Facile preparation of a long-term durable nano- and micro-structured polymer blend membrane for a proton exchange membrane fuel cell

Abstract

The blend of sulfonated poly(ether sulfone)s (PESs) with different degrees of sulfonation is used to prepare a membrane for fuel cell applications. Considering their facile preparation and excellent performance, such sulfonated PES blends constitute good candidates as polymer electrolyte for fuel cell applications. The long-term durability of the blend membrane is improved compared to unblended sulfonated polymer membranes with one degree of sulfonation. The nano-to-micron scale morphology of the polymer blend is investigated by transmission electron microscopy. The long-term stability of the blend membrane is related to its morphology, as confirmed by ultra small angle and small angle neutron scattering (USANS and SANS, respectively) analyses. The blend membrane comprises better-defined interconnected nano-sized ionic phases that enhance proton conductivity and water flow and micro heterogeneous domains, which could be reasons for the long-term durability.

---

Introduction

Nafion-type perfluorosulfonated polymers (PFSAs) are popular materials used as electrolytes in proton exchange membrane fuel cells (PEMFCs). These polymers have high proton conductivity, good mechanical strength, and excellent thermal and chemical properties. However, they are not without problems. Their high oxygen permeability increases the formation of hydrogen peroxide radicals that cause polymer degradation.^1,2 In addition, the conductivity of PFSAs is low under low relative humidity and elevated temperature conditions. Moreover, they are very expensive. To overcome these disadvantages, low cost hydrocarbon-based polymer electrolyte membranes such as sulfonated poly(ether sulfone)s have been investigated.^3–7

Sulfonated poly(ether sulfone)s can be easily synthesized by the condensation reactions of di-fluoro and di-hydroxyl compounds. They exhibit high oxygen barriers, reasonable proton conductivity, and high mechanical strength.^1,7 Although these properties suggest them as possible candidates to replace Nafion-type PFSAs, the polymer electrolytes do not exhibit good long-term durability, and dissolve or decompose during fuel cell operation. To improve their long-term stability, we cross-linked the polymers,^8–12 which resulted in improved dimensional stability and water insolubility. However, the cross-linking process was very complicated and difficult to control. Another approach to improve dimensional stability and water insolubility involves blending the sulfonated polymer with several tough polymers.^13–17 However, this method was also imperfect because the polymers were not sufficiently compatible or able to maintain stability during fuel cell operations.

Most sulfonated polymers vary in their molecular weight distributions and degrees of sulfonation. Even with strict control of polymerization and sulfonation, low molecular weight polymers that have unexpectedly high degrees of sulfonation are obtained unintentionally. These polymers dissolve during fuel cell operation, causing membrane thinning and pinhole formation. In this report, we describe a novel polymer blend system that mitigates these problems. We blended several sulfonated poly(ether sulfone)s having sulfonation degrees of 10, 20, 30 and 40%. In this case, even though there are highly sulfonated low molecular weight polymers, which have relatively low stability toward water, they are held tightly through hydrogen bonding with other dimensionally stable polymers that are high in molecular weight but low in degree of sulfonation. These results in the formation of highly durable polymer electrolyte blend membranes.

When miscible polymers are well blended, the blends show properties that are different from the constituent polymers. Our blend system also shows totally different morphological and long-term durability properties compared to the pristine polymers that comprise it. We characterized the morphology of the blend and pristine polymer electrolytes by transmission electron microscopy (TEM), ultra small neutron scattering (USANS) at micrometer scale, and small angle neutron scattering (SANS) at nanometer scale. Finally, membrane electrode assemblies (MEAs) using blend and pristine polymer membranes were tested for long-term durability.

Results and discussion

Polymer characterization

A blend membrane of sulfonated polymers with different degrees of sulfonation, PES 10, 20, 30, and 40, in a 0.5 : 1 : 1 : 1 weight ratio was prepared in DMAc (Fig. 1). Individual membranes of PES 30 and 40 were also prepared from DMAc solution in the same manner. Basic membrane properties for PES 30, 40, and the blend are shown in Table 1. Also, basic properties of PES 10, 20, 30, and 40 are listed in ESI Table S1.† The mechanical properties of the blend membrane were similar to those of PES 30 and 40 under low and high relative humidities. The ion exchange capacity (IEC) of the blend membrane was 1.14 meq g⁻¹, the lowest value among the three membranes. In case of water uptake (WU), PES 40 has the highest value because of its high sulfonation degree.

