Phase Analysis of Hydraulic Cements by X-Ray Powder Diffraction: Precision, Bias, and Qualification

Abstract

Abstract Estimates of cement phase composition have traditionally been made using bulk chemical analysis and the Bogue calculations, developed over 75 years ago. X-ray powder diffraction analysis provides an alternative with direct phase determination, qualitatively and quantitatively. The Rietveld method to refine multi-phase X-ray powder diffraction patterns of cement uses a whole-pattern approach with determination of the best-fit set of calculated diffraction patterns based on crystal structure models. The pattern intensities calculated in the analysis are used to estimate relative phase abundances. To estimate precision and bias and to establish qualification criteria, an inter-laboratory study involving eleven laboratories was undertaken. Four cement reference specimens were prepared using NIST SRM clinkers compounded with known amounts of one or more of gypsum, bassanite, anhydrite, and calcite. Results of the study were used to estimate the inter- and intra-laboratory precision and bias of phase abundance determinations. The data were also used to establish criteria for qualification of a laboratory, using prediction intervals that reflect the collective performance characteristics of the participants. Results show an improvement in repeatability over previous cement inter-laboratory studies utilizing internal standard-based, peak area measurement methods.