Structure and Thermoelastic Properties of Al- and Fe- Bearing High Pressure Silicate Minerals

University Of California-Berkeley, Berkeley CA

Investigators

Abstract

Bukowinski EAR-0003456 Although the gross outlines of the mineral composition of the Earth's mantle are reasonably well known, geophysics remains unable to interpret seismic shear properties and tomographic maps of lateral heterogeneity in the lower mantle and transition zone with much confidence. Laboratory measurements of elastic wave velocities are still confined to pressures corresponding to the upper mantle. In addition, little is known about the effects of Al and Ca on the properties of (Mg,Fe)SiO3-perovskite, likely the dominant component of the lower mantle. Recent experiments indicate that a small amounts of Al in silicate perovskite significantly enhances its compressibility, and hence forces a re-examination of compositional models that ignore Al altogether, or assume that it has little effect on the elasticity of silicates. The effect of Al is known to be sensitive to the concurrent presence of Fe3+, but the mechanism is not fully understood. The role of Al and Ca in stabilizing garnets below the 660 km discontinuity also needs to be better understood. Even less is known about the mineralogy of the D'' zone, whose properties gain in complexity with every new seismic examination. We propose to develop a model of silicate minerals that will allow an accurate examination of elastic properties by combining low pressure data with density functional theory. This semi-empirical model will be based on the Variationally Induced Breathing (VIB) theory of bonding in "ionic" materials. In addition to the proven density-functional energy, the new model will incorporate parametric covalent contributions, which will compete with inter-ionic charge transfer in response to electronegativity equalization. The parameters will be constrained with low pressure data on elastic velocities and vibrational spectroscopy. Preliminary investigations show this to be a very promising approach. The unique strength of the method lies in its high efficiency relative to fully first-principles approaches. Furthermore, the method promises to effectively bootstrap high quality low pressure elasiticity data into deep Earth conditions, thus greatly enhancing their utility. It will thus be possible to efficiently search for mineral structures and to examine their dependence on various geophysical conditions. The fully developed model will be used to generate phonon spectra, elastic constants and equations of state. It will also be used to examine the effects of coupled Al and Fe3+ substitution into (Mg,Fe)SiO3 perovskites. This will allow the generation of seismic velocities for candidate lower mantle mineral assemblages, and an examination of how they are affected by crystal structure and the presence of relatively low abundance components like Al, Ca and Fe3+. An attempt will also be made to evaluate the temperature dependence of these properties. The minerals to be investigated include silicate perovskites, garnets, various SiO2 phases, and other minerals composed of MgO, SiO2, CaO, Al2O3, and FeO that may be present at lower mantle and D'' zone conditions.