E. Hywel Evans*a, John B. Dawsonb, Andrew Fishera, Steve J. Hilla, W. John Pricec, Clare M. M. Smithd, Karen L. Suttone and Julian F. Tysonf
aDepartment of Environmental Sciences, University of Plymouth, Drake Circus, Plymouth, UK PL4 8AA
bDepartment of Instrumentation and Analytical Science, UMIST, PO Box 88, Manchester, UK M60 1QD
cEllenmoor, East Budleigh, Budleigh Salterton, Devon, UK EX9 7DQ
dDepartment of Chemistry, University College Cork, Ireland
eProcter and Gamble Technical Centres Ltd., Rusham Park, Whitehall Lane, Egham, Surrey, UK TW20 9NW
fDepartment of Chemistry, University of Massachusetts, Box 34510, Amherst, MA 01003-4510, USA
First published on 22nd May 2001
Several trends have been evident during the review period. There was a marked decrease in the number of papers describing applications of chemometric procedures, speciation using gas chromatography (though not speciation using liquid chromatography) and solid sampling using arcs and sparks. The number of publications describing applications involving some form of sample pretreatment, either on-line or batch, such as matrix removal, preconcentration or vapour generation, was similar to previous years. Similarly, applications involving lasers maintained their popularity. There were very few novel developments in instrumentation, with the only significant trend of note being that towards miniaturisation.
The micro-sampling characteristics of FI introduction continue to be examined and exploited. Dispersion in a FI-ICP-AES system was modelled by analysis of the impulse response function.8 Dispersion was minimized by the incorporation of air bubbles before and after the sample bolus. Even compact introduction systems produced significant dispersion. For the determination of alkaline earth metals in soils by FAAS,9 samples were introduced via a single-line FI manifold to minimize the errors due to the high dissolved solids content, and for the determination of Cr in lubricating oils,10 on-line emulsification with the aid of ultrasound produced a fluid that could be introduced directly into the atomizer for determination by ETAAS. In the determination of Fe in soil extracts by ICP-OES,11 the high dissolved solids content was handled by FI sample introduction. A sequential injection system was employed for the determination of Fe, Mn and Zn in wines,12 the volume of sample introduced being varied according to the analyte concentration so as to produce the desired degree of dilution.
Several novel FI calibration and dilution procedures have been developed. An FAA spectrometer was calibrated13 by generating concentration gradients from a standard and the sample. The solutions were delivered by a computer-controlled pump. Standard was delivered at an increasing rate and sample at a decreasing rate; in each case, the nebulizer total flow was made up by aspiration of water. The sample concentration was found from the intersection of the linear absorbance–time plots. Off-range samples could be diluted; for a calcium standard of 30 mg l−1, the working range was 1–200 mg l−1. The accuracy and precision of flow-based automatic dilutions has been compared with those of manual procedures.14,15 As might be expected, the accuracy and precision of the manual methods were higher, but the automatic procedures were faster. A version
of the generalized standard additions method has been implemented in a two-channel flow system.16 The procedure was evaluated by the determination of Ca, K and Na by FAES in solutions containing varying amounts of ethanol or glycerol. Dilutions up to 10000-fold17 with an RSD of 1.5% have been produced through delivery by electroosmotic pumping. For the determination of Mg by FAAS, the results were comparable with those of EDTA titration. Diffusional transport of metal salts through a track-etch membrane filter18 has been exploited for the calibration of a FAA spectrometer, and interferences in the determination of Zn in multivitamin tablets by FAAS were removed by electrodialysis.19
Dissolution or leaching procedures have been implemented in a FI or continuous flow mode. Anodic dissolution FI procedures have been reviewed.20 Tool steels were analysed for Cr, Fe, Mo, V and W by ICP-OES following on-line anodic dissolution21 in a cell directly coupled to the spectrometer. The dissolution of rates of Cr, Cu, Fe, Mn, Mo and Ni from 304 stainless steel were followed22 by ICP-AES in a study of the fundamentals of the process. Mild steel and brass were rapidly dissolved23 by first creating a colloidal suspension in 5–10 ml of water by spark ablation followed by the addition of 100 µl of concentrated acid. On-line leaching of Fe from iron ore24 was followed by determination by molecular absorption spectrometry; As was determined in soils by on-line microwave-assisted leaching from an injected slurry followed by HG-AAS;25 and Cd, Cr, Mn, Ni and Pb were determined by ICP-MS26 following microwave-assisted extraction in a system which incorporated a double pumping action instead of a back-pressure regulator. Problems associated with dispersion and dilution of the acid were also addressed.
Several indirect methods have been adapted to the FI format. Iodide was determined via Cr by the injection of a CrVI solution into a stream of iodide with on-line retention of the CrIII, produced by the reduction of CrVI, on a poly(aminophosphonic acid) column. When applied to the analysis of tap and sea waters,27 the CrIII retained on the column was eluted with 0.5 mol l−1 HCl and determined by FAAS. For a 3 min preconcentration, the LOD was 2 µg l−1, and the throughput was 17 h−1. For the analysis of iodized table salt,28 the excess CrVI was determined directly (also by FAAS) to give an increased throughput of 80 h−1. Ascorbic acid in vitamin C tablets was determined29 by a very similar procedure; the CrIII was retained on a cation-exchange resin prior to elution with 3 mol l−1 HNO3. The calibration range was 0.3–60 mg l−1, and the throughput was 20 h−1. Citric acid was determined,30via Pb, in confectionery by an on-line precipitation–filtration procedure. The calibration range was 2–40 mg l−1. This procedure has also been used to determine flavone in bee chrysalis,31 and amoxycillin.32 In the latter procedure, quantification was based on measurement of the residual manganese after reaction of the analyte with permanganate in the presence of barium chloride in alkali. Batch indirect procedures are described in a later section.
Preconcentration in a flame by trapping on a refractory oxide surface was used in the determination of Te in water.33 The analyte was collected on an alumina-coated slotted silica tube for 2–5 min and then atomized by a step change in the acetylene flow rate. The sensitivity was over two orders of magnitude greater than that of conventional nebulization. For the determination of Cu in eggs34 a slotted tube atom retarder was placed in the flame.
The first report of the performance of Pb-Spec (a crown ether in a long-chain aliphatic alcohol) in the pores of a solid support41 shows that LOD for Pb of about 0.5 mg kg−1 in high purity iron and steel samples could be obtained for a 30-ml sample volume. The Pb retained from 1 mol l−1 HNO3 solution was eluted with 0.1 mol l−1 ammonium oxalate solution directly into the nebulizer of the FAA spectrometer. A similar macrocylic compound, immobilized on silica gel,42 retained lead over a range of acid concentrations (0.075 to ≫3 mol l−1 HNO3) to give a solution LOD of 5 µg l−1. Elution was with 0.03 mol l−1 EDTA solution at pH 10.5. Some biological and environmental CRM were accurately analysed. Chromosorb 102, a gas chromatography stationary phase, retained43 the DDC complex of Pb from waters and soil digests to give an LOD of 5 µg l−1. The complex retained from pH 9 solutions was eluted with ethanol.
Polyurethane foam was the solid phase for the retention of Cd complexed with 2-(2-benzothiazolylazo)-2-p-cresol.44 For Cd the LOD was 0.3 µg l−1 (1 min loading). The retained complex was eluted with mineral acid solution. Silica-immobilized thiourea retained Ag, Au and Pd45 to give LODs of 1, 10 and 20 µg l−1, respectively, for a loading time of 1 min and elution with 5% thiourea solution. The procedure was applied to the analysis of a nickel alloy, an anode slime, an electrolytic solution (cobalt chloride) and three CRM ores. When the ligand was changed to rhodanine,46 the LODs were 4, 20 and 20 µg l−1, respectively. Metal dithiocarbamates used as fungicides were determined by retention on C60 fullerene47 with LODs of 1–5 µg l−1. The procedure was applied to the determination of ziram, maneb and ferbam in spiked grain samples. A column of activated carbon retained Ni as the complex with alpha-diphenylglyoxime48 prior to elution with dilute HNO3. The optimization strategy was a mixed-level orthogonal array design. The LOD was 0.05 µg l−1. The method devised was validated by spike recoveries (4–40 µg l−1) from Zayandeh-hood river. The potential of synthetic zeolites as sorbents has been investigated.49 The LODs were 0.1–0.4 µg l−1 for 2 min loading with elution into 300 µl of IBMK.
Gold has been retained on VION chemisorption fibres50 by a combination of ion-exchange and chelation. The chlorocomplex of Ag was retained51 on VS-II anion-exchange fibres and eluted with a 2% thiourea solution to give an LOD of 0.2 µg l−1. The same chemistry was applied52 to the separation of traces of Zn from a nickel matrix by retention on a strongly basic anion-exchange material. The eluent was water. The LOD was 2 µg l−1 for a throughput of 30 h−1. The chlorocomplexes of Pd, Pt and Rh were retained53 by a solid phase consisting of highly cross-linked polystyrene impregnated with 4-(n-octyl)diethylenetriamine (which would have been positively charged through protonation). The LODs were 3–8 µg l−1. The chlorocomplexes of Cu, Fe and Zn were retained54 on coarse-particle (>0.5 mm) anion-exchange resin from solutions of multivitamin tablets prepared on-line by passing the ground sample through a heated PVC coil (1.4 mm id). The analytes were eluted with dilute HNO3 and determined by FAAS (Cu and Zn) or molecular absorption spectrometry (Fe). Amberlite XAD-2 resin loaded with TAM55 retained Cd from rice flour and lobster hepatopancreas RM digests to give an LOD by FAAS of 1 µg l−1. The eluent was dilute HCl. Aluminium, chromium, cobalt, copper, iron, lead, manganese, nickel and zinc at concentrations of 200 µg l−1 did not interfere with the determination of 10 µg l−1 Cd. Amberlite XAD-2 loaded with calmagite56 retained Cu from sea-water and biological materials. For 3 min loading (13.5 ml of sample), the LOD was 0.2 µg l−1. This is a later publication of a previously published batch procedure57 which had the same detection limit. In a comparison of procedures for the preconcentration of Cd, Cu and Pb by retention on an iminodiacetate-modified silica,58 the flow version with a microcolumn was considerably faster. Tannin resin prepared from Eucaplytus saligna Sm retained Cu59 from digests of plant materials and foodstuffs. The eluent was again dilute HCl, the LOD was 1 µg l−1, and the throughput 48 h−1. Another biomaterial, dealginated seaweed, retained Cd, Cr, Cu and Pb60 from various water samples (including two Lake Ontario water RM). As CrVI was not retained by the material, a (somewhat inaccurate) speciation procedure for Cr was developed.
A precipitation–dissolution procedure was developed61 for the determination of Cd in mussels. The analyte was precipitated as the ion-pair between the tetraiodo Cd complex and quinine. The precipitate was dissolved in ethanol. A preconcentration of 32-fold was obtained for a sample volume of 15 ml, loaded at 3 ml min−1. The range of concentrations determined was 0.25–5.5 mg kg−1. Nickel DDTC co-precipitated Cd, Cu, Fe and Pb62 from 0.3 mol l−1 HNO3 with retention on the inner walls of a knotted tubular reactor. The precipitate was dissolved in IBMK. Enhancement factors at a sampling frequency of 60 h−1 were 65, 60, 59 and 58, and the LODs were 0.2, 0.5, 2 and 3 µg l−1 for Cd, Cu, Fe and Pb, respectively. The procedure was also applied to the analysis of mussels as well as hair, soils and waters. Electrochemical preconcentration on graphite particles63 led to an improvement in the LOD for Cu by FAAS of two orders of magnitude.
A number of papers describing preconcentration prior to determination by chemical vapour generation are covered later in Section 1.4, as are papers describing HG with ETAAS.
Thallium was retained on polyurethane foam modified with phosphomolybdate100 with elution by HNO3; the LOD was 20 µg l−1. Nickel has been determined in fruits by FAAS after preconcentration on active carbon101 as part of a larger study of the speciation of Ni in soils. Sequential extraction with various reagents was used to assign nickel to the following categories: exchangeable, carbonate-bound, iron and manganese oxide-bound and organically bound. A chemically modified chloromethylated polystyrene pyridyl-betanaphthol collected Au, Cd, Cr, Cu, FeIII, Mn, Ni, Pb, Pd and Zn from aqueous solutions at pH 8–9;102 the metals were eluted with 1 mol l−1 mineral acid solution and the procedure was applied to the analysis of waters, urine, milk and ores. A similar list of metals (Cd, Co, Cu, FeIII, Hg, Ni, Pb and Zn) was retained on silica gel modified with 5-amino-1,3,4-thiadiazole-2-thiol.103 The eluent was 2 mol l−1 HCl and the procedure was applied, with prior UV photodegradation, to waste waters and organic-rich natural waters. A column of activated carbon loaded with zinc piperazinedithiocarbamate collected Cd, Cu and Ni.104 The retained analytes were eluted with mercury solution. Under optimized conditions, the LOD were 0.1, 0.1 and 0.2 mg l−1, repectively. Silica gel, modified with covalently bound 8-hydroxyquinoline,105 retained Cu, Cd, FeIII, Pb and Zn.
Lead has been speciated106 by first precipitating with chromate, filtering and dissolving, followed by measurement of total inorganic lead. Alkyllead species were complexed with DDC and retained on a C60 column from which trimethyllead and triethyllead could be selectively removed by hexane and IBMK, respectively. For a 50-ml sample, LODs were 1–2 µg l−1. The possibility of using the surfactant cloud-point phenomenon for the speciation of metals has been demonstrated107 for the two oxidation states of Cr. On heating, an aqueous solution of a non-ionic surfactant becomes turbid (the cloud point) and then separates into two phases: a surfactant-rich phase composed almost entirely of the surfactant and an aqueous phase in which the surfactant concentration is close to the critical micelle concentration. If the original solution contains hydrophobic metal complexes, these may be extracted into the surfactant rich phase on heating. The procedure was applied to the selective determination of Cr by first complexing CrIII with 8-hydroxyquinoline, and then CrVI with APDC with Triton X-114 as the surfactant. The LOD were 1 and 0.6 µg l−1, respectively. Free Cd, Cu and Pb, in the presence of humic substances, were determined108 following dialysis with cuprophan planar membranes.
A non-woven polypropylene fabric was phosphonylated to improve the collection efficiency for Eu, Hg and Pb109 over that of ion-exchangers containing sulfur binding sites. Poly(ethylene terephthalate) fibres grafted with methacrylic acid collected Cd, Co, Ni and Zn110 from river- and synthetic sea-water samples. E.coli immobilized on sepiolite (the calcareous inner shell of the cuttlefish) collected Cd, Cu, Fe, Ni and Zn111 from alloy solutions prior to determination by FAAS. A somewhat similar procedure, in which the metals were collected on Aspergillus niger on sepiolite, was applied to the analysis of some geological materials.112 The ability of modified Slovak zeolites to collect CrIII was measured by FAAS.113
In the determination of Fe by FAAS114 preconcentration by co-flotation with lead hexamethylenedithiocarbamate gave an LOD of 0.3 µg l−1. A variety of electrode materials were able to preconcentrate Cd and Cu115 to an extent comparable to that obtained with complexing agents (8-hydroxyquinoline, dithizone and rubeanic acid immobilized on filter paper). Precipitation by reducing the solubility of polyacrylic acid116 on the addition of acetone preconcentrated Cd, Cu and Zn by a factor of 10 (the metals in a 50-ml sample were precipitated and redissoved in 5 ml of water). The LOD was 2 µg l−1 for each element.
Palladium was extracted into chloroform117 as the ion-association complex between the anionic thiocyanato complex and the cationic potassium complex with dicyclohexyl-18-crown-6. The procedure was applied to the determination of Pd in choroplatinic acid and rhodium chloride; after back extraction, the Pd was determined by FAAS. Silver was extracted into a liquid membrane118 made from trioctylamine, added as an emulsion to the solution to be analysed. Ammonia solution was then added as inner phase desorption agent, followed by vortexing and removal of the outer aqueous phase while the remaining emulsion was de-emulsified with a 2–4 kV spark. The Ag in the resulting aqueous phase, enriched 80-fold, was determined.
Three strains of sponges have been tested119 for their ability to adsorb metals (Cd, Co, Cu and Ni). Polystyrene microspheres modified with dithiocarbamate selectively removed mercury120 from solutions also containing cadmium, copper and lead, and octadecyl silica disks modified with tri-n-octylphosphine oxide separated uranium121 from thorium and other elements in soil samples.
