Polymerized imidazolium ionic liquids crosslinking sulfonated poly(ether ether ketone) (SPEEK) for high-temperature proton exchange membrane

Abstract

An ionic liquid (IL) monomer of (acryloyloxy)propanylimidazolium chloride with unsaturated carbon–carbon double bonds was synthesized. The IL monomer clung to sulfonated poly(ether ether ketone) (SPEEK) polymer owing to the electrostatic attraction between its cations and sulfonic groups. Polymerized ionic liquid (PIL) crosslinked with the SPEEK matrix after the in situ polymerization of the unsaturated carbon–carbon double bonds in the system of SPEEK/PIL. As a result, the SPEEK/PIL membrane displays a better performance in terms of tensile stress and chemical stability compared to those for the SPEEK/IL membrane. In order to improve the proton conductivity of the SPEEK/PIL membrane, phosphoric acid (PA) molecules were introduced through hydrogen bonds between imidazolium cations and phosphoric acid molecules. In the case of the SPEEK/PIL/PA membrane, a proton-conducting network was formed with more free and continuous phosphoric acid molecule chains owing to more phosphoric acid molecules being doped relative to the SPEEK/IL/PA membrane. The maximum proton conductivity of the SPEEK/PIL/PA membrane was 4.5 × 10⁻² S cm⁻¹ at 160 °C under anhydrous conditions and a value of 2.5 × 10⁻² S cm⁻¹ was even stable for more than 300 h at 140 °C. Moreover, the SPEEK/PIL/PA membrane possesses satisfactory mechanical properties, such as 2.64 MPa for tensile stress at 120 °C.

---

1. Introduction

Over the past decade, proton exchange membrane fuel cells (PEMFCs) as a favorable candidate in the field of clean energy sources have attracted much attention owing to their merits of high efficiency, high power density, low emission of pollutants and simplicity.^1,3 Although PEMFCs have made great progress, a sustained effort is needed to improve the performance to that required for their participation in people's lives. Significantly, elevation of the operation temperature was considered to be an effective strategy for achievement.^4,5 As reported, high temperature operation provides some advantages for PEMFCs, including accelerating electrochemical dynamics, simplifying water and heat management, and enhancing Pt catalyst tolerance of CO poisoning.^2,6,7 Proton exchange membranes (PEMs) as the core of PEMFCs were thus expected to be durable in the temperature range of 373–473 K. At present, some advanced membrane electrolytes, including modified Nafion® membranes,^8–10 acid–base polymer membranes,^11,12 inorganic composite membranes,^13,14 polymer/IL membranes^15,17 and PIL membranes,^3,18 have been synthesized in order to cater for the needs of high temperature operation.

In the case of ILs, they have some advantages of high ionic conductivity and immeasurably low vapor pressure.^19–23 ILs could donate protons by maintaining a channel for proton transfer and thus they are used to prepare conductive membranes in combination with flexible polymers. As a critical factor, the compatibility of ILs with polymers dominated the properties and even the application prospects of membrane electrolytes.²⁴ The research thus far has been confined to ILs with imidazolium cation and some flexible polymers.^5,25 Besides that, the polymerization of vinyl groups in IL monomers to form PIL membranes is another strategy that has been used to prepare membrane electrolytes.^19,26–29 The first report can be traced back to 1998, by Ohno.³⁰ After the pioneering work, a large number of membranes based on polycation-type ILs,^31,32 polyanion-type ILs,^33,34 copolymers,^35,36 poly(zwitterion)s³⁷ and microemulsion polymerization^38,39 were reported in succession. The research and application of PIL membranes thus entered a rapid development period. However, there was an inherent risk of weak mechanical strength owing to the lack of a supporter in the PIL membranes. Meanwhile, the movement of ions was inevitably limited after polymerization, resulting in a decline in conductivity.^31,40 For example, the ion conductivity of the N-vinylimidazolium tetrafluoroborate (VyImBF₄) monomer was 1.0 × 10⁻⁴ S cm⁻¹ at 30 °C, while the value was reduced to 2.0 × 10⁻⁹ S cm⁻¹ after polymerization.³¹ The introduction of proton carriers was an effective strategy to improve the conductivity, but a drop in tensile stress unavoidably occurred.⁴¹ The proton conductivity–tensile stress dilemma drove further research to explore PIL membranes with balanced properties.

In this paper, we attempted to solve this dilemma by crosslinking PIL with a SPEEK matrix via in situ polymerization of IL monomer. PIL was expected to enhance the tensile stress through cross-linking with the SPEEK polymer and also to improve proton conductivity through absorbing phosphoric acid molecules with hydrogen bonds. As a model system, the IL monomer with the imidazolium cation of (acryloyloxy)propanylimidazolium chloride was synthesized in consideration of the compatibility with SPEEK. In the case of the SPEEK/PIL membrane, the IL monomer clung to the SPEEK matrix owing to the electrostatic attraction and PIL crosslinked the SPEEK matrix owing to the polymerization of carbon–carbon unsaturated double bonds with 2,2′-azodiisobutyronitrile (AIBN) as the initiator in solution. The SPEEK/PIL/PA membrane was prepared by immersing the SPEEK/PIL membrane into pure phosphoric acid at room temperature. High proton conductivity and satisfactory mechanical properties were obtained as a result of phosphoric acid molecule chains charging the proton conduction and PIL crosslinking SPEEK matrix in the SPEEK/PIL/PA membrane.

2. Experimental

2.1. Materials

Poly(ether ether ketone) (PEEK) powder was purchased from Jida High Performance Materials Co., Ltd. China. Acryloyl chloride, 3-chloro-1-propanol and N-methylimidazole as the starting materials to synthesize IL monomer were purchased from J&K Scientific Ltd, China. Polyphosphoric acid, 2,2′-azobisisobutyronitrile (AIBN), ether, triethylamine, diethyl ether and dichloromethane were purchased from Sinopharm Chemical Reagent Co., Ltd, China. All chemicals were analytical reagent grade and were used as received except AIBN, which was purified via re-crystallization in methanol. Moreover, pure phosphoric acid was prepared by dissolving polyphosphoric acid (62.6 g) in de-ionized water (10 mL).

