Sulfonated poly(etheretherketone) and sulfonated polyvinylidene fluoride-co-hexafluoropropylene based blend proton exchange membranes for direct methanol fuel cell applications

Abstract

In this work, blend membranes based on sulfonated poly(etheretherketone) (SPEEK) and sulfonated polyvinylidene fluoride-co-hexafluoropropylene (SPVDF-co-HFP) with a 10 wt% concentration of a polymer blend in solution were prepared by a solution casting method. The PEEK and PVDF-co-HFP were sulfonated with a direct sulfonation method to reach sulfonation degrees of 68% and 31%, respectively. The different weight ratios of the blend polymers were synthesized and a weight ratio of 20 wt% PVDF-co-HFP and SPVDF-co-HFP exhibited the best properties in terms of a few essential membrane properties such as membrane swelling, liquid uptake, thermal and mechanical stability, methanol permeability, proton conductivity and DMFC performance. The prepared blend membranes presented excellent swelling and methanol barrier properties while the proton conductivities were acceptable. Also, the effect of the casting solvent on the morphology and properties of the prepared membranes was investigated. The membranes cast on N-methyl-2-pyrrolidone (NMP) showed the best results compared to the others. The SPEEK/SPVDF-co-HFP blend membrane cast in NMP had good interaction and showed the highest proton conductivity of 35.7 S cm⁻¹ for 10 wt% SPVDF-co-HFP and the highest power density of 43.02 mW cm⁻² for 20 wt% SPVDF-co-HFP, among all of the prepared membranes, at room temperature.

---

Introduction

Today with the worldwide increasing attention for clean, environment friendly and high efficiency power sources, fuel cells play an important role as renewable sources of energy. Among the various types of fuel cells, proton exchange membrane fuel cells (PEMFCs) are most investigated for stationary, transportation and portable applications.^1,2 Direct methanol fuel cells (DMFCs) are one of the most applied categories of PEMFC, that have a promising future as power supplies for portable electronic devices such as mobile phones, laptops etc.³ DMFCs have several advantages such as a simplified system design, convenient fuel storage, high energy density at low operating temperature and independence of the reformer.^4,5 The proton exchange membrane (PEM) is an essential component of DMFC that should have desired characteristics such as low methanol permeability, high proton conductivity, chemical and physical stability and long durability.⁶ Nowadays perfluorosulfonic acid (PFSA) membranes, e.g., Nafion, are the most used membranes in DMFCs due to their excellent thermal and chemical stability as well as high proton conductivity and also high mechanical strength. Nafion membrane is one of the best low temperature operating polymer electrolyte membranes for DMFCs application and exhibits good performance and durability during fuel cell operation.⁷ However, Nafion membranes have some drawbacks such as low proton conductivity at temperature above 80 °C, high methanol crossover and high cost.⁸

In order to increase the proton conductivity and diminution the effect of methanol crossover on DMFC performance, researchers focused on modifying Nafion membranes.^9–11 In other hand, several efforts have been made to produce new alternative membranes for DMFCs. Different polymeric membranes have been modified for use in DMFC such as poly(vinyl alcohol) (PVA),¹² polybenzimidazole (PBI),¹³ sulfonated polyethersulfone (sPES),¹⁴ sulfonated poly(phenylene oxide) (sPPO),¹⁵ sulfonated polyimide (sPI)¹⁶ and sulfonated poly(etheretherketone) (SPEEK).¹⁷

Among the various PEMs that have been reported, SPEEK is considered as a promising candidate for DMFC applications because of its good thermal and mechanical stability, low cost and appropriate proton conductivity.¹⁸ SPEEK membranes have hydrophobic backbone with hydrophilic branches and also hydrophobic/hydrophilic domains in this polymer are smaller than theirs in Nafion. Thus, methanol permeability and electro-osmotic drag reduce drastically. The proton conductivity of SPEEK membranes has direct relation to degree of sulfonation (DS). The DS of SPEEK can be controlled by acid concentration, temperature and sulfonation time. Sulfonated membranes such as SPEEK with high DS, in spite of high proton conductivity, have much swollen in aqueous solution and are even soluble in aqueous methanol solution. This drawback may decrease its performance in DMFCs.^19,20 To eliminate this problem, in recent years, researchers focused on modification of SPEEK with blending with other polymers. Recently, our group prepared novel blend membrane based on SPEEK and PVA for DMFC applications.¹⁷ With blending of SPEEK and PVA proton conductivity of membrane was decreased. To eliminate this problem sulfonated graphene oxide/Fe₃O₄ nanosheets were added to blend membranes to improve proton conductivity and performance of membranes. Molla et al. modified SPEEK membrane by blending it with PVA and poly(vinyl butyral) (PVB) for use in DMFCs.²¹ The prepared membranes presented excellent mechanical and methanol barrier properties while proton conductivity was low. Prasad et al. have developed a blend membrane based on SPEEK and polyvinylidene fluoride-co-hexafluoropropylene (PVDF-co-HFP) in the presence of cloisite 30B for DMFC, which exhibited higher proton conductivity and methanol permeability than pristine SPEEK.²² Inan et al. prepared a blend membrane of SPEEK and PVDF-co-HFP and discussed the effect of type and molecular weight of PVDF-co-HFP on the membrane properties.²³ The results showed that with increase of PVDF content proton conductivity, water uptake, methanol permeability and ion exchange capacity (IEC) decrease.

