Direct polymerization of novel functional sulfonated poly(arylene ether ketone sulfone)/sulfonated poly(vinyl alcohol) with high selectivity for fuel cells

Abstract

Novel cross-linked membranes based on sulfonated poly(arylene ether ketone sulfone) containing pendant amino groups (Am-SPAEKS) and sulfonated poly(vinyl alcohol) (SPVA) as macromolecular cross-linkers were fabricated for potential application value in fuel cells. Water uptake, dimensional stability, thermal stability, gel fraction, methanol permeability and proton conductivity were thoroughly studied. Both atomic force microscopy (AFM) phase images and transmission electron microscopy (TEM) were used to evaluate the morphology of the membranes. The methanol permeability coefficients of the Cr-Am/S-80 membrane, Am-SPAEKS membrane cross-linked with 80 wt% SPVA, are only 0.32 × 10⁻⁷ cm² s⁻¹ at 20 °C and 0.96 × 10⁻⁷ cm² s⁻¹ at 60 °C, respectively. These values are much lower than that of the Am-SPAEKS membrane (9.86 × 10⁻⁷ cm² s⁻¹ at 20 °C and 19.91 × 10⁻⁷ cm² s⁻¹ at 60 °C). The proton conductivities of the cross-linked membranes were above 10⁻² S cm⁻¹ at 80 °C. In addition, the cross-linked membrane exhibited high selectivity, further proving its potential application value in proton exchange membrane fuel cells.

---

Introduction

Proton exchange membrane fuel cells (PEMFCs) have received extensive attention on account of simple system features, superb energy conversion rates and desirable operating temperatures.^1,2 The proton exchange membrane (PEM), as a core component, plays an important role in providing proton transfer and preventing fuel penetration between anode and cathode.^3,4

At the present, Nafion, perfluorosulfonic membranes, has been regarded as the benchmark PEM due to its good electrochemical stability and remarkable proton conductivities.^5,6 Nafion has numerous advantages such as desirable mechanical property, distinguished chemical stability, long service life and outstanding proton conductivity at 100% relative humidity.^7–9 However, Nafion still has some disadvantages.^10–13 For example, (i) the synthesis of Nafion is very complex, (ii) the proton conductivities of Nafion decrease rapidly with increasing temperature and decreasing humidity, (iii) the methanol crossover increases sharply at elevated temperature and high methanol concentration. Therefore, the wide application of Nafion has been limited.^14–18

At present, sulfonated aromatic polymers have been widely studied. Sulfonated aromatic polymers material has great potential application value, and it is expected to replace the Nafion in the fuel cell application. For example, sulfonated poly(arylene ether sulfone),^19–21 sulfonated poly(arylene ether ketone sulfone)s,^22,23 sulfonated polyimides,²⁴ sulfonated poly(arylene thioether sulfone)s^25,26 and sulfonated poly(ether ether ketone)²⁷ etc. have been developed and studied in detail. The potential application value of the above material in PEMFC and direct methanol fuel cell (DMFC) has been reported widely. For example, Tsai et al. reported novel sulfonated poly(arylene ether sulfone) membranes which exhibited excellent thermal properties, desirable dimensional stability and satisfactory proton conductivity. In their work, the methanol permeability coefficients of the reported membranes were lower than that of Nafion 117. The membranes showed a great advantage for use in DMFC.¹⁹ Yang et al. reported the synthesis of sulfonated poly(ether ether ketone) based membranes with excellent performance for DMFC. The membranes exhibited lower methanol permeability coefficients and desirable fuel cell performance.²⁷ Li et al. reported sulfonated polyimides with good mechanical property, high proton conductivity which revealed their potential value to be applied for PEMFC.²⁴ Yoonessi et al. reported sulfonated poly(arylene thioether sulfone) membranes as potential PEM materials for hydrogen fuel cells with high proton conductivities at 85% relative humidity, desirable water uptake as well as comprehensive analysis of the microstructure.^25,26 The research progress of sulfonated poly(arylene ether ketone sulfone) is also wide and fast. Wang et al. reported a series of sulfonated poly(arylene ether ketone sulfone) membranes with different degree of sulfonation (DS). The membranes exhibited very low methanol diffusion and good prospects in PEMFC usages.²² Bae et al. reported novel multiblock copolymers based on sulfonated poly(arylene ether ketone sulfone) with high mechanical strength and low hydrogen and oxygen permeability.²³

High DS is necessary for sulfonated polymers obtaining high proton conductivities. However, high DS brings about some problems such as poor mechanical property, undesirable dimensional stability and high methanol crossover. Therefore, it is necessary to take some measures to solve these problems. Cross-linking is generally classified as ionic and covalent cross-linking. Cross-linking is an effective method to hinder methanol crossover and improve the dimensional stability of highly sulfonated polymers. However, ionic cross-linking is instable and covalent cross-linking reduces the proton conductivity.^11,28–30 Na et al. reported a new type of cross-linked membrane based on Nafion and sulfonated poly(arylene ether ketone) bearing carboxylic acid groups. In that paper, the methanol crossover decreased sharply after cross-linking. The methanol permeability coefficients were in the range of 4.56 × 10⁻⁷ cm² s⁻¹ to 2.57 × 10⁻⁷ cm² s⁻¹ at 25 °C. Moreover, the membranes showed suitable swelling ratio and water absorption which result in desirable dimensional stability.²⁸ Lee et al. reported the application of cross-linked PVA membranes containing sulfonic acid groups. The membranes exhibited good thermal stabilities, desirable solubility and satisfactory proton conductivities.³⁰ In our previous work, a series of sulfonated poly(arylene ether ketone sulfone) with amino groups (Am-SPAEKS) cross-linked membranes were prepared.¹⁷ In that work, covalent cross-linking reaction occurred through thermal treatment in single polymer chain. The treated membranes exhibited good thermal property, excellent mechanical property and lower methanol permeability. However, large amounts of –SO₃H were consumed during the reaction which leads to low proton conductivity. Therefore, it is necessary to maintain desirable proton conductivity during cross-linking process. Kreuer speculate that the capacity of carriers combine and dissociate with a proton jointly decide the proton migration rate.^31,32 Honma et al. and Zawodzinski et al. blend the acidic and alkaline carrier in order to build proton jump sites.^33,34 In such sites, acidic groups (–SO₃H) and alkaline groups (–NH₂) can form acid–base pair (–SO₃H⋯H₂N–). When the proton dissociates from –SO₃H groups, the –NH₂ groups produce suction which can promote proton dissociation and accept the proton to form ⁺H₃N–; –SO₃⁻ by dissociated can also promote the proton dissociation of ⁺H₃N– and accept the proton. In this way, the proton transfer between the carriers can be promoted by attractive force. The results of our previous study accorded with this hypothesis.¹⁷