Several speciation procedures, based on selective solid-phase reaction chemistry, have been developed. The pH dependence of the retention of SeVI and SeIV on algae was exploited as the basis of a method of speciation.125 At pH 5.0, SeVI was retained followed by retention of the SeIV at pH 1.0. The LOD (40 ml of sample solution) were 0.2 and 0.1 µg l−1 for SeIV and SeVI, respectively, and the procedure was applied to sediment and water samples. Inorganic As was speciated in drinking water126 by selective retention of AsIII as the complex with 2,3-dimercaptopropane-1-sulfonate on two C18 Sep-Pak cartridges, connected in series, followed by elution with methanol. Total As was determined as AsIII after reduction of AsV with L-cysteine (1% m/V in 0.2% HNO3) and retention on 5 cartidges in series. The LOD was 0.1 µg l−1. The modifier was nickel. Apparently, both CrIII and CrVI were retained from pH 7.0 solution by “acidic activated” alumina.127 The alumina was pretreated with 1 M HNO3, followed by 1 M NH3 solution, and equilibrated with NH4Cl–NH4OH buffer at pH 7.0. The analytes were sequentially eluted with 1 M HNO3 (CrIII) and 1 M NH3 (CrVI) solutions. The modifier was magnesium nitrate. Heme Fe in pork was determined128 by extraction with 80% acetone acidified with 2% HCl, retention (after dilution of the acetone content to 40%) on a C18 Sep-Pak cartridge and elution with 80% acetone. Recoveries, based on spiking several amounts of hemoglobin and myoglobin, were nearly 100%. Antimony(III) was determined129 in spring- and sea-water samples (pH 2.5) after retention on silica gel loaded with pyrocatechol violet and trioctylmethylammonium chloride and elution with 0.1 mol l−1 HCl. The interference from iron was masked by reduction to FeII with ascorbic acid. The procedure had been previously used for the preconcentration of Sn in sea-water;130 the analyte from a large sample volume (adjusted to pH 5) was eluted with a 10- to 20-times smaller volume of 1 mol l−1 HCl solution.
A number of liquid–liquid extraction procedures have been devised. Lead was determined in blood, homogenized by sonication,131 after extraction of the APDC complex into IBMK. A 20 µl subsample was atomized from a W-coil for an LOD of 0.6 µg l−1. The spectrometer contained a Pb HCL and a charge-coupled detector. In comparison with the results obtained for the analysis of 20 samples by graphite furnace AAS, the W-coil results were 8% low. The same extractant and solvent separated132 As from food samples; the modifier was a mixture of magnesium and nickel nitrates and the LOD was 2 µg l−1. The same solvent extracted the 8-hydroxyquinoline complex of Al133 from food sample digests to give an LOD of 90 µg kg−1. For the determination of Cs,134 the analyte was extracted as its 18-crown-6 complex into chloroform. The organic phase was mixed with a solution containing 0.1% KNO3 and 2% NH4NO3 as modifier with injection of 10 µl of the mixture into the furnace. Indium was extracted135 as the complex with 5-sulfato-8-hydroxyquinoline into carbon tetrachloride (or xylene) from a solution, adjusted to pH 5, containing zephiramine. The analyte was back extracted into concentrated HNO3 followed by dilution with Pd modifier solution.
Chitosan (the deacylated derivative of chitin), an amine polysaccharide containing amine and hydroxy functional groups that has metal complexing abilities, collected Cr136 from water samples adjusted to pH 3.0. The collection efficiency for CrVI was 100%, whereas that for CrIII was only 20%; however, as the CrVI could be selectively eluted with 0.1 mol l−1 NaOH solution, both species could be determined in a two-stage procedure in which total Cr was determined following oxidation with persulfate, and CrIII was determined by difference. The LOD for CrVI was 0.05 µg l−1. Soluble chitosan was precipitated137 at pH 8 from up to 1 l of solution and the co-precipitated Ru determined after dissolution in 1 ml of 1 mol l−1 acetic acid. The LOD was 0.06 µg l−1. The ability of hydrated iron(III) oxide to coprecipitate Mn, Pb and Zn was enhanced by the addition of dithiocarbamates.138 The LOD for Mn was 0.02 µg l−1.
A liquid–liquid extraction procedure was devised150 for the determination of Th and U in apatite minerals with introduction of the organic solvent (diisobutyl ketone) into the plasma for measurement by ICP-OES whereby the interference from calcium was overcome. The analytes were complexed with 1-phenyl-3-methyl-4-trifluoroacetyl-5-pyrazolone (pH > 2). To remove the interference of plutonium in the determination of 45 analytes by ICP-MS and ICP-OES,151 the matrix element was retained on an anion-exchange resin. Of the analytes, only Ag, Th, Tl and U were retained to any extent by the resin, but these could be readily determined by ICP-MS without matrix separation.
Liquid pumped through a 50 µm jet nozzle emerged in a thin stream that was directed into a heated tube mounted in an air–C2H2 flame,178 where impaction on the tube wall produced droplets fine enough to give rise to a significant atom population within the tube. This beam injection flame furnace AAS had LODs 6–200 times better than for conventional nebulization for 17 elements (Ag, As, Au, Bi, Cd, Cu, Hg, In, K, Pb, Pd, Rb, Sb, Se, Te, Tl and Zn). For 50 µl samples the throughput was 240 h−1. The addition of thermospray (by positioning the nozzle tip in the flame179) gave slight improvements in sensitivity. The LODs for 10 µl samples for Cd, Cu, Hg, Pb and Tl were 1, 5, 20, 10 and 5 µg l−1. The effect of replacing the Pt–Ir nozzle of a hydraulic high-pressure
nebulizer (HHPN) system with a diamond nozzle fixed in a threaded titanium socket was studied.180 A HHPN delivered cellulose bleaching liquors181 for analysis by FAAS. The system was used in the determination of Cu, Fe and Mn. The LOD were 2–3 times better than those of pneumatic nebulization. Cigarettes were analysed for Cr by FAAS with HHPN introduction.182 Total Cr (as CrVI) was determined following ashing at 800°C,and CrIII was determined by selective SPE on C18 in the presence of potassium hydrogenphthalate. During smoking, between 0.8 and 1.2% of the Cr gets into the smoke in the form of CrVI.
Several research groups have reported the nebulization of microemulsions. A “remarkable sensibilization” effect has been reported183 for the determination of Pb in copper and waste water. Milk powder was analysed for Fe, Mg and Zn184 and for Ca185 after formation of a stable emulsion with “emulsifier OP”. Methylcyclopentadienyl manganese tricarbonyl in gasoline was determined by FAAS186 with the formation of an emulsion by the addition of butanol, water and sodium dodecyl sulfonate. Edible oils were analysed for the P content by ICP-OES187 by the formation of an emulsion in water with a surfactant, ethoxynonylphenol. Calibration with aqueous standards was possible. Some interferences in FAAS due to the presence of chloride and sodium have been studied.188
The generation of the hydrides of Ag, Au, Cu and Zn from aqueous solution on reaction with sodium tetrahydroborate(III) has been reported194 by Sturgeon and co-workers. The elements were detected by AAS with a quartz tube atomizer; losses during transport were minimized by using a short length of Teflon tube. The HG-AAS determination of Cu, in the presence of trace amounts of o-phenanthroline, has been applied195 to the analysis of hair and rice RM. The LOD was 2 µg l−1. The same research group also reported the HG-AAS determination of Au.196 This time the solution contained DDC and the LOD was 20 µg l−1. A quartz tube atomizer was used in each case. The determination of Ni by ICP-OES following HG has been reported. The LOD was 0.5 µg l−1. Again, the transfer tube length was kept as short as possible. Phosphine was generated197 by passing the aqueous sample solution containing phosphate through an incandescent “copper silica tube”. The LOD was 2 µg l−1. Sulfide was determined by CV molecular absorption spectrometry of radiation at 202.6 nm from a Mg HCL, following HG.198 The LOD was 20 µg l−1 and some steels were analysed, though it was not clear how H2S was generated from these solid samples.
Electrolytic HG with a flow through cell incorporating a reticulated vitreous carbon cathode was applied to the determination of As, Se and Sb by quartz tube atomization AAS.199 The large cathode surface area (120 cm2) resulted in generation efficiencies of >90%. The volume of the cell was only 390 µl, allowing FI as well as continuous flow introduction. The LOD were 5, 2 and 1 µg l−1, respectively; this is still significantly poorer than the LOD which can be obtained by conventional HG based on reaction with borohydride, even for FI introduction. A cell consisting of platinum electrodes in compartments separated by a Nafion membrane200 generated H2Se for subsequent determination by AAS in a quartz tube atomizer heated by an air–liquid-petroleum-gas flame. The LOD was 10 µg l−1. A batch electrochemical HG procedure followed by reaction of the arsine with silver diethyldithiocarbamate and determination by molecular absorption spectrometry (Talanta, 2000, 52, 1007) had an LOD of 50 µg l−1. The cell contained a “pre-activated” graphite rod cathode. No signal was obtained for AsV. The moveable reduction bed hydride generator, developed by researchers at Xiamen, has been adapted201 as the interface between capillary zone electrophoresis (CZE) and ICP-OES for the determination of four As species. The LODs were 300–400 µg l−1.
The uses of Brindle's reagent (L-cysteine) in HG atomic spectrometry have been reviewed (38 references), though the reagent hardly deserves the epithet “novel” in the title of the article. The masking action of dithiophosphoric acid diacylester on copper for the determination of Bi by HG-AAS202 removed the interference from up to a 50000-times excess of copper. The use of fluoride as a selective complexant allowed speciation of Sb by FI-HG-AES. At high concentrations of fluoride (100 mg l−1) in the presence of 1.2% potassium iodide the signal from SbV and SbIII was suppressed,203 allowing determination of trimethylantimony oxide (TMSbO), whereas in the presence of fluoride but absence of KI, only the signal from SbV was suppressed.
The “simultaneous” determinations of several hydride forming elements have been reported. A sequential injection procedure with AAS detection was devised for the determination of As and Hg in water with LODs of 0.07 and 0.2 µg l−1, respectively. A four-step procedure was developed204 with a programmable intermittent vapour generation reactor for As, Bi, Cd, Ge, Hg, Pb, Sb, Se, Sn, Te and Zn by FI-AFS with hydrogen diffusion flame atomization. The LOD were <8 ng l−1 for Cd and Hg, <90 ng l−1 for As, Bi, Ge, Pb, Sb, Sn and Te and <5 µg l−1 for Zn. The determination of As, Bi, Hg, Sb and Se in human hair by continuous flow HG-AFS (hydrogen diffusion flame atomization) following microwave digestion with HCl and H2O2 has been described.205
Drinking water was analysed206 for As and Sb by the same technique. A preliminary reduction step with KI allowed distinction of AsIII and SbIII from the total element contents. The technique has also been applied to the determination of these elements in brass.207 An AFS procedure was developed for the determination of As, Bi, Hg, Sb, Se and Te208 with, apparently, LOD of <1 ng l−1. The analysis of lead ingots for As, Bi and Sb by HG-AFS involved209 removal of the majority of the lead by precipitation with chloride and the addition of a thiourea–ascorbic acid mixture to adjust the oxidation states. A procedure for the simultaneous determination of As, Bi, Se and Sn, and some non-hydride-forming elements, by HG-ICP-OES without a gas–liquid phase separator
has been described.210 The presence of calcium, magnesium, potassium and sodium suppressed the analyte signals. Arsenic, Bi, Sb and Sn were determined in steels by HG-ICP-OES211 with calibration against aqueous standards. The LODs were 0.1–0.2 µg l−1. A batch HG-ETAAS was devised for the determination of AsIII and AsV, BiIII, SbV and SeIV.212 The procedure was applied to the analysis of pure and brackish water samples for which quite different optimized conditions were obtained. The optimized trapping temperature for pure water was 600°C and for brackish water was 150
°C. The reaction vessel temperature was also controlled, and the optimized values were 16 and 0
°C, respectively.
A sequential injection system224 for the determination of total inorganic As delivered sample, borohydride and acid–ascorbic acid–KI solutions to a gas–liquid separation cell. Detection was by FAFS, and the LOD was 0.7 µg l−1 with a throughput of 33 h−1. Fumaric acid food additive was analysed by HG-AFS225 with an LOD of 0.2 µg l−1, the same LOD was obtained for the analysis of lead ingots209 and for brass.207 Values of between 0.005 and 0.02 µg l−1 were obtained206 for the analysis of drinking water by continuous flow HG-AFS, though LODs down to <0.001 µg l−1 have been reported.226
An argon MIP with AES detection227 gave an LOD of 5 µg l−1 for a power of 100 W when As was introduced from a continuous flow HG system. The stability of the plasma towards H2 was improved when O2 was introduced into the outer tube of the torch. An LOD of 0.5 µg l−1 was obtained228 by HG-ICP-AES in the analysis of iron and steels. The sensitivity was improved 1–2 orders of magnitude compared with that of conventional nebulization. A somewhat lower value (0.2 µg l−1) was obtained211 by HG-ICP-AES for the analysis of the same matrix, though the judicious use of Brindle's reagent allowed calibration against aqueous standards. A HG apparatus for the determination of As in gasoline has been developed,229 although the final quantification was by coulometric titration.
Several multi-element HG schemes have been devised which include As in the list of analytes. An electrolytic method has been developed for the determination of As, Sb and Se.199 An AFS method for several elements in hair has been devised,205 and an ETAAS procedure for As, Bi, Sb and Se has been optimized with the aid of statistical software.212
With regard to the determination of mercury species in air, an interlaboratory study involving four laboratories from North America and seven laboratories from Europe255 evaluated recently marketed automated analysers and obtained good agreement with results from a procedure based on noble-metal amalgamation AFS. However, agreement was less good for the results for the analysis of particulate material. The Hg in the atmosphere near Kagoshima City was measured256 daily for one year by collection on a porous gold collector followed by AAS. It was concluded that the nearby Sakurajima volcano contributed to the Hg in the air. The performance of a method based on this procedure has been evaluated257 in one German laboratory and in another258 two commercial automated instruments were compared. The first, based on AFS, had two gold absorbers in parallel, which
were switched every 5 min. The second, based on AAS, had two gold absorbers in series. The Hg collected for 6.5 min on the first was transferred to the second and further desorbed for determination, which required a further 3.5 min. Both instruments performed satisfactorily in the tests. The specifications and features of five instruments for the continuous monitoring of Hg emissions from waste incineration have been described.259 Six further instruments under development were also considered and compared with multi-metal continuous emission monitors. The determination of a species referred to as “gaseous divalent mercury” has been described.260 The species was collected on KCl coated denuder tubes and thermally desorbed at 450°C. The species was converted to an atomic vapour at 900
°C and collected on a gold trap for subsequent determination by AFS. Apparently, the
composition of mercurous chloride vapour (what might be referred to as gaseous monovalent mercury) is not well known.261 Application of AAS showed that at temperatures below 473 K mercurous chloride is completely dissociated into mercury and mercuric chloride. Atomic spectrometry techniques monitored262 the efficiency of several activated carbon and other sorbents for the removal of mercury from flue gases.
Organomercury compounds were decomposed263 with ozone prior to determination by AAS, thereby decreasing the amount and number of reagents needed. The reductant was tin(II). Urine was treated264 with concentrated HNO3 prior to determination of inorganic Hg by CV-AAS with tin(II) as reductant. The LOD was 0.1 µg l−1, and results similar to those of a method in which samples were decomposed with boiling KMnO4–H2SO4 were obtained. However, if the sample contained organoHg compounds the latter procedure was needed.265 For the determination of Hg in river water266 a remote monitoring device has been developed. Samples were treated on-line with a bromide–bromate digestion mixture and irradiated with UV light. Calibration standards of 50 and 100 ng l−1 were prepared by sequential injection dilution of stock 500 and 1000 ng l−1 solutions. With only UV digestion Hg was determined in sea-water267 with a LOD, after amalgam trapping, of 0.5 ng l−1 by AAS. The same system was used to monitor the Hg in river water.268 Again UV digestion was employed, as it was found to be superior to chemical and microwave-assisted procedures.
