Research Article

Volatile and Organic Compositions of Sedimentary Rocks in Yellowknife Bay, Gale Crater, Mars

See allHide authors and affiliations

Science  24 Jan 2014:
Vol. 343, Issue 6169, 1245267
DOI: 10.1126/science.1245267


H2O, CO2, SO2, O2, H2, H2S, HCl, chlorinated hydrocarbons, NO, and other trace gases were evolved during pyrolysis of two mudstone samples acquired by the Curiosity rover at Yellowknife Bay within Gale crater, Mars. H2O/OH-bearing phases included 2:1 phyllosilicate(s), bassanite, akaganeite, and amorphous materials. Thermal decomposition of carbonates and combustion of organic materials are candidate sources for the CO2. Concurrent evolution of O2 and chlorinated hydrocarbons suggests the presence of oxychlorine phase(s). Sulfides are likely sources for sulfur-bearing species. Higher abundances of chlorinated hydrocarbons in the mudstone compared with Rocknest windblown materials previously analyzed by Curiosity suggest that indigenous martian or meteoritic organic carbon sources may be preserved in the mudstone; however, the carbon source for the chlorinated hydrocarbons is not definitively of martian origin.

Curiosity landed in Gale crater on 6 August 2012 (UTC) with a goal to explore and quantitatively assess a site on Mars’ surface as a potential habitat for past or present life. A topographic low informally named Yellowknife Bay located about 0.5 km northeast of the landing site was chosen as the first major exploration target because strata exposed were inferred to be fluvio-lacustrine deposits (1). Fluvio-lacustrine depositional systems are thought to preserve measurable evidence of paleo-habitability (2), e.g., factors such as mineral associations, elemental inventory, redox state, and character of light elements and compounds.

Curiosity entered Yellowknife Bay on sol 125 (12 December 2012) and began a drilling campaign to obtain powder samples from a mudstone located near the base of an exposed stratal succession (3, 4). Two drill samples informally named John Klein (drilled on sol 183) and Cumberland (drilled on sol 279) were extracted for delivery to the Sample Analysis at Mars (SAM) (5) and Chemistry and Mineralogy (CheMin) (6) instruments. The samples were obtained from the lowermost stratigraphic unit in the Yellowknife Bay formation, informally named the Sheepbed member (1). The Sheepbed member and an overlying medium- to coarse-grained sandstone, named the Gillespie Lake member, appear to have a complex postdepositional aqueous history. Apparent low matrix permeability in these units suggests that they were lithified during an early diagenetic event followed by at least one additional aqueous episode when Ca-sulfate minerals precipitated in a network of intersecting fractures (7). Millimeter-sized nodules and hollow nodules in the Sheepbed member are interpreted, respectively, to be concretions and void spaces possibly formed as trapped gas bubbles during early diagenesis and lithification (1). The geology, stratigraphy, and diagenetic history of Yellowknife Bay are described in detail in companion papers (1, 7, 8).

The John Klein (JK) and Cumberland (CB) targets were drilled about 3 m apart in the Sheepbed mudstone and within ≈10 cm of the same stratigraphic position. The JK drill hole intersected thin Ca-sulfate–rich veins. The CB sample was collected from an area rich in nodules and poor in Ca-sulfate–rich veins, to aid in mineralogical and geochemical characterization of the nodules. Powders extracted from both holes were gray in color, suggesting a relatively unoxidized material (1, 8), in contrast to the red-colored, oxidized materials observed earlier by Curiosity at the Rocknest aeolian deposit (9, 10) and other surface soils (11) encountered by previous missions (12). Additional details on the drill holes are described in (1) and (8), including maps of light-toned fractures in the drill hole walls (8).

Here we describe the volatile and organic C content of the Sheepbed mudstone and evaluate its potential for preservation of organic C. Volatile-bearing phases (including possible organic material) in Sheepbed are indicators of its past environmental and geochemical conditions and can shed light on whether the environment recorded in this mudstone once was habitable, i.e., met the requirements for microbial life as known on Earth (13). The volatile and organic compositions of JK and CB materials were characterized by the SAM instrument’s evolved gas analysis (EGA), gas chromatography–mass spectrometry (GCMS), and tunable laser spectroscopy (TLS) experiments (14). Four JK subsamples (JK-1, JK-2, JK-3, and JK-4) and four CB subsamples (CB-1, CB-2, CB-3, and CB-5) of the <150-μm-size fraction of drill fines were delivered to SAM for EGA and GCMS analyses (15).

Evolved H2O

The most abundant gas evolved from JK and CB materials was H2O. H2O abundances released from JK [1.8 to 2.4 weight percent (wt %) H2O] and CB (1.7 to 2.5 wt % H2O; Table 1) were similar to those of Rocknest (1.6 to 2.4 wt % H2O) (10). Other major evolved gases, in descending order of abundance, were H2, CO2, SO2, and O2 from JK and H2, O2, CO2, and SO2 from CB (Table 1 and Figs. 1 and 2).

