Planning Considerations Related to the Organic Contamination of Martian Samples and Implications for the Mars 2020 Rover

Rationale

Recommended approach

Contaminant levels of concern

Implementation strategy

2.1. Introduction to the Mars 2020 OCP

The scientific successes of recent NASA Mars missions, including the rovers Spirit, Opportunity, and Curiosity, have provided important insights into our understanding of the past history and habitability of the Red Planet (e.g., Fairén et al., 2010; Hurowitz et al., 2010; Ehlmann et al., 2011; Grotzinger et al., 2014; see also the 500 abstracts recently submitted to the 2014 8^th International Mars Conference). These research findings include evidence of liquid water at the ancient planetary surface and the presence of environments that would be conducive to the existence of microbial life (Haskin et al., 2005; Ehlmann et al., 2011; Grotzinger et al., 2014). The Mars science community, via input through the NRC Planetary Science Decadal Survey Committee, is emphatic that NASA scientific missions in the upcoming decade include an MSR mission to support a broad array of scientific and other goals (see NRC, 2007, 2011; MEPAG ND-SAG, 2008; Carr et al., 2012; McLennan et al., 2012). Returned martian samples could profoundly change our understanding of the evolution of Mars and the distribution of life in the universe. Consequently, despite the technical challenges of returning samples, the NRC gave a Mars sample-collecting rover its highest priority within the flagship class.

Subsequently, the Mars 2020 SDT was chartered by the Mars Exploration Program to define a Mars caching rover mission that would make significant and concrete steps toward possible return of samples from Mars (NRC, 2011). They recommended that the 2020 rover should carefully select and cache samples for potential future return to Earth, and noted that in order for the samples to be worth returning, they should be scientifically compelling (Mustard et al., 2013). The SDT also noted that organic contamination of the samples was a significant and complex issue that should be investigated by a future committee. Accordingly, NASA has chartered the Mars 2020 OCP. The OCP comprised 12 members with expertise in astrobiology, PP, organic chemistry and geochemistry, analytical chemistry, and contamination control/containment engineering. Ex officio members from the Mars 2020 Project Team, the NASA Office of Planetary Protection, and the Mars Exploration Program Office also provided input for the committee's research.

Within this report, the term “organic geochemistry” is sometimes used, and this may not be familiar to all readers. We use this to refer to the study of organic molecules in geological samples, such as in rocks, soils, and meteorites. Organic geochemistry is a component of the broader field of “astrobiology,” which also encompasses other kinds of investigations.

2.1.2. Charge to the OCP

It is assumed that the Mars 2020 rover would assemble a returnable cache of martian samples that may be returned to Earth, contingent on future decision-making by NASA and other sectors of the U.S. Government (see Mustard et al., 2013). The state of the samples, required to meet these potential returned sample science goals, would thus dictate the conditions for how the samples would be collected, encapsulated, and possibly returned to Earth for laboratory analysis.

One of the reasons why sample return is a high priority to the science community is because of the unique capabilities of Earth-based instrumentation and analysis (see NRC, 2011, and references therein). High among these are the abilities to undertake high-precision, spatially resolved chemical and isotopic analyses and to perform wet-chemical extraction, derivatization, and separation of organics. Both would allow the detection of specific analytes of interest with greatly increased sensitivity and specificity over robotic instrumentation (NRC, 2007, 2011; Blake et al., 2012; Mahaffy et al., 2012). This substantial improvement in measurement sensitivity is especially important to the general scientific goals of MSR, which require that any organic evidence of habitability (past or present) is both verifiable and supportable in terms of its provenance and origin; the extraordinary claim of life on Mars would require, as first noted by Marcelo Truzzi, and reiterated by Carl Sagan, “extraordinary proof” (Sagan et al., 1980).

The OCP charter outlines a specific and logical approach for generating recommendations about allowable contamination levels as input to the eventual requirement-setting process. It is expected that these recommendations would flow from sample-based measurement objectives, as illustrated in Fig. 1. The OCP was chartered with four primary technical tasks (see Appendices 1 and 2 online at http://mepag.jpl.nasa.gov/reports/OCP_2014_final_report_docx.pdf). A key ground rule is that the charter asks for a survey of existing instruments capable of measuring organics in geological samples—the OCP did not put any effort into forecasting analytic capabilities of the future:OPEN IN VIEWER

FIG. 1.

Graphic representation of the logical flow of the 2014 OCP study. M2020 (M-2020), Mars 2020 Mission.

• Based on current knowledge and capabilities, construct a list of measurements (and associated instruments/methods) anticipated to be made on the returned samples in support of objectives related to martian organic geochemistry.

• Determine the types and quantities of Earth-sourced organic contaminants of greatest concern, if they were found on the samples. Also, specify a TOC constraint.

• Assess possible implementation approaches for recognizing and distinguishing Mars-sourced organic molecules in the samples from Earth-sourced organic molecular contamination.

• Evaluate draft Mars 2020 sample organic contamination requirements and draft verification methodologies (to be provided by the Mars 2020 Project).

A note from OCP about its charter: OCP has been asked to develop quantitative recommendations in response to a question for which the answer is intrinsically unknowable. The quantity and character of terrestrial organic contamination that would not significantly interfere with future measurement objectives depend strongly on what is actually in the samples, and exactly how it is measured in the future, and both of these are indeterminate. Thus, it is more realistic to think about the limit for maximum allowable contamination of samples as having some uncertainty, rather than a single number. In order to help with NASA's requirements-setting process, the Panel ultimately has proposed a quantitative definition of “how clean is clean enough,” but in doing so, we encountered significant differences of opinion within the Panel, and we had few criteria to evaluate these differences. In addition, although science and PP both have a vested interest in the organic cleanliness of returned samples, the OCP's results are most concrete in the area of measuring and interpreting organic molecules in rocks/soils. The issues around PP's needs to implement policy and to comply with broader requirements (some of which relate to risk issues) are in part non-technical, and the OCP was not provided with criteria to evaluate this.

By focusing on levels of contamination in returned samples, our deliberations subsume a number of related activities. The Mars 2020 sampling rover would not be the only event in the “lifetime” of these samples (Fig. 2). The problem is simplified somewhat by the facts that the process we are trying to minimize is terrestrial organic contamination and that the samples are assumed to be contained while on Mars and in transit (see, e.g., McLennan et al., 2012; Mustard et al., 2013; Beaty et al., 2014). To the extent that OCP discussed flight hardware, we focused on the baseline Mars 2020 rover, including sampling apparatus and caching system. However, organic contamination issues would eventually need to be considered for the entire lifetime of the samples, particularly in regard to design of the SRF. A second consideration is that of cross-contamination (of organics) between martian materials. Although of lesser concern for the scientific objective of detecting Mars-sourced organic compounds, it is nevertheless essential to the eventual conduct of scientific research. For the purpose of limiting the scope of this study, the OCP was asked to accept the assumption that cross-contamination (from one martian sample to another) on Mars would be insignificant.

OPEN IN VIEWER

FIG. 2.

Stages in the history of the samples throughout the notional MSR campaign. The yellow box illustrates the focus of this study, although a critical point is that levels set at this stage of the campaign would impact everybody downstream. Note that within the scale of the lifetime of the Mars 2020 rover, the samples are assumed to be vulnerable to organic contamination only until each sample is sealed, which would be a matter of days, not years.

2.1.3. Introduction to the proposed Mars 2020

The SDT (Mustard et al., 2013) outlined three broad objectives for Mars 2020: (1) understand the local geology, (2) assemble a returnable cache of samples, and (3) make a contribution to preparing for the eventual human exploration of Mars. Of these, only (2) is relevant to the charge to this Panel. In this context, a “returnable” cache is one that could reasonably fulfill both scientific objectives for its study and PP needs for the protection of the public safety. Highly contaminated samples would presumably violate both requirements. In order to achieve these objectives, the proposed Mars 2020 rover would explore a site interpreted to have high astrobiology potential, conduct preliminary analyses using its onboard instrumentation, and carry sampling equipment needed to populate a cache with the most compelling samples.

As currently envisioned, the sample collecting and caching system would include a robotic arm with a turret-mounted sample acquisition tool (“drill”), a cache canister with sample tubes and plugs (Fig. 3), bit boxes for coring, brushing and abrading bits, and a blank material for quality control (Mustard et al., 2013). The SDT suggested that Mars 2020 have a minimum capability to collect 31 samples of 15–16 g each. Mustard et al. (2013) specifically called for the sealing of individual samples as quickly as possible after their acquisition. The purpose of this sealing is primarily to prevent the gain or loss of dust or volatile components, most importantly water and organic molecules. At the time of the writing of the SDT report (Mustard et al., 2013), it was not clear how tight a seal was technically possible, and the SDT recommended a threshold requirement that the seals be “dust-tight,” with a baseline requirement that they be air-tight (see their Finding 6-6). Clearly, from the point of view of the investigation of organic molecules in the samples, gas-tight seals on each sample would be a very high priority, so as to prevent their contamination after collection. If it is technically possible, we recommend that gas-tight sample sealing be a threshold-level requirement (with the quantification of the leak rate to be determined). Also, note that interpretation of potential eventual returned sample data may have less ambiguity if the sample tubes were also sealed during the voyage to Mars (this would eliminate some possible vectors for contamination), although the implementation planning is left for others.

OPEN IN VIEWER

FIG. 3.

One notional concept for the cache canister, sample tubes, and plugs to be carried on the proposed Mars 2020 rover (from Mustard et al., 2013). Alternative hardware configurations are possible (e.g., Zacny et al., 2014). The essential aspect for the purpose of this report is the left half of this diagram, illustrating an encapsulated sample.

Although Mars 2020 is a heritage-driven mission, the organic contamination considerations may depart significantly from heritage. For example, a central part of MSL's strategy involved dilution of contaminants (aka “dilution cleaning”), and given the sampling concepts for Mars 2020, it may not be possible to use this approach to clean the sampling chain. In addition, the instruments used to evaluate the quantity and character of contamination on the samples collected by Mars 2020 would be broader and more comprehensive than those used by MSL.

2.2. Previous work on organic contamination control of acquired samples

Several previous groups have studied the question of permissible levels of organic contamination in returned samples in general as well as in past/ongoing sample-based investigations at Mars. Although our current thinking on this subject certainly benefits from recent geochemical measurements of martian materials (e.g., Freissinet et al., 2014) that were not available to earlier groups, it is instructive to understand how thinking on this subject has evolved through time:

Apollo (1969–1972). The first requirements for the organic cleanliness of spacecraft-acquired extraterrestrial samples originated in the Apollo Program (see recent summary by Calaway et al., 2014). Although there was a desire to make samples available for measurement with organic contamination below 1 ng/g (Flory et al., 1969; Simoneit and Flory, 1971; Simoneit et al., 1973), this was not technically possible at the time (although this number is sometimes quoted in the post-Apollo literature). Prior to the flight of Apollo 11, organic contamination levels within the Lunar Receiving Laboratory (LRL) were observed as high as a sample equivalent of 1,000 ppm. Revamping laboratory sample handling and cleaning procedures reduced this level of organic contamination to <1 ppm for the Apollo 11 samples (Flory and Simoneit, 1972; Simoneit et al., 1973). After Apollo 11, Simoneit and Flory (1971) prepared a comprehensive analysis of potential pathways of surface contamination of Apollo samples, including contaminants arising from the LRL, Apollo lunar sample return container (ALSRC) and contents, Apollo lunar hand tools (ALHT), exhaust products from the Lunar Module (LM) [LM outgassing, venting of tanks, and Primary Life Support System (PLSS)], astronauts' suit leakage, astronauts' suit abrasion, and cleaning at the White Sands Test Facility (WSTF). Within the then ongoing Apollo program, this led to improved wet-chemical cleaning methods that led to an interpreted average of 10–100 ng/cm² of flight hardware organic contamination for Apollo 12–15 [based on analysis of the ALSRC Aluminum York Mesh witness plates (Flory and Simoneit, 1972)]. Most Apollo LRL sample handling tools and containers were cleaned at WSTF to between 1 to 10 ng/cm², although the state of their cleanliness at the time they were used to interact with samples is unknown. For Apollo 12, the Ottawa sand organic monitors showed some levels <1 ng/cm² for the high vacuum complex in the LRL, and organic contamination levels on lunar samples were reduced to below 0.1 μg/g. The level of detection for many organic geochemistry research laboratories at the time was about 1 ng/g (Calaway et al., 2014). After Apollo 14, the Apollo missions did not mitigate against organic contamination. However, it is known that some types of sample containment added organics; for example, the SNAP line used for Apollo 14 and later missions, after abandoning the high vacuum complex, added as much as 10 μg/g.

