Synthesis of sulfonated poly(bis(phenoxy ...

Solid State Ionics 296 (2016) 127–136

Contents lists available at ScienceDirect

Solid State Ionics journal homepage: www.elsevier.com/locate/ssi

Synthesis of sulfonated poly(bis(phenoxy)phosphazene) based blend membranes and its effect as electrolyte in fuel cells Rambabu Gutru, S. Gouse Peera, Santoshkumar D. Bhat ⁎, Akhila Kumar Sahu CSIR-Central Electrochemical Research Institute-Madras Unit, CSIR Madras Complex, Taramani, Chennai 600113, India

a r t i c l e

i n f o

Article history: Received 2 February 2016 Received in revised form 9 September 2016 Accepted 13 September 2016 Available online xxxx Keywords: Polyphosphazenes Fuel cross-over PEMFC sPEEK

a b s t r a c t In the present study, acid–base blend membranes are synthesized to imply as electrolytes in polymer electrolyte membrane fuel cells (PEMFCs). Poly(bis(phenoxy)phosphazene) (POP) is sulfonated with sulfuric acid to impart the ionic conductivity and then blended with sulfonated poly(ether ether ketone) (sPEEK). The blend membranes are fabricated by varying the sulfonated poly(bis(phenoxy) phosphazene) (sPOP) content from 2 to 4 wt.% in relation to sPEEK. The strong hydrophobic backbone of POP improves the mechanical strength of the membrane which is critical for the long term operation of PEMFCs. SAXS analysis suggests the change and enhanced dimension of ionic clusters in sPOP-sPEEK blend membrane in comparison with pristine sPEEK. Hydrophobic and hydrophilic phase distinction in the blend is determined by AFM analysis. Blend membranes exhibit higher ionic conductivity than pristine sPEEK membrane due to the acid–base interactions between sPEEK and sPOP. When subjected to cell polarization, blend membranes of sPOP-sPEEK attain peak power density of 935 mW cm−2 in PEMFCs which is on par with the peak power density observed for commercial Nafion-212 membrane. Apart from its higher performance, the sPOP (3 wt.%)-sPEEK blend also shows low fuel crossover and comparable stability to Nafion-212 membrane under cell operation. © 2016 Elsevier B.V. All rights reserved.

1. Introduction Fuel cells are considered to be one of the best alternative energy conversion systems due to their low environmental pollution, high efficiency and quick start up [1]. Among all the other types of fuel cells, polymer electrolyte membrane fuel cells (PEMFCs) are extensively studied in recent times due to their wide range of transport and stationary applications. Till date, the major hindrance in commercialization of PEMFCs lies in the cost and their long term durability of the components which include Pt based catalyst, proton exchange membrane (PEM) and bipolar plates [2–4]. Among these components, PEM plays a major role in determining the performance of PEMFCs; and Nafion® is the widely used membrane for the same due to its superior proton conductivity in humid environment in conjunction with its high mechanical strength and electrochemical stability [5,6]. However, the practical applications of these membranes are determined by several factors. Primarily, the cost of Nafion® membrane is still too high for commercial applications and its dehydration at elevated temperature significantly reduces proton conductivity [7,8]. To address these issues, immense efforts are in progress to modify Nafion® membranes for improved conductivity and stability. Other approach is to develop cost effective ⁎ Corresponding author. E-mail address: [email protected] (S.D. Bhat).

http://dx.doi.org/10.1016/j.ssi.2016.09.011 0167-2738/© 2016 Elsevier B.V. All rights reserved.

alternative hydrocarbon membranes with better ionic conductivity and stability in relation to Nafion® membranes [9–13]. Among the wide range of alternative membranes studied for PEMFCs, hydrocarbon polymer based membranes like sulfonated poly(ether ether ketone) (sPEEK) is considered to be a suitable candidate due to its ease of processing and unique physico-chemical properties [13,14]. However, the proton conductivity of the sPEEK membrane is less as it possesses many dead-end ion conducing channels which limits the proton transport across the membrane in comparison with Nafion®, wherein the ionic domains are well inter-connected [15]. Hence it is necessary to modify the sPEEK membranes in order to open up these dead end channels and also for enlarging the channel interconnectivity intended for enhanced conductivity. In polymers like sPEEK, the required proton conductivity can be achieved with increased degree of sulfonation, but its mechanical properties tend to deteriorate when the sulfonation content is higher than an optimal level. Moreover, higher sulfonation increases the water uptake due to which the membrane dimensional stability is gradually lost on long term operation and this phenomenon will still be serious when the fuel cell operates at higher temperatures and humidity levels [16]. To clearly understand and tackle these issues, one of the best approaches would be to disperse the inorganic additives like TiO2, ZrO2 and zeolites in sPEEK matrix to form organic/inorganic nanocomposite systems with superior thermomechanical properties [17–19].

128

R. Gutru et al. / Solid State Ionics 296 (2016) 127–136

Another unique approach to improve the properties of sPEEK is to form blend membranes with basic polymers. Basic polymers fine tunes the ionic domains of sulfonated polymers in the form of blend/ composite [20]. Among the basic polymers, we have selected poly(phosphazenes) which contain\\P_N\\inorganic moiety as backbone and are known for their better chemical and mechanical stabilities. These polymers can also be modified precisely by attaching the ion conducting sulfonic acid groups through substitution reaction on the aromatic ring [21–24]. Polymer electrolytes with additives having basic character lead to improved physicochemical and electrochemical properties [25]. Previously, composite membranes of sPEEK and poly bis(phenoxy)phosphazene] (POP) were studied as PEM for direct methanol fuel cells (DMFCs). POP being the inert/hydrophobic (no charge) component of the blend membrane, greatly reduced the methanol permeation in DMFCs with appreciable compromise in the proton conductivity [26]. The present study establishes the sulfonation of POP to form a compatible blend with sPEEK and is applied as PEM in H2\\O2 fuel cells. The blend membranes exhibit improved proton conductivity, lower reactant crossover (H2 and O2) in comparison with pristine sPEEK membranes. The membrane electrode assemblies (MEAs) comprising these blend membranes exhibit higher peak power density in PEMFCs in relation to pristine sPEEK and Nafion-212 membranes. 2. Experimental

