Enhanced Performance of WO3/SnO2 Nanocomposite Electrodes with Redox-Active Electrolytes for Supercapacitors

For effective supercapacitors, we developed a process involving chemical bath deposition, followed by electrochemical deposition and calcination, to produce WO3/SnO2 nanocomposite electrodes. In aqueous solutions, the hexagonal WO3 microspheres were first chemically deposited on a carbon cloth, and then tin oxides were uniformly electrodeposited. The synthesized WO3/SnO2 nanocomposite was characterized by XRD, XPS, SEM, and EDX techniques. Electrochemical properties of the WO3/SnO2 nanocomposite were analyzed by cyclic voltammetry, galvanostatic charge-discharge tests, and electrochemical impedance spectroscopy in an aqueous solution of Na2SO4 with/without the redox-active electrolyte K3Fe(CN)6. K3Fe(CN)6 exhibited a synergetic effect on the electrochemical performance of the WO3/SnO2 nanocomposite electrode, with a specific capacitance of 640 F/g at a scan rate of 5 mV/s, while that without K3Fe(CN)6 was 530 F/g. The WO3/SnO2 nanocomposite catalyzed the redox reactions of [Fe(CN)6]3/[Fe(CN)6]4− ions, and the [Fe(CN)6]3−/[Fe(CN)6]4− ions also promoted redox reactions of the WO3/SnO2 nanocomposite. A symmetrical configuration of the nanocomposite electrodes provided good cycling stability (coulombic efficiency of 99.6% over 2000 cycles) and satisfied both energy density (60 Whkg−1) and power density (540 Wkg−1) requirements. Thus, the WO3/SnO2 nanocomposite prepared by this simple process is a promising component for a hybrid pseudocapacitor system with a redox-flow battery mechanism.


Introduction
Clean and renewable energy is the basic criterion for sustainable development of the global economy and human lives throughout the world [1]. The massive increase in fossil fuel usage has brought two significant problems: the first is rapidly reduced fossil fuel reserves, and the second is environmental degradation, such as air/water pollution and climate change. Considering these issues, the need for renewable energy and new technology for energy storage are currently the top concerns around the world. In response to rising ecological concern and modern civilization, new environmentally friendly and cost-effective energy storage devices with excellent performance are required for customer products, such as hybrid automobiles, electronic vehicles, mobile phones, and laptop computers [2,3]. Supercapacitors, batteries, and fuel cells are the most common electrochemical energy storage devices used for this purpose. Among them, supercapacitors are emphasized because of their excellent power densities, fast charge/discharge, long-term cycling life, cost-effectiveness, good reversibility, and wide working temperature range compared to those of batteries; however, in terms of commercial application, supercapacitors are behind batteries [4][5][6]. The fundamental issue of supercapacitors is their lower energy density, especially when compared to batteries.
To solve this problem of supercapacitors, three different energy storage mechanisms have been developed: electric double-layer capacitors (EDLCs), pseudocapacitor, and Crystals of tungsten oxide (WO 3 ) are polymorphic and often adopt a hexagonal system [1,15]. Various synthetic aspects, such as the reaction temperature, process duration, precursor type, and chemical agent (e.g., (NH 4 ) 2 SO 4 ), can influence the structure and morphology. In comparison to other structures, this structure provides large tunnels for ion penetration, which improves the electrochemical performance [13]. Thus, WO 3 is usually synthesized as a hexagonal system using reagents containing sulfate ions; in particular, ammonium sulfate is preferred. The NH 4 + and SO 4 2− ions act as stabilizing and capping agents, respectively, during the process. The radius of the NH 4 ion is larger than those of alkali metal ions, and these differences in size influence the penetration of numerous ions, which implies that many SO 4 2− ions can readily be adsorbed on the surface parallel to the WO 3 axis [13,30]. The morphology and crystal structure of WO 3 are also affected by the annealing temperature. The orthorhombic phase changes into an anhydrous hexagonal phase at 400 • C and turns into stable monoclinic WO 3 with aggregation above 500 • C [31]. Therefore, annealing in this work was done at 400 • C for 2 h.

Characterization of WO 3 , SnO 2 , and WO 3 /SnO 2 Nanocomposite Thin Films
The structural characteristics and crystal phases of the WO 3 , SnO 2 , and WO 3 /SnO 2 nanocomposites were analyzed by XRD ( Figure 1). The XRD peaks for WO 3 were assigned to the hexagonal system, except for the one marked with an asterisk; this came from the carbon cloth substrate (Figure 1a H2WO4·nH2O → WO3 + (n + 1)H2O (2) Crystals of tungsten oxide (WO3) are polymorphic and often adopt a hexagonal system [1,15]. Various synthetic aspects, such as the reaction temperature, process duration, precursor type, and chemical agent (e.g., (NH4)2SO4), can influence the structure and morphology. In comparison to other structures, this structure provides large tunnels for ion penetration, which improves the electrochemical performance [13]. Thus, WO3 is usually synthesized as a hexagonal system using reagents containing sulfate ions; in particular, ammonium sulfate is preferred. The NH4 + and SO4 2-ions act as stabilizing and capping agents, respectively, during the process. The radius of the NH4 ion is larger than those of alkali metal ions, and these differences in size influence the penetration of numerous ions, which implies that many SO4 2− ions can readily be adsorbed on the surface parallel to the WO3 axis [13,30]. The morphology and crystal structure of WO3 are also affected by the annealing temperature. The orthorhombic phase changes into an anhydrous hexagonal phase at 400 °C and turns into stable monoclinic WO3 with aggregation above 500 °C [31]. Therefore, annealing in this work was done at 400 °C for 2 h.

Characterization of WO3, SnO2, and WO3/SnO2 Nanocomposite Thin Films
The structural characteristics and crystal phases of the WO3, SnO2, and WO3/SnO2 nanocomposites were analyzed by XRD ( Figure 1). The XRD peaks for WO3 were assigned to the hexagonal system, except for the one marked with an asterisk; this came from the carbon cloth substrate (Figure 1a   The XRD pattern for the synthesized SnO 2 showed sharp diffraction peaks at 30.7 • , 32.1 • , 43.9 • , 45.1 • , and 55.5 • , which corresponded to the (221), (101), (111), (210), and (220) planes of the tetragonal phase (JCPDS 41-1445), respectively, except for the one marked with an asterisk, which came from the carbon cloth substrate (Figure 1b) [32]. Figure 1c shows the diffraction peaks for the WO 3 (221), (101), (111), (210), and (220) planes of SnO 2 , respectively. The diffraction peaks, denoted with diamond symbols, were indexed to the tetragonal structure of SnO 2 [33]; similarly, the other peaks were associated with the hexagonal crystal phase of WO 3 [1]. Therefore, this diffraction pattern was composed of characteristic peaks for both WO 3 and SnO 2 , as observed in (Figure 1a,b), which confirmed successful preparation of the WO 3 /SnO 2 nanocomposite. No new significant peak was observed, and thus, the WO 3 and SnO 2 components were separately crystallized in the nanocomposite. The presence of sharp diffraction peaks suggested the highly crystalline nature of the materials deposited on the carbon cloth substrate. The average sizes of the WO 3 and SnO 2 crystallites in the WO 3 /SnO 2 nanocomposite were determined by the Scherrer equation, as shown below in (Equation (3)) [6,[34][35][36]: where D is the crystallite size (nm), β is the full width half maximum (FWHM, radians) of the peak corresponding to the plane, K = 0.9 (Sherrer constant), λ = 0.15406 nm (wavelength of the X-ray source), and θ is obtained from the 2θ value corresponding to the diffraction peak [35]. These findings showed that the average crystallite sizes for the WO 3 synthesized in the nanocomposite were 7.5 nm, with diffraction peaks of 2θ = 23.28 • and 28.2 • for the corresponding (110) and (200) planes, respectively. The average crystallite size of the SnO 2 synthesized in the nanocomposite was estimated to be 3.9 nm, with diffraction peaks at 2θ = 32.10 • , with a corresponding (101) plane [37,38]. The chemical compositions and surfaces of the WO 3 , SnO 2 , and WO 3 /SnO 2 nanocomposites were further investigated using XPS ( Figure 2 and Table 1). The survey spectrum for the WO 3 /SnO 2 nanocomposite showed the coexistence of W, Sn, O, and C elements, and no other impurity peak was significantly detected (Figure 2a).