In this paper, conductive sulfonated poly(vinyl alcohol) (SPVA) played a role of macromolecular cross-linker due to its outstanding methanol diffusion resistance properties, simple synthetic method, desirable solubility, suitable proton conductivity and film-forming properties.³⁰ Moreover, macromolecular cross-linker has a high molecular weight and contains large amounts of cross-linking points on its backbone. Therefore, the cross-linked membranes can get high degree of cross-linking.^35,36 The introduction of amino groups and sulfonic acid groups can form acid–base pair (–SO₃H⋯H₂N–), thus providing new proton transport pathway and promoting proton conduction. In addition, the cross-linking structures forming between amino groups in Am-SPAEKS polymer and sulfonic acid groups in SPVA polymer can effectively reduce the methanol permeation. Novel Am-SPAEKS-based cross-linked membranes which contain different content of SPVA were fabricated. The properties were studied on the subject of dimensional stability, water uptake, oxidative stability, methanol permeability coefficients, ion exchange capacity (IEC), proton conductivity and the micro-morphology.

Experimental

Materials

SPVA (M_n = 105923, degree of sulfonation = 2.38%), 4-aminophenyl hydroquinone (4Am-PH) and 3,3′-disulfonated-4,4′-dichlorodiphenylsulfone (SDCDPS) were synthesized in our lab.^17,37 4,4′-Difluorodiphenylmethanone, 2,2-bis(4-hydroxypheny)propane (bisphenol A) (AR grade), anhydrous potassium carbonate (AR grade), N-methyl-2-pyrrolidone (NMP) (99%), toluene and dimethyl sulfoxide (DMSO) (AR grade) were from commercial sources.

Synthesis of Am-SPAEKS copolymers

As shown in Scheme 1, Am-SPAEKS polymer was synthesized by polycondensation reaction. The DS (60%) is constant by controlling the mole ratio between SDCDPS and 4,4′-difluorobenzophenone (3 : 7). The fixed proportion between 4Am-PH and bisphenol A is 2 : 8. Experimental procedures were as follows:

Scheme 1 Synthesis route of Am-SPAEKS membranes.

First of all, 4,4′-difluorobenzophenone (4.578 g, 21 mmol), SDCDPS (4.419 g, 9 mmol), 4Am-PH (1.206 g, 6 mmol), bisphenol A (5.472 g, 24 mmol), potassium carbonate (4.14 g, 30 mmol), NMP (40 mL), toluene (30 mL) were carefully put into 500 mL three-necked flask armed with mechanical agitator, reflux condenser apparatus under N₂ protection. The mixture was kept stirring at room temperature for 2 h and then elevated the temperature to 130 °C slowly for 4–5 h to exclude the water from the system. Afterwards, the toluene was removed by distillation. Then, the reaction temperature was elevated and kept at 170 °C for 30 h. The system temperature was cooled down to 25 °C when the mixture reached a certain viscosity. The viscous mixture was dumped into a beaker loaded with 500 mL water to precipitate the polymers. Polymers were minced and washed by boiling deionized water repeatedly. The purpose of doing this is to exclude the residual organic solvents. Finally, the product was put into oven and dried for 48 h at 60 °C.

Membrane preparation

Am-SPAEKS and SPVA were used for preparing composite membranes. The Am-SPAEKS/SPVA composite membranes were recorded as Am/S-xx., where xx represents the SPVA weight ratio (20%, 40%, 60% and 80%). Am/S-20 was prepared as follows: firstly, Am-SPAEKS (1.6 g) and SPVA (0.40 g) were dissolved in 16 mL and 4 mL of DMSO, respectively. And then the solutions were stirred until homogeneous liquid are obtained. Finally, the above two kinds of solution were blended and continued to stir for 6 h, cast onto a glass plate, dried at 60 °C for 48 h to exclude the residual solvent. The cross-linked membrane was denoted as Cr-Am/S-xx through thermal treatment at 180 °C for 1 h. The above membrane was soaked in 1 M HCl for 24 h to transform the membrane into acid type. The acidic membranes surface was washed with distilled water over and over again till neutral. The preparation process of the cross-linked membranes was shown in Scheme 2.

Scheme 2 The preparation process of cross-linked membranes.