were done to confirm the sulfonation of POP. Mechanical strength of all the membranes was estimated as described in our previous literature using universal testing machine (UTM) (Model AGS-J, Shimadzu, Japan) with an operating head-load of 10 kN [27]. FT-IR spectra for sPEEK and sPOP-sPEEK membranes were recorded following similar protocols mentioned above. Thermo-gravimetric analysis (TGA) of pristine sPEEK and sPOP-sPEEK blend membranes was carried out by using a NETZSCH STA 449F3 TGA-DSC instrument in the temperature range between 30 °C and 1000 °C at a heating rate of 5 °C min−1 with nitrogen flushed at 60 mL min−1. Surface and cross-sectional morphologies of pristine sPEEK and sPOP-sPEEK blend membranes were analysed through a JEOL JSM 35CF Scanning Electron Microscope along with elemental mapping for the samples. Topological and phase images for pristine sPEEK and its blend membranes were determined by tapping mode atomic force microscopy (AFM, PicoSPM-Picoscan 2100, Molecular Imaging, USA). Small angle X-ray scattering (SAXS) analysis was done on a Bruker Nanostar machine equipped with a Cu rotating anode with a tungsten filament. The SAXS was operated at a voltage of 45 kV and current of 20 mA with Cu Kα radiation (wavelength = 1.54 Å). Detector calibration was done with silver behenate. Samples were sandwiched between Kapton films and pasted on a metallic holder. Scattering data were recorded from the multiwire gas filled Hi-star 2D area detector and were reduced to 1D using Bruker offline software. To understand the better identification of the peak shift in the membrane samples, the scattering vector (q) scale was restricted to 0.22 Å−1.

2.1. Materials 2.4. IEC, water uptake and proton conductivity Sulfonated poly(ether ether ketone) (sPEEK, Mw = 50,000 g mol−1, Mn = 14,000, with IEC of 1.54 meq g− 1 and degree of sulfonation (DS) 54%) was procured from FuMA-Tech GmbH, Germany. Poly(bis(phenoxyphosphazene) was purchased from Sigma Aldrich, India. Dimethyl acetamide was obtained from Acros organics India. Commercial GDL (SGL-DC-35) was supplied by Nikunj Exim Pvt. Ltd., India. Pt/C (40 wt.% Pt on Vulcan XC-72R carbon) was obtained from Alfa Aesar (Johnson Matthey, India) chemicals.

Ion exchange capacity, water uptake and proton conductivity of aforesaid membranes were measured as per the procedure reported in our earlier studies [28,29]. IEC was also measured for sPOP (3 wt.%)-sPEEK membrane at different time intervals till 130 h to rule out the change in IEC due to the leaching of sulfonic acid groups of sPOP. Ion exchange capacity of membranes was measured by acid– base titration method and the following equation was used to calculate the IEC values.

2.2. Sulfonation of POP and synthesis of blend membranes Poly(bis(phenoxyphosphozene) was sulfonated through electrophilic substitution reaction as reported in the literature [24]. The typical sulfonation process is as follows: 0.2 g of POP was dissolved in mixture of conc. H2SO4 and DMAc (1:2 vol. ratio) at 80 °C with continues stirring for 2 h. The clear transparent solution was then neutralized with icecold water to quench the reaction to from the product precipitate. The precipitate is then filtered and washed with distilled water to remove the residual acid and the white precipitate obtained was dried in hot air oven at 60 °C to get sPOP particles. The blend membranes of sPOP-sPEEK were synthesized by varying the amount of sPOP (2–4 wt.%) in sPEEK. Initially the required amount of sPOP was dissolved in DMAc at 80 °C. Then the transparent solution obtained is mixed with 2 wt.% sPEEK that was already dissolved in DMAc. The mixture was stirred for 3 h for blending and the blend solution was cast uniformly on to a Plexi glass plate and dried in vacuum at 80 °C for 12 h. Finally the membranes were peeled out from the glass plate and dipped in 0.5 M H2SO4 for the activation of ion conducting groups. The membranes were then repeatedly washed with de-ionized water to remove any residual acid till neutral pH and then stored in the same for further studies. Thickness of all the membranes was in the range of 50–60 μm. For ease of comparison, Nafion-212 of similar thickness was used. 2.3. Characterization FT-IR spectra for POP and sPOP were recorded using a Nicolet IR 860 Spectrometer (Thermo Nicolet Nexus-670). Elemental analysis for POP and sPOP was performed on Elementarvario EL 111-Germany. These

IEC ¼

Volume of NaOH Normality of NaOH meq g−1 Dry weight of the sample

ð1Þ

Water uptake of the membranes was determined after equilibrating the membrane sample in a sorption chamber filled with distilled water and by measuring the weight difference between the wet and dry membrane samples with the following equation. Water uptake ¼

Wwet −Wdry 100 Wwet

ð2Þ

Proton conductivity of the membranes was measured at different temperatures (30 °C to 70 °C) under fully humidified condition using four-probe method. Subsequently activation energy for the membranes was also calculated by implementing the Arrhenius behaviour described in the literature [30,31] . 2.5. Oxidative stability (Fenton's test) Oxidative stability for the membranes was evaluated by Fenton's test. Fenton's solution (3 ppm) containing Fe+2 was prepared using 3 wt.% hydrogen peroxide and iron (II) chloride. The membrane samples were cut in equal size and dried to remove the absorbed moisture and weighed before dipping into Fenton's solution. The solution was heated to 70 °C for 5 h and then the membrane samples were removed from the solution and washed with water and dried at 60 °C before noting the final weight of the membrane sample. The difference between initial and final weight gives the weight loss due to the attack of hydroxyl and peroxy radicals during the Fenton's test [32].


2.6. Gas permeability measurements Air permeability for pristine sPEEK and sPOP-sPEEK blend membranes was analysed using a capillary flow porometer (Porous Materials, Inc. NY) and compared with Nafion-212. Typically 5 cm2 test membrane sample was placed on O-ring of the adapter and then the appropriate adopter was placed on top of the sample, compressive stress was applied on the membrane sample, pressure drop (P) across the sample and the flow rate (F) of gas through the pores of the membrane were measured. Average permeability was calculated using Darcy's law as follows [33]. C¼