2.2. Synthesis of IL monomer

The synthesis of the IL monomer was done according to the literature.⁴² Specifically, this procedure includes the synthesis of the (acryloyloxy)propanol chlorine intermediate (eqn (1)) and the (acryloyloxy)propanylimidazolium chloride monomer (eqn (2)). First, a mixture of 3.3 mL of 3-chloro-1-propanol and 6 mL of triethylamine was dissolved in 50 mL of dichloromethane under magnetic stirring. 3.5 mL of acryloyl chloride was then added slowly under dry N₂ atmosphere in an ice-water bath. The mixture was stirred at 30 °C for an additional 48 h. Afterwards, dichloromethane was removed in a rotary vacuum evaporator at 50 °C and the precipitate was dissolved in diethyl ether. Finally, 5.7622 g of a white solid deposit of (acryloyloxy)propanol chlorine intermediate was obtained after removing the diethyl ether and drying under vacuum at room temperature. The synthesis of (acryloyloxy)propanylimidazolium chloride was inspired by the synthesis of imidazolium ionic liquids. A portion of 5.6024 g of (acryloyloxy)propanol chlorine intermediate was firstly dissolved into 7.0 mL of N-methylimidazole and the reaction lasted for 72 h at 45 °C under magnetic stirring with dry N₂ protection. The product was then washed with ether for 6–7 times. After that, a yellow viscous liquid of (acryloyloxy)propanylimidazolium chloride as the IL monomer was obtained.

2.3. Preparation of SPEEK/PIL and SPEEK/PIL/PA membranes

SPEEK polymer was prepared by PEEK sulfonation with sulfuric acid and its degree of sulfonation (DS) was determined to be 64.3% with titration method. 2.4996 g pf SPEEK polymer was dissolved into 150 mL of N,N-dimethylacetamide (DMAc) and then reacted with 1.6673 g of IL monomer, expecting to prepare SPEEK/IL in an equimolar amount. After heating at 60 °C for an additional 2 h with dry N₂ protection, the homogenous SPEEK/IL solution was refluxed at 60 °C for 24 h with 0.0511 g of AIBN to complete the carbon–carbon double bond polymerization and thus formed SPEEK/PIL. The SPEEK/PIL membrane was then obtained after spreading the solution on a leveled glass plate and evaporating solvent at 120 °C in a vacuum overnight. The produced HCl volatilized during the membrane forming process. Moreover, the SPEEK/PIL/PA membrane was prepared by immersing the SPEEK/PIL membrane into 30 mL of pure phosphoric acid in a sealed container for 48 h at room temperature. Similarly, SPEEK/IL and SPEEK/IL/PA membranes were also prepared for comparison. 2.1518 g of SPEEK polymer reacted with 1.4405 g of IL monomer to prepare a SPEEK/IL membrane with an equal molar ratio. The SPEEK/IL/PA membrane was prepared through immersing the SPEEK/IL membrane into 30 mL of pure phosphoric acid.

2.4. Fourier transform infrared spectra

The (acryloyloxy)propanol chlorine intermediate and (acryloyloxy)propanylimidazolium chloride monomer were characterized in the range of 4000–500 cm⁻¹ using a Vertex 70 Spectrometer (Bruker Optics Company, Germany). Carbon–carbon double bonds were expected to be observed in the intermediate and IL monomer.

2.5. NMR

The nuclear magnetic resonance spectrum of the polymerization of (acryloyloxy)propanylimidazolium chloride was obtained using a Bruker Avance 600 MHz NMR Spectrometer. Trichloromethane-D (CDCl₃) and tetramethyl silane (TMS) were used as the solvent and the internal standard, respectively.

2.6. Thermal and chemical stability

The thermal stabilities of IL monomer, PIL, and SPEEK/IL, SPEEK/PIL, SPEEK/IL/PA, and SPEEK/PIL/PA membranes were determined with a TGA 290C thermogravimetric analyzer (Netzsch Company, Germany). Approximately 5 mg of each sample was analyzed in the temperature range of 25–600 °C at a heating rate of 10 °C min⁻¹ with an air flow rate of 30 mL min⁻¹. In the case of the chemical stability test, the membrane samples of pristine SPEEK, SPEEK/IL and SPEEK/PIL without phosphoric acid were immersed into Fenton reagent (H₂O₂, 3 wt%; Fe²⁺, 4 ppm) at 68 °C. We recorded the weight and volume variations over a period of 36 h to evaluate their chemical stability and analyze the degradation behavior. Moreover, Fenton reagent was renewed after each measurement.

2.7. Fine surface structure

Surface and cross-sectional morphologies of the SPEEK/IL, SPEEK/PIL, SPEEK/IL/PA and SPEEK/PIL/PA membranes were investigated on an SSX-550 (Shimadzu Company, Japan) scanning electron microscope (SEM). Membrane samples were vacuum sputtered with a thin layer of gold nanoparticles prior to this examination.

2.8. X-ray diffraction

X-ray diffraction (XRD) test was performed to characterize the nanostructure variation of membranes with PIL using a Part Pro-MPD X-ray diffractometer (Panalytical B.V., The Netherlands) with Cu Kα generated at 40 kV and 100 mA. The test angle was at a range of 5–80° with a scanning speed of 5° min⁻¹.

