JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JUNE 1993 VOL. 8 ATOMIC SPECTROMETRY UPDATE-ATOMIC EMISSION Barry L. Sharp* Chemistry Department Loughborough University of Technology Loughborough LE1 1 3TU Simon Chenery 1 151R SPECTROMETRY Leices te rs hire U K Analytical Geochemistry Group British Geological Survey Keyworth Nottingham UK NE72 5GG Raymond Jowitt British Steel Technical Teesside Laboratories P. 0. Box 1 1 Grangetown Middlesbrough Cleveland UK TS6 6U5 Simon T. Sparkes and Andrew Fisher Department of Environmental Sciences University of Plymouth Drake Circus Plymouth Devon UK PL4 8AA Summary of Contents 1 Arcs Sparks Low-pressure Discharges and Lasers 1 .l. Arcs 1.2. Sparks 1.3. Low-pressure Discharges 1.3.1. Glow discharge lamps 1.3.2. Hollow cathode discharges 1.3.3.Other sources 1.4. Lasers 2 Inductively Coupled Plasmas 2.1. Fundamental Studies 2.2. Sample Introduction 2.2.1. Nebulizers 2.2.2. Flow injection 2.2.3. Chromatography 2.2.4. Electrothermal vaporization 2.2.5. Solid sampling procedures 2.2.6. Chemical vapour generation 2.3.1. Torch and generator design 2.3.2. Spectrometers 2.3.3. Instrument control and chemometrics 2.3. Instrumentation 3 Microwave-induced Plasmas 3.1. Fundamental Studies 3.2 3.3 3.4 Instrumentation Sample Introduction 3.3.1. Direct nebulization 3.3.2. Electrothermal vaporization 3.3.3. Chemical vapour generation 3.3.4. Direct analysis of solids Chromatography 3.4.1. Instrumentation 3.4.2. Gas chromatography-microwave-induced plasma applications 3.4.3. Supercritical fluid chromatography 4 Direct Current Plasmas This review describes developments in all aspects of atomic emission spectrometry including fundamental processes and instrumentation reported in the Atomic Spectrometry Updates References in JAAS Volume 6 (91 /C1688-91/4050) and Volume 7 (92/1-92/1447).The full references names and addresses of authors can be readily found from the Atomic Spectrometry Update References in the relevant issues of JAAS. However as an additional service to readers an abbreviated form of each reference quoted (except those to Conference Proceedings) is given at the end of the review. The most significant developments in emission spectrometry relate to the introduction of commercial instruments that employ two-dimensional solid state detectors (CCDs or CIDs) in place of the conventional photomultiplier tubes.These instruments permit considerable flexibility in wavelength selection allow simultaneous background correction and the use of internal standards to improve precision. Undoubtedly they will be a spur to the routine implementation of more advanced forms of spectral processing which hitherto have been only of academic interest. The early promise of discharges in graphite furnaces remains largely unfulfilled and it seems unlikely that these devices will find the universal application of the established source such as the ICP. As mentioned last year the glow discharge continues to be developed and undoubtedly its simplicity and modest cost have encouraged attempts to extend the range of its application. Solid sampling in general continues to be an active area and there has been renewed interest in laser ablation as a sampling technique for optical spectrometry.152R JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JUNE 1993 VOL.8 1. ARCS SPARKS LOW-PRESSURE DISCHARGES AND LASERS Four reviews of the atomic emission literature have been produced in the last year. Morozov (92/3085) concentrated on regression coupling equations for the analysis of light alloys Busch and Busch (92/3894) collated analytical applications of flame-furnace IR emission spectrometry Chen (92/3976) covered all aspects of AE citing 581 references although in Chinese and Mermet (93/ 1090) reviewed plasmas as a source of photons and ions in chemical analysis 1.1. Arcs The effects of the presence of chlorine in an arc plasma have been calculated by Radic-Peric (92/4625) in respect of B Ca and Si atomic ionic and molecular forms and their spectral line intensities and Zayakina et al.(93/1170 93/117 1) have shown similar reduced emission intensities for Mg Sn Ti and Zn with increasing concentrations of NaCl in the graphite collector. The effects of NaCl on the electrical parameters of an arc burning in a rotating magnetic field have also been studied (92/ 1782). Electro- thermal vaporization of NaCl into a d.c. arc has been achieved using a tungsten spiral (93/703) and the technique has been used to determine Cu with a detection limit of 2 . 7 ~ 10-Io g pl-1 and an RSD of 4.1%. Florian and Terpakova (92/2983) have proved the halogenating ability of CuCl in a study of the evaporation processes of REE in a d.c.arc whilst the influence of discharge media (Ar-N2 Ar-02) on the line intensities of REE has been investigated by You et al. (92/4345). Characteristics of arc AE such as speed simplicity and high sensitivity have been exploited in two specific methods. Xu (93/1031) determined Au in geological ma- terials down to 0.5 ng g-' after separation of the Au by sorption on foamed plastic in a U-shaped glass tube and using NH41 as a spectrochemical carrier for excitation in gas chamber profile electrodes. Detection limits from 2 x lo-' to 5 x for Al Ca Cu Cr Fe Ga In Mg Mn Ni Pb and Si in special purity red phosphorus were achieved by Zolotoreva (93/796). Reported equipment developments have been limited to a study of electrode geometries (93/C82) aimed at designing a hand held spectrometer to provide a more stable d.c.arc excitation for scrap metal analysis and the introduction of a d.c. arc spectrograph with charge injection device (CID) solid state detector (9212055 92/C4180 93/C 178 93/C295). 1.2. Sparks Instrumentation capable of time and spatial resolution has been developed by Lograsso and Coleman (93/C176) to determine number densities of species in a high voltage uni-directional spark. Bye and Scheeline (93/C378) have used similar equipment to make fundamental measure- ments of single discharges and produce electron number density maps and obtained apparent excitation tempera- tures for both analyte and support gas species. They concluded that limitations in the precision of spark source analysis result from the metallurgy of the sample owing in part to the thermal effect of the spark rather than the reproducibility of the plasma and that method develop- ment rather than fundamental research represented the most fruitful development path for spark methods. Stux (93/C995) has followed this road in the use of time-resolved spark analysis to improve the accuracy of metals analysis.Details of a number of spark based analytical methods have been published. Nygaard et a/. (92/3075) correlated results of the well established rotating disc electrode spark technique for oil analysis with those for ICP. Non- equivalence between the techniques was attributed to the diifferent response of each source to viscosity mismatches between samples and calibration standards and suspended particulates.Improved performance over the rotating disc rnethod was claimed by Saba and Byrd (93/1185) who pipetted 5 p1 oil samples onto a paraffin-coated stationary lower electrode eliminated any matrix effect by ashing and then sparked for 20 s. The wide ranging applications of an aerosol-spark technique have been demonstrated by Zhe- leznova and Kuzmenko (93/C958) where sample solution aerosols passed through a channel in the lower electrode after pneumatic or ultrasonic nebulization. Results pre- sented included the concentrations of Al Cd Co Cu Fe IMn Mo and Ni in copper alloys ocean and still waters aqueous solutions of boric acid chloroform extracts and the concentrations of Br Cl and I in solutions of halogen- containing organic substances.A technique for the rapid spark analysis of small solid samples based on pressing specimens into holes in a copper support disc has been developed by Puttman (93/1087). The eternal problem of the rapid determination of acid-soluble A1 in steel has been ,tackled by workers at Thyssen Stahl (92/1918) by the peak integration method. Success was limited by the samples not being homogeneous with respect to A1 and A1203 distribu- tion rather than by deficiencies in the spectrometric method. Spark ablation (SA) coupled with ICP excitation has been used for tin-lead solders (92/252) low-alloyed steel (92/38 19) and ferro-vanadium (92/4627). Each of these publications demonstrated SA-ICP to be a valid analytical technique. Continued development of spark source spectrometry equipment has been described in four presentations (92/C373 1 92/C3737 92/C3789 93/C393).These deal with long-term stability (92/C373 I) determination of nitro- gen in steel (921C3737 93/C393) and digital techniques in the control of source and data acquisition parameters (92/C 3 7 8 9). 1.3. Low-pressure Discharges 1.3.1. Glow discharge lamps Fundamental aspects of the ionization and excitation characteristics of GDLs have been reported by several workers. Horlick (93/C233) used a Fourier transform spectrometer and atomic absorption measurements in a study of neutral atom and ionic species. Straaten et a/. (92/1629) related etching rate to power density and ob- tained good agreement between calculated and measured results at three Ar pressures.They also reported (93/C1328) on the plasma processes occurring in the cathode dark space and negative glow regions. Emissions for three Cu transi- tions were measured by Levy et al. (92/1627) to study charge-exchange processes. Copper ionic emissions were also used by Wagatsuma and Hirokawa (92/1628) to show the effects of He additions to Ar and Ne GD plasmas. Harrison et al. (92/C37 17 92/4598) highlighted the separa- tion of the sputtering and excitation processes within the GD as the phenomenon that separated it from most other spectroscopic sources. Ionization processes in a pulsed GD were also investigated (93/C359). Marcus and co-workers have again made a significant contribution to GD knowledge particularly in the area of r.f. powered GDs. The effects of support gas flow (92/C3424) anode geometry (92/C3425) line selection (921C3459 92/C3727) plasma characteristics (93/C336) the use of Langmuir probes (93/C230 93/C231) and the bulk analysis of high-purity metals and complex alloys (92/C3730 93/C357) were all reported.An r.f GD was used by Kawaguchi et a/. (92/C3526,93/59 1) as both an emissionJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JUNE 1993 VOL. 8 153R and ion source. The effects of r.f. power gas pressure and sample thickness on emission and mass spectra were studied. Heintz and Hieftje (921C3729) characterized the effects of changing the r.f. frequency in terms of emission characteristics and effective power for both conductive and non-conductive samples. A magnetic field was also added (931C358) to increase sample sputtering and modify the glow region in order to improve analyte excitation.Plasma electron temperatures in a magnetron r.f. GD were reported by Heco et al. (92/C3647) together with experimental conditions for quantitative analysis. Background equiva- lent concentrations were compared with conventional GD- AES. Sacks et al. (9211626 9212058 93/C232 93/C337 93/C356) have continued their development of a planar magnetron GD. Cathode current densities of greater than 100 mA cme2 were easily achieved for pressures in the range 0.0007-2.5 Torr (lTorr= 133.322Pa) with a plasma voltage below 500 V (92/1626 92/C232 93/C337). Spati- ally resolved emission spectral data showed the effects of plasma current and pressure on line intensities for the plasma gas and cathodically sputtered species from a pure copper cathode and a zinc-based alloy reference material.Analytical application of the source was demonstrated by the determination of Al Mg Mn and Ni in a series of six zinc-based NIST standards (92/2058 93/C356). Detection limits of 0.006 and 0.00055% were achieved for Mn and Ni respectively . Steers et al. (93/C23,93/C84) have shown that microwave boosting of GDs enhances strongly the resonance lines of elements in the sample. A high-resolution Fourier trans- form spectrometer was used to enable mechanisms such as excitation by charge transfer to be shown by comparison of spectra obtained with different carrier gases. Gas tempera- tures and the concentration of excited states of Ar have been calculated by Leis et al.(93/C234) from the shape of lines emitted by a microwave boosted GD through which a tuned diode laser beam was directed. The laser wavelength was scanned through & 5 half-widths of the Ar line at 8 10.6 nm. The jet-assisted GD has been further investigated by Broekaert et al. (93/C354); increased sputter rates were confirmed and the nature of sample ablation studied with the aid of electron microprobe and X-ray analysis. Self absorption problems associated with high sputter rates and axial viewing have been tackled by Banks and Blades (93/C355) with a source modified to allow side viewing of the jet-induced plasma plume. Park et al. (931828) capital- ized on the high proportion of ground state atoms in ajet- assisted GD and optimized its design as a source for direct solid analysis by AA.Larkins (9211630) used AA measure- ments to detect the reduced number of free atoms produced when 140 ppm of water vapour was present in the Ar used for a GD. Steel aluminium brass and zinc alloys all showed the same effect; the reduction in absorbance varied from 12% for Ni in steel to 77% for Cr in aluminium. The effect was more pronounced at lower sputtering currents. In- creases in sputtering rate for Si and C resulting from H2 additions to Ar have been reported (92/4523). Glow discharges have for some time been regularly used as ion sources for mass spectrometry; most commonly in d.c. form combined with quadrupole MS. Duckworth (93lC1329) has now developed an r.f. powered GD as an ion source for a magnetic sector mass spectrometer to be used for the direct analysis of insulators.A GD has also been used as a detector for GC. Piepmeier and Kizuya (93/C280) found that a GD cell made to oscillate in the frequency range 0.1-10 MHz and operating at 1.8 Torr could be used to detect fmol amounts of analyte. Hieftje and Starn (93/C334) have described an atmospheric sampl- ing GD used as a detector for SFC. A particularly novel GD development has been described by Sacks et al. (92/1625). Ion bombardment heated a graphite cathode to temperatures suitable for the vaporiza- tion of solution residues from the cathode surface. Cylindri- cal post- and hollow-cathode designs gave maximum temperatures of 2500 and 2100 "C respectively for a 250 mA discharge in Ar at 4.0 Torr. Longer vapour residence times were given by the hollow cathode form.The effect of an axial magnetic field on cathode heating was also considered. Quantification of depth profile measurements using GD have been investigated by Nickel et al. (9213864 92/4626). Reference pellets were produced from powder mixtures of copper with Cr Mn and/or Ti and from Cr203 Mn02 and Ti02. Sputtering rate and discharge current corrections were determined using the base metal as reference and the method was applied to the analysis of oxide scales on Ni-Cr alloys. Another major problem in depth profile analysis crater shape has prompted an examination of the influence of anode geometry on the electric field distribution and crater profile of a GDL (93/1019 93/1187). A ceramic spacer was used for the anode tube to restrict the burn spot which together with optimization of the discharge voltage and pressure resulted in an almost flat crater profile.Kliment (93/C884) also determined optimal experimental conditions for depth profiling of different elemental films on metallic samples. Two Japanese steel companies Kawa- saki (92/3122) and Nippon (9213262) have reported on the use of GDs for depth profiling. In the former work the discharge gas supply was cut off and a preliminary dis- charge used to remove extraneous substances before surface analysis was performed. Workers at Nippon (92/3262) confirmed iron oxide film measurements by GD-AES using Auger electron spectroscopy XPS or SIMS. Bulk analysis of a wide range of iron and nickel alloys has been achieved by Glick and Hieftje (93/723) using an artificial neural network and multivariate calibration of GD emission spectra.More conventional narrow range calibra- tions were used in the anlysis of an iron-based alloy (92/1858) and the effects of microstructure were examined (92/1934). Steel analysis by GD-AES was used to illustrate the performance of a photodiode array spectrometer (92/1688). Limits of detection were poorer by an order of magnitude than those reported for multi-channel photomul- tiplier spectrometers. Sputter rate correction based on the Bengtson model ( see J. Anal. At. Spectrum. 