Voltage Cycling Degradation Dependence on O2 Pressure: A Comparative Voltage-Loss Analysis

Abstract

Polymer-electrolyte-membrane fuel cells (PEMFCs) have proven their applicability as a clean energy conversion technology. At the same time, a clear understanding of their degradation under dynamic operation is critical to project and ultimately improve long-term durability. In order to mimic real-life operation stressors, accelerated stress tests (ASTs) are used to enhance a particular type of degradation mechanism. After the AST application, various physico- and electrochemical diagnostic tools aid to understand and catalogue the observed performance losses. Since the ultimate goal is to develop in situ mitigation and recovery strategies for automotive applications, a comprehensive voltage loss study describing the degradation effects on performance losses is desired. Particularly inside ASTs, voltage cycling tests reproduce conditions that mimic the dynamic operation of a fuel cell stack, and which are directly linked to the degradation of catalyst layers1 in a membrane electrode assembly (MEA). In a semi-empirical study, voltage cycling tests were applied by Kneer et al. 2 to promote electrochemical surface area (ECSA) losses of a carbon-supported platinum catalyst (Pt/C) for the oxygen reduction reaction (-ORR), establishing an ECSA loss semiempirical model for a wide range of voltage cycling conditions. Subsequently, this model was used to project performance losses in H2/air curves based on a theoretical relationship between ECSA evolution and the ORR observed kinetic and transport losses. However, the authors observed that their model prediction resulted in a substantial unaccounted voltage loss (compared to experimental data), suggesting that deeper analyses are required to understand voltage cycling induced performance losses, with the ultimate goal to develop in situ mitigation and recovery strategies. Therefore, our study will make use of voltage cycling ASTs, which will correlate voltage losses to degradation phenomena measured by various diagnostic tests3. These include the ORR mass activity (via H2/O2 measurements), proton transport resistance through the membrane and catalyst layer (via impedance analyses), and O2 transport resistance (via limiting current measurements). By performing a standardized protocol, the differences in MEA aging during voltage cycling tests conducted with H2/air vs. the more frequently used H2/N2 (anode and cathode, respectively) will be separately studied. Furthermore, as observed in Figure 1a, lowering the relative humidity (RH) of the reactants significantly reduces ECSA losses during voltage cycling tests, in comparison to the same protocol performed at 100% RH. More importantly, the H2/air performance losses are substantially lowered after low RH testing (in comparison to 100% RH) (Figure 1b). The unaccounted voltage losses vs. specific current density (ispec ) (the difference between the theoretical 70 mV/dec. line (black line in Figure 1c) and the corrected experimental cell voltages) are lowered, as well. Therefore, the effect of further operational stressors (e.g. cell temperature) on voltage cycling aging for H2/N2 vs. H2/air will be presented. Ultimately, we will seek to correlate the observed losses to those under automotive operating conditions. This insight leads to a better understanding of catalyst degradation and its subsequent suppression via appropriate mitigation strategies through the correct application of operating conditions as described by single cell experiments. Figure 1: a) ECSA losses vs. number of cycles. The insert depicts the used voltage profile (1 cycle equals to 8s at the lower voltage limit (LVL) followed by 8s at the higher voltage limit (HVL), voltage ramp >4V/s). Voltage cycling tests were performed as described by Harzer et. al. in a single-cell PEMFC with 5 cm2 active area (loading: 0.1 and 0.4 mgPt cm-2 in the anode and cathode, respectively; 100 kPaabs, 80°C and H2/N2 flows of 200 nccm/75 nccm on anode/cathode, respectively)3. b) H2/air performance curves (conditions: differential flow, 80°C, 170 kPaabs inlet-controlled pressure, anode: 2000 nccm of H2, cathode: 5000 nccm of air). c) Tafel representation of I-V curves after voltage cycling AST at 30% RH (20k cycles) and 100% RH (10k cycles). Ecell was corrected for HFR, H+ resistance through the catalyst layer, and O2 mass transport resistance; specific current density (ispec ) is corrected for cell shorting resistance and H2 crossover current. References Kneer, A.; Wagner, N.; Sadeler, C.; Scherzer, A. C.; Gerteisen, D., Effect of Dwell Time and Scan Rate during Voltage Cycling on Catalyst Degradation in PEM Fuel Cells. J Electrochem Soc 2018, 165 (10), F805-F812. Kneer, A.; Wagner, N., A Semi-Empirical Catalyst Degradation Model Based on Voltage Cycling under Automotive Operating Conditions in PEM Fuel Cells. J Electrochem Soc 2019, 166 (2), F120-F127. Harzer, G. S.; Schwämmlein, J. N.; Damjanovic, A. M.; Ghosh, S.; Gasteiger, H. A., Cathode Loading Impact on Voltage Cycling Induced PEMFC Degradation: A Voltage Loss Analysis. J Electrochem Soc 2018, 165 (6), F3118-F3131. Figure 1