A field-screening procedure for the determination of Hg in soils269 involved thermal desorption of the Hg onto a gold collector with further thermal desorption onto a gold film Hg vapour analyser. Results comparable to those of AAS were obtained. A sequential injection AAS procedure was developed218 for the determination of As and Hg in geochemical materials. The LOD for Hg was 0.2 µg l−1. For the analysis of biological and environmental samples a procedure based on pyrolysis in a combustion tube at 750°C in the presence of O2 has been developed.270 The resulting mercury was trapped by amalgamation with gold and subsequently desorbed for determination by AAS. The concentrations in the materials examined ranged from 0.035–7.1 mg kg−1. Much the same approach has been used for the analysis of
crude oil.271 With AFS detection after amalgam preconcentration, the LOD was 0.2 µg kg−1. The procedure was also applied to the analysis of biological, environmental and general merchandise materials. For the analysis of adipose tissue,272 samples were first lyophilized to facilitate homogenization and the accurate weighing of 50 mg samples. Closed-vessel, microwave-assisted digestion with HNO3–H2SO4 was followed by oxidation of the liberated mercury species with potassium bromate–bromide reagent. Finally, Hg vapour was produced on the addition of SnII and quantified by AFS. The solution LOD was 2 ng l−1, and in the solid was 1 µg kg−1. Chinese workers have developed a “steam generator” AAS procedure273
for the analysis of feeds. The LOD was 0.2 µg l−1. Problems in the analysis of plant material and biological tissues have been discussed.274 A positive bias in the AAS method was caused by the presence of nitrous oxides with absorption bands close to the Hg wavelength. The problem was overcome by the addition of sulphamic acid prior to injection into the FI manifold. The problem only arose if HCl was present in the digest and was only seen when HNO3 and H2O2 were used. Fish tissue was analysed by CV-AAS following digestion with H2SO4–HNO3 in a sealed vessel. Following a 5-min purge with air, Hg0 was generated by the addition of tin(II). The method was validated by the analysis of two NIST SRMs (1577a Bovine Liver and 1566 Oyster Tissue). The LOD in fresh fish was 5 µg kg−1.
Brazilian workers analysed various fish from a river basin area where there is intense gold mining activity.275 Concentrations in carnivorous species ranged from 110 to 2300 mg kg−1, whereas in the detritivorous, herbivorous and omnivorous species, the concentrations ranged from 3 to 310 mg kg−1. A CV-AAS procedure for Hg in zinc ore276 has been developed. The dried, powdered sample was first treated with HCl to remove H2S, and then with HCl and HNO3. A reversible intermittent-flow FI procedure was developed277 for the analysis of sediments and vinasses (the residues remaining after distillation). In the forward direction, Hg vapour was carried through the cell, and in the reverse direction solution remaining in the reaction vessel was discarded. The procedure was extended to include SPE preconcentration278 A FI system containing a column of Chelite-S cationic resin preconcentrated Hg ions from solution and allowed a number of components of the agro-industrial samples to be discarded. The retained HgII was eluted with tin(II)–HCl directly into the gas–liquid separator. The LOD was 0.8 ng l−1 for 120 s loading (10 ml sample). A variety of silica gel materials, modified with thioacetal derivatives, were evaluated279 for their ability to retain metals. It was found that the materials had relatively high selectivity for HgII and were able to selectively remove Hg from spiked sea- and tap water samples at concentrations of 1 and 10 µg l−1. The same research group also evaluated silica gel modified with dithizone.280 A 20 ng l−1
natural tap water was preconcentrated by a factor of 200, allowing analysis by CV-AAS after elution. Preconcentration on a polyamide resin281 formed the basis of a procedure for the analysis of waters by CV-AAS. The eluent was 2% thiourea.
Mercury species in soils were determined282 by AAS on the basis of selective sequential extraction with 1 mol l−1 ammonium acetate, 1 mol l−1 hydroxylamine hydrochloride in 25% acetic acid, 0.1 mol l−1 hydrochloric acid, 0.5 mol l−1 NaOH and 8 mol l−1 nitric acid. Capel et al. have shown283 that ultrasonic irradiation at room temperature will convert organoHg compounds to inorganic mercury, which could then be determined by FI-CV-AAS. Interestingly enough, the presence of HCl was required to obtain complete conversion; in the presence of HNO3 or H2O2, oxidation was incomplete. In the presence of 1 mol l−1 HCl, complete conversion was obtained in 3 min on the application of 20 kHz ultrasound from a 100 W probe device operated at 40% sonication amplitude. The mechanism was thought to involve OH radicals but could tolerate species acting as scavengers at concentrations up to 100 mg l−1. Organic material equivalent to a chemical oxygen demand of 1000 mg l−1 could also be tolerated. The procedure was applied, in combination with selective reduction, to the determination of inorganic and total Hg in simulated wastewaters and spiked environmental waters. OrganoHg compounds were extracted284 from soils by 0.1 mol l−1 ammonium acetate–sodium acetate buffer at pH 4 and converted to the corresponding hydrides on the addition of 6% potassium borohydride solution. The hydrides were collected by headspace solid phase microextraction and separated by GC with AAS detection. The absolute LODs were 16, 12, and 7 ng for methyl-, ethyl- and phenylmercury, repectively. The method was applied to soils from orchards and wheat fields with concentrations ranging from 0.04 to 0.64 mg kg−1. Russian researchers285 have devised a liquid extraction procedure for the speciation of Hg compounds in river water. OrganoHg compounds were extracted into chloroform leaving inorganic Hg in the aqueous phase (determined by CV-AAS). Alkylmercury chlorides were back extracted with 0.1% cysteine solution. DialkylHg compounds and organic complexes were hydrolysed with HCl to alkylHg and inorganic Hg, respectively. Rio-Segade and Bendicho discriminated between inorganic Hg and organoHg species in mussel tissue on the basis of reactivity towards tin(II). It was assumed that only inorganic Hg gave a volatile derivative with this reagent. Total Hg was determined after microwave-assisted digestion with nitric acid and hydrogen peroxide. The authors concluded that the mercury in the mussels was methylmercury, which means that no signal was observed for inorganic Hg; however, the extraction involved ultrasonic irradiation in the presence of HCl which, from the findings of Capelo et al. discussed above,283 might be expected to lead to conversion of organomercury to inorganic mercury. Clearly there are some aspects of the speciation of mercury, and of the reactions of methylmercury, which are not yet fully understood.
An approach to the evaluation of data quality in direct solid sampling for ETAAS was described by Lucker et al.286 Calibration was performed close to the limits of decision, detection and quantitation in order to estimate the lower limits of these parameters. The values obtained were 2–20 times lower than those obtainable by decomposition methods. In two papers in German, Kurfuerst287,288 discussed accuracy in solid sampling ETAAS. Random and systematic errors were discussed in relation to the results of a collaborative study. Systematic error due to uncertainties in the calibration and in the sample measurement and also in the quality control implemented in the participating laboratories was highlighted. It was concluded, however, that the accuracy achievable by direct solid analysis was no poorer than that obtained using liquid sample introduction after acid decomposition.
Slurry sampling has become a widely accepted form of solids analysis. The substantial body of work that exists on slurry sampling ETAAS is of value in itself and as a basis for slurry sampling for ETV-ICP-AES and for ETV-ICP-MS. However, in the period covered by this review, few descriptions of new applications have been published. The analysis of inorganic materials that are difficult to digest is a perfect application for the slurry approach, but also highlights one of the disadvantages caused by introducing high masses of non-volatile matrix into the furnace. Belarra et al.289 considered the analysis of inorganic samples when the analyte and matrix have similar vaporisation temperatures. The use of non-stop gas flow conditions was proposed as the preferred method of compensating for potential gas phase interferences.
Several reviews have been published during the period of this review on various aspects of the use of glow discharge (GD) atomic spectrometry for solid sampling. A review of GD sources by Caroli et al.290 provided a historical description and concentrated on the application of these useful techniques to the analysis of solid samples. In another paper, the major developments in cathodic sputtering techniques were reviewed by Gough (36 references).291
Several reviews of the use of GD-AES in particular application areas were published. Environmental analysis was reviewed by Baude et al.,292 surface and depth profile analysis were described in an article by Deraed and Backhaus,293 and the direct analysis of polymeric materials using rf-GDS was discussed by Marcus.294 A patent by Marcus describes a method for introduction of airborne particulate matter into a low pressure plasma.295 Particle beam sampling allows real time analysis in either continuous mode or after collection for a specified time.
An interesting application of GD-AES for the analysis of deposited SiO2 films was described by Dorka et al.296 The effect of interference due to reflectance from the Si substrate through the transparent SiO2 layer causes intensity oscillations of certain spectral lines. Investigation of this intereference allows measurement of film thickness and refractive index, measurements normally made using ellipsometry. Together with the chemical characterization obtained using GD-AES, the technique provides a unique combination of analytical measurements.
The effect of Ca, Sr, Li, Mg and Al matrices on plasma conditions in laser ablation ICP-AES was investigated by Chan et al.297 The ratio of ionic to atom line intensities of a test element were monitored to evaluate the effect of sample matrix on a number of excitation characteristics including plasma temperature and electron number density. The matrix effect of Sr, Ca and Mg was more pronounced than that of the more easily ionizable elements. The authors concluded that the matrix effects might be due to a mechanism involving the 2nd ionization level of the matrix elements.
It is now established that key parameters in ETAAS, namely gas phase temperature, source beam intensity and analyte vapour distribution, are generally non-uniform over the absorption volume. This non-uniformity only leads to analytical error when atomization conditions differ for samples and standards; unfortunately, this is usually the case. Matrix modifiers are conventionally used to minimize differences in standard and sample atomization conditions. Gilmutdinov and co-workers continue to provide evidence for the benefits of spatially resolved ETAAS analysis based on the use of a linear solid state detector.300 In this work, an instrument was described that allowed spatially and temporally resolved detection. Cd and Pb were determined in a range of SRMs to evaluate matrix effects. Large variations in analyte distribution in the different matrices were observed, illustrating the advantages of analysis using spatial resolution. L'vov also endorsed the application of spatial resolution, and cited the approach as a significant step towards absolute analysis in ET-AAS.301 Measurement of absorbances integrated in terms of both wavelength and time is seen as a parallel requirement for standardless analysis. Continuum source and boosted-output hollow cathode lamps are compared for this task, and L'vov concludes, at this time, that the simpler approach, technologically, is the use of line sources.
Schlemmer and Petek302 described the use of a class 100 clean room cabinet for ET-AAS. Optimized extraction conditions were established with special emphasis on particle emission from the furnace during the pyrolysis and atomization stages. When optimized flow conditions were used, class 100 conditions were maintained in the vicinity of the autosampler regardless of furnace emissions. This helped avoid blank fluctuations for elements such as Ca, Fe and Zn, where blank contamination is usually the limiting noise source.
As discussed in previous Updates (1998, 1999) Torsi et al. have promoted their concept of standardless analysis in ETAAS using a specially designed atomizer system that is ideal for the determination of volatile elements. Rapid atomization is an essential part of the approach, which assumes uniform appearance in the gaseous phase of all atoms present in the condensed phase. Re-design of the atomizer system, permitting higher atomization temperatures, has now allowed the determination of elements of medium volatility.303
An overview of the use of permanent modification in ETAAS was provided by Tsalev et al.304 This approach to modification currently receives the most attention in the literature. The benefits include increased sample throughput with ‘fast’ programs, reducing reagent blanks, preliminary elimination of unwanted modifier components, compatibility with on-line equipment and in situ enrichment. There are disadvantages to the use of permanent modification, and these are cited as: poor reproducibility in the treatment technologies; impaired efficiency compared with modifier addition to each sample aliquot; relatively short lifetimes; limitations imposed on temperature programs, applicability to relatively simple sample solutions rather than to high-salt matrices and acidic digests; and the potential to overstabilize. The authors concluded that in the niche areas of coupled VG-ETAAS techniques, analysis of organic solvents and extracts, concentrates and fractions obtained after enrichment and/or speciation separations and direct ETAAS determinations of highly volatile analytes in relatively simple sample matrices, permanent modification has the most promising future.
The two common vapour exit points in an ETV for ICP-AES were compared by Kantor.306 In one version the vapour streams upward through the transverse hole of the longitudinally heated tube (UPS), while with the other version the vapour streams towards one of the ends of the tube (ENS). When pure argon is used as the internal gas, the ENS is less effective than UPS for medium and low volatile elements because of the possible condensation of the vapour on the cooler parts of the graphite tube, close to the outlet end. The ENS, however, is more effective for the volatile elements due to the faster mixing of the hot vapour with the carrier gas within the graphite tube. When carbon tetrachloride is added to the internal furnace gas, the UPS and ENS provide similar transport efficiencies, as no condensation of chloride vapours takes place. A significant increase in the relative sensitivities for volatile elements (Cd and Zn) was
obtained with carbon tetrachloride additive, due to the carrier effect of the particles formed from this halocarbon above 500°C. This kind of carrier effect was not observed when toluene was used as a gaseous phase additive.
Rybak and Salin used a closed-system for the measurement of transport efficiency in ETV sample introduction.307 The products of an ET vaporization event were collected in a commercially available gas sampling bag. The bag was then manually agitated and its liquid contents decanted and analysed by ICP-MS. Analyte transport efficiencies were found to be in good statistical agreement with those determined on the same ETV-ICP-AES system for a range of elements. Gras et al.308 investigated the limiting effect on transport efficiency of matrix load in ETV-ICP-AES. Several certified reference materials were analysed and the results obtained for samples with low inorganic matrix concentration were within 50% of the certified value. For samples with a high inorganic matrix concentration or high element concentrations the results were only qualitative, reflecting the transport efficiency changes due to matrix load.
An update333 to an earlier review of atomic spectra databases on the World Wide Web from the US National Institute of Standards and Technology includes a number of significant changes and a much expanded version of the NIST database. A fifth volume of “Advances in Atomic Spectroscopy” (291 pp.) has been published by JAI, Stamford, CT, USA, 1999,334 and there was a review in Japanese335 of Nobel prize-winning techniques applicable to the study of atoms at the atomic level, including scanning tunnelling microscopy and AFM. There are also tutorial reviews by Lagalante on AAS336 and on AES337 with just twelve and eight references, respectively.
A vacuum UV Fourier transform spectrometer, operating at wavelengths from 140 nm upwards (limited only by the cut-off of the CaF2 optics), was described by Griesman et al.339 Capabilities in several fields, including AES, were illustrated.
A method for measuring diffraction grating spacing with a stabilized laser in an interferometric system was patented in Korea.340 The standard deviation of measurement at 463 nm spacing was given as 0.05%.
There were several patents concerning atomic emission spectrometers. Two of these from Shimadzu341,342 relate to portable analysers in the steel industry and are based on stepwise variable positioning of the diode array/CCD image sensors. In one case, the range is expanded by using a second spectrometer equipped with a UV sensitive PMT in order to measure C, P and S at wavelengths down to 178 nm. A Chinese patent343 for a spectrometer with wavelength modulation specified a quartz plate mounted at the entrance slit on an electromagnetic driving device which gave a maximum rotation angle of ±15°. The signal was displayed via a PMT and phase-locked amplifier.
A rapid scanning spectrometer was developed by Kitagawa et al.344 They replaced the sine bar of their grating instrument with a laser encoder and the pulses so produced when the shaft was driven were counted to detect the grating angle and equivalent wavelength. Wavelength accuracy and precision were ±0.26 nm and ±0.013 nm, respectively. The instrument was successfully used for multi-element measurements, background correction and characterization of an ICP.
Two patents have been lodged for features which make a spectrometer especially suitable for use with plasma sources. From Israel comes a shutter assembly345 which is synchronized with the source pulse frequency. This prevents matter ejected from the source from accumulating on the optics. A Japanese patent346 claims the incorporation of an “independent diaphragm” which, too, prevents adhesion of impurities and consequent corrosion at what are taken to mean the entrance optics of the spectrometer.
Commercial échelle grating spectrometers now available for ICP-AES have been reviewed by Olesik.347 The basic physics and operating principles of such spectrometers are covered and there is special emphasis on the use of échelle gratings in conjunction with charge transfer detectors (10 references). A 22-channel Paschen–Runge mounted spectrometer with an ICP source and individual PM detectors has been constructed in-house and described by Shukla et al.348 It was observed that each element emits in a different region of the plasma and this effect had to be studied in order to achieve best detection limits, which were typically in “the sub-ppm levels” for trace elements in water.
