Decay in Mechanical Properties ofCatalyst Coated Membranes Subjected toCombined Chemical and MechanicalMembrane DegradationA. Sadeghi Alavijeh1, M.-A. Goulet1, R. M. H. Khorasany1, J. Ghataurah1, C. Lim1,M. Lauritzen2, E. Kjeang1*, G. G. Wang1, R. K. N. D. Rajapakse1

1 School of Mechatronic Systems Engineering, Simon Fraser University, 250-13450 102 Avenue, Surrey, BC, V3T 0A3, Canada2 Ballard Power Systems, 9000 Glenlyon Parkway, Burnaby, BC, V5J 5J8, Canada

Received March 14, 2014; accepted November 20, 2014; published online November 28, 2014

Abstract

The mechanical stability of catalyst coated membranes(CCMs) is an important factor for the overall durability andlifetime of polymer electrolyte fuel cells. In this article, theevolution of the mechanical properties of degraded CCMs iscomprehensively assessed. A combined chemical andmechanical accelerated stress test (AST) was applied to simu-late field operation and rapidly generate partially degradedCCM samples for tensile and expansion experiments underboth room and fuel cell conditions. The tensile results indi-cated significant reductions in ultimate tensile strength,toughness, and fracture strain as a function of AST cycles,accompanied by a mild increase in elastic modulus. The

increased brittleness and reduced fracture toughness of theCCM, caused primarily by chemical membrane degradation,is expected to play an important role in the ultimate failure ofthe fuel cell. The expansion tests revealed a linear decay inhygrothermal expansion, similar in magnitude to the loss ofmechanical strength. The decline in CCM sensitivity to envir-onmental changes leads to non-uniform swelling and con-traction that may exacerbate local degradation. Interestingly,the hygrothermal expansion in the late stages of degradationcoincided with the fracture strain, which correlates to in situdevelopment of fractures in chemically weakened membranes.

Keywords: Accelerated Stress Test, Degradation, Durability,Fuel Cell, Mechanical Properties, Membrane

1 Introduction

One of the key challenges facing the commercialization ofpolymer electrolyte fuel cell (PEFC) technology is the develop-ment of membrane electrode assemblies (MEAs) that can meetindustry durability targets [1]. PEFCs are currently the princi-pal fuel cell type for automotive applications because of theirhigh performance at low temperature [1]. Perfluorosulfonicacid (PFSA) ionomer membranes (e.g., Nafion� from DuPont)are commonly used in PEFCs due to high proton conductivityand chemical stability inherited from the polytetrafluoroethy-lene (PTFE) base structure [1, 2]. However, in order to addressindustrial requirements and commercially compete withhybrid engines, the durability and lifetime of PFSA mem-branes must be enhanced.

Chemical and mechanical degradation are the primary rea-sons for membrane failure in fuel cells. Each of these degrada-tion mechanisms can be intensified by the presence of theother [1]. Chemical degradation is caused by the attack of radi-

cal species such as hydroxyl radical (.OH) and hydroperoxylradical (.OOH) originating from hydrogen peroxide formedduring fuel cell operation where the membrane is subjected tosevere oxidizing and reducing environments on the cathodeand anode sides. High temperature, low relative humidity,and high reactant gas pressures accelerate the chemical degra-dation [1, 3–6]. Open circuit voltage (OCV) conditions arewidely used in accelerated stress testing to intensify chemicalstressors caused by the reactant crossover mechanism. UnderOCV conditions, the reactant gases are at the highest level ofpartial pressures, which enhances reactant crossover andaccelerates membrane deterioration [7, 8]. Several researchgroups have studied different aspects of chemical membranedegradation and suggested mitigation strategies [9–14]. Forinstance, Yuan et al. [9] studied the performance of Nafion�

–[*] Corresponding author, ekjeang@sfu.ca

204 ª 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim FUEL CELLS 15, 2015, No. 1, 204–213

ORIG

INALRES

EARCH

PAPER

DOI: 10.1002/fuce.201400040

Sadeghi Alavijeh et al.: Decay in Mechanical Properties of Catalyst Coated Membranes

membranes with different thicknesses under idling conditionfor 1000 hours at constant current of 0.5 A. Results of variouselectrochemical tests on degraded samples showed that thinmembranes lead to more rapid OCV loss due to higher gascrossover [9]. Trogadas et al. [11] demonstrated mitigation ofchemical membrane degradation under OCV conditions usingcerium oxide nanoparticles embedded in the membrane. Pear-man et al. [12] further investigated this concept and showeddramatically reduced OCV decay and fluoride emission rate[12]. In our group, mitigation of membrane degradation by aplatinum band in the membrane was investigated by Macau-ley et al. [14]. According to electrochemical leak test and fluo-ride emission rate results, the platinum band reduced thechemical degradation significantly and enhanced the mem-brane durability [14].

Mechanical membrane degradation, on the other hand, ori-ginates from hygrothermal cycles during fuel cell operation.Variations in temperature and relative humidity result in sig-nificant dimensional changes due to water sorption in PFSAmembranes [15–17]. Inside a fuel cell, membrane expansionand contraction due to water sorption is constrained by otherfuel cell components, which may exacerbate mechanical deg-radation. Hydration and dehydration accompanied bychanges in temperature during start up and shut down andcyclic operating conditions induce mechanical stresses such ashygrothermal fatigue, creep, and tension in the membrane.

