Atomic Spectrometry Update- Advances in Atomic Absorption and Fluorescence Spectrometry and Related Techniques 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 MK60 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 3451 0 Amherst M A 01 003-451 0 USA SUMMARY OF CONTENTS 1. Atomic Absorption Spectrometry 1.1. Flame Atomizers 1.1.1. Fundamental studies 1.1.2. Interference studies 1.1.3. Sample introduction 1.1.3.1. Discrete procedures 1.1.3.2.Atom-trapping techniques 1.1.3.3. Sample introduction by flow injection 1.1.3.4. Solid sample introduction 1.1.3.5. Thermospray introduction 1.1.3.6. High-pressure nebulization 1.1.4. Sample pre-treatment 1.1.5. Chromatographic detection 1.1.6. Plasma atomizers 1.2. Electrothermal Atomizers 1.2.1. 1.2.1.1. Graphite atomizers 1.2.1.2. Metal atomizers and metallic coatings 1.2.2. Sample introduction 1.2.2.1. Slurry sampling 1.2.2.2. Solid sampling 1.2.2.3. Gas sampling 1.2.2.4. Coupled techniques and preconcentration 1.2.2.5. Electrodeposition 1.2.3. Fundamental processes 1.2.4. Interferences 1.2.4.1. Spectral interferences 1.2.4.2. Chemical modifiers-general Atomizer design and surface modification This review follows on from last year’s review (J. Anal.At. Spectrom. 1994 9 213R) 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 JAAS Volume 10 Issue 4. With respect to section 1.2 Electrothermal Atomizers some changes have been “Review Coordinator to whom correspondence should be addressed. Atomic Spectrometry Update 1.2.4.3. Chemical modifiers-palladium 1.2.4.4. Other chemical modifiers 1.2.5. Developments in technique 1.3. Chemical Vapour Generation 1.3.1.Hydride generation 1.3.1.1. General studies of fundamentals techniques and 1.3.1.2. Determination of individual elements instrumentation 1.3.2. 1.3.3. 1.4. 1.4.1. 1.4.2. 1.4.3. 1.5. 1.5.1. 1.5.2. 1.5.3. 2. 2.1. 2.2. 2.3. 3. Mercury by cold vapour generation Volatile organo-metallic compound generation and metal vapour separation Spectrometers Light sources Continuum source and simultaneous multi-element AAS Background correction in AAS Instrument Control and Data Processing Instrument control Data processing Chemome t rics Atomic Fluorescence Spectrometry Discharge Lamp-excited Atomic Fluorescence Laser-excited Atomic Fluorescence Studies of Flames Plasmas and Atomic Vapours Using Laser-induced Fluorescence Laser Enhanced Ionization implemented.The larger sections have been divided into sub- sections to allow easier and quicker access to material covering a specific area. Much of the work previously reported under section 1.2.5 Developments in technique was being duplicated in section 1.4 Spectrometers. In the present Update a more rigorous division of the subject matter bas been applied in an attempt to reduce the size of the Electrothermal Atomizers section which in the opinion of the Update authors was becoming rather long. Consequently some work that was previously discussed under section 1.2.5 will now be found under sections 1.4 and 1.5. The full references names and addresses of authors can be readily found from the Atomic Spectrometry Updates References in the relevant issue of JAAS. However as an additional service to readers an abbreviated form of each reference quoted (except for those of Journal of Analytical Atomic Spectrometry August 1995 Vol.10 199 RConference 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. carbon black to remove the interference of aluminium on the determination of Mg in a Mg-A1 alloy but found that the procedure was ineffective in the cases of Ba Ca and Sr. The effects of a number of other organic additives have been 1. ATOMIC ABSORPTION SPECTROMETRY 1.1. Flame Atomizers The material published during this review period is very similar to that of the previous year (see J. Anal. At. Spectrom. 1994 9,213R) in that the amount of published work of a fundamental nature was relatively small in comparison with that describing particular determinations and novel aspect of sample pretreat- ment.If the sample pretreatment was directly coupled to the instrument then the work has been included in section 1.1.3. if the pretreatment was not directly coupled then the paper is included in section 1.1.4. only if the work contains some feature that could be exploited in an on-line situation. 1.1.1. Fundamental studies The imaging of OH (by LIF) and the study of temperature distributions (from Rayleigh scattering) in turbulent diffusion flames were described (94/3279) by Kelman and Masri. A laminar flow air-CH flame was used for calibration purposes. The fuels studied were compressed natural gas (CNG) a mixture of CNG and hydrogen and a mixture of hydrogen and argon.These latter two mixtures were chosen because their Rayleigh cross section is close to that of air (used in all cases as the oxidant). The authors validated the results of their study by showing that good agreement was obtained between measured values and calculated values for the model (air-CH,) flame system. This flame was also the object of a study of CO and temperature distributions (94/2240). A tuneable diode laser (output between 2090 and 2160cm-’) was used as the light source for absorbance and wavelength modulation (second derivative) studies. Good agreement with the results of previous studies based on LIF and quartz microprobe MS experiments was obtained. Useful data concerning the production of CO the agent largely responsible for loss of life in the ceiling layers of room fires was provided. 1.1.2.Interference studies Further work on the interference of aluminium on the alkaline earth elements in the air-C,H flame has been reported by Welz and Lueke (94/2399 94/3349). The phenomenon under investigation is the restoration of the analyte signal (to a value approaching that obtained in the absence of aluminium) as the aluminium concentration is increased beyond that needed to produce a signal suppression. In the first of these studies (94/2399) the effect of aluminium nitrate and chloride on Ca Mg and Sr was investigated and it was concluded that the influence of the chloride was less pronounced because the analyte chloride must be converted to the oxide before it reacts with the aluminium oxide.Any compound that delays the formation of oxide will alleviate the depressing effect of alu- minium and when higher concentrations of aluminium (as chloride) are introduced the larger particles produced isolate the analyte atoms from the oxygen in the flame owing to the production of a hydrogen chloride atmosphere. The effect of caesium chloride addition a further release from the depressive effect of aluminium was interpreted as supporting the proposed mechanism. In the second study (94/2399) the effect of a h - minium on Ba and Sr was reported. The roles of the releasing agents lanthanum EDTA and quinolin-8-01 were investigated and it was found that although lanthanum and ETDA were effective in the N20-C2H2 flame quinolin-8-01 was the best for both analytes.None of the releasing agents was effective in the cooler air-C,H flame. Japanese workers (95/124) used studied. Brilliant green was shown (95/38) to act a a releasing agent in the determination of Rh. The interferences of a number of anions and cations in the determination of Au have been quantified (94/3244) and a number of possible releasing agents evaluated. The most effective releasing agent was triethanolam- ine which was also used in the froth flotation preconcentration and separation of Au from spiked natural water samples. An overview of the role ofsurfactant-based ordered media in analytical atomic spectrometry has been provided by Sanz- Medel et al. (94/3036). This paper briefly describes some of the recent results on the use of surfactants to improve the sample introduction characteristics of FAAS from which the authors concluded that ‘much more work is needed to unveil such mechanisms of surfactant signal enhancement or surfac- tant interference elimination in flames’.Disappointingly it was also concluded that ‘trial and error techniques are mandatory’. The use of surfactants to form oil-in-water emulsions as a means of introducing non-aqueous samples was also discussed. A study (94/2203) of the effect of acetone ethanol ethylene glycol glycerol and 1,4-dioxane on the signals for Ca Cd Cu and Fe provides a good example of this trial and error approach. Changes in signal intensity as a function of density surface tension and viscosity in accordance with an empirical equation derived by Greenfield et al.(Anal. Chim. Acta 1976 84 67) were found. For Cd Cu and Fe signal enhancements of up to 1.8-times were obtained with 40% ethanol solutions. For Ca none of the solvents produced much of an enhancement or a depression. The effects were slightly more pronounced for introduction into a plasma spectrometer. Chinese workers (95/291) have studied the effects of a number of surfactants on the signal for Yb. Sodium hexylsulfonate and sodium laurylsul- fonate were found to be helpful in overcoming interferences from some concomitants. In a study of the sequential extraction oftrace elementsfrom soils according to standard protocols it was found that the presence of ammonium acetate gave high results in the determi- nation of Cd Cr and Ni (95/119).It was suggested that matrix matching be used in this case. For multiplicative interferences in the analysis of poultry feed use of the method of standard additions was advocated (95/412). 1.1.3. Sample introduction Work discussed in the previous section on the use of organic additives is relevant to this section as are some of the studies described later in section 1.1.4. in which an organic solvent was nebulized directly into the flame. Apart from the small amount of work on high pressure nebulization and thermo- spray devices (see sections 1.1.3.5. and 1.1.3.6.) which seems to be at a lower level of activity than for the previous review period (see J. Anal. At. Spectrom. 1994 9 213R) most of the research reported in the literature has been concerned with the application of commercially available pneumatic nebulizers.Chinese workers described (94/3052) a vertically mounted nebulizer which used gravity to discriminate between the larger droplets and those small enough to be entrained into the surrounding gas stream. 1.1.3.1. Discrete procedures. Conceptually there is only a small difference between the use of FI sample introduction and the use of a discrete micro-sampling system. A system which could be considered to be either was described by Xu and Fang (94/C3378) in that the sample was transported by air and introduced into the nebulizer via hydrodynamic injec- tion. The system was used for the determination of Cu in serum at a rate of about 100 h-’ with a total sample consump- 200R Journal of Analytical Atomic Spectrometry August 1995 Vol.10tion of 800 pl for triplicate injections. The same matrix has been analysed by other Chinese workers (94/2625). Calcium Cu and Mg were determined at ppm concentrations with precisions between 2 and 6% RSD. The technique has also been applied (95/402) to the determination of these analytes and Zn in whole blood. The direct nebulization of organic solvents that would extinguish the flame under normal circumstances has been described (95/304). Eight chelating agents were used in con- junction with three chlorinated solvents (dichloromethane chloroform and carbon tetrachloride). Volumes injected ranged from 20 to 50 p.1 and both the air-C2H and N20-C2H flames were used. The best performance was obtained for Ag Cd Pb and Zn but acceptable results were also found for Co Cu Fe Mg Mn and Ni.For the air-C2H2 flame there was a depen- dence of signal shape on the element that correlated with both the ash/atomize behaviour in a graphite furnace atomizer and the metal-oxygen bond strength. The height of the initial transient signal to the steady-state signal depended on the effect of flame temperature on the degree of atomization which in turn was reflected in the ease of atomization of the metal from an oxide precursor. The authors have also described (95/55) some further applications of the procedure in the analyses of biological botanical and silicaceous materials following microwave digestion. An adjustable volume (10-100 p l ) sampling device has been described by Russian workers (95/174) and Chinese workers have combined discrete sampling with a stainless steel slotted tube atom retarder for the determination of Cd Pb and Zn in atmospheric particulates (94/C3414).The increased sensitivity available allowed sample collection times to be reduced from 10 to 2 h. 1.1.3.2. Atom-trapping techniques. Despite the sustained efforts of a number of Chinese research groups there would appear to be few significant improvements in slotted tube atom retarder (STAR) performance since the original work by Watling (Anal. Chim. Acta 1977 94 181; and 1978 97 395). Further work on metallurgical analyses (94/3336) by Thorburn Burns et al. has demonstrated the improved precision that is possible with a STAR. Tin was accurately determined in a number of alloys at concentrations ranging from 0.7 to 9.8 % m/m with precisions ranging from 0.2 to 1.4 YO RSD which was 2-3 times better than that obtained with conventional nebulization. A sensitivity enhancement of a factor of 2 was obtained and it was observed that the presence of copper ( 1000 ppm) decreased the interferences from antimony arsenic bismuth chromium lead manganese nickel and zinc.As might be expected for easily atomized elements prelimi- nary results (94/3296) for the introduction of organometullic compounds in IBMK have demonstrated that enhancement effects due to the organic solvent are also observed with the STAR. Sensitivity increases of factors of 11 6 and 9 were obtained for Ag Cu and Pb respectively. No improvement was obtained for Fe and Ni which was interpreted as being due to the poorer degree of atomization in the cooler gases within the quartz tube.Chinese workers have used a STAR in the determination of Cu in rice (95/391) and a combined STAR and atom trap (94/1360) for the determination of Ag Au Cu Pb and Zn. Sensitivity enhancements by factors of 62 106 28 83 and 77 were obtained respectively. The device was used in the accurate determination of Ag and Cu in standard oyster samples. A stainless steel STAR has been used (94/C3414) in a procedure for the determination of Cd Pb and Zn in atmospheric particulates which also used discrete nebulization (see section 1.1.3.1 .). 1.1.3.3. Sample introduction by flow injection. As has now been the pattern for several years much of the novel work on the development of methods with flame AAS detection involves FI.Flow injection techniques with a variety of spectrometric detectors have been reviewed (95/5) and recent developments in the particular application for chemical vapour generation AAS surveyed (94/C3383). Welz and Sperling reviewed the FI atomic spectrometry combination (FIAS) under the title of 'the ultimate approach to automation in analytical atomic spectroscopy' (95/152). In addition to the usual examples of microsampling preconcentration and matrix removal a pro- cess analysis application was described in which the Bi concen- tration in the acid cleaning solution applied to aircraft engine blades was automatically monitored. Provided full automation is not needed the hardware needed to implement an FI sample introduction system can be very inexpensive indeed.The use of low cost components (home made injection valve and aquarium pump) was shown (94/3015) to give adequate performance for the determination of Ca K and Na by FAES. At the other end of the cost scale a new high pressure nebulizer device (see section 1.1.3.6.) with potential for use in LC applications has been evaluated in the FI mode (Todoli et a!. Spectrochim Acta Part B 1993 48 1461 ). Several applications which exploited the microsampling characteristics of FI have been described. Calcium has been determined along with phosphate in serum (95/399) and Hg measured in blood saliva and urine (94/1838). In the latter procedure an off-line microwave digestion was used. A pro- cedure involving on-line microwave digestion for the determi- nation of Cu and Zn in whole blood was developed (94/2228) in which the patient was connected directly into the FI manifold.After mixing with an anticoagulant and acid the sample passed through a microwave oven and after degassing introduced into the nebulizer of a flame spectrometer. On-line microwave digestion has also been applied (94/3034) in the determination of Cu Mn Pb and Zn in a variety of solid matrices including vegetables and sewage sludge. The samples were slurried (20 mg ml-') in concentrated nitric acid and 1 0 0 ~ 1 portions of slurry injected and merged with the same volume of 30% hydrogen peroxide solution. After passage through the 650 W oven the solution was delivered to a specially designed vessel which allowed evolved gases to escape and the nebulizer to aspirate solution at the optimum rate.Accurate results were obtained apart from low recoveries of Zn from sewage sludge. Flow injection slurry nebulization has been applied for a variety of determinations including Ca Fe Mg Mn and Zn in vegetables (94/3287) Ca Fe and Mg in materials with a high silica content (94/3035) and Fe and Mn in cement (94/3330). In the first of these studies (94/3287) Tyson's reversed FI standard additions method was used as well as calibration against aqueous standards. A sample volume of 150 pl was injected into a carrier flowing at 1.8 ml min-'. Air compensation was used. For the final study (94/3330) volumes of up to 200 pl were injected into a carrier stream flowing at 4ml min-'.No air compensation was used. All three studies employed a high dispersion manifold consisting of a single well-stirred mixing tank for the dilution of off- range concentrations. A method for dealing with the high dissolved solids arising from metal and rock samples has been devised (94/2623) in which gelatin was added to the sample solution. Detection limits of 20 and long g-' were obtained for Ag and Cu respectively. Highly concentrated sample solutions were delib- erately prepared for the determination of Pb in paint chips by FI-FAAS (94/C1938) so as to achieve a lower limit of detection in the solid by avoiding the dilution necessary to avoid clogging of the nebulizer and burner when conventional nebulization is employed. A number of flow-based systems for sample dilution and calibration have been described (94/2254 94/2922 95/305 95/451).In the first of these (94/2254) a small volume (down Journal of Analytical Atomic Spectrometry August 1995 Vol. 10 201 Rto 0.7 pl) was introduced into the loop of the injection valve by controlled rotation of the stepper-motor driven peristaltic pump head. On switching the valve to the inject position this small volume was then washed out into a knotted tubular reactor to give dilution factors up to 1330 (with a precision of 2% RSD). The power of the procedure was demonstrated by the accurate determination of Mg in an aluminium reference alloy at a concentration of 7.17% m/m. A similar approach (95/305) has been used with a voltage controlled peristaltic pump.The same pump was used both to aspirate sample solution into the loop of the injection valve and to deliver the carrier stream to the spectrometer. With the valve in the fill position and a low applied voltage a small volume (down to 1.5 pl) was drawn into the loop. The valve was switched to the inject position and the voltage ramped linearly up to a suitable value for introduction to the spectrometer. A T-piece located in front of the nebulizer allowed compensation by a diluent so as to keep the flow into the nebulizer constant. Dilution factors up to 2000 were obtained with precisions for peak area measurements of between 0.6 and 2.3% RSD. Results of experiments designed to test the hypothesis that the problems associated with poor repeatability of small volumes metered by peristaltic pumps can only be solved by the use of combinations of pump rate and metering times such that the product is an integer multiple of the period of the rollers' pulse showed that this idea had some merit.A variable injection volume device for the implementation of the standard additions method has been described by Brazilian workers (94/3112). The use of a dialysis membrane as a method of sub-sampling is well known and further evidence of the practical use of this method for the analysis of clinical samples has been provided (95/451). Dialysis and stream splitting were used to bring the amount of Ca and Na in serum samples introduced into the spectrometer into the working range. A full description of the use of discontinuous flow (DF) sample handling and calibration for FAAS has now appeared (94/2922).Three methods of DF implemented by a commer- cially available device were described. In the first the standard solution was delivered from a cam-driven piston pump and merged with diluent which was aspirated from a reservoir by the suction provided by a peristaltic pump running at constant speed downstream of the confluence point. The standard and diluent solutions were merged at flow rates which produced four standards (0 30 71 and 100% of the standard concen- tration). In the second method the manifold was modified so that the method of standard additions could be implemented by the introduction of sample solution from an open reservoir downstream of the standard-diluent confluence point but upstream of the peristaltic pump. This method was not tested in the presence of an interferent but for a 10 ppm Ca standard a bias of about 2% was found.In the third method a system of three cam-driven piston pumps was used to generate a linear concentration gradient between a low and a high standard. Sample was introduced via a fourth cam-driven pump. Solution was withdrawn from the manifold upstream of the merging point of the two standard streams at a flow rate equal to that at which the high standard was added. Efficient mixing was provided by a vibrating reed mixer (see Sperling et al. Anal. Chem. 1991 63 151). For the replicate determination of 10 ppm Ca using standards of 2 and 20 ppm the bias and precision were both 0.5%. Methods 2 and 3 took about 30 s to implement and eliminated the effects of instru- mental drift as the system was constantly recalibrated for each sample. As in previous review periods there is still considerable activity in the development of on-line solid-phase extraction (SPE) procedures.Several papers have been concerned with the Cr"'/Crv1 speciation problem (94/1605 94/C1937 94/C3489). Some evaluative comments on the method of Sperling et al. (Anal. Chem. 1992 64 3103) have been made (94/C1937). In this procedure alumina was used as the solid phase extractant whose charge was controlled by the pH of the buffer carrier stream. The procedure was found to be very sensitive to this pH and thus control of the buffer capacity of the carrier and eluent (and a knowledge of that of the sample) were shown to be important.Interferences from iron sulfate and phosphate were found to be greater than originally reported. It was confirmed that the method was useful for determinations down to 10 ppb of either Cr species in relatively clean water samples of low ionic strength. An HPLC procedure (see section 1.1.5.) has been described for the same analysis (94/3280). In this procedure CrV' could be preconcentrated by changing the composition of the carrier with respect to the tetrabutylammonium acetate concentration. The eluent was a 60/40 (YO v/v) methanol-water solution. The limit of detection for CrV' in drinking water was 0.5 ppb. Workers at Strathclyde have investigated the use of a number of solid phases for this analysis (94/C3489). It was found that Muromac A-1 and quinolin-8-01 immobilized on controlled pore glass (CPG) would preconcentrate Cr"' but not Crvl whereas the Dow resins XFS-415 and XFS-43084 retained CrV' but not Cr"'.A procedure in which the CrV1 was reduced by hydroxylamine prior to retention on Muromac A-1 or on quinolin-8-01 on CPG was applied to the analysis of a number of estuarine and coastal waters. Covalently bound chelating agents have been used in a number of other studies (94/C1940,94/2188,94/3123,94/3156). A DETATA chelating agent bound to fibrous cellulose was used (94/C1940) to preconcentrate Cd by factors of between 50 and 100 while still retaining a throughput of 20 h-'. Fang and Dong determined the same analyte in urine (94/3156) with a procedure involving retention by quinolin-8-01 immobilized on CPG.An enrichment factor of 30 a sampling rate of 60 h-l and a detection limit of 0.3 ppb were obtained. Cysteine immobilized on CPG was used (94/2188) to preconcentrate Cd Cu Co Hg Pb and Zn. Improvements in detection limits by factors ranging from 500 to 2000 were obtained. Acids of various concentrations were used to elute the analytes with the exception of Hg for which thiourea had to be added to the eluent. The effects of the pH of the solution prior to loading and of high concentrations of potentially interfering cations were evaluated. The columns had useful lifetimes of at least 336 h. Gold has been preconcentrated (94/3123) on Cellex-T (cellulose powder with quaternary amine functional groups) followed by elution with 300 pl of 0.1 mol 1-l thiourea in 0.1 mol 1-l HCl solution.Several experimental variables were studied including possible interference effects. It was found that Pt" and Pd" as well as the common ionic constituents of natural waters did not interfere. This paper is an excellent source of information on previous efforts to preconcentrate Au by SPE procedures. A number of sorbent extraction procedures have been described (94/1644 94/3 106 94/3289 94/C3408 95/192) in which the analyte is derivatized as an uncharged chelate complex which is then retained on a non-polar stationary phase. In a high pressure procedure (94/1644) 10 elements were retained as their APDC complexes on a C-18 material followed by elution with methanolic nitric acid (0.01 mol 1-l). The method was applied to the analysis of aluminium and drinking water.A low pressure version of very similar chemistry has been applied (94/3289) to the determination of Cd Cu and Pb in ash soils and sediments. The complexing agent used was either diethylammonium-N,N-diethyldithiocarbamate or ammonium diethyldithiophosphate and elution was with methanol. A number of experimental variables were studied including the effects of pH reagent concentration use of masking agents nature of the eluent and the flow rates for loading and elution. Accurate analyses of some reference 202 R Journal of Analytical Atomic Spectrometry August 1995 Vol. 10materials were obtained with the exception of the determi- nation of Cd in the presence of high concentrations of iron. Valcarcel et al. (94/3106) determined Cd by retention of the dithizone complex on activated carbon followed by elution with IBMK.This protocol resulted from a comprehensive study of the possibilities of other chelating agents and eluting solvents as well as a number of experimental variables and a survey of published work on the SPE-FIAS determination of Cd. Preconcentration factors of between 40 and 130 were obtained with a detection limit between 0.3 and 1.3 ppb. The procedure was used to determine Cd accurately in a city waste incineration ash and in some biological reference materials. The same analysis was performed (94/C3408) by retention of the Cd-DDC complex on the interior walls of a PTFE knotted tubular reactor followed by elution with IBMK. About 80% of the analyte was retained at a loading flow rate of 5.2 ml min-'.The interferences from copper and iron were overcome by the addition of thiourea and ascorbic acid-1,lO-phenanthroline respectively. An enhancement factor of 66 and a detection limit of 0.1 ppb were obtained. In the determination of Co in oriental tobacco leaves and an artificial ocean water (95/192) the analyte was retained on Brockmann activity I acid alumina as the 1-nitroso-2-naphthol- 3,6-disulfonate with elution by dilute sodium hydroxide solution. A variety of experimental parameters were studied as well as possible interference effects from other metals. A 97% recovery of 1 ppb from the water sample was obtained. A detection limit of 0.44 ppb was obtained for a 20 min loading time. Alumina has also been used for the retention of As in water samples (94/C3388).Unlike most procedures for the determination of As the method required oxidation of the As"' to AsV. The loading flow rate was 16ml min-' with elution by 8 mol 1-' HC1 at 8 ml min-'. An enrichment factor of 48 was obtained. Some preliminary results on the possibilities of fibrous alumina for the determination of Pb in waters have been reported (95/194) by workers in a laboratory with extensive experience of using powdered alumina. It was found that the new material performed as well as the powder giving a detection limit of 0.7 ppb for a 15 ml sample. The sample was loaded at pH 8 (0.1 mol I-' ammonia) and eluted at pH 0 (1 mol 1-' nitric acid). There is also an interest in the use of immobilized mico- organisms for SPE (94/2878,94/2967).The yeast Saccharomyces cereuisiae immobilized on CPG has been used to preconcen- trate Cd Cu Fe Pb and Zn (94/2878). Again a variety of experimental variables were studied. Under optimized con- ditions enhancement factors of as high as 2000 (for Zn) were obtained. It proved difficult to remove Fe from the material and an eluent consisting of 0.5 mol 1-' perchloric and nitric acids was used. Despite this the lifetime was reported as 400 h provided microbial growth was prevented by storage at 4°C when not in use. Detection limits for all the analytes were below 1 ppb and Cd and Cu were accurately determined in a sediment reference material (BCR 144) with the aid of the standard additions method of calibration. The properties of two immobilized algae (Chlamydonomous reinhartii and Selenestrum capricornutum) for the preconcentration of Ag Cr"' Crv' and Cu have been investigated (94/2967).The algae were immobilized on silanized CPG and the study followed the protocols of previous work (see Elmahadi and Greenway J . Anal. At. Spectrom. 1991,6 643). The two algae were found to have significantly different retention characteristics for Cr"' and Crvl and this was exploited as a method of Cr speciation. The selectivity of each column was controlled by the pH as Chlamydonomous reinhartii preferentially retained CrV' from pH 4.0 phosphate buffer and Selenestrum capricornutum prefe- rentially retained Cr"' from pH 8.5 tris buffer. Chromium(Ix1) was eluted with a solution of thiourea in 0.1 rnol I-' perchloric acid and Crv~ with 0.1 mol 1-' sodium hydroxide solution.There is still a low level of interest in the use of FI liquid- liquid extraction (LLE) procedures directly coupled to FAAS though it is not clear what advantage such a system would have over an appropriately designed SPE system for the same analysis. Nonetheless Chinese workers have reported (95/241) a method for the determination of Mn in hair water and herbal medicine based on extraction with APDC into IBMK. A detection limit of 70 ppb was obtained. A detailed study of the FI extraction of Au and Cu as thiocyanates into IBMK has been described (94/1650). The flow rates were 12.6ml min-' and 0.42 ml min-' for the sample and IBMK respect- ively and segmentation was achieved at a three-way connector with an angle of 120" between each channel.The phase separator was an asymmetric hour-glass shaped glass vessel with an upper chamber volume of 1 ml and a lower chamber volume of 10ml. The mixed phases were introduced at the narrow neck between the chambers. The dimensions of the constriction between the chambers were not given. Detection limits of 2 ppb and 1 ppb for Au and Cu were obtained and the method was used for the accurate determination of these elements in some standard ore samples. In a separate publi- cation (95/220) the same system was used for the determi- nation of Cd in ores down to a detection limit of 0.5 ppb by extraction of the biphenylenebisazoaminoazo-benzene complex into IBMK. A mixed solvent (toluene-xylene 1 4) has been used (94/2250) for the extraction of Cu as the complex with 5,7-dibromo-quinolin-8-01. A T-shaped segmentor was used and phase separation was achieved with a PTFE membrane separator.The detection limit was 0.44 ppb for a preconcen- tration factor of 16 and a throughput of 180 h-'. Yet another system (95/283) for the determination of Ag in geological samples based on extraction with diphenylthiourea into IBMK had a detection limit of 0.54 ppb and a throughput of 30 h-'. A further contribution (94/1821) in the series of indirect methods developed by Valcarcel's group described the determination of amylocaine and bromhexine at concen- trations between 3 and 120ppm by the extraction of the reineckates into 1,2-dichloroethane and determination of the Cr in the organic phase. Kuban has reviewed the use of continuous precipitation techniques in FI analysis (94/2049).A total of 75 references were cited covering applications for FAAS UV-visible spectro- photometry and potentiometry. It was concluded that precipi- tation/dissolution techniques were simple rapid precise and easy to automate requiring lower sample volumes and leading to reduced reagent consumption (in other words all the usual advantages of the FI format). It was observed that crystalline precipitates were harder to handle than gelatinous or curdy material because the former tended to block the filter media and also required longer induction periods necessitating the use of longer reaction coils and/or stopped flow procedures. As the neutral metal complexes used in sorbent extraction procedures (see earlier) are insoluble in water the same reagents can be used in methods of preconcentration and matrix isolation based on precipitation or coprecipitation.Fang and coworkers have described methods for the determination of Ag in geological materials (95/334) and for the determination of Cd Co and Ni in biological materials (95/398). In the former study the Ag was collected by co-precipitation with the iron salt of diethyldithiocarbamate in the presence of 1,lO-phenanthroline and retention of the precipitate on the walls of a knotted tubular reactor without a filter. The precipi- tate was dissolved in IBMK and a detection limit of 0.5 ppb was obtained with an enhancement factor of 26 and a through- put of 62 h-'. A very similar procedure was used in the second study (95/398) with the exception that the analytes were coprecipitated with the APDC complex of iron.Detection limits of 0.3 1 and 3 ppb for Cd Co and Ni respectively were obtained. A cylindrical stainless-steel filter was used (94/2982) Journal of Analytical Atomic Spectrometry August 1995 Vol. 10 203 Rfor the collection of Ni as the 1-nitroso-2-naphthol salt for the analysis of silicate rocks. The sensitivity was considerably increased by precipitation in the presence of hydrogen peroxide and the iron interference was masked with Tiron. The precipi- tate was dissolved in ethanol. A 50-fold preconcentration was obtained from a 10 ml sample with a throughput of 15-20 h-'. Several indirect FI precipitation methods have been described. In a procedure (94/2204) very similar to that based on liquid- liquid extraction (LLE) (see J.Anal. At. Spectrom. 1994 9 213R) papaverine strychnine and cocaine were determined by precipitation with Dragendorff's reagent (Bi14-) The pro- cedure was based on the decrease in the bismuth signal obtained on the introduction of the sample. A number of experimental variables were examined and the final procedure developed applied to the determination of papaverine in a number of pharmaceutical preparations. An indirect method (94/3063) for the determination of cyanide and thiocyanate based on the precipitation and redissolution of the silver salts has been developed by Esmadi et al. and workers in Taiwan have determined sulfate (94/3058) by monitoring the change in the signal due to barium. Also of relevance to this section are the considerable number of papers describing the use of FI techniques for the implemen- tation of a chemical vapour generation procedure (see sec- tion 1.3.) and in conjunction with ETAAS (see section 1.2.2.4.) 1.1.3.4.Solid sample introduction. Further evidence from Spain of the usefulness of the introduction of slurries by FI has been provided (94/3034,94/3035,94/3287). No other work on solid sampling has been reported during this review period. Workers at Murcia (94/3287) have devised a procedure for the determination of Ca Fe Mg Mn and Zn in vegetables. After drying and ashing the residue was ground and for the determination of Fe Mn and Zn was suspended with the aid of ultrasonication in a 10% glycerol 1% hydrochloric acid solution to give a slurry concentration of not more than 1% m/v.For the determination of Ca and Mg the ash was suspended in a 1% hydrochloric acid 1% lanthanum and 0.02% Triton X-100 solution. For the determination of the former group of elements a 150 p1 sample was injected into a water carrier. The concentrations of Ca and Mg were too high for direct introduction and an on-line dilution procedure using a single well-stirred mixing chamber was used (see Lopez Garcia et al. J. Anal. A t . Spectrom. 1992 7 1291). A similar procedure (9413035) was used by these workers for the determi- nation of Ca Fe and Mg in samples with a high silica content (such as diatomaceous earth). Slurries were prepared by grind- ing in an agate ball mill followed by suspension in a 2% hydrochloric acid 1% hydrofluoric acid solution at a concen- tration appropriate for the element.The particle size distri- bution of the ground sample showed that 90% of the particles were below 11 pm in size. The maximum slurry concentration was 1% m/v. Accurate analyses of some geological reference materials were obtained. Workers at Valencia (94/3330) have determined Fe and Mg in cement by a procedure in which the analytes were leached from the solid by sonication in an hydrochloric-nitric acid solution prior to injection into the FI manifold. For the determination of Fe an on-line dilution was effected by a high dispersion manifold containing a single well-stirred mixing chamber. The initial slurry concentration was 0.2% m/v. The same researchers have also developed (94/3034) an on-line microwave digestion procedure for leaching Cu Mn Pb and Zn from vegetables powdered dietary products and sewage sludges A 100 p1 portion of slurry (0.2% in concentrated nitric acid) was injected into a water carrier and merged with a 30% vjv hydrogen peroxide solution in a merging zones manifold. After passage through 100 cm of tubing in a microwave oven the solution was aspirated into the spectrometer by nebulizer suction from a cooled interface cup which allowed the evolved gases to escape.Apart from the determination of Zn in a sewage sludge sample accurate results were obtained. 1.1.3.5. Thermospray introduction. The silica T-tube micro- atomizer device described for HPLC applications in the pre- vious update (see J.Anal. A t . Spectrom. 1994 9 213R) has now been used (95/303) for the determination of Hg Pb and Zn as well as the Cd reported previously. The HPLC effluent was vaporized by a thermospray into an 02-H2 diffusion flame maintained in a heated combustion chamber. A number of operating parameters were optimized and it was found that the optimum conditions for all four elements were similar. Only the Hg response depended on the composition of the mobile phase that for methanol being 6-times greater than that for water. 1.1.3.6. High-pressure nebulization. Berndt and coworkers continue to publish applications of the hydraulic high pressure nebulizer (HHPN) device (94/1605 94,4644 9413280 95,418). In this device the solution is nebulized by pumping at about 2.5 ml min-' through a 20 pm nozzle.This requires a pressure of about 17 MPa and results in nebulization efficiencies of up to 50% with consequent improvements in sensitivity. The system is also more tolerant of high dissolved solids than conventional nebulization. Procedures for the determination of Cr"' and CrV1 incorporating on-line separation with an HPLC column and preconcentration of CrV' were described (94/3280). A very similar experimental set-up was also used for the preconcentration of a number of trace elements from aluminium and in drinking water samples (94/1644). For the aluminium matrix the metals were retained as their APDC complexes on a C-18 solid phase extractant. Detection limits for Au Bi Cd Co Cu Fe Ni Pb T1 and Zn ranged from 0.1 to 1 pg g-'. For trace elements in drinking water (including aluminium) the metals were retained as their quinolin-8-01 complexes.The eluent was methanol in both cases. Improvements in detection limits compared with those for conventional nebulization of between 1 and 2 orders of magni- tude were obtained. The HHPN has also been used as an interface for coupling size-exclusion chromatography with FAAS detection (95/118). The possibilities of the instrumen- tation were demonstrated by the chromatography of ferritin cytochrome C metallothionein and cyanocobalamin with element specific detection. The system was able to handle aqueous solutions as mobile phases with ionic strengths up to 0.1 mol 1-'. The system requires enough pressure from the pump so that after the pressure drop due to the column there is enough pressure to drive the nebulizer.Even at a flow rate of 0.5 ml min-' the nebulization efficiency was sufficent for detection limits of between 5 and 15ng in a 5011 sample volume. A second paper in the series (see J. Anal. At. Spectrom. 1994 9 213R) describing the performance of the single-bore high pressure nebulizer (SBHPN) has been published (Todoli et al. Spectrochim Acta Part B 1993 48 1461). The performance of the device with discrete introduction (200 pl) was studied. Five solvents (water methanol ethanol 2-propanol and 1-butanol) were used as both solvent for the analyte solution (20 ppm Mg) and for the carrier. Two configurations were used; a vertical downward mounting with a Scott-spray chamber and a horizontal mounting with the AA spectrometer's standard spray chamber.As might be expected from the work of Berndt on the HHPN the device performs satisfactorily at low flow rates and with high concentrations of dissolved solids thereby showing potential as an interfacing device for HPLC. As for the previous work the performance of the SBHPN was com- pared with that of a standard Meinhard nebulizer rather than that of a conventional AA nebulizer. The concentration of the 204R Journal of Analytical Atomic Spectrometry August 194'5 Vol. 10test solution used strongly suggests that the SBHPN has significantly poorer sensitivity than a conventional nebulizer. 1.1.4. Sample pre-treatment Although many publications which feature AAS and other atomic spectrometry techniques appear in the original litera- ture because of some novel aspect of the sample pre-treatment procedure only a relatively small number are surveyed in this section. Papers in which the novelty resides in the particular combination of analyte(s) and matrix are not considered relevant to this update.Details of this kind of work may be found in the relevant subject-based updates. The material included in this section has both some novelty about the pre- treatment chemistry and the potential for adaptation to an on-line format. Chinese workers (94/2298) described a co-precipitation pre- concentration procedure for the determination of Cd Cu and Pb in water samples by precipitation of the hydroxides with magnesium hydroxide and re-dissolution in nitric acid. Detection limits of 4 5 and 17 ppb respectively were obtained.For the preconcentration of Cu Fe and Pb from dialysis solutions (94/1665) the APDC complexes were retained on an Amberlite XAD-4 column (60-80 mesh) at 10-15 ml min-'. The analytes were eluted with 1 moll-' nitric acid in acetone. In the determination of Cu Fe and Zn in powdered milk (95/349) the analytes were retained from 6mol 1-' hydro- chloric acid solution as the chloro complexes on an anion exchange column and eluted with dilute acid. Retention on a column of cotton macerated for 24 h with 4 mol 1-' hydro- chloric acid formed the basis of a method for the determination of Cd Cu and Pb (94/1624). River sea and bog water have been analysed (94/2266) for 15 elements collected simul- taneously with trie th ylene tetraminepentaacetate functionali ties immobilized on cellulose.The system had distribution coefficients in the range 104-105ml g-' (pH3-8) and fast exchange kinetics so that even with contact times of less than 1 s collection efficiencies of 88-99.5% were obtained. A column containing silanized diatomite modified with dioctyl tin dichloride was used in the speciation of As in potable and waste waters (94/3329). At pH 2.5-3.5 AdV) was retained whereas As"' was not. Arsenic' was eluted with 2mol 1-1 hydrochloric acid solution. It was found that columns with sorbed AsV could be stored for at least a week without any change thus allowing field sampling with elution and determi- nation in the laboratory at a later time. Polyether foam was used (95/286) to retain Ag and Au from cyanide solutions followed by elution with thiourea solution.A similar procedure using polyurethane foam has been developed (94/3 175) for the determination of Ag Au Cd and T1 in mineral samples. After dissolution in acid the addition of potassium iodide solution and ascorbic acid 50ml of the sample solution was shaken with 200 mg of the foam loaded with IBMK. After separation the metals were eluted with 10 ml of 0.1 mol 1-' hydrochloric acid containing 2% thiourea. The detection limits were 4 40 8 and 85 ng MI-' respectively. Preconcentration in a liquid membrane has been applied (95/135) in the determination of Zn in beverages. The sample was mixed with P507-N,05-liquid paraffin-n-heptane (5 + 3 + 6 + 86) acetate buffer at pH 4.7 and centrifuged after adjusting the ionic strength with sodium chloride solution.The inner phase was de-emulsified by application of a high voltage static electric spark (3000 V for 10 min) diluted and sprayed into an air-C,H flame. The detection limit was 6 ppb. A jotation procedure for the separation of Cu as the sulfide with oleic acid has been devised (95/39). The method was relatively free from interferences from concomitant ions and ionic strength effects. An indirect method for the determination of fluoride has been developed for the analysis of workplace air and waste water (94/2275). The sample solution was passed through a column containing granular lead zirconate-titanate and the leached lead determined at 670.8 nm. Fluoride concentrations between 20 and 20mg 1-' were determined with precisions around 9% RSD.The Pd content of a catalyst has been determined by electrolytic collection on a copper cathode followed by insertion of the cathode into an air-C,H flame (94/2294). 1.1.5. Chromatographic detection As something of a contrast to the previous review period when most of the publications featured GC methods this year's review period is concerned mainly with LC methods. Several overviews of both techniques have appeared concerned with coupling to a variety of spectroscopic techniques including atomic spectrometry (95/126) the use of element specific detection in GC (95/41) the use of micellar or vesicular mobile phases (94/3036) and the use of GC techniques for the determination of organolead compounds (95/1). A variety of coupling devices have been described ranging from the simple connection of the HPLC effluent to the nebulizer of a flame spectrometer with a short length of heat shrink tubing (94/2985) to a thermospray pyrolysis chamber flame-in-tube atomizer (95/303).As described earlier (sec- tion 1.1.3.6.) the operations of two high pressure nebulizer devices have been described (Todoli et al. Spectrochim Acta Part B 1993 48 1461 94/1605 95/118). Only the potential of the device described in the first of these papers was postulated as a result of studies of the FI behaviour of the system with organic solvents. Several size-exclusion chromatographic appli- cations (95/118) were used to illustrate the performance of the hydraulic high pressure nebulizer. Biologically derived species of Co Cu Fe and Zn were determined.The device was also used (94/3280) in the ion-pair chromatographic separation of Cr"' and Cr". A 5 cm column containing 5 pm C-18 packing was used with tetrabutylammonium phosphate-phosphoric acid as the eluent. A pump capable of producing 40 MPa was used which provided enough pressure after the column to drive the nebulizer (about 17 MPa is required for a flow rate of 2.5 ml min-I). Detection limits of 30 and 20 ppb for Cr"' and Crvl respectively were obtained. Gel permeation separation of free and bound species of Ag Cd Cu and Zn in lobster digestive gland extracts followed by either flame or graphite furnace detection has been described (95/98). However with an elution rate of 15 ml min-' the separation was surely not directly coupled to the spectrometer.The thermospray device (95/303) discussed earlier in section 1.1.3.5. was also used for detection following gel permeation separation of metal species from aquatic organisms. In this case Cd Cu and Zn metallo- thioneins from a fresh water mollusc were the target analytes. Post-column chemical vapour generation has been implemented in the determination of As species following LC separation (94/2926 94/3036). In the first of these procedures (94/2926) the species were separated on a Hamilton PRP-X 100 column (with an upstream Waters IC-H anionic cartridge when the cationic species arsenobetaine and arsenocholine were to be determined) by elution with a solution of sodium dihydrogen phosphate and disodium hydrogen phosphate. The eluent was merged with 5% m/v potassium persulfate in 2.5% m/v sodium hydroxide solution (for the cationic species only) and passed through 1.5m of 0.5mm id tubing in a 700 W domestic microwave oven.After cooling in an ice-bath the effluent was merged successively with hydrochloric acid and tetrahydraborate solution before entering a Philips V-tube gas- liquid separator. For a 100 pl injection volume detection limits of between 0.3 and 0.9 ng for the two cationic species and four anionic species (arsenite arsenate dimethylarsinate mono- methylarsonate) were obtained. A vesicular mobile phase has Journal of Analytical Atomic Spectrometry August 1995 Vol. 10 205Rbeen used to separate the anionic species (94/3036) with post- column chemical vapour generation of the volatile derivative directly from the organized medium.Studies of the GC separation of lead alkyls with detection by AAS with a quartz tube atomizer (95/13) and of the tin species liberated as pyrolysis products from solid samples including PVC (94/C1951) have been described. 1.1.6. Plasma atomizers An ultrasonic nebulizer has been used (94/3304) for sample introduction into an MIP as a source for AAS (see J. Anal. At. Spectrom. 1994,9,213R). The resulting aerosol was desolvated with a heating tube and a combination of water-cooled con- denser and concentrated sulfuric acid desiccator. The atom source replaced the burner in a conventional FAA spec- trometer. The characteristic concentration for 12 elements under optimized conditions ranged from 1 (Ag Zn and Cd) to 30 ppb (Ni).Interference effects due to stable compound formation and ionization suppression similar to those of the air-C,H2 flame were observed and the authors concluded that MIP-AAS is similar to FAAS in terms of sensitivity and operating cost. Some fundamental studies on an rf glow discharge atomizer have been reported (94/2216). It was found that for sputtering from oxygen-free Cu the GD plume was constricted by an increase in the argon pressure. It was concluded that the GD source was stable enough to allow sequential AA determinations and depth profiling of metal alloys. Chinese workers have reported on the determination of Cd Cr Cu Fe Mg and Mn by sputtering from a graphite cathode into a glow discharge atomizer operated in a transient mode (94/1628). Marcus (94/2057) has surveyed the use of glow discharge sources for AA and AF.1.2. Electrothermal Atomizers There have been few major developments within the field of electrothermal atomizers during the period covered by this review. With respect to instrumentation the commercial intro- duction of a new simultaneous multi-element ETAAS instru- ment is a major development and is discussed in sections 1.2.5 and 1.4 of this Update. Recently a product overview of ETAAS appeared in Analytical Chemistry (Anal. Chem. 1995,67 51A). Although fairly non-technical it does neatly summarize the development of ETAAS instrumentation in a few pages and concludes that ‘with ETAAS you get a mature technique for a modest investment’. Even though ETAAS is considered a mature technique both the previous Update (94/3317) and the biennial review by Jackson and Mahmood (Anal.Chem. 1994 66,252R) indicated that the technique is still a thriving research area and that there are fresh opinions and ideas appearing in the literature. However apart from these publications there appear to have been no other major reviews published in the field of ETAAS recently. In an inter-laboratory comparison of instruments for the determination of elements in acid digestates of solids Kimbrough and Wakakuwa (Analyst 1994,119,383) appeared to suggest that for environmental analysis ETAAS is no longer a viable technique compared with ICP-AES and even hydride generation AAS is obsolete. However close reading of this paper reveals that at high concentrations ICP-AES is the preferred technique whereas at low concentrations and when matrix interferences are suspected ETAAS is preferred.These conclusions were based on a consideration of the determination of As Cd Mo Se and TI in five soil samples with concen- trations ranging from 5000 to 5 mg kg-’ well above concen- trations that would be considered appropriate for ETAAS. In addition no information was given as to the quality control procedures adopted by the laboratories that produced data for this study which would provide the reassurance that the laboratories were in control before submitting results to the study. For those who have an interest in both the efficiency of detection and efficiency of measurement in atomic spectrometry the paper from Winefordner et al. (see J.Anal. At. Spectrom. 1994 9 131) is recommended reading. This paper discussed how these figures of merit affect the limit of detection and limit of guarantee of purity. The 18 concluding paragraphs concisely assessed every major elemental analytical technique with respect to efficiency of detection and measurement. The possibilities of each technique what should theoretically be achievable and how all the techniques relate to each other were admirably presented. This paper is essential reading for every atomic spectroscopist who has an ‘axe to grind‘ or who holds exaggerated views about the capabilities of his/her own particular technique. 