A scanning échelle spectrometer having dynamic wavelength stabilization, with detection and background correction using a specially designed, back illuminated CCD, was described by Becker-Ross et al.349 The double monochromator uses a CaF2 prism to preselect the equivalent of a single order of the spectrum for entry to the échelle. All wavelengths between 165 nm and 900 nm can thus be unambiguously accessed. The wavelength is effectively stabilized, even over an ambient temperature range of 15–35°C, by simultaneous monitoring and measurement of a Ne spectrum. The design results in very low levels of stray light. Excellent detection limits were claimed for P, Tl and Se in the low wavelength region.
Hitachi has patented an AA spectrometer350 in which the entry window is tilted at an angle between 3° and 60°, but typically about 10°, so that reflected light from the ETA does not enter the measurement channel. Two papers from Zimmermann et al.351,352 described their micro-flame ionization detector and easily portable micro-flame spectrometer. At the heart of both devices is a miniaturized burner unit which utilizes an oxy-hydrogen flame, the gas mixture for which was generated as required in a small battery-operated electrolysis cell. The system is still under development, but present performance shows detection limits only about 100× those expected with conventional systems.
Two papers from Chinese workers353,354 describe work leading towards the application of narrow band acousto-optic tunable filters as wavelength selectors in AAS. Atomizers cited in the two papers were ETA and MIP, respectively, but the only application reported was the measurement of Na using the MIP, with a detection limit of 0.23 µg ml−1.
Patents from Italy357 and from Japan358 described improvements to the ICP torch and sample introduction system, and another from Japan359 used a stream of coolant gas for protecting the spectrometer from source-generated heat.
The construction of a microplasma cell for measuring metals in solution was described by McCallum et al.360 The plasma was generated as a bubble of noble gas confined to a particular location within the analyte solution by means of a low power applied acoustic field. It could function on a few µl of gas for several hours at a time and offered a virtually non-destructive technique with very low running costs. In preliminary studies on Cu, K and Na, calibrations extending over three orders of magnitude were linear. The extension of the ICP's wavelength range down to 120 nm in the far UV region for measuring halogens has been successfully achieved by Tyler.361 Both oxygen and quartz must be eliminated from the optical train, the latter being replaced by CaF2 or MgF2 optics.
Bilgic, Engel and others, in several papers, gave details of a low power 245 GHz miniaturized MIP,365,366 the source of which was generated inside a 0.9 mm gas channel “grown into” a 30 mm square sapphire wafer. Power was transferred by means of microstrips, which were delineated by a standard photolithographic process and designed both for compactness and high field strength in the plasma channel. The microwave input power was 5–30 W for He and 1–10 W for Ar, with gas flows of 50–1000 ml min−1. An earlier paper367 described the application of the device (at that time constructed on a wafer of fused silica) to the measurement of Hg in soil leachates using a flow injection cold vapour technique for which a detection limit of the order of 50 pg ml−1 was claimed.
Chinese workers investigated the difference between end-on and side-on viewing of a MIP-AES system.368 A shielding gas was passed between the torch and entrance optics to prevent damage from heat and analyte sputtering. End-on viewing proved to be better for multi-element applications as there was no need to compromise operating conditions. It also resulted in better detection limits, improved sensitivity and freedom from matrix effects. However, a disadvantage was that the hot tail flame caused the dynamic range to be narrower.
A ring discharge plasma was used by Wrembel370 to measure Hg at ultra-trace levels in “different bodies” in the Baltic Sea area, but no details of the system were given in the abstract available. Workers in the University of Florida described the generation of a plasma in Cs vapour by a continuous wave resonance excitation process.371 A highly ionized plasma is produced when a heated Cs vapour cell is subjected to a continuous wave Ti:sapphire laser of relatively low irradiance, tuned to the 852.110 nm Cs resonance transition line. The paper presents preliminary results on the coupling of low irradiance CW laser light to a plasma via a resonance process. From Minsk, Belarus, comes the suggestion372 that a compact quasi-steady-state plasma accelerator has the potential for use as a source for AES. Both gas discharge and erosion types of accelerator were studied, generating a compressed plasma of specific composition.
Eijkel et al.376 described an atmospheric pressure glow discharge in a nanolitre-sized chamber formed on a microchip. So far, this has only been used for molecular emission (e.g., methane was detected at 10−14 g s−1) but it is believed that the device will be suitable for incorporation into microanalysis systems.
“Metastable transfer emission spectroscopy” was the term used378,379 to describe a method for Cd determination in which 5 µl of acidified sample solution was placed upon a W-coil and atomized by passage of a current pulse from the discharge of a capacitor. The vapour was transported and mixed with active N2, which had been generated in a microwave discharge. A high degree of emission was observed for a period of 60 ms. The detection limit for Cd in emission at 326.1 nm was less than 5 pg and was unaffected by the presence of higher concentrations of co-existing elements, particularly Pb.
A Ta coil atomizer for AAS was described by Bai et al.381 Ar was used as the shielding gas and 3 µl aliquots of sample solution were atomized, giving characteristic concentrations for Cd, Cu and Zn of 2, 30 and 7 ng ml−1, respectively, and good results for standard materials. Shimadzu sought a patent382 on improvements to the sample holder of their flameless AA spectrometer in order correctly to maintain its position in the radiation beam.
A theoretical study of the electric field distribution in the cathode fall region of a spherical HCL was made by Mixon387 and, in two papers, Gil'mutdinov established the spatial distribution of the radiation emitted by single,388 double and combined389 HF discharge lamps of the type used in AAS. A “control” lamp, containing only Xe, the fill gas, was used to compensate for radiation of this gas. Considerable non-uniformity of radiation was shown to occur, thus, the recorded intensity is always low, and this fact has to be taken into account when AA signals are being evaluated.
Analyte | Matrix | Comment/LOD | Ref. |
---|---|---|---|
As | Environmental samples | Comparison with ICP-MS for As speciation. Results for both methods were comparable | 391 |
As, Hg, Se | Coal | Acidified H2O extraction, vapour generation. RSDs: 4.7–12% | 392 |
As, Bi, Hg, Sb, Se, Te | Food/environmental samples | Evaluation of commercial instrumentation | 393 |
Hg | Fish | Review of atomic spectroscopic techniques | 394 |
Hg | Natural gas | Evaluation of commercial instrumentation | 395 |
Hg | Sediment/sludge | Laser ablation of pelleted sample into commercial Hg detector | 396 |
Pb | Tea | Evaluation of commercial instrumentation | 397 |
Se | Tea | Evaluation of in-house developed instrument. LOD: 0.1 µg l−1 | 398 |
Se | Waters | Speciation by LC-UV derivatization-HG-AFS | 399 |
Se | Nutritional supplements/shampoos | Focused microwave acid digestion reduced all forms of Se to SeIV in 10 min. LOD: 1.4 µg l−1 | 400 |
A system for the preconcentration of Ag, Bi, Cd, Pb and Tl in a microcolumn crucible has been reported.390 The crucible serves as an electrothermal atomizer for AF and AA following preconcentration on DETATA adsorbents. The LODs ranged from 0.3 ng 1−1 (Ag) to 10 ng 1−1 (Tl). The method was accurate for the determination of trace elements in river and sea-water and in atmospheric precipitates.
Developments in solid state detectors for use in atomic spectroscopy have been reviewed by Pennebaker et al. (43 refs.).402 Particular attention was given to multi-element or multi-channel spectrometers. A review (70 refs.) of multi-element solid state detectors and their application to AAS has been prepared by Shelpakova et al.403 Methods of interpreting spectra and calculating the intensities of spectral lines were considered. Progress in charge-coupled devices and their applications in atomic spectroscopy has been the subject of a review with 83 references by Zhang et al.404 Advances in the design of charge injection devices as detectors for atomic emission techniques have also been reviewed.405
The merits of CCDs include higher quantum efficiency and lower detector noise than PMTs and consequently they are increasingly used in both AAS and ICP-AES instruments. The design and use of some commercially-available CCDs has been considered by Ohta and Tanibayashi and a Hg spectrum obtained with a CCD linear sensor presented.406 The incorporation of a CCD into an ICP-AES instrument provided simultaneous background correction to sequential elemental analysis.407 There is an analytical need to make measurements of spectral line intensities from ICPs in the spectral range 120–190 nm, notably for the determination of Cl and Br. CCDs are not very efficient below 200 nm: however, a lumogen coating to convert UV into visible radiation improves sensitivity. A UV-enhanced CCD detector has been incorporated into an axially viewed ICP-AES spectrometer with a view to comparing the performance of the CCD with that of PM tubes.408 The LODs achieved with the UV-enhanced CCD compared favourably with those obtained with PMT detection.
All diode array detectors integrated on a single Si chip suffer to some extent from non-uniformity of sensitivity. This problem has been examined by Birkenshaw et al.409 A correction algorithm derived from spectra with known characteristics was developed to solve the problem. The incorporation of a position-sensitive photodetector into a commercial ETAAS instrument combined with a specially developed algorithm for processing and interpreting the results facilitated the analysis of complex samples.410 Software for instrument control and data processing for use in analytical atomic spectrometry is constantly being developed. The development of Windows-based software for multi-channel ICP-AES multi-element transient signal acquisition and processing has been described by Zhang et al.411 The communication program was written in VISUAL BASIC and smoothing of a noisy signal was by digital filtration. The advantages of this software include the easy transfer of data to other application programs for further data treatment.
Ultimately LODs in AAS are set by the shot noise of the light beams. Experiments by Schwob and co-workers have demonstrated a possible way of overcoming the shot noise limitation.412 The authors used an optical parametric oscillator to generate light beams in which quantum intensity fluctuations were correlated.
A kinetic model for an atomic absorption signal has been elucidated for the case of first-order kinetics.414 The formation energy of free copper ions and other kinetic parameters of the copper signal were determined from experimental data. The role of oscillator strength in atomic absorption spectroscopy has also been discussed,415 where a universal method for the determination of atomic lifetimes and hence absolute oscillator strengths of atomic resonance lines for almost any element in the Periodic Table was given.
Concentration measurements made over long time periods, as obtained in quality control charts, provided a possibility of gaining supplementary information about an analytical method.416 Quality control charts for soil analysis by AAS with measurements lasting for more than 1000 d were studied and Fourier spectra at low frequencies obtained. When no distinguishable uncertainty sources were present, the noise power spectral density dependence on frequency was calculated.
Atomic emission spectroscopy of a continuous atmospheric pressure microwave sustained plasma in an undiluted slipstream of stack exhaust was developed for real-time monitoring of EPA regulated hazardous metals.420 The detection limit in the plasma was found to depend on the pathlength between the plasma and the detection system. Detection limits using a long pathlength axial view versus short pathlength side views, at different points along the plasma column, showed a clear diminution of the signal for the axial view.
A fundamental comparison between non-equilibrium aspects of ICP and MIP discharges has been published this year.421 A conventional ICP torch and an axial injection MIP source were compared in terms of construction, geometry, dimensions, gas flow rates, operating frequencies and the size of the radial scale length which, together with the power density and pressure, determines the internal plasma properties such as the electron number density and the temperature. Good agreement was obtained with experimental values of temperature and electron number density for the Ar ICP, but for the Ar MIP, the predicted values were too low and too high, respectively, with respect to experimental values.
Helium–argon working gas systems in rf GD-AES have been described.424 The mixed gas plasmas were used to determine carbon in steel samples. It was found that for pure mixed gas plasmas, the excitations by helium could produce excited states of carbon more readily. The detection sensitivity obtained with the mixed gas plasma was enhanced compared with that of the 100% argon plasma.
Several papers this year have described argon glow discharges containing small quantities of hydrogen.425–428 The progressive addition of hydrogen causes intensity changes specific to individual lines of different species such as atomic and ionic copper and argon.427 New spectral features, such as emission bands of new compounds (hydrides) and a continuous background in the range ∼220–440 nm was observed. Introducing hydrogen externally in gaseous form or sputtered as a sample constituent causes very similar effects in terms of discharge processes. The addition of hydrogen has also been predicted to cause a drop in Ar ion and electron density and a change in the electron energy distribution function.425
The analytical performance of a novel high voltage neon plasma in GD-AES has been described this year.429 The use of singly ionized lines and atomic resonance lines was compared. Better sensitivity and detection limits were found for the former.
The majority of glow discharge systems currently operate in a continuous direct current (dc) or radiofrequency (rf) mode so that a steady state discharge is obtained. The amount of power normally applied to a glow discharge is 1–3 orders of magnitude smaller when compared with an ICP source. Increasing the power of the GD can easily lead to sample overheating; therefore, a pulsed high current mode is used so that high instantaneous power may be applied to yield an intensified transient signal.
Investigations have been performed using microsecond pulsed glow discharge atomic emission spectrometry to determine the effect of the glow discharge parameters, including applied pulse voltage, pulse width, pulse frequency and gas pressure on the resolution of depth profiling.430 All parameters were investigated and optimized and the ideal settings used to analyze the surfaces of a number of materials, including electroplated steel. Research has also been performed to investigate possible gains in signal to background ratio (SBR) and detection limits for a pulse glow discharge compared with a conventional direct current mode of operation.431 It was found that a high instantaneous pulse power gives an increase in the SBR while the use of time resolved detection also gives significant gains. It was also found that certain nearly non-conducting materials, which cannot be run in dc mode, could be run in the microsecond pulse mode.
A hybrid model has been developed for a microsecond pulsed analytical glow discharge432 consisting of a Monte Carlo model for fast electrons and argon species and a fluid model for slow electrons and argon ions, in the cathode dark space. The results of the model enabled calculation of electrical characteristics, electric potential distributions, argon ion density profiles and the rates of ionisation and recombination mechanisms. The results compared well with experimental data when available. Further work by the same group discussed the use of this three-dimensional modeling network to compare similarities and differences between dc and rf glow discharges.433–435
A comparison between Boltzmann plots of ionic vanadium lines using an Ar dc glow discharge plasma as well as an Ar ICP has been described.436 Local thermodynamic equilibrium conditions were not achieved, illustrated by departures from linear Boltzmann plots. The differences were attributed to resonance charge transfer collisions between vanadium atoms and argon ions.
The effects of drifting instrument sensitivity in GD-AES on instrument calibration have been investigated.437 It was found that the rate of drift might be faster than desired for adequate calibration of the instrument, resulting in invalid calibration data and analytical bias. A new drift correction method was used to enhance the utility of the technique for several applications.
The implications of water loading into the central channel of the plasma were considered when an ultrasonic nebuliser was used in axial view ICP-AES.170 Interferences occurred as a result of changes in the sample introduction system. Signal suppression was observed for the highest water loading and vice versa. The desolvation of aerosols in the USN saved plasma energy but influenced the excitation conditions by changing the thermal environment of the plasma. Interferences were minimized using solvent removal. A study was made of mineral acid interferences in the trace element determination of Fe, Mn and Zn by ICP-AES.441 Nitric and sulfuric acids were found to depress the emission signal of the analyte elements. The effects were neither caused by the change in aerosol flow rate nor the amount of aerosol reaching the plasma. The study of the influence of organic solvents on the characteristics of ICP-AES with the aid of flow injection has also been addressed in a publication this year.442
A number of papers have discussed the need for optimization of the operating parameters for improved signal and signal to background response. The influence of the power and carrier gas flow rate on the net signal and SBR values was studied using simultaneous measurements with multi-channel detection and axial viewing of an ICP with AES detection.443 Although no simple relationship could be suggested between experimental behavior and the line characteristics such as wavelength, ionization state and excitation energy, it was possible to find compromise conditions for most atomic and ionic emission lines where 80% of the net signal could be obtained. The same group went on to discuss the influence of power and carrier gas flow rate on time correlation between lines of the same element using the same multi-channel detection and axial viewing of the plasma.444 Both correlation between lines and improvements in the RSD values were calculated, not only to evaluate the link between correlated signals and improved RSD values, but also to examine the influence of the magnitude of the relative fluctuations of the signals. In a similar study, a systematic approach to optimum working conditions with ICP-AES has also been reported.445 The present status of axially viewed ICPs has been reviewed with special emphasis placed on the analytical performance of currently available systems.446
Various plasma sources are used for the analysis of samples and a way to characterize these plasma sources has been recommended.447 This involves the determination of the absolute intensities of several atomic lines of the main plasma. From these measurements, the population densities of excited states can be deduced and, via a Boltzmann plot, the electron temperature and electron number density can be calculated. The paper states that the method works well as long as the plasma is relatively close to equilibrium. The reason why spectrochemical plasmas can show strong deviations from equilibrium is discussed.
