Generating quantitative phase maps is unvaluable to access the phase distribution in a material. X-ray hyperspectral mapping using an energy dispersive spectrometer (EDS) attached to a scanning electron microscope (SEM) is the most practical way to collect these data, mainly due to its relatively ease of operation and availability around the world. In this work, we demonstrate an innovative technique to produce high-quality phase maps based on the correlation between each hyperspectral image pixel spectrum and a set of standards spectra. The standards spectra can be obtained experimentally from bulk specimens of known materials but also via analytical computations. The method is not only robust against electron beam current and spectrometer instabilities but also to statistical noise in the standards as well as in the hyperspectral image spectra. Equally, the method is reliable to produce phase distribution maps at low operating voltage where SEM-EDS quantitative x-ray microscopy is limited by many factors like peak overlapping, continuous background approximations and produces inaccurate phase maps. The results presented in this study will certainly open a new path to produce useful x-ray quantitative data and potentially break the locks of conventional x-ray microanalysis using EDS.
Keywords: Cross-correlation; Geological samples; Monte Carlo; Phase identification; Principal component analysis; Spectrum computing; X-ray mapping.
Copyright © 2022. Published by Elsevier B.V.