Quantitative compositional analysis of sedimentary materials using thermal emission spectroscopy: 1. Application to sedimentary rocks

Abstract

Thermal emission spectroscopy is used to determine the mineralogy of sandstone and mudstone rocks as part of an investigation of linear spectral mixing between sedimentary constituent phases. With widespread occurrences of sedimentary rocks on the surface of Mars, critical examination of the accuracy associated with quantitative models of mineral abundances derived from thermal emission spectra of sedimentary materials is necessary. Although thermal emission spectroscopy has been previously proven to be a viable technique to obtain quantitative mineralogy from igneous and metamorphic materials, sedimentary rocks, with natural variation of composition, compaction, and grain size, have yet to be examined. In this work, we present an analysis of the thermal emission spectral (~270–1650 cm−1) characteristics of a suite of 13 sandstones and 14 mudstones. X‐ray diffraction and traditional point counting procedures were all evaluated in comparison with thermal emission spectroscopy. Results from this work are consistent with previous thermal emission spectroscopy studies and indicate that bulk rock mineral abundances can be estimated within 11.2% for detrital grains (i.e., quartz and feldspars) and 14.8% for all other mineral phases present in both sandstones and mudstones, in comparison to common in situ techniques used for determining bulk rock composition. Clay‐sized to fine silt‐sized grained phase identification is less accurate, with differences from the known ranging from ~5 to 24% on average. Nevertheless, linear least squares modeling of thermal emission spectra is an advantageous technique for determining abundances of detrital grains and sedimentary matrix and for providing a rapid classification of clastic rocks.

Quantitative compositional analysis of sedimentary materials using thermal emission spectroscopy: 2. Application to compacted fine-grained mineral mixtures and assessment of applicability of partial least squares methods

“…Figure compares NNLS‐derived mineral abundances with PLS‐derived abundances to that of XRD abundances for the mudstone samples from paper 1 [ Thorpe et al , ]. The PLS method retrieved known abundances to within ±10% for the majority of the mudstones, with no more than three samples falling outside this range.…”

“…Mineral abundance is determined by full pattern X‐ray diffraction quantitative refinements from Thorpe et al [].…”

“…In a companion paper [ Thorpe et al , , hereafter as “paper 1”], we characterize the spectral mixing behavior of a suite of terrestrial sandstones and mudstones to assess the accuracy of quantitative mineral abundance estimates derived from thermal emission spectra of sedimentary rocks. In this paper, we characterize the spectral properties of compacted, very fine grained (<10–25 µm) mineral mixtures to assess spectral mixing behavior of very fine grained materials.…”

Incorporation of Portable Infrared Spectral Imaging Into Planetary Geological Field Work: Analog Studies at Kīlauea Volcano, Hawaii, and Potrillo Volcanic Field, New Mexico

“…Next, we modeled the spectra with a set of 19 laboratory spectra using a nonnegative linear least squares minimization routine (Lawson & Hanson, ; Rogers & Aharonson, ). For thermal emission spectra of rocks and coarse particulates (>~60 μm), and using a library that captures the range of minerals present in the mixture, this technique is capable of retrieving areal mineral abundances to within 15% absolute accuracy (Feely & Christensen, ; Hamilton & Christensen, ; Ramsey & Christensen, ; Thorpe et al, ). Mineral detection limits depend on the mixture but generally range between ~5 and 15% (Christensen et al, ).…”

Radiative‐Transfer Modeling of Spectra of Planetary Regoliths Using Cluster‐Based Dense Packing Modifications

“…Additionally, 1,000-particle clusters preserve the balance between improvements in modeled spectra and the increasing computational burden with the use of more particles. We investigate midinfrared wavelengths in this work as the analytical methods for this wavelength range require substantial improvement when analyzing spectra of fine-grained regolith surfaces (e.g., Ramsey & Christensen, 1998;Rogers et al, 2007;Thorpe et al, 2015). Ito et al (2017) observed that the use of the superposition T-matrix method to compute the single-scattering characteristics of an elementary volume results in a spurious forward-scattering feature, wherein the magnification of the "diffraction" peak caused by the far-field effect of forward-scattering interference (Mishchenko, Figure 1.…”