2.9. Weight gain and dimension swelling of membranes loading with phosphoric acid

The values of weight gain and dimension swelling were obtained through measuring the weight and volume of the membranes before and after loading with phosphoric acid. The rectangular SPEEK/IL and SPEEK/PIL membrane samples with volumes of 0.08–0.11 cm³ and weights of 0.10–0.15 g were immersed into 20 mL of pure phosphoric acid in an airtight glass bottle at room temperature. We measured the data of the SPEEK/IL/PA membrane rapidly after wiping the surface-attached phosphoric acid with a tissue paper. The weight and volume values were measured to be 0.21–0.23 g, 0.15–0.17 cm³ and 0.32–0.41 g, 0.15–0.20 cm³ for the SPEEK/IL/PA membrane and the SPEEK/PIL/PA membrane, respectively. Then the weight gain and volume swelling values were calculated according to eqn (3) and (4). Each sample was tested with at least three specimens and the results are reported as the average.W₀ and W represent the weights of the original membrane and the phosphoric acid doped membrane, respectively; similarly, V₀ and V are the volumes for these mentioned membranes.

2.10. Proton conductivity

A pair of platinum probes supplying about 2 kHz frequency alternating current with a constant value were placed on the both sides of the membrane sample. The voltage drop caused by the current through the membrane part between the probes was recorded manually and the resistance (R) was thus calculated by Ampere's law. The through-plane proton conductivity (σ) was determined according to eqn (5):

In this equation, l represents the distance of current going through the membrane part; s means the cross-sectional area of the membrane, calculated from the product of width (w) and thickness (t). Phosphoric acid doped membranes were pre-heated at 80 °C until a stable value was reached. All proton conductivity measurements were performed in an oven without humidification.

2.11. Mechanical properties

The mechanical properties of these membranes including tensile stress (E) and elongation at break were tested using a CMT6502 tensile strength instrument (SANS Company, China). A closed oven was used for tests at 120 °C and 140 °C. All measurements were performed at a constant separating speed of 5 mm min⁻¹. Membrane samples were cut into initial dimensions of 25 mm in length and 4 mm in width with a mold. The initial cross-sectional area A₀ was 4L mm², where L represents the thickness of the membrane in mm. E values are calculated by dividing the force at break (F) by A₀ as in eqn (6).

3. Results and discussion

3.1. Synthesis of membranes

In the case of the SPEEK/PIL membrane, the IL monomer was firstly bound into the SPEEK matrix through electrostatic attraction between the imidazolium cations and the sulfonic groups in solution (Fig. 1(b)). The vinyl groups in the IL monomer were polymerized with AIBN as the initiator (Fig. 1(c)). Although the free ionic motion was reduced after polymerization, there was little effect on the proton conduction because the disubstituted imidazolium cations hardly conduct protons. PIL actually played a role in two aspects in the SPEEK/PIL/PA membrane: crosslinking the polymer of SPEEK and absorbing phosphoric acid molecules. Phosphoric acid molecules as proton carriers were thus introduced through hydrogen bonds in the SPEEK/PIL/PA membrane. Specifically, phosphoric acid molecules were adsorbed owing to the hydrogen bonds between the phosphoric acid molecules and the imidazolium groups. Subsequently, more free phosphoric acid molecule chains formed with more phosphoric acid molecules doped owing to intermolecular hydrogen bonds among phosphoric acid molecules (Fig. 1(d)).

Fig. 1 Presentation of the synthesis procedures of the SPEEK/PIL/PA membrane and photographs of the SPEEK/PIL and SPEEK/PIL/PA membranes.

Photographs of the SPEEK/PIL and SPEEK/PIL/PA membranes with transparency and homogeneity are shown in Fig. 1. According to the bent behavior, they probably have high tensile stress and enough flexibility. The thickness was about 80–90 μm for the SPEEK/PIL membrane, which increased to nearly 100–110 μm for the SPEEK/PIL/PA membrane as determined by microcallipers (5 measurements).

3.2. FTIR

As shown in Fig. 2, the FTIR spectra confirmed that the IL monomer was synthesized successfully. For IL intermediate and IL monomer, the peak at 1730 cm⁻¹ was caused by C=O double bond in carbonyl group.⁴³ The two peaks at 1634 cm⁻¹ and 1618 cm⁻¹ correspond to the stretching vibration of the C=C double bond.⁴⁴ In the case of the C=H double bond, the peak at 3030 cm⁻¹ was caused by the stretching vibration; the peaks at 1409 cm⁻¹ and 1361 cm⁻¹ corresponded to the in-plane bending vibration. Notably, the peaks at 1233 cm⁻¹ and 860 cm⁻¹ were caused by the stretching mode and the in-plane deformation vibration of the C–Cl bond in the IL intermediate.⁴⁵ However, these characteristic peaks were not observed in the IL monomer, resulting from the reaction of the C–Cl bond with the N-methylimidazolium cation. Moreover, the peaks at 1169 cm⁻¹, 3171 cm⁻¹ and 3125 cm⁻¹ were determined to be caused by the in-plane vibration and stretching vibration of the C–H group in the imidazolium groups.⁴⁶ These results depicting the basic characteristic groups of C=C, C=O, C=H double bonds and imidazolium ring confirmed the successful synthesis of the IL monomer.

Fig. 2 FTIR spectra of IL intermediate and IL monomer.

3.3. NMR

The ¹H NMR spectrum of the PIL cation is illustrated in Fig. 3. The bands at 8.74 ppm, 7.44 ppm and 7.46 ppm correspond to the H-1, H-2 and H-3 protons in the imidazolium ring. The band from the H-4 proton in the N-bound methyl group was observed at 3.97 ppm. Moreover, the bands from the protons H-5, H-6 and H-7 in methyl groups emerged at 4.51 ppm, 1.57 ppm, 4.30 ppm. The bands at 2.03 ppm and 1.59 ppm were attributed to H-8 in methylene and H-9 in methylidyne of the polymerization cation chains, respectively.⁴⁷ Moreover, the band at 3.20 ppm was from the residual ether and the strong bands at 7.64 ppm were caused by the solvent CDCl₃.

Fig. 3 ¹H NMR spectrum of the PIL cation in CDCl₃.

