The specific morphology of porous materials, as well as their physical properties, crucially affect their applications, e.g. their use in fuel cells, batteries, or electrolysers. A key point is the correlation of their transport properties with the spatially distributed pore structure. In polymer electrolyte membrane water electrolysers (PEMWE), the drainage of oxygen, counter-currently to the water supply inside the porous anode, can significantly control the performance of the system. More precisely, the temporal and spatially distributed oxygen coverage of the catalyst layer obstructs the water supply towards it, which can result in a decrease of reaction rate. At the same time affects the heterogeneous distribution of liquid water the temperature distribution inside porous electrodes. This can have a severe impact on the performance of the PEMWE as well as on the degradation of the catalyst material.

In the presented study, mathematical pore network modeling (PNM) was used to investigate the correlation of the porous transport layer (PTL) pore structure with the distribution of water (wetting) and oxygen (non-wetting). The pore network was reconstructed based on the information about porosity, pore size distribution, pore connectivity and pore shape, which were obtained from micro-computed tomography of commercial materials (Fig. 1a). Compared to existing macroscopic models, PNM is suitable for low capillary numbers, takes local wetting characteristics into account and resolves the invasion process at the pore scale. The quasi-static nature of the model provides an edge over other modelling techniques in terms of computational power and additionally the ease of implementation of locally distributed properties. With the help of PNM, it was studied how a change of structure, i.e. the spatial grading of the pore size distribution and porosity, affects the local oxygen and water distribution, as well as the transport properties. This phenomenon was also analyzed experimentally by the use of a non-electrochemical aluminum cell (Fig. 1b and 1c) which was employed in an in-situ imaging setup installed at ILL, Grenoble, France. The recorded water saturations can be compared to the simulation results with the purpose of model validation and further model development. The outcomes of this study can be useful for the development of structural grading strategies with the target of improved transport properties for the efficient removal of the reactant oxygen and the homogeneous distribution of water.