- Email: [email protected]
One-pot solvent-free synthesis of cross-linked anion exchange membranes for electrodialysis
Journal of Membrane Science 515 (2016) 115–124
Contents lists available at ScienceDirect
Journal of Membrane Science journal homepage: www.elsevier.com/locate/memsci
One-pot solvent-free synthesis of cross-linked anion exchange membranes for electrodialysis Qi Pan a, Md. Masem Hossain a, Zhengjin Yang a, Yaoming Wang a,b, Liang Wu a,b,n, Tongwen Xu a,nn a CAS Key Laboratory of Soft Matter Chemistry, Collaborative Innovation Center of Chemistry for Energy Materials, School of Chemistry and Materials Science, University of Science and Technology of China, Hefei 230026, PR China b Hefei Chemjoy Polymer Materials Co. Ltd., Hefei 230601, PR China
art ic l e i nf o
a b s t r a c t
Article history: Received 8 January 2016 Received in revised form 27 May 2016 Accepted 28 May 2016 Available online 31 May 2016
Solvent-free strategy has attracted a broad interest in preparation of ion exchange membranes in recent years stemming from the no need of organic solvents and facilitated construction of highly crosslinking networks. However, post-functionalization is commonly required for further introducing ion-exchange groups. In this study, we report a one-pot solvent-free synthesis of cross-linked anion exchange membranes (AEMs) without the need for post-functionalization. We firstly dissolved brominated poly (2, 6dimethyl-1, 4-phenylene oxide) (BPPO) in liquid monomers mixture of 4-vinylbenzyl chloride (VBC) and styrene without any organic solvent, then added appropriate amount of N-vinylimidazole and N-methylimidazole to introduce imidazolium groups into both of polyelectrolyte and solvents. The transparent, robust and crosslinked AEMs were obtained by the thermal crosslinking of the unsaturated moieties during the membrane forming process. This approach, distinct from the classical post-functionalization processes, performs crosslinking and functionalization simultaneously. Particularly, imidazolium cations locating on cross-links endow the resultant membrane a high concentration of charge carriers (ionexchange capacity) for target low resistance, while maintain a high mechanical stability. & 2016 Elsevier B.V. All rights reserved.
Keywords: Anion exchange membrane Solvent-free Electrodialysis Crosslinking Desalination
1. Introduction Ion-exchange membranes (IEMs) are polymer electrolytes that conduct anions or cations, as they contain counter charged groups bound covalently to a polymer backbones [1]. Over the last decade or so, there is an increasing worldwide interest in the use of IEMs on a large industrial scale in processes such as electrodialysis, electrodeionization, diffusion dialysis, and energy conversion and storage systems [2–5]. Among them, electrodialysis (ED), in which a series of alternating anion- and cation-exchange membranes arranged between two electrodes, is today one of the most stateof-the-art ion-exchange membrane process. It has found industrial scale application in water desalination and recovery of useful substances from industrial waste water [3]. As the key component in ED devices, cation-exchange membranes (CEMs) have successfully been commercialized for the reliable manufacturing n Corresponding author at: CAS Key Laboratory of Soft Matter Chemistry, Collaborative Innovation Center of Chemistry for Energy Materials, School of Chemistry and Materials Science, University of Science and Technology of China, Hefei 230026, PR China. nn Corresponding author. E-mail addresses: [email protected] (L. Wu), [email protected] (T. Xu).
http://dx.doi.org/10.1016/j.memsci.2016.05.050 0376-7388/& 2016 Elsevier B.V. All rights reserved.
technique, however, anion-exchange membranes (AEMs) still face technique challenges in large-scale preparation and further improving their performance, i. e. high ions selective permeability and stability [6,7]. AEMs have a selective permeability for anions as they contain fixed positive charged ions, such as ammonium, phosphonium, guanidinium ions [8–15]. Traditional preparation of AEMs usually employs the post-modification of pristine polymers or direct polymerization of functionalized monomers. For post-modification method, the most common approach is to carry out a chloromethylation or bromomethylation on the polymer so as to further cationize them with trimethylamine. A variety of polymers, such as polysulfone [16], poly (phthalazinone ether sulfone ketone) [17], poly(ether-imide) [18], poly(phenylene oxide) [19], poly (phenylene) [20], could act as scaffolds for preparing AEMs. The direct polymerization is usually performed by polymerization of monomers with moieties bearing ammonium, halomethyl-containing or amine-containing groups, that can be conveniently converted into anionic exchange groups in organic solvent. Coates et al. reported their work on the facile synthesis and ring-opening metathesis polymerization of a tetraalkylammonium-functionalized norbornene with dicyclopentadiene as a cross-linkable monomer to yield strong AEMs [21,22]. Zhang et al. reported that
116
Q. Pan et al. / Journal of Membrane Science 515 (2016) 115–124
partially fluorinated poly (arylene ether sulfone) with pendant ammonium groups were prepared by copolymerization of bisphenol monomers containing amine groups and partially fluorinated monomers, followed by quaternization [23]. We have reported preparation of high performance AEM by polyacylation of pre-quaternized monomers to endow the resultant AEMs with excellent hydroxide ion conductivity [24]. Although these strategies represent a promising progress in the development of AEMs in recent years, they all require excess reagents and long reaction time to obtain highly functionalized polymers. Particularly, the large amount of organic solvents in both reaction and membrane formation processes will carries toxic risk to the environment. Hence, to further industrial-scale manufacture, it is important to develop simple, rapid and environmental friendly methods for production of AEMs with excellent performances. To overcome these obstacles, we previously developed a solvent-free strategy for AEMs by the in-situ polymerization [25]. Different from the above mentioned post-modification and direct polymerization, it starts from the dissolving of polymer electrolytes by liquid monomers, other than organic solvent, followed by in-situ polymerization to obtain un-charged membranes. It should be noted that post-quaternization was required to convert halomethyl groups to ammonium cations by immersing base membrane into trimethylamine (TMA) aqueous solution. The disadvantages of this strategy include: a) the reaction efficiency is very low due to the solid-liquid heterogeneous reaction. b) The quality of the membrane is hardly to be controlled for different production batch, and membranes usually exhibit a loose morphology due to the erosion of TMA. In view of the above mentioned issues, we herein presents our current efforts toward developing upgradable solvent-free strategy for AEMs. This approach, distinct from the classical post-quaternization processes, performs crosslinking and quaternization simultaneously. We firstly dissolved brominated poly (2, 6-dimethyl-1, 4-phenylene oxide) (BPPO) in liquid monomers mixture of 4-vinylbenzyl chloride (VBC) and styrene without any organic solvent, then added appropriate amount of N-vinylimidazole and N-methylimidazole to introduce imidazolium cations into both of polyelectrolyte and solvents. The transparent, robust and crosslinked AEMs were obtained by the thermal crosslinking of the unsaturated moieties during the membrane forming process. Particularly, the imidazolium cations locating on cross-links endow the resultant membrane a high concentration of charge carriers for target high ions permeability, while maintain a high mechanical stability. Properties characterizations, such as ion exchange capacity (IEC), membrane area resistance (Rm), electrodialysis related performance, etc., are discussed extensively to show the advantages of the solvent-free strategy over the conventional post-modification method.
2. Experimental methods 2.1. Materials Brominated poly (2, 6-dimethyl-1, 4-phenylene oxide) (BPPO) was kindly supplied by Tianwei Membrane Corporation Ltd. of Shandong (China). 4-Vinylbenzyl chloride (VBC), which was purchased from Changzhou Wujin Linchuan Chemical Co. Ltd. N-vinylimidazole (N-VI, 98%) and N-methylimidazole (N-MI, 99%) were purchased from Energy Chemical Co. Ltd. Styrene, sodium sulfate (Na2SO4), sodium chloride (NaCl), N-Methyl pyrrolidone (NMP, AR grade), N, N-dimethylformamide (DMF, AR grade), chlorobenzene (CB, AR grade), dimethyl sulfoxide (DMSO, AR grade) were purchased from Shanghai-Sinopham Chemical Reagent Co. Ltd. (China). Commercial anion exchange membrane Neosepta AMX
(thickness 120–180 mm, IEC 1.4–1.7 meq g 1, area resistance 2.0– 3.5 Ω cm2) and cation exchange membrane Neosepta CMX (thickness 220–260 mm, IEC 1.5–1.8 meq g 1, area resistance 2.0–3.5 Ω cm2) were purchased from ASTOM, Japan. 2.2. Membranes preparation As shown in Scheme 1, BPPO (1 g/0.00298 mol –CH2Br) was dissolved in a liquid monomers mixture; then appropriate amount of N-Vinylimidazole was added to convert all bromomethyle of BPPO and half amount of chloromethyl groups of VBC to imidazolium cations. The remaining chloromethyl groups were converted to imidazolium cations by adding N-Methylimidazole dropwise. The composition of reaction mixture was determined according to the solubility of polymer electrolytes (BPPO) in the liquid monomers. In this study, the content of BPPO (1 g) keeps constant for the preparation of all membranes. Styrene, which is employed as solvent, need a critical content (18 ml) to completely dissolve BPPO. Hence, the contents of BPPO and styrene are kept constant during the preparation. The content of VBC was varied to tune the amount of functional groups in membranes. Accordingly, the contents of N-VI and N-MI were varied with the change of VBC content to convert functional groups (–CH3Br and –CH3Cl) to imidazolium cations, also to keep the membrane with an appropriate crosslinking degree. Table 1 shows the components of the different reaction mixtures. The resulting membranes are designated as M-x, where x represents the amount of VBC (0.25 ml. 0.5 ml, 0.75 ml and 1.0 ml). Before characterization, all membranes were immersed in ether at room temperature for 4 h to remove unreacted monomers (styrene, VBC, N-VI and N-MI, which are all soluble in ether), followed by thorough washing with deionized water to remove imidazolium salts (possibly resulting from the reaction of unreacted VBC with N-VI and N-MI). 2.3. Membranes characterizations 2.3.1. Fourier transform infrared spectroscopy (FT-IR) FT-IR spectra were measured using a Vector 22 Fourier transform infrared spectrometer (Bruker). 2.3.2. Measurement of gel content Gel content was measured by calculating the ratio of mass remaining after immersing the membrane in organic solvent for 7 days at room temperature, typical procedure is as follows: a) Four pieces of membranes (2 4 cm, BPPO-0.25, BPPO-0.5, BPPO-0.75, BPPO-1.0)were dried at 60 °C for 24 h, then weighted (m0). b) The dried membranes were soaked in NMP, DMSO, DMF and CHCl3 at room temperature for 7 days, respectively. After that, they were washed and immersed in distilled water for 1 day. c) Finally, they were dried at 60 °C for 24 h, and weighed (mt). The gel content was calculated according to the following equation:
DC =
mt × 100% m0
(1)
2.3.3. Thermal analysis TGA and DTG thermograms were recorded using a TGA Q5000 V3.15 analyzer at a heating rate of 10 °C min 1 under a N2 atmosphere in order to evaluate the short-term thermal stability of membrane.
Q. Pan et al. / Journal of Membrane Science 515 (2016) 115–124
117
Scheme 1. Schematic diagram of the preparation process.
