Charge Transfer Landscape Manifesting Structure-Rate Relationship in the Condensed Phase via Machine Learning

In this work, we develop a machine learning (ML) strategy to map molecular structure to condensed-phase charge transfer (CT) properties including CT rate constants, energy levels, electronic couplings, energy gaps, reorganization energies, and reaction free energies, which are called CT fingerprints. The CT fingerprints of selected landmark structures covering the conformation space of an organic photovoltaic molecule dissolved in explicit solvent are computed and used to train ML models using kernel ridge regression. The ML models show high predictive power with R2>0.97, and both mean absolute error and root mean square error within chemical accuracy. The CT landscape for millions of molecular dynamics sampled structures is thus constructed, which allows for instant prediction of CT rate properties given any molecular structure. The unprecedented CT landscape will shed light on real-time CT dynamics in nanoscale and mesoscale condensed-phase systems, and the optimal fabrication design for homogeneous and heterogeneous optoelectronic devices.