2022, Vol.95, No.5

We designed and synthesized a series of sulfonated terpolymers, SPP-PQP and SPP-BQP, containing sulfophenylene, quinquephenylene (QP) and perfluoroalkylene (PAF) or hexafluoroisopropylidene (BAF) groups whose compositions were optimized to achieve high-performance proton conductive membranes. In both series, the terpolymers were obtained as high-molecular-weight (Mn = 19.0–32.9 kDa, Mw = 99.9–198 kDa) providing bendable and transparent membranes with supposed ion exchange capacity (IEC) values (2.41–2.68 meq. g−1) by solution casting. SPP-PQP membranes exhibited higher water uptake, higher proton conductivity, and larger strain at break with increasing the PAF composition. In contrast, the membrane properties were less sensitive to the composition in the SPP-BQP membranes. Among the terpolymer membranes investigated, SPP-PQP50 showed the best-balanced properties in terms of low water uptake, high proton conductivity, and high mechanical properties probably because highly hydrophobic aliphatic PAF and more rigid QP groups both contributed to those relevant properties. Overall, the combination of different hydrophobic components in the terpolymers was effective in improving the properties of proton conductive membranes, that could not be achieved with the corresponding copolymer membranes.

Proton conductive membranes or proton exchange membranes (PEMs) have been utilized in various energy devices such as fuel cells, batteries, water and chlor-alkali electrolyzers. Perfluorosulfonic acid (PFSA) ionomer membranes are commercialized and have been most used for these purposes due to high ion conductivity and reasonable chemical/physical stability under the device operating conditions.13 PFSA ionomers, however, do not carry much freedom in synthesis and chemical structures and thus, may not be feasible in properties alternations required for each different application.

As one of the alternative PEMs, acid-functionalized aromatic polymers have been extensively investigated in the last two decades.46 Those include sulfonated poly(arylene ether)s,710 poly(phenylene)s,11,12 and acid-doped poly(benzimidazole)s.13,14 Among them, poly(phenylene)-based PEMs have markedly progressed in recent years with improved proton conductivity, gas impermeability, and chemical stability.12,15,16 We have successfully developed sulfonated polyphenylenes (SPP-QP) composed solely of sulfonic acid groups and phenylene rings, that showed thin membrane forming capability.11 Polyphenylenes devoid of heteroatom linkages in the main chain led to highly oxidative stability of the resulting membranes, that had not been achieved with other aromatic ionomers.11,12 More recently, we found that some fluorinated groups (-CF2-, -CF3) were effective in tuning the membrane properties of the sulfonated polyphenylenes. In fact, sulfonated polyphenylenes (SPAF) containing perfluoroalkylene (PAF) groups provided membranes with much improved mechanical properties because of increased freedom in rotational degrees of the main chain.17 Introducing PAF components, however, resulted in excessive swelling in water and accordingly decrease in the ion conductivity. In contrast, sulfonated polyphenylenes (SBAF) containing hexafluoroisopropylidene (BAF) groups provided membranes with better dimensional stability and higher ion conductivity at the expense of some favorable mechanical properties.18

For further tuning and improving the membrane properties, terpolymer seems an attractive approach where two different hydrophobic components can be incorporated. For example, Holdcroft et al. reported that water affinity of sulfonated polyphenylene terpolymers could be controlled by changing ortho- and meta-biphenylene composition.19 Kim et al. developed terpolymer-based PEMs, that exhibited high chemical stability by introducing branched structure to reduce the accessible sites for the highly oxidative radical attack in the hydrophilic components.20 We also reported terpolymer SPAF membranes containing quinquephenylene (QP) hydrophobic groups to achieve improved proton conductivity by suppressed water uptake, where the hydrophobic composition was not well-investigated.21

In the present study, we designed a series of sulfonated terpolymers, SPP-PQP and SPP-BQP, containing sulfophenylene, QP, and PAF or BAF groups in the main chain. In particular, effect of the hydrophobic compositions was investigated in detail. Phase-separated morphology, water absorbability, proton conductivity, gas permeability, and mechanical properties were measured to optimize terpolymer structure and composition for better performing PEMs.

