- Email: [email protected]
Electrochemical performance of a thermally rearranged polybenzoxazole nanocomposite membrane as a separator for lithium-ion batteries at elevated temperature
Journal of Power Sources 305 (2016) 259e266
Contents lists available at ScienceDirect
Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour
Electrochemical performance of a thermally rearranged polybenzoxazole nanocomposite membrane as a separator for lithium-ion batteries at elevated temperature Moon Joo Lee a, 1, Jun-Ki Hwang a, 1, Ji Hoon Kim b, Hyung-Seok Lim a, Yang-Kook Sun b, Kyung-Do Suh a, *, Young Moo Lee b, ** a b
Department of Chemical Engineering, College of Engineering, Hanyang University, Seoul 133-791, Republic of Korea WCU Department of Energy Engineering, Hanyang University, Seoul, Republic of Korea
h i g h l i g h t s TR-PBO membranes with high thermal and electrochemical stability were prepared. The structure of TR-PBO membranes has an effect on the electrochemical performance. TR-PBO membranes show excellent Li-ion cell performance at high operating temperature. Sea-squirt structured nanoparticles help the membrane wettability and ionic transport.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 3 April 2015 Received in revised form 28 October 2015 Accepted 18 November 2015 Available online 13 December 2015
Shape-tunable hydroxyl copolyimide (HPI) nanoparticles are fabricated by a re-precipitation method and are coated onto electrospun HPI membranes, followed by heat treatment to prepare thermally rearranged polybenzoxazole (TR-PBO) composite membranes. The morphology of HPI nanoparticles consisted of sphere and sea-squirt structures, which is controlled by changing the concentration of the stabilizer. The morphological characteristics of TR-PBO nanoparticles convert from HPI nanoparticles by heat treatment and their composite membranes is confirmed by scanning electron microscopy (SEM), transmission electron microscopy (TEM), infrared spectroscopy (ATR-IR), thermogravimetric analysis (TGA) analysis, and contact angle measurements. TGA and DSC measurements confirm the excellent thermal stability compared to Celgard, a commercial PP separator for lithium-ion batteries (LIBs). Further, TR-PBO nano-composite membranes used in coin-cell type LIBs as a separator show excellent high power density performance as compared to Celgard. This is due to the fact that sea-squirt structured nanoparticles have better electrochemical properties than sphere structured nanoparticles at high temperature. © 2015 Published by Elsevier B.V.
Keywords: Lithium-ion batteries Composite membrane Thermally rearranged polybenzoxazole Re-precipitation method
1. Introduction Lithium-ion batteries (LIBs) with high electrochemical performance are the most common choice of power source for portable
* Corresponding author. Department of Chemical Engineering, College of Engineering, Hanyang University, Seoul 133-791, Republic of Korea. ** Corresponding author. WCU Department of Energy Engineering, Hanyang University, Seoul 133-791, Republic of Korea. E-mail addresses: [email protected] (K.-D. Suh), [email protected] (Y.M. Lee). 1 Equal contributions. http://dx.doi.org/10.1016/j.jpowsour.2015.11.068 0378-7753/© 2015 Published by Elsevier B.V.
electronic devices such as cellular phones, notebooks, and cameras [1e3]. In the future, LIBs must be safe to use and have an increased cell capacity, as required for large, high current density batteries for rapidly charge/discharge of electric vehicles and energy storage systems. The safety of a battery is closely related to the thermal stability of the separator and the flammability of the liquid electrolyte; this is because an explosion can result from the thermal shrinkage of the separator and the subsequent ignition of the flammable liquid electrolyte under demanding conditions [4,5]. Therefore, the separator plays an important role in regulating cell kinetics, allowing ionic flow, blocking a short circuit between two electrodes, and acting as a safety device. The separators most
260
M.J. Lee et al. / Journal of Power Sources 305 (2016) 259e266
generally used are based on polyolefins such as polyethylene (PE) and polypropylene (PP), and these show good performance due to their mechanical strength and chemical stability. However, they are unfavorable for use in next generation batteries due to their relatively low thermal stability and poor electrolyte wettability [6e8]. Organic or inorganic particles coated with polyolefin membranes have been investigated by some researchers to improve the thermal stability of the separator. Choi et al. prepared an Al2O3 powder coated PE membrane with a hydrophilic poly(lithium 4styrenesulfonate) binder [9]. These separators exhibited enhanced thermal stability and a 3.8% shrinkage ratio at 105 C, thereby largely retaining their initial dimensions. Preparation of polyimidecoated PE membranes by a simple dipping method was reported by Song et al. [10]. They found that the polyimide coating was very effective in suppressing thermal shrinkage without sacrificing its inherent battery performance. The thermal shrinkage ratio for the bare PE membrane was 83.3% after storage at 140 C for 30 min, while that for polyimide-coated PE membranes decreased markedly to 10%. Another approach is using a nonwoven separator such as polyethylene terephthalate (PET) [11,12], poly(phthalazinone ether sulfone ketone) (PPESK) [13], polyimide (PI) [14,15], and thermally rearranged polybenzoxazole (TR-PBO) [16e26] membranes by an electrospinning method to develop thermally stable separators. Among these, thermally rearranged polybenzoxazoles (TR-PBOs) have proved to be promising materials for gas separation membranes due to their inherent ultra-pores, which induce both extraordinary permeability and selectivity [16e26]. Furthermore, in our previous study, we have developed the first TR-PBO LIB separators that show excellent thermal and dimensional stability and low cell resistance [27]. In this study, to prepare LIB separators with excellent electrochemical properties, we first made HPI nanoparticles with sphere and sea-squirt structures that were subsequently coated on an electrospun HPI nonwoven membrane. Then, the HPI composite membranes were converted into TR-PBO composite membranes. It ̊ is our aim to tune the shape of HPI and TR-PBO nanoparticles under controlled conditions and to investigate particle formation mechanisms. The ultimate goal of this study is to determine the relationship between the particle shape and cell performance in TRPBO composite membranes as a potential separator for LIBs. TRPBO membrane separators were compared with Celgard® 2400 in terms of electrochemical performance.
solution. Then, 60 ml of o-xylene was poured into the HPAA solution while the temperature was maintained between 160 and 180 C for 6 h to fully imidize the amic acid groups into imide groups. After removing the eliminated water, the solution became a viscous, brown HPI solution. The solution was precipitated in a 3:1 water:methanol solution by using a mechanical mixer to effectively remove solvents. After washing the precipitated HPI powder overnight, a 3:1 water:methanol solution was again added followed by stirring 4 h. Subsequently, the solution was washed with deionized several times. Finally, the HPI powder was obtained and was vacuum dried at 150 C for 12 h. Average molecular weights (Mw) of two batches of the synthesized polymer were measured by gel permeation chromatography (GPC) in tetrahydrofuran (THF), and were 250,000 and 200,000. 2.3. Polymer nanoparticle fabrication HPI nanoparticles were prepared by the re-precipitation method, as described in the literature [28]. HPI powders were dispersed into NMP at a concentration of 2 wt% with mechanical stirring. After dissolving in NMP, the HPI solution was filtered by using a 0.5-mm PTFE syringe filter to remove any impurities. To prepare nanoparticles using stabilizers, two selected stabilizers, PVP and PVA, were added into the prepared solution with the same weight content of polymers. Then, a PVA or PVP solution was inserted into the polymer solution at a rate of 0.05 ml/min using a syringe pump (LB-200, Longer pump, China) in a four neck round bottom flask with no purge gas, which was pre-heated to 70 C by an oil bath with mechanical stirring at 300 rpm. Depending on the total volume of the final solution, additional stirring was conducted. After the nanoparticles were fully prepared and stabilized, the flask was quenched with ice water to prevent any aggregation of nanoparticles. The particle solution was filtered with a pressurized paper filter and was centrifuged to wash off the residual using a 3:1 ratio of ethanol and DI. The particles were obtained by freeze drying at 0 C for 30 h or more. 2.4. Preparation of electrospun HPI membrane
To synthesize the hydroxyl copolyimide, three monomers were used: 4,40 -oxydiphthalic anhydride (ODPA, Shanghai Resin Factory Co. Ltd., China), 3,30 -dihydroxy-4,40 -diamino-biphenyl (HAB, Central Glass Co. Ltd., Japan), and 4,40 -oxydianiline (ODA, Central Glass Co. Ltd., Japan). N-Methyl-2-pyrrolidinone (NMP, Aldrich Chemical Co., Milwaukee, WI, USA) and o-xylene were used as solvents for polymer synthesis. For particle production, NMP was used as a solvent, while methanol, ethanol, and isopropanol (SigmaeAldrich Corp., St. Louis, MO, USA), and deionized water (DI water) were used as non-solvents. Poly(vinyl alcohol) (PVA, average molecular weight 85,000) and poly(vinyl pyrrolidone) (PVP, average molecular weight 40,000) were also purchased from SigmaeAldrich.
HPI powders were dissolved into dimethyl acetamide (DMAc, Aldrich Chemical Co. Milwaukee, WI, USA) in 8 wt% solution. A single nozzle system with an automated syringe pump (LB-200, Longer pump, China) was used to fabricate the electrospun HPI membrane. A rotating drum, with a diameter of 12 cm, was used as a collector, and was covered with polyethylene terephthalate (PET) as a backing material (BM), and was rotating at a speed of 10 rpm. The tip-to-collector distance was 15 cm. A syringe (Normject) with a nozzle (diameter of 6 mm) was attached. Furthermore, the syringe was attached to an ES-robot (Nano NC, Korea), which moves in the x-direction in a range of 45e230 mm at a speed of 100 mm/min. The HPI solution was poured into the syringe and was directly electrospun at a speed of 0.5 ml/min to the collector for a total volume of 8 ml. During the injection, 10 kV and 4 kV were applied to the needle and the collector, respectively. After the electrospinning process, the electrospun HPI membrane was dried for 2 h at room temperature, and then was detached from PET BM. The membrane was then was pressed at 130 C using a pressure of 150 kgf/cm2 to form the final electrospun membrane with a thickness of 18e20 mm.
