Stochastically Modeled Gas Diffusion Layers: Effects of Binder and Polytetrafluoroethylene on Effective Gas Diffusivity

An improved stochastic reconstruction method for a gas diffusion layer (GDL) of proton exchange membrane fuel cell is developed to promote the accuracy in evaluating effective gas diffusivity. Carbon fibers are generated using stochastic algorithm within a representative element volume. Structural characteristics, porosity distribution and fiber orientation distribution are set as constraints in reconstructing the microstructure. Morphological opening of image processing with structuring element is employed to add binder and polytetrafluoroethylene (PTFE), with disk and sphere binder configurations. Pore-scale simulations are subsequently carried out to compute the anisotropic, effective gas diffusivities of these reconstructed GDLs. Simulation results show that the reconstructed GDL with binder and PTFE produces significant decrease of the effective gas diffusivity. The disk-shape binder appears to match the real GDL geometry visually, and the predicted effective gas diffusivity is also in good agreement with the reported experimental data in the literature. This demonstrates the importance of binder and PTFE in GDL reconstruction. Moreover, the correlations of the effective diffusivities in the through-plane and in-plane directions as functions of porosity and volume fraction of binder and PTFE are determined for the reconstructed GDLs.