- Email: [email protected]
Recyclable cross-linked anion exchange membrane for alkaline fuel cell application
Journal of Power Sources xxx (2017) 1e8
Contents lists available at ScienceDirect
Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour
Recyclable cross-linked anion exchange membrane for alkaline fuel cell application Jianqiu Hou, Yazhi Liu, Qianqian Ge, Zhengjin Yang*, Liang Wu, Tongwen Xu** CAS Key Laboratory of Soft Matter Chemistry, Collaborative Innovation Center of Chemistry for Energy Materials, School of Chemistry and Material Science, University of Science and Technology of China, Hefei 230026, PR China
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
A recyclable cross-linked AEM was prepared by mild one-pot reaction. Exploiting disulfide chemistry, the cross-linked membrane can be reprocessed. The repeatedly recycled membrane exhibits acceptable conductivity.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 8 May 2017 Received in revised form 21 June 2017 Accepted 22 June 2017 Available online xxx
Cross-linking can effectively solve the conductivity-swelling dilemma in anion exchange membranes (AEMs) but will generate solid wastes. To address this, we developed an AEM cross-linked via disulfide bonds, bearing quaternary ammonium groups, which can be easily recycled. The membrane (RC-QPPO) with IEC of 1.78 mmol g1, when cross-linked, showed enhanced mechanical properties and good hydroxide conductivity (24.6 mS cm1 at 30 C). Even at higher IEC value (2.13 mmol g1), it still has low water uptake, low swelling ratio and delivers a peak power density of 150 mW cm2 at 65 C. Exploiting the formation of disulfide bonds from -SH groups, the membrane can be readily cross-linked in alkaline condition and recycled by reversibly breaking disulfide bonds with dithiothreitol (DTT). The recycled membrane solution can be directly utilized to cast a brand-new AEM. By washing away the residual DTT with water and exposure to air, it can be cross-linked again and this process is repeatable. During the recycling and cross-linking processes, the membrane showed a slight IEC decrease of 5% due to functional group degradation. The strategy presented here is promising in enhancing AEM properties and reducing the impact of artificial polymers on the environment. © 2017 Elsevier B.V. All rights reserved.
Keywords: Membrane recycling Anion exchange membrane Reversible cross-linking Disulfide linkage Fuel cell
1. Introduction Fuel cell, which can directly convert chemical energy into electric energy without pollution, has attracted intensive research attention and is believed to be an effective approach in solving the
* Corresponding author. ** Corresponding author. E-mail addresses: [email protected] (Z. Yang), [email protected] (T. Xu).
energy crisis [1]. However, high material cost has impeded the practical application of fuel cell, especially the proton exchange membrane fuel cell (PEMFC), which is operated in acidic electrolyte using noble metal catalyst and Nafion membrane [2]. By switching the acidic media to a basic one, the electrode reaction in anion exchange membrane fuel cell (AEMFC) can be catalyzed by inexpensive catalyst (such as Ag) and that will greatly reduce the cost of fuel cells [3]. Currently, the major challenge in AEMFC is developing stable (both chemically and mechanically), highly conductive and low-cost anion exchange membranes (AEMs). Most AEMs are not
http://dx.doi.org/10.1016/j.jpowsour.2017.06.073 0378-7753/© 2017 Elsevier B.V. All rights reserved.
Please cite this article in press as: J. Hou, et al., Journal of Power Sources (2017), http://dx.doi.org/10.1016/j.jpowsour.2017.06.073
2
J. Hou et al. / Journal of Power Sources xxx (2017) 1e8
stable in alkaline media due to high water swelling, which facilitates the degradation of functional groups [4,5] and the breakdown of polymer backbones [6,7] by allowing easier access of OH to the membrane matrix. Cross-linking, as an effective strategy, is commonly utilized and can be readily carried out by in-situ heat treatment [8] and/or by adding extra cross-linking agents [9] during the preparation of AEMs. Cross-linked AEMs generally deliver better mechanical properties, improved thermal and chemical stability, reduced water swelling and methanol permeability [10]. Typically, three types of interactions i.e. hydrogen bonding [11], ionic bonding [12] and covalent bonding [13] are involved in AEM cross-linking. Among them, cross-linking via covalent bonds is more stable and widely used in the preparation of AEMs [14]. However, AEMs cross-linked via covalent bonds cannot be dissolved anymore and that makes it difficult to recycle or reprocess the disposed AEMs, thus generating solid wastes, especially when fuel cell is widely implemented. In this case, recycling “retired” AEMs is of critical importance. Inspired by the folding/unfolding of protein with the formation/ break of disulfide bonds [15], we seek similar strategy to make recyclable AEMs in this contribution. Disulfide bond, with dissociation energies of 210e270 kJ mol1, can be formed by oxidation of sulfhydryl groups (-SH) with oxygen and broken by reduction with dithiothreitol (DTT) [16]. It has been utilized to design recyclable polymer network with improved mechanical properties. Otsuka et al. demonstrated the efficient reprocessing of cross-linked epoxy resins by breaking disulfide bonds. The disulfide bond cross-linked resin exhibited comparable tensile strength (24 MPa) to other permanently cross-linked resin (29 MPa) and can be completely broken into soluble fragments [17]. However, the -SH groups are too reactive to survive in the homogeneous membrane-casting solution when exposed to oxygen. Protection is therefore necessary to avoid forming insoluble gel. It is found that releasing -SH groups from thioacetate groups by alkaline hydrolysis and subsequently forming disulfide bonds by heating and/or exposing to oxygen is an effective approach in membrane self-crosslinking [18]. In this contribution, we synthesized a PPO-based membrane (AcS-QPPO) bearing quaternary ammonium functional groups and thioacetate groups via mild one-pot reaction. Hydrolyzing the thioacetate groups in aqueous NaOH solution and exposing them to air lead to the formation of cross-linked, insoluble RC-QPPO membrane, which exhibited lower water uptake (37.4 wt% at 30 C), lower water swelling ratio (13.4% at 30 C), enhanced mechanical property, similar hydroxide conductivity (24.6 mS cm1 at 30 C) and comparable alkaline stability in comparison with the traditional QPPO membrane. Moreover, RC-QPPO can be re-dissolved in DTT/DMF solution, and the resulting solution can be further utilized to cast a brand-new AEM. This membrane was then cross-linked again by washing off the residual DTT with water and exposure to oxygen. After two successive recycling and cross-linking processes, acceptable hydroxide conductivity (20.0 mS cm1 at 30 C) was observed for membrane RC2-QPPO. The reversible crosslinking strategy demonstrated here can be widely adopted to prepare recyclable cross-linked AEMs.
sulfate, sodium chloride, sodium hydroxide, sodium bicarbonate, 1methyl-2-pyrrolidinone (NMP), dimethyl formamide (DMF), ethanol and trimethylamine alcohol solution (TMA) were purchased from Sinopharm Chemical Reagent Co. Ltd. (Shanghai, P. R. China). Potassium thioacetate and dithiothreitol (DTT) were purchased from Energy Chemical (Shanghai, P. R. China). All reagents are of analytical grade and used as received. 2.2. QPPO membrane BPPO (0.25 g) was dissolved in NMP (5 mL). Trimethylamine alcoholic solution (33 wt%) was then slowly added under vigorous stirring (the amount of trimethylamine might be varied to adjust the ion exchange capacity). The resulting solution was stirred at room temperature for 6 h and then cast in a 60 mm-diameter glass petri dish. After solvent evaporation, QPPO membrane in bromide form was obtained and transformed into OH form by immersing in aqueous NaOH solution (1 mol L1). Finally, the membrane samples (QPPO-n, n indicates the IEC value, mmol g1) were thoroughly washed with deionized water to remove residual NaOH prior to characterization. 2.3. Reversibly cross-linked membrane (RC-QPPO) BPPO (0.25 g) was dissolved in NMP (5 mL). Potassium thioacetate (7 mg, 0.1 equiv. to the benzyl bromide groups in BPPO) was added and stirred at room temperature for 2 h. Trimethylamine alcoholic solution (33 wt%) was then slowly added (the amount of trimethylamine might be varied to adjust the ion exchange capacity) and the resulting solution was stirred at room temperature for 6 h. Crude polymer, namely AcS-QPPO, was obtained by precipitation in acetone (40 mL) and centrifugation. It was then washed with aqueous HCl solution (pH ¼ 3) three times and finally dried in vacuum oven at 30 C overnight. The structure of AcS-QPPO was confirmed by 1H NMR spectroscopy. Similarly, dissolving AcSQPPO polymer in DMF (10 wt%) and subsequent solution casting lead to AcS-QPPO membrane. Its cross-linking was conducted by immersion in aqueous NaOH solution (1 mol L1) and exposure to air. These disulfide bond cross-linked AEMs were designated as RCQPPO-n (n indicates the IEC value, mmol g1) and transformed into Cl or HCO 3 form by immersing in aqueous NaCl or NaHCO3 solution (1 mol L1) before characterization. 2.4. RC-QPPO recycling and recasting RC-QPPO membrane samples (25 mg) were cut into pieces and suspended in 5 mL DMF containing DTT (50 mg, 20 equiv. to the disulfide bonds in the membrane). The mixture was stirred at room temperature until the samples are completely dissolved. Similar solution casting procedure was used to prepare AEMs from the recycled membrane solution and the membranes were then crosslinked by washing off the residual DTT and exposing to air. These recycled membranes were designated as RCn-QPPO (n implies the recycling times).
2. Experimental
2.5. Membrane characterization
2.1. Materials
2.5.1. Nuclear magnetic resonance (NMR) 1 H NMR spectra were conducted on a Bruker Avance III 400 MHz spectrometer, using CDCl3 or DMSO-D6 as solvent.
