Durability of Long-Life Low Power Fuel Cells

Abstract

Fuel cells have the potential to provide very low uninterrupted power for long periods (several decades) of time. The specifications for these applications can vary tremendously from transportation fuel cells that are a large driving force behind fuel cell development. For these applications, the fuel and oxidant supply are often required to be stored for the entire lifetime of the application; these systems have higher inherent safety and overall higher system energy density than batteries for these prolonged operational times. Many of the technical challenges are application dependent but can include: Fuel/oxidant loss (due to membrane/plates/seals permeability)Membrane degradation and thinning (due to high cell voltages)Passive system water managementOperation in uncontrolled environments (including operation a large temperature range and unpredictable/uncontrollable ambient atmospheres)Control of power transients (microwatt to watt power levels) The primary degradation modes expected for these applications involve the electrode layer, loss of catalytic activity, membrane thinning, changing hydrophobicity of carbon materials (which impacts water management) and corrosion of metallic bipolar plates (including increasing contact resistance). Electrode degradation involving performance loss of the cathode catalyst layer includes catalyst particle agglomeration, catalyst support corrosion, losses of catalyst layer porosity and loss of catalyst layer proton conductivity due to ionomer degradation. Depending on the systems load requirements, potential cycles can be many millions of cycles over decades long continuous operation. To minimize catalyst degradation, we expect that a Pt-black electrode will be superior compare to the Pt supported on carbon electrodes that are the preferred for transportation applications. To examine the durability of these catalysts we are conducting Accelerated Stress Testing (ASTs); Figure 1 shows an on-going test of a Pt-black catalyst over 1.3 million cycles. The square wave cycling induces Pt dissolution and re-precipitation (Ostwald ripening) of the platinum thus decreases the catalyst surface area. From Figure 1, catalyst surface area loss appears to have reached 30-35% loss, but has not substantially increased after about 750,000 cycles. To avoid large amounts of hydrogen loss due to membrane cross-over, we are exploring multiple layers of membrane to make a thick, mechanically reinforced and chemically stabilized membrane. A primary method to examine membrane degradation is to examine radical attack of the membrane at OCV conditions. Membrane thinning was measured at approximately 1.7 - 4.2 micron/year over a 2.4 year test with H2/air, at 25 C, which over a 20 year operational life extrapolates to34 - 83 micron. This suggests that a membrane of > 200 micron thick will be required for a 20 year operational life at low power levels. Figure 1