3.4. Thermal stability

Fig. 4 shows the TGA curves of the IL monomer, the PIL and the membrane samples. Although no mass loss was observed up to 190 °C for the SPEEK membrane, there was a two-stage decomposition of the sulfonic groups and the polymeric backbone.²¹ For the SPEEK/PIL membrane, the initial mass loss was observed at 310 °C. Although PIL began to decompose at 180 °C, the mass loss of the SPEEK/PIL membrane at 310 °C was probably caused by the decomposition of the imidazolium groups in the PIL as a result of the electrostatic attractions between the imidazolium cations and the sulfonic groups. In the case of phosphoric acid doped membranes, there were always additional mass loss stages, such as the obvious mass loss below 100 °C caused by the evaporation of water absorbed by phosphoric acid molecules. Moreover, the evaporation of water molecules released from phosphoric acid polymerization also caused the following mass stages in the phosphoric acid doped membranes.^41,48 The SPEEK/PIL/PA membrane possessed satisfactory thermal stability with no mass loss occurring until 240 °C. The high thermal decomposition temperature ensured its application in high temperature PEMFCs.¹⁸

Fig. 4 TGA curves of IL monomer, PIL and SPEEK/IL, SPEEK/PIL, SPEEK/IL/PA, and SPEEK/PIL/PA membranes.

3.5. Chemical stability

Fenton reagent can generate strong oxidizers such as HO˙, HOO˙, Fe³⁺ and O₂ via the decomposition of H₂O₂ with the catalysis of Fe²⁺. These radicals attack the heteroatom in polymeric chains, further breach the polymeric skeleton and, ultimately, cause the degradation of membrane samples.¹² Fenton reagent is thus frequently used to evaluate the chemical stability of membranes. In consideration of phosphoric acid leaking from membranes in Fenton reagent, the SPEEK/IL and SPEEK/PIL membranes rather than the phosphoric acid doped membranes were selected for the test.

The pristine SPEEK membrane degenerated within the first 3 h because these radicals repeatedly attacked sulfonic groups.²¹ In the SPEEK/PIL membrane, these radicals preferentially attacked any one of the three carbon atoms in the imidazolium ring.⁴⁹ In fact, PIL functioning as a protective layer outside the skeleton of the SPEEK polymer experienced persistent attacking from these radicals. Although the decomposition of PIL or IL was inevitable, the skeleton decomposition of the SPEEK polymer was delayed consequently.¹² As a result, the SPEEK/IL and SPEEK/PIL membranes kept their integrities for over 36 h. Notably, the SPEEK/PIL membrane was believed to maintain more residuals compared to the SPEEK/IL membrane after 36 h of testing. Specifically, the values were 64.2% in mass and 71.7% in volume for the SPEEK/PIL membrane. The corresponding values, by contrast, were 58.8% and 65.8% for the SPEEK/IL membrane, as shown in Fig. 5(A) and (B). Besides the decomposition of SPEEK and ILs, there was a potential risk of IL leaking from the polymer/IL membranes.⁵⁰ It is thus inferred that PIL has an advantage in residing in membranes relative to the IL monomer, further improving the chemical stability of membranes.

Fig. 5 Mass (A) and volume (B) variations of SPEEK/IL and SPEEK/PIL membranes immersed in Fenton reagent at 68 °C.

3.6. Fine structure of membranes

Some obvious variations in the microstructure after IL monomer polymerization were observed in the SEM images. For the SPEEK/IL and SPEEK/PIL membranes, they displayed different morphologies even with the same contents of IL and PIL. The SPEEK/PIL membrane had a coherent and uniform structure, albeit with some drawbacks in the surface and cross section structure (Fig. 6B and F). In contrast, some clusters were found in the microstructure of the SPEEK/IL membrane (Fig. 6A and E). The homogeneity of the SPEEK/PIL membrane is owing to the good compatibility between SPEEK and PIL. Unlike the discrete distribution of the IL monomer in the SPEEK/IL membrane, PIL was found to form one dimensional (1D) chains. The PIL and the SPEEK skeleton thus overlapped and interwove to form a whole, which may favor the formation of homogeneity in the microstructure. A similar phenomenon was also observed in these phosphoric acid doped membranes. As a result of phosphoric acid swelling the membranes, there was a tendency towards component separation, leading to some cracks being generated on the surface and some holes in the cross section of the SPEEK/IL/PA membrane (Fig. 6C and G). The SPEEK/PIL/PA membrane showed a uniform and compact microstructure owing to PIL crosslinking with the SPEEK matrix (Fig. 6D and H). It was inferred that the homogeneity in the SPEEK/PIL/PA membrane resulted in more efficient conducting pathways, besides an improvement on tensile stress.

Fig. 6 SEM images of membranes: (A and E) SPEEK/IL, (B and F) SPEEK/PIL, (C and G) SPEEK/IL/PA, (D and H) SPEEK/PIL/PA. (A–D) surface images, (E–H) cross-section images.

3.7. X-ray diffraction

The microstructures of these composite membranes were investigated by XRD analysis and the XRD patterns are shown in Fig. 7. As reported, although PEEK is highly crystalline, SPEEK presents an amorphous structure by showing a wide diffraction peak after sulfonation because of random incorporation of –SO₃H disordering in the hydrophobic polymeric skeleton.⁵¹ These composite membranes based on SPEEK thus showed low broad peaks located in the range of 2θ = 18–26°, typical of very low crystalline materials. In the case of the SPEEK/PIL membrane, a shoulder peak at 2θ = 22° shifted by about 2° to a higher angle relative to the SPEEK/IL membrane. According to Bragg's equation (eqn (7)), an increase in the θ value of the SPEEK/PIL membrane indicated a smaller d-spacing in the microstructure as a result of the fixed value of a product of n and λ.