Fig. 1 Chemical structure of sulfonated PES copolymers: PES 10: n/m = 1/9; PES 20: n/m = 2/8; PES 30: n/m = 3/7; and PES 40: n/m = 4/6.

Table 1 Physical properties of PES 30, 40, and blend membranes

	IEC^a [meq g^−1]	WU^b [%]	Mw^c [g mol^−1]	Tg [°C]	@80 °C, 30% RH		@80 °C, 90% RH
					Tensile strength [MPa]	Elongation [%]	Tensile strength [MPa]	Elongation [%]
a The ion exchange capacity (IEC) was determined by titration.b Water uptake (WU) was measured at room temperature.c GPC measurement.
PES 30	1.36	10	1.2 × 10^5	220	53	91	33	164
PES 40	1.64	24	0.9 × 10^5	202	45	100	28	184
Blend	1.14	12	1.3 × 10^5		46	99	35	157

---

The morphologies of the polymer membranes were characterized by TEM (Fig. 2). The membranes were stained in aqueous lead acetate solution for the TEM analysis. The dark areas indicate the hydrophilic domains containing sulfonic acid groups and the bright areas reveal the hydrophobic areas of the membrane. The morphological features of PES 30 and 40 were unclear, and immaturely developed phase separation was observed, similar to that of many random copolymers.¹⁸ In contrast, the blend membrane clearly showed well-defined phase separation similar to that observed for Nafion membranes. That kind of morphology is observed sulfonated poly(arylene ether sulfone) block copolymers.^19–21 The phenomenon occurs because each sulfonated polymer with its different sulfonation degree plays a strong role during the blending process. PES 10 and 40 formed opaque solution when they were dissolved in DMAc, which means they are not miscible. However, when PES 20 and 30 were added to the DMAc solution of PES 10 and 40, we were able to obtain transparent solution. PES 20 and 30 work as a kind of compatibilizer. The hydrophilic regions consist of polymers with sulfonated group, whereas the hydrophobic regions contain non-sulfonated moiety (Fig. 3). Each region stands out because its property is remarkably different. Polymers with different degrees of sulfonation blend well because of the interactions of their sulfonic acid functional groups. Nevertheless, the characteristic properties of each polymer due to its low or high degree of sulfonation are retained during the blending process and are shown prominently in the TEM images.

Fig. 2 TEM images of lead exchanged membrane (a) PES 30, (b) PES 40, (c) blend, and (d) N212.

Fig. 3 Schematic diagram of the blend membrane that has nano-scale hydrophilic and hydrophobic phase separation.

The TGA curves of the three membranes are shown in Fig. 4. Dehydration from the PES 30, 40, and blend membranes occurred up to 300 °C. However, the blend membrane exhibited a pattern different from the PES 30 and 40 membranes. Water loss from the blend membrane occurred in the two temperature regions around 150 and 250 °C. The latter can be ascribed to very tightly bound water²² in the polymer blend membrane. The water molecule was not easily removed by drying at 60 °C under reduced pressure. The tightly bound water in the blend is also observed in the SANS profiles, which will be discussed later. This is why the PES 30 and blend membranes exhibit similar IEC and water uptake (WU) properties, even though water evaporation from the blend membrane is much larger than that from PES 30 in Fig. 4a. In addition, the amount of water evaporation from the blend was very similar to that from the PES 40 membrane as shown in TGA analysis.

Fig. 4 Thermogravimetric analyses of polymer and blend membranes. (a) TGA curves and (b) differential thermogravimetry (DTG) curves of the PES 30, 40, and blend membranes. The arrow indicates water bound to hydrophilic domains.

