similar results for As were obtained as for Se.It was found that the addition of potassium permanganate effectively reduced the unexpected atomic vapour temperature increase and allowed a distinct absorbance signal for the atomization of As in urine to be obtained. In another study (95/3645) these workers examined the relationship between the formation of intermetallic com- pounds for Pd-Sn and Pt-Pb by chemical modifiers and observation of the vaporization of intermetallic compounds by means of electron microscopy. For the determination of In Journal of Analytical Atomic Spectrometry August 1996 Vol. 11 303RTable 3 CHEMICAL MODIFIERS ~~ Matrix REEs ~~ Sample treatmentlcomments signal magnitude by 190% and no interferences observed for other rare earth elements.Chemical modifier used to remove chloride interferences worse LODs found with Zeeman-effect background correction than with continuum source. Pyrolysis temperature increased to 850 "C with this chemical modifier mixture. Signal magnitude also increased compared with phosphate chemical modifier combinations. Ammonium vanadate purified in situ by heating to 1700 "C. Platform atomization and method of analyte additons employed. LOD 0.024 pg I-' but lifetime of tubes and platforms only 50 firings. coated graphite tubes. magnitude by 5.7-fold. eliminate interferences. atomization signals to match those of aqueous calibration solutions at an atomization temperature of 2800 "C. LOD 0.03 pg I-'. From 1 1 sample Hg extracted into 3 ml CCl with dithizone organic phase preconcentrated into 0.1 ml DMSO.DMSO also considered as the chemical modifier. LOD 1.3 ng 1-'. Interferences from chloride and phosphate eliminated by use of glucose. LOD for Sb 3 pg 1-'. Results compared well with Mixed chemical modifier of hafnium and tartaric acid for Bi LOD 1 pg 1-' and ammonium dihydrogen phosphate for Ag LOD 0.06 pg I-'. Nickel-based matrix matched calibration solutions. Examination of a range of fluoride compounds as chemical modifiers for P. Sodium fluoride found to be the best. Free from memory effects comparable in thermal stabilization to palladium and lanthanum. Silver extracted from sea water then back extracted into nitric acid plus In. Investigation of suitability of a pure metal layer of Ir deposited by cathodic sputtering in a low-pressure Ar discharge.Lifetime of up to 700 firings Cd Pb and Se stabilized to 800 1200 and 1400"C respectively in absence of matrix. Russian workers examined the use of lanthanum as a chemical modifier for Nb was found to increase the signal magnitude. Aluminium was employed as a chemical modifier for the determination of Zr in the presence of Nb. magnitude and reduced interferences from serum. Continuum source background correction and wall atomization LOD 6.7 pg I-'. Samples prepared in aqua regia LOD 2.3 pg 1-'. Range of REEs examined as chemical modifiers for Dy. Lanthanum 5 pg or gadolinium 10 pg found suitable. Gadolinium preferred reduced LOD memory effects and improved precision. LOD (3s) 1.1 pg 1-'. Discussed the role of metallic chemical modifiers in ETAAS.Addition of ammonium chloride enhanced Mixed chemical modifier for molybdate- Addition of calcium fluoride enhanced signal Use of chromium nitrate claimed to Mixed chemical modifier caused Pb HG-AAS. Use of lanthanum gave a higher signal Chemical modifier Element Ammonium chloride Eu Reference 9511963 9514579 9511918 9514742 9513394 9 5/4 528 9513791 9513379 9513 5 10 9511673 9514696 9513517 95/3376 9514689 9513487 951469 1 9513632 9514692 9514381 Ammonium nitrate Al Cu Water from mud volcanoes Environmental and biological samples Ammonium sulfate-Tris Cd Ammonium vanadate Zn Sea water Ammonium vanadate and Ni Calcium fluoride Mo lanthanum Serum Serum and hair Alkaline sludge River water Chromium nitrate V Co" 0.02% m/v plus NH,H2P04 Pb 1% m/v Snow and river waters DMSO Glucose Sb Se Ag Bi Sea water Nickel-based alloys and steel Hafnium and tartaric acid NH4H2P04 HF NaF KF CsF and NH4F P Biological RMs Indium Ag Ir deposited by cathodic sputtering Various (5) Sea water Aqueous Lanthanum aluminium Hf,Nb Zr - Lanthanum Ge Serum Lanthanum and calcium Lanthanum or gadolinium Si DY High-temperature alloys Serum Metallic elements 304 R Journal of Analytical Atomic Spectrometry August 1996 Vol.11Table 3 (continued) Chemical modifier Element Mixed chemical modifier of ascorbic acid plus 5 pg Re adn 1 pg Rh Au Matrix Biological fluids Sample treatment/comments Mixed chemical modifier allowed interference free measurements with STPF conditions. Maximum pyrolysis temperature 1150 "C LOD 0.12 pg 1-' with a Re-Rh-thiocyanate chemical modifier.modifiers to overcome sulfate interference on Se. With continuum source background correction Pd-Sr or Pd-Ba are recommended the addition of Mg( NO,) considered unnecessary. Slurry sampling. Use of the mixed chemical modifier provided some improvement in the signal magnitude and eliminated need for sample digestion and separation. Use of platform atomization. Interferences due to large excess of tellurium and bismuth matrix removed by use of mixed chemical modifier. LOD (3s) 0.6 pg 1-'. digested with HN03/HF/HC104. Interference from A1 required matrix matching of calibration solutions and samples. Quantitative recoveries of organotin compounds LOD 0.27 pg g-'. nickel chemical modifier allowed a maximum pyrolysis temperature of 2200 "C.After acid digestion of samples 10 pl of nickel solution taken platform atomization. Aluminium concentration > 20% required matrix-matching. LOD 0.27 pg g-'. Investigation of a series of mixed chemical Determination of total Sn sediment samples Urine diluted 1 + 349 with water. Use of Mixed chemical modifier proposed. Mixed chemical modifier in zirconium- coated graphite tube. Addition of nitric acid as chemical modifier eliminated the interference effects from calcium chloride increased the signal magnitude and shifted Mn signal to a later appearance temperature. Review with 257 references concerning organic chemical modifiers. Russian workers studied factors which predict efficiency of organic chemical modifiers though sadly the best not identified! Addition of oxalic acid necessary to reduce background absorption from sea water matrix.Wall atomization with a pyrolytic graphite coated graphite tube at 1300 "C. Graphite tube soaked in PdCl solution. Determination of As"' and AsV by pre- atomization removal of volatile AsCl prior to atomization of As". Comparison of wall platform and probe atomization. For As palladium added with sample. For Cd and Pb palladium reduced on graphite surface prior to sample introduction. Examined possibility of absolute analysis for the determination of Be. Use of palladium allows higher pyrolysis temperature. LOD 0.7 pg 1-'. Palladium found to enhance thermal stability of Pb in urine matrix. Determination of Pb at 217.0 nm peak absorbance measurements and method of analyte additions employed.Shown by XRD and XPS that Te reacts with palladium to form intermetallic compound. Use of optimized conditions to eliminate chloride interferences. Reference 9511496 9512156 9513501 9512251 9514693 961354 96/70 961354 9 5/43 52 9511877 951235 9514410 9514453 95/3386 951976 9 511 48 5 9512150 9512769 9511 116 951540 9514250 95/c725 Mixed Pd-Sr or Pd-Ba Se Sulfate matrix Mixture of Ni V and agar-agar Mixture of strontium and Ge Ga T1 ammonium nitrates Geological Blood Mixture of tartaric ascorbic acids T1 and MgWO,) Bi,Te Sediments Nickel Sn Nickel chloride Nickel solution 2% m/v Si Sn Urine Sediments Nickel 3000 mg 1-' and palladium Nickel 3000 mg 1 - ' and zirconium Nitric acid 1% v/v Mn Se Te B 100 mg 1-' 1000 mg I-' Iron- and nickel-based Iron- and nickel-based alloy alloy Organic chemical modifiers - Organic chemical modifiers Ni Water Sea water Oxalic acid Cd Palladium As Environmental Palladium As Cd Pb Sea water Palladium Palladium Palladium Palladium Be c o Pb Pb Environmental and Serum geological Urine Food materials Co metal and sea water Palladium Te Palladium T1 Journal of Analytical Atomic Spectrometry August 1996 Vol.11 305RTable 3 (continued) Chemical modifier Palladium-Triton X- 100 Palladium +tartaric acid Palladium or palladium-magnesium nitrate Palladium palladium and ascorbic acid and palladium and magnesi um nit rate Palladium-coated platform Palladium Palladium or magnesium nitrate Palladium plus magnesium nitrate Palladium nitrate magnesium nitrate plus barium nitrate and sodium vanadate Palladium 0.1-0.5 mg 1-' +2% m/v citric acid in 0.01 mol 1-' HN0,-0.1% v/v Triton X-100 Palladium-ammonium acetate Palladium-ascorbic acid Palladium-ascorbic acid Platinum Platinum-ascorbic acid Sodium hydroxide 20% m/v Element Cr Ni Cd As Cd Pb Various (16) Te Se Various (6) Se Cr Various (6) Se Se Te Mo Si c o Mn Matrix Whole blood Environmental Chloride matrix Waters and urine Urine Sediment andl biological Biological samples Serum High-purity iron Urine Electrolyte solutions Cerebrospinal fluid Seawater Sample treatmentlcomments Whole blood or serum diluted 2- or 3-fold with chemical modifier.Sample preconcentrated by injecting 5-10 times with oxygen ashing prior to atomization. Optimum conditions found for absolute analysis with tube wall atomization.Use of method of analyte additions and successive dilution whereby dilutions of both sample and sample plus addition are measured. Method was successful in achieving correct analytical results. Pd-ascorbic acid Pd-Mg( N03)2 in terms of pyrolysis temperatures characteristic mass etc. Performance was found to be similar background absorption greater with Pd-Mg(N03)2 hence not recommended. See also ref. 95lC711. on liquid anion exchanger (Amberlite LA.2) followed by ETAAS. Use of grazing incidence X-ray absorption spectroscopy to elucidate Pd-induced stabilization of Se. Transverse HGA system used for a variety of metals. Gas-phase interferences from chloride and sulfate were eliminated with the THGA system. LODs were improved in the THGA system compared with a conventional atomizer.Found to be the optimum for Se in urine. LOD 6 pg 1-'. Comparison of the three chemical modifiers. Sodium vanadate identified as an effective chemical modifier for the determination of Cr in botanical samples. River sediment and sludge no chemical modifier was required. Wall atomization with electrographite tubes. 0.1 % Triton X-100. Various concentrations of the chemical modifier applied depending upon which element was being determined. sample and chemical modifier soak in and hence reaction is facilitated. Ammonium acetate added to remove interferences from chloride and nitrate. Analyte elements preconcentrated by coprecipitation with palladium. Various reductants examined ascorbic acid preferred. Pd-hyroxylamine hydrochloride as chemical modifiers.LOD 0.2-0.3 pg 1- ' . by 15-fold. Absolute LOD (3s) 8 pg. Solutions matrix matched with 0.1 mol 1-' lithium chloride. No interference from matrix. LOD 2.23 pg 1- ' . Sample 80 p1 and chemical modifier 4 pl injected and dried slowly 3-step pyrolysis at 300 600 and 1400"C atomization at 2000°C. LOD 0.051 pg 1-'. Comparison of performance of Pd Te separated from matrix by HG collected Samples diluted to varying degrees with Porous carbon plate used as platform Comparison of BaF Pd-Mg(NO,) and Presence of Pt improved signal magnitude Reference 9512668 9513795 9514510 9514697 9511674 95/C600 9513442 9 5/25 5 8 9512392 9511227 96/97 95/21 10 961304 961280 95/45 16 951899 (96/353) the action of silver gold copper or palladium as chemical modifiers was examined with positive and negative ion SIMS by sputtering with O,+ and Cs' respectively. As the pyrolysis temperature was increased intermetallic com- pounds having progressively higher melting-points were formed.Not surprisingly this effect was found to suppress the vaporization of In. Mahmood et al. (95/549) used SEM and visual examination to study the physical changes occurring with temperature when copper and nickel were used as chemical modifiers for the 306 R Journal of Analytical Atomic Spectrometry August 1996 determination of Se. Previous work from this group (95/C659 and J. Anal. At. Spectrom. 1995 10 199R) has suggested a physical mechanism of stabilization for a palladium chemical modifier whereby the analyte is entrapped in the chemical modifier metal and not released until the temperature is sufficiently high for the analyte to migrate out of the molten chemical modifier.The present work investigated whether a similar mechanism could account for the stabilization effects of copper and nickel. It would appear that both chemical and physical effects can contribute to the stabilization of Vol. 11Se by copper and nickel. At pyrolysis temperatures below 700-800 "C small amounts (0.1 pg) of copper or nickel are sufficient to stabilize 2 ng of Se. This stabilization is thought to be due to the formation of selenides with copper or nickel or the entrapment of Se species in the excess of chemical modifier. At a pyrolysis temperature of 800 "C a dip in the curve is seen and this could be improved either by the addition of an excess of chemical modifier or by the introduction of H2 during the drying step of the atomizer program.The H2 appears to block the active sites and prevent the formation of the thermally unstable Se-C-Ni alloy. The strongest evidence for a physical mechanism of stabilization was obtained from the delay in analyte appearance temperature during atomiz- ation and the decreased absorption when the chemical modifier was present in large excess. While these results appear to agree with the data from Hinds and co-workers (see section 1.2.3. Fundamental processes) concerning the atomization of Si from a gold matrix there is also work that suggests a strong interaction with the surface graphite layers. Clearly there is still room for further investigations to attempt to clarify the mechanism of stabilization with metal chemical modifiers.Many of the SEM micrographs used to support one mechanism or the other show distinct droplets of the metal chemical modifier on the graphite surface. It is still difficult to appreciate how the analyte atoms can migrate to these droplets. If this is really happening then clearly there is some fascinating chemis- try or physical effects taking place on the graphite surface within an electrothermal atomizer at the drying and pyrolysis temperatures typically employed. Scheie and Holcombe (95/C565) discussed the use of chemi- cal modifiers to enhance atomization for direct ETV-MS studies. In these studies the analyte is vaporized and then trapped on a cooled tantalum surface.This second surface is then intro- duced into the vacuum where the analyte is thermally desorbed and detected by MS. However some matrices require the addition of a chemical modifier such as palladium to remove the matrix and hence eliminate the need to determine numerous molecular species along with the atomic signal. 1.2.4.3. Chemical modifiers-palladium. The theme that pal- ladium is a universal chemical rnodijier continues to generate a large amount of literature and comparative studies. Shan and Wen (95/4697 95/C711) made a comprehensive comparison between the performance of palladium palladium plus ascorbic acid and palladium with magnesium nitrate as chemical modi- fiers for the determination of Ag As Au Bi Cd Ga Ge Hg In Mn Pb Sb Se Sn Te and TI in terms of the maximum pyrolysis temperatures characteristic mass values background absorption permissible interference range capability of improving the atomization signal shapes and the precision of determining these elements in real samples.Generally they found that the performance of all three combinations were similar except that the background absorption of the palladium-magnesium nitrate mixed chemical modifier is 1-2 orders of magnitude greater than that of palladium and palladium and ascorbic acid. However this appears to be the only significant difference between the three possibilities. These workers concluded that palladium is a universal chemical modifier. The need for the addition of ascorbic acid appears to be matrix dependent while the addition of magnesium nitrate is not essential though this compound could assist as an ashing aid in the drying and pyrolysis of biological samples. It is well known that the presence of sulfate has a strong suppressive effect on the determination of Se with ETAAS.Ni et al. (95/2156) investigated the use of a mixture of chemical modifiers such as the nitrate salts of palladium lanthanum strontium barium in various combination and with or without magnesium nitrate. Lanthanum nitrate was not able to remove the interference from sulfate on Se while palladium barium and strontium nitrates were able to stabilize Se to a pyrolysis temperature of 800 "C and allow the determination of 1 ng of Se in the presence of 5 pg of sulfate. Improved results were obtained with the use of a mixed chemical modifier of either palladium and barium or palladium and strontium.Recoveries of 85-105 and 90-107% respectively were found when 20 pg of sulfate were present as sodium sulfate. These workers did not find any advantage with the addition of magnesium nitrate to the mixed chemical modifier. Though in contrast Drake and Hain (95/2558) found the combination of palladium magnesium and barium nitrates the optimum chemical modifier for the determination of Se in urine. As discussed above (see section 1.2.4.2. the work of Jackson and co-workers (95/C659) suggested that the stabilizing efSect ofpalladium could be a multi-stage process whereby the analyte is first vaporized from the slurry or solution particle then collides with excess palladium becomes physically entrapped in molten palladium droplets and is slowly released from these droplets into the gas phase.In contrast to this work that of Styris et al. (95/C600) applied grazing incidence X-ray absorp- tion spectroscopy (XAS) to elucidate the palladium induced stabilization of Se. This was performed on polished pyrolytic graphite having monolayer surface concentrations of palladium and Se. These measurements indicated that the palladium and Se are present in the interstitial regions between the basal graphite plates. These findings appeared to show a clear role for the graphite surface in the stabilization mechanism for palladium which is ignored in the heterogeneous mechanism proposed by Jackson and co-workers. Itai et al. (96/97) pro- posed the use of a porous carbon plate as a platform.The porosity of the plate facilitates the reaction between the palladium chemical modifier and the analyte as the solutions permeate into the graphite plate. The method was applied to the determination of Se in serum and it was claimed that the use of a porous graphite plate eliminated interferences. 1.2.4.4. Other chemical modijiers. In an effort to develop a 'universal' atomizer programme for the determination of Al As Cd Co Cr Cu Mn Ni Pb Sb Sn and V in pharmaceuticals Arpadjan and Alexandrova (95/4698) found that the sample matrix was an effective chemical modifier and it was not necessary to add any other reagent. The chemical modification effect of the sample matrix was Lhought to be due to the reducing nature of the samples.Sramovka et al. (95/4693) determined T1 as a dopant in layered monocrystals of bismuth telluride. The AA signal for T1 was found to be influenced by the large excess of H,TeO and Bi(NO,) obtained by dissolv- ing the samples in dilute nitric acid. Losses of T1 were considered to be due to the volatility of TlNO above 300 "C and the volatility of T102 and TeO which were thought to participate in recombination reactions at the start of atomiz- ation. These interfering effects were removed by the use of a chemical modifier consisting of an optimized mixture of tartaric acid ascorbic acid and magnesium nitrate. French workers (95/4579) compared the determination of A1 and Cu in water from mud volcanoes using ammonium nitrate as a chemical modifier with continuum source and Zeeman-effect back- ground correction systems.Surprisingly these workers found that the LODs obtained with continuum source background correction for A1 and Cu 0.3 and 0.6 pg l-' respectively were slightly lower than those obtained with Zeeman-effect back- ground correction 0.5 and 0.8 pg l-' respectively. This is contrary to what one would expect given the nature of the matrix and the generally accepted performance of both back- ground correction systems though no conclusions were pre- sented in the abstract. The comparison of different chemical modifiers for the determination of Au discussed last year (see J. Anal. At. Spectrom. 