2.1 Materials.

Hydrazine monohydrate, benzene-1,4-diboronic acid, potassium carbonate, N,N-dimethylformamide (DMF), ethylene glycol dimethyl ether (DME), sodium chloride (NaCl), sodium hydrogen sulfite (NaHSO3), sodium hydroxide (NaOH), Cu powder, dimethyl sulfoxide (DMSO), dimethyl sulfoxide (DMSO) (dehydrated), toluene (dehydrated), sulfuric acid (96%), hydrochloric acid (35%), potassium carbonate (K2CO3), bis(1,5-cyclooctadiene)nickel(0) (Ni(cod)2), ethanol, methanol, ethyl acetate, hexane, chloroform, dichloromethane, and dimethyl sulfoxide-d6 (DMSO-d6 with 0.03% tetramethylsilane (TMS), 99.9 atom% D) were purchased from Kanto Chemical Co. and used as received. Palladium(II) chloride (PdCl2), triphenylphosphine, 3-bromo-3′-chloro-1,1′-biphenyl, 3-iodoaniline, sodium nitrite (NaNO2), 2,5-dichlorobenzenesulfonic acid dihydrate (SP), and 2,2′-bipyridyl were purchased from TCI Inc. and used as received. Copper(II) sulfate pentahydrate, 4,4′-(hexafluoroisopropylidene)dianiline were purchased from FUJIFILM Wako Pure Chemical Corp. and used as received. 6-Diiodo-dodecafluorohexane was kindly provided by Tosoh Finechem Corp. and used as received. Quinquephenylene (QP), perfluoroalkylene (PAF), and hexafluoroisopropylidene (BAF) monomers were prepared according to the literature.

2.2 Synthesis of SPP-PQP and SPP-BQP.

SPP-PQP and SPP-BAP of three different compositions were synthesized. A typical procedure is as follows. QP, PAF, SP (see Table 1 for detail), 2,2′-bipyridine (9.69–9.85 mmol), DMSO (12.0 mL), toluene (10.0 mL), and K2CO3 (1.71 mmol) were added into a 100-mL three-neck flask equipped with a magnetic stirring bar, a Dean-Stark trap, a reflux condenser, and a nitrogen inlet/outlet. After azeotropic dehydration at 170 °C for 3 h, the mixture was cooled to 80 °C and Ni(cod)2 (6.46–6.56 mmol) was added. After the polymerization at 80 °C for 3 h, the mixture was cooled to room temperature, diluted with DMSO and poured into a large excess of 6 M HCl aqueous solution. The precipitate was washed with 6 M HCl twice and deionized water three times. Drying in a vacuum oven at 45 °C for 12 h gave SPP-PQP as a dark brown solid. SPP-BQP were synthesized using BAF in place of PAF in a similar manner.

Table
Table 1. Summary of polymerization reaction for SPP-PQP and SPP-BQP.
Table 1. Summary of polymerization reaction for SPP-PQP and SPP-BQP.
Sample QP
(M)
PAF or BAF
(M)
SP
(M)
Yield
(%)
Mn
(kDa)
Mw
(kDa)
IECNMRa
(meq. g−1)
IECtitrationb
(meq. g−1)
SPP-PQP33 0.033 0.016 0.52 96 23.4 166 2.83 2.53
SPP-PQP50 0.024 0.024 0.52 99 20.4 163 2.55 2.41
SPP-PQP67 0.015 0.031 0.52 99 32.9 198 2.80 2.68
SPP-BQP33 0.037 0.019 0.52 99 20.1 116 2.07 2.58
SPP-BQP50 0.029 0.029 0.52 99 25.0 162 2.34 2.66
SPP-BQP67 0.020 0.040 0.52 97 19.0 99.9 2.25 2.47

aCalculated from 1H NMR spectra. bObtained by titration.

2.3 Membrane Preparation.

DMSO solutions of the SPP-PQP and SPP-BQP polymers (5.0 wt.%) were cast onto a flat glass plate and dried in air for 24 h. The membranes were carefully removed from the glass plate, immersed in 1 M H2SO4 aqueous solution for 24 h, washed with deionized water thoroughly, and dried at ambient temperature.

