Ternary hybrid (SPEEK/SPVdF-HFP/GO) based membrane electrolyte for the applications of fuel cells: profile of improved mechanical strength, thermal stability and proton conductivity

Abstract

Ternary hybrid membranes composed of sulfonated (poly ether ether ketone) (SPEEK), sulfonated polyvinylidene fluoride-co-hexafluoropropylene (SPVdF-HFP) and 1, 3, 5 or 7 wt% graphene oxide (GO) were fabricated using a facile solution casting method. The reinforcement due to the existence of SPVdF-HFP and GO afforded good mechanical and thermal stabilities to the hybrid membranes, which was confirmed by dynamic mechanical analysis (DMA) and thermogravimetric analysis (TGA). The surface morphological properties and roughness of the hybrid membranes were scrutinized using field emission scanning electron microscopy (FE-SEM) and atomic force microscopy (AFM), whereas the clenched structure with even intercalation of GO sheets in the polymer matrix was observed. The temperature dependent changes in proton conductivity, as well as the mass, length, and thickness of membranes were measured; the ternary hybrid exhibited more significant changes compared to other membranes. The chemical structure, intermolecular bond stretching, structural reorganization, and crystallinity of the membranes were analyzed using proton nuclear magnetic resonance spectroscopy (¹H-NMR), Fourier transform infrared spectroscopy (FT-IR), differential scanning calorimetry (DSC), and X-ray diffraction (XRD) instrumentation techniques. In the ternary hybrid membranes, SPVdF-HFP increased the per cluster volume of SO₃H groups and GO increased the number of directional hydrogen bonds (H-bonds), which collectively provided good impact in proton conductivity. At 90 °C, the peak proton conductivity attained by the SPEEK was 68 mS cm⁻¹, while that of the ternary hybrid was 122 mS cm⁻¹, 1.7 times better conductivity. Furthermore, the ternary hybrids exhibited much lower H₂ permeability compared to that of SPEEK and Nafion-117 membranes.

---

1. Introduction

Hydrogen is a renewable energy source abundant in nature, while proton exchange membrane fuel cells (PEMFCs) are considered to have considerable potential for recognizing hydrogen economy.^1–4 The substantial benefits of PEMFC, such as energy conversion in a single step and the low production of green-house gases make them effectual alternatives to the traditional internal combustion engines.^5–7 In PEMFC, the proton exchange membrane (PEM) is an inherent module that performs as the proton conductor to conduct the protons from the anode to the cathode, and as the fuel barrier to prevent the contact of fuel with oxidants.⁸ Nafion membrane (the sulfonic acid (SO₃H) group terminated perfluorovinyl ether unit-grafted teflon chains) has been considered as the standard benchmark for PEMFC applications, due to its good oxidative resistance properties, high dimensional stability, and high proton conductivity. However, high cost remains as the main limitation in such systems.^9–12 Poly ether ether ketone (PEEK) is a comparatively low cost aromatic hydrocarbon polymer with good mechanical and thermal properties. The aromatic chains are very sensitive to electrophilic substitution reactions and hence, desirable proton conductivity can be achieved easily by varying the sulfonation quantity.^13–17 Therefore, this polymer has been extensively exploited for PEM applications. However, at high degrees of sulfonation, the hydrophobic backbones of the polymer bear vast numbers of hydrophilic groups, which significantly weaken the co-connectivity of backbone chains and consequently generate wide gaps in between the chain structures. This leads to the generation of pinholes and tears in the membrane during the casting process.^18–20 Consequently, the mechanical strength of the membranes and their electrochemical selectivity toward H⁺ ions are diminished.

PVdF-HFP is a partially fluorinated copolymer, widely scrutinized for both physical reinforcements of membranes, and aqueous, acid and oxidant resistant applications, owing to its immense mechanical strength, considerable hydrophobic behavior and oxidative resistant properties.^2,21,22 PVdF-HFP blending can effectively alter the thermo-mechanical and electrochemical properties of PEM by preventing the generation of pinholes and tears. The peculiarity of PVdF-HFP is its partially fluorinated fragments that enhance the compatibility of PEM with binders like Nafion and PTFE, leading to the fabrication of highly durable MEA.²³ Although PVdF-HFP has been exploited as a potential reinforcing material, the inadequate solubility and improper miscibility with aromatic SPEEK, due to the lack of interfacial interactions, impede the intention of blending. The direct grafting of sulfonic acid groups in the backbones of PVdF-HFP has been considered as a better approach to alleviate the aforementioned constraints, which can lead to the threshold synergism of PVdF-HFP with SPEEK. A PVdF-HFP can perform two significant roles: (i) SO₃H moiety grafting backbones to increase the per cluster volume ratio of SO₃H in SPEEK and (ii) as a mixing polymer to heighten the mechanical stiffness and thermal properties of SPEEK. Again, the wt% of SPVdF-HFP in the membrane is the problem that effectually influences the physiochemical, electrochemical, and thermomechanical properties of the hybrid membranes. Recently, Bagheri et al. modified the SPEEK membrane by blending it with different amounts of SPVdF-HFP, i.e., 10, 15, 20 and 25 wt%, for use as electrolytes in fuel cell applications. Based on their report, 10 wt% of SPVdF-HFP provided the highest water uptake (70%), ion exchange capacity (1.40), and proton conductivity (68 mS cm⁻¹), compared to those of the other membranes; this membrane also retained good dimensional stability and considerable oxidative stability. Thus, blending 10 wt% of SPVdF-HFP into the SPEEK matrix assisted in fabricating electrolyte membranes with promising performances for fuel cell applications. On the other hand, functionalized carbon nanomaterial saturated SPEEK composites have received considerable attention as a PEM for PEMFC application, owing to their good compatibility with protons, their good thermal stability, physiochemical properties, Young's modulus, and surface treatment characteristics. Gahlot et al. incorporated and electrically aligned f-CNT within the SPEEK matrix. They described that aligning f-CNT in the SPEEK matrix effectively improve the proton conduction, water uptake, and mechanical properties of composite membranes.¹⁵ Rambabu et al. utilized pSSA functionalized multi-wall carbon nanotubes (pSSA-MWCNTs) as the filler for the SPEEK matrix for the fabrication of composite membranes. With the intercalation of pSSA-MWCNTs, the tensile stiffness and elongation of SPEEK have been improved, owing to strong electrostatic and π–π interactions.²⁴

Graphene is a single, two-dimensional layered material composed of hexagonally arranged sp² hybrid carbons, and is more specifically associated with exceptional extended surfaces and good electrical, mechanical, and thermal properties.^25–27 While conjugated sp² hybrid networks of graphene need to be functionalized/oxidized to tailor the material as an electronic insulator and protonic conductor, this could facilitate their use as fillers in PEMs. The blending of GO as an ingredient in PEM has been endorsed to amplify thermal properties and mechanical strength significantly.²⁸ However, the main issue remaining is proton conductivity. Kumar et al. reported that the presence of GO in the Nafion matrix improves the proton conductivity by increasing the number of hydrophilic domains and H-bonds.²⁹ He et al. revealed that well-defined microstructures and well-connected proton transport paths in GO/SPI composite membranes can typically be obtained by strong interfacial interactions, such as π–π interactions and H-bonds.³⁰ Thus, we proposed that nano-fluidic channels of GO, such as hydroxyl, epoxy, and carboxylic groups formed directional H-bond networks with the SO₃H groups of SPEEK and SPVdF-HFP, facilitating the facile and smooth travel of protons throughout the membranes. Also, GO adjusts the aromatic chains of SPEEK through π–π interactions, thereby enhancing the thermal stability and mechanical strength of the hybrid membrane.

