Multiscale Simulations for Organic Mixed Ionic Electronic Conducting Polymer Performance

Organic redox active semiconducting polymers are promising channel materials for organic electrochemical transistors in neuromorphic computing owing to their high transconductance, low power operation, and inherent biocompatibility. Despite this potential, systematic high throughput computational strategies for establishing structure–property relationships remain underdeveloped, largely due to the slow experimental cycle of synthesis, device fabrication, and characterization.

We develop a multiscale simulation framework capable of screening candidate polymers orders of magnitude faster than experiments. Charge carrier mobility and volumetric capacitance are identified as key figures of merit, since transconductance scales with their product ([μ][C*]). Volumetric capacitance is estimated using molecular dynamics with machine learned force fields combined with semi empirical electronic structure methods, yielding voltage dependent, morphologically relevant capacitance consistent with experiment.

Limited experimental data preclude large scale machine learning; instead, we extract physically grounded descriptors from the same simulations, including features derived from simulated X-ray scattering and small scale electronic structure calculations. These descriptors capture charge transport trends across diverse redox active polymer families. This framework provides a computationally efficient pathway for predicting and understanding the molecular origins of device performance in organic electrochemical transistors.