Comparison of MEA Degradation through FCV Actual Driving Test and Load-Cycle Durability Test

Abstract

Introduction In research and development of high-durability polymer electrolyte fuel cells (PEFCs) for fuel cell vehicles (FCVs), accelerated stress tests (ASTs) are employed to evaluate durability of materials with a short time [1-3]. In order to validate an AST, it is necessary to confirm that the degradation mechanisms in the AST and in actual FCV operation are consistent, and to elucidate quantitative correspondence of their degradation rates. In this study, we analyzed degradation of a membrane electrode assembly (MEA) after FCV real-world driving and compared the degradation factor with that of a load-cycle durability test in a laboratory. Experiments MEA samples were extracted from an FC stack of a Toyota MIRAI 2014 model which was operated in real-world driving of 200,000 km, about 6000 h. MEA performance and electrochemical properties of them were investigated using a small differential cell with a straight flow field. As a comparison, another MEA at beginning of life (BOL) was subjected to a load-cycle durability test following a protocol encouraged by New Energy and Industrial Technology Development Organization (NEDO), Japan. Cell temperature and relative humidity were controlled at 80°C and 100%, respectively. A square-wave potential cycle shown in Figure 1 was adopted. MEA performance was investigated after cycles of 500, 1000, 3000, 10,000, and 30,000. Structures of catalyst layers and catalyst particles were analyzed for the MEAs after driving test and 30,000 cycles of the load-cycle test using scanning electron microscope (SEM), transmission electron microscope (TEM), and X-ray absorption fine structure (XAFS). Results and discussion The performance of the MEA after 200,000 km driving is close to that of the MEA after 30,000 cycles of the load-cycle test as shown in Figure 2. The performance degradation is mainly attributed to degradation of the catalyst layer, and properties of a gas diffusion layer and a proton exchange membrane are unchanged from those of the BOL sample. Degradation mechanisms of the catalyst layers in the driving test and the load cycle test are well corresponding. First, mass activities of PtCo catalyst decay to about a half of the BOL value mainly due to the decrease in electrochemically effective surface areas (ECSAs) as represented in Figure 3. Decrease rates of the ECSAs are consistent with the increase in average particle sizes of the catalyst. The catalyst size distributions of the two samples are also comparable. Then, oxygen transport resistance in the catalyst layers is increased. The resistance of the samples after the driving test and 30,000 cycles of the load-cycle test are twice as large as the BOL value. As shown in Figure 4, the transport resistance obviously exhibits inverse proportion to the catalyst surface area. This relationship represents that the primary factor of the resistance is one at vicinity of catalyst particles including ionomer permeation [4], and increase in it is attributed to decrease in the catalyst surface area. A performance simulation based on a simple 1-dimensioal MEA model supports that the decrease in the ECSA causes most of the MEA performance degradation through the deterioration in the catalyst activity and the oxygen transport resistance. Conclusions The MEA performance degradation after the 200,000 km real-world driving with the system of MIRAI 2014 model is primarily caused by catalyst metal coarsening and is well simulated by NEDO load-cycle durability test of 30,000 cycles. This result has validated the test protocol and should contribute to a target-setting of durability of future MEAs.