Correlated electronic structure methods for point defects, grain boundaries, and interfaces

Computer simulations can accurately predict the properties of a hypothetical, perfect, and pure substance. However, real materials are imperfect: they contain trace amounts of the wrong types of atoms, atoms in the wrong location, dislocations, and other types of defects. Predicting the impact of these imperfections is a challenge for computer simulation. Moreover, the exotic properties of these defects can be exploited for technological applications. For example, some localized point defects in solids behave as 'qubits' and thus represent a promising platform for quantum computers. As another example, some rough surfaces with atomic imperfections are efficient catalysts, which speed up chemical reactions occurring near the surface. This proposal will develop new computer simulation tools that can accurately and affordably predict the properties of these defects, boundaries, and interfaces. The methods are designed to focus the computational effort on the interesting regions without completely ignoring the rest of the atoms in the material. The developed tools will be implemented in open-source scientific software and made freely available to community.