Table 1 Membrane compositions. Sample
BPPO/g
Styrene/ml
VBC/ml
N-VI/ml
N-MI/ml
M-0.25 M-0.5 M-0.75 M-1
1 1 1 1
1.8 1.8 1.8 1.8
0.25 0.5 0.75 1.00
0.35 0.45 0.53 0.62
0.075 0.15 0.23 0.30
2.3.4. Ion exchange capacity The ion exchange capacities (IEC) were measured according to our previous work [26]. The samples were firstly converted into Cl form after immersing in the NaCl aqueous solution (0.5 mol L 1) for 24 h. Then, the samples were washed and immersed in distilled water for 24 h. After that, the samples were dried to a constant weight and then immersed in a Na2SO4 aqueous solution (0.5 mol L 1) for 24 h. The concentration of Cl
released from the membrane was measured using a REX PHS-3C analyzer (China) and an attach Cl ion-selective electrode (model pCl-1). The IEC value was calculated from the measured amounts released Cl ions and is expressed as mmol g 1 of the dry membrane. 2.3.5. Water uptake Water uptake was calculated by taking the difference between the wet weight (mwet) and the dry weight (mdry) of the AEMs. To obtain the wet weight, AEM samples were equilibrated in distilled water at room temperature for 2 days. The water uptake was calculated according to the following equation:
WR=
mwet −mdry × 100% mdry
(2)
118
Q. Pan et al. / Journal of Membrane Science 515 (2016) 115–124
2.3.6. Membrane area resistance Membrane area resistance was measured using a commercial cell-assembly ((CJ-MPMD-1, Hefei Chemjoy Co., Ltd., China)) under a constant current mode [27]. It is composed of five compartments: two electrode chambers at both ends of the cell; two intermediate compartments equipped with two reference electrodes; one quadrate clip for membrane. Particularly, two pieces of Nafion-117 membranes were assembled between electrode and intermediate chambers, separately, to eliminate the influence of electrode reaction. Titanium (Ti) coated stainless steel sheets were used as cathode and anode. Two intermediate chambers are separated by the membrane clip, and tips of the reference electrodes should be closely to the center of membrane. During the measurement, Na2SO4 solution (0.3 mol L 1) was fed to electrode chambers, and NaCl (0.5 mol L 1) is fed to intermediate chambers. A constant current is supplied by a direct current power supply (SHEKONIC, Yangzhou Shuanghong Co., Ltd.) and the potential between electrodes is read by a digital multimeter (model: GDM8145, Good will instrument Co. Ltd., Taiwan). Then, by subtraction of the electrolyte resistance (Rsol) from the membrane resistance equilibrated in the electrolyte solution (Rcell), the membrane resistance (Rmem) was calculated according to the equation Rmem ¼Rcell Rsol, and final membrane area resistance was calculated as follows:
R = Rmem × A
(3)
where Rmem is the resistance of membranes (Ω), A is the effective area for membrane (7 cm2). 2.3.7. Membrane potential and static transport number The membrane potential (Em) and the transport number t were determined using the same device for membrane resistance. Two intermediate chambers, which are equipped with two reference electrodes, are separated by the membrane clip. Tips of the electrodes should be closely to the center of membrane. During the measurement, continuous flows of aqueous KCl (0.01 mol/L) and KCl (0.05 mol/L) were pumped into intermediate chambers, separately. The potential difference between the electrodes (Em) was determined by a digital multimeter (model: GDM-8145, Good will instrument Co. Ltd., Taiwan). The static transport number ťi of the anion in the membrane can be calculated as follows:
Em=(2ti′−1)
⎛C ⎞ RT ln ⎜ 1 ⎟ ⎝ C2 ⎠ nF
Fig. 1. ED cell configuration (a) and schematic diagram of the test in ED system.
connected to the Autolab (PGSTAT 30, Eco Chemie, Netherland). The chronopotentiogram was carried out at a constant applied current density of 3.5 mA cm 2 and the corresponding potential was recorded in every 0.01 s for a period of 200 s, automatically. 2.3.9. Continuous-mode electrodialysis (ED) A schematic diagram of continuous-mode ED stack is illustrated in Fig. 1. It consists of five compartments. Each chamber provides 7.07 cm2 active area for membranes with an o-ring to prevent leakage during the testing. There were three streams in the ED stack, dilute, concentrated and electrode wash. ED experiments were conducted in the recirculation mode of operation, in which each stream (150 cm3) was recirculated. Typically, a feed solution (0.1 mol L 1 NaCl) is pumped into the dilute and concentrate cells with a constant flow rate of 30 ml min 1. Meanwhile, two electrode cells were fed by 0.3 mol L 1 Na2SO4 solution and were connected together to prevent pH change. The commercial cation-exchange membrane Neosepta CMX were placed at the position C1, C2 and C3 in Fig. 1. The tested membrane was placed at the position A.
(4)
3. Results and discussion
where R is universal gas constant, F is the Faraday constant, T is the room temperature, n is the electrovalence (1 in this case), and C1 and C2 are the concentrations of electrolyte solutions in the testing cell, respectively.
3.1. Membrane preparation
2.3.8. Current-voltage (i-v) characteristic and choronopotentiometry The i-v curves of the membranes were measured at room temperature using the same three-compartment cell in 0.1 M NaCl. The choronopotentiograms of the membranes were measured using the same device without auxiliary membranes in 0.025 M NaCl. For i-v curves measurement, the cell was connected to a DC power supply (Shekonic, China, Model WWL-LDX) by the Pt electrodes and to the multimeter by the reference electrodes. Experiments were conducted in recirculation mode of operation, in which each stream was recirculated with a flow rate of 30 ml min 1. 1.5 wt% Na2SO4 was used as an electrode rinse solution and circulated in electrode compartments from a single reservoir. 0.1 M NaCl electrolyte solution was recirculated in the central compartment. A stepwise current was applied and the corresponding potential difference across the membrane was measured. For chronopotentiogram measurement, the cell was
Solvent-free method is widely considered as one of the most state-of-art strategy for preparing IEMs due to the advantages of easily forming interpenetrating polymer networks (IPNs) and no need of organic solvent. It is commonly performed by dissolving polymer electrolytes in liquid monomers followed by in-situ polymerization. However, post-modifications, e.g. quaternary amination or sulfonation, are required for introducing ion-exchange groups, and usually cause unexpected corrosion to polymer electrolytes. To overcome this shortcoming, this study presents an alternative one-pot solvent-free strategy with no need of post-modification process. As shown in Scheme 1a, this approach, distinct from the classical post-functionalization processes, performs crosslinking and quaternization simultaneously. we firstly dissolved brominated poly (2, 6-dimethyl-1, 4-phenylene oxide) (BPPO) in liquid monomers mixture of 4-vinylbenzyl chloride (VBC) and styrene without any organic solvent, then added appropriate amount of N-Vinylimidazole and N-methylimidazole to introduce imidazolium cations into both of polyelectrolyte and solvents. The
Q. Pan et al. / Journal of Membrane Science 515 (2016) 115–124
transparent, robust and crosslinked AEMs were obtained by the thermal crosslinking of the unsaturated moieties during the membrane forming process. Styrene, which is a favorable solvent as VBC, is the crucial component during the process. The resulting membrane would swell excessively by water and brittle into pieces if no styrene being involved (as shown in Scheme 1f) for the large amount of hydrophilic imidazolium cations converted from chloromethyl groups in VBC. Hence, an appropriate amount of styrene can maintain the content of imidazolium cations in a proper level and significantly enhance the toughness of the resulting membrane. Moreover, the stepwise adding of N-VI and N-MI is another key step for the ionic crosslinked membrane preparation. If only N-VI being employed, excessive amount of vinyl groups would be introduced to BPPO and VBC. The resultant membrane is brittle for the excessive crosslinking. On the contrary, the resultant membrane would be highly water swelling if only employing N-MI. To prepare membrane with a proper crosslinking degree, an appropriate content of N-VI was firstly required to convert all bromomethyle of BPPO and half amount of chloromethyl groups of VBC to imidazolium cations. Then the rest of chloromethyl groups of VBC were converted to imidazolium cations by adding an appropriate content of N-MI. As shown in Scheme 1e, we perform performed the following polymerization at 100 °C according to our previous study on self-crosslinked membrane. It was performed by attaching alkene pendant groups onto polymer mainchains via the Menshutkin reaction, followed by thermal crosslinking of the unsaturated side chains during the membrane formation process [28]. In-situ FTIR analysis indicated that the self-crosslinking of the pendant alkene groups was temperature dependence. The intensity of the alkene peaks decreases while increasing temperature treatment (30–100 °C), suggesting that the polymerization of alkene groups happens during heating; the higher temperature, the higher crosslinking degree. In this study, after temperature-curing at 100 °C, a yellowish, transparent and flexible AEM was obtained and easily being cut in any shape. 3.2. Investigation of the crosslinking process FTIR analysis was performed to confirm the quaternization reaction. Fig. 2 shows the FTIR spectra of the pure BPPO membrane, and M-0.25, M-0.5, M-0.75 and M-1, respectively. All solvent-free membranes show strong bands at 3060 cm 1 which are attributed to –HC ¼ N in imidazolium group, suggesting that quaterization reaction occurs successfully during membranes
119
preparation. Furthermore, comparing with the FTIR spectrum of BPPO, new peaks at 3425 cm 1 appear in the spectra of solventfree membranes. These peaks are attributed to the vibration of the hydroxyl groups in water molecules bonded to hydrophilic imidazolium groups, further indicating the quaterizaiton reaction. According to the reaction, the crosslinking generates methylene groups. The overlapped stretching vibration peaks in the range of 2854–2923 cm 1 are attributed to the stretching vibration of carbon hydrogen bonds in methylene (–CH2–), suggesting that crosslinking happens during membrane preparation. However, there are still residual alkene groups (–CH ¼ CH2) in the membrane after crosslinking. As shown in the spectra of all solvent-free membranes, the double peaks at 1604 and 1653 cm 1 attributed to the vibration of the carbon carbon double bonds in the pendant alkene groups (–CH ¼CH2) were clearly detected. It should be noted that a proper crosslinking degree is expected to maintain the membrane excellent flexibility and stability. Hence, further solubility and thermal stability tests were performed to determine the advantage of the moderate crosslinking. Solubility test is a common method to determine the formation of crosslinking networks in membrane. As shown in Table 2, M-1, M-0.75 and M-0.5 membranes are insoluble in NMP, and can maintain over 90% of the mass retention in other solvents, such as DMF, DMSO and CHCl3, while pure BPPO membrane is completely soluble in NMP and CHCl3, and slightly soluble in DMF, but insoluble in DMSO. Hence, it can be concluded that the crosslinking network has constructed in the membrane successfully. In addition to the organic solvent resistance, the crosslinking network is also beneficial to the improvement of thermal stability of membrane. Fig. 3 shows the TGA curves of BPPO and crosslinked membranes. These curves exhibited three main degradation stages, arising from the processes of thermal desorption of water, thermal deamination, and thermal oxidation of the polymer matrix. Notably, original BPPO exhibit higher thermal stability than the crosslinked membranes at the temperature over 375 °C, at which crosslinks degrade rapidly. In the temperature range of 300–375 °C, crosslinked structure endows the membrane better thermal stability than BPPO. Basically, the normal working temperature of the membrane for ED application is lower than 100 °C. as shown in Fig. 3, crosslinked membranes exhibit a slightly weight loss in this temperature range of 0–100 °C, which is attributed to the loss of water molecules absorbed by hydrophilic imidazolium groups, other than the degradation of membrane. TGA result suggests that the crosslinked membranes exhibit sufficient thermal stability at temperature below 100 °C for ED application. 3.3. Electrochemical properties of the membranes The content of charged functional groups (commonly reflected as IEC) in an IEM plays an important role in providing hydrophilic environments for ions transport. This study performs crosslinking and quaternization simultaneously. Particularly, hydrophilic imidazolium groups locating on cross-links is expected to endow the Table 2 Solubility properties of solvent-free crosslinked membranes. NMP M-1 M-0.75 M-0.5 M-0.25 BPPO
Fig. 2. ATR-FTIR spectra of BPPO membrane and solvent-free membranes.
a
100% 100% 100% 95.2% 0%
DMF
DMSO
CHCl3
98.5% 97.2% 94.8% 79.6% 17.4%
97.8% 98.9% 89.1% 74.6% 99.0%
96.4% 93.3% 92.1% 80.8% 0%
a The data indicate the percentage of the mass retention of each membrane after being immersed in organic solvent for 7 days at room temperature.