Characterization

FT-IR spectra were measured using a Vector-22 spectrometer equipped with the attenuated total reflectance (ATR) device in the range of 650–4000 cm⁻¹. The ¹H NMR spectra were obtained using a 400 MHz Bruker Avance III spectrometer using deuterated dimethyl sulfoxide (DMSO-d₆) as the solvent and tetramethylsilane (TMS) as the reference. The thermogravimetric analysis (TGA) was conducted to evaluate the thermal stability by a Perkin Elmer Pyris 1 analyzer in N₂ atmosphere. The heating rate was 10 °C min⁻¹ from 40 °C to 600 °C. For microstructure characterization, the morphology was observed with a JEM-1011 transmission electron microscope and Agilent 5100 atomic force microscope. Tapping mode AFM was performed in this work.

Measurement

Mechanical property

An Instron 5965 instrument was used to evaluate the mechanical properties of the membranes. The stretching rate of the test was 1 mm min⁻¹. The specimens were dumbbell-shaped accompany with the size of 25 mm × 4 mm. For each test, at least three measurements were obtained. Finally, in the experiment, an average was adopted.

Ion exchange capacity (IEC)

Traditional titration method was employed to evaluate the IEC values of cross-linked membranes. First, square dry specimens were soaked in NaCl aqueous (1.0 M) and kept at room temperature for 24 h to ensure the ionic exchange from H⁺ to Na⁺ completely. And then the above solution was titrated with NaOH (0.05 M). During the test, phenolphthalein plays the role of an indicator. The IEC can be calculated using the formula:where V is the exhausted volume of NaOH during the titration; M is the concentration of NaOH solution, which is a constant (0.05 mol L⁻¹); W_dry is dry membranes mass.

Water uptake and swelling ratio

The water uptake and swelling ratio were measured from the changes in weight and dimension, respectively. The specific experiment steps were reported in our previous work.³⁷ The water uptake and swelling ratio were calculated as follows:where W_dry and W_wet are the weight of the dry membranes and wet membranes, and T_dry and T_wet are the thickness of dry and wet membranes, respectively.

The number of water molecules per sulfonic groups (λ) can be calculated by the water uptake and the IEC via the following relationship:³⁸

where n(H₂O) is the mole number of H₂O, n(SO₃⁻) is the mole number of SO₃⁻ groups, and 18 corresponds to water molecular weight.

Gel fraction

The dry specimens were taken out, dried and weighted after immersing them in DMSO at 80 °C for 24 h. Gel fraction was examined from the ratio of residual mass after extraction from DMSO and the initial mass.

Oxidative stability test

The samples (2 × 1 cm²) were soaked in Fenton's reagent (3% H₂O₂ solution containing 2 ppm FeSO₄) at 80 °C for 1 h to evaluate their oxidative stability. After 1 h, the specimens were taken out, dried and weighted. The oxidative stability values were calculated from the ratio of residual mass after oxidative stability test and the initial mass.

Measurement of proton conductivity

In this part of test, four-probe alternating current (AC) impedance technique was used to obtain the proton conductivity. The instrument type and other information were the same as our previous work.¹⁷ The specimens (5 × 1 cm²) were immersed in deionized water for 24 h, and then taken out and sandwiched between Teflon sheets. The processed specimen was soaked in deionized water to obtain 100% relative humidity. The proton conductivity (σ) was calculated using the following formula:where σ is the proton conductivity (S cm⁻¹), L (cm) is the distance between the two electrodes, R (Ω) is the membrane resistance, and S (cm²) is the cross-sectional area of the membrane.

Methanol permeability

The custom glass cell used in the test composed of two reservoirs, one reservoir is used for storing the deionized water, and the other reservoir is used for storing methanol. The specimen membrane was sandwiched between two reservoirs. The solutions in both sides were magnetically stirred during the test. The schematic illustration of the device has been shown in Scheme 3. Methanol concentration in the water reservoir was tested by Agilent-6890N gas chromatograph. The methanol diffusion coefficients at 20 °C and 60 °C (ref. 37) were calculated using the following formula:where c_B (mol L⁻¹) and c_A (mol L⁻¹) are the methanol concentrations of B and A reservoirs, respectively; A (cm²) is the membrane's effective area; L (cm) is the membrane's thickness; V_B (mL) is the volume of the reservoirs; and DK (cm⁻² s⁻¹) is the methanol permeability coefficient.

Scheme 3 The schematic illustration of the diffusion device used for measuring the methanol permeability of the membranes.

Results and discussion

¹H NMR spectra of 4Am-PH and Am-SPAEKS

¹H NMR spectrum of the 4Am-PH is shown in Fig. 1(a). The peak at δ_H 5.0 ppm was attributed to NH₂ group. The peaks at δ_H 8.5 ppm were observed and assigned to the phenolic hydroxyl groups. Chemical shifts of the benzene ring unit are observed at δ_H 6.5 to δ_H 7.5 ppm.

Fig. 1 ¹H NMR spectra of (a) 4Am-PH and (b) Am-SPAEKS.

The ¹H NMR spectrum of the Am-SPAEKS is shown in Fig. 1(b). In Fig. 1(b), the chemical shifts at δ_H 8.26 and δ_H 7.05 ppm were attributed to the H⁺ surrounding the –SO₃H and –NH₂, respectively.

FT-IR spectra of Am/S-20 and Cr-Am/S-20 membranes

The FT-IR spectra of Am/S-20 and Cr-Am/S-20 are shown in Fig. 2. In Am/S-20 FT-IR spectrum, the absorption peaks were attributed as follows: the symmetrical stretching of S=O in –SO₃H groups at 1014.41 cm⁻¹, the existence of C=O at 1654.22 cm⁻¹, vibrational mode of the C=C in benzene ring at 1593.40 and 1499.44 cm⁻¹, C–O stretching vibrations in polymer chain at 1306.60 cm⁻¹, O=S=O bond in polymer main chain at 1161.44 cm⁻¹, C–C stretching vibration at 1243.42 cm⁻¹, N–H stretching vibration at 3373.92 and 3461.54 cm⁻¹. However, the N–H stretching vibration disappeared in the IR spectrum of Cr-Am/S-20 membrane. This is due to the cross-linking reaction between –NH₂ and –SO₃H. In addition, the different color between Am/S-20 and Cr-Am/S-20 membranes showing in Fig. 2 further illustrates the existence of cross-linking reaction. It is obvious that the color of Cr-Am/S-20 membrane is deeper than that of Am/S-20 membrane.

