Diagnosis of hydrogen crossover and emission in proton exchange membrane fuel cells G. Mousa a , J. DeVaal b , F. Golnaraghi a,* a School of Mechatronic Systems Engineering, Simon Fraser University, Surrey, BC V3T 0A3, Canada b Ballard Power Systems, 9000 Glenlyon Parkway, Burnaby, BC V5J 5J9, Canada article info Article history: Received 12 September 2014 Accepted 22 September 2014 Available online 25 October 2014 Keywords: EIS PEM fuel cells Hydrogen leak Hydrogen pumping Air starvation Fuzzy logic abstract When hydrogen leaks through holes or cracks in membrane-electrode assemblies (MEAs) in Proton Exchange Membrane (PEM) fuel cells, it recombines directly with air. This recombination results in a reduction in oxygen concentration on the cathode side of the MEA. In this paper, the signatures of electrochemical impedance spectroscopy (EIS) are analyzed in different multi-cell stack configurations to show the relation between hydrogen leak rate and reduced oxygen concentrations. The reduction in concentration was made by mixing oxygen with nitrogen at different rates, and the increase in hydrogen leak rate was made by controlling the differential pressure (dP) between anode and cath- ode. To analyze the impedance signatures, we fit the data of oxygen concentration and dP with the parameters of a Randles circuit. The correlation between the parameters of the two data sets allows us to understand the change in impedance signatures with respect to reduction of oxygen in the cathode side. To have a better insight on the effect of insuffi- cient oxygen at the cathode, a model that establishes a relationship between impedance and voltage was considered. Using this model along with the impedance signatures we were able to detect the reduction of oxygen concentrations at the cathode with the help of fuzzy rule-base. However, resolution of detection was reduced with the reduction of leak rate and/or increases in the stack cell count. Copyright © 2014, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved. Introduction When the fuel cell starves of hydrogen, its anode and cathode potentials rise and drop respectively. In such a fault, the voltage of a single starved cell is dropping to zero [1]. If a hydrogen starving cell was placed in a stack, its voltage drops to negative values due to the high anode potential [2]. When the cell is air starved however, its voltage drops to almost zero with its anode potential little higher than the cathode [3]. Zhang et al. [1] studied the effect of fuel and air starvation on the voltage along a single PEM fuel cell. Reducing the air stoich to 1.5, the cell voltage drops due to the lack of oxygen con- centration. As concentration drops along the flow channel, lower local current density was observed at the downstream of the cell. Running the cell at lower hydrogen stoich, the current density dropped sharply to zero at the downstream of the cell due to the insufficient hydrogen along the cell. In a similar study, Schneider et al. [4] considered the effect of air starvation on the impedance along a 63 cm 2 single PEM fuel cell. At normal air stoich, the impedance spectra had two frequency loops. They suggested that the presence of the * Corresponding author. E-mail addresses: gmousa@sfu.ca (G. Mousa), jake.devaal@ballard.com (J. DeVaal), mfgolnar@sfu.ca (F. Golnaraghi). Available online at www.sciencedirect.com ScienceDirect journal homepage: www.elsevier.com/locate/he international journal of hydrogen energy 39 (2014) 20116 e20126 http://dx.doi.org/10.1016/j.ijhydene.2014.09.116 0360-3199/Copyright © 2014, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.