8FTV πD P2 −1 2

ð3Þ

where C is the Darcy's permeability constant, F is the flow rate of the permeate, T is the thickness of the sample (50 μm), V is the viscosity of permeate (1.87 × 10−5 kg m−1-s), D is the diameter (2.5 cm) and P is the pressure drop across the membrane. 2.7. Fabrication of membrane electrode assembly (MEA) The performance of the aforesaid membranes was evaluated in PEMFCs comprising of membrane electrode assemblies (MEAs) prepared using decal process [34]. In Brief, required amount of Pt/C catalyst was dispersed in the mixture of Nafion® ionomer, ethanol, water and propylene glycol with prerequisite composition of binder (Nafion® ionomer) to carbon ratio being 0.75. The resultant composition is ultrasonicated to form an ink. The prepared ink was dropped onto the pre-treated Teflon (washed with de-ionized water and acetone) sheets and the doctor blade of 120 μm thickness was run over the Telfon sheets pasted on the uniform surface bed of K control coater procured from RK print coat instruments Ltd. U.K. Model No. K303 Multi-coater. During this, the catalyst ink is uniformly coated on the surface termed as catalyst coated Teflon sheets (CCTS). The resulting CCTS were dried at 130 °C for 4 h under the vacuum. After drying, CCTS were cut to the desired size, sandwiched with the pre-treated membrane and hot-pressed at a pressure of 40 kg cm−2 at 130 °C for 8 min. After cooling, the Teflon sheets were carefully peeled off from either side to form catalyst coated membrane (CCM). The CCMs were again sandwiched between gas diffusion layers (SGL DC-35) and hot-pressed at a pressure of 20 kg cm−2 at 130 °C for 2 min to form the overall MEA. The Pt loading calculated was around 0.21–0.22 mg cm−2 at both anode and cathode. The detailed process schematic for the fabrication of MEA is represented in Fig. S1. 2.8. Hydrogen cross-over measurements The electrochemical measurement of fuel permeation is used to determine the hydrogen cross over rate. The method is based on measuring the mass-transfer limiting current which is directly proportional to the concentration of hydrogen permeating from anode to cathode which gets oxidized under mass-transfer limiting conditions. The hydrogen crossover was determined by linear sweep voltammetry (LSV) technique using a potentiostat (Biologic Science Instruments model: VMP3B-20) at room temperature (25 °C) as described elsewhere [35]. In Brief, the cells were purged with hydrogen at the anode and nitrogen at the cathode. In this mode, the fuel cell anode serves as reference as well as counter electrode and the fuel cell cathode act as the working electrode. Hydrogen supplied to the anode permeates through the membrane and reaches the cathode where it is electrochemically oxidized. The detected current resulting from the oxidation of molecular hydrogen at fuel cell cathode is termed as hydrogen-crossover limiting current. For LSV, the potential was scanned from 0.1 V to 0.4 V at a sweep rate of 2 mV s−1. The crossover current was determined at the

129

steady-state voltage range between 0.3 V and 0.35 V. Finally, permeability coefficient for hydrogen crossover was calculated using the following equation [36].

Ki ¼

J max L nFP

ð4Þ

where Jmax is the maximum current density (mA cm−2), L is the thickness of the membrane (μm), n is the number of electrons involved in oxidation process, F is the Faraday constant (96,500 C mol−1) and P is the applied pressure (Pa). 2.9. Fuel cell performance evaluation and durability studies MEAs were coupled with Teflon gas-sealing gaskets and placed in single-cell test fixtures with parallel serpentine flow-field machined on graphite plates. At a cell temperature of 80 °C, galvanostatic polarization data were obtained using Biologic fuel cell test station (Model: FCT50/150S) at 100% RH with gaseous H2 and O2 fed to the anode and cathode of the PEFCs respectively. All the MEAs were evaluated with an active area of 25 cm2 in PEFCs. After the cell polarization, electrochemical impedance analysis is performed to record the cell resistance at 0.6 V between the frequencies 10 kHz to 100 mHz. The optimized blend membrane along with pristine sPEEK membrane and Nafion-212 was subjected to stability test by measuring the open circuit voltage (OCV) as function of time (t) for 50 h. The measurements were done at start–stop cycles. 3. Results and discussion 3.1. Sulfonation of POP POP is an extremely hydrophobic polymer and the use of POP as such to form polymer electrolyte in H2\\O2 fuel cells may not bring the appreciable improvement in the performance due to the deficiency of ion conducting groups. Sulfonation is the most common approach to impart ion conducting groups on the polymers. Hence POP is sulfonated with conc. H2SO4 through electrophilic substitution and the extent of sulfonation is controlled by selecting the optimum reaction time reported by Wycisk et al. [24], to avoid excessive swelling and dissolution of the polymer. Ion exchange capacity and degree of sulfonation are the important evidences to confirm the sulfonation of POP. In the present study, ion exchange capacity for sPOP is observed as 1.4 meq·g−1 and to further confirm the sulfonation, elemental analysis of POP and sPOP is carried out using CHNS analyser and reported in Table 1. For sPOP, the content of sulphur is 1.41% which indicates that the \\SO3H group attached to the aromatic ring of POP and degree of sulfonation calculated from CHNS analysis (Table 1) is found to be 42% suggesting the sulfonation of POP. Fig. 1 (a) shows the FT-IR analysis of POP and sPOP, wherein for both the spectra, the peaks appeared at 1068 cm−1 and 1160 cm−1, attributed to P-OAr and P_N stretching vibrations respectively and the peak at 940 cm−1 is due to P\\N\\P stretching. Further for the typical FT-IR spectra of sPOP, apart from all the characteristic peaks of POP, two more additional peaks at 1094 cm− 1 and 1125 cm−1are observed which are due to the symmetric and asymmetric stretching vibrations of O_S_O of sulfonic acid group [14,26]. Table 1 Elemental analysis of POP and sPOP. Sample name

Sample Carbon Hydrogen Nitrogen Sulphur Degree of weight (mg) (%) (%) (%) (%) sulfonation (%)

POP sPOP

3.18 5.98

62.14 56.83

4.401 4.423

6.017 5.532

– 1.401

– 42

130


[37]. On the other hand, elongation of blend membranes decreases compared to pristine sPEEK which may be due to the dominant inorganic backbone of sPOP which may affect the flexibility of polymeric chain. Fig. 2 shows the TG profiles for sPOP, pristine sPEEK and sPOP-sPEEK blend membranes. Prior to TG analysis, all the membrane samples were pre-heated in vacuum oven at 70 °C for 12 h to remove physically adsorbed water. In case of sPOP, major weight loss is observed from 220 °C to 450 °C which includes the decomposition of sulfonic acid and phenoxy groups and the degradation after 450 °C is attributed to the degradation of polymer backbone. Interestingly, pristine sPEEK and sPOP-sPEEK membranes show similar degradation pattern with three steps of weight loss. The first being in the range of 30–150 °C, due to the loss bound water wherein sPOP (3 wt.%)-sPEEK blend membrane shows slightly higher weight loss due to higher bound water content than pristine sPEEK and is in good agreement with water uptake values. Second loss observed between 280 and 400 °C, is attributed to the desulfonation of sPEEK and sPOP and the third loss observed from 460 to 630 °C, may be due to the degradation of main chain of sPEEK and sPOP [12]. It is observed that, compared to pristine sPEEK membrane, the blend membrane has relatively higher weight losses even in the second stage (desulfonation) attributing to the presence of higher sulfonic acid content than pristine sPEEK. These results are similar to the one as reported in the literature [38]. 3.3. Morphology of the membranes Fig. 1. (a) FT-IR spectra for POP and sPOP, (b) FT-IR spectra for pristine sPEEK and sPOPsPEEK blend membranes.

