An impact origin for hydrated silicates on Mars: A synthesis


  • Livio L. Tornabene,

    Corresponding author
    1. Centre for Planetary Science and Exploration, Department of Earth Sciences, Western University, London, Ontario, Canada
    • Corresponding author: L. L. Tornabene, Centre for Planetary Science and Exploration, Department of Earth Sciences, Western University, 1151 Richmond St., London, ON N6A 5B7, Canada. (

    Search for more papers by this author
  • Gordon R. Osinski,

    1. Centre for Planetary Science and Exploration, Department of Earth Sciences, Western University, London, Ontario, Canada
    2. Department of Physics and Astronomy, Western University, London, Ontario, Canada
    Search for more papers by this author
  • Alfred S. McEwen,

    1. Lunar and Planetary Laboratory, The University of Arizona, Tucson, Arizona, USA
    Search for more papers by this author
  • James J. Wray,

    1. School of Earth and Atmospheric Sciences, Georgia Institute of Technology, Atlanta, Georgia, USA
    Search for more papers by this author
  • Michael A. Craig,

    1. Centre for Planetary Science and Exploration, Department of Earth Sciences, Western University, London, Ontario, Canada
    Search for more papers by this author
  • Haley M. Sapers,

    1. Department of Natural Resource Science, McGill University, Sainte-Anne-de-Bellevue, Quebec, Canada
    Search for more papers by this author
  • Philip R. Christensen

    1. School of Earth and Space Exploration, Arizona State University, Tempe, Arizona, USA
    Search for more papers by this author


[1] Recent Mars-orbiting spectrometers continue to detect surface materials containing hydrated silicates, particularly clays and amorphous phases (e.g., silica glasses), concentrated within the heavily cratered Noachian highlands crust. This paper provides a review, summary, and synthesis of observations from terrestrial impact structures with current Martian data. It is suggested that numerous and frequent impacts into the volatile-rich silicate crust of Mars, through direct and indirect impact-generated mechanisms, represent a plausible hypothesis that can explain the widespread distribution of hydrated silicates in the surface and subsurface of the heavily cratered Noachian highlands crust largely independent of climate. In addition to impact-generated hydrothermal activity, devitrification, autometamorphism, and the voluminous production of impact “damaged” materials that are susceptible to alteration must be considered. When taken together, a drastically different early climate on Mars, in which water is stable at the surface for extended periods of time, cannot be ruled out; however, it is noted here that these additional impact mechanisms can operate and thereby extend the range of possible alteration settings to include climate conditions that may have been predominately colder and drier. Such a climate would not be dissimilar to the conditions of today, with the important exceptions of a higher geothermal gradient, and punctuated thermal disturbance to the cryosphere and hydrosphere from igneous activity and an exponentially higher impact flux.

1 Introduction

[2] During the early history of the solar system, impacts of all sizes occurred more frequently than at present [e.g., Hartmann, 2005; Strom et al., 2005]. During a period known as the late heavy bombardment [Turner et al., 1973; Tera et al., 1974; Cohen et al., 2000; Kring et al., 2002], impact basins thousands of kilometers in diameter formed (e.g., the ~1100 km diameter Imbrium Basin on the Moon and the ~2300 km Hellas Basin on Mars) greatly influencing the formation and geologic evolution of early planetary crusts. Radiometric dating of returned samples and meteorites from the Moon as well as crater populations on the ancient surfaces of the Moon, Mercury, Venus, and Mars suggest that the formation of large basins ceased after ~3.85 Ga (i.e., during the Noachian period on Mars) [Strom et al., 2005]. Interestingly, both valley networks [e.g., Carr, 2006] and hydrated silicates (i.e., H2O- and OH-bearing phases) are widespread within the ancient Noachian crust [Poulet et al., 2005; Bibring et al., 2006; Mustard et al., 2008; Ehlmann et al., 2009, 2011; Murchie et al., 2009; Wray et al., 2009; Carter et al., 2011a, 2013], in which this period of intense impact bombardment is preserved. Previous models suggest that the cumulative effects of heavy impact bombardment by mid-sized to basin-sized impacts may account for global climate change and the formation of valley networks [Segura et al., 2002, 2008], but it is unclear whether the duration and intensity of such climatic episodes were sufficient to cause widespread alteration; a more recent model by Segura et al. [2012] suggests that basin-sized impacts may generate a warmer and wetter climate on Mars as a consequence of a runaway greenhouse state and would persist until the escape process reduces water vapor in the atmosphere and forces the planet to return to a colder and drier climate.

[3] On Earth, hydrated silicates typically form from prolonged exposure of silicate rocks to abundant water. Therefore, their widespread detection within the heavily cratered Noachian highlands crust has often been interpreted as further evidence for an early warmer and wetter climate on Mars. However, Martian climate models fail to explain widespread stable water [e.g., Kasting, 1991; Squyres and Kasting, 1994; Colaprete and Toon, 2003; Gaidos and Marion, 2003], including a fainter early Sun [e.g., Gough, 1981; Kasting, 1991], and mounting geologic evidence suggests that only sporadic and transient water-related events occurred on early Mars [McEwen et al., 2007; Christensen et al., 2008]. If this is correct, then hypotheses for hydrated silicate formation on Mars should also include mechanisms that would operate under colder and drier surface conditions and perhaps not even require stable water at or near the surface for extended periods of geologic time. So the question is: is it possible to generate such pervasive alteration on a predominately cold and dry Mars?

[4] An analogous line of evidence for subsurface formation of hydrated phases under conditions that do not require warm and wet surface conditions comes from the CI, CR, and CM carbonaceous chondrite meteorites [e.g., McSween, 1999; Cloutis et al., 2011a, 2011b, 2012]. These aqueously altered meteorites originated in/on small atmosphere-free parent bodies that were abundant during planetary accretion. Interestingly, all of the carbonaceous chondrite varieties and the chondritic interplanetary dust particles related to them contain significant amounts of serpentine and smectite clays, which are common phyllosilicate minerals [Bearley, 2005; Mutschke et al., 2008]. The identification of phyllosilicates comes from the study of meteorites as well as the spectral observations of CI, CR, and CM like asteroids that may be the meteorites parent bodies [Bunch and Chang, 1980; Noguchi et al., 2002; Rivkin et al., 2006]. Clearly, and as evidenced from the petrographic studies of these meteorites, the decay of Al26, and oxygen isotopes [e.g., Clayton and Mayeda, 1977; Zolensky et al., 1989; Weisberg et al., 1993; Kallemeyn et al., 1994], these alteration materials did not form under warm and wet surface conditions in the presence of an atmosphere. Although this comparison is likely extreme for Mars, it serves to prove that hydrated silicates do not necessarily require warm and wet surface conditions to be produced.

[5] One recent suggestion is that while Mars may have been largely cold and dry on the surface, extensive aqueous alteration may have occurred in the warmer, deep subsurface and was later exposed by subsequent impacts [Ehlmann et al., 2011]. But is endogenic heat supplied to the global hydrosphere and cryosphere on early Mars sufficient to drive these systems? The dominance of impact cratering on early Mars and the observation that hydrated silicates appear to spatially correlate with the heavily cratered portions of Martian crust suggest the possibility of a genetic link between the impact process and formation of hydrated silicates. Indeed, ongoing observations continue to indicate that most of the occurrences of hydrated silicates are associated with impact craters. Previous studies explain this correlation in either one of two ways: via impact “sampling” (i.e., excavation and/or uplift of preexisting hydrated materials) and/or impact-generated hydrothermal alteration, with these two explanations not necessarily being co-incident [Poulet et al., 2005; Bibring et al., 2006; Ehlmann et al., 2009, 2011; Schwenzer and Kring, 2010; Carter et al., 2011b; Schwenzer et al., 2012]. Given our current understanding of terrestrial impact structures, both mechanisms are indeed plausible, but they are not mutually exclusive and thus often occur together in the same impact structure. For example, there are terrestrial impact structures with clay-rich sedimentary lithologies (e.g., shales) exposed in their central uplifts, but these also provide evidence of overprinting hydrothermal activity [Naumov, 2005; Osinski, 2005; Osinski et al., 2012]. There are also examples where ejecta have both excavated preexisting clays and generated them through a variety of impact-generated mechanisms [Hörz et al., 1983; Osinski et al., 2004; Osinski, 2005; Muttik et al., 2008]