Brominated poly (2, 6-dimethyl-1, 4-phenylene oxide) (BPPO) was kindly provided by Tianwei Membrane Corporation Ltd. (Shandong, P.R. China). It was purified by dissolving in 1-methyl-2pyrrolidone (NMP), precipitating in deionized water, washing with ethanol and finally drying in vacuum oven at 30 C for 48 h 1H NMR indicates 48 mol% benzyl bromide per repeating unit. Sodium
2.5.2. Gel permeation chromatography (GPC) Molecular weight (Mn, Mw) and polymer dispersity index (PDI ¼ Mw/Mn) of the original (RC-PPO) and recycled (RC1-PPO) polymer without being quaternised was investigated by gel
Please cite this article in press as: J. Hou, et al., Journal of Power Sources (2017), http://dx.doi.org/10.1016/j.jpowsour.2017.06.073
J. Hou et al. / Journal of Power Sources xxx (2017) 1e8
permeation chromatography (GPC) using a PL 120 Plus (Agilent Technologies co., Ltd, China) equipped with differential refractive index detector. The PL Gel Mixed Carbon 18 SEC columns connected in series were used. Samples were freshly prepared in THF (HPLC) and HPLC grade THF was used as the eluent at a flow rate of 1.0 mL/ min. The detection system was calibrated with a 2 mg/mL polystyrene standard (Mw ¼ 110 K). 2.5.3. Water uptake and swelling ratio Membrane samples (1 cm 4 cm, OH form) were immersed in deionized water for 24 h. They were then taken out and water on the surface was quickly wiped with tissue paper. The weight and length of wet membrane samples were recorded. The weight and length of dry membrane samples were also recorded after drying the membrane samples at 60 C for 24 h. Water uptake (WU) of membrane samples was calculated by eq. (1):
WU ¼
Wwet Wdry 100% Wdry
(1)
where Wwet is the weight of wet membrane sample and Wdry is the weight of dry membrane sample. Swelling ratio (SR) of membrane samples was characterized by linear expansion ratio, which was calculated by eq. (2):
SR ¼
Lwet Ldry 100% Ldry
(2)
where Lwet is the length of wet membrane sample and Ldry is the length of dry membrane sample. To evaluate the temperature-dependent water swelling, QPPO and RC-QPPO membrane samples were immersed in water with predetermined-temperature for 4 h. The WU and SR of these membrane samples were recorded and plotted versus temperature. 2.5.4. Ion exchange capacity (IEC) IEC of the membrane samples was measured by the Mohr's method. Membrane samples in Cl form were immersed in aqueous Na2SO4 solution (0.5 mol L1) for 8 h. After removing membrane samples, Cl content in the solution was titrated with aqueous AgNO3 solution (0.01 mol L1) and K2CrO4 was used as the indicator. The IEC values were calculated by eq. (3):
IEC ¼
V C Wdry
(3)
where V is the consumed volume of AgNO3 in titration, C is the concentration of aqueous AgNO3 solution (0.01 mol L1) and Wdry is the weight of dry membrane sample in Cl form. 2.5.5. Conductivity measurement Ion conductivity of membrane samples was measured by the four-point probe technique (Zahner Zemmiun E) with potentiostatic mode (amplitude 50 mV, frequency range 10 KHz-100 Hz). The membrane sample (1 cm 4 cm, OH, Cl or HCO 3 form) was placed in a Teflon cell which has two current collecting electrodes and two potential sensing electrodes (the distance between couple electrodes is 1 cm). The membrane was completely immersed in deionized water. The ionic conductivity was calculated according to the following equation:
k¼
L RWd
(4)
where R is the membrane resistance, L is the distance between
3
potential sensing electrodes. W and d are the width (here 1 cm) and thickness of the membrane, respectively. Before measurement, the samples were equilibrated at least 30 min at predetermined temperature. 2.5.6. Mechanical analysis Tensile strength measurements of hydrated and dry membrane samples were performed on a Q800 dynamic mechanical analyzer (DMA, TA Instruments) at a stretching rate of 0.5 N min1. 2.5.7. Alkaline stability To evaluate the chemical stability, QPPO and RC-QPPO membrane samples were immersed in aqueous NaOH solution (2 mol L1) at 60 C for 240 h. The hydroxide conductivity of these membranes was monitored at 30 C during this period and was plotted versus time. 2.5.8. H2/O2 fuel cell measurements Pt/C or PtRu/C catalysts (60 wt% in metal content) were mixed with the QPPO/isopropanol solution under ultrasonic to yield inks (20 wt% of polymer and 80 wt% of catalyst). The membrane electrode assemblies (MEAs) were fabricated using the catalyst-coatedmembrane (CCM) method: the PtRu anode ink and the Pt cathode ink were then sprayed on the correct sides of the membrane samples (with a thickness of 50 mme70 mm when fully hydrated). The metal loading in anode or cathode was controlled to be 0.5 mg/ cm2, and the electrode area was 12.25 cm2. The prepared CCM was then converted to OH form by immersing in aqueous NaOH (1 mol L1) solution for 12 h followed by thorough washing with deionized water. The obtained CCM was placed between two pieces of carbon paper (Toray TGP-H-060) to make the MEA without hotpressing. Single cell was tested using an 850E Multi Range fuel cell test station (Scribner Associates, USA) in a galvanic mode at 65 C. H2 and O2 were humidified at 65 C (100% RH) and fed with a flow rate of 1 L min1 and a backpressure of 0 MPa was kept on both sides. The cell voltage at each current density was recorded. 3. Results and discussion 3.1. AcS-QPPO synthesis AcS-QPPO bearing quaternary ammonium and thioacetate groups was prepared in a mild one-pot reaction, as outlined in Fig. 1. Although the conversion ratio of benzyl bromide into benzyl thioacetate shall be controlled to leave enough benzyl bromide for quaternization, as a proof of concept we tried to convert benzyl bromide into equal amount of quaternary ammonium and thioacetate in order for them to have obvious signals in the 1H NMR spectrum (Fig. 2). This quantified conversion can be conducted by adding equal amount of trimethylamine and potassium thioacetate (0.5 equiv. with respect to the molar content of benzyl bromide) while synthesizing AcS-QPPO. The peak at 3.85 ppm (H4 in Fig. 2b) belongs to the methylene protons in benzyl thioacetate group and the peak of methyl protons in thioacetate group appears at 2.20 ppm. As benzyl thioacetate was attached, the peak intensity of methyl protons has been changed and compared with the starting polymer (BPPO, Fig. 2a), the ratio of two different methyl protons, (H2þ H6)/H7 was greatly increased. As shown in Fig. 2b, a new peak at 3.05 ppm is observed, which belongs to the -CH3 group attached to the positive-charge quaternary ammonium group. As benzyl bromide (-CH2Br) reacted with trimethylamine, -CH2- peak shifted from 4.32 ppm to 4.43 ppm. It indicates the successful quaternization of AcS-QPPO polymer. Additionally, the integration ratio of peaks at 4.43 ppm to that at 3.85 ppm is close to 1:1, which is expected since the equal
Please cite this article in press as: J. Hou, et al., Journal of Power Sources (2017), http://dx.doi.org/10.1016/j.jpowsour.2017.06.073
4
J. Hou et al. / Journal of Power Sources xxx (2017) 1e8
Fig. 1. Preparation of recyclable AEM, RC-QPPO. The reaction starts from commercial polymer, brominated poly (2, 6-dimethyl-1, 4-phenylene oxide) (BPPO), in one pot. The obtained AcS-QPPO membrane can be cross-linked under alkaline condition to form RC-QPPO membrane. The disulfide bonds in RC-QPPO membrane can be reduced and broken by DTT, leading to the recycling of RC-QPPO.
Fig. 2. 1H NMR spectra of (a) BPPO (in CDCl3) and (b) AcS-QPPO (in DMSO-D6).
amount of trimethylamine and potassium thioacetate was added. It implies that by adjusting the amount of trimethylamine and potassium thioacetate added, the ratio of quaternization, i.e. IEC, and the number of potential cross-linking sites can be readily tuned. 3.2. AcS-QPPO membrane cross-linking and recycling We firstly dissolved AcS-QPPO in DMF and fabricated AcS-QPPO membrane by solution casting, as described in section 2.3. The obtained AcS-QPPO membrane can be repeatedly dissolved in DMF, implying a linear polymer structure in the membrane (Fig. 3a). As we immersed the membrane in aqueous NaOH solution (1 mol L1), the thioacetate groups will be hydrolyzed into the sulfhydryl groups (-SH), which can be quickly coupled to form disulfide groups when exposed to air. Therefore, when alkali treated, AcS-QPPO membrane will be gradually cross-linked and cannot be dissolved in DMF anymore, corresponding to a network polymer structure (Fig. 3b). By taking advantage of the reversible formation and break of disulfide groups, we were able to reprocess and recycle the cross-
linked AcS-QPPO membrane (namely, RC-QPPO). Treating RC-QPPO with DTT can break these disulfide bonds, leading to re-dissolving of RC-QPPO in DMF (Fig. 3c). The resulting solution can be utilized to cast a brand-new AEM, which can be cross-linked after removing the residual DTT (washed off with water), as described in section 2.4. The entire process can be carried out repeatedly and two repeated cross-linking and recycling processes were performed in the lab. It should be noted that the crosslinking degree has a great impact on the recycling process. We initially prepared a cross-linked membrane by adding 10 mol% of potassium thioacetate with respect to the molar content of benzyl bromide in the starting BPPO. The resulting RC-QPPO membrane can be completely re-dissolved and re-cast. However, as we increased the degree of cross-linking by adding more reactive -SH groups, it was slightly challenging to completely recycle all the starting polymers. This is possibly caused by the incomplete breaking of disulfide bonds due to restricted diffusion of DTT into the membrane matrix. We believe that adding more DTT, extending dissolution time and gently heating the solution would enable us to
Please cite this article in press as: J. Hou, et al., Journal of Power Sources (2017), http://dx.doi.org/10.1016/j.jpowsour.2017.06.073
J. Hou et al. / Journal of Power Sources xxx (2017) 1e8
5
Fig. 3. Images demonstrating the membrane (a) can be dissolved in DMF before cross-linking; (b) cannot be dissolved in DMF after alkaline treatment and (c) can be recycled when treated with dithiothreitol (DTT).
access 100% recycling. To demonstrate the effectiveness of our strategy, RC-QPPO membrane with a cross-linking degree of 10% was presented and thoroughly characterized (see following sections for property discussions).