Viking (1976). The first missions to Mars to measure organic carbon in collected samples were the Viking missions. In terms of contamination control, the Viking missions were implemented with significant influences from the Apollo Program. The goal was to measure levels as low as 0.1 μg/g of total organics in small samples (100 mg) of collected soil, and a total contamination level of less than 1 μg/g (Flory et al., 1974). However, the requirement for high-temperature “sterilization” of the entire lander system resulted in an extensive screening program for all materials, which could release potential contaminants by outgassing, as well as an inventory of materials flown and analysis of their compositions. A detailed process was developed for cleaning and processing the hardware (Flory et al., 1974; Bionetics Corporation, 1990). The final cleaning of the sample path hardware in the organic analysis instrument [gas chromatography–mass spectrometry (GC-MS)] was done at the WSTF (see Seger and Gillespie, 1974): “The GCMS PDA and the collector head/shroud assemblies are sent to the NASA White Sands Test Facility (WSTF) to be organically cleaned to less than 1 ng/cm² with the procedures developed for the Apollo Moon Missions.” After sonic cleaning in triple-distilled ultrapure Freon solvent, the units were purged with low-pressure helium gas at high temperature (125°C) for 96 hr. Although there is evidence that during at least some parts of the Apollo program metal surfaces were cleaned to <1 ng/cm² of TOC, there is very little documentation for either Apollo or Viking regarding recontamination after cleaning. A second concern is that the cleanliness of the GC-MS instrument was assessed by solvent rinses followed by GC-MS analysis of the rinse fluid; thus they measured what could be removed from the surfaces, not what remained on them. It is uncertain whether a layer of adventitious carbon (AC) would be fully detected by this approach. Although a transfer coefficient was not quantified, the capability of the system to achieve its organic cleanliness goals was verified by using pyrolyzed soil (500°C) as a blank using prototype hardware. The Viking system apparently achieved significantly lower levels of contamination than the formal requirement, based on the assessment of the actual results from the GC-MS instrument during the mission, which measured residues of cleaning solvents at levels less than 1 to 50 ng/g [depending on the compound (Biemann et al., 1976; Biemann and Bada, 2011)].

Mars Phoenix Mission (2007). The Phoenix mission utilized heritage hardware not originally intended for use in organic-sensitive analyses. As a result, no organic contamination requirements were levied on the project, and organic cleanliness was approached on a “best-effort” basis. Important surfaces, including those of the thermal and evolved gas analyzer (TEGA) instrument, were open to the ambient atmosphere; thus recontamination by AC was all but assured. These problems notwithstanding, analyses of the assembled spacecraft surfaces indicated typical TOC loadings of 40–600 ng/cm². Much of this contamination, up to 85%, was a perfluorinated lubricant (“Bray-type oil”) used in the spacecraft. Remaining contaminants were largely phthalates and aliphatic alcohols, acids, and esters. Palmitic (hexadecanoic) acid was a substantial contaminant on all of the surfaces tested, but was not quantified (M. Anderson, unpublished data, 2006 ¹ ).

MSL (2011). Planning for contamination control requirements. The Organic Contamination Science Steering Group (OCSSG) was chartered in 2004 to specifically consider issues of organic contamination for the MSL mission in particular, and martian surface lander missions in general. Their report (Mahaffy et al., 2004) considered a similar range of subjects as the OCP, but focused particularly on (1) identifying contaminants of most concern, (2) understanding methods of quantifying residual contamination, (3) possible mitigation strategies, and (4) use of controls and standards.

Regarding contaminants of concern, the OCSSG considered a broad array of compounds of potential scientific interest. These included all major biomolecules, aromatic and aliphatic hydrocarbons, nitrogen-, sulfur-, and oxygen-containing compounds, sulfonic and phosphonic acids, and others. They did not explicitly include any halogenated organic compounds, as the potential importance of these compounds was not appreciated at that time. Compounds rated as being of highest scientific interest (ranked H or VH) included aromatic and aliphatic hydrocarbons, carboxylic acids, amines and amino acids, purines and pyrimidines, carbonyl compounds, and alcohols (Mahaffy et al., 2004).

To derive proposed quantitative cleanliness levels (i.e., bare minimum levels), the OCSSG relied on the calculations of Benner et al. (2000), who considered average rates of delivery of organics to Mars by meteorites, the depth of mixing of these organics into surface materials, and potential mechanisms and rates of oxidation. They concluded that “reduced organic groups such as aromatics or their oxidation products at mixing ratios of hundreds of parts per billion to hundreds of parts per million are expected.” These investigators noted that to define individual molecular species within these groups, and to definitively understand oxidation pathways, measurements down to ng/g levels might be necessary. They (Mahaffy et al., 2004) thus concluded:

Keeping terrestrial contamination to below 1–10 parts per billion in Mars samples should allow significant scientific conclusions to be reached concerning the fate of organic material delivered by meteorites. The total molecular carbon contamination allowed could be substantially higher (for example, 40 ng/g) if the contamination by specific critical species or classes was maintained at dependably constant levels. Although extinct or extant life on Mars has the potential to leave signature organic material in either much higher abundance than the parts per billion levels discussed above, the OCSSG concluded that a definitive search for such signatures could be implemented on MSL by maintaining terrestrial contamination below levels of 1–10 ng/g for relevant biomarkers.

The OCSSG went on to propose specific limits for those relevant biomarkers and for total reduced (organic) carbon in their Table 1. They also provided example calculations translating those sample-based requirements into hardware surface requirements, with the caveat that example requirements “…may be modified as the fidelity of contamination migration models increases.” Requirements for levels of specific compound classes were initially adopted essentially verbatim by the Project based on these recommendations, but were later waived as not being implementable. Reasons for this determination included limitations upon existing contamination measurement capabilities within laboratories of NASA Goddard Space Flight Center (GSFC) and JPL and the lack of resources to obtain equipment and carry out analytical methods development within the original project development time frame.

Early on, the MSL contamination control engineers conducted a trial to demonstrate analysis capability. Various analytical methods [optical and Fourier transform infrared (FTIR) microscopy, diffuse reflectance infrared (IR) Fourier transform (DRIFT)/FTIR spectroscopy, and GC-MS] were used to assemble a comprehensive set of results against the compound classes identified in Table 1. Although successful, the logistics of performing this elaborate suite of analyses were quite time-consuming and were deemed impractical for routine cleanliness verification on hundreds of spacecraft assays. The MSL team concluded that DRIFT/FTIR spectroscopy was adequate to verify the cleanliness of spacecraft hardware—including the sample transfer chain—to the level required to assure the integrity of the Sample Analysis at Mars (SAM) science goals [GC-MS, pyrolysis-GC-MS, and direct analysis in real time–mass spectrometry (DART/MS) were also used as needed on a case-by-case basis]. Rather than attempting to modify the OCSSG-proposed limits into a set of implementable requirements, MSL proceeded with a waiver against the requirements as written: the specific verifications of the subcompounds in Table 1 were waived, while the overall 40 ng/g of TOC requirement was retained. In the context of the overall validation and verification program, the verification approach implemented by MSL was deemed robust enough to constitute a low risk to the mission science goals, SAM instrument objectives, and hardware safety.

A significant contribution of the OCSSG effort was to conduct and report experiments designed to understand the efficiency of physical transfer of organic contaminants from spacecraft surfaces to a simulated martian regolith. They considered a variety of organic compounds, regolith analogs, and metal surfaces, over temperatures ranging from −40°C to 25°C. In general, they found that the most significant factor affecting transfer efficiency was the level of abrasion of the metal surfaces: strong abrasion of surfaces resulted in up to 60% transfer, moderate abrasion produced 1.3–7.6% transfer, and passive transfer without abrasion yielded levels of 0.1–1.3%.

Actual performance. The history of the state of organic contamination of MSL's sample contact surfaces as a function of time is as follows:

1. Initial cleaning. MSL sample acquisition, processing, and handling (SA/SPAH) hardware surfaces were cleaned to an initial level of approximately ≤25 ng/cm² (average) as determined by FTIR analysis (Herrick et al., 2002) of dichloromethane solvent rinses of witness coupons that accompanied individual piece-parts through the cleaning process. Surface residue was not quantified below 20 ng/cm² because the requirement was 100 ng/cm².

Since direct verification of in-sample contamination levels is not technically possible at Mars, the predicted values for contamination quoted above are necessarily model-based. The model used predicts contributions of the sample transfer chain only, and does not include additional contamination within the SAM instrument itself (the MSL Project had an internal allocation of 4 ng/g out of the total 40 ng/g requirement for SAM).

Note that the SAM instrument had contamination issues during the first 2 years of the MSL mission as a result of leakage of the derivatization solvent within the SAM instrument (and these issues are continuing up to the time of writing). Although this is a contamination-related concern, it is the result of a mechanical anomaly that is unique to SAM, and it has no significance to Mars 2020. This is not discussed further in this report.

Origins Spectral Interpretation Resource Identification Security—Regolith Explorer (OSIRIS-Rex) (launch planned in 2016). A Level-1 requirement for OSIRIS-Rex (see Lauretta et al., 2014) is to return >60 g of pristine asteroid regolith. During initial formulation (pre-Phase A), OSIRIS-REx adopted the OCSSG contamination limits, but it became clear during Phase A that these were not achievable on a New Frontiers budget. OSIRIS-REx formed a working group with representation from the science team, flight system, instruments, systems engineering, and project management to ensure a broad and balanced consensus. In the development of contamination control requirements the first obstacle was to define pristine as having contamination levels from the addition of species that do not interfere with the scientific conclusions. It was determined that alteration via loss of material (e.g., water), was beyond the scope of the mission. The OSIRIS-REx contamination control requirements for organic and inorganic species were drawn from lessons learned from Stardust (Sandford et al., 2010), from which OSIRIS-REx derives significant heritage. In addition, to avoid complicated analytical testing, proxies for compounds and species of concern were identified based on carbonaceous chondrite abundances and Stardust performance. These were converted from nanograms per gram of sample into nanograms per square centimeter of collector [touch-and-go sample acquisition mechanism (TAGSAM)] surface area assuming 100% transfer efficiency (although a vibration test show ppm levels of amino acid transfer under dry conditions), and shown in Table 2. It is important to remember that the OSIRIS-REx target is expected to be orders of magnitude richer in organic compounds than found on Mars.

To implement these levels, specific testing for amino acids on sampling hardware via witness plates is used. Amino acid requirements led to a ban on nylon, rubber, and latex use on all parts of the spacecraft. The hydrazine is limited by spacecraft design and operations. The other species are controlled by converting these surface requirements to particle and film requirements for theoretical IEST-STD-CC1246D and fractions of A. This results in a requirement of 100 A/2 under reasonable conditions.

OSIRIS-REx has a witness plate plan that includes assembly, test, and launch operations (ATLO) activities, gas samples, curation, and a series of sapphire and aluminum witness plates on the flight hardware. The exposure of these witness plates is adjacent to the sample collector and timed by mission phase. The flight witnesses will be returned with the samples. Analyses of ground witnesses will be performed off the critical path and under the watch of the Sample Analysis Working Group lead. Ground and flight witnesses will be distributed under the same rules as the asteroid samples.

A materials archive of ≥1 g of components and polymers with a reasonable path to the samples is planned. Determining which materials to archive required a high-fidelity outgassing model of the spacecraft. To minimize the diversity of materials and extent of the archive, the diversity of polymerics is minimized and complex polymers (e.g., silicones) are only used by waiver when no viable substitute can be found. Since the curation of hydrazine under flight-like conditions is impractical, flight hydrazine is analyzed before launch so that hydrazine that is identical both isotopically and in trace components can be mixed and fired through an archived thruster and catalyst bed if required by future researchers.

ExoMars (launch planned in 2018). The ExoMars 2018 Rover is planned to be the first life detection mission to Mars to be undertaken since Viking (for details, see Vago et al., 2013). Its plans are to employ certain contamination control-related practices that were common to Viking's design and integration, including:

• Design to protect the sensitive surfaces based on segregation (sealed sample path) and overpressure (for example, pre-launch pressurization of the canister containing the drill)

As of August 2014, the requirements for the maximum amount of terrestrial organic contamination per gram of martian samples to be used for life detection on the ExoMars 2018 rover are as follows (Kminek, written communication, 2014):

• Material from biological sources ≤50 ng/g

This is the contamination level applied to the subsystems involved in the acquisition, delivery, and analysis of martian samples for life detection. The philosophy of structuring the list this way is to have a lower level for biological organics and a higher level for engineering sources that have previously been characterized with standard outgassing testing, specific GC-MS analysis [protocol developed with the Mars organic molecule analyzer (MOMA)], and samples sent and tested by the MOMA team.

The ExoMars Rover instrument contents and sample pathway are planned to be assembled in an aseptic, International Organization for Standardization (ISO) 3 environment with ISO airborne molecular contamination class-9 (AMC-9) cleaning. In order to have empirical data regarding assembly and testing contamination data, the build philosophy contains two models that are planned to be employed to measure contamination sources during assembly and environmental qualification testing. During integration, indirect verification are planned to take place using witness plates, and routine cleanliness checks would be conducted with gravimetric analysis. End-to-end testing of the flight model is planned to verify cleanliness levels prior to launch. The MOMA instrument on the ESA ExoMars rover mission is being built (by NASA GSFC) with a required level of surface cleanliness of 10 ng of non-volatile residue (NVR)/cm².

MSR (MSR SSG II). The MSR Science Steering Group II (MSR SSG II) was formed in 2004 and chartered with re-examining the MSR goals and mission concepts in light of the recent (at the time) results of the Mars Exploration Rovers (Spirit and Opportunity). This group comprised four subpanels tasked with studying a broad range of issues; one of those was forward organic contamination (MacPherson and the MSR-SSG II, 2005). As with the OCSSG, they began by compiling a list of potential “indigenous martian organic molecules” (Table 6 in MacPherson and the MSR-SSG II, 2005) that included most of those from the earlier report, plus humic and fulvic acids, complex macromolecules, phospholipids, N- and S-containing gases, and hydrocarbon gases. They also ranked compounds in terms of relative scientific interest, and here they departed significantly from the OCSSG report (Mahaffy et al., 2004). The highest-rated compounds of interest (ranked VH) were C1–C4 hydrocarbon gases, saccharides, amino acids, porphyrins, glycerides and phospholipids, and nucleic acids. The fact that only amino acids were ranked of “very high” scientific interest by both the MSR SSG II and OCSSG panels emphasizes the inherent difficulty in prioritizing the most interesting molecular targets. The MSR SSG II further stated that, considering the reactivity and volatility of these species, strict temperature control of samples during their return is likely necessary. They proposed a limit of less than −5°C as being desirable, <20°C as highly recommended, and <50°C as essential.