1995,10 199R) has now been published (95/1496). Glucose was examined as a chemical modifier for Journal of Analytical Atomic Spectrometry August 1996 Vol.11 307 Rthe determination of Sb and Se by Perez-Corona et al. (95/1673). Applying wall atomization these workers found that glucose allowed the thermal stabilization of Sb and Se up to 1300 "C. In addition this chemical modifier significantly reduced the interference by phosphorus on the determination of Se and chloride on the determination of Sb. Yuan et al. (95/4256) compared the use of ammonium diphosphate ammonium fluoride ammonium sulfate pal- ladium chloride chloroplatinic acid and various combinations of chloroplatinic acid with organic acids as chemical modifiers for the determination of Cd in pigs liver. Chloroplatinic acid was the preferred chemical modifier. 1.2.5. Developments in technique Gilmutdinov and co-workers (95/C722 95/C3738 95/C4233) have continued their work on spatially resolved detection of analytical signals in ETAAS.Based on the work discussed last year (see J. Anal. At. Spectrom. 1995 10 199R) experimental results from a detection system using a linear photodiode array located vertically along the exit slit of the monochromator have been presented. The spatially resolved radiant intensities from the light sources before and after passing through the graphite atomizer have been assessed (95/C722). In addition the system also allows the simultaneous detection of the transient radiation from the tube walls and platform which can be used to monitor the respective temperatures. What the data in these presentations show is that very few of the prerequisites for Beer's law and its application to AA are fulfilled.In principle non-monochromatic and highly non- uniform radiation from the emission source interacts with non- uniformly distributed atoms in a non-isothermal environment (95/C3738). Spatially resolved detection of the atomization process in a graphite atomizer allows these factors to be taken into consideration and will extend the linear dynamic range of ETAAS which in a conventional spectrometer is considerably reduced due to the fact that none of the requirements for Beer's law are satisfied. The investigation into the effect of atomization temperatures on the ability to determine several elements simultaneously with a THGA and discussed last year (see J. Anal. At. Spectrom. 1995 10 199R) has now been published by Harnly and Radziuck (95/1493).Clearly what this publication shows is that while the over-all signal magnitude may be reduced at higher temperatures if the system is photon shot noise limited and the integration interval is optimized then there is no reduction in the S/N at higher temperatures hence there is no compromise in performing simultaneous multi-element deter- minations for elements of widely differing volatilities with respect to detection limits. Fernando and Jones (95/1031) have improved the continuum source ETAAS system with photodi- ode array detection discussed previously (see J. Anal. At. Spectrom. 1995 10 199R). For single-element determinations the detection limits are comparable to those obtained with HCLs and Zeeman-effect background correction at wave- lengths down to 213 nm and in most instances are better than those previously reported for continuum source AAS. It appears that the detector on this system covers a 10nm window and that five elements having absorption lines within this window can be determined simultaneously. At wavelengths <240 and > 250 nm the limiting noise sources are detector readout noise and signal shot noise respectively.Smith and Harnly (95/2359) discussed the analytical figures of merit obtained for ETAAS with continuum source and linear photo- diode array detection for As Cd Co Cu Fe Mn Ni Pb Se Sn and Zn. The intrinsic masses were calculated based on the wavelength and time integrated absorbance in units of pm s-'. A simple triangular model based on the experimental values for the HCL linewidths and calculated widths for the absorp- tion profiles was used to convert published characteristic mass values for conventional ETAAS to intrinsic mass values.With the exception of As and Se the intrinsic mass values for continuum source and conventional ETAAS agreed within a factor of two. The continuum source detection limits obtained were similar to those reported for line source ETAAS. The work described by Smith and Harnly on the use of a double atomizer for ETAAS mounted in a modified vacuum cross and discussed last year (see J. Anal. At. Spectrom. 1995 10 199R) has now been extended and published (95/1492). The absorbances for Al Cd Cr Mn and Pb integrated with respect to wavelength and time were determined at pressures as high as 607 kPa from either the platform of an ICC or the atomizer wall of an end-capped ICC.In a standard ICC the maximum increase in signal was limited to a factor of three at 607 kPa owing to a strong convection component which was reduced but not eliminated for Cd and Pb when the injection port was plugged. Signal magnitudes were improved by an average factor of 3.2 with an end capped ICC but the convection component was not reduced. Agreement was obtained between the experimental data and a simple model that incorporated diffusion and convection loss mechanisms. Recoveries of Pb from 3% m/v sodium chloride increased from 50% at 101 kPa to approximately 100% at 607 kPa in a standard ICC with the injection hole plugged. Detection limits were 50 100 30 and 90 fg for Cd Cr Mn and Pb respectively at 607 kPa with the injection port of an end-capped ICC plugged.While it is clear from the work performed by the two research groups led by Holcombe and Harnly that a lot of interesting information and excellent analytical perfomance can be obtained by placing an ETA inside a modified vacuum cross one has to ask whether such an approach will ever be more than a research tool. The problems of automating the sample introduction and operation of the ETA within the vacuum cross would appear to be a major limitation in moving forward. Hungarian workers (95/2317) have improved the simul- taneous multi-element capabilities of a FANES system by applying a micro hollow cathode cup instead of the macro graphite tube normally used.The S/N was improved by using a modulated excitation current reduced gas pressure and a correction for the interference signals produced during the excitation pulses. Although not AA the use of a graphite electrothermal device for molecular absorption measurements is considered to be part of the subject matter considered in this Update. Butcher and Fender (95/C3025) applied Smith- Hieftje background correction for the determination of fluoride chloride and bromide. The technique involves adding a reagent metal to the sample solution to allow the formation of a diatomic molecule. Chloride was determined by adding alu- minium to allow the formation of the AlCl molecule. By employing a lead HCL at 261.4nm the absorption by this molecule can be determined.However one of the disadvantages of this technique is that the high levels of reagents produce large spectral background levels. While most studies have used continuum source background correction this cannot cope with the large levels of background produced. In this study both continuum source and Smith-Hieftje background correc- tion were compared. For the determination of fluoride no signals were obtained by Smith-Hieftje due to the large width of the molecular line but for the other halides improved sensitivity and precision were obtained compared with con- tinuum source. However although the procedures developed were applied to the determination of halides in real samples severe chemical interferences were observed. The topic of standardless or absolute analysis continues to generate a steady trickle of publications and presentations though the comments made in last years Update (see J.Anal. At. Spectrom. 1995 10 199R) still hold true. Apart from 308R Journal of Analytical Atomic Spectrometry August 1996 Vol. 11Chinese workers an additional country is represented this year Russia by L‘vov (96/C234). In a XXIX CSI presentation L‘vov (96/C234) discussed the prospects and problems of absolute analysis by ETAAS. Although the possibility was originally noted by Walsh in 1955 (Spectrochim. Acta 1955 7 108) 40 years later it is clear that not all the problems have been solved especially with respect to providing the formation of a sample vapour column of known geometry total atomiz- ation of the sample and obtaining a background-free analytical signal. Although many technical developments have moved in these directions additional uncertainties in the theoretical calculations and experimental studies required to determine the atomic constants such as diffusion coefficients oscillator strengths a-parameters and the hyperfine structure of analyt- ical lines remain.However these problems notwithstanding workers continue to pursue the aim of absolute analysis. Li et al. (95/4374) considered the effects of the slit width the lamp current the sample injection volume and the cooling step prior to atomization on the experimental characteristic mass values for As Cd Co Cr Cu Mn Mo Ni Pb and V. They concluded that factors which have significant effects on the experimental characteristic mass values must be eliminated or controlled to obtain more stable values.These same workers (95/3657) considered the stability of the experimental charac- teristic mass values obtained by different workers from differ- ent samples and on different days and with three different commercial AA spectrometers. Sadly there are no conclusions presented in the abstract. Yang and Ni (95/2150) investigated the possibility of absolute analysis for the determination of Be in environmental and geological samples. They examined the factors that influence the characteristic mass for Be including lamp current pyrolysis and atomization temperatures the use of chemical modifiers and atomization from either the tube wall or platform. The atomization efficiency [defined as the ratio m,(calc)/m,(exp) x 1001 of Be on the tube wall or platform was found to be temperature dependent and to be about 90% at the atomization temperature of 2700K.The experimental data showed that Be could be determined in environmental and geological samples using the experimental characteristic mass values and with palladium as chemical modifier. The problem with using the experimental character- istic mass value is that this has to be determined and from the point of view of this author (I.L.S.) this seems to miss the point of absolute analysis in that such a procedure should not require calibration standards to perform the analysis. Ma et al. (95/3795) used a chemical modifier of palladium and tartaric acid to determine Cd in environmental samples by absolute analysis.Zheng and Su (94/1175) applied absolute analysis to the determination of Ag and Cd in sediment and geological samples. Good correlation was observed between the exper- imental characteristic mass values for commercial instruments and the theoretical values derived from a mathematical model. The control of temperature in an ETA is important and Histen and Holcombe (95/C604) discussed a new method for temperature control. The ability to control and measure the temperature is essential in the determination of atomization mechanisms. The method allows the use of different atomizer designs as the system is ‘calibrated’ by applying a known constant voltage and measuring the temperature for a fixed time period. The measured temperature is then used in a series of calculations that give values that describe the various atomizer heating parameters.These parameters enable power calculations to be made for the desired heating rate. Booth (96/318) discussed a variety of procedures for method develop- ment in ETAAS including dual background correction both continuum source and Zeeman-effect in the same instrument the ability to generate pyrolysis and atomization curves auto- matically automatic quality control software and ‘graphite furnace television’ which allows real-time video images from within the graphite tube to be displayed on the computer screen. Chernyakhovskiy et al. (95/2828) applied a ‘fast furnace’ program to the determination of Pb in microwave digests obtained from dust wipe and air filter samples. The atomizer program was 36 s long and included heated injection fast drying and sufficient removal of matrix prior to atomization to avoid interferences and no cool down step before atomiz- ation.However with a 1000-fold dilution of the digest prior to analysis little matrix would be taken into the atomizer and in such cases ‘fast furnace’ programs can be successfully applied. Wang and Demshar (95/51) described a quick and simple procedure for determining Cu in serum and urine with Zeeman- effect background corrected ETAAS. Serum was diluted 10-fold and urine 2-fold and a short atomizer program (30 s) employed. It is interesting to note that such dilutions are normally employed for the determination of Cu in serum by FAAS. The pyrolysis temperature for both the conventional and fast furnace program was 600 “C which is rather low for Cu and this matrix; consequently it is not surprising that the workers did not find any differences between a 30 or 5 s hold time at this temperature.The within- and between-batch precisions were 2.6 and 3.4% and 0.9 and 2.5% for serum and urine Cu at concentrations of 30.4 and 2.70 pmol l-’ respectively. 1.3. Chemical Vapour Generation There is a sustained interest in the use of methods incorporating a chemical vapour generation (CVG) stage for the determi- nation of As Hg and Se and a large number of publications concerning these three elements alone have appeared during this review period. In addition there is still an interest in utilizing the general approach for Bi Ge Sb and T1 and there is a growing interest in the use of the methodology for the determination of Pb and Sn.Reports have appeared during this period in which convincing evidence for the determination of Cd by cold vapour is presented. There is continued growth in the application of CVG methods as part of speciation studies the majority of which concern the coupling of chroma- tographic separation with element specific detection. Also a growing number of publications describe methodology in which spectrometries other than absorption have been used. Information about applications of plasma sources for emission and mass spectrometry are to be found in other Updates but work involving AFS is included in this section. Chemical vapour generation with ETAAS is considered in section 1.2.2.3.1.3.1. Hydride generation The atomic spectrometry community has been done an enor- mous service by Dedina and Tsalev whose book ‘Hydride Generation Atomic Absorption Spectrometry’ (95/3514) appeared in 1995. The authors have done a wonderful job in collecting and categorizing nearly 1600 references. To quote from a forthcoming review in J. Am. Chem. Soc. ‘The text moves smoothly from tutorial introductory material concerned with basic principles of atom formation and spatial distribution in tubular atomizers to atomization mechanisms and then a very important chapter concerned with the mechanisms of the various types of interference that are encountered. The first part of the book concludes with a general overview of hydride generation and a brief survey of other volatile metal com- pounds that have been used as atom precursors.An appendix to this part deals with some basic atomic absorption theory as it relates to the magnitude of the absorption coefficient. The second part of the text is devoted to methodology and analyt- ical applications and consists of ten chapters dealing with the elements antimony arsenic bismuth germanium indium lead selenium tellurium thallium and tin. Each chapter is organized along the same lines and consists of sections on general Journal of Analytical Atomic Spectrometry August 1996 Vol. I 1 309Rcharacteristics optimization of instrumental and chemical par- ameters control of interferences preconcentration and separa- tion techniques speciation studies and the analysis of real samples.This latter section is further subdivided into sections concerned with environmental samples agricultural samples food and beverages biological and clinical samples geochem- ical samples and metallurgical and industrial samples. The final section of the book is devoted to several appendices which provide lists for several methods of classification of the references. Thus there are lists for technique and analyte (containing no less than 20 different categories) speciation sample matrix (divided into hydride generation and other volatile derivatives) and finally a list of other relevant key words'. Reviews by Chinese workers have also appeared (95/946 and 95/1869). The second (95/1846) is in English by workers from the Research Centre for Eco-Environmental Science Academica Sinica and covers the material contained in nearly 300 references.1.3.1.1. General studies of fundamentals techniques and instrumentation. A brief review (95/3417) covering 22 references covered topics specifically related to the chemistry of the generation reaction such as the non-nascent hydrogen mechan- ism the alkaline mode of generation and the reaction sequence. Less commonly used techniques were reviewed (95/2242) by Huang who covered HG from non-aqueous media as well as electrochemical generation and the performance of a nebulizer HG device (which used borohydride as the generant) for ICP- AES. Sanz-Medel and co-workers have reviewed the use of organized media in CVG procedures (95/4584). Most of these concern HG. Evidence for the presence of Cd atoms at room temperature was presented (see later).A number of papers (95/3455 96/289 95/C4327) have reported on the use of microwave assisted reductions from higher to lower oxidation states. The procedure was applied (95/3455) to the on-line reduction of SeV' to Se" as the basis of a speciation procedure. In the second method (96/289) selenium species were precon- centrated by SPE and detected by AFS with an LOD of 40 ppt. The procedure has also been developed (95/C4327) for discrimination between SbV and Sb"' . Microwave energy has also been used (95/539) in the on-line digestion of wine beer and fruit prior to the determination of Pb. The LOD for wine was 10 ppb. For fruit the hydride was generated from slurries of the fresh sample. The use of a slotted tube atom retarder (STAR) has been described (95/2502,95/2248,95/3635,95/4467 and 95/4621).In the determination of As (95/2248 95/3635) the STAR increased the sensitivity by a factor of 3.5 to give an LOD of 1.8 ppb. For the determination of Sn (95/2502) the use of a STAR and HG produced a sensitivity enhancement of a factor of 200 compared with conventional nebulization. The procedure was applied to the analysis of copper-based alloys. The work was later extended (95/4261) to the determi- nation of Bi in the same materials. The LOD was 10 ppb in solution. The STAR gave improved tolerance