120
Q. Pan et al. / Journal of Membrane Science 515 (2016) 115–124
(in the range of 0.93–0.95). As discussed above, imidazolium cations locate on cross-links in the membrane. Hence, the membrane with higher VBC content possess higher IEC as well as much denser crosslinking networks, which is expected to resistant against water swelling and maintain a proper ions selective transport property (transport number). Moreover, M-0.75 exhibits the highest transport number of 0.95 among the crosslinked membranes. It can be explained by the effect of the distribution of conducting region on membrane surface. In addition to the dense ionic crosslinking networks, M-0.75 might possesses a highly ordered alignment of conducting regions on the surface, which is likely to endow the membrane excellent selective ions transport property. 3.4. Electrodialysis related performances of the membrane
Fig. 3. TGA and DTG (inset) traces for the solvent-free membranes recorded with a heating rate of 10 °C min 1 and under flowing nitrogen.
Fig. 4. IEC and water uptake of M-1, M-0.75, M-0.5 and M-0.25.
Table 3 Transport number and area resistance of the solvent-free membranes and commercial AEM (Neosepta AMX). Name
Transport number
Area resistance (Ω/cm2)
M-0.25 M-0.5 M-0.75 M-1 Neosepta AMX
0.93 0.93 0.95 0.94 0.98
1.40 1.12 1.12 0.98 2.52
Electrodialysis is generally operated under direct current which provides the driving force for ions transport through membrane. However, concentration polarization often happens at the interface between an IEM and an electrolyte solution (boundary layer) when an electric current passes through the system, leading to the performance degradation of the ED device. Chronopotentiometry is a powerful tool for studying transport phenomena and concentration polarization near membrane interfaces [29]. In this study, it was used for investigating the polarization phenomena and the micro-scale distribution of conducting and non-conducting regions on the membrane surface. The typical chronopotentiograms with three-stage potential differences were obtained and shown in Fig. 5. After a constant current density of 3.5 mA cm 2 is applied to the system, a potential (E0) onset was observed due to the resistance of the membrane and working solution. Corresponding to the concentration polarization, this initial potential gradually increased for a certain period because the counter ions pass through IEMs and gather on the other side of the membrane interface. The potential significantly increased and finally reached a steady state when the ionic concentration of depleting solution nearby the membrane interface dropped to zero [30]. The difference between the initial potential and the maximum potential from chronopotentiograms (ΔE) was believed to relate indirectly to the thickness of the concentration gradient established at the membrane interfacial zone. The time at which the potential transition occurs is called the transition time (τ) and is determined by the intersection of the tangents with the first and second stages of the curve. It can also be defined by the well-known Sand's equation based on the assumption of entire conducting surface of the membrane.
τ=
( C0 zi F )2πD 2 4i2 ( ti′ − ti )
(5) 2
resultant membrane a high concentration of charge carriers (IEC) for target low resistance, while maintain a high ions selective transport property. Fig. 4 presents the change of IEC and water uptake along with the varying of VBC content. As expected, the higher VBC content, the more imidazolium groups were introduced into membrane and finally resulting in a higher IEC and water uptake. Accordingly, the membrane area resistances decrease from 1.4 Ω cm2 to 0.98 Ω cm2, which are much lower than that of commercial membrane (Neosepta AMX, 2.52 Ω cm2, measured at the same conditions). Basically, a high IEC often leads to excessive membrane swelling on hydration, and tends to cause a poor selective ions transport property (usually reflected by transport number). However, as shown in Table 3, the crosslinked membranes with different IECs exhibit similar transport numbers
where i is the current density (mA cm ), C0 is the concentration of electrolyte, Zi is the valence of the ith ion, F is the Faraday constant and D is the diffusion coefficient (cm2 s 1). t'i the transport number of the counter-ion in the membrane and ti is that of the same ion in free solution at the same concentration. The transition times deprived from chronopotentiogram (Fig. 5b) are listed in Table 4 and compared with the theoretical values from Eq. (5). It clearly shows that the experimental data are slightly lower than the corresponding theoretical ones. It is worth mentioning that the Eq. (5) is based on the assumption of the entire conducting membrane surface. For an IEM, the non-conducting regions, which do exist on the real membrane surface, possibly result in the lower values. While comparing M-0.25, M-0.5 and M-1 in Table 4, the τ values decrease alone with the increasing of VBC contents. And the more VBC content shows
Q. Pan et al. / Journal of Membrane Science 515 (2016) 115–124
121
Fig. 5. Chronopotentiograms (a) and their derivation dE/dt curve (b) of the synthesized membranes compared with commercial membranes tested at room temperature in 0.025 M NaCl. Table 4 Characteristic values from chronopotentiograms. Membrane
τa (s)
τb (s)
ε
E0 (V)
Emax (V)
ΔE (V)
Δt (s)
M-1 M-0.75 M-0.5 M-0.25 Neosepta AMX
49.4 46.6 52.4 52.4 39.6
49.1 45.1 51.3 51.1 39.5
0.994 0.968 0.979 0.975 0.997
0.209 0.211 0.209 0.211 0.223
0.779 0.875 0.856 0.794 1.147
0.570 0.664 0.647 0.583 0.924
19 16 28 29 13
a b
Estimated from Eq. (5). Deprived experimentally from chronopotentiogram.
shorter defined transition time with smaller interval time (Δt). Considering the difference between conducting and non-conducting regions in an IEM, the local current density of the conducting regions (i*) will be higher than the overall current density (i) applied to the system [31]. i* can be expressed as a function of i and the fraction of the conducting region (ε) as shown below:
i*=
i ε
(6)
ε can be estimated by the derived Sand's equation
Thus the below [31]: 1
ε=
2iτ 2 ( t′i−ti ) 1/2
C0 Z i F ( π D)
(7)
The ε as well as other characteristic values from chronopotentiogrames are summarized in Table 4. Basically, the higher content of charged functional groups in an IEM (reflected by IEC) is supposed to result in a higher ε value. While comparing M-0.25, M-0.5 and M-1 in Table 5, the ε value increases along with the amount of VBC content in the membrane. It is in accordance with the result in Fig. 4, in which the higher VBC content, the more imidazolium groups were introduced into membrane and finally Table 5 Characteristic values from i–v curve. a
Membrane
ilim
M-1 M-0.75 M-0.5 M-0.25 Neosepta AMX
20.1 22.8 21.5 20.4 14.2
a
(mA cm 2)
Experimental LCD values recorded from Fig. 5.
ΔE (V)
Ror (Ω cm2)
0.48 0.89 1.00 1.05 0.71
9.94 7.28 9.10 11.83 14.63
resulting in a higher IEC. It is interesting to note that M-0.75 does not follow this principle. It exhibits the lowest values among the crosslinked membrane. Considering the high transport number of M-0.75, it is reasonable to explain by the effect of geometrical heterogeneity of membrane surface reported by previous study [30], which suggested that not only the ε value but also the alignment of conducting regions were needed to be taken into account for explaining the ions transport performance of the membrane. Accordingly, M-0.75 might possess highly ordered distribution between conducting and non-conducting regions on the surface even at a low ε value, further resulting in a shorter transition time (τ) as well as and interval time (Δt). In addition to the chronopotentiogram, limiting current density (LCD) under a flowing solution is another important parameter to determine the ED performance of the membrane. As well known, a boundary layer close to the desalting surface of an IEM will be formed when an electric current is applied across the membrane. When the applied current density is lower than LCD, the mass transfer is balanceable in both boundary layer and membrane matrix. However, the mass transfer across the membrane is much faster than that in boundary layer when it is higher than LCD, causing concentration polarization as well as the poor ED desalination performance and high energy consumption. Herein, the i–v characteristic was used to investigate the concentration polarization of the membranes and the limiting operating parameter for ED operation. The i–v curves were obtained by recording the corresponding current density with the stepwise potential difference. As shown in Fig. 6, all curves show the typical three characteristic stages: the beginning stage follows an approximately ohmic behavior, subsequently, the second plateau stage is reached resulting from the concentration polarization near the membrane interface, followed by the third region of rapid current increase. The experimental LCD can be estimated from intersection of the tangents from first and second regions [32]. To get precious LCD values from the curves, the derivative dE/di is derived from the original i–v curves, and plotted as a function of current density and E in Fig. 6b and c, respectively. The corresponding LCD values are listed in Table 5. All solvent-free crosslinked membranes exhibit excellent LCD in the range of 20.1–22.8 mA cm 2, which are much higher than commercial Neosepta AMX (14.2 mA cm 2). According to the concentration polarization theory, theoretical LCD can be expressed in terms of diffusion boundary layer thickness (δ), diffusion coefficient (D), and transport number (listed in Table 3):
122
Q. Pan et al. / Journal of Membrane Science 515 (2016) 115–124
Fig. 6. Current-voltage curve (a), derivative dE/di as a function of current density (b), and derivation dE/di as a function of potential (c) of the membranes measured at room temperature in 0.1 M NaCl.
ilim=
zi C0 FD δ ( ti′−ti )
(8)
The Eq. (8) clearly shows that the LCD depends on the transport number while other parameters being considered constant in this work for the same experimental condition. Note that the δ was considered constant in this work as this value is a function of hydrodynamic property such as feed flow or stirring rate which was also controlled in the experiment. Accordingly, the LCD values (listed in Table 5) are higher than that of commercial membrane for the lower transport numbers (listed in Table 3). However, the data in Tables 3 and 4 show that the crosslinked membranes with higher transport number exhibit the higher LCD. For example, M-0.75 with a transport number of 0.95 exhibits the highest LCD of 22.8 mA cm 1. As discussed earlier on ε value, M-0.75 exhibits a highly ordered distribution between conducting and non-conducting regions on membrane surface. It is likely to balance the mass transfer in both boundary layer and membrane matrix, and finally result in the concentration polarization occurring at a high current density (high LCD). 3.5. Desalination performance of the solvent-free crosslinked membranes Continuous-mode ED test has been performed to evaluate the desalination performance of M-0.25, M-0.5, M-0.75 and M-1, while commercial Neosepta AMX membrane being tested at the same condition for comparison. The feed solution (0.1 mol L 1 NaCl) is pumped into the dilute chamber with a constant flow rate of 30 cm3 min 1 and a direct current of 0.21 A (corresponding current density of 30 mA cm 2). The change in conductivity of NaCl solution in dilute cell and potential over the stack during the test are recorded. It is worth mentioning that the applied current density (30 mA cm 2) is higher than LCDs of membranes according to Fig. 6 and Table 5, indicating that concentration polarization would happen during the desalination process. As shown in Fig. 7, all desalination curves can be divided into two stages. In the first time range of 0–75 min, all membranes exhibit similar desalination rate. However, in the following stage, M-0.5 and M-0.25 membranes show significantly higher rates than others. The result is according with that from chronopotentiograms. As shown in Fig. 5, Neospta AMX and M-0.75 reach LCD in a much shorter time than M-0.5 and M-0.25, suggesting that their concentration polarization happen earlier than that of M-0.5 and
Fig. 7. The change of conductivity in dilute cell during electrodialysis test.