To derive quantitative limits for contamination, this group first considered those that had been recently proposed by the OCSSG. They comment (MacPherson and the MSR-SSG II, 2005):

In light of the great variety of state of the art analytical techniques that could be applied to organics analysis of returned samples; it would be desirable to realize even lower thresholds in a returned sample than these proposed thresholds for in situ analysis. For example, thresholds of less than 10 ng/g in total organic contamination and ng/g or below in the proxy compound classes listed above could enable robust conclusions to be drawn regarding the origin and processing of indigenous Martian organics found in the ng/g or higher abundances.

Next, in an effort to set a lower bound on what might be required for analyses of Martian rocks, the group considered analogous organic-poor rocks on Earth, as follows. Oxidized red sandstones and mudstones (“redbeds”) commonly contain 0.10–0.01 weight percent TOC. Of this, typically less than 10 μg/g is extractable bitumen, and sometimes as low as 1 μg/g. Major compound classes, such as aliphatic hydrocarbons, often represent just 1% of that total extract, or 10 ng/g. Individual biomarker molecules, for example, n-alkanes and hopanes, are themselves typically only 1–10% (or less) of aliphatic hydrocarbons, and so are present at 0.1–1.0 ng/g (MacPherson and the MSR-SSG II, 2005, pp 29–30). Inferring that similarly low levels might be present in returned martian samples, the MSR-SSG II proposed organic contamination thresholds that were lower than those of the OCSSG by a factor of 4 (i.e., 10 ng/g of TOC), and as low as 0.25 ng/g amino acids. Critical to their evaluation and conclusions is the assumption that terrestrial TOC-poor rocks are representative of what we might find on—and return from—the surface of Mars. The validity of this assumption was not explicitly discussed by the SSG II report, and was questioned by several panelists in the OCP.

MSR (Mars 2020 SDT). The most recent group to take on the subject of organic contamination was the Mars 2020 SDT (Mustard et al., 2013). This group provided a technical analysis of the Mars 2020 objectives, and looked closely at the susceptibility of those objectives to terrestrial organic contamination. They concluded that Objective C (sample return) was very highly susceptible to organic contamination, and provided a high-level strategy for controlling and characterizing organic contaminants. With regard to specific contaminant thresholds, the SDT did not perform any further analysis of scientific requirements, and instead recapitulated the proposals of earlier studies:

The degree to which interpretation of analyses of martian samples would be compromised by the presence of organic contaminants in samples containing indigenous martian organic material is unknown. Thus, we do not know what level of cleanliness would be appropriate. Contamination should be kept as low as reasonably possible and within the guidelines proposed by the MEPAG OCSSG and the MSR SSG report. In these reports a total of 40 ng/g reduced organic compounds, with sub-allocations of 1–10 ng/g for specific compound classes was proposed by OCSSG (2004) (this spec was specifically intended for in situ investigations, including MSL). The MSR SSG (2005) proposed a total of 10 ng/g of reduced organic compounds, with sub-allocations for specific compound classes—proposed for at least some MSR samples. These figures are estimates only of contamination levels needed to achieve the science objectives. As discussed, different levels may be required to meet planetary protection requirements, and those levels would be specified by advisory groups specifically chartered for that purpose.

The SDT ultimately proposed adopting the lower of the two earlier limits (10 ng/g of TOC) as the baseline (i.e., desired) requirement, and the higher of the two (40 ng/g of TOC) as the threshold (i.e., bare minimum) requirement. Although the text of their report implies that they supported the inclusion of separate thresholds for individual classes of compounds, their finding (Finding 6-12) does not explicitly state this. (See Section 4.1.3 of this report for the hybrid resolution to this issue proposed by the OCP.)

Summary. In summary (Table 3), three conclusions can be drawn from the above history: (1) It is technically possible to clean spacecraft sample-contacting surfaces (on Earth, prior to launch) to TOC burdens as low as 1 ng/cm², or perhaps even lower. (2) Previously cleaned sampling surfaces become recontaminated up to the point they are used to interact with samples, such that the state of their cleanliness is significantly worse (and hard/impossible to measure) than at the time of original cleaning. (3) Samples have been collected and analyzed by prior missions, or are proposed by future missions, with Earth-sourced organic contamination as low as about 40–50 ng/g.

2.3. Key concepts

2.3.3. Not all contaminants are equal

One approach to limiting organic contamination of the returned samples is to adopt limits on TOC contamination. However, certain contaminants can be more problematic than others if they directly interfere with analytes of interest. Without knowing the chemistry of the returned samples, it is difficult to propose specific molecules to avoid. Nonetheless, complex contaminant profiles generally make it much more difficult to detect target analytes than a few well-characterized contaminants. Therefore, both the complexity of contaminating compounds as well as the total amount of organic contamination (Fig. 4) can affect the interpretability of the data.

OPEN IN VIEWER

FIG. 4.

Complex contaminant profiles are, in general, much worse than a few well-characterized contaminants.

Major Finding #4:

Reducing those specific contaminants that interfere with compounds of scientific interest is as important as reducing the total contamination burden.

2.3.4. Contamination control versus contamination knowledge

The Panel believes it is important to distinguish between the concepts of “contamination control” and “contamination knowledge.” The former represents the efforts needed to reduce contamination on the spacecraft (and eventually the samples), and to maintain that level of cleanliness (Fig. 5). It is inherently part of the project design, engineering, and fabrication effort, and is implemented to achieve quantitative requirements (but generally no more). In contrast, contamination knowledge represents our need to understand in as much detail as possible what remains on—or is added to—the spacecraft. It is performed by the project team, but would provide foundational knowledge for the science team studying the samples. In many ways it should be considered as the first stage of scientific investigation. It has neither easily quantifiable limits, nor easily proscribed methodology. Indeed, when conducted to its fullest extent, it should more closely resemble scientific research than engineering design and testing.

OPEN IN VIEWER

FIG. 5.

Graphic representation of the elements of a viable organic contamination management plan for Mars 2020.

Two important points emerge from this discussion. The first is that both contamination control (strategies #1 and #3 in Fig. 5) and contamination knowledge (strategies #2 and #5 in Fig. 5) are necessary and complementary. The second is that they would likely require fundamentally different approaches. Adequate consideration should be given to the necessary personnel, resources, and strategies required for both to be successful.

Finding #5:

Our ability to correctly interpret data from partially contaminated samples correctly depends on a combination of (1) minimizing contamination at the start, (2) characterizing residual contamination, and (3) minimizing and characterizing recontamination. All three are equally important.

2.4. Science and PP objectives both drive need for organic analyses

Some of the scientific and PP objectives of MSR would both rely heavily on careful measurements of organic molecules in the returned samples (Fig. 6). There are different ways of organizing the way various interpreters think about the data [e.g., the null hypothesis (see Kminek et al., 2014)], but regardless of how the hypotheses are structured, all would make use of the same analytic data from the samples. Once samples are available for Earth-based analysis, they would be analyzed for their organic contents to the maximum degree possible. For both science and PP applications, a central part of the problem is to reliably interpret the geochemical/biological context of organic molecules on Mars. Whether or not they are indigenous to Mars is a key interpretation, and for which it would be very important to understand the degree of definitiveness. Given that sample mass would be limited, we anticipate that it would neither be possible nor necessary to make separate measurements by “science” and “planetary protection” teams—both would have a shared need for accurate data from precious samples.

OPEN IN VIEWER

FIG. 6.

Science and PP objectives would both drive the need for analysis of potential returned samples for organic molecules.

In this context, the OCP carefully considered ways to organize a set of measurement goals related to organic molecules in any future returned samples that would constitute an approach that may be useful for both science investigators and PP technical evaluators (Fig. 6). A key uncertainty lies in the fact that the samples would arrive in terrestrial laboratories as inadequately characterized objects—we would have at best fragmentary information about the concentration and identity of contained organic molecules. This leads to two logical phases of investigation (illustrated by the upper and lower parts of the simple table in Fig. 6): (1) Do the samples contain organic compounds?, and (2) What is the nature and origin of any compounds discovered? The purpose of pointing out this intrinsic sequential approach is that the instruments and detection limits are somewhat different in the two phases, as are the sensitivities to organic contamination (all amplified in the next section). It is not our intent to comment on the important details that make up the life detection and hazard assessment protocols, both of which are well outside the scope of the Panel. However, we note that biomolecular organic analyses are now sufficiently sensitive to detect even the smallest possibly viable contaminant populations.

3.1. Need for early survey measurements

Because of the wide spectrum of available instrumentation, and the unknown character of the samples, it would not be possible to describe in advance a specific analytical plan for each sample. The investigation pathways would, of necessity, need to be driven by discovery (Fig. 7). The analytical scheme for any given sample would be contingent on its size, organic content, environmental and stratigraphic context, lithology, surface exposure age, thermal history, etc. Different samples may take quite different analytical pathways. It is for these reasons, as well as those of PP, that initial survey measurements would be needed to determine whether organics are present and to provide first-pass characterization. This information would then be used to prioritize (presumably via peer-reviewed proposals) a more targeted series of measurements geared toward more specific research objectives. Given the recommended integration of science and PP measurements, early survey measurements would need to be performed while the samples are still isolated in the SRF, which would pose technological challenges. Ensuring that the SRF would be capable of supporting such survey measurements should thus be a key consideration in its design.

OPEN IN VIEWER

FIG. 7.

Diagrammatic illustration of the relationship between early survey measurements and later targeted analyses.

3.2. Potential analytical methods for returned samples

As part of its research, the OCP compiled a list of analytical methods that are typically employed to investigate samples for organic constituents (Table 4). A confusing but important aspect to consider in comparing these methods concerns the units of measurement. For example, different techniques would yield results that are properly expressed in units of mass/area, mass/mass, mass/volume, ppb, etc. Depending on details of sample introduction, some of the instruments and techniques are inherently sensitive to the total mass of molecules in a sample aliquot (e.g., mass spectrometry), while others are sensitive to their concentration per unit area or volume (e.g., optical spectroscopy). As a concrete example, the absolute detection limits for mass spectrometers are most properly expressed in units of mass (nanogram) or moles (nanomole). Thus, to apply a mass spectrometric detection limit to a returned sample, we require knowledge of how much sample would be assayed to derive ppm (μg/g) or ppb (ng/g) by weight. For surfaces, the area of sample to be assayed is needed to derive a contamination limit (e.g., ng/cm²).

OPEN IN VIEWER

Table 4.

Science/Planetary Protection Questions, Measurement Objectives, and Instrument Techniques

a

See Appendix 4 for instrument sensitivities, preparation requirements, and notes.

A further complication is that when we specify the mass of an amino acid (e.g., 1 ng of glycine) detected in a mass spectrometry assay, we are specifying the total mass of that compound. Knowledge of the molecular formula then allows conversion to an equivalent mass of carbon. For example, a molecule of squalene (common contaminant from fingerprints) weighs almost six times as much as a molecule of glycine and, molecule for molecule, has 15 times more carbon. Considerations such as these must be factored into understanding the relationships between measurements such as TOC (which measures carbon mass) and those such as analyses of specific amino acids (which target total analyte mass).

One of the key outcomes of this exercise was recognition that there are so many potential measurement techniques, in many cases with extraordinarily low detection limits, that it is impossible to avoid contamination that would affect any of them. We cannot know in advance every contaminant that might matter, but we can predict ones that definitely would matter (see discussion in Section 4). Without some means to narrow our focus on a smaller subset of methods, they cannot serve as a useful basis for establishing appropriate limits on contamination. Any attempt to simultaneously consider all such methods must necessarily reach the conclusion that virtually all organic compounds must be limited to vanishingly small concentrations.

As a means to broach this apparent conundrum, the OCP decided to parse prospective analytical methods into those most suitable for use in initial survey measurements (hereafter, referred to as “survey methods”) versus those more suitable for targeted measurements (in the sense of Fig. 7; “targeted methods”). This is an admittedly arbitrary distinction, in that many instruments could be used to make either survey or targeted analyses. For example, the same mass spectrometer operating in full-scan mode can provide a useful survey of what is present, whereas when operating in selected-ion or tandem mass spectrometry (MS-MS) mode would provide a more sensitive and detailed picture of particular analytes. Nevertheless, this distinction is useful in that we know we would utilize survey methods on the returned samples, but the use of targeted methods would be contingent on what is present. Ideal characteristics of survey methods would be (1) broad sensitivity to a wide range of organic analytes, (2) minimal sample consumption, and (3) suitability for use in a containment facility. The latter consideration rules out, for example, synchrotron x-ray spectroscopy.

3.3. Survey versus targeted analytical methods

Most analytical methods exhibit an inherent trade-off between specificity and sensitivity. The ability to confidently detect and identify a given analyte is governed by the measured signal-to-noise (S/N) ratio. Signal is largely a function of the sample concentration and physics (ionization cross-section, quantum fluorescence yield, etc.), so the most effective way to boost sensitivity is to reduce noise. Noise in the context of a complicated, geologic sample is mostly determined by the number and abundance of other (i.e., non-analyte) molecules being detected. The primary means to reduce noise (and so increase S/N) is thus to screen out signals from everything other than the analyte of interest (e.g., only looking at one particular mass or wavelength). In other words, the act of simultaneously looking for everything would generally decrease our ability (sensitivity) to detect any one thing in particular. We conclude that survey analytical methods are likely to be inherently less sensitive than are targeted analytical methods.