3.3. Hydroxide conductivity of RC-QPPO We have prepared a QPPO membrane (as control) and series of cross-linked AcS-QPPO membranes (RC-QPPO) with different IEC by adding different amount of trimethylamine in AcS-QPPO synthesis, as described in section 2.2 and 2.3. Fig. 4 shows the hydroxide conductivity of QPPO-1.78, RC-QPPO-1.27, RC-QPPO-1.78 and RC-QPPO-2.13 as a function of operating temperature. As the operating temperature increases, the hydroxide conductivity increases because of enhanced OH mobility. For instance, as temperature increases from 30 C to 80 C, the hydroxide conductivity of RC-QPPO-1.27 increases from 13.3 mS cm1 to 29.7 mS cm1 and an increase from 24.6 mS cm1 to 64.5 mS cm1 is observed for RCQPPO-1.78. For membrane with a higher IEC, i.e. RC-QPPO-2.13, the conductivity stretched from 38.0 mS cm1 at 30 C to 69.6 mS cm1 at 80 C. However, when the operating temperature was increased to 80 C, a severe deformation of RC-QPPO-2.13 was observed due to the excessive membrane swelling. Additionally, the increase in hydroxide conductivity with an increase in IEC is expected since more exchangeable functional groups are available inside the membrane matrix. Comparing RCQPPO-1.78 membrane with QPPO-1.78 membrane that has the
Fig. 4. Hydroxide conductivity of QPPO-1.78, RC-QPPO-1.27, RC-QPPO-1.78 and RCQPPO-2.13 as a function of temperature.
same IEC, we did not observe significant difference in hydroxide conductivity, indicating that the cross-linking process has no obvious influence on membrane conductivity. These conductivity results suggest that RC-QPPO membranes show sufficient hydroxide conductivity (>10 mS cm1 at 30 C [19]) for a practical application in fuel cell. The conductivity of RC-QPPO membrane also depends on the type of counter anions. Fig. 5 presents the OH, Cl and HCO 3 conductivity of RC-QPPO (IEC ¼ 1.78 mmol/g) as a function of operating temperature. RC-QPPO membrane in OH form shows the highest conductivity, while in HCO 3 form it exhibits the lowest conductivity over the full operating temperature window because of the intrinsic difference in ion mobility [20]. As the operating temperature is elevated from 30 C to 80 C, the Cl conductivity increases from 11.4 mS cm1 to 30.5 mS cm1 and the HCO 3 conductivity climbs from 5.0 mS cm1 to 11.9 mS cm1. Lower ionic conductivity of RC-QPPO membrane in HCO 3 form implies the importance of CO2 removal in alkaline fuel cells. 3.4. Water swelling Water uptake and swelling ratio of QPPO and RC-QPPO with different IECs are measured and presented in Table 1. We observed that the water swelling of either membrane is greatly influenced by the content of functional groups, i.e. the IEC values. The RC-QPPO membrane with an IEC of 1.27 mmol g1 had a water uptake of 21.6 wt% and swelling ratio of 7.8%, while RC-QPPO membrane with
Fig. 5. The conductivity of RC-QPPO membrane with different counter ions (OH, Cl and HCO 3 ) as a function of temperature.
Please cite this article in press as: J. Hou, et al., Journal of Power Sources (2017), http://dx.doi.org/10.1016/j.jpowsour.2017.06.073
6
J. Hou et al. / Journal of Power Sources xxx (2017) 1e8
Table 1 IEC value, water uptake and swelling ratio of QPPO-1.78, QPPO-2.15, RC-QPPO-1.27, RC-QPPO-1.78 and RC-QPPO-2.13 at 30 C.
Table 2 Yielding stress and elongation at break of QPPO-1.78 and RC-QPPO-1.78 in both dry and wet state.
Membrane
IEC (mmol/g)
Water uptake (wt%)
Swelling ratio (%)
Membrane
QPPOdry
QPPOwet
RC-QPPOdry
RC-QPPOwet
QPPO-1.78 QPPO-2.15 RC-QPPO-1.27 RC-QPPO-1.78 RC-QPPO-2.13
1.78 2.15 1.27 1.78 2.13
41.3 163.1 21.6 37.4 70.0
15.4 36.5 7.8 13.4 21.8
Yielding stress (MPa) Elongation at break (%)
12.95 5.21
12.31 6.46
19.49 6.73
23.14 17.09
higher IEC (2.13 mmol/g) exhibited higher water uptake (70.0 wt%) and higher swelling ratio (21.8%). It is not surprising since the functional groups are hydrophilic and more water will be absorbed in membranes with higher content of functional groups, i.e. higher IEC. As we expected, RC-QPPO membranes, which are cross-linked by disulfide bonds, generally show less water uptake and increased swelling resistance compared with the linear counterpart, i.e. the QPPO membrane. For the QPPO-1.78 membrane, its water uptake and swelling ratio at 30 C are 41.3 wt% and 15.4%, respectively. However, for the cross-linked membrane with the same IEC, i.e. RC-QPPO-1.78 shows a decrease of 9.4% water uptake and a decrease of 13.0% swelling ratio. When the IEC of RC-QPPO membrane is further increased to 2.13 mmol g1, much more decrease in water uptake (57.1%) and swelling ratio (40.2%) are observed in comparison with QPPO-2.15 that has the similar IEC (2.15 mmol g1, Table 1). It is an obvious conclusion that disulfide bond cross-linking can enhance the water swelling capability of RCQPPO membranes, especially at higher IEC (>2.0 mmol g1). To facilitate a practical application, which usually involves different operating temperature, the swelling of RC-QPPO membranes are investigated at varied temperatures (Fig. 6). The water uptake of RC-QPPO-1.78 at 30 C is 37.4 wt%, which is increased to 59.7 wt% at 80 C, while the swelling ratio of RC-QPPO-1.78 is slightly increased over the temperature range 30 Ce80 C. Concerning the effect of elevated temperature on membrane water uptake and swelling ratio, we conclude that the reversibly crosslinked RC-QPPO shows good swelling resistance at elevated operating temperature in spite of the increased water content.
3.5. Mechanical properties The mechanical properties of cross-linked RC-QPPO (RC-QPPO-
Fig. 6. Water uptake and swelling ratio of RC-QPPO-1.78 as a function of temperature.
1.78) membrane and its linear counterpart QPPO (QPPO-1.78), with the same IEC, are investigated in both dry and wet (hydrated) state to illustrate the impact of cross-linking on membrane robustness. The results are presented in Table 2. The yielding stress of QPPOdry, QPPOwet, RC-QPPOdry and RC-QPPOwet are 12.95 MPa, 12.31 MPa, 19.49 MPa and 23.14 MPa, respectively. It is obvious that RC-QPPO1.78 exhibits enhanced robustness in comparison with QPPO-1.78 due to disulfide bond cross-linking. The elongations at break of QPPOdry, QPPOwet, RC-QPPOdry and RC-QPPOwet are 5.21%, 6.46%, 6.73% and 17.09%, respectively. Surprisingly, the elongation at break of wet RC-QPPO-1.78 is much higher than that of the rest, indicating a “significantly flexible” polymer network, which is not very common for a cross-linked membrane. It is attributed to the special nature of disulfide bonds, which is in dynamic equilibrium and can self-exchange in hydrated polymer network [21]. The results suggest that the RC-QPPO membranes show appropriate mechanical strength for a practical AEMFC.
