The Role of Redox on Bridgmanite Crystal Chemistry and Calcium Speciation in the Lower Mantle: Supporting Information

Published: 12-10-2020| Version 3 | DOI: 10.17632/hj3wg6s7yd.3
Neala Creasy


Authors: Neala Creasy1,2*, Jennifer Girard1, James Eckert1, and Kanani K. M. Lee1,3 1Department of Geology and Geophysics, Yale University, New Haven, CT, USA, 2Department of Geophysics, Colorado School of Mines, Golden, CO, USA, 3Lawrence Livermore National Laboratory, Livermore, CA, USA Corresponding author: Neala Creasy ( *Additional Supporting Information (Files uploaded separately in Mendeley data repository) Tables S1-S5 as .xlsx file Raw .tif files for all data used in this study Integrated .chi files for all data Integrated CeO2 .chi files for calibration of XRD data In this supporting information document, we include a wide variety of raw data tables, additional analysis, assumptions made in our modeling, and our Monte Carlo results: • Raw data for Mossbauer and starting materials; • Methods and raw EPMA results of our samples post-experiment; • Equations of state calculated for each phase identified and measured; • Rietveld refinements of two chosen x-ray diffraction patterns; • Our Monte Carlo methods, assumptions, and results; • Raw XRD data and integrated files Abstract: The amount of ferric iron Fe3+ in the lower mantle is largely unknown and could be influenced by the disproportionation reaction of ferrous iron Fe2+ into metallic Fe and Fe3+ triggered by the formation of bridgmanite. Recent work has shown that Fe3+ has a strong effect on the density and seismic wave speeds of bridgmanite and the incorporation of impurities such as aluminum. In order to further investigate the effects of ferric iron on mineral behavior at lower mantle conditions, we conducted laser-heated diamond-anvil cell (LHDAC) experiments on two sets of samples nearly identical in composition (an aluminum-rich pyroxenite glass) except for the Fe3+ content; with one sample with more Fe3+ (“oxidized”: Fe3+/ΣFe ~ 55%) and the other with less Fe3+ (“reduced”: Fe3+/ΣFe ~11%). We heated the samples to lower mantle conditions, and the resulting assemblages were drastically different between the two sets of samples. For the reduced composition, we observed a multiphase assemblage dominated by bridgmanite and calcium perovskite. In contrast, the oxidized material yielded a single phase of Ca-bearing bridgmanite. These Al-rich pyroxenite samples show a difference in density and seismic velocities for these two redox states, where the reduced assemblage is denser than the oxidized assemblage by ~1.5 % at the bottom of the lower mantle and slower (bulk sound speed) by ~2%. Thus, heterogeneities of Fe3+ content may lead to density and seismic wave speed heterogeneities in the Earth’s lower mantle.