Noise-robust analysis of X-Ray absorption near-edge structure based on poisson distribution

Abstract

We demonstrate that the noise tolerance of X-ray absorption near-edge structure (XANES) analysis can be improved by reconsidering the noise model. XANES measurement is a powerful technique to investigate the structure and electronic state of materials. In XANES measurements, the incident photon number and the absorbed photon number are measured, and the ratio of the two (called the absorption factor) is analyzed. The conventional analytical method involves fitting by the least-squares method, in which the stochastic noise is assumed to conform to a Gaussian distribution. We propose a framework for XANES analysis using a Poisson distribution as a likelihood model, focusing on the number of absorbed photons rather than the absorption factor. We validate the effectiveness of our method in numerical experiments. Moreover, we apply our method to actual solid electrolyte material data and demonstrate its practicality.

Keywords

• Bayesian statistics
• XANES
• poisson distribution
• exchange Monte Carlo method
• solid electrolyte