3.6. RC-QPPO recycling and reusing The most important characteristic of RC-QPPO membrane is its capability to be recycled. Therefore, we sequentially repeated the cross-linking and recycling process twice to see the impact of such process on membrane properties. As a demonstration, these experiments and characterizations were conducted on RC-QPPO-2.13 membrane, which is the most promising candidate for alkaline fuel cell application according to conductivity measurements. The membranes recovered and recast for the first time and the second time were designated as RC1-QPPO and RC2-QPPO. To examine the impact of the recycling process on membrane properties, we firstly evaluated the IEC of both RC1-QPPO and RC2-QPPO and found that the IEC of the recycled membranes are similar, i.e. 1.99 mmol g1 for RC1-QPPO and 2.04 mmol g1 for RC2-QPPO. Compared with the pristine RC-QPPO membrane, the recycled membranes, both RC1QPPO and RC2-QPPO, show slightly lower IEC. It could possibly be ascribed to the degradation of functional groups in the process of recycling. We anticipate that with more stable functional groups, this could be mitigated. Still, the recycled membrane exhibited proper conductivities for fuel cell application. For example, RC1-QPPO and RC2-QPPO show hydroxide conductivity of 18.1 mS cm1 and 20.0 mS cm1 at 30 C, respectively and it lies in the acceptable window for a fuel cell application. It should also be mentioned that as the operating temperature increases, the conductivity can be further enhanced (42.0 mS cm1 at 80 C for RC1-QPPO and 39.5 mS cm1 at 80 C for RC2-QPPO), projecting even better fuel cell performance (Fig. 7). To further understand the recycling process, gel permeation chromatography (GPC) measurements were conducted. Because GPC measurements do not allow polymers with quaternised ammonium groups in case of blocking the column, the measurements were performed on PPO polymers without being quaternised. Results (Table 3) show that the recycled polymer (RC1-PPO) has similar molecular weight (Mn ¼ 31302 g/mol) as that of the pristine one (RC-PPO) (Mn ¼ 24514 g/mol), while the molecular weight distribution of RC1-PPO (PDI ¼ 4.03) is higher than that of the pristine RC-PPO (PDI ¼ 2.34). We presume that was due to
Please cite this article in press as: J. Hou, et al., Journal of Power Sources (2017), http://dx.doi.org/10.1016/j.jpowsour.2017.06.073
J. Hou et al. / Journal of Power Sources xxx (2017) 1e8
7
Fig. 9. H2/O2 single cell performance of RC-QPPO-2.13 and RC1-QPPO at 65 C. Fig. 7. Hydroxide conductivity of RC-QPPO, RC1-QPPO and RC2-QPPO as a function of temperature.
Table 3 GPC results of RC-PPO and RC1-PPO. HPLC grade THF was used as the eluent. Membrane
Mn (g/mol)
PDI
RC-PPO RC1-PPO
24514 31302
2.34 4.03
QPPO-2.13 membrane in 2 mol L1 NaOH at 60 C for 240 h, the IEC was decreased by 31% (from 2.13 mmol g1 to 1.48 mmol g1). This can be solved by incorporating more stable functional groups into the membrane, e.g. the imidazolium [22], phosphonium [23] and so on. The obtained results tend to suggest that the crosslinking strategy does not have much interference with the membrane alkaline stability.
3.8. H2/O2 fuel cell performance several residual disulfide bonds after recycling, which may not prevent the polymer from being soluble but increase the PDI of samples. 3.7. Alkaline stability Ex-situ alkaline stability tests are performed to evaluate the effect of introducing disulfide cross-linkage on membrane alkaline stability. Fig. 8 presents the hydroxide conductivity of QPPO-2.13 and RC-QPPO-2.13 as a function of time during alkaline treatment. The hydroxide conductivity decreased by 20% over the first 60 h and was slightly changed over the next 130 h. After 240 h of alkaline treatment, 65% of the conductivity was maintained. Similar behavior was observed for both QPPO-2.13 and RC-QPPO-2.13 since they have the same kind of functional groups (quaternary ammonium groups). The decline in hydroxide conductivity was due to the degradation of quaternary ammonium groups, which has been rigorously proved by IEC measurements. After immersing RC-
The original membrane (RC-QPPO-2.13, IEC ¼ 2.13 mmol g1) and the corresponding recycled membrane (RC1-QPPO, IEC ¼ 1.99 mmol g1) were selected to evaluate the effect of recycling process on fuel cell performance (Fig. 9). In a H2/O2 cell operated at 65 C, the peak power densities of RC-QPPO-2.13 and RC1-QPPO are 150 mW cm2 (at 0.6 V with a current density of 265 mA cm2) and 123 mW cm2 (at 0.5 V with a current density of 252 mA cm2), respectively. The slightly decrease in fuel cell performance after recycling is ascribed to the decrease in IEC and the resulting decrease in hydroxide conductivity. The performance we obtained here is comparable to previous studies [3,24], suggesting the potential application of RC-QPPO membrane and the recycling strategy in AEMFC. And we believe that the fuel cell performance of the pristine and the recycled membranes can be further improved by optimizing the fabrication of MEA and by incorporating more stable functional groups.
4. Conclusions
Fig. 8. Changing trends in hydroxide conductivity of QPPO-2.13 and RC-QPPO-2.13 at 30 C after immersing in 2 mol L1 NaOH at 60 C.
A PPO-based membrane (AcS-QPPO) bearing quaternary ammonium and thioacetate groups was successfully synthesized via one-pot reaction at mild condition. The thioacetate groups serve as potential cross-linkage and result in a reversibly cross-linked anion exchange membrane. The obtained cross-linked membrane, RC-QPPO exhibits enhanced mechanical strength, suppressed water swelling and comparable hydroxide conductivity in comparison with the control, QPPO membrane. Disulfide bond cross-linking strategy developed here leads to a cross-linked RC-QPPO membrane that can be repeatedly recycled and recast. After two repeated cross-linking and recycling processes, the membrane still showed acceptable hydroxide conductivity. The acquired facts prove that the strategy to fabricate a cross-linked AEM, which can be recycled, is very promising and could be extended to other types of membranes. Further study will be focused on developing recyclable cross-linked AEM with increased alkaline stability and its application in a practical AEMFC.
Please cite this article in press as: J. Hou, et al., Journal of Power Sources (2017), http://dx.doi.org/10.1016/j.jpowsour.2017.06.073
8
J. Hou et al. / Journal of Power Sources xxx (2017) 1e8
Acknowledgements Financial support received from the National Science Foundation of China (Nos. 21506201, 21490581, 91534203) and Postdoc Foundation Support (No. 2014M560521, 2015T80667) is gratefully acknowledged. We deeply appreciate the help Mr. Xian Liang gave us in investigating the fuel cell performance. References [1] B.C.H. Steele, A. Heinzel, Nature 414 (2001) 345e352. [2] J. Pan, C. Chen, Y. Li, L. Wang, L. Tan, G. Li, X. Tang, L. Xiao, J. Lu, L. Zhuang, Energy Environ. Sci. 7 (2014) 354e360. [3] J. Pan, S. Lu, Y. Li, A. Huang, L. Zhuang, J. Lu, Adv. Funct. Mater 20 (2010) 312e319. [4] X. Dong, B. Xue, H. Qian, J. Zheng, S. Li, S. Zhang, J. Power Sources 342 (2017) 605e615. [5] A.D. Mohanty, C. Bae, J. Mat. Chem. A 2 (2014) 17314e17320. [6] O.D. Thomas, K.J. Soo, T.J. Peckham, M.P. Kulkarni, S. Holdcroft, J. Am. Chem. Soc. 134 (2012) 10753e10756. [7] C.G. Arges, V. Ramani, Proc. Natl. Acad. Sci. 110 (2013) 2490e2495. [8] L. Wu, Q. Pan, J.R. Varcoe, D. Zhou, J. Ran, Z. Yang, T. Xu, J. Membr. Sci. 490 (2015) 1e8. [9] X. Chen, P. Chen, Z. An, K. Chen, K. Okamoto, J. Power Sources 196 (2011) 1694e1703.
[10] H. Han, M. Liu, L. Xu, J. Xu, S. Wang, H. Ni, Z. Wang, J. Membr. Sci. 496 (2015) 84e94. [11] J. Li, W. Cai, L. Ma, Y. Zhang, Z. Chenb, H. Cheng, Chem. Commun. 51 (2015) 6556e6559. [12] Z. Yue, Y.-B. Cai, S. Xu, J. Power Sources 286 (2015) 571e579. [13] W. Xu, Y. Zhao, Z. Yuan, X. Li, H. Zhang, I.F.J. Vankelecom, Adv. Funct. Mater 25 (2015) 2583e2589. [14] L. Zhu, T.J. Zimudzi, N. Li, J. Pan, B. Lin, M.A. Hickner, Polym. Chem. 7 (2016) 2464e2475. [15] J.-H. Ryu, R.T. Chacko, S. Jiwpanich, S. Bickerton, R.P. Babu, S. Thayumanavan, J. Am. Chem. Soc. 132 (2010) 17227e17235. [16] E.-K. Bang, M. Lista, G. Sforazzini, N. Sakai, S. Matile, Chem. Sci. 3 (2012) 1752e1763. [17] A. Takahashi, T. Ohishi, R. Goseki, H. Otsuka, Polymer 82 (2016) 319e326. [18] J. Kim, K. Baek, D. Shetty, N. Selvapalam, G. Yun, N.H. Kim, Y.H. Ko, K.M. Park, I. Hwang, K. Kim, Angew. Chem. Int. Ed. Engl. 54 (2015) 2693e2697. [19] G. Merle, M. Wessling, K. Nijmeijer, J. Membr. Sci. 377 (2011) 1e35. [20] L. Zhu, J. Pan, Y. Wang, J. Han, L. Zhuang, M.A. Hickner, Macromolecules 49 (2016) 815e824. [21] G. Deng, F. Li, H. Yu, F. Liu, C. Liu, W. Sun, H. Jiang, Y. Chen, ACS Macro Lett. 1 (2012) 275e279. [22] X. Lin, J.R. Varcoe, S.D. Poynton, X. Liang, A.L. Ong, J. Ran, Y. Li, T. Xu, J. Mat. Chem. A 1 (2013) 7262e7269. [23] S. Gu, R. Cai, T. Luo, Z. Chen, M. Sun, Y. Liu, G. He, Y. Yan, Angew. Chem. Int. Ed. Engl. 48 (2009) 6499e6502. [24] K.H. Lee, D.H. Cho, Y.M. Kim, S.J. Moon, J.G. Seong, D.W. Shin, J.-Y. Sohn, J.F. Kima, Y.M. Lee, Energy Environ. Sci. 10 (2017) 275e285.