Recent efforts have been directed to reduce the membranethickness in order to minimize ohmic losses and improve fuelcell performance. Therefore, membrane mechanical durabilityand integrity has become a critical challenge. Mechanicalproperties of PFSA membranes [18–22] and CCMs [23–25]were comprehensively investigated in the literature. Underdifferent temperature and relative humidity conditions, Tanget al. [18] studied the tensile properties of Nafion� 112 PFSAmembrane and expressed that yield stress and Young’s modu-lus decreased as temperature and humidity increased. Tanget al. [19] investigated the degradation in mechanical proper-ties of Nafion� 111 samples and determined the safety limitfor fatigue strength to be 10% of the tensile strength. Using adynamic mechanical analyzer (DMA), Bauer et al. [20] studiedthe effect of temperature and relative humidity on Nafion�

117 elastic modulus. They discovered that the maximum elas-tic modulus with respect to humidity is temperature depen-dent and therefore concluded that water mainly acts as amembrane plasticizer and stiffener at low and high tempera-tures, respectively [20].

Further research revealed that the mechanical properties ofthe membrane are altered by the presence of catalyst layerscoated on the membrane. Through experiments and modeling,Huang et al. [23] studied the tensile properties of an MEA as afunction of relative humidity and reported that reduction inductility combined with constrained drying strain led to me-chanical failure [23]. Mechanical durability of MEA sampleswere studied by Aindow et al. [24] via fatigue lifetime experi-ments to evaluate the remaining lifetime of as-received anddegraded MEAs. The mechanical properties of CCMs and

PFSA membranes were systematically compared in a previousstudy by our research group [25]. The membranes showedhigher yield stress, yield strain, and elastic modulus whencompared with the equivalent CCMs at all environmental con-ditions in the tested range. For both materials, the yield stressand elastic modulus decreased when temperature and/or rel-ative humidity increased. Moreover, the swelling and contrac-tion of the membrane with respect to relative humidity werenearly two times greater than for the CCM. It was thus con-cluded that the catalyst layers reinforce the membrane [25]and that the mechanical properties of the composite CCMshould generally be measured and applied for in situ model-ing, as recently demonstrated by Khorasany et al. [26].

Accelerated stress testing (AST) is a powerful method tobenchmark the durability of fuel cells and determine mitiga-tion approaches at a reasonable time and cost [3, 21, 27–29]. Inthese tests, certain stressors known to cause degradation areelevated beyond regular duty cycle levels [3]. These stressorsmay include operating temperature and humidity, cell voltage,and current density [30]. The US Department of Energy’shydrogen program established AST protocols for membranedegradation in PEFCs [27]. An open circuit voltage (OCV)AST protocol was developed to intensify the chemical degra-dation mechanisms, while an RH cycling AST protocol wasdesigned to accelerate the mechanical degradation mecha-nisms [5]. In the context of mechanical membrane properties,Patil et al. [31] observed a significant reduction in the modulusand failure strain of Nafion� membranes subjected to chemicaldegradation under the OCV AST. Similar results wereobtained by Huang et al. [23], who compared the failuremodes of the chemical and mechanical AST protocols. Micro-structural examination of the membrane fracture surfacerevealed that crack initiation sites formed close to the anodeafter mechanical AST while chemical AST led to a clean andbrittle fracture surface [23]. In another study, Patil et al. [32]reported improvements in toughness and modulus accompa-nied with less creep deformation of Titania-reinforced Nafion�

membranes compared to Nafion� 112 prior and after subject-ing to OCV test. Furthermore, the reinforced membraneshowed enhanced resistance against dimensional changescaused by contractile stresses due to variation in humidity[32]. In our group, combined chemical and mechanical degra-dation mechanisms and failure modes were recently investi-gated via introduction of Ballard Power Systems’ cyclic opencircuit voltage (COCV) AST protocol [33]. The joint action ofchemical and mechanical stressors was found to intensify themembrane degradation and lead to more rapid failures thanthe independent mechanisms. Furthermore, results of fluorideemission, cross-sectional microscopy, and solid state 19F NMRanalysis revealed a combination of side chain and main chaindegradation in the polymer. The failure modes were consistentwith those observed in field operation, and comprised ofmembrane thinning, pinholes, and microcracks. Room-tem-perature tensile traction tests indicated a significant loss in me-chanical strength, which is expected to induce local membranedamage [33].

