Quaternized adamantane-containing poly(aryl ether ketone) anion exchange membranes for vanadium redox flow battery applications

Journal of Power Sources 325 (2016) 801e807

Contents lists available at ScienceDirect

Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour

Quaternized adamantane-containing poly(aryl ether ketone) anion exchange membranes for vanadium redox flow battery applications Bengui Zhang a, *, Shouhai Zhang b, **, Zhihuan Weng b, Guosheng Wang a, Enlei Zhang a, Ping Yu a, Xiaomeng Chen a, Xinwei Wang a a b

College of Chemical Engineering, Shenyang University of Chemical Technology, Shenyang 110142, China State Key Laboratory of Fine Chemicals, School of Chemical Engineering, Dalian University of Technology, Dalian 116024, 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


New adamantane-containing poly(aryl ether ketone) membranes are prepared for VRFB.
QADMPEK membranes show significantly low vanadium ions permeability.
High battery efficiency of VRFB with QADMPEK membranes are obtained.
QADMPEK membranes show high stability in VRFB.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 8 April 2016 Received in revised form 13 June 2016 Accepted 24 June 2016

Quaternized adamantane-containing poly(aryl ether ketone) anion exchange membranes (QADMPEK) are prepared and investigated for vanadium redox flow batteries (VRFB) application. The bulky, rigid and highly hydrophobic adamantane segment incorporated into the backbone of membrane material makes QADMPEK membranes have low water uptake and swelling ratio, and the as-prepared membranes display significantly lower permeability of vanadium ions than that of Nafion117 membrane. As a consequence, the VRFB cell with QADMPEK-3 membrane shows higher coulombic efficiency (99.4%) and energy efficiency (84.0%) than those for Nafion117 membrane (95.2% and 80.5%, respectively) at the current density of 80 mA cm2. Furthermore, at a much higher current density of 140 mA cm2, QADMPEK membrane still exhibits better coulombic efficiency and energy efficiency than Nafion117 membrane (coulombic efficiency 99.2% vs 96.5% and energy efficiency 76.0% vs 74.0%). Moreover, QADMPEK membranes show high stability in in-situ VRFB cycle test and ex-situ oxidation stability test. These results indicate that QADMPEK membranes are good candidates for VRFB applications. © 2016 Elsevier B.V. All rights reserved.

Keywords: Vanadium redox flow battery Anion exchange membrane Adamantane Poly (aryl ether ketone)

1. Introduction

* Corresponding author. ** Corresponding author. E-mail addresses: [email protected] (B. Zhang), [email protected] (S. Zhang). http://dx.doi.org/10.1016/j.jpowsour.2016.06.101 0378-7753/© 2016 Elsevier B.V. All rights reserved.

Vanadium redox flow batteries (VRFB) are promising energy storage devices for renewable energy and power grids [1]. As key materials in VRFB, membranes are used to prevent cross-mixing of the positive and negative electrolytes while allowing the transport of ions to complete the electrical circuit. Of these membranes, the

