A method has been developed for the determination of arsenite and arsenate in natural water samples based on the generation of arsine (AsH3) from the reaction between the arsenic species in the injected solution and tetrahydroborate immobilized on a strong anion-exchange resin (Amberlite IRA-400). Speciation was based on two different measurement conditions: (i) acidification to 0.7 M with HCl, and (ii) acidification to 0.1 M with HCl in the presence of 0.5% L-cysteine, which produced two calibration equations with different sensitivities for each species. The LOD for a 0.5 mL sample volume was 13 ng L−1 and 15 ng L−1 for arsenite and arsenate, respectively. The precision, expressed as % relative standard deviation of the measurement of 0.5 μg L−1 As was 4.3% and 4.1% for determination of arsenite and arsenate, respectively, in 0.7 M HCl; and 3.8% and 3.6%, for the determination in 0.1 M HCl and 0.5% L-cysteine. Interferences from transition metals and hydride-forming elements were eliminated by the addition of L-cysteine. The method was evaluated by the analysis of spiked natural waters. The recoveries for 0.5 and 1 μg L−1 arsenite were 92–108% and 88–112%, respectively; the recoveries for 0.5 and 1 μg L−1 arsenate were 94–111% and 95–112%, respectively. This method was also validated by the accurate analysis of a seawater certified reference material, NASS-6, which contains 1.43 ± 0.12 μg L−1 (total arsenic). The method was applied to the analysis of a number of real water samples, none of which contained arsenic below the method detection limit. The time required per measurement was less than 4 min and the procedure consumes about 100-times less hydrochloric acid that the conventional continuous-flow procedure.

Mercury is a potent neurotoxin with which food and beverages may be contaminated from a number of sources, both natural and anthropogenic. The determination of mercury at concentrations close to instrumental detection limits suffers from problems related to memory effects and loss, both during sample preparation and within sample introduction systems. L-cysteine (1%) was added to rice samples, standards, and rinse solutions in order to keep the mercury in solution and decrease the memory effect. Gold (1 μg L−1) was added online as an internal standard to improve accuracy and precision, while further decreasing the memory effect. A comparison of methods involving microwave digestion or acid extraction showed that both were capable of detecting single-digit μg kg−1 concentrations of mercury in rice. The microwave digestion ICP-MS procedure was further validated by a comparison of results with those obtained with a solid-sampling mercury analyser, based on CV-AAS, for which no significant differences were found. Both instrumental techniques were also validated by recoveries of spikes at various stages of the procedures and by the analysis of rice flour CRMs (NIST SRM 1568a and NIST SRM 1568, containing 5.8 μg kg−1 Hg and 6.0 μg kg−1, respectively). Recoveries between 80 and 120% were obtained and the concentrations measured in the CRM 1568a were not significantly different from the certified value.

A quartz crystal microbalance sensor has been developed for the determination of inorganic arsenic species in water. The gold electrode surface was modified by a self-assembled layer of dithiothreitol, and the frequency change of the modified crystal was proportional to the arsenic concentration from 0 to around 50 μg L−1, a range which spans the current US EPA maximum contaminent level of 10 μg L−1 in drinking water. As dithiothreitol is capable of reducing arsenate to arsenite, the sensor detects both species. The method was applied to the determination of arsenic in spiked rain, tap, pond and bottled water; recoveries not significantly different from 100% were obtained for a number of spike additions of less than 10 μg L−1. Arsenic was only detected in the bottled water sample, at a concentration of 8 μg L−1. This method is simple, fast, and inexpensive compared with other conventional arsenic detection methods, and has the potential to be used in the field.

