燃料电池阳极催化层缺陷在循环工况过程中的扩展Anode defects’ propagation in polymer electrolyte membrane fuel cells

Salah Touhami
Marie Crouillere
Julia Mainka
Jérôme Dillet
Christine Nayoze－Coynel
Corine Bas
Laetitia Dubau
Assma El Kaddouri
Florence Dubelley
Fabrice Micoud
Marian Chatenet
Yann Bultel
Olivier Lottin

Abstract

Defects－propagation in polymer electrolyte membrane fuel cells membrane electrode assemblies （MEA） is investigated via Accelerated Stress Tests （AST） combining load （hence potential） and load－driven humidity cycling， and open－circuit voltage． Customized MEA with lack of anode catalyst layer at two different locations －near the hydrogen inlet or outlet－are fabricated and subjected to the AST． Periodical electrochemical characterizations are performed using a segmented cell， enabling to track the cell performance and anode／cathode electrochemical surface area （ECSA） over the test period with a spatial resolution along the gas channels． These observations are completed by post mortem analyses of the MEA．

The MEA accelerated degradation is obvious， with multiple impacts on the cell performance and materials． More specifically， the results brought first evidence of defects propagation， in term of anode ECSA loss， in the direction of the hydrogen flow． The cathode ECSA is also impacted， although seemingly homogeneously． Significant membrane thinning is observed for the defective segments， without propagation to the adjacent ones． Anode and cathode local potential monitoring during the AST reveals the absence of cathode high－potential excursion， in both the segments with／without initial defects： the membrane and anode accelerated degradation is governed by chemical mechanisms like gas crossover rather than electrochemical mechanisms induced by high－potential excursions．

Fig． 1． Pictures and operation principle of the linear segmented cell． Up： segmented cathode flow－field with the auxiliary hydrogen channel feeding the RHE used to measure the local potentials． Middle： anode flow－field plate． Bottom： cross－sectional view of the cell． The reference electrodes give access to the local electrolyte potential Ve（i） and thus to the anode and cathode potentials of each segment

Fig． 2． Up： anode －before transfer on the membrane－with －or without－a lack of CL at one end． Below： longitudinal sectional view of the segmented cell． The location of the defect is situated either at the anode inlet （in green） or outlet （in red）．

在感兴趣的位置刻意制造的催化层缺陷

The distributor plate on the cathode side is made up of 20 segments （1．5 cm×1 cm each） electrically insulated along the channel length， which allows individual current collection on each of them．

分区电流测试区块大小1．5 cm×1 cm，缺陷大小1．5 cm×2 cm

defects have an area of 3 cm2， corresponding to the complete area of segment ＃3 and half of area of segments ＃2 and ＃4 （when close to the hydrogen outlet） or to the whole area of segment ＃18 and half of that of segments ＃17 and ＃19 （when close to the hydrogen inlet）．

Fig． 3． Current density and voltage profiles during the RH and load cycling AST． The hydration stage consists of alternating 1 s at low current （0．25 A／cm2） and 3 s at high current （1．3 A／cm2） sequences and lasts 52 s． The dehydration stage consists of alternated sequences of 3 s at low current and 1 s at high current， followed by 27
s at OCV （total duration ＝ 105 s）

Fig． 4． Variation of the FC voltage at 0．5 A／cm2 （top， left）， of the OCV （top， right） and of the hydrogen permeation current （bottom） during AST performed with a reference MEA （without defect）， a MEA with a lack of anode CL close to the hydrogen inlet， and a MEA with a lack of anode CL close to the hydrogen outlet． All data were measured during the characterization stage performed every 24 h． The voltage degradation rates were estimated using a linear interpolation of the dots．

看完后面的图8，图4这种图已经没有意义了，也无法做任何推测了。

Fig． 5． Variation of the local ECSA during the AST at the cathode （a， c and e） and the anode （b， d and f） in the case of the MEA without defect （a and b）， the MEA with a lack of CL close to the anode outlet （c and d） and the MEA with a lack of CL close to the anode inlet （e and f）， respectively． The segments corresponding to the anode defects are marked in red （d and f）． The segments marked in grey are those where the cathode ECSA measurements cannot be considered as reliable due to the absence of a counter electrode．

