a modified decal method for preparing the membrane electrode assembly of proton exchange membrane...

8
A modified decal method for preparing the membrane electrode assembly of proton exchange membrane fuel cells Xiaolu Liang a,b , Guoshun Pan a,b,, Li Xu a,b , Jiashu Wang a a State Key Laboratory of Tribology, Tsinghua University, Beijing 100084, China b Shenzhen Key Laboratory of Micro-nano Manufacturing, Research Institute of Tsinghua University in Shenzhen, Shenzhen 518057, China highlights The complete transfer is achieved by optimizing the catalyst inks’ composition and a hot peeling technique. The thickness of catalyst layers (before transfer) is consistent. The effects of drying methods, transfer pressure, test temperatures are explored and discussed. The MEAs show better performance. article info Article history: Received 4 May 2014 Received in revised form 4 August 2014 Accepted 8 September 2014 Available online 18 September 2014 Keywords: Proton exchange membrane fuel cells Membrane electrode assembly Decal method Hot peeling abstract Decal transfer is an effective method for fabricating membrane electrode assemblies (MEAs), due to its low interfacial resistance and applicability for mass production. Here we introduce a modified decal method which makes it possible and convenient to achieve complete decal transferring. The composition of catalyst inks, drying process and transfer pressure are optimized in detail. During catalyst ink prepa- ration, the viscosity is adjusted by altering the composition of the solvents to obtain catalyst layers (before transfer) with continuous thickness. In addition, MEAs whose catalyst layers (before transfer) were dried in four different ways are tested for application in a proton exchange membrane fuel cell. The transfer pressure is also optimized on the basis of the two previous conditions and the MEAs fabri- cated by this modified method show simplicity to achieve complete transfer in decal method, good repeatability and improvement in cell performance. Ó 2014 Elsevier Ltd. All rights reserved. 1. Introduction Proton exchange membrane fuel cells (PEMFCs) have attracted a great deal of attention due to several significant properties as a promising power sources for residential applications, vehicles, and portable electronic devices [1–3]. As the key component of PEMFC, the membrane electrode assembly, where both hydrogen oxidation and oxygen reduction reactions take place, is the major research area, especially studies on the fabrication. The decal method, in which catalyst inks are coated on decal substrates, dried and transferred to a membrane, is currently regarded as the most suitable approach for the commercial production of MEAs. Most studies of the decal method usually focused on catalyst inks, drying methods, complete transfer and hot-pressing conditions. Attention on catalyst inks can be classified as follows: (i) Coating methods, which are well known as spraying, doctor blading, and screen printing [1,4,5]. (ii) Dispersion solvents, whose volatility, viscosity and dielectric constant determine, to a great degree, the porosity of catalyst layers, the paintability of catalyst inks and the aggregation of Nafion ionomer [2,6–11]. (iii) Nafion- to-carbon ratio and solvent-to-carbon ratio. Plenty of studies focused on influence of Nafion content in catalyst layers on the resulting cell performance, and the reported Nafion loading varied from 20% to 50% along with Pt loading, weight percentage of Pt/C and preparation methods of MEAs [12–17]. As to the solvent-to- carbon ratio, both porosity of catalyst layers, which plays an important role on activation area and the mass transport rate, and residual organic solvents are research emphasis, especially the decal transfer method used [9,18]. (iv) Additives to improving water management by diminishing flooding in cathode catalysts layer, include pore formers, hydrophobic materials, oxide nanoparticles, and carbon materials which have fibrous morphology [3,19–24]. http://dx.doi.org/10.1016/j.fuel.2014.09.022 0016-2361/Ó 2014 Elsevier Ltd. All rights reserved. Corresponding author at: State Key Laboratory of Tribology, Tsinghua Univer- sity, Beijing 100084, China. Tel.: +86 0755 26957243; fax: +86 0755 26957301. E-mail addresses: [email protected] (X. Liang), [email protected] (G. Pan). Fuel 139 (2015) 393–400 Contents lists available at ScienceDirect Fuel journal homepage: www.elsevier.com/locate/fuel

Upload: jiashu

Post on 16-Feb-2017

212 views

Category:

Documents


0 download

TRANSCRIPT

Fuel 139 (2015) 393–400

Contents lists available at ScienceDirect

Fuel

journal homepage: www.elsevier .com/locate / fuel

A modified decal method for preparing the membrane electrodeassembly of proton exchange membrane fuel cells

http://dx.doi.org/10.1016/j.fuel.2014.09.0220016-2361/� 2014 Elsevier Ltd. All rights reserved.

⇑ Corresponding author at: State Key Laboratory of Tribology, Tsinghua Univer-sity, Beijing 100084, China. Tel.: +86 0755 26957243; fax: +86 0755 26957301.

E-mail addresses: [email protected] (X. Liang), [email protected](G. Pan).

