Enhanced single-ion conduction and free-standing properties of solid polymer electrolyte by incorporating a polyelectrolyte

A free-standing solid polymer electrolyte (SPE) composed of poly(ethylene oxide) (PEO), bis(trifluoromethanesulfonyl)imide (LiTFSI), and poly(lithium 4-styrenesulfonate) (PLSS) was developed in this work. Thermal analysis indicated that the melting points of PEO were depressed with increasing the salt additives, LiTFSI and PLSS. At the composition of [EO]:LiTFSI:[LSS] = 14:1:1, the SPE exhibited a crystallinity of 7.75%, and a crystallite size of 30.62 nm on the (120) crystallographic plane. [EO] and [LSS] represent the structural unit of PEO and PLSS, respectively. This SPE also exhibited an ionic conductivity (σ) of 1.70 × 10−5 S cm−1 at 25 °C and 1.04 × 10−4 S cm−1 at 45 °C. For analyzing the temperature dependence of σ, the Vogel-Tammann-Fulcher equation was employed. Resultantly, a pseudo activation energy (E a = 0.1552 eV), a prefactor (A = 206.0338 S cm−1 K1/2), and an empirical constant (B = 1800.5879 K) were obtained using the optimized [EO]:[Li+] = 7:1 complex. The SPE showed an electrochemical stability window of ∼4.7 ± 0.1 V versus Li/Li+. Through DC polarization and AC impedance, the Li-ion transference number of 0.66 was obtained at 70 °C. Finally, when a Li/SPE/LiFePO4 cell was prepared, the device exhibited a discharge capacity of 121 mAh g−1 at 50 °C with a coulombic efficiency close to 100%.


Introduction
Li-ion batteries (LIBs) have been the most commercially successful primary power source for portable electronic devices due to their high energy density, low rate of self-discharge, low weight and zero memory effect [1][2][3]. Their applications have also extended to electric vehicles and various grid-scale energy storage applications [4][5][6][7]. However, LIBs have various limitations including safety problems related with liquid electrolytes [8][9][10][11]. Hence, attempts have been made to replace this liquid electrolyte with a solid polymer electrolyte (SPE) to solve this leakage problem. The origin of SPE could be ascribed to the fundamental observation that there is an ionic nature in poly(ethylene oxide) (PEO)-salt complexes [12]. It was suggested that such complexes could be used as a solid electrolyte for electrochemical devices [13]. Since then, there have been tremendous research efforts to develop SPEs for thin and safe LIBs and supercapacitors [14]. SPEs have many practical advantages such as minimized dendritic formation, design flexibility, high voltage usage, low cost fabrication, and size scalability, which are desirable for the safe operation of energy storage devices.
The single-ion conducting polymer electrolyte has received great attention because a regular dual-ion conductor leads to a concentration gradient and cell polarization, resulting in a lower power output during discharge, higher internal impedance, and premature battery failure [48][49][50]. For solving these problems, the polyelectrolyte (also called an ionomer) with counter ions bound to the backbone has been a topic of interests [20,51]. Through this approach, it is possible to achieve a cation transference number close to unity in principle. However, in general, many single-ion conductors have inferior ionic conductivity compared to dual-ion conductors because of reduced charge concentration and/or mobility [52]. For example, a PEO-poly(lithium 4-styrenesulfonate) (PLSS; also called LiPSS) complex exhibited a very low ionic conductivity (σ=3.0×10 −8 S/cm at 25°C), although the Li + -transference number ( + t Li =0.85 at 70°C) was high enough [53]. To overcome this low σ, some researchers added a plasticizer into the PEO-PLSS complexes, and others designed completely new polyelectrolyte [20,[54][55][56][57][58][59].
In this study, a ternary blend composed of PEO, LiTFSI and PLSS for improving the performance of binary PEO-LiTFSI and PEO-PLSS complexes was investigated. A conventional PEO-LiTFSI electrolyte is very sticky without dimensional stability in the amorphous state [42], whereas a single-ion conductor PEO-PLSS has a small ionic conductivity (σ=∼10 −8 S cm −1 ) [53]. To solve these problems, the ternary complexes with different [EO]:[Li + ] molar ratios were prepared in order to have a balance property among ionic conductivity, transference number and mechanical stability. Then, the optical, thermal, structural, and electrochemical properties of SPE materials were examined. The melting point depression phenomena [60,61] were observed when the ternary composition was changed. The temperature dependence of ionic conductivity was analyzed in terms of the empirical Vogel-Tammann-Fulcher (VTF) equation [62][63][64]. As a result, the prefactor (A), empirical constant (B), and pseudo activation energy (E a ) were obtained. Then the electrochemical stability window and Li + -transference number were characterized. Finally, the charge-discharge properties were examined using the Li/SPE/LiFePO 4 cell.
Preparation of polymer electrolyte PLSS was synthesized using ion exchange method according to literature report [53]. The solutions of PEO/ LiTFSI/PLSS were prepared in the glove box. PEO was first dissolved in acetonitrile overnight. Then LiTFSI and PLSS were added into the solution and then stirred further for 24 h. Then, this viscous electrolyte solution was poured on a Teflon plate and left to evaporate acetonitrile slowly at 30°C, leading to a self-standing film. The prepared thin-film was dried under high vacuum at 60°C for 24 h to remove any residual solvent. The thickness of the film ranged from 85-120 μm. PEO, LiTFSI and PLSS were mixed in molar ratio of [LSS]=14:1:1 complex (as a binder and Li-ion conductor), and Super-P with the ratio of 70:20:10 (wt.%). After ball milling for 3 h, the slurry solution was coated on 15 μm thick aluminum foil and dried under vacuum at 110°C for 12 h. The electrode was pressed with a roll pressed machine (WV-60, Samyang 60, Korea) and then punched into a circular disc of 7 mm radius. The active mass loading of the positive electrodes was 4.2 mg cm −2 . Together with SPE and the Li metal as a counter/reference electrode, the composite positive electrode was assembled with a CR-2032 coin-type half-cell in the argon glove box for charge-discharge cycling performance measurements.