A semi-empirical equation to describe the electron number density in the analytical zone of the ICP as a function of rf power, height above the load coil and water flow rate, has been described by Nakamura et al.448 Results showed that electron number density linearly increased with increased rf power, exponentially decreased with increased observation height and decreased with increasing amounts of carrier gas and H2O molecules.
The evaluation of hydrogen line emission and argon plasma electron concentrations resulting from the gaseous sample injection involved in hydride generation-ICP-AES analysis has been reported.449 Line and background intensities, as well as the FWHM of the hydrogen gamma and delta lines, were measured. Electron number densities were estimated from Stark broadening of the line profiles.
Spectral interferences in the determination of trace elements in environmental materials by ICP-AES have been addressed.450 The referenced Q-concept of Boumans and Vrakking was used to quantify the spectral interference for several elements enabling the selection of analytical lines subject to the least interference and correction of the line for background interference without the necessity for a matrix blank. The relative spectral response of a commercially available ICP has been determined over a wide spectral range (190–900 nm) using overlapping sets of radiative branching ratios of several atomic and ionic species.451 Branching ratios determined from the ICP were compared to those previously published and good agreement was found.
The effect of charge transfer reactions on analyte excitation and ionisation in the inductively coupled plasma was studied by two independent techniques.452 The techniques were used to study the effects of charge transfer reactions on the third row elements Ca–Cu. With the exception of Cr and Mn, all of the elements studied showed positive evidence of excitation and ionisation by charge transfer collision with argon.
The measurement error values generated by the Abel inversion for Ar ICP diagnostics have been calculated.453 A simple mathematical model to simulate the conditions in the Ar ICP was used. The measurement errors were found to increase with increasingly greater peak position error. To keep measurement error values less than 5%, the peak position error must be kept at less than 1% of the plasma diameter. The effect of pulsations of a plasma torch on the radial distribution of temperature454 and the analysis of line widths to obtain improved Voigt parameters and optical thickness in the ICP455 have also been described in the literature this year.
The formation and fundamental characteristics of a free-running helium ICP have been described this year.456 The shape and physical appearance of this free running plasma differed from traditional He ICP sources. Although the free-running helium plasma appeared to be annular, high-speed video studies revealed that the plasma rotated. Rotational and excitational temperatures were calculated for the plasma along with the electron number density. These characteristics were directly compared to previously reported properties of crystal controlled He ICPs and Ar ICPs.
Fundamental properties of aerosol produced in helium by a direct injection nebuliser have been determined on a whole-field and spatially resolved basis using optical patternation and phase-Doppler particle analysis.457 The helium-generated aerosol was found to generally exhibit larger droplet sizes, lower mean droplet velocities and lower droplet number densities than aerosols produced with argon.
An arc two-jet plasmatron with an argon flow, predominantly giving a spark spectrum, was used as a source of excitation to evaluate spectral interferences in the atomic emission determination of rare-earth elements and yttrium in natural samples.459 Spectral interferences were quantified using emission line characteristics.
Fundamental studies of plasma parameters for an argon dc arc have been reported this year.460,461 Plasma temperature and electron density of the plasmas were measured by atomic emission spectroscopy using various plasma conditions. A radiative–collisional model was formulated for the arc near the cathode region where a departure from LTE was observed.
Oxygen/tetraethoxysilane (O2/TEOS) plasmas created in a low-pressure rf helicon reactor have been studied by atomic emission spectroscopy as a function of the rf power injected into the plasma.462 By varying the rf power, different molecular spectra were observed to give an indication of the composition of the plasma gas phase.
The effect of easily ionized elements in a dc plasma has been modelled and experimentally investigated.463 The introduction of EIEs in the U-shaped DCP decreases a ‘potential barrier’ due to an internal radial electric field, which allows the partially ionized analyte better entry into the hot plasma zone and produces enhancement of the analyte spectral emission. By comparing model predictions to experimental data, the validity of the model was tested. The main features of the effect were a large enhancement at small addition of EIE and almost equal enhancement of ion and atom line intensities. The results were believed to be applicable to the inverted Y-shaped DCP.
Results from further investigations into the temporal oscillations of analytical signals for elements in ETAAS from L'vov, Sturgeon et al.465 have been reported. Microgram amounts of Al, Mn and Yb oxides were slowly heated in both isothermal and non-isothermal furnaces. The signals obtained in both types of furnace were identical. The origin of temporal oscillations appears to lie with the primary sample residue, which was interpreted by the authors to provide support for the gaseous carbide reduction mechanism. The mechanism of the carbothermal reduction of FeO, CoO, NiO and Cu2O was the subject of an additional paper by L'vov.466 A thermal dissociation model based on the dissociative evaporation of the reactant with simultaneous condensation of the low-volatile product (metal vapour) was proposed. This model was used to interpret the kinetics of the reduction of the metal oxides by carbon. Analysis of literature data and comparison with theoretical calculations were used to show that the main kinetic characteristics of carbothermal reduction, including the initial decomposition temperature and activation energy, appear to confirm the proposed mechanism. In an extension of the conclusions found in the L'vov/Sturgeon paper,465 the observed temporal oscillations are considered due to condensation of metal vapour in the reaction zone and partial transport of condensation energy to the oxide.
Katskov and co-workers continue to publish interesting and apparently controversial results of their studies into nitrate vaporization in ETAAS.467–469 As discussed in last years ASU, Kastkov observed that the vaporisation of Mg(NO3)2 in a pyrolytic-graphite coated graphite furnace was accompanied by light scattering only when Ar was used as the purge gas and not when He was used, when a Ta-lined tube or a filter-furnace was used. This was attributed to an exothermic reaction between the large amounts of generated MgO and the graphite tube, leading to a higher gas temperature near the tube wall compared with the tube centre. This leads to higher vaporization rates at the tube wall into the cooler tube axis. Condensation at the axis leads to light scattering. At higher tube temperatures, scattering was attributed to condensation outside of the tube ends. L'vov, however, disputed this mechanism, claiming that mistakes in theoretical calculations and experimental interpretation led to erroneous conclusions.470 Katskov et al. responded to this criticism,469 claiming that L'vov had misinterpreted the experimental evidence and had applied principles based on the vaporization of small samples, which were therefore not applicable to the model used by Katskov.
In previous work (see J. Anal. At. Spectrom., 1999, 14, 1245–1285) it was established that in a furnace heated at a rate of ∼10 K ms−1, the processes of atomization and dissipation of the atomic vapor from the analytical volume are separated by time if the gaseous atoms do not interact with the wall of the furnace. The absorbance front is defined by the supply function, and the absorbance decay reflects the removal function. In a rapidly heated furnace it is possible to study atomic dissipation using standard analytical signals. If the effective vapour temperature is known, it is possible to not only obtain a qualitative study of atomic dissipation, but it is also practicable to generate accurate measurements of the dissipation parameters. Using a fast heating furnace, diffusion coefficients for 12 metals were measured by Sadagoff et al.471 Good agreement between experimental data and theoretical values (calculated using the Chapman–Enskog theory) was obtained for all elements studied. This verified that for Ag, Au, Bi, Cd, Ga, In, Mn, Pb, Sb, Sn, Tl and Zn, diffusion is the dominant loss mechanism and also that only weak interaction between the atomic vapour with the graphite surface is exhibited for these metals. Torsi et al. also used a specially designed atomizer and power supply to produce exceptionally fast heating rates.472 With this system, absorbance versus time curves are obtained which show a very steep increase of the signal from the baseline at the beginning of the atomization step, followed by a more or less extended flat region and then an exponential decay. A simple model was developed describing absorbance versus time curves considering diffusion as the sole loss mechanism. Based on this model, experimental curves obtained agreed well with the theoretical results and this was given as evidence of the existence of the flat region in the absorbance profile.
Liang and Xia conducted a comparison of two approaches to the elucidation of kinetic parameters for atom formation in ET-AAS.473 Experimental data collected for the probe atomisation of Tl were used to compare orthogonal polynomial regression and cubic spline function. Polynomial regression tended to produce larger values of kinetic order and activation order, compared with those obtained using cubic spline function, and the kinetic orders were found to be less temperature dependent, whereas in the latter case, values for order decreased with increasing temperature.
The formation of intercalation compounds is significant in ETAAS, affecting analysis in both positive and negative fashion. The formation of very thermally stable carbides results in poor detection of elements such as lanthanides and actinides, among others. The action of Pt-group modifiers is based on the formation of stable analyte–modifier compounds. The formation of intercalation compounds of graphite in the atomization process for ETAAS was investigated by Bulska et al.474 The graphite surface and subsurface was investigated using EPMA with EDX detection. Chloride was found to form thermally stable graphite intercalation products on any graphite surface, modified or not. The importance of compound formation at a number of levels was illustrated.
A number of groups of workers choose to study the chemical reactions of one or two elements at a time in ETAAS. The diverse results obtained in these studies emphasise the importance of these studies. Chekalin et al.475,476 used laser-excited fluorescence spectrometry (LIFS) to study vaporization and atomisation processes of In and Ga in a variable pressure atomizer. The high sensitivity of LIFS permitted the use of analytical concentrations of the element under study. Atomisation mechanisms for In were found to be dependent on analyte mass, gas pressure and heating rate. Mechanisms were proposed although no detail was given in the abstract. The chemical reactions of Te in a number of furnace types were investigated by Mueller-Vogt et al.477 On drying, TeO2 was formed, which was reduced by the graphite to volatile TeO, detected in the gas phase by molecular
absorption spectroscopy. In coated tubes, the reduction process was observed to occur at lower temperatures and with a faster reaction rate compared with uncoated and Zr-treated tubes. TeO is reduced to elemental Te on collision with the graphite tube wall. This transport process via the gas phase was observed by vaporization from L'vov platforms and subsequent separate atomization from wall and platform. In uncoated tubes, Te could be stabilized to temperatures above 1200°C due to intercalation, illustrated using re-oxidation experiments.
Se atomization has been the subject of a number of investigations over the years. The low temperature loss mechanisms of this element are well documented (see previous ASUs). Cabon and Le Bihan have made a systematic study of Se atomization in the presence of nitrates, chlorides and sulfates478 In the absence of modidier, about 50% of Se was lost as molecular species at low temperatures. The effect of salts on the shape of the atomization signal, the integrated absorbance and the stabilizing effect of Pd and Sr were highly dependent both on their nature and concentration. For low nitrate and chloride concentrations, significant increase of the integrated absorbance was observed. For higher concentrations of nitrate and chloride, the sequestration of Se in the corresponding oxides caused a shift to higher atomization temperatures and, consequently, a decrease of the integrated absorbance. The effect of sulfates was mainly dependent on their concentration and decomposition/vaporization mechanisms. In the presence of nitrate or chloride salts, Pd had a similar delaying effect, signal shape and integrated absorbance; however, Pd was less efficient in the presence of sulfate. Good recovery of Se was obtained in the presence of Sr, through a modification of the decomposition/vaporization mechanism of sulfate salts; this modification was accompanied by a dramatic decrease of the corresponding background absorbance signal. Therefore, Sr can be recommended as a chemical matrix modifier in the presence of high sulfate concentrations, alone or mixed with Pd.
The formation of molecular species containing B, Si or Ca in a graphite furnace was studied by Radic-Peric et al.479 Gas composition (Ar–O2–C) was varied and it was found that the equilibrium gas phase molecular composition was very sensitive to the amounts and ratio of amounts of O2 and C. Unsurprisingly, increasing the O∶C ratio favoured oxide formation of the analyte elements and led to higher atomization temperatures. Increasing the C∶O ratio favoured atom formation and also carbide formation at lower temperatures. Interestingly, however, it was found that over the relatively wide temperature range of 1000–3500 K, the composition of the Ar–O2–C–analyte system, with comparable amounts of C and O present, did not significantly depend on the presence of a solid phase.
The modifying effect of platinum group metals (PGM) in ETAAS is well established and these modifiers are widely used. However, the mechanism of modification is not yet completely understood. A review of the literature concerning the modifying mechanisms of the PGMs was presented by Volynsky480 (284 references). Despite the substantial body of work on this subject, and the general agreement on the mechanism of modification [low-temperature stabilization of volatile analyte species as a result of their chemisorption on the surface of the modifiers and/or on the graphite surface, followed by the PGM catalysed reduction by the graphite (or thermal dissociation) of analyte species at higher temperatures], the significant differences in the efficiency of Pd, Pt, Rh, Ru and Ir when used as modifiers has not been explained to date.
Yamamodo et al.481 used transmission electron microscopy to observe lattice structures in binary alloys to study the modifiying action of Ag and Pd on the atomization of Sb. Radical vaporization was observed in the Ag–Sb system but was not seen in the Pd–Sb system through the changing phases of the Pd–Sb intermetallic compounds. This means Pd does not have the function of a matrix modifier in the analysis of Sb. In the Ag–Sb system, however, an intermetallic compound was formed that has a lower value of the activity coefficient of Sb. As a result, Ag was found suitable to act as a matrix modifier.
The atomisation and modification of P in transversely heated ETAAS was studied by Caraballo et al.482 In agreement with previous studies, a palladium–magnesium nitrate mixed modifier was found to perform better for phosphorus than the other simple and mixed modifiers evaluated. SEM analysis of Pd-based modifiers deposited onto integrated platforms showed that palladium distribution over the platform is non-homogeneous, either when added alone or together with ascorbic or oxalic acid. This leads to partial analyte–modifier interactions and P losses. Magnesium nitrate appears to help in increasing Pd coverage of the deposition surface, thus allowing higher and more homogeneous coverage and more effective analyte–modifier interactions. This seemingly simple observation explains the benefit of the use of magnesium nitrate: however, the mechanism is not explained.
The mechanism of atomization and modification of Cr was investigated by Thomaidis and Piperaki.483 Kinetic parameters and atom dissipation processes were evaluated. In the absence of chemical modifiers, Cr atoms are produced in the gas phase by thermal desorption of adsorbed atoms, a first-order process. Pyrolysis temperatures over 1300°C were observed to promote the formation of Cr3C2, leading to a two-step atomization process: (i) the thermal decomposition of Cr3C2; and (ii) the thermal desorption of adsorbed Cr(ads). In the presence of Mg(NO3)2 and Ca(NO3)2 modifiers, atomization of Cr was observed to proceed through a two-precursor mechanism, with the first step including the thermal decomposition of the oxide. In the presence of Na2WO4, Rh and Pt,
a single-precursor mechanism was observed which was suggested to be a diffusion process through the bulk of the modifier.
Dual cavity platforms have been used by several groups of workers to aid elucidation of interference effects in ETAAS. Oezcan et al.484 studied the interference effects of sulfate and chloride salts of Na, K and Ni on the atomization of Sn using this method. It was observed that all chloride and sulfate salts studied influence Sn. In contrast to many other elements, the effect of sulfates was found to be more significant than that of chlorides, probably due to a greater extent of gas-phase reactions between the analyte and decomposition products of sulfates, which sometimes completely suppress the atomization signal.
Reviews on the use of lasers in analytical chemistry have been presented by Lytle,486 Gooijer and Mank487 and Winefordner et al.488 The last paper is directed specifically at the application of lasers in atomic spectroscopy, while the other two are of a more general nature. A review (62 refs.) of new types of tunable lasers has been published by Hou et al.489 It concludes that “much of the success of diode lasers lies in the future commercial availability of long-lived, blue diode lasers”. Other workers have “locked” an external-cavity diode laser to the D2 transition of atomic Cs using low frequency wavelength modulation with voltage tuning and rotation of the end mirror in the laser cavity with a piezoelectric stepper motor.490 With 6f-harmonic detection, the SNR was 460. A ring Ti:A12O3 laser has been “locked” to atomic absorption lines by bidirectional passive self-injection locking.491 The control relied on the imbalance between two opposite wavelength scanned self injections that was produced by absorption at the reference line. The emitted spectrum was centred at the absorption maximum and was much narrower then the absorption linewidth.