Please cite this article in press as: J. Hou, et al., Journal of Power Sources (2017), http://dx.doi.org/10.1016/j.jpowsour.2017.06.073
Contents lists available at ScienceDirect
Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour
Recyclable cross-linked anion exchange membrane for alkaline fuel cell application Jianqiu Hou, Yazhi Liu, Qianqian Ge, Zhengjin Yang*, Liang Wu, Tongwen Xu** CAS Key Laboratory of Soft Matter Chemistry, Collaborative Innovation Center of Chemistry for Energy Materials, School of Chemistry and Material Science, University of Science and Technology of China, Hefei 230026, PR China
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
A recyclable cross-linked AEM was prepared by mild one-pot reaction. Exploiting disulfide chemistry, the cross-linked membrane can be reprocessed. The repeatedly recycled membrane exhibits acceptable conductivity.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 8 May 2017 Received in revised form 21 June 2017 Accepted 22 June 2017 Available online xxx
Cross-linking can effectively solve the conductivity-swelling dilemma in anion exchange membranes (AEMs) but will generate solid wastes. To address this, we developed an AEM cross-linked via disulfide bonds, bearing quaternary ammonium groups, which can be easily recycled. The membrane (RC-QPPO) with IEC of 1.78 mmol g1, when cross-linked, showed enhanced mechanical properties and good hydroxide conductivity (24.6 mS cm1 at 30 C). Even at higher IEC value (2.13 mmol g1), it still has low water uptake, low swelling ratio and delivers a peak power density of 150 mW cm2 at 65 C. Exploiting the formation of disulfide bonds from -SH groups, the membrane can be readily cross-linked in alkaline condition and recycled by reversibly breaking disulfide bonds with dithiothreitol (DTT). The recycled membrane solution can be directly utilized to cast a brand-new AEM. By washing away the residual DTT with water and exposure to air, it can be cross-linked again and this process is repeatable. During the recycling and cross-linking processes, the membrane showed a slight IEC decrease of 5% due to functional group degradation. The strategy presented here is promising in enhancing AEM properties and reducing the impact of artificial polymers on the environment. © 2017 Elsevier B.V. All rights reserved.
Keywords: Membrane recycling Anion exchange membrane Reversible cross-linking Disulfide linkage Fuel cell
1. Introduction Fuel cell, which can directly convert chemical energy into electric energy without pollution, has attracted intensive research attention and is believed to be an effective approach in solving the
* Corresponding author. ** Corresponding author. E-mail addresses: [email protected] (Z. Yang), [email protected] (T. Xu).
energy crisis [1]. However, high material cost has impeded the practical application of fuel cell, especially the proton exchange membrane fuel cell (PEMFC), which is operated in acidic electrolyte using noble metal catalyst and Nafion membrane [2]. By switching the acidic media to a basic one, the electrode reaction in anion exchange membrane fuel cell (AEMFC) can be catalyzed by inexpensive catalyst (such as Ag) and that will greatly reduce the cost of fuel cells [3]. Currently, the major challenge in AEMFC is developing stable (both chemically and mechanically), highly conductive and low-cost anion exchange membranes (AEMs). Most AEMs are not
http://dx.doi.org/10.1016/j.jpowsour.2017.06.073 0378-7753/© 2017 Elsevier B.V. All rights reserved.
Please cite this article in press as: J. Hou, et al., Journal of Power Sources (2017), http://dx.doi.org/10.1016/j.jpowsour.2017.06.073
2
J. Hou et al. / Journal of Power Sources xxx (2017) 1e8
stable in alkaline media due to high water swelling, which facilitates the degradation of functional groups [4,5] and the breakdown of polymer backbones [6,7] by allowing easier access of OH to the membrane matrix. Cross-linking, as an effective strategy, is commonly utilized and can be readily carried out by in-situ heat treatment [8] and/or by adding extra cross-linking agents [9] during the preparation of AEMs. Cross-linked AEMs generally deliver better mechanical properties, improved thermal and chemical stability, reduced water swelling and methanol permeability [10]. Typically, three types of interactions i.e. hydrogen bonding [11], ionic bonding [12] and covalent bonding [13] are involved in AEM cross-linking. Among them, cross-linking via covalent bonds is more stable and widely used in the preparation of AEMs [14]. However, AEMs cross-linked via covalent bonds cannot be dissolved anymore and that makes it difficult to recycle or reprocess the disposed AEMs, thus generating solid wastes, especially when fuel cell is widely implemented. In this case, recycling “retired” AEMs is of critical importance. Inspired by the folding/unfolding of protein with the formation/ break of disulfide bonds [15], we seek similar strategy to make recyclable AEMs in this contribution. Disulfide bond, with dissociation energies of 210e270 kJ mol1, can be formed by oxidation of sulfhydryl groups (-SH) with oxygen and broken by reduction with dithiothreitol (DTT) [16]. It has been utilized to design recyclable polymer network with improved mechanical properties. Otsuka et al. demonstrated the efficient reprocessing of cross-linked epoxy resins by breaking disulfide bonds. The disulfide bond cross-linked resin exhibited comparable tensile strength (24 MPa) to other permanently cross-linked resin (29 MPa) and can be completely broken into soluble fragments [17]. However, the -SH groups are too reactive to survive in the homogeneous membrane-casting solution when exposed to oxygen. Protection is therefore necessary to avoid forming insoluble gel. It is found that releasing -SH groups from thioacetate groups by alkaline hydrolysis and subsequently forming disulfide bonds by heating and/or exposing to oxygen is an effective approach in membrane self-crosslinking [18]. In this contribution, we synthesized a PPO-based membrane (AcS-QPPO) bearing quaternary ammonium functional groups and thioacetate groups via mild one-pot reaction. Hydrolyzing the thioacetate groups in aqueous NaOH solution and exposing them to air lead to the formation of cross-linked, insoluble RC-QPPO membrane, which exhibited lower water uptake (37.4 wt% at 30 C), lower water swelling ratio (13.4% at 30 C), enhanced mechanical property, similar hydroxide conductivity (24.6 mS cm1 at 30 C) and comparable alkaline stability in comparison with the traditional QPPO membrane. Moreover, RC-QPPO can be re-dissolved in DTT/DMF solution, and the resulting solution can be further utilized to cast a brand-new AEM. This membrane was then cross-linked again by washing off the residual DTT with water and exposure to oxygen. After two successive recycling and cross-linking processes, acceptable hydroxide conductivity (20.0 mS cm1 at 30 C) was observed for membrane RC2-QPPO. The reversible crosslinking strategy demonstrated here can be widely adopted to prepare recyclable cross-linked AEMs.
sulfate, sodium chloride, sodium hydroxide, sodium bicarbonate, 1methyl-2-pyrrolidinone (NMP), dimethyl formamide (DMF), ethanol and trimethylamine alcohol solution (TMA) were purchased from Sinopharm Chemical Reagent Co. Ltd. (Shanghai, P. R. China). Potassium thioacetate and dithiothreitol (DTT) were purchased from Energy Chemical (Shanghai, P. R. China). All reagents are of analytical grade and used as received. 2.2. QPPO membrane BPPO (0.25 g) was dissolved in NMP (5 mL). Trimethylamine alcoholic solution (33 wt%) was then slowly added under vigorous stirring (the amount of trimethylamine might be varied to adjust the ion exchange capacity). The resulting solution was stirred at room temperature for 6 h and then cast in a 60 mm-diameter glass petri dish. After solvent evaporation, QPPO membrane in bromide form was obtained and transformed into OH form by immersing in aqueous NaOH solution (1 mol L1). Finally, the membrane samples (QPPO-n, n indicates the IEC value, mmol g1) were thoroughly washed with deionized water to remove residual NaOH prior to characterization. 2.3. Reversibly cross-linked membrane (RC-QPPO) BPPO (0.25 g) was dissolved in NMP (5 mL). Potassium thioacetate (7 mg, 0.1 equiv. to the benzyl bromide groups in BPPO) was added and stirred at room temperature for 2 h. Trimethylamine alcoholic solution (33 wt%) was then slowly added (the amount of trimethylamine might be varied to adjust the ion exchange capacity) and the resulting solution was stirred at room temperature for 6 h. Crude polymer, namely AcS-QPPO, was obtained by precipitation in acetone (40 mL) and centrifugation. It was then washed with aqueous HCl solution (pH ¼ 3) three times and finally dried in vacuum oven at 30 C overnight. The structure of AcS-QPPO was confirmed by 1H NMR spectroscopy. Similarly, dissolving AcSQPPO polymer in DMF (10 wt%) and subsequent solution casting lead to AcS-QPPO membrane. Its cross-linking was conducted by immersion in aqueous NaOH solution (1 mol L1) and exposure to air. These disulfide bond cross-linked AEMs were designated as RCQPPO-n (n indicates the IEC value, mmol g1) and transformed into Cl or HCO 3 form by immersing in aqueous NaCl or NaHCO3 solution (1 mol L1) before characterization. 2.4. RC-QPPO recycling and recasting RC-QPPO membrane samples (25 mg) were cut into pieces and suspended in 5 mL DMF containing DTT (50 mg, 20 equiv. to the disulfide bonds in the membrane). The mixture was stirred at room temperature until the samples are completely dissolved. Similar solution casting procedure was used to prepare AEMs from the recycled membrane solution and the membranes were then crosslinked by washing off the residual DTT and exposing to air. These recycled membranes were designated as RCn-QPPO (n implies the recycling times).
2. Experimental
2.5. Membrane characterization
2.1. Materials
2.5.1. Nuclear magnetic resonance (NMR) 1 H NMR spectra were conducted on a Bruker Avance III 400 MHz spectrometer, using CDCl3 or DMSO-D6 as solvent.