The intention of the current work is to both enhance the proton conductivity and reinforce the aromatic chains of SPEEK through blending SPVdF-HFP and GO. The fabricated membranes such as SPEEK, SPEEK/SPVdF-HFP (10 wt%), SPEEK/SPVdF-HFP (10 wt%)/GO (1 wt%), SPEEK/SPVdF-HFP (10 wt%)/GO (3 wt%), SPEEK/SPVdF-HFP (10 wt%)/GO (5 wt%), and SPEEK/SPVdF-HFP (10 wt%)/GO (7 wt%) have been designated as SP, SPSPV, SPSPVG-1, SPSPVG-3, SPSPVG-5, and SPSPVG-7, respectively. The influences of blending on the mechanical, thermal, and structural properties of the membrane were scrutinized with the corresponding instrumentation techniques. Membranes were effectually studied for their water uptake, dimensional stability, IEC, oxidative stability, and H₂ gas permeability. Proton conductivities of membranes under both hydrated and anhydrous conditions were also estimated.

2. Experimental section

2.1. Materials

Poly(ether ether ketone) (PEEK) was purchased from the Victrex Company. Potassium permanganate (KMnO₄) was supplied by Samchun Chemicals, South Korea. Hydrogen peroxide (H₂O₂), concentrated sulfuric acid (H₂SO₄-95%), hydrochloric acid (HCl) and 1,2-dichloroethane were received from Daejung Chemicals, South Korea. Natural graphite powder and sodium nitrate (NaNO₃) were obtained from the Alfa Aesar Company. Poly vinylidene fluoride-co-hexafluoropropylene (PVdF-HFP) and chlorosulfonic acid (HSO₃Cl) were procured from Aldrich chemicals. N,N-dimethylformamide (DMF) was supplied by Duksan Reagents, South Korea.

2.2. PEEK sulfonation

PEEK was directly sulfonated by electrophilic substitution reaction with conc. H₂SO₄.^15,31–33 Before performing the reaction, the PEEK powder was dried at 70 °C for 5 h to remove the moisture content. The 28 g of dry PEEK powder was dissolved in 300 mL of conc. H₂SO₄ at 40 °C using a motor stirrer (300 rpm speed). The solution temperature was then elevated to 60 °C, and the reaction was carried out for the designated time. The sulfonation was terminated by precipitating the polymer solution in an excess of cold water. The formed precipitate was repeatedly washed using deionized water (DI water) to obtain a neutral pH, and the sulfonated polymer was collected and dried. The degree of sulfonation (DS) was 65% for SPEEK, and was evaluated by using ¹H NMR spectra. The degree of sulfonation of polymer depends on the concentration and volume of acids, reaction time, and temperature.

2.3. PVdF-HFP sulfonation

Sulfonation of PVdF-HFP was performed with HSO₃Cl as the precursor to the SO₃H moiety.^22,34–37 Prior to the reaction, the PVdF-HFP pellets were dried at 70 °C for 5 h to remove the moisture content. Dry pellets (∼3 g) were mixed with the 20 mL of HSO₃Cl, and the mixture was stirred at 60 °C for 7 h. The sulfonated pellets were washed using methanol, 1,2-dichloroethane, and DI water. The degree of sulfonation of the polymer was 31%, which was determined by acid–base titration method.

2.4. Graphite (Gr) oxidation

Graphite oxidation was performed based on the modified Hummer's method.^38,39 In brief, 4 g of natural graphite powder and 2 g of NaNO₃ were first added to the solution of 70 mL conc. H₂SO₄. To this mixture, 12 g of KMnO₄ were added, while maintaining the mixture in an ice-bath to prohibit augmentation of solution temperature. After this addition, the reaction mixture was stirred vigorously for 1 h at 40 °C, then, 150 mL DI water were added to the above-mentioned mixture and the reaction temperature was raised to 100 °C, while maintaining the reaction for 5 h. Then, a 30% solution of H₂O₂ was added to this mixture, followed by the addition of excess DI water. Finally, the solid GO was washed, centrifuged and dried. The solid sheets of GO were exfoliated through ultrasonication in ethanol before further characterization.

2.5. Membrane preparation

The hybrid membranes were fabricated through the solution casting method, in which the solution of DMF (10 wt%) containing SPEEK (90%) and SPVdF-HFP (10%) was stirred for 12 h at 60 °C. Then, 1, 3 5 or 7 wt% of GO was mixed with the above blend, and the mixture was ultrasonicated until a homogeneous dispersion of GO was attained. The mixture was cast over a clean glass matrix and dried at 70 °C for 12 h, followed by 120 °C for 2 h to remove the remaining solvent residues.²² The dried membranes were then peeled off by dipping the glass matrix into DI water.⁸ The thicknesses of the fabricated membrane specimens were in the range of 120–125 μm. Scheme 1 illustrates the preparation of the ternary hybrid membrane.

Scheme 1 Preparation process of ternary hybrid membranes.

2.6. Process of the membrane activation

All the prepared membrane specimens were activated prior to the proton conductivity measurement, using the following procedure. First, the specimens were treated with 0.5 M H₂SO₄ solution for 1 h to activate the acidic groups, and were brought to boiling with DI water for 1 h to cause hydration. The specimens were then saturated with excess DI water at room temperature for 12 h to remove the free acid contents.^40,41

3. Characterization and apparatus

3.1. Structural and morphological characterization

The crystallinities of the Gr, GO, polymers and membranes were evaluated using a high resolution X-ray diffractometer, the X'pert-MRD model, having a radiation of CuKα in the scattering 2 theta angle range of 5–80°. The chemical structures of the constituent materials and membranes were characterized using an FT-NMR spectrometer (JNM-ECA600-600 MHz), with DMSO-d₆ solvent. The functional groups of specimens of the constituent materials and membranes were characterized by Fourier transform infrared (FT-IR) spectra, using a spectrum GX model spectrometer in the frequency range of 4000–400 cm⁻¹, where the KBr pellet method was exploited for sample preparation. The surface morphological images of specimens were captured by SUPRA 40VP microscope instrumentation (FE-SEM-Schottky type), comprising direct and cross detectors, following osmium sputtering. The roughness in the surface of the prepared specimens was investigated with an atomic force microscope (AFM), using the multimode-8 model in the tapping mode.