PVDF-co-HFP is a hydrophobic copolymer used for blending with SPEEK that its fluorocarbon base and high C–F chemical bond provide desired characteristics such as good mechanical strength, thermal and chemical stability for blend membranes.²⁴ PVDF-co-HFP is inherently resistant to methanol crossover and this property helps to increase performance of blend membrane.²⁵ In addition, hydrophobic nature of this copolymer can efficiently reduce water uptake and membrane swelling.²⁶ Also, the absence of –SO₃H groups in structure of PVDF-co-HFP can decrease proton conductivity of blend membrane. To overcome these drawbacks sulfonation of PVDF-co-HFP is one of the best ideas to increase the hydrophilic nature of PVDF-co-HFP copolymer and subsequently improves water uptake and proton conductivity of blend membranes.

In this work, the SPEEK/PVDF-co-HFP and SPEEK/SPVDF-co-HFP blend membranes were prepared by solution casting method in different solvents. SPEEK based membranes were prepared using few solvents such as N,N-dimethyl acetamide (DMAc), N-methyl-2-pyrrolidinone (NMP) and dimethylsulfoxide (DMSO). Each kind of mentioned solvents can affect the properties of blend membranes. The morphology and proton conductivity of membranes depend on the different casting solvents. The solvents can form hydrogen bonding with –SO₃H groups of polymer, which reduces the mobility of polymeric chains and the dissociation of sulfonic acid groups of polymers, resulting in the lower proton conductivity. The physical properties, thermal and oxidative stability, water and methanol uptake, swelling ratio, electrochemical characteristics and performance of prepared blend membranes were investigated.

Experimental

Materials

Poly(etheretherketone), PEEK, (MW = 20 800) was obtained from Sigma-Aldrich. Polyvinylidene fluoride-co-hexafluoropropylene (PVDF-co-HFP) (MW = 455 000) was supplied by Sigma-Aldrich. N,N-Dimethyl acetamide (DMAc), N-methyl-2-pyrrolidinone (NMP) and dimethylsulfoxide (DMSO) used as solvents were purchased from Merck. Deionized water was used in all experiments.

Sulfonation of polymers

Sulfonation of PEEK was conducted in presence of sulfuric acid. In brief, dried PEEK (2 g) powder was first added into concentrated sulfuric acid (20 mL) under continuous stirring condition at room temperature. Then, obtained solution was stirred at 60 °C for 4 h under nitrogen atmosphere, cooled to room temperature, and then added into large excess of ice water under continuous mechanical agitation. The precipitate was washed with deionized water until pH was 7 and then dried at 70 °C overnight. The DS was calculated to be 68% using the acid–base titration method.^27–29

Sulfonation of PVDF-co-HFP was carried out according to following procedure: first, co-polymer pellets were dried in vacuum oven for 12 h at 60 °C. After that, 20 mL of chlorosulfonic acid was heated at 60 °C in a round-bottom flask and subsequently co-polymer pellets were added into acid solution carefully in continuous stirring condition. After 7 h black pellets were obtained and washed with 1,2-dichloroethane, methanol and deionized water, respectively, and finally dried in vacuum oven at 60 °C. The DS was determined to be 31% by acid–base titration method.²⁴

Preparation of blend membranes

Three types of SPEEK, SPEEK/PVDF-co-HFP and SPEEK/SPVDF-co-HFP membranes were prepared by solution casting method; these membranes are identified as M_S, M_SPx and M_SSPx respectively, where x represents the weight percentage of PVDF-co-HFP and SPVDF-co-HFP in blend membranes. To achieve the best solvent for preparing of membranes, M_SP10 and M_SSP10 were selected as representative membranes and nominated by M_SP10-y and M_SSP10-y that y denotes type of solvent. In order to prepare polymer solution with 10 wt% of the total weight of polymers, suitable quantity of dried SPEEK was dissolved in different solvents (DMAc, DMSO and NMP) separately in 60 °C for 6 h. After that several weight ratios of PVDF-co-HFP and SPVDF-co-HFP were dissolved in above solvents in the same condition, separately. Preparation of blend membranes was followed by adding PVDF-co-HFP and SPVDF-co-HFP solutions into SPEEK solution and stirred at 60 °C for 6 h. The concentration of the blend solution was kept constant at 10% w/v solution. The resulting solutions were casted onto a clean glass plate, successively dried at room temperature overnight and then 70 °C for 12 h and 120 °C for 2 h to remove any residual solvent and completion of probable cross-linking processes. The thickness of prepared membranes was measured about 60–100 μm in their dried state. It should be noted that, increase in PVDF and SPVDF contents of the blend membranes higher than 25 wt% was not possible, for the reason of weakness in mechanical stability and phase separation of blend membranes.

Characterization of membranes

The water uptake (WU) and methanol uptake (MU) of the prepared membranes were calculated in the following way: a cut piece from each membrane was dried in oven for 24 h. Then the membranes were soaked in distilled water or 2 M methanol in aqueous solution for 24 h and weighed (W_w). Next, the wet membranes were dried in the drying oven at high temperature and weighed again (W_d). The liquid uptake (LU) was calculated using the following equation:

The swelling ratio (SR) was calculated using the wet and dry thickness of membranes immersing in deionized water, designated as L_w and L_d, respectively. The membrane swelling was determined by following equation:

Titration technique was used to determine the ion exchange capacity (IEC) of the prepared membranes. First, the membrane was immersed in 1 M NaCl solution at 25 °C for 24 h to exchange the H⁺ ions with the Na⁺ ions in the solution. Then, obtained solution was titrated with a 0.01 M NaOH solution using phenolphthalein as an indicator. The IEC values (meq g⁻¹) can be determined from the following equation

where, V_NaOH is the volume of NaOH solution (mL), C_NaOH is the concentration of NaOH solution (mol L⁻¹) and W_dry is the mass of the dry sample (g).

