Patching Force Fields of Organic Materials Through Open Scientific Software Development
Additional Funding Sources
This project is supported by a 2021-2022 STEM Undergraduate Research Grant from the Higher Education Research Council.
Abstract
Solar cells made from organic photovoltaic (OPV) materials have the potential to provide sustainable solar power generation due to their low manufacturing cost and processability. Molecular dynamics (MD) simulations allow for the pre-screening of OPVs more efficiently than wet lab experimentation alone. Doing so requires the description of interaction potentials between simulation elements within the system. Therefore, the first step of simulating a new compound with MD is deciding how to apply forcefield parameters based on the chemical structure of the molecule, and if the chemical environment is not defined in the forcefield, then new parameters must be created. Here we develop and use computational tools for identifying bond, angle, and dihedral constraints that are missing from a forcefield, and perform quantum chemical calculations to parameterize these missing components. We set up the QUBEKit software stack on the Borah high performance computing (HPC) cluster, utilize SMILES strings to specify minimal molecular snippets, and parameterize models of Y6 and BTO, which have recently demonstrated power conversion efficiency over 17%.
Patching Force Fields of Organic Materials Through Open Scientific Software Development
Solar cells made from organic photovoltaic (OPV) materials have the potential to provide sustainable solar power generation due to their low manufacturing cost and processability. Molecular dynamics (MD) simulations allow for the pre-screening of OPVs more efficiently than wet lab experimentation alone. Doing so requires the description of interaction potentials between simulation elements within the system. Therefore, the first step of simulating a new compound with MD is deciding how to apply forcefield parameters based on the chemical structure of the molecule, and if the chemical environment is not defined in the forcefield, then new parameters must be created. Here we develop and use computational tools for identifying bond, angle, and dihedral constraints that are missing from a forcefield, and perform quantum chemical calculations to parameterize these missing components. We set up the QUBEKit software stack on the Borah high performance computing (HPC) cluster, utilize SMILES strings to specify minimal molecular snippets, and parameterize models of Y6 and BTO, which have recently demonstrated power conversion efficiency over 17%.