Developing a powdR-based workflow for quantitative powder XRD analysis

Abstract

Powder XRD is a popular method for mineral quantification, but it often remains semi-quantitative due to difficulties such as preferred orientation in clays and feldspars. Preferred orientation effects are minimised when a full pattern summation quantification method is used. The quantitative powder XRD workflow developed here is centred around such a full pattern summation method: the R application powdR (Butler 2020). PowdR requires a reference library consisting of reference patterns of all to-quantify minerals. A reference library consisting of quartz, calcite, Ca-montmorillonite, labradorite, and microcline was developed here, and used to optimize the workflow. This optimized workflow consists of: 1) homogenising the samples in a McCrone mill; 2) scaling the reference patterns; 3) keeping the difference in intensity magnitude between reference patterns and sample pattern at a minimum; 4) using peak height and averaging multiple peak pairs to obtain the reference intensity ratio (RIR); 5) using manual instead of automatic full pattern summation, and an internal instead of an external standard, whenever possible; 6) using your own laboratory-specific reference library instead of an external one. The success of the peak-height approach is inconsistent with previous research—which has suggested a peak area-based RIR approach—and the cause of this success is unknown. When these recommendations are followed, quantification results with a relative bias of ≤10% can be obtained. When an external library is used instead, relative biases of ≥50% are reported. These large errors result in incorrect interpretations of paleoenvironment, paleoclimate, and diagenetic degree.