The proton conductivities of membranes were measured using an AC impedance spectroscopy at 25 °C and 100% relative humidity (RH). Before test, the fully hydrated membranes were kept in 1 M H₂SO₄ solution for 12 h. The proton conductivity (σ, S cm⁻¹) of membranes was calculated from the respective impedance data using the following equation:

where, L is the thickness of samples (cm), A is the surface area of samples (cm²), and R is the ionic resistance of samples (Ω).

Oxidative stability of prepared membranes was investigated by Fenton reagent (3 wt% H₂O₂ + 2 ppm FeSO₄) at 80 °C. The membrane was immersed in Fenton reagent until the membrane began to break.

Methanol permeability (P, cm² s⁻¹) of membranes was measured by using a diaphragm diffusion cell. The cell consisted of two identical compartments separated with test membrane. Compartment A was filled with 20 wt% methanol solution and the compartment B was filled with deionized water. Before test, the membranes were equilibrated in deionized water for 12 h. Both A and B compartments were under stirring during the experiment. The increase in the methanol concentration in the water compartment with respect to time was recorded using a density meter. The methanol permeability was calculated from the following equation:

where, C_A and C_B are the methanol concentration in compartment A and B (mol L⁻¹) respectively, L is thickness of the membranes (cm), A is the membrane area (cm²) and V_B is the volume of the compartment B (cm³).

The selectivity (S, S s cm⁻³) is an important parameter in DMFC and evaluates membrane performance. The selectivity of membranes can be obtained using the following equation:

where, σ is the membrane proton conductivity (S cm⁻¹) and P is the methanol permeability (cm² s⁻¹).

Small angle X-ray scattering (SAXS) technique was used to study the clustering structure of the ionic groups that are usually indicated by the existence of a scattering maximum in the polymer matrix. The prepared membrane was immersed in a 1.0 M CsCl solution for 24 h to exchange the protons of functional groups with the ionic cesium and then washed with deionized water, and dried in an oven at 90 °C for 24 h. The SAXS data was plotted as relative intensity versus scattering vector (q) according to eqn (7). The scattering profiles had relation to the scattering vector, which is a function of the scattering angle (θ);

q = (4π/λ)sin θ (7)

where, θ is half of the scattering angle (2θ) and λ is the X-ray wavelength (0.136 nm). The average mean separation distance of the membrane ion clusters was calculated from the following equation:

d = 2π/q (8)

where, d is the average ion cluster dimension of the membrane and q is the scattering vector.

Platinum (Pt)/ruthenium (Ru)/carbon (Pt/Ru/C, 2 mg (PtRu) per cm²) and Pt/C (1 mg (Pt) per cm²) were used as anode and cathode electrodes, respectively. The catalyst solution was prepared via mixing catalyst (20 wt% of metals), 5 wt% Nafion solution as an electrode ionomer, isopropyl alcohol (IPA), suitable amount of deionized water and glycerol. The catalyst ink was painted onto carbon cloth (E-tek, HT 2500-W), followed by drying in an oven at 80 °C, 120 and 320 °C for 2 h at each temperature. The membranes were sandwiched between two prepared electrodes and then hot pressed under pressure of 135 kg cm⁻² at 120 °C for 3 min to get the final MEA. The DMFC was operated at 30 °C with 1 M methanol with flow rate of 3 mL min⁻¹ at the anode side and pure oxygen with a flow rate of 300 mL min⁻¹ at the cathode side under ambient pressure.

Apparatus

The Fourier transform infrared (FT-IR) spectra were obtained using a Bruker Equinox with ultra-dry compressed air in the wave number range from 4000–400 cm⁻¹. The SAXS experiments were carried out by placing the sample cell in the path of the X-ray beam using a PANalytical-X'Pert PRO MPD instrument. The thermal stability of membranes was evaluated by thermogravimetric analysis (TGA) under nitrogen atmosphere using a Hi-Res TGA 2950 thermogravimetric analyzer with a temperature range of 25–600 °C at a heating rate of 20 °C min⁻¹. The morphology of the membranes was obtained by scanning electron microscope (SEM, TESCAN). Mechanical properties of the prepared membranes were determined at room temperature on a SANTAM DBBP-100 testing machine with an operating rate of 1 mm min⁻¹. Proton conductivity of prepared membranes was measured by the ac impedance spectroscopic technique, employing an Autolab potentiostat/galvanostat with a frequency range of 0.1 Hz to 1 MHz and voltage amplitude of 50 mV.

Results and discussion

Selection of the casting solvent

Fig. 1 shows surface and cross section SEM images of representative M_SP10 and M_SSP10 blend membranes prepared in different solvents. The morphology of membranes may change due to the solvent effects.³⁰ It is evident that the morphology of blend membranes casted from NMP solution is better than those of membranes casted from DMAc and DMSO solutions. This result may be from two possible reasons. First, strong interaction of DMSO and DMAc solvents with sulfonic acid groups of SPEEK likely prevent crosslinking of SO₃H groups of SPEEK with C–F and SO₃H groups of PVDF and SPVDF polymers, respectively.^31–33 Second, the high dielectric constant of DMSO (46.7) and DMAc (37.8) compared to NMP (32.2) makes the distribution and mixing of the polymers unsatisfactory, leading to phase separation and heterogeneous morphology of the formed membranes.³²

Fig. 1 (a), (b) and (c) Surface and (d), (e) and (f) cross-sectional images of M_SP10 membrane casted in DMAc, DMSO and NMP, respectively. (g), (h) and (i) Surface and (j), (k) and (l) cross-sectional images of M_SSP10 membrane casted in DMAc, DMSO and NMP, respectively.

