Specimen-mass sensitivity in geotechnical index testing of expexted Martian samples

Future Mars Sample Return missions will provide extremely limited quantities of rock and soil, posing challenges for conventional geotechnical testing, which typically requires larger specimen masses (e.g., 100s to 1000s of grams). Depending on the return material type, a single sample may contain only ~10 g of loose or regolith-like material. This study investigates how reduced specimen mass affects the results of common geotechnical index tests using the Mars regolith simulant MGS-1, focusing on water content, particle density, particle-size distribution (PSD), and loose/tapped bulk density. Reference values were first established using standard specimen masses (typically ~100 g or ASTM-recommended quantities). Tests were then repeated while systematically reducing the mass (10 g, 2 g, and 0.5 g) to evaluate sensitivity to specimen size. Results demonstrate that sensitivity to reduced mass is strongly method-dependent. At 10 g, most tests (water content, particle density, wet-prepared sieving, and bulk density) remained consistent with reference values and showed low variability. At 2 g, deviations and uncertainty increased, although some methods still followed expected trends. At 0.5 g, results were largely dominated by handling effects and measurement limitations. Dry sieving underestimated fines content compared to wet-prepared sieving and is not recommended as a stand-alone method. Hydrometer testing was identified as the main limitation for full PSD characterization, requiring larger masses for reliable results.

Overall, 10 g represents the lowest broadly reliable specimen mass for MGS-1 across most index tests, with limitations for fine-fraction PSD. While smaller masses may still yield useful data, confidence becomes increasingly method-dependent. The study demonstrates that meaningful geotechnical characterization is feasible with Mars Sample Return-scale materials, provided that deviations from standard methods and associated uncertainties are clearly documented.