2.4 Measurements.

Apparent molecular weights were measured with gel permeation chromatography (GPC) equipped with a Shodex K-805L column and a Jasco 805 UV detector and calibrated with standard polystyrene samples. DMF containing 0.01 M LiBr was used as eluent. 1H (500 MHz) and 19F (471 MHz) NMR spectra were recorded on a JEOL JNM-ECA 500 at 80 °C in DMSO-d6 as solvent and TMS as an internal reference. Ion exchange capacity (IEC) of the membranes was measured as follows: a small piece of dried membrane (ca. 20 mg) was immersed in a large excess of 2 M NaCl aqueous solution for 24 h; the H+ in the solution released from the membrane was titrated with standard 0.01 M NaOH aqueous solution. For transmission electron microscopic (TEM) observation, the membranes were stained with lead ions by immersing the samples in 0.5 M Pb(OAc)2 aqueous solution and rinsed with water. The membranes were sectioned into 50 nm slices using a Leica microtome Ultracut UCT. The slices were placed on a copper grid and analyzed with a Hitachi H-9500 TEM at the acceleration voltage of 200 kV. The size of the hydrophilic domains was calculated by averaging diameters of more than 200 domains. Proton conductivity and water uptake were measured at 80 °C by a solid electrolyte analyzer system (MSBAD-V-FC, Bel Japan Co.) equipped with a temperature and humidity controllable chamber and magnetic suspension balance. The water uptake was calculated as follows. Water uptake = ((weight of hydrated membrane) − (weight of dry membrane))/(weight of dry membrane) × 100%. The membranes were kept at 80 °C for 3 h under vacuum to obtain the dry weight and exposed to the set humidity for at least 2 h to obtain the hydrated weight. In-plane proton conductivity (σ) of the membranes was measured by ac impedance spectroscopy (Solartron 1255B and 1287) simultaneously in the same chamber, and calculated according to the following equation: σ = l/(A × R), where l is the distance between the two inner wire electrodes (l = 1 cm) and A is the conducting area, respectively. Ion conducting resistances (R) were determined from the impedance plot measured over the frequency range from 1 to 105 Hz. The tensile strength was measured with a Shimadzu AGS-J 500 N universal testing instrument attached with a Toshin Kogyo Bethel-3A temperature and humidity controllable chamber at a tensile rate of 10 mm min−1. Samples cut into a dumbbell shape [DIN-53504-S3, 35 mm × 6 mm (total) and 12 mm × 2 mm (test area)] were used. Before the test, the samples were equilibrated at 80 °C and 60% RH for at least 3 h. Gas permeability was measured with a GTR-Tech 110XADF gas permeability measurement apparatus. A membrane sample was set in a cell that had gas inlet and outlet on both sides of the membrane. Prior to the measurement, both sides were exhausted by vacuum pump for 10 h. Then, test gas was supplied to one side of the membrane at a pressure of 201.30 kPa. The change of the pressure on the other side was monitored for 10 min at 80 °C. The gas permeability coefficient, (Q (cm3 (STD) cm cm−2 s−1 cmHg−1) was calculated using the following equation: Q = dp/dt × (V × 273)/(760 × (273 + T)) × 1/A × 1/P × l, where dp/dt is a slope of the pressure-time plot, V (cm3) is the volume of the cell in high pressure side, T (K) is the absolute temperature, A (cm2) is the permeation area, P (cmHg) is the pressure of applied gas, and l (cm) is the thickness of the membrane.

3.1 Synthesis and Characterization of Terpolymers.

Series of SPP-PQP and SPP-BQP were synthesized from QP, PAF or BAF, and SP monomers via C-C coupling polymerization reaction using Ni(cod)2 as a promoter (1.5 equimolar to the terminal chlorine groups in the monomers, Scheme 1). The feed composition of QP/PAF or BAF as the hydrophobic components was set at 1/0.5, 1/1 and 0.5/1, while the feed composition of SP monomer was constant in order to obtain the terpolymers containing the same concentration of sulfonic acid groups (or ion exchange capacity: IEC = 3.10 mequiv g−1). The polymerization proceeded well within a few hours and the terpolymers were obtained in reasonably high yields (>96%, Table 1). The terpolymers were soluble in polar organic solvents such as DMSO and DMF. GPC analyses indicated that the polymers were of high-molecular-weights (Mn = 19.0–32.9 kDa and Mw = 99.9–198 kDa) with reasonable polydispersity. From the molecular weight data, the reactivity of each monomer was roughly estimated in the order of PAF > QP > BAF > SP. The chemical structure of SPP-PQP and SPP-BQP was analyzed by 1H and 19F NMR spectra (Figures 1, S1 and S2). 1H NMR spectra of SPP-PQP were the same as those of our previous polymers with different terpolymer composition,21 and the integrals of the aromatic protons enabled estimation of the IECs of the terpolymers to be 2.55–2.83 meq. g−1. 19F NMR spectra of SPP-PQP were well-assigned. 1H NMR spectra of SPP-BQP were also assignable, from which the compositions of the components were estimated. The resulting IEC values calculated from the 1H NMR spectra were ca. 2.07–2.34 meq. g−1 and somewhat lower than those calculated from the feed compositions probably because of the lower reactivity of the SP monomer than the hydrophobic monomers. (Please also note that the aromatic proton peaks in the 1H NMR spectra were overlapped and thus, the IEC from the peak integrals might contain errors.)

