Meridiani Planum hematite deposit and the search for evidence of life on Mars—iron mineralization of microorganisms in rock varnish

Icarus 171 (2004) 20–30 www.elsevier.com/locate/icarus

Meridiani Planum hematite deposit and the search for evidence of life on Mars—iron mineralization of microorganisms in rock varnish Carlton C. Allen a,∗ , Luke W. Probst b , Beverly E. Flood c , Teresa G. Longazo d , Rachel T. Schelble c , Frances Westall e a NASA Johnson Space Center, Houston, TX 77058, USA b Rice University, Houston, TX 77005, USA c University of Southern California, Los Angeles, CA 90089, USA d University of Arizona, Tucson, AZ 85721, USA e Centre de Biophysique Moleculaire, Orleans 45071 cedex 2, France

Received 18 June 2003; revised 26 January 2004 Available online 19 June 2004

Abstract The extensive hematite deposit in Meridiani Planum was selected as the landing site for the Mars Exploration Rover Opportunity because the site may have been favorable to the preservation of evidence of possible prebiotic or biotic processes. One of the proposed mechanisms for formation of this deposit involves surface weathering and coatings, exemplified on Earth by rock varnish. Microbial life, including microcolonial fungi and bacteria, is documented in rock varnish matrices from the southwestern United States and Australia. Limited evidence of this life is preserved as cells and cell molds mineralized by iron oxides and hydroxides, as well as by manganese oxides. Such mineralization of microbial cells has previously been demonstrated experimentally and documented in banded iron formations, hot spring deposits, and ferricrete soils. These types of deposits are examples of the four “water–rock interaction” scenarios proposed for formation of the hematite deposit on Mars. The instrument suite on Opportunity has the capability to distinguish among these proposed formation scenarios and, possibly, to detect traces that are suggestive of preserved martian microbiota. However, the confirmation of microfossils or preserved biosignatures will likely require the return of samples to terrestrial laboratories. Published by Elsevier Inc. Keywords: Rock varnish; Mars; Exobiology; Hematite; Meridiani Planum; Microfossil

1. Introduction The hematite deposit in Meridiani Planum is the landing site for Opportunity, one of the two Mars Exploration Rover (MER) spacecraft (Fig. 1). This site was chosen because it shows “strong evidence for surface processes involving water and appear(s) capable of addressing the science objectives of the missions, which are to determine the aqueous, climatic, and geologic history of sites on Mars where conditions may have been favorable to the preservation of evidence of possible prebiotic or biotic processes” (Golombek et al., 2003). We are investigating the potential for hematite, as well as other iron oxide and hydroxide min* Corresponding author. Fax: 281-483-5347.

E-mail address: [email protected] (C.C. Allen). 0019-1035/$ – see front matter Published by Elsevier Inc. doi:10.1016/j.icarus.2004.04.015

erals, to preserve microfossils and physical biomarkers— actual evidence for life on Mars. Christensen et al. (2000), using data from the Mars Global Surveyor Thermal Emission Spectrometer (TES), identified gray crystalline hematite [α-Fe2 O3 ] in a 350 km by 750 km region near Meridiani Planum. This deposit corresponds closely to the low-albedo highlands unit “sm,” mapped as a wind-eroded, ancient, subaqueous sedimentary deposit (Edgett and Parker, 1997). Christensen et al. (2001) interpreted the deposit to be “an in-place, rock-stratigraphic sedimentary unit characterized by smooth, friable layers composed primarily of basaltic sediments with approximately 10 to 15% crystalline gray hematite.” Christensen et al. (2000) discussed five possible scenarios for the formation of the martian hematite deposit: (1) direct precipitation from standing, oxygenated, iron-rich water; (2) precipitation from iron-rich hydrothermal fluids;

Rock varnish and the Meridiani Planum hematite deposit

(3) low-temperature dissolution and precipitation through mobile groundwater leaching; (4) surface weathering and coatings; (5) thermal oxidation of magnetite-rich lavas. The first four of these scenarios involve the interactions of rock with water, and thus have implications in the search for evidence of martian life. Christensen et al. (2001) assessed the scenarios for forming the hematite deposits and argued for chemical precipitation from aqueous fluids, under either ambient or hydrothermal conditions. Newsom et al. (2003) reported a chain of paleolake basins and associated layered deposits along the margin of the hematite region. Hynek et al. (2002) suggested that the hematite may have been formed by precipitation from circulating fluids within layered volcanic materials. Catling and Moore (2003) favored a hydrothermally charged aquifer as the original setting for developing coarse-grained crystalline hematite. They also recognized that hematite with the observed crystallographic signature could be produced from sedimentary deposits subsequently buried to a depth of several kilometers, in agreement with the formation hypothesis described by Lane et al. (2002). Iron oxide and hydroxide minerals, including hematite, can mineralize and preserve microfossils and physical biomarkers. Previous research by ourselves and others, summarized below, has demonstrated such mineralization and preservation in deposits from three of the four water–rock interaction scenarios listed by Christensen et al. (2000). The current study is focused on the fourth scenario— mineralization of microorganisms in iron-rich surface weathering and coatings, using the specific example of rock varnish from desert sites in the United States and Australia. This work also addresses the capabilities and limitations the MER spacecraft in the search for evidence of life at Meridiani Planum.