Table 1 Evolved gas abundances released during SAM pyrolysis runs of samples (<150 μm fraction) obtained from the John Klein and Cumberland drill holes.

Rocknest aeolian material evolved gas abundances are provided for comparison with the Sheepbed mudstone materials.

View this table:
Fig. 1 Evolved gas analysis for a continuous ramp SAM pyrolysis of drill material (<150 μm) from the Sheepbed mudstone JK-4 sample.

(A) Most-abundant evolved gases [CO2 has been scaled up 3 times; SO2 and O2 have been scaled up 10 times]. (B) Evolved gas traces for H2, H2S, HCl, CH3Cl, and CH2Cl2 [H2S and HCl traces have been scaled up 35 times; CH3Cl and CH2Cl2 have been scaled up 100 times]. Isotopologs were used to estimate species that saturated the QMS detector (m/z = 12 for CO2 and m/z = 20 for H2O).

Fig. 2 Evolved gas analysis for a continuous ramp SAM pyrolysis of drill material (<150 μm) from the Sheepbed mudstone CB-2 sample.

(A) Most-abundant evolved gases [CO2 and SO2 counts have been scaled up 3 and 15 times, respectively]. (B) Evolved gas traces for H2, H2S, HCl, CH3Cl, and CH2Cl2 [H2S and HCl traces have been scaled up 35 times; CH3Cl and CH2Cl2 have been scaled up 100 and 300 times, respectively]. Isotopologs were used to estimate species that saturated the QMS detector (m/z = 12 for CO2 and m/z = 20 for H2O).

An independent estimate of the volatile inventory of JK and CB can be obtained from measurements made by the Alpha Particle X-ray Spectrometer (APXS) (16). The measurement calculates the bulk concentration of the aggregate of excess light elements (including H2O, CO2, C, F, B2O3 and Li2O) using the relative intensities of Compton- and Rayleigh-scattering peaks (14, 17). Estimates of the average excess light-element concentrations for the JK drill tailings (APXS measurement on sol 230) and the CB drill tailings (sol 287) were 4.3 (±5.5) and 6.9 (±6.2) wt %, respectively. The two methods for determining volatile abundances in the mudstone are consistent within uncertainties.

The JK and CB samples showed similar releases of H2O in EGA experiments with a continuous temperature ramp (Figs. 1A and 2A) (18, 19). Evolved H2O from JK (JK-4) resulted in two major H2O releases, with very broad peaks at about 160° and 725°C (Fig. 1A). Cumberland samples exhibited similar behavior (Fig. 2A). The majority (~70%) of H2O was driven off in the lower-temperature peak.

CheMin results constrain the potential phases releasing H2O in the lower peak in JK and CB samples. CheMin detected basaltic silicate minerals (feldspar, pyroxene, olivine), magnetite (maghemite), anhydrite, bassanite, akaganeite, sulfides, and ~30 wt % x-ray amorphous components in addition to a 2:1 trioctahedral phyllosilicate in JK and CB (8). Therefore, candidates for the lower-temperature water release are H2O adsorbed on grain surfaces, interlayer H2O associated with exchangeable cations in 2:1 phyllosilicates (e.g., smectite), structural H2O (e.g., bassanite), structural OH (e.g., Fe-oxyhydroxides such as akaganeite), and occluded H2O in glass or minerals. Adsorbed H2O and interlayer H2O in 2:1 phyllosilicates will generally release water below 300°C. Bassanite (CaSO4·½H2O) dehydrates at ~150°C. Akaganeite [FeO(OH,Cl)] undergoes dehydroxylation at ~250°C (fig. S2). H2O incorporated into the amorphous components (e.g., nanophase Fe-oxides, allophane/hisingerite) may also evolve below 450°C. Water as liquid or vapor inclusions in glass or minerals would be released over a wide range of temperatures. Additional sources of evolved H2O at low temperatures not constrained by CheMin include structural H2O in oxychlorine compounds (e.g., hydrated perchlorates), structural OH in organics, and H2O formed during organic reactions in the SAM pyrolysis oven. Organic matter can release H2O over a wide range of temperatures from structural O and H as a consequence of reactions that take place in the SAM oven.