Xiaolu Liang a,b, Guoshun Pan a,b,⇑, Li Xu a,b, Jiashu Wang a

a State Key Laboratory of Tribology, Tsinghua University, Beijing 100084, Chinab Shenzhen Key Laboratory of Micro-nano Manufacturing, Research Institute of Tsinghua University in Shenzhen, Shenzhen 518057, China

h i g h l i g h t s

� The complete transfer is achieved by optimizing the catalyst inks’ composition and a hot peeling technique.� The thickness of catalyst layers (before transfer) is consistent.� The effects of drying methods, transfer pressure, test temperatures are explored and discussed.� The MEAs show better performance.

a r t i c l e i n f o

Article history:Received 4 May 2014Received in revised form 4 August 2014Accepted 8 September 2014Available online 18 September 2014

Keywords:Proton exchange membrane fuel cellsMembrane electrode assemblyDecal methodHot peeling

a b s t r a c t

Decal transfer is an effective method for fabricating membrane electrode assemblies (MEAs), due to itslow interfacial resistance and applicability for mass production. Here we introduce a modified decalmethod which makes it possible and convenient to achieve complete decal transferring. The compositionof catalyst inks, drying process and transfer pressure are optimized in detail. During catalyst ink prepa-ration, the viscosity is adjusted by altering the composition of the solvents to obtain catalyst layers(before transfer) with continuous thickness. In addition, MEAs whose catalyst layers (before transfer)were dried in four different ways are tested for application in a proton exchange membrane fuel cell.The transfer pressure is also optimized on the basis of the two previous conditions and the MEAs fabri-cated by this modified method show simplicity to achieve complete transfer in decal method, goodrepeatability and improvement in cell performance.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Proton exchange membrane fuel cells (PEMFCs) have attracted agreat deal of attention due to several significant properties as apromising power sources for residential applications, vehicles,and portable electronic devices [1–3]. As the key component ofPEMFC, the membrane electrode assembly, where both hydrogenoxidation and oxygen reduction reactions take place, is the majorresearch area, especially studies on the fabrication. The decalmethod, in which catalyst inks are coated on decal substrates, driedand transferred to a membrane, is currently regarded as the mostsuitable approach for the commercial production of MEAs.

Most studies of the decal method usually focused on catalystinks, drying methods, complete transfer and hot-pressing

conditions. Attention on catalyst inks can be classified as follows:(i) Coating methods, which are well known as spraying, doctorblading, and screen printing [1,4,5]. (ii) Dispersion solvents, whosevolatility, viscosity and dielectric constant determine, to a greatdegree, the porosity of catalyst layers, the paintability of catalystinks and the aggregation of Nafion ionomer [2,6–11]. (iii) Nafion-to-carbon ratio and solvent-to-carbon ratio. Plenty of studiesfocused on influence of Nafion content in catalyst layers on theresulting cell performance, and the reported Nafion loading variedfrom 20% to 50% along with Pt loading, weight percentage of Pt/Cand preparation methods of MEAs [12–17]. As to the solvent-to-carbon ratio, both porosity of catalyst layers, which plays animportant role on activation area and the mass transport rate, andresidual organic solvents are research emphasis, especially thedecal transfer method used [9,18]. (iv) Additives to improvingwater management by diminishing flooding in cathode catalystslayer, include pore formers, hydrophobic materials, oxidenanoparticles, and carbon materials which have fibrousmorphology [3,19–24].

394 X. Liang et al. / Fuel 139 (2015) 393–400

It is worth mentioning that, with the aid of optimized catalystinks in the past several decades, improvement on performance ofMEAs prepared by different methods is obvious to all [2,7,9–11].In decal transfer method, catalyst inks are generally applied ontopolymer films to form catalyst layers at first, by means of a doctorblade in general, and then transferred onto both sides of Nafionmembrane by hot pressing. Numerous studies have concentratedon maximizing the transfer yield to cut loss of catalyst duringthe MEAs fabrication, since it is a critical and challenging targetin decal method. Ha et al. [25,26] discussed the influence of decalselections on complete transfer, and the result proved that Telfonwas far superior to Kapton. By employing a novel liquid nitrogenfreezing method, transfer yield was also increased from 95.4% to99.2%, presented by Cho et al. [27]. Moreover, after it was discov-ered that a thin dense Nafion ionomer ‘‘skin’’ formed during hot-pressing process, on account of strong hydrophobicity of the decalsurface, Park and Cho et al. [25,28,29] made use of a carbon break-ing layer to insure a high transfer ratio which was applied to decalbefore catalyst inks coating. Liquid treatment on electrolyte mem-brane was also proposed to achieve complete transfer which alsomade the process complicated [30]. It can be concluded that, it willbe inspiring to achieve complete transfer via a convenient produc-tion process at the aid of optimizing catalyst inks coated on asuitable decal substrate.

