Relation between oxygen gas diffusivity and porous characteristics under capillary condensation of water in cathode catalyst layers of polymer electrolyte membrane fuel cells

Abstract

We present a coupled simulation method for evaluating oxygen diffusivity in cathode catalyst layers of polymer electrolyte membrane fuel cells under capillary condensation of water partially filling the pore network. The voxel data of a catalyst layer formed with platinum-supported carbon and Nafion ionomer are prepared from sequential cross-sectional images taken by a cryo-focused ion beam scanning electron microscope. Capillary condensation of water, which effectively alters the pore network, is simulated using the lattice density functional theory. Then, the effective porosity εeff, chord length distribution, and characteristic length of the partially wet catalyst layer are evaluated. The effective diffusion coefficients of oxygen in the Knudsen regime are evaluated using the mean-square displacement method. The ratio εeff/τ, with τ being the tortuosity factor, roughly exhibits intermediate behavior between the Bruggeman correlations of sphere- and cylinder-filling for εeff ≳ 0.2, while εeff/τ drastically deviates from the Bruggeman correlation for εeff ≲ 0.2. The effective oxygen diffusion coefficients at 350 K are predicted as a function of relative humidity and total pressure based on the kinetic theory of gases. The proposed method gives deeper insights into oxygen diffusivity under capillary condensation, contributing to more efficient material design.