Characteristics, distribution, origin, and ... - Wiley Online Library

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 116, E00F23, doi:10.1029/2010JE003767, 2011

Characteristics, distribution, origin, and significance of opaline silica observed by the Spirit rover in Gusev crater, Mars Steven W. Ruff,1 Jack D. Farmer,1 Wendy M. Calvin,2 Kenneth E. Herkenhoff,3 Jeffrey R. Johnson,3 Richard V. Morris,4 Melissa S. Rice,5 Raymond E. Arvidson,6 James F. Bell III,5 Philip R. Christensen,1 and Steven W. Squyres5 Received 27 October 2010; revised 19 January 2011; accepted 4 February 2011; published 22 April 2011.

[1] The presence of outcrops and soil (regolith) rich in opaline silica (∼65–92 wt % SiO2)

in association with volcanic materials adjacent to the “Home Plate” feature in Gusev crater is evidence for hydrothermal conditions. The Spirit rover has supplied a diverse set of observations that are used here to better understand the formation of silica and the activity, abundance, and fate of water in the first hydrothermal system to be explored in situ on Mars. We apply spectral, chemical, morphological, textural, and stratigraphic observations to assess whether the silica was produced by acid sulfate leaching of precursor rocks, by precipitation from silica‐rich solutions, or by some combination. The apparent lack of S enrichment and the relatively low oxidation state of the Home Plate silica‐rich materials appear inconsistent with the originally proposed Hawaiian analog for fumarolic acid sulfate leaching. The stratiform distribution of the silica‐rich outcrops and their porous and brecciated microtextures are consistent with sinter produced by silica precipitation. There is no evidence for crystalline quartz phases among the silica occurrences, an indication of the lack of diagenetic maturation following the production of the amorphous opaline phase.

Citation: Ruff, S. W., et al. (2011), Characteristics, distribution, origin, and significance of opaline silica observed by the Spirit rover in Gusev crater, Mars, J. Geophys. Res., 116, E00F23, doi:10.1029/2010JE003767.

1. Introduction [2] Hydrothermal environments have long been suspected for Mars based upon orbital observations, terrestrial analogs, and Martian meteorites [e.g., Newsom, 1980; Gulick and Baker, 1990; Farmer, 1996; Greenwood et al., 2000]. In Gusev crater, the Mars Exploration Rover (MER) Spirit observed multiple outcrops and an isolated occurrence of soil rich in silica adjacent to Home Plate, a low plateau of layered volcanic materials [Squyres et al., 2008] (Figure 1). The silica phase is opal‐A (hydrated amorphous silica), determined with thermal infrared spectra from the Miniature Thermal Emission Spectrometer (Mini‐TES). The SiO2 abundance is >90 wt % in soil measured by Spirit’s Alpha Particle X‐ray Spectrometer (APXS). Abundant opaline silica in a volcanic setting is compelling evidence for a hydrothermal system in

1 School of Earth and Space Exploration, Arizona State University, Tempe, Arizona, USA. 2 Department of Geological Sciences, University of Nevada, Reno, Nevada, USA. 3 U.S. Geological Survey, Flagstaff, Arizona, USA. 4 NASA Johnson Space Center, Houston, Texas, USA. 5 Department of Astronomy, Cornell University, Ithaca, New York, USA. 6 Department of Planetary Sciences, Washington University, St. Louis, Missouri, USA.

Copyright 2011 by the American Geophysical Union. 0148‐0227/11/2010JE003767

this location [Squyres et al., 2008]. This discovery, which represents the first definitive identification of opaline silica on Mars, was followed by identifications elsewhere using near‐infrared (NIR) spectra from the Compact Reconnaissance Imaging Spectrometer for Mars on the Mars Reconnaissance Orbiter [e.g., Milliken et al., 2008; Ehlmann et al., 2009]. Previously, Mini‐TES spectra from the Opportunity rover in Meridiani Planum showed indications of as much as 25% (by area) opaline silica and/or high‐silica glass in outcrop rocks [Glotch et al., 2006]. A high‐silica phase, perhaps opal, has been suggested as a major component of the surface type 2 material identified with spectra from the Thermal Emission Spectrometer on the Mars Global Surveyor [e.g., Bandfield et al., 2000; Kraft et al., 2003; Michalski et al., 2005]. But the role of water in the formation of these phases remains ambiguous given the uncertainty regarding their origin by primary (igneous) or secondary (alteration) processes. [3] Water in some form was required to produce the highly enriched concentrations of opal‐A in the Home Plate occurrences. An enrichment of Ti observed in the silica‐rich soil was interpreted as evidence for the dissolution of basaltic materials by low pH fluids that concentrated the relatively insoluble SiO2 and TiO2 components [Squyres et al., 2008]. Coupled with the presence of S‐rich soils encountered elsewhere in the Columbia Hills, the role of low pH waters, perhaps an acid sulfate condensate, was the favored explanation for the silica occurrences. Unresolved was the issue