3.2. Characterization of blend membranes FT-IR spectra for pristine sPEEK and sPOP (3 wt.%)-sPEEK membranes are represented in Fig. 1(b) and the characteristic peak values corresponding to their functional groups are given in Table S1. In the spectra of sPEEK, the typical bands observed at 1020, 1077 and 1216 cm−1 are attributed to symmetric and asymmetric stretching vibrations of O_S_O from the sulfonic acid attached to the aromatic ring [37]. In the spectra of sPOP-sPEEK, almost all the peaks correspond to sPEEK are retained. It is important to note that the bands corresponding to sulfonic acid group are shifted to slightly higher wavelengths in blend membranes with reduced intensity compared to pristine sPEEK due to the interaction between sulfonic acid group (acid) and nitrogen moiety of sPOP (base) in the blend membrane. It is noteworthy that P\\N\\P stretching vibration of sPOP shifts from 940 to 974 cm− 1 when blended with sPEEK, may be due to the interaction of lone pair electrons of nitrogen with sulfonic acid groups of sPEEK through hydrogen bonding. Tensile strength and elongation of sPOP-sPEEK blend membranes are measured and compared with pristine sPEEK membrane as shown in the Table 2. It is noteworthy that tensile strength of the blend membranes increase as the content of POP increases in the membrane, which can be attributed presumably to the interfacial hydrogen bonding and ionic crosslinking between acid–base pairs in the blend membrane

Fig. 3 shows the surface and cross-sectional morphologies of sPOPsPEEK blend (c and d) along with pristine sPEEK membranes (a and b). From the surface morphology of blend membrane (Fig. 3c), it is seen that sPOP incorporation in sPEEK forms distinct phase separation whereas smooth surface is seen for pristine sPEEK (Fig. 3a). Interestingly, it is also seen from the surface morphology that sPOP molecules are self-organised into globular morphologies (Fig. 3c) with good interconnectivity, similar to invert missile structure of Nafion® [4]. Further, cross-sectional images show that dense structure of the membranes (Fig. 3b and d) is well maintained without any pores and aggregation of sPOP. It is to be noted that, these globular morphologies of sPOP can act as water reservoirs that contributes towards enhanced proton conductivity for blend membranes. The self-organization of the sPOP molecules is due to the difference in polarities between hydrophilic (polar) and hydrophobic (nonpolar) moieties, which induces efficient phase separation between sPOP and sPEEK [39]. Eigen and Weller proposed that the acid–base interactions are bi-molecular in nature [40]. These interactions are possible between acidic and basic moieties in polymer solution via diffusion phenomenon, wherein ionic transport takes place between the encounter pairs to form strong hydrogen bonding [41]. In the present scenario, SPEEK and the sPOP can act as acidic and basic moieties respectively, to form hydrogen bonded acid–base pairs via the available electron pair of \\N in sPOP with sulfonic acid groups of sPEEK. It is also proposed that this proton transfer is temperature

Table 2 Properties of the membranes.

Membrane type Nafion-212 Pristine sPEEK sPOP (2 wt.%)-sPEEK sPOP (3 wt.%)-sPEEK sPOP (4 wt.%)-sPEEK

Tensile strength (MPa)

Elongation (%)

Water uptake (%)

Activation energy (kJ mol−1)

16.4 ± 0.2 12.9 ± 0.4 16.1 ± 0.2

28.3 ± 0.3 8.3 ± 0.1 6.8 ± 0.4

19.3 ± 0.1 24.6 ± 0.1 31.8 ± 0.2

10.8 ± 0.2 12.2 ± 0.1 11.6 ± 0.1

19.1 ± 0.3

7.5 ± 0.1

34.3 ± 0.1

9.1 ± 0.3

21.3 ± 0.3

7.0 ± 0.2

35.8 ± 0.2

10.0 ± 0.1 Fig. 2. TGA for sPOP polymer, sPEEK and sPOP-sPEEK membranes.


dependent and the acid–base interactions transforms in to electrostatically interacted acid–base pairs. It is presumed that sPOP molecules are self-organized and sulfonic acid moieties of the sPOP concentrates at the center of the globular structure and exposes the N-back bone to the outer surface to form hydrogen bond with sPEEK. To further confirm the above proposed phenomenon, elemental mapping is done for the blend membrane to verify the globular structure formed by sPOP molecules. It is evident that the\\P of sPOP molecules form the globular morphology center and\\N of the backbone is left for hydrogen bonding with sulfonic acid groups of sPEEK. In addition, carbon, oxygen and sulphur are uniformly distributed which suggest the formation of globular morphology due to the presence of sPOP. Similar type of globular morphology is also observed for acid– base blend membranes based on polyimides [42]. Atomic force

131

microscopy analysis for the membranes further confirms the above observation. Fig. 4 shows topography and phase images of pristine sPEEK and sPOP-sPEEK-blend membranes, wherein the light regions are assigned to the hydrophilic sulfonic acid groups and the dark regions are assigned to the hydrophobic polymer backbone [36,43]. The blend membrane exhibited distinct hydrophilic/hydrophobic phase providing facile proton transport along the interface [39]. 3.4. Small angle X-ray scattering (SAXS) analysis SAXS analysis is done to understand the size of the ionic clusters present in the aforesaid membranes. Fig. 5 shows the SAXS analysis for pristine sPEEK and sPOP (3 wt.%)-sPEEK membranes, wherein both the membranes exhibits clear ionomer peak at 0.2–0.21 A−1 range

Fig. 3. Surface and cross-sectional SEM morphology for (a & b) pristine sPEEK and (c & d) sPOP-sPEEK blend membrane, (e–j) Elemental mapping for sPOP(3 wt.%)-sPEEK blend membrane.

