Quantitatively reconstructing the microstructure of gas diffusion electrode for CO2 reduction combining WAXS tomography, FIB-SEM and generat

The performance of gas diffusion electrodes (GDE) for CO2 electroreduction (eCO2R) depends on the catalyst's position within the GDE, as this defines the contact between catalyst, CO2 gas, and aqueous electrolyte and thus governs reactant transport. This initial distribution changes during eCO2R through dissolution/re-deposition phenomena, and is complicated by the close presence of e.g. polymer additives or undesired (bi)carbonate crystals. Knowing the distribution of catalyst and additional phases within the GDE, both before and during eCO2R, is thus key to understand and improve the GDE's function. Here we propose to combine WAXS tomography with high-res FIB-SEM and machine learning by generative adversarial networks to reconstruct detailed tomographic images of GDEs containing microstructural and chemical information. This will provide a quantitative model of a GDE's morphological features, revealing their correlation to the eCO2R performance and enabling advanced electrode design.