Polymer electrolyte membrane fuel cells (PEMFCs) are considered as promising energy-delivery devices for automotive and portable applications. Commercialization of PEFCs relies on cost reduction through the curtailment in platinum loading (L_Pt, mg cm^-2), as well as the cell performance and durability improvements. Usually, PEMFCs operate at high current density to achieve high power output in the absence of high L_Pt. Nevertheless, large voltage loss accompanying low L_Pt caused by the vexing rise in oxygen transport resistance has been reported (Weber, 2014). The mass transport resistance localized to Pt particles, termed as the local transport resistance (R_Pt, s m^-1), has been identified as a main contributor (Greszler, 2012). Minimization of such resistance is imperative to accomplish better cell performance and electrode design.
In conventional PEFCs, CL is a complex porous structure, which consists of carbon supports for Pt particles providing pathways for electrons, ionomer networks for the transport of protons, and pore voids for the transport of reactants and water. Microscopy of CL reveals that the grape-like clusters containing all three constituents tend to form agglomerates with primary pores inside them much smaller than secondary pores between them. Oxygen moves from gas diffusion layer (GDL) to secondary pores, and then diffuses to primary pores. Electrochemical reactions take place in the vicinity of reaction sites, where oxygen is consumed together with electrons and protons, generating water as well as waste heat. Therefore, the local transport resistance depends on the detailed CL microstructures, as well as the operation conditions and the liquid water amount.
The present work introduces a multiscale framework that integrates microstructure reconstruction and lattice Boltzmann modeling to capture interfacial transport dynamics. Specifically, it quantifies how factors such as ordered ionomer alignment, accessible mesoporous carbon, hollow carbon supports, Pt loading gradients, the gas diffusion layer, and gas channels influence oxygen transport resistance. The catalyst layer’s microstructure, enhanced by controllable generative AI, is incorporated to establish strong structure-performance correlations. Additionally, the study investigates the time-sequential degradation of the catalyst layer’s transport performance. This research provides mesoscopic insights into reactive transport mechanisms through the development of a novel mesoscopic modeling framework, offering an efficient tool for the multiscale optimization of next-generation MEAs.

Fuel cell,Lattice Boltzmann simulation,catalyst layer