P. Carmichael, S. Kutagulla, Y. Lee, D. Akinwande
University of Texas at Austin,
United States
Keywords: fuel cells, graphene, defect engineering
Summary:
The insertion of graphene and other 2D materials into the ionomer membrane of a proton exchange membrane fuel cell (PEMFC) has recently gained significant attention in academia.1-2 The angstrom scale porosity of graphene is used as a size selective membrane for the prevention of hydrogen crossover. However, research in this field ubiquitously notes a tradeoff between improvements in hydrogen crossover and decreased membrane conductivity.1 While decreases of hydrogen crossover by up to 50% have been observed, this comes with decreases in the conductivity of the membrane by up to 50% as well.1 While the ohmic selectivity, defined by the ratio of ohmic conductivity to crossover, improves with the addition of the graphene membrane, the total current output decreases due to unexpected catalyst activation losses. By elucidating the mechanisms behind the activation loss, researchers in this field will be able to develop targeted solutions. In this study, we fabricate graphene/Nafion-211 membranes and pretreat using 0.1M HCl in an attempt to remove excess copper ions and protonate the Nafion. There are four categories of membranes tested: Baseline Nafion-211, acid soaked Nafion-211, single layer graphene (SLG) on Nafion, and acid soaked SLG/Nafion-211. Figure 1 shows the electrochemical data for each category of membrane. A decrease in crossover up to 30% is seen through linear sweep voltammetry in figure 1b, which is consistent with studies in the field. Polarization curves in Figure 1a reveal the conductivity of the membranes increase after acid treatment while maintaining low crossover; however, examination of the Tafel slopes in Figure 1c reveals the graphene layer causes a decrease in catalyst activation even after acid soaking. This is a curious effect, as the graphene membrane is not in contact with the catalyst layers. Further, while several studies in the field show this effect, there is no concerted consensus on why it occurs. We believe the graphene membrane acts as a barrier to water flow, causing a buildup of water on the anode where there is no water outlet. This buildup of water could be preventing molecular hydrogen from accessing the active sites of the platinum catalyst. As a result, the magnitude of the Tafel slope increases by about 30%, marking a significant loss in catalyst activation. The utilization of different 2D materials and defect engineering are the two most promising methods for preventing the drop in catalyst activity. We exposed graphene to UV-Ozone to introduce selective defects, facilitating greater fuel accessibility for the catalyst. Figure 2a demonstrates the three-fold increase in overpotential at 10mA/cm2 from open circuit voltage when graphene is introduced. Upon ozonation and UV exposure, this overpotential drops to baseline values. In combination with the improved selectivity seen in Figure 2b, this suggests the defect creation in the graphene membrane mitigates the impact on catalyst activation. These results mark an important improvement in the application of graphene in PEMFCs.