COCV加速测试工况下的燃料电池膜降解：膜均匀减薄叠加形成针孔

Membrane degradation during combined chemical and mechanical accelerated stress testing of polymer electrolyte fuel cells

C． Lim

L． Ghassemzadeh

F． Van Hove

M． Lauritzen

J． Kolodziej

G． G． Wang

S． Holdcroft

E． Kjeang

Abstract

A cyclic open circuit voltage （COCV） accelerated stress test （AST） is designed to screen the simultaneous effect of chemical and mechanical membrane degradation in polymer electrolyte fuel cells． The AST consists of a steady state OCV phase to accelerate chemical degradation and periodic wet／dry cycles to provide mechanical degradation． The membrane degradation process induced by COCV AST operation is analyzed using a standard MEA with PFSA ionomer membrane． The OCV shows an initially mild decay rate followed by a higher decay rate in the later stages of the experiment． Membrane failure， defined by a threshold convective hydrogen leak rate， is obtained after 160 h of operation． Uniform membrane thinning is observed with pinhole formation being the primary cause of failure． Mechanical tensile tests reveal that the membrane becomes stiffer and more brittle during AST operation， which contributes to mechanical failure upon cyclic humidity induced stress． Solid state 19F NMR spectroscopy and fluoride emission measurements demonstrate fluorine loss from both side chain and main chain upon membrane exposure to high temperature and low humidity OCV condition．

Fig． 1． Dynamics of the open circuit voltage and high frequency impedance （per cell） recorded during the 4th COCV AST cycle， covering a selected portion of the OCV and RH cycling phases： （a） open circuit voltage； and （b） high frequency impedance．
每一个循环中，前段氢空OCV，后段氢氮干湿循环

即使切换到N2，电压的下降也不是立刻产生的。

Fig． 2． Measured open circuit voltage， （a） high frequency impedance， and （b） stack coolant temperature differential （difference between outlet and inlet temperature） tracked during a complete AST experiment． Impedance logging started at the 4th cycle．

图2的HFR和图1的HFR各表示了一部分HFR。图1最高HFR25Ohmcm2／cell，而图2最低HFR0．2Ohmcm2／cell。

膜的串漏造成冷却水温度上升。

Fig． 3． Variation of fluoride emission rate （FER） measured from the anode and cathode effluents during a complete COCV AST experiment

FER并不是随着循环逐渐升高。在第7个循环100hr达到最高。为什么表现出最高点文章没有解释。似乎第7个循环出现针孔后失效的模式发生了显著变化。

Fig． 4． Changes in average cell OCV and concurrent electrochemical leak detection test （ELDT） results with COCV AST cycles．

Fig． 5． Fuel cell polarization curves measured in hydrogen／air at atmospheric pressure， 75 C， and 100％ RH after different numbers of COCV AST cycles．

Fig． 6． Cumulative fluoride loss as a function of the COCV AST cycles．

Fig． 7． Chemical structure of PFSA ionomer．

Fig． 8． Solid state 19F NMR of PFSA membrane at 30 kHz and the corresponding peak assignments． Peak assignments are based on former solid－state and liquid state NMR investigations．

Fig． 9． Integrals of the 19F NMR peaks of polymer side chain before and after COCV AST tests．

Fig． 10． SEM micrographs of （a） cross－sectional MEA at the beginning of life （BOL） stage， （b） cross－sectional MEA at the end of life （EOL） stage， and （c） pinhole on the surface of the membrane at EOL．

Fig． 11． Normalized membrane thickness as a function of COCV AST cycles obtained from SEM cross－sectional image analysis of MEA samples extracted at different AST cycles．

稳态极化曲线第11个循环出现OCV下降，此时平均膜厚下降到初始的60％。

没有说用的是什么型号的膜。

Fig． 12． Stressestrain curves obtained at 23 C and 38％ RH from membrane samples extracted at different numbers of COCV AST cycles： （a） BOL； （b） 1st cycle； （c） 2nd cycle；（d） 5th cycle； （e） 8th cycle； and （f） 11th cycle． The stress was calculated based on the thickness of the BOL membrane excluding catalyst layer．

Fig． 13． Changes in （a） elastic modulus， （b） ultimate tensile stress， and （c） fracture strain of the membrane as a function of COCV AST cycles． The stress was calculated based on the SEM measured membrane thickness after each cycle （Fig． 11）．

Conclusions

In－situ degradation of PFSA membranes was evaluated using a
cyclic open circuit voltage （COCV） accelerated stress test （AST）
protocol consisting of a combined chemical OCV phase and a mechanical RH cycling phase． The cell OCV showed a mild decay rate of
0．7 mV／h in the early stage of membrane degradation during the
first half of the experiment， followed by an increased decay rate up
to 3．9 mV／h in the later degradation stage leading to membrane
failure． The stack coolant outlet temperature and the OCV drop
upon a small hydrogen overpressure differential， proportional to
the amount of hydrogen crossover from anode to cathode， were
found to increase rapidly in the later stage of the process towards
failure． Fluoride emission rate （FER） was measured in－situ from
anode and cathode effluents during COCV AST． The average FER was
progressively increased as a function of OCV operation time from
0．36 to 0．85 mmol／h／cm2 up to the 7th OCV phase． Analysis of
solid state 19F NMR spectra on partially degraded membrane
samples showed that the total amount of fluoride in the side chain
decreased by 23％ while the SCF2 group in side chain and CF in main chain decreased by 36％ and 23％， respectively， representing the occurrence of fluorine losses in both side chain and main chain
regions． This result was supported by a cumulative 30％ reduction in
membrane fluorine content and a 27％ reduction in membrane
thickness at the midpoint of the AST experiment．

指第7个循环

SEM morphological analysis demonstrated that the membrane experienced uniform thinning as a function of COCV AST cycles and eventually developed pinholes， which were the main source of hydrogen
crossover leaks and ultimate MEA failure． From tensile tests on
degraded membranes， the fracture strain was found to decrease
rapidly during the COCV AST operation while the elastic modulus
increased， indicating that the membrane developed a stiffer， more
brittle structure expected to be vulnerable to initiation of pinholes．
In summary， while the combined chemical and mechanical COCV
AST procedure was deliberately designed to generate rapid membrane
failures， the present results demonstrated reliable signs of
both chemical and mechanical membrane degradation mechanisms，
and perhaps more importantly， revealed that the simultaneous
presence of both mechanisms significantly accelerated the
overall rate of degradation．