摘要:

Atomic Spectrometry Update— Advances in Atomic Absorption and Fluorescence Spectrometry and Related Techniques STEVE J. HILL*a JOHN B. DAWSONb W. JOHN PRICEc IAN L. SHUTTLERd CLARE M. M. SMITHe AND JULIAN F. TYSONf aDepartment of Environmental Sciences University of Plymouth Plymouth Devon UK PL 4 8AA bDepartment of Instrumentation and Analytical Science UMIST P.O. Box 88 Manchester UK M60 1QD cEllenmoor East Budleigh Budleigh Salterton Devon UK EX9 7DQ dBodenseewerk Perkin-Elmer GmbH Postfach 101761 D-88647 U� berlingen Germany eDepartment of Chemistry University College Cork Ireland fDepartment of Chemistry University of Massachusetts Box 34510 Amherst,MA 01003-4510 USA SUMMARY OF CONTENTS 1. 1.5.3. Chemometric methods 1.5.3.1. Calibration 1.5.3.2.Interference eects 1.5.3.3. Optimization of conditions 1.5.3.4. Validation of methodology Atomic Fluorescence Spectrometry Discharge-excited Atomic Fluorescence Laser-excited Atomic Fluorescence 2. 2.1. 2.2. Laser-based Spectroscopy Laser Ablation and Excitation Laser Enhanced Ionization 3. 3.1. 3.2. References 4. This review follows on from last year’s (J. Anal. At. Spectrom. 1996 11 281R) and describes the developments in atomic absorbance and fluorescence spectrometry since that time. Included in this review are fundamental processes and instrumentation in the areas of atomic absorption and atomic fluorescence spectrometry together with advances in related techniques such as atomic magneto-optical rotation spectrometry and laser-enhanced ionization.The review of Atomic Emission Spectrometry may be found in J. Anal. At. Spectrom. 1997 12 263R. Once again this year a number of changes have been implemented to aid presentation. Following the inclusion of tables to present information in Section 1.2. Electrothermal Atomization last year we have continued the practice in this year’s review to facilitate easier and quicker access to material. The sub-headings have also been modified to reflect the material presented in the review although in general the format is the same as in previous years. This year for the first time references are present in the standard format. Comments as to these changes and possible improvements for future reviews are welcomed by the review coordinator.1. ATOMIC ABSORPTION SPECTROMETRY Atomic Absorption Spectrometry 1.1. Flame Atomization 1.1.1. Studies of flames 1.1.2. Sample introduction 1.1.2.1. Atom trapping techniques 1.1.2.2. Nebulization/vaporization 1.1.2.3. Sample introduction by flow injection 1.1.3. Interference studies 1.1.4. Sample pre-treatment 1.1.5. Chromatographic detection 1.2. Electrothermal Atomization 1.2.1. Atomizer design and surface modification 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 and solid sampling 1.2.2.2. Gas sampling 1.2.2.3. Coupled techniques and preconcentration 1.2.2.4. Electrodeposition 1.2.3. Fundamental processes 1.2.4. Interferences 1.2.4.1.Spectral interferences 1.2.4.2. Chemical modifiers—general 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 instrumentation 1.3.1.2. Determination of individual elements 1.3.2. Mercury by cold vapour generation 1.3.3. Volatile organic compound generation and metal vapour separation 1.4. Spectrometers 1.4.1. Light sources 1.4.2. Continuum source and simultaneous multi-element Instrument Control and Data Processing AAS 1.4.3. Background correction 1.4.4. Detectors 1.5. 1.5.1. Instrument control 1.5.2. Data processing *Review Co-ordinator to whom correspondence should be addressed.1.1. Flame Atomization Over the past year there have been few developments of note in the field of pure FAAS. Nevertheless flames continue to play a major part in the applications of AAS to analytical problems. In last year’s Atomic Spectrometry Updates1 flames were used in 42% of papers devoted to applications; in the same period electrothermal atomization was used in 45% and 327R Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 (327R–379R) cold vapour/hydride generation in 13%. The choice of technique is determined by the combination of analyte element the sample matrix and the resources available. Arising from the simplicity of their operation flame techniques are particularly suitable for pioneering new applications of AAS.Hyphenated techniques is one such area of development. Chromatography combined with atomic spectrometry has been the subject of two reviews where the principles and applicability of several designs of sample introduction interfaces were featured.2,3 This topic is also reviewed in depth in Section 1.1.5. From time to time indirect methods for the determination of analytes by FAAS are developed. These methods will generally be recorded in ASU reviews directed at the analysis of specific matrices. However some methods of general interest or having some novelty will also be presented here. For example a method has been described wherein V was determined indirectly via Cr.4 The sample containing <20 mg ml-1 of V was ‘doped’ with a standard Cr solution (10–14 mg ml-1) and the Cr determined by FAAS.Not surprisingly interferences were serious and recoveries variable! In the same vein B and I were determined via complexation with [Cd(phen)3]2+ followed by extraction with nitrobenzene.5 Recoveries were 92.8–102.4% for B and 97.2–106.7% for I. Two papers have reported indirect methods for the determination of cyanide. In one method Ag was precipitated from AgNO3 in a weak acid solution. The Ag in the precipitate was measured by FAAS and the CN- content derived.703 The response was linear up to 5.2 mg CN- with an RSD of 1.5% and recovery of 97–103%. The other method was based on an ion association compound between either Ag(CN)2- or Cu(CN)2- and benzyldimethylhexadecylammonium followed by extraction with IBMK.6 The detection limits were 0.6 mg l-1 and 1.7 mg l-1 of CN- using Cu and Ag respectively.The recovery when Ag was employed was 99% while that with Cu was 93%. The Ag-based approach suered from less interference and the precision was much better (2.5% for Ag 6.0% for Cu). Other indirect methods reported include the determination of glycyrrhizic acid via precipitation of its Pb compound7 and of a-aminoacids by forming Schi ’s base–CuII chelates and measurement of Cu.8 T riton X-100 was determined indirectly by reaction with cadmium tetrathiocyanate in acetate buer.9 The Cd complex was extracted into xylene and the Cd determined in the organic phase by FAAS. No interference was observed in the analysis of polluted waters and the analytical range was 0.1–1.0 mg ml-1.Two attempts to improve the sensitivity of FAAS by instrumental means may be noted here. The first was an electronic system designed to generate the first time-dierential of the absorbance signal i.e. a derivative signal.10 It was claimed that ‘the strength of the derivative signal is much higher than that of a conventional signal’ but it was not demonstrated that in practice this would lead to better analyses than are already 1.1.1. Studies of Flames achieved by other means. Wavelength modulation of tunable 3 diode laser radiation was used in the measurement of Cr species by FAAS.11 The initial light beam (850 nm) was frequency doubled using an LiIO crystal and modulated at 5.5 kHz with an amplitude approximately equal to the width of the Cr 425.44 nm flame absorption line.The LOD using a water– 45% methanol mixture and high-pressure nebulization was 1 ng ml-1. This LOD is approximately one fifth of that usually achieved by conventional instrumentation using the same spectral line. There is still much to be learnt on the possible uses of lasers in AAS particularly in relation to the benefits of an intense t somewhat unstable light source of narrow line width and further research should be encouraged. Most studies to be reported here relate to fundamental aspects of flames e.g. temperature and free radical concentrations. The 328R Journal of Analytical Atomic Spectrometry August 1997 Vol.12 techniques employed and results obtained may facilitate the understanding of atomization processes in FAAS. However over the past 40 years considerable eort has been devoted to the latter topic; it is therefore highly unlikely that a novel development will emerge to revolutionize present practices in FAAS. The velocity of atoms passing through the observation zone of a flame is a crucial factor in determining sensitivity in FAAS. 2H2 flame is 1.6 m s-1 and The burning velocity of the air–C has been used to estimate the residence time of atoms in the optical absorption path. A recent study12 using laser assisted ionization and laser excited fluorescence of Na in an air–C2H2 flame reported atom velocities of 8–23 m s-1 with a measurement error of <10%.These velocities are an order of magnitude greater than the burning velocity of the flame and merit further consideration when attempting to reconcile theoretical interpretations of the behaviour of atoms in flames with experimental observations. The flame expansion factor i.e. the ratio of the volumes of flame gases before and after burning has been determined experimentally for several designs of burner using Hg vapour as the tracer.13 The method was based on measurement by AAS of the ratio of the Hg atom concentration in the gas flows with and without the flame burning. The observed expansion factors agreed with those calculated by the gas laws. Coherent anti-Stokes Raman spectroscopy (CARS) is a technique which is commonly used for the measurement of flame temperature.The presence of temperature gradients in the measurement volume leads to spatially averaged temperatures when derived by CARS. When the dimensions of the sampled volume are comparable with the laser beam coherence length simulated spatially averaged spectra show that significantly dierent temperatures result when coherent additions of electric field amplitudes of the component spectra are used rather than the most generally used incoherent addition of intensities.14 L asers continue to be one of the most valuable aids for the study of combustion processes in flames and OH one of the most commonly studied radicals. Hydroxyl concentrations and lifetimes in laminar CH4–O2–N2 flames have been measured using picosecond time resolved laser induced fluorescence (LIF).15 The results obtained were comparable with those predicted from quenching cross sections and calculated species concentration.The temperature and OH radical density in a turbulent flame have been determined by a simple LIF measurement technique.16 Radiation from a narrow-band tunable excimer laser (l=308 nm) was focused into a turbulent atmospheric non-premixed flame the OH radical excited and by vibrational energy transfer the population of excited upper electronic energy levels redistributed. By analysis of the broadband emission spectrum the temperature and OH density were determined. The eect of laser radiation intensity on the results of flame diagnostic studies by LIF has been studied.17 It was found that even when the laser was exciting out of the lower state many more OH radicals than were originally present owing to rapid energy transfer the lower state was continuously replenished and the signal could still vary linearly with laser intensity.At the higher laser intensities a significant fraction of the measured OH arose from photodissociation of the water formed from the flame reactions. 4 Factors aecting the validity of measurements of NO in flames by LIF have been examined in two papers. Polarization spectroscopy of NO in a pre-mixed H2–N2O flame was found to be a relatively simple procedure which based on the recorded spectra could be used to estimate temperature and to produce two-dimensional images of NO distribution.18 The interference of O2 on the LIF measurements of NO in an air–CH flame has been studied.19 An optimum excitationdetection scheme which minimized interferences in the broadband (filter) measurement of NO was identified.T unable diode laser absorption spectroscopy has been used to characterize a low pressure pre-mixed CH 3H) and HFC-125 (C2F5H).20 Carbon 3 4–O2–Ar flame inhibited with Halon 1.0 ng ml-1 for Cd and Pb respectively with a 2 min collection 1301 (CF Br) FE 13 (CF monoxide two-line thermometry was used to profile flame temperature. Concentration profiles of various flame species including CH4 H2O CO CF2O CF2H2 CF3H and CF4 were observed. The data were used in validating the detailed mechanisms of chemical flame inhibition.Chemical species in an 2 air–C2H2 flame have been studied by XANES (X-ray absorption near edge spectrometry).21 Metal nitrate solutions were nebulized into the flame. The XANES spectra of free atoms of Cu and Ni were detected but not those of Sr or Rb. The analytical possibilities of NO as the oxidant in acetylene propane and butane based flames have been re-investigated by Pupyshev et al.22 They found that the degree of atomization of B Hf Nb and P in the NO–C2H2 flame was higher than in an N O–C2H2 flame. There were no significant dierences in atomization when NO and N2O were used as oxidants for propane and butane flames. The findings for the NO–C2H2 flame are consistent with those noted earlier by Slavin and co-workers.23 Improvements in the safety and convenience of flame systems continue to be made and a patent has been granted for equipment with facilitates change-over between gas mixtures.24 A special spray–evaporation chamber and burner has been designed for the determination of Pb directly in petrols by using the petrol itself as the fuel for the flame in an AAS system.25 Matrix eects were minimized by decreasing the concentration of O2.Under these conditions Pb was converted into PbO prior to atomization and fluctations in the optical density of the flame were considerably reduced. Optimization of flame conditions is particularly critical when simultaneous multi-element analysis by FAAS is attempted. This problem was investigated by Choi et al.26 with a view to measuring 6 elements simultaneously.Measurements at 4 mm above the burner were found to be generally satisfactory except for Cr and Co. Surfactants as reagents for enhancing sensitivities in FAAS have been examined over many years; however their usefulness has so far proved to be limited. A study of this subject has been published by Pharr.27 1.1.2. Sample introduction While it is generally recognized that introduction of the sample into the flame is still a very inecient process long experience has demonstrated that significant improvements of universal applicability are most elusive. Current research activity in this field is therefore modest compared with that into electrothermal methods of atomization. 1.1.2.1. Atom trapping techniques.Atom retarding tubes increase the AA sensitivity of conventional flame systems by delaying the loss of analyte atoms from the optical absorption path. In addition the tube may be cooled to collect and concentrate (‘trap’) the analyte atoms which are then released by rapid heating to generate a pulse of atomic vapours. The most common configuration employs a quartz tube with an aperture usually a slot mounted above the burner of an FAAS instrument and on its optical axis. This system is known as the slotted tube atom retarder (STAR); when the tube is cooled it becomes a slotted tube atom trap (STAT). In a series of papers Roberts and co-workers reported on their studies of three designs of water cooled quartz tube atom traps viz.a single water cooled tube twin water cooled tubes and a single water cooled tube inside a slotted quartz tube. In a comparison of the three designs using Cd and Pb as test elements they found that in general the water cooled traps were 5–10 times more sensitive than the simple slotted quartz tube atom trap and that the combined system of the slotted tube water cooled atom trap were a further 1.5–4 times more sensitive. In the latter case the sensitivites were 0.1 and Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 period.28 The merits of the slotted tube water cooled atom trap have been noted by other workers.29 A twin tube water cooled atom trap system was optimized for the determination of As Sb Cu and Mn.30 The greatest sensitivity was achieved with high fuel flow the optical axis midway between and close to the trapping tubes and the tubes themselves close to the burner with a coating on the tubes appropriate to the analytical element.The characteristic concentrations for As Sb Cu and Mn were 8.6 8.0 0.93 and 0.38 ng ml-1 respectively with a collection time of 2 min. These sensitivities were 40–80-fold better than those obtained by conventional FAAS. When Hg was determined with the same apparatus and gold used as the tube coating the greatest sensitivity was then obtained with a low fuel gas flow and the tubes high above the burner.31 Sensitivity was 60 ng ml-1 for a 2 min collection period. A water-cooled stainless-steel tube has been used to collect atoms from an air–C2H2 flame for the determination of Pb in alcoholic drinks and water samples.32 A sensitivity of 5.5 ng ml-1 was achieved with a 5 min collection period.As an alternative to expelling the water from the tube of a water cooled atom trap it has been demonstrated by Ertas et al. that atoms were released from the tube surface when a small volume (50 ml ) of organic solvent was aspirated into the flame.33 Isobutyl methyl ketone was generally used and it was suggested that in addition to raising the temperature of the silica surface the reductive properties of the flame played an important role in atom generation. Sensitivity enhancement factors over conventional FAAS for Au Bi In Mn and Te were 180 94 15 7 and 10 respectively with a collection period of 2 min.Bismuth and Sb in pure tin have been determined by slotted quartz tube atom trapping.34 A 1 min trapping period yielded characteristic concentrations of 3.5 and 1.8 mg ml-1 for Bi and Sb respectively. Sensitivity gains were of the order of 250-times over those of conventional FAAS. The gain in sensitivity when Co in vitamin B12 was determined using a STAT was approximately 5-fold with an LOD at 3.8 ng ml-1.35 A novel approach to atom trapping was used for the determination of Sb and Pb in gunpowder smoke.36 The smoke was first trapped in a doubleslotted quartz tube the tube was then wiped clean with cotton soaked in 5% HNO3. The tube and cotton were washed in 1% HNO3 and the resulting solution analysed by FAAS. The gains in sensitivity were 4-fold for Pb and 9-fold for Sb.A simple continuous flow STAR system was applied to the determination of Ag Au Bi Cd Cu Pb Sb and Zn.37 Modest improvements in performance over conventional FAAS of 3- to 5-fold were reported with LODs of 0.5 and 5 ng ml-1 for Cd and Pb respectively. A specially designed hydride atomizer suitable for operation in either the flame-in-tube mode or miniature diusion flame atomization mode has been used to compare the extent of atomization interference of As on Se in the two modes.38 The principal source of interference in both modes was ‘decay’ of free analyte atoms caused by gas phase or surface chemical reaction. The decay occurring in the quartz tube was the more serious. A study of the SNR for a coupled GC–quartz tube-AAS system concluded that the critical factors determining the SNR were the standard deviation of the chromatographic peak and the limiting aperture of the optical train.39 For given dimensions of the quartz tube it was necessary to optimize the gas flow rates and composition.More information on GC–AAS systems will be found in Section 1.1.5. 1.1.2.2. Nebulization/vaporization. Surprisingly little information on developments in nebulization/vaporization techniques have been received for this review. This situation may reflect a general recognition that after many years of study most options for nebulizing samples have been fully evaluated and their roles in practical analysis determined. High pressure and thermospray nebulizers employed in coupled systems may however reward further study.329R The advantages of discrete sampling have been exploited for the determination of Fe in serum.40 By careful adjustment of the sample size according to uptake rate and by measurement of peak height or area as appropriate the results obtained were superior with respect to speed sensitivity and precision to those obtained using sample pretreatment. Silver was selectively extracted into a small volume of chloroform from solutions of high purity Cu and Pb using thioether at pH 1.0–1.5 in the presence of picric acid without using a masking agent.41 Direct nebulization of 50 ml of the chloroform extract into a fuel-lean air–C2H2 flame gave a sensitive signal without the need for background correction for the chloroform present.A high pressure nebulization system was developed by Berndt et al. for the determination of Al Cr Cu Fe K Na Ni Pb and Si and V in oils by FAAS.42 The sample for nebulization was introduced into a carrier stream of methanol–IBMK– white spirit using an HPLC sample injection valve. An HPLC pump fed the sample at a pressure of 100–400 bar to the 20 mm diameter nebulizing orifice. The results were comparable with those obtained by XRF. The thermospray eect was utilized to volatilize supercritical CO2 containing alkylammonium halide complexes of Cr Cu and Se extracted from aqueous solution.43 A heating coil was placed around the sample capillary of a modified nebulizer. The air flow to both the nebulizer and spray chamber was heated to 200 °C.This system gave a 10-fold improvement over conventional FI operation. Recoveries of Cr Cu and Se were 89% 98% and 93% respectively. The operation of a previously described thermospray quartz T-tube interface44 was optimized for the detection of metals bound to metallothioneine and separated by size exclusion or ion exchange HPLC.45 The LODs for Ag Cd Cu and Zn in a 10 mmol 1-1 Tris buer were 1.0 2.6 1.3 and 1.1 ng respectively. 1.1.2.3. Sample introduction by flow injection. Developments and trends in FI–AAS in the period 1972 to early 1995 were surveyed by Fang et al.46 A total of 645 references were classified in a variety of ways such as technique application field and analyte species. Flame AAS accounted for 63% of the publications vapour generation 24% with ETAAS accounting for 11% (including 3% concerned with chemical vapour generation with in-atomizer trapping).Despite its maturity reports of the performance of the basic FI–FAAS combination are still appearing in the literature.47 Several of the leading researchers in the field have made overview presentations at conferences.48–50 As was apparent from the material surveyed for the previous Update,1 and is again evident from this year’s material FI is being used routinely as a sample introduction method that provides an appropriate on-line dilution. Several papers describing work in which FI has been used are clearly applications papers i.e. they report on new or improved methods for the determination of certain analyte elements in particular matrices.These papers contain reports of the following determinations (a) K in human blood serum51 with dilution by split and confluence; (b) Ca and Mg in gelatin52 with dilution by oset merging zones (two unequal sample volumes merge asynchronously at a confluence point giving three points of measurement the two peak heights and the valley in between); (c) Ca and K in plant digests53 with dilution by merging streams and control over the sample volume by computer controlled stream-switching valves; (d) Ca K Mg and Na in wines54 with dilution by zone sampling (a sub-sample of the dispersing sample zone is injected by a second valve into the carrier for transport to the spectrometer); (e) Ca Cu Fe K Mg Na and Zn in waters55 with dilution again by zone sampling; and ( f ) Ca Li and Mg in urine56 with dilution by merging streams.The stoichiometry of small copper indium diselenide crystals was determined57 by coupling a closed dissolution system to an FAA spectrometer by an FI system 330R Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 with merging streams dilution. Flow injection–FAAS was used in a method for the determination of Ca Cu Mg and Zn in whole blood58 in which sample preparation consisted of the addition of a non-ionic surfactant followed by ultrasonic irradiation. It was found that the surfactant with the shortest chain length (Triton X-100) increased the signals by a factor between 2 and 2.5 depending on the element.It was noted that the sensitivity was considerably improved by ultrasonic agitation. A simple flow system was used in the determination of Ca Fe and Mg in soils and sediment RMs59 in which the sample was aspirated at a T-junction in a single line manifold as a slurry prepared by warming in a microwave oven with 1% HF–2% HCl followed by the addition of solid lanthanum nitrate. Recoveries of 100% were obtained with calibration against aqueous standards provided that the sample was ground and heated in the acid mix to at least 80 °C. A single standard calibration procedure was used. The carrier stream (water) was delivered via a computer-controlled pump whose speed was varied such that a linear absorbance versus time plot was obtained for the solution aspirated at the T-junction.The ratio of the slope obtained when a sample was aspirated to that for a standard was used to calculate the concentration in the sample. It is interesting to note that the method is accurate despite the insolubility of calcium and lanthanum fluorides. The basic hardware for FI (peristaltic pump rotary injection valve with external loop 0.8 mm id PTFE tubing) is widely used and is now rarely described in detail in publications. Some innovations in FI hardware have appeared during this review period. Gine� and colleagues53 devised a system based on a network of computer-controlled two-way solenoid valves for the dilution of discrete sample volumes by factors of up to 20 with precisions of less than 1% RSD.Researchers at Perkin- Elmer have described60,61 the performance of a computercontrolled stepper-motor driven dual-channel syringe pump. Such pumps deliver a flow rate which is independent of sample viscosity and can therefore provide reliable metering of small volumes of viscous solutions needed to provide large on-line dilution factors. Dilutions up to 1000-fold at precisions of better than 2% RSD were obtained. The system was used in a procedure for the determination of Cu in a plating bath solution and in a standard copper alloy. An injection device made from tubing clamp valves has been used62 instead of a six-port valve for sample injection and it has been pointed out63 that the use of 0.5 mm id tubing instead of 0.8 mm id tubing results in a 50% reduction in reagent waste for the same analytical performance.The back-pressure in a typical manifold was increased by only 20%. Further developments in a system of variable-volume dilution chambers have been described.64 The manifold achieved dilution in two stages. A discrete volume was injected into a recirculating loop manifold which contained a variable volume (10–20 ml) dilution chamber. After homogenization a 100-ml sub-sample was injected into a single-line manifold containing a single wellstirred variable-volume tank (made from a plastic syringe). Dilution factors between 2 and 130 000 were obtained at precisions up to 3% RSD. A dilution system based on a new design of peristaltic pump head has been widely reported by researchers at Varian.65–68 Dilution factors up to 200 were possible and the system was used in a method for the determination of Ca Fe K Mg and Na in antacid and vitamin preparations.A two-channel version was used to perform standard additions. Van Staden and Hattingh69 devised a dialyser for on-line dilutions and removal of particulates in the determination of copper in industrial euent. Excellent precisions ( less than 0.3% RSD) were obtained for an injection volume of 75 ml with a sample throughput of 70 h-1. The authors also combined preconcentration with dialysis70 by passing the dialysate through a cation-exchange resin (see later). The procedure was used for the accurate determination of Cu in pet-food supplements. A review of 30 references relating to discontinuous flow analysis DFA,71 concluded that the application of the technique with atomic spectrometry was particularly promising with performance rivalling that of FI in terms of versatility and precision.Further information on DFA can be found in the 1995 Update.72 With the additional feature of control over the extent of dilution FI may be used to generate calibration standards. Fang et al.73 developed a system based on the use of stepper-motor driven peristaltic pumps to introduce small volumes into a larger sample loop followed by injection into the carrier stream. Dilution factors up to 144 allowing calibration for Mg over the range 0–20 mg l-1 were obtained. The various dilutions were achieved by introducing discrete volumes by running the pump for periods of between 2 and 10 s at a flow rate of 0.21 ml min-1.The peak concentrations were calculated on the assumption that the dispersion eects gave exponential peaks. Over an 8 h period the calibrations produced were not significantly dierent from those constructed from the data obtained when conventionally prepared standards were introduced. Thus the success of the procedure further validates Tyson’s single well-stirred model for dispersion in FI–FAAS.74 Some further applications of the gradient ratio method described originally by Sperling et al.75 have been described.76 The procedure is the FI analogue of the method of successive dilutions and was applied to the determination of Ca and Mg in the presence of aluminium and of Ca in the presence of silicon.The resulting stable compound interferences were overcome except for the eect of aluminium on Ca for which the additional help of a releasing agent lanthanum was needed to obtain accurate results. A merging zones manifold has been used77 to implement standard addition calibrations based on electronic dilution. In this version of the procedure standard and sample were simultaneously injected from two loops (in parallel) synchronously merged down stream and about 150 measurements were made on the tail of the peak at 0.1 s intervals. These data were then processed by software which accounted for dispersion and any asymmetry eects at the confluence. The procedure was applied to the determination of K (by flame photometry) in water–ethanol and water–glycerol mixtures and to the determination of Cu (by FAAS) in spirits.The pulsations produced by the rotation of the peristaltic pump head have been used78 as the basis of a novel signal processing procedure. In this method the Fourier transform of the signal was convoluted with an empirically derived function to give a signal in the Fourier domain that was directly proportional to the analyte concentration. As this also improved the S/N the calibration could be extended to lower concentrations than would be possible with conventional signal handling and as the response was related to the magnitude of the variation of the signal and not the magnitude of the signal itself the technique permitted extension of the calibration range to higher concentrations.By using a compensation stream introduced at a T-piece prior to the nebulizer a variety of concentrations could be generated by varying the flow rate of a single concentrated standard. As has been apparent for several years w a major aspect of the published work in FI–FAAS is the use of solid phase extraction (SPE) for preconcentration of analyte(s) and separa- 18 tion from potentially interfering matrices. This year sees similar contributions to the literature as for the previous years which may be summarized as follows there is considerable interest in (a) the use of chelating agents immobilized on dimensionally stable backbones (b) the retention of neutral metal complexes on C and (c) the use of ion exchange materials for the speciation of Cr.The topic of SPE has been reviewed by Mentasti,79 but as only 28 references were cited the review must be somewhat selective. Greenway has compared a number of chelating extractants for the determination of Cd Cu Mn Ni and Zn.80 Further details are in a special issue of Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 Spectrochimica Acta Part B devoted to flow analysis (see reference 81). Four materials were compared three of which are commercially available (Chelamine tetraethylenepentamine on reticulated polyacrylamide resin; Chelosolve tetraethylenepentaminehexaacetic acid on reticulated polyacrylamide resin and PROSEP Chelating-1 iminodiacetate on controlled pore glass) while the fourth was synthesised in-house (quinolin-8-ol on controlled pore glass).The polymer-based resins were found to have better capacities than those based on controlled pore glass although lengthier conditioning was required making them less suitable for rapid procedures. The iminodiacetate material had much higher capacity than the quinolin-8-ol material but as with all iminodiacetate materials some retention of calcium was observed. Analytes that have attracted attention during this review period were Au,82 Cd,83–85 Co,83,86 Cr,87–91 Cu,83 Ir,92 Pb,83,93–98 Pd,92,99,100 Pt,92 Sr101 and Zn.83,102,103 Most of the samples analysed were biological or environmental in nature. Two groups have published results for the determination of Pb using extractants consisting of a crown ether immobilized on an inert support.Sperling et al.104,105 used silica-based ‘Pb-02’ and Grudpan and co-workers96 used ‘Sr.Spec’ (bis-tertbutyl-cis-dicyclohexano-18-crown-6 coated on Amberlite XAD-7 polystyrene). Both of these materials are commercially available. The materials demonstrated high selectivity for Pb due to the stereoselective nature of the binding reaction. The sample solutions were acidified with nitric acid and the retained lead was eluted either with EDTA (0.3 mol l-1 pH 10.5)104 or 0.05 mol l-1 ammonium oxalate.106 An enrichment factor of 52 was obtained with the Pb-02 material104 with a sample throughput of 63 h-1 and for the Sr.Spec material,107 a detection limit of 7 mg l-1 was obtained for a sample volume of 12 ml with a sample throughput of 17 h-1 and an enrichment factor of 22.This latter material was used in a procedure for the determination of Pb in soils at concentrations around 20 mg kg-1. Two reference soils were analysed accurately. Chinese workers used 85 a piperidine resin for the preconcentration of Cd from 2–20% HCl media with elution by 4–30% nitric acid. The procedure was applied to the determination of Cd in geological materials at concentrations down to 0.1 mg kg-1. A chelating resin with an a-aminopyridine functionality has been used92 for the preconcentration of Ir Pd and Pt. The analytes could be selectively eluted by appropriate eluents. Chinese workers reported108 on a similar procedure for the determination of Pd and Sr.Enrichment factors of 145 and 78 were obtained for these two elements respectively. A two-column manifold design was used configured so that the columns were connected in parallel in the preconcentration mode and in series in the elution mode. This eectively doubled the enrichment factor obtained compared with that for a single column loaded at the same flow rate per column. The manifold arrangement is described in detail in another paper.99 Elmahadi and Greenway83 described the performance of a procedure using a chelating column containing 5-chloroquinolin-8-ol immobilized on controlled pore glass. The column was loaded at 5 ml min-1 for 2 min with elution with 40 ml of 1 mol l-1 nitric acid injected upstream of the column resulting in enrichment factors of between 50 and 140 for Cd Co Cu Pb and Zn and LODs of about 0.5 mg l-1 were obtained for all analytes except Pb for which the LOD was 4 mg l-1.The sample throughput was 20 h-1. Chinese workers,82 who continue to be active in the development of procedures for the determination of Au described the use of the ‘324H fibrous chelating ion-exchanger’. After roasting dissolution in aqua regia and the removal of silicon by precipitation and filtration the sample solution was passed through the column at 5 ml min-1. The usual gold eluent was used namely an aqueous solution of thiourea and HCl (both 1%) resulting in over 92% recovery from gold ore samples. 331R In a conference presentation by Russian workers,109 concentration eciencies (i.e.enrichment factors per minute of preconcentration) of up to 100 were reported for the determination of Cd Co Cu Ni Pb and Zn in foodstus by a procedure which used retention on ‘DETATA-sorbent’. Further details will it is hoped be available in a forthcoming publication. A number of new reagents for sorbent extraction have been described. In this procedure a reagent is added to form a nonpolar neutral complex with the analyte which is retained by the hydrophobic surface of the solid phase extractant in the column. The derivative is subsequently eluted by a solvent that can be introduced directly into the flame (thus solvents such as ethanol or IBMK may be used as well as aqueousbased eluents).Workers at the University of Antwerp102 investigated the performance of a method in which Zn was determined via the retention of a dialkyldithiophosphonate (RO)2P(S)S-.Out of the eight dierent reagents studied the best performance was obtained with di-sec-butyldithiophosphate (0.5% m/v) at pH 3 with 0.1 M citrate as a masking agent for iron. The complex was eluted with methanol. Accurate analyses of some biological and environmental CRMs were obtained with an LOD of 0.5 mg l-1 and an enhancement factor of 35 for 20 s loading at 8.7 ml min-1. The work has recently been extended to the determination of Cd Cu and Pb (Ma and Adams110) for which diethyldithiophosphate was found to be suitable. Citrate was again used to mask iron and accurate analyses of a loam soil and a lake sediment reference material were reported.For 20 s loading at 8.6 ml min-1 LODs of 0.8 1.4 and 10 mg l-1 for Cd Cu and Pb respectively were obtained. Tao and Fang described95 a procedure for the sorbent extraction of Pb as the ion pair between the iodoplumbate anion and the tetrabutylammonium cation on a C18 microcolumn (10×3 mm). The samples were rice tea peach leaves and human hair each of which was wet ashed with nitric acid followed by treatment with perchloric acid and dissolution of the residue in hydrochloric acid. The complex was eluted with ethanol. An enhancement factor of 36 was obtained for 60 s loading at a combined reagent and sample flow rate of 8.5 ml min-1 of which the sample flow rate was 7.0 ml min-1.An alternate approach to the selective retention of both The LOD was 3 mg l-1 the sample throughput 48 h-1 and the oxidation states of Cr is to use a system which selectively precision for 50 mg l-1 was 3% RSD. A similar approach was retains CrIII. The speciation is then performed in two stages adopted86 for the determination of Co in biological materials. first the CrIII is determined and then after reduction of any The analyte was retained as the ion-pair between tetrabutylam- CrVI to CrIII total Cr is determined from which the CrVI is monium and the Co-complex with nitroso-R salt. An enhance- calculated by dierence. This approach was used by Cesponment factor of 40 a sample throughput of 45 h-1 and an LOD Romero et al.90 and Gaspar et al.91 In the first of these of 3 mg l-1 were obtained.Other Chinese workers described94 procedures,90 a column of a poly(aminophosphonic acid) a procedure based on the retention of the quinolin-8-ol complex chelating resin (PAPhA) was used. The eluent was 0.5 M nitric and Brazilian workers described97 a procedure based on the acid and for a 6.6 ml sample the LOD was 0.2 mg l-1 with a diethyldithiocarbamate (DDC) complex with elution by a sample throughput of 30 h-1. Ascorbic acid was used to reduce discrete volume (380 ml ) of IBMK for which the LOD was CrVI to CrIII and the method was applied to the analysis of 3 mg l-1. Both of these chemistries have been described already tap mineral and river waters. The second method114 is best in the literature (see for example reference 111).Kradtap and described as high-performance FI (HPLC equipment is used) Tyson93 compared methanol and acetonitrile as eluents in the though the authors refer to their procedure as ‘chromatogradetermination of Cd and Pb by retention of the DDC com- phy’. The method is based on the finding that in the presence plexes on C18. A manifold design was used in which the of potassium hydrogenphthalate a C18 column (50×4.6 mm confluence between the sample and reagent streams as well as with 5 mm particles) selectively retains CrIII over all other the column was in the sample loop. This ensured that any cations by a mechanism that the authors were unable to complex adhering to the tubing walls was also ‘eluted’. explain. The CrIII retained from a 5 ml sample loop was eluted Acetonitrile was the preferred eluent.For the determination of with a discrete volume (100 ml ) of methanol–water (80520 v/v). Pd,99 the chloro-complex was retained on a column packed The LOD was 0.7 mg l-1. Ascorbic acid was used to reduce with an activated carbon fibre. The double-column technique CrVI to CrIII and the method was applied to the determination mentioned earlier was used to give an enhancement factor of of total Cr and Cr species in a variety of samples including 145 leading to an LOD of 0.3 mg l-1 with a sample throughput nutritional supplements ground and drinking waters whole of 15–20 h-1. Recoveries of spikes from two rock matrices blood urine cigarette tobacco tuna fish and a medication for were described as ‘satisfactory’. A system consisting of four the treatment of tobacco virosis.columns of ion-exchange resin (AG50W-X3) has been devel- Several Chinese research groups appear fascinated by the oped84 for the determination of Cd. The sample throughput was 80 h-1 with an LOD of 0.7 mg l-1. It was not clear whether the columns were used in a manner which increased the sensitivity (over that obtained with a single column) or 332R Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 which increased the sample throughput. For the preconcentration of Cu after removal of particulates by dialysis van Staden and Hattingh112 used a styrene-based sulfonic acid cation-exchange resin (Dionex OnGuard-H). The optimization of the system was complicated by the fact that the loading flow rate was also that of the acceptor in the dialyser.The column was back-flushed with a mixed nitric–hydrochloric acid eluent (1 M of each). The procedure was applied to the analysis of cat and dog food supplements containing vitamin–mineral conditioners. For the speciation of Cr most workers have used ionexchange or both ion-exchange and chelating materials though Littlejohn et al.87 pursued a procedure based on two chelating resins. CrIII was immobilized on an iminodiacetate resin and they have synthesized a number of nitrogen-based chelating materials as candidates for the retention of CrVI. Sule and Ingle89 developed a procedure in which the CrIII was retained on a chelating resin (Chelex-100) and the CrVI was retained on an anion-exchange resin (AG MP-1).In the latter procedure a 2 ml sample was loaded onto the columns in series and each column was then separately eluted. For CrIII the eluent was 2 M nitric acid while for CrVI a mixture of 2 M ammonium nitrate and 0.5 M ammonia solution was used. The LOD was 2 mg l-1 and the method was applied to the analyses of tap water and humic acid. Chinese workers88 developed a somewhat similar procedure based on the retention of CrIII on anion exchange material and CrVI on activated alumina in the basic form. The eluents were 1 M nitric acid and 1 M ammonium nitrate (pH 8) for CrIII and CrVI respectively. The corresponding LODs for 25 s loading at 9 ml min-1 were 0.7 and 1.5 mg l-1. Mena et al.113 have pointed out that the elution of both CrIII and CrVI from alumina columns is more dicult if the columns have aged.In fact it was not possible to get useful recoveries under FI conditions and the Cr had to be recovered by digestion with nitric acid in a batch procedure. These low recoveries have implications for the use of alumina microcolumns in field sampling procedures and as candidate reference materials. analysis of hair. Workers in Shanghai103 devised a method for the determination of Zn in hair by retention of the chlorocomplexes formed when the hydrochloric acid concentration exceeds 1.0 M on an anion exchange resin. The zinc was eluted with dilute (<0.005 M) HCl. Complete freedom from interferences was claimed and the LOD was 0.1 mg l-1. The Cd and Pb content of hair has been determined by workers in Hefei who used an FI liquid–liquid extraction (LLE) procedure115 based on the extraction of the APDC complexes into IBMK.Enhancement factors of 21 and 23 were obtained for Cd and Pb respectively. This would appear to be the only FI–LLE extraction paper apart from that by Lin et al. who described the performance of a new design of phase separator. 116 Finally workers at Shenyang117 used co-precipitation in the presence of lanthanum hydroxide to determine Cr in hair (and waste water). The precipitate was retained on the walls of a knotted tubular reactor and then eluted with 0.5 M hydrochloric acid. For a 110 s loading period at a sample flow rate of 4 ml min-1 the LOD was 0.8 mg l-1. Other Chinese workers118 used co-precipitation with iron hydroxide to determine Cu and Ni in caustic soda.The precipitate was retained on a filter and dissolved with 2 M HCl. The LODs were 5 and 60 mg l-1. A procedure119 for the determination of Sr by co-precipitation with lead sulfate has been developed (by Chinese researchers but published in English) in which an enhancement factor of 61 was obtained at a sample throughput of up to 30 h-1. The LOD was 0.9 mg l-1 and the procedure was applied to the analysis of tea mineral water and rock. The precipitation of Co as the acetylacetonate has been used120 as the basis of an FI method for the determination of Co in steels. The precipitate from a 500 ml sample was retained on a sintered glass filter of pore size 10–16 mm and after collection for 30 s was dissolved in 80 ml of IBMK.Dittfurth et al.121 determined Mn in a number of biological reference materials at concentrations between 8 and 65 mg kg-1 by a procedure based on the precipitation of hydrated MnIV oxide. Hydrogen peroxide was added to the sample digest to oxidize MnII to MnIV. The sample stream was merged with a 1 M ammonium chloride–ammonia stream at pH 9 and the resulting precipitate was collected on a cylindrical stainless steel filter with pore size 0.5 mm and inner volume 580 ml. The precipitate was dissolved in a stream of 0.1 M ammonium hydrogen oxalate solution or 2 M nitric acid. For 24 ml of sample loaded at 4 ml min-1 an enhancement factor of 55 and an LOD of 1 mg l-1 were obtained. Interestingly the LOD was calculated as the concentration giving a signal equal to 3 s of the response to a standard in the middle of the linear range which would be about 15 mg l-1.The interferences due to cobalt and iron were discussed in some detail and the possible role for masking agents (fluoride thioglycollic acid) evaluated. Precipitation reactions have also been exploited as the basis of some indirect procedures. Esmadi et al.122 devised procedures for the determination of binary mixtures of anions based on the precipitation of silver salts. The mixtures contained chloride and one of the following anions carbonate chromate or oxalate. The precipitates were first dissolved in dilute acid to determine the non-chloride anion and then the chloride was determined by dissolution in 0.5 M ammonia solution.Valcarcel and co-workers123 used a similar procedure for the indirect determination of saccharin. The silver salt was precipitated filtered and dissolved in ammonia solution. Chloride therefore must be absent. The method was applied to the determination of saccharin in the range 5–75 ppm in some mixtures of artificial sweeteners and pharmaceutical preparations. The same authors124 also developed a method for the determination of tannins in wine and tea based on the Folin–Ciocalteu reagent. The tannins were precipitated with copper (as acetate) and the excess copper determined. Tannic acid was determined over the range 1–15 mg l-1. A further application of the polymer-entrapped copper carbonate reagent (see reference 125) has been described126 by Rivas and Martinez Calatayud for the determination of salicylic acid in pharmaceuticals.The basis of the procedure is that the salicylic acid Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 is able to mobilize some of the copper from the solid phase reagent (presumably by complexation) and thus the copper signal is related to the concentration of the salicylate in the injected sample. The method was applicable to the concentration range 4–75 mg l-1 and had a sample throughput of 257 h-1. A method for the determination of Chemical Oxygen Demand based on the determination of the unreacted CrVI has been developed.127 The problem in using FAAS to determine the unreacted CrVI is that the CrIII interferes.To overcome this problem an anion-exchange resin (Absorbex Merck) was used to collect the CrVI remaining after passage of the sample (to which dichromate and sulfuric acid were added) through a microwave oven. After washing the resin the CrVI was eluted with 10 M nitric acid. The method had a high sample throughput for COD determination of 50 h-1 with an LOD of 7 mg l-1 (based on potassium hydrogenphthalate) and a working range of between 25 and 5000 mg l-1. In addition to the FI methods described in this section there have been a considerable number of FI procedures reported for chemical vapour generation. These procedures are discussed in Section 1.3 if a quartz tube atomizer was used or in Section 1.2.2.4. if a graphite furnace atomizer was used.1.1.3. Interference studies In any analysis the detection of interferences is one of the most important and often dicult steps in validating the method. A continuous dilution procedure has been proposed as a means of assessing interference by the determination of Ca in the presence of Si and Al. The technique however is only applicable to systems where dilution modifies the degree of interference. A more general approach to the detection of interference has been suggested by Luterotti and applied to the determination of Cu Mn and Zn in simulated whole liver tissue homogenates. 128 In this approach seven factors were varied simultaneously in a series of eight experiments according to an experimental scheme devised by Youden and Steiner.129 Computer data processing was used to extract the statistical significance of the results.Interferences in the determination of Zn in tissue homogenate were identified by a short programme of experiments and agreed with those found in the course of a comprehensive systematic investigation. A procedure for the prediction of possible interference based on information obtained from a study of 17 binary elemental systems has been proposed.130 The interferences were expressed as empirical equations using a computerized least squares fitting method. The equations can then be used to predict interferences likely to arise in the analysis of real samples containing the interfering elements. Principal component regression and partial least squares calibration methods were applied to the quantitative determination of trace elements in 2 by FAAS.131 The results obtained were comparable with TiO those achieved by XRF.Interferences can often be minimized or even eliminated by optimization of the flame conditions and observation height in the flame. This approach was followed for the determination of Zn in a variety of matrices.132 Residual interferences were eliminated by preparing sample and standard solutions in 0.25 M ethanolamine–1.0 M HCl. Matrix matching is another practical approach to overcoming errors caused by interferences. This method was used in the determination of K and 3O4 created problems.133 A Na in phosphorus where Al and H third technique to overcome interferences is the use of suppressing agents.T riton X-100 and ammonium chloride were used as interference suppressors in the determination of Cr in steel.134 No interference was observed from 16 co-existing ions and the calibration curve was linear up to 40 mg ml-1 of Cr. For the determination of Co and Cu in solutions containing various inorganic species and organic solvents addition of 333R ethanolamine (for Co) or triethanolamine with HNO 2 were eective in eliminating interferences.135 Interferences by cationic and organic compounds on the FAAS determination of Ba Ca and Mg were studied by Mostafa.136 By adding 4-aminosalicylic acid to both samples and standards the interfering eects were completely eliminated. The same worker used a mixture of MgCl and NH ferences of organic and inorganic compounds in the determination of Sr.137 The mechanism of and procedure for overcoming interferences when determining Pt in aluminium have been studied in detail by Tang et al.138 Some interference eects can be turned to analytical advantage.The enhancement eect of Y on Y b in the air–C 3 (for Cu) Bi Cd Co Cr Cu Fe Ge In Mn Ni Pb Sb Sn Te Tl V and Zn was obtained all of which could be eluted in a twostage procedure with 1.0 M nitric acid followed by 0.1 M EDTA. The use of the material in a procedure for the determination of these elements in waters was demonstrated with LODs down to 10 mg l-1 for a 1-l sample eluted with 3 ml of each of 4Cl to overcome the inter- the eluents. The authors discussed the mechanism of retention.A procedure for the determination of Ga in bauxite and waste water151 has been developed. The analyte was retained on a containing 1% ascorbic acid and eluted with 0.5 M HCl. The procedure used a water-cooled silica tube atom trap to increase 2H2 flame the sensitivity. An anion-exchange column was also used152 to has been used to lower the detection limit of Yb 2- to 3-fold down to 20 ng ml-1.139,140 The enhancement eect of Al on Fe was used in the indirect determination of Al.141 A known amount of Fe (3 mg) was added to the sample and the Fe absorbance signal observed. The calibration curve of the Fe signal enhancement against Al was linear in the range of 0.5–1.3 mg ml-1. The method was applied to the analysis of Al bronze with an RSD of 1.4–2.2%.1.1.4. Sample pre-treatment This section contains papers that have either the potential to be implemented in the FI format or some novel feature concerning analyte and matrix separation or analyte preconcentration. Papers which report the determination of a particular analyte in a particular matrix will be found in the Update devoted to the relevant matrix. The majority of the work covered in this section is concerned with some form of solid phase extraction. Several papers142–146 are concerned with the use of activated carbon. Activated carbon coated with DDC was used142 to collect the Pb in a natural water sample. Up to 125 mg of Pb could be collected on 2 g of the coated carbon and the LOD was 7 mg l-1.Silicate rocks were analysed for Mo145 at concentrations down to 1 mg kg-1 by preconcentration of the APDC complex on active carbon. The same chemistry was used144 to determine Ni in urine. After filtering the residue (from a 300-ml sample) was dried and then taken up in 3 ml of 2 M nitric acid. The LOD was around 1 mg l-1. Yu and co-workers devised procedures for the determination of Cu Fe Mn and Zn in aluminium143 and Cu Co Ni and Zn in molybdenum.146 In the first of these the metals were collected as their complexes with 1-(2-pyridyl)-2-naphthol (FeIII was reduced to FeII with ascorbic acid) at pH 10 while aluminium was masked with sulfosalicylic acid. The complexes were dissolved in 8 M hydrochloric acid. In the second method,146 the metals were collected as their complexes with xylenol orange at pH 7 (hexamine) and dissolved in 10 M HCl.The speciation of Cr147 based on the use of pH to control the uptake by dried algae (Anabaena-1058) has been developed. At pH 5 CrIII was taken up whereas at pH 7 CrVI was taken up (the interference from a number of heavy metals was prevented by the addition of citrate). The Cr was released with 6 M HCl. In a method for the determination of Cd and Pb in potable and surface waters,148 the complexes with 4-(4-nitrophenylazo)-1-naphthol were retained as the ion pairs with cetyltrimethylammonium on silica gel. The LODs were 0.5 and 2 mg l-1 respectively. A new chelating resin has been synthesised149 by the immobilization of bis-(N,N-salicylidene)-1,3-propanediamine functional groups onto XAD-4 resin.The performance of the material was evaluated in terms of pH capacity and equilibration times. Cations were adsorbed from neutral solution (the pH varied from 5 to 8.5 depending on the element) and eluted with 1 M HCl. Quantitative recoveries of Cd Cr Co Cu Fe Mn Ni Pb and Zn were reported. The uptake was unaected by up to 50 mg l-1 of Ca or Mg. Anatase (titanium dioxide) was used150 to retain a large number of heavy metal ions with a capacity of about 5 mg g-1. At pH 8 quantitative uptake of 334R Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 Type 717 strongly basic anion-exchange resin from acid solution retain Cd Fe and Pb from solutions containing high concentrations of chloride.In a procedure for the determination of these elements in soils the Cd and Pb were further preconcentrated using a chelating ion-exchanger (Ostsorb DTTA). Chinese workers153 used a chelating ion-exchange fibre (324H) to preconcentrate Au from ore solutions produced by dissolution in aqua regia followed by removal of silicon by the addition of gelatin. The Au was eventually eluted with a solution containing 1% HCl and 1% thiourea. Workers in Slovakia154 also used a chelating ion-exchanger Spheron- Thiol to preconcentrate Au. A batch procedure was used in which the solid phase extractant was dry ashed and the residue taken up in aqua regia. Japanese workers155 have republished a procedure for the determination of Fe by the retention of the ion pair between the Fe complex with tetradecyldimethylbenzylammonium and iodide on microcrystalline naphthalene followed by dissolution of the naphthalene and Fe complex in DMF.Full details of the procedure can be found in reference 156. Chinese workers157 determined Ag in ores by retention on P350 stationary phase followed by elution with 0.1 M HCl–0.6 M sodium thiocyanate. Komarek et al. described158 a procedure for speciating Cu in ketone. which free CuII ions were transported across a tubular nafion cation-exchange membrane. The method was applied to the determination of Cu species in the presence of humic acids. A liquid membrane was used by Huifen et al.159 for the preconcentration of Co Cu and Ni in a procedure for the determination of these elements in alkali metal salts.The membrane was described as containing 5% P204 as the mobile carrier 4% N205 as surfactant and 2 M HCl. The stripping agent was 1.5 M sulfuric acid in kerosine. Surfactants were used160 in conjunction with ultrasound in a procedure for the determination of Zn in milk. Three surfactants were evaluated and sodium tetrapropylenebenzosulfonate was the preferred reagent. Triton-X 114 was used by Garcia Pinto et al.161 for the preconcentration of Cd in a procedure for the analysis of tap- and seawater based on the behaviour of a non-ionic surfactant at a concentration above the critical micellar concentration (CMC) when heated above the so-called ‘cloud point’. This is the temperature at which the surfactant solution separates into two distinct phases; a surfactant-rich phase of very small volume and an aqueous phase containing the surfactant at close to its CMC.On cooling the surfactant-rich phase becomes viscous and may be separated from the aqueous phase. In the procedure used here Cd was retained in the viscous phase which formed on heating a solution containing 0.05% Triton-X 114 at pH 8 (borax buer) as the complex with 1-(2-pyridylazo)-2-naphthol. Methanol containing 0.1 M nitric acid (100 ml ) was added and 100 ml were aspirated into the spectrometer. The LOD was 0.4 mg l-1. A viscous phase also featured in yet another method for the determination of Au.162 The ground ashed ore sample was heated with aqua regia until a thick syrup formed which was then dissolved in 1 M HCl and extracted with isopentyl alcohol.The organic phase was aspirated into the spectrometer. Korean workers163 extracted Au as the complex with 3-thiophenaldehyde- 4-phenyl-3-thiosemicarbazone into diisobutyl Pakistani workers164 extracted Cd Cu Ni and Zn as the complexes with 6-methyl-2-pyridinecarboxaldehyde-4-phenylsemicarbazone into chloroform. The organic solvent was evaporated and the residue taken up in nitric acid. The same solvent was used by Czech workers165 to extract the complex of B into 2-ethylhexane-1,3-diol. After back extraction into sodium hydroxide solution the B was determined by FAAS using an N O–C 2 2H2 flame. It is of interest to note a new procedure for the automation of liquid–liquid extraction in a sequential injection system.166 In this procedure the interior wall of a PTFE tube was coated with an organic solvent (benzene was used) then an aqueous solution of the analyte (bromothymol blue) was passed through the tube and finally the retained analyte was back-extracted into a small volume of sodium hydroxide solution.The authors (Ruzicka and Christian and co-workers) speculated on the possible applications with AAS. The development of a new solid phase reactor for FI should also be noted. Troccoli and coworkers167 determined CrVI by the well-known reaction with 1,5- diphenylcarbazide using an FI system in which the reagent was added by dissolution of the solid from a reactor packed with a mixture of the reagent and 25% m/m acid-washed silica.A column containing a total of 260 mg of material was stable for up to 300 injections (of 300 ml). 1.1.5. Chromatographic detection As has been the case for some years now there is relatively little activity concerning the use of FAAS for chromatographic detection. In a review (51 references) of the speciation of Se168 in biological and environmental samples by HPLC with selenium-specific detection FAAS was listed as one of the techniques used along with ETAAS and ICP-MS. The range of applications appeared limited to the determination of SeIV SeVI and the trimethylselenonium cation (an important component in urine). Quevauviller169 has presented an overview of organizational aspects of inter-laboratory studies and discussed possible errors in speciation analyses.A variety of ion-pairing reagents have been evaluated170 for the determination of SeIV and SeVI in selenium supplements for animals by reversedphase HPLC. Of those studied the best performance was obtained with 5 mM tetrabutylammonium dihydrogenphosphate in a mobile phase of 50+50 methanol–water. For a 100-ml injection volume the LODs were 31 and 51 ng for SeIV and SeVI respectively. Ion-pairing with tetraalkylammonium salts has also been applied for the speciation of Cr171 using a polymer-based reversed-phase column (Hamilton PRP1). Satisfactory results were obtained with (a) tetrabutylammonium phosphate (1 mM) in methanol–water (60540) and (b) tetraethylammonium nitrate (2 mM) in water at pH 3–4.The LODs were 24 and 40 mg l-1 for CrIII and CrVI respectively. The procedures were also used with ICP-MS detection and were applied to the analysis of aqueous extracts of coal fly ash and ash from a wood treatment company. Reversedphase chromatography has also been used172 for the separation of vitamin B12 analogues. Gradient elution was used to separate 4 compounds in the water extracts of pharmaceutical preparations. Over a period of 5 min the mobile phase composition was changed from 25575 methanol–0.085 phosphoric acid (buered at pH 5.2 with triethanolamine) to a 60540 mixture. The eluent (1.5 ml min-1) was connected directly to the nebulizer of an FAA spectrometer operating with air compensation and for a 100 ml injection volume the LODs were 4–5 mg l-1 for the various cobalamins.Indirect detection was used by the same research group173 following the separation of alkali metals (and ammonium) on a strong cation exchange stationary phase with 60 mg l-1 copper(II ) nitrate as the mobile phase. Again direct coupling to the nebulizer with aircompensation was used. Some further applications of the silica T-tube interface–atomizer developed by Marshall174 have been Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 described for the further characterization of mammalian and molluscan metallothioneins based on the use of size exclusion and ion-exchange HPLC. The elements determined were Cd Cu and Zn and the authors studied silver saturation as a means of determining the total amount of the polypeptides.Size exclusion chromatography has also been used175 in a study of the Zn-proteins in various milk samples from which fat and casein micelles had been removed. 1.2. Electrothermal Atomization As in previous years there have been few major developments to note during the time period covered by this Update. In general large numbers of ETAAS applications are published in the scientific literature though many of these simply repeat work that has been previously published. This Update concentrates on activities directly related to ETAAS be it atomizer design or novel chemical modifiers. Applications are only covered when these show some new or interesting aspect of work that relates to ETAAS research. Workers continue to compare ETAAS with other techniques.Singer176 asked ‘Can ICP AES replace GFAAS?’ and discussed the LODs for As Cd Co Cr Ni Pb Sb and Se in environmental samples obtained with simultaneous-sequential mode operation of an ICP spectrometer. Although the LODs were at or near the 10s LOQs generally required for assessing environmental samples it has to be remembered that this does not necessarily mean that one can perform reliable measurements at these levels. The LODs for ETAAS instruments are often well below the 10s LOQs indicating that measurements at or near these limits can easily be performed by ETAAS. Schlemmer177 discussed the LODs precision and long-term stability of ETAAS and showed that when utilizing STPF conditions improved chemical modifiers and graphite tube atomizer and platform designs ETAAS is a powerful tool for trace and ultra-trace determination of elements in samples with high matrix content.As many workers continue to acknowledge a major advantage of ETAAS is the comparatively low requirement for sample preparation; consequently it is possible to obtain LODs in samples that do not dier significantly from the instrumental LODs. In addition the stability and freedom from drift is reliable over long periods even in the presence of strong acids or corrosive matrices. Ivaldi et al.178 discussed ‘Can axial viewing of ICP emission replace GFAA?’ and presented a balanced view of the advantages and disadvantages of the two techniques. The choice of analytical tool really depends upon the application.While ICP oers a faster and cheaper means to handle routine high volume analysis compared with ETAAS and has fast multielement capability with a wide dynamic range ETAAS has the advantage that it can cope with a wide variety of sample types and sample introduction is rarely a problem. In addition the LODs are generally an order of magnitude better than axial ICP. These workers concluded that both ETAAS and ICP have their place in the arsenal of tools available to the analytical scientist. A few reviews were published during this period. Farah and Sneddon179 reviewed the developments and applications of multi-element ETAAS from an instrumental development perspective (52 references) and the biennial Analytical Chemistry fundamental review180 was published.The previous Update1 was also published. 1.2.1. Atomizer design and surface modification 1.2.1.1. Graphite atomizers. During the period of this Update there appears to have been little work published on graphite atomizers compared with the previous one.1 Another group from Western Carolina University USA,181–183 appears to be looking in detail at the graphite surface. Butcher et al.,181 Vandervoort et al.182 and McLain et al.183 investigated the 335R graphite substrates used in electrothermal atomizers with scanning tunneling microscopy (STM) which allows atomic level imaging of the graphite surface. For pyrolytic graphite platforms the STM scans indicated the presence of individual crystallites with a maximum grain size of 5×5 nm2.These grains have a high degree of disorientation leading to a high density of exposed edge sites which appear to be oxidised and may serve as active areas for chemical reactions during ETAAS determinations. For pyrolytic graphite coated electrographite approximately 25% of the STM scans show considerably less height variation than those of the pyrolytic graphite platforms. However the atomic structures of the pyrolytic graphite coated surface tend to be more amorphous. The surfaces of the ETAAS graphite substrates do not even approximate the uniform structure of highly orientated pyrolytic graphite with grain sizes of the order of 100×100 nm2. In subsequent work Vandervoort et al.182 obtained STM images from pristine pyrolytic graphite coated and uncoated polycrystalline graphite tubes and from pure pyrolytic graphite platforms.For comparison images of highly orientated pyrolytic graphite not used in ETAAS work were also obtained. Polycrystalline tubes were characterized by disordered surfaces with extensive oxidation. Pyrolytic graphite coated tubes and pure pyrolytic graphite platforms were characterized by scaled structures island columns or smoothly varying contours. Scaled structures and island columns presented the greatest abundance of exposed carbon edge sites and seemed to be the probable areas for analyte reactivity and intercalation. In areas displaying smoothly varying contours atomic imaging revealed basal plane surface layers with moderate curvature but regular spacing between the carbon atoms within the layers.The results imply that the weak ordering between graphite layers along the carbon axis does not preclude good atomic ordering within the layers and this factor may be sucient for discouraging analyte–substrate interactions. This work is in contrast to that from Professor Ortner’s group discussed in the last Update,1 and appears that it is intended to provide atomic scale information about ETAAS atomization mechanisms whereas Ortner and co-workers are examining the changes that take place during the lifetime of a graphite atomizer. However both are valid approaches and likely to enhance considerably our knowledge of the graphite surface within an electrothermal atomizer.Further work by McLain et al.183 considered the eect of heat on these surfaces. As the surfaces of polycrystalline and pyrolytic graphite were considered too rough to monitor subtle changes in surface topography following heat treatment pristine highly orientated pyrolytic graphite (HOPG) was used since single layer atomic scale defects are readily observed by STM on HOPG. When HOPG was heated in air atomic defects in the basal plane grew into etch pits. HOPG heated in the argon atmosphere of a graphite atomizer also developed etch pits from the oxygen present as an impurity in the argon purge gas. With argon gas flow stopped etch pits were formed at lower temperatures and at shorter heating times. The activation energy for the oxidation of graphite at high temperatures was found to be less than a value reported in the literature for low temperature reaction.At high temperatures in addition to pit growth at defects carbon atoms were abstracted from the basal plane and the rate of abstraction was measured and used to determine the partial pressure of oxygen present during ETAAS measurements. Hoenig and Dheere184 evaluated the use of end-capped transversely heated graphite electrothermal atomizer tubes (THGA) with an aperture of 3.5 mm compared with standard THGA graphite tubes (6 mm) with respect to characteristic mass LODs and RSDs for the determination of Cd Cr Pb and Mo in environmental samples. With end-capped THGA graphite tubes the characteristic masses for Cd Cr and Pb were enhanced by approximately 1.4-fold the absolute LODs 336R Journal of Analytical Atomic Spectrometry August 1997 Vol.12 being 0.9 0.8 and 4.8 pg respectively compared with 1.8 2.9 and 7.5 pg respectively for the standard THGA graphite tube and the RSDs were generally lower. Interferences were negligible for both types of tube. These results presented an independent confirmation of the work discussed last year from Hadgu and Frech (see ref. 1). In the previous Update,1 the conference reports on temporal and spatial gas-phase temperature measurements with coherent anti-stokes Raman scattering (CARS) within a transversely heated graphite electrothermal atomizer (THGA) were discussed. This work has now been published by Sperling et al.185,186 This has to be one of the definitive publications concerning the temporal and spatial temperature distributions within a graphite atomizer.In detail the analytical characteristics are assessed and interpreted with respect to the temperature distribution. The results from both longitudinally heated graphite atomizers (HGA) and THGAs are compared. The eect of the internal gas flow on the size of the heated atmosphere was studied with steady-state temperature measurements. Temporally and spatially resolved measurements made it possible to study the temperature field within the atomizer volume in all three dimensions during the rapid heating of the furnace to final temperatures in the region of 2173–2673 K. The role of the integrated platform in the THGA on the temperature was investigated by temperature measurements of the gas phase in the presence and absence of the platform.The platform was identified as a major source of temperature gradients inside the tube volume internal gas which may be as high as 1000 K in the radial direction during rapid heating. These gradients were most pronounced for heating cycles starting at room temperature and gradually decreased with increasing starting temperatures. Shortly after the tube wall reaches its final temperature the gas phase temperature equilibrates and approaches the wall temperature. Because of the unavoidable contact with the cold environment at the open ends of the tube minor temperature gradients are observed in the gas phase in the longitudinal direction which can be further reduced by restricting the openings with end caps.The results obtained for the THGA were compared with those obtained earlier for HGAs including some analytical applications of these two atomizer types. Both the temperature characteristics and the resulting analytical characteristics substantiate the superiority of the THGA in comparison with the HGA as an atomizer for ETAAS. The work of Smith and Harnly concerning the application of a double atomizer for ETAAS measurements was discussed last year1 and has now been published.187 Optimized cup and integrated contact cuvette (ICC) temperatures were determined for Cd Cr Cu Fe and Pb. Ecient analyte transportation (up to 91%) from the cup to the ICC was observed for an ICC–cup separation of up to 3 mm.Analyte transport was aided by an apparent convective flow of Ar into the hole in the bottom of the ICC. Despite this flow the losses of Cd and Pb from the atomizer appeared to be limited by diusion. The relative sensitivity for the modified double atomizer was approximately 50% worse than that for the same ICC with atomization from a platform. Chinese workers188 studied the atomization mechanisms for over 20 elements from a probe atomizer. 1.2.1.2. Metal atomizers and metallic coatings. The work produced during the period covered in this Update is summarized in Table 1 and much of it follows that of the previous years with few new developments. A review of metallic atomizers (115 references) was published in Portuguese189 and covered developments from 1970 to 1994.Volynskii190 questioned the IUPAC recommended nomenclature191 ‘tantalum (or other element) carbide-coated tube’ for describing the treatment of the surface of a graphite-based atomizer as being valid only in Table 1 Metal atomizers and surface modifications Atomizer material/metallic coating Comparison of electrographite pyrolytic graphite coated electrographite W-coated electrographite and Zr-coated electrographite Comparison of Ir Nb Pd–Ir Taand W-coated platforms and graphite tubes Comparison of Ir- Nb- Pd–Ir- Ta- and W-coated platforms and graphite tubes Comparison of Ir- W- Zr-coated graphite tubes Comparison of Mo- Pd- and W-coated platforms Comparison of Nb- W- or Zr-coated graphite tubes Hf-coated tube Ir-coated graphite tube Ir–W- and Ir–Zr-coated platforms Ir–Zr-coated platform La-coated tube La-coated tube La-coated tube Mo-coated graphite tube Matrix Element Acids Cr Sediment SRM Ge Low-alloy steels Sn Cd Water and seawater — Cd Steel standards Sn River water Cr Sea-water Hg — Various (6) — Various (4) — Au Sn Cu alloys Sn Fish and mussels Sn Waste water Br Cl F Journal of Analytical Atomic Spectrometry August 1997 Vol.12 Reference Sample treatment/comments 203 Optimum sensitivity and precision with pyrolytic coated graphite tube [%RSD (n=8) 1.8%] longer lifetimes obtained with metal-coated tubes.Eect of increasing concentrations of HCl HClO4 H2SO4 and HNO3 studied. Investigation of long-term stable 214 215 in-atomizer hydride trapping reagents Zr coatings preferred. Good signal stability for over 400 firings. Zr coatings provided LOD of 18 pg for 1 ml sample volume. See also ref. 211 215. Sn trapped as hydride on coatings. Zr and W coated tubes preferred. LOD 25 pg for 1 ml sample volume. See also refs. 211 214. 207 Cd trapped on coatings. Co Ga and Si used as catalysts for Cd species generation. Ir-coating and Ga catalyst gave best results. LOD 4 ng l-1 for 500 ml sample volume. RSD 1.3% for 0.5 mg l-1 Cd in sea-water. 202 Comparison made in the presence or absence of Mo Pd and W chemical modifiers. Cd–Mo and Cd–Pd interactions noted with modifier and Mo coating but no interaction with Pd coating.211 Coatings used for in-atomizer hydride trapping. Single coating lasted 400 firings. Best sensitivities obtained for W- and Zr-carbide coatings. Samples digested in aqua regia. LOD 25 ng l-1 Sn. No advantage gained by coating platforms rather than tubes. See also refs. 214 215. 200 CrIII and CrIV determined separately after co-precipitation with Ga phosphate. LOD (2s) 0.1 ng l-1 for 500 ml initial volume. 216 212 Hg vapour trapped on coating. LOD 25 ng l-1 for 1.5 ml sample volume. RSD 1.0% for 2.5 mg l-1. See also ref. 207. In-atomizer trapping of hydrides on coatings. Optimum conditions for hydride generation studied. Ir–Zr preferred combination.Best characteristic masses close to those obtained with direct injection. 213 In-atomizer trapping of hydrides on metal coating. Eect of L-cysteine on hydride generation studied. Best characteristic masses for As Bi DMA MMA Sb Sn were 35 110 32 31 83 104 pg respectively. See also ref. 212. 208 Comparison of La-coated and uncoated graphite tubes. Electron microscopy AAS and ESCA appeared to show that the formation of La2O3 catalyses the oxidation of carbon changing the atomization mechanism of Sn but not that of Au. See Section 1.2.3. Fundamental processes. 201 Eect of tartaric acid chemical modifier studied. La shown to change atomization mechanism of Sn. RSD 4.4% recovery 101%. See also ref.208. 199 Samples digested with HNO3–K2Cr2O7–diammonium hydrogenphosphate. Interference due to S studied. LOD 0.17 mg g-1 in samples 10 ml sample volume. 217 Al-halides molecular absorption spectrometry used to determine Br Cl and F. Fe Pb and Pt HCLs used as light sources. LODs 74 86 1.29 mg l-1 for Br Cl and F respectively. 10 ml sample volume. 337R Table 1 (Continued) Atomizer material/metallic coating Mo-coated graphite tube Mo-sputtered platform Mo atomizer Mo atomizer Mo atomizer Ni-tube atomizer Ti–W-coated graphite tube W-coated tube W atomizer WETA 90 W coil W coil W coil W coil W coil 338R Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 Element Ga In Tl Various (4) Au Au Pd Hg B As — Pb Cd Cu Pb Various (5) Various (5) Cr Matrix Geological — Biological River and seawater Biological Environmental and biological Metallic standards — Various Paint and blood — — Blood paint soil and urine River water Reference Sample treatment/comments 205 Mo coating helped avoid loss of analytes as volatile oxides prior to atomization.RSDs 1.4 4.8 and 3.0% for Ga In Tl respectively. 204 Homogeneous layer of Mo deposited on graphite platforms by cathodic sputtering. Performance of permanent Mo modifier compared with modification using Mo solution modification for Cd Cu Ge and Pb. Good signal stability over 200 firings. Elevated thermal stability observed for Cd and Pb.Enhanced sensitivity observed for Ge. 218 2 Samples decomposed with 14 mol l-1 HNO3–30% H2O2 evaporated to dryness and dissolved in HCl. Cu (10 mg) as copper nitrate used as modifier Ar–H purge gas LOD 38 ng l-1 for 1 ml sample volume RSD (n=10) 2.7% for 100 pg Au. 219 Au preconcentrated on W wire before being atomized in Mo atomizer. LOD 1.6 ng l-1 RSD (n=10) 3.2% for 250 mg l-1 Au. No interferences observed from 25 mg l-1 Al Ca Cu Fe K Mg Na Pb or Zn. See also ref. 218. 221 222 Sample pre-digestion with HNO3–H2O2 chemical modifier thiourea (5 g l-1) 2 LOD 15 ng l-1 1 ml purge gas Ar–H sample volume. See also refs. 219 220. Ni atomizer replaced solid sampling device in Grun SM20 ZAAS aqueous and solid samples analysed.Solids weighed in platinum crucibles and placed directly into nickel atomizer calibration prepared against solid standards. LOD 0.1 ng; sample mass of up to 300 mg used. 223 Chemical modifiers of Sr and Ni nitrates and ascorbic acid atomization at 2900 °C. LOD 0.053 mg l-1 RSD 2.6% for 2 mg l-1 B 20 ml sample volume. 224 W coating provided thermal stabilization for As species. Improved performance of PdCl modifier observed with W coating especially for organic species. 225 Study of W atomizer surfaces. Thermal conditioning of atomizer stabilizes sensitivity. W atoms only observed at temperatures 3000 °C significant corrosion observed only in the presence of H2SO4 and H2O2. 195 Prototype instrument aim was to develop a low cost ($6000) portable system.Miniature spectrometer–CCD detector mounted on a PC card used. LOD 20 pg in sample volume of 20 ml limiting noise was detector noise. 194 193 Construction and application of system described in ref. 195. Use of multi-element HCLs for simultaneous multi-element analysis using portable system described in reference 195. LODs in low mg l-1 range. 226 W coil atomizer described in ref. 195 used for AAS and for electrothermal sample introduction into ICP-AES with LODs at sub-ppm level for simultaneous determination of As Cd Cu and Pb. 227 Laboratory made W coil atomizer used. LOD 8 pg in 20 ml sample volume RSD 3% for 0.4 ng Cr. Table 1 (Continued) Atomizer material/metallic coating W coil W coil W coil W coil Zr-coated tube Zr-coated tube Zr- Pd- and Zr–Pd-coated tubes certain circumstances.Options were oered for atomizers depending on the method of modification. When vapour deposition is used the term ‘metal carbide-coated graphite tubes’ was proposed. For the more common practice of treating the graphite tube with solutions of compounds of high-melting elements ‘graphite tubes modified with tantalum (tungsten etc.)’ was proposed. For tubes coated with lanthanum a special term ‘graphite tubes modified with lanthanum compounds’ was suggested. This was proposed because lanthanum carbides unlike other high-melting carbides are easily hydrolysed and are therefore transformed into the respective hydroxide during contact with analyte solution.However the authors of this Update (I.L.S. and C.M.M.S.) consider that the IUPAC recommendations do not exclude such terminology. The recommendations suggest that a modified atomizer be specified by indicating the material and form. This seems suciently wide enough to include the proposals of Volynskii.190 The most prominent development in the area of metallic atomizers has been the increase in reports describing the development of low-cost portable systems utilizing a tungstencoil atomizer mainly for the determination of Pb in blood. Three research groups have reported work on these systems. 192–197 Work completed on the instrument developed by the group led by Parsons and Slavin and discussed in last Matrix Element Whole blood Pb Whole blood Pb Whole blood Pb Various (6) Drinking water wine and soft drinks — As Se Co-based alloys B — Ge Journal of Analytical Atomic Spectrometry August 1997 Vol.12 Reference Sample treatment/comments 196 Prototype instrument portable Pb in blood analyser. Parameters optimized filament position gas flow temperature. Peak height measurements used 42 human samples analysed and good agreement between graphite furnace AAS and W coil atomizer LOD 20 mg l-1 10 ml sample volume. 197 Prototype instrument. Low cost portable system incorporating W coil atomizer and solid-state detector. LOD 30 mg l-1 10 ml sample volume RSD 9% for 100 mg l-1 Pb. 228 Calibration based on method of standard additions using blood SRM.LOD 19 mg l-1 for 10 ml sample volume. Results agreed well with those obtained using graphite atomizer. 229 Unenclosed W coil atomizer operated in Ar–H2 (90+10) mixed gas. Used for waters after preconcentration using Chelex-100 LODs in waters (after preconcentration) ranged from 0.01 mg l-1 (Cd) to 0.13 mg l-1 (Co) LODs in wines ranged from 0.01 mg l-1 (Cd) to 382.7 mg l-1 (Mn) and in soft drinks LODs were 1 mg l-1. Sample volumes of 10 ml used. Recoveries ranged from 82 to 118%. 209 Simultaneous determination of As and Se by hydride generation and trapping on Zr coating. Pretreated tubes found to have lifetimes of up to 80 firings. LODs 13 and 17 ng l-1 for As and Se respectively 7.1 ml sample volume.230 Mixed chemical modifiers of Ni–Zr used interference eects due to major component cobalt and eight minor components were studied. RSD 3.3% for 0.048% B. 210 Ge deposited as solution or trapped as GeH4 (Pd-coating only). Ge–Pd interaction stronger than Ge–Zr interaction Zr–Pd coating produced double peaks. Hydride trapping led to stronger interaction than solution deposition. years review,198 has now been published.196 What appears to be a commercially available instrument fitted with a solidstate detector has been described,197 although there appears to be no means of background correction which is essential for this application. A third instrument developed by Jones and co-workers192–195 at Wake Forest University also makes uses of a tungsten-coil atomizer and incorporates a miniature spectrometer and CCD detector mounted on a PC card.This system has been used to determine a range of elements in several sample types194 in addition to Pb in blood. By using multi-element HCLs the system has also been used for simultaneous multi-element analysis.192,193 There continues to be considerable interest in the use of permanent metallic coatings as chemical modifiers199–208 and as in-atomizer trapping agents for hydrides209–215 and Hg vapour.216 A number of studies comparing the performance of long-term chemical modifiers and trapping agents have appeared in the literature. Alvarez et al.202 compared the eects of dierent atomizer surfaces including metal-coated atomizers on the atomization of Cd.Diering degrees of interaction between Cd and the various surfaces were observed. With platforms coated with either molybdenum or palladium strong Cd–Mo interaction was observed but no Cd–Pd interaction. These studies appear to have been carried out in the absence 339R of a matrix which would presumably aect the analyte–coating interaction. Pyrzynska203 made a similar comparison of atomizer surfaces for the atomization of Cr. It was found that while coating a graphite atomizer with tungsten or zirconium signifi- cantly increased the lifetime of the atomizer it also resulted in decreased sensitivity and precision. It was suggested that the metallic coatings be used therefore when high sensitivity was not critical.For in-atomizer hydride trapping iridium and zirconium are the coatings most often used usually in combination. Continuing their work comparing atomizer coatings for hydride trapping Haug and Liao211 considered the trapping of SnH4. Coatings of the carbide forming elements zirconium tungsten and niobium were compared. Eective trapping was possible on tungsten- and zirconium-coated atomizers at trapping temperatures of 500–600 °C. Good signal stability was observed for over 400 trapping–atomization cycles. It was noted that there was no advantage to be obtained from coating the platform instead of the atomizer tube. Bermejo-Barrera et al. compared iridium tungsten and zirconium coatings for the in-atomizer trapping and preconcentration of cadmium vapour.207 Cobalt gallium and silicon were used as catalysts for cadmium vapour generation.A combination of an iridium coated atomizer and gallium catalyst provided the best detection limit (4 ng l-1 for a 500 ml sample volume). 