3.2. Thermal and mechanical characterizations

The thermal behaviors of the specimens of materials and membranes were analyzed using the TA instruments Q-50 model and the analysis was conducted from 30 to 800 °C at a heating rate of 5 °C min⁻¹, under N₂ gas atmosphere. Analysis of the glass transition properties of the specimens was performed with the TA instruments Q-20 model in the temperature range of 80–260 °C, at a heating rate of 10 °C min⁻¹. The temperature variant mechanical properties of the fabricated specimens of membranes were analyzed using DMA instruments Q800, having the tan delta sensitivity of 0.0001 and the analysis was conducted from 20 to 300 °C.

3.3. Apparatus used

The membrane specimen thickness was measured using an ABSOLUTE digimatic indicator, Mitutoyo, Model: ID-S112X, Japan. The mass of the membrane specimen was evaluated by the Denver four digit micro balance (Model: S-234). The length and width of the membrane specimens were measured using the ABSOLUTE digital vernier caliper (150 mm range), model: 500-196-20.

4. Measurements

4.1. Changes in scales of thickness, length and mass

The previously dried (at 100 °C for 5 h) membrane specimens were measured for mass, length and thickness, and then dipped in DI water at 30 °C, 50 °C, 70 °C or 90 °C for 24 h. Then, the water on the surface of the specimens was carefully blotted with tissue paper and the changes in mass, length, and thickness of the wet specimens were measured. Each measurement was repeated three times and the reported value is the average of the measurements with an error of ±1.0. The following equations were used to calculate the changes:⁵


where T_wet, L_wet, and M_wet are the thicknesses (μm), lengths (mm) and masses (mg) of the wet specimens, respectively, and T_dry, L_dry, and M_dry are the thicknesses, lengths and masses of the dry specimens, respectively.

4.2. Oxidative stability

To measure the oxidative stability of membrane specimens, Fenton's test was conducted based on the literature.⁴² Previously dried membrane specimens (at 100 °C for 5 h) were weighed and dipped into the Fenton's reagent (3% H₂O₂ + 2 ppm FeSO₄) at 80 °C until it started to rupture. The stabilities of the specimens were estimated from the rupture time (RT) and remaining weights (RW).

4.3. Ion exchange capacity (IEC)

The IECs of the membrane specimens were measured by acid–base titration method.⁴³ A piece of dry specimen (90 mg) was saturated in 0.1 M solution of NaCl for 1 day for the ion exchange of Na⁺ by H⁺. The generated HCl was then titrated against the 0.01 M NaOH solution using phenolphthalein as the indicator. Each titration was repeated several times and the reported value is the constant of two consecutive titrations with the error of ±0.05. IEC was calculated using the following equation:
where M_dry is the relevant mass of dry specimens in grams, V_NaOH is the volume of the NaOH in mL, and C_NaOH is the concentration of NaOH in M.

4.4. Hydration number

The amount of water molecules adsorbed per cluster volume of SO₃H moieties at different temperatures was calculated from the following equation:^15,43
where λ is the hydration number, and 18.01 is the molecular weight of water in g mol⁻¹.

4.5. Proton conductivity

The proton conductivities for the in-plane directions of the fabricated membrane specimens were measured using alternating current impedance spectroscopy (Sci Tech instrumentation, Keithley-2400 source meter with the conjugation of Bekk Tech four electrode conductivity cell).^41,44–46 The pre activated and hydrated specimen (30 mm × 5 mm) was placed perpendicular to the current–voltage Pt electrodes of a Bekk Tech cell, separated by equal distances of 4.20 mm. The current was passed through the two Pt electrodes and the remaining two Pt electrodes were used to assess the voltage drops. The parameters such as frequency, relative humidity, and temperature were controlled by the system. The proton conductivities of the specimens were calculated from the following equation:
where σ is the proton conductivity in mS cm⁻¹, L is the distance between the Pt electrodes in cm, R is the resistance in ohms, and T and W are the thickness and width of the membrane specimen in cm. Each measurement was repeated several times for the purposes of reproducibility.

4.6. H₂ gas permeability

The permeability of H₂ gas in the membrane specimens was measured to examine the fuel resistance properties by traditional constant volume/variable pressure method.^47–49 Before performing the measurements, the specimens were dried at 100 °C for 2 h to remove the moisture content. The H₂ gas was fed with the pressure rate of 1 barrer, and the temperature was kept at 30 °C throughout the experiment. Each measurement was repeated three times and the reported value is the average of the measurements with an error of ±0.1. The following equation was used to calculate the H₂ gas permeability:

P = DS = V_pl(P_p2 − P_p1)/[ARTΔt(P_f − (P_p2 + P_p1)/2)]

where P is the gas permeability in barrer, D is the diffusivity coefficient in cm² S⁻¹, S is the solubility coefficient in (cm³ (cm² scm Hg)⁻¹), l is membrane thickness in cm, V_p is the constant permeation volume in cm³, A is the active area of membranes in cm², R is the gas constant in J mol⁻¹ K⁻¹, T is the temperature in kelvin, Δt is the time of pressure change from P_p1 to P_p2 and P_f is feed pressure in cm Hg.

5. Results and discussion

5.1. Morphological properties

The morphologies of the natural Gr and GO were analyzed by FE-SEM and the micrographs are shown in Fig. 1. The natural Gr exhibits a morphology of compressed, thicker, and darker structure, due to the stronger van der Waals interactions (π–π stacking) exerted between the layers (Fig. 1a). Compared to Gr, thin layers are observed for GO (Fig. 1b and c) because of oxidation. The typical oxygen related functionalities, including carboxyl, epoxy, and hydroxyl have been directly grafted at interlayers of graphite during the oxidation, which hindered the restacking of layers and formation of flakes.³⁹ However, some unoxidized domains of GO still show entangled thick sheets. Fig. 1d shows the photographs of Gr and GO. It can be seen that compared to Gr, GO is effectively dispersible in water, due to its surface hydrophilic functionalities.

Fig. 1 FE-SEM images of (a) Gr, (b and c) GO, and (d) dispersibility images of Gr and GO.

