3D classification of polymer electrolyte membrane fuel cell materials from in-situ X-ray tomographic datasets

• Published in: International journal of hydrogen energy 2020-04, Vol.45 (21), p.12161-12169
• Main Authors: Ince, Utku U., Markötter, Henning, Ge, Nan, Klages, Merle, Haußmann, Jan, Göbel, Martin, Scholta, Joachim, Bazylak, Aimy, Manke, Ingo
• Format: Article
• Language: English
• Subjects: Image processing; Material classification; Polymer electrolyte membrane fuell cell; Quasi in-operando synchrotron X-Ray tomography; Water management
• ISSN: 0360-3199; 1879-3487
• DOI: 10.1016/j.ijhydene.2020.02.136

To conduct simulations of transport properties within fuel cell materials 3D-data are required as input. For a functional simulation of flow and thermal characteristics the morphological features of the gas diffusion layer (GDL) materials must be well-defined. In this regard, the distribution of the distinct substances in the GDL, each of which is represented by a different parameter set, plays a decisive role. By means of synchrotron X-Ray computed tomography a fuel cell equipped with SGL® 28BC GDL material is recorded in 3D. The segmentation of the solid structural components and the identification of water agglomerations in the components are fulfilled by image processing techniques. This step is often realized using a lot of simplifications, like the reduction to only one or two different materials or idealized structural characteristics. In the presented work 8 different phases are defined according to the cell materials, which are the catalyst layer, the carbon fibers, the Teflon (PTFE)-binder, the micro porous layer (MPL), the flow field ribs, the pore spaces as well as the water in pore spaces of the GDL substrate and in the MPL. The image processing steps used for the classification of each voxel into these phases are described in detail in this work. Benefiting from this classification methodology, the macroscopic properties of the GDL materials such as water saturation, diffusivity, thermal and electrical conductivity can be obtained in simulations. •Operated PEMFC recorded by synchrotron X-Ray tomography.•Described 3D image processing allows for classification of eight different phases.•Classified 3D data highly suitable for simulations of water transport properties.