Micro- to milli-watt fuel cell stacks and systems are designed and tested for applications that require continuous operation for multiple decades and at very wide ambient temperature ranges. Development presumes generic load profiles consisting of small baseline currents with frequent, periodic excursions to higher currents, both delivered at the 3.3 V standard integrated circuit voltage level. Stack design and testing was done between -55˚C to 80˚C using a baseline output current of 10 uA and periodic pulses of 4.5 mA lasting 100 ms, every 100 s. To simplify testing, the stacks are paired with 3.3V output miniature dc-dc converters to supply the respective 33 uW and 15 mW power requirements, with the stack then initially providing about 40 uW and 17 mW reflecting the converter efficiencies at the two sets of output currents/input voltages. Since maximum cell power tends to occur at >0.5 V/cell, 8-cell stacks were chosen so that, if called upon, the maximum power point still exceeds the ~4 V minimum buck converter input. Concerns about air availability and purity impel the use of H₂ & O₂ reactants; accommodating the fittings and other design constraints lead to a 13x19mm footprint (active area = 0.2 cm²), as shown in Figure 1. Since the stack reactant supplies are “dead-ended”, provision needs to be made to remove product water.

In contrast to typical (higher power) fuel cell applications, the primary system challenges with the microwatt system are reactant cross-over and leakage current. As an example of the severity of the issue, a 2020 technical target¹ for membrane leakage current is an equivalent DC resistance of 1000 Wcm². A similar value in this case would yield a cumulative leakage current of about 1.4 mA, causing an otherwise 30 y reactant supply for this system to last less than a year. Moreover, cross-over losses are generally hand-in-hand with leakage. The simple straightforward solution is to use thick membranes, done here primarily by stacking multiple layers of DuPont XL, which provides the benefits of radical scavengers and PTFE reinforcement scrims to enhance long-term chemical and dimensional stability. Other loss mitigation configurations and strategies are being pursued in tandem.

The need to provide 17 mW pulses not only limits the maximum membrane thicknesses (and stack resistances), but subjects the individual cells to voltage cycles between about 0.6V and OCV, similar to a catalyst durability AST (accelerated stress test). These issues can be alleviated by pairing the stack with a supercapacitor to minimize the voltage swings during the 17 mW pulses, thus allowing the stack to supply a fairly constant 65 uW (accounting for converter efficiency). Consequently, life and environmental testing are underway to demonstrate the utility of the configuration, but the preferred goal is to achieve the desired results without introducing additional components with potential lifetime and temperature range issues. Long-term room temperature life-testing is underway on stacks with and without supercaps. As of March 2018, a stack subjected to the full 17 mW pulses is approaching 5000 h of operation and has shown little if any drop in performance beyond an initial modest decline. A second stack paired with supercaps has been in operation nearly as long and likewise shows minimal sustained losses, even though it is being subjected to 10x the standard pulse rate.

Environmental chamber testing has demonstrated prolonged operation at temperatures ranging from -40˚ to 80˚C. Stack resistance decreases between room and high (80˚C) temperatures, overtly because the proton transport kinetics are enhanced in the already extremely dry membranes. While such dry operation was believed warranted to avoid product water freezing in the catalyst layer causing transport issues, the ability of the stacks to provide 17 mW pulses is challenged by the high stack resistance. Consequently, we have begun investigating the degree to which we can allow water content to increase while maintaining reactant access. Additional future efforts include environmental testing down to -55˚C and further stack design evolution to improve performance and functionality.

Figure 1. An 8-cell stack with nitrided titanium bipolar and endplates.

Reference

¹DOE Technical Targets for Polymer Electrolyte Membrane Fuel Cell Components. https://www.energy.gov/eere/fuelcells/doe-technical-targets-polymer-electrolyte-membrane-fuel-cell-components