2.2. Hydroxy copolyimide (HPI) synthesis
2.5. Fabrication of TR-PBO nano-composite membrane
Hydroxypolyamic acid (HPAA) was synthesized by an azeotropic imidization method as described in the literature [20]. Following the azeotropic imidization, the HPAA solution becomes an HPI
The preformed HPI nanoparticles and nano-composite membranes were thermally rearranged into polybenzoxazole as follows. First, the temperature was increased by ramping at 10 C/min to
2. Experimental 2.1. Materials
M.J. Lee et al. / Journal of Power Sources 305 (2016) 259e266
300 C followed by equilibration for 1 h in a furnace under a nitrogen atmosphere to evaporate residual solvents in the membranes. Membranes were then thermally rearranged into poly(benzoxazole) at 400 C for 2 h at a ramp rate of 10 C/min. Fabrication of a TR-PBO particle coated electrospun membrane was conducted in two steps. First, HPI nanoparticles, prepared by using PVP/PI wt% ratios of 2/2 and 5/5, designated as aP-PVP2/PI2-HPI and aP-PVP5/PVP5-HPI, were dispersed in ethanol as 1-wt% solutions and were sonicated 60 min. Then, the solution was sprayed on the front side of the azeotropic imidized electrosun HPI membrane using a spray gun. Spraying distance between the gun and the membrane was less than 30 cm. After the dispersion solution was sprayed onto the membrane, the particle-coated membrane was dried in an oven. Finally, the particle coated composite HPI membranes were treated at 400 C for 2 h to induce thermal rearrangement, as described in the previous section. 2.6. Characterization of HPI and TR-PBO nanoparticles and TR-PBO composite membranes The morphologies of sub-micrometer HPI nanoparticles were verified with scanning electron microscopy (SEM, JSM-6300, JEOL, Tokyo, Japan) and transmission electron microscopy (TEM, JEOL, JEM-2000EXII, Tokyo, Japan). Additionally, the average particle diameters were determined from SEM images. After thermal treatment of the nanoparticles and composite membranes, Fourier transform infrared spectroscopy (FT-IR, Nicolet 6700, Thermo, MA, USA), attenuated total reflectance infrared spectroscopy (ATR-IR, Nicolet 6700, Thermo, MA, USA), and thermogravimetric analysis (TGA, Q50, TA Instruments, DE, USA) with mass spectrometry (MS, GSD30173, Thermostar, Asslar, Germany) were conducted to determine whether they were fully converted or not. Further, to measure the hydrophobicity of the separators, contact angles were measured using a contact angle analyzer (Phoeaix 300, S.E.O, Ansung, Korea) by dropping a liquid electrolyte on the separators. To clarify the compatibility between particle coating layers and the non-woven supporting layer of the TR-PBO nano-composite membrane, the membranes were subjected to sonication (5510EDTH, Bransonic Ultrasonic Co. Ltd., USA) for 60 min and 100 min. Thickness changes were estimated from SEM images. Furthermore, the porosity of the separators was measured by using a capillary flow porometer (CFP-1500-AE, Porous Materials Inc.). 2.7. Electrochemical measurements Coin-type full-cells (CR 2032) were employed for the electrochemical measurements. Natural graphite (Wellcos, Japan) was used for the anode, LiCoO2 (LCO, Wellcos, Japan) for the cathode, and 1 M LiPF6 in ethylene carbonate (EC)/diethyl carbonate (DEC) (1/1, v/v, Techno Semichem Co. Ltd., Korea) was used as a liquid electrolyte. A PP separator (Celgard 2400, Charlotte, USA) was used as a reference separator. A non-coating TR-PBO non-woven membrane (pristine membrane), HPI1, and HPI4 particle-coated nanocomposite membranes (TR-PBO 1 membrane and TR-PBO 4 membrane, respectively) were also used as separators for
261
electrochemical testing (see Table 2). The electrodes and the separators were dried at 100 C for 20 min. Subsequently, assembly of cells was performed in a glove box under an argon atmosphere by stacking the anode, separator, and cathode and filling the liquid electrolyte to activate the cell. To test the cycle performance of the cells, the cells were cycled at a constant charge/discharge current density of 1 C/1 C in a voltage range from 3.0 to 4.2 V at 30 C. 3. Results and discussion 3.1. Structure control by stabilizer The HPI particle size and shape were controlled by changing the concentration of two kinds of stabilizers, NMP and PVA. Table 1 provides a summary of sample names along with the corresponding concentration of stabilizers and their characteristics. Fig. 1 shows SEM images of as-prepared HPI nanoparticles. HPI1 nanoparticles showed a smooth surface and spherical shape with an average diameter of 700 nm, as shown in Fig. 1(a); these nanoparticles were prepared with 2 wt% PVP stabilizer in water. However, nanoparticles produced in a 2 wt% PVA solution (Fig. 1(b)) exhibited interesting sea-squirt shapes. When the HPI concentration in water changed from 2 wt% to 5 wt%, as shown in Fig. 1(c) and (d), respectively, the shape of the sea-squirt nanoparticles changed dramatically. TEM images of spherical spheres and sea-squirt structures are shown in Fig. 1(e) and (f). TEM images of other seasquirt structures, i.e., HPI3 and HPI4, were similar to the TEM image of HPI2. As shown in Fig. 1(e) and (f), TEM images of HPI1 nanoparticles exhibited whole spherical particles, where HPI2 nanoparticles originate from agglomeration of primary particles to form sea-squirt shaped nanoparticles. The stabilizer was adsorbed on the surface of primary particles during particle formation and provided a barrier to further particle growth due to the steric hindrance of the long chains of the stabilizer molecules. A consequence of this is that a low concentration of the stabilizer or a low stabilizer molecular weight induces coalescence between primary particles. On the other hand, a high concentration or a high molecular weight leads to agglomeration between primary particles because of steric hindrance. This paper is focused on the particle structure at the membrane surface for high electrochemical performance in LIBs. So, we performed a comparative analysis of two samples having a sphere structure (HPI1) and a sea-squirt structure (HPI3). TR-PBO nanoparticles were successfully prepared by thermal treatment of HPI nanoparticles at 400 C for 2 h. As shown in Fig. S1, the morphologies of HPI and TR-PBO nanoparticles can be categorized into two groups by their shapes, i.e., (a) and (b) spheres, and (c) and (d) sea-squirts. Even after HPIs were fully converted into TRPBOs, their shape was maintained. Three TR-PBO membranes were prepared. First, an HPI membrane was prepared by azeotrope imidization, and a pristine electrospun membrane was made without particle coating and was thermally rearranged at 400 C for 2 h. Second, spherically shaped TR-PBO 1 nanoparticles were coated onto a pristine membrane with thermal treatment at 400 C for 2 h. Third, sea-squirt shaped
Table 1 Characteristics of particle size and shape prepared with different concentrations of stabilizers. Sample name
Stabilizers
HPI (wt%)
Size (nm)
Shape
TR-PBO particles
Membrane
HPI1 HPI2 HPI3 HPI4
PVP 2% PVA 2% PVP 5% PVP 5%
2 2 2 5
550 900 900 700
Sphere Sea-squirt Sea-squirt Sea-squirt
TR-PBO TR-PBO TR-PBO TR-PBO
TR-PBO TR-PBO TR-PBO TR-PBO
Adding speed ¼ 0.05 ml/min, temperature was 70 C, stirring speed of 300 rpm.
1 2 3 4
1 2 3 4
membrane membrane membrane membrane
262
M.J. Lee et al. / Journal of Power Sources 305 (2016) 259e266
Table 2 Summary of basic properties of all separators.
Composition Thickness (mm) Pore size (mm) Porosity (%) Contact angle ( ) (Avg/STD)
Celgard® 2400
Pristine membrane
TR-PBO 1 membrane
TR-PBO 4 membrane
PP 25 0.043 41 ± 1.2 42.83/1.83
PBO 18 0.7563 63 ± 0.7 23.75/2.75
PBO e PBO 33 0.643 62 ± 1.5 24/3.02
PBO e PBO 40 0.5579 62 ± 0.3 N/A
Fig. 1. SEM and TEM images of spherical and sea-squirt shaped nanoparticles for (a) HPI1, (b) HPI2, (c) HPI3, and (d) HPI4. TEM images are provided for (e) HPI1 and (f) HPI2. Spherical HPI nanoparticles with average size of 900 nm (a) and sea-squirt shaped nanoparticles with average size of 700 nm in diameter can be seen in (b).
Fig. 2. SEM images of HPI and TR-PBO nanocomposite membranes: (a) front and (b) cross-section of pristine membrane, (c) front and (d) cross-section of the TR-PBO 1 membrane, and (e) front and (f) cross-section of the TR-PBO 4 membrane.
