Science Priorities for Mars Sample Return

V.J.Other

VI.FACTORS THAT WOULD AFFECT THE SCIENTIFIC VALUE OF THE RETURNED SAMPLES

VI.A.Sample size

Preliminary examination is necessary to make decisions on what actions to take with each sample, including how each sample is subdivided. The samples from the Apollo, Antarctic Meteorite, Cosmic Dust, Stardust, and Genesis collections provide excellent precedents for planning this step for samples from Mars. Accordingly, the discussion here is based on nearly 30 years of experience gained from such activity at Johnson Space Center. As part of preliminary examination, techniques that are non-destructive or require minimal sample mass (e.g., Raman spectroscopy, XRF, FTIR spectroscopy, laser desorption-mass spectrometry, optical microscopy, SEM, EMPA, TOF-SIMS) could be used to classify and characterize the samples (table 5). The use of non-destructive techniques would maximize the quantity of sample available for subsequent investigations by the planetary science community. Preliminary determination of mineralogy would also be required, in part to place the biohazard tests in context (XRD, XRF, EDX, electron microprobe)--toxicity of the samples to biology requires a knowledge of the inorganic species present to ensure any toxic effects are linked to a biohazard (e.g. presence of As, Cl, Br etc.)

In addition, thin sections could be prepared and curated as is done for lunar and meteorite samples, using standard thin sectioning methods for small rocks and coarse fines. Focussed ion beam milling would be used to prepare small sections if necessary; this technique is being used for all kinds of samples from the lunar (Noble et al., 2007), meteorite (Goldstein et al., 2006), pre-solar grain (Stroud et al., 2006), Stardust (Nakamura-Messenger et al., 2007) and Mars (Clemett et al., 2006) communities. For very small samples, ultramicrotomy would be used to prepare thin slices that could be distributed to multiple scientists (Figure 12 in Nakamura-Messenger et al., 2006). Destructive techniques used during preliminary examination for sample preparation should be limited to those required to prepare the thin sections and slices by these three techniques.

Life detection and biohazards testing

The most recent analysis of the test protocol for life detection and biohazard testing for returned Martian samples was published by Rummel et al. (2002; based on technical analysis done in 2000-01). There have been significant improvements in analytic methodology since then, so the list of analytical methods and the required sample sizes must be updated substantially (for example, many techniques could be performed on a thin section, and the more extensive destructive techniques could be performed on sample splits on the order of 50 to 100 mg; dependent on the concentration of organic material (Glavin et al., 2006; Elsila et al., 2005). These tests would be grouped into two categories: non-destructive (e.g., Raman and confocal Raman spectroscopy, XRF, FTIR spectroscopy, laser-desorption mass spectrometry (LDMS), and 3D tomography) and destructive techniques designed to look for carbon compounds and their molecular structures (e.g. GC-MS, LC-MS, Py-GC-MS LAL, TOF-SIMS), and nucleic acids via amplification techniques (i.e. PCR--up to 1 gram of sample may be needed for this analysis). Since the volatile inventory is critical for assessing the presence of extant or extinct biomass, we would need some way to determine the abundance of the four light elements (C, H, N, and S) likely to co-occur in biosynthesized organic matter. In addition, the draft test protocol specifies plant and animal challenge tests, which would also be destructive.

The total amount of sample to carry out the life-detection and biohazard tests was estimated by Rummel et al. (2002) as 15-25 g, although has sometimes been represented subsequently as ~10% of the returned sample. Given the near certain that the total quantity of returned sample would be relatively small, it is important that only the amount absolutely needed be used for such purposes. We need to plan for the sample size and packaging that would be needed to carry out the hazard assessment protocol.  A specific open issue is how to achieve statistically significant subsampling of the returned collection, particularly involving the rock samples.

There are two alternative strategies for allocating enough sample mass for these tests. Both strategies need further discussion by the community.

1. 
   Collect most of the incremental mass in the form of larger regolith samples (e.g. one or more samples >30 g). Since the regolith is composed of components derived from multiple geologic sources, the regolith samples would contain a mixture of rocks, dust, volcanic ash, ejecta, decomposed bedrock, etc. Moreover, all of these have interacted with the Martian atmosphere and obliquity-driven climate change. In short, they may represent an integration of Martian surface geologic processes. This might best kind of sample in which to test for the possibility of forms that proliferate on the surface during intermittent warmer/wetter intervals, and then become wind-blown constituents of the regolith. If there are significantly warmer periods during extreme obliquity there may be the possibility of intermittent proliferation of a surface microbial community that is adapted to long periods of inactivity.  Searching for spores or biopolymers (something equivalent to extracellular polymeric substances) could be a goal for regolith studies. If there is an extant microbial organism or community on Mars, it would need to be encased in desiccation, oxidation, and radiation resistant molecules. This collection plan could allow for processing individual samples through the entire test protocol.
   