Two procedures to improve the performance of the Hach EZ test kit for quantifying inorganic arsenic concentrations in drinking water have been investigated. In the first, a digital image of the colored spot formed on the test strip, obtained with a flat-bed scanner was analyzed, by the computer program Colors, for the R, G, and B values. Calibrations were constructed by plotting the B values as a function of concentration. Agreement between the experimentally determined B-values and those of the printed chart was only obtained by either increasing the reaction time (to 40 min) or increasing the reaction temperature. The precision as a function of concentration was quantified. A comparison with previously estimated values for visual comparison of the colours, showed that the improved precision of the digital analysis would produce fewer false positive and fewer false negative results at the important threshold values of 10 and 50 μg L−1. By running the test for 24 h, improved performance at the low concentration (around 10 μg L−1) end of the response scale was obtained.

An inductively coupled plasma atomic fluorescence spectrometry (ICP-AFS) instrument, was modified so that it was capable of monitoring transient chromatographic or flow-injection profiles and that sulfur molecular emission and selenium atomic fluorescence could be monitored simultaneously in an argon–hydrogen diffusion flame on a glass burner. The analytes were introduced as hydrogen selenide and hydrogen sulfide, generated on a flow-injection manifold. Selenate was reduced to hydride-forming selenite by microwave-assisted on-line reaction with hydrochloric acid, and sulfate, or sulfite, was reduced to hydride-forming sulfide by a mixture of hydriodic acid, acetic acid and sodium hypophosphite. The effects of the nature of reducing agent, flow rate, microwave power and coil length were studied. The limit of detection (3 s) for selenium was 10 μg L−1, and for sulfide was 70 μg L−1 (200-μL injection volume). The calibration was linear for selenium up to 2 mg L−1 and to 10 mg L−1 for sulfide. The throughput was 180 h−1. The three sulfur species could be differentiated on the basis of reactivity at various microwave powers.

Several procedures, involving various solvents and ultrasound, were evaluated for the extraction of four arsenic species, arsenite (As(III)), arsenate (As(V)), monomethylarsonate (MMA) and dimethylarsinate (DMA), from a silt loam soil to which species had been added at a concentration of 20 mg kg−1. The best extraction was by a two-stage procedure: shaking for 24 h in the presence of 0.1 mol l−1 phosphoric acid followed by shaking for 24 h in 1.0 mol l−1 sodium hydroxide solution. The arsenic species in the extracts were separated by high performance ion-exchange liquid chromatography, derivatized to hydrides by reaction with tetrahydroborate(III) in a multi-mode sample introduction system (MSIS) and quantified by ICP-OES. Detection limits in solution ranged from 0.4 (As(III) and DMA) to 1 (MMA and As(V)) µg l−1, corresponding to 10 and 25 µg kg−1 in a 0.2 g soil sample and 5 ml of extractant. The most significant change over time was that As(III) was converted to As(V). When each species was added individually, arsenic was 100% recovered over a period of several months. When all four species were added together, the recovery was 89%. As the precipitation of humic acids was slow, the sodium hydroxide extract could be acidified and analyzed without loss of analyte species.

A quartz tube atomic absorption spectrometric method for the determination of antimony by FI-HG was developed in which stibine (SbH3) was generated from the reaction between antimony in the injected solution and tetrahydroborate immobilized on a strong anion-exchange resin (Amberlite IRA-400). Several samples could be injected before the column was reloaded. The LOD (3s) in 4 mol l−1 HCl and 10% cysteine, and 4 mol l−1 HCl and 10% thiourea, were 0.55 and 0.54 μg l−1, respectively. The precision, expressed as %RSD (n = 5), was 5.9, 4.9, 3.8, 2.6 and 1.0 for 5.0, 10.0, 20.0, 40.0 and 60.0 μg l−1, respectively, in 4 mol l−1 HCl and 10% (m/v) L-cysteine; and 8.1, 5.0, 1.5, 1.2 and 1.4 for the same concentrations in 4 mol l−1 HCl and 10% (m/v) thiourea. The throughput was 60 h−1. Interferences from transition metals and hydride-forming elements, and signal suppression due to high ionic strength, were eliminated by the addition L-cysteine or thiourea to the samples, which also allowed the acid concentration to be decreased to 0.61 mol l−1. The method was evaluated by the analysis of spiked sea and well waters, for which no matrix effects were observed: the recoveries for 3.0 and 5.0 μg l−1 were 102% and 110–114%, respectively.

