Evaluation of the Galvanostatic Charge Technique for the in Situ Characterization of Aged Meas in a PEMFC

Abstract

Voltage-controlled in situ diagnostics like cyclic voltammetry (CV) are commonly used to characterize membrane electrode assemblies (MEAs) in single-cell proton exchange membrane fuel cells (PEMFCs) and to monitor MEA degradation over operating time. Their application, however, is more complex for characterizing individual cells in a PEMFC stack, depending on stack design and the capability of the fuel cell test station. Consequently, several studies have demonstrated a facile and easily adaptable alternative by using a galvanostatic (current-controlled) technique for diagnosing MEAs in a PEMFC stack, where all the cells are connected in series. It was demonstrated that a galvanostatic charge curve applied to a PEMFC stack permits the quantification of the electrochemical surface area (ECSA, m2/gPt) and the double layer capacitance (C dl, mFcm2 MEA ) of the cathode electrode as well as of the hydrogen crossover current (i H2,X, mA/cm2 MEA) and the electrical shorting resistance of the membrane (R short, mA/cm2 MEA) of each individual MEA in a stack.1, 3, 4 However, most of these studies only investigated healthy MEAs that represent their beginning-of-life (BoL) condition, and it is not clear whether the galvanostatic charge method can also yield quantitative measures for aged MEAs in a PEMFC stack, which may exhibit significant H2 permeation and shorting currents as well as very low electrode roughness factors (in units of cm2 Pt/cm2 MEA).In this study, we aim to investigate and understand the viability of the galvanostatic charge technique in the presence of substantial H2 permeation and shorting currents. For this, single cells with 50 cm2 MEAs exhibiting different characteristics are examined: 1) low i short, representing an MEA at BoL; 2) high i H2,X and i short, representing an MEA with a degraded membrane; and, 3) low electrode roughness factors (i.e., low ECSAs), representing, e.g., an MEA after a voltage-cycling AST (accelerated stress test). The MEAs were prepared by either ex situ or in situ methods. To quantify the ECSA, R short, and i H2,X, an anodic current (I) is supplied to the N2-filled cathode of the cell (using the H2-filled anode as counter and reference electrode), and the corresponding cell voltage profiles (V) within a defined time scale (t) are analyzed by equations 1 and 2:3 I⋅dt = Q Pt/H + C dl⋅dV + (i short + i H2,X)⋅dt equation 1 I = (Q Pt/H/dv+ C dl)⋅dV/dt + (i short + i H2,X) equation 2Here, Q Pt/H is the hydrogen desorption charge from Pt catalyst (i.e., from Pt-H → Pt + H++ e-) and C dl is the double-layer capacitance of the cathode electrode; further contributions are from i short and i H2,X. The results obtained from this galvanostatic analysis approach will be compared with those obtained from commonly used potentiodynamic methods (i.e., from CV and linear scan voltammetry (LSV)). An exemplary data set is shown in Figure 1. Additionally, the impact of operating conditions (e.g., cathode gas flow rate during the analysis) is systematically investigated. To conclude, we will propose an improved and general methodology for using a galvanostatic analysis approach for the diagnosis of MEAs in a PEMFC stack.Reference Y. Chatillon, C. Bonnet and F. Lapicque, Journal of Applied Electrochemistry, 43, 1017 (2013). K. Chen, Y. Hou, C. Jiang, X. Pan and D. Hao, International Journal of Hydrogen Energy, 46, 38469 (2021). P. Pei, H. Xu, X. Zeng, H. Zha and M. Song, Journal of Power Sources, 245, 175 (2014). I. Hartung, S. Kirsch, P. Zihrul, O. Müller and T. Von Unwerth, Journal of Power Sources, 307, 280 (2016). Figure 1