2. 
   Collect rock samples 1-2 g above what would be needed for scientific purposes, so that a split could be taken from each rock for destructive hazard assessment testing. The hazard assessment protocol consists of a package of tests, each of which would have different mass requirements. Thus, in this strategy it would be possible to run individual samples through some of the tests, but other more mass-intensive tests (e.g. plant growth experiments?) may require the use of composite samples. As input to future more detailed discussion of this topic, ND-SAG offers that the rocks themselves could be the most probable habitat for Martian life. The protective coating of the rock could help retain water, protect the interior from radiation, and reduce exposure of the endolithic (rock interior) habitat to surface oxidants. Thus, ND-SAG would be uncomfortable with a strategy that did not test for biohazards in at least some of the rocks. 
   

In order to estimate the mass of rock sample that must be collected to meet analytical needs for various scientific investigations, we can turn to experience gained from the Martian meteorite collection. In 1994, a 12.02 g meteorite, now referred to as QUE-94201, was found in the Queen Alexandra Range of the Transantarctic Mountains. This sample is a basaltic rock that also contains hydrous minerals (phosphate), and evaporites. Both of these mineral types could provide information about Martian volatiles and igneous processes. Since 1994, this sample has been subdivided into 63 splits, including 27 bulk samples (4.416 g) for destructive analysis, and 13 thin sections (using 2.2 g). To date 23 principal investigators have studied the first set of splits (sub-samples), and 29 principal investigators examined splits that were created subsequently. In addition, 5.16 g of material is still available for study using new techniques or by a new generation of scientists. Of relevance to any sample return mission is the attrition measured during sample processing and in the case of QUE 94201, 0.346 g (or ~3%) were lost during processing.

The manner in which QUE 94201 was subdivided and the number of investigators involved provides a relevant analog situation that might be expected for Martian samples of similar size in a collected suite of rocks such that a rock sample could be divided into subportions that are subsequently divided for various analyses. This would allow application of single analytical techniques on one portion of a sample or multiple analyses for techniques that have low mass requirements that may reveal spatial distributions. Also, an estimate of mass required for destructive techniques part of scientific investigations is provided by the QUE 94201 example: the average mass of QUE 94201 used for destructive analysis by individual PIs is 0.2 g (based on analysis in Table 1). Therefore, if 12-15 PIs were allocated material from an individual sample from a suite, that would require ~ 2.5 to 3.0 g. Notably, QUE 94201 was not tested for organic composition. Consequently, either additional sample mass would be necessary for organic tests for science investigations that extend beyond life-detection and biohazard screening by the SRF, or all the destructive tests applied would be limited to a select number of techniques determined based on the sample

Adding up all of the currently understood proposed uses of the returned Martian samples, the minimum size for the purpose of the mission’s scientific objectives would be about 8g for both rock and regolith samples. If we assume an additional 1-2g of sample needs to be taken from each rock and regolith sample to support biohazard testing, a good standard sample size would be 10g each. Alternatively, if most of the biohazard testing is to be done on regolith samples, it may be possible to standardize on 8g rock samples, and 20g regolith samples. A very similar conclusion (10-20g samples) was reached in Appendix III by MacPherson et al. (2005).

Occasionally, rocks and sediments exhibit fabrics and textures at the mm to cm scale that are highly diagnostic of their formation and/or subsequent alteration. For example, the MER rover Opportunity documented the shapes and sizes of both grains and laminations that were consistent with the former presence of a shallow playa lake (Grotzinger et al., 2005), and these features are of a scale that is best observed in larger samples. On Earth, other rock types (e.g., igneous cumulates and high grade metamorphic rocks) also locally exhibit large-scale textures having high diagnostic value (e.g. foliation, flow features, layering, segregations, etc.). Having the capability of collecting one or more samples of about 20 g may help to correctly interpret such features. This may be achievable from two 10-g samples collected adjacent to each other (e.g. 1-2 cm apart). Alternatively, we may need to put a priority on documenting larger-scale textures in situ, so that the local context within heterogeneities larger than the sample size is documented.

Special note about the size of sedimentary rock samples

The minimal mass of samples of sedimentary deposits depends on the specific nature of the intended investigation. Experience from Earth suggests that sedimentological and stratigraphic studies normally need at least 5 g per sample in order to have a sufficient area of bed surface and internal structure to observe and document orientation of stratification, sedimentary structures, grain-size distributions, grain contacts, and mineral composition. Although we don’t know the concentration of organic molecules that might be present in returned Martian samples, studies on terrestrial samples commonly involve 10-20 g per sample. Solvent-extractable organic compounds are present in many samples in low concentrations that approach instrumental detection limits. In such cases, 1-2 g of sample is needed per measurement; however, multiple analyses are commonly required to verify molecular structures. Careful documentation of geological context is required for samples of sedimentary materials in order to relate their interpretation to the regional scale.