to the inter- ferences by transition metals and other hydride forming elements. Chinese workers (95/4467) used a similar procedure for the determination of Bi in geological materials.Despite the fundamental role for the gas-liquid separation process there appears to have been relatively little work on this topic published in the review period. The relative perform- ances of membrane and U-tube devices have been evaluated (95/3626) for AFS. No significant differences were found with respect to sensitivity precision LOD. suppression of inter- ferences and ease of operation. Brindle reported (95/C801) on the evaluation of two commercial and one home-build gas- liquid separators. A separator containing a glass frit gave better peak height response than the other designs. Several different research groups have published schemes for non-chromatographic speciation. Ruede and Puchelt (95/2842) devised a scheme for the determination of four As species based on the effects of different acid concentrations on the FI sensitivity of AsV AS'" monomethylarsonic acid (MMAA) and dimethylarsinic acid (DMAA).Bryce et al. (96/289) and Pitts et al. (95/3455) determined Se" and SeV' sequentially by first determining SeIV only followed by the determination of total Se after on-line microwave assisted reduction with hydrochloric acid. Burguera et al. applied (95/C4327) the same procedure for the determination of Sb"' and SbV. A procedure for preconcentration and matrix separation of As (95/C2946) involving evolution of the hydride trapping on Chromosorb P impregnated with iron(m) chloride and elution with hot dilute acid has been described. The resulting solution was then analysed by a variety of atomic spectrometry tech- niques including HGAAS.Lead has been preconcentrated on a sulfonated dithizone resin (95/2328) and determined by HG directly from the surface of the solid in a medium containing hydrochloric acid hydrogen peroxide and borohydride. A detection limit as low as 25 ppt and improved tolerance to interferences were obtained. A comprehensive study of the interference eflects of a number of elements on the HG determination of Sb has been made (96/275). A scheme covering all mechanisms of interference was proposed and guidelines established based on the use of conditional reduction potentials and their variation with pH for ascertaining whether a particular species will interfere. It has been shown possible to reduce the extent of interferences on Se (95/4699) by using the alkaline mode of generation.Masking agents were added followed by raising the pH to 12- 13 with sodium hydroxide solution. On adding borohyd- ride Se" was reduced to selenide Se". The hydride was produced on acidification with hydrochloric acid in a flow system. Interferences have also been discussed in conference presentations (95/C3745 95/C4238). A study (95/2401) of the kinetics ofthe reduction ofSeV' to SeIV in 6 mol 1-1 hydrochloric acid showed that complete reduction could be obtained in as little as 6 min by heating to 70 "C. The activation energy was measured to be 90.4 kJ mol-'. In contrast it was found (95/1672) that better performance (in terms of precision) in the determination of TI could be obtained by cooling the reaction manifold to 0 "C.It was thought that the volatile species was T1' hydride. A quantitative evaluation of the performance of a single- strand Nafion dryer used to remove the water vapour in the transfer line between the gas-liquid separator and the atom cell has been made (95/4583). It was found that up to 2.3 mg rnin-l of water vapour could be removed representing an efficiency of over 90%. Decreases in sensitivity of between 3 and 5% were observed for As Hg and Se. Mercury loss through the membrane was estimated to be no more than 0.04%. Chinese workers have described (95/4247) a novel intermittent flow procedure described as a hybrid of FI and continuous flow methods. Slovakian workers have confirmed (95/C4205) the utility of Brindle's reagent (L-cysteine) for the reduction of AsV to As"'.Other operating parameters were optimized. Optimization was also at the heart of work (95/3099) on the determination of organotin compounds. Application of a simplex procedure showed that the design of the quartz cell atomizer was not critical provided that the flow of helium carrying the organotin hydrides was not mixed with hydrogen in as small a volume as possible and that the introduction of oxygen should be made as close as possible to the light path. The same researchers have reported on a similar study (95/541) of relevant parameters in the determination of Se. Interestingly no factor was found 'really essential for the S/N. For the determination of As (95/1889) response surfaces 31 0 R Journal of Analytical Atomic Spectrometry August 1996 VoE.11which mapped the role of various operating parameters on the sensitivity were established. 1.3.1.2. Determination of individual elements. As far as poss- ible this section is organized alphabetically by determinand. In some publications more than one analyte is measured and such papers are discussed once only in the section correspond- ing to the element first in the alphabet. By far the largest number of papers are devoted to the determination of As and to Se. Compared with last year there has been an increase in the numbers of papers concerned with some aspect of speci- ation studies. Arsenic has been determined in waters (95/4426 95/3497) and in soils and waters (94/2891). In the second report an on-line stopped flow reduction with KI solution was used to reduce AsV to As"' in 40 s.The LOD of the FI procedure was 0.1 ppb. A FI procedure has been developed (95/2884) for the determination of As and Se in spinach and tomato leaf RMs by workers at the National Institute of Standards and Technology. A batch wet ashing procedure was followed by reduction of AsV to As"' with ascorbic acid and KI solution in 60 min before injection into the FI system. For the reduction of SeV' to Se" the solution was heated with hydrochloric acid (1 + 1) at 90 "C for 25 min. The LODs were 0.15 and 0.17 ppb for As and Se respectively. In a follow-up paper (95/3091) a microwave digestion procedure was used in a very similar procedure for the same elements in a sludge RM. For the determination of As in blood and serum (95/1112) it was shown that the digestion and reduction could be performed on-line in an FI system.A focused microwave reactor was used and L-cysteine was used as reductant. As the sampling frequency was 10 h-l the total residence time in the system including that needed for reduction of AsV to As"' was 6 min. The LOD was 0.25 ppb for a 500 pl sample volume. For the determination of As in steel and geological material a pro- cedure incorporating a STAR has been described. The LOD was 1.8 ppb. For the determination of As in thermally cracked gasoline (95/1087) a combustion procedure followed by HG had a LOD of 5 ppb (m/m). Combustion in oxygen has been used (95/4507) as the first step in procedures for the determi- nation of As in pond sediment mussel seaweed and rice flour RMs.The arsenic(1v) oxide produced was dissolved in water and reduced to As"' with KI and hydrochloric acid. Procedures for the determination of As and Sb in traditional Chinese medicines have been developed (95/3618) and well as As in high purity germanium oxide (95/3616). In the determination of As in urine it is useful to be able to discriminate between the non-toxic forms of As (such as arsenobetaine and arsenocholine) and the toxic inorganic and methylated forms. A procedure based on controlling the acidity and KI concentrations to values at which only inorganic As species form arsine with borohydride has been described (96/149). Recoveries of inorganic spikes were more than 90% but the method was not validated by any other procedure. A method (95/2842) for four As species (arsenate arsenite MMAA and DMAA) based on a similar procedure but also including the addition of oxidizing agents has been developed.An off-line low-pressure cation-exchange chromatographic method has been developed (95/2117) for the determination of the same four species with LODs of around 1 ppb. Again validation was by spike recovery. Workers at the Netherlands Energy Research Foundation have devised a procedure in which the inorganic As species are selectively co-precipitated and determined. The methylated species were converted to volatile hydrides separated passed through a Nafion dryer cryogenically trapped on a GC stationary phase and sequen- tially determined by selective volatilization. A somewhat simi- lar procedure was described by Japanese workers (95/2266) who also included the measurement of the trivalent methlylated species monomethylarsonous acid and dimethylarsinous acid produced in sediments by the action of hydrogen sulfide from the corresponding pentavalent species.A two-stage HPLC procedure for the determination of six As species has been developed (94/2926). A Hamilton PRP-X100 anion exchange column was used. A cartridge was first placed in front of the column to retain all anionic species allowing cationic species (arsenobetaine and arsenocholine) to be separated and quant- ified after microwave assisted decomposition with potassium persulfate. The anionic species (arsenate arsenite MMAA and DMAA) were separated after removal of the cartridge and the substitution of water for the persulfate solution.A very similar procedure (95/1043) based on the use of a Polysphere SAW anion exchange column has been developed for the determi- nation of these As species in urine. The method has also been used with ICP-MS. A better separation was achieved by ion- pair chromatography on Phenomenex CI8 with 10mmol 1-' heptanesulfonate in 0.1% methanol at pH 3.5 as the mobile phase. An indirect procedure for the determination of I based on the distillation of iodine and its interaction with As"' has been developed (95/4499). The calibration covered the range 25-200 nmol 1- Two reports (95/4467 and 95/4261) on the determination of Bi by HG procedures incorporating a STAR have appeared. The methods were used for the determination of Bi in geological materials and copper-based alloys respectively.It is not clear what advantages this procedure offers over conventional HG in either the batch or FI format. An indirect procedure (95/4375) for the determination of Bi based on interference in the AF signal from mercury has been described. The LOD was 0.5 ppb in water. Convincing evidence for the formation of cadmium atoms at room temperature has now been provided by Sanz-Medel's group in Oviedo (Sanz-Medel et al. Anal. Proc. 1995 32 49 95/4349 and 95/4584). Much the same data are published in all three references showing the absorption profile of the absorbing species obtained with a high resolution continuum source spectrometer. A comparison with the absorption profile for Hg obtained with the same apparatus is also provided.It is proposed that the species stripped from solution is an unstable cadmium hydride and this decomposes at room temperature to release Cd atoms which can be detected by AAS as the formation of Cd metal appears to be relatively slow although favoured thermodynamically. The Sanz-Medel group reported that the presence of an organized medium (vesicles of didodecyldimethylammonium bromide) was essen- tial but this finding is not supported by another more recent report (96/277) of cold vapour Cd. The LODs were less than 100 ppt. In the latter study the sensitivity was enhanced by the addition of thiourea and cobalt. This result had been previously reported (96/279) in a procedure for the determi- nation of Cd by ICP-AES. Reviews of the determination of Pb have appeared in the Chinese (95/2499) and western literature (95/3649).An oxidiz- ing medium is required for the generation of plumbane presum- ably to maintain the Pb in the quadrivalent state. It is also known that the process depends critically on the acid concen- tration. A procedure (95/1978) which used a citric acid- potassium hexacyanoferrate(n1) medium gave an LOD of 0.23 ppb and some enhanced tolerance to interference due to the masking action of the citrate although copper still posed a severe problem. The use of nitroso-R salt in an FI procedure (95/2822) gave an LOD of 1 ppb and an increased resistance to copper interference compared with the corresponding batch procedure. The method was applied to blood hair rice and peach leaves. A nitric acid-hydrogen peroxide medium was used (95/539) in an FI procedure involving the introduction of slurries and on-line microwave assisted digestion.An LOD of Journal of Analytical Atomic Spectrometry August 1996 Vol. 11 31 1 R10 ppb was obtained for the analysis of wine (for which the microwave digestor was not needed). A spectacular improve- ment in the LOD to 25 ppt was obtained by workers at Osaka University (95/2328) who preconcentrated the Pb on sulfonated dithizone loaded resin (Amberlite IRA 400 anion exchange resin) followed by generation of the plumbane from the solid material in a medium containing hydrochloric acid and hydrogen peroxide. A batch procedure for the determination of Sb (95/2863) has been applied to the analysis of copper concentrates. A mixture of hydrochloric acid and KI was used to reduce SbV to Sb"'.An FI-AFS procedure (95/3686) gave an LOD of 0.19 ppb. A continuous flow procedure (96/274) with non-dispersive AFS detection had an LOD of 22 ppt. L-Cysteine was used to reduce Sb" to Sb'" and to decrease interferences. The method was applied to the analysis of sediment copper and riverine water CRMs. An HPLC method for the separation and quantification of SbV and Sb"' has been described (95/4701). A strong anion- exchange column was used with a phthalic acid mobile phase. Post-column hydride generation from hydrochloric acid solu- tion showed interestingly enough the same sensitivity for SbV as for Sb"'. The LODs were 6 and 50 ppb for SbV and Sb'" respectively. Selenium has been determined in soils (95/C4214) and together with Te in some health foods (95/C4308) by HGAAS and in geological materials by HGAFS (95/4377).Detection limits down to l o n g g-' were possible. For some types of mineral samples (95/3675) interferences were removed by the addition of EDTA and extraction with dithizone in chloroform. The determination of total Se in urine is difficult because some of the Se species notably trimethylselenonium (TMS) are resistant to degradation by acidified borohydride solution (nor do they form a volatile hydride) and so a preliminary oxidative degradation is needed. A procedure using sulfuric acid and hydrogen peroxide was applied to the analysis of blood serum and urine (95/2736) with good agreement with the results obtained by use of an IUPAC procedure incorporating per- chloric acid.For the determination of Se in water sediment coal and some biological materials (95/2544) the pretreatment involved UV irradiation of a solution containing bromate bromide and hydrochloric acid. After further acidification and the addition of hydroxyammonium chloride the solution was heated with nitric and perchloric acids. After decomposition of the acids further hydroxylamine hydrochloride was added and the solution heated to boiling. The hydrogen selenide formed on the addition of borohydride was trapped cryo- genically before determination by AAS. For the determination of Se in vegetables and fruits (95/4555) the pretreatment was not quite so harsh. Even so a nitric-perchloric digestion was used with reduction of Sev' to Se" effected by hydrochloric acid with heating to 100 "C for 10 min.A batch generation procedure was used with an LOD of 0.21 ppb. A procedure which used HG as a separation step (95/4508) followed by trapping and subsequent determination by HG has been applied to the determination of Se in volcanic rocks. An alternative approach to the elimination of interferences from high concentrations of other elements is to use the alkaline mode of generation (95/4699). After adding masking agents the solution was made alkaline so that on addition of the borohyd- ride the Se" was reduced to selenide. When the solution was acidified the volatile hydride was formed. The method was applied to the analysis of nickel oxide and low alloy steel CRMs. An FI procedure (95/4384) with on-line SPE precon- centration on alumina was applied to the analysis of natural waters.For a 25 ml sample volume a preconcentration factor of 50 and an LOD of 6 ppt were obtained. As part of a speciation study (96/289) SeIV and SeV1 were retained on a strong anion exchange resin (Dowex-1) and eluted sequentially with formic acid and hydrochloric acid respectively. An on- line microwave reactor was used to reduce SeV' to Se". The LODs were 4 ppt and the method was applied to the analysis of waters and haemodialysis samples by AFS. The use of microwave assisted reduction was also reported by Pitts and co-workers (95/3455) for use in conjunction with an AAS method for Se in a water CRM. The procedure was sub- sequently developed (95/3725) as an HPLC post-column reac- tion scheme for the determination of SeV' and Se" after separation on a Benson BAX-1 0 high-capacity anion exchange column. The LODs were 0.2 and 0.3 ppb for Se" and Se" respectively.An HPLC method has been developed (95/3850) for the determination of TMS SeV' and SetV in tap water. After separation with a phosphate buffer at pH6.8 on a Hamilton PRP X-100 anion exchange column the eluent was mixed with persulfate passed through a microwave reactor cooled and merged with borohydride solution. Recoveries of around 100% were obtained for all three species. An HPLC method for the selective determination of selenomethionine in nutritional supplements has been developed (95/2303). The procedure involved derivatization with 1-fluoro-2,4- dinitrobenzene (Sanger's reagent) removal of the excess reagent and extraction from acid solution into ether.The derivatized selenomethionine was separated by HPLC on a silica stationary phase modified with p-nitrophenyl moieties with a methanol mobile phase. The Se was determined by thermo- chemical HGAAS. A comparison of methods for the determi- nation of inorganic Se in soil-water-plant systems (95/1153) in which methods using HGAAS ion chromatography or visible spectrophotometry were evaluated was somewhat inconclusive. It was noted that the AAS method had the best sensitivity and detection limit. Tin has been determined in various steels and nickel superal- loys (95/3806) and in geological materials and lead ingots (95/3385) by batch HGAAS procedures. Thorburn Burns and co-workers (95/2502) used a STAR for improved sensitivity and precision in the determination of Sn in copper-based alloys.Organotin species have been determined (95/3099) by GC with quartz furnace AAS detection. Detection limits were between 30 and 130 pg depending on the species. A commercial purge and trap dynamic headspace concentrator has been modified (95/C2905) so that volatile Sn species were swept from the sparging cell and cold trapped. The thermally desorbed Sn species were quantified by flame photometric detection after GC separation. The quality control issues in the determination of tributyltin in various seafoods by a procedure involving HG cryogenic trapping GC separation and AAS detection have been discussed by Pannier and co-workers (96/74). An HG procedure for the determination of Te in a variety of environmental and biological materials has been outlined (95/C4304).Although the toxicity of Te has not been widely studied its relation to other metalloids suggests that it may be harmful. 