Casting the terpolymer solutions in DMSO provided brown, bendable membranes. The IEC values of the SPP-PQP and SPP-BQP membranes were obtained by acid/base titration to be 2.41–2.68 meq. g−1, which were in fair agreement with those obtained from the 1H NMR spectra.

3.2 Morphology of Terpolymer Membranes.

Morphology of the SPP-PQP and SPP-BQP membranes was observed by TEM images for samples stained with Pb2+ ions (Figures 2 and 3). Both membranes exhibited phase-separated morphology where the dark areas represent hydrophilic domains and the bright areas represent hydrophobic domains. The average domain size was plotted as a function of the composition of the fluorinated component (PAF and BAF) in Figure 4. In both series of the terpolymer membranes, the domain size exhibited volcano-type dependence on the fluorinated component composition with the largest hydrophilic and hydrophobic domains at 50 mol%. It is noted that not only the hydrophobic domain but also the hydrophilic domain was affected by the hydrophobic composition. While the differences were rather minor, IEC values seemed to affect more the morphology taking into account our previous data on SPP-PQP membranes with different IECs.21

The morphology was further investigated by SAXS analyses under humidified conditions (Figure S3), where a broad and small peak was observed at d = 8–10 nm for all terpolymer membranes. The d value was larger than the domain sizes observed in the TEM images because of the presence of the water, however, the peak diminished with increasing humidity indicating that the periodic structure became randomized.

3.3 Water Uptake and Proton Conductivity.

Water uptake of SPP-PQP and SPP-BQP membranes was measured at 80 °C and different humidity (Figure 5). For comparison, data of the parent copolymers, SPP-QP, SPAF and SBAF membranes, are also included. The terpolymer membranes exhibited similar water uptake and humidity dependence because of the similar IEC value. Then, the water uptake was re-plotted as a function of the PAF and BAF composition in Figure 7a and b. SPP-PQP exhibited a volcano-type dependence of the water uptake on the PAF composition, where the maximum water uptake was confirmed at 67 mol% of the PAF composition. The dependence was more significant at higher humidity. In fact, the water uptake was 84 wt% for SPP-PQP67 membrane at 95% relative humidity (RH), compared to 56 wt% for SPP-PQP33 and 52 wt% for SPP-PQP50 membranes. It is considered that co-existence of QP and PAF components with large differences in the molecular structures must have produced larger free volume (in which excess water was located) compared to that of the parent copolymers (SPP-QP and SPAF). Similar dependence was observed in the water uptake of SPP-BQP terpolymers, while the effect of the BAF composition was relatively smaller compared to SPP-PQP. The results suggest that PAF groups would have larger conformation than that of BAF groups.

The proton conductivity of SPP-PQP and SPP-BQP membranes was measured under the same conditions as for the water uptake (Figure 6). Among the terpolymer membranes, SPP-PQP67 membrane exhibited the highest proton conductivity at any humidity investigated; 4.55 mS cm−1 at 20% RH and 408 mS cm−1 at 95% RH. The proton conductivity is also plotted as a function of the PAF and BAF composition in Figure 7c and d. The conductivity showed similar (but not exactly the same) dependence on the PAF and BAF composition to water uptake in Figure 7a and b. The results are reasonable since hydrated protons are the ion carriers in those membranes. Since SPP-BQP67 membrane did not show periodic structure and ionic clusters (in the SAXS profile), proton diffusion would be slower and thus, proton conductivity was lower than those of SPP-BQP50 membrane. In both series, the terpolymer membranes exhibited higher proton conductivity than that of their parent copolymer membranes, proving the effective molecular strategy of terpolymers for better-performing proton conductive membranes.

