Preliminary results of an inter-laboratory study on quantitative phase analysis

Abstract

Quantitative phase analysis (QPA) of clayey materials using X-Ray diffraction data has always been a challenge. While quantitative methods based on Rietveld analysis offer excellent results for non-clay minerals, they become problematic when high crystallochemical and order degree variations are present in the phases, as is the case of most clay minerals. In the summer of 2014, the Spanish Clay Society (Sociedad Española de Arcillas, SEA) proposed a collective QPA experiment to its members. The aim was to compare the results obtained in different clay laboratories studying the same samples by X-Ray diffraction. This comparative study is a starting point for a long-term work focused on the development of an optimized QPA guide for clay minerals adapted to modern equipment and software. The results presented in this talk are those from the 2014 preliminary study in which the 22 participant laboratories analyzed three samples following their own laboratory routines. The only condition was not to use other complementary techniques for the quantification (e.g. IR spectroscopy, thermal or chemical analysis). The samples were mixtures of nearly pure quartz, feldspar, calcite, alunite, and different clays. Clay minerals included two different smectites (with low and high crystallinity), sepiolite, kaolinite and palygorskite. The set of procedures, technical characteristics of the equipment, and software used in each laboratory was very different and, as a consequence, comparison of the results is not straightforward. The results were classified into "Reflecting Powers Method" (RPM) and “Relative Intensity Ratio (RIR) – Rietveld methods” groups. As it is well known RPM is based on the peaks areas while RIR is based on the peak intensities. As expected, a high dispersion of the results was found both globally and per group. At first glance, RPM returned, on average, better approximations to the true composition of the samples. Interestingly that trend was clearer in samples with clay minerals having higher variability in crystallochemistry and crystalline range order.