2. Iron mineralization of microorganisms and biosignatures in rock varnish 2.1. Previous studies Rock varnish, also known as desert varnish, is a dark, hard coating that forms on rocks in many arid environments (Dorn, 1998). Varnish coatings can develop on exposed rock surfaces in a matter of decades, and noticeably darken on a time scale of centuries (Dorn, 1991; Krinsley et al., 1990). Rock varnish is a complex combination of clays with iron and manganese oxides and hydroxides. The dominant clay minerals are illite, smectite, interstratified illitesmectite, and chlorite (Potter and Rossman, 1977). Hematite is the major iron oxide phase in rock varnish, and birnessite [Na4 Mn14 O27 ·9H2 O] is the major manganese oxide phase (Potter and Rossman, 1979; McKeown and Post, 2001). Many rock varnish samples contain living microorganisms. The most common rock varnish inhabitants are slowgrowing, melanin-pigmented microcolonial fungi along

21

with typical soil-inhabiting actinomycetes and nonmotile endospore-forming gram-positive cocci (Krinsley and Dorn, 1991; Nagy et al., 1991; Staley et al., 1992; Sterflinger, 2000; Gorbushina et al., 2002; Gorbushina, 2003). Dozens of culturable strains of varnish microorganisms oxidize manganese and/or iron. These microorganisms include members of the bacterial genera Micrococcus, Arthrobacter, Bacillis, and the actinomycetes Geodermatophilis (Hungate et al., 1987; Adams et al., 1992; Staley et al., 1992). Several studies have indicated limited mineralization and preservation of the remains of microbial life in rock varnish. Mineralized microcolonial fungi have been recognized in varnish layers (Taylor-George et al., 1983; Dragovich, 1993; Gorbushina et al., 2002; Gorbushina, 2003). Nagy et al. (1991) reported textures resembling “microstromatolites” within varnish layers. Dorn (1991) found botryoidal structures in varnish samples. He interpreted the structures as bacterial colonies, which served as nucleation centers during varnish formation. Krinsley (1998) and Krinsley and Rusk (2000) used high-resolution TEM techniques to study varnish layers on rocks from deserts in California, Hawaii, Peru, and Antarctica. They reported micrometer-scale coccoidal and granular structures within the varnish that had much higher concentrations of iron and manganese than the surrounding matrix. They suggested that these structures were bacterial casts, hyphae, buds, or bacterial precipitates. 2.2. Samples 2.2.1. Sonoran Desert Samples of varnish coatings on granite were collected for the present study from a road cut off Interstate Highway 17 just north of Camp Verde, AZ. Additional samples were collected from outcrops at South Mountain Park in Phoenix, and from an outcrop in a wash just east of Gates Pass adjacent to the Tucson Mountain Park. Initial results were reported by Probst et al. (2002). 2.2.2. Pilbara We also examined varnish coatings on foliated rocks of granitoid composition from the Pilbara region of Western Australia. These rocks have been exposed to weathering since the Permian glaciation, approximately 280 myr ago, though the age of the varnish layers is unknown. Initial results were reported by Flood et al. (2003). 2.3. Laboratory methods The Sonoran Desert samples were collected with flamesterilized tools and immediately placed in sterile plastic bags for transport. The Pilbara samples were collected and transported in a non-sterile manner. All samples were preserved in a laboratory desiccator at room temperature. Subsamples were prepared and mounted in open-face hoods to minimize surface contamination. The varnish samples were initially examined in reflected light using a Nikon ME 600 binocular microscope with a

22

C.C. Allen et al. / Icarus 171 (2004) 20–30

Fig. 1. Meridiani Planum landing site viewed by the Mars Exploration Rover Opportunity. NASA image.

digital imaging system. Based on this examination, subsamples were selected for additional study by scanning electron microscopy (SEM) and transmission electron microscopy (TEM). Exterior surfaces and interior regions of Sonoran Desert varnish samples were mounted on SEM stubs and coated with 5 to 10 nm of conductive platinum. These samples were examined with a JEOL 6340F SEM and an IXRF (Gresham Scientific Instruments, Ltd.) light element energy dispersive X-ray spectrometer (EDS). Other subsamples were examined uncoated in an FEI XL-30 environmental SEM (ESEM), operated at a pressure of 1.7 torr. Several varnish chips, approximately 20 µm in size, were prepared for TEM study. These were embedded in epoxy, ultramicrotomed, and examined with a JEOL 2000 FX TEM equipped with a dedicated Link EDS system using a windowless detector. Chips of the Pilbara samples were mounted on SEM stubs using carbon paste and coated with 100 Å of platinum. Preliminary electron microscope investigations of these samples were conducted utilizing a JEOL 5910 SEM equipped with IXRX EDS. The primary investigation utilized the JEOL 6340F SEM and IXRF light element EDS described above.

Fig. 2. Pilbara sample cross section showing varnish layer (black), alteration rind (orange), and underlying rock (white); reflected light optical micrograph.

layer coated heavily weathered, orange rinds 200 to 300 µm thick. 2.5. Composition and mineralogy

2.4. Structure 2.4.1. Sonoran Desert Varnish coatings on the Sonoran Desert samples were uniformly hard and dark, similar to material from the same area previously studied by Allen (1978) and Nagy et al. (1991). Characteristic varnish thicknesses ranged from 50 to 100 µm. 2.4.2. Pilbara The dark, smooth varnish from the Pilbara samples was lamellate and typically 75 to 150 µm thick (Fig. 2). The dark

2.5.1. Sonoran Desert The predominant clay mineral in the Sonoran Desert varnish samples was illite [KAl2 (Si3 AlO10 )(OH)2 ], identified by EDS elemental ratios and a regular basal spacing of approximately 1.0 nm in high-resolution TEM images. The other dominant mineral was a sheet-like manganese oxide phase containing trace concentrations of barium, likely birnessite (approximately 0.6 to 0.7 nm basal spacing). Birnessite is known to accommodate water and readily undergo cation exchange reactions to accommodate a variety of large cations of elements including potassium, sodium, calcium,

Rock varnish and the Meridiani Planum hematite deposit

Fig. 3. Sonoran Desert varnish layer showing lath-shaped illite; SEM secondary electron micrograph.