3.4 Mechanical Properties.

Mechanical properties of SPP-PQP and SPP-BQP membranes were investigated by tensile test at 80 °C and 60% RH, and compared to those of the SPP-QP, SPAF and SBAF membranes (Figure 8). Young’s modulus, maximum stress, and maximum strain are plotted as a function of the terpolymer composition in Figure 9. In SPP-PQP membranes, Young’s modulus and maximum stress decreased and the maximum strain increased as increasing the PAF composition due to more flexible nature of the aliphatic components. In SPP-BQP membranes, the mechanical properties were much less sensitive to the terpolymer composition probably because both BAF and QP were composed mainly of phenylene groups. In particular, Young’s modulus and maximum stress were very similar for all SPP-BQP membranes. Among SPP-PQP and SPP-BQP terpolymer membranes, SPP-PQP50 exhibited high mechanical properties, especially large strain at break. Since the molecular weight was not correlated well with these mechanical parameters, the hydrophobic composition was responsible for the improved tensile properties.

3.5 Gas Permeability.

SPP-PQP50 membrane with balanced conductivity/mechanical stability was selected for gas permeability measurement. As summarized in Table 2, SPP-PQP50 membrane showed similar hydrogen permeability and slightly lower oxygen permeability compared to those of SPP-QP copolymer membrane at 80 °C. The permeability coefficient was then divided into diffusion and solubility coefficients. The oxygen diffusion coefficient declined by introducing the PAF component, while the solubility coefficient was less dependent on the hydrophobic component. The perfluorohexylene groups tended to have more crystalline structure with closer molecular packing, which might have caused slower diffusion of larger oxygen molecules.

Table
Table 2. Gas permeability coefficient, diffusion coefficient, and solubility coefficient of SPP-PQP50 membrane.
Table 2. Gas permeability coefficient, diffusion coefficient, and solubility coefficient of SPP-PQP50 membrane.
Membrane Gas Permeability
coefficient
(cm3 cm cm−2 s−1 cmHg−1)
Diffusion
coefficient
(cm2 s−1)
Solubility
coefficient
(cm3 cm−3 cmHg−1)
SPP-PQP50 H2 1.50 × 10−9 1.18 × 10−6 1.28 × 10−3
O2 9.10 × 10−11 7.62 × 10−8 1.20 × 10−3
SPP-QP H2 1.18 × 10−9 1.26 × 10−6 9.40 × 10−4
O2 2.86 × 10−10 2.43 × 10−7 1.20 × 10−3

Two series of sulfonated terpolymers, SPP-PQP and SPP-BQP, were designed and synthesized as proton conductive membranes. Compared to their parent copolymer membranes, SPP-PQP containing perfluoroalkyl (PAF) and quinquephenyl (QP) groups as hydrophobic components exhibited higher water uptake, higher proton conductivity, and larger strain at the break with increasing the PAF composition. The highest proton conductivity (408 mS cm−1 at 80 °C and 95% RH) was achieved with SPP-PQP67 containing 67 mol% PAF. In contrast, effect of the hydrophobic composition on the membrane properties was less pronounced for SPP-BQP containing hexafluoroisopropylidene (BAF) and QP groups as the hydrophobic components. Among the terpolymer membranes investigated, SPP-PQP50 showed the best-balanced membrane properties in terms of water uptake, proton conductivity, and mechanical properties probably because highly hydrophobic aliphatic PAF and more rigid QP groups both contributed to those relevant properties. Unlike the commercial perfluorinated PEMs (such as Nafion), SPP-PQP50 exhibited low hydrogen and oxygen permeability typical for aromatic ionomer membranes. As a conclusion, we have confirmed that terpolymer molecular strategy is promising for designing better functioning proton conductive membranes than copolymer membranes.

This work was partly supported by the New Energy and Industrial Technology Development Organization (NEDO), the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan, through Grants-in-Aid for Scientific Research (18H05515), Japan Science and Technology (JST) through SICORP (JPMJSC18H8), JKA promotion funds from AUTORACE, Iwatani Naoji Foundation, and the thermal and electric energy technology foundation. We thank Tosoh Fine Chem for kindly supplying 1,6-diiodoperfluorohexane.

Synthesis and NMR spectra of monomers and polymers, and SAXS profiles. This material is available on https://doi.org/10.1246/bcsj.20220057.

Kenji Miyatake

Prof. Kenji Miyatake received his PhD degree in polymer chemistry from Waseda University, Japan in 1996. He was a postdoc (JSPS overseas research fellow) at McGill University, Canada, from 1999 to 2001. In 2001, he was appointed an associate professor in Clean Energy Research Center at the University of Yamanashi, where he currently serves as a professor. He also holds a professor position at Waseda University. He is a Fellow of the Royal Society of Chemistry.