Two categories of likely survey methods can be recognized. Surface imaging and surface spectroscopic/spectrometric techniques are minimally destructive and so likely to be used at an early stage. Their strengths include minimal sample consumption, in situ analysis with excellent spatial resolution, and broad sensitivity to many analytes. In situ observations also generate context for the organics that can be related to the bulk measurements. However, identifying and quantifying specific individual molecules present in complex mixtures are limited using these techniques, and so other methods are needed for complete molecular characterization. Mass spectrometric techniques used with bulk or extracted samples are destructive of samples, but currently provide the best ability to look sensitively for a broad range of organics, and to identify them and their specific characteristics (e.g., isomeric form). Most recent studies of meteorites have used mass spectrometry in some form. Ultimately, some combination of non-destructive and destructive techniques would likely be used to establish the survey analyses of samples.

Based on the information compiled in Appendix 4, the OCP concludes that the detection limits for surface spectroscopy are typically in the ppm range, whereas for mass spectrometry, detection limits are in the ppb range. An important implication is that surface spectroscopic measurements should not drive the contamination limits. In contrast, mass spectrometric methods cover a very wide range of potential analytes (and indeed are valuable for this ability), and their detection limits are sufficient to identify the principal organic molecular components in geological samples at the nanogram level.

“Targeted” methods for organic characterization would be those that look more sensitively for a reduced number of specific components. The decision of whether or not to employ any particular targeted method would be based on a combination of what is known to be present in the samples, perceived importance of the target analytes, sample availability, and other factors. Some targeted analyses (e.g., for amino acids) might be deemed so essential that they would be undertaken no matter what is found by initial surveys. Others (e.g., for hydrocarbons) may not. Regardless, knowledge from the survey measurements of which compound types are present would enable prioritization of which targeted methods to use, and so are more essential than targeted methods that may or may not be used. The OCP therefore decided to focus on survey methods for the purpose of establishing tractable contamination limits (Figs. 7 and 8).

OPEN IN VIEWER

FIG. 8.

A two-step analytic process for organics—survey versus targeted.

3.4. Survey analytical methods

The most likely initial survey measurements would employ one or more of a range of surface spectroscopy tools. In part, this would be driven by the need for the initial investigations of returned samples to be non-destructive. These include FTIR, IR reflectance spectroscopy, confocal Raman spectroscopy, and deep UV Raman/fluorescence spectroscopy. These methods are non-destructive, but in most cases do not allow the identification of individual molecules in complex mixtures. On the other hand, they generally do not require significant sample preparation and work at small spatial scales. Therefore, they can be very useful for detecting organic carbon aggregates on the surface or within inclusions in samples.

Surface spectroscopic techniques can be complementary and can precisely survey a sample for surface organic contamination, including contaminant heterogeneity and general type. In decreasing order of organic sensitivity, these are (1) deep UV Raman and fluorescence spectroscopy, (2) confocal Raman spectroscopy, (3) FTIR spectroscopy, and (4) IR reflectance spectroscopy. In conjunction with their capabilities for organics detection, each has strengths for mineralogical assessment, which is critical not only for distinguishing between native organics and contamination, but also for determining detection limits and detection depths.

4.1. General considerations

4.1.2. S/N threshold for acceptable contamination

A common misconception is that robust analytical measurements require background (blank) levels of a particular analyte to be well below those one is trying to measure. As an example, measuring glycine with an S/N level of 3 (a commonly accepted threshold) might be misconstrued as meaning glycine concentration in the blank must be threefold lower than that in the sample. This is incorrect. Our ability to confidently resolve the presence or absence of an analyte requires only that we can determine at some statistical confidence interval (say, 99.7% or 3σ) that the measured concentration is greater than that of the blank. Thus, as illustrated in Fig. 9, we need to know both the concentration of a particular contaminant in our blank and the uncertainty in our knowledge of that concentration. This uncertainty could result from incomplete characterization, or from physical variability (whether spatial or temporal) in its abundance, or both. This leads to two important points. First, contaminant concentrations do not necessarily need to be lower than detection limits for a measurement to be successful. Second, a stable and well-characterized background is as important as a low background.

OPEN IN VIEWER

FIG. 9.

A diagram showing how different levels of signal and background support or deny a successful measurement. The vertical arrows indicate variability in background levels.

Finding #11:

Confirming a measurement to be statistically significant does not necessarily require that contamination be below the detection limit, only that it be acceptably low, stable, and well characterized. Consequently, the contamination knowledge program should address the composition, abundance, and variability of organic molecules present on the spacecraft.

In practical terms, the variability of background contamination is very difficult to control, and is likely to be strongly dependent on the particular analyte. Certain volatile organic compounds that are ubiquitous in clean room air as a result of degassing from permanent equipment (flooring, duct work, paint, etc.) might prove to be present on sampling hardware at very predictable levels. Other compounds, for example, those present in cleaning products used intermittently, might be much more variable. Particulate organic contamination would have a statistical spatial variability determined by its Poisson mean concentration. Obviously, the more variable a particular contaminant is, the lower its average concentration must be to ensure that a particular sample concentration could be confidently measured.

Variability is also influenced by the scale of sampling. For example, contaminant concentrations can vary between samples, as well as spatially within a single sample. An organic particle represents a relatively large local concentration, but if there are few such particles, then could also represent a small concentration averaged over the entire sample. In recognition of such issues, we recommend that the appropriate scale for averaging contamination levels to compare with specified limits is that of the individual sample core. In other words, if the specified limit is 1 ng/g, then each and every sample that is collected should meet that limit (rather than the average of all samples). Conversely, we do not recommend that subportions or aliquots of any given sample be required to fall below the limit. As long as the average concentration for a sample falls below the limit, then portions of that sample could reasonably exceed the limit.

4.2. Considerations related to specific contaminants

The OCP considered two distinct but related questions in relation to contamination by individual organic molecules: which ones do we care about, and at what level? Although the two are intimately linked, and need to be (and have been) considered iteratively, we separate them here for the purpose of clarity.

4.2.2. Allowable levels of contamination

In order to arrive at plausible and defensible limits on individual organic contaminants, the panel considered three related criteria as described above and in Fig. 10. Of these, expected concentrations of targeted analytes are the most relevant to scientific goals, but are unfortunately also the least constrained. We therefore considered available data for all three criteria in the following sections. In doing so, we have made very broad generalizations about concentrations and detection limits for all organic compounds in all sample matrices. We point out that this is a very substantial oversimplification, for example, detection limits can vary by at least an order of magnitude between compounds or matrices (Watson and Sparkman, 2007). On the other hand, the Panel did not feel it was productive to undertake the huge effort required to develop detection limits and cleaning standards for every different class of analyte and sample material, given the large uncertainties in what we expect to find in the final samples. The reader should therefore treat the following discussion as best presumed, rather than precise values.

OPEN IN VIEWER

FIG. 10.

Determining allowable levels of contamination (see Sections 4.2.2.1–4.2.2.3).

4.2.2.1. What analyte concentrations do we expect?

Preliminary estimates of Mars in situ organic concentrations and molecular distributions can be derived from four sources: studies of martian meteorites found on Earth, lander- and rover-based in situ measurements, analogous terrestrial environments, and estimates of delivery rates of meteoritic organics. Studies of martian meteorites are the most detailed and comprehensive of in situ martian concentrations, but suffer from at least one significant drawback; existing meteoritic samples exclusively represent igneous martian rocks, which are not the principal rock types likely to be examined for life detection on Mars (Grotzinger et al., 2014). On one hand, martian igneous rocks might be expected to contain far less organic material than a sedimentary rock bearing the remains of abundant biota. On Earth, igneous rocks commonly contain <0.01% organic carbon (primarily as graphite), whereas organic-rich sedimentary rocks can contain 10% or more TOC. Altered igneous rocks are also reasonable candidates as they might have partially altered minerals that might have served as sources of redox energy. On the other hand, martian meteorites arriving on Earth were likely blasted into space from well below the martian surface, where any organics present would have been shielded from the highly oxidizing and radiolytic surface environment. While it is true that near-surface exposure to cosmic radiation likely degrades organic molecules, recent work on Curiosity shows that at least some locations on Mars are eroding fairly rapidly (Farley et al., 2014). Such locations are likely targets for sample collection.

Amino acids constitute the best studied, and among the most interesting, classes of molecules in martian meteorites. Table 6 summarizes existing data on measured amino acid concentrations in seven martian meteorites; some of these data are unpublished, because of the difficulties inherent in distinguishing indigenous martian organics from terrestrial contaminants in these samples. Of these examined meteorites, RBT-04262 appears to be the least contaminated by terrestrial compounds, and thus provides the most reliable indications for what to expect. This sample contained glycine, β-alanine, and γ-aminobutyric acid at levels from 8 to 28 ng/g; other amino acids and nucleobases were below method detection limits in the low nanogram per gram range (reported by Callahan et al., 2013). Other meteorites have generally yielded higher concentrations of amino acids, but again we cannot be certain that these are not terrestrial contaminants. The Antarctic meteorites ALHA 770051 and EETA 79001 have yielded similar concentrations of amino acids. Two caveats to these results are noteworthy: first, none of them is a sedimentary rock, and so might be expected to have lower amino acid contents than sedimentary rocks bearing life (although see the discussion above about surface oxidation and radiolysis); and second, these amino acids are not definitive evidence of life, as a shock-metamorphic origin has also been suggested. Collectively, these data suggest that some amino acids may be present in martian sedimentary rocks at concentrations (from RBT04262 as an example) ranging from a few to a few tens of nanograms per gram, and would be detectable above a background contamination level of <1 ng/g per compound.

OPEN IN VIEWER

Table 6.

Summary of the Average Procedural Blank-Corrected Free (Non-hydrolyzed) and Total (6 M HCl acid Hydrolyzed) Amino Acid Concentrations (in ng/g) in the Hot-Water Extracts of the SNC Meteorites

	Tissint meteorite		Tissint soil		NWA 7034		RBT 04262		ALHA 77005	EETA 79001	ALH 84001	MIL 03346	Antarctic ice
Amino acid	Free	Total	Free	Total	Free	Total	Free	Total	Total	Total	Total	Total	Free	Total
d-Aspartic acid	3±2	48±9	105±69	1648±442	21±5	199±50	<1	<1	2	2	<2	10	0.01E-3	0.01E-3
l-Aspartic acid	27±1	196±32	1331±76	4341±1489	233±150	531±152	<6	<1	<2	<2	<2	27	0.01E-3	0.02E-3
d-Glutamic acid	1±1	59±29	56±5	1501±761	14±3	500±53	<1	<1	2	2	<2	10	<0.01E-3	0.01E-3
l-Glutamic acid	12±1	548±187	630±17	9485±4312	157±26	1,756±503	<2	<1	10	17	<2	84	<0.01E-3	<0.01E-3
d-Serine	17±5	32±5	423±319	1811±610	149±57	183±26	<1	<2	1	2	<2	4	<0.01E-3	<0.01E-3
l-Serine	99±1	273±15	3771±112	5841±1276	871±683	771±373	<19	<2	<2	<2	<2	25	0.02E-3	0.05E-3
Glycine	97±11	432±120	3,311±695	10,524±597	498±410	1,572±500	28±9	22±3	2	74	60	256	0.02E-3	0.03E-3
β-Alanine	7±1	42±16	382±110	1697±482	45±1	305±156	8±1	64±17	11	14	50	16	<0.01E-3	<0.01E-3
γ-Amino-n-butyric acid	8±1	58±8	290±69	1812±680	43±13	109±90	14±2	133±12	51	53	178	18	0.01E-3	<0.01E-3
d-Alanine	8±1	28±10	162±2	1244±444	73±48	168±46	<2	<6	<2	1	<2	3	0.01E-3	0.01E-3
l-Alanine	41±1	167±29	1459±85	6117±1600	347±148	910±224	<7	4±1	<2	<2	8	11	0.01E-3	0.01E-3
d-β-Amino-n-butyric acid	<3	<5	<60	<60	<2	<2	<4	<4	<2	<2	<2	<2	0.01E-3	0.05E-3
l-β-Amino-n-butyric acid	<3	<5	<60	<60	<2	<2	<4	<4	<2	<2	<2	<2	<0.01E-3	<0.01E-3
α-Amino-isobutyric acid	<3	<6	<80	<80	<1	<2	<4	<5	<2	<2	<2	<2	<0.01E-3	<0.01E-3
dl-β-Amino-isobutyric acid	<5	<5	<80	<80	<5	<5	<5	<5	ND	ND	ND	ND	<0.01E-3	<0.01E-3
dl-α-Amino-n-butyric acid	<4	<5	<90	<90	<8	<4	<5	<5	<2	<2	<2	<2	<0.01E-3	<0.01E-3
dl-Isovaline	<4	<8	<95	<95	<2	<3	<5	<8	<2	<2	1	1	<0.01E-3	<0.01E-3
3-Amino-2,2-dimethylpropanoic acid	<1	14	<48	76	<1	11	<1	<1	ND	ND	ND	ND	<0.01E-3	<0.01E-3
5-Aminopentanoic acid	1	7	41	770	17	36	1	4	2	1	14	1	<0.01E-3	<0.01E-3
ɛ-Amino-n-hexanoic acid a	5±4	25±7	151±16	2146±771	246±146	690±527	53±17	1,200±780	244	153	1,439	126	<0.01E-3	<0.01E-3
d-Valine	2±2	10±7	99±19	143±113	16±19	22±20	<1	<1	<2	<2	<2	<2	<0.01E-3	<0.01E-3
l-Valine	13±1	121±16	280±272	3558±1200	230±74	400±150	<2	<1	<2	7	3	23	0.01E-3	<0.01E-3

Data are from Tissint and Moroccan soil (Steele et al., 2014), NWA 7034 (Steele et al., 2013), RBT 04262 (Callahan et al., 2013), ALHA77005, EETA79001, ALH 84001, MIL 03346 (Callahan, unpublished data), and Graves Nunataks Antarctic ice (Burton et al., 2012). All values are reported in parts per billion (ppb) on a bulk sample basis except where noted. Extracts were analyzed by OPA/NAC derivatization and high-performance liquid chromatography separation with UV fluorescence detection and by LC-ToF-MS at NASA GSFC. The uncertainties are based on the standard deviation of the average value of two separate measurements.

a

Also known as ɛ-amino-n-caproic acid (EACA), which is also the hydrolysis product of Nylon-6.