The efficiencies of four types of strongly basic anion-exchange resins in the tetrahydroborate form for the generation of lead hydride, used in the determination of lead by quartz tube (QT) atomization AAS, were investigated. Amberlyst A-26 gave the highest peak-height and peak-area sensitivities. The effects of column dimensions, tetrahydroborate concentration, loading time, loading direction, carrier reagent flow rate, carrier gas flow rate, sample acidity, and stripping coil length were studied. Without the argon carrier gas, the sensitivity was improved almost five times, though the precision was dependent on carrier agent and waste flow rate. Three procedures for the purification of potassium hexacyanoferrate(III) were investigated; a batch procedure, based on plumbane generation, was found to be the most effective. The concentration of acid was critical. Interferences from coexisting cations and anions were investigated. Hydride forming elements and phosphate interfered, but less suppression than for conventional HG-AAS was observed for some species. The limit of detection (3s) in 3% K3Fe(CN)6 and 0.10 mol l−1 HNO3 was 0.25 μg l−1, with a sampling frequency of 20–40 h−1. The precision, expressed as RSD, was 6.4% and 3.5% (n = 5) at concentrations of 3.0 and 5.0 μg l−1, respectively. The method was applied to the analysis of different types of biological matrices including natural waters, wine, human saliva and human urine. The detection limits (3s) were 0.20–0.85 μg l−1 for natural waters and 3.1–5.2 μg l−1 for the wine, human saliva and human urine. Recoveries for spiked samples were 86–110%. The results of the analyses of NIST standard reference materials, freeze-dried urine (SRM 2670), and apple leaves (SRM 1515) were in agreement with the certified values and the results for San Joaquin soil (SRM 2709) agreed with the leaching recovery values.

The abilities of various extractants to recover four arsenic species [As(III), As(V), dimethylarsinic acid (DMA), and monomethylarsonic acid (MMA)] from soils spiked with 20 µg g−1 As were investigated. The extractants were water, buffer solutions (citrate and ammonium dihydrogen phosphate), acidic solutions (phosphoric acid and acetic acid), a basic solution (sodium hydroxide) and household chemicals (vinegar and Coca Cola). Gentle shaking at room temperature with each extractant for 24 h gave different recoveries for the different arsenic species. With 0.1 M NaOH solution 46% As(III), 53% DMA, 100% MMA and 84% As(V) were recovered. A rapid extraction procedure using a sonicator probe has been developed to obtain higher extraction efficiencies. Extracts of arsenic-spiked soil, SRM 2711 Montana soil and SRM 2709 San Joaquin soil were analyzed by HPLC-ICP-MS. In the SRM water extracts, DMA and MMA were identified in addition to inorganic arsenic. The solution detection limits (3s) were 0.1, 0.12, 0.13 and 0.15 ng mL−1 for As(III), DMA, MMA and As(V), respectively for HPLC-ICP-MS.

Two manifold designs were evaluated. Water samples and wine digests in 10% nitric acid were pumped through a column containing a commercially available resin (Pb-Spec®), an immobilized crown ether with a cavity size selective for Pb2+. The column was rinsed with 2% HNO3 and the eluent, 0.1 mol l−1 ammonium oxalate was injected via a six-port rotary valve. The eluted lead was delivered to the flame atomic absorption spectrometer at 4.0 ml min−1. The following flow-injection (FI) parameters were optimized: sample acidity and volume, loading and elution flow rates, and eluent composition and volume. The detection limit for the water samples, estimated from the noise on the signal obtained for 250 ml of 10 μg l−1 loaded at 19.1 ml min−1 was 1 μg l−1. For 50 ml of wine digest loaded at 4 ml min−1, the value was 3 μg l−1. The roles of loading flow rate and sample volume were investigated in detail. The variation in retention efficiency with loading flow rate showed that the amount of lead retained (during a fixed loading time) increased with flow rate until the upper performance limit of the peristaltic pump was reached. The variation of detection limit with sample volume followed the expected hyperbolic relationship and showed that only small improvements in LOD were obtained for volumes greater than 50 ml. The method was evaluated through spike recovery for both water and wine. The lead contents of tap (0.24 μg l−1), pond (0.40 μg l−1), and river waters (not detected) were determined. The concentrations of lead in three Port wine samples ranged from not detected to 190 μg l−1. No significant matrix suppression effects were observed.

