Predicting MEA catalyst layer performance at 120 °C using a pressurized GDE testing platform and machine learning

A pressurized gas diffusion electrode (GDE) platform that extends half-cell testing to 120 °C is employed to provide a medium-throughput evaluation of proton exchange membrane fuel cell (PEMFC) catalyst layers under conditions relevant to heavy-duty PEMFC operation. Using a straight-forward catalyst layer fabrication procedure in combination with data-driven modeling, we systematically optimized five critical fabrication parameters and highlighted their individual influence on performance. Importantly, we systematically validated whether trends observed in pressurized GDE testing transfer to membrane electrode assembly (MEA) operation, an assumption often made but rarely validated experimentally. We show that the performance trends identified in the pressurized GDE setup can be consistently reproduced in MEA testing, with catalyst layers exhibiting higher GDE performance also displaying superior MEA performance in the same relative order. Together, these results demonstrate that medium-throughput pressurized GDE screening, when combined with machine-learning analysis, provides a robust framework to accelerate the transfer of basic to applied research. Catalyst Layer Optimization via Pressurized GDE Testing at 120 °C, Confirmed Transferable to MEA Performance . • Pressurized GDE platform enables catalyst testing at 120 °C and low humidity. • Machine learning reveals key fabrication parameters for catalyst layers. • Optimized catalyst layers retain high activity under dry and humid MEA conditions. • GDE-derived trends reliably predict MEA performance at 120 °C. • Framework accelerates catalyst-layer design with reduced experimental cost.