Frontier Molecular Orbital Offset as an Empirical Descriptor for Predicting Short Circuit Current of Nonfullerene Organic Solar Cells

An efficient computational approach, in contrast to the trial‐and‐error experiment process, for predicting, characterizing, and optimizing the macroscopic performance parameters (e.g., short‐circuit current density (Jsc)) of nonfullerene acceptors‐based organic solar cells (OSCs) remains a rarely addressed and complicated challenge. In this work, a data‐driven approach is used to predict the electrical performance of nonfullerene OSCs and reveal their charge transfer behaviors. The eXtreme Gradient Boosting (XGBoost) model within empirical descriptors is used to understand the governing feature for enhancement of Jsc, which is vital for the design and discovery of new donor/nonfullerene acceptor photoactive layers for photovoltaic applications. Through the well‐trained XGBoost model and SHapley Additive exPlanations theory, the descriptors impacting the Jsc of nonfullerene OSCs are further explained and analyzed. Remarkably, the XGBoost model combines four empirical descriptors to achieve an impressive prediction accuracy (R2 > 0.8). The results from data‐driven approaches prove that the lowest unoccupied molecular orbital (LUMO) offset (between donor and acceptor) plays the most significant role in increasing the Jsc values of nonfullerene OSCs. Moreover, this study highlights the effect of LUMO offset on the photoinduced charge transfer process of donor/non‐fullerene acceptor blends, which might pave the way toward rapid and precise energy‐level tuning of efficient OSC materials.