ND, not determined.

Other studies of Martian meteorites have constrained reduced refractory carbon concentration in the low ppm range (Wright et al., 1988, 1992; Grady et al., 1997a,b, 2004; Steele et al., 2012). In the Mars meteorites studied to date, this material is highly refractory and does not yield appreciable levels of extractable molecules. Further work on the Tissint meteorite has shown that -CN, -COOH, CO, CH, and C-C bonding and aliphatic and aromatic carbon are a component of this material, but the concentrations of these functional groups and the specific molecules concerned are not yet fully identified. Benzene, naphthalene, phenanthrene, benzonitrile, and chloromethane, among, other species have been identified in pyrolysis products (Steele et al., 2012, 2013).

Studies by Sephton et al. (2002) and Jull et al. (2000) have also examined the molecular and carbon-isotopic composition of organics released by pyrolysis of martian meteorites. The primary structures released include aromatic rings (benzene, toluene, biphenyl, etc.) plus phenol and benzonitrile, with δ¹³C values similar to those from carbonaceous chondrites (but also terrestrial organic matter). Although terrestrial contamination cannot be rigorously excluded, these authors argue that such compounds likely originate from aromatic, high-molecular-weight organic matter derived from meteoritic input to Mars. The concentrations of organic molecules in the pyrolysates were not quantified, but based on stated detection limits are likely >10 ng/g.

The Panel was not aware of any other reported results for measurements of individual organic compounds from martian meteorites. If we assume that martian organics derive primarily from delivery by carbonaceous meteorites (but note the contrary possibility that the amino acids have shock-metamorphic origins), then we can use the composition of the Murchison meteorite as a rough guide to their expected abundance. This analogy suggests that monocarboxylic acids should be present at roughly an order of magnitude higher abundance than amino acids, aliphatic and aromatic hydrocarbons, alcohols, aldehydes, and ketones at a similar order of magnitude abundance, and amines, purines, pyrimidines, and polypyrroles at 1–2 orders of magnitude lower abundance (Table 7).

OPEN IN VIEWER

Table 7.

Distribution of Carbon in the Murchison CM2 Meteorite

Substance	Abundance
Insoluble carbonaceous phase	1.3–1.8%
Carbonate and CO2	0.1–0.5%
Aliphatic hydrocarbons	12–35 ppm
Aromatic hydrocarbons	15–28 ppm
Monocarboxylic acids (C2–C8)	∼170 ppm
Hydroxy acids (C2–C5)	∼6 ppm
Amino acids	10–20 ppm
Alcohols (C1–C4)	∼6 ppm
Aldehydes (C2–C4)	∼6 ppm
Ketones (C3–C5)	∼10 ppm
Ureas	∼20 ppm
Amines (C1–C4)	∼2 ppm
Pyridines and quinolones	0.04–0.4 ppm
Pyrimidines	∼0.05 ppm
Purines	∼1 ppm
Polypyrroles	< <1 ppm
Sum	1.43–2.35%
Total carbon	2.0–2.5%

Data are from Wood and Chang (1985) and Cronin and Chang (1993).

In situ measurements of martian organics are available from the Viking, Phoenix, and Curiosity missions. Both Viking and Phoenix studied loose sediment scooped from the martian regolith (expected to be highly oxidized), and reported no volatile organic compounds above detection limits of <1–10 ng/g (Biemann et al., 1977). The SAM instrument on Curiosity has not yet completed its work or formally reported all the results in hand; however, recent abstracts and papers in preparation describe the “tentative” detection of dichloropropane and chlorobenzene at levels of a few tens of nanograms per gram in a core drilled into a mudstone at Yellowknife Bay (e.g. Summons et al., 2014a,b). A significant caveat here is that all three missions used thermolysis and pyrolysis to volatilize organic compounds so that they are amenable to gas chromatographic and mass spectrometric analysis. A series of studies (Navarro-González et al., 2006, 2010, 2011; Navarro-Gonzalez and McKay, 2011) suggested that the presence of perchlorate and other oxidants in the martian regolith renders the Viking results difficult to interpret because of the high probability of either oxidizing or chlorinating indigenous organic molecules during heating. These observations have been contested (Biemann, 2007). As a result of these complications, the Panel did not rely on the earlier non-detections of organics in Mars regolith. The more recent detections of chlorobenzene (or possibly its aromatic precursors) and other chlorohydrocarbons by MSL (e.g., Freissinet et al., 2014) are considered to provide likely lower limits for these compounds.

The OCP briefly considered organic concentrations in analog terrestrial rocks as another constraint on what to expect on Mars. Fine-grained sedimentary rocks that have not been oxidized or weathered constitute the type of sample that we might hope to return from Mars. On Earth, similar rocks commonly contain >0.1% (100 ppm) TOC, and contain many individual biomarkers at levels >1 ppm. Even those sediments considered to be relatively poor in organics contain >0.01% (10 ppm) TOC (Mayer, 1994), and yield individual biomarkers at levels >10 ng/g (e.g., Lipp et al., 2008). Soils from the Atacama Desert are reported to have 32 ppm TOC (Navarro-González et al., 2010), an order of magnitude below typical “organic poor” marine sediments. Subcritical water extraction of subsurface Atacama soils (Jungay region) followed by derivatization and capillary electrophoresis of the fluorescently labeled amines has demonstrated individual amines and amino acids at the 50–100 ng/g level (Skelley et al., 2007). The earlier MSR SSG II panel considered a more extreme example of organic-poor sediments (i.e., a highly oxidized sedimentary rock that had been buried and undergone thermal maturation). Although such rocks have similar levels of TOC (approximately 0.01%), diagenesis has rendered most biomarkers into macromolecular kerogen, which is not extractable. They thus estimated expected concentrations for hydrocarbon biomarkers of 0.1–1.0 ng/g in such a rock (see Section 2.4 for details of this calculation).

There are several limitations to using such terrestrial analogs to predict concentrations of martian organics. First and foremost, there is no a priori reason to expect that concentrations of organics in terrestrial rocks would be indicative of those on Mars. Indeed, valid arguments can be made for their being either higher or lower than indigenous martian concentrations. For example, with lower organic input and more oxidizing subsurface conditions, martian rocks might have lower organic concentrations. If microbial activity were more limited (or absent) on Mars, residual organic concentrations might be higher. Second, it is unclear which terrestrial rocks, sediments, or soils we should choose as appropriate analogs. Even considering that oxidized rocks represent a reasonable lower bound for those found on Mars, the goal of the mission is clearly not to sample and return the most oxidized martian rocks. Indeed, it is unclear whether scientific goals could be met with such a rock even given zero organic contamination. The third concern expressed by the OCP is that hydrocarbons (the dominant biomarkers in thermally mature terrestrial rocks, and those considered primarily by the MSR SSG II report) may not be the class of organics that are most abundant or interesting on Mars. With no active tectonics to deeply bury sediments under reducing conditions, biomolecules (or even meteoritic organic compounds) might be transformed to more oxidized species rather than more reduced ones. In summary, consideration of terrestrial analog rocks indicates that organic concentrations in martian rocks might span a huge range around those directly measured in meteorites. We therefore conclude that this line of argument provides little firm footing on which to construct quantitative limits.

A final constraint on expected concentrations in the absence of martian biota, previously considered by the OCSSG report and by Benner et al. (2000), can be derived from estimated rates of delivery of organic carbon to the surface of Mars by meteorites. Meteorites deposit an estimated 2.4×10⁸ g/year of organic carbon to the martian surface (Flynn, 1996). If allowed to accumulate over 3 billion years, and given a Martian surface area of 3.6×10¹³ m², this would result in 20 kg/m² of organic carbon. Assuming a mixing depth of 1 m and rock density of 4 g/mL results in a predicted TOC concentration of 5 mg/g. A much more conservative mixing depth of 100 m would lower this to 50 μg/g. If we presume that this organic carbon has a molecular makeup similar to that of Murchison (see Table 7), where functional classes of molecules represent approximately 0.05% of TOC, we predict approximately 2.5 μg/g of each class of organics (1 m mixing depth). Further assuming that each class comprises 10–100 compounds, this yields a final prediction of approximately 20–200 ng/g per compound (or 0.2–2 ng/g for the 100-m mixing depth). These estimates span the measured concentrations described above (calculations presented here follow those in Benner et al., 2000).

A significant complication is that much of this meteoritic carbon is likely to be oxidized on the martian surface. Benner et al. (2000) considered that many of the likely oxidation products are organic molecules that would be metastable under martian conditions (e.g., benzenecarboxylic acids). Working from the assumption that meteoritic TOC would be converted to such molecules with an efficiency of approximately 10%, they predicted (using the same delivery rate and mixing depths as above) an accumulation of 5–500 μg/g of benzenecarboxylic acids. The concentrations of individual species were not predicted, but again assuming 10–100 major compounds would yield a range of estimated concentrations of 50 ng/g–50 μg/g per compound. These values are higher than those estimated for individual molecules in the preceding paragraph due to the assumption that meteoritic TOC (as opposed to just extractable carbon) is converted into measurable molecules. Moreover, laboratory analog experiments show that heating functionalized aromatics such as benzenecarboxylic acids in the presence of perchlorate generates chlorobenzene as identified by Curiosity (Miller et al., 2013).

4.2.2.2. What concentrations can we measure?

Instrument detection limits provide a useful lower bound for setting organic contamination requirements in two ways. First, it is not feasible to set requirements that are below the detection limit of any analytical method, because there would be no way of verifying that requirements have been met. Second, there would be little practical benefit in protecting analyte concentrations that are themselves too low to measure, although here one needs to account for the inevitable improvements in sensitivity that would exist by the time samples are returned. Regardless, it is not necessarily true that contamination limits need to be lower than detection limits to protect all measurements (see Section 4.1.2). Another key aspect is that the ability to characterize contamination levels (e.g., their reproducibility and chemical nature) might improve in the future.

As described in Section 3, there is a huge diversity of analytical techniques that could be brought to bear on returned martian samples, each differing in their sensitivity towards analytes (limits of detection) and their selectivity (ability to measure a wide range of compounds). Based on the general trade-off between selectivity and sensitivity, we argue that it is not practical to consider the detection limits of all possible analytical techniques, and instead focus on those with the broadest analytical window (least selective) that would be used for initial assessment of the types and concentrations of organics present in samples (survey measurements). As described above, detection limits for the most sensitive of these methods are typically in the range of 0.1–10 ng/g. We thus adopt 1 ng/g as a representative value.

A limit of 1 ng/g for individual organic compounds of primary concern appears to strike a good compromise between protecting the survey measurements that would undoubtedly be conducted on returned samples, while still being detectable by more targeted measurements for validation of spacecraft contamination levels. The primary perceived downside to this strategy is that a 1 ng/g limit would not fully protect even the most sensitive techniques existing today, let alone those that might exist 20 years in the future. Clearly, lower contamination limits would not be “wasted” in the sense that they would enable even more sensitive measurements to be made. If lower contamination levels (for example, down to 0.01 ng/g) can be achieved at reasonable expense, they would be scientifically valuable. On the other hand, the few concrete measurements of martian samples that do exist (described above) do not indicate that such stringent limits would be absolutely required to meet mission scientific objectives.

4.2.2.3. What level of cleanliness can we achieve?

The third criterion considered by the Panel is what levels of background contamination can be achieved using current technology. At the level of individual compounds, this is a difficult issue to grapple with for a number of reasons. First, contamination varies widely in space and time, as a function of analytical conditions, target analytes, sample matrices, etc. Compiling estimates for every compound of interest, under even a few sets of conditions and samples, is prohibitive. Second, eliminating contamination is difficult and time-consuming, and so typically is driven only to the levels that are needed, rather than those that are technologically possible. Sub-nanogram per gram levels of organic analysis are uncommon in terrestrial geologic samples, so similar levels of cleanliness are seldom needed. Third, quantitative background concentrations are not often reported, rather simply that they are below the levels of detected analytes. Fourth, methods of reducing contamination are often specific to the particular analytes of interest, and so it is unclear whether it is possible to simultaneously achieve reported limits for all analytes of interest.

With these caveats, we can use the minimum levels of detection reported for various analytes in geologic samples as a maximum limit for the blank concentrations that must have been achieved by those studies. Amino acids have been measured in martian meteorites at levels down to 1 ng/g (Callahan et al., 2013), fatty acids in deep-subsurface rocks have been measured down to 10 pg/g (Onstott et al., 2003), DNA has been measured down to 10 fg/g by fluorescence detection (Zhao et al., 2003), and PAHs have been measured down to 10 fg/mL by laser-induced fluorescence (Yan et al., 1995). It thus appears likely that most, if not all, measurable organic compounds can readily be reduced to at least <1 ng/g, and in many cases to orders of magnitude below that. We thus foresee no significant difficulties in achieving the levels of cleanliness indicated in preceding sections, although it is conceivable that achieving such limits simultaneously for all Tier-I compounds may be more difficult.