2d sin θ = nλ (7)

where n is an integer, λ is the wavelength of incident wave, d is the space between the planes in the atomic lattice, and θ is the angle between the incident ray and the scattering planes. This result was indicative of a closed microstructure or efficient packing in the SPEEK/PIL membrane.⁵² Besides that, the addition of PIL could destroy the crystalline structure of SPEEK owing to the formed hydrogen bonds, which may contribute an increment to the proton transport in the amorphous phase and thus resulted in the improvement of proton conductivity.⁵³ Moreover, a similar phenomenon of 2θ shifting by about 4° to a higher angle was observed in the SPEEK/PIL/PA membrane. The SPEEK/PIL/PA membrane was thus expected to possess the compact microstructure, further favoring the reduced swelling volume.

Fig. 7 XRD patterns of the SPEEK, SPEEK/IL, SPEEK/PIL, SPEEK/IL/PA and SPEEK/PIL/PA membranes.

3.8. Weight gain and volume swelling

Indeed, we found that more phosphoric acid molecules were doped in the SPEEK/PIL/PA membrane compared to in the SPEEK/IL/PA membrane, leading to higher values of weight, as shown in Fig. 8. The average value of phosphoric acid uptake was 126% for the SPEEK/PIL/PA membrane, higher than the 109% for the SPEEK/IL/PA membrane at static equilibrium. Correspondingly, the average volume swelling values were 67.6% and 42.0%. The data indicated that the SPEEK/PIL membrane absorbed more phosphoric acid molecules even though its structure became compact. In the SPEEK/PIL/PA membrane, PIL is arranged in order along the SPEEK skeleton, which favored the absorption of more phosphoric acid molecules owing to an ordered arrangement. Although these phosphoric acid molecules occupied some space in these membranes, it was inferred that the ordered arrangement of phosphoric acid molecules along with PIL could form a compact accumulation in the SPEEK/PIL/PA membrane. The density of the SPEEK/PIL/PA membrane was calculated to be 2.14 g cm⁻³, which is higher than the 1.50 g cm⁻³ for the SPEEK/IL/PA membrane according to the weight and volume parameters. Moreover, the densities of these phosphoric acid doped membranes were higher than the values of 1.21 g cm⁻³ and 1.30 g cm⁻³ for the SPEEK/IL and SPEEK/PIL membranes while phosphoric acid molecules could occupy the space in the membranes. A higher content of phosphoric acid molecules in the membranes ensures an outstanding ability to conduct protons for these phosphoric acid doped membranes.

Fig. 8 Weight and dimension swelling of the SPEEK/IL and SPEEK/PIL membranes immersed in pure PA.

3.9. Proton conductivity

Fig. 9 shows that the conductivities of the SPEEK/PIL, SPEEK/IL/PA and SPEEK/PIL/PA membranes increased with rising temperature. As is well known, the viscosity of a membrane is reduced and the vibration of the polymeric chains is accelerated when the temperature rises. The conductivities of membranes are thus improved obviously, benefiting from these variations in the property of membranes. Notably, a conductivity value of 1.10 × 10⁻² S cm⁻¹ at 160 °C was observed for the SPEEK/PIL membrane, which was obviously higher than our reported value of 7.90 × 10⁻³ S cm⁻¹ at 160 °C for the SPEEK/50%BMIMPF₆ membrane.¹² In the SPEEK/PIL membrane, PIL absorbed phosphoric acid molecules through the intermolecular hydrogen bonds, endowing the SPEEK polymer with the ability to conduct protons. As reported, phosphoric acid molecules with excellent thermal stability and low vapor pressure at elevated temperature could function as both proton donors and proton acceptors, and even form dynamic hydrogen bond networks in an anhydrous condition.⁵⁴ Phosphoric acid molecules as main carriers have always been used in the exploration of high temperature PEMs. The proton conductivity of 100 wt% phosphoric acid was 8.0 × 10⁻¹ S cm⁻¹ at 160 °C,⁴⁸ which was obviously higher than the 4.5 × 10⁻² S cm⁻¹ for the SPEEK/PIL/PA membrane as well as the 2.3 × 10⁻² S cm⁻¹ for the SPEEK/IL/PA membrane. In the SPEEK/PIL/PA membrane, phosphoric acid as a proton carrier was deemed to dominate the proton conductivity. Like polyvinyl chloride (PVC) matrix in a phosphoric acid doped methylimidazolium (MIM) modifying PVC membrane,⁴¹ the SPEEK matrix actually acted as a frame for immobilizing phosphoric acid molecules in the membrane system. So it is more difficult for phosphoric acid molecules to rotate and move in membranes compared to in 100 wt% phosphoric acid. However, the proton conductivity of the SPEEK/PIL/PA membrane was lower than the reported 2.17 × 10⁻¹ S cm⁻¹ at 200 °C for the OPBI/PA membrane with a phosphoric acid doping level of 24.6, which was prepared through the flexible ether linkage (–O–) and side group (phenyl) grafting onto a polybenzimidazole (PBI) backbone by Guiver.⁵⁵

Fig. 9 Conductivities of the SPEEK/PIL, SPEEK/IL/PA and SPEEK/PIL/PA membranes as a function of temperature.

Although a great motive of using PIL was to improve the tensile stress of the SPEEK/PIL/PA membrane, we indeed found an obvious improvement in proton conductivity relative to the SPEEK/IL/PA membrane. What was the underlying reason for this discrepancy in the ability to conduct protons? Of key importance was the distinction of phosphoric acid content in the membranes. As discussed in Section 3.8, the values of phosphoric acid uptake were 126% and 109% at static equilibrium for the SPEEK/PIL/PA and SPEEK/IL/PA membranes, respectively. Moreover, the PIL stability in the membrane presumably favored maintaining higher proton conductivity as well. As a matter of fact, IL are frequently reported to have a tendency towards leaking from membranes due to the lack of strong interaction.²¹ There was thus a risk of IL leakage while the SPEEK/IL membrane was immersed in the phosphoric acid solution or even for the pure phosphoric acid system. More PIL, by contrast, was retained in the SPEEK/PIL/PA membrane owing to a reduction of its flowability.²¹ The massive amount of PIL ensured the absorption of an adequate amount of phosphoric acid molecules, forming an interconnected and fast proton conduction network in the SPEEK/PIL/PA membrane. Therefore, the proton conductivity of the SPEEK/PIL/PA membrane was improved obviously as expected.

