(262e) Long-Time Molecular Simulations for Linking Organic Semiconductor Morphologies to Carrier Mobilities

Organic semiconducting molecules have the potential to transform photovoltaic and optoelectronic device fabrication due to their low manufacturing costs and high abundance relative to inorganic semiconductors. A key challenge to realizing the transformative potential of organic semiconductors is understanding how to control the nanostructures they self-assemble into such that they are optimized for a given application. In this work we apply molecular dynamics simulations to understand how the ingredients in organic photovoltaic blends determine the structures they self-assemble into, and use kinetic Monte Carlo simulations to infer the degree to which these structures would make favorable or poor photovoltaic devices.

We develop coarse-grained models of poly(benzodithiophene-thienopyrrolodione) (pBDT-TPD) and fullerene derivatives and perform accelerated time sampling using rigid-body approximations for conjugated systems and graphics processing units (GPUs) to parallelize independent computational threads. Order-disorder transition temperatures for the photovoltaic blends are determined by high-throughput sampling, and the resulting thermodynamically stable morphologies are characterized by domain size, interfacial area, and network connectivity metrics. We compare our predicted morphologies and carrier mobilities against experimental scattering data and charge transport measurements and reason that the observed agreement positions our computational pipeline as a promising tool for broad structure-property studies of organic semiconductors.