M.J. Lee et al. / Journal of Power Sources 305 (2016) 259e266
TR-PBO 4 nanoparticles were coated onto a pristine membrane with thermal treatment at 400 C for 2 h. The morphologies of the membranes were investigated by SEM, as shown in Fig. 2. The pristine membrane displayed a macro-porous nonwoven surface and interconnected nanofibers merged together, as seen in the front view (Fig. 2(a)), while no particles were observed in a crosssectional view (Fig. 2(b)). In contrast, the TR-PBO 1 membrane had a closely arranged surface (Fig. 2(c)), and the particle layer had a thickness of approximately 5e6 mm (Fig. 2(d)). Finally, the TR-PBO 4 membrane exhibited (Fig. 2(e)) a relatively more porous surface than the TR-PBO 1 (Fig. 2(c)). This is due to the polydisperse size distribution of TR-PBO 4 nanoparticles, and the TR-PBO 4 sample showed that the particle coating layer had a thickness of approximately 8e9 mm (Fig. 2(f)). As shown in Fig. S2, the structural differences between HPI and TR-PBO were clarified using ATR-IR. First, there was no broad and strong band observed between 3200 cm1 and 3600 cm1 because the hydroxyl groups in HPI copolymers were previously converted into polybenzoxazole. Second, there were two imide bands (eCeO) at 1777 cm1 and 1718 cm1. ODPA-HAB5-ODA5 is a co-polymer of polybenzoxazole (ODPA-HAB) and polyimide (ODPA-ODA). In addition, because the TR-PBO copolymer membranes were not derived from hexafluorinated monomers, the benzoxazole bands did not appear at 1059 cm1, but shifted to 1052 cm1. However, a bezoxazole band at 1475 cm1 appeared at the same position in both HPI and TR-PBO samples. Above all, though the pristine membrane indicated an explicit benzoxazole band at 1052 cm1, the TR-PBO 1 and TR-PBO 4 membranes displayed ambiguous bands in this region, because these were in the form of particles [28] and contained the TR-co-polymer, as shown in Fig. S2. The wettability of the separators is critical to the electrochemical performance. Wettability was evaluated using a contact angle analyzer by dropping non-aqueous liquid electrolyte used in all electrochemical tests (ED:DEC ¼ 1:1, w:w, 1 M LiPF6) onto the membrane surface. The wettability is influenced by the surface structure and chemical properties. As already discussed, these TRPBO series membranes have similar functional groups, as shown in Fig. S2. Fig. 3 shows that the contact angle was low for the TRPBO 4 membrane (d) as compared to pristine membrane (b). By contrast, even though the TR-PBO 1 membrane (Fig. 3(c)) was particle coated, the contact angle was similar to that of the non-
263
coated pristine membrane. Because the surface area of the TRPBO 4 membrane increased greatly and the surface pore size decreased substantially, the electrolyte was absorbed instantly due to an induced capillary effect. However, while the TR-PBO 1 membrane had increased surface area, the particle layer was more closely packed, and the capillary effect was not observed. In any case, all the TR-PBO membranes show higher wettability than Celgard because of the polarity between the microporous TR-PBO copolymer separators and the highly polar liquid electrolyte [29]. Therefore, the electrolyte uptake of the TR-PBO copolymer separators was measured to be greater than that of the Celgard. Table 2 provides a summary of the basic properties of all separators, including porosity. Thermal stability of the separators evaluated by TGA and DSC showed similar results in comparison with our previous report. As can be seen in Figure S3(a), the TR-PBO 1 and the TR-PBO 4 membranes showed no weight loss until degradation occurred at 600 C, while the Celgard was decomposed below 400 C. TGA analysis showed that not only were the TR-PBO membranes fully converted, but the TR-PBO membranes also showed superior thermal properties compared to the Celgard. Additional thermal analysis was conducted using DSC, and these data are provided in Figure S3(b). The TR-PBO membranes showed a straight line in the DSC data until 250 C, while the Celgard membrane displayed an endothermic peak at 160 C, characteristic of the melting point of polypropylene. Consequently, the TR-PBO membranes provide superior thermal stability in comparison with the Celgard based on DSC and TGA analyses. Fig. S4 shows digital pictures of (a and d) Celgard 2400, (b and e) pristine, (c and f) TR-PBO 1, and (d and g) TR-PBO 4 membranes before and after a hot oven test at 220 C for 30 min. TR-PBO series membrane maintained its integrity even after a hot oven test at high temperature, while the Celgard 2400 membrane made of polypropylene was completely contracted, as shown in Figure S4(e). 3.2. Electrochemical properties The battery performances of the two different TR-PBO based cells were evaluated with a LiCoO2/graphite electrochemical couple. Fig. S5 show the chargeedischarge curves of the TR-PBO 4, TR-PBO 1, and the Celgard 2400 based cells, where the voltage
Fig. 3. Contact angle analysis of the separator membranes. Drops were stabilized in 0.42 s and images were taken. The membranes were (a) Celgard 2400, (b) pristine membrane, (c) TR-PBO 1 membrane, and (d) TR-PBO 4 membrane. The hydrophobicity trend is (d) > (c) > (b) > (a).
264
M.J. Lee et al. / Journal of Power Sources 305 (2016) 259e266
ranged between 3.0 and 4.2 V with charge/discharge current rates at two initial cycles (0.05 C/0.05 C) up to the 3rd cycle (1.0 C/1.0 C). All samples show typical charge/discharge curves of the coin-cell assembled with LiCoO2 and graphite electrodes as cathode and anode. Fig. 4 shows the cycling performances of the TR-PBO 4, TRPBO 1, and the Celgard 2400 based cells, where the voltage ranged between 3.0 and 4.2 V with charge/discharge current rates at two initial cycles (0.05 C/0.05 C) up to the 100th cycle (1.0 C/1.0 C). Initial discharge capacities of the three cells are not significantly different at a low current rate (0.05 C). At a higher C-rate (1.0 C), however, the discharge capacities of TR-PBO 4, TR-PBO 1, and Celgard 2400 based cells were 103 mAh g1, 78 mAh g1, and 56 mAh g1, respectively. Two different TR-PBO based cells exhibited higher capacities than the Celgard 2400 based cell due to their excellent wettability, which can improve the Li-ion mobility [19]. As shown in Fig. 3, the wettability of the TR-PBO series membranes is greater than that of the Celgard 2400 membrane. Further, the discharge capacity of TR-PBO 4 based cell is 25% higher than that of the TR-PBO 1 based cell. This result is very closely related to the porosity of the membranes. In this experiment, two types of TR-PBO particles having different morphologies were utilized to control the porosity of electrospun TR-PBO membranes having macropores caused by deposition of one-dimensional TRPBO strings. As shown in Table 1, the TR-PBO 4 membrane coated with sea-squirt TR-PBO nanoparticles has a smaller average pore size than that of the TR-PBO 1 membrane. In addition, the narrow pore size distribution (at a maximum diameter, TR-PBO 4: 94.2% and TR-PBO 1: 93.7%) of the TR-PBO 4 membrane has an influence on the higher capacity. These characteristics of the TR-PBO 4 membrane result in better lithium-ion transport during chargeedischarge processes because the sea-squirt structures of the TRPBO 4 nanoparticles can make small porous channels that accelerate ion transport. In order to understand the effect of membrane pore size on the lithium-ion transport, electrochemical impedance spectroscopy (EIS) was employed. Fig. 5 shows Nyquist plots of the electrochemical impedance spectra obtained from the fresh TR-PBO 4, TRPBO 1, and Celgard 2400 based cells (before cycling) and those of the cells after cycling at a 1 C rate. The coin cells used for this measurement were made by the same fabrication process. EIS analysis can provide electrochemical information such as charge and mass transfer resistances, and correlation between the electrochemical performance and electrode kinetics. Fig. 5(a) shows the
Fig. 4. Cycle performance of cells assembled with TR-PBO 4, TR-PBO 1 and Celgard 2400 at a 1 C rate. The cells were operated in the voltage range of 3.0e4.2 V at 30 C.
