In silico materials design: porous high-pressure intermetallics and rattling Heusler compounds.
I will give a talk on my recent reserach results at the Cornell University. The talk will be aimed at anyone interested in how to use computer simulations effectively to investigate materials. The focus will be on structure prediction methods, and how they can be applied to high-pressure materials discovery, but also in the context of large-scale materials screening projects.
Abstract: With increasing computational resources and novel, efficient algorithms, it has become possible to predict materials solely based on computer simulations prior to their experimental synthesis. A key challenge in materials prediction is to find the correct arrangement of atoms in a crystal lattice, the crystal structure, which fundamentally defines the physical, chemical and mechanical properties of a compound. Reliable crystal structure prediction methods are especially valuable to study materials at conditions where experimental structure determination is impossible or very difficult, for example at extreme pressures. In this talk, I will present a crystal structure prediction algorithm based on the minima hopping method (MHM), and examples of its successful application in conjunction with density functional theory calculations. On one hand, I will focus on high-pressure materials discovery, particularly in alloy systems that are known to exhibit strong immiscibility at ambient conditions. Elements that commonly don’t bond at all can form compounds once a sufficiently high pressure is applied, resulting in intermetallics that exhibit superconducting or exotic, low-dimensional structural properties with potential use in metal-ion batteries. On the other hand, I will show how structure prediction can be combined with high-throughput and “big-data” efforts to screen a large chemical space for new energy materials. In particular, I will present our recent discovery of ultralow thermal conductivity in a new class of Heusler semiconductors for thermoelectric applications.
The announcement can be found here.