Fig. 4. Hyphae on the surface of Sonoran Desert varnish layer; SEM back-scattered electron micrograph.

and barium (Post, 1999). Illite occurred as large regions of well-crystallized laths (Fig. 3) and as individual packets mixed with the fine-grained, poorly-crystalline, birnessitelike phase. Discrete anhydrous mineral grains, tentatively identified by their EDS spectra as hematite, ilmenite, quartz and rutile, were embedded in the clay minerals. 2.5.2. Pilbara The surface layer consisted mainly of clays with low concentrations of potassium and iron. Platy iron-rich minerals and discrete sub-micrometer manganese-rich minerals were distributed throughout the clay matrix. 2.6. Microbiology 2.6.1. Sonoran Desert The Sonoran Desert samples were partially coated with black patches that corresponded in scale and morphology to the microcolonial fungi identified by Palmer et al. (1986). Fungal hyphae were common features of these samples. Hyphae were visible both on the surface (Fig. 4) and within the varnish layers. The ESEM images show rod-shaped objects, approximately 0.5 to 2 µm in length, located within thin continuous

23

Fig. 5. Bacterial biofilm with cells (arrows) in Sonoran Desert varnish; ESEM secondary electron micrograph.

layers incorporated in the varnish coatings (Fig. 5). These are interpreted as examples of the bacterial biofilms, i.e., conglomerations of individual cells and water-rich extracellular polysaccharide (EPS) produced by diverse microorganisms (Costerton et al., 1994). The morphologies of individual cells and EPS layers were preserved from desiccation and deformation by the 1.7 torr pressure in the ESEM sample chamber. Cells were also recognized in SEM images, though they were characteristically dehydrated and possibly contorted by the high vacuum of the microscope chamber (Fig. 6a). An EDS spectrum of one group of cells (Fig. 6b) indicates high concentrations of carbon and oxygen, along with minor sulfur, indicative of cellular material. The spectrum also contains peaks corresponding to aluminum, silicon, and potassium from the underlying illite, as well as manganese, iron, and minor barium from the birnessite. 2.6.2. Pilbara The Pilbara samples hosted multiple species of black microcolonial fungi. Some species appeared to be specific to the varnish, while the other species inhabited only the non-varnished substrate. These fungi apparently contributed to the weathering of the non-varnish substrate by creating and inhabiting large micropits. Gorbushina et al. (2002) and Gorbushina (2003) also reported widespread microcolonial fungi on varnished quartzite surfaces from Australia and Namibia. While fungi were common, confirmation of bacterial presence within the Pilbara samples was rare. One cell, possibly deformed by the SEM vacuum, is shown in Fig. 7a. The EDS spectrum of this feature (Fig. 7b) indicates a high concentration of carbon and oxygen, along with minor sulfur, in the cell. The aluminum, silicon, manganese, and iron peaks are attributed to the varnish matrix. 2.7. Mineralization 2.7.1. Sonoran Desert Partial mineralization of a fraction of the fungal population is apparent on the surfaces and within these samples.

24

C.C. Allen et al. / Icarus 171 (2004) 20–30

(a) Fig. 8. Hymenium of microcolonial fungi in Sonoran Desert varnish sample; SEM secondary electron micrograph.

(b) Fig. 6. (a) Deformed cells in Sonoran Desert varnish; SEM secondary electron micrograph. (b) EDS spectra of cells and substrate; Pt peak from conductive coating.

(a)

(b) Fig. 7. (a) Deformed cell in Pilbara varnish; SEM secondary electron micrograph. (b) EDS spectra of cell and substrate; Pt peak from conductive coating.

Figure 8 shows a portion of the hymenium (the layer of cells originally containing the spore-bearing cells) of one microcolonial fungal body. The hymenium is partially coated with interlocking platy minerals, characteristically 3 to 5 µm in width. This relationship closely matches the mineralized fungal colonies documented in rock varnish by Gorbushina et al. (2002). Groups of rod-shaped cavities, each 1 to 2 µm in length, are occasionally observed in various states of degradation and mineralization within the varnish layers (Figs. 9a, 10a). These objects, similar in scale and morphology to the carbon-rich cells found in these samples (Fig. 6a), are interpreted as mineralized cell molds. EDS spectra of these molds (Figs. 9b, 10b) demonstrate that most of the organic material, composed chiefly of carbon, oxygen, and sulfur, has been lost. The spectra include different ratios of the major elements corresponding to illite and iron and manganese oxides. 2.7.2. Pilbara These samples contained small numbers of partiallymineralized microcolonial fungi, very similar to those documented in the Sonoran Desert material. Figure 11a shows a portion of the hymenium from one microcolonial fungal body, comparable in size and morphology to the more extensive hymenium in Fig. 8. EDS element mapping of the Pilbara material (Fig. 11b) demonstrates that the hymenium was partially mineralized by platy iron-rich minerals 2 to 3 µm across, as well as finer-scale manganese-rich minerals. The samples also exhibited elevated concentrations of manganese and/or iron in discrete nodules, typically one micrometer or less in length, either within the matrix of the varnish or loosely bound to detrital grains. Some of the nodules had higher carbon contents than the surrounding varnish, suggesting that they may have been mineralized cells or cell fragments. These nodules resembled the mineralized coccoidal and granular structures reported by Krinsley (1998) and Krinsley and Rusk (2000). One completely intact bacterial cast, containing a significant concentration of carbon, was documented by SEM.

Rock varnish and the Meridiani Planum hematite deposit

25

(a) (a)

(b) Fig. 9. (a) Mineralized cells in Sonoran Desert varnish; SEM back-scattered electron micrograph. (b) EDS spectra of mineralized cells and substrate; Pt peak from conductive coating.

(b) Fig. 11. (a) Mineralized fungal hymenium in Pilbara varnish; SEM secondary electron micrograph. (b) EDS map overlay on SEM secondary electron micrograph showing concentrations of iron (green) and manganese (purple) on and around fungal hymenium in Pilbara varnish.

2.8. Abundance and preservation

(a)

(b) Fig. 10. (a) Mineralized cells in Sonoran Desert varnish; SEM secondary electron micrograph. (b) EDS spectra of mineralized cells and substrate; Pt peak from conductive coating.