Brominated poly (2, 6-dimethyl-1, 4-phenylene oxide) (BPPO) was kindly provided by Tianwei Membrane Corporation Ltd. (Shandong, P.R. China). It was purified by dissolving in 1-methyl-2pyrrolidone (NMP), precipitating in deionized water, washing with ethanol and finally drying in vacuum oven at 30 C for 48 h 1H NMR indicates 48 mol% benzyl bromide per repeating unit. Sodium
2.5.2. Gel permeation chromatography (GPC) Molecular weight (Mn, Mw) and polymer dispersity index (PDI ¼ Mw/Mn) of the original (RC-PPO) and recycled (RC1-PPO) polymer without being quaternised was investigated by gel
Please cite this article in press as: J. Hou, et al., Journal of Power Sources (2017), http://dx.doi.org/10.1016/j.jpowsour.2017.06.073
J. Hou et al. / Journal of Power Sources xxx (2017) 1e8
permeation chromatography (GPC) using a PL 120 Plus (Agilent Technologies co., Ltd, China) equipped with differential refractive index detector. The PL Gel Mixed Carbon 18 SEC columns connected in series were used. Samples were freshly prepared in THF (HPLC) and HPLC grade THF was used as the eluent at a flow rate of 1.0 mL/ min. The detection system was calibrated with a 2 mg/mL polystyrene standard (Mw ¼ 110 K). 2.5.3. Water uptake and swelling ratio Membrane samples (1 cm 4 cm, OH form) were immersed in deionized water for 24 h. They were then taken out and water on the surface was quickly wiped with tissue paper. The weight and length of wet membrane samples were recorded. The weight and length of dry membrane samples were also recorded after drying the membrane samples at 60 C for 24 h. Water uptake (WU) of membrane samples was calculated by eq. (1):
WU ¼
Wwet Wdry 100% Wdry
(1)
where Wwet is the weight of wet membrane sample and Wdry is the weight of dry membrane sample. Swelling ratio (SR) of membrane samples was characterized by linear expansion ratio, which was calculated by eq. (2):
SR ¼
Lwet Ldry 100% Ldry
(2)
where Lwet is the length of wet membrane sample and Ldry is the length of dry membrane sample. To evaluate the temperature-dependent water swelling, QPPO and RC-QPPO membrane samples were immersed in water with predetermined-temperature for 4 h. The WU and SR of these membrane samples were recorded and plotted versus temperature. 2.5.4. Ion exchange capacity (IEC) IEC of the membrane samples was measured by the Mohr's method. Membrane samples in Cl form were immersed in aqueous Na2SO4 solution (0.5 mol L1) for 8 h. After removing membrane samples, Cl content in the solution was titrated with aqueous AgNO3 solution (0.01 mol L1) and K2CrO4 was used as the indicator. The IEC values were calculated by eq. (3):
IEC ¼
V C Wdry
(3)
where V is the consumed volume of AgNO3 in titration, C is the concentration of aqueous AgNO3 solution (0.01 mol L1) and Wdry is the weight of dry membrane sample in Cl form. 2.5.5. Conductivity measurement Ion conductivity of membrane samples was measured by the four-point probe technique (Zahner Zemmiun E) with potentiostatic mode (amplitude 50 mV, frequency range 10 KHz-100 Hz). The membrane sample (1 cm 4 cm, OH, Cl or HCO 3 form) was placed in a Teflon cell which has two current collecting electrodes and two potential sensing electrodes (the distance between couple electrodes is 1 cm). The membrane was completely immersed in deionized water. The ionic conductivity was calculated according to the following equation:
k¼
L RWd
(4)
where R is the membrane resistance, L is the distance between
3
potential sensing electrodes. W and d are the width (here 1 cm) and thickness of the membrane, respectively. Before measurement, the samples were equilibrated at least 30 min at predetermined temperature. 2.5.6. Mechanical analysis Tensile strength measurements of hydrated and dry membrane samples were performed on a Q800 dynamic mechanical analyzer (DMA, TA Instruments) at a stretching rate of 0.5 N min1. 2.5.7. Alkaline stability To evaluate the chemical stability, QPPO and RC-QPPO membrane samples were immersed in aqueous NaOH solution (2 mol L1) at 60 C for 240 h. The hydroxide conductivity of these membranes was monitored at 30 C during this period and was plotted versus time. 2.5.8. H2/O2 fuel cell measurements Pt/C or PtRu/C catalysts (60 wt% in metal content) were mixed with the QPPO/isopropanol solution under ultrasonic to yield inks (20 wt% of polymer and 80 wt% of catalyst). The membrane electrode assemblies (MEAs) were fabricated using the catalyst-coatedmembrane (CCM) method: the PtRu anode ink and the Pt cathode ink were then sprayed on the correct sides of the membrane samples (with a thickness of 50 mme70 mm when fully hydrated). The metal loading in anode or cathode was controlled to be 0.5 mg/ cm2, and the electrode area was 12.25 cm2. The prepared CCM was then converted to OH form by immersing in aqueous NaOH (1 mol L1) solution for 12 h followed by thorough washing with deionized water. The obtained CCM was placed between two pieces of carbon paper (Toray TGP-H-060) to make the MEA without hotpressing. Single cell was tested using an 850E Multi Range fuel cell test station (Scribner Associates, USA) in a galvanic mode at 65 C. H2 and O2 were humidified at 65 C (100% RH) and fed with a flow rate of 1 L min1 and a backpressure of 0 MPa was kept on both sides. The cell voltage at each current density was recorded. 3. Results and discussion 3.1. AcS-QPPO synthesis AcS-QPPO bearing quaternary ammonium and thioacetate groups was prepared in a mild one-pot reaction, as outlined in Fig. 1. Although the conversion ratio of benzyl bromide into benzyl thioacetate shall be controlled to leave enough benzyl bromide for quaternization, as a proof of concept we tried to convert benzyl bromide into equal amount of quaternary ammonium and thioacetate in order for them to have obvious signals in the 1H NMR spectrum (Fig. 2). This quantified conversion can be conducted by adding equal amount of trimethylamine and potassium thioacetate (0.5 equiv. with respect to the molar content of benzyl bromide) while synthesizing AcS-QPPO. The peak at 3.85 ppm (H4 in Fig. 2b) belongs to the methylene protons in benzyl thioacetate group and the peak of methyl protons in thioacetate group appears at 2.20 ppm. As benzyl thioacetate was attached, the peak intensity of methyl protons has been changed and compared with the starting polymer (BPPO, Fig. 2a), the ratio of two different methyl protons, (H2þ H6)/H7 was greatly increased. As shown in Fig. 2b, a new peak at 3.05 ppm is observed, which belongs to the -CH3 group attached to the positive-charge quaternary ammonium group. As benzyl bromide (-CH2Br) reacted with trimethylamine, -CH2- peak shifted from 4.32 ppm to 4.43 ppm. It indicates the successful quaternization of AcS-QPPO polymer. Additionally, the integration ratio of peaks at 4.43 ppm to that at 3.85 ppm is close to 1:1, which is expected since the equal
Please cite this article in press as: J. Hou, et al., Journal of Power Sources (2017), http://dx.doi.org/10.1016/j.jpowsour.2017.06.073
4
J. Hou et al. / Journal of Power Sources xxx (2017) 1e8
Fig. 1. Preparation of recyclable AEM, RC-QPPO. The reaction starts from commercial polymer, brominated poly (2, 6-dimethyl-1, 4-phenylene oxide) (BPPO), in one pot. The obtained AcS-QPPO membrane can be cross-linked under alkaline condition to form RC-QPPO membrane. The disulfide bonds in RC-QPPO membrane can be reduced and broken by DTT, leading to the recycling of RC-QPPO.
Fig. 2. 1H NMR spectra of (a) BPPO (in CDCl3) and (b) AcS-QPPO (in DMSO-D6).
amount of trimethylamine and potassium thioacetate was added. It implies that by adjusting the amount of trimethylamine and potassium thioacetate added, the ratio of quaternization, i.e. IEC, and the number of potential cross-linking sites can be readily tuned. 3.2. AcS-QPPO membrane cross-linking and recycling We firstly dissolved AcS-QPPO in DMF and fabricated AcS-QPPO membrane by solution casting, as described in section 2.3. The obtained AcS-QPPO membrane can be repeatedly dissolved in DMF, implying a linear polymer structure in the membrane (Fig. 3a). As we immersed the membrane in aqueous NaOH solution (1 mol L1), the thioacetate groups will be hydrolyzed into the sulfhydryl groups (-SH), which can be quickly coupled to form disulfide groups when exposed to air. Therefore, when alkali treated, AcS-QPPO membrane will be gradually cross-linked and cannot be dissolved in DMF anymore, corresponding to a network polymer structure (Fig. 3b). By taking advantage of the reversible formation and break of disulfide groups, we were able to reprocess and recycle the cross-
linked AcS-QPPO membrane (namely, RC-QPPO). Treating RC-QPPO with DTT can break these disulfide bonds, leading to re-dissolving of RC-QPPO in DMF (Fig. 3c). The resulting solution can be utilized to cast a brand-new AEM, which can be cross-linked after removing the residual DTT (washed off with water), as described in section 2.4. The entire process can be carried out repeatedly and two repeated cross-linking and recycling processes were performed in the lab. It should be noted that the crosslinking degree has a great impact on the recycling process. We initially prepared a cross-linked membrane by adding 10 mol% of potassium thioacetate with respect to the molar content of benzyl bromide in the starting BPPO. The resulting RC-QPPO membrane can be completely re-dissolved and re-cast. However, as we increased the degree of cross-linking by adding more reactive -SH groups, it was slightly challenging to completely recycle all the starting polymers. This is possibly caused by the incomplete breaking of disulfide bonds due to restricted diffusion of DTT into the membrane matrix. We believe that adding more DTT, extending dissolution time and gently heating the solution would enable us to
Please cite this article in press as: J. Hou, et al., Journal of Power Sources (2017), http://dx.doi.org/10.1016/j.jpowsour.2017.06.073
J. Hou et al. / Journal of Power Sources xxx (2017) 1e8
5
Fig. 3. Images demonstrating the membrane (a) can be dissolved in DMF before cross-linking; (b) cannot be dissolved in DMF after alkaline treatment and (c) can be recycled when treated with dithiothreitol (DTT).
access 100% recycling. To demonstrate the effectiveness of our strategy, RC-QPPO membrane with a cross-linking degree of 10% was presented and thoroughly characterized (see following sections for property discussions).
