Mars-Analog Calcium Sulfate Veins Record Evidence of Ancient Subsurface Life

2.1. Gypsum vein sample

The gypsum vein analyzed here was collected in hand samples underground in situ from the Kirkby Thore gypsum mine, which accesses the Eden Shales Formation of the Cumbrian Coast Group in the Vale of Eden Basin, Cumbria, NW England. Gypsum has been mined here for over 100 years (Hughes, 2003). It occurs in numerous veins parallel to bedding in dark grey mudstone and derives from gypsum originally deposited in an evaporitic lacustrine basin or coastal sabkha in the late Permian. The gypsum is white and fibrous (“satin spar”), with fibers oriented normal to the vein walls, which are themselves also impregnated with gypsum cement. The veins are displacive and grew antitaxially, that is, by successive additions at the reopening and resealing margin between the vein and the host rock, causing fragments of the wall rock to be included along irregular median planes within the veins.

The age of the veins and their emplacement mechanism are poorly constrained, although the fibrous habit suggests a key role for fluid overpressure (i.e., fluid pressure in excess of hydrostatic pressure) and hydraulic fracturing (Parnell et al., 2000). Burgess and Holliday (1974) interpreted the veins in the Eden Shales Formation as mineralized hydraulic fractures formed by compaction-related pore-fluid overpressure during the burial of the host sediment. However, the wall rock fragments included within the veins are as fully compacted as the outside matrix, suggesting that veining postdated compaction. Moreover, it is now widely thought that gypsum veins can form during uplift as a consequence of fluid overpressure during tectonic compression and/or the rehydration (dissolution and reprecipitation) of anhydrite beds, which causes a volume expansion of >60% and associated hydraulic fracturing by sulfate-rich fluids. This mechanism has been invoked to explain the occurrence of gypsum veins in other Permo-Triassic evaporite basins in the west of England, which experienced Cretaceous and Cenozoic compression and uplift (Murray, 1964; Philipp, 2008; Meng et al., 2017). In the Vale of Eden, uplift began in the Cretaceous (Fig. 1). Fluid-inclusion microthermometry and palaeomagnetic data from the Penrith Sandstone Formation conformably underlying the Eden Shales Formation indicate a cooling trend beginning at that time from a maximum temperature of ∼120°C (equivalent to a depth of about 3–4 km; Turner et al., 1995). The Eden Shales Formation was uplifted through the 50°C isotherm and to depths of <1 km approximately 50 million years ago. Since anhydrite typically rehydrates to gypsum under these conditions (Zanbak and Arthur, 1986), it is likely that vein emplacement occurred at this time (and at this temperature), although an earlier or later origin cannot be ruled out.

OPEN IN VIEWER

FIG. 1.

Sketch of the burial history of the Eden Shales and Penrith Sandstone Formations. Gypsum veins in the Eden Shales Formation could have formed either during burial or during uplift as a consequence of hydraulic fracturing. Adapted from Turner et al. (1995).

2.2. Optical microscopy

Optical microscopy was undertaken with a Leica DMLP reflected/transmitted light polarizing microscope with DFC 420C camera and Leica Application Suite v 4.00.

2.3. Scanning electron microscopy and energy-dispersive X-ray spectroscopy

Scanning electron microscopy (SEM) and energy-dispersive X-ray spectroscopy (EDS) of carbon-coated polished, uncovered thin sections were undertaken at the Imaging Spectroscopy and Analysis Centre (ISAAC) at the University of Glasgow using a Zeiss Sigma analytical scanning electron microscope equipped with an Oxford Instruments X-Max silicon drift energy dispersive X-ray detector and an operating voltage of 20 kV.

2.4. Total organic carbon

The carbon contents of the host mudstone were measured by a LECO CS225 elemental analyzer, after decarbonatization with hydrochloric acid.

2.5. Raman spectroscopy

Raman spectra were acquired from an uncoated chip of the gypsum vein and host rock by using a Renishaw inVia Raman microscope at the University of Edinburgh. The 514 nm, 2 mW laser beam was focused onto the samples with a 100 × objective lens, providing an excitation spot of about 2 μm diameter. Raman point spectra were taken at different positions on the samples over the range 100–3200 cm⁻¹. The spectra were acquired with a 30 s exposure time at 50% laser power. Renishaw Wire 2.0 software was used for data acquisition.