Microanalysis of rock varnish samples from widelyseparated locations demonstrates similar patterns of mineralization. Microcolonial fungi become coated with iron- and manganese-minerals. Bacterial forms can be preserved as iron- and manganese rich casts of the individual cells. These patterns have been reported by previous authors, in samples from desert regions in the United States, Peru, Australia, and Antarctica. Rock varnish may therefore be a suitable medium for the preservation of biosignatures indicative of microbial life. However, it is not yet clear that such preservation is common, nor that biosignatures are preserved in rock varnish over geologic time. While semi-intact microcolonial fungi were common in varnish samples from the present study, mineralization of fungi and microbial cells proved to be rare. Many hours of SEM traverses yielded only a few clear examples of such mineralization. Krinsley (1998), in a comprehensive TEM study of varnish samples from four locations, similarly

26

C.C. Allen et al. / Icarus 171 (2004) 20–30

found very few objects that could confidently be interpreted as preserved cells, and no preserved fungal bodies. Mineralization of microorganisms on rock surfaces in arid environments may be rare, relative to the total bioloads typical of the varnish layers. Varnish samples hundreds to thousands of years old, dated by archaeological methods, are common (Dorn, 1991; Krinsley et al., 1990). Varnish has been shown to persist in arid-alkaline environments for periods of 100,000 years (Liu, 1994). However, varnish can be destroyed by a variety of mechanisms, including biogeochemical leaching and aeolian erosion (Krinsley and Dorn, 1991; Krinsley et al., 1990). Very little evidence of rock varnish is apparent in the sedimentary record (Krinsley, 1998). The potential for long-term preservation of evidence of life in rock varnish may thus be limited by erosion of the thin varnish layers.

3. Iron mineralization of microorganisms and biosignatures—other Mars-relevant scenarios Christensen et al. (2000) proposed three other scenarios, besides surface weathering, by which water–rock interactions could have produced the martian hematite deposits. These include: (1) direct precipitation from standing, oxygenated, iron-rich water; (2) precipitation from iron-rich hydrothermal fluids; (3) low-temperature dissolution and precipitation through mobile groundwater leaching. Previous research, summarized below, has demonstrated that iron oxide and hydroxide minerals, including hematite, can mineralize and preserve evidence of microorganisms in controlled experiments as well as in natural examples from each of these proposed scenarios. 3.1. Experiments Laboratory and field experiments have demonstrated that iron oxides and hydroxides can mineralize living cells, as well as the EPS that binds cells into biofilms. Ferrihydrite [Fe5 O7 (OH)·4H2 O] has been shown to mineralize specific types of microbial cells (Thomas-Keprta et al., 1998). Cell walls were completely coated with this mineral on a time scale of weeks, in experimental microcosms utilizing basalt and groundwater. Allen et al. (2000) studied the formation of iron-mineralized, amorphous coatings in hot spring deposits at 60 ◦ C. In situ experiments demonstrated that the coatings were in fact biofilms, and that their formation was essentially prevented by passing the water through 0.2 µm filters, thus excluding microorganisms. 3.2. Direct precipitation from standing, oxygenated, iron-rich water. Example: banded iron formations Banded iron formations (BIFs), as defined by Klein and Beukes (1992), are “chemical sediments, typically thinlybedded or laminated, whose principal chemical characteris-

tic is an anomalously high content of iron, commonly but not necessarily containing layers of chert.” BIFs, which range in age from 3.8 to 0.8 Ga, are among the most distinctive sedimentary deposits of the Precambrian. BIFs are composed principally of hematite and magnetite [Fe3 O4 ] interbedded with chert. Investigations of many BIFs have revealed evidence of ancient life. For example, stromatolitic sections of the Gunflint Iron Formation, which spans the US/Canada boundary through Minnesota and Ontario, contain a variety of microorganisms that have been preserved in chert for approximately 1.8 Ga (Barghoorn and Tyler, 1965; Awramik and Barghoorn, 1977; Strother and Tobin, 1987). Barghoorn and Tyler (1965) originally classified the Gunflint microfossils, and documented numerous filamentous and coccoidal forms. With the aid of an electron microprobe, Tazaki et al. (1992) documented trace quantities of iron within silicified microorganisms from the Gunflint. Allen et al. (2001) investigated microfossils mineralized by iron oxides in samples from the Gunflint Iron Formation. This initial study has been continued in more detail by Schelble et al. (2004). Filaments and coccoids were documented by analytical SEM in samples etched with hydrofluoric acid vapor. Many of the coccoids and filaments contained elevated concentrations of iron and carbon, compared to the surrounding matrix material. These structures were interpreted as microfossils mineralized and preserved by iron oxides, though the specific mineral or minerals could not be identified. DeGregorio and Sharp (2002) used TEM electron diffraction analysis to demonstrate that some filamentous microfossils in the Gunflint Formation were mineralized by hematite. 3.3. Precipitation from iron-rich hydrothermal fluids. Example: hot spring deposits Hematite is one of several iron oxide and hydroxide minerals that can precipitate when large volumes of hydrothermal fluids move through rocks. Terrestrial hydrothermal deposits range from simple systems (predominantly calcite, silica, or iron oxides) to complex assemblages with mineralogies reflecting local changes in temperature, pH, and fluid composition (Guilbert and Park, 1986). A study by Wade et al. (1999) of iron-depositing hot springs in Colorado and Yellowstone National Park (48 to 55 ◦ C) showed that iron was partitioned among hematite, ferrihydrite, goethite [α-FeO·OH], siderite [FeCO3 ], and nontronite [Na0.3 Fe2 (Si,Al)4 O10 (OH)2 ·nH2 O]. Microorganisms are abundant in many hot springs, and are often preserved by mineralization. They are most commonly preserved in hydrothermal systems dominated by silica (Cady and Farmer, 1996; Cady et al., 2003), sometimes in association with iron-bearing minerals. Wade et al. (1999) noted the remains of rod-shaped bacteria and dehydrated biofilm in a silica-dominated iron hot spring that precipitated hematite and other iron oxides and carbonates. Ferris et al.