USANS and SANS analysis

Dry PES 30 and 40 membranes showed only incoherent scattering while the dry blend membrane showed a weak hump around Q ≈ 0.09 Å⁻¹ (Fig. 5a). This hump indicates the presence of the residual water bound in the hydrophilic domain of the dry blend that causes the contrast with the hydrophobic phase, since the small angle scatterings can be observed only when there is a difference in the scattering length density (SLD) between two phases.²³ When the membranes were soaked (i.e., saturated) with D₂O, the scatterings appear over four orders of magnitude, from nanometer to micrometer, and fifteen orders in absolute intensity (Fig. 5b), which reveal several structural features. The structure peaks from all the samples, which represent the distance (d_ion) between the ionic channels (i.e., the hydrophilic domains), appeared as shown in Fig. 5c in linear x–y scale. The scattering peak of ionic channel in the blend shifted to lower Q, from ∼0.09 Å⁻¹ for the dry blend to 0.07 Å⁻¹ for the wet blend, which corresponds to the increase of an averaged correlation distance d_ion (=2π/Q) from 70 ± 1 Å to 90 ± 2 Å. This indicates the expansion of ionic channels by as much as ∼20 Å (i.e., approximately 29% swelling of the hydrophilic domain). The peak intensity is proportional only to the amount of water since the neutron contrast between the hydrophilic and the hydrophobic domains are the same for the saturated membranes. Thus, the narrower width and higher intensity of the blend around Q ≈ 0.07 Å⁻¹ (Fig. 5c), indicates that the blend has better defined ionic channels in nano size and takes up more water than the other membranes. Another distinctive feature of the wet blend is the Guinier scattering at Q ≈ 5 × 10⁻⁴ to 1 × 10⁻³ Å⁻¹ region that represents a large heterogeneous micron size domain. The generalized Guinier power-law fit²⁴ including very low angle scattering shows a radius of gyration R_g = 3187 Å corresponding to a diameter of 0.821 μm, assuming a spherical domain. PES 30 and 40 membranes do not show such Guinier scatterings. The extraordinary characteristic of the blend membrane provides excellent electrochemical resistance which will be discussed in fuel cell operation section.

Fig. 5 USANS–SANS profiles of (a) dry membranes showing residual water bounded in the blend, (b) membranes wetted with D₂O in double logarithm scale showing structural features of both micro domain and nano ionic channel as well as sharp interfaces between the hydrophobic and hydrophilic phases, and (c) linear scale showing the distance between ionic channels. The unit of total cross section is cm⁻¹ sr⁻¹.

The blend, PES 30, and PES 40 showed the slope of −4 at high Q region (Fig. 5b) that indicates smooth, sharp interfaces between the hydrophobic matrices and hydrophilic ionic channels with water. Hydrated sulfonated poly(etherketone) showed the same sharp interface.²⁵

Fuel cell operation

Proton conductivities for the PES 30, 40, and blend membranes were determined as a function of relative humidity at 65 and 80 °C (Fig. 6a and b). The proton conductivities of all the membranes decreased as the relative humidity decreased. Proton conductivity increased with increasing IEC values (Table 1), similarly to universal sulfonated polymer membranes.^5,26 However, a distinctive behaviour was observed. The proton conductivity of the blend membrane was similar to that of PES 40 under all relative humidity and temperature conditions, even though the blend membrane exhibited an IEC similar to PES 30. This can be well explained with the TGA and small angle neutron scattering results. Since the proton conductivity is believed to be dominated by the structure of the proton passages (i.e., water pathways) in the membrane, the well-defined, interconnected ionic channels and high water uptake capability (as shown in the small angle scattering profiles) can explain the high proton conductivity of the blend. There are more water molecules in the blend membrane, which is why the blend membrane shows higher proton conductivity than PES 30.

Fig. 6 Proton conductivity of the prepared membranes as a function of relative humidity at (a) 65 °C and (b) 80 °C. Fuel cell cycling performance: (c) OCV decays for PES 30, PES 40, and blend membranes during cycling, and (d) hydrogen cross-over currents for PES 30, 40, and blend membranes.