132


Fig. 4. AFM topography and phase images for (a & b) pristine sPEEK and (c & d) sPOP (3 wt.%)-sPEEK blend membranes.

confirming the existence of ionic clusters [44,45]. By incorporation of sPOP in sPEEK, the peak shifts towards lower q value suggesting the larger ionic cluster in blend membrane in comparison with pristine sPEEK. The dmax values calculated from the Bragg's equation are 3.06 nm and 3.12 nm for pristine sPEEK and sPOP-sPEEK membranes respectively and these results are in good agreement with earlier literature reports [44,45]. It is noteworthy that larger ionic clusters of sPOPsPEEK blend membranes will in turn improve proton transport.

IEC, water uptake and proton conductivity are three critical factors in determining the performance of PEM. As expected, IEC of the blend membranes improved with the increased content of sPOP in sPEEK matrix (Fig. 6) due to the presence of sulfonic acid groups present in sPOP

in addition to the sulfonic acid groups of sPEEK. These results follow the similar trend of increased IEC values reported for sPEEK based membranes [46,47]. Further the optimized blend membrane sPOP (3 wt.%) is also subjected to IEC test at different time intervals till 130 h (longtime measurements) to assess any leaching of the sulfonic acid groups of sPOP. IEC of the sPOP (3 wt.%)-sPEEK membrane is measured at every interval of 24 h as shown in the inset to Fig. 6. As observed, IEC of the sPOP (3 wt.%)-sPEEK membrane did not show any change in IEC over a period of 130 h which clearly confirms no leaching of sulfonic acid groups from the blend membrane. This is possibly due to the strong interfacial hydrogen bonding and ionic crosslinking between acid–base pairs in the blend membrane. The water uptake is related to the proton conductivity and mechanical strength of the membrane. The presence of water in the membrane facilitates the proton transfer thereby increasing the ionic conductivity of the membrane. However, if the

Fig. 5. SAXS analysis for pristine sPEEK and sPOP (3 wt.%)-sPEEK blend membranes.

Fig. 6. IEC of pristine sPEEK and sPOP-sPEEK blend membranes. Inset to the figure shows the long-time measurements of IEC for sPOP (3 wt.%)-sPEEK blend membrane.

3.5. IEC, water uptake and proton conductivity


water uptake is too high, mechanical stability of the membrane is affected due to the high degree of swelling and may affect the overall performance of MEAs [48]. For blend membranes, water uptake is improved when compared with pristine sPEEK due to the enhanced hydrophilicity of the blend membranes with sPOP. Proton conductivity of the membranes is measured from 30 °C to 70 °C as shown in Fig. 7a. Higher proton conductivity is observed for blend membranes in comparison with pristine sPEEK due to i) contribution of additional ion conducting groups from sPOP providing well connected hydrophilic pathways for proton conduction and ii) sulfonic acid groups coupled with water molecules at the sPOP/sPEEK interfaces might form additional pathway for proton conduction [49,50]. In sPOP-sPEEK blend, water absorbed by hydrophilic regions help in dissociation of sulfonic acid groups and forms hydronium ions which diffuses through each hydrophilic domain well connected for the facile proton transport [51]. In case of higher addition of sPOP (4 wt.%), there is a phase segregation and reduced conductivity as seen in Fig. 7. However, in SPEEK matrix, these domains are not well connected which leads to less proton conductivity [52]. Aggregation of sulfonic acid groups into larger hydrophilic-ion clusters is possible with flexible side chains of sPOP in turn increasing the proton conductivity of sPOP-sPEEK blend [53,54]. In addition, due to the crosslinking between acidic and basic domains, the blend membrane exhibits higher proton conductivity due to the delocalization of protons from sulfonic acid (− SO3H) group [6]. + Basic character of sPOP favors the formation of\\SO− 3 ⋯NH ⋯ bridges leading to fast dissociation. Two mechanisms can be presumed to explain this unique proton transport in sPOP-sPEEK blends; (a) proton hopps across a whole array of water molecules [55] and (b) a vehicular transport mechanism where proton diffuses in solution in the form of hydrated protons [56].

Fig. 7. (a) Variation of proton conductivity with temperature for Nafion-212, pristine sPEEK and sPOP-sPEEK blend membranes and (b) Arrhenius plots for Nafion-212, pristine sPEEK and sPOP-sPEEK blend membranes.

133

Scheme 1. Acid–base interaction between sPOP and sPEEK.

The high proton conductivity of blend membranes of sPOP-sPEEK is attributed to both the mechanisms. The vehicle-type mechanism is possi+ ble in the hydrophilic regions where protons (H3O+, H5O+ 2 , H9O4 ) were transported between the clusters of the sulfonic acid groups. While, the Grotthuss-type mechanism is possible in the regions where there is weak acid–base interactions (SO3 ⋯ HN and SO3H ⋯(H2O)n ⋯ N) between the sulfonic acid groups and nitrogen atoms through hydrogen bonding in sPOP-sPEEK [57,58]. The presumable interaction between the sPOP and sPEEK is represented in Scheme 1. Variation of proton conductivity in relation to acid–base interactions is also studied by Vito Di Noto et.al. wherein Nafion® is doped with basic dopants such as ZrO2, HfO2 and ZrHf. Proton conductivity of these organic–inorganic hybrid membranes increases by forming acid–base interactions between sulfonic acid groups of Nafion® and nano-additives due to the proton migration through delocalized bodies [6,25,39,58,59]. Similarly in the present study, the proton migration and delocalization in hydrophilic and hydrophobic phase can be assumed for improved proton conductivity of blend membranes of sPOP-sPEEK. Fig.7 (b) shows the Arrhenius plots for all the membranes. The activation energy for proton transport in the membranes is calculated and represented in Table 2. It is noteworthy that the activation energy for sPOP-sPEEK blend membranes is lower than pristine sPEEK membrane for the proton transport [60,61]. This may be due to the facile ion

Fig. 8. Oxidative stability for pristine sPEEK and sPOP-sPEEK blend membranes.