ORIG

INALRES

EARCH

PAPER

FUEL CELLS 15, 2015, No. 1, 204–213 ª 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 205www.fuelcells.wiley-vch.de

Sadeghi Alavijeh et al.: Decay in Mechanical Properties of Catalyst Coated Membranes

The objective of the present work is to systematically char-acterize the mechanical properties of catalyst coated mem-branes subjected to combined chemical and mechanical degra-dation, indicative of field operation where both stressors occursimultaneously. This article is a continuation of our previouslypublished study by Lim et al. [33], in which the overall degra-dation mechanism and failure modes were determined. Incontrast, the present work focuses exclusively on determiningthe evolution patterns of mechanical CCM properties duringdegradation. Tensile mechanical properties are measuredunder a wide range of precisely controlled temperature andhumidity conditions as a function of membrane health state.The emphasis is on the properties obtained at fuel cell condi-tions (elevated temperature and humidity), which is a novelcontribution of this article and expected to be more representa-tive of in situ operation and degradation. Additionally, thedegradation induced changes in CCM thermal and humidityexpansion properties are determined, which contributes a pre-viously missing essential link between in situ swelling andcontraction due to hygrothermal variations and the ex situmeasured tensile properties under an externally applied force.Unique correlations are provided between the degradationinduced evolution of expansion and tensile properties that canbe used to inform future advances in MEA durability. The roleof the obtained property changes in the joint chemical and me-chanical membrane degradation mechanism is discussed. Thepresent scope is limited to CCMs based on PFSA ionomermembranes and the results are expected to be generally validfor this category of materials.

2 Experimental

MEAs used for this study were fabricated by coating car-bon paper based gas diffusion layer (GDL) substrates withmicro-porous layers, followed by coating carbon supportedplatinum catalyst layers containing perfluorosulfonic acidionomer to form gas diffusion electrodes (GDEs). Then, MEAswere prepared by hot pressing the anode and cathode GDEswith PFSA membranes. The obtained MEAs were subjected tocombined chemical and mechanical degradation under BallardPower Systems’ cyclic open circuit voltage (COCV) acceler-ated stress test (AST) protocol [33]. Research-scale stacks wereassembled in a standard test hardware including five 45 cm2

cells separated by graphite bipolar plates with quasi-uniformconditions over the active area. Co-flow parallel straight chan-nels and high gas flow conditions were utilized, ensuring tominimize pressure drop between channels and avoid changesin reactant composition from stack inlet to outlet. Uniformcompression between MEAs and bipolar plates was achievedusing a pressurized bladder. A Ballard fuel cell test stationequipped with computer controlled mass flow controllers(MFCs), back pressure control system, pre-heated water circu-lating coolant loop, electronic load, and two water-injectedevaporator type humidifiers was used to apply acceleratedstress tests on research-scale stacks. Prior to AST operation,the MEAs were conditioned by operating the cells at constant

current for a few hours. Each AST cycle, as shown in Figure 1,included a high temperature/low RH steady state OCV phaseto accelerate chemical degradation followed by a set of wet/dry humidity cycles in N2 to provide mechanical degradation.In the wet period, the inlet gases were oversaturated by inject-ing water into the evaporator type humidifier in an amountlarger than that attaining the saturated vapor pressure. In thedry period, the inlet gases were kept in a relatively dry stateby closing the water supply pumps of the humidifiers. Mem-brane failure was determined by measuring a threshold con-vective hydrogen leak rate via an electrochemical leak detec-tion test. Further information regarding the COCV AST isprovided elsewhere [33].

In order to investigate mechanical properties, partiallydegraded MEAs were removed from the AST stack after con-ditioning, two, five, eight, and ten AST cycles and were pre-pared for mechanical testing. Degraded MEAs were cut to rec-tangular shapes (25 · 2 mm) along the transverse direction toensure consistency. The GDLs were removed from both sidesof the MEAs and the remaining CCM specimens were storedbetween two microscope glass slides at room conditions untiltesting. A dynamic mechanical analyzer (TA InstrumentsQ800 DMA) equipped with an environmental chamber (TAInstruments DMA-RH Accessory) was used for both tensileand expansion experiments. Tensile test specimens wereloaded with a 5:1 length to width aspect ratio to avoid edgestress concentration [25]. The remaining portion of the speci-men (15 mm in length) was positioned under the tensile grips.A specimen mounted for tensile and expansion characterizationin the DMA chamber is illustrated schematically in Figure 2.

Fig. 1 Graphical illustration of the cyclic open circuit voltage acceler-ated stress test (COCV AST) applied to generate combined chemical andmechanical membrane degradation [33].

Fig. 2 Schematic of a CCM specimen undergoing tensile and expansiontests.

ORIG

INALRES

EARCH

PAPER

206 ª 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim FUEL CELLS 15, 2015, No. 1, 204–213www.fuelcells.wiley-vch.de

Sadeghi Alavijeh et al.: Decay in Mechanical Properties of Catalyst Coated Membranes

The exact gauge length and width of each specimen were mea-sured prior to test initiation by the DMA and optical micro-scope, respectively, while the average membrane thicknesswas measured by micrometer through subtracting the catalystlayer thicknesses from the CCM thickness.