1990 5 563) was demonstrated by O'Gram et al. (93/C83) and similar equipment was used by McGeorge et al. (92/C3724) for cast iron analysis. Small samples of powdered steel ( 1 - 100 mg) were analysed by Luft (91/3549) after diluting to more than 200 mg with copper powder at ratios of between 1:20 and 1:lOO; disks 10 mm in diameter and 0.3 mm thick were pressed surrounded by more copper and pressed again.Measurements of Al Cr Fe Mn Mo Nb Ni Si Ti and V were made using copper as an internal standard. A universal powder technique based on specimens mixed and pressed with copper powder was reported by Ehrlick (92/1623). A comparison of GD and spark AES for the analysis of heat-treated steels (92/4572) illustrated the effectiveness of the GD in producing universal calibrations for a wide range of steel types after sputter rate correction. Precision values for the techniques were similar. A novel technique in which a device containing an aqueous sample was moved towards a GD generated with He flowing at atmospheric pressure was reported (92/4485) to provide solvent vaporization charring atomization excitation and emission.Elements determined were Ag Cd Cr Cu Hg K Na Pb and Zn. 1.3.2. Hollow cathode discharges A short review of plasma processes and some analytical applications of hollow cathode discharges has been pro-154R JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JUNE 1993 VOL. 8 duced in Hungary (92/2974). The authors also reported their work on the changes in plasma characteristics caused by EIEs (92/2683). Reduced line intensities were observed for Sr I1 at 430.54 and 407.7 1 nm A1 I at 494.40 and 396.15 nm and He I at 41 2.1 nm. Senofonte et al. (92/2662) have continued their work on the basic phenomena of the microwave boosted hollow cathode by the quantification of mass variations of cathode bottoms and walls over a 2 h operation with and without microwave boost.Overall mass loss was considerably lower with microwave coupling but emission output for Ag Al C Cu Fe and Ni was much increased. Improved performance was also demonstrated by the determination of Cr in steel (92/2973). Detection limit improvement from the region of 1 x 10-4-1 x loe7 to 1 x 10-6-1 x lo-* mass-% by optimization of excitation processes using magnetic fields and pulsed power has been shown by Maximov et al. (93/C975) for metals oxides and semiconductor materials. Hollow anode FANES has been further characterized by Riby et al. (93/C229 92/C246) with an investigation of the effects of cathode temperature and thermionic electrons on emission signals.1.3.3. Other sources The tandem emission techniques FANES and FAPES have been examined by Falk (92/26 17) and their character- istics compared with other techniques such as ICP. He concluded that separation and hence optimization of the atomization stage and excitation process could not be completely achieved; GDs were more prone to matrix influences than exitation sources working at atmospheric pressure; analyte and matrix concentrations in the excita- tion source were relatively high; and tandem sources needed elaborate procedures for analytical application. Dittrich (92/2496) compared FANES with ETAAS and reported lower detection limits for FANES. Demeny and Bernard (93/1188) however obtained similar detection limits for Cd by the two techniques although FANES provided a wider dynamic range of 6 orders of magnitude.A comprehensive investigation of the influence of experi- mental variables on the analytical characteristics of FAPES has been made by Sturgeon et al. (9212750 93/C157). Forward powers of 50 W were used to establish He plasmas at 13.6 27 40 and 54 MHz and Ar plasmas at 27 and 40 MHz. Excitation temperatures for the Ar and He plasmas were similar = 3400 K. Detection limits were independent of frequency whereas sensitivities increased with frequency and were consistently greater in Ar plasmas for Ag Cu Fe Mn Ni and Pb. Further work on the excitation and detection of molecular species was also reported (92/4642 93/C243). Temporal emission behaviour of Ag Al Mn and Pb in a FAPES source were examined by Blades and Hettipathirana (93/C247) and related to simultaneous absorption measurements.A FAPES type source developed by Kitagawa and Katoh (92/4571) used r.f. power to heat a graphite cup located centrally in a grounded stainless-steel cylinder. The sample in solution or powder form was placed into the cup that was heated to a temperature of about 1900 "C resulting in thermal vaporization/atomization of the sample into the reduced pressure Ar plasma surrounding the cup. Analytical capability was demonstrated by the direct determination of Cl Cu and Zn in standard biological samples. A similar device has been proposed by Winefordner et al. (92/1994) in which a microwave plasma surrounded the graphite cup heating it to vaporize the sample into the plasma.Detection limits of 10-20 pg were obtained for Ag Ba Cd Cu Ga Ge In Li Mg Mn Rb and Zn with RSDs of less than 12%. Goldberg and Robinson have continued development of a plasma gun (9212393) which atomizes the sample in a discharge tube and emits a plasma plume created by 6 kV 50 pF discharges. Emission lines of Ag and V were used to characterize the source which was later coupled to ICP and :MIP for excitation of the atomized sample (93lC333 9 3 x 3 7 7). Plasmas for AE have been generated at atmospheric and reduced pressures using a standard d.c. plasma spraying torch with Ar-H mixtures (93/833). A current of 600 A at 50 V with pressures of 100 and 53 kPa produced local thermodynamic equilibrium conditions in what would appear to be a very substantial source.An a.c. plasma utilizing He support gas between two copper electrodes was used by Colon and Barry (92/2407 92/2630) for the determination of 14 elements in aqueous solution intro- duced via a glass frit nebulizer or a thermospray interface. The role of charge transfer in He plasmas was examined by Camanan (93/C264) who noted the intense line emission of non-metal positive ions not observed in Ar plasmas. Huang and Blades (92/2753 93/C335) tabulated the intensities of near-IR atomic and ionic lines of Br C C1 F I N and 0 in an He capacitively-coupled r.f. plasma between parallel plates. Flames have been used as AE sources for the determina- tion of A1 at the ppb level (92/1692) B via diborane generation and emission in a hydrogen diffusion flame (92/ 1709) and K using a sophisticated statistical method for calibration (92/4309).Cresser et al. (92/2455) attempted to overcome deficiencies in the flame source by movement of the burner height with synchronized AE and AA measure- ment to ensure detection of chemical or incomplete atomization interferences. Segmented flow techniques for cool flame emission were investigated by the same group (93/C88). A combined flame-arc high frequency discharge developed by Prudnikov (92/2280) gave improved stability and a three-fold improvement in the detection limits of A1 and Si. The convenience of using a flame emission source for AA measurements has been demonstrated by Calloway and Jones (921C3750) with results for multi-element anlay- sis and internal standardization. Detectors for gas chromatography have been proposed by Wu et al.(92/4619) and by Webster and Boss (93K338). The former work used an He r.f. plasma for selective determination of N and the latter electric field measure- ments of surface wave launched plasmas. A universal ionization/spectral emission detector was developed by Vasnin et al. (92/C3721) which used a pulsed high-voltage discharge in He to ionize all substances including the permanent gases. Hofmann et al. (93/ 107 1) have added a constriction fabricated from molybdenum to a deuterium lamp to intensify the plasma and improve this standard UV light source. Detailed spectroscopic measurements of elec- tron densities and gas temperatures were reported. 1.4.Lasers Selected applications of lasers in atomic spectrometry have been reviewed by Thiem Lee and Sneddon (93/478) who gave 208 references. To further the understanding of laser- material interactions which are fundamental to the use of lasers in spectroscopic analysis Chan and Russo (931 1 184) have used ICP-AES to study the effects of laser power density the pressure pulse generated by expansion of the laser plasma and sample transport characteristics. Reduced self absorption of the Cu I lines 324.7 and 327.4 nm has been demonstrated by Kuzuya and Mikami (92/4564) with reduced Ar pressure in a laser microprobe. A linear calibration for Cu in aluminium was obtained over the range 1-9.8% at a pressure of 150 Torr. High pressure He has also proved effective as a medium for laser-induced plasmas (92/2752) used as an AE source. High energy densities 1 x 10'' W ern- produced by 80 fs laser pulseshave excited X-ray emissions in the 0.7-0.9 nm range from solid targets (9212699). The possibilities for CW laser radiation combined with an arc as an AE source have been established by Toktogonov et al.(93/793). Laser-induced plasma AE methods have been reported for a wide range of materials. Particulate matter in fluids has been analysed in Japan (92/C3555) and France (9213858 92/4021). A time gated multichannel analyser was used by Petit et al. (921C3352) to determine Mg in aluminium alloys over the range 1 - 10% with a 2% RSD. Microanalysis of steel has been performed by Niemax (92/4335) with particular reference to Cr and Si and similar work on alloys and minerals by Pershin (92/3083) included a comparison with arc excitation of the products of LA.Laser atomic emission and LA-ICP-AES were used in the microanalysis of geolo- gical samples (9213849). Grant et a/. studied iron ore analysis (92/ 1690) and produced calibrations for Al Ca Mg Si and Ti with precision values from 2 to 25% and detection limits of the order of 0.01%. Transfer of the technique to on-site application was discussed. Equipment developed for in situ liquid steel analysis has been adapted by Lorenzen and Carlhoff (92/C335 I ) to scan a pulsed laser over rubber slabs to determine spatial element distribution and provide on- line control of compound homogeneity. Contaminants on electronic microcircuits fabricated on A1203 substrates have been analysed by laser emission (931725) with a degree of depth profile information also being obtained.Experimental characterization of photon detection based on pulsed laser enhanced ionization (LEI) and photo- ionization has been carried out by Winefordner’s group (92/4563) using Mg in a miniature air-C2H2 flame. This resonance ionization detector was used to detect weak Raman scatter of CCl. CHC13 and CH3SOCH3. Interference of the OH radical in the detection of trace metals such as Pb in air-C2H2 flames by LEI was observed by Yan et al. (93/676). Electrothermal vaporization into an air-C,H flame has been used for the determination of T1 in water by LEI (931559). A 200-fold preconcentration by non-aqueous extraction yielded a detection limit of 0.043 ng ml-I.High- purity orthophosphoric acid germanium Cd-Hg-Te alloys and silver nitrate have been analysed by LEI combined with flame and graphite furnace atomizers1ionizers (92/4623). For standard aqueous solution both systems gave detection limits down to the 1 x 10-I6 g level but these were imporoved 1 00-fold when solid specimens ofthe Cd-Hg-Te alloys were used owing to reduced background signal. Detection limits of 2 ng ml-’ of Sm were achieved by Zhang (92/3276) using LEI. A number of other novel uses of lasers have been reported. Electrothermal atomization in a graphite furnace into which a pulsed Nd:YAG laser was focused has yielded detection limits of 5 and 50 pg for Co and Cd with an RSD of 5% using AE detection (93/C331). A pulsed ion gun for laser atomic ionization spectrometry was constructed by Kuz’mina et al.(9213232) and achieved a detection limit of for Ga in indium. Geological samples were evapo- rated with a CW laser (931C927) and introduced into a flame or inert gas jet to be transported for analysis by AA or AE using flame arc spark or ICP excitation. The mecha- nisms involved in multiphoton absorption spectroscopy were described by Ashford (92/2597). Laser-induced fluo- rescence has been used by Westblom et al. (92/2687) to detect N in flames whilst Hollberg et al. (931C250) described the characteristics of the rapidly developing diode lasers and their potential applications to a range of spectrometries. 155R JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JUNE 1993 VOL. 8 2. INDUCTIVELY COUPLED PLASMAS 2.1.Fundamental Studies A number of reviews of ICP-AES have been produced in the last year covering the subject in either a general fashion or concentrating on its fundamental properties and analyti- cal abilities. Gerasimov et al. (9213084) reviewed 107 references on the physical properties and production of the plasma as well as its analytical applications with special reference to sample introduction. Sanz-Medel (92/4520) commented on the past present and future of ICP-AES citing six references and also (9213068) assessed the latest developments and expectations in eliminating the current limitations of the technique (28 references). All plasmas whether ICP d.c.-arc or LA are potential sources of photons and ions. Mermet (93/I 090) reviewed excitation mecha- nisms sample introduction and analytical parameters of different plasmas but also (92/C37 18) discussed why ICP-AES still continues to dominate when compared with AAS and ICP-MS and suggested its future lay with CCD and CID detector systems that allow real-time background correction.Dale (92/C4 194) also compared and contrasted the advantages of ICP-AES and ICP-MS and concluded that AES is more versatile for geochemical environmental and metallurgical analysis when major minor and trace element analyses are all required as opposed to ultra-trace analysis. This will no doubt be agreed upon by exponents of both techniques and is the result of the superior ability of AES to handle high levels of matrix. Wagatsuma (9213093) discussed the effect of r.f generator frequency on the spectral characteristics of the ICP and its analytical per- formance with 12 references while Stern (931 12 19) briefly reviewed (3 references) the application of the ICP-AES to environmental monitoring and process control.Hieftje and his group looked forward to where they believe improve- ments in ICP-AES will be made in the future (92/4597). Topics such as overcoming matrix effects automating and minimizing sample preparation adding diagnostics to instrumentation as well as precision were all considered targets for research and development. Yang and Barnes (9212544) have produced a very comprehensive (1 25 references) review of the advances in plasma process modelling and computer simulation since 1984. Blades (92/C3392) specifically reviewed plasma excitation mechanisms and observed that comparisons between models and experimental results are complicated by three types of plasma inhomogeneity spatial temporal and compositional. Zheng et al.