1.2.1. Atomizer design and surface modijication 1.2.1.1. Graphite atomizers. A few reports have been pro- duced during this Update period though generally these were on variations to commercially available atomizer designs rather than anything radically new.Katskov et al. (94/2950 94/2595) discussed the use of Massmann-style graphite tube atomizers containing a graphite Jilter. Within a conventional pyrolytic graphite coated electrographite tube was positioned a tubular graphite insert around which was wrapped 50 mm of graphite thread. The sample was pipetted into the cavity between the tube wall and the tubular insert. When vaporized the atomic vapour diffused through the walls of the filter into the light path of the spectrometer. The performance of the graphite filter atomizer was assessed by measurements with Al Bi Cd and Cu and Bi Cd and Pb in the presence of NaCl and CuC1 matrices.The advantages of such a design were an increase in signal magnitude of the order of 1.6-2.8-fo1dy the ability to increase the sample volume to 100 pl and at the same time reduce the drying period for such a volume to 15s and attainment of a lower level of spectral background and chemi- cal interferences without the need for chemical modification. The graphite filter atomizer was successfully applied to the determination Cd and Pb in whole blood and steel. However some anomalous results discussed in the paper indicated that memory effects were a problem and that the technique needs to be refined. In further work concerning high-temperature metal-uapour elution Kitagawa et al. (95/324) developed an electrothermal atomizer which consisted of a long glassy carbon tube packed with activated charcoal. Solid samples of biological materials were inserted into the atomizer which was maintained at a constant temperature of 2280°C and the atomic vapours of Cu Mg and Mn passed through the activated charcoal with a stream of argon into the light beam of the spectrometer.The advantage of this was the absence of any background absorp- tion and isothermal operation. Streck et al. (94/C3451) considered the use of integrated contact cuvettes (ICC) with either no platform a forked platform or a ring platform for use in a transverse heated graphite atomizer. The system was applied to the determination of Pb in urine and Cr and Se in serum. No indication was given as to which type of ICC-platform combination was employed for these determinations.Wall platform and probe atomization were compared by Cimadevilla et al. (94/2401) for the determination of Cr in biological fluids with and without the use of a mixed chemical modifier of calcium and magnesium nitrates. The best con- ditions were found to be atomization from the wall of pyrolytic graphite coated electrographite tubes and use of the chemical modifier for only water and serum samples. Detection limits 206R Journal of Analytical Atomic Spectrometry August 1995 Vol. 10for peak absorbance measurements were 0.04 0.22 and 0.05 pg 1-1 in water serum and urine respectively. 1.2.1.2. Metal atomizers and metallic coatings. Perez and Sanz-Medel(94/2400) examined the use of the carbide-forming elements molybdenum tantalum tungsten and zirconium for the treatment of pyrolytic graphite platforms and graphite tubes for the determination of Si in biological fluids to prevent the formation of silicon carbide.Tungsten was found to give the best results with a detection limit (3s) of 3.5 pg 1-' and a precision of 2% (n= 10) at a concentration of 90 pg 1-' and with the procedure described it was not found necessary to use any chemical modifiers. Tungsten-metal platforms and atomizers continue to receive attention. Silva et al. (94/3245) used a tungsten coil for the determination of Ba in waters. A purge gas of 10% v/v hydrogen in argon was found to be essential for the atomization of Ba providing a source of hydrogen for the reduction of Ba containing oxygen species. No atomization of Ba was observed when pure argon was used as the purge gas.The lifetime of the tungsten coil depended upon the acidity of the solutions analysed and the presence of organic compounds in the sample matrix. In 0.014mol 1-l nitric acid the lifetime was found to be 400 firings. For sample volumes of 10 pl the interference from Ca above 10 ng was minimized by the use of EDTA and the system was found to tolerate 8000 ng of K lo00 ng of Mg 6000 ng of Na and 10 ng of Sr. An absolute detection limit of 2 pg and a characteristic mass of 3.6 pg were obtained from peak absorbance measurements. Good agreement was found between tungsten coil ETAAS and ICP-AES for a series of water samples. A similar tungsten-coil deoice was used by Barth and Krivan (94/3023) in conjunction with ICP-AES for the determination of Al Cr Fe Mg Mn Ni and Ti in aqueous suspensions of silicon dioxide powder.For the determination of Li in molybdenum oxide Docekal and Krivan (94/3139) used slurry sampling and a transverse heated tungsten atomizer (WETA-82); further details are given in section 1.2.2. A tung- sten-metal atomizer was used by Minamisawa et al. (95/200) for the determination of Cu preconcentrated onto chitin. Chinese workers (94/2281) described the use of a tungsten probe for the determination of Sn in drinking and mineral water. Yizai et al. (94/3104) used a tungsten-foil platform inside a pyrolytic graphite coated electrographite tube in conjunction with a palladium and tartaric acid mixed chemical modifier for the determination of Cd in sea water and a variety of biological and environmental RMs by standardless analysis.These workers reported that the lifetime of the tungsten foil platform 200-300 firings was longer than that of tantalum foil which they used previously. Oimatsu et al. (94/2565) used a tantalum carbide coated graphite atomizer in conjunction with slurry sampling to deter- mine Sn in titanium dioxide. A tantalum platform within a transverse heated graphite atomizer was used by Skroce et al. (95/245) to determine La. This gave better sensitivity than atomization from a pyrolytic graphite surface. The detection limit (3s) was 40 pg 1-1 with a characteristic mass of 220 pg for integrated absorbance measurements. For a recommended method for the determination of A1 in serum French workers (Ann. Biol. Clin. 1992 50 577) some- what surprisingly preferred tantalum coated graphite tubes to graphite tubes.This was claimed to be due to the fact that poor quality graphite tubes can seriously affect the measure- ments. Giordano et al. (95/435) compared the use of either pyrolytic graphite platforms or titanium coated graphite tubes for the determination of Sn in tricyclohexyltin hydroxide; both performed well. Ohta et al. (94/2299) have continued their work on the use of molybdenum-tube atomizers in this case for the determination of K in biological materials. After digestion sample volumes of 1 p1 were taken into the 25 mm long 2 mm i. d. molybdenum foil tube atomizer. A detection limit of 0.12 ng was reported equivalent to 0.12 mg 1-' which appears rather high compared with flame emission measurements for this element.Ammonium molybdate impregnated graphite tubes were used by Lin et al. (94/C3379) for the determination of Ge in herbs plants and sediments. After addition of ammonium molybdate to the 0.3 mol 1-l nitric acid sample solutions the Ge was extracted as the germanomolybdate into butanol. Ammonium molybdate impregnated graphite tubes enhanced the sensitivity. Only limited work on zirconium atomizers has been reported. Yan et al. (94/C3371 95/252) used a zirconium-coated tube for mechanistic studies on the atomization of In and the atomiz- ation of Sn (94/331,94/C3371). This work is discussed in more detail in section 1.2.3. 1.2.2. Sample introduction 1.2.2.1. Slurry sampling. Every other year the numbers of papers and conference reports concerning solid and/or slurry sampling with ETAAS are given a boost by the publication of the papers from the Solid Sampling Symposium held in Germany (Fresenius' J.Anal. Chem. 1993 346). Kurfurst (94/1638) discussed the information that can be obtained from imprecise solid sampling data if these represent the heterogen- eity of the analyte in the sample. In an effort to demonstrate the reliability of slurry sampling Miller-Ihli (94/C1943) dis- cussed a current international collaborative study. The first phase focused on the determination of Cr and Pb in sediment materials. No doubt the results from such a study will go a long way to improving the reputation of slurry sampling ETAAS as a routine analytical procedure. Dobrowolski and Mierzwa (94/2580) considered some of the analytical problems associated with slurry sampling for the determination of trace elements in plant materials.They compared the effects of slurry composition and mechanical and ultrasonic slurry preparation. The use of 5% nitric acid as a solvent and ultrasonic mixing improved the stability of the slurries. Results for CRMs indicated that the slurry method was accurate and comparable to a conventional wet-ashing procedure. The conference report from Miller-Ihli considering the influence of slurry preparation on accuracy was discussed last year (see J. Anal. At. Spectrom. l994,9,213R) and has now been published (95/300). Ultrasonic slurry sampling was applied by Bendicho and Sancho (94/2886) to the determination of Se in wheat flour.A mixed chemical modifier of palladium and magnesium nitrates was used along with a transverse heated atomizer. Good agreement with the certified value for NIST SRM 1567a Wheat Flour was obtained and recoveries on five commercial wheat flour samples were in the range 92-96% with a calibration established against aqueous solutions. The repeatability of the method was 8.2% with a 5% m/v slurry (n= 10) and the detection limit (3s) was 0.06 pg g-'. The method overcame the need for acid digestion or ashing of the samples with the attendant risk of loss of Se. Krivan's research group from the University of Ulm con- tinued their investigations into the suitability of slurry sampling for the determination of trace elements in refractory materials. From this group Hauptkorn et a!.(94/2955) applied slurry sampling into a transverse heated atomizer equipped with an automated ultrasonic mixing accessory to the determination of Si in titanium dioxide and zirconium dioxide powders. Calcium nitrate was found to be the best chemical modifier because with magnesium nitrate (generally recommended for Si) poor signals were obtained from the titanium dioxide matrix. The retention of calcium and zirconium in the atomizer was examined with the aid of 47Ca and 97Zr as radiotracers. Even though the retention of zirconium increased linearly with the number of determinations this did not have any effect on the Si signal and up to 300 firings could be achieved before Journal of Analytical Atomic Spectrometry August 1995 Vol. 10 207 Rany signal degradation occurred.Calibration was performed by the methods of analyte additions using aqueous standards. The limits of detection (3s) were 7 and 2pg g-' in titanium and zirconium dioxide respectively. Good agreement with results from d.c. AES slurry ICP-AES and ICP-AES after sample digestion was obtained. However comparison of the results obtained allowed an estimation of the heterogeneity from the ETAAS results due to the larger standard deviations found. Hauptkorn and Krivan (94/3363) determined Si in boron nitride by slurry sampling with a magnesium nitrate chemical modifier. Calibration was established either by the method of analyte additions or with a calibration curve prepared from aqueous solutions. For the determination of 20 pg g-' of Si the RSDs were 8.3 and 10.5% for the method of analyte additions or a calibration curve respectively.The detection limit (3s) was 3 pg g-' and good agreement was obtained with results from XRF spectrometry. Docekal and Krivan (94/3 139) have continued their work on slurry sampling for the determination of various elements in molybdenum oxide (see J. Anal. At. Spectrom. 1994 9 213R). In previous work poor detection limits were obtained for Li. This has now been improved by the use of a transverse heated tungsten atomizer (WETA-82) and the detection limit (3s) was improved to 2 ng g-' compared with 40 ng g-' for a graphite atomizer. These same workers (94/1612) described a procedure to remove the molybdenum matrix from a graphite atomizer after such analyses using a purge gas of CF4 after the atomizer clean-out step.The procedure described restored the signals for Co Cr Cu Fe Mg Mn and Ni to their original values and extended the analytical lifetime of the graphite tube to 200-300 firings. Oimatsu et al. (94/2565) used a tantalum carbide coated graphite atomizer in conjunction with slurry sampling to determine Sn in titanium dioxide. The powdered sample (0.5 g) was mixed ultrasonically with 2.5 ml of 0.1% m/v sodium hexametaphosphate and diluted to 50ml with water. After shaking for 30 min aliquots of the slurry were taken for analysis along with lop1 of 20% m/v ascorbic acid as the chemical modifier. A detection limit of 0.15 ng was reported (though no indication was given of the sample volume) and the method was claimed to eliminate most of the interferences from the matrix and produced results in agreement with those obtained by hydride generation AAS or ICP-MS after acid digestion.Lopez Garcia et al. (94/3286) have continued their work on slurry sampling ETAAS for the determination of Cd Mn and Zn in high silica content diatomaceous earth. The same procedure as that described previously (see J. Anal. At. Spectrom. l994,9,213R) was employed. Good agreement with results obtained by acid dissolution of the samples was found but the precision was improved by slurry sampling. For 0.1 pgg-' of Cd 36pg g-' of Mn and 9 pg g-' of Zn precisions (n=7) of 7.8 2.8 and 3.3% RSD were obtained respectively. Chinese workers (95/457) determined Pb in soil by suspending the powdered sample after passage through an 180-mesh sieve in 0.2% agar solution.Russian workers (95/445) determined Cd and Pb in soil suspensions after ultrasonic mixing with 1% v/v nitric acid. The determination of Pb in marine sediments by Bermejo-Barrera et al. and discussed last year (see J. Anal. At. Spectrom. 1994 9 213R) has now been published (94/2956). These workers (94/C3465) applied a similar procedure for the determination of As"' and AsV species in marine sediment by slurry sampling ETAAS. The As'" was determined by a prior extraction method using NaDDC and 4-methylpentan-2-one before injection into the graphite atomizer using palladium as the chemical modifier. The AsV was determined by difference from the total As determined by slurry sampling and As"'. The precision and accuracy of the method were studied using PACS-1 (Marine Sediment) RM.To avoid the time-consuming digestion pro- cedure for the determination of As in plant materials Kukier et al. (95/373) examined the use of automated slurry sampling into a transverse heated graphite atomizer. The results were in agreement with those from hydride generation after a nitric acid-perchloric acid digestion. These workers concluded that this illustrated that a slurry sampling ETAAS procedure could eliminate the time and labour involved in the plant digestion step. For the determination of Cu in glass samples Morita et al. (94/2563) mixed 1-30 mg samples of powdered glass with 5 ml of 10% glycerine solution and applied ultrasonic agitation for 30 s. From this homogeneous suspension 10 pl aliquots were taken into the graphite atomizer along with 10 pl of 5% m/v ammonium fluoride as the chemical modifier.The precision (n=6) for 11.5 pg g-' of Cu was 2.4%. Chinese workers (94/3096) developed a magnetic auto- stirrer which was designed to be mounted on the sample tray of a Perkin-Elmer AS-40 autosampler. The system was applied to the analysis of plant food coal fly ash and geological samples. An underwater spark system was applied by Bendicho (94/3339) to produce colloidal suspensions from electrolytic iron samples for the determination of Cr Cu and Mn impurit- ies. The system was operated at 75 W for 30-60 min and the amount of sample dispersed found by weighing the rods before and after the spark procedure. The colloidal suspension was analysed by ETAAS. The procedure was evaluated by examin- ation of NIST SRM 665 Electrolytic Iron and results in agreement with the certified values were obtained.Japanese workers (94/3306) determined Pd by adsorption onto an anion-exchange resin containing an immobilized sulfonic acid derivative of dithizone at pH 1-10 for 1-15 min. The Pd-containing resin was suspended in water and the suspension injected into the graphite atomizer. The detection limit for Pd" was 5 pg 1-' and the procedure was applied to the determination of Pd" in natural waters. The use of micro-organisms to preconcentrate elements fol- lowed by slurry sampling of the residue has been examined by Robles et al. for the determination of Au (94/2189) Be (94/3247 94/C3498) and Cd (94/C3436). Preconcentration with Escherichia coli or Pseudomonas putida either living (batch) or dead (immobilization) was examined.The optimum procedure for all elements appeared to be batch preconcentration with E . coli at a pH of 6-9 followed by centrifugation and suspension of the bacterial mass in 3.5 moll-' nitric acid and introduction of the slurry into a graphite atomizer equipped with a platform. For 0.1 pg 1-' of Au and 10 pg 1-' of Be the precisions (n=9) were 7 and 3% respectively and detection limits (3s) of 0.004 and 0.05 pg 1-' were found respectively. For Cd a detection limit (3s) of 0.04 pg 1-' was found. For Au and Be little interference from major and minor elements of environ- mental interest was found. In spite of this the method appears to have little advantage over other procedures as the precon- centration is subject to so many variables such as pH solution volume to bacterial mass ratio the bacterial mass used in preconcentration and interferences by elements such as Ag Ca and Fe.For the determination of B in cell suspensions Papaspytou et al. (94/3026) used magnetic stirring of the cell suspensions in the autosampler cups. To prevent clogging of the autosampler injection capillary a slightly larger capillary was employed. This method was an improvement over other procedures because it eliminated a time-consuming digestion procedure and any risk of losing B. The method was developed in conjunction with studies concerned with boron neutron capture therapy for tumour treatment. Calibration for B was by the method of analyte additions using boric acid solutions and the between-batch precision (n = 6) for 0.57 pg mg-' of protein was found to be 5.7%.The limit of detection (3s) was 208R Journal of Analytical Atomic Spectrometry August 1995 Vol. 100.05 mg 1-' of cell suspension. This compared well with the detection limits achieved by DCP-AES and ICP-AES. 1.2.2.2. Solid sampling. In another example of turning one of the disadvantages of solid sampling into an advantage Pauwels et al. (94/1631) determined the homogeneity of Cd in four polyethylene CRMs with nominal Cd concentrations in the range 40-400 mg kg-' with solid sampling ETAAS. From each of the four CRMs 60 analyses were performed with sample masses ranging from 60 to 250pg. Leuker et al. (94/258 1 94/2582) applied solid sampling Zeeman ETAAS to assess the distribution of Cd and Pb in mallard (94/2581) and equine (94/2582) livers. With respect to the mallard livers Cd in the concentration range 0.09-2.4 ng g-' was found to have an increased degree of heterogeneity though for Pb the variance due to heterogeneity was distinctly lower than the residual experimental error.For equine livers the analysis of variance showed a negligibly low influence of heterogeneity on the results. In both papers the importance of a proper sampling strategy for solid sampling procedures was emphasized. In general these workers recommended the analysis of at least six microsamples from different parts of a liver in order to reduce the influence of minor heterogeneity. Using such a strategy results by solid sampling agreed well with those obtained by conventional sample decomposition procedures after sample homogenization. The application of solid sampling to the determination of Si in gold discussed last year (see J.Anal. At. Spectrom. 1994 9 213R) has now been published (94/2953). Hinds and Kogan (94/2953) compared solid sampling ETAAS with aqueous calibration solution-based ETAAS and ICP-AES procedures with matrix matched calibration solutions and spark ablation ICP-AES. The first three procedures gave comparable results though solid sampling ETAAS was more error prone. However detection limits were of the order of 3 pg g-' for all the four procedures. While spark ablation ICP-AES is somewhat more sensitive and convenient the absence of gold reference mate- rials for calibration which up till recently have not been available limits its use.Russian workers (94/3 147) applied solid sampling ETAAS to the determination of Al Bi Cd Co Cu Fe Mn Pb Si and Sn in high-purity tantalum pentoxide. The direct solid introduction for the determination of Ag Cd Cu and V in fly ash using pressure regulated ETAAS discussed previously (see J. Anal. At. Spectrom. 1993 8 197R) has now been published (94/2579). Calibration for Cd and Cu was against a calibration curve established with aqueous solutions though for V the method of analyte additions had to be used owing to analyte expulsion by the fly-ash matrix. The pro- cedures were verified by analysis of NIST SRM 1633a Coal Fly Ash and good agreement was obtained with the certified values.1.2.2.3. Gas sampling. Though not directly concerned with solid or slurry sampling work continues in a few groups on the direct sampling of air into a graphite atomizer for the purposes of assessing airborne contamination due to particu- late matter. Sneddon (92/C3369) discussed the use of a single- stage impacter connected to a graphite atomizer. The air or atmosphere was pulled through the atomizer by vacuum and collected directly on the graphite surface. Sampling for several minutes followed by the analysis stage of a few minutes allowed a near real-time determination of metals in air. This was applied to the determination of Cd Cu Hg and Pb in a laboratory atmosphere. Similarly Liidke et al. (94/3 105) drew air through the wall of a porous graphite tube to collect particulates on the inner surface.This tube was then used as the graphite electrothermal atomizer in furnace atomization non-thermal excitation spectrometry (FANES). Allied with the use of an Cchelle polychromator simultaneous multi-element determinations were possible. Detection limits ranged from 0.1 to 1 ng m-3 for Ca Cd Cr Cu Fe Ni Pb and Sr. Flow rates through the graphite tubes ranged from 140 to 200dm3 h-' and air volumes of between 1.1 andl.5 m3 were sampled. Given that Ni had the worse detection limit a minimum sampling time of 2 h was required to achieve a reliable determination at eight times the detection limit. There was satisfactory agree- ment with results obtained previously by established methods. Baaske et al. (94/3246) described a novel procedure for the determination of trace amounts of Fe in gaseous hydrogen chloride with a modified ETAAS instrument.Pyrolytic graphite coated electrographite tubes designed for use with a probe atomizer were employed. This allowed a gas sample to be injected into the atomizer via the probe slot in the tube using a graphite capillary attached to a gas-tight syringe. Gas sample volumes of 0.1 ml were injected into the atomizer with no internal gas flowing at a temperature of 1100°C followed immediately by atomization at 2300°C. The use of integrated absorbance measurements allowed calibration by means of standards prepared in hydrochloric acid purified by sub-boiling distillation. A detection limit (3s) of 0.7 ngml-' was found and the Fe content of hydrogen chloride calculated to be 50.1 ng ml-' with a precision of 4% ( n = 6 ) .1.2.2.4. Coupled techniques and preconcentration. For the coupling of FI techniques to ETAAS the group of Welz and Sperling have been very active. These workers have applied automatic on-line column preconcentration and separation to ETAAS (94/1636) for the determination of Cd and Pb in high- purity reagents. Following retention on a microcolumn con- taining 9 pl of solid sorbent and the introduction of air to remove the wash solution the analytes were eluted quantitat- ively with 80 pl of methanol. The total eluate volume was transferred into the graphite atomizer which was pre-heated to 80°C using a flow rate of 0.08 ml min-' which allowed the partial evaporation of the solvent during eluate introduction.The enrichment factors were 62 and 64 for Cd and Pb respectively compared with the direct introduction of 30 pl of an aqueous solution. The sample throughput was 13 h-' for a sample loading time of 60 s and the detection limits (3s) were 0.7 and 4.5 ng 1-' for Cd and Pb respectively. The precision of the procedure in the optimum working range was approximately 3% (n = 6 ) . A number of high-purity and analyt- ical-grade reagents were analysed and recoveries were of the order of 97-104%. Wang et al. (94/3073) described the determination of ultra- trace amounts of Cd and Cu in sea-water with FI microcolumn preconcentration coupled with ETAAS. These workers used a manual flow injection procedure followed by collection of 80 pl of ethanol eluate into an autosampler cup from which it was loaded into the furnace in two 40 pl aliquots. It was claimed that a new type of C18 column loaded with NaDDC was used to extract Cd and Cu from sea-water as DDC chelates however the procedure seems very similar to that employed by Sperling et al.(J. Anal. At. Spectrom. 1991,6,295). Although the procedure required only small volumes of sea-water 400 and 600 pl for Cd and Cu respectively compared with 3000 pl taken by Sperling et al. lower preconcentration factors of 10- and 15-fold for Cd and Cu respectively were found. Detection limits (3s) for the preconcentration from aqueous solutions were 0.004 and 0.024 pg 1-' for Cd and Cu respectively. Results for the determination of Cd and Cu in CASS-2 Nearshore Sea Water RM showed no significant difference between the certified values and experimental results.Liu and Huang (94/1616) also determined Cd and Cu in sea-water by automatic on-line preconcentration in the autosampler capil- lary tip. A silica gel CI8 microcolumn was mounted near the tip of the autosampler capillary. The preconcentration pro- cedure was performed by using the pumps of the autosampler Journal of Analytical Atomic Spectrometry August 1995 VoE. 10 209 R(a Perkin-Elmer AS-40) and a four-way valve controlled by a programmable controller. The chelating agent APDC was added to the sample solutions and only 0.2-2ml of sample were required to determine Cd and Cu in sea-water. Detection limits were found to be 1.26 and 6.5ng I-' for Cd and Cu respectively.Heng-bin et al. (94/2 198) applied %-atomizer trapping' within a graphite atomizer coated with palladium to determine As in fractions eluted from a Dionex CASl bifunctional column to determine various As-containing species in lake and river water. From 0.1 ml samples and using a mobile phase of 5 mmol 1-l ammonium dihydrogen phosphate at pH 5.8 the first 2 ml of eluate were discarded and then 40 fractions of ten drops each were collected and analysed. Recoveries were >96% and detection limits for the different species ranged from 1.6 to 1.9 pg 1-' Other Chinese workers (94/2411) applied 'in-atomizer trapping' with an FI chemical vapour generator coupled to an electrothermal graphite atomizer to determine Hg in environmental samples. For the determination of Se in fruit juices Arruda et al.(94/2981) applied an automatic FI method coupled with ETAAS. The flow manifold used which was coupled to the autosampler cup enabled dilution of the slurry addition of a chemical modifier and the determination of Se in the filtered liquid phase. In fruit and tomato juices diluted ten-fold Se was successfully determined using the method of analyte additions with a detection limit (3s) of 5 pg I-'. Kuo et al. (94/2697) increased the sample loading within a graphite atomizer by automatically pipetting and drying nine multiple 90 p1 aliquots of the samples for the determination of Co Cr Cu Fe Mn and Ni. Detection limits were 9 8 4 6 3 and 9 ng 1-' for Co Cr Cu Fe Mn and Ni respectively. While these are certainly impressive detection limits this performance was achieved at the cost of 21.8 min being required for a single measurement.The use of a hydraulic high-pressure nebulizer sample intro- duction system for ETAAS was discussed previously (see J. Anal. At. Spectrom. 