High performance direct methanol fuel cells withmicro/nano-patterned polymer electrolyte membrane

Yoon-Hwan Cho a,b, Jin Woo Bae b, Ok-Hee Kim a,b, Jae Young Jho b, Namgee Jung a,b,Kyusoon Shin b, Hyelim Choi c, Heeman Choe c, Yong-Hun Cho c,n, Yung-Eun Sung a,b,1

a Centre for Nanoparticle Research, Institute for Basic Science (IBS), Seoul National University, Seoul 151-744, South Koreab School of Chemical and Biological Engineering, Seoul National University (SNU), Seoul 151-744, South Koreac School of Advanced Materials Engineering, Kookmin University, 861-1 Jeongneung-dong, Seoul 136-702, Seongbuk-gu, South Korea

a r t i c l e i n f o

Article history:Received 14 November 2013Received in revised form17 March 2014Accepted 30 March 2014Available online 12 April 2014

Keywords:Direct methanol fuel cell (DMFC)Thermal imprint lithography (TIL)Nafion 115 membranePatterningThree-phase boundary

a b s t r a c t

The effect of an enlarged specific surface area of the membrane with well-defined line patterns on theperformance of a direct methanol fuel cell (DMFC) is investigated and compared with the baselinepristine Nafion 115 membrane. Line patterns with dimensions ranging from several tens of nanometersto several micrometers were fabricated on Nafion 115 membranes with high reliability using thermalimprint lithography to ensure an uncollapsible structure. In the case of quasi-nano-patterned membranethe cell performance increased about 35% compared with that of the pristine Nafion 115 membraneowing to an increased effective three-phase boundary caused by an enlarged specific surface area. Thusthe performance of DMFCs can be improved further by controlling the shape and size of the line patternsfor sufficient formation of the three-phase boundary.

& 2014 Elsevier B.V. All rights reserved.

1. Introduction

The direct methanol fuel cell (DMFC) is an eco-friendly energyconversion system, generating electricity directly through theelectrochemical reaction of methanol and oxygen [1,2]. Moreover,the DMFC is one of the best alternative portable power sourcesowing to its high energy density as well as the ease with which itcan be applied to small electronic devices [3,4]. On the other hand,resolving its cost issue, which stems primarily from the use of a Ptcatalyst in the membrane–electrode assembly (MEA), remains thelargest obstacle to the widespread application of DMFCs to mobileand stationary platforms [5,6]. Accordingly, many attempts havebeen made to reduce the amount of Pt in the catalyst layer withoutdeteriorating the DMFC performance [7–11].

Several methods of creating as large a ‘three-phase boundary’area as possible, where the reactants, catalysts and electrolytecoexist in the MEA, have attracted considerable interest becausethey might improve the electrochemical performance of DMFCsand reduce the amount of Pt catalyst needed. For example,roughening the surface of the electrolyte membrane comprisingthe three-phase boundary in the MEA has been attempted

recently. Since Sheppard et al. first reported that the electrolytemembrane could be roughened by SiC paper [12,13], additionalinvestigations into modifying the surface structure of the electro-lyte membrane have been performed by plasma etching or ionbeam bombardment [14–17]. Yildirim and coworkers reported theincreased MEA performance with micro-patterned Nafion 117membrane using a hot embossing process [18]. DMFC performanceof electron beam patterned membrane was further investigated byOmosebi and Besser [19]. Nevertheless, previously modified elec-trolyte membranes with large specific surface areas had somelimitations regarding the electrochemical fuel cell performance.The whisker-like structure of the surface-modified electrolytemembranes collapsed readily during MEA preparation, whichdisrupted the three-phase boundary where the electrochemicalreaction occurred. Therefore, it is essential to design an electrolytemembrane with both an enlarged surface area and robust surfacestructure so that an effective three-phase boundary with theloaded catalysts and reactants can be formed in the MEA, whichthus improves the conversion efficiency of electrochemical energyin DMFCs.