(931495) considered the factors influencing matrix ion1electron number density and plasma temperature. The effects of concentration observa- tion height plasma power and carrier gas flow on the continuum from calcium magnesium and aluminium matrices were noted. Results showed that the continuum intensity varied in a complex way with these parameters. Sun et al. (93/789) produced axial profiles of the degree of ionization in plasmas. The effect of plasma power and carrier gas flow rate on the profiles were determined and they concluded that the degree of ionization was not related to the excitation characteristics of the ICP directly and surprisingly the influence of matrix elements was insignifi- cant.Huang and Hieftje (92K4110) used laser light scattering to study fundamental excitation parameters such as electron temperature electron number density and gas kinetic temperature. Galley and Hieftje (93/C26 1) calcu- lated spatial distributions of excited atomic and ionic states using Abel inversion asymmetric Abel inversion and computer tomography. The computer tomography although time consuming and complex was considered to provide the most accurate models. Computer simulation was also used by Cai and Montaser (931C2 10) to predict the effect of156R JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JUNE 1993 VOL.8 He ICP operating conditions on fundamental plasma properties. This was then compared with experimental data. A different approach to measuring fundamental ICP parameters was taken by Strother et al. (9212627) who measured temperature profiles using two colour IR optical pyrometry. Position- and power-dependent temperatures between 1668 and 2543 K were recorded. The spatial dependence of the temperature closely resembled that of the electron number density suggesting that this plays a significant role in determining the thermodynamic temper- ature in an ICP. The importance of the atomization step (occurring on a millisecond time scale) as opposed to the excitation stage (on a nanosecond time scale) has perhaps been underplayed in the past. Mermet observed (931C207) that departures from LTE and matrix effects can be related to parameters influencing the atomization step.These parameters include the size and state of the test material its residence time in the plasma and the efficiency of energy transfer. Some of the most significant work this year on the fundamental pro- cesses of atomization and excitation in the ICP has been described by Olesik and co-workers (9211978 9211979 and 931C206). They demonstrated how the presence of incom- pletely desolvated droplets resulted in changes in the Ca I and Ca I1 emission intensity by factors of 25 and 2.5 respectively. In the region near a solvated droplet Ca I emission was enhanced while Ca I1 was depressed and this directly affected the vertical emission profile. Further to this at the observation height of peak emission intensity (either Ca I or Ca 11) the number of desolvated droplets remained constant regardless of plasma power or injector gas flow rate.The result of these effects was that the time resolved Ca 1:Ca I1 ratio varied by a factor of 70 and was controlled by the fraction of time during which an incom- pletely desolvated droplet was in the observation zone. Farnsworth and Ogilvie (931C204) and Wu and Hieftje (931C258) also studied excitation mechanisms by observing the response of atom and ion lines when large droplets passed through the plasma. In both studies correlation spectroscopy was used to show the relationship between atom and ion line intensities. Wu and Hieftje estimated how droplet size distributions changed vertically in the central channel with respect to plasma power and injector gas flow rate.The measured vaporization rate of particles was compared with that predicted and modelling suggested that particle vaporization was rate limited. Browner et al. (931C259) and Blades and Wier (931C260) both considered the effect of solvent load on fundamental ICP parameters particularly with reference to organics and this has pro- vided useful baseline data. The study of noise in ICP-AES has been of perennial interest and its reduction is the key to improving the quality of analytical data. Snook (931C22) has characterized noise and separated additive and multiplicative components. The dominant noise was multiplicative and resulted from pneumatic nebulization and desolvation processes.Sharp (931C3 931C266) discussed the origins of noise and the influence of the interaction between the sample and analytical system. It was noted that flicker noise cannot be removed by increasing measurement time. Previous noise studies have concentrated on noise above 10 Hz however measurements were made typically on the 1 - 10 s time scale. Both conventional and novel signal processing strategies for reducing the effect of very low frequency noise were discussed. Of practical help Easley et al. (9212633) demon- strated that noise in the 200-500 Hz region could be reduced by the use of a chimney over the plasma. Unfortunately this had Iittle effect on the magnitude of the 11’ noise. Novel fundamental investigations have included the measurement of space-time emission profiles from a pulsed plasma (9311084).A comparison of the performance of various mixed gas plasmas that appear to offer better detection limits has been presented (92/2975) whilst Barnes (931 10 1 1) has described a sealed and static ZCP-AES system for the analysis of gases specifically arsine. A Fourier transform spectrometer has been used to character- ize spectral sources (93/C3394) its rapid acquisition of multi-line intensity data aided the excitation temperature measurements. An AES facility has been added to an ICP- MS instrument by coupling the radiative output with a fibre optic (921 1984) and it was concluded that compromise ICP parameters had to be used. The use of the ICP as a fluorescence source has never really taken off but Ng et al.(931668) have demonstrated that dye lasers are an ideal excitation source for optimal sensitivity. If dual laser double resonance excitation is used the result is highly element specific. The analytical performance of ICP- LEAFS was evaluated for 17 elements using both atomic and ionic lines. Detection limits ranged from 0.2 ppb (Sc) to 364 ppb (W). Freedom from spectral interference can give this technique special advantages for trace element analysis in complex matrices. 2.2. Sample Introduction Borsier (9214726) has conducted a review of sample introduction with respect to the so called hyphenated techniques (spark and LA FI ETV and chromatography) with 12 references. Cassagne et al. (921C4183) have also reviewed the factors limiting the precision of ICP-AES analysis with particular reference to sample introduction.2.2.1. Nebulizers The fundamentals of sample introduction in particular nebulizers have preoccupied Browner and co-workers for many years. They considered (921C3283) that much of the current understanding of nebulizer sample introduction is based on unsound theory and described the importance of a correct understanding of aerosol generation because of its effects on transport parameters. It was felt that optimal aerosol generation would have a significant impact on future developments in ICP/MIP spectrometry. Data were presented linking the droplet size distribution with preci- sion and accuracy. It was suggested (931C183) that for nebulization the key parameters are optimum mean drop size optimum size distribution and the ratio of solvent to vapour.Using the best available data predictions were made on the ideal aerosol and how it might be achieved. Zheng et al. (921C4113 92/C4114) reported on the use of Monte Carlo techniques in nebulization and aerosol tran- sport. They simulated different solution properties operat- ing conditions and nebulizer dimensions to obtain data on nebulization efficiency and total mass transport rate. Simulations of the rate of aerosol droplet evaporation suggested that this has a minimal effect on transport processes and four fifths of the total transported mass is evaporated in the plasma. Luan et al. (921C3286) studied the noise characteristics of various nebulizers and spray chambers using laser light scattering from the aerosol.Power spectra were produced from the scattered radiation by a spectrum analyser. Evidence suggested that pumping was the main source of discrete noise frequencies and that this could be eliminated by the use of pulse-free pumps. The Scott type spray chamber was found to suppress white noise while llf noise was a major cause of instability with ultrasonic nebulization. Ultrasonic nebulizers (USN) have provided much inter- est this year and have continued their revival. A number of new devices with particular characteristics have beenJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JUNE 1993 VOL. 8 157R produced. For example a geyser type made from a low cost humidifier (92/1671) has been described. The USNs are often expensive but equivalent performance to conven- tional designs has been obtained from a device with simplified cooling and reduced production costs (921C3289). An ultra-sonic transducer sprayed with a jet of test solution that allowed increased aerosol production while minimizing wash-out time was described and the fundamentals of aerosol production and performance dis- cussed (921C3290 921C3759 93/C40 1).A multi-stage cryogenic method for removing volatile organic solvents such as methanol and acetonitrile after ultrasonic nebuliza- tion has been reported (9213 105). Several commercial units have been described and evaluated (9212460 921C33 13 921C3809 931C309 92x33 1 1 9213801 92K4184 93/C95 1). Applications of USNs have included the analysis of metals in waters by Bannister and Te Hennepe (92/C1953) where the use of a USN avoided preconcentra- tion a time-consuming and potentially contaminating process.It was concluded that great care was needed with optimization and QA/QC procedures should be used. Interestingly the USN was found to yield different recover- ies of metal present in solution and associated with suspended matter when compared to a pneumatic nebu- lizer. Chan and Geil(921C3803) used a USN to achieve the EPA required detection limits for As Pb Se and TI under a QA/QC regime and reported (92/C3312) that As Cr Mo Se and Sn could all be easily detected in biological samples produced by a veterinary laboratory. They also compared oil samples analysed after acid digestion organic solvent dissolution and ultrasonic emulsification (921C3359).Botto (921C3316) investigated the use of a USN for petroleum and petrochemicals and concluded that the high nebulization efficiency offset the degradation of analytical quality arising from residual solvent loading after desolva- tion. Fukaza et al. (921C3708) found that this technique could reach the detection limits of ETV-AAS but was easier faster and more precise for the determination of As Bi Pb Sb Se Sn and Te in copper after indium coprecipitation. A significant amount of work is now being produced on the thermospray nebulizer. It is hoped that this will result in more efficient coupling of HPLC to ICP-AES as this is an area ripe for exploitation. Coetzee and Robinson (9212554) developed a thermospray nebulizer using a stainless-steel capillary heated spray chamber and a desolvator.This produced an 8-fold increase in sensitivity over conventional nebulization but direct injection resulted in flicker at liquid flow rates above 0.1 ml min-l. Elgersma (9211988) investi- gated a thermospray nebulizer specifically designed for coupling either a micro-HPLC or an FI system to ICP-AES. This used a 50 pm fused silica capillary expansion chamber and desolvator. The system could be optimized for a 120 pl min-l flow rate of a methanol-water mixture (8:2). When using FI carrier flow rates of 100-500 pl min-l could be used very reproducibly. An optimum flow rate of 400 pl min-l for carrier and 80 pl sample injections yeilded ppm- level detection limits for 10 elements. Koropchak et al. (93/C399) used a silica capillary with a diameter of 150 pm but this had a short length of 20 m capillary fused to the end.The characteristics and capabilities of this nebulizer were discussed for a variety of matrices. Peng et al. (931649) found that although the thermospray nebulizer had superior detection limits compared with a Babington type nebulizer it suffered from serious matrix effects. A number of new nebulizers have been described. Ivaldi et al. (92/1641) designed and evaluated a conespray nebu- lizer as proposed by Sharp. Short term precisions of < 1% RSD and long term (8 h) precisions of approximately 1 Yo RSD for both simple and high salt matrices were reported. Sensitivity and detection limits were similar to those produced with a cross-flow nebulizer. It was considered to provide a similar performance to the best of the Babington type nebulizers.Guo and Li (93/5 15) tested the GMK stop- flow nebulizer with high salt solutions and found it successful in avoiding clogging and memory effects. Beres et al. (921C3805) have produced a new cross-flow nebulizer with polished sapphire tips inserted into polyether ether ketone (PEEK) bodies. It was found to resist attack from all common acids and solvents including HF and HC104. This was an interesting use of an old methodology im- proved by the use of new materials. Berndt and Schaldach (92/2086) used hydraulic high-pressure nebulization for both aqueous and organic solvents. Relative signal intensi- ties were compared with those from conventional nebuliza- tion. Todoli et al. (93/C79) also used a high pressure pneumatic nebulizer where gas and liquid streams both passed through the same orifice.The droplet size distribu- tion and detection limits were measured at various gas and liquid flow rates. Fischer and Rademeyer (93/C972) used a heated nebulizer system to determine directly metal concen- trations in waxes greases and fats. Problems and para- meters affecting efficiency and sample transport were considered as were the problems of reference materials and calibration. Spray chambers are often considered the junior partners of a sample introduction system but they make an important contribution. Wu and Fu (931650) produced a multi-function spray chamber that could be used in a variety of modes. The spray path was selectable to obtain the best S/B and stability.No matter which combination of nebulizer and spray chamber is used there is always a certain amount of wasted time while the signal stabilizes prior to analysis and decays afterwards. Brown et al. (93/C2 1 1) suggested that this wasted time could be used if a dual multiplexed sample introduction system was used. This configuration overlaps the data aquisition time of one sampling system with the equilibrium cycle of another. 