Simplex optimisation of conditions for the determination of antimony in environmental samples by using electrothermal atomic absorption spectrometry

Analysis of the total antimony in plant material was unsuccessful using the electrothermal atomic absorption spectrometry (ETAAS) conditions recommended by the instrument manufacturer. For this reason, an optimisation procedure utilising the Plackett-Burman method, simplex optimisation and visualisation of the generated response surface via principal components analysis, was carried out. The Plackett-Burman method was used to eliminate four of the initial variables chosen. Four variables (atomisation temperature, atomisation time, ash temperature and modifier concentration) were subsequently optimised using the composite modified simplex method and the results were visualised as a contour diagram, after reduction to two principal components. The optimised conditions were used for the analysis of both an acid digested pine needle standard reference material (NIST 1575) and a pond weed sample, collected from a contaminated site at Yellowknife Bay, Yellowknife, NWT, Canada. The total concentration of antimony present in the pine needles was statistically indistinguishable from the non-certified value, as was the value for the pond weed sample, compared with a value determined by neutron activation analysis (NAA). The results for the analysis of the pond weed sample by ETAAS agreed with those obtained from a subsequent analysis by inductively coupled plasma-mass spectrometry. (C) 1997 Elsevier Science B.V.