802

B. Zhang et al. / Journal of Power Sources 325 (2016) 801e807

most common membrane investigated for VRFB is commercial available Nafion membrane due to its high conductivity and excellent chemical stability. However, Nafion membrane has suffered from high cost and severe permeability of vanadium ions. Recently, VRFB have benefited from a large number of alternative membrane studies, where cation exchange membranes [2e4], anion exchange membranes [5e14], porous membranes [15e18] and amphoteric ion exchange membranes [19e21] have been reported for VRFB application. All of these alternative membranes demonstrated great promise for VRFB applications and laid a firm foundation for the further development of alternative membranes with high performance. However, enhancing durability in oxidative vanadium electrolyte still remained a significant challenge in this fields. It has been demonstrated that restricting water uptake and swelling behavior of the membrane can improve the chemical stability of the ion exchange membrane by inhibiting the absorption of VOþ 2 and further delaying the membrane degradation [4,14,22]. Adamantane has rigid spherical cycloaliphatic-cage structure, which imparts it with high melting point (269  C) and excellent thermal stability [23]. The bulky and rigid adamantane incorporated in to the backbone of polymer resulted in much higher Tg than that of polysulfone derived from bisphenol A, because the former has greater resistance to segmental motion [24]. In this paper we reported on an unique anion exchange membrane derived from adamantane-containing poly(aryl ether ketone), where the bulky, rigid and highly hydrophobic adamantane segment can restrict the water uptake and swelling of membrane, the as-prepared membrane was employed for VRFB application, and showed high stability attribute to the introduction of adamantane segment, the other properties of membrane and VRFB cell performances were also investigated in detail. 2. Experimental 2.1. Materials 1,3-Dibromoadamantane (Luzhou Dazhou chemical Ltd., China), chloromethyl ethyl ether (CMEE) (Henan wanxiang technology & trade co., Ltd), tin tetrachloride (SnCl4) (Sinopharm Chemical Reagent Co., Ltd), vanadyl sulfate (VOSO4) (Shanghai Luyuan Fine Chemical Plant), 4,40 -difluorobenzophenone (DFB) (Changzhou Huashan Chemical Co., Ltd). All chemicals were chemically pure grade or better and were used as supplied. 2.2. Synthesis of adamantane-containing poly(aryl ether ketone) The monomer 1,3-(4-hydroxyphenyl)adamantane (1,3-ADM) was synthesized from 1,3-dibromoadamantane and phenol according to the literature procedure [25]. As depicted in Scheme 1, the adamantane-containing poly(aryl ether ketone) (ADMPEK) were synthesized as following method: 4, 40 -Difluorobenzophenone (DFB) (0.1 mol), 1,3-ADM (0.1 mol), anhydrous potassium carbonate (0.14 mol), sulfolane (40 mL), and toluene (20 mL) were charged to a three-neck flask equipped with a mechanical stirrer, DeaneStark trap, and reflux condenser. Under nitrogen purge, the reaction mixture was refluxed for 2 h at 140  C. After refluxing, the DeaneStark trap was drained to remove the toluene, and the temperature was maintained at 180  C for another 20 h. Following the reaction, the viscous solution was diluted with 60 mL of sulfolane, and then the solution was precipitated into water. The intrinsic viscosity for ADMPEK in chloroform at 25  C was measured on an Ubbelohde capillary viscometer as an indication of molecular weight.

2.3. Preparation of anion exchange membrane An illustrative synthesis for chloromethylated adamantanecontaining poly(aryl ether ketone) (CADMPEK) was provided as following. Adamantane-containing poly(aryl ether ketone) (5.0 g) had been dissolved in nitrobenzene (75 mL), then, chloromethyl ethyl ether (CMEE) (5 mL) and calculated SnCl4 were added to the polymer solution. The temperature of polymer solution was maintained at 30  C for 10 h. Then, the polymer solution was precipitated with excess of ethanol. The obtained polymer was dried at 60  C under vacuum for 48 h. The degree of substitution (DS) of CADMPEK was calculated from 1H NMR spectroscopy (Bruker Avance 500 M). A series of CADMPEK with different DS were prepared by altering the amount of SnCl4. Preparation of membrane: CADMPEK was dissolved in 1methyl-2-pyrrolidone to form a 10 wt % solution. Then, the solution was casted and heated on a smooth glass plate (60  C, 9 h), after that, the resulting base membrane was immersed in trimethylamine solution (33 wt%) (40  C, 72 h) to obtain quaternized adamantane-containing poly(aryl ether ketone) membrane (QADMPEK) (Scheme 1). The thickness of QADMPEK membranes was in the range of 40e45 mm. 2.4. Characterization of membrane 1 H NMR (Bruker Avance 500 M) and FT-IR analysis (Varian 6400 FT-IR spectrometer) were used to confirm the chemical structure of ADMPEK and CADMPEK, respectively. The ion exchange capability (IEC) of membrane was determined by titration. QADMPEK membrane (in chloride form) was dried under vacuum (60  C, 48 h), and then the membrane was weighed and immersed in 25 mL of 0.1 mol L1 NaNO3 solution for 48 h at room temperature. After that the solution was titrated with 0.05 mol L1 AgNO3 over a potentiometric titrator (ZD-2, Shanghai INESA & Scientific Instrument Co., Ltd, China). IEC was calculated according to the following equation:

IEC ¼

VAgNO3  CAgNO3 M

where VAgNO3 was the volume of AgNO3 solution, CAgNO3 was the concentration of AgNO3 solution, and M was the weight of dried membrane, respectively. Water uptake was calculated according to the following equation:

Water uptakeð%Þ ¼

Wwet  Wdry  100% Wdry

where Wwet and Wdry were the weights of the membrane in wet and dry state, respectively. The swelling ratio was calculated according to the following equation:

Swelling ratio ¼

L1  L0  100% L0

where L1 and L0 were the length of the soaked membrane and dry membrane, respectively. The area resistances of the membranes were measured by electrochemical impedance spectroscopy (EIS) over a frequency range from 100 kHz to 1 Hz (CH Instruments Electrochemical Workstation, CHI 660E) [26]. The conductivity cell was separated with membrane, and each cell was filled with 1.5 M VOSO4 in 3 M H2SO4. The effective area S of membrane was 1 cm2 r1 and r2

B. Zhang et al. / Journal of Power Sources 325 (2016) 801e807

803

Scheme 1. Synthesis of quaternized adamantane-containing poly(aryl ether ketone) anion exchange membrane (QADMPEK).

were the electric resistances of the conductivity cell with membrane and without membrane, respectively. The area resistance of the membrane r (U cm2) was calculated by the following equation:

r ¼ ðr1  r2 Þ  S The permeability of VO2þ through membrane was evaluated by the methods described in literature [27,28]. The diffusion cell was separated by a membrane (active area of the membrane was 1.76 cm2), and the left half-cell was filled with 1.5 M VOSO4 in 3 M H2SO4 (30 mL), while the right half-cell was filled with 1.5 M MgSO4 in 3 M H2SO4 (30 mL). The concentration of VO2þ ion in MgSO4 solution was measured on an UV/Vis spectrophotometer (764 nm), and the ion permeability was calculated by using the Fick Law:

VB

dCB ðtÞ P ¼ S ðCA  CB ðtÞÞ dt L

where VB was the volume of solution in the right half-cell; S and L

were the effective area and thickness of the membrane, respectively; P was the permeability of VO2þ; CA was the concentration of VO2þ in the left half-cell; CB was the concentration of VO2þ in the right half-cell; and t was test time. Ex-situ chemical stability of the membranes was evaluated according to the methods described in literatures [8,29]. Dried membrane sample (30 mm  35 mm) was immersed in 50 mL of 2þ  0.15 M VOþ in the 2 solution at 40 C. The concentration of VO solution was measured by UV/Vis spectroscopy with VO2þ as an indicator, as the VOþ 2 can be produced during the oxidation of membrane by VO2þ. The ex-situ chemical stability test in higher concentration of VOþ 2 also was carried out, where the QADMPEK membrane was immersed in 1.5 M VOþ 2 in 3 M H2SO4 solution for 30 days under room temperature. The VRFB cell was assembled according to our previous report [10,12]. Two pieces of carbon felts (5 mm of thickness, 5 cm2) were used as electrodes, and two pieces of graphite plates were served as current collectors. The thickness of electrode frame was 4 mm. The negative electrolyte (30 mL) and positive electrolyte (30 mL) were

804

B. Zhang et al. / Journal of Power Sources 325 (2016) 801e807

1.5 M V2þ/V3þ in 3.0 M H2SO4 and 1.5 M VO2þ/VOþ 2 in 3.0 M H2SO4, respectively. Charge-discharge tests were conducted by using a battery test system CT-3008-5V/3A (Neware Co., Ltd, China), and the upper limit of charge voltage and lower limit of discharge were 1.65 V and 0.8 V, respectively.

Table 1 Degrees of chloromethylation for CADMPEK. Polymer

CMEE (mL)

SnCl4 (mL)

DS

CADMPEK-1 CADMPEK-2 CADMPEK-3

5 5 5

0.2 0.3 0.4

0.75 0.91 1.08

3. Results and discussion 3.1. Polymer synthesis and membrane properties Adamantane-containing poly (aryl ether ketone) (ADMPEK) was polymerized from monomers (DFB and 1,3-ADM) (Scheme 1). 1H NMR spectra of ADMPEK was shown in Fig. 1. The peaks around 2.0 ppm were assigned to the protons (H(6e13)) of adamantane [30], which indicated that the adamantane had been incorporated into the backbone. The viscosity of ADMPEK was tested at 25  C in chloroform, and the result (intrinsic viscosity, hinh ¼ 0.87 dL g1) indicated that ADMPEK had high molecular weight and could be used as an available membrane material. Chloromethylated adamantane-containing poly(aryl ether ketone)s (CADMPEKs) with various degrees of substitution were synthesized with CMEE as the chloromethylation reagent and SnCl4 as the catalyst (Table 1). 1H NMR spectrum of CADMPEK-3 was shown in Fig. 1, the peak at 4.64 ppm was assigned to the methylene protons of chloromethyl groups [31]. Due to the ortho-para directing effect of oxygen atom on the benzene rings of polymer repeat units, the electron-rich positions (H1, ortho relative to oxygen atom) were substituted by chloromethyl group, and the change of chemical shift of protons (H2) also confirmed that substitution occurred in the positions (H1). Degree of substitution (DS) of CADMPEK was calculated from 1H NMR spectrum by taking the integral ratio of eCH2Cl (H(5)) at 4.64 ppm to the protons at 7.79 ppm (H(4)) which were intact during the chloromethylation reaction, and the calculation equation was shown as following: DS ¼ 2A5 =A4 where A5 was the integration value of H(5), and A4 was the integration value of H(4), respectively. FT-IR spectra of ADMPEK, CADMPEK and QADMPEK were shown in Fig. 2, the absorbance around 1241 cm1 was attributed to the aryl ether bond, and the peaks at 2902 cm1 and 2848 cm1 were