Fig. 2 FT-IR spectra of Am/S-20 and Cr-Am/S-20 membranes.

Thermal property

Fig. 3 presents the TGA curves from 40 °C to 600 °C of Am-SPAEKS and cross-linked membranes. Am-SPAEKS membrane had three weight loss stages from 80 °C to 600 °C. The first stage ranged from 80 °C to 170 °C was related to the evaporation of remain solvent and water. The second loss stage begin at 235 °C was related to the desulfonation (–SO₃H). The last loss stage started at 435 °C was related to polymer backbone decomposition. All the cross-linked membranes exhibited three degradation stages. The first weight loss stage (100–195 °C) was related to the evaporation of the residual water and solvent. The second loss stage was corresponded to the loss of –SO₃H and –SO₂–NH– groups (at ∼265 °C). The third loss step (400–470 °C) was attributed to the decomposition of SPVA and Am-SPAEKS backbone. In addition, the temperatures leading to 5% weight loss (T_d5%) of the membranes are shown in Table 1. Obviously, the cross-linked membranes show higher T_d5% than that of Am-SPAEKS. Moreover, T_d5% values of cross-linked membranes increased with increasing SPVA. It is because the degree of cross-linking may increase with increasing SPVA, which leading to the incremental cross-linking sites. Cross-linking structure can restrict the movement of the polymer chain and effectively improves the thermal stability of the cross-linked membranes.¹¹ From the TGA curves and T_d5% data in Table 1, it is obvious that the cross-linked membranes exhibit better thermal property than that of Am-SPAEKS and show satisfactory thermal performance for application.

Fig. 3 TGA curves of Am-SPAEKS and cross-linked membranes.

Table 1 The properties of composite membranes and cross-linked membranes

Samples	Td5% (°C)	Gel fraction (%)	λ (H2O molecule/SO3^−)	IEC (mmol g^−1)	Oxidative stability (%)
Am-SPAEKS	179	0	2.8	1.34 ± 0.03	87.2 ± 0.7
SPVA	—	0	—	0.53 ± 0.01	12.8 ± 0.2
Am/S-20	—	0	10.3	1.25 ± 0.03	78.3 ± 0.3
Am/S-40	—	0	20.1	1.12 ± 0.02	67.8 ± 0.5
Am/S-60	—	9.1 ± 0.2	39.3	0.95 ± 0.02	49.5 ± 0.3
Am/S-80	—	12.3 ± 0.4	67.8	0.75 ± 0.01	28.9 ± 0.2
Cr-Am/S-20	266	58.7 ± 0.5	3.3	1.17 ± 0.03	92.8 ± 0.7
Cr-Am/S-40	271	63.1 ± 0.5	2.7	1.01 ± 0.03	94.7 ± 0.6
Cr-Am/S-60	279	72.9 ± 0.8	2.6	0.83 ± 0.01	95.0 ± 0.3
Cr-Am/S-80	288	78.6 ± 0.9	1.5	0.62 ± 0.01	99.0 ± 0.5

---

IEC, water uptake and swelling ratio

IEC is the key factor to determine the proton conductivity, and it plays a very important role in the application of PEMFC. The IEC values are show in Table 1. The values of IEC decreased with incremental SPVA content. The cross-linked membranes showed lower IEC values than that of composite membranes. That is because that the –SO₃H groups were consumed partially in cross-linking reaction.

Water molecules act as carriers in the proton transfer process. However, superabundant water can lead to poor mechanical property and undesirable dimensional stability. Therefore, suitable water absorption is essential for membranes applications. Water uptake and swelling ratio of the composite membranes increased with incremental SPVA content at 25 °C (Fig. 4(a)). The Am/S-60 and Am/S-80 composite membranes lost their size severely at 80 °C (Fig. 4(c)). Therefore, the water uptake and swelling ratio of Am/S-60 and Am/S-80 membranes can not be obtained. Conversely, the cross-linked membranes have more stable shape than that of composite membranes. The water absorption and swelling ratio of the cross-linked membranes reduced with increasing SPVA both at 25 °C and 80 °C (Fig. 4(b) and (d)). It is noteworthy that the cross-linked membranes had much lower λ values than the composite membranes. This is because the lower water uptake. Cross-linking reaction can bring about denser structures which seriously restricts the movement of water molecules. The reticular structures can make the transmission hydrophilic channels for water absorption become smaller, thus decreasing the water absorption.^17,28–30

Fig. 4 Water uptake and swelling ratio of all the membranes at 25 °C and 80 °C, respectively, (a) composite membranes at 25 °C, (b) cross-linked membranes at 25 °C, (c) composite membranes at 80 °C, (d) cross-linked membranes at 80 °C.