2.2.2. Flow injection The advantages limitations principles and applications of flow injection methods are discussed in a review by Fan and Fang (92/2253) and again by Fang in another paper (9212105). Flow injection has been used by several workers to effect analyte preconcentration. For example Moss and Salin (93/722 92K3281) used a chelating column and direct sample insertion incorporating a cup specifically designed for liquids. The overall detection limits improved by a factor of 140-1200 for Cu Pb and Zn.Gold at the ng ml-I level in natural waters has been preconcentrated on microcolumns of either Amberlyst-A26 or sulfydryl cotton. Detection was by ICP-AES and ICP-MS (93K27). A new and interesting approach to sample collection and preservation has been described by Gomez et al. (921C3484). Microcolumns of ion-exchange resins were used to collect analyte species at the sampling point and the columns were then returned to the laboratory for elution and quantification. Sulfur in steels has been determined using a microcolumn of activated alumina (9311012). The S was collected as the sulfate ion whilst the iron passed through the column.The detection limit was 0.3 pg g-l. Caroli et al. (9212515) have preconcentrated Cd Co Cu and Pb from waters and urine on a column of iminodiacetic acid-ethyl cellulose and found that the performance of this system compared favourably with other resins. Limits of detection were improved by 1-2 orders of magnitude and 12 samples per hour could be analysed. Iodine as I- and 103- has been preconcentrated by factors of 207 and 15 respectively on a membrane disc containing an anion-exchange resin (92/2396). After preconcentration the I- and 103- were oxidized to I2 to enhance analyte158R JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JUNE 1993 VOL. 8 transport to the plasma. Other methods of FI preconcentra- tion have also been reported.Copper has been preconcen- trated from water samples by FI solvent extraction using dithizone-CCI followed by detection by ICP-AES (9212740). Good agreement with the certified value of NIST SRM 1643 water was obtained. Ten-fold preconcentration was achieved yielding a limit of detection of 0.1 ng ml-I. A novel method of preconcentration using Donnan dialysis ion-exchange membranes has been described where enrich- ment factors of up to 100 were obtainable for a 5 min dialysis period (93/C402). Flow injection has also been combined with HG as a means of sample introduction to an ICP-AES system. Gao and Li (931686) used a mixture of KI ascorbic acid and thiourea to eliminate interferences in the on-line FI-HG- ICP-AES determination of As Bi and Sb in geological samples.Limits of detection were 0.2,O. 1 and 0.2 pg g-' for As Bi and Sb respectively and the analysis rate was 60 samples per hour. Standard additions methods have been used with FI techniques. Shen et al. (92/2738) nebulized sample solu- tions continuously and made discrete injections of aqueous calibrants into the sample stream. Recovery for some REEs was 85-1 12% and precision was 4.2% RSD. A program- mable system for dilutions and standard additions has been described (921C3279). As an application Ca Mg and P in digested plant materials were diluted and determined simultaneously. Standard additions were used to determine Al B Cu Fe Mn and Zn in the digests. Good agreement with certified values was obtained. On-line digestion has been reported by Karanassios et al.(92K3346). A sample slurry and acid were pumped into a coil and the flow was stopped. The coil was then sealed and the enclosed sample exposed to microwave radiation for 2 min. The system was reported to provide rapid precise and quantitative digestion of biological and geological samples. A similar procedure has been described by Gluodenis and Tyson (9214636). These workers used an oven and high- pressure conditions for the on-line digestion of cocoa powder. Dickinson et al. (92/CI 959) have used FI to enable direct determination of Cr Fe Mn and Zn in samples such as 30% sodium chloride and 1 mol 1-* ammonium acetate which have a high dissolved solids content. Zhang and Zeng (93/496) studied the matrix effects of sodium chloride on the determination of 11 elements by FI-ICP-AES.Flow injection reduced signal drift and improved the reproduci- bility. The effects of the NaCl varied with r.f. power and it was concluded that the interferences existed in the evapora- tion-atomization-excitation process of the plasma. Discon- tinuous flow analysis and the advantages it offers over continuous flow has been described by Kimber et al. (921C4 195). 2.2.3. Chroma tograph y There has been a large increase in the number of reports using ICP-AES as a detector for chromatography. Howard and Hunt (92/4729) reviewed (14 references) the use of AAS AES and AFS coupled with GC SFC and LC. The problems of interfacing were discussed. Liquid chromato- graphy remains the most popular of the separation tech- niques for coupling with ICP-AES.A chelating resin was used to preconcentrate analytes selectively from matrices such as sea-water whilst background and spectral interfer- ences were reduced (93K238). Porta et al. (9213821) also used a chelating resin (XAD-2 functionalized with 1-(2-thi- azoylazo)-2-naphthoI to preconcentrate Cd Cu Fe Mn Ni and Zn from waters. Limits of detection were 2-40 ng 1-1 and precision was less than 5% RSD. Lanthanides in rock samples have been separated using a cation-exchange resin and ammonium lactate or a-hydroxyisobutyric acid in the mobile phase and detected simultaneously by Sawatari et a/. (931588). In a similar paper the workers determined 15 Nanthanides in less than 40 min per run and obtained detection limits of 0.4-30 ppb (921C3559). Chromium Mn Mo and Ni have been determined in steel samples with results close to the certified values for all except Cr (93/434).The REE were determined in terbium and terbium oxide using a reversed-phase column and an eluent of 2-hydroxyisobutyric acid with recoveries in the range 85-1 00% (9213045). Watkins and Nolan (9213852) deter- mined Hf Sc and Y in geological samples using a cation- exchange column. Limits of detection were 0.2 0.1 and 0.2 eg g-l,. respectively. Boron has been determined in iron and iron disilicide by Yamada et a/. (92/44 17) using an anion- exchange column. Good agreement with certified values was obtained and the limit of detection was 0.05 pg g-' in the iron sample. Metallothioneins have been separated successfully and characterized in biological (9318 52) and marine samples (9212556).Liquid or ion chromatography has been used extensively for speciation studies. Carney et a/. (931C340) used HPLC-ICP-AES to separate butyltin trichloride from tin tetrachloride. Derivatization was not necessary and as an application industrial air samples were monitored. Vanadi- um(v) and VIv species have been determined and precon- centrated from natural waters using a two column system (921260 1 ). Immobilized silica gel bonded with ethylenedi- amine was used to separate Vv and silica gel bonded with ethylenediaminetriacetate was used to collect both Vv and VIv. Recoveries of 9 1-105% and enrichment factors of 40 were obtained. Sulfur speciation has been achieved by Shan et al. (92/C4 124). For the determination of total S in soils a digestion using HN03-HClO under pressure was rec- ommended. The speciation consisted of a sequential extrac- tion procedure in which water soluble adsorbed dilute HCl volatile and soluble pyrite HI reducible ester and carbon- bonded S were determined.Sulfur speciation has also been reported by Halmos and Borszeki (921C4230). Arsenic speciation has again been reported by Rauret et al. (92/2007). Arsenic(rIr) methylarsonic acid dimethylarsinic acid and AsV were eluted from an anion-exchange column using a phosphate mobile phase. The eluate passed into a hydride generator and the hydrides evolved were deter- mined by ICP-AES. Limits of detection were 3.5 3.8 21.3 and 9.2 pg l - l respectively. Recoveries were close to 100% for all but As1[' which was 83.5%.Wiederin (921C3470) separated and quantified various inorganic forms of As Cr and Se using a microcolumn and a direct injection nebulizer. Detection limits were reported to be in the low- ppb range. An ultrasonic nebulizer and various tetraalky- lammonium salts as ion-pairing reagents have been used for the speciation of the same analytes (9212839). Again detection limits were in the low-ppb range. The interface between a microcolumn used for speciation and an ICP- AES detector has been studied by Jinno (921C3465). It was concluded that the column should be coupled directly to the torch to achieve maximum sensitivity and repeatability and to avoid dilution effects produced by the spray chamber. A similar interface was found to be oecessary for coupling SFC with ICP-AES.Laborda et a/. (9211989) also made a comparison of interfaces for speciation purposes. A ther- mospray nebulizer was found to have three times the sensitivity and to be three times more tolerant of organic solvents than a cross-flow nebulizer. Initial results for the speciation of Se as (CH3)3Se+ Se03*- and Se0,2- were promising. The ICP is less commonly used as a detector for other chromatographic techniques. Kato et al. (9213820) have determined methylmercury species by capillary column GC using an axially viewed ICP-AES system with an echelleJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JUNE 1993 VOL. 8 159R monochromator as a detector. The methylmercury species were converted to the iodide form and separated on a chemical bonded fused silica capillary column coated with methylsilicone.The detection limit was 3 pg and the linear range covered 3 orders of magnitude. Gas chroma- tography-ICP-AES has also been used to study the uptake of 17 elements from sea-water by oysters (921C3573). Size exclusion chromatography has been used successfully to study Fe concentrations as a function of molecular size in bitumens (9311142). A thorough review of packed micro- column SFC coupled with ICP-AES has been presented by Jinno (931610). The problems in the coupling of these two techniques were discussed and overcome. 2.2.4. Electrothermal vaporization A review containing 54 references on the analysis of biological materials by ICP and MIP using thermal vapori- zation as a means of sample introduction has been presented by Matusiewicz (92/4284).Huang et al. (93/68 1) have studied the vaporization mechanisms of analytes of different volatilities. Effects such as vaporizing current and the volume and geometry of the atomizer were studied. Matousek and Mermet (921C3401) have used ETV to determine the effects on the plasma of hydrogen that does not originate from water. Huang et al. (92/2735) studied the effects of transport efficiency between the vaporizer and the ICP. Parameters such as tube material length temperature and diameter were all studied. In a similar paper the mechanism of analyte condensation in the transport pro- cess was explored (93/ 1 123) and non-rare earth impurities in lanthanum oxide were enriched by diethyldithiocarba- mate-CC1 extraction and determined by this technique.The same authors also developed a device that prevented sample expansion between the vaporizer and the ICP (921C4128). As a result increased signal intensity was achieved. The authors also concluded that platform vapori- zation was not suitable for ETV-ICP-AES. The use of halogen-containing compounds to vaporize analytes into an ICP is still receiving a lot of attention with solid liquid and gaseous reagents all being used extensively. The most common solid fluorinating agent used is PTFE. Huang et al. (9212486) used a PTFE slurry to aid the vaporization of carbide forming analytes such as the REE. Detection limits were two orders of magnitude better when fluorination was used and memory effects were negligible. Hu et al. (9211807 9212019 9212615) have produced several papers in this field.Chromium in serum with a detection limit of 1.4 ng ml-l (92/1807) B in plant leaves by a standard additions method with no interferences from Ca K Mg or Na (9212019) and Mo in food slurries with validation by the use of CRMs (92/2615) have been determined. Both solid and gaseous halocarbons have been used by Jian et al. (921C33 18) to volatilize away the matrix. The system used a modified ETV-ICP-AES system that enabled both solid and liquid samples to be analysed. Two types of interface were compared by Kantor (92/4622) using cadmium as the best analyte. Matousek and Wu (921C 1944) have used CC14 and CF3CH20H to volatilize V and W. Memory effects were completely eliminated and the limits of detection were improved over the use of argon-chlorine mixtures. Tungsten filaments continue to be used as electrothermal vaporizers.Fujimoto et al. (921C3584 and 93/592) com- pared direct sample insertion (DSI) using Freon-assisted volatilization with ETV. It was found that DSI was more resistant to oxidative attack but that ETV gave better detection limits. Graphite cup DSI devices have been used successfully by Abdullah et al. (92/C4 199) who managed to obtain superior LODs and precision than with conventional nebulization. Preconcentration of analytes from sea-water and analysis of both solid and liquid matrices was also achieved. Umemoto and Kubota (9211998) compared DSI with normal nebuli- zation. The DSI plasma was found to have a different structure yield a slightly reduced linear range and an electron density 4.5 times lower than when conventional nebulization was used.Background intensities were found to be inversely proportional to the thickness of the graphite cup with a cup of wall thickness 0.5 mm producing an intensity 1.7 times lower than conventional nebulization. Umemoto and Haraguchi (921C3591) used DSI to deter- mine Cu and Fe in lead and zinc metals. Detection limits were 5 ppb for Cu and 20 ppb for Fe in a 2 mg sample. Results for the analysis of CRMs were in good agreement with certificate values. Bir and Rybarczyk (92/C3301) have described a FAPES system constructed from commercially available compo- nents. The performance in terms of detection limits and interferences were discussed. Liang et al. (92/C38 14) have again reported a graphite furnace plasma source for atomic spectrometry with the developments anticipated previ- ously (see J.Anal. At. Spectrom. 1992 7 165R) described. The salt-induced matrix effects in a niobium carbide coated ETV-ICP-AES has been reported by Soman and Gilbert (92lC3293). Transport efficiency was found to be affected by a sea-water matrix. Numerous applications of ETV-ICP-AES have been reported. Mikasa et al. (92/C3609 and 92/4722) determined B in silicon and silica. After dissolution of the sample B was extracted into toluene as the tetrafluoroborate ion associated with Ethyl Violet. The B was then determined by ETV-ICP-AES giving a detection limit of 1 ng ml-I. Another interesting way of determining B in pure iron was described by Yoshikawa and Funabiki (93/C50).They reacted B with methanol and sulfuric acid in a pyrolytic carbon coated tube and determined the evolved methyl borate. Wear metals in lubricating oils have been deter- mined and the results compared with those obtained by rotrode emission spectroscopy (92K33 1 5). Plutonium in urine (931C30) and Cu in pure iron reference materials have also been determined (9213260). In this latter paper 25 analytes were measured simultaneously but precision was poor. This was attributed to the optimum heating pro- gramme being different for each analyte. A method of preconcentration by repeated injection and desolvation has been described by Zhang et al. (931697). Cadmium Pb and Zn have been determined in botanical samples using pelletized solids (92/2404 and 93/C398).The sample was mixed with graphite and pressed into a pellet. The pellet was then heated resistively and the vapour swept to the plasma by a flow of Ar. Detection limits were I 0.6 and 3 ppb for Cd Pb and Zn respectively. 