Machine Learning-Guided Gradient Dual-Proton Conducting Catalytic Layers for High Temperature Proton Exchange Membrane Fuel Cells in Aviation

• Machine Learning highlights PA volume concentration in CL for HT-PEMFC performance. • Multiphysics simulations and experiments reveal a detrimental PA gradient from membrane to GDL. • A novel EDTMPA-PA system enables continuous proton conduction in CL via hydrogen bonding. • Achieves a peak power density of 2.16 W cm −2 at 160 °C, advancing HT-PEMFCs for aviation. State-of-the-art proton exchange membrane fuel cell (PEMFC) for aviation application requires large radiator to avoid overheating. High temperature PEMFCs (HT-PEMFCs) using phosphoric acid (PA) as a proton conductor enable operating over 160 °C and anhydrous conditions, solving the heat rejection problems. Nevertheless, the dynamic PA redistribution in catalytic layers severely compromises their electrochemical performance. This study, using machine learning, clarifies that the PA volume concentration within the catalytic layers is the dominant performance factor, contributing 25.5% of fuel cell performance. Furthermore, multiphysics simulations for the catalytic layers reveal a detrimental PA gradient formation from membrane to the gas diffusion layer, causing progress deactivation of Pt catalyst by 50% in membrane-distal electrode regions. That has also been confirmed by electrochemical performance of gradient electrode layers with a 111.5% loss of Pt utilization in PA-deficient zones. To address this critical challenge, we propose an innovative dual proton conductor system combining ethylenediamine tetramethylene with PA (EDTMPA). The strong hydrogen bonding interactions between the phosphonic groups of EDTMPA and PA molecules create continuous proton conduction pathways and limit PA migration. This modification achieves a record peak power density of 2.16 W·cm − 2 at 160 °C, significantly advancing HT-PEMFCs for aviation application.