The high-resolution W 4f spectrum for WO 3 was deconvoluted into two peaks ( Figure 2b). The peaks at 36.08 eV and 38.21 eV were assigned to W 4f 7/2 and W 4f 5/2 binding energies, respectively, corresponding to the W 6+ oxidation state. The separation between the W 4f 7/2 and W 4f 5/2 peaks was 2.13 eV, which was in good agreement with an earlier report [1]. Figure 2c shows the O 1s peak at 531.1 eV, which was assigned to the lattice oxygens of W-O bonds in WO 3 . The peak at 532.01 eV was due to an oxygen deficiency, the peak at 532.7 eV was assigned to chemisorbed oxygen, and the peak at 533.6 eV was attributed to water and hydroxide. Figure 3d shows the high-resolution Sn 3d spectrum of pure SnO 2 , which was deconvoluted into two peaks with binding energies of 487.6 eV and 496.05 eV. These peaks were assigned to the Sn 3d 5/2 and Sn 3d 3/2 binding energies, respectively, for the oxidation state Sn 4+ . The splitting between the Sn 3d 5/2 and Sn 3d 3/2 peaks was 8.45 eV, which was in good agreement with a previous study [4]. As displayed in (Figure 2e), the O 1s spectrum was deconvoluted into three peaks at 531.5 eV for lattice SnO 2 , chemisorbed oxygen at 532.7 eV, and water and hydroxide at 534.3 eV. In the WO 3 /SnO 2 nanocomposite, the W 4f spectrum showed two peaks at 35.74 eV and 38.08 eV, which corresponded to W 4f 7/2 and W 4f 5/2 binding energies, respectively, and an oxidation state of W +6 (Figure 2f). The spin-orbit separation of W 4f 7/2 and W 4f 5/2 was 2.24 eV, which agreed with previous reports [39]. The difference in the W 4f doublet separation for the WO 3 and WO 3 /SnO 2 nanocomposites (2.13 eV and 2.24 eV, respectively) was also reported in an earlier study [2]: the W 4f 7/2 peak was shifted by −0.24 eV, and the W 4f 5/2 peak was shifted by −0.13 eV, which suggested an interaction between WO 3 and SnO 2 (Table 1). Moreover, the Sn 3d spectrum of the WO 3 /SnO 2 composite had two peaks at 487.69 eV and 496.12 eV, which corresponded to the characteristic Sn 3d 5/2 and Sn 3d 3/2 binding energies, respectively. The spin-orbit separation between the Sn 3d 5/2 and Sn 3d 3/2 peaks was 8.43 eV (Figure 2g), which agreed with earlier reports [40]. Comparing pure SnO 2 and WO 3 /SnO 2 , the Sn 3d 5/2 and Sn 3d 3/2 peaks were slightly shifted by 0.02 eV, which could have been due to an interaction between WO 3 and SnO 2 . The O 1s spectrum was deconvoluted into four peaks (Figure 2h) at 531.09 eV, 531.76 eV, 532.8 eV, and 534.3 eV. The main peak at 531.76 eV was attributed to lattice oxygens in the metal oxides WO 3 and SnO 2 [33,37,40], the peak at 531.09 eV was assigned to oxygen defects or vacancies, the peak at 532.8 eV was related to chemisorbed oxygen species [41], and the peak at 534.3 eV was attributed to water and hydroxides on the surface of the WO 3 /SnO 2 nanocomposite, as reported previously [6,32,40]. The binding energies of the lattice oxygens and oxygen vacancies in the WO 3 /SnO 2 nanocomposite were shifted from those of the single components by −0.34 eV and 0.25 eV, respectively, which suggested an interaction between WO 3 and SnO 2 in the nanocomposite. The binding energy for absorbed water/hydroxides in the nanocomposite was similar to that of SnO 2 , which suggested that the surface of the nanocomposite was covered by SnO 2 .
