AI3SD Project: Deep-Learning-Enhanced Quantum Chemistry: Pushing the Limits of Materials Discovery

Abstract

Discovery of new functional materials is central to achieving radical advances in societally important challenges (efficient energy materials, organic solar cells, etc). The vast space of possible materials combinations and compositions gives hope that useful materials exist, but finding ways to navigate this vast space remains a fundamental challenge to materials research. Quantum theoretical computational materials research based on density functional theory has revolutionised the search for new materials over the last twenty years with its ability to predict materials properties based on atomic structure. However, the computational cost of solving quantum mechanical equations at high-performance computing centres remains a severe bottle-neck to its commonplace use in research. In this project, we use machine learning to develop an accurate quantum mechanical simulation method of structural, optical, and electronic properties of hybrid organic-metallic materials used in modern solar cells that is efficient enough to run on standard desktop computers for systems that include many thousands of atoms.

More information

Identifiers

ORCID for Samantha Kanza: orcid.org/0000-0002-4831-9489

ORCID for Jeremy G. Frey: orcid.org/0000-0003-0842-4302

Catalogue record

Export record

Altmetrics