P12B-01:
Redox State of Iron in Lunar Glasses using X-ray Absorption Spectroscopy and Multivariate Analysis

Monday, 15 December 2014: 10:20 AM
Melinda Darby Dyar1, Molly C McCanta2, Antonio Lanzirotti3, Stephen R Sutton3, C J Carey4, Sridhar Mahadevan4 and Malcolm J Rutherford5, (1)Mount Holyoke College, South Hadley, MA, United States, (2)Tufts University, Medford, MA, United States, (3)University of Chicago, Argonne, IL, United States, (4)University of Massachusetts Amherst, School of Computer Science, Amherst, MA, United States, (5)Brown Univ, Providence, RI, United States
Abstract:
The oxidation state of igneous materials on a planet is a critically-important variable in understanding magma evolution on bodies in our solar system. However, direct and indirect methods for quantifying redox states are challenging, especially across the broad spectrum of silicate glass compositions found on airless bodies. On the Moon, early Mössbauer studies of bulk samples suggested the presence of significant Fe3+ (>10%) in lunar glasses (green, orange, brown); lunar analog glasses synthesized at fO2 <10-11 have similar Fe3+. All these Mössbauer spectra are challenging to interpret due to the presence of multiple coordination environments in the glasses.

X-ray absorption spectroscopy (XAS) allows pico- and nano-scale interrogation of primitive planetary materials using the pre-edge, main edge, and EXAFS regions of absorption edge spectra. Current uses of XAS require availability of standards with compositions similar to those of unknowns and complex procedures for curve-fitting of pre-edge features that produce results with poorly constrained accuracy.

A new approach to accurate and quantitative redox measurements with XAS is to couple use of spectra from synthetic glass standards covering a broad compositional range with multivariate analysis (MVA) techniques. Mössbauer and XAS spectra from a suite of 33 synthetic glass standards covering a wide range of compositions and fO2(Dyar et al., this meeting) were used to develop a MVA model that utilizes valuable predictive information not only in the major spectral peaks/features, but in all channels of the XAS region. Algorithms for multivariate analysis t were used to "learn" the characteristics of a data set as a function of varying spectral characteristics.

These models were applied to the study of lunar glasses, which provide a challenging test case for these newly-developed techniques due to their very low fO2. Application of the new XAS calibration model to Apollo 15 green (15426, 15427 and 15425), Apollo 17 orange (74220 ), and other lunar glasses confirms the Fe3+ contents from Mössbauer and shows that higher-Si glasses tend to have slightly elevated Fe3+.