Optimal gradient designs of catalyst layers for boosting performance: A data-driven-assisted model

Improving the platinum (Pt) utilization is essential to reduce its loading in proton exchange membrane fuel cells (PEMFCs). The gradient design in cathode catalyst layers (CLs) is reported to improve the performance of PEMFCs, but lacks general criteria. To this end, we investigate the performance of CLs with gradients in ionomer and Pt loading along the thickness direction under different relative humidity (RH) conditions based on the agglomerate model. The homogeneity of reaction rate in CLs is improved due to the gradient design. A data-driven model integrated with genetic algorithms is then developed to determine the RH-dependence optimal structure parameters for both the non-gradient and gradient CLs. We reveal how variations in Pt and ionomer loading within gradient cathode CLs improve the performances of PEMFCs. Leveraging RH-independence insights from the data-driven optimization model, we propose a general approach for fast predictions of optimal structures for both the non-gradient and gradient CLs, boosting both the power density and limiting current density simultaneously.