M-0.25. According to Fig. 7, concentration polarization is supposed to be not occurred in the time period of 0–75 min. Desalination is driven by the concentration difference between the dilute and concentrate cells. Ions transport is balanceable in both boundary layer and membrane matrix. Accordingly, all membranes show comparable desalination rates in this time period. However, in the following time stage, concentration polarization of Neospta AMX and M-0.75 happen prior to that of M-0.5 and M-0.25. Ions transport across the membrane is much faster than that in boundary layer, resulting in a lower desalination rate. Moreover, it is interesting to note that M-0.5 exhibits a higher desalination rate than N-0.25 in the second time stage. The homogeneous ionic crosslinking networks in M-0.5, which endow membranes lower area resistance and higher LCD, are responsible for it. Hence, it can be concluded that the solvent-free crosslinking strategy presents significantly advantage of maintaining high ionic content, whilst minimizing increase in membrane area resistance and decrease in LCD. Accordingly, the resultant crosslinked membrane, e.g. M-0.5, is expected to have a great potential in ED application while overall consideration of their all properties. 4. Conclusion In summary, we herein present our current efforts toward
Q. Pan et al. / Journal of Membrane Science 515 (2016) 115–124
developing upgradable solvent-free strategy for crosslinked AEMs. This approach, distinct from the classical post-quaternization processes, performs crosslinking and quaternization simultaneously. Particularly, imidazolium cations locating on cross-links endow the resultant membrane a high IEC (over 2.0 mmol g 1) for target low area resistance (lower than 1.4), while maintain a high ion selective transport property. The resulting membranes exhibit much higher limiting current densities (over 20 mA cm 2) than commercial AEM (Neosepta AMX, 14.2 mA cm 2, measured at the same condition). Accordingly, when being applied in ED application, the optimized M-0.5 membrane exhibits higher desalination efficiency than commercial Neosepta AMX, suggesting its potential application in ED.
Acknowledgements We thank the National Natural Science Foundation of China (Nos. 91534203, 21376232, 21490581), National High Technology Research and Development Program 863 (2015AA03A601) and One Hundred Person Project of the Chinese Academy of Sciences (2014, D type).
Nomenclature AEM BPPO
anion exchange membrane brominated poly (2, 6-dimethyl-1, 4-phenylene oxide) VBC 4-vinylbenzyl chloride IEM ion exchange membrane ED electrodialysis CEM cation exchange membrane TMA trimethylamine IEC ion exchange capacity N-VI N-vinylimidazole N-MI N-methylimidazole NMP N-Methyl pyrrolidone DMf N, N-dimethylformamide DMSO dimethyl sulfoxide TGA thermogravimetric analysis DTG differential thermal gravity mwet mass of hydrated membrane (g) mdry mass of dehydrated membrane (g) Rsol electrolyte resistance (Ω cm2) Rcell resistance of membrane equilibrated in electrolyte solution (Ω cm2) Rmem membrane resistance (Ω cm2) Em membrane potential A effective area of the membranes (cm2) R gas constant (J K 1 mol 1) F Faraday constant (C mol 1) C1 and C2 concentrations of electrolyte solutions t′i transport number of the counter-ion in the membrane ti transport number of the counter-ion in free solution DC direct current IPN interpenetrating polymer network τ transition time (s) i current density (mA cm 2) Zi valence of the ith ion π mathematical constant (3.141) D diffusion coefficient (cm2 s 1)
ΔE Δt i*
ε ilim
δ
Ror
123
plateau length (V) interval time (s) local current density in the conducting regions (mA cm 2) fraction of conducting region and LCD limiting current density (mA cm 2) diffusion boundary layer thickness (cm) resistance in ohmic region (Ω cm2)
References [1] W. Juda, W.A. McRae, Coherent ion-exchange gels and membranes, J. Am. Chem. Soc. 72 (1950) 1043–1044. [2] M.G. Buonomenna, Membrane processes for a sustainable industrial growth, RSC Adv. 3 (2013) 5694–5740. [3] L. Jingyi, W. Cuiming, X. Tongwen, W. Yonghui, Diffusion dialysis-concept, principle and applications, J. Membr. Sci. 366 (2011) 1–16. [4] D.H. Kim, A review of desalting process techniques and economic analysis of the recovery of salts from retentates, Desalination 270 (2011) 1–8. [5] H. Strathmann, A. Grabowski, G. Eigenberger, Ion-exchange membranes in the chemical process industry, Ind. Eng. Chem. Res. 52 (2013) 10364–10379. [6] L. Hong-Joo, H. Min-Kyoung, H. Sang-Don, M. Seung-Hyeon, Influence of the heterogeneous structure on the electrochemical properties of anion exchange membranes, J. Membr. Sci. 320 (2008) 549–555. [7] B.M. Asquith, J. Meier-Haack, C. Vogel, W. Butwilowski, B.P. Ladewig, Sidechain sulfonated copolymer cation exchange membranes for electro-driven desalination applications, Desalination 324 (2013) 93–98. [8] C.G. Arges, J. Parrondo, G. Johnson, A. Nadhan, V. Ramani, Assessing the influence of different cation chemistries on ionic conductivity and alkaline stability of anion exchange membranes, J. Mater. Chem. 22 (2012) 3733–3744. [9] K.K. Stokes, J.A. Orlicki, F.L. Beyer, RAFT polymerization and thermal behavior of trimethylphosphonium polystyrenes for anion exchange membranes, Polym. Chem. 2 (2011) 80–82. [10] S. Gu, R. Cai, Y.S. Yan, Self-crosslinking for dimensionally stable and solventresistant quaternary phosphonium based hydroxide exchange membranes, Chem. Commun. 47 (2011) 2856–2858. [11] S. Gu, R. Cai, T. Luo, Z.W. Chen, M.W. Sun, Y. Liu, G.H. He, Y.S. Yan, A soluble and highly conductive ionomer for high-performance hydroxide exchange membrane fuel cells, Angew. Chem. Int. Ed. 48 (2009) 6499–6502. [12] K.J.T. Noonan, K.M. Hugar, H.A. Kostalik, E.B. Lobkovsky, H.D. Abruna, G. W. Coates, Phosphonium-functionalized polyethylene: a new class of basestable alkaline anion exchange membranes, J. Am. Chem. Soc. 134 (2012) 18161–18164. [13] D.S. Kim, A. Labouriau, M.D. Guiver, Y.S. Kim, Guanidinium-functionalized anion exchange polymer electrolytes via activated fluorophenyl-amine reaction, Chem. Mater. 23 (2011) 3795–3797. [14] J.H. Wang, S.H. Li, S.B. Zhang, Novel hydroxide-conducting polyelectrolyte composed of an poly(arylene ether sulfone) containing pendant quaternary guanidinium groups for alkaline fuel cell applications, Macromolecules 43 (2010) 3890–3896. [15] X.C. Lin, L. Wu, Y.B. Liu, A.L. Ong, S.D. Poynton, J.R. Varcoe, T.W. Xu, Alkali resistant and conductive guanidinium-based anion-exchange membranes for alkaline polymer electrolyte fuel cells, J. Power Sources 217 (2012) 373–380. [16] G.G. Wang, Y.M. Weng, D. Chu, R.R. Chen, D. Xie, Developing a polysulfonebased alkaline anion exchange membrane for improved ionic conductivity, J. Membr. Sci. 332 (2009) 63–68. [17] J. Fang, P.K. Shen, Quaternized poly(phthalazinon ether sulfone ketone) membrane for anion exchange membrane fuel cells, J. Membr. Sci. 285 (2006) 317–322. [18] G.G. Wang, Y.M. Weng, D. Chu, D. Xie, R.R. Chen, Preparation of alkaline anion exchange membranes based on functional poly(ether-imide) polymers for potential fuel cell applications, J. Membr. Sci. 326 (2009) 4–8. [19] T.W. Xu, Ion exchange membranes: state of their development and perspective, J. Membr. Sci. 263 (2005) 1–29. [20] M.R. Hibbs, C.H. Fujimoto, C.J. Cornelius, Synthesis and characterization of poly (phenylene)-based anion exchange membranes for alkaline fuel cells, Macromolecules 42 (2009) 8316–8321. [21] T.J. Clark, N.J. Robertson, H.A. Kostalik, E.B. Lobkovsky, P.F. Mutolo, H. D. Abruna, G.W. Coates, A ring-opening metathesis polymerization route to alkaline anion exchange membranes: development of hydroxide-conducting thin films from an ammonium-functionalized monomer, J. Am. Chem. Soc. 131 (2009) 12888–12889. [22] N.J. Robertson, H.A. Kostalik, T.J. Clark, P.F. Mutolo, H.D. Abruna, G.W. Coates, Tunable high performance cross-linked alkaline anion exchange membranes for fuel cell applications, J. Am. Chem. Soc. 132 (2010) 3400–3404. [23] J.H. Wang, Z. Zhao, F.X. Gong, S.H. Li, S.B. Zhang, Synthesis of soluble poly (arylene ether sulfone) ionomers with pendant quaternary ammonium groups for anion exchange membranes, Macromolecules 42 (2009) 8711–8717.
124
Q. Pan et al. / Journal of Membrane Science 515 (2016) 115–124
[24] Z. Zhenghui, W. Liang, J. Varcoe, L. Chuanrun, O. Ai Lien, S. Poynton, X. Tongwen, Aromatic polyelectrolytes via polyacylation of pre-quaternized monomers for alkaline fuel cells, J. Mater. Chem. A 1 (2013) 2595–2601. [25] L.A. Wu, G.F. Zhou, X. Liu, Z.H. Zhang, C.R. Li, T.W. Xu, Environmentally friendly synthesis of alkaline anion exchange membrane for fuel cells via a solvent-free strategy, J. Membr. Sci. 371 (2011) 155–162. [26] T.W. Xu, W.H. Yang, Fundamental studies of a new series of anion exchange membranes: membrane preparation and characterization, J. Membr. Sci. 190 (2001) 159–166. [27] C. Larchet, S. Nouri, B. Auclair, L. Dammak, V. Nikonenko, Application of chronopotentiometry to determine the thickness of diffusion layer adjacent to an ion-exchange membrane under natural convection, Adv. Colloid Interface Sci. 139 (2008) 45–61. [28] L. Wu, Q. Pan, J.R. Varcoe, D. Zhou, J. Ran, Z.J. Yang, T.W. Xu, Thermal
[29]
[30]
[31]
[32]
crosslinking of an alkaline anion exchange membrane bearing unsaturated side chains, J. Membr. Sci. 490 (2015) 1–8. C. Klaysom, R. Marschall, S.-H. Moon, B.P. Ladewig, G.Q.M. Lu, L. Wang, Preparation of porous composite ion-exchange membranes for desalination application, J. Mater. Chem. 21 (2011) 7401. P.V. Vyas, P. Ray, S.K. Adhikary, B.G. Shah, R. Rangarajan, Studies of the effect of variation of blend ratio on permselectivity and heterogeneity of ion-exchange membranes, J. Colloid Interface Sci. 257 (2003) 127–134. J.H. Choi, S.H. Kim, S.H. Moon, Heterogeneity of ion-exchange membranes: the effects of membrane heterogeneity on transport properties, J. Colloid Interface Sci. 241 (2001) 120–126. J.H. Choi, H.J. Lee, S.H. Moon, Effects of electrolytes on the transport phenomena in a cation-exchange membrane, J. Colloid Interface Sci. 238 (2001) 188–195.
