Rapid Estimation of the Intermolecular Electronic Couplings and Charge-Carrier Mobilities of Crystalline Molecular Organic Semiconductors through a Machine Learning Pipeline

Organic semiconductors offer tremendous potential across a wide range of (opto)electronic applications. However, the development of these materials is limited by trial-and-error design approaches, as well as computationally heavy modeling approaches to evaluate/screen candidates using a suite of materials descriptors. For the latter, for instance, density functional theory (DFT) methods are widely used to derive descriptors such as the oxidation and reduction potentials, molecular relaxation and reorganization energies, and intermolecular electronic couplings; these calculations are compute-intensive, often requiring hours to days to determine. Such bottlenecks slow the pace and limit the exploration of the vast chemical space that can comprise organic materials. Here, we introduce a machine learning (ML) model to predict intermolecular electronic couplings in organic, molecule-based crystalline materials that take a few seconds, as compared to hours by DFT. Further, we use the ML model in conjunction with mathematical formulations to rapidly screen the charge-carrier mobilities and associated anisotropies of over 60,000 molecular crystal structures. The ML models and pipeline are made fully available on the open-access OCELOT ML infrastructure.