Tensile tests were conducted at standardized room condi-tions (23 �C and 50% RH) and fuel cell conditions (70 �C and90% RH) on beginning of life (BOL), BOL conditioned, andAST degraded CCMs. To ensure consistency and statisticalrelevance, four identical specimens were tested at each condi-tion. The measurement variability presented in the form oferror bars represents two standard deviations, which statisti-cally covers about 95% of the data. Specimens were loadedand equilibrated to desired environmental conditions while asmall preload tensile force (0.003 N) was applied to ensuretension. After equilibration, a fixed strain rate (0.01 min–1) wasapplied to slowly stretch the specimens until the fracture pointor until reaching the maximum DMA clamp travel length of26 mm (160% elongation for 10 mm initial length). Elastic modu-lus, ultimate tensile strength (UTS), and final strain were calcu-lated from the obtained stress-strain curves using the membranethickness for normalization, assuming that the membrane wasthe main load bearing part of the CCM [33]. The elastic moduluswas determined from the maximum slope of a 5th order poly-nomial fit to the initial section between 0 and 0.5% strain [25].The UTS and final strain were calculated as the maximum stressand total strain (elastic and plastic), respectively.

Expansion tests similar to those reported by Goulet et al.[25] were performed on CCM specimens extracted after BOLconditioning, five (half-life), and ten (end of life) AST cycles.Specimens were equilibrated at the desired initial conditionunder a constant tensile force (0.003 N). Humidity expansiontests were conducted by measuring the length of the specimenafter equilibration at fixed temperature (23 �C and 70 �C) whilethe humidity was raised from 25% to 50%, 50% to 70%, and70% to 90% RH. Temperature expansion tests were performedin the same manner at fixed humidity settings (50% and 90%RH) while the temperature was increased from 20 �C to 40 �C,40 �C to 55 �C, and 55 �C to 70 �C in discrete steps.

3 Results and Discussion

The cyclic open circuit voltage accelerated stress test(COCV AST) was applied on research-scale hardware to simu-late the chemical and mechanical degradation mechanisms ob-served in regular duty cycle operation of fuel cell stacks, andto generate degraded catalyst coated membrane (CCM) sam-ples for characterization of mechanical properties. TensileCCM properties were systematically measured at differentstages of degradation from beginning of life (BOL) to end oflife (EOL) under relevant environmental conditions rangingfrom room conditions to standard fuel cell conditions. Theobtained results, featuring tensile stress-strain curves, elasticmodulus, ultimate tensile strength, final strain, and thermaland humidity expansion properties, are presented and dis-cussed in this section.

3.1 AST Results

Firstly, the overall results of the combined chemical andmechanical COCV AST experiments are described. The OCVdecay behaviour and the results of the electrochemical leakdetection test (ELDT) ΔV are plotted in Figure 3 as func-tions of the number of COCV AST cycles. The OCV decay rateincreased gradually with respect to AST cycles and was ob-served to correlate well with the ELDT results which indicateda gradually rising rate of hydrogen leaks across the membranedue to the combined chemical and mechanical membrane deg-radation process. Major hydrogen leaks were detected afterten AST cycles (EOL) in accordance with the MEA failure cri-terion of the COCV AST [30]. The leaks were facilitated by theinitiation of small membrane pinholes and cracks during theintermediate AST phase (third to eighth cycle) and acceleratedsignificantly due to rapid growth of localized fractures in thefinal phase of the test (eighth cycle to EOL). Increased gascrossover through the degraded membrane was responsiblefor the concurrent OCV decay as a result of mixed electrodepotentials and subsequently lead to local heat generation viathe exothermal water formation reaction. Figure 4 shows the

Fig. 3 Variation in open circuit cell voltage (OCV) and electrochemicalleak detection test (ELDT) voltage differential results with respect to thenumber of COCV AST cycles.

Fig. 4 Normalized membrane thickness variation as a function ofCOCV AST cycles.

ORIG

INALRES

EARCH

PAPER

FUEL CELLS 15, 2015, No. 1, 204–213 ª 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 207www.fuelcells.wiley-vch.de

Sadeghi Alavijeh et al.: Decay in Mechanical Properties of Catalyst Coated Membranes

variation in the average membrane thickness as a function ofthe number of AST cycles, which indicates significant mem-brane thinning due to chemical degradation caused by radicalattack on the polymer structure [33]. While the original mem-brane thickness was retained during conditioning and the firsttwo AST cycles, uniform thinning proceeded from the thirdcycle and onwards until the final failure of the MEA after thetenth cycle, with 66% of the original thickness remaining. Thecumulative amount of anode, cathode, and total fluoride lossgathered from effluent water as a function of AST cycles is pre-sented in Figure 5. Fluoride loss, as a strong evidence of globalchemical membrane degradation during the OCV phase,shows a relatively steady increase during the intermediatephase of the experiment, in agreement with the rate of mem-brane thinning observed in Figure 4. The fluoride loss detectedthrough the cathode side was somewhat higher than at theanode for the entire duration of the experiment.