Contents lists available at ScienceDirect
Journal of Membrane Science journal homepage: www.elsevier.com/locate/memsci
One-pot solvent-free synthesis of cross-linked anion exchange membranes for electrodialysis Qi Pan a, Md. Masem Hossain a, Zhengjin Yang a, Yaoming Wang a,b, Liang Wu a,b,n, Tongwen Xu a,nn a CAS Key Laboratory of Soft Matter Chemistry, Collaborative Innovation Center of Chemistry for Energy Materials, School of Chemistry and Materials Science, University of Science and Technology of China, Hefei 230026, PR China b Hefei Chemjoy Polymer Materials Co. Ltd., Hefei 230601, PR China
art ic l e i nf o
a b s t r a c t
Article history: Received 8 January 2016 Received in revised form 27 May 2016 Accepted 28 May 2016 Available online 31 May 2016
Solvent-free strategy has attracted a broad interest in preparation of ion exchange membranes in recent years stemming from the no need of organic solvents and facilitated construction of highly crosslinking networks. However, post-functionalization is commonly required for further introducing ion-exchange groups. In this study, we report a one-pot solvent-free synthesis of cross-linked anion exchange membranes (AEMs) without the need for post-functionalization. We firstly dissolved brominated poly (2, 6dimethyl-1, 4-phenylene oxide) (BPPO) in liquid monomers mixture of 4-vinylbenzyl chloride (VBC) and styrene without any organic solvent, then added appropriate amount of N-vinylimidazole and N-methylimidazole to introduce imidazolium groups into both of polyelectrolyte and solvents. The transparent, robust and crosslinked AEMs were obtained by the thermal crosslinking of the unsaturated moieties during the membrane forming process. This approach, distinct from the classical post-functionalization processes, performs crosslinking and functionalization simultaneously. Particularly, imidazolium cations locating on cross-links endow the resultant membrane a high concentration of charge carriers (ionexchange capacity) for target low resistance, while maintain a high mechanical stability. & 2016 Elsevier B.V. All rights reserved.
Keywords: Anion exchange membrane Solvent-free Electrodialysis Crosslinking Desalination
1. Introduction Ion-exchange membranes (IEMs) are polymer electrolytes that conduct anions or cations, as they contain counter charged groups bound covalently to a polymer backbones [1]. Over the last decade or so, there is an increasing worldwide interest in the use of IEMs on a large industrial scale in processes such as electrodialysis, electrodeionization, diffusion dialysis, and energy conversion and storage systems [2–5]. Among them, electrodialysis (ED), in which a series of alternating anion- and cation-exchange membranes arranged between two electrodes, is today one of the most stateof-the-art ion-exchange membrane process. It has found industrial scale application in water desalination and recovery of useful substances from industrial waste water [3]. As the key component in ED devices, cation-exchange membranes (CEMs) have successfully been commercialized for the reliable manufacturing n Corresponding author at: CAS Key Laboratory of Soft Matter Chemistry, Collaborative Innovation Center of Chemistry for Energy Materials, School of Chemistry and Materials Science, University of Science and Technology of China, Hefei 230026, PR China. nn Corresponding author. E-mail addresses: [email protected] (L. Wu), [email protected] (T. Xu).
http://dx.doi.org/10.1016/j.memsci.2016.05.050 0376-7388/& 2016 Elsevier B.V. All rights reserved.
technique, however, anion-exchange membranes (AEMs) still face technique challenges in large-scale preparation and further improving their performance, i. e. high ions selective permeability and stability [6,7]. AEMs have a selective permeability for anions as they contain fixed positive charged ions, such as ammonium, phosphonium, guanidinium ions [8–15]. Traditional preparation of AEMs usually employs the post-modification of pristine polymers or direct polymerization of functionalized monomers. For post-modification method, the most common approach is to carry out a chloromethylation or bromomethylation on the polymer so as to further cationize them with trimethylamine. A variety of polymers, such as polysulfone [16], poly (phthalazinone ether sulfone ketone) [17], poly(ether-imide) [18], poly(phenylene oxide) [19], poly (phenylene) [20], could act as scaffolds for preparing AEMs. The direct polymerization is usually performed by polymerization of monomers with moieties bearing ammonium, halomethyl-containing or amine-containing groups, that can be conveniently converted into anionic exchange groups in organic solvent. Coates et al. reported their work on the facile synthesis and ring-opening metathesis polymerization of a tetraalkylammonium-functionalized norbornene with dicyclopentadiene as a cross-linkable monomer to yield strong AEMs [21,22]. Zhang et al. reported that
116
Q. Pan et al. / Journal of Membrane Science 515 (2016) 115–124
partially fluorinated poly (arylene ether sulfone) with pendant ammonium groups were prepared by copolymerization of bisphenol monomers containing amine groups and partially fluorinated monomers, followed by quaternization [23]. We have reported preparation of high performance AEM by polyacylation of pre-quaternized monomers to endow the resultant AEMs with excellent hydroxide ion conductivity [24]. Although these strategies represent a promising progress in the development of AEMs in recent years, they all require excess reagents and long reaction time to obtain highly functionalized polymers. Particularly, the large amount of organic solvents in both reaction and membrane formation processes will carries toxic risk to the environment. Hence, to further industrial-scale manufacture, it is important to develop simple, rapid and environmental friendly methods for production of AEMs with excellent performances. To overcome these obstacles, we previously developed a solvent-free strategy for AEMs by the in-situ polymerization [25]. Different from the above mentioned post-modification and direct polymerization, it starts from the dissolving of polymer electrolytes by liquid monomers, other than organic solvent, followed by in-situ polymerization to obtain un-charged membranes. It should be noted that post-quaternization was required to convert halomethyl groups to ammonium cations by immersing base membrane into trimethylamine (TMA) aqueous solution. The disadvantages of this strategy include: a) the reaction efficiency is very low due to the solid-liquid heterogeneous reaction. b) The quality of the membrane is hardly to be controlled for different production batch, and membranes usually exhibit a loose morphology due to the erosion of TMA. In view of the above mentioned issues, we herein presents our current efforts toward developing upgradable solvent-free strategy for AEMs. This approach, distinct from the classical post-quaternization processes, performs crosslinking and quaternization simultaneously. We firstly dissolved brominated poly (2, 6-dimethyl-1, 4-phenylene oxide) (BPPO) in liquid monomers mixture of 4-vinylbenzyl chloride (VBC) and styrene without any organic solvent, then added appropriate amount of N-vinylimidazole and N-methylimidazole to introduce imidazolium cations into both of polyelectrolyte and solvents. The transparent, robust and crosslinked AEMs were obtained by the thermal crosslinking of the unsaturated moieties during the membrane forming process. Particularly, the imidazolium cations locating on cross-links endow the resultant membrane a high concentration of charge carriers for target high ions permeability, while maintain a high mechanical stability. Properties characterizations, such as ion exchange capacity (IEC), membrane area resistance (Rm), electrodialysis related performance, etc., are discussed extensively to show the advantages of the solvent-free strategy over the conventional post-modification method.
2. Experimental methods 2.1. Materials Brominated poly (2, 6-dimethyl-1, 4-phenylene oxide) (BPPO) was kindly supplied by Tianwei Membrane Corporation Ltd. of Shandong (China). 4-Vinylbenzyl chloride (VBC), which was purchased from Changzhou Wujin Linchuan Chemical Co. Ltd. N-vinylimidazole (N-VI, 98%) and N-methylimidazole (N-MI, 99%) were purchased from Energy Chemical Co. Ltd. Styrene, sodium sulfate (Na2SO4), sodium chloride (NaCl), N-Methyl pyrrolidone (NMP, AR grade), N, N-dimethylformamide (DMF, AR grade), chlorobenzene (CB, AR grade), dimethyl sulfoxide (DMSO, AR grade) were purchased from Shanghai-Sinopham Chemical Reagent Co. Ltd. (China). Commercial anion exchange membrane Neosepta AMX
(thickness 120–180 mm, IEC 1.4–1.7 meq g 1, area resistance 2.0– 3.5 Ω cm2) and cation exchange membrane Neosepta CMX (thickness 220–260 mm, IEC 1.5–1.8 meq g 1, area resistance 2.0–3.5 Ω cm2) were purchased from ASTOM, Japan. 2.2. Membranes preparation As shown in Scheme 1, BPPO (1 g/0.00298 mol –CH2Br) was dissolved in a liquid monomers mixture; then appropriate amount of N-Vinylimidazole was added to convert all bromomethyle of BPPO and half amount of chloromethyl groups of VBC to imidazolium cations. The remaining chloromethyl groups were converted to imidazolium cations by adding N-Methylimidazole dropwise. The composition of reaction mixture was determined according to the solubility of polymer electrolytes (BPPO) in the liquid monomers. In this study, the content of BPPO (1 g) keeps constant for the preparation of all membranes. Styrene, which is employed as solvent, need a critical content (18 ml) to completely dissolve BPPO. Hence, the contents of BPPO and styrene are kept constant during the preparation. The content of VBC was varied to tune the amount of functional groups in membranes. Accordingly, the contents of N-VI and N-MI were varied with the change of VBC content to convert functional groups (–CH3Br and –CH3Cl) to imidazolium cations, also to keep the membrane with an appropriate crosslinking degree. Table 1 shows the components of the different reaction mixtures. The resulting membranes are designated as M-x, where x represents the amount of VBC (0.25 ml. 0.5 ml, 0.75 ml and 1.0 ml). Before characterization, all membranes were immersed in ether at room temperature for 4 h to remove unreacted monomers (styrene, VBC, N-VI and N-MI, which are all soluble in ether), followed by thorough washing with deionized water to remove imidazolium salts (possibly resulting from the reaction of unreacted VBC with N-VI and N-MI). 2.3. Membranes characterizations 2.3.1. Fourier transform infrared spectroscopy (FT-IR) FT-IR spectra were measured using a Vector 22 Fourier transform infrared spectrometer (Bruker). 2.3.2. Measurement of gel content Gel content was measured by calculating the ratio of mass remaining after immersing the membrane in organic solvent for 7 days at room temperature, typical procedure is as follows: a) Four pieces of membranes (2 4 cm, BPPO-0.25, BPPO-0.5, BPPO-0.75, BPPO-1.0)were dried at 60 °C for 24 h, then weighted (m0). b) The dried membranes were soaked in NMP, DMSO, DMF and CHCl3 at room temperature for 7 days, respectively. After that, they were washed and immersed in distilled water for 1 day. c) Finally, they were dried at 60 °C for 24 h, and weighed (mt). The gel content was calculated according to the following equation:
DC =
mt × 100% m0
(1)
2.3.3. Thermal analysis TGA and DTG thermograms were recorded using a TGA Q5000 V3.15 analyzer at a heating rate of 10 °C min 1 under a N2 atmosphere in order to evaluate the short-term thermal stability of membrane.
Q. Pan et al. / Journal of Membrane Science 515 (2016) 115–124
117
Scheme 1. Schematic diagram of the preparation process.
