(177aa) Exploring the Contributions of Vehicular and Grotthuss Diffusion Mechanisms in Anion Exchange Membranes Using Molecular Dynamics Simulations

Anion exchange membrane fuel cells are of significant fundamental and industrial relevance due to their ability to use cheaper nonprecious catalysts compared to proton exchange membranes. However, the low conductivity of anion exchange membranes remains an ongoing challenge that has prevented their widespread adoption. An understanding of the hydroxide transport mechanisms through the polymer can help in the design of membranes with improved conductivity. To achieve this, we leverage molecular dynamics (MD) simulations for efficient screening of different polymer candidates, enabling a deep understanding of molecular-level transport mechanisms. In prior work, we evaluated the conductivity of AEMs from the self-diffusion coefficient of hydroxide ions, and while we were able to capture the right trends, our results showed a notable deviation from experimental values. We hypothesize that our force field’s inability to capture the reactive Grotthuss transport (proton hopping) is a reason for the discrepancy between simulations and experiments.

We attempt to bridge this gap by quantifying proton hopping using the Reacter protocol, that allows us to model chemical reactions in classical molecular dynamics simulations. We do this by employing the LAMMPS fix bond/react tool. This allows us to capture the contribution of proton hopping on ion diffusion, thereby improving our model's accuracy in predicting ionic conductivity. The simulations are performed on Tetraalkylammonium-Functionalized Polyethylene and poly(arylene ether sulfone) with tetra(quaternary ammonium). Chemistries are parameterized with GAFF2 force field and constructed using the Pysimm algorithm. Our findings underscore the significant contribution of Grotthuss mechanisms, especially in polymers with narrow hydration channels, consistent with prior research. Nonetheless, discrepancies with experimental values persist, partially due to inaccuracies in how conductivity is calculated in our simulations. Despite these limitations, our approach offers valuable insights into Grotthuss contribution without requiring the use of costly explicit reactive force fields, thus facilitating integration into high-throughput protocols. Future work will include improvements in estimating conductivity of AEMs from MD simulations.