1.2.2. Sample introduction 1.2.2.1. Slurry and solid sampling. Slurry sampling for electrofactors which contribute to the uncertainty in the direct analysis A general trend concerning investigations carried out to assess how eective permanent metallic coatings are is the use of aqueous standard solutions or water RMs to study analyte stabilization. As significant dierences in analyte behaviour can be observed in the presence of a sample matrix it is to be hoped that subsequent studies consider the eects in the presence of such matrices.thermal atomization again proved to be a very active area with a large number of reports. Most were applications based and are summarized in Table 2. However a number of reports with a more fundamental trend have appeared during the period covered by this Update. Jackson231 presented the results of a study into slurry atomization processes using a two-step atomizer. This atomizer by separating the volatilization and atomization processes can be used to measure the temperatures at which analyte elements leave the slurry particles. The conclusions from this work should be interesting. Many of the fundamental-based reports have considered the of solids and slurries. Belarra and co-workers232 reported an improvement in RSD from originally 10% to 5% for the determination of Cd in solid PVC samples simply by using the median rather than the mean value in a set of data.Use of the median value has the same eect on precision as has removal of ‘suspect’ values using Dixon’s Q-test i.e. it reduces the influence of outliers. Improved precision in the analysis of slurry samples was reported by Schaer and Krivan233 through the use of large volume (15 ml compared with the usual 2 ml ) autosampler cups. These larger volume cups allowed larger sample masses to be used which resulted in a more representative and homogeneous slurry. In a conference presentation Sonntag and Rossbach234 described a method of improving detectability in solid sampling when using a commercial graphite boat sampler.By pressing solid samples into pellets the sample mass used could be increased from 1–2 mg to 15–20 mg. An instrumental development for direct solids analysis was described by Freedman et al.271 Solid samples were introduced to the graphite atomizer using a micro-piston. Kurfurst et al.235,236 assessed the systematic and random eects that give rise to uncertainty in analytical measurements 340R Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 for samples analysed both as slurries and after decomposition. The various components that contribute to uncertainty (defined as those which occur in the calibration unknown sample measurement and quality control stages of analysis) contribute to varying degrees depending on the form of sample introduction and whether calibration is made against aqueous or solid samples.The use of solid RMs for calibration leads to greater uncertainty in the calibration response function but lower uncertainty in the quality control measurement as the quality control component is ‘built-in’. Despite these variations the overall uncertainty ranges for Cd and Pb in digested and slurried samples were of the same order. A study of error sources in solid sampling was also carried out by Luecker et al.237 for biological samples with high water content. The study concentrated on errors due to evaporation of water from samples. This concentration eect led to an increase in apparent analyte concentration of 10% for cadmium solutions (0.1 and 2.0 ng mg-1 in 0.1 mol l-1 HNO3).Use of an automated solid-sampling system incorporating a microbalance reduced operator time by 80%. The method was recommended for use as a screening tool for residue control in fresh animal tissues. Berglund and Baxter238 used the generalized standard additions method of calibration for the analysis of solids. The useful calibration range was extended using linearization algorithms. These workers considered the approach to be generally successful and suitable for use in multi-element ETAAS. It was stated however that although good results were obtained for Ag and Cd in RMs the applicability of the technique to the determination of other elements in real samples remains to be demonstrated.Arruda et al.239 reviewed (89 references) the analysis of foods by solid and slurry sampling methods. Slurry sampling was found to be more popular than direct solid analysis for foods and animal tissues and vegetables were more commonly analysed than fish and fruit. A comparison of solid versus slurry sampling for the analysis of food samples was also reported by Lucker.240 A very large number of samples (20 000) were analysed for five elements and some interesting conclusions were drawn. For food monitoring applications direct solids analysis was found to be preferable to slurry analysis. The improved sample homogeneity obtained as a result of slurry preparation oered relatively slight improvements in analytical error when compared with the increased sample preparation costs and contamination possibilities.The use of slurry sampling for the analysis of soils was considered by de Loos- Vollebregt.241 Simultaneous multi-element analysis was used for groups of elements and 20 samples were analysed. The results compared well with those obtained after sample decomposition. Krivan242 used both solid and slurry sampling for the analysis of high-purity materials. The benefits of direct sampling methods for these materials particularly in terms of reduced contamination was demonstrated. The eect of chemical modifiers in slurry ETAAS was studied by Miller-Ihli.243 A problem was identified with the concentration of modifier used. The concentrations that are suitable for solution sampling were shown to be too low to prevent pre-atomization losses when higher analyte concentrations are present in slurries.A high modifier concentration was also shown to be problematic when it leads to stabilization of the matrix such that it co-volatilizes with the analyte. In two presentations244,245 the same worker compared ETAAS and ETV–ICP-MS for solid sampling using automatic slurry preparation with an ultrasonic probe. Several advantages and disadvantages of both techniques were illustrated. A method of solid sampling by sputtering sample onto a graphite platform which is subsequently placed in a graphite tube atomizer was described by Wendl and co-workers.246 The Table 2 Slurry and solid sampling Type of sampling Slurry Slurry Slurry activated charcoal Slurry sample preconcentrated on activated charcoal Slurry prepared from ground sample Slurry prepared from ground sample Slurry prepared from ground sample particle size 0.08 mm Slurry prepared from ground samples Slurry prepared from ground samples particle size 20 mm Slurry prepared from powdered sample particle size 10 mm Slurry from freeze dried and ground sample Slurry prepared from ground sample Slurry prepared from sample shaken with zirconia beads Slurry prepared from sample ground with acetone in mortar Matrix Element Cd Biological and environmental Hair Cd — — Se Sediments and natural waters Hair Cd Mn Pb Hair Pb Rice Cd Cd Pb Tl Soils and sediments Soils Cr Ni Geological Cu Se Lyophilized seafood Se Soils and sediments Fruit Various (5) Sediments Al Fe Si Journal of Analytical Atomic Spectrometry August 1997 Vol.12 Sample treatment/comments Escherichia coli and Pseudomonas putida used for preconcentration pH range optimized. LODs 0.04 and 0.5 mg l-1 for E. coli. and P. putida respectively. Samples 0.5 g pulverized and a 0.1 g portion suspended in 25 ml H2O. 100 ml portion of slurry adjusted to contain 2 g l-1 NH4H2PO4 and 4 g l-1 glycerol and diluted to 1 ml. 20 ml aliquot taken for ETAAS measurement. LOD 22.5 ng g-1 RSDs in the range 2.5–6.9%. Preconcentration and separation on activated carbon. Sediment digest and waters extracted with 50 mg activated carbon to preconcentrate and collect SeIV complex.Mixture filtered and the activated carbon phase suspended in 5 ml H2O containing 0.5 ml 2000 mg l-1 PdCl2. A 20 ml portion of the magnetically stirred slurry was analysed by ETAAS. RSDs in range 4–5% recorded. Chemical modifier of Pd used for Mn and of Pd–Mg(NO3)2 used for Cd and Pb. ‘Fast furnace’ program used. LODs 14.2 9.9 and 13.4 ng g-1 for Cd Mn and Pb respectively. Results agreed well with certified values. Slurry prepared in H2O+4% glycerol. ‘Fast furnace’ program used with mixed modifier of Pd–Mg(NO3)2. LOD 0.2 mg g-1 RSD (n=11) in range 1.7–5.3% for 0–30 mg l-1 of PbII. See also ref. 252. Samples mixed with xanthan gum and 0.1% HNO3 and stirred for 10 min.RSD 5%. Samples suspended in 5% v/v HF (30% v/v HF used for Tl) and magnetically stirred. 10 ml aliquot taken for analysis by ‘fast furnace’ ETAAS program. Aqueous calibration for Cd and Pb standard additions used for Tl. Results agreed well with certificate values. Samples mixed with 5% HNO3 and sonicated in autosampler cup. 20 ml aliquot analysed. Good agreement obtained between slurry results and those obtained by FAAS after aqua regia digestion. See also ref. 250. Samples 10 mg+5 ml of 10% v/v glycerol sonicated for 30 s. 10 ml aliquot of slurry taken for ETAAS measurements with 10 ml of 5% m/v NH4F. Up to ten additions of slurry were made if preconcentration was required. RSD (n=6) 8.6% for 1.5 mg g-1 Cu.3 Samples suspended in 5 ml H2O containing 10% v/v EtOH 10% v/v H2O2 1% v/v HNO and 0.08 g Ni(NO3)2. 20 ml aliquot taken for analysis by ‘fast furnace’ ETAAS program. LOD 0.2 mg g-1 Se with a 1% m/v slurry. See also ref. 258. Samples suspended in 1000 ml HF containing 1% Ni(NO3)2 and manually shaken. 10 ml aliquots taken for analysis by ‘fast furnace’ ETAAS program. LOD 0.1 mg g-1 Se with a 10% m/V slurry. Results agreed well with certificate values. See also refs. 255 259. Fresh samples+Triton X-100 shaken to form slurry. Results agreed well with certificate values. 3 Samples 3–5 mg +100 ml 65% HNO left overnight then diluted with 1 ml H2O. 100 ml aliquot diluted with 1.5 ml H2O agitated ultrasonically and analysed by ETAAS using non-pyrolytic graphite platform atomization.Method validated using SRMs. Reference 248 249 250 251 252 253 254 255 256 257 259 260 261 262 341R Table 2 (Continued) Type of sampling Slurry prepared by vortex mixing Slurry prepared by high-speed blending and homogenization of frozen samples Slurry prepared by high-speed blending and homogenization Slurry prepared from powdered sample by vortex mixing or manual ultrasonic probe agitation Slurry automatic preparation with an ultrasonic probe Slurry automatic preparation with an ultrasonic probe Solid Solid Solid Solid Solid Solid dried and ground Solid samples pressed into pellets 342R Journal of Analytical Atomic Spectrometry August 1997 Vol.12 Matrix Element Rice flour Cd Cd Cu Pb RMs liver and kidney Various (6) Zoological and botanical RMs and animal feeds Hg Fluorescent lamp cullet–phosphor Fe Na Graphite and quartz powders Zr-based materials Various (15) PVC Mg Hg Environmental and biological RMs Various (4) Various (4) Ferrous and nickel alloys Geochemical Various (4) Cd Cu Ground water colloids — — Reference Sample treatment/comments 263 Samples 25–75 mg mixed with 500 mg 3% tapioca flour suspension in 0.1% Triton X-100 for 30 s. Aliquots of 1 ml of slurry mixed with 5 ml 1000 mg l-1 Pd solution and taken for ETAAS measurement. LOD 0.28 ng g-1 RSDs in the range 3.4–9.6%. Results agreed well with those obtained after digestion.Samples 2 g or RMs 0.1 g mixed with 20 ml 264 of EtOH–water (1+9 v/v) containing 0.25% m/m TMAH. Mixture blended at 20 000 rev min-1 for 60 s resulting suspension processed through homogenizer four times. NH4H2PO4 used as modifier for Cd NH4NO3 for Cu and Pd(NO3)2 for Pb. Results in good agreement with analysis by ICP-MS after acid digestion. Slurries were stable for 6 d. 10 ml aliquots taken for analysis by ‘fast furnace’ ETAAS programs. No significant dierences observed between aqueous calibration and method of analyte additions. Method limited by Pb and Cu contamination from homogenizer. 265 Slurries prepared and determined as described in reference 264. Capping flat valve head of homogenizer with a ruby disc reduced but did not eliminate contamination.Addition of EDTA to solvent increased mobilization of Fe but increased contamination from homogenizer. 266 267 Samples 10–40 mg mixed with 0.05% HNO3 and Pd(NO3)2 (50 mg Pd per g of slurry) using vortex mixing or ultrasound. Ultrasound was preferred. LOD 24 ng g-1 RSD (n=9) 8.3% for 1.55 mg g-1 Hg. 233 Autosampler modification allows preparation of large slurry volumes in 15 ml cylindrical beakers. Improved precision compared with use of conventional 2 ml cups. 268 Samples 5–150 mg suspended in 10 ml H2O. Chemical modifier of H3BO3–HF used for Al NH4H2PO4–Mg(NO3)2 for Cd and Ca(NO3)2 for Si. Slurry homogenized in an ultrasonic bath and then with automated ultrasonic probe in autosampler tray.20 ml aliquot taken for analysis. Instrumental parameters for 15 elements given. LODs for Ca Cd K Mg Na and Zn in the range 1–20 ng g-1. See also ref. 233. 269 Sample 1–3.5 mg analysed directly in furnace. RSD (n=10) 6.4% for 1.95% Mg. Results compared well with FAAS analysis. The five samples h-1 throughput for ETAAS was considered better than FAAS and digestion. See also ref. 232. 222 238 Sample mass up to 300 mg analysed in nickel atomizer. LOD 0.1 ng. See also Table 1. The generalized standard additions method and linearization algorithms were tested for use in simultaneous multi-element analysis. The combination yielded acceptable analytical results for the elements studied.247 Bi and Pb determined using aqueous standards Se and Sn required calibration using solid RMs. LODs 0.20 and 0.03 mg g-1 for Bi and Pb. Method recommended for screening for Se and Sn. 270 271 Chemical modifier of Pd–Mg(NO3)2 used. Good agreement between solids analysis and analysis after acid extraction. Mixed Pd–Mg modifier used. Samples directly introduced into the graphite atomizer using a micro-piston sampler. 234 Graphite boat used pellet technique enables use of sample mass of 15–20 mg compared with the #2 mg sample mass used normally. LODs improved 10-fold. Table 2 (Continued) Matrix Element Type of sampling Industry dust Various (5) Solid samples cut from trapping filters Biological Various (4) Solid and slurry prepared from dried samples ground using agate beads mean particle size 10 mm technique was proposed as being a useful one for surface analysis and quantification was reported.Direct solids analysis was used by Hinds et al.247 for the semi-quantitative determination of volatile elements in alloys. Aqueous calibration was possible for Bi and Pb but incomplete release of Sn and matrix eects in the determination of Se resulted in the need for calibration against solid reference samples for accurate determination of these two metals. It was concluded that the technique was a valuable screening tool for materials that may contain abnormal levels of Sn and Se. 1.2.2.2. Gas sampling. Sneddon’s group at McNeese State University USA appears to be the most active group working on gas sampling with an electrothermal atomizer.In two conference presentations from this research group Smith et al.274 and Indurthy et al.275 discussed the use of a single-stage impactor for sample collection coupled with ETAAS for quanti- fication. This was applied to the determination of Cr Hg and Pb in cigarette smoke. Chakrabarti et al.276 developed a method to determine Se in atmospheric particulate matter. The particulates were collected by filtration through a porous electrographite plate which was then inserted using a probe into a modified graphite electrothermal atomizer (HGA-500). The use of chemical modifiers such as a mixture of palladium and magnesium nitrates or ascorbic acid was necessary in order to obtain stable integrated absorbance signals.Employing one of these mixed chemical modifiers allowed aqueous Se solutions to be used to establish the calibration hence the method of analyte additions was not necessary. The absolute LOD was 36–41 pg although there was no indication in the abstract of the typical volumes of air that were sampled. For the determination of Se in NIST SRM 1648 Urban Particulate Matter recoveries of 88% were obtained. 1.2.2.3. Coupled techniques and preconcentration. The publication of reviews covering a relatively new technique tends to be a good indication that the technique has become accepted or at least sucient work has been generated to make it worth the authors preparing a review! Matusiewicz and Sturgeon277 have reviewed ‘in-atomizer trapping’ with a graphite atomizer for the determination of hydride forming elements coupled with ETAAS CVAAS MIP-AES ICP-MS and FANES after both batch and FI sample introduction (82 references).Similarly Chinese workers278 assessed the progress of ‘inatomizer trapping’ with ETAAS (36 references). What is clear from the literature presented on ‘in-atomizer trapping’ is that the use of a graphite atomizer coated with metallic trapping reagent to serve as both the hydride trapping medium and atomization cell has become well established as the method of choice for ultratrace applications of the HG technique. Ding and Sturgeon279 evaluated electrochemical HG for the determination of total Sb in natural waters by ETAAS with ‘in-atomizer trapping’.A continuous flow electrochemical HG (timed injection) system with lead as the cathode material was Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 Reference Sample treatment/comments 272 Cup-in-tube method used. Samples deposited in graphite cup with 10 ml 0.2% HNO3 for Cr Cu and Ni or 10 ml Pd for Cd and Pb. Reasonable agreement between solid analysis and analysis after wet ashing. 273 Graphite cup atomizer and oxygen ashing used. Solid sample discs punched out of plant material and deposited in cup. Ground samples 20–200 mg plus 2 ml H2O sonicated before analysis. Results agreed well with certified values. developed for the production of stibine. Both SbIII and SbV were equally converted into their hydrides by electrochemical means with 92±4% eciency and the stibine was trapped in a pyrolytic graphite coated electrographite tube coated with palladium prior to atomization.An absolute LOD of 45 pg (3s) and a concentration LOD of 0.02 mg l-1 were obtained using 2 ml sample volumes. For measurements at a concentration of 0.2 mg l-1 RSDs were <6%. The accuracy of the procedure was assessed by the analysis of SRMs SLRS-2 and SLRS-3 river waters and NASS-4 open ocean sea-water. The advantage of using electrochemical HG compared with the traditional use of sodium tetrahydroborate in acidic media is that electrochemical HG is less prone to contamination problems and avoids the need for a pre-reduction step of SbV to SbIII.However these systems are acknowledged to be very sensitive to sample matrix and it was indicated that this is to be studied in subsequent work. Matusiewicz et al.280 applied HG and ‘in-atomizer trapping’ for the determination of Bi in clinical samples with trapping at 350 °C in a graphite atomizer coated with reduced palladium and iridium. Samples were digested with nitric acid and hydrogen peroxide in PTFE bombs in a microwave oven. An LOD of 100 pg was found though no indication was given of the sample volume taken. The results for a urine RM agreed well with the certified value. Hydride generation and ‘in-atomizer trapping’ with a commercially available system was applied by Erber et al.281 for the determination of Pb. Neither the HG of plumbane nor the atomization of Pb within a conventional heated quartz atomizer is easy and the application of an FI technique allied with ‘in-atomizer trapping’ appears to overcome some of these problems.The HG process was optimized by using an ammonium peroxodisulfate–hydrochloric acid system as oxidant and sodium tetrahydroborate as reducing agent. The addition of sodium cumol sulfonate as a surfactant was advantageous to ecient plumbane generation. An LOD of 0.7 mg l-1 for a 500 ml sample volume was found. Another element which is not generally determined by HG is Cd; however Infante et al.282 from Sanz-Medel’s group at the University of Oviedo Spain have extended the work concerning the generation of cadmium hydride to combining this with ‘in-atomizer trapping’ in a graphite atomizer.The cadmium hydride was generated from a vesicular medium of didodecyldimethylammonium bromide with sodium tetrahydroborate and trapped on a palladium coating pre-heated to 150 °C on a pyrolytic graphite L’vov platform followed by atomization at 1600 °C. An LOD of 60 ng l-1 was found with a sample volume of 1.4 ml and an RSD (n=10) of 1.7% for 500 ng l-1 of Cd. The procedure was performed manually with the cadmium hydride transferred into the graphite tube via a quartz capillary tube placed with its end in contact with the L’vov platform and using a continuous flow (CF) procedure. The LOD was comparable with a typical ETAAS value (20 ng l-1 for a 20 ml sample volume) although it remains to 343R be seen how reliable the HG procedure is in the presence of typical HG interfering elements.‘In-atomizer trapping’ procedures are not solely the preserve of elements that form volatile hydrides. Erber and Cammann283 described ‘inatomizer trapping’ of nickel carbonyl. The sample in a carrier stream of 0.08 mol l-1 sulfuric acid was merged with a stream of 1% (m/v) sodium tetrahydroborate solution followed by CO and the mixture passed through a 20 cm reaction coil. The nickel carbonyl produced was carried by an Ar stream to a gas–liquid separator and then to the THGA graphite atomizer pre-heated to 250 °C where the nickel carbonyl was decomposed and the Ni produced trapped on the graphite surface; atomization was at 2400 °C. Both FI and CF modes were used for sample introduction with sample volumes of 0.1–1.0 and 1–50 ml respectively.LODs were 0.18 and 0.01 mg l-1 for the FI and CF modes respectively although no indication was given as to which sample volumes were employed for the detection limit calculations. The procedure was applied to the determination of Ni in two CRMs water and rye grass and the results obtained agreed with the certified values. The three publications concerning ‘in-atomizer trapping’ of elements such as Cd,282 Pb281 and Ni283 found LODs not dissimilar to those for conventional ETAAS measurements. In general workers using ‘in-atomizer trapping’ for traditional HG elements found improved LODs (typically low ng l-1 values). The fact that for Cd Pb and Ni this is not the case indicates either low eciencies in generating the volatile analyte species or trapping in the atomizer.Clearly these aspects and the eects of interfering elements warrant further investigation. Tao and Fang284,285 proposed an alternative procedure for the determination of Ni using on-line FI solvent extraction coupled with ETAAS. The nickel–pyrolidin-1-yl dithioformate chelate was extracted on-line into 4-methylpentan-2-one. The organic phase was separated from the aqueous phase with a novel gravity phase separator with a small conical cavity and stored in a collector tube from which 50 ml of organic solution were introduced into the graphite atomizer by a flow of air. An enrichment factor of 25 was obtained in comparison with the direct introduction of 50 ml of aqueous Ni solutions and an LOD of 4 ng l-1 (3s) found with an RSD of 1.5% for 1 mg l-1 of Ni.The method was applied to the determination of Ni in body fluids and other biological samples. The ‘in-atomizer trapping’ of volatile compounds in a graphite atomizer can be readily automated owing to the ease with which gases or vapours can be pumped or passed into the atomizer and there are no volume constraints on the amount of gas. However when considering the use of on-line column preconcentration procedures coupled with a graphite electrothermal atomizer the experimental parameters become more complex as the volume injected into the atomizer has to be carefully controlled and consequently the sequence timing becomes more critical.Sperling et al.286 have been active in investigating this area of concern for some time. They compared dierent manifolds for preconcentration coupled with ETAAS. Even with stringent control of dispersion by using narrow bore tubing avoidance of dead-volumes and miniaturization of columns etc. the eluate volume of such manifolds often exceeds the capacity of the graphite electrothermal atomizer. Several procedures have been developed to ensure that only the eluate fraction with the highest analyte content is passed into the atomizer. While FI oers the reproducibility to obtain the necessary precision the technique suers from the problem of long-term reproducibility and the need to optimize the ‘zone-sampling’ window without the help of on-line detection of the elution profile.This is extremely tedious. These workers have solved this problem by developing an FI manifold that allows the introduction of the complete eluate volume into the graphite atomizer avoiding the previous 344R Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 diculties. After merging chelating compounds with the sample the analyte complexes formed on-line are either sorbed onto microcolumns packed with silica C18 material or on the inner walls of knotted reactors (KR) made of PTFE tubing. After loading the sample an air flow is introduced to remove the residual solution from the KR or column and the adsorbed chelate is quantitatively eluted with a pre-defined volume of 50 ml of ethanol propelled by an air flow.This leads to improved recoveries and precision. Di and Davey287 investigated on-line column preconcentration procedures for Au Ir Pd and Pt. Clearly the interface between an FI unit and a graphite atomizer system is more dicult than an FI and a flame interface and these workers found that manual control for the sample introduction into the graphite atomizer was preferable over automated introduction. The sample injection time into the atomizer needed to be carefully controlled to maintain the precision. Chinese workers288 described an on-line liquid flow extraction system with a specially designed phase separator coupled with ETAAS. It appears that a common three-way confluence connector was used for segmentation but further details are not available.This approach sounds similar to the separation system described by Tao and Fang,284,285 but insucient data are available to elaborate on any dierences between these two descriptions. Yan et al.289 applied on-line sorption preconcentration in a KR coupled with ETAAS for the determination of SbIII in water. The SbIII was selectively complexed with APDC at an acidity of 1 mol l-1 HCl the complex sorbed onto the inner walls of a KR made from 100 cm×0.5 mm id PTFE tubing followed by quantitative elution with 35 ml of ethanol into the graphite atomizer. A similar procedure to that described by Sperling et al.,286 was employed for complex elution from the KR i.e. an air flow was used to drive the eluent elution and deliver all the eluate into an uncoated electrographite tube without a platform.With a preconcentration time of 30 s and a sample loading flow rate of 5.4 ml min-1 an enhancement factor of 30 was obtained in comparison with a direct injection of 35 ml of aqueous solution resulting in a sampling frequency of 42 samples h-1. The LOD (3s) was 0.021 mg l-1 and the RSD was 2.9% at 1 mg l-1 (n=11). The introduction of an ecient wash step before elution with the addition of APDC to the wash solution permitted the use of simple aqueous solutions for calibration for the analysis of sea-water. As the SbV complex does not form at an acidity of 1 mol l-1 HCl this procedure allows selective preconcentration and determination of SbIII in the presence of SbV.There were no interferences from the alkali and alkaline earth elements due to the specific nature of the complexing agent although CuII FeIII NiII and SeIV interfered in large excess owing to competition for the chelating agent. Schuster and Schwarzer290 showed what can be achieved with on-line column preconcentration coupled with ETAAS. The 18 work described a fully automated system for the determination of very low levels of Pd in 1 g l-1 AAS and ICP standard solutions of ruthenium rhodium osmium iridium platinium silver and gold. A variety of N,N-diethyl-N¾-benzoylthiourea complexing agents were examined and N,N-diethyl-N¾¾- benzoylthiourea (DEBT) was found to be the optimum. Initially the microcolumn [polymethylpentene (TPX) cylinders 1.0 mm id 13 mm long filled with 10.2 ml of C Polygosil 40–63 mm] was reversibly loaded with the complexing agent.Subsequently the sample solution was moved through the column. After complex formation a wash step with dilute nitric acid 6.5% (v/v) was performed to remove interfering matrix constituents. The palladium complex and excess ligand were eluted with 60 ml of ethanol and directly introduced into the graphite THGA atomizer. Sensitivity enhancements of 40 70 and 200 were obtained from sample volumes of 0.8 1.86 and 4.85 ml respectively. The corresponding LODs were 51 25 and 13 ng l-1 calculated from the calibration graph (n=10 P=95%). In 6.5% (v/v) nitric acid solutions alkaline alkaline earth and base metal ions such as CoII CuII FeII and NiII as well as precious metals were tolerated up to concentrations of 10 g l-1.The work of Yuan and Shuttler291 concerning on-line preconcentration of Al coupled with ETAAS discussed previously, 125 has now been published. Two preconcentration materials were examined quinolin-8-ol immobilized on controlled pore glass and Amberlite XAD-2 and while both systems were suitable for the preconcentration of Al neither showed a clear advantage over the other. For the determination of Al in drinking water recoveries of 100–115% were found for the quinolin-8-ol immobilized on controlled pore glass system and 90–100% for the Amberlite XAD-2 system. However on applying the method developed to the determination of Al in potable fresh river and sea-water samples varying results were obtained.Malcus et al.292 discussed an alternative approach to on-line column preconcentration and constructed a computer controlled system which consisted of four parallel supported liquid membrane devices for analyte enrichment. The system was tested with 40% v/v di(2-ethylhexyl)phosphate (which is an ecient extractant for many metals) in kerosene as the liquid in the porous PTFE membrane and analyte elements of Al Cd and Cu at concentrations below the mg l-1 level. The driving force for transport across the membrane system was pH with a pH of 4 on the donor side and 0.3 on the acceptor side. After preconcentration the acceptor solution was eluted and collected in the autosampler prior to ETAAS measurements.Clearly while the enrichment procedure is fully automatic the measurement step requires collection and subsequent sampling into the electrothermal atomizer which is a disadvantage compared with the procedures described by Sperling et al.286 and Schuster and Schwarzer,290 wherein the enriched solution containing the analytes is passed directly into the electrothermal atomizer. Malcus et al.292 found enrichment factors of 7.3 7.7 and 8.5 for Al Cd and Cu respectively using a 30 min period for processing the samples in the supported liquid membrane step. The RSDs were 6.5% at 0.5 mg l-1 and 6.3% at 1.0 mg l-1 based on 108 independent measurements. The blank for Al was 0.15 mg l-1 with an RSD (n=16) of 11%. For the purposes of speciation studies a non-flow through system such as ETAAS does not recommend itself but on the other hand the selectivity and detection power is what one needs for such studies and consequently a small number of workers each year attack the dicult challenge of coupling HPL C with ETAAS.In general it would be fair to state that as yet there is no ideal solution and most workers end up with a system whereby the HPLC eluate is collected in aliquots and these are then sampled into the ETAAS system. This does have the disadvantage that the HPLC resolution is degraded owing to the time- or volume-based collection of the eluate. d’Haese et al.293 reviewed the HPLC–ETAAS hybrid method for studying protein binding and speciation of trace elements in biological materials.The method involved the use of a BioGel TSK-DEAE-SPW anion-exchange column and a similar guard column with ETAAS determination of Al Fe and Si with CV determination of Hg. The system was kept metal-free to reduce interferences and a silica-based scavenger column was placed prior to the injection valve to retain Al and Fe present in the buer solutions. Trace element recoveries were of the order of 100% and RSDs <10%. The method was suciently sensitive to allow determinations at clinically relevant levels. Gilon et al.294,295 optimized the chromatographic parameters for the determination of Se species using HPLC coupled with ETAAS. The selenite selenate selenocystine and selenomethionine species were separated using either a cationic or an anionic reagent and ion-exchange with either a reversed- Journal of Analytical Atomic Spectrometry August 1997 Vol.12 phase Hamilton PRP-1 column or an anion-exchange one the Hamilton PRP-X100. The nature of the mobile phase strongly aected the ETAAS measurements. Addition of a palladium chemical modifier was required in the case of a heptane sulfonate HPLC eluate whereas the nickel acetate mobile phase on the anion-exchange mechanism was on its own a good chemical modifier. Unfortunately in the published work the conclusions were that separation using ion-pair chromatography with sodium heptanesulfonate as the anionic counter ion and performing the ETAAS measurements with palladium as the chemical modifier was preferred. However the conference abstract indicated that the nickel acetate solution was preferred.For the on-line determination of tributyltin and dibutyltin in sea-water Bermejo-Barrera et al.296 coupled a mBondapak C18 column with methanol–water–acetic acid (25+18+7) containing 0.01% tropolone as the mobile phase. Column fractions (98 ml ) containing the two Sn species were automatically injected with air as carrier gas into the pyrolytic graphite coated electrographite tube fitted with a pyrolytic graphite L’vov platform. It is dicult to understand how such a large volume could be injected into the atomizer automatically as 98 ml exceeds by three-fold the maximum capacity of the L’vov platform used. Unfortunately the exact details of the experimental procedure were not given in the abstract although a comparison was made with an o-line HPLC and ETAAS procedure and the results compared favourably.The LODs for the on-line system were 3.1 and 2.8 ng for tributyltin and dibutyltin respectively based on sampling 2 l of water. Vinas et al.297 identified vitamin B12 analogues by liquid chromatography coupled with ETAAS. However this appears to have been an o-line procedure as the HPLC eluate was collected every 15 s (equivalent to 375 ml ) and 25 ml aliquots were then taken for ETAAS measurements for the determination of Co. 1.2.2.4. Electrodeposition. As noted in previous Updates there is very little work being performed on electrodeposition although it appears to be of interest for some special cases.A major drawback to this technique is the length of time electrodeposition takes. In general most procedures described require 5–10 min for this thus the overall analysis time is exceedingly slow and automation is dicult. Vrana and Komarek298 determined Cd and Cu by electrodeposition on a graphite disc electrode at -1 and -0.7 V respectively with 0.01 mol l-1 sodium nitrate at a pH of 4–4.5 as the supporting electrolyte. Almost quantitative recoveries were obtained with deposition times of 2–10 and 2–4 min for Cd and Cu respectively. The electrode was then placed by the autosampler into a graphite tube atomizer for measurement by ETAAS. These workers found that sensitivity increased linearly with deposition time with LODs of 5.9 ng l-1 for Cd after 10 min and 100 ng l-1 for Cu after 4 min.The method was applied to the analysis of sea-water. In addition the speciation of Cu free and bound to EDTA was also described. Chinese workers299 have continued their work1 on the use of a tungsten loop upon which Pb is electrodeposited before insertion into a graphite tube atomizer ETAAS system. In the current work,299 labile Pb in foodstus was determined and the sensitivity for Pb was increased 80-fold compared with conventional ETAAS. The LOD was 0.01 mg l-1 and recovery of labile Pb from foods was in the range of 96–103%. 1.2.3. Fundamental processes In a provocative title L’vov300 asked ‘Gaseous carbide theory. Has it been buried prematurely?’ and reviewed the present state of the reduction of oxides by carbides (ROC) theory in ETAAS.He concluded that the counter arguments of Holcombe et al. can be accommodated. While L’vov admits that errors were made in some of the indirect arguments used 345R in the course of the development of the gaseous carbide model he believes that analysis of quadrupole MS studies of spikes does not weaken the position of this model. The main thesis of the theory namely the presence of an over-equilibrium content of gaseous carbides he believes is confirmed by direct MS evidence and that all arguments presented against the ROC theory can be answered with an extended version of the model. This worker considers that the ROC hypothesis remains the only working model capable of explaining the formation mechanism and the various features of spikes under dierent conditions though this does not mean that it can account for all features of the process and that further improvements are needed.L’vov300 concludes by stating that in his opinion the studies of spike formation are as attractive now as they were 15 years ago when they were observed for the first time. It would appear that this Update author’s (ILS) comments in 199572 that the ROC theory appeared to have been discarded was presumptuous! Lamoureux et al.301 applied ETV–ICP-MS in a configuration which also allows simultaneous measurement of atomic and molecular absorption signals and MS signals and shadow spectral filming (SSF) and shadow spectral digital imaging (SSDI) to investigate the mechanisms of Al spike formation and dissipation in ETAAS.Aluminium sub-oxide (AlO and Al appearance of both Al atom spikes and Al containing molecular 2O) and CO(g) spikes in ICP-MS correlated with the values for Ea of 120 and 250 kJ mol-1 were obtained in the low- and high-temperature region respectively. However at a spikes in absorption spectrometry. The Al carbide (AlC2) signal higher heating rate of 700 K s-1 an Ea that increased from in ICP-MS was not coincident with the appearance of either Al atom spikes or Al containing molecular spikes in absorption spectrometry. The results from the SSF and SSDI experiments provided temporal and spatial resolved absorption profiles of Al atoms and Al containing molecules during Al spike formation and dissipation.The transverse cross-sectional distribution of Al atoms and of Al containing molecules in the graphite atomizer were complementary to one another for both wall and platform atomization. The highest concentration of Al atoms was near the graphite surface whereas the highest concentration of Al containing molecular species was at the centre of the graphite tube. The Al containing molecules This work was further extended by Rojas304 to consider the observed in both wall and platform atomization consisted of eects of copper ascorbic acid and Triton X-100 on the both gaseous Al molecules and a non-uniformly distributed atomization processes of Ag as a function of the initial mass of cloud of finely dispersed Al2O3 (s l ) particles. It was proposed analyte and the heating rate of atomization.In general a that Al spikes are formed from gaseous Al2O precursors and double-pulse structure was detected at a heating rate of that this reaction is triggered by the formation of a molten condensed phase Al4C3 melt. Further evidence for the presence of condensed phase particles was presented by Hughes et al.302 also working in Chakrabarti’s group at Carleton University Ottawa Canada. Using SSDI the condensation of Au and matrices of various chemical modifiers and compounds in ETAAS was investigated. Spatial and temporal non-uniformity in light scattering was observed and these eects were attributed to the formation of condensedphase particles during the process of high-temperature vaporization and atomization.The materials investigated were Au magnesium chloride sodium chloride diammonium hydrogenphosphate lanthanum nitrate and a mixture of palladium and magnesium nitrate. The non-uniform distributions of the condensed phase particle clouds were attributed to the gas flow patterns that developed in the graphite tube atomizer during heating as well as the steep temperature gradients that developed along the longitudinal axis of the end-heated tube of the Massmann-type graphite atomizer. Dierences observed with fast and slow rates of heating were related to thermal expansion of gas and diusion eects. Use of either a graphite platform or a low Ar purge gas flow during the high temperature heating of the graphite atomizer was found to reduce condensation of matrix vapour and to improve the accuracy of continuum source background correction.These results con- firmed some of the earlier studies by Frech and co-workers 346R Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 concerning the condensation of matrix vapours within a graphite atomizer. It is clear that the ROC theory is an emotive issue for all those concerned in its development or rebuttal and the paper by L’vov300 gives a feel for the passion of the pro and counter arguments. Considerable work continues to be performed with respect to atomization mechanisms. A clear trend in this area is dicult to discern as many parameters aect the final result. However it is now clear from both surface studies and those examining particle formation in the gas phase that the atomization mechanisms are complex and can in fact vary under dierent electrothermal atomizer conditions such as heating rate and presence of matrix components.Rojas and Olivares303 continue to work on atomization mechanisms and have now turned their attention to Ag. They proposed a two-precursor kinetic mechanism to account for the double peak structure detected in the atomization signals for Ag. The rate constants for the dissipation of the atomic vapour were determined from a single absorbance pulse under conditions of extended time with varying temperature. From numerical simulations two temperature regions of atomization with well defined atomization energies (Ea) were predicted by the Arrhenius plots for systems in which the vaporization processes were at least partially time resolved.Employing a heating rate of 300 K s-1 average 125 to 240 kJ mol-1 as the initial mass of the analyte increased was obtained in the low-temperature region whereas a mass average atomization energy of 117 kJ mol-1 with a first kinetic order of release was obtained in the high-temperature region. The results seem to indicate simultaneous first-order atomization from dispersed particles and also from small clusters whose size increased as the initial mass of Ag increased. In the opinion of these workers the agreement between the experimental atomization energies and those predicted from the numerical simulations provided an excellent description of the entire absorbance profiles.a a 300 K s-1 in the absorbance signal. In the presence of copper and Triton X-100 a low Ea was obtained in the lowtemperature region and a high Ea which approached the heat of vaporization of Ag was obtained in the high-temperature region. However in the presence of ascorbic acid desorption energies that were too low were obtained in both temperature regions suggesting a higher dispersion of particles owing to the presence of a higher number of active sites. At a heating rate of 700 K s-1 a single atomization step with an atomization energy of 233 kJ mol-1 and a first order of release was detected in the presence of copper. In the presence of ascorbic acid and Triton X-100 two temperature regions of atomization were obtained from the Arrhenius plots even though the absorption profiles appeared continuous.In these cases a mass dependent E was obtained in the low-temperature regions and a low Ea with a first kinetic order of release was obtained in the hightemperature region. It would appear that the low value of Ea indicates vaporization from disperse particles whereas the mass dependent higher value of E indicates atomization from small clusters the size and energy of which increase as the initial mass of Ag increases. These results agree with those found in the previous study and correlate well with those predicted by the two-precursor atomization model proposed. It has to be noted that the experimental heating rates during atomization are considerably slower than those generally recommended for generating analytical data and for this reason it is often dicult to correlate the results of such studies with what is often seen or found during routine measurements.Progressing along similar lines Slaveykova et al.305 applied the Kelvin equation to the vaporization of Ag and Au in a graphite electrothermal atomizer. They found a quantitative relationship which was derived from the Kelvin equation between the enthalpy of vaporization and the size of the released particles and applied this over a wide range of working conditions to study the vaporization mechanisms. From these studies indirect evidence for the island-structure droplet model for the vaporization behaviour of these metals emerged.Alvarado306 presented a traditional study of the atomization mechanisms of Co Fe and Ni in a graphite electrothermal atomizer using thermogravimetric and X-ray powder diraction studies of the residues formed on heating metal nitrate solutions with and without the presence of powdered graphite. In the absence of graphite the oxides were seen to be stable to a temperature that was at least 150 °C higher than the temperature at which rapid carbon reduction was observed. The mechanism proposed was that the atomization of Co Fe and Ni occurs via direct evaporation of the respective metals. Reduction of the metal oxides to solid metal by carbon was identified as a precursor to atomization. Imai et al.307 examined the atomization signals of In and The sensitivity enhancement was attributed primarily to the reduction of In2O3(s) to In( l ) owing to active sites and trapping of the In( l ) as smaller sized droplets on the active sites.The amorphous carbon layer plays an important role as an area for movement collision and coalescence of In droplets and produces a decrease in the tendency of larger sized droplets to diuse on the graphite surface. Benzo et al.309 studied the atomization of V from a graphite surface by experimental and theoretical modelling to assess the eects of mineral acids on the memory eects usually encountered during the atomization of V in a longitudinally heated graphite electrothermal atomizer. This appears to be a continuation of previous work from this group concerning Mo,111,125 applying the complete neglect of dierential overlap (CNDO) method to performing quantum chemistry calculations of V–graphite interactions to aid the understanding of the chemistry associated with the atomization of V.The eect of mineral acids on the atomization of V was investigated by modelling co-adsorption of protons on the graphite surface. The most important variables and interactions associated with the atomization of V were investigated using factorial designs and response surface methodology. These workers found that the presence of nitric and hydrochloric acids enhanced the signal for V in the concentration range studied and that the presence of these acids favoured the formation of V species that can be more easily atomized.The optimum V signal was found with an acid concentration of 3–4% v/v and a pyrolysis temperature of 1400–1700 °C; the duration of the pyrolysis step had no eect. Memory eects were confirmed by the release of some V species chemisorbed onto the pyrolytic graphite tube surface when an acidified solution was used as a blank. The theoretical calculations indicated that the presence of protons allows the migration of dispersed V atoms hence the V atoms can agglomerate in the centre or in sites located far from the co-adsorbed protons and the agglomerated V atoms can be atomized more readily than the dispersed V More workers are utilizing ET V–ICP-MS to explore atomizally been considered. In the previous Update workers from Chakrabarti’s group (Carlton University Canada) investigated the atomization mechanism of U.1 During the period covered by the present Update Byrne and Carambassis311 combined theoretical and experimental approaches to study the atomization processes of Nd in graphite ETAAS.Experimental results using graphite ETAAS and ETV–ICP-MS showed that atomization occurred by two distinct atomization processes gaseous oxide dissociation and molecular carbide dissociation. The oxide dissociation begins at a slightly lower temperature but otherwise these processes occur concurrently. These results appear to explain the improvement in sensitivity that is observed when carbide formation is precluded by the use of a tantalum atomizer lining or a non-graphite tube and suggests that optimum atomization eciency could be achieved by minimizing both oxide and carbide formation by including hydrogen in the purge gas.A thermodynamic equilibrium model based on the thermal dissociation of gaseous rare earth monoxides was proposed that accounts for the wide variation in ETAAS sensitivities among the REEs and shows a good correlation between the measured characteristic masses and the corresponding monoxide bond dissociation energies. observed double and single peaks when electrographite and pyrolytic graphite atomizers respectively were used. As the Arrhenius plots for the single-peak signal in the pyrolytic graphite atomizer were composed of two straight line segments it was proposed that the single peak possibly consisted of two unresolved signals.The eects of zirconium and tungsten treated atomizers and the addition of oxygen and carbon monoxide on the pyrolysis temperature AA signal signal shape appearance temperature of the AA signal and kinetic data were investigated. Based on the interpretation of these results the following processes were proposed for the atomization mechanism in both types of atomizer. For the first signal direct heterogeneous reduction of In2O(g) on the hot graphite atoms on the graphite surface which improved the atomizwall to form In(g) for the second signal direct reduction of ation signal. In2O3 on the hot graphite wall to form the metal followed by Mazzucotelli and Grotti310 examined the appearance tematomization via gaseous In dimers.They also showed that the perature and activation energy of atomization of Se both with sensitivity loss reaction proposed as thermal dissociation of and without palladium as chemical modifier and calcium In2O3(s) to form In2O(g) was catalysed by carbon atoms from potassium magnesium and sodium as matrix elements to gain the furnace wall. Although double peaks were observed in the more information about the stabilization of Se by chemical electrographite atomizer single peaks were observed in the modification with palladium and interfering elements. pyrolytic graphite atomizer. The single-peak signal was attributed to an unresolved double peak-signal involving dierent ation mechanisms and there seems to be a trend towards using pathways of atomization one for gaseous In2O and another this combination to examine REEs which have not traditionfor condensed In metal.The rate-determining step in the atomization mechanism for the first peak is the direct heterogeneous reduction of In2O(g) on the hot graphite surface and for the second peak it is the dimer dissociation process. These results do not appear to contradict those presented previously by Gilmutdinov72 and Majidi and Xu1 and it could be that there are two processes occurring. Further work from Imai et al.308 examined the atomization of In from a pyrolytic graphite surface pyrolysed with an organic matrix solution and when deposited in a bare pyrolytic graphite atomizer with the addition of a matrix solution of ascorbic acid and sucrose.In both cases an unresolved double-peak signal was observed with significant enhancement of sensitivity. As the pyrolysis temperature increased the integrated absorbance signal increased owing to an increase in the second peak. The treated pyrolytic graphite atomizer showed better thermal stability in comparison with the added organic chemical modifier. The double peaks were assessed by means of Arrhenius plots. Smaller sized droplets of In( l ) on either active sites or the thermally stable amorphous carbon from the pyrolysed organic matrix were considered to be the atomizing species in the ratedetermining step for the first and second signals respectively. 347R Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 In contrast to examining the atomization mechanism of REEs Goyal et al.312 investigated the atomization mechanisms and determination of Ag Be Cd Li Na Sn and Zn in uranium–plutonium matrices.The eects of the uranium– plutonium matrix composition on analyte absorbance by ETAAS was studied. The amount of plutonium in the 20 mg l-1 uranium matrix was varied in the range 0–100%. The absorbance signals for Ag Cd Na and Zn increased with increasing amounts of plutonium whereas the signals for Li and Sn did not change and Be displayed up to 75% reduction in absorbance on addition of 10% plutonium to the matrix. The fall in the absorbance signal for Ag Cd Na and Zn in a uranium matrix and its restoration in the presence of plutonium was thought to be correlated with the change in the partial pressure of oxygen released from the matrix at or below the signal appearance temperature.In the case of Li and Sn the signal remains unaected irrespective of the uranium–plutonium matrix which may be because of the high appearance temperatures for these elements. It was considered that the suppression of the Be signal was due to the formation of stable Pu–Be compounds. However this author (ILS) considers it unlikely that plutonium will become a widely adopted chemical modifier! The formation of Pb and Sn molecular species from slurries of their compounds was examined by Tittarelli et al.313 Molecular vaporization can constitute a factor in the atomization characteristics of these elements and can lead to signifi- cant analyte loss with consequent inaccuracy.From an examination of the atomic and molecular absorption spectra obtained simultaneously during atomization the only molecular species observed were SnO SnS and PbS. The Chinese group of Deng et al.188 continues to examine the atomization mechanisms of elements from the graphite probe atomizer. From 20 elements investigated atomization was found to originate from either the thermal dissociation of chlorides oxides or carbides or through the reduction of oxides. Recent years have seen an increasing acceptance that the surface and bulk of an atomizer play an important role in the vaporization and atomization mechanisms. Majidi and co-workers have applied Rutherford backscattering spectroscopy (RBS) to examine the graphite surfaces and the interactions between the surface layers and analyte elements.125 Continuing with this work Rulon et al.314 have investigated the interactions of Pb with pyrolytically coated graphite platforms both with and without hydrogen treatment.Conventional ETAAS measurements show enhancement and a shift to earlier appearance times of the absorption peaks on a hydrogen treated platform whereas delays in the absorption peaks for Pb have been reported for oxygen treated surfaces. An initial RBS study indicated that the surface of a hydrogen treated platform at room temperature was covered by a thin film of deposited Pb solution. Further studies showed that the desorption of Pb was shifted to an earlier time as most of the Pb had vaporized by 420 °C.Previous studies have suggested that chemisorption of hydrogen maintains the carbon surface in a hydrophobic state. Because of this hydrophobicity the liquid should assume a spherical shape to minimize surface area in order to decrease excess surface free energy. An SEM micrograph confirmed the RBS results and showed complete wetting of the surface instead of the anticipated agglomeration. The interaction of Pb with the substrate was expected to be weaker because of the stability of the carbon–hydrogen surface complex. Since most of the active sites are taken up with hydrogen the Pb must migrate towards less reactive sites resulting in weaker bonds. An alternative possible explanation was that hydrogen creates a more strongly reducing environment which favours the dissociation of metal oxides at lower temperatures.The same group continues to explore analytical techniques to examine the surface chemistry of an electrothermal atomizer and recent conference reports315,316 have 348R Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 discussed the use of laser desorption mass spectrometry to evaluate the interaction of Group II metal nitrates with the graphite substrate. The group led by Professor Holcombe University of Texas at Austin USA have also been involved in surface investi- Shepard et al.321 applied continuum source absorption spectrometry to investigate the mechanisms of molecular and atom gations in electrothermal atomizers for some time.Recent work 2 by Jackson et al.317 investigated the migration of Ag Cd and Cu from solutions of the nitrates deposited on highly orientated pyrolytic graphite (HOPG) using ETAAS measurements with platform atomization. Metal migration was verified by removing the top layers of the HOPG platform after the sample had dried and performing the analysis with ETAAS. It was considered that the metals or their salts migrate into HOPG only when the sample solutions are deposited on those areas of the platform that have surface imperfections. The surface blemishes can be seen as tiny lines on the otherwise smooth surface of the HOPG platform. One possible driving force for the migration could be simple capillary action; however additional information is needed to establish the true mechanism.The eect of metal migration was also studied by comparing signals obtained after the sample had been deposited either on the imperfections or on the smooth areas of the HOPG platform. Samples were atomized from both sides of a pyrolytic graphite coated electrographite platform in which one of the sides had been roughened with an abrasive material to expose the electrographite. The main eect of metal migration on the absorption profiles seems to be an increasing tailing of the latter part of the signals. This possibly suggests a secondary generation function limited by the rate of diusion of the metal back to the substrate surface and subsequent vaporization. Continuing with this work Jackson et al.318 examined the surface migration of palladium in an eort to elucidate some of the conflicting theories that have been proposed with respect to the functional role of palladium as a chemical modifier.The use of SIMS allowed data to be collected while the graphite surface was heated. It was clear that migration of palladium below the surface of the graphite occurs at relatively low temperatures while the palladium is still considered as a solid. For some analyte metals e.g. Tl the palladium appears to alter the chemical form on the surface by the time the dry cycle is complete and the analyte appears to migrate below the surface with the palladium. Lamoureux et al.319 have continued to explore the use of X-ray absorption fine structure spectrometry to investigate the mechanism of palladium stabilization.It is well known that palladium either in its reduced or non-reduced form stabilizes Se such that higher pyrolysis temperatures can be used for the determination of Se with graphite ETAAS. Condensed-phase species of Se with or without palladium as a chemical modifier were investigated at elevated temperatures and in situ. At 500 °C an Se K-edge energy shift was measured for Se samples (as H SeO3) with either reduced or non-reduced palladium and this corresponded to an average change in the oxidation state of 4. Structural analysis also showed the presence of palladium and oxygen in the coordination sphere of Se. These results suggest the formation of a compound similar to a palladium selenide.Data analysis of the first derivative of the near-edge X-ray absorption spectrum indicated that Se samples plus PdII underwent chemical changes more readily than Se samples plus Pd0. Giglio et al.320 employed a two-step graphite electrothermal atomizer to provide independent control of the vaporization and atomization processes in ETAAS. As atomization in electrothermal atomizers may proceed through step-wise vaporization condensation or adsorption and re-vaporization processes the two-step atomizer permits these processes to be studied through measurement of the temperature dependency of secondary analyte interaction with the atomizer wall. formation for AlF and InF as molecular absorption spec-trometry (MAS) is a sensitive technique for the determination of halides.Spectral scans were performed to identify the species present as a function of time during the vaporization stage. In addition the sizes of the MAS signals were compared for mixed and physically separated analyte and reagent solutions to establish whether molecule formation occurs in solution or in the gas phase. Hadgu et al.322 developed a vapour transfer model for an analytical techniques including ion chromatography. Croft et al.326,327 examined the mechanisms of chloride interferences on the determination of In. A reduction in the ETAAS signal of In in the presence of sodium chloride and magnesium chloride was observed. The amounts of analyte cation and anion (from the salt) remaining on the platform when heated to various char temperatures were determined using ICP-MS FAAS and ion chromatography.The eects of ascorbic acid and 5% v/v nitric acid as chemical modifiers were studied. In R) of Hg addition visual dierences in the residues left on the graphite platform in the presence of the chemical modifiers were observed with SEM. Grotti et al.328 proposed a new approach based on the end-capped THGA atomizer. The model assumes that the atom density distribution is step-wise linear along the tube axis and there is no injection hole. With a quartz tube system providing controlled experimental conditions at room temperature the time constant of the diusion removal function (t vapour was determined for various open and end-capped tube geometries.These results were also described by an empirical multiple regression equation. The theroretically predicted tR values corrected with an empirical factor of 1.33 agreed well with the experimentally obtained values for the end-capped quartz tubes. For commercial graphite THGA tubes the diusional transfer model was evaluated using the integrated atomic absorbance ratio between various end-capped and open tubes. This is meaningful because the signal ratio for graphite atomizers is closely equal to the corresponding tR ratio. For the recommended atomization temperatures the average deviation between the experimental and theoretically predicted signal ratios for the elements Ag Cd Co Cu Hg and In was 1–5% for various end-capped tube geometries. The results for the individual elements deviated more from the theoretically predicted ratios mainly because of small dierences in the mean gas-phase temperature between open and end-capped tubes.For elements which tend to form molecules in the gas phase at low temperatures and for which the atomization eciency is increased with the atomization temperature the experimental ratios tended to be higher than the theoretically predicted values (Al As In Se and Sn) whereas experimental ratios were slighty lower for other elements (Cd Co and Cu). With respect to computer modelling of events within electrothermal atomizers and in particular Monte Carlo simulations in the past this has mainly been the preserve of scientists with access to supercomputing facilities.Histen and co-workers,323 from Professor Holcombe’s group at the University of Texas at Austin USA have addressed this issue with the development of a C-based program that can be used on a Windows-based PC. With this program basic parameters such as furnace geometry sample placement and kinetic parameters i.e. activation energy and adsorption energy can be changed to show the absorbance profile dependence on these parameters. In addition data such as the time dependent spatial distribution of analyte inside the atomizer can be collected. Rayson and Sae-teung324 have applied computer simulation of the local thermodynamic equilibrium (LTE) and diusion controlled mass transport to model the complex gas-phase chemical environment within a graphite electrothermal atomizer during the thermal pre-treatment stage of an atomizer time–temperature program.As this stage is often relatively long it is thought that these assumptions are valid and free energy minimization calculations of the gas-phase composition of separate segments of the furnace volume can be calculated. The calculations used a database of 5000 possible species. The calculation of diusioncontrolled mass transport between adjacent segments followed by the recalculation of the ‘new’ equilibrium composition enabled the simulation of the environment within the graphite electrothermal atomizer. Russian workers325 reviewed (66 references.) the use of thermodynamics for the investigation prediction and control of thermochemical processes in sources for atomization and spectrum excitation including a tungsten coil and graphite electrothermal atomizers flames a high frequency torch and arc discharges.Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 1.2.4. Interferences Workers at Strathclyde University UK have continued their investigations into chloride interferences using a variety of concepts of experimental design and empirical modelling to investigate interferences in graphite electrothermal atomization studies. The interferences of sodium potassium calcium and magnesium on the ETAAS determination of Te were studied. In the absence of a chemical modifier two peaks were obtained; the first peak was found to be independent of the concentrations of the concomitant elements while the second peak was due to the presence of the interfering elements according to an additive linear model whereby the integrated absorbance could be calculated based on the logarithmic ratios of the interfering element concentration to Te.All four elements had a significant eect. The interference eects depended upon the presence of carbon (from the graphite platform) and could be eliminated by employing a palladium or nickel coated graphite platform. The main purpose of this work appeared to be stressing the advantages of using experimental design to obtain data of good statistical quality with a limited number of experimental measurements. Although not directly concerned with interferences French workers329 applied experimental design theory to the determination of Cu in butter.A twolevel four-factor plan was used to show that drying and heating steps at 50 and 900 °C respectively were not required for optimum ETAAS conditions which required atomization at 2450 °C with a pyrolysis ramp temperature of 40 s. The time– temperature program involved heating from 75 to 2450 °C. The LOD was 5 mg g-1 and the RSDs were 1.1–13% (n=10) over the range 25–200 mg g-1. The interference eects of bone matrix on the determination of Pb were studied by Zong et al.330 using a transverse Zeemane ect ETAAS instrument with a longitudinally heated graphite electrothermal atomizer. It was shown that Pb could be accurately measured using aqueous Pb calibration solutions containing NH4H2PO4 as chemical modifier and Ca(NO3)2 to compensate for the matrix and Mg(NO3)2 was as eective as 3)2 for this purpose.In contrast Cabon and Le Bihan331 Ca(NO examined the influence of various salts on the atomization of Pb with a transversely heated graphite electrothermal atomizer equipped with longitudinal Zeeman-eect background correction. To obtain more information about the interference mechanisms the volatilization of the salts was studied by ion chromatography of the residue remaining in the atomizer after drying and pyrolysis both with and without the application of palladium palladium and magnesium nitrate mixture or magnesium nitrate as chemical modifiers. In 0.1 mol l-1 chloride medium sodium magnesium and calcium chlorides did not interfere significantly; however their dierent behaviour in the atomizer and particularly the hydrolysis of magnesium chloride did influence the pyrolysis curves for Pb.The use of a palladium plus magnesium chemical modifier appeared to be necessary only in the presence of sodium chloride and was able to stabilize Pb suciently to permit the removal of sodium chloride by charring. In the case of magnesium chloride Pb 349R was not suciently stabilized to remove chloride through hydrolysis or volatilization of magnesium chloride. In the presence of calcium chloride the Pb signal was delayed and coincided with the background absorption signal of calcium chloride; the stabilization eect was not sucient to eliminate calcium chloride by charring before atomization.With 0.1 mol l-1 nitrate concentrations the presence of sodium magnesium or calcium nitrates greatly modified the atomization signal for Pb. The stabilization of Pb was greater in the nitrate medium but losses were observed at the decomposition step of the nitrate salts. In this medium the stabilization eect of Pb leads to a single peak and permits elimination of nitrate decomposition products before atomization. Interference eects are more important in the presence of 0.1 mol l-1 sulfate salts and increase with the acidity of the medium. Sodium sulfate which is reduced to Na2S on the graphite surface does not interfere significantly. However decomposition products of magnesium and calcium sulfates induce an important interference eect on the determination of Pb which is stabilized in the atomizer.In the case of sodium sulfate the use of mixed palladium and magnesium nitrates chemical modifier delays the atomization signal such that it coincides with the background absorption signal leading to an important interference eect which cannot be eliminated by charring. In the presence of magnesium sulfate and calcium sulfate the stabilizing eect of palladium permits the elimination of the decomposition products of sulfate salts before atomization and suppresses the chemical interference eect. Along with Pb Al continues to be an element which receives more attention with respect to interference studies and the conclusions drawn were often conflicting.Wieteska and Drzewinska332 discussed the interferences from calcium on the determination of Al in bone and biological samples. Both integrated absorbance and peak absorbance were employed for signal quantification. In samples with low calcium concentrations and use of a magnesium nitrate chemical modifier the interference from calcium was moderate. However with bone samples there were dierences between the results obtained by direct calibration against aqueous calibration solutions and the method of analyte additions. Agreement was obtained after modification of the atomizer time–temperature program and the use of a mixed chemical modifier which contained the bone matrix. These results conflict with those discussed in the previous Update1 from Tang et al.who found that a calcium nitrate chemical modifier in nitric acid was the optimum chemical modifier for determining Al in a bone matrix in preference to magnesium nitrate. Clearly the various parameters studied and the type of instrumentation employed appear to have an eect on the results found in such studies and it is vital that workers in this area state exactly what has been studied how and why. Increasing attention is focusing on the interferences on individual organometallic species especially with respect to the determination of Se in biological samples. Nixon and Moyer333 compared the determination of Se in serum and urine with ICP-MS and Zeeman-eect background corrected ETAAS. When trimethlyselenonium was present in the urine samples ICP-MS gave good recoveries whereas ETAAS gave poor recoveries.In a more detailed study Leblanc334 studied the eect of urinary phosphates on the ETAAS Se signal obtained from solutions containing the trimethylselenonium ion (TMSe+). Following urine digestion with nitric acid–hydrogen peroxide and use of a mixed copper (100 mg) and magnesium nitrate (200 mg) chemical modifier the eect of specific gravity the digestion procedure the amounts and nature of the various inorganic anions pre-dilution of the urine prior to testing and chemical modifier concentration on the TMSe+ signal was clearly demonstrated. It was found that the suppression eect of the phosphates on the TMSe+ signal was more pronounced 350R Journal of Analytical Atomic Spectrometry August 1997 Vol.12 and in test samples spiked with phosphates (in the same urine matrix) correlated with the phosphate concentration than in various dierent samples where the phosphate concentration was experimentally determined. This showed that phosphates are only partly responsible for this suppression eect and that most probably some unknown chemical species is involved. Detailed studies of this nature are required especially as more interest is being shown in speciation studies. The determination of Si in organosilicon species was studied by Durfee135 after in situ oxidation of the species which is known to reduce volatility and improve the signal response. It was found that the Si signal was dependent upon the relative molecular mass and ligand identity of the various species and this appears to be the first report of the Si signal being dependent upon the organosilicon species.Low relative molecular mass species tended to give a lower response suggesting that volatility was still a problem. Akman and Doner336,337 have continued to investigate interference eects using dual cavity platforms. In these latest studies the nickel chloride interferences on Co and Zn336 and magnesium chloride interferences on Zn337 were examined. As was found in their previous studies,1 the major interference is due to the formation of volatile analyte chlorides and depends upon the pyrolysis temperature. In the presence of excess of nickel chloride above the temperature (approximately 600 °C) at which thermal hydrolysis occurs and HCl is formed analyte chlorides are generated both in the condensed phase and by reaction with the HCl vapour and are lost during pyrolysis or at the beginning of the atomization step.At low pyrolysis temperatures where nickel chloride is not completely hydrolysed the loss in signal response is considered to be due to the expulsion of analyte together with the rapidly expanding matrix gases and/or gas-phase reactions during the atomization step. In the presence of magnesium chloride,337 a similar assessment is made. At high pyrolysis temperatures Zn is lost as zinc chloride formed from the hydrolysis of magnesium chloride at low temperatures and analyte expulsion is considered the predominant loss mechanism along with gas-phase reactions between Zn and chlorine atoms.1.2.4.1. Spectral interferences. Heitmann et al.338 discussed the use of an experimental high-resolution double echelle spectrometer coupled with a linear CCD array that produced a monochromator with an extremely high spectral resolution of approximately 300 000 which enabled the separation of atomic line splittings that lie in the range of a few pm for magnetic field strengths of up to 1 Tesla. This was applied to the study of unexpected spectral interferences or unspecific background radiation with Zeeman-eect background correction systems. For clarification of the spectral processes during atomization the need for highly resolved measurements of analytical spectra is essential.In addition the inclusion of a continuum source allows for the possibility of simultaneous recording of the absorption spectrum at and also in the vicinity of the absorption wavelength. This enables a detailed observation of line broadening and shifting eects as well as the appearance of structured background signals. With this arrangement the splittings of analytical lines in a magnetic field were measured and found to be in good agreement with theory. It is to be hoped that this work will be published in the near future as these investigations may go some way to explaining those rare instances where Zeeman-eect background correction appears to suer from spectral interferences. Shepard et al.339 compared the use of continuum source and Smith–Hieftje background correction for electrothermal atomization molecular absorption spectrometry.The latter was found to be superior for the correction of large background levels obtained from the aluminium chloride molecule used for the determination of chloride. 1.2.4.2. Chemical modifiers—general. Once again a large number of reports concerned with and making use of a range of chemical modifiers appeared in the period covered by this Update. Many of the reports are applications based and consequently are discussed more fully in the appropriate application-related Update. The work concerning chemical modifiers is listed in Table 3. One trend that appeared in last year’s review1 is still apparent this year. A number of groups are working on ‘permanent’ modification of graphite atomizers by applying a metallic coating to the furnace or platform.This topic is covered in Section 1.2.1.2 of this review with no new trends to report. In two conference presentations340,341 Volynsky described a catalytic approach to chemical modification. The catalytic approach refers to the acceleration of the reduction of inorganic compounds or the thermal dissociation of organometallic compounds and hydrides in the graphite furnace due to the use of chemical modifiers. It was proposed that the similarities in the mechanism of modification observed for platinum group metals added as solutions and for transition metal carbide coatings on graphite tubes is due to their similarity in catalytic properties.Considering the modifying action of the platinum group metals on the atomization of Se it was found that the losses of Se in the pyrolysis step were proportional to the melting point temperature of the modifier used. The diering properties of the platinum group metal chlorides were also shown to be the source of dierences for the modifying behaviour of these chlorides. This appears to be an interesting approach not unlike that considered by Tsalev and co-workers111 and it is to be hoped that this work will be published. 1.2.4.3. Chemical modifiers—palladium. Palladium is still the most commonly reported chemical modifier used either alone or in combination with an increasingly large range of compounds. There are conflicting theories as to the mechanism of the action of palladium see also Section 1.2.3.Fundamental processes. Jackson et al.342 have continued their work on the proposed heterogeneous mechanism and in this report used wall-to-platform vaporization to study the mechanism of chloride interference on thallium in the presence of palladium. This technique involves the deposition of sample (analyte in a sodium chloride matrix) onto the graphite tube wall followed by the drying stage of the furnace program. A graphite platform is then inserted into the graphite atomizer and pretreated with palladium solution. During the subsequent pyrolysis and atomization stages the analyte is selectively vaporized from the wall condensed on the palladium-coated platform and re-vaporized.The results of this work appear to substantiate the step-wise mechanism of modification by palladium described previously by the same worker.343 Using wall-toplatform vaporization with palladium as chemical modifier interference-free determination of several metals in the presence of up to 700 mg chloride were made. The results from a number of studies comparing the performance of palladium with and without various reducing agents have appeared in the literature.344–351 In one such study concerned with the determination of arsenic in sea-water Bermejo-Barrera et al.349 found that a mixed modifier of palladium nitrate and zirconium oxychloride provided the best stabilizing eect and quantitative recovery of arsenic in seawater. The same group compared chemical modifiers for the reduction of interferences in the determination of mercury in sea-water.350 It was found that although a palladium–ascorbic acid mixture provided the best detection limit the best %RSD was obtained with a palladium–citric acid mixture.Ni et al.346 have continued their investigations into the eect of high sulfate concentrations in samples such as sea-water by ETAAS. The eect on the determination of Pb was studied.346 Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 Various chemical modifiers; lanthanum palladium alone palladium and magnesium nitrate mixture and palladium and strontium nitrate mixture were investigated. The mixture of palladium and strontium nitrate was capable of stabilizing Pb to the highest pyrolysis temperature of 1300 °C.This temperature permitted the determination of 0.5 ng of lead in the presence 200 mg of sulfate. In the recent literature on the use of palladium as a chemical modifier there are no major trends and there is little consensus on which combination makes the ‘ideal’ modifier. The only general theme running through the research this year is the increasing use of palladium-containing mixed chemical modi- fiers with dierent combinations providing the best performance depending on the analyte and matrix studied. 1.2.4.4. Other chemical modifiers. Several organic chemical modifiers have been applied either alone or in conjunction with a metal. Ascorbic acid citric acid and sucrose have been popular as chemical modifiers due to the formation of active carbon species and reductive gases such as H2 CH4, CO and CO2.An investigation of the ability of a number of organic acids to reduce chloride interference in the determination of Cd Pb and Zn was undertaken by Kantor.359 The eectiveness of the chemical modifier was found to be dependent on the chloride type present. For example when MgCl2 was present ascorbic and oxalic acids were equally eective but when CaCl2 or NaCl were present oxalic acid was the preferred modifier. In a Chinese publication Li et al.360 compared the eectiveness of several organic modifiers to reduce sea-water interferences in the determination of Cd Co Cr Cu Mn Mo Ni and Pb. Recoveries of 95–107% were reported although no specific modifier was recommended in the abstract.Russian workers361 compared organic nitrogen-containing modifiers for the determination of Ni. The thermal stability of the chemical modifier itself and that of the analyte–modifier complex were reported to be the most significant factors in the determining the eectiveness of the chemical modifier. However no specific chemical modifiers are identified in the abstract. A range of metals (the nitrates of magnesium calcium strontium scandium yttrium and lanthanum the metals rhenium palladium rhodium platinum and ruthenium the compounds sodium tungstate and zirconium oxychloride and the mixtures rhodium–magnesium nitrate and platinum–magnesium nitrate) were compared as chemical modifiers for the determination of Cr.357 The chemical modifiers that gave the best results in terms of thermal stabilization and sensitivity were magnesium rhodium and platinum.The determination of Cr in serum samples was reported to require calibration by the method of standard additions regardless of the modifier used. Again no major trends or conclusions can be drawn from the work on chemical modifiers reviewed this year but mixed chemical modifiers are being increasingly suggested to solve particular problems in individual determinations. 1.2.5. Developments in technique Gilmutdinov and co-workers386–388 have continued their work on spatially resolved detection of analytical signals in ETAAS. The conference reports discussed last year1 have now been published.389,390 Investigating and modelling the threedimensional distributions of free atoms and condensed particles inside a tubular electrothermal atomizer,387 and the threedimensional distribution of radiant intensity from HCLs and EDLs389 has been the stimulus to the development and proposal of spatially resolved detection of analytical signals in ETAAS.390 This latter paper is to be recommended as it is a concise summary of spatial non-uniformity in the gas phase analyte distribution atomizer gas phase temperature and radiant 351R Table 3 Chemical modifiers Chemical modifier Ascorbic acid and nickel Ascorbic acid and sucrose Ascorbic acid magnesium nitrate platinum and rhodium Calcium nitrate Europium Lanthanum Nitrates of Ca La Mg Sc and Sr compounds Na2WO4 and ZrOCl2 single modifiers of Pd Pt Re Rh Ru and mixed modifiers Pt–Mg(NO3)2 and Ru–Mg(NO3)2 Magnesium nitrate Organic chemical modifiers Organic chemical modifiers Organic nitrogen containing chemical modifiers Rhodium Palladium Inorganic and organic forms of palladium lecithin and niobium Palladium plus NaPH2O2 or NaBH4 352R Journal of Analytical Atomic Spectrometry August 1997 Vol.12 Element Various (8) In V Si Yb P Cr Be Cd Pb Zn Various (8) Ni Au Bi Cd Ge Sb Sn Matrix Geochemical — Water and oil samples Potassium dihydrogen phosphate Mixture of rare earth oxides Organic soils Rain water and serum Water Chloride matrices Sea-water — Cobalt oxide catalyst Water Edible oils and fats High-purity iron Reference Sample treatment/comments 352 Extraction used to improved LODs for As Sb and Te as diethyl dithiophosphates.Ni in extracts acts as chemical modifier. Ascorbic acid preferred modifier in overcoming interference due to boric acid. Cd Co Cu Ni and Pb were determined after sorbent extraction with sorbent analysed directly as slurry. 308 Addition of mixed modifier enhanced sensitivity with an unresolved double peak. Thermal pre-treatment of atomizer with mixed modifier was recommended. See Section 1.2.3. Fundamental processes. 353 Comparison of chemical modifiers. Ascorbic acid removed chloride interference.Small enhancements in sensitivity with no improvement in LOD observed. Modifer used to prevent carbide formation. 354 355 Interference-free determinations were possible. Mechanism of Yb atomization from Yb–Eu system studied using powder XRD. 356 357 Samples extracted with NaOH. LODs 0.18 0.091 and 0.14 mg l-1 (20 ml sample volume) with Mg Pt and Rh modifiers respectively. Pt preferred modifier for serum analysis; though calibration required matrix matching with serum samples no modifier was required for rain water analysis. 358 359 Addition of modifier enhanced sensitivity by a factor of 2–3 LOD 27.1 mg l-1 sample volume 20 ml. Investigation of the use of organic modifiers to overcome chloride interferences.Ascorbic or oxalic acid removed interferences from matrices based on MgCl2. Oxalic acid recommended for matrices based on CaCl2 or NaCl. Mechanisms of modification studied see Section 1.2.4. 360 Investigation of the use of organic modifiers to overcome sea-water matrix eects. LODs for Cd Cr Cu Mn and Pb were in the range 0.02–0.2 mg l-1 with recoveries of 95–107%. 361 Study of the factors aecting the eciency of organic nitrogen-containing modifiers. Thermal stability of modifier and that of analyte–modifier complex considered to be most important factors. 362 Addition of modifier enhanced sensitivity and provided freedom from interference. Direct analysis with chemical modification proved to be a more rapid method than extracting analyte from matrix using IBMK 363 Addition of 5 mg l-1 Pd stabilized Bi to 1300 °C.Sensitivity and reproducibility also enhanced. 364 Chemical modifiers compared and evaluated based on maximum pyrolysis temperature possible sensitivity background absorption and atomization signal shape. Palladium chloride was the preferred modifier. Samples analysed directly from sample tray heated to 60 °C. RSD 3% for 6–10 ng g-1 recoveries of 101±12% found. 365 3 Samples acid digested pH adjusted and sealed in PTFE beakers with 0.6 g l-1 Pd–NaPH2O2 (for Ge and Sb) or NaBH2 (for Ge and Sn). Palladium-containing precipitate was collected dissolved in 0.66 mol l-1 tartaric acid –HNO –HCl diluted with H2O and analysed.Sensitivities enhanced by factors of 1.5 3.7 and 4.5 for Ge Sb and Sn respectively. Table 3 (Continued) Chemical modifier Palladium Palladium Palladium Palladium and ascorbic acid Palladium and ascorbic acid Palladium–nickel Palladium–nickel Palladium–magnesium nitrate Palladium nitrate–magnesium nitrate Palladium–rhodium Palladium and tartaric acid Palladium lanthanum or silver Mixed modifiers of palladium– magnesium nitrate and iron chloride–magnesium nitrate Palladium nitrate and magnesium nitrate Palladium–magnesium or nickel Palladium and palladium–1% NH4OH Mixed palladium and hydroxyammonium chloride Mixed palladium–tungsten and palladium–tungsten– ammonium nitrate Element I V Various Ag Se Bi Pb Sb Se Cd Ni Pb Si Sr Cd As Ge As Ge Se Various (6) Matrix Tap water Serum Sodium chloride matrices Sea-water Food samples Ni-based alloys Ni-based alloys Serum Plant and animal SRMs Serum Environmental samples Ni-based alloys — — Seafood RMs Water Biological food and clinical samples Biological samples Journal of Analytical Atomic Spectrometry August 1997 Vol.12 Reference Sample treatment/comments 366 I determined as the iodide– mercury(II)–phenanthroline complex. Addition of Pd had no stabilizing eect therefore no chemical modifier was required. 367 20 ml 2.76 mmol l-1 Pd added to sample. LOD 11 ng l-1 240 ml sample volume used (6×40 ml injections) RSD (n=10) 5.5% for 1.5 mg l-1 V recoveries 99–103%.The benefits of a palladium modifier in this determination of V were not given. 342 Wall-to-platform vaporization used. Addition of modifier allowed interference free determination of several metals in the presence of up to 700 mg chloride. 369 Mixed modifier used. LOD 0.1 mg l-1 20 ml sample volume used recoveries from seawater 95.8%. RSD (n=4) 1.4–3.4% for 0.4–11.7 mg l-1 Ag. See also references 349 350 368. 370 371 Use of mixed modifier enhanced sensitivity. Determination possible without prior separation of Bi. RSD 3–5.9% for 0.3–0.7% Bi. Recoveries 87–119%. 372 Determination possible without prior separation of Pb and Sb from matrix.RSD 6.7–7.6% for Pb and 4.7–7.5% for Sb. Recoveries 94–103% for Pb and 95–110% for Sb. 373 Comparison of ETAAS with chemical modification against FI–HGAAS. Best LOD 6.5 mg l-1 obtained with ETAAS for 100 ml serum volume. 374 Conditions for simultaneous multi-element analysis optimized pyrolysis temperature of 350 450 °C and atomization temperature of 2100 °C used. Samples prepared by microwave digestion. 375 5 mg Pd+3 mg Rh in 10 ml injected into THGA coating lasted up to 60 firings. Pd–Rh reduced formation of oxides and carbides and provided freedom from interferences. Addition of rhodium stabilized Si to higher temperature up to 1600 °C than with palladium alone and a four-fold increase in sensitivity obtained. LOD 40 pg 20 ml sample volume.Recoveries 85–110%. Mixed chemical modifier proposed. 376 377 Investigation of the eciency of three modifiers in reducing matrix and foreign ion interferences. 378 Performance was found to be similar for the two mixed modifiers. The formation of a stable ferric arsenide was proposed as the mechanism of modification. 379 XPS suggested formation of Ge–Pd and Ge–Mg bonds. Intermetallic compounds and germanates detected by XRD. See Section 1.2.3. Fundamental processes. 380 381 Pd–Mg modifier recommended. Addition of 10 mg ml-1 Pd stabilized Ge to 1000 °C. Sensitivity was not enhanced. Pd–NH4OH was preferred modifier. See also ref. 363. 383 Comparison of hydroxyammonium chloride and ascorbic acid as reducing agents for palladium.Hydroxyammonium chloride preferred for sample types studied. See also ref. 382. 384 Tissue samples were acid digested hair solubilized in TEAH and urine diluted. Palladium–tungsten–ammonium nitrate used for urine analysis. LODs in urine 14 1.2 and 24 pg for Bi Cd and Pb respectively with 20 ml sample volumes. 353R Table 3 (Continued) Chemical modifier Palladium magnesium nitrate mixed palladium–magnesium nitrate and palladium mixed with ascorbic acid hydroxyammonium chloride or citric acid Palladium lanthanum palladium–magnesium nitrate palladium–strontium nitrate Palladium iron nickel and the mixtures of each with magnesium nitrate Palladium and mixed modifiers of palladium–tungsten palladium–tungsten–citric acid palladium–zirconium and palladium–zirconium– citric acid Palladium nitrate palladium and ascorbic acid lanthanum chloride magnesium nitrate silver nitrate and zirconium oxychloride PdCl2 NH4H2PO4 NH4F (NH4)2SO4 H2PtCl6 and H PtCl6+organic acids 2 Palladium chloride–magnesium nitrate palladium chloride– hydroxyammonium chloride and magnesium nitrate Palladium chloride palladium nitrate ammonium dioxalatopalladate(II ) colloidal palladium and a mixture of palladium nitrate and diammonium hydrogencitrate Organopalladium intensity distributions and their eect on analyte detection.The conventional detection of analyte atoms based on the use of PMTs provides excellent temporal resolution and sucient wavelength isolation but totally ignores the spatial aspects of the interaction of the probing radiation beam with the analyte in the atomizer.Using a monochromator equipped with a linear photodiode array located vertically along the exit slit the temporally and spatially resolved detection of Cd atomization and sodium chloride vaporization was investigated. It was shown that severe non-uniformity in atomic and/or background absorbance may be a potential source of analytical error especially if the gas phase analyte distribution generated from the atomization of an aqueous reference solution and that from a sample were significantly dierent. A further claimed advantage of the spatially resolved approach is the possibility of extending the dynamic range by using the regions of the atomizer where the analyte concentration is highest for 354R Journal of Analytical Atomic Spectrometry August 1997 Vol.12 Matrix Element Sea-water Hg Sea-water Pb — Se Water Various (7) Sea-water As Biomaterials Cd Sea-water Cr Mn Mo Se Sodium chloride matrices Biological samples Hg Se Reference Sample treatment/comments 350 Comparison of the three modifiers. Best LOD obtained using palladium–ascorbic acid. Best RSD of 2.3% obtained with palladium–citric acid. Recoveries close to 100% obtained with reduced palladium and magnesium nitrate. See also ref. 349. 346 Comparison of modifier eciency in overcoming interference eect due to sulfate. Palladium–strontium nitrate preferred.Recoveries of 94–103% from spiked sea-water were obtained. 351 Investigation of the eciency of the modifiers to overcome chloride interference eects. Eects of stable chlorides could be overcome using metals alone; magnesium nitrate was required when less stable chlorides were present. 344 Comparison of the five modifiers each containing 2 mg Pd. Modifier mechanisms investigated using SEM and electron microprobe analysis. Interference eects of ten compounds were investigated. Interference by Mo Si Ti and V could not be eliminated. 349 Comparison of the six modifiers. Palladium nitrate and zirconium oxychloride either alone or mixed with magnesium nitrate or reducing agent provided best recoveries and LODs in the range 1.1–3 mg l-1 20 ml sample volume.348 6 2 Comparison of modifiers. H PtCl was preferred as it showed the strongest modifying ability and provided recoveries of 90–115%. 347 2 Best modifier for Cr was PdCl2–magnesium nitrate or PdCl –hydroxyammonium for Mn magnesium nitrate and for Mo providing an LOD of 0.3 mg l-1 with a 30 ml sample volume palladium nitrate. Method of standard additions was required for Cr and Mn determinations. 345 Study of the eciency of the various forms of palladium to overcome interferences due to sodium chloride. Pre-reduced palladium was found to be most ecient modifier. 