In the past two decades, much attention has been focused onthe high viscosity solvents used in catalyst inks, particularly glyc-erol [4,9,18,28,31,32]. Effects of addition, extraction and propor-tion of glycerol on cell performance were investigated in detail.However, as another organic solvent with high dielectric constantand viscosity, ethylene glycol which ever acts as the secondarypore former is less studied [33,34]. In this present research, deion-ized water, isopropyl alcohol and ethylene glycol are chosen as dis-persion solvents. Influence of weight proportion of the three onviscosity is explored. On the basis of high-performing catalyst inks,a hot peeling technique is raised to achieve complete transfer with-out any auxiliary process. To optimize this modified decal method,effect of drying ways and transfer pressure on the single cell per-formance of MEAs is also discussed.

2. Experimental

2.1. Preparation of catalyst inks

The Pt/C catalyst (40 wt.% Pt content, Johnson Matthey),isopropyl alcohol, ethylene glycol, deionized water, and 5% Nafionsolution (Dupont, Inc.) were mixed by an ultra turrax and in anultrasonic bath at room temperature. The weight ratios of Pt/Ccatalyst, isopropyl alcohol, ethylene glycol and deionized waterwere varied. The viscosities of mixed solutions with differentweight ratios were measured at 25 �C using a rotational rheometer(Malvern, Inc.). In all MEAs, the weight content of Nafion in catalystlayers is adjusted to 30% for all samples, and the Pt loading of cath-ode and anode is kept at 0.25 mg cm�2 and 0.10 mg cm�2

respectively.

2.2. MEA fabrication using the decal method

The prepared catalyst inks were coated onto decal substrates byan automatic film applicator. The fluorinated ethylene propylene(FEP) films were chosen as decal substrates. The blade gap of theautomatic film applicator was set as 200 lm for the cathode and100 lm for the anode. Meanwhile, the blade speed was kept at38.6 mm/s for the cathode and 98.3 mm/s for the anode. After this,the decal substrates were dried in a vacuum oven or air dryingoven at different temperatures (60 �C or 120 �C) for at least 4 h.

The Nafion HP membrane (Dupont, Inc.) was selected for the pro-ton exchange membrane. The coated decal substrates were assem-bled by sandwiching the Nafion HP membrane and hot pressingunder different pressures (4 MPa, 6 MPa, 8 MPa or 10 MPa) at120 �C for 3 min. After the decal substrates were peeled off, themargin of MEAs was sealed by Kapton tape to shape a 25 cm2 reac-tion area, which enhanced the MEAs’ air tightness to a large degree.To obtain complete transfer, the whole hot peeling process wasoperated on the hot plate of the hot press machine whose temper-ature was still maintained at 120 �C. The two decal substrates werepeeled off one by one in 15 s after the hot press machine stoppedworking. In contrast, decal substrates were also peeled off at roomtemperature, according to the traditional transfer process. Fig. 1shows the whole MEA preparation process during this experiment.

2.3. Characterization

The morphological and elementary characteristics of catalystlayers and decal substrates were investigated using the FE-SEM(MIR3 XMH, Tescan, Ltd.). To obtain cross-sectional images, thedecal substrates and MEAs were split by a knife and then stuckto a vertical surface. The pore size distribution of the catalyst layerswas measured with a pore size analyzer (V-Sorb 2800P, Gold APPInstruments, Ltd.). The sealed MEAs were sandwiched betweentwo carbon papers with microporous layers (GDS 3260, Avcarb,Ltd.) and then assembled with composite, one-serpentine flowfield plates at a pressure of 20 kg cm–2. The performance of sin-gle-cells was tested using a Fuel Cell Test System (TEC-TS500,Top Energy, Ltd.) at 65 �C and ambient pressure. H2 and air weresupplied to the anode and cathode at flow rates of 700 ml min–1

and 3000 ml min–1, respectively. The humidification temperaturefor both gases was maintained at 85 �C. The single-cells were acti-vated at a constant voltage of 0.6 V for 4 h, then at 0.5 V for at least20 h before characterization. The polarization curves wereobtained by measuring the voltage using stepwise increments ofcurrent density, at intervals of 100 mA cm–2; each measurementlasted for 60 s. The EIS was tested at i = 1000 mA cm–2, and thescanning frequency varied from 0.1 Hz to 10,000 Hz. The hydrogenhumidification temperature, air humidification temperature andcell temperature were kept at 85 �C, 85 and 65 �C, respectively.The cell performance under different humidities was tested byaltering the three above temperatures.