Gel fraction and oxidative stability

The gel fraction can indirectly reflect the degree of cross-linking.¹² As shown in Table 1, the composite membranes and Am-SPAEKS membranes showed good solubility and easily dissolved in DMSO. The cross-linked membranes exhibited serious swelling behavior and only partially dissolved. The experiment values of the cross-linked membranes were increased from 58.7% to 78.6% with increasing SPVA from 20 wt% to 80 wt%. The reason is that the degree of cross-linking reaction enhanced with increasing SPVA which contains large amount of cross-linking points.^35,36

Table 1 shows the resistance-to-oxidation data of the composite and cross-linked membranes. For the composite membranes, with the increasing SPVA content (from 20% to 80%), an obvious decline in weight was observed (from 78.3% to 28.9%). The lower oxidative stability of composite membrane with more SPVA probably resulted from the instability of PVA aliphatic chain. Oppositely, the cross-linked membranes possessed higher resistance-to-oxidation data. The residual weight of all the cross-linked membranes was above 90%. This is because the oxidative stability has a close relationship with the water absorption.⁴³ Generally speaking, excessive water absorption and extremely swelling mainly decrease the density of the molecular chain. Thereby the free radical (HO· and HO₂·) could attack on the molecular chain in less time.³⁷ Experiment data show that the dense cross-linking structures effectively prevent the oxidation free radical from attacking the polymer chain.³⁹ Therefore, the oxidative stability of the polymers can be improved obviously by cross-linking reaction.

Mechanical property

Table 2 shows the mechanical property data of the membranes. For composite membranes, Young's modulus were in the range of 256.16–936.61 MPa; tensile strength ranged from 14.57 MPa to 34.66 MPa. For cross-linked membranes, the Young's modulus and tensile strength are higher than that of the composite membrane, which ranged from 361.16 MPa to 1098.41 MPa and 16.10 MPa to 39.23 MPa, respectively. It is obvious in Table 2 that the elongation-at-break of the cross-linked membranes decreased sharply. The elongation at break of Cr-Am/S-80 is only 1/33 of Am/S-80.

Table 2 Mechanical properties of Am-SPAEKS/SPVA membranes

Samples	Young's modulus (MPa)	Tensile strength (MPa)	Elongation-at-break (%)
Am-SPAEKS	1685.45 ± 4.02	35.70 ± 1.25	7.68 ± 0.98
Am/S-20	936.61 ± 2.56	34.66 ± 1.43	116.64 ± 5.75
Am/S-40	529.78 ± 1.65	19.60 ± 0.85	148.68 ± 7.92
Am/S-60	317.91 ± 1.81	14.57 ± 0.78	130.85 ± 6.66
Am/S-80	256.16 ± 2.17	14.76 ± 0.69	234.45 ± 12.18
Cr-Am/S-20	1098.41 ± 3.03	39.23 ± 1.88	7.60 ± 1.05
Cr-Am/S-40	796.33 ± 2.86	31.55 ± 1.69	6.09 ± 0.65
Cr-Am/S-60	584.79 ± 2.57	26.32 ± 1.75	5.21 ± 0.52
Cr-Am/S-80	361.16 ± 3.32	16.10 ± 0.0.92	7.15 ± 0.87

---

Morphology

The morphology of PEMs is directly related to their physical and electrochemical properties. Eisenberg's theoretical hypothesis is that sulfonic acid groups may gather into ion clusters, thus forming ionic transport channels.⁴⁰ Both TEM and AFM analyses were performed to evaluate the microstructure of the membranes.

As shown in Fig. 5, the black dots represent the hydrophilic sulfonic acid clusters stained by Ag⁺ ions, whereas the brighter area represents the hydrophobic polymer backbones.⁴¹ The similar structures are also reported by Yoonessi.²⁵ Yoonessi believe the silver was attributed to proton exchange at the sulfonic sites. Ag⁺ or H⁺ in ionic aggregates was localized in near-spherical.²⁵ In Fig. 5, both Am/S-20 and Cr-Am/S-20 membranes have spherical ionic clusters (dark areas). In Cr-Am/S-20 membranes, the ionic clusters are larger than that of Am/S-20 membrane. The reasons are as following: (i) the acidic groups (–SO₃H) and alkaline groups (–NH₂) can form acid–base pair (–SO₃H⋯H₂N–); (ii) the new proton jump sites are constructed through acid–base pair, thus making the ionic clusters larger; (iii) the existing hydrogen bonds promote the formation of the larger ionic clusters.

Fig. 5 TEM images of (a) Am/S-20 and (b) Cr-Am/S-20 membranes.

The AFM phase images (tapping mode) of Cr-Am/S-40 are shown in Fig. 6. For Cr-Am/S-40 membrane, black, cluster-like structures were obviously visible in the phase image. The black areas represent the soft region containing hydrophilic sulfonic acid and hydroxyl groups which accompany with small amounts of water. The light yellow areas express the hydrophobic polymer main chains (tough region).³⁹ The distinct phase separate structure is beneficial to the proton conductivity.

Fig. 6 AFM tapping phase images of Cr-Am/S-40 (a) 2D view and (b) 3D view.

Proton conductivity and methanol permeability

The proton conductivities of the Am-SPAEKS and cross-linked membranes are shown in Table 3. The highest proton conductivity (0.061 S cm⁻¹) was attributable to Am-SPAEKS. Regrettably, from Table 3, the proton conductivity of the cross-linked membranes decreased with the increasing SPVA. This is related to the morphology of the membranes. From the TEM images of cross-linked membrane, larger ionic clusters formed obviously which is conducive to proton transfer;⁴⁰ however, cross-linking reactions occupy the primary position and a certain amount of –SO₃H groups were consumed during the reaction. Thus, the proton conductivity decreased. Moreover, SPVA showed low IEC which result in low proton conductivity. When the content of SPVA reached 80%, the IEC value of Cr-Am/S-80 deceased dramatically, this directly led to a sharp drop in the proton conductivity.