Materials characterization
Chemical structures (functional groups) were identified using Fourier-transform infrared spectroscopy (FT-IR, Nicolet iS5, USA). Thermal analyses were carried out using differential scanning calorimetry (DSC, Mattler-Toledo STAR e system, version 12.1) under N 2 at the scanning rate of 10°C min −1 . An x-ray diffractometer (XRD, Rigaku Smartlab, Japan) was used to figure out the crystallinity of SPE under Cu-Kα (=1.5406 Å) radiation at 40 kV and 40 mA. The morphologies of SPE were investigated using field-emission scanning electron microscope (FE-SEM, Hitachi S-4800, Japan) at 5 kV electron beam energy. All the electrochemical performances were carried out using Autolab electrochemical workstation (ECO CHEMIE, PGSTAT 100). Some properties of materials were characterized while the electrochemical cell was placed in an oven. Electrochemical impedance spectroscopy (EIS) measurements were also carried out to obtain Nyquist plots, leading to the calculation of ionic conductivity of a sample. The EIS spectra were characterized at different temperatures (∼25-90°C) with an applied potential difference of 10 mV and frequency range of 0.01 Hz to 10 6 Hz. The free-standing SPE samples were inserted between two stainless-steel electrodes with an active area of 1×1 cm 2 and then compacted inside a pouch cell using vacuum sealer. Linear sweep voltammetry (LSV) was carried out to determine the electrochemical potential window of the SPE samples, for which stainless steel and Li metal foil were used as working and counter (reference) electrodes, respectively. The voltage was swept from 0 to 7 V versus Li/Li + with a scanning rate of 1 mV s −1 at room temperature. The transference number of Li-ion ( + t Li ) was estimated using the steady-state current method. The electrolyte was inserted between two Li metal electrodes and assembled with a CR-2032 coin type cell in the glove box. The Li metal surface was scraped and cleaned with a scalpel prior to putting the components together in the coin cell. The AC impedance measurement was conducted before and after potentiostat polarization at 70°C [53,54]. The symmetric cell was polarized by applying a constant voltage of 10 mV. The charge-discharge dynamic performance of the Li/SPE/LiFePO 4 coin type half-cell was analyzed using a TOSCAT-3000 battery charge/discharge test system (Toyo System Co., Ltd, Japan). The half-cell was assembled in the glove box and then stored in a vacuum oven at 50°C for 12 hr in order to enhance interfacial contact between the SPE and the composite positive electrode. The charge-discharge cycle property of the cell was analyzed using a constant current and constant voltage (CC-CV) input charging method with different C-rates (0.1 C, 0.2 C, 0.5 C and 1.0 C) and then discharges at the same rate with a voltage range of 3-4.5 V.

