Self-cross-linked quaternary phosphonium based anion exchange membranes: assessing the influence of quaternary phosphonium groups on alkaline stability

Abstract

A series of novel quaternary phosphonium functionalized poly(aryl ether sulfone)s were successfully synthesized using a new methyl containing monomer via a multi-step process including bromination, quaternary phosphination and alkalization. The study of the chemical stability of these alkaline anion exchange membranes (AAEMs) using different spectroscopic techniques (¹H NMR, ³¹P NMR and ATR-FTIR) revealed the excellent alkaline stability of the polymer backbone and the partial degradation of phosphonium groups after alkaline exposure at 60 °C for 288 hours. Self-cross-linked membranes with different cross-linking degrees were also synthesized via Friedel–Crafts electrophilic substitution C-alkylation. Despite the fact that the IEC values of the cross-linked membranes are very low, the nano-phase separated morphologies observed by TEM enable the formation of ionic clusters, thus facilitating hydroxide conductivity. Indeed, the normalized IEC hydroxide conductivity for self-cross-linked AAEMs is 13.1 mS g cm⁻¹ meq.⁻¹ at 70 °C, which is comparable to that of other cross-linked AAEMs. The results of ATR-FTIR spectroscopy, TGA, and IEC demonstrated the excellent alkaline stability of the self-cross-linked AAEMs even after 40 days of immersion in 1 M KOH at 60 °C or even under harsher conditions such as 4 M KOH at 80 °C for 16 days, suggesting that the self-cross-linking strategy was beneficial for the protection of phosphonium groups from hydroxide attack.

---

Introduction

There is an intensified interest worldwide in the use of Alkaline Anion Exchange Membranes (AAEMs) in many electrochemical systems, such as alkaline polymer electrolyte fuel cells (APEFCs), flow and metal/air batteries, alkaline polymer electrolyte electrolysers, reverse electrodialysis cells and bioelectrochemical systems.^1–7 Particularly, the main advantage of APEFCs over PEM technology is the possibility of using less expensive, non-precious electrocatalysts due to the enhanced oxygen reduction/evolution^8–11 kinetics under alkaline pH environment. However, the primary concerns with the use of AAEMs in APEFCs are the low hydroxide conductivities^12,13 and low chemical stability in OH⁻ form,^14–16 since both determine to a great extent the performance and lifetime of APEFCs. Many studies directed towards improving the hydroxide conductivity of AAEMs, a property that is inherently limited by the lower intrinsic mobility of the hydroxide anion. Indeed, up to date the hydroxide conductivity of AAEMs has been greatly improved (more than 0.1 S cm⁻¹ at 80 °C).^17–20

However, the alkaline stability still remains the most important and challenging issue to be solved. AAEMs traditionally suffer from poor chemical stability in alkaline environments, as their degradation is induced by the presence of hydroxide ion, a strong nucleophilic ion.

Though quaternary ammonium (QA) functionalized poly(arylene ethers) have been the most extensively studied AAEMs, QA alkaline stability is severely deteriorated under high pH conditions, as evidenced by several studies.^1–5,21,22 The main degradation mechanisms for ammonium groups are via direct nucleophilic substitution^1,3–5 and/or Hofmann elimination.²³ This problem triggered the study of alternative cationic head group chemistries including guanidinium,²⁴ imidazolium,^21,25 diazacyclooctane-based (DABCO),²⁶ pyridinium,²⁷ phosphonium.²⁸

Although the lifetime studies indicate that the cation stability is of paramount importance for stable operation at AAEMs, recent studies proved that equally important is the stability of the polyaromatic backbone under alkaline media.^22,29,30 For example, polysulfone, one of the many available aromatic polymeric backbones, has been a very attractive candidate for preparing AAEMs, due to its high oxidative stability and thermal stability, ease of preparation and low cost. However, although it is stable when exposed to alkaline media, its functionalization with relative stable cationic groups tethered via CH₂ linkage on the diphenyl propane group in PSF resulted in backbone degradation upon exposure to alkaline solutions by both quaternary carbon and ether hydrolysis.²⁹ This is happening because the hydrophilicity of the polymer backbone increases after functionalization, allowing close approach of the OH⁻ anions. Therefore a stable polymeric backbone is also required to ensure AAEMs stability improvement.

To enhance the backbone stability, a new strategy is suggested by our group in order to move the cation away from the polysulfone backbone where an additional benzyl group is located between the cationic group and the polysulfone backbone.

Recently, this successful strategy was also reported by other groups.^31–34 Regarding the cation group chemistry, tris(2,4,6-trimethoxyphenyl)phosphonium groups were selected due to their high ionic conductivity, solubility and alkaline stability.^28a–c

This work focuses on the development and the systematic property evaluation of new AAEMs membranes based on polysulfone backbone where tris(2,4,6-trimethoxylphenyl)phosphonium groups are attached. The initial PSF copolymer was synthesized via nucleophilic aromatic substitution reaction while functionalization of PSF with bromomethyl groups took place via radical bromination of benzylic groups followed by Menshutkin reaction between tris(2,4,6-trimethoxylphenyl)phosphine and benzyl bromide and the subsequent alkalization provided the AAEM. The chemical stability of AAEMs after exposure to 1 M KOH at room temperature and 60 °C for different time periods was assessed using different spectroscopic techniques (ATR-FT-IR, ¹H NMR, ³¹P NMR) in order to conclusively identify the degradation mechanism since the precise identification of the degraded products is vital to gain mechanistic insights that could aid in the design more robust materials.

Surprisingly, these AAEMs show moderate alkaline stability at 60 °C and poor mechanical stability, thus covalent cross-linking has been considered as a promising technique to improve the mechanical properties and the alkaline stability. Thiol–ene click chemistry is used as a strategy to covalently cross-link AAEMs.³⁵ Another strategy includes in situ cross-linked AAEMs that have been developed via Friedel–Crafts electrophilic substitutions^{28a,34a,36–38} or olefin metathesis techniques,^17,39,40 allowing cross-linking during solution casting to afford AAEMs with homogeneous morphologies.

Inspired by the in situ cross-linking strategy, which is a simple route without the use of any separate cross-linker or catalyst, the Friedel–Crafts electrophilic substitution C alkylation was used to prepare the cross-linked AAEMs. In specific, the self cross-linked membranes were prepared in a single step where quaternization takes place first and subsequent Friedel–Crafts C alkylation reaction of bromomethylated groups containing PSF with the aromatic rings of the quaternary phosphonium groups. The properties of the cross-linked membranes, such as alkaline stability, water uptake, thermal stability, dimensional stability, ion exchange capacity, hydroxide conductivity were studied in details.

Experimental

Materials

Bisphenol A (99+%), bis(4-fluorophenyl)sulfone (99%), tris(2,4,6-trimethoxyphenyl)phosphine (99%), were purchased from Sigma-Aldrich and used as received without further purification. p-Tolyl boronic acid,⁴¹ 2,5-dibromohydroquinone,⁴² 2,5-dibromo-1,4-bis(tetrahydro-2H-pyran-2-yloxy)benzene⁴³ and palladium(II) tetrakis triphenyl phosphine [Pd(PPh₃)₄]⁴⁴ were prepared according to literature procedures.

Instrumentation

¹H NMR spectra were obtained on a Bruker Advance DPX 400 MHz spectrometer. The samples were dissolved either in deuterated chloroform (CDCl₃) or dimethylsulfoxide (DMSO-d₆) with tetramethylsilane (TMS) used as internal standard. Samples for ³¹P NMR were dissolved in dimethylsulfoxide (DMSO-d₆). Thermogravimetric analysis (TGA) was performed using a Labsys TG (Setaram Instrumentation). The samples were heated at 20 °C min⁻¹ to 800 °C under nitrogen atmosphere. ATR-IR spectra were recorded on Platinum ATR Bruker spectrometer. The surface and the cross-sectional morphology of the membranes were studied by Scanning Electron Microscopy (SEM) using a LEO Supra 35VP microscope. The specimens for the cross-section study were prepared by fracturing the membranes in liquid nitrogen. An elemental analysis was performed via Energy Dispersive X-ray Microanalysis (EDX, Bruker Quantax 200) to evaluate the distribution of the surface concentration across the membrane. Transmission Electron Microscopy (TEM) images were recorded using a JEM-2100 electron microscope (JEOL, Japan) at a working voltage 120 kV. The specimen was prepared by casting an AAEM thin film onto a Cu grid then exchanging the anions for I⁻ in KI solutions. Gel permeation chromatography (GPC) measurements were performed on a Polymer Lab chromatograph equipped with two PL gel 5 μm mixed columns and a UV detector (254 nm), using CHCl₃ as eluent with a flow rate of 1 ml min⁻¹ at 25 °C and polystyrene standards.