Table 3 The properties of Am-SPAEKS membranes and cross-linked membranes

Samples	Proton conductivity (S cm^−1)		Methanol permeability coefficient (×10^−7 cm^−2 s^−1)		Relative selectivity (×10^4 S scm^−3)		Ea (KJ mol^−1)
	20 °C	80 °C	20 °C	60 °C	20 °C	60 °C
Am-SPAEKS	0.027	0.061	9.86 ± 0.49	19.91 ± 0.79	2.74	2.56	12.45
Cr-Am/S-20	0.023	0.054	7.23 ± 0.36	14.09 ± 0.56	3.18	3.19	13.20
Cr-Am/S-40	0.017	0.045	4.52 ± 0.22	9.53 ± 0.38	3.76	3.25	14.19
Cr-Am/S-60	0.011	0.029	1.61 ± 0.08	3.38 ± 0.13	6.83	6.51	14.48
Cr-Am/S-80	0.005	0.011	0.32 ± 0.01	0.96 ± 0.03	15.63	8.33	15.62

---

The Arrhenius plots of conductivity vs. temperature of the Am-SPAEKS and cross-linked membranes at different temperatures are shown in Fig. 7. It is as expected that proton conductivity considerably depended on temperature and increased with increasing temperature. Arrhenius relationship can appropriately expound the temperature dependence of proton conductivity.

Fig. 7 Arrhenius plot of conductivity vs. temperature.

The activation energy (E_a) was calculated using the following formula:⁴²

where R′ equals to 8.314 J K⁻¹ mol⁻¹ representing the universal gas constant, σ is the proton conductivity, σ_o is the pre-exponential factor, T is the temperature (K) and E_a is activation energy.

Generally, proton transport occurs mainly by the Grotthus mechanism and vehicular mechanism.¹¹ In Grotthus mechanism, the H⁺ forms as H₃O⁺ ion. The H₃O⁺ jumps to the neighboring lone pair of electrons of water molecules, and E_a should range from 14 kJ mol⁻¹ to 40 kJ mol⁻¹.^42,43 In vehicular mechanism, H⁺ combines with H₂O and form H₃O⁺ or H₅O₂⁺ etc. in water.⁴⁴ Then, the protons transfer through the diffusion and movement of these hydronium ions.

Table 3 shows the E_a values which calculated from the slopes of the Arrhenius plot in Fig. 7. The E_a of the cross-linked membranes ranged from 13.20 to 15.62 KJ mol⁻¹. The protons transfer mechanism in the Am-SPAEKS and cross-linked membranes may mainly be vehicular mechanism. The E_a values of Cr-Am/S-40, Cr-Am/S-60 and Cr-Am/S-80 were in the range of 14–40 KJ mol⁻¹. In Cr-Am/S-40, Cr-Am/S-60 and Cr-Am/S-80 membranes, the protons may transport partially by Grotthus mechanism. The reasons may be as follow. On the one hand, the acid–base pair (–SO₃H⋯H₂N–) can be formed between –SO₃H and –NH₂. When the proton dissociates from –SO₃H groups, the –NH₂ groups produce suction which can promote proton dissociation and accept the proton to form the ⁺H₃N–; while the –SO₃⁻ by dissociated can also promote the proton dissociation of ⁺H₃N– and accept the proton. In this way, the proton transportation between the carriers can be promoted by attractive force.^31–34 On the other hand, the H⁺ can jumps between the hydrogen bonds. Nevertheless, due to the low IEC of SPVA, the cross-linked membranes showed a low-grade level of the proton conductivity. Scheme 4 shows the proton transport mechanism in the cross-linked membranes.

Scheme 4 Schematic diagram of proton transport mechanism in cross-linked membranes.

Methanol concentrations were measured both at 20 °C and 60 °C, respectively. As shown in Table 3, the SPVA content increased from 20% to 80%, the methanol permeability coefficients reduced from 7.23 × 10⁻⁷ to 0.32 × 10⁻⁷ cm² s⁻¹ at 20 °C and 14.09 × 10⁻⁷ to 0.96 × 10⁻⁷ cm² s⁻¹ at 60 °C, respectively. The methanol permeability coefficients of Cr-Am/S-80 membrane (0.32 × 10⁻⁷ cm² s⁻¹ at 20 °C and 0.96 × 10⁻⁷ cm² s⁻¹ at 60 °C) in this paper were much lower than that of C-Am-SPAEKS-1 membrane (5.56 × 10⁻⁷ cm² s⁻¹ at 25 °C and 10.93 × 10⁻⁷ cm² s⁻¹ at 60 °C) in our previous work.¹⁷ This result indicates the importance of the introduction of SPVA. The introduced SPVA has significant contribution to the low methanol crossover. As shown in Fig. 8, with increasing SPVA content, a gradual decline in the methanol permeability coefficients were exhibited. This has a close relationship with the introduction of SPVA, which directly leads to the decrease of the methanol permeability coefficients. It is obvious that the cross-linked membranes possess better methanol-rejecting ability than that of Am-SPAEKS, which was possibly originated from the enhanced cross-linking reaction and dense cross-linked network. It is worth mentioning that methanol is poor solvent for PVA. The chain of PVA molecular would curl up in methanol solution. This causes the transmission channel of methanol molecular hampered.^8,37

Fig. 8 Effects of SPVA content on the DK and relative selectivity value of membranes at 20 °C and 60 °C.