3.3. Hydroxide conductivity of RC-QPPO We have prepared a QPPO membrane (as control) and series of cross-linked AcS-QPPO membranes (RC-QPPO) with different IEC by adding different amount of trimethylamine in AcS-QPPO synthesis, as described in section 2.2 and 2.3. Fig. 4 shows the hydroxide conductivity of QPPO-1.78, RC-QPPO-1.27, RC-QPPO-1.78 and RC-QPPO-2.13 as a function of operating temperature. As the operating temperature increases, the hydroxide conductivity increases because of enhanced OH mobility. For instance, as temperature increases from 30 C to 80 C, the hydroxide conductivity of RC-QPPO-1.27 increases from 13.3 mS cm1 to 29.7 mS cm1 and an increase from 24.6 mS cm1 to 64.5 mS cm1 is observed for RCQPPO-1.78. For membrane with a higher IEC, i.e. RC-QPPO-2.13, the conductivity stretched from 38.0 mS cm1 at 30 C to 69.6 mS cm1 at 80 C. However, when the operating temperature was increased to 80 C, a severe deformation of RC-QPPO-2.13 was observed due to the excessive membrane swelling. Additionally, the increase in hydroxide conductivity with an increase in IEC is expected since more exchangeable functional groups are available inside the membrane matrix. Comparing RCQPPO-1.78 membrane with QPPO-1.78 membrane that has the
Fig. 4. Hydroxide conductivity of QPPO-1.78, RC-QPPO-1.27, RC-QPPO-1.78 and RCQPPO-2.13 as a function of temperature.
same IEC, we did not observe significant difference in hydroxide conductivity, indicating that the cross-linking process has no obvious influence on membrane conductivity. These conductivity results suggest that RC-QPPO membranes show sufficient hydroxide conductivity (>10 mS cm1 at 30 C [19]) for a practical application in fuel cell. The conductivity of RC-QPPO membrane also depends on the type of counter anions. Fig. 5 presents the OH, Cl and HCO 3 conductivity of RC-QPPO (IEC ¼ 1.78 mmol/g) as a function of operating temperature. RC-QPPO membrane in OH form shows the highest conductivity, while in HCO 3 form it exhibits the lowest conductivity over the full operating temperature window because of the intrinsic difference in ion mobility [20]. As the operating temperature is elevated from 30 C to 80 C, the Cl conductivity increases from 11.4 mS cm1 to 30.5 mS cm1 and the HCO 3 conductivity climbs from 5.0 mS cm1 to 11.9 mS cm1. Lower ionic conductivity of RC-QPPO membrane in HCO 3 form implies the importance of CO2 removal in alkaline fuel cells. 3.4. Water swelling Water uptake and swelling ratio of QPPO and RC-QPPO with different IECs are measured and presented in Table 1. We observed that the water swelling of either membrane is greatly influenced by the content of functional groups, i.e. the IEC values. The RC-QPPO membrane with an IEC of 1.27 mmol g1 had a water uptake of 21.6 wt% and swelling ratio of 7.8%, while RC-QPPO membrane with
Fig. 5. The conductivity of RC-QPPO membrane with different counter ions (OH, Cl and HCO 3 ) as a function of temperature.
Please cite this article in press as: J. Hou, et al., Journal of Power Sources (2017), http://dx.doi.org/10.1016/j.jpowsour.2017.06.073
6
J. Hou et al. / Journal of Power Sources xxx (2017) 1e8
Table 1 IEC value, water uptake and swelling ratio of QPPO-1.78, QPPO-2.15, RC-QPPO-1.27, RC-QPPO-1.78 and RC-QPPO-2.13 at 30 C.
Table 2 Yielding stress and elongation at break of QPPO-1.78 and RC-QPPO-1.78 in both dry and wet state.
Membrane
IEC (mmol/g)
Water uptake (wt%)
Swelling ratio (%)
Membrane
QPPOdry
QPPOwet
RC-QPPOdry
RC-QPPOwet
QPPO-1.78 QPPO-2.15 RC-QPPO-1.27 RC-QPPO-1.78 RC-QPPO-2.13
1.78 2.15 1.27 1.78 2.13
41.3 163.1 21.6 37.4 70.0
15.4 36.5 7.8 13.4 21.8
Yielding stress (MPa) Elongation at break (%)
12.95 5.21
12.31 6.46
19.49 6.73
23.14 17.09
higher IEC (2.13 mmol/g) exhibited higher water uptake (70.0 wt%) and higher swelling ratio (21.8%). It is not surprising since the functional groups are hydrophilic and more water will be absorbed in membranes with higher content of functional groups, i.e. higher IEC. As we expected, RC-QPPO membranes, which are cross-linked by disulfide bonds, generally show less water uptake and increased swelling resistance compared with the linear counterpart, i.e. the QPPO membrane. For the QPPO-1.78 membrane, its water uptake and swelling ratio at 30 C are 41.3 wt% and 15.4%, respectively. However, for the cross-linked membrane with the same IEC, i.e. RC-QPPO-1.78 shows a decrease of 9.4% water uptake and a decrease of 13.0% swelling ratio. When the IEC of RC-QPPO membrane is further increased to 2.13 mmol g1, much more decrease in water uptake (57.1%) and swelling ratio (40.2%) are observed in comparison with QPPO-2.15 that has the similar IEC (2.15 mmol g1, Table 1). It is an obvious conclusion that disulfide bond cross-linking can enhance the water swelling capability of RCQPPO membranes, especially at higher IEC (>2.0 mmol g1). To facilitate a practical application, which usually involves different operating temperature, the swelling of RC-QPPO membranes are investigated at varied temperatures (Fig. 6). The water uptake of RC-QPPO-1.78 at 30 C is 37.4 wt%, which is increased to 59.7 wt% at 80 C, while the swelling ratio of RC-QPPO-1.78 is slightly increased over the temperature range 30 Ce80 C. Concerning the effect of elevated temperature on membrane water uptake and swelling ratio, we conclude that the reversibly crosslinked RC-QPPO shows good swelling resistance at elevated operating temperature in spite of the increased water content.
3.5. Mechanical properties The mechanical properties of cross-linked RC-QPPO (RC-QPPO-
Fig. 6. Water uptake and swelling ratio of RC-QPPO-1.78 as a function of temperature.
1.78) membrane and its linear counterpart QPPO (QPPO-1.78), with the same IEC, are investigated in both dry and wet (hydrated) state to illustrate the impact of cross-linking on membrane robustness. The results are presented in Table 2. The yielding stress of QPPOdry, QPPOwet, RC-QPPOdry and RC-QPPOwet are 12.95 MPa, 12.31 MPa, 19.49 MPa and 23.14 MPa, respectively. It is obvious that RC-QPPO1.78 exhibits enhanced robustness in comparison with QPPO-1.78 due to disulfide bond cross-linking. The elongations at break of QPPOdry, QPPOwet, RC-QPPOdry and RC-QPPOwet are 5.21%, 6.46%, 6.73% and 17.09%, respectively. Surprisingly, the elongation at break of wet RC-QPPO-1.78 is much higher than that of the rest, indicating a “significantly flexible” polymer network, which is not very common for a cross-linked membrane. It is attributed to the special nature of disulfide bonds, which is in dynamic equilibrium and can self-exchange in hydrated polymer network [21]. The results suggest that the RC-QPPO membranes show appropriate mechanical strength for a practical AEMFC.