This communication reports the improved single-cell performanceof electrolyte membranes containing well-defined micro-sized, quasi-nano-sized and nano-sized line patterns. The line patterns wereintroduced to the electrolyte membrane to ensure an uncollapsiblestructure with a high-specific surface area. Uniform line patterns,ranging in size from several tens of nanometers to several tens of

Contents lists available at ScienceDirect

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

Journal of Membrane Science

http://dx.doi.org/10.1016/j.memsci.2014.03.0690376-7388/& 2014 Elsevier B.V. All rights reserved.

n Corresponding author. Tel.: þ82 2 910 5672; fax: þ82 2 910 5674.E-mail addresses: yhun00@kookmin.ac.kr (Y.-H. Cho),

ysung@snu.ac.kr (Y.-E. Sung).1 Tel.: þ82 2 880 1889; fax: þ82 2 888 1604.

Journal of Membrane Science 467 (2014) 36–40

micrometers, were developed to measure quantitatively the effect ofan enlarged specific membrane surface area on the performance ofDMFCs. The well-defined line patterns were imprinted on both sidesof an electrolyte membrane by thermal imprint lithography (TIL)[20–22]. TIL is expected to increase the reactive surface area of thethree-phase boundary and enhance the performance of DMFCsconsiderably because a wide range of well-defined line patterns canbe fabricated on an electrolyte membrane without collapsing thestructure.

2. Experimental methods

2.1. Preparation of micro-patterned master molds [23]

Periodic micro-scaled and quasi-nano-scaled Si line patternswere fabricated using photolithography at the Korea AdvancedNano Fab Center (KANC). An n-type Si (100) wafer (MEMC Electro-nic Materials, Inc.) was etched by piranha solution (3:1 mixture ofconcentrated sulfuric acid and 30% hydrogen peroxide) at roomtemperature for 1 h and ultrasonicated with acetone to clean thesurface of Si wafer. Photoresist (AZ7210) was spun-cast on thecleaned Si wafer substrate at 3000 rpm resulting in 1 μm thicklayer, followed by baking at 100 1C for 90 s. Exposure was per-formed with a manual mask aligner (EVD620, EVD) at 40 mJ cm�2

and followed by post-exposure bake at 110 1C for 60 s. Afterdevelopment, the patterned photoresist on the Si wafer was bakedat 120 1C for 90 s and the Si wafer substrate was selectively etchedwith inductively coupled plasma (Multiplex ICP, STS). Leftoverphotoresist on the etched Si wafer was removed with the Micro-wave Asher (ALA-0601E, AMS). Height and line intervals of the twopatterns are 1 μm and 15 μm, and 700 nm and 700 nm for micro-scaled and quasi-nano-scaled patterns, respectively.

2.2. Preparation of nano-patterned master mold [24]

Periodic nano-scaled Si line pattern was fabricated using e-beamlithography (JBX9300FS, JEOL) at the Korea Advanced Nano FabCenter (KANC). Height and line intervals of the periodic nano-scaledSi master mold were 70 nm and 200 nm, respectively, upon whichseveral Rigiflex master molds were replicated. ZEP 520A was usedas an e-beam resist and the exposure does used was 300 μC cm�2.UV-curable polymer solution (MINS 311, Minuta tech.) was droppedon the Si master line pattern template. Adhesion promoter-coatedPET film (Minuta tech.) was then placed on top of the MINS 311 as asupporting layer. After UV-curing for 20 s, rigiflex master mold wasphysically detached from the Si master. Prior to use, the rigiflexmaster molds were aged for several hours.

2.3. Fabrication of MEAs with the patterned Nafion membranes andthe pristine Nafion 115 membrane

Fabrication of the MEAwith the line-patterned Nafion membraneis shown schematically in Fig. 1. Nafion 115 membrane (equivalentweight value¼1100 g eq.�1, thickness: 120 μm, DuPont) was used asa polymer electrolyte membrane for the DMFC. First, line-patternedmaster molds were in contact with both sides of Nafion 115membrane, and pressed with 200 Pa at 180 1C for 1 h in a vacuumoven. Following slow cooling to room temperature and release ofmechanical pressure, the resulting Nafion membranes wereimmersed in distilled water for 24 h at 50 1C so that the patternedNafion membrane can naturally detach from the master molds. Toremove remaining impurities prior to applying the patterned Nafionmembrane to the MEA, the patterned Nafion membranes weretreated sequentially in a 3 wt% solution of hydrogen peroxide,distilled water, 0.5 M sulfuric acid, and distilled water, each at 70 1C