Relative selectivity is a key factor for membrane application in PEMFC. Usually, the ratio of proton conductivity to methanol permeability coefficient is defined as the relative selectivity. In general, if the membrane has higher selectivity, it has a better the comprehensive performance.⁴⁴ The relative selectivity values are shown in Table 3 and the curves of methanol permeability coefficient and relative selectivity values are shown in Fig. 8. It is obvious in Fig. 8 that the relative selectivity sharply increased with increasing SPVA content and the methanol permeability coefficients decreased with incremental SPVA content. Low methanol crossover contributed to the high relative selectivity. The methanol crossover increased with incremental temperature. However, from the selectivity values, the cross-linked membranes show better comprehensive performance compared to Am-SPAEKS, which is mainly caused by much lower methanol permeability.⁴⁴

Solubility parameter

The solubility parameter is the characterization of interactions between polymer and solvent. The solubility parameters values can be estimated using the following formula according to their chemical structures:^45,46F represents the groups' molar attraction constant of solvent or polymer repeat unit, ρ is density, M₀ is molecular weight of repeat unit and δ is solubility parameter.

According to the previous literatures, the solubility parameters of methanol and water are 14.5 and 23.2 (cal^1/2 cm^−3/2), respectively.^45,46 The F can be obtained from the literatures.^45–47 The molecular weight of repeat unit of PVA and SPVA are 44 and 124, respectively. The densities of PVA (1.23 g cm⁻³) and SPVA (1.09 g cm⁻³) have been obtained by the density calculation formula. According to the solubility parameters formula, the solubility parameters of PVA 12.39 (cal^1/2 cm^−3/2) and SPVA 11.37 (cal^1/2 cm^−3/2) are obtained. The solubility parameters of PVA and SPVA are similar. PVA and SPVA are very difficult to dissolve in methanol but easily dissolve in water, which may be related to their strong polarity.^8,37,45–47

Single cell performance

The single fuel cell performance of Cr-Am/S-20, Cr-Am/S-40 and Cr-Am/S-60 membranes was studied with polarization curve in a single cell test. Fig. 9 shows the polarization and power density curves of Cr-Am/S-20, Cr-Am/S-40 and Cr-Am/S-60 membranes under fully humidified inlet gas conditions at 90 °C. From Fig. 9, in the region of low current density, the potential decreased greatly which caused by activation control. In the region of intermediate current density, the potential was decreased due to the intrinsic ohmic resistance.⁴⁸ The maximum power density of Cr-Am/S-20, Cr-Am/S-40 and Cr-Am/S-60 membranes were 65.92 mW cm⁻², 57.55 mW cm⁻² and 44.23 mW cm⁻², respectively. The power density of the cross-linked membranes decreased with increasing SPVA content. This is mainly due to the low IEC of SPVA which results in low proton conductivity.^49,50 These results demonstrate that the cross-linked membranes with low SPVA content have better fuel cell performance.

Fig. 9 Polarization and power density curves for PEMFC with Cr-Am/S-20, Cr-Am/S-40 and Cr-Am/S-60 membranes.

Conclusion

Novel Am-SPAEKS/SPVA cross-linked membranes with various SPVA content were fabricated. Water uptake and methanol crossover of cross-linked membranes decreased sharply with the increasing SPVA. The thermal stability, dimensional stability and mechanical property were improved obviously after cross-linking. Especially, Cr-Am/S-80 membrane displayed a much lower methanol permeability coefficient of 0.96 × 10⁻⁷ cm² s⁻¹ at 60 °C which was only one twentieth of Am-SPAEKS (19.91 × 10⁻⁷ cm² s⁻¹) at the same condition. The E_a values approached to 14 KJ mol⁻¹ which indicate that Grotthus mechanism may be partly happen in the cross-linked membranes. According to the relative selectivity, the cross-linked membranes exhibited desirable comprehensive performance. Furthermore, the cross-linked membranes show desirable fuel cell performance to prove its application value as PEM in PEMFC.

Acknowledgements

We are thankful for the financial support from the National Natural Science Foundation of China (Grant no. 51273024 and 51303015) and Department of Education of Jilin Province (Grant no. 2014119).