Ion exchange capacity (IEC)

The IECs of the AAEMs were measured via the back titration method. Specifically, the membrane sample was equilibrated with 3 ml 0.1 M HCl standard solution and 27 ml deionized water for 2 days, followed by back titration with 0.05 M NaOH standard solution with phenolphthalein used as the indicator. A solution of 3 ml 0.1 M HCl standard solution and 27 ml deionized water was used as a blank sample for the control experiment. The IEC was calculated by the following equation:
where V_b and V_s are the consumed volumes (L) of the NaOH solution for the blank sample and the membrane sample respectively. W_dry is the mass, in grams of the membrane after drying under vacuum at 60 °C overnight and C_NaOH is the concentration of NaOH standard solution used.

Water uptake (WU) and swelling ratio (SR)

To measure water uptake and swelling ratio, the membranes where immersed in deionized water for 1 day at 60 °C. After that, the excess water was removed by wiping with tissue paper and the membranes were weighed and measured. The wet membranes were then dried under vacuum at 60 °C overnight and the dry membranes were weighed and measured. The water uptake (%) was calculated using the following equation:
where W_wet and W_dry are the mass of wet and dry membrane sample, respectively.

The swelling ratio (%) was calculated with the following equation

where L_wet and L_dry are the length of wet and dry membrane sample, respectively.

Conductivity measurements

Ionic conductivity measurements were conducted using a two-point probe AC impedance spectroscopy technique on Autolab PGSTAT302N potentiostat/galvanostat. The impedance spectrum was carried out in the frequency range from 0.1 Hz to 100 kHz with a constant voltage of 10.0 mV at the temperature range 30–80 °C. All samples were fully hydrated in N₂-saturated deionized water for at least 24 h prior to their conductivity measurements. Membrane samples of 1 × 2 cm² were sandwiched between two Pt electrodes and the measurements were performed at 25, 40, 60, 70 °C in a vessel filled with degassed deionized water under N₂ flow to maintain the relative humidity at 100% and to remove CO₂ as well.

The ionic conductivity of the membrane samples can be calculated by the following equation

σ is the conductivity (S cm⁻¹), L is the thickness of the membrane, R is the membrane resistance (Ω), A is the area of the electrodes (cm²).

Alkaline stability

The alkaline stability of the prepared QPOH(x)PAES membranes was evaluated after immersion in 1 M aqueous KOH solution by using ¹H NMR, ³¹P NMR and ATR-FT-IR spectroscopy. Pieces of the membranes were stored in a sealed vessel at 25 and 60 °C respectively. After 6 days and 12 days of immersion, samples were taken out, thoroughly washed with deionized water to remove the absorbed KOH solution and dried under vacuum at 40 °C for 24 h prior to measurements. The QPOH(x)PAES membranes were dissolved in DMSO-d₆ for the ¹H NMR measurements.

The alkaline stability of the self-crosslinked CL-QPOH(x)PAES could not be studied via ¹H NMR and ³¹P NMR spectroscopy due to their insolubility in DMSO-d₆ solvent. The alkaline stability was investigated by monitoring the changes of the ion exchange capacity, ATR-FT-IR spectra and TGA curves before and after treatment in aqueous 1 M and 4 M KOH solutions at different temperatures (60 °C and 80 °C) for 40 and 16 days, respectively.

Synthesis of 2,5-bis(4-methylphenyl)benzene-1,4-diol monomer

In a degassed, filled with argon 500 ml flask, a mixture of p-tolyl boronic acid (6.26 g, 46.0 mmol), 2,5-dibromo-1,4-bis(tetrahydro-2H-pyran-2-yloxy)benzene (6.69 g, 15.3 mmol), Pd(PPh₃)₄ (0.5318 g, 0.46 mmol), toluene (92 ml) and aqueous 2 M Na₂CO₃ solution (46 ml) were added. The reaction mixture was left at reflux for 72 hours under argon atmosphere. The organic layer was extracted with toluene and washed twice with water and once with a saturated NaCl solution. A grey-orange solid was obtained after evaporation of the solvent. The solid (THP-protected diol) was treated with ethanol under reflux, filtered and subsequently dissolved in THF. The THF solution passed through celite in order to remove palladium residues. In the next step, deprotection of the THP-protected diol took place, through addition of 2 ml of 37% HCl solution to this solution. The solution was first stirred for 4 hours at room temperature, then was concentrated under reduced pressure to a small volume and finally was poured dropwise in 500 ml of deionized water. A white precipitate was formed, filtrated and washed with cold water and hexane. The obtained diol, a white powder, was dried under vacuum at 40 °C overnight. Yield: 3.7 g (83%).

¹H NMR (DMSO-d₆, δ, ppm): 2.33 (s, 6H, –CH₃), 6.82 (s, 2H, Ar-H), 7.2–7.4 (d.d., 8H, Ar-H), 8.84 (s, 2H, –OH).

Synthesis of dimethylphenyl substituted poly(aryl ether sulfone)s (DMe(x)PAES)

The parent copolymer DMe(x)PAES is synthesized by nucleophilic aromatic substitution polymerization as shown in Scheme 1. A typical procedure for the synthesis of DMe(x)PAES, where x denotes the molar ratio of 2,5-bis(4-methylphenyl)benzene-1,4-diol is described below. For example, DMe(80)PAES contains 0.8 mol of 2,5-bis(4-methylphenyl)benzene-1,4-diol and 0.2 mol of bisphenol A per 1 mol of bis(4-fluorophenyl)sulfone. In a degassed, 100 ml round bottomed flask, equipped with a dean-stark trap, a mixture of 2,5-bis(4-methylphenyl)benzene-1,4-diol (2.0000 g, 6.89 mmol), bisphenol A (0.3931 g, 1.72 mmol), bis(4-fluorophenyl)sulfone (2.1892 g, 8.61 mmol), dimethylformamide (32 ml), toluene (12 ml) and K₂CO₃ (1.3804 g, 9.99 mmol) was added. The mixture was heated at 155 °C and left overnight under argon atmosphere. The following day the reaction temperature was raised at 185 °C to remove the azeotropic mixture, and the reaction mixture was left at this temperature overnight. The mixture was cooled down and poured dropwise in ×10 V of MeOH to obtain a white-pink precipitate in the form of small spheres. The precipitate was filtered and left stirring in H₂O at 60 °C for 1 day. Finally, the polymer was filtered, washed with H₂O and hexane thoroughly and dried under vacuum at 60 °C overnight.

Scheme 1 Synthetic route for preparation of quaternary phosphonium based QPOH(x)PAES.

Bromination of dimethylphenyl substituted poly(aryl ether sulfone)s (Br(x)PAES)

The bromination of the DM(x)PAES was carried out in chloroform or 1,2-dichloroethane via radical reaction using NBS as the bromination agent and BPO as initiator. A typical procedure is the following: in a degassed, filled with argon atmosphere, 100 ml round bottomed flask, DM(80)PAES (2 g, 4.063 mmol) was completely dissolved in 40 ml of fresh, distilled CHCl₃ followed by the addition of N-bromosuccinimide (NBS) (1.2728 g, 7.1514 mmol) and benzoylperoxide (BPO) (0.0866 g, 0.3576 mmol). The reaction mixture was heated at 85 °C overnight under argon atmosphere and then precipitated in 10× V of MeOH to obtain a yellow powder. The resulting product was filtered, washed with MeOH and hexane and was dried under vacuum at 40 °C overnight. The polymer is named Br(80)PAES, where Br denotes the bromination reaction.

Synthesis of quaternary phosphonium-functionalized poly(aryl ether sulfone)s (QPOH(x)PAES)

The quaternary phosphonium-functionalized membranes were synthesized as described previously.^28b A typical procedure was as follows: the brominated copolymer Br(80)PAES (0.1292 g, 0.3208 mmol CH₂Br groups) was dissolved in 7 ml NMP under argon. Then, tris(2,4,6-trimethoxyphenyl)phosphine (TTMPP) (0.1708 g, 0.3208 mmol) was added. The reaction mixture was kept under inert atmosphere for 48 h at 80 °C and then was poured in an aluminum casting plate and left for 48 h at 80 °C, so that the NMP evaporated completely. Finally, the casted membrane was peeled off after being immersed in water. The final membranes in the hydroxide form (QPOH(80)PAES) were obtained after soaking in 1 M KOH solution at room temperature for 48 h, where the solution was exchanged several times. The membranes were washed with deionized water to remove residual KOH until the pH was neutral and kept in deionized water before use.

Synthesis of self cross-linked membranes (CL(y)-QPOH(x)PAES)

Both quaternary phosphination and self cross-linking were achieved via a simple, one step reaction, as described by Gu et al.^28a For the quaternary phosphination reaction, as mentioned before, stoichiometric molar ratio of –CH₂Br groups/TTMPP was used while for the fabrication of the self-cross linked membranes the –CH₂Br groups are in excess compared to TTMPP. The –CH₂Br groups react first with TTMPP to form the quaternary phosphonium groups and the residual bromomethylated groups (–CH₂Br) will subsequently react with the aromatic rings of the already formed quaternary phosphonium groups via Friedel–Crafts reaction. Self cross-linked membranes CL(y)-QPOH(x)PAES (y denotes the degree of cross-linking) with different cross-linking degrees (DCL) were obtained by varying the TTMPP/CH₂Br groups ratio (0.8–0.33). The self cross-linked membranes in bromide form were converted to the hydroxide form (CL(y)-QPOH(x)PAES) after soaking in 1 M KOH solution at room temperature for 4 days, where the solution was exchanged several times. The self-cross-linked membranes were washed with deionized water to remove residual KOH until the pH was neutral and kept in deionized water before use.

