Development of multiple core-level XPS spectra decomposition method based on the Bayesian information criterion

Abstract There is a need to develop an automatic spectral analysis method integrated with reference database as the reference database builds up. At the time of spectral analysis, the compound ratio is often estimated by comparing a measured spectrum with reference spectra of known single-phase compound samples. However, it is difficult to automate all processes, and there is the problem that the operator's arbitrariness is included in the analysis results. The present paper proposes a method that analyzes the X-ray photoelectron spectroscopy (XPS) spectrum of a multiphase compound using modeled reference XPS spectra. The proposed method estimates the component ratio of compounds in a fully automatic manner. To use reference spectra measured by different devices or cited from the literatures for spectra analysis, reference spectra are represented by a mathematical function using a peak separation method based on the Bayesian information criterion (BIC). In particular, it is clarified that the estimation accuracy is improved by simultaneously analyzing multiple core-level spectra rather than by independently analyzing only single core-level spectrum.