3.2 Stress-Strain Curves

Figure 6 shows a graphical summary of the tensile stress-strain curves obtained at (a) room and (b) fuel cell conditionsfor partially degraded CCM specimens extracted after differ-ent numbers of COCV AST cycles. As shown previously [25],the BOL CCM exhibits ductile behaviour under a wide rangeof temperature and relative humidity conditions. While noconsiderable changes in mechanical behaviour occurred dur-ing fuel cell assembly and conditioning, rapid reductions inelongation and ultimate tensile strength (UTS) were observedduring degradation from the onset of the AST until the EOLstate after ten AST cycles. Both BOL and BOL-conditionedCCMs exhibited excellent elongation and did not reach frac-ture within the maximum travel length of the DMA; however,after only two AST cycles, the elongation diminished and thefirst fracture was observed. As noted previously, at this earlystage of degradation, membrane damage such as pinholes and

cracks were not detected and the initial decay in mechanicalstrength (primarily ductility) was likely caused by polymerchain degradation and disentanglement. After merely fiveAST cycles, the specimen fractured immediately after the yieldpoint. With continued membrane degradation, the fracturestress decreased below the yield point where the materialremained in the viscoelastic region. By inspection of thecurves, it appears that although the elastic modulus was rela-tively constant during the AST process, both yield stress andfinal strain decreased substantially. Most notably, the fracturetoughness, which represents the total amount of energyrequired for fracture and can be estimated from the area belowthe stress-strain curve, dropped rapidly during the first fiveAST cycles. Membrane chain scission and reduction in molecu-lar weight due to chemical degradation are expected toweaken the macromolecular entanglements and cause theductile to brittle transition and rapid decay in fracture tough-ness observed in the stress-strain curves. In the second half ofthe experiment (from five to ten cycles), the fracture energywas negligible due to specimen failure before reaching theyield point.

3.3 Tensile Properties

The elastic modulus of the membrane (in the CCM) wascalculated from the slope of the stress-strain curves in theinitial, elastic region and presented in Figure 7 as a function ofCOCV AST cycles. In principle, the elastic modulus defines theresistance of the material against elastic deformation. Due tothe non-linear viscoelastic behaviour of the membrane, themaximum slope of the viscoelastic region in the stress-straincurve was selected as the elastic modulus. The average elasticmodulus was found to increase slightly from BOL to EOL atboth room and fuel cell conditions. These results wereobtained based on the actual thickness of the membrane aftereach corresponding AST cycle and therefore takes membranethinning into account in order to extract the true properties ofthe remaining material. On the contrary, should the BOLmembrane thickness be used and assumed to be constant, theelastic modulus would decrease from BOL to EOL. Previouswork has reported a reduction in elastic modulus due tochemical membrane degradation [31], although the calculationmethod was not disclosed. The increasing behaviour of theelastic modulus determined here indicates that the membranebecame progressively stiffer during exposure to combinedchemical and mechanical membrane degradation.

The total elongation of the CCM at the termination of thetensile test is displayed in Figure 8 as a function of AST cycles.As mentioned previously, the maximum measurable strain is160% due to the DMA travel length limitation. Therefore, thefinal strain obtained for both pristine materials (BOL andBOL-conditioned CCMs) reached 160% due to the absence offracture. However, the final strain of the degraded CCMsdecreased dramatically from ductile behaviour with high elon-gation values to brittle behaviour with almost no elongation atfinal fracture. At fuel cell conditions, the partially degraded

Fig. 5 Cumulative fluoride loss at anode side, cathode side, and total asa function of COCV AST cycles.

ORIG

INALRES

EARCH

PAPER

208 ª 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim FUEL CELLS 15, 2015, No. 1, 204–213www.fuelcells.wiley-vch.de

Sadeghi Alavijeh et al.: Decay in Mechanical Properties of Catalyst Coated Membranes

CCM extracted after two AST cycles exhibited relatively highvariability, as some specimens fractured at 91–130% elonga-tion while others reached the DMA displacement limit with-out fracture. This pattern suggests that the first few AST cyclesresulted primarily in localized degradation with potential frac-ture sites located far apart and still requiring substantial frac-ture energy to fail. However, the half-life CCM extracted afterfive cycles appeared to have lost almost all elongation capacitybefore failure, exhibiting less than 10% final strain. In addition,the half-life CCM had much more consistent failures than the

previously extracted CCMs, as demonstrated by the reducedvariability, suggesting that the degradation process had pro-ceeded to become more global in nature with more closelyspaced fracture sites. The continued degradation process inthe second half of the AST experiment led to further reduc-tions in final strain, down to very small elongation values ofless than 2% at EOL. The high hydrogen leak rates and severedamage of these CCMs suggest that pre-existing membraneholes and cracks contributed to rapid mechanical failures inthe tensile test.

a)

b)

Fig. 6 Stress-strain curves measured at (a) room conditions and (b) fuel cell conditions for partially degraded CCM specimens extracted after differentnumbers of COCV AST cycles (as indicated). For graphical clarity, the origin was shifted horizontally by 5% for each curve, although all measurementswere started at zero strain.

