The operation of polymer electrolyte membrane (PEM) fuel cells requires a sensitive balance of the amount of water inside the cell. There must be enough water so that the membrane is well hydrated and has a high proton conductivity, but not too much so that liquid water accumulates and blocks the flow of reactants to the catalyst. In practice, this is achieved by increasing the flow rates of the gases, so that any excess water is simply blown out of the cell, and then humidifying the feeds to counteract any drying of the membrane, especially near the inlets. This solution adds a need for humidifiers and a recycling system to keep the fuel efficiency high, both increasing the cost and complexity of the fuel cell system. Our work takes a different approach by using low flow rates and exploiting the material properties and fundamental physics of the system.

Previous work in our group has shown the importance of two key aspects of the operation of fuel cells high fuel utilization: the structure of the gas diffusion layer (GDL) and the effects of gravity (Kimball, et al., AIChE J. 54(5), 2008). A model fuel cell with a single transparent straight flow channel and segmented anode allowed for the direct correlation of liquid water movement with the dynamic current distribution along the channel. The stability of operation was determined largely by the degree with which gravity assisted in removing the liquid water from the cell, ranging from small periodic fluctuations to huge oscillations with periods of only seconds. These effects were then coupled with the effects of the hydrophobicity of the GDL. Water flows through only the largest pores of the GDL with the size of the droplets that emerge on the surface determined by the size of the pore. This was exploited by forcing the largest pore to be in a certain location along the channel and either under the land or under the channel. The mechanism of how the water flowed into and along the channel affected the magnitude and frequency of fluctuations in the local currents as well as the uniformity of the current distribution along the channel. The effects of gravity were amplified when the large pore was under the channel, but diminished with the large pore under the land. These results have implications for application-specific system design, an issue that is not widely discussed, and for improving the durability of the membrane and electrodes, one of the major hindrances to increasing the mainstream use of fuel cells.