Rock varnish and the Meridiani Planum hematite deposit

(1986, 1988, 1989a) described the fossilization of bacteria by iron and silicon oxide crystallites in a geothermal setting. Microorganisms can also be preserved in a variety of hot spring deposits dominated by iron minerals (Allen et al., 2000; Pierson and Parenteau, 2000). Chafetz et al. (1998) reported recognizable bacterial remains in iron and manganese oxide deposits within hot spring travertines in Morocco. These microfossils were found in centimeter-scale black dendrites. Filamentous and coccoidal bacteria were observed densely packed within the dendrites, but no microfossils were found in the enclosing aragonite and calcite laminae. 3.4. Low-temperature dissolution and precipitation through mobile groundwater leaching. Example: ferricrete soils Ferricretes are formed on Earth when acidic groundwaters extensively leach soil and rock. Mafic minerals are dissolved in low-pH, reducing environments and form colloidal oxides and hydroxides. These can be transported by water and redeposited as hydrated silica and iron oxide minerals (Guilbert and Park, 1986). Fortin et al. (1997) reported that Thiobacillus cells can act as substrates for the precipitation of iron oxides in acidic mine tailings. Ferris et al. (1989b) identified hematite, ferrihydrite, and goethite as the dominant iron minerals associated with bacteria in acidic environments. Furniss et al. (1999) noted rod-shaped structures, apparently bacteria mineralized by ferrihydrite and goethite, in both modern and ancient (8840 years b.p.) ferricrete samples from a mining district in Montana. Westall and Kirkland (2002) and Kirkland et al. (2002) studied a ferricrete sample from Western Australia using SEM and emission spectroscopy. The sample was collected from an extremely dry soil surface adjacent to Shark Bay. Filamentous microfossils, as well as abundant EPS, were present in the ferricrete sample. Some filaments appeared collapsed and split for portions of their lengths. They were interpreted as microbial filaments, bacterial or possibly fungal, embedded in films of EPS. Iron oxide precipitates engulfed the networks of filaments. The degree of iron oxide mineralization varied locally over a scale of micrometers. Mineralization apparently occurred over a long enough period to produce several generations of microfossils.

4. Meridiani Planum hematite deposit and the search for evidence of life on Mars 4.1. Martian hematite The large hematite deposit at Meridiani Planum is important for understanding many aspects of early martian history. This deposit represents the only place on the planet where a specific mineral has been identified and correlated with a mappable geologic unit. On Earth, hematite is formed by

27

several mechanisms, most of which involve water. The martian hematite may prove to be the only large-scale mineralogical evidence for water–rock interactions on early Mars. If the Meridiani Planum hematite was formed by a mechanism or mechanisms involving water, the deposit may be significant in the search for life on Mars. Liquid water is the one prerequisite accepted by most researchers for the existence of terrestrial life (e.g., Jakosky, 1998). On Earth, diverse microbiota have been reported in each of the four “water–rock” scenarios discussed by Christensen et al. (2000) for hematite deposition. The present study, as well as many others, provides evidence for the mineralization of microorganisms in each of these scenarios. 4.2. Exploring the Meridiani Planum hematite deposit Opportunity, one of NASA’s twin Mars Exploration Rovers, is exploring Meridiani Planum in 2004. Each rover carries six major science instruments (Squyres et al., 2003). Pancam, a mast-mounted stereo multispectral imager, is sensitive to visible and near-infrared (0.4 to 1.1 µm) wavelengths. The instantaneous field of view is 0.28 mrad per pixel, approximately the resolving power of the human eye (Bell et al., 2003). This camera is bore-sighted to Mini-TES, an infrared (5 to 29 µm) spectrometer that provides medium resolution (20 or 8 mrad per pixel) spectra that complement those from the orbiting TES on the Mars Global Surveyor spacecraft (Christensen et al., 2003). Thermal emission spectra can identify many rock-forming minerals, including silicates and oxides. Each MER rover also deploys an articulated arm carrying four instruments (Fig. 12). An Alpha Particle X-Ray Spectrometer (APXS) can determine abundances of major and minor rock-forming elements, including carbon (Rieder et al., 2003). The Mössbauer Spectrometer can determine iron valence states and thus help identify iron-bearing minerals in rocks and soils (Klingelhofer et al., 2003). The Microscopic Imager returns monochromatic reflected light images at a resolution of 30 µm per pixel (Herkenhoff et al., 2003). The Rock Abrasion Tool exposes interior surfaces of rocks for analysis by grinding to a depth of 5 mm (Gorevan et al., 2003). Terrestrial hematite deposits produced by various mechanisms differ significantly from one another in their mineralogies, textures, and spectra. Such differences are distinguishable with the combination of analytical capabilities on Opportunity. For example, the Pancam and Microscopic Imager can detect sedimentary layering by color and texture on the scale of centimeters to millimeters. The Mini-TES can distinguish among different iron oxides and can specifically identify hematite, in the same manner as its orbiting counterpart. The APXS and Mössbauer Spectrometer, when combined, can identify and quantify the iron-bearing phases in a rock or soil. Rock surfaces can be analyzed by the APXS, Mössbauer Spectrometer, and Microscopic Imager. The Rock Abrasion Tool can then grind into the rock, after which the analyses can be repeated. Comparison of the two

28

C.C. Allen et al. / Icarus 171 (2004) 20–30

Fig. 12. Mars Exploration Rover arm, carrying the Alpha Particle X-Ray Spectrometer, Mössbauer Spectrometer, Microscopic Imager, and Rock Abrasion Tool. NASA image.

