Nanoscale properties of polymer fuel cell materials - A selected review

Kurzfassung

The properties of the components of a membrane electrode assembly in a polymer electrolyte fuel cell (PEFC) determine its efficiency and performance. This paper aims at demonstrating the importance of nanoscale properties of PEFC membranes and electrodes and discussing the information obtained by various experimental techniques. The nanostructure and conductivity of freshly prepared as well as artificially degraded Nafion membranes and Pt/C electrodes are investigated by contact atomic force microscopy (AFM), conductive AFM, pulsed force mode (PFM)-AFM, in situ scanning tunnelling microscopy (STM), and scanning electron microscopy. The different techniques can provide complementary information on structure and conductivity. With in situ STM on Pt catalyst covered graphite, a layer of very small Pt particles between the catalyst particles is imaged, which is probably not visible with TEM and can explain a systematic discrepancy between TEM and XRD in particle size distribution. Conductive AFM is used to investigate the conductivity of Nafion. The images show a quite inhomogeneous distribution of current at the surface. The percentage of conductive surface increases with humidity, but regions without any current still present up to 80% of relative humidity (RH). Comparison with PFM-AFM images, where differences in adhesion forces are measured, indicates that hydrophobic regions are present at the surface with comparable dimensions, which are attributed to non-conductive PTFE-like polymer backbone. The changes in hydrophilic and hydrophobic parts after artificial degradation by plasma etching in air plasma can be imaged by PFM. High-resolution current images of the membrane were used to directly compare the measured nanostructure of the single conductive channels with model predictions from the literature. Recent models in the literature propose the formation of water-filled inverted micelles, with a mean diameter of 2.4 nm, and their agglomeration into clusters agrees well with the current images