385 2 Three organopalladium compounds [including Pd(acetylacetonate) and PdCl2(PhCN)2] were evaluated as modifiers in the determination of organomercury and organoselenium compounds.LODs (20 ml sample volume) 0.25 and 0.12 ng for Hg and Se respectively low mass measurements and the atomizer parts where the analyte concentration is lowest for measurements of high masses. In addition simultaneous monitoring of tube emission provides the opportunity for exact furnace alignment and for on-line monitoring of the wall and platform temperatures. Similarly detection of the whole intensity profile of the radiation from the primary source could be useful in source alignment. Belarra et al.391 discussed the aim of extending the working concentration range for Sn in graphite ETAAS by studying several non-resonance lines for this element and with atomization from either a platform or the graphite tube wall.A number of parameters were assessed such as LOD LOQ characteristic mass and either measuring peak or integrated absorbance along with the respective %RSDs. The use of the non-resonance lines enabled the working range to be extended up to 140-fold greater than the upper limit for the most sensitive line without degrading precision. This work appears to have been performed solely with aqueous solutions although in the conclusion it was indicated that at high levels wall atomization is required which may have an adverse eect through the presence of interference eects on real samples. A high current power supply for ETAAS has been described by Torsi et al.392 consisting of two high current batteries and associated electronics and software.The power delivered can be changed from 0 to 100% in 1% steps which permits both the temperature at low power and the heating rate to be controlled. Using the system absorption signals for Ag Pb Tl and Zn have been recorded. Simultaneous multi-element detection continues to receive considerable attention although for the period of this Update mainly in conference presentations. It is to be hoped that some of these will be published in due course. Russian workers393 developed a simultaneous multi-element AA spectrometer with a continuum source to identify multi-component alloys by ETAAS. The system allows up to 12 elements to be monitored simultaneously in one sample. In general it seems that solid sampling was used and calibration achieved by the use of CRMs.Levine et al.394 appear to have constructed a multielement ETAAS system using an ICP as the light source. This has been discussed previously.72 An aqueous solution containing a mixture of elements with concentrations in the range 10–10 000 g l-1 was aspirated into the ICP and the plasma emission focused through the electrothermal atomizer. The unabsorbed radiation exiting the atomizer was focused onto the entrance slit of a small monochromator and detected with a linear photodiode array. Background correction was achieved by monitoring the absorbance at a second plasma emission line in the vicinity generally 5 nm of each analytical line which was either a weakly absorbing wavelength emitted by the analyte or an emission line from a dierent element.The system continuously monitored a 90 nm spectral window and was characterized using two dierent spectral windows for monitoring eight biologically important elements (Al As Cd Cu Pb Sb Se and Tl). These workers claimed analytical figures of merit similar to those reported for Zeeman-eect background corrected ETAAS. While this approach may be easier for any research laboratory to assemble while not having access to high intensity continuum source lamps and associated power supplies it does seem to require an additional level of complexity given the running requirements of an ICP. In addition ‘near line’ background correction measurements are acceptable when simple broad-band background is present but are unable to compensate for the complex interferences found in electrothermal atomization measurements.Demers and Almeida395 described the design and operation of the Leeman Labs Analyte 5 multi-element ETAAS system. An alternative approach to achieving simultaneous multielement determinations with ETAAS was described by Edel et al.,396 whereby frequency modulation of multiple light sources was used to allow Cd Cr Cu Fe Mn and Pb to be measured simultaneously in digested sewage sludges and sediments. Deuterium arc continuum source background correction was employed. Measurements were compared with those obtained on a conventional single-element ETAAS instrument with a transversely heated graphite atomizer and longitudinal Zeeman-eect background correction.A large number of conference reports have been presented by those working with the Perkin-Elmer simultaneous multielement ETAAS instrument. White et al.397 decribed the development of methods for the simultaneous determination of Cd and Pb in whole blood Cr and Ni in serum and Co and Mn in urine. From the same group Iversen et al.398 described the application of simplex optimization to establish the electrothermal atomization conditions for simultaneous determination of Co and Mn in urine. In all cases the methods have been Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 applied to the determination of these elements on nonoccupationally exposed populations within the European Community.Schlemmer et al.399 and Carnrick et al.400 discussed the principles of method development for simultaneous multi-element ETAAS measurements. In determining up to 6 elements simultaneously common atomizer conditions and chemical modifiers are required compared with the traditional approach which focused on the individual optimization of parameters for each element. Schlemmer et al.401 and Shuttler et al.402 showed that modern atomizer and instrument developments allow the attainment of simultaneous multi-element ETAAS measurements without major compromises on the accuracy precision and detection limits obtained compared with conventional single-element determinations. The amount of work related to furnace atomization plasma excitation spectrometry (FAPES) varies from year to year.In this Update period there seems to have been a resurgence though all in the form of conference presentations. Sturgeon403 reviewed the characteristics of the source and FAPES for diagnostic purposes within a graphite atomizer to investigate vaporization and atomization processes. The potential for multi-element trace analysis of both metals and non-metals at the pg level with ml sample volumes has been demonstrated and the absolute LODs rival those of conventional graphite ETAAS. Pavski et al.404 have continued their work on imaging the atomic and molecular species in the FAPES source with a CCD camera. Because analyte atomization takes place within a conventional graphite electrothermal atomizer the atomization mechanisms underlying the formation of FAPES emission transients are in general similar to those of graphite ETAAS.However significant dierences do exist particularly in that analyte appearance temperatures are sometimes less than those observed in ETAAS implying significant plasma-induced dissociation of molecular oxide species. By using a CCD camera two-dimensional distributions of analyte atoms and oxide species in the FAPES were obtained in an eort to gain further understanding of the atomization mechanisms taking place. Le Blanc and Blades405 continued to investigate the FAPES source and measured emission intensities from OH N2+ In Mg and Pb. The results showed that there is a steep thermal gradient in the source.Spatial and temporally resolved In emission profiles were measured to gain a better understanding of the plasma discharge and to quantify analyte spatial distributions within the source. These measurements showed that there is a significant degree of inhomogeneity in the FAPES source related to excitation rather than the distribution of the source atoms. From the same group Hettipathirana and Blades406 developed a FAPES system to allow simultaneous measurement of AA and emission signals. The Pb excitation temperature of this plasma source operating at 50 W is about 4500 K. Time resolved Pb excitation temperatures showed a relatively stable plasma excitation environment during the analyte atomization and excitation. Simultaneously measured AA and emission signals also showed that the temporal profile of the Pb emission signal was determined by the atomization and vaporization characteristics of the analyte rather than by excitation characteristics.Little additional work is reported in this Update on standardless or absolute analysis. The conference presentation from L’vov reviewing the status of absolute analysis and discussed in the last Update1 has now been published.407 Yang and Ni408 investigated the possibility of standardless determination of Au in geological samples. A simplex optimization procedure was employed to develop the atomizer time–temperature conditions. The atomization eciency for Au was found to be temperature independent and approach 100% at higher atomization temperatures.The standardless determination of Au in geological RMs using palladium as chemical modifier showed an accuracy of better than 7% on applying the experimental 355R characteristic mass obtained as a result of the simplex optimization procedure. Other Chinese workers409 found that a comparison of experimental versus theoretical characteristic mass values for Ag Al Bi Cd Cu Fe Hg Mn Ni Pb Sn and V under STPF conditions gave a ratio of two for the atomization eciency. This is in contrast to other workers in this field who have generally found ratios of approximately one,111 although the reason for this discrepancy is unclear. These elements were determined in human hair and peach leaf RMs using the experimental characteristic mass values and the results obtained were found to be in good agreement with the certified values.1.3. Chemical Vapour Generation The research literature concerned with chemical vapour generation (CVG) continues to grow. It is often dicult to draw a boundary between papers which are reporting some development in technique or methodology and those which are ‘merely’ application papers. Thus the coverage of the literature is relatively more comprehensive in this section than in some other sections of this Update. Although the main structure of this Update is to classify work according to the spectroscopic technique determinations by AAS and AFS are both included in this section (unless the final step is ETA in which case the work is reviewed in Section 1.2.2.3).It should be borne in mind that many of the developments reported for CVG in conjunction with ICP optical emission spectrometry (OES) and ICP-MS (which are covered in other Updates) could probably be implemented with either AAS or AFS and thus for a full picture of recent developments in the field of CVG for atomic spectrometry these other Updates should be consulted. Most of the published work relates to the determination of As Se and Hg. There is a sustained interest in the development of methods involving CVG which give information about the speciation of the analytes. Such studies often involve the coupling of chromatographic separation with element specific detection. Last year the determination of Cd by a CV procedure was reported,1 and this year saw the report of the generation of a volatile Cu derivative on the addition of borohydride,410 although the detection method was ICP-OES.The generation of hydrogen selenide from an anion-exchange resin on which SeVI and borohydride had been co-immobilized was reported.411 Most of the work on determination of the total hydride-forming element (or mercury) appears to have been motivated by the need to achieve either an improved LOD or greater freedom from interferences (or both). 1.3.1. Hydride generation A number of review articles concerning hydride generation have appeared412,413 which cover the use of the technique with all atomic spectrometries. Flow injection HG–AAS was discussed46,414 in two review articles by Fang and co-workers.The capability of FI to discriminate kinetically against interferences by transition metals was highlighted as well as the variety of preconcentration chemistries (such as co-precipitation or solid phase extraction) that may be linked with the HG technique. Howard gave an overview presentation415,416 of the use of HG in speciation studies of As Sb and Se. 1.3.1.1. General studies of fundamentals techniques and instrumentation. In a review417 of electrolytic sample pretreatment Beinrohr covered electrochemical HG. Hueber and Winefordner reported on an extensive study of electrochemical generation of As Sb and Se for detection by FAAS and ICPOES. A variety of cathode materials (lead palladium platinum and silver) were evaluated.Palladium platinum and silver gave comparable sensitivity for arsine generation lead was considered the best for the generation of stibine and palladium 356R Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 was the best for hydrogen selenide. It was possible to generate arsine and hydrogen selenide simultaneously. No detection limits for AAS were given but those for ICP-OES were rather disappointing (0.7 and 3 mg l-1 for As and Se respectively). The conversion eciency of the cell was 50–98% depending on the flow rate. The eect of several possible interfering species were investigated. Ding and Sturgeon discussed418 the interference of copper and nickel on the electrochemical generation of hydrides at a variety of cathodes including platinum carbon and lead.Detection was by ETAAS. In the introduction to this paper seven previous references to electrochemical HG are cited. Ridgway419 reported that for the generation of arsine both gold and palladium electrodes were better than those made of nickel or platinum which gave a higher hydrogen to hydride ratio. Schaumloeel and Neidhart420 used electrochemical generation of arsine with o-line reduction of AsV to AsIII with Brindle’s reagent (L-cysteine). Several other groups of workers have used this reagent421–424 for pre-reduction sensitivity enhancement and interference control. Mierzwa and Dobrowolski425 have provided a further example of the generation of arsine directly from a slurry sample. Recoveries from some sediment SRMs were enhanced by treatment with microwaves followed by ultrasound.Bye and co-workers426 showed that matrix elements including cobalt chromium nickel and iron were transported from the gas– liquid separator to the atom cell and concluded that one possible mechanism for interference by these elements was the inactivation of sites in the quartz tube atomizer involved in the radical decomposition reactions which eventually lead to the formation of free analyte atoms. The mechanism of transport was thought to be by aerosol droplets as the transport was reduced when a glass frit was inserted into the transfer line. It was pointed out that such a frit rapidly saturated with liquid and needed frequent cleaning. The same authors used generation from alkaline medium427 as part of the strategy for overcoming the interference of cobalt copper and nickel on the determination of As and Te.The cobalt and nickel interferences were removed by the addition of DTPA (diethylenetriaminepentaacetic acid) and the solution made alkaline by the addition of sodium hydroxide followed by the addition of borohydride. The precipitated copper hydroxide was removed by filtration and the hydrides generated by mixing with an acid solution in a continuous flow system. Both AAS and ICP-OES were used. Vuchkova and Ardjapan428 showed it was possible to generate the hydrides of As Bi Se and Sn from solutions of their dithiocarbamate complexes in methanol on the addition of borohydride. Mercury vapour was also released from HgII but no signal for Pb was obtained.The detection technique was ICP-OES. Chinese workers reported429 on ‘a new hydride generation and sample introduction technique’ which they described as intermittent flow being a hybrid of FI and continuous flow. The authors claimed less cross-contamination for the new system although it is not clear whether air-bubbles were used to separate samples. There is continued interest in coupling 2 microwave-assisted sample pretreatment with HG procedures. 430–433 The element of interest currently appears to be Se. Procedures have been developed for the determination of Se in a variety of matrices but as yet the decomposition of trimethylselenonium ion (the major metabolite in human urine) in a flow-through microwave system has not been achieved.Siska et al.434 improved the sensitivity and LOD of an FI–continuous flow system by incorporating a cryotrap between the gas–liquid separator and the atom cell. Water was removed by a magnesium perchlorate dryer and on warming the hydrides were transported in a He–1% O stream to the quartz tube atomizer. The LODs claimed were 10–20 ng l-1. The authors cite the use of high purity reagents and the use 2). Detection limits between 26 pg (Te) and to this paper contains a useful collection of references to of a closed system as the reason for the ‘surprisingly’ low blank co-workers434,447 were able to achieve LODs of 10–20 ng l-1 values. Zang and co-workers435 described a low pressure for As and Se.The methods were applied to a variety of CRMs atomizer for the determination of As Bi Sb Se Sn and Te. and foodstus containing the analytes at concentrations of The hydrides were formed and retained in a reaction vessel only a few ng g-1. To overcome the interferences from transheld at below atmospheric pressure. A metered volume of air ition elements Wickstrom et al.427 added diethylenetriamine- (1–5 ml depending on the element) was then admitted and the pentaacetic acid (DTPA) to overcome the eect of cobalt and mixture drawn into the evacuated quartz tube atomizer which nickel. The interference from copper was overcome by precipiwas electrically heated. It was found that the sensitivity was tation of copper hydroxide. Generation from alkaline solution enhanced by the presence of O2 (by a factor of approximately was used and the procedure was applied to the determination 2 for all elements except Sn for which the signal was negligible of As and Te in some nickel and copper alloys.The introduction in the absence of O 610 pg (As) for a sample volume of 1–2 ml were obtained. Ellis and Tyson found436 that the addition of surfactant (didodecyldimethylammonium bromide) had no eect on the HG determination of As using a Perkin-Elmer FIAS system with flameheated quartz tube atomization. The Varian ‘SIPS’ system has been adapted for HG.437 Castillo and co-workers reported438 on a simple gas–liquid separator. The device was made from The reduction by L-cysteine was used448 as the basis of an a 5 ml plastic syringe and was configured with the inlet and FI speciation procedure for various water samples.After drain tubes in the base and the outlet at the top. The drain treatment of the sample with Brindle’s reagent the As detertube was mounted so that the separator always contained a mined was considered to be ‘total’ and the As determined after certain volume of liquid. The gas flow used was apparently the addition of 0.166 M acetic acid was considered to be ‘AsIII’. 0.12 ml min-1 a value which is so low as to suggest that there is a typographical error in the paper and the flow should be 120 ml min-1. An improvement in LOD for Se by a factor of 3 was reported although the final value of 6.5 mg l-1 is still high for an FI–HG–AAS system. 1.3.1.2. Determination of individual elements.In general every paper dealing with the determination of an element by hydride generation has three sections of interest. These sections are (a) how the sample was pre-treated in order to release an appropriate inorganic precursor (b) how potential interference eects (including analyte oxidation state eects) were overcome and (c) how the hydride was generated separated transported and atomized. As not all forms of a particular element react to form hydrides speciation may be based on the basic generation reaction and a number of publications have described this procedure as a basis for speciation. Increasingly chromatography is being used to separate analyte compounds followed by HG as the interface with element specific detection.Microwave digestion of seafood for the determination of As and Se has been described.439 The author suggested that large amounts of nitric acid commonly used in microwave digestion procedures can interfere with hydride formation. Indian researchers440 accurately determined As and Se in a coal flyash SRM using a method which involved microwave digestion. They also examined a sodium hydroxide fusion dissolution method. Spanish workers441 determined As and Se in soils and plants using a procedure involving Kjeldahl digestion and Japanese workers442 devised a procedure for the determination of the same elements (in biological CRMs) in which the interferences from transition metals were removed by passing the sample digest through a column of Chelex-100.An FI system was used with a Pyrex glass mixing coil (16 cm×2 mm id) although the hydride was collected and dehydrated before being swept to the atomizer. Italian researchers443 determined As Bi Sb and Se in steels and found that LODs were an order of magnitude lower than those of the corresponding ETAAS determinations. All four of these publications440 –443 are in English. The possible use of a commercial HG system for the determination of some of the toxic elements in toys covered by the European Standard EN 71 Part 3 (As Ba Cd Cr Hg Pb Se and Sb) was discussed.444 Some preliminary data445,446 on the analyses of biological RMs foods and environmental samples by AFS have been presented. The LODs for As and Se were a few ng l-1. This allows dilution to be used as a simple approach to the removal of interference eects.With preconcentration in a cold trap Siska and Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 previous work on (a) the use of HG from alkaline solution and (b) the use of masking agents to overcome interferences. One such masking agent is Brindle’s reagent L-cysteine which will also reduce AsV to AsIII. This was used in a procedure for the determination of As in geological RMs421 and of As by electrochemical HG420 with an LOD of 0.4 mg l-1. An FI microwave digestion procedure for speciating As in human urine as ‘toxic’ and ‘non-toxic’ has been developed by Lopez Gonzalvez et al.431 The basis of the procedure was to find two sets of reaction conditions under which (a) all arsenic species gave a response and (b) inorganic As monomethylarsenic (MMA) and dimethylarsenic (DMA) gave responses but arsenobetaine (AsB) and arsenocholine (AsC) did not.Total As was determined by merging the sample with an alkaline persulfate stream and passing through the microwave digestor. The stream was then acidified and merged with a borohydride stream. For the determination of ‘toxic’ arsenic the experiment was repeated with an acid carrier stream instead of the alkaline persulfate stream. ‘Non-toxic’ arsenic was determined by dierence. The LOD was 6 mg l-1. The authors’ results are at variance with previously reported work in regard to the need for additional reducing agents (such as Brindle’s reagent) to reduce AsV to AsIII.The authors also discussed the eect of acid concentration when ‘using 3% NaBH4’. However in comparing their work with previously reported procedures it would seem important to account for both concentrations of reagents and flow rates as both of these parameters control the concentrations of species in the manifold. Japanese workers449 determined AsIII AsV and organic arsenic in soft drinks by a sequential liquid–liquid extraction procedure. First AsIII was extracted into toluene then AsV was reduced with iodide and extracted and the remaining As in the aqueous layer considered ‘organic As’. Chinese workers450 determined inorganic As MMA and DMA in natural waters by generation of the hydrides from all species in a 100 ml sample followed by trapping on 50 cm of GC stationary phase (3% OV-101 on Chromosorb GAW-DMCS) in a liquid nitrogen cold trap.The trap was allowed to warm up and the hydrides eluted sequentially. To distinguish between AsIII and AsV the generation reaction was performed at dierent acidities. When the sample was buered with a 0.1 M acetate–acetic acid buer only AsIII reacted with borohydride. Detection limits of 4–7 ng l-1 were obtained. A similar procedure was described by Teague and co-workers451 who determined AsIII at pH 6–7 (TRIS buer) and total As at an acidic pH with reduction of the AsV with Brindle’s reagent which also enhanced the signal and allowed determination at low acid concentration. Detection limits between 6 and 24 ng l-1 were obtained which exceed by more than an order of magnitude the expected revised Maximum Contaminant Level for As in drinking water set by the US Environmental Protection Agency.A number of speciation studies based on HPLC separation have been reported. Stummeyer et al.452 separated AsIII AsV MMA and DMA on 357R an ion-exchange column (Hamilton PRP X-100) with phosphate mobile phase adjusted to pH 6 with HCl. The eluted components were merged with 15% KI in 2 M HCl and either heated to 100 °C or irradiated with UV light prior to merging with acid and borohydride. For a 1-ml injection loop the LOD was 500 ng l-1. The post-column decomposition was used because the authors claimed that ‘the formation of volatile arsines from organoarsenicals is quite uneective (sic)’.However other workers appear to be able to generate arsines from MMA and DMA quite satisfactorily.452–454 Cornelis and co-workers have developed procedures for the determination of As species in blood serum.455,454 In the first of these,455 cation-exchange HPLC was used to separate MMA DMA arsenobetaine (AsB the main component in serum) and arsenocholine (AsC). The eluent was merged with argon and alkaline persulfate solution prior to passage through a UV reactor. The segmentation produced by the argon reduced the extra-column broadening and the UV reactor decomposed AsB and AsC which otherwise would not give a response with borohydride. The LODs were around 1 mg l-1. The second procedure454 was designed for the determination of AsIII AsV MMA and DMA and anionexchange HPLC was used (though ion-pairing with tetrabutylammonium hydroxide with chromatography on C18 was also investigated).The eluent was merged with acid (20% HCl) and borohydride (2.0%) for HG. 18 There is some interest in the determination of As species in seafood products453,456–458 Montoro and co-workers453 developed a procedure for the determination of MMA and DMA in cockles anchovy octopus sardine tuna and sole as well as a dogfish muscle CRMby anion-exchange HPLC with gradient elution (step change in phosphate buer composition). The column eluent was merged with acid then borohydride. The analytes were extracted with methanol–water (1+1 v/v). A variety of ‘de-fatting’ procedures both pre- and post-extraction were investigated.The LODs were around 0.2 ng g-1 based on fresh mass. The work has been extended456 to the determination of AsB. Samples were again extracted with methanol– water and the chromatography was very similar but with post-column microwave-aided digestion by alkaline persulfate. A preliminary clean-up through a strong cation-exchanger (Dowex 50W-X8) was employed to remove all As species except AsB and DMA (other species passed through the column and the retained species were eluted with 4 M ammonia) thus the overlap between AsB and AsIII which would otherwise occur was avoided. A method for the determination of up to 11 As compounds,457 based on HPLC at elevated temperature (up to 70 °C) has been developed.The separation was based on ion-pairing with various reagents on a C column. A post-column microwave-assisted digestion with alkaline persulfate was used followed by HG-AFS. The procedure was applied to the analysis of a seaweed food product and of the urine of a volunteer who ingested the material. The fate of arsenosugars ingested was studied and it was found that contrary to previous reports arsenosugars were not excreted unmodified. However the authors were unable to identify all the metabolites though it was clear that DMA was one. Lamble and Hill458 devised a procedure for the determination of AsB MMA DMA and total inorganic As in marine biological tissues (CRM TORT-1 and DORM-1). The species were separated by HPLC on a strong anionexchange column with 0.1 mM potassium sulfate as mobile phase.Post-column microwave-assisted breakdown with alkaline persulfate followed by HG-AAS was used for detection. After the microwave reactor the eluent stream was merged with a stream of nitric acid and L-cysteine (to reduce AsV to AsIII) prior to HG with borohydride. Rubio et al.459 have reviewed post-column photo-decomposition procedures for the determination of AsB and AsC. There has been only one report of work relating to the 358R Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 determination of Cd during this review period. Tyson’s group compared460 the determination of Cd and Pb by hydride generation with their determination by ethylation. The cold vapour Cd species was confirmed.The methods had similar LODs (a few mg l-1) but the reaction with tetraethylborate appeared less susceptible to matrix interferences than the reaction with borohydride. In neither case were surfactants found to have any beneficial eects. Iodine was measured461 by vapour-phase molecular absorption of I2 released on adding concentrated sulfuric acid to the sample (iodized table salt). The light source was a hollow cathode lamp and the absorption at 518 nm (most likely a neon fill-gas line) was measured. The LOD was 0.9 mg l-1. For the determination of Pb Yi462 used the reaction with borohydride in the presence of ferricyanide in a flow system. Tyson’s group460 confirmed that with an FI system ferricyanide is a suitable oxidizing agent (the best of several tested).However Camara and co-workers reported463 that with their FI system the best results were obtained with a mixture of lactic acid and potassium dichromate. The LOD was 1.5 mg l-1 with a sample throughput of 180 h-1. The method was applied to the analysis of sea river tap and well water. The results obtained were statistically indistiguishable from those of an ETAAS method. Hydrogen peroxide has also been used as an oxidant.464 Continuous-flow HG was compared with ETAAS for the determination of Sb.423 It was found that the HG method was five-times more sensitive but took longer due to sample pretreatment. Brindle’s reagent was used to reduce SbV to SbIII and was pronounced superior to the use of KI in that smaller amounts of chemicals could be used and there were no problems from sparingly soluble iodides.In an HG-AFS method465 for the determination of Sb in gold-bearing ore powder KI was used along with thiourea an iron salt EDTA and aqua regia. Increased tolerance to interference from Hg AsV and BiIII in the presence of FeII were reported. Burguera and coworkers466 devised a method for the determination of SbIII and SbV in liver tissue. For the determination of SbIII the slurried sample was digested (microwave-assisted) with 1 M acetic acid whereas for SbV the carrier was a mixture of sulfuric acid potassium iodide and ascorbic acid. The LODs were around 0.1 mg l-1. In the determination of Se in urine by HG-AAS,467 a digestion procedure involving nitric and perchloric acids was used for sample pretreatment.Hydrochloric acid was used to reduce SeVI to SeIV. The LOD was 5 mg l-1 and the results were compared with those obtained by a method involving microwave digestion with nitric acid and vanadium pentoxide. On-line microwave digestion has been used430 to decompose some selenoamino acids in blood and serum prior to determination by FI–HG-AAS. A mixture of nitric and perchloric acids has been used373,468,469 as the first step for the determination in serum with an LOD of 2 mg l-1. Conventional HCl reduction of SeVI to SeIV was used. In the first of these reports,373 a comparison between HG-AAS and ETAAS was made. The authors reported that precision and sensitivity were the same but that ETAAS had a lower LOD 6 mg l-1 compared with 12 mg l-1 for HG-AAS.Various methods for the determination of Se in whole blood have been compared by Mestek et al.470 Although the ICP-MS method was preferred for routine analysis the LOD of 6 mg l-1 was only marginally better than that of HG-AAS (8 mg l-1). For the HG-AAS procedure the same pretreatment as above was used whereas for the ICP-MS method the pretreatment consisted of the addition of Triton X-100 water and dilute nitric acid. A batch procedure has been used471 for the determination of Se in term and preterm infant formulae. The nitric–perchloric digestion was used and the LOD was 0.15 mg l-1. This method was also used472 as the first stage in a method for the determination of Se and Te in ginseng ganoderma and garlic.The values obtained for Se were 200 300 and 300 ng g-1 respectively and the LOD was 0.1 mg l-1 in the solution. Chinese workers473 appear to have duplicated the work of Oey et al.702 in devising a method for the determination of Se in copper with removal of the copper by cation-exchange. Neilsen et al.474 have further developed the procedure of Tao and Hansen475 for the preconcentration of Se by co-precipitation with lanthanum hydroxide so that the precipitation and dissolution were carried out in an FI manifold. The precipitate was collected on the interior walls of a knotted tubular reactor and was dissolved in HCl for detection by HG-AAS. For a sample volume of 14.5 ml the LOD was 4 ng l-1 and the sample throughput was 33 h-1.Sanz-Medel and co-workers433 devised a procedure for the determination of total Se and inorganic Se speciation (SeIV and SeVI) based on reactions in an FI system with a focused microwave digester. Sample volumes of 500 ml containing selenomethionine selenoethionine selenocystine SeIV and SeVI were injected into a water carrier which was merged with 47% HBr and 0.015 M potassium bromate solutions then passed through 4 m of tubing in the microwave cavity. Following passage through a cooling coil a borohydride stream was merged and the hydrogen selenide detected by AAS. Speciation was achieved by the simple expedient of turning the microwave oven o when only SeIV gave a response. The LOD was around 1 mg l-1. A conceptually similar procedure was used by Burguera et al.432 for the determination of SeIV and SeVI in lemon juices and geothermal waters.Microwave heating of the sample in the presence of HCl reduced SeVI to SeIV thus allowing the determination of total Se. Without the on-line reduction only SeIV was determined. Lemon juice was found to contain 5.3 mg l-1 of SeIV and 2.0 mg l-1 of SeVI. The LODs were 1 and 2 mg l-1 for SeIV and SeVI respectively. Sanz- Medel’s group have further developed their method476 by the incorporation of an HPLC column between the injection valve and the confluence point with the HBr stream. The stationary phase was C18 and the mobile phase was 0.1 M ammonium acetate buer (pH 4.5). No ion-paring reagent was used. The method was applied to the speciation of Se in urine with somewhat inconclusive results.At least one unknown Se compound was detected and the fate of trimethylselenonium could not be unambiguously ascertained. A comparison of HG-AAS HG-ICP-OES and HG-ICP-MS as detectors was made and it was found that the order of detection power was MS>AAS>OES; the respective LODs for inorganic Se being 0.2 7 and 30 mg l-1. Lei and Marshall have described477 developments to their quartz thermochemical HG atomizer so that it would operate with either aqueous-based or organicbased HPLC eluent. Several approaches to the speciation of inorganic Se and selenoamino acids were described including separation on a cyanopropyl-bonded phase with acetic acid– ammonium acetate mobile phase.A procedure for the extraction from water or feed-supplement into liquefied phenol gave acceptable performance apart from the recovery of SeVI. In the Proceedings of the 5th International Symposium on the uses of Se and Te Koelbl et al.168 provided a review (51 references) of methods for the speciation of Se in biological and environmental samples. A comparison of detection methods (ETAAS FAAS and ICP-MS) was also made. In the FI–HG-AAS determination of Sn in biological materials Burguera et al.478 used a system with a time-based injection device. The eects of various acids used in the digestion procedure were studied and it was found that a nitric acid–perchloric acid mixture could be tolerated only in the range 0.5–2 M whereas hydrogen peroxide had no eect.The LOD was 2 mg l-1. The same value was obtained by Brindle and co-workers424 for the determination of Sn in steels by FI–HG-AFS for a 500-ml sample. Naturally Brindle’s reagent was used which enhanced the sensitivity (30% increase for 1% Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 L-cysteine) and reduced the interferences from matrix elements particularly that due to nickel. The reagent also made the procedure more robust with respect to the eect of acid concentration. 1.3.2. Mercury by cold vapour generation 2. Workers at Hebei University479 found agreement between the experimentally observed absorbance–time relationship and that derived on the basis of theoretical calculations. Golloch and Goetzen480 have disclosed some preliminary information about a system for the determination of Hg in soils and sludges.Up to 200 mg of sample were heated in an ETV unit and the vapours passed through a solution of stannous chloride. Amalgamation trapping of Hg on a gold-coated gauze was employed followed by thermal desorption and measurement by AAS. An anonymous author481 has described a device in which the sample was dried and burnt in a stream of O The gaseous products were decomposed over a potassium permanganate–cobalt catalyst and the Hg collected by amalgamation with gold. The LOD was 10 pg but the sample mass was not given. Conference presentations482,483,484 by representatives of several instrument companies (PS Analytical Thermo Jarrel Ash Varian) have highlighted developments in dedicated instrument performance.Determinations at a few ng l-1 would now seem to be routinely possible by AAS and if AFS is used,482 the LOD drops to 0.2 ng l-1. The design features of a new gas–liquid separator from CETAC for flow-based determinations have been explained485 The incoming stannous chloride–HCl–Hg solution flows over a convex protrusion around which the carrier gas flows in an upward spiral. In the analysis of digests of biological materials the device showed excellent response time with minimal foaming. Brindle and Sheng486 compared four designs of gas–liquid separator device for AAS determination by sequential injection. A detuned cross-flow nebulizer and spray chamber gave the worst performance; the Perkin-Elmer FIAS device performed the best (though it had to be operated without the water droplet filter).An in-house frit-based device (for which full constructional details are given in the paper) had similar performance but suered from memory eects with high concentrations of Hg. The authors also pointed out that quartz tube atom cells without end windows (as are used for HG measurements) are significantly inferior to cells with end windows. Esholz62 described the application of a device which used tubing clamp valves to simulate a six-port rotary injection valve to the determination of Hg in suspended particulate matter. A considerable number of conference presentations featured methods for the decomposition of sample materials (mainly clinical and biological materials).Grandillo and co-workers487,488 described methods based on acid attack in closed vessels in microwave ovens. Saskai and Pacey489 compared the use of ozone with the permanganate–peroxodisulfate methods. McIntosh and Hanna490 used nitric acid in a closedvessel pressure-controlled microwave oven to digest fish tissue. Schlemmer and Erler491 developed procedures for the analysis of blood urine and wastewater in which on-line microwave digestion was employed. Adeloju and Dhindsa492 considered a nitric–sulfuric acid mixture to be the best of several mineral acid cocktails for the digestion of biological and environmental materials. They found that sulfuric acid increased the response in CV-AAS significantly whereas perchloric acid depressed the response.Nitric and hydrochloric acids had no eect. Hill and coworkers493 used an o-line microwave assisted digestion followed by FI–CV-AAS in which borohydride was used as the reductant. The method was applied to two CRMs and to river sediments and canned tuna. Stockwell and co-workers494 reported on the use of CV-AFS in process analysis. For 50% m/v sodium hydroxide the Hg 359R was generated by the addition of hydroxylamine hydrochloride after oxidation with alkaline permanganate; and for concentrated sulfuric acid an acid permanganate oxidation followed by reduction with SnII was used. Researchers on opposite sides of the globe (Plymouth UK and Hangzhou China) have come up with the same chemistry and spectroscopy (AFS) for the determination of Hg in sea water495,496 The UK group495 used an FI procedure in which the sample was first acidified and then merged with an oxidant stream of potassium bromide– bromate.After heating to 70 °C the excess oxidant was removed by reaction with hydroxylamine hydrochloride and Hg0 generated by the addition of SnII with detection by AFS. The LOD was 20 ng l-1. The procedure provided complete oxidation of humic-bound and organic mercury species. The Chinese group used batch reactions without heating to achieve an LOD of 2 ng l-1. Another Chinese group497 described an AAS method for the analysis of water in which the sample was treated first with permanganate and then successive additions of small volumes of sulfuric acid.The excess oxidant was removed by reaction with hydrogen peroxide. This method was claimed to be superior to one in which hydroxylamine hydrochloride was used to reduce the excess permanganate but no indication of the basis for the comparison was given. Lamble and Hill498,499 have developed the UK FI method described above so that slurry samples (fish tissue and sediment) could be processed. After merging with the bromate–bromide oxidant stream the sample zone was passed through a microwave reactor. As part of the study an open-vessel microwave procedure was developed based on digestion with hydrogen peroxide nitric and sulfuric acids. A variation of this latter procedure was used500 for the analysis of similar samples by FI–CV-AAS using borohydride as the reductant.Possible interferences from antimony arsenic cadmium copper lead nickel selenium and silver were investigated and it was noted that only selenium and silver depressed the Hg signal whereas the other elements caused an enhancement. The concentrations of the interferents needed to cause significant eects were too high for there to be problems with the majority of environmental and biological samples. Schnitzer et al.501 compared the use of open and closed vessels in focused microwave digestions for the determination of total Hg by CV-AAS. Nitric and sulfuric acids and hydrogen peroxide were used to attack fish kidney and milk powder RMs. Although accurate results were obtained with the open vessel method fat was not digested.Borohydride was used as reductant. Fostier et al.502 developed closed vessel microwave procedures for the decomposition of fish and sediments for the determination of Hg by AAS. They found that fish fat content had no eect on the solubilization of the analyte. Guo et al.503 determined Hg in saliva in an FI–AAS procedure in which the sample was first attacked o-line by bromide–bromate and then on-line with permanganate. Borohydride was used as reductant. The dedicated Perkin-Elmer FIMS instrument was used and a method LOD of 50 ng l-1 in 500 ml of saliva diluted 1+9 was achieved with a sample throughput of 80 h-1. Chinese workers504 developed a method for the determination of Hg by AAS in 18 dierent biological samples (including of course hair) though no details of the digestion procedure were given.Polish workers505 devised a method for the analysis of coal in which a ground sample was covered with alumina and ignited at 1000 °C in a stream of O2. The combustion products were scrubbed by a solution of nitric acid and permanganate. Hydrogen peroxide was then added to the scrubber from which Hg0 was generated by the addition of SnII. The LOD was 0.1 mg kg-1. Chinese workers506 used nondispersive AFS for the determination of Hg in milk powder. A signal enhancement was obtained which was attributed to the decomposition of Hg(NxOy)z (formed in the nitric acid 360R Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 digestion) on the addition of SnII releasing Hg0 and NxOy both of which give a response with the detector used.In what appears to be the only paper describing the use of amalgamation trapping Ombaba 507 reported a method for the analysis of several biological and environmental materials. High pressure microwave digestion with nitric acid was used to attack samples some of which required further digestion (HF was added for sediments and coal fly ash and persulfate was added to urine). The reductant was SnII and after trapping on gold the collected Hg was remobilized by heating to 500–800 °C. Just as for As and Se there is considerable activity related to the development of methods for Hg speciation. Antonovich508 has reviewed (in Russian) methods for the speciation of Hg in environmental samples.Chromatography and CV-AAS feature prominently in the methods used. Jones et al.509 have developed methods for the determination of total and organic Hg in water soil and tissue by AFS. For total Hg soil sediment and fish samples were digested with nitric acid in sealed glass ampoules. Water samples were attacked with bromide–bromate. Organomercury compounds were adsorbed from water onto sulfhydryl cotton eluted with potassium bromide and copper sulfate and then extracted with dichloromethane. Sediment soil and tissue samples were extracted first with acidic potassium bromide and copper sulfate which in turn was extracted with dichloromethane. Initial extracts were subjected to thiosulfate clean up and the organomercury compounds converted to chlorides by the addition of cupric chloride followed by extraction into dichloromethane.The organic extracts were analysed by capillary column GC with AFS detection. Somewhat similar extraction procedures were used by Lester and coworkers510 for the determination of total and organic mercury in dogfish liver eel and roach. Samples were digested with 20% tetramethylammonium chloride for 4 h at 60 °C in a closed vessel. The digest was extracted with toluene to which 1 ml of 0.1 mM cysteine or thiosulfate was added and the aqueous phase was analysed for Hg by FI–CVAFS. Total Hg was determined by digestion with hydrogen peroxide and nitric acid. In the FI manifold a stream of peroxodisulfate and copper sulfate was used for the on-line release of HgII.Nixon and co-workers511 devised FI procedures for speciating Hg in blood and urine. For the determination of total Hg a persulfate digestion was used whereas for inorganic mercury acid permanganate was used. Organic Hg was determined as the dierence. A total of 902 samples from individuals with no extraordinary Hg exposure were analysed for which the total Hg content was in the range 0–10 mg l-1 for both samples. Pervaporation has been used512 as the basis for speciation in solid samples. Inorganic mercury was determined by sealing the sample and stannous chloride solution in the pervaporation cell in a microwave oven whereas organic Hg was released by the action of bromide–bromate and stannous chloride. The released mercury vapour was swept into a gas–liquid separator and detected by AFS.The method was applied to an RM (CRM 145R) whose total Hg content was 2 mg kg-1. A Japanese group513 used CrII as a reducing agent in a postcolumn reaction to generate Hg0 from various organomercury compounds separated by reversed-phase HPLC on C18. The mobile phase was a 40560 (v/v) mixture of methanol and 0.02 M acetate buer (pH 5) containing 0.02% (v/v) mercaptoethanol. Mercury vapour was separated by a porous membrane and detected by AFS. Falter et al.514 devised a screening procedure for the determination of methylmercury and inorganic Hg in hair by HPLC with AAS detection. Samples were extracted with 50% acetonitrile containing 50 mM sodium pyrrolidinedithiocarbamate SPTC (adjusted to pH 5.5 with acetate buer).A 50-ml sub-sample was injected into a reversedphase HPLC system and the components separated on a C18 column with 65% acetonitrile containing 0.5 mM SPTC. Postcolumn UV photodegradation was used to release inorganic Hg prior to reduction with borohydride. 1.3.3 Volatile organic compound generation and metal vapour separation Ni515 has reviewed (88 references) the application of chromatography with atomic spectrometric detection for the characterization of trace organometallic compounds in environmental and biological samples. She points out the importance of the provision of reference materials certified for the various chemical forms of an element rather than for just the total element. Examples of recent developments such as the certification of a dogfish RM for arsenobetaine content and a fish tissue for tributyltin were given.Unfortunately the article is in Chinese. Welz and Sucmanova516 presented preliminary results of a procedure for the determination of mercury species by the volatilization of derivatives formed in the reaction with tetraethylborate. The method involved on-line derivatization cryotrapping and chromatographic separation. A quartz tube furnace atomizer was used for AAS detection. The authors suggested that a ‘soft focused’ microwave field might be useful for the extraction of analytes from biological and environmental matrices. A similar procedure was suggested by Fischer and coworkers517 The LOD for a 200-ml sample was 0.5 ng l-1.A supercritical CO2 extraction of methylmercury from seafood has been developed518 It was not clear how the methylmercury was atomized though the final measurement was by AAS. A GC method519 which separated methylmercury and inorganic Hg based on extraction of water samples with dithiocarbamate resin has been applied to some lake and tap water samples. The extracted Hg species were transferred to hexane and derivatized with butylmagnesium chloride in tetrahydrofuran. The LOD for methylmercury was claimed to be 30 ppq whereas that for inorganic Hg was a mere 400 ppq. Several separations of Sn compounds by GC have used AAS detection. An overview of methods involving leaching and extraction has been provided520 Microwave-assisted procedures had advantages of speed over conventional extractions.A number of examples of capillary GC separation with AAS flame photometric or MIP-AES detection were given. More specifically Shawky et al.521 used GC–AAS for the determination of various organotin compounds in fish. Following leaching with acidified methanol the Sn compounds were ethylated prior to GC separation. The LOD was 2–4 ng g-1 (fresh mass). The authors point out that it was easy to lose analytes if samples were dried. Chinese workers522 have determined various alkyltin species in water samples by GC–AAS. A packed column (OV-101 on Chromosorb G) and temperature programming was used. Donard and co-workers523 developed a method for the determination of butyl- and phenyltin compounds in sediments based on microwave-assisted leaching with acetic acid derivatization with tetraethylborate extraction into iso-octane and capillary GC separation.Detection was by flame photometry at 610 nm. Sulfur (S6) interfered and was removed by precipitation from the iso-octane phase with tetrabutylammonium sulfite. The method was validated by accurate analysis of two sediment CRMs. Ohta et al.524 determined Zn in biological materials by a technique they called ‘sequential metal vapor elution analysis’. The digested sample solution was introduced by ETV (at 1880 K) at one end of a molybdenum capillary tube (250×1.22 mm) containing a tungsten spiral at a temperature of 1150 K. Detection was by AAS with the light beam passing through a 0.8 mm hole at the end of the tube.With a carrier flow of H2 at 1.3 ml min-1 Zn was separated from aluminium calcium copper iron lead potassium and sodium. Sample pretreatment involved digestion with HCl and the method Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 was applied to several CRMs (bovine liver citrus leaves and non-fat milk powder) with precisions ranging from 4–10% RSD. This may have been partly due to sampling errors as the sample mass taken was only 1–10 mg. Fourier transform spectroscopy was used by Chakrabarti’s 1.4. Spectrometers Patents filed during the year for AAS equipment include from Japan an AAS with automatic sample enrichment,525 with a nebulizer having improved eciency,526 with an ETA and data processor527 and with an ‘oven’ design (for ETA) giving enhanced sensitivity.528 An AAS for Hg determination using a sample cell with two or more connectable chambers to widen the dynamic range was filed in Germany.529 Spectrometers designed for special purposes were also described.A compact self-contained optical spectrometer530 could be hand-held and used a diode-array detector giving a wavelength accuracy of ±2 nm at 434 nm. A portable AAS with W-filament ETA and modified Smith–Heiftje background correction for blood-lead screening was described by Aldous.531,532 An ‘inexpensive tungsten coil AA spectrometer’ the combination monochromator–CCD detector being mounted on a PC card was developed for measuring Cd533 and Pb,195 both to about 1 ng ml-1.The atomizer was a standard 12 V projector lamp bulb without the glass envelope. A ‘multielement’ version using a Cu–Zn HCL was also described.534 An AAS was modified for making elemental determinations in a radiological environment535 at the Argonne Laboratories at Idaho Falls USA. The light source optical and electronic components are mounted outside the ‘hot’ area and radiation to and from the sample atomization system inside is directed through two ports in the two feet thick concrete separating wall. The application of AAS to in situ monitoring of deposition of dielectric materials536 involved measurements of Al atoms from decomposed Al2O3 molecules during RF sputtering processes. The technique is non-intrusive and the absorption signal is directly related to the Al2O3 deposition rate enabling film thickness to be precisely controlled.A double-encoding Hadamard transform spectrometer was developed in Wuhan China,537 to improve S/N ratio and light throughput. It is said to be capable of multiplexed simultaneous detection with one detector. In a conference paper the fundamentals of acousto-optic devices were described538 and some possible applications to analytical atomic spectometry were discussed. Further developments will now be awaited with interest. 1.4.1. L ight Sources An erratum has been added539 to the ‘Compendium and Critical Review of Neutral Atom Resonance Line Oscillator Strengths for AAS’540 cited in last year’s update.1 Two papers from the National Institute of Standards and Technology Maryland USA concern wavelengths541 and irradiances542 of spectral lines in Hg pencil lamps.These are low pressure discharge sources commonly used for alignment and calibration of spectrometers in UV and visible spectral regions. The values recommended should enable these lamps to be used as secondary standards for both wavelength and radiometric purposes. Some types of line broadening and their causes were discussed by Ball.543 The phenomenon is complex particularly in gas-phase samples and a knowledge of the processes is useful especially when the normal Gaussian or Lorentzian function does not describe a particular situation well. group in Ottawa to examine atomic line profiles in both HCL and GD sources.544 Sensitivity and linear calibration range 361R are more adversely aected by high lamp currents in glow discharge AAS than with flame or ETAAS using HCLs.Geometry of light source optics was treated in several papers. The German Perkin-Elmer team showed545 that nonuniformities dier significantly between HCLs EDLs and D lamps. The eect of this on the response to Beer’s law and the performance of AA spectrometers is discussed. Characteristics of the discharges from HCLs of dierent shapes and sizes were outlined by Williams et al.546 and some practical advice concerning the use of HCLs with dierent diameters in AAS is oered from Poland,547 the recommendations seeming to involve some masking of the tube itself. A newly developed ‘Hyper’ HCL developed by Shimadzu for As,548 is claimed to provide a ten-fold increase in intensity as compared with a conventional lamp.Moett from Varian Australia also describes the use of a new range of high intensity boosted HCL s,549 developed especially for the measurement of the volatile elements As Pb Sb Se and Tl in environmental samples. This paper appears to be an extension of the work by Schrader550 cited in last year’s update.1 Fe and Pt are measured simultaneously by Chinese workers,551 using an Fe HCL the method being based upon the interference by Pt on Fe. No indications of concentrations or accuracy are given in the abstract of the original Chinese paper. Two papers mention the use of diode lasers as sources. In one from Cardi UK,552 the laser is used to probe alkali metals in HCLs.In the other from San Diego USA,553 the laser is used to provide Doppler-free spectral resolution for measuring trace elements. In combination with this type of source low pressure discharge atomizers oer better spectral resolution (because of low Doppler and Lorentzian broadening) and ETAs give lower background noise than flames. The laser power requirements are low so several lasers can be used simultaneously for dierent elements. 1.4.2. Continuum source and simultaneous multi-element AAS A general paper on CSAAS from Smith in Cork Ireland,554 was an update on the use of the linear photodiode array (LPDA) which is claimed to give better detection limits than those in conventional line-source AA.This is due to the higher quantum eciency of the LPDA and also to the multiplex advantage of measuring intensities simultaneously over the entire relevant wavelength region. The advantages of a graphite furnace atomizer capable of operating from vacuum to pressures of 10 atmospheres were also enlarged upon. Multi-element graphite furnace AAS is also reviewed by Farah and Sneddon.179 Instrumental developments and applications are covered with 52 references. A 128-page review of multi-element line sources in AAS by Calloway is available on microfilm.555 Theoretical calculations were presented by Becker-Ross et al.,556 showing how characteristic mass can depend upon the spectral bandwidth in a CSAAS spectrometer. It was shown that spectral bandwidth should be less than the bandwidth of the absorption line (which was why Walsh specified the use of fine line sources in AAS 40 years ago).Three conference papers from Harnly,557–559 outlined convincingly the advantages of a CCD array detector over the LPDA in CSAAS. These lie in the fact that a much narrower entrance slit can be utilized to attain the same degree of sensitivity and that dierent parts of the absorption line profile can be used to establish a series of overlapping calibration curves thus extending the calibration range to about five or six orders of magnitude. The second paper described how these principles may be realized with the two-dimensional field of an echelle spectrometer. The third paper shows how calibration for high concentrations may be carried out using the wings of the absorption line profile where absorbance tends to be 362R Journal of Analytical Atomic Spectrometry August 1997 Vol.12 dependent on the square root of concentration (see also Section 1.5.2). The frequency-modulated simultaneous measurement of As Sb and Se using one EDL was described by Edel et al.560 Both hydride generation and ETA methods were used and good results obtained with standard reference materials. A German patent filed by Roedel et al. of Perkin-Elmer described multi-element AA spectrometers and techniques for their use.561 Radiation from several sources can be directed through the sample atomizer. Zeeman background correction was used in conjunction with an echelle spectrometer with a photodiode array detector.Programming cycles for individual analyses were detailed. In multi-element AAS it is unlikely that one set of atomization conditions can be chosen to be most eective for each of the elements to be measured. The compromises that have to be made may aect characteristic mass linearity of response and even accuracy. In an informative conference paper Shuttler et al. 562 described how optimum conditions for a set of elements may be established. The eects of integration time atomization site (e.g. wall or platform) and temperature were also discussed. Ammonium dihydrogenphosphate for example added to blood reference samples563 enables an elevated ashing temperature to be programmed so that both Cd and Pb can be measured simultaneously.Looking at the problems and potential of FAAS from the point of view of another manufacturer Moseley of Thermo Jarrell Ash,564 focused particularly on new hardware technology and improved software. In particular improved high performance servo systems developed originally to improve computer disk drive head positioning but now used in conjunction with optical encoders provide an ideal grating drive system for multi-element AA spectrometers. The device is highly repeatable and scans the entire relevant spectrum in 25 ms. 1.4.3. Background correction Two papers compared deuterium arc with Smith–Heiftje (also referred to as SR—‘self-reversal’) background correction methods. Fender and Butcher565 used the methods in molecular absorption spectrometry in a graphite furnace determining F by measuring the AlF absorption with a Pt HCL and Cl by measuring AlCl with a Pb HCL.The Smith–Heiftje method was not successful with the Pt HCL presumably because the Pt line being measured does not undergo self-reversal but otherwise it was believed to be more accurate. Lee and Kim566 used the two methods in determining Pb in blood. Their conclusion is not clear as it is simply stated that absorbance is higher for the D arc method but background correction is higher (sic) for the SR method. Mention might also be made here of a Chinese paper on ‘self absorption eect AAS’.567 Zn was satisfactorily measured in alcoholic drinks using selfabsorption eect (background) correction and the eect of boat (sic) and narrow pulse current were investigated for eight elements.In a paper already cited in Section 1.4 Aldous et al.531 described the use of the Smith–Heiftje method in their portable ETA AA spectrometer for measuring blood Pb. The eects of HCL pulsing frequency and lamp pulse current waveforms necessary to achieve eective background correction were investigated. The remainder of the abstracts to hand in this category concern Zeeman-eect background correction. Two papers by the same authors Sholupov et al.568,569 described a new version of Zeeman AA (ZAAS) with high frequency modulation of the light polarization. It is claimed that this enables the spectrometer transmission to be increased thus lowering detection limits and extending the analytical range.L’vov et al. continue the work on the algorithm for linearization of analytical curves in ZAAS reported in last year’s updates.1 It is now claimed to be applicable up to the rollover point for any element under any measurement conditions. 570 Only one blank and four calibration standards are required to evaluate the three relevant parameters and the linearization error is less than the random scatter error in replicate measurements. A knowledge of the roll-over absorbance is not as important as the Zeeman sensitivity ratio (R)-related coecient for accuracy in linearization of calibration curves according to Berglund,571 who has proposed an improved method for calculating R. A linearization program is presented on two disks for IBM compatible PCs.Background over-correction problems in the measurement of Pb in the presence of phosphate and various metals was the subject of a paper noted in last year’s update.1 Two more papers on this subject have appeared from the same source. In the first Parsons and Zong572 stated that the error is a spectral interference as it can be observed in the absence of Pb. Its magnitude is also dependent upon the field strength of the Zeeman system used suggesting that a molecular absorption band perhaps caused by the species PO is split in the magnetic field. In the other paper Qiao Parsons and Slavin573 reported how their blood Pb method developed and optimized for Zeeman AA spectrometers has now been successfully transferred to an instrument with continuum correction.The same method with both types of instrumentation gave results that agreed within limits similar to those found with reference materials. From China came two papers on theoretical aspects of ZAAS. A theoretical method of calculation of absorbance in ZAAS methods using constant and transverse magnetic fields574 was also applied to improving analytical performance. A study of the behaviours of two Te lines of diering sensitivities in dierential modulated ZAAS575 showed dierent splitting patterns. This apparently does not aect line sensitivity although a knowledge of the pattern overlap may be important. 1.4.4. Detectors Three conference papers came from manufacturers of detectors and instruments.Toriyama et al. (Hamamatsu)576 described recent developments in PMTs. A ‘metal package metal channel dynode’ PMT is much smaller than a conventional tube and was claimed to give an ultra-fast response time with wide dynamic range. It is small enough to be incorporated into an array. A cooled PMT a newly designed tube with small thermo-cooler was said to exhibit significantly improved S/N characteristics. The use of special solar-blind photomultipliers was one of the stratagems recommended by Sanders of Varian577 in the optimization of instrument hardware for the measurement of trace elements. Such PMTs have a very low response in the visible region of the spectrum where emission from the hot ETA tube and its gases create increased noise.The other recommendation was not surprisingly the use of the boosted discharge HCLs noted earlier in Section 1.4.1. Both of these would of course result in improved S/N and thence lower detection limits. A spatially resolved AA spectrometer was described by Gilmutdinov and his hosts in Perkin-Elmer Germany.578 Not only is the spatial structure of the primary radiation beam non-uniform but so also is the distribution of atoms in the atomizer cell. Spatial resolution was achieved by placing solidstate detectors along the length of the spectrometer exit slit. The data output is a two-dimensional matrix of temporally and spatially resolved absorbances and requires new methods of data handling. A charged coupled device (CCD) camera was used by Chakrabarti’s group579 in Ottawa to measure the two- Journal of Analytical Atomic Spectrometry August 1997 Vol.12 dimensional distribution of analyte atom and oxide species in FAPES. While the atomization mechanisms underlying the formation of FAPES transients are generally similar to those in graphite furnace AAS the appearance temperatures in FAPES are sometimes lower suggesting that there is significant plasma induced dissociation of oxide species. The more sensitive CCD camera helps in the understanding of these atomization mechanisms. 1.5. Instrument Control and Data Processing Three general papers are worthy of note before this section is sub-divided. A technical report from IUPAC580 on absolute methods in analytical chemistry describes definitive and reference methods (and discusses the misuse of the term ‘absolute method’).In a truly absolute method the analyte is determined using numerical calculations derived from the theoretical model of the phenomena giving rise to the analytical signal which is itself based only on universal quantities and fundamental physical constants. The possibility of fully automating spectrometry was discussed by Scheeline.581 It is believed that full automation with ‘direct reading’ instrumentation (probably referring more particularly to AES) is feasible only for repetitive solution samples. A collection of Internet resources linking spectrochemists worldwide has been compiled by O’Haver.582 Included in the list are home pages of individual spectrocopists of journals and of academic departments research centres and institutes; also on-line software course materials papers posters reviews tutorials and more.1.5.1. Instrument Control The design of a peristaltic pump from Varian was described,583 and this has been incorporated into a sample introduction pumping system which is also used for on-line preparation of standards and dilution of over-range samples. Another paper about the same Varian equipment from Epstein et al. of the National Institute of Standards and Technology584 included studies on the determination of trace elements in urine and environmental samples and the evaluation of reference methods for serum electrolytes. The objective was to confirm that fast automated methods of dilution are capable of the same precision and accuracy as traditional manual methods.A conference paper from China585 described software control programs for a GFAAS system. These include modules for methodology diagnostics optimization of atomization sequences and a ‘knowledge acquisition’ interface. A comparison of sample dilution and automatic burner rotation methods for bringing absorbances into the best measuring range was made by McCrum586 using a GBC AA spectrometer with automatic burner rotation facility. It was concluded that with manual dilution of samples there is a greater likelihood of introducing systematic and random errors. 1.5.2. Data processing The interpolative standard addition method of Tyson587 was investigated by Koscielniak,588 who claims that in the determination of Mg in the presence of Al this method is not as eective as conventional calibration procedures in eliminating the interference eect.In comparing a number of ETA methods Kurfurst et al.235 pointed out that the complex expressions for uncertainty in results experienced when solid samples are analysed can be evaluated easily by modern computer programs such as MS-Excel. Kale and Voigtman589 claimed that in the processing of transient ETAAS signals peak-area measurements always give best precision regardless of the type source and statistics of 363R the noise. However variation in atomization conditions is more serious than noise as a cause of poor precision.These results were derived from digitally simulated ETAAS signals treated with various kinds of synthetic noise evaluated with peak height and peak area measurements under various conditions. Custom manipulation of sample data is treated in two conference papers from Varian.590,591 Their ‘Dynamic Data Exchange’ concept enables data to be transferred automatically between dierent applications on one PC or between PCs (i.e. between dierent software packages) or to a commercial spreadsheet. Kneisel et al.592 successfully used multivariate statistics to match prehistoric pottery sherds in South Arizona to the source of clay most likely to have been used for their production. Precision by atomic spectroscopy was shown to be better than that reported for neutron activation analysis.Lithium metaborate fusion and HF digestion methods were applied to clays and sherds and accuracy was checked with standard reference materials. After measuring the metal contents of welding fumes in a Spanish shipyard by FAAS following collection on paper filters Bellido-Milla593 subjected the results to multivariate analysis of variance to study 11 factors that could aect the metal levels. It was thus possible to establish the most favourable working conditions consistent with safeguarding the health of the workers. An automated GFAAS analysis system using the method of standard additions was described by Oatts et al.594 The OBEY computer program written for a PE Zeeman Model 5100 follows the US EPA quality assurance/control protocol for samples requiring analysis by the method of standard additions.A method for standardless analysis in ETAAS595 has been applied to the determination of Pb in air aerosols.596 A very high heating rate was used. Comparison with standard methods for Pb shows a dierence of 20%. Extending the calibration range by using the wings of absorption profile is a method described in two papers by Harnly559,597 already referred to in Section 1.4.2. With HCL sources limited movement of the measurement wavelength is possible allowing extension to about three orders of magnitude. With a continuum source calibration more than 6 orders of magnitude are possible. The limitations to this technique are purely practical more likelihood of contamination remaining in the furnace when higher concentrations are measured and more probability of line overlap interferences within the broader spectral region examined.Data handling methods relating to the transient signals characteristic of graphite furnace AAS were discussed in a conference paper appearing twice from Schrader et al. of Varian.598,599 Taking detection limit as the most useful figure of merit it is found though not expected that peak height measurements give better figures than peak area. The latter includes points which contribute to noise rather than to the signal and which may also be adversly aected by the cut-o of tailing peaks in the measurement sequence. Alternative strategies for increasing the S/N ratio were also investigated.The problems caused by non-linearity and roll-over of cali- 1.5.3.1. Calibration. The slope-ranking method was used by Vankeerberghen et al.607 to check the linearity of the first order bration curves in Zeeman GFAAS continued to exercise L’vov’s calibration line in GFAAS. It was able to discriminate very team and others. L’vov et al.600 studied the eects of HCL eciently between curvature and blank problems and the current and spectrometer slit width on characteristic mass mo technique was illustrated with Monte Carlo simulations. Zeeman sensitivity ratio and roll-over absorbance. They con- Rational design of linear calibration experiments were aimed cluded that self absorption of analytical lines and presence of ultimately at estimating the molar absorption coecient of non-absorbable radiation aect mo dierently and must there- chlorophyll by first measuring Mg by AAS.608 The original fore be corrected for separately.The correction procedure procedure was referred to in last year’s update,609 see reference given involves additional measurement of roll-over absorbance 1. Main sources of error in the calibration lack-of-fit dilution at a narrow slit width but is ineective for Mn and Ni of the and instrumental errors were accounted for by the design of nine elements investigated because resonance lines other than the procedure adopted. the analytical line pass through the slit. L’vov’s team also 364R Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 discuss in detail601 the reasons for systematic deviation of linearized calibration graphs from true linearity.Increasing the slit width and reducing HCL current are eective for some elements but other ways are suspected to exist. Yuzefovsky Lonardo Michel and others602 discussed mainly from a mathematical viewpoint the linearization model of L’vov.603 The assumption that the Zeeman sensitivity ratio R should be unity made by L’vov at that time to simplify the calculations was shown to be unsatisfactory because the linearized slope in the upper portion of the calibration curve diers from the slope of the normal linear region by a factor of about 30%. Better results have been obtained by incorporating the experimental Zeeman sensitivity ratio of 0.67 (for Pb) at the roll-over point by a method of successive approximations.Further development of this work was described by Yusefovsky et al. in a conference paper.604 The developed algorithm extended the working range by several orders of magnitude for Ag Cd Cu Mn Pb and Tl. The non-linear fit between calibration points was achieved with a third-order regression though the overall computational procedure was semi-empirical. Main limitations in the method are low atomization eciency of non-volatile analytes at higher concentrations and limited intensity of the light sources employed. The eect of the L ’vov calibration curve linearization method on the precision of analytical results obtained with Sn 224.6 nm in GFAAS was investigated by Belarra et al.605 Amounts of Sn present in the calibration aliquots injected were in the range 3.8–23.0 ng.The linearization model reduces to a fairly simple equation relating found absorbance with theoretical absorbance and absorbance of a limiting high concentration standard. It extends linearity over a concentration range some two to three times that of the normal linear portion of the calibration graph. While this would be helpful for many purposes the authors considered it to be inadequate for dealing with measurements in solid samples. Some major sources of uncertainty in the determination of Pb in blood by ETAAS were identified by Kristiansen Christensen and Nielsen.606 A detailed statistical treatment was described in which the overall ‘uncertainty’ was compared with the summation of the individual experimental standard deviations.The major factor was found to be random fluctuations of the analytical and blank signals at low analyte concentrations while tube wear and matrix eects increased with concentration. 1.5.3. Chemometric methods Where advanced mathematical procedures have been applied to the interpretation of results (e.g. the probability of matching a sample of clay from one of a number of dierent areas to a particular prehistoric sherd) we have included them in the previous section. Such methods,when used in actually obtaining results or in validating them are included in this section. We can recognize four principle aspects calibration and linearization; interference eects their utilization and correction; optimization of instrumental conditions; method validation.Chinese workers have applied artificial neural networks to non-linear response in spectrophotometry and in AAS.610 Three optimized neural network models were proposed with an extended delta-bar-delta learning algorithm. The models gave predictions agreeing with experimental values with relative errors of 1.9–5%. 1.5.3.2. Interference eects. A calibration model to overcome non-spectral interference eects was developed by Paz Carril et al.611 Prior knowledge of matrix composition is not required as the absorbance measured is a function of the amount of sample and the amount of analyte added. Results for Fe in the presence of interfering Cu gave an error of -20% for this method (compared with -42% for direct calibration and +10% for the method of standard additions).Some refinement in technique would seem still to be desirable. A study on the computation elimination of multi-interferent eects came from Chinese workers.612 The relationship between analyte and interferences was discussed the optimal mathematical model being obtained with a ‘descent searching algorithm’. The method was applied to the multivariate statistical analysis of a multicomponent system at a single wavelength. The indirect determination of V by computational FAAS4 is based on the interference eect of V on Cr which is described in terms of a multinomial. The concentration of V is calculated using the multinomial from the Cr absorbance in a solution of known Cr concentration.Multi-wavelength linear regression has been applied by Chinese workers to compensate for the overlap of Ga 403.298 nm and Mn 403.37 nm in the determination of both of these elements613 and in a similar method for measuring Fe and Pt using only an Fe HCL at 271.903 nm.614 Good results were obtained for ‘synthetic ores’. The same group has also published work on the Mn–Ga overlap between 403.08–403.45 nm using only an Mn HCL and employing the Kalman filter method with a PC.615 Recoveries on 16 synthetic samples ranged from 92 to 107%. Experimental design and empirical modelling were proposed as the basis of an approach to the quantification of interferences aecting GFAAS determinations.328 The interferences of Ca K Mg and Na on Te was the system studied by multivariate analysis and besides a better characterization of matrix eects the atomization mechanism of the analyte in the presence of these interferents was also clarified.1.5.3.3. Optimization of conditions. In the GFAAS measurement of Cd and Pb in seafood and environmental samples Stalikas et al.616 used a composite modified simplex optimization procedure to set ramp and ashing temperatures atomization ramping time and temperature and modifier concentration. The method took into account possible interactions among the parameters involved while minimizing the number of experiments needed to access the optimum parameter values. The way of introducing the matrix modifier into the graphite furnace proved to be one of the most important factors to be considered.It was injected onto a L’vov platform in the tube prior to injection of the sample leaving the modifier wet. Simplex optimization was also used by Iversen et al.398 to select conditions for the simultaneous determination of Co and Mn in urine by Zeeman GFAAS. Araujo et al.617 showed how multivariate methodology is important for the optimization of conditions in systems like platform ETAAS in which a number of parameters aect the experimental response. A fractional factorial design was applied to evaluate systematically and simultaneously the optimal drying ashing and atomization conditions for determining Cd in cocoa beans. The accuracy was checked using standard reference materials.Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 1.5.3.4. Validation of methodology. There are several papers from authors representing ocial bodies and of these two are from workers involved in the European Commission’s Standards Measurement and Testing Programme. A conference paper from Quevauviller and Hill618 set out the strategy used in the M & T programme aimed at harmonizing and improving methods of measurement and analysis when these are believed not to be accurate enough or when dierent laboratories obtain dierent results. The programmes involve intercomparisons on (i ) simple solutions to test methods of final determination (ii ) raw or cleaned extracts to test performance of separation methods (iii ) spiked samples to test extraction recoveries and (iv) real samples to evaluate overall performance of the method.The two examples quoted were methylmercury and butyltin compounds. Quevauviller et al.619 described the step by step approach adopted to improve the quality of determination of butyltins. A structured programme was carried out by about 20 laboratories leading to the production of CRM 462 a certified reference material for diand tributyltin. A similar study on the quality control of SeIV and SeVI determinations in simulated fresh water was reported by Cobo-Fernandez et al.620 This too led to the availability of two CRMS. Hasselbarth BAM Germany (a national institute for research and testing of materials) proposed plans for defining uncertainty in reported analytical results,621 as required by ISO Guide 25.An upper uncertainty limit is set which is then monitored using appropriate QC samples a scheme used in the field of industrial quality control but not much pursued in chemical analysis. An expression relating the uncertainty to be quoted to other statistical values and constants is given together with the most appropriate values of the latter. The correction of measurement errors in studies on method valuation is discussed by Christensen (National Institute of Occupational Health Copenhagen).622 Reference materials were used in an evaluation of blood Pb methods by Zeeman ETAAS and As and metabolites in urine by an FI–AAS method. A decision support system for ETAAS which now runs under MS Windows 3.x on a 486-based PC was described by Vankeerberghen et al.623 The run-suitability check investigates whether the quality of calibration and standard addition lines meets the requisite standards.If not the problem is identified and remedies put forward. The explorative method validation which is based on parts of the suitability check investigates whether proposed methods are promising or need further development. The most important parameter investigated is the calibration line itself e.g. blank correction non linearity presence of outliers and spread of data points; and this is done objectively and automatically. Standard addition lines are used to check for matrix interferences.These are evaluated by comparing the standard addition lines with the aqueous calibration line. The system used as a run suitability check was itself evaluated by comparing its diagnoses with the conclusions of experts on a large set of calibration lines. In a later paper from the same group Penninckx et al.624 presented a strategy for the evaluation of experimental results obtained by AAS. This is again focused on calibration and is based on the assumption that the analyst knows the approximate linear range of the calibration line which should extend to three or four concentration levels. It then goes on to investigate a second degree model which requires a set of points covering at least four concentration levels. Tests to trace single and paired outliers variances at dierent concentration levels and lack-of-fit were included.The results were used to construct a general validation strategy for use in AAS and other measurement techniques. Dundar and Haswell625 reported a study on the comparison of Cu speciation results from ETAAS and from computer modelled data. The latter were predicted using the NSPEC 365R computer speciation program which identified the same dominant anionic species as the experimental data. Conference papers from two instrument manufacturers described their approaches to quality control and system validation. Shuttler and Schlemmer626 from PE discussed the possibilities of on-line multivariate quality control in atomic spectroscopy (as compared with existing standard methods which have to be applied retrospectively).The work described was based on a scheme developed by Delves of Southampton University which was performed manually while the analyses were being run. This was extended and fully automated for use with a single element ETAAS instrument. Further extension to multi-element analysis was envisaged here as an intelligent automated quality control system which would be essential if unnecessary and time consuming failure actions were to be avoided. Oey and Morton from Unicam described a validation package627 which ensured compliance of their instrumentation with the dicta of various regulatory bodies (FDA EPA ISO etc.). System qualification covers instrument speci- fication and design installation operation and performance.Method validation was also discussed it being shown how the analyst could be helped to develop working methods and to validate the data generated confirming their accuracy reliability and fitness for purpose. 2. ATOMIC FLUORESCENCE SPECTROMETRY The reports to be reviewed in this section of the ‘Update’ have been selected for their emphasis on AFS. Where the AFS contribution is incidental to the prime purpose of the study e.g. hydride generation flow injection or chromatography such papers will be reviewed elsewhere (e.g. Sections 1.1.2.3 1.1.2.5 1.3). The most significant areas of activity in AFS appear to be attempts at commercial exploitation of cold vapour/ hydride generation with some multi-element measurement capability using discharge lamp excitation and research studies employing electrothermal atomization and laser excitation.A number of reviews of AFS its principles and practice have been published. These reviews include one by Greenfield wherein future developments are considered.628 One review concludes that the detection limits of a number of transition metals by GF–LEAFS are superior to those by GFAAS and ICP-MS,629 while another finds that HG–AFS is superior to other related techniques for determining hydride forming elements in environmental samples.630 Analytical atomic/ionic fluorescence spectrometry has been the subject of a review in Chinese by Gong et al.631 A general review containing 528 references to AAS AFS and FAES publications in the Chinese literature during 1983 and 1994 has been prepared by He and Shu.632 2.1.Discharge-excited Atomic Fluorescence Several papers have descibed the application of commercial vapour generation-AFS instruments. The linear range for the determination of Hg by a vapour generation technique was extended by means of a computer controlled discrete sample injection system.633 When the analytical signal exceeds a predetermined level the software re-sets the sampling valve to increase the dilution of the sample. The linear dynamic range was from 1 ppt to 10 ppm Hg. The performance of a fully automated multi-channel (2 or 3 elements) AF spectrometer incorporating hydride generation and pulsed high intensity hollow cathode lamps has been reported.634,635 The linear dynamic range was over 3 orders of magnitude with RSDs of less than 1%.The detection limits for As Pb Ge Sn Se Te Bi and Sb were below 9×10-11 g ml-1 and for Cd and Hg below 8×10-12 g ml-1. A similar instrumental system has 366R Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 been described636 and applied to the determination of Pb in whole blood serum plasma and red blood cell samples.637 Recoveries were 95–105% with an LOD of 1 ng ml-1. Most interferences in vapour generation systems occur during the chemical reaction phase and are reviewed in depth elsewhere in this ‘Update’ (Section 1.3). In a study of the determination of As in environmental samples by HG–AFS Guongong et al.638 minimized interference eects by the addition of a thiourea–ascorbic acid mixture.No interferences on 5 ng ml-1 As by co-existing hydride-forming elements at concentrations up to 10 mg ml-1 were observed. At higher concentrations of such elements (200 mg ml-1) however spectral interferences due to molecular fluorescence or scattering could occur. A commercial ICP-AFS instrument was modified by Yehl and Tyson to monitor the transient concentration profiles produced by chromatographic separation and flow injection.639 A liquid chromatographic–hydride generation interface was developed and applied to the speciation of As in contaminated soils. Gong et al. have studied an ICP-AFS system using high current (15 A) microsecond pulsed (1.5 ms) hollow cathode lamps.640 Using an 850 W ICP as atomizer the LOD for Ca was 0.07 ng ml-1 which was 40 times better than with a conventionally pulsed HCL at the same ICP power.A lowpower (40–80 W forward power) Ar microwave plasma torch with a desolvating pneumatic nebulization system has been modified for use as an atomization cell for AFS.641 The modified design allows atmospheric pressure Ar to be easily ignited and maintained at low power and gas flow rates. Using pulsed hollow cathode lamps LODs ranged from 0.25 ng ml-1 (Cd) to 121 ng ml-1 (As) for the 10 elements examined. The dynamic range was over 3 orders of magnitude with RSDs from 0.42% (Zn) to 4.9% (Ag). A single paper on flame atomic magneto-optic rotation (AMORS) has been received for review.642 The Faraday con- figuration was used for the determination of Ag and Mg by FAMORS and FAAS with the same apparatus.The best AMORS results were obtained when the angle between the planes of polarization of the polarizer and analyser prisms was 45°. Under these conditions the detection limits and linear ranges of FAMORS were slightly better (~40%) than those of FAAS. and characterization of a flowing electrolytic hydride generator 2.2. Laser-excited Atomic Fluorescence In addition to its application to the quantitative determination of elements laser induced atomic fluorescence is used to probe atomic and ionic states and distributions. Photon-induced 2 atomic fluorescence from alkali dimers Na and K2 at 500 K has been studied theoretically by means of quantum mechanics.643 It was predicted that the induced fluorescence intensity would vary with the energy of the exciting photons. The ion beam in an ICP-MS has been characterized using a laser excited ionic fluorescence probe.644 The probe generated axial and radial maps of the ion beam with 1 mm spatial resolution at positions >20 mm downstream of the skimmer cone. The test element was Ba. The Ba ion density in the centre of the beam and 20 mm from the tip of the skimmer cone was >16 ions mm-3 ppm-1. The FWHM of the beam 28 mm from the cone was 7 mm. A comparison of flame ICP and electrothermal atomizers for the determination of Sb Se and other metalloid elements by LEAFS found that ETA achieved the lowest detection limit.645 Due to the far-UV wavelengths of Se transitions a thermally assisted LIF approach was adopted utilizing laser excitiation at 206.279 nm and fluorescence detection at 196.026 nm and 203.985 nm.Using ETA the LOD for Se was 15 ppt and for Sb 2 ppt. A dissertation on the design for As and Se analysis by excimer LEAFS has been prepared by Hueber.646 The eciency of sample utilization in ICP–LEAFS was considerably improved when the laser pulse was synchronized with the injection of single dried microparticles into the ICP.647 This device a Monodisperse Dried Microparticulate Injector (MDMI) may in theory improve the eciency of sample utilization by up to 10 000 times but in the opinion of this reviewer is unlikely to be of great significance in LEAFS as lack of sample is not a common problem particularly when ETA–LEAFS is employed.A see-through hollow cathode glow discharge lamp was used as the atom cell in experiments directed at the determination of actinides and lanthanides by diode-laser LEAFS.648 Preliminary trials achieved an LOD for Na of <1 ng. A miniature glow discharge atom reservoir has been developed and applied to the determination of Eu Pb Tu and Y.649,650 The analyte was deposited on an Ni rod cathode which after drying was inserted into an Ar filled chamber at a pressure of 5.5 Torr. The chamber served as the anode and a current of 20 mA was passed. Atoms sputtered from the cathode were excited by a Cu vapour laser-pumped dye laser. The LODs for Pb were 0.03 pg (peak area measurement) and 0.6 pg (peak height); the RSDs were 25 and 14% respectively.The calibration graph was linear over 6 orders of magnitude. Graphite furnace L EAFS is the most sensitive and practical form of LEAFS and has been the subject of two reviews.100,651 The first review also includes reference to laser excited molecular fluorescence while the second predicts future developments. Background and interferences can be serious problems when applying ETA–LEAFS to the analysis of untreated real samples. Studies of alternative optical systems have been carried out by Michel and co-workers with a view to maximizing the collection of fluorescences while rejecting black body radiation from the furnace.652 The best results were obtained by replacing the plane mirror–lens combination with a single 90° o-axis ellipsoidal mirror fragment with a dispersive detection system.With the ellipsoidal mirror system the LODs for Co and P were 20 fg and 7 pg respectively. Matrix eects were reduced considerably by a double furnace system developed by Winefordner and co-workers.653 This device is based on the temporal separation of the analyte signal from that of the matrix. The sample is placed in the inner chamber of the furnace from which on heating the sample vapours diuse at dierent rates through the graphite walls to enter the optical path. This approach to the reduction of matrix interference has been attemped from time to time over a number of years one of the earliest being by L’vov in 1976.654 In the present work the LOD for Ag was 40 fg (=4 ppt) with a dynamic range of 6 orders of magnitude and RSDs 1.1–0.4%.When Ag in seawater was determined there was a 2-fold depression of the analytical signal compared with that from an aqueous sample of the same concentration. No signal depression was observed when analysing soil SRMs. An investigation of the eect of furnace operating pressure on sensitivity in ETA–LEAFS found that the optimum was close to atmospheric pressure.655 A theoretical model based on two functions one the supply of analyte by vaporization the other its removal by diusion was developed. The model was tested for the direct determination of P and Te in Ni alloys and Co in glass. No change in sensitivity was observed when the pressure was reduced from atmospheric to 70 kPa reduction to 7 Pa decreased sensitivity by 2 orders of magnitude but the peak shape was improved.Several analytical applications of L EAFS have been reported. A tungsten furnace and a modified commercial graphite furnace have been applied to the determination of As in reference materials and lake water.656 The modifications were designed to reduce the stray light problem and facilitate synchronization of the Zeeman correction system with the laser operation. As the ArF laser line overlaps the 193.76 nm As line no dye laser was required. Background correction by subtraction of the Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 signal between laser pulses and by simultaneous on-line/o- line measurement was investigated.Selenium in blood samples was measured by ETA–LEAFS.657 To generate the required tunable laser radiation in the VUV spectral region a laser system comprising two dye lasers pumped by an Nd5YAG laser was developed. The laser radiations were frequency doubled and sum frequency mixed by non-linear optical KDP or BBO crystals respectively. The output in the VUV consisted of 5 ms pulses at 20 Hz with up to 100 mJ of energy. The detection limit was 1.5 fg (=1.5 ng ml-1) of Se. L ead in blood has been determined by ETA–LEAFS.658 No matrix modifi- cation was required. The sample was diluted 20+1 with high purity water. Fluorescence was excited at 283.3 nm and detected at 405.8 nm. The LODs of Pb were 0.4 and 10 pg ml-1 in aqueous and blood samples respectively.Mercury in Ag has been determined by ETA-LEAFS without separation or preconcentration. 