The degree of cross-linking (% DCL) was calculated based on the insoluble fraction of the membranes after immersion in DMAc for 2 days at 60 °C (the polymer concentration was the same for all samples). The insoluble fraction was filtered, washed thoroughly with deionized water and then dried under vacuum at 60 °C overnight. The dried membranes were weighed and % DCL was calculated by the following equation:

where F_ins is the insoluble fraction of the dried membrane after 2 days in DMAc and F is the weight of the dried membrane before solubility test.

Results and discussion

Ideally, AAEMs should have sufficient mechanical and chemical stability, while providing high OH⁻ conductivity at moderate water swelling. A strategy to achieve this goal is to introduce cationic functional groups onto an aromatic polyether backbone which combines thermal and chemical stability with excellent mechanical properties. In specific, our idea is to use a mechanical robust poly(arylene ether sulfone) backbone where no cationic links are directly attached that could activate the ether links for cleavage through nucleophilic attack by hydroxide ions, thus resulting in an alkaline and mechanical stable polymeric backbone that will be modified to produce AAEMs.

Synthesis of dimethylphenyl substituted poly(aryl ether sulfone)s

Poly(arylene ether sulfone)s containing dimethylphenyl groups (DMe(x)PAES) were synthesized via nucleophilic aromatic substitution polycondensation and functionalized with quaternary hydroxide phosphonium groups in three steps including (a) bromination of pendant methyl groups and conversion to bromo methylated analogues (b) quaternary phosphination using tris(2,4,6-trimethoxyphenyl)phosphine, (c) alkalization with 1 M KOH solution. The synthetic path followed for the preparation of quaternary phosphonium-functionalized poly(aryl ether sulfone)s (QPOH(x)PAES) is depicted in Scheme 1.

Monomer 2,5-bis(4-methylphenyl)benzene-1,4-diol was synthesized via suzuki coupling of p-tolyl boronic acid and 2,5-dibromo-1,4-bis(tetrahydro-2H-pyran-2-yloxy)benzene, followed by treatment with HCl to cleave THP protecting groups, as shown in Scheme 2. p-Tolyl boronic acid⁴¹ and 2,5-dibromo-1,4-bis(tetrahydro-2H-pyran-2-yloxy)benzene⁴³ were synthesized according to literature procedures. The structure of the monomer was identified by ¹H NMR spectroscopy (Fig. 1).

Scheme 2 The synthesis of 2,5-bis(4-methylphenyl)benzene-1,4-diol.

Fig. 1 ¹H NMR spectrum of 2,5-bis(4-methylphenyl)benzene-1,4-diol.

Poly(arylene ether sulfone)s containing dimethylphenyl groups (DMe(x)PAES) were synthesized via nucleophilic aromatic substitution polycondensation of bis(4-fluorophenyl)sulfone with 2,5-bis(4-methylphenyl)benzene-1,4-diol monomer and bisphenol-A. By varying the feed ratios of diols, copolymers with different molar percentages of pendant methyl groups were obtained. The chemical composition of the synthesized copolymers was determined by ¹H NMR spectroscopy, based on the methyl protons a of p-tolyl groups at 2.17 ppm and the methyl protons b of bisphenol A unit at 1.65 ppm (Fig. 2). The molar ratios of the two comonomers calculated by ¹H NMR are in good agreement with the theoretical feed ratio, as shown in Table 1. The synthesized copolymers were soluble in common organic solvents such as chloroform, DMF, NMP and showed high molecular weights as estimated by GPC, with very good film forming properties except for the case of copolymer DMe(95)PAES which forms membranes with poor mechanical stability. The molecular characteristics of the synthesized copolymers are shown in Table 1.

Fig. 2 ¹H NMRs of the synthesized DMe(70)PAES, brominated analogues Br(70)PAES and quaternary phosphonium based QPOH(70)PAES.

Table 1 Molecular weight, polydispersity index, membrane quality and degree of bromination of DMe(x)PAES

Sample	Theoretical feed ratio	Feed ratio (^1H NMR)			PDI	Membrane quality	DB^a (%)
a Degree of bromination as determined by ^1H NMR.b The bromination reaction was carried out in chloroform.c The bromination reaction was carried out in 1,2-dichloroethane.
DMe(95)PAES	95/5	95/5	11 000	16 000	1.42	+	81^b
DMe(90)PAES	90/10	88/12	39 000	57 000	1.43	++	79^b
DMe(90)PAES	90/10	86/14	56 000	79 000	1.40	++	30^c
DMe(80)PAES	80/20	79/21	51 000	72 000	1.41	+++	47^c
DMe(80)PAES	80/20	81/19	37 000	57 000	1.52	+++	95^b
DMe(70)PAES	70/30	68/32	47 000	67 000	1.43	+++	92^b
DMe(70)PAES	70/30	65/35	41 000	59 000	1.43	+++	92^b

---

Bromination and quaternary phosphination

The bromination of poly(arylene ether sulfone)s containing dimethylphenyl groups (DMe(x)PAES) was carried out via radical-mediated reaction using N-bromosuccinimide, benzoylperoxide and chloroform or 1,2-dichloro ethane as solvents.⁴⁵ The comparison of the ¹H NMR spectra of the native DMe(70)PAES and the bromomethylated analogue, as depicted in Fig. 2, confirms the successful bromination reaction. In particular, a new peak k appears at 4.59 ppm in the brominated copolymer which is assigned to the protons in bromomethyl groups (–CH₂Br) while the peak a at 2.17 ppm attributed to methyl protons of p-tolyl groups is much reduced compared to the unbrominated DMe(x)PAES. In the aromatic region, all peaks corresponding to aromatic protons are down-shifted compared to the parent copolymer DMe(70)PAES due to the deshielding effect induced by the introduction of bromomethylated groups. The degree of bromination was calculated by ¹H NMR based on the integral ratio of the bromomethyl groups (–CH₂Br) k protons to b protons of bisphenol A methyl groups and is given in Table 1. It should be noticed that the solvent plays a critical role in the yield of bromination, e.g. high yields (over 80%) were obtained when chloroform used as a solvent compared to 1,2-dichloroethane.

The reaction of the bromomethylated precursor copolymers with tris(2,4,6-trimethoxyphenyl)phosphine (TTMPP) and casting of the solution at 80 °C, resulted in quaternary phosphonium based membranes in Br⁻ form. These membranes were subsequently converted to OH⁻ form (AAEM) after treatment with 1 M KOH solution at room temperature for 2 days. The comparison of ¹H NMR spectra of the quaternary phosphonium (QP) based copolymer in hydroxide form (QPOH(70)PAES) and its parent copolymer Br(70)PAES in Fig. 2 reveals the appearance of three new peaks at 6.14, 3.73 and 3.48 ppm after quaternary phosphination. The first peak is assigned to the n aromatic protons in tris(2,4,6-trimethoxyphenyl) quaternary phosphonium group, while the second and the third are assigned to para-methoxy protons o and ortho-methoxy protons m in tris(2,4,6-trimethoxyphenyl) quaternary phosphonium group, respectively. The broad peak at 4.56 ppm is assigned to the positively charged phosphorous adjacent methylene protons l. The successful quaternary phosphination was also evidenced by ³¹P NMR spectroscopy. Free tris(2,4,6-trimethoxyphenyl)phosphine (TTMPP) has a strong signal at −65.6 ppm, while after phosphination only a new signal at +5.5 ppm is observed, which is in agreement with ³¹P NMR data reported by others,^28b indicating that a positively charged phosphorous species was formed (Fig. S2†). Clearly, all above NMR data confirm that QPOH(x) PAES have been successfully synthesized.

Further evidence for the successful introduction of the phosphonium groups was also collected by ATR-FT-IR spectroscopy. In Fig. 3, it is shown the pristine copolymer (DMe(80)PAES), the bromomethylated analogue (Br(80)PAES) and the corresponding quaternary phosphonium based AAEM (QPOH(80)PAES).

Fig. 3 ATR-FT-IR spectra of the synthesized DMe(80)PAES, brominated analogues Br(80)PAES and quaternary phosphonium based QPOH(80)PAES.