Arsenic-containing species were extracted from soil and sediment SRM by a mixture (1 + 1) of acetone and hydrochloric acid (10% v/v) in a sealed vessel in a microwave oven during heating to 160 °C at pressures up to 1150 kPa (160 psi). Following separation by anion-exchange HPLC those species which gave a volatile derivative on reaction with borohydride in acid were detected by plasma-source mass spectrometry. The procedure was used to determine the monomethyl arsonate species in SRM 2704 Buffalo River sediment (0.30 ppm), SRM 1944 New York–New Jersey waterway sediment (0.23 ppm) and in SRM 2710 highly elevated Montana soil (1.03 ppm). The method was developed by investigating the recovery of dimethylarsinate added to Buffalo River sediment as a function of various experimental parameters, including the composition of the solvent. For 1 + 1 mixtures of acetone and 5% HCl, methanol and 5% HCl, and isopropanol and 5% HCl, recoveries ranged from 91% to 112%. Similar recoveries were obtained for ultrasound-assisted extractions with the same solvents. The chromatographic eluent was not directly introduced into the mass spectrometer as, compared to the post-column hydride generation procedure, the sensitivity was too low for reliable quantitative measurements, although the chromatographic resolution was better. Problems with signal pulsations were overcome by incorporating pulse dampers into the reagent delivery lines.

The determination of trace elements in lead by inductively coupled plasma (ICP) source mass spectrometry (MS) is not possible without the removal of a substantial proportion of the lead matrix. This was achieved by the retention of lead from a 130-μl sample solution (100 mg l−1 lead in 2% v/v nitric acid) injected into a single-line (3% v/v nitric acid) flow injection manifold, on 100 mg of Pb-Spec® packed into a cylindrical column (6 cm×4 mm internal diameter). The analytes, Ag, As, Bi, Cd, Cu, Sb and Sn, passed through the column and were quantified against matrix-matched standards. Only Ag showed significant retention, but could still be measured in an 8-min run. The column was rinsed by flushing with 0.1 M ammonium citrate solution. Lead was monitored by flame atomic absorption spectrometry in preliminary experiments concerning column capacity and breakthrough. Although the capacity of the material in the dynamic, flow-through mode was less than the literature value based on equilibrium studies, the lead from up to 13 successive injections was sufficiently retained to allow accurate determination of the analytes without intermediate rinsing of the column. The precision [percentage relative standard deviation (%R.S.D.), n=5] of the procedure ranged from 1.7% (100 ng ml−1 copper) to 2.8% (5 ng ml−1 cadmium), and detection limits were in the range 0.2–10 ng ml−1. The accuracy of the procedure was assessed by the analysis of three National Institute of Standards and Technology standard reference materials (SRM 2416 bullet lead, SRM 2415 battery lead, and SRM 2417 lead base alloy). For each SRM, duplicate determinations of seven analytes were made. Of the 42 determinations, 36 fell within the confidence interval around the accepted value. Three real bullets were analyzed for seven elements by both the flow injection solid-phase extraction ICP-MS method and by aspiration of the bullet solutions (10 000 mg l−1 lead) directly into an ICP emission spectrometer. A linear least squares regression of these two sets of results gave a line with slope 1.01±0.04 and an intercept of −5±100 μg g−1, where the ± terms are 95% confidence intervals. The column lifetime was in excess of 5 months of daily use.

