Optimized Compton fitting and modeling for light element determination in micro-X-ray fluorescence map datasets

The Planetary Instrument for X-ray Lithochemistry (PIXL) is an X-ray fluorescence instrument scheduled to fly to Mars on NASA's 2020 rover (Allwood et al., 2015). It will be capable of quantifying elements with atomic number of at least 11 using X-ray fluorescence (XRF), but the detector window blocks fluorescence from lighter elements. Important elements otherwise invisible include carbon, oxygen, and nitrogen, which can make up anions in minerals of scientific interest. X-rays scattered by all elements can be detected, so the ratio of Compton to Rayleigh scatter may be measured and used to infer the presence of elements for which there is no detectable fluorescence. We have refined a fundamental parameters model to predict the Compton/Rayleigh ratio for any given composition that can be compared to an experimentally measured ratio. We compare with a published Monte Carlo model (Schoonjans et al., 2012) and to experimental values for a set of seven materials. Compton/Rayleigh ratios predicted by the model are in good, though imperfect, agreement with experimental measurements. A procedure for consistently computing the Compton/Rayleigh ratio from a noisy spectrum has also been developed using a variation on a common background removal method and peak fitting.