Nyquist plots of the fresh cells containing TR-PBO 4, TR-PBO 1, and Celgard 2400 separators. The Nyquist plots are composed of one semicircle in the high frequency range and a sloping straight line at low frequency, which can be assigned to the impedance of the charge transfer resistance (Rct) coming from the interface between the electrode and electrolyte, and the Warburg impedance (W), related to the diffusion of lithium ions into the graphite anode phase. Charge-transfer resistances of TR-PBO 4, TR-PBO 1, and Celgard based cells were 48.2, 53.1, and 57.5 U, respectively. The TRPBO 4 based cell exhibited the lowest charge transfer and mass transfer resistances because of the excellent wettability and high porosity. After 10 cycles, all impedance spectra were composed of two partially overlapping semicircles at high and middle frequency ranges, which can be assigned to the impedance of the solid electrolyte interface (RSEI) and the charge transfer resistance, respectively, with a sloping straight Warburg line at low frequency. After the first cycle, byproducts are generated by decomposition of the electrolyte, which can act as resistive layers for delaying ionic transport. Thus, the impedance of the solid electrolyte interface was obtained at low frequency, as shown in Fig. 5(b). However, the charge transfer resistances of all cells decreased after 10 cycles. The charge transfer resistances of TR-PBO 4, TR-PBO 1, and Celgard 2400 based cells were 5, 6 and 7 U, respectively, which indicates faster charge transfer kinetics in the TR-PBO-based cells. Such a phenomenon is because the irreversible SEI layer formed on the electrode slows the increase in cell impedance by protecting the electrode from chemically reacting with the electrolyte. After cycling, the charge transfer resistances of TR-PBO based cells were still smaller than that of the Celgard 2400 based cell. In particular, the TR-PBO 4 based cell exhibited lower charge and mass transfer resistances than those of the TR-PBO 1 based cell after 10 cycles because the TR-PBO 4 based cell has a high porosity caused by seasquirt like nanoparticles as well as the excellent affinity of the TRPBO with the organic electrolyte. These results suggest that the porosity and wettability of the separator have an effect on the kinetics of the cell electrochemical reaction and lithium-ion transport during cycling. The TR-PBO polymer has good thermal stability, as shown in Fig. S3. These thermal properties can significantly affect the electrochemical properties such as cycle and capability at high temperature. Fig. 6 shows the cyclability of three different cells containing TR-PBO 4, TR-PBO 1, and Celgard 2400 separators at same current rate of 1 C at a constant temperature of 50 C. At high temperatures, the cyclability of LIB cells is generally poor because the reactivity between the electrodes and electrolyte rapidly increases. In this test, interesting results were obtained. Initial discharge capacities of TR-PBO 4, TR-PBO 1, and Celgard 2400 based cells operated at 50 C were 291, 283, and 272 mAh g1, respectively. These values were similar to the initial discharge capacities of cells operated at room temperature. However, after 100 cycles at 50 C, the Celgard based cell showed lower capacities during continuous cycling up to the 100th cycle because of the poor thermal stability of the monolayer polypropylene. On the other hand, the TR-PBO 4 based cell exhibited much higher capacities from the 4th cycle up to the 100th cycle than even that observed for room temperature testing. These results are attributed to the excellent thermal stability of the TR-PBO polymer and the high porosity of the membrane. Since testing at high temperature leads to better Li-ion diffusivity due to the low electrolyte viscosity at high temperature, TR-PBO 4 based cells showed higher capacities even at high current rate. To compare the thermal stability of the membranes, we carried out a hot oven test at 220 C for 30 min Fig. S4 shows digital pictures of Celgard 2400 and TR-PBO series membranes having a uniform size (2 cm2) before and after hot oven testing. TR-PBO series membranes maintained their original
M.J. Lee et al. / Journal of Power Sources 305 (2016) 259e266
265
Fig. 5. Electrochemical impedance (EIS) Nyquist plots of the lithium cells of the TR-PBO separators compared with Celgard 2400 as a reference (frequency range 5e1000 Hz at amplitude of 5 mV) at (a) first cycle and (b) after 10 cycles with 1 C current charge/discharge.
Fig. 6. Cycle performance of cells assembled with selected separators in 1 C current density and compared with Celgard 2400 as a reference. The cells were operated in the voltage range between 3.0 and 4.2 V at 50 C.
membrane size even above 200 C, in contrast with a contracted Celgard 2400 membrane. Furthermore, the porous structure of the Celgard 2400 membrane completely disappeared after hot oven testing, as shown in Figure S6(d). On the other hand, the intrinsic surface structure of TR-PBO series membranes was not changed, as shown in Figures S6(e) and (f). Therefore, TR-PBO separators coated with TR-PBO particles significantly improve battery safety and exhibited high capacity retention during high temperature testing.
of the electrolyte. Thermal stability of these membranes was excellent when compared to Celgard, as confirmed by TGA and DSC analyses. The particle coating on TR-PBO membranes decreased the overall pore size, and the sea-squirt shaped particle coated membrane (TR-PBO 4 membrane) displayed smaller pore size and broader pore size distribution than the spherically shaped particle coated membrane (TR-PBO 1 membrane). The wettability of the TRPBO membranes was superior to the Celgard, while the TR-PBO 4 membrane exhibited instant absorption of electrolytes due to its large surface area. Moreover, the thermal stability of the TR-PBO membranes surpassed that of the Celgard. TR-PBO membranes were stable up to 490 C, while the Celgard decomposed at 340 C, as evidenced from TGA analysis. TR-PBO membranes did not show any melting up to 250 C, whereas the Celgard showed melting at 160 C based on DSC analysis. Finally, the cyclability test validated the superiority of the TR-PBO 4 membrane among the TR-PBO membranes due to the fact that a sea-squirt shaped particle coating layer facilitated ion transport and improved wettability of the electrolyte. Meanwhile, all the TR-PBO membranes displayed better cycle performance compared to the Celgard membrane, which demonstrates that TR-PBO membranes are excellent candidates for high density LIB separators. Acknowledgments This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (2012M3A7B4049745). This work was supported by the Global Frontier R&D Program (2013M3A6B1078875) on Center for Hybrid Interface Materials (HIM) funded by the Ministry of Science, ICT & Future Planning.
4. Conclusions
Appendix A. Supplementary data
Shape tunable TR-PBO nanoparticles were successfully fabricated by a re-precipitation method and subsequent thermal treatment. The parameters influencing the particle shape were investigated, and a possible particle formation mechanism was suggested. Meanwhile, spherically shaped HPI1 and sea-squirt shaped HPI4 nanoparticles were coated onto the pristine membrane and were successfully converted to TR-PBO 1 and TR-PBO 4 membrane nano-composite membranes, respectively. Both HPI and TR-PBO nanoparticles and TR-PBO membranes were investigated by ATR-IR and TGA analysis to confirm the structural evolution. Moreover, the TR-PBO membranes, including the pristine membrane, showed good pore size distribution and excellent wettability
Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.jpowsour.2015.11.068. References [1] [2] [3] [4] [5] [6] [7] [8]
M.H. Ryou, Y.M. Lee, J.K. Park, J.W. Choi, Adv. Mater. 23 (2011) 3066e3070. K. Hwang, B. Kwon, H. Byun, J. Membr. Sci. 378 (2011) 111e116. J. Hassoun, B. Scrosati, Adv. Mater. 22 (2010) 5198e5201. C.J. Orendorff, T.N. Lambert, C.A. Chavez, M. Bencomo, K.R. Fenton, Adv. Energy Mater. 3 (2013) 314e320. K. Xu, Chem. Rev. 104 (2004) 4303e4418. H. Maleki, J.N. Howard, J. Power Sources 191 (2009) 568e574. J.Y. Kim, Y.B. Lee, D.Y. Lim, Electrochim. Acta 54 (2009) 3714e3719. J.Y. Lee, Y.M. Lee, B. Bhattacharya, Y.C. Nho, J.K. Park, Electrochim. Acta 54
266
M.J. Lee et al. / Journal of Power Sources 305 (2016) 259e266
(2009) 4312e4315. [9] J.-A. Choi, S.H. Kim, D.-W. Kim, J. Power Sources 195 (2010) 6192e6196. [10] J. Song, M.H. Ryou, B. Son, J.N. Lee, D.J. Lee, Y.M. Lee, J.W. Choi, J.-K. Park, Electrochim. Acta 85 (2012) 524e530. [11] J.L. Hao, G.T. Lei, Z.H. Li, L.J. Wu, Q.Z. Xiao, L. Wang, J. Memb. Sci. 428 (2013) 11e16. [12] J. Ding, Y. Kong, P. Li, J. Yang, J. Electrochem. Soc. 159 (2012) A1474eA1480. [13] W. Qi, C. Lu, P. Chen, L. Han, Q. Yu, R. Xu, Mater. Lett. 66 (2012) 239e241. [14] W. Jiang, Z. Liu, Q. Kong, J. Yao, C. Zhang, P. Han, G. Cui, Solid State Ion. 232 (2013) 44e48. [15] Y.-E. Miao, G.-N. Zhu, H. Hou, Y.-Y. Xia, T. Liu, J. Power Sources 226 (2013) 82e86. [16] H.B. Park, C.H. Jung, Y.M. Lee, A.J. Hill, S.J. Pas, S.T. Mudie, E.V. Wagner, B.D. Freeman, D.J. Cookson, Science 318 (2007) 254e258. [17] J.I. Choi, C.H. Jung, S.H. Han, H.B. Park, Y.M. Lee, J. Memb. Sci. 349 (2010) 358e368. [18] S.H. Han, N. Misdan, S. Kim, C.M. Doherty, A.J. Hill, Y.M. Lee, Macromolecules 43 (2010) 7657e7667.
[19] S.H. Han, H.J. Kwon, K.Y. Kim, J.G. Seong, C.H. Park, S. Kim, C.M. Doherty, A.W. Thornton, A.J. Hill, A.E. Lozano, K.A. Berchtold, Y.M. Lee, Phys. Chem. Chem. Phys. 14 (2012) 4365e4373. [20] H.B. Park, S.H. Han, C.H. Jung, Y.M. Lee, A.J. Hill, J. Memb. Sci. 359 (2010) 11e24. [21] S. Kim, S.H. Han, Y.M. Lee, J. Memb. Sci. 403e404 (2012) 169e178. [22] M. Calle, Y.M. Lee, Macromolecules 44 (2011) 1156e1165. [23] M. Calle, Y. Chan, H.J. Jo, Y.M. Lee, Polymer 53 (2012) 2783e2791. [24] C.H. Jung, J.E. Lee, S.H. Han, H.B. Park, Y.M. Lee, J. Memb. Sci. 350 (2010) 301e309. [25] M. Calle, A.E. Lozano, Y.M. Lee, Eur. Polym. J. 48 (2012) 1313e1322. [26] H. Wang, T.-S. Chung, J. Memb. Sci. 385e386 (2011) 86e95. [27] M.J. Lee, J.H. Kim, H.-S. Lim, S.Y. Lee, H.K. Yu, J.H. Kim, J.S. Lee, Y.-K. Sun, M.D. Guiver, K.D. Suh, Y.M. Lee, Chem. Commun. 51 (2015) 2068e2071. [28] J.-Y. Xiong, X.-Y. Liu, S.B. Chen, T.-S. Chung, J. Phys. Chem. B 109 (2005) 13877e13882. [29] Y.-E. Miao, G.-N. Zhu, H. Hou, Y.-Y. Xia, T. Liu, J. Power Sources 226 (2013) 82e86.
