X-ray fluorescence analysis of silicate rocks using fused glass discs and a side-window Rh source tube: accuracy, precision and reproducibility

In this paper, we provide a detailed account of our sample fusion, calibration and instrumentation methods for major-element whole-rock analysis by XRF, and we discuss several aspects of sample preparation and instrument performance that are important considerations for accurate analysis. The fusion procedure involves moderate capital costs and is easy to apply, yielding flat, polished, homogeneous glass discs as cast. The calibration method utilizes a least-squares procedure that rigorously fits data according to both compositional and counting statistical uncertainties. We use a side-window Rh tube for analyzing major elements (including Na) and employ real-time testing for constant count rate to reject spurious results. The methods result in excellent analytical precision and reproducibility. The standards used for calibration lie within compositional and counting statistical uncertainties of best-fit straight lines. Analyses of replicate discs and repeated analyses of single discs show excellent long-term reproducibility over several months, approaching counting statistical uncertainties in several cases. Comparison with independent measurements made by other laboratories using instrumental neutron activation and X-ray fluorescence analyses shows excellent agreement with our results. A side-window Rh tube gives increased detection limits for most major elements, but otherwise shows little difference in precision compared to a Cr tube. This means that major and trace elements can be analyzed without changing X-ray sources, which provides saving in terms of time and money, as well as being a convenience to the analyst.