2.2.5. Solid Sampling Procedures Dean (9211460) provided the sole general review of solid sample introduction devices along with those for liquid and gaseous introduction (22 references). Van Loon et al. (921C4126) appraised direct injection techniques such as probe ETV and slurry nebulization for different types of geological industrial and biological test material. Their important conclusion was not surprisingly that no one method was suitable for all sample types. Zaray et al. (92/2507) observed that when comparing direct and indi- rect analysis of trace elements in pure alumina the most suitable solid sampling method depended on the physical properties of the Al,03 such as the mean grain size.Chenery (92/C4080) when reviewing his work on small scale solid sample introduction stressed the importance of consider- ing absolute instrumental detection limits rather than the more conventional concentration based limits.160R JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JUNE 1993 VOL. 8 A small but significant increase of activity in laser ablation for ICP-AES has occurred this year but few new workers appear to be entering the field. Perhaps this will change if the explosion of interest in LA for ICP-MS crosses back to ICP-AES. Unfortunately the cost of an LA system is still relatively high when compared with the cost of an ICP-AES.The use of LA was reviewed ( I 1 references) by Sneddon (92/4472) while Monke-Blankenburg reviewed her own extensive work (921C4107). In this field funda- mental and methodological work is of great importance. Thus Chan and Russo (9212785) monitored the laser- material interaction using spatial and temporal ICP-AES response. Observations on the effect of laser power density condition of sample surface pressure pulses and sample transport were all reported. This work was further devel- oped (931C352) by comparing the effect of the power density of the laser pulse using excimer and Nd:YAG lasers of similar wavelength but different pulse lengths (20 ns and 30 ps respectively). The results were assessed in terms of mass of material ablated and elemental ratios from test materials as measured by the ICP-AES. Mermet et al.(93/C92 1) have investigated the effects of laser parameters such as wavelength (Nd:YAG 1064 532 266 nm and excimer 308 nm) mode (normal and Q-switched) energy density and spot size. Both the amount of material ablated and the ICP-AES response were investigated. The influence of target type (metals and glasses) surface state and transmittance was also observed. They concluded the current lack of understanding does not preclude the use of LA because of its successful performance. Greenhill (9113404) also varied the wavelength of a laser system using harmonic generating crystals from the near-IR to the UV to optimize ablation of various test materials including ores aluminium alloys and carbon based samples.Monke- Blankenburg (92/C3353) reported that transient ICP-AES signals from single shot LA can be described in time by mathematics analogous to FI. Results from models were compared with those found experimentally and new methods of quantification were evaluated. Furuta (92/2634) when analysing pond sediment powder pellets described the effects of transport tubing on emission. It was observed not surprisingly that a 30% variation in response could be reduced to 4% by the use of an Fe internal standard and translation of the sample. Various new analytical methodologies have been pro- posed. When the ablation cell is remote from the ICP-AES by several metres a loss of test material is observed. Lui and Horlick (921C3373) tried to resolve the problem by three alternative methods placing the sample immediately below the ICP torch; placing the sample in the ICP torch on a graphite rod; and ablating material remotely onto a graphite ring and subsequently introducing this into the torch.Iada et al. (9212406) have ablated material under water such that all the material was captured. This slurry was then either nebulized directly or first dissolved. This methodology also allowed filtration of material for observa- tion by SEM. Lin et al. (921 1790) have used a continuous CO laser to ablate totally small samples of test material thus overcoming any matrix effects from phase separation. A detection limit of 1 ppm of Yb was obtained in pressed powder pellets of REE ores and graphite binder.This laser was also used for the ablation of liquid droplets and detection limits twenty times better than solution nebuliza- tion were obtained for Ag Cd and Yb on 20 pl samples. Jowitt and Whiteside (931 1 106) described the analysis of molten steel by observing the real-time emission of the ablation microplasma or by carrier gas transport of the ablated material to an ICP-AES instrument. Applications of LA sampling have continued to be dominated by geochemistry. Li and Duan (9213946) ana- lysed ion-exchange paper on which REE from rocks had been deposited. Thompson et al. (93K38) investigated the ferro-manganese oxide coatings on stream pebbles and trace elements concentrations delineated known copper and gold deposits in North Wales. Ramsey et al.(9311000) have now validated their analyses of large fluid inclusions (> 30 pm) in topaz and halite. Calibration was performed by using sensitivities obtained from aqueous nebulization. Validation was by comparison with qualitative SEM semi-quantitative crushed leach ICP-AES and quantitative synchrotron XRF. Monke-Blankenburg and Gunther (92/3849) used their solid-liquid method of calibration and a Ti internal standard for the determination of La in pseudo-brookite. Fomenkov (921C3376) choose to analyse trace elements in LiF crystals by LA in preference to nebulization because of the low solubility of some fluorides. Slurry nebulization although an established technique for the introduction of solid samples in ICP-AES failed to make a significant impact in the literature.This is partially explained in a review of the advantages and disadvantages of the technique by Jarvis (92/3848) who suggested the major problem is poorer precision compared with solution nebulization. Quantification of the problems was provided by Laird et al. (9114036) (see also Ebdon,L. and Collier AX. Spectrochim. Acta Part B 1988 43 355) who centrifuged suspension of clays into different size fractions. These size fractions were either nebulized directly or after dissolution. Recoveries from 2- 100% were observed depending on size fraction and ICP parameters. For size fractions less than 2 pm recoveries better than 90% were obtained and these fractions also gave reproducible analy- ses. Larger size fractions gave poor recoveries. Both particle size and composition limited the quantitative analysis of Ti02 and A1203 carried out by Broekaert (9212494 and 93/C376).Solutions to the problems of slurry nebulization will only be found after fundamental studies such as that of Ebdon and Foulkes (92/2499). They measured the rotation temperature T of the ICP using the zero-zero vibrational band of the OH radical while introducing aqueous solu- tions humidified argon and aqueous slurries of A1203 or BN. The measured temperatures varied between 2200-3600 K depending on ICP parameters. Somewhat surprisingly no significant decrease in T,o was observed when solutions or slurries up to 1% by mass were intro- duced. Isozaki (931565) optimized the ICP-AES conditions for the determination of Fe in silicon nitride. Ohls et al.(921C3292) have used activated carbon and a complexing agent to separate trace elements from a matrix. This material was then dried and milled before nebulization. Calibration standards were produced by the same method. Both Carrion et al. (9212642) and Liu and Li (921C4146) analysed bio-materials for trace elements by slurry nebuli- zation using aqueous calibration the former validated their work with reference materials. Gomez-Coedo et al. (92138 19) compared spark source (SS)-ICP-AES with aqueous nebulization-ICP-AES and SS- AES for the analysis of low-alloy steel reference materials. The objective criteria studied included detection limit S/B and repeat ability. Spark sou rce-I CP-A ES was subsequently used for the analysis of ferro-vanadium (9214627). The test material was diluted with pure iron to avoid fracturing and Al Cu Si and V were determined with Fe used as an internal standard to improve precision. Reference materials were used to validate accuracy.Fengdi and Mingjun (931C969) also evaluated this technique using reference steels. Interference corrections were calculated by ablating high-purity metals. The advantage of the speed of the method over aqueous nebulization was discussed. Vujicic and Steffan (9212 1 14 and 931525) analysed conducting solids directly by spark ablation and also non-conducting geological materials by placing them in a graphite cup or briquetting them with graphite and wax (for strength).JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JUNE 1993 VOL. 8 161R Oxygen (10%) was mixed with the outer gas and the background equivalent concentrations observed were better than those from aqueous nebulization.A cyclone chamber placed between the SS and the ICP improved repeatability and reduced deposition in the torch. Novel SS procedures have included (93/1254) a spark discharge to molten steel to produce ultra-fine particles for analysis. The discharge plasma observation height and particle content of the carrier gas were studied with respect to the precision of the analysis. Lin and Xu (9212240) loaded powder into a graphite electrode and then arced this directly in the plasma. This appeared to give better S/B than direct powder introduction. Elias-Eljuri et al. (921C3370) determined P in bronze by using an r.f. spark under water and then nebulizing the dispersion.This allowed aqueous calibration and the method was validated with reference materials. Direct sample insertion continued to receive some inter- est despite known matrix effects associated with the differential atomization efficiencies of elements. Umemoto et al. (9211998) compared the spectral characteristics of a plasma with aqueous nebulization and alternatively graph- ite cup insertion. Key features noted included a lower electron number density and lower background intensities for direct sample insertion. The intensity ratio of ion to atom lines was observed to change for some elements and this was attributed to cup position. They went on to insert cups containing pieces of lead and zinc metal into the plasma (93/805). Matrix evaporation occurred before atomization of the Cu and Fe analytes.Time-resolved integration allowed minimization of matrix eflects and interferences. Calibration was performed using standard solutions; the method was validated using reference materials and the RSD of the method was 5-20%. To overcome some of the problems of this method Blain and Salin (931534 93/C2 12) investigated probe design. Sedi- ment reference materials (MESS-1 and PACS-1 from the National Research Council of Canada) were inserted for analysis. The use of internal standards and standard additions allowed accurate calibration for elements of high and intermediate volatility (Cd Cu Hg Mn Pb and Zn) and sub-ppm detection limits for these elements were obtained. Roberts and Snook (93/C81) have made initial studies on the determination of Mn in terephthalic acid resulting in linear calibrations of peak area versus sample mass and RSDs of 6-10%.Direct powder insertion research has been led by Guevre- mont and De Silva (9211 99 1,92/2786,93/538 92/C332 l) starting with the premise that errors can occur in direct powder introduction if there is a size bias in either the introduction system or in the efficiency of vaporization/ excitation in the ICP. The magnitude of the bias from a fluidized bed system was then estimated by labelling two sizes of silica particles with different elements. The chemi- cal labels on the particles from the carrier stream were then removed by acid extraction. A 50+ 50 mixture of 2 I and 34 pm particles became a 65+35 mix after transport (921 199 1).Subsequently dry Chelex resin was introduced containing spikes of elements from I 1 to 1470 ppm. Element ratios gave precisions of ( 5 % . However when the Chelex resin was mixed with a complex test materials i.e. geological reference materials the precision of element ratios degraded to between 2 and 16% (92/2786,92/C332 1). Finally a new device for introducing powders was de- scribed (93/538). Lin ef al. (93/675) increased elemental response and excitation temperature in the plasma when directly introducing powders by applying an auxiliary arc discharge to the plasma thus moving the optimum observa- tion zone higher into the plasma. Liu devised a new method of directly introducing powder particles (92/C4 1 32) by vibrating the carrier gas. The carrier gas was driven by a diaphragm allowing variation of the hydrodynamic field- induced forces and therefore sufficient control of the rate of particle introduction and residence time in the plasma to optimize the analytical conditions.Meyer (92/C3360) sub- sampled the gas stream from a commercial spray drier used to make fine powders and introduced this stream directly into an ICP-AES system and thus avoided the time- consuming process of redissolving the powders when determining metals. 2.2.6. Chemical vapour generation Not surprisingly research into chemical vapour generation as a means of sample introduction into an ICP has focused mainly on the hydride forming elements. There have been several reviews in this area. Zhang et al. (92/2257) have reviewed (55 references) chemical interferences in HG. Several mechanisms were proposed and discussed.The applications of HG techniques to ICP-AES have been reviewed (4 1 references) by Nakahara (92/3889). A critical survey of HG techniques in atomic spectrometry containing 134 references has been presented by Campbell (93/641). A hydride generator made in-house using a peristaltic pump was described by Brooks (92/163 I). Although no gas-liquid separator was required and the response for Se was linear up to 1000 ,ug I - I boron contamination of the torch was problematic. The use of both continuous and discrete hydride generators was described by Steffan and Vujicic (92/209 1). The advantages of using HG as a means of sample introduction were discussed. Several papers have described the presentation of hydrides directly to the nebulizer of an ICP without the need for a normal gas-liquid separator.Li et al. (93/C240) used on-line reduction followed by HG to determine As Sb and Se. The procedure was relatively interference free although copper in excess of 40 pg ml-1 was problematic. Certified reference materials were used to validate the technique. Noelte (92/2644) evaluated the effects of parameters such as pH stoichiometry reaction time and sample introduction into the plasma on the hydride-forming elements and Hg signal intensity. Arsenic Bi Sb Se and Sn have been determined in biological and environmental CRMs with good agree- ment with the certified values (92/2003). Arsenic has been determined in a variety of samples with a detection limit of 4 pg kg-' and recoveries of 99-104% (92/1764). Arsenic and Se levels in mussel and shrimp CRMs (93/680) and As Se and Sb in waters (93/769) have also been determined.Ozaki (92/C3319) determined As Bi and Sb in steels and nickel alloys using high concentrations of HCl and Fell1 to overcome the interference of nickel. Matrix interferences have been eliminated in the determination of Bi in waste samples by the use of a chelating resin to remove ions of Fe and Cu (92/4732). Qiu et al. (92/C4131) noted that the hydrides of the different elements were formed at slightly different times in the same solution. An alternative method of chemical vapour generation has been reported by Kantor and Zaray (92lC4214) in which two versions of halogenation-vaporization were described.The first used a graphite furnace