Analysis of active neutron data for in-situ planetary bulk geochemistry

Introduction: Active neutron spectrometers have been used extensively in Earth applications (e.g. in borehole/well logging analysis) to provide ~meter scale bulk geochemical information of rocks near the instrument. The Dynamic Albedo of Neutrons (DAN) instrument onboard the Mars Science Laboratory (MSL) Curiosity rover at Gale crater, Mars is the only such instrument deployed on a planetary body. Active experiments provide considerable advantages over traditional ‘cosmic-ray’ (passive) neutron experiments, albeit with a lifetime-limited neutron source. Given the high priority science targets currently traversed by Curiosity rover (e.g. Vera Rubin Ridge and the ‘Clay unit’) and future geologic units (e.g. the ‘Sulfate unit’), it is critical that DAN continues to provide statistically robust geochemical constraints (e.g. bulk water content) and is readily utilized for tactical decision making (e.g. rapid identification of hydrated phases). Using publicly-available data from MSL DAN [1], we investigate the effects of decreasing neutron generator output and provide solutions for maintaining sufficient uncertainty levels in current and future operational environments. We present a novel methodology for Bayesian analysis of active neutron data that provides a mapping of non-unique solutions and robust uncertainty quantification from random error and model noise. We also present data-driven analysis of DAN data for rapid assessment of anomalous rock geochemistries to allow for tactical decision making during operations and in future missions. This study provides several lessons learned for future active nuclear spectrometers deployed on planetary surfaces. Neutron spectroscopy: In both active and passive neutron spectrometers, ~1-3 detectors accumulate neutron count rates across ~1-3 energy bins. For active detector systems the count rates are also acquired as a function of time after the pulse of a neutron source. This increases the number of variables that can be interpreted, as the time of arrival of neutrons is sensitive to several factors such as the depth of hydrated materials. In both passive and active techniques, low-energy neutrons provide important constraints on the abundance of neutron scatters (e.g. H) and neutron absorbers (e.g. Cl and Fe) in the substrate. The common passive, i.e. ‘cosmic-ray’, neutron spectrometers are often deployed at orbit and depend on the relatively low flux of Galactic Cosmic Rays (e.g. H+) to produce neutrons in the subsurface; thus, they require long integration times and feature large sensing areas (>~100 km) when in orbit. Although active spectrometers must be deployed at the surface, requiring mobile platforms for widespread geochemical assay, they feature a neutron generator which allows for high signal-to-noise data and thus lower uncertainties. When paired with a gamma-ray detector, an even greater wealth of geochemical information is provided via time-resolved spectroscopy. In both cases, neutron spectrometers fill a ‘gap’ in geochemical remote sensing, as neutrons and gamma-rays penetrate to tens of centimeters, as compared to X-ray fluorescence (XRF) analyses which sense at microns depth, and radar which senses tens of meters to kilometers depth. Despite the many advantages of active neutron instruments, neutron generators are a lifetime-limited resource with variable output, introducing new complexities. Non-unique solutions: Most detector packages provide count rates across 1-3 energy bins, yet planetary surfaces of unknown geochemistry can include variable abundances of many more than a few neutron scatters and absorbers, making the measurements inherently subject to degeneracy or non-unique solutions. Active neutron spectrometers overcome this limitation by producing time-resolved spectroscopy using pulses of highenergy neutrons and integrating the returned low-energy neutron counts across several time bins. This analysis allows for additional degrees of freedom and thus provides a more rich array of geochemical information from the instrument. Dynamic Albedo of Neutrons Instrument: The Dynamic Albedo of Neutrons (DAN) instrument features a Pulse Neutron Generator (PNG) which produces 14.1 MeV neutrons across a <2 microsecond pulse through the following interaction: