Side Chain Engineering of Sulfonated Poly(arylene ether)s for Proton Exchange Membranes

Proton conductivity of proton exchange membranes (PEMs) strongly relies on microscopic morphology, which can be modulated by engineering the distribution of ionic groups. Herein, poly(arylene ether)s with densely distributed allyl functionalities are polymerized from a tetra-allyl bisphenol A monomer. The subsequent thiol-ene addition with sodium 3-mercapto-1-propanesulfonate yields comb-shaped sulfonated fluorinated poly(arylene ether)s (SFPAEs) with ion exchange capacities (IECs) ranging from 1.29 mmol·g−1 to 1.78 mmol·g−1. These SFPAEs exhibit superior proton conductivity over the whole temperature range, which is attributed to the enhanced hydrophilic/hydrophobic phase separation as evidenced by small angle X-ray scattering characterizations. The SFPAE-4-40 with an IEC of 1.78 mmol·g−1 shows the largest proton conductivity of 93 mS·cm−1 at room temperature under fully hydrated condition, higher than that of Nafion 212. Furthermore, the vanadium redox flow battery (VRFB) assembled with SFPAE-4-40 separator exhibits higher energy efficiency than the VRFB assembled with Nafion 212.


INTRODUCTION
Proton exchange membranes (PEMs) have gained application in more and more electrochemical devices recently, [1−5] where they serve not only as the conducting matrix for protons, but also as the barrier for electrical insulation between the electrodes. Generally, PEMs should meet several basic requirements, such as high proton conductivity, good chemical stability, robust mechanical strength, and low cost. [6−8] The state-of-the-art PEM is Nafion ® , a perfluorinated ionomer from DuPont, whose sulfonate groups are placed at the end of the long and flexible side chains. These sulfonate groups are capable of self-aggregating into ion-conductive channels for efficient proton transport. [9,10] However, Nafion ® suffers from high fuel permeability and manufacturing cost, which limit its wide application. [11,12] Therefore, lots of works have been devoted to the development of alternative PEMs.
Recently, sulfonated hydrocarbon polymers with nonfluorinated or partially fluorinated rigid backbones, such as sulfonated poly(arylene ether sulfone)s, [13,14] poly(arylene ether ketone)s, [15,16] poly(phenylene)s, [17] polyimides [18,19] and polybenzimidazoles, [20] have emerged as low-cost and high-performance alternatives to Nafion ® . It has been found that the proton conductivities of the randomly sulfonated polymers are usually low at certain ion exchange capacity, because of the lack of nanophase-separated morphology. [21,22] To enhance the nanophase separation, many strategies on molecular design have been proposed, such as dense sulfonation at particular segments of polymers or the attachment of lengthy and flexible ionic side chain. [23−25] Indeed, distinct nanophase separation can be achieved by such design, as evidenced by small-angle X-ray scattering, transmission electron microscopy, or atomic force microscopy, and improved proton conductivity can be obtained. [26−28] While many successes have been achieved for high performance PEMs, new synthetic approaches to attach lengthy, especially densely located, ionic side chains for enhanced phase separation are still highly desirable. Previously, we have reported the polymerization of sulfonated fluorinated poly(arylene ether)s (SFPAEs) featuring with (i) low polymerization temperature, (ii) 7 atoms spacing between sulfonic acid group and aromatic backbones, and (iii) two alkyl sulfonate side chains per monomer unit. [29] To further promote the phase separation, herein, we report the synthesis of a new tetra-allyl bisphenol A monomer and the resulting sulfonated fluorinated poly(arylene ether)s copolymers containing doubled density of alkyl sulfonate side chains per monomer unit compared with the previous ones.

EXPERIMENTAL Materials
All reactants and solvents were purchased from local chemical suppliers and used as received.

Membrane Formation
Solution-casting method was employed to prepare the membranes. 0.08 g/mL of SFPAEs in DMAc were cast onto leveled glass plates, which was conditioned in an oven at 80 °C. After 24 h of drying, the membranes were peeled off the glass surface by immersing in ultrapure water. Then, the membranes were immersed in 1 mol·L −1 H 2 SO 4 solution to change the counter cation to proton. The excess H 2 SO 4 in the membranes was removed by immersing in ultrapure water at 80 °C with replacing fresh water several times until pH of water is neutral. The thickness of all samples is ~50−55 μm.