Results and discussion
Chemical structures and FT-IR spectra analysis Figure 1 shows the chemical structures of PEO, LiTFSI, and PLSS, in which PEO is a polar polymer with the donicity of ∼20 to provide excellent solubility for Li salts [16], LiTFSI is a plasticizing salt with well delocalized anions, and PLSS is a polyelectrolyte/ionomer with bound anions in the sulfonate group. These molecules comprise the main components of the ternary SPE in this work. Figure 2(a) shows the FT-IR spectra for PLSS sample, in which the prominent peaks at 1500 cm −1 and 1430 cm −1 reveal the S=O antisymmetric vibration. The peaks at 1070 cm −1 and 833 cm −1 are attributed to the symmetric vibration of S-O. This result suggests that PLSS was successfully synthesized [53,65]. Figure 2(b) shows the FT-IR spectra of various polymer electrolytes at different salt, PLSS and LiTFSI composition. From pure PEO, the characteristic vibrational peaks exhibits an asymmetric aliphatic C-H stretching mode at 2877 cm −1 , CH 2 asymmetric scissoring mode at ∼1466 and 1454 cm −1 , CH 2 wagging mode at 1340 and 1360 cm −1 , CH 2 asymmetric twisting at 1279 and 1240 cm −1 , CH 2 asymmetric rocking at 960 and 946 cm −1 (shoulder type) , C-O stretching and CH 2 rocking mode at 841 cm −1 , and strong C-O-C stretching bands which are splits into three peaks at ∼1143 cm −1 , 1093 cm −1 , and 1059 cm −1 , respectively [66]. The presence of CH 2 bending vibrations and C-O-C triplet peaks demonstrates the presence of PEO crystallinity [67].
With the addition of LiTFSI and PLSS into PEO, the change in intensity and position of spectral patterns are notably observed. The CH 2 asymmetric scissoring peak has excessively broadened, implying interaction between PEO chains and the added salts. It was observed that the spectra is first largely influenced by the concentration of PLSS, which corresponds to PEO-PLSS blend and it was found to be broader as the concentration of PLSS and LiTFSI increases [68]. The triplet stretching peaks from C-O-C ether linkages became wider and shifted down to 1132 cm −1 , 1093 cm −1 , and 1055 cm −1 with increasing salt concentration which is attributed to the complexation of cations with ether oxygen. As a result, the intensity of CH 2 asymmetric twisting spectra also decreased and downshifted to lower wavenumbers and the CH 2 asymmetric rocking modes became a single less intense band that appears at about 949 cm −1 . These observations indicate that the mixing remarkably enhanced the overall disordered conformation of PEO chain molecules [69,70].
The new peak observed at 1184 cm −1 is ascribed to the TFSI anion's plasticizing effect within the polymer electrolyte complex [71]. The sharp peak at 841 cm −1 which corresponds to C-O stretching and CH 2 rocking modes became broaden and split into two peaks, appearing at 866 and 843 cm −1 . In the spectrum of the PEO-LiTFSI complex, a set of three peaks were observed at 785 cm −1 , 761 cm −1 , and 734 cm −1 due to the S-N-S asymmetric vibration, the CF 3 asymmetric bending vibration mode, and the S-N-S symmetric stretching mode, respectively. This absorption slightly shifted to higher wavenumbers in the PEO-LiTFSI-PLSS blend and appeared at 786 cm −1 , 761 cm −1 , and 738 cm −1 . The little broadening of the symmetric S-N vibration indicates LiTFSI is well dissolved into the host polymer, PEO.
Thermal property and melting point depression Solid to liquid (S-L) phase transition is the melting of crystalline regions, increasing additional active segmental motions in amorphous regions. Figure 3 shows the DSC thermograms for the SPE samples in the range of -50 to 100°C, in which -50°C was limited by the DSC instrument itself although PEO's T g is known to be about -60°C [13,[15][16][17][18][19]. The melting points were taken at the peak maximum. Figure 3(a) shows the different roles between  causes the binary PEO-PLSS to become phase-separated when mixed, which allows semicrystalline PEO to survive with some amounts of crystalline regions. Here, it is noteworthy that not only single-phase but also phase-separated polymer blends could be used in many practical applications [72][73][74]. It is clearly observed that the melting temperature of PEO in a ternary system was depressed when mixed with both LiTFSI and PLSS. The reason is that the presence of Li salts, acting as diluent or impurity from the viewpoint of PEO, may shift the S-L phase equilibrium point between the liquid and the crystalline chains of PEO, although LiTFSI and PLSS have different intermolecular interactions with PEO chains. Finally, the results of the melting point depression in the ternary PEO-LiTFIS-PLSS system were clearly plotted in figure 4 based on the data in figure 3 Crystal structure analysis through XRD patterns The XRD data for the blend polymer electrolytes is illustrated in figure 5. The x-ray pattern of pure PEO reveals two strong crystalline peaks at 19.1°and 23.2°, corresponding to (120) and (032) crystallographic planes, respectively [76]. This observation indicates that the PEO host is semicrystalline. The peak intensity was reduced and widen upon the addition of LiTFSI and PLSS salts. For low concentrations of LiTFSI and PLSS salt, the intensity of the second characteristic peak of PEO at 2θ=23.2°was mainly affected and turned into broaden and short peak. As can be seen from the figure 5, the two PEO peaks became relatively broader as well as less prominent as the LiTFSI and PLSS content increased in PEO. This reveals that Li + ion and C-O-C segment coordination interaction alters the ordered arrangements PEO chains while promoting salt dissociation and increasing charge carriers. The absence of sharp peaks in the [EO]:LiTFSI:[LSS]=14:1:1 indicates that PEO exhibits no crystalline characteristics and suggests a complete dissolution of the salts into PEO host matrix. This observation is consistent with the information obtained from DSC. Note that the crystallite size (t) could be calculated based on Scherrer's equation [77,78], where λ (=0.154 nm) is the x-ray wavelength, β is the full width at half maximum (FWHM), and θ is the diffraction angle of x-ray light. As shown in table 2, although other component can destroy the PEO's crystalline regions (i.e. decrease of crystallinity), the crystallite size (average single crystalline domains in polycrystals) slightly increases when blended.       (=∼55°C-75°C). On the other hand, when c < 20%, c the transition of ionic conductivity is not readily discernable at the aforementioned melting range, owing to a relatively minor increase of amorphous regions.