1992 7 215R) but has now been published by Berndt and Schaldach (94/2215). The samples were introduced into an aerosol deposition module equipped with a hydraulic high-pressure nebulizer whereby the aerosol was deposited within the graphite atomizer tube maintained at a temperature of 140 "C. Using such a high-pressure system the efficiency of aerosol transport was increased with maximum aerosol yields of the order of 60%. As a high-pressure flow system is required for this type of nebulization on-line matrix separation and preconcentration procedures directly linked to ETAAS are possible.The technique was applied to the determi- nation of Au Co Cu and Pb complexed with APDC within an aluminium chloride matrix. Detection limits of 0.019 0.17 0.021 and 0.05 pg g-' were obtained for Au Co Cu and Pb respectively. Chinese workers (95/282) also appear to have examined pneumatic nebulization for sample introduction for ETAAS though few details were given in the abstract. 1.2.2.5. Electrodeposition. There is still a low level of interest in the use of electrodeposition for separation of analytes from appropriate matrix solutions. Matousek and Powell (94/3037) described a procedure whereby a 50 pl sample volume was pipetted into the V-shaped well of a pyrolytic graphite platform. The solution was electrolysed at an uncontrolled potential of 4-6 V with the platform as the cathode and a platinum wire of 0.5 mm diameter positioned 0.63 mm from the platform as the anode.The rate of removal of Pb from the solution was monitored by withdrawing pl samples and analysed using differential pulse anodic-stripping voltammetry. After removal of deposited Pb with 3% v/v nitric acid and drying the platform was ready for the next sample. Complete deposition of the Pb was achieved in 4 min at 6 V for concentrations of Pb of up to 500yg 1-'. This rapid rate of deposition was considered to be due to the convective stirring caused by gas evolution. A single platform could be used for up to 50 electrolyses. Sadly even from the original paper it is difficult to work out exactly what is going on in this process.Presumably the platform is present within the atomizer after electrodeposition the matrix containing solution must be removed from within the atomizer leaving the electrodeposited Pb on the platform. If the Pb is then determined by atomiz- ation why is it necessary to remove the Pb by washing the platform? These workers appear to have developed a fully automated system for loading the samples electrolysis electro- lyte removal washing and pre-treatment of the deposit prior to ETAAS (94/3037) and for which a patent has been applied (94/3083). The other worker active in the field of electrodeposition is Beinrohr (see J. Anal. At. Spectrom. 1992 7 215R and 1994 9 213R) who has taken a different approach to developing a flow-through electrolytic cell.In the most recent work Beinrohr et al. (94/1643) determined Pt in tobacco beans slag and dust samples after a digestion procedure. Platinum from the sample solutions was deposited at -0.9 to 1.2 V uersus an SCE in a flow-through system incorporating a three-electrode flow-through cell with a graphite counter and Ag-AgC1 refer- ence electrodes. The working electrode was reticulated vitreous carbon packed into a pyrolytic graphite coated electrographite tube. The reticulated vitreous carbon filled the tube and had a hole bored through to allow the light beam to pass through the tube for ETAAS measurements. During the preconcen- tration step this hole was plugged with a glass rod. After electrodeposition the tube was washed and placed into the atomizer and an atomization temperature of 2700 "C applied.For a 2 g sample digested and diluted to 25 ml and with 5 ml taken for electrodeposition the detection limit was found to be 1 ng g-'. Samples containing Fe"' required reduction with ascorbic acid before the deposition step. The results obtained with this procedure agreed with those obtained by a micro- column preconcentration procedure followed by ETAAS. However it has to be said that this cannot be the fastest procedure ever developed and remains a manual procedure. Pannier et al. (94/1690) developed a method to determine butyltin species by ion chromatography linked on-line with ETAAS. Few details were given as to how this was achieved though mono- di- and tributyltin could be determined in a single experiment with detection limits of 0.5 1.1 and 0.8 ng of Sn respectively.Koelbl et al. (95/101) discussed the use of computerized data treatment for an HPLC-ETAAS system for the identification and quantification of trace element compounds. This is needed as the quantification of the dis- continuous signals from an ETAAS instrument defining a chromatography band is difficult. 1.2.3. Fundamental processes Judging by the absolute lack of reports considering A1 'spikes' and the reduction ofoxides by carbides (ROC) theory it would appear that the ROC theory or mechanism has been discarded. However that does not mean that A1 'spikes' have been forgotten! The main participants in the discussion now have new theories to account for the phenomenon. L'vov (95/244) commented that the condensation scheme developed by himself in conjunction with Frech (see J. Anal.At. Spectrom. 1994 9 213R) can be applied to interpret the observations of Gilmutdinov et al. (J. Anal. At. Spectrom. 1992 7 675) and Chakrabarti et al. (Anal. Chem. 1993 65 716) in particular the doughnut structure of the cloud produced by evaporation of pg amounts of A1203. L'vov considered that the doughnut phenomenon correlated well with the condensation concept and that condensation of either metallic A1 or A1203 is 2 t 0 R Journal of Analytical Atomic Spectrometry August 1995 Vol. 10occurring at the cooler tube ends and also around the tube injection hole. L’vov questioned whether the partial pressure of oxygen in the tube is comparable to that of A1,O which is necessary for the reaction to A1203 to proceed hence it would appear that condensation of metallic particles is favoured.A series of papers has appeared from Gilmutdinov’s group concerned with the application of shadow spectral Jilming to investigate the non-steady-state structure of atomic and mol- ecular absorbing layers in ETAAS. These reports concerned the atomization of Cd Hg and Zn (94/3040); Cu Fe and Mn (94/3146); Ga and In (95/255); Bi Ge and T1(95/256); and A1 (95/260). While these are all in Russian it is hoped that English translations appear soon. It is clear that shadow spectral filming and the work of Gilmutdinov’s group has had a major impact on our understanding of what is happening inside a graphite electrothermal atomizer. While it is still not fully understood there is a better appreciation of the atomic and molecular distributions than there was 4-5 years ago in at least that they are complex and often element specific. The cross-sectional distribution of Cd Hg and Zn (94/3040) under interrupted gas flow conditions at the atomization stage showed a weak inhomogeneity with a maximum near the sample in the atomizer and the structure appeared to be mainly conditioned by physical factors.The longitudinal non- uniformity of the absorbing layer in the graphite atomizer could be increased by gas expulsion from the dosing hole. For Cd and Zn an anisotropic distribution and a cascade mechan- ism in their transport was observed. In the case of Cu Fe and Mn (94/3146) these were atomized from a graphite platform in a pyrolytic graphite coated electrographite tube.The shadow spectral filming results indicated a significant non-uniformity of the absorbing layer which was attributed to adsorption of the atoms on the furnace walls the diffusion transfer processes and non-uniformity of the distribution of the sample residue. The results for the atomization of Ga and In (95/255) showed a non-uniform distribution with respect to both the cross- sectional and longitudinal sections of the atomizer. The com- pounds were localized in the central part of the atomizer over the platform and disappeared at the hotter walls. The absorp- tion bands observed with maxima at 255 and 295nm upon the atomization of gallium and indium nitrates were assigned to gaseous Ga,O and In,O respectively.The distribution of free In and Ga atoms differs from that of their compounds. These enter the gas phase not from the platform but from the upper wall of the atomizer (inverse atomization). This is due to sample evaporation as M 2 0 molecules which are dissociated or reduced to metal atoms on the hotter graphite walls of the atomizer. In the case of atomization of Ga and In the effects of chemical reactions between the vapour and the walls of the atomizer on the formation of absorbing layers are stronger than that of diffusion and convective mass-transfer processes which are common to all of the elements. Atom propagation from the centre to the ends of the atomizer proceeds through the cascade mechanism because of the atomizers relatively low rate of heating and strong longitudinal non-homogeneous temperature distribution.The differences in the atomic and molecular distributions were examined for the atomization of Si Ge and TI (95/256) by shadow spectralfilming with the sample deposited either on the atomizer wall or platform. With deposition on the atomizer wall the atomic distribution decreases near the wall with platform atomization the effect of inverse atomization is seen atoms enter the gas phase not from the platform but from the opposite atomizer wall. The nitrates of Bi Ge and T1 evaporate as volatile sub-oxides which are reduced on the graphite atomizer wall and re-enter the atomizer gas phase as free atoms. Atoms diffusing within the atomizer volume are deposited on the colder platform and re-enter the atomizer gas phase as the platform warms up.Thus the atomic and molecu- lar vapour densities of these elements increase and decrease in the furnace cross-section when absorbing layers of these elements are formed. Atom propagation from the centre of the atomizer to its ends proceeds progressively by the cascade mechanism as a result of the non-isothermal temperature distribution within the atomizer. The structure of the absorp- tion layer becomes uniform at the end of sample evaporation. In the final report covered in this Update period Gilmutdinov et al. (95/260) turned their attention once more to study of the atomization of A1 by shadow spectralJilming. Aluminium was injected as the nitrate sulfate or metal. The structures of the atomic and molecular layers were significantly inhomogeneous with respect to both the atomizer cross- and longitudinal-sections.Free atoms were found to be localized near the atomizer walls and under the platform. By contrast the A1 compounds were located along the axis of the central part of the atomizer and found to disappear near the walls. The absorption band with a peak at 255 nm was assigned to gaseous A120 molecules. The structural heterogeneity of the atomic layer and of the molecular layer of A1 sub-oxides are caused by oxidation reactions between A1 compounds and oxygen and by the reduction of gaseous oxides on the graphite walls of the atomizer. During atomization of pg amounts of A1 in the gas phase inside the graphite atomizer a cloud of finely divided A1203 forms as a result of vapour oxidation with air entering the atomizer through the injection hole.It would appear that reactions involving oxygen-containing molecules have a major effect on the atomization of Al. The use of cross-beam sampling with ETV-MS to investigate the decomposition of metallic nitrates to oxides was discussed last year (see J. Anal. At. Spectrom. 1994 9 213R) and has now been published (94/2949). It is clear that the nature of the surface from which atomiz- ation takes place has a major effect on the atomization process. Up to now most workers have concentrated on the gas phase though these investigations have resulted in a general recog- nition of surfaces as principal players in the atomization processes but few have actually examined the surfaces involved. This is not because the surface was being ignored but more a case of the severe difficulties in trying to investigate the surface under the conditions present in an electrothermal atomizer.Styris et al. (94/C3420) discussed some of the experimental approaches that can be used and presented recent work on the elucidation of electrothermal atomizer mechanisms by applying extended X-ray absorption fine structure spec- troscopy in particular to the selenium-palladium mixture. Results were compared with MS data. Chinese workers (94/1825) described a model which accounted for the dissi- pation process of the atomic vapour and for the interaction between the elements determined and the wall of the graphite atomizer tube. Katskov et al. (94/2932) reviewed the results obtained with a two-step atomizer and a scanning spectrometer to investigate the processes of sample evaporation and atomization in ETAAS.The apparatus consisted of a tube furnace held at a constant temperature which was used for vapour decomposition. Sample vapours enter the furnace from a vaporizer situated next to the furnace via a very low stream of carrier gas or by diffusion. To enable a stationary absorbing layer to be formed within the furnace the vaporizer temperature was controlled by means of an automatic system using the analytical signal feedback for determining the vapour flux. Absorption was measured by one of two spectral channels of a spectrometer at a set wavelength. The other channel scanned the spectrum over a range of wavelengths. The evaporation characteristics of Ag Au Bi Cd Co Cu Ga Ge Pb Sn Te T1 and Zn in the presence of halide and oxygen containing atmospheres and Ag Bi Cd Ga In Mn Pb Sb Sn Te T1 and Zn in both inert and oxygen containing atmospheres were presented.An Journal of Analytical Atomic Spectrometry August 1995 Vol. 10 21 1 Rexplanation of the results obtained based on Langmuir's theory of evaporation was developed for the experimental conditions of ETAAS. The paper is recommended reading as it presents an alternative experimental approach and contains a great deal of information. The conference report from Tittarelli et al. and discussed last year (see J. Anal. At. Spectrom. 1994 9 213R) concerning the vaporization of Ge and Si as molecular species within electrothermal atomizers has now been published (94/2952).Chinese workers (95/202) investigated the atomization of Ge from a graphite probe surface as germanium fluoride and sodium germanate using XRD X-ray photoelectron and Auger electron spectrometry. The following mechanism for the forma- tion of Ge atoms was proposed. Both GeF and Na3Ge03 decompose first to GeO and then to GeO followed by thermal dissociation of GeO to form free atoms of Ge at a temperature greater than or equal to 2400 K. The decomposition route from GeO to GeO tends to agree with that found by Doidge and McAllister (see J. Anal. At. Spectrom. 1994 9 213R) who used ETV-MS and thermochemical equilibrium calculations which was discussed last year. Wiltshire et al. (95/321) investigated the atomization of B from a total pyrolytic graphite tube coated with either tungsten carbide or lanthanum carbide.These coatings increased the optimum pyrolysis temperature of B from 850 to >22OO"C. Addition of a calcium-magnesium chemical modifier increased the pyrolysis temperature to 1200 "C whereas a titanium- ascorbic acid modifier had no significant effect. The low- temperature loss of B without a chemical modifier was investigated using dynamic SIMS which confirmed that vapor- ization of B species occurs above 900°C. Atomic emission signals were obtained at < 800 "C by hollow-cathode FANES. Molecular dissociation and excitation of B atoms apparently occurred through a one-step collisional process. The results suggested that the atomization of B in ETAAS was likely to occur through molecular dissociation rather than solely by sublimation of B.The chemical modifiers probably prevented low-temperature dissociative desorption of B,O occurring at active carbon sites and so increased the optimum pyrolysis temperature to >85O"C. The poor detection limit for B in ETAAS was due to the inefficient thermal dissociation of the B-containing species (probably oxides and carbides) produced by dissociative desorption of B203. In addition once formed B atoms apparently undergo a series of condensation-vaporiz- ation steps which caused a persistent plateau in the tail of the AA signal and resulted in severe memory effects. Doner and Akman (94/2933) investigated the effect of cobalt chloride on the atomization of Zn with the use of dual cavity graphite platforms to allow differentiation between gas-phase and condensed-phase interferences.When the analyte and interferent were separated on the dual cavity platform con- densed-phase interferences disappeared whereas gas-phase interferences continued to affect the sensitivity of the analyte. From this study the dominant interference mechanism appeared to be the formation of volatile zinc chloride upon reaction between the analyte and interferent at low tempera- tures which was expelled from the atomizer by the gaseous hydrogen chloride generated in large amounts during the thermal hydrolysis of cobalt chloride hexahydrate. In addition a condensed-phase gas-phase reaction between the gaseous hydrogen chloride and the analyte could also cause the forma- tion of zinc chloride.Although gas-phase reactions and/or expulsion mechanisms would seem plausible their effects did not seem to be very pronounced. Russian workers (94/2277) reviewed (55 refs.) the problems of mathematical simulation of electrothermal atomization in AAS. They proposed a method for an equilibrium thermo- dynamic simulation based on dividing a non-equilibrium dynamic system into successive local quasi-equilibrium zones. The method was applied to the description of the pyrolysis and atomization of more than 40 elements in graphite tube and tungsten spiral atomizers. Zheng (95/277) determined the atomization eficiency of the Hitachi GA-3 atomizer for Ag Cd Cr Ge and In by a method based on the measurement of residence time of analyte atoms peak absorbance and integrated absorbance.The atomization efficiencies for Ag and Cd appeared to be stable at low temperatures whereas those for Cr Ge and In reached a stable value at higher temperatures. For In the efficiency was increased in the presence of palladium and the ammonium salt of EDTA as chemical modifiers. The method proposed by Yan et al. and discussed last year (see J. Anal. At. Spectrom. 1994,9 213R) for the determination of the kinetic order and activation energy from a single absorbance profile has been applied by Ni and Yan (94/C3371) to the atomization of In Pb and Sn and in detail by Yan et a!. (95/252) to the atomization of In and by Quan et al. (95/331) to the atomization of Sn. A calculation based on x-order kinetics and the Arrhenius equation was given which leads to a function of absorbance temperature and time and which allows activation energy and order of the atomization pro- cess(es) to be found from the linear plot(s).For both the atomization of In (95/252) and Sn (95/331) the influence of the atomizer surface and a palladium chemical modifier was considered. The atomization mechanism of In seemed to be identical for the pyrolytic graphite coated and uncoated electro- graphite tubes and the rate-limiting step for the atomization changes from 1st-order kinetics at lower temperatures to nearly 3rd-order at higher temperatures. These workers suggested that the In moves from a dispersed state to agglomerates with increasing temperature. However for the zirconium-treated tubes the atomization of In is controlled by a single mechanism with a kinetic order of approximately 2/3 and an activation energy of 186+ 13 kJ mol-l.It was suggested that this is due to relatively weak indium-zirconium interactions and the release of In from the sphere of molten In metal on the zirconium surface. In the presence of a palladium chemical modifier based on the observed 1st-order kinetics and an activation energy of 421 & 27 kJ mol- ' a simple mechanism was proposed in that the In was released from a solid solution of In and palladium on the pyrolytic graphite surface. For the atomization of Sn (95/331) results from a pyrolytic graphite coated electrographite tube gave two linear regions corresponding to kinetic orders and activation energies of 1.10 and 450 kJ mol-' at lower temperatures and 0.47 and 210 kJ mol-' at higher ones respectively. It was proposed that carbon reduction of SnO,(l) to Sn(g) was the rate-determining step at low temperatures.At higher temperatures the rate- determining step was Sn,(g) to Sn(g). Zirconium-coated graph- ite gave a single line corresponding to an order of 0.99 and an E of 429 kJ mo1-'. Similar values to those obtained in the low-temperature region for the pyrolytic graphite coated sur- face but with a lower appearance temperature for the Sn signal indicated a different mechanism. It was proposed that zir- conium effectively prevented the formation of gaseous SnO,. A palladium-coated graphite surface also gave a single line with an order of 1.07 and an E of 838kJ mol-' and an elevated appearance temperature indicating a strong inter- action between the analyte and palladium. It was suggested that palladium promotes the reduction of SnO or SnO to Sn which also forms a solid solution in palladium.From a practical point of view improved sensitivity with zirconium- and palladium-coated graphite tubes was found. Similar results were also obtained for the atomization of In and Pb (94/C3371) and in addition the mode of sample deposition did not influence the atomization mechanism for Pb. With either hydride vapour or aqueous solution methods of sample intro- 212 R Journal of Analytical Atomic Spectrometry August 1995 Vol. 10duction the same mechanism in either zirconium- or palladium- coated graphite tubes was found. Chinese workers (94/2407 95/56) have studied the role of the n-component in Zeeman-effect AAS paying particular atten- tion to the interrelation between the n-component and polar- izer in transverse magnetic field modulation Zeeman-effect AAS.The calculated and experimental values for the relative sensitivity coefficient for 32 elements and 34 lines in the absence of a polarizer were also presented. 1.2.4. Interferences A review appeared in Russian (94/2594) containing 217 refs. covering the last five years on the use of chemical modifiers for the determination of trace impurities in complex samples by ETAAS. The role of chemical modifiers in the analytical process as well as the chemical modifier influence on all elements in the sample and the use of different chemical modifiers for analyses were broken down into the following groups of sample types geological medical biological technological ecological and pure materials and chemicals.1.2.4.1. Spectral interferences. The conference reports on the effects of chemical and spectral interferences in the determi- nation of Zn in sea-water and spectral interference on the Pb 283.3 nm line in Zeeman-effect AAS and discussed last year (see J. Anal. At. Spectrom. 1994 9 213R) have now been published (94/2957 and 94/2964 respectively). Aller (94/3288) has examined the roll-over effect with Smith- Hieftje background correction for an Au HCL at 242.8 267.6 and 274.8 nm. Hardly surprisingly this worker comes to the same conclusions as those of de Loos Vollebregt and de Galan in their study of roll-over effects published in Spectrochim. Acta Part B 1986 41 597.The roll-over effect was affected mainly by the relative sensitivity and stray light level. It was found that the stray light produced in the HCL was different for each wavelength range and its level during the low current pulse was modified by the presence of vanadium and copper in the atomizer. Chinese workers (94/3048 95/224) examined the nature and source of the background absorption of samarium (94/3048) and europium (95/224) matrices. For samarium the back- ground absorption can be reduced by using nitrate salts. With a europium matrix there is one peak for EuC1 which arises from the evaporation of EuCl,. Two peaks were observed in the background absorption due to Eu(NO,),. The first peak arose from NO and the second from Eu20,.These workers found that the background interferences could be decreased or eliminated by the choice of an appropriate signal measure- ment time optimum pyrolysis temperature and the use of chemical modifiers though no indication was given as to which were used. 1.2.4.2. Chemical modijiers-general. Workers at Strath- Clyde have continued their investigations into the effectiveness of various chemical modifiers by the use of ion chrornato- graphy. Belazi and Littlejohn (94/C3482) examined the effects of ammonium dihydrogen orthophosphate ammonium nitrate nitric acid and ascorbic acid on the removal of chloride and sulfate from sea-water deposits in a graphite atomizer. Without a chemical modifier sodium chloride was vaporized at a temperature of 800-1000 "C and sulfate was lost at 1000 "C.When equimolar amounts of the chemical modifiers were added to 10 p1 of two-fold diluted sea-water chloride was vaporized at temperatures < 800 "C except with ascorbic acid. Up to 5% m/v ascorbic acid had no effect on the vaporization of chloride. With nitric acid chloride was almost totally removed at a pyrolysis temperature of 200°C. Addition of ammonium nitrate removed some chloride at 200"C but the remainder was not vaporized until 1000 "C. Ammonium dihydrogen orthophosphate allowed almost complete removal of chloride at 500-600 "C. The decomposition of sulfate was not affected significantly by the presence of these modifiers with the exception of ammonium dihydrogen orthophosphate which reduced the vaporization temperature from 1000 to 700°C.These results tended to confirm with direct experimen- tal evidence what others had suggested and it is clear that many of the reactions between interfering matrix components and the chemical modifier take place at lower temperatures than previously thought possible. The role of ascorbic acid is one that many find confusing. Some workers have found the use of ascorbic acid to be of benefit while others have found that it has no beneficial effect whatsoever. Rock (94/C3453) considered the use of hydroxylamine hydrochloride as a reducing agent instead of ascorbic acid. The advantage of using hydroxylamine hydrochloride is that it can be mixed with the palladium solution unlike ascorbic acid which causes precipi- tation of palladium.A more consistent performance with reduced palladium was claimed though during the presentation it became clear that wall atomization and peak absorbance measurements were being used for assessment purposes. Workers from Hitachi (94/3311 94/3334) have continued their studies into the use of alloy phase diagrams in an attempt to elucidate the role ofchemical modijiers and the atomization process. Using the two-line method of establishing the tempera- ture within a graphite tube atomizer these workers (94/3311) investigated the effective vapour temperature for Se in the presence of palladium or rhodium. In the presence of these metals the effective vapour temperature was increased by 400°C and this effect was attributed to the formation of intermetallic compounds with palladium or rhodium resulting in the activity coefficients of Se becoming < 1 thus raising the atomization temperature.The effective vapour temperature of Se atomized from the platform was 200 "C higher than atomiz- ation from the tube wall. A urine matrix was used to investigate excitation interferences and the presence of an oxidizing agent increased the effective atomic vapour temperature of Se to >8OOO"C. It is difficult however to understand how the vapour-phase temperature can be so much higher than the atomizer. This interference could be reduced by increasing the amount of palladium or rhodium or by the addition of a reducing agent. Similarly these workers (94/3334) assessed the role of Ag Au Cd Cu Mg Pd Pt and Sb as chemical modifiers for Pb and Sn.The effective temperature profiles of the analytes were obtained in the presence of various modifiers and the phase diagrams for the alloys formed constructed. When the activity coefficient of the analyte was < 1 intermet- allic compounds were formed and the atomization was shifted to higher temperatures. Atomization was not changed if the chemical modifiers formed alloys with Pb that had a melting- point lower than the pyrolysis temperature and the initial atomization temperature for Pb. For chemical modifier salts such as those containing magnesium which were neither reduced to metal nor formed alloys with the analyte during ashing and atomization the role involved suppression of vaporization of metal halides with higher vapour pressure. 