Table 1 Membrane compositions. Sample
BPPO/g
Styrene/ml
VBC/ml
N-VI/ml
N-MI/ml
M-0.25 M-0.5 M-0.75 M-1
1 1 1 1
1.8 1.8 1.8 1.8
0.25 0.5 0.75 1.00
0.35 0.45 0.53 0.62
0.075 0.15 0.23 0.30
2.3.4. Ion exchange capacity The ion exchange capacities (IEC) were measured according to our previous work [26]. The samples were firstly converted into Cl form after immersing in the NaCl aqueous solution (0.5 mol L 1) for 24 h. Then, the samples were washed and immersed in distilled water for 24 h. After that, the samples were dried to a constant weight and then immersed in a Na2SO4 aqueous solution (0.5 mol L 1) for 24 h. The concentration of Cl
released from the membrane was measured using a REX PHS-3C analyzer (China) and an attach Cl ion-selective electrode (model pCl-1). The IEC value was calculated from the measured amounts released Cl ions and is expressed as mmol g 1 of the dry membrane. 2.3.5. Water uptake Water uptake was calculated by taking the difference between the wet weight (mwet) and the dry weight (mdry) of the AEMs. To obtain the wet weight, AEM samples were equilibrated in distilled water at room temperature for 2 days. The water uptake was calculated according to the following equation:
WR=
mwet −mdry × 100% mdry
(2)
118
Q. Pan et al. / Journal of Membrane Science 515 (2016) 115–124
2.3.6. Membrane area resistance Membrane area resistance was measured using a commercial cell-assembly ((CJ-MPMD-1, Hefei Chemjoy Co., Ltd., China)) under a constant current mode [27]. It is composed of five compartments: two electrode chambers at both ends of the cell; two intermediate compartments equipped with two reference electrodes; one quadrate clip for membrane. Particularly, two pieces of Nafion-117 membranes were assembled between electrode and intermediate chambers, separately, to eliminate the influence of electrode reaction. Titanium (Ti) coated stainless steel sheets were used as cathode and anode. Two intermediate chambers are separated by the membrane clip, and tips of the reference electrodes should be closely to the center of membrane. During the measurement, Na2SO4 solution (0.3 mol L 1) was fed to electrode chambers, and NaCl (0.5 mol L 1) is fed to intermediate chambers. A constant current is supplied by a direct current power supply (SHEKONIC, Yangzhou Shuanghong Co., Ltd.) and the potential between electrodes is read by a digital multimeter (model: GDM8145, Good will instrument Co. Ltd., Taiwan). Then, by subtraction of the electrolyte resistance (Rsol) from the membrane resistance equilibrated in the electrolyte solution (Rcell), the membrane resistance (Rmem) was calculated according to the equation Rmem ¼Rcell Rsol, and final membrane area resistance was calculated as follows:
R = Rmem × A
(3)
where Rmem is the resistance of membranes (Ω), A is the effective area for membrane (7 cm2). 2.3.7. Membrane potential and static transport number The membrane potential (Em) and the transport number t were determined using the same device for membrane resistance. Two intermediate chambers, which are equipped with two reference electrodes, are separated by the membrane clip. Tips of the electrodes should be closely to the center of membrane. During the measurement, continuous flows of aqueous KCl (0.01 mol/L) and KCl (0.05 mol/L) were pumped into intermediate chambers, separately. The potential difference between the electrodes (Em) was determined by a digital multimeter (model: GDM-8145, Good will instrument Co. Ltd., Taiwan). The static transport number ťi of the anion in the membrane can be calculated as follows:
Em=(2ti′−1)
⎛C ⎞ RT ln ⎜ 1 ⎟ ⎝ C2 ⎠ nF
Fig. 1. ED cell configuration (a) and schematic diagram of the test in ED system.
connected to the Autolab (PGSTAT 30, Eco Chemie, Netherland). The chronopotentiogram was carried out at a constant applied current density of 3.5 mA cm 2 and the corresponding potential was recorded in every 0.01 s for a period of 200 s, automatically. 2.3.9. Continuous-mode electrodialysis (ED) A schematic diagram of continuous-mode ED stack is illustrated in Fig. 1. It consists of five compartments. Each chamber provides 7.07 cm2 active area for membranes with an o-ring to prevent leakage during the testing. There were three streams in the ED stack, dilute, concentrated and electrode wash. ED experiments were conducted in the recirculation mode of operation, in which each stream (150 cm3) was recirculated. Typically, a feed solution (0.1 mol L 1 NaCl) is pumped into the dilute and concentrate cells with a constant flow rate of 30 ml min 1. Meanwhile, two electrode cells were fed by 0.3 mol L 1 Na2SO4 solution and were connected together to prevent pH change. The commercial cation-exchange membrane Neosepta CMX were placed at the position C1, C2 and C3 in Fig. 1. The tested membrane was placed at the position A.
(4)
3. Results and discussion
where R is universal gas constant, F is the Faraday constant, T is the room temperature, n is the electrovalence (1 in this case), and C1 and C2 are the concentrations of electrolyte solutions in the testing cell, respectively.
3.1. Membrane preparation
2.3.8. Current-voltage (i-v) characteristic and choronopotentiometry The i-v curves of the membranes were measured at room temperature using the same three-compartment cell in 0.1 M NaCl. The choronopotentiograms of the membranes were measured using the same device without auxiliary membranes in 0.025 M NaCl. For i-v curves measurement, the cell was connected to a DC power supply (Shekonic, China, Model WWL-LDX) by the Pt electrodes and to the multimeter by the reference electrodes. Experiments were conducted in recirculation mode of operation, in which each stream was recirculated with a flow rate of 30 ml min 1. 1.5 wt% Na2SO4 was used as an electrode rinse solution and circulated in electrode compartments from a single reservoir. 0.1 M NaCl electrolyte solution was recirculated in the central compartment. A stepwise current was applied and the corresponding potential difference across the membrane was measured. For chronopotentiogram measurement, the cell was
Solvent-free method is widely considered as one of the most state-of-art strategy for preparing IEMs due to the advantages of easily forming interpenetrating polymer networks (IPNs) and no need of organic solvent. It is commonly performed by dissolving polymer electrolytes in liquid monomers followed by in-situ polymerization. However, post-modifications, e.g. quaternary amination or sulfonation, are required for introducing ion-exchange groups, and usually cause unexpected corrosion to polymer electrolytes. To overcome this shortcoming, this study presents an alternative one-pot solvent-free strategy with no need of post-modification process. As shown in Scheme 1a, this approach, distinct from the classical post-functionalization processes, performs crosslinking and quaternization simultaneously. we firstly dissolved brominated poly (2, 6-dimethyl-1, 4-phenylene oxide) (BPPO) in liquid monomers mixture of 4-vinylbenzyl chloride (VBC) and styrene without any organic solvent, then added appropriate amount of N-Vinylimidazole and N-methylimidazole to introduce imidazolium cations into both of polyelectrolyte and solvents. The
Q. Pan et al. / Journal of Membrane Science 515 (2016) 115–124
transparent, robust and crosslinked AEMs were obtained by the thermal crosslinking of the unsaturated moieties during the membrane forming process. Styrene, which is a favorable solvent as VBC, is the crucial component during the process. The resulting membrane would swell excessively by water and brittle into pieces if no styrene being involved (as shown in Scheme 1f) for the large amount of hydrophilic imidazolium cations converted from chloromethyl groups in VBC. Hence, an appropriate amount of styrene can maintain the content of imidazolium cations in a proper level and significantly enhance the toughness of the resulting membrane. Moreover, the stepwise adding of N-VI and N-MI is another key step for the ionic crosslinked membrane preparation. If only N-VI being employed, excessive amount of vinyl groups would be introduced to BPPO and VBC. The resultant membrane is brittle for the excessive crosslinking. On the contrary, the resultant membrane would be highly water swelling if only employing N-MI. To prepare membrane with a proper crosslinking degree, an appropriate content of N-VI was firstly required to convert all bromomethyle of BPPO and half amount of chloromethyl groups of VBC to imidazolium cations. Then the rest of chloromethyl groups of VBC were converted to imidazolium cations by adding an appropriate content of N-MI. As shown in Scheme 1e, we perform performed the following polymerization at 100 °C according to our previous study on self-crosslinked membrane. It was performed by attaching alkene pendant groups onto polymer mainchains via the Menshutkin reaction, followed by thermal crosslinking of the unsaturated side chains during the membrane formation process [28]. In-situ FTIR analysis indicated that the self-crosslinking of the pendant alkene groups was temperature dependence. The intensity of the alkene peaks decreases while increasing temperature treatment (30–100 °C), suggesting that the polymerization of alkene groups happens during heating; the higher temperature, the higher crosslinking degree. In this study, after temperature-curing at 100 °C, a yellowish, transparent and flexible AEM was obtained and easily being cut in any shape. 3.2. Investigation of the crosslinking process FTIR analysis was performed to confirm the quaternization reaction. Fig. 2 shows the FTIR spectra of the pure BPPO membrane, and M-0.25, M-0.5, M-0.75 and M-1, respectively. All solvent-free membranes show strong bands at 3060 cm 1 which are attributed to –HC ¼ N in imidazolium group, suggesting that quaterization reaction occurs successfully during membranes
119
preparation. Furthermore, comparing with the FTIR spectrum of BPPO, new peaks at 3425 cm 1 appear in the spectra of solventfree membranes. These peaks are attributed to the vibration of the hydroxyl groups in water molecules bonded to hydrophilic imidazolium groups, further indicating the quaterizaiton reaction. According to the reaction, the crosslinking generates methylene groups. The overlapped stretching vibration peaks in the range of 2854–2923 cm 1 are attributed to the stretching vibration of carbon hydrogen bonds in methylene (–CH2–), suggesting that crosslinking happens during membrane preparation. However, there are still residual alkene groups (–CH ¼ CH2) in the membrane after crosslinking. As shown in the spectra of all solvent-free membranes, the double peaks at 1604 and 1653 cm 1 attributed to the vibration of the carbon carbon double bonds in the pendant alkene groups (–CH ¼CH2) were clearly detected. It should be noted that a proper crosslinking degree is expected to maintain the membrane excellent flexibility and stability. Hence, further solubility and thermal stability tests were performed to determine the advantage of the moderate crosslinking. Solubility test is a common method to determine the formation of crosslinking networks in membrane. As shown in Table 2, M-1, M-0.75 and M-0.5 membranes are insoluble in NMP, and can maintain over 90% of the mass retention in other solvents, such as DMF, DMSO and CHCl3, while pure BPPO membrane is completely soluble in NMP and CHCl3, and slightly soluble in DMF, but insoluble in DMSO. Hence, it can be concluded that the crosslinking network has constructed in the membrane successfully. In addition to the organic solvent resistance, the crosslinking network is also beneficial to the improvement of thermal stability of membrane. Fig. 3 shows the TGA curves of BPPO and crosslinked membranes. These curves exhibited three main degradation stages, arising from the processes of thermal desorption of water, thermal deamination, and thermal oxidation of the polymer matrix. Notably, original BPPO exhibit higher thermal stability than the crosslinked membranes at the temperature over 375 °C, at which crosslinks degrade rapidly. In the temperature range of 300–375 °C, crosslinked structure endows the membrane better thermal stability than BPPO. Basically, the normal working temperature of the membrane for ED application is lower than 100 °C. as shown in Fig. 3, crosslinked membranes exhibit a slightly weight loss in this temperature range of 0–100 °C, which is attributed to the loss of water molecules absorbed by hydrophilic imidazolium groups, other than the degradation of membrane. TGA result suggests that the crosslinked membranes exhibit sufficient thermal stability at temperature below 100 °C for ED application. 3.3. Electrochemical properties of the membranes The content of charged functional groups (commonly reflected as IEC) in an IEM plays an important role in providing hydrophilic environments for ions transport. This study performs crosslinking and quaternization simultaneously. Particularly, hydrophilic imidazolium groups locating on cross-links is expected to endow the Table 2 Solubility properties of solvent-free crosslinked membranes. NMP M-1 M-0.75 M-0.5 M-0.25 BPPO
Fig. 2. ATR-FTIR spectra of BPPO membrane and solvent-free membranes.
a
100% 100% 100% 95.2% 0%
DMF
DMSO
CHCl3
98.5% 97.2% 94.8% 79.6% 17.4%
97.8% 98.9% 89.1% 74.6% 99.0%
96.4% 93.3% 92.1% 80.8% 0%
a The data indicate the percentage of the mass retention of each membrane after being immersed in organic solvent for 7 days at room temperature.