Contents lists available at ScienceDirect
Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour
Electrochemical performance of a thermally rearranged polybenzoxazole nanocomposite membrane as a separator for lithium-ion batteries at elevated temperature Moon Joo Lee a, 1, Jun-Ki Hwang a, 1, Ji Hoon Kim b, Hyung-Seok Lim a, Yang-Kook Sun b, Kyung-Do Suh a, *, Young Moo Lee b, ** a b
Department of Chemical Engineering, College of Engineering, Hanyang University, Seoul 133-791, Republic of Korea WCU Department of Energy Engineering, Hanyang University, Seoul, Republic of Korea
h i g h l i g h t s TR-PBO membranes with high thermal and electrochemical stability were prepared. The structure of TR-PBO membranes has an effect on the electrochemical performance. TR-PBO membranes show excellent Li-ion cell performance at high operating temperature. Sea-squirt structured nanoparticles help the membrane wettability and ionic transport.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 3 April 2015 Received in revised form 28 October 2015 Accepted 18 November 2015 Available online 13 December 2015
Shape-tunable hydroxyl copolyimide (HPI) nanoparticles are fabricated by a re-precipitation method and are coated onto electrospun HPI membranes, followed by heat treatment to prepare thermally rearranged polybenzoxazole (TR-PBO) composite membranes. The morphology of HPI nanoparticles consisted of sphere and sea-squirt structures, which is controlled by changing the concentration of the stabilizer. The morphological characteristics of TR-PBO nanoparticles convert from HPI nanoparticles by heat treatment and their composite membranes is confirmed by scanning electron microscopy (SEM), transmission electron microscopy (TEM), infrared spectroscopy (ATR-IR), thermogravimetric analysis (TGA) analysis, and contact angle measurements. TGA and DSC measurements confirm the excellent thermal stability compared to Celgard, a commercial PP separator for lithium-ion batteries (LIBs). Further, TR-PBO nano-composite membranes used in coin-cell type LIBs as a separator show excellent high power density performance as compared to Celgard. This is due to the fact that sea-squirt structured nanoparticles have better electrochemical properties than sphere structured nanoparticles at high temperature. © 2015 Published by Elsevier B.V.
Keywords: Lithium-ion batteries Composite membrane Thermally rearranged polybenzoxazole Re-precipitation method
1. Introduction Lithium-ion batteries (LIBs) with high electrochemical performance are the most common choice of power source for portable
* Corresponding author. Department of Chemical Engineering, College of Engineering, Hanyang University, Seoul 133-791, Republic of Korea. ** Corresponding author. WCU Department of Energy Engineering, Hanyang University, Seoul 133-791, Republic of Korea. E-mail addresses: [email protected] (K.-D. Suh), [email protected] (Y.M. Lee). 1 Equal contributions. http://dx.doi.org/10.1016/j.jpowsour.2015.11.068 0378-7753/© 2015 Published by Elsevier B.V.
electronic devices such as cellular phones, notebooks, and cameras [1e3]. In the future, LIBs must be safe to use and have an increased cell capacity, as required for large, high current density batteries for rapidly charge/discharge of electric vehicles and energy storage systems. The safety of a battery is closely related to the thermal stability of the separator and the flammability of the liquid electrolyte; this is because an explosion can result from the thermal shrinkage of the separator and the subsequent ignition of the flammable liquid electrolyte under demanding conditions [4,5]. Therefore, the separator plays an important role in regulating cell kinetics, allowing ionic flow, blocking a short circuit between two electrodes, and acting as a safety device. The separators most
260
M.J. Lee et al. / Journal of Power Sources 305 (2016) 259e266
generally used are based on polyolefins such as polyethylene (PE) and polypropylene (PP), and these show good performance due to their mechanical strength and chemical stability. However, they are unfavorable for use in next generation batteries due to their relatively low thermal stability and poor electrolyte wettability [6e8]. Organic or inorganic particles coated with polyolefin membranes have been investigated by some researchers to improve the thermal stability of the separator. Choi et al. prepared an Al2O3 powder coated PE membrane with a hydrophilic poly(lithium 4styrenesulfonate) binder [9]. These separators exhibited enhanced thermal stability and a 3.8% shrinkage ratio at 105 C, thereby largely retaining their initial dimensions. Preparation of polyimidecoated PE membranes by a simple dipping method was reported by Song et al. [10]. They found that the polyimide coating was very effective in suppressing thermal shrinkage without sacrificing its inherent battery performance. The thermal shrinkage ratio for the bare PE membrane was 83.3% after storage at 140 C for 30 min, while that for polyimide-coated PE membranes decreased markedly to 10%. Another approach is using a nonwoven separator such as polyethylene terephthalate (PET) [11,12], poly(phthalazinone ether sulfone ketone) (PPESK) [13], polyimide (PI) [14,15], and thermally rearranged polybenzoxazole (TR-PBO) [16e26] membranes by an electrospinning method to develop thermally stable separators. Among these, thermally rearranged polybenzoxazoles (TR-PBOs) have proved to be promising materials for gas separation membranes due to their inherent ultra-pores, which induce both extraordinary permeability and selectivity [16e26]. Furthermore, in our previous study, we have developed the first TR-PBO LIB separators that show excellent thermal and dimensional stability and low cell resistance [27]. In this study, to prepare LIB separators with excellent electrochemical properties, we first made HPI nanoparticles with sphere and sea-squirt structures that were subsequently coated on an electrospun HPI nonwoven membrane. Then, the HPI composite membranes were converted into TR-PBO composite membranes. It ̊ is our aim to tune the shape of HPI and TR-PBO nanoparticles under controlled conditions and to investigate particle formation mechanisms. The ultimate goal of this study is to determine the relationship between the particle shape and cell performance in TRPBO composite membranes as a potential separator for LIBs. TRPBO membrane separators were compared with Celgard® 2400 in terms of electrochemical performance.
solution. Then, 60 ml of o-xylene was poured into the HPAA solution while the temperature was maintained between 160 and 180 C for 6 h to fully imidize the amic acid groups into imide groups. After removing the eliminated water, the solution became a viscous, brown HPI solution. The solution was precipitated in a 3:1 water:methanol solution by using a mechanical mixer to effectively remove solvents. After washing the precipitated HPI powder overnight, a 3:1 water:methanol solution was again added followed by stirring 4 h. Subsequently, the solution was washed with deionized several times. Finally, the HPI powder was obtained and was vacuum dried at 150 C for 12 h. Average molecular weights (Mw) of two batches of the synthesized polymer were measured by gel permeation chromatography (GPC) in tetrahydrofuran (THF), and were 250,000 and 200,000. 2.3. Polymer nanoparticle fabrication HPI nanoparticles were prepared by the re-precipitation method, as described in the literature [28]. HPI powders were dispersed into NMP at a concentration of 2 wt% with mechanical stirring. After dissolving in NMP, the HPI solution was filtered by using a 0.5-mm PTFE syringe filter to remove any impurities. To prepare nanoparticles using stabilizers, two selected stabilizers, PVP and PVA, were added into the prepared solution with the same weight content of polymers. Then, a PVA or PVP solution was inserted into the polymer solution at a rate of 0.05 ml/min using a syringe pump (LB-200, Longer pump, China) in a four neck round bottom flask with no purge gas, which was pre-heated to 70 C by an oil bath with mechanical stirring at 300 rpm. Depending on the total volume of the final solution, additional stirring was conducted. After the nanoparticles were fully prepared and stabilized, the flask was quenched with ice water to prevent any aggregation of nanoparticles. The particle solution was filtered with a pressurized paper filter and was centrifuged to wash off the residual using a 3:1 ratio of ethanol and DI. The particles were obtained by freeze drying at 0 C for 30 h or more. 2.4. Preparation of electrospun HPI membrane
To synthesize the hydroxyl copolyimide, three monomers were used: 4,40 -oxydiphthalic anhydride (ODPA, Shanghai Resin Factory Co. Ltd., China), 3,30 -dihydroxy-4,40 -diamino-biphenyl (HAB, Central Glass Co. Ltd., Japan), and 4,40 -oxydianiline (ODA, Central Glass Co. Ltd., Japan). N-Methyl-2-pyrrolidinone (NMP, Aldrich Chemical Co., Milwaukee, WI, USA) and o-xylene were used as solvents for polymer synthesis. For particle production, NMP was used as a solvent, while methanol, ethanol, and isopropanol (SigmaeAldrich Corp., St. Louis, MO, USA), and deionized water (DI water) were used as non-solvents. Poly(vinyl alcohol) (PVA, average molecular weight 85,000) and poly(vinyl pyrrolidone) (PVP, average molecular weight 40,000) were also purchased from SigmaeAldrich.
HPI powders were dissolved into dimethyl acetamide (DMAc, Aldrich Chemical Co. Milwaukee, WI, USA) in 8 wt% solution. A single nozzle system with an automated syringe pump (LB-200, Longer pump, China) was used to fabricate the electrospun HPI membrane. A rotating drum, with a diameter of 12 cm, was used as a collector, and was covered with polyethylene terephthalate (PET) as a backing material (BM), and was rotating at a speed of 10 rpm. The tip-to-collector distance was 15 cm. A syringe (Normject) with a nozzle (diameter of 6 mm) was attached. Furthermore, the syringe was attached to an ES-robot (Nano NC, Korea), which moves in the x-direction in a range of 45e230 mm at a speed of 100 mm/min. The HPI solution was poured into the syringe and was directly electrospun at a speed of 0.5 ml/min to the collector for a total volume of 8 ml. During the injection, 10 kV and 4 kV were applied to the needle and the collector, respectively. After the electrospinning process, the electrospun HPI membrane was dried for 2 h at room temperature, and then was detached from PET BM. The membrane was then was pressed at 130 C using a pressure of 150 kgf/cm2 to form the final electrospun membrane with a thickness of 18e20 mm.