Lead hydride was generated from acid solution, containing potassium ferricyanide as an oxidizing agent, by the reaction with alkaline borohydride solution. The effects of reaction conditions (hydrochloric acid, ferricyanide and borohydride concentrations), and the lengths of reaction and stripping coils were studied. The effects of trapping temperature and argon flow rate were also investigated. Under the conditions giving the best peak area sensitivity, the detection limit (concentration giving a signal equal to three S.D. of the blank signal) was 0.12 μg l−1 for a 1000 μl injection volume. The detection limit was improved to 0.03 μg l−1 when the ferricyanide was purified by passage through a cation-exchange resin. Two calcium supplement materials were analyzed by the flow injection (FI)-hydride generation (HG)-electrothermal atomization atomic absorption spectrometry (ETAAS) method, giving values of 0.55 and 0.66 μg g−1, in agreement with results obtained by previously validated methods. For a 500-mg sample the limits of detection and quantification were 0.006 and 0.02 μg g−1, respectively.

Several procedures for the determination of Ca, Mg and Sr in soils have been compared on the basis of the accuracy of analysis of two NIST reference materials (Montana Soils SRM 2710 and SRM 2711). Samples were dissolved in a mixture of hydrofluoric and nitric acids in sealed vessels in a microwave oven and in teflon beakers on a hot plate. The digests obtained from both dissolution methods were evaporated to dryness in an attempt to remove silicon. Boric acid was added to prevent the precipitation of the lanthanum releasing agent (as lanthanum fluoride) and potassium was added as an ionization buffer. Determinations were made by flame atomic absorption spectrometry with both the nitrous oxide–acetylene flame and the air–acetylene flame, with calibration either by standard additions or against external standards matrix matched with respect to nitric acid, boric acid, lanthanum and potassium. The silicon remaining in the solution was also determined by external calibration. A single-line flow injection manifold was used to overcome any problems due to the presence of high dissolved solids. A volume of 300 μl was injected into a water carrier stream flowing at 8 ml min−1. To determine Ca in the air–acetylene flame, it was necessary to remove silicon. Magnesium was determined in either flame without complete removal of the silicon, however, for the determination of Sr, it was necessary to remove the silicon and use the nitrous oxide–acetylene flame. The indicative value for Sr in SRM 2710 was too low: the value determined was 360±30 μg g−1.

Several procedures, involving various solvents and ultrasound, were evaluated for the extraction of four arsenic species, arsenite (As(III)), arsenate (As(V)), monomethylarsonate (MMA) and dimethylarsinate (DMA), from a silt loam soil to which species had been added at a concentration of 20 mg kg−1. The best extraction was by a two-stage procedure: shaking for 24 h in the presence of 0.1 mol l−1 phosphoric acid followed by shaking for 24 h in 1.0 mol l−1 sodium hydroxide solution. The arsenic species in the extracts were separated by high performance ion-exchange liquid chromatography, derivatized to hydrides by reaction with tetrahydroborate(III) in a multi-mode sample introduction system (MSIS) and quantified by ICP-OES. Detection limits in solution ranged from 0.4 (As(III) and DMA) to 1 (MMA and As(V)) µg l−1, corresponding to 10 and 25 µg kg−1 in a 0.2 g soil sample and 5 ml of extractant. The most significant change over time was that As(III) was converted to As(V). When each species was added individually, arsenic was 100% recovered over a period of several months. When all four species were added together, the recovery was 89%. As the precipitation of humic acids was slow, the sodium hydroxide extract could be acidified and analyzed without loss of analyte species.

A quartz crystal microbalance sensor has been developed for the determination of inorganic arsenic species in water. The gold electrode surface was modified by a self-assembled layer of dithiothreitol, and the frequency change of the modified crystal was proportional to the arsenic concentration from 0 to around 50 μg L−1, a range which spans the current US EPA maximum contaminent level of 10 μg L−1 in drinking water. As dithiothreitol is capable of reducing arsenate to arsenite, the sensor detects both species. The method was applied to the determination of arsenic in spiked rain, tap, pond and bottled water; recoveries not significantly different from 100% were obtained for a number of spike additions of less than 10 μg L−1. Arsenic was only detected in the bottled water sample, at a concentration of 8 μg L−1. This method is simple, fast, and inexpensive compared with other conventional arsenic detection methods, and has the potential to be used in the field.