The high-temperature H2O release between 450° and 835°C is consistent with the dehydroxylation of the octahedral layer of a 2:1 phyllosilicate (e.g., smectite), although other phases that contain OH in octahedral layers may also evolve H2O in this temperature region. The basal (001) spacing of the JK phyllosilicate measured by CheMin was mostly collapsed to ~10 Å suggesting a 2:1 phyllosilicate with little interlayer H2O (8). The 2:1 phyllosilicate in the CB sample was expanded, with a basal (001) spacing of ~13 to 14 Å, consistent with several possible interpretations; the CB phyllosilicate could be smectite with an interlayer partially occupied by metal-hydroxyl groups or smectite with high–hydration-energy cations (e.g., Mg2+), facilitating retention of H2O (8). The high-temperature H2O releases at 750°C can be ascribed to dehydroxylation of Mg- and Al-enriched octahedral sheets in 2:1 phyllosilicates (Fig. 3A). The most likely candidate for the high-temperature H2O release (i.e., ~750°C) in Sheepbed material is saponite or Fe-saponite based on CheMin measurement of the 02ℓ diffraction band (4.58 Å) as consistent with a trioctahedral 2:1 phyllosilicate (8) and the water-release peak temperature (Fig. 3A). If all of the H2O released between 450° and 835°C during SAM pyrolysis runs resulted from dehydroxylation of 2:1 phyllosilicates, the proportions of 2:1 phyllosilicate present in the JK and CB samples are 17 (±12) wt % and 16 (±11) wt %, respectively. These values are consistent with the independent estimates from CheMin x-ray diffraction semiquantitative data, which give 22 (±11) and 18 (±9) wt % for 2:1 phyllosilicate in JK and CB, respectively (8).

Fig. 3

Major evolved gases from the JK and CB samples compared with EGA of analog minerals under SAM-like oven operating conditions (14). (A) Evolved H2O for nontronite, montmorillonite, and saponite compared with JK and CB evolved H2O. (B) Evolved O2, SO2, HCl, and H2S for CB-2 compared with pyrrhotite mixed with a Ca-perchlorate run under SAM-like operating conditions in laboratory experiments (HCl evolution coincides with O2 release for CB-2 and laboratory pyrrhotite/perchlorate experiments). (C) Evolved O2 for several perchlorate salts compared with JK and CB evolved O2.

High-temperature release of H2 occurs over roughly the same temperature regions as the high-temperature releases of H2O and H2S (Figs. 1 and 2). The origin of the high-temperature evolved H2 is unknown but is likely associated with the dehydroxylation of the most thermally stable OH groups in the 2:1 phyllosilicates.

Evolved O2

The JK and CB samples have distinctly different O2 releases (Figs. 1A and 2A and Table 1). The onset of O2 evolution from JK (~150°C) was lower than for CB (~230°C). O2 abundances released from CB (0.3 to 1.3 wt % Cl2O7) were nearly eight times the abundances for JK (0.07 to 0.24 wt % Cl2O7); Rocknest O2 abundances (0.4 wt. % Cl2O7) were about four times those for JK (Table 1). By comparison, the perchlorate anion (ClO4) was present in soil at the Phoenix landing site at the 0.4 to 0.6 wt % level (20). The JK-4 sample had two distinct peaks, suggesting different or additional O2-evolving phases in the JK sample or consumption of O2 during combustion of organic materials (see below) or thermal oxidation of ferrous-containing phases (e.g., magnetite to maghemite transition). O2 evolution in the Rocknest aeolian material occurred at a higher temperature (onset ~300°C with a peak temperature ~400°C) than in JK and CB (10, 21).

Evolved O2 from JK and CB is inferred to result from decomposition of perchlorate or chlorate salts, based on analogy with other analyses on Mars, on bulk compositions of the JK and CB samples, on the timing of chlorinated hydrocarbon and HCl releases, and on laboratory experiments with perchlorate salts. Perchlorate was definitively identified in soil at the Mars Phoenix landing site (20), and O2 release from the Rocknest aeolian material is roughly consistent with decomposition of Ca-perchlorate (10, 21). The CB sample contains three times the Cl of the JK sample as measured by APXS (1.41 wt % versus 0.4 wt % for CB and JK, respectively) (7), and the abundance of O2 released from CB (splits CB-2 and CB-3) is nearly eight times the abundance from JK (Table 1), suggesting that the substance responsible for the release of O2 from CB was a perchlorate or chlorate. Similarly, HCl and chlorinated hydrocarbons (chloromethane and dichloromethane) were released in conjunction with O2 (Figs. 1 and 2, see below), suggesting that Cl and O were hosted by the same compound in CB and JK.

The coincident release of HCl with O2 in CB is consistent with several types of perchlorate salt (Fig. 2). HCl is evolved during thermal decomposition of Mg- and Fe-perchlorate, caused by reaction between Cl2 gas and water vapor (2224). Thermal decomposition of Ca-perchlorate alone does not yield substantial HCl at temperatures <450°C (22, 25). However, thermal decomposition of Ca-perchlorate in the presence of an Fe-bearing mineral such as pyrrhotite can also yield simultaneous releases of O2 and HCl (Fig. 3B).