659 Two wavelength excitation (l1=253.7 nm l2=435.8 nm) with detection at 546.1 nm was employed and particular eort made to optimize the optical alignment. The LOD was 0.5 pg (=50 ppt) of Hg. The P content of polymers has been determined by direct solid analysis.660 The sample (0.1–1.0 mg) was placed on the L’vov platform of an ETA–LEAF instrument excitation was at 213.618 nm and fluorescence was collected at 253.4 and 253.6 nm. The LOD was 8 pg of P; the linear dynamic range was over 5 orders of magnitude and the RSD was 11–12%. Gold in air particulates has been determined by ETA–two-colour LEAFS.661 An absolute instrumental LOD of 1 fg of Au was achieved but owing to noise in the blank the practical LOD was 1 pg of Au; with a typical sampling volume of 10 m3 air while the method was limited to Au concentrations above 0.1 pg m-3.The determination of Au Ir Pd Pt and Rh in silver nitrate by ETA– LEAFS is subject to strong interferences by the formation of condensation products. This problem has been investigated by Masera et al.662 Two kinds of condensation were identified diatomic molecules (AgH Ag2) and condensed particles. The interferences were reduced by use of neon as the purge gas two-step vaporization to remove part of the matrix before atomization of the analyte elements and tranverse heating of the atomizer.663 3. LASER-BASED SPECTROSCOPY Developments in laser technology are the key to progress in the utilization of lasers in analytical atomic spectrometry.The principal limitations are wavelength range stability complexity and cost. It is likely to be some time before all these problems have been overcome. The uses of the excimer laser in atomic spectroscopy have been reviewed by Thiem et al.664 Its main applications were found to be in sample introduction of solids and as a pumping source for other lasers. The design and performance characteristics of a pulsed tunable dye laser for the simultaneous generation of 2 independently tunable UV wavelengths has been described by Schuetz et al.665 Two such systems were applied to the simultaneous determination of Cd Mn Ni and Pb by ETA–LEAFS.Michel and co-workers have been active in exploring the possibilities of tunable solid state lasers notably a 20Hz optical parametric oscillator (OPO) laser.666,667 The OPO laser can be easily tuned by computer throughout the visible spectrum. The main technological challenges lie in the pointing stability and high energies required of the YAG pump laser. The OPO laser system was applied to sequential multielement determination of Co Cu Mn Pb and T l in much diluted Bualo River sediments by flame LEAFS.104 A wavelength slew scan rate of 0.125 nm s-1 was employed to tune the laser to individual analytical lines. A measurement speed of 45 s measurement-1 was achieved. Detection limits were in the 367R femtogram range. No matrix interference was observed and the RSD was <5%.Frequency doubled tunable diode lasers have been investigated as sources for the monitoring and control of vapour deposition processes by AAS. Aluminium vapour density in a vacuum dc sputtering chamber was measured by AAS at 394 nm and used to regulate the deposition rate of Al to 900 A ° min-1.539 The control of the deposition of Y based on AAS and employing frequency doubled diode lasers has also been investigated.668 An element selective detector for GC based on wavelength modulation diode laser AAS in an He MIP was used to measure chlorinated carbons by the absorption of excited metastable chlorine atoms.669 The MIP completely dissociated the halocarbons present in the GC euent and LODs were of the order of 1 mg ml-1 or 80 pg s-1.Electrothermal atomization intra-cavity laser AAS has been applied by Barakov et al. to the determination of Cs Mn and Tb 670 and Al Cr Fe and Mn.671 A flashlamp-pumped tunable multi-mode dye laser operating in the visible region of the spectrum was used as the primary light source. Detection limits of Cs Mn and Tb were 5 20 and 1000 mg l-1 respectively with an RSD at the middle of the dynamic range of 9.7%. The analytical possibilities of a new technique intracavity laser frequency dispersion spectroscopy were investigated theoretically and experimentally by Zheltukhin.672 The method was based on the linear repulsion of frequencies of a 2-node laser with intracavity absorption. A graphite furnace tube was mounted along the axis of the laser resonator and could be operated under vacuum or He filled at 2 kPa.When 300 mg of Na were atomized the frequency change was of the order of 1011 Hz. The LOD was approximately 3 ng of Na. 3.1. Laser Ablation and Excitation The interaction of laser beams with matter is a process of great complexity and is of considerable interest to material scientists and spectroscopists alike. A theoretical study of laser irradiated Al particles predicted that at the start of the heating process the contribution of heat exchange with the environment is dominant but as the boiling point is approached vaporization plays the main role.673 A comprehensive review and experimental study of the laser ablation of Cu has been presented by Gill et al.674 At laser fluences of 35 mJ cm-2 the material removal rate by low power resonant laser ablation was typically <10–3 A ° while atom velocity distributions were Maxwellian with peak velocities of 1–2×105 cm s-1.Diraction-like surface features were noted along with non-thermal desorption mechanisms. A laser multi-element scanning AA instrument has been developed which employs solid sample vaporization using a continuous-wave laser.675 High sensitivity in multi-element analysis was demonstrated. L aser-induced breakdown spectroscopy (L IBS) is a technique in which a laser pulse is used to create microplasma from a solid or in a gas. The light from the plasma is collected and spectrally analysed to determine the elemental composition of the sample.The technique requires minimal sample preparation and facilitates in-situ analysis. A spectrally-resolved imaging study of the dynamic evolution of the laser spark as used in LIBS has been carried out by Multari et al.676 The results indicated that analytical measurements could be improved by use of spatial resolution. The relative merits of spatial and temporal resolution were examined. Similar studies were carried out by Vordillo et al. using photovoltaic solar cells as the test material.677 The samples were irradiated at 18×1012 W cm-2 with the second harmonic (532 nm) of a Nd5YAG laser. The distributions of atomic and ionic species in the plume were studied. Many of the reports of L IBS are directed at demonstrating the feasibility of a given analysis rather than presenting practi- 368R Journal of Analytical Atomic Spectrometry August 1997 Vol.12 2 cal applications. An attempt has been made to analyse soils as airborne dust.678 The shock wave generated on or in a loose soil surface by a laser creates a dust cloud; the particles of the cloud are then vaporized and excited in the laser spark formed in the air over the sample. The method was evaluated for the analysis of contaminated soils. Soils have also been analysed by forming the sample into an annulus and rotating (1 rpm) the annulus in a focused KrF excimer laser beam (repetition rate 1 Hz).679 The optimum delay between laser pulse and spectrum acquisition was 0.2 to 1 ms. Data was collected for 25 to 100 pulses and gave LODs for Cr and Eu of 2 and 100 ppb respectively.Geological samples were analysed by L IBS with an Nd5YAG laser at 532 nm and 0.012 mm2 spot size.680 The method detected minor components contained in silicate vanadate and sulfide rocks allowing the characterization and identification of the minerals. The eect of electron density and plasma temperature on the determination of As Sb and Sn by LIBS has been studied.681 By averaging the signals from 50 laser pulses RSDs of 5–10% were obtained. The LODs were 2.5 times lower in a He rather than an N atmosphere. Under optimal conditions accuracies of the order of 10% were obtained when metals in vitrified and melt glass samples were determined by LIBS.682 Sample rotation was used for the determination of Pb in concrete by time-resolved LIBS.683 The calibration function was the log ratio of the integrated intensities of the lead 405.78 nm and oxygen 407.59 nm lines against the log concentration of Pb in the concrete; the LOD was 10 ppm.The suitability of LIBS for the determination of Cu in steel has been examined using hard optics to deliver the laser radiation to the sample and fibre optics to carry the plasma emission to the spectrometer.684 The LIBS results were comparable with those by AAS with errors ranging from 30% at 0.02% (m/m) to 13% at 1.0% (m/m) of Cu in steel. The possibility of in-situ measurement is an attractive feature of LIBS and there are two reports of instruments developed for that purpose.One instrument was designed for the detection of metals in the environment.685 This instrument weighed 14.6 kg fitted into a case 46×33×24 cm and operated on 115 V ac. The small hand-held sampling probe contained a laser which was attached to the main analytical unit by optical and electrical cables. The instrument was evaluated for the analysis of soils. The LODs for Ba Be Pb and Sr were 265 93 298 and 42 ppm respectively. The coupling of high power laser pulses into fibre optics without damaging the fibres was found by Marquardt et al.686 to be a key issue in the development of an instrument for in-situ LIBS measurements. The problem was examined in depth. A fibre probe was developed and used as a laser ablation sample introduction system for an ICP-MS instrument.This arrangement facilitated simultaneous OES and MS measurements and was used to study the relationship between laser power reproducibility and sensitivity of the methods. The LIBS instrument was also applied to the in-situ measurement of Pb in paint.687 The LOD was 0.014% Pb in dried paint with an RSD of 5–10%. Two groups have studied the L IBS analysis of aqueous solutions of metals. Knopp and co-workers688 used an excimer (308 nm) pumped dye laser (500 nm) system with pulse energy of 22 mJ and duration of 28 ns. The emission was collected at 90° to the laser beam and resolved with a polychromator– intensified time gated photodiode-array detector. Calibration curves for Li Na Ca Ba Pb and Cd were linear with detection limits of 0.013 0.0075 0.13 6.8 12.5 and 500 mg l-1 respectively.For the analysis of metal submersed in water Cremers et al.33 found it necessary to use pulsed pairs. The first pulse formed a vapour cavity while the second pulse 20–200 ms later vaporized the sample metal to give similar excitation to that obtaining in air. Chromium Ca Mn and Si in steel were measured by this approach. When solutions or suspension of metals were to be analysed best results were obtained by generating laser sparks at the surface of the liquid.689 The LODs for Al Be Cr Ni Pb and Zn were 7 0.1 7 22 and 87 ppm respectively; the RSD being <12%. 2 L aser induced breakdown in H gas irradiated by an excimer 2 laser operating at 248 nm with a 20 ns pulse of up to 250 mJ was investigated by Yagi and Huo.690 The breakdown threshold showed a strong dependence on gas pressure and radiation flux.Similarly the breakdown of the gaseous metal hydrides of As and Sn was found by Singh et al.691 to depend on (1) the gas present (there was a significant dierence between N and He) (2) the gas pressure and (3) the laser beam intensity. Chemical reactions in the laser plume that consumed the free analyte atoms caused a signal decrease with time in a static system; hence a flowing system was preferred for analytical applications. 3.2. Laser Enhanced Ionization It is claimed by its advocates that LEIS is among the most promising techniques for meeting the current challenges in ultratrace elemental analysis. To date the application of the technique to the analysis of real samples has been slow.A book on L EIS has been written by pioneer workers in this field Travis and Turk.692 The use of LEI materials has been reviewed by Nasimov et al.693 By combining flame–LEI-MS with conventional flame–LEI the eciency of MS sampling and detection has been calculated.694 The technique was demonstrated using Ca Fe K and Na. A computer program has been published oering fully time-resolved simulation of twostep pulsed laser excitation of atoms in highly collisional media which can be used predictively in two-colour LEIS.695 In the search for the optimum atom reservoir suitable for the analysis of real samples an LEIS scheme was developed which involved graphite furnace atomization with L EI detection in an air–C2H2 flame.696 An attempt was made to characterize all stages from sample vaporization to ion detection in the flame with a view to achieving standardless or absolute analysis.The system was evaluated using Mg in 2% HNO3.697 The overall eciency of the system was 0.0025%. The experimental LOD was 20 pg (=2 ngml-1) of Hg and as is usually the case was determined by the noise in the blank signal. In the absence of such noise it was claimed the LOD would be reduced to 5.9 fg (=590 fg ml-1). A laser microprobe was used for the direct introduction of solid samples into a flame for LEI analysis.698 In the determination of Li in Al samples the surface resolution was better than 100 mm but the spatial distribution of Li in the laser plume was not uniform.Atomization of refractory elements in flames is limited by the formation of stable metal oxides. It has been proposed that L EI molecular spectrometry in a cool flame may be a way round the problem.699 This approach was applied to the measurement of L aO molecules by LEI in an air–natural gas flame. Without optimization of conditions an LOD of 10 ng ml-1 was achieved. The feasibility and application of 3-step excitation and ionization schemes for the determination of Ir in an air–C2H2 flame and Hg in a gas cell was studied by Matveev et al.700 The LOD for Ir was 0.2 ng ml-1 and for Hg it was 107 atoms cm-3 of air. T race impurities in natural and high purity substances (fluorides) were determined by L EI with flame atomization.701 This approach provided direct and rapid analysis of samples.J. F. J. Anal. At. Spectrom. 1996 11(8) 281R. 4. REFERENCES 1 Hill S. J. Dawson J. B. Price W. J. Shuttler I. L. Tyson 2 Ko H. F. Huang C. C. Yang M. H. Lan C. R. Huaxue 1995 53(4) 396. 3 Seubert A. Nachr. Chem. T ech. L ab. 1996 44(1) 40. Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 4 Wu Y.-L. Zhang C.-H. Zhu D.-H. Guangpuxue Yu Guangpu 5 Long S.-H. Liu Q.-L. Liao H.-X. Fenxi Huaxue 1995 6 Fullana-Barcelo B. Bosch-Serrat F. Marin-Saez R. M. Mauri- 7 Lang H.-Y. Xie Z.-H. Zhao Y. Huaxue Yanjiu Yu Y ingyong Fenxi 1995 15(2) 85. 23(4) 430. Aucejo A. R. Anal. L ett. 1995 28(12) 2247. 1995 7(1) 72. 8 Long S. Hu Q. Xiangtan Daxue Ziran Kexue Xuebao 1994 16(4) 32.9 Zhang J. L. Yao T. S. Li H. B. Guangpuxue Yu Guangpu 1996 16(1) 121. 10 Sun H.-W. Wang W.-X. Zhang D.-Q. Yang L.-L. (Dept. Chem. Hebei Univ. Baoding 071002 China). Presented at Proceedings of the 2nd International Symposium of Worldwide Chinese Scholars on Analytical Chemistry Shenzhen China November 15–18 1995. 11 Groll H. Schaldach G. Berndt H. Niemax K. Spectrochim. Acta Part B 1995 50(10) 1293. 12 Barker P. Thomas A. Rubinsztein-Dunlop H. Ljungberg P. Spectrochim. Acta Part B 1995 50(11) 1301. 13 Lakatos J. Bezur L. Lakatos I. Fresenius’ J. Anal. Chem. 1996 355(3–4) 312. 14 Parameswaran T. Snelling D. R. Appl. Opt. 1996 35(27) 5461. 15 Reichardt T. A. Klassen M. S. King G. B. Laurendeau N. M.Appl. Opt. 1996 35(12) 2125. 16 Neuber A. A. Janicka J. Hassel E. P. Appl. Opt. 1996 35(21) 4033. 17 Rothe E. W. Gu Y.-W. Reck G. P. Appl. Opt. 1996 35(6) 934. 18 Lo� fstedt B. Fritzon R. Alden M. Appl. Opt. 1996 35(12) 2140. 19 Partridge W. P. Jr. Klassen M. S. Thomsen D. D. Laurendeau 20 Daniel R. G. McNesby K. L. Miziolek A. W. Appl. Opt. 21 Nakai I. Terada Y. Nomura M. Uchida T. Physica B 22 Pupyshev A. A. Gubanova A. N. Byl’chenko K. F. Zh. Anal. N. M. Appl. Opt. 1996 34(24) 4890. 1996 35(21) 4018. (Amsterdam) 1995 208(1–4) 209. Khim. 1995 50(2) 173. 23 Slavin W. et al. At. Absorption Newsl. 1966 5 84. 24 Yamano Y. Jpn. Kokai Tokkyo Koho JP 07,120,382 [95,120,382] (Cl. G01N21/31) 12 May 1995 Appl. 93/292,793 27 Oct 1993; 6 pp.25 Dolinski A. A. Lavrinenko A. I. Zh. Prikl. Spektrosk. 1996 63(1) 26. 26 Choi S. K. Lee H. J. Park Y. S. Kim H. J. Yakhak Hoechi 1995 39(2) 148. 27 Pharr D. Y. Yen Teh Fu (Ed.) Enhancement eects of surfactants in flame atomic absorption analysis. Adv. Appl. Membr.-Mimetic Chem. 1995 79–94. 28 Turner A. D. Roberts D. J. J. Anal. At. Spectrom. 1996 11(3) 231. 29 Matusiewicz H. Kopras M. (Dept. Anal. Chem. Poznan Univ. Technol. 60-965 Poznan Poland). Presented at 5th Polish Conference on Analytical Chemistry Gdansk Poland 3–9 September 1995. 30 Ellis L. A. Roberts D. J. J. Anal. At. Spectrom. 1996 11(4) 259. 31 Ellis L. A. Roberts D. J. J. Anal. At. Spectrom. 1996 11 1063. 32 Sun H.-W. Yang L.-L. Zhang D.-C. J. Anal. At. Spectrom. 1996 11(4) 265.33 Pichahchy A. E. Cremers D. A. Ferris M. J. (Chem. Sci. and Technol. Div. Los Alamos Natl. Lab. Los Alamos NM 87545 USA). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 34 Huang G. Qian S. Gao F. Yejin Fenxi 1995 15(3) 16. 35 Yang H.-Y. Huang G. Q. Qian S. H. Fenxi Kexue Xuebao 1995 11(2) 35. 36 Guo M. Liu Y.-G. Lu Y.-Q. Hao T.-H. Ye D.-H. Guangpuxue Yu Guangpu Fenxi 1996 16(1) 113. 37 Huang Y. Mu G.-R. Zhong X.-F. L ihua Jianyan Huaxue Fence 1996 32(2) 84. 38 Ulivo A. Dedina J. Spectrochim. Acta Part B 1996 51(5) 481. 39 Kotova V. N. Zh. Anal. Khim. 1996 51(3) 336. 40 Shiowatana J. J. Sci. Soc. T hailand 1995 21(3) 197. 41 Kojima I. Takayanagi A. J. Anal. At. Spectrom. 1996 42 Berndt H. Schaldach G.Kaegler S. H. Fresenius’ J. Anal. 11(8) 607. Chem. 1996 355(1) 37. 43 Wang J. Marshall W. D. Analyst (Cambridge U.K.) 1996 121(6) 817. 369R 44 Tan Y. Momplaisir G.-M. Wang J. Marshall W. D. J. Anal. At. Spectrom. 1994 9(10) 1153. 45 O’Connor G. Ebdon L. Evans E. H. Ding H. Olson L. K. Caruso J. A. J. Anal. At. Spectrom. 1996 11 1151. 46 Fang Z.-L. Xu S.-K. Tao G.-H. J. Anal. At. Spectrom. 1996 11(1) 1. 47 Ma R.-L. Van Mol. W. Adams F. Mikrochim. Acta 1995 119(1–2) 1995. 48 Ruzicka J. (Dept. Chem. Univ. Washington Seattle WA 98115 USA). Presented at Colloquium Analytische Atomspektroskopie Huaxue Fence 1996 32(2) 82. Chim. Acta 1996 323 47. (CANAS ’95) Konstanz Germany 2–7 April 1995. 49 Tyson J. F. (Chem. Dept. Univ. Massachusetts Amherst MA 01003-4510 USA).Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 50 Hansen E. H. (Dept. Chem. Tech. Univ. Denmark 2800 Lyngby Denmark). Presented at The 15th Nordic Atomic Spectroscopy and Trace Element Conference Ebeltoft Denmark June 2–6 1996. 51 de Paiva A. C. de Oliveira Neto G. Tubino M. Anal. L ett. 1996 29(10) 1719. 52 Zhang M. Chen S.-Y. Lin S.-Q. Cheng L. L ihua Jianyan 53 Gine� M. F. Packer A. P. Blanco T. Freire dos Reis B. Anal. 54 Lapa R. A. S. Lima J. L. F. C. Santos J. L. M. Food Chem. 1996 55(4) 397. 55 Garrido J. M. P. J. Lapa R. A. S. Lima J. L. F. C. Delerue- Matos C. Santos J. L. M. J. Autom. Chem. 1996 18(1) 17. 56 Wabner C. L. Sears D. C. At. Spectrosc. 1996 17(3) 119. At. Spectrosc.1996 17(3) 107. 57 Matousek de Abel de la Cruz A. Burguera J. L. Burguera M. Wasim S. Rivas C. Fresenius’ J. Anal. Chem. 1996 354(2) 184. 58 Rondon C. Burguera M. Burguera J. L. Brunetto M. R. Carrero P. J. T race Elem. Med. Biol. 1995 9(1) 49. 59 Lopez-Garcia I. Sanchez-Merlos M. Hernandez-Cordoba M. 60 Hanna C. P. Baasner J. Leyrer M. Sperling M. Fang Z. (Inorg. Anal. Div. Perkin-Elmer Corp. Norwalk CT 06097-0215 USA). Presented at Pittsburgh Conference (PITTCON ’96) Chicago IL USA March 3–8 1996. 61 Sperling M. Fang Z. Tao G.-H. Welz B. (Dept. Applied Res. Bodenseewerk Perkin-Elmer GmbH 88647 Ueberlingen Germany). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 62 Elsholz O. Fresenius’ J. Anal. Chem. 1996 355(3–4) 295.63 Diamond D. (Lachat Instruments Milwaukee WI 53218 USA). Presented at Pittsburgh Conference (PITTCON ’96) Chicago IL USA March 3–8 1996. 64 El Azouzi H. Perez-Jordan M. Y. Salvador A. de la Guardia M. Spectrochim. Acta Part B 1996 51B(14) 1747. 65 Projahn H. D. Keim A. L aborPraxis 1995 19(6) 48. 66 Projahn H. D. (Varian GmbH Darmstadt Germany). Presented at Analytica 1996 Munich Germany April 23–26 1996. 67 Shrader D. Flajnik-Rivera C. (Varian Optical Spectroscopy Instruments Wood Dale IL 60191 USA). Presented at Pittsburgh Conference (PITTCON ’96) Chicago IL USA March 3–8 1996. 68 Shrader D. Vanclay E. (Varian Optical Spectroscopy Instruments Wood Dale IL 60191 USA). Presented at Pittsburgh Conference (PITTCON ’96) Chicago IL USA J.F. J. Anal. At. Spectrom. 1995 10(8) 199R. March 3–8 1996. 69 Van Staden J. F. Hattingh C. J. J. Anal. At. Spectrom. 1995 10(10) 727. 70 Olinic A. Pantea A. Coca M. Onisor I. Anca Z. Kovats A. Clujul Med. 1994 67(3–4) 225. 71 Cattrall R. W. Scollary G. R. Anal. Methods Instrum. 1995 2(2) 61. 72 Hill S. J. Dawson J. B. Price W. J. Shuttler I. L. Tyson 73 Fang Z. Xu S. Bai X. Anal. Chim. Acta 1996 326 49. 74 Tyson J. F. Idris A. B. Analyst (L ond.) 1981 106(10) 1125. 75 Sperling M. et al. Anal. Chem. 1991 63 151. 76 Koscielniak P. Sperling M. Welz B. Chem. Anal. (Warsaw) 1996 41(4) 587. 77 Silva E. C. Araujo M. C. U. Honorato R. S. Costa Lima J. L. F. Guidetti Zagatto E. A. Boscolo Brienza S. M. Anal. Chim. Acta 1996 319 153.78 Lopez-Garcia I. Sanchez-Merlos M. Vinas P. Hernandez- Cordoba M. Spectrochim. Acta Part B 1996 51B(14) 1761. 79 Mentasti E. Ann. Chim. (Rome) 1995 85(7–8) 443. 80 Greenway G. M. Nelms S. M. Skhosana I. K. (Sch. Chem. 370R Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 Univ. Hull Hull UK HU6 7RX). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 81 Greenway G. M. Nelms S. M. Skhosana I. K. Dolman S. J. L. Spectrochim. Acta Part B 1996 51B(14) 1909. 82 Sun S. Song X. Kang W. Youkuangye 1995 14(3) 194. 83 Elmahadi H. A. M. Greenway G. M. Microchem. J. 1996 53(2) 188. 84 Saraviva Miranda C. E. Freire dos Reis B. Krug F. J. J. Braz. Chem. Soc. 1995 6(4) 387. 85 Li Y. Chen D. Yankuang Ceshi 1995 14(4) 285. 86 Liu X.-Z.Fang Z.-L. Anal. Chim. Acta 1995 316 329. 87 Littlejohn D. Pasullean B. Anllo-Sendin R. McLaughlin L. Davidson C. M. (Dept. Pure and Appl. Chem. Univ. Strathclyde Glasgow UK G1 1XL). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 88 Kong Q. A.,Wu Q. F. Guo X. H. Fenxi Huaxue 1996 24(1) 1. 89 Sule P. A. Ingle J. D. Jr. Anal. Chim. Acta 1996 326 85. 90 Cespon-Romero R. M. Yebra-Biurrun M. C. Bermejo-Barrera M. P. Anal. Chim. Acta 1996 327 37. 91 Gaspar A. Posta J. Roth R. J. Anal. At. Spectrom. 1996 11 1067. 92 Di P. Davey D. E. T alanta 1995 42(5) 685. 93 Kradtap S. Tyson J. F. (Dept. Chem. Univ. Massachuse Amherst MA 01003-4510 USA). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 94 Wang A.-X. Liu J.-L. Yu H.-F.Huanjing Huaxue 1995 14(3) 268. 95 Tao G. H. Fang Z. L. At. Spectrosc. 1996 17(1) 22. 96 Sooksamiti P. Geckeis H. Grudpan K. Analyst (Cambridge U.K.) 1996 121(10) 1413. 97 Lima R. Leandro K. C. Santelli R. E. T alanta 1996 43(6) 977. 98 Sperling M. Yan X.-P. Welz B. (Dept. Applied Res. Bodenseewerk Perkin-Elmer GmbH 88647 Ueberlingen Germany). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 99 Lin S.-L. Zheng C.-S. Yun G.-Z. Talanta 1995 42(7) 921. 100 Butcher D. J. Adv. At. Spectrosc. 1995 2 1. 101 Lin S. Zheng Q. Zhu H. Yin G. Fenxi Huaxue 1995 23(7) 835. 102 Ma R.-L. Van Mol W. Adams F. Anal. Chim. Acta 1995 309(1–3) 395. 103 Zhang J. Wang S. Qu W. Huadong L igong Daxue Xuebao 1995 21(4) 507. 104 Zhou J.X. Hou X. D. Yang K. X. Tsai S.-J. J. Michel R. G. (Dept. Chem. Univ. Connecticut Storrs CT 06269 USA). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 105 Sperling et al. Spectrochim Acta Part B 1996 51 1895. 106 Donard O. F. X. (Lab. Chim. Bio-Inorg. Environ. Univ. Pau Pays 1 Pau France). Presented at 15th Nordic Atomic Spectroscopy and Trace Element Conference Ebeltoft Denmark June 28–6 1996. 107 Sooksamiti P. Geckers H. Grudpan K. Analyst (Cambridge U.K.) 1996 121(10) 1413. 108 Husain A. Baroon Z. Al-khalafawi M. Al-Ati T. Sawaya W. Environ. Int. 1995 21(6) 803. 109 Sedykh E. M. Tsysin C. Bannykh L. Sorokina N. Zolotov Yu. (Vernadsky Inst. Geochem. and Anal. Chem. Moscow Russia). Presented at Colloquium Analytische Atomspektroskopie (CANAS ’95) Konstanz Germany 2–7 April 1995.110 Ma and Adams Spectrochim. Acta Part B 1996 51 1917. 111 Hill S. J. Dawson J. B. Price W. J. Shuttler I. L. Tyson J. F. J. Anal. At. Spectrom. 1993 8 197R. 112 van Staden J. F. Hattingh C. J. Anal. Chim. Acta 1995 308(1–3) 214. 113 Mena M. L. Morales Rubio A. Cox A. G. McLeod C. W. Quevauviller P. Quim. Anal. (Barcelona) 1995 14 164. 114 de Koster C. G. Vanwersch J. M. H. Rousseau J. P. G. (Dept. Phys. Anal. and Computational Chem. DSM Res. Netherlands). Presented at European Conference on Energy Dispersive X-ray Spectrometry Lisboa Portugal June 23–29 1996. 115 Chen S.-Y. Zhang M. Lin S.-Q. Peng Z.-H. Huanjing Huaxue 1995 14(2) 169. 116 Lin et al. Spectrochim.Acta Part B 1996 51 1769. 117 Zou H. F. Xu S. K. Fang Z. L. At. Spectrosc. 1996 17(3) 112. 118 Zhang Y.-L. L ihua Jianyan Huaxue Fence 1996 32(1) 16. 119 Lin S.-L. Zheng C.-S. Zu H.-J. Talanta 1995 42(8) 1143. 120 Mir J. M. Jimenez M. S. Castillo J. R. At. Spectrosc. 1995 16(3) 127. Spectrochim. Acta Part B 1996 51B(14) 1935. J. Flow Injection Anal. 1995 12(1) 35. 121 Dittfurth C. Ballesteros E. Gallego M. Valcarcel M. 122 Esmadi F. T. Khasawneh I. M. Kharoaf M. A. Attiyat A. S. 123 Yebra M. C. Gallego M. Valcarcel M. Anal. Chim. Acta 1995 308(1–3) 275. 124 Yebra M. C. Gallego M. Valcarcel M. Anal. Chim. Acta 1995 308(1–3) 357. 125 Hill S. J. Dawson J. B. Price W. J. Riby P. Shuttler I. L. Tyson J. F. J. Anal. At. Spectrom. 1994 9 213R.126 Rivas G. A. Martinez Calatayud J. T alanta 1995 42(9) 1285. 127 Cuesta A. Todoli J. L. Canals A. Spectrochim. Acta Part B 1996 51(14) 1791. 128 Luterotti S. J. Anal. At. Spectrom. 1996 11 973. 129 Youden Steiner Statistical Manual of the Association of Ocial Analytical Chemists Association of Ocial Analytical Chemists Washington DC USA 1975. 130 Jin S. Shiyou Huagong Gaodeng Xuexiao Xuebao 1995 8(2) 7. 131 Majcen N. Fresenius’ J. Anal. Chem. 1996 355(3–4) 315. 132 Kabil M. A. El-Kourashy A.-G. El-Hagrassy M. A. Anal. Methods Instrum. 1995 2(4) 190. 133 Li X. L ihua Jianyan Huaxue Fence 1995 31(4) 229. 134 Yan Z. Fenxi Shiyanshi 1996 15(1) 86. 135 Kabil M. A. El-Kourashy A.-G. El-Hagrasy M. A. J. Anal. At. Spectrom. 1996 11(5) 379.136 Mostafa M. A. J. Anal. At. Spectrom. 1996 11(6) 453. 137 Mostafa M. A. Al-Azhar J. Pharm. Sci. 1994 14 88. 138 Tang Y.-W. Song J.-Y. Mo S.-J. Guangpuxue Yu Guangpu Fenxi 1995 15(4) 1995. 139 Zhang Z.-Z. Zhang W.-M. Zhang J. Fenxi Shiyanshi 1995 15(1) 13. 140 Zhang Z. Z. Zhang W. M. Zhang J. Fenxi Shiyanshi 1996 15(1) 13. 141 Yang M. H. Fang J. L ihua Jianyan Huaxue Fence 1996 32(4) 233. 142 Snigirev A. Rev. Sci. Instrum. 1995 66(2 Pt. 2) 2053. 143 Yu T.-F. Zhou L.-F. Guangpuxue Yu Guangpu Fenxi 1995 15(5) 57. 144 Aydemir T. Gucer S. Anal. L ett. 1996 29(3) 351. 145 Boaventura G. R. Hirson J. da R. Santelli R. E. J. Braz. Chem. Soc. 1995 6(3) 317. 146 Yu T. F. Fenxi Shiyanshi 1996 15(3) 79. 147 He J.-L. Fenxi Huaxue 1995 23(9) 1047.148 Rychlovsky P. Bily J. Bazakasova� E. Collect. Czech. Chem. Commun. 1995 60 76. 149 Dev K. Rao G. N. T alanta 1996 43(3) 451. 150 Vassileva E. Proinova I. Hadjiivanov K. Analyst (Cambridge U.K.) 1996 121(5) 607. 151 Yan Q. Y. L ihua Jianyan Huaxue Fence 1996 32(1) 42. 152 Svoboda L. Safarik T. Schmidt U. Collect. Czech. Chem. Commun. 1995 60(6) 938. 153 Sun S.-Q. Song X. Kang W.-D. Youkuangye 1995 14(1) 43. 154 Stresko V. Medved. J. Polakovicova J. Kubova J. Celkova A. (Fac. Nat. Sci. Comenius Univ. SK-842 15 Bratislava Slovakia). Presented at Colloquium Analytische Atomspektroskopie (CANAS ’95) Konstanz Germany 2–7 April 1995. 155 Nagahiro T. Satake M. Puri B. K. Chim. Acta T urc. 1995 23(2) 125. 156 Satake M. Nagahiro T.Puri B. K. J. Anal. At. Spectrom. 92 7 183. 157 Yu J.-A. Zhang A.-X. Lian N. Fenxi Huaxue 1995 23(6) 740. 158 Koma�rek J. Novotny K. Kuba�n V. (Dept. Anal. Chem. Masaryk Univ. 61137 Brno Czech Republic). Presented at Colloquium Analytische Atomspektroskopie (CANAS ’95) Huaxue Yanjiu Yu Y ingyong 1995 7(4) 382. Konstanz Germany 2–7 April 1995. 159 Yu H.-F. Wang A.-X. Wang Y.-P. Xu S.-S. Yang G.-Q. 160 Bermejo-Barrera P. Dominguez-Gonzalez R. Soto-Ferreiro R. Bermejo-Barrera A. Analusis 1995 23(3) 135. 161 Huang H.-P. Lin S.-L. Fenxi Shiyanshi 1994 13(3) 73. 162 Kosturkova P. Alexandrov S. Pancheva N. Anal. L ab. 1995 4(2) 108. 163 Lee S.-S. Uesugi K. Thoru N. Choi W.-H. Kim K.-T. Choi S.-Y. Anal. Sci. T echnol. 1995 8(3) 391. 164 Khuhawar M.Y. Das P. J. Chem. Soc. Pak. 1996 18(1) 6. 165 Lexa J. Krauler V. Vackova� J. Chem. L isty 1995 89 58. 166 Luo Y.-Y. Al-Othman R. Ruzicka J. Christian G. D. Analyst (Cambridge U.K.) 1996 121(5) 601. 167 Andrade F. J. Tudino M. B. Troccoli O. E. Analyst (Cambridge U.K.) 1996 121(5) 613. Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 168 Koelbl G. Krachler M. Kalcher K. Irgolic K. J. Carapella S. C. (Ed.) Oldfield J. E. (Ed.) Palmieri Y. (Ed.) Palmieri Y. Determination of selenium compounds in biological and environmental samples using HPLC with selenium-specific detection. Proc. Int. Symp. Uses Selenium T ellurium 5th Selenium- Tellurium Dev. Assoc. Grimbergen Belgium 94 291. 169 Quevauviller P. J. Anal. At. Spectrom. 1996 11 1225.M. Fresenius’ J. Anal. Chem. 1995 351(7) 604. M. Anal. Chim. Acta 1996 318 319. M. Fresenius’ J. Anal. Chem. 1996 354(4) 497. 170 Koelbl G. Lintschinger J. Kalcher K. Irgolic K. J.,Mikrochim. Acta 1995 119(1–2) 113. 171 Lintschinger J. Kalcher K. Goessler W. Koelbl G. Novic 172 Vinas P. Campillo N. Lopez Garcia I. Hernandez Cordoba 173 Vinas P. Campillo N. Lopez Garcia I. Hernandez Cordoba 174 Tan Y. Ager P. Marshall W. D. Chan H. M. J. Anal. At. Spectrom. 1996 11 1183. 175 Bermejo-Barrera P. Dominguez-Gonzalez R. Bermejo-Barrera A. Cocho de Juan J. A. Fraga-Bermudez J. M. (Dept. Anal. Chem. Nutrition and Bromatol. Fac. Chem. Univ. Santiago de Compostela Santiago de Compostela Spain). Presented at Eighth Biennial National Atomic Spectroscopy Symposium Univ.East Anglia UK July 17–19 1996. 176 Singer R. L aborPraxis 1995 19(12) 48. 177 Schlemmer G. At. Spectrosc. 1996 17(1) 15. 178 Ivaldi J. C. Carnrick G. Grosser Z. (Perkin-Elmer Corp. Norwalk CT 06859 USA). Presented at Pittsburgh Conference (PITTCON ’96) Chicago IL USA March 3–8 1996. 179 Farah Kddon J. Appl. Spectrosc. Rev. 1995 30(4) 351. 180 Jackson K. W. Chen G. Anal. Chem. 1996 68(12) 231R. 181 Butcher D. J. Vandervoort K. G. Brittain C. T. Lewis B. B. (Dept. Chem. and Phys. Western Carolina Univ. Cullowhee NC 28723 USA). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 182 Vandervoort K. G. Butcher D. J. Brittain C. T. Lewis B. B. Appl. Spectrosc. 1996 50(7) 928. 183 McLain K. N. Vandervoort K.G. Butcher D. J. (Dept. Chem. and Phys. Western Carolina Univ. Cullowhee NC 28723 USA). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 184 Hoenig M. Dheere O. Mikrochim. Acta 1995 119(3–4) 259. 185 Sperling M. Welz B. Hertzberg J. Rieck C. Marowsky G. Spectrochim. Acta Part B 1996 51B 897. 186 Sperling M. Welz B. Hertzberg J. Rieck C. (Bodenseewerk Perkin-Elmer GmbH D-88647 U� berlingen Germany). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. Guangpuxue Yu Guangpu Fenxi 1996 16(1) 99. 187 Smith C. M. M. Harnly J. M. J. Anal. At. Spectrom. 1996 11 1055. 188 Deng B. Luo Y.-F. Wang J.-P. Liu Q. Gao Y.-G. 189 Nobrega J. de A. Silva M. M. Vitoriano de Oliveira P. Krug F. J. Baccan N. Quim. Nova 1995 18(6) 555.190 Volynskii A. B. Spectrochim. Acta Part B 1995 50(11) 1417. 191 Ure T. A. M. Butler L. R. P. L’vov B. V. Rubeska I. Sturgeon R. Pure Appl. Chem. 1992 64 253. 192 Knor I. B. Naumova E. N. Trounova V. A. Dolbnya I. P. Zolotarev K. V. Nucl. Instrum. Methods Phys. Res. Sect. A 1995 358(1,2) 324. 193 Salido A. L. Sanford C. L. Batchelor J. D. Jones B. T. (Dept. Chem. Wake Forest Univ. Winston-Salem NC 27109 USA). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 194 Jones B. T. Sanford C. L. Thomas S. E. Batchelor J. D. (Dept. Chem. Wake Forest Univ. Winston-Salem NC 27109 USA). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 195 Sanford C. L. Thomas S. E. Jones B. T. Appl. Spectrosc. 1996 50(2) 174. 196 Parsons P.J. Qiao H. Aldous K. M. Mills E. Slavin W. Spectrochim. Acta Part B 1995 50(12) 1475. 197 Brockman P. J. Drislane W. F. (Exeter Analytical Inc. N. Chelmsford MA 01863 USA). Presented at Pittsburgh Conference (PITTCON ’96) Chicago IL USA March 3–8 1996. 198 Parsons P. J. Aldous K. Mills E. Qiao H.-C. Slavin W. (Wadsworth Center New York State Dept. Health Albany NY 12201 USA). Presented at Vth COMTOX Symposium on Toxicology and Clinical Chemistry of Metals Vancouver B.C. Canada 10–13 July 1995. 371R 205 Bao C. Cheng X. Changchun Dizhi Xueyuan Xuebao 1995 199 Takeo M. Tsunekage N. Takata E. Maeda Y. Bunseki Kagaku 1995 44(9) 697. 200 Kagaya S. Ueda J. Bull. Chem. Soc. Jpn. 1995 68 2843. 201 Qu R.-L. Her F. Fenxi Shiyanshi 1995 14(6) 53.202 Alvarez M. A. Carrion N. Gutierrez H. Spectrochim. Acta Part B 1995 50(13) 1581. 203 Pyrzynska K. Spectrochim. Acta Part B 1995 50(13) 1595. 204 Carrion N. Alvarez M. A. Quim. Anal. (Barcelona) 1996 15(2) 167. 25(2) 232. 206 Denton M. B. (Dept. Chem. Univ. Arizona Tucson AZ USA). Presented at Pittsburgh Conference (Pittcon ‘96) Chicago IL USA March 3–8 1996. 207 Bermejo-Barrera P. Moreda-Pineiro J. Moreda-Pineiro A. Bermejo-Barrera A. J. Anal. At. Spectrom. 1996 11 1081. 208 Qu R. He F. Fenxi Ceshi Xuebao 1995 14(5) 73. 209 Garbos S. Walcerz M. Bulska E. Hulanicki A. Spectrochim. Acta Part B 1995 50(13) 1669. 210 Ni Z. M. Zhang D. Q. Spectrochim. Acta Part B 1995 50(14) 1779. 211 Haug H. O. Liao Y. P. Spectrochim.Acta Part B 1995 50(11) 1311. 212 Tsalev D. L. D’Ulivo A. Lampugnani L. Di Marco M. Zamboni R. J. Anal. At. Spectrom. 1996 11 19979. 213 Tsalev D. L. D’Ulivo A. Lampugnani L. Di Marco M. Zamboni R. J. Anal. At. Spectrom. 1996 11 989. 214 Haug H. O. Liao Y.-P. (Inst. Tech. Chem. Forschungszentrum Karlsruhe D-76021 Karlsruhe Germany). Presented at Colloquium Analytische Atomspektroskopie (CANAS ’95) Konstanz Germany 2–7 April 1995. 215 Liao Y.-P. Haug H. O. (Inst. Tech. Chem. Forschungszentrum Karlsruhe D-76021 Karlsruhe Germany). Presented at Colloquium Analytische Atomspektroskopie (CANAS ’95) Konstanz Germany 2–7 April 1995. 216 Bermejo-Barrera P. Moreda-Pineiro J. Moreda-Pineiro A. Bermejo-Barrera A. (Dept. Anal. Chem. Nutrition and Bromatol.Fac. Chem. Univ. Santiago de Compostela Santiago de Compostela Spain). Presented at Eighth Biennial National Atomic Spectroscopy Symposium Univ. East Anglia UK July 17–19 1996. 217 Zhou S. X. Guangpuxue Yu Guangpu Fenxi 1996 16(2) 67. T. J. Chromatogr. A 1996 723(2) 337. 223 Matusaki K. Yamaguchi T. Yamamoto Y. Anal. Sci. 1996 218 Ohta Z. Yokoyama M. Ogawa J. Mizuno T. Mikrochim. Acta 1996 122(1–2) 61. 219 Ohta K. Yokoyama M. Ogawa J. Mizuno T. Analusis 1996 24(1) 22. 220 Pedersen-Bjergaard S. Semb S. I. Brevik E. M. Greibrokk 221 Ohta K. Yokoyama M. Ogawa J. Mizuno T. Ann. Chim. (Rome) 1996 86(5–6) 307. 222 Grobecker K. H. (Inst. Reference Materials and Measurements 2440 Geel Belgium). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996.12(2) 301. 224 Slaveykova V. I. Rastegar F. Leroy M. J. F. J. Anal. At. Spectrom. 1996 11 997. 225 Cernohorsky T. Spectrochim. Acta Part B 1995 50(13) 1613. 226 Jones B. T. Batchelor J. D. Levine K. Salido A. Wagner K. (Dept. Chem. Wake Forest Univ. Winston-Salem NC 27109 USA). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 227 Knochen M. Saritsky E. Dol I. Quim. Anal. (Barcelona),1996 15(2) 184. 228 Krug F. J. Silva M. M. Oliveira P. V. Nobrega J. A. Spectrochim. Acta Part B 1995 50(12) 1469. 229 Bruhn C. G. Ambiado F. E. Cid H. J. Wuerner R. Tapia J. Garcia R. Quim. Anal. (Barcelona) 1996 15(2) 197. 230 Gong B.-L. Liu Y.-M. Xu Y.-L. Li Z.-H. Lin T.-Z. Talanta 1995 42(10) 1419.231 Jackson K. W. Giglio J. J. Mahmood T. M. (Wadsworth Center State Univ. New York Albany NY 12201-0509 USA). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 232 Belarra M. A. Lavilla I. Castillo J. R. Analyst (Cambridge U.K.) 1995 120(12) 2813. 233 Schaer U. Krivan V. J. Anal. At. Spectrom. 1996 11 1119. 234 Sonntag T. M. Rossbach M. (Inst. Applied Phys. Chem. Res. Res. Centre Juelich 52425 Juelich Germany). Presented at 372R Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 FACSS XXIII Kansas City MO USA September 29–October 4 1996. 235 Kurfuerst U. Rehnert A. Muntau H. Spectrochim. Acta Part B 1996 51(2) 229. 236 Kurfurst U. (Univ. Fulda 36039 Fulda Germany). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996.237 Luecker E. Schuierer O. Spectrochim. Acta Part B 1996 51(2) 201. 238 Berglund M. Baxter D. C. Mikrochim. Acta 1995 119(3–4) 311. 239 Arruda M. A. Z. Gallego M. Valcarcel M. Quim. Anal. (Barcelona) 1995 14(1) 17. 240 Lucker E. (Inst. Vet. Food Hygiene Justus-Leibig Univ. 35392 Giessen Germany). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 241 de Loos-Vollebregt M. T. C. (Lab. Materials Sci. Univ. Technol. 2628 AL Delft Netherlands). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 242 Kriyan V. (Sektion Anal. Hocstreinigung Univ. Ulm 89069 Ulm Germany). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 243 Miller-Ihli N.J. (US Dept. Agric. Food Comp. Lab. Beltsville Human Nutr. Res. Center Beltsville MD 20705 USA). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 244 Miller-Ihli N. J. (Food Composition Lab. Beltsville Human Nutrition Res. Center Beltsville MD USA). Presented at Eighth Biennial National Atomic Spectroscopy Symposium Univ. East Anglia UK July 17–19 1996. 245 Miller-Ihli N. J. (U.S. Dept. Agric. Food Composition Lab. Beltsville MD 20705 USA). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 246 Wendl W. Huwe A. Maiser T. Muller-Vogt G. (Univ. Karlsruhe 76128 Karlsruhe Germany). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 247 Hinds M. W. Ozaki E. Oliviera E. Curtius A. (Royal Canadian Mint Ottawa Ontario Canada K1A 0G8).Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 248 Robles L. C. Aller A. J. Talanta 1995 42(11) 1731. 249 Bermejo-Barrera P. Moreda-Pineiro A. Romero-Barbeito T. Moreda-Pineiro J. Bormejo-Barrera A. Clin. Chem. (Washington D.C.) 1996 42(8) 1287. 250 Dobrowolski R. (Central Lab. Univ. Maria-Curie-Sklodowska 20-031 Lublin Poland). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 251 Kubota T. Suzuki K. Okutani T. Talanta 1995 42(7) 949. 252 Bermejo-Barrera P. Moreda-Pineiro A. Moreda-Pineiro J. Bermejo-Barrera A. (Dept. Anal. Chem. Nutrition and Bromatol. Fac. Chem. Univ. Santiago de Compostela Santiago de Compostela Spain). Presented at Eighth Biennial National Atomic Spectroscopy Symposium Univ.East Anglia UK July 17–19 1996. 253 Bermejo-Barrera P. Moreda-Pineiro A. Romero-Barbeito T. Moreda-Pineiro J. Bermejo-Barrera A. Talanta 1996 43(7) 1099. 254 Fu X.-Q. Qiu Y.-H. Fenxi Huaxue 1995 23(9) 1109. 255 Lopez-Garcia I. Sanchez-Merlos M. Hernandez-Cordoba M. Anal. Chim. Acta 1996 328 19. 256 Dobrowolski R. Spectrochim. Acta Part B 1996 51(2) 221. 257 Morita Y. Isozaki A. Bunseki Kagaku 1995 44(9) 703. 258 Tsuji K. Sasaki A. Hirokawa K. Jpn. J. Appl. Phys. Part 1 1994 33(11) 6316. 259 Garcia I. L. Vinas P. Campillo N. Cordoba M. H. J. Agric. Food Chem. 1996 44(3) 836. 260 Lopez-Garcia I. Sanchez-Merlos M. Hernandez-Cordoba M. J. Anal. At. Spectrom. 1996 11 1003. 261 Cabrera C. Lorenzo M.L. Lopez M. C. J. AOAC Int. 1995 78(4) 1061. 262 Spellmaeker M. Ouddane B. Fischer J. C. Wartel M. Hoenig M. Analusis 1996 24(3) 76. 263 Shiowatana J. Siripinyanond A. At. Spectrosc. 1996 17(3) 122. 264 Tan Y.-X. Marshall W. D. Blais J.-S. Analyst (Cambridge U.K.) 1996 121(4) 483. 265 Tan Y. Blais J.-S. Marshall W. D. Analyst (Cambridge U.K.) 1996 121(10) 1419. 266 Dobrowolski R. Mierzwa J. Analyst (Cambridge U.K.) 1996 121 897. 267 Dobrowolski R. Mierzwa J. Analyst (Cambridge U.K.) 1996 121(7) 897. 268 Schneider G. Krivan V. Spectrochim. Acta Part B 1995 50(13) 1557. 269 Belarra M. A. Lavilla I. Castillo J. R. Anal. Sci. 1995 11(4) 651. 270 Long G. L. Ronen D. Freedman Y. (Dept. Chem. Virginia Tech. Blacksburg VA 24061-0212 France).Presented at FACSS A. J. Anal. At. Spectrom. 1996 11(8) 571. XXII Cincinnati USA 15–20 October 1995. 271 Freedman Y. E. Ronen D. Long G. L. Environ. Sci. T echnol. 1996 30(7) 2270. 272 Almeida A. A. Lima J. L. F. C. At. Spectrosc. 1995 16 261. 273 Stupar J. Dolinsek F. Spectrochim. Acta Part B 1996 51 665. 274 Smith M. V. Sneddon J. Indurthy S. Lee Y. I. Deval A. (Dept. Chem. McNeese State Univ. Lake Charles LA 70609 USA). Presented at Pittsburgh Conference (PITTCON ’96) Chicago IL USA March 3–8 1996. 275 Indurthy S. Sneddon J. Smith M. Lee Y. I. Deval A. (Dept. Chem. McNeese State Univ. Lake Charles LA 70609 USA). Presented at Pittsburgh Conference (PITTCON ’96) Chicago IL USA March 3–8 1996. 276 Chakrabarti C. L. Marchand B. Vandernoot V.Walker J. Schroeder W. H. Spectrochim. Acta Part B 1996 51(1) 155. 277 Matusiewicz H. Sturgeon R. E. Spectrochim. Acta Part B 1996 51(4) 377. 278 Yao J. Y. Peng L. Q. L ihua Jianyan Huaxue Fence 1996 32(2) 118. 279 Ding W.-W. Sturgeon R. E. J. Anal. At. Spectrom. 1996 11(3) 225. 280 Matusiewicz H. Kopras M. Suszka A. Microchem. J. 1995 52(3) 282. 281 Erber D. Quick L. Winter F. Cammann K. Talanta 1995 42(7) 927. 282 Goenaga Infante H. Fernandez Sanchez M. L. Sanz-Medel 283 Erber D. Cammann K. Analyst (Cambridge U.K.) 1995 120(11) 2699. 284 Tao G.-H. Fang Z.-L. (Flow Injection Anal. Res. Center Inst. Appl. Ecol. Acad. Sin. Shenyang 110015 China). Presented at 6th International Beijing Conference and Exhibition on Instrumental Analysis Beijing China 24–27 October 1995.285 Tao G. Fang Z. Spectrochim. Acta Part B 1995 50(14) 1747. 286 Sperling M. Yan X.-P. Welz B. (Dept. Appl. Res. Bodenseewerk Perkin-Elmer GmbH D-88647 U� berlingen Germany). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 287 Di P. Davey D. E. (Sch. Chem. Technol. Univ. South Australia The Levels SA 5095 Australia). Presented at 10th National Convention of the Royal Australian Chemical Institute 27 September–2 October 1995 Adelaide Australia. 288 Lin S. Zhao C. Yu G. Fenxi Kexue Xuebao 1994 10(1) 24. 289 Yan X.-P. Van Mol W. Adams F. Analyst (Cambridge U.K.) 1996 121(8) 1061. 290 Schuster M. Schwarzer M. Anal. Chim. Acta 1996 328 1. 291 Yuan D.-X. Shuttler I. L. Anal. Chim. Acta 1995 316 313.Chim. Acta 1996 327 295. Organomet. Chem. 1995 9(7) 623. 292 Malcus F. Djane N.-K. Mathiasson L. Johansson G. Anal. 293 d’Haese P. C. Van Landeghem G. F. Lamberts L. V. De Broe M. E. Mikrochim. Acta 1995 120(1–4) 83. 294 Gilon N. Astruc A. Astruc M. Potin-Gautier M. Appl. 295 Gilon N. Potin-Gautier M. Astruc M. (Lab. Chim. Anal. Univ. Pau et Pays L’Adour 64000 Pau France). Presented at Fourth International Symposium on Hyphenated Techniques in Chromatography St John’s Conference Center Belgium February 7–9 1996. 296 Bermejo-Barrera P. Soto-Ferreiro R. M. Dominguez-Gonzalez R. Aboal-Somoza M. Bermejo-Barrera A. Analusis 1996 297 Vinas P. Campillo N. Lopez-Garcia I. Hernandez-Cordoba 298 Vrana A. Komarek J. Fresenius’ J. Anal. Chem.1996 299 Li S.-X. Huang G.-Q. Qian S.-H. Fenxi Kexue Xuebao 1994 24(3) 79. M. Chromatographia 1996 42(9–10) 566. 355(3–4) 321. 10(3) 48. 300 L’vov B. V. Spectrochim. Acta Part B 1996 51(5) 533. 301 Lamoureux M. M. Chakrabarti C. L. Hutton J. C. Gilmutdinov A. Kh. Zakharov Y. A. Gregoire D. C. Spectrochim. Acta Part B 1995 50(14) 1847. 302 Hughes D. M. Chakrabarti C. L. Goltz D. M. Sturgeon R. E. Gregoire D. C. Appl. Spectrosc. 1996 50(6) 715. Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 J. Anal. At. Spectrom. 1996 11(8) 601. J. Anal. At. Spectrom. 1996 11(6) 445. J. Anal. Chem. 1996 354(3) 311. 303 Rojas D. Olivares W. Spectrochim. Acta Part B 1995 50(9) 1011. 304 Rojas D. Spectrochim. Acta Part B 1995 50(9) 1031.305 Slaveykova V. T. Manev S. Lazarov D. Spectrochim. Acta Part B 1995 50(13) 1725. 306 Alvarado J. Quim. Anal. (Barcelona) 1996 15(2) 173. 307 Imai S. Hasegawa N. Hayashi Y. Saito K. J. Anal. At. Spectrom. 1996 11(7) 515. 308 Imai S. Hasegawa N. Nishiyama Y. Hayashi Y. Saito K. 309 Benzo Z. Montero T. Quintal M. Sierraalta A. Ruette F. 310 Mazzucotelli A. Grotti M. Spectrochim. Acta Part B 1995 50(14) 1897. 311 Byrne J. P. Carambassis A. L. Spectrochim. Acta Part B 1996 51(1) 87. 312 Goyal N. Purohit P. J. Page A. G. Sastry M. D. Fresenius’ 313 Tittarelli P. Anselmi A. Bi C. Kmetov V. Spectrochim. Acta Part B 1995 50(13) 1687. 314 Rulon L. L. Robertson J. D. Majidi V. (Dept. Chem. Univ. Kentucky Lexington KY 40506 USA). Presented at FACSS XXII Cincinnati USA 15–20 October 1995.315 Kirkland J. P. Elam W. T. Rev. Sci. Instrum. 1995 66(2 Pt. 2) 1709. 316 Majidi V. Smith R. Pogue R. (Dept. Chem. Univ. Kentucky Lexington KY 40506 USA). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 317 Jackson J. G. Fonseca R. W. Holcombe J. A. Spectrochim. Acta Part B 1995 50(14) 1837. 318 Jackson J. Garg K. Holcombe J. A. (Dept. Chem. and Biochem. Univ. Texas Austin TX 78712 USA). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 319 Lamoureux M. M. Hutton J. C. Styris D. L. Gordon R. L. (Dept. Chem. Saint Mary’s Univ. Halifax NS Canada B3H 3C3). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 320 Giglio J. J. Chen G. Jackson K. W.(Wadsworth Center Sch. Pub. Health State Univ. New York Albany NY 12201-0509 USA). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 321 Chearn V. Lechner J. Desrosiers R. (Natl. Water Res. Inst. Burlington ON Canada L7R 4A6). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 322 Hadgu N. Ohlsson K. E. A. Frech W. Spectrochim. Acta Part B 1995 50(9) 1077. 323isten T. Holcombe J. A. Gu� ell O. A. (Dept. Chem. and Biochem. Univ. Texas Austin TX 78712 USA). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 324 Rayson G. D. Kowit S.-T. (Dept. Chem. and Biochem. New Mexico State Univ. Las Cruces NM 88003 USA). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 325 Pupyshev A. A. Muzgin V. N.J. Anal. Chem. (Engl. T ransl.) 1995 50(7) 632. 326 Croft L. J. Littlejohn D. Marshall J. Gardiner P. H. E. (Dept. Pure and Applied Chem. Univ. Strathclyde Glasgow UK G1 1XL). Presented at Eighth Biennial National Atomic Spectroscopy Symposium Univ. East Anglia UK July 17–19 1996. 327 Hensman C. E. Rayson G. D. (Dept. Chem. and Biochem. New Mexico State Univ. Las Cruces NM 88003 USA). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 328 Grotti M. Magi E. Leardi R. Anal. Chim. Acta 1996 327 47. 329 Lelievre C. Hennequin D. Barillier D. Lemasson F. Analusis 1996 24(2) 39. 330 Zong Y.-Y. Parsons P. J. Slavin W. J. Anal. At. Spectrom. 1996 11(1) 25. 331 Cabon J. Y. Le Bihan A. Spectrochim. Acta Part B 1996 51 619.332 Wieteska E. Drzewinska A. Chem. Anal. (Warsaw) 1995 40(2) 207. 333 Nixon D. E. Moyer T. P. (Metals Lab. Div. Clin. Biochem. and Immunol. Mayo Clinic Rochester MN 55905 USA). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 334 LeBlanc A. J. Anal. At. Spectrom. 1996 11 1093. 335 Durfee L. D. (Dow Corning Corp. Midland MI 48686-0994 USA). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 373R 336 Akman S. Do� ner G. Spectrochim. Acta Part B 1995 50(9) 19975. 337 Doner G. Akman S. Spectrochim. Acta Part B 1996 51(1) 181. 338 Heitmann U. SchBtz M. Becker-Roa H. Florek S. (Gesellschaft Forderung Angewandter Optik und Spektroskopie (GOS) e.V. 12489 Berlin Germany). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996.339 Shepard M. R. Jones B. T. Butcher D. J. (Dept. Chem. Wake Forest Univ. Winston-Salem NC 27109 USA). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 340 Volynsky A. B. (Section Anal. and Hoechstreinigung Univ. Ulm 89069 Ulm Germany). Presented at Eighth Biennial National Atomic Spectroscopy Symposium University of East Anglia UK July 17–19 1996. 341 Volynsky A. B. Krivan V. (Sektion Anal. und Hoechstreinigung Univ. Ulm 89069 Ulm Germany). Presented at Eighth Biennial National Atomic Spectroscopy Symposium Univ. East Anglia UK July 17–19 1996. 342 Jackson K. W. Zhang D. Mahmood T. M. Castillano M. T. (Wadsworth Center Sch. Pub. Health State Univ. New York Albany NY 12201-0509 USA). Presented at FACSS XXII Cincinnati USA 15–20 October 1995.343 Jackson K. W. Chen G. (Wadsworth Center Sch. Pub. Health State Univ. NY and NY State Dept. Health P.O. Box 509 Albany NY 12201-0509 USA). Presented at Federation of Analytical Chemistry and Spectroscopy Societies (1994 FACSS) St Louis Missouri USA October 2–7 1994. 344 Havezov I. Detcheva A. Rendl J. Mikrochim. Acta 1995 119(1–2) 147. 345 Volynsky A. B. Krivan V. J. Anal. At. Spectrom. 1996 11(2) 159. 346 He B. Ni Z.-M. J. Anal. At. Spectrom. 1996 11(2) 165. 347 Huang S.-D. Lai W.-R. Shih K.-Y. Spectrochim. Acta Part B 1995 50(10) 1237. 348 Raith A. Sigsworth P. T. Henmry R. (Fisons Instruments Elemental Analysis Winsford Cheshire UK). Presented at Pittsburgh Conference (PITTCON ‘96) Chicago IL USA March 3–8 1996.349 Bermejo-Barrera P. Moreda-Pineiro J. Moreda-Pineiro A. Bermejo-Barrera A. Fresenius’ J. Anal. Chem. 1996 355(2) 174. 350 Bermejo-Barrera P. Moreda-Pineiro J. Moreda-Pineiro A. Bermejo-Barrera A. Mikrochim. Acta 1996 124(1–2) 111. 351 Skwara W. Pszonicki L. (Inst. Nucl. Chem. and Technol. Warsaw Poland). Presented at 5th Polish Conference on Analytical Chemistry Gdansk Poland 3–9 September 1995. 352 Sedykh E. M. Volynsky A. B. Baranova N. N. (Vernadsky Inst. Geochem. and Anal. Chem. 117334 Moscow Russia). Presented at Colloquium Analytische Atomspektroskopie (CANAS ’95) Konstanz Germany 2–7 April 1995. Blank A. B. J. Anal. Chem. (Engl. T ransl.) 1995 50(9) 914. C. E. J. Anal. At. Spectrom. 1996 11(1) 31. 353 Thomaidis N.S. Piperaki E. A. Analyst (Cambridge U.K.) 1996 121(2) 111. 354 Zolotovitskaya E. S. Potapova V. G. Grebenyuk N. N. 355 Zhang J. Guo S.-X. Analyst (Cambridge U.K.) 1995 120(6) 1661. 356 Hanafi M. M. Saifulbahari A. R. Peli M. Commun. Soil Sci. Plant Anal. 1996 27(5–8) 1479. 357 Thomaidis N. S. Piperaki E. A. Polydorou C. K. Efstathiou 358 Li H. T. Zheng L. M. Guangpuxue Yu Guangpu Fenxi 1995 15(5) 61. 359 Kantor T. Spectrochim. Acta Part B 1995 50(13) 1599. 360 Li G.-K. Zhang Z.-X. Yang X.-H. Guangpuxue Yu Guangpu Fenxi 1995 15(3) 91. 361 Alemasova A. S. Shevchuk I. A. Rybak I. A. Ukr. Khim. Zh. (Russ. Ed.) 1995 61(1–2) 48. 362 Rozanska B. Piwko A. (Dept. Anal. Chem. Warsaw Univ. Technol. Warsaw Poland). Presented at 5th Polish Conference on Analytical Chemistry Gdansk Poland 3–9 September 1995.363 Kim Y.-S. Choi J.-M. Kim Y.-M. Anal. Sci. T echnol. 1995 8(4) 435. 364 Van Dalen G. J. Anal. At. Spectrom. 1996 11 1087. 365 Ashino T. Takada K. J. Anal. At. Spectrom. 1996 11(8) 577. 366 Bermejo-Barrera P. Annlo-Sendin R. M. Aboal-Somoza M. Bermejo-Barrera A. (Dept. Anal. Chem. Nutrition and Bromatol. Fac. Chem. Univ. Santiago de Compostela Santiago de Compostela Spain). Presented at Eighth Biennial National Atomic Spectroscopy Symposium Univ. East Anglia UK July 17–19 1996. 374R Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 Contam. T oxicol. 1996 56 860. Fenxi Shiyanshi 1995 14(1) 64. 367 Heinemann G. Vogt W. Clin. Chem. (Washington D.C.) 1996 42(8) 1275.368 Gaggi C. Zino F. Duccini M. Renzoni A. Bull. Environ. 369 Bermejo-Barrera P. Moreda-Pineiro J. Moreda-Pineiro A. Bermejo-Barrera A. Talanta 1996 43 35. 370 Ou H. Hua L. Gong H.-L. Xie Y.-S. He H.-K. Fenxi Shiyanshi 1995 14(5) 66. 371 Xie W.-B. Yao J.-Y. Hu Q.-L. Meng X.-M. Lian C.-Z. 372 Xie W.-B. Yao J.-Y. Hu Q.-L. Y ingyong Huaxue 1995 373 Van Dael P. Van Cauwenbergh R. Robberecht H. Deelstra 374 Gottelt U. Henrion G. Kalaehne R. Stoyke M. Nahrung 375 Bulska E. Chelmecki G. Hulanicki A. Can. J. Appl. Spectrosc. 376 Sun D.-J. Wang D.-H. Ma Y.-Z. Huanjing Huaxue 1995 377 Tsai S.-J. J. Chang L.-L. (Dept. Appl. Chem. Providence 12(4) 108. H. A. Calomme M. At. Spectrosc. 1995 16(6) 251. 1996 40(2) 87. 1996 41(1) 5. 14(6) 524.Univ. Taichung Hsien 43301 Taiwan). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 378 Dudek J. Pszonicki L. (Inst. Nucl. Chem. and Technol. Warsaw Poland). Presented at 5th Polish Conference on Analytical Chemistry Gdansk Poland 3–9 September 1995. 379 Xuan W. Eects of Pd and Mg nitrate on the atomization of Ge in graphite furnace atomic absorption spectrometry Report ISTIC-TR-93019; Order No. PB94-166493 1993 17. 380 Julshamn K. Maage A. Larsen E. H. Fresenius’ J. Anal. Chem. 1996 355(3–4) 304. 381 Choi H.-S. Choi J.-M. Kim Y.-S. J. Korean Chem. Soc. 1996 40(2) 109. 382 Rock P. E. Booth P. K. (ATI Unicam Cambridge UK CB1 2PX). Presented at 5th Polish Conference on Analytical Chemistry Gdansk Poland 3–9 September 1995.383 Rock P. E. Tuckerman R. (ATI Unicam Cambridge UK CB1 2PX). Presented at Colloquium Analytische Atomspektroskopie (CANAS ’95) Konstanz Germany 2–7 April 1995. 384 Tsalev D. L. Slaveykova V. I. Georgieva R. B. Anal. L ett. 1996 29(1) 73. 385 Li H. Nagasawa H. Matsumoto K. Anal. Sci. 1996 12(2) 215. 386 Gilmutdinov A. Kh. Sperling M. Welz B. (Dept. Appl. Res. Bodenseewerk Perkin-Elmer GmbH D-88647 U� berlingen Germany). Presented at Colloquium Analytische Atomspektroskopie (CANAS ’95) Konstanz Germany 2–7 April 1995. 387 Gilmutdinov A. Kh. Mrasov R. M. Somov A. R. Chakrabarti C. L. Hutton J. C. Spectrochim. Acta Part B 1995 50(13) 1637. 388 Welz B. Gilmutdinov A. Kh. Sperling M. (Dept. Applied Res. Bodenseewerk Perkin-Elmer GmbH 88647 Uberlingen Germany).Presented at 5th Na Conference on Atomic Spectroscopy May 12–16 1996 Wuhan China. Spectrochim. Acta Part B 1996 51B 1023. Spectrochim. Acta Part B 1995 50(13) 1679. 389 Gilmutdinov A. Kh. Radziuk B. Sperling M. Welz B. Nagulin K. Yu. Spectrochim. Acta Part B 1996 51B 931. 390 Gilmutdinov A. Kh. Radziuk B. Sperling M. Welz B. 391 Belarra M. A. Resano M. Castillo J. R. Spectrochim. Acta Part B 1996 51 697. 392 Torsi G. Valcher S. Reschiglian P. Cludi L. Patauner L. 393 Bezsudnov I. V. Sevryukov V. A. Tatsy Y. G. (Sci. and Prod. Enterprise ‘Nauka-Service’ 103473 Moscow Russia). Presented at Pittsburgh Conference (PITTCON ’96) Chicago IL USA March 3–8 1996. 394 Levine K. Wichems D. N. Salido A. L. Jones B. T. (Dept. Chem.Wake Forest Univ. Winston-Salem NC 27109 USA). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 395 Demers D. R. Almeida M. C. Am. Environ. L ab. 1995 7(5) 13. 396 Edel H. Quick L. Cammann K. Anal. Chim. Acta 1995 310(1) 181. 397 White M. A. Panayi A. Iversen B. S. Sabbioni E. (Environ. Inst. European Commission Joint Res. Centre Ispra Italy). Presented at Eighth Biennial National Atomic Spectroscopy Symposium Univ. East Anglia UK July 17–19 1996. 398 Iversen B. S. Panayi A. Camblor J. P. Sabbioni E. J. Anal. At. Spectrom. 1996 11(8) 591. 399 Schlemmer G. Feuerstein M. Portala F. (Bodenseewerk Perkin-Elmer GmbH D-88647U� berlingen Germany). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 400 Carnrick G. Schlemmer G.Portala F. Feuerstein M. (Perkin- Elmer Corp. Norwalk CT 06859 USA). Presented at Pittsburgh Conference (PITTCON ’96) Chicago IL USA March 3–8 1996. 401 Schrader W. Shuttler I. L. Schlemmer G. Portala F. (Bodenseewerk Perkin-Elmer GmbH D-88647 U� berlingen Germany). Presented at 6th International Beijing Conference and Exhibition on Instrumental Analysis Beijing China 24–27 October 1995. 402 Schlemmer G. Shuttler I. L. (Bodenseewerk Perkin-Elmer GmbH D-1-887647 U� berlingen Germany). Presented at 10th National Convention of the Royal Australian Chemical Institute 27 September–2 October 1995 Adelaide Australia. 403 Sturgeon R. E. (Inst. Natl. Measurement Standards Natl. Res. Council Canada Ottawa Canada K1A 0R9). Presented at The 15th Nordic Atomic Spectroscopy and Trace Element Conference Ebeltoft Denmark June 2–6 1996.404 Pavski V. Sturgeon R. E. Chakrabarti C. L. (Dept. Chem. Ottawa-Carleton Chem. Inst. Carleton Univ. Ottawa Ontario Canada K1S 5B6). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 405 Le Blanc C. W. Blades M. W. (Dept. Chem. Univ. British Columbia Vancouver BC Canada V6T 1Z1). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 406 Hettipathirana T. D. Blades M. W. (Dept. Chem. Univ. British Columbia Vancouver British Columbia Canada V6T 1N4). Presented at 10th National Convention of the Royal Australian Chemical Institute 27 September–2 October 1995 Adelaide Australia. 407 L’vov B. V. Fresenius’ J. Anal. Chem.1996 355(3–4) 222. 408 Yang W. M. Ni Z. M. Spectrochim. Acta Part B 1996 51(1) 65. 409 Long S. Liu D. Yin H. Guangpuxue Yu Guangpu Fenxi 1996 16(1) 107. 410 Sturgeon R. E. Liu J. Boyko V. J. Luong V. T. Anal. Chem. 1996 68(11) 1883. 411 Tyson J. F. Carrero P. Ellis R. I. (Dept. Chem. Univ. Massachusetts Amherst MA 01003 USA). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 412 Nakahara T. Adv. At. Spectrosc. 1995 2 139. 413 Narasaki H. CACS Forum 1995 15 9. 414 Fang Z. L. Tao G. H. Xu S. K. Liu X. Z. Wang J. Microchem. J. 1996 53(1) 42. 415 Howard A. G. (Chem. Dept. Univ. Southampton Southampton UK). Presented at Eighth Biennial National Atomic Spectroscopy Symposium Univ. East Anglia UK July 17–19 1996. 416 Howard A.G. J. Anal. At. Spectrom. 1997 12(3) 267. 417 Beinrohr E. Mikrochim. Acta 1995 120(1–4) 39. 418 Ding W.-W. Sturgeon R. E. J. Anal. At. Spectrom. 1996 11(6) 421. 419 Ridgway T. H. (Dept. Chem. Univ. Cincinnati Cincinnati OH 45221 USA). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 420 Schaumloeel D. Neidhart B. Fresenius’ J. Anal. Chem. 1996 354(7–8) 866. 421 Celkova A. Kubova J. Stresko V. Fresenius’ J. Anal. Chem. 1996 355(2) 150. 422 Yin X. Homann E. Luedke C. Fresenius’ J. Anal. Chem. 1996 355(3–4) 324. 423 Pohl B. Weichbrodt G. Fraunhofer S. GIT Fachz. L ab. 1995 39(12) 1134. 424 Chen H. W. Yao W. Wu D. X. Brindle I. D. Spectrochim. Acta Part B 1996 51B(14) 1829. 425 Mierzwa J. Dobrowolski R. (Central Lab.Univ. Maria 20-031 Lublin Poland). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 426 Wickstrom T. Lund W. Bye R. Analyst (Cambridge U.K.) 1996 121(2) 201. 427 Wickstrom T. Lund W. Bye R. Analyst (Cambridge U.K.) 1995 120(11) 2695. 428 Vuchkova L. Arpadjan S. T alanta 1996 43 479. 429 Terada S. Kawabata J. Yamashita N. Jpn. Kokai Tokkyo Koho JP 06,223,750 [94,223,750] (Cl. H01J35/10) 12 Aug 1994 Appl. 93/10,254 25 Jan 1993 7 pp. 430 Hanna C. P. McIntosh S. A. Tyson J. F. Sundin N. (Perkin- Elmer Corp. Norwalk CT 06859-0219 USA). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 431 Lopez-Gonzalvez M. A. Milagros Gomez M. Camara C. Palacios M. A. Mikrochim. Acta 1995 120(1–4) 301. Journal of Analytical Atomic Spectrometry August 1997 Vol.12 432 Burguera J. L. Carrero P. Burguera M. Rondon C. Brunetto M. R. Gallignani M. Spectrochim. Acta Part B 1996 51B(14) 1837. 433 Gonzalez LaFuente J. M. Fernandez Sanchez M. L. Marchante-Gayon J. M. Sanchez Uria J. E. Sanz-Medel A. Spectrochim. Acta Part B 1996 51B(14) 1849. 434 Siska R. Borszeki J. Gegus E. Can. J. Appl. Spectrosc. 1995 40(5) 117. 435 Zhang B.-G. Wang Y. Wang X.-S. Chen X.-L. Feng J.-X. T alanta 1995 42(8) 1095. 436 Ellis R. I. Tyson J. F. Spectrochim. Acta Part B 1996 51B(14) 1859. 437 Shrader D. Moett J. Vanclay E. (Varian Optical Spectroscopy Instruments Wood Dale IL 60191 USA). Presented at Pittsburgh Conference (PITTCON ’96) Chicago IL USA March 3–8 1996. 438 Laborda F.Gomez M. T. Jimenez M. S. Mir J. M. Castillo J. R. J. Anal. At. Spectrom. 1996 11 1121. M. Nara-ken Eisei Kenkyusho Nenpo 1995 29 140. 439 Barnes B. S. (Forensic Chem. Center US Food and Drug Admin. Cincinnati OH 45202 USA). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 440 Kumar S. J. Meeravali N. N. At. Spectrosc. 1996 17(1) 27. 441 Jimenez de Blas O. Rodriguez Mateos N. Garcia Sanchez A. J. AOAC Int. 1996 79(3) 764. 442 Narasaki H. Cao J. Y. Microchem. J. 1996 53(1) 18. 443 Bettinelli M. Spezia S. Baroni U. Bizzarri G. Proc. Chem. Conf. 1993 45 67. 444 Shrader D. Vanclay E. (Varian Optical Spectroscopy Instruments Wood Dale IL 60191 USA). Presented at Pittsburgh Conference (PITTCON ’96) Chicago IL USA March 3–8 1996.445 Wei L. Gupta P. Farhat F. Duboise S. (Chem. Services Div. Bur. Lab. Texas Dept. Health Austin TX 87856-3199 USA). Presented at Pittsburgh Conference (PITTCON ’96) Chicago IL USA March 3–8 1996. 446 Fernando R. Goldberg M. Pellizzari E. D. Sheldon L. S. Lang M. (Res. Triangle Inst. Research Triangle Park NC 27709 USA). Presented at Pittsburgh Conference (PITTCON ’96) Chicago IL USA March 3–8 1996. 447 Siska R. Borszeki J. Gegus E. Magy. Kem. Foly. 1996 63(1) 248. 448 Fisera M. Hladky Z. Risova J. Chem. Pap. 1996 50(1) 8. 449 Tanaka T. Aoki Y. Ujike E. Okayama A. Tahara S. Sasaki 450 Chen. P. H. Chen W. Q. Zhang J. H. Dai S. G. Fenxi Huaxue 1996 24(1) 69. 451 Teague S. M. Wang H.-C. Davis M. K. (Metropolitan Water District of Southern California La Verne CA 91750 USA).Presented at Pittsburgh Conference (PITTCON ’96) Chicago 452 Stummeyer J. Harazim B. Wippermann T. Fresenius’ J. Anal. 453 Velez D. Ybanez N. Montoro R. J. Anal. At. Spectrom. 1996 454 Zhang X. Cornelis R. de Kimpe J. Mees L. J. Anal. At. 455 Zhang X. Cornelis R. De Kimpe J. Mees L. Anal. Chim. 456 Velez D. Ybanez N. Montoro R. J. Anal. At. om. 1997 IL USA March 3–8 1996. Chem. 1996 354(3) 344. 11(4) 271. Spectrom. 1996 11 1075. Acta 1996 319 177. 12(1) 91. 457 Le X. C. Ma M. Wong N. A. Anal. Chem. 1996 68 4501. 458 Lamble K. J. Hill S. J. Anal. Chim. Acta 1996 334 261. 459 Rubio R. Alberti J. Padro A. Rauret G. T rends Anal. Chem. 1995 14(6) 274. 460 Ellis R. Tyson J. Kradtap S. Vargas C.McIntosh S. Hanna C. P. (Dept. Chem. Univ. Massachusetts Amherst MA 01003 USA). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 461 Duan Y. X. Zhang H. Q. Jiang X. M. Jin Q. H. Spectrosc. L ett. 1996 29(1) 69. 462 Yi J. Huanjing Baohu (Beijing) 1995(11) 25. 463 Madrid Y. Chakraborti D. Camara C. Mikrochim. Acta 1995 120(1–4) 63. 464 Fang Y. (Fuzhou Health and Quarantine Bur. Fuzhou Fujian 350001 China). Presented at Pittsburgh Conference (PITTCON ’96) Chicago IL USA March 3–8 1996. 465 Yin S. H. L ihua Jianyan Huaxue Fence 1996 32(4) 238. 466 Rondon C. Burguera J. L. Burguera M. Brunetto M. R. Gallignani M. Petit de Pena Y. Fresenius’ J. Anal. Chem. 1995 353(2) 133. 375R T. J. Anal. At. Spectrom. 1997 12(1) 85.467 Navarro M. Lopez H. Lopez M. C. Perez V. J. AOAC Int. 1996 79(3) 773. 468 Navarro M. Lopez H. Ruiz M. L. Gonzalez S. Perez V. Lopez M. C. Sci. T otal Environ. 1995 175(3) 245. 469 Hao D.-Q. Xie G.-H. Zhang Y.-M. Tian G.-J. T alanta 1996 43 595. 470 Mestek O. Suchanek M. Vodickova Z. Zemanova B. Zima 471 Foster L. H. Sumar S. Food Chem. 1996 55(3) 293. 472 Cha K.-W. Park S.-H. Park K.-W. Anal. Sci. T echnol. 1995 8(4) 419. 473 Zhou Q.-W. Zheng Y.-Z. Fenxi Shiyanshi 1995 14(6) 49. 474 Nielsen S. Sloth J. J. Hansen E. H. Analyst (Cambridge U.K.) 1996 121(1) 31. 475 Tao G. Hansen E. H. Analyst (Cambridge U.K.) 1994 119(2) 333. 476 Vela N. P. Caruso J. A. J. Anal. At. Spectrom. 1996 11 1129. 477 Lei T. Marshall W. D. Appl. Organomet.Chem. 1995 9(2) 149. 478 Burguera M. Burguera J. L. Rivas C. Carrero P. Brunetto R. Gallignani M. Anal. Chim. Acta 1995 308(1–3) 339. 479 Sun H.-W. Zhang D.-Q. Yang L.-L. (Dept. Chem. Hebei Univ. Baoding 071002 China). Presented at Proceedings of the 2nd International Symposium of Worldwide Chinese Scholars on Analytical Chemistry Shenzhen China November 15–18 1995. 480 Golloch A. Goetzen A. (Gerhard-Mercator Univ. D-47048 Duisburg Germany). Presented at 6th International Beijing Conference and Exhibition on Instrumental Analysis Beijing China 24–27 October 1995. 481 L aborPraxis 1996 20(2) 42. 482 Stockwell P. B. Corns W. T. (PS Analytical Ltd. Orpington Kent UK BR5 3HP). Presented at Pittsburgh Conference (PITTCON ’96) Chicago IL USA March 3–8 1996.483 Pfeil D. Ballantyne K. (Thermo Jarrell Ash Franklin MA 02038 USA). Presented at Pittsburgh Conference (PITTCON ’96) Chicago IL USA March 3–8 1996. 484 Shrader D. (Varian OSI Victoria 3170 Australia). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 485 Zimmerman C. L. Rivers J. (CETAC Technol. Inc. Omaha NE 68107 USA). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 486 Brindle I. D. Zheng S. Spectrochim. Acta Part B 1996 51B(14) 1777. 487 Rodriguez M. C. Hernandez M. Sanchez J. M. Cubillan H. S. Semprn B. I. Barrios L. C. Granadillo V. A. (Anal. Instrumentation Lab. Univ. Zuila Maracaibo 4011 Venezuela). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996.488 Sanchez J. M. Rodriguez M. C. Cubillan H. S. Hernandez M. Semprn B. I. Barrios L. C. Granadillo V. A. (Anal. Instrumentation Lab. Fac. Sci. Univ. Zuila Maracaibo 4011 Venezuela). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 489 Sasaki K. Pacey G. E. (Dept. Chem. Miami Univ. Oxford OH 45056 USA). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 490 McIntosh S. A. Hanna C. P. (Perkin-Elmer Corp. Norwalk CT 06859-0219 USA). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 491 Schlemmer G. Erler W. (Bodenseewerk Perkin-Elmer GmbH D-88647 U� berlingen Germany). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 492 Adeloju S. B. Dhindsa H. S. (Centre Electrochem. Res.and Anal. Technol. Dept. Chem. Univ. Western Sydney Kingswood NSW 2747 Australia). Presented at 10th National Convention of the Royal Australian Chemical Institute 27 September–2 October 1995 Adelaide Australia. 493 Murphy J. Jones P. Hill S. J. (Dept. Environ. Sci. Univ. Plymouth Plymouth Devon UK PL4 8AA). Presented at Eighth Biennial National Atomic Spectroscopy Symposium Univ. East Anglia UK July 17–19 1996. 494 Brahma N. Stockwell P. B. Corns W. T. Evans E. H. Ebdon L. (PS Anal. Ltd. Orpington Kent UK BR5 3HP). Presented at Eighth Biennial National Atomic Spectroscopy Symposium Univ. East Anglia UK July 17–19 1996. 495 Bloxham M. J. Hill S. J. Worsfold P. J. J. Anal. At. Spectrom. 1996 11(7) 511. 496 Wei L. H. Fenxi Huaxue 1996 24(2) 247.497 Hu G.-L. L ihua Jianyan Huaxue Fence 1995 31(6) 366. 376R Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 498 Lamble K. J. Hill S. J. (Dept. Environ. Sci. Univ. Plymouth Plymouth UK PL4 8AA). Presented at Eighth Biennial National Atomic Spectroscopy Symposium Univ. East Anglia UK July 17–19 1996. 499 Lamble K. J. Hill S. J. J. Anal. At. Spectrom. 1996 11 1099. 500 Murphy J. Jones P. Hill S. J. Spectrochim. Acta Part B 1996 51B(14) 1867. 501 Schnitzer G. Soubelet A. Testu C. Chafey C. Mikrochim. Acta 1995 119(3–4) 199. 502 Fostier A. H. Ferreira J. R. Oeterrer de Andrade M. Quim. Nova 1995 18(5) 425. 503 Guo T. Baasner J. Gradl M. Kistner A. Anal. Chim. Acta Lester J. N. Int. J. Environ. Anal. Chem. 1996 63(3) 187.Chim. Acta 1996 324 69. 1996 320(2–3) 171. 504 Gu H.-Y. Fenxi Ceshi Xuebao 1995 14(2) 76. 505 Mniszek W. Chem. Anal. (Warsaw) 1996 41(2) 269. 506 Feng J. Pan Z. Liu H. Fenxi Huaxue 1996 24(1) 74. 507 Ombaba J. M. Microchem. J. 1996 53(2) 195. 508 Antonovich V. P. Bezlutskaya I. V. Zh. Anal. Khim. 1996 51(1) 116. 509 Jones R. D. Jacobson M. E. Jae R. West-Thomas J. Arfstrom C. Alli A. Water Air Soil Pollut. 1995 80(1–4) 1285. 510 Edwards S. C. MacLeod C. L. Corns W. T. Williams T. P. 511 Nixon D. E. Mussmann G. V. Moyer T. P. J. Anal. T oxicol. 1996 20(1) 17. 512 Bryce D. W. Izquierdo A. Luque de Castro M. D. Anal. 513 Yoshino M. Tanake H. Okamoto K. Bunseki Kagaku 1995 44(9) 691. 514 Falter R. Schoeler H.-F. Fresenius’ J. Anal. Chem.1996 354(4) 492. 515 Ni Z.-M. Guangpuxue Yu Guangpu Fenxi 1995 15(6) 87. 516 Welz B. Sucmanova M. (Dept. Applied Res. Bodenseewerk Perkin-Elmer GmbH 88662 Ueberlingen Germany). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 517 Fischer R. G. Rapsomanikis S. Andreae M. O. (Biogeochem. Dept. Max Planck Inst. Chem. 55020 Mainz Germany). Presented at Pittsburgh Conference (PITTCON ’96) Chicago IL USA March 3–8 1996. 518 Holak W. J. AOAC Int. 1995 78(4) 1124. W. Spectrochim. Acta Part B 1996 51 829. T rends Anal. Chem. 1996 15(4) 181. O. F. X. Analyst (Cambridge U.K.) 1995 120(11) 2665. 519 Emteborg H. Sinemus H. W. Radziuk B. Baxter D. C. Frech 520 Szpunar J. Schmitt V. O. Donard O. F. X. Lobinski R. 521 Shawky S.Emons H. Du�rbeck H. W. Anal. Commun. 1996 33(3) 107. 522 Xu F.-Z. Jiang G.-B. Han H.-B. Fenxi Huaxue 1995 23(11) 1308. 523 Lalere B. Szpunar J. Budzinski H. Garrigues P. Donard 524 Ohta K. Koike Y. Mizuro T. Anal. Chim. Acta 1996 329 191. 525 Honda A. Jpn. Kokai Tokkyo Koho JP 07,128,227 [95,128,227] (Cl. G01N21/31) 19 May 1995 Appl. 93/293,888 30 Oct 1993; 6 pp. 526 Karasawa H. Kojima S. Tagushi G. Jpn. Kokai Tokkyo Koho JP 07,151,675 [95,151,675] (Cl. G01N21/31) 16 Jun 1995 Appl. 93/299,411 30 Nov 1993; 3 pp. 527 Nakano T. Jpn. Kokai Tokkyo Koho JP 07,128,228 [95,128,228] (Cl. G01N21/31) 19 May 1995 Appl. 93/294,740 29 Oct 1993; 7 pp. 528 Satsukawa N. Hioki S. Kubota M. Jpn. Kokai Tokkyo Koho JP 07,190,926 [95,190,926] (Cl.G01N21/31) 28 Jul 1995 Appl. 93/353,159 27 Dec3; 4 pp. 529 Huber B. Ger. Oen. DE 4,411,441 (Cl. G01N21/71) 5 Oct 1995 Appl. 4,411,441 31 Mar 1994; 12 pp. 530 Baird W. Nogar N. S. Appl. Spectrosc. 1995 49(11) 1699. 531 Aldous K. M. Mills E. Slavin W. Qiao H. Parsons P. J. (Wadsworth Center New York State Dept. Health Albany NY 12201-0509 USA). Presented at Pittsburgh Conference (PITTCON ’96) Chicago IL USA March 3–8 1996. 532 Aldous K. M. Mills E. Slavin W. Qiao H. Parsons P. J. (Wadsworth Center New York State Dept. Health Albany NY 12201-0509 USA). Presented at Eighth Biennial National Atomic Spectroscopy Symposium Univ. East Anglia UK July 17–19 1996. 533 Batchelor J. D. Thomas S. E. Jones B. T. (Dept. Chem. Wake Forest Univ. Winston-Salem NC 27109 USA).Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 534 Salido A. Jones B. T. (Dept. Chem. Wake Forest Univ., Winston-Salem NC 27109 USA). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 535 Barker S. A. Johnson S. G. Knighton G. C. Sayer M. T. Candee B. M. Dimick V. D. Appl. Spectrosc. 1996 50(6) 816. 536 Hollocher K. Ruiz J. Geostand. Newsl. 1995 19(1) 27. 537 Liang F. Mei E. Zeng X. Wang L. Chen G. Fenxi Kexue Xuebao 1995 11(2) 56. 538 Carnahan J. W. Bucher E. Gillespie S. Spudich T. (Dept. Chem. Northern Illinois Univ. DeKalb IL 60115 USA). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 539 Doidge P. S. Spectrochim. Acta Part B 1996 50(11) 1421. 540 Doidge P.S. Spectrochim. Acta Part B 1995 50(3) 209. 541 Sansonetti C. J. Salit M. L. Reader J. Appl. Opt. 1996 35(1) 74. 542 Reader J. Sansonetti C. J. Bridges J. M. Appl. Opt. 1996 35(1) 78. 543 Ball D. W. Spectroscopy (Eugene Oreg.) 1996 11(1) 29. 544 Pavski V. Chakrabarti C. L. Appl. Spectrosc. 1995 49(7) 927. 545 Gilmutdinov A. Kh. Radziuk B. Sperling M. Welz B. Nagulin K. Yu. (Dept. Appl. Res. Bodenseewerk Perkin-Elmer GmbH D-88647 U� berlingen Germany). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 546 Williams J. C. Kung J.-Y. Chen Y.-X. Cai X.-J. Grin S. T. Appl. Spectrosc. 1995 49(11) 1705. 547 Wycislik A. Gralewska K. Chemik 1995 48(12) 350. 548 Takasaka M. Shimadzu Hyoron 1995 52(2) 129. 549 Moett J. (Varian Australia Mulgrave VIC 3170 Australia).Presented at 10th National Convention of the Royal Australian Chemical Institute 27 September–2 October 1995 Adelaide Australia. 550 Shrader D. Moett J. Vanclay E. (Varian Optical Spectroscopy Instruments 201 Hansen Ct. Ste. 108 Wood Dale IL 60191 USA). Presented at Pittsburgh Conference (Pittcon ’95) New Orleans LA USA March 5–10 1995. 551 Qiu X. Zhong M. Huang Z. Huaxue Shijie 1995 36(3) 150. 552 Telle H. H. Acosta Ortiz S. E. Proc. SPIE-Int. Soc. Opt. Eng. 1996 2730 (Second Iberoamerican Meeting on Optics 1995) 34. 553 Weed K. M. Maniaci M. J. Tong W. G. Proc. SPIE-Int. Soc. Opt. Eng. 1995 2385 157. 554 Smith C. M. M. (Univ. Coll. Cork Cork Ireland). Presented at Joint Meeting of the Republic of Ireland Sub-Region of the Analytical Division and the Atomic Spectroscopy Group of the Royal Society of Chemistry Dublin Ireland March 28–29 1996.555 Calloway C. P. Jr. Diss. Abstr. Int. B 1995 56(6) 3162. 556 Becker-Ross H. Florek S. Heitmann U. Weisse R. Fresenius’ J. Anal. Chem. 1996 355(3–4) 300. 557 Harnly J. M. (Beltsville Human Nutrition Res. Center Food Composition Lab. USDA Beltsville MD 20705 USA). Presented at Eighth Biennial National Atomic Spectroscopy Symposium Univ. East Anglia UK July 17–19 1996. 558 Shrader D. Webb C. Kim S. Vanclay E. Mika J. (Varian OSI Wood Dale IL 60191 USA). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 559 Shrader D. Rivera C. Vanclay E. Moett J. (Varian OSI Wood Dale IL 60191 USA).Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 560 Edel H. Erber D. Lehnert R. Buscher W. Cammann K. Fresenius’ J. Anal. Chem. 1996 355(3–4) 292. 561 Roedel G. Radziuk B. Zeiher M. Stenz H. Ger. Oen. DE 4,413,096 (Cl. G01J3/42) 19 Oct 1995 Appl. 4,413,096 15 Apr 1994; 25 pp. 562 Shuttler I. L. Schlemmer G. Portala F. Feuerstein M. (Bodenseewerk Perkin-Elmer GmbH 88647 Ueberlingen Germany). Presented at Eighth Biennial National Atomic Spectroscopy Symposium Univ. East Anglia UK July 17–19 1996. 563 Deval A. Sneddon J. Microchem. J. 1995 52(1) 1996. 564 Moseley R. Z. Schleicher R. (Thermo Jarrell Ash Franklin MA 02038 USA). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. Xuebao Ziran Kuxueban 1995 15(2) 25.565 Fender M. A. Butcher D. J. Anal. Chim. Acta 1995 315(1–2) 167. 566 Lee S. Kim P. Anal. Sci. T echnol. 1994 7(4) 427. 567 Sun H.-W. Zeng Z.-M. Zhang Y.-L. Shao J.-H. Hebei Daxue 568 Sholupov S. E. Ganeev A. A. Spectrochim. Acta Part B 1995 50(10) 1227. 569 Sholupov S. E. Ganeev A. A. Timofeev A. D. Ivankov V. M. J. Anal. Chem. (Engl. T ransl.) 1995 50(6) 589. Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 570 L’vov B. V. Polzik L. K. Novichikhin A. V. Borodin A. V. Dyakov A. O. Spectrochim. Acta Part B 1996 51(6) 609. 571 Ma H.-B. Ren J.-S. Zhang G.-Z. Fenxi Huaxue 1995 23(6) 728. 572 Parsons P. J. Zong Y. Y. (Dept. Environ. Health and Toxicol. New York State Dept. Health State Univ. New York Albany and Wadsworth Center Albany NY 12201-0509 USA).Presented at Eighth Biennial National Atomic Spectroscopy Symposium Univ. East Anglia UK July 17–19 1996. 573 Qiao H.-C. Parsons P. J. Slavin W. Clin. Chem. (Washington D.C.) 1995 41(10) 1451. 574 Yang X.-T. He H.-K. Fenxi Ceshi Xuebao 1995 14 20. 575 Chen J.-H. He Z.-R. Gong H.-L. Shu Y.-H. Xie Y.-S. (China Natl. Anal. Centre Guangzhou 510070 China). Presented at Proceedings of the 2nd International Symposium of Worldwide Chinese Scholars on Analytical Chemistry Shenzhen China November 15–18 1995. 576 Toriyama N. Kyushima H. Tateishi N. Uchino T. Ikedo T. Itoh Y. Matsui R. Shimazu Y. (Hamamatsu Photonics K. K. Shizuoka 438-01 Japan). Presented at Analytica 1996 Munich Germany April 23–26 1996. 577 Sanders J.B. (Varian Australia Pty Ltd. Mulgrave 3170 Victoria Australia). Presented at 10th National Convention of the Royal Australian Chemical Institute 27 September–2 October 1995 Adelaide Australia. 578 Gilmutdinov A. Kh. Radziuk B. Sperling M. Welz B. Nagulin K. Yu. (Dept. Appl. Res. Bodenseewerk Perkin-Elmer GmbH D-88647 U� berlingen Germany). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 579 Pavski V. Sturgeon R. E. Chakrabarti C. L. (Dept. Chem. Ottawa-Carleton Chem. Inst. Carleton Univ. Ottawa Ontario Canada K1S 5B6). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 580 IUPAC Pure Appl. Chem. 1995 67(11) 1905. 581 Scheeline A. Spectroscopy (Eugene Oreg.) 1996 11(1) 14. 582 O’Haver T. C. Spectroscopy (Eugene Oreg.) 1996 11(1) 12.583 Flajnik C. Shrader D. Vanclay E. (Varian Optical Spectroscopy Instruments Wood Dale IL 60191 USA). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 584 Epstein M. S. Buehler N. B. Bullard M. F. (Anal. Chem. Div. Natl. Inst. Standards and Technol. Gaithersburg MD 20899 USA). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 585 He J.-Y. Chen J.-H. Li H.-W. He Z.-R. Xie Y.-S. He H.-K. (China Natl. Anal. Centre Guangzhou 510070 China). Presented at Proceedings of the 2nd International Symposium of Worldwide Chinese Scholars on Analytical Chemistry Shenzhen China November 15–18 1995. 586 McCrum M. (GBC). Presented at The 15th Nordic Atomic Spectroscopy and Trace Element Conference Ebeltoft Denmark June 2–6 1996.587 Tyson J. F. Anal. Proc. 1981 18(12) 542. 588 Koscielniak P. Analusis 1996 24(1) 24. 589 Kale U. Voigtman E. Spectrochim. Acta Part B 1995 50(12) 1531. 590 Shrader D. Delles F. Vanclay E. (Varian Optical Spectroscopy Instruments Wood Dale IL 60191 USA). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 591 Vanclay E. (Varian Australia Pty Ltd. Mulgrave 3170 Victoria Austra. Presented at 10th National Convention of the Royal Australian Chemical Institute 27 September–2 October 1995 Adelaide Australia. 592 Kneisel E. A. Bowyer W. J. Walker F. B. Huntsberger T. G. Bell G. Foust R. D. (Dept. Chem. Hobart and William Smith Coll. Geneva NY 14456 USA). Presented at FACSS XXII Cincinnati USA 15–20 October 1995.593 Bellido-Milla D. Hidalgo-Hidalgo de Cisneros J. L. Hernandez-Artiga M. P. Jimenez-Jimenez A. (Dpto. Quim. Anal. Univ. Cadiz 11510 Puerto Real Spain). Presented at Eighth Biennial National Atomic Spectroscopy Symposium Univ. East Anglia UK July 17–19 1996. Acta Part B 1993 48 681. 594 Oatts T. J. Hamilton L. G. Buddin N. P. III At. Spectrosc. 1995 16(4) 145. 595 Torsi G. Reschiglian P. Fagioli F. Locatelli C. Spectrochim. 596 Torsi G. Reschiglian P. Lippolis M. T. Toschi A. Microchem. J. 1996 53(4) 437. 597 Harnly J. M. (US Dept. Agric. Beltsville Human Nutr. Res. 377R Koltracht I. Spectrochim. Acta Part B 1996 51 713. Center Beltsville MD 20705 USA). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 598 Schrader D.Webb C. Mika J. Vanclay E. (Varian Optical Spectroscopy Instruments Wood Dale IL 60191 USA). Presented at Pittsburgh Conference (PITTCON ’96) Chicago IL USA March 3–8 1996. 599 Goldberg J. M. Gluodenis R. (Dept. Chem. Univ. Vermont Burlington VT 05405-0125 USA). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 600 L’vov B. V. Polzik L. K. Fedorov P. N. Novichikhin A. V. Borodin A. V. Spectrochim. Acta Part B 1995 50(13) 1621. 601 L’vov B. V. Polzik L. K. Novichikhin A. V. Borodin A. V. Dyakov A. O. Spectrochim. Acta Part B 1995 50(14) 1757. 602 Yuzefovsky A. I. Lonardo R. F. Zhou J. X. Michel R. G. 603 L’vov B. V. Polzik L. K. Kocharova N. V. Spectrochim. Acta Part B 1992 47 889. 604 Yuzefovsky A. I. Lonardo R.F. Michel R. G. McCarey J. T. (Res. Center Philip Morris USA Richmond VA 23234 (Barcelona) 1996 15 32. USA). Presented at Pittsburgh Conference (PITTCON ’96) Chicago IL USA March 3–8 1996. 605 Belarra M. A. Resano M. Castillo J. R. Quim. Anal. 606 Rippert E. D. Song S. Maglic S. R. Lomatch S. Wang J. SPIE-Int. Soc. Opt. Eng. 1994 2157 300. (Cambridge UK) 1995 120 295. Chen J. Thomas C. Ketterson J. B. Ulmer M. P. Proc. 607 Vankeerberghen P. Smeyers-Verbeke J. Massart D. L. Analusis 1995 23(6) 247. 608 Rattray R. Salin E. D. J. Anal. At. Spectrom. 1995 10(12) 1053. 609 Araujo P. W. Kavianpour K. Brereton R. G. Analyst 610 Zhang R. S. Chen X. G. Chen S. Y. Hu Z. D. Fenxi Huaxue 1996 24(6) 724. 611 Paz Carril M. Soledad Corbillon M.Madariaga J. M. Mikrochim. Acta 1996 124(1–2) 1. 612 Wu Y. L. Yu B. Zhou T. T. Fenxi Huaxue 1996 24(2) 202. 613 Zhong M.-H. Chang H.-Q. Zheng Y.-S. Fenxi Ceshi Xuebao 1996 15(1) 41. 614 Qiu X.-T. Zhong M.-H. Xiao L. Fenxi Shiyanshi 1996 15(1) 40. 615 Zhong M. H. Chang H. Q. Fenxi Shiyanshi 1996 15(3) 72. 616 Stalikas C. D. Pilidis G. A. Karayannis M. I. J. Anal. At. Spectrom. 1996 11(8) 595. 617 Araujo P. W. Marcano E. Gomez C. V. Benzo Z. An. Quim. 1994 90(5–6) 343. 618 Quevauviller P. Hill S. J. (Eur. Comm. Measurements and Testing Prog. B-1049 Brussels Belgium). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 619 Quevauviller P. Astruc M. Ebdon L. Muntau H. Cofino P. Quim. Anal. (Barcelona) 1995 14 169. W. P. Morabito R.Griepink B. Mikrochim. Acta 1996 123(1–4) 163. 620 Cobo-Fernandez M. G. Camara M. A. P. y C. Quevauviller 621 Haesselbarth W. (Bundesanstalt Mater. und-pruefung (BAM) D-12200 Berlin Germany). Presented at Pittsburgh Conference (PITTCON ’96) Chicago IL USA March 3–8 1996. 622 Christensen J. M. Mikrochim. Acta 1996 123(1–4) 231. J. J. Anal. At. Spectrom. 1996 11(4) 237. 623 Vankeerberghen P. Smeyers-Verbeke J. Massart D. L. J. Anal. At. Spectrom. 1996 11(2) 149. 624 Penninckx W. Hartmann C. Massart D. L. Smeyers-Verbeke 625 Dundar M. S. Haswell S. J. Anal. Proc. (L ondon) 1995 32 133. 626 Shuttler I. L. Schlemmer G. (Bodenseewerk Perkin-Elmer GmbH D-88647 U� berlingen Germany). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 627 Oey S.G. Morton S. F. N. (Unicam Atomic Absorption Cambridge UK CB1 2SU). Presented at Eighth Biennial National Atomic Spectroscopy Symposium University of East Anglia UK July 17–19 1996. 628 Greenfield S. T rends Anal. Chem. 1995 14(9) 435. 629 Butcher D. Spectroscopy (Eugene Oreg.) 1993 8(2) 14. 630 Guo X.-W. Guo X.-M. Shanghai Huanjing Kexue 1995 14(7) 28. 631 Gong Z.-B. Yang P.-Y. Wang X.-R. Lin Y.-J. Huang B. L. Fenxi Kexue Xuebao 1995 11(3) 87. 632 He H.-K. Shu Y.-H. Fenxi Shiyanshi 1995 14(1) 76. 633 Stockwell P. B. Corns W. T. Stockwell P. M. Poling J. (PS Analytical Ltd. Orpington Kent UK BR5 3HP). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 634 Liang D. C. (Aurora Instruments Ltd. Vancouver British 378R Journal of Analytical Atomic Spectrometry August 1997 Vol.12 Columbia Canada V5Y 1K3). Presented at Pittsburgh Conference (PITTCON ’96) Chicago IL USA March 3–8 1996. 635 Liang D. C. Au-Yeung P. Lee P. (Aurora Instruments Ltd. Vancouver B. C. Canada V5Y 1K3). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 636 Wang P.-X. (Varsal Instrum. Inc. Chalfont PA 18914 USA). Presented at FACSS XXII Cincinnati OH USA 15–20 October 1995. 637 Wang P.-X. (Varsal Instrum. Inc. Chalfont PA 18914 USA). Presented at FACSS XXII Cincinnati OH USA 15–20 October 1995. 638 Li G.-G. Liu J. Cao J.-S. Qi W.-Q. (China Natl. Environ. Monitoring Centre Beijing 100012 China). Presented at 6th International Beijing Conference and Exhibition on Instrumental Analysis Beijing China 24–27 October 1995.639 Yehl P. M. Tyson J. F. (Dept. Chem. Univ. Massachusetts Amherst MA 01003-4510 USA). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. Xuexiao Huaxue Xuebao 1995 16(6) 865. 640 Gong Z.-B. Yang P.-Y. Wang X.-R. Huang B.-L. Gaodeng 641 Duan Y. X. Du X. G. Li Y. M. Jin Q. H. Appl. Spectrosc. 1995 49(8) 1079. 642 Ince A. T. Dawson J. B. Snook R. D. J. Anal. At. Spectrom. 1996 11 967. 643 Bhattacharyya S. Bhattacharyya D. K. J. Phys. Chem. 1996 100(27) 11246. 644 Farnsworth P. B. Duersch B. S. Chen Y. B. (Dept. Chem. and Biochem. Brigham Young Univ. Provo UT 84602-5700 USA). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 645 Simeonsson J. B. Ezer M.Pacquette H. L. Preston S. L. Swart D. J. (Dept. Chem. Univ. Iowa IA 52242-1294 USA). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 646 Hueber D. M. Diss. Abstr. Int. B 1996 56(11) 6078. 647 Farnsworth P. Lazar A. (Dept. Chem. and Biochem. Brigham Young Univ. Provo UT 84602 USA). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 648 Lee S. C. Song K. S. Cha H. G. Lee J. M. Lipert R. J. Edelson M. C. (Dept. Chem. Kyungnam Univ. Masan South Korea). Presented at Pittsburgh Conference (PITTCON ’96) Chicago IL USA March 3–8 1996. 649 Davis C. L. Smith B. W. Winefordner J. D. Microchem. J. 1995 52(3) 383. 650 Davis C. L. Diss. Abstr. Int. B 1996 56(11) 6074. 651 Wang M. Kuangwu Yanshi 1995 15(3) 98. 652 Yuzefovsky A.I. Lonardo R. F. Michel R. G. Anal. Chem. 1995 67(13) 2246. 653 Gornushkin I. B. Smith B. W. Winefordner J. D. (Dept. Chem. Univ. Florida Gainesville FL 32611-7200 USA). Presented at Pittsburgh Conference (PITTCON ’96) Chicago IL USA March 3–8 1996. 654 L’vov T alanta 1976 23 109. 655 Lonardo R. F. Yuzefovsky A. I. Irwin R. L. Michel R. G. Anal. Chem. 1996 68(3) 514. 656 Cheam V. Desrosiers R. Lechner J. Sekerka I. Mudroch A. (Natl. Water Res. Inst. Burlington Ontario Canada L7R 4A6). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 657 Heitmann U. Hese A. Schoknecht G. Gries W. JAERIConf. 1995 95(005 Vol. 2) 444. 658 Wagner E. P. II Smith B. W. Besteman D. A. Winefordner J. D. (Dept. Chem. Univ. Florida Gainesville FL 32611-7200 USA).Presenteat Pittsburgh Conference (PITTCON ’96) Chicago IL USA March 3–8 1996. 659 Baker C. L. Smith B. W. Winefordner J. D. Bolshov M. A. (Dept. Chem. Univ. Florida Gainesville FL 32611 USA). Presented at Pittsburgh Conference (PITTCON ’96) Chicago IL USA March 3–8 1996. 660 Lonardo R. F. Yuzefovsky A. I. Yang K. X. Michel R. G. Skelly Frame E. M. Barren J. J. Anal. At. Spectrom. 1996 11(4) 279. 661 Beissler H. Petrucci G. A. Baechmann K. Panne U. Cavalli P. Omenetto N. Fresenius’ J. Anal. Chem. 1996 355(3–4) 345. 662 Masera E. Mauchien P. Remy B. Lerat Y. J. Anal. At. Spectrom. 1996 11(3) 213. 663 Hill S. J. Dawson J. B. Price W. J. Shuttler I. L. Tyson J. F. J. Anal. At. Spectrom. 1995 10 199R. 664 Thiem T.L. Lee Y. III Sneddon J. Adv. At. Spectrosc. 1995 2 179. 665 Schuetz M. Heitmann U. Hese A. Appl. Phys. B 1995 B61(4) 339. 666 Michel R. G. Zhou J. X. Hou D. X. (Dept. Chem. Univ. Connecticut Storrs CT 06269-3060 USA). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 667 Zhou J. X. Hou X. Yang K. X. Michel R. G. (Dept. Chem. Univ. Connecticut Storrs CT 06269-3060 USA). Presented at Pittsburgh Conference (PITTCON ’96) Chicago IL USA March 3–8 1996. 668 Levenson M. D. Commercialize an atomic absorption system based on frequency-doubled tunable diode lasers for rate monitoring process control and spectroscopy of physical vapor deposition Report AD-A299180 1995 5 pp. P. Y. Fresenius’ J. Anal. Chem. 1996 355(3–4) 317. 669 Schnuerer-Patschan C.Niemax K. Spectrochim. Acta Part B 1995 50(9) 963. 670 Burakov V. S. Isaevich A. V. Meleshchenko L. A. Misakov 671 Burakov V. S. Misakov P. Y. Raikov S. N. Fresenius’ J. Anal. 672 Zheltukhin A. A. J. Anal. Chem. (Engl. T ransl.) 1995 Chem. 1996 355(3–4) 361. 50(9) 865. 673 Prishivalko A. P. Astafieva L. G. Leiko S. T. Appl. Opt. 1996 35(6) 965. 674 Gill C. G. Allen T. M. Anderson J. E. Taylor T. N. Kelly P. B. Nogar N. S. Appl. Opt. 1996 35(12) 2069. 675 Ashikhjmina E. I. Oshemkov S. V. Petrov A. A. Khait O. V. Shlyaev F. E. Artamanova E. O. Guletskii N. N. Petrov S. Ya. Opt. Zh. 1995(6) 37. 676 Multari R. A. Cremers D. A. Foster L. E. (Los Alamos Natl. Lab. Los Alamos NM 87545 USA). Presented at FACSS XXII Cincinnati USA 15–20 October 1995.677 Vadillo J. M. Milan M. Laserna J. J. Fresenius’ J. Anal. Chem. 1996 355(1) 10. 678 Cardell K. Pichahchy A. E. Cremers D. A. (Chem. Sci. and Technol. Div. Los Alamos Natl. Lab. Los Alamos NM 87545 USA). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 679 Jensen L. C. Langford S. C. Dickinson J. T. Addleman R. S. Spectrochim. Acta Part B 1995 50(12) 1501. 680 Vadillo J. M. Laserna J. J. T alanta 1996 43(7) 1149. 681 Singh J. P. Zhang H. Yueh F.-Y. (Diagnostic Instrumentation and Anal. Lab. Mississippi State Univ. Mississippi State MS 39762 USA). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 682 Singh J. P. Yueh F.-Y. Su C. F. (Diagnostic Instrumentation and Anal. Lab. Mississippi State Univ. Mississippi State MS 39762 USA). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 683 Pakhomov A. V. Nichols W. Borysow J. Appl. Spectrosc. 1996 50(7) 880. Journal of Analytical Atomic Spectrometry August 1997 Vol. 12 Appl. Spectrosc. 1996 50(2) 222. 684 Ernst W. E. Farson D. F. Sames D. J. Appl. Spectrosc. 1996 50(3) 306. 685 Yamamoto K. Y. Cremers D. A. Ferris M. J. Foster L. E. 686 Marquardt B. J. Goode S. R. Shaw T. J. Angel S. M. (Dept. Chem. and Biochem. Univ. South Carolina Columbia SC 29208 USA). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 687 Marquardt B. J. Goode S. R. Angel S. M. Anal. Chem. 1996 68(6) 19977. 688 Knopp R. Scherbaum F. J. Kim J. I. Fresenius’ J. Anal. Chem. 1996 355(1) 16. 689 Eppler A. S. Pichahchy A. E. Cremers D. A. (Chem. Sci. and Technol. Div. Los Alamos Natl. Lab. Los Alamos NM 87545 USA). Presented at FACSS XXIII Kansas City MO USA September 29–October 4 1996. 690 Yagi T. Huo Y. Appl. Opt. 1996 35(18) 3183. 691 Singh J. P. Zhang H. Yueh F.-Y. Carney K. P. Appl. Spectrosc. 1996 50(6) 764. 692 Travis J. C. Turk G. C. Laser-enhanced ionization spectroscopy John Wiley and Sons Ltd Chichester UK 1996 0 471 57684 320. 693 Nasimov A. M. Khalmanov A. T. Tursunov A. T. Chekalin N. V. Zavod. L ab. 1995 61(4) 21. 694 Turk G. C. Yu L.-J. Koirtyohann S. R. (Anal. Chem. Div. Natl. Inst. Stand. Technol. Gaithersburg MD 20899 USA). Presented at FACSS XXII Cincinnati USA 15–20 October 1995. 695 Boudreau D. Ljungberg P. Axner O. Spectrochim. Acta Part B 1996 51(4) 413. 696 Clevenger W. K. Riter K. L. Mordoh L. S. Smith B. W. Winefordner J. D. (Dept. Chem. Univ. Florida Gainesville FL V. I. AIP Conf. Proc. 1995 329 535. Akad. Nauk Resp. Uzb. 1994(12) 20. 32611-7200 USA). Presented at Pittsburgh Conference (PITTCON ’96) Chicago IL USA March 3–8 1996. 697 Riter K. L. Clevenger W. L. Mordoh L. S. Smith B. W. Matveev O. I. Winefordner J. D. J. Anal. At. Spectrom. 1996 11(6) 393. 698 Gorbatenko A. A. Kuzyakov Y. Y. Murtazin A. R. Zorov N. B. AIP Conf. Proc. 1995 329 105. 699 Kuzyakov Y. Ya. Zorov N. B. Gorbatenko A. A. Beketov 700 Matveev O. I. Cavalli P. Omenetto N. AIP Conf. Proc. 1995 329 269. 701 Nasimov A. M. Kholmanov A. T. Tursunov A. T. Dokl. 702 Oey S. G. Seare N. J. Tyson J. F. Kibble H. A. B. J. Anal. At. Spectrom. 1991 6(3) 133. 703 Xie Z.-h. Lang H.-y. Gao Y. Huaxue Yanjiu Yu Y ingyong 1995 7(2) 219. Paper 7/03983B Received June 9 1997 379R