O 1

1 2 13 2

9 10 11

1

3 O

2

CDCl3

3 CH2Cl 5

4 O 4 C

3

4

3

4

3395

1241

QADMPEK-3

6,7,8,9,10,11,12,13

5

4 2 1,3

2902 2848

CADMPEK-3

Absorbance

2 76 8 12

assigned to the CeH absorption of adamantane [30], which also confirmed that the adamantane had been incorporated into the backbone as expected. QADMPEK membrane displayed absorbance around 3395 cm1, which was attributed to water bound to the ionic groups [10,32], indicating that hydrophilic quaternized ammonium groups had been successfully introduced into the membrane. The properties of QADMPEK membranes were presented in Table 2. Ion exchange capacity (IEC) reflects the amount of hydrophilic anion exchange groups in membrane, here the IEC values of resulting membranes were in the range of 1.34e1.75 mmol g1. The water uptake of QADMPEK membranes correlated with IEC. Water uptake values were enhanced from 10.2 to 18.8% with the increasing of IEC from 1.34 to 1.75 mmol g1. Compared to QADMPEK membranes, Nafion117 membrane showed much higher water uptake due to its highly hydrophilic sulfonic acid groups. Furthermore, the water uptake of QADMPEK membrane was significantly lower than that of other anion exchange membrane based on poly(phenyl sulfone) with comparable IEC [10,31], which demonstrated that this adamantane-containing membrane material had higher water repellency due to the increase of the hydrophobic hydrocarbon character of the main chain. This behavior is similar to what has been observed in adamantane-containing polyimides [33]. For swelling ratio, in general, it should be increased with the increase of IEC value due to the hydrophilicity of ion exchange groups. In our cases, QADMPEK membranes showed lower swelling ratios (ranging from 5.5 to 8.6%) than those of anion exchange membranes based on poly(phenyl sulfone) [10]. The swelling ratio of QADMPEK membrane may be restricted by the combined effects from strong hydrophobic effect and interference from bulky and rigid adamantane segment. QADMPEK membranes showed lower swelling ratio values than that of Nafion117 membrane due to the more rigid aromatic backbone and lower water uptake of QADMPEK membranes. Moreover, the low swelling ratio

CADMPEK-3

ADMPEK ADMPEK

8

7

6

5

4

3

2

1

Chemical Shift / ppm 1

Fig. 1. H NMR spectra of adamantane-containing poly(aryl ether ketone) (ADMPEK) and chloromethylated adamantane-containing poly(aryl ether ketone) (CADMPEK-3) in CDCl3.

4000

3500

3000

2500

2000

1500

1000

500

-1

Wave numbers (cm ) Fig. 2. FT-IR spectra of ADMPEK, CADMPEK-3 and QADMPEK-3 membrane.

B. Zhang et al. / Journal of Power Sources 325 (2016) 801e807

805

Table 2 Properties of QADMPEK and Nafion117 membranes. Membranea

IEC (mmol g1)

Water uptake (%)

Swelling ratio (%)

Area resistance (U cm2)

Vanadium permeability P(107 cm2 min1)

QADMPEK-1 QADMPEK-2 QADMPEK-3 Nafion117

1.34 1.53 1.75 0.96

10.2 14.5 18.8 31.8

5.5 6.4 8.6 17.1

1.57 0.95 0.49 0.60

0.06 0.51 7.64 31.30

a Membranes with quaternized ammonium groups derived from the chloromethylated adamantane-containing poly(aryl ether ketone)s (CADMPEK-1, CADMPEK-2 and CADMPEK-3) were named QADMPEK-1, QADMPEK-2 and QADMPEK-3, respectively.