The potential for LA sampling of minerals in remote and extraterrestrial locations has been studied with particular attention to the understanding and control of fractionation.495 Ablated material was deposited as a film on a graphite substrate and analysed by Rutherford backscattering spectrometry and X-ray photoelectron spectroscopy. Laser ablation sampling has also been used in the analysis of sintered silicon nitride.496 In this case the sampled aerosol particles were collected in acid solution prior to analysis by ICP-MS. The efficiency of the particle transfer was ∼7%. The method was successfully applied to the determination of Co, Mg, Mn, Ti and W in a sintered body of silicon nitride. The design and operation of a laser ablation mass analyser has been described.124 The sample spot size was 30–50 µm diameter. The instrument was used for semiquantitative analysis of alloys and non-conducting powders. Mass resolution was 250 with LODs of 50 µg g−1 in the mass range <250 u. A novel use of LA has been in the study of the temporal characteristics of a pulsed glow discharge.497 It was found that decreasing the glow discharge working gas pressures increased the transport efficiency of laser ablated Cu from the target into the negative glow.
Several reports have described the coupling of laser ablation with ICP-AES. A patent has been granted to Shimadzu Corp. for an instrument based on that concept.498 The attraction of LA-ICP-AES lies in its solid sampling depth profiling multi-element capability. These qualities have been exploited by Nolte and Paul for the analysis of coins.499 Ablated material was swept via a 6 mm id, 1 m long Tygon tube into an atmospheric plasma. A low pressure ICP system was used for the LA analysis of alloy samples.500 The ablation chamber and ICP torch were directly connected via a three-step nozzle which focused the sample stream into the centre of the torch. LODs ranged from 1.8 µg g−1 (Mn) to 71 µg g−1 (Cu).
A study of the depth profiling of Sn coated glass using LA-ICP-AES found that while the Sn-coating mainly evaporated, glass crumbled into a powder.501 It was also noted that an acoustical signal had the same temporal profile as the Sn emission line. In the analysis of tungsten carbide–cobalt coatings a steep increase in the acoustic signal was observed when the laser was focused in the argon above the surface.502 The mass ablation rate increased as a function of the square root of the irradiance. This effect was attributed to plasma shielding. The Co∶W line intensity ratio was practically constant over the range 1.5–250 GW cm−2. Surface precipitates on pebbles and shells have been successfully determined by LA-ICP-AES.503 The heterogeneity of inorganic additives in organic polymers has been investigated by LA-ICP-AES and found to give results comparable with those obtained by solid sampling ZAAS but, by virtue of the former's multi-element capability, in a much shorter time.504
In some systems a second laser, focused into the ablated material and fired ∼30 ns after the first pulse, is used to excite the emission signal of the ablated material, thereby avoiding the need to transport the ablated material to, say, an ICP-AES instrument. An instrument based on the 2-laser system has been patented by Sabsabi and Cielo.505 The application of such a system to the determination of Na in KI pellets achieved an LOD five times lower (200 pg) than that of a single laser LIPS system.506 Using a 2-laser system 11–33 fold signal enhancement was observed when the first laser beam was aligned parallel to the sample surface and a few mm above it to generate a pre-ablation spark.507 In all cases where enhanced emission signals were seen greatly enhanced sample ablation occurred.
Studies of specific laser/target systems have been undertaken by a number of researchers. A study of the emission characteristics of a Cu alloy target in an Ar atmosphere at reduced pressure excited by YAG laser found that, compared with operation in air at atmospheric pressure, line intensity was increased and self absorption reduced.512 Other workers studied the interaction between radiation from a high-repetition-rate copper vapour laser and a Cu target with several buffer gases at varying pressures.513 In addition to maximizing the emission intensity of the 327.75 nm line temporal studies of the changes in SBR were undertaken. Air was found to play an important part in the process of plasma formation. Using a Q-switched Nd:YAG laser the excitation mechanism of Cu and Zn atoms and ions in a shock wave plasma have been investigated by Bude et al.514 The shock wave excitation process was found to be dominant for laser energies between 8 and 86 mJ at air pressures from 2 to 50 Torr. Below 2 Torr collision-induced excitation became dominant. In all cases emission from neutral atoms was greater than that from ions and had a higher SBR which, however, tended to decrease at higher laser energy and air pressure. When a mechanically “soft” sample was to be analysed, it was found to be necessary to incorporate a metal (Cu) sub-target in order to generate an adequate shock wave plasma.515 An atmospheric CO2 laser (100 mJ, 50 ns) was used to irradiate the sample in He at 760 Torr pressure.
Time resolved studies of the early phase of LIP generation revealed a rapidly changing situation greatly influenced by laser fluence, probe distance and chamber gas and pressure.516 In the first 200 ns spectral lines appear superimposed on a continuum spectrum: subsequently the latter gradually disappears while close to the substrate surface spectral line widths decrease and their peaks shift to shorter wavelengths. Ion lines disappear before atom lines due to plasma recombination. Spectral line intensities increase with laser fluence but tend to saturate at high laser fluence owing to a plasma shielding effect. The LIP from silicon undergoes a significant change in plasma characteristic beyond a 20 GW cm−2 threshold.517 From measurement of electron number density and temperature it was concluded that laser self-focusing and a critical temperature could explain the threshold effect.
A theoretical study of the problems associated with the fluid dynamics of the expanding plasma plume using time-dependent collisional-radiative models to describe the population densities of excited states and the time-dependent Boltzmann equation for the electron energy distribution concluded that all these factors should be taken into account when developing calibration-free LIPS methodology.518 Species distribution have been determined experimentally by means of temporal and spatial resolved LIP spectra.519 It was found that at the head of the plasma plume, ions were the dominant species. Atoms had two distribution peaks, one at the head of the plume, the other at the tail. These distributions changed with time while the proportion of ions to atoms decreased with decreasing laser fluence.
Bulatov et al. have obtained effectively time-resolved LIP spectra of the plasma front by viewing the plasma perpendicularly to the plasma expansion vector.526 Each point along the vector corresponds to a given time delay, e.g., 4.5 mm was equivalent to a delay of 4 µs and an integration time of 2 µs. The performance of the orthogonally mounted fibre optic, non-gated detector system observing a narrow solid angle was similar to that of a gated detector observing the whole plasma.
LIPS has the potential for effecting very rapid analysis. This attribute can be of considerable benefit for the sorting of scrap metal or the identification of geological materials. The identification of unknown materials is based on the comparison of the spectrum of the unknown with library spectra of known origin and similar composition. Gornushkin et al. found rank correlation to be the most reliable method giving a probability of correct identification close to unity.531,532 Several statistical methods (principal components analysis, cluster analysis, multiple discriminant analysis and spectral matching) were investigated for alloy identification by Goode et al.533 Spectral matching gave the best results and in a trial of 234 individual spectra against a library of 39 average spectra the identification of specific alloys was ∼80% correct.
Specific applications of LIPS are summarized in Table 2. Where the original publication contains information of more general interest it will be commented upon in this text. In a study of LIPS of glass the plasma emission front and shock wave front were observed simultaneously at air pressures up to 760 Torr.534 Initially, the two fronts coincided, but later the emission front was left behind. No dispersion of elements of different atomic masses was observed and it was concluded that the shock wave generated the plasma. The suitability of LIPS as a tool for space exploration has been investigated by Knight et al.535 The attraction of LIPS lies in its speed of analysis at a distance (<20 m). In a simulated Martian atmosphere (5–7 Torr CO2) an analytically useful plasma was generated from soils at a distance of 19 m with 35 mJ pulse−1. Increasing the pressure to 590 Torr increased the signal at least 3–4 fold.
Matrix solid | Analyte | Comment/LOD | Ref. |
---|---|---|---|
Solids— | |||
Bulk analysis— | |||
Steel (low alloy) | C, Cr, Mn, Ni, P, S, Si | Optimization of sensitivity. Performance comparable with spark AES. LODs: <10 µg g−1 | 546 |
Steel (low alloy) | C, Cr, Ni, Si | Time-gated electronics, Fe line as internal standard. RSD ∼0.5%; LODs 6–80 ppm | 547 |
Steel | C | Vacuum UV, time-integrated space-resolved measurements. LOD: 90 ppm for the C2+ 97.7 nm line | 548 |
Steel | C, P, S | Vacuum UV, multipulse excitation, time resolved electronics. LODs: <10 µg g−1 | 549 |
Al alloys, brass | Cu | Self absorption reduced by increasing laser pulse energy with time delay measurement (0.4 µs). Linear calibration curve: 4.5–20% Cu | 550 |
Al, Zn melts | Signal area measurement preferred. Calibration by molten standards | 551 | |
PuO2, UO2 | Impurity elements | 12 elements in PuO2 at ∽100 ppm, 18 elements in UO2 at ∽500 ppm | 717 |
Glass | Elemental composition | 10 elements determined qualitatively: quantitative elemental ratios could be derived | 718 |
Teeth | Trace metals | Elements from tooth fillings and toothpaste detected | 719 |
Iron ore | Mn, Si | Samples pelleted or applied to double-sided tape. Time-gated electronics. Evaluation of optimized conditions | 720 |
Soils | Metals | Estimation of plasma temperature | 721 |
Mineral drill cores | Cr, Cu, Fe, Mn | Semi-quantitative analysis. Remote signal acquisition. Correction for uneven surface of rock necessary | 722 |
Lignite | Al, Ca, Fe, Mg, Na, Si | Pelleted sample. LODs: 60–200 ppm | 708 |
Surface analysis— | |||
Polymer | Cl, F | Analysis of polluted surfaces | 552 |
Coated Ti microparticles | Study of stability of coating | 553 | |
Polychromes | Identification of pigments | 554 | |
Painted artworks | Identification of pigments. LIPS and Raman microscopy employed | 555 | |
Automotive exhaust catalysts | Rh, Pd, Pt | Spatial distribution analysis | 556 |
Photovoltaic cells | Ag, Al | Surface mapping, depth profiling | 557 |
Coated steel | Al, Fe, Zn | Depth profiling | 558 |
Liquid analysis— | |||
Water, oil | Cr | Laser radiation focused on liquid surface | 559 |
Gas/aerosol particle analysis— | |||
Incinerator gases | Be, Cd, Cr, Hg, Pb | Improved sampling and signal processing. LODs: 2–100 μg dscm−1 | 560 |
Droplets | Cr | Optimized conditions. LODs: 12–60 µg dscm−1 | 561 |
Industrial exhaust stream aerosols | Cr | Continuous sampling through heated flow cell. LOD: 14 µg m−3 | 562 |
Industrial gases | HgBr2 | Heated quartz flow cell. HgBr2 photo-dissociated. Hg emission detected. LOD: 13 µg m−3 | 562, 563 |
A study of the ablation products from graphite and organic materials found that C2 as well as C was released directly from the targets while CN was formed later by interaction of C2 with atmospheric N2.536 In the case of organic compounds a clear relationship was found between C2 emission and the presence of aromatic rings in the compounds.
Mapping of surface topography and elemental composition have been achieved by combining a sharp fibre-optic scanning probe microscope with LIPS in a single instrument.537 Mapping elemental composition of heterogeneous surfaces using LIPS was speeded up when the incident laser beam was focused by a cylindrical lens to produce a micro-line image. The emission from the line of laser plasma was imaged along the length of the spectrometer entrance slit while in the focal plane of the spectrometer a two-dimensional detector array generated the analytical signals.538 This system speeded-up the compositional mapping of solar cells 25 times. The depth resolution of the depth profiling of coated materials (Sn on steel, Cr on foil) was significantly improved to less than 2 nm pulse−1 by illuminating the sample with a “conditioned” laser beam at non-normal incidence, thereby reducing the energy density at the surface.539 “Conditioning” was used to produce a more homogeneous energy profile and thus generate a crater with a more rectangular profile. Sun and co-workers have used LIPS as a tool to investigate the effectiveness of barrier creams in preventing the penetration of hydrophilic and lipophilic agents through skin in vivo.540,541 Zinc was used as the tracer element as ZnCl2 solution and ZnO paste. Multiple biopsies 2–3 µm thick were taken from the forearm skin of volunteers for analysis by LIPS. Barrier creams were found to be effective against both ZnCl2 and ZnO penetration.
A comprehensive study of the application of LIPS to real-time, in-situ and remote analyses of liquid samples has been reported.542 Internal standardization utilized the Hα, Hβ or Hγ lines of H2 or “spiking” with reference elements. Numerous elements were measured over a wide concentration range with LODs of the order of a few ppm. As an alternative to the direct measurement of liquids, with its attendant problems, Wall et al. deposited 1 ml of liquid on a C planchet and evaporated the sample to dryness prior to LIPS analysis.543 Some 15 metals were studied and detection limits in the range of 10 ppb to 10 ppm measured.
The spectral emission of LIPS in gases (Ar, He, N2 and air) has been studied to determine the sources of the emission (continuum, atomic and ionic) and the electron temperature and density of the plasmas.544 The effect of gas pressure on the plasmas was also examined. A system based on LIPS for the detection and determination of toxic molecular compounds has been developed by Lancelin et al.545 The high temperature of the plasma decomposes the compounds and excites the emission of atoms such as Cl, F, P and S from which the original organohalogenated molecule may be identified. The LOD was of the order of 10 ppm (w/w).
Podshivalov and co-workers have investigated the fluorescence of 202Hg vapours when excited by two lasers tuned to two connected atomic resonant frequencies.564 Some conditions of the laser power and tuning were observed such that there was a decrease in fluorescence yield and increased transmittancy of the ground state transition. As these effects only occurred when the two laser pulses were temporally coincident photoionization was excluded as a possible mechanism. Stark splitting of energy levels and electromagnetically-induced transparency were suggested as possible explanations of the observations. These effects can be of significance in multi-step excitation when used for the determination of elements by atomic fluorescence. A two laser system has been used by Paquette et al. to excite the fluorescence of atomic Cd and Zn.565 When an ICP was used to atomize the sample the LODs for Cd and Zn were 5 and 1.7 µg 1−1, respectively, and when a L'vov platform in a graphite furnace was the atomizing system, the LODs were 4 and 70 ng 1−1, respectively. Calibration graphs were linear over 2–3 orders of magnitude. The same workers compared the performance of ICP and ETA atomization for the determination of As and Se by LEAF following hydride generation.566 The LODs were: ICP, As, 1.0 µg 1−1, Se, 0.06 µg 1−1; ETA, As, 0.04 µg 1−1, Se, 0.16 µg 1−1. Precision was 3–10% using the ICP and 5–30% for ETA. The performance of ETA was worse than expected; this was attributed to inefficient trapping of hydrides in the furnace.
For the determination of Au, double resonance LEAF with ETA gave LODs at the low fg level.567 As a result of such high sensitivity there was a significant problem of blank signals which had to be addressed. A variety of furnace coatings and linings were investigated for the determination of ultratrace levels of Tl by ETA-LEAFS.568 The best results were obtained with rhenium lined graphite tubes using a stop-flow regime. Under optimized SBR conditions the LOD was 200 fg with a precision of 6% at the 10 µg 1−1 level. By using a Katskov filter Pb in whole blood was determined directly by ETA-LEAFS employing aqueous standards for calibration.569 Palladium in airborne particulates has been determined by ETA-LEAFS after dissolution of the sample in aqua regia followed by selective extraction.570 The concentration of Pd in ambient air was found to be in the range 0.2–14.6 ng m−3.
Laser ablation was coupled with LEAF for the determination of the distribution of Na on the surface of Na-doped polymethyl methacrylate.571 The ablation rate was 4.4 nm shot−1 with an LOD of 35 fg.
The most interesting studies in LAAS have been carried out by Hannaford and McLean, where laser-atom cooling has been used to create idealized AAS experiments.574 Under these conditions the resolution and sensitivity for a given atom density are greatly enhanced (∼3 orders of magnitude) compared with conventional AAS. The experimental system used laser-cooled Cs atoms and a narrow-band diode laser. The temperature of the ultra-cold atoms was ∼6 µK.
