COMPARISON OF THE TECHNIQUES OF ATOMIC-ABSORPTION SPECTROMETRY AND INDUCTIVELY-COUPLED PLASMA-MASS SPECTROMETRY IN THE DETERMINATION OF BI,SE AND TE BY HYDRIDE GENERATION

This study compares the performance of hydride generation quartz tube atomic absorption spectrometry (HG-QTAAS) and inductively coupled plas ma mass spectrometry (HG-ICP-MS) in the determination of Bi, Se and Te in 38 geological reference materials (GRMs). Samples are decomposed i n aqua regia and analytes separated from potential liquid-phase interf erents (e.g., Cu, Ni, Co) by coprecipitation with La(OH)(3). Method de tection limits using HG-QTAAS are 10 ng g(-1) for all three elements a nd are improved using the more sensitive HG-ICP-MS, to 1 ng g(-1) for Bi and Te, and 6 ng g(-1) for Se. Results for Bi by both techniques co mpare well, with typical relative standard deviations (RSD) of 3-4% fo r the triplicate digestions and analyses of the GRMs. For the most par t, results are also in agreement for Se but diverge for those samples where As is present in excess over Se by about 150-fold. In these case s, mutual interference by As suppresses the Se signal in measurement b y HG-QTAAS. Mutual interferences in HG-QTAAS are worse for the element of lowest natural abundance, Te. In the presence of As at 300-fold ex cess, recovery of the Te signal is only 50%. In the drainage sediment GSD-6 which contains As at 13.6 mu g g(-1), the value for Te by HG-QTA AS is as low as 68 ng g(-1) compared to that by HG-ICP-MS of 157 ng g( -1) and a proposed literature value of 140 +/- 40 ng g(-1). An excess of 120-fold of Se-IV also reduces the Te signal by 50%, but this is of less concern given the relative levels of As and Se in geological mat erials and can be negated in any case by preferential reduction to the non-reactive Se-0 with a small amount of KI. The only mutual interfer ence warranting attention in determination of these elements by HG-ICP -MS is that of As-V on Te where an excess of As of 5000-fold can suppr ess the Te signal by 35%. This can be negated by reduction to the non- interfering As-III by addition of KI (to a concentration of 0.005%) pr ior to analysis. Thus, HG-ICP-MS is clearly the preferred method for t he determination of Te, with precision typically at 5-7% RSD at levels of Te above 10 ng g(-1). Analysis of a subset of 16 GRMs using the mi xed-acid attack, HF-HClO4-HNO3-HCl, demonstrates that while aqua regia can be expected to result in full recovery for Bi, it provides incomp lete digestion for some sample matrices for Se and, to a greater exten t, for Te.