Using X-Ray Computed Tomography (XCT) to Meet the Instrument Needs in a Sample Receiving Facility: Insights from Analogue Mars Samples

Introduction: The NASA Mars 2020 Perseverance rover is collecting samples intended for Earth through the Mars Sample Return (MSR) programme [1]. Approximately 20–30 samples are expected to arrive in the 2030s, providing a unique opportunity to investigate Mars’ geologic history, mineralogy, and surface processes. These samples will be processed at a dedicated Sample Receiving Facility (SRF), a Biosafety Level-4 biological containment and curatorial facility [2,3]. The SRF will conduct a three-phase characterization process: Pre-Basic Characterization (Pre-BC) for non-destructive analyses before opening the tubes, Basic Characterization (BC) for standard data collection in pristine environments, and Preliminary Examination (PE) for detailed study and subsampling. Efficient processing is critical, particularly for time-sensitive studies such as life detection and organic analyses [3,4]. SRF goals include documenting the state of the tubes and contents before opening, monitoring sample inventories, performing initial lithological assessments, and creating a detailed sample catalogue [2,3]. X-ray Computed Tomography (XCT) has been identified as a key non-destructive tool for Pre-BC, offering the potential to visualize internal structures and inform subsequent sample handling [2]. However, the specific requirements and capabilities of XCT for MSR samples remain under evaluation. This study uses terrestrial analogue cores to assess how XCT can support SRF objectives.

Methods: Terrestrial core samples were collected from the Pilbara Craton in Western Australia during the 2022 Pilbara Sample Return Campaign through the MARSnet program [5]. Samples were selected based on similarities to units at Jezero crater. The University of Alberta (UAb) received nine cores (6 sedimentary rocks and 3 regolith samples). Abrasions were made near sampling sites during sample collection, and Raman spectra were acquired. Raman Match software [6] was used to interpret the spectra and characterize the potential mineralogy of the cores. All core tubes contained sample material within an inner steel tube sealed with Teflon caps, and an outer Teflon tube sealed with caps. These sealed tubes were scanned using a Nikon XT H 225ST Industrial CT scanner at UAb, with settings of 220 kV voltage, 65–160 µA beam current, and voxel sizes of 15–19 µm. Following preliminary XCT characterization using Dragonfly software [7], five cores (Fig. 1) were opened in a clean lab at UAb for sub-sectioning. Selected fragments were rescanned individually at 110 kV and 36–61 µA, achieving a voxel size of 7.5 µm. XCT data were analyzed to visualize internal structures and determine fragment size distributions. Fragments containing features of interest were prepared into thin sections for mineralogical analysis. Quantitative point analyses were performed on thin sections using a JEOL JXA-8900 Electron Probe Microanalyzer (EPMA) at UAb, with a 20 kV voltage and 20 nA beam. Elemental X-ray mapping was conducted using a Cameca SX100 EPMA with a 20 kV voltage and 100 nA beam.

Using XCT to meet the goals of the SRF: To assess the condition of the cores before opening, initial XCT scans were performed which revealed that most samples had fragmented during sampling, and that the tubes were only partially filled (Fig. 1). However, banding features were visible in samples such as Kulja (laminated mudstone) and Monkey Mia (stromatolitic boulder), prompting the decision to rescan fragments of those cores. A fragment size distribution was generated using Dragonfly software to help plan subsampling strategies for fragments of approximately equal mass. XCT images were used to track fragment recovery when opening the core tubes. In many cases, it was possible to reorient grains to match their original position within the core, an important step for preserving contextual information critical to future magnetic analyses [2,3]. While original XCT scans helped guide initial assessments, the need for high source energy when scanning whole cores resulted in lower image resolution and diminished X-ray attenuation contrast (Figs. 2, 3), complicating phase identification [8]. To overcome this, selected fragments were rescanned at lower voltages, achieving higher resolution (Fig 3). This allowed mineral grains and cement phases to be distinguished, particularly in samples such as Koorda. The mineralogy of fragments from the sample tubes Koorda, Kulja, and Monkey Mia was initially estimated using Raman scans from abrasion patches taken near the sampling locations [5]. Once thin sections from these samples were analyzed using EPMA, it was possible to compare the Raman-generated mineralogy with the true mineralogy, and we found that they differed significantly (Table 1). This analysis allowed us to confidently retroactively colour sample fragments based on their true mineralogy (Fig. 4). Together, these results demonstrate that XCT, when paired with complementary mineralogical analyses, provides a powerful foundation for achieving the SRF’s science and curation objectives. High-resolution XCT data will be essential for guiding subsampling strategies, optimizing material allocation, and enabling high-priority, time-sensitive investigations once the Mars samples arrive on Earth.

References: [1] Farley K. A. et al. (2020) Space Sci. Rev. 216, 8. [2] Measurement Definition Team for the MSR Sample Receiving Facility Report (Forthcoming). [3] Tait K. T. et al. (2022) Astrobio., 22, S1. [4] Tosca N. J. et al. (2022) Astrobio., 22, S1. [5] Benaroya S. et al. (2024) LPSC LVI #1698. [6] Berrada M. et al. (2024) Am. Min. 0003-004X. [7] Dragonfly 2022.2, Comet Tech. Canada. [8] Hanna R. D. and Ketcham R. A. (2017) Geochem. 77, 4.