Thermal infrared spectral analysis of compacted fine-grained mineral mixtures: implications for spectral interpretation of lithified sedimentary materials on Mars

Characterizing the thermal infrared (TIR) spectral mixing behavior of compacted fine-grained mineral assemblages is necessary for facilitating quantitative mineralogy of sedimentary surfaces from spectral measurements. Previous researchers have demonstrated that TIR spectra from igneous and metamorphic rocks as well as coarse-grained (>63 micron) sand mixtures combine in proportion to their volume abundance. However, the spectral mixing behavior of compacted, fine-grained mineral mixtures that would be characteristic of sedimentary depositional environments has received little attention. Here we characterize the spectral properties of pressed pellet samples of <10 micron mineral mixtures to 1) assess linearity of spectral combinations, 2) determine whether there are consistent over- or under-estimations of different types of minerals in spectral models and 3) determine if model accuracy can be improved by including both fine- and coarse-grained end-members. Major primary and secondary minerals found on the Martian surface including feldspar, pyroxene, smectite, sulfate and carbonate were crushed with an agate mortar and pestle and centrifuged to obtain less than 10 micron size. Pure phases and mixtures of two, three and four components were made in varying proportions by volume. All of the samples were pressed into pellets at 15000PSI to minimize volume scattering. Thermal infrared spectra of pellets were measured in the Vibrational Spectroscopy Laboratory at Stony Brook University with a Thermo Fisher Nicolet 6700 Fourier transform infrared Michelson interferometer from ~225 to 2000 cm-1. Our preliminary results indicate that some pelletized samples have contributions from volume scattering, which leads to non-linear spectral combinations. It is not clear if the transparency features (which arise from multiple surface reflections of incident photons) are due to minor clinging fines on an otherwise specular pellet surface or to partially transmitted energy through optically thin grains in the compacted mixture. Inclusion of loose powder (<10 μm) sample spectra improves mineral abundance estimates for some mixtures. In general, mineral abundances are predicted to within +/- 10% (absolute) for approximately 60% of our samples; thus far, there are no clear trends in which cases produce better model results. With the exception of pyroxene/feldspar ratios being consistently overestimated, there are no consistent trends in over- or under-estimation of minerals. The results described here are based on the unsubstantiated assumption that areal abundance on the pellet surface is equal to the volume abundance. Thus future work will include micro-imaging of our samples to constrain areal abundance. We will also prepareclay mixtures using a wetting/drying sequence rather than pressure, and expand our set of samples to include additional mixture combinations to further characterize the spectral behavior of compacted mixtures. This work will be directly applicable to analysis of TES and Mini-TES data of lithified sedimentary deposits.