1.2.4.3. Chemical modijiers-palladium.Once again with respect to chemical modijiers palladium is the reagent that has attracted most attention in the period covered by this Update. It seems that various groups are looking in more detail at how palladium functions and also how applicable palladium either alone or in conjunction with other reagents is for stabilizing elements such as As and Se when present in different inorganic and organo-forms. This work has significant relevance to speciation studies. Matsumoto (94/2547) discussed the use of palladium as a chemical modifier for Group 3 and 6 elements in a review containing 42 refs. The use of MS to elucidate the Journal of Analytical Atomic Spectrometry August 1995 Vol. 10 213Rvaporization and atomization processes of As and Se in the presence of palladium was also discussed.The work of Qiao et al. (94/3346) concerning the mechanism of action of palladium in reducing chloride interferences was discussed in detail last year (see J. Anal. At. Spectrom. 1994 9 213R). In contrast to this study and many others over the years which have shown how effective palladium is as a chemical modifier in overcoming chloride interferences Pszonicki and Essed (94/3155,95/216) claimed that palladium is only a suitable chemical modifier in aqueous nitric acid (95/216). They found that the interference effect of chloride on Pb depended upon the properties of the chlorides present and on the solution in which the sample was dissolved. With a large excess of chloride especially in dilute hydrochloric acid the palladium was converted into an unstable chloride which was responsible for some losses of Pb during the pyrolysis step.However in subsequent work the same workers (94/3155) found that the addition of magnesium nitrate eliminated the losses of Pb during the pyrolysis step and overcame the chloride interference problems. Not surprisingly this is the same result as was found by most other workers with this chemical modifier combination and which has been in the literature for the last six years. Liang and Ni (94/3364) considered the release mechanisms of Ag Co Mn and T1 with a palladium chemical modifier. For Co and Mn no change in the activation energies was observed by the presence of pal- ladium. The activation energy was decreased for Ag and increased for T1 when palladium was present.They found that the palladium stabilizing effect was more pronounced for the relatively volatile elements which can be stabilized to a higher tolerable pyrolysis temperature owing to the formation of a relatively strong palladium-analyte interaction. The relatively weak interaction between palladium and Co or Mn accounts for the similar activation energies obtained both in the absence and presence of palladium. Johannessen et al. (94/2186) compared a series of chemical modijiers i.e. nickel nitrate magnesium nitrate copper nitrate copper nitrate mixed with magnesium nitrate palladium nitrate and palladium nitrate mixed with magnesium nitrate for their effectiveness in stabilizing selenite selenate seleno- methionine and trimethylselenonium. None of the examined chemical modifiers stabilized all the different Se species to the same extent.In aqueous solutions addition of 2.7 pg of pal- ladium or 7.5 pg of palladium plus 5 pg of magnesium nitrate stabilized selenite selenate and selenomethionine equally though the signal response for trimethylselenonium was only 55% using palladium and 85% with palladium and magnesium nitrate. Various chemical modifier combinations and mixtures were applied to the determination of Se in serum and urine. The recommended chemical modifier combination was that of 7.5 pg of palladium plus 5 pg of magnesium nitrate which gave results that were in good agreement with the certified values for Se in NIST SRM 2670 Freeze-Dried Urine and a serum RM though these are known to be spiked with selenate.This chemical modifier was considered to be the best of those investigated despite the impaired sensitivity towards trimeth y lselenonium. Sneddon and Farah (95/23) found that a combined chemical modif er of palladium and magnesium nitrates allowed compro- mise pyrolysis temperatures to be used for the simultaneous determination of As Cd Pb and Se. However no details were given of the instrumentation used in this study. For the determination of Co in serum (95/153) palladium nitrate was used as the chemical modifier. Chinese workers (94/1785) found that a combined nickel nitrate and palladium nitrate mixture was the optimum chemical modifier for the determi- nation of Se in drinking water. Zhixia et al. (94/2202) used palladium as a chemical modifier for the determination of Si in low-alloy steels.Sensitivity was improved by the use of palladium. For the determination of Sn in serum (95/33) palladium and ascorbic acid were compared as chemical modifiers. The optimum chemical modifier was palladium and XRD measurements of pyrolysed residues indicated the pres- ence of Pd3Sn2. For the determination of trimethylselenonium in urine samples Tsunoda et al. (95/438) applied a palladium chemical modifier after solvent extraction of the trimethyl- selonium as an ion pair with tetrakis( 4-fluorophenyl) borate. Fenbutatin oxide in fruit (95/456) was determined indirectly by extraction with ethanol and subsequent determination of the Sn with the use of palladium chloride as a chemical modifier.Sachsenberg et al. (94/1611) applied a mixed chemical modifier of ammonium oxalate and tetraamminepalladium(n) chloride to overcome the interferences in the determination of trace elements in sea water. For Mn and Pb this chemical modifier combination allowed direct determinations but for Cd only ammonium oxalate could be used. These workers examined the geometry of the micro-distribution of palladium on the palladium-conditioned graphite platform to investigate the possible stabilizing effects of palladium on the analytes that was observed with the different chemical modifiers. Fractal characteristics of the palladium elemental distributions were ascertained by SEM and energy dispersive X-ray spectrometer image box counting analysis. These investigations indicated differences in the structure of palladium deposits on the platform.It was felt that these variations in the palladium geometry affected the analytes to a certain extent. The increasing importance of determining organotin com- pounds in potable water supplies means that several workers have turned their attention to the optimum chemical modifiers for the determination of Sn in organic extracts. Previous recommendations for this determination relied upon coating of the graphite tube with refractory elements to prevent interaction with Sn even though this is a time-consuming and non-reproducible technique. Li et al. (95/32) considered the effect of oxygen atoms in the solvent and ligand by examining the sensitivity for Sn for tetrabutyltin tributyltin chloride dibutyltin diacetate and dibutyltin dilaurate in the presence of toluene ethyl acetate or tributyl phosphate.As a chemical modifier PdC12(CH3CH)2 was employed. The results showed that the presence of tributyl phosphate and oxygen in the organotin compound enhanced the sensitivity for Sn. Dadfarnia et al. (94/2210) applied a 4-methylpentan-2-one solution of palladium as chemical modifier for the determi- nation of Sn in toluene-extractable organotin compounds from sewage effluents and a complex simulated effluent. 1.2.4.4. Other chemical modifiers. Although palladium with or without another reagent appears to be the chemical modifier of first choice there are situations and elements where pal- ladium is not the optimum choice for a chemical modifier. In these situations a mixed metal chemical modifier could be more appropriate. Dahl et al.(94/2209) investigated the effectiveness of metal chemical modifiers for the determination of Sb in aqueous samples blood and urine. A radiotracer technique was used to investigate the effectiveness of the thermal stabiliz- ation. A purge gas of hydrogen was introduced only during the pyrolysis step. After pyrolysis the activity of lz5Sb was counted and compared with that of control samples which had been allowed to dry in air at room temperature. The best results were obtained with a chemical modifier solution con- taining 0.025% m/v each of Pd Pt Rh and Ru in 1% m/v ascorbic acid. Applying this chemical modifier mixture pyrol- ysis temperatures up to 1200 "C gave quantitative recoveries of Sb with an overall RSD of 3.2% (n=20).Matsusaki et al. (95/210) found that a mixture of aluminium and cobalt nitrates proved useful as a chemical modifier for the determination of Ge. An increase in sensitivity of 40-fold 214R Journal of Analytical Atomic Spectrometry August 1995 Vol. 10over that found in the absence of the mixture of metal nitrates was claimed. For the removal of chloride interferences in the determination of Ge a mixture of ammonium acetate with these metal nitrates was suitable as a chemical modifier because chloride formation of aluminium was masked by acetate and the volatile ammonium chloride formed could be easily removed from the atomizer. The tolerable concentration of co- existing chlorides was approximately ten-fold or more greater than in the absence of ammonium acetate.The sulfate inter- ference could also be removed by the addition of a mixture of ascorbic acid and ammonium EDTA with cobalt and alu- minium nitrates. The tolerable concentration of co-existing sulfates was also 200-fold greater than in the absence of these reagents. Chinese workers (95/83) compared the chemical modification capabilities of nickel and calcium nitrates for the determination of Ge in plasma and urine. Unfortunately it is not clear from the abstract which chemical modifier they found to be the most effective however these workers have developed a method and have finished the determination of Ge in the preclinical animals and are about to start experiments with 132Ge. Several different metals have been examined as chemical modijiersfor Au by Thomaidis et al.(94/C3502). Only Eu Pd and Rh increased the thermal stabilization of Au to 900-1000°C. These workers also examined the stability of Au solutions in the autosampler cups and found that the Au solutions were only stable in the presence of 0.25 g I-' of ammonium thiocyanate or 0.5 g 1-l of ascorbic acid. Biological and geological samples were analysed. Botelho et al. (95/322) compared diferent chemical modijiers atomization surfaces and potential interferences for the determi- nation ofB. Calcium as chloride and nitrate nickel as chloride and nitrate magnesium yttrium and lanthanum all as chlorides were compared. From a consideration of the pyrolysis tempera- ture and sensitivity calcium chloride was considered to be the best.Ammonium hydrogen phosphate and manganese and iron chlorides all showed severe signal depression though sodium and magnesium chlorides had little effect. Severe memory effects were observed which required the measurement of a blank solution after each sample. No advantages were found for the use of alternative atomization surfaces such as tungsten carbide or tantalum foil and atomization from the wall of a pyrolytic graphite coated electrographite tube was preferred. The recommendation of a calcium chemical modifier for B agrees with the results of Wiltshire et al. (95/321) who found a mixed calcium-magnesium chemical modifier to be the best and whose work was discussed earlier (see section 1.2.3). For the determination of Bi in kidney liver brain and bone platinum was employed as the chemical modifier (94/1047).The samples were digested and calibration was established in matrix matched solutions from digested tissues obtained from unex- posed animals. Magnesium nitrate was used as the chemical modifier for the determination of Mo in serum (95/450) and a detection limit of 0.3 pg 1-' was found. Vanadium and vanadium plus molybdenum chemical modijiers were applied by Manzoori and Saleemi (94/2934) in the determination of Cr in serum and lake water samples with high salinity. These workers examined several other chemical modifiers suggested in the literature for Cr but found that ammonium vanadate (20 pg of vanadium) was the most effec- tive for this study with improved recoveries and longer tube lifetimes when the chemical modifier was employed.The detec- tion limit (3s) was found to be 0.2 pg 1-'. However the results were all based on peak absorbance measurements and with wall atomization. Koshino and Narukawa (94/3000) investigated the elimin- ation of the sodium nitrate-borate matrix interference on the determination of Mn in geological samples. Chemical modifiers of magnesium nitrate nickel nitrate and palladium nitrate were examined and for this particular matrix interference (0.45 mol 1-' sodium nitrate and 0.05 mol 1-l boric acid) which produced an 80% reduction in the integrated absorbance signal for Mn. A nickel nitrate chemical modifier was found to give the best sensitivity and smallest background signal. Zu and Li (94/2412 94/2888 95/235) examined the inter- ference eflects of calcium chloride on the determination of Mn.A chemical modifier of nitric acid shifted the Mn signal to a higher temperature increased the sensitivity and improved the signal shape. The interferences by magnesium strontium and barium chlorides were found to be similar to calcium chloride. He (94/3323) applied lanthanum as a chemical modifier to the determination of Sc in titanium and iron ores. Lin et al. (94/2410) determined Sn in blood with potassium nitrate as the chemical modifier. Japanese workers (95/437) used a mixed chemical modifier of nickel phosphoric and ascorbic acids for the determination of Sn in whole blood and found a detection limit of 2.5 pg I-'. Indian workers (94/3018 95/333) investi- gated the chemical modification of Sn with silver gold zirconium molybdenum and tungsten from powder X-ray diffraction (94/3018) and ETAAS measurements (95/333).The carbide-forming chemical modifiers (molybdenum tungsten and zirconium) shifted the Sn signals to lower atomization times. Palladium was found to be the best of the non carbide- forming chemical modifiers (gold palladium and silver) giving increased atomization times the best sensitivity and to be the most efficient in preventing losses of Sn. This was attributed to the formation of SnPd as an intermediate. Chinese workers recommended ascorbic acid as a chemical modifier for the determination of Ru (94/2889) in water. A detection limit of 0.27 pg 1-' was found. Yet another group of workers (94/1667) have decided to investigate the role of the chemical rnodiJier for the determination of Pb in whole blood.Not surprisingly these workers found exactly the same results as everybody else and recommended diammonium hydrogen phosphate. They also compared the use of this reagent in conjunction with palladium chloride and found worse precisions with this combination of 3.1-9.1 YO in the concentration range 31-624 mg 1-' compared with 2.2-6.3% with diammonium hydrogen phosphate. Indian workers (95/43) suggested the use of bismuth nitrate as a chemical modifier for the determination of Pb in fish tissue. These workers claimed that the detection limit of 0.9 pg 1-l obtained with bismuth nitrate was better than that found when using a palladium nitrate chemical modifier. However no further details were available to allow further assessment of this work.Russian workers (95/163) proposed a new class of chemical modifiers aminopolyphosphonic acids for the deter- mination of Pb in copper chloride copper-based alloys fresh water sediments and blood plasma. These chemical modifiers were claimed to allow a 10- to 100-fold improvement in the detection limit though quite how this was achieved is not at all clear from the abstract. Huang (94/2211) examined a variety of chemical modijiers for the determination of Si in bone and other tissues. A mixed chemical modifier of lanthanum nitrate and calcium chloride was used. Ammonium dihydrogen phosphate was added to sample digests to eliminate the interferences arising from the soft tissue samples (brain kidney liver spleen and heart).This mixture was not able to overcome the interferences arising from the bone digests and for these samples the mixed lantha- num and calcium chemical modifier plus the addition of disodium hydrogen tartrate was successfully applied although wall atomization was used. Concentrations of Si in the sample digests were determined against a calibration curve established with aqueous Si solutions diluted with the same chemical modifier mixture and integrated absorbance measurements. A characteristic mass of 37 pg and detection limits (3s n = 10) of Journal of Analytical Atomic Spectrometry August 1995 Vol. 10 215R0.9 and 0.14 pg g-' wet mass for bone and soft tissue respectively were found with recoveries of the order of 100%. Volynsky et al.(95/190) examined the use of organic chemical modijers such as ascorbic acid and glucose under STPF conditions with thermal and elemental techniques. Ascorbic acid did not affect the integrated absorbance values for Pb and Sn but increased that for Ga by 1.37-fold. The appearance times for all three elements increased in the presence of ascorbic acid. It was found that the ascorbic acid melt wets Ga and Pb oxides better than the glucose melt. Electron microscope studies showed that after drying and pyrolysis at 500°C of a solution containing ascorbic acid Ga and Pb only 14 and 7% of the corresponding metal was retained on the surface of the carbon residue. For glucose the values were 30 and 25% respectively. These workers proposed that the increase in appearance time in the presence of ascorbic acid was a result of diffusion from the carbon residue at the atomization stage.Indian workers (95/327) employed ascorbic acid as the chemi- cal modifier for the determination of T1 in nickel-based alloys after a microwave digestion procedure with sulfuric nitric and hydrofluoric acids with a continuum source background cor- rected ETAAS system. They found that the use of STPF conditions improved the absorbance signal and the precision of the measurements (2-10% without a platform and 0.2-3% with a platform) and that the sulfuric and ascorbic acids overcame matrix interferences. The detection limit (3s) was 1.0 pg 1-' and good agreement with certified values was obtained for a range of nickel-based RMs. Pergantis et al. (95/345) applied simplex optimization to establish the optimum experimental conditions for the determi- nation of As in a marine RM.Four experimental variables were considered pyrolysis temperature; atomization tempera- ture; chemical modifier concentration; and atomization ramp times. These workers found that this approach provided a pow- erful means of rapidly improving the experimental conditions used. Excellent recoveries of As were obtained when the optimum conditions established by the simplex process were used to determine As in standard solutions of arsenobetaine arsenocholine and tetramethylarsonium iodide. 1.2.5. Developments in technique Material presented in a thought provoking lecture at the XXVIII CSI Post-Symposium on Graphite Atomizer Techniques in Analytical Spectroscopy in 1993 and discussed last year (see J.Anal. At. Spectrom. 1994 9 213R) has now been published by Gilmutdinov et al. (94/2979) in the form of a full paper. The work discussed the signal obtainedfrom ETA systems in terms of the number of absorbing atoms as well as their spatial distribution in the atomizer volume. The results obtained showed that cross-sectional distribution of both radiant intensity and the analyte can be highly non-uniform within the atomizer. This results in the fact that the absorbance measured by a PMT-based detection system is dependent not only on the number of free atoms but also on their gradients and the gradient of the radiation beam. To obtain an analytical signal that would be independent of these non-uniformities spatially resolved detection is necessary.To achieve this a solid-state detector should be used. The absorbance recorded by such a system depends only on the number of absorbing atoms. Such a system would have the advantage of both spatially resolved detection and the fact that these spatially resolved absorbances could then be integrated. In principle the absorbance should be integrated not only over time but also with respect to the monochromator slit and this should be used as a measure of analyte concentration in the sample. This would provide more analytical information and improved linearity for AA measurements. In the previous Update the publication of this work was awaited with interest and the full paper lives up to the expectations of the conference report.Harnly (94/C3468) has continued the work on simultaneous multi-element E TAAS with an investigation into the well-known problem of atomizing different elements under identical conditions. The effect of furnace parameters on simultaneous multi-element measurements was examined in a commercially available transverse heated graphite atomizer (THGA). The sensitivities and S/N ratios for Cd Cu Cr Pb and V selected to provide a representative range of atomization temperatures were characterized as a function of atomization temperature with the use of a mixed palladium and magnesium nitrates chemical modifier and with normal or end-capped THGA tubes to restrict analyte loss. Using a normal open tube without chemical modifier and an atomization temperature of 2500"C the minimum for an acceptable V signal produced sensitivity losses ranging from 9 to 28% for the other four elements.The loss of Cd Cu and Pb was caused by a strong convection component though with end-capped tubes losses without a chemical modifier were less strongly influenced by convection. With a chemical modifier losses from end-capped tubes were controlled mainly by diffusion. Simultaneous multi-element detection has again been receiving attention particularly with the release of a new commercial simultaneous multi-element ETAAS system from Perkin- Elmer. The search for such a technique was compared with the 'search for the Holy Grail' by Shuttler et al. (94/C3430). The system uses multiple lamps as line sources an Cchelle polychromator with a solid-state detector and a THGA with longitudinal Zeeman-effect background correction.Similar detection limits were obtained with the instrument running in either single- or multi-element mode with up to six elements being determined simultaneously. Smith and Harnly (94/C3467) examined the use of a double atomizer for ETAAS. The double atomizer consisted of an integrated contact furnace (ICF) located directly above a graphite cup. The cup was placed 3 mm below the ICF and hence was not in contact and the ICF had a hole drilled in the bottom to admit the gases from the cup. Both the cup and ICF had separate power supplies and were operated indepen- dently and mounted in a modified vacuum cross that permitted operation at low 0.1333 Pa and high pressures 10.133 x lo5 Pa. It was found that when the ICF was at a constant temperature the vapour evolved from the cup was drawn into the inlet hole of the ICF.The sensitivity for Cd was found to be greater than that for an HGA-500 equipped with a platform when using a cup and ICF temperatures of 200 and 1000"C respectively. Parameters such as inlet hole size spacing between the cup and the ICF and the respective temperatures were examined. Further work on furnace atomization plasma excitation spec- trometry (FAPES) has been reported by Hettipathirana and Blades (94/2178). The interference effects of sodium chloride and sodium nitrate were discussed and further details are given in the Atomic Emission Update. The emission intensity in the rf plasma was decreased for both Ag and Pb signals in the presence of both salts.These workers also suggested that both salts acted as carriers resulting in early atomization of both Ag and Pb. In an attempt to produce multiple line sources for ETAAS Wichems et al. (94/C2020) suggested the use of an ICP to replace the HCLs. Solutions of the elements to be determined were aspirated into the ICP and the resulting emission focused through a graphite atomizer (or flame) followed by a mono- chromator and detected with a linear photodiode array. While this could overcome the problems associated with using a continuum source it does seem a rather elaborate way of producing a multiple line source. A few reports appear every year on the use of an electrother- 21 6 R Journal of Analytical Atomic Spectrometry August 199.5 Vol.10ma1 atomizer for molecular absorption measurements. Chinese workers (95/80) determined S by volatilization as Ins in a pyrolytic graphite tube. A Be HCL was used to measure the Ins absorption band at 234.5 nm and this method improved the S/N ratio compared with the use of a continuum source. Satisfactory results were claimed for the determination of S however more details were not available. With respect to absolute analysis Chinese workers are the most active. The current trend appears to be one of establishing both the stability of the atomization efficiency with respect to temperature and the stability of the characteristic mass over time. Ensuring that these are stable both in different sample types and over time are important issues for the ability to achieve absolute analysis.Zheng et al. (94/1272 94/3049) considered the factors that influenced the characteristic masses (mo) for Ag Cd Cry Gay Ge and In over a temperature range of 1800-2600 K. The experimental results showed that the mo values appeared to be stable to better than 10-15%. The influences of atomization temperature on the atomization efficiency [defined as ratio of mo(calc)/m,(exp)] were studied using wall platform and V-shaped boat atomization. For Ag and Cd the atomization efficiency did not change significantly when either platform or V-shaped boat atomization was used. However for Cr Gay Ge and In the atomization efficiency was found to increase with the increase in atomization temperature although values of 20-50% were observed. The use of a chemical modifier such as palladium or nickel increased the efficiency values.Further work from the same research group (94/2892) investigated the day-to-day stability of mo obtained from one instrument. For the determination of As Cd Co Cry Cu Mn Ni and Pb in sea-water and As in toys the results showed that the characteristic mass values obtained by different workers from different samples and on different days were generally stable. Ma et al. (94/1625 95/82) discussed the determination of Cr (94/1625) and Cd (95/82) in environmental samples by absolute analysis. In both instances the use of a chemical modifier dilute ammonia solution for Cr and palla- dium and ascorbic acid for Cd were necessary to overcome interferences. Subject matter previously discussed under this category in previous Updates such as linearization are now found under section 1.5.The desire to increase the sample throughput of ETAAS systems is an issue that concerns many. Halls (95/307) has been one of the more active workers in this area and has developed an interface device that enables an older Perkin- Elmer AS1 autosampler to be started before the HGA-400 furnace programme has ended. This allowed a reduction in the total cycle time of 10-15 s without any degradation in analytical perfomance. Shrader et al. (94/C1946) discussed the use of new hardware and software which allows the use of fast furnace time-temperature program. Minimum program times and hot injection were employed to produce a two-fold improvement in productivity. 