4.3. Considerations related to TOC

A limit for TOC contamination can serve two purposes. First, it serves to limit background levels that must underlie measurements of TOC (whether of concentration, elemental or isotopic composition, molecular weight, etc.) in returned samples. Second, it serves as a simple way to effectively limit contamination at the molecular level by all possible compounds, without the analytical effort required to quantify them individually. Given that TOC is almost universally more abundant than individual compounds by orders of magnitude, the second consideration serves as the more stringent constraint. The OCP therefore focused primarily on setting TOC limits that would achieve effective protection of measurements at the molecular level. Previous missions seem to have adopted a similar philosophy in setting TOC limits, particularly given that levels for individual molecules were rarely established. In this context it is important to remember that our proposed TOC limit is not an end unto itself, but rather must be consistent with the molecular limits proposed above. Relaxing requirements for individual molecules would likely lead us to recommend more stringent limits on TOC.

4.3.1. Allowable levels of contamination

As for individual contaminants, the Panel considered three lines of evidence in choosing appropriate limits. The discussion below deals exclusively with data for TOC, primarily the background levels required for protecting TOC measurements. We return to the question of protecting measurements of individual molecules in Section 4.4.

4.3.1.1. What analyte concentrations do we expect?

The most recent and complete study of TOC in martian meteorites reported levels ranging from 4 to 26 μg/g (ppm) in all seven samples subjected to this analysis (Steele et al., 2012). To avoid analyzing terrestrial contaminants, the authors first heated samples to 600°C to drive off volatile and semi-volatile compounds. The reported concentrations thus represent a minimum limit for the true total organic load of the samples prior to landing on Earth. Further analysis by Raman spectroscopy and laser-desorption mass spectrometry indicated discrete blebs of macromolecular carbon trapped within mineral grains, consistent with an indigenous martian origin. Analyses of ¹³C and ¹⁴C in the samples provided further support for a dominantly martian origin. Similar, albeit non-spatially resolved, data have previously been obtained by Grady et al. (2004). Given that these samples are all martian basalts, it appears quite likely that most unaltered martian rocks contain a background level of at least 1 μg/g (ppm) of TOC. Whether or not sedimentary rocks and oxidized regolith contain more or less TOC is subject to assumptions about the relative importance of biotic and meteoritic inputs versus oxidative loss (see above), and appears unknowable at the current time.

Finding #14:

Although TOC from Martian igneous minerals can reach 20 μg/g (ppm), and are >4 μg/g in all samples measured to date, the concentration in non-igneous samples is currently unknown, and may be much lower.

4.3.1.2. What concentrations can we measure?

TOC is an operationally defined fraction subject to all the ambiguities that accompany the definition of “organic carbon.” Consequently, analytical methods for measuring TOC concentration (Table 8) do not all report the same value from the same sample. For example, methods based on oxidative combustion at temperatures >800°C include all carbon that can be oxidized to CO₂. Such methods commonly have detection limits near 1 μg/g without concentrating samples, although this appears to be limited more by terrestrial blanks than by instrument sensitivity. Some methods based on vibrational spectroscopy (FTIR, Raman, etc.) are typically more sensitive (down to 1 ng/g levels), but may not detect all carbon species; for example, FTIR cannot detect graphitic or amorphous macromolecular carbon. On the other hand, surfaces can be mapped using Raman spectroscopy to survey for minute concentrations of these species. However, these techniques require delicate surface preparation and are susceptible to topographic and matrix effects. With so many analytical dependencies, they are generally not quantitative.

OPEN IN VIEWER

Table 8.

Compilation of Analytical Methods, and Representative Sensitivity, Potentially Suitable for Measuring TOC in Samples and/or on Spacecraft Surfaces

*

It should be noted that all methods require specialized hardware sampling and/or witness plates.

**

Solvent extracts may use a surface rinse or specialized solvent swabs of hardware surfaces.

GATR, grazing angle attenuated total reflection (IR); MW, molecular weight.

Mass spectrometric methods for analyzing TOC in solvent extracts are equally sensitive, but have more restrictive analytical windows and would likely not include macromolecular carbon. Other techniques, such as secondary ionization mass spectrometry (SIMS) and x-ray photoelectron spectroscopy (XPS) are highly sensitive, but are more complex and expensive to implement. A second complication is that some TOC methods are amenable only to analyses of bulk sample materials (oxidative combustion), whereas others are amenable only to analyses of surfaces (SIMS, FTIR, Raman, XPS). Some sample-based measurements can be applied to spacecraft surfaces by means of solvent rinses or swabs.

Beyond simple measurements of TOC concentration, it is likely that scientists would be interested in assaying the elemental and isotopic composition, average molecular weight, aromaticity, optical activity, ion-exchange capacity, surface-area loading, and other characteristics of total organic matter in returned samples. Such measurements all use specialized analytical techniques, only some of which overlap with those described in Section 3. Nevertheless, it is the Panel's belief that all such measurements would be less sensitive than those used to measure simple TOC concentrations, and so they are not considered further here.

The OCP discussed these analytical issues in detail, but ultimately decided that it was beyond our purview to provide a detailed intercomparison of methods, or to propose specific methods for monitoring TOC. For the purpose of setting minimum contamination limits, we point out that several surface-based techniques can routinely achieve detection limits of <1 ng/cm², and so could be employed to verify contamination limits on spacecraft surfaces down to this level. Translating this into a bulk contamination level for samples requires consideration of both transfer efficiency and sample-contact surface (see below), but leads to plausible levels of 2–20 ng/g TOC. To our knowledge, no bulk TOC analyses of terrestrial geologic samples have yet reported detection limits this low. On the other hand, no terrestrial geologic materials have such low levels of TOC, so these detection limits have presumably not been needed. It is our belief that if a martian sample were returned that contained only 2–20 ng/g TOC, a suitable analytical method for quantifying this level of carbon could be readily developed.

4.3.1.3. What level of cleanliness can we achieve?

There was a very substantial debate among OCP panelists as to what are the lowest levels of TOC contamination that are readily achievable. In considering this issue, a key concept is to distinguish between what is possible on spacecraft surfaces, and what is achievable in returned samples. The former can be readily reduced to arguments about nanograms per unit area, but the latter requires additional understanding of contaminant contributions from the entire sampling train, and potentially the entire landed spacecraft. Hardware geometry and mode of operation, as well as surface cleanliness, become relevant. We consider the issue of cleaning surfaces first and that of the integrated sample contamination second.

Methods for cleaning spacecraft surfaces vary widely, and with highly variable results. As an optimal solution, the Panel agreed that extended (hours) heating to >500°C in an oxidizing atmosphere (commonly air) typically renders a surface free of organic carbon at or below picograms per gram (ppt) levels. This procedure is commonly used in most organic geochemical laboratories today. The question then becomes how rapidly the surface becomes recontaminated with organic carbon. Earth's atmosphere contains an appreciable amount of organic carbon, both in the form of volatile organic compounds and as organic particulates. The latter are readily filtered out of clean-room air, but the former are harder to control to extremely low levels. Thus a bare metal surface exposed to even “clean” air will quickly (within minutes/hours) acquire a layer of AC, typically approximately 20–100 ng/cm². The rate of recontamination is highly dependent upon the type of organic carbon present, their partial pressures within the environment, and the temperature differential between the environment and the spacecraft surfaces. Deposition rates differing by orders of magnitude can be achieved by covering cleaned parts with a clean impermeable wrapping such as aluminum foil. The phenomenon of accretion of contaminants is well documented in the literature (e.g., Siegbahn et al., 1967; Swift, 1982; Barr and Seal, 1995; Piao and McIntyre, 2002). Studies indicate that this layer asymptotically approaches a stable, approximately monolayer equivalent film (Fig. 11), with the equilibrium TOC content of this film depending both on surface composition and on volatile organic concentrations, among other factors. Typical mean deposition rates for clean-room environments are reputed to be near 0.15 ng/cm²/hr (0.1 mg/ft²/month), suggesting that a stable AC film typically requires 6–9 days to develop. Limiting contact of a metal surface to air shortly after baking should thus provide a straightforward and efficient way of achieving very low (ppm) TOC levels. Although this can be done with minimal effort on Earth, “unwrapping” the protected parts for use once on Mars would presumably introduce additional engineering risks. Although lower levels of organic contamination could be achieved by combusting (at >500°C in an oxidizing atmosphere) and then isolating the entire sampling hardware chain, hermetically sealing all elements of the sample chain has serious implications for mission failure modes (and opportunities to compromise science).

OPEN IN VIEWER

FIG. 11.

JPL test data suggest an asymptotic time dependence for the formation of an AC film on precleaned metal surfaces. Rate and amount of formation are highly variable, and depending on precursor concentrations, substrate, configurations, and environmental conditions. The quantification of the y-axis is TOC. Std, standard.

4.3.1.3.1. Contamination pathways

In the case of MSL, the vectors for the transfer of Earth-sourced organic contamination to rock/soil samples have been interpreted by the MSL project engineers to include the following (Fig. 12):OPEN IN VIEWER

FIG. 12.

Graphic representation of three important pathways for contamination transport from spacecraft to samples. VOC, volatile organic compound.

1. Direct contact: Direct contact of Mars sample material with sampling hardware is thought to be the largest contributor to sample C contamination. Based on contamination modeling, direct transfer from spacecraft contact surfaces has been interpreted to be quantitatively the most significant component for both MSL and OSIRIS-REx (Harstad and Bellan, 2006; Blakkolb et al., 2008; ten Kate et al., 2008; Anderson et al., 2012; Dworkin, unpublished data, 2014).

2. Particle transport: Dislodgment of particulate (potentially microbe-laden) contamination from the exterior of the Rover by saltation is not believed to be a significant source of TOC in samples. Beaudet (2000) calculated particle dislodgement rates, not taking into account particle adhesion. Harstad and Bellan (2006) performed calculations that included adhesion forces and concluded the probability of particle removal “…is estimated as corresponding to an adhesion half-life of O (10⁴) years, and is thus not important.”

3. Outgassing from rover hardware: Engineering contamination transport calculations [using MSL transport models and source rates (Blakkolb et al., 2008; ten Kate et al., 2008)] show that outgassing from Rover hardware contributes less than approximately 1 ng/g TOC during the sample acquisition process (Blakkolb et al., 2008).

A comparable analysis for Mars 2020 has not yet been carried out. For the purpose of this report, we confine our analysis to that of direct contaminant transfer from hardware surfaces to samples, while noting that other transport vectors need to be rigorously evaluated by the Mars 2020 Engineering Team using deterministic and probabilistic methods (see, for example, Hudson et al., 2010). This reduces the problem to two variables: the hardware surface area contacting samples, and the efficiency with which the contamination is transferred.

A comparable analysis for Mars 2020 has not yet been carried out. For the purpose of this report, we confine our analysis to that of direct contaminant transfer from hardware surfaces to samples, while noting that other transport vectors need to be rigorously evaluated by the Mars 2020 Engineering Team. This reduces the problem to two variables: the hardware surface area contacting samples, and the efficiency with which the contamination is transferred.

However, changes in payload, sampling philosophy, and cache placement dictate that the Mars 2020 Project Team will need to undertake contamination transport analyses and models of transport modes that are specific to the Mars 2020 system recontamination profile. Factors influencing the induced contamination environment include, but are not limited to, the relative configuration of the cache and the instrument payload, especially potential outgassing sources from the instruments that have heaters on them, and the fact that the dilution cleaning strategy cannot be applied to Mars 2020.

Because the baseline configuration of the sample contact surfaces for Mars 2020 is very different from those of MSL, the dilution cleaning process appears not to be available to Mars 2020. However, it is important to note that in comparison with Mars 2020, MSL has far higher sample contact surface area, and far lower sample masses, so the effects of surface contamination are greatly magnified.

Finding #15:

In order to achieve contamination levels for sample contact surfaces lower than 20 ng/cm², a more effective strategy for avoiding recontamination after initial cleaning than that used by MSL would need to be implemented.

4.3.1.4. Contamination transfer from sample-contacting surfaces

Hardware surfaces relevant to cached samples comprise two main components: the sample (cache) container itself, and the sample collection (drilling) and transfer apparatus. Sample containers are not intended to be re-used, so we need consider only their surface area. In contrast, current design ideas for drilling apparatus include both re-usable and single-use designs. Re-useable drill strings have the advantage of being able to undergo “dilution cleaning” [i.e., dilution of terrestrial contaminants by repeated processing of martian samples (Anderson et al., 2012)]. For our analysis, we considered only the most conservative case of single-use drill strings (i.e., no dilution of contaminants). Several hardware designs representing a range of sample-contacting surface areas are currently under consideration by the Mars 2020 Project Team. To constrain the problem, we use 300 and 30 cm² as the respective upper and lower limits on the area of sample-contacting surfaces.

Although it is likely that most of the sample-contacting surfaces would be subject to only moderate or slight abrasion, for the purposes of our analysis we consider two cases: a non-realistic but conservative bounding case, in which 100% of the contaminants on sample contact surfaces are assumed to transfer to the samples, and a “mean expected” case, where the available experimentally determined transfer coefficients are used. If further experiments by the project on actual sampling hardware can demonstrate lower transfer efficiencies, then lowering this assumption would be reasonable.

We also assume the sample would contact the bit (a strong abrasion environment) and its sample tube (a “slight abrasion” environment), and that 25% of the sample contact surface area is bit, and 75% is tube. Such a model would imply a mean transfer efficiency of approximately 20%. This figure would drop if the bit were re-used for multiple samples due to dilution of contaminants (because a portion of the organic contaminants initially present on the bit would end up in the first sample, and be unavailable to be transferred to the second sample).