The new peak at 606 cm⁻¹ observed in the Br(80)PAES is assigned to the newly formed C–Br bond vibration,⁴⁶ confirming the successful bromination reaction. After reaction with the tris(2,4,6-trimethoxyphenyl)phosphine at least three new peaks arose at 917, 950 and 1124 cm⁻¹, respectively. These new peaks are assigned to phosphonium groups as already reported by other groups,^47,48 thus denoting that successful incorporation of the phosphonium groups. The characteristic absorption for hydroxide groups appears as a strong broad band ranging from 3100–3700 cm⁻¹, while the new absorption bands at 2833 and 2933 cm⁻¹ which result from the stretching vibration of aliphatic C–H bonds of methoxy groups, further confirm the successful conversion into the quaternary phosphonium groups.

Another indication for the successful functionalization was the solubility of QPOH(x)PAES in methanol (the brominated analogue is not soluble in methanol) which is a signature of the quaternary phosphonium groups.^28b This is important since they can be used as ionomers for improving the performance of APEFCs.

IEC, water uptake and swelling ratio of QPOH(x)PAES

Ion exchange capacity (IEC) is an important parameter which dominates AAEM properties such as water uptake, dimensional stability, ionic conductivity, and mechanical strength. Table 2 compares the IEC, water uptake, swelling ratio of the QPOH(x)PAES membranes. The IEC values (IEC_m) determined by titration were lower compared to the theoretical ones (IEC_t) calculated by ¹H NMR based on the degree of bromination and assuming 100% quaternary phosphination conversion and 100% conversion in hydroxide form. A possible explanation for the lower IEC_ms can be either the incomplete conversion of bromomethyl groups to quaternary phosphonium groups which is commonly observed and reported for quaternized poly(arylene ether sulfones)³⁰ or the partial ion exchange ability (incomplete conversion in hydroxide form). The incomplete conversion of bromomethyl groups to quaternary phosphonium groups is probably due to the sterical hindrance of TTMPP which makes it difficult to come in contact with the benzyl bromide groups.

Table 2 Physico-chemical properties of QPOH(x)PAES and their self cross-linked analogues

Membrane	Molar ratio TTMPP/CH2Br	DCL^a (%)	IECt^b	IECm^d	Water uptake (wt%)		SR (%)
					20 °C	60 °C	20 °C	60 °C
a Degree of cross-linking (DCL) was calculated based on solubility test.b Calculated by ^1H NMR (meq. g^−1).c Determined by the molar ratio CH2Br/TTMPP.d Measured by titration (meq. g^−1).
QPOH(70)PAES	1.0	—	1.07	0.66	32.0	45.0	10.0	17.8
QPOH(80)PAES	1.0	—	1.18	0.73	44.0	71.1	11.5	24.2
CL66QPOH(70)PAES	0.6	66	0.83^c	0.41	20.2	25.0	7.5	11.2
CL19QPOH(80)PAES	0.8	19	1.07	0.53	38.0	40.1	9.1	12.2
CL52QPOH(80)PAES	0.6	52	0.93	0.42	25.0	31.0	8.1	11.5
CL81QPOH(80)PAES	0.5	81	0.85	0.23	17.1	19.2	6.2	8.6
CL89QPOH(80)PAES	0.33	89	0.66	0.20	12.0	18.1	5.3	6.5
PEEKQPOH 70% (ref. 28e)	1.0	—	1.03	0.89	∼30.0	∼32.0	∼16.0	∼16.0

---

Generally, the water uptake of the membranes increased with IEC value due to increased hydrophilicity. The highest water uptake of the QPOH(80)PAES membrane was 44.0 wt% at 20 °C for the highest IEC_m (0.73 meq. g⁻¹) which is increased compared to that of quaternary phosphonium poly(ether ether ketone) (PEEKQPOH) with IEC_m = 0.89 meq. g⁻¹.

The water uptake obviously increases with temperature, e.g. the water uptake of the QPOH(80)PAES membrane was 71.1 wt% at 60 °C while for PEEKQPOH had very little change.

The swelling ratio reflects the wet dimensional stability of the AAEMs at various conditions. Similar to water uptake behavior, the swelling ratio also increased with IEC and temperature. The swelling ratio was 11.5% for QPOH(80)PAES at 20 °C which is lower compared to the one of PEEKQPOH (16%). However, at higher temperatures the swelling ratio is increased to 24.2%.

Alkaline stability of QPOH(x)PAES

The chemical stability of the prepared AAEMs in alkaline media is one of the main issues addressed in this work. Thus, a detailed study using ATR-FT-IR and NMR as analytical tools for the precise identification of the degraded chemical structures/moieties has been conducted since it is essential to gain mechanistic insights that could help to the design of more robust materials.

The alkaline stability of the prepared AAEMs membranes was evaluated by immersion of the AAEM into 1 M aqueous KOH solution at 25 and 60 °C, respectively. Samples were removed after 144 and 288 hours treatment, washed thoroughly with water until the pH was neutral, and studied via ¹H NMR, ³¹P NMR and ATR-FT-IR spectroscopy. All membranes maintain their mechanical properties and did not become brittle even after exposure in alkali at 60 °C, supporting that the high chemical stability of polymeric backbone is preserved.

Moreover, the DMe(x)PAES by itself was resilient in alkaline media as revealed by ¹H NMR and GPC analysis (data not shown), further confirming the excellent poly(aryl ether sulfone) backbone alkaline stability.

This result is contradictory to the findings by Ramani et al.²⁹ who established that the direct tethering of cation groups to the benzyl position of polysulfone backbone caused the polymeric backbone to chemically degrade through several different hydrolysis mechanisms in alkaline media. However, in our case an additional phenyl ring acts as a spacer between the cation and the polymer backbone. The presence of the spacer suppresses the electron withdrawing effect of the affixed cation, thereby minimizing the delocalization of the electronic structure of the polymer backbone, rendering it less susceptible to hydroxide ion attack.

Regarding the phosphonium group stability in alkaline media, the comparison of ¹H NMR spectra of QPOH(70)PAES membrane before and after alkaline exposure for 144 and 288 hours at 25 °C (Fig. S1†) although did not reveal the appearance of any new peaks, however, the two single peaks at 3.73 and 3.48 ppm which are assigned to para-methoxy protons and ortho-methoxy protons in tris(2,4,6-trimethoxyphenyl) quaternary phosphonium group respectively, after alkaline treatment were split into several peaks in the region 3.4–3.9 ppm, implying the chemical environment change of the methoxy groups. Moreover, the intensity ratio of I_o-OMe/I_p-OMe was reduced over time (the integral of the peaks corresponding to methoxy groups is difficult to be measured due to overlapping with the peak of water at 3.5 ppm). It was also observed that the intensity of the signal of methylene protons directly attached to QP groups at 4.56 ppm (I_QP(–CH2), l protons) is decreased over time in relation to the intensity of the peak at 1.65 ppm (I_BisA(–CH3), b protons) assigned to the methyl groups of the bisphenol unit. However, as evidenced by ³¹P NMR (Fig. S2†), except of the signal at +5.5 ppm, a new signal at +24.8 ppm also appeared, suggesting that two different positively charged phosphorous species are present after alkaline exposure. Thus, the CH₂–P⁺ groups were partially degraded via a mechanism that should be identified. The signal at +24.8 ppm is believed to correspond to the stabilized phosphonium phenoxide group^49,50 as detailed below.

ATR-FT-IR spectroscopy (Fig. S3†) revealed that after alkaline treatment, the peaks at 2840 and 2940 cm⁻¹ attributed to C–H stretching vibration of methoxy groups are reduced over time. These evidence derived from the different analytical methods should be combined to avoid resulting in an incorrect interpretation of the chemical degradation products, and hence the incorrect degradation mechanism.

Fig. 4 depicts the ¹H NMR spectra of QPOH(80)PAES before and after alkaline exposure at 60 °C for 144 and 288 hours. Having a closer inspection, after alkaline treatment, four new peaks with increasing intensity over time appeared in the region 6.25–6.55 ppm, probably belonging to aromatic protons of a high electron density substituted benzene ring as a degradation product. The same trends, regarding the splitting of methoxy groups into several peaks in the region 3.4–3.9 ppm and the reduced ratio of I_o-OMe/I_p-OMe were also observed at 60 °C after alkaline exposure. The observed reduction of the intensity of the signal of methylene protons adjacent to QP groups at 4.56 ppm (I_QP(–CH2), l protons) in relation to the intensity of the peak at 1.65 ppm is much more pronounced over time and with increasing temperature. The percentage (%) of the remaining quaternary phosphonium groups of different AAEMs after exposure to 1 M KOH at different temperatures and time periods calculated by ¹H NMR, are given in Table S1.† The IEC for the QPOH(70)PAES membrane after 144 hours exposure in 1 M KOH at 60 °C was also measured. This membrane retained 89% of its initial IEC (initial IEC = 0.73 meq. g⁻¹ and after alkaline stability test IEC = 0.65 meq. g⁻¹), a result which is in good agreement with the percentage of the remaining cationic groups under the same conditions, as calculated by ¹H NMR (94%, Table S1†).

Fig. 4 ¹H NMR spectra of QPOH(80)PAES before and after exposure in 1 M KOH at 60 °C for 144 and 288 hours.