1.3. Chemical Vapour Generation Developments in chemical vapour generation have once again been directed at improving analytical sensitivity and accuracy and to give additional information notably speciation.Im- provements in vapour generation are not only of benefit in AAS but also in AES AFS and MS therefore where appro- priate experiences in vapour generation in the range of spectro- scopies will be recorded in this section of the ‘Updates’. No single outstanding development or strong general trend has been identified since the last ASU review (J. Anal. At. Spectrom. 1994 9 213R) and publications once again reflect repetition consolidation and modest developments in technique and understanding. Most relate to specific applications rather than to general methodology. 1.3.1.Hydride generation Though a significant number of workers continue to discover the benefits of FI-HG the technique has in fact become firmly established as regular practice in many laboratories. Overcoming interferences in HG continues to exercise many workers. It has been argued (Stockwell P. Analysis Europa 1994 Dec. 41) that the simplest way of overcoming inter- ferences is to dilute the sample and that AFS provides the greater sensitivity required to compensate for increased dilution. This view has some support in the fact that there has been a modest increase in the number of reports on the determination of As Sb and Se by HG-AFS. 1.3.1.1. General studies of fundamentals techniques and instrumentation. Interferences are an ever present potential problem in hydride generation methods of analysis and much effort continues to be spent on their identification and elimin- ation.An investigation of the interfering eflects of cadmium and zinc on the determination of As Bi Hg Sb Se Sn and Te found that the effect of cadmium was greater than that of zinc (94/2587). A procedure was developed that overcame these interferences except that Te could not be determined when cadmium was present. The mutual interferences of As Sb and Se along with the effect of bismuth tellurium and tin on the determination of the these elements have been studied by means of a twin channel HG flow system (94/1639). It was found that the interactions of As Sb Se and Sn on each other were similar and occurred in the gas phase. Bismuth and tellurium also interfered in the liquid phase.Antimony and As could also interfere spectrocopically in the wavelength range 190-235 nm owing to absorption by their molecular bands. Alkaline mode’ operation has been claimed to reduce inter- ferences by Group 10 and 1 1 elements and to have the potential for speciation analysis of hydride forming elements deriving from differences in the stability of species in different pH media (94/C1913). In this mode of operation acid is mixed with an alkaline solution of the sample containing NaBH to generate the hydride. The technique was evaluated using FI-HG-ICP and gave sensitivities for As Bi Ge Pb Sb Se Sn and Te comparable with those of the traditional method. Vapour generation of a oolatile species of Cd by reacting an aqueous solution of the buffered sample with NaBEt is a very sensitive technique when used with AFS detection. This tech- nique is also applicable to Pb.Thompson (94/C3432) has reported detection limits of 5 ng 1-1 Cd and 150ng 1-1 Pb. Generation of volatile species of Cd and Zn using NaBH in an FI system has also been reported (94/C3384). The addition of an organic reagent greatly enhanced sensitivity and with AFS detection achieved detection limits of 8 ng I-’ Cd and 8 pg 1-1 Zn. Surfactant-based ordered media are of increasing interest in analytical atomic spectrometry. A volatile Cd species has been readily formed in cationic vesicles by reaction with with NaBH (94/3036). This Cd species increased the detect- ability of Cd by ICP-AES by a factor of 5. A 2-fold improve- ment in the ICP-AES sensitivity of As and Pb was also noted.The publication also reviewed the general applications and benefits of surfactant-based ordered media in analytical atomic spectrometry. In HG or CV techniques loss ofsensitivity and blockage of transfer lines due to excessive moisture carried to the atom cell or lodged in the lines is not uncommon. Sundin and Tyson (94/C1939) have described a system whereby water was removed by passing a dry sheath gas around the outside of a hygroscopic Nafion membrane transfer line. The Nafion dryer removed water at the rate of 1.7 to 4.9 mg min-’ depending upon operating conditions. There was a slight loss in sensitivity Journal of Analytical Atomic Spectrometry August 1995 Vol. 10 21 7 Rin the determination of As and Hg due to the larger internal volume of the drier.Two types of gas-liquid separators have been examined with the aid of radiotracers (95/112). The techniques studied were the 'classical' separation by gravity and that based on diffusion through a permeable tube. The As species studied were As"' AsV' monomethylarsenic acid and dimethylarsenic acid. Yield and response were investigated as a function of experimental variables. It was concluded that only the classical method with a cold trap was satisfactory. The latter approach was the basis of an automated procedure where electromagnetic relays and timers were used to control the hydride feed to an electrically heated quartz tube atomizer The construction and application of a laboratory-made double input hydride generator has been reported (94/3100). Recoveries of 98.4-103.0% for As and 97.2-108.0% for Sn in steel were achieved. A review of the development state-of-the- art and future trends in FI-HG-AAS from the point of view of automation has been published by Tao and Fang (94/2066).( 94/c 3 3 7 5 ) . 1.3.1.2. Determination of individual elements. In this section of the 'Update' some references will be made to papers where several elements have been studied. In such cases the reference will be presented once only in the paragraph relating to the element that is first alphabetically. Antimony and As are frequently determined in the same matrix. L-Cysteine first introduced by Brindle (Anal. Chem. 1992 64 667) has been comprehensively studied as a reducing and releasing agent for Sb and As by Welz and Sucmanova using FI-HG-AAS (94/3002 94/3003).Under optimized conditions the detection limits were 0.05 pg 1-' Sb and 0.01 pg 1-' As with linear calibration graphs up to 10 pg 1-' Sb and 5 pg 1-' As. The method was developed for the determination of Sb and As in copper and steel. The tolerance limits (<lo% interference) in the determination of Sb were 250 mg 1-' nickel and 500 mg 1-' copper in the determination of As the limits were 200mg 1-' nickel and >lOOOmg I-' copper. In addition to As and Sb Jiao and Lin (94/2710) have determined Bi in geological samples. Using a dual channel instrument with AFS detection As and Sb were determined simultaneously. Bismuth was determined separately but by the same procedure.Detection limits were As 0.79 pg 1-'; Bi 0.46 pg 1-l and Sb 0.29 pg 1-'. Selenium was another element determined in combination with Sb (95/199). In this work tetrahydroborate ( H I ) was bound to an anion-exchange resin. Antimony was reduced to Sb"' with 0.1 mol I-' KI and Se to Se" with 6 moll-' HC1 and heating. Transition elements were not separated from the sample. In the analysis of SRMs results were within certified values. A significant step in the determination of As in workplace air based on HG-AAS has been the validation of a proposed international standard method (94/2923). Particulate As com- pounds were collected on a filter followed by acid digestion. Analysis was by both continuous flow and flow injection HG. For the determination of As in orange juice (94/3170) and in milk (95/419) Cervera and co-workers used ashing aids in a carefully controlled dry ashing procedure.The detection limit of the method was 0.1 ng g-' As in the beverage. The accuracy of the method was checked using tomato leaf and powdered milk standards and found to be within 5% of the certified values. Arsenic was determined in landJill waste waters con- taining high levels of copper (10 g 1-') iron (0.2 g 1-') lead (0.2 g 1-') and nickel (0.2 g 1-') by means of pre-reduction of Cu" to Cu' with hydroxylammonium chloride and a KI-ascorbic acid reagent to reduce AsV to As"' and to precipi- tate CuI and PbI (94/1642). A detection limit of 1 pg 1-' As was achieved. An FI stop-flow-HG technique was used to determine As in tap water and soil extracts (94/2891).The sample solution in 4mol 1-' HCl was merged in a knotted reactor and stored in a sampling loop for 40 s to complete reduction to AsV prior to injection. The detection limit was 0.1 pg 1-' As. Cysteine added to urine has been reported to overcome the problem of the pH dependence of the HG sensitivities of As species (95/191). In the presence of cysteine the optimum condition for the determination of all the As species was in the same pH range. A single As species could be used for calibration and for the determination of total As. Increasingly interest is being directed at the species themselves as well as total As. Such information may be obtained either by selective chemical pre-treatment or by chromatography. A comprehensive chemical pre-treatment procedure for the deter- mination of As"' AsV monomethylarsenic and dimethylarsenic in urine has been described by Hanna et al.(94/1729). Recoveries of added As compounds ranged from 95-112% and detection limits were between 0.1 and 0.2 pg 1-' As in the solution injected into the FI system. Ion exchange chromatogra- phy was used to separate As species in urine mineralized with HN03-H2S04-H202 prior to determination by FI-HG-AAS (95/414). Recoveries of added As species were in the range 8543%. The method could be applied to urine samples relating to arsenic exposure. Arsenic species in lake and river waters have been separated by ion chromatography without pre-treatment and in one instance (94/1626) the column output was coupled directly to an HG-AAS system to give detection limits of 0.2-0.8 pg I-' As.In another report fractions were collected and a thermal decomposition procedure prior to NaBH reduction (94/2198) was used; the hydrides were then trapped on a palladium-coated graphite tube prior to ETAAS determination. Detection limits were of the order of 2 pg 1-' As in the injected sample. Two directly coupled LC-HG systems have been described which incorporate on-line oxidation. Photo-oxidation by low power UV irradiation in conjunction with persulfate has been shown to be effective for a wide range of arsenic species (94/1609). Detection limits were typically 50 pg As corresponding to 0.3 pg 1-' As. Anion-exchange HPLC was coupled via on-line thermo-oxidation to an HG-AA spectrometer for the determination of As species in environ- mental samples (94/1640).The thermo-reactor consisted of a loop of PTFE tubing dipped in a powdered graphite oven heated to 140°C. Samples and persulfate were run together into the reactor. Recoveries were ~ 1 0 0 % and the detection limit was < 1 ng As. An automated FI system for the interference free determination of As and Se in serum and urine has been described (94/3167). The sample with digestion reagent was passed through a coil in a microwave apparatus prior to injection into the HCl carrier stream for reaction with NaBH and subsequent detection by AAS. A continuous frow HG system has been applied to the determination of As and Se in highly mineralized water (94/2925). Atomization was in either a heated silica tube or graphite furnace.Significant matrix effects were found only with high concentrations of magnesium sodium and sulfate. The lowest detection limits (0.02 pg 1-l As) were obtained with in-atomizer preconcentration in the graphite furnace. Methods for the determination of Bi Ge In and Pb have been studied by several groups of workers. A complex sample treatment procedure was used for the determination of Bi in geochemical samples (94/3319). The hydride was generated and trapped in an 12-KI solution prior to reaction with Rhodamine B and spectrophotometric measurement. The influence of O2 and H2 in the electrically heated quartz tube on the absorption signal of Ge has been studied (95/165). The atomization of GeH was attributed to collision with hydrogen free radicals while a reduction in signal resulted from the formation of GeO.The mechanisms of hydride generation and atomization of In in the quartz tube have been studied by Liao and Li (94/3320). Under optimized conditions a peak height sensitivity of 0.007 pg was obtained. For the determination of 21 8 R Journal of Analytical Atomic Spectrometry August 1995 Vol. 10Pb by HG a series of 22 chelating agents were tested for the generation of plumbane (94/3030). The preferred reagent was 1-( 2-pyridylazo)-2-napthol-6-sulfonic acid (PAN-S) which gave injection of a 3% solution of NaBH in DMF. The detection limit was 8.1 pg 1-' Sn. a sensitivity of 1.3 pg 1-' using a 5 ml sample. Lead and Sn in food have been determined by dry ashing the sample followed by dissolution of the ash and adsorption of the analyte elements onto sulfhydryl cotton from which they were desorbed for hydride generation with KBH and atomization in an air-C2H2 flame (94/C3387).Detection limits were 48 pg 1-l Pb and 50 pg 1-' Sn. Methods including HG for the determination of Se have been the subject of a review in Chinese (95/12). A method using FI-HG-AAS for the determination of Se in blood had a throughput of 120 samples h-' and an LOD of 0.06 pg I-' (95/396). Several reports have been devoted to sample prep- aration prior to HG. Four microwave methods for the digestion of fish tissue have been compared (95/344). The preferred method employed HN03-H2S0,-H,02 digestion with potass- ium hexacyanoferrate(II1) as a masking agent to eliminate interferences. The detection limit was 0.03 pg g-' Se dry mass using a 100mg sample; there was good agreement (within 3.5%) with a mussel tissue reference material.For the determi- nation of Se in pure copper following dissolution of the sample the copper was removed by precipitation the hydride generated with NaBH from the filtrate and the Se determined by AFS to give a detection limit of 0.04 pg 1-' Se (94/2712). Inorganic Se (Se" and total Se) in drinking water has been determined by preconcentration and separation on Dowex 1x8 prior to HG-AAS (95/253). Total Se was determined using permanga- nate digestion and anion-exchange preconcentration. The efect of matrices reducing conditions and interferences on the measurement of total Se and SeIV in environmental materials has been investigated by Blackwell et al.(94/C3462). Conditions affecting the reduction of Sew to Se" the effective- ness of the partial digestion of soils sediments and plant materials with Mg( N03)2 + HN03 and of the total digestion using HF + HNO + HClO were also examined. In-situ concen- tration of Se and Te by trapping the hydrides in a graphite atomizer coated with silver has been examined (94/2185). The trapping temperatures were 200-400 "C for Se and 200-800 "C for Te. The atomization of the Se and Te trapped on the silver surface occurred at 1800°C compared with >2000°C from palladium. Interferences in the speciation of organotin have been studied extensively by Martin and co-workers (94/3292 95/302). The method employed HG cryogenic trapping selective volatiliz- ation followed by quartz furnace AAS.A harbour sediment was characterized in order to identify the agents that could interfere in the speciation. Organic pollutants did not interfere; inorganic elements produced severe problems. A masking agent (L- cysteine) overcame most problems. Several sample pre- treatment procedures have been compared for the determination of butyltin compounds in shellfish (95/338). Digestion with 0.083 mol 1-' HC1 in 16.7% methanol with sonication for 1 h was preferred. The method used HG cryogenic trapping selective volatilization and quartz furnace AAS The detection limit was 2-3 ng g-' Sn wet tissue. For the ultra-trace determi- nation of Sn (94/C3377) the element was first extracted from a 2 moll-' HCl solution onto a strongly basic anion exchanger D-201 following elution of Sn with 0.05mol I-' HNO the hydride was generated and trapped in a palladium-coated graphite tube preheated to 300°C then atomized at 2300°C.A throughput of 25 samples h-' with an LOD of 0.05 pg 1-' Sn was achieved. The determination of Sn in rigid PVC necessitated the generation of the hydride from a sample dissolved in an organic solvent (hot N,N'-dimethylformamide DMF) (94/2249). The sample solution was mixed with H2S04 in DMF and formamide in the hydride generator prior to the 1.3.2. Mercury by cold uapour generation The determination of mercury continues to be the most fre- quently reported analysis by vapour generation techniques. In this section of the 'Update' (1.3) there are as many reports of Hg determination as there are of all other elements combined.In some 15% of the Hg determinations AFS was used as the measurement technique. Most of the papers reviewed here relate to development of techniques for specific analytical applications. These papers have been selected to be representa- tive of techniques available and to encourage their wider use. Full accounts of the determination of Hg in particular sample types will be found in other 'Updates'. A review of Hg determination in China has been prepared by Li and Chen (95/23 1). Refinements of instrumentation and techniques continue to be made. Two on-line digestion CVAFS systems which achieve the sensitivity of a laboratory unit with minimum servicing and reagent consumption have been described (94/C2033 94/C3469). The interference of iodides when tin@) chloride is used as the reducing agent in the CV generation of Hg can be eliminated by the addition of chromium(11).It has now been reported that the addition of an excess of iodide to such a system improves reproducibility and detection limit by stabiliz- ing the Hg (94/3041)! An intermittent second peak observed by Scott (94/1806) in the CVAAS determination of Hg was attributed to HC1 gas. This effect was overcome by adsorption of the HCl gas onto glass wool. Interferences in the efficiency of trapping of CV-generated Hg on an iridium-coated graphite surface were avoided by an additional heating step to clean the furnace following each Hg determination (95/250). Workers in the University of Oviedo (94/2200) investigating the potential of organized assemblies (micelles and vesicles) have recommended on-line drying of the Hg vapour.When the cationic surfactant didodecyldimethylammonium bromide (DDAB) and a membrane drier tube were used detection limits were improved 3-fold over the conventional aqueous technique. The effect of surfactant media on Hg CV generation in both FI and continuousJEow modes has been investigated by Gutierrez and Camara (94/3027). Different mechanisms appeared to be involved in the two flow modes. The best results were obtained in the continuous mode with sodium dioctylsulfosuccinate (SDSS) which gave a 50% increase in sensitivity over a conventional aqueous system with reduced noise level and improved signal stability. The same group of workers have also reported that the micellar media tetradecyl- trimethylammonium bromide (TTAB) and SDSS also reduces inter$erences by arsenic (III and v) cadmium copper lead nickel and selenium on the continuous flow CV determination of Hg (94/3360).Speciation of inorganic Hg" and methylmer- cury has been facilitated by the use of vesicle mediated HPLC coupled to CVAAS (95/325). The vesicular mobile phase was DDAB in acetate buffer containing 5% v/v of modifier methyl cyanide (acetonitrile) and 0.005% v/v 2-mercaptoethanol. The stationary phase was a C18 bonded silica column. The system has been applied to spiked sea-water and urine. Recoveries were 91-103% with detection limits of 0.1-0.2 pg 1-' Hg after off-line preconcentration of aqueous samples. Aqueous-phase ethylation of Hg species with NaBEt has proved to be advantageous in the speciation of Hg in environ- mental samples.In one application the Hg compounds were extracted from sediments by distillation followed by ethylation precollection on a Carbotrap isothermal GC and detection by AFS (94/1615). The distillation approach avoided the matrix effects in the ethylation stage which can arise from alkali digestion. Recoveries were 95 4% and the LOD was Journal of Analytical Atomic Spectrometry August 1995 Vol. 10 219R0.001 ng 1-1 MeHg as Hg. Other workers (95/369) who digested biological materials in methanolic KOH solution prior to ethyl- ation with NaBEt trapping and GC separation do not appear to have noticed any interference and achieved an LOD of 0.3 ng g-' MeHg. Refinement of the GC operating conditions in a Hg speciation system using the ethylation step eliminated the need for a liquid nitrogen cooled trap and avoided some of the thermal decomposition that had occurred during desorp- tion (95/254).The LOD for MeHg was 0.6 pg; up to 80 samples could be analysed in 8 h. High pressure liquid chroma- tography has been used to separate organomercury compounds (94/1689). The separated compounds were converted to Hg(0) in a continuous-flow system by an oxidizing solution (peroxo- disulfate and Cu" in dilute H2S04) followed by a reducing solution (alkaline SnC1,) both of optimized composition. The elemental Hg generated was swept into an AF spectrometer with argon. The determination of total Hg in whole blood by FI-CVAAS has been reported by Guo and Baasner (94/3142).Two oxi- dation stages were employed firstly a KBr-KBrO solution in a microwave heated knotted digestion coil was used followed when cool by a solution of KMn0 and finally Hg" was reduced to Hg(0) with NaBH in NaOH solution. Recovery of Hg was 96-100% and the LOD was 1 pg 1-l Hg in whole blood. The same workers developed a method for the determi- nation of Hg in urine (94/1610). The procedure was similar to that used for whole blood but was operated at room tempera- ture throughout without microwave heating. The LOD was 0.1 pg 1-' Hg. Other workers (94/1767) determining Hg in urine noted that SnC1 reduction leads to greater sensitivity than NaBH but careful sample decomposition was necessary if organomercury was to be measured. Tahan et al.(94/2187) used Purr digestion bombs heated either by convection or microwaves for acid decomposition of whole blood urine and tissue. There was no significant difference in the results from the two heating methods. Convection was slow (12 h) but could handle large numbers of samples microwave digestion required only 70 s per sample. The LOD was 6.6ng g-' Hg in the solid sample. The accuracy of the method was verified with NIST SRMs. Four methods of acid digestion of tissue prior to CVAAS determination of Hg viz ( i ) HN03 ( i i ) HN03-H2S04 ( i i i ) HN03-HC104 and (iu) HN03-H202 have been compared by Adeloju et al. (94/3130). Recovery of added Hg was nearly quantitative in all instances but that of MeHg was best with method ( i i ) . For the determination of total Hg in scalp hair dissolution of the sample with HN03 under pressure prior to CV generation of Hg has been used (94/2965). The Hg vapour was collected on a Au-Pt grid and de-amalgamated at 700°C into the AA spectrometer.The LOD was 0.13 ng Hg. An accuracy of 2.5% in the analysis of several biological and environmental CRMs and SRMs was achieved. Speciation of Hg in human hair has been effected by means of the selective leaching of MeHg with HCl (94/2928). The results agreed with those obtained by solvent extraction methods. Continuous flow and FI techniques achieve typical LODs for Hg of 10-50ng 1-' in aqueous samples (94/C2035 94/C2036). However levels of Hg in natural waters may be as low as 0.2 ng 1-' (94/C3458) hence means of improving sensi- tivity have been developed.A sorption method using the chelat- ing ion exchanger Spheron Thiol (ST) has been used for the preconcentration of Hg (95/306). Both organic and inorganic Hg were adsorbed directly from aqueous samples (1 1) onto the ion exchanger (0.1 g). The ST was separated by filtration and placed in a sample boat. The Hg was released thermally followed by trapping in an amalgamator prior to final thermal release into the absorption cell. These preconcentration pro- cedures gave an effective lo4 increase in sensitivity with an LOD of 0.2ng 1-' Hg in the original sample. Though the procedure just described does not use CV generation of Hg it has been presented here to draw attention to the possibility of high sensitivity Hg determination by alternative methods.Several reports describe methods using amalgamation to improve the sensitivity of CV Hg analysis with detection by AAS or AFS (94/C2032 94/C2035 94/C2036). Fully auto- mated CV-amalgamation/atomization systems have been developed which have both reduced blank values and lowered detection limits to 0.07 ng 1-' Hg (94/3285 94/C3458). The exploitation of such low detection limits is critically dependent upon the elimination of contamination and reagent blanks. 1.3.3. Volatile organo-metallic compound generation and metal vapour separation Traces of Co Fe and Ni in pure water have been determined by conversion to their carbonyls followed by atomization in a microwave-induced plasma and detection by AFS (94/3134). The sample was mixed with thiourea and tris( hydroxymethy1)amin- omethane in a carbonylator at 50°C and NaBH in NaOH solution added.Pure CO was passed through the carbonylator and the generated carbonyl swept through drying tubes into a cold trap (- 60 "C). After 20 min collection the trap was heated at 15 "C min- up to 50 "C. The released carbonyls were passed through the microwave plasma atomizer of an AF spec- trometer. Detection limits were 20 5 and 3 pg for Co Fe and Ni respectively. Metal vapour elution analysis has been applied to the separa- tion of Ag Cu Fe Mn and Nu prior to detection of the atomic vapours by AAS (94/2258). The sample was injected into a molybdenum tube (25 cm long 1.12 mm id) 7 cm of which was packed with tungsten powder. Following drying and pyrolysis of the sample the tube was heated to 1820°C and a gas flow of Ar-H initiated.Silver Mn and Na were successfully detected with retention times of 7.7 2.1 and 12.6 s and LODs of l.Opg 34pg and 25 pg respectively. The metal vapour elution technique was applied by Ohta et al. (94/3128) to the determination of Zn in aluminium metal. The Zn vapour was separated from Al Cu Fe Mg Si and Ti at a column temperature of 1650 K. The LOD was 12pg and the results were in agreement with those obtained by a spectrometric method and certified values. 