2.6. Ion microprobe sulfur isotope analysis

Spatially resolved pyrite sulfur isotope compositions were measured on a sample of the Kirkby Thore gypsum vein with the Cameca IMS-1270 (#309) instrument at the School of GeoSciences at the University of Edinburgh, UK (the NERC Ion Microprobe Facility). Sulfur isotopes (³²S and ³⁴S) were analyzed as S^- ions produced by bombardment of the target by a ∼1 nA, ¹³³Cs+ primary beam accelerated at +10 kV resulting in a net impact energy of 20 keV at the sample surface. To eliminate charging during the analysis, the sample was coated with a ∼30 nm layer of gold, and the exact position of the primary cesium beam was flooded with low-energy electrons produced by the normal incidence electron gun. Secondary ions were accelerated at -10 kV and analyzed at a mass resolution sufficient to resolve molecular interference by ³³S¹H^- and ³⁴S¹H₂ ^- on the ³⁴S^- peak. An energy window of 40 eV was used. After pre-sputtering for 30 s, automated secondary ion beam alignment was performed by using the DT_xy deflection plates to center the beam in the mass spectrometer field aperture (2,500 μm) and entrance slit (90 μm) positions.

Measurements of sulfur isotopes were made simultaneously in multicollection mode using two off-axis Faraday cups (L′2 for ³²S and H′2 for ³⁴S), which were calibrated at the start of the analytical session with in-built Cameca hardware and software. Count rates were typically ∼2 × 10⁸ cps of ³²S and ∼9 × 10⁶ cps of ³⁴S. The total acquisition time was 100 s comprising 20 cycles, split into two blocks of 10, with each cycle comprising a 5 s counting period. A single ∼4 min analysis (including 30 s pre-sputtering time and beam centering routines) resulted in an internal error of <0.25‰. Sulfur isotope compositions are reported as δ³⁴S_VCDT, calculated as

(VCDT = Vienna Canyon Diablo Troilite).

The Balmat pyrite standard (δ³⁴S_VCDT = +14.6‰; Crowe et al., 1990) from the University of Wisconsin was mounted in the same epoxy block as the sample and was analyzed five times at the start and five times at the end of the session so that any change in instrumental mass fractionation could be monitored and corrected for. This gave mean δ³⁴S_VCDT values of +14.73‰ and +14.97‰ before and after the sample analysis respectively, with standard deviations of 0.12 and 0.24. The average difference in the standard values before and after measuring the sample was used to apply a small instrumental drift correction to the sample values.

3.1. Occurrence of pyrite and dolomite

Study of polished hand samples and thin sections of fibrous, bedding-parallel gypsum (“satin spar”) veins from Kirkby Thore revealed the presence of brassy, reflective mineral linings <1 mm thick embedded in gypsum at the interfaces between the gypsum veins and the wall rock (mudstone) above and below (Fig. 2). These linings do not occur either in the vein or in the host mudstone except near this interface. The mineral linings displayed the optical characteristics expected of a metal sulfide (opacity, brassy reflective luster; Fig. 3A–3C) but an unusually complex morphology consisting of disseminated patches of micrometer-scale crystals, rounded ∼5–25 μm globules, and growth structures 100 μm in scale consistently oriented outward from the vein margin, showing complex, micro-digitate, interfingering boundaries with the surrounding gypsum (e.g., top of Fig. 3C; right of center of Fig. 3D). The mineral was identified as pyrite (FeS₂) by EDS analysis (Fig. 3E). Subhedral rhombohedral crystals of dolomite (CaMg(CO₃)₂; identified by EDS; Fig. 3F) were also observed close to the pyrite. In one instance a dolomite crystal appeared to enclose a pyrite grain (Fig. 3D, upper right).

OPEN IN VIEWER

FIG. 2.

Macroscopic distribution of pyrite across a polished vertical cross section through a gypsum vein. (A) Photograph showing the gross appearance of the vein. Note the scale bar. (B) Labeled sketch drawn from photograph (A) showing the distribution of pyrite visible to the unaided eye as a brassy reflective material.