References

1. H. X. Xie, D. Tao, X. Z. Xiang, Y. X. Ou, X. J. Bai and L. Wang, J. Membr. Sci., 2015, 473, 226–236.
2. N. W. Li, D. W. Shin, D. S. Hwang, Y. M. Lee and M. D. Guiver, Macromolecules, 2010, 43, 9810–9820.
3. D. W. Shin, S. Y. Lee, N. R. Kang, K. H. Lee, M. D. Guiver and Y. M. Lee, Macromolecules, 2013, 46, 3452–3460.
4. J. F. Zheng, Q. Y. He, C. L. Liu, T. Yuan, S. B. Zhang and H. Yang, J. Membr. Sci., 2015, 476, 571–579.
5. N. W. Li, D. S. Hwang, S. Y. Lee, Y. L. Liu, Y. M. Lee and M. D. Guiver, Macromolecules, 2011, 44, 4901–4910.
6. Y. Yang, H. Gao and L. Zheng, RSC Adv., 2015, 5, 17683–17689.
7. X. Xu, L. Li, H. Wang, X. Li and X. Zhuang, RSC Adv., 2015, 5, 4934–4940.
8. D. S. Kim, H. B. Park, J. W. Rhim and Y. M. Lee, J. Membr. Sci., 2014, 240, 37–48.
9. J. H. Pang, S. N. Feng, H. B. Zhang, Z. H. Jiang and G. B. Wang, RSC Adv., 2015, 5, 38298–38307.
10. X. P. Li, C. Liu, S. H. Zhang, L. S. Zong and X. G. Jian, J. Membr. Sci., 2013, 442, 160–167.
11. J. M. Xu, H. L. Cheng, L. Ma, H. L. Han and Z. Wang, Int. J. Hydrogen Energy, 2013, 38, 10092–10103.
12. H. T. Li, G. Zhang, J. Wu, C. J. Zhao, Y. Zhang, K. Shao, M. M. Han, H. D. Lin, J. Zhu and H. Na, J. Power Sources, 2010, 195, 6443–6449.
13. H. Y. Yao, Y. H. Zhang, Y. Liu, K. Y. You, N. N Song, B. J. Liu and S. W. Guan, J. Membr. Sci., 2015, 480, 83–92.
14. M. Y. Li, G. Zhang, S. Xu, C. J. Zhao, M. M. Han, L. Y. Zhang, H. Jiang, Z. G. Liu and H. Na, J. Power Sources, 2014, 255, 101–107.
15. H. L. Cheng, J. M. Xu, L. Ma, L. S. Xu, B. J. Liu, Z. Wang and H. X. Zhang, J. Power Sources, 2014, 260, 307–316.
16. C. J. Zhao, H. D. Lin, K. Shao, X. F. Li, H. Z. Ni, Z. Wang and H. Na, J. Power Sources, 2006, 162, 1003–1009.
17. J. M. Xu, H. L. Cheng, L. Ma, H. L. Han, Y. S. Huang and Z. Wang, J. Polym. Res., 2014, 21, 423.
18. L. M. Wang, J. H. Zhu, J. F. Zheng, S. B. Zhang and L. Y. Dou, RSC Adv., 2014, 4, 25195–25200.
19. J. C. Tsai, C. K. Lin, J. F. Kuo and C. Y. Chen, J. Power Sources, 2010, 195, 4072–4079.
20. F. Wang, M. Hickner, Y. S. Kim, T. A. Zawodzinski and J. E. McGrath, J. Membr. Sci., 2002, 197, 231–242.
21. K. Miyatake, D. Hirayama, B. Bae and M. Watanabe, Polym. Chem., 2012, 3, 2517–2522.
22. Z. Wang, X. F. Li, C. J. Zhao, H. Z. Ni and H. Na, J. Power Sources, 2006, 160, 969–976.
23. B. Bae, K. Miyatake and M. Watanabe, Macromolecules, 2010, 43, 2684–2691.
24. N. W. Li, Z. M. Cui, S. B. Zhang, S. H. Li and F. Zhang, J. Power Sources, 2007, 172, 511–519.
25. M. Yoonessi, T. D. Dang, H. Heinz, R. Wheeler and Z. W. Bai, Polymer, 2010, 51, 1585–1592.
26. M. Yoonessi, H. Heinz, T. D. Dang and Z. W. Bai, Polymer, 2011, 52, 5615–5621.
27. T. Yang, L. Y. Meng and N. Huang, J. Power Sources, 2013, 224, 132–138.
28. H. D. Lin, C. J. Zhao and H. Na, J. Power Sources, 2010, 195, 3380–3385.
29. M. Higa, S. Y. Feng, N. Endo and Y. Kakihana, Electrochim. Acta, 2015, 153, 83–89.
30. J. W. Rhim, H. B. Park, C. S. Lee, J. H. Jun, D. S. Kim and Y. M. Lee, J. Membr. Sci., 2004, 238, 143–151.
31. M. Schuster, T. Rager, A. Noda, K. D. Kreuer and J. Maier, Fuel Cells, 2005, 5, 355–365.
32. S. J. Paddison, K. D. Kreuer and J. Maier, Phys. Chem. Chem. Phys., 2006, 8, 4530–4542.
33. D. M. Yamada and I. Honma, Angew. Chem., Int. Ed., 2004, 43, 3688–3691.
34. R. Subbaraman, H. Ghassemi and T. A. Zawodzinski, J. Am. Chem. Soc., 2007, 129, 2238–2239.
35. S. Gu, G. H. He, X. M. Wu, Y. J. Guo, H. J. Liu, L. Peng and G. K. Xiao, J. Membr. Sci., 2008, 312, 48–58.
36. C. J. Zhao, H. D. Lin and H. Na, Int. J. Hydrogen Energy, 2010, 35, 2176–2182.
37. J. M. Xu, H. Z. Ni, S. Wang, Z. Wang and H. X. Zhang, J. Membr. Sci., 2015, 492, 505–517.
38. T. Tamura, R. Takemori and H. Kawakami, J. Power Sources, 2012, 217, 135–141.
39. J. M. Xu, L. Ma, H. L. Han, H. Z. Ni, Z. Wang and H. X. Zhang, Electrochim. Acta, 2014, 146, 688–696.
40. A. Eisenberg, Macromolecules, 1970, 3, 147–514.
41. M. Lee, J. K. Park, H. S. Lee, O. Lane, R. B. Moore, J. E. McGrath and D. G. Baird, Polymer, 2009, 50, 6129–6138.
42. S. Gu, G. H. He, X. M. Wu, Z. W. Hu, L. L. Wang, G. K. Xiao and L. Peng, J. Appl. Polym. Sci., 2010, 116, 852–860.
43. T. J. Peckham and S. Holdcroft, Adv. Mater., 2010, 22, 4667–4690.
44. H. W. Zhang and P. K. Shen, Chem. Rev., 2012, 112, 2780–2832.
45. M. J. He, D. H. Zhang, X. W. Chen and X. X. Dong, Polymer physics [M], Fudan University press, 3rd edn, 2007, p. 54.
46. A. F. M. Barton, Chem. Rev., 1975, 75, 731–753.
47. P. A. Small, J. Appl. Chem., 1953, 3, 71–80.
48. J. C. Tsai, C. K. Lin, J. F. Kuo and C. Y. Chen, J. Power Sources, 2010, 195, 4072–4079.
49. T. Yang, J. Membr. Sci., 2009, 342, 221–226.
50. P. O. Osifo and A. Masala, J. Power Sources, 2010, 195, 4915–4922.

---