3D Visualization of Liquid Water Penetration from MPL to Substrate Layer in PEFC Using in-Situ X-Ray Tomographic Microscopy

Abstract

Polymer electrolyte fuel cells (PEFCs) play an important role in solving environmental problems and energy problems. It is required to perform them at higher current density to achieve the cost reduction. Under the circumstance, however, the cathode gas diffusion layers (GDLs) are saturated with liquid water, which is generated from the electrochemical reaction and condenses. In other words, water is one of the factors that prevents the oxygen diffusion, and clarifying the water management in GDLs is essential to perform at higher current density. At present, synchrotron-based X-ray tomographic microscopy (XTM) is the common way to observe water distribution. While its high intensity makes it possible to get high contrast data by imaging in short exposure time, it decomposes the polymer electrolyte. It is the crucial point to keep cell performance during imaging. In addition, the experiment time using synchrotron radiation facility is limited. To inspect other ways than synchrotron-based XTM, here we aim to achieve 3D visualization of liquid water penetration in GDLs using laboratory-based in-situ XTM. Besides, we aim to evaluate the relationship between water distribution and GDL structure. To get 3D data of liquid water penetration in GDLs using Rigaku nano3DX, which is the laboratory-based X-ray imaging, the fuel cell with an active area 1.0 × 1.0 mm2 was applied. The cell equipped with Toray TGP-H-060 coated with micro porous layer (MPL) was used. 3D data that spatial resolution was 2.16 um was taken. By applying phase retrieval process and image processing, MPL, fiber, water and air were classified. As a result, we have succeeded in the 3D visualization of liquid water penetration from MPL to substrate layer under room temperature. Then, the volume of water captured by XTM was compared with that detected by X-ray radiography to verify that XTM could capture water precisely. This is because some water which is moving during imaging cannot be captured after reconstruction. The total scan time was about 3 hours, which is longer than that of synchrotron-based XTM. It took about 14 second to get an image by X-ray radiography. As a result, though the amount of water near the separator captured by XTM was fewer than that detected by X-ray radiography, the amount of water near MPL and substrate layer was almost the same. The results showed that the liquid water distribution obtained by laboratory-based XTM is reliable. Figure shows 3D images of cathode GDL and water at current density of 0.75, 1.0, 1.25, 1.5 A/cm2. We have identified that water passed through the cracks of MPL and remained in substrate layer where was close to the boundary between MPL and substrate layer. In addition, the effective pore diameter of substrate layer which water remained in was analyzed, and it was identified that water tended to remain in the bigger pore diameter. This is because substrate layer we used is hydrophobic. The result indicates that the existence of cracks on MPL has a strong influence on liquid water distribution and oxygen transport to the catalyst layer. Figure 1