The Artificial Intelligence Guided Search for New CHIPs Comonomers

Chalcogenide Hybrid Inorganic/Organic Polymers (CHIPs) have the potential to revolutionize infrared (IR) optics and create sustainable and recyclable devices. CHIPs combine elemental sulfur with organic comonomers via inverse vulcanization to create a high-sulfur content polymer, with optical properties that rival state-of-the art inorganic solids with the processability and recyclability of plastic materials. However, the optimal comonomer for these applications remains unknown. This manuscript presents a gradient boosted tree model that determines which comonomers merit further consideration as high-performing CHIPs materials. After training models on previously calculated IR-absorption data we apply them to a larger set 960,966 molecules from the GDB dataset and validate the predictions for both highly transparent molecules and a set of one thousand randomly selected molecules. We then look at the 199,511 molecule subset of the expanded search space with chemical moieties eligible for inverse vulcanization and found 2,942 possible comonomers predicted to have better optical properties than the state-of-the-art comonomer Stillene. Finally, we calculate optical properties of all 2,942 comonomers in the gas phase and in a configuration to approximate the polymer films to find a new set of target comonomers.