44th Lunar and Planetary Science Conference (2013) 1602.pdf

CONSTRAINING THE FERROUS IRON CONTENT OF SILICATE MINERALS IN MERCURY’S CRUST. Rachel L. Klima ([email protected])1, Noam R. Izenberg1, Scott Murchie1, Heather M. Meyer1, Karen R. Stockstill-Cahill2, David T. Blewett1, Mario D’Amore3, Brett W. Denevi1, Carolyn M. Ernst1, Jörn Helbert3, Timothy McCoy2, Ann L. Sprague4, Faith Vilas5, Shoshana Z. Weider6, and Sean C. Solomon7. 1Johns Hopkins University Applied Physics Laboratory, Laurel, MD 20723, USA; 2National Museum of Natural History, Smithsonian Institution, Washington, DC 20013, USA; 3Institute for Planetary Research DLR, Berlin, Germany; 4Lunar and Planetary Laboratory, University of Arizona, Tucson, AZ 85721, USA; 5Planetary Science Institute, Tucson, AZ 85719, USA; 6Department of Terrestrial Magnetism, Carnegie Institution of Washington, Washington, DC 20015, USA; 7Lamont-Doherty Earth Observatory, Columbia University, Palisades, NY 10964, USA.

Introduction: Since the spectrum of Mercury was Fresh Crater Spectra: The VIRS instrument is a first observed by telescope, the surface has been point spectrometer, and global coverage is thus built up interpreted as being relatively low in iron when by accumulating tracks of spectra (Fig. 2). Whereas large compared with the lunar maria and much of the inner craters and other features of interest can be targeted, the solar system. Early telescopic measurements suggested majority of bright, fresh craters are small, which limits that the Moon and Mercury share a similarly reddened the number that are traversed by VIRS tracks. We spectral slope, perhaps indicating that Mercury’s crust identified fresh craters in the northern volcanic plains has a similar composition to the lunar highlands [1,2]. (NVP) by searching all spectral data within a defined MESSENGER flyby data [3] and ground-based geographic range and located those with albedos that are telescopic observations at visible and near-infrared (NIR) at least five times brighter than the surrounding terrain. wavelengths [4] indicated and at mid-infrared [5] A total of 710 spectra were identified, and the 100 wavelengths confirmed that the silicate surface is very brightest were selected for more detailed analysis. The low in ferrous iron, but the upper limit remained illusive. geographic coordinates of each spectrum were plotted on Since MESSENGER entered orbit, the X-Ray an MDIS map to verify that they are indeed craters and Spectrometer (XRS) has been used to quantify the bulk not instrumental artifacts (see Fig. 3). Several fresh iron content of the surface, giving an initial upper limit craters found within other terrains were also identified of ~4 wt % Fe [6]. However, the XRS data do not manually (Fig. 3). explicitly determine the iron oxidation state, nor in which mineral(s) it resides. The interior of Mercury is interpreted to be extremely reducing [7], and surface iron should not be present as Fe3+. It is also difficult to retain any Fe2+ in silicates at low oxygen fugacity, so the iron observed at the surface could be bound in sulfides and/or iron metal deposited via meteorite impacts. NIR Spectroscopy of Low-Fe Minerals: The low iron content of Mercury’s surface severely complicates the identification of specific minerals using NIR reflectance spectroscopy. This is because the diagnostic crystal field bands used to identify and characterize mineralogy at these wavelengths require transition Fig. 1. Example laboratory spectra for mafic minerals that metals to be structurally bound within a silicate crystal petrologic modeling [8] has shown should be present on 2+ lattice (see Fig. 1). The absence of a crystal field band Mercury. When Fe is present, even in small amounts, in the crystal lattice, absorption bands are strong; when it is absent, near 1000 nm can, however, be used to constrain how spectra are bright and featureless and essentially much iron is hosted in the silicates at the surface of indistinguishable in the NIR from nominally iron-free minerals Mercury. From telescopic measurements of Mercury, it (e.g., quartz and plagioclase). was suggested that up to a generous maximum of 6 wt % FeO could be present in the silicates, even with no measurable 1000 nm absorption band, but that a minimum of 3 wt % FeO had to be present in order to produce the observed surface darkening by space weathering [9]. With high spectral resolution, spatially resolved data from the Mercury Atmospheric and Surface Composition Spectrometer (MASCS) Visible and Infrared Spectrograph (VIRS), we reexamine the spectral constraints on the amount of iron in Mercury’s surface silicates by examining fresh craters that have Fig. 2. False color composite of VIRS spectral parameters experienced substantially less space weathering than the showing surface coverage. Red: 575 nm reflectance; green: average surface of Mercury. vis/IR (415 nm/750 nm); blue: UV/vis (310 nm/390 nm). 44th Lunar and Planetary Science Conference (2013) 1602.pdf