In order to gain mechanistic insight into the degradation mechanism, ³¹P NMR spectroscopy was also employed. The ³¹P NMR spectra of QPOH(80)PAES before and after alkaline treatment at 60 °C for 288 hours are depicted in Fig. 5. The signal of QPOH(80)PAES at +5.5 ppm which corresponds to tris(2,4,6-trimethoxyphenyl)phosphonium group, after alkaline treatment, appeared along with two new signals, one in the positive field +24.8 ppm and the other in the negative field −14.3 ppm, respectively. As mentioned before, the signal at +24.8 ppm is attributed to phosphonium phenoxide while the signal at −14.3 ppm corresponds to a free, neutral phosphine. The exclusive appearance of the signal at −14.3 ppm at 60 °C along with the signals at +5.5 ppm and +24.8 ppm, implies that an additional but neutral phosphorus species is formed with increasing temperature.

Fig. 5 ³¹P NMR spectra of QPOH(80)PAES before and after treatment in 1 M KOH at 60 °C for 288 hours.

Considering the above data, Scheme 3 proposes several degradation mechanisms including the typical direct nucleophilic attacks, the Cahours–Hofmann reaction and the formation of a stabilized phosphonium phenoxide fostered by yield intermediates.

Scheme 3 Proposed degradation mechanisms for QPOH(x)PAES.

It is believed that the SN2 reaction of the phosphonium cation to TTMPP did not happen. The peak at 4.5 ppm in ¹H NMR (Fig. 4) – signature character of the CH₂-bounded to an alcohol group (degradation mechanism A) could not be easily detected due to overlapping with the peak at 4.56 ppm assigned to methylene protons adjacent to positively charged phosphorous (CH₂–P⁺) and also the TTMPP was not detected (absence of the peak at −65.6 ppm in ³¹P NMR, Fig. 5). Moreover, the evidence of the splitting of methoxy groups as well as the presence of new peaks in the region 6.25–6.55 ppm cannot be correlated with the proposed degradation products.

Based on the literature data, the degradation for small molecule phosphonium cations in alkaline conditions results in phosphine oxide and a hydrocarbon as byproducts, the as referred Cahours–Hofmann reaction.⁵¹ This reaction is a four step process as reported by McEwen and coworkers⁵² and is illustrated as mechanism C. In this case, one would expect to yield to the p-methyl phenyl substituted aromatic polymeric backbone with a characteristic peak at 2.17 ppm in ¹H NMR, which has not been observed in Fig. 4, suggesting that this byproduct has not been formed. Moreover, the characteristic single signal at +10.8 ppm of tris(2,4,6-trimethoxyphenyl)phosphine oxide⁵³ was not detected in ³¹P NMR spectrum after 288 h alkaline exposure at 60 °C. The presence of the four new peaks in the region 6.25–6.55 ppm in ¹H NMR spectrum (see inset, Fig. 4) in combination with the splitting of the two single peaks attributed to p- and o-methoxy groups into several peaks after alkaline treatment probably can be explained if one considers that yield formation (production of methylene carbanion that is in the presence of a positively charged phosphorus atom) takes place. Yields are highly reactive, thus methylene carbanion attacks to methoxy group, leading to the formation of the stabilized phosphonium phenoxide (mechanism D). In the ³¹P NMR spectra (Fig. 5) the signal observed at +24.8 ppm is due to the formation of the stabilized phosphonium phenoxide as already reported,^49,50 thus confirming the above speculation. Additionally, methine protons adjacent to positively charged phosphorus is expected to be detected around 4.55 ppm,⁵⁴ but due to overlapping with the signal of methylene protons adjacent to positively charged phosphorus (–CH₂–P⁺), it is difficult to be distinguished. However, the reduced intensity of the peak at 4.56 ppm (–CH₂–P⁺) after alkaline exposure as evidenced by ¹H NMR data (Fig. 4 and S1†) could also be attributed to the presence of methine groups due to their reduced number of integrated protons in relation to methylene protons. The methyl group (–CH₃CH–P⁺) should be observed around 1.7 ppm,⁵³ which is also difficult to be detected due to overlapping with the peak centered at 1.65 ppm assigned to the methyl groups of the bisphenol unit. Finally, the presence of the four new doublet peaks could be assigned to the aromatic protons of phenoxide unit which are downshifted compared to the corresponding aromatic protons of tris(2,4,6-trimethoxyphenyl)phosphonium group centered at 6.1 ppm.

Regarding the assignment of the exclusive signal at −14.3 ppm after alkaline exposure at 60 °C in Fig. 5, is believed to correspond to a neutral polymer bounded phosphine, which is the main degradation product of mechanism B. The 2,4,6-trimethoxy phenol byproduct was not detected either due to overlapping of its aromatic protons at 6.0 ppm with the aromatic protons of tris(2,4,6-trimethoxyphenyl)phosphonium group or because it was removed upon rinsing of the membrane with water after alkaline treatment. Thus, it can be concluded that the appearance of two positively charged phosphorous at +5.5 ppm and +24.8 ppm which correspond to the non degraded tris(2,4,6-trimethoxyphenyl)phosphonium group and the stabilized phosphonium phenoxide respectively, along with the signal at −14.3 ppm assigned to a neutral polymer bounded phosphine suggests that at higher temperatures there is a strong nucleophilic attack at trimethoxy benzene ring of phosphonium group resulting to the formation of a neutral polymer bounded phosphine. It should be noted that the presence of the stabilized phosphonium phenoxide probably enables ionic conductivity since the electrostatic attraction between OH⁻ and positive phosphorous is not so strong due to the presence of phenoxide unit (Ph–O⁻), thus leading to improved hydroxide mobility.

Very recently, Zhang et al.⁵⁵ studied the degradation mechanism and kinetics of a series of quaternary phosphonium cations including benzyl tris(2,4,6-trimethoxyphenyl)phosphonium (small-molecule model) revealing that the later has much higher alkaline stability than the benchmark cation, benzyl trimethyl ammonium. However, a multi step methoxy triggered degradation mechanism of benzyl tris(2,4,6-trimethoxyphenyl)phosphonium was proposed, although the electronic and steric effects of methoxy substituents on the benzene rings effectively protect the P center against OH⁻ nucleophilic attack. Thus, these data correlate well with the proposed degradation mechanisms of this work. By replacing methoxy substituents with methyl groups, the tris(2,4,6-trimethylphenyl)phosphonium cation showed superior alkaline stability.

Synthesis of self-cross linked CL(y)QPOH(x)PAES

In order to overcome the problem of extensive swelling and limited mechanical integrity as well as to improve further the chemical stability, covalent cross-linking via Friedel–Crafts electrophilic substitution C-alkylation was chosen for the preparation of self-cross-linked membranes. As reported by Gu et al.,^28a first, the more active quaternary phosphination reaction takes place followed by the self cross-linking reaction in one step procedure. The –CH₂Br groups react initially with TTMPP to form the quaternary phosphonium groups and the residual bromomethylated groups (–CH₂Br) will subsequently react with the aromatic rings of the already formed quaternary phosphonium groups via Friedel–Crafts reaction (Scheme 4).

Scheme 4 Self-cross linked AAEMs.

By varying the molar ratio of TTMPP to –CH₂Br groups of brominated copolymer (Br(x)PAES) (0.8–0.33), cross-linked membranes with different cross-linking degrees (DCL) were obtained (Table 2). The gel fraction (%, solubility test) was used as an indirect method to measure the cross-linking degrees (DCL%). The theoretical degrees of cross-linking (DCL_t) were calculated by assuming that all the residual bromomethyl groups completely cross-linked with quaternary phosphonium groups and are in good agreement with the experimental degrees of cross-linking (Table 3). This is a confirmation that self cross-linking reaction has high efficiency.

Table 3 Properties of self-cross linked membranes CL(y)QPOH(x)PAES

Sample	Molar ratio TTMPP/CH2Br	DCLt^a(%)	DCL (%)	Solubility				Membrane properties
				DMAc	NMP	MeOH	DMSO
a DCLt, theoretical degree of cross-linking, assuming that all of the residual bromomethyl groups completely cross-linked with quaternary phosphonium groups. + soluble, − insoluble.
QPOH(70)PAES	1	−	−	+	+	+	+	Brittle
QPOH(80)PAES	1	−	−	+	+	+	+	Brittle
CL66QPOH(70)PAES	0.6	67	66	−	−	−	−	Tough, flexible
CL19QPOH(80)PAES	0.8	25	19	Partial soluble	−	−	−	Tough, flexible
CL52QPOH(80)PAES	0.6	67	52	Swollen, partial soluble	Swollen	−	Swollen	Tough, flexible
CL81QPOH(80)PAES	0.5	100	81	Swollen	Swollen	−	Swollen	Tough, flexible
CL89QPOH(80)PAES	0.33	200	89	Swollen	Swollen	−	Swollen	Rigid, flexible

---

The resulting membranes had improved mechanical properties compared to their non-crosslinked counterparts, were not soluble in MeOH and other common solvents, and were only slightly soluble in solvents like DMAc and NMP (Table 3).