Ionic conductivity and Vogel-Tammann-Fulcher equation
When the temperature is higher than 1.15×T g (e.g. PEO: 1.15×213.15=245.15 K), it is known that VTF equation could be applied to the explanation of s T log versus1000 plot. As shown in figure 6(b), the plots have non-constant slopes along each line of the measured conductivities, suggesting that the ionic conductivity is supported by the inter-and intra-chain segmental motion of the polymer, i.e. obeys the VTF conduction mechanism. Due to this observation, to understand the ionic transport process in the PEO-LiTFSI-PLSS electrolyte, we employed VTF equation [62,63], where A, B, and T 0 are a prefactor, an empirical constant with unit of temperature (related with expansivity), and the equilibrium glass transition temperature (called Vogel temperature), respectively. Here, T 0 is known to be smaller than the glass transition temperature (T g ) by ∼45-50 K [43,62,63]. In this work, we used T 0 =T g −50 K [43,62]. At this moment, it is noteworthy that, until now, most authors have obtained the three VTF parameters (A, B, and T 0 ) through simple fitting methods [16,62,79]. However, in this work, we used PEO's T g (=213.15 K) [15][16][17][18][19] for estimating T 0 (=213.15−50=163.15 K), and obtained the other two VTF parameters (A and B) through a fitting method. From B, we may estimate pseudo activation energy (E a =B×k) analogous to the activation energy, in which k=8.61739×10 −5 eV K −1 is Boltzmann constant. Thus, based on the data in figure 6(b), we fitted the data using the plot of ( · ) ( and in which n is the concentration of charge, q is charge, μ is mobility, D is diffusion coefficient, and C is a constant. Hence, A is physically related with the concentration of charges. However, the constant C is unknown, making it difficult to compare across systems as a function of composition.

Li-ion transference number
The lithium-ion transference number ( + t Li ) was calculated according to the equation initially proposed by Evans et al [80]  DV is the DC potential applied across the cell, I 0 and I S are the initial and steady-state currents, R 1 0 and R S 1 are the initial and steady-state interfacial resistance measured by the AC impedance method before and after DC polarization, respectively. The ion transference number, which describes the fraction of current carried by a particular ionic species, is just as important as the total conductivity. Increasing the cation transference number close to one may provide a long-lasting battery lifetime. A conventional PEO-LiTFSI complex has shown a poor lithium ion transference number between 0.16-0.30, being independent of temperature [43], whereas PEO-PLSS has a high transference number, 0.85 at 70°C [53]. and after a steady state current has indicating that the cation mobility is higher than that of counter anions. This is because PSS anions are fixed on the polymeric chains and partially disrupt the ordering of the PEO matrix, although PEO and PLSS are in a phase-separated state. Recall that, because of entropic reasons, the polymer-salt PEO-LiTFSI complex is much more miscible than the polymer-polyelectrolyte PEO-PLSS complex.