1.3.2. Mercury by cold vapour generation There is still a considerable interest in the determination of mercury by the cold vapour method. Developments in instru- mentation appear to be largely the results of efforts by instru- ment companies whereas developments in sample pretreatment have been reported primarily by academic or applications laboratories. Clinical and geochemical laboratories are to be found in this latter category. As is the case for the hydride- forming elements not all of the determinations have been performed by AAS; a substantial minority were performed by AFS.There is a growing interest in speciation studies especially with regard to environmental materials. There are some difficulties in the classification of published work. However the following basis has been used for this Update. If the method involves reduction of the mercury 312R Journal of Analytical Atomic Spectrometry August 1996 Vol. 11species in solution to Hgo followed by stripping and measure- ment then the work is included in the present section. If the mercury is volatilized from solution as a volatile organic species the work is discussed in section 1.3.3. The determination of mercury and organomercury compounds in fish and other aquatic food sources has been reviewed (96/67 j. Representatives from several instrument manufactur- ing companies have reported on developments in instrumen- tation (95/C3000 95/3656 95/1506 95/C2896) that is capable of providing rapid automated determinations down to ppt concentrations.For the determination of total mercury it is necessary to convert all forms of mercury in the sample to the free mer- cury(1r) ion in solution. Typically an oxidative attack is used followed in some cases by destruction of the excess oxidant prior to the generation of Hgo by the addition of a reducing agent. A number of papers in the present review period have been concerned with the oxidative attack stage of the method. For the determination of Hg in waters Cossa and co-workers (95/1506) used nitric acid as the oxidant in a method that involved AFS with amalgam (gold-platinum) preconcentration in the vapour phase.A new design of gas-liquid separator was used in which the solution was purged with argon delivered via a sintered glass disk. For a 45 ml sample the LOD was 0.1 ppt. It was necessary to adopt strict control measures to ensure a mercury-free environment around the instrument. Hanna et al. have devised a procedure (95/4493 95/C2957) which involved on-line microwave heating following the addition of sulfuric or hydrochloric and nitric acids potassium permanga- nate and potassium persulfate for the determination of Hg in waters and sediments. Excess oxidant was reduced by hydroxy- ammonium chloride solution prior to the generation of Hgo by tin@). The detection limit was 35 ppt. Duplicate analyses were carried out in less than 2 min.The method has also been described in the German literature (95/3777). Russian workers have reported a batch method (95/2889) in which nitric acid was used as the oxidant followed by addition of borohydride solution to generate the Hg'. The LOD was 1 ppt for a 100 ml sample. For the analysis of soils sediments and sludges McLeod's group at Sheffield Hallam University devised an FI procedure (96/300) in which a 400 pl slurry sample in nitric acid was digested on-line with the aid of microwave heating. After de-gassing and cooling Hgo was generated by the merging of a tin(I1) stream. Measurement by AFS gave a detection limit of 0.23 ppb. A comprehensive study of the analysis of reference soils sediments and sludges has been made by Landi and Fagioli (95/2235).Their method was based on an oxidative attack by potassium dichromate in sulfuric acid. They showed that the presence of chloride in a sample was a potential problem as low recoveries ascribed to the reoxidation of Hgo by chlorine were obtained. Chlorine could be removed by heating the reaction vessel with the air condenser removed but care was needed with samples containing high concentrations of organic matter or carbonate as the rapid release of carbon dioxide on the addition of sulfuric acid to the digestion vessel could cause a loss of Hg. The final method involved oxidation with the condenser in place followed by a period of heating with the condenser removed. The Hg contents of 9 SRMs were accurately determined by AAS with recirculation of the vapour. For the determination of Hg in urine French workers (95/2828) used a procedure which involved attack with a nitric-sulfuric acid mixture followed by oxidation with per- manganate.The excess permanganate was reduced with oxalate and Hgo produced by reduction with tin(r1). A linear AAS calibration over the range 10-1000 ppb for a 2 0 0 ~ 1 sample was obtained. Corns and co-workers (95/3406) described an AFS procedure in which oxidative attack was by bromine generated by the reaction between bromate and bromide in acid solution. The detection limit was 1 ppt and accurate analyses of two RMs were reported. A comparison between convective heating and microwave heating of samples (urine blood lyophilized tuna and pond sediment) in the presence of nitric acid was made (95/1179) by Romero's group at Zulia University in Venezuela.The mercury contents of the resulting digests were determined by AAS. Both methods gave accurate results and no significant differences between the results of the two methods were observed. The convective heating procedure required 12 h with two 4 h cooling intervals. The microwave heating stage required about 70 s. A collaboration between three clinical laboratories in Alberta Canada produced a method (95/935) for the analysis of urine blood breast milk faeces and monkey kidney cortex. Samples were mineralized with nitric-sulfuric acids in sealed vessels at 70 "C for 16 h cooled and bromine chloride solution added to convert organo- mercury to Hg". After 2 h a subsample was transferred to a batch reactor for generation of Hgo with tin@).The mercury was stripped with a stream of helium and trapped on gold- coated sand which was then sealed and removed and placed in a desorption apparatus. The trap was heated by conduction from a nichrome wire coil heated to 550 "C by electrical resistance heating and the thermally desorbed Hg retained on a second trap. The second trap was heated in a similar manner and the released Hgo finally measured by AFS. A detection limit of about 0.25 ppt was claimed. A somewhat similar procedure was used by Soater and Blattmann (95/2532) for the determination of total Hg in water and biological materials. Only a single gold trapping stage was used. The AFS LOD was 70 ppt for waters and 1 ppb (m/m) for biological materials.Bergdahl and coworkers (95/4165) used inductive heating of a gold wire trap in a procedure for the determination of inorganic Hg in whole blood and erythrocytes by AAS. Total mercury was determined when the samples were digested with a nitric- perchloric acid mixture whereas only inorganic Hg was released when the samples were attacked with sulfuric acid. Both digestions lasted 16 h (overnight). The difference between the two values was interpreted as the concentration of methyl- mercury and a strong correlation between this value and that obtained from a GC method for the erythrocytes was obtained. The detection limit in blood for total Hg was 60 ppt (m/m). Other Swedish workers (95/1949) who used the Magos reagent for the CVAAS determination of Hg in brain tissue found that about 5% of the methyl Hg was rapidly degraded to inorganic Hg.A correction procedure was proposed. In an attempt to speed up sample digestion Pacey and Sasaki have suggested (95/C767 95/C2893) the use of ozone as the oxidant. Results presented so far show that methyl Hg can be determined in the absence of a matrix down to 100 ppt in a flow injection procedure. It was found that passage of a tin@) solution through a glass capillary tube left tin(11) on the walls so that 100 determinations could be made with this open tubular reactor. Publication of this work is awaited with interest. For the determination of Hg in geological materials Kingston and McIntosh described a procedure (95/4494) in which samples were heated with aqua regia (2 min) and then permanganate (30 min).Excess oxidant was reduced with hyroxyammonium chloride and a sub-sample of the digest injected into an FI system. Water vapour was removed from the gaseous phase by passage though a single-strand Nafion dryer. The detection limit was 23 ppb in the solid. Terashima has described (95/2312) a procedure for the same sample materials in which the Hg was formed in the vapour phase by combustion in a quartz boat at 650°C for 3 min and the evolved gases were then passed through quartz sand and sodium carbonate. Samples which contained high concen- trations of volatile halogen or sulfur species were covered with magnesium oxide. The Hg vapour formed was trapped on a gold wire and determined by AAS after thermal desorption at Journal of Analytical Atomic Spectrometry August 1996 Vol.11 313R600 "C. The Hg contents of 118 geochemical reference materials were determined. Frech and co-workers have reported (95/4694) on the eficacy of noble metal collectors in procedures for the determination of Hg in natural gases. They found that collection by a gold- platinum wire was more efficient than collection by either gold or silver wire. Quantitative collection of both Hg and dimethyl- mercury was obtained with the gold-platinum collector at 80 "C and a sampling rate of less than 2 1 min-'. After collection the mercury was desorbed by furnace heating to between 790 and 850 "C and measured by AAS. An LOD of 30 ng mP3 was obtained for a 10 1 sample. Preliminary infor- mation (95/C742) about a commercial system for measuring air pollution by mercury either as elemental Hg or some inorganic species has been presented.There are numerous potential interferents in the CV .method when borohydride is used as the reductant including all elements that can either be reduced to the elemental form or form a volatile hydride. Two groups of Polish workers have reported on studies of these (95/1930 95/2847). In the first study (95/1930) the methods of standard additions and suc- cessive dilution were successful in overcoming these effects. It was noted that some interferents were retained in the equip- ment used and interfered with subsequent determinations. In the second study (95/2847) it was shown that the interferences from copper gold palladium platinum and silver could be overcome by the use of 1,lO-phenanthroline and thiosemicarba- zide in hydrochloric acid as masking agents.Tolerance to selenium was increased by prior oxidation to SeV1 whereas the interference from lead was decreased by digestion in an acid mixture containing sulfuric acid so as to precipitate lead sulfate. Cold vapour generation of Hgo has been used as a post- column reaction scheme for the detection of mercury species separated by reversed-phase HPLC separation on a C18 stationary phase (95/1944 95/1976). The procedure used in a method for the determination of the dimethyl and diethyl species (95/1944) was to decompose the dialkyls with hydrogen peroxide and UV irradiation and then to release Hg from the solution by the addition of borohydride. For the determination of methyl ethyl phenyl and inorganic Hg (95/1976) only UV irradiation was used followed by borohydride addition.In both cases detection was by AAS. Several other reports of the use of CV generation as a post column reaction scheme have appeared (95/1183,95/C2954 95/C3751). An indirect procedure for the determination of aldehydes based on the reduction of Hg" to Hg' by methanal ethanal or propanal has been proposed (95/3674). Russian workers (95/3880) have reported a differential AA method for the determination of Hg isotopes. Chinese workers claim (95/3871) to have liberated mercury hydride on the addition of a solution of potassium borohydride to an aqueous sample. It is more likely that the liberated species is Hg'. 1.3.3. Volatile organometallic compound generation and metal vapour separation An approach to the determination of mercury species in soils and sediments based on the temperature at which mercury was released from the sample has been described (95/2840).The sample was mixed with quartz powder and heated to 1000 "C in an iron capsule while the evolution of Hg was continuously monitored by CVAAS. It was possible to distinguish between metallic mercury mercury associated with organic matter and mercury associated with sulfide. Temperatures of 150,400 and 850 "C were used to differentiate between these types of Hg. In the analysis of samples the mercury evolved at each characteristic temperature was trapped on gold and sub- sequently released at 700 "C. A procedure for the determination of various mercury species in estuarine grasses has been described (95/1872 4363).The procedure is based on the evolution of derivatives with borohy- ride followed by trapping on a gas chromatography stationary phase cooled with liquid nitrogen. The trap was electrically heated to evolve the trapped species in order of increasing relative molecular mass the species being sequentially deter- mined by AAS with a heated quartz tube atomizer. The method was applied to the determination of Hg" monomethylmercury dimethylmercury and diethylmercury. The results indicated that dimethylmercury is more common in the aquatic environ- ment than is commonly supposed. A similar method for the determination of the monomethyl and dimethyl species has been described (95/4361). Derivatization with borohydride converted the monomethyl species to a volatile hydride and both methylmercurials were determined by GC with a purge and trap procedure.Both FTIR and AAS detectors were used. The method was applied to the thiosulfate extraction of fish and sediment. Derivatization of organomercury species with tetraethyl borate has been used by several groups (95/2532 95/369 95/937 95/4360) prior to the determination of the derivatized organomercury compounds by GC. Two groups (95/2532,95/937) used similar procedures involving potassium hydroxide and methanol for the extraction of the organomer- cury compounds from biological samples (hair monkey blood marine fish liver and muscle). The evolved ethyl species were trapped and thermally desorbed prior to GC separation and AFS detection.One group (95/937) found no dimethyl mercury in the biological materials. Reports of the determination of Pb by volatilization of the tetraethyl species have appeared from groups in Spain (95/2113) and the UK (95/C564,95/538). While both methods used a continuous flow procedure with AAS detection there are substantial differences between the results obtained. The Spanish group used potassium dichromate as oxidant in a lactic acid medium giving an LOD of 18 ppb where as the UK group used hydrogen peroxide in 0.55 mol 1-' sodium hydroxide solution to give an LOD of 0.07 ppb. The Spanish group used a de-tuned grid nebulizer as gas-liquid separator whereas the UK workers used a U-tube device. The UK workers used a flame heated quartz tube atomizer and it appeared as though the Spanish workers used a flame atomizer.The UK group claimed to have used rigorous optimization based on a simplex method whereas the Spanish group's approach to optimization was unclear. The results confirming the use of CV cadmium have been discussed earlier in the HG section 1.3.1.2. (determination of individual elements). 1.4 Spectrometers 'Building the perfect spectrometer' is the title of a review by Ciurczak (95/1038) which no doubt many people including the instrument designers will be anxious to read. However the article rather than presenting definitive instructions for the construction of such an ideal device summarizes the required components and deals with idealized modes of oper- ation. The description of a multi-tasking AAS by Schrader of Varian appears twice (95/2522 and 95/3483).The instrument described is for flame vapour and furnace-Zeeman AAS applications. It will run up to 999 samples for each of 20 methods in a fully automated sequence with complete random access sampling in each mode. Software allows users to review previous results develop the next method or process data while the instrument is still running samples. The study of molecular spectra using AAS instruments is referred to in section 1.1 of this review. An attachment to an AAS to convert it for 'spectrophotometry' is described by Russian workers (95/3429) who claim 4 times better sensitivity 314R Journal of Analytical Atomic Spectrometry August 1996 Vol. 11for direct nitrate measurements at 240 nm than with a standard UV spectrophotometer.Preliminary studies on a sequential multielement method for direct analysis of solid conducting samples by AAS are reported by Gillette (95/3819). Some of the disadvantages of XRF and OES are said to be avoided (particularly the need for matched standards) and undoubtedly more details of such a method will be awaited with considerable interest. A compact khelle monochromator with external order separa- tion was developed by Florek and Becker-Ross (95/1484) particularly for AAS with a continuum source. It overcomes the disadvantage of restriction in detector height giving lower stray light and less aberration. It is used in conjunction with a pre-monochromator featuring a variable-angle liquid reflec- tion prism.A resolving power of 140000 is obtained with a 23 pm slit. Fundamental theory of monochromators is dealt with in several papers. The influence of an intermediate slit on the performance of a double-pass parabolized Ebert monochroma- tor is considered in detail by Mattei and Gill (95/1236) and the scattered light properties of digraction gratings were studied by Woods et al. (95/1226). The latter found that replicas of gratings from the same master exhibit significantly more scattered light than a preceding replica by factors of 1.1-2.0. A patent has been lodged for astigmatism-corrected gratings for plane grating and spherical mirror spectrographs based on Czerny-Turner and Ebert-Fastie configurations (96/170). The grating is ruled directly on to an aspheric surface.Looking into the future Fulton and Horlick (95/C691) consider whether Acousto-Optic Tunable Filters (AOTFs) are yet good enough for use in atomic spectroscopy. AOTFs are electronically tunable light filters with a wavelength range of greater than 100 nm. There are no moving parts they are light in weight and can have an aperture up to 49 mm2. They are already in use in near-IR. They can be used to build simple and compact all-electronic rapid scanning monochromators and simple AA spectrometers. There is still evidently room for improvement in performance levels but the concept seems so ideal that perhaps this is the direction in which those seeking to build the perfect spectrometer should now be looking. 