ISSN:0267-9477    

DOI:10.1039/a703983b

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Establishing an SI-traceable Copper Concentration in the Candidate Reference Material MURST ISS A1 Antarctic Sediment Using Isotope Dilution Applied as a Primary Method of Measurement I. PAPADAKIS†, P. D. P. TAYLOR AND P. DE BIE` VRE Institute for ReferenceMaterials and Measurements, European Commission-JRC, B-2440 Geel, Belgium Traceability is a term that is heavily debated world-wide in the material, are the subject of dierent reports prepared by the analytical chemistry community.This paper describes an organisers of the project. attempt to obtain an SI-traceable value for the Cu According to the International Vocabulary of Basic and concentration in the Candidate Reference Material MURST- General Terms in Metrology (VIM) definition,2 traceability is: ISS A1 Antarctic Sediment. This material was collected by the ‘property of the result of a measurement whereby it can be Instituto Superiore di Sanita (ISS, Rome, Italy) and was related to stated references, usually national or international processed at the Institute for Reference Materials and standards, through an unbroken chain of comparisons all Measurements (IRMM, Geel, Belgium) into a homogeneous having stated uncertainties’.and dried powder. The analytical method used was isotope In our opinion, such ‘stated uncertainties’ should be as dilution (ID) combined with inductively coupled plasma mass complete and detailed as possible. Here, we present two spectrometry (ICP-MS).Microwave pressurised digestion and approaches to this subject using the EURACHEM and separation of Cu by ion-exchange chromatography were used. International Organisation for Standardisation (ISO)/Bureau The International Vocabulary of Basic and General Terms in International des Poids et Measures (BIPM) guides for the Metrology (VIM) definition of traceability requires ‘stated establishment of ‘full’ uncertainty. uncertainties’. Because a primary method of measurement A complete uncertainty statement is also needed for the (ID) is used, an attempt was made to make these ‘stated result obtained by primary methods of measurement.A priuncertainties’ as detailed as possible, thereby using the mary method has been defined by the Comite� Consultatif pour International Organisation for Standardization (ISO)/Bureau la Quantite� de Matie`re (CCQM)3 of the BIPM as: ‘a method International des Poids et Measures (BIPM) guide and taking having the highest metrological qualities, whose operation into account all possible sources of uncertainty (Type A and can be completely described and understood, for which a Type B).The established value for the Cu concentration is complete uncertainty statement can be written down in terms 86.7 nmol g-1 with an expanded uncertainty of 4.9 nmol g-1 of SI units, and whose results are, therefore, accepted without (coverage factor k=2). reference to a standard of the quantity being measured’.Practically, this means ‘being able to write down an explicit Keywords: T raceability; primary method; isotope dilution mass equation describing what is measured to what is intended to spectrometry; copper isotope ratio; inductively coupled plasma be measured without containing any (significant) empirical mass spectrometry; SiCl+ interference; Antarctic sediment correction factors’. The BIPM concept of primary methods, although new in chemistry, is applied in other measurement Traceability is a tool to achieve comparability of measuresciences. ments.This is important in the context of many border crossing Isotope dilution (ID or IDMS for isotope dilution mass issues related to measurement values in trade and industry. spectrometry) was recognised3 by CCQM as a method that When traceability is towards a common base such as the SI, has the potential to be primary if carried out correctly (e.g., it it is called SI-traceability. Linked with the concept of does not automatically produce correct results).Primary SI-traceability is the notion that values are stable in place and methods, and hence ID are, however, unique tools in chemical time and not dependent on the procedure used. This is, for measurements which can, from an a priori point of view, lead instance, considered important in many legal, environmental to results with small expanded uncertainty.4,5 The influence of and climatological issues.1 matrix eects caused by complicated matrices, such as for this In this paper we propose a way to achieve this traceability sediment, is far more dicult to identify and estimate quantitat- to the SI in the determination of the Cu concentration in the ively when using procedures not based on measuring isotope candidate reference material MURST ISS A1 Antarctic amount ratios.Moreover, measuring isotope amount ratios Sediment. The results presented are only the contribution of has the additional advantage of making the procedure inde- the Stable Isotope Measurement (SIM) Unit of the Institute pendent of other sources of possible uncertainties, such as for Reference Materials and Measurements (IRMM) to the analyte loss, occurring during the sample treatment (digestion, certification campaign of the material, which was carried out separation, transfer of the sample, etc.).under the responsibility of the Instituto Superiore di Sanita It is critical when using isotope mass spectrometry to verify (ISS).Other methods were also used for this purpose. Those the absence of possible isobaric interference. This problem results, as well as the certified value of the new reference exists independently of the calibration strategy used (e.g., IDMS, external calibration, standard additions) and usually causes many problems in analyses. In this work, a quadrupole † EC fellow. Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12 (791–796) 791inductively coupled plasma mass spectrometry (ICP-MS) sample, Ny=known number of atoms of the element in the spike, Ry=isotope amount ratio of the chosen isotope pair in instrument was used.It is well known that the Cu mass spectrum suers from many isobaric interferences. A well the spike, Rb=isotope amount ratio of the chosen isotope pair in the blend, Rx=isotope amount ratio of the chosen isotope known case is 40Ar23Na+ interfering on 63Cu,6 when Na is present at high concentration in the samples.Ion-exchange pair in the sample, SRix=sum of the isotope amount ratios in the sample, SRiy=sum of the isotope amount ratios in the chromatography was applied to the samples in order to remove any such interferences. Since, however, other possible inter- spike, cx=unknown concentration of the element in the sample, cy=known concentration of the element in the spike, my= ferences are not well documented in the open literature, a detailed study of interferences was performed on the samples mass of spike material used for the blend preparation and mx=mass of sample material used for the blend preparation.before and after the separation. When applied to the Cu isotope ratio R=n(63Cu)/n(65Cu), where n is the amount expressed in moles, the concentration EXPERIMENTAL cx of this element in the samples can be evaluated from eqn. (2). It takes into account mass discrimination eects during Instrumentation the measurement (mass discrimination correction factor=K) For the measurements of the isotope amount ratios in the as well as the moisture content of the sediment [moisture blends (spiked samples), ICP-MS was used.The instrument correction factor w=1-(mass of water/mass of sample)]. used is manufactured by Fisons (Fisons VG PlasmaQuad 2+) and was slightly modified at IRMM. It is equipped with cx= Ry-K Rb(obs) K Rb(obs)-Rx cy my w mx Rx+1 Ry+1 (2) Balzers turbo pumps and a V-groove nebuliser.Argon was used for the operation of the plasma torch. Further technical characteristics and operating conditions are presented in Table 1. For the digestion of the samples, a microwave oven Weighing of the Samples and Blend Preparation (Milestone MLS-1200 MEGA) with a microwave digestion The weighing of the samples was performed in a humidityrotor (1000/6/100/110) (six-position, pressure up to 110 bar) conled environment (humidity stable between 50 and 60%) was used.in order to avoid absorption of water content into the sediment powder from the humidity of the atmosphere. In addition, the Sample sample was immediately weighed in the PFA digestion vessels to avoid transfer to dierent containers and eliminate possible The candidate reference material MURST ISS A1 Antarctic losses and additional uncertainties to the final result. Two Sediment originates from the Terra Nova Bay (Antarctica). dierent bottles of the material were used and three fractions The certification campaign was carried out by the ISS (Italy), of 100 mg were taken from each bottle.whose delegation performed the sampling. The material was For the spiking of the samples, the certified spike reference dispatched to IRMM where it was processed7 into a fine material IRMM-63212 enriched in 65Cu having a ratio (<150 mm) sediment powder. The moisture content of the n(63Cu)/n(65Cu)=0.0028921 with an expanded uncertainty of material was measured to be 0.35% with a standard uncertainty 0.0000086 (k=1) and concentration cy=0.0437 mmol g-1 Cu of 0.05% using Karl-Fisher titration in ten dierent bottles of with an expanded uncertainty of 0.0002 mmol g-1 (k=1) was the material.7 A homogeneity study8 for Cu was performed used.The concentration of the spike was calculated graviusing solid sampling Zeeman-eect background corrected metrically and confirmed using reverse ID, when the isotopic atomic absorption spectrometry (SS-ZAAS), carried out composition was measured by means of thermal ionisation on 20 dierent bottles of the material, which gave for Cu a mass spectrometry (TIMS).To minimise possible problems homogeneity factor HE=5.8%Ómg.9 due to multiplier dead time the blend isotope amount ratio (Rb) was kept close to unity. Procedure The spike was added to the sample material, which was weighed in the PFA vessel, before any further treatment of the The general IDMS equation [eqn. (1)] as can be found in the sample.For each blend, approximately 0.1 g of sample material literature10,11 is: and 0.08 g of IRMM-632 material were used. All the weighings were performed by the IRMM mass Nx Ny = (Ry-Rb) (Rb-Rx) S Rix S Riy metrology department using substitution measurements against operational mass standards, which are calibrated against an IRMM stainless steel mass standard, which itself is u cx = (Ry-Rb) (Rb-Rx) (cy my) mx S Rix S Riy (1) calibrated against a stainless steel BIPM mass standard.The latter is finally calibrated against the platinum–iridium primary where Nx=unknown number of atoms of the element in the mass standard at BIPM. A certificate accompanied each individual weighing where the weighed mass and its total uncertainty (taking into account buoyancy correction) was Table 1 Operating conditions of the quadrupole ICP-MS instrument given. used for the isotopic measurements Electron multiplier Galileo 4870V Digestion Procedure Dead time/ns 13±1 Nebulizer type V-groove The optimised microwave digestion procedure consisted of a Nebulizer flow/l min-1 #0.8 simplified one-step procedure to ensure homogeneity between Sample aspiration rate/ml min-1 #1 sample and spike materials.Forward power/W 1400 Plasma gas flow/l min-1 #14 The acid mixture used was 5 ml of HNO3 (#14 M) and HF Intermediate flow/l min-1 #1.5 (#20 M) in the proportion 1+1 (v/v). The heating programme Data acquisition mode Peak jump, 3 points per peak was a five-step, 7 min each, programme with 250W power Dwell time/ms 10.24 used in the odd steps and no power used in the even steps. Acquisition time/s 6×60 No pressure indication was available but maximum admissible 792 Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12pressure in the vessels is 110 bar. Both acids used were of subboiled quality and prepared in the IRMM Ultra Clean Chemical Laboratory (UCCL). After heating under pressure in the digestion vessels, the samples were evaporated to dryness.Finally, the blends were dissolved in about 50 g of 0.14 M HNO3. Separation The aim of the separation was to remove the alkali and alkaline earth metals from the blends. Bio-Rad AG 1-X8 anionexchange resin in the chloride form was used. The columns were about 2 ml in volume. The separation was carried out in concentrated HCl (7.5 M) of sub-boiled quality: a 5 g fraction of each blend (in 0.14 M HNO3 matrix) was first evaporated and then redissolved in 2 ml of 7.5 M HCl prior to the separation.The column was flushed with 10 ml of 7.5 M HCl (to remove alkali and alkaline earth metals). About 10 ml of Fig. 1 Theoretical eect of SiCl+ isobaric ions on the measurement 0.14 M HNO3 were used to elute the Cu from the column. of the ion current ratio I15I2 (I1: m/z=63; I2: m/z=65) in natural and non-natural Cu (blend) and on the apparent Cu concentration (the arrows are indicate the y-axis used for each trend-line). 2 Ion current Measurements ratio on mass positions 63 and 65 in natural Cu solution (slope= 0.47). & Ion current ratio on mass positions 63 and 65 in blends The sample acquisition time was 6×60 s and between samples (slope=1.54). + Apparent Cu concentration calculated using IDMS. there was a minimum of 3 min washout time, which is adequate to avoid carry-over from one sample to the next. At the beginning of the measurement, a blank solution (0.14 M HNO3) was measured and from all subsequent measurement values, this blank value was subtracted. Every two blends, a natural Cu solution (ASARCO metal dissolved in sub-boiled 0.14 M HNO3) with an approximate concentration of 100 nmol g-1, prepared at IRMM, was measured in order to determine and apply the correction factor for mass discrimination.Interference Study Because of the possible interference of ArNa+, Na was first removed from the sample as described above. After this separation, some other possible interferences were studied in detail.The interference most dicult to detect, not widely reported in the literature, is the interference of SiCl+. Only Fig. 2 Observed eect of Si on the ratio of ion currents at mass positions 63 and 65, as measured in the blends (second-order Vanhaecke et al.13 have reported it. The eect of this interpolynomial fit). ference can be simulated. This is eected by calculating the ‘abundance’ of the SiCl+ species at the dierent mass positions problem to a level where it was no longer detectable as by multiplying the abundances of the Si isotopes by the substantiated by measurements on the separated samples.abundances of the Cl isotopes according to combination Additionally, after removal of the two influencing interferences theory. The natural abundances of Cu and the ‘theoretical (ArNa+ and SiCl+), the samples were measured for isotopic abundances’ of SiCl+ are found to be very similar.Fig. 1 composition (unspiked) and this was found to be identical (calculated from theoretical isotope abundance ratios14) clearly with the natural isotopic composition, which indicates the shows the problem of detecting this interference, which absence of other interferences (e.g., ZnH, SO2). approaches the natural Cu isotope ratio. Moreover, Fig. 1 shows that the eect of this interference is much more severe on blend (close to unity) Cu ratios (factor of 3) and it is shown RESULTS that the eect on the apparent Cu concentration is even larger.This was confirmed experimentally by spiking six blends with The Cu concentration in the material established with this procedure is 86.7 (4.9) nmol g-1 (the number in parentheses is dierent amounts of Si. The results are presented in Fig. 2. In addition, high resolution (HR)-ICP-MS was used to investigate the expanded uncertainty with k=2). The results from the measurements on each individual this problem, as has been described in a companion paper,15 in more detail.The study was carried out at the University of digestion are given in Table 2. These results are not corrected for the procedural blank. The Cu blank of the procedure used Gent, Laboratory of Analytical Chemistry, and confirmed our indirect prediction of this interference. was determined using ID as being 2.5 nmol g-1 with a standard uncertainty of 0.1 nmol g-1. This value was then subtracted The presence of SiCl+ species in the solution, after ionexchange, can be explained by the presence of Si in the from the measured Cu concentration in the sediments and its standard uncertainty added to the combined uncertainty of sediments (mostly silicate minerals) and Cl from using an ionexchange column in the chloride form.Early digests (using the result. In the following sections, the calculation of the uncertainty, dierent digestion procedures, still leaving trace amounts of Si and Cl) showed variable concentrations up to 120 nmol g-1 according to the ISO/BIPM guide,16 is described in detail.The ISO/BIPM document is a consensus document developed for Cu. The final digestion procedure as described (which uses HF in the digestion, forming volatile SiF4, which volatilises when measurements in general, which was only finalised after many years of discussion. It gives guidance on how to evaluate and the solution is evaporated to near dryness) minimised this Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12 793Table 2 Results obtained from the measurement on each individual isobaric interference correction factor Kiso digestion (blend) after moisture content correction. For the relevant uncertainties consult the uncertainty budgets cx=H CRy-K Kiso Rb(obs) K Kiso Rb(obs)-Rx cy my w mx Rx+1 Ry+1D (3) Sample name Bottle No. c (Cu)/nmol g-1 As a basis for the H factor, the independent homogeneity I960620A 92.9 study was taken.8 For a sample size of 100 mg, the variability I960620B 50 89.3 I960620C 87.1 of the results obtained due to the homogeneity of the material I960620D 88.5 is expected to be 0.6%; hence, H=1.000±0.006. I960620E 75 86.6 The estimation of Kiso is more complicated.From the I960620F 90.6 HR-ICP-MS scans, the ratio of the SiCl5Cu signal was estimated (based on the same assumption as Fig. 1). Assuming that this is representative of the amount ratio, this value was then introduced into Fig. 1 and the influence on Rb was calculated. This was found to be 1.4%; hence, Kiso= express uncertainty. One of the important aspects introduced 1.000±0.014. is the distinction between Type A and Type B sources of un- With these data, a new uncertainty budget (Table 5) was certainty. ‘Type A evaluation is a method of evaluation of a produced using eqn. (3), resulting in a combined uncertainty standard uncertainty by the statistical analysis of series of similar to the original uncertainty. This confirms that the observations’ whereas ‘Type B is a method of evaluation of a magnitude of the reproducibility (Approach I) can be explained standard uncertainty by means other than by the statistical by taking into account the homogeneity and remaining varia- analysis of series of observations’ (e.g., generated by certificates, bility caused by the spectral interference of SiCl+ (Approach or when assessing the influence of parameters, as was carried II).out in this work for SiCl+ interference). The important steps in the uncertainty estimation process are: defining the measurand in a mathematical equation [in DISCUSSION our case eqn. (2)], identifying the sources of uncertainty, The somewhat vague VIM definition for traceability mentions estimating their magnitude and finally, combining the uncer- ‘stated uncertainties’. In our interpretation, to be useful, this tainty contributions. The first three steps are critical and often means a complete uncertainty budget where all uncertainty the most dicult, whereas the final step is merely a trivial sources are described is needed.Obviously, this requirement mathematical problem, for which there often exist good needs to be translated into reality, as completeness can only approximations (see Appendix). be reached asymptotically. Indeed, as an example, in this work, In the following paragraphs, the combined uncertainty for the isobaric interferences ArNa+ and SiCl+ were investigated the measured Cu concentration in this material is determined in depth, but many others, which can be found in the inter- using two dierent approaches.In the first approach, part of ference tables (e.g., ZnH, SO2), were not investigated in depth the uncertainty of the final value is simply derived from the because they seem to be negligible based on the experimental reproducibility of the six obtained IDMS results. In a second, evidence that the isotopic composition of the samples was more refined approach, all available knowledge on uncertainty found to be identical with the natural isotopic composition.sources (e.g., material inhomogeneity and variability of the This is why, in the context of the ISO/BIPM uncertainty SiCl+ interference, both of which were studied and quantified philosophy, the responsibility for the uncertainty finally as described above) were incorporated in the uncertainty remains with the person performing the analysis.Obviously, budget. the major advantage of the uncertainty budget (Tables 3–5) is to have a list of sources considered, which can be the starting Uncertainty Calculation Approach I point of discussion should problems arise. Tables 3 and 4 list the dierent uncertainty sources considered. As can be seen, an uncertainty statement is made for each of the quantities as defined in the IDMS equation [eqn. (2)]. Table 3 Uncertainty budget for Approach I using the simplified Most of these are Type A uncertainties.When these sources approach of addition of relative variances are combined, this yields an uncertainty of 2.8%, whereby the Parameter Typical value SU* RSU† (%) reproducibility of IDMS on the six digestions is by far the dominant contribution. Therefore, in an empirical, pragmatic Weighing data— approach, one could simply consider this reproducibility as mx (g) 0.106 0 0.000 5 0.5 my (g) 0.094 5 0.000 5 0.5 reflecting the uncertainty originating from all undefined Certificate data— sources. cy (Cu/mmol g-1) 0.043 7 0.000 2 0.4 Ry 0.002 9 9×10-6 0.3 Measurement data— Uncertainty Calculation Approach II Rb 0.894 3 0.001 8 0.2 Rx 2.243 6 0.002 2 0.1 In an attempt to refine the uncertainty budget further, an eort K 1.055 8 0.001 3 0.1 was made to identify the uncertainty sources as apparent in Other data— the large reproducibility (from the repeated IDMS).In the w 0.996 5 0.000 5 0.05 meantime, data from SS-ZAAS on the homogeneity of the Measurement blank subtraction <0.01 material became available.8 Furthermore, the theoretical esti- Reproducibility 2.64 Preliminary uncertainty on mate of the extent of isobaric interference was used to assess cx (Cu/mmol g-1) 0.089 2 0.002 4 2.79 the uncertainty introduced from this source.Although no Procedural blank (mmol g-1) 0.002 5 0.000 1 4.75 correction was carried out on the Cu concentration due to the Combined uncertainty uc 2.79 isobaric interference (because experimentally a change in iso- Expanded uncertainty U (k=2) 5.58 tope ratio could not be detected), an uncertainty was added to the combined uncertainty.Eqn. (2) was fine-tuned, resulting *Standard uncertainty. †Relative standard uncertainty. in eqn. (3), which includes a homogeneity factor H and an 794 Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12Table 4 Uncertainty budget for Approach I, combining uncertainty contributions using the EURACHEM spreadsheet approach Parameter Reproducibility Blank Rb Ry Rx cy mx my K w Typical value 1 0.002484 0.8943 0.002892 2.2436 0.043723 0.106 0.0945 1.0558 0.9965 SU* 0.0264 0.000118 0.0018 8.6×10-6 0.0022 0.000183 0.0005 0.0005 0.0013 0.0005 Reproducibility 1.0264 1 1 1 1 1 1 1 1 1 1 Blank 0.002484 0.002602 0.002484 0.002484 0.002484 0.002484 0.002484 0.002484 0.002484 0.002484 0.002484 Rb 0.8943 0.8943 0.8961 0.8943 0.8943 0.8943 0.8943 0.8943 0.8943 0.8943 0.8943 Ry 0.002892 0.002892 0.002892 0.002901 0.002892 0.002892 0.002892 0.002892 0.002892 0.002892 0.002892 Rx 2.2436 2.2436 2.2436 2.2436 2.2458 2.2436 2.2436 2.2436 2.2436 2.2436 2.2436 cy 0.043723 0.043723 0.043723 0.043723 0.043723 0.043906 0.043723 0.043723 0.043723 0.043723 0.043723 mx 0.106 0.106 0.106 0.106 0.106 0.106 0.1065 0.106 0.106 0.106 0.106 my 0.0945 0.0945 0.0945 0.0945 0.0945 0.0945 0.0945 0.095 0.0945 0.0945 0.0945 K 1.0558 1.0558 1.0558 1.0558 1.0558 1.0558 1.0558 1.0558 1.0571 1.0558 1.0558 w 0.9965 0.9965 0.9965 0.9965 0.9965 0.9965 0.9965 0.9965 0.9965 0.997 0.9965 cx 0.091518 0.089046 0.089483 0.089162 0.089071 0.089547 0.088734 0.089649 0.089359 0.089118 0.089164 SUi 0.002354 -0.00012 0.00032 -1.6×10-6 -9.3×10-5 0.000384 -0.00043 0.000485 0.000195 -4.6×10-5 (SUi)2 5.54×10-6 1.39×10-8 1.02×10-7 2.63×10-12 8.62×10-9 1.47×10-7 1.85×10-7 2.35×10-7 3.82×10-8 2.11×10-9 S(SUi2) 6.27×10-6 cx/mmol g-1 SU/mmol g-1 RSU(%) 0.089164 0.002505 2.80905 * Standard Uncertainty.Table 5 Uncertainty budget for Approach II, combining uncertainty contributions using the EURACHEM spreadsheet approach Parameter Homogeneity Blank Kiso Rb Ry Rx cy mx my K w Typical value 1 0.002484 1 0.8943 0.002892 2.2436 0.043723 0.106 0.0945 1.0558 0.9965 SU* 0.006 0.000118 0.0144 0.0018 8.6×10-6 0.0022 0.000183 0.0005 0.0005 0.0013 0.0005 Homogeneity 1.006 1 1 1 1 1 1 1 1 1 1 1 Blank 0.002484 0.002602 0.002484 0.002484 0.002484 0.002484 0.002484 0.002484 0.002484 0.002484 0.002484 0.002484 Kiso 1 1 1.0144 1 1 1 1 1 1 1 1 1 Rb 0.8943 0.8943 0.8943 0.8961 0.8943 0.8943 0.8943 0.8943 0.8943 0.8943 0.8943 0.8943 Ry 0.002892 0.002892 0.002892 0.002892 0.002901 0.002892 0.002892 0.002892 0.002892 0.002892 0.002892 0.002892 Rx 2.2436 2.2436 2.2436 2.2436 2.2436 2.2458 2.2436 2.2436 2.2436 2.2436 2.2436 2.2436 cy 0.043723 0.043723 0.043723 0.043723 0.043723 0.043723 0.043906 0.043723 0.043723 0.043723 0.043723 0.043723 mx 0.106 0.106 0.106 0.106 0.106 0.106 0.106 0.1065 0.106 0.106 0.106 0.106 my 0.0945 0.0945 0.0945 0.0945 0.0945 0.0945 0.0945 0.0945 0.095 0.0945 0.0945 0.0945 K 1.0558 1.0558 1.0558 1.0558 1.0558 1.0558 1.0558 1.0558 1.0558 1.0571 1.0558 1.0558 w 0.9965 0.9965 0.9965 0.9965 0.9965 0.9965 0.9965 0.9965 0.9965 0.9965 0.997 0.9965 cx 0.089699 0.089046 0.091471 0.089483 0.089162 0.089071 0.089547 0.088734 0.089649 0.089359 0.089118 0.089164 SUi 0.000535 -0.00012 0.002307 0.00032 -1.6×10-6 -9.3×10-5 0.000384 -0.00043 0.000485 0.000195 -4.6×10-5 (SUi)2 2.86×10-7 1.39×10-8 5.32×10-6 1.02×10-7 2.63×10-12 8.62×10-9 1.47×10-7 1.85×10-7 2.35×10-7 3.82×10-8 2.11×10-9 S(SUi2) 6.34×10-6 cx/mmol g-1 SU/mmol g-1 RSU(%) 0.089164 0.002518 2.824015 * Standard Uncertainty.Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12 795CONCLUSION Measurements unit for his help in the development of the separation procedure and to F.Vanhaecke from Gent This paper attempts to explain how SI-traceable values can be University for performing some very useful high resolution obtained for the measurement of the Cu concentration in a ICP-MS measurements. The whole project is in the framework sediment sample. This is established by making explicit the of the collaboration between IRMM and ISS. The MURST relationship between the measurand and the measured quantit- ISS-A1 Antarctic Sediment was prepared and certified under ies (isotope ratios, amounts), thereby including the relevant Italy’s National Programme for Research in Antarctica uncertainties in as much detail as possible.In this way, the (PNRA, Programma Nationale per la Ricerca in Antartide). CCQM definition of a primary method is applied in a real life situation. This highlights the transparency and relative ease with which uncertainty calculation can be performed, only for REFERENCES primary methods of measurement. 1 De Bie`vre, P., Kaarls, R., Peiser, H.S., Rasberry, S. D., and Reed, As is well known, ICP-MS, especially using quadrupole W. P., Accred. Qual. Assur. 1996, 1, 3. instruments, is very susceptible to isobaric interferences. In 2 International Vocabulary of Basic and General T erms in Metrology, particular, for bi-isotopic elements this can be problematic. In International Organisation for Standardisation, Gene`ve, 1993. this case, a relatively unknown interference from SiCl+ could 3 Comite� Consultatif pour la Quantite de Matie`re, Rapport de la have caused serious errors, if not properly identified.This was 1re session, 1995, E� dite� par le BIPM, Pavillon de Breteuil, F-92312 avoided by a metrological investigation, involving a careful, Se`vres Cedex, France. 4 Fassett, J. D., and Paulsen, P. J., Anal. Chem., 1989, 61, 643A. detailed and systematic study of the matrix and the subsequent 5 Moody, J. R., and Epstein, M. S., Spectrochim.Acta, Part B, 1991, removal of the interference. The uncertainty due to the possible 46, 1571. presence of undetectable amounts of SiCl+ interference was 6 Lyon, T. D. B., Fell, G. S., J. Anal. At. Spectrom., 1990, 5, 135. quantified by means of a Type B uncertainty evaluation. 7 Kramer, G., personal communication. 8 Grobecker, K. H., personal communication. 9 Kurfu� rst, U., Grobecker, K. H., and Stoeppler, M., In: T race APPENDIX Elements—Analytical Chemistry in Medicine and Biology, ed.There are several ways to combine uncertainties mathemat- Bra�ter, P., and Schramel, P., Walter de Gruyter, Berlin, 1984, Vol. 3, pp. 591–601. ically. The strictest way is a vigorous application of the 10 De Bie`vre, P., Fresenius’ J. Anal. Chem., 1990, 337, 766. uncertainty propagation law, which is cumbersome. An alterna- 11 De Bie`vre, P., Fresenius’ J. Anal. Chem., 1994, 350, 277. tive simplified approach is to add all the relative variances 12 Certificate for IRMM-632 spike isotopic reference material, in (Table 3) into a combined uncertainty uc. In this case, this preparation. approach gives a slightly overestimated combined uncertainty. 13 Vanhaecke, F., Vanhoe, H., Moens, L., and Dams, R., Bull. Soc. A third way consists of using the spreadsheet approach (Tables Chim. Belg., 1995, 104, 653. 14 International Union of Pure and Applied Chemistry, Commission 4 and 5) of the EURACHEM guide17 for uncertainty calcuon Atomic Weights and Isotope Abundances, Isotopic lation. This spreadsheet approach is a numerical method, Composition of Elements, Pure Appl. Chem., 1991, 63, 991. whereby the influence of a change in one of the uncertainty 15 Vanhaecke, F., Moens, L., Dams, R., Papadakis, I., and Taylor, components is calculated, for each of these components. P., Anal. Chem., 1997, 69, 268. Finally, these uncertainties are combined, again by adding the 16 Guide to the Expression of Uncertainty in Measurement, variances. International Organisation for Standardisation, Gene`ve, 1993. 17 Quantifying Uncertainty in Analytical Measurement, EURACHEM, London, 1995. Special thanks are due to B. Dijckmans from IRMM mass metrology for her help in the weighings, to P. Conneely and P. De Vos from the IRMM Management of Reference Paper 7/00750G Received February 3, 1997 Materials unit for their help in treating the solid sediment and its weighing, to K. Raptis from the IRMM Stable Isotope Accepted May 12, 1997 796 Journal of Analytical Atomic Spectrometry, August 199