of QADMPEK membranes would benefit the stability of membrane in oxidative electrolyte. The area resistance of membrane also is a critical issue for VRFB applications. As expected, with the increase of IEC, the area resistance of QADMPEK was declined from 1.57 U cm2 to 0.49 U cm2. The area resistance values of QADMPEK-1 and QADMPEK-2 membranes were higher than that of Nafion117 membrane. And, compared to Nafion117, the QADMPEK-3 membrane with the highest IEC value (1.75 mmol g1) showed lower area resistance. Membrane intended for VRFB must be an effective barrier for vanadium ions crossover from the anode to the cathode compartment, because crossover of vanadium ions would lower the voltage and reduces the system capacity. The permeability of VO2þ against diffusion time through QADMPEK and Nafion117 membranes were measured under identical conditions and the results were summarized in Fig. 3. It is clear that the permeability of VO2þ ions was depressed greatly in QADMPEK membranes compared with Nafion117 membrane, and the vanadium permeability P values for QADMPEK membranes were much lower than that of Nafion117 (Table 2). These results were in common with other aromatic anion exchange membranes, which derives from a combination of Donnan exclusion effect of positively charged ion exchange groups with vanadium ions and narrower, less connected hydrophilic channels in aromatic polymer based ion conducting membranes [8,10]. 3.2. Ex-situ chemical stability of the membrane Ex-situ chemical stability of the membranes in 0.15 M VOþ 2 solutions was tested by using VO2þ as an indicator. The VO2þ concentration for QADMPEK membranes slightly increased against membrane immersion time (Fig. 4), whereas the values for Nafion117 remained relatively low because of the high stability of perfluorinated polymer backbone. The concentration of VO2þ for

30

To develop suitable membranes for VRFB running at high current density, the performances of VRFB cell with obtained membranes were investigated at a wide range of current densities. VRFB working at higher current density would need smaller VRFB stack size and reduce the area of electrode and membrane, which would lead to a lower overall cost for the VRFB system, and benefit for the practical application of VRFB [17]. The typical charge-discharge curves for VRFB with QADMPEK and Nafion117 membranes were outlined in Fig. 5a. The VRFB with QADMPEK-3 membrane showed slightly lower charge voltage and higher discharge voltage than

4.5

-1

15 0 3

2+

2

VO

2+

-1

VO concentration (mmol L )

45

3.3. VRFB cell performance

concentration (mmol L )

Nafion117 QADMPEK-1 QADMPEK-2 QADMPEK-3

QADMPEK-3 membrane was slightly higher than those of QADMPEK-1 and QADMPEK-2 membranes, indicating that QADMPEK-1 and QADMPEK-2 membranes had higher chemical stability than QADMPEK-3 membrane. This phenomenon has also been observed in sulfonated poly(phenylene) membranes [34] and anion exchange membranes [10], because membrane with higher IEC and swelling ratio provides more opportunity for the VOþ 2 ions to get into the hydrophilic domains and attack the membrane [36]. The concentration of VO2þ in the case with QADMPEK-3 was changed from 2.1 to 3.3 mmol L1 after the membrane was immersed for 768 h, whereas under the similar test conditions, this value was increased from 0 to 3.8 mmol L1 on highly stable sulfonated poly(ether ether ketone) membranes (about 580 h) [4], indicating that QADMPEK membranes would have much higher chemical stability. These results demonstrated that QADMPEK membranes were stable in acid and oxidant 0.15 M VOþ 2 solutions, therefore they would be good alternative membranes for VRFB application.

1

4.0 3.5 3.0 2.5 2.0 1.5

0

0

10

20

30 40 Time ( h )

50

60

Fig. 3. Concentration of VO2þ in MgSO4 solution versus diffusion time.

70

Nafion117 QADMPEK-1 QADMPEK-2 QADMPEK-3

0

100

200

300

400

500

600

700

800

Time (h) Fig. 4. VO2þ concentration of 0.15 M VOþ 2 solution versus membranes immersion time at 40  C.