[6] Previously, most studies dealing with the impact association of hydrated silicates on Mars have typically only considered alteration via the development of hydrothermal systems within individual impact craters on Mars [Newsom, 1980; Berkley and Drake, 1981; Allen et al., 1982; Schwenzer and Kring, 2010; Osinski et al., 2012; Schwenzer et al., 2012]. One of the major goals of this paper is to account for all possible alteration pathways developed as a consequence of the impact process and to draw attention to the role of impact ejecta deposits in distributing material widely. In addition to alteration from the development of a sustained hydrothermal system, the spatial, textural, and mineral assemblage phase relationships (including various hydrated silicates) within specific impactites collected from well-studied terrestrial impact sites [Naumov, 2005; Osinski, 2005; Osinski et al., 2012] indicate alteration of the target rocks via devitrification of hydrated impact glass-bearing materials, autometamorphism, and post-impact alteration of variously shocked (i.e., “impact-damaged”) materials. Devitrification refers to the solid-state transformation/crystallization of amorphous materials into crystalline materials—commonly recognized in volcanic glasses (e.g., “snowflake” obsidian is volcanic glass that has partially recrystallized into crystobalite). Autometamorphism pertains to the in situ metamorphism of rocks by the action of their own heat and composition, which includes volatiles [Osinski, 2005]. These two impact-induced alteration processes may have the most important implications for early Mars, particularly devitrification, because they operate under limited exposure to post-lithification alteration fluids (e.g., short lived, limited, or even in the absence of a hydrothermal system). These two alteration mechanisms are emphasized herein, including some discussion related to impact-generated hydrothermal systems. For more details specifically on the role of impact-generated hydrothermal on the alteration of rocks on Earth and Mars, please see Osinski et al. [2012].

[7] This paper provides a review and synthesis of these various impact alteration mechanisms and their byproducts based on terrestrial field studies that include many of the same alteration phases identified on Mars. The main idea being that the numerous impacts into the early Martian crust may have greatly contributed to the formation, burial, exposure, and re-distribution of hydrated materials within the Noachian crust. This hypothesis, referred to hereafter as the impact alteration model, is supported by six observations that we emphasize in this contribution: (1) Mars possesses a volatile- and water/ice-rich silicate upper crust; (2) numerous hypervelocity impacts have provided abundant energy to the crust resulting in pressure-temperature (P-T) conditions that cause disequilibrium and thermodynamically irreversible changes to target materials; (3) the most common hydrated silicates formed by impact-generated hydrothermal activity, autometamorphism and devitrification in terrestrial impact structures are similar to those most commonly observed on the surface of Mars; (4) Martian hydrated silicates are concentrated in regions of thick, highly elevated sections of the heavily cratered highlands crust that are circumferential to some of the largest Martian impact basins, and generally occur in mid-sized to large craters that superimpose them; (5) some hydrated silicates on Mars are associated with the occurrence of megabreccias: and (6) there is a close stratigraphic relationship and sharp delineation between deposits rich in hydrated silicates and unaltered volcanic rocks. We explore each of these observations in further detail below, and at the end of the paper, we include a discussion and summary of some considerations on how the impact alteration model may be tested with data collected in situ.

2 Synthesis of Hydrated Silicate Occurrences From Terrestrial Impact Studies and Martian Observations

2.1 A Volatile-Rich, Silicate Upper Crust on Mars

[8] Hypervelocity impacts into Mars' volatile- and water/ice-rich upper crust provides three essential components for hydrated silicate formation: fluids, silicates, and heat. A volatile-rich crust not only provides fluids for impact-generated hydrothermal systems, but also allows for the generation of hydrated impact melt glasses that are highly altered in many terrestrial impact structures [Naumov, 2002, 2005; Osinski, 2005; Osinski et al., 2012]. Several lines of evidence suggest that the silicate upper crust of Mars is volatile-rich and that it acquired volatiles during its early formation [e.g., Carr, 2006]. Theoretical calculations suggest that Mars acquired less water during its formation than Earth, but still a significant amount, equivalent to ~6–27% of Earth's present oceans [Lunine et al., 2003]. Presently, Mars is cold and dry and its regolith is presently frozen; liquid water at the surface is generally unstable (only possible under highly localized and transient conditions). Water on Mars in the form of surface and/or subsurface ice is supported by observational evidence, which includes (1) orbital remote sensing data, including both gamma ray and neutron spectroscopy [Boynton et al., 2002; Feldman et al., 2002; Holt et al., 2008] and both visual and visible-near infrared (VNIR) spectral evidence [Byrne et al., 2009]; (2) geomorphological landforms that are indicative of glacial and/or ground ice (e.g., lineated valley fill, polygons, thermokarst depressions) [Costard and Kargel, 1995]; and (3) direct surface observations and sampling of water-ice by the Phoenix mission [Smith et al., 2009]. In addition, layered ejecta blankets [e.g., Carr et al., 1977; Squyres et al., 1992; Barlow, 2005] and the recently discovered crater-related pitted materials suggest that volatiles are not just a shallow subsurface phenomenon (<1 km) and are likely accessible at deeper (>1 km) levels [e.g., Tornabene et al., 2012a]. Models of the global cryosphere/hydrosphere of Mars support this suggestion [Clifford, 1993; Clifford et al., 2010], while small-scale channels with multiple distributaries associated with well-preserved Martian craters suggest that the mobilization of target volatiles likely occurs during the impact process [Williams et al., 2004; McEwen et al., 2007; Tornabene et al., 2008a; Morgan and Head, 2009; Williams and Malin, 2009; Harrison et al., 2010; Jones et al., 2011].

2.2 Formation of Impactites and Impact “Damaging” of Target Rocks on Earth and Mars

[9] Hypervelocity impacts generate P-T conditions that can vaporize, melt, metamorphose, and/or deform a substantial volume of rocks within the target sequence [e.g., Melosh, 1989]. Such effects directly give rise to various alteration paths and provide voluminous “damaged” target materials that are generally more susceptible to alteration than the surrounding unaffected country rocks. The volume of the target affected by an impact scales exponentially with the total impact energy; larger craters produce substantially higher volumes of impact-damaged materials—collectively termed impactites—than smaller ones [e.g., Melosh, 1989; Grieve and Cintala, 1992; Cintala and Grieve, 1998].

[10] At lower shock pressures (generally <10 GPa), rocks are faulted and fractured, and features such as shatter cones, as well as both planar fractures and planar deformation features in individual mineral grains, are produced (for a recent review of shock metamorphic effects, see French and Koeberl [2010]). Fractures and faults are particularly prevalent in crater rims, wall terraces, central uplifts, and subcrater basement rocks [Melosh and Ivanov, 1999; Collins et al., 2004, 2005; Collins, 2013; Osinski and Spray, 2005]. The extent of the fracture system in terrestrial structures, ascertained from drill cores and various remote sensing and geophysical techniques (e.g., gravity anomalies, seismic surveys), may extend up to one crater diameter beyond the topographic crater rim [Pilkington and Grieve, 1992; Smith et al., 1999; Collins et al., 2004, 2005; Collins, 2013; Gurov et al., 2007]. This extensive system of fractures, faults, and associated incipient brecciation is specifically relevant to promoting alteration of impact-damaged rocks as these effects increase porosity and permeability [e.g., Collins et al., 2004, 2005; Collins, 2013] in the target by providing subsurface conduits for enhanced fluid flow. For example, a terrestrial impact structure tens of kilometers in diameter has an estimated volume of ~600 km3 of target rock with enhanced permeability [Naumov, 2005]. Brecciation (i.e., comminution) also enhances alteration by increasing the surface area-to-volume ratio of materials due to the production of clasts with increasingly smaller grain sizes. Recent, shock and recovery experiments performed on olivine in the presence of water indicate that shock effects such as comminution and mosaicism may enable alteration of olivine into serpentine without invoking post-impact alteration mechanisms [Furukawa et al., 2011].

[11] At higher shock pressures (~30–45 GPa), diaplectic glasses (e.g., maskelynite) form from the shock disorder/amorphization of minerals, particularly framework silicates (e.g., quartz and plagioclase) [e.g., Chao, 1967; Engelhardt and Stöffler, 1968; Stöffler, 1984; French, 1998; French and Koeberl, 2010]. Diaplectic glasses are homogenous mineral glasses formed from quenching of a high-pressure phase before the shock pressure is released and without mineral melting [e.g., Stöffler, 1984]; hence, they do not show any indication of disruption or flow and characteristically retain the outer form of the original crystal [e.g., French and Koeberl, 2010]. However, because they differ from melt glasses with respect to their texture, density and structure (e.g., they are known to retain short-range structural order, unlike thermally produced melt glasses) [Stöffler, 1984; French, 1998; French and Koeberl, 2010], they may alter differently from glasses that are created in the lab or formed from endogenic processes.

[12] At shock pressures above ~40 GPa, temperatures exceed the melting temperature, with most common rock-forming minerals melting at pressures >60 GPa [Grieve et al., 1977]. Collectively, amorphization and melting produce abundant impact glasses, which range from glassy fragments, included in various impact breccias, to extensive meters to kilometers thick coherent bodies of igneous-textured impact melt rocks (Figure 1). Amorphous phases and glasses are metastable and are amongst the first phases to be altered in terrestrial impact structures [Naumov, 2005]. Impact melt rocks also range from anhydrous to hydrous, which is relevant with respect to how they alter and will be discussed in the following section. They are found within a variety of settings within and around the source crater, but are concentrated within three main settings: (1) within the interior crater-fill deposits (including both the crater floor and terraces of complex craters); (2) within, or superimposed on, the impact ejecta deposits; and (3) within dikes that crosscut the subcrater floor and central uplift of complex craters (Figure 1).

Figure 1.

Distribution and characteristics of terrestrial impact melts and hydrated silicate-bearing lithologies. (a) Cross section of a complex crater in the ~10–200 km size range showing the distribution of melt-bearing impactites (red) [after Osinski, 2006 and references therein]. Letters for each sub-image mark the approximate locations of the various melt and hydrated silicate-bearing lithologies in the structure. (b) Hydrated silicate-rich impact breccias (suevites) (yellow-green) overlying melt-free ballistic ejecta (red-brown) at the ~24 km diameter Ries impact structure, Germany. (c) Close-up of suevite; impact-derived hydrated silicates (clays) comprise ~70 vol % of the groundmass of this sample. The black clasts are impact glasses; 6 cm camera lens cap for scale. (d) Coherent impact melt rocks from the interior of the ~28 km diameter Mistastin impact structure, Canada, overlying hydrated silicate-rich impact breccias (suevites). Yellow-dotted line demarks the approximate boundary between the two units (image courtesy of Cassandra Marion). (e) Coherent impact melt rocks intercalated with hydrated silicate-bearing impact breccias/megabreccias (suevites) inside the crater rim of the ~100 km diameter Popigai impact structure, Russia (image courtesy of Richard Grieve). (f) Close-up of suevites, which underlie coherent impact melt rocks shown in Figure 1d. The black clasts are flow-banded impact glasses; 6 cm camera lens cap for scale (image courtesy of Cassandra Marion). (g) Clast-rich, glassy impact melt dike (arrows) crosscutting the rocks of the crater floor of the Mistastin structure. Field notebook for scale (image courtesy of Cassandra Marion).

[13] Impact melt rocks can represent a significant component of the impactites produced by impact events, as the volume of melt scales exponentially with crater diameter [Grieve and Cintala, 1992; Cintala and Grieve, 1998]. Generally, the melt volume (Vm; in km3), based on a chondritic impactor traveling at 10 km s−1, follows the relationship of

display math(1)

where Dt is the transient crater diameter in kilometers is calculated from the final rim diameter using the Croft [1985] equation and using a simple-complex crater transition diameter of ~7 km [e.g., Boyce and Garbeil, 2007; Stewart and Valiant, 2006]. This scaling relationship implies that a crater hundreds of kilometers in diameter would produce thousands of cubic kilometers of impact melt. For such large-scale impact events, the melt volume produced exceeds the depth of the transient crater as it overtakes the displaced zone and extends further into the excavation zone (i.e., very large craters produce both melt-rich crater-fill and ejecta deposits). Such large impact craters, which formed more frequently in the Noachian period [Hartmann, 2005; Strom et al., 2005], produced the largest contribution of impact melts to the early Martian crust. Hence, a modeled cumulative crater melt production for an early Noachian Mars (~4 Ga), based on this relationship and the Hartmann (2004 iteration) production function for craters >10 km [Hartmann, 2005], is estimated at ~3.61 × 107 km3, or ~25% of the upper 1 km of heavily cratered early Noachian crust. However, note that other abundant impact byproducts formed from lower shock levels, which are still relevant with respect to the alteration history of surface materials, are not included in this estimate. Furthermore, heat from numerous impacts, and/or active volcanism would have disrupted the cryosphere/hydrosphere, mobilizing fluids and further enhancing alteration kinetics, especially for the abundant metastable impact-damaged materials contributed by impacts to the early crust. Schwenzer et al. [2012] recently demonstrated the plausibility of this idea by modeling the consequence of the formation of mid-sized impact craters (tens of kilometers in diameter) into a crust with a 2–6 km thick cryosphere.

[14] Overall, silicate glasses are more susceptible to aqueous alteration than crystalline materials [Berger et al., 1994, 2002; Browning et al., 2003; Stopar et al., 2006; Hausrath et al., 2008]; this includes impact-generated glasses such as diaplectic glass [e.g., Stöffler, 1984] and particularly, hydrated impact melt glasses [e.g., Naumov, 2005; Osinski, 2005]. Alteration models for simulated anhydrous Martian glasses (mafic composition) show that clays (>5 wt %) can form within days, but only when starting with small grain sizes (<1 mm), higher pH (>8), and higher temperature (>100°C) conditions [Browning et al., 2003]. The process is much slower at cooler surface temperatures of ~5°C (still warm for Mars), taking ~105 years. Under very low or moderate pH conditions (2 > pH > 8) [Hausrath et al., 2008], most crystalline materials react at rates that are at least 1 order of magnitude slower than glasses, requiring up to 106 years to alter into clays [Stopar et al., 2006]. However, it has been shown in experiments that thick glass deposits exhibit a marked decrease in dissolution rate with time due to the formation of a leached zone [e.g., Leturcq et al., 1999], which has implications with respect to the aqueous alteration of thick non-permeable coherent impact melt sheets. However, porous and permeable volatile-rich impact melt breccias (suevite) may be more common on Mars than thick coherent melt sheets [Pope et al., 2006; Boyce et al., 2012; Tornabene et al., 2012a], which would not form a coherent layer of leached glass.

[15] Unfortunately, experiments on the alteration of glass have only been conducted on anhydrous synthetic and volcanic glasses, which are texturally, compositionally, and structurally distinct from both diaplectic and impact melt glasses [e.g., French and Koeberl, 2010]. Depending on the target composition, impact melt glasses are much more diverse in composition; as such, experiments comparing the dissolution and alteration rates of volcanic and various impact glasses (diaplectic and both anhydrous and hydrous melt glasses and breccias) are needed to truly assess the susceptibility of impact various glasses to aqueous alteration. Regardless, these experiments do demonstrate the susceptibility of glasses versus minerals to aqueous alteration. Because the upper Noachian crust on Mars possesses the highest volume of impact glass and impact-damaged materials, it is most likely that the alteration of these impactites played an important role in the production of the abundant and widespread hydrated silicates we observe today. This is likely whether or not the subsequent mechanisms that formed the hydrated silicates in these impactites were directly or indirectly related (e.g., endogenic forms of heating and aqueous activity).

3 Hydrated Silicates in Terrestrial Impact Structures: Mechanisms, Types, and Relationships to Those Observed on Mars

[16] A synthesis of the common secondary mineral groups or alteration assemblages within 70 terrestrial impact structures ranging from ~1.8 to 250 km in diameter is summarized in Table 1 [see Naumov, 2002, 2005; Osinski et al., 2012]. Although various hydrated silicates occur as cavity and fracture fillings within parautochtonous rocks of the central uplifts, subcrater floor, terrace, and rims, the most common occurrences are within crater-fill impactites and ejecta deposits. In fact, they are particularly abundant when associated with highly shocked and melt-rich materials, particularly the amorphous hydrated melt glasses generated during the impact process [Naumov, 2005; Osinski, 2005; Osinski et al., 2012].

Table 1. Common Hydrated Silicates on Mars and in Terrestrial Impact Structuresa
Hydrated SilicatesExample LocalitiesOccurrences: MartianTerrestrial
  • a

    (Martian) Mustard et al. [2008], Murchie et al. [2009], Ehlmann et al. [2011], and Carter et al. [2013]; (Terrestrial) Naumov [2005], [Osinski, 2005; Osinski et al., 2012].

Fe-Mg-rich smectitesNili, Mawrth (Fe-rich), Tyrrhena, Sabaea, Sirenum, Noachiscraters, megabreccias, both massive and layered fine-grained outcropsmelt-bearing bodies, breccia deposits, and target rocks (including the central uplift)
Chlorite/PrehniteNili, Marineriscraters, particularly central uplifts; massifsall rock types and crater localities, but most commonly associated with lower zone of central uplift, and high-T hydrothermal deposits
Al-rich smectiteMawrthlayered fine-grained outcropsmelt-bearing bodies, breccia deposits, and target rocks (including the central uplift)
Silica glassesMawrth, Nilicraters, layered fine-grained outcropsremnant unaltered hydrated impact glasses (injection dikes into the target rocks—including the central uplift, crater-fill, and ejecta deposits)
ZeolitesNilicraters, massifsall rock types and crater localities, but most commonly associated with veins, upper zone of central uplifts, and low-T hydrothermal deposits
Illite-muscoviteNilicraterscommonly associated with lithic breccias and target rocks, but may also occur in melt-bearing breccias

[17] Studies of impact melt rocks and breccias from several terrestrial craters show that these hydrated silicates are produced by a variety of mechanisms. Impact-generated hydrothermal alteration is perhaps the most documented and best understood mechanism in terrestrial impact structures (see review by Osinski et al., [2012]). As on Earth, large volumes of melt and geothermal heat from uplifted rocks produced by sizable impacts into a volatile-rich target, in addition to increased permeability from fracturing and comminution, are expected to generate a sustained hydrothermal system on Mars [Newsom, 1980; Newsom et al., 1996; Rathbun and Squyres, 2002; Abramov and Kring, 2005; Schwenzer and Kring, 2010; Osinski et al., 2012; Schwenzer et al., 2012]. Models of impact-generated hydrothermal systems on Mars suggest that they may persist for ~104 and up to 107 years for a ~30 km diameter crater and a Hellas Basin-sized event (D ~ 1000 km+), respectively [Abramov and Kring, 2005], providing ample time for the alteration of impactites and their immediate surroundings.

[18] The extent of influence of hydrothermal alteration can be discerned from the models and also observations from the field. The circulation cells in models [Abramov and Kring, 2005] and the depth of alteration observed in drilled terrestrial structures can span on the order of ~10% of the crater diameter in depth (e.g., 5 km deep for a 50 km sized structure) [Naumov, 2005]. However, it is important to note that current models do not account for the heat contributed from substantial volumes of melt emplaced in the ejecta of very large impact craters (>100 km), which are needed to properly gauge the extent of influence of large impacts on the alteration history of the heavily cratered southern highlands.

[19] Not all impact-related alteration associated with terrestrial impact structures comes from hydrothermal alteration; we delve into several examples from the Ries impact structure, located in southern Germany, which is consistent with other impact sites (for more details specifically on the role of impact-generated hydrothermal systems on the alteration of rocks on Earth and Mars, please see Osinski et al. [2012]). Ries is ~24 km in diameter and ~14 Myr old, and is one of the best preserved and best exposed impact structures on Earth [e.g., Osinski, 2005]. At the Ries, hydrated silicates are concentrated within impact melt-bearing breccias or so-called “suevites” (Figure 2), which comprise the uppermost component of the ejecta deposits (surficial suevite; Figures 1b, 1c, and 2) and the bulk of the crater-fill deposits (crater suevite; Figures 1d–1f) [Engelhardt, 1990; also see Osinski, 2005, Figure 2]. The Ries suevites, as characterized by Engelhardt [1990], consist of lithic and mineral clasts in a groundmass that occupies ~80 vol % of the sample, which in turn is dominated by melt glass and abundant phyllosilicates. Compositional and textural analyses of both surficial and crater suevites indicate that they were altered via a combination of processes including devitrification or autometamorphism, impact-generated hydrothermal activity, and post-impact diagenesis [e.g., Osinski, 2005]. Hydrous impact glasses in Ries surficial suevites are significantly altered, but less severely when compared to crater-fill suevites. Indeed, some pristine and unaltered glasses can be found, but these impactites still typically contain secondary phases including Al-Fe montmorillonite, zeolites, and calcite that comprise up to ~70 vol % of their groundmass [Osinski et al., 2004; Osinski, 2005]. These phases have been suggested to form by post-impact hydrothermal alteration in the ejecta blanket [e.g., Muttik et al., 2008], but this is at odds with the presence of some pristine and relatively unaltered glasses when compared to crater-fill suevites.

Figure 2.

Various hydrated silicate-rich impact melt breccias (suevites) from the Aumuhle quarry, Ries impact struture, Germany. Scale bar equal to 2 cm. (a) Sample Ries/003. Note the large glass clast with a flow-banded texture surrounded by a green-grey matrix. (b) Sample Aumuhle RI 01 whole rock. This is a flow-banded glass clast. (c) Sample RI_09_008b glass clast. Note the beige-colored alteration of the originally dark grey glass. (d) Sample Aumuhle RI 01. White materials in this sample represent a mixture of partially devitrified melt glass into hydrated phyllosilicates. (e) Sample Aumuhle 2005. Note the color and texture of this minimally altered glass. (f) Sample Aumuhle September 2005. Various fragments of glass-bearing breccias. Note the sharp margin between the dark glass clasts and the beige clastic matrix.

[20] Recent textural and compositional analyses suggest that the phyllosilicates in the surficial suevites are largely the product of devitrification or autometamorphism, with a minor hydrothermal and digenetic component (~10–15 vol %; Figure 3) [Osinski et al., 2004; Osinski, 2005]. Hydrated glasses can devitrify in the solid state into phyllosilicates (e.g., smectites) because all the necessary elements to form these crystalline hydrated silicates are already included in the melt glasses (e.g., Si, Al, Fe, Mg, Ca, Na, K), including sufficient H2O [Stöffler, 1984; Engelhardt et al., 1995; Osinski, 2003]. Unlike endogenic glasses (e.g., volcanic), impact glass may become oversaturated with water and internally strained due to rapid adiabatic cooling on decompression [Engelhardt et al., 1995], and thus may incorporate significant amounts of H2O from the target materials (e.g., in some cases up to ~20 wt %) [Stöffler, 1984; Engelhardt et al., 1995; Osinski, 2003]. In fact, the bulk of the phyllosilicates in the Ries surficial suevites derived their H2O from the hydrated glasses themselves and not an external source [Osinski et al., 2004]. As such, in lieu of impact-generated hydrothermal activity, residual heat and volatiles from these melt-bearing deposits will enhance and facilitate both devitrification [e.g., Marshall, 1961] and autometamorphic reactions [e.g., Osinski, 2005].

Figure 3.

Comparison of hydrated silicates formed from devitrification/autometamorphism and hydrothermal alteration (see Osinski et al. [2004] for a complete discussion of the origin of these hydrated silicates). (a) Globules of devitrified clay (Dv-c) and interstitial volatile-poor silicate glass (Gl) showing textural evidence for deformation and flow. A large vesicle is present in the right of the image and a calcite (Cc) melt globule is present in the left of the image. (b) Spectacular flow textures developed between volatile-poor silicate glasses (Gl) and what are now clay minerals (Dv-c), indicating that both phases were in a fluid state at the same time. These textures in Figures 3a and 3b are not consistent with alteration by hydrothermal fluids, but are indicative of recrystallization in the solid state (devitrification) of hydrated impact melt glasses into clays. (c) Fine-grained groundmass-forming montmorillonite clays formed via devitrification of hydrated glasses (Dv-c) crosscut by stacks of platy montmorillonite formed from hydrothermal alteration (Hy-c), which displays clear open-space cavity filling textures. (d) Another example of stacks of platy hydrothermal montmorillonite (Hy-c).

[21] Although hydrothermal alteration is minor in the Ries ejecta deposits, this may not be the case for very large complex craters; larger craters (>100 km), which incorporate greater volumes of impact melt in their ejecta, may develope an ejecta blanket-based hydrothermal system. It is important to note that the presence of preserved hydrated glasses at Ries is also likely a function of its relatively young age (i.e., ~14 Ma versus ~Ga), and that over extended periods of time, these materials are expected to devitrify further into phyllosilicates [e.g., Marshall, 1961]. Moreover, volatile-rich impact melts may be very common on Mars throughout its history [Boyce et al., 2012; Tornabene et al., 2012a]; as a consequence, alteration of hydrous impact glass into clays may be particularly relevant on Mars.

[22] Studies of the Ries surficial suevite indicate that Al-Fe montmorillonite and halloysite are the most common devitrified phases (note that the preferred cation composition of these phases reflects the predominant Si/Al-rich target sampling/composition for the melt component at Ries) [Osinski et al., 2004; Osinski, 2005; Muttik et al., 2008]. Considering these phases may occupy up to 70 vol % of the groundmass-dominated surficial suevite, it is not surprising that the VNIR bulk rock spectra are dominated by these phases (Figure 4). We collected bulk rock spectra of surficial suevite samples (see Figure 2) with a FieldSpec Pro FR Portable Spectroradiometer produced by Analytical Spectral Devices, Inc. The FieldSpec Pro possesses three sensors over a range of 0.35–2.5 µm and has a spectral resolution of 10 nm. Multiple rock spectra of surficial suevite samples were then matched to mineral library spectra in the Environment for Visualizing Images (ENVI) using the provided mineral and rock spectral libraries (e.g., U.S. Geological Survey (USGS) mineral and rock spectral libraries). Figure 4 shows surficial suevite bulk rock spectra compared to USGS mineral library spectra. These spectra are comprable to the features associated with Al-montmorillonites, but lack any sign of the doublet associated with the kaolin group mineral halloysite [e.g., Ehlmann et al., 2009]. This spectrum is generally consistent with previous VNIR spectra of suevites from the Ries and Brent impact structures [Allen et al., 1982], and is even comparable to some of the Al-Fe-smectite spectra collected from Mars [cf. Ehlmann et al., 2009, Figure 7].

Figure 4.

NIR spectra of Ries hydrated silicate-rich impact melt breccias (suevites) from Figure 2 (dashed lines), compared to spectrally pure montmorillonite samples from the USGS spectral library (solid lines). In each panel, the lowermost spectrum has been normalized to 1 at 1.65 µm, and each spectrum above it has been offset +5% at 1.65 µm. Diagnostic absorption features are noted at ~1.41 due to OH, ~1.9 due to H2O, and ~2.2 µm which is a combination of Al-OH/Mg-OH and OH absorptions [Clark et al., 1990; Bishop et al., 1993; Cloutis et al., 2008]. (a) NIR spectra, representative of terrestrial hydrated phyllosilicates. (from top to bottom) A breccia sample from the Aumühle quarry, Ries; the spectrum of a sliced section of an impact bearing melt breccia (surficial suevite) hand specimen collected from the ejecta blanket of the Ries impact structure, and the USGS montmorillonites CM27 and Stx-1 [Clark et al., 1990]. (b) NIR spectra, akin to what one might expect from the same types of phyllosilicates, but desiccated as they might be on Mars [Bishop and Pieters, 1995; Cloutis et al.; 2008]. (from top to bottom) Four samples collected from the Aumühle quarry, Ries. These spectra also exhibit contributions from the presence of other metal-OH bonds (likely Fe-OH) that are observed in the ~2.1 and ~2.23 to 2.3 µm regions [Bishop et al., 1993; Cloutis et al., 2008]. Fe-rich montmorillonite has been confirmed by previous microprobe analyses in the surficial suevite sample. Compositional and textural evidence suggest that these matrix-dominated samples (up to 90 vol % in some cases) are predominantly devitrified clay (up to ~70 vol %) (see Figure 2a) [Osinski et al., 2004; Osinski, 2005].

[23] The dominance of the Al-Fe montmorillonite spectral signature in these suevite samples demonstrates that a similar smectite clay signature could be readily detectable from orbit given good exposures that are free from coatings and other surface mantles within the bounds of the spatial and spectral resolution of orbiting spectrometers. Thus, we submit that impact melt-bearing breccias such as those from the Ries structure should be studied extensively in the field and the lab as analogs for Martian hydrated silicate-bearing rocks, so that their physical, compositional, and spectral properties can be recognized and applied to future roving missions.

[24] Underlying the melt-rich surficial suevite at Ries is the Bunte Breccia, which is a poorly sorted, melt-free polymict breccia, derived predominantly from the uppermost sedimentary target lithologies and interpreted to be the ballistic component of the Ries ejecta [Hörz et al., 1983]. Some excavated preexisting phyllosilicates are observed within this unit. These Bunte Breccia clays have been traced to the pre-impact target materials that were incorporated into the Bunte Breccia from outside the initial transient cavity by the action of secondary cratering formed via the impact of primary ejecta [Hörz et al., 1983]. They are not, therefore, linked with impact-generated alteration mechanisms. This also serves as an excellent example of how craters can excavate preexisting clays on Mars and not destroy them (i.e., due to excavation flow, not all ejecta are highly shocked and heated).

[25] Moving to the interior deposits, the alteration of the Ries crater-fill suevites is extensive and no unaltered hydrated glasses remain. The alteration minerals included in Ries crater suevite includes phyllosilicate clays (kaolinite, montmorillonite, mixed layer clays, illite, saponite, chlorite), zeolites (analcite, clinoptilollite, erionite, phillipsite, stilbite), carbonates (calcite), feldspars (K-feldspar, albite), sulfides (pyrite), and iron hydroxides [Osinski, 2005]—many of which are consistent with commonly detected phases on Mars (see Table 1). The alteration of the crater-fill suevites are more extensive than the surficial suevites. This is a consequence of the comparatively greater heat budget from higher melt volumes found in the crater-fill and the presence of an overlying crater lake, thus allowing alteration of all three types, particularly hydrothermal, to persist for a longer period of time than in the surficial suevites [Osinski, 2005]. This suggests that the predominant alteration mechanism and the extent of the alteration associated with impact cratering are largely dependent on heat sources (e.g., from the melt volume/concentration and thickness) and water availability.

[26] Because some impact melt glasses are inherently volatile rich, devitrification may continue to form hydrated silicate phases, albeit at much slower rates, even after heat and fluid circulation has ceased [Marshall, 1961]. This is an important point that has implications for the Martian climate history and the timing of alteration. This is because devitrification of hydrous silicate impact melt glasses can occur even under a prevailing colder and drier climate, which may provide a possible explanation for the widespread distribution of smectite-rich rocks within the heavily cratered Noachian crust. Hydrated glasses will completely devitrify within ~109 years of their formation even if no additional heat and fluids are introduced to the glasses after they are formed [Marshall, 1961]. As a consequence, the timing of “cold and dry” devitrification implies that although the host rock may be Noachian in age, the hydrated silicates formed within them may, in some cases, be formed much more recently.

[27] Many terrestrial impact sites, despite variations in target composition, develop alkaline hydrothermal environments (pH ~6–8) that typically form an alteration assemblage of smectite-zeolite with calcite-pyrite, including dioctahedral and trioctahedral smectites, some mixed layer smectite-chlorite, and zeolites [Naumov, 2005]. Given that the bulk of the Martian crust is mafic to ultramafic in composition [e.g., Christensen et al., 2008], abundant Fe/Mg/Ca-rich hydrated phases are predicted to form. Higher temperature phases (e.g., chlorite and anhydrite with Fe/Ca silicates such as prehnite) are also expected, but may be rarer because the cooling history of the impact-induced hydrothermal system occurs predominantly below 350°C, so that these phases are typically overprinted retrograde by reactions forming the lower temperature smectite-zeolite-calcite-pyrite assemblage [Naumov, 2005]. Lower water/rock ratios under Martian conditions may produce high-salinity fluids resulting in carbonate and sulfate precipitation followed by Fe-rich smectites and halloysite. The most commonly detected hydrated silicates on Mars are Fe/Mg-rich smectites (~89% of all occurrences) [Carter et al., 2013], followed by chlorite/prehnite, Al-rich smectites (see Table 1; see summary by Murchie et al. [2009] and recent updates from Ehlmann et al. [2011] and Carter et al. [2011b, 2013]. If Al-rich hydrated silicates are combined (e.g., smectites, micas, and kaolins), then this group constitutes the second most abundant occurrence (~17%) after the Fe/Mg-rich smectites (~49%) [Carter et al., 2013]. There are several cases of kaolin group clays (possibly halloysite), illite and/or muscovite, zeolites, and carbonates [Mustard et al., 2008; Ehlmann et al., 2008, 2013; Murchie et al., 2009; Niles et al., 2013], all of which are consistent with the common assemblages observed in terrestrial impact structures (Table 1). Mg-Fe carbonates have been observed in Martian meteorite ALH8400l and were also recently discovered in situ on Mars by the Spirit rover [Morris et al., 2010; Niles et al., 2013]. Mg-bearing serpentine has been recently reported on Mars, but is quite rare [Ehlmann et al., 2010]. This contrasts with the results of recent geochemical models of Martian hydrothermal systems [Schwenzer and Kring, 2010], which suggest that serpentine should be an abundant phase. Smectites, rather than serpentine, dominate terrestrial impact structures, including those produced within mafic targets [Naumov, 2005], and where they are commonly produced by other impact-generated mechanisms (i.e., devitrification and autometamorphism) as well as hydrothermally. Interestingly, kinetic models dealing with the formation and continued diagenesis of Martian materials indicate that “water did not persist much beyond their [clays] initial deposition” and that the clay deposits on Mars are “diagenetically underdeveloped” and “juvenile” [Tosca and Knoll, 2009]. This is consistent with the impact alteration model and the mechanisms described herein.

3.1 Hydrated Silicate Associations With Circumbasin Terrains and Their Superimposed Impact Craters

[28] Observations of Martian hydrated phases continue to be detected by the Observatoire pour la Mineralogie, l'Eau, les Glaces et l'Activitie´ (OMEGA) and Compact Reconnaissance Imaging Spectrometer for Mars (CRISM) VNIR spectrometers [Bibring et al., 2006; Murchie et al., 2009]. Ongoing detections of hydrated silicate phases continue to support the general observation that many of these phases are associated with the ancient heavily cratered southern highlands of Mars [Carter et al., 2011b, 2013; Ehlmann et al., 2011]. In particular, they occur in regions of thick, highly elevated sections of the heavily cratered southern highlands crust that are circumferential to some of the largest Martian impact basins (e.g., Hellas and Isidis) formed during the Late Heavy Bombardment (Figures 5 and 6). Lunar and terrestrial evidence indicates that the proximal ejecta from these basins may in some cases extend to hemispherical or nearly global scales [e.g., Lowe et al., 1989; Haskin, 1998; Smit, 1999; Korotev, 2005]. Generally, this concentration of hydrated silicates associated with circumbasin terrains, especially within overprinting craters (Figures 5 and 6), supports the likelihood that many of these hydrated silicates occurrences may be predominately preexisting (i.e., excavated) in origin. Additional evidence of large impacts playing a role in the distribution of hydrated phases comes from the differentiated body Vesta [Prettyman et al., 2012], where it has been suggested that concentrations of hydrogen-rich terrains may be associated with hydrated minerals ejected from the Rheasilvia and the Veneneia impact basins. Despite this, it is important to note that the overprinting impact in these regions may still be excavating materials that were altered by impact-related mechanisms (i.e., basin formation). Based on previous discussions, a reasonable assumption is that the circumbasin high-elevation terrains contain ejecta composed of meters to kilometers thick melt-rich and impact-damaged rocks produced by the basin formation event, and such rocks would be subject to the alteration as described herein.

Figure 5.

Hydrated silicate detections from the Compact Reconnaissance Imaging Spectrometer for Mars (CRISM; after Ehlmann et al. [2011]) on a MOLA colorized relief map of the Isidis Basin and surrounding regions. Surface dust increases dramatically to the east and north, effectively obscuring the spectral properties of bedrock in these areas (above and right of the thick dashed line). Hydrated silicate occurrences are concentrated in the ancient heavily cratered basin ejecta-bearing higher elevation terrains surrounding the Isidis Basin. Many occurrences correspond with the presence of megabreccias observed in images from the High-Resolution Imaging Science Experiment (HiRISE) camera images (red squares outlined in green).

Figure 6.

Hydrated silicate detections from the Compact Reconnaissance Imaging Spectrometer for Mars (CRISM; after Ehlmann et al. [2011]) on a MOLA colorized relief map of the Hellas Basin and surrounding regions projected to the approximate centroid of the basin. Hellas is a noncircular basin, which likely resulted in asymmetrical ejecta distribution. Lines of latitude and longitude have a spacing of 30°.

[29] Global mapping of 1680 high-resolution observations from OMEGA and CRISM by Carter et al. [2011b, 2013] indicates that ~70% of the occurrences correlate specifically with mid-sized to large craters (tens up to a few hundred kilometers in diameter). This was confirmed by a nearly simultaneous, independent study of 629 CRISM high-resolution observations by Ehlmann et al. [2011]. Ehlmann et al. [2011] specifically classify both the crater-exposed and the less common tectonic-exposed occurrences together as “crustal clays.” In general, Carter et al. [2011b, 2013] and Ehlmann et al. [2011, 2013] contend that most of these occurrences are merely exposures of previously existing hydrated silicate phases. Indeed, the correlation between spectral units and geologic features, due to complexities of the impact process and Martian geologic history, does not automatically equate to an impact origin. Essentially, the hydrous silicates associated with craters can be pre-impact, syn-impact, and/or post-impact. Osinski et al. [2012] discuss how all three scenarios should be considered collectively and carefully. In the case between pre-impact and syn-impact scenarios, one should not be quickly dismissed in favor of the other. Depending on the specific impact conditions and the history of a crater, hydrated silicate spectral units may have originated from any one of these scenarios. This appears to be the case for Toro Crater, a ~40 km diameter complex crater in Syrtis Major [Marzo et al., 2010], for which there is observational evidence in support of both pre-impact and syn-impact formed hydrous silicates. Although excavation is more likely for smaller simple craters, terrestrial studies show that hydrothermal alteration is still possible for craters down to ~2 km in diameter [Hagerty and Newsom, 2003; Naumov, 2005; Osinski et al., 2012].

[30] The impact alteration model also does not rule out, but rather is compatible with the occurrence of hydrated silicates in other non-impact geologic settings (e.g., fluvial, lucustrine, burial diagenesis) . Essentially, these deposits may be composed of eroded and transported materials that may have originated from a heavily cratered and impact-altered crust. A complicating factor that is important to acknowledge here is that these settings may further promote alteration, which could even overprint any previous impact-generated alteration signature. These combined factors may explain the widespread distribution of hydrated silicates that continue to be detected from orbiting spectrometers predominantly within or proximal to Noachian-aged surfaces [e.g., Murchie et al., 2009; Carter et al., 2011b, 2013; Ehlmann et al., 2011]. Untangling such a complex petrogenetic and alteration history may require detailed in situ analyses or perhaps even Mars sample return.

3.2 Hydrated Silicate Associations With Craters and Megabreccia Deposits

[31] The association of hydrated silicate spectral signatures particularly with megabreccias, interpreted to be impact in origin [McEwen et al., 2008; Tornabene et al., 2009, 2010; Osinski et al., 2012], also supports that impact cratering may have had an important role in generating hydrous silicate phases on Mars. Martian megabreccias were discovered within High Resolution Imaging Science Experiment (HiRISE) images of Holden Crater and within outcrops in Uzboi Valles that incised into the Holden Crater basement. These were initially described as a “variably rounded, poorly sorted, chaotically arranged blocks up to 50 m across within a finer matrix often characterized by clastic dikes” [Grant et al., 2008]. Other occurrences have now been identified throughout, not only the heavily cratered southern highlands, but also in craters within the northern lowlands of Mars [e.g., Carter et al., 2010]. Deep bedrock exposures of megabreccias have been specifically observed within the floor of Uzboi Valles and in parts of the Valles Marineris canyon system [Murchie et al., 2009]. They are also found within or around structural uplifts of large complex craters [Tornabene et al., 2010, 2012b], and are commonly observed in the structural uplifts in large complex craters in or near Noachian terrains, and especially proximal to large impact basins [Tornabene et al., 2010, 2012b]. Hydrated minerals, especially phyllosilicates, are commonly associated with the occurrence of megabreccia, which in some cases also contains unaltered megaclasts (i.e., large blocks) consisting of mafic lithologies [Tornabene et al., 2009; Osinski et al., 2012]. Megabreccias, particularly their clasts, may represent some of the oldest rocks exposed on the surface of Mars—possibly including fragments of the original “primary” Martian crust [e.g., Skok et al., 2010, 2012]. However, there are some clear cases of megabreccias that are associated with younger, post-Noachian craters. The relatively youthful, Amazonian-aged crater Mojave Crater is one such example, where megabreccias with relatively small clasts (~10 m) are observed as part of the pitted crater-fill materials [see Tornabene et al., 2012a, Figure 4]. Megabreccias with larger megaclasts (> 10 m) are typically found in deep bedrock exposures, such as crater central structural uplifts [Tornabene et al., 2009, 2010, 2012b]. Such megabreccias are interpreted pre-impact, which have a complex geologic history due to additional brecciation, and the potential alteration associated with the formation of the host crater. Mars may possess an ~1 km thick lunar-like layer of porous and permeable megaregolith [Hartmann and Neukum, 2001]; however, unlike the Moon, the Martian version of this layer may be only partially exposed, been eroded and modified by subsequent geologic processes, and may have been largely cemented by alteration and diagenesis.

[32] By combining the highest spectral and spatial resolution VNIR and visible image data for Mars [e.g., Delamere et al., 2010], in addition to our current understanding of the impact process, it may be possible to identify and possibly differentiate between different impact scenarios for hydrated silicate phases. For example, by combining CRISM with other higher resolution Mars Reconnaissance Orbiter (MRO) data sets such as Context Camera (CTX) and HiRISE images, the precise location and type of occurrence of hydrated silicate signatures can be determined (e.g., within megaclasts, injection dikes, megabreccia mesostasis, etc.), which can yield important clues to their most probable origins [Tornabene et al., 2009, 2010, 2012b; Osinski et al., 2012]. These techniques have been particularly insightful in several localities, including: Toro Crater, Syrtis Major, Holden Crater, and Uzboi Vallis [Marzo et al., 2010; Osinski et al., 2012].