The comparative WU, SR, MU and proton conductivity of blend membranes prepared in different casting solvents were investigated and presented in Table 1 for selecting the suitable solvent. The M_SP10-NMP and M_SP10-DMAc membranes showed higher hydrophilicity and methanol uptake than M_SP10-DMSO membrane. The higher water content of membrane led to higher methanol uptake.³⁴ A similar trend was observed for M_SSP10 membranes. This could be due to the strong interaction between the DMSO and DMAc solvents with sulfonic acid groups of SPEEK.^21,31 It is noteworthy that water uptake effects on proton conductivity of the membranes. The proton conductivity of blend membranes in different solvents is in the increasing order of DMSO, DMAc and NMP which is consistent with the sequence of the water uptake and swelling ratio of blend membranes as displayed in Table 1. With changing solvent from DMSO to NMP, proton conductivity increased from 23 to 43 S cm⁻¹. As a consequence, NMP is designated as the suitable solvent for membrane solution casting process in this paper.

Table 1 Properties of blend membranes prepared in different solution casting solvents

Membranes	WU (%)	SR (%)	MU (%)	σ (mS cm^−1)
MSP10-NMP	28.87	12.58	30.14	20.01
MSP10-DMAc	27.38	13.69	31.10	16.01
MSP10-DMSO	25.36	12.37	30.04	15.10
MSSP10-NMP	29.33	11.02	31.58	37.03
MSSP10-DMAc	28.54	12.47	32.1	35.71
MSSP10-DMSO	24.32	11.09	31.35	24.03

---

FTIR study

The FTIR spectra of M_S, M_SP20 and M_SSP20 membranes were shown in Fig. 2. All membranes showed the same characteristic peaks in 500 to 1700 cm⁻¹. The absorption peaks at 1023, 1074 and 1247 cm⁻¹ were assigned to the asymmetric and symmetric stretching vibration of O=S=O and stretching vibration of S=O on sulfonic acid (–SO₃H) groups and absorption peak at 1644 cm⁻¹ was assigned to the carbonyl band of SPEEK.²³ In M_SP20 and M_SSP20 membranes, the low intensity absorption peak around 2890 cm⁻¹ was assigned to the symmetric –CH₂ groups of PVDF-co-HFP based polymers.³⁴ In M_S membrane, the broad peak around 3400 cm⁻¹ was assigned to hydroxyl groups from sulfonic acid groups of SPEEK. The intensity of this peak severely decreased with addition of PVDF-co-HFP based polymers, because of direct condensation reaction between –OH of sulfonic acid groups of SPEEK and PVDF-co-HFP based polymers chains thus establishing cross-linking bonds.

Fig. 2 FTIR spectra of M_S, M_SP20 and M_SSP20 membranes.

Ion cluster dimension of membranes

SAXS analysis was used to study the morphology and clustering structure of the ionic groups in the M_S and M_SSP20 membranes.^35–37 As clearly showed in Fig. 3, a broad peak in the range of scattering vector of 1.7–4.0 nm⁻¹ demonstrated the existence of the ionic cluster in these membranes. It was found that the M_S and M_SSP20 membranes showed small-angle peaks at 2.86 and 2.91 nm⁻¹, respectively. Blending SPEEK with SPVDF-co-HFP shifted the ionomer peak to slightly higher q values and caused to a small size of clusters associated with interactions between the SPEEK and SPVDF-co-HFP functionalized groups.³⁸ The SAXS peak intensity increased in the case of M_SSP20 membrane due to increase in the number of scattering center at the inter-ionic domains distance.³⁹ The incorporation of SPVDF-co-HFP decreased the average dimension of the ionic clusters of the SPEEK membranes from 2.2 to 2.16 nm. From this result, it can be assumed that the decreased size of ionic clusters combined with the low swelling ratio of blended membrane caused to reduce methanol permeability through the membrane.⁴⁰

Fig. 3 SAXS patterns of M_S and M_SSP20 membranes.

Water uptake, swelling ratio and methanol uptake

The WU and SR are two crucial parameters for PEM fuel cells. In other hand, these properties are interrelated with proton conductivity and dimensional stability values of membranes.^41–43 In this regard, ionic clusters of hydrophilic sulfonic acid functional groups provide channels in polymer backbones and protons get transported through the membrane via formation of H₃O⁺ ions, either by hopping along bound water molecules as Grotthuss mechanism or by using free water molecules as vehicles (vehicular mechanism).^44–46 By this reason, water content provides an indirect measurement for proton conductivity of PEM. Increase in water uptake causes enhancement in proton conductivity. However, too much water absorption by membrane promotes its swelling ratio and leads to decrease in mechanical stability of membrane.⁴⁷ Fig. 4 and 5 shows WU and SR of prepared membranes in temperature range of 20–80 °C. For M_S membrane, WU and SR are measured to be 35% and 13.4% at 20 °C, respectively. In pristine SPEEK membrane, the hydrophilic and hydrophobic domains are small and hydrophilic domain of SPEEK can expand by hydration comfortably.⁴⁸ With incorporation of SPVDF-co-HFP and PVDF-co-HFP polymers into SPEEK polymer, both WU and SR decreased. Water uptake of M_SSPx membranes is higher than M_SPx membranes due to more hydrophilicity of SPVDF-co-HFP than PVDF-co-HFP. But, a little decrease in swelling of M_SSPx membranes than M_SPx membranes can be due to hydrogen bonding between SO₃H groups of SPEEK and SPVDF-co-HFP chains. By increasing temperature until 80 °C both WU and SR were raised due to increasing in chain mobility and trapping of water molecules in voids formed in the membrane.^22,23 But in above 80 °C, M_S membrane was dissolved. It is important to note that hydrogen bonding between SO₃H groups of SPEEK and SPVDF-co-HFP polymers prevents excessive swelling of membrane and increases stability of blend membrane.