[33]; similarly, the other peaks were associated with the hexagonal crystal phase of WO3 [1]. Therefore, this diffraction pattern was composed of characteristic peaks for both WO3 and SnO2, as observed in (Figure 1a,b), which confirmed successful preparation of the WO3/SnO2 nanocomposite. No new significant peak was observed, and thus, the WO3 and SnO2 components were separately crystallized in the nanocomposite. The presence of sharp diffraction peaks suggested the highly crystalline nature of the materials deposited on the carbon cloth substrate. The average sizes of the WO3 and SnO2 crystallites in the WO3/SnO2 nanocomposite were determined by the Scherrer equation, as shown below in (Equation (3)) [6,[34][35][36]: where D is the crystallite size (nm), β is the full width half maximum (FWHM, radians) of the peak corresponding to the plane, K = 0.9 (Sherrer constant), λ = 0.15406 nm (wavelength of the X-ray source), and θ is obtained from the 2θ value corresponding to the diffraction peak [35]. These findings showed that the average crystallite sizes for the WO3 synthesized in the nanocomposite were 7.5 nm, with diffraction peaks of 2θ = 23.28° and 28.2° for the corresponding (110) and (200) planes, respectively. The average crystallite size of the SnO2 synthesized in the nanocomposite was estimated to be 3.9 nm, with diffraction peaks at 2θ = 32.10°, with a corresponding (101) plane [37,38]. The chemical compositions and surfaces of the WO3, SnO2, and WO3/SnO2 nanocomposites were further investigated using XPS ( Figure 2 and Table 1). The survey spectrum for the WO3/SnO2 nanocomposite showed the coexistence of W, Sn, O, and C elements, and no other impurity peak was significantly detected (Figure 2a).   The elemental compositions of the synthesized materials were measured with ED ( Figure 4). In the pure WO3 microspheres, the atomic ratio of W to O was 1:3.09, whic agreed with the theoretical ratio of 1:3. In the SnO2 nanorods, the atomic ratio of Sn to O was 1:1.58, which was lower than the theoretical ratio of 1:2, suggesting that unreacte Sn 2+ provided defects or oxygen vacancies in the crystal lattice. The WO3/SnO2 nanocom posite consisted of W, Sn, and O, with atomic percentages of W (18.45%), Sn (6.31%), an O (75.24%). The Sn content was significantly lower than the W content, which was sup ported by the SEM observation (Figure 3e  The morphologies of the materials were observed by SEM. WO 3 was composed of aggregated monodisperse microspheres (diameter:~1 µm) with rough surfaces (Figure 3a,b). The spaces between these microspheres would allow electrolyte ions to diffuse into the porous nanostructure [15,42]. The roughness of the microsphere surface increased the number of active sites available for electrochemical reactions. Moreover, the SnO 2 exhibited nanorod structures with~60 nm widths and~200 nm lengths (Figure 3c,d). The applied potential and the deposition time are crucial for successful formation of SnO 2 nanorods [43,44]. In this work, SnO 2 nanorods were synthesized from SnCl 2 on a cathode (−0.9 V vs. Ag/AgCl), and the formation of SnO 2 crystals can be attributed to electrochemical deposition of Sn(OH) 2 , followed by air oxidation during calcination, as shown below (Equations (4)- (7)) [45]: The SEM images in (Figure 3e,f) indicate the surface morphology of the WO 3 /SnO 2 nanocomposite. Thin and wrinkled films were uniformly formed on the substrate. These films seemed to cover microspheres, which could be the WO 3 first deposited on the carbon cloth substrate. This morphology suggested that SnO 2 could not form coarse nanorods on the rough surfaces of WO 3 , so smaller SnO 2 crystals formed a network on the surface [46]. In the cross-sectional views (Figure 3g,h), the surface of WO 3 was covered by nanorods of SnO 2 [47]. These small crystals and the fine nanorods provide a large surface area, which could enhance the electrochemical performance of this nanocomposite [32].
The elemental compositions of the synthesized materials were measured with EDS ( Figure 4). In the pure WO 3 microspheres, the atomic ratio of W to O was 1:3.09, which agreed with the theoretical ratio of 1:3. In the SnO 2 nanorods, the atomic ratio of Sn to O was 1:1.58, which was lower than the theoretical ratio of 1:2, suggesting that unreacted Sn 2+ provided defects or oxygen vacancies in the crystal lattice. The WO 3 /SnO 2 nanocomposite consisted of W, Sn, and O, with atomic percentages of W (18.45%), Sn (6.31%), and O (75.24%). The Sn content was significantly lower than the W content, which was supported by the SEM observation (Figure 3e

Electrochemical Measurements without a Redox-Active Electrolyte
The electrochemical behaviors of the synthesized materials were analyzed with CV and GCD measurements using the non-redox electrolyte Na2SO4 ( Figures 5 and 6).