The electrochemical evaluation of the blend membrane directly shows an improvement in the long-term durability of PEMFC performance. All the MEAs (membrane electrode assemblies) using the PES 30, PES 40, and blend membranes were fabricated with the same components and operated under identical conditions. Weidner et al. reported that the open circuit voltage (OCV) is decreased by membrane thinning or micro-hole formation during fuel cell operation.¹ Therefore, comparing the rate of OCV decrease can be good way to evaluate the chemical and mechanical stabilities of the membranes. Each MEA was operated according to the potential cycle displayed in ESI Fig. S1† until the OCV decreased below 0.9 V. The potential cycle contained low, middle, and high voltage operations, which generate wet conditions at 0.4 V and dry conditions at 0.9 V. The cycle operation accelerates the fatigue of the fuel cell membrane. Various electrochemical characteristics as a function of cycles are listed in ESI Table S2,† as well as in ESI Fig. S2–S4.† Ohmic resistance increased and electroactive surface area (EAS) decreased as potential cycles repeated. Also, the OCVs for all the cells were reduced with repeating potential cycles. The single cell with the blend membrane failed after 48 000 cycles, whereas cells using PES 30 and 40 membranes failed after 32 000 and 16 000 cycles, respectively (Fig. 6c). OCV decrease is strongly related to the hydrogen cross-over current density. The results for hydrogen cross-over current density assessment are shown in Fig. 6d. The initial hydrogen cross-over current densities were similar for the three membranes. The PES 30, PES 40, and blend membranes were quite stable until 13 000, 24 000, and 40 000 cycles, respectively. However, at the point of OCV failure, the hydrogen cross-over current density was greatly increased. At the end point, the hydrogen cross-current exceeded 15 mA cm⁻².

In the potential cycling tests, the blend membrane exhibited the most durable properties, demonstrating good chemical stability with reasonable performance. According to the TEM analysis and USANS–SANS measurements, the blend membrane showed distinct hydrophilic and hydrophobic segregation at nano- and micro-scales. During the blending process, the sulfonated and non-sulfonated regions of each polymer (PES 10, 20, 30, and 40) are gathered separately; the solid and rigid matrix prevents severe swelling of the hydrophilic domains and results in good dimensional stability for fuel cell voltage cycle operation. The long-term durability of the blend may be due to the heterogeneous structure observed at microscale as shown in Fig. 5b. The micro heterogeneity of the blend may result from the hydrophobic matrix composed of rigid chains that may be responsible for the long life cycle, while the nanosized hydrophilic domains that are responsible for water and proton flow are interconnected by hydrogen bonding between polymers of low and high sulfonation degrees. The microscale heterogeneity were not observed in the pristine membranes. Considering the facile preparation of blend membranes, such sulfonated poly(ether sulfone) blends constitute good candidates for polymer electrolyte membranes for fuel cell applications.

Experimental

Materials

Poly(arylene ether sulfone)s with sulfonation degrees of 10, 20, 30, and 40% (PES 10, PES 20, PES 30, and PES 40) were purchased from Yanjin Technology Co., Ltd., China, and were used as received. The structures are shown in Fig. 1. The numbers n and m indicate the molar ratio of hydrophilic segments (with sulfonic acid groups) and hydrophobic segments (without sulfonic acid groups) in the polymer backbone, respectively (PES 10: n/m = 1/9; PES 20: n/m = 2/8; PES 30: n/m = 3/7; and PES 40: n/m = 4/6). N,N-Dimethylacetamide (DMAc) was obtained from Aldrich Chemical Co. and used without further purification.

Membrane fabrication

Each polymer solution was prepared by dissolving a sulfonated poly(arylene ether sulfone) in DMAc (15 wt%) at room temperature. For the blend membrane, the weight ratio of PES 10, 20, 30, and 40 was controlled at 0.5 : 1 : 1 : 1, respectively. Each sulfonated polymer (PES 10, 20, 30, 40) or blend polymer membrane was prepared by the following method. The polymer solution was poured onto a glass plate and flattened by a homemade spin-coater. The solvent was dried at 40 °C for 4 h and then at 60 °C under vacuum for 24 h. The obtained membrane was removed from the glass surface by immersion in a water bath. The membrane was acidified by soaking in 10 wt% HCl aqueous solution at 60 °C for 1 h and then washed with deionized water several times at 60 °C.