134

R. Gutru et al. / Solid State Ionics 296 (2016) 127–136 Table 3 Air and hydrogen permeability for the membranes.

Fig. 9. H2 cross-over analysis for Nafion-212, pristine sPEEK and sPOP (3 wt.%) - sPEEK blend membranes.

transport in blend membranes due to the wider ionic channels as inferred from the SAXS analysis. However when the addition of sPOP is beyond 3 wt.%, ionic conductivity is slightly reduced due to dominant hydrophobic phase of POP than hydrophilic sulfonic acid groups [62]. 3.6. Oxidative stability

Membrane type

Thickness (um)

Average permeability of Air (Darcy's constant) (m2 × 10−17)

Pristine sPEEK sPOP (2 wt.%)-sPEEK sPOP (3 wt.%)-sPEEK sPOP (4 wt.%)-sPEEK Nafion-212

50 ± 5 60 ± 5

12.92 ± 0.02 12.73 ± 0.01

3.0 –

60 ± 5

12.13 ± 0.01

1.0

50 ± 5

11.44 ± 0.02

–

50 ± 5

22.10 ± 0.02

3.8

Permeability constant for hydrogen (mol m−1 s−1 × 10−14)

compared to Nafion® and pristine sPEEK membranes where addition of sPOP would eventually restrict the fuel cross over as the molecular hydrogen is made to follow tortuous path across the blend membranes [35]. Permeability coefficient for hydrogen crossover is shown in Table 3 wherein Nafion®-212 shows highest permeability coefficient than other membranes. Lowest permeability co-efficient is observed for sPOP (3 wt.%)-sPEEK membrane suggesting the restricted hydrogen cross-over compared to sPEEK and Nafion-212. In addition to hydrogen, O2 crossover across PEM also has significant effect on MEA performance and durability. The diffused O2 form hydroxyl (HO•) and hydroxyl-peroxyl (HOO•) radicals affect the membrane structure and conductivity. Therefore O2 cross-over from the cathode to anode needs to be controlled [64–68]. To assess the resistance of the membranes to oxidant permeability, the air permeability for the membranes were done by using capillary flow porometer and the

Hydroxyl radicals produced at electrode-electrolyte interface during the course of fuel cell operation will attack the membrane resulting in pin holes that initiate membrane degradation process. Hence membrane subjected to fuel cells should be stable enough towards hydroxyl radicals in order to achieve stable fuel cell performance. In the present study, all the prepared membranes are subjected to Fenton's test to analyse their oxidative stability as shown in Fig. 8. It is noted that for pristine sPEEK, 20% of its initial weight is lost during Fenton's test for 5 h, whereas blend membranes relatively show lower weight loss and the optimized blend membrane i.e. sPOP (3 wt.%)-sPEEK has shown only 7.6% of loss from its initial weight. The better oxidative stability for the blend membranes compared to sPEEK is attributed to the relatively stable backbone of sPOP and ionic crosslinking between sulfonic acid groups of sPEEK and nitrogen moiety of sPOP. Nafion-212 has not shown any significant weight loss attributing to the high stability of \\C\\F backbone. 3.7. Fuel and oxidant cross-over studies The hydrogen cross-over leads to reduction in H2\\O2 fuel cell efficiency due to parasitic hydrogen consumption and mixed potentials on the cathode [63]. Hence the hydrogen crossover affects the life time of fuel cell and quantification of the crossover becomes important for membrane to be applied in fuel cell. Among all the techniques for the measurement of fuel permeation, monitoring the change of mass transfer limiting current by in situ electrochemical measurement is the most direct and easiest method [35]. The hydrogen molecules that cross over from anode to cathode through membrane get oxidized at the Pt/C catalyst. When the applied potential is high enough then all the hydrogen molecules get oxidized and the oxidation current reaches its maximum due to the mass-transfer limitation. This mass-transfer limiting current is proportional to the hydrogen permeation rate assuming the hydrogen concentration gradient within the PEM membrane is linear. Fig. 9 shows the hydrogen crossover curves of Nafion-212, pristine sPEEK and sPOP (3 wt%)-sPEEK membranes. From Fig. 9, it is clear that the cross-over oxidation current for sPOP-sPEEK blend membranes is lower when

Fig. 10. (a) PEMFC performance for pristine sPEEK, sPOP-sPEEK blend and Nafion-212 membranes and (b) Ohmic region plot for sPEEK, sPOP-sPEEK blend and Nafion-212 membranes.


average Darcy's permeability constant are given in Table 3. It is seen that the sPOP-sPEEK blend membranes have lesser Darcy's permeability constant values indicating the reduced air permeability for the same compared to sPEEK and Nafion® membranes. 3.8. PEMFC studies and durability evaluation Steady state cell polarization for pristine sPEEK and sPOP-sPEEK blend membranes along with Nafion-212 is represented in Fig. 10 (a) at ambient pressure. It is seen that MEAs with blend membranes exhibits improved Ohmic region which is a characteristic of enhanced ionic conductivity of the blend membranes. PEFC with sPOP (3 wt.%)sPEEK membrane exhibits peak power density of 986 mW cm−2 at a load current density of 2276 mA cm− 2, while the PEFC with pristine sPEEK attains a peak power density of only 713 mW cm−2 under a load current density of 1775 mA cm−2 at 70 °C with atmospheric pressure. However incorporation of sPOP beyond optimized content in the blend slightly reduces the cell performance may be due to the bulk resistance contribution towards the blend. Interestingly, Nafion-212 exhibits peak power density of 907 mW cm−2 which is similar to the performance of sPOP-sPEEK blend membranes. To understand the significance of the blend membranes, Ohmic regions of the polarization curves were plotted separately and presented in the Fig. 10 (b). Maximum total current of 31, 29 and 21 A is obtained for sPOP (3 wt.%)sPEEK, Nafion 212 and pristine sPEEK membrane respectively. Overall, sPOP (3 wt.%)-sPEEK performance is enhanced by 33% compared to pristine sPEEK membrane. An analysis of iR corrected voltage vs. current density plot is given in Fig. 11 (a). Activation region remains similar for both pristine sPEEK and sPOP-sPEEK based MEAs pointing towards the similar ORR activity of the catalyst layer with significant improvement