B. Zhang et al. / Journal of Power Sources 325 (2016) 801e807

those of Nafion117 membranes, which was attributed to the lower area resistance of QADMPEK-3. In contrary, the VRFB with QADMPEK-1 and QADMPEK-2 membranes showed higher charge voltage and lower discharge voltage than those of Nafion117 and QADMPEK-3 membranes due to the lower conductivity of QADMPEK-1 and QADMPEK-2 membranes. Furthermore, the charge and discharge capacity of QADMPEK-3 higher than those of VRFB with Nafion117, QADMPEK-1 and QADMPEK-2 membranes. For QADMPEK-3 and Nafion117 membranes, the charge capacities and discharge capacities decreased with the increasing of current density, which was mainly attributed to the higher overpotential and ohmic resistance at a higher current density. Coulombic efficiency (CE) of VRFB with QADMPEK membranes were remained above 98.6% at all current densities investigated (Fig. 5b), indicating that QADMPEK membranes had high ion selectivity. The QADMPEK membranes and Nafion117 membrane shared a similar dependency of the CE upon current densities, but the formers exhibited higher CE values as a result of their much lower vanadium ions permeability. Voltage efficiency (VE) values of the cells both with QADMPEK and Nafion117 membranes were declined with the increasing of current densities (Fig. 5b). This phenomenon was mainly attributed to the higher overpotential and ohmic resistance at a higher current density. It is noted that at a relative low current density (from 40 mA cm2 to 80 mA cm2), the VE values of QADMPEK-3 membrane were slightly higher than those of Nafion117 membrane. This was mainly attributed to the lower area resistance of QADMPEK-3 (0.49 U cm2, Table 2) compared with Nafion117 membrane (0.60 U cm2). However, at higher current density (from 100 mA cm2 to 140 mA cm2), QADMPEK-3 membrane and Nafion117 membrane exhibited comparable performance. This phenomenon may be attributed to the Donnan exclusion effect between quaternized ammonium groups and protons that the ion conductivity of anion exchange membrane decreased at a high

(a)

1.6

1.4

-2

80 mAcm

1.2 1.0 Nafion117 QADMPEK-1 QADMPEK-2 QADMPEK-3

0.8 0.6 0

Voltage (V)

Voltage (V)

1.6

1.4 1.2 1.0

-2

QADMPEK-3 120mAcm

0.8

Nafion117

-2

QADMPEK-3 140mAcm

0.6

200 400 600 800 1000

Nafion117

0

200

-2

CE of QADMPEK-3 EE of QADMPEK-3 CE of Nafion117 EE of Nafion117

90

80

70

60

0

20

40

60

80

100

Cycle number Fig. 6. Charge-discharge cycling performance (80 mA cm2) of VRFB with QADMPEK-3 and Nafion117 membranes.

current density [9]. Moreover, these results also indicated that the QADMPEK-3 membrane would be a good candidate for a VRFB running at high current density (80 mA cm2 to 140 mA cm2). In addition, the VE values of QADMPEK-1 and QADMPEK-2 membranes were lower than those of QADMPEK-3 and Nafion117 membranes due to their higher area resistance (1.57 and 0.95 U cm2, respectively). Energy efficiency (EE) for QADMPEK and Nafion117 membranes were outlined in Fig. 5b. In the entire range of test current density, the EE values of cell with QADMPEK-3 membrane were higher than those with Nafion117 membrane. Particularly, at a high current density of 80 mA cm2, few membranes reported would display EE values higher than 83% [11,15,35], however, here QADMPEK-3 showed an EE of 84%, which demonstrated that the obtained membrane showed excellent performance in VRFB. Moreover, at a much higher current density of 140 mA cm2, the cell with QADMPEK-3 exhibited superior EE value (76%) to that with Nafion117 membrane (74%). The above results also confirmed that QADMPEK-3 would be a good candidate for VRFB running at high current density (80 mA cm2 to 140 mA cm2). Membrane durability in VRFB is a critical factor for non-

140mAcm

400 600 800

100

Capacity (mAh)

Capacity (mAh)

(b)

-2

120mAcm

100

Efficiency ( % )

806

Nafion 117 QADMPEK-1 QADMPEK-2 QADMPEK-3

90 80

EE ( % )

VE ( % )

100 90 80 70 100 90

90 Efficiency ( % )

CE ( % )

100

80

CE of QADMPEK-3 EE of QADMPEK-3 CE of QADMPEK-3 after immersing in 1.5 M VO2+ for 30 days

70

EE of QADMPEK-3 after immersing in 1.5 M VO +for 30 days

60

2

80 70

40 40

60

80

100

-2

120

140

Current density (mA cm ) Fig. 5. (a) Charge-discharge curves of VRFB with QADMPEK and Nafion117 membranes; (b) efficiencies of VRFB with membranes at different current densities.