图5e的1、2不明情况得下降、图f的1不明情况得下降。

图5df的ECSA测量BOL状态与图5b的结果差异很大（指测量结果的分散状态）。

不是很理解无缺陷样品氢侧ECSA的增加是怎么回事。

图5的ef的1区ECSA都为0，在图8区域还能测试出性能也是挺奇怪的现象。

作者文中并没有做解释。

文章标题的缺陷扩展指的是图5d的1区，图5f的10－16

Table 1 Beginning of Life （BoL） and Enf of Test （EoT） average anode and cathode ECSA of the MEA without defect， the MEA with a lack of CL close to the anode outlet， and the MEA with a lack of CL close to the anode inlet．

Fig． 6． Overall thickness of the membrane at different segments （＃18， ＃16， ＃10 or ＃11， ＃6 and ＃3） after 240 h of AST． Homogeneous MEA without defect （segments ＃3， ＃4 and ＃18）， MEA with lack of CL at the anode outlet （segments ＃1， ＃3， ＃5， ＃6， ＃11， ＃18） and at anode inlet （segments ＃3， ＃6， ＃12， ＃14， ＃16， ＃18 and ＃20）． Ten measurements were made in the segments with a defect， i．e． five in the channel region and five below the lands． Only three measurements were made
in the other segments because it quickly appeared that the membrane thickness was unaffected． Data plotted on the graphs correspond to the average values and the error bars stand for the standard deviation．

和图4总体的渗氢电流对比可以发现，阳极催化层存在缺陷时膜容易减薄，但整体的渗氢电流并未显示警示信号。

The MEA were made with reinforced Gore 735．18 membranes．缺陷位置出现减薄，减薄到初始值的72％。

Fig． 7． EoT membrane thickness compared to BoL （100％）， below the channels and below the lands， in the segments with a lack of anode CL （i．e． ＃3 in the case of the MEA with a lack of anode CL near the hydrogen outlet， and ＃18 in the case of the MEA with a lack of anode CL near the hydrogen inlet）． Data plotted on the graphs correspond to the average values of five measurements， with standard deviation．

膜的减薄主要发生在流道内部，而且当催化层缺陷出现在阳极出口时，流道内部和脊部均出现减薄。

Fig． 8． Local polarization curves （left）， cathode local potentials （middle） and anode local potentials （right） as functions of the current density in the case of the MEA with a lack of anode CL near the hydrogen inlet （above） and near the hydrogen outlet （below）． All data are plotted in black， except those measured in the segments with defect． The defect was either centered on segment ＃18 or on segment ＃3 and spread to half the area of the adjacent segments． The dots appear horizontally
aligned on the polarization curves because all segments are electrically connected in parallel once the local currents －which vary from one segment to the other are measured．

Conclusion

The impact and propagation of anode defects taking the form of a
lack of CL was investigated using an AST combining potential and humidity cycles， and OCV holds． Customized MEA were intentionally
prepared with anode defects close to the hydrogen inlet or outlet． The
measurements were performed using a segmented and instrumented
cell， making it possible to follow how the local performances varied and to track the anode and cathode local ECSA． The results were compared to those of a reference MEA， without defects．

The results clearly showed an accelerated degradation of the MEA when there is a lack of anode CL in some segments， with multiple impacts on FC performance， and electrode CL and membrane degradations．The results also suggest that in term of anode ECSA， the anode defect may propagate in the direction of hydrogen flow． The cathode ECSA was also impacted， although seemingly homogeneously． A significant membrane thinning was observed in the defective segments， without significant propagation to the adjacent ones． However， one cannot exclude， that propagation would happen for longer AST， especially considering the degradation of the anode CL in the direction of the hydrogen flow． This point will be the subject of future work．

作者也指出时间效应问题。

The monitoring of anode and cathode local potential during the AST
did not reveal any excursion of the cathode potential to abnormally high values， neither in the defective segments nor in the others． This allowed to propose some initial interpretations of the results， which will have to be confirmed by additional post－mortem analyses： in particular， the membrane and anode CL accelerated degradation is more likely governed by chemical mechanisms （i．e． gas crossover） than by electrochemical mechanisms （i．e． high potentials）．

文章的结论和四类催化层缺陷对燃料电池局部膜降解失效的影响中相关部分结论不同，不清楚相关原因，可能影响这件事情和工况强相关。既然不同应该做个阴极侧缺陷的实验。

我暂无权限下载支撑材料。