Glass and Phyllosilicate Obfuscation (and Clarification) on Mars

Several authors have postulated that contributions from glasses and phyllosilicates to thermal emission spectra of the Martian surface may be difficult to distinguish TES spectra. We explore this presumption by generating 336 sets of two-component numerical mixtures of glass and phyllosilicate spectra and then deconvolving these mixtures using a linear least squares algorithm with a variety of end member suites. Three end member scenarios are explored for laboratory, TES, and THEMIS-resolution mixtures: 1) a variety of glasses and phyllosilicates including those used to make the mixture are included in the end member suite (a control case), 2) the glass and phyllosilicate spectra used to make the mixture are excluded from the end member suite, and 3) only the phyllosilicate used to make the mixture is excluded. For each mixture and each end member scenario, the modeled abundances of glass and phyllosilicate are tracked, as are the compositions of the mineral end members used in substitutions. Model results indicate that exchange of glasses for phyllosilicates and vice versa does occur when one or both end members used to make the mixture is excluded. The degree of substitution is variable depending on the specific excluded members, however, a few trends are observed. In end member scenario two, glass is underestimated by 10-64% (absolute) if Si-K₂O glass is in the mixture, and it is overestimated by 11-50% (absolute) if pure Si-glass is in the mixture. Scenario three shows up to 20% absolute difference from the known fraction of glass content, but most analyses show less than 10% absolute difference. This result suggests that when the correct glass end member is included in the end member suite, phyllosilicate end members only minimally substitute for glass end members.