Determination of cadmium by flow injection atomic absorption spectrometry with cold vapor generation by a tetrahydroborate-form anion-exchanger

Abstract

A volatile derivative was formed on passage of an acidified cadmium solution through a strong anion-exchange resin in the tetrahydroborate(III) form. The efficiency of four resins, Amberlite IRA-400, Amberlite IRA-410, Amberlyst A-26 and Dowex-1, were investigated. Amberlite IRA-400 (160 mg, 50 mm × 3 mm id) gave the highest sensitivity for both peak height and peak area. The effects of column dimensions, concentration of tetrahydroborate, loading time, carrier reagent flow rate, carrier gas flow rate, acidity of sample, stripping coil length, the direction of loading and injection were studied and the relationship between sample volume and limit of detection were evaluated. The column was regenerated after each determination. From a study of masking agents, thiourea, L-cysteine, sodium fluoride, cupferron, EDTA, tartaric acid, 5-sulfosalicylic acid for the plant and soil matrices, it was found that thiourea and L-cysteine were suitable, whereas cupferron, EDTA, tartaric acid and 5-sulfosalicylic acid suppressed the signal. Passing the sample through an anion-exchange column before the immobilized tetrahydroborate column improved the signal for plant and soil matrices. Less suppression by some elements was observed than for generation by reaction with tetrahydroborate in aqueous solution, and smaller amounts of reagents were required. The limits of detection (3s) in 0.2 mol l⁻¹ nitric acid solution with 1% thiourea, and with 1% thiourea plus 1 mg l⁻¹ Co, were 40 and 30 ng l⁻¹, respectively, with a sampling frequency of 30–45 h⁻¹. The relative standard deviation (n = 5) was 5%, 2.0% and 3.2% at 0.5, 1.0 and 3.0 μg l⁻¹, respectively. The method was applied to the analysis of natural waters, wine, human saliva, orange juice, human urine and human hair, for which 1% thiourea and 1 mg l⁻¹ Co were added. For plant and soil analysis, 0.1% L-cysteine and 1 mg l⁻¹ Co were added. Pre-digestion spike recoveries were in the range of 81–110%. The detection limits (3s) were in the range of 0.04–0.05 μg l⁻¹ in natural water samples and 0.02–0.40 μg g⁻¹ in human hair and plant samples. The results obtained for the analysis of four NIST standard reference materials, freeze-dried urine (SRM 2670), Montana soil (SRM 2711), spinach leaves (SRM 1570a), and tomato leaves (SRM 1573a), were in agreement with the certified values.