Multivariate analysis of pixelated diffraction data

Abstract

A novel pixelated ASIC detector using Cadmium Telluride is applied to a combined energy dispersive- and angular dispersive- X-ray diffraction system. This system is designed to obtain multiple diffraction signatures of powdered materials simultaneously. The diffraction data is analyzed using multivariate partial least squares regression utilizing the diffraction spectra at multiple scatter angles, and material concentration in a three-way regression analysis. The calibration models are used to predict unknown samples, and show that utilizing the angular information can help improve concentration prediction in samples of mixtures and has potential in material identification systems. © 2011 IOP Publishing Ltd and SISSA.