3. Results and discussions

3.1. Effect of the catalyst inks’ composition on coating

In this experiment, paintability for catalyst inks is achievedwhen the carbon-to-solvent weight ratio (CSWR) is above 0.06.As the Nafion content is kept at 30 wt.%, the CSWR values onlyinclude isopropyl alcohol, ethylene glycol, and deionized water.To investigate further the effect of catalyst ink composition oncoating, the viscosity of catalyst inks and cross-sections of catalystlayers before transfer are characterized. All the catalyst inks wereapplied to FEP films by the automatic film applicator. As samples#1, #2, and #3 in Table 1 show, the viscosity of catalyst inksincreases as the CSWR changes from 0.06 to 0.12. Sample #4 sub-stitutes ethylene glycol with the same weight of isopropyl alcoholthat is in sample #2, whereas the opposite is the case for sample#5. Comparison of samples #2, #4 and #5 makes it clear that eth-ylene glycol increases the catalyst inks’ viscosity, which influencestheir flow velocity when they are being coated onto decal sub-strates. It is evident that the viscosity of catalyst inks also can beadjusted through solvent selections, and paintability in the decalmethod can be improved by selecting high-viscosity solvents

Fig. 1. Schematic diagram of the modified decal method for preparing MEAs.

Table 1Composition and viscosity of catalyst inks.

Sample Solvent weight (g) CSWR Viscosity (Pa s)

Isopropyl alcohol Ethylene glycol Deionized water

#1 2.39 2.39 3.54 0.06 0.01124#2 1.59 1.59 2.36 0.09 0.01244#3 1.20 1.20 1.77 0.12 0.01464#4 – 3.18 2.36 0.09 0.01346#5 3.18 – 2.36 0.09 0.01147

X. Liang et al. / Fuel 139 (2015) 393–400 395

instead of low-viscosity ones, as the latter leads to high CSWR val-ues in inks and low porosity in catalyst layers. To achieve goodpaintability, the proportion of Pt/C catalyst was great for thelow-viscosity which led to high CSWR values. As the solventsreduced, pores formed by solvents volatilizing in catalyst layersdecreased in the drying process. To research the effect of ethyleneglycol on the consistency of catalyst layer thickness, samples #1,#2, #3, #4 and #5 are coated onto decal substrates under the sameconditions and then dried in an air drying oven at 120 �C for 4 h.The cross-sections of catalyst layers before transfer are character-ized with the aid of FE-SEM (Fig. 2). As the CSWR of catalyst inksadded, the thickness of catalyst layers increased as shown inFig. 2(a–c). Because of the low viscosity, it is difficult to form con-tinuous a catalyst layer for sample #1. For the catalyst ink contain-ing ethylene glycol, the catalyst layer on the transfer film ishomogenous and maintains a continuous thickness of 6.91 lm(Fig. 2(d) and (e)). In the case of sample #5, in which ethylene gly-col is replaced with isopropyl alcohol, whose viscosity is quite low,the catalyst ink tends to flow on the transfer film, leading to incon-sistent thickness; the maximum thickness is approximately threetimes larger than the minimum, as Fig. 2(f) shows. It can be con-cluded that the viscosity of catalyst inks influences the uniformityand thickness of the catalyst layers, and the presence of ethyleneglycol facilitates the even spreading of the catalyst inks.

3.2. Forced air drying and vacuum drying methods

This section addresses the influence of drying methods and tem-perature on the single-cell performance of MEAs. Several researchgroups have chosen to use atmospheric drying during MEA fabrica-tion, when solvents with low boiling points, such as isopropyl alco-hol, are used for catalyst inks [25,27,35,36]. Cho et al. [27]investigated atmospheric drying and vacuum drying and found that

small pores formed in the catalyst layers during vacuum drying,enhanced the single-cell performance of their MEAs. Regarding eth-ylene glycol, which has a boiling point of 197.3 �C, there are no stud-ies on the effect of drying method on the single-cell performance. Inthis present study, four drying conditions are chosen: in an air dryingoven at 60 �C or 120 �C, and in a vacuum oven at 60 �C or 120 �C. Allof the drying processes last 4 h and the pressure in the hot pressingprocess is 6 MPa. Sample #2 is used to prepare the MEAs. Fig. 3shows the surface and cross-sectional micrographs of the catalystlayers dried under these four different conditions. As their surfacemicrostructure shows, the porosity increases along with the changesin temperature and drying conditions, proving that vacuum dryingat 120 �C promotes the evaporation of ethylene glycol and the for-mation of a porous structure. However, this difference is less sharplyvisible in the cross-sectional micrographs due to the depth-of-fieldand surface irregularities caused by cutting the catalyst layers. Inthe analysis of pore size distributions of catalyst layers in Fig. 4,the difference of small pores in the range of 1–100 nm can be found.These graphs illustrate that vacuum drying and high temperaturepromote the formation of small pores in the MEA. Fig. 5 presentspolarization curves of the MEAs fabricated under the four differentdrying conditions. It is obvious that the cell performance of thefour MEAs is similar when the current density is lower than400 mA cm–2. However, the difference in both cell voltage andpower density becomes larger in the high current region, especiallyabove 800 mA cm–2. For air drying at 60 �C and 120 �C, the maxi-mum current densities hardly reach 1200 mA cm–2 and the cell volt-ages of the MEAs are lower than those of the MEAs prepared withvacuum drying at the same temperatures. In addition, the cell per-formance levels of the MEAs whose catalyst layers were dried at120 �C are more favorable than those dried at 60 �C. These resultscan be explained in terms of different ethylene glycol residuesresulting from the four drying conditions. As Fig. 5 shows, the gap

