Dresden 2020 – scientific programme

The DPG Spring Meeting in Dresden had to be cancelled! Read more ...

Parts | Days | Selection | Search | Updates | Downloads | Help

MM: Fachverband Metall- und Materialphysik

MM 30: Poster Session II

MM 30.32: Poster

Tuesday, March 17, 2020, 18:15–20:00, P4

The Process for Creating A General-Purpose Machine Learned Potential for Silicon Carbide — •Harry Tunstall, James Kermode, and Gabrele Sosso — The University of Warwick, Coventry CV4 7AL, United Kingdom

SiC is a prototypical material for high temperature applications (e.g aerospace, automotive and thermoelectric) involving complex microscopic processes typically inaccessible to experiments. To gain insight into the functional properties of e.g. SiC nanostructures, computationally expensive quantum mechanical methods such as density functional theory (DFT) must be employed. This is because less computationally demanding methods are almost always not accurate enough. In fact, similar to Si and C alone, various empirical interatomic potentials have been developed for SiC, such as Tersoff or Stillinger-Weber. These potentials are designed to reproduce specific features of the material, at the expense of transferability to a wider range of functional properties.

The aim of this project is to build a general purpose interatomic potential for SiC (e.g using machine-learning regression, in the form of Gaussian approximation potentials (GAPs) and neural networks (NNs), starting from a DFT dataset of representative configurations), enabling accurate large scale simulations into defects and grain boundaries. The methodology for creating and maintaining a machine learning database for atomic systems from the ground up will be discussed, from the perspective of this ongoing project.