Deep Transfer Learning: A Fast and Accurate Tool to Predict the Energy Levels of Donor Molecules for Organic Photovoltaics

Molecular engineering is driving the recent efficiency leaps in organic photovoltaics (OPVs). A presynthetic determination of frontier energy levels makes screening potential molecules more efficient, exhaustive, and cost-effective. Here, a convolutional neural network developed to predict highest occupied lowest unoccupied molecular orbital (HOMO/LUMO) donor for OPV. The model takes 2D structure image returns prediction its HOMO/LUMO comparable experimental values. Insufficient datasets are overcome with transfer learning where initially trained on large Harvard Clean Energy Project dataset then fine-tuned using data from Organic Photovoltaic dataset. Error margins predicted below 200 meV achieved, without any chemical knowledge implemented. Noticeably, outputs have higher accuracy precision than corresponding density functional theory (DFT) estimations. limitations further tested home-built commercially available polymers reported OPVs (e.g., P3HT, PTB7-Th, PM6, D18). results demonstrate both practical utility this model, foster rational OPV optimization, deep techniques, general, revolutionize materials research development sector.