3D Microstructure Reconstructions of Solid Oxide and Proton Exchange Membrane Fuel Cell Electrodes With Applications to Numerical Simulations of Reacting Mixture Flows Using LBM

Abstract

Computational modeling of fuel cell electrode-catalyst layers is an important tool in understanding the different electrochemical reactions and transport phenomena occurring within fuel cell electrodes. Proper modeling of this layer is required for an accurate prediction of cell behavior which in turn can be used for the development of more efficient fuel cells. In macroscopic CFD approaches such layers are typically modeled as infinitely thin interfaces populated by sources and sinks or as very thin homogeneous porous layers. However, these layers are neither infinitely thin nor homogeneous and, thus, modeling in this fashion leads to a loss of information about the microstructure and its varying effects on the reacting mixture flows which pass through and into the structure. Thus, the utility of relying only on such macroscopic representations limits the general applicability of these macroscopic models as tools for design and for predicting fuel cell performance over a wide range of conditions. Furthermore, such macroscopic models cannot aid in the design of the electrode-catalyst layer itself. In order to address this latter point, a microscopic/mesoscopic modeling approach can be used, e.g., the Lattice Boltzmann Method (LBM), which models the reacting mixture flow through the porous microstructure of the electrode-catalyst layer. However, to do so requires reconstructing the porous geometry of this layer which can be done by using 2D microscopic images of cross-sections of the layer to generate 3D geometries from, for example, stochastic models which are relatively cost efficient and lead to similar structures with approximately the same characteristics of porosity, catalyst loading, three-phase boundaries, etc. as the original structure. Two such 3D reconstruction methods, i.e. one based on the granulometry law (one-point statistics) and the other on two-point statistics, are applied to a 2D SEM (scanning electron microscope) image of an SOFC electrode-catalyst layer and to the 2D SEM and TEM (transmission electron microscope) images for such a layer in a PEMFC. Results for these reconstructions are presented as are results for reacting mixture flow simulations through the two different reconstructed 3D SOFC structures using a 3D LBM approach. The development and application of a 3D LBM model for two-phase reacting mixture flows in PEMFC electrode-catalyst layer structures is in progress and will be reported in a future paper.