MID-INFRARED SPECTRAL CHARACTERIZATION OF SEDIMENTARY ROCKS AND THEIR CONSTITUENT PHASES

Sedimentary rocks and structures have recently been recognized as common features on the Martian surface. To help characterize these deposits and the aqueous environments in which they might have formed, quantitative mineralogical analysis from remote spectral measurements is required. Here, we assess the accuracy of mineral abundance estimates determined from mid-infrared (MIR) spectral measurements of a suite of clastic and chemical sedimentary rocks. In the process of analyzing terrestrial sedimentary materials, we are improving existing mineral spectral libraries with the addition of new phases, as well as exploring best practice data reduction techniques for quantitative mineralogical analysis from microspectroscopic image data sets. For a set of 50 rocks from Stony Brook sedimentology collections, classic sedimentary petrology techniques were compared to modern remote sensing applications. Traditional sedimentary petrology point-counts were conducted to establish estimates of true modal mineral abundance as well as crystal size. Powder X-Ray Diffraction (XRD) was also used to determine specific mineral phases in each sample as well as to corroborate mineral abundances estimated from point counts. Bulk rock thermal emission spectra, similar to those acquired on Mars by the Thermal Emission Spectrometer (TES), were collected from each sample. Bulk rock spectra will be modeled with a mineral spectral library using a linear least squares minimization routine and compared with known abundances determined from point counts and XRD. Preliminary results indicate a positive correlation between the traditional sediment mineralogy methods with that of thermal emission spectral characterization for fine grained sandstones; detailed results will be reported at the meeting. Portions of thin sections (~3-14 mm diameter) were also analyzed using a microscopic FTIR reflectance imager (~4 to 50 microns range). Micro-FTIR images were processed using principal components analysis and classification techniques, to identify major and/or unique components in the sample and their abundances. Preliminary results show that these standard data reduction techniques are useful and accurate means of rapidly deriving abundance from microspectroscopic images.