sets of analyses allows characterization of surface alteration, including possible rock varnish layers. Such in situ observations may lead to a consensus on the origin of hematite on the martian surface. If the Meridiani Planum hematite deposit carries a record of martian life, traces that are suggestive of life may also be revealed by the instruments on Opportunity. The distinct colored banding characteristic of many banded iron formations (Klein and Beukes, 1992) should be obvious in MER Pancam images. The fossil record contains many examples of macroscopic fossil biofilms and microbial structures (Westall et al., 2000, 2001, 2004), observable at the resolution of the Microscopic Imager and in some cases the resolution of the Pancam. Many hot springs precipitate minerals that fossilize and preserve characteristically colored microbial mats, as well as masses of organisms many centimeters across (Allen et al., 2000; Chafetz et al., 1998). Mineralized microbes sometimes contain traces of carbon in abundances that could be detected by APXS analysis, though this instrument cannot distinguish between organic and inorganic carbon. The instruments on Opportunity, however, are not designed to provide unambiguous evidence of life, particularly at the microscopic scales characteristically observed in terrestrial iron oxide and hydroxide deposits. The mineralized microfossils and associated biofilms described above have typical dimensions of micrometers to tens of micrometers. Almost all such features are significantly smaller than the spatial resolution of any MER instrument. Most fossil microorganisms in terrestrial rocks are only detectable by optical and electron microscopy (Westall, 1999). The association of carbon with putative microfossils can only be confirmed using microbeam techniques.

Thus, while evidence suggestive of martian microbial life may be detected in situ, confirmation by direct fossil evidence will probably require the return of samples to terrestrial laboratories. A key function of the next generation of Mars landers will be to discover and certify prime sites for future sample return missions. The Meridiani Planum hematite deposit may well be among those prime sites.

5. Conclusions Microbial life, including microcolonial fungi and bacteria, is documented in rock varnish layers from desert regions of the American southwest and Australia. Analytical and environmental SEM data show that portions of the fungi, as well as molds of individual cells, are mineralized by iron oxides and hydroxides, as well as by manganese oxides. The number of microorganisms thus mineralized is small compared to the microbiota on and within the varnish coating these samples. Iron oxide and hydroxide mineralization of microbial cells has previously been demonstrated experimentally and documented in banded iron formations, hot spring deposits, and ferricrete soils. These deposits, along with rock varnish, are examples of the four “water–rock interaction” scenarios proposed for formation of the hematite deposits detected on Mars. The extensive Meridiani Planum deposit is the landing site for one of the twin Mars Exploration Rover spacecraft. The lander instrument suite may have the capability to distinguish among these proposed formation scenarios. In addition, the lander instruments may be able to detect spectral, morphological, or compositional evidence suggestive of preserved martian biosignatures. The confirmation of microfossils or

Rock varnish and the Meridiani Planum hematite deposit

preserved biosignatures, however, will likely require the return of samples to terrestrial laboratories.

Acknowledgments We gratefully acknowledge support from the NASA Astrobiology Institute through Principal Investigator David McKay (CA, FW, TL) and the Lunar and Planetary Institute Summer Intern Program (TL, RS, LP, BF). Craig Schwandt, Kathie Thomas-Keprta, and Susan Wentworth provided invaluable assistance in the electron microscope laboratories. Everett Gibson generously donated the Pilbara samples for this study. Thoughtful comments by David Krinsley significantly improved the manuscript.

References Adams, J.B., Palmer, F., Staley, J.T., 1992. Rock weathering in deserts: mobilization and concentration of ferric iron by microorganisms. Geomicrobiol. J. 10, 99–114. Allen, C.C., 1978. Desert varnish of the Sonoran Desert—optical and electron probe microanalysis. J. Geol. 86, 743–752. Allen, C.C., Albert, F.G., Chafetz, H.S., Combie, J., Graham, C.R., Kieft, T.L., Kivett, S.J., McKay, D.S., Steele, A., Taunton, A.E., Taylor, M.R., Thomas-Keprta, K.L., Westall, F., 2000. Microscopic physical biomarkers in carbonate hot springs: implications in the search for life on Mars. Icarus 147, 49–67. Allen, C.C., Westall, F., Schelble, R.T., 2001. Importance of a martian hematite site for astrobiology. Astrobiology 1, 111–123. Awramik, S.M., Barghoorn, E.S., 1977. The Gunflint microbiota. Precambrian Res. 5, 121–442. Barghoorn, E.S., Tyler, S.A., 1965. Microorganisms from the Gunflint Chert. Science 147, 563–577. Bell III, J.F., 24 colleagues, 2003. Mars Exploration Rover Athena Panoramic Camera. J. Geophys. Res. 108. 4-1–4-30. doi:10.1029/ 2003JE002070. Cady, S.L., Farmer, J.D., 1996. Fossilization processes in siliceous thermal springs: trends in preservation along thermal gradients. In: Bock, G.R., Goode, J.A. (Eds.), Evolution of Hydrothermal Systems on Earth (and Mars?). Wiley, New York, pp. 150–170. Cady, S.L., Farmer, J.D., Grotzinger, J.P., Schopf, J.W., Steele, A., 2003. Morphological biosignatures and the search for life on Mars. Astrobiology 3, 351–368. Catling, D.C., Moore, J.M., 2003. The nature of coarse-grained crystalline hematite and its implications for the early environment of Mars. Icarus 165, 277–300. Chafetz, H.S., Akdim, D., Julia, R., Reid, A., 1998. Manganese- and ironrich black travertine shrubs: bacterially (and nanobacterially) induced precipitates. J. Sediment. Res. 68, 404–412. Christensen, P.R., 15 colleagues, 2000. Detection of crystalline hematite mineralization on Mars by the Thermal Emission Spectrometer: evidence for near-surface water. J. Geophys. Res. 105, 9623–9642. Christensen, P.R., Morris, R.V., Lane, M.D., Bandfield, J.L., Malin, M.C., 2001. Global mapping of martian hematite mineral deposits: remnants of water-driven processes on early Mars. J. Geophys. Res. 106, 23873– 23885. Christensen, P.R., 19 colleagues, 2003. Miniature Thermal Emission Spectrometer for the Mars Exploration Rovers. J. Geophys. Res. 108. 5-1–523. doi:10.1029/2003JE002117. Costerton, J.W., Lewandowski, Z., DeBeer, D., 1994. Biofilms, the customized microniche. J. Bacteriol. 176, 2137–2142.

