Polymer electrolyte membrane fuel cells (PEMFCs) are energy conversion devices that offer high power densities at low operating temperatures making PEMFCs the most promising technology for many applications such as automobiles, back-up power generating units, and portable devices. While design and material considerations for PEMFCs have a large impact on cost, it is also necessary to consider a transition to high volume production of fuel cell systems, including MEA components, to enable economies of scale and reduce per-unit cost. The fuel cell industry has identified quality control as a critical barrier for continuous production of MEA components, i.e. membranes, electrodes, and GDLs. One of the critical manufacturing tasks is developing and deploying techniques to provide in-process measurement of fuel cell components for quality control. This work focuses on a necessary subsidiary task: The study of the effect of manufacturing irregularities on performance with the objective to establish validated manufacturing tolerances for fuel cell electrodes. Membrane electrode assemblies with nominal active areas of 50 cm2 were prepared by spraying a catalyst ink formulation directly onto NRE212 polymer membrane material held at 80°C. The spray system was an ExactaCoat from Sono-Tek with an ultrasonic spray head. Catalyst loading variations were created by masking off 0.0625 – 1 cm2 areas. Within the defect area the nominal catalyst loading was reduced by various degrees up to 100% of the nominal value. The sample pool includes MEAs with different locations, shapes, and severity of coating irregularities, different nominal loading, as well as CCM vs. GDE structures. For spatial interrogation NREL’s high resolution segmented cell system was employed. The system consists of 121 segments of 0.41 cm2 area each arranged along the path of a quadruple serpentine flow-field. The segmented cell system was operated with a state-of-the-art single cell fuel cell test station. Figure 1 shows one sample set used to understand the performance effect of reducing total catalyst loading by 1% of the nominal loading. Rather than reducing catalyst loading in a single location only, variations of the 1% reduction were introduced to investigate if (i) location, (ii) shape, and (iii) intensity of the 1% coating variation matter. Differential spatial data was computed by subtracting the current distribution measured with a sample MEA from that of a pristine MEA. Figure 2 shows two such differential data sets. Areas with reduced performance are red, those with increased performance are blue, and those with unchanged performance are black. The data shows results for a sample with 2x 0.25 cm2 coating variations each having a 100% loading reduction, one near the inlet and one near the outlet (left) and for a second sample with a 1 cm2 50% reduction coating variation in the center. In addition, Figure 3 shows the total cell performance of the sample set. The total cell performance alone, as shown in Figure 3, does not indicate the presence of the defect. Instead, the detection of the impact of the defect requires a high resolution spatial diagnostic tool. As shown in Figure 2, the effect of the defect fades when reducing defect size. The areas that are 0.25 cm2 and have no catalyst loading show a smaller performance impact than the 1 cm2 area that only has a loading reduction of 50%. Obviously the impact of the coating variation has been distributed to more than one location. In any case, with respect to the total cell performance impact, the coating variation may not be classified as a defect, since no impact has been detected. A second criteria to classify the studied coating variation as defects is the effect on lifetime. These studies are currently underway and will be addressed in a second presentation. Figure 1
Read full abstract