The likely diversity of regolith materials, particularly at a geologically complex landing site, means that a number of separate regolith samples e.g. 3, each of 1 to 25g, are preferred. A regolith sample of this mass is also likely to be appropriate for biohazard testing at the Sample Return Facility. More detailed information on sampling involving trenching or drilling to depths on the order of tens of cm is given in Appendix II. For the purpose of MEPAG Investigation IVA-5 (possible toxic effects of Martian dust/regolith on humans), it is currently estimated that a minimum of 20 grams may be necessary, although this kind of test can make use of composite samples.

Given the global homogeneity of dust on Mars (Christensen et al., 2004; Yen et al., 2005), a single sample from anywhere would likely be representative of the planet as a whole. However, because relatively pure dust deposits often are only mm thick, scooping a pure sample may be challenging in some locations. It is recommended that enough material be acquired to satisfy the needs of the various scientific investigations, as well as to provide an amount material sufficient to allow its potential hazard to humans and machines to be assessed. As discussed in Appendix III, for human toxicity studies, we need to plan for enough material to be able to conduct intratracheal, corneal, dermal and ingestion studies that would allow assessment of toxic effects. Past experience with lunar sample material and with lunar stimulant has shown that 20 grams is likely to be sufficient, but these tests could be carried out with either dust or regolith. The fraction of interest for toxicity studies is in the <20 m size fraction, and especially the <5 m fraction.

Because of the wide range of concentration of the various gas species in the Martian atmosphere, the quantity of atmospheric gas needed for measurement varies greatly among the different major species (Table 2). Also, higher analytic precision would be possible with larger samples, and multiple analyses of most species would be desirable.  Consideration should also be given to possible gas sample contamination during return to Earth and distribution of sub-samples of gas to various analytic labs. We suggest that a minimum returned gas sample should be 10 cm³ at a pressure of 0.5 bar (since ambient martian atmospheric pressure in about 0.006 bar, this would require a compressed gas sample)—this would provide enough gas material for a robust analytic program. However, if it is not possible to collect a pressurized Martian gas sample, it would be possible to make 10 determinations with a 20 cc sample of gas at Martian ambient pressure, and achieve four high priority measurements (Table 2)—although this is lower priority than a compressed sample, it is well worth doing. Finally, it should be possible to recover the headspace gas within the sample canister, although this gas will be significantly less useful for scientific purposes than a sample that has been isolated. For example, the headspace gas may be contaminated by welding byproducts during the sealing of the canister.

Atmospheric species probably would occur in some form and in widely varying concentrations in nearly all returned solid samples, either as trapped volatiles or as condensed phases such as hydrates, carbonates, or sulfates. One important property of Martian rocks is that several components are present, including primitive trapped gases and atmospheric components, and these must be resolved.  This is important for atmospheric gases, as these may have been incorporated at different times (paleoatmospheres) and may provide samples of the evolving Martian atmosphere.  Therefore, the precision of the measurements must permit these components to be resolved. Unfortunately their concentrations are typically much lower in rocks from Mars, compared to those from Earth. For example, in nakhlite NWA998, the observed gas release is typically 0.2 ppm of N per temperature step, giving an uncertainty of ~0.5‰ from zero to +150 (Mathew and Marti, 2005). The release of xenon (¹³²Xe 0.1 to 5 e^-12 cm³/g) gave (one sigma) precision of 1% for rare isotopes (¹²⁴Xe, ¹²⁶Xe) and < 5‰ for the abundant isotopes (e.g. ¹³¹Xe).  When highly variable anomalies, due to radiogenic (¹²⁹Xe), fission (e.g. ¹³⁶Xe) and spallation components (e.g. 126Xe), are observed the uncertainties increase. ND-SAG concludes from all of this that it is not feasible to set the minimum sample size of the rock samples based on their proposed use in gas-release experiments—we simply don’t have enough information to know how to set the thresholds.

A final note

A

• 
  A full program of scientific investigations (12-15 PI allocations, multiple thin sections, wide diversity of applied instrumentation, save 50% for future researchers) is expected to require samples of rock and regolith at least 8 g in size. However, for study of some kinds of heterogeneities, there may be value in one or more larger samples of ~20 g.
  
• 
  To support the sample mass required biohazard testing, either some of the samples need to be larger (e.g. 30 g), or each sample should be increased by about 2 g (endorsed) leading to an optimal sample size of about 10 g.
  

• 
  Because of the importance of the trace atmospheric species, it would be scientifically valuable to have a gas sample that is both compressed (to get more mass), and isolated from rock and mineral samples.