1.4.1 Light sources From the cradle of AAS (CSIRO in Australia) Hannaford (95/2214) has made an important contribution on the influence of oscillator strength on analytical sensitivity.In an ideal situation (atoms at rest in a collision-free environment) oscil- lator strength,f has little effect at the line centre. The peak absorption coefficient is dependent on f only through Doppler and collisional broadening. A summary of recent measurements off is included and the results are used to revise significantly the values of theoretical characteristic mass used hitherto in ‘absolute’ ETAAS. Dating from the same time and clearly with mutual knowl- edge of each others’ work Doidge of Varian (95/4717) has compiled a compendium and critical review of neutral atom resonance line oscillator strengths for use in ETAAS. The importance of the newly derived values which can vary by very significant factors from those hitherto in use is stressed asf-value data are needed to derive solar and stellar elemental abundances as well as for absolute ETAAS.The spatial distribution of radiant intensity from HCLs EDLs and D lamps used as light sources in AAS was investigated by Gilmutdinov Radzuik Sperling and Welz (95/C607). The distribution patterns for these three different types of source were found to be quite different and the effect of these differences on measured absorbance signals and on D lamp background correction was discussed. In a subsequent publi- cation (95/3879) the same workers described their work specific to HCLs in detail. The distribution pattern is strongly depen- dent on hollow cathode diameter though for a given lamp the pattern is similar for all currents applied. The manufacture of long-life multielement HCL sources for AAS is the subject of a Japanese patent (95/858).A mixture of alkali or alkaline earth halides with a ductile metal particle filler (Ag Al Cu or Mg) is compacted into a hollow cathode to form a layer with 10-55% porosity. The output intensity of an HCL may be increased by the addition of nitrogen to theJill gas according to Niemczyk et al. (95/2168). The intensity increase occurs without any line broadening but relative intensities can be dramatically changed within the spectrum of a given element. Care must be taken that the atomic lines to be used are isolated from the N bands. The effects of conditioning over time on the shape of a hollow cathode in a demountable spherical HCL were investi- gated by Yixin Chen et al.in Memphis USA (95/C705). Argon gas pressure and pulse width of the conditioning current have significant effects. The result of conditioning seems to be that the upper hemisphere of the cathode is etched away while the lower hemisphere is built up. Proper conditioning of the hollow cathode by sputtering is according to further work from the same laboratory (95/C707) critical to analytical performance. The hollow bottom of the cathode after conditioning was mirror-like smooth and shiny and gave rise to significant improvements in emission intensity and repeatability (96/343). Perhaps this is a point which other HCL manufacturers may wish to note. Shrader (95/C3023) described a new range of boosted HCLs based on a design originally conceived by Walsh and Sullivan (Sullivan J.V.Prog. Anal. At. Spectrosc. 1981 4 311). They feature a separate excitation circuit using a second anode and operation is as easy as for conventional HCLs. In other designs using a common anode for discharge and for excitation the lamp current has to be optimized by the operator using a peaking meter. This problem is now avoided and the new lamps can therefore be used in automatic systems. AAS with a flame emission source instead of an HCL was described by Calloway and Jones (95/2224). The source is obtained by aspirating a solution containing a high concen- tration of the analyte element an N,O-acetylene flame provid- ing best S/N ratios for most elements.Limits of detection were sometimes comparable with those given by traditional HCLs and linear dynamic ranges usually span 2-3 orders of magni- tude. However it has to be realised that the source bandwidth is up to twice that of the analyte. The system is inexpensive and can quickly be changed for different analytes. It may not be environmentally safe for very toxic elements and is not successful for elements which do not emit well in flames at good absorption wavelengths. From the same laboratory (95/C697 and 95/C2997) come the descriptions of simultaneous multi-element AAS using an ICP as light source. Two mixtures each of four elements were used to generate the source but actual results were not available in the conference abstracts to hand.As is to be expected diode lasers are being used more as AAS primary sources. This year we have seven reports on their use. However the first five of these have a common co-author in K. Niemax whose work in Dortmund Germany in collaboration with the Institute of Spectroscopy of the Russian Academy of Sciences is well known. Following last year’s report (J. Anal. At. Spectrom. 1995 10 221R) on the improvement in detection limits in ETAAS with a diode laser source the method has been extended to flames and to low pressure discharges. The apparatus described at that time has been modified to allow it to be used with laminar burners (95/2204). Cs was determined in an air-propane flame at 852.1 1 nm with a wavelength modulation frequency of 2 kHz and modulation amplitude of 18 GHz (for theory of wavelength Journal of Analytical Atomic Spectrometry August 1996 Vol.11 31 5 Rmodulation AAS see Silver J.A. Appl. Opt. 1992 31 707). Ti was measured best at 399.86 nm by frequency doubling in an LiIO crystal as was also Cr at 425.44nm. Because very low absorbance values can be measured using the diode laser the detection limits obtained in flame AAS are significantly improved and alternative less sensitive lines are more readily utilized. However lines involving excited states are more sensitive to flame parameters than those involving the ground state. Second harmonic generation of fundamental laser diode radiation frequencies can be used to increase the number of elements which can be measured and now more than 50 metals can be determined by this method.Further information was promised in a conference paper (96/C259). Extension of the method to the measurement ofelements in low pressure DCPs or MIPS provides a very sensitive detector for halogen-containing species in gas chromatography (95/4169 and 96/C258). Such species are completely dissociated and metastable halogen atoms are generated and excited. The improvement of analytical figures of merit in the coupling of GC and wavelength modulated-laser AAS (WM-LAAS) is the result of the additional modulation of the absorption by switching the plasma on and off with a frequency of a few kHz. This both increases the density of atoms in the metastable states and suppresses wavelength-dependent changes of the laser intensity measured with the detector such as Ctalon effects.Detection limits 3s for CHCl and CC1 were below 3 ng ml-I in the MIP and slightly higher in the DCP. For C1 the laser was tuned to 837.60 nm and for Br to 827.24 nm. No monochromator is necessary of course and the detector can be either a photomultiplier or a photodiode. A further paper on all of this work from Niemax’s group has been published in Russian (95/3678). In work from another German laboratory (96/C260) approaches with wavelength modulation using semiconductor diode lasers are compared with Zeeman magnetic field modu- lation and measurements without modulation. Flames and electrothermal atomization were investigated and again the signals were detected with photodiodes a monochromator not being required. Burakov et al.(96/C257) working in Minsk use an intracau- ity laser spectroscopy technique in combination with electro- thermal atomization to measure trace amounts of Li and Sr which have analytical lines within the visible region. The tunable flashlamp pumped dye laser produces broadband (10-15 nm) spectra with a pulse duration of 3-10 ys. The ETA system is located within the laser cavity. 1.4.2 Continuum source and simultaneous multi-element AAS After much pioneering work on continuum source AAS (CSAAS) Harnly (96/C216) confidently predicts that multi- element AAS (in which he believes the continuum source to be an essential ingredient) must ultimately succeed because the new instruments will provide better detection limits and more predictable and extended calibration than most single element instruments currently available on the commercial market.There are now at least six Cchelle spectrometers commercially available and these allow focusing of multiple wavelengths within compact dimensions. Furthermore solid state detectors either individually or in arrays are more compatible with Cchelle spectrometer geometry and also provide greater quan- tum efficiency. Harnly also evaluated the effects of graphite furnace atomization conditions on simultaneous multi-element AAS (95/C606). Sensitivities and SNRs of Cd Cr Cu Pb and V were characterized for a transversely heated atomizer. With open and end-capped tubes the minimum satisfactory atomiz- ation temperature for V was 2500 “C but this resulted in some diffusion loss of the other elements. Nevertheless at this temperature too the optimum SNR was experienced for all five elements.From the same laboratory work is reported using a seg- mented array CCD detector (95/C3027). This seems to have been in co-operation with Perkin-Elmer a P-E ‘Optima’ spectrometer utilizing a much smaller aperture than the CS-LPDA instrument was employed. The CCD offers shot- noise limitation and thus potentially improved detection limits. It also retains the advantage of high resolution. Meanwhile in a paper which came through too late for last year’s report Smith Nichol and Littlejohn (95/1122) demon- strated that CS-LPDA with an ichelle spectrometer gave a performance at least as good as that experienced with mechan- ical wavelength modulation and a photomultiplier.Good analytical results were obtained with certified reference materials too. The use of a photodiode array detector for simultaneous multi- element ETAAS was described by Tong and Chin (95/1020). HCLs for Co Cr Cu Fe Mn and Ni were used in the multichannel analyser system and sensitivities for Ni Co and Fe were in the same range as those expected with a conven- tional single element instrument. Background correction was by the two-line method. Fernando and Jones (95/103) using a recently set up system for CS-LPDA AAS reported limits of detection for a number of elements which are up to 2 x better than for previously reported CSAAS and close to the best for single element determinations. They used a 300 W xenon arc lamp and a 1.33 m focal length McPherson 209 monochromator. Multi- element determinations could be made for groups of metals having absorption lines in a single 10 nm spectral window.In a conference report (95/C666) the use of this equipment for the determination of Pb in the blood of house-painters is detailed. From the same laboratory came two more conference reports (95/C697 and 95/C2997 already cited in section 1.4.1) describ- ing a simultaneous multi-element AAS using an ICP as the radiation source. Light from the ICP dosed with a mixture of the elements to be determined is passed through a graphite furnace thence on to the slit of a small monochromator and detected with an LPDA. Figures of merit were promised but not anticipated in the abstracts. It is good to see that development of these systems is taking place in Europe as well as in the USA.Becker-Ross et al. (95/552) described the combination of an kchelle spectrometer and a linear array CCD with graphite furnace and xenon arc flashlamp for AAS measurements. The repetition rate is 10 Hz triggered by the atomization step of the graphite furnace. Although flashlamps offer a high intensity in the UV it can fluctuate strongly. This problem was overcome by simul- taneously recording the spectral vicinity of the absorption line for correction hence the need for a high resolution spec- trometer and a multi-pixel detector. T1 was measured for example with a characteristic mass of 28 pg and a detector limit in sea-water of 20 pg I-’. Radziuk et al. (95/1482) and Shuttler and Schulze (95/2537) all of Perkin-Elmer Germany now provide further details of their spectrometer system for simultaneous multi-element ETAAS. Up to four or six elements are determined.An kchelle optical system is used in conjunction with a custom-made solid-state detector. The transversely heated atomizer with integrated platform and longitudinally orientated Zeeman background correction increases luminosity by avoiding the need for a polarizer. Performance characteristics are quoted. The performance of the same equipment with respect to environmental samples (95/C2904) and complex matrices (95/C4230) was the subject of two conference papers. Jointly workers from Perkin-Elmer in Germany and Hamamatsu in Japan described a solid-state detector for 3 16 R Journal of Analytical Atomic Spectrometry August 1996 Vol.11simultaneous multi-element E TAAS with Zeeman background correction (95/305 1). Photodiodes giving a high quantum efficiency are married to a complementary metal oxide semi- conductor charge amplifier array which gives a low noise output. At the high intensities used in AAS measurements dark current is relatively unimportant (in contrast to ICP emission for example). This detector is evaluated and its performance relative to ETAAS is discussed. In another conference paper (95/C724) Shuttler et al. discuss whether recent developments in ETAAS instrumentation including the stabilized temperature platform in the transverse furnace Zeeman effect background correction availability of echelle monochromators and solid-state detectors make poss- ible simultaneous multi-analysis without compromising accu- racy precision or detection limits.It appears that identical conditions can provide satisfactory results for both volatile and non-volatile elements in the same sample matrix as compared with a conventional Massmann-type atomizer. Compromised conditions were quoted in the determination of trace elements in body fluid reference materials by simul- taneous ETAAS (95/C4320) as part of an interlaboratory trial. The problem of finding atomization conditions suitable for a range ofelements was also recognized by Dulude Moseley and Martin (95/C3028). Their equipment also measured up to four elements in a single electrothermal atomization and a ‘univer- sal’ modifier using combinations of Pd Mg and Ni was found to enable elements of differing volatilities to be measured simultaneously.Also from Germany Edel et al. (95/3857) described a frequency modulated simultaneous multi-element AAS using ETA and D background correction. Procedures for separating elements into groups having similar atomization conditions were given and results were said to be comparable with those obtained using conventional single channel AAS instruments. The same group (96/307) go on to apply the method to standard reference materials (dust pine needles aquatic moss and rye grass) with results ‘compatible with those obtained’ with a conventional commercial instrument. 1.4.3 Background correction After more than a quarter of a century of background correc- tion (BC) in AAS and more specifically a decade and a half of correction using the Zeeman effect one might have hoped that a best or even an optimum technique might by now have emerged. The plethora of papers on this subject and particu- larly on aspects of Zeeman correction suggests otherwise.Indeed newcomers to the scene may well wonder if Zeeman methods introduce more problems than they solve. A conference paper from Shrader of Varian (95/C645) reviewed commonly used BC techniques and compared the results of applying D lamp and Zeeman methods to real samples. Going (refreshingly) back to basics Thomas et al. (95/C702) presented a simple AAS with near-line BC for measuring Cd in biological and environmental samples (note that this method was recommended among last year’s reviews J. Anal. At. Spectrom.1995 10 223R for measuring Pb in blood). Where a suitable closely lying non-absorbing line is available this is clearly an attractive solution. Little is heard of the Smith-Hieftje method at the present time. However from China (95/66) came details of a power supply for hollow cathode lamps being used for this method. Selected wide and narrow pulse currents for 24 common elements are recommended. Several papers aim to compare or evaluate D lamp and Zeeman methods. Details of differences in method optimization wavelengths linear range final results and detection limits were promised by Delles (95/C643) in the measurement of Cd Cr and Pb in cementitious materials. The methods were also compared by Flajuik (95/3981) in a clinical study measuring Pb in blood. Commercial proof that there is as yet no ‘best’ method for BC surely came from Wassall of Unicam in two papers (95/883 and 95/980) showing how the two methods D lamp and Zeeman should be considered to be complementary.He described an AAS spectrometer in which either method is available simply by switching from one mode to the other. The system can run identical analyses automatically and serially within a single task selection of correction mode being made by the spectrometer software. It is claimed that this approach improves reliability and productivity but presumably a human decision has to be made at some stage on which technique is the better for each individual determination. In papers which specifically concern Zeeman correction a number of different aspects were brought up.In a patent application (95/1902) a switch and power supply to provide a periodic mugneticjield with a switching time of 1-1.5 ms was described. The applied voltage was 400-800 V. The debate about transverse and longitudinal effects was taken up in Bulgaria (95/3075) the pros and cons of both being discussed. Background over-correction problems in the measurement of lead in the presence of phosphate and various metals was investigated by Zong Parsons and Slavin (95/2220). Recoveries of less than 90% were found at a number of wavelengths with a longitudinal Zeeman system but only at 217.0nm and 239.4 nm with a transverse system. An explanation involving molecular absorption by PO species was invoked. Stray light can also cause problems in Zeeman BC and Voigtman et al.