[33] Figure 7 shows a composite of CRISM and HiRISE data covering an outcrop of megabreccias in the footwalls of one of the troughs (i.e., graben) of Nili Fossae. The megabreccias specifically correlate with Fe/Mg-OH and H2O-bearing materials consistent with smectite. Given the general regional geology and proximity to the Isidis Basin, we interpret these phyllosilicate-bearing megabreccias as ejecta originating from Isidis. If true, much of the alteration in this region may be attributed to pervasive alteration of thick, voluminous and melt-rich basin ejecta deposit (and surrounding materials) as previously suggested by Tornabene et al. [2008b]. This is consistent with abundant occurrences of megabreccias around the Isidis Basin that correlate with various hydrated silicate phases.

Figure 7.

Megabreccia occurrences along Nili Fossae trough near the Isidis Basin. (a) A THEMIS daytime infrared image with colorized MOLA elevation for context. Massive outcrops of megabreccias (approximately hundreds to thousands of meters thick) noted in HiRISE (Primary Science Phase images—PSP) correlate specifically with Fe-Mg bearing phyllosilicates detected by CRISM. Hargraves (D ~ 60 km; just off center) and Toro craters are particularly noted here as they are interpreted to have excavated hydrated silicate-rich rocks in their ejecta and structural uplifts, respectively. (b) CRISM phyllosilicate summary product using the D2300-BD2210-BD1900 band parameters [see Murchie et al., 2009] in R-G-B space highlights the presence of Fe-Mg phyllosilicates in magenta and red and is overlain on the infrared brightness summary product (1.3 µm) derived from a CRISM Half-Resolution Long observation (HRL000095A2) with an ~37.7 m/pixel ground resolution. (c) A close up of a HiRISE red mosaic image (PSP_006844_2000) showing an outcrop of megabreccias that occur along the trough. Megaclasts are light toned with respect to their surroundings and on the order of tens of meters in diameter with the largest block measuring over 100 m in diameter. (d) The same close-up from Figure 7c with a precise overlay of the CRISM phyllosilicate summary product. Areas highlighted by magenta correlate with the megablocks specifically and exhibit the strongest absorptions indicative of Fe-Mg phyllosilicates (smectite clays). Aeolian bed forms in the red area indicate that a considerable component of the red unit is surficial fines, which obscure the underlying bedrock.

3.3 The Juxtaposition of Hydrated Silicates and Unaltered Volcanics on Mars

[34] In the vicinity of the Nili Fossae, unaltered olivine-rich and olivine-poor [Mustard et al., 2007, 2008; Tornabene et al., 2008b; Ehlmann et al., 2011] materials, including Syrtis Major volcanics, unconformably overlie these phyllosilicate-bearing megabreccias. In Mawrth Vallis, almost a hemisphere away, phyllosilicate-bearing outcrops are also capped by a thinner, clay-free, mafic unit [Loizeau et al., 2010; Noe Dobrea et al., 2010]; similarly, altered megabreccias in contact with unaltered mafic materials have also been observed within some of the deepest bedrock exposures in the Valles Marineris system [Murchie et al., 2009]. Although there may be some exceptions [e.g., Meunier et al., 2012], the juxtaposition of what are interpreted to be unaltered volcanic rocks with pervasively altered rocks and megabreccias is difficult to reconcile with most regional endogenic alteration processes, but is in fact predicted by an impact model where the alteration is driven by the heat and volatiles from within the impact deposits, and thus spatially and stratigraphically restricting the alteration.

4 Discussion

4.1 Determining a Relationship Between Hydrated Silicates and an Impact Origin on Mars

[35] The most vital step toward determining a relationship between hydrated silicates and an impact origin would be to constrain if whether an occurrence of hydrated silicates is hosted in or derived from impactites. Impactites are defined as “any rock affected by impact metamorphism” [Stöffler and Grieve, 2007]. Impactites formed from single or multiple impact events and are classified into three major groups: (1) shocked rocks, (2) impact melt rocks, and (3) impact breccias. Even on Earth, identifying impactites is challenging, often requiring intensive and multitechnique laboratory analysis of hand specimens [e.g., French and Koeberl, 2010]. As such, it is even more challenging to recognize such materials in remote data sets [e.g., Johnson et al., 2002; Clark et al., 2007; Arvidson et al., 2008; Wright et al., 2011]. Nonetheless, these challenges need to be met, as it is vitally important for understanding the role and interaction between volatiles and impacts with respect to the geologic history and early habitability of Mars.

[36] Typically, the effects of shock metamorphism are characteristic and thus can be used as criteria for the identification of an impact origin for various geologic materials (see the summary and review by French and Koeberl [2010] and references therein). On Earth, several mineral, lithologic, and geochemical characteristics may be used to identify a potential impact setting. Unfortunately, with the exception of shatter cones, the most diagnostic features are microscopic and require analysis in the lab for unambiguous determinations. Recognition of these microscopic features, in addition to impactites having similar textures to volcanic and volcanoclastic rocks, are the greatest challenges for identifying impactites on Mars.

[37] In lieu of sample return, using a synthesis of chemical, mineralogical, contextual and textural in situ and orbital data is the current best method for identifying impactites on Mars. Recent results from the Mars Exploration Rovers (MERs) demonstrate how this was accomplished with the Athena science package [Squyres et al., 2012]. The results from the Spirit rover particularly demonstrate the various challenges, including some successes, with respect to recognizing possible impactites when the setting and the lithologies are often complicated and potentially ambiguous [Clark et al., 2007; Arvidson et al., 2008]. Here below we summarize and discuss an example from the Spirit rover, and then we make suggestions for general strategies toward identifying impactites in situ with the Mars Science Lab (MSL) Curiosity rover's instruments; we note that this general strategy can be adapted, modified, or even used to plan the type of instrumentation desired, for future in situ exploration missions.

4.1.1 General Strategies for Identifying a Relationship Between Hydrated Silicates and an Impact Origin In Situ

[38] During Spirit's campaign in the Columbia Hills, outcrops of bedrock encountered during Sols 549–572, named the Assemblee outcrop, were interpreted to be altered impactites [Arvidson et al., 2008]. The interpretation of these materials as impactites required a comprehensive characterization via a synthesis of multiple data sets collected from the various instruments, including the Navigation Camera (Navcam), Panoramic Camera (Pancam), Microscopic Imager (MI), Miniature Thermal Emission Spectrometer (Mini-TES), Mössbauer Spectrometer (MB), Alpha-particle X-Ray Spectrometer (APXS), and the Rock Abrasion Tool (RAT) [see Arvidson et al., 2008]. The first and most striking clue consistent with an impact origin was the unusual (with respect to previous observations in the Columbia Hills) fractured and clastic texture of the rocks. Although they are described as “conglomerates,” this does not preclude an origin via impact processes (e.g., impact breccias). In general, breccia, a rock consisting of angular mineral or lithic clasts, is a term that is pervasively used to describe clastic impactites in both the terrestrial and planetary literature [e.g., French, 1998; Reimold, 1998; Stöffler and Grieve, 2007]. The extensive use of “breccia” as being synonymous with a clastic impactite is problematic and has led some workers to dismiss an impact origin for clastic rocks on Mars that contain rounded clasts (e.g., at the Pathfinder landing site) [Matijevic et al., 1997]. Rounded clasts are an essential component of conglomerates—a term seldom used to describe impactites; however, rounded clasts are also found in numerous impactites [e.g., Reimold, 1998]. As such, an impact origin for rocks on Mars should not be dismissed based on the presence of rounded clasts alone. However, these rock textures alone, even in light of strong support from local and regional geologic context, are insufficient to recognize lithologies as impactites. Additional clues supporting an impact origin for the Assemblee clastic rocks came from spectral and composition data sets that indicated the presence of Fe-poor amorphous glass and relatively high siderophile elements (Ni, Cr, Co) [Clark et al., 2007; Arvidson et al., 2008; Ming et al., 2008]. Furthermore, the inclusion of maskelynite (diaplectic plagioclase glass) was noted to improve modeled Mini-TES spectral matching of the Assemblee outcrop. The results of the Mini-TES linear spectral deconvolution suggest the presence of both ~15–20% maskelynite and ~60% amorphous aluminosilicate, possibly representing an impact melt component [Clark et al., 2007]. Recent work on the VNIR and TIR spectral characteristics of exogenic and endogenic glasses by Craig et al. [2011] indicates that impact glass may have spectral distinct features, which may further improve deconvolutions of the Assemblee clastic rocks and be used to identify a possible impact glass component remotely.

