Quality Control Diagnostics for Mass Produced Fuel Cell Gas Diffusion Electrodes
The National Renewable Energy Laboratory has developed a QC diagnostic method for gas diffusion electrodes that is applicable to roll-to-roll fuel cell materials during production. The experimental setup of this Infrared Reactive Impinged Flow (IR/RIF) method is shown in Figure 1. It provides three main functionalities: (i) gas supply and delivery, (ii) GDE material transport, and (iii) IR thermographic imaging. A non-flammable gas mixture that contains small amounts of hydrogen and oxygen (typically 2% and 1%, respectively) is impinged on a gas diffusion media in an ambient environment. The hydrogen reacts in an exothermic reaction at the platinum catalyst of the electrode with the oxygen. The resulting heat signature contains information about the homogeneity of the sample and is observed with an infrared sensing device.
Results will be presented that demonstrate the feasibility of the IR/RIF method using NREL’s industrial webline at 10 ft min-1and higher. Production defects were intentionally created in-house using various scratch and/or spray coating methods. Figure 2 shows a sample set consisting of 100% loading reduction defects placed at various locations. Defect sizes of this sample were 5 mm x 5 mm and 2 mm x 2 mm. Results will be discussed using false color temperature movies and images as well as raw and processed temperature data.
Figure 1: Schematic of Reactive Impinged Flow Method for detecting defects in the catalyst layer coating of a fuel cell gas diffusion electrode.
Figure 2: Example of GDE with catalyst loading defects: 100% loading reduction, 2 mm x 2 mm and 5 mm x 5 mm.
Figure 3: False color image of the sample’s heat signature with the sample moving at 10 ft min-1.
Figure 4: Temperature data along the dashed line in Figure 3 indicating the detection of two coating defects.
The authors would like to acknowledge funding from the U.S. DOE Fuel Cell Technology Office under contract number DE-AC36-08GO28308. The authors would further like to thank D. Bittinat and J. Porter of Colorado School of Mines for participation in the early development stages of this method.