(95/2217) investigated these by modelling a transverse Zeeman AAS using a polychromatic optical calculus simulation software program. Calibration curve behaviour in the presence of three different types of stray light was simulated. It seems most likely that the major source of problems is light source polychromaticity even when the spectral bandwidth profile is relatively narrow. The work is continuing in the hope of extending both linear and dynamic ranges of ZAAS cali- bration curves beyond the roll-over region. Fifteen years after her doctoral thesis on zeeman atomic absorption spectroscopy (Delft University Press 1980) de Loos-Vollebregt is still investigating enhanced non-linearity of calibrations caused by the application of Zeeman BC.Following on work by L‘vov Slavin and others (e.g. B.V. L‘vov et al. Spectrochim. Acta Part B 1992,47 141 1) to introduce lineariz- ation techniques and to reduce the initial slope of the analytical curve in the 3-field ac Zeeman method de Loos-Vollebregt (95/C605) now believes that this approach offers only a 2-5 fold extension of the dynamic range. She went on to discuss the particular limitations for ETAAS with Zeeman BC. Also following on from work of L‘vov and others (Spectrochim. Acta Part B 1992 47 889 and 1992 47 1187) the original linearization algorithm has been exploited (95/1016) and later extended (95/2218) by Slavin and co-workers to reduce curvature at the upper end of Zeeman ETAAS calibrations. The assumption in the L‘vov treatment that the Zeeman sensitivity ratio should have a value of unity is not borne out and a method of successive approximation was used to derive a solution to L‘vov’s original theoretical expression. Both the original and extended algorithms were used for the linearization of calibration curves of Ag Bi Cu Mn and T1 and the derived characteristic masses for the two models were compared to test the validity of each algorithm.The original algorithm was found to be inadequate particularly for elements having Zeeman sensitivity ratios less than or equal to 0.5. From the same group of workers (95/C602) also came a conference paper on the eflect of spectral interferences on the linearization of calibration curves in Zeeman E TAAS. The new algorithm extended the linear dynamic range of calibration Journal of Analytical Atomic Spectrometry August 1996 Vol.11 31 7 Rcurves by a factor of 3-5 beyond roll-over for a number of elements. The behaviour of this algorithm in the presence of molecular spectral interferences was also discussed in detail. L'vov himself (95/C3736) discussed the results of his investi- gations on linearization and stabilization of Zeeman ETAAS calibration curves. He predicted that calibration procedures can be simplified and that characteristic mass ualues will be established with more certainty for most elements following a series of inter-laboratory studies. The latter theme is developed in two further papers from L'vov et al. (95/2216 and 95/4684) which described theoretical and experimental studies of the effect of photometric measuring error on detection limits in Zeeman ETAAS.In the second of these a simple method was proposed for calculating detection limits from integration time light flux energy and background absorbance. In a comparison of calculated and experimentally measured detection limits for a number of elements in pure water solutions and in blood saline and sea-water matrices good agreement was found. The possible fundamental advantages of either the linearized method or the three-field method over the conventional approach to ZAAS calibration were investigated by Harnly (95/2820). The methods were evaluated using a computer model. This assumed that either photon shot noise or the combination of this and analyte fluctuation noise was the limiting factor and that stray light was the only cause of deviation from linearity.The reduced sensitivity curve for three-field Zeeman gives a limited advantage in relative concentration error. Linearized Zeeman was shown effectively to extend the analytical range when integrated absorbance was used. In spite of the problems touched upon in the foregoing work computer controlled Zeeman scanning AAS has been mirror configuration was used by Bedenbender et al. (95/4436) to improve detection limits of a flame coherent forward- scattering (CFS) spectrometer. Designed with a xenon high pressure source for automated fast sequential multi-element analysis the results even with a continuum light source were said to be better than any previously reported for CFS. Comparison is made with FAAS.Photon-counting techniques in germanium avalanche diodes were described by Owens et al. (95/2231). To overcome ampli- fication noise limitations the diodes were biased beyond break- down and quenched with a simple series resistance. They then showed moderate quantum efficiencies and good dark count rates. Bristow et al. (96/171) presented the results of work to identify PMT divider networks and gating circuitry for use in the current measuring mode to limit variations in signal gain over a 25 ps gate period and improve linearity up to 20 mA to better than +0.1%. The work was done for differential absorption lidars but could be interesting to sequential multi- element AAS analysts. Charge transfer device detectors are playing an increasingly important role in analytical spectroscopy.In section 1.4.2 some examples have already been quoted. In particular a segmented array charge coupled detector (CCD) was used to limit shot noise and hence improve detection limits in multi-element AAS (95/C3027). Bonner Denton (95/C823) discussed the impact of CCDs and charge injection devices (CIDs) on various spectro- scopic techniques including rapid scan fluorescence and atomic spectroscopy. Different design and software approaches were considered. 1.5 Instrument Control and Data Processing From Chinese workers Liang Ni and Yang (95/4683) comes an authoritative review on the progress of software development is that while commercially available chemometric software packages can assist spectroscopists in the collection and interpretation of data they often fail to meet the more specialized needs of research workers.The latter therefore inevitably have to develop their own software to supplement available systems. Thus analytical scientists are becoming one of the most creative groups of conventional program users. The review the significant advances in conven- tional and intelligent software packages for AAS as well as some of the better commercially available packages. successfully employed in many fields of analysis. A less usual application requiring high precision is the measurement of The spectrometer was a modified direct ZAAS instrument with a low pressure graphite furnace. Complete isotope distributions of certified lead sulfide samples and recoveries of spiked natural samples were determined with an s of 1%.Thus the method is reasonably suited to some applications of tracers or isotope dilution methods although the precision is not adequate for dating geological specimens. The method would be applicable to other elements having isotope shifts of the order of the Doppler width or more. An intriguing paper (95/3623 unfortunately in Chinese) describes the correction of background interference by Kalman lead isotope distributions by Grimm and Hermann (95/213 )' for atomic spectroscopy during the past decade. Their verdict filtering. The work reported more specifically concerned ICP- OES but is said to be also applicable to spectrophotometry and to AAS. The background could be evaluated from mean square sums and normalized autocorrection coefficients lying outside the 95% confidence limit.This computational method is claimed to provide background correction without the need for background measurement. What a lot of time expense and argument that could save! 1.4.4 Detectors In a review of theoretical and practical limits in atomic spec- troscopy Winefordner et al. (94/2899) gave expressions for efficiency of detection and measurement. These were then combined with expressions for noise giving expected detection limits in AAS AES AFS etc. within an order of magnitude. These methods are then compared in respect of a number of analytical figures of merit. On a more practical level Batey Liezers and Tye (95/C578) described experiments to investigate why the plateau required when an electron multiplier is used in the pulse-counting mode is often ill-defined.They suggested how that mode of operation can be improved. Photon-counting detection in conjunction with a multi-pass 1.5.1 Instrument control The interface problems to be overcome in 'designing instrument control software that works' were treated in a general way in a paper reviewed last year (J. Anal. At. Spectrom. 1995 10 223R). The same title was used over a report (95/2543) on a user interface for a flow injection system for measuring Hg by AAS from Perkin-Elmer. The design principles and iterative development of the interface were illustrated. The implementation of instrument control and data analysis software for a fully automated unattended FAAS system for 'analysis of trace metals' (sic) is described by Lahiri Yuan and Stillman (95/1863).A fully automated analysis system is com- bined with an error-correction device AA-QC. The latter is able to identify from the absorption profile four of the main problems which can be encountered in FAAS a blocked burner a too-viscous solution a blocked capillary and the effects of a long capillary tube. AA-QC is in Microsoft VISUAL BASIC. File handling and the calculation of signal parameters are performed by FORTRAN routines called up from the VISUAL BASIC program. Almost every component of an AAS has been subjected to computer control source lamp gas supplies furnace controls 318R Journal of Analytical Atomic Spectrometry August 1996 Vol. 11spectrometer programming readout calculation and interpret- ation. The last bastion has now fallen. The measurable range in FAAS has been extended by software-controlled automation of burner rotation.Faugeron et al. (95/2356) described how this device can also simplify the analytical procedure and maintain accuracy in particular for their work done on determining Ca Cu and Fe. 1.5.2 Data processing The computer modelling work of Harnly on various approaches to Zeeman background correction (95/2820) was mentioned in section 1.4.3. Again using computer modelling Harnly (95/C603) has made an interesting comparison of calibration curves for Zeeman (normal linearized and 3-field approaches as before) self-reversed source and continuum source-array detection AAS. The same basic assumptions were made as in 95/2820. Continuum source-array detection seems to offer advantages at high concentrations in respect of both peak height and peak area absorbance. With PCs now in operation in most laboratories it is often necessary to transfer data between software programs on one PC and between PCs.‘Dynamic Data Exchange’ from Varian (95/C3021) is a system which makes this possible providing at the same time additional advantages such as applying alternative calibration algorithms to the original raw data. Data are presented on a spread-sheet in any required form and as replicate values are obtained they can be passed on to other software e.g. a commercial spread-sheet. An extensive range of QC protocols is provided and the system can be used to carry out further user-specific QC operations. Password access is required to ensure the security of data and parameters.Much work done on the calculation ofdetection limits from instrumental parameters has been cited in section 1.4.3. in connection with Zeeman background correction. A priori calculation of detection limits in ETAAS is described by Piso and de Boer (95/2823) using light source energy absorbance and concentration of the standard solution and the standard deviation of a blank step in the furnace program. Again reasonable agreement between these theoretical values and experimental values was claimed. Le Bihan et a/. (96/279) studied the eflect ofnoise on detection limits for some ETAAS determinations and showed that the detection limit can be calculated from the real experimental noise of a single atomization by use of an equation derived from a statistical approximation.The possibilities and limitations of internal standardization a technique hitherto used very extensively in both solid sample and ICP-AES methodology has been considered by Radziuk et al. (96/C205) in relation to ETAAS. This approach is made possible by the advent of simultaneous or quasi-simultaneous multi-element instrumentation. It could help to improve accu- racy when sample pipetting errors become significant and also when matrix effects influence the atomization process pro- vided of course that the internal standard element has been chosen to have similar volatility and atomization character- istics to the analyte elements themselves. A review in Hungarian (95/2869) presents an overview of spectroscopic analysis in that country.OES AAS and ‘hyphen- ated’ methods with hydride generation and cold vapour tech- niques are included. Hungary has always been a centre of excellence in analytical spectroscopy and it is a pity that the review is not available to a wider readership not least for its sections on computerized data handling and processing. 1.5.3 Chemometr ics The knowledge-based computer system from Penninckx et al. cited from a conference paper in last year’s review (J. Anal. At. Spectrom. 1995 10 224R) has now been published (95/1494). Matrix interferences it will be remembered are detected by variations between the slopes of aqueous standard and stan- dard addition calibration lines. The experimental design guarantees an acceptably low probability of non-detection of important interferences.A linearization is now proposed so that curved calibration lines can be compared. The Windows software is one component of a more complete Hypertext system. Somewhat pre-dating this work and unfortunately not able to have been reported last year is from the same group (95/931) the development of a system for the detection of interferences during routine analysis by ETAAS. Different methods for the characterization of the absorption signal in order to indicate the variations in position and shape caused by interferences were evaluated. These were then summarized in a multivariate test statistic Hotellings T2. Values of this parameter falling outside established limits show the presence of interferences. An inverse least squares multivariate calibration procedure is a key element in the method for speciation and simultaneous determination of As”’ AsV monomethylarsenate and dimethyl- arsinate described by Torralba et al.(95/2128). The influence of pH and reaction matrix on the continuous generation of the respective hydride forms with sodium tetrahydroborate followed by AAS detection and measurement was studied. In speciation by selective reduction univariate calibration does not usually work because of low selectivity between the species. The multivariate calibration procedure was therefore adopted a prediction matrix being used for simulated conditions in natural waters. The recovery of all four species was between 94-105% and detection limits between 0.8 and 4.9 ng m1-l. A major advantage is the speed of analysis and procedural simplicity leading to minimal risk of contamination.Chemometric experimental design is said to be the feature of a method for measuring Mg in chlorophyll-a by AAS (95/3100). Briefly it was shown that the simpler method of extraction with acetone-water mixtures gave the same result as an acid digestion metal recoveries being 99 and 98% respectively. Both methods agreed with Mg calculated from chlorophyll measured spectrophotometrically. The application of genetic algorithms (GAS) to the detection of outliers in non-linear models is discussed in some detail by Vankeerberghen et al. (95/4512). One of the examples used was a non-linear calibration in FAAS. Among the advantages of the GA approach are robustness towards local optima and no need for assumption about the geometry of the search- space; however configuration problems and premature conver- gence are disadvantages. 2.ATOMIC FLUORESCENCE SPECTROMETRY The analytical attraction of AFS lies in its high sensitivity and long linear range but matrix interference effects seriously limit its practical use; consequently developments in AFS are chiefly directed at capitalizing on the former and minimizing the latter. Compared with AES and AAS the range of applications of AFS is likely to remain limited for the forseeable future though there may continue to be some modest growth in its use in speciation studies and where the analyte element is readily separated from the sample matrix e.g. by hydride generation. 2.1. Discharge-excited Atomic Fluorescence Greenfield (95/1491) has discussed the use of the ICP as light source and atomizer for AFS.Earlier work demonstrated that there were fewer spectral interferences in AFS than in AES or even AAS. A study of the performance of a high power plasma as light source demonstrated lower LODs than the low power Journal of Analytical Atomic Spectrometry August 1996 Vol. 1 1 319Rsource but those were still not as low as were achieved using HCLs as sources. As these measurements were made using different instruments however it is difficult to draw firm conclusions from the data. At the present time there are no substantial advantages in ICP source ICP-AFS that are likely to lead to its widespread adoption as a general analytical technique.Hollow cathode lamps have been used in successful multi-element ICP-AFS. Ten elements were determined in soils (95/C2934). Analytical performance in AFS is greatly improved when the analyte element is separated from its matrix. It was proposed to couple the Baird AFS instrument incorporating a modified lamp pulse control system to a HG-LC system for the determination of volatile hydride generating elements in soil (95/C576). Zinc was determined in acid rain by ICP-AFS with HCL excitation following separation on an active carbon packed column (95/2246). The LOD was z 0.09 ng ml - The most common applications of AFS use flame-quartz tube atomization and discharge lamp-HCL excitation. The analyte elements are usually those which can be separated from the sample matrix by volatilization. Several instruments based on these principles have been described.A twin-channel instrument was presented at Pittcon '95 by Wang (95/C2953 95/C2955). The performance of a single channel instrument designed for the determination of As Cd Hg Sb and Se has been evaluated by Stockwell (95/3648). The analytical potential of a non-dispersive AF spectrometer with a Ta-coil atomizer for gas monitoring was investigated by Khvostikov et al. (95/1487). Iron pentacarbonyl in argon was used as the test system. The LOD was 4 ng I-' and the RSD 3.0%. Mercury appears to be the element most frequently deter- mined by AFS. The subject was reviewed in depth by Morita et al. (95/4443). In addition to presenting the principles and practice of the AFS determination of Hg information on CRM's was also recorded.Methods have been published for the determination of Hg in urine (95/3406) and freshwater ecosystems (96/387). In both cases the samples required chemi- cal pre-treatment and achieved LODs of z 1.0 ng m1-I. For the determination of Hg in sediments sample preparation was minimized by directly releasing Hg from the sample by heating. The Hg was then collected in a preconcentration trap prior to AFS determination (95/C709). The method was found to be 'simple reliable and accurate'. The determination of As in soil was simplified by suspending a soil homogenate in agar solution adding reagents and injecting the suspension into 1% KBH for generation of the hydride prior to detection by AFS (95/1126). In As speciation studies an ultrasonic nebulizer was used to provide the interface between an HPLC column and the AFS detection of system (95/4176).Detection limits for As"' DMAs MMAs and AsV were in the range 20-50ng. When it is not practical to release the analyte element from the interfering matrix it is sometimes possible to remove the principal interferent from the sample. This approach was adopted for the determination of Se and Te in copper ores (95/2370). In this analysis Cu was removed by cation exchange resin prior to AFS determination of the analyte. 