Computational Design of Novel Multiferroics
In the past decade, first principles computational tools, in particular the implementations of density functional theory (DFT), have achieved the power to not only support and explain experimental results but also to make predictions and design new materials. This so-called materials by design approach has been intensively used, especially in the field of oxides, to either come up with new compounds or optimize superlattice structures that give rise to new functionalities.
This research project focuses on magnetoelectric multiferroics, compounds that exhibit both magnetism and a macroscopic, switchable dipole moment. By using well-established evolutionary structure prediction algorithms (implemented, for example, in the USPEX package) interfaced with standard DFT (implemented, for example, in the Vienna Ab Initio Simulation Package), these researchers are determining new candidate compounds/structures and studying their properties, such as the magnetic order, magnetoelectric coupling coefficient, etc. Later stages of the project might possibly involve approaching these compounds with more state-of-the-art computational methods for strongly correlated systems, such as the Dynamical Mean Field Theory (implemented, for example, in the DMFT-Wien2K package). While this group's DFT calculations are not highly parallelizable, especially above couple of dozen cores, and have comparatively low memory needs, the evolutionary structure prediction is a length process, which requires considering thousands of different possible crystal structures to find the lowest energy one. As a result, the computation power required for this project can easily be very large.
Return to this PI's main page.