Ion exchange capacity (IEC), water uptake, and proton conductivity

The IECs of the membranes were determined by the acid–base titration method. A dried membrane (acid form) was immersed in brine for 24 h with stirring. NaOH solution (0.01 M) was used to back titrate to determine the IEC using phenolphthalein as an indicator.²⁷ Water uptake in a membrane was determined by measuring its change in weight before and after immersion in deionized water. In this study, the membrane samples were soaked in deionized water at 30 °C for 24 h after drying under reduced pressure at 60 °C for 24 h. Each sample was weighed immediately after removing the water on the membrane surface.

The proton conductivity of each membrane sample (1 × 4 cm²) was measured using a four-probe-conductivity cell in a chamber with controlled temperature and relative humidity via an IM6 electrochemical workstation (ZAHNER-Elektrik GmbH & Co. KG). The proton conductivity σ was calculated using eqn (1). L is the distance between the two electrodes, R is the resistance of the membrane, and A is the cross-sectional area (thickness × width) of the membrane.²⁸ Impedance spectra were recorded from 10 MHz to 1 Hz.

σ = L/(R × A) (1)

Transmission electron microscopy (TEM)

To observe phase separation, the acid forms of the membranes were treated with 1 M aqueous lead acetate solution for 24 h, rinsed with deionized water several times, and dried in a vacuum oven for 24 h. The membranes were embedded in epoxy, cross-sectioned to 90 nm with a cryo-ultramicrotome system, and examined with a Cryo-TEM (FEI Tecnai G2 F20).

Thermal analysis

The glass transition temperatures (T_g) were obtained using a Q2000 differential scanning calorimeter (TA Instruments). Scans were conducted under nitrogen atmosphere at a heating rate of 10 °C min⁻¹. Thermogravimetric analysis (TGA) and TGA-mass spectrometry (TGA-MS) were performed under nitrogen atmosphere at a heating rate of 10 °C min⁻¹ using a Q50 analyzer (TA Instruments).

Ultra small angle neutron scattering (USANS) and small angle neutron scattering (SANS) measurements

Membranes were attached in several layers to increase the scattering intensity. Transmission was more than 80% from both SANS and USANS measurements, which assures no multiple scattering. SANS was measured by both the 40M and 18M instruments at the cold neutron guide hall of HANARO, the Korean Research Reactor installed at the Korea Atomic Energy Research Institute (KAERI), Daejeon, Korea. All measurements were corrected to the absolute scale (i.e., total scattering cross-section) by normalizing the sample thickness, transmission, dark current, and detector sensitivity. The scattering vector Q, defined as, Q = 4π sin θ/λ where 2θ is the scattering angle, covered the Q range from 4 × 10⁻³ to 7 × 10⁻¹ Å⁻¹.

USANS was measured with the KIST-USANS instrument at the HANARO cold neutron facility. The instrument consists of a pair of modified Bonse–Hart–Agamalian channel-cut crystals,^29,30 monochromator, and analyzer, with a wavelength of 4 Å. The minimum Q-resolution can reach 2 × 10⁻⁵ Å⁻¹. The combination of the USANS and SANS covers a wide Q range, from 2 × 10⁻⁵ to 7 × 10⁻¹ Å⁻¹. The measured one-dimensional-smeared USANS intensity was converted to the pinhole SANS intensity by multiplying by Δq_v/Q, where Δq_v is the vertical divergence of the one-dimensional detector. The data reduction was performed with the data reduction package²⁵ modified for the KIST-USANS. The dry membranes were thoroughly wet by placing the membranes in D₂O water to increase the neutron scattering contrast.

