Distinctive impact responses of extraterrestrial regolith simulants: Insights from crater morphology classification and granular dynamics through machine learning and x-ray computed tomography

Impact responses of granular materials remain poorly understood, posing significant challenges to extraterrestrial exploration activities such as landing, sampling, drilling, and construction. We studied the crater morphology formed in the granular media and their granular behavior on low-velocity impact cratering by integrating three-dimensional surface scanning, machine learning-based classification, and x-ray computed tomography (X-CT) imaging. Our laboratory experiments focused on lunar (LHS-1 and LMS-1) and Martian regolith simulants (MGS-1) and terrestrial fine silica sand (T-8) as studied granular materials. The profiles of formed craters were analyzed under various experimental conditions, defined by different spherical projectile diameters and fall heights, accounting for the low kinetic energy at the granular surface contact of 0.1–0.5 mJ. In machine learning, we applied logistic regression models to classify target granular materials based on crater profile features, revealing primary morphological differences influenced by the given granular properties. The particle distribution from X-CT imaging revealed notable differences in the granular behavior of regolith simulants due to impact compared with T-8. These results underscored significant differences between the granular properties of terrestrial fine sand and regolith simulants. Additionally, our X-CT data emphasized that the high cohesion of LMS-1 significantly enhanced its resistance to impact, resulting in the porosity decline beneath the crater bottom of 0.20 and 0.03 on average in the LHS-1 and LMS-1 media, respectively, when the kinetic energy was 0.3 mJ. Our findings highlight the distinctive granular properties of regolith particles, advancing our understanding of their granular responses to impact.