Examples of two spectra from fresh craters within the interpretations from an absorption’s non-detection. NVP and one from the plains surrounding Calvino crater However, a space weathered, low-iron, high phase angle are shown in Figure 4. The VIRS spectra have been (α =85°) measurement of the lunar surface by M3 with a photometrically corrected to a common viewing low signal-to-noise ratio (~6) still shows an appreciable geometry (incidence=45°, emittance=45°, phase=90°) absorption band. Petrologic modeling of Mercury’s using a Lommel-Seeliger correction with a data-derived surface compositions has shown that the dominant mafic phase function, and then binned to 10 nm spectral mineral on Mercury is low-iron orthopyroxene [8], resolution. For comparison, a low signal-to-noise, high- which exhibits an absorption band centered at 900 nm latitude (85°N) lunar spectrum, measured by the Moon that is well within the range of the higher signal-to-noise Mineralogy Mapper (M3), is also shown in Figure 4. This visible detector on VIRS. This absorption, if present, spectrum was obtained from a region with ~2 wt% FeO would also be clearly discernable in multiple filters of (derived from Lunar Prospector gamma-ray spectrometer the MDIS color data, yet none has so far been identified data [10]). The low-iron (but not iron-free) olivine and [12]. This does not imply that orthopyroxene is not pyroxene spectra from Figure 1 are included in Figure 4, abundant, but rather that its composition must be almost but are scaled to 10% of their laboratory reflectance. iron-free. Although we are attempting to mitigate some of the effects of space weathering by searching for optically fresh exposures, Mercury’s high surface temperature creates an additional complication. Spectral changes for some minerals have been observed when they are heated to Mercury-relevant temperatures in a vacuum [13]. However, this thermal processing does not appear to weaken the iron absorption bands in pyroxene or olivine [13]. Implications: The orthopyroxene spectrum shown in Figure 4 contains 1.83 wt % FeO. Nonlinear mixing of Fig. 3. Locations of fresh craters identified in the NVP (blue stars) and other terrains (yellow stars). The mapped smooth this pyroxene with an optically neutral mineral, such as plains [11] are overlaid on the Mercury Dual Imaging system plagioclase, in the proportions modeled by Stockstill- (MDIS) monochrome base map. Sites outside of the NVP were Cahill et al. [10] does not diminish the strength of its selected to be within other smooth plains units. absorption band. Darkening the spectrum by decreasing the overall spectral contrast to 10% of its laboratory value still results in a clear absorption band at 900 nm. Our results suggest that the silicate minerals on Mercury contain about an order of magnitude less iron than was previously interpreted [9]. This figure is lower than the most recent Fe abundance estimates made from XRS data that suggest the bulk surface contains ~1% Fe [14] as well as the Fe content measured by the Gamma-Ray Spectrometer [15], supporting the conclusion that this iron is not bound in silicates but rather is contained in sulfide phases and/or is present as a late veneer from impact debris [14].

Fig. 4. Representative spectra from two volcanic plains regions Acknowledgements: We are grateful to the NASA on Mercury compared with a lunar highlands spectrum from Discovery and PMDAP (NNX12AQ67G to RLK and M3 and scaled laboratory spectra. NNX12AQ73G to NRI) programs for supporting this work. Even fresh crater spectra (Fig. 4) display no evidence References: [1] Irvine, W. M. (1968) Astron. J., 73, 807- of a 1000 nm band in either the NVP or the Calvino 828. [2] McCord, T. B. and J. B. Adams (1972) Science, 178, smooth plains. Similarly, none of the other ~720 bright 745-747. [3] McClintock, W. E., et al. (2008) Science, 321, 62- crater spectra identified so far in the VIRS data has a 65. [4] Vilas, F. (1985) Icarus, 64, 133-138. [5] Sprague, A. L., 1000 nm band. This result further suggests that the et al. (2002) MAPS, 37, 1255-1268. [6] Nittler, L. R., et al. silicates on Mercury’s surface have a ferrous iron content (2011) Science, 333, 1847-1850. [7] McCubbin, F. M., et al. that is substantially lower than the 3-6% previously (2012) GRL, 39, L09202. [8] Stockstill-Cahill, K. R., et al. (2012) JGR, 117, E00L15. [9] Hapke, B. (1977) Phys. Earth estimated [9]. Planet. Inter., 15, 264-274. [10] Lawrence, D. J., et al. (2002) There has been a long-standing concern that the JGR, 107, 5130. [11] Denevi, B. W., et al. (2012) JGR, restricted viewing geometry of VIRS (no phase angles submitted. [12] Blewett, D. T., et al. (2009) EPSL, 285, 272- lower than ~78°) or the reduced signal-to-noise of the 282. [13] Helbert, J. et al. (2011) AGU. [14] Weider, S. Z., et al. VIRS IR detector could obscure any absorption band that (2013), LPS, 44, this mtg. [15] Evans L. G. et al. (2012) JGR, might otherwise be observed. It is difficult to make E00L07.