Fig. 2. Cross-sections of catalyst layers for samples #1 (a), #2 (b), #3 (c), #4 (d, f) and #5 (f).

396 X. Liang et al. / Fuel 139 (2015) 393–400

of polarization curves as a function of drying conditions increaseswith the current density. It can be concluded that the vacuum andhigh temperature drying method facilitates the extraction of ethyl-ene glycol and improves the cell performance. To verify this idea,Fig. 6 presents further information on the AC impedance of the fourMEAs. It should be noted that the cathode charge transfer resistanceincreases as the extraction of ethylene glycol decreases, which iscaused by drying conditions. This result is consistent with the abovepolarization curves and confirms that catalyst layers possess enoughpores when they are dried in a vacuum oven at 120 �C. This fine porestructure provides appropriate paths for gas flow and waterremoval, both of which increase the number of reaction sites onthe Pt catalyst. Therefore, the cell performance is improved andthere is less overpotential in the high current region.

3.3. Hot peeling for complete transfer

Several studies have focused on the effect of polymer films suchas Teflon and Kapton on transfer yield. In this experiment, trans-parent FEP film with good flexibility is first introduced as the decalsubstrate to achieve complete transfer with the aid of the preparedsample #2. As shown in step 5 of Fig. 1, the decal substrates are

peeled quickly before the catalyst layers cool down. By contrast,in traditional transfer the peeling step is carried out at room tem-perature. Here, complete transfer MEAs are achieved through hotpeeling, as shown in Fig. 7. During traditional transfer, residueresults because polymer films with higher hardness at room tem-perature are more difficult to peel. In this modified decal method,a transfer yield of 100% is realized just with the help of suitablecatalyst inks and a hot peeling technique that is easy to processand highly repeatable. This method will make mass productionefficient and realizable, especially using a roll-press to fabricateMEAs. Moreover, the transfer yield and Pt loading can be calculatedaccurately by weighing decal substrates before and after transfer,since the thickness of the catalyst layers is consistent (Fig. 2).

3.4. Effects of hot-pressing conditions

This section focuses on the effects of hot-pressing pressure onthe surface topography of the catalyst layers, the thickness of MEAs,and single-cell performance, since hot-pressing temperaturesbetween 100 �C and 140 �C are available for complete transfer inthis modified decal method. Most reported studies involving hot-pressing conditions have sought to determine the relationship

Fig. 3. Surface (a–d) and cross-sectional (e–h) micrographs of catalyst layers dried under four different conditions: atmospheric drying at 60 �C; atmospheric drying at120 �C; vacuum drying at 60 �C and vacuum drying at 120 �C.

Fig. 4. Comparison of the pore size distribution of catalyst layers dried under fourdifferent conditions: atmospheric drying at 60 �C; atmospheric drying at 120 �C;vacuum drying at 60 �C and vacuum drying at 120 �C.

Fig. 5. Comparison of the single-cell performance of MEAs fabricated usingdifferent drying methods.

Fig. 6. Comparison of the EIS spectra of MEAs fabricated using different dryingmethods.

Fig. 7. Decal substrates before and after catalyst layers were transferred ontomembranes.

X. Liang et al. / Fuel 139 (2015) 393–400 397

Fig. 9. Comparison of the single-cell performance of MEAs fabricated underdifferent pressures.

Fig. 10. Comparison of the EIS spectra of MEAs fabricated under different pressures.