ISSN:0267-9477    

DOI:10.1039/a700750g

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Microwave Digestion of Plant and Grain Reference Materials in Nitric Acid or a Mixture of Nitric Acid and Hydrogen Peroxide for the Determination of Multi-elements by Inductively Coupled Plasma Mass Spectrometry SHAOLE WU*, XINBANG FENG AND ADOLPH WITTMEIER Alberta Research Council, P.O. Bag 4000, Vegreville, Alberta, Canada T 9C 1T 4 The closed-vessel microwave digestion of four plant standard HNO3, HClO4, H2SO4, HCl, HF and H2O2. The solutions reference materials (SRMs) and two grain reference materials are then heated to near-dryness and the resulting residues are (RMs) in nitric acid or in a mixture of nitric acid and re-dissolved in an appropriate solvent (usually HNO3).These hydrogen peroxide was explored for the direct determination ashing procedures eectively decompose the organic materials of 26 elements, Al, As, B, Ba, Ca, Cd, Co, Cr, Cu, Fe, K, Li, in biological samples and minimize the dissolution reagents in Mg, Mo, Ni, Pb, Sb, Se, Sn, Sr, Ti, Tl, Th, U, V and Zn, by the final solution. Thus, the possible matrix and spectral ICP-MS.In 2.5 or 5 ml of HNO3 with or without the presence interferences in the subsequent ICP-MS determinations are of 2 ml of H2O2, 0.5 g of grain or plant sample was pre- eliminated or reduced. However, ashing procedures are usually digested overnight at room temperature, then at an elevated slow and tedious, needing constant attention, and are subject temperature of 120–165 °C with a pressure limit of 13.8 bar to possible contamination and potential loss of some volatile (200 psi) for 20 min. The presence of H2O2 helped to maintain elements.Direct wet dissolution procedures are usually a higher temperature under the pressure limit and reduced the achieved in concentrated HNO3 or in a mixture of HNO3 and carbon content in the digestates; but its impurities hampered H2O2 at an elevated temperature and pressure in closed vessels the ICP-MS analysis for certain elements at low levels. The or bombs heated thermally or by microwave energy.Many ICP-MS system was calibrated by the method of external studies have used this dissolution approach and have reported standards prepared in reagent blank solutions with In as the good recoveries for certain elements in various biological internal standard. Interferences from the sample matrices were materials analysed not only by ICP-MS6–12,15–18 but also by eliminated, corrected or reduced by subtracting the blank GFAAS and ICP.15,19,20 Compared with other acids such as signals, selecting suitable isotopes and applying the appropriate HClO4, HCl and H2SO4, HNO3 produces the least backinterference correction equations.Using this method, nearly all ground spectral interference in ICP-MS analysis21,22 and has of the predigestion spike recoveries for the 26 elements were become the most preferred sample preservative and dissolution within 90–110%. For the grain RMs studied including Corn reagent for ICP-MS analysis.With H2O and reactive oxygen Bran and Wheat Flour, the majority of the recoveries for most being the decomposition products, H2O2 has become the elements studied were within 85–115%. For the plant SRMs preferred additional oxidizing reagent used in digestion for studied including Pine Needles, Tomato Leaves, Apple Leaves subsequent ICP-MS analysis. This dissolution approach is and Peach Leaves, the majority of the recoveries were within simple, rapid and subject to less potential contamination and 90–115% for the determination of As, B, Ba, Ca, Cd, Cu, loss of volatile elements.However, using this approach the Mg, Mn, Mo, Pb, Sr, and Zn, within 70–100% for Al, Co, digestion of the siliceous materials in the samples is incomplete. Cr, Fe, K, Sb and V, but were 40–80% for Ni, Th, Ti and U. The organic materials in the samples may not be as completely The low recoveries were caused by the siliceous materials in decomposed as in the dry and wet ashing approaches.Also, these samples which were not decomposed during digestion. the biological materials studied using this digestion approach Keywords: Closed-vessel microwave digestion; inductively were mostly various animal tissues and biofluids, only a few coupled plasma mass spectrometry; biological material; spectral being plant materials. Most digestion bomb systems used were interference; environmental either heated thermally or heated by microwave energy without temperature and pressure control.The performance characteristics of the closed-vessel microwave digestion systems used in Inductively coupled plasma mass spectrometry (ICP-MS) has these studies varied widely from simple domestic microwave become one of the most attractive detection systems for the ovens to advanced laboratory microwave systems equipped determination of trace and ultra-trace elements because of its with temperature and pressure regulation and automatic power excellent detection limits, wide linear dynamic range, multicontrol. The maximum high-pressure of the closed vessel used element capability and the ability to measure isotope ratios.in these studies also varied from the (low) high-pressure of In recent years ICP-MS analytical methods have been increas- 15.8 bar (220 psi) to the (medium) high-pressure of 30–41 bar ingly applied to the determination of trace elements in biologi- (600 psi)11,12 or the (high) high-pressure of 117 bar (1700 psi).13 cal materials.1–18 The dissolution procedures used for biological In this study, the direct ICP-MS determination of 26 materials include dry ashing,1–3 wet ashing,4–6,15 or direct wet elements in plant SRMs and grain RMs digested in concen- dissolution.6–15 The dry ashing procedures include drying the trated HNO3 or in a mixture of HNO3 and H2O2 at an sample in a mue furnace for 8 h prior to wet dissolution,2 or elevated temperature and pressure was explored.A closed- using Mg(NO3)2 as an ashing aid in a tedious dissolution vessel laboratory microwave system equipped with temperature procedure.3 In wet ashing procedures, samples are initially digested in a mixture of acids or oxidizing reagents, such as and pressure regulation with the maximum (low) high-pressure Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12 (797–806) 797of 15.2 bar (220 psi) was used for the digestion under optimized sample with 5.0 ml being added to the sensor vessel (a minimum of 5 ml of solution is required to immerse the temperature conditions.The objectives were to investigate the potential and limitations of applying this simple method to analyse probe). Samples were pre-digested overnight (16 h) in a class 100 clean fumehood at room temperature. The grain and plant plant and grain samples collected in environmental monitoring, assessment and remediation projects, which often require a materials were digested in separate batches so that the sensor and the sample vessels contained similar matrices.A maximum high throughput of samples. The challenges of the study were the low analyte content of the samples and the relatively severe of twelve vessels including a reagent blank and a sensor were sealed and digested using the following heating program: heat spectral interferences arising from the biological matrix subjected to this direct wet dissolution procedure.to 165 °C within a ramp time of 10 min at full power (1000 W), hold for 20 min at 165 °C under the recommended maximum working pressure of 13.8 bar (200 psi). After the vessel had EXPERIMENTAL cooled, 97.5 and 95 ml of DDW were added to each vessel Standard Reference Materials and Reagents containing grain and plant samples, respectively. Sensor vessel contents were discarded. Prior to ICP-MS analysis, aliquots Four plant NIST SRMs, 1515 (Apple Leaves), 1547 (Peach of the digestates were further diluted 1.25-fold for the grain Leaves), 1573 (Tomato Leaves) and 1575 (Pine Needles), and samples and 2.5-fold for the plant samples.The overall dilution two grain RMs, 8433 (Corn Bran) and 8436 (Wheat Flour) was 250 or 500 (v/m) for the grain and plant samples, respect- were used for this study. The two grain RMs were prepared ively. The final solution for the ICP-MS analysis contained and characterized by Agriculture Canada and distributed by 2% HNO3 for both grain and plant samples.NIST with only best estimated values or estimated values The procedures used in the HNO3–H2O2 digestion were the provided. same as described above, with the addition of 2–5 ml of H2O2 High-purity concentrated HNO3 (68–71%, sub-boiling (30%) into each PFA liner either immediately after the addition double distilled in quartz, Seastar Chemicals, Sidney, British of HNO3, or after the overnight predigestion with HNO3 but Columbia, Canada) and, whenever required, certified 30% 1–2 h prior to the microwave digestion, or after the microwave H2O2 (analytical reagent grade, BDH, Poole, Dorset, UK) digestion with HNO3 followed by a second-stage digestion were used for the sample digestion.De-ionized distilled water using the same heating procedure (this will be called two-stage (DDW) was obtained from a three-column NANOpure water digestion hereafter). purification system with 18 mV cm specific resistivity Field samples spiked with standard solutions and reagent capability.blanks were digested in the same manner. Microwave Digestion System ICP-MS Measurement and Data Reporting A QWAVE-1000 microwave digestion system (Questron The ICP-MS system used was the Perkin-Elmer (Norwalk, Corporation, Mercerville, NJ, USA) equipped with tempera- CT, USA) ELAN Model 5000. This system and the operating ture and pressure regulation (through a sensor vessel) was parameters used were described in detail in previous work.23 used. This system is equipped with one sensor and up to eleven The ICP-MS system was calibrated by the method of sample digestion vessels in a 12 position rotating carousel.external standards with In as the internal standard. The reagent Each vessel consists of a PFA rupture disk, an outer pressure blank solution contained 2% of concentrated HNO3. Mixed vessel and a PFA inner liner with 100 ml capacity. Each vessel standard solutions containing 23 elements, Al, As, B, Ba, Cd, is capable of withstanding pressures up to 15.2 bar (220 psi) Co, Cr, Cu, Fe, Mn, Mo, Ni, Pb, Sb, Se, Sn, Sr, Th, Ti, Tl, U, and temperatures up to 200 °C.The pressure and temperature V and Zn, at the concentration levels of 1, 10 and 100 mg l-1, are measured by a transducer and a thermocouple, respectively, were prepared in reagent blank solutions. A single standard which are located in the sensor vessel. The heating program is solution containing 100 mg l-1 of Sn, 100 mg l-1 of Ca and controlled and monitored by a personal computer with QW405 177 mg l-1 of Cl was similarly prepared.This standard solution software installed, where power, ramp and dwell times, temwas used to derive the correction coecients for Ca and Cl perature or pressure reached, and the temperature and pressure interferences presented in Table 1, as well as to calibrate the limits in each of 9 heating steps can be programmed.After the system for the determination of Sn, Ca and Cl. In some cases, pre-set temperature (or pressure) in the sensor vessel is reached, a single mixed standard solution containing 50 mg l-1 of Mg the power is automatically regulated to maintain this value and K was used for the calibration of these elements using the unless the pre-set pressure (or temperature) limit is reached. ‘Omni’ range (i.e., reduced detection voltage). A 100 ml aliquot Prior to their use, the PFA inner liners were first soaked in of In solution (3.5 mg l-1) was added to 10 ml of each blank, 1+2+9 HNO3–HCl–H2O solution overnight.The following standard and sample solution. All signals were corrected by microwave cleaning procedure was carried out: a 40 ml aliquot subtraction of reagent blank signals. The calibration curves of 10% HNO3 was added to each liner; the digestion vessels were plotted linearly through zero for each isotope tested with were sealed and the temperature was raised to 165 °C in 10 min and without using the interference correction equations pre- and held at 165 °C for 10 min (the dwell time); after cooling, sented in Table 1.A unique number was assigned to each the contents of the vessels were discarded and the liners were correction equation used for the same isotope. The default thoroughly rinsed with DDW. equation number was zero, indicating that no correction was applied. Sample Digestion The signals at 43Ca and 35Cl were constantly measured to estimate the possible interferences from Ca and Cl and to Several digestion parameters were tested, including the amount of concentrated HNO3 (1, 2, 2.5 and 5 ml ), the dwell time derive the coecients used in the interference correction equations.The results for the analysed sample solution (in mg l-1) (10, 20, 30 and 40 min), and the use of H2O2 in the digestion. Unless otherwise specified, the following digestion procedure and for the sample solid (in mg g-1) obtained with and without interference corrections were simultaneously calculated and was used for the HNO3 digestion.Approximately 0.5 g (dry mass) of grain or plant sample was stored in the database. For field samples, these results were inspected prior to data reporting. When the sample solution weighed, to the nearest 0.1 mg, in the PFA inner liners including that of the sensor vessel. Then 2.5 ml (grain samples) or 5.0 ml analysed by ICP-MS contained relatively high concentrations of Ca or Cl that aected the results of relevant isotopes, the (plant samples) of concentrated HNO3 was added to each 798 Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12Table 1 Interference correction equations Element m/z Eqn. No. Correction equation Interferences Corrections for Cl interferences: V 51 1 51M-3.127 53M+0.3535 52M 35Cl16O V 51 2 51M-0.0015 35M* 35Cl16O Cr 53 1 53M-0.0005 35M* 37Cl16O As 75 1 75M-3.127 77M+2.548 82M 40Ar35Cl As 75 2 75M-0.0003 35M* 40Ar35Cl Se 77 1 77M-0.0001 35M* 40Ar37Cl Corrections for Ca interferences: Fe 57 1 57M-0.090 43M* 40CaOH Co 59 1 59M-0.00160 43M* 43CaO, 42CaOH Ni 60 1 60M-0.00600 43M* 43CaOH, 44CaO Correction for isobaric elemental interferences: Cd 114 1 114M-0.02747 118M 114Sn In 115 1 115M-0.016 118M 115Sn Pb 208 1 206M+207M+208M Isotope abundance variation * Coecient measured in each run.results calculated using the interference correction equations were selected and reported.Otherwise, the results obtained without correction were selected and reported. However, to demonstrate the interferences, results for the same element obtained using dierent isotopes or using the same isotope with dierent correction equations will be presented in this paper. RESULTS AND DISCUSSION HNO3 Digestion The minimum amount of concentrated HNO3 required for digestion was studied, in order to use a low dilution factor and to have a suciently low concentration of HNO3 (1–2%) in final solutions for ICP-MS analysis.Tests showed that undigested particles remained in the digestates when 1 or 2 ml of concentrated HNO3 for a 0.5 g of grain sample was used. While 2.5 ml of concentrated HNO3 was sucient for digesting the grain samples, 5.0 ml of concentrated HNO3 was required for the digestion of 0.5 g of SRM 1547 (Peach Leaves) to eliminate PFA inner liner damage by localized overheating. Thus, 2.5 and 5.0 ml of HNO3 were used for each 0.5 g of grain and plant samples, respectively.Fig. 1 The typical temperature and pressure profiles of samples using The microwave heating dwell time was studied using NIST the closed vessel microwave digestion: (a), digested in HNO3 alone; SRM 1515 (Apple Leaves). There are no significant dierences (b), digested in HNO3–H2O2, 2 ml H2O2 was added 16 h after the among the results obtained at dwell times of 10, 20, 30 and addition of 5 ml HNO3 but 1 h prior to the microwave digestion. 40 min, except for slightly lower Sb results at longer dwell times. Since only one type of sample, SRM 1515, was tested, which improved but did not guarantee the reproducibility of a dwell time of 20 min was chosen to minimize the possibility the temperature profiles. In addition, variations in temperature of incomplete digestion of any sample types being tested in profile can exist between digestion batches as well as between this study without a large increase in total time for microwave samples in the digestion vessels and the sensor vessel in the digestion and the cooling of the digestion vessels.same batch. All these factors could cause variations in digestion Fig. 1(a) demonstrates the variance in typical temperature eciencies for certain elements. To overcome this, a longer and pressure profiles of samples (in sensor vessels) with HNO3 predigestion time, or predigestion at an elevated temperature, digestion. After the temperature of the solution reached 100 °C, or using the HNO3–H2O2 digestion as discussed below may the pressure inside the vessel increased rapidly due to the be required.generation of CO2 and other gases from the decomposition of organic materials. When the pressure did not exceed the preset pressure limit of 13.8 bar (200 psi), the microwave power HNO3–H2O2 Digestion was regulated by the pre-set temperature of 165 °C. Otherwise, it was regulated by the pressure limit and the temperature Using HNO3 digestion, the diluted sample digestates had a light-yellow colour, indicating that the organic materials in started to decrease slowly to about 140–145 °C [Temperature 1 in Fig. 1(a)] or even 120 °C [Temperature 2 in Fig. 1(a)]. these samples were not completely decomposed. Also, as mentioned above, the digestion temperature profiles were not The digestion process controlled by pressure limit was not desirable since the temperature reached could vary greatly very reproducible.In addition, the recoveries of several elements such as Ti, Th, U, Ni, Al and V in the plant SRMs depending on sample organic matter content and result in variations in digestion eciencies. By pre-digesting samples were generally low. Thus digestion with both HNO3 and H2O2 was tested to find out if the use of H2O2 would improve the overnight, the organic materials were partially decomposed, Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12 799Table 2 Elemental contaminations in 2% solution of H2O2 (30% digestion eciencies for these elements.The addition of 2–5 ml m/v) analysed with ICP-MS of H2O2 guaranteed good contact between the temperature probe and the digestion solution. Tests (as described in the Concentration/mg l-1* experimental section) showed that in the presence of H2O2, the diluted digestate was lighter in colour after one-stage H2O2, (BDH H2O2, (BDH digestion and colourless after two-stage digestion.However, Element m/z Lot 116837–48528) Lot 95885–5063) no improvement in the accuracy of the analytical results was Al 27 0.8 32 observed with the HNO3–H2O2 digestion, except that carbon B 11 0.2 0.26 interferences on 52Cr by 40Ar12C and on 82Se by 35Cl35Cl12C, Ba 137 —† 0.06 Ca 43 16 62 although still severe, were reduced. Thus the use of H2O2 Cr 52 0.13 1.3 indeed reduced the C content in the digestate as reported by Cu 65 0.18 0.47 other studies,14 but a colourless diluted digestate is not an Mn 55 —† 0.1 indicator of good analyte recovery as determined by ICP-MS.Mo 98 —† 0.99 Using two-stage HNO3–H2O2 digestion or the one-stage diges- Ni 60 0.1 1 tion in which H2O2 was added after the overnight predigestion Pb 208 0.03 0.06 Sb 121 —† 0.02 with HNO3 but 1–2 h prior to the microwave digestion, the Sn 118 0.58 1.87 analytical precision was improved, especially for the elements Sr 86 —† 0.16 with low recoveries. This is probably due to the fact that, using Th 232 —† 0.059 those digestion procedures, the microwave power is tempera- Ti 47 —† 1.0 ture limit controlled during the entire dwell time as shown in Tl 205 —† 1.0 Fig. 1(b), resulting in less variations in digestion eciencies. Zn 66 0.8 2.44 This advantage was not observed using one-stage digestion * Concentrations of As, Cd, Co, Fe, Se, U and V were below DLs. after overnight predigestion with HNO3 and H2O2. No signifi- † Below DL. cant dierence was observed when 2 or 5 ml of H2O2 was used.The major concern in the use of H2O2 for digestion is the Table 3 Basic spectral interferences in ICP-MS for elements in biological samples (percentage natural relative abundance in parentheses) Isobaric Isobaric polyatomic ions elemental Element ions H2O, HNO3 Biological matrix 24Mg (78.99) 12C2 (97.817) 25Mg (10.0) 12C2H (97.797), 12C13C (2.176), 26Mg (11.0) 12CN (99.504), 13C2 (0.012), 28Si (92.23) 14N2 (99.202) 12CO (98.702) 29Si (4.67) 14N2H (99.187) 30Si (3.10) 14NO (99.401) 35Cl (75) 34SH (4.199) 43Ca (0.135) 44Ca (2.086) 14N216O (99.503) 12CO2 (98.505) 45Sc (100.0) 14N216OH (99.988) 12CO2H (98.49), 13CO2 46Ti (8.0) 46Ca (0.004) 14NO2(99.202) 47Ti (7.3) 35Cl12C (74.966), 31P16O (99.800) 48Ti (73.8) 48Ca (0.187) 32S16O (94.81), 31P16OH (99.785) 49Ti (5.5) 37Cl12C (23.934, 35Cl14N (75.497), 32S16OH (94.796) 50Ti (5.4) 50V (0.25) 35Cl14NH 51V (99.75) 35Cl16O (75.648), 37Cl14N (24.103) 52Cr (83.789) 40Ar12C (98.504) 35Cl16OH (75.637) 53Cr (9.501) 37Cl16O (24.303) 54Fe (5.80) 54Cr (2.365) 40Ar14N (99.202) 37Cl16OH (24.148) 55Mn (100.0) 40Ar14NH (99.187) 39K16O (93.11) 56Fe (91.72) 40Ar16O (99.401) 40Ca16O (96.776) 57Fe (2.20) 40Ar16OH (99.386) 40Ca16OH (96.762) 58Ni (68.077) 58Fe (0.33) 42Ca16O (0.639) 59CO (100.0) 42Ca16OH (0.639), 43Ca16O (0.14) 60Ni (26.223) 44Ca16O (2.096), 43Ca16OH (0.14) 61Ni (1.14) 44Ca16OH (2.095) 62Ni (3.63) 46Ca16O (0.003) 63Cu (39.17) 31P16O2 (99.60),, 35Cl14N2 (75.195) 64Zn (48.6) 64Ni (0.926) 32S16O2 (94.62), 31P16O2H, 32S32S (90.25), 48Ca16O (0.18) 65Cu (30.83) 32S16O2H (94.606), 33S16O2 (0.873), 32S33S (0.76) 37Cl14N2 (24.007), 48Ca16OH (0.18) 66Zn (27.9) 34S16O2 (7.986), 32S34S2 (4.565) 67Zn (4.1) 35ClO2 (75.497) 68Zn (18.8) 40ArN2 (99.805) 35Cl16O2H (75.486), 40Ar12C16O (98.307) 75As (100.0) 36Ar38ArH (0.001) 40Ar35Cl (75.52), 37Cl2H (5.855), 40Ar34SH (4.183), 77Se (7.63) Ar2H (0.598) 40Ar37Cl (24.103), 40Ca37Cl 82Se (8.73) 40Ar2H2(99.172) 12C35Cl2 (56.824), 34S16O3(4.175), 33S16O3H (0.795) 86Sr (9.86) 12C37Cl2(5.792) 88Sr (82.58) 94Mo (9.3) 94Zr (17.38) 95Mo (15.9) 96Mo (16.7) 98Ru (5.52) 97Mo (9.6) 98Mo (24.13) 98Ru (1.99) 63Cu35Cl (52.378) 800 Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12contamination from its impurities as encountered in this study sate for multiplicative noises, but it also compensates for matrix eects and instrument drift. In most ICP-MS quantitat- (Table 2). Thus, the majority of samples were digested in HNO3 alone. However, if high purity H2O2 reagent is available, ive analyses, multi-internal standards are used to cover dierent mass regions.25 Scandium, Y and In are usually used in the then the addition of H2O2 to samples several hours after the addition of HNO3 and use of one-stage HNO3–H2O2 micro- ‘light and medium’ mass region, while heavy elements such as Tb, Ho and Bi are used in the heavy mass region.1 Since SRMs wave digestion is recommended. 1515 (Apple Leaves) and 1573 (Tomato Leaves) contain Sc, Y and Tb, the use of these elements as internal standards was Calibration of the ICP-MS System eliminated. After comparison with other internal standards, In was found to be an eective internal standard covering the It was the intention of this study to apply the simple method whole mass range. Thus, as in previous work,23 In was chosen of calibration by external standards in conjunction with as the single internal standard in this study.internal standardization, as opposed to calibration by standard additions or the isotope dilution technique, in order to improve the sample throughput rate. Tests showed that there was no Interferences Correction and the Selection of Primary Isotopes significant dierence between the analytical results of a digested and a non-digested standard solution for the elements tested The spectral interferences in ICP-MS analysis have been investigated thoroughly in many studies.21,22,25–28 The basic in this study.This indicates that none of the tested elements were lost in the closed-vessel microwave digestion procedure, spectral interferences from reagent background and biological sample matrix are listed in Table 3. The background inter- which is in agreement with our previous findings using a dierent digestion reagent system.23 Hence, undigested stan- ferences from the plasma gases, air entrainment and solvent can be corrected by subtracting the blank signals.The isobaric dards prepared in reagent blank were used for the calibration. It is recognized that in actual samples the elements are not spectral interferences originating from the polyatomic ion species involving the sample matrix elements present more present in pure ionic form but are present as oxides, organometallic complexes and various minerals and thus may not neces- challenges. They may be eliminated by selecting a suitable isotope, or may be corrected or reduced by applying inter- sarily behave the same as in the standard solution.24 The internal standard method has been used in almost all ference correction equations.Although theoretically predictable, the actual extent of the interference at a given m/z value quantitative ICP-MS determinations. Not only does it compen- Table 4 The detection limits (DLs) and method detection limits (MDLs) for the primary isotopes and associated equations DL MDL Element m/z Eqn.No. Solution*/mg l-1 Solid†/mg g-1 Solution§/mg l-1 Solid‡/mg g-1 Al 27 0 0.2 0.1 3 1.5 As 75 0 0.05 0.025 0.4 0.2 As 75 1 0.2 0.1 0.6 0.3 As 75 2 0.1 0.05 0.3 0.15 B 10 0 0.5 0.25 2 1 Ba 137 0 0.02 0.01 0.6 0.3 Ca 43 0 10 5 22 11 Cd 114 1 0.04 0.02 0.12 0.06 Co 59 0 0.01 0.005 0.2 0.1 Co 59 1 0.01 0.005 0.3 0.15 Cr 53 0 0.1 0.05 0.6 0.3 Cr 53 1 0.1 0.05 0.6 0.4 Cu 65 0 0.1 0.05 0.5 0.25 Fe 57 0 3 1.5 5 2 Fe 57 1 3 1.5 5 2 K 39 0 — — 100¶ 50¶ Li 7 0 0.2 0.1 0.3 0.15 Mg 25 0 — — 3¶ 1.5¶ Mn 55 0 0.02 0.01 0.05 0.025 Mo 98 0 0.02 0.01 0.2 0.1 Ni 60 0 0.08 0.04 0.8 0.4 Ni 60 1 0.08 0.04 0.8 0.4 Pb 208 1 0.02 0.01 0.08 0.04 Sb 121 0 0.01 0.005 0.016 0.008 Se 77 0 0.2 0.1 0.3 0.15 Se 77 1 0.2 0.1 0.8 0.4 Sn 118 0 0.05 0.025 0.12 0.06 Sr 86 0 0.04 0.02 0.2 0.1 Th 232 0 0.01 0.005 0.1 0.05 Ti 47 0 0.2 0.1 0.6 0.3 Tl 205 0 0.003 0.0015 0.016 0.008 U 238 0 0.005 0.0025 0.012 0.006 V 51 0 0.02 0.01 0.3 0.15 V 51 2 — — — — Zn 66 0 0.2 0.1 0.3 0.15 * Derived from three times the standard deviation of undigested reagent blank solutions, n=10. † Calculated from the DL (in solution) in mg l-1 with the overall dilution factor of 500 (v/m).‡ Derived from three times the within-run standard deviation of duplicate digestion/analysis of the reagent blanks, SRMs and field samples (n=7–44) with the overall dilution factor of 500 (v/m). § Calculated from the MDL (in solid) in mg g-1 using the overall dilution factor of 500 (v/m). ¶ ‘Omni’ range used.Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12 801depends on the design and the operational conditions of the which will be simply called the ‘certified’ values hereafter. Other major isotopes suering interference from C species ICP-MS system as well as the concentrations of the sample matrix constituents. Hence, interferences vary greatly among include 44Ca(12C16O2), 45Sc(12C16O2H), 47Ti(12C35Cl) and 49Ti(12C37Cl), and the minor ones include 48Ti(36Ar12C), dierent operational conditions of the same ICP-MS system, among dierent ICP-MS systems and among dierent sample 53Cr(40Ar13C) and 60Ni(12C16O3) (Table 3).By selecting 10B, 53Cr and 77Se as the primary isotopes and applying matrices. For this reason, an interference investigation over the whole mass range was carried out prior to this study for suitable equations to correct for Cl interference on 53Cr and 77Se, as discussed below, the C interference on the determi- about 36 elements in 1% HNO3.Operational conditions were similar to the ones used in this study. The primary isotope nations of B, Cr and Se by ICP-MS was eliminated or reduced. The major Cl interferences in ICP-MS analysis include the and the correction equation were then selected (Table 4), based on the relative isotope abundance of the analyte, the calculated isotopes of 47Ti(35Cl12C), 49Ti(37Cl12C), 51V(35Cl16O), 53Cr(37Cl16O), 75As(35Cl40Ar) and 77Se(37Cl40Ar).The extent of relative abundance of the interfering species, the measured apparent analyte concentration produced by the interfering the Cl interference at m/z 51, 53, 75 and 77 is reflected in the corresponding measured interference correction coecients element at a given concentration and the matrix constituents of the biological SRMs studied, especially, C, Cl, Ca, K, Mg, listed in Table 1, which were the percentage ratios of the signals of a Cl standard at these m/z values to that at m/z 35 (35Cl).P and S. As already mentioned, the spectral interferences from poly- Without correction for the Cl interference (i.e., with equation number 0), the results of 75As and 77Se in NIST SRM 1573 atomic ion species involving C were more severe using this digestion procedure compared with sample dissolution by wet (Tomato Leaves) which contains #1% Cl were elevated (Table 6). Using the corresponding correction equation 2 given ashing procedures. This is reflected in Tables 5–7 by the elevated results for 52Cr(40Ar12C), 82Se(12C35Cl35Cl) and, to a in Table 1, the results for 51V and 75As were in good agreement with the certified values (Table 6).However, when the lesser degree, for 11B (tail interference by the strong 12C signal) compared with the certified or the best estimated values or corresponding equation 1 given in Table 1 was used for the correction, the results for 51V and 75As were all biased high compared with the reference, estimated or consensus29 values Table 5 Analytical results (mg g-1) for NIST SRM 1515 (Apple Leaves) and NIST SRM 1547 (Peach Leaves) digested in HNO3 alone; results of 75As (eqn 0 and 1), 11B, 52Cr, 57Fe (eqn. 0), 82Se and 51V (eqn. 1) are listed to demonstrate interferences NIST SRM 1515 NIST SRM 1547 Certified Found Certified Found Element m/z Eqn. No. Mean±Uncert. Mean±Uncert. Mean±Uncert. Mean±Uncert. Al 27 0 286±9 235±26 249±8 188±20 As 75 0 0.038±0.007 0.159±0.006§ 0.06±0.018 0.13±0.006§ As 75 1 0.038±0.007 0.132±0.028§ 0.06±0.018 0.18±0.02§ As 75 2 0.038±0.007 — 0.06±0.018 —‡ B 10 0 27±2 29.9±2.2 29±2 27.5±2.0 B 11 0 27±2 32±1.3 29±2 33.5±1.4 Ba 137 0 49±2 49.6±0.3 124±4 124±2.8 Ca 43 0 15260±150 16827±262 15600±200 16692±378 Cd 114 1 0.014* 0.016±0.002 0.026±0.003 0.027±0.002§ Co 59 0 0.09* 0.113±0.006§ 0.07* 0.074±0.008§ Co 59 1 0.09* 0.086±0.004§ 0.07* 0.048±0.006§ Cr 52 0 0.3* 1.56±0.22 1* 2.14±0.24 Cr 53 0 0.3* 0.38±0.06 1* 0.97±0.04 Cr 53 1 0.3* 0.311±0.010 1* 0.95±0.04 Cu 63 0 5.64±0.24 5.77±0.08 3.7±0.4 3.73±0.12 Cu 65 0 5.64±0.24 5.74±0.12 3.7±0.4 3.75±0.18 Fe 57 0 83±5 148±2 218±14 276±12 Fe 57 1 83±5 75.9±2.6 218±14 198±5 K 39 0 16100±200 14180±150 24300±300 19682±2500 Li 7 0 — —‡ — 0.113±0.02§ Mg 25 0 2710±80 2665±116 4320±80 4021±130 Mn 55 0 54±3 54.8±0.8 98±3 94.1±2.0 Mo 98 0 0.094±0.013 0.091±0.004§ 0.06±0.008 0.055±0.004§ Ni 60 0 0.91±0.12 1.37±0.014 0.69±0.09 1.0±0.06 Ni 60 1 0.91±0.12 0.782±0.026 0.69±0.09 0.37±0.04 Pb 208 1 0.47±0.024 0.443±0.010 0.87±0.03 0.847±0.024 Sb 121 0 0.013* 0.011±0.002 0.02* 0.019±0.002 Se 77 0 0.05±0.009 0.827±0.036 0.12±0.009 0.43±0.10 Se 77 1 0.05±0.009 0.793±0.034 0.12±0.009 0.51±0.10 Se 82 0 0.05±0.009 0.52±0.08 0.12±0.009 0.74±0.10 Sn 118 0 0.02* 0.050±0.028§ 0.2* 0.092±0.030§ Sr 86 0 25±2 26.1±0.2 53±4 54.9±1.2 Th 232 0 0.03* 1.37±0.014 0.05* 0.031±0.002§ Ti 47 0 — 9.02±3.6 — 7.92±1.0 Tl 205 0 — 0.018±0.008 — 0.023±0.002 U 239 0 0.0006* 0.0053±0.001 0.015* 0.0084±0.0006 V 51 0 0.26±0.03 0.236±0.018 0.37±0.03 0.321±0.022 V 51 1 0.26±0.03 0.324±0.006 0.37±0.03 0.69±0.06 Zn 66 0 12.5±0.3 12.8±0.4 17.9±0.4 17.7±1.0 * NIST reference value.‡ Below DL, refer to Table 4. § Above DL, but below MDL, refer to Table 4. 802 Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12Table 6 Analytical results (mg g-1) for NIST SRMs 1573 (Tomato Leaves) and 1575 (Pine Needles) digested in HNO3 alone; results of 11B, 52Cr, 82Se and 51V (eqn. 1) and results of 75As (eqn. 0 and 1) for SRM 1573 are listed to demonstrate interferences NIST SRM 1573 NIST SRM 1575 Certified Found Certified Found Element m/z Eqn. No. Mean±Uncert. Mean±Uncert. Mean±Uncert. Mean±Uncert. Al 27 0 1200* —d 545±30 482±58 As 75 0 0.27±0.05 0.46±0.04 0.21±0.04 0.22±0.010 As 75 1 0.27±0.05 0.46±0.04 0.21±0.04 0.28±0.034§ As 75 2 0.27±0.05 0.31±0.016 0.21±0.04 0.21±0.012 B 10 0 33.4¶ 35.8±2 17±2¶ 17.1±1.0 B 11 0 33.4¶ 37.8±1.4 17±2¶ 17.9±0.8 Ba 137 0 57±9¶ 55.1±1.8 7.2±0.8¶ 6.04±0.28 Ca 43 0 30000±300 30819±680 4100±200 4457±146 Cd 114 1 3*, 2.5±0.2¶ 2.72±0.08 0.22±0.06¶ 0.183±0.008 Co 59 0 0.6* 0.510±0.024 0.1*, 0.122±0.014¶ 0.100±0.006 Co 59 1 0.525±0.046¶ 0.462±0.024 0.1*, 0.122±0.014¶ 0.093±0.006 Cr 52 0 4.5±0.5 4.60±0.12 2.6±0.2 2.97±0.26 Cr 53 0 4.5±0.5 4.34±0.18 2.6±0.2 2.31±0.09 Cr 53 1 4.5±0.5 3.80±0.04 2.6±0.2 2.30±0.08 Cu 63 0 11±1 11.3±0.8 3±0.3 2.77±0.12 Cu 65 0 11±1 10.3±0.4 3±0.3 3.09±0.18 Fe 57 0 690±25 663±36 200±10 177±20 Fe 57 1 690±25 528±38 200±10 159±24 K 39 0 44600±300 41681±3523 3700±200 3469±560 Li 7 0 — 0.44±0.06 — 0.127±0.04§ Mg 25 0 7000* 7328±1130 1220±160¶ 1077±144 Mn 55 0 238±7 216±8 675±15 —d Mo 98 0 0.53±0.09¶ 0.573±0.034 0.15±0.05¶ 0.112±0.02 Ni 60 0 1.3±0.2¶ 2.06±0.12 3.5* 2.33±0.14 Ni 60 1 1.3±0.2¶ 0.95±0.12 3.5* 2.18±0.14 Pb 208 1 6.3±0.3 6.05±0.14 10.8±0.5 10.3±0.34 Sb 121 0 0.036±0.007¶ 0.027±0.004 0.2* 0.154±0.014 Se 77 0 0.054±0.006¶ 0.74±0.19 0.047±0.005¶ 0.069±0.10§ Se 77 1 0.054±0.006¶ 0.056±0.012§ 0.047±0.005¶ 0.050±0.016§ Se 82 0 0.054±0.006¶ —‡ 0.047±0.005¶ —‡ Sn 118 0 — — — 0.26±0.08 Sr 86 0 44.9±0.3 41.1±0.6 4.8±0.2 4.41±0.068 Th 232 0 0.17±0.03 0.056±0.012 0.037±0.003 0.018±0.004§ Ti 47 0 56±39¶ 23.7±1.8 13.7¶ 6.95±0.52 Tl 205 0 0.05* 0.032±0.002 0.05* 0.046±0.002 U 238 0 0.061±0.003 0.026±0.002 0.02±0.004 0.0128±0.0026 V 51 0 1.2±0.2¶ 1.14±0.14 0.39±0.07¶ 0.335±0.064 V 51 1 1.2±0.2¶ 4.55±6.4 0.39±0.07¶ 0.57±0.12 V 51 2 1.2±0.2¶ 1.16±0.06 0.39±0.07¶ 0.32±0.02 Zn 66 0 62±6 61.9±0.2 67±9¶ 64.3±2.4 * NIST reference value.‡ Below DL. § Above DL, but below MDL, refer to Table 4. ¶ Consensus value.29 d Out of the linear calibration range, further dilution required. (Tables 5–7). This is because in both equations 1 (Table 1), and hydroxides species at m/z values 57, 59 and 60 is reflected in the corresponding interference correction coecients listed the signal at m/z 52 or 82 was involved, which suers severe C interference.Thus, although successfully applied to the in Table 1. These were the measured percentage ratios of the signals of a Ca standard at these m/z values to that at m/z 43 ICP-MS analysis for sediment and soil samples,23 correction equations 1 were not suitable for the determination of V and (43Ca). In this study, 57Fe was selected for the Fe determination over 54Fe and 56Fe which suered severe interference from As in biological samples digested with this procedure because of the relatively high C concentrations.The Cl interference on 40Ar14N or 40Ar16O (Table 3). The Ca interferences at m/z 57 were corrected using equation 1 (Table 1). Otherwise, the Fe 53Cr was relatively low and there was little dierence between the 53Cr results in SRM 1573 obtained with and without the results could be biased high, especially for SRMs 1515 and 1547 (Table 5). All Ni isotopes suer from Ca interference use of the correction.The Cl levels were much less in SRMs 1515, 1547, 1575 and in the two grain RMs (up to 0.04%) than (Table 3). Because of the isobaric interference of 58Fe on 58Ni and the much lower isotope abundance of 61Ni and 62Ni, 60Ni in SRM 1573. Consequently, correction equations 2 for the determination of 51V and 75As and equations 1 for 53Cr and was selected as the primary isotope and equation 1 listed in Table 1 was used to correct the Ca interference on its determi- 77Se were either not applied or, when applied, produced little dierence (Tables 5–7).Apparently, unknown interferences nations. The generally low recoveries for Ni obtained using 60Ni with equation 1 in Tables 5–7 might be caused by caused the biased high results for 77Se in SRMs 1515 and 1547. Because the digestion eciencies for Ti were generally low incomplete digestion. A similar equation was also applied to correct the Ca interference on 59Co (Table 1).The interference (#50%) as shown from the results at 47Ti in Table 6, no eort was made either to compare the Ti results using various Ti of Ca on 65Cu was much weaker and the correction was not applied. isotopes or to correct for the possible interferences at 47Ti. The Ca levels in biological samples may be as high as Plant samples may contain relatively high K concentrations, which, as 39K16O, may interfere in the determination of Mn at 1.5–3%(Tables 5–7). The extent of interference from Ca oxides Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12 803Table 7 Analytical results (mg g-1) for NIST RMs 8433 (Corn Bran) and 8436 (Durum Wheat Flour) digested in HNO3 with the overall dilution factor of 250 (v/m); results of 11B, 52Cr, 57Fe, 82Se and 51V (eqn. 1) are listed to demonstrate inferferences NIST RM 8433 NIST RM 8436 Best estimated Found Best estimated Found Element m/z Eqn. No. Mean±Uncert. Mean±Uncert. Mean±Uncert. Mean±Uncert.Al 27 0 1.01±0.55 0.49±0.14§ 11.7±4.7 9.5±0.9 As 75 0 0.002±0.002 —‡ 0.03* 0.018±0.004§ As 75 1 0.002±0.002 —‡ 0.03* —‡ B 10 0 2.8±1.2 3.11±0.12 — 0.82±0.08 B 11 0 2.8±1.2 3.77±0.56 — 1.21±0.76 Ba 137 0 2.4±0.52 2.42±0.10 2.11±0.47 2.31±0.08 Ca 43 0 420±38 467±23 278±26 266±10 Cd 114 1 0.012±0.005 0.0135±0.0016§ 0.11±0.05 0.119±0.004 Co 59 0 0.006±0.006 0.003±0.004§ 0.008±0.004 0.0044±0.0035§ Co 59 1 0.006±0.006 —‡ 0.008±0.004 0.0039±0.0006§ Cr 52 0 0.101±0.087 0.46±0.10 0.023±0.009 0.44±0.20 Cr 53 0 0.101±0.087 0.11±0.04§ 0.023±0.009 0.11±0.11§ Cu 63 0 2.47±0.4 2.71±0.10 4.3±0.69 4.51±0.04 Cu 65 0 2.47±0.4 2.71±0.10 4.3±0.69 4.45±0.14 Fe 57 0 14.8±1.8 15.9±1.8 41.5±4 42.4±2.0 Fe 57 1 14.8±1.8 13.9±1.6 41.5±4 41.2±1.8 K 39 0 566±75 433±479 3180±140 3249±490 Li 7 0 — —‡ — 0.096±0.010 Mg 25 0 818±59 797±115 1070±80 1037±64 Mn 55 0 2.55±0.29 2.58±0.12 16±1 15.8±0.76 Mo 98 0 0.252±0.039 0.258±0.044 0.7±0.12 0.762±0.024 Ni 60 0 0.158±0.054 0.086±0.082§ 0.17±0.08 0.13±0.02§ Ni 60 1 0.158±0.054 0.069±0.082§ 0.17±0.08 0.12±0.02§ Pb 208 1 0.14±0.034 0.138±0.010 0.023±0.006 0.028±0.001§ Sb 121 0 0.0045* —‡ — —‡ Se 77 0 0.045±0.008 —‡ 1.23±0.09 1.39±0.04 Se 82 0 0.045±0.008 0.22±0.22§ 1.23±0.09 1.49±0.01 Sn 118 0 — † — —‡ Sr 86 0 4.62±0.56 4.97±0.15 1.19±0.09 1.22±0.02 Th 232 0 — 0.003±0.006§ — 1.21±0.02 Ti 47 0 — 4.89±0.08 5* 6.21±1.07 Tl 205 0 — —‡ — —‡ U 238 0 — —‡ — —‡ V 51 0 0.005±0.002 0.006±0.003§ 0.021±0.006 0.022±0.002§ V 51 1 0.005±0.002 0.114±0.028§ 0.021±0.006 0.112±0.008§ Zn 66 0 18.6±2.2 19.3±0.76 22.2±1.7 24.0±0.8 * Estimated value.‡ Below DL, DLs are half of those listed in Table 4 due to a 2-fold dilution factor dierence (250 v/m versus 500 v/m). § Above DL, but below MDL, MDLs are half of those listed in Table 4 due to a 2-fold dilution factor dierence (250 v/m versus 500 v/m). m/z 55. The certified K concentrations in the plant SRMs using this ICP-MS system. For the isotopes at higher m/z values, where spectral interferences are generally less, the range between 0.4–4.4%, which is equivalent to 8–88 mg l-1 of K in the diluted sample solutions.Using this ICP-MS isotopes with higher abundance were usually preferred. To be consistent with the isotopes used for water and sediment system, these concentrations would produce apparent Mn concentrations of below 0.3 mg l-1 (or 0.15 mg g-1 in sample analysis,23 the isotopes 47Ti, 114Cd, 121Sb, 137Ba, 205Tl and 238U were selected, although 111Cd, or 123Sb, or 135Ba and 138Ba, or solids) at m/z 55.Because the SRMs and RMs studied contain relatively high Mn concentrations (2.5–675 mg g-1), the K 203Tl can also be used for accurate determinations of these elements in biological samples. interference on 55Mn was negligible (Tables 5–7). The main considerations in the selection of the other primary isotopes listed in Table 4 are given below briefly. The plant Accuracy and Precision and grain samples contain relatively high Mg concentrations.Thus, an isotope with low relative abundance is preferred and The method was applied to the analyses of field samples from environmental monitoring, assessment and remediation pro- 25Mg was selected instead of 26Mg, which may be subject to interference from the tail of a strong Al signal at m/z 27. For jects, in which plant SRMs and grain RMs were analysed for quality control (QC) purposes. The data presented in Tables the Cu determinations, the results measured at both m/z 63 and 65 show good agreement with the certified values 5–7 represent all these QC results performed over a six month period.The accuracy of the method (for the primary isotopes (Table 5–7). To be consistent with this laboratory’s analysis of water samples, with potentially high Na concentrations, 65Cu and within the analytical range of the method) can be judged from the spike recoveries presented in Table 8 and the results was selected to avoid 23Na40Ar interference on 63Cu.For Zn determination, 66Zn was selected because the interferences at for SRMs and RMs presented in Tables 5–7. It is realized that for a given isotope, a low digestion eciency in conjunction m/z 66 are much less severe than those at m/z 64, 67 and 68 (Table 3). For Mo determination, the interferences of with the existence of an interference may produce an acceptable recovery. The precision of the method can be judged from the 40Ar39K16O on 95Mo and of 40Ar41K16O on 97Mo reported in another study26 were negligible for the plant and grain samples uncertainties (at the 95% confidence level) of the results presented in Tables 5–7, and the detection limits (DLs) and tested using this ICP-MS system. The isotope 98Mo was selected because it has the highest relative abundance among method detection limits (MDLs) listed in Table 4.The MDLs are 2–10 times greater than the corresponding DLs. The DLs Mo isotopes.The interference from Ru at 98Mo was negligible 804 Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12Table 8 Predigestion spike recoveries for field plant samples Field Samples Spike* % Recovery Element m/z Eqn. No. Conc. range/mg l-1 Conc./mg l-1 Mean±Uncert. Al 27 0 95–238 50 99.5±3.2 As 75 0 0.4–1.4 5 98.6±4.3 As 75 0 0–0.7 5 98.5±10 B 11 0 117–787 50 99.8±4.4 Ba 137 0 81–139 50 99.5±3.2 Ca 43 0 10000–22000 5000 101±3 Cd 114 1 0.5–5 5 93.6±2.2 Co 59 0 2–10 5 101±4 Co 59 1 2–10 5 101±4 Cr 52 0 3–6 5 100.4±4.7 Cr 53 0 0.8–2 5 99.7±5.3 Cu 63 0 7–17 5 97.0±4.8 Cu 65 0 7–17 5 96.9±4.3 Fe 57 0 200–450 50 108±5 Fe 57 1 200–400 5 98.0±6 K 39 0 4000–20000 1000 95.0±15 Li 7 0 9–30 50 103.3±1.9 Mg 25 0 5000–8000 1000 97.7±3.2 Mn 55 0 300–1500 50 96.2±3.1 Mo 98 0 2–30 5 102±2 Ni 60 0 9–45 5 100±5 Ni 60 1 9–45 5 98.8±5.3 Pb 208 1 0.4–1.2 5 96.2±3.4 Sb 121 0 0.015–0.04 5 95.1±3.8 Se 77 0 0.1–1 5 95.4±6 Se 82 0 -25–-2 5 106±20 Sn 118 0 0.03–0.3 5 106.5±4.8 Sr 86 0 100–170 50 101.1±2.5 Th 232 0 0.01–0.08 5 97.1±7.5 Ti 47 0 6–25 5 108±5 Tl 205 0 0.01–0.04 5 96.4±2.1 U 238 0 0.008–0.03 5 102.1±2.5 V 51 0 0.8–2.3 5 103.6±5.1 V 51 1 1.6–3.7 5 103.5±5.2 Zn 66 0 110–260 5 99.2±3.0 * Spike recovery is calculated as (conc.of spiked sample)/(conc. of sample+spike). obtained with the HNO3–H2O2 digestion (H2O2, BDH Lot samples as discussed previously, which produced nonreproducible digestion eciencies for elements that are dicult 116837–48528) were similar to those listed in Table 4, except the DL for Sn was poorer (0.2 mg l-1).to digest. In contrast to the plant SRMs, most of the mean recoveries Based on Tables 5 and 6, the elements tested for the four plant SRMs may be loosely grouped as follows. The first group for the majority of the elements tested in the grain RMs were within 85–115%. However, the MDLs for the determination contains 12 elements, As, B, Ba, Ca, Cd, Cu, Mg, Mn, Mo, Pb, Sr and Zn, with the majority of mean recoveries within of As, Co, Cr, Ni and Se were not suciently low.Although the results for several isotopes/equations in these 90–115%. The 7 elements, Al, Co, Cr, Fe, K, Sb and V, with most mean recoveries varying from 70–100% make up the SRMs and RMs are biased high because of interference or are biased low because of poor digestion eciencies, the pre- second group. The third group consists of 4 elements, U, Th, Ti and Ni with mean recoveries mainly within 40–80%.The digestion spike recoveries shown in Table 8 are all excellent. This is a perfect example illustrating that additive spectral fourth group includes elements such as Li, Sn and Tl that do not have available certified data. The fifth group consists of interferences and low digestion eciencies cannot be detected from spike recoveries. Therefore, acceptable spike recoveries Se and Cl. The recoveries for Se determined at m/z 77 with correction equation 1 varied so greatly that the accuracy for alone may not be sucient to prove the accuracy of a method.Se analysis of field samples is unknown. The Cl was not a target analyte in this study and its results were lacking in both CONCLUSION accuracy and precision. This was caused primarily by the variable HCl residues in the digestion PFA liners and sample The closed-vessel microwave digestion of plant and grain materials in concentrated HNO3 alone or in a mixture of tubes which were soaked in 1+2+9 HNO3–HCl–H2O solution for cleaning.It seems that the plant leaves contain siliceous HNO3 and H2O2 is favorable for trace element analysis by ICP-MS. The digestion procedure is simple. With 3 sets of 12 material to which the elements listed in the second and third groups are partially bound. The siliceous materials were not digestion vessels and 3 additional sets of 12 inner liners, 6 batches of samples were processed by overnight predigestion decomposed by the HNO3 digestion, resulting in lower recoveries.The precision for the determination of the elements listed with microwave digestion the following day. None of the elements tested were lost in the digestion. With the overall in the second and third group is generally poorer than for those listed in the first group. This is probably caused by the dilution factor of 500 (v/m) for the plant SRMs and of 250 (v/m) for the grain RMs, the samples could be analysed by a variations in the actual digestion temperature profiles of the Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12 8053 Buckley, W. T., and Ihnat, M., Fresenius’ J. Anal. Chem., 1993, simple ICP-MS method in which undigested external standards 345, 217. prepared in reagent blanks were used for calibration with In 4 Munro, S., Ebdon, L., and McWeeny, D. J., J. Anal. At. Spectrom., as the internal standard. The background interferences were 1986, 1, 211.corrected by subtracting the blank signals. The isobaric elemen- 5 Goossens, J., De Smaele, T., Moens, L., and Dams, R., Fresenius’ tal and polyatomic ionic interferences, especially those from J. Anal. Chem., 1993, 347, 119. 6 Beauchemin, D., McLaren, J. W., and Berman, S. S., J. Anal. At. C, Ca and Cl, were eliminated, corrected, or reduced by Spectrom., 1988, 3, 775. selecting the suitable primary isotopes and applying the appro- 7 Lyon, T. D. B., Fell, G. S., McKay, K., and Scott, R.D., J. Anal. priate interference correction equations. Although this diges- At. Spectrom., 1991, 6, 559. tion procedure did not decompose the siliceous material or 8 Gu� nther, K., von Bohlen, A., Paprott, G., and Klockenka�mper, minimize the C concentration in the samples as the wet ashing R., Fresenius’ Z. Anal. Chem., 1992, 342, 444. 9 Ebdon, L., Fisher, A. S., Worsfold, P. J., Crews, H., and Baxter, procedure does, the majority of recoveries for the 22 elements, M., J.Anal. At. Spectrom., 1993, 8, 691. Al, As, B, Ba, Ca, Cd, Co, Cr, Cu, Fe, K, Li, Mn, Mg, Mo, 10 Evans, S., and Kra�henbu� hl, U., Fresenius’ Z. Anal. Chem., 1994, Pb, Sb, Sn, Sr, Tl, V and Zn in the pre-digested spikes and/or 349, 454. in the four plant SRMs and two grain RMs were within 11 Krachler, M., Radner, H., and Irgolic, K. J., Fresenius’ Z. Anal. 70–115%. Chem., 1996, 355, 120. The limitations of the method include the low recoveries for 12 Liu, H., Montaser, A., Dolan, S.P., and Schwartz, R. S., J. Anal. At. Spectrom., 1996 11, 307. the determination of U, Th, Ti and Ni in the plant SRMs. 13 Amarasiriwardena, D., Krushevska, A., Argentine, M., and Barnes, Also, if the Ca and Cl concentrations in the solutions are so R. M., Analyst, 1994, 119, 1017. high that the ICP-MS signals for 43Ca and 35Cl are saturated 14 Krushevska, A., La� sztity, A., Kotrebai, M., and Barnes, R. M., or the corresponding correction equations are no longer valid, J.Anal. At. Spectrom., 1996, 11, 343. then the results for the determination of 57Fe, 60Ni and 59Co, 15 Subramanian, K., Spectrochim. Acta, Part B, 1996, 51, 291. 16 Beary, E. S., and Paulsen, P. J., Anal. Chem., 1993, 65, 1602. or for the determination of 51V, 75As, 77Se and 53Cr will be 17 Pepelnik, R., Prange, A., and Niedergesa�ß, R., J. Anal. At. biased. In addition, the method detection limits are not suc- Spectrom., 1994, 9, 1071. iently low for the determination of Se in all SRMs and of As, 18 Evans, S., and Kra�henbu� hl, U., J. Anal. At. Spectrom., 1994, 9, 1249. Co, Cr and Ni in some plant SRMs and grain RMs tested. 19 Matusiewicz, H., Sturgeon, R. E., and Berman, S. S., J. Anal. At. The detection limit can be further improved by reducing the Spectrom., 1989, 4, 323. 20 De Bra�tter, V. E. N., Bra�tter, P., Reincke, A., Schulze, G., Alvarez, overall dilution factor, as long as the increased concentrations W. O. L., and Alvarez, N., J. Anal. At. Spectrom., 1995, 10, 487. of the interfering elements in the solution do not hamper the 21 Vaughan, M. A., and Horlick, G., Appl. Spectrosc. 1986, 40, 434. accuracy of the determination. Furthermore, the maximum 22 Tan, S. H., and Horlick, G., Appl. Spectrosc., 1986, 40, 445. amount of the biological sample to be eciently digested is 23 Wu, S., Zhao, Y., Feng, X., and Wittmeier, A., J. Anal. At. limited by the maximum working pressure of 13.8 bar (or 200 Spectrom., 1996, 11, 287. psi) allowed in the closed-vessel microwave digestion system 24 Nadkarni, R. A., Anal. Chem., 1984, 56, 2233. 25 Tothill, P., Matheson, L. M., Smyth, J. F., and McKay K., J. Anal. used in this study. At. Spectrom., 1990, 5, 619. Overall, the explored method is simple, rapid, and suitable 26 Vanhoe, H., Goossens, J., Moens, L., and Dams, R., J. Anal. At. for the analysis of at least 22 elements in a variety of plant Spectrom., 1994, 9, 177. and grain samples for environmental monitoring and assess- 27 Vaughan, M. A., and Templeton, D. M., Appl. Spectros., 1990, ment projects. 44, 1685. 28 Shao, Y., and Horlick, G., Appl. Spectros., 1991, 45, 143. 29 Compilation of Elemental Concentration Data for NBS Clinical, Biological, Geological, and Environmental Standard Reference REFERENCES Materials, U.S. Department of Commerce, National Bureau of Standards, 1987. 1 Shiraishi, K., Takaku, Y., Yoshimizu, K., Igarashi, Y., Masuda, K., Mclnroy, J. F., and Tanaka, G., J. Anal. At. Spectrom., 1991, Paper 6/07217H 6, 335. Received October 22, 1996 2 Thompson, J., and Ward, N. I., Journal of Micronutrient Analysis., 1989, 6, 85. Accepted April 15, 1997 806 Journal of Analytical Atomic Spectrometry,