Fig. 4 Water uptake of (a) M_SPx and (b) M_SSPx membranes at different temperatures and comparison with M_S membrane.

Fig. 5 Swelling ratio of (a) M_SPx and (b) M_SSPx membranes at different temperatures and comparison with M_S membrane.

MU capacity is the affinity of a membrane toward methanol that provides an indirect measure of methanol permeability.^26,49 As shown in Fig. 6, MU value for each membrane was determined in 0.5, 1, 2 and 4 M of methanol solution. It can be observed that the MU values were higher compared to the corresponding WU values and MU increased with increasing methanol concentration. The main reason behind this observation is the higher molecular weight of methanol than water, the sorption of both water and methanol through the ionic cluster of the membranes and also affinity of membranes towards methanol.²⁴ In contrast, with incorporation of PVDF-co-HFP and SPVDF-co-HFP polymers into SPEEK matrix, blend membranes showed less affinity toward methanol due to hydrophobic nature of these polymers.

Fig. 6 Methanol uptake of (a) M_SPx and (b) M_SSPx membranes at different temperatures and comparison with M_S membrane.

IEC and proton conductivity

The IEC is a measure of the ability of ion exchanging and ion transferring of membranes. The IEC value plays an important rule for determination of proton conductivity of membrane.⁵⁰ The IEC value for prepared membranes was measured and results were presented in Table 2. Incorporation of PVDF-co-HFP and SPVDF-co-HFP in SPEEK matrix led to a decrease in IEC from 1.61 meq g⁻¹ to 0.92 and 1.28 meq g⁻¹ respectively, due to the decrease of ionic sites in the blend membranes. This can be attributed to the decrease in the amount of sulfonic acid groups in the structure of membranes and hydrophobic nature of blended membranes. It is noteworthy to mention that the IEC values for M_SSPx membranes are higher than M_SPx blend membranes due to presence of –SO₃H groups in the PVDF-co-HFP structure that provides proton transfer sites in M_SSPx blend membranes.²²

Table 2 IEC, proton conductivity (σ), methanol permeability (P), selectivity (S) and oxidative stability of prepared membranes

Membranes	IEC (meq g^−1)	σ (mS cm^−1)	P × 10^−7 (cm^2 s^−1)	S × 10^4 (S s cm^−3)	Oxidative stability (h)
MS	1.61	42.02	5.71	7.37	1.5
MSP10	1.20	16.01	3.63	4.44	2
MSP15	1.15	12.00	2.31	5.22	2.5
MSP20	1.10	8.02	1.10	7.27	3.5
MSP25	0.92	6.03	0.90	6.67	4
MSSP10	1.40	35.71	4.01	8.94	2
MSSP15	1.37	33.12	3.40	9.73	2
MSSP20	1.35	32.71	2.11	15.51	3
MSSP25	1.28	25.03	2.00	12.53	3.5

---

Proton conductivity is an essential factor for determining potential of PEM in DMFCs that is affected by the WU, IEC and DS of membranes. The proton conductivity of prepared membranes mainly occurs by two methods which are vehicular and Grotthuss mechanisms. In the vehicular mechanism proton transfers by the hydronium ions, while in the Grotthuss mechanism protons jump from one ionic site (SO₃⁻H₃O⁺) to any neighboring ionic structure. As shown in Fig. 7d, the sulfonic groups play an important role in proton transferring in the Grotthuss mechanism, while water molecules are essential in both mechanisms.⁵¹ Proton conductivity of prepared membranes at 20 °C was presented in Table 2. It is well known that proton conductivity of prepared membrane is specially affected by the ionic clusters and fluorocarbon matrix of PVDF-co-HFP and SPVDF-co-HFP.^22,52 The M_S membrane exhibited the highest proton conductivity (42 mS cm⁻¹ at 20 °C) among all of the prepared membranes due to existence of sufficient sulfonic acid groups in the polymer chain. As seen, the proton conductivity reduced by increasing PVDF-co-HFP and SPVDF-co-HFP amounts. Compared to M_SPx membranes, M_SSPx membranes have higher proton conductivity due to existence of sulfonic acid groups, hydrophilic nature of SO₃H groups of SPVDF-co-HFP and more connectivity of the hydrophilic domains in M_SSPx than M_SPx which create more direct pathways for protons as shown in Fig. 7. It was observed that, among all of the blend membranes, M_SSP10 membrane showed the highest conductivity of 37.03 mS cm⁻¹ while M_SP10 membrane had a proton conductivity of 20.01 mS cm⁻¹ at 20 °C. As mentioned in previous section, this difference in the proton conductivity could be due to the presence of sulfonic acid groups in the PVDF-co-HFP backbone.¹¹ The proton conductivities of all prepared membranes at different temperature (20–80 °C) are shown in Fig. 8. The proton conductivity of all prepared membranes increased with increasing the temperature due to increase in mobility of polymer chains and water molecules at higher temperatures.

Fig. 7 Schematic presentation of water and methanol crossover from (a) M_S, (b) M_SSPx and (c) M_SPx membranes and (d) proton transferring through membrane.

Fig. 8 Proton conductivity of (a) M_SPx and (b) M_SSPx membranes at different temperatures and comparison with M_S membrane.