Fig. 1. Schematic illustration of the fabrication of the MEA using the line-patterned Nafion membrane. (a) Line patterns on both sides of the Nafion 115 membrane werereplicated by the master molds using TIL. And then, (b) the MEA was fabricated by directly spraying the Pt/C catalyst ink onto the both sides of line-patterned Nafionmembrane without a hot-pressing process.

Y.-H. Cho et al. / Journal of Membrane Science 467 (2014) 36–40 37

for 1 h. After treatment, the patterned Nafion membranes werepreserved in distilled water to keep the patterned Nafion membranefully hydrated. The catalyst ink was prepared by ultrasonication, fromthe following elements: 75 wt% carbon-supported PtRu (JohnsonMatthey) for anode and 60 wt% carbon-supported Pt (JohnsonMatthey) for cathode, 5 wt% Nafion solution (DuPont), isopropylalcohol (Aldrich), and deionized water. The content of Nafion solutionwas 30 wt%, and the ratio of IPA and water was 3:1. Finally, MEAswith an active area of 5 cm2 were fabricated by directly spraying thecatalyst ink on both sides of the patterned Nafion membraneswithout a hot-pressing process [25]. The amounts of PtRu/C andPt/C catalyst used in the spraying technique were 2.0 mg cm�2 and1.5 mg cm�2, respectively. An MEA with a pristine Nafion 115membrane was fabricated by the above-described procedure, exceptfor the thermal imprinting lithography. After drying the MEAs for 1 hat room temperature, the single cell was assembled with the MEA,gas diffusion layer (GDL), and graphite plates. All the single cells wereclamped with torque of 60 kgf cm.

2.4. Characterization

The morphology of the line-patterned master molds, the inverselyline-replicated Nafion membranes, and the membrane–electrodeinterface were investigated using a FE-SEM (JSM-6700F, JEOL) withan acceleration voltage of 5 kV. In order to investigate changes in thechemical properties of the surface of the thermal imprinted mem-brane, FT-IR ATR spectra (Thermo Scientific, Nicolet 6700) wereobtained from 32 scans at a resolution of 4 cm�1 from 4000�1 to400 cm�1, as seen in Fig. S1. Wide Angle X-ray Scattering (WAXS,Bruker) patterns were recorded in the diffraction angular range of 10–401 2θ, using a GD 2000 transmission diffractometer produced by ItalStructures, working in the Seemann-Bohlin geometry, and with aquartz crystal monochromator of the Johansson type on the primaryX-ray beam using CuK radiation. Small Angle X-ray Scattering (SAXS)patterns were recorded by anMBraun system, using the CuK radiationfrom a Philips PW 1830 X-rays generator in the diffraction angularrange of 1–91 2θ. The membrane through-plane proton conductivitywas determined according to the impedance technique by means ofsolartron 1260 Impedance/Gain Phase Anayser (IM6, Zahner) in the

frequency range 100 mHz–10 kHz at a signal amplitude 100 mV.Membrane disc, 8 mm in diameter was sandwiched between twoPt plates. To evaluate DMFC performance, single cells were operatedat a constant temperature of 60 1C and ambient pressure, by feeding1.0 M methanol solution (0.5 ml min�1) and dry air (200 sccm) forthe anode and cathode, respectively. The voltages of operating singlecells were obtained with increasing current density, by using a fuelcell test station (CNL Energy) to measure the current–voltage char-acteristics [26,27]. The single cell test was performed with threedifferent cells, of which the mean values were taken with their errorbars based on 7standard deviation values.