E00F23

1 of 48

E00F23

RUFF ET AL.: SILICA IN GUSEV CRATER

E00F23

Figure 1. A portion of HiRISE image PSP_001513_1655_RED showing Home Plate and vicinity with Spirit parked at Low Ridge during its second Martian winter. Examples of silica‐rich nodular outcrops are: Tyrone Nodules (TN), Kobal (K), Clara Zaph (CZ), Elizabeth Mahon (EM), Nancy Warren (NW), Darlene Mickelsen (DM), Virginia Bell (VB), Innocent Bystander (IB), Norma Luker (NL), Philomena Zale (PZ), and Stapledon. Gertrude Weise (GW) is the silica‐rich soil. Halley (H) is sulfur‐rich outcrop. and Eileen Dean (ED) is silica and sulfur‐poor light‐toned soil. Olive Little (OL), Good Question (GQ), Kadabra (Ka) and other labeled features are described in the text. Dashed lines indicate approximate extent of recognized and likely silica‐rich nodular outcrops. White ellipses enclose speculative candidates for this material. of whether the silica was produced by leaching from condensates of fumaroles or precipitation as silica sinter in a hot spring environment. Although both represent hydrothermal settings that could have led to locally habitable environments [Squyres et al., 2008], the lack of specificity hinders our ability to constrain the nature of the associated aqueous conditions. [4] On Earth, opaline silica is produced over a range of temperature, pH, and water‐to‐rock ratio conditions, factors that must be taken into account in evaluating habitability [e.g., Knoll et al., 2005; Tosca et al., 2008; Tosca and Knoll, 2009]. In the case of hot springs and geysers, precipitation from silica‐rich solutions produces sedimentary rock known as silica sinter. Siliceous hot spring deposits form over a broad range of pH, from acidic to alkaline, although most typically they precipitate from near neutral to alkaline pH, alkali chloride water. In the case of fumaroles, silica residue forms as a leaching product from the reaction between

silicate‐rich rocks and acid sulfate steam condensates, sometimes described as solfataric alteration [e.g., Payne and Mau, 1946; White et al., 1956; Bignall and Browne, 1994; Herdianita et al., 2000; Rodgers et al., 2004]. Although a single hydrothermal system can produce both silica residue and precipitate, chemical and textural variations and stratigraphic relationships are sometimes evident in terrestrial examples that serve to distinguish them. [5] In this study, we present a range of observations intended to better understand the geologic environment in which the Home Plate silica formed. In section 2 we describe the data sets and their sources, and laboratory samples and measurements used in our analyses. Section 3 presents the results of the various observations, which are then used in section 4 to establish the distribution and stratigraphy of the silica occurrences. All of the observations are discussed in section 5 as they relate to an acid sulfate leaching origin or precipitation origin for Home Plate silica and their

2 of 48

E00F23


significance with respect to climate and habitability. Section 6 contains a set of conclusions that represents our most confident results.

2. Data Sets and Materials 2.1. Miniature Thermal Emission Spectrometer [6] Miniature Thermal Emission Spectrometer (Mini‐TES) remote sensing observations span the spectral range from ∼340 to 2000 cm−1 (∼5–29 mm) with ∼10 cm−1 spectral sampling and a nominal spatial resolution of 20 mrad [Christensen et al., 2003]. Typical observations involve a single “stare” of the target of interest in which 200 individual spectral scans plus internal blackbody measurements for calibration are acquired. Each of the individual scans is downlinked and calibrated to emissivity including a correction for mirror dust contamination (section 3.1.1). The Mini‐ TES spectra presented in this paper are an average of the individual mirror dust corrected spectra. Those showing evidence for an opaline silica component are available in the auxiliary material.1 Although the more useable spectral range of ∼360 to 1800 cm−1 is shown in text images, the spectra given in Data Sets S1 and S2 span the full measured range. 2.2. Panoramic Camera [7] Panoramic Camera (Pancam) is a multispectral stereo imager that views surface targets over the spectral range from 432 to 1009 nm through 13 glass interference filters, of which two are redundant to allow red/blue stereo imaging [Bell et al., 2003]. Single frames or subframes are acquired from a single pointing of the Pancam Mast Assembly (PMA). Mosaics of two or more frames up to full 360° panoramas are built up from multiple pointings, which in the case of the largest examples, were typically over multiple sols (Martian day). In this paper, both approximate true color and false color images are presented that have been obtained from the Pancam website (http://marswatch.astro.cornell.edu/ pancam_instrument/), which includes a description of how they were processed. The Web site includes the processed mosaics and panoramas that are presented at reduced resolution in this paper. Both panoramas that we discuss are available in anaglyph stereo from the Website (not shown in this paper). Pancam spectra also are presented in this paper and are described more fully in section 3.2.2. 2.3. Alpha Particle X‐Ray Spectrometer [8] The Alpha Particle X‐ray Spectrometer (APXS) is one of the three scientific instruments mounted on the end of Spirit’s articulated robotic arm known as the Instrument Deployment Device (IDD). A fourth instrument called the Rock Abrasion Tool (RAT) also is mounted on the IDD, although since sol 417 it has only been used for brushing because of significant wear of its abrasive bits. The APXS is a contact instrument that characterizes the abundance of elements from Na to Y via X‐ray radiation excited by alpha particles and X rays from its radioactive sources [Rieder et al., 2003]. This interaction is analogous to that used in particle‐ induced X‐ray emission (PIXE) and X‐ray fluorescence 1 Auxiliary materials are available at ftp://ftp.agu.org/apend/je/ 2010je003767.

