Constraining Lunar Surface Mineralogy with Combined Thermal- and Near- Infrared Spectral Data

Introduction: The current understanding of lunar surface mineralogy largely comes from two sources: samples (Apollo and meteorite [e.g. 1]) and remote near-infrared spectroscopic measurements (telescopic and Clementine [e.g. 2]). Samples provide detailed compositional analysis of a limited number of sites across the lunar surface and NIR measurements provide global, high-resolution coverage of Fe-bearing mineralogy. The upcoming Diviner Lunar Radiometer Experiment (DLRE) on the Lunar Reconnaissance Orbiter (LRO) will provide the first global coverage maps of thermal-infrared derived compositions and physical properties. Diviner has only three mineralogy spectral channels centered at 7.8, 8.2, and 8.6μm so it is important to integrate Diviner with other compositional data sets. We examine two approaches in this study. First, thermal infrared laboratory spectral measurements of mineral endmembers, a mineral mixture, Apollo 16 highlands, and Apollo 17 mare soil samples are convolved to Diviner spectral bands. With these laboratory spectra, we investigate how Diviner data can be applied to distinguish lunar surface materials. Second, the thermal infrared measurements of minerals and a mineral mixture are integrated with an adapted nearinfrared spectral curvature parameter developed for mafic minerals and Clementine data [2]. Here, we examine the extent to which combined analyses can be used to constrain the mineralogy of immature lunar surface lithologies (i.e. small outcrops and crater central peaks). Samples and Methods: Laboratory emissivity spectra of < 25 μm grain size fractions of plagioclase, lowand high-Ca pyroxenes, and olivine used in this work are from the Berlin emissivity database (BED). Emissivity measurements were made with a Fouriertransform infrared spectrometer Bruker VERTEX 80V [3]. A 50/50 wt% mineral mixture of end-members anorthite and olivine is also examined. Lunar soil samples are characterized by the Lunar Soil Characterization Consortium (LSCC) [4-5]. Apollo 17 mare soil samples chosen for this study include 71061, 71501, 70181, and 79221 [4] and Apollo 16 highlands soil samples include 61141, 61221, 62331, 64801, 67471, and 67481 [5]. Apollo lunar soils are plotted in Figure 1 on a plagioclaseorthopyroxene-clinopyroxene ternary diagram [6]. Highland soil samples are classified as anorthosite and mare soil samples as gabbro. Brown University’s Reflectance Experiment Laboratory (RELAB) Nexus FTIR spectrometer was used to measure thermal infrared spectra of each lunar soil sample for the 10 – 20 μm grain size fraction. Thermal infrared RELAB spectra are converted to emissivity using the approximation to Kirchoff’s relation E=1R. All thermal infrared spectra are convolved to Diviner’s three spectral bands using ENVI’s spectral resampling tool. The Diviner spectral bands were chosen specifically to measure the location of the Christiansen Feature (CF). The CF is an emission maximum, or reflectance minimum, first described as an indicator of compositions by [8]. The CF shift to shorter wavelengths for particulate materials in a vacuum environment is well constrained [9-10]. In this study, we calculate three band ratios (7.8/8.2, 7.8/8.6, and 8.2/8.6) for each spectrum, assume that the CF shift applied to each spectral band is the same, and apply the ratios to accurately identify lunar lithologies.