2.2. Hydroxy copolyimide (HPI) synthesis
2.5. Fabrication of TR-PBO nano-composite membrane
Hydroxypolyamic acid (HPAA) was synthesized by an azeotropic imidization method as described in the literature [20]. Following the azeotropic imidization, the HPAA solution becomes an HPI
The preformed HPI nanoparticles and nano-composite membranes were thermally rearranged into polybenzoxazole as follows. First, the temperature was increased by ramping at 10 C/min to
2. Experimental 2.1. Materials
M.J. Lee et al. / Journal of Power Sources 305 (2016) 259e266
300 C followed by equilibration for 1 h in a furnace under a nitrogen atmosphere to evaporate residual solvents in the membranes. Membranes were then thermally rearranged into poly(benzoxazole) at 400 C for 2 h at a ramp rate of 10 C/min. Fabrication of a TR-PBO particle coated electrospun membrane was conducted in two steps. First, HPI nanoparticles, prepared by using PVP/PI wt% ratios of 2/2 and 5/5, designated as aP-PVP2/PI2-HPI and aP-PVP5/PVP5-HPI, were dispersed in ethanol as 1-wt% solutions and were sonicated 60 min. Then, the solution was sprayed on the front side of the azeotropic imidized electrosun HPI membrane using a spray gun. Spraying distance between the gun and the membrane was less than 30 cm. After the dispersion solution was sprayed onto the membrane, the particle-coated membrane was dried in an oven. Finally, the particle coated composite HPI membranes were treated at 400 C for 2 h to induce thermal rearrangement, as described in the previous section. 2.6. Characterization of HPI and TR-PBO nanoparticles and TR-PBO composite membranes The morphologies of sub-micrometer HPI nanoparticles were verified with scanning electron microscopy (SEM, JSM-6300, JEOL, Tokyo, Japan) and transmission electron microscopy (TEM, JEOL, JEM-2000EXII, Tokyo, Japan). Additionally, the average particle diameters were determined from SEM images. After thermal treatment of the nanoparticles and composite membranes, Fourier transform infrared spectroscopy (FT-IR, Nicolet 6700, Thermo, MA, USA), attenuated total reflectance infrared spectroscopy (ATR-IR, Nicolet 6700, Thermo, MA, USA), and thermogravimetric analysis (TGA, Q50, TA Instruments, DE, USA) with mass spectrometry (MS, GSD30173, Thermostar, Asslar, Germany) were conducted to determine whether they were fully converted or not. Further, to measure the hydrophobicity of the separators, contact angles were measured using a contact angle analyzer (Phoeaix 300, S.E.O, Ansung, Korea) by dropping a liquid electrolyte on the separators. To clarify the compatibility between particle coating layers and the non-woven supporting layer of the TR-PBO nano-composite membrane, the membranes were subjected to sonication (5510EDTH, Bransonic Ultrasonic Co. Ltd., USA) for 60 min and 100 min. Thickness changes were estimated from SEM images. Furthermore, the porosity of the separators was measured by using a capillary flow porometer (CFP-1500-AE, Porous Materials Inc.). 2.7. Electrochemical measurements Coin-type full-cells (CR 2032) were employed for the electrochemical measurements. Natural graphite (Wellcos, Japan) was used for the anode, LiCoO2 (LCO, Wellcos, Japan) for the cathode, and 1 M LiPF6 in ethylene carbonate (EC)/diethyl carbonate (DEC) (1/1, v/v, Techno Semichem Co. Ltd., Korea) was used as a liquid electrolyte. A PP separator (Celgard 2400, Charlotte, USA) was used as a reference separator. A non-coating TR-PBO non-woven membrane (pristine membrane), HPI1, and HPI4 particle-coated nanocomposite membranes (TR-PBO 1 membrane and TR-PBO 4 membrane, respectively) were also used as separators for
261
electrochemical testing (see Table 2). The electrodes and the separators were dried at 100 C for 20 min. Subsequently, assembly of cells was performed in a glove box under an argon atmosphere by stacking the anode, separator, and cathode and filling the liquid electrolyte to activate the cell. To test the cycle performance of the cells, the cells were cycled at a constant charge/discharge current density of 1 C/1 C in a voltage range from 3.0 to 4.2 V at 30 C. 3. Results and discussion 3.1. Structure control by stabilizer The HPI particle size and shape were controlled by changing the concentration of two kinds of stabilizers, NMP and PVA. Table 1 provides a summary of sample names along with the corresponding concentration of stabilizers and their characteristics. Fig. 1 shows SEM images of as-prepared HPI nanoparticles. HPI1 nanoparticles showed a smooth surface and spherical shape with an average diameter of 700 nm, as shown in Fig. 1(a); these nanoparticles were prepared with 2 wt% PVP stabilizer in water. However, nanoparticles produced in a 2 wt% PVA solution (Fig. 1(b)) exhibited interesting sea-squirt shapes. When the HPI concentration in water changed from 2 wt% to 5 wt%, as shown in Fig. 1(c) and (d), respectively, the shape of the sea-squirt nanoparticles changed dramatically. TEM images of spherical spheres and sea-squirt structures are shown in Fig. 1(e) and (f). TEM images of other seasquirt structures, i.e., HPI3 and HPI4, were similar to the TEM image of HPI2. As shown in Fig. 1(e) and (f), TEM images of HPI1 nanoparticles exhibited whole spherical particles, where HPI2 nanoparticles originate from agglomeration of primary particles to form sea-squirt shaped nanoparticles. The stabilizer was adsorbed on the surface of primary particles during particle formation and provided a barrier to further particle growth due to the steric hindrance of the long chains of the stabilizer molecules. A consequence of this is that a low concentration of the stabilizer or a low stabilizer molecular weight induces coalescence between primary particles. On the other hand, a high concentration or a high molecular weight leads to agglomeration between primary particles because of steric hindrance. This paper is focused on the particle structure at the membrane surface for high electrochemical performance in LIBs. So, we performed a comparative analysis of two samples having a sphere structure (HPI1) and a sea-squirt structure (HPI3). TR-PBO nanoparticles were successfully prepared by thermal treatment of HPI nanoparticles at 400 C for 2 h. As shown in Fig. S1, the morphologies of HPI and TR-PBO nanoparticles can be categorized into two groups by their shapes, i.e., (a) and (b) spheres, and (c) and (d) sea-squirts. Even after HPIs were fully converted into TRPBOs, their shape was maintained. Three TR-PBO membranes were prepared. First, an HPI membrane was prepared by azeotrope imidization, and a pristine electrospun membrane was made without particle coating and was thermally rearranged at 400 C for 2 h. Second, spherically shaped TR-PBO 1 nanoparticles were coated onto a pristine membrane with thermal treatment at 400 C for 2 h. Third, sea-squirt shaped
Table 1 Characteristics of particle size and shape prepared with different concentrations of stabilizers. Sample name
Stabilizers
HPI (wt%)
Size (nm)
Shape
TR-PBO particles
Membrane
HPI1 HPI2 HPI3 HPI4
PVP 2% PVA 2% PVP 5% PVP 5%
2 2 2 5
550 900 900 700
Sphere Sea-squirt Sea-squirt Sea-squirt
TR-PBO TR-PBO TR-PBO TR-PBO
TR-PBO TR-PBO TR-PBO TR-PBO
Adding speed ¼ 0.05 ml/min, temperature was 70 C, stirring speed of 300 rpm.
1 2 3 4
1 2 3 4
membrane membrane membrane membrane
262
M.J. Lee et al. / Journal of Power Sources 305 (2016) 259e266
Table 2 Summary of basic properties of all separators.
Composition Thickness (mm) Pore size (mm) Porosity (%) Contact angle ( ) (Avg/STD)
Celgard® 2400
Pristine membrane
TR-PBO 1 membrane
TR-PBO 4 membrane
PP 25 0.043 41 ± 1.2 42.83/1.83
PBO 18 0.7563 63 ± 0.7 23.75/2.75
PBO e PBO 33 0.643 62 ± 1.5 24/3.02
PBO e PBO 40 0.5579 62 ± 0.3 N/A
Fig. 1. SEM and TEM images of spherical and sea-squirt shaped nanoparticles for (a) HPI1, (b) HPI2, (c) HPI3, and (d) HPI4. TEM images are provided for (e) HPI1 and (f) HPI2. Spherical HPI nanoparticles with average size of 900 nm (a) and sea-squirt shaped nanoparticles with average size of 700 nm in diameter can be seen in (b).
Fig. 2. SEM images of HPI and TR-PBO nanocomposite membranes: (a) front and (b) cross-section of pristine membrane, (c) front and (d) cross-section of the TR-PBO 1 membrane, and (e) front and (f) cross-section of the TR-PBO 4 membrane.