Measurement of fuel cell performance

The MEAs (membrane electrode assemblies) were prepared by the catalyst-coated membrane (CCM) method.³¹ A catalyst/ionomer slurry was prepared by mixing a carbon-supported platinum catalyst (Pt/C, 45.5 wt% Pt, Tanaka Kikinzoku Kogyo K.K), 5 wt% Nafion ionomer dispersion (EW 1100, Dupont Inc.), isopropyl alcohol (Baker Analyzed HPLC Reagent), and deionized water. The slurry was sprayed on the membrane using an automated spraying machine. The weight ratio of the Pt/C catalyst and Nafion ionomer was controlled at 80 : 20. The active electrode area was 25 cm² with a platinum loading of 0.4 mg cm⁻¹ for both the anode and cathode sides. The cell temperature was 65 °C, and fully humidified hydrogen and air were supplied after passing through a bubble humidifier under ambient pressure. The stoichiometries of hydrogen and air were controlled at 1.5 and 2.0, respectively. Before potential cycling, the single cell was activated at a constant voltage (0.45 V) for 24 h. Potential cycling tests were carried out for three membranes (PES 30, PES 40, and the blend). The experiments were terminated when the open circuit voltage (OCV) was lower than 0.9 V. The cell potential was cycled at 0.9, 0.6, and 0.4 V for 1 min each. The voltage profile is shown in ESI Fig. S1.† Electrochemical analysis, which included I–V curves, electrochemical impedance spectroscopy (EIS), cyclic voltammetry (CV), and linear sweep voltammetry (LSV), was performed every 2000 cycles. These electrochemical parameters were obtained according to methods described earlier.³²

Conclusions

Good chemical and mechanical properties with reasonable proton conducting material was obtained by blending series of sulfonated poly(ether sulfone)s. The blend membrane exhibited clear hydrophobic/hydrophilic phase separation and high water affinity compared to pristine sulfonated polymers that have single sulfonation degree. USANS and SANS measurement showed the blend has two characteristic features, the ionic channels in nanoscale and heterogeneity in microscale. The nano ionic channels, the pathway of water molecules and protons, are better phase-separated and interconnected each other with high specific surface area, which can enhance proton conductivity. Heterogeneous domains that were not observed on other membranes can be a reason of the long term durability. The unique property of the membrane produces excellent durability for fuel cell cycling operation.

Acknowledgements

This work was supported by the projects “KIST Institutional Program”. Also, it was partially supported by the Korea Institute of Energy Technology Evaluation and Planning (KETEP) and the Ministry of Trade, Industry & Energy (MOTIE) of the Republic of Korea (No. 20153010031920).