The O2-release profiles and temperatures for JK and CB do not match exactly those of common perchlorate salts. Although the best matches are with Fe-perchlorates (Fig. 3C), the presence of Fe-oxides/oxyhydroxides may lower the decomposition temperature of perchlorate salts (26). Chlorate salts may also be stable on the martian surface (27), and mixtures of K-chlorate and hematite can decompose at temperatures consistent with O2-release temperatures observed in JK and CB (26). Other possible sources of the low-temperature O2 release—e.g., peroxides and superoxide radicals—cannot be ruled out (2830).

Evolved CO2

The CO2 releases for the JK and CB samples peaked at temperatures below 300°C (Figs. 1A and 2A), distinct from the CO2 release between 400° and 512°C from the Rocknest aeolian materials (10). The CO2 releases in the Rocknest samples were interpreted to derive largely from carbonates (10), and the CO2 release shoulder around 400° to 450°C in the JK samples could also derive from carbonate minerals, specifically fine-grained Fe/Mg-carbonate (10, 24, 31). The 400° to 450°C release shoulder is absent from the CB samples. Another possible CO2 source, given the inferred presence of akaganeite and substantial proportions of perchlorate or chlorate phases in the samples, is that HCl evolved at lower temperatures and then reacted with carbonate minerals (23). The onset of evolved HCl is nearly simultaneous with CO2 releases in JK and CB (Figs. 1 and 2), suggesting that low-temperature acid dissolution and subsequent thermal decomposition of carbonates may be responsible for some of the evolved CO2 (fig. S2). Total CO2 evolved is equivalent to <1 wt % carbonate and, if present, carbonates are at abundance below the detection limit by CheMin. Adsorbed CO2 is an unlikely candidate for the CO2 peak near 300°C because most adsorbed CO2 is expected to be desorbed from smectite and palagonite-like material surfaces at temperatures <200°C (32). The low-temperature shoulder around 100° to 200°C in JK materials could reflect adsorbed CO2, although it was not seen in CB materials.

Although there are several possible CO2 sources in JK and CB materials, the simultaneous evolution of CO2 and O2, in conjunction with a possible O2 inversion (i.e., O2 consumption) in JK-4 and the similar CO2 and O2 releases in CB samples, suggest combustion of C compounds. It is nearly certain that at least some of the CO2 produced is derived from the combustion of vapor from N-methyl-N-(tert-butyldimethylsilyl) trifluoroacetamide (MTBSTFA, a derivatization agent carried in SAM) and its reaction products that were identified by the EGA and GCMS experiments and adsorbed onto the samples and sample cups inside SAM during sample transfer in Curiosity’s sample acquisition and processing system (10, 21, 33). The background-derived C detected in the blank runs was up to ~120 and 30 nmol of C attributed to MTBSTFA and dimethylformamide (DMF) (33), respectively. If the background was similar for the analyzed samples of the mudstone, another source of C for combustion to CO2 during pyrolysis is required to account for the >2 μmol of evolved CO2 (Table 2). The estimated amount of MTBSTFA + DMF C in the blank runs is only 1 to 3% of total evolved CO2-C from JK and CB analyses. Also, lower amounts of MTBSTFA C (~18 nmol C) and DMF C (~15 nmol C) were detected in the CB-5 analysis (Table 2), suggesting that substantially less MTBSTFA and DMF were available for combustion to CO2 in this run due to implementation of the MTBSTFA-reduction protocol (14). These results indicate that most of the evolved CO2 from CB is not related to the known terrestrial C background in SAM, and therefore additional C sources are required.

Table 2

Abundances of terrestrial C from N-methyl-N-(tert-butyldimethylsilyl) trifluoroacetamide (MTBSTFA) reaction products and dimethylformamide (DMF) compared with the measured abundances of chloromethane (CM), dichloromethane (DCM) and CO2 detected during SAM evolved gas analysis (EGA) by the mass spectrometer.

View this table:

The initial amount of MTBSTFA and DMF C in the JK and CB analyses could have been higher than the levels measured in the empty-cup blank because of additional adsorption of these volatiles to the solid-sample surface area after sample delivery. A triple-sized sample portion (~135 mg of sample) of JK (JK-3) was delivered to SAM to explore the effects of adsorption on measured CO2 releases. Estimates of the amount of C from MTBSTFA and DMF sources that could contribute to evolved CO2 for the JK-3 triple-portion sample show that the levels are similar (within error) compared to the single-portion JK sample analyses (column 4 of Table 2), despite the potential for at least three times as much MTBSTFA and DMF adsorption to the sample due to greater surface area. Also, the triple-portion sample evolved 2.4 to 3.5 times as much CO2 (Table 2). MTBSTFA and DMF C likely contributed to a small portion of the evolved CO2 (21). The order-of-magnitude more C observed as CO2 in all samples and the near threefold increase in CO2 observed for the triple portion run further demonstrate that the dominant C source for CO2 in the JK analyses came from the mudstone itself and not from known background C sources in SAM or Curiosity’s sample acquisition and delivery system.