Methanol permeability and selectivity

As known, methanol crossover through the polymeric membranes is a very important problem that changes the DMFCs performance.^11,24 Generally, membrane with high proton conductivity provides ionic cluster regions consisting of alcohols, protons and sulfonic acid groups for penetration of methanol from anode side that oxidized at cathode and thereby reduces the performance of the DMFC.⁵⁰ Although the methanol permeability of pristine SPEEK is acceptable for DMFC, further improvement can be made by blending with PVDF-co-HFP and SPVDF-co-HFP.²² The methanol permeability for pristine SPEEK and blended membranes was evaluated at 25 °C and presented in Table 2. The methanol permeability for prepared membranes was obtained in the range of 5.7 × 10⁻⁷ to 9.0 × 10⁻⁸ cm² s⁻¹. As shown in Fig. 7, methanol crossover in M_SPx and M_SSPx is lower than M_S membrane. The lower methanol permeability of blend membranes is due to hydrophobic nature of PVDF-co-HFP and SPVDF-co-HFP and low affinity towards methanol.⁵³ The PVDF-co-HFP based polymers are inherently resistant to methanol crossover. The methanol permeability results are conformity with MU results that were presented in Fig. 6. A gradual decrease in the methanol crossover is observed with an increasing amount of PVDF-co-HFP in the M_SPx and M_SSPx blend membranes. These results suggested that both proton conductivity and methanol permeability decreased with increasing of PVDF-co-HFP and SPVDF-co-HFP.

The selectivity of membranes was defined as the ratio between the proton conductivity and methanol permeability of the membranes. This parameter is essential to identify the optimum relation between proton conductivity and methanol permeability of membrane and employed to evaluate the potential performance of membrane in DMFCs.¹¹ A careful investigation of the data shown in Table 2 revealed that although methanol permeability of M_SPx blend membranes is lower than M_SSPx blend membranes, but proton conductivity values for M_SSPx blend membranes are much higher than M_SPx blend membranes. However, considering selectivity of membranes indicates that highest amount of selectivity was obtained in 20 wt% of PVDF-co-HFP and SPVDF-co-HFP in M_SPx and M_SSPx membrane. Then M_SP20 and M_SSP20 membranes were selected for more analysis.

Membrane morphology

The cross-section SEM images of prepared membranes are an effective way to study the compatibility in polymer blend systems. In Fig. 9a it is observed that the M_S membrane has relatively uniform and dense morphology. The influence of incorporation of PVDF-co-HFP and SPVDF-co-HFP into SPEEK matrix was shown in Fig. 9b and c respectively. In comparison to pristine SPEEK membrane, this possessed uniform cross-sectional morphology the M_SP20 membrane noticeably observable possessed lower miscibility and significant phase separation. A probable reason for this phenomenon is the hydrophobic nature and absence of hydrophilic group in PVDF-co-HFP structure. As shown in Fig. 9c M_SSP20 membrane demonstrated negligible phase separation. It could be clearly seen that sulfonation of PVDF-co-HFP increases compatibility of SPEEK and SPVDF-co-HFP polymers due to strong polar–polar interactions.²²

Fig. 9 The cross-sectional SEM images of (a) M_S (b) M_SP20 and (c) M_SSP20 membranes.

Mechanical and thermal stability

One of the important properties of PEMs is mechanical stability of membranes that strongly effects on stability, durability and performance of the prepared MEA. The effect of blending reaction on mechanical stability of prepared membranes was investigated. Table 3 shows the tensile strength (TS), elongation at break (E_b) and modulus of the prepared membranes at room temperature. From the results it is observed that the TS and modulus of the M_S membrane increased with blending with PVDF-co-HFP and SPVDF-co-HFP polymers. This may be resulted from two different reasons. First, PVDF-co-HFP based polymers have higher TS compared to SPEEK.²⁴ Then, addition of PVDF-co-HFP to M_S membrane increases TS of this membrane. Second, the mechanical stability of membranes is directly related with intermolecular interaction. Then, the TS and modulus of blend membranes increases compared to M_S membrane due to hydrogen bonding between –SO₃H groups of SPEEK and PVDF-co-HFP chains which leads to more compact polymer structure.³² The M_SSPx membranes showed higher TS and modulus than M_SPx membranes due to increase in hydrogen bonding between –SO₃H groups of SPEEK and SPVDF-co-HFP. Unlike TS, the E_b of M_S membrane decreased with blending with PVDF-co-HFP and SPVDF-co-HFP. Blending of SPEEK with PVDF co-polymers decreased flexibility of polymer chains and decreased E_b of the blend membranes.

Table 3 Mechanical properties of prepared membranes at room temperature

Membranes	TS (MPa)	Eb (%)	Modulus (MPa)
MS	17.3	11.37	723
MSP20	24.14	10.67	760
MSSP20	25.63	10.95	825

---

TGA was performed to investigate the effect of blending of SPEEK with PVDF-co-HFP based polymer on thermal stability of prepared membranes. Fig. 10 shows TGA curves of M_S, M_SP20 and M_SSP20 membranes. The initial weight loss at below 200 °C is because of loss of residual solvent and absorbed water. The second weight loss that occurred between 300 and 350 °C can be attributed to the loss of sulfonic acid groups of sulfonated polymers.²³ The weight loss observed above 400 °C is because of degradation of SPEEK, PVDF-co-HFP and SPVDF-co-HFP polymeric chains. As shown in Fig. 10, the thermal stability of membranes increased with addition of PVDF-co-HFP based polymers. The thermal stability improvement was because of increase of intermolecular interaction with formation of hydrogen bonding between PVDF-co-HFP based polymers and SPEEK chains. Then, thermal stability of M_SP20 and M_SSP20 blend membranes is better than M_S membrane.

