Combining quantum chemistry, machine learning and rate theory for organic luminescent materials

The theoretical design of highly efficient, low roll-off and full-color emission organic materials is of great interest, although there are great challenges due to the limitations of the present-day methodology. In this review, we present progress achieved in our group on the theoretical and computational investigation for the structure-property relationships and screening strategy for organic fluorescent molecules, selection of thermally activated delayed fluorescence (TADF) and multi-resonance TADF (MR-TADF) molecules for optically and electrically pumped lasing application, and high-throughput virtual screening of phosphorescent organometallic complexes. We combined a quantum chemistry method with the molecular representation learning model Uni-Mol and rate theory-based molecular material property prediction package (MOMAP) developed in our group. Finally, we outline the limitation of current computational protocols and the future directions for organic luminescent materials.