In order to further improve power density of PEFC stack, the operations at high current density and high voltage are required in addition to downsizing. To achieve 6.0kW/L around 2030, the extreme targets such as 0.7V@3.0A/cm² and 0.66V@3.8A/cm² being proposed. However, such operations cannot be achieved by the currently common MEA and GDL, and it is difficult to experimentally investigate the issues of internal states at such very high current density. In the present study, we carried out simulation of fuel cell performance by our originally developed software P-Stack. The results indicated that improving GDL properties is very important in addition to enhance performances of PEM and catalyst layer.

The most important feature of the software is the employed macroscopic model comprehensively dealing with all the relevant transport phenomena coupled with electrochemical reactions. In anode and cathode channels and GDLs, mass transport of gas and liquid phase are solved by considering the two-phase fluid dynamics with taking into account the effects of phase changing. Here mass transport is coupled with heat and chemical species transport equations. Those all transport phenomena are coupled with electrochemical reactions in the MEA. Electrochemical reactions are modeled by Butler-Volmer equation. The parameters in the equation and other models need to be fitted to reproduce the fuel cell performance with target MEA and GDL for different operating conditions.

First, we assembled 1cm² cell with a typical MEA (PEM: Nafion NR211, Ca/An CL: TKK TEC10E50E 0.3mgPt/cm², Ca/An GDL: SGL28BC), and then measured I-V and I-R curves under various operating conditions and transient behavior when GDL flooding occurred. These experiments were carried out with massive flow rate and constant temperature, and therefore it can be assumed the in-plane distributions are uniform. In order to reproduce the measured results of the 1cm² cell by simulation, we second fitted the various simulation parameters including electrochemical parameters and mass transport resistance of cathode catalyst layer and the parameters related to GDL flooding. As a validation of the fitted parameters, we compared measured and calculated I-V and I-R curves for 25cm² dual serpentine cell with the reference MEA. The figure (a) and (b) show the measured and calculated I-V curves under the different RHs, stoichiometry and back pressure, and it is indicated that the calculated results are in good agreement with the measured values.

Figure (c) shows the simulated I-V curves of 300cm² cell under the conditions: anode RH is 80 %, cathode RH is 40%, cathode stoichiometry is 2.5, back pressure is 150kPa(g) and coolant flow rate is 1.0L/min/cell. In order to clarify the influence of GDL flooding, we carried out calculations with and without the consideration of liquid phase changing. The blue broken line and dots represent the results of the reference MEA. The calculations with GDL flooding (dots) were performed at the few points because they are very time consuming. The results indicated that the cell voltage of the reference MEA decreases more than 200mV at 3.0A/cm² due to the influence of GDL flooding. In figure (c), red broken line and dots represent the result of hypothetically modified MEA in which proton conductivity, catalyst activity, less mass transport resistance, through-plane electric/thermal conductivity of GDL and gas diffusivity of GDL are highly enhanced to achieve 0.7V@3.0A/cm². The figure (d) and (e) show the in-plane distribution of temperature of PEM and liquid water saturation in cathode GDL at 3.0A/cm². Although the details will be discussed in the presentation, it was indicated that the further improvement of gas diffusivity and through-plane thermal conductivity are very important to reduce the GDL flooding and overheat of PEM, respectively.

References

[1] Y. Ishigami, W. Waskitoaji, M. Yoneda, K. Takada, T. Hyakutake, T. Suga, M. Uchida, Y. Nagumo, J. Inukai, H. Nishide, M. Watanabe, J. Power Sources 269 (2014) 556-564