Experimental Results with Fuel Cell Start-up and Shut-down. Impact of Type of Carbon for Cathode Catalyst Support

Wednesday, October 14, 2015: 14:00
211-B (Phoenix Convention Center)
O. Lottin, J. Dillet, G. Maranzana (LEMTA, Université de Lorraine, Vandoeuvre-lès-Nancy), S. Abbou (LEMTA, Université de Lorraine, Vandoeuvre-lès-Nancy), S. Didierjean (LEMTA, CNRS, Vandoeuvre-lès-Nancy), A. Lamibrac (Paul Scherrer Institut), R. L. Borup, R. Mukundan (Los Alamos National Laboratory), and D. Spernjak (Los Alamos National Laboratory)
The main impediment to the wide-range spread of proton exchange membrane fuel cells is most probably their low durability – at a reasonable production cost: highly dispersed and carbon supported catalysts developed to lower the cost of PEM systems suffer from a lack of stability due to carbon corrosion and catalyst dissolution. Several specific working conditions have been identified as responsible for accelerated catalyst degradation [1, 2]. Among them, the harshest one may be the fuel cell startup or shutdown -without any particular mitigation. During fuel cell startup and shutdown, a part of the anode compartment is filled with hydrogen while the complementary part remains occupied with oxygen-rich gases. In this case, the electric potential of the cathode facing the oxygen-rich portion can reach values as high as 1.6 V [3, 4] which entails accelerated carbon corrosion and catalyst degradation in the local regions exposed to air in the anode compartment the longest: degradations will be more severe near the anode outlet (inlet) in the case of startup (shutdown) [5] and heterogeneities were also observed recently between the channel itself and the rib [6, 7]. This corrosion phenomenon is now relatively well characterized thanks to segmented cells [8].
In the experimental part of this work, separate testing protocols for fuel cell startup and shutdown were developed to distinguish between their effects on performance degradation. The internal currents and the local potentials (Figure 1) were measured with different membrane-electrode assemblies (MEAs): we examined the influence of the cathode and anode Pt loading, the type of carbon for cathode catalyst support and monitored the time evolution of spatially-resolved performance decrease and electrochemical active surface area (ECSA). Both the CO2 emissions and the charge exchanged –between the passive and the active regions of the cell- increased with the common residence time of air and hydrogen in the anode compartment. However, the evolved CO2 accounted for less than 25 % of the total exchanged charge indicating the predominance of other reactions: water electrolysis, Pt oxidation... Startups were also consistently more damaging than the shutdowns, evidenced by more evolved CO2, severe ECSA decrease, and higher performance losses.
The objectives of the modelling part of the work were to quantify mathematically the redox reactions occurring during startups and shutdowns in order to understand in detail the influence of the experimental parameters varied above. The numerical approach is based on a model that takes account variations in gas concentration and platinum oxide coverage between the cell inlet and outlet. Mass transport in the direction perpendicular to the membrane and electrochemical phenomena are modeled locally (along parallel hydrogen and air channels) while the concentration of gases in the channels are imposed as boundary conditions, as functions of space and time, so that this model can be considered as "pseudo2D" [9].

[1] R. Borup et al., Chem. Rev. 2007, 107, 3904-3951.
[2] L. Dubau et al., Wiley interdisciplinary reviews-energy and environment, Vol 3, Issue 6, pp 540-560, 2014.
[3] Q. Shen et al., J. Power Sources, 189, (2009).
[4] C.A. Reiser et al., Solid-State Lett., 8, A273 (2005).
[5] A. Lamibrac et al., J. Power Sources, 196, (2011).
[6] J. Durst et al., App. Catalysis B. : env. , 138-139, (2013).
[7] I.A. Schneider and S. von Dahlen, Electrochem. Solid-State Lett., 14, B30 (2011).
[8] J. Dillet et al., J. Power Sources 250C (2014).
[9] G Maranzana et al., J. Electrochem. Soc., 162 (7), F694-F706 (2015).