3.6. RC-QPPO recycling and reusing The most important characteristic of RC-QPPO membrane is its capability to be recycled. Therefore, we sequentially repeated the cross-linking and recycling process twice to see the impact of such process on membrane properties. As a demonstration, these experiments and characterizations were conducted on RC-QPPO-2.13 membrane, which is the most promising candidate for alkaline fuel cell application according to conductivity measurements. The membranes recovered and recast for the first time and the second time were designated as RC1-QPPO and RC2-QPPO. To examine the impact of the recycling process on membrane properties, we firstly evaluated the IEC of both RC1-QPPO and RC2-QPPO and found that the IEC of the recycled membranes are similar, i.e. 1.99 mmol g1 for RC1-QPPO and 2.04 mmol g1 for RC2-QPPO. Compared with the pristine RC-QPPO membrane, the recycled membranes, both RC1QPPO and RC2-QPPO, show slightly lower IEC. It could possibly be ascribed to the degradation of functional groups in the process of recycling. We anticipate that with more stable functional groups, this could be mitigated. Still, the recycled membrane exhibited proper conductivities for fuel cell application. For example, RC1-QPPO and RC2-QPPO show hydroxide conductivity of 18.1 mS cm1 and 20.0 mS cm1 at 30 C, respectively and it lies in the acceptable window for a fuel cell application. It should also be mentioned that as the operating temperature increases, the conductivity can be further enhanced (42.0 mS cm1 at 80 C for RC1-QPPO and 39.5 mS cm1 at 80 C for RC2-QPPO), projecting even better fuel cell performance (Fig. 7). To further understand the recycling process, gel permeation chromatography (GPC) measurements were conducted. Because GPC measurements do not allow polymers with quaternised ammonium groups in case of blocking the column, the measurements were performed on PPO polymers without being quaternised. Results (Table 3) show that the recycled polymer (RC1-PPO) has similar molecular weight (Mn ¼ 31302 g/mol) as that of the pristine one (RC-PPO) (Mn ¼ 24514 g/mol), while the molecular weight distribution of RC1-PPO (PDI ¼ 4.03) is higher than that of the pristine RC-PPO (PDI ¼ 2.34). We presume that was due to
Please cite this article in press as: J. Hou, et al., Journal of Power Sources (2017), http://dx.doi.org/10.1016/j.jpowsour.2017.06.073
J. Hou et al. / Journal of Power Sources xxx (2017) 1e8
7
Fig. 9. H2/O2 single cell performance of RC-QPPO-2.13 and RC1-QPPO at 65 C. Fig. 7. Hydroxide conductivity of RC-QPPO, RC1-QPPO and RC2-QPPO as a function of temperature.
Table 3 GPC results of RC-PPO and RC1-PPO. HPLC grade THF was used as the eluent. Membrane
Mn (g/mol)
PDI
RC-PPO RC1-PPO
24514 31302
2.34 4.03
QPPO-2.13 membrane in 2 mol L1 NaOH at 60 C for 240 h, the IEC was decreased by 31% (from 2.13 mmol g1 to 1.48 mmol g1). This can be solved by incorporating more stable functional groups into the membrane, e.g. the imidazolium [22], phosphonium [23] and so on. The obtained results tend to suggest that the crosslinking strategy does not have much interference with the membrane alkaline stability.
3.8. H2/O2 fuel cell performance several residual disulfide bonds after recycling, which may not prevent the polymer from being soluble but increase the PDI of samples. 3.7. Alkaline stability Ex-situ alkaline stability tests are performed to evaluate the effect of introducing disulfide cross-linkage on membrane alkaline stability. Fig. 8 presents the hydroxide conductivity of QPPO-2.13 and RC-QPPO-2.13 as a function of time during alkaline treatment. The hydroxide conductivity decreased by 20% over the first 60 h and was slightly changed over the next 130 h. After 240 h of alkaline treatment, 65% of the conductivity was maintained. Similar behavior was observed for both QPPO-2.13 and RC-QPPO-2.13 since they have the same kind of functional groups (quaternary ammonium groups). The decline in hydroxide conductivity was due to the degradation of quaternary ammonium groups, which has been rigorously proved by IEC measurements. After immersing RC-
The original membrane (RC-QPPO-2.13, IEC ¼ 2.13 mmol g1) and the corresponding recycled membrane (RC1-QPPO, IEC ¼ 1.99 mmol g1) were selected to evaluate the effect of recycling process on fuel cell performance (Fig. 9). In a H2/O2 cell operated at 65 C, the peak power densities of RC-QPPO-2.13 and RC1-QPPO are 150 mW cm2 (at 0.6 V with a current density of 265 mA cm2) and 123 mW cm2 (at 0.5 V with a current density of 252 mA cm2), respectively. The slightly decrease in fuel cell performance after recycling is ascribed to the decrease in IEC and the resulting decrease in hydroxide conductivity. The performance we obtained here is comparable to previous studies [3,24], suggesting the potential application of RC-QPPO membrane and the recycling strategy in AEMFC. And we believe that the fuel cell performance of the pristine and the recycled membranes can be further improved by optimizing the fabrication of MEA and by incorporating more stable functional groups.
4. Conclusions
Fig. 8. Changing trends in hydroxide conductivity of QPPO-2.13 and RC-QPPO-2.13 at 30 C after immersing in 2 mol L1 NaOH at 60 C.
A PPO-based membrane (AcS-QPPO) bearing quaternary ammonium and thioacetate groups was successfully synthesized via one-pot reaction at mild condition. The thioacetate groups serve as potential cross-linkage and result in a reversibly cross-linked anion exchange membrane. The obtained cross-linked membrane, RC-QPPO exhibits enhanced mechanical strength, suppressed water swelling and comparable hydroxide conductivity in comparison with the control, QPPO membrane. Disulfide bond cross-linking strategy developed here leads to a cross-linked RC-QPPO membrane that can be repeatedly recycled and recast. After two repeated cross-linking and recycling processes, the membrane still showed acceptable hydroxide conductivity. The acquired facts prove that the strategy to fabricate a cross-linked AEM, which can be recycled, is very promising and could be extended to other types of membranes. Further study will be focused on developing recyclable cross-linked AEM with increased alkaline stability and its application in a practical AEMFC.
Please cite this article in press as: J. Hou, et al., Journal of Power Sources (2017), http://dx.doi.org/10.1016/j.jpowsour.2017.06.073
8
J. Hou et al. / Journal of Power Sources xxx (2017) 1e8
Acknowledgements Financial support received from the National Science Foundation of China (Nos. 21506201, 21490581, 91534203) and Postdoc Foundation Support (No. 2014M560521, 2015T80667) is gratefully acknowledged. We deeply appreciate the help Mr. Xian Liang gave us in investigating the fuel cell performance. References [1] B.C.H. Steele, A. Heinzel, Nature 414 (2001) 345e352. [2] J. Pan, C. Chen, Y. Li, L. Wang, L. Tan, G. Li, X. Tang, L. Xiao, J. Lu, L. Zhuang, Energy Environ. Sci. 7 (2014) 354e360. [3] J. Pan, S. Lu, Y. Li, A. Huang, L. Zhuang, J. Lu, Adv. Funct. Mater 20 (2010) 312e319. [4] X. Dong, B. Xue, H. Qian, J. Zheng, S. Li, S. Zhang, J. Power Sources 342 (2017) 605e615. [5] A.D. Mohanty, C. Bae, J. Mat. Chem. A 2 (2014) 17314e17320. [6] O.D. Thomas, K.J. Soo, T.J. Peckham, M.P. Kulkarni, S. Holdcroft, J. Am. Chem. Soc. 134 (2012) 10753e10756. [7] C.G. Arges, V. Ramani, Proc. Natl. Acad. Sci. 110 (2013) 2490e2495. [8] L. Wu, Q. Pan, J.R. Varcoe, D. Zhou, J. Ran, Z. Yang, T. Xu, J. Membr. Sci. 490 (2015) 1e8. [9] X. Chen, P. Chen, Z. An, K. Chen, K. Okamoto, J. Power Sources 196 (2011) 1694e1703.
[10] H. Han, M. Liu, L. Xu, J. Xu, S. Wang, H. Ni, Z. Wang, J. Membr. Sci. 496 (2015) 84e94. [11] J. Li, W. Cai, L. Ma, Y. Zhang, Z. Chenb, H. Cheng, Chem. Commun. 51 (2015) 6556e6559. [12] Z. Yue, Y.-B. Cai, S. Xu, J. Power Sources 286 (2015) 571e579. [13] W. Xu, Y. Zhao, Z. Yuan, X. Li, H. Zhang, I.F.J. Vankelecom, Adv. Funct. Mater 25 (2015) 2583e2589. [14] L. Zhu, T.J. Zimudzi, N. Li, J. Pan, B. Lin, M.A. Hickner, Polym. Chem. 7 (2016) 2464e2475. [15] J.-H. Ryu, R.T. Chacko, S. Jiwpanich, S. Bickerton, R.P. Babu, S. Thayumanavan, J. Am. Chem. Soc. 132 (2010) 17227e17235. [16] E.-K. Bang, M. Lista, G. Sforazzini, N. Sakai, S. Matile, Chem. Sci. 3 (2012) 1752e1763. [17] A. Takahashi, T. Ohishi, R. Goseki, H. Otsuka, Polymer 82 (2016) 319e326. [18] J. Kim, K. Baek, D. Shetty, N. Selvapalam, G. Yun, N.H. Kim, Y.H. Ko, K.M. Park, I. Hwang, K. Kim, Angew. Chem. Int. Ed. Engl. 54 (2015) 2693e2697. [19] G. Merle, M. Wessling, K. Nijmeijer, J. Membr. Sci. 377 (2011) 1e35. [20] L. Zhu, J. Pan, Y. Wang, J. Han, L. Zhuang, M.A. Hickner, Macromolecules 49 (2016) 815e824. [21] G. Deng, F. Li, H. Yu, F. Liu, C. Liu, W. Sun, H. Jiang, Y. Chen, ACS Macro Lett. 1 (2012) 275e279. [22] X. Lin, J.R. Varcoe, S.D. Poynton, X. Liang, A.L. Ong, J. Ran, Y. Li, T. Xu, J. Mat. Chem. A 1 (2013) 7262e7269. [23] S. Gu, R. Cai, T. Luo, Z. Chen, M. Sun, Y. Liu, G. He, Y. Yan, Angew. Chem. Int. Ed. Engl. 48 (2009) 6499e6502. [24] K.H. Lee, D.H. Cho, Y.M. Kim, S.J. Moon, J.G. Seong, D.W. Shin, J.-Y. Sohn, J.F. Kima, Y.M. Lee, Energy Environ. Sci. 10 (2017) 275e285.
Please cite this article in press as: J. Hou, et al., Journal of Power Sources (2017), http://dx.doi.org/10.1016/j.jpowsour.2017.06.073