135

in the Ohmic region for sPOP-sPEEK blend. Fig. 10 (b) shows the improvement of Ohmic region in the blend is attributed to the membrane property as the other components contributing towards electronic conductivity remains similar. From this, it is conclusive that overall performance of sPOP (3 wt.%)-sPEEK membrane is on par with Nafion-212. The improvement in single cell performance is further supported by electrochemical impedance spectroscopy (EIS). The EIS spectra obtained at 0.6 V for the single cell for sPEEK and sPOP (3 wt.%)-sPEEK blend membranes are shown in Fig. 11 (b). It is to be noted that charge transfer resistance is almost similar for the MEAs, however there is a significant decrease in the high frequency intercept for sPOP (3 wt.%)-sPEEKblend membranes indicating the lower Ohmic resistance and higher ionic conductivity in comparison with other blend membranes. In the present study, attempts are made to understand the stability of the membranes in real fuel cell environmental conditions. The stability of optimized blend membrane along with the pristine sPEEK and Nafion212 is studied under OCV condition by measuring the cell voltage as function of time for 50 h at start–stop cycles of 10 h per day as shown in Fig. 11 (c). sPOP-sPEEK blend membrane and Nafion-212 shows higher OCV with less voltage drop in comparison to pristine sPEEK during the 50 h of operation, which is attributed to the less fuel and oxidant cross over as evident from hydrogen crossover and air permeability data described above. It is conclusive that sPOP-sPEEK blend membrane and Nafion-212 show higher stability after 50 h than pristine sPEEK. After 50 h of OCV operation, the MEAs containing pristine sPEEK and the optimized sPOP (3 wt.%)-sPEEK blend membranes are subjected to polarization studies to observe any change in the performance as shown in Fig. 11 (d). It is noteworthy that pristine sPEEK shows 25% of decrement in power density whereas the blend membrane shows hardly 6% decrement in comparison with its initial power density attributed to the

Fig. 11. (a) Polarization plots (iR corrected) for pristine sPEEK, sPOP(3 wt.%)-sPEEK and Nafion-212 membranes, (b) Nyquist plots for pristine sPEEK and sPOP (3 wt.%)-sPEEK blend membranes, (c)variation of open circuit voltage (OCV) as function of time for pristine sPEEK, sPOP(3 wt.%)-sPEEK blend and Nafion-212 membranes, (d) PEMFC performance for sPEEK and sPOP(3 wt.%)-sPEEK blend membranes, initial and after 50 h of OCV operation.

136


better oxidative stability and no leaching of sulfonic acid groups from the blend. 4. Conclusions sPOP-sPEEK blend membranes are fabricated to understand its electrolyte properties in PEMFCs. Acid–base interactions between sPOP and sPEEK facilitate new pathways for proton conduction and these interfacial interactions also improve the mechanical properties. Blend membranes shows distinct phase separation (hydrophilic/hydrophobic) leading to the improved properties in terms of proton conductivity and mechanical stability. SAXS analysis reveals that the ionic cluster size improves with sPOP content providing the facile proton transport. This also provides the better insight towards the stability of membranes in terms of water uptake, IEC and conductivity which has the major impact on the fuel cell performance. Long-time measurements of IEC for the blend membrane also confirm that there is no leaching of sulfonic acid groups from sPOP. Fuel and oxidant cross-over is seen restricted when globular characteristic of sPOP is implanted with sPEEK. Optimized sPOP-sPEEK blend membrane showed comparable peak power density to Nafion-212. Acknowledgements Authors acknowledge CSIR for the financial support under 12th Five Year Plan HYDEN programme (CSC 0122). Authors also thank ScientistIn-Charge, CECRI Madras Unit and Director, CSIR-CECRI Karaikudi for their support and encouragement. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ssi.2016.09.011. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19]

A. Shirdast, A. Sharif, A. Mahdi, Int. J. Hydrog. Energy 39 (2014) 1760. E. Ali Mistri, S. Banerjee, RSC Adv. 4 (2014) 22398. J. Li, M. Pan, H. Tang, RSC Adv. 4 (2014) 3944. L. Wang, B.L. Yi, H.M. Zhang, Y.H. Liu, D.M. Xing, Z.G. Shao, Y.H. Cai, J. Power, Sources 164 (2007) 80. R. Devanathan, Energy Environ. Sci. 1 (2008) 101. V.D. Noto, N. Boaretto, E. Negro, G.A. Giffin, S. Lavina, S. Polizzi, Int. J. Hydrog. Energy 37 (2012) 6199. A.K. Sahu, S. Pitchumani, P. Sridhar, A.K. Shukla, Bull. Mater. Sci. 32 (2009) 285. J. Tian, P. Gao, Z. Zhang, W. Luo, Z. Shan, Int. J. Hydrog. Energy 33 (2008) 5686. M. Lavorgna, L. Sansone, G. Scherillo, R. Gu, A.P. Baker, Fuel Cells 11 (2011) 801. K.T. Adjemian, S.J. Lee, S. Srinivasan, J. Benziger, A.B. Bocarsly, J. Electrochem. Soc. 149 (2002) A256. Z. Shao, P. Joghee, I. Hsing, J. Membr. Sci. 229 (2004) 43. S. Zhong, T. Fu, Z. Dou, C. Zhao, H. Na, J. Power, Sources 162 (2006) 51. E. Yan, J. Wang, Z. Jiang, H. Feng, L. Nie, T. Xu, X. Yang, X. Zhang, J. Mater. Chem. A 1 (2013) 11762. H. Beydaghi, M. Javanbakht, A. Bagheri, P. Salarizadeh, H.G. Zahmatkesh, S. Kashefi, E. Kowsari, RSC Adv. 5 (2015) 74054. K.D. Kreuer, J. Membr. Sci. 185 (2001) 29. D.M. Xing, B.L. Yi, F.Q. Liu, Y.Z. Fu, H.M. Zang, Fuel Cells 5 (2005) 406. E. Sengül, H. Erdener, R.G. Akay, H. Yücel, N. Bac, I. Eroglu, Int. J. Hydrog. Energy 34 (2009) 4645. M.L.D. Vona, Z. Ahmed, S. Bellitto, A. Lenci, E. Raversa, S. Licoccia, J. Membr. Sci. 296 (2007) 156. Y. Zhang, H. Zhang, X. Zhu, C. Bi, Phys. Chem. B 111 (2007) 6391.