As the membranes were not dissolved completely in any deuterated solvent, ¹H NMR spectroscopy was not used for their study.

The successful cross-linking reaction was confirmed by ATR-FT-IR spectroscopy (Fig. 6). It was observed that upon increasing cross-linking degree, the intensity of the peaks at 874 and 834 cm⁻¹ which are characteristic absorptions of Ar-H bonds in the penta substituted benzene rings of tris(2,4,6-trimethoxyphenyl)phosphonium groups and in para substituted benzene rings of side groups (the one substituent is 2,4,6-trimethoxy benzene) respectively, was increased. Moreover, the peak centered at 876 cm⁻¹ which is attributed to tetra-substituted benzene rings of 2,5-bis(4-methylphenyl)benzene-1,4-dioxy moieties overlaps with the peak at 874 cm⁻¹. Meanwhile, an obvious decrease is observed of the band at 855 cm⁻¹ which is assigned to the tetra-substituted benzene rings of tris(2,4,6-trimethoxyphenyl)phosphonium groups.

Fig. 6 ATR-FT-IR spectra of the cross-linked CLQPOH(80)PAES with different degrees of cross-linking.

These are the direct proofs that successful cross-linking reaction at the meta positions of trimethoxyphenyl phosphonium groups took place.

IEC, water uptake and swelling ratio of CL(y)QPOH(x)PAES

In Table 2 are listed IECs, water uptakes, swelling ratios of QPOH(x)PAES and cross-linked CL(y)QPOH(x)PAES. The IEC_ms of the cross-linked membranes CL(y)QPOH(80)PAES are significantly reduced compared to uncross-linked QPOH(80)PAES membranes and vary from 0.53 to 0.20 meq. OH⁻ per g, depending on the degree of cross-linking. The lower IEC_m values of the cross-linked membranes can be attributed to the partial consumption of the bromomethyl groups in the cross-linking reaction. Moreover, the formation of a tight cross-linking network in which the ions are trapped and cannot be easily ion exchanged also has a negative impact on the IEC. This was evidenced by EDX analysis where it was demonstrated that bromide ions (Br⁻) were present at the surface of the membrane even after ion exchange to hydroxide form (Fig. 7b). After cross-linking, the water uptake shows a sharp decline: at 60 °C the uncross-linked membrane QPOH(80)PAES had a water uptake 71.1 wt% and for the CL81QPOH(80)PAES the value was only 19.2 wt%, as evidenced in Table 2. As expected, due to the drastic water uptake decrease, the swelling ratio was also lowered, indicating that the formation of a cross-linked network improved the dimensional stability of the membranes. In specific, the swelling ratio at 60 °C was decreased from 24.2% for the uncross-linked QPOH(80)PAES to 8.6% for the cross-linked membrane CL(81)QPOH(80)PAES. This is one of the main reasons to do the cross-linking treatment.

Fig. 7 (a) Cross-sectional SEM micrograph of CL66QPOH(70)PAES membrane, (b) EDAX analysis of the surface of CL66QPOH(70)PAES membrane, (c) TEM micrograph of non cross-linked QPOH(70)PAES membrane stained with KI.

Morphology of CL(y)QPOH(x)PAES

To investigate the phase separation and ionic aggregation of uncross-linked and their cross-linked counterparts, cross-sectional SEM micrographs were studied. In Fig. 7a is depicted the SEM micrograph of the CL66QPOH(70)PAES membrane which showed a phase separated morphology where the dark regions represent ionic clusters composed of phosphonium cations and the bright areas represent hydrophobic domains composed of the hydrophobic polymer backbone. To further support the formation of ionic clusters, QPOH(70)PAES based membranes were treated with KI solution to exchange OH⁻ to I⁻and studied with Transmission Electron Microscopy (TEM). As it is shown in Fig. 7c, the membranes showed a phase separated morphology where the dark regions represent ionic clusters composed of phosphonium cations with iodide counterions (ion-stained electron rich hydrophilic domains) and the bright areas represent hydrophobic domains composed of the hydrophobic polymer backbone. These clusters are evenly dispersed in the polymeric matrix and their size is between 3 to 6 nm. Thus, the same phase separated morphology was also observed in the case of the non-cross linked counterpart suggesting that the cross-linking had no adverse effect on the development of hydrophilic hydrophobic phase separation and the formation of ionic clusters. This distinctive phase separated morphology is expected to provide efficient ion transport pathways^1,19 where phosphonium ions form the OH⁻ transport channels, thus ensuring the high ionic conductivity of cross-linked based membranes.

Ionic conductivity of CL(y)QPOH(x)PAES

The ionic conductivities of the cross-linked AAEMs were measured in chloride anion form as this form is not sensitive to exposure to atmospheric CO₂. It is very difficult to exclude all CO₂ from the standard rapid 2-probe test cells, which makes the measurement of hydroxide conductivity difficult and less repeatable due to the rapid conversion of the membranes to the HCO₃⁻/CO₃²⁻ anion forms. The Cl⁻ conductivities as a function of temperature are depicted in Fig. 8. It is observed, as expected, that the conductivity is increased with temperature and IEC. In specific, the Cl⁻ conductivity for CL19QPCl(80)PAES membrane with IEC_m = 0.53 meq. g⁻¹ is 1.36 mS cm⁻¹ at room temperature and increases to 2.67 mS cm⁻¹ at 70 °C, while for commercial available Fumasep FAA-3-PK-130 membrane with IEC = 1.43 meq. g⁻¹ the conductivity value is higher, reaching up to 9.5 mS cm⁻¹ at 70 °C. The Cl⁻ conductivities values for these cross-linked membranes are comparable with other phosphonium based AAEMs.^28g Interestingly, the hydroxide ion is known to have about 2.6 times greater ion mobility in infinitely dilute solutions than the chloride ion. Thus, hydroxide conductivities can be estimated by multiplying the Cl⁻ conductivities by a factor of 2.6. A similar approach, for the indirect estimation of hydroxide conductivity by using the measured HCO₃⁻ conductivity and multiplying by 3.8 (derived from the differences in ion mobilities) has been recommended by Hickner and Yan⁴⁵ and validated by others.²¹ The Cl⁻ conductivities as well as the calculated hydroxide conductivities are shown in Fig. 9. The calculated hydroxide conductivity values for CL19QPOH(80)PAES and CL52QPOH(80)PAES are 6.94 mS cm⁻¹ and 4.26 mS cm⁻¹ at 70 °C, respectively. However, these conductivity values correspond to very low IECs (0.52 meq. g⁻¹ and 0.43 meq. g⁻¹ for C19QPOH(80)PAES and CL52QPOH(80)PAES, respectively). Since IEC dominates ion conductivity with almost a linear relationship, thus IEC-normalized hydroxide conductivity could be more objective to assess the intrinsic hydroxide conduction. The IEC normalized hydroxide conductivity value for CL19QPOH(80)PAES and some other reported cross-linked AAEMs are listed in Table 4 for the convenience of the comparison.

Fig. 8 Conductivities of cross-linked membranes and state of the art Fuma-Tech reinforced membrane Fumasep FAA-3-PK-130 in chloride form as a function of temperature.

Fig. 9 The chloride and calculated hydroxide conductivities of cross-linked membranes as a function of temperature.

Table 4 Hydroxide conductivity and ion exchange capacity-normalized hydroxide conductivity of cross-linked CL19QPOH(80)PAES and some reported AAEMs^a

Cross linked AAEMs	IECm^b (meq. g^−1)	σ^c (mS cm^−1)	σIEC^d (mS g cm^−1 meq.^−1)
a Q, QA refer to quaternary ammonium.b Measured by titration (meq. g^−1).c Hydroxide conductivity.d Ion exchange capacity normalized hydroxide conductivity.
CL19QPOH(80)PAES	0.53	3.5(30 °C)	6.7
SCL-TPQPOH(phosphonium)^28a	1.23	38(20 °C)	30.9
Q-PPEK-3 (ref. 56)	2.31	9(30 °C)	3.89
GA-PVA/chitosan-QA-OH^57	0.35	2.8(25 °C)	8
CPAES-Q-80 (ref. 34a)	1.43	23.2(25 °C)	16.2
x-PP-TMA-20 (ref. 58)	1.34	14.6(20 °C)	10.9
M2/4EP-6FPSF-QAOH^59	0.73	4(25 °C)	5.5
x-PEEK-Q-100 (ref. 60)	1.18	14(25 °C)	11.86
Q-PEN-0.4 (ref. 61)	1.56	15(30 °C)	9.6

---

Clearly, when the same (2,4,6-trimethoxyphenyl)phosphonium group used, the IEC normalized hydroxide conductivity of SCL-TPQPOH (PSf is the polymeric backbone) is about 5 times higher than that of the prepared CL19QPOH(80)PAES. This behavior can be probably attributed to the different microstructure of SCL-TPQPOH, in which larger ionic clusters are formed, thus facilitating the formation of interconnected broad ionic channels, consequently, leading to higher conductivity. In another comparison, with ammonium based AAEMs, the IEC normalized hydroxide conductivity of CL19QPOH(80)PAES is two times higher than that of Q-PPEK-3.⁵⁶ However, in general, the IEC normalized hydroxide conductivity values for most ammonium based AAEMs (e.g. CPAES-Q-80,^34a x-PP-TMA-20,⁵⁸ x-PEEK-Q-100 (ref. 60)) are higher than that of the prepared CL19QPOH(80)PAES. Moreover, the IEC normalized hydroxide conductivity value for CL19QPOH(80)PAES at 70 °C is 13.1 mS g cm⁻¹ meq.⁻¹.