OPEN IN VIEWER

FIG. 3.

Pyrite and dolomite at the margins of a gypsum vein. (A) Petrographic thin section viewed in transmitted light, showing the contact between the gypsum vein (clear) and the wall rock mudstone (variably opaque). (B) The same field of view as (A) viewed in reflected light, showing the bright, reflective pyrite. (C) A more magnified view of the pyrite in reflected light, showing its finely disseminated, globular, and digitate morphologies. This sample was an epoxy-mounted wafer subsequently used for sulfur isotope analyses (shown here prior to gold coating). (D) Backscattered electron scanning electron microscope image from a region similar to (C) on a carbon-coated thin section, showing gypsum (flat grey, “gyp”), dolomite (darker grey, “dol”), pyrite (bright, “pyr”), and clay (dark, bottom left, “clay”). The pyrite shows a complex, interfingering relationship with the gypsum (e.g., right of center). The dolomite in one instance appears to enclose the pyrite (upper right). Circles indicate the positions of the EDS spectra shown in (E) and (F). (E) EDS spectrum from pyrite at point indicated on (D). (F) EDS spectrum from dolomite at point indicated on (D). Iron is present as a minor impurity.

3.2. Carbonaceous and organic matter

Samples of the mudstone hosting the gypsum veins were analyzed for total organic carbon content, yielding values of 0.22, 0.64, and 0.74 wt %; palynomorphs and phytoclasts were also observed within the mudstone using optical microscopy. The reflective pyrite exposed by polishing extends into the gypsum vein beneath the polished surface of the thin section, forming an opaque region bordered by diffuse dark material embedded within the gypsum (Fig. 4A, 4B). This material was investigated with Raman spectroscopy, which consistently detected bands at c. 1350 and c. 1600 cm⁻¹, both in thin section and in rock chips, in addition to bands attributable to pyrite and gypsum (Fig. 4C). As discussed below, these observations suggest that the dark material included disordered carbonaceous material of probable organic origin. Clear gypsum within the veins was also analyzed, yielding only the expected peaks for gypsum.

OPEN IN VIEWER

FIG. 4.

Pyrite-associated carbonaceous material. (A) Reflected light photomicrograph showing surface of polished thin section, with reflective pyrite (left) abutting nonreflective gypsum. (B) Transmitted light photomicrograph of the same field of view as (A), showing further opaque pyrite within the vein beneath the surface of the thin section, surrounded by diffuse dark material. (C) Typical Raman spectrum from the opaque region adjacent to the vein-margin pyrite, in this instance from a small chip of the rock. This spectrum shows peaks at about 1350 and 1600 cm⁻¹ consistent with the presence of disordered carbonaceous material. Additional peaks are attributed to gypsum and pyrite on the basis of comparison with reference spectra in the RRUFF database (Lafuente et al., 2015).

3.3. Pyrite sulfur isotope composition

Sulfur isotope compositions (δ³⁴S_VCDT) were measured at 13 points on a sample of the vein-margin pyrite, yielding values from -36.47‰ to -28.50‰ with a mean of -30.70‰ and a standard deviation of 2.19‰ (Fig. 5).

OPEN IN VIEWER

FIG. 5.

Reflected-light photomicrograph indicating sulfur isotope compositions (δ³⁴S_VCDT) measured at individual points on pyrite from a Kirkby Thore gypsum vein. The sample is a gold-coated, epoxy-mounted wafer with a pyrite grain from the Balmat standard mounted at its center (“Standard”). The edge of the vein is toward the bottom of the field of view in this image and is encrusted with pyrite (high relief). A well-preserved dolomite crystal also occurs in the lower right (low relief, “dol”). The dark marks associated with each measurement are pits made in the sample surface by the ion beam. The inset image shows a wider field of view; note the scale bar.

