Microstructure and Transport Analyses in Polymer Electrolyte Membrane Fuel Cells: Comparing Modeling Approaches

Abstract

Despite the recent entrance of proton exchange membrane fuel cells (PEMFC) into the automotive industry, performance under low platinum loading is still hampered by prohibitive and poorly understood transport losses inside the heterogeneous catalyst layer (CL). Predictive models do not yet sufficiently capture the transport of oxygen, water, and protons in the CL ionomer phase, due to the complexity of the CL microstructure. These transport rates are highly dependent on the substrate, film thickness, and relative humidity. Confinement effects limit water uptake and transport rates below roughly 100 nm thickness, and substrate effects can induce ionomer structures (i.e. lamellae) that further impact transport rates. Implementing these complexities into computational models allows for the microstructural and transport effects on PEMFC performance to be studied and better understood, especially in devices with low platinum loading. Research presented here examines two pseudo-2d, Newman-type model approaches for examining microstructure and transport in PEMFC functional layers: particle-shell and flooded-agglomerate. Figures 1 and 2 show the differences in model domains for these two approaches, respectively. The particle-shell approach approximates a physically realistic structure by simulating a carbon particle at the center of a thin shell of Nafion. This thin shell is discretized to resolve the complex transport that takes place within it. At each vertical discretization, a single Nafion-wrapped carbon particle is taken to be representative of others at that same functional layer depth. The flooded-agglomerate model also simulates a representative spherical domain for each vertical discretization; however, the domain is made up of a cluster of Nafion-wrapped carbon particles enclosed in an additional layer of electrolyte. The models predict cell polarization to examine the effect of CL microstructure as a function of RH- and thickness-dependent transport properties. Figure 3 shows an example overpotential curve from the flooded-agglomerate model in which interesting behavior can be seen with respect to the Nafion thickness. It shows that the CL overpotentials decrease with decreasing ionomer thickness down to 7 nm. Below 7 nm thickness, decreased proton conductivity limits transport, leading to increasing overpotentials at higher currents. The trends indicate that the balance between proton and oxygen transport rates in the CL Nafion can lead to possible trade-offs in optimizing CL design. These and other results will be presented and discussed, to demonstrate the coupled impacts of Nafion thickness and transport on PEMFC performance, and the impact of modeling approach on CL microstructure design. Keywords – Nafion, Transport, PEM, Fuel Cell, Modeling, Electrochemistry, Flooded Agglomerate Figure 1