Characterizations
The small angle X-ray scattering (SAXS) was measured on an Anton Paar SAXSess mc 2 instrument. The 1 H-NMR spectra were obtained on a Bruker AVANCE 400 spectrometer. The thermogravimetric analysis (TGA) was performed on a NETZSCH STA449C analyzer with a ramping rate of 10 °C·min −1 and under the protection of N 2 atmosphere (100 mL·min −1 ). FTIR spectra were achieved on a Thermo Nicolet 5700 FTIR spectrometer. The tensile properties were measured on a Suns universal testing instrument UTM 6502X with stretching rate of 5 mm·min −1 . Gel permeation chromatography (GPC) analysis was performed on a Waters Breeze system equipped with a Waters Styragel column, Waters 515 HPLC pump, and Waters 2414 refractive index detector. Tetrahydrofuran was used as elution solvent (flow rate: 1 mL·min −1 ) and polystyrene as standard for calibration. Inherent viscosities of ionomers in NMP (0.5 g·dL −1 ) were measured by an Ubbelohde viscometer. Electrochemical impedance spectroscopy (EIS) of membranes was measured by a Solartron 1260A impedance/Gain-phase analyzer in the frequency range from 1 Hz to 10 MHz at the voltage amplitude of 10 mV. Samples are fully hydrated in bulk water during measurements. Proton conductivity (σ) of membranes was calculated by where L is the distance between electrodes, R is the resistance of the sample, which was obtained from the diameter of semicircle in EIS, and A is the transverse area of the sample. The ion exchange capacity (IEC) was determined by acidbase titration according to literature protocol. [29] The water uptake and swelling ratio were defined as the weight and length changes of the dry membranes, respectively, after being soaked in ultrapure water at each testing temperature for 24 h. The oxidative stability was evaluated by immersing in Fenton's reagent (3% H 2 O 2 + 2 ppm FeSO 4 ) at 80 °C with regular shaking and recording the time for the samples starting to break. VO 2+ permeability was measured according to previously established method. [30] The concentration of VO 2+ was determined by a UV-5800HPC UV-Vis spectrophotometer. The vanadium redox flow batteries (VRFBs) were assembled according to literature. [31] 100 mL of 1.0 mol·L −1 VO 2+ in 2.0 mol·L −1 H 2 SO 4 solution was used as the anolyte; 50 mL of 1.0 mol·L −1 V 2+ in 2.0 mol·L −1 H 2 SO 4 solution was used as the catholyte. The flow rate is 200 mL·min −1 . The voltage window for testing VRFBs was 0.7−1.7 V. The membrane active area in the VRFBs was 15 cm 2 . The electrolytes were pumped unceasingly through the cell during the tests. The Coulombic efficiency (CE), voltage efficiency (VE), and energy efficiency (EE) of the VRFBs were calculated by the following equations: where I d and I c are the discharge and charge currents, respectively; V d and V c are the average discharge and charge voltages, respectively; t is the time taken for charge/discharge.