2.2. Laser-excited Atomic Fluorescence Laser-excited atomic fluorescence has generated much more research interest than other means of exciting fluorescence. The attraction of laser sources lies in their high intensity and well dejined optical properties.The use of a 90" off-axis ellip- soidal mirror to replace lenses for collecting fluorescence radi- ation in an ETA-LEAFS system has been examined theoretically and experimentally (95/C832,96/248). The mirror system operated with high optical efficiency and very little aberration throughout the spectral range 200-700 nm. Background correction in ETA-LEAFS by wavelength modu- lation and the feasibility of absolute analysis have been the subjects of a dissertation by Su (95/1917). Pulsed lasers are commonly used for AF measurements hence in continuous flow atomizing systems much of the sample passes through the atom reservoir without exposure to the laser radiation. When the sample is small such inefficiency may be unaccept- able.The problem can be overcome if the sample is injected into the atom cell as a pulse in synchronism with the laser pulse. A monodisperse dried microparticulate injector capable of injecting single droplets into an ICP at frequency up to 6 kHz has been used in conjunction with a two-colour excimer- pumped dye laser for the ultra-trace analysis of small samples by LEAFS (96/C174). The system was studied with a view to optimizing the degree of droplet desolvation plasma operating conditions and viewing height. The importance of optimizing both the timing in multi-step excitation and the signal integration period has been demonstrated by Omenetto and Matveev (95/2209). These workers investigated laser-excited time- resolved waveforms to evaluate the essential parameters describ- ing the dynamics of the interactions between the radiation and the analyte atoms.The elements studied were As Hg Mg Na Pb Sr and T1. A double-pulse technique was developed by Cignoli et al. (95/4743) for the examination of fluorescence saturation. The first pulse was of low intensity in order that the atomic system could recover before the second pulse occurred. By this approach the population densities of the atomic levels involved were not disturbed and a theoretical model could be used to interpret the experimental results obtained for Pb in an air-C,H,Jame. One paper has reported a study of multi-element LEAFS (96/C252). It was based on a high power laser (Nd:YAG laser) pumping two dye lasers whose outputs were frequency doubled in non-linear optical crystals.Graphite furnace atomization and spectral filtering by monochromator were employed. The following wavelengths were used and LODs achieved Cd 228.2 nm 0.20 pg (10 ppt); Mn 279.5 nm 0.09 pg (4.5 ppt); Ni 232.0 nm 2.8 pg (140 ppt); Pb 283.3 nm 1.5 pg (75 ppt). Two elements could be determined simultaneously. Laser-induced fluorescence has been applied to gas and atmospheric analysis. The topic was reviewed by Omenetto and Panna (96/C230) with particular reference to the determination of pollutants (SO2 DMS NO,) radicals (OH HO,) and particulate metals (Au Pb). Detection limits of the order of 10 ppt in air were suggested. Krypton was detected in He at 40 ppt with an S/N of 500 (95/3887). In order to determine atmospheric temperatures and wind speed 100 km above the earth's surface using fluorescence of the Na D transition it was necessary to determine its Doppler-free fluorescence fea- tures for reference purposes (95/2260).From a simulated spectrum the frequencies of the dominant Doppler-free features were determined as -651.4 157.8 and 1068.0 MHz for D2a crossover and D2b resonances respectively relative to the weighted centre of the six D2 hyperfine transition lines. Axner and co-workers have published a series of papers based on a LEAFS system incorporating cup-tube graphite furnace atomization and a spectrometer fitted with an intensified charge-coupled device (ICCD) for radiation detection (95/2199). The advantage of the ICCD lies in its simultaneous detection over a wide spectral range. This feature facilitates the selection of the most sensitive excitation-detection wavelengths and provides for comprehensive background correction.There was a 50% degradation of the LOD for Ni in water compared with that achievable with photomultiplier detection (from 0.01 pg with PMT to 0.015 pg with ICCD detection). The results for the determination of Ni in river and estuarine waters agreed with the certified values. Aluminium and Pb in Arctic atmospheric aerosols were determined with the ICCD system and found to be in the range 1-Song rne3 (95/3106). Applications of the instrument have been extended to include Sb in biological and environmental samples (95/3729). An 320R Journal of Analytical Atomic Spectrometry August 1996 Vol. I ILOD of 5fg of Sb in pure water was achieved as well as good agreement between measured and certified values in thepgml-' tongml-1 range.The detection limit for T i in pure deionized water was higher at 1 pg (96/C246). The spectroscopic and analytical characteristics of double resonance laser inducedjuorescence of Au atoms in a graphite furnace and in a flame were studied by Petrucci et al. (95/4713). Gold was selected for study by virtue of its high ionization potential (9.23 eV) and its importance in geological medical and environmental sciences. Even after excitation with laser radiation at 267.6 nm and 406.5 nm the final fluorescent level lies more than 1 eV below the ionization continuum hence ionization loss by collision was minimal. The LODs were 3 fg (0.15 pg ml - ') and 8 pg (0.4 ng ml - ') for the furnace and flame respectively.Two-dimensional laser imaging with a gated charge coupled device (CCD) camera was used to investigate the efSect of a silver matrix on Au atomization in the graphite furnace (95/1483 95/C251). The laser was tuned to the 242.8nm Au line for imaging absorption by Au atoms and the non-specific background. The distribution of the Au atoms was observed by placing a Hg line interference filter centred at 313 nm in front of the CCD detector to select the Au fluorescence at 312.3 nm. The laser was pulsed and the CCD gated in synchronism to decrease the blackbody interference. When 100-2OOpg Ag were present the Au atom distribution was similar to that from a pure water matrix. However when z 1 mg of Ag was present Ag condensations were formed and the distribution of Au was greatly modified.The atomization of Ag in a graphite furnace was studied under low pressure by LEAFS (95/2200). It was concluded that the atomization process was both pressure and mass dependent. Under low pressure (1 mbar) and low analyte mass (0.1 ng) Ag atomizes as individual 'adatoms'. Under atmospheric conditions repeated adsorption and desorption processes delay the appearance of the signal in time as well as being responsible for the formation of 2-D and 3-D microstructures on the surface. The heating rate of the furnace was found to be important but was not fully investigated. Mercury was determined by ETA-LEAFS using two-step excitation with 253.7 nm and 435.8 nm radiation and detection of direct line fluorescence at 546.2nm (95/3398). The LOD was 14 fg (1.4 pg ml-') and the calibration graph was linear up to 90pg Hg.This same excitation and detection scheme has been studied in detail and laser-like collimated emission at 546.074 nm observed along the axis of the atom cell (95/1544). This radiation exhibits a pronounced non-linear dependence upon the number of Hg atoms in the cell and is attributed to amplified spontaneous emission (ASE) resulting from a transi- ent population inversion. Time-resolved observation of ASE and spontaneous fluorescence signal confirmed the different characteristics of the two emission processes. Dye lasers capable of resonance excitation of As (< 200 nm) are complicated and expensive. This problem has been over- come by using a fixed frequency ArF excimer laser which generates a broad-band output centred at 193.0-193.2 nm and overlaps the As line at 193.7nm (95/3078).An air-H flame and Ar-ICP were used as atomizers for the LEAFS determi- nation of As with the ArF excimer laser and both gave LODs of ~ 2 0 ng ml-'. Laser-induced scatter was the source of the limiting noise in both cases. A miniature glow discharge (GD) (internal volume 1.5 cm3) was evaluated as an atom reservoir for the LEAFS determi- nation of Eu Tu and Y (96/C262 96/344). Nanolitre-sized samples were deposited and dried on an Ni cathode and atomized by ion sputtering. The temporal profiles of the fluores- cence were found to consist of a transient spike ( ~ 2 0 ms) followed by a long tail (60 s). Signal area integration over 6 s gave LODs of 9 0.3 and 100 fg for Eu Tu and Y respectively.The application of LEAFS to the detection of elemental pollution of the environment was reviewed by Bolshov at the XXIX CSI Leipzig (96/C208). Arising from the high sensitivty of LEAFS direct analysis of environmental samples at and below the ppt level is possible and sample pre-treatment can be simplified or even eliminated. Bismuth was studied in some depth by Bolshov and co-workers (95/2201). A graphite cup ETA was employed and the lowest LOD (50 fg ml-' Bi) was achieved by combining excitation at 223.061 nm with detection at 299.334 nm. When this method was applied to the determi- nation of Bi in snow without pre-treatment strong matrix interference arising from molecular fluorescence of unidentified species was found (95/2202 95/3803). To overcome the inter- ference a two-step scheme of excitation was suggested. A tunable dye laser operating at 440-680 nm and pumped by an XeCl excimer laser was used to excite fluorescence of Cd and Pb in melted snow samples atomized from a graphite cup (95/2819).The lines at 228.8 and 405.8 nm were used for the detection of Cd and Pb respectively with corresponding LODs of 0.07 and 0.18 pg m1-I. Thallium in Antarctic snow was determined by LEAFS and by high resolution ICP-MS (95/1420). Levels of the order of tenths of a pg g-' were found with good agreement between the two methods. 3. LASER-BASED SPECTROSCOPY This is a new section within the ASU reviews and was created to bring together the uses of lasers in spectroscopy outside the LEAFS remit.It does not signal a significant increase in the activity in these fields but rather an attempt to clarify the presentation of this review. The principal subject areas are laser ablation and laser-enhanced ionization. 3.1. Laser Ablation and Excitation Laser ablation has been used to atomize all forms of sample solid liquid or gas prior to determination of the analyte element by AES AAS or AFS. Its particular advantages are vaporization of small samples with little or no sample prep- aration suitable for surface bulk and spot analysis and applicable to non-conducting as well as conducting materials. The advantages and limitations of coupled laser ablation-atomic spectrometry for the elemental analysis of solids have been reviewed by Mauchiev et al.(95/2877). The physical character- istics and instrumentation appropriate to the determination of alloy constituents were examined by Ahmad and Goddard (95/3800). Factors influencing elemental analysis by laser induced breakdown spectroscopy (LIBS) were studied in relation to the analysis of a NIST nickel alloy (SRM 1160) by AES (95/C3010). Laser power varied between 50-360 mJ pulse-' (107-109 W cmP2) at repetition rates from 1-1 5 Hz. The atmospheric compositions investigated included air N and Ar at pressures ranging from 740 Torr (atmospheric) to 2 x lov4 Torr. Laser induced non-resonance AF was used in conjunction with AE and AA to study the spatial and temporal distribution of atoms in a laser microprobe plume (95/2870). Microplumes were formed in Ar He and Ne atmos- pheres at pressures from 50 to 500Torr using stainless steel and Nb alloy samples.Iron Hf Ti and Zr were studied. The S/N for Fe was 3 times greater for Stokes direct line fluores- cence than for resonance detection. Atmospheric density strongly affected the interaction of the microplume with the atmosphere and inner filter effects were more severe in higher density gases. The dispersal of the microplasma was much faster in He than in Ar. Laser AA was employed to measure the density of N b atoms in a laser induced uapour plume (95/3815). The density near the plume centre was Nb atoms cm-3 and the electron excitation temperature in the range 3500-4500K. It was estimated that Nb atoms were ablated per laser pulse. Laser ablation was used to Journal of Analytical Atomic Spectrometry August 1996 Vol.11 321 Rgenerate fast atoms of Cu Zn and Ni (95/3072). Samples of pure metal and of compounds were irradiated with lOns 532 nm laser pulses in high vacuum. The metal atom intensities were 10 times greater when compounds rather than pure metals were irradiated. At laser powers below 109 W cm- the metal atom velocity distributions were shifted Maxwellian; above that power the distribution became distinctly non-Maxwellian. At the highest laser power densities very fast Rydberg atoms were observed and attributed to 3-body recombination colli- sions following plasma formation. Laser ablation LEAFS was applied to the determination of Co in pure tin (95/1058). The measurements in an atmosphere of Ar at atmospheric pressure achieved an LOD of 2 pg (0.2 pg g-') of Co.The calibration was linear from 1 pg g-' to 1 mg g-'. Laser ablation was used as the sampling procedure for the multi-element analysis of solids by ETA-coherent forward scattering (95/2219). The laser generated aerosol was deposited by corona discharge into a graphite tube furnace which was mounted between the poles of a magnet. Polarized radiation from an Xe continuum source passed through the furnace and into a spectrograph via an analyser prism. An optical multich- annel analyser provided simultaneous measurement of all spectral lines within a 50 nm range. The results for Ca Cr and Fe in incinerator waste agreed reasonably well with those obtained by AAS. Trace element determination in pure water by excimer laser ablation with LEAF detection has been demonstrated (95/3677 95/4160). A theoretical calculation predicted detection limits in the range 0.01-0.1 ppt. In a low pressure He atmosphere the LOD for Na was 1.7 ppt but at atmospheric pressure and in a He supersonic jet the detection limit was E 1 ppb.Trace levels of brominated compounds in air or N have been determined by laser induced photofragment- ation-atomization (95/2211). Laser radiation at 260.634 nm was used both to fragment the brominated compounds (CH3Br CHBr CHBr,Cl) and to excite the Br atom via a 2-photon transition. Detection was by resonance enhanced multiphoton ionization (REMPI) laser induced fluorescence (LIF at 844 nm) or stimulated emission (SE at 751 nm). The SE signal was coaxial with the laser beam and 100 times greater than the LIF signal.The most sensitive technique was REMPI with LODs of the order of ppb; those for LIF and SE were in the ppm range. 3.2. Laser Enhanced Ionization Laser enhanced ionization has raised the prospect of single atom detection. This concept has been considered theoretically for some time. Conference review papers by Omenetto (95/C2272) and Winefordner (95/C3018) have been devoted to the topic and have directed particular attention to the possibility of absolute and standardless analysis. To date there have been few practical analyses based on LEI. Severe problems arise when analysing samples containing high concentrations of easily ionized elements. A rod-jame system with two exciting lasers has been described which goes some way towards overcoming these problems (95/4735 96/C250).In this arrangement the sample liquid or solid is placed on an electrically heated graphite rod which is mounted above a slot burner. The advantage of this configuration over a simple burner was that the several components of the sample vapor- ized at different times. The system was applied to the analysis of NH4F and NaF. The detection limits for the analysis of solid samples were Co 1 ng g-' Cr 0.2 ng g-' Mn 0.3 ng g-' and Ni 0.08 ng g-'. The use of an F I system in place of direct nebulization to feed the sample into an air-C,H flame was found to extend the analytical range and reduce Na ionization interference in the determination of In (95/3644). Direct experimental evaluation of the ionization rates of Nu induced by laser radiation in an air-C,H flame has been reported (95/2123).The method is based on observing the rise time of the current pulse induced on the electrode by the negative charge created in the ionization process. The 'effective' ionization rate of the 4 d levels reached in 2-step excitation via the 3p level was found to vary from 0.27 to 3.4GHz. Laser enhanced ionization of Na in an air-H flame was used to replace the plasma source of ions in an ICP-MS system (95/2210). This arrangement overcame the problem of the non- selectivity of the current measurement in LEI and demon- strated the feasibility of MS detection of an LEI signal. LOCATION OF REFERENCES The full list of references cited in this lJpdate have been published as follows 9412868-9412994 J.Anal. At. Spectrom. 1994 9( lo) 307R-312R. 95/1-951182 J. Anal. At. Spectrom. 1995 10( l ) 1R-7R. 951183-951469 J. Anal. At. Spectrom. 1995,10( 2) 49R-59R. 95/470-95/C849 J. Anal. At. Spctrom. 1995 10(4) 113R-125R. 951850-9511159 J. Anal. At. Spectrom. 1995 10( 5 ) 127R-138R. 95/1160-95/1597 J. Anal. At. Spectrom. 1995 10(6) 155R-171R. 95/1598-95/C2275 J. Anal. At. S,oectrom. 1995 10( 7) 173R-198R. 9512276-9512891 J. Anal. At. Spectrom. 1995 10( S) 229R-251R. 95/C2892-95/3361 J. Anal. At. Spectrorn. 1995 10( lo) 31 1R-328R. 9513362-9514189 J. Anal. At. Spectrom. 1995 10( 1 l) 329R-358R. 95/C4190-95/4746 J. Anal. At. Spectrom. 1995 10( 12) 402R-422R. 9611-96/416 J. Anal. At. Sptxtrom. 1996 11( l ) 1R-17R. Abbreviated forms of the literature references quoted (excluding those to Conference Proceedings) are given on the following pages for the convenience of the readers.The full references names and addresses of the authors and details of the Conference presentations can be found in the appropriate issues of JAAS cited above. 322 R Journal of Analytical Atomic Spectrometry August 1996 VoZ. 1 195/51 Clin. Chem. ( Winston-Salem N.C.) 1993 39 1907. 95/66 Fenxi Shiyanshi 1994,13( l ) 90.95/100 J. Anal. Toxicol. 1994 18 36. 951103 J. Electrochem. SOC. 1994 141 718. 951108 J. Korean Chem. SOC. 1994,38,214.95/235 Guangpuxue Yu Guangpu Fenxi 1993,13( 5 ) 85.951281 Xiyou Jinshu Cailiao Yu Gongcheng 1993 22( l) 63. 951305 J. Anal. At. Spectrom. 1994,9 1167. 951334 Anal. Chim. Acta 1994 294 185. 951339 Anal.Chim. Acta 1994 288 17. 951369 Clin. Chem. (Washington D.C.) 1994 40 602. 951390 God. Zb. Zemjod. Fak. Univ. 0 Kiril Metodij ’-;Skopje 1993 38 59. 951401 Hebei Daxue Xuebao Ziran Kexueban 1992,12(2) 60. 951537 J. Anal. At. Spectrom. 1994 9 1411. 951538 J. Anal. At. Spectrom. 1994 9 1417. 951539 J. Anal. At. Spectrom. 1994 9 1423. 951540 J. Anal. At. Spectrom. 1994 9 1427. 951541 J. Anal. At. Spectrom. 1994 9 1433. 951548 J. Anal. At. Spectrom. 1995 10 37. 951549 J. Anal. At. Spectrom. 1995 10 43. 951550 J. Anal. At. Spectrom. 1995 10 49. 951551 J. Anal. At. Spectrom. 1995 10 55. 951552 J. Anal. At. Spectrom. 1995 10 61. 951858 Jpn. Kokai Tokkyo Koho JP 05,275,061 [93,275,061] (Cl. H01J61/09) 22 Oct 1993 Appl. 92/65,511 24 Mar 1992; 6 pp. 951883 Am. Lab.(Fairjeld Conn.) 1994,26,48JJJ-48LLL. 951889 Anal. Chim. Acta 1994 288 247. 951892 Anal. Sci. 1994 10 59. 951899 Analyst (London) 1994 119 1033. 951900 Analyst (London) 1994 119,1119.95/916 At. Spectrosc. 1994,15,126.95/931 Chemom. Intell. Lab. Syst. 1994 23 137. 951935 Clin. Chem. (Winston- Salem N.C.) 1994 40 206. 951936 Clin. Chem. (Winston- Salem N.C.) 1994 40 431. 951937 Clin. Chem. (Winston- Salem N.C.) 1994 40 602. 951943 Fenxi Huaxue 1994 22 423.951945 Fenxi Huaxue 1994,22,512.95/946 Fenxi Huaxue 1994 22 523. 951948 Fenxi Shiyanshi 1994 13(2) 72. 951952 Fresenius’ J. Anal. Chem. 1994 348 41 1. 951956 Fresenius’ J. Anal. Chem. 1994 349 306. 951957 Fuel 1994 73 629. 951966 Guangpuxue Yu Guangpu Fenxi 1993 13(5) 107. 951976 Guangpuxue Yu Guangpu Fenxi 1994,14( l) 81.951980 Int.Lab. 1994 24(1) 10. 951982 J. Agric. Food Chem. 1993 41,2024.95/1012 Scr. Fac. Sci. Nut. Univ. Masarykianae Brun. 1993 (23) 27. 9511016 Spectrochim. Acta Part B 1994 49 367. 9511017 Spectrochim. Acta Part B 1994,49 399. 9511019 Spectrochim. Acta Part B 1994 49,445.9511020 Spectrochim. Acta Part B 1994 49 459. 9511024 Spectrochim. Acta Part B 1994 49 537. 9511028 Spectrochim. Acta Part B 1994 49 583. 9511031 Spectrochim. Acta Part B 1994 49 615. 9511032 Spectrochim. Acta Part B 1994,49,627. 9511038 Spectroscopy (Eugene Oreg.) 1994 9(2) 12 14. 9511039 Spectrosc. Lett. 1993,26,1613.95/1041 Spectrosc. Lett. 1993,26,2017.95/1042 Spectrosc. Lett. 1994 27 817. 9511043 Talanta 1994 41 495. 9511058 Zh. Prikl. Spektrosk. 1993 59 135. 9511079 Bunseki Kagaku 1994 43 39.9511087 Bunseki Kagaku 1994 43 85. 9511112 Talanta 1993,40 1917. 9511116 Gongye Weisheng Yu Zhiyebing 1992 18 364. 9511122 J. Anal. At. Spectrom. 1993 8(7) 987. 9511125 Fresenius’ J. Anal. Chem. 1993 346 1072. 9511126 Shanghai Huanjing Kexue 1992 011(5) 22. 9511131 Fresenius’ J. Anal. Chem. 1993 346 1054. 9511153 J. Enuiron. Qual. 1993,22 851.9511170 Latv. Zinat. Akad. Vestis B 1992 12 57. 9511171 Heat Transfer Radiat. Combust. Syst. Proc. EUROTHERM Semin. 17th 1991 332. 9511175 Appl. Geochem. 1993 8 265. 9511179 J. Anal. At. Spectrom. 1993 8 1003. 9511183 Fresenius’ J. Anal. Chem. 1993 346 648. 9511188 Fenxi Shiyanshi 1992 11(6) 5. 9511213 Indian J. Chem. Sect. A 1994 33 86. 9511217 Ann. Chim. (Rome) 1993 83 397. 9511219 Appl. Organomet. Chem. 1993 7 593.9511226 Appl. Opt. 1994 33 4273. 