Although the IECs of the cross-linked CL(y)QPOH(x)PAES are very low, their OH⁻ conductivities at higher temperatures are acceptable for APEMFC application, but of course can be further optimized by many strategies; e.g. the IEC increase would lead to increased hydroxide conductivities. However, in this work we are more interested in the alkaline stability of these membranes before further improvement in conductivity is addressed.

Alkaline stability of CL(y)QPOH(x)PAES

The alkaline stability of the cross-linked membranes was determined by observing changes in conductivities, FT-IR, TGA and IEC values with time in 1 M KOH solution at 60 °C for 40 days as well as in even more harsh conditions such as 4 M KOH solution at 80 °C for 16 days. All the membranes remained tough and flexible throughout the test due to the exceptional stability of polymeric backbone. The degradation is typically ex situ assessed by a decrease in conductivity or IEC values with aging time. All cross-linked membranes showed excellent alkaline stability over 40 days in 1 M KOH solution at 60 °C. The IEC_m values of the treated cross-linked membranes at various conditioning times are shown in Fig. 10. It was expected that in the first days the IEC and ion conductivity would increase slightly since in the beginning a further exchange of hydroxide ions should have taken place. Indeed, in the case of CL52QPOH(80)PAES the IEC was increased after 16 days while on the contrary for the CL19QPOH(80)PAES was decreased initially and again increased after 40 days. The nearly constant behavior of IEC (within experimental errors) suggests the excellent alkaline stability. This speculation regarding the higher absorption of hydroxide ions in the membranes in the first days was further supported by the ionic conductivities increase after 25 days aging time (not shown) and also by EDX analysis where it was demonstrated that bromide ions (Br⁻) were present at the surface of the membrane before aging test (Fig. 7b). The CL19QPOH(80)PAES membrane was also studied when exposed to harsher conditions, in 4 M KOH at 80 °C for 16 days. In specific, the membrane almost totally retained its initial IEC (initial IEC = 0.53 meq. g⁻¹ of the pristine sample and IEC = 0.52 meq. g⁻¹ after 16 days exposure to 4 M KOH at 80 °C), further confirming the excellent alkaline stability of these AAEMs.

Fig. 10 Time courses of IEC in 1 M KOH at 60 °C.

The comparison of the FT-IR results before and after alkaline exposure to 1 M and 4 M KOH at 60 and 80 °C respectively, provides further evidence supporting its excellent alkaline stability. As shown in Fig. 11, no obvious variation in intensity of absorption bands at 915, 950, 1124 cm⁻¹, corresponding to phosphonium groups was observed even after 40 days alkaline exposure in 1 M KOH at 60 °C or 16 days in 4 M KOH at 80 °C. This remarkable stability is mainly ascribed to the formation of strong cross-linked networks which leads to reduced water uptakes and consequently hindering the permeation of OH⁻ ions into the hydrophilic channels. These results demonstrate that the combination of cross-linked network and the side chain quaternary phosphonium groups is an effective strategy to maintain the performance of the ion transport channels.

Fig. 11 FT-IR spectra of CL52QPOH(80)PAES before and after exposure to 1 M and 4 M KOH solution at 60 and 80 °C, respectively.

The thermal stability of the cross-linked samples before and after alkaline exposure was also studied demonstrating that the membranes preserved their high thermal stability (the onset decomposition temperature at 174 °C attributed to the progressive decomposition of phosphonium groups) even after 40 days alkaline exposure in 1 M KOH at 60 °C or 16 days in 4 M KOH at 80 °C (Fig. 12). Such thermal stability enhancement should benefit from the stronger electrostatic attractions developed through the formation of more interconnected broad ionic clusters due to the swelling of the membrane (and further exchange to OH⁻ form) during the alkaline stability experiment at 60 °C.

Fig. 12 TGA curves of CL52QPOH(80)PAES before and after exposure to 1 M and 4 M KOH solution at 60 and 80 °C, respectively.

Conclusions

A novel 2,5-bis(4-methylphenyl)benzene-1,4-diol monomer and a series of novel tris(2,4,6-trimethoxyphenyl)phosphonium based poly(aryl ether sulfone)s and their self cross-linked counterparts were synthesized successfully. Assessment of the alkaline stability of the prepared AAEMs by using different spectroscopic techniques (¹H NMR, ³¹P NMR and ATR-FTIR) revealed that the polymer backbone has excellent alkaline stability while phosphonium groups degrade only after alkaline exposure at 60 °C for 288 h via direct nucleophilic attack of hydroxide ions to the benzene rings of tris(2,4,6-trimethoxyphenyl)phosphonium group. Self cross-linked membranes were also synthesized to overcome the poor mechanical properties and the extensive swelling of the non cross-linked analogues. The study of the effect of the cross-linking degree on IEC, swelling and water uptake showed that all these properties were mitigated with increasing degree of cross-linking. Moreover, due to the formation of the tight cross-linked network, the cross-linked AAEMs showed exceptional alkaline stability even after 40 days alkaline exposure in 1 M KOH at 60 °C or 16 days in 4 M KOH at 80 °C (via protection of phosphonium groups against hydroxide attack) and improved mechanical properties as well. Although IEC values are very low, the cross-linked AAEMs possess normalized IEC hydroxide conductivity values up to 13.1 mS g cm⁻¹ meq.⁻¹ at 70 °C.

Acknowledgements

The authors are indebted to Dr M. Daletou (FORTH/ICE-HT) and to Assist. Prof. G. Rassias (Department of Chemistry, University of Patras) for their stimulating discussions and helpful suggestions. This work was supported by Grant E052 from the Research Committee of the University of Patras (Programme K. Karatheodori).