Notes and references

1. V. A. Sethuraman, J. W. Weidner, A. T. Haug and L. V. Protsailo, J. Electrochem. Soc., 2008, 155, B119–B214.
2. M. Inaba, T. Kinumoto, M. Kiriake, R. Umebayashi, A. Tasaka and Z. Ogumi, Electrochim. Acta, 2006, 51, 5746–5753.
3. N. Li, D. S. Hwang, S. Y. Lee, Y.-L. Liu, Y. M. Lee and M. D. Guiver, Macromolecules, 2011, 44, 4901–4910.
4. H.-S. Lee, O. Lane and J. E. McGrath, J. Power Sources, 2010, 195, 1772–1778.
5. F. Wang, M. Hickner, Y. S. Kim, T. A. Zawodzinski and J. E. McGrath, J. Membr. Sci., 2002, 197, 231–242.
6. C. Wang, N. Li, D. W. Shin, S. Y. Lee, N. R. Kang, Y. M. Lee and M. D. Guiver, Macromolecules, 2011, 44, 7296–7306.
7. H.-S. Lee, A. Roy, O. Lane, S. Dunn and J. E. McGrath, Polymer, 2008, 49, 715–723.
8. Y. S. Oh, H. J. Lee, M. Yoo, H. J. Kim, J. Han, K. Kim, J. D. Hong and T. H. Kim, Chem. Commun., 2008, 208–2030,  10.1039/b800064f.
9. B. G. Kim, H.-J. Kim, J. H. Jang, E. A. Cho, D. Henkensmeier, S.-K. Kim, I.-H. Oh, S.-A. Hong and T.-H. Lim, Polym. Int., 2011, 60, 685–691.
10. K. B. Heo, H. J. Lee, H. J. Kim, B. S. Kim, S. Y. Lee, E. Cho, I. H. Oh, S. A. Hong and T. H. Lim, J. Power Sources, 2007, 172, 215–219.
11. Y.-S. Oh, H.-J. Lee, M. Yoo, H.-J. Kim, J. Han and T.-H. Kim, J. Membr. Sci., 2008, 323, 309–315.
12. R. L. Thankamony, M.-G. Lee, K. Kim, J.-D. Hong, T.-H. Kim, H.-J. Lee, H.-J. Kim, S. Nam and Y.-B. Lim, Macromol. Res., 2010, 18, 992–1000.
13. Y. Fu, A. Manthiram and M. D. Guiver, Electrochem. Commun., 2006, 8, 1386–1390.
14. N. Y. Arnett, W. L. Harrison, A. S. Badami, A. Roy, O. Lane, F. Cromer, L. Dong and J. E. McGrath, J. Power Sources, 2007, 172, 20–29.
15. S. Swier, M. T. Shaw and R. A. Weiss, J. Membr. Sci., 2006, 270, 22–31.
16. D. H. Kim, J. Choi, Y. T. Hong and S. C. Kim, J. Membr. Sci., 2007, 299, 19–27.
17. S. Swier, V. Ramani, J. Fenton, H. Kunz, M. Shaw and R. Weiss, J. Membr. Sci., 2005, 256, 122–133.
18. Y. A. Elabd and M. A. Hickner, Macromolecules, 2011, 44, 1–11.
19. H. Ghassemi, J. E. McGrath and T. A. Zawodzinski, Polymer, 2006, 47, 4132–4139.
20. Y. Li, A. Roy, A. S. Badami, M. Hill, J. Yang, S. Dunn and J. E. McGrath, J. Power Sources, 2007, 172, 30–38.
21. B. Bae, K. Miyatake and M. Watanabe, Macromolecules, 2010, 43, 2684–2691.
22. M. K. Mistry, N. R. Choudhury, N. K. Dutta, R. Knott, Z. Shi and S. Holdcroft, Chem. Mater., 2008, 20, 6857–6870.
23. M.-H. Kim, C. J. Glinka, S. A. Grot and W. G. Grot, Macromolecules, 2006, 39, 4775–4787.
24. B. Hammouda, J. Appl. Crystallogr., 2010, 43, 716–719.
25. S. R. Kline, J. Appl. Crystallogr., 2006, 39, 895–900.
26. K. Miyatake, H. Zhou, H. Uchida and M. Watanabe, Chem. Commun., 2003, 368–369,  10.1039/b210296j.
27. P. Krishnan, J.-S. Park and C.-S. Kim, J. Membr. Sci., 2006, 279, 220–229.
28. H.-J. Lee, J. Choi, J. Y. Han, H.-J. Kim, Y.-E. Sung, H. Kim, D. Henkensmeier, E. Ae Cho, J. H. Jang and S. J. Yoo, Polym. Bull., 2013, 70, 2619–2631.
29. M. Agmalian, G. D. Wignall and R. Triolo, J. Appl. Crystallogr., 1997, 30, 345–352.
30. J. G. Barker, C. J. Glinka, J. J. Moyer, M. H. Kim, A. R. Drews and M. Agamalian, J. Appl. Crystallogr., 2005, 38, 1004–1011.
31. N. N. Krishnan, H.-J. Lee, H.-J. Kim, J.-Y. Kim, I. Hwang, J. H. Jang, E. A. Cho, S.-K. Kim, D. Henkensmeier, S.-A. Hong and T.-H. Lim, Eur. Polym. J., 2010, 46, 1633–1641.
32. Y. Y. Jo, E. A. Cho, J. H. Kim, T.-H. Lim, I.-H. Oh, J. H. Jang and H.-J. Kim, Int. J. Hydrogen Energy, 2010, 35, 13118–13124.

---

Footnote

† Electronic supplementary information (ESI) available: Electrochemical characterization for fuel cell cycling operation. See DOI: 10.1039/c6ra05624e

---