The porosity, the synergism between individual constituents, and the filler intercalation are significant factors in determining the thermal, mechanical, fuel barrier, and proton conduction properties of membranes. The microstructures of fabricated membrane specimens were investigated through FE-SEM and Fig. 2 presents the obtained micrographs. The SP specimen (Fig. 2a and b) shows the morphologies of pores and pinholes, caused by the incompatibility between interconnected hydrophilic channels and hydrophobic polymer backbones, while a morphology with less pinholes was observed for SPSPV (Fig. 2c and d), owing to the physical reinforcements of SPVdF-HFP. However, the pores still persisted in the SPSPV specimen, which perhaps provide more selective permeation for H⁺ ions than gas molecules. It can be clearly observed that the SPSPVG-5 specimen (Fig. 2e and f) exhibits a crumpled morphology where the aromatic and aliphatic chains of polymers were clenched or folded with the GO sheets. The attained crumpling was due to the π–π stacking and H-bonding between the polymer chains and GO, which enable the effectual dispersion of GO sheets throughout the polymer matrix. Scheme 2 shows the possible interactions exerted between the constituents in the hybrid membrane. The elemental mappings and energy dispersive X-ray spectra of SPSPVG-5 are shown in Fig. S1a–f.† As observed from the figure, carbon, oxygen, fluorine, and sulfur are present in the hybrid membrane. The 3D AFM images of membrane specimens are displayed in Fig. 3a–e. The surface roughness calculated for SP was considerably reduced when it was blended with SPVdF-HFP, whereas, it was again increased by increasing the wt% of GO in the membrane, owing to the formation of a crumpled morphology. Fig. S2a–e† displays the optical photographs of the membrane specimens. The photograph of SP is fully transparent, compared to the SPSPV hybrid membrane. However, after the addition of GO in the membrane, the transparency is effectively decreased and is found to further decrease upon increasing the wt% of GO in the membrane.

Fig. 2 Surface FE-SEM images of membranes of (a and b) SP, (c and d) SPSPV, and (e and f) SPSPVG-5.

Scheme 2 Possible interactions between the individual constituents in the hybrid membranes.

Fig. 3 3D AFM images of membranes of (a) SP, (b) SPSPV, (c) SPSPVG-1, (d) SPSPVG-3, and (e) SPSPVG-5.

5.2. Structural properties

The ¹H-NMR spectra for analyzing the chemical structures of polymers and composite membranes are shown in Fig. 4. Fig. 4A shows the ¹H-NMR spectra of SPVdF-HFP in DMSO-d₆. The H atoms that exist in the PVdF fragment of SPVdF-HFP show two sharp peaks related to the head to head (hh) and head to tail interactions (ht).^50,51 Fig. 4B shows the ¹H-NMR spectra of SPEEK in DMSO-d₆. In this case, sulfonation appears at the phenyl rings between the ether linkages.⁵² The SO₃H group causes the down field shift for the H_E proton (7.4 ppm), compared to the other protons in the hydroquinone ring. The DS can be evaluated by taking the ratio of the area of H_E peak and the sum of the peak areas of the all other aromatic protons. The following equation was used to calculate the DS:
where A is the area of the peak, n is the number of H_E present per unit; H_A, H_A′, H_B, H_B′, H_C, and H_D are symbols referring to the corresponding aromatic protons, as shown in Fig. 4B.⁵³

Fig. 4 ¹H-NMR spectra of (A) SPVdF-HFP and (B) SPEEK.

FT-IR spectra were recorded to investigate the intermolecular bond stretching of the specimens. Fig. S3† reveals the FT-IR spectra of the constituent materials. GO formation from Gr was confirmed by the formation of new characteristic peaks at 3407 cm⁻¹ (–OH stretching of carboxyl or intercalated water molecules), 1721 cm⁻¹ (C=O stretching of COOH group), 1223 (C–O stretching of C–OH) and 1040 cm⁻¹ (C–O stretching from epoxy groups).^54–56 The peak appearing at 1601 cm⁻¹ is common to both GO and Gr, which is attributed to the unoxidized C=C stretching. Compared to PVdF-HFP, SPVdF-HFP shows additional significant peaks at 1637 cm⁻¹ and 1396 cm⁻¹, which are related to the symmetric and asymmetric stretching of sulfone in SO₃H groups, respectively; the broad band appearing in the region of 3455 cm⁻¹ is ascribed to the asymmetric stretching of –OH in SO₃H groups.^57,58 The peak found at 2895 cm⁻¹ is common to both PVdF-HFP and SPVdF-HFP and is correlated to the symmetric stretching of –CH₂ in aliphatic chains. For the aromatic polymers, the band observed at 1644 cm⁻¹ is associated with the asymmetric stretching of –C=O groups in the polymer chains.²² The SO₃H group present in SPEEK showed new characteristic peaks at 1222, 1077, and 1012 cm⁻¹, which correlate to the asymmetric and symmetric stretching of sulfone in SO₃H groups, and the absorption peak at 3418 cm⁻¹ is attributable to the asymmetric stretching of –OH in the SO₃H groups.⁴⁹ However, these peaks are invisible in the PEEK case. This result verifies the successful sulfonation of PEEK to SPEEK. The FT-IR spectra of the membrane specimen are shown in Fig. 5. In comparison to SP, the hybrid membranes exhibit the new low intensity band at 2895 cm⁻¹ that corresponds to the symmetric stretching of –CH₂ in the SPVdF-HFP polymer.²² The blending of SPVdF-HFP can also be confirmed by the –OH peak shift from 3418 to 3477 cm⁻¹ for SPSPV, which is possibly due to the overlapping of –OH stretching of SPVdF-HFP with –OH stretching of SPEEK. After the addition of GO, the intensity of this peak was slightly reduced because of non-covalent, cross H-bond interactions between hydrophilic functionalities of GO and SO₃H groups of SPEEK and SPVdF-HFP.

Fig. 5 FT-IR spectra of membranes of SP, SPSPV, and SPSPVG-5.

The influence of oxidation and sulfonation on the crystallinity of constituent materials is analyzed by XRD and the diffractions patterns obtained are shown in Fig. S4.† From the diffraction patterns of Gr and GO, it was observed that the 2θ angle of Gr appearing at 26.3° shifted to 11.6° after the oxidation process. This shift could be due to the increased inter-layer distance caused by the presence of diverse oxygen related functionalities.⁵⁹ PVdF-HFP has the characteristic sharp 2θ peaks at 18.4°, 19.9° and 26.8°, which correspond to the diffractions of α, β, and γ phases, respectively.⁶⁰ After sulfonation, the peak at 19.9° (β phase) became relatively amorphous, demonstrating that the direct grafting of SO₃H moieties in the aliphatic chains diminishes the crystallinity of the polymer.³⁶ The semicrystalline nature of PEEK was verified by the sharp 2θ peaks found at 18.6° (110), 20.6° (111) and 22.5° (200). However, the crystalline nature was significantly reduced after sulfonation, evidenced by the single broad peak of SPEEK at 19.9°.^49,61 The XRD patterns of membrane specimens were recorded to study the influence of blending and are displayed in Fig. 6. Compared to the SP specimen, the SPSPVG-5 has the additional three sharp 2θ peaks at 21.5°, 23.9° and 26.9°, associated with the α, β and γ phases of fluorine in SPVdF-HFP. However, after blending, these peaks shifted to higher 2θ because of direct electrostatic interactions of fluorine with the polar hydrogens of OH, COOH and SO₃H groups in GO and SPEEK.

Fig. 6 XRD patterns of membranes of SP and SPSPVG-5.