ORIG

INALRES

EARCH

PAPER

FUEL CELLS 15, 2015, No. 1, 204–213 ª 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 209www.fuelcells.wiley-vch.de

Sadeghi Alavijeh et al.: Decay in Mechanical Properties of Catalyst Coated Membranes

The ultimate tensile strength (UTS) of the CCMs is depictedin Figure 9 as a function of AST cycles. The UTS was found todecrease in a linear pattern from BOL to EOL, where the CCMhad lost 49% and 69% of its initial stress tolerance under roomand fuel cell conditions, respectively. Interestingly, thedecreasing UTS trend follows qualitatively the trends ob-served previously for membrane thinning and fluoride release.This suggests that the decay in UTS was caused by globalchemical degradation. Reductions in polymer molecularweight and main chain length due to chemical degradationare expected to be the main source of UTS decay, in agreementwith previous literature [31]. However, the overall impact ofthe molecular changes is likely exacerbated by the applicationof mechanical stress, acting to disentangle the ionomer bun-dles and disintegrate the ionic clusters, either in the form of exsitu stress applied by the tensile experiment or in the form ofin situ stress applied by the intermittent wet/dry cycles in theCOCV AST. Indeed, the small UTS resulting in tensile failureof heavily degraded CCMs would likely be equivalent to thestress generated in situ during wet/dry cycling, where actualfailures were evident. In addition, the variability observed inthe UTS data was found to increase gradually with degrada-tion, in particular under fuel cell conditions, suggesting thatpre-existing, randomly located membrane holes and cracksmay lead to lower local stress tolerance and earlier tensile frac-tures. This effect would also contribute to the in situ growth ofholes and cracks during wet/dry cycling and more rapid fail-ures during combined chemical and mechanical degradation

a)

b)

Fig. 7 Evolution of CCM elastic modulus during combined chemical andmechanical membrane degradation applied by the COCV AST, as mea-sured by tensile test at (a) room condition and (b) fuel cell condition.

a)

b)

Fig. 8 Evolution of CCM final strain during combined chemical and me-chanical membrane degradation applied by the COCV AST, as mea-sured by tensile test at (a) room condition and (b) fuel cell condition.

a)

b)

Fig. 9 Evolution of CCM ultimate tensile strength during combined chem-ical and mechanical membrane degradation applied by the COCV AST,as measured at (a) room condition and (b) fuel cell condition.

ORIG

INALRES

EARCH

PAPER

210 ª 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim FUEL CELLS 15, 2015, No. 1, 204–213www.fuelcells.wiley-vch.de

Sadeghi Alavijeh et al.: Decay in Mechanical Properties of Catalyst Coated Membranes

(e.g., COCV AST) compared to isolated chemical degradation(e.g., OCV hold), as demonstrated in our previous work [33].

3.4 Expansion Properties

While tensile properties are convenient to measure ex situand useful in the context of membrane degradation, the exter-nally applied mechanical stress is fundamentally different tothe stresses generated in situ due to variations in hygrothermalconditions. Understanding the CCM in situ mechanical behav-ior is crucial due to the time-dependent response under varia-tions in hygrothermal conditions, inherited from the elastic-viscoplastic characteristics of PFSA ionomer membranes.Furthermore, the effect of catalyst layers coated on the mem-brane on the swelling and contraction of membrane was eval-uated by Goulet et al. [25] and confirmed the significant rein-forcing role of catalyst layers limiting the membranedimensional changes. It was previously observed that thetime-dependent response of a PFSA membrane may be alteredby chemical degradation, as shown by Patil et al. [31, 32] viatensile creep experiments. In order to capture the correcttrends for the in situ fuel cell environment, complementaryhygrothermal expansion tests were conducted to assess theevolution of the mechanical behaviour of the degraded CCMsin response to changes in temperature and humidity. Theexpansion tests were performed under a constant preloadforce by applying a step change in either temperature or rel-ative humidity (RH) and measuring the resulting strain uponspecimen equilibration. The obtained thermal expansionresults at constant RH of 50% and 90% are provided in Figures10 and 11, respectively. The data points represent thermalexpansion in specimen length from 20 �C to the desired tem-perature (as indicated). A linear trend was found for thermalexpansion in all cases. Thermal elongation at fixed RH is likelydue to a combination of thermal expansion and increasedwater uptake, as the moisture content of the air increases withtemperature despite the constant RH. Depending on mem-brane preparation conditions, Hinatsu et al. [15] observed lin-ear and second order curves for water uptake with respect tovariation in temperature in Nafion� 117 and 125 samples, in

reasonable qualitative agreement with the present results. Thehighest thermal expansion was obtained with the pristineCCM (at BOL), both in terms of total expansion and expansionrate, as shown in Table 1. The thermal expansion of the CCMwas significantly reduced at the half-life degradation state andfurther reduced at the final EOL state. The overall decay inthermal expansion from BOL to EOL was approximately 80%at both humidity conditions (from 7.9% to 1.5% at 50% RHand from 6.4% to 1.4% at 90% RH). Similarly, Figures 12 and13 represent the corresponding hygral expansion results atconstant temperatures of 23 �C and 70 �C, respectively. In thiscase, the response of the CCM to changes in humidity was lesssignificant than for temperature changes and was only mar-ginally reduced by the AST degradation (25–30% overalldecay).

Fig. 10 CCM thermal expansion at 50% RH for BOL, half-life (five ASTcycles), and EOL (ten AST cycles) degradation states. Expansion isshown for heating from 20 �C to 40 �C, 55 �C, and 70 �C.

Fig. 11 CCM thermal expansion at 90% RH for BOL, half-life (five ASTcycles), and EOL (ten AST cycles) degradation states. Expansion isshown for heating from 20 �C to 40 �C, 55 �C, and 70 �C.

Table 1 Thermal expansion rates at different levels of degradation andrelative humidity.