[39] Additional supportive evidence from bulk rock and clast compositions for the Assemblee clastic rocks appear to be consistent impact mixing. Clasts compositions are consistent with the regional rocks previously characterized by Spirit (e.g., Wishstone class), while the bulk composition of these putative impactites shows that these materials plot close to compositional origin (i.e., the average) of rocks analyzed in the Columbia Hills. Although not strongly emphasized in the literature, the distinctive structure of the Voltaire outcrop, compared to the other outcrops in the Columbia Hills [McCoy et al., 2008], may also be an important clue supporting an impact origin for these materials [Arvidson et al., 2008].

[40] Although most of the instruments onboard Spirit and Opportunity are distinct from those on Curiosity, the above example demonstrates the effectiveness of data synthesis that can lead to the identification and characterization of impactites on Mars. It is clear from this example that characterization of impactites cannot be as easily accomplished as it is commonly done on Earth. Any one data set alone is insufficient for the recognition of impactites, thus a similar data synthesis approach should be adopted and adjusted accordingly to make similar identifications with the Curiosity's suite of instruments. The Curiosity instrument suite consists of: the Mars Hand Lens Imager (MAHLI) [Edgett et al., 2012], Mast Camera (Mastcam) [e.g., Malin et al., 2010; Bell et al., 2013], Chemical Camera (Chemcam)—including laser induced breakdown spectroscopy (LIBS) and remote micro-imager (RMI) [Maurice et al., 2012], CheMin—which includes X-ray diffraction (XRD) and X-ray fluorescence (XRF) [Blake et al., 2012], Alpha-Particle X-ray Spectrometer (APXS) [Campbell et al., 2012], Sample Analysis at Mars (SAM) [Mahaffy et al., 2012], and Dynamic Albedo of Neutrons (DAN) [Litvak et al., 2008]. Our recommendations involve the following observations and instruments:

  1. [41] Look for rocks containing shatter cones or consisting of highly fractured, fragmental textures, and/or glassy materials that may also include flow-banded textures (i.e., schlieren) and spherules. Impact melt rocks may be highly vesiculated and clast-poor impact melt rocks may generally lack intact phenocrysts—Mastcam, MAHLI, ChemCam(RMI).

  2. [42] Perform textural analyses of images of clastic rocks similar to the approach of [Chanou et al., 2011] for additional context—ChemCam (RMI), Mastcam, and MAHLI.

  3. [43] Determine if siderophile enrichments are present, possibly representing a meteoritic component—ChemCam (LIBS), APXS, and CheMin (XRF).

  4. [44] Determine if amorphous or partially amorphous phases represent impact melt glasses (may be associated with a generally higher Si content), and/or the presence of diaplectic glass (e.g., maskelynite)—CheMin (XRD) [see Hörz and Quaide, 1973; Heymann and Hörz, 1990].

    1. [45] If hydrated melt glasses are present, quantify the abundance and determine the host phase of the hydration—SAM, DAN, APXS, ChemCam (LIBS), and CheMin.

  5. [46] Determine if impact-generated high-pressure polymorphs are present—CheMin (XRD)

  6. [47] Determine if the local and regional context strongly support an impact origin—a synthesis of orbital data in addition to Mastcam multiband images.

  7. [48] Use spectral data sets and techniques (band ratios, etc.) to provide additional supporting evidence—ChemCam (passive spectroscopy), Mastcam.

[49] On Mars, the observation of breccias and glassy lithologies would be the most obvious first clue that warrants further investigations. The amorphous component, such as diaplectic glass and impact melt should be detectable with the XRD in the CheMin package. Although it may not be distinct from volcanic glasses with respect to its XRD pattern, the recognition of glass in addition to other corroborative evidence is useful. Impact-generated high-pressure polymorphs (e.g., ringwoodite) [French and Koeberl, 2010] maybe detectable but only if highly shocked rocks are encountered and have concentrations at the >3% detection limit of CheMin [Blake et al., 2012]. Major and trace element composition can be a key indicator of impact origin, but there are some caveats that need to be accounted for when making such determinations. Impactites will often possess nonequilibrium compositions (e.g., a variety of mixed lithologies from the target) or will have elevated siderophile elements (Ni, Cr, Co) or Platinum Group Elements (PGEs; specifically Ir and Os). Siderophile and PGE enrichments are highly dependent on impactor chemistry and are sometimes ambiguous with some endogenic processes that can lead to enrichments in these elements (e.g., hydrothermal, elemental partitioning—like Ni into olivine, etc.).

5 Summary and Conclusions

[50] In addition to the development of a hydrothermal system, we propose a series of additional impact-generated mechanisms for the production of hydrated silicates on Mars that do not necessarily rely on assumptions about a sustained warmer and wetter climate during the early history of the planet. Numerous and frequent large impacts into a volatile-, water/ice-, and silicate-rich upper crust are the two main requirements for these mechanisms to operate. Abundant impact glasses (including hydrated varieties) and impact-damaged crustal materials produced during the late heavy bombardment are most voluminous in the Noachian crust. The impact alteration model predicts that these crustal materials would be most susceptible to any subsequent aqueous alteration. In addition to impact-generated hydrothermal activity, devitrification and autometamorphism are significant impact-generated alteration processes in terrestrial craters. Autometamorphic reactions may dominate in some craters where water may be limited (e.g., the hydrothermal system is either stunted or even absent), while devitrification provides an even more extreme case. Hydrated impact glasses will form hydrated silicates (e.g., smectite clay) in the solid state by devitrification even under cold and dry conditions because it does not require additional water. This process may be particularly significant on Mars given the recent supporting evidence for volatile-rich impact deposits on Mars [Boyce et al., 2012; Tornabene et al., 2012a].

[51] We note that devitrification of hydrated impact glass into crystalline hydrated silicates can proceed at very slow rates in the absence of subsequent heat and aqueous activity, forming crystalline hydrated silicates over periods on the order of 109 years. Hence, if the Martian crust has been volatile rich throughout its history, then it is possible to produce hydrated silicates in any geologic period since the Noachian via this process. In spite of numerous claims that hydrated silicates were formed exclusively during the earliest, most ancient period on Mars, and suggestions that phyllosilicate formation may even define the earliest era (i.e., Phyllosian) [e.g., Bibring et al., 2006], it is vitally important to recall that model age estimates from crater counting statistics date surfaces (i.e., the host rocks) and not the alteration events that overprint them. This is not to say that many hydrated silicates were not necessarily formed during the Noachian, but to emphasize that they need not have formed during this period. However, for younger terrains (i.e., those possessing less abundant and generally smaller craters), these deposits would be less voluminous and more difficult to detect from orbit. Since the cratering rate has exponentially decreased over time, the most voluminous deposits predicted by this model—and those that would be most easily detectable by orbiting instruments—occur within the heavily cratered Noachian-aged surfaces. The impact alteration model is therefore consistent with the observations from orbiting spectrometers that continue to detect these materials primarily in the heavily cratered Noachian highlands.

[52] Finally, we emphasize that the impact alteration model does not preclude climate change, or other endogenic alteration settings over Martian geologic history, but expands the realm of possibilities to include extensive hydrated silicate formation on a colder and drier Mars that was only transiently “wet”. We note that even if Mars experienced a warmer and wetter climate, impact-damaged materials are still a significant component of the Martian crust and would have likely played an important role in the alteration history of the planet. Unraveling the origin and redistribution of hydrated materials resulting from impacts and other geologic processes will be difficult, but may be possible with detailed mapping by combining spectral data sets such as CRISM with the high-resolution (submeter) color capabilities of the HiRISE instrument, and via future in situ analysis by rovers (e.g., MSL).


[53] We thank Jeffrey Moersch for use of the FieldSpec Pro FR Portable Spectroradiometer and the reduction of the data collected. We thank the engineers and managers, and operations and science teams of the MRO project. We would like to thank Oleg Abramov and Gilles Berger for their comments and suggestions, which greatly improved our manuscript. L.L.T. acknowledges funding from G.R.O.’s Industrial Research Chair in Planetary Geology, which is supported by NSERC, CSA, and MDA. L.L.T. would also like to personally thank Ryan Hopkins, Albert Ortiz, and Gisela Telis for some additional insights that were helpful toward the completion of this manuscript. G.R.O. thanks NSERC for funding.