Figure 13 presents the expected level of bulk contamination in a sample (y-axis) given an assumed hardware surface area, transfer efficiency, and level of cleanliness (x-axis). For an assumed level of surface contamination of 20 ng/cm² (see above), a sample contact surface area of 300 cm² and 100% transfer efficiency results in >>100 ng/g TOC in the sample. In contrast, given the same assumptions but with only a 30 cm² contact area, sample contamination of approximately 40 ng/g is predicted. If a 60% transfer efficiency were assumed, that level would fall to only 8 ng/g. We thus believe that TOC contamination levels in the sample of 20–40 ng/g are achievable even when sampling surfaces are exposed to air. We note that such levels would place significant constraints on the viability of some hardware designs with large surface areas: \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} ppb = \frac {\hbox{hardware cleanliness level}{(\rm ng/\rm cm^{2})} \times \hbox{surface area of the sample that contacts hardware} \ ({\rm cm}^{2}) \times \hbox{transfer efficiency}} {\hbox{mass of sample core (g )}}\end{align*} \end{document}

OPEN IN VIEWER

FIG. 13.

Translation from sample-contacting surface values to sample values, assuming sample mass=16 g. Contaminant contact transfer efficiency is dependent on sample-hardware configuration; a range of 10–100% is presented to illustrate the proportional dependence of in-sample contamination to this parameter.

Finding #16:

In the case of a system with sample contact surfaces of 30 cm², and contaminated with 20 ng/cm² organic carbon, direct transfer could result in a theoretical maximum of 40 ppb organic contaminants on collected samples assuming 100% transfer efficiency. Actual concentrations in samples could be either higher or lower than this, depending on actual transfer efficiencies and the importance of other contaminant transfer pathways.

4.4. Considerations related to particulate organic matter

4.5. Implementation

5.1. Introduction

The OCP recognizes four broad strategies for recognizing and distinguishing organic contamination. Some of these strategies are also beneficial in the characterization of contamination:

1. Use of witness plates

5.2. Witness plates

Witness plates would be critically important to the integrity of the science if samples are returned. Witness plates are objects, such as metal or ceramic plates, that are positioned in a way that allows them to collect the same chemical contamination or debris as an object of interest. When exposed sequentially during the course of a mission, witness plates can establish a record of the history of contamination events. The witness plates can be analyzed to identify the type and quantity of contamination at various times and three-dimensional positions. In the case of the MSR mission, they could be used to provide information about both the flight and the ground environment, including operations on the ground during ATLO of the Mars 2020 rover, the flight environment, the recovery environment at the Earth landing site, the containment environment, and curation. This information would be a key input into distinguishing native compounds from introduced contaminants. Experience with the managing of other sample collections suggests that allocation requests for its witness plates are as common as requests for the samples themselves (for example, as of this writing there have been 600 allocations of Genesis solar wind samples and 300 allocations of the accompanying flight-like reference collectors).

Since samples and sampling hardware have different physical properties with respect to different contaminants, it is typically best to employ multiple (at least two) different types of witness plate material (Sandford et al., 2010). This allows the witness plates to account for different adsorption and absorption properties among contaminants. It may be desirable to place one or more witness plates inside the sample tubes. In addition, some effort should be put into understanding witness plate materials and sizes, in order to optimize eventual analysis. Time-of-flight (TOF)–SIMS (static SIMS with “in situ” virtually non-destructive surface analysis of both organic compounds and inorganics) could be very useful for sample analysis (however, there are questions given the possible low organic content in the samples). However, it would be simply invaluable for in situ contamination characterization on hardware, witness plates, cache tubes, and other pieces of hardware.

While it is desirable for the witness plates to maximize surface area and mimic the chemistry of the sample, it is more important that the witness plates (1) do not endanger the mission, (2) do not contaminate the sample, (3) are amenable to a variety of analyses, and (4) can be divided between different labs without further contaminating them. For example, while zeolite is more martian-like than sapphire windows and is easier to divide, it may be harder to contain, and particles could get into the sample, or even endanger spacecraft mechanisms. A larger number of smaller witness plates may be preferable in the sense that not only can a larger number of witness plates be used with multiple instruments, they can provide a more statistically robust sampling (or a combination of both factors). Experience with previous missions (e.g., Sandford et al., 2010) shows that witness plates cannot be easily divided after the mission without risking further contamination. Witness plates should be deployed with a geometry that provides a molecular “view” of the sample, or hardware they are representing, and thus, they should be located as close to the actual sample, or hardware as possible. Also, consideration should be given to the possibility that cache hardware components might suffice as supplemental witness plates. Other important considerations include dimensions of the plates, mounting, transport to analytical instrumentation, and the development of procedural controls that ensure confidence in the analysis. These multiple trade-offs should be examined by the project team.

The flux of contamination onto the Mars 2020 spacecraft and rover would not be constant, and accrued contamination may change through exposure to different environments. Also, not all contaminants are retained by the samples equally over time due to chemical variation and volatility. Thus, the more temporal/spatial information gleaned from control samples, the more precisely that information can be applied to understanding the samples. This temporal variability in contamination leads to the need for phased witness plates, and the sequence of witness plate employment can be roughly divided by mission phase. The detailed employment strategy of witness plates depends on the number of witness plates required to record the various phases of the mission, how many time periods should be sampled during the mission, the witness plate geometry and composition, and, ultimately, how many witness plates would be needed for post-flight analysis. We recommend that a witness plate campaign should be explicitly designed and implemented as part of contamination control for the mission. For Mars 2020, key time periods for sequential witness plate employment include (see Fig. 14):

OPEN IN VIEWER

FIG. 14.

An example of a witness plate deployment plan to provide information on the nature and timing of contamination events/processes. Note that for MSR this plan would need to be established early, since some of the witness plates that would be valuable later need to be deployed early, and make the round trip to Mars. SRV, sample recovery vehicle.

Beyond the considerations for the Mars 2020 sampling rover, witness plates should be used to record the contamination environment of individual samples in the SRF, and those witness plates should be periodically analyzed to maintain a contamination knowledge record for individual samples and sample subsets. One key item with respect to witness plates is the differentiation between witness samples collected (during build and up to launch) for future analysis and those collected that require analysis to validate cleanliness levels immediately so any mitigation steps can be performed. Witness plates integrated with the assembly of the sample retrieval and storage system would be the most useful. A subset of these plates can be pulled off and archived prior to launch and analyzed when the samples are returned and compared with the plates that made the trip. The idea of trying to put some of these archived plates through the same environmental changes the Mars plates are experiencing would give a good view of the changes various molecules go through during the trip.

An example of the structure of a witness plate plan is shown in Fig. 14. A key point is that this plan needs to be established early in the MSR mission, and then implemented systematically throughout—some of the exposure periods begin in the mission development phase. Representative witness plates should be archived for eventual analysis in parallel with returned martian samples. However, it would also be useful to analyze the assembly and test witness plates on short time scales to mitigate contamination events, as part of ATLO contamination control/knowledge. If direct sampling of contamination on flight system surfaces is prohibited owing to concern about recontamination or geometry, then witness plates of identical material (e.g., 6061 aluminum) can be processed in parallel with the same cleaning and testing procedures to serve as a contamination control proxy.

5.3. Blanks and blank standards

In contrast to witness plates, which provide an accumulation of contaminant particles and films during a defined period of time, a blank serves as a measurement of non–sample-related inputs to an instrument signal. Blank standards may turn out to be the most sought-after material collected on Mars. Without them, measurements of organic compounds in all other returned samples would be unconstrained and therefore suspect. In the case of the possible returned martian samples, a blank can be envisioned in one of two ways: (1) A system (collection, analysis, other) is operated as it would if a sample were present, except that no sample is inserted. This would be a crucial step for the instruments used to analyze the possible returned martian samples. For the sampling system of the Mars 2020 rover, this could mean that a sample tube is opened and then closed without inserting a sample. (2) A synthetic sample known to have zero concentration of the analyte of interest is inserted. Any instrument response to the analysis of such a sample can be interpreted as contamination.

As an example, Phoenix flew an organic-free ceramic blank that was to be used to characterize the cleanliness of the sampling system by using the TEGA to detect organic molecules (Ming et al., 2008). MSL was prepared to use this strategy to control its sampling and analysis operations by means of what is referred to as the organic check material (OCM) (Conrad et al., 2012), although as a practical matter, the OCM has not been sampled on Mars as of this writing (Fig. 15). The OCM is a block of ceramic (a non-Mars material) that was fired at high temperature to drive off all organic molecules. (The OCM was then doped with a single non-Mars organic molecule, so that it can serve both as a positive and a negative control standard.)

OPEN IN VIEWER

FIG. 15.

NASA's Mars rover Curiosity carries five cylindrical blocks of OCM for use in a control experiment if the rover's SAM laboratory detects any organic compounds in samples of Martian soil or powdered rock. The blocks are carried on the front of the rover, within reach of the sample-collecting drill on the rover's arm, and are sealed under foil until needed. This image centered on the foil that covers one of the bricks was taken by the rover's Mars Hand Lens Imager during Sol 34 of Curiosity's work on Mars (September 9, 2012).

As shown in Fig. 16, there are several potential strategies for inserting blank standards into the sequence of unknown martian samples. The decisions regarding how many, and when blank standards should be added, are left to the Mars 2020 Science Team. Regardless of the actual strategy, it is important that the blank standard be collected periodically as the sample (e.g., the geological unit examined) and/or the sampler (e.g., the drill bit) changes (Fig. 16). It is desirable that sampled cores of the negative control standard bracket (in time) the collection of actual samples, with the allowable caveat that a sample collection system failure could interrupt collection of the final negative control. The key consideration is that in order for the potential detection of organic molecule that may imply past life on Mars, to have maximum credibility, such samples should be bracketed by standards.

OPEN IN VIEWER

FIG. 16.

In order to bracket all unknown martian samples with blank standards, the first sample taken should probably be a blank standard. The use of blank standards in the rest of the sample series depends on the decision of the Mars 2020 Science Team. Some suggestions are offered here.

Past studies have assumed that in a 31-slot cache, at least three slots would be reserved for blanks (see, for example, McLennan et al., 2012; Mustard et al., 2013). However, as shown in Fig. 16, it is not hard to envision scenarios in which this number appears to be too low, and it would seem more prudent to have the capability to be able to collect at least five blank standards, and possibly six. How this capability is actually used during the mission would be up to the Mars 2020 Science Team, and would presumably depend on what is encountered by the mission during its surface operations phase.

Note that positive control standards are also typically a critically important part of the sample analysis process. However, their value lies in validating the reality or quantification of a positive detection, not in interpreting contamination, so they are not discussed further here (positive controls are typically synthetic materials that contain a known concentration of an analyte of interest).

5.4. Archive of organic and trace biological materials

The storage of archived materials is of great importance, and some careful thought needs to go into that aspect (since we need to be able to go back to these materials after the cache is returned and be reasonably certain that these materials and the contamination level recorded in them has not changed significantly over the decade or more between launch of Mars 2020 and the return of the cache). Materials storage would support legacy analyses (Figs. 17 and 18). To support analyses, storage contamination effects must be clearly understood.

OPEN IN VIEWER

FIG. 17.

Examples of MSL material samples, many of which are organic-bearing, stored in the JPL Planetary Protection Archive: (a) heat shield, (b) heat shield in ESD packaging, (c) heat shield panel, (d) backshell, and (e) harness. (f) The samples are stored in a cabinet under well-controlled temperature conditions. Source: Melissa Jones, JPL Biotechnology and Planetary Protection Group.

OPEN IN VIEWER

FIG. 18.

Samples of organic matter on outbound spacecraft, including swab samples of trace microbial contamination, organic-bearing materials used in the spacecraft build, and witness plates, should be stored for future reference. Some portion of these samples presumably would need to be analyzed shortly after collection, but the main value would be analysis with the potential future martian samples and blank standards. (Modified after Mustard et al., 2013).

Appropriate witness plates, blanks, and flight-like reference materials are important and require adequate documentation and preservation. To correctly interpret round-trip contamination in returned samples, a facility and systematic approach for storing several kinds of materials and samples associated with the Mars 2020 rover would need to be established. All of the samples described below need to be preserved, monitored, and made available to analysts in a suitably clean, equipped, and staffed curation facility. In the case of organic analytes, particular attention needs to be given to the conditions of storage in the archive, requiring a well-documented systematic approach. Stored genomic samples would allow for capability to compare against returned samples to avoid false positives. The OCP is aware of two relevant capabilities already existing within the NASA system: (1) the facilities managed by the Astromaterials Acquisition and Curation Office at Johnson Space Center, and (2) the Planetary Protection Archive, managed by JPL Biotechnology and Planetary Protection Group. The latter archive currently houses approximately 200 organic material samples, approximately 100 flight parts/components (see, e.g., Fig. 17), approximately 200 nucleic acid samples from MSL (Venkateswaran et al., 2012; La Duc et al., 2014), and approximately 3500 microbial isolates (Schubert et al., 2003; Schubert and Benardini, 2013, 2014).

Within the archive facility (and on the flight mission as well), attention to packaging is extremely important. Avoid plastic bags as primary containers for cleaned hardware and returned samples. Packaging of cleaned parts is often done by heat-sealing inside plastic bags. The plastic films liberate plasticizers just by their presence. Even more plasticizers are given off when heat sealing. Of particular concern to this mission is outgassing of caprolactam by nylon bags, as there is some indication caprolactam reacts with water to form amino acids.

