Molecular Dynamics Simulation of Polymer Electrolyte Membrane for Understanding Structure and Proton Conductivity at Various Hydration Levels Using Neural Network Potential

Alkyl sulfonated polyimides (ASPIs), as alternative polymer electrolytes for fuel cells, are known to exhibit lyotropic liquid crystalline behavior upon water uptake, forming organized lamellar structures and achieving high proton conductivity. Previous experimental studies have shown that ASPIs with planar backbones exhibit enhanced proton conductivity (0.2 S/cm) compared to those with bent backbones (0.03 S/cm). To explain this difference at the atomistic level, molecular dynamics simulations were conducted using a universal neural network potential. The appearance of monomer unit length in planar ASPIs, indicating higher molecular order, was found to correlate with higher proton conductivity compared to that of bent ASPIs. Despite the similar deprotonation and solvation of sulfonic acid groups in both planar and bent ASPIs, the proton conductivity was independent of these factors. Directional mean square displacement analysis provided further insights into the differences in proton conductivity between planar and bent types.