120
Q. Pan et al. / Journal of Membrane Science 515 (2016) 115–124
(in the range of 0.93–0.95). As discussed above, imidazolium cations locate on cross-links in the membrane. Hence, the membrane with higher VBC content possess higher IEC as well as much denser crosslinking networks, which is expected to resistant against water swelling and maintain a proper ions selective transport property (transport number). Moreover, M-0.75 exhibits the highest transport number of 0.95 among the crosslinked membranes. It can be explained by the effect of the distribution of conducting region on membrane surface. In addition to the dense ionic crosslinking networks, M-0.75 might possesses a highly ordered alignment of conducting regions on the surface, which is likely to endow the membrane excellent selective ions transport property. 3.4. Electrodialysis related performances of the membrane
Fig. 3. TGA and DTG (inset) traces for the solvent-free membranes recorded with a heating rate of 10 °C min 1 and under flowing nitrogen.
Fig. 4. IEC and water uptake of M-1, M-0.75, M-0.5 and M-0.25.
Table 3 Transport number and area resistance of the solvent-free membranes and commercial AEM (Neosepta AMX). Name
Transport number
Area resistance (Ω/cm2)
M-0.25 M-0.5 M-0.75 M-1 Neosepta AMX
0.93 0.93 0.95 0.94 0.98
1.40 1.12 1.12 0.98 2.52
Electrodialysis is generally operated under direct current which provides the driving force for ions transport through membrane. However, concentration polarization often happens at the interface between an IEM and an electrolyte solution (boundary layer) when an electric current passes through the system, leading to the performance degradation of the ED device. Chronopotentiometry is a powerful tool for studying transport phenomena and concentration polarization near membrane interfaces [29]. In this study, it was used for investigating the polarization phenomena and the micro-scale distribution of conducting and non-conducting regions on the membrane surface. The typical chronopotentiograms with three-stage potential differences were obtained and shown in Fig. 5. After a constant current density of 3.5 mA cm 2 is applied to the system, a potential (E0) onset was observed due to the resistance of the membrane and working solution. Corresponding to the concentration polarization, this initial potential gradually increased for a certain period because the counter ions pass through IEMs and gather on the other side of the membrane interface. The potential significantly increased and finally reached a steady state when the ionic concentration of depleting solution nearby the membrane interface dropped to zero [30]. The difference between the initial potential and the maximum potential from chronopotentiograms (ΔE) was believed to relate indirectly to the thickness of the concentration gradient established at the membrane interfacial zone. The time at which the potential transition occurs is called the transition time (τ) and is determined by the intersection of the tangents with the first and second stages of the curve. It can also be defined by the well-known Sand's equation based on the assumption of entire conducting surface of the membrane.
τ=
( C0 zi F )2πD 2 4i2 ( ti′ − ti )
(5) 2
resultant membrane a high concentration of charge carriers (IEC) for target low resistance, while maintain a high ions selective transport property. Fig. 4 presents the change of IEC and water uptake along with the varying of VBC content. As expected, the higher VBC content, the more imidazolium groups were introduced into membrane and finally resulting in a higher IEC and water uptake. Accordingly, the membrane area resistances decrease from 1.4 Ω cm2 to 0.98 Ω cm2, which are much lower than that of commercial membrane (Neosepta AMX, 2.52 Ω cm2, measured at the same conditions). Basically, a high IEC often leads to excessive membrane swelling on hydration, and tends to cause a poor selective ions transport property (usually reflected by transport number). However, as shown in Table 3, the crosslinked membranes with different IECs exhibit similar transport numbers
where i is the current density (mA cm ), C0 is the concentration of electrolyte, Zi is the valence of the ith ion, F is the Faraday constant and D is the diffusion coefficient (cm2 s 1). t'i the transport number of the counter-ion in the membrane and ti is that of the same ion in free solution at the same concentration. The transition times deprived from chronopotentiogram (Fig. 5b) are listed in Table 4 and compared with the theoretical values from Eq. (5). It clearly shows that the experimental data are slightly lower than the corresponding theoretical ones. It is worth mentioning that the Eq. (5) is based on the assumption of the entire conducting membrane surface. For an IEM, the non-conducting regions, which do exist on the real membrane surface, possibly result in the lower values. While comparing M-0.25, M-0.5 and M-1 in Table 4, the τ values decrease alone with the increasing of VBC contents. And the more VBC content shows
Q. Pan et al. / Journal of Membrane Science 515 (2016) 115–124
121
Fig. 5. Chronopotentiograms (a) and their derivation dE/dt curve (b) of the synthesized membranes compared with commercial membranes tested at room temperature in 0.025 M NaCl. Table 4 Characteristic values from chronopotentiograms. Membrane
τa (s)
τb (s)
ε
E0 (V)
Emax (V)
ΔE (V)
Δt (s)
M-1 M-0.75 M-0.5 M-0.25 Neosepta AMX
49.4 46.6 52.4 52.4 39.6
49.1 45.1 51.3 51.1 39.5
0.994 0.968 0.979 0.975 0.997
0.209 0.211 0.209 0.211 0.223
0.779 0.875 0.856 0.794 1.147
0.570 0.664 0.647 0.583 0.924
19 16 28 29 13
a b
Estimated from Eq. (5). Deprived experimentally from chronopotentiogram.
shorter defined transition time with smaller interval time (Δt). Considering the difference between conducting and non-conducting regions in an IEM, the local current density of the conducting regions (i*) will be higher than the overall current density (i) applied to the system [31]. i* can be expressed as a function of i and the fraction of the conducting region (ε) as shown below:
i*=
i ε
(6)
ε can be estimated by the derived Sand's equation
Thus the below [31]: 1
ε=
2iτ 2 ( t′i−ti ) 1/2
C0 Z i F ( π D)
(7)
The ε as well as other characteristic values from chronopotentiogrames are summarized in Table 4. Basically, the higher content of charged functional groups in an IEM (reflected by IEC) is supposed to result in a higher ε value. While comparing M-0.25, M-0.5 and M-1 in Table 5, the ε value increases along with the amount of VBC content in the membrane. It is in accordance with the result in Fig. 4, in which the higher VBC content, the more imidazolium groups were introduced into membrane and finally Table 5 Characteristic values from i–v curve. a
Membrane
ilim
M-1 M-0.75 M-0.5 M-0.25 Neosepta AMX
20.1 22.8 21.5 20.4 14.2
a
(mA cm 2)
Experimental LCD values recorded from Fig. 5.
ΔE (V)
Ror (Ω cm2)
0.48 0.89 1.00 1.05 0.71
9.94 7.28 9.10 11.83 14.63
resulting in a higher IEC. It is interesting to note that M-0.75 does not follow this principle. It exhibits the lowest values among the crosslinked membrane. Considering the high transport number of M-0.75, it is reasonable to explain by the effect of geometrical heterogeneity of membrane surface reported by previous study [30], which suggested that not only the ε value but also the alignment of conducting regions were needed to be taken into account for explaining the ions transport performance of the membrane. Accordingly, M-0.75 might possess highly ordered distribution between conducting and non-conducting regions on the surface even at a low ε value, further resulting in a shorter transition time (τ) as well as and interval time (Δt). In addition to the chronopotentiogram, limiting current density (LCD) under a flowing solution is another important parameter to determine the ED performance of the membrane. As well known, a boundary layer close to the desalting surface of an IEM will be formed when an electric current is applied across the membrane. When the applied current density is lower than LCD, the mass transfer is balanceable in both boundary layer and membrane matrix. However, the mass transfer across the membrane is much faster than that in boundary layer when it is higher than LCD, causing concentration polarization as well as the poor ED desalination performance and high energy consumption. Herein, the i–v characteristic was used to investigate the concentration polarization of the membranes and the limiting operating parameter for ED operation. The i–v curves were obtained by recording the corresponding current density with the stepwise potential difference. As shown in Fig. 6, all curves show the typical three characteristic stages: the beginning stage follows an approximately ohmic behavior, subsequently, the second plateau stage is reached resulting from the concentration polarization near the membrane interface, followed by the third region of rapid current increase. The experimental LCD can be estimated from intersection of the tangents from first and second regions [32]. To get precious LCD values from the curves, the derivative dE/di is derived from the original i–v curves, and plotted as a function of current density and E in Fig. 6b and c, respectively. The corresponding LCD values are listed in Table 5. All solvent-free crosslinked membranes exhibit excellent LCD in the range of 20.1–22.8 mA cm 2, which are much higher than commercial Neosepta AMX (14.2 mA cm 2). According to the concentration polarization theory, theoretical LCD can be expressed in terms of diffusion boundary layer thickness (δ), diffusion coefficient (D), and transport number (listed in Table 3):
122
Q. Pan et al. / Journal of Membrane Science 515 (2016) 115–124
Fig. 6. Current-voltage curve (a), derivative dE/di as a function of current density (b), and derivation dE/di as a function of potential (c) of the membranes measured at room temperature in 0.1 M NaCl.
ilim=
zi C0 FD δ ( ti′−ti )
(8)
The Eq. (8) clearly shows that the LCD depends on the transport number while other parameters being considered constant in this work for the same experimental condition. Note that the δ was considered constant in this work as this value is a function of hydrodynamic property such as feed flow or stirring rate which was also controlled in the experiment. Accordingly, the LCD values (listed in Table 5) are higher than that of commercial membrane for the lower transport numbers (listed in Table 3). However, the data in Tables 3 and 4 show that the crosslinked membranes with higher transport number exhibit the higher LCD. For example, M-0.75 with a transport number of 0.95 exhibits the highest LCD of 22.8 mA cm 1. As discussed earlier on ε value, M-0.75 exhibits a highly ordered distribution between conducting and non-conducting regions on membrane surface. It is likely to balance the mass transfer in both boundary layer and membrane matrix, and finally result in the concentration polarization occurring at a high current density (high LCD). 3.5. Desalination performance of the solvent-free crosslinked membranes Continuous-mode ED test has been performed to evaluate the desalination performance of M-0.25, M-0.5, M-0.75 and M-1, while commercial Neosepta AMX membrane being tested at the same condition for comparison. The feed solution (0.1 mol L 1 NaCl) is pumped into the dilute chamber with a constant flow rate of 30 cm3 min 1 and a direct current of 0.21 A (corresponding current density of 30 mA cm 2). The change in conductivity of NaCl solution in dilute cell and potential over the stack during the test are recorded. It is worth mentioning that the applied current density (30 mA cm 2) is higher than LCDs of membranes according to Fig. 6 and Table 5, indicating that concentration polarization would happen during the desalination process. As shown in Fig. 7, all desalination curves can be divided into two stages. In the first time range of 0–75 min, all membranes exhibit similar desalination rate. However, in the following stage, M-0.5 and M-0.25 membranes show significantly higher rates than others. The result is according with that from chronopotentiograms. As shown in Fig. 5, Neospta AMX and M-0.75 reach LCD in a much shorter time than M-0.5 and M-0.25, suggesting that their concentration polarization happen earlier than that of M-0.5 and
Fig. 7. The change of conductivity in dilute cell during electrodialysis test.