ISSN:0267-9477    

DOI:10.1039/a607217h

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Studies of Analyte Particle Transport in a Particle Beam–Hollow Cathode Atomic Emission Spectrometry System JIANZHANG YOU, MELISSA A. DEMPSTER AND R. KENNETH MARCUS* Department of Chemistry, Howard L . Hunter L aboratory, Clemson University, Clemson, SC 29634-1905, USA The particle beam–hollow cathode glow discharge atomic on the GD plasma operation and energetics. The most direct emission spectrometry system appears to be a viable analytical way to accomplish this task is the application of an aliquot method for volume-limited liquid samples. Systematic (1–100 ml ) of the sample onto a target cathode ex situ, followed enhancements and reductions in instrumental response were by a solvent evaporation step prior to introduction into the related to the analyte transport eciency through the particle plasma source/cell.3,4,6,7 beam interface.The eects of liquid flow rate and analyte Because of the increasing need for chemical speciation input concentration on analyte transport eciency were information in both environmental and biological analyses, evaluated in an attempt to explain these eects on the the use of chromatographic separation techniques has seen analytical signal.Two specifically designed collectors were phenomenal growth over the last decade. While direct coupling mounted within the glow discharge source for sampling analyte of liquid chromatography (LC) with flame atomic absorption particles passed through the particle beam interface.Atomic and inductively coupled plasma (ICP) sources is relatively absorption spectrometry and scanning electron microscopy straightforward, LC solvents (and buers) can strongly aect were applied to determine analyte transport eciencies and the nebulization and spectral characteristics of these particle size distributions, respectively. In general, transport devices.13–15 In addition, many of these sorts of separations eciencies of 4–18% were achieved.Improvements in analyte present the need to detect elements that are not readily response, noted in previous studies, on addition to the sample determined in atmospheric pressure sources, including nitroof concentrated HCl (1+5, v/v) can now be attributed to gen, oxygen and halides. The low pressure GD environment enhanced analyte transport through the particle beam interface presents the possibilities to look at these ‘atmospheric’ by virtue of an increase in the size of the desolvated analyte elements.Clearly, the use of a deposition-type approach to particles. The size distribution of analyte particles appears to GD detection of LC eluents is not practical. To be most change with the distance from the center to the edge area of pragmatic, one would look to choose an interface that possesses the sample collector as most particles follow a straight the capability of operating with a variety of solvent polarities, pathway to the hollow cathode, with very little evidence of over a wide range of liquid flow rates and reasonable analyte dispersion.Typical particle sizes lie in the range 2–8 mm. transport eciency without the deleterious eects of solvent carry-over to the plasma region. In fact, some approaches have Keywords: Particle beam; hollow cathode atomic emission; looked to capitalize on methods developed for liquid chromato- transport eciency; particle size graphy–mass spectrometry (LC–MS) to introduce liquid samples into the GD.Among these have been the use of The glow discharge (GD) has a long history of application in moving belt5,9 and particle beam (PB) interfaces.10–12 the area of metal and alloy analyses.1 The application of GD The use of a PB interface for LC–MS, generally known as techniques in atomic and mass spectrometries has been the MAGIC LC–MS interface, was first developed by expanded to cover the range of direct solids elemental analysis Willoughby and Browner.16 The development and function of of both conducting and non-conducting materials by use of the PB interface for LC–MS meets many important criteria radiofrequency (rf ) powered devices.2 In all of these applifor low pressure sample presentation to an ionization volume, cations, the combination of cathodic sputtering (atomization) including: (1) highly ecient removal of solvent, (2) mechanical and electron and metastable species collisions in the gas phase simplicity and ease of operation, (3) analysis capabilities for a (excitation and ionization) present a very ecient means of wide range of sample/solvent volatility, thermal lability and direct solids analysis. The possibility of developing a GD polarity, (4) choice of detection (ionization) techniques and device as an analytical tool for both liquid and solid sample (5) preservation of sample and chromatographic integrity.The analyses has been a challenge for many chemists in this area.PB interface involves three distinct processes: (1) nebulization Researchers have looked to utilize the easily controlled negative and aerosol formation, (2) desolvation and (3) momentum glow of the plasma to eect excitation and ionization of analyte separation. The production of the primary aerosol from the species originating in the liquid state.3–11 In each of these LC flow can in principle be accomplished by any number of applications, the sample is presented to the GD as a dried nebulizers common in atomic spectrometry such as concentric solution residue brought into the gas phase through cathodic and cross-flow pneumatic, ultrasonic and thermally-assisted sputtering.In the ideal case though, sputtering would not be devices. The desolvation chamber is a very important region needed, and the plasma geometry and conditions could be in which the aerosols produced by the nebulizer evaporate at optimized solely for optical emission and ionization. An or near atmospheric pressure.Obviously, heat transfer has a example of this approach has recently been presented by Olson great impact on the aerosol desolvation process, as pressure et al.,12 who used an rf-GD as an ionization source for volatile and temperature changes occur in this region. Helium is usually analytes by gas chromatography–mass spectrometry chosen as the nebulizer or sheath gas because of its high (GC–MS). The major hindrance to the direct analysis of thermal conductivity, enhancing heat transfer to the aerosol solution-based samples is the inability of the low pressure from the heated chamber components.The rate of solvent (single Torr), low temperature (slightly above room temperavaporization will vary depending on the solvent properties, an ture) plasma to eect analyte desolvation and then subsequently to eliminate the deleterious eects of solvent vapors especially important consideration for solvents such as water Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12 (807–815) 807which demand high temperatures. The momentum separator input concentration, the use of chelated analytes and the addition of HCl as a carrier were evaluated to explore the serves to enrich the PB by largely removing the solvent vapors. A typical momentum separator is composed of skimmers relationship between the analyte transport eciency and response. Direct observation of the introduced particles using separating two dierential pumping stages.The mixture of sample particles and solvent vapors is forced through the scanning electron microscopy (SEM) allows assessment of the role of particle size in the transport process. By establishing a skimmers and attains momentum along the orifice axes directly related to their mass. Subsequently, the enrichment of sample relationship between sample transport eciency and analyte response, attention can be focused on improving the cumulative particles is accomplished based on the fact that light species (vapors) gain o-axis momenta and are skimmed from the nebulization–desolvation–transport processes towards the goal of higher transport eciencies, better calibration quality and expanding beam in preference to the solute (analyte) particles.The application of PB-LC–MS interfaces relies most heavily minimized matrix eects. As such, the practical use of the PB-HC-AES approach to element-specific chromatographic on the ability to operate with almost any LC solvent at flow rates of up to 2 ml min-1, with the resulting pressure in the detection can be further developed. MS ion volume easily being less than 1 mTorr.Because of this ecient solvent removal, the PB interface is attractive for EXPERIMENTAL sample introduction in LC–MS applications because of its compatibility with various ionization techniques including Glow Discharge Source electron impact, chemical ionization and fast atom bombard- The basic design of the PB-HC-AES interface and source has ment (FAB).16–22 However, studies of commercial PB interfaces been described previously,10,11 so only a cursory description is have indicated some problems for quantitative analysis at low presented here.The 3.25 mm diameter graphite hollow cathode concentrations. Bellar et al.17 first reported the existence of (HC) is mounted in the center of a heated stainless-steel cube, non-linear behavior in the region of the detection limit in the termed the thermoblock. The entire thermoblock is heated to form of depressed analyte responses. This phenomenon could #220 °C by a pair of commercially available cartridge heaters be remedied by what was described as a carrier eect that (Scientific Instrument Service, Model SC 2515, Ringoes, NY, resulted in increased target analyte ion abundance in the USA), with the block temperature measured by a W–Re presence of co-eluting compounds.The PB carrier process is thermocouple.Analyte particles enter the plasma area through postulated to be initiated by neutral molecules capable of a 1.53 mm aperture in the wall of the HC perpendicular to the forming molecular clusters in solution and/or in the latter cathode axis. Particles impacting on the opposite wall are stages of the transport processes. Apel and Perry18 have vaporized and swept into the HC–GD region by the perpen- explained the observed non-linear behavior using a high-pass dicular flow of the He discharge gas.Discharge gas inlets, filter model which addresses a hypothetical cut-o level where electrical feedthroughs, PB interface fittings and vacuum pump- the analyte particles suer a dramatic loss of their transport ing ports are also fixed to the thermoblock. The He discharge eciencies if their particle size is below the cut-o level. Simply, gas pressure within the block is monitored by a thermocouple the size/mass of the particles at low concentration is not gauge (Teledyne Hastings-Raydist, Model DV-4D, Hampton, sucient to ensure their direct transport through the momen- VA, USA).The GD is powered by a Kepco (Flushing, NY, tum separator. Further investigation by Ho et al.19 suggested USA) Model BHK 2000 supply operating in a constant-current that many compounds exhibit linear calibration graphs with mode. Typical source operating conditions are 3.5 Torr He particular co-eluting compounds. The trend of structural simisource pressure and a discharge current of 30 mA (#450 V).larity between a carrier and the analytes of interest is generally accepted and employed in choosing co-eluting carriers. The complex mechanism of the carrier eect, which involves physi- Particle Beam Interface cal and chemical processes, clearly indicates that no single universal additive exists and therefore there is much room for The PB-LC–MS interface employed here consists of a thermal concentric nebulizer used to generate a finely dispersed aerosol, further studies.Based on the above-cited studies, it would not be surprising a stainless-steel spray chamber for desolvation, and a twostage momentum separator which serves to remove residual to see non-ideal transport responses in all PB-based interfacing schemes. In previous studies of the feasibility of the PB sample solvent vapor and reduce the backing pressure. The liquid sample passes through a fused-silica capillary (110 mm id) introduction for flow injection-type liquid analysis by hollow cathode atomic emission spectrometry (HC-AES),10,11 there housed within a stainless-steel capillary (0.5 mm id).A fine aerosol is generated as He is introduced into the gap between was a definitive signal enhancement observed by the addition of concentrated HCl (1+5, v/v). This carrier eect yielded as the fused-silica and steel capillary (i.e., concentric, pneumatic nebulization).A dc potential applied across the steel capillary much as a 1000-fold signal enhancement in aqueous solutions, overcoming severe depressive eects seen for samples of high causes resistive heating and adds a thermal component to the nebulization process. The generated aerosol is directed into salt content (5 M NaNO3–0.1 M KOH).10 As with nebulizationbased atomic spectrometric methods employing flames and the stainless-steel desolvation chamber which is heated by electrical tape and the temperature (typically 220 °C) is moni- atmospheric pressure plasmas, there is a need to delineate the possible eects of sample transport and source operation tored by a W–Re thermocouple. In the system under study here, the commercial desolvation chamber has been modified conditions in the observed enhancement.In the absence of any changes in the HC discharge electrical characteristics (voltage with an additional gas inlet to introduce a supplementary He gas flow into the desolvation chamber.11 A vacuum gauge and current), it was suggested that addition of the acid enhanced analyte transport in particle beam–hollow cathode (TFS Technologies, Model PDR-D-1, Albuquerque, NM, USA) is mounted to the desolvation chamber to monitor the glow discharge atomic emission spectrometry (PB-HC-AES) in a manner similar to that observed in ‘organic’ LC–MS.pressure (typically#400 Torr). The flow rate of the supplemental He is measured by a mass flow meter (AALBORG Therefore, the understanding of analyte transport through the PB interface is vital to understanding the mechanism(s) of Instruments & Control, GFM-1700, Monsey, NY, USA).A thermocouple (Omega Engineering, Type J, Stanford, CT, signal enhancement (or suppression in the example above) for the future application of PB-HC-AES. USA) is mounted through the chamber walls to allow direct measurement of the gas-phase temperature (#150 °C) within In this paper, we present analyte transport studies in the PB-HC-AES system.The influence of liquid flow rate, analyte the desolvation chamber. 808 Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12Sample Particle Collectors and Sampling 200, Buck Scientific, East Norwalk, CT, USA) was used with a 1+1 air–acetylene flame mixture. The alignment of the flame Two types of sample particle collectors were designed for these between the hollow cathode lamp and photomultiplier tube transport studies, shown in Fig. 1. The first was designed to was optimized for maximum absorbance during the aspiration act as an integrating collector of the analyte introduced in a of the sample solution. The amount of sample collected, and continuous-flow mode for a specified time period. This cylindrithus the eciency of the interface, was calculated based on the cally-shaped sample collector was machined from Teflon to absorbance of the sample solution as determined from calithe same diameter (3.25 mm) as the analytical HC.Analyte bration graphs prepared from a series of standard elemental particles from sample solutions were collected by positioning solutions in the range 0.1–10.0 ppm. this sample collector in the place of the HC. The analyte residues deposited in the cylindrical sample collector were then Scanning Electron Microscopy dissolved by placing the collector in 10 ml of a 1% HNO3 solution for 1 h, followed by sonication for 30 min.This Knowledge of analyte particle sizes and structure is very procedure ensures that all analyte sample deposited on the important in understanding the mechanism of analyte particle sample collector will be dissolved and transferred into solution. transport in this PB-HC-AES system. SEM oers sucient The solution was diluted to 25 ml in a calibrated flask for resolving capabilities, as well as a wide viewing angle and high measurement by atomic absorption spectrometry. The second magnification. A JEOL 848 (JEOL USA, Peabody, MA, USA) particle collector was designed for the purpose of subsequently scanning electron microscope was employed to observe the observing the particle residues with the scanning electron images of analyte particles collected in the GD area.Sample microscope. In this experiment, the flat sample collector, which preparation for SEM requires that the sample residues must is made of aluminium, is inserted in the middle of the HC be: (1) devoid of water and other solvents that could vaporize mount, perpendicular to the PB entrance direction for optimum in the vacuum system, (2) firmly mounted and (3) electrically analyte particle collection eciency.Because of the deleterious conductive. The flat aluminium sample collectors were loaded eects of residual moisture on the performance of the SEM within a sputter-deposition apparatus for coating with a thin instrument, the flat sample collectors were stored in a desic- Au film to reduce the eects of space charging by the noncator after sampling to reduce contamination.conductive particles. The SEM samples were then mounted firmly on the sample stage (stub) with metallic tape. The accelerating voltage and magnification were selected to be Solution Preparation and Delivery 15 kV and ×1000, respectively, for stable imaging. Type 55 The aqueous stock solutions of NaNO3, CuNO3, Fe(NO3)3 Kodak films and exposure times of 20–25 s were chosen for and HCl were prepared with de-ionized, distilled water from photographic recording.analytical-reagent grade inorganic salts and acid. The solvent delivery system was a Waters (Division of Millipore, Milford, RESULTS AND DISCUSSION MA, USA) Model 510 high-performance liquid chromatogra- Eects of Solution Flow Rate and Concentration phy pump. The solutions were passed through 1.53 mm stainless- steel tubing between the pump and the nebulizer, with the The understanding of flow rate as an important parameter in solution flow rate fixed at 1.5 ml min-1.For all of the experithis PB-HC-AES system not only includes its role in the ments performed here, the delivery system was operated in the nebulization process but also its eect on the particle transport continuous-flow mode. In this way, the amount of sample that through the interface. An increase in solution flow rate should ideally enters this PB-HC-AES source can be calculated with produce a corresponding, proportional increase in analytical consideration of the solution concentration, liquid flow rate, signal response.Previous studies11,12 demonstrated the existand the time of solution delivery. ence of an optimum sample liquid flow rate of 1.5–2.0 ml min-1 in this PB-HC-AES system for a range of analyte species. The general explanation for this phenomenon has been two-fold: Atomic Absorption Spectrometry (1) analyte mass transport to the GD source increases with In order to evaluate the transport characteristics of the PB increasing flow rate in the lower flow rate range interface, atomic absorption spectrometry was used to measure (<1.5 ml min-1) and (2) a continuous increase of the flow rate the actual amount of analyte reaching the GD region.The beyond 2.0 ml min-1 decreases the analyte signal because of analytical wavelengths employed were 324.7, 589.0 and degraded desolvation eciency caused by solvent overloading. 385.9 nm for Cu, Na and Fe measurements, respectively.The The eect of liquid flow rate on analyte transport, in terms of standard slotted burner supplied with the instrument (Buck the amount of Cu reaching the sample collector, is illustrated in Fig. 2. The appearance of the optimum flow rate at approximately 2 ml min-1 is consistent with the previous optical emission results,11,12 suggesting that the amount of analytical signal is an accurate reflection of the aerosol formation and eciency of the transport process.Certainly, the liquid flow rate has an eect on the nebulization processes, leading to droplets that have dierent desolvation eciencies. Harris20 characterized PB-LC–MS nebulizers with respect to the mean Sauter diameter (volume5surface area ratio) and the droplet size distributions. Those studies pointed to the need to generate the smallest diameter droplets while preventing aerosol/particle losses in the desolvation chamber and momentum separator due to turbulence, impaction and gravitational forces.The trends described here probably reflect these characteristics also, as will be discussed in a subsequent section. As noted previously, observations of non-linear behavior of Fig. 1 Sample particle collectors designed for insertion into the PB interfaces have been attributed to low desolvation hollow cathode volume for use in subsequent, 1, atomic absorption and 2, SEM analyses. eciencies, as well as turbulence, impaction and gravitational Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12 809size and shape of the particles. The particles delivered at the lower two concentrations [Fig. 4(a) and (b)] are fairly spherical in nature, with the largest particles seen in each increasing from #5 to 10 mm in diameter as the analyte concentration is increased from 10 to 50 ppm. The particles collected for the higher concentration solutions are dierent. First, the particles in Fig. 4(c) (100 ppm CuNO3) are less monodisperse than seen for the lower concentrations. Clearly, large, irregular-shaped particles are now being delivered in the presence of smaller spherical particles. The general shape seems to suggest the agglomeration of particles which could take place in either the gas phase or on the collector surface. In contrast to the analytical application of the PB-HC-AES source, the collector is not heated, so ‘wet’ particles could be envisioned to coalesce on the target surface.The departure from the ideal case of discrete particle introduction is even more dramatic in the micrograph of the ‘particles’ resulting from the 200 ppm solution [Fig. 4(d)]. In this case, the residue has the appearance Fig. 2 Eect of liquid flow rate on Cu transport expressed as the of a dried slurry. This definitively suggests a case where relative Cu concentration retrieved from the sample collector and complete desolvation has not taken place and ‘wet’ particles analyzed by atomic absorption spectrometry. are being delivered to the collector surface.The trends observed in these micrographs are consistent with conditions that would losses within the dierential pumping stages of the momentum lead to reduction in transport eciencies, most probably due separator. More specifically, these phenomena reflect that to impaction and gravitational losses. aerosol properties change with analyte concentration, basically shifting the aerosol size distribution.22 Even for 100% desolvation eciency, the dierences in initial droplet sizes will HCl Carrier Eect manifest themselves in shifting the analyte particle size distri- The basic concept of the PB is to accomplish analyte particle butions.Thus, at low analyte concentrations (small particle transport and enrichment in the detection stage via simple sizes) losses due to turbulence may play a factor, while at high physical and mechanical techniques.The comprehensive eects concentrations (large particles) impaction and gravitational of the processes of nebulization, desolvation and momentum losses would be important. These eects are analogous to separation are directly related to fine aerosol formation, solvent those seen in tertiary aerosol/particle size measurements in vaporization and analyte separation from the solvent. The nebulizer/desolvation systems used in atomic spectrometry.23,24 term ‘carrier eect’ is applied to phenomena that appear to In an attempt to evaluate the particle size/concentration eects, increase the analytical signal with the addition of certain a study was performed using the first type of sample collector physical and chemical modifiers. Bellar et al.17 reported the by continuously introducing CuNO3 solutions of dierent observation of a carrier eect associated with the LC mobile concentration. The atomic emission responses shown in Fig. 3 phase composition in the PB-LC–MS experiments.Their do not suggest suppression eects (non-linearity) at the low research explored the function of the carrier eect in the analyte concentrations, but do indeed reflect a declining transanalyte transport process and how it influenced the analytical port eciency at the higher Cu concentrations. Based on an performance of the mass spectrometer source. The addition of extrapolation of the response curve between 5 and 50 ppm chemical modifiers such as ammonium acetate and malic acid CuNO3, the relative transport eciency of the particles for to mobile phases has been shown to be an excellent approach the 100 ppm solution is depressed by approximately 30%.for improving sensitivity and linearity.24 Evidence supporting a particle size eect leading to the In our previous work,11 analyte emission signal enhancement observed non-linearity exhibited in Fig. 3 is provided through was observed by addition of HCl in the determination of Cs, scanning electron micrographs of the particles reaching the where the signal suppression caused by a high salt concen- HC volume.Fig. 4 presents the SEM images obtained for the tration matrix is in fact eliminated. In that work, it was the particles collected (30 min, 1.5 ml min-1 flow rate) for 10, 50, presence of the chloride ion which was deemed the important 100 and 200 ppm CuNO3 solutions [Fig. 4(a)–(d), respectfactor, as use of other acids (HNO3, H2SO4) did not yield ively].The micrographs reflect distinct changes in both the such enhancements. By the same token, the addition of other chloride salts did not prove to be eective at enhancing transport. The eect of carrier addition was studied by comparing the transmission eciencies of metal nitrate salts with and without HCl addition. In each case, known concentrations (10 ppm) of the metal ions were introduced under continuous- flow conditions (1.5 ml min-1) for a period of 30 min.The sample particles were collected on the Teflon sample collector (sample collector 1, Fig. 1) and analyzed by atomic absorption spectrometry. Table 1 presents the results of these transport studies in terms of the percentage of the total analyte input for each solution. The data point to a number of interesting trends. First, and probably most important in terms of analytical applications of the PB-HC-AES approach, is the fact that in every case the addition of HCl increases the analyte transport Fig. 3 Eect of Cu input concentration on Cu transport expressed eciency. The extent of the improvements ranges from #43% as the relative Cu concentration retrieved from the sample collector for the Fe3+ solution to 73% for Cu2+. On the other hand, and analyzed by atomic absorption spectrometry (1.5 ml min-1 liquid flow rate). there are also appreciable dierences in the transport 810 Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12(a) (c) (b) (d) Fig. 4 Scanning electron micrographs of Cu particles resulting from dierent CuNO3 input concentrations: (a) 10, (b) 50, (c) 100 and (d) 200 ppm (all micrographs taken at ×1000 magnification). Table 1 Multielement transport studies of the PB–HC-AES interface is often used as a drying agent in humid atmospheres. Based on the same argument of relative solubility, the improvements Without addition of HCl — seen on addition of the chloride are easily explained.In each Analyte Cu Fe Na case, the metal chlorides are less soluble than the corresponding Transport eciency (%)* 6.61 3.83 11.67 nitrates. Therefore, as the droplet containing both the NO3- With addition of HCl (155) — and Cl- counter ions begins to desolvate, the vastly lower Analyte Cu Fe Na solubility of the metal chlorides initiates their formation in Transport eciency (%)* 11.47 5.47 17.80 lieu of the nitrates. As with the metal nitrates, the observed transport eciencies are inversely related to the solubilities of * Transport eciency(%)=(Caa×F1)/(F2×Co×LFR×T )×100%, the metal chlorides.While the knowledge of using solution where Caa is AA measured concentration, mg ml-1; Co input analyte concentration, 10 mg ml-1; F1 volume dilution factor, 25; F2 volume chemistry to improve analyte transport can be used to analytdilution factor with 551 HCl, 5/6; LFR liquid flow rate, 1.5 ml min-1; ical advantage, the comparison between the dierent metals T time of sample input, 30 min. points to sources of discrepancies in the analytical performance (sensitivity) of dierent metal ions in solution.It is interesting to compare the obtained transport eciencies eciencies between each of the metal nitrates and the HClwith nebulization systems commonly employed in atomic added solutions. Let us first consider the dierences between spectrometry. While the eciencies depicted in Table 1 (nom- the respective metal nitrate solutions.As will be seen in the inally 4–18%) may not, at first glance, seem impressive, they next section, the most eective analyte transport will be are substantially higher than those typically quoted for pneu- achieved for those cases where completely dry particles of matic nebulizers operating in the #1 ml min-1 solution flow finite size are passed through the PB interface. Thus, desolvrate regime, #1%.26,27 On the other hand, the use of nebulizer ation eciency will be a key determining factor in transport systems which operate at very low solution uptake rates processes.To a first approximation, desolvation eciency will (1–100 ml min-1) has shown eciencies in the 50–100% be related to the solubility characteristics of the salt in question. range.28,29 The values reported for this thermoconcentric For the Na, Cu and Fe nitrates, there is a large variation in nebulizer-PB introduction system reflect optimization of nebul- the solubilities, ranging from 39 g ml-1 (at 100 °C) for NaNO3 ization and desolvation parameters and may be improved to infinite solubility for the Fe3+ salt.25 The relative solubilities further through optimization of the PB interface itself.For are inversely related to the observed transport eciencies. This example, no eort has been made to optimize the spacing of makes sense as the higher the solubility, the greater the propensity for the solute to ‘hold’ solvent. In fact, Fe(NO3)3 the skimmer cones, nor their aperture sizes.Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12 811SEM Images reflect some degree of hydration on the surface, which was followed by further dehydration. Second, there appears to be Because analyte transport in the PB interface is a mechanical much more material for the Cu salts. The relative inhomogenprocess, it is obvious that the size of the particles will have a eity of the distribution in Fig. 6(a) could be due to a process bearing on the overall eciency.Studies by Harris20 suggest whereby the primary particles may be scattered from the that the ideal particle size distribution is in the range 2–5 mm, central portion of the target and are re-deposited to form the so that particles can possess sucient momentum to cross observed pattern. As with the Na particles, a comparison over into the source region while at the same time promoting between the center and edge of the deposition [Fig. 6(b) and vaporization. Scanning electron micrographs are powerful (d), respectively] reveals that the central portion has a contools in studies involving detailed observation of such particles. densed appearance while the image taken from the edge of the For the sake of comparison with the previously presented deposit is composed of dispersed, discrete particles. Also constudies, Na and Cu were chosen as the analyte species in the firming the observations in Fig. 5, the micrographs for HCl SEM studies.Inspection of dierent areas of the sample addition [Fig. 6(c), central portion, and 6(e), edge] indicate collector provides information on both the particle sizes and the presence of larger particles, with some evidence that the their spatial distribution, giving insights into carrier eects. It larger particles are the result of agglomeration of smaller should be emphasized that the particle collection and imaging particles. The presence of HCl seems to induce aggregation of process employed here is not ideal as the samples are brought small sample particles to yield larger particles.Interestingly, into contact with ambient atmosphere on transfer between the the space charging observed for the sodium salts [Fig. 5(c) and GD source, the deposition chamber and the electron micro- (e)] is not observed in Fig. 6(c) and (e), possibly because the scope. Obviously, in situ observations of the sample surface copper-containing particles have better conductive character.would be preferable. Fig. 5(a) is a low magnification (×20) image of NaNO3 CONCLUSIONS particles on the sample collector. The image indicates that the sample particles basically concentrate in the center of the In the development of the PB-HC-AES system, it is important target area, with some dispersed particles surrounding the to understand the eect of the analyte transport phenomenon central region. The fact that the majority of the deposition on analytical performance.The high eciency solvent removal occurs in an approximately 1 mm diameter region reflects the of the PB interface eliminates interferences with regard to well collimated nature of the PB, as the spot size is smaller plasma processes and the observed optical emission spectra. than the entrance aperture through the side of the HC. Fig. 5(b) The absence of solution residues in the desolvation chamber (magnification=×1000) depicts a fairly homogeneous sample suggests that the sensitivity of the technique is controlled by particle distribution in the center of the collector; however, the the dynamics of the analyte transport across the PB interface.size of the sample particles is still not clear because of particle Studies of the overall transport eciencies reflect the cumulatcondensation. The cause of the condensation appearance could ive aspects of solvent loading, particle sizes, and the eects of either be due to hydration of the surface as it was brought scattering, impaction and gravitational forces on the resulting into ambient atmosphere, or by continuous impaction of the particles. Many of the sample introduction parameters that high momentum particles.Alternatively, the image may be aect the observed analyte emission intensities are shown to blurred due to space charging eects. The inability to discern be the direct result of analyte transport characteristics.accurately the particle sizes is overcome by examining the edge Inspection of collected analyte particles gives insights into the area around the concentrated particle accumulation where roles of analyte concentration and carrier eects on particles there is a sparse particle distribution as shown in Fig. 5(c) sizes and transport eciencies. (magnification=×1000). This image reveals discrete particles An optimum liquid flow rate in the range 1.5–2.0 ml min-1 in this area that are in the size range from 2 to 5 mm in can be seen in the curve of transport eciency aected by diameter.Deposits of the Na sample with HCl carrier addition liquid flow rate. Heavy solvent loading at high liquid flow were also examined through SEM images in both the center rates causes analyte transport eciency to decrease as the and edge areas and are presented in Fig. 5(d) and (e), respect- result of low desolvation eciency. The deviation from linearity ively. By comparing Fig. 5(b) with (d) and Fig. 5(c) with (e), of analyte transport eciency at high analyte input concenthe dierences caused by addition of HCl are readily observed. trations is observed and attributed to a large-particle cut-o. First, in the micrographs of the center area, Fig. 5(b) and (d), As expected, the eects of liquid flow rate and analyte input space charging eects of the incident electron beam appear as concentration on analyte transport eciency are the same as lines across Fig. 5(d), although both images show the same their eects on emission signal. This coincidence demonstrates condensed and homogeneous surface. The space charging the direct relationship between analyte transport eciency and suggests the presence of either more non-conductive material analytical signal. The function of HCl as an analyte carrier (i.e., more analyte) or a particle deposition that is more non- was evaluated by studying its eects on both analyte transport conductive than in the simple nitrate salts. It can be seen that and particle size changes.Without exception, experimental the addition of HCl in the sampling process seems to have a data show analyte transport enhancements for elements tendency to cause particles to be more closely packed on the with the addition of HCl, although this transport eciency sample surface. A comparison of the edge region micrographs, improvement is analyte-dependent. Fig. 5(c) and (e), reveals more details of larger particle sizes The analyte particle size and particle size distribution were produced by HCl addition.While there is still some evidence examined by SEM analysis at both the center and edge areas of charging eects, the presence of particles in the 10 mm size of the sample collectors. The micrographs of Na and Cu at range is clear. the edge area show clear, discrete particles, with sizes on the In order to avoid false assumptions based on scanning single micrometer level. However, condensed and clustered electron micrographs of Na samples, the same experiments particle aggregation is observed in the center areas of the were performed for Cu samples, with the results presented in collected Na and Cu samples. The dierence between the Fig. 6(a)–(e). Fig. 6(a) shows the Cu sample particles in the center and edge area micrographs indicates that the analyte center portion of the collector under low magnification (×20). particles travel in a very straight path with high momentum.Comparison of the low magnification images for the sodium Based on this observation, the HC, which is perpendicular to and copper nitrate salts [Fig. 5(a) and 6(a)] is interesting. First, the PB entrance, is being redesigned with a longer hollow cylinder for more ecient acceptance of the analyte particles the cracking of the residue of the Cu deposition appears to 812 Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12(a) (c) (e) (b) (d) Fig. 5 Scanning electron micrographs of Na particles: (a) low magnification (×20), (b) central portion of target area at high magnification (×1000), (c) edge area at high magnification (×1000), (d) central portion with HCl addition at high magnification (×1000) and (e) edge area with HCl addition at high magnification (×1000). and more ecient atomization/excitation. Scanning electron The observed transport eciencies are in the 4–18% range, which is higher than values typically quoted for other atomic micrographs also depict the particle size increasing with HCl addition for both Na and Cu samples.At this point, it is easy spectrometry systems employing pneumatic nebulization and desolvation chambers operating at similar solution uptake to see that the improvement of analyte transport eciency by HCl addition is due to an increase in analyte particle size rates. It is believed that this number may be improved as no eort has been made to date to optimize the spacing of the which overcomes the PB interface cut-o limitation.The formation of larger particles would seem to be related to the skimmers in the interface or in the sizes of the apertures. Future studies of this PB-HC-AES system will focus on its relative insolubility of the analyte metal chlorides versus nitrates and the like. Measurements of the primary aerosol application to organic compound analyses. For example, research using this system for amino acid analyses is currently droplets sizes would lend insights into whether the improvements are related more to nebulization characteristics or underway in an eort to determine molecular composition with atomic emission information.desolvation. Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12 813(a) (c) (e) (b) (d) Fig. 6 Scanning electron micrographs of Cu particles: (a) low magnification (×20), (b) central portion of target area at high magnification (×1000), (c) edge area at high magnification (×1000), (d) central portion with HCl addition at high magnification (×1000) and (e) edge area with HCl addition at high magnification (×1000). 2 Marcus, R. K., Harville, T. R., Mei, Y., and Shick, C. R., Anal. The authors acknowledge the financial support from the Chem., 1994, 66, 902A. National Science Foundation under grant No. CHE-9420751 3 Daughtrey, E. H., Jr., Donohue, D. L., Slevin, P. J., and Harrison, and the technical support from the Electron Microscopy W.W., Anal. Chem., 1975, 47, 683. Laboratory of Clemson University for use of the SEM facilities. 4 Harrison, W. W., and Prakash, N. J., Anal. Chim. Acta, 1970, Donation of the Thermabeam nebulizer and particle beam 49, 151. 5 Bracket, J. M., and Vickers, T., Spectrochim. Acta, Part B, 1982, interface by Extrel Corporation is greatly appreciated. 37, 841. 6 Chen, F., and Williams, J. C., Anal. Chem., 1990, 62, 489. 7 Barshick, C. M., Duckworth, D. C., and Smith, D. H., J. Am. Soc. REFERENCES Mass Spectrom., 1993, 4, 47. 1 Broekaert, J. A. C., Bricker, T., Brushwyler, K. R., and Hieftje, 8 Bracket, J. M., and Vickers, T., Spectrochim. Acta, Part B, 1983, 38, 979. G. M., Spectrochim. Acta, Part B, 1992, 47, 131. 814 Journal of Analytical Atomic Spectrometry, August 1997, Vol. 129 Strange, C. M., and Marcus, R. K., Spectrochim. Acta, Part B, 22 Browner, R. F., Canals, A., and Hernandis, V., Spectrochim. Acta, Part B, 1992, 47, 659. 1991, 46, 517. 10 You, J., Fanning, J. C., and Marcus, R. K., Anal. Chem., 1994, 23 Browner, R. F., and Cressner, M. S., Spectrochim. Acta, Part B, 1980, 35, 73. 66, 3916. 11 You, J., DePalma, P. A., Jr., and Marcus, R. K., J. Anal. At. 24 Browner, R. F., Boorn, A. W., and Smith, D. D., Anal. Chem., 1982, 54, 1411. Spectrom., 1996, 11, 483. 12 Olson, L. K., Belkin, M., and Caruso, J. A., J. Anal. At. Spectrom., 25 CRC Handbook of Chemistry and Physics, ed., Weast, R. C., Astle, M. J., and Beyer, W. H., CRC Press, Boca Raton, FL, 69th 1996, 11, 491. 13 Hausler, D. W., and Taylor, L. T., Anal. Chem., 1981, 53, 1227. edn., 1988. 26 Browner, R. F., and Boorn, A. W., Anal. Chem., 1984, 56, 786A. 14 Morgan, C. A., Smith, B. W., and Winefordner, J. D., J. Anal. At. Spectrom., 1993, 8, 539. 27 Olesik, J. W., and Bates, L. C., Spectrochim. Acta, Part B, 1995, 50, 285. 15 Todorovic, W., Vidovic, S., and Ilic, Z., J. Anal. At. Spectrom., 1993, 8, 1113. 28 Olesik, J. W., Kinzer, J. A., and Harkelroad, B., Anal. Chem., 1994, 66, 2022. 16 Willoughby, R. C., and Browner, R. F., Anal. Chem., 1984, 56, 2626. 17 Bellar, T. A., Behymer, T. D., and Budde, W. L., J. Am. Soc. Mass 29 Liu, H., Cliord, R. H., Dolan, S. P., and Montaser, A., Spectrochim. Acta, Part B, 1996, 51, 27. Spectrom., 1990, 1, 92. 18 Apel, A., and Perry, M. L., J. Chromatogr., 1994, 554, 103. 19 Ho, J. S., Behymer, T. D., Budde, W. L., and Bellar, T. A., J. Am. Soc. Mass Spectrom., 1992, 3, 662. 20 Harris, W. E., Dissertation, Georgia Institute of Technology, 1991. Paper 7/02703F 21 Kim, I. S., Sasinos, F. I., Stephens, R. D., and Brown, M. A., J. Agric. Food Chem., 1990, 38, 1223. Received April 21, 1997 Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12 815

ISSN:0267-9477    

DOI:10.1039/a702703f

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Microwave-induced Plasma Boosted Microsecond-pulse Glow Discharge Optical Emission Spectrometry YONGXUAN SU, PENGYUAN YANG*, DENGYUN CHEN, ZHIGANG ZHANG, ZHEN ZHOU, XIAORU WANG AND BENLI HUANG Department of Chemistry, L aboratory of Analytical Science, Xiamen University, Xiamen 361005, China A microsecond-pulse (ms-pulse) glow discharge (GD) source When in combination with a GD source, free sample atoms produced by cathode-sputtering in the GD enter the MIP by boosted by a microwave-induced plasma (MIP) has been developed and studied for optical emission spectrometry diusion and convection.However, most previous studies on MIP boosted GDs operated under the dc GD mode. (OES). The excitation processes of the tandem GD source were investigated. The analytical characteristics of the As a new member type of GD, ms-pulse GD has recently been paid more attention in analytical areas, including OES, GD-OES source in the presence and absence of the MIP were compared, including the operating parameters, signal-to- AAS, AFS and MS techniques.6–9 Experimental results have shown that the ms-pulse GD mode has an analytical perform- background ratios (S/B) and relative standard deviation (RSD).The results show that under a relatively low discharge ance that is much better than the dc GD mode.7,8 The pulsing technique is also a useful diagnostic tool for the excitation and pressure (<180 Pa), the ms-pulse GD can couple fairly well with the MIP and emit intense analytical lines.When the GD ionization processes in a GD plasma.6,10–12 In order to evaluate the possibility of using ms-pulse GD as an ion source for solid source is operated under a pressure higher than 200 Pa, two emission peaks appear, independent in time, for a given sample analysis as well as depth profile analysis on a solid surface, experiments on ms-pulse GD time of flight mass resonance atomic line, because sample atoms are independently structurally excited, first by the ms-pulse GD spectrometry (TOFMS), with some interesting results have been carried out.8,9 One of the most significant characteristics and then by the MIP.The time interval between the two peaks for Zn I 213.8 nm is longer than that for Cu I 324.7 nm, which of ms-pulse GD is that the sputtering rate of ms-pulse GD is more than two orders of magnitude higher than in dc GD is believed to be due to the faster diusing velocity of copper atoms. When the ms-pulse GD lamp is operated under a gas during the pulse-on regime, and that the average discharge power is relatively low.9 Therefore, the number of sample pressure higher than 220 Pa, the ion population is so high that Cu II ionic line at 224.7 nm ‘becomes’ two peaks because of a atoms produced in the very short pulse time is significantly high.possible self-absorption. The results show that the supplementary use of an MIP can eliminate the self-absorption In the present work, an MIP plume is superimposed on the sputtered sample atom–ion cloud in order to obtain atomic of ionic and atomic lines.When the ms-pulse GD source is coupled with the MIP, S/Bs are improved by a factor of more and ionic lines of high intensity. Based on the work of ms-pulse GD-TOFMS, an MIP boosted ms-pulse GD source has been than one order of magnitude for several analytical lines. A short-term RSD of 0.2% is achieved for the ‘ms-pulse designed in which the MIP is sustained in a tunable Beenakker TM010 resonator.The excitation processes of the OES source GD+MIP’ configuration compared with that of 1.0% for ‘mspulse GD only’ mode. The experimental results show that the were studied. The eects of the operating parameters on ms-pulse GD in the presence and absence of the MIP were MIP boosted ms-pulse GD is a promising technique for solid sample and surface analysis. investigated. A comparison of analytical characteristics for the ms-pulse GD alone and the ms-pulse GD-MIP tandem source Keywords: Microsecond-pulse glow discharge; microwaveis presented, in terms of the signal-to-background (S/B) ratios induced plasma; optical emission spectrometry; tandem source; and relative standard deviation (RSD).excitation process Glow discharge (GD) has been widely used in optical emission EXPERIMENTAL spectrometry (OES), atomic absorption spectrometry (AAS) MIP Boosted ms-Pulse GD-OES Source and atomic fluorescence spectrometry (AFS) for analytical applications, and also has served extensively as an ion source A schematic diagram of the MIP boosted GD source is shown in Fig. 1. The structure of the GD source is similar to the one for mass spectrometry (MS).1 Various techniques have been used to obtain high intensity without losing the advantage of used as an ion source for TOFMS and has been described in detail previously.9 Both the GD source and the resonator are narrow spectral lines. Leis and Steers have reviewed articles on boosted GD sources.2 The use of an MIP to enhance the made of brass.The discharge gap in GD lamp is designed to be relatively thin, such that the distance between the microwave performance of a GD has attracted increased interest since the 1980s.3–5 The advantage of a microwave-boosted GD lamp is cavity and the sample cathode is only 7 mm. Water cooling is used directly to cool the sample cathode. The anode body and that the atomization of a solid sample is performed by the process of cathode-sputtering in the GD while highly ecient the microwave cavity are earthed.A quartz tube with an id of 8 mm, separates the inner low pressure region from the outer excitation of analytical lines can be carried out in an MIP at low pressure. atmospheric pressure. Two tuning screws are used to adjust the minimum reflected power. A GD source with microwave boosting has been presented by Leis et al.3 They combined a conventional Grimm-type GD The lamp is evacuated by a two-stage mechanical pump (4 dm3 s-1, Shanghai Vacuum Plant, Shanghai, China) and source with an MIP sustained in a tunable Beenakker TM010 resonator.It is known that an MIP shows good excitation the working carrier gas is pumped continuously through the lamp. The working gas pressure is regulated by an inlet needle properties but is not particularly suitable for atomization. Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12 (817–822) 817Comparison of the analytical characteristics of the GD source in the presence and absence of theMIP An ICP-AES instrument (sequential ICP-AES, Model 2070, Baird, Bedford, MA, USA) was utilized, with a built-in 1 m Czerny–Turner monochromator with a grating of 3600 grooves mm-1 and a blaze wavelength of 400 nm. The ICP torch was removed and replaced by the GD–MIP tandem source. The slit-widths of both the entrance and the exit are 17 mm. Signals are integrated for 0.2 s, and then digitized with a OS-2 microcomputer.The signal processing part of this ICPAES instrument is designed for continuous emission rather than signal pulses, such as a microsecond wide signal obtained in the present study. Thus, the signal pulse has to be averaged against a background signal of about 500 ms for a pulse rate of 2 kHz that is used in this study. Therefore, the S/B obtained in this experiment can only be used as reference values for a Fig. 1 Structure of the MIP boosted ms-pulse GD source.The solid comparison of some of the results obtained for a ms-pulse GD sample is sputtered and atomized in the GD, and the excitation occurs in the MIP tandem ms-pulse GD discharge. in the absence and presence of the MIP. Experimental Considerations valve and monitored with a thermocouple gauge (ZDO-54, The lamp can be operated in two optical configurations: ‘ms- Chengdu Instrument Factory, Chengdu, China). pulse GD only’ and ‘ms-pulse GD+MIP’. The current was The ms-pulse GD power supply is laboratory built with regulated for ms-pulse GD and argon was used as the carrier the following adjustable parameters: discharge frequency gas.A dc voltage (-750 V, with a current of several tens of (10–5000 Hz); discharge duration (0.2–5 ms); and discharge mA) was utilized to pre-burn the sample cathode. After several current (0.03 mA–3.1 A). The voltage output of the ms-pulse days operation, the reflected microwave power could increase GD power supply is kept at -750 V.owing to the redeposition of sputtered material on the inner The microwave generator produces a maximum power of surface of the quartz tube. In the case of brass samples, the 200W at a frequency of 2450 MHz (Hai Guang Instruments, quartz tube should be cleaned frequently because of the high Beijing, China). It is connected to the resonator via a 2m sputtering rate and the thin metallic film hampers the ecient length of coaxial cable with 50 V impedance.Two meters coupling of the microwave energy. indicate the forward and reflected microwave powers. RESULTS AND DISCUSSION Study of the Excitation Processes of the GD-OES Source Experimental Set-up Superimposition of anMIP on a GD Investigation of the excitation processes in the GD source When sucient microwave power (50W forward power in the A block diagram of the experimental set-up used for investigapresent experiment) is superimposed on a running GD plasma, ting of the excitation processes is shown in Fig. 2. The radiation a microwave discharge in the quartz tube can be ignited, and source is imaged by a quartz lens onto the entrance slit of a forms a plasma extention to the positive column of the GD. 0.5 m Czerny–Turner monochromator (Acton Research, Acton, No boundary between the microwave discharge and the MA, USA), with 1800 grooves mm-1 and a blaze wavelength ms-pulse GD can be observed visually. A minimum reflected of 500 nm.Both the entrance and exit slit-widths are 20 mm. power of about 5 W can be obtained by adjusting the two Signals from the preamplifier were observed with a 40 MHz tuning screws. Because charged particles are also produced by oscilloscope (Model BS-5504, Aron, Korea), they can also be the MIP, the transient pulse voltage drop between the anode integrated with a boxcar integrator (laboratory made) and and cathode decreases significantly, and the radiation of the then recorded with a recorder (Model 9176, Varian, Palo Alto, lamp is obviously enhanced when the microwave generator is CA, USA) or a compatible PC-386 computer with a 20 MPS switched on.analogue to digital converter. Eect of Discharge Parameters on Excitation Processes Dependence of discharge pressure on the coupling of an MIP with ms-pulse GD The argon pressure can clearly aect the coupling of an MIP with ms-pulse GD. As shown in Fig. 3, the sample atoms emit a strong radiation peak when the MIP boosted ms-pulse GD lamp is operated at a discharge pressure lower than 180 Pa.However, if this GD lamp is operated at a pressure higher than 200 Pa, two emission peaks appear independently in time for a given resonance atomic line (see Fig. 3). Depending on the increase in argon pressure, the MIP plume tends to shrink away from the GD region. The sample atoms are excited first by ms-pulse GD and then by the MIP when atoms move from Fig. 2 Schematic diagram of the MIP boosted ms-pulse GD-OES the GD region to the MIP region.As shown in Fig. 3(a), the system. A commercial ICP-AES is utilized to replace the monochroma- first peaks (peak I) correspond to the emission excited by tor and the signal processing system, when the analytical characteristics of this tandem GD source are evaluated. See text for details. ms-pulse GD and the second peaks (peak II) by the MIP. The 818 Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12Fig. 4 Eect of discharge pressure on the time interval between peak I and peak II for dierent resonance atomic lines: &, Zn I 213.8 nm; and %, Cu I 324.7 nm. surface of the sample cathode under the operating conditions used. Eect of discharge frequency and pulse duration on emission intensities The eect of discharge frequency on the intensities of peaks I and II for Zn I 213.8 nm and Cu I 324.7 nm are illustrated in Fig. 5. It can be seen that the peak height for peak I reduces Fig. 3 Eect of discharge pressure on the time-resolved resonance with an increase in discharge frequency. If the discharge atomic lines for a brass sample: (a) Cu I 324.7 nm; (b) Zn I 213.8 nm. frequency is operated between 800 and 1200 Hz, peak II can reach a maximum value. Above this frequency range, the sputtering rate increases in accordance with discharge pressure, higher the frequency, the lower the peak height will be. It and the high pressure may also tend to slow the loss of could be supposed that a very high discharge frequency would sputtered atoms owing to diusion, and as a result peak I eliminate the sputtered and excited processes because the increases.However, because the gap between the sample and laboratory made power supply is unable to recover for the the MIP plume lengthens, the density of the atom cloud next discharge pulse owing to its limited power capacity.9 This decreases when the atom cloud passes through the MIP excited finite capacity of the discharge power could also limit operation region.For ionic lines with high energy transition radiation, of the GD pulse to a relatively wide pulse duration. Walden only the first peak appears with the superimposition of the et al. have found such results in their experiments on ms-pulse MIP. In general, the radiation emitted by the microwave GD-OES.7 The present work also indicated that intensities of boosted lamp can be characterized by those spectral lines with both peak I and peak II increased according to the duration low and mean excitation energies.of the discharge pulse, although not obviously (the maximum The time duration for peak I at its maximum value is 15 ms pulse duration can be adjusted only to within 5 ms in the pulse after the discharge pulse starts and is almost the same for both generator). Zn I 213.8 nm and Cu I 324.7 nm. This time value hardly varies with the change in discharge pressure because the shift of the most excited region in ms-pulse GD is insignificant.Eect of discharge current on emission intensities However, the duration of peak II at its peak value changes The relationship between the transient current and the radi- according to the discharge pressure, because the MIP plume ation intensities of peak I and peak II is given in Fig. 6. Visual shrinks further away from the GD excited area with an increase observation through the quartz window showed that the glow in argon pressure.The relationship between time interval of the two emission peaks and discharge pressure is illustrated in Fig. 4. It can be seen that the higher the discharge pressure, the longer the time interval, and that the time interval between the two peaks for Zn I 213.856 nm is longer than that of Cu I 324.7 nm under identical operating parameters, which is believed to be due to the faster diusing velocity of the copper atoms. When the discharge pressure is 300 Pa, the times for peak II at the peak values are about 120 and 140 ms for Cu I 324.7 nm and Zn I 213.8 nm, respectively.The distance between the sample and the central part of the quartz tube is about 18 mm (the thickness of the quartz restrictor ring is 3 mm). Therefore, the average diusing velocities of copper and zinc atoms can be estimated to be 150 and 129 m s-1, respectively, and the most reactive region for excitation in this ms-pulse GD is 1.94–2.25 mm (estimated from the diusing velocity multiplied by the time interval between the falling edge of the discharge Fig. 5 Dependence of emission intensities of peak I and peak II on discharge frequency: &, Zn I 213.8 nm; and %, Cu I 324.7 nm. pulse and peak I at its maximum value) from the sputtering Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12 819tation mechanism. Such an experiment is currently in progress in this laboratory. Eects of MIP on Sputtering Rate and Sputtered Surface The addition of an MIP has an obvious eect on sputtering rate as well as on the topography of the sputtered surface.The sputtering rate of ms-pulse GD has been studied previously9. The transient sputtering rate (9.5 mg s-1 mm-2) for ‘ms-pulse GD+MIP’ is only about half of that of ‘ms-pulse GD only’ (21 mg s-1 mm-2) under the same operating conditions for a brass sample, with a discharge frequency of 1.8 kHz, pulse duration of 3 ms, current of 2 A and pressure of 180 Pa.The crater on the surface of the erosion area also becomes finer, when observed with a scanning electron microscope (Hitachi Fig. 6 Eect of discharge current on the intensities of peak I and S-300, Tokyo, Japan). The decrease in sputtering rate and the peak II: &, Zn I 213.8 nm; and %, Cu I 324.7 nm. improvement in crater shape are believed to be due to the lower voltage drop between the anode and cathode.14 For turns white and then green with an increase in discharge more details of the results of the sputtering rate and on the current, indicating very intensive sputtering and excitation sputtered surface for ms-pulse GD, see ref. 9. processes. Peak I increases sharply correspondingly with discharge current and reaches its peak when the transient current Comparison of the Analytical Characteristics of the OES is about 0.8 A, then the emission intensity tends to decrease. Source in the Presence and Absence of the MIP In this experiment, it was observed that when the discharge current increases, the line profile of peak I widens with hardly Dependence of emission intensities on discharge parameters any increase in peak intensity.It could be that a self-absorption As described above, an MIP can couple well with ms-pulse occurs, which is a result of the high density of ‘cool’ atoms GD if the discharge pressure is maintained at less than 200 surrounding the ‘hot’ atoms in the center of the plasma.Pa. In order to evaluate the analytical characteristics of the However, peak II increases steadily and tends to saturation boosted GD source, the argon pressure in the discharge only when the current is higher than 2 A. The high current chamber was kept relatively low throughout the following can result in an increase in the sputtering rate, and more experiments. sample atoms can be excited by the MIP. The self-absorption in the MIP excited region is not severe because of a relatively Eect of discharge frequency on spectral intensity.As shown in low density of ‘cool’ atoms. Fig. 7, emission intensities of the integrated signals increase steadily in accordance with the increase in discharge frequency without an MIP, and increase sharply (especially for Cu I Eect of microwave power on emission intensities The forward microwave power was optimized at 50Wthroughout the experiment. The microwave power has a minor eect on emission peaks when the power is varied between 50 and 80 W.When the forward power is high, the reflected power will increase accordingly, and is rather dicult to tune unless the discharge gas pressure is lowered for improved coupling. However, a low discharge pressure can eliminate sputtering. The experimental results are also supported by previous reports from Li et al.5 They found that the MIP and dc GD can be optimized to an ideal coupling by adjusting the relationship between discharge voltage, microwave power and discharge pressure under dierent discharge conditions.In their experiments, the maximum S/B levels could also be achieved with a microwave power of 50 W, if the GD lamp is operated under a low discharge pressure.5 Eect of MIP on the Excitation of Ionic Lines The excitation and ionization processes are very intensive when the ms-pulse GD lamp works under a high transient current and relatively high pressure (>220 Pa). Owing to its high ion populations, the ionic line of Cu II 224.7 nm appears to become two peaks because of a possible self-absorption.Supplementing with the MIP aects the excitation of sample ions and results in the alleviation of this self-absorption of Cu II 224.7 nm. Steers and Leis have reviewed many examples of charge exchange (CE) excitation between sample atoms and argon ions in a GD source,13 which has been believed to be a very important process for the excitation of ionic lines in a Fig. 7 Relationship between discharge frequency and line intensities GD with inert gases.The change in the ion density of argon of sputtered atoms, (a) in the absence of the MIP and (b) in the can be observed by a mass analyser coupled with an MIP presence of the MIP: ×+, Cu I 324.7 nm, see right y-scale; &, Cu II 224.7 nm; +, Zn I 213.8 nm; and %, Zn II 202.5 nm. boosted ms-pulse GD for a better understanding of the exci- 820 Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12324.7 nm) when the microwave generator is switched on. Owing to the relatively low discharge pressure, self-absorption of emitted radiation by a cloud of ‘cool’ atoms in the outer region of the discharge is fairly serious when the discharge frequency is increased to as high as 2 kHz. The emission intensities decrease when the discharge frequency is adjusted to higher than 2 kHz. Two possible reasons account for the results: one is the possible self-absorption of radiation owing to an increase in the surrounding cold atoms, and the other results from the limited power capacity of the laboratory made ms-pulse power supply, as discussed previously.9 Eect of discharge pressure on spectral intensity.The eect of discharge gas pressure on the line intensity is shown in Fig. 8. The working carrier gas pressure is an important factor for it has a clear eect on the GD. As shown in Fig. 8(a), with the ‘ms-pulse GD only’ mode, the intensity of Cu I 324.7 nm rises rapidly in accordance with the gas pressure.The relationship between the argon gas pressure and the line intensities for Cu II 224.7 nm and Zn II 202.5 nm are similar in the two discharge configurations. However, self-absorption of spectral line Cu I 324.7 nm occurs in the ms-pulse GD+MIP mode when the argon pressure is higher than 180 Pa. As a result, the emission intensity decreases. It is interesting to note that the intensity of the Zn I 213.8 nm line is about four orders of magnitude lower than the intensity of the Cu I 324.7 nm line when the ‘ms-pulse GD only’ mode is utilized, while it is eight times lower when the ‘ms-pulse GD+MIP’ configuration is used, because the line transition for Zn I 213.8 nm originates from Fig. 9 Relationship between discharge current and signal intensities, a higher energy state. Similar results can be seen in Figs. 7 (a) in the absence of the MIP (b) in the presence of the MIP: +×, Cu I 324.7 nm, see right y-scale; &, Cu II 224.7 nm; +, Zn I 213.8 nm and and 9.%, Zn II 202.5 nm. Eect of discharge current on spectral intensity. The dependence of emission intensities on discharge current is depicted in sputtering rate. Coupling with the MIP results in a decrease Fig. 9. As shown in Fig. 9(a), a possible self-absorption of in the sputtering rate and particularly enhances the excitation selected spectral lines occurs when the discharge current is processes of the low energy lines.As a result, the self-absorption higher than 1.5 A, which is believed to be due to the high of Cu I 324.7 nm and Zn I 213.8 nm decrease because of the smaller number of ‘cool’ atoms and the intensities of Cu II 224.7 nm and Zn II 202.5 nm become lower with the increase in discharge current. Signal-to-background Ratios and the Relative Standard Deviation Signal-to-background ratios As shown in Table 1, the S/B are enhanced for the selected lines when the microwave generator is coupled to the ms-pulse GD.As the generator is turned on, the emission intensities increase sharply but the background level rises only slightly. Therefore, detection power can be improved by using a ms-pulse GD–MIP source. Leis et al. have also shown that detection powers can be significantly increased by supplementing a dc GD with an MIP discharge.3,14 Shown in Fig. 10 is the spectral profile for Cu I 324.7 nm (brass sample, 69.4% Cu), obtained with both the ‘ms-pulse GD only mode’ and ‘ms-pulse GD+MIP’ mode.It was found that the profile is still very narrow in the presence of the MIP and the S/B is fairly good with a relatively strong line intensity. Relative standard deviation The short-term stability of the lamp operated under ‘ms-pulse GD only’ and ‘ms-pulse GD+MIP’ modes was studied by measuring the intensity of the Mn I 257.610 nm line for a low alloy steel sample containing 0.122% Mn. About 2 min after ignition, the intensity of Mn I 257.610 nm becomes constant Fig. 8 Eect of discharge pressure on signal intensities, (a) in the and then the RSD measurements were carried out. The RSDs absence of the MIP and (b) in the presence of the MIP: +×, Cu I of the ‘ms-pulse GD only’ and ‘ms-pulse GD+MIP’ configur- 324.7 nm, see right y-scale; &, Cu II 224.7 nm; +, Zn I 213.8 nm; and %, Zn II 202.5 nm. ation were 1.0 and 0.2%, respectively, obtained from line Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12 821Table 1 Comparison of the S/Bs of the GD source in the absence and presence of the MIP Spectral line/nm Mode Parameter Cu I 324.7 Cu II 224.7 Zn I 213.8 Zn II 202.5 ms-pulse GD only Signal (count) 113071 20987 27908 4479 Background (count) 1910 191 185 176 S/B ratio 59 110 151 25 ms-pulse GD+MIP Signal (count) 8096826 96030 1430311 96821 Background (count) 4212 331 752 234 S/B ratio 1922 290 1902 414 operation of the lamp under ‘ms-pulse GD only’ mode, the S/B ratios of ‘ms-pulse GD+MIP’ mode have been enhanced several to several tens of times for selected lines.The RSD of the level of emission intensity for the GD lamp has also been improved from 1.0 to 0.2% in the presence of MIP boosting. Further work will focus on the determination of detection limits for the boosted source and on investigating the excitation processes using TOFMS. The application of the technique to surface analysis of solid samples is also an attractive area. The Laboratory of Analytical Science is run by the State Education Commission of China (SEDC). This work was supported by the National Nature Science Foundation of China under grant number CHEM-29235110-II, and partially by the Outstanding Youth Fellowship of SEDC.The authors thank Baird for their generous donation of the ICP-AES Fig. 10 Spectral profile of Cu I 324.7 nm: discharge frequency 1.7 kHz; instruments to the laboratory and also thank Liang Feng for pulse width of 2 ms; current 1.5 A; argon pressure 180 Pa; and microwave forward power 50 W.Signals were integrated for 0.2 s, helpful discussion in designing the microwave resonator. recorded with a 3600 groves mm-1 high-resolution monochromator. A, In the absence of the MIP, see left y-scale; and B, in the presence of the MIP, see right y-scale. REFERENCES 1 Broekaert, J. A. C., Appl. Spectrosc., 1995, 49, 12A. intensities (ten measurements) with a 0.2 integration time for 2 Leis, F., and Steers, E.B. M., Spectrochim. Acta, Part B, 1994, a 3 min interval. The application of microwaves could lead to 49, 289. an improvement in the background equivalent concentration, 3 Leis, F., Broekaert, J. A. C., and Laqua, K., Spectrochim. Acta, Part B, 1987, 42, 1169. because the RSD and S/B of the analytical emission lines are 4 Outred, M., Ru�mmeli, M. H., and Steers, E. B. M., J. Anal. At. enhanced in the MIP boosted ms-pulse GD lamp. Spectrom., 1994, 9, 381. 5 Li, Y. M., Du, Z.H., Duan, Y. X., Zhang, H. Q., Jin, Q. H., and Liu, R. S., Chem. J. Chin. Univ., 1996, 17, 215. CONCLUSION 6 Hang, W., Walden, W. O., and Harrison, W. W., Anal. Chem., Results of the study on an MIP boosted ms-pulse GD-OES 1996, 68, 1148. 7 Walden, W. O., Hang, W., Smith, B. W., Winefordner, J. D., and instrument show that an MIP can couple fairly well with Harrison, W. W., Fresenius’ J. Anal. Chem., 1996, 354, 442. ms-pulse GD when the discharge pressure is less than 200 Pa, 8 Hang, W., Yang, P. Y., Wang, X. R., Yang, C. L., Su, Y. X., and and that this hyphenated discharge can produce strong emis- Huang, B. L., Rapid Commun. Mass Spectrom., 1994, 8, 590. sion intensities from sample atoms. However, under a relatively 9 Su, Y. X., Zhou, Z., Yang, P. Y., Wang, X. R., and Huang, B. L., high discharge pressure (>240 Pa), the ms-pulse GD and MIP Spectrochim. Acta, Part B, 1997, 52, 633. tend to excite the sputtered atoms separately and produce two 10 Klingler, J. A., Savickas, P. J., and Harrision, W. W., J. Am. Soc. Mass Spectrom., 1990, 1, 138. emission peaks, independent in time, for a given resonance 11 Klingler, J. A., Barshick, C. M., and Harrision, W. W., Anal. atomic line. Chem., 1991, 63, 2571. The eects of the discharge parameters on the emission 12 Pan, C., and King, F. L., Anal. Chem., 1993, 65, 3187. intensities of the analytical lines of interest have been studied 13 Steers, E. B. M., and Leis, F., Spectrochim. Acta, Part B, 1991, and optimized as 1.8 kHz for discharge frequency, 200 Pa for 46, 527. argon pressure and 1.5 A for discharge current. Spectral self- 14 Leis, F., and Broekaert, J. A. C., Spectrochim. Acta, Part B, 1991, 46, 243. absorption appears to exist in the GD and can be reduced when the microwave generator is switched on, especially for Paper 7/00913E low energy emission lines. Received February 10, 1997 The net intensities of the spectral radiation increase sharply Accepted April 11, 1997 when the lamp is coupled with the MIP. Compared with 822 Journal of Analytical Atomic Spectrometry, August 1997, Vol.

ISSN:0267-9477    

DOI:10.1039/a700913e

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Laser Ablation-assisted Radiofrequency Atomization Excitation Source for Direct Determinations of Elements in Ceramics TAKAHIRO OGURIa , HIROFUMI INOUEa , SHIN TSUGEa , KUNIYUKI KITAGAWA*b AND NORIO ARAIb aDepartment of Applied Chemistry, School of Engineering, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464–01, Japan bResearch Center for Advanced Energy Conversion, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464–01, Japan A radiofrequency He discharge atomization/excitation source the powder ceramics sample is loaded.When an rf power of 300–500 W at 13.56 MHz (5) is applied to the central electrode which was assisted by laser ablation was developed for the direct elemental analysis of ceramics by atomic emission (11), a stable He discharge plasma is formed under reduced pressure around the electrically conductive Ta hollow lid. spectrometry. Eects of laser irradiation, such as sensitivity enhancement and alleviation of matrix eects, were found in Tantalum is also preferable for attaining higher temperatures because of lower radiative heat loss due to its smaller emissivity the direct atomization/excitation of ceramics.Calibration using powder reference materials became feasible for sintered (e#0.4) than that of graphite (e#1).2 For sintered ceramic samples, a dierent type of sample holder (6) was used in ceramic samples. order to attain a stable discharge, to prevent the sample from Keywords: Direct atomization/excitation; solid sampling; melting with the Ta hollow lid and to fix the sample location.ceramics; radiofrequency discharge; helium plasma; atomic The pulsed laser radiation of about 124 mJ per 10 ns passes emission spectrometry; laser ablation through the Ta hollow lid and successively hits the samples on the graphite disk, at a rate of 10 pulses per second, resulting Direct elemental analysis is attractive in that laborious chemi- in enhancement of the atomization eciency.The irradiation cal pre-treatments and analyte contamination from the position on the sample surface was adjusted by the focusing reagents and the environment can be minimized. However, lens (8). The atomic emission is detected by the spectrometer sintered ceramics are among the most dicult matrices for (9) and the analyte signal from the photomultiplier tube is direct atomization/excitation because of their thermally high acquired by a microcomputer with an A/D interface.The stability. computer was operated with a program written in C-language, In previous work,1,2 a radiofrequency atomization/excitation to trigger the laser Q-switch and to synchronize the boxcar source using a He discharge with a hot electrode, which also integration of the signals. The boxcar window was 13 ms, serves as the atomizer under reduced pressure, was developed during which atomic emission appeared and decayed. for direct elemental analysis and its basic characteristics were The experimental components employed in this work were studied. The rf He plasma is of interest for attaining the basically the same as those used in the previous work,1,2 except eective excitation of elements with high energy levels, such as for the Nd5YAG laser (NEC SL1200).The experimental halogens. The source has been successfully applied to direct conditions are given in Table 1. simultaneous determinations of elements including chlorine in biological samples.However, its application to ceramic samples Sample and Procedures was found to be dicult because the thermal atomization of the sample was insucient even at a maximum temperature Samples of 25 mg of powdered (>500 mesh) silicon nitride, of approximately 2200 °C, attainable with the hot electrode. alumina and boron carbide were dispersed in 1 ml of 0.1 M On the other hand, the laser ablation technique has been nitric acid by a mechanical vibrator. A 5 ml volume of the applied as a promising technique for sample introduction in resulting slurry or several milligrams of sintered silicon nitride inductively coupled plasma mass spectrometry (ICP-MS) for were loaded in the cavity of the graphite bottom disk.The the direct analysis of ceramics.3–5 In this work, the laser sample was then covered with the hollow lid and the sample ablation technique was applied to assist the rf He discharge cup was fitted on the top of the central electrode (11).The air source in the atomization process in order to develop a simple in the discharge chamber (12) was pumped out for 5 min to a tool for direct determinations of elements in sintered ceramics pressure lower than 0.1 mmHg. During this procedure, the by atomic emission spectrometry. The eects of ablation by nitric acid solution in the slurry was evaporated. Subsequently, an Nd5YAG laser were studied in conjunction with sensitivity helium was introduced into the chamber to attain a discharge and matrix eects.pressure. Finally, the rf power was switched on, and concurrently the computer started a program to trigger the laser and to acquire the data in the boxcar integration mode. EXPERIMENTAL The powdered samples contain known concentrations of the Instrumentation analytes of interest and one of the objectives is to investigate their use as standards. A schematic diagram of the experimental system is shown in Fig. 1. The components were basically the same as those used in previous work,1 except that the sample in the sample holder RESULTS AND DISCUSSION (1) was irradiated with a Q-switch Nd5YAG laser (7) through Eect of Laser Irradiation the quartz window (10), and that the computer data acquisition system was operated in a boxcar integration mode.Fig. 2 demonstrates typical signal peaks of aluminium emission, obtained for powdered silicon nitride (SN-BL JFCC, Al The sample holder (1), which also serves as the hot electrode and the atomizer, consists of a hollow Ta lid (2), a graphite 0.209%), in (a) the presence and (b) the absence of laser irradiation and Table 2 gives comparisons of the peak area cup (3) and a graphite bottom disk (4) with a cavity in which Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12 (823–826) 823Fig. 1 Schematic diagram of the experimental system: (1) sample holder; (2) tantalum lid; (3) graphite cup; (4) graphite disk; (5) rf power source; (6) dierent type of sample holder for sintered ceramics; (7) Nd5YAG laser; (8) laser beam-focusing lens; (9) spectrometer; (10) quartz window for laser irradiation; (11) central electrode; (12) discharge chamber; and (13) quartz window for optical observation.Table 1 Experimental conditions Rf power 500 W (2100 °C),* or 350 W (1800 °C),† SWR= 1.5–3 Laser pulse 10 pulses s-1, 124 mJ per pulse,* 270 mJ/pulse,† Q-switch mode He pressure 5.3 kPa He flow rate 1.0 l min-1 (NTP) Bandpass of 0.045 nm spectrometer Observation 0–25 mm above the sample holder top height PMT voltage -810 V (max.) Flow rate of 0.4 l min-1 coolant water * For powdered samples.† For sintered samples. and the precision in terms of relative standard deviation. Although both the peak height and area are increased, the eect of the laser irradiation is not as significant as expected from the eect of laser ablation, as exemplified by LA–ICP-MS, where no signal is observable in the absence of laser irradiation.The main cause of the relatively small enhancement could be explained as follows. It is seen from Fig. 2 that the peak profile is narrowed by the laser irradiation. The ratio of the full width at half maximum of the peak height is decreased by a factor of about 0.36. The peak narrowing suggests that the rate of sample powder introduction from the sample holder into the He plasma is increased, since powdered samples are readily Fig. 2 Eect of laser irradiation on the emission peak in direct blown o by a laser pulse-induced plume plasma, and the excitation of powder ceramics (a) with and (b) without laser irradiation.Sample, silicon nitride powder (SN-BL, Japan Fine Ceramics Center). samples are rapidly consumed, leading to a short duration Wavelength=Al I 396.153 nm. signal with a sharp decline. In the atomization/excitation of powder samples, such an eect seems to prevail over the laser ablation. In practice, no sample residue existed in the presence Table 2 Eect of laser irradiation on Al emission peak in direct of the laser irradiation whereas it was occasionally found excitation of powder ceramics (n=4).The sample and the conditions without the laser. are the same as in Fig. 1. The peak area is enhanced by about 2.6 fold despite the shorter residence time of the sample powder in the plasma. A Without laser With laser Ratio possible cause is the increase in eciency of analyte atomiz- Average peak area 930* 1600 2.6 ation6 and excitation by the laser pulse-induced plume plasma.(arbitrary units) It is seen from Table 2 that the precision is also increased RSD (%) 5.1 2.8 0.55 in the presence of laser irradiation. This is probably because the large fluctuation in thermal vaporization of the analyte is * Sample powder residue was occasionally found in the cup. This value is based on 100% introduction. relatively decreased by the increased rate of sample introduc- 824 Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12tion due to the blowing o of the samples by the laser pulse He plasma. In fact, a blue–white emission point was observed visually on the sample in the cup. Such an eect may be as described above. Fig. 3 illustrates the results for the sintered ceramics sample clarified by measuring the excitation temperatures or taking the Boltzmann plots of the analyte elements.7 The enhancement (Si3N4-9 JFCC, Mg 0.71%).In this case, the eect of the laser irradiation is observed more significantly and a noticeable of the normalized peak area by the ablated mass ratio of the ceramic (see the value 2.0 in Table 3) is lower than that of the enhancement in the emission signal is attained. Table 3 lists the eect quantitatively in terms of the changes in the sample powdered sample (see the value 2.5 in Table 3). This suggests that ablation of the ceramic forms relatively large particulates residue in the sample holder, the peak area, the relative standard deviation and the normalized peak area by ablated or agglomerates that are not atomized by the plasma and make no contribution to the emission signal.mass. The peak area is increased about 4.9-fold, which is greater than that obtained for the powder sample (about The reproducibility is also improved, as can be seen from Table 3. The contribution of laser ablation to the process of 2.6-fold). The net sample mass introduced into the plasma, estimated as the loaded mass minus the residue mass, is sample introduction into the He plasma becomes greater and consequently the fluctuation in the thermal vaporization increased about 2.5-fold on average by the laser irradiation.This is mainly due to the laser ablation eect. The normalized becomes less significant. peak area by ablated mass (Table 3) is increased about 2.0-fold, which suggests that eects other than the laser irradiation are Determination of Elements in Sintered Ceramics involved.The most likely phenomenon is tiny plasma plumes produced on the sample surface by the Q-switch laser radiation, The eect of the laser irradiation was tested in the determiwhich assist the atomization/excitation of the analyte in the nation of elements in practical samples, sintered ceramics. Direct analyses of sintered ceramics, made with calibrations using solution standards, are ideal in that no complex matrix matching is necessary.Such a direct determination of Mg in a sample of sintered silicon nitride was tentatively made with an Mg (as nitrate) standard solution (from Wako, Osaka, Japan). The analytical results in the presence of laser irradiation indicated an improvement in the relative error, but there was still a large negative error in the improved results (0.14%) from the reference value (0.71%). In order to promote the decomposition of the sample matrix in the atomization process, a solution of NH4F HF and (NH4)2S2O3 was added as a chemical modifier.However, problems such as an elevated blank level and instability of the He discharge arose. Fig. 4 shows the calibration curves using powdered standards of silicon nitride (SN-BL JFCC, Fe 0.310%) and boron carbide (B4C JFCC, Fe 0.24%), and the analytical results for silicon nitride (Si3N4-8 JFCC, Fe 0.048%) are listed in Table 4. It can be seen that the relative errors are significantly improved compared with those for the solution standards. It might be fortuitous that the accuracy (relative error) appears better than the precision (RSD) for boron carbide.However, it should be noted that the fairly good result was obtained using the powder standard in a dierent matrix, boron carbide from the specimen matrix. Hence, it is not necessary to use sintered standards of dierent analyte concentrations, which are sometimes dicult to obtain, and a powdered standard of one concentration is promising for the calibration.In conclusion, eects of laser irradiation such as improvements in sensitivity and precision were observed in direct Fig. 3 Eect of laser irradiation on emission peak in direct excitation of sintered ceramics (a) with and (b) without laser irradiation. Sample, sintered silicon nitride (Japan Fine Ceramics Center) cut to about 1 mm3 (6–11 mg) with a diamond cutter. Wavelength=Mg I 279.553 nm. Table 3 Eect of laser irradiation on Mg emission peak in direct excitation of sintered ceramics (n=3) Without laser With laser Ratio Average peak area 1360 6660 4.9 for 1 mg of the initial sample (arbitrary units) RSD (%) 22.8 3.6 0.16 Average vaporized 17.7 43.6 2.5 mass* (%) Normalized peak 7680 15300 2.0 Fig. 4 Calibration curve for direct determination of Fe in silicon area by ablated mass (arbitrary units) nitride using powder standards: # and A, with the powdered standard of silicon nitride; % and B, with the powdered standard of boron carbide.* Calculated as 100(1-residue mass/mass introduced). Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12 825Table 4 Direct determination of Fe in sintered silicon nitride using The authors thank Dr. M. Yanagisawa for the reference powder standards (n=3). Wavelength=Fe II 259.940 nm* materials and helpful discussions, Mr. S. Takahashi and Mr. K. Tachibana for constructing the atomization/excitation Powder standard source and Mr. T.Watanabe and Mr. T. Imura for skilful machining of ceramics. Silicon nitride Boron carbide Found (%) 0.039 0.042 Relative error (%) 23 14 REFERENCES from the reference of 0.048% 1 Kitagawa, K., and Katoh, T., J. Anal. At. Spectrom., 1992, 7, 539. RSD (%) 5.6 18 2 Inoue, H., Fukuda, M., Kitagawa, K., and Tsuge, S., J. Spectrosc. Soc. Jpn., 1996, 45, 232. * The sample of sintered silicon nitride and standards are those 3 Mochizuki, H., Sakashita, A., Iwata, H., Ishibashi, Y., and available from the Japan Fine Ceramic Center. Gunji, N., Anal. Sci., 1991, 7, 151. 4 Westheide, J. T., Becker, J. S., Ja�ger, R., Dietze, H.-J., and Broekaert, J. A. C., J. Anal. At. Spectrom., 1996, 11, 661. determinations of elements in sintered ceramic samples by 5 Arrowsmith, P., Anal. Chem., 1987, 59, 1437. atomic emission spectrometry using an rf atomization/exci- 6 Yanagisawa, M., and Takeuchi, T., Anal. Chim. Acta, 1974, 68, 212. tation source. Further improvements in accuracy and precision 7 Kaga, Y., Tsuge, S., Kitagawa, K., and Arai, N., J. Anal. At. are necessary for practical use. Among the promising tech- Spectrom., submitted for publication. niques for increasing the sample introduction eciency are optimization of the chemical modifiers and elevation of the Paper 6/07937G laser power. Received November 22, 1996 Accepted February 24, 1997 826 Journal of Analytical Atomic Spectrometry, August 1997, Vol.