4.1. Evidence for ancient subsurface gypsum-hosted life

To our knowledge, the present study is the first to report pyrite lining the walls of gypsum veins. This disseminated, microcrystalline pyrite occurs only in close proximity to the vein–wall interface, where it is largely embedded in the gypsum; it does not appear to be a preexisting constituent of the wall rock passively entrained into the vein by antitaxial growth. Rather, the pyrite appears to have precipitated authigenically in situ and thus retains its original growth habit, including complex interfingering with the surrounding gypsum. Although the two minerals appear to be intergrown, it is unlikely that pyrite and gypsum formed competitively at the same time, since they require reducing and oxidizing conditions respectively. We suggest instead that pyrite precipitated as a replacement of preexisting gypsum, which dissolved into reducing aqueous fluids percolating through microporosity along the vein margin, perhaps long after the vein formed.

Pyrite formation in low-temperature (<80°C) depositional and postdepositional (diagenetic) settings on Earth is typically a consequence microbial sulfate reduction (MSR) coupled to the respiration of organic matter or the oxidation of H₂ (Machel, 2001; Seal, 2006). MSR produces H₂S, which reacts with iron minerals to form the metastable precursors of pyrite (Raiswell and Canfield, 1998). Sulfate-reducing microorganisms are globally abundant in subsurface porewaters, where sulfate is commonly supplied by the dissolution of calcium sulfate minerals (Machel, 2001). Hence, we infer that pyrite in the Kirkby Thore veins derives from the microbial reduction of sulfate dissolved out of the gypsum veins. Microbial consumption of sulfate ions may in turn have promoted further gypsum dissolution, creating additional microporosity into which pyrite precipitated. Such a feedback process may explain the unusual dendritic morphology of the pyrite–sulfate interface, although this texture should not be regarded as a biosignature in itself (e.g., Fig. 3C, 3D). Sulfate reduction was probably coupled to the oxidation of organic matter in the pore-fluid or derived from the host mudrock, which contains ∼0.5 wt % total organic carbon and visible palynomorphs and phytoclasts. Exposure to water when this rock was hydraulically fractured would likely have facilitated diffusion of low-molecular-weight organic acids from the vein walls and entrained fragments into the sulfate-bearing fluid (Fredrickson and Balkwill, 2007). There is no evidence for the migration of exogenous hydrocarbons through these veins.

The authigenic, subhedral dolomite intimately associated with the pyrite may also be a by-product of sulfate reduction (Baldermann et al., 2015); speculatively, the necessary Mg²⁺ and CO₃ may have derived from dissolution of dolostone elsewhere in the stratigraphic sequence (Hughes, 2003). Sulfate anions in solution can inhibit the precipitation of dolomite by forming complexes with Mg²⁺, but this effect requires medium-to-high sulfate concentrations and would therefore be suppressed by sulfate reduction, which may also favor dolomite precipitation by increasing alkalinity (Brady et al., 1996; Sánchez-Román et al., 2009). Indeed, Mg-sulfate complexes can also promote dolomite precipitation (Brady et al., 1996). Other factors favorable to dolomite formation include microbially induced local supersaturation of Ca²⁺ and Mg²⁺ and the presence of carboxylated organic matter (Petrash et al., 2017). The dolomite in the Kirkby Thore gypsum veins generally does not contain inclusions of pyrite, but there is one possible exception (Fig. 3D), which implies that dolomite may have precipitated (at least somewhat) later than the pyrite. In principle this could have occurred at any time in the later alteration history of the veins.