and CCl at 1800 "C to distil trace constituents from samples such as alumina silicon carbide and silicon nitride into an ICP. The second version utilized a laboratory constructed quartz furnace equipped with either a condensation or chemical absorp- tion device. In this version a cold finger condenser was used if the analyte was more volatile than the matrix or an absorption bottle was used if the matrix was more volatile than the analyte. Sulfur has been determined in gallium phosphide crystal (92jC3767). After dissolution of the crystal the S content was determined by reducing sulfate to H2S quantitatively using NaI-HI-H,PO at 130 "C. Sodium sulfate was used to calibrate the procedure. Sulfur contents of 2-10 pg g-' in162R JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JUNE 1993 VOL.8 gallium phosphide was determined with a precision of 1-3% The results obtained were in good agreement with those obtained by GDMS. 2.3. Instrumentation 2.3.1. Torch and generator design This has been a relatively quiet area during this review period. A review with seven references discussing the advantages of self-oscillating ICP generators has been presented by Fono et al. (92/2542). A new ICP with a ‘direct serial coupled’ r.f. generator has been developed (93K288) and has been found to be both more efficient and more stable than previous designs. This has the advantage that organic solvents can be run more easily. Data were presented for Pb and S in gasoline andLODs for wear metals in oils diluted with a variety of organic solvents were also given.Ishii et al. (92/198 I ) have modified a 27.1 MHz generator to produce one capable of forming He ICPs at 6.8 27.1 and 40.7 MHz at powers of up to 2.5 kW. Detection limits for Br C1 I and S spectroscopic temperatures and electron number densities were measured. The results were compared with those obtained from a 27.1 MHz Ar ICP operating at 1.1 kW. In addition a reduced pressure He ICP operating at 27.1 MHz was investigated and its perform- ance compared with the atmospheric pressure ICP. Gamage et al. (93/C992) have described a new solid state generator which has a novel method of power control. This has been used with an axially viewed ICP and a Paschen Runge polychromator to show that gas flows nebulizer spray chamber drain and sample presentation have less effect on precision than they do in a radially viewed ICP.Chan and Geil (92K3287) described the development and performance of a high solid sample torch which when used in conjunction with a USN produced encouraging preliminary results for continuous determinations of trace metals in matrices such as sea-water brackish water and beverages. A comparison of torch designs has been made by Atherton et al. (92K4203). A new all glass demountable torch was compared with a fixed and a fixed-base demoun- table torch. A variety of samples including those contain- ing high solids were analysed. Nygaard et al. (92K3328 92/C3799) used an axially viewed ICP operating with low gas flows and at 40 MHz to obtain better detection limits and precision compared with a radially viewed ICP.In addition interferences and linearity were un-affected by the end-on viewing. Intra-alkali interferences were also exam- ined and methods for minimizing them studied. 2.3.2. Spectrometers A review (62 references) of ICP instrumentation including charge transfer detectors Cchelle grating and Fourier transform spectrometers sample treatment and introduc- tion devices and expert systems has been presented by Wang et al. (92/1712). Kolizynski et al. (92/2461) have reviewed (27 references) detection by charge injection devices (CID). The advantages of such systems are dis- cussed. The wavelength positioning accuracy of a sequential ICP-AES has been determined by Grosser and Collins (92/2052).The long-term thermal errors were found to be only a few picometres and short term instability had no significant effect on the wavelength. An improved wave- length Cali brat ion algorithm was described. A lot of interest has focused on photodiode array (PDA) spectrometers. An on-line intelligent background correction system for ICP-AES using a laboratory-made PDA spectro- meter has again been described by Huang et al. (92K3585 931593). The system enabled acquisition of data with concurrent background correction. In a related paper (93/507) the same workers presented data for LODs which ranged from 0.002-0.1 mg 1-1 for a variety of analytes. Chang (92/C4 108) found that a PDA spectrometer made in- house gave LODs slightly inferior to those obtained using a PMT.Brushwyler et al. (92/1622) have characterized a PDA spectrometer which uses a series of dispersion gratings and an optical mask to block unwanted parts of the spectrum thereby optimizing resolution. Limits of detec- tion were reported to be comparable with those obtained using a PMT. Interferences were minimal but scattered light was problematic. McGeorge et al. (92/C3796) have described improvements to a commercial PDA echelle spectrometer. The design was said to be versatile and can easily be re-configured for specific tasks. The spectral range was improved to include analytes such as K Li and Na. In a related paper (92/C3320) hydride-forming elements were determined using the same system. The low noise characteristics excellent quantum effici- ency large number of pixels and large dynamic range of a CCD has been used by Bilhorn (9212697).The system was an improved design which was described as ‘anti-bloom- ing’ i.e. saturated areas of the detector do not spill charge to adjacent areas. The custom built spectrometer was based on an Cchelle grating with a CaF prism that operates from 180 to 700 nm. Pomeroy et al. (92/C3811) used this system to resolve the As 228.812 nm and the Cd 228.812 nm lines. Barnard et al. (92/C3706 92/C3793) described a new solid state detector which reportedly provides the same flexibility and data acquisition rate of PDA and CCD spectrometers but with the same photometric performance as a PMT. An echelle spectrometer made from materials that elimi- nate thermal expansion effects has been described by Cassagne et al.(92X3104). Closely spaced lines such as B/Fe Cd/As and P/Cu were resolved and the performance compared with a larger 1 m focal length Paschen Runge spectrometer. Cassagne et al. (92/C4 18 1) described the structure and performance of an echelle spectrometer with a novel multiplexing of detectors as well as an axially viewed ICP source. An automated echelle cross-prism spectrometer was described by Gower et al. (92K3783 see also 93/C40). Performance in terms of BEC detection limits and stability were reported. A discussion of Czerny- Turner monochro- mators and a theoretical study was presented by Wuensch et al. (92/2787). The use ofjfibre optics has gained some attention. Fibre optics have been placed between the ICP source and the entrance slit of the spectrometer (92K4182).The fibre optic can be several metres long and therefore enable the spectrometer to be placed away from the source. An enclosed ICP-AES system for toxic or radioactive samples that can also utilize fibre optics has been described by Marty et al. (921C3330). The novel optics of this instrument allowed coupling to between one and three spectrometers. A spectrometer with fibre optic coupling between the poly- chromator and the PMTs has been described by Quillfeldt and Notzdd (92K3481). One hundred and thirty two optical fibres for 70 elements were positioned to 12 PMTs. Two papers (92lC4133 92X4186) described a new mechanism for wavelength scanning in sequential ICP-AES spectrometers. A harmonic drive mechanism was claimed to have a much improved performance because it provides very fast rotation very good positional repeatability no backlash long operating life and continuous spectrum access.A sequential ICP-AES instrument has been developed (92/C3566) which incorporates a double monochromator and several diffraction gratings thus enabling shorter run times and simultaneous background correction. High resolution spectrometers and the parameters influencing instrumental broadening and the consequences for the optimization of S/B have been reported by Mermet et al.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY. JUNE 1993 VOL. 8 163R (9211673). High-resolution ICP-AES has also been used to determine REE resolve the three isotopes of U and to analyse various steels and geological matrices (92/C362 7).2.3.3. Instrument Control and Chemometrics Automated analysis received surprisingly little attention in the literature and this is unfortunate in times when the analyst must in most cases strive to be as cost effective as possible. Shields and Peipmeier (9212837) have used a PC to control a mirror that images light from anywhere in a plasma source onto a high resolution echelle spectrometer. This allowed characterization of the source but would also allow optimization of the observation point for different spectral lines. Agnes and Horlick (92/C3297) integrated an automated robotic sample preparation system an ICP source where all the usual parameters including generator frequency were programmable and a direct reading spectro- meter. This system allowed not only variable analytical strategies based on results but also automated optimization for a given matrix.Borsier (92K3296) only considered automated sample preparation for ICP-AES describing a dedicated system based on conveyor belts. It was suggested that a more modern flexible system would be based on robots. The most common method of optimizing ICP-AES after that of a simple univariate search is the use of sequential simplex optimization and this method was extensively reviewed (82 references) by Golightly and Leary (92/3890). A new computer program for the modified simplex approach has also been published by Moore and Bohmer (9211 835). Brenner and Le Marchand (92K3339) suggested that simplex optimization can only provide minimal improvement with complex matrices and in this situation LODs are mainly dependent on spectrometer resolution.Automatic control of the spectrometer entrance and exit slits allowed them to maximize line emission intensity and minimize the effect of interferences when using a scanning ICP-AES. An example of the determination of trace elements in tungsten was given. Matherny and Eckschlager (93/1122) assessed the efficiency of individual optimization steps using information theory with simulated data. Tyler and Shkolnik (92K3336 92K3630) considered the effect of optimizing to different analytical criteria i.e. a set of analytical conditions that improved precision at high levels or improved peak location at low levels. They suggested that in sequential ICP-AES the importance of peak location is underestimated and therefore optimization for peak response might be better than optimization of SIB.Bauer et al. (92/ 1993,92/2006) have performed multivar- iate calibrations and estimated a ‘real’ LOD derived from error propagation theory. Contributions from calibration error were included in the calculations. Interconnections between derived error in concentration and selectivity were found theoretically. However investigations into different definitions of selectivity revealed only limited correlation to errors in concentration. Carre and Mermet (921C3335) also investigated the problems of errors in calibration. They considered the measurement of a concentration in an unknown test material as a two part process. Calibration where a series of standards are used with regression anlaysis to produce a calibration graph followed by counter calibra- tion where the graph is used to calculate concentrations in unknown samples.They observed that both parts contain errors and confidence bands were used to test the adequacy of curve fitting and to evaluate the consequence for measured concentrations. Wegscheider et al. (92K3549) compared multivariate calibration with least squares and Kalman filtering. Both approaches were shown to yield improved detection limits because of the averaging advan- tage of measuring at several wavelengths and because explicit modelling of background was better than discrete measurement. Chen (92IC4130) has proposed a new itera- tive technique known as analyte standard additions that minimizes the effect of matrix and background by giving a better estimate of the ‘true’ background. Starn et al.(92/C3713) have suggested a practical method of on-line calibration and standard additions by the use of a computer controlled HPLC pump. This type of approach where samples and standards are modified chemically within an analysis run could offer a great deal in the future particularly now that advanced computer control and data processing are available. Drift correction is normally performed on both a high- and low-response value. Gueldner (92/ 1878) suggested that rather than using test solutions this process can be performed by using the spectral background and the photo- multiplier dark current as the two points. Data suggested that the applicability of this method depends on spectro- meter properties and favours certain elements.Fredeen and Ivaldi (92/C3758) and Krushevska et al. (92K3337) have both successfully used Myers- Tracy signal compensation but it was noted that ionic lines close in energy sum to that of the internal standard were those best corrected. Mermet (92/C3295) considered the long-term stability limitations of ICP-AES. No commercial software can monitor the ICP parameters that can cause drift and therefore test elements must be used for drift diagnostics. These can be used as internal standards if carefully chosen or might be used as part of a feedback network. Quality assurance and quality control (QA/QC) are two phrases that will be increasingly familiar to analysts working in a more commercial environment.Olsen and Holst (92/4579) described the calculation of a ‘method evaluation function’ (MEF) and they defined this function as the expected value of an analysis as a function of the ‘true’ content of the analyte in the test materials. This generalized function also provides additional information intended to help write tight QA/QC protocols something that is likely to become an increasing demand on the analyst. The determination of trace elements in cellulose filters was given as an example specific to ICP-AES. Borszeki et al. (92/4577) grouped lead alloy samples from ‘good’ and ‘bad’ manufactured products. Linear discrimi- nate analysis of ICP-AES elemental determinations was able to make the classification accurately and was consi- dered to be useful for quality control.Shaw (93/C21) described software to implement a QC protocol for a scanning ICP-AES instrument. The checks it makes are switchable and have user set limits. It also allows remedial actions. These are useful features that many manufacturers could consider for their future software. ArtiJicial intelligence (AI) expert systems and neural networks are an area of chemometrics that is gradually developing with most information currently only appear- ing in conference abstracts. Ugolev and Sokolova (93/1264) have suggested that chemometricians can learn much from many branches of science particularly the ‘hot areas’ of computer science. These techniques could be used to solve problems when there is a degree of uncertainty in experi- mental information.They described new aspects of chemo- metrics with examples obtained when analysing biological material for Cu. Pomeroy et al. (92/2054 and 93/C300) have demonstrated that the large amount of information that array devices in spectrometers provide has allowed the development of expert systems to use this data for ‘on the fly’ matrix-dependent line selection. However