10kW金属双极板燃料电池电堆动态工况衰减分析［失效分析其八］

Performance degradation and processengineering of the 10 kW proton exchange membrane fuel cell stack

Tiankuo Chu

Ruofan Zhang

Yanbo Wang

Mingyang Ou

Meng Xie

Hangyu Shao

Daijun Yang

Bing Li

Pingwen Ming

Cunman Zhang

Abstract

Insufficient durability of proton exchangemembrane fuel cells （PEMFCs） remains one of the important factors hinderingtheir large－scale commercial applications． To investigate the degradationmechanism， we describe the durability test of 10－kW metal plate fuel cell stack containing 30 cells under dynamic driving cycles． After 600 h of testing， the mean voltage decay percentage of the stackunder the rated current densities of 1000 mA cm−2 is 2．67％． A semi－empirical modelis introduced to predict the remaining useful life of the stack， and the resultsatisfies the 5000 h target set by the department of energy （DOE）． Three cells with the highest， moderate， andlowest rate of decay are disassembled and characterized by electrochemicaland physical methods． Scanning electron microscopy （SEM） shows that thecross－section of the cathode catalystlayer （CL） of the 30＃ MEA has the lowest thickness of 8．45 μm compared with thefresh sample and other samples． Transmission electron microscopy （TEM）shows serious agglomeration of the 30＃catalyst． These observations led to serious performance degradation in the30＃ cell． The defects in the design of the stack structure leads to theattenuation of the consistency of the stack and further explains stackperformance degradation．

The 30－cell metal bipolar plate PEMFC  stack tested in this study consists of endplates， current collector plates，  seals， bipolar plates and MEA， which was supplied by YQ Power Co．， Ltd．

The catalyst， MEA， and metal bipolar  plate were all independently developed by Tongji University．

the active electrode area was 340 cm－2

Fig． 1． （a） PEMFC stack assembly， （b）  Physical graph of the PEMFC stack and test bench， （c） Dynamic driving cycles．

Table 1 Nominal operating conditions

Fig． 2． （a） The polarization curve for  durability test of the fuel cell stack， （b） The mean cell voltage decay at  different current densities， （c） The 13 ＃ ， 20 ＃ ， 30 ＃ and mean cell voltage  decay at 1000 mA cm－2 ． （d） The voltage of individual cells of the stack  before and after 600 h at 1000 mAcm－2

衰减并不多，衰减模式也是电堆两侧低，尤其是进出气口单节电压低。

Table 2 The decay percentages of cell  voltage under different current densities

Fig． 3． （a） Validation result with  prognostic model， （b） Prediction result of the mean cell voltage under 1000  mA cm－2

如果有能力预测电压的衰减，预测平均电压的衰减还不如预测最低电压的衰减。

Fig． 4． Electrochemical characterization  results for the 13 ＃ ， 20 ＃ ， and 30 ＃ and fresh MEA： （a） LSV curves， （b） CV  curves， （c） ECSA． EIS spectra at different current density： （d）300 mA cm－2 ，  （e） 1000 mA cm－2 ， and （f） 1500 mA cm－2

Fig． 5． The contact pressure distribution  over the （a） 1 ＃ ， （b） 15 ＃ ， and （c） 30 ＃ MEA．

Fig． 6． The contact angles of the （a）  fresh anode， （b） fresh cathode， （c） 30 ＃ anode， and （d） 30 ＃ cathode metal  bipolar plate．

Table 3 The contact angles of each  component inside MEA．

Fig． 7． SEM cross－sectional micrograph of  （a） fresh， （b） 13 ＃ ， （c） 20 ＃ ， and （d） 30 ＃ CCM．

Fig． 8． TEM micrograph of cathode  catalyst in the （a） fresh， （b） 13 ＃ ， （c）20 ＃ ， and （d） 30 ＃ MEA．

Conclusion

We report durability test of a 10－kW  metal plate fuel cell stack containing 30 cells under dynamic driving cycles．  The average voltage decay percentages of the stack under current densities of

300， 1000， and 1500 mA cm－2 are 0．78％，  2．67％， and 6．97％， respectively． The  standard deviation of individual cell voltage under 1000 mA cm－2 increased  from 7 to 17 mV． A semi－empirical model was introduced to evaluate the  remaining useful life of the stack． The stack does not reach the specified  end－point of life （mean cell voltage decline reaches 10％ under rated current  density） before 5000 h， which meets the DOE standards． The MEA  electrochemical results showed that the  cathode ECSA of the 30 ＃ MEA was the smallest， and the charge transfer resistance was the highest． The SEM images  showed that the cathode CL of the 30 ＃ MEA has the

lowest thickness of 8．45 m m compared  with the fresh and the other samples． The TEM images showed severe  agglomeration of the 30 ＃ cathode catalyst with a maximum average particle  size of 5．5 nm．

Combined with XRD and XPS characterization  results， the cathode catalyst degradation caused the ECSA drop and charge  transfer resistance increase of the 30 ＃ MEA， which is the main reason for  its worst voltage performance．

The massive difference in the percentage  of decay between every single cell is due to defects in the design of the  stack structure that attenuates the consistency of the stack． Stack  performance degradation can also be explained by this result． We note that  the 600 h of durability testing is not long enough， and the degradation mechanism  of some core components is still to be explored． The long－term durability  test of high－power fuel cell stacks requires more research in order to study  the relationship between the model based life prediction and the real  lifetime of the stack． The results present in this study adds useful  information to the field of PEMFC durability， which is of great importance  for specific applications of PEMFC．