Another possible C source for the evolved CO2 is combusted martian indigenous and/or exogenous (meteoritic) organic matter in the mudstone. The Sheepbed mudstone has trace-element compositions consistent with a meteoritic contribution that may have delivered 300 to 1200 parts per million (ppm) organic C (7) and/or from weathering of igneous material (34). Also, metastable partially oxidized weathering products of martian organics of indigenous and/or exogenous origins such as mellitic acid (35) may have undergone decarboxylation during analysis in SAM. Laboratory analog experiments using SAM-like instrument conditions where mellitic acid and perchlorate salts were heated together have shown that the primary degradation products detected during pyrolysis are CO2 and CO (36, 37). If mellitic acid or other benzenecarboxylates were present in Sheepbed, the organic degradation products may have decomposed to CO2 and other less-volatile degradation products, which could have gone undetected by SAM GCMS under the oven conditions used for the pyrolysis experiments.

Overall, the potential CO2 sources include combustion of terrestrial organics resident in SAM (for a small portion of the evolved CO2), low-temperature acid dissolution of martian carbonates, and combustion and/or decarboxylation of indigenous and/or exogenous organic materials.

Evolved SO2 and H2S

Sulfur dioxide and H2S evolved from the JK and CB samples during pyrolysis, and both also evolved from the Rocknest aeolian material (10). The release of both reduced and oxidized S volatiles suggests that reduced and oxidized S species were present in all samples, or that redox reactions in the SAM oven affected S speciation. The total abundance of S-bearing gases and the ratio of SO2/H2S observed at JK were both <30% of the values measured in Rocknest samples. Abundances of these gases were even lower in CB than in JK, consistent with the lower bulk S composition measured by APXS [1.57 (±0.03) wt % SO3 for CB compared with 5.52 (±0.21) wt % SO3 for JK (7)].

The evolved SO2 had a release peak temperature at ~600° to 625°C, with a shoulder at ~675°C (Figs. 1A and 2A). Several mineral sources of S are possible. CheMin detected anhydrite, bassanite, and pyrrhotite in both JK and CB and possible pyrite in JK (8). Pyrrhotite and possibly pyrite are candidate S sources for the SO2 and H2S releases from CB and JK. The lower-temperature SO2 and HCl evolutions occurred simultaneously with the release of O2 in CB (Fig. 2), and two additional SO2 release peaks were observed in the 500° to 800°C range. Laboratory pyrolysis experiments of mixtures of pyrrhotite and Ca-perchlorate exhibited similar release patterns for SO2, O2, and HCl (Fig. 3B); however, the onset temperatures for their release is lower in CB, consistent with a different perchlorate/chlorate salt (i.e., lower O2 release) or complex chemistry occurring in the SAM ovens that lowers the decomposition temperature of oxychlorine compounds (as discussed above). Thermal decomposition of Ca-sulfate is not likely to have contributed to the SO2 releases from JK and CB because they typically break down at higher temperatures than the maximum that can be achieved by the SAM oven used in these experiments (>835°C).

Although pyrrhotite and pyrite are candidate S-bearing phases in CB and JK, the evolved SO2 data are not uniquely diagnostic of these or any specific S-bearing phases. Fe-sulfate minerals will evolve SO2 at 500° to 800°C under conditions similar to SAM operational conditions. However, formation of Fe-sulfates requires strongly acidic conditions (38, 39), and Sheepbed is interpreted to record depositional and diagenetic environments at near-neutral pH’s (1).

Evolution of H2S occurred nearly simultaneously with evolution of H2 and high-temperature H2O resulting from the dehydroxylation of the 2:1 phyllosilicate (Figs. 1 and 2). H2S may be a by-product of the reaction of H2O with a Fe-sulfide such as pyrrhotite (Fig. 3B); however, it is possible that SO2 evolved at high temperatures is reduced in the presence of H2 to H2S (40, 41). HCl also evolves at higher temperatures (Figs. 1B and 2B), which can react with reduced S phases to form H2S (42, 43).

Evolved N-Bearing Species

Potential N-bearing compounds evolved from JK and CB include NO, HCN, CH3CN, ClCN, CF3CN, and C3H4F3NO. Evolved NO [mass/charge ratio (m/z) = 30] in the JK and CB materials had abundances of 129 to 190 nmol and 190 to 389 nmol NO, respectively (Table 1). The abundances of NO in JK and CB blank runs were 70 and 12 nmol, respectively, and the predicted level of N contributed by MTBSTFA and DMF based on background measurements was typically less than 20 nmol N. Hence, the source for the evolved NO appears to be within the mudstone. Other N compounds detected by SAM (HCN, CH3CN, and ClCN) are present at substantially lower abundances, and they may be contributed by the MTBSTFA and DMF background in the SAM instrument. Both wet chemistry reagents contain one N atom per molecule of reagent. CF3CN is almost certain to be from decomposition of MTBSTFA because possible sources of F in the martian samples (i.e., fluorapatite) will not decompose in the SAM temperature range. In addition, HCN and CH3CN have also been identified during laboratory pyrolysis of MTBSTFA and DMF in the presence of perchlorate run under similar operating conditions as SAM (21).