29

DeGregorio, B., Sharp, T., 2002. Kerogen distribution in oxide bearing microfossils from Gunflint formation. In: Second Annual Astrobiology Science Conference. Abstract, p. 56. Dorn, R.I., 1991. Rock varnish. Am. Sci. 79, 542–553. Dorn, R.I., 1998. Rock Coatings. Elsevier, New York. 429 p. Dragovich, D., 1993. Distribution and chemical composition of microcolonial fungi and rock coatings from arid Australia. Phys. Geogr. 14, 323– 341. Edgett, K.S., Parker, T.J., 1997. Water on early Mars: possible subaqueous sedimentary deposits covering ancient cratered terrain in western Arabia and Sinus Meridiani. Geophys. Res. Lett. 24, 2897–2900. Ferris, F.G., Beveridge, T.J., Fyfe, W.S., 1986. Iron-silica crystallite nucleation by bacteria in a geothermal sediment. Nature 320, 609–611. Ferris, F.G., Fyfe, W.S., Beveridge, T.J., 1988. Metallic ion binding by Bacillus subtilis: implications for the fossilization of microorganisms. Geology 16, 149–152. Ferris, F.G., Schultze, S., Witten, T.C., Fyfe, W.S., Beveridge, T.J., 1989a. Metal interactions with microbial biofilms in acidic and neutral pH environments. Appl. Environ. Microb. 55, 1249–1257. Ferris, F.G., Tazaki, K., Fyfe, W.S., 1989b. Iron oxides in acid mine draining environments and their association with bacteria. Chem. Geol. 63, 321– 330. Flood, B.E., Allen, C., Longazo, T., 2003. Microbial fossils detected in desert varnish. Lunar Planet. Sci. XXXIV. Abstract #1633, p. 126. Fortin, D., Ferris, F.G., Beveridge, T.J., 1997. Surface-mediated mineral development by bacteria. In: Bandfield, J.F., Nealson, K.H. (Eds.), Geomicrobiology: Interactions Between Microbes and Minerals. In: Rev. Mineral., vol. 35, pp. 161–180. Furniss, G., Hinman, N.W., Doyle, G.A., Runnells, D.D., 1999. Radiocarbon-dated ferricrete provides a record of natural acid rock drainage and paleoclimatic changes. Environ. Geol. 37, 102–106. Golombek, M., 21 colleagues, 2003. Selection of the Mars Exploration Rover landing sites. J. Geophys. Res. 108. 13-1–13-48. doi:10.1029/ 2003JE002074. Gorbushina, A., 2003. Microcolonial fungi: survival potential of terrestrial vegetative structures. Astrobiology 3, 543–554. Gorbushina, A.A., Krumbein, W.E., Volkmann, M., 2002. Rock surfaces as life indicators: new ways to demonstrate life and traces of former life. Astrobiology 2, 203–213. Gorevan, S., 13 colleagues, 2003. Rock Abrasion Tool Mars Exploration Rover mission. J. Geophys. Res. 108. 9-1–9-8. doi:10.1029/ 2003JE002061. Guilbert, J.M., Park Jr., C.F., 1986. Geology of Ore Deposits. Freeman, New York. 985 p. Herkenhoff, K.E., 23 colleagues, 2003. Athena Microscopic Imager investigation. J. Geophys. Res. 108. 6-1–6-23. doi:10.1029/2003JE002076. Hungate, B., Danin, A., Pellerin, N.B., Stemmler, J., Kjellander, P., Adams, J.B., Staley, J.T., 1987. Characterization of manganese-oxidizing (MnII → MnIV) bacteria from Negev Desert rock varnish: implications in desert varnish formation. Can. J. Microbiol. 33, 939–943. Hynek, B.M., Arvidson, R.E., Phillips, R.J., 2002. Geologic setting and origin of Terra Meridiani hematite deposit on Mars. J. Geophys. Res. 107, 5088. doi:10.1029/2002JE001891. Jakosky, B., 1998. The Search for Life on Other Planets. Cambridge Univ. Press, New York. 336 p. Kirkland, L.E., Herr, K.C., Adams, P.M., Westall, F., Salisbury, J.W., 2002. Spectra of cemented, hematite-rich material and TES spectra of Sinus Meridiani, Mars. Lunar Planet. Sci. XXXIII. Abstract #1218, p. 20. Klein, C.K., Beukes, N.J., 1992. Time distribution, stratigraphy, sedimentologic setting, and geochemistry of Precambrian iron-formations. In: Schopf, J.W., Klein, C.K. (Eds.), The Proterozoic Biosphere. Cambridge Univ. Press, New York, pp. 139–146. Klingelhofer, G., 14 colleagues, 2003. Athena MIMOS II Mössbauer spectrometer investigation. J. Geophys. Res. 108. 8-1–8-18. doi:10.1029/ 2003JE002138. Krinsley, D., 1998. Models of rock varnish formation constrained by high resolution transmission electron microscopy. Sedimentology 45, 711– 725.

