Development of Machine-Learned Interatomic Potentials to Predict Structure, Transport, and Reactivity in Platinum-Based Fuel Cells

This research details the development of robust machine-learned interatomic potentials (MLIPs) specifically designed for complex multicomponent systems, with a primary focus on hydrated Nafion ionomers and platinum catalysts, which are critical components in advanced fuel cell technologies. The methodology involved constructing a diverse and comprehensive training set, meticulously designed to accurately describe both the bulk polymer properties and the intricate interfacial interactions between the catalyst and polymer. The developed MLIP was subsequently utilized to systematically investigate various properties of the platinum-Nafion system. Key findings demonstrated excellent predictive capabilities for structural characteristics and reaction pathways, particularly for phenomena contained within the training set. Furthermore, the model effectively captured the nuances of

proton transport mechanisms, including both vehicular transport and Grotthuss hopping. While converged calculations of diffusivities were not fully achieved due to the extensive computational demands of tens of nanoseconds, the model provided valuable insights into transport phenomena. The combined understanding derived from this computational approach offers significant potential for optimizing fuel cell performance, and the established methodology is broadly applicable to other chemical processes and devices where atomic-level understanding of structure, transport, and reactivity is essential for design and improvement.

Message: Hi FE, I need help with “Development of Machine-Learned Interatomic Potentials to Predict Structure, Transport, and Reactivity in Platinum-Based Fuel Cells” in “Chemical Engineering”