INFRARED SPECTRAL IDENTIFICATION OF UNUSUALLY FELDSPAR-RICH ROCKS ON MARS

Introduction: Feldspar-rich, mafic-poor rocks dominate Earth’s upper continental crust and the ancient lunar highlands. On Earth, felsic rocks (>65% SiO2), including granite and its volcanic equivalent, rhyolite, form most commonly in subduction zones through partial melting or fractional crystallization in the shallow crust. Anorthosite (>90% feldspars) forms in low-density cumulate layers in a slowly cooling mafic magma chamber, or on a global scale (e.g., on Earth’s Moon) in a primordial magma ocean. Mars appears to lack plate tectonics, yet has experienced too much erosion and burial to retain abundant surface evidence for any early magma ocean. The inferred dominance of basaltic compositions [1] is thus relatively unsurprising. Some regions spectrally resemble andesite [2], but have alternatively been interpreted as weathered basalt or glass [3,4]. Localized quartz was first attributed to evolved magmas [5] but can instead be explained as a secondary phase formed hydrothermally or via diagenetic maturation of opaline silica found in the same locations [6,7]. The only widely accepted example of intermediate-to-felsic igneous rocks on Mars is a dacitic unit in the Nili Patera caldera of Syrtis Major, identified in the thermal IR [8]. Here we describe a new method for identifying very mafic-poor materials on Mars, present detections, and discuss possible interpretations and implications. Data and Methods: We used data from the Compact Reconnaissance Imaging Spectrometer for Mars (CRISM). Spectral features in CRISM’s 0.4–3.9 μm wavelength range result mostly from electronic transitions in transition metals (primarily Fe) and vibrational modes of light molecular groups such as OH, H2O, or CO3. Pure quartz and feldspar are nominally undetectable, but some feldspars exhibit a broad absorption centered at ~1.25–1.3 μm due to minor substitution of Fe for Ca. This absorption is typically much weaker than the ~1 μm absorption of Fe-rich phases such as olivine and pyroxene, so it is detectable only in very feldspar-rich rocks with <5% associated mafics (as seen in some areas on the Moon [9]). Such rocks were not expected on Mars, so none of CRISM’s standard mapping parameters is designed to identify them. Yet they were found in small outcrops in Xanthe Terra (N. of Valles Marineris) [10] and elsewhere [11]. We sought broader exposures to better characterize the distribution and physical properties of these materials. We focused our analyses on Syrtis Major (the only large Martian volcano whose bedrock compositions are not obscured by mantling dust) and northeast Noachis Terra in the southern highlands, which has the densest concentration of exposed intercrater and craterfloor bedrock on Mars [12,13]. Other regions may contain analogous materials beneath surficial dust. Noachis Terra: In broad outcrops spanning tens of km each across several large crater floors and intercrater plains in Noachis Terra (Fig. 1), a ~1.3 μm absorption band is the strongest spectral feature observed in the 0.7–2.6 μm range (Fig. 2). We attribute this feature to Fe in feldspar. The only other minerals in our spectral libraries with similarly broad absorptions centered at 1.23–1.32 μm are Fe-bearing garnets, such as almandine (Fig. 2b). However, such garnets have a comparably broad (and half as deep) absorption at ~1.7 μm, which is not observed in our CRISM spectra. The lack of any absorptions near 1 μm (Fig. 2a) suggests very low pyroxene and olivine abundances, possibly <2% [9]. Spectrally these rocks resemble lunar anorthosites, but they could alternatively be truly felsic, as CRISM cannot uniquely constrain their quartz content nor Ca,Na,K-feldspar composition (Fig. 2b).