Sulfide minerals can also form abiotically via thermochemical sulfate reduction (TSR) at temperatures of ∼100–140°C and above (Machel, 2001; Watanabe et al., 2009; Meshoulam et al., 2016). Given that they most likely formed by anhydrite rehydration during uplift, the Kirkby Thore veins have probably not experienced such high temperatures, and our spatially resolved sulfur isotope measurements confirm that the pyrite is microbial in origin. MSR is associated with a kinetic sulfur isotope fractionation such that sulfur-34 is preferentially excluded from the reduced (i.e., sulfide) product. Where the supply of sulfate is not limiting, this leads to the formation of sulfides with markedly higher ³²S/³⁴S ratios than the original sulfate (usually expressed relative to a standard using the δ³⁴S notation; see the Materials and Methods section). This effect can lead to sulfide δ³⁴S values negatively offset by ∼20–45‰ (and occasionally more) compared to the original sulfate (Machel, 2001; Sim et al., 2011). TSR is also isotopically selective but induces a maximum sulfur isotope fractionation of only about 20‰ at ∼100°C (Thomazo et al., 2009; Watanabe et al., 2009; Meshoulam et al., 2016). Thus, highly negative δ³⁴S values in sulfide minerals, especially where demonstrably >20‰ more negative than the precursor sulfate, are commonly considered diagnostic of MSR, that is, a biosignature (e.g., Thomazo et al., 2009). The Kirkby Thore pyrite yielded sulfur isotope values (δ³⁴S_VCDT) consistently around -30‰. These compare with δ³⁴S_VCDT values for the host gypsum veins of +7.6‰, +7.5‰, and +10.4‰ (mean 8.5‰), which were measured on samples of the same gypsum vein for a previous study (Armstrong et al., 2020; our Fig. 6) and are broadly in line with the expected sulfur isotope composition of Permian marine evaporites (Fig. 6; Claypool et al., 1980). Armstrong et al. (2020) reported similar values (δ³⁴S_VCDT = +8.8‰ and +9.6‰) from coeval veins at the Barrowmouth gypsum mine about 70 km west of Kirkby Thore, suggesting that regional variation is small. The sulfide–sulfate offset of about -38‰ at Kirkby Thore is incompatible with TSR and is interpreted here as a robust signature of ancient microbial activity. By contrast, sulfide veins of known hydrothermal origin in Northwest England have published δ³⁴S_VCDT values mostly between 0 and +10 (Fig. 6), much heavier than the pyrite analyzed here and consistent with TSR (Solomon et al., 1971).

OPEN IN VIEWER

FIG. 6.

Box plot showing the sulfur isotope composition of pyrite and gypsum in the veins from Kirkby Thore, Cumbria. The gypsum data are reported in Armstrong et al. (2020). Sulfide mineral veins of known hydrothermal origin in the same region of England are shown for comparison (Solomon et al., 1971). Typical values for late Permian marine sulfate deposits are also indicated (Claypool et al., 1980). The lower panel illustrates the magnitude of the offset between gypsum and pyrite in the Kirkby Thore veins (bold arrow) compared to magnitudes known to be achievable by biotic and abiotic sulfate reduction (dotted arrows).

An apparent association between the pyrite and preserved carbonaceous material adds further weight to our interpretation of the sulfide as biotic in origin. Pyrite-rich opaque regions within the margins of the gypsum veins showed broad Raman bands at c. 1350 and c. 1600 cm⁻¹ in addition to other bands attributable to pyrite and gypsum (Fig. 4). We identify the feature at 1350 cm⁻¹ as the D band characteristic of disordered carbonaceous material, including organic matter (Beyssac et al., 2002). The feature at about 1600 cm⁻¹ peaked between 1580 and 1620 cm⁻¹ (e.g., Fig. 4C) and is likely to represent the graphitic G band produced by polyaromatic carbon (Beyssac et al., 2002), with a variable contribution from the D2 band. The carbonaceous material indicated by the Raman data is likely to be of ancient biological origin and appears to have been preserved in close association with the pyrite.

4.2. Comparison with calcium sulfate veins on Mars

Calcium sulfate veins observed crosscutting fine-grained siliciclastic rocks on Mars have important contextual, textural, and inferred genetic features in common with those from Kirkby Thore. For example, veins encountered by MER Opportunity in Endeavour Crater appear to be antitaxial and composed of fibrous gypsum, with fibers oriented normal to the vein walls (Squyres et al., 2012; Arvidson et al., 2014). White veins of calcium sulfate—anhydrite, bassanite, and possibly gypsum—also crosscut hundreds of meters of fluviolacustrine/eolian stratigraphy traversed by MSL Curiosity in Gale Crater, including the Yellowknife Bay and Murray formations (Nachon et al., 2014; Vaniman et al., 2014; Kronyak et al., 2019; Minitti et al., 2019). Most of these veins are oriented perpendicular to bedding (like those at Endeavour Crater), but at higher stratigraphic levels bedding-parallel and subparallel vein sets also occur (Nachon et al., 2014; Fedo et al., 2018; Kronyak et al., 2019). Like those in Kirkby Thore, the veins in the Yellowknife Bay formation are thought to postdate lithification and probably formed at depths of at least 1 km in rocks that underwent fairly rapid exhumation (Caswell and Milliken, 2017). In the stratigraphically higher Murray Formation, diverse mutually crosscutting vein sets record multiple generations of fracture opening and subsurface fluid circulation through a long window of geological time (Kronyak et al., 2019). Many of these veins are seemingly fibrous and are interpreted as generally syntaxial rather than antitaxial, although some appear to show entrained fragments of older material at boundaries between consecutive generations of sulfate, an indication of crack-seal growth shared by veins from Kirkby Thore (Fig. 7).