3. Results and discussion

Before the electrolyte membranes were patterned by TIL, mastermolds of the line patterns with a range of periodicities underseveral tens of micrometers had been prepared by photolithographyfor micropatterning and quasi-nanopatterning [23] as well as e-beam lithography and imprinting process for nanopatterning [24].Electrolyte membranes with enlarged specific surface areas werefabricated by sandwiching a Nafion membrane between the mastermolds with well-defined periodic line patterns, followed by thermalimprinting and removal of the master molds. Fig. 2 shows cross-sectional images of the line-patterned master molds (top) alongwith their corresponding electrolyte membranes (bottom), asobserved by field-emission scanning electron microscopy (FE-SEM). As shown in Fig. 2a–c (top), the width/spacing/height of thethree master molds were 10/5/1 μm, 700/700/700 nm, and 90/110/70 nm, respectively. The top images show well-defined lines foreach type of line pattern on both the micro and nano-scalesindicating the quality of the master molds.

The inverse structure of each master mold was transferred to thesurface of a thin Nafion membrane by TIL. Fig. 2d–f (bottom) showsthese line-patterned surface morphologies. The imprinted Nafionmembranes have well-defined line patterns with width/spacing/height values of 5/10/1 μm, 700/700/700 nm, and 110/90/70 nm,respectively. Based on the FE-SEM images, the uniform line patternsfrom the master molds, which range in size from several tens ofnanometers to several tens of micrometers, were well-replicated on

Fig. 2. FE-SEM images of line-patterned master molds (top) and the replicated Nafion membranes (bottom). Various line-patterned master molds with well-defined linepatterns (a)–(c) were prepared. The structure of the master molds are inversely replicated on the Nafion substrates by TIL. The lines on the Nafion substrate are (d) 5 μm inwidth, 10 μm in spacing, and 1 μm in height; (e) 700 nm, 700 nm, and 700 nm; (f) 110 nm, 90 nm, and 70 nm, respectively.

Y.-H. Cho et al. / Journal of Membrane Science 467 (2014) 36–4038

the Nafion membranes by TIL without collapsing the structure. Basedon estimations of enlarged specific surface areas of the imprintedNafion membranes, the specific surface areas of the imprinted Nafionmembranes were approximately 113%, 200%, and 170%, compared tothat of the pristine Nafion 115 membrane, when their width/spacing/height values were 5/10/1 μm, 700/700/700 nm, and 110/90/70 nm,respectively. These observations suggest that uncollapsed line patternscan be prepared readily on the Nafion membranes by TIL and thesurface area to volume ratio can be controlled straightforwardly. Theseresults may provide solutions to improving the formation of aneffective three-phase boundary in the MEA, which leads to DMFCswith better performance. In previous studies, surface roughening ofthe electrolyte membrane for extremely enlarged surface area was notsuccessful in improving the formation of a stable three-phase bound-ary during the preparation of MEA owing to the instability ofelectrolyte membrane surface structure [14,15], which made it difficultto demonstrate the effect of an enlarged specific surface area of themembrane on the performance of DMFCs. In this study, Fouriertransform infrared (FT-IR) spectroscopy in ATR mode, wide-angleX-ray scattering (WAXS), and small-angle X-ray scattering (SAXS)were performed to examine the changes in the chemical propertiesof the surface of the thermal-imprinted membrane. As shown in Figs.S1–S3, no peak shift was observed in the FT-IR spectrum of thepatterned Nafion membrane, which was measured by considering allwave numbers similar to that of the untreated Nafion 115 membrane.This suggests that TIL used to improve the MEA performance did notcause any major changes in the chemical properties of the Nafionmembrane. Furthermore, the proton conductivity and methanol cross-over rate of the patterned Nafionmembranewere unaffected by TIL, asshown in Fig. S4 and Table 1.