Organic Compounds

Pyrolysis of the JK and CB samples led to low-temperature (125° to 350°C) release of chloromethane and dichloromethane that correlated with the release of O2 and 1,3-bis(1,1-dimethylethyl)-1,1,3,3-tetramethyldisiloxane, a known reaction product of MTBSTFA and H2O (fig. S3). This correlation suggests that thermal degradation of the O2 source (most likely an oxychlorine compound, see above) is contributing to chlorination of C phases and/or the release of chlorinated hydrocarbons. Identification and quantification of these trace organic species by their characteristic m/z values in EGA mode was enhanced and confirmed by bulk collection and GCMS identification (44) of volatiles on an adsorbent-resin hydrocarbon trap (Fig. 4 and fig. S4). Detection of chlorinated hydrocarbons by SAM in Rocknest samples (10, 21), as well as supporting laboratory EGA and GCMS experiments conducted under SAM-like conditions, have shown that both chloromethane and dichloromethane are produced when MTBSTFA and DMF are heated in the presence of Ca- and Mg-perchlorates (21). Therefore, reaction of MTBSTFA and DMF carbon with an oxychlorine compound is a likely source of some of the chloromethane and dichloromethane detected in JK and CB.

Fig. 4 SAM gas chromatogram of the major m/z values of the chlorinated hydrocarbons detected in JK and CB samples (cps, counts/s).

(A) JK Sample 3 (JK-3; triple portion) compared with JK-2 (single portion) and the JK blank. (B) CB Sample 2 (CB-2, single portion) compared with the CB blank. (C) The mass spectra generated for the GC peaks detected in JK-3 are shown in red and compared with the mass spectra for chloromethane (CH3Cl), dichloromethane (CH2Cl2), and trichloromethane (CHCl3) in black from NIST/EPA/NIH Mass Spectral Database (44).

Empty-cup blank analyses before each sample set (Rocknest, JK, and CB) showed low chloromethane C abundance (<7 nmol C, including estimated error, via EGA) and no detection of dichloromethane. In comparison to the blanks, single-portion runs of JK and CB consistently showed more abundant chloromethane and dichloromethane C [up to about six times the abundance; range: 21 (± 4) to 66 (±13) nmol C, Table 2], whereas lower amounts of chloromethane and dichloromethane [12 (± 2) nmol C] were released from Rocknest single-portion samples, even though the total MTBSTFA C amounts measured during pyrolysis of Rocknest were similar (within error) to those measured in JK and the CB-1 and CB-2 runs (Table 2). Furthermore, the JK-3 EGA analysis of a triple-portion sample released approximately twice as much chloromethane + dichloromethane [127 (±25) nmol C] as a single-portion JK analysis [38 (±8) to 66 (±13) nmol C], and JK-3 released trichloromethane (CHCl3) and carbon tetrachloride (CCl4), which were not previously detected in the blank or single-portion JK experimental runs (table S2 and Fig. 4). However, the increase in chlorinated products in JK-3 could also be explained by three times as much sample (i.e., more oxychlorine compounds) compared to the single-portion JK runs. Trichloromethane and carbon tetrachloride were also detected by GCMS in both CB-3 and CB-5 at similar levels (table S2). These data indicate that greater chlorinated hydrocarbon production is associated with the Sheepbed mudstone (compared with the Rocknest aeolian deposit) and with larger sample mass. However, the substantial reduction (~50%) in C abundance from MTBSTFA (and presumably DMF) in the CB-5 experiment compared with CB-3 (fig. S5 and Table 2) was matched by a ~50% drop in chloromethane and dichloromethane detected by EGA [10.2 (±2.0) and 0.7 (± 0.2) nmol C, respectively]. This suggests that martian organics may not be substantial contributors to these chloromethanes detected in CB samples.

Trace quantities of chloromethane and dichloromethane were also identified by the Viking 1 and 2 lander GCMS instruments (45, 46). The Viking GCMS team attributed the chlorinated hydrocarbons to terrestrial sources including cleaning solvents, although the possibility that some of the chloromethane C was indigenous to Mars was not ruled out (46). Unlike SAM, the Viking GCMS instruments did not include an MTBSTFA and DMF wet chemistry experiment. Following the Phoenix perchlorate detection and subsequent laboratory experiments in which Atacama soils mixed with 1 wt % of Mg-perchlorate produced chloromethane and dichloromethane at 500°C, the Viking chloromethane compounds have been attributed to the presence of perchlorates and indigenous organic carbon at the Viking landing sites (47). The hypothesis has been challenged (48) and debated (49).