E00F23

(XRF) measurements. Although its maximum field of view (FOV) is 38 mm, its effective FOV is a circular spot ∼25 mm in diameter [Gellert et al., 2004]. All the APXS data we present are from the published work of Ming et al. [2008] except in one case (Stapledon, section 4.1.3). 2.4. Mössbauer Spectrometer [9] The IDD‐mounted MIMOS II Mössbauer spectrometer (MB) is a contact instrument that uses radioactive sources to generate the Mössbauer effect, the recoilless emission and resonant absorption of gamma rays by, in this application, the 57 Fe nuclei in solid materials [Klingelhöfer et al., 2003]. Mössbauer spectra contain information on the identity and abundance of Fe‐bearing phases within its “beam diameter” of ∼14 mm on the sample. The sensor head of the MB has an aluminum contact plate with switches designed to trigger at ∼1 N of pressure applied by the IDD to the surface on which it is placed. This ability to locate a surface target in a coordinate frame used by the other IDD instruments means that the MB was deployed in some cases without acquiring its own data. All MB data we present are from the published work of Morris et al. [2008]. 2.5. Microscopic Imager [10] The Microscopic Imager (MI) is a camera mounted opposite the MB on the IDD. Its design is similar to Pancam and the engineering cameras (section 2.6) but makes use of a “macro” lens to acquire broadband (400–700 nm) images with a spatial resolution of 30 mm/pixel across its 31 × 31 mm FOV using only solar or skylight illumination of the target surface [Herkenhoff et al., 2003]. Because of its fixed focus, it uses small movements of the IDD to acquire several images (typically a 5 or 7 image “stack”) from varying distances to achieve a best focus position. A 42 mm long spring‐mounted shaft is used as a contact sensor to place the MI slightly closer to the target surface than its best focus distance (about 66 mm) from which IDD adjustments can be made. Although the contact sensor is outside the instrument’s FOV, it produces a thin shadow visible in some MI images. Its thin aspect ratio is not suited to soil targets, so the MB contact sensor is used to locate the surface in these cases. IDD movements also are used to produce MI mosaics of varying size, for example 2 × 2 frames centered on a RAT brushed spot. Similarly, stereo images are produced using a routinely acquired MI image laterally offset from the main stack. All MI single frame images and some stereo anaglyphs presented in this paper are available from the Planetary Data System Analyst’s Notebook website (http://an.rsl.wustl.edu/mer/), although mosaics are not. 2.6. Engineering Cameras [11] Two Navigation cameras (Navcam) are mounted on the PMA and a total of four Hazard Avoidance cameras (Hazcam) are mounted fore and aft on the rover body. All have a common design with the other rover cameras but with different lenses intended for their specific use [Maki et al., 2003]. Both Navcam and Hazcam provide gray scale stereo images used to guide the rover and orient its IDD (front Hazcam). A combination of filters produces a relatively narrow band pass from ∼600 to 800 nm centered at ∼650 nm (red). All Navcam images presented in this paper are available as single frame monoscopic images from the Planetary Data

3 of 48

E00F23


System Analyst’s Notebook website (http://an.rsl.wustl.edu/ mer/). 2.7. Laboratory Spectra and Sample Materials [12] Laboratory thermal emission spectra were measured at the Arizona State University (ASU) Mars Space Flight Facility using a Nicolet 670 Fourier transform infrared spectrometer modified to make emission measurements [Ruff et al., 1997]. Data were collected over the range ∼200–2000 cm−1 with ∼2 cm−1 sampling and calibrated to spectral emissivity using the technique of Ruff et al. [1997]. Spectra were obtained from a range of synthetic and natural silica‐rich materials. Synthetic amorphous silica, variously referred to as vitreous silica, fused quartz, or fused silica, is manufactured in different forms for various scientific and industrial applications. Fused quartz, as its name implies, is produced from the melting and quenching of quartz grains, although we have received some fused quartz samples that appear to be borosilicate glass. Our fused quartz samples are in the form of discs 25 mm in diameter and 4 mm thick from Technical Glass Products, Ohio. Among these are microporous frits that typically are used for gas filtration or diffusion and are available with a range of pore sizes. The frit spectrum shown in this paper is from one with 15–40 mm pores. [13] In addition to a precious opal specimen from the ASU teaching collection, we have measured the spectral emissivity of a range of silica sinter and residue samples that vary widely in their texture and porosity, and to a lesser extent, mineralogy (i.e., crystallinity and accessory phases). Samples were acquired from hydrothermal environments in Yellowstone National Park, Wyoming; Steamboat Springs, Nevada; Sulphur Banks, Hawaii; and sites in the Taupo Volcanic Zone in New Zealand. [14] In order to facilitate comparisons between the terrestrial and Martian materials, some samples (natural and synthetic) were measured over a range of emission angles (up to ∼60°) to better simulate the viewing geometry of the Mini‐ TES observations (see section 3.1.2). Natural samples with a relatively flat unprepared surface and in some cases, a sawn surface, were used for these measurements. The synthetic samples all had a flat surface. Small samples (1300 cm−1), the features of the corrected spectrum are slightly deeper than those of the reference. This may be due in part to the effect of greater emission angle during the measurement of the Husband Hill dust. But there is misfit at low wave numbers (1 near 550 cm−1 in the case of 50% background subtraction serves as an indication of a limiting case. Given that emis-