The MEAs were fabricated by spraying the well-dispersed 75 wt%PtRu/C and 60 wt% Pt/C catalysts in the anode and cathode sides,respectively, on the pristine and line-patterned Nafion membranes,followed by single-cell tests to measure the performance of theprepared MEAs in a 1 M methanol/air DMFC. Fig. 3 shows thesingle-cell performance of the MEAs fabricated using the pristineand patterned Nafion membranes as a polymer electrolyte. The open-circuit voltages of almost all the single cells were held constant at0.76 V, regardless of the membrane surface structure, which suggeststhat methanol crossover from the anode to cathode across thepatterned Nafion membrane did not affect the performance of theDMFC, despite the patterning introduced to both sides of the Nafionmembrane. The current density of the pristine Nafion 115 membranewas measured to be 0.173 A cm�2 at a cell voltage of 0.4 V, and itsevaluated maximum power density was 0.104W cm�2. The currentdensities of the MEAs with the micro-patterned and quasi-nano-patterned membranes were respectively 0.179 and 0.193 A cm�2 at0.4 V, whereas that of the MEA with the nano-patterned membranewas only 0.145 A cm�2. Furthermore, the similar trend was observedin the maximum power density measurements; the maximum powerdensities were 0.122, 0.138, and 0.109W cm�2 for the line-patternedmembranes with width/spacing/height values of 5/10/1 μm, 700/700/700 nm, and 110/90/70 nm, respectively. The performance results aresummarized in Table 1. Compared with the MEA of the pristine Nafion115 membrane, the MEA of the Nafion membrane line-patterned inquasi nanoscale showed remarkably improved performance (�35%

increase in maximum power density) in the DMFC owing to theenlarged specific surface area of the membrane. This dramatic increasein performance without the additional use of more Pt electrocatalyst issubstantial. Adequately spaced line patterns, particularly on the quasinanoscale, as shown in Fig. 4(a–c), appear to increase the interfacecontact area of the three phase boundary in the same geometric size,thereby improving the fuel cell performance significantly and increas-ing the Pt utilization effectively. Therefore, a properly patternedelectrolyte membrane with an uncollapsible structure and high-specific surface area are the key parameters determining the perfor-mance of DMFCs.

On the other hand, the MEA with the Nafion membrane line-patterned on the nanoscale exhibited the worst single-cell perfor-mance. The performance at 0.4 V was even poorer than that of thesingle cell with the pristine Nafion 115 membrane. Indeed, weexpected that single cell performance of the MEA with the Nafionmembrane line-patterned on the nanoscale could exhibit betterthan those of the MEAs with the pristine membrane and micro-patterned membrane, in anticipation for the large specific surfacearea of the nano-patterned Nafion membrane with a stablestructure formation as shown in Fig. 2(f). The Pt/C catalystparticles often agglomerated to form particles of more than severalhundreds of nanometers when sprayed over the line-patternedNafion membrane (Fig. S1) [25]. As such, the agglomerated Pt/Ccatalyst particles could not penetrate into the narrow intervals ofthe nano-patterned membrane despite its good stability, deterior-ating the formation of an effective three-phase boundary in theMEA, as shown in Fig. 4(c). On the other hand, the reactive area ofthe three-phase boundary could increase when the surface struc-ture of the Nafion membrane was modulated appropriately on themicroscale level as compared with the nanoscale level. Moreover,the patterning scale was further optimized on the quasi nanoscalelevel (several hundreds of nanometers).

Table 1Summary of the performance characteriztion results for the tested membranes. The data were obtained from Supplementary Fig. S4 and Fig. 3.

Membrane Proton conductivity(S cm�1)

Methanol crossover(cm2 s�1)

Current density at 0.4 V(A cm�2)

Maximum power density(W cm�2)

Pristine Nafion 115 0.100 2.02�10�6 0.17370.003 0.10470.002Patterned Nafion membrane with 5/10/1 μm 0.105 2.00�10�6 0.17970.001 0.12270.003Patterned Nafion membrane with 700/700/700 nm 0.100 2.01�10�6 0.19370.001 0.13870.003Patterned Nafion membrane with 110/90/10 nm 0.106 2.02�10�6 0.14570.002 0.10970.002

Fig. 3. Polarization and power density curves of MEAs fabricated using the pristineNafion 115 membrane (black square: ■), and patterned Nafion membranes withline patterns of 5 μm inwidth, 10 μm in spacing, and 1 μm in height (green triangle:); 110 nm, 90 nm, and 70 nm (red circle: ); 700 nm, 700 nm, and 700 nm (blue

triangle ), respectively. (For interpretation of the references to color in this figurelegend, the reader is referred to the web version of this article.)

