Dielectric constant prediction of polymers for organic solar cells and generation of library of new organic compounds

Abstract

This work is based on a rapid framework that has ability to design novel polymers for organic solar cells. Dielectric constant is predicted using machine learning (ML) models. In organic solar cells, the dielectric constant is critical because it influences the efficiency of charge separation and reduces recombination losses by stabilizing charge carriers. A higher dielectric constant can enhance exciton dissociation and improve the overall power conversion efficiency of the solar cell. 10,000 new polymers were generated, and their dielectric constant values were predicted using ML. Generated database of polymers is visualized using various measures. Polymers with higher dielectric constant values were selected and their synthetic accessibility was assessed to aid future empirical measurements. Additionally, chemical similarity analysis revealed structural resemblance among the selected polymers. This framework provides a quick and easy method for finding the efficient materials.