Synthesis of Monomers and Polymers
Claisen rearrangement has been widely used to make di-allyl bisphenol monomers; [32,33] however, repeated Claisen rearrangement is seldom reported, probably because of the potential difficulty of the second step reaction in the presence of large steric hindrance. In order to develop multi-allyl bisphenol monomers as dense reaction sites for the introduction of sulfonic acid groups, a tetra-allyl bisphenol monomer TABPA was designed from repeated Claisen rearrangement, i.e. reacting DABPA with allyl bromide to yield the intermediate a, which was then subjected to a second Claisen rearrangement (Scheme 1). The proposed structures of compound a and TABPA were confirmed by 1 H-NMR, as shown in Fig. 1. In the spectrum of compound a (Fig. 1a), the signals at 5.08, 5.27, and 5.88 ppm have an integral ratio of 2:2:1, which suggest the presence of O-CH 2 -CH＝CH 2 . After thermal rearrangement, a new signal at 5.47 ppm is observed (Fig. 1b), which can be assigned to the protons of phenol groups (Ar-OH). Also, the signals for the allyl ether groups disappear. All the other protons are well assigned in the 1 H-NMR spectroscopy (Fig. 1b), suggesting the successful synthesis of TABPA. The polymerization of bisphenols with decafluorobiphenyl catalyzed by CsF can be carried out at room temperature without crosslinking because of the high reactivity of decafluorobiphenyl. [29,34] The chemical structures of FPAE-4-ks and SFPAE-4-ks were confirmed by 1 H-NMR. The representative 1 H-NMR spectrum of FPAE-4-25 is shown in Fig. 1(c). The signals 2, 3, and 4 prove clearly the existence of allyl group after polymerization, which cannot be achieved by conventional K 2 CO 3 -activated polymerizations. [35] The FPAE-4-ks were sulfonated via thiol-ene addition at 80 °C to give SFPAE-4-ks. The 1 H-NMR spectrum of SFPAE-4-25 is shown in Fig. 1(d). No resonance signals for the allyl protons can be observed, but new resonance signals corresponding to the alkyl protons appear. Therefore, the thiol-ene addition, resulting in alkyl sulfide products, is successful and efficient. 19 F-NMR spectra were used to confirm the structure of the decafluorobiphenyl units in FPAE-4-25 and SFPAE-4-25 (Fig. 2). It can be seen that both FPAE-4-25 and SFPAE-4-25 have two sets of signals at around −137 and −152 ppm, which demonstrate the successful incorporation of decafluorobiphenyl moieties into the linear chain of the polymers in the manner depicted in Scheme 2.
The attachment of sulfonated side chains was also valid-ated by the FTIR, as shown in Fig. 3. Compared with the control sample FPAE-0 without allyl group, the FPAE-4-25 shows a new absorption band at 914 cm −1 in the FTIR spectrum, which can be attributed to the vibration of the allyl groups. After the attachment of sulfonated side chains (SFPAE-4-25), the band for allyl groups disappears, while two absorptions bands at 1238 and 1045 cm −1 show up, which are the symmetric and asymmetric vibration of the sulfonate groups. Therefore, these FTIR results are consistent with the above NMR results, demonstrating the successful synthesis of SFPAE-4-ks.
Four FPAE-4-ks with TABPA loadings being 25%, 30%, 35%, Chemical shift (ppm) Chemical shift (ppm)  and 40%, respectively, were polymerized. It can be seen from Table 1 that the inherent viscosities of FPAE-4-ks are relatively high, and the number-average molecular weights are in the range of 38 kDa to 54 kDa. After thiol-ene addition, the inherent viscosities become higher (Table 2) because of stronger intermolecular interactions. The FPAE-4-ks are readily soluble in CHCl 3 , CH 2 Cl 2 , and aprotic solvents, while the SFPAE-4-ks are only soluble in aprotic solvents. The ion exchange capacities (IECs) of SFPAEs are titrated to be ranging from 1.29 mmol·g −1 to 1.78 mmol·g −1 . Higher content of TABPA unit in the polymer backbones leads to higher IEC, as expected.

Membrane Morphology
To study the effect of side chain distribution on the properties of membranes, a previously reported sulfonated fluorinated poly(arylene ether) bearing di-alkylsulfonate side chains (SFPAE-2-50), [29] which has identical backbone and same content of side chains to SFPAE-4-25 but different distributions of the side chains, is selected as control sample, as depicted in Fig. 4(a). The aggregation of ionic groups usually leads to the presence of an ionomer peak in SAXS patterns, which corresponds to the distance (d) between the ionic aggregates. [25,36−38] This d can be calculated by the equation of d = 2π/q, where q refers to the scattering vector of the ionomer peaks. The impact of side chain density on the microscopic morphology of SFPAEs was investigated by small angle X-ray scattering (SAXS), as shown in Fig. 4(b). As can be seen, the qs of SFPAE-4-25 and SFPAE-2-50 are 1.34 and 1.84 nm −1 , respectively, corresponding to the ds of 4.7 and 3.4 nm, respectively. The larger d of SFPAE-4-25 demonstrates better aggregation of its sulfonic acid groups, suggesting that the molecular design of densely distributed ionic side chains is successful.