Y.-H. Cho et al. / Journal of Membrane Science 467 (2014) 36–40 39

4. Conclusions

In summary, this study demonstrates the effect of an enlargedspecific surface area of the membrane with well-defined linepatterns on the performance of a DMFC, compared with thebaseline pristine Nafion 115 membrane. The well-defined linepatterns with high reliability, ranging in size from several tens ofnanometers to several micrometers, are developed on a Nafionmembrane by soft lithography without collapsing the structure. Inparticular, the current density of the quasi-nano-patterned Nafionmembrane is increased by approximately 35% in a DMFC, com-pared with that of the pristine Nafion 115 membrane. This isattributed to an increase in the effective three-phase boundarycaused by the enlarged specific surface area of the electrolytemembrane in the MEA. On the other hand, the nano-patternedNafion membrane exhibits relatively poor performance at 0.4 Vbecause the agglomerated Pt/C catalyst particles inhibit contactwith the nano-patterned electrolyte membrane, causing insuffi-cient formation of the three-phase boundary. This suggests thatthe performance of DMFCs can be improved further by controllingthe shape and size of the line patterns on the electrolyte mem-brane; more importantly, an appropriate choice of line pattern canreduce effectively the amount of Pt catalyst used while maintain-ing the same performance level.

Acknowledgments

This work was supported by the Institute for Basic Science (IBS)in Korea. Y.-H. C. acknowledges financial support by the PriorityResearch Centre Program (2009-0093814) and Basic ScienceResearch Program (2013R1A1A2061636) through NRF funded bythe Ministry of Education, Republic of Korea.

Appendix A. Supplementary material

Supplementary data associated with this article can be found inthe online version at http://dx.doi.org/10.1016/j.memsci.2014.03.069.

References

[1] B.C.H. Steele, A. Heinzel, Materials for fuel-cell technologies, Nature 414 (2001)345–352.

[2] O. Diat, G. Gebel, Fuel cells: proton channels, Nat. Mater. 7 (2008) 13–14.[3] S. Litster, N. Djilali, Mathematical modelling of ambient air-breathing fuel cells

for portable devices, Electrochim. Acta 52 (2007) 3849–3862.[4] X.G. Yang, C.Y. Wang, Nanostructured tungsten carbide catalysts for polymer

electrolyte fuel cells, Appl. Phys. Lett. 86 (2005) 224104-1–224104-3.[5] S.S. Mao, X. Chen, Selected nanotechnologies for renewable energy applica-

tions, Int. J. Energy Res. 31 (2007) 619–636.

[6] M. Michel, A. Taylor, R. Sekol, P. Podsiadlo, P. Ho, N. Kotov, L. Thompson, High-performance nanostructured membrane electrode assemblies for fuel cellsmade by layer-by-layer assembly of carbon nanocolloids, Adv. Mater. 19(2007) 3859–3864.

[7] K. Sawai, N. Suzuki, Highly active nonplatinum catalyst for air cathodes, J.Electrochem. Soc. 151 (2004) A2132–2137.

[8] D.C. Papageorgopoulos, M. Keijzer, F.A.D. Bruijn, The inclusion of Mo, Nb andTa in Pt and PtRu carbon supported 3electrocatalysts in the quest for improvedCO tolerant PEMFC anodes, Electrochim. Acta 48 (2002) 197–204.

[9] H.A. Gasteiger, S.S. Kocha, B. Sompalli, F.T. Wagner, Activity benchmarks andrequirements for Pt, Pt-alloy, and non-Pt oxygen reduction catalysts forPEMFCs, Appl. Catal. B 56 (2005) 9–35.

[10] Y.H. Cho, B. Choi, Y.H. Cho, H.S. Park, Y.E. Sung, Pd-based PdPt(19:1)/Celectrocatalyst as an electrode in PEM fuel cell, Electrochem. Commun. 9(2007) 378–381.

[11] L. Xiong, A. Manthiram, High performance membrane–electrode assemblieswith ultra-low Pt loading for proton exchange membrane fuel cells, Electro-chim. Acta 50 (2005) 3200–3204.