1.4. Spectrometers A little bit of the early history of atomic absorption spec- trometry has been filled out by Willis (94/2245) who was working close to Walsh in the early 1950s. Apparently Walsh with a background of spectrochemical emission and IR absorp- tion began to wonder in 1952 why atomic spectra were always measured in emission while molecular spectra were studied in absorption.So promising were the experiments which he then initiated that the following year CSIRO filed a patent appli- cation for an AA spectrometer and in March 1954 the technique was demonstrated at an exhibition in Melbourne Australia. Walsh's classic paper (Spectrochim. Acta. 1955 7 108) soon appeared. This led several 'alert minds' to try the method for some trace elements and Mg in plant materials. This was the so-called 'do-it-yourself' period. It was only after Willis had demonstrated that calcium could advantageously be measured in blood serum ('one of the most difficult projects I could have undertaken') that a major manufacturer was interested in taking out a licence to manufacture.In the forty years to the present day the lowering ofdetection limits in A A and in AE has been a challenge to workers in many fields from environmental and clinical chemists to instru- ment designers and academics. Two papers from Winefordner's group on theoretical and practical limits in atomic methods (Spectrochim. Acta Part B 1993 48 757 94/2899 given also as a plenary lecture at the York CSI July 1993) showed how 220 R Journal of Analytical Atomic Spectrometry August 1995 Vol. 10limits of detection depended not only upon spectral selectivity and background noise but also on other analytical figures of merit such as efficiency of detection and efficiency of measure- ment. These relationships were discussed in detail as well as their significance to single atom detection.Advice on lowering detection limits in AAS was also offered by Gadsby (94/1849) although the methods recommended use of EDLs rather than HCLs where these are applicable; elimin- ating background; and selection of appropriate matrix modi- fiers must all be well known to alert minds of the present time. A new type of Doppler-free spectroscopy was heralded in a paper from Japan (95/111). Optical heterodyne detection was applied to the Doppler-shifted backward light scattering caused by the inhomogeneity of Rb atoms velocity-selected by a single-mode diode laser. It would seem however that consider- able development will be required before this principle can be used to reduce line-broadening in AA sources.1.4.1. Light sources Hollow cathode lamps boosted glow discharges electrode-free HF discharges and diode lasers have all been the subject of reports this year. Characteristics of HCLs in the pulsed current mode at low and high powers have been examined by Prudnikov and Shapkina (95/182). It appears that higher frequency pulsing at low power improves accuracy and sensitivity in FAAS particu- larly for Cu determinations in rock. The selection of power supply conditions for HCLs for use with Smith-Hieftje back- ground correction was investigated by Chinese workers (95/66) and recommendations for some 24 elements were tabulated. In a review with 72 references of boosted glow discharges for atomic spectroscopy including their use as primary sources in AAS Leis and Steers (94/3366) described the different types of such sources now known and illustrated their fundamental properties with specific examples.The review is particularly valuable because the amount of work published in this field is limited. Boosting can be achieved with an auxiliary discharge a magnetic field or by structural modifications. Lines of special analytical interest are often greatly increased in intensity though other modifications to the spectra may be complex. Analytical applications and limitations for each type of source were outlined. A series of patents concerning discharge lamps for A A was filed in Russia in 1992 and 1993. An electrode-free HF discharge lamp using a quartz-glass bulb filled with an inert gas was the subject of one of these (94/3082).It seems that some line background ratios may be improved by using ‘transition glass’ as the bulb material. Spectral gas discharge lamps were covered by two further patents (94/3081,95/123). However the descrip- tions of the lamps appear to be quite complicated and their operation equally so. No indications of possible analytical performance were given. Niemax and co-workers continued with their investigations into the application of diode lasers to AAS. Schnurer-Patschan et al. (94/2201) reported on the improved detection limits obtained in single-element ETAAS measurements by applying wavelength modulation techniques. The use of wavelength modulation reduced the influence of various sources of noise on the absorption signal. These workers found that the low absorbances that can be handled by modulation techniques allowed laser AAS measurements of either weak transitions from ground states or strong transitions from thermally popu- lated excited states to compete readily with traditional AAS measurements from atomic ground states. The laser radiation was detected with either a photodiode or photo-multiplier. For La and Rb measurements were made by absorption of the fundamental diode laser radiation at 670.95 and 780.03 nm from excited and ground states respectively whereas A1 was measured by deep blue light at 396.15 nm produced from the second harmonic generation of diode laser radiation in a lithium iodate crystal.Detection limits (3s) based on the total noise were found to be 0.1 ng ml-’ 7.5 ng ml-I and 1 pg ml-’ for Al La and Rb respectively.1.4.2. Continuum source and simultaneous multi-element AAS Although the principles of continuum source and of multi- element AAS are most usually combined in one instrument an interesting ‘inverse polychromator’ approach has been described by Kitagawa et a!. (95/30). The idea is not new of course more than 25 years ago Mavrodineanu and Hughes (Appl. Opt. 1968 7 1281) constructed a double polychromator system arranged back-to-back so that resonance radiation from a number of sources placed around the focal curve of one polychromator were combined into a single beam to pass through the sample then re-dispersed in the second polychrom- ator. Kitagawa (95/30) used the first polychromator but dis- pensed with the second by modulating the sources at different frequencies then detecting with a single PMT.Individual signals were then discriminated by means of fast Fourier transformation. Analytical performances were evaluated for simultaneous measurement of 10 elements in test solutions and in real samples. A technical breakthrough of considerable interest is described in detail by Shuttler et al. (94/C3430). This is the SIMAA 6000 from Perkin-Elmer the first simultaneous multi- element electrothermal atomic absorption system to be com- mercially available. Up to six elements can be measured using four single or multi-element line sources. The system also incorporates a stabilized temperature platform furnace longi- tudinal Zeeman-effect background correction and an echelle optical system with a custom-made solid state detector array.There are claimed to be no sacrifices in sensitivity or accuracy as compared with a single-element system. Radiation from the sources is combined in a tetrahedral prism before passing through the THGA. The monolithic detector incorporates diodes located at the line positions of the 38 most important AA elements and includes a number of secondary lines. The instrument is fully controlled from a PC using a Windows- based software package which includes analytical quality control protocols. The alternative approach to multi-element AA i.e. a con- tinuum source followed by polychromator and photodiode array detection has received more attention in several groups. Smith et al. (94/2184) evaluated such a system equipped with an electrothermal atomizer getting accurate results for a number of different SRMs.The polychromator was an echelle spectrometer with easily interchangeable PMT and LPDA cartridges. It was claimed that detection limits for a number of common elements were similar with both PMT and LPDA detection. Analytically however the latter was more useful as integration of the signal across the line profile increased the linear range of the calibration curves. This point was made again by Smith this time working with Harnly (94/C1942). The graphite furnace used in the later work was capable of operation both under vacuum and at pressures up to 10 atmospheres. Atomization at elevated pressures produced a significant increase in sensitivity for a number of elements though it also caused line broadening and a shift in the absorption profile.The absorbance as measured by the CSAAS/LPDA detection system was not affected by such shifts and therefore represents an advantage over line source AAS. This work is also reported in ref. 94/C3467 (see Section 1.2.5.). Harnly and co-workers continue to explore the application of xenon arc continuum source lamps to simultaneous multi- element E TAAS. The conference presentation concerning the use of pulsed continuum sources and discussed last year (see Journal of Analytical Atomic Spectrometry August 1995 Vol. 10 221 RJ. Anal. At. Spectrom. 1994 9 213R) has now been published (94/2948). A series of 300 and 500 W xenon arc lamps normally operated at 20 and 35 A respectively were pulsed as high as 300 W (0.5 ms pulse 3.75% duty cycle) to achieve higher intensities in combination with linear photodiode array detec- tion.Comparing dc and normal ac operation the 500 W lamp was in general a factor of 1.5 times more intense except below 210 nm where the 300 W lamp was more intense. Pulsing at high current provided a significant increase in the integrated signal from the LPDA and resulted in a factor of up to 80 times improvement in the detection limit for a new lamp. However pulsing at 200 A did induce failure of both the 300 and 500 W lamps after 10 h of operation the equivalent of about 200 atomizations. While reducing the pulse current increased the lamp lifetime there were less significant improve- ments in the detection limits. After 50 h of operation at 100 A the performance of a 300 W lamp was comparable to dc operation.In view of these findings the authors conclude that an improved lamp design is necessary before pulsing of xenon short arc lamps will be economically attractive for CSAAS instruments. Also from Harnley's group (94/C3468) came investigations on the eflects of furnace parameters on simultaneous multi- element AA measurements using THGAs. Using Pd-Mg (NO,) as matrix modifier it was established that at an atomization temperature of 2500 "C the minimum atomization temperature for V losses of Cd Cu Cr and Pb ranged from 3-34% depending on the tube arrangement (see also Section 1.2.5.). Of the two papers from Sneddon's group in Los Angeles USA the first (94/1649) was on a simultaneous theme. It described the optimization using a variable size simplex pro- cedure of the overall response of a simultaneous multi-element flame AAS.Seven instrumental parameters were subjected to the procedure for measurement of Cu Fe Mn and Zn. Although the selected conditions were necessarily a compro- mise it was found that sensitivity was in no case reduced by a factor of more than two. The second paper (94/3072) is a review with 50 references of the development performance and application of multi-element AAS covering multi- element HCLs continuum sources multichannel systems etc. 1.4.3. Background correction in AAS Further investigations into the subject of linearization ofanalyt- ical curues obtained by Zeeman-effect background corrected ETAAS have been published during the period covered by this update.A number of these were discussed in the previous Update as conference reports (see J. Anal. At. Spectrom. 1994 9,213R). L'vov et al. (94/3029) have automated their procedure whereby the dip formed in the absorbance at the pulse maxi- mum after roll-over can be corrected for. Previously the correction procedure had to be applied after operator evalu- ation of the absorption signal. A fully automated correction procedure has now been implemented. Written in BASIC the peak is first subjected to Savitzky-Golay smoothing and then searched for local maxima at a minimum height and spatial separation. The program then chooses the two largest maxima as representing the peaks on either side of the peak dip and then applies the corrections.The program was applied to the determination of Bi in the presence of a palladium chemical modifier and Ag and Cd. However problems were encountered with peaks for which a dip was barely discernible near the region where the dip first occurs. The workers proposed that this could be overcome either by repeating the determination with a smaller sample size or improving the algorithm. The work of L'vov et al. regarding the calculation of photometric error and discussed in the previous Update (see J . Anal. At. Spectrom. l994,9,213R) has now been published (Spectrochim. Acta Part B 1994 49 1609). The workers conclude that the contribution of the photometric error to the total error of analyte mass determined for moderate and large signals is insignificant and consequently the dip restoration method provides approximately the same error in mass measurement for both pre- and post- roll-over absorbance signals.They believe that the proposed theoretical approach to the evalu- ation of precision and detection limits may be considered as a new step in the development of absolute or standardless analysis as to date only parameters such as the characteristic mass and curvature of the analytical curves have been theoretically considered. Another group active in investigating the algorithm devel- oped by L'vov and co-workers for linearization of Zeeman- eflect ETAAS analytical curues is that based at the University of Connecticut. The conference report from Slavin et al. discussed last year has now been published (see J . Anal.At. Spectrom. 1994 9 213R). Su et al. (94/3070) applied the linearization approach developed by L'vov and co-workers to assess the eflect of stray light on the characteristic mass in Zeeman-effect ETAAS. The characteristic mass (mo) effective stray light (a) Zeeman-roll-over absorbance (A,) and the Zeeman-effect sensitivity ratio (R) were studied systematically as a function of lamp current and slit-width. The characteristic mass was found to be directly related to the effective stray light and the Zeeman sensitivity ratio while the Zeeman-effect roll-over absorbance was a quantitative measure of the effective stray light. By taking a and R into account a stable character- istic mass value 'm0 the 'corrected characteristic mass' was obtained. Without correction for these parameters the RSDs for measured m values were 26,20 8 and 19% for Ag Cr Cu and T1 respectively for a wide range of lamp currents and/or slit-widths.After correction for these parameters the corrected characteristic mass values were more stable with RSDs of only 7% for Ag and 4% for Cu Cr and T1. Yuzefovsky et al. (Spectrochim. Acta Part B 1994 49 1643) have refined the L'vov algorithm to include Newton's method of successive approximations to approach a solution to the theoretical expression of L'vov' and co-workers. The L'vov algorithm to linearize analytical curves is based on three parameters namely the roll-over absorbance (Ar) the limiting absorbance (Alh) and the Zeeman sensitivity ratio (R) but does not have an exact mathematical solution; rather a simplifying assumption is made that the Zeeman sensitivity ratio has a unity value and this makes it possible to linearize the analytical curves.This work introduced an additional parameter the sensitivity ratio at the rollover point (R') which was found using Newton's method and considered in the algorithm along with the roll-over absorbance. The extended algorithm and that of L'vov and co-workers were compared with analytical curves generated for Ag Bi Cu Mn and T1. For elements with relatively high R values (approximately 0.6 or greater) such as Ag Bi Mn and T1 the linearized parts of the analytical curves were practically the same for both models. The discrep- ancy in the values of the corrected characteristic mass for the linearized part of the analytical curve for both models for a large variety of experimental conditions remained within several per cent after taking the actual value of R into account in the case of the extended model or without it where R= 1 for the L'vov model. For Cu with a relatively low R value (< 0.9 the extended method improved the experimental characteristic mass value by 23% compared with the L'vov model.However it is clear from the conclusions that a major problem with the linearization models namely the accurate and reproducible measurement of the various parameters at the roll-over point may be aided by the application of the extended algorithm since the R value generally differs from 222R Journal of Analytical Atomic Spectrometry August 1995 Vol. 10unity and varies with analyte concentration for almost all elements.de Loos-Vollebregt et al. (94/2051) discussed the extension of the dynamic range that can be obtained by applying a 3-jield ac Zeeman-efect. Up to now it has generally been considered that the application of the 3-field Zeeman-effect would extend analytical curves by factor of 10. This study is the first to investigate thoroughly the use of this technique by using a commercially available transverse heated graphite atomizer with longitudinal Zeeman-effect modified to allow 3-field measurements and assessing a variety of representative elements with respect to analytical importance and Zeeman- effect splitting. At best the 3-field Zeeman-effect offers a 5-fold extension of the dynamic range. The slope of the 3-field analytical curve can also be selected by adjusting the intermedi- ate field strength. However acceptable precision of the meas- ured analyte concentrations over the increased concentration range was obtained only when the slope of the 3-field curve was reduced by a factor of about 2-3 in comparison with the corresponding 2-field Zeeman-effect analytical curves.These workers concluded that the extension of the dynamic range of the analytical curve was obtained only at the cost of less precision for the measured analyte. A new contributor to the field of extending the analytical dynamic range was Harnly (94/C3452) who provided a timely theoretical consideration of calibration for Zeeman-effect back- ground corrected ETAAS applying either normal linearized or 3-field methods.Computer modelling was applied to com- pare the working ranges and relative concentration errors (inverse of the S/N ratio) for all three methods. For peak absorbance measurements all three methods had approxi- mately the same relative errors although the 3-field method offered a slight advantage when the most sensitive analytical curve approached a maximum absorbance. For integrated absorbance measurements the linearized method was found in theory to extend the analytical curve by at least 1.5 orders of magnitude compared with the other two methods. However it is clear that this study from a theoretical consideration of the errors obtained from generating the signals and the correc- tion procedures which will limit how far the procedures can be applied does not find the large increases in dynamic range that have been previously claimed by the proponents of the linearization procedure.It is to be hoped that this material will eventually be published and contribute to the debate about the viability of linearizing analytical curves. In further considering the theoretical aspects of Zeeman- effect AAS instruments and the effects of various instrumental parameters on the linearity of analytica2 curves Voigtman et al. (Spectrochim. Acta Part B 1994 49 1629) have modelled a transverse Zeeman-effect background corrected ETAAS spec- trometer using a polychromatic optical calculus simulation software program. The behaviour of the analytical curves in the presence of three different types of stray light was modelled. Similar ‘modelled’ curves to those found by experiment were obtained.From the initial results it seems most likely that the major source of stray light in real Zeeman-effect measurements is due to the polychromaticity of the light source even when the light source is relatively narrow in spectral bandwidth. The workers intend to examine with the aid of this model how a sound theoretical basis can be established for roll-over dip correction and linearization algorithms. Stephens (94/3 103) described a field-on-source design for Zeeman-effect background correction. Although conventional Zeeman-effect background correction systems use magnetic fields applied around the atomizer this has the disadvantage that the magnet must be big enough to produce a suitably strong field in the area containing the atomizer. In this work the magnet size on the specially designed source was reduced and so mounted that a stable dc discharge could be maintained.As the magnetic fields were essentially parallel to the optical axis over the discharge region the longitudinal Zeeman-effect (no n component and circularly polarized 0 components) was observed. In order to use this configuration with the steady field of the bar magnet a photoelastic modulator was placed in the optical path to allow selective modulation of the intensity difference between the 0 and 0- components. The system was tested with Cu and found to function with satisfactory background correction capabilities. However the occurrence of spectral interferences due to line background could be a problem in such a system.It was concluded that the need for a specially designed source is a significant disadvantage but that this could be overcome as the size of the magnet used in this study is sufficiently small that the whole assembly includ- ing the magnet can fit inside the envelope of a conventional hollow cathode lamp. A conference paper from Dulude et al. of Jarrell Ash (94/C 1948) reports further investigations of the Smith-Hieftje technique which aim at the elimination of potential dc emission interference. Methods of temporal averaging of the dc radiation before and after signal measurement were evaluated. The results were incorporated into a modified system for measuring Pb and Cd in blood and the analytical performance was discussed.Background correction in a frequency-modulated simul- taneous AAS described by Lehnert et al. (94/1633) relates to a specific instrument. Radiation from three HCLs and a deuterium lamp was combined using fibre optics passed through the flame atomizer then separated with fibre optics and interference filters. Modulation demodulation and signal selection were achieved with lock-in amplifiers. The inclusion of the deuterium lamp in this simple system apparently resulted in an efficient multi-element method applicable to the analysis of complex matrices and represents an interesting alternative to known instrumentation. A graphite furnace AAS with a ‘dual option background correction facility’ was presented as a conference paper from AT1 Unicam (94/C3479). No one single system they believe provides the unique solution for every analysis though what options are made available was not clear from the abstract.In contrast an inexpensive portable AAS using ‘near-line’ back- ground correction by monitoring two closely lying wave- lengths from the same lamp (94/C2019) has been successfully used for measuring Pb in blood at the ppb level. 1.5. Instrument Control and Data Processing 1.5.1. Instrument control ‘News on fundamental reference data’ is the title of a review with 12 references (94/3140) of recent publications on this aspect of atomic spectroscopy. Rapid repetitive scanning of spectrometer pass wavelengths using a rotating 10mm quartz cube is described by McNeill et al (94/2544). The width of scan achieved was much larger than the spectrometer bandpass.Scanning frequencies greater than 300 Hz were possible. The output waveform was computer processed and displayed using software written in TURBO PASCAL. Designing AA instrument control software that works would seem to be a desirable and basic objective of any reputable instrument manufacturer and a joint paper on this subject from Perkin-Elmer and Cognetics Corporations (94/C1936) discusses the problems that a team designing user interfaces needs to solve in order to assure convincing performance. To deal with the complexity of those problems the team members need very diverse skills including chemistry graphics cognitive psychology communications and computer programming and the ability to interpret requirements voiced at customers’ group discussions and training courses or suggested by experience of Journal of Analytical Atomic Spectrometry August 1995 Vol.10 223Rprevious versions of software and instrumentation. This is an iterative process the result of which in this particular instance is an effective Windows-based system which has better capabili- ties yet is easier to use than earlier systems because of the emphasis on ‘usability engineering’ in its design. 1.5.2. Data processing All the papers in this section originated in China. Two happily were published in English language journals. The first of these claimed that a computer program of general applicability AAS-TOOLS (94/3102) facilitates ( i ) fitting and filtering of analytical data; ( i i ) on-line data collection in a specific analysis; ( i i i ) determination of kinetic parameters for a specific element; and (iv) theoretical simulation of an AAS signal based on an exponentially modified Gaussian function.It was used in conjunction with a PE 4000 AAS with deuterium-lamp back- ground correction and an HGA 400. The work aimed to present integrated software for on-line data collection and processing and to provide algorithms that are concise and as widely valid as possible. Analysis without calibration curves for the measurement of In using ETAAS was the subject of the second paper (95/270). The characteristic mass values for furnace wall platform and boat were compared at 1600-2400 “C. Satisfactory results were claimed for In in standard and geological reference materials. A third Chinese paper (this time not available in English 95/280) described an experimentally designed simplex optimiz- ation method which was claimed to be simpler more reliable and more accurate than the common orthogonal experimental design. The method was recommended when there were many factors to be considered and especially when there were significant cross effects between them.As so often happens with abstracts from Chinese papers no detail is available and so it is difficult to see in what way there can be an improvement over the simplex systems now commonly used by western researchers. 1.5.3. Chemometrics Chemometrics is the application of advanced statistical and mathematical techniques to the field of analytical chemistry. Inexpensive but powerful microcomputers will become it is said in an American environmental review (95/120) a standard part of the coming generation of analytical instruments.Sophisticated pattern recognition algorithms can be incorpor- ated into such instruments liberating them from the laboratory and turning them into ‘smart analysers’ for a variety of chemical identification quantification control and alarm tasks. Various types of on-line and in situ instrumentation were quoted. Although AAS and AFS specifically were absent from the list they could well have been included because as it is suggested these developments will permit better control and in particular a closer degree of monitoring of potential sources of pollution by enforcement agencies and the like. A conference paper from Varian (94/C1935) described new generation automated real time quality control software for the validation of data in AAS.Thirteen tests each with its own selectable limits could be programmed. Practical appli- cations to both flame and furnace atomization methods were discussed. The concept of self-monitoring systems based on the extended Kalman filter has been applied to ETAAS by Wienke et al. from Nijmegen (94/2044). The method was based upon on-line state estimation by the Kalman filter extended by quality control sampling. Advantages of the proposed method over conventional approaches were that it performed simul- taneous calibration and recalibration detection and correction of drift and detection and repair of outliers. It also required only a minimal number of QC samples for parameter updating as compared with the existing Kalman filter algorithm.An analytical procedure with ETAAS was validated after exhaustive testing. From Penninckx et al. in Brussels (94/C3434) come details of a ‘knowledge-based‘ computer system which assists vali- dation of AAS methods by detection of matrix interferences. The method was based on comparison of the slopes of standard additions and aqueous standard calibrations. As in practice these slopes are evaluated from a limited number of experi- ments there is always some probability that a significant difference may not be detected. In the system described this probability was shown to be acceptably small. The method was run using Windows software. 