As a candidate for PEMs, an understanding of proton transport in these materials is academically and practically essential. We then calculated the activation energy (E_a) values of the SPEEK/IL/PA and SPEEK/PIL/PA membranes according to the Arrhenius equation, as shown in eqn (8) and (9), expecting to get more information to understand the proton conduction mechanism.

Plotting the proton conductivity (σ) on a logarithmic scale versus the inverse of temperature in Kelvin gives an estimate of the E_a value from the slope of the linear curve (k) multiplied by constant R (8.314 J mol⁻¹ K⁻¹) in eqn (9). Specifically, the calculated E_a values of 17.3 kJ mol⁻¹ and 14.8 kJ mol⁻¹ corresponded to the SPEEK/IL/PA and SPEEK/PIL/PA membranes, respectively. It was inferred that there was presumably no discrepancy in the proton conduction mechanism between the SPEEK/IL/PA and SPEEK/PIL/PA membranes owing to their similar E_a values. Moreover, we reported that the E_a value was 15.5 kJ mol⁻¹ for a SPEEK/50%BMIMPF₆/4.6PA membrane with free phosphoric acid molecules chains dominating the proton conduction process.²¹ In phosphoric acid doped imidazolium IL membranes, phosphoric acid molecules were firstly bounded by occupying the active sites in imidazolium cations and then free phosphoric acid molecule chains emerged with more phosphoric acid molecules doping via intramolecular phosphoric acid hydrogen bonds. Importantly, the massive amount of free phosphoric acid molecule chains further formed the proton conduction network, which dominated the proton conduction in the SPEEK/IL/PA and SPEEK/PIL/PA membranes.

3.10. Proton conductivity stability

As one of many possible demonstrations of the practical relevance of new membrane materials, the proton conductivity stabilities of the SPEEK/IL/PA and SPEEK/PIL/PA membranes were investigated. We expected to obtain evidence to approve the function of the PIL crosslinked SPEEK matrix according to their discrepant performance. As shown in Fig. 10, the proton conductivity of the SPEEK/PIL/PA membrane gradually decreased to 2.7 × 10⁻² S cm⁻¹ within the first 60 h at 140 °C under anhydrous conditions. In consideration of our previous reports, the local dimensional change of the membrane sample in contact with Pt wire electrodes was responsible for the drop in proton conductivity.¹⁶ It was worth noting that a stable proton conductivity of 2.2 × 10⁻² S cm⁻¹ lasting 400 h was achieved, obviously higher than the 1.3 × 10⁻² S cm⁻¹ of the SPEEK/IL/PA membrane as the experiment was terminated voluntarily. Xu reported that a phosphoric acid doped tetrazole-based sulfonated poly(phthalazinone ether sulfone ketone) (PA-AtSPPESK) membrane possessed a proton conductivity of 1.5 × 10⁻² S cm⁻¹ at 150 °C. The proton conductivity was 1.1 × 10⁻² S cm⁻¹ at 130 °C while the test lasted 10 h.⁵⁶ Therefore, the outstanding proton conductivity stability performance provided the SPEEK/PIL/PA membrane with the potential for application as high temperature PEMs.

Fig. 10 Anhydrous proton conductivities of the SPEEK/IL/PA and SPEEK/PIL/PA membranes as a function of time at 140 °C.

3.11. Mechanical property

We then systematically investigated and compared the mechanical properties of these prepared membranes, namely tensile strength (E) and elongation (ε) at room temperature (RT) and high temperature (HT) under anhydrous conditions. As shown in Table 1, the SPEEK/IL and SPEEK/PIL membranes showed satisfactory E and suitable ε values, demonstrating their overall high mechanical performance. Notably, the E values of the SPEEK/IL and SPEEK/PIL membranes were close and even superior to the SPEEK membrane at RT. Specifically, the E values were 38.6 MPa and 52.7 MPa for the SPEEK/IL and SPEEK/PIL membranes, compared to 41.2 MPa for the SPEEK membrane. The data showed that there was a negligible effect on the E value with the introduction of the IL monomer in the SPEEK/IL membrane. Furthermore, it was determined that the PIL crosslinked SPEEK matrix led to an improvement in the E value of the SPEEK/PIL membrane. But in the case of their performance at HT, the E values of the SPEEK/IL and SPEEK/PIL membranes were lower than that of SPEEK membrane at 140 °C. We believe that the accelerating vibration of polymeric molecular chains from adsorbing more energy during the heating process played a critical role in the decrease of E, while the structural parameters were secondary. So the E values for all measured membranes decreased, but the SPEEK membrane possessed the maximum E value owing to its good integrity.

Table 1 Mechanical properties of the SPEEK/IL, SPEEK/PIL and PA doped membranes at room temperature and high temperature under anhydrous conditions

Membrane	Tensile stress (MPa)		Elongation at break (%)
	RT	HT	RT	HT
SPEEK	41.2 (ref. 16)	24.8 (140 °C)^16	15.8	47.1
SPEEK/IL	38.6	5.11 (140 °C)	12.0	213
SPEEK/PIL	52.7	15.5 (140 °C)	4.86	127
SPEEK/IL/PA	17.3	1.83 (120 °C)	90.5	105
SPEEK/PIL/PA	18.4	2.64 (120 °C)	53.6	88.7