30

C.C. Allen et al. / Icarus 171 (2004) 20–30

Krinsley, D., Dorn, R.I., 1991. New eyes on eastern California rock varnish. California Geology 44, 107–115. Krinsley, D., Rusk, B.G., 2000. Bacterial presence in layered rock varnish—possible Mars analog? In: Second International Conference on Mars Polar Science and Exploration. Abstract #4001, pp. 98–99. Krinsley, D., Dorn, R.I., Anderson, S., 1990. Factors that may interfere with the dating of rock varnish. Phys. Geogr. 11, 97–119. Lane, M.D., Morris, R.V., Mertzman, S.A., Christensen, P.R., 2002. Evidence for platy hematite grains in Sinus Meridiani, Mars. J. Geophys. Res. 107. 9-1–9-15. Liu, T., 1994. Visual microlaminations in rock varnish: a new paleoenvironmental and geomorphic tool in drylands. PhD thesis. Arizona State University. McKeown, D.A., Post, J.E., 2001. Characterization of manganese oxide mineralogy in rock varnish and dendrites using X-ray absorption spectroscopy. Am. Mineral. 86, 701–713. Nagy, B., Nagy, L.A., Rigali, M.J., Jones, W.D., Krinsley, D.H., Sinclair, N.A., 1991. Rock varnish in the Sonoran Desert: microbiologically mediated accumulation of manganiferous sediments. Sedimentology 38, 1153–1171. Newsom, H.E., Barber, C.A., Hare, T.M., Schelble, R.T., Sutherland, V.A., Feldman, W.C., 2003. Paleolakes and impact basins in southern Arabia Terra, including Meridiani Planum: implications for the formation of hematite deposits on Mars. J. Geophys. Res. 108. doi:10.1029/ 2002JE001993. Palmer, F.E., Staley, J.T., Murray, R.G.E., Counsell, T., Adams, J.B., 1986. Identification of manganese-oxidizing bacteria from desert varnish. Geomicrobiol. J. 4, 343–360. Pierson, B.K., Parenteau, M.N., 2000. Phototrophs in high iron microbial mats; microstructure of mats in iron-depositing hot springs. FEMS Microb. Ecol. 32, 181–196. Post, J.E., 1999. Manganese oxide minerals; crystal structures and economic and environmental significance. Proc. Nat. Acad. Sci. 96, 3447– 3454. Potter, R.M., Rossman, G.R., 1977. Desert varnish: the importance of clay minerals. Science 196, 1446–1448. Potter, R.M., Rossman, G.R., 1979. The manganese- and iron-oxide mineralogy of desert varnish. Chem. Geol. 25, 79–94. Probst, L.W., Allen, C.C., Thomas-Keprta, K.L., Clemett, S.J., Longazo, T.G., Nelman-Gonzalez, M.A., Sams, C., 2002. Desert varnish— preservation of biofabrics and implications for Mars. Lunar Planet. Sci. XXXIII. Abstract #1764, p. 78. Rieder, R., Gellert, R., Bruckner, J., Klingelhofer, G., Dreibus, G., Yen, A., Squyres, S., 2003. The new Athena alpha particle X-ray spectrom-

eter for the Mars Exploration Rovers. J. Geophys. Res. 108. 7-1–7-13. doi:10.1029/2003JE002150. Schelble, R.T., Westall, F., Allen, C.C., 2004. ∼ 1.8 Ga iron-mineralized microbiota from the Gunflint Iron Formation, Ontario, Canada: implications for Mars. Adv. Space Res. 33, 1268–1273. Squyres, S.W., 11 colleagues, 2003. Athena Mars rover science investigation. J. Geophys. Res. 108. 3-1–3-21. doi:10.1029/2003JE002121. Staley, J.T., Adams, J.B., Palmer, F.E., 1992. Desert varnish: a biological perspective. In: Stotzky, G., Bollag, J.M. (Eds.), Soil Biochemistry. Dekker, New York, pp. 173–194. Sterflinger, K., 2000. Fungi as geologic agents. Geomicrobiol. J. 17, 97– 124. Strother, P.K., Tobin, K., 1987. Observations of the Genus Huroniospora Barghoorn: implications for Paleoecology of the Gunflint Microbiota. Precambrian Res. 36, 323–333. Taylor-George, S., Palmer, F.E., Staley, J.T., Borns, D.J., Curtiss, B., Adams, J.B., 1983. Fungi and bacteria involved in desert varnish formation. Microb. Ecol. 9, 227–245. Tazaki, K., Ferris, F.G., Wiese, R.G., Fyfe, W.S., 1992. Iron and graphite associated with fossil bacteria in chert. Chem. Geol. 95, 313–325. Thomas-Keprta, K.L., McKay, D.S., Wentworth, S.J., Stevens, T.O., Taunton, A.E., Allen, C.C., Coleman, A., Gibson Jr., E.K., Romanek, C.S., 1998. Bacterial mineralization patterns in basaltic aquifers: implications for possible life in martian meteorite ALH84001. Geology 26, 1031–1034. Wade, M.L., Agresti, D.G., Wdowiak, T.J., Armendarez, L.P., Farmer, J.D., 1999. A Mössbauer investigation of iron-rich terrestrial hydrothermal vent systems: lessons for Mars exploration. J. Geophys. Res. 104, 8489– 8507. Westall, F., 1999. The nature of fossil bacteria: a guide to the search for extraterrestrial life. J. Geophys. Res. 104, 16437–16451. Westall, F., Kirkland, L., 2002. Hematite on Mars: investigations of a terrestrial desert ferricrete analogue. Lunar Planet. Sci. XXXIII. Abstract #1179, p. 118. Westall, F., Steele, A., Toporski, J., Walsh, M., Allen, C., Guidry, S., McKay, D., Gibson, E., Chafetz, H., 2000. Polymeric substances and biofilms as biomarkers in terrestrial and extraterrestrial materials: implications for extraterrestrial samples. J. Geophys. Res. 105, 24511–24527. Westall, F., De Wit, M.J., Dann, J., Van Der Gaast, S., De Ronde, C., Gerneke, D., 2001. Early Archaean fossil bacteria and biofilms in hydrothermally influenced, shallow water sediments, Barberton Greenstone Belt, South Africa. Precambrian Res. 106, 91–112. Westall, F., Hofmann, B., Brack, A., 2004. Searching for fossil microbial biofilms on Mars: a case study using a 3.46 billion-year old example from the Pilbara in Australia. In: ESA SP, vol. 545. In press.