A statistical approach to removing diffraction from X-ray fluorescence spectra

Abstract

Diffraction peaks can lead to inaccurate elemental abundances in X-ray fluorescence maps. However, manually removing diffraction peaks is laborious while existing automated methods unnecessarily remove significant volumes of fluorescence data. Here we propose a new automated method to remove diffraction from multiple-detector spectra based on a statistical threshold. This method eliminates only the diffraction peaks from the spectra, retaining more of the fluorescence data and increasing the information content available for diffraction-free elemental quantification and mapping. By retaining the majority of the fluorescence data, the proposed method does not require an increase in dwell times or additional data to be collected to compensate for the removed fluorescence data. This method is therefore particularly valuable in instances where measurement time and data volumes are highly restricted, such as for instruments on planetary exploration missions like the Planetary Instrument for X-ray Lithochemistry on NASA's Mars 2020 mission Perseverance Rover.