Designing promising molecules for organic solar cells via machine learning assisted virtual screening

Abstract

Navigating chemical space for organic photovoltaics (OPVs) is in high demand for further increasing the device efficiency, which can be accelerated through virtual screening of a large number of possible candidate molecules using a computationally cheap and efficient model. However, predicting the efficiency of an OPV is quite challenging due to the complex correlations between factors influencing the energy conversion process. In this work, we performed high-throughput virtual screening of 10 170 candidate molecules, constructed from 32 unique building blocks, with several newly built, computationally affordable and high-performing (Pearson's correlation coefficient = 0.7–0.8) machine learning (ML) models using relevant descriptors. Important building blocks are identified, and new design rules are introduced to construct efficient molecules. The critical molecular properties required for high efficiency are unraveled. Also, 126 candidates with theoretically predicted efficiency >8% are proposed for synthesis and device fabrication. Similar ML-assisted virtual screening studies may reveal hidden guidelines to design promising molecules and could be a breakthrough in the search for lead candidates for OPVs.