sivity values never exceed unity, removing 50% background is unrealistic. Even the 40% case with its feature very near 1.0 probably is unrealistic. We use the more conservative 30% case for the comparative analysis of section 3.1.6. [33] There are 17 Mini‐TES observations of outcrops that show features attributable to opaline silica acquired before the major accumulation of additional dust on the Mini‐TES optics began on approximately sol 1220. All show some combination of contributions from contaminating soil and dust as well as features attributable to viewing angle and optically thin dust [Ruff and Bandfield, 2010] (Figure 9). We use the prominent emissivity minimum near 475 cm−1 as a unique indicator of the presence of an opaline phase. No other rocks observed by Mini‐TES in Gusev crater have this feature with the same position and shape. The low wave number region also is the least affected by dust in any form [Ruff et al., 2006]. Eight of the spectra are on outcrops within 1 m of an IDD target in which the APXS measured silica > 60 wt %. As shown in section 3.4, the silica‐rich outcrops have a distinctive morphology. Among these 17 Mini‐TES spectra, all occur on targets where a Navcam image, and in some cases a Pancam image, shows this distinctive morphology. 3.1.6. Comparison With Other SiO2 Phases and Forms [34] On Earth, opal‐A is susceptible to diagenesis, becoming increasing crystalline (paracrystalline) via short‐range ordered domains resembling cristobalite (C) and tridymite (T) under a wide range of temperature and pressure conditions [e.g., Williams et al., 1985; Herdianita et al., 2000; Lynne and Campbell, 2004]. With increasing diagenesis, the paracrystalline varieties of opal lead to microcrystalline (20 mm) quartz (mega-

11 of 48

E00F23


E00F23

Figure 9. Mini‐TES spectra of outcrops/rocks with features attributable to opaline silica and viewing angle. Three vertical lines represent the location of features present in opal‐A. Soil and dust contribute their spectral features with varying intensity, obscuring the opal‐A features. The emissivity minimum near 475 cm−1 is least affected by contaminants so is most diagnostic. quartz) [Hesse, 1989]. All of these phases are spectrally distinguishable in the thermal infrared wavelengths as shown in Figure 10a. [35] Michalski et al. [2003] demonstrated that the spectral emissivity of paracrystalline opal‐CT and opal‐CT/C (borderline between opal‐CT and opal‐C) are only subtly different from that of opal‐A. We recognize a feature in these spectra that possibly serves as an indicator of increasing crystallinity. A secondary Christiansen feature (an emissivity

maximum) is prominent near 550 cm−1 in the spectra of the fully crystalline polymorphs tridymite, cristoballite, microquartz, and quartz (Figure 10a). It also is evident in the paracrystalline opal presented by Michalski et al. [2003] where it occurs as a prominent inflection point leading to the strong emissivity minimum near 475 cm−1. In opal‐A there is no prominent secondary Christensen feature, and although there is a strong minimum near 475 cm−1, the inflection point is shifted to ∼520 cm−1 (Figure 10b). With

12 of 48

E00F23


regard to this spectral characteristic, the Mini‐TES spectra of silica‐rich Kenosha Comets soil and Clara Zaph4 outcrop are more similar to fully amorphous opal‐A. Although more work is needed to demonstrate the robustness of this possible crystallinity indicator, given the absence of a fully crystalline SiO2 phase, the Home Plate silica appears to have experienced little diagenesis. [36] Opal‐A is the dominant phase in both silica sinter (from precipitation) and silica residue (from leaching).

E00F23

Despite their different origin, both types of occurrences are spectrally similar and provide a good fit to the Martian silica (Figure 11). However, there are subtle but notable differences that may prove to be significant with the analysis of additional samples. The silica residue samples in Figure 11 are from Sulphur Banks, HI where another sample presented by Morris et al. [2000] was shown to have both SiO2 and TiO2 enrichment similar to some of the silica‐rich materials at Home Plate [Squyres et al., 2008]. Their spectra provide a better fit to the contrast of the ∼475 cm−1 feature of the Clara Zaph4 outcrop than the sinter from Excelsior Geyser Crater, Yellowstone (Figure 11b). However, the situation is reversed at high wave numbers (>1400 cm−1) where the sinter provides a better fit to the contrast. These variations in contrast may be less significant than the appearance of a feature near 550 cm−1 in the silica residue spectra. There is no comparable feature in the sinter and Mini‐TES spectra in this location. Given that acid sulfate alteration is known to produce phyllosilicate and sulfate phases in addition to the silica [e.g., Macdonald, 1944; Payne and Mau, 1946; Morris et al., 2000; Rodgers et al., 2002], these are candidate phases for producing the feature. [37] We note that the Clara Zaph4 outcrop spectrum shows evidence for scattering due to microporosity as described in section 3.1.3. The spectra of microporous silica sinters provide a better fit than nonporous silica in the region around 900 cm−1 and to a lesser extent above ∼1400 cm−1 where transparency features arise (compare Figures 11b and 11c). These spectral characteristics are consistent with the Class 2 spectra representing porous, poorly consolidated sinter material described by Goryniuk et al. [2004]. The microtexture of the silica residue samples shown in Figures 11b and 11c has not yet been characterized. Microporosity in silica sinter can arise from the dissolution of detrital and authigenic grains as well as entombed microbes [e.g., Hinman and Lindstrom, 1996; Farmer and Des Marais, 1999; Jones and Renaut, 2003; Lynne and Campbell, 2004]. Because the scale of this porosity is below the resolving power of the MI, we have not been able to characterize the shapes of micropores in the Home Plate silica. 3.2. Pancam Spectral Characteristics 3.2.1. Visible Color [38] The Gertrude Weise soil target Kenosha Comets represents the purest example of silica‐rich material observed by Spirit and displays a much whiter hue relative to the Figure 10. Laboratory spectra of silica polymorphs compared with Mini‐TES spectra of silica‐rich soil and outcrop. (a) The crystalline silica phases involved in the diagenetic transformation of opal‐A are all spectrally distinguishable. (b) Laboratory spectra of transitional (paracrystalline) forms of opal that incorporate cristobalite (C) and tridymite (T) are subtly different from opal‐A. A secondary Christensen feature (emissivity maximum) near 550 cm−1 (vertical line) is common to the paracrystalline and crystalline phases and lacking in opal‐A. Mini‐TES spectra of silica‐rich Kenosha Comets soil (KC) and Clara Zaph4 outcrop (CZ4; minus 30% background) more closely resemble opal‐A in this regard. All laboratory spectra except microquartz are from Michalski et al. [2003] and have been normalized and offset.

