化学－机械衰减条件下燃料电池膜电极活性区域电解质膜的原位失效分析4D imaging of chemo－mechanical membrane degradation in polymer electrolyte fuel cells － Part 2： Unraveling the coupled degradation mechanisms within the active area

Yixuan Chen
Yadvinder Singh
Dilip Ramani
Francesco P． Orfino
Monica Dutta
Erik Kjeang

Abstract

This work is a two－part article series on chemo－mechanical membrane degradation in fuel cells， wherein Part 1 investigated edge failure and the present work in Part 2 investigates failure within the active area of an edge－protected cell． X－ray computed tomography based 4D visualization with three spatial dimensions plus one time dimension is applied to capture the progress of membrane degradation under chemo－mechanical accelerated stress testing． Buckling driven membrane cracks are found to be the predominant failure mechanism， occurring exclusively under the uncompressed flow field region． In situ electrochemical diagnostics show significant gas crossover and dramatic open－circuit cell voltage loss accompanying membrane crack formation． The observed root cause of membrane buckling is facilitated by pre－existing cracks and surface pores in the catalyst and microporous layers， which enable hygral expansion and locally amplified compressive－tensile stress cycles in the membrane． Moreover， permanent membrane creep into gas diffusion layer pores is identified exclusively under the compressed land region， which is driven by locally elevated compressive stress in the membrane． Chemical stress is inferred to be a contributing， accelerating factor in this degradation process as a uniform， global effect rather than a locally amplified cause of failure．

相关文献

化学－物理衰减条件下燃料电池膜电极封装区域电解质膜的失效分析

原位可视精确跟踪燃料电池电介质膜机械衰减的变化

四类催化层缺陷对燃料电池局部膜降解失效的影响

X射线断层扫描技术用于燃料电池失效电极的三维微米纳米分析

同向供气循环开路加速测试下的燃料电池质子交换膜失效

这张图在化学－物理衰减条件下燃料电池膜电极封装区域电解质膜的失效分析的文章中使用过。

ECSA在循环工况中显著下降，HFR和双电层电容变化不大。

BTW：HFR出现那么大的一个异常点也不重复一下实验？

短路电阻变化不大There was no observable tilt in the CV curve throughout the entire MEA lifetime （not shown for brevity）， indicating that the shorting resistance of the MEA remained relatively large with negligible electrode shorting within the MEA．

这个极化曲线的失效状态和本文主要说的膜失效似乎不太对应。但在OCV工况中碳腐蚀和离聚物降解一定程度上也能解释该结果。

可以看出在该设计中脊和流道内展现出很明显不同的失效形态。

在流道区域内

简单说流道区域内膜的破损很厉害，催化剂平面内不均一性增加，催化剂和膜分离。

因为是原位实验，可以获得在EOL状态下干态和湿态不同状态下的图像，这在使用其他方法是无法获得的，

在脊区域内

在化学－机械衰减工况，膜的状态不是很友好。实验中the MEA was compressed between two flow field plates made of compressed carbon／graphite with cured resin。不知道使用其它类型的双极板，比如脊很光滑的金属双极板会是什么样的结果。

在设计中观察区都是有CCM的，不知道不同的设计有什么差异。

实验中使用的是non－reinforced Nafion® NR211 PFSA ionomer membrane，使用其它类型的膜的结果是如何的很令人感兴趣。

实验中使用化学－机械衰减工况，单独的OCV工况不知道是什么样的结果。单独的RH工况可以参见原位可视精确跟踪燃料电池电介质膜机械衰减的变化。

Conclusions

In this work， small－scale fuel cell testing and micro XCT visualization
was cooperatively used to perform 4D in situ membrane degradation
analysis of the mechanisms and root causes of combined chemical and
mechanical membrane degradation． The cell was edge－protected based on the recommendations of Part 1 of this work in order to facilitate indepth investigation of degradation within the active area． The test was composed of steady state OCV hold to generate chemical stress and RH cycling between wet and dry states to produce mechanical stress． Membrane cracks were the dominant failure mode in the channel region while membrane creep was the main deformation observed in the land region． According to in situ diagnostic results， most OCV losses occurred in the last six AST cycles （26th to 31st cycle） which was accompanied by and attributed to convective hydrogen crossover． Based on 4D image analysis， through－thickness membrane cracks were first observed at the 25th AST cycle and rapidly propagated thereafter． Therefore， it was confirmed that membrane cracks were the dominating degradation mode causing convective gas crossover． Most membrane cracks were driven by membrane buckling during RH cycling when membrane underwent
cyclic swelling and shrinking． The MEA frame design limited inplane
membrane expansion to a great extent so that through－plane
membrane expansion was dominant． Based on cross－sectional image
analysis， most membrane buckling was due to weak bonding between CCM and MPL， large pores in GDL， missing catalyst layer， and membrane deformation that created gaps so that the stresses acting on anode and cathode sides of the CCM were not balanced． As a result， buckling occurred and created bending stress inside the membrane． Through morphological analysis， all membrane cracks related to buckling were initiated from the catalyst layer surface under compression， and gradually propagated into the membrane． It was further discovered that cyclic membrane swelling and shrinking can cause catalyst layer material to detach from the membrane surface and accumulate into large，densified particles that are pushed into adjacent GDL pores during membrane expansion． Comparing to previous pure chemical and combined chemo－mechanical degradation studies， membrane thinning and exclusive membrane cracks were not observed in this work．

Moreover， branching during in－plane crack propagation that was
discovered in GDE－based MEAs subjected to COCV AST was not observed in the present work． Referring to these outcomes， chemical
stress was considered moderate in the current combined chemical and mechanical AST protocol． All membrane cracks seem to be mechanically initiated and only chemically enhanced even if the AST induced both chemical and mechanical stressors．

Instead of cracks， membrane creep was exclusively observed in the
land region where the plates constrained through－plane MEA expansion． For most creep spots， membrane filled the GDL macro pores beneath the MPL and formed through－plane protrusions． During the wet phase of RH cycling， stress concentration was present at the CCM－MPL interface，which appeared to cause MPL collapse into underlying GDL pores． Most creep spots were found at the anode side， where the CL was thinner than at the cathode． Catalyst layer material accumulation and densification was also observed at the creep spots． Thus， ECSA， gas diffusion， and water removal may be negatively affected by the action of creep protrusion．However， the impact of creep on hydrogen crossover was likely negligible compared to the impact of the membrane cracks in the channel region．

The application of 4D in situ visualization technique enabled the
acquisition of both dry and wet MEA images in the same FOV as well as the evolution of membrane damage， which provided a thorough understanding of the membrane degradation mechanisms and the interactions between the membrane and other MEA components in the
fuel cell environment． The capability of tracking failure evolution also
enabled the correlation between in situ electrochemical diagnostic results and MEA images at each corresponding life stage． For instance， it
was discovered that the membrane edge cracks were not critical to
performance loss while active area cracks were the dominating failure． Such conclusions could not have been drawn without the support of intermediate MEA images． The present 4D visualization technique
adopted in both Part 1 and Part 2 of this work is therefore recommended for future research on fuel cell degradation and durability in order to appropriately capture the internal processes within the native fuel cell environment．