The EGA and GCMS observations of varying chlorinated hydrocarbon abundances in JK and CB could result from any combination of the following: (i) chlorination of MTBSTFA and DMF or other unknown terrestrial C sources in SAM (instrument background) that were not identified during EGA or GCMS in the empty-cup blank runs; (ii) chlorination of C contamination from the drill and/or sample handling chain; and (iii) chlorination of martian or exogenous C phases in the Sheepbed mudstone.

Terrestrial contamination from the sample handling chain is unlikely because it was scrubbed multiple times with Rocknest scooped material before the first drilled sample at JK (9, 10). The GCMS abundances of chloromethane and dichloromethane measured in JK-4 were similar within error to the abundances of the chlorinated hydrocarbons identified in the second drill sample at CB (CB-2) run under identical conditions (table S2). Therefore, if terrestrial C contamination from the drill was the primary source of C for these chlorinated hydrocarbons in JK-4, their abundances measured in the CB-2 sample should have been reduced compared to JK-4 due to sample scrubbing of hardware surfaces and disposal of the JK sample from the sample processing system. This was not observed. Moreover, swabbed surfaces of Curiosity’s sample acquisition and processing system were found to be organically clean before launch (50, 51), and only trace quantities (~0.1 to 0.3 nmol) of perfluoroethene, a known pyrolysis product of Teflon that is within the drill system (52), were identified in the JK and CB EGA analyses.

Chlorinated hydrocarbons are likely produced in the SAM oven during heating of samples in the presence of Cl from the thermal decomposition of oxychlorine compounds. Some of the chlorinated hydrocarbon detected at JK and CB may be derived from the sample. The SAM data do not allow us to prove, or disprove, organic C contributions from Sheepbed to evolved chlorinated hydrocarbons and CO2 from the JK and CB samples. There are no conclusive EGA or GCMS observations of other organic molecules indigenous to the Sheepbed mudstone.

Preservation of Organics

Although the detection and identification of possible organic compounds in the Sheepbed mudstone is complicated by reactions in the SAM oven, the lack of a definitive detection of martian organics suggests that, if organics were deposited in the Sheepbed mudstone, organic alteration and destruction mechanisms may present the single most fundamental challenge to the search for organics on Mars. Some organic compounds are expected on the surface of Mars. Exogenous delivery of meteoritic organics to the martian surface has been estimated as 2.4 × 108 g C/year (53) and may have been higher during periods of higher impact flux on the surface. Benzenecarboxylates derived from the oxidation of meteoritic organic matter on Mars could contribute up to 500 ppm organic C by weight in the top meter of the martian regolith (35). Abiotic organic matter formed by igneous and/or hydrothermal processes (34, 54) is also expected to be embedded within basaltic minerals where it is protected from chemical oxidants in the environment. Consequently, at least a low concentration of organic compounds is likely in the source material for the Sheepbed mudstone. In addition, the distal fluvial to lacustrine depositional environment of Sheepbed (1) makes it a prime site for concentration of organic matter through sedimentary processes (2).

If organics were present, evaluating their fate during diagenesis becomes critical for understanding their molecular structures, distribution in sediments, and SAM observations. The JK and CB samples experienced diagenesis, including the alteration of basaltic minerals to smectite and magnetite after deposition (7, 8). Organics may have been released during basaltic mineral alteration, and any reactive organic molecules present may have been incorporated into or adsorbed onto the forming smectite and magnetite. These types of mineral-organic associations could have enhanced the early preservation of organic matter. In addition, reducing conditions during deposition, as suggested by the presence of magnetite (7), would have favored early preservation. Subsequently, and at any time during burial, oxidants already present in disequilibrium with reduced phases or circulated into the rock may have contributed to the oxidative degradation of organic compounds that were deposited with the sediment or released from the altering igneous minerals. There is no evidence of mineral oxidation associated with the second diagenetic event that precipitated calcium sulfate minerals from fluids into the Sheepbed fracture network (7), and the bulk rock remained largely reducing based on the presence of magnetite and pyrrhotite in JK and CB, and possibly pyrite in JK (8). However, akaganeite in the samples may reflect oxidative weathering of pyrrhotite, which could indicate exposure of JK and CB to an oxidizing fluid in earlier diagenetic event(s) that may have degraded any organic compounds.