13 of 48

E00F23


Figure 11. Mini‐TES spectra of the silica‐rich Kenosha Comets soil and Clara Zaph4 outcrop (with correction for a 30% contribution from background soil) compared to best fit examples of natural silica residue and silica sinter. Laboratory spectra have been resampled and contrast reduced (except where noted) to fit the Mini‐TES spectra at ∼1100 cm−1. (a) Both crushed, unsieved terrestrial samples measured at a 45° emission angle provide a good fit to the Martian material. (b) Solid samples of microporous sinter and apparently microporous residue (the latter with no constrast change) measured at a 60° emission angle provide a better fit to the Mini‐TES spectrum than (c) nonporous sinter and apparently nonporous residue.

E00F23

reddish hues of the adjacent soil in approximate true color images (Figure 3). The silica‐rich outcrops show a reddish hue in approximate true color images except where they are broken open by the rover, at which point they are much whiter (Figure 12a). False color images emphasize the lighter tone and reddish hue of the undisturbed outcrops relative to the bluer hues of silica‐poor rocks and outcrops throughout the Eastern Valley (Figure 12b). Dark‐toned, silica‐poor soils, even where covered by thin or patchy accumulations of dust, also appear relatively blue in false color images. [39] The redder hue of undisturbed silica‐rich outcrop rocks relative to the whiter hue where they are broken open is due in part to air fall dust accumulation. Rocks that were broken open on sol 1234 were reimaged on sol 1294 after the fallout from regional dust storms [Malin et al., 2007]. Their initial whitish hue was replaced by a redder hue similar to that of adjacent undisturbed silica‐rich outcrop (Figure 12, inset). It is likely that the porous texture evident on both disturbed and undisturbed surfaces of these rocks (section 3.5) is an effective dust trap, explaining the reddish hue. [40] Where no Pancam color images exist, Navcam gray scale images provide some capacity for distinguishing between silica‐rich and silica‐poor rocks. Sunlit surfaces of silica‐rich outcrops commonly represent the brightest component in a Navcam scene that contains them (Figure 13). 3.2.2. Multispectral Characteristics [41] The silica‐rich soil and outcrops are spectrally distinct from other materials in Gusev crater in visible to near‐ infrared (Vis‐NIR) wavelengths. Wang et al. [2008] identified a strong spectral downturn from 934 to 1009 nm that characterizes the Pancam spectra of all known high‐silica targets. This feature is not a direct indication of the silica‐rich nature of soil and outcrops (amorphous silica is typically featureless in Vis‐NIR wavelengths), but is likely due to the 2n 1 + n 3 H2O combination band centered near ∼1000 nm [Rice et al., 2010]. The silica‐rich soil and outcrops in the Home Plate vicinity also are characterized by: red‐to‐blue (436 to 754 nm) slopes that are greater than those of the surrounding surface dust; positive 864 to 934 nm slopes; and “flat” near‐infrared profiles from 864 to 934 nm (Figure 14). Using spectral criteria based on these four spectral parameters (934 to 1009 nm slopes less than −2.0 × 10−4 nm−1; 754 to 864 nm slopes greater than 0.0 nm−1; 436 to 754 nm slopes greater than 4.0 × 10−4 nm−1; and 864 to 934 nm slopes between −1.0 and 1.0 × 10−4 nm−1), Rice et al. [2010] defined a “hydration signature” that characterizes the Pancam spectra of all recognized opaline‐silica‐rich targets in Gusev crater. [42] Figure 14 shows where the spectral criteria of the hydration signature are met in a Pancam multispectral observation of the Mini‐TES target Clara Zaph4 mapped over a Pancam R2 filter (754 nm) image. The color scale in these maps indicates the magnitude of the 934 to 1009 nm slope, which weakly correlates with SiO2 content in targets with APXS observations [Rice et al., 2010]. The hydration signature does not map continuously over the entire extent of the Clara Zaph4 outcrop. This likely results from variable dust cover, illumination effects such as shadowing, and/or compositional heterogeneities within the deposits. [43] Although the hydration signature alone does not uniquely identify silica‐rich materials (spectra of a few other hydrated minerals, such as gypsum and epsomite, also meet the hydration signature criteria), it can be used along