Water Affinity and Proton Conductivity
The water uptake and swelling ratio of SFPAE-4-ks, SFPAE-2-50 and Nafion 212 as a function of temperature are shown in Figs. 5(a) and 5(b), respectively. Both the water uptake and swelling ratio for SFPAE-4-ks increase with the increase of IEC or temperature. All of the SFPAE-4-ks exhibit modest water uptake (20.7%−45.0%) and swelling ratio (12.4%−27.9%) at room temperature, while these values are higher than those of Nafion 212. Notably, SFPAE-2-50 has slightly lower water uptake and swelling ratio than SFPAE-4-25 at all the tested temperatures. This result suggests that the increase of side chain density slightly increases the water affinity of the polymer. The proton conductivities of SFPAE-4-ks, SFPAE-2-50, and Nafion 212 measured under fully hydrated conditions are gathered in Figs. 5(c) and 5(d). It can be seen that the proton conductivity of SFPAE-4-ks increases gradually either with the increase of IEC at room temperature or with the increase of temperature. The SFPAE-4-40 with the highest IEC of 1.78 mmol·g −1 exhibits the highest proton conductivity of 93 mS·cm −1 at room temperature among all SFPAE-4-ks.
The SFPAE-4-25 has a nearly identical IEC to SFPAE-2-50; however, its room temperature proton conductivity is 24% higher than that of the latter. It has been well known that the excellent proton conductivity of Nafion 212 is a result of its highly aggregated ionic groups, which forms proton conductive channels for efficient proton transport. [10] Indeed, the proton conductivity of Nafion 212 is measured to be as high as  85 mS·cm −1 at room temperature. Therefore, the higher proton conductivity of SFPAE-4-25, compared with that of SFPAE-2-50, can be attributed to its more aggregated ionic groups. For sulfonated poly(arylene ether)s to possess higher proton conductivity than Nafion, much larger IEC is usually required. [1,39] The SFPAE-4-40 achieves higher proton conductivity at an IEC of 1.78 mmol·g −1 , suggesting its promising prospect as alternative to Nafion.

Thermal Stability, Mechanical Property, and Oxidative Stability
The thermal stabilities of FPAE-4-ks and SFPAE-4-ks were investigated by TGA, and the 5% weight-loss temperatures (T d-5% ) are listed in Table 3. It can be seen that all of the FPAE-4ks exhibit excellent thermal stability, a typical property of the robust poly(arylene ether) backbones. The higher content of tetra-allyl groups leads to a slight reduction in T d-5% for FPAE-4ks. After thiol-ene addition, the T d-5% s are lowered dramatically to about 240−280 °C, which is the typical decomposition temperature range of sulfonic acid groups. [40,41] The higher content of tetra-alkylsulfonate groups leads to an obvious reduction in T d-5% for SFPAE-4-ks, as expected. All of SFPAE-4ks are thermally stable up to > 200 °C, suggesting their excellent thermal stability for applications in vanadium redox flow batteries (VRFBs) or H 2 /O 2 fuel cells.
The mechanical strength of PEMs is an essential requirement for device fabrications and operations. The tensile strength and elongation at break of SFPAE-4-ks freshly taken out from deionized water are listed in Table 3. Both the tensile strength and elongation at break increase with the decrease of IEC. This is because more water was absorbed into the mambranes at higher IEC, resulting in more loosely packed macromolecular chains. The effect of water on mechanical properties can also be supported by the sigfinicantly higher tensile strength of dry membranes than wet membranes in literature. [42−44] All of the SFPAE-4-ks show good mechanical properties with tensile strength and elongationat-break in the range of 10-17 MPa and 7%-11%, respectively.
The oxidative stability of PEMs is an essential requirement for long-term device operations. The oxidative stability of SFPAE-4-ks determined by immersing in Fenton's reagent (3 wt% H 2 O 2 + 2 ppm FeSO 4 ) at 80 °C with constant shaking is listed in Table 3. It can be seen that the oxidative stability of SFPAE-4-ks increases with the decrease of IEC, which can be attributed to the higher uptake of degradative radicals in Fenton's reagent at higher IECs. [29] Notably, the oxidative stabilities of the SFPAE-4-ks are significantly higher than those of non-fluorinated membranes under similar condi- tions. [40,45,46] Therefore, the presence of fluorine in the macromolecular structures is highly beneficial to enhanced oxidative stabilities.