2. ATOMIC FLUORESCENCE SPECTROMETRY Each year a limited number of reports incorporating AFS are received for inclusion in this section of ‘Updates’ however the technique has not achieved the widespread acceptance enjoyed by its rival spectroscopic techniques despite 30 years of elo- quent advocacy by its supporters.It is therefore reasonable to conclude that it is unlikely that the technique will ever become widely used. There are however some analyses e.g. determi- nation of hydride forming elements in environmental samples for which AFS is well suited. Reviews of AFS analyses have been published by Stockwell and Corns (94/2684) and Guo (94/2627). Commercial AFS detectors suitable for the determination of As Hg and Se species when separated by chromatography have also been described (94/C203 1). 2.1.Discharge Lamp-excited Atomic Fluorescence Several methods of atomizing samples for AFS analysis have been examined. The simplest approach was that used for the determination of Hg in sediments (94/C3490). A small sample was heated in a quartz tube and the released Hg trapped prior to its rapid release into the AF detector. The method proved to be simple reliable and accurate. A most comprehensive review of the use of the ICP in AFS has been prepared by Greenfield (94/2969 94/C3426). Three systems were con- sidered; the ICP as ( 1) light source (2) atomizer (3) light source and atomizer. The advantages and limitations of these systems were fully examined and compared with those of AAS and AES. The author concludes that there is a place for ICP-AFS but it is unlikely to be generally exploited until commercial systems are established as viable.In the meantime in-house combinations of ICP and atomizer systems may be used to further define the benefits of ICP-AFS. A microwave plasma torch has been evaluated as an atomizer ofAg As Cd Fe Hg Mg Mn Pb and Zn for AFS determination (94/1622,94/3044). The detection limits for Cd and Zn were 0.26 and 1.2 pg l-l respectively. A spiral tungsten atomizer has been constructed for the determination of trace impurities in micro-volume liquid samples (94/2106). The LODs for Ag Bi Cd Cu Mn Pb and Zn ranged from 1 ng 1-’ to pg 1-’ in 2 p1 samples. The glow discharge (GD) has been shown to be a useful atom source for the direct sputtering of solid samples for AFS (94/3271). A xenon arc lamp and xenon flash-lamp were used as excitation sources and gave an improved S/B ratio for GD-AFS compared with GD-AES.The S/B was further improved by pulse operation of both the GD and the flash- lamp; the pulsing of the latter being delayed by 3 ms after completion of the GD pulse. 2.2. Laser-excited Atomic Fluorescence Lasers are used in atomic spectroscopy not only as excitation sources in AFS but also to atomize and ionize the analyte and as a diagnostic tool to probe atom vapours. Two reviews have examined these broader aspects of laser usage (94/2060 224 R Journal of Analytical Atomic Spectrometry August 1995 Vol. 1094/2068) while a third (95/147) has considered the potential uses of a new generation of semi-conductor diode lasers in atomic and molecular spectroscopy.As in all atomic spectroscopies the production of the atomic uapour for AFS is a key step in achieving sensitive and accur- ate analysis and consequently stimulates research activity. Electrothermal atomizers are currently popular devices for use with LEAFS. The operation of a graphite furnace at reduced pressure ( x 100 mTorr) for the determination of Co P Pt and Te in solid samples (nickel alloys and glass) has been investi- gated by Lonardo et al. (94/C1949). Though interferences were reduced and peak shape improved there was a decrease in sensitiuity of 2 orders of magnitude which was attributed to the use of low pressure. Other workers have developed an electrothermal atomizer in which the sample uapour remains in the atomizer for up to 10 min thereby facilitating the sequential determination of Fe Ni and Co (94/3092).A capacitive dis- charge heatedfurnace has been investigated as an atomizer for the determination of T1 (94/1853). An LOD of 5 fg of T1 was obtained with a laser repetition rate of 500Hz and peak integration time of 80 ms. Using this system TI was determined in NIST biological samples at levels of 1-2 orders of magnitude below those obtained with ETAAS with a precision of 8-20%. A laser system has been developed for the ETAFS determi- nation of As and Se in blood samples (94/2951). The system which generated radiation in the VUV region consisted of two dye lasers pumped by a Nd:YAG laser the output radiations were frequency doubled by a non-linear optical KDP crystal and sum frequency mixed by a BBO crystal.The LODs were 54 fg As and 15 fg Se using 10 p1 aliquots. Graphite furnace AFS has been demonstrated to be suitable for the determi- nation of trace elements in sea-water and polar ice. Lead was determined directly in sea-water without separation concen- tration or the addition of a chemical modifier (94/2931). The practical detection limit was 3 fg of Pb. Cobalt in sea-water required preconcentration (95/311). A bonded silica column with octadecyl functional groups (CIS) was used and gave a concen- tration factor of 12.5; the detection limit in ocean water RMs was 1.0 ng 1-’. Detection limits lower than those of other analytical methods have been reported by Daskin et al. (94/3039) for the determination of Si in high purity indium when the solid sample was atomized in a planar magnetron discharge.Laser excited molecular Juorescence has been used by Michel and co-workers to determine C1(94/2960) and F (95/312). For the determination of C1 mixed solutions of indium and C1 in an elemental ratio between 100 and 1OOO 1 were placed on a L‘vov platform in the furnace. Indium monochloride was believed to be formed at the vaporization stage. Fluorescence was excited at 267.21 nm and detected at 269 nm to give an LOD of 1Opg C1. In the determination of F an earlier study by Butcher et al. (J. Anal At. Spectrom. 1991 6 9) found that barium when added to excess magnesium promoted the forma- tion of magnesium monofluoride and enhanced the sensitivity 100-fold. The purpose of the present work (95/312) was to elucidate the mechanism of the barium enhancement.Theoretical and experimental observations suggested that barium changed the mechanism of MgF formation by forming BaF as the first stage of the reaction during the heating process to be followed at a higher temperature by the appearance of Mg from the thermal decomposition of MgO and of F from BaF,. 2.3. Studies of Flames Plasmas and Atomic Vapours using Laser-induced Fluorescence A technique for obtaining a temporally resolved two-line NO temperature image of a complex combustion-related flow field has been described (94/2992). The temperature measurements were obtained with two lasers and two intensified cameras with a two-line ratio of planar laser-induced fluorescence from NO seeded either in the fuel jet or in both the jet and cross flow.Factors affecting the accuracy of the measurement were studied in depth. The OH density in an atmospheric air furnace has been studied by means of LIF of the OH radical (94/2994). The measured OH density was found to agree well with the computed OH chemical equilibrium density over a temperature range of 1500-1850K. The system could be used for the calibration of OH measurements. A computer controlled time-resolued laser-induced breakdown spectrometer has been used by Nemet et al. (95/148) to measure the plasma emission and fluorescence of ores metals and alloys. A review of the properties ofthe hollow cathode discharge as revealed by LIF and high-resolution spectra has been published (95/87). Previously undetected spectral lines were found spontaneous emission branching ratios were measured and other fundamental atomic processes studied.Omenetto (94/2062) has reviewed the use ofLIF and LEI as tools for the study of atomic and molecular reseruoirs. Although each technique has in itself many attractive analytical and diagnostic capabilities their interaction results not only in better understanding of the dynamics of the excited states but also provides new spectroscopic information and fundamental parameters of the system under investigation. High resolution spectroscopy of Rb atoms with a line-width-narrowed diode laser at 780.1 nm was used to study hyperfine spectra and magnetic field effects (94/2065). It was found that the hyperfine spectra were very sensitive to the geomagnetic field.3. LASER ENHANCED IONIZATION Two-step LEI is believed to be more sensitive and selective than one-step LEI. Su and Lin (94/2983) have examined that proposition both theoretically and experimentally. The optim- ization of a two-step system is complex as it is multifactorial. A simple theoretical model was developed to assist in the optimization process. They concluded that any enhancement of sensitivity in the two-step system was critically dependent upon second laser intensity the transition probability of the second step excitation collisional ionization rates and the transition line width. The step-wise resonant 3-photon ionization spectrum ofthe neutral Z r atom using three separately tuneable pulsed visible dye lasers has been used by Page et al. (94/2061) to probe the lifetimes of even parity levels.They found the direct ionization cross section to be <1O-”cm2 and lifetimes of the order of 10-100 ns. The analytical applications of LEI have been reviewed by Letokhov (95/40) with particular reference to trace element determination. Trace amounts of Lu (0.1 pg g-’) in geological samples have been determined by two-step laser excitation (94/2191). The sample was vaporized from a graphite crucible under vacuum. By extrapolation of the calibration curve measurements down to 0.1 ng of Lu were claimed to be possible. Epler et al. (94/2394) have combined LC and LEI to give a very sensitive method for the measurement of organolead species. The LC stage separates interferents from analytes and the great sensitivity of LEI facilitates the determination of the latter in dilute solutions after flame atomization.The method was applied to NIST SRM 1566a Oyster Tissue. A new laser ion source has been developed with a cavity made of high purity pyrolytically coated graphite (94/2678). The source achieved a 14% ionization efficiency in the trace analysis of Tc an efficient path for the resonance ionization of Sn and the determination of actinides in environmental trace analysis. Actinides have also been determined by resonance ionization following electrolytic deposition onto a rhenium substrate and thermal release (94/2723). Surface contamination of silicon by Nu was measured by laser ablation of the surface followed by LEI (95/273). The technique was applied to the Journal of Analytical Atomic Spectrometry August 1995 Vol.10 225 Rdetermination of the Na impurity flux in the edge region of a tokamak plasma by collecting Na on silicon plates placed in that zone of the plasma. The high sensitivity of LEI facilitated the analysis of single discharges. The desorbtion of H molecules from clean and sulfur-covered polycrystalline nickel was stud- ied by Pozgainer (95/158) using one-step LEI. The experiments revealed a strong rotational cooling of the desorbing molecules while the vibrational energy was similar to that of the ther- malized beam. LOCATION OF REFERENCES The full list of references cited in this Update have been published as follows 94/961-94/1264 J. Anal. At. Spectrom. 1994,9( 4) 135R-146R. 94/1265-94/1830 J. Anal. At. Spectrom.1994 9( 5) 149R-169R. 94/1831-94/2175 J. Anal. At. Spectrom. 1994,9(6) 189R-200R. 94/2176-94/2412 J. Anal. At. Spectrom. 1994,9( 7) 203R-212R. 94/2413-94/2867 J. Anal. At. Spectrom. 1994 9( 8) 249R-265R. 94/2868-94/2994 J. Anal. At. Spectrom. 1994,9( lo) 307R-312R. 94/2995-94/3279 J. Anal. At. Spectrom. 1994,9( 1 l) 307R-318R. 94/328&-94/C3502 J. Anal. At. Spectrom. 1994,9( 12) 357R-364R. 95/1-95/182 J. Anal. At. Spe,ctrom. 1995 10( l ) 1R-60R. 95/183-95/469 J. Anal. At. Spectrom 1995 10( 2) 49R-59R. 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. Abbreviated List of References Cited in Update 9411047 Clin.Chem. (Washington DC) 1993,39,800.94/1272 Appl. Spectrosc. 1993,47 1222.9411360 Carbon 1993,31,273. 9411605 Anal. Chem. 1993 65 2590. 9411609 Anal. Chem. 1993 65 3098. 9411610 Anal. Chim. Acta 1993 278 189. 9411611 Anal. Chim. Acta 1993 279,241.9411612 Anal. Chim. Acta 1993 279 253. 9411615 Anal. Chim. Acta 1993,281 135. 9411616 Anal. Chim. Acta 1993 281 185. 9411622 Fenxi Huaxue 1993 21 610. 9411624 Fenxi Huaxue 1993 21 740. 9411625 Fenxi Huaxue 1993,21 745. 9411626 Fenxi Huaxue 1993 21 761. 9411628 Fenxi Huaxue 1993 21 825. 9411631 Fresenius’ J. Anal. Chem. 1993 345 475. 9411633 Fresenius’ J. Anal. Chem. 1993 346 392. 9411636 Fresenius’ J.Anal. Chem. 1993,346 550.9411639 Fresenius’ J. Anal. Chem. 1993 346 622. 9411640 Fresenius’ J. Anal. Chem. 1993 346 643. 9411642 Fresenius’ J. Anal. Chem. 1993 346 683. 9411643 Fresenius’ J. Anal. Chem. 1993 346 689. 9411644 Fresenius’ J. Anal. Chem. 1993 346 711. 9411649 Talanta 1993 40 879. 9411650 Talanta 1993 40 1077. 9411665 Anal. Lett. 1993 26 1997. 9411667 Anal. Sci. Technol. 1992 5 349. 9411689 Appl. Organomet. Chem. 1993 7 173. 9411690 Appl. Organomet. Chem. 1993 7 213. 9411729 Clin. Chem. (Washington D.C.) 1993,39,1662.94/1767 GIT Fachz. Lab. 1993,37,411.94/1785 Huanjing Baohu (Taipei) 1992 15 29. 9411806 J. Anal. Toxicol. 1993 17 196. 9411821 J. Pharm. Biomed. Anal. 1993 11,301.94/1825 Jilin Daxue Ziran Kexue Xuebao 1992,3,85. 9411838 Laboratoriumsmedizin 1992 16 419.941 1849 Muter. World 1933 1 397. 9411853 Microchem. J. 1993 47 363. 9412044 Anal. Chem. 1994,66,841.94/2049 Fresenius’ J. Anal. Chem. 1993,346,873.9412051 Spectrochim. Acta Part B 1993 48,1505.9412057 Glow Discharge Spectrosc. 1993,67.94/2060 Hwahak Sekye 1993 33 658. 9412061 Inst. Phys. Conf. Ser. 1992 128 63. 9412062 Inst. Phys. Con$ Ser. 1992 128 151. 9412065 Jpn. J. Appl. Phys. Part 1 1993 32 3291. 9412066 Lab. Rob. Autom. 1993 5 299. 9412068 Lebensm.- Biotechnol 1993,10,125.94/2106 Prib. Tekh. Eksp. 1993,2,120.94/2178 J. Anal. At. Spectrom. 1993 8 955. 9412184 J. Anal. At. Spectrom. 1993 8 989. 9412185 J. Anal. At. Spectrom. 1993 8 995. 9412186 J. Anal. At. Spectrom. 1993 8 999. 9412187 J. Anal. At. Spectrom. 1993 8 1005. 9412188 J. Anal.At. Spectrom. 1993 8 1011. 9412189 J. Anal. At. Spectrom. 1993 8 1015. 9412191 J. Anal. At. Spectrom. 1993,8 1029.9412198 J. Anal. At. Spectrom. 1993 8 1085. 9412200 J. Anal. At. Spectrom. 1993 8 1097. 9412201 J. Anal. At. Spectrom. 1993 8 1103. 9412202 J. Anal. At. Spectrom. 1993,8 1109.94/2203 J. Anal. At. Spectrom. 1993 8 1113. 9412204 J. Anal. At. Spectrom. 1993 8 1117. 9412209 J. Anal. At. Spectrom. 1994 9 1.9412210 J. Anal. At. Spectrom. 1994,9 7.9412211 J. Anal. At. Spectrom. 1994,9 11.9412215 J. Anal. At. Spectrom. 1994 9,39.94/2216 J. Anal. At. Spectrom. 1994,9,45. 9412240 Appl. Optics 1993 32 6082. 9412245 Clin. Chem. 1993 39 155. 9412249 At. Spectrosc. 1993 14 95. 9412250 At. Spectrosc. 1993 14 99. 9412254 Anal. Chem. 1993 65 1682. 9412258 Analyst (London) 1993,118 1031.9412266 Fresenius’ J. Anal. Chem. 1993,346,414. 9412275 Zh. Anal. Khim. 1993,48,648. 9412277 Zh. Anal. Khim. 1993 48 774. 9412281 Fenxi Ceshi Xuebao 1993 12 64. 9412294 J. Chromatogr. 1993 641 337. 9412298 Lihua Jianyan Huaxue Fence 1993 29 179. 9412299 Microchem. J. 1993 48 184. 9412394 J. Anal. At. Spectrom. 1994 9 79. 9412399 J. Anal. At. Spectrom. 1994 9 105. 9412400 J. Anal. At. Spectrom. 1994 9 111. 9412401 J. Anal. At. Spectrom. 1994,9 117.9412407 Fenxi Ceshi Xuebao 1993 12 5. 9412410 Fenxi Shiyanshi 1993 12 ( 5 ) 51. 9412411 Guangpuxue Yu Guangpu Fenxi 1993 13 110. 9412412 Guangpuxue Yu Guangpu Fenxi 1993 13 85. 9412544 Anal. Proc. (London) 1993,30,401.94/2547 Anal. Sci. 1993,9,447. 9412563 Bunseki Kagaku 1993 42 551. 9412565 Bunseki Kagaku 1993,42,599.94/2579 Fresenius’ J.Anal. Chem. 1993 346 1047. 9412580 Fresenius’ J. Anal. Chem. 1993 346 1058. 9412581 Fresenius ’ J. Anal. Chem. 1993 346 1062. 9412582 Fresenius’ J. Anal. Chem. 1993 346 1068. 9412587 Fresenius’ J. Anal. Chem. 1993 347 303. 9412594 Zh. Anal. Khim. 1993 48 1254. 9412595 Zh. Anal. Khim. 1993 48 1345. 9412623 Fenxi Shiyanshi 1993,12(4) 19.9412625 Fenxi Shiyanshi 1993 12(4) 63. 9412627 Fenxi Shiyanshi 1993 12(4) 92. 9412670 Inst. Phys. Conf. Ser. 1992 128 255. 9412678 Inst. Phys. Con$ Ser. 1992,128 313.9412684 J. Autom. Chem. 1993,15(3) 79. 226R Journal of Analytical Atomic Spectrometry August 1995 Vol. 1094/2697 J. Radioanal. Nucl. Chem. 1993 172 117. 94/2710 Lihua Jianyan Huaxue Fence 1993 29 265. 94/2712 Lihua Jianyan Huaxue Fence 1993 29 280. 94/2723 Muter.Res. SOC. Symp. Proc. 1992 260 157. 94/2878 Anal. Chem. 1994 66 1462. 94/2886 At. Spectrosc. 1993 14 187. 94/2888 Guangpuxue yu Guangpufenxi 1993,13(6) 100.94/2889 Lihua Jianyan Huaxue Fence 1994 30(1) 15. 94/2891 Fenxi Shiyanshi 1994 13(2) 20. 94/2892 Fenxi Ceshi Xuebao 1994 13( l) 49. 94/2899 J. Anal. At. Spectrom. 1994,9 131.94/2922 J. Anal. At. Spectrom. 1994 9 267. 94/2923 J. Anal. At. Spectrom. 1994 9 273. 9412925 J . Anal. At. Spectrom. 1994 9 285. 94/2926 J. Anal. At. Spectrom. 1994 9 291. 94/2928 J. Anal. At. Spectrom. 1994 9 303. 94/2931 J. Anal. At. Spectrom. 1994 9 315. 94/2932 J. Anal. At. Spectrom. 1994 9 321. 94/2933 J. Anal. At. Spectrom. 1994 9 333. 94/2934 J. Anal. At. Spectrom. 1994 9 337. 94/2948 J. Anal.At. Spectrom. 1994 9 419. 9412949 J. Anal. At. Spectrom. 1994 9 427. 94/2950 J. Anal. At. Spectrom. 1994 9 431. 94/2951 J. Anal. At. Spectrom. 1994 9 437. 94/2952 J. Anal. At. Spectrom. 1994 9 443. 94/2953 J. Anal. At. Spectrom. 1994 9 451. 94/2955 J. Anal. At. Spectrom. 1994 9 463. 94/2956 J. Anal. At. Spectrom. 1994 9 469. 94/2957 J. Anal. At. Spectrom. 1994 9 477. 94/2960 J. Anal. At. Spectrom. 1994 9 501. 94/2964 J. Anal. At. Spectrom. 1994 9 531. 94/2965 J. Anal. At. Spectrom. 1994 9 535. 94/2967 J. Anal. At. Spectrom. 1994 9 547. 94/2969 J. Anal. At. Spectrom. 1994 9 565. 94/2979 J. Anal. At. Spectrom. 1994 9 643. 94/2981 J. Anal. At. Spectrom. 1994 9 657. 94/2982 J. Anal. At. Spectrom. 1994,9,663. 94/2983 Appl. Spectrosc. 1994,48,241. 94/2985 Appl. Spectrosc.1994 48 273. 94/2992 Appl. Opt. 1993 32 7532. 94/2994 Appl. Opt. 1994 33 1115. 94/3000 Analyst 1993 118 1027. 94/3002 Analyst 1993 118 1417. 94/3003 Analyst 1993 118 1425. 94/3015 Fresenius’ J. Anal. Chem. 1993 346( 10-11) 882. 94/3018 Fresenius’ J . Anal. Chem. 1993 346( 10-1 l) 461. 94/3023 J. Anal. At. Spectrom. 1994 9 773. 94/3026 J. Anal. At. Spectrom. 1994 9 791. 94/3027 Spectrochim. Acta Part B 1993 48 1551. 94/3029 Spectrochim. Acta Part B 1993,48( 13) 1625.94/3030 Talanta 1993 40 1147. 94/3034 Talanta 1993 40(11) 1609. 94/3035 Talanta 1993 40( 1 l) 1677. 94/3036 Talanta 1993 40( 1 l) 1759. 94/3037 Talanta 1993 40( 12) 1829. 94/3039 Zh. Anal. Khim. 1993,48(4) 715.94/3040 Zh. Anal. Khim. 1993,48(5) 813. 94/3041 Anal. Sci. 1993 9(6) 859. 94/3044 Chin.Chem. Lett. 1993 4(7) 615. 94/3048 Fenxi Shiyanshi 1993 12(5) 47. 94/3049 Gaodeng Xuexiao Huaxue Xuebao 1993 14(7) 937. 94/3052 Guangdong Gongxueyuan Xuebao Ziran Kexueban 1991 8(4) 27. 94/3058 Inst. Phys. Conf. Ser. 1992 1218( Resonance Ionization Spectroscopy 1992) 225. 94/3070 Microchem. J. 1993 48(3) 278. 94/3072 Microchem. J. 1993 48(3) 318. 94/3073 Microchem. J. 1993 48(3) 326. 94/3081 U.S.S.R. SU 1,737,561 (Cl. HOlJ61/02) 30 May 1992 Appl. 4,813,072 12 Apr 1990. 94/3082 U.S.S.R. SU 1,737,565 (Cl. H01J65/04) 30 May 1992 Appl. 4,814,974 17 Apr 1990. 94/3083 PCTInt. Appl. WO 93 17,321 (Cl. GOlN1/34) 02 Sep 1993 AU Appl. 92/1,069 25 Feb 1992; 31 pp. 94/3092 Vysokochist Veshchestva 1993 (6) 114. 94/3100 Yejin Fenxi 1993 13(4) 55. 94/3102 J. Anal. At. Spectrom.1994 9 669. 94/3103 J. Anal. At. Spectrom. 1994 9 675. 94/3104 J. Anal. At. Spectrom. 1994 9 679. 94/3105 J. Anal. At. Spectrom. 1994 9 685. 94/3106 J. Anal. At. Spectrom. 1994 9 691. 94/3112 J. Anal. At. Spectrom. 1994 9 727. 94/3123 J. Anal. At. Spectrom. 1994 9 801. 94/3128 Anal. Chim. Acta 1994 285 53. 94/3130 Anal. Chim. Acta 1994 285 359. 94/3134 Anal. Chim. Acta 1993 283(2) 895. 94/3139 Spectrochim. Acta Part B 1993 48( 13) 1645. 9413140 Spectrochim. Acta Part B 1993 48(13) 1651. 94/3142 Talanta 1993 40(12) 1927. 94/3146 Zh. Anal. Khirn. 1993 48(12) 1906. 94/3147 Zh. Anal. Khim. 1993 48(12) 2012. 94/3155 Chem. Anal. (Warsaw) 1993 38(6) 771. 94/3156 Fenxi Shiyanshi 1993 12(6) 55. 94/3167 Labor Praxis 1993 117(4) 44. 94/3170 Microchem. J. 1994 49( l) 20.94/3175 Yankuang Ceshi 1993 12(2) 85.94/3244 J.Anal. At. Spectrom. 1994,9 857.94/3245 J. Anal. At. Spectrom. 1994 9 861. 94/3246 J. Anal. At. Spectrom. 1994 9 867. 94/3247 J. Anal. At. Spectrom. 1994 9 871. 94/3271 J. Anal. At. Spectrom. 1994 9 1039. 94/3279 Appl. Opt. 1994 33 18 3992. 94/3280 Anal. Chem. 1 Oct 1993 65(19) 2590. 94/3285 Anal. Chim. Acta 5 Oct 1993 282( l) 109. 94/3286 Anal. Chim. Acta 15 Nov 1993 283( l) 167. 94/3287 Anal. Chim. Acta 15 Nov 1993 283(1) 393. 9413288 Anal. Chim. Acta 30 Dec 1993 284(2) 361. 94/3289 Anal. Chim. Acta 20 Jan 1994,285( l) 33.94/3292 Anal. Chim. Acta 28 Feb 1994,286( 3) 343.94/3296 Anal. Proc. Mar 1994 31(3) 89. 94/3304 Appl. Spectrosc. Nov 1993 47(11) 1871. 94/3306 Bunseki Kagaku Nov 1993 42( 1 l) T147. 94/3311 Bunseki Kagaku Feb 1994 43(2) 105. 94/3317 J. Anal. At. Spectrom. Aug l994,9,213R 94/3319 Fenxi Huaxue Oct 1993 21(10) 1239. 94/3320 Fenxi Huaxue Jan 1994 22(1) 19. 94/3326 Fresenius’ J. Anal. Chem. Sep-Oct 1993 347(3-4) 82. 94/3329 Fresenius’ J. Anal. Chem. Nov 1993 347(8-9) 320-323. 94/3330 Fresenius’ J. Anal. Chem. Nov 1993 347( 8-9) 356-360. 94/3334 Fresenius’ J. Anal. Chem. 1994 348(3) 195. 94/3336 Fresenius’ J. Anal. Chem. 1994 348(3) 248. 94/3339 Fresenius‘ J. Anal. Chem. Feb 1994 348(5-6) 353. 94/3346 Spectrochim. Acta Part B 18 Oct 1993 48(12) 1495. 94/3349 Spectrochim. Acta Part B 20 Dec 1993 48( 14) 1703. 94/3360 Spectrochim. Acta Part B 25 Feb 1994 49(2) 163. 94/3363 Spectrochim. Acta Part B Mar 1994,49(3) 221. 94/3364 Spectrochim. Acta Part B Mar 1994 49(3) 229. 94/3366 Spectrochim. Acta Part B Mar 1994,49( 3) 289.95/1 Anal. Chim. Acta 1994 286 381. 9515 Fenxi Shiyanshi 1993 12 97. 95/12 Guangpuxue Yu Gangpu Fenxi 1993 13(4) 103. 95/13 Guangpuxue Yu Guangpu Fenxi 1994 14(1) 99. 95/23 Spectrosc. Lett. 1994 27 257. 95/30 Anal. Sci. 1993 9 663. 95/32 Anal. Sci. 1993 9 707. 95/33 Anal. Sci. 1993 9 723. 95/38 Asian J. Chem. 1993 5 1047. 95/39 Asian J. Chem. 1993 5 1078. 95/40 Ber. Bunsen.-Ges. Phys. Chem. 1993 97 1512. 95/41 Analusis 1994,22 37. 95/43 Bull. Chem. SOC. Jpn. 1993 66 2233. 95/55 Dojin News 1993 66 3. 95/56 Fenxi Ceshi Xuebao 1993,12 1. 95/66 Fenxi Shiyanshi 1994,13( l) 90. 95/80 Guangpuxue Yu Guangpu Fenxi 1993 13(1) 121. 95/82 Guangpuxue Yu Guangpu Fenxi 1993 13(3) 51. 95/83 Guangpuxue Yu Guangpu Fenxi 1993 13(3) 57. 95/87 Guangpuxue Yu Guangpu Fenxi 1993. 13(4) 23. 95/101 J. Autom. Chem. 1993 15 37. 951111 Jpn. J. Appl. Phys. Part 1 1993 32 5717. 95/112 J. Radioanal. Nucl. Chem. 1993 174 133. 95/118 Int. J. Enuiron. Anal. Chem. 1993 52 195. 95/119 Int. J. Environ. Anal. Chem. 1993 51 219. 95/120 Int. SAMPE Environ. Cons. 1991 l(Environment in the 1990’s A Global Concern) 290. 95/123 U.S.S.R. SU 2,804,597 (CI. GO1 J3/2O) 23 Mar 1993 Appl. 4,891,880 1993 ( l l ) 226. 95/124 Keikinzoku 1993 43 353. 95/126 LaborPraxis 1993 17( 8 ) 12. 95/135 Lihua Jianyan Huaxue Fence 1994 30(1) 26. 95/147 Proc. SPIE-Int. SOC. Opt. Eng. 1992,1711,197.95/148 Proc. SPIE-Int. SOC. Opt. Eng. 1993 1983 977. 95/152 Pure Appl. Chem. 1993 65 2465. 95/153 Quim. Anal. (Barcelona) 1993,12,30.95/158 Surf. Sci. 1994,307-309,344.95/163 Ukr. Khim. Zh. (Russ. Ed.) 1993 59 618. 95/165 Yankuang Ceshi 1993 12(2) 101. 95/174 Zavod. Lab. 1993,59 25. 95/182 Zh. Prikl. Spektrosk. 1993,58 560. 95/190 Anal. Chim. Acta 1993 284 367.95/191 Anal. Chim. Acta 1994,285,277.95/192 Anal. Chim. Acta 1994 287 247. 95/194 Anal. Proc. 1994 31 61. 95/199 Bunseki Kagaku 1994,43,169.95/200 Bunseki Kagaku 1993 42 767. 95/202 Fenxi Huaxue 1993 21 1258. 95/210 Anal. Sci. 1994 10 281. 95/216 Chem. Anal. (Warsaw) 1993 38,759.95/220 Diqiu Kexue 1993,18,810.95/224 Fenxi Ceshi Xuebao 1993 12(4) 32. 95/231 Guangpuxue Yu Guangpu Fenxi 1993 13 121. 95/235 Guangpuxue Yu Guangpu Fenxi 1993 13(5) 85. 95/241 Guangpuxue Yu Guangpu Fenxi 1994 14( l) 91,42.95/244 Spectrochim. Acta Part B 1993,48 1633. 95/245 Spectrochim. Acta Part B 1993 48 1639. 95/250 Spectrochim. Acta Part B 1993 48 1719. 95/252 Talanta 1993 40 1839. 95/253 Talanta 1994 41 67. 95/254 Talanta 1994 41 371. 95/255 Zh. Anal. Khim. 1994 49 150. 95/256 Zh. Anal. Khim. 1994 49 157. 95/260 Zh. Anal. Khim. 1994 Journal of Analytical Atomic Spectrometry August 2 995 Vol. 10 227 R49 361. 95/270 Mikrochim. Acta 1994 112 237. 95/273 Opt. Eng. (Bellingham Wash.) 1993 32 2487. 951277 Spectrosc. Lett. 1994 27 353. 95/280 Xibei Daxue Xuebao Ziran Kexueban 1993 23 109. 95/282 Yankuang Ceshi 1993 12( l ) 6. 95/283 Yankuang Ceshi 1993 12(1) 14. 951286 Yankuang Ceshi 1993 12(1) 57. 951291 Zhongguo Xitu Xuebao 1992 10 183. 951300 J. Anal. At. Spectrom. 1994 9 1129. 95/302 J. Anal. At. Spectrom. 1994 9 1143. 95/303 J. Anal. At. Spectrom. 1994 9 1153. 95/304 J. Anal. At. Spectrom. 1994 9 1161. 95/305 J. Anal. At. Spectrom. 1994 9 1167. 95/306 J. Anal. At. Spectrom. 1994 9 1173. 95/307 J. Anal. At. Spectrom. 1994 9 1177. 95/311 J. Anal. At. Spectrom. 1994 9 1195. 95/312 J. Anal. At. Spectrom. 1994 9 1203. 95/321 J. Anal. At. Spectrom. 1994 9 1255. 95/322 J Anal. At. Spectrom. 1994 9 1263. 951324 J. Anal. At. Spectrom. 1994 9 1273. 951325 J. Anal. At. Spectrom. 1994 9 1279. 95/327 J. Anal. At. Spectrom. 1994 9 1289. 95/331 Can. J. Appl. Spectrosc. 1994 39(2) 54. 951333 Can. J. Appl. Spectrosc. 1994 39(3) 61. 951334 Anal. Chim. Acta 1994 294 185. 95/338 Anal Chim. Acta 1994 287 17. 951344 Talanta 1994 41 195. 951345 Talanta 1994 41 205. 951349 Anal. Lett. 1994 27 411. 95/369 Clin. Chem. (Washington D.C.) 1994 40 602. 95/373 Commun. Soil Sci. Plant Anal. 1994 25 1149. 95/391 Guangpuxue Yu Guangpu Fenxi 1993 13(4) 97 92. 95/396 Guangpuxue Yu Guangpu Fenxi 1993 13(5) 103. 95/398 Guangpuxue Yu Guangpu Fenxi 1994 14(1) 85 98. 95/399 Guangpuxue Yu Guangpu Fenxi 1994,14( 1). 95.95/402 Hebei Daxue Xuebao Ziran Kexueban 1993,13(2) 34.951412 J. AOAC Int. 1993 76 1400. 95/414 J. AOAC Int. 1994 77 441. 951419 J. Dairy Res. 1994 61 83. 95/435 Microchem. J. 1994 49 69. 951437 Microchem. J. 1994 49 275. 951438 Microchem. J. 1994 49 282. 951445 Pochuouedenie 1993 (8) 118. 951450 Quim. Anal. (Barcelona) 1992 11 35. 95/451 Quim. Anal. (Barcelona) 1992 11 55. 95/456 Sendai-shi Eisei Kenkyushoho 1992 (Pub. 1993) 22 226. 95/457 Shanghai Huanjing Kexue 1993 12( l) 31. 228 R Journal of Analytical Atomic Spectrometry August 199.5 Vol. 10