Electrochemical Measurements without a Redox-Active Electrolyte
The electrochemical behaviors of the synthesized materials were analyzed with CV and GCD measurements using the non-redox electrolyte Na 2 SO 4 ( Figures 5 and 6). and (g) Sn.

Electrochemical Measurements without a Redox-Active Electrolyte
The electrochemical behaviors of the synthesized materials were analyzed with CV and GCD measurements using the non-redox electrolyte Na2SO4 (Figures 5 and 6).  Figure 5a shows the CV curves for WO3 in the potential window −0.2 to 0.7 V vs. Ag/AgCl with various scan rates of 5−50 mV/s. These CV curves demonstrated nonrectangular shapes and medium oxidation peaks at approximately 0.1 V, which indicated a typical combination of pseudocapacitive behavior and electrical double-layer capacitance [48,49]. The oxidation peak clearly increased as the scan rate was increased, indicating the high reactivity of the WO3 electrode [50,51]. The specific capacitance was determined from the CV curves by using (Equation (8)  where Ccv is the gravimetric specific capacitance (F/g) calculated from the measurement, i is the voltammetric current, m is the mass of the active material (mg), ν is the scan rate (mV/s), Vf and Vi are the bounds of the potential window. The Ccv of the WO3 electrode was 240, 180, 150, 120, 98, and 84 F/g at scan rates of 5, 10, 20, 30, 40, and 50 mV −1 , respectively.
Under the same conditions, the CV curves for SnO2 displayed quasi-rectangular shapes, with redox peaks at 0.54−0.69 V (oxidative) and 0.12−0.37 V (reductive), which indicated the more pseudocapacitive nature of the SnO2 electrode ( Figure 5b). As the scan rate was increased, the CV curves for SnO2 exhibited more rectangular shapes, and the difference between the redox peaks increased, which suggested relatively slow redox reactions occurring on the surfaces of the SnO2 nanorods. The specific capacitance values for the SnO2 electrodes were calculated from the CV curves and were 140, 120, 90, 60, 50, and 45 F/g at scan rates of 5, 10, 20, 30, 40, and 50 mV/s, respectively, as shown in (Figure 5e) The CV curves for the WO3/SnO2 nanocomposite showed no significant redox peak (Figure 5c), which suggested the complementary work of the two components and a synergistic effect arising from interactions between WO3 and SnO2 (see Figure 2 and Table 1) This WO3/SnO2 nanocomposite electrode generated greater areas under the CV curves which indicated a higher specific capacitance than the single-component electrodes (WO3 and SnO2) (Figure 5e). The specific capacitances of the WO3/SnO2 nanocomposite were calculated as 530, 410, 280, 210, 160, and 150 Fg −1 at scan rates of 5, 10, 20, 30, 40, and 50 mV s −1 , respectively. Comparing the CV curves and the specific capacitances of the three electrodes, the WO3/SnO2 nanocomposite electrode had a higher specific capacitance, especially at low scan rates (Figure 5d,e). This enhanced performance for the nanocomposite electrode could be explained by activation of the SnO2 via the synergistic effect and the smaller crystals of SnO2 formed on WO3 (Figure 4). The small SnO2 crystals could have enhanced the surface area of the electrode and shortened the electron pathway from the surface of SnO2 to WO3 [55].   Figure 5a shows the CV curves for WO 3 in the potential window −0.2 to 0.7 V vs. Ag/AgCl with various scan rates of 5-50 mV/s. These CV curves demonstrated nonrectan-gular shapes and medium oxidation peaks at approximately 0.1 V, which indicated a typical combination of pseudocapacitive behavior and electrical double-layer capacitance [48,49]. The oxidation peak clearly increased as the scan rate was increased, indicating the high reactivity of the WO 3 electrode [50,51]. The specific capacitance was determined from the CV curves by using (Equation (8)) [9,[52][53][54]: where C cv is the gravimetric specific capacitance (F/