398 X. Liang et al. / Fuel 139 (2015) 393–400

between transfer yield and hot-pressing conditions [26,27,34]. Yimet al. [37] evaluated water management capability of catalyst layersand cell performance under different pressure conditions. For thehot pressing process, Suzuki et al. [38] also focused on effects ofpressure on the structure of catalyst layers. They indicated thatthe cell performance was improved as the increased pressureenhanced the ionomer and carbon network and also found theporosity and the pore size reduced voltage in higher pressing pres-sure range. However, in this experiment it is necessary to discoverthe optimal pressure when the solvents in the catalyst inks and thedrying methods are varied. Here, 4 MPa, 6 MPa, 8 MPa, and 10 MPaare selected for fabricating MEAs. Also, to understand the micro-structures of the catalyst layers pressed under these four pressures,surface and cross-sectional micrographs are obtained by SEM. Asshown in Fig. 8, well-developed porous structures can be identifiedin the surface images, even when the catalyst layers have beenpressed. Comparison of Fig. 8(a) and (b) shows that the apertureclearly decreases as the pressure increases; however, this differenceis not apparent when comparing (b)–(d). The thickness of the MEAs,as (e)–(h) show, decreases by 30.8%, 11.1%, and 5.6%. This differencealso decreases significantly when the pressure is above 6 MPa. Itcan be concluded that the influence of pressures over 6 MPa on cat-alyst layers diminishes gradually. Fig. 9 illustrates the cell perfor-mance of the four MEAs fabricated under different pressuresduring the decal process. Comparing the polarization curves for4 MPa and 6 MPa shows that the difference in cell voltage is evidentin both high and low current regions. This demonstrates that theconnection between the catalyst layers and the electrolyte mem-brane is weak under 4 MPa, since both of them contain the samepolymer. When the hot-pressing pressure is above 6 MPa, the cellvoltage as a function of pressure drops slightly. It can be demon-strated that small pore structures, as the SEM images in Fig. 8 show,reduce the rate of mass transport that increases overpotential lossand lowers the cell performance. To further evaluate the differencebetween the polarization curves in Fig. 9, EIS tests are performed forthe MEAs fabricated under four different pressures. As the real axisintercept of the EIS curve shows in Fig. 10, the ohmic resistance of4 MPa is significantly higher than that of 6 MPa, which is due tothe interface contact resistance mentioned above, and to thereduced proton conductivity and increased electrical resistance,both of which are attributable to the large pore structure. Althoughthe charge transfer resistance of this MEA is obviously less than thatof the other three, the ohmic resistance lowers the cell

Fig. 8. Surface (a–d) and cross-sectional (e–h) micrographs of MEAs pressed under 4 MPa, 6 MPa, 8 MPa, and 10 MPa, respectively.

Fig. 11. Effects of cell and humidification temperatures on cell performance ofsamples #a and #b. (The order of the temperatures is: hydrogen humidificationtemperature, air humidification temperature and cell temperature.)

X. Liang et al. / Fuel 139 (2015) 393–400 399

performance, as indicated by the polarization curves in Fig. 9. Forthe MEAs fabricated using 6 MPa, 8 MPa, or 10 MPa, the differencesin ohmic resistances are very minor, but the charge transfer resis-tance increases due to the reduction in catalyst active sites causedby the added pressure. The SEM images, polarization curves, and EISspectra demonstrate that the cell performance depends on variouslosses caused by ohmic, activation, and concentration overpoten-tials. There is a trade-off between the advantages and disadvan-tages of the hot-pressing pressure. For the MEA fabricated using4 MPa, ohmic loss is the main reason for the reduced cell perfor-mance. When the pressure is above 6 MPa, the ohmic resistancechanges slightly, and the mass transfer resistance becomes themajor factor influencing the cell performance. Therefore, it can beconcluded that 6 MPa of pressure yields a favorable catalyst elec-trode, appropriate reactant accessibility to the reaction region,and better reaction kinetics for catalyst utilization.

3.5. Effect of test temperatures

It is well known that test temperatures, including both fuel tem-perature and cell temperature, largely determine cell performance.To research the cell performance of the MEAs fabricated by thismodified decal method under both low and high humidity condi-tions, cell temperature, hydrogen humidification temperature,and air humidification temperature are varied as Fig. 11 shows.The samples #a and #b are fabricated under the same condition.Both of them are dried in the vacuum oven at 120 �C and hotpressed under 6 MPa. In the whole current density range, the cellperformance of the two samples improves as the cell and humidifi-cation temperatures increase. This can be attributed to the catalystactivity which changes with temperature. Moreover, there are onlysmall differences between samples #a and #b under different tem-perature conditions, which indicates good repeatability of the MEAsfabricated by this modified decal method.

4. Conclusions

This is the first report to describe using hot peeling in decaltransfer to achieve complete transfer without any additional pro-cess. The success of the technique is based on the right combina-tion of catalyst inks and decal substrate. EDX characterizationshows that hot peeling significantly improves transfer yieldcompared with the traditional peeling method. Hot peeling leaves