---

The proton conductivity–tensile stress dilemma was observed in the SPEEK/IL/PA and SPEEK/PIL/PA membranes. A decrement in the E value at break was accompanied by an increase in proton conductivity with phosphoric acid doping or temperature rising.⁴¹ Specifically, the E values were 52.7 MPa at RT and 15.5 MPa at 140 °C for the SPEEK/PIL membrane, dropping down to 18.4 MPa at RT and 2.64 MPa at 120 °C for the SPEEK/PIL/PA membrane. Zheng reported that the maximum tensile strength of a poly(ionic liquid)/polyvinyl alcohol/caprylic acid (PIL/PVA/CA) membrane was 8.0 MPa at RT, which was higher than 6.9 MPa for a poly(ionic liquid)/polyvinyl alcohol membrane at room temperature.⁵⁷ A blend membrane based on PBI and poly(vinylidene fluoride-co-hexafluoro propylene) (PVDF-HEP) doping phosphoric acid was reported to have a maximum proton conductivity of 1.7 × 10⁻¹ S cm⁻¹ at 160 °C and the tensile strength was 5.25 MPa at RT.⁵⁸ In phosphoric acid doped membranes, phosphoric acid molecules would result in big volume swelling and would therefore separate the polymer skeleton. The E values were consequently decreased owing to a reduction of intermolecular forces.⁵⁹ Although the SPEEK/IL/PA and SPEEK/PIL/PA membranes have similar E values at RT, an obvious discrepancy in E values was observed in the test performed at 140 °C. For a phosphoric acid doped Nafion® membrane, the E value at break was reported to be 2.3 ± 0.2 MPa at 130 °C,⁶⁰ which was similar to the 2.64 MPa at 120 °C for the SPEEK/PIL/PA membrane. Our group has also reported E values of 1.88 MPa for a PVC-MIMCl/PA (1/5) membrane at 130 °C (ref. 41) and 0.10 MPa for a SPEEK/50%BMIMPF₆/4.6PA membrane at 130 °C.¹² Therefore, in consideration of its high performance in terms of mechanical property and proton conductivity, the SPEEK/PIL/PA membrane is probably a suitable high temperature PEM candidate.

4. Conclusions

In this research, an improvement in the mechanical properties of phosphoric acid doped membranes is achieved owing to the PIL crosslinked SPEEK matrix. The SPEEK matrix captured the imidazolium IL monomer by the interaction between the imidazolium cations and the sulfonic groups. The SPEEK/PIL membrane was formed when the PIL crosslinked with the SPEEK matrix owing to the polymerization of the unsaturated double bonds in the IL monomer. Phosphoric acid molecules were absorbed through the intermolecular hydrogen bonds with the imidazolium cations of the PIL. The high thermal stability, satisfactory and stable proton conductivity, and favorable mechanical properties guaranteed SPEEK/PIL/PA membrane as an electrolyte for application in high temperature PEMFCs. Our future work will be focused on the synthesis of new PILs to crosslink sulfonated polymers, expecting to prepare polymer/PIL/PA membranes with a balance between proton conductivity and tensile stress. Furthermore, as a membrane electrolyte for use in PEMFCs, the SPEEK/PIL/PA membrane also holds promise for advanced membrane materials for other energy storage devices that require proton conductivity.

Acknowledgements

We are grateful for the financial support from the Scientific Research Fund of Liaoning Provincial Education Department (L20150166) and the China Scholarship Council (201208210023).