50% RH 90% RH

BOL 0.00157 0.00117

Half-life 0.00108 0.00042

EOL 0.00030 0.00031

Fig. 12 CCM hygral expansion at 23 �C for BOL, half-life (five ASTcycles), and EOL (ten AST cycles) degradation states. Expansion isshown for humidification from 25% RH to 50%, 70%, and 90% RH.

ORIG

INALRES

EARCH

PAPER

FUEL CELLS 15, 2015, No. 1, 204–213 ª 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 211www.fuelcells.wiley-vch.de

Sadeghi Alavijeh et al.: Decay in Mechanical Properties of Catalyst Coated Membranes

Correlations between the trends in CCM expansion proper-ties and tensile properties may provide valuable linkagesbetween the ex situ measured mechanical behaviour and thecombined chemical and mechanical membrane degradationprocess occurring in situ. The overall decay in hygrothermalexpansion properties appears to correlate reasonably wellwith the decay in membrane thickness, remaining fluoridecontent, and ultimate tensile strength. Additionally, the loss ofhygrothermal expansion is analogous to the loss of ductilityobserved in the stress-strain curves after only a few ASTcycles. Although loss in hygrothermal expansion capabilitiescan be favourable in order to decrease the mechanical stressesoccurring in situ, localized degradation may lead to non-uni-form dimensional changes that can accelerate crack initiationand propagation in these areas. The main implication of thisloss, as mentioned previously, is the reduced fracture tough-ness, which makes the CCM more susceptible to crack forma-tion. More importantly, while the net hygrothermal expansionat BOL could easily be accommodated within the elastic re-gime without major implications, the ~2% hygrothermalexpansion observed at EOL is within the same range of strainas the final tensile strain, which indicates that the membrane issufficiently weakened for ruptures to form as a direct result ofthe hygrothermal variations during wet/dry cycling. Thisfinding is expected to play a key role in the development ofmembrane damage and ultimate failure due to combinedchemical and mechanical membrane degradation.

4 Conclusions

The decay in mechanical properties of catalyst coated PFSAmembranes was investigated in the context of combinedchemical and mechanical membrane degradation representa-tive of regular duty cycle fuel cell operation. Ballard’s COCVAST protocol, which deploys elevated chemical and mechani-cal stressors, was utilized to generate partially degraded CCMmaterials for analysis. Tensile test results demonstrated a mildincrease in elastic modulus accompanied by dramatic reduc-tions in final strain and ultimate tensile strength during the

degradation process, which indicates a gradual membranetransformation from a soft and ductile material to a stiff andbrittle object with local variations in properties. The rapiddecay in fracture toughness observed during the early stagesof degradation was likely caused by locally elevated chemicaldegradation leading to potential sites for fracture initiation.Moreover, the brittleness caused by chemical degradationreduced the membrane resistance against crack propagationdue to mechanical degradation via wet/dry cycles. Expansiontest results revealed linear reductions in CCM hygrothermalexpansion characteristics during the membrane degradationprocess. A critical linkage between hygrothermal expansionand fracture strain was identified in the late stages of degrada-tion. In this regime, given the underlying chemical degrada-tion and weakened ionomer structure, the regular hygrother-mal cycles occurring in situ are deemed to be sufficient togenerate fractures and to proliferate the overall membranedamage towards ultimate failure. Overall, the decay in me-chanical properties determined in this work forms the core forthe interaction between chemical and mechanical membranedegradation during fuel cell operation. Prevention of suchdecay is essential for development of suitable membrane/CCM stabilization strategies toward enhanced fuel cell dur-ability and lifetime.

Acknowledgements

The authors gratefully acknowledge Automotive Partner-ship Canada (APC), Ballard Power Systems, Natural Sciencesand Engineering Research Council, and Simon Fraser Univer-sity for supporting this project. The authors also thank JoannaKolodziej, Spencer Arbour, and Frederick Van Hove for techni-cal assistance.

References

[1] C. S. Gittleman, F. D. Coms, Y. Lai, in Polymer ElectrolyteFuel Cell Degradation (Eds: M. M. Mench, E. C. Kumbur,T. N. Veziroglu), Elsevier Inc., 2012, pp. 15–88.

[2] F. D. Coms, ECS Transactions 2008, 16, 235.[3] S. Zhang, X. Yuan, H. Wang, W. Merida, H. Zhu, J. Shen,

S. Wu, J. Zhang, Int. J. Hydrogen Energ. 2009, 34, 388.[4] A. Collier, H. Wang, X. Zi Yuan, J. Zhang, D. P. Wilkin-

son, Int. J. Hydrogen Energ. 2006, 31, 1838.[5] M. P. Rodgers, L. J. Bonville, H. Russell Kunz, D. K. Slat-

tery, J. M. Fenton, Am. Chem. Soc. 2012, 112, 6075.[6] J. Healy, C. Hayden, T. Xie, K. Olson, R. Waldo,

M. Brundage, H. Gasteiger, J. Abbott, Fuel Cells 2005, 5,302.

[7] S. Zhang, X.-Z. Yuan, J. N. C. Hin, H. Wang, J. Wu, K. A.Friedrich, M. Schulze, J. Power Sources 2010, 195, 1142.