M.J. Lee et al. / Journal of Power Sources 305 (2016) 259e266
TR-PBO 4 nanoparticles were coated onto a pristine membrane with thermal treatment at 400 C for 2 h. The morphologies of the membranes were investigated by SEM, as shown in Fig. 2. The pristine membrane displayed a macro-porous nonwoven surface and interconnected nanofibers merged together, as seen in the front view (Fig. 2(a)), while no particles were observed in a crosssectional view (Fig. 2(b)). In contrast, the TR-PBO 1 membrane had a closely arranged surface (Fig. 2(c)), and the particle layer had a thickness of approximately 5e6 mm (Fig. 2(d)). Finally, the TR-PBO 4 membrane exhibited (Fig. 2(e)) a relatively more porous surface than the TR-PBO 1 (Fig. 2(c)). This is due to the polydisperse size distribution of TR-PBO 4 nanoparticles, and the TR-PBO 4 sample showed that the particle coating layer had a thickness of approximately 8e9 mm (Fig. 2(f)). As shown in Fig. S2, the structural differences between HPI and TR-PBO were clarified using ATR-IR. First, there was no broad and strong band observed between 3200 cm1 and 3600 cm1 because the hydroxyl groups in HPI copolymers were previously converted into polybenzoxazole. Second, there were two imide bands (eCeO) at 1777 cm1 and 1718 cm1. ODPA-HAB5-ODA5 is a co-polymer of polybenzoxazole (ODPA-HAB) and polyimide (ODPA-ODA). In addition, because the TR-PBO copolymer membranes were not derived from hexafluorinated monomers, the benzoxazole bands did not appear at 1059 cm1, but shifted to 1052 cm1. However, a bezoxazole band at 1475 cm1 appeared at the same position in both HPI and TR-PBO samples. Above all, though the pristine membrane indicated an explicit benzoxazole band at 1052 cm1, the TR-PBO 1 and TR-PBO 4 membranes displayed ambiguous bands in this region, because these were in the form of particles [28] and contained the TR-co-polymer, as shown in Fig. S2. The wettability of the separators is critical to the electrochemical performance. Wettability was evaluated using a contact angle analyzer by dropping non-aqueous liquid electrolyte used in all electrochemical tests (ED:DEC ¼ 1:1, w:w, 1 M LiPF6) onto the membrane surface. The wettability is influenced by the surface structure and chemical properties. As already discussed, these TRPBO series membranes have similar functional groups, as shown in Fig. S2. Fig. 3 shows that the contact angle was low for the TRPBO 4 membrane (d) as compared to pristine membrane (b). By contrast, even though the TR-PBO 1 membrane (Fig. 3(c)) was particle coated, the contact angle was similar to that of the non-
263
coated pristine membrane. Because the surface area of the TRPBO 4 membrane increased greatly and the surface pore size decreased substantially, the electrolyte was absorbed instantly due to an induced capillary effect. However, while the TR-PBO 1 membrane had increased surface area, the particle layer was more closely packed, and the capillary effect was not observed. In any case, all the TR-PBO membranes show higher wettability than Celgard because of the polarity between the microporous TR-PBO copolymer separators and the highly polar liquid electrolyte [29]. Therefore, the electrolyte uptake of the TR-PBO copolymer separators was measured to be greater than that of the Celgard. Table 2 provides a summary of the basic properties of all separators, including porosity. Thermal stability of the separators evaluated by TGA and DSC showed similar results in comparison with our previous report. As can be seen in Figure S3(a), the TR-PBO 1 and the TR-PBO 4 membranes showed no weight loss until degradation occurred at 600 C, while the Celgard was decomposed below 400 C. TGA analysis showed that not only were the TR-PBO membranes fully converted, but the TR-PBO membranes also showed superior thermal properties compared to the Celgard. Additional thermal analysis was conducted using DSC, and these data are provided in Figure S3(b). The TR-PBO membranes showed a straight line in the DSC data until 250 C, while the Celgard membrane displayed an endothermic peak at 160 C, characteristic of the melting point of polypropylene. Consequently, the TR-PBO membranes provide superior thermal stability in comparison with the Celgard based on DSC and TGA analyses. Fig. S4 shows digital pictures of (a and d) Celgard 2400, (b and e) pristine, (c and f) TR-PBO 1, and (d and g) TR-PBO 4 membranes before and after a hot oven test at 220 C for 30 min. TR-PBO series membrane maintained its integrity even after a hot oven test at high temperature, while the Celgard 2400 membrane made of polypropylene was completely contracted, as shown in Figure S4(e). 3.2. Electrochemical properties The battery performances of the two different TR-PBO based cells were evaluated with a LiCoO2/graphite electrochemical couple. Fig. S5 show the chargeedischarge curves of the TR-PBO 4, TR-PBO 1, and the Celgard 2400 based cells, where the voltage
Fig. 3. Contact angle analysis of the separator membranes. Drops were stabilized in 0.42 s and images were taken. The membranes were (a) Celgard 2400, (b) pristine membrane, (c) TR-PBO 1 membrane, and (d) TR-PBO 4 membrane. The hydrophobicity trend is (d) > (c) > (b) > (a).
264
M.J. Lee et al. / Journal of Power Sources 305 (2016) 259e266
ranged between 3.0 and 4.2 V with charge/discharge current rates at two initial cycles (0.05 C/0.05 C) up to the 3rd cycle (1.0 C/1.0 C). All samples show typical charge/discharge curves of the coin-cell assembled with LiCoO2 and graphite electrodes as cathode and anode. Fig. 4 shows the cycling performances of the TR-PBO 4, TRPBO 1, and the Celgard 2400 based cells, where the voltage ranged between 3.0 and 4.2 V with charge/discharge current rates at two initial cycles (0.05 C/0.05 C) up to the 100th cycle (1.0 C/1.0 C). Initial discharge capacities of the three cells are not significantly different at a low current rate (0.05 C). At a higher C-rate (1.0 C), however, the discharge capacities of TR-PBO 4, TR-PBO 1, and Celgard 2400 based cells were 103 mAh g1, 78 mAh g1, and 56 mAh g1, respectively. Two different TR-PBO based cells exhibited higher capacities than the Celgard 2400 based cell due to their excellent wettability, which can improve the Li-ion mobility [19]. As shown in Fig. 3, the wettability of the TR-PBO series membranes is greater than that of the Celgard 2400 membrane. Further, the discharge capacity of TR-PBO 4 based cell is 25% higher than that of the TR-PBO 1 based cell. This result is very closely related to the porosity of the membranes. In this experiment, two types of TR-PBO particles having different morphologies were utilized to control the porosity of electrospun TR-PBO membranes having macropores caused by deposition of one-dimensional TRPBO strings. As shown in Table 1, the TR-PBO 4 membrane coated with sea-squirt TR-PBO nanoparticles has a smaller average pore size than that of the TR-PBO 1 membrane. In addition, the narrow pore size distribution (at a maximum diameter, TR-PBO 4: 94.2% and TR-PBO 1: 93.7%) of the TR-PBO 4 membrane has an influence on the higher capacity. These characteristics of the TR-PBO 4 membrane result in better lithium-ion transport during chargeedischarge processes because the sea-squirt structures of the TRPBO 4 nanoparticles can make small porous channels that accelerate ion transport. In order to understand the effect of membrane pore size on the lithium-ion transport, electrochemical impedance spectroscopy (EIS) was employed. Fig. 5 shows Nyquist plots of the electrochemical impedance spectra obtained from the fresh TR-PBO 4, TRPBO 1, and Celgard 2400 based cells (before cycling) and those of the cells after cycling at a 1 C rate. The coin cells used for this measurement were made by the same fabrication process. EIS analysis can provide electrochemical information such as charge and mass transfer resistances, and correlation between the electrochemical performance and electrode kinetics. Fig. 5(a) shows the
Fig. 4. Cycle performance of cells assembled with TR-PBO 4, TR-PBO 1 and Celgard 2400 at a 1 C rate. The cells were operated in the voltage range of 3.0e4.2 V at 30 C.