[12] S.A. Sheppard, S.A. Campbell, J.R. Smith, G.W. Lloyd, T.R. Ralph, F.C. Walsh,Electrochemical and microscopic characterisation of platinum-coated per-fluorosulfonic acid (Nafion 117) materials, Analyst 123 (1998) 1923–1929.

[13] R. O’Hayre, S.J. Lee, S.W. Cha, F.B. Prinz, A sharp peak in the performance ofsputtered platinum fuel cells at ultra-low platinum loading, J. Power Sources109 (2002) 483–493.

[14] M. Prasanna, E.A. Cho, H.J. Kim, T.H. Lim, I.H. Oh, S.A. Hong, Effects of platinumloading on performance of proton-exchange membrane fuel cells usingsurface-modified Nafions membranes, J. Power Sources 160 (2006) 90–96.

[15] S.A. Cho, E.A. Cho, I.H. Oh, H.J. Kim, H.Y. Ha, S.A. Hong, J.B. Ju, Surface modifiedNafions membrane by ion beam bombardment for fuel cell applications, J.Power Sources 155 (2006) 286–290.

[16] P. Millet, T. Alleau, R. Durand, Characterization of membrane–electrodeassemblies for solid polymer electrolyte water electrolysis, J. Appl. Electro-chem. 23 (1993) 322–331.

[17] S.Y. Cha, W.M. Lee, Performance of proton exchange membrane fuel cellelectrodes prepared by direct deposition of ultrathin platinum on themembrane surface, J. Electrochem. Soc. 146 (1999) 4055–4060.

[18] H. Hakan Yildirim, J. te Braake, H. Can Aran, D.F. Stamatialis, M. Wessling,Micro-patterned Nafion membranes for direct methanol fuel cell applications,J. Membr. Sci. 349 (2010) 231–236.

[19] A. Omosebi, R.S. Besser, Electron beam patterned Nafion membranes for DMFCapplications, J. Power Sources 228 (2013) 151–158.

[20] S.Y. Chou, P.R. Krauss, P.J. Renstrom, Imprint lithography with 25-nanometerresolution, Science 272 (1996) 85–87.

[21] E. Woo, J. Huh, Y.G. Jeong, K. Shin, From homogeneous to heterogeneousnucleation of chain molecules under nanoscopic cylindrical confinement,Phys. Rev. Lett. 98 (2007) 136103-1�–136103-4.

[22] K. Shin, S. Obukhov, J.T. Chen, J. Huh, Y. Hwang, S. Mok, P. Dobriyal,P. Thiyagarajan, T.P. Russell, Enhanced mobility of confined polymers, Nat.Mater. 6 (2007) 961–965.

[23] S. Franssila, Introduction to Microfabrication, John Wiley and Sons Ltd.,Chichester, England, 2004.

[24] G. Cao, Nanostructures and Nanomaterials, Imperial College Press, England,2004.

[25] S.J. Shin, J.K. Lee, H.Y. Ha, S.A. Hong, H.S. Chun, I.H. Oh, Effect of the catalyticink preparation method on the performance of polymer electrolyte membranefuel cells, J. Power Sources 106 (2002) 146–152.

[26] Y.H. Cho, H.S. Park, Y.H. Cho, D.S. Jung, H.Y. Park, Y.E. Sung, Effect of platinumamount in carbon supported platinum catalyst on performance of polymerelectrolyte membrane fuel cell, J. Power Sources 172 (2007) 89–93.

[27] Y.H. Cho, S.J. Yoo, Y.H. Cho, H.S. Park, I.S. Park, J.K. Lee, Y.E. Sung, Enhancedperformance and improved interfacial properties of polymer electrolytemembrane fuel cells fabricated using sputter-deposited Pt thin layers, Electro-chim. Acta 53 (2008) 6111–6116.

Fig. 4. Cross-sectional FE-SEM images of membrane–electrode interface prepared by patterned Nafion membranes with line patterns of (a) 5 μm in width, 10 μm in spacing,and 1 μm in height; (b) 700 nm, 700 nm, and 700 nm; (c) 110 nm, 90 nm, and 70 nm, respectively.

Y.-H. Cho et al. / Journal of Membrane Science 467 (2014) 36–4040