no catalyst residue because the transparent FEP film peels moreeasily above room temperature. Catalyst ink coating, drying meth-ods, transfer pressure, and testing temperatures are also studied inrelation to this modified decal method. We find that ethylene gly-col increases the viscosity of catalyst inks, and also facilitates evenspreading of the catalyst inks, yielding catalyst layers with consis-tent thickness after drying. In addition, due to the high boilingpoint of ethylene glycol, vacuum drying at 120 �C facilitates extrac-tion of this solvent to form pores. The MEA fabricated by thisdrying method shows the best cell performance, as illustrated bypolarization curves and EIS spectra. Furthermore, transfer pressureis optimized by using 4 MPa, 6 MPa, 8 MPa and 10 MPa. The SEMphotographs and polarization curves show that the influence ofpressure above 6 MPa on the catalyst layers falls gradually. Whenthe pressure is lower than 6 MPa, the cell voltage is lower thanfor other MEAs at the same current density in both high and lowcurrent regions, which illustrates that the connection betweenthe catalyst layers and the electrolyte membrane may be weak.With the optimized drying method and transfer pressure, the cur-rent density in this experiment reaches to 1000 mA cm–2 when thevoltage is 0.6 V, the average Pt loading is 0.175 mg cm–2, and thereactant gas in cathode is air.

Acknowledgements

This work was supported by National Natural Science Foundationof China (No. 91223202), International Science & TechnologyCooperation Program of China (No. 2011DFA73410), National KeyBasic Research Program of China (973 Program, No.2011CB013102), and Tsinghua University Initiative ScientificResearch Program (No. 20101081907). All the financial support isgratefully acknowledged.

References

[1] Lee HK, Park JH, Kim DY, Lee TH. A study on the characteristics of the diffusionlayer thickness and porosity of the PEMFC. J Power Sources 2004;131:200–6.

[2] Yang TH, Yoon YG, Park GG, Lee WY, Kim CS. Fabrication of a thin catalyst layerusing organic solvents. J Power Sources 2004;127:230–3.

[3] Zhang X, Shi P. Dual-bonded catalyst layer structure cathode for PEMFC.Electrochem Commun 2006;8:1229–34.

[4] Bender G, Zawodzinski TA, Saab AP. Fabrication of high precision PEFCmembrane electrode assemblies. J Power Sources 2003;124:114–7.

[5] Hwang DS, Park CH, Yi SC, Lee YM. Optimal catalyst layer structure of polymerelectrolyte membrane fuel cell. Int J Hydrogen Energy 2011;36:9876–85.

[6] Shin SJ, Lee JK, Ha HY, Hong SA, Chun HS, Oh IH. Effect of the catalyst inkpreparation method on the performance of polymer electrolyte membrane fuelcells. J Power Sources 2002;106:146–52.

[7] Fernández R, Ferreira-Aparicio P, Daza L. PEMFC electrode preparation:influence of the solvent composition and evaporation rate on the catalyticlayer microstructure. J Power Sources 2005;151:18–24.

[8] Sung KA, Jung HY, Kim WK, Cho KY, Park JK. Influence of dispersion solvent forcatalyst ink containing sulfonated poly (ether ether ketone) on cathodebehaviour in a direct methanol fuel cell. J Power Sources 2007;169:271–5.

[9] Jung CY, Kim WJ, Yi SC. Optimization of catalyst ink composition for thepreparation of a membrane electrode assembly in a proton exchangemembrane fuel cell using the decal transfer. Int J Hydrogen Energy2012;37:18446–54.

[10] Ngo TT, Yu TL, Lin HL. Influence of the composition of isopropyl alcohol/watermixture solvents in catalyst ink solutions on proton exchange membrane fuelcell performance. J Power Sources 2013;225:293–303.

[11] Ngo TT, Yu TL, Lin HL. Nafion-based membrane electrode assemblies preparedfrom catalyst inks containing alcohol/water solvent mixtures. J Power Sources2013;238:1–10.

[12] Wilson MS, Gottesfeld S. Thin film catalyst layers for polymer electrolyte fuelcell electrodes. J Appl Electrochem 1992;22:1–7.

[13] Passalacqua E, Lufrano F, Squadrito G, Patti A, Giorgi L. Nafion content in thecatalyst layer of polymer electrolyte fuel cells: effects on structure andperformance. Electrochim Acta 2001;46:799–805.

[14] Antolini E, Giorgi L, Pozio A, Passalacqua E. Influence of nafion loading in thecatalyst layer of gas-diffusion electrode for PEFC. J Power Sources1999;77:136–42.

[15] Sasikumar G, Ihm JW, Ryu H. Dependence of optimum nafion content incatalyst layer on platinum loading. J Power Sources 2004;132:11–7.

400 X. Liang et al. / Fuel 139 (2015) 393–400

[16] Sasikumar G, Ihm JW, Ryu H. Optimum nafion content in PEM fuel cellelectrodes. Electrochim Acta 2004;50:601–5.

[17] Kim KH, Lee KY, Kim HJ, Cho E, Lee SY, Lim TH, et al. The effect of nafion�

ionomer content in PEMFC MEAs prepared by a catalyst-coated membrane(CCM) spraying method. Int J Hydrogen Energy 2010;35:2119–26.