OPEN IN VIEWER

FIG. 7.

Photograph of the “Palmwag” calcium sulfate vein in the Murray formation in Gale Crater. This image was taken by the Mars Hand Lens Imager (MAHLI) aboard NASA's rover Curiosity on Sol 1275 of the MSL mission (Hurowitz et al., 2017). In common with gypsum veins from Kirkby Thore, England, this vein displays an apparent crystallographic preferred orientation normal to the vein walls, a variably distinct and asymmetric median line, and entrained fragments of older material. MAHLI image 1275MH0005550010404415C00. Scale bar (lower left) = 2 cm (after Kronyak et al., 2019).

4.3. Implications for biosignature detection on Mars

Discoveries in Earth's deep biosphere and its fossil record have inspired widespread speculation that life on Mars could also have colonized the subsurface, perhaps surviving there for billions of years after the surface became inhospitable (e.g., Boston et al., 1992; Fisk and Giovannoni 1999; Cockell, 2014). This line of research has been encouraged by multiple detections of methane as a trace gas (at ppbv level) in the martian atmosphere, where it is strongly out of equilibrium (Krasnopolsky et al., 2004; Mumma et al., 2009; Webster et al., 2015; Giuranna et al., 2019). It is unknown whether this methane is biological in origin, but methane-producing microorganisms are known to be globally abundant in Earth's subsurface. However, because Mars is a predominantly basaltic planet, studies of terrestrial analogs to a possible martian deep biosphere have mostly focused on porous basalt, where subsurface habitability could be maintained by water–rock reactions (serpentinization) that liberate hydrogen and thereby fuel both biotic and abiotic methanogenesis (e.g., McKinley et al., 2000; Ivarsson et al., 2008; McMahon et al., 2013). Although there is abundant evidence for a sediment- and sedimentary-rock-hosted deep biosphere on Earth, only a few studies have reported signatures of ancient subsurface life in sedimentary rocks or considered them relevant to Mars (Kretzschmar, 1982; Trewin and Knoll, 1999; Parnell et al., 2013; McMahon et al., 2018b). Our results imply that the porous margins of calcium sulfate veins in the subsurface can serve as conduits for the flow of sulfate-rich groundwater and therefore as potential habitats for prokaryotes capable of utilizing this sulfate to oxidize organic carbon (including methane) and/or molecular hydrogen. Such habitats may disappear and reappear several times over long spans of geological time as veins reopen under changing stress regimes and re-seal as further sulfate precipitates. Although in our samples MSR was probably stimulated by the organic content of the host rock, other sources of carbon and electrons may be available in organic-poor rocks. On Mars, pulses of groundwater circulation through sedimentary rocks could introduce methane and hydrogen produced by deeper serpentinization, and could perhaps also facilitate the production of hydrogen in situ through radiolysis or fracture-driven processes (e.g., McMahon et al., 2016; Dzaugis et al., 2018). Lithoautotrophic methanogens are found in sedimentary as well as basaltic rocks on Earth (e.g., Kimura et al., 2005). Thus, MSR coupled to hydrogen or methane oxidation could potentially have occurred in sulfate-bearing fracture or vein-margin fluids on Mars even if the host rock was relatively organic-poor, although organic-rich rocks are clearly preferable as sampling targets.