References

1. P. Costamagna and S. Srinivasan, J. Power Sources, 2001, 102, 242–252.
2. N. Zhang, B. L. Wang, C. J. Zhao, S. Wang, Y. R. Zhang, F. Z. Bu, Y. Cui, X. F. Li and H. Na, J. Mater. Chem. A, 2014, 2, 13996–14003.
3. Q. T. Che, N. Chen, J. M. Yu and S. C. Cheng, Solid State Ionics, 2016, 289, 199–206.
4. B. Singh, M. S. Hundal, G. G. Park, J. S. Park, W. Y. Lee, C. S. Kim, K. Yamada and S. S. Sekhon, Solid State Ionics, 2007, 178, 1404–1410.
5. Q. T. Che, B. Y. Sun and R. H. He, Electrochim. Acta, 2008, 53, 4428–4434.
6. Q. F. Li, R. H. He, J. O. Jensen and N. J. Bjerrum, Chem. Mater., 2003, 15, 4896–4915.
7. C. X. Xu, X. T. Liu, J. G. Cheng and K. Scott, J. Power Sources, 2015, 274, 922–927.
8. Z. G. Shao, H. F. Xu, M. Q. Li and I. M. Hsing, Solid State Ionics, 2006, 177, 779–785.
9. M. A. Dresch, R. A. Isidoro, M. Linardi, J. F. Q. Rey, F. C. Fonseca and E. I. Santiago, Electrochim. Acta, 2013, 94, 353–359.
10. Y. Devrim and A. Albostan, Int. J. Hydrogen Energy, 2015, 40, 15328–15335.
11. Q. F. Li, R. H. He, R. W. Berg, H. A. Hjuler and N. J. Bjerrum, Solid State Ionics, 2004, 168, 177–185.
12. Q. T. Che, L. Zhou and J. L. Wang, J. Mol. Liq., 2015, 206, 10–18.
13. S. P. Jiang, J. Mater. Chem. A, 2014, 2, 7637–7655.
14. V. S. Rangasamy, S. Thayumanasundaram, N. de Greef, J. W. Seo and J. P. Locquet, Eur. J. Inorg. Chem., 2015, 7, 1282–1289.
15. J. Fuller, A. C. Breda and R. T. Carlin, J. Electrochem. Soc., 1997, 144, L67–L70.
16. Q. T. Che, R. H. He, J. S. Yang, L. Feng and R. F. Savinell, Electrochem. Commun., 2010, 12, 647–649.
17. X. J. Bao, F. Zhang and Q. T. Liu, Int. J. Hydrogen Energy, 2015, 40, 16767–16774.
18. M. Díaz, A. Ortiz, M. Vilas, E. Tojo and I. Ortiz, Int. J. Hydrogen Energy, 2014, 39, 3970–3977.
19. J. Fuller, R. T. Carlin and R. A. Osteryoung, J. Electrochem. Soc., 1997, 144, 3881–3886.
20. M. A. B. H. Susan, T. Kaneko, A. Noda and M. Watanabe, J. Am. Chem. Soc., 2005, 127, 4976–4983.
21. M. Armand, F. Endres, D. R. MacFarlane, H. Ohno and B. Scrosati, Nat. Mater., 2009, 8, 621–629.
22. J. L. Wang, L. L. Wang, R. J. Feng and Y. Zhang, Solid State Ionics, 2015, 278, 144–151.
23. Y. Yang, H. J. Gao and L. Q. Zheng, RSC Adv., 2015, 5, 17683–17689.
24. F. Yan, S. M. Yu, X. W. Zhang, L. H. Qiu, F. Q. Chu, J. B. You and J. M. Lu, Chem. Mater., 2009, 21, 1480–1484.
25. S. Livi, J. Duchet-Rumeau, J. F. Gérard and T. N. Pham, Macromol. Chem. Phys., 2015, 216, 359–368.
26. J. Y. Yuan and M. Antonietti, Polymer, 2011, 52, 1469–1482.
27. B. Huber, L. Rossrucker, J. Sundermeyer and B. Roling, Solid State Ionics, 2013, 15–21, 247–248.
28. A. S. Shaplov, R. Marcilla and D. Mecerreyes, Electrochim. Acta, 2015, 17, 18–34.
29. Y. Yang, N. Sun, P. P. Sun and L. Q. Zheng, RSC Adv., 2016, 6, 25311–25318.
30. H. Ohno and K. Ito, Chem. Lett., 1998, 8, 751–752.
31. M. Hirao, K. Ito and H. Ohno, Electrochim. Acta, 2000, 45, 1291–1294.
32. L. C. Tomé, A. S. L. Gouveia, C. S. R. Freire, D. Mecerreyes and I. M. Marrucho, J. Membr. Sci., 2015, 486, 40–48.
33. H. Ohno, M. Yoshizawa and W. Ogihara, Electrochim. Acta, 2004, 50, 255–261.
34. Y. Kohno, S. Saita, Y. J. Men, J. Y. Yuan and H. Ohno, Polym. Chem., 2015, 6, 2163–2178.
35. A. Narita, W. Shibayama, N. Matsumi and H. Ohno, Polym. Bull., 2006, 57, 109–114.
36. H. L. Ricks-Laskoski and A. W. Snow, J. Am. Chem. Soc., 2006, 128, 12402–12403.
37. M. Yoshizawa, M. Hirao, K. Ito-Akita and H. Ohno, J. Mater. Chem., 2001, 11, 1057–1062.
38. J. Zhao, F. Yan, Z. Z. Chen, H. B. Diao, F. Q. Chu, S. M. Yu and J. M. Lu, J. Polym. Sci., Part A: Polym. Chem., 2009, 47, 746–753.
39. J. Texter, V. A. Vasantha, R. Crombez, R. Maniglia, L. Slater and T. Mourey, Macromol. Rapid Commun., 2012, 33, 69–74.
40. S. Washiro, M. Yoshizawa, H. Nakajima and H. Ohno, Polymer, 2004, 45, 1577–1582.
41. Q. T. Che, Z. F. Zhu, N. Chen and X. Zhai, Mater. Des., 2015, 87, 1047–1055.
42. M. Yoshizawa and H. Ohno, Electrochim. Acta, 2001, 46, 1723–1728.
43. B. Tsuchiya, Y. Konishi, S. Nagata and T. Shikama, Solid State Ionics, 2008, 179, 1128–1132.
44. L. Wu, Q. Pan, J. R. Varcoe, D. Zhou, J. Ran, Z. J. Yang and T. W. Xu, J. Membr. Sci., 2015, 490, 1–8.
45. S. Ramesh, K. H. Leen, K. Kumutha and A. K. Arof, Spectrochim. Acta, Part A, 2007, 66, 1237–1242.
46. J. Stejskal, J. Dybal and M. Trchová, Synth. Met., 2014, 197, 168–174.
47. H. Erdemi, Ü. Akbey and W. H. Meyer, Solid State Ionics, 2010, 181, 1586–1595.
48. Y. Ma, J. Wainright, M. Litt and R. F. Savinell, J. Electrochem. Soc., 2004, 151, A8–A16.
49. E. M. Siedlecka, W. Mrozik, Z. Kaczyński and P. Stepnowski, J. Hazard. Mater., 2008, 154, 893–900.
50. S. S. Sekhon, J. S. Park, E. K. Cho, Y. G. Yoon, C. S. Kim and W. Y. Lee, Macromolecules, 2009, 42, 2054–2062.
51. M. Prasad, S. Mohanty and S. K. Nayak, RSC Adv., 2014, 4, 61178–61186.
52. T. H. Weng, H. H. Tseng and M. Y. Wey, Int. J. Hydrogen Energy, 2008, 33, 4178–4182.
53. H. T. Pu, M. D. Luo and Z. L. Yang, Eur. Polym. J., 2007, 43, 5076–5083.
54. J. C. Chen, J. A. Wu and K. H. Chen, RSC Adv., 2016, 6, 33959–33970.
55. X. B. Li, H. W. Ma, H. L. Wang, S. T. Zhang, Z. H. Jiang, B. J. Liu and M. D. Guiver, RSC Adv., 2015, 5, 53870–53873.
56. J. F. Pan, B. Wu, L. Wu, Y. B. He, J. B. Miao, L. Ge and T. W. Xu, Int. J. Hydrogen Energy, 2016, 41, 12337–12346.
57. Y. Yang, N. Sun and L. Q. Zheng, RSC Adv., 2016, 6, 58890–58897.
58. M. Hazarika and T. Jana, Eur. Polym. J., 2013, 49, 1564–1576.
59. R. H. He, Q. F. Li, A. Bach, J. O. Jensen and N. J. Bjerrum, J. Membr. Sci., 2006, 277, 38–45.
60. D. Aili, M. K. Hansen, C. Pan, Q. F. Li, E. Christensen, J. O. Jensen and N. J. Bjerrum, Int. J. Hydrogen Energy, 2011, 36, 6985–6993.

---