60

80

100

120

140

-2

Current density ( mA cm ) Fig. 7. The performance of VRFB with QADMPEK-3 before and after immersing in 1.5 M VOþ 2 in 3 M H2SO4 electrolyte solution for 30 days under room temperature.

B. Zhang et al. / Journal of Power Sources 325 (2016) 801e807

fluorinated ion exchange membrane. The cycling test was conducted to investigate the durability of QADMPEK-3 membrane in VRFB (Fig. 6). Due to the time constrains, the cycling test was conducted for 100 cycles (about 15 days). During the cycling test, QADMPEK-3 membrane maintained higher CE and VE than those for Nafion117 membrane indicating that QADMPEK-3 membrane was durable during the cycling test. In general, membranes with higher IEC would display lower stability, because more oxidative vanadium ions would permeate into the membrane and lead to the degradation of the membrane [36]. In order to further investigate the stability of QADMPEK membrane, an accelerating oxidizing experiment was carried out, the membrane with the highest IEC (QADMPEK-3) was immersed in 1.5 M VOþ 2 in 3 M H2SO4 electrolyte solution for 30 days under room temperature, then the performance of VRFB cell with the pretreated membrane was tested. Under the above conditions, SPES [37] or SPEEK [38] membranes would readily break into pieces. In addition, Polysulfone based anion exchange membrane, which showed stable battery performance over 1600 cycles [14], was broken up after soaking in 1.5 M VOþ 2 for 18 days [36]. On the contrary, QADMPEK-3 membrane was unbroken and could be assembled in VRFB to test the cell performance after soaking in 1.5 M VOþ 2 for 30 days, indicating that QADMPEK-3 membrane has better stability. As showed in Fig. 7, there was no significant change in CE and EE before and after treatment, indicating that the QADMPEK membrane was endurable in oxidative 1.5 M VOþ 2 electrolyte. The results further confirmed the high stability of QADMPEK membranes and suggested that the QADMPEK membranes could be promising candidates for VRFB application. 4. Conclusions Novel quaternized adamantane-containing poly(aryl ether ketone) anion exchange membranes (QADMPEK) were synthesized and characterized. QADMPEK membranes exhibited low water uptake and swelling ratio. Compared with Nafion117 membrane, QADMPEK membranes also displayed significantly lower permeability of vanadium ions. As a result, the VRFB with QADMPEK-3 membrane exhibited better coulombic efficiency and energy efficiency (CE 99.4%, EE 84.0%) than those of Nafion117 (CE 95.2%, EE 80.5%) at 80 mA cm2. Moreover, QADMPEK membranes showed high stability in the ex-situ chemical stability test and the in-situ cycle test. The combination of good VRFB cell performance and high stability in oxidative vanadium electrolyte solution makes QADMPEK membranes attractive as alternative for VRFB applications. More investigations on anion exchange membrane based on adamantane-containing poly(aryl ether)s are in progress and will be reported in the near future. Acknowledgment We greatly acknowledge financial support from the Natural

807

Science Foundation of China (No. 21444006).