Notes and references

1. J. R. Varcoe, P. Atanassov, D. R. Dekel, A. M. Herring, M. A. Hickner, P. A. Kohl, A. R. Kucernak, W. E. Mustain, K. Nijmeijer, K. Scott, T. Xu and L. Zhuang, Energy Environ. Sci., 2014, 7, 3135.
2. K. D. Kreuer, Chem. Mater., 2014, 26(1), 361.
3. R. C. T. Slade and J. R. Varcoe, Fuel Cells, 2005, 5, 187.
4. G. Merle, M. Wessling and K. Nijmeijer, J. Membr. Sci., 2011, 377, 1.
5. G. Couture, A. Alaaedine, F. Boschet and B. Ameduri, Prog. Polym. Sci., 2011, 36, 1521.
6. P. G. Bruce, S. A. Freunberger, L. J. Hardwick and J.-M. Tarascon, Nat. Mater., 2012, 11(1), 19.
7. B. Schwenzer, J. Zhang, S. Kim, L. Li, J. Liu and Z. Yang, ChemSusChem, 2011, 4, 1388.
8. B. E. Logan and M. Elimelech, Nature, 2012, 488, 313.
9. S. Lu, J. Pan, A. Huang, L. Zhuang and J. Lu, Proc. Natl. Acad. Sci. U. S. A., 2008, 105, 20611.
10. T. Lee, E. K. Jeon and B.-S. Kim, J. Mater. Chem. A, 2014, 2, 6167.
11. J. Ohyama, Y. Okata, N. Watabe, M. Katagiri, A. Nakamura, H. Arikawa, K.-I. Shimizu, T. Takeguchi, W. Ueda and A. Satsuma, J. Power Sources, 2014, 245, 998.
12. J. R. Varcoe, Phys. Chem. Chem. Phys., 2007, 9, 1479.
13. D. Stoica, F. Alloin, S. Marais, D. Langevin, C. Chappey and P. Judeinstein, J. Phys. Chem. B, 2008, 112, 12338.
14. H. Long and B. S. Pivovar, Phys. Chem. C, 2014, 118(19), 9880.
15. H. Long, K. Kim and B. S. Pivovar, J. Phys. Chem. C, 2012, 116(17), 9419.
16. S. Chempath, J. M. Boncella, L. R. Pratt and N. Henson, J. Phys. Chem. C, 2010, 114(27), 11977.
17. N. J. Robertson, H. A. Kostalik, T. J. Clark, P. F. Mutolo, H. D. Abruña and G. W. Coates, J. Am. Chem. Soc., 2010, 132(10), 3400.
18. H.-S. Dang and P. Jannasch, Macromolecules, 2015, 48, 5742.
19. J. Pan, C. Chen, Y. Li, L. Wang, L. Tan, G. Li, X. Tang, L. Xiao, J. Lu and L. Zhuang, Energy Environ. Sci., 2014, 7, 354.
20. L. Zhu, J. Pan, Y. Wang, J. Han, L. Zhuang and M. A. Hickner, Macromolecules, 2016, 49(3), 815.
21. O. I. Deavin, S. Murphy, A. L. Ong, S. D. Poynton, R. Zeng, H. Herman and J. R. Varcoe, Energy Environ. Sci., 2012, 5, 8584.
22. S. A. Nũnez and M. A. Hickner, ACS Macro Lett., 2013, 2, 49.
23. T. Sata, M. Tsujimoto, T. Yamaguchi and K. Matsusaki, J. Membr. Sci., 1996, 112, 161.
24. (a) L. Liu, Q. Li, J. Dai, H. Wang, B. Jin and R. Bai, J. Membr. Sci., 2014, 453, 52; (b) D. S. Kim, A. Labouriau, M. D. Guiver and Y. S. Kim, Chem. Mater., 2011, 23(17), 3795; (c) J. Wang, S. Li and S. Zhang, Macromolecules, 2010, 43, 3890; (d) X. Lin, L. Wu, Y. Liu, A. Ong, S. D. Poynton, J. R. Varcoe and T. Xu, J. Power Sources, 2012, 217, 373.
25. (a) B. Qiu, B. Lin, L. Qiu and F. Yan, J. Mater. Chem., 2012, 22, 1040; (b) X. Yan, S. Gu, G. He, X. Wu and J. Benzinger, J. Power Sources, 2014, 250, 90; (c) B. Lin, H. Dong, Y. Li, Z. Si, F. Gu and F. Yan, Chem. Mater., 2013, 25(9), 1858; (d) K. M. Hugar, H. A. Kostalik IV and G. W. Coates, J. Am. Chem. Soc., 2015, 137(27), 8730.
26. (a) M. G. Marino and K. D. Kreuer, ChemSusChem, 2015, 8, 513; (b) A. Katzfuß, S. Polynton, J. Varcoe, V. Gogel, U. Storr and J. Kerres, J. Membr. Sci., 2014, 465, 129; (c) A. Katzfuß, V. Gogel, L. Jörissen and J. Kerres, J. Membr. Sci., 2013, 425–426, 131.
27. (a) J. Miyake, K. Fukasawa, M. Watanabe and K. Miyatake, J. Polym. Sci., Part A: Polym. Chem., 2014, 52, 383; (b) A. Vöge, V. Deimede and J. K. Kallitsis, RSC Adv., 2014, 4, 45040.
28. (a) S. Gu, R. Cai and Y. Yan, Chem. Commun., 2011, 47, 2856; (b) S. Gu, R. Cai, T. Luo, Z. W. Chen, M. W. Sun, Y. Liu, G. H. He and Y. S. Yan, Angew. Chem., Int. Ed., 2009, 48, 6499; (c) S. Gu, R. Cai, T. Luo, K. Jensen, C. Conbtreras and Y. Yan, ChemSusChem, 2010, 3, 555; (d) C. G. Arges, J. Parrondo, G. Johnson, A. Nadhan and V. Ramani, J. Mater. Chem., 2012, 22(9), 3733; (e) X. Yan, S. Gu, G. He, X. Wu, W. Zheng and X. Ruan, J. Membr. Sci., 2014, 466, 220; (f) C. Jangu and T. E. Long, Polymer, 2014, 55, 3298; (g) Y. Ye, K. K. Stokes, F. L. Beyer and Y. A. Elabd, J. Membr. Sci., 2013, 443, 93.
29. C. G. Arges and V. Ramani, Proc. Natl. Acad. Sci. U. S. A., 2013, 110, 2490.
30. Cy. H. Fujimoto, D.-S. Kim, M. Hibbs, D. Wrobleski and Y. S. Kim, J. Membr. Sci., 2012, 423–424, 438.
31. L. Zhu, J. Pan, C. M. Christensen, B. Lin and M. A. Hickner, Macromolecules, 2016, 49, 3300.
32. H. S. Dang, E. A. Weiber and P. Jannasch, J. Mater. Chem. A, 2015, 3, 5280.
33. M. R. Hibbs, J. Polym. Sci., Part B: Polym. Phys., 2013, 51, 1736.
34. (a) G. Nie, X. Li, J. Tao, W. Wu and S. Liao, J. Membr. Sci., 2015, 474, 187; (b) X. Li, G. Nie, J. Tao, W. Wu, L. Wang and S. Liao, ACS Appl. Mater. Interfaces, 2014, 6, 7585; (c) A. D. Mohanty, Y.-B. Lee, L. Zhu, M. A. Hickner and C. Bae, Macromolecules, 2014, 47, 1973.
35. L. Zhu, T. J. Zimudzi, N. Li, J. Pan, B. Lin and M. A. Hickner, Polym. Chem., 2016, 7, 2464.
36. J. Ni, C. Zhao, Y. Zhang, J. Wang, W. Ma, Z. Liu and H. Na, Chem. Commun., 2011, 47, 8943.
37. A. Jasti and V. K. Shahi, RSC Adv., 2014, 4, 19238.
38. J. Pan, Y. Li, L. Zhuang and J. Lu, Chem. Commun., 2010, 46, 8597.
39. M. W. Li, L. Z. Wang and M. Hickner, Chem. Commun., 2014, 50, 4092.
40. L. Z. Wang and M. A. Hickner, Polym. Chem., 2014, 5, 2928.
41. C. Morfopoulou, A. K. Andreopoulou and J. K. Kallitsis, J. Polym. Sci., Part A: Polym. Chem., 2011, 49, 4325.
42. A. Fkyerat, G. M. Dubin and R. Tabacchi, Helv. Chim. Acta, 1999, 82, 1418.
43. A. K. Andreopoulou and J. K. Kallitsis, Macromolecules, 2002, 35, 5808.
44. D. R. Coulson, Inorg. Synth., 1972, 13, 121.
45. J. Yan and M. Hickner, Macromolecules, 2010, 43, 2349.
46. Y. Z. Zhuo, A. N. Lai, Q. G. Zhang, A. M. Zhu, M. L. Ye and Q. L. Liu, J. Membr. Sci., 2015, 491, 138.
47. X. Lin, K. Wang, Y. Feng, J. Z. Liu, X. Fang, T. Xu and H. Wang, J. Membr. Sci., 2015, 482, 67.
48. S. T. Hemp, M. Zhang, M. Tamani and T. E. Long, Polym. Chem., 2013, 4, 3582.
49. Y.-H. Liu, L.-H. Chen and M. Shi, Adv. Synth. Catal., 2006, 348, 973.
50. M. Shi, L.-H. Chen and C.-Q. Li, J. Am. Chem. Soc., 2005, 127, 3790.
51. Z. Wang, Comprehensive Organic Name Reactions and Reagents, Wiley-Interscience, John Wiley & Sons, Hoboken, 1st edn, 2009.
52. W. E. McEwen, C. A. VanderWerf and M. Zander, J. Am. Chem. Soc., 1959, 81, 3806.
53. K. R. Dunbar and S. C. Haefner, Polyhedron, 1994, 13, 727.
54. M. M. Kayser, K. L. Hatt and D. L. Hooper, Can. J. Chem., 1991, 69, 1929.
55. B. Zhang, R. B. Kaspar, S. Gu, J. Wang, Z. Zhaung and Y. Yan, ChemSusChem, 2016, 9, 2374.
56. Z. Liu, X. Zhu, G. Wang, X. Hou and D. Liu, J. Polym. Sci., Part B: Polym. Phys., 2013, 51, 1632.
57. Y. Xiong, Q. L. Liu, Q. G. Zhang and A. M. Zhu, J. Power Sources, 2008, 183, 447.
58. M. Zhang, J. Liu, Y. Wang, L. An, M. D. Guiver and N. Li, J. Mater. Chem. A, 2015, 3, 12284.
59. J. F. Zhou, M. Unlu, I. Anestis-Richard and P. A. Kohl, J. Membr. Sci., 2010, 350, 286.
60. S. Xu, G. Zhang, Y. Zhang, C. Zhao, L. Zhang, M. Li, J. Wang, N. Zhang and H. Na, J. Mater. Chem., 2012, 22, 13295.
61. A. N. Lai, L. S. Wang, C. X. Lin, Y. Z. Zhuo, Q. G. Zhang, A. M. Zhu and Q. L. Liu, J. Membr. Sci., 2015, 481, 9.

---

Footnote

† Electronic supplementary information (ESI) available: Evaluation of AAEM stability under alkaline conditions using ¹H NMR, ³¹P NMR, ATR-FT-IR; percentage of remaining QP groups after alkaline stability tests calculated by ¹H NMR. See DOI: 10.1039/c6ra24102f

---