Optimization of Transports in a Proton-Exchange Membrane Fuel Cell Cathode Catalyst Layer at High Current Densities: A Coupled Modeling/Imaging Approach

Abstract

For PEMFC to attain sustainable commercialization, the amount of platinum (catalyst used in conventional PEMFCs) must be significantly reduced, while the performance must be improved. At high current densities, it is found that the nature of the losses is mainly mass transport related [1]. Since the catalyst layer is a highly heterogeneous composite material, understanding its structure is major in order to proceed with the optimization [2]. Therefore, in this work, an approach coupling both numerical modeling and electron microscopy characterization is adopted. An agglomerate/pore-scale model is developed and coupled with a MEA-scale model [3]. The model includes a description of the ionic and gas transport, plus the oxygen reduction reaction under a classical Butler-Volmer formulation. Multiple agglomerate structures are reproduced on the basis of a multiscale characterization setup. From 2-nm isotropic resolution FIB-SEM images, the carbon phase arrangement is built, while from HAADF-STEM a representative platinum particle size histogram is extracted and the particles are distributed on the surface of the carbon particles accordingly. On the HRTEM images, it is possible to observe the Nafion thin film covering the Pt/C agglomerates, and therefore, an average layer thickness can be extracted from these. The local pore size distribution is then obtained using the method of maximum sphere inscription [4], which allows to then determine the local gas diffusivities through the computation of local Knudsen diffusion coefficients. The agglomerate scale structures are upscaled to the catalyst layer scale through a mirroring procedure along the catalyst layer thickness, allowing therefore to analyze the cathode catalyst layer performance for different structure arrangements. In this work, the influence of the operating conditions and of certain structural aspects is analyzed: namely the Pt particle size, carbon aggregate size and Nafion layer thickness. It is found that decreasing the Pt particle size leads to a performance improvement since it is expected that more surface is available for the electrochemical reactions to take place. Still, the improvement margin is dependent on the Pt particle spatial distribution since particles located in zones having large Nafion agglomerates and low Nafion/pore contact simultaneously are found to severely underperform when comparing to particles located in other regions. Increasing the carbon aggregate size is found to be highly detrimental for the overall performance, even though an improved electrical contact is ensured. Such performance decay is linked to the large oxygen concentration drop that arises mainly from the porosity reduction, and to a certain extent, from the mean pore size reduction and the pore tortuosity increase. The Nafion layer thickness parametrization requires a more complex analysis since there is a delicate trade-off that must be explored in the sense that it is highly structure-dependent. On the one hand, an improvement in the ionic conductivity is expected upon increasing the Nafion layer thickness, though on the other hand, an increased resistance to oxygen diffusion is also expected. This trade-off is explored and discussed along with the introduction of the Nafion swelling effect, which is surprisingly not explored in the literature, neither on agglomerate/pore-scale models, nor macroscale models. It is found that increasing the Nafion layer thickness is detrimental for performance only when the porosity is not high enough. To complement the analysis, 'close-to-ideal' structures are built while holding the composition of the reference structures (i.e. electron microscopy based structures). It is found that these idealized structures lead to large performance improvements, showing that there is still much ground for further improvements on the PEMFC cathode catalyst layer.