Nyquist plots of the fresh cells containing TR-PBO 4, TR-PBO 1, and Celgard 2400 separators. The Nyquist plots are composed of one semicircle in the high frequency range and a sloping straight line at low frequency, which can be assigned to the impedance of the charge transfer resistance (Rct) coming from the interface between the electrode and electrolyte, and the Warburg impedance (W), related to the diffusion of lithium ions into the graphite anode phase. Charge-transfer resistances of TR-PBO 4, TR-PBO 1, and Celgard based cells were 48.2, 53.1, and 57.5 U, respectively. The TRPBO 4 based cell exhibited the lowest charge transfer and mass transfer resistances because of the excellent wettability and high porosity. After 10 cycles, all impedance spectra were composed of two partially overlapping semicircles at high and middle frequency ranges, which can be assigned to the impedance of the solid electrolyte interface (RSEI) and the charge transfer resistance, respectively, with a sloping straight Warburg line at low frequency. After the first cycle, byproducts are generated by decomposition of the electrolyte, which can act as resistive layers for delaying ionic transport. Thus, the impedance of the solid electrolyte interface was obtained at low frequency, as shown in Fig. 5(b). However, the charge transfer resistances of all cells decreased after 10 cycles. The charge transfer resistances of TR-PBO 4, TR-PBO 1, and Celgard 2400 based cells were 5, 6 and 7 U, respectively, which indicates faster charge transfer kinetics in the TR-PBO-based cells. Such a phenomenon is because the irreversible SEI layer formed on the electrode slows the increase in cell impedance by protecting the electrode from chemically reacting with the electrolyte. After cycling, the charge transfer resistances of TR-PBO based cells were still smaller than that of the Celgard 2400 based cell. In particular, the TR-PBO 4 based cell exhibited lower charge and mass transfer resistances than those of the TR-PBO 1 based cell after 10 cycles because the TR-PBO 4 based cell has a high porosity caused by seasquirt like nanoparticles as well as the excellent affinity of the TRPBO with the organic electrolyte. These results suggest that the porosity and wettability of the separator have an effect on the kinetics of the cell electrochemical reaction and lithium-ion transport during cycling. The TR-PBO polymer has good thermal stability, as shown in Fig. S3. These thermal properties can significantly affect the electrochemical properties such as cycle and capability at high temperature. Fig. 6 shows the cyclability of three different cells containing TR-PBO 4, TR-PBO 1, and Celgard 2400 separators at same current rate of 1 C at a constant temperature of 50 C. At high temperatures, the cyclability of LIB cells is generally poor because the reactivity between the electrodes and electrolyte rapidly increases. In this test, interesting results were obtained. Initial discharge capacities of TR-PBO 4, TR-PBO 1, and Celgard 2400 based cells operated at 50 C were 291, 283, and 272 mAh g1, respectively. These values were similar to the initial discharge capacities of cells operated at room temperature. However, after 100 cycles at 50 C, the Celgard based cell showed lower capacities during continuous cycling up to the 100th cycle because of the poor thermal stability of the monolayer polypropylene. On the other hand, the TR-PBO 4 based cell exhibited much higher capacities from the 4th cycle up to the 100th cycle than even that observed for room temperature testing. These results are attributed to the excellent thermal stability of the TR-PBO polymer and the high porosity of the membrane. Since testing at high temperature leads to better Li-ion diffusivity due to the low electrolyte viscosity at high temperature, TR-PBO 4 based cells showed higher capacities even at high current rate. To compare the thermal stability of the membranes, we carried out a hot oven test at 220 C for 30 min Fig. S4 shows digital pictures of Celgard 2400 and TR-PBO series membranes having a uniform size (2 cm2) before and after hot oven testing. TR-PBO series membranes maintained their original
M.J. Lee et al. / Journal of Power Sources 305 (2016) 259e266
265
Fig. 5. Electrochemical impedance (EIS) Nyquist plots of the lithium cells of the TR-PBO separators compared with Celgard 2400 as a reference (frequency range 5e1000 Hz at amplitude of 5 mV) at (a) first cycle and (b) after 10 cycles with 1 C current charge/discharge.
Fig. 6. Cycle performance of cells assembled with selected separators in 1 C current density and compared with Celgard 2400 as a reference. The cells were operated in the voltage range between 3.0 and 4.2 V at 50 C.
membrane size even above 200 C, in contrast with a contracted Celgard 2400 membrane. Furthermore, the porous structure of the Celgard 2400 membrane completely disappeared after hot oven testing, as shown in Figure S6(d). On the other hand, the intrinsic surface structure of TR-PBO series membranes was not changed, as shown in Figures S6(e) and (f). Therefore, TR-PBO separators coated with TR-PBO particles significantly improve battery safety and exhibited high capacity retention during high temperature testing.
of the electrolyte. Thermal stability of these membranes was excellent when compared to Celgard, as confirmed by TGA and DSC analyses. The particle coating on TR-PBO membranes decreased the overall pore size, and the sea-squirt shaped particle coated membrane (TR-PBO 4 membrane) displayed smaller pore size and broader pore size distribution than the spherically shaped particle coated membrane (TR-PBO 1 membrane). The wettability of the TRPBO membranes was superior to the Celgard, while the TR-PBO 4 membrane exhibited instant absorption of electrolytes due to its large surface area. Moreover, the thermal stability of the TR-PBO membranes surpassed that of the Celgard. TR-PBO membranes were stable up to 490 C, while the Celgard decomposed at 340 C, as evidenced from TGA analysis. TR-PBO membranes did not show any melting up to 250 C, whereas the Celgard showed melting at 160 C based on DSC analysis. Finally, the cyclability test validated the superiority of the TR-PBO 4 membrane among the TR-PBO membranes due to the fact that a sea-squirt shaped particle coating layer facilitated ion transport and improved wettability of the electrolyte. Meanwhile, all the TR-PBO membranes displayed better cycle performance compared to the Celgard membrane, which demonstrates that TR-PBO membranes are excellent candidates for high density LIB separators. Acknowledgments This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (2012M3A7B4049745). This work was supported by the Global Frontier R&D Program (2013M3A6B1078875) on Center for Hybrid Interface Materials (HIM) funded by the Ministry of Science, ICT & Future Planning.
4. Conclusions
Appendix A. Supplementary data
Shape tunable TR-PBO nanoparticles were successfully fabricated by a re-precipitation method and subsequent thermal treatment. The parameters influencing the particle shape were investigated, and a possible particle formation mechanism was suggested. Meanwhile, spherically shaped HPI1 and sea-squirt shaped HPI4 nanoparticles were coated onto the pristine membrane and were successfully converted to TR-PBO 1 and TR-PBO 4 membrane nano-composite membranes, respectively. Both HPI and TR-PBO nanoparticles and TR-PBO membranes were investigated by ATR-IR and TGA analysis to confirm the structural evolution. Moreover, the TR-PBO membranes, including the pristine membrane, showed good pore size distribution and excellent wettability
Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.jpowsour.2015.11.068. References [1] [2] [3] [4] [5] [6] [7] [8]
M.H. Ryou, Y.M. Lee, J.K. Park, J.W. Choi, Adv. Mater. 23 (2011) 3066e3070. K. Hwang, B. Kwon, H. Byun, J. Membr. Sci. 378 (2011) 111e116. J. Hassoun, B. Scrosati, Adv. Mater. 22 (2010) 5198e5201. C.J. Orendorff, T.N. Lambert, C.A. Chavez, M. Bencomo, K.R. Fenton, Adv. Energy Mater. 3 (2013) 314e320. K. Xu, Chem. Rev. 104 (2004) 4303e4418. H. Maleki, J.N. Howard, J. Power Sources 191 (2009) 568e574. J.Y. Kim, Y.B. Lee, D.Y. Lim, Electrochim. Acta 54 (2009) 3714e3719. J.Y. Lee, Y.M. Lee, B. Bhattacharya, Y.C. Nho, J.K. Park, Electrochim. Acta 54
266
M.J. Lee et al. / Journal of Power Sources 305 (2016) 259e266
(2009) 4312e4315. [9] J.-A. Choi, S.H. Kim, D.-W. Kim, J. Power Sources 195 (2010) 6192e6196. [10] J. Song, M.H. Ryou, B. Son, J.N. Lee, D.J. Lee, Y.M. Lee, J.W. Choi, J.-K. Park, Electrochim. Acta 85 (2012) 524e530. [11] J.L. Hao, G.T. Lei, Z.H. Li, L.J. Wu, Q.Z. Xiao, L. Wang, J. Memb. Sci. 428 (2013) 11e16. [12] J. Ding, Y. Kong, P. Li, J. Yang, J. Electrochem. Soc. 159 (2012) A1474eA1480. [13] W. Qi, C. Lu, P. Chen, L. Han, Q. Yu, R. Xu, Mater. Lett. 66 (2012) 239e241. [14] W. Jiang, Z. Liu, Q. Kong, J. Yao, C. Zhang, P. Han, G. Cui, Solid State Ion. 232 (2013) 44e48. [15] Y.-E. Miao, G.-N. Zhu, H. Hou, Y.-Y. Xia, T. Liu, J. Power Sources 226 (2013) 82e86. [16] H.B. Park, C.H. Jung, Y.M. Lee, A.J. Hill, S.J. Pas, S.T. Mudie, E.V. Wagner, B.D. Freeman, D.J. Cookson, Science 318 (2007) 254e258. [17] J.I. Choi, C.H. Jung, S.H. Han, H.B. Park, Y.M. Lee, J. Memb. Sci. 349 (2010) 358e368. [18] S.H. Han, N. Misdan, S. Kim, C.M. Doherty, A.J. Hill, Y.M. Lee, Macromolecules 43 (2010) 7657e7667.
[19] S.H. Han, H.J. Kwon, K.Y. Kim, J.G. Seong, C.H. Park, S. Kim, C.M. Doherty, A.W. Thornton, A.J. Hill, A.E. Lozano, K.A. Berchtold, Y.M. Lee, Phys. Chem. Chem. Phys. 14 (2012) 4365e4373. [20] H.B. Park, S.H. Han, C.H. Jung, Y.M. Lee, A.J. Hill, J. Memb. Sci. 359 (2010) 11e24. [21] S. Kim, S.H. Han, Y.M. Lee, J. Memb. Sci. 403e404 (2012) 169e178. [22] M. Calle, Y.M. Lee, Macromolecules 44 (2011) 1156e1165. [23] M. Calle, Y. Chan, H.J. Jo, Y.M. Lee, Polymer 53 (2012) 2783e2791. [24] C.H. Jung, J.E. Lee, S.H. Han, H.B. Park, Y.M. Lee, J. Memb. Sci. 350 (2010) 301e309. [25] M. Calle, A.E. Lozano, Y.M. Lee, Eur. Polym. J. 48 (2012) 1313e1322. [26] H. Wang, T.-S. Chung, J. Memb. Sci. 385e386 (2011) 86e95. [27] M.J. Lee, J.H. Kim, H.-S. Lim, S.Y. Lee, H.K. Yu, J.H. Kim, J.S. Lee, Y.-K. Sun, M.D. Guiver, K.D. Suh, Y.M. Lee, Chem. Commun. 51 (2015) 2068e2071. [28] J.-Y. Xiong, X.-Y. Liu, S.B. Chen, T.-S. Chung, J. Phys. Chem. B 109 (2005) 13877e13882. [29] Y.-E. Miao, G.-N. Zhu, H. Hou, Y.-Y. Xia, T. Liu, J. Power Sources 226 (2013) 82e86.