14 of 48

E00F23


E00F23

Figure 12. Pancam images of undisturbed and disturbed silica‐rich outcrops among silica‐poor soil, dust, and float rocks. (a) The approximate true color mosaic (sol 1234, P2378) highlights the whitish hue of the silica‐rich rocks crushed by the rover wheel on the right side of the image. (b) The false color version emphasizes the redder hues of the undisturbed silica‐rich outcrops relative to the bluer hues of silica‐poor rocks and soil. Disturbed silica‐rich rocks and soil appear bluish white in this color stretch. The scene is ∼1 m across. Nancy Warren (NW) and Darlene Mickelsen (DM) were the targets of IDD measurements. The inset false color image (sol 1294, P2581) acquired from a different position shows that the hue of the disturbed rocks Innocent Bystander (IB) and Norma Luker (NL) is nearly the same as undisturbed rock Virginia Bell (VB) following major atmospheric dust fallout during the 60 sols after the rocks were broken open. Parts of VB were exhumed by vigorous winds during this time. with morphological evidence to identify candidate silica‐rich outcrops that have not been targeted by APXS and/or Mini‐ TES. It should only be used as a tool for identifying potential silica‐rich materials when Pancam observations of the targets are made at low local incidence angles. Rice et al. [2010] have

shown that the hydration signature presents false positives on surfaces tilted away from the rover’s line of sight, and on flat surfaces in image sequences acquired at very low Sun angles. This effect likely is due to calibration inaccuracies at low solar elevations at longer wavelengths. The Pancam calibration

Figure 13. Portion of a Navcam mosaic (sol 1232) showing an expanded view of the scene in Figure 12. In direct sunlight, the silica‐rich nodular outcrops appear light‐toned in the ∼600–800 nm band pass of Navcam relative to surrounding materials. CS is the location of IDD measurements on the target Calibration Soil on sol 1230 near the Nancy Warren (NW) silica‐rich nodular outcrop. 15 of 48

E00F23


E00F23

Figure 14. Pancam images and spectra of silica‐rich nodular outcrop (sol 1174, P2588) that includes the Mini‐TES target Clara Zaph4 (circled) and IDD target Elizabeth Mahon (EM). (a) Approximate true color image (blue, 482 nm; green, 535 nm; red, 601 nm). (b) False color image (blue, 432 nm; green, 535 nm; red, 753 nm). The blue and red boxes show regions from which the spectra in Figure 14d were extracted. (c) Hydration signature map defined by four spectral parameters (see text) highlights silica‐rich rocks. The color scale corresponds to the magnitude of the 934 to 1009 nm slope, with purple indicating slopes above the threshold limit of −2.0 × 10−4 nm−1 and red indicating slopes less than −4.0 × 10−4 nm−1. Map is overlain on the R2 (754 nm) Pancam image. (d) Pancam spectra of the Gertrude Weise silica‐rich soil target (green diamonds; from sol 1158, P2581) and the Clara Zaph4 silica‐rich nodular outcrop (blue squares; from sol 1174, P2588), compared to the spectrum of surface dust on undisturbed soil (red triangles; from sol 1174, P2588). R* is equivalent to Lambert albedo. model assumes that the calibration target behaves as a Lambertian scatterer, an assumption that for the specific calibration target materials used on the rovers, is increasingly incorrect as incidence angles get closer to 90° or at strongly forward scattering geometries [Bell et al., 2003, 2006; Kinch et al., 2007]. 3.3. APXS and MB Observations 3.3.1. APXS [44] In the Eastern Valley (Figure 1), the APXS measured two undisturbed outcrops, two disturbed outcrop rocks, and two undisturbed soil targets with SiO2 > 60 wt %. Squyres

et al. [2008] noted that the outcrop targets all were contaminated by soil as evident in MI images, such that measured SiO2 values should be regarded as lower limits. In that work, contamination of the Kenosha Comets silica‐rich soil target was recognized from the examination of enrichments and deficiencies of its major elements relative to a typical soil, rather than from an MI image. The sampling depth of the APXS measurement is function of atomic number, with X rays from greater depth (50–100 mm) arising from higher atomic number elements [Rieder et al., 2003]. Kenosha Comets data show an exponential downward trend with

16 of 48

E00F23


E00F23

Figure 15. APXS data from Ming et al. [2008] showing the ratio of chemical elements in silica‐rich materials versus a typical undisturbed soil in the Eastern Valley (sol 1230, Calibration Soil), shown from lowest to highest atomic number. An increase in Si relative to the soil is the only enrichment all have in common. increasing atomic number of most major elements except for Si, Ti, Cr, and Zn [Squyres et al., 2008]. This was attributed to the admixture of ∼30% soil grains with the silica‐rich material. Among the other silica‐rich materials in the Eastern Valley, the same trend is not evident (Figure 15) yet images clearly indicate soil contamination (some examples are presented in section 3.5). [45] In an effort to determine a more accurate composition for the measured silica‐rich materials, we have attempted to remove a soil component assuming simple linear mixing between a representative soil and the silica‐rich material. The “typical soil” of Squyres et al. [2008] is used here. It is undisturbed regolith called Calibration Soil (sol 1230) within 0.5 m of the silica‐rich outcrop Nancy Warren (Figure 13). MI images of this soil reveal a relatively uniform grain size of what appears to be windblown sand, likely with some air fall dust coating the grains as determined from Pancam color images. All APXS data presented here are from Ming et al. [2008]. We apply soil corrections to the Si‐rich float rock Fuzzy Smith observed on top of Home Plate that is notably different from other silica‐rich rocks in the Eastern Valley (see section 4.1) and Good Question, a weakly Si‐enriched outcrop on the southeastern edge of Home Plate (see Figure 1 for locations). Fractional amounts of the soil data were subtracted from the silica‐rich materials according to the equation c ¼ ðm xsÞ=ð1 xÞ