Fig. 10 TGA curves of the M_S, M_SP20 and M_SSP20 membranes.

Oxidative stability

It is well known that hydroxyl (HO˙) or hydroperoxy (HOO˙) radicals formed on both anode and cathode side may attack to the polymer and caused chemical degradation.⁴⁴ In this regard, oxidative stability of prepared membranes were investigated using Fenton's reagent at 80 °C. As shown in Table 2 the chemical stabilities of prepared blend membranes were improved with the addition of PVDF-co-HFP and SPVDF-co-HFP. The M_S membrane began to break after 1.5 h while M_SPx and M_SSPx blend membranes at the same conditions were broken in 2–4 h and 2–3.5 h, respectively (Table 2). This is due to hydrophobic nature and chemical stability of PVDF-co-HFP and SPVDF-co-HFP polymers which decrease WU and SR of the blend membranes and consequence reduce the attack possibilities of free radicals to polar groups of polymer.^18,22,23

DMFC test

The prepared membranes were used to prepare MEAs to carry out DMFC performance test. The performance curves for DMFCs, comprising M_S, M_SP20 and M_SSP20 membrane at operating temperature of 30 °C, are shown in Fig. 11. Generally in DMFCs, the open circuit voltage (OCV) is depended on the methanol crossover.⁴⁹ The results showed that OCV of DMFC consists of M_S membrane increases with blending with PVDF base polymers, because of decrease in methanol crossover with addition of mentioned polymers. The DMFC with M_S, M_SP20 and M_SSP20 membranes gave a maximum power density of 36.12, 28.51 and 43.02 mW cm⁻² with highest current density of 194.1, 169.1 and 215.3 mA cm⁻², respectively. The results showed that effect of higher proton conductivity of M_S membranes overcome with lower methanol crossover of M_SP20 membrane and membrane performance decreases with blending of two polymers. On the other hand, the better performance of M_SSP20 membrane compared to M_S membrane is due to higher selectivity of M_SSP20 membrane.

Fig. 11 Current density–potential (I–V) and power density curves of the DMFC assembled with different prepared membranes at 30 °C.

Conclusions

In this study, the blend membranes based on SPEEK with varied PVDF-co-HFP and SPVDF-co-HFP based polymers contents were prepared by solution casting method using different solvents including DMAc, DMSO and NMP and effects of solvents on morphology and properties of membranes were investigated. Compared to membranes casted from DMSO and DMAc, membranes casted from NMP exhibited better proton conductivity, water uptake and morphology under ambient condition. The hydrogen bonding between sulfonated groups of SPEEK and PVDF-co-HFP based polymers chains confirmed by FTIR spectra. SEM images indicated that M_SP20 and M_SSP20 blend membranes have homogeneous cross-section morphology. The thermal and mechanical stability can be improved by formation of hydrogen bonding between SPEEK and PVDF-co-HFP based polymers during blending reaction. However, in view of the improving swelling properties, fuel barrier, selectivity, mechanical and thermal stability of SPEEK with blending with SPVDF-co-HFP, the M_SSP20 blend membrane showed good potential for DMFC application.

Acknowledgements

The authors are grateful to the Renewable Energy Research Center (Amirkabir University of Technology, Tehran, Iran) for the technical support of this work.