[18] Chisaka M, Daiguji H. Effect of glycerol on micro/nano structure of catalystlayers in polymer electrolyte membrane fuel cells. Electrochim Acta2006;51:4828–33.

[19] Chao WK, Lee CM, Tsai DC, Chou CC, Hsueh KL, et al. Improvement of theproton exchange membrane fuel cell (PEMFC) performance at low-humidityconditions by adding hygroscopic c-Al2O3 particles into the catalyst layer. JPower Sources 2008;185:136–42.

[20] Reshetenko TV, Kim HT, Kweon HJ. Modification of cathode structure byintroduction of CNT for air-breathing DMFC. Electrochim Acta2008;53:3043–9.

[21] Reshetenko TV, Kim HT, Kweon HJ. Cathode structure optimization for air-breathing DMFC by application of pore-forming agents. J Power Sources2007;171:433–40.

[22] Cho YH, Jung N, Kang YS, Chung DY, Lim JW, et al. Improved mass transferusing a pore forming in cathode catalyst layer in the direct methanol fuel cell.Int J Hydrogen Energy 2012;37:11969–74.

[23] Li A, Chan SH, Nguyen NT. Anti-flooding cathode catalyst layer for highperformance PEM fuel cell. Electrochem Commun 2009;11:897–900.

[24] Therdthianwong A, Saenwiset P, Therdthianwong S. Cathode catalyst layerdesign for proton exchange membrane fuel cells. Fuel 2012;91:192–9.

[25] Cho JH, Kim JM, Prabhuram J, Hwang SY, Ahn DJ, Ha HY, et al. Fabrication andevaluation of membrane electrode assemblies by low-temperature decalmethods for direct methanol fuel cells. J Power Sources 2009;187:378–86.

[26] Mehmood A, Ha HY. An efficient decal transfer method using a roll-press tofabricate membrane electrode assemblies for direct methanol fuel cells. Int JHydrogen Energy 2012;37:18463–70.

[27] Cho HJ, Jang H, Lim S, Cho E, Lim TH, Oh IH, et al. Development of a novel decaltransfer process for fabrication of high-performance and reliable membraneelectrode assemblies for PEMFCs. Int J Hydrogen Energy 2011;36:12465–73.

[28] Xie J, Garzon F, Zawodzinski T, Smith W. Ionomer segregation in compositeMEAs and its effect on polymer electrolyte fuel cell performance. JElectrochem Soc 2007;151:A1084–93.

[29] Park HS, Cho YH, Cho YH, Park IS, Jung N, Ahn M, et al. Modified decal methodand its related study of microporous layer in PEM fuel cell. J Electrochem Soc2008;155:B455–60.

[30] Yoon YJ, Kim TH, Kim SU, Yu DM, Hong YT. Low temperature decal transfermethod for hydrocarbon membrane based membrane electrode assemblies inpolymer electrolyte membrane fuel cells. J Power Source 2011;196:9800–9.

[31] Xie J, More KL, Zawodzinski TA, Smith WH. Porosimetry of MEAs made by‘‘thin film decal’’ method and its effect on performance of PEFCs. J ElectrochemSoc 2004;151:A1841–6.

[32] Chun YG, Kim CS, Peck DH, Shin DR. Performance of a polymer electrolytemembrane fuel cell with thin film catalyst electrodes. J Power Sources1998;71:174–8.

[33] Yoon YG, Park GG, Yang TH, Han JN, Lee WY, Kim CS. Effect of pore structure ofcatalyst layer in a PEMFC on its performance. Int J Hydrogen Energy2003;28:657–62.

[34] Saha MS, Paul DK, Peppley BA, Karan K. Fabrication of catalyst-coatedmembrane by modified decal transfer technique. Electrochem Commun2010;12:410–3.

[35] Jeon S, Lee J, Rios GM, Kim HJ, Lee SY, Cho E, et al. Effect of ionomer content andrelative humidity on polymer electrolyte membrane fuel cell (PEMFC)performance of membrane-electrode assemblies (MEAs) prepared by decaltransfer method. Int J Hydrogen Energy 2010;35:9678–86.

[36] Mehmood A, Ha HY. An efficient decal transfer method using a roll-press tofabricate membrane electrode assemblies for direct methanol fuel cell. Int JHydrogen Energy 2012;37:18463–70.

[37] Yim SD, Sohn YJ, Park SH, Yoon YG, Park GG, Yang TH. Fabrication ofmicrostructure controlled cathode catalyst layers and their effect on watermanagement in polymer electrolyte fuel cell. Electrochim Acta 2011;56:9046–73.

[38] Suzuki T, Tsushima S. Pore structure and cell performance analysis of PEMFCcatalyst layers fabricated by decal transfer method with variation of hotpressing pressure. ECS Trans 2011;41:909–14.