ð3Þ

where c is the corrected composition, m is the measured composition, x is a fractional value, and s is the Calibration

Soil. The fractional value was increased in 5% increments until one or more elements became negative, at which point the composition at the previous increment represents the corrected composition. The results are given in Table 2. [46] The SiO2 value increased in all materials except the Kenosha Comets soil. Because its measured K concentration is at the detection limit, it is not possible to subtract any amount of the K‐bearing Calibration Soil without yielding a negative value. Ignoring this element, it is possible to remove nearly 10% of the soil before Fe, Mn, and Na values become negative. This is still well below the ∼30% soil contamination identified by Squyres et al. [2008]. So either the Calibration Soil we used is inappropriate or the assumption of linear mixing is wrong. Alternatively, the analysis of Squyres et al. [2008] is flawed. It is noteworthy that the Lefty Ganote silica‐ rich soil, following subtraction of 40% contaminant, has many elements in close agreement with uncorrected Kenosha Comets including SiO2 at ∼92%. This material is in the same trench containing Kenosha Comets soil, but based on MI images, clearly contains a darker soil mixed in by the rover wheels (section 3.5). The assumption of linear mixing of two components in this case likely is correct, and given the proximity of the Calibration Soil within 4 m, the contaminating soil composition likely is a good proxy. These results suggest that the Kenosha Comets soil has very little contamination. [47] The measured compositions of the two undisturbed outcrop targets Elizabeth Mahon and Nancy Warren are notably similar to each other with the exception of Ni and Zn. Thus soil correction was achieved using the same fractional

17 of 48

1230 Calibration Soil

1179 Good Question outcrop

1251 Innocent Bystander rock

770 Fuzzy Smith rockc

1288 Norma Luker rock

1157 Elizabeth Mahon outcrop

1190 Kenosha Comets soilb 1225 Nancy Warren outcrop

a Data compiled from Ming et al. [2008] and Morris et al. [2008] are sorted from highest to lowest corrected SiO2 value. MAI, Mineralogical Alteration Index (see text); NA, not applicable; ND, not determined; NM, no MB measurement. b No fraction of soil could be subtracted from Kenosha Comets without producing negative values. c Fuzzy Smith occurs on the top of Home Plate rather than in the Eastern Valley.

26 ND 36 27 ND 38 ND NM ND 63 ND 24 ND 38 ND NM 0.36 ND 0.36 0.44 ND 0.54 ND NM ND 0.00 ND 0.64 ND 0.47 ND NM 1024 1563 278 177 157 828 1158 249 267 679 761 768 829 1030 1845 215 562 657 151 185 58 1477 2046 294 226 272 246 612 633 996 1572 420 28 27 17 32 34 43 51 42 49 21 20 91 98 245 461 29 0.30 0.17 0.10 0.50 0.50 0.50 0.50 0.50 0.50 0.60 0.62 0.50 0.50 1.50 2.50 0.50 2.90 2.10 1.00 3.70 3.48 3.78 3.61 3.40 3.02 3.40 3.28 3.60 3.54 4.30 4.50 4.10 0.53 0.33 0.30 0.65 0.55 0.60 0.48 0.53 0.37 0.68 0.65 0.45 0.41 0.72 0.61 0.83 0.18 0.01 0.00 0.27 0.18 0.24 0.13 0.24 0.13 2.76 3.17 0.06 0.02 0.42 0.40 0.44 1.30 0.03 0.30 1.70 0.89 1.40 0.43 1.60 0.74 2.90 2.85 0.40 0.09 2.00 0.80 3.20 2.70 0.17 0.70 3.00 1.12 2.70 0.65 3.10 1.27 1.90 1.09 1.60 1.06 3.90 1.30 6.50 4.20 1.33 2.30 5.30 3.58 5.50 3.88 6.50 5.42 4.20 3.44 12.80 13.28 10.80 13.10 8.50 0.13 0.01 0.03 0.11 0.00 0.11 0.00 0.11 0.00 0.15 0.12 0.08 0.05 0.20 0.09 0.31 6.20 0.33 1.40 5.80 0.85 6.90 2.54 8.50 5.00 6.80 5.35 13.50 13.33 14.80 14.60 15.00 0.61 0.80 0.33 0.26 0.22 0.26 0.22 0.24 0.19 0.06 0.01 0.25 0.24 0.48 0.63 0.33 4.80 0.60 1.70 5.20 2.02 4.70 1.25 5.20 2.02 6.30 5.45 2.80 1.88 6.00 0.90 11.10 1.35 1.74 1.21 0.54 0.42 0.75 0.74 0.78 0.79 1.71 1.88 0.62 0.60 0.73 0.69 0.77 74.60 92.13 90.50 72.80 85.99 72.40 85.38 69.20 80.45 68.40 71.95 63.10 64.74 53.90 59.50 48.30 0 40 0 0 35 0 35 0 35 0 15 0 10 0 50 NA 1199 Lefty Ganote soil

x (%)

SiO2 (%)

TiO2 (%)

Al2O3 (%)

Cr2O3 (%)

FeO (%)

MnO (%)

MgO (%)

CaO (%)

Na2O (%)

K2 O (%)

P 2 O5 (%)