5.3. Thermal behaviors

TGA thermographs of Gr, GO, PVdF-HFP, SPVdF-HFP, and PEEK are given in Fig. S5.† As shown in the thermographs, GO is thermally less stable than Gr and diverse substantial weight drops were revealed at different ranges, including the following: (i) below 100 °C, ascribed to the elimination of unbound water molecules; (ii) 100–150 °C, owing to the release of the bound water molecules; (iii) 150–230 °C, which is associated with the decay of labile oxygen related functionalities; (iv) 600–800 °C, caused by the destruction of the carbon skeleton.^43,62 The third weight drop in particular, demonstrates the oxidation of Gr to GO. In the PVdF-HFP case, a single step weight drop found from 400 to 470 °C is attributed to the decay of the aliphatic chains. Unlike PVdF-HFP, three step weight drops were observed for SPVdF-HFP, including the following: (i) 60–150 °C, related to the bound and free water loss; (ii) 200–350 °C, caused by the loss of SO₃H moieties grafted in the aliphatic chains; (iii) 380 to 470 °C, ascribed to the decay of aliphatic backbone chains.⁶³ The PEEK exhibited the steep single step drop over 500 °C, attributed to the disintegration of aromatic chains,⁶⁴ while the SPEEK provided three different steps of weight drops.^15,49 From the thermographs of the membrane specimens (Fig. 7a), it can be clearly observed that the SP and hybrids show similar types of three step weight drops: (i) from 30 to 230 °C, (ii) from 240 to 350 °C, and (iii) from 470 to 800 °C. (i) and (ii) are due to the evacuation of free and bound water molecules and the detachment of SO₃H moieties, and (iii) is due to the decay of the aromatic SP backbones. However, the hybrids revealed a retarded manner of weight drops, compared to SP. Therefore, the addition of SPVdF-HFP and GO is justified, as they effectively improve the thermal stability of SP by prohibiting the movement of aromatic chains through interfacial interactions. At 800 °C, the SP showed the remaining weights of 37.24%, whereas SPSPV, SPSPVG-1, and SPSPVG-5 exhibited the weights of 40.10%, 41.24%, and 41.83%, respectively. The Nafion-117 membrane exhibited backbone degradation over the range of 400 °C, which is relatively lower than the prepared SP membranes.

Fig. 7 (a) TGA and (b) DSC curves of bare SP and hybrid membranes.

Thermograms (DSC) of the constituent materials are presented in Fig. S6.† PEEK shows the endothermic inflexion (T_g) at 340 °C because of the effectual reorganization in the chain structure attained by the crystalline nature of the polymer.⁶⁵ After sulfonation, the T_g was shifted to a lower temperature (203 °C) with successive broadening, demonstrating that the crystallinity of PEEK was considerably reduced by the hydrophilic nature of SO₃H groups. For PVdF-HFP, the endothermic inflexion was observed at 164 °C, while SPVdF-HFP showed the inflexion at 157 °C. This change is due to the hydrophilic behavior of aliphatic SPVdF-HFP backbones.³⁶ As shown in Fig. 7b, a relatively lower T_g is noticed for the hybrids, compared to the SP membrane, due to the sluggish reorganization of aromatic SPEEK chains in the hybrids caused by interfacial H-bonding and π–π interactions of SPVdF-HFP and GO. The specimens of SPSPV, SPSPVG-1 and SPSPVG-5 provided the respective T_g at 161 °C, 150 °C and 145 °C, whilst SP was shown at 203 °C.

5.4. Temperature reliant mechanical stability and oxidative stability

Temperature reliant storage modulus curves are demonstrated in Fig. 8a. From this, it can be observed that modulus is substantially increased upon the addition of SPVdF-HFP and GO in SP. The maximum value of storage modulus showed by the SPSPVG-5 (3492 Mpa) is 2.2 times higher than that of SP (1533 Mpa) specimen. The higher storage modulus attained for the hybrid membrane is due to the high mechanical strength of C–F bonds of SPVdF-HFP and the high Young's modulus of the GO carbon skeleton. Furthermore, the strong H-bonds formed due to the existence of common hydrophilic functional groups in SPEEK, SPVdF-HFP and GO, effectively confine the movements of polymeric chains and it further extends the strength of the membrane.¹⁵ The tan δ DMA curves for the membrane specimens are shown in Fig. 8b. The tan δ value of SPSPVG-5 is increased by 2.9 times, compared to that of the SP membrane.

Fig. 8 DMA curves of (a) storage modulus and (b) tan δ of bare SP and hybrid membranes.

In order to evaluate the oxidative stability of the membrane specimens, the Fenton's test was conducted. Fig. 9 displays the remaining weights and rupture time of the specimens. It can be seen that SP exhibited 94% of the remaining weight, while the SPSPV, SPSPVG-1, SPSPVG-3, SPSPVG-5 and SPSPVG-7 revealed 98, 97.7, 97.1, 96.8 and 96% of the remaining weights, respectively. Besides, all the hybrid membranes are stable up to 5 h. The results that were obtained indicate that the radicals have little effect on the hybrid membranes, due to the high oxidative resistivity of the C–F bond of SPVdF-HFP. Furthermore, the mutual H-bonds formed by GO and SPVdF-HFP protect the functionalities of SP from the radical attacks and further enhance the stabilities of the hybrids.⁸

Fig. 9 Oxidative stabilities of bare SP and hybrid membranes.

5.5. Percentage of changes in mass, length, and thickness of membrane, hydration number and IEC

The changes in mass, length, and thickness, as well as the hydration number and the IEC are critical properties of the PEM to determine the proton conductivity.¹⁵ Fig. 10a–c show the changes that occurred in mass, length, and thickness while treating the membranes at 30 °C, 50 °C, 70 °C, and 90 °C. Usually, the percent changes in mass refer to the water uptake of membranes and a higher water uptake consequently leads to the higher swelling profile. At 90 °C, the SP membrane showed a maximum water uptake of 39.47% and the related changes in length and thickness were 26.23% and 12.71%, respectively. Even at high temperature, the SP membrane is not completely soluble; this may be due to the preheating at 100 °C, thus delaying the so called plasticization time of polymeric chains. Compared to SP, the SPSPV, SPSPVG-1, SPSPVG-3, SPSPVG-5 and SPSPVG-7 showed higher% water uptakes of 47.36%, 49.96%, 52.98%, 56.95% and 50.90%, while the corresponding changes in lengths were 28.30%, 27.39%, 25.42%, 24.22% and 23.39%, respectively. The acquired results demonstrate that the ternary hybrid membranes exhibited low swelling in the length even after the uptake of more water, which is due to the strong hydrophobic backbone interaction of GO with the aromatic chains of SPEEK. The maximum swelling in the length of the SPSPVG-5 hybrid membranes is slightly larger than that of the Nafion-117 membrane (21.7%).⁸ However, the hybrid membranes demonstrated increased swelling in the thickness, as shown in Fig. 10c. IEC of the membranes is directly proportional to the charge densities in the membranes; it is an essential factor to influence the proton conduction properties.⁸ Table 1 lists the values of IECs of membranes. IEC for SP was 1.59, which is lower than the IECs of SPSPV (1.66), SPSPVG-1 (1.72), SPSPVG-3 (1.76), SPSPVG-5 (1.81) and SPSPVG-7 (1.78). Enhanced IEC of the SPSPVG-X hybrid membranes is associated with the higher volume per cluster of SO₃H groups and excess density of OH and CO₂H groups attained, due to the existence of SPVdF-HFP and GO. In each membrane, each cluster of SO₃H moieties adsorbs a certain amount of water molecules, which is known as the hydration number (λ). The peak value of λ (Fig. 10d) yielded by the SPSPVG-5 hybrid is 17.45, whereas the SP specimen affords the value of 13.76. The 1.26 time enhancement in λ is due to higher IEC of the hybrid membrane. The overall significant improvements in all the above mentioned factors consequently enhance the proton conductivities of the ternary hybrid membranes.

