Artem R. Oganov
Laboratory of Crystallography, ETH Zurich, Switzerland
“Wonders Under Pressure: Novel Ab Initio Simulation Methodologies as a Tool to Study Matter at Extreme Conditions”
Thursday, February 8, 2007 @ 12:00 PM.
Earth and Space Sciences Bldg., Hanson Seminar Rm. 123
Abstract: Under extreme compression, matter often behaves counter-intuitively, and theoretical understanding of this behavior and our ability to predict it are still not fully developed. I will focus on some methodological breakthroughs that took place in the last 3-4 years, and their first applications:
(1) Studies of MgSiO3 post-perovskite phase, from its discovery to extensive calculations of its properties and chemical behavior. This discovery has clarified the nature of the mysterious Earth’s D” layer (~2700-2890 km depths).
(2) First applications of the ab initio meta-dynamics to geophysical problems: phase transition and plastic deformation mechanisms of the major Earth’s mantle-forming minerals – MgSiO3 perovskite and post-perovskite. This study has clarified the origin of seismic anisotropy of the D” layer, and predicted a polytypic series of structures intermediate between perovskite and post-perovskite as possible minerals in the Earth’s mantle.
(3) New high-pressure forms of the elements and compounds: new chain structure of sulphur, unusual ionic form of elemental boron, unique phases of calcium, new polymorphs of CaCO3 and MgCO3. These studies used a new powerful method for crystal structure prediction - the ab initio evolutionary algorithm developed by Oganov and Glass. Crystal structure prediction has long been considered an impossible task, but this new method turns out to solve this problem very efficiently and using only the knowledge of the chemical composition.
In spite of the breathtaking advances, many major problems remain unsolved. For instance, given just the constituent elements, it is not yet possible to predict simultaneously all stable compositions and structures. Transport properties of solids (e.g. thermal conductivity and viscosity), crystallization and nucleation dynamics are still beyond the reach of simulation methods. These and many other challenges will keep the field of computational mineral physics active for many decades to come.