SO3 (%)

Cl (%)

Br ppm

Ni ppm

Zn ppm

Fe#+/FeT

MAI (%)


Material (Sol and Name)

Table 2. Composition and Alteration Parameters of Silica‐Rich Materials in the Eastern Valley Before (x = 0) and After (x > 0) Subtraction of the Maximum Allowable Fraction of a Soil Component (Calibration Soil) Applied in 5% Incrementsa

E00F23

E00F23

amount (35%). The resulting SiO2 values (∼85%) are much closer to those of the silica‐rich soils. [48] On sol 1232, Spirit was commanded to perform a drive sequence intended to break open the silica‐rich nodular outcrop target Virginia Bell within 1 m of Nancy Warren (Figure 12). With the RAT grinding capability unavailable, wheel crushing offered the opportunity to expose a fresh interior surface. The inoperative right front wheel was pushed toward the target and then steered to the left and right twice via the functioning steering actuator followed by two ∼5 cm backward drive segments with the steering actuation at the end of each segment. The result failed to crush the intended target but instead broke open several smaller rocks, one of which became known as Innocent Bystander. An adjacent broken rock is called Norma Luker (Figure 16). The success of the operation was not immediately apparent based on gray scale images from the Navcam and Hazcam, so Spirit was commanded to drive to a light‐toned soil target (Eileen Dean) ∼1 m away. The potential importance of the crushed rocks was subsequently recognized based on the color view afforded by the Pancam image, which was received later in the tactical process (Figure 16). The whitish hues of the crushed rocks prompted a return 19 sols later. By that time, vigorous wind activity [Sullivan et al., 2008] had contaminated the freshly exposed rock surfaces with darker sand grains, which is clearly evident in the MI images (Figure 17 and section 3.5). [49] Correction for the contamination of the Norma Luker crushed rock accommodated 35% soil subtraction, yielding a SiO2 value of ∼80%. This rock is 50% [Morris et al., 2008]. The Fe3+/Fetotal values range more widely, from a Gusev record low of 0.00 (Fuzzy Smith) to 0.64 (Innocent Bystander) with the rest in the low to intermediate range of values for Gusev crater materials. The significance of redox state with regard to the origin of the silica‐rich materials is presented more fully in section 5. 3.4. Morphology of the Outcrops [54] All of the silica‐rich rocks identified by Spirit have a characteristic morphology that appears unique among the

rocks in the Home Plate region. Squyres et al. [2008] used the adjective “nodular” in an effort to describe these rocks and characterized them as “outcrops.” Silica‐rich nodular outcrops is the description adopted in subsequent work [e.g., Arvidson et al., 2008; Wang et al., 2008; Rice et al., 2010]. Hereafter we also adopt this terminology and demonstrate why both “nodular” and “outcrop” are appropriate descriptors. [55] The first indication of opaline silica by Mini‐TES came from a target called Tyrone Nodules (sols 1100 and 1101), so named because of its proximity to the Tyrone S‐rich soil deposit (Figure 1) and the somewhat bulbous, rounded forms of the rocks (Figure 18a). Although it has the appearance of a pile of loose cobbles, other examples proved to be resistant to deformation under the weight of the rover. The largest example named Kobal (∼4 m long and ∼30 cm high) was driven over by Spirit without any apparent deformation (Figure 18b and 19). In a much smaller example < 1 m from opal‐bearing rocks measured by Mini‐TES, Spirit drove over a light‐toned, likely silica‐rich cobble a few cm in length perched on a mound of smaller ones, then onto two darker “bluer” cobbles (consistent with low silica) of comparable size. The perched rock was undisturbed by the rover wheels while the darker ones were pressed into the soil (Figure 20). These examples and others throughout the Eastern Valley and east to Tyrone demonstrate that the nodular rocks commonly behave as coherent, undeformable masses. A straightforward interpretation is that they are outcrops that erode into nodular, cobble‐like forms. We have found evidence for the role of wind in shaping the silica‐rich outcrops (sections 3.5 and 4.2). However, no bedding or layering has been observed. Erosion of a massive, nonbedded rock unit could satisfy the observations, but the possibility exists that the nodular form is in part a primary characteristic rather than exclusively erosional. This issue is explored further in section 5. [56] The silica‐rich nodular outcrops commonly form meter‐scale long knobs or discontinuous ridges several to a few tens of centimeters high. Kobal is the largest known

20 of 48

E00F23


E00F23

Figure 19. Sol 1113 Navcam mosaic showing the silica‐rich outcrop that contains the Mini‐TES target Kobal (white circle ∼16 cm across). The highest points on this outcrop are ∼30 cm above the surrounding platy, buff‐colored outcrops seen here and elsewhere in the region. The inset frame shows an anaglyph from the same mosaic demonstrating that the rover did not deform the outcrop during its drive (toward the viewer) on sol 779. Rover tracks are ∼1 m apart. example and is visible from orbit with the High Resolution Imaging Science Experiment (HiRISE) (Figure 1). No preferred orientation of the ridges or knobs is apparent from orbital or surface images, but their relationship to an under-

lying buff‐colored, platy unit has been described previously [Arvidson et al., 2008; Squyres et al., 2008] and is discussed extensively in section 5. Their small scale makes them relatively obscure in the landscape, but the density of the nodular

Figure 20. Pancam false‐color image (sol 1164, P2584) of nodular outcrops comparable to the nearby (