Fig. 10 (a) mass change, (b) length change (c) thickness change, and (d) hydration number of bare SP and hybrid membranes.

Table 1 Properties of ternary hybrid membranes

Membrane	IEC (meq. g^−1) (±0.05)	(H2O/SO3H) (±0.1)	Water uptake (%) (±1.0)	σ (mS cm^−1) (±0.1)	Ea (kJ mol^−1) (±0.5)	H2 permeability (barrer) (±0.1)
SP	1.59	13.76	39.47	68.1	19.31	5.83
SPSPV	1.66	15.83	47.36	90.3	16.30	3.35
SPSPVG-1	1.72	16.11	49.96	101.4	15.46	2.95
SPSPVG-3	1.76	16.70	52.58	115.2	15.31	—
SPSPVG-5	1.81	17.45	56.95	122.5	14.71	1.64
SPSPVG-7	1.78	15.85	50.90	118.3	15.79	—
Nafion-117	0.95	—	—	141.3	13.55	—

---

5.6. Proton conductivity of membranes with various temperatures and Arrhenius plots

Proton conductivity of PEM is a substantial process, which directly influences the final power efficiency of fuel cells. Proton conduction in PEM mainly appears through two different methods: (i) Grotthuss mechanism and (ii) vehicular mechanism. In the Grotthuss mechanism, the protons are transferred from one ionic site to a nearer site via the H-bond networks between them, while in the vehicular mechanism, they diffuse with the carrier ions such as H₃O⁺ generated from the water molecules.^22,49,66 The SO₃H, OH, and CO₂H groups that exist in the membrane perform a vital role in the Grotthuss type conduction, while water molecules adsorbed by the SO₃H groups play a significant role in both types of proton conduction. Both mechanisms usually coexist while operating the cell below 100 °C, and above this temperature, the Grotthuss mechanism solely persists because of the evaporation of water molecules.

The proton conductivities of the membrane specimens were measured at various temperatures under both hydrated and anhydrous conditions. Fig. 11a demonstrates all obtained proton conductivity plots up to 90 °C (hydrated condition). The proton conductivity attained by the SP specimen was 19 mS cm⁻¹ at 30 °C, which was considerably raised to 68 mS cm⁻¹ upon reaching the temperature of 90 °C. The improved proton conductivity obtained is due to rapid proton movement caused by the high temperature operation. Since the per cluster volume of SO₃H groups and the number of other hydrophilic groups increases in the membrane, the protons can easily move through the H-bonds formed between these functional groups. The addition of additives such as SPVdF-HFP and GO further enhances the SP conductivity by increasing the per cluster volume of SO₃H groups and increasing the density of OH and CO₂H functionalities. The SPSPVG-X hybrids afforded the maximum conductivities are in the range between 101–122 mS cm⁻¹, a few fold larger, compared to that of the bare SP specimen. Under identical conditions, the Nafion-117 membrane exhibited a conductivity of 141 mS cm⁻¹, indicating that the peak conductivity of the ternary hybrid membrane is still 1.15 fold lower than Nafion-117. This may due to the relatively higher swelling profile of the prepared membranes than Nafion-117. As mentioned in the above section, both types of mechanisms coexist in the hybrid membranes at low temperatures. However, the Grotthuss mechanism is highly responsible for the increased conductivity of hybrid membranes, where the H-bonds are directionally reorganized with the assistance of functionalities in GO.²⁹ The nano-fluidic path channels of GO, such as –OH, –O– and –CO₂H formed extended H-bond networks with the –SO₃H groups of SPEEK and SPVdF-HFP, facilitating the facile and smooth travel of protons throughout the hybrid membrane. At high temperature operation (Fig. 11b), all the membrane specimens lose their proton conductivity, due to the dehydration. However, even at 130 °C, the SPSPVG-X hybrid specimens yielded considerable proton conductivity in the range of 13–21 mS cm⁻¹, which is higher than that of the Nafion-117 specimen, implying that the prepared ternary hybrid membrane can also be useful for the high temperature PEMFC applications. Table 2 lists the proton conductivity results of various GO based membranes reported in literature, for comparison with the present study.

Fig. 11 Proton conductivities of bare SP and hybrid membranes, under (a) hydrated conditions and (b) anhydrous conditions, Arrhenius plots of proton conductivities of bare SP and hybrid membranes under (c) hydrated conditions and (d) anhydrous conditions.

Table 2 Comparisons of GO based hybrid membranes

Membrane	Proton conductivity (mS cm^−1)	IEC (meq. g^−1)	Water uptake (%)	Reference
SPSPVG-5	122.5	1.81	56.95	Present work
Nafion/GO-4	170.0	1.38	37.2	29
SPI/GO	158.0	1.37	23.0	30
Nafion/F-GO	47.0	0.96	29.2	40
SPES/S-GO-5	58.0	1.63	17.4	43
SPEEK/SSGO-5	52.3	1.66	30.1	49
Nafion/GO	80.9	—	42.0	67
Chitosan/S-GO	61.0	1.95	59.7	68
Nafion/SPEEK/HGO	322.2	—	71.2	69
SPEEK/Ssi-GO	162.6	1.87	50.9	70
SPEEK/SDBS-GO-8	162.6	1.83	53.2	71