9511227 Anal. Chim. Acta 1994 295 187. 9511228 Anal. Chim. Acta 1994 295 199. 9511236 Appl. Opt. 1994 33 5187. 9511237 Appl. Opt. 1994 33 6514. 9511238 Appl. Opt. 1994 33 6068. 9511420 Int. J. Enuiron. Anal. Chem. 1994 55 211. 9511482 J. Anal. At. Spectrom. 1995 10 127. 9511483 J. Anal. At. Spectrom. 1995 10 137. 9511484 J. Anal. At. Spectrom. 1995 10 145. 9511485 J. Anal. At. Spectrom. 1995 10 149. 9511486 J. Anal. At. Spectrom. 1995 10 155. 9511487 J. Anal. At. Spectrom. 1995 10 161. 9511491 J. Anal. At. Spectrom. 1995 10 181. 9511492 J. Anal. At. Spectrom. 1995 10 185. 9511493 J. Anal. At. Spectrom. 1995 10 195. 9511494 J. Anal. At. Spectrom. 1995 10 205. 9511496 J. Anal. At. Spectrom. 1995 10 219. 9511499 J. Anal. At.Spectrom. 1995 10 239. 9511500 J. Anal. At. Spectrom. 1995 10 245. 9511506 J. Anal. At. Spectrom. 1995 10 285. 9511544 Appl. Phys. B Lasers Opt. 1994 58 303. 9511672 J. Anal. At. Spectrom. 1995 10 317. 9511673 J. Anal. At. Spectrom. 1995 10 321. 9511674 J. Anal. At. Spectrom. 1995 10 325. 9511675 J. Anal. At. Spectrom. 1995 10 329. 9511814 Proc. SPIE-Int. SOC. Opt. Eng. 1994 2011 108. 9511838 Anal. Chim. Acta 1994,296 181. 9511839 Anal. Chim. Acta 1994 296 195. 9511840 Anal. Chim. Acta 1994 296 205. 9511846 Anal. Chim. Acta 1994 297 231. 9511861 An. Asoc. Quim. Argent. 1994,82 1.9511863 Anal. Chem. 1994,66 2954. 9511869 Anal. Chim. Acta 1994 291 89. 9511872 Anal. Chim. Acta 1994,292 175. 9511877 Anal. Chim. Acta 1994 292 325. 9511880 Anal. Sci. 1994 10 481. 9511882 Analusis 1994 22 257.9511884 Analyst (London) 1994,119,627.95/1889 Analyst (London) 1994 119 1403.9511891 Analyst (London) 1994 119 1459.9511894 Ann. Chim. (Rome) 1994,84,275.95/1898 Appl. Spectrosc. 1994 48 713. 9511902 Ger. Offen. DE 4,335,592 (Cl. GOlN21/71) 11 May 1994 AU Appl. 92/5,666 05 Nov 1992; 1Opp. 9511907 Ger. Offen. DE 4,243,766 (Cl. GOlN21/74) 30 Jun 1994 Appl. 23 Dec 1992; 12 pp. 9511908 Eur. Pat. Appl. EP 603,814 (Cl. GOlN21/74) 29 Jun 1994 DE Appl. 4,243,767 23 Dec 1992; 1Opp. 9511909 Ger. Offen. DE 4,240,934 (Cl. GOlN21/74) 09 Jun 1994 Appl. 04 Dec 1992; 7pp. 9511917 Diss. Abstr. Int. B 1994 54 5123. 9511918 Beijing Keji Daxue Xuebao 1993,15(4) 429.9511923 Bol. SOC. Quim. Peru 1993 59 247. 9511924 Bol. SOC. Quim. Peru 1993 59 253. 9511925 Bol.SOC. Quim. Peru 1993 59 259. 9511930 Chem. Anal. (Warsaw) 1994 39 205. 9511940 Fenxi Huaxue 1994,22,586.95/1941 Fenxi Shiyanshi 1993,12(4) 19.9511944 Fresenius’ J. Anal. Chem. 1994 348 253. 9511949 Fresenius’ J. Anal. Chem. 1994 348 815. 9511951 Fresenius’ J. Anal. Chem. 1994,348,839.9511957 Guangpuxue Yu Guangpu Fenxi 1993,13(6) 87.9511963 Guangpuxue Yu Guangpu Fenxi 1994 14(2) 97. 9511976 J. Chromatogr. A 1994 675 253. 9511978 J. Enuiron. Sci. Health Part A 1994 A29 867. 9512110 Anal. Chim. Acta 1994 297 443. 9512111 Anal. Chim. Acta 1994 298 167. 9512113 Quim. Anal. (Barcelona) 1994 13 67. 9512114 Quim. Anal. (Barcelona) 1994 13 15. 9512115 Quim. Anal. (Barcelona) 1994 13 73. 9512117 Quim. Anal. (Barcelona) 1994 13 138. 9512119 Spectrochim. Acta Part B 1994 49 577.9512121 Spectrochim. Acta Part B 1994 49 665.9512123 Spectrochim. Acta Part B 1994,49,691.95/2127 Spectrochim. Acta Part B 1994,49 875. 9512128 Spectrochim. Acta Part B 1994 49 893. 9512131 Spectrochim. Acta Part B 1994 49 925. 9512150 Spectrochim. Acta Part B 1994 49 1067.9512156 Spectrochim. Acta Part B 1994,49,947.95/2168 Appl. Spectrosc. 1994,48 896.9512199 Spectrochim. Acta Part B 1994,49 1385.9512200 Spectrochim. Acta Part B 1994,49 1411. 9512201 Spectrochim. Acta Part B 1994 49 1437. 9512202 Spectrochim. Acta Part B 1994 49 1445. 9512204 Spectrochim. Acta Part B 1994,49,1463.95/2209 Spectrochim. Acta Part B 1994 49 1519. 9512210 Spectrochim. Acta Part B 1994,49 1537.9512211 Spectrochim. Acta Part B 1994,49 1545. 9512213 Spectrochim. Acta Part B 1994 49 1569.9512214 Spectrochim. Acta Part B 1994 49 1581. 9512215 Spectrochim. Acta Part B 1994,49 1595.9512216 Spectrochim. Acta Part B 1994 49 1609. 9512217 Spectrochim. Acta Part B 1994,49 1629.9512218 Spectrochim. Acta Part B 1994 49 1643. 9512220 Spectrochim. Acta Part B 1994 49 1667. 9512221 Spectrochim. Acta Part B 1994 49 1681. 9512222 Spectrochim. Acta Part B 1994,49 1689.9512223 Spectrochim. Acta Part B 1994 49 1695. 9512224 Spectrochim. Acta Part B 1994,49 1707.9512225 Spectrochim. Acta Part B 1994,49 1717. 9512226 Spectrochim. Acta Part B 1994 49 1733. Abbreviated List of References Cited in Update Journal of Analytical Atomic Spectrometry August 1996 Vol. 11 323 R9512231 Appl. Opt. 1994 33 6895. 9512232 Appl. Opt. 1994 33 7777. 9512235 Anal.Chim. Acta 1994 298 363. 9512242 Can. J . Appl. Spectrosc. 1994,39 117.9512246 Fenxi Shiyanshi 1995,14,50.95/2248 Fenxi Ceshi Xuebao 1995,14,68.95/2251 Fenxi Ceshi Xuebao 1995,14,60.95/2253 Fenxi Ceshi Xuebao 1995 14 72. 9512257 Appl. Opt. 1995,34 1117. 9512258 Appl. Opt. 1995,34,1108.95/2259 Appl. Opt. 1995,34,1103.95/2261 Appl. Opt. 1995,34,973. 9512266 Anal. Chem. 1994,66,3247. 9512267 Anal. Chem. 1994 66 3533. 9512303 Analyst (Cambridge UK) 1995 120 395. 9512310 Jilin Daxue Ziran Kexue Xuebao 1994 1 119. 9512312 Kinzoku 1994 64 4. 9512317 Magy. Kem. Foly. 1994,100 119.9512328 Microchem. J. 1994 49 368. 9512331 Mikrochim. Acta 1994 113 325. 9512350 Rev. Chim. (Bucharest) 1994,45,135.95/2356 Spectra Anal. 1993 22 33. 9512359 Spectrochim. Acta Part B 1994 49 387.9512368 Yankuang Ceshi 1993 12 241. 9512370 Yankuang Ceshi 1993 12 287. 9512377 Zh. Anal. Khim. 1994 49 477. 9512388 Zhejiang Daxue Xeubao Ziran Kexueban 1993 27 542. 9512390 J. Anal. At. Spectrom. 1995 10 343. 9512392 J. Anal. At. Spectrom. 1995 10 353. 9512393 J. Anal. At. Spectrom. 1995 10 359. 9512401 J. Anal. At. Spectrom. 1995 10 409. 9512499 Guangpuxue Yu Guangpu Fenxi 1994 14(2) 113. 9512501 Fresenius’ J. Anal. Chem. 1994 349 527. 9512502 Fresenius’ J. Anal. Chem. 1994 349 530. 9512508 Fresenius’ J. Anal. Chem. 1994 349 442. 9512514 Lihua Jianyan Huaxue Fence 1994,30 144. 9512522 Int. Lab. 1994 24 10. 9512528 Food Chem. 1994 50 317. 9512532 Anal. Chem. 1994 66 2031. 9512534 Int. Lab. 1994 24(6) 27. 9512537 Analysis Europa 1994 1 44. 9512543 At. Spectrosc.1994 15 156. 9512544 At. Spectrosc. 1994 15 151. 9512558 Anal. Biochem. 1994,220 336.9512623 Anal. Chim. Acta 1994 291 115. 9512668 Biomed. Res. Trace Elem. 1993 4 51. 9512684 Chem. Res. Chin. Univ. 1993 9 256. 9512736 J. Trace Elem. Electrolytes Health Dis. 1993 7 21 1. 9512741 Lab. Rob. Autom. 1993 5 277. 9512769 Quim. Anal. (Barcelona) 1993 12 30. 9512799 Zh. Anal. Khim. 1994 49 639. 9512819 Analusis 1994 22 M44. 9512820 Appl. Spectrosc. 1994 48 1156. 9512822 At. Spectrosc. 1994 15 229. 9512823 At. Spectrosc. 1994 15 220. 9512824 At. Spectrosc. 1994 15 245. 95/2828 At. Spectrosc. 1994 15 250. 9512834 Fenxi Huaxue 1994 22 973. 9512837 Fenxi Huaxue 1994 22 971. 9512839 Fresenius’ J. Anal. Chem. 1994 349 779. 9512840 Fresenius’ J. Anal. Chem. 1994 350 18. 9512842 Fresenius’ J.Anal. Chem. 1994 350 44. 9512847 Fresenius’ J. Anal. Chem. 1994 350 178. 9512849 Fresenius’ J. Anal. Chem. 1994 350 359. 9512861 Lihua Jianyan Huaxue Fence 1994 30 270. 9512862 Lihua Jianyan Huaxue Fence 1994 30 278. 9512863 Lihua Jianyan Huaxue Fence 1994,30 268. 9512868 Lihua Jianyan Huaxue Fence 1994,30 244. 9512869 Magy. Kem. Foly. 1994 100,327.95/2870 Microchem. J. 1994,50,253.95/2877 Spectra 2000 [Deux Mille] 1994 180 33. 9512884 Talanta 1994 41 1785. 9512889 Zavod. Lab. 1994 60(8) 26. 9513051 J. Anal. At. Spectrom. 1995,10,415.95/3066 Appl. Opt. 1995,34 1606. 9513072 J. Phys. Chem. 1994 98 11931. 9513075 Anal. Lab. 1993 2 127. 9513078 Appl. Spectrosc. 1994 48 1213. 9513091 Analyst (Cambridge UK) 1995 120 95. 9513095 Analyst (Cambridge UK) 1995,120,101.95/3096 Analyst (Cambridge UK) 1995 120 89.9513099 Analyst (Cambridge UK) 1995 120 79. 9513102 Analyst (Cambridge UK) 1995 120 403. 9513106 Analyst (Cambridge UK) 1995 120 635. 9513369 Spectroscopy (Eugene Oreg.) 1995,10,26.95/3371 Mikrochim. Acta 1994 117 49. 9513372 Fresenius’ J. Anal. Chem. 1994 350 651. 9513374 Fresenius’ J. Anal. Chem. 1995 351 139. 9513375 Zh. Anal. Khim. 1994 49 755. 9513376 Bunseki Kagaku 1994 43 1009. 9513379 Bunseki Kagaku 1994 43 1149. 9513381 Bunseki Kagaku 1994 43 1197. 9513382 Fenxi Huaxue 1994 22 1135. 9513384 Fenxi Shiyanshi 1994 13 52. 9513385 Lihua Jianyan Huaxue Fence 1994 30 329. 9513386 Lihua Jianyan Huaxue Fence 1994 30 341. 9513392 Guangpuxue Yu Guangpu Fenxi 1994 14(4) 79. 9513394 Guangpuxue Yu Guangpu Fenxi 1994 14(4) 97.9513395 Guangpuxue Yu Guangpu Fenxi 1994 14(4) 101. 9513398 Talanta 1994 41 2073. 9513401 Talanta 1994 41 2165. 9513406 Analyst (London) 1994 119 2481. 9513413 Anal. Chem. 1994 66 3900. 9513417 TrAC Trends Anal. Chem. (Pers. Ed.) 1995 14 76. 9513429 Zavod. Lab. 1994 60(12) 26.9513436 Fenxi Huaxue 1994,22 1289.9513442 Microchem. J. 1995 51 39. 9513451 Talanta 1995 42 211. 9513452 Talanta 1995 42 263. 9513455 Analyst (London) 1994 119 2785. 9513459 Anal. Chem. 1994 66 4375. 9513462 Ciencia 1994,2 87. 9513467 Spectrochim. Acta Part B 1994 49 1211. 9513482 Anal. Sci. 1994 10 779. 9513483 Am. Lab. (Fairjield Conn.) 1994,26(11) 24C. 9513487 Zh. Anal. Khim. 1994 49 825. 9513494 Fenxi Huaxue 1994 22 1037. 9513497 Guangpuxue Yu Guangpu Fenxi 1994 14(3) 87. 9513499 Guangpuxue Yu Guangpu Fenxi 1994 14(3) 101.9513501 Yankuang Ceshi 1994 13(2) 96. 9513503 Yankuang Ceshi 1994 13(2) 125. 9513508 Fenxi Huaxue 1994 22 1002. 9513510 Bunseki Kagaku 1994 43 971. 9513511 Anal. Chim. Acta 1994 296 115. 9513512 Academic Press Inc. San Diego CA USA 1995. 0 12 226700 1. 6208pp. 9513513 JAI Press Ltd. London UK 1995. 1 55938 701 7. 300 pp. 9513514 John Wiley and Sons Ltd. Chichester W. Sussex UK 1995. 0 471 95364 4. 544 pp. 9513516 J. Anal. At. Spectrom. 1995 10 479. 9513517 J. Anal. At. Spectrom. 1995 10 483. 9513519 J. Anal. At. Spectrom. 1995 10 493. 9513528 Anal. Chim. Acta 1995 304 121. 9513534 Analyst (Cambridge UK) 1995 120 629. 9513595 J. Trace Microprobe Tech. 1994 12 323. 9513603 Fresenius’ J. Anal. Chem. 1995 351 393. 9513616 Fenxi Shiyanshi 1995 14 59.9513618 Fenxi Shiyanshi 1995 14 11. 9513620 Guangpuxue Yu Guangpu Fenxi 1995,15,89.95/3621 Guangpuxue Yu Guangpu Fenxi 1995 15 97. 9513623 Guangpuxue Yu Guangpu Fenxi 1995 15 61. 9513626 Guangpuxue Yu Guangpu Fenxi 1995,15,69.95/3627 Gaodeng Xuexiao Huaxue Xuebao 1995,16,536.95/3629 Fenxi Huaxue 1995 23 180. 9513630 Huaxue Xuebao 1995 53 57. 9513631 Huaxue Xuebao 1995 53 68. 9513632 Fenxi Huaxue 1995 23 284. 9513635 Fenxi Ceshi Xuebao 1995 14 66. 9513637 Lihua Jianyan Huaxue Fence 1995 31 75. 9513638 Lihua Jianyan Huaxue Fence 1995,31,21.95/3639 Faming Zhuanli Shenqing Gongkai Shuomingshu CN 1,076,278 (Cl. G01N21/31) 15 Sep 1993 Appl. 92,101,560 09 Mar 1992; 11 pp. 9513644 Anal. Chem. 1994 66 2180. 9513645 Anal. Sci. 1994 10 623.9513648 Analyst (London) 1994 119 1641. 9513649 Analyst (London) 1994 119 1647. 9513654 Analusis 1994 22 350. 9513657 Fenxi Ceshi Xuebao 1994 13 44. 9513659 Fresenius’ J. Anal. Chem. 1994 348 418. 9513660 Fresenius’ J. Anal. Chem. 1994,348,422.9513667 Guangpuxue Yu Guangpu Fenxi 1994 14(3) 79. 9513668 Guangpuxue Yu Guangpu Fenxi 1994 14(3) 121. 9513674 Mikrochim. Acta 1994 116 57. 9513675 Nanjing Huagong Xueyuan Xuebao 1994,16,53.95/3676 Nanjing Huagong Xueyuan Xuebao 1994 16 84. 9513677 Opt. Commun. 1994 110 298. 9513678 Opt. Spektrosk. 1994 76 502. 9513681 Talanta 1994 41 565. 9513682 Talanta 1994 41 1185. 9513686 Yankuang Ceshi 1994 13 113. 9513692 Analyst (Cambridge UK) 1995 120 799.9513699 Analyst (Cambridge UK) 1995,120,925.9513725 J. Anal. At. Spectrom. 1995 10 519.9513726 J. Anal. At. Spectrom. 1995 10 521. 9513727 J. Anal. At. Spectrom. 1995 10 527. 9513732 J. Anal. At. Spectrom. 1995 10 199. 9513766 Appl. Phys. B Lasers Opt. 1994,59 543. 9513767 Anal. Chem. 1995,67,354.95/3772 Anal. Sci. 1994,10,901.95/3777 Chem.- Anlangen Verfahren 1994,27,92.95/3785 Fenxi Huaxue 1994 22 752. 9513790 Fresenius’ J. Anal. Chem. 1994 349 738. 9513791 Fresenius’ J. Anal. Chem. 1994 349 774. 9513792 Gaodeng Xuexiao Huaxue Xuebao 1994 15 809. 9513795 Huanjing Huaxue 1994 13 332. 9513800 J. Fiz. Malays. 1993 14 43. 9513803 J. Phys. IV&hairsp; 1994 4 661. 9513806 Kovine Zlitine Tehnol. 1993 27 301. 9513807 Lihua Jianyan Huaxue Fence 1994 30 74. 9513808 Mikrochim. Acta 1994 113 287. 9513809 Microchem. J. 1994 50 145. 9513814 Proc. Pak. Acad. Sci. 1994 31 39. 9513815 Proc. SPIE-Int. SOC. Opt. Eng. 1994 2306 127. 9513816 Quim. Nova 1994 17 369. 9513819 Spectroscopy (Eugene Oreg.) 1994 9 42. 9513840 324R Journal of Analytical Atomic Spectrometry August 1996 Vol. 11At. Spectrosc. 1995 16 86. 95/3842 Collect. Czech. Chem. Commun. 1994 59 2227. 9513846 Fresenius’ J. Anal. Chem. 1994 350 662. 95/3847 Fresenius’ J. Anal. Chem. 1995 351 403.95/3850 Fresenius’ J. Anal. Chem. 1995,351,438.9513857 Fresenius’ J. Anal. Chem. 1995 351 479. 95/3858 Fresenius’ J. Anal. Chem. 1995 351 577. 95/3861 Guangpuxue Yu Guangpu Fenxi 1994 14(5) 63. 9513863 Guangpuxue Yu Guangpu Fenxi 1994,14( 5 ) 79.95/3871 Lihua Jianyan Huaxue Fence 1995 31 40. 95/3872 Lihua Jianyan Huaxue Fence 1995 31 41. 95/3874 Lihua Jianyan Huaxue Fence 1995 31 48. 95/3878 Springer Ser. Chem. Phys. 1994 59 83. 95/3879 Appl. Spectrosc. 1995 49 413. 95/3880 Appl. Spectrosc. 1995 49 499. 9513882 Appl. Spectrosc. 1995 49 655. 95/3887 Appl. Opt. 1995,34,3250.95/3889 Appl. Opt. 1995,34,3318.95/3966 Tech. Instrum. Anal. Chem. 1994,15 53.95/3967 Tech. Instrum. Anal. Chem. 1994 15 87. 95/3981 Am. Clin. Lab. 1994 13 45. 9513985 Anal. Sci. 1995 11 17. 95/4028 Fresenius’ J. Anal. Chem. 1994,350,504.95/4032 Guangpuxue Yu Guangpu Fenxi 1994 14(5) 69. 95/4160 Bunko Kenkyu 1995 44 10. 95/4165 Analyst (Cambridge UK) 1995 120 1205. 95/4169 J. Anal. At. Spectrom. 1995 10 563. 95/4176 J. Anal. At. Spectrom. 1995 10 609. 95/4246 Guangpuxue Yu Guangpu Fenxi 1995 15(4) 85.95/4247 Guangpuxue Yu Guangpu Fenxi 1995,15(3) 97. 95/4250 Lihua Jianyan Huaxue Fence 1995 31 207. 95/4256 Fenxi Shiyanshi 1995 14( 3) 26. 95/4261 Anal. Chim. Acta 1995,311,93.95/4345 Analusis 1994,22(9) 440.95/4352 Anal. Chim. Acta 1995,304( l) 115. 95/4356 Anal. Chim. Acta 1995 306(2-3) 183. 95/4360 Appl. Organomet. Chem. 1994 8( 7-8) 659. 95/4361 Appl. Organomet. Chem. 1994 8( 7-8) 687. 95/4363 Appl. Organomet. Chem. 1994 8( 7-8) 709. 95/4365 At. Spectrosc. 1994 15(5) 205. 95/4369 Bunseki Kagaku 1994,43(9) 661.95/4372 CLB Chem. Labor Biotech. 1995 46(2) 68. 95/4373 At. Spectrosc. 1994 15(6) 250. 95/4374 Fenxi Ceshi Xuebao 1994 13(5) 1. 95/4375 Fenxi Ceshi Xuebao 1994 13(6) 92. 95/4377 Fenxi Shiyanshi 1994 13(4) 65. 95/4378 Fenxi Shiyanshi 1994 13(4) 76. 95/4381 Fresenius’ J. Anal. Chem. 1994,350(6) 410.95/4384 Fresenius’ J. Anal. Chem. 1994 350( 12) 667. 95/4392 GIT Fachz. Lab. 1995 39(2) 100. 95/4397 Guangpuxue Yu Guangpu Fenxi 1994,14( 5 ) 59.95/4405 Huaxue Tongbao 1994,6,50.95/4406 Indian J. Chem. Sect. A Inorg. Bio-inorg. Phys. Theor. Anal. Chem. 1995,34( 3) 246.95/4409 Jisuanji Yu Yingyong Huaxue 1994 11(4) 293. 95/4410 Zh. Anal. Khim. 1995 50(1) 2. 95/4418 Liaoning Shifan Daxue Xuebao Ziran Kexueban 1994 17 (2) 134.9514421 Lihua Jianyan Huaxue Fence 1994,30( l) 9.95/4424 Magy. Kem. Foly. 1994 loo( lo) 450.95/4426 Mie- ken Kankyo Kagaku Senta Kenkyu Hokoku 1995 15 81. 9514428 Neimenggu Daxue Xuebao Ziran Kexueban 1994 25(2) 187. 95/4432 Proc. SPIE-Int. SOC. Opt. Eng. 1994,2124 145. 95/4436 Rev. Sci. Instrum. 1995 66(3) 2423. 95/4443 Spectrochim. Acta Part B 1995 50(1) 69. 95/4452 Talanta 1995 42(3) 375. 95/4453 Ukr. Khim. Zh. (Russ. Ed.) 1994 60( l ) 56. 95/4467 Yankuang Ceshi 1994 13(4) 298. 95/4469 John Wiley and Sons Ltd. Chichester W. Sussex UK 1995. 0 471 95331 8. 240pp. 95/4482 Anal. Chim. Acta 1995 303(2-3) 341.95/4493 At. Spectrosc. 1995,16(3) 106.95/4494 At. Spectrosc. 1995 16(3) 115. 95/4499 Anal. Sci. 1995,11(3) 363. 95/4506 Bunseki Kagaku 1995 44(5) 391. 95/4507 Bunseki Kagaku 1995 44( 5 ) 405. 95/4508 Bunseki Kagaku 1995 44(6) 437. 95/4510 Chem. Anal. (Warsaw) 1995 40(3) 351.95/4511 Chem. Anal. (Warsaw) 1995,40(3) 361.95/4512 Chemom. Intell. Lab. Syst. 1995 28(1) 73. 95/4516 Fenxi Huaxue 1995,23( 3) 367.95/4519 Fenxi Shiyanshi 1994,13( 5 ) 23. 95/4522 Fenxi Shiyanshi 1994 13(5) 66. 95/4526 Fenxi Shiyanshi 1995 14(2) 76. 95/4528 Fenxi Shiyanshi 1995 14( 3) 77. 95/4532 Fresenius’ J. Anal. Chem. 1995,351 (8) 798. 95/4550 Guangpuxue Yu Guangpu Fenxi 1994 14(6) 99. 95/4555 J. Agric. Food Chem. 1994 42(12) 2848. 95/4564 Microchem. J. 1995,51(3) 360.95/4579 Spectra Analyse 1995 182,42. 95/4583 Spectrochim. Acta Part B 1995 50(4-7) 369. 95/4584 Spectrochim. Acta Part B 1995,50(4-7) 377.95/4594 Spectrochim. Acta Part B 1995,50(4-7) 517.95/4597 Talanta 1995,42(4) 527.95/4621 Chem. Unserer Zeit 1994,28( 5 ) 222. 95/4683 J. Anal. At. Spectrom. 1995 10(10) 699. 95/4684 J. Anal. At. Spectrom. 1995 10(10) 703. 95/4685 J. Anal. At. Spectrom. 1995 10( lo) 711. 95/4686 J. Anal. At. Spectrom. 1995,10( lo) 721.95/4688 J. Anal. At. Spectrom. 1995,10( lo) 733. 95/4689 J. Anal. At. Spectrom. 1995 10( lo) 739. 95/4691 J. Anal. At. Spectrom. 1995 10(10) 753. 95/4692 J. Anal. At. Spectrom. 1995 10(10) 757. 95/4693 J. Anal. At. Spectrom. 1995,10( lo) 763.95/4694 J. Anal. At. Spectrom. 1995 10( lo) 769. 95/4695 J. Anal. At. Spectrom. 1995,10( lo) 777.95/4696 J. Anal. At. Spectrom. 1995 10(10) 785. 95/4697 J. Anal. At. Spectrom. 1995 10(10) 791. 95/4698 J. Anal. At. Spectrom. 1995,10( lo) 799.95/4699 J. Anal. At. Spectrom. 1995,10( lo) 803. 95/4701 J. Anal. At. Spectrom. 1995,10( lo) 815.95/4713 J. Anal. At. Spectrom. 1995 10( lo) 885. 95/4717 Spectrochim. Acta Part B 1995,50( 3) 209.95/4723 Spectrochim. Acta Part B 1995,50(4-7) 477.95/4731 Spectrochim. Acta Part B 1995 50(8) 707. 95/4732 Spectrochim. Acta Part B 1995 50(8) 713. 95/4733 Spectrochim. Acta Part B 1995 50(8) 725. 95/4735 Spectrochim. Acta Part B 1995 50( 8) 753. 95/4736 Spectrochim. Acta Part B 1995 50( S) 763. 95/4739 Spectrochim. Acta Part B 1995 50(8) 803. 95/4742 Spectrochim. Acta Part B 1995 50( 8) 837. 95/4743 Spectrochim. Acta Part B 1995 50( 8) 847. 95/4744 Spectrochim. Acta Part B 1995 50( 8) 857. 96/67 VCH New York NY USA 1994. 175. 96/70 Anal. Sci. 1995 11(2) 267. 96/74 Appl. Organomet. Chem. 1994 8(7-8) 595. 96/78 At. Spectrosc. 1995 16(2) 82. 96/97 Environ. Sci. (Tokyo) 1995 3(2) 113. 96/103 Fresenius’ J. Anal. Chem. 1995 351(7) 695. 96/111 Guangpuxue Yu Guangpu Fenxi 1995,15(2) 81.96/145 Rev. Esp. Cienc. Tecnol. Aliment. 1994 34(4) 441. 961149 Shojinkai Igakushi 1994 33( l) 79. 96/154 Shanghai Huanjing Kexue 1994 13(2) 25. 96/156 Talanta 1995 42(4) 557. 96/170 5,384,656 24 Jan 1995 (CI. 359-659) 18 May 1993. Continuation of Ser. 778,302 16 Oct 1991 abandoned. 961171 Appl. Opt. 1995 34(21) 4437. 96/274 J. Anal. At. Spectrom. 1995 10(11) 969. 96/275 J. Anal. At. Spectrom. 1995 10(11) 975. 96/277 J. Anal. At. Spectrom. 1995 10( 1 l) 987. 96/279 J. Anal. At. Spectrom. 1995 10(11) 993. 96/280 J. Anal. At. Spectrom. 1995 10(11) 999. 961281 J. Anal. At. Spectrom. 1995,10( 1 l ) 1003.96/289 J. Anal. At. Spectrom. 1995,10( 12) 1059. 96/290 J. Anal. At. Spectrom. 1995 10( 12) 1065. 961291 J. Anal. At. Spectrom. 1995 10(12) 1069. 961292 J. Anal. At. Spectrom. 1995 10( 12) 1077. 96/300 Anal. Chim. Acta 1995 308( 1-3) 364. 96/301 Anal. Chim. Acta 1995 308( 1-3) 85. 96/304 Anal. Chim. Acta 1995,310( l ) 189.96/307 Anal. Chim. Acta 1995 310(1) 181. 96/311 Analusis 1995 23(3) 117. 96/312 Analusis 1995,23(4) 159.96/313 Analusis 1995,23(4) 168. 96/318 Am. Lab. (Fairfield Conn.) 1995 27(4) 48X. 96/343 Appl. Spectrosc. 1995 49(7) 890. 96/344 Appl. Spectrosc. 1995 49(7) 907. 96/353 Bunseki Kagaku 1995 44(7) 521. 96/354 Bunseki Kagaku 1995 44(8) 617. 96/358 Fenxi Huaxue 1995 23(6) 733. 96/363 Fenxi Huaxue 1995 23(4) 378.96/368 Fresenius’ J. Anal. Chem. 1995,352(6) 604. 96/387 J. Autom. Chem. 1995 17(2) 65. 96/412 Zh. Anal. Khim. 1994 49(10) 1083. 96/414 Zh. Anal. Khim. 1995 50(3) 242. Journal of Analytical Atomic Spectrometry August 1996 Vol. 11 325R