Determination of arsenic in a nickel-based alloy by flow injection hydride generation atomic absorption spectrometry incorporating continuous-flow matrix isolation and stopped-flow pre-reduction procedures

Abstract

A flow system incorporating a microcolumn of strongly acidic cation-exchange resin (Dowex 50W), to achieve continuous-flow matrix isolation, was used to eliminate nickel interference in the determination of arsenic in a nickel-based alloy by flow injection hydride generation atomic absorption spectrometry. A stopped-flow iodide pre-reduction procedure within the matrix isolation unit converted all arsenic present in the sample into As^III prior to determination with a tube-in-flame atomizer. The flow injection valve interface between the matrix isolation and hydride generation manifolds allowed separate optimization of each chemistry. After removal of the nickel, the sample stream, flowing at 2.2 ml min^–1, was merged first with a stream of 12 mol dm^–3 hydrochloric acid flowing at 2.2 ml min^–1 and then with a stream of potassium iodide solution (30% m/v) flowing at 1.4 ml min^–1. For stop times between 5 and 30 s, an average recovery of 97% with respect to As^III was obtained. A 400 µl volume of solution was then injected into a water carrier stream (flowing at 11.0 ml min^–1), merged with 3.6 mol dm^–3 hydrochloric acid flowing at 3.2 ml min^–1 and a stream of 1.0% m/v sodium tetrahydroborate solution. After passage through a 600 mm open-tubular reactor argon was merged at 400 dm³ min^–1 and the arsine separated in a glass U-tube separator. A 0.2 µm polytetrafluoroethylene membrane filter in the gas transfer line removed aerosol droplets with consequent improvement in the performance of the tube atomizer. The procedure was successfully applied to the determination of arsenic in a nickel-based alloy reference material (BCS-346) containing 50 µg g^–1 of arsenic for which a sealed-vessel microwave digestion procedure, involving nitric and hydrofluoric acids, was found to produce arsenic in the + 5 oxidation state. A characteristic concentration of 2.0 ng ml^–1 of arsenic and a limit of detection of 3.9 ng ml^–1 of arsenic were obtained with the pre-reduction procedure (which diluted the samples by a factor of 2.7) for a set of operating parameters optimized with respect to interference tolerance and throughput in addition to sensitivity. Under these conditions a sample throughput of 54 h^–1 was obtained.