[20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] [58] [59] [60] [61] [62] [63] [64] [65] [66]

[67] [68]

E. Quartarone, P. Mustarelli, Energy Environ. Sci. 5 (2012) 6436. R. Wycisk, P.N. Pintauro, J. Membr. Sci. 119 (1996) 155. Q. Guo, P.N. Pintauro, H. Tang, S. O'Connor, J. Membr. Sci. 154 (1999) 175. R. Wycisk, P.N. Pintauro, Adv. Polym. Sci. 216 (2008) 157. R. Wycisk, J.K. Lee, P.N. Pintauro, J. Electrochem. Soc. 152 (2005) A892. V.D. Noto, M. Piga, E. Negro, G.A. Giffin, S. Polizzi, T.A. Zawodzinski, RSC Adv. 3 (2013) 18961. S.G. Peera, S. Meenakshi, K.H. Gopi, S.D. Bhat, P. Sridhar, S. Pitchumani, RSC Adv. 3 (2013) 14048. S. Mohanapriya, S.D. Bhat, A.K. Sahu, S. Pitchumani, P. Sridhar, A.K. Shukla, Energy Environ. Sci. 2 (2009) 1210. G. Rambabu, S.D. Bhat, Chem. Eng. J. 243 (2014) 517. G. Rambabu, S.D. Bhat, Electrochim. Acta 176 (2015) 657. S. Mohanapriya, S.D. Bhat, A.K. Sahu, A. Manokaran, R. Vijayakumar, S. Pitchumani, P. Sridhar, A.K. Shukla, Energy Environ. Sci. 3 (2010) 1746. W. Ma, C. Zhao, J. Yang, J. Ni, S. Wang, N. Zhang, H. Lin, J. Wang, G. Zhang, Q. Li, H. Na, Energy Environ. Sci. 5 (2012) 7617. A. Chromik, J.A. Kerres, Solid State Ionics 252 (2013) 140. S. Alwin, S.D. Bhat, A.K. Sahu, A. Jalajakshi, P. Sridhar, S. Pitchumani, A.K. Shukla, J. Electrochem. Soc. 158 (2011) B91. M.S. Saha, D.K. Paul, B.A. Peppley, K. Karan, Electrochem. Commun. 12 (2010) 410. C. Felice, S. Ye, D. Qu, Ind. Eng. Chem. Res. 49 (2010) 1514. X. Li, D. Chen, D. Xu, C. Zhao, Z. Wang, H. Lu, H. Na, J. Membr. Sci. 275 (2006) 134. J. Wang, J. Liao, L. Yang, S. Zhang, X. Huang, J. Ji, J. Membr. Sci. 415-416 (2012) 644. J.C. Tsai, H.P. Cheng, J.F. Kuo, Y.H. Huang, C.Y. Chen, J. Power, Sources 189 (2009) 965. G.A. Giffin, G.M. Haugen, S.J. Hamrock, V.D. Noto, J. Am. Chem. Soc. 135 (2013) 822. M. Eigen, W. Kruse, L. De Maeyer, Prog. React. Kinet. 2 (1964) 285. T. Steiner, Angew. Chem. 114 (2002) 50. W. Jang, S. Sundar, S. Choi, Y. Shul, H. Han, J. Membr. Sci. 280 (2006) 321. F. Wang, M. Hickner, Y.S. Kim, T.A. Zawodzinski, J.E. McGrath, J. Membr. Sci. 197 (2002) 231. D.X. Luu, E.B. Cho, O.H. Han, D. Kim, J. Phys. Chem. B 113 (2009) 10072. S. Min, D. Kim, Solid State Ionics 180 (2010) 1690. Z. Jiang, X. Zhao, Y. Fu, A. Manthiram, J. Mater. Chem. 22 (2012) 24862. Y. Heo, H. Im, J. Kim, J. Membr. Sci. 425–426 (2013) 11. S. Deabate, G. Gebel, P. Huguet, A. Morin, G. Pourcelly, Energy Environ. Sci. 5 (2012) 8824. Y. Li, G. He, S. Wang, S. Yu, F. Pan, H. Wu, Z. Jiang, J. Mater. Chem. A 1 (2013) 10058. Z. Li, G. He, Y. Zhao, Y. Cao, H. Wu, Y. Li, Z. Jiang, J. Power, Sources 262 (2014) 372. V.D. Noto, R. Gliubizzi, E. Negro, G. Pace, J. Phys. Chem. B 110 (2006) 24972. K.D. Kreuer, J. Membr. Sci. 185 (2001) 29. P. Xing, G.P. Robertson, M.D. Guiver, S.D. Mikhailenko, S. Kaliaguine, Polymer 46 (2005) 3257. Y. Gao, G.P. Robertson, M.D. Guiver, S.D. Mikhailenko, X. Li, S. Kaliaguine, Macromolecules 37 (2004) 6748. N. Agmon, Chem. Phys. Lett. 244 (1995) 456. A. Venkatnathan, R. Devanathan, M. Dupuis, J. Phys. Chem. B 111 (2007) 7234. L. Wu, C. Huang, J.J. Woo, D. Wu, S.H. Yun, S.J. Seo, T. Xu, S.H. Moon, J. Phys. Chem. B 114 (2010) 13121. V.D. Noto, N. Boaretto, E. Negro, P.E. Stallworth, S. Lavina, G.A. Giffin, S.G. Greenbaum, Int. J. Hydrog. Energy 37 (2012) 6215. V.D. Noto, M. Piga, G.A. Giffin, K. Vezzu`, T.A. Zawodzinski, J. Am. Chem. Soc. 134 (2012) 19099. G.M. Shashidhara, K.N. Kumar, Polym.-Plast. Technol. Eng. 49 (2010) 796. C. Zhao, Z. Wang, D. Bi, H. Lin, K. Shao, T. Fu, S. Zhong, H. Na, Polymer 48 (2007) 3090. T. Kim, J. Choi, S.C. Kim, J. Membr. Sci. 300 (2007) 28. S.S. Kocha, J.D. Yang, J.S. Yi, AICHE J. 52 (2006) 1916. V.A. Sethuraman, J.W. Weidner, A.T. Haug, S. Motupally, L.V. Protsailo, J. Electrochem. Soc. 155 (2008) B50. Y. Shao, G. Yin, Y. Gao, J. Power, Sources 171 (2007) 558. R. Borup, J. Meyers, B. Pivovar, Y.S. Kim, R. Mukundan, N. Garland, D. Myers, M. Wilson, F. Garzon, D. Wood, P. Zelenay, K. More, K. Stroh, T. Zawodzinski, X.J. Boncella, J.E. McGrath, O.M. Inaba, K. Miyatake, M. Hori, K. Ota, Z. Ogumi, S. Miyata, A. Nishikata, Z. Siroma, Y. Uchimoto, K. Yasuda, K. Kimijima, N. Iwashita, Chem. Rev. 107 (2007) 3904. J. Yu, B. Yi, D. Xing, F. Liu, Z. Shao, Y. Fu, H. Zhang, Phys. Chem. Chem. Phys. 5 (2003) 611. D.A. Schiraldi, Polym. Rev. 46 (2006) 315.