The emission wavelengths of laser diodes are very gradually being extended toward shorter (violet) wavelengths. A commercial 5 mW CW diode laser operating at 404 nm has been used by Gustafsson to study the spectroscopy of the 404.5 and 404.8 lines of K.575 Frequency modulation of the output of diode lasers can be used to provide ultrasensitive absorption measurements. Again, K has been used as a test element to demonstrate the analytical potential of a laser diode technique.576 Detection sensitivities of 2 × 10−6 and 5 × 10−6 absorption for wavelength and two-tone frequency modulation, respectively, using a 120 Hz bandwidth were observed. The dynamic range of wavelength modulated diode laser absorption spectrometry (WM-DLAS) has been extended by utilizing the fact that the sample optical thickness is a multivalued function of the WM-DLAS signal.577 This approach was applied to the detection of Rb at 780 nm in a graphite furnace. The dynamic range of the 2f-WM-DLAS technique was extended to ∼20 integrated absorbance units (s) (i.e., detection of Rb atoms in graphite furnace from femtograms to nanograms) with no need for change of the laser settings.
Isotope analysis by 2f-WM-DLAS has been described in several publications. In a study by Wizemann of natural Li isotopes using low pressure graphite tube atomization, Li isotope ratios larger than 2000 could be measured though at the cost of detection power.578 Resonant Doppler-free two photon DLAS has been used by the same workers as an alternative approach to Li isotope analysis.579 A low pressure graphite furnace atomizer was employed with co- and counter-propagating laser beams. The former created large optical saturation conditions while the latter created small optical saturation. The LODs under both conditions were ∼1 ng while the selectivity for 6Li was 4.8 × 103 in high saturation and 4.6 × 104 in the low saturation case. Rubidium isotope ratios, 85Rb∶87Rb, in a solid CaCO3 sample were determined by low pressure laser ablation with wavelength scanning laser AAS of the 780.02 nm transition.580 The ratio was found to be 2.7 ± 0.2 with LODs for each isotope of ∼25 ppm.
Isotope dilution has been applied to the determination of 85Rb and 87Rb in water using diode laser low pressure graphite furnace AAS.581 Two wavelength modulated diode lasers were employed, one to detect each isotope. In the analysis of NIST Reference Water (SRM 1640), 87Rb was used as the tracer isotope and the results at the 200 µg 1−1 level agreed with the reference value to within 1%. A two laser system has been employed for the determination of Hg.582 Mercury, generated by a CV technique, was carried in a stream of Ar into a discharge tube. The wavelength of a transition at 365.11 nm with a high oscillator strength was obtained by sum frequency generation of the two diode lasers. An LOD of 100 ng 1−1 was achieved. Chlorine in polymers was determined by diode laser AAS in a low pressure plasma following ablation of the sample and GC separation of compounds.583 The LOD for Cl was 10 pg with a determination range for PVC of 0.1–50%.
Laser ionization of neutral material sputtered from a sample surface by a Ga+ ion beam has been proposed as a means of 3-D surface analysis.586 The photoions were selected by a TOF-MS. The depth resolution of 20 nm and the lateral resolution of 3 µm do not represent the limits achievable by the technique. Two tunable pulse lasers have been used to selectively photoionize Ni atoms in a mixed beam of Ni and Cu atoms.587 The photoionized Ni ions were extracted by an electric field and deposited on a Ag substrate. Starting from a Ni(36%)–Cu(64%) alloy a film containing Ni(93.3%)–Cu(6.7%) was obtained.
A comparison of traditional and multivariate calibration techniques applied to complex matrices using ICP-AES has been reported by Griffiths et al.596 The study used inter-element correction and matrix matching for the determination of platinum group metals in automobile catalyst digests. A multivariate study of matrix effects in ICP-AES using a comparison of pneumatic and ultrasonic nebulisation has also been reported438 using the relationship between interference effects and matrix composition. In a further paper,440 matrix effects induced by Ca and Mg in ICP-AES were investigated using a 22 factorial design and multiple linear regression to correct for the matrix. Average deviations from the expected values in the range of 61–6% were reduced to 4.8–0.5% after correction. Other reports based on matrix interference diagnostics include an extension to the total interference level concept,597 where measurements with different interferents are used to determine a set of interference coefficients, and the application of factorial designs for correction.598 Correction of spectral interferences in ICP-AES has also been reported using a window-adaptive noise-cancelling method.599
Systematic errors in both ICP-AES and ICP-MS have been evaluated by Hill's group at Plymouth. Principal component analysis and experimental designs were first evaluated600 and the study was then extended to the application of parallel factor analysis.601 In both cases, the study was based on the multi-element determination of a range of analyses in an environmental matrix. The same group have also described a drift correction procedure for ICP-AES.602 The suggested procedure utilised the drift pattern of intrinsic plasma lines to reduce errors for up to 20% to below 2%. The application of Kohonen neural networks to diagnose calibration problems in AAS has been reported by Vander Heyden et al.,603 whilst Haaland and Chambers604 have used a multi-window classical least-squares multivariate calibration method for ICP-AES.
The partial least squares approach has been used by several authors. Full spectrum modelling rather than traditional utilisation of individual calibration lines has been reported605 for application with ICP-AES. Partial least-squares regression has also been used in conjunction with analysis by laser induced breakdown spectroscopy for the determination of noble metals in jewellery.606
Various specific applications based on data obtained using atomic spectroscopy have been reported. These include the geographical origin of brown rice,607 the classification of beers,608 classification of Roman glazed ceramics using neural network self organizing maps,609 essential elements in hair, teeth and nails,610 and spectrochemical analysis of high-temperature superconductor materials.611
A topic which is sure to attract more interest in future is that of transalkylation reactions between organometallic species. One group of workers studied such reactions between tetraethyllead and inorganic mercury618 and found that the latter was quickly and almost completely ethylated to EtHg+ by the former under laboratory conditions. They also found that EtHg+ and MeHg+ were present in soils from sites that had been contaminated with Et4Pb and Hg2+, indicating that such reactions occur quite readily in the environment. The only other application addressing organometallic speciation was the determination of methylcyclopentadienylmanganese tricarbonyl in gasolines and octane boosters.619
The determination of compounds containing hetero-atoms is the other major application of GC-MIP-AES that usually dominates the literature, and this year was no exception, with applications detailing the selective detection of Cl-, S-, N-, P-, O-, As-, F-, Br- and I-containing compounds.620–626 Of particular note was: the coupling of a programmable pyrolyzer with MIP-AES and MS for the analysis of pressure sensitive adhesives in adhesive tapes,627 such that additives were selectively identified by the combined analysis of the two pyrograms; determination of compounds in nerve agents624 and chemical warfare agents,625 wherein the element selective nature of the detection allowed the analytes to be quantified even if no standards were available or were too dangerous to use; determination of methyl sulfonyl PCB and DDE in grey seal tissues;622 and multi-element detection during simulated distillation of diesel oil,621 resulting in LOD of 0.002 and 0.04% for S and N at 181.379 and 174.2 nm, respectively.
A significant instrumental development was made during the review period, namely the development of a dc microplasma on a chip.628 A 180 nl plasma chamber was micromachined into the chip, and an atmospheric dc plasma was generated in the chamber at 9 mW power (770 V, 12 µA) and 320 nl s−1 He gas flow rate. Element selective detection of carbon at 519 nm was effected for a number of compounds introduced by GC, resulting in a LOD of 800 ppb and a LDR of 102; however, marked peak broadening was observed, which was attributed to dead volume. Application of a 60 W argon microwave plasma torch (MPT) for GC detection has also been investigated,626 with LODs of 0.24, 4.7 and 0.13 µg for chlorobenzene, bromoethane and iodoethane, respectively.
Some interest in the use of glow discharges as element specific detectors was evident during the review period. A novel slant on this was described in four consecutive papers by the same group, which appeared in the Journal of Chromatography, on the development and application of a pulsed discharge source for element specific detection of Cl, Br, I and S.629–632 The mode of operation for Cl-selective detection630 is facilitated by doping the He discharge with Kr to form Kr ions, which react with chlorinated compounds to form excited KrCl which emits at 221–222 nm. The advantages of this, listed by the authors, are that: the ion–molecule reaction is extremely fast, and hence sensitive; the excited KrCl emission is spectrally separated from interferences by C emission at 193 and 247.3 nm; emission in this part of the spectrum does not require a monochromator purge; the pulsed discharge has a volume of only 0.35 µl and is sufficiently narrow to replace the monochromator entrance slit, thereby increasing light throughput; it requires a flow rate of only 5 ml min−1. The ‘minimum detectability’ of the detector is quoted as 50 fg s−1, with a selectivity of 1000 with respect to carbon. The detector has been used for the analysis of organochlorine pesticide and polychlorinated biphenyl test mixtures, and the response was shown to be independent of the nature of the compound.631,632
Other papers of note include the development of a radiofrequency hollow cathode He GD633 for speciation of methyl-, ethyl- and inorganic Hg, with LOD of 0.2–0.3 pg; a dc hollow cathode He GD for detection of chlorinated and brominated hydrocarbons,634 with LOD of 3 and 5 pg s−1, respectively; flame photometric detection of white phosphorus residues in the gizzards of ducks635 after extraction with isooctane; and headspace SPME of butyltin compounds from sea-water after generation of their hydrides with NaBH4, then desorption from the SPME fibre directly into a GC and detection using FPD, with LODs of 19.4, 1.5 and 0.5 ng l−1 for mono-, di- and tributyltin, respectively.
Another indication of the problems that arise in speciation analysis was evident from a study of artefact formation during the determination of organomercury species.644 These workers found that methylmercury formation can occur from the methylation of Hg2+ in the GC column itself, as well as in the extraction and sample preparation steps, due to the silanizing agent, dimethyldisilazene, acting as a methyl donor when high concentrations of Hg2+ (>1 ppm) were present.
Other reviews have concentrated on specific techniques. One by Vela and Caruso650 reviewed (with 33 refs.) the coupling of supercritical fluid chromatography (SFC) with both ICP-AES and ICP-MS instruments for the analysis of organometallic compounds. An introduction to the principles and a discussion of the requirements for the coupling was also given. The last part of the review concentrated on a comprehensive description of SFC-ICP applications for the analysis of organometallic compounds containing As, Cr, Fe, Hg, Pb, Sb, Sn and Ti. Guerin et al.651 have prepared a review of high-performance liquid chromatography (HPLC) separation techniques hyphenated to specific detectors for the determination of As and Se compounds in environmental samples (152 refs.). This publication contained numerous tabulated data and included ion pairing reversed phase and anion exchange separations, details of the columns, mobile phases, matrices, the species determined, the detection system and limits of detection.
There have been numerous element specific reviews; some of which have been more critical and comprehensive than others. Arsenic speciation and toxicity has been reviewed by three sets of workers.652–654 The last of these contained 137 references, but also included an assortment of detection systems including spectrometric and electrochemical methods. The first of the As reviews652 contained 114 references and took the reader through most steps of the analysis, including sample preparation, species separation (including a range of different types of HPLC separation) and detection by several different atomic spectrometric methods. Tin speciation has been reviewed (98 refs.) by Takeuchi et al.655 who compared GC and HPLC separation/detection systems. The authors stated that the HPLC detection systems tend to be AAS, ICP-MS and fluorescence based, but that ICP-MS is used most frequently because of its inherent sensitivity and selectivity. A survey of the literature for Cr speciation in liquid samples has been published.656 The article contains 404 references, but only 134 of them focus on atomic spectrometric techniques, although details of types of pre-treatment, species, samples, techniques and analytical features of the methodologies are given. An assessment of the methods currently used for the determination of Cr and Se species in solution has been published.657 It focused on the results obtained during an inter-laboratory study and preparation of candidate reference materials. Several different methodologies were detailed along with the results obtained from each. Selenium speciation has also been overviewed and the analytical techniques critically evaluated, with particular emphasis on the element specific detection and identification of the detected Se compounds.658 A review of Sb determination in environmental samples has been made by Nash et al.659 This article critically evaluated the methodologies available for the determination of total Sb and also its aqueous chemical speciation. The potential drawbacks with HG based techniques are discussed and the advantages of the on-line coupling of HPLC with ICP instrumentation are evaluated. The article also states that far more work has to be performed in the area of Sb speciation so that a better understanding of the environmental chemistry can be obtained. Mercury has been the most popular element for speciation reviews this year. Liu et al.660 described an assortment of different methods including AAS, AFS, AES, chromatographic techniques such as HPLC and GC and also other less commonly used techniques such as X-ray fluorescence (XRF) and neutron activation analysis (NAA). The paper is in Chinese and contains only 68 references. Other, potentially more helpful, reviews include one by Carro and Mejuto,661 who discussed the application of chromatographic and electrophoretic methodology to the speciation of organomercury compounds in food (218 refs.). The review concentrated very much upon AAS and ICP-MS detection systems. Two other relevant papers have been published. Harrington662 reviewed numerous HPLC separations of Hg and organomercury compounds (79 refs.) and gave the relevant technical information, i.e., column type, mobile phase composition, method of detection and LOD, in a tabular form for each of the sample types. It was concluded that the majority of methods use reversed-phase chromatography, usually with a buffer, organic modifier and some form of counter ion or complexing agent. Another review, also containing 79 references and also focusing on HPLC methodology, has been produced.663 Again, the authors gave the merits of the various detection methods and a brief description of the isolation techniques.
Two other applications of As speciation have been published. One determined As species in assorted types of Slovenian mushrooms, using a variety of detection systems including HPLC-AAS, HPLC-ICP-MS and HPLC-ETAAS.670 The chromatography column used was Hamilton PRP X-100. The paper listed the proportion of As species found in each mushroom type. It was found that HPLC-AAS was insufficiently sensitive to quantify the As species in mushrooms from uncontaminated sites. The other application671 determined total and inorganic As in fresh and processed fish products. Unfortunately, the abstract gave no experimental details. Both As and Sn species have been determined using HPLC and UV oxidation-HG-AAS.672 The UV oxidation was assisted by alkaline persulfate for As species (with >90% transformation of MMA, DMA, AsB, AsChol, TMAO and tetramethylarsonium to arsenate) and by acidic persulfate for Sn (with >80% transformation for most Sn species). Tetrabutyltin had the lowest transformation efficiency (15%). Limits of detection for As and Sn were 4 and 2 µg l−1, respectively. Precision for 5 ng of As and 10 ng of Sn were 8.5 and 6.4%, respectively. Urine CRMs were analysed for their total As content to validate the method.
Numerous other analytes have undergone speciation analysis using LC-AAS detection. These include Sb.673 Several anion exchange columns were used in an attempt to separate the Sb species. Dionex AS14 used with a 1.25 mM EDTA solution at pH 4.7 gave the best separation between SbV and SbIII, and an ION-120 column with 2 mM NH4HCO3 and 1 mM tartaric acid at pH 8.5 gave the best separation between trimethylantimony dichloride (TMSbCl2) and SbV. The column eluates were then mixed on-line with acidic sodium tetrahydroborate and the hydrides so formed detected by AAS. It was noted that each Sb species required a very different optimum concentration of acid and tetrahydroborate, but that 0.5% w/v HCl and 0.6% w/v NaBH4 was the best compromise. Although the inorganic Sb species yielded sharp peaks, the TMSbCl2 was very broad. This, however, seemed not to affect the LODs, which were 0.4, 0.7 and 1.0 µg l−1 for TMSbCl2, SbIII and SbV, respectively. A comparison of Cr speciation using CZE with UV detection and ion exchange coupled with AAS has been made.674 The LC-AAS method used was a standard method published in 1974 (Pankow and Janauer, Anal. Chim. Acta, 1974, 69, 97), but the results were in good agreement with those obtained by the CZE method. Several other more novel Cr speciation papers have been published: however, most of them involve an “off-line” detection system. Hexavalent chromium in lime-treated sewage sludge has been determined by anion exchange fast protein liquid chromatography with detection by ETAAS.675 A Tris-HCl buffer (0.005 M, pH 8) and the same buffer but containing 0.5 M NaCl, were used to elute the species within 15 min. The LOD was 1.5 ng ml−1, which correlated to 20 ng g−1 in the sewage sludge. Method validation was achieved using three different methods, i.e., analysis of a CRM, spiking/recovery experiments (93–101%) and analysis using an alternative technique. Different oxidation states of Cr have been speciated in water samples.676 An anion complex of the Cr species with EDTA was formed and these complexes were retained on a column of strong anion exchange resin. Elution was by 0.5 M NaCl. The retention and elution characteristics were optimised prior to analysis. The LODs of CrIII and CrVI were 0.4 and 1.1 µg l−1, respectively. An interference study for the ions MgII, MnII, SnII, FeIII, BaII, AlIII, CaII, chloride, bromide, nitrate, fluoride, sulfate, phosphate and bicarbonate was also performed. The results obtained using this method for the analysis of a wastewater were in agreement with those obtained using a reference method.
