SéminaireComputational Mineral Physics: Predicting New Minerals in Earths Deep Interior

Résumé: Last 5 years have witnessed the discovery of a growing number of new mineral phases relevant to the Earths deep interior. These have dramatically changed our understanding of the Earths structure and dynamics, as well as the geochemical behaviour of some elements. A growing role in these discoveries is played by theory, especially after the development of an ab initio evolutionary algorithm USPEX which allows one to predict the stable crystal structure at any given P-T conditions, given just the chemical formula. Here I will discuss the following works done in my group: 1. Joint theoretical-experimental discovery of the post-perovskite phase of MgSiO3 by two groups, and an analogous phase of Al2O3. Post-perovskite is likely to be the major phase of the Earths D" layer, and its properties explain many of the known seismic anomalies of that region. 2. Prediction of a series of polytypic phases intermediate between perovskite and postperovskite , and of an unusual mechanism of plastic deformation of perovskite and post-perovskite. Both predictions have recently been confirmed experimentally. The predicted polytypes may become stable mantle phases due to the effects of impurities and temperature. The new mechanism of plastic deformation quantitatively reconciles seismological observations of D" anisotropy with mineral physics. 3. Confirmation of the counterintuitive experimental result that free metallic Fe should be present in the Earths lower mantle. 4. Prediction of two new structures of CaCO3 stable at pressures of the lower mantle, and of two new phases of FeS stable at pressures of Martian and Earth's cores. These predictions were confirmed by recent experiments. -------------------------------------------------------------------------------- Séminaires généraux de l'Institut de Physique du Globe de Paris --------------------------------------------------------------------------------