References

1. J. Pandey and A. Shukla, Solid State Ionics, 2014, 262, 811.
2. Q. He, J. Zheng and S. Zhang, J. Power Sources, 2014, 260, 317.
3. Y.-W. Chang and G. Shin, J. Ind. Eng. Chem., 2011, 17, 730.
4. Y. Iwai, S. Ikemoto, K. Haramaki, R. Hattori and S. Yonezawa, J. Supercrit. Fluids, 2014, 94, 48.
5. H. Ahmad, S. K. Kamarudin, U. A. Hasran and W. R. W. Daud, Int. J. Hydrogen Energy, 2011, 36, 14668.
6. H.-Y. Jung and J. W. Kim, Int. J. Hydrogen Energy, 2012, 37, 12580.
7. B. B. R. Silva, J. B. Soares, C. F. Malfatti and M. M. C. Forte, J. Membr. Sci., 2011, 374, 12.
8. H. Beydaghi, M. Javanbakht and E. Kowsari, Ind. Eng. Chem. Res., 2014, 53, 16621.
9. C. Bonis, D. Cozzi, B. Mecheri, A. D'Epifanio, A. Rainer, D. Porcellinis and S. Licoccia, Electrochim. Acta, 2014, 147, 418.
10. A. Alvarez, C. Guzman, S. Rivas, L. A. Godinez, A. Sacca, A. Carbone, E. Passalacqua, L. G. Arriaga and J. L. Garcia, Int. J. Hydrogen Energy, 2014, 39, 16686.
11. P. Kumar and P. P. P. Kundu, Electrochim. Acta, 2015, 173, 124.
12. H. Beydaghi, M. Javanbakht, H. Salar Amoli, A. Badiei, Y. Khaniani, M. R. Ganjali, P. Norouzi and M. Abdouss, Int. J. Hydrogen Energy, 2011, 36, 13310.
13. A. Shabanikia, M. Javanbakht, H. S. Amoli, K. Hooshyari and M. Enhessari, J. Electrochem. Soc., 2014, 161, 1403.
14. S. C. Gil, J. C. Kim, D. Ahn, J.-S. Jang, H. Kim, J. C. Jung, S. Lim, D.-H. Jung and W. Lee, J. Membr. Sci., 2012, 417, 2.
15. V. Deimede, G. A. Voyiatzis, J. K. Kallitsis, L. Qingfeng and N. J. Bjerrum, Macromolecules, 2000, 33, 7609.
16. K. Miyatake, H. Zhou, T. Matsuo, H. Uchida and M. Watanabe, Macromolecules, 2004, 37, 4961.
17. H. Beydaghi, M. Javanbakht, A. Bagheri, P. Salarizadeh, H. G. Zahmatkesh, S. Kashefi and E. Kowsari, RSC Adv., 2015, 5, 74054.
18. Y. Liang, C. Gong, Z. Qi, H. Li, Z. Wu and Y. Zhang, J. Power Sources, 2015, 284, 86.
19. G. Rambabu and S. D. Bhat, Chem. Eng. J., 2014, 243, 517.
20. H. Ilbeygi, A. F. Ismail, A. Mayahi, M. M. Nasef, J. Jaafar and E. Jalalvandi, Sep. Purif. Technol., 2013, 118, 567.
21. S. Molla and V. Compan, Int. J. Hydrogen Energy, 2014, 39, 5121.
22. M. Prasad, S. Mohanty and S. K. Nayak, RSC Adv., 2014, 4, 61178.
23. T. Y. Inan, H. Dogan, E. E. Unveren and E. Eker, Int. J. Hydrogen Energy, 2010, 35, 12038.
24. S. Neelakandan, D. Rana, T. Matsuura, A. Muthumeenal, P. Kanagaraj and A. Nagendran, Solid State Ionics, 2014, 268, 35.
25. J. Wootthikanokkhan and N. Seeponkai, J. Appl. Polym. Sci., 2006, 102, 5941.
26. S. Das, P. Kumar, K. Dutta and P. P. Kundu, Appl. Energy, 2014, 113, 169.
27. Z. Mai, H. Zhang, X. Li, C. Bi and H. Dai, J. Power Sources, 2011, 196, 482.
28. P. Salarizadeh, M. Javanbakht and S. Pourmahdian, Solid State Ionics, 2015, 281, 12.
29. J. Wang, H. Bai, H. Zhang, L. Zhao, H. Chen and Y. Li, Electrochim. Acta, 2015, 152, 443.
30. M.-S. Jun, Y.-W. Choi and J.-D. Kim, J. Membr. Sci., 2012, 396, 32.
31. L. Yu, H. M. Shen and Z. L. Xu, J. Appl. Phys., 2009, 113, 1763.
32. X. Liu, S. He, Z. Shi, L. Zhang and J. Lin, J. Membr. Sci., 2015, 492, 48.
33. W. Li, F. Zhang, S. Yi, C. Huang, H. Zhang and M. Pan, Int. J. Hydrogen Energy, 2012, 37, 748.
34. S. Mondal, S. Soam and P. P. Kundu, J. Membr. Sci., 2015, 474, 140.
35. D. X. Luu and D. Kim, J. Membr. Sci., 2011, 371, 248.
36. B. Yang and A. Manthiram, J. Power Sources, 2006, 153, 29.
37. S. Min and D. Kim, Solid State Ionics, 2010, 180, 1690.
38. Z. Jiang, X. Zhao, Y. Fu and A. Manthiram, J. Mater. Chem., 2012, 22, 24862.
39. J. M. Song, J. Shin, J. Y. Sohn and Y. C. Nho, Macromol. Res., 2011, 19, 1082.
40. Z. Jiang, X. Zhao and A. Manthiram, Int. J. Hydrogen Energy, 2013, 38, 5875.
41. M. Yoo, M. Kim, Y. Hwang and J. Kim, Ionics, 2014, 20, 875.
42. S. Zhong, X. Cui, Y. Gao, W. Liu and S. Dou, Int. J. Hydrogen Energy, 2014, 39, 17857.
43. X. Liang, G. Pan, L. Xu and J. Wang, Fuel, 2015, 139, 393.
44. G. Rambabu and S. D. Bhat, Electrochim. Acta, 2015, 176, 657.
45. Z. Di, Q. Xie, H. Li, D. Mao, M. Li, D. Zhou and L. Li, J. Power Sources, 2015, 273, 688.
46. Q. Xie, Y. Li, X. Chen, J. Hu, L. Li and H. Li, J. Power Sources, 2015, 282, 489.
47. A. Kumar, N. Hoon, D. Jung and J. Hee, J. Membr. Sci., 2014, 458, 128.
48. A. Mayahi, A. F. Ismail, H. Ilbeygi, M. H. D. Othman, M. Ghasemi and M. N. A. M. Norddin, Sep. Purif. Technol., 2013, 106, 72.
49. K. Dutta, S. Das, P. Kumar and P. P. Kundu, Appl. Energy, 2014, 118, 183.
50. T. Gumus, A. Ari and H. Deligoz, J. Membr. Sci., 2011, 376, 25.
51. K. Hooshyari, M. Javanbakht, A. Shabanikia and M. Enhessari, J. Power Sources, 2015, 276, 62.
52. K.-Y. Cho, J.-Y. Eom, H.-Y. Jung, N.-S. Choi, Y. M. Lee, J.-K. Park, J.-H. Choi, K.-W. Park and Y.-E. Sung, Electrochim. Acta, 2004, 50, 583.
53. A. Shabanikia, M. Javanbakht, H. S. Amoli, K. Hooshyari and M. Enhessari, Electrochim. Acta, 2015, 154, 370.

---