Integrated probing of cycling-induced degradation of multi-component electrode in hydrogen fuel cells via machine learning-empowered spectroscopic imaging

To improve the performance of proton-exchange membrane fuel cells (PEMFCs), the control of the spatial distribution of ionomer–Pt alloy catalysts on porous carbon supports is crucial because changes in their morphological and geometrical distributions are relevant to the performance degradation of PEMFCs upon operation. However, their changes remain poorly understood due to the absence of characterization tools with sufficient chemical sensitivity and spatial resolution. Here, an efficient machine learning-assisted electron energy loss spectroscopy is introduced to interpret cycling-induced morphological changes of the cathode at the nanoscale. This approach allows the reliable visualization of the three distinctive components of Pt alloy catalysts, ionomers, and carbon in the electrode. Furthermore, based on large data interpretation, changes in the ionomer–Pt alloy distribution and ionomer coverage on the carbon support can be statistically assessed in relation to the degree of structural degradation of the components upon cycling.