Towards high-throughput exciton diffusion rate prediction in molecular organic semiconductors

Accurate property prediction is paramount to high-throughput computational screening of organic photovoltaic materials with superior solar power conversion efficiencies. This is challenging because many key processes depend on both intra- and inter-molecular properties. Here, we study exciton diffusion as an example of a key intermolecular process that exerts a significant influence on the functionality of organic photovoltaics and other organic semiconductor devices. The goals of this paper are: (1) to examine the accuracy of DFT + kinetic Monte Carlo (kMC) for predicting the exciton diffusion coefficients of fused-ring electron acceptors (FREAs) against experimental results, and (2) to explore computational simplifications with a view towards high-throughput screening. We perform a computational study of a series of six FREA materials with experimentally determined exciton diffusion coefficients, spanning 24 distinct FREA crystal structures. Based on these results, we: (a) show that kMC can indeed predict experimental exciton diffusion with reasonable accuracy based on FREA crystal structures, (b) derive a simple diffusion rate equation capable of approximately reproducing kMC simulations in the low disorder regime occupied by modern molecular semiconductors, and (c) demonstrate multiple ways to reduce the number of computationally-expensive quantum chemistry calculations required to calculate exciton diffusion coefficients. Our results reveal a viable path toward high-throughput virtual screening of organic semiconductors. Accurate property prediction is paramount to high-throughput screening of organic photovoltaics. Here, the accuracy of predicting exciton diffusion computationally is examined, and several simplifications towards high-throughput screening are explored.