Electrochemical stability window analysis through linear sweep voltammetry
The electrochemical stability of the polymer electrolyte was measured by linear sweep voltammetric (LSV) measurements at 50°C and 60°C, respectively. Figure 8 presents the LSV curve of the blended electrolyte (sandwiched) on a Li/SPE/SS coin type cell assembled inside a glove box. It is observed that the anodic stability of the blend solid polymer electrolyte (namely, the onset voltage for oxidative decomposition) is at ∼4.7±0.1 V (Li versus Li + ), indicating that the electrochemical stability window of this polymer electrolyte is suitable for practical applications.

Charge-discharge cycling performance
To assess the performance of PEO-LiTFSI-PSS electrolyte in a practical battery, we assembled a coin type halfcell inside a glove box with the Li/SPE/LiFePO 4 configuration. Hence, the cell has a Li metal electrode, composite positive electrode (LiFePO 4 ), and free-standing SPE as electrolyte and separator. Figure 9(a) shows the charge discharge curve of the cell for different cycle numbers at 50°C. The smooth horizontal potential plateau starts at about 3.4 V, related with the intercalation/deintercalation of Li + into the LiFePO 4 structure, leading to stable and reversible cycles. The charge-discharge specific capacities at different C-rate are shown in figure 9(b). Here, C-rate is defined by = C I f, in which I is current and f is a multiple or a fraction of C. For example, if discharged at a current of 12.1 mA in the 121 mAh/g cell, the C-rate would be 0.1C, indicating C-rate is equivalent to current [82]. The specific capacities were found to be 121 mAh g −1 , 105 mAh g −1 , 102 mAh g −1 , and 100 mAh g −1 at 0.1C, 0.2C, 0.5C and 1C, respectively. The corresponding coulombic efficiency began at 87%, which was lower compared to the subsequent cycle efficiencies due to irreversible processes associated with electrolyte decomposition during solid-electrolyte-interface formation. The efficiency improved and increased quickly above 98%, attributed to the high Li + ion transference number and ionic conductivity of the PEO-LiTFSI-PLSS polymer electrolyte. It reached close to 100% when the discharge current density returned to 0.1 C, indicating excellent capacity recovery and reversible lithium deposition-stripping performance.
Single-ion conducting solid polymer electrolyte: optimization Figures 10(a) and (b) display the macroscopic and microscopic SEM images of the ternary PEO-LiTFSI-PLSS film, in which the former demonstrates a semi-translucent free-standing property and the latter exhibits a microscale wrinkled morphology coming from phase-separated semicrystalline PEO and a polyelectrolyte PLSS [recall DSC data in figure 3(b)]. Finally, figure 10(c) sums up this research through the schematic illustration about the concept of this work. In this triangle, the maximum area indicates the ideal single-ion conducting SPE with characteristics of high ionic conductivity, high transference number, and free-standing properties. However, note that in this triangle, we do not compare other important factors such as the stability window and electrolyte-electrode interfacial kinetics. As shown in this scheme, the binary polymer electrolyte systems, PEO-PLSS and PEO-LiTFSI, have serious drawbacks in ionic conductivity and mechanical properties, respectively. Therefore, by combining these together through a ternary approach, we could optimize the property of this SPE, demonstrating that the ternary SPE approach is promising in the development of SPEs.

Conclusions
We demonstrated that a ternary PEO-LiTFSI-PLSS complex is an optimized system in terms of overall ionic conductivity, transference number and free-standing mechanical properties, when compared to a binary PEO-LiTFSI (a sticky film without free-standing property in amorphous state) or PEO-PLSS (with small σ∼10 −8 S cm −1 ). In this study, the major findings are as follows: (1) In the DSC data, we observed melting point depression and the crystallinity of the [EO]:[Li + ]=7:1 complex is 7.75%. (2) Through XRD patterns, we observed that the small molecule LiTFSI salt/ions effectively destroyed the crystalline regions of PEO when compared to a polyelectrolyte PLSS. Note that polymer-polyelectrolyte is not as miscible as observed in other The ternary PEO-LiTFSI-PLSS showed the lithium transference number, ∼0.66 at 70°C when tested with Li/SPE/Li cell using DC polarization and AC impedance methods. (6) The ternary system showed a wide enough electrochemical stability window, ca. 4.7±0.1 V in the Li/SPE/SS device when analyzed by the linear sweep voltammetry method. Finally, (7) a coin type half-cell with the Li/SPE/LiFePO 4 configuration displayed the specific capacity of 121 mAh/g at a 0.1 C-rate, demonstrating that this free-standing electrolyte is appropriate for a practical Li-ion battery. Therefore, the ternary blends incorporating a polyelectrolyte, just like versatile polymer-polymer blend systems in commercial markets, appears promising as a material of choice, applicable for energy storage devices such as batteries and supercapacitors.