[8] S. Kundu, M. Fowler, L. C. Simon, R. Abouatallah,J. Power Sources 2008, 182, 254.

[9] X. Yuan, S. Zhang, H. Wang, J. Wu, J. C. Sun, R. Hiesgen,K. A. Friedrich, M. Schulze, A. Haug, J. Power Sources2010, 195, 7594.

Fig. 13 CCM hygral expansion at 70 �C for BOL, half-life (five ASTcycles), and EOL (ten AST cycles) degradation states. Expansion isshown for humidification from 25% RH to 50%, 70%, and 90% RH.

ORIG

INALRES

EARCH

PAPER

212 ª 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim FUEL CELLS 15, 2015, No. 1, 204–213www.fuelcells.wiley-vch.de

Sadeghi Alavijeh et al.: Decay in Mechanical Properties of Catalyst Coated Membranes

[10] J. Wu, X. Z. Yuan, J. J. Martin, H. Wang, D. Yang, J. Qiao,J. Ma, J. Power Sources 2010, 195, 1171.

[11] P. Trogadas, J. Parrondo, V. Ramani, Electrochem. Solid.ST 2008, 11, B113.

[12] B. P. Pearman, N. Mohajeri, R. P. Brooker, M. P. Rodgers,D. K. Slattery, M. D. Hampton, D. A. Cullen, S. Seal,J. Power Sources 2013, 225, 75.

[13] L. Ghassemzadeh, S. Holdcroft, J. Am. Chem. Soc. 2013,135, 8181.

[14] N. Macauley, L. Ghassemzadeh, C. Lim, M. Watson,J. Kolodziej, M. Lauritzen, S. Holdcroft, E. Kjeang, ECSElectrochemistry Letters 2013, 2, F33.

[15] J. T. Hinatsu, M. Mizuhata, H. Takenaka, J. Electrochem.Soc. 1994, 141, 1493.

[16] K. Broka, P. Ekdunge, J. Appl. Electrochem. 1997, 27, 117.[17] T. A. Zawodzinski, C. Derouin, S. Radzinski, R. J. Sher-

man, V. T. Smith, T. E. Springer, S. Gottesfeld, J. Electro-chem. Soc. 1993, 140, 1041.

[18] Y. Tang, A. M. Karlsson, M. H. Santare, M. Gilbert,S. Cleghorn, W. B. Johnson, Mater. Sci. Eng. A 2006, 425,297.

[19] H. Tang, S. Peikang, S. P. Jiang, F. Wang, M. Pan, J. PowerSources 2007, 170, 85.

[20] F. Bauer, S. Denneler, M. Willert-Porada, J. Polym. Sci.Pol. Phys. 2005, 43, 786.

[21] M. F. Mathias, R. Makharia, H. A. Gasteiger, J. J. Conley,T. J. Fuller, C. J. Gittleman, S. S. Kocha, D. P. Miller,C. K. Mittelsteadt, T. Xie, S. G. Yan, P. T. Yu, U. S. H.Hoover, Electrochem. Soc. Interface 2005, 24.

[22] S. Kundu, L. C. Simon, M. W. Fowler, Polym. Degrad. Sta-bil. 2008, 93, 214.

[23] X. Huang, R. Solasi, Y. Zou, M. Feshler, K. Reifsnider,D. Condit, S. Burlatsky, T. Madden, J. Polym. Sci. Pol.Phys. 2006, 44, 2346.

[24] T. T. Aindow, J. O'Neill, J. Power Sources 2011, 196, 3851.[25] M. A. Goulet, R. M. H. Khorasany, C. De Torres,

M. Lauritzen, E. Kjeang, G. G. Wang, N. Rajapakse,J. Power Sources 2013, 234, 38.

[26] R. M. H. Khorasany, M.-A. Goulet, A. S. Alavijeh,E. Kjeang, G. G. Wang, R. K. N. D. Rajapakse, J. PowerSources 2014, 252, 176.

[27] N. Garland, T. Benjamin, J. Kopasz, ECS Transactions2007, 11, 923.

[28] B. Sompalli, B. A. Litteer, W. Gu, H. A. Gasteiger, J. Elec-trochem. Soc. 2007, 154, B1349.

[29] M. Liu, C. Wang, F. Xie, Z. Mao, Int. J. Hydrogen Energ.2013, 38, 11011.

[30] S. J. Bae, S. J. Kim, J. I. Park, J.-H. Lee, H. Cho, J. Y. Park,Int. J. Hydrogen Energ. 2010, 35, 9166.

[31] Y. P. Patil, W. L. Jarrett, K. A. Mauritz, J. Membrane Sci.2010, 356, 7.

[32] Y. Patil, S. Sambandam, V. Ramani, K. Mauritz, J. Electro-chem. Soc. 2009, 156, B1092.

[33] C. Lim, L. Ghassemzadeh, F. Van Hove, M. Lauritzen,J. Kolodziej, G. G. Wang, S. Holdcroft, E. Kjeang,J. Power Sources 2014, 257, 102.

______________________

ORIG

INALRES

EARCH

PAPER

FUEL CELLS 15, 2015, No. 1, 204–213 ª 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 213www.fuelcells.wiley-vch.de