Finally, experience from prior missions (e.g., Genesis) has shown that nearly every handling step (measurement technique or cleaning step) can add contamination—this needs to be planned for.

5.5. Spatially resolved measurements on returned samples

Spatial information is one of the most powerful means of determining whether organics are contaminants or indigenous to the sample (e.g., Allwood et al., 2009; Steele et al., 2012, 2013). Spatial information is also critically important for determining the origin of indigenous organics [biotic, abiotic, etc. (e.g., Schopf, 1993, 2006; Ueno et al., 2001; Sugitani et al., 2007, 2010; Allwood et al., 2009, 2013; Steele et al., 2012)]. For example, if organics were distributed across only the exterior surface of a core, the probability that they are contaminants would be high. Similarly, if the organics reside in cracks that are connected to the exterior environment, it is possible that the organics are contaminants. As an illustrative example, Figure 19 shows evidence of terrestrial microbes growing in recent cracks within the martian meteorite NWA 7034. An indigenous origin for organic carbon would be indicated if the carbon is embedded in the mineral matrix, in an environment protected from exposure to contamination sources (e.g., see Agee et al., 2013). Similarly, if organic carbon deposits are cross-cut by geologic features in the sample, such as a vein or the broken edge of a clastic sedimentary grain, then those organic deposits are indigenous to the rock.

OPEN IN VIEWER

FIG. 19.

The spatial arrangement of organisms within cracks and fissures of the martian meteorite NWA 7034 shows that has been exposed to terrestrial conditions for quite some time lead to the irrevocable conclusion that these cells (indicated by arrows) are terrestrial microbial contamination. (A) A reflected light micrograph. (B) An enlargement of the boxed region in (A). (C) A Raman peak image of β-carotene showing the location of microbial contaminants at the same scale as in (A). (D) Raman spectrum of β-carotene from a single spot from the map in (C). For additional information, see Steele et al. (2013).

As noted by Mustard et al. (2013), the location of organic molecules on/in the sample(s) and with respect to host mineral assemblages at the microscopic scale would provide crucial insight into the origin of these organic molecules (whether terrestrial contamination versus martian). Organic molecules located in the rock core interior may have a very different meaning than the same molecules found on the surface of the core. Surface removal or excavation (e.g., ion beam milling/sputtering) combined with microanalytical capabilities (e.g., nanoSIMS or TOF-SIMS) would be essential technologies for the analyses of returned samples. Mustard et al. (2013; Section 6.3.5.2) presented an important case history involving the Tissint martian meteorite that illustrates these points.

One aspect of this is that a priority for interpreting organic contamination in rock samples is to make them available to the analysts in good mechanical condition. If samples are pulverized, or even worse, the fragments move relative to each other, a significant amount of interpretive power would have been lost. In analogous terrestrial studies, it is common for the spatial distribution and morphology to play a significant role in the interpretation of the origin of organics (Allwood et al., 2013). Therefore, maintaining the original structure of martian samples to the extent possible is crucial to enabling confident interpretation of the origin of any detected organics. Some important potential strategies to accomplish this goal could include minimizing fracturing of the sample during collection, restraining the cores within the sample tube to minimize vibration/shock damage, and marking cores in order to record the original geometry of pieces relative to the rest of the core. These issues are discussed further by Beaty et al. (2014) and Liu et al. (2014a,b).

6.1. The case for cleaner

Sample return missions have a very special place in humans' exploration of the universe, in that they provide “ground truth” for all other types of investigations—computational, laboratory, and flight missions. As seen repeatedly in previous sample return missions, they often prompt fundamental shifts in our understanding. Apollo lunar samples revolutionized our understanding of the evolution and current state of the Moon (e.g., Taylor, 1975). The Stardust mission provided the first physical evidence of pan–Solar System radial mixing in its formative stages (Wooden et al., 2007). The Genesis mission has delivered the first direct measurement of the oxygen isotopic composition of the Sun, among other measurements (McKeegan et al., 2011). All of these findings have proven invaluable in addressing current hypotheses about our Solar System.

Sample return missions provide uniquely valuable information that cannot be obtained in other ways. Fundamentally, sample return missions provide three important things: (1) the mission science team gets to select their samples, (2) sample context is well known on a range of scales, from knowing the originating body all the way to understanding the specific sampling site(s), and (3) the mission can carefully control and document the contamination history of the sample. Meteorites also provide samples, but they are biased towards tough materials that survive impact-generated transfer, arrive on Earth without geologic context, and become contaminated with terrestrial materials upon impact. So while they are valuable, from an organic standpoint, they are not a replacement for sample return.

Although sample return missions are very scientifically valuable, they are also inherently very difficult, and MSR will be especially so. There is thus a strong case to be made that the MSR program is not merely an opportunity, but an extraordinary opportunity. Although we often talk about scientific success or failure, the reality is considerably more complex. The more scientific measurements we are able to make (i.e., measuring signals that clearly exceed contamination backgrounds), the more we will learn. Indeed, it is conceivable that we could learn a great deal more from these samples than the minimum required to declare scientific “success.” There is thus a strong case to be made that we should not set contamination requirements at the highest possible level that still allows scientific success, but rather should strive to reach levels that are commensurate with the extraordinary nature of these samples.

As discussed earlier, results of previous missions (Viking 1 and 2, Phoenix) were originally interpreted to show that native soils contain vanishingly small amounts of carbon (Biemann et al., 1976; Boynton et al., 2009; Ming et al., 2009). The confirmed presence of perchlorate, however, complicates all past interpretations because perchlorate can act as an oxygen source for combustion and a chlorine source to chlorinate molecules. Furthermore, the results of experiments on Mars analog soils cast further doubt on the earlier interpretations of miniscule organic carbon contents (Navarro-González et al., 2006, 2010, 2011). MSL has now analyzed several different, martian sedimentary materials. The initial results for the aeolian drift Rocknest soil and the John Klein and Cumberland mudstone samples from Yellowknife Bay suggested that their carbon contents were either too low, or the data were rendered ambiguous by the chlorination and oxidation chemistry acting on the leaked derivatization reagents (Leshin et al., 2013; Ming et al., 2014). Subsequent data for the evolved gas and combustion experiments conducted by SAM, however, allow for the possibility that some organics detected in analyses of the Cumberland mudstone are martian (e.g., Glavin et al., 2013; Freissinet et al., unpublished data). MSL is an ongoing mission, and as new results from the SAM instrument suite develop, our understanding of Mars organics is likely to change. We must also keep in mind that Gale Crater is not likely representative of the natural variability of Mars sedimentary environments. It seems clear, therefore, that we should remain open minded about both the upper and lower bounds of organic carbon contents of Mars sediments.

Based on this and other evidence, the current OCP and previous organic contamination-related panels have recommended upper limits in the single nanogram/gram range for especially important compounds, and upper limits overall in the 10–40 ng/g range. It seems likely that such levels will allow many of the most abundant organic compounds to be measured confidently. On the other hand, it is also likely that such levels will be problematic for at least some (and possibly many) trace organic compounds. Could we do more? Fortunately, the state of the art for contamination control has improved considerably over recent years, led in part by the semiconductor industry's need for particulate- and NVR-reduced materials and processes. Commercially available sample handling hardware regularly achieves ISO 1/2 particle cleanliness levels (as per ISO 14644-1). In terms of NVR cleanliness, lunar sampling boxes used in the Apollo program achieved 10–100 ng/cm² NVR cleanliness over 40 years ago (Calaway et al., 2014). More modern publications state detection limits of various organic compounds to ppt levels, so analytical techniques are currently available to meet verification needs (e.g., Fujimoto et al., 2007). Also, NASA has the benefit of recent contamination control experience in contamination-sensitive missions such as Genesis, the James Webb Space Telescope, and the MOMA instrument for the ESA ExoMars rover. It therefore seems likely that achieving much cleaner levels of contamination is possible at reasonable expense. Although they may not be absolutely necessary, they could mean the difference between successful versus extraordinary scientific results.

6.2. Summary and conclusions

We do not know what organics we would find in martian samples that are returned to Earth. This fact alone makes predicting the required cleanliness of returned samples highly ambiguous. Although we have attempted to constrain the problem to within an order of magnitude, we emphasize that much uncertainty remains. Given the high sensitivity of modern analytical techniques (detection limits of <1 pg), setting contamination limits that are low enough to remove any risk of interfering with any scientific measurements is not achievable. Thus choosing a discrete limit within the bounds we provide would implicitly require accepting some level of scientific risk (i.e., that analyte concentrations in the returned samples are lower than we have anticipated). The panel members varied substantially in their opinions about the level of scientific risk that is appropriate for this mission, and our recommendations fall somewhere in the middle of this range. Several panelists in particular felt that the limits should be substantially more conservative, to ensure a higher likelihood of ultimate scientific success, albeit at a (presumably) higher cost. Resolving such questions should involve a full consideration of costs and benefits associated with different levels of cleanliness, but our group was not able to evaluate the cost side of that equation.

The first logical step in setting limits is to determine what kinds of organics we care most about, and here our recommendation is that measurements of the individual organic molecules that would be the most scientifically valuable must be made. These are the measurements that should be most carefully protected from interference by Earth-sourced organic contaminants. A limit on TOC can play a supporting role as a blanket insurance policy against all possible contaminants, but should not be viewed as an end unto itself. Rather, any limit on TOC should be tightly linked to the perceived need to limit individual contaminants. We also agreed that virtually all organic compounds are of potential concern, either because we might potentially find them on Mars, or because they might plausibly interfere with measurements of martian organics. At the same time we recognize that certain molecules (mainly those known to be associated with terrestrial life, and those that have been detected on Mars) are of greater concern, and so recommend a two-tiered strategy for controlling individual contaminants: Tier I (see Table 4) contains compounds of greater concern, and so should have lower limits. Tier II includes everything else.

Our goal is to limit individual contaminants to levels that are below those we hope or expect to measure, but unfortunately we do not know which organic compounds would be present in those samples, nor at what concentrations. Previous groups have attempted to constrain the problem using estimates of meteoritic input fluxes, or by comparison with organic-poor terrestrial rocks. This Panel chose to rely more strongly on recently published measurements of martian meteorites and from the MSL rover; although few and far between, these measurements are at least demonstrably representative of martian materials. The Panel also considered the analytical techniques likely to be used on returned samples, and their sensitivities to various contaminants. However, because of the huge diversity of techniques, with detection limits spanning many orders of magnitude, this does not appear to provide useful constraints on allowable contamination levels. Our recommendations are consistent with protecting those measurements most likely to be used for initial characterization of samples (survey techniques), but would still be visible to other more targeted techniques.

The most valuable data for predicting expected concentrations, in our opinion, are the recent measurements of amino acids and TOC in martian meteorites, and the “tentative” detection of chlorobenzene by MSL. We infer that at least some, and perhaps most, martian rocks would contain important organic compounds at levels of a few to tens of nanograms per gram, and TOC at tens of micrograms per gram. If we assume a distribution of compound classes similar to that seen in the Murchison meteorite, then many other types of compounds should also be present at similar levels. These levels could be confidently measured against a background comprising <1 ng/g per compound. Given the small number of data, we point out that there is much uncertainty about whether these are typical values, or whether returned samples might have higher or lower concentrations. Clearly having cleaner samples would provide the opportunity to measure more (lower abundance) compounds in more organic-poor rocks, thus maximizing scientific return. Nevertheless, adopting <1 ng/g per compound as a guideline should allow robust measurements in some rocks and so provide the ability to meet stated scientific goals.

A significant problem in implementing contamination limits lies in the vast number of potential contaminant compounds, which are far too numerous to quantify individually. One potential approach would be to limit TOC itself to below 1 ng/g. This is the most conservative approach, and thus the most scientifically desirable one. Although technologically feasible, it is difficult at the levels required and thus expensive. A second approach would be to explicitly monitor and limit Tier-I compounds at <1 ng/g, and TOC at <10 ng/g. This would ensure that all Tier-II compounds are present at <10 ng/g. The third and least conservative approach is to explicitly monitor Tier-I compounds, passively monitor Tier-II compounds at <10 ng/g, and then relax the TOC limit to 40 ng/g in order to limit the number of contaminants that can simultaneously be present.

There are several important strategies for recognizing terrestrial contamination in returned samples. First and foremost is a comprehensive plan of contamination monitoring and characterization during spacecraft construction. We would not achieve non-detectable levels of cleanliness, and it is essential to know—at the molecular level—what residual contaminants are being carried along. Second, witness plates can be used to monitor ongoing contamination during all stages of the project, and some of these would need to make the round trip to Mars. Experience dictates that these would be in high demand by scientists studying returned samples. Third, we deem the sampling (on Mars) and return of negative control standards (blanks) to be absolutely essential to building convincing evidence for the identity of martian organic molecules. We recommend planning to include multiple blanks on the rover. Fourth, an archive of all materials used to construct the spacecraft should be developed to provide future scientists with the ability to look for novel sources of contamination. Maintaining such an archive for decades, without further contamination, would be nontrivial. Fifth, a strategy that would undoubtedly be employed to assess the indigenous nature of any organics is their spatial distribution in the samples. For example, contaminants transferred from hardware are more likely to be present on sample surfaces. Preserving the original structure of samples during the caching and return operations would be of the utmost importance.

6.3. Topics for future work

During its deliberations, the OCP recognized several issues that would benefit from either further technology development, or further discussion by a subsequent group. Note that some of the items in the lists that follow are directly relevant to the proposed Mars 2020 rover, and others relate to other aspects of potential returned sample science. The lists that follow are not listed in priority order.