References [1] R. Darling, K. Gallagher, J. Kowalski, S. Ha, F. Brushett, Energy Environ. Sci. 7 (2014) 3459e3477. [2] S. Winardi, S.C. Raghu, M. Oo, Q. Yan, N. Wai, T. Lim, M. Skyllas-Kazacos, J. Memb. Sci. 450 (2014) 313e322. [3] W. Dai, Y. Shen, Z. Li, L. Yu, J. Xi, X. Qiu, J. Mater. Chem. A 2 (2014) 12423e12432. [4] Z. Yuan, X. Li, Y. Duan, Y. Zhao, H. Zhang, Polym. Chem. 6 (2015) 5385e5392. [5] C. Sun, Z. Tang, C. Belcher, T. Zawodzinski, C. Fujimoto, Electrochem. Comm. 43 (2014) 63e66. [6] Z. Mai, H. Zhang, H. Zhang, W. Xu, W. Wei, H. Na, X. Li, ChemSusChem 6 (2013) 328e335. [7] Y. Li, X. Lin, L. Wu, C. Jiang, M. Hosain, T. Xu, J. Memb. Sci. 483 (2015) 60e69. [8] S. Yun, J. Parrondo, V. Ramani, J. Mater. Chem. A 2 (2014) 6605e6615. [9] B. Zhang, S. Zhang, D. Xing, R. Han, C. Yin, X. Jian, J. Power Sources 217 (2012) 296e302. [10] B. Zhang, E. Zhang, G. Wang, P. Yu, Q. Zhao, F. Yao, J. Power Sources 282 (2015) 328e334. [11] S. Zhang, B. Zhang, D. Xing, X. Jian, J. Mater. Chem. A 1 (2013) 12246e12254. [12] S. Zhang, B. Zhang, G.F. Zhao, X.G. Jian, J. Mater. Chem. A 2 (2014) 3083e3091. [13] S. Zhang, C. Yin, D. Xing, D. Yang, X. Jian, J. Memb. Sci. 363 (2010) 243e249. [14] W. Xu, Y. Zhao, Z. Yuan, X. Li, H. Zhang, I. Vankelecom, Adv. Funct. Mater. 25 (2015) 2583e2589. [15] Z. Yuan, Y. Duan, H. Zhang, X. Li, H. Zhang, I. Vankelecom, Energy Environ. Sci. 9 (2016) 441e447. [16] Y. Li, X. Li, J. Cao, W. Xu, H. Zhang, Chem. Commun. 50 (2014) 4596e4599. [17] X. Wei, Z. Nie, Q. Luo, B. Li, B. Chen, K. Simmons, V. Sprenkle, W. Wang, Adv. Energy Mater. 3 (2013) 1215e1220. [18] B. Gindt, D. Abebe, Z. Tang, M. Lindsey, J. Chen, R. Elagmmal, T. Zawodzinski, T. Fujiwara, J. Mater. Chem. A (2016), http://dx.doi.org/10.1039/C6TA00698A. [19] S. Liu, L. Wang, D. Li, B. Liu, J. Wang, Y. Song, J. Mater. Chem. A 3 (2015) 17590e17597. [20] Y. Wang, S. Wang, M. Xiao, S. Song, D. Han, M. Hickner, Y. Meng, Int. J. Hydrogen Energy 39 (2014) 16123e16131. [21] A. Chromik, A. Santos, T. Turek, U. Kunz, T. Haring, J. Kerres, J. Memb. Sci. 476 (2015) 148e155. [22] A. Mohanty, Y. Lee, L. Zhu, M. Hickner, C. Bae, Macromolecules 47 (2014) 1973e1980. [23] L. Mathias, C. Lewis, K. Wiegel, Macromolecules 30 (1997) 5970e5975. [24] M.R. Pixton, D.R. Paul, Polymer 36 (1995) 3165e3172. [25] Y.T. Chern, H.C. Shiue, Macromolecules 30 (1997) 4646e4651. [26] Z. Mai, H. Zhang, X. Li, C. Bi, H. Dai, J. Power Sources 196 (2011) 482e487. [27] T. Sukkar, M. Skyllas-Kazacos, J. Appl. Electrochem. 34 (2004) 137e145. [28] J. Qiu, M. Li, J. Ni, M. Zhai, J. Peng, L. Xu, H. Zhou, J. Li, G. Wei, J. Memb. Sci. 297 (2007) 174e180. [29] S. Kim, T. Tighe, B. Schwenzer, J. Yan, J. Zhang, J. Liu, Z. Yang, M. Hickner, J. Appl. Electrochem. 41 (2011) 1201e1213. [30] Z. Geng, Y. Lu, S. Zhang, X. Jiang, P. Huo, J. Luan, G. Wang, Polym. Int. 63 (2014) 333e337. [31] M. Hibbs, M. Hickner, T. Alam, S. Mcintyre, C. Fujimoto, C. Cornelius, Chem. Mater. 20 (2008) 2566e2573. [32] H. Zhang, H. Zhang, F. Zhang, X. Li, Y. Li, I. Vankelecom, Energy Environ. Sci. 6 (2013) 776e781. [33] Y. Chern, H. Shiue, Macromolecules 30 (1997) 5766e5772. [34] C. Fujimoto, S. Kim, R. Stains, X. Wei, L. Li, Z. Yang, Electrochem. Commun. 20 (2012) 48e51. [35] F. Zhang, H. Zhang, C. Qu, ChemSusChem 6 (2013) 2290e2298. [36] Z. Yuan, X. Li, Y. Zhao, H. Zhang, ACS Appl. Mater. Interfaces 7 (2015) 19446e19454. [37] D. Chen, M. Hickner, Phys. Chem. Chem. Phys. 15 (2013) 11299e11305. [38] Z. Yuan, X. Li, J. Hu, W. Xu, J. Cao, H. Zhang, Phys. Chem. Chem. Phys. 16 (2014) 19841e19847.