Mercury is a potent neurotoxin with which food and beverages may be contaminated from a number of sources, both natural and anthropogenic. The determination of mercury at concentrations close to instrumental detection limits suffers from problems related to memory effects and loss, both during sample preparation and within sample introduction systems. L-cysteine (1%) was added to rice samples, standards, and rinse solutions in order to keep the mercury in solution and decrease the memory effect. Gold (1 μg L−1) was added online as an internal standard to improve accuracy and precision, while further decreasing the memory effect. A comparison of methods involving microwave digestion or acid extraction showed that both were capable of detecting single-digit μg kg−1 concentrations of mercury in rice. The microwave digestion ICP-MS procedure was further validated by a comparison of results with those obtained with a solid-sampling mercury analyser, based on CV-AAS, for which no significant differences were found. Both instrumental techniques were also validated by recoveries of spikes at various stages of the procedures and by the analysis of rice flour CRMs (NIST SRM 1568a and NIST SRM 1568, containing 5.8 μg kg−1 Hg and 6.0 μg kg−1, respectively). Recoveries between 80 and 120% were obtained and the concentrations measured in the CRM 1568a were not significantly different from the certified value.

A method has been developed for the determination of arsenite and arsenate in natural water samples based on the generation of arsine (AsH3) from the reaction between the arsenic species in the injected solution and tetrahydroborate immobilized on a strong anion-exchange resin (Amberlite IRA-400). Speciation was based on two different measurement conditions: (i) acidification to 0.7 M with HCl, and (ii) acidification to 0.1 M with HCl in the presence of 0.5% L-cysteine, which produced two calibration equations with different sensitivities for each species. The LOD for a 0.5 mL sample volume was 13 ng L−1 and 15 ng L−1 for arsenite and arsenate, respectively. The precision, expressed as % relative standard deviation of the measurement of 0.5 μg L−1 As was 4.3% and 4.1% for determination of arsenite and arsenate, respectively, in 0.7 M HCl; and 3.8% and 3.6%, for the determination in 0.1 M HCl and 0.5% L-cysteine. Interferences from transition metals and hydride-forming elements were eliminated by the addition of L-cysteine. The method was evaluated by the analysis of spiked natural waters. The recoveries for 0.5 and 1 μg L−1 arsenite were 92–108% and 88–112%, respectively; the recoveries for 0.5 and 1 μg L−1 arsenate were 94–111% and 95–112%, respectively. This method was also validated by the accurate analysis of a seawater certified reference material, NASS-6, which contains 1.43 ± 0.12 μg L−1 (total arsenic). The method was applied to the analysis of a number of real water samples, none of which contained arsenic below the method detection limit. The time required per measurement was less than 4 min and the procedure consumes about 100-times less hydrochloric acid that the conventional continuous-flow procedure.

Two procedures to improve the performance of the Hach EZ test kit for quantifying inorganic arsenic concentrations in drinking water have been investigated. In the first, a digital image of the colored spot formed on the test strip, obtained with a flat-bed scanner was analyzed, by the computer program Colors, for the R, G, and B values. Calibrations were constructed by plotting the B values as a function of concentration. Agreement between the experimentally determined B-values and those of the printed chart was only obtained by either increasing the reaction time (to 40 min) or increasing the reaction temperature. The precision as a function of concentration was quantified. A comparison with previously estimated values for visual comparison of the colours, showed that the improved precision of the digital analysis would produce fewer false positive and fewer false negative results at the important threshold values of 10 and 50 μg L−1. By running the test for 24 h, improved performance at the low concentration (around 10 μg L−1) end of the response scale was obtained.