The veins from Kirkby Thore are younger than 0.25 Ga; by the time they formed, habitable conditions on Mars were already severely restricted, perhaps to depths of several kilometers on average; any early biosphere may have been extinguished by this time. However, the well-preserved calcium sulfate mineral veins observed on Mars formed close to 3.5 Ga ago when habitable conditions were probably widespread much nearer the surface (Westall et al., 2013; Grotzinger et al., 2014). If the vein-hosted martian habitats we envisage were actually inhabited, could they have produced durable, detectable biosignatures analogous to those described in the present study? Encouragingly, iron sulfide minerals have now been detected on Mars, although they are not yet well characterized. The Mössbauer instrument aboard MER Spirit made a tentative detection of FeS₂ (as either pyrite or marcasite) in a hydrothermally altered basaltic rock at the Home Plate site in Gusev Crater (Morris et al., 2008). More recently, and more significantly for the present study, the CheMin XRD instrument aboard MSL Curiosity has suggested the presence of both pyrrhotite (Fe_(1-x)S where 0 ≤ x ≤ 0.17) and pyrite in the sulfate-bearing Sheepbed mudstone (Yellowknife Bay formation) in Gale Crater; pyrrhotite was detected at 1.0 wt % in the “Cumberland” and “John Klein” drill samples, and pyrite was found at ∼0.3% (close to detection limits) in the latter sample, which is crosscut by hairline calcium sulfate veins (Vaniman et al., 2014). The sources of these sulfide minerals are unknown and may include igneous detritus and meteoritic material as well as sulfate reduction in the crater lake or its underlying sediment column (McAdam et al., 2014; Franz et al., 2017). Importantly, the Sheepbed mudstone is not pervasively oxidized and also preserves ancient organic matter (Ming et al., 2014; Freissinet et al., 2015). Thus, despite mineralogical evidence for sulfide oxidation in Gale Crater and elsewhere on Mars (i.e., the presence of jarosite; Zolotov and Shock, 2005; Rampe et al., 2017) and widespread oxidation of the martian surface in general, a small quantity of pyrite has survived locally alongside calcium sulfate and organic matter in at least some martian mudstones. The results of the present study recommend such materials for further in situ analysis and indeed return to Earth should they be encountered by future missions. We would additionally recommend that veins should be sampled where still encased in their host rocks rather than where they form free-standing “raised ridges,” since the vein–wall-rock interface is of key astrobiological interest. Drilling below the surface or sampling recently exposed materials may improve the probability of obtaining unoxidized martian sulfides in future missions.

On Earth, the δ³⁴S biosignature has been identified in rocks as old as 3.47 Ga, although it did not become widespread until after 2.5 Ga (Shen et al., 2009; Thomazo et al., 2009). It must be conceded that sulfur isotope systematics are not well understood in martian contexts, and the isotopic fingerprints of any indigenous martian organisms may well have differed from life on Earth. Nevertheless, the Curiosity rover has already detected isotopically light sulfide (by evolved gas analysis) in Gale Crater sediments, where it could ultimately have originated either from abiotic or from biotic processes (Franz et al., 2017). The best available abiotic interpretation invokes the slow equilibration between aqueous sulfide and sulfate in an impact-driven subsurface hydrothermal system followed by the upward migration of fluids and the precipitation of sulfide minerals in the overlying sediment (Franz et al., 2017). Such a process would induce ³⁴S-depleted sulfide mineralization throughout the accessible porosity of the host rock and not exclusively within microenvironments localized to the margins of preexisting calcium sulfate veins, as seen in our samples. Hence, multiscale analysis of the distribution of martian pyrite through its host rock could be essential for determining its origin. The analysis of ³³S and ³⁶S alongside δ³⁴S could also constrain the biogenicity of martian sulfide and its relationship to adjacent sulfate minerals versus exogenous fluids (Ueno et al., 2008); these signatures would also discriminate between sulfide produced by sulfate reduction and that produced by sulfur disproportionation (Johnston et al., 2005; Ueno et al., 2008). Finally, evidence for ancient biological activity may also be obtained through the analysis of carbonaceous material associated with metal sulfides, as in the present study; both the Perseverance and Rosalind Franklin rovers will be equipped with Raman laser spectrometers as well as other instruments to investigate organic matter.