In a polymer electrolyte fuel cell a catalyst layer is built typically of platinum nano-particles supported on a high surface area carbon support and resides between a solid-state electrolyte membrane and a highly porous gas diffusion layer. One performance test for a catalyst layer uses a rotating disk electrode (RDE), with a liquid electrolyte. The liquid electrolyte conditions are very different from solid-state, due to mobile anions. Also, reactions become mass transport limited quickly due to the relatively low concentration and diffusion of reactant gas in the liquid; therefore, extending the results from the RDE to a catalyst in a fuel cell is less than ideal.
A novel approach has been developed to study catalyst layers in realistic fuel cell conditions but carries the simplicity of the RDE. This used a three-electrode half-cell with a wall jet configuration, to blow humidified gas directly onto the back of the working electrode, allowing direct transport of the reactant gas to the electrode. The catalyst layer was pressed onto a polymer electrolyte, and a three-electrode electrochemical cell was completed using an iridium oxide reference electrode sandwiched within the polymer electrolyte between the working electrode and a coarse gold mesh counter electrode. In the presence of hydrogen, oxygen and nitrogen it also gave a stable potential (< 3 mV h-1 drift), although this potential changed with respect to the gases. An ultra-thin and ultra-low loading catalyst layer was optimised for studying the oxygen reduction reaction. This involved vacuum filtration of a catalyst ink onto a porous substrate, forming layers which were uniform and as thin as 1 μm, with a platinum loading as low as 5 μgPt cm-2. The catalyst was supported on a gold sputtered polycarbonate membrane with Teflon AF coated pores, to act as an ultra-thin gas diffusion layer and low resistance current collector.
These ultra low loading catalyst layers supported on a porous substrate were first tested in contact with an aqueous electrolyte, achieving a current density of 165 mA cm-2Spec or 680 mA cm-2Geo at 0.3 V vs. RHE. Above 0.6 V vs. RHE (fuel cell relevant potentials), a clear curve was visible. Below this, the curve straightened, showing the beginnings of mass transport and resistive losses. Upon transferring the electrode to the solid state electrochemical cell, a current density of only 12 mA cm-2Spec was attained. This 10 fold drop in performance was believed to be because of poor bonding of the PFSI membrane across the reference electrode leading to non-uniform current distribution across the working electrode.