If any organic compounds remained throughout burial, they may have been altered by other mechanisms. The martian surface is subjected to strong ionizing radiation that can alter organic molecules (55) depending on the chemical and physical microenvironment of the host sediments (56). Ionizing radiation directly breaks chemical bonds in organic molecules and other chemicals, producing a reactive pool of radicals and oxidants (e.g., OH•, H2O2, oxychlorine compounds). In the presence of mineral catalysts, these reactants can fully oxidize organic matter to CO, CO2, and carbonates or produce partially oxidized organics such as acetate, oxalates, and other carboxylates that may survive in a metastable state on Mars (35). If exogenous or martian organic matter survived largely intact until removal of the overlying Gillespie sandstone, it may still have degraded or oxidized during surface exposure of the sampling site to ionizing radiation. Therefore, surface-exposure age is also an important variable in the preservation of organics. Alternatively, the organics may have survived all of the martian processing that occurred naturally, only to be oxidized by oxide minerals or oxychlorine compounds at elevated temperatures in the SAM oven or, if sufficiently refractory, survive pyrolysis and pass undetected by SAM, as does most of the kerogen-like material in meteorites.

The complicated story of carbon on Mars is poorly understood. As of sol 370, SAM results support the presence of carbon source(s) in samples from the Sheepbed mudstone that contribute(s) to the production of chlorinated hydrocarbons and evolved CO2. There may be organic matter in these samples, but it has not been confirmed as martian.

Supplementary Materials

Materials and Methods

Figs. S1 to S6

Tables S1 to S3

References (5765)

MSL Science Team Author List

References and Notes

  1. By definition, a mudstone is a fine-grained sedimentary rock composed of 50% or more of particles <62 μm (4). The Sheepbed mudstone is composed of ~20% clay minerals and ~30% amorphous materials (8).
  2. The term ‘‘soil’’ is used here to denote any loose, unconsolidated materials that can be distinguished from rocks, bedrock, or strongly cohesive sediments. No implication for the presence or absence of organic materials or living matter is intended.
  3. Materials and methods are available as supplementary materials on Science Online.
  4. The mass of each sample portion delivered to SAM was estimated to be 45 ± 18 mg, except for JK-3, where a triple portion (135 ± 31 mg) was delivered to the sample cup. Due to the known presence of the reaction products of MTBSTFA and possibly DMF in the SAM background due to a broken seal (10, 21), the analytical conditions of the first three JK samples were adjusted to boil off these molecules and associated reaction product by heating to 200° to 300°C and holding the temperature for about 20 min. All evolved gases were analyzed directly by the quadrupole mass spectrometer (QMS), but boil-off gases were vented thereafter to simplify GCMS analyses (14). The JK-4 sample was run under nominal conditions in which the pyrolysis oven was steadily heated from Mars ambient to ~835°C. Three samples of CB (CB-1, CB-2, and CB-3) were run under the same conditions that were used for the JK-4 sample, i.e., under a continuous temperature ramp starting at Mars ambient conditions. Sample CB-5 SAM cup was preheated to ~120°C and then moved into position directly under the solid sample inlet tube to receive the CB-5 portion from Curiosity’s sample delivery system (14).
  5. Only the John Klein sample 4 (JK-4) EGA data set is discussed in the text and shown in Fig. 1. The other three experiments were conducted with a preconditioning boil-off that prevented systematic examination of the temperature of evolution below the boil-off temperature [~300°C (15)].
  6. The three SAM continuous temperature ramp analyses of the Cumberland material (CB-1, CB-2, and CB-3) revealed similar evolved gas patterns (fig. S1), with variations in the abundances of gases evolved. Only the CB-2 sample is shown in Fig. 2 to represent the Cumberland evolved gas analysis. CB-5 EGA is not shown in this figure because it did not have the sample continuous temperature ramp conditions that were used in the previous three CB runs.
  7. MTBSTFA (N-methyl-N-tert-butyl-dimethyl-silyl-trifluoro-acetamide) and DMF have been identified in both empty-cup blank and Rocknest runs. See supplementary materials on Science Online.
  8. The National Institute of Standards and Technology/Environmental Protection Agency/National Institutes of Health (NIST/EPA/NIH) Mass Spectral Database (NIST SRD Database No. lA, Gaithersburg, MD, 2011) was used to identify chloromethane, dichloromethane, trichloromethane, and tetrachloromethane in GCMS measurements based on spectral lines in the gas chromatograms. The confirmation of these compounds aided in their identification and quantification in EGA.
  9. Acknowledgments: We are indebted to the Mars Science Laboratory Project engineering and management teams for making this mission possible and enhancing science operations. Much of this research was carried out at the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration (NASA). NASA provided support for the development of SAM. Data from these SAM experiments are archived in the Planetary Data System ( Essential contributions to the successful operation of SAM on Mars and the acquisition of this data were provided by the SAM development, operations, and testbed teams. Development and operation of the SAM and APXS instruments were also supported by funds from the French Space Agency (CNES) and the Canadian Space Agency. Work in the UK was funded by the UK Space Agency. B.L.E., J.L.E., K.F., D.P.G., J.E.G., K.E.M., S.M.M., J.M., P.B.N., and R.E.S. acknowledge funding support from the NASA ROSES MSL Participating Scientist Program.

Stay Connected to Science

Navigate This Article