Gen dataset of polymer donor materials of organic solar cells

Repetitive units of polymers are used to represent polymer molecules. The fragments are generated from the experimental molecules. The framework is used to generate polymers. Specifically, BRICS is used to decompose existing molecules into constituent fragments. These substructures are then categorized into core (C) (MolWt > 500) and spacer (S) (MolWt < 500). These fragments are then recombined through BRICS to generate new molecules in the order S-C-S-C'.C and C' represent different fragments from the core. The BRICS dataset is then used as input to VAE, which is used to generate more diverse structures. The Gen database is obtained from SMILES generated by BRICS and VAE. The Gen database includes a total of ~3.9 million polymers extracted from existing donor materials, which can be used for the discovery of new high-performance donor materials.