Flow Battery Application
To demonstrate the applicability of SFPAE-4-ks as high performance PEMs, the SFPAE-4-40 was evaluated as separator for VRFB. Ideally, VRFBs requires separators to be able to completely block the permeation of vanadium ions to ensure low self-discharge rate. [47−49] It is found that the VO 2+ permeability of SFPAE-4-40 is 4.1 × 10 −12 m 2 ·s −1 , lower than the 5.36 × 10 −12 m 2 ·s −1 of Nafion 212. The selectivity of membranes, which is defined as the ratio of proton conductivity over VO 2+ permeability, is calculated to be 3.78 × 10 4 S·min·cm −3 for SFPAE-4-40 and 2.64 × 10 4 S·min·cm −3 for Nafion 212. In conclusion, the SFPAE-4-40 has higher proton conductivity, lower VO 2+ permeability, and thus higher selectivity than Nafion 212. The charge/discharge curves of the VRFBs assembled with SFPAE-4- 40 and Nafion 212 at a current density of 40 mA·cm −2 are shown in Fig. 6. It can be seen that the VRFB assembled with SFPAE-4-40 has longer charge/discharge time than the VRFB assembled with Nafion 212. The CE of the VRFB assembled with SFPAE-4-40 is calculated to be 90.9%, higher than the 86.9% of the VRFB assembled with Nafion 212. The VE of the VRFB assembled with SFPAE-4-40 battery is calculated to be 87.0%, also higher than the 84.3% of the VRFB assembled with Nafion 212. Thus, the overall EE of the VRFB assembled with SFPAE-4-40 (79.1%) is higher than that of the VRFB assembled with Nafion 212 (73.3%). Therefore, the SFPAE-4-40 can be a promising separator for VRFBs. It should be noted that the proton conductivity of SFPAE-4-ks may keep increasing with the increase of IEC, but the water uptake would increase as well, which would decrease the mechanical strength and increase the vanadium permeability. Besides, the thickness of membrane, treatment of electrode, concentration of electrolyte, configuration of cell, and operation parameters of VRFB would affect the overall device performance of the membrane. Therefore, the further optimization of the separator is needed for practical applications. This study provides the synthesis of a new series of comb-shaped proton exchange membranes from a reaction route developed by our lab previously for the establishment and commercialization of potential electrochemical systems.

CONCLUSIONS
Comb-shaped sulfonated fluorinated poly(arylene ether)s bearing tetra-alkylsulfonate side chains (SFPAE-4-ks) were successfully synthesized from room temperature condensation polymerization of a newly developed tetra-allyl bisphenol A monomer followed by thiol-ene addition with sodium 3mercapto-1-propanesulfonate. The IECs were in the range of 1.29-1.78 mmol·g −1 . It is found that the SFPAE-4-25 had higher proton conductivity than a previous reported SFPAE-2-50, which has identical backbone and same content of side chains to SFPAE-4-25 but different distributions of the side chains. This was rationalized by the SAXS results where the SFPAE-4-25 had better aggregation of sulfonic acid groups to form proton conductive channels than SFPAE-2-50. The SFPAE-4-40 with an IEC of 1.78 mmol·g −1 exhibited higher proton conductivity, lower VO 2+ permeability, and higher selectivity than Nafion 212. Accordingly, the vanadium redox flow battery (VRFB) assembled with SFPAE-4-40 exhibited higher Coulombic efficiency, voltage efficiency and energy efficiency than the VRFB assembled with Nafion 212 at a current density of 40 mA·cm −2 . Therefore, side chain engineering is a versatile protocol for the development of high performance PEMs.