---

The activation energy required for the proton travel is a paramount factor for PEM, and low activation energy usually allows effortless proton conduction in the PEM.⁴⁹ The acquired Arrhenius plots (Fig. 11c and d) of membranes have been evaluated from the corresponding temperature reliant proton conductivity data. The reported activation energy E_a (Table 1) of membranes was derived from the Arrhenius plots using the following equation:⁷¹

ln σ = ln σ₀ − E_a/RT

where σ and σ₀ are the values of proton conductivity and pre-exponential factor, respectively, measured in mS cm⁻¹; E_a is the activation energy required for protons to travel, measured in kJ mol⁻¹; R is the gas constant in J mol⁻¹ K⁻¹ and T is temperature in kelvin. According to the calculated E_a values for low temperature operation, the SPSPVG-X hybrid specimens exhibited E_a in the range of 15.79–14.71 kJ mol⁻¹, which is nearer to that of Nafion-117 (13.55 kJ mol⁻¹). These results indicate that the hybrid membrane demonstrates comparable activation energy to the Nafion 117 membrane. However, under the same conditions, the SP specimen yielded the E_a of 19.31 kJ mol⁻¹, which is 1.2 times higher than that of the hybrid membrane. Therefore, it can be determined that the presence of GO is a substantial factor for facile and swift proton conduction in the membrane. The E_a values of the proton conduction at high temperatures were also derived from the corresponding Arrhenius plots. The obtained E_a values for SPSPVG-X membranes are in the range of 39.02–34.65 kJ mol⁻¹, higher than the values obtained for low temperature, showing that Grotthuss type proton conduction only persists in all the membranes while operating the cell at high temperature under anhydrous conditions.⁴⁹

5.7. H₂ gas permeability

Fuel permeation in PEM is the crucial drawback that diminishes the power performance of fuel cells through reducing the fuel efficiency and causing the mixed potential of fuel and oxidant. For improved PEMFC efficiency, the PEM should have minimum fuel permeability and high proton selectivity. Fig. 12 displays the obtained values of H₂ permeability in prepared membranes. The SP reveals the permeability of 9.83 barrer, which is higher in comparison to the hybrid membranes; this is because a higher number of pinholes and tears exist in SP membranes, as shown in the FE-SEM image. SPSPV revealed the permeability of 5.35 barrer, indicating that blending of SPVdF-HFP significantly reduced the permeability by reducing the numbers of pinholes and tears in the membranes. In the case of SPSPVG-X membranes, GO further resized the holes in the membranes by connecting the hydrophilic domains and forming supramolecular structures with the aromatic polymer chains, thus leading to lower H₂ gas permeability (3.95–1.64 barrer), compared to the non-filled membranes.

Fig. 12 H₂ permeability of bare SP and hybrid membranes.

6. Conclusion

Ternary hybrid membranes, incorporating different wt% of GO, were successfully fabricated through a facile solution casting method. In the ternary hybrids, the per cluster volume of the SO₃H groups was increased using SPVdF-HFP, and the H-bond networks of SO₃H groups were extended using GO. The ternary hybrids show good water uptake, in-length dimensional stability, IEC, λ, and proton conductivity, compared to those of the SP membrane. The introduction of GO reorganized the aromatic polymer chains through H-bonding and π–π interactions, which strengthened the membrane thermally and mechanically by 1.1 and 2.2 fold, respectively. H₂ gas permeability decreased from 9.83 to 1.64 barrer upon increasing the content of GO from 0 to 5% in the membrane. The present study affords valuable information about the fabrication and physicochemical, thermomechanical, fuel permeation, and proton conduction properties of ternary hybrid type electrolytes. Therefore, this study might help to realize the significance of ternary hybrid type electrolytes in the applications of PEMFC.

Acknowledgements

This work was 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. 20164030201070).

