Optimizing the Composition of the PEMFC Catalyst Layer

Abstract

PEMFC performance is very sensitive to the design of the catalyst layers of the membrane electrode assembly, in particular the cathode catalyst layer. The overall performance of the cell is affected by the catalyst density, porosity, electrolyte distribution, agglomerate size, agglomerate volume fraction, etc. The ideal values of these parameters vary throughout the catalyst layer, requiring a non-uniform 3-dimensional distribution. Therefore, it is useful to be able to determine this 3-dimensional catalyst layer composition at the time of design, based on anticipated operating conditions, cell physical geometry, and other design constraints.Previously we developed a gradient-based technique to optimize the catalyst distribution within the PEMFC catalyst layer. Our technique requires the computation of a sensitivity function for the platinum distribution using an adjoint space method originally developed by the applied mathematics community to solve 1-D optimization problems in fluid dynamics, climate, and heat transfer problems. The sensitivity functions are then applied using an iterative method to find the optimum platinum distribution throughout the catalyst layer. The computational cost required for computing the sensitivity functions using the adjoint space method is relatively small since this method requires solving only one sparse system of linear equations instead of performing multiple fuel cell simulations.Now we have extended this technique to allow for the simultaneous optimization of multiple catalyst layer design parameters, including electrolyte volume fraction, carbon volume fraction, and porosity, in addition to platinum. Our technique allows for constrained optimization (where the overall quantity of material being optimized is constant) and unconstrained optimization (where the overall quantity of material being optimized is allowed to vary). Constrained optimization is applied to platinum optimization, as the cost of the material limits how much can be used in practical applications. Unconstrained optimization is applied to optimization of electrolyte and carbon, as the material cost of these components in the catalyst layer is negligible.Using this technique, we are able to estimate the optimum 3-D distribution of platinum, carbon agglomerate, electrolyte, and void fraction within the catalyst layer to maximize the power density of the cell at different operating current densities. The optimum distributions depend on the positions of the landings and openings and the geometry and dimensions of the cell.The figure below shows optimization of electrolyte fraction, only, for a PEMFC. More details about the technique, the numerical implementation, and a number of fully optimized structures will be presented at the meeting.