There has not been as much interest in Se speciation as in previous review periods. In the only paper published, a HPLC-microwave digestion-HG-AAS procedure was used to determine selenomethionine, selenocystine, selenate and selenite.677 The procedure underwent multivariate optimization and it was found that the HBr∶BrO3− ratio was a very important parameter because it both controlled the conversion efficiency of the Se compounds to a reducible form and also diminished the effects of an interferent. The procedure was applied to the analysis of extracts of CRM 402. The authors concluded that the procedure was sensitive to interferences, but that with LODs of 1.0–1.6 ng g−1 the detection capability was good. Selenate seemed to be the only Se compound found in the extract, but this constituted only 22% of the total Se. The extraction method was blamed for this low result.
A conference presentation has described selenium speciation in infant milk formulas and milk whey using size exclusion chromatography (SEC)-HG-ETAAS with an iridium coated graphite tube.678 After ultracentrifugation to remove milk fat and casein micelles, the whey milk was passed through a column of TSK gel 2000 SW and the fractions collected were digested with nitric acid and hydrogen peroxide in a microwave oven and analysed using HG-ETAAS. The Ir impregnated tube enabled Se collection at 800°C and atomisation at 2100
°C. The abstract did not enter into much more detail.
The same authors have undertaken studies of Fe speciation in the same sample type.679 A TSK gel SW guard column and a TSK gel G2000 glass column were used to separate the Fe species associated with the low molecular weight fractions. The mobile phase was water, which facilitated the subsequent analysis by ETAAS. The method was described as being sensitive, yielding a LOD of 1.4 µg l−1, as well as being precise (RSD typically <10%). It was found that the Fe was associated with proteins between 3 and 76 kDa in breast milk, but that it was irregularly distributed in infant milk formula. Copper has also been determined by these same authors in the same sample types.680 In this presentation, the mobile phase was NH4NO3 (0.2 M) and NH3 solution (pH 6.7), but the same analytical column was utilised.
The determination of inorganic Hg and methylmercury in brackish waters and digests of biological samples (CRMs DOLT-2 and TORT-2) by FI-HPLC-CVAAS has been described.681 Experimental details in the abstract were sparse to say the least, but a comparison was made with a GC-MIP-AES method. Absolute LODs for the FI-HPLC-CVAAS method were 1.7 pg and 3.4 pg for methyl- and inorganic Hg, respectively. Cadmium speciation has been described in two publications. In one,682 metallothioneins from rat livers were saturated with Cd and then separated using HPLC-AAS. The extraction method of the metallothioneins, the Cd saturation process and the protein separation were all optimized. The other relevant publication determined Cd species of environmental importance, i.e., fractions able and not able to be taken up by aquatic organisms.683 Chelex-100 was used to accumulate the Cd ions and weak Cd complexes, whereas a silica C18 column was used to retain the stable Cd complexes. The analytes were then eluted and were detected by ETAAS. The authors went to great length to ensure that the eluant had both optimal efficiency and was compatible with the detection system. Speciation of inorganic and trialkyl Pb compounds has been reported.106 The proposed flow system consisted of two units. In the first unit inorganic Pb at concentrations from 8 to 200 ng ml−1 were precipitated continuously as the chromate. The precipitate was then redissolved in dilute acid and sent to the detector. The filtrate containing the organolead compounds was passed to the second unit where they were complexed with NaDDTC and subsequently retained on a C60 fullerene column. The trimethyllead (TML) and triethyllead (TEL) were then resolved by treating the column with either n-hexane or IBMK. The LODs for a 50 ml sample volume were 1–2 ng ml−1. Trimethyllead was found to give the poorest results because of its low adsorption onto the fullerene column. Despite this, the three lead species can be determined accurately in proportions from 1∶1∶1 to 30∶12∶1 (for Pb2+∶TML∶TEL) with relative errors of <10%.
Aluminium has also proved to be a popular element in this review period. Its speciation in blood serum has been reported in two papers by the same authors.684,685 In both papers fast protein liquid chromatography (FPLC) was used with an NH4NO3 mobile phase to separate the species, fractions of 0.2 ml were collected and then detection was “off-line” using ETAAS. The work reported in these two papers was a follow up of the work reported previously.686 It was found that the main species found were Al–citrate, Al–phosphate and ternary Al–citrate–phosphate complexes. In another study of Al speciation in serum, HPLC-AAS detection was used to determine what Al compounds were formed after deferoxamine, a compound that chelates with Al forming aluminoxamine, had been administered.687 Several species other than aluminoxamine were found. The paper was published in a biomedical journal and so the abstract is short on experimental details. It is known, however, that ultra-filtration techniques were also used in an attempt to clarify the results. Another application of Al speciation has been during the analysis of tea infusions.688 An attempt was made to fractionate Al bound to hydrolyzable polyphenols and its cationic species and then quantify them using flame AAS. Chelex-100 was used to retain the inorganic Al, whilst amberlite XAD-7 resin was used to retain the polyphenol bound Al species. Both columns were used on-line, with the effluent from the amberlite column passing straight into the chelex column. It was found that between 10 and 19% of the Al in the tea infusion existed as cationic species, whereas 28–33% was polyphenol bound. The abstract did not specify what the remaining 50–60% existed as. A very interesting paper describing the speciation of Mn has been published.689 Methylcyclopentadienyl manganese tricarbonyl (MMT) is an additive to unleaded gasoline and is suggested to be a health risk. This compound, and other metabolites (cyclopentadienylmanganese tricarbonyl (CMT) and inorganic Mn) were determined by HPLC-diode laser (DL)-AAS. The DLAAS technique was very sensitive for “total” Mn determination with a LOD of 1 ng ml−1 and a linear range that extended to five orders of magnitude. The analytical figures of merit for HPLC-DLAAS included a LOD of 2 ng ml−1 (as Mn), a linear dynamic range spanning three orders of magnitude and an analysis time of 3 min. The system had sufficient resolution to determine a 20-fold excess of CMT when compared with MMT. The process was applied to spiked gasoline, human urine and tap water.
Speciation of Cu, Fe and Mn in beer has been achieved by Svendsen and Lund.690 Cation and anion forms of the metals were collected on solid-phase extraction cartridges, whereas the molecular weights of the metal complexes were determined using SEC. Electrothermal AAS was used “off-line” for the measurement step. It was found that Mn was present as Mn2+ in beer, whereas Fe existed as a negatively charged complex and Cu was found as both positive and negative species. The SEC was calibrated using globular proteins and the results from the analysis of the beer found that the metal complexes were in the range 4–12 kDa. Spiking of the beer with Cu2+ and FeIII led to the formation of negatively charged complexes, indicating that the beer contained free metal complexing ligands. Other papers have also determined more than one element. An investigation into the transport of metal ions in xylem sap of cucumber plants has been performed using SEC-ETAAS, again in an “off-line” mode.691 A simple element specific detector using a solar blind photocell instead of a dispersion system has been described for the speciation of As, Hg and Se.401 Spectrometric studies of AAS background lamps for As and Se measured with a CsI cathode photocell showed its quality as a narrow band detector. Species determination was carried out subsequent to GC or HPLC separation. The LODs for alkylated Hg species were <1 ng l−1, but for methylated As species, they were marginally less impressive, although still yielding an LOD of <1 µg l−1. Precision was better than 10%. The authors concluded that with the components described, the production of cheap, automated, dedicated speciation spectrometers was possible. An apparatus consisting of three columns packed with three different substrates (an anion exchange resin, a chelating resin and a reversed-phase sorbent) has been developed to distinguish between species with different charge or chemical behaviour.692 Experiments were performed with free metal ions, their anionic complexes with EDTA and 8-hydroxyquinoline–5-sulfonic acid, and their neutral complexes with 8-hydroxyquinoline. The results obtained were compared with those computed according to theoretical species distribution of all species as a function of pH. The applicability of the technique to sea-water was demonstrated.
Micellar SEC has been used for the simultaneous multi-elemental determination of a novel glutathione–As–Se ion.696 A Sephadex G-25 column used with Tris buffers (0.1 M, pH 8) containing various surfactants, e.g., hexadecyltrimethylammonium bromide (HDTAB), dodecyltrimethylammonium bromide (DDTAB) and sodium lauryl sulfate, and with ICP-AES detection was used for the determination of As, S and Se. By comparison of the different chromatograms, it was possible to determine the relative strengths of association of the molecule with the positively charged micelles and hence elucidate the structure of the molecule. Data strongly supported the spectroscopically derived formula of (glutathione)2–As–Se.
As in previous years, a combination of LC with HG followed by ICP-AES detection has yielded a small crop of speciation papers because of the improvement in sensitivity obtained arising from the increased transport efficiency of the analyte to the atom cell. Several As species (AsIII, AsV, MMA and DMA) in water have been determined using this methodology.697 Several anion exchange columns were used (both silica based and polymeric) with a mobile phase of EDTA. The eluate was mixed with sodium tetrahydroborate in an acidic medium and the hydrides so formed detected by ICP-AES at 193.7 nm. Complete elution and resolution of all four species in approximately 6 min was reported, with linear ranges spanning 0.05–2 µg ml−1 for AsIII, AsV and MMA and 0.1–2 µg ml−1 for DMA. For a 200 µl sample injection, LODs were in the ng range, with DMA being the least sensitive. The analysis of natural water samples indicated that recovery values may be dependent upon sample composition. The same four As species have been determined in the urine of patients that have been intravenously treated with As2O3.698 In this paper, a reversed-phase HPLC separation was used but, unfortunately, the abstract was less than informative so details of mobile phase composition were not given. It was reported, however, that determinations were sensitive and that the As species could be detected below the µg l−1 range.
There has been far less interest in using LC-AES for the speciation of other analytes. A report (in Hungarian) described the speciation of CrIII and CrVI using HPLC with either MIP-AES or ICP-AES detection and sample introduction to the plasma instrumentation via a hydraulic high pressure nebuliser (HHPN).699 Again, few details are available. Another paper that has used sample introduction via HHPN has determined S in plant extracts.700 An anion exchange column with a mobile phase of 2.5 mM phthalic acid adjusted to pH 4.2 with tris was used to separate organic S compounds from sulfate. The organic S compounds eluted in the solvent front (after 2 min) whilst the sulfate was retained until 4.2 min. Other S containing species, e.g., sulfite, sulfide, thiosulfate, etc., were not detected in the plant extracts, but the sulfite and sulfide standards did elute between the organic S and sulfate. Determination at the 180.734 nm line yielded LODs of 0.02 mg l−1 and 0.085 mg l−1 for sulfate S and organic S, respectively. The extraction of butyltins from sediments and their determination by LC-ICP-AES has been reported.701 The sediment was extracted with methanol containing tropolone and glacial acetic acid or hydrochloric acid. The extracts were separated using a Partisil 10-SCX column using ammonium acetate in methanol–water containing tropolone as the mobile phase. After separation, the butyltins underwent hydride generation prior to ICP-AES detection. Recoveries of tributyltin (TBT), dibutyltin (DBT) and monobutyltin (MBT) were 86, 80 and 42 %, respectively. Detection limits were 27–62 ng Sn g−1 dried sediment.
An “off-line” method of matrix removal has been reported.702 Rare earth element impurities were determined in CeO2 by HPLC followed by ETV-ICP-AES detection. Using an ethylhexyl hydrogen 2-ethylphosphonate (P507) resin as a stationary phase and dilute nitric acid as mobile phase, the cerium could be eluted to waste. Elution of the analytes was achieved using EDTA, but the separation took 60 min.
Particle beam–hollow cathode–glow discharge–atomic emission spectrometry (PB-HC-GD-AES) as a means of detection for the liquid chromatographic separation of aromatic amino acids as well as organomercury and organolead compounds has been evaluated.703 Sample was introduced to a heated hollow cathode glow discharge source for subsequent vaporization, atomization and excitation via a high efficiency thermoconcentric nebuliser. Emission responses for H and N of amino acids for a 200 µl sample injection volume yielded LODs of 0.13 and 3.6 ppm, respectively. Limits of detection for a 200 ml sample injection of the organometallic species were in the sub-ppm range.
One of the more interesting papers produced in this topic area compared a number of systems for the simultaneous determination of AsIII, AsV, MMA and DMA as well as SeIV and SeVI.710 The system comprised a liquid chromatograph coupled with two atomic fluorescence detectors via either HG, an ultrasonic nebuliser (USN) or a HHPN. The separation of the six species was achieved using a gradient elution. The best LODs were obtained using LC-HG-AFS (1, 0.5, 1, 0.3 and 0.7 pg for SeIV, AsIII, DMA, MMA and AsV, respectively, using a 100 µl sample injection volume), although SeVI was not detected using this method since it does not form a hydride. A comparison between LC-UV irradiation-HG-ICP-MS and LC-UV irradiation-HG-AFS has been made for the determination of MMA, DMA, AsIII, AsV and AsB.391 The UV irradiation was required to photolyse the AsB since this species will not form a hydride. Detection limits ranged from 0.1 to 0.3 µg l−1 and the precision of the two methods was comparable. Both instrumental couplings were applied to the analysis of several environmental and biological materials including sea-water, freshwater, sediments, bivalves and bird eggs. The method was validated by the analysis of a CRM (NRCC TORT-1). No influence of the sample matrix on the results obtained using external calibration was observed. A similar sort of paper by the same research group determined the same species in biological samples, but on this occasion HPLC-UV-HG-AFS was compared with HPLC-UV-HG-AAS.711 Unsurprisingly, the AFS detection system yielded better performance, with LODs for this method being 0.06–0.14 µg l−1, a factor of approximately 20 times better than AAS, and a much longer linear range (0.25–8000 µg l−1 compared with 5–500 µg l−1). The methodology was applied to the analysis of several biological samples using a methanol/water (1∶1) extraction medium and the analysis of CRM TORT-1 as a method of validation. Another paper that has utilised LC-UV-HG-AFS has used two columns (in separate analyses) to separate the As species discussed above and also AsChol.712 A new software package was developed that enabled quantification of very different concentrations of As species in the same sample. After optimisation of the instrumental performance, LODs lower than 4 µg l−1 were obtained and the precision was better than 5% RSD for all the species studied.
As well as UV irradiation, there have also been a number of other on-line sample treatment procedures that have been used to transform species into a reducible form. Liquid chromatography-thermo-oxidation-HG-AFS has been described for the determination of a number of As species (AsB, AsChol, TMAO and tetramethylarsonium) in seafood.713 A methanol–water medium was used to extract the species and each stage of the determination was optimised. The best LOD was achieved by TMAO (0.0009 µg g−1) and the worst by AsChol (0.0063 µg g−1). Recovery was better than 97% for all species and the proposed procedure was applied to the analysis of CRMs DORM-2 and NFA shrimp and NFA plaice. The results obtained were comparable to those obtained by other workers.
A column switching method for Se speciation in urine has been reported.714 After on-line microwave assisted digestion of the samples, the species selenocysteine (SeCys), selenoethionine (SeEt) and selenomethionine (SeMet) were separated on a reversed-phase column using water as the eluent and SeIV and SeVI were separated on an anion exchange column using gradient elution of an acetate mobile phase. Separation of the five species was complete in 15 min. Limits of detection ranged from 0.6 to 0.9 µg l−1.
There have been several applications of LC-AFS published. These include a paper that determined Se compounds in mushrooms715 using a variety of different chromatographic separations (size exclusion, anion and cation exchange) and two different detection systems (NAA and AFS). It was found that the mushrooms contained mainly low molecular weight (<6 kDa) Se compounds. After proteolysis, only a small percentage of the extracted species could be identified as selenite, selenocystine or selenomethionine. The majority of the Se compounds remained unidentified. Another paper has monitored the concentration in bile and the urinary excretion of As compounds in rats that had been exposed to AsIII and AsV.716 It was concluded that metabolism of the As species occurred since several other As species were found. An interesting paper has investigated the transalkylation mechanisms of organolead and organomercury compounds.618 It was found that under laboratory conditions, inorganic mercury could be ethylated to ethylmercury+ in the presence of tetraethyllead. Apparently, the transalkylation reactions take place very rapidly and almost completely. The authors concluded that transalkylation reactions occur at environmental sites where both organolead compounds and Hg2+ ions exist.
This journal is © The Royal Society of Chemistry 2001 |