References

1. B. E. Logan, Environ. Sci. Technol., 2004, 38, 160.
2. R. P. Pandey, A. K. Das and V. K. Shahi, ACS Appl. Mater. Interfaces, 2015, 7, 28524.
3. X. Cheng, Z. Shi, N. Glass, L. Zhang, J. Zhang, D. Song, Z. S. Liu, H. Wang and J. Shen, J. Power Sources, 2007, 165, 739.
4. B. C. H. Steele and A. Heinzel, Nature, 2001, 414, 345.
5. A. M. Baker, L. Wang, W. B. Johnson, A. K. Prasad and S. G. Advani, J. Phys. Chem. C, 2014, 118, 26796.
6. C. Y. Wang, Chem. Rev., 2004, 104, 4727.
7. R. Borup, Chem. Rev., 2007, 107, 3904.
8. S. Neelakandan, N. J. Kaleekkal, P. Kanagaraj, R. M. Sabarathinam, A. Muthumeenal and A. Nagendran, RSC Adv., 2016, 6, 51599.
9. J. Lu, H. Tang, C. Xu and S. P. Jiang, J. Mater. Chem., 2012, 22, 5810.
10. C. H. Wirguin, J. Membr. Sci., 1996, 120, 1.
11. K. A. Mauritz and R. B. Moore, Chem. Rev., 2004, 104, 4535.
12. B. R. Matos, M. A. Dresch, E. I. Santiago, L. P. R. Moraes, D. J. Carastan, J. Schoenmaker, I. A. V. Davalos, A. Ruediger, A. C. Tavares and F. C. Fonseca, Electrochim. Acta, 2016, 196, 110.
13. A. N. Mondal, B. P. Tripathi and V. K. Shahi, J. Mater. Chem., 2011, 21, 4117.
14. V. S. Silva, B. Ruffmann, S. Vetter, A. Mendes, L. M. Madeira and S. P. Nunes, Catal. Today, 2005, 104, 205.
15. S. Gahlot and V. Kulshrestha, ACS Appl. Mater. Interfaces, 2015, 7, 264.
16. S. Kaliaguine, S. D. Mikhailenko, K. P. Wang, P. Xing, G. Robertson and M. Guiver, Catal. Today, 2003, 82, 213.
17. S. D. Mikhailenko, K. Wang, S. Kaliaguine, P. Xing, G. P. Robertson and M. D. Guiver, J. Membr. Sci., 2004, 233, 93.
18. E. M. W. Tsang, Z. Zhang, A. C. C. Yang, Z. Shi, T. J. Peckham, R. Narimani, B. J. Frisken and S. Holdcroft, Macromolecules, 2009, 42, 9467.
19. B. T. J. Peckham and S. Holdcroft, Adv. Mater., 2010, 22, 4667.
20. J. Roziere and D. J. Jones, Annu. Rev. Mater. Res., 2003, 33, 503.
21. G. Gnanakumar, J. Shin, Y. C. Nho, I. S. Hwang, G. Fei, A. R. Kim and K. S. Nahm, J. Membr. Sci., 2010, 350, 92.
22. A. Bagheri, M. Javanbakht, H. Beydaghi, P. Salarizadeh, A. Shabanikia and H. S. Amoli, RSC Adv., 2016, 6, 39500.
23. A. K. Mohanty, E. A. Mistri, S. Banerjee, H. Komber and B. Voit, Ind. Eng. Chem. Res., 2013, 52, 2772.
24. G. Rambabua and S. D. Bhat, Chem. Eng. J., 2014, 243, 517.
25. A. K. Geim, Nature, 2009, 324, 1530.
26. M. J. Allen, V. C. Tung and R. B. Kaner, Chem. Rev., 2010, 110, 132.
27. B. Y. Zhu, S. Murali, W. Cai, X. Li, J. W. Suk, J. R. Potts and R. S. Ruoff, Adv. Mater., 2010, 22, 3906.
28. T. Ko, K. Kim, M. Y. Lim, S. Y. Nam, T. H. Kim, S. K. Kim and J. C. Lee, J. Mater. Chem. A, 2015, 3, 20595.
29. R. Kumar, C. Xu and K. Scott, RSCAdv., 2012, 2, 8777.
30. Y. He, C. Tong, L. Geng, L. Liu and C. Lu, J. Membr. Sci., 2014, 458, 36.
31. P. Xing, G. P. Robertson, M. D. Guiver, S. D. Mikhailenko, K. Wang and S. Kaliaguine, J. Membr. Sci., 2004, 229, 95.
32. S. M. J. Zaidi, S. D. Mikhailenko, G. P. Robertson, M. D. Guiver and S. Kaliaguine, J. Membr. Sci., 2000, 173, 17.
33. R. K. Nagarale, G. S. Gohil and V. K. Shahi, J. Membr. Sci., 2006, 280, 389.
34. K. Dutta, S. Das and P. P. Kundu, J. Membr. Sci., 2014, 468, 42.
35. K. Dutta, S. Das, P. Kumar and P. P. Kundu, Appl. Energy, 2014, 118, 183.
36. S. Das, P. Kumar, K. Dutta and P. P. Kundu, Appl. Energy, 2014, 113, 169.
37. P. P. Sharma, S. Gahlot, B. M. Bhil, H. Gupta and V. Kulshrestha, Macromol. Symp., 2015, 357, 154.
38. B. W. S. Hummers and R. E. Offeman, J. Am. Chem. Soc., 1958, 80, 1339.
39. R. Kannan, A. R. Kim, K. S. Nahm, H. K. Lee and D. J. Yoo, Chem. Commun., 2014, 50, 14623.
40. H. Zarrin, D. Higgins, Y. Jun and Z. Chen, J. Phys. Chem. C, 2011, 115, 20774.
41. A. K. Sahu, K. Ketpang, S. Shanmugam, O. Kwon, S. Lee and H. Kim, J. Phys. Chem. C, 2016, 120, 15855.
42. M. Vinothkannan, R. Kannan, A. R. Kim, G. Gnanakumar, K. S. Nahm and D. J. Yoo, Colloid Polym. Sci., 2016, 294, 1197.
43. S. Gahlot, P. P. Sharma, V. Kulshrestha and P. K. Jha, ACS Appl. Mater. Interfaces, 2014, 6, 5595.
44. C. Xu, Y. Cao, R. Kumar, X. Wu, X. Wang and K. Scott, J. Mater. Chem. A, 2011, 21, 11359.
45. S. Ma, A. Kuse, Z. Sirmo and K. Yasuda, Measuring proton conductivity, Espec Technology Report-2, 2000, p. 12.
46. M. Mamlouk, X. Wang, K. Scott, J. A. Horsfall and C. Williams, Proc. Inst. Mech. Eng., Part A, 2011, 225, 152.
47. A. Car, C. Stropnik, W. Yave and K. V. Peinemann, J. Membr. Sci., 2008, 307, 88.
48. S. Wang, Z. Tian, J. Feng, H. Wu, Y. Li, Y. Liu, X. Li, Q. Xin and Z. Jiang, J. Membr. Sci., 2015, 473, 310.
49. L. Zhao, Y. Li, H. Zhang, W. Wu, J. Liu and J. Wang, J. Power Sources, 2015, 286, 445.
50. L. Xie, X. Huang, K. Yang, S. Li and P. Jiang, J. Mater. Chem. A, 2014, 2, 5244.
51. M. Rahimabady, L. Q. Xu, S. Arabnejad, K. Yao, L. Lu, V. P. W. Shim, K. G. Neoh and E. T. Kang, Appl. Phys. Lett., 2013, 103, 262904.
52. H. L. Wu, C. C. M. Ma, H. C. Kuan, C. H. Wang, C. Y. Chen and C. L. Chiang, J. Polym. Sci., Part B: Polym. Phys., 2006, 44, 565.
53. R. N. Muthu, S. Rajashabala and R. Kannan, Int. J. Hydrogen Energy, 2015, 40, 1836.
54. N. Jiang, Z. Xiu, Z. Xie, H. Li, G. Zhao, W. Wang, Y. Wu and X. Hao, New J. Chem., 2014, 38, 4312.
55. X. Li, J. Feng, Y. Du, J. Bai, H. Fan, H. Zhang, Y. Peng and F. Li, J. Mater. Chem. A, 2015, 3, 5535.
56. X. Liu, N. Wen, X. Wang and Y. Zheng, Nanomater. Nanotechnol., 2015, 5, 1.
57. S. Subianto, N. R. Choudhury and N. Dutta, Nanomaterials, 2014, 4, 2.
58. Y. Lei, Y. Luo, F. Chen and L. Mei, Polymers, 2014, 6, 1914.
59. F. T. Johra, J. W. Lee and W. G. Jung, J. Ind. Eng. Chem., 2014, 20, 2883.
60. Y. Feng, W. L. Li, Y. F. Hou, Y. Yu, W. P. Cao, T. D. Zhanga and W. D. Fei, J. Mater. Chem. C, 2015, 3, 1250.
61. A. Rahnavard, S. Rowshanzamir, M. J. Parnian and G. R. Amirkhanlou, Energy, 2015, 82, 746.
62. H. Yang, C. Shan, F. Li, D. Han, Q. Zhang and L. Niu, Chem. Commun., 2009, 3880.
63. S. Neelakandan, D. Rana, T. Matsuura, A. Muthumeenal, P. Kanagaraj and A. Nagendran, Solid State Ionics, 2014, 268, 35.
64. J. H. Chang, J. H. Park, G. G. Park, C. S. Kim and O. O. Park, J. Power Sources, 2003, 124, 18.
65. O. Coban, M. O. Bora, E. Avcu and T. Sinmazcelik, Polymer Compos., 2011, 32, 1766.
66. L. Vilciauskas, M. E. Tuckerman, G. Bester, S. J. Paddison and K. D. Kreuer, Nature, 2012, 4, 461.
67. D. C. Lee, H. N. Yang, S. H. Park and W. J. Kim, J. Membr. Sci., 2014, 452, 20.
68. Y. Liu, J. Wang, H. Zhang, C. Ma, J. Liu, S. Cao and X. Zhang, J. Power Sources, 2014, 296, 898.
69. A. K. Mishra, N. H. Kim, D. Jung and J. H. Lee, J. Membr. Sci., 2014, 458, 128.
70. Z. Jiang, X. Zhao and A. Manthiram, Int. J. Hydrogen Energy, 2013, 38, 5875.
71. Z. Jiang, X. Zhao, Y. Fu and A. Manthiram, J. Mater. Chem., 2012, 22, 24862.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra22295a

---