Two important flow field designs are parallel and interdigitated flow fields. In contrast to parallel flow fields, interdigitated flow fields cause convective transport between channels by forcing flow through the gas diffusion layer via pressure differences, also known as cross flow. Cross flow reduces the diffusion length between reactants and the catalyst layer, reducing performance losses, and helping to remove water from land areas. However, cross flow raises pumping losses.
This study experimentally examines the cathode bipolar plate channel/land width and channel depth while taking into account pumping losses. This has been experimentally examined for parallel flow fields, but has only been computationally explored for interdigitated flow fields, such as in Lin and Nguyen (A Two-Dimensional Two-Phase Model of a PEM Fuel Cell, Journal of the Electrochemical Society, 2006). This study seeks to understand how electrochemical and overall performance is affected by channel dimension in interdigitated flow fields, relative to parallel flow fields, from an experimental perspective.
Six bipolar plates with different combinations of channel/land width x channel depth were tested: 1 x 1 mm, 1 x 0.5 mm, 1 x 0.25 mm, 0.5 x 0.5 mm, 0.5 x 0.25 mm, and 0.25 x 0.25 mm. Channel and land widths were equal for the experiment, to keep a constant channel to land ratio. The bipolar plates are 20 cm long and 1.5 cm wide, for 30 cm2 cell area. The bipolar plates are aluminum coated in rhodium for corrosion protection. The bipolar plates fit into a cell superstructure that can use valves to switch from a parallel to interdigitated flow field in situ. Both parallel and interdigitated flow fields were tested, as well as two stoichiometries - high stoichiometry at 2.0 anode / 4.0 cathode and low stoichiometry of 1.5 anode / 2.0 cathode. Each combination of cathode bipolar plate, flow field and stoichiometry was tested three times in random order.
The polarization curve and associated pressure drop through the cell were measured for each bipolar plate combination. Based on these, the raw and net power density, and limiting current density of each condition are determined. The electrochemical performance results will be examined with respect to individual channel size, and overall bipolar plate permeability. The results will also be examined using relevant diffusive and convective length criteria.
Initial testing for this study has been completed and analysis is underway. In the analysis, it was discovered that the 1x0.25 mm plates suffered from GDL intrusion, and so they were removed. Some preliminary results have been obtained, such as the peak raw and net power density results in Figure 1. The peak raw power density and peak net power density, after accounting for pumping power losses, are plotted against the hydraulic diameter (HD) of the channels. The results show an optimal HD for net power density to be approximately 0.4 mm. There are still gains in performance for smaller channels, but these gains are overwhelmed by the pumping losses due to increased pressure to drive flow through the cell. The parallel flow fields experienced a smaller net power decrease than the interdigitated at low HD, as they incur a reduced pressure increase, due to the lack of cross flow.
Major remaining analysis includes determining the performance gains due to increased pressure in the cell, ohmic losses and the relative permeability of the bipolar plates. By subtracting off these gains (and losses) the actual electrochemical performance of the cells can be examined, as well as the net system performance. This will clarify what sort of diffusion and flow benefits are available at lower channel dimensions.
Once this study is completed, it will provide a guide for PEM bipolar plate designers on how the size & shape of channels, and the permeability of bipolar plates affect the performance of resulting fuel cell. It will also provide critical information on which design parameters are most important for each flow field. This can be used to design bipolar plates to fit a wide range of channel dimension or sizes in the future.