ISSN:0267-9477    

DOI:10.1039/a607937g

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Pattern Recognition for Sample Classification Using Elemental Composition—Application for Inductively Coupled Plasma Atomic Emission Spectrometry CHRISTINE SARTOROS AND ERIC D. SALIN Department of Chemistry, McGill University,Montreal, Quebec, Canada Three pattern recognition techniques were investigated as tools the possibility of using pattern recognition with elemental concentrations for automatic sample identification. for automatic recognition of samples: k-Nearest Neighbors, Bayesian Classification and the C4.5 inductive learning algorithm. Their abilities to classify 20 geological reference EXPERIMENTAL materials were compared.Each training and test example used 13 elemental concentrations. The data set was composed of The general method for classifying samples is illustrated in 2582 examples obtained from CANMET in the form of Fig. 1. The pattern recognition technique is applied to a set of results of analyses performed on these reference materials by examples, the training set, to develop classification rules.These dierent laboratories. It was found that all three pattern rules are then tested using another set of examples, the test recognition techniques performed extremely well with a large set. The sample types of the examples in the test set are known data set of real samples. Bayesian Classification and k- and used to evaluate the performance of the classification Nearest Neighbors worked very well with small data sets.techniques. Keywords: Pattern recognition; autonomous instrument; ICPAES k-Nearest Neighbors The k-Nearest Neighbors5 algorithm is a simple statistical technique which does not exactly follow the general method The Autonomous Instrument Project involves developing described in Fig. 1. An example in the training set is denoted software and methodologies which allow instruments to run as the vector Eclass,i(aj)=(a1, a2, . . . , am) where class is the ‘intelligently’ with an absolute minimum of supervision.1–4 One sample type of the ith example in the training set, and aj is characteristic of the ideal autonomous instrument would be the jth attribute in the example.An example in a test set will the ability to classify samples. The classification can be used be the vector T(aj)=(a1, a2, . . . , am). To classify an example in for two purposes: (1) selecting operating conditions and cali- the test set, the Euclidean distance between it and every bration methodology (e.g., internal standards, standard example in the training set is calculated.Since large changes additions), and (2) identification of the sample (e.g., 440 in attributes of large values would influence the calculation stainless steel ). The first case consists of (a) performing a more than big changes in attributes of small values, a ‘relative’ preliminary analysis (a semi-quantitative scan), ( b) recognizing Euclidean distance was used. that the apparent elemental composition is similar to a sample class that has been successfully run with the use of a certain d=S.m j=1 C(Eclass,i(aj)-T(aj)) T(aj) D2 methodology and (c) adopting the same methodology.3 The second case consists of (a) performing a preliminary analysis, These distances are then used to determine the closest neigh- and (b) finding an exact (or very close) match in the sample bors (i.e., smallest Euclidean distance) to the test example. The database using the apparent elemental composition of the nearest k neighbors are selected and the frequency of each is sample.In a recent paper,3 we evaluated the potential of determined. The class to which the majority of training several numeric processing techniques for the identification of examples belong is assigned to the test example. If a tie should samples. A wide variety of reference materials were selected, occur, the class with the closest neighbors is selected. ranging from clinical through botanical to geological, using their reported elemental concentrations and RSDs to generate validation (test) sets and training sets.The pattern recognition techniques that were used were k-Nearest Neighbors (kNN),5 Bayesian Classification6 and inductive learning.7,8 These three techniques were described in some detail.3 The reference materials used to study the three pattern recognition techniques were dissimilar and would not have posed as much of a problem as in a study in which the materials were similar.In addition, the data was not ‘real’ in the sense that the standard deviations were not obtained experimentally but were fabricated as described. In this study, geological reference material data were used for test and training sets in the comparison of the three pattern recognition techniques. These geological materials were much more similar than those in the previous Fig. 1 Flow chart of steps in pattern recognition. study and consequently provide a much more rigorous test of Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12 (827–831) 827Bayesian Classification concentration for each of the 13 most commonly analyzed elements in the reference materials. The data set contained Bayesian Classification6 is a probabilistic technique of pattern unknown attributes since the concentration of each of the 13 recognition. It is based on the assumption that the classification elements was not available for each reference material.The problem is posed in probabilistic terms. It also assumes that C4.5 inductive learning algorithm is capable of handling all the probability values are known. Using the mean and unknown attributes with their values set to a question mark standard deviation of examples of a class in the training set, (‘?’). Since the other two techniques, kNN and Bayesian the Bayesian Classification technique determines the prob- Classification, require all attributes to be known, the unknown ability that an example in the test set belongs to a particular attributes needed to be set to some value. For kNN, the class.The class with the highest probability is assigned to the unknown attributes (concentrations) were set to zero. Fig. 2 example in the test set. In this study the distributions of the depicts the frequency of samples in each class that belong in examples were assumed to be Gaussian in nature. the dierent Euclidean distance ranges relative to the zero point.For Bayesian Classification, the attributes (concentrations) in a class that were sometimes known were set to the C4.5 Inductive Learning value of the mean of that attribute in that class; the attributes Inductive learning has been mainly used to generate rule-based in a class that were completely unknown were given a mean expert systems although it is also a powerful technique for (999 999.5) far removed from any possible value and a very pattern recognition.By evaluating examples in a training set, small standard deviation, thereby eliminating that element as inductive learning has the ability to infer general relationships a classifier. The average concentrations of each element in each about these examples. In this study, the C4.5 induction engine class are shown in Fig. 3, which illustrates the relative similarity developed by Quinlan7 was used. The output of the C4.5 of some materials. (Note: The maximum concentration on the induction algorithm is in the form of a decision tree.The graph is 600 000 mg g-1 so that the lower concentrations can algorithm determines which data attribute best divides the be seen.) Fig. 4 shows both the RSD with a material type as examples in the training set into distinct classes. It separates well as deviations across sample types for a given element. The the examples in such a way that any pattern in the data is made apparent. This results in a hierarchical structure of decisions, i.e., a decision tree.The decision tree can then be used to classify the examples in the test set. The Data Sets The data set used in this study consisted of the elemental compositions for 20 reference materials obtained from CANMET, Mineral Science Laboratories (Ottowa, Canada) (Table 1). The data set was used for all three techniques, kNN, Bayesian Classification and the C4.5 inductive learning algorithm. Computer programs for kNN classification and Bayesian Classification were written in Borland International’s (Otis Valley, CA, USA) Turbo Pascal 7.0.C4.5 was written in Watcom International’s (Waterloo, Ontario, Canada) C/C++ v.10.0 for OS/2. All computer programs were written in our laboratory and were executed on a 66-MHz 486 PC type computer. The data set was composed of results of multiple repeat analyses obtained for each reference material by dierent laboratories. Each training and test example consisted of the 200 100 0 05010035050075010001500500010 000100 000200 000300 000400 000500 000700 000 CCU-1b CPB-1 CZN-1 KC-1a MP-1a MP-2 MW-1 PTC-1a PTM-1a RTS-1 RTS-2 RTS-3 RTS-4 SCH-1 SU-1a TBD-1 WGB-1 WMG-1 WMS-1 SY-4 Classes Euclidean distance Frequency of samples Fig. 2 Frequency of samples from each class in each Euclidean distance range. Table 1 Reference materials from CANMET Number of examples in Reference material data set Copper concentrate CCU-1b 133 Lead concentrate CPB-1 274 Zinc concentrate CZN-1 298 Zinc–lead–tin–silver ore KC-1a 90 Zinc–tin–copper–lead ore MP-1a 119 Tungsten–molybdenum ore MP-2 79 Iron ore MW-1 82 Noble metals bearing sulfide concentrate PTC-1a 98 Noble metals bearing nickel–copper matte PTM-1a 97 Sulfide ore mill tailings RTS-1 30 Sulfide ore mill tailings RTS-2 43 Sulfide ore mill tailings RTS-3 58 Sulfide ore mill tailings RTS-4 47 Iron ore SCH-1 271 Nickel–copper–cobalt ore SU-1a 216 Diorite gneiss SY-4 248 Diabase rock PGE material TDB-1a 101 Gabbro rock PGE material WGB-1 96 Mineralized gabbro PGE material WMG-1 102 600 000 500 000 400 000 300 000 200 000 100 000 0 PTM-1a RTS-2RTS-4SU-1a CCU-1b CZN-1MP-1a MW-1 WGB-1 WMS-1 RTS-1 RTS-3SCH-1 TDB-1 CPB-1KC-1aMP-2 PTC-1a WMG-1 SY-2 Classes Elements AsCdCoCuAuFePbNiPdPtAgSZn Average concentration/mg g–1 Massive sulfide PGE material WMS-1 100 Fig. 3 Average concentration of each element in each class. 828 Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12Table 3 Classification results for kNN Test set Size relative to data set (%) Rate of success (%) 1 25 97.1 2 25 97.6 3 25 97.3 4 25 97.3 5 26 97.2 6 22 97.9 7 50 95.7 8 50 96.3 9 54 95.4 10 51 95.4 11 52 96.1 12 51 96.2 Table 4 Results for Bayesian Classification Test set Size relative to data set (%) Rate of success (%) 1 25 95.8 2 25 95.9 100 80 0 20 40 60 PTM-1a RTS-2 RTS-4SU-1a CCU-1b CZN-1 MP-1a MW-1 WGB-1WMS-1 RTS-1 RTS-3 SCH-1 TDB-1 CPB-1KC-1a MP-2 PTC-1a WMG-1 SY-4 Standard deviation (%) Classes Elements AsCdCoCuAuFePbNiPdPtAgSZn 3 25 95.6 Fig. 4 RSD of the concentration of each element in each class. 4 25 96.6 5 26 96.1 6 22 95.6 Table 2 Reference materials used in subset 7 50 99.3 8 50 96.5 Number of examples in 9 52 97.8 Reference material data set 10 52 98.7 Sulfide ore mill tailings RTS-1 18 11 52 98.4 Sulfide ore mill tailings RTS-2 25 12 54 98.5 Sulfide ore mill tailings RTS-3 34 Sulfide ore mill tailings RTS-4 28 Table 5 Classification results of C4.5 inductive learning Test set Size relative to data set (%) Rate of success (%) examples used in the test sets were selected at random from 1 25 92.3 the data set.In addition, the sizes of the test sets were varied 2 25 91.3 to provide a variety of testing environments. 3 25 93.5 Due to the inconvenience of the missing data, further studies 4 25 92.6 were done on a subset of the CANMET data consisting of 5 26 92.0 only the data from the Sulfide Ore Mill Tailings (RTS-1, 6 22 93.2 RTS-2, RTS-3 and RTS-4).The number of elements studied 7 50 91.5 8 50 90.7 was reduced from 13 to 9 and all examples with missing data 9 51 91.8 were eliminated. The numbers of remaining examples for each 10 51 92.4 sample type are listed in Table 2. This particular set was 11 52 91.7 selected for the similarity of the materials. 12 54 90.7 RESULTS AND DISCUSSION extremely well, with Bayesian Classification and kNN performing the best. Fig. 2 is a depiction of the number of examples lying in a Euclidean distance range with respect to the origin. The Even though these results are extremely good and demonstrate how well these three techniques can perform with large examples in the data set are grouped with their respective sample types. Some of the sample types can be seen to fall in data sets, the data and the tests performed are not necessarily representative of all situations. The Autonomous Instrument the same range (e.g., CCU-1b and CPB-1) which could make them dicult to distinguish with a technique like kNN.Fig. 3 system that we are developing will not be initially supplied with such a complete set of data. It will have to build it in illustrates the average concentration of the elements in the various sample types. This figure can be read in two dierent time with experiments. In addition, this system will always perform a fast analysis (a semiquantitative scan) at standard ways.The first is to start at the axis labeled ‘Elements’ and to go across in parallel with the ‘Classes’ axis; the concentration conditions of a sample to provide an estimate of the concentrations of all the elements. Since this will be performed every of an element can be viewed and compared for the various sample types. The second method is to start at the ‘Classes’ time and the concentrations of all the elements will be known, there will not be any missing data in the training or test set.axis and go in parallel with the ‘Elements’ axis; the concentration of all elements in a sample type can be viewed. The The system would start with an empty training set the first time it is run. This first analysis would be the first example in pattern for a sample type seen by this method can be used for comparison with other sample types. Fig. 4 has a similar layout the training set and for every other analysis performed an example would be added to the training set as long as the except that it shows standard deviations rather than concentrations.This figure can be read in the same manner as Fig. 3. sample type is known. The system will use this training set (or database) when running an unknown sample to find the closest A casual examination of both Figs. 3 and 4 would suggest that it might be dicult to dierentiate between some of the sample match; that information can then be used to select operating conditions and calibration methodology.Since the number of types with a technique such as Bayesian Classification. The classification results of the three techniques are listed in Tables examples of each class in the training set can vary and could be as low as one example, a set of experiments was conducted 3, 4 and 5. As can be seen, all three classifications perform Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12 829Table 6 Classifications based on number of examples in training set Average rate of success of pattern recognition technique (%) Number of Number of examples in dierent cases k-Nearest Bayes C4.5 Inductive training set tests neighbor classification learning 1 10 92 88 44 2 10 92 96 61 3 10 97 83 65 5 5 97 100 83 10 3 100 100 86 in which only one neighbor (k=1) was used for k-Nearest training examples selected for the RTS-1 sample type were so similar that the standard deviations of some of their attributes Neighbors.For Bayesian Classification, the one-example experiments were conducted using a small percentage of the were zero or close to it.When the standard deviation is very low or zero, the probability of a test example’s attributes mean as the standard deviation (e.g., 1%). To simulate the way our system would work, we started with an empty training set. belonging to that particular class becomes zero or close to it. This results in very few test examples being assigned to the We used a subset of the CANMET data which only consisted of the four types of Sulfide Ore Mill Tailings with no missing RTS-1 sample type.The kNN technique did not exhibit the same diculties since the technique does not rely on standard values. A random example of each sample type was put into the training set. The remaining examples were put into the test deviations. This same phenomenon was observed in one of the cases for the three training examples of the sample types set. Ten dierent combinations of training and test sets were created and all three pattern recognition techniques were RTS-1, RTS-2 and RTS-3.The training examples of the RTS-4 sample type, on the other hand, had relatively large standard applied to them, searching for the closest match. This described the first time the system was used for classification of an deviations which resulted in most of the examples in the test set being assigned to the RTS-4 sample type. Bayesian unknown sample with respect to four sample types; this was also repeated using two, three, five and ten examples of each Classification was only able to correctly classify 54% of the test set.sample type in the training set. The average success rates are listed in Table 6. Between the two techniques it would seem that kNN would suit our purposes best for small numbers of examples. The Upon examination of the average success rate of the three techniques (Table 6), C4.5 inductive learning did not perform minimum requirements in our system are that there be at least one example of each sample type in the training set and that well with very few examples in the training set; however, its classification performance did improve significantly as the the examples in the training set be representative of their class.number of examples in the training set increased. With a small number of examples in the training set it seems that both kNN CONCLUSION and Bayesian Classification performed well (Fig. 5).In the tests based on two examples of each type in the training set, It has been shown that for the classification of 20 reference there were two cases where one technique outperformed the materials, with the use of their elemental compositions obtained other. In the first, kNN had a success rate of 70% whereas experimentally by various laboratories, C4.5 inductive learning, Bayesian Classification performed at 96%. The two training kNN and Bayesian Classification all perform extremely well.examples selected for the RTS-4 sample type were not represen- In choosing a pattern recognition technique, there are several tative of the whole class. In fact, these two were situated important characteristics to consider: speed, classification accusomewhere in between the RTS-2 sample type and the RTS-4 racy and clarity of results. C4.5 and Bayesian Classification sample type such that the Euclidean distances calculated were generate their rules and means and standard deviations, so close that the test examples were sometimes classified as respectively, once, and hence take approximately the same the RTS-2 sample type.Bayesian Classification, since it uses amount of time to classify an example. The kNN technique means and standard deviations to calculate the probabilities, can be much slower since the test example must be compared did not misclassify the test examples. The probabilities of the with every training example and the amount of time needed is test examples belonging to either class were nonetheless com- dependent on the number of examples in the training set.We petitive. In the second case, Bayesian Classification performed have seen that in classification accuracy all three techniques at 71%, whereas kNN had a success rate of 94%. The two performed extremely well when large training sets were used. As for clarity of results, it depends on what information is required.C4.5 provides a decision tree demonstrating which attributes are being used in the classification. Bayesian Classification provides a statistical probability of a test example belonging to a particular class and can be used as a measure for possible misclassifications. kNN provides the Euclidean distance between the test example and its nearest neighbors, which can be used as a measure of goodness. Using small training sets, both kNN and Bayesian Classification performed well whereas C4.5 required larger training sets.For the Autonomous Instrument system, kNN seems to be the appropriate choice for the pattern recognition module. The hypothesis from our last paper3 (i.e., that pattern recognition for sample identification is highly practicable) has been strongly reinforced by this more recent work. It would seem that modern ICP-AES spectrometer systems, using readily available Fig. 5 Classification of test samples based on number of examples in training set. pattern recognition techniques, have a high probability of 830 Journal of Analytical Atomic Spectrometry, August 1997, Vol. 122 Webb, D. P., and Salin, E. D., Intelligent Instruments and Computers, being able to extract considerably more useful information 1992, 5, 185. than is presently extracted from data which is already in 3 Branagh, W., Yu, H., and Salin, E. D., Appl. Spectrosc., 1995, the system. 49(7), 964. 4 Branagh, W., and Salin, E. D., Spectroscopy, 1995, 10, 20. The authors gratefully acknowledge William Bowman at 5 Sharaf, M. A., Illman, D. L., and Kowalski, B. R., Chemometrics, CANMET for providing the data set, and financial support John Wiley, New York, 1986. from the National Sciences and Engineering Research Council 6 Duda, R. O., and Hart, P. E., Pattern Classification and Scene Analysis, John Wiley, Toronto, 1973. of Canada. C. S. would like to acknowledge financial support 7 Quinlan, J. R., C4.5 Algorithms for Machine L earning, Morgan from Fonds pour la Formation de Chercheurs et l’Aide a` la Kaufmann Publishers, San Mateo, 1993. Recherche. 8 Salin, E. D., and Winston, P. H., Anal. Chem., 1992, 64, 49A. REFERENCES Paper 6/08166E Received December 3, 1996 1 Webb, D. P., Hamier, J., and Salin, E. D., T rends Anal. Chem., 1994, 13, 44. Accepted May 15, 1997 Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12 831

ISSN:0267-9477    

DOI:10.1039/a608166e

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Determination of Titanium in Fine Gold by Electrothermal Atomic Absorption Spectrometry MICHAEL W. HINDS*a , IAN L. SHUTTLERb AND CYNTHIA PRIEST BOSNAKc aRoyal Canadian Mint, 320 Sussex Dr., Ottawa, Ontario, K1A 0G8, Canada bBodenseewerk Perkin-Elmer GmbH, Postfach 10 17 61, D-88647 U� berlingen, Germany cPerkin-Elmer Corporation, 2000 York Road, Suite 132, Oak Brook, IL 60522-3608 The transversely heated graphite atomizer (THGA) proved to expected that the absence of lower temperature sites would eliminate or substantially reduce analyte carry-over.be better suited to the task of determining Ti in gold owing to the reduction in analyte carry-over compared with a Massman This paper presents a brief comparison between THGA and HGA Massman atomizers for the atomization and the determi- type atomizer. Carry-over was also reduced by using a second maximum power atomization step (2600 °C) as a clean-out nation of Ti in a gold matrix, mainly in terms of analyte carryover.Results of experiments to establish optimal conditions step and cycling through the temperature program twice with no sample between solution samples. Platform atomization for titanium atomization (THGA) are presented and accuracy is assessed by analyses of gold reference materials with certified with the THGA was preferred because of the slightly higher observed sensitivity, especially as increasing amounts of gold concentration values for the titanium content.matrix per aliquot attenuates the analytical signal. The tube lifetime was lower owing to the high temperatures and long EXPERIMENTAL atomization times (10 s). The limit of detection (3s) was Apparatus estimated to be 5 mg l-1 based on 20 ml aliquots containing 40 mg of gold. The accuracy of the method was verified by Experiments were conducted on two dierent spectrometer/ determining titanium in gold reference materials with known graphite electrothermal atomizer combinations.Titanium was concentrations of titanium. Typical precision was determined in a Massman atomizer using a Perkin-Elmer approximately 4% (RSD) for 0.2% gold solutions containing HGA 500 atomizer in conjunction with an AS 40 autosampler 42 mg l-1 Ti (20 ml aliquot). (Perkin-Elmer Bodenseewerk, U� berlingen, Germany). Only wall atomization was measured from non-grooved pyrolytically Keywords: Atomic absorption spectrometry; electrothermal coated graphite atomizers (P-E part no.B010-5197). Atomic atomization; precious metal analysis; titanium absorption measurements were made on a P-E Model 5000 spectrometer with continuum source background correction. High purity gold (99.99%) refined at the Royal Canadian Mint The spectrometer was interfaced to a personal computer is assayed by determining the concentration of a suite of through a DAS8, 12 bit analogue-to-digital converter (Keithley elements that are known to be common impurities and sub- Metrabyte, Taunton, MA, USA).Data collection software was tracting these values from 100% to arrive at a purity value. similar in design to the work of Allen and Jackson.5 The Because of the monetary value of gold, rapid assays are atomizer was a pyrolytic graphite coated electro-graphite tube. required to maintain the flow of gold between the Mint and The other spectrometer employed was a Perkin-Elmer its customers. Solid sample atomic spectrometric techniques Model 4100 ZL atomic absorption spectrometer which such as spark ablation ICP-AES and laser ablation ICP-MS, included a transversely heated graphite electrothermal atomare routinely used for this purpose.However, solid sample izer and longitudinal Zeeman-eect background correction. calibration standards (gold matrix) are required for these Samples were dispensed by a Model AS 70 autosampler which methods. A set of gold reference materials were manufactured is an integral part of the system. Both platform and wall with dierent concentrations of the elements of interest and atomization were used.Platform atomization occurred from analysed by the Mint for the purpose of calibrating solid the graphite tube with an integrated platform (P-E part no. sample atomic spectrometers. Titanium is one of the elements B050-4033) and wall atomization occurred from the same monitored in the refining process and was therefore included atomizer with the platform manually removed. The temperain the suite of elements added during the manufacture of these ture programs used from both atomizer designs are summarreference materials.ized in Table 1. Atomic absorption measurements were eected Titanium was determined in plant materials1 and coal fly by PEAALABS software, version 6.2, running on an IBM ash2 by ETAAS with Massman type atomizers. In both cases, Personal System 2 computer, Model 70 386. high temperatures were used for atomization and clean-out In both instruments, a titanium hollow cathode lamp was steps.Because of the refractory nature of titanium, multiple used with a lamp current of 15 mA. The most sensitive clean-out steps1 or a set of multiple empty temperature wavelength, 364.3 nm, was used with a slit width of 0.2 nm. program cycles2 were used to minimize carry-over. The analyte atomic signal, from a Massman type atomizer, Reagents was variable mainly owing to the high carry-over between sample injections. It has been shown by Frech et al.3 that the High purity hydrochloric and nitric acids (Seastar Chemicals, Sidney, BC, Canada) were used for dissolving gold samples.cooler ends of the end-heated Massman atomizers (HGA) act as condensation sites for more refractory elements, from where Water used in these experiments was distilled and de-ionized using a Nanopure II system (Barsted/Thermolyne, Dubuque, re-atomization can occur. The transversely heated graphite atomizer (THGA) is virtually isothermal over its length owing IA, USA).Calibration standards were diluted from 1000 mg ml-1 stock solution made from high purity titanium to the electrical contacts at either side of the atomizer.4 It was Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12 (833–836) 833Table 1 Temperature programs for the determination of Ti in dier- any possible analyte carry-over interfering with the measureent atomizers ment of the next sample. For the determination of Ti in FAU6 the method of standard Part A Massman T ype Atomizer— additions was used by adding a 10 ml aliquot of 50 ng ml-1 Ti Step Temperature/°C Ramp Time /s Hold Time /s solution to the 30 ml sample aliquot. The temperature program 1 200 10 10 outlined in Table 1, Part A was used for the determination, 2 20 1 5 although two more clean-out steps at 2700 °C were added to 3 2700 0 10* minimize analyte carry-over. 4 20 1 10 5 2700 1 5 RESULTS AND DISCUSSION Part B THGA atomizer (two maximum power heating step)— Step Temperature/°C Ramp Time /s Hold Time /s Atomizer and Temperature Program Studies 1 120 1 30 Analyte carry-over: Massman versus transersely heated 2 130 10 40 atomizers 3 1500 10 20 4 2600 0 10† The analyte carry-over in a Massman-type atomizer is illus- 5 20 1 30 trated in Fig. 1(a). The integrated absorbance observed for 6 2600 0 7 7 20 1 30 each pair of empty firings increases as Ti is introduced into the atomizer. It is also evident that the precision for samples * BOC sequence started 1 s before maximum power heating; argon with measureable amounts of Ti is variable.As noted in the gas flow stopped (otherwise at 250 ml min-1). introduction, analyte carry-over can be explained by the work † BOC sequence started 2 s before maximum power heating; argon done by Frech et al.,1 who showed that analyte atoms condense gas flow stopped (otherwise at 250 ml min-1). along with the metal (Pd) matrix modifier at the cooler tube ends of a Massman type atomizer.During the next atomizer (99.99%) dissolved in hydrochloric acid (High Purity heating cycle the condensed analytes at the tube ends are Standards, Charleston, SC, USA). Calibration solutions also re-atomized because, at the early stage of the atomization contained 0.1 g of high purity gold (Metalor, North cycle, the high temperature zone extends almost the length of Attelborough, Massachusetts) in 50 ml (0.2) and 1% the atomizer.8 However, as the atomization cycle progresses hydrochloric acid to match the sample solution matrix.the high temperature zone moves back from the tube ends which sets up another cycle of analyte/matrix condensation at the tube ends. Titanium does form refractory titanium carbide Samples and Sample Preparation through interaction with the graphite upon pyrolysis and Samples were fine gold reference materials produced and atomization. The titanium carbide formation9 particularly near analysed by the Royal Canadian Mint.Each of these reference the atomizer ends and the refractory nature of titanium contribmaterials was made by adding set amounts of 16 elements to utes to analyte carry-over. molten gold. Further treatment of each material resulted in Results from a transversely heated atomizer are shown in the trace elements being homogeneously distributed in the gold matrix. A more complete description is summarized by Kogan et al.6 The added elements are the same as the elements required in the American Society for Testing Materials specifi- cation for high purity gold,7 plus other elements that are of interest to the Mint.Samples for analysis were prepared by taking 1 g of turnings which were taken from dierent areas of the reference material bar and combined. Weighed samples were dissolved in 10 ml aqua regia (3 parts concentrated hydrochloric acid and 1 part concentrated nitric acid) in a closed vessel microwave system (Model MDS81D, CEM, Matthews, NC, USA).Prior to sealing, each Teflon vessel was flushed with argon gas to minimize the risk of igniting the hydrogen released during the dissolution process. A pressure controlling device maintained the pressure at 100 psi (1 psi#6.895 kPa). The dissolved gold solution was transferred quantitatively into 50 ml precalibrated plastic volumetric flasks and diluted to volume with water. A 10-fold dilution was performed prior to analysis by taking 100 ml of the gold solution and mixing with 900 ml of water directly in the autosampler cups.Analytical Procedures The Ti signals for each sample and standard were measured with duplicate injections and in a few instances in triplicate. Aliquot volumes varied between 20–30 ml depending on the analyte concentration. In half the determinations an additional 10 ml of water was also injected with the sample to assist in Fig. 1 Integrated absorbance observed for a series of atomizer firings spreading the sample on the atomizer surface.There did not with 20 ml aliquots of 0.2% Au solutions with titanium present. Two appear to be a dierence with or without this added water. empty atomizer firings occur between each duplicate sample firings: Between each sample or standard, two temperature program (a) Massman atomizer, with 1 ng Ti per aliquot; (b) THGA with 0.8 ng per aliquot. cycles were performed without any sample in order to minimize 834 Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12Fig. 1(b). There is a marked decrease in the amount of carry- of a gold matrix. The HGA atomizer was observed to have a higher lifetime (90–110) but this was partially oset by the over observed for the empty firings between samples containing 0.8 ng Ti per aliquot and improved reproducibility for repli- high number of atomizer firings required to clean out the atomizer after the injection of solutions containing more than cates.The integrated absorbance from the second empty firing nearly returns to the values observed for the Au blank. 50 mg l-1 Ti. Eect of Amount of Gold on Integrated Absorbance Comparison between wall and platform atomization in a transversely heated atomizer As noted previously, the integrated absorbance observed for titanium decreases as the amount of gold per aliquot increases Titanium can be atomized eciently from a platform owing to the faster heating characteristics of the integrated platform (Fig. 2). A sharp background absorbance peak occurs about one second into the atomization step which corresponds to a of the transversely heated atomizer. As with Massman atomizers, wall atomization can also be considered, especially since rapid evolution of vapour (or smoke) through the dosing hole. The Ti signal just begins to be observed above the baseline as atomization of titanium does not occur until most of the gold matrix vaporizes and near the time when the atomizer tempera- the background signal peaks.This is similar to the eect observed in a paper by Frech et al.,10 in which the phenomenon ture is isothermal. A comparison of sensitivity and analyte carry-over for wall was explained by the gas phase analyte atoms adsorbing on metal matrix particles which form in the gas phase by cooler and platform atomization in the THGA is given in Table 2. For aqueous solutions of Ti there does not appear to be a atmospheric gas entering the atomizer.The amount of analyte carry-over in the atomizer is approxi- substantial dierence between the two signals produced from the atomization sites. However, the use of the platform results mately 7% of the analyte signal without the presence of the gold matrix and slightly increases with the amount of gold in slightly higher integrated absorbance (Qa) when titanium is determined in a gold matrix. This is particularly important matrix atomized. This is denoted from the titanium peak area measured for the first empty atomizer firing immediately because of the reduced sensitivity of titanium in the presence of large amounts of gold matrix.Consequently, platform following a gold sample containing a fixed concentration of titanium. Since there are no cooler ends in the transversely atomization is preferred despite the lower carry-over demonstrated by wall atomization. Carry-over can be minimized (for heated atomizer, it is likely that carry-over occurs as a result of titanium carbide formation within the atomizer graphite or either wall or platform) by including empty firings between samples.on the surface because of surface and graphite lattice imperfections. The increase of Ti carry-over in the presence of gold From Table 2, the characteristic mass (mo) value for titanium without the presence of gold obtained by platform THGA is may occur owing to the increased carbide formation from a greater number of active sites (surface imperfections) formed lower (more sensitive) than the instrument manufacturer’s expected mo value of 70 pg.This occurs because of the higher from more material interacting with the graphite. Another explanation comes from the work of Eloi et al.11 who found atomization temperature and the longer atomization/read time used in this study. that analytes in aqueous solution migrate into the graphite. It is probable that for a 0.2% Au solution, gold will migrate into the graphite and will take Ti with it into the lattice.The T emperature program for the transversely heated atomizer portion of Ti in the graphite lattice is less likely to be totally removed after the normal temperature program cycle. It was thought that analyte carry-over could be reduced by having two clean out steps in the HGA temperature program at the maximum temperature of 2600 °C (Part A Table 1) as previous demonstrated at 2700 °C with a Massman type atomizer. 1 The carry-over was about 14% of the signal obtained from 200 mg l-1 Ti in 0.1% m/v gold. Temperature programs with three or four clean-out steps did not substantially reduce the carry-over and often caused the furnace power supply to overheat, drawing more current than the internal circuit breakers permit, causing an interuption in the analysis run. A second maximum power heating step is permitted by the P-E 4100 ZL software and firmware. This was used as indicated in Table 1 (Part B).Although the inclusion of a second maximum power heating step (as a clean out step) does not eliminate analyte carry-over, it does require less time and energy, and does not appear to tax the power supply as is the case for multiple clean out steps. The atomizer lifetime is quite low at 60 temperature cycles for the THGA. This is not Fig. 2 Eect of the amount of gold matrix on the integrated unexpected because of the high atomization temperature, long absorbance from gold solutions containing 100 mg l-1 Ti (20 ml aliquot) atomization time (10 s), and second maximum power ramp using the THGA (platform): &, Ti integrated absorbance from gold step used for clean-out.This appears to be a reasonable trade- solutions; +, Ti integrated absorbance from first empty atomizer firing after the gold containing solution. o for the higher sensitivity of Ti determinations in the presence Table 2 Comparison between platform and wall atomization from THGA atomizers for titanium Sample Atomizer type Mean Qa/s m0/pg Residual Qa/s 2 ng Ti Platform 0.152 58 0.013 Wall 0.166 53 0.013 2 ng Ti+100 mg Au Platform 0.066 130 0.010 Wall 0.053 170 0.003 Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12 835Table 3 Comparison of the determined concentration of Ti (mg g-1) Recent work by Goltz et al.12 has shown that 0.3% Freon in gold reference materials by ETAAS to the reference concentration 23 (CHF3) in Ar was eective in reducing uranium carbide values formation (as monitored by ETV-ICP-MS).A similar approach may be useful in reducing analyte carry-over due to titanium Reference material Determined concentration* Reference value† carbide formation. FAU9 6.1±1.1 (n=7) 5.9±1.1 FAU11 18.2±1.4 (n=9) 16.5±3.1 FAU12 21.6±0.7 (n=8) 22.29±0.25‡ Analytical Results * ±Standard deviation. Assessment of accuracy and precision † ±95% confidence interval. Titanium was determined in three fine gold reference materials ‡ ICP-AES determination±standard deviation.and the results are presented in Table 3. The determined concentration value with noted standard deviation overlaps has been verified with reasonable precision and limits of with the reference concentration values. Although analyte detection. carry-over was minimal, duplicate empty firings occurred between each sample and standard. As noted previously, this REFERENCES decreased the atomizer lifetime but this ensured the lowest amount of carry-over. 1 Lopez Garcia, I., Vinas, P., and Herna�ndez Co�rdoba, M., J. Anal. Typically, the relative standard deviation for two replicates At. Spectrom., 1992, 7, 529. was found to be approximately 4% for a 20 ml aliquot of 2 Bhattacharyya, S. S., Chakraborty, R., and Das, A. K., Anal. 42 mg l-1 Ti within a 0.2% gold solution. L ett., 1993, 26, 341. 3 Frech, W., Li, K., Berglund, M., and Baxter, D. C., J. Anal. At. Spectrom., 1992, 7, 141. L imit of detection 4 Sperling, M., Welz, B., Hertzberg, J., Reick, C., and Marowsky, G., Spectrochim.Acta, Part B, 1996, 51, 897. The limit of detection for titanium in a 0.2% gold solution 5 Allen, E., and Jackson, K. W., Anal. Chim. Acta, 1987, 192, 355. was estimated to be 5 mg l-1 based on 10 replicate determi- 6 Kogan, V.V., Hinds, M.W., Ocampo, G., and Valente, G., Precious Metals 1993, ed. Mishra, R., International Precious Metal nations of gold blank solution (k=3). For a 20ml aliquot Institute, 1993, p. 101–116. (40 mg Au), this corresponds to 2.5 mg g-1 in solid. This limit 7 Annual Book of ASTM Standards, Nonferrous Metal Products, B of detection appears to be acceptable to determine the concen- 562–86, vol. 02.04, American Society for Testing and Materials tration of Ti at higher levels of between 6 and 10 mg g-1 which Philadelphia, PA, 1991, p. 425. was a requirement of this research project. 8 Falk, H., Glismann, A., Bergann, L., Minkwitz, G., Shubert, M., and Skole, J., Spectrochim. Acta, Part B, 1985, 40, 533. 9 Weast, R. C., Editor, Handbook of Chemistry and Physics, 68th CONCLUSION edition, CRC Press, Boca Raton, FL, p. B-140. 10 Frech, W., L’vov, B. V., and Romanova, N. P., Spectrochim. Acta, It is evident from the reduction in analyte carry-over that the Part B, 1992, 47, 1461. THGA atomizer is preferred for the determination of Ti in a 11 Eloi, C., Robertson, J. D., and Majidi, V., J. Anal. At. Spectrom., gold matrix. High atomization temperatures (2600 °C), longer 1993, 8, 217. atomization times and a second maximum power step maxim- 12 Goltz, D. M., Chakrabarti, C. L., Gre�goire, D. C., and Byrne, J. P., Spectrochim. Acta Part B, 1995, 50, 803. ize sensitivity and minimize carry-over, but unfortunately reduce atomizer lifetimes. This appears to be an excellent trade-o in view of the diculty of determining such a refrac- Paper 7/00268H Received January 10, 1997 tory element in a relatively large amount of gold matrix that acts to suppress the sensitivity. The accuracy of the method Accepted April 15, 1997 836 Journal of Analytical Atomic Spectrometry, August 1997, V

ISSN:0267-9477    

DOI:10.1039/a700268h

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Improved Determination of Aluminium in Port Wine by Electrothermal Atomic Absorption Spectrometry Using Potassium Dichromate Chemical Modification and End-capped Graphite Tubes AGOSTINHO A. ALMEIDA, M. ISABEL CARDOSO AND JOSE� L. F. C. LIMA* CEQUP/Departamento de Quý�mica-Fý�sica, Faculdade de Farma�cia, Rua Aný�bal Cunha, 164–4050 Porto, Portugal Improvements in the determination of aluminium in Port Wine was evaluated as a chemical modifier in the determination of aluminium in blood serum8 and several advantages over other by electrothermal atomic absorption spectrometry, namely in chemical modifiers, including magnesium nitrate, have been terms of precision and sensitivity, are described. Chemical stressed regarding accuracy, precision and reduction of the modification with potassium dichromate was evaluated and reagent blank levels.Moreover, it is reported that sensitivity compared with the performance of the most commonly used increases 2-fold when compared with that obtained with no chemical modifiers.The behaviour of end-capped transverse chemical modification.8 heated graphite tubes (recently commercialized) and the On the other hand, in an attempt to improve the sensitivity standard tubes was compared. The developed method proved to of ETAAS instrumentation with THGAs, end-capped THGA be accurate and an improvement in sensitivity was achieved, tubes (EC-THGAs) were developed,9 which, by lowering the when compared with that obtained without chemical diusive losses of the atomic vapour, have already proved to modification and using standard graphite tubes.be capable of increasing sensitivity by a factor of 30–40% for Keywords: Aluminium determination; Port Wine; some elements.6 This aspect associated with an improved electrothermal atomic absorption spectrometry; potassium repeatability enables lower detection limits to be obtained. dichromate chemical modification; end-capped transverse Based on the aforementioned assumptions, and to improve heated graphite tubes the method of aluminium determination in Port Wine by ETAAS previously described,5 and also to make it more suitable for routine control purposes, the use of potassium The determination of aluminium in Port Wine is still very dichromate as a chemical modifier and EC-THGA tubes was important, both for producers and control institutions, mainly investigated.for three reasons. First, regarding toxicology, several adverse Optimum temperatures were established for ashing and eects of aluminium are known, which has led to an attempt atomization.The linear response range and precision of the to reduce its intake through foodstus.1 Furthermore, from an results as well as the detection limit were assessed. The accuracy enological perspective, the presence of aluminium can produce of the developed methodology was evaluated by comparing a bad taste and the clouding of wines.2 Finally, whereas for the results with those provided by the standard additions other wine types aluminium has not proved to be relevant method.regarding taxonomic classification,3,4 for Port Wine such a Sensitivity was compared using potassium dichromate as possibility cannot be completely excluded. Therefore, laborachemical modifier versus magnesium nitrate, magnesium tories involved in the control of Port Wine production and nitrate+palladium nitrate and no chemical modification.marketing must be provided with prompt and reliable analyt- Using potassium dichromate as chemical modifier, the senical methodologies in order to be able to preserve the quality sitivity and reproducibility obtained with EC-THGA and and authenticity of Port Wine. standard THGA tubes were also assessed. In previous work,5 we have described a methodology for the determination of aluminium in Port Wine by electrothermal EXPERIMENTAL atomic absorption spectrometry (ETAAS) using longitudinal heated graphite tubes and deuterium lamp background correc- Instrumentation tion (a Perkin-Elmer Model 1100B atomic absorption spec- A Perkin-Elmer Model 4100 ZL atomic absorption spectrometer). In order to minimize the matrix eects and therefore trometer with longitudinal Zeeman-eect background correcavoid the use of a chemical modifier, a single high sample tion was used.It was equipped with a THGA and an AS 70 dilution was carried out (1+24). autosampler, also from Perkin-Elmer.All equipment was con- Despite theoretically providing results of better quality, trolled by Perkin-Elmer software installed in a personal com- ETAAS instrumentation equipped with transverse heated puter and the results were printed by an Epson LX-800 printer. graphite atomizers (THGAs) and longitudinal Zeeman-eect The standard THGA tubes (Part. No. B050–4033) and the background correction (e.g., the Perkin-Elmer Model 4100 ZL EC-THGA tubes (Part.No. B300–0644) were from Perkinatomic absorption spectrometer) presents lower sensitivity.6 Elmer. Other instrumental conditions used are summarized Hence, some methodological adjustment, namely a lower in Table 1. dilution of the samples, is necessary and, therefore, the use of a chemical modifier also. Reagents and Solutions Several chemical modifiers have been used in the determination of aluminium, and magnesium nitrate is usually con- Doubly de-ionized water with a conductivity less than 0.1 mS cm-1 was used throughout.Standard solutions were sidered the most suitable.7,8 However, potassium dichromate Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12 (837–840) 837Table 1 Instrumental conditions Source of radiation Al HCL (Perkin-Elmer) operated at 25 mA Wavelength/nm 309.3 Slit-width/nm 0.7 Inert gas Argon at a 250 ml min-1 flow rate* Background correction Longitudinal Zeeman-eect Sample injection volume/ml 10 Measurement mode Integrated absorbance (A s) Integration time/s 5 * Stop flow in the atomization step.prepared by accurate dilution of a 1000 mg l-1 aluminium stock solution (SpectrosoL, BDH). A solution of 0.2% v/v HNO3, prepared from concentrated nitric acid (Merck, Suprapur grade), was used as the diluting solution. The potassium dichromate chemical modifier solution was prepared by Fig. 1 Assessment of the ashing (a) and atomization (b) temperatures dissolving the extrapure reagent (Merck) in doubly de-ionized for both a diluted Port Wine sample and an aluminium standard water.Other chemical modifying solutions tested were presolution. pared from palladium nitrate (Merck) and magnesium nitrate (Merck). Triton X-100 was supplied by Sigma. Comparison Between Potassium Dichromate and Other Matrix Sample Preparation and Calibration Solutions Modifiers Port Wine samples were diluted 1+9 with 1.0 g l-1 potassium The results in Fig. 1 show that potassium dichromate has a dichromate and 0.2% Triton X-100 directly in the cups of the slight eect on the aluminium charring temperature, allowing autosampler.The same procedure was used in the preparation a temperature of 1400 °C, which is higher than that recof the calibration solutions from standard solutions. ommended and checked by us when using magnesium nitrate All glass and plastic material was kept immersed in a or magnesium nitrate+palladium nitrate. On the other hand, solution of 15% v/v HNO3 and rinsed with doubly de-ionized the atomization temperature decreases slightly (about 100 °C) water prior to use.with potassium dichromate. This decreased temperature gap justifies the sensitivity improvement.8 As stated earlier, one of the main purposes of this work was RESULTS AND DISCUSSION to evaluate the increase in sensitivity attainable using potass- Graphite Furnace Programme Optimization ium dichromate as a chemical modifier for Port Wine, compared with no chemical modification or other chemical The programme for the graphite furnace temperature control modifiers usually used in the determination of aluminium.All was optimized using a pool of Port Wine samples and a chemical modifiers tested improved the sensitivity, which was standard solution (500 mg l-1), both diluted 1+9 with the higher by a factor of 2.6 when using 15 mg of magnesium chemical modifying solution (1.0 g l-1 potassium dictrate or 5 mg of magnesium nitrate+3 mg of palladium or by and 0.2% Triton X-100).In order to establish the temperature a factor of 2.8 when using 9 mg of potassium dichromate. control programme used in the subsequent evaluations and Hence, considering that potassium dichromate proved to be determinations (Table 2), a compromise between minimization better in terms of sensitivity improvement and considering also of the programme time and background level and also the the other reported advantages of its use,8 all subsequent work best repeatability of the signals was attempted.Considering was carried out with this chemical modifier. We have verified the analytical signal variation in relation to the ashing and that the blank levels were lower with potassium dichromate atomization temperatures (Fig. 1), 1400 and 2200 °C were and the precision was similar for the three chemical modifiers selected, respectively. tested. Hence, considering the better sensitivity obtained with For the comparison between potassium dichromate and potassium dichromate, it is possible to conclude that the other chemical modifiers (or no chemical modification) the detection limit would also be better with this chemical modifier same instrumental conditions and graphite furnace programme than with the other chemical modifiers.were used, except for the ashing and atomization temperatures. These were 1200 and 2300 °C, respectively. These values have been checked by us as being the optimum ashing and atomiz- Comparison Between EC-THGA and Standard THGA Tubes ation temperatures, in terms of sensitivity achieved, and they are in good agreement with the manufacturer’s recommen- Regarding optimization of the methodology of aluminium dations and with data from the literature5,8 (if we consider determination in Port Wine, another purpose of this work was that instrumentation with THGAs operates at lower nominal to compare the sensitivity and precision provided by the temperatures than previous instrumentation).EC-THGA versus the standard THGA tubes. Using EC-THGA tubes there was a linear relationship between the analytical signal and concentration up to about Table 2 Graphite furnace programme for the determination of aluminium in Port Wine using potassium dichromate as chemical modifier 100 mg l-1. Sensitivity, expressed as characteristic mass (m0), was usually close to 20 pg per 0.0044 A s. The detection limit Step Temperature/°C Ramp/s Hold/s (DL), expressed as the concentration corresponding to three Drying 1 90 1 40 times the standard deviation of the blank solution, was 1.3 mg Drying 2 130 5 40 l-1 (13 mg l-1 in the wine samples).Repeatability was assessed Ashing 1400 10 30 by performing ten replicate injections of a standard solution Atomization 2200 0 4 (500 mg l-1) and a Port Wine sample (660 mg l-1), diluted Cleaning 2400 0 3 as already described (1+9 with diluent/chemical modifier 838 Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12solution, in the autosampler cups). Expressed as RSD, repeat- systematic error was also insignificant (about 0.85% of the mean value of the aluminium concentration in the analysed ability was less than 1.0% for both sample and standard solutions. samples). The results were also compared using the paired Student’s It was found that EC-THGA tubes gave an improvement of all the tested parameters in relation to the standard THGA t-test at a 95% confidence level, which also showed that there were no significant dierences between both sets of results. All tubes which provided a repeatability of 2.0%, a m0 of 36 pg per 0.0044 A s and a DL of 2.8 mg l-1.This represents a parameters of the statistical evaluation of the results are also listed in Table 3. sensitivity improvement of 1.8-fold. The absorbance–time profiles were similar; however, using The reproducibility of the results was assessed by analysing the same set of ten samples in three independent runs.EC-THGA tubes there was a slight delay in the production of the analytical signal (Fig. 2), which can be ascribed to the fact Expressed as RSD, the between-run variation for the concentration of the samples obtained in each run ranged from 0.81 that the end caps represent an increase of 30% in the heated part of the graphite tubes,6 thereby aecting the heating rate to 4.73% (mean=2.70%). in the atomization step.CONCLUSIONS Determination of Aluminium in Port Wine Samples The use of potassium dichromate as chemical modifier for the determination of aluminium in Port Wine presents the same The procedure developed for the determination of aluminium advantages as those obtained when it is used in the determi- in Port Wine, using ETAAS with Zeeman-eect background nation of aluminium in other matrices, namely, very low correction, EC-THGA tubes and potassium dichromate chemireagent blank levels (generally <0.004 A s) and a good repeat- cal modification, was then tested regarding the accuracy and ability (RSD less than 1.0%).The good repeatability associated reproducibility (between-run precision) of the results obtained. with increased sensitivity (2.8 times higher than that obtained Therefore, determinations were carried out by the calibration with no chemical modification) enabled the DL to be lowered graph method (CGM) in ten Port Wine samples, and the to 1.3 mg l-1.results were compared with those obtained by the standard Compared with other commonly used and recommended additions method (SAM) (Table 3). chemical modifiers, potassium dichromate seems to be a good The relationship CGM=a×SAM+b was established by alternative. The sensitivity achieved is better and the blank linear regression, and a good correlation (r=0.991) was levels are lower, which is particularly relevant since the com- observed. The slope of the regression straight line (a=0.991) pared chemical modifiers are specially pure chemicals (and shows that there were no significant proportional systematic therefore fairly expensive).The ashing temperature to which errors and the intercept (b=4.383) shows that the constant samples can be submitted (1400 °C) is also higher than that recommended for magnesium nitrate (1200 °C), which enables the total furnace programme time to be shortened. A lowering of the sample background to levels similar to those obtained with standard solutions was also accomplished, thereby emphasizing the eectiveness of this chemical modifier.Concerning the use of EC-THGA tubes, it was found that these tubes represent a good improvement in graphite furnace technology. The repeatability achieved with these tubes was better than that obtained with standard THGA tubes (about 2-fold better) and the sensitivity was also better (about 2-fold also). Both factors led to a DL of 1.3 mg l-1 using EC-THGA tubes, whereas with standard THGA tubes the DL was 2.8 mg l-1 (13 and 28 mg l-1 in the samples prior to dilution, respectively).Although this aspect may appear to be irrelevant in the Fig. 2 Peak profiles for a Port Wine sample atomized in a standard determination of aluminium in Port Wine samples (as the THGA tube and in an EC-THGA tube. aluminium levels are considerably higher than these DLs), it is important considering that Port Wine is a fairly complex Table 3 Comparison between the results (mg l-1) obtained in the matrix.The improvement in the DL allows higher sample determination of aluminium in Port Wine samples by the calibration dilutions (and therefore the minimization of some matrix graph (CGM) and the standard additions methods (SAM) eects) as well as the injection of smaller sample volumes (10 ml was used). Hence, shorter temperature programmes (an import- Sample CGM SAM RD (%)* ant aspect in routine analysis) and a lower probability of 1 803 795 +1.0 carbon build-up occurring inside the tube, which tends to 2 551 586 -6.0 deteriorate precision, introduce memory eects and prevent 3 332 335 -0.90 4 523 551 -5.1 the passage of light through the tube, are advantages presented 5 555 534 +3.9 by the use of EC-THGA tubes. 6 627 621 +0.97 Moreover, our experience shows that EC-THGA tubes 7 506 516 -1.9 enable more constant sensitivity levels to be obtained through- 8 535 508 +5.3 out their use, whereas using standard THGA tubes sensitivity 9 273 263 +3.8 was aected and decreased significantly.Therefore, frequent 10 416 420 -0.95 recalibrations are not necessary, which is also very important b a R† t0.025‡ t0.025§ in routine analysis. 4.383 0.991 0.991 0.128 2.262 The combination of both aspects studied in this work (performing potassium dichromate chemical modification and * Relative deviation. using EC-THGA tubes) enabled us to accomplish our purpose † Correlation coecient.of optimizing the method of aluminium determination in Port ‡ Calculated values for a paired two-tailed t-test. § Tabulated t-values.10 Wine (previously developed5 by some of the authors for Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12 8392 Eschnauer, H., Jacob, L., Meierer, H., and Neeb, R., Mikrochim. instrumentation with longitudinal heated graphite tubes and Acta, 1989, III, 291. deuterium arc background correction). In particular, it was 3 Martin, G., Zhang, B., Day, M., and Lees, M., Oce International possible to avoid the loss of sensitivity provided by the de la Vigne et du Vin, FV No. 917, 1992. instrumentation with Zeeman-eect background correction 4 Herrero-Latorre, C., and Medina, B., J. Int. Sci. V igne V in, 1990, and transverse heated graphite tubes. 24, 147. The proposed method provides good precision (both in 5 Almeida, A. A., Bastos, M. L., Cardoso, M. I., Ferreira, M. A., terms of repeatability and reproducibility).Accuracy (assessed Lima, J. L. F. C., and Soares, M. E., J. Anal. At. Spectrom., 1992, 7, 1281. by comparison with the SAM) and sensitivity were also 6 Hoenig, M., and Dheere, O., Mikrochim. Acta, 1995, 119, 259. adequate. 7 Slavin, A., Graphite Furnace AAS—A Source Book, Perkin-Elmer, The entire time of the furnace programme (135 s) was lower Ridgefield, 1984. than that previously proposed5 and the sample treatment was 8 Xiao-quan, S., Shen, L., and Zhe-ming, N., J. Anal. At. Spectrom., a simple 1+9 dilution with the chemical modifier solution. 1988, 3, 99. For these reasons, the proposed method, incorporating the 9 Frech, W., and L’vov, B. V., Spectrochim. Acta, Part B, 1993, most recent advances in graphite furnace technology, seems to 48, 1371. 10 McCormick, D., and Roachn, A., Measurement Statistics and be reliable for the determination of aluminium in Port Wine Computation, Wiley, Chichester, 1987. and is particularly suitable for routine laboratory work. Paper 7/01604B The authors acknowledge the financial support obtained Received March 7, 1997 through the CE-AAIR 3-CT 94–2468 project. One of us Accepted April 24, 1997 (A.A.A.) also acknowledges the PRODEP project. REFERENCES 1 World Health Organization, T oxicological Evaluation of Certain Food Additives and Contaminants, WHO Food Additives Series 24, Cambridge University Press, Cambridge, 1989, p. 113. 840 Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12

ISSN:0267-9477    

DOI:10.1039/a701604b

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Automated On-line Preconcentration System for Electrothermal Atomic Absorption Spectrometry for the Determination of Copper and Molybdenum in Sea-water YU-HSIANG SUNG, ZHEN-SHAN LIU† AND SHANG-DA HUANG* Department of Chemistry, National T sing Hua University, Hsinchu, T aiwan 30043, Republic of China A flow injection accessory for electrothermal atomic chelating groups immobilized on copolymer matrices that absorption spectrometry was developed. The performance of retain transition metal ions.After the sample loading step, the the on-line preconcentration system was tested by determining metal ions are stripped from the column with a suitable eluent Cu and Mo in sea-water. Calibration graphs constructed from and directly injected into the nebulizer of a flame atomic the preconcentration of standards in 0.2% HNO3 solution were absorption or a plasma emission spectrometer for detection. used. On-line preconcentration is computer-controlled. A On-line preconcentration systems are better than o-line batch miniature Muromac A-1 resin column was inserted at the tip preconcentration methods, because the former are more of the AS-60 autosampler arm.A modification of the AS-60 ecient, reproducible, easily automated, have low consumption autosampler in the tubing line and circuit allowed either flow of sample and reagent and low risk of contamination.18 On-line of the sample through the column or operation of the flow injection column preconcentration in atomic spectrometry autosampler in the normal mode.Retention of the metal ions was reviewed by Fang et al.19 The on-line column preconcenas complexes on the microcolumn was achieved by using tration systems with liquid delivery forced by air oer a Muromac A-1 as the chelating resin; 20% v/v HNO3 was then number of advantages:20–22 ease of automation; lower conused for elution. With a 198.6 ml sample loop, the throughput sumption of reagent; reduced risk of contamination; and the is 14 samples h-1.Detection limits are 0.009 mg l-1 for Cu peristaltic pump can be used as the air drive. Beinrohr et al.20 (606.9 ml sample loop) and 0.06 mg l-1 for Mo (50 ml sample were the first to introduce air transportation of sample/eluent loop and repeated four times). The accuracy of the method streams obtained from the air support of the atomic absorption was confirmed by the analysis of certified reference saline spectrometer. Azeredo et al.21 introduced the peristaltic pump waters.as the air drive and integrated column preconcentration with ETAAS successfully by using a column packed with quinolin- Keywords: Atomic absorption spectrometry; preconcentration; 8-ol immobilized on silica. Based on solid-phase extraction copper; molybdenum; sea-water with C18 silica gel, Sperling and co-workers23–26 modified the on-line flow injection system of a flame atomic absorption Although electrothermal atomic absorption spectrometry spectrometer to achieve feasible determinations with (ETAAS) has very low detection limits for trace metals in ETAAS.23–26 We reported that trace metals (Cu, Cd, Pb) in aqueous solution,1 the direct determination of trace metals in sea-water could be determined using C18 silica gel and APDC sea-water by ETAAS is dicult even with sophisticated backwith an automated on-line preconcentration system coupled ground correction and chemical modification. This is due to with ETAAS.27,28 Porta et al.12 used dierent materials in the the low concentrations and strong interference from the sample preconcentration column.Hirata and co-workers,9,10 Taylor matrix. ETAAS with on-line sorbent extraction separation and et al.11 and Sung et al.29 used Muromac A-1 chelating resin preconcentration can solve the two problems mentioned above for on-line column preconcentration coupled with ICP-AES,9 and lead to easy determination. flame AAS,10 ICP-MS11 and ETAAS.29 Sorbents used successfully as packing materials for on-line In this work, some commercially available hardware and column preconcentration include chelating ion exchangers,2–17 software components were used to automate the preconcen- C18-bonded silica gel, polymer sorbents, strongly basic anion tration procedure and decrease human intervention. A flow exchangers, strongly acidic cation exchangers, and activated injection accessory for atomic spectrometry was developed; the alumina.rotation of the pump, the stop and go intervals, the actuation Of these, column preconcentration using chelating resins as of the valves, and the time at which the thermal program of packing materials is simpler and less time consuming than the the atomic absorption measurement was started were con- other options. Chelating resins such as Chelex-100,2–6 trolled automatically by an IBM PC-compatible computer. A Muromac A-1,7–11 quinolin-8-ol immobilized on porous miniature Muromac A-1 column was inserted at the tip of the glass,4,12–14 or on silica,15,16 and poly(dithiocarbamate)17 have AS-60 autosampler arm.The functional group of Muromac been used for the enrichment of natural waters and biological A-1 is similar to that of Chelex-100 which contains the materials. iminodiacetic acid[-CH2-N(CH2COOH)2] group. The chelat- A number of on-line chelating resin preconcentration systems ing ability of Muromac A-1 is comparable to that of quinolin- for trace metal determinations have been reported.2–17 In these 8-ol or Chelex-100.9–11,29 Muromac A-1 resin is more highly systems, a controlled volume of sample is passed through a purified and does not swell or shrink.9 A modification of the column containing a cationic resin, a chelating resin, or AS-60 autosampler in the tubing line and circuit allowed either the flow of the sample through the column or the operation † Present address: National Institute of Environmental Analysis, of the autosampler in the normal mode.The retention of the Environmental Protection Administration Government of Republic of China. metal ions in the form of complexes on the microcolumn was Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12 (841–847) 841achieved by using Muromac A-1 as the chelating resin; 20% platform were used. The main reason for using the integrated platform is that it is slightly larger than the standard L’vov plat- v/v HNO3 was then used for the elution.The accuracy of the method was confirmed by the analysis of certified reference form and it is curved, allowing the loading of larger sample volumes, up to 50 ml, without diculty. saline waters. Preconcentration System EXPERIMENTAL The major components and construction of the on-line ion- Reagents and Samples exchange preconcentration system are depicted schematically High-purity water (18 MV cm) was prepared with a de-ionized in Fig. 2(a) and (b). The smallest available sample loop volume water system (Milli-Q, Millipore). Nitric acid (Merck, suprapur in Fig. 2(b) is 135.1 ml. For Mo determination, the required grade) was purified by sub-boiling distillation. A 0.2% v/v sample volume for Mo is only 50 ml; hence, the system in HNO3 solution, prepared from sub-boiled HNO3, was further Fig. 2(b) was modified to that shown in Fig. 2(a) and a 50 ml purified by passing it through the Muromac A-1 column. sample loop was used.Commercial Cu and Mo standards (1000 mg l-1, Merck) were The microcolumn was mounted near the tip of the sampler used. The standard solution (1000 mg l-1) of Mo or Cu was capillary [Fig. 2(a) and (b)]. The solution was delivered by a diluted to the desired concentration with purified HNO3 peristaltic pump (Ismetac, MC-MS/CA4); this is a foursolution (0.2%, v/v), which was also used as the conditioning channel, variable-speed tubing pump. The four channels of the and washing solution for the column in the preconcentration pump were connected as shown in Fig. 2(a) and (b) to draw steps.A sea-water sample used for blank sea-water preparation air, sample, washing solution and elution solution, respectively. was collected from coastal surface water near Hsinchu, Taiwan. With this design, loading the sample, eluting and washing the The sea-water was filtered through a membrane filter column are carried out sequentially with only one variable- (Millipore, 0.45 mm), acidified with HNO3 and stored at 4 °C.speed peristaltic pump. Sea-water reference materials such as SLEW-1 (Estuarine The switch governing the pump rotation speed is computer- Water), CASS-2 (Nearshore Seawater), and NASS-4 (Open controlled. The design is as follows. In the Ismatec instruction Ocean Seawater) were obtained from the Marine Analytical manual, the peristaltic pump oers the user a set pump speed Chemistry Standards Program of the National Research and flow rate either by using the push-button numerical Council of Canada.controller on the front panel of the pump or by using an analog input signal to the pump, which accepts a 0–4.7 V dc, 0-10 V dc, 4–20 mA or 0–20 mA signal from a DB15 female Blank Sea-water Preparation connector on the back panel of the pump. The latter control Blank sea-water was prepared by passing the collected seamethod (using a 4–20 mA signal) is compatible with our water sample through a column packed with the Muromac computerized automated system design.For remote flow con- A-1 resin. The residual Cu and Mo concentrations in the blank trol, an adjustable flow controller was constructed using a sea-water were less than 0.009 and 0.06 mg l-1, respectively. digital potentiometer and three adjustable linearizing resistors (1 kV). Before starting the preconcentration procedure, the desired input current values are established by using a screw- Microcolumn Preparation driver to turn the adjust knob of the three adjustable linearizing The Muromac A-1 microcolumn, shown in Fig. 1, was prepared resistors.The linearizing characteristic of the adjustable linusing a PTFE capillary tube of an AS-40 autosampler earizing resistor versus the input current to the pump permits (2.5 cm×0.94 mm id, Perkin-Elmer), packed with Muromac a linear increment of 0.12 mA per turn, and direct reading of A-1 resin (Muromachi Chemicals, #7 ml, 100–200 mesh for the current from the liquid crystal display (LCD) of the digital Cu; #4 ml, 200–400 mesh for Mo).A resin with a larger mesh potentiometer. Pump speeds used in dierent preconcentration size and smaller volume was used for Mo determination to steps (e.g., washing, sample loading and elution steps) are improve the recovery. Polyethylene frits (porosity 0.5 mm, switch selectable for three choices and controlled by computer. taken from a Sep-Pak C18 cartridge, Waters) were fixed in It is necessary to meter accurately the volumes of the both ends of the microcolumn. solutions used in sample loading, microcolumn washing, microcolumn conditioning and elution.This is eected by means of sampling loops. Three six-port rotary valves [Omnifit, V1–V3 Instrumentation in Fig. 2(a) and (b)] and two seven-port distribution valves An atomic absorption spectrometer (Perkin-Elmer Model [Omnifit, V4 and V5 in Fig. 2(b)] were used to construct the Zeeman 5100 PC), equipped with a graphite furnace sampling loops in the preconcentration system. The lengths of (HGA-600), Zeeman-eect background correction, and a lab- the loops on the valves were chosen so as to obtain injected oratory-built automated on-line preconcentration system were volumes between 50 and 600 ml.The volumes of the loops were used. Pyrolytic graphite coated graphite tubes without a calibrated by weighing. V1–V3 in Fig. 2(a) and (b) were actuplatform were used for Mo determination.For Cu determi- ated pneumatically by connecting the valves to the in-house nation, heated graphite atomizer (HGA) tubes with an inte- high-pressure air line. V4 and V5 in Fig. 2(b) were actuated grated platform as opposed to HGA tubes with a L’vov electrically by connecting the valves to stepper motors. The volumes of the sample loops in V1 (for column conditioning and washing) and V3 (for elution) are 103 and 50 ml, respectively. The volumes of the six sampling loops of V2, V4 and V5 [Fig. 2(b)] for sample loading cover the range 135.1–900 ml. These sampling loops are made with 0.3 and 0.8 mm id Teflon tubing. Three 30 ml poly(propylene) bottles (Nalgene) served as sample (standard or sea-water), washing (0.2% HNO3) and elution (20% HNO3) solution reservoirs. The connections between the pump, valves, column and reservoir bottles are made with 0.3 or 0.5 mm id Teflon tubing (Omnifit) and Fig. 1 Schematic diagram of the microcolumn assembly.chemically inert type fittings (Omnifit). 842 Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12Fig. 2 Automated on-line preconcentration system. An IBM PC-compatible computer was used to control water. The acidified sea-water samples (pH=1.68–1.80) and aqueous standards (pH=1.74) were preconcentrated directly various components of the preconcentration system. The control program was written in TURBO C. The rotation of the without further adjustment of the sample pH.The six preconcentration steps using the automated system processes are pump, stop and go intervals and the configuration of the valves are controlled by switching the dc power supply to as follows. Step 1 for stand-by [Fig. 2(b)]: the PC configures the sample these devices with solid-state relays activated by signals from I/O lines of the PC. The electrical communication between the loops in V1, V2 and V3 to the ‘load’ position such that the washing, sample and elution solutions in the solution reservoirs experimental apparatus (valves, pump, etc.) and the computer was accomplished with an 8255-interface card (Yi Zhong Co.), are drawn through the sampling loops of the valves by the peristaltic pump until the loops are filled.which was installed in the computer. A simplified diagram of the system control circuit is shown in Fig. 3. Step 2 for conditioning the microcolumn [Fig. 2(c)]: the PC configures V1 to the ‘inject’ position, and 0.2% HNO3 solution in the sampling loop of V1 is delivered to the microcolumn by Preconcentration Procedure the peristaltic pump [flow as depicted by the broken lines in Fig. 2(c)]. After all the washing solution (0.2% HNO3; 100 ml ) The flow injection manifold and the sequence of its operation in the loop has been drawn through the microcolumn, V1 is are shown in Fig. 2(a)–(e). The duration and function of each switched to the ‘load’ position. step are shown in Table 1. The conditions used here for the Step 3 for sample loading on the microcolumn [Fig. 2(d)]: preconcentration of Cu and Mo from sea-water with Muromac the PC configures V2 to the ‘inject’ position, and the standard A-1 were similar to those given elsewhere,29 in which Muromac A-1 was also used to preconcentrate Cu and Mo from sea- or sample solution in the sampling loops of V2, V4 and/or V5 Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12 843Fig. 3 Simplified diagram of preconcentration system circuit.Table 1 Operating parameters and sequence of on-line flow injection automated preconcentration system for ETAAS using Muromac A-1 chelating resin Step Fig.* Duration/s Flow rate/ml s-1 Valve position† Purpose 1 (b) 15 5.0 Fill (1), (2), (3) Fill the loops 2 (c) 50 5.0 Inject (1) Condition column Fill (2), (3) 3 (d) [76, 90, 116, 4.5 Inject (2) Load sample 137, 181, 263]‡ Fill (1), (3) 4 (c) 50 5.0 Inject (3) Wash sample matrix Fill (1), (2) in column 5 (e) 55 1.6 Inject (3) Elute analyte into Fill (1), (2) graphite tube 6 (b) 15 5.0 Fill (1), (2), (3) Fill the loops (e) 55 1.6 Inject (3) Elute residual Fill (1), (2) analyte to waste * See Fig. 2. † (1), (2) and (3) are the valves V1, V2 and V3 in Fig. 2(a) and (b), respectively. ‡ Correspond to volume (ml ) of sample loop as follows: 135.1, 198.6, 311.9, 406.9, 606.9, 976.0. is delivered to the microcolumn by the peristaltic pump [flow in Fig. 2(e)], and the euent from the microcolumn containing the trace metals of interest is directed into the graphite tube.as depicted by the broken lines in Fig. 2(d)]. After all the standard or sample solution in the sampling loop has been After all the elution solution (20% HNO3; 50 ml ) in the loop has been drawn through the microcolumn, the position of the delivered to the microcolumn, V2 is switched to the ‘load’ position; the position of V4 and V5 is converted to ‘zero’. AS-60 autosampler arm is switched to ‘waste’ and V3 is switched to the ‘load’ position.Simultaneously, the thermal Step 4 for washing the microcolumn [Fig. 2(c)]: the process is the same as for step 2. measuring cycle of the furnace is initiated by means of the PC. The transient absorbance was recorded and quantified by peak Step 5 for microcolumn elution [Fig. 2(e)]: the PC configures the AS-60 autosampler arm to the ‘inject’ position (the tip of area measurement. Step 6 for cleaning the microcolumn [Fig. 2(b) and (e)]: after the sampler capillary is inserted into the dosing hole of the graphite tube), which is held in place during the introduction step 5 (elution), the PC configures the system as in Fig. 2(b) in order to refill the elution loop of V3 (50 ml ). Next, the PC and retracted at the end of the step. The PC then configures V3 to the ‘inject’ position, and 20% HNO3 solution in the configures the system as in Fig. 2(e) to elute the residual metal retained on the microcolumn to waste.sampling loop of V3 (50 ml ) is delivered to the microcolumn by the peristaltic pump [flow as depicted by the broken lines A complete cycle of preconcentration and eluate introduc- 844 Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12Table 2 Test of the relationship between flow rate versus the input current to the pump Experimental Correlation conditions* Calibration range† Equation coecient A x=4.80–9.20: y=2.07–7.09 y=1.0584x-2.9996 0.9983 B x=4.20–5.80: y=2.60–7.61 y=3.0963x-10.3999 0.9995 C x=4.10–5.00: y=3.10–7.98 y=5.3789x-18.8705 0.9987 * A: Using 0.89 mm id pump tubing; B and C: both using 1.33 mm id pump tubing, but with dierent tubing occlusion.† x: Electric current input to pump (mA); y: flow rate through microcolumn (ml s-1); microcolumn: 100–200 mesh and 7.0 mm length. tion, consisting of six steps, takes 260 s with a sample loading period of 90 s (corresponding to a 198.6 ml volume of the sample loop).RESULTS AND DISCUSSION Test of Flow Rate Control The flow rate was controlled by adjusting the input current to the pump. A test of flow rate control, in terms of the relationship between flow rate versus the input current to the pump, was performed. Table 2 shows that the calibration graphs were reasonably linear within the range of interest. Graphite Furnace Heating Program Fig. 5 Ashing and atomization curves for Mo. An atomization tem- The eects of ashing temperature on the atomic absorbance perature of 2650 °C was used to establish the ashing curve and an and background absorbance are shown in Figs. 4 and 5; the ashing temperature of 1800 °C was used for the atomization curve. A, 50 ml of 20% HNO3 eluate after preconcentration of aqueous standard atomization temperature was 2300 °C for Cu and 2650 °C for (10 mg l-1 Mo) with a 50 ml loop. B, 50 ml of 20% HNO3 eluate after Mo. The sea-water matrix had been removed eectively preconcentration of CASS-2 Sea-water with a 50 ml loop.through the preconcentration steps; therefore, varying the ashing temperature (Cu5300–1400 °C; Mo5300–1800 °C) has little eect on the atomization and background signals. The lapping Zeeman splitting components of the analyte line) and background signals of Cu and Mo remained in the ranges 0.001–0.005, respectively. Ashing temperatures of 1400 and 0.052–0.056 (most of the background signal arises from over- 1800 °C were chosen for Cu and Mo, respectively.For Cu, the eect of atomization temperature on the absorption signal is shown in Fig. 4; the absorption signal remained constant over the range 2300–2600 °C. An atomization temperature of 2300 °C for Cu was selected. For Mo, the eect of atomization temperature on the absorption signal is shown in Fig. 5; the absorption signal of Mo increases with increasing atomization temperature (2000–2650 °C). An atomization temperature of 2650 °C for Mo was chosen for maximum sensitivity.The temperature programs for Cu and Mo determinations are shown in Table 3. Eect of Sample Loading Flow Rate on Relative Recovery The eect of sample loading flow rate on relative recovery was evaluated by extracting the heavy metal ions from an aqueous Fig. 4 Ashing and atomization curves for Cu. An atomization tem- standard at flow rates varying over the range 2.07–5.61 ml s-1 perature of 2300 °C was used to establish the ashing curve and an for Cu and 2.20–5.82 ml s-1 for Mo.The results are shown in ashing temperature of 1400 °C was used for the atomization curve. A, Fig. 6. The data for Cu and Mo were normalized to the values 50 ml of 20% HNO3 eluate after preconcentration of aqueous standard at flow rates of 2.07 and 2.20 ml s-1, respectively. The sample (1 mg l-1 Cu) with a 198.6 ml loop. B, 50 ml of 20% HNO3 eluate after preconcentration of CASS-2 Sea-water with a 198.6 ml loop. loading rate over the range 2.07–5.61 ml s-1 did not aect the Table 3 Graphite furnace temperature program Cu Mo Temperature/°C Ramp/s Hold/s Temperature/°C Ramp/s Hold/s Drying 150 1 70 150 1 70 Ashing 1300 5 20 1600 1 20 Cooling 20 1 15 20 1 15 Atomization 2300 0 5 2650 0 5 Clean-out 2650 1 5 2650 1 5 20 1 5 20 1 5 2650 1 5 2650 1 5 Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12 845in blank sea-water were almost identical (shown in Table 4); hence Cu and Mo in sea-water can be determined using calibration graphs constructed from the preconcentration of standards in 0.2% HNO3 solution.The accuracy of the method was examined by the determination of Cu and Mo in certified reference saline waters (SLEW-1, CASS-2 and NASS-4). Table 5 shows that the method provides analytical results within the ranges of the certified values. Detection Limit and Precision The method detection limits, based on three times the standard deviation of eight replicate measurements of blank sea-water using dierent sample loop volumes, are shown in Table 6.The sample volume for Cu [Fig. 2(b)] is adjusted by selecting Fig. 6 Eect of sample loading flow rate on relative recovery for extracting the heavy metal ions from an aqueous standard (Cu: 2 mg the sample loop. The sample volume for Mo [Fig. 2(e)] is l-1, 134.8 ml loop; Mo: 10 mg l-1, 50 ml loop). adjusted with a multi-injection mode. The multi-injection mode allows the sample solution to be repeatedly loaded and injected on to the microcolumn for the number of times specified by relative recovery significantly.A sample loading flow rate of the user in the control program. The technique improves the 5.0 ml s-1 was selected for Cu and Mo determination. detection limit and the dynamic range of the sample loop. The detection limit of the method decreased with increasing Sea-water Analysis sample volume. The average integrated absorbance obtained with the pre- The slopes of calibration graphs established from the preconconcentration procedure for eight replicate measurements of centration of standards prepared in 0.2% HNO3 solution and CASS-2 (198.6 ml ) is 0.063±0.002, and the relative standard deviation is 3.8%.The results show that this system can Table 4 Calibration graph constructed from preconcentration of stanexecute a series of preconcentration and determination pro- dards (Cu and Mo prepared in 0.2% HNO3 solution and blank seacedures with high precision.The system using a column packed water, respectively) with Muromac A-1 chelating resin (4–7 ml ) can be used for Correlation over 150 preconcentration cycles without any noticeable deteri- Element Equation* coecient Matrix of standard oration in performance. Cu† y=0.0060x+0.00035 0.9987 0.2% HNO3 solution y=0.0060x+0.00033 0.9986 Blank sea-water CONCLUSION Mo‡ y=0.0158x-0.0009 0.9984 0.2% HNO3 solution Although on-line preconcentration systems such as the FIA y=0.0159x-0.0010 0.9988 Blank sea-water 200 (Perkin-Elmer) are commercially available, our inexpensive laboratory-built preconcentration system (which cost US * y and x are integrated absorbances and metal concentrations $2870) coupled with a Muromac A-1 microcolumn performed (mg l-1), respectively.† Cu standard solution (1–2 mg l-1, 135.1–311.9 ml ). well and is fully automated. The laboratory-built preconcen- ‡ Mo standard solution (5–10 mg l-1, 50 ml ). tration system, which is completely computer-controlled, not only enables instrument operation to be preprogrammed rather than performed manually, but also permits rapid reprogram- Table 5 Trace element determination in sea-water reference materials using on-line preconcentration and ETAAS ming when it is necessary to change the preconcentration procedure for the system.Cu/mg l-1 Mo/mg l-1 We thank the National Science Council of the Republic Sample Certified Found*,† Certified Found*,‡ of China for grants (NSC-85–2113-M007–028 and SLEW-1 1.76±0.09 1.82±0.0006 — 4.10±0.08 NSC-84–2621-M007–001ZA) in support of this work.CASS-2 0.675±0.039 0.675±0.007 9.01±0.28 9.05±0.09 NASS-4 0.228±0.011 0.224±0.0017 8.84±0.60 8.29±0.16 REFERENCES * Mean and standard deviation of triplicate runs. † Sea-water sample volume: SLEW-1 (100 ml ), CASS-2 (198.8 ml ), 1 Tsalev, D. L., Slaveykova, V. I., and Mandjukov, P. B., Spectrochim. Acta Rev., 1990, 13, 225. NASS-4 (406.1 ml ). ‡ Sea-water sample volume: SLEW-1 (50 ml ), CASS-2 (50 ml ), 2 Olsen, S., Pessenda, L.C. R., Ru°z¡ ic¡ka, J., and Hansen, E. H., Analyst, 1983, 108, 905. NASS-4 (50 ml ). Table 6 Detection limits Cu Mo Sample volume/ Detection limit/ Enrichment Sample volume/ Detection limit/ Enrichment ml mg l-1 factor* ml mg l-1 factor* 135.1 0.0345 3 50 0.24 1 198.6 0.0218 4 100 0.14 2 311.9 0.0149 6 150 0.08 3 406.9 0.0131 8 200 0.06 4 606.9 0.0088 12 * Compared with direct introduction of 50 ml aqueous solution. 846 Journal of Analytical Atomic Spectrometry, August 1997, Vol. 123 Novikov, E. A., Shpigun, L. K., and. Zolotov, Yu. A., Anal. Chim. 19 Fang, Z.-L., Xu, S., and Tao, G., J. Anal. At. Spectrom., 1996, 11, 1. 20 Beinrohr, E., Cakrt, M., Rapta, M., and Tarapci, P., Fresenius’ Z. Acta, 1990, 230, 157. Anal. Chem., 1989, 335, 1005. 4 Fang, Z.-L., Ru°z¡ ic¡ka, J., and Hansen, E. H., Anal. Chim. Acta, 21 Azeredo, L. C., Sturgeon, R. E., and Curtius, A. J., Spectrochim. 1984, 164, 23.Acta, Part B, 1993, 48, 91. 5 Liu, Y., and Ingle, J. D., Jr., Anal. Chem., 1989, 61, 520. 22 Yang, C.-L., Zhuang, Z.-X., and Yang, P.-Y., Henliang Fenxi, 6 Liu, Y., and Ingle, J. D., Jr., Anal. Chem., 1989, 61, 525. 1993, 91(2), 32. 7 Ikeda, M., Anal. Chim. Acta, 1985, 170, 217. 23 Fang, Z.-L., Sperling, M., and Welz, B., J. Anal. At. Spectrom., 8 Kumamaru, T., Matsuo, H., Okamoto, Y., and Ikeda, M., Anal. 1990, 5, 639. Chim. Acta, 1986, 181, 271. 24 Sperling, M., Yin, X., and Welz, B., J. Anal. At. Spectrom., 1991, 9 Hirata, S., Umezaki, Y., and Ikeda, M., Anal. Chem., 1986, 58, 2602. 6, 295. 10 Hirata, S., Honda, K., and Kumamaru, T., Anal. Chim. Acta, 25 Sperling, M., Yin, X., and Welz, B., J. Anal. At. Spectrom., 1991, 1989, 221, 65. 6, 615. 11 Taylor, D. B., Kingston, H. M., Nogay, D. J., Koller, D., and 26 Sperling, M., Yin, X., and Welz, B., Spectrochim. Acta, Part B, Hutton, R., J. Anal. At. Spectrom., 1996, 11, 187. 1991, 46, 1789. 12 Porta, V., Abollino, O., Mentalti, E., and Sarzanini, C., J. Anal. 27 Liu, Z.-S., and Huang, S.-D., Spectrochim. Acta, Part B, 1995, At. Spectrom., 1991, 6, 119. 50, 197. 13 Malamas, F., Bengtsson, M., and Johansson, G., Anal. Chim. 28 Liu, Z.-S., and Huang, S.-D., Anal. Chim. Acta, 1993, 281, 185. Acta, 1984, 160, 1. 29 Sung, Y.-H., Liu, Z.-S., and Huang, S.-D., Spectrochim. Acta, Part 14 Fang, Z.-L., and Welz, B., J. Anal. At. Spectrom., 1989, 4, 543. B, 1997, in the press. 15 Beauchemin, D., and Berman, S. S., Anal. Chem., 1989, 61, 1857. 16 Yamane, T., Watanabe, K., and Mottola, H. A., Anal. Chim. Acta, Paper 7/01657C 1988, 207, 331. ReceivedMarch 10, 1997 17 Wang, X.-R., and Barnes, R. M., J. Anal. At. Spectrom., 1989, 4, 509. 18 Tyson, J. F., Spectrochim. Acta Rev., 1991, 14, 169. Accepted April 22, 1997 Journal of Analytical Atomic Spectrometry, August 1997, Vol. 12 847

ISSN:0267-9477    

DOI:10.1039/a701657c

出版商:RSC    

年代:1997

数据来源: RSC