M-0.25. According to Fig. 7, concentration polarization is supposed to be not occurred in the time period of 0–75 min. Desalination is driven by the concentration difference between the dilute and concentrate cells. Ions transport is balanceable in both boundary layer and membrane matrix. Accordingly, all membranes show comparable desalination rates in this time period. However, in the following time stage, concentration polarization of Neospta AMX and M-0.75 happen prior to that of M-0.5 and M-0.25. Ions transport across the membrane is much faster than that in boundary layer, resulting in a lower desalination rate. Moreover, it is interesting to note that M-0.5 exhibits a higher desalination rate than N-0.25 in the second time stage. The homogeneous ionic crosslinking networks in M-0.5, which endow membranes lower area resistance and higher LCD, are responsible for it. Hence, it can be concluded that the solvent-free crosslinking strategy presents significantly advantage of maintaining high ionic content, whilst minimizing increase in membrane area resistance and decrease in LCD. Accordingly, the resultant crosslinked membrane, e.g. M-0.5, is expected to have a great potential in ED application while overall consideration of their all properties. 4. Conclusion In summary, we herein present our current efforts toward
Q. Pan et al. / Journal of Membrane Science 515 (2016) 115–124
developing upgradable solvent-free strategy for crosslinked AEMs. This approach, distinct from the classical post-quaternization processes, performs crosslinking and quaternization simultaneously. Particularly, imidazolium cations locating on cross-links endow the resultant membrane a high IEC (over 2.0 mmol g 1) for target low area resistance (lower than 1.4), while maintain a high ion selective transport property. The resulting membranes exhibit much higher limiting current densities (over 20 mA cm 2) than commercial AEM (Neosepta AMX, 14.2 mA cm 2, measured at the same condition). Accordingly, when being applied in ED application, the optimized M-0.5 membrane exhibits higher desalination efficiency than commercial Neosepta AMX, suggesting its potential application in ED.
Acknowledgements We thank the National Natural Science Foundation of China (Nos. 91534203, 21376232, 21490581), National High Technology Research and Development Program 863 (2015AA03A601) and One Hundred Person Project of the Chinese Academy of Sciences (2014, D type).
Nomenclature AEM BPPO
anion exchange membrane brominated poly (2, 6-dimethyl-1, 4-phenylene oxide) VBC 4-vinylbenzyl chloride IEM ion exchange membrane ED electrodialysis CEM cation exchange membrane TMA trimethylamine IEC ion exchange capacity N-VI N-vinylimidazole N-MI N-methylimidazole NMP N-Methyl pyrrolidone DMf N, N-dimethylformamide DMSO dimethyl sulfoxide TGA thermogravimetric analysis DTG differential thermal gravity mwet mass of hydrated membrane (g) mdry mass of dehydrated membrane (g) Rsol electrolyte resistance (Ω cm2) Rcell resistance of membrane equilibrated in electrolyte solution (Ω cm2) Rmem membrane resistance (Ω cm2) Em membrane potential A effective area of the membranes (cm2) R gas constant (J K 1 mol 1) F Faraday constant (C mol 1) C1 and C2 concentrations of electrolyte solutions t′i transport number of the counter-ion in the membrane ti transport number of the counter-ion in free solution DC direct current IPN interpenetrating polymer network τ transition time (s) i current density (mA cm 2) Zi valence of the ith ion π mathematical constant (3.141) D diffusion coefficient (cm2 s 1)
ΔE Δt i*
ε ilim
δ
Ror
123
plateau length (V) interval time (s) local current density in the conducting regions (mA cm 2) fraction of conducting region and LCD limiting current density (mA cm 2) diffusion boundary layer thickness (cm) resistance in ohmic region (Ω cm2)
References [1] W. Juda, W.A. McRae, Coherent ion-exchange gels and membranes, J. Am. Chem. Soc. 72 (1950) 1043–1044. [2] M.G. Buonomenna, Membrane processes for a sustainable industrial growth, RSC Adv. 3 (2013) 5694–5740. [3] L. Jingyi, W. Cuiming, X. Tongwen, W. Yonghui, Diffusion dialysis-concept, principle and applications, J. Membr. Sci. 366 (2011) 1–16. [4] D.H. Kim, A review of desalting process techniques and economic analysis of the recovery of salts from retentates, Desalination 270 (2011) 1–8. [5] H. Strathmann, A. Grabowski, G. Eigenberger, Ion-exchange membranes in the chemical process industry, Ind. Eng. Chem. Res. 52 (2013) 10364–10379. [6] L. Hong-Joo, H. Min-Kyoung, H. Sang-Don, M. Seung-Hyeon, Influence of the heterogeneous structure on the electrochemical properties of anion exchange membranes, J. Membr. Sci. 320 (2008) 549–555. [7] B.M. Asquith, J. Meier-Haack, C. Vogel, W. Butwilowski, B.P. Ladewig, Sidechain sulfonated copolymer cation exchange membranes for electro-driven desalination applications, Desalination 324 (2013) 93–98. [8] C.G. Arges, J. Parrondo, G. Johnson, A. Nadhan, V. Ramani, Assessing the influence of different cation chemistries on ionic conductivity and alkaline stability of anion exchange membranes, J. Mater. Chem. 22 (2012) 3733–3744. [9] K.K. Stokes, J.A. Orlicki, F.L. Beyer, RAFT polymerization and thermal behavior of trimethylphosphonium polystyrenes for anion exchange membranes, Polym. Chem. 2 (2011) 80–82. [10] S. Gu, R. Cai, Y.S. Yan, Self-crosslinking for dimensionally stable and solventresistant quaternary phosphonium based hydroxide exchange membranes, Chem. Commun. 47 (2011) 2856–2858. [11] S. Gu, R. Cai, T. Luo, Z.W. Chen, M.W. Sun, Y. Liu, G.H. He, Y.S. Yan, A soluble and highly conductive ionomer for high-performance hydroxide exchange membrane fuel cells, Angew. Chem. Int. Ed. 48 (2009) 6499–6502. [12] K.J.T. Noonan, K.M. Hugar, H.A. Kostalik, E.B. Lobkovsky, H.D. Abruna, G. W. Coates, Phosphonium-functionalized polyethylene: a new class of basestable alkaline anion exchange membranes, J. Am. Chem. Soc. 134 (2012) 18161–18164. [13] D.S. Kim, A. Labouriau, M.D. Guiver, Y.S. Kim, Guanidinium-functionalized anion exchange polymer electrolytes via activated fluorophenyl-amine reaction, Chem. Mater. 23 (2011) 3795–3797. [14] J.H. Wang, S.H. Li, S.B. Zhang, Novel hydroxide-conducting polyelectrolyte composed of an poly(arylene ether sulfone) containing pendant quaternary guanidinium groups for alkaline fuel cell applications, Macromolecules 43 (2010) 3890–3896. [15] X.C. Lin, L. Wu, Y.B. Liu, A.L. Ong, S.D. Poynton, J.R. Varcoe, T.W. Xu, Alkali resistant and conductive guanidinium-based anion-exchange membranes for alkaline polymer electrolyte fuel cells, J. Power Sources 217 (2012) 373–380. [16] G.G. Wang, Y.M. Weng, D. Chu, R.R. Chen, D. Xie, Developing a polysulfonebased alkaline anion exchange membrane for improved ionic conductivity, J. Membr. Sci. 332 (2009) 63–68. [17] J. Fang, P.K. Shen, Quaternized poly(phthalazinon ether sulfone ketone) membrane for anion exchange membrane fuel cells, J. Membr. Sci. 285 (2006) 317–322. [18] G.G. Wang, Y.M. Weng, D. Chu, D. Xie, R.R. Chen, Preparation of alkaline anion exchange membranes based on functional poly(ether-imide) polymers for potential fuel cell applications, J. Membr. Sci. 326 (2009) 4–8. [19] T.W. Xu, Ion exchange membranes: state of their development and perspective, J. Membr. Sci. 263 (2005) 1–29. [20] M.R. Hibbs, C.H. Fujimoto, C.J. Cornelius, Synthesis and characterization of poly (phenylene)-based anion exchange membranes for alkaline fuel cells, Macromolecules 42 (2009) 8316–8321. [21] T.J. Clark, N.J. Robertson, H.A. Kostalik, E.B. Lobkovsky, P.F. Mutolo, H. D. Abruna, G.W. Coates, A ring-opening metathesis polymerization route to alkaline anion exchange membranes: development of hydroxide-conducting thin films from an ammonium-functionalized monomer, J. Am. Chem. Soc. 131 (2009) 12888–12889. [22] N.J. Robertson, H.A. Kostalik, T.J. Clark, P.F. Mutolo, H.D. Abruna, G.W. Coates, Tunable high performance cross-linked alkaline anion exchange membranes for fuel cell applications, J. Am. Chem. Soc. 132 (2010) 3400–3404. [23] J.H. Wang, Z. Zhao, F.X. Gong, S.H. Li, S.B. Zhang, Synthesis of soluble poly (arylene ether sulfone) ionomers with pendant quaternary ammonium groups for anion exchange membranes, Macromolecules 42 (2009) 8711–8717.
124
Q. Pan et al. / Journal of Membrane Science 515 (2016) 115–124
[24] Z. Zhenghui, W. Liang, J. Varcoe, L. Chuanrun, O. Ai Lien, S. Poynton, X. Tongwen, Aromatic polyelectrolytes via polyacylation of pre-quaternized monomers for alkaline fuel cells, J. Mater. Chem. A 1 (2013) 2595–2601. [25] L.A. Wu, G.F. Zhou, X. Liu, Z.H. Zhang, C.R. Li, T.W. Xu, Environmentally friendly synthesis of alkaline anion exchange membrane for fuel cells via a solvent-free strategy, J. Membr. Sci. 371 (2011) 155–162. [26] T.W. Xu, W.H. Yang, Fundamental studies of a new series of anion exchange membranes: membrane preparation and characterization, J. Membr. Sci. 190 (2001) 159–166. [27] C. Larchet, S. Nouri, B. Auclair, L. Dammak, V. Nikonenko, Application of chronopotentiometry to determine the thickness of diffusion layer adjacent to an ion-exchange membrane under natural convection, Adv. Colloid Interface Sci. 139 (2008) 45–61. [28] L. Wu, Q. Pan, J.R. Varcoe, D. Zhou, J. Ran, Z.J. Yang, T.W. Xu, Thermal
[29]
[30]
[31]
[32]
crosslinking of an alkaline anion exchange membrane bearing unsaturated side chains, J. Membr. Sci. 490 (2015) 1–8. C. Klaysom, R. Marschall, S.-H. Moon, B.P. Ladewig, G.Q.M. Lu, L. Wang, Preparation of porous composite ion-exchange membranes for desalination application, J. Mater. Chem. 21 (2011) 7401. P.V. Vyas, P. Ray, S.K. Adhikary, B.G. Shah, R. Rangarajan, Studies of the effect of variation of blend ratio on permselectivity and heterogeneity of ion-exchange membranes, J. Colloid Interface Sci. 257 (2003) 127–134. J.H. Choi, S.H. Kim, S.H. Moon, Heterogeneity of ion-exchange membranes: the effects of membrane heterogeneity on transport properties, J. Colloid Interface Sci. 241 (2001) 120–126. J.H. Choi, H.J. Lee, S.H. Moon, Effects of electrolytes on the transport phenomena in a cation-exchange membrane, J. Colloid Interface Sci. 238 (2001) 188–195.