such a strat- egy needs a large database of lines and the use of fundamental spectroscopic principles. An example of using such an expert system for environmental monitoring of chemical waste was given. The choice of line selection prior to analysis led Chen el al. (92K4134) to produce a database164R JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JUNE 1993 VOL. 8 that could intelligently associate spectroscopic analysis literature information and acquired knowledge.The pres- ence or absence of an element in ICP-AES spectra should be a simple matter of looking for prominent lines. However in real matrices this is complicated by interferences and statistical error. Wegscheider et al. (921C42 17 921C3306) have used fuzzy logic to make decisions of this nature on a neural network by using Bayes like probabilities. The system worked satisfactorily for trace elements in a tung- sten matrix. Nikdel (921C3307) had previously character- ized the trace element content of orange juice from around the world and a pattern recognition program ‘ARTHUR’ had been used to classify orange juice based on this training set. However this program had no learning ability.In contrast a neural network subsequently tried could learn. The network proved to be both very fast and have a better success rate. Salin et al. (92/C3304) have been developing an expert system to automatically run an ICP-AES instru- ment and they have now evaluated what the system provides as opposed to what the analyst appears to need. Spectral manipulation has provided many reports in the last year because with the wide availability of low cost scanning monochromator systems and powerful personal computers the use of chemometric techniques to manipu- late spectra particularly to correct interferences is much easier than in the past. In a series of papers Yang et al. (9211983 9212718 931698) have developed a numerical derivative technique. This was first evaluated with respect to improving selectivity and they defined a measure of this as the ‘interferent equivalent concentration (IEC)’.Factors that affect derivative spectra such as line shape were discussed in terms of the IEC (921 1983). Noise on raw data also hampered derivative calculations and smoothing and spectral step size criteria were investigated to minimize the IEC (92127 18). They concluded that derivative spectros- copic methods improved SIB by 1-2 orders of magnitude. However this was not accompanied by an improvement in detection limits in simple spectra but when spectral interference occurs the higher resolution can avoid deter- ioration of detection limits (931698). The method was tested on the determination of trace elements in the difficult Y20J matrix and found to give more reliable results than conventional off-peak/on-peak correction methods.Sun et al. (9212246) returned to first principles to solve the problems of spectral interference describing the philosophy behind their computer simulation of ICP-AES spectra and spectral interferences. Formulae for the calcula- tion of spectral linewidths were then proposed for the simulation. An example of the determination of V in an iron matrix was used to test the simulation against real spectra (931487). A modified model was also used to describe line profiles. Voigt profiles were decomposed into Gaussian and Lorentzian components. By comparing simu- lated and real line profiles they determined that the instrumental profile (as opposed to the natural line profile) of a medium resolution spectrometer was approximately Gaussian (921C4118). This modified model was then used to simulate the interference of iron on the Cr 283.6 nm line and the result compared with experimental spectra.Both were unable to resolve the interference but importantly the model was able to predicate spectrometer parameters that could (921C4119). Boumans (9315 16) has also produced a simulation of ICP-AES spectra from fundamental prin- ciples and a database of information on 350 prominent lines of 65 elements. The database of the computer program was designed to be customized by the user. The Kalman filter has been of particular interest as a means of correcting for spectral interferences. Van Veen and De Loos-Vollebregt (9211818) discussed its use and concluded that a spectral characterization should be performed prior to filtering but that it would give only a 1-3-fold improvement in detection limit and it would not correct for non-spectroscopic matrix effects.Karpate (921C4242) compared single element evaluation multivari- ate calibration and Kalman filtering for the determination of trace elements in bauxite and suggested that Kalman filtering gave more reliable results when determining trace elements in the presence of spectral interferences. Ma et al. (921C4 I 16) also evaluated the use of the Kalman filter for a number of complex matrices. Zhang et al. (921C4 136) have used factor analysis to correct for spectral interferences. This required both test spectra and a number of pure spectra to allow a single decomposition.Tests on real data suggested that errors could be due to non-linear combina- tions of pure spectra. A large amount of research into spectral characteristics to aid quantification is being reported only at conferences. Perhaps this suggests that these methods are not yet robust and routine. Improved methods of peak fitting and smooth- ing have been reported (921C3596 93/C2 14 931C298 931C43 1 931C699). A better estimation of background (natural and spectral interference) and its subtraction have been discussed ( 9 2 ~ 4 1 1 6 92/C4 1 20 9 2 ~ 4 1 2 I 92K4 123) 931C80 931C952). Specifically Caughlin and Blok (921C3280) have addressed the problem of FI for complex matrices where accurate background correction is difficult because of the constantly changing signal.They considered two instrumental approaches to background correction. The use of photodiode arrays allows simulta- neous acquisition of peak and background or alternatively high speed data acquisition that allowed accurate on1off peak correction during the period of the changing FI signal. 3. MICROWAVE-INDUCED PLASMAS Several reviews regarding MIPs have been published most notably those by Uden (9214268 931820) and Bulska (9214620) who give a robust assessment of the development and application of MIPs as element selective detectors for chromatography. Coulombe et al. (931609) reviewed the fundamental aspects of MIPs as did Mermet (9311090) who also discussed other electrically generated plasmas. Other review and summary papers published pertinent to chroma- tographic applications included those by Sullivan (92/4527 93/603) Webster and Carnahan (931602) and by Deruaz and Brazier (93/1248).the spatial emission characteristics of a 450 W atmospheric pressure He MIP. The TMolo cavity with side viewing port was mounted on an x-y stage allowing movement of the plasma with respect to the optical axis of the spectrometer and the authors found that emission maxima for different elements were distributed along the length of the cavity with Ca nearest and Cl furthest from the bottom of the discharge. The authors suggested that excitation mecha- nisms for metals and non-metals differed with C1 I1 emission being excited via a charge transfer mechanism. Wu and Carnahan have described the characteristics of a high-power MIP (931724).Their device referred to as a kilowatt-plus MIP used a re-designed torch and cavity and was reported to give improved plasma energy coupling and excitation characteristics. On a practical level the limit of detection for C1 was improved by two orders of magnitude. 3.1. Fundamental Studies Literature reports in this area appear to be declining in number. Pak and Koirtyohann (9212628) have reported onJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JUNE 1993 VOL. 8 165R High power MIPS were also discussed by Furuta and Koga (92lC3302 92lC3593) who described an annular nitrogen plasma which tolerates an aqueous aerosol in a similar manner to the ICP. The authors measured an excitation temperature of 5400 K and suggested that LTE conditions prevailed in this plasma.The effects of EIEs were reported by Jin et al. (92/3051). Aqueous samples were introduced via a USN with a desolvation device and the eflects of EIE upon the atomic emission and atomic absorption of a range of analyte elements were studied. The interaction of power gas flow rates and discharge tube diameter were investigated for a range of EIE (Li Na K Rb Cs) and analyte emission and absorption lines. Although detailed observations were reported it is disappointing that the authors did not interpret the data in mechanistic terms. 3.2. Instrumentation Broekaert et al. (92/2489) have published a comparison between a Beenakker cavity and a surfatron device (see J. Anal At Spectrom. 1992 7 169R). The devices were each operated at optimum conditions using electrothermal sam- ple introduction and the authors found that for Cd and Cu the surfatron device gave lower detection limits and a greater linear dynamic range and showed lower interference from EIE.There is however some contradiction of this last point in the text of the paper in that the authors also mention that the calibration curves with sodium present when using the surfatron device were not parallel to those of pure solutions. Duan et al. (92/38 18) have evaluated a 50 W argon MIP as an atomizer for Hg determination using atomic fluores- cence spectrometry. The device was operated using Ar at 600 ml min-I. Mercury vapour was generated following reduction of Hg with 5% SnCI2 solution. Excitation of Hg was by a pulsed HCL and the instrument gave an LOD of 3 ng m1-I.Evaluation of the system indicated that the viewing height was a critical parameter however the assessment of this system would have benefitted from the use of a rigorous multivariate optimization. 3.3. Sample Introduction 3.3.1. Direct nebulization Aerosol desolvation remains a priority for some workers in this field (92/C3322 92/C3680 93/C180 93/C262) how- ever there has been no substantive literature contribution for some time perhaps indicating a decline in research activity. 3.3.2. Electrothermal vaporization Although sample introduction by direct nebulization is declining the popularity of ETV continues because the solvent loading to the plasma in a typical low power cavity is reduced to a level that does not quench the discharge.Matusiewicz and Kurzawa (92/2228) described the determi- nation ofAs and Se with detection limits of 150 and 200 pg respectively. Evans et al. (92/2754) described a novel tantalum tip ETV device suitable for both MIP-AES and MIP-MS. For volatile elements and using MS detection LODs were generally sub-pg and limited by contamination from the materials used to build the device. Electrothermal vaporization sample introduction was also used by Broekaert et al. (9212489) in the comparison of two MIP sources. system. This was achieved using 10 mmol 1-1 potassium persulfate in 5 mol 1-1 sulfuric acid. The limits of detection for the 447.78 470.49 and 734.86 nm Br emission lines were 29.5 7.46 and 18.4 ng ml-l respectively. Sanz-Medel et al. (921C3294) have outlined methods for the determina- tion of halogens by continuous vapour generation. Oxida- tion of Br C1 and I was achieved using NaCIO KMnO and H202 respectively and detection limits of 20 ng ml-1 for I and of 2 ng m1-l for Br cited. If multivariate optimization was used more regularly (92/C3334) effective comparisons could be made between these systems.3.3.4. Direct analysis of solids In an interesting paper Gelhausen and Carnahan (9212395) reported the determination of the elemental ratios of C H Cl and S in coal following direct sample injection into a 500 W MIP operated in a TMolo cavity. Sample was introduced in 1 mg amounts via a stopcock in the carrier gas line. The key criterion the ratio between C and H which is associated with the rank of the coal could be determined with an accuracy of lo% provided measurements were simultaneous and an appropriate solid calibrant was used.It is unfortunate that no data were reported regarding the transport or atomization efficiencies of the system particu- larly as this type of plasma is not normally considered appropriate for ‘difficult’ matrices and that particles of up to 20 mesh size were introduced. The technique was also able to detect S and C1 although quantification was limited by matrix-induced background shifts. 3.4. Chromatography 3.4. I . Instrumentation Platzer et al. (93/527 931608) evaluated a 140W 27 MHz stabilized capacitative plasma in helium at atmospheric pressure. The discharge was contained within a liquid- cooled silica tube with power coupled via two annular electrodes. Detection limits for the device were reported to be in the low pg s-l range.Wu et al. (92/46 19) described a He r.f. plasma for N specific determination using an avalanche photodiode detector the detection limit for N was 57 pg s-l. The ability to perform simultaneous multi-element deter- minations of analytes is a major advantage of a GC-MIP system. This point was made by several workers (92K3303 92lC3468 92/C3723) but is particularly relevant with regard to the determination of empirical formulae. This has been discussed in earlier reviews in this series (see J. Anal. At. Spectrom. 1992 7 170R) and was again dealt with by Huang et al. (92/4369 92/4500) who appear to be perform- ing a detailed assessment of the capabilities of this application. An obvious extension of this work is the application of rigorous statistical methods to the data as by Kosman and Lukco (92/C3762) and this writer expects this to be a major theme in future updates.3.4.2. Gas chromatograph y-micro wa ve-induced plasma applications Organotin compounds have been determined by GC-MIP at levels of 1 pg 1-l (92/1653) and at an absolute detection limit of 6 pg (92/4271) indicating that this technique is capable of achieving low detection levels. Other organome- t a l k species that have attracted attention include organo- lead species (92K3467) and organomercury species (9212909) and Bulska et al. (9214105) have developed a robust method for the speciation of mercury in whole blood. Element specific analysis remains a major theme for GC-MIP.There have been reports regarding the determi- nation of halogens (93/819 92/C3432 93/C/199) for N 3.3.3. Chemical vapour generation Nakahara et al. (92/462 I 921C3587) described the determi- nation ofBr by oxidation of bromide in a continuous flow166R JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JUNE 1993 VOL. 8 and 0 (92/4298 92/C3434 92/C38 13) and for deuterated compounds (92/2526). 3.4.3. Supercrit ical fluid chroma tograp h y Webster and Carnahan (92/2057) have evaluated 250 W and 500 W He MIPS as SFC detectors. The higher power plasma showed a greater tolerance of variation of mobile phase pressure with stable He and C11I emission intensities for varying flow rates. These authors also evaluated the system for Cl and S determinations using C 0 2 and N20 mobile phases (9212972). In the UV S/Ns were degraded because of molecular band interferences for both mobile phases whilst in the near IR molecular bands were problematic for the N 2 0 mobile phase only.4. DIRECT CURRENT PLASMAS An indication of the body of work that has been performed using DCPs was given by Keliher (92/3892) who reviewed the application of the DCP to geochemical analysis (45 references) and Krull(93/822) who reviewed chromatogra- phic couplings to the DCP (76 references). Developments in instrumentation for DCP tend to focus on modification of the plasma source. This frequently involves the addition of more electrodes which in the most recent variation is six. McGuire and Piepmeier (92/4592) have evaluated a six electrode DCP. The plasma consisted of a horizontal a.c. plasma formed between three electrodes at the top of the device and three d.c. arcs formed between the base electrodes and the top electrode.Sample was introduced through the centre of the three arcs A simplex procedure was used to optimize the system gas flows and d.c. current however detection limits for several elements were approximately 100 times worse than those for ICP- AES. Brindle et al. (9214094) described a continuous jlow hydride generation system for the determination of As in water. Sample and NaBH were pumped into the hydride generator in separate flows. An Ar flow rate of 0.4 1 min-' through the base of the separator both mixed the sample and stripped analyte from solution into a 2 1 min-' carrier Ar flow. Also on sample introduction Thompson and Boss 92/C3768) found that small volume samples (40 pl) could be introduced without significant degredation of signal intensity.Solid sample introduction using slurries has been re- ported frequently for the DCP and in this review period there has been a further upsurge in interest. This may be a reflection upon the potential for higher transport efficiency of larger solid materials although this remains a subject for further work as does the assessment of particle atomization. Jerrow et al. (92/383 1) reported the determination of major elements in soils following fine grinding of the material with zirconia beads and confirmed the importance of reducing particle size to the range 2-5 pm in order to achieve high transport efficiency of the sample to the plasma.Yoon and Long (92/3 102) did not appear to show such clear under- standing of the importance of particle size reduction to achieve high transport efficiency however their proposed addition of propane gas to the nebulizer gas flow may be advantageous for the reduction of refractory oxides. A particulary exciting slurry sample introduction applica- tion has been developed by Fairman et al. (93/C26). Elemental data from the analysis of aqueous slurry samples of kaolin along with multi-component linear regression analysis was used to predict the important product quality characteristics of abrasiveness montmorillonite content and viscosity in china clay production. The methodology employed was rapid and was able to predict accurately (98.3% reliability) the variation in sample abrasiveness and has the capability to be employed for on-line process control in a clay extraction plant.The combination of such analytical capability with information based expert systems (9212054) is clearly an area where DCPs could become a valuable tool. Other more conventional applications of DCPs included the determination of Au (92/2518); of Si in urine (93/1096); of REE in ores (92/2230); the analysis of biological materials (92/4285); and the analysis of waste oils used as secondar-v fuels (92/C3678) where the sample might contain varying ratios of solvent and aqueous phases. LOCATION OF REFERENCES The full list of references cited in this Update have been published as follows 9211448-9212589 J. Anal. At. Spectrom.1992 7(4) 173R-2 13R. 9212590-92lC3494 J. Anal. At. Spectrom. 1992 7(5) 247R-277R. 92lC3495-9214073 J. Anal. At. Spectrom. 1992 7(6) 329R-348R. 9214074-9214734 J. Anal. At. Spectrom. 1992 7(8) 389R-411R. 93lC1-93lC997 J. Anal. AI. Specr'rorn. 1993 8( l) 45R-78R. 931998-93lC1354 J. Anal. At. Spwcfrom. 1993 8(3) 137R-149R. Abbreviated forms of the literature references quoted (excluding those to Conference Proceedings) are given on the following pages for the convenience of the readers. The full references names and addresses of the authors and details of the Conference presentations can be found in the appropriate issues of JAAS cited above. Abbreviated List of References Cited in Update 9211084. Methods Phys. Rex Sect. B 1990 B51 133. 9211625. Spectrochim. Acta Part B 199 1,46,2 17.9211626.9211219. Anal. Chem. 1990,62,2 158.9211460. Lab. Pract. Spectrochim. Acta Part B 199 1 46 229. 9211627. Spectro- 1990 39 7 1 75. 9211622. Spectrochim. Acta Part B 199 1 chim. Acta Part B 199 1 46 253. 9211628. Spectrochim. 46 85. 9211623. Spectrochim. Acta Part B 1991 46 115. Acta Part B 1991 46 269. 9211629. Spectrochim. Acta,JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JUNE 1993 VOL. 8 167R Part B 1991 46 283. 9211630. Spectrochim. Acta Part B 1991,46 291. 9211631. At. Spectrosc. 1991,12 1.9211641. Spectrochim. Acta Part B 199 1,46 1063. 9211653. Colloq. Atomspektrom. Spurenanal 5th 1989 429. 9211671. U.S. Pat. Appl. US 592,489 15 Apr 199 1 Appl. 03 Oct 1990; 29 pp. Avail. NTIS Order No. PAT-APPL-7-592 489.9211673. Proc. SPIE-Int.Soc. Opt. Eng. 1990 1318 88. 9211688. Appl. Spectrosc. 199 1 45 682. 9211690. Appl. Spectrosc. 1991 45 701. 9211692. Microchem. J. 1991 43 213. 9211709. Anal. Chim. Acta 1991 246 413. 9211712. Yankuang Ceshi 1990 9 245. 9211764. J. Assoc. OJf Anal. Chem. 1991 74 516. 9211782. J. Serb. Chem. SOC. 1990 55 457. 9211790. Guangpuxue Yu Guangpu Fenxi 1990 10(6) 54. 9211807. Anal. Sci. 1991 7 433. 9211818. Anal. Chem. 1991 63 1441. 9211835. Chemom. Intell. Lab. Syst. 1991 11 97. 9211858. Hutn. Listy 1990 45 355. 9211878. Neue Huette 199 1,36 109.9211978. Spectrochim. Acta Part B 1991 46 851. 9211979. Spectrochim. Acta Part B 1991 46 869. 9211981. Spectrochim. Acta Part B 199 1,46 90 1.9211983. Spectrochim. Acta Part B 199 1,46 953. 9211984. Spectrochim. Acta Part B 1991 46 967.9211988. Spectrochim. Acta Part B 1991 46 1073. 9211989. Spectrochim. Acta Part B 1991 46 1089. 9211991. Spectrochim. Acta Part B 1991 46 1149. 9211993. Spectrochirn. Acta Part B 1991 46 1185. 9211994. Spectrochim. Acta Part B 1991 46 1207. 9211998. Spectrochim. Acta Part B 1991 46 1275. 9212003. Fresenius’ J. Anal. Chem. 199 1 340,4 1.9212006. Fresenius’ J. Anal. Chem. 199 1 340 135. 9212007. Fresen- ius’ J. Anal. Chem. 1991 340 157. 9212019. Fresenius’J. Anal. Chem. 1991 340 435. 9212052. Appl. Spectrosc. 1991 45 993. 92/2054. Appl. Spectrosc. 1991 45 1 1 11. 9212055. Appl. Spectrosc. 199 1 45 1 120. 9212057. Appl. Spectrosc. 1 99 1 45 1285. 9212058. Appl. Spectrosc. 1 99 1 45 1 327. 9212086. Colloq. Atornspektrom. Spurenanal. 5th 1989 2 1. 9212091.Colloq. Atomspektrom. Spurenanal. 5th 1989 9 1. 9212105. Colloq. Atornspektrom. Spurenanal. 5th 1989 3 13. 9212114. Colloq. Atornspektrom. Spurenanal. 5th 1989 523.9212228. Acta Chim. Hung. I99 1,128,401. 9212230. Chem. Listy 199 1,85,654.9212240. Guangpuxue Yu Guangpu Fenxi 1990 10(6) 30. 9212246. Guangpuxue Yu Guangpu Fenxi 1991 11( I) 33. 9212253. Guangpuxue Yu Guangpu Fenxi 1991 11(2) 33 41. 9212257. Guang- puxue Yu Guangpu Fenxi 1991 11(2) 68. 9212280. Zh. Prikl. Spektrosk. 1991 55 310. 9212393. Anal. Chem. 1991 63 2357. 9212395. Anal. Chem. 1991 63 2430. 9212396. Anal. Chem. 1991,63 2539. 9212404. J. Anal. At. Spectrom. 1991 6 527. 9212406. J. Anal. At. Spectrom. 199 1 6 541. 9212407. J. Anal. At. Spectrom. 199 1 6 545. 9212455. Microchem. J. 199 1 44 1 17.9212460. Am. Lab. (Faitfield Conn.) 1990 22(18) 22 24 28. 9212461. Am. Lab. (Fairjield Conn.) 1991 23(8) 48 50. 9212486. Anal. Sci. 1991 7 773. 9212489. Talanta 1991 38 863. 9212494. Chern. Anal. (Warsaw) 1990 35 5. 9212496. Chem. Anal. (Warsaw) 1990 35 33. 9212499. Chern. Anal. (Warsaw) 1990 35 109. 9212507. Chem. Anal. (Warsaw) 1990 35 31 1. 9212515. Anal. Chim. Acta 1991 248 241. 9212518. Anal. Chim. Acta 199 1 248 569. 9212526. Anal. Lett. 1991 24 1531. 9212542. LaborPraxis 1991 15 512 514 516. 9212544. Spectrochim. Acta Rev. 1990 13 275. 9212554. Spectrosc. Lett. 199 1 24 607. 9212556. Anal. Proc. 1991 28 79. 9212597. Anal. Proc. 1991 28 416. 9212601. Analyst 1992 117 13. 9212615. J. Anal. At. Spectrom. 1991 6 623. 9212617. J. Anal. At. Spectrom. 1991 6 631.9212628. Appl. Spectrosc. 1991 45 1132. 9212630. Appl. Spectrosc. 199 1 45 1225. 9212633. Appl. Spectrosc. 199 1 45 1368. 9212634. Appl. Spectrosc. 199 1 45 1372. 9212642. At. Spectrosc. 1991 12 162. 9212644. At. Spectrosc. 1991 12 199. 9212662. Can. J. Appl. Spectrosc. 1991 36(5) 114. 9212683. Talanta 1991 38 1265. 9212687. Appl. Opt. 1991 30 2990. 9212697. Proc. SPIE-Int. SOC. Opt. Eng. 1991 1448 74. 9212699. Proc. SPIE-Int. SOC. Opt. Eng. 1991 1502 31 1. 9212718. Fenxi Huaxue 1991 19 993. 9212735. Fenxi Shiyanshi 1991 10(2) 12. 9212738. Fenxi Shiyanshi 1991 10(2) 45. 9212740. Fenxi Shiyanshi 199 1 10(3) 7. 9212750. Appl. Spectrosc. 199 1 45 14 13. 9212752. Appl. Spectrosc. 1 99 1 45 1463. 9212753. Appl. Spectrosc. 1991 45 1468. 9212754. Appl. Spectrosc.1 99 1 45 1478.9212785. Spectro- chim. Acta Part B 199 1 46 147 1. 9212786. Spectrochirn. Acta Part B 1991 46 1499. 9212787. Spectrochim. Acta Part B 199 1,46 15 17.9212837. J. Autom. Chem. 199 1,13 129. 9212839. J. Chin. Chem. SOC. (Taipei) 1991 38 327. 9212909. Water Air Soil Pollut. 1991 56 103. 9212945. Zavod. Lab. 1991 57(2) 35. 9212972. Anal. Chem. 1992 64 50. 9212973. Acta Chim. Hung. 1991 128 455. 9212974. Acta Chim. Hung. 1991 128 463. 9212983. Acta Chim. Hung. 1991 128 699. 9213045. Bunseki Kagaku 1991 40 T125. 9213051. Microchem. J. 1991 44 153. 9213068. Mikrochim. Acta 199 1 2 265. 9213075. ASTM Spec. Tech. Publ. 1991 1109 77. 9213083. Zh. Prikl. Spektrosk. 199 1 54 10 1 1. 9213084. Zh. Prikl. Spektrosk. 1991 55 7. 9213093. Bunseki 1991 7 553. 9213102.Report 1989 DOE/PC/80532-T4; Order No. DE910013728 165 pp. Avail. NTIS. From Energy Res. Abstr. 1991 16(4) Abstr. No. 9469.9213105. Report 1991 IS-T- 1528; Order No. DE9 1009868 130 pp. Avail. NTIS. From Energy Res. Abstr. 1991 16(6) Abstr. No. 14967. 9213122. Eur. Pat. Appl. EP 448,061 (CI. GOlN21/67) 25 Sep 1991 JP Appl. 90/67,364 19 Mar 1990; 16 pp. 9213232. Prib. Tekh. Eksp. 1991 3 120. 9213260. Tetsu to Hagane 199 1 77 1936. 9213262. Tetsu to Hagane 199 1 77 1985. 9213276. Zhongguo Jiguang 1991 18 558 544. 9213818. J. Anal. At. Spectrom. 1992,7 7.9213819. J. Anal. At. Spectrom. 1992 7 11. 9213820. J. Anal. At. Spectrom. 1992 7 15. 9213821. J. Anal. At. Spectrom. 1992 7 19. 9213831. Anal. Proc. 1992 29 45. 9213848. Chem. Geol. 1992 95 73. 9213849. Chern.Geol. 1992,95 85. 9213852. Chem. Geol. 1992 95 13 1. 9213858. Geochim. Cosmo- chim. Acta 199 1 55 9 17.9213864. Spectrochim. Acta Part B 199 1,46 125.9213889. Spectrochim. Acta Rev. 199 1 14 95. 9213890. Spectrochim. Acta Rev. 1991 14 1 1 1. 9213892. Spectrochim. Acta Rev. 199 1 14 16 1. 9213894. Spectrochim. Acta Rev. 1991 14 303. 9213946. Fenxi Huaxue 1990 18 1 158. 9213976. Fenxi Shiyanshi 199 1 10 29 68. 9214021. Spektr. Anal. Tr. Mosk. Kollok. PO Spektr. Anal. 1989-90. AN SSSR. Otd-nie Obsh. Fiz. z. Astron. Nauch Sov. PO Spektroskopii M. 1990 1 1 65. 9214094. Analyst 1992 117 407. 9214105. Analyst 1992 117 657. 9214268. Anal. Appl. Spectrosc. 2 1990 (Pub. 199 I) 165. 9214284. Biol. Monit. Exposure Chem. Met. 1991 145. 9214285. Biol. Monit. Exposure Chem. Met.1991 163. 9214298. CLB Chem. Labor Betr. 1990 41 200 203 206. 9214309. Chemom. Intell. Lab. Syst. 1991 11 12 1. 9214335. Ettore Majorana Int. Sci. Ser. Phys. Sci. 199 1 54 229.9214345. Gaodeng Xuexiao Huaxue Xuebao 199 1 12 183. 9214369. Huaxue Xuebao 199 1 49 232. 9214417. Mater. Trans. JIM 199 1 32 480. 9214472. Proc. Int. Conf Lasers 1989 (Pub. 1990) 750. 9214485. Rev. Colomb. Quirn. 1990 19 81. 9214500. Sepu 1991 9 141. 9214520. Tee. Lab. 199 1 13( 16 l) 24. 9214523. Thin Solid Films 199 1 205,6.9214527. Trends Anal. Chem. 199 1 10 23.9214563. J. Anal. At. Spectrom. 1992,7,481.9214564. J. Anal. At. Spectrorn. 1992 7 493. 9214571. J. Anal. At. Spectrom. 1992 7 539. 9214572. J. Anal. At. Spectrom. 1992 7 545. 9214577. Can. J. Appl. Spectrosc. 1992 37 1. 9214579.Analyst 1992 117 707. 9214592. Can. J. Appl. Spectrosc. 1991 36 127. 9214597. J. Anal. At. Spectrorn. 1992 7 69. 9214598. J. Anal. At. Spectrorn. 1992 7 75. 9214619. J. Anal. At. Spectrorn. 1992 7 197. 9214620. J. Anal. At. Spectrom. 1992 7 201. 9214621. J. Anal. At. Spectrorn. 1992 7 21 1. 9214622. J. Anal. At. Spectrorn.,168R JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JUNE 1993 VOL. 8 1992 7 219. 9214623. J. Anal. At. Spectrom. 1992 7 225. 9214625. J. Anal. At. Spectrom. 1992 7 235. 9214626. J. Anal. At Spectrom. 1992 7 239. 9214627. J. Anal. At. Spectrom. 1992 7 247. 9214636. J. Anal. At. Spectrom. 1992 7 301. 9214642. J. Anal. At. Spectrom. 1992 7 339. 9214722. Bunseki Kagaku 199 1 40 749. 9214726. Spectra 2000 (Deux Mille) 1 99 I 159 5 9214729. Spectrosc. Int. 1991 3 26. 9214732. Appl. Spectrosc. 1991 45 312. 931434. Anal. Chim. Acta 1991 254 109. 931478. Micro- chem. J. 1992 45 1. 931487. Guangpuxue Yu Guangpu Fenxi 1 99 1 11(4) 23. 931495. Guangpuxue Yu Guangpu Fenxi 199 1 11(5) 25. 931496. Guangpuxue Yu Guangpu Fenxi 199 1 11( 5 ) 32. 931507. Guangpuxue Yu Guangpu Fenxi 199 1 11(6) 22. 931515. Guangpuxue Yu Guangpu Fenxi 1991 11(6) 69. 931516. Spectrochim. Acta Part B 199 1 46 1545E. 931525. Spectrochim. Acta Part B 1992 47 61. 931527. Spectrochim. Acta Part B 1992 47 95. 931535. Spectrochim. Acta Part B 1992 47 239. 931538. Spectrochim. Acta Part B 1992 47 37 1.931559. Anal. Sci. 1991 7 1053. 931588. Anal. Sci. 1991 7 477. 931591. Anal. Sci. 1991 7 537. 931592. Anal. Sci. 1991 7 549. 931593. Anal. Sci. 1991 7 553. 931602. ACS Symp. Ser. 1992 479 (Elem. Specific Chromatogr. Detect. At. Emiss. Spectrosc.) 25. 931603. ACS Symp. Ser. 1992 479 (Elem. Specific Chromatogr. Detect. At. Emiss. Spectrosc.) 62. 931608. ACS Symp. Ser. 1992,479 (Elem. SpeciJc Chroma- togr. Detect. At. Emiss. Spectrosc.) 152.931609. ACSSymp. Ser. 1992 479 (Elem. Specific Chromatogr. Detect. At. Emiss. Spectrosc.) 189. 931610. ACS Symp. Ser. 1992 479 (Elem. Specific Chrornatogr. Detect. At. Emiss. Spectrosc.) 218. 931641. Pure Appl. Chem. 1992 64 227. 931649. Yankuang Ceshi 199 1 10 254. 931650. Yankuang Ceshi 1991 10 274. 931668. Proc. Int. Conf Lasers 1990 (Pub 199 l) 634. 931675. Fenxi Huaxue 199 1 19 1 137.931676. Fenxi Huaxue 1991 19 1141. 931680. Fenxi Huaxue 1991 19 1183. 931681. Fenxi Huaxue 1991 19 1192. 931686. Fenxi Huaxue 1991 19 1285. 931697. Fenxi Huaxue 1992 20 114. 931698. Fenxi Huaxue 1992 20 153. 931703. Fenxi Huaxue 1992 20 348. 931722. Appl. ,Spectrosc. 199 1 45 158 1. 931724. Appl. Spectrosc. 1992 46 163. 931725. Appl. Spectrosc. 1992 46 593. 931769. Fresenius’ J. Anal. Chem. 1992 342 391. 931789. Gao- deng Xuexiao Huaxue Xuebao 199 l 12 1022.931793 Zh. Prikl. Spektrosk. 1992 56 112. 931796. Zh. Anal. Khim. 1991,46 2274.931805. Anal. Chem. 1992,64,257.931819. J. Chroma togr. 1992 594 395. 931820. J. Chromatogr. Libr. 199 1 47 (Trace Met. Anal. Speciation) 1. 931822. J. Chromatogr. Libr. 199 1 47 (Trace Met. Anal. Speciation). 931828. J. Korean Chem. Soc. 1992 36 273. 931833. J. Phys. D Appl. Phys. 1992 25 425. 931852. Methods Enzymol. 199 1 205 (Metallobiochem. Pt. B) 1 98. 9311000. J. Anal. At. Spectrom. 1992 7 587. 9311001. J. Anal. At. Spectrom. 1992 7 595. 9311012. J. Anal. At. Spectrom. 1992 7 661. 9311019. J. Anal. At. Spectrom. 1992 7 707. 9311031. J. Anal. At. Spectrom. 1992 7 775. 9311071. Appl. Opt. 1992 31 3540.9311087. Spectrochim. Acta Part B 1992,47 1045.9311090. Actual. Chim. 199 1 4 279. 9311096. Ann. Clin. Lab. Sci. 1991 21 328. 9311106. ATB Metal!. 1990 30 85. 9311122. Chem. Pap. 1991 45 745. 9311123. Chem. Res. Chin. Univ. 1991 7 27.9311 142. Fuel Sci. Technol. Int. 1992,10,3 13.9311170. Izv. Sib. Otd. Akad. Nauk SSSR Ser. Khim. Nauk 1991,3 47. 9311171. Izv. Sib. Otd. Akad. Nauk SSSR Ser. Khim. Nauk 1991 3 52. 9311184. Lect. Notes Phys. 1991 389 53. 9311185. Lubr. Eng. 1992 48 227. 9311187. Magy. Kem. Foly. 1991 97 466. 9311188. Magy. Kem. Fol-v. 1992 98 33. 9311248. Spectra 2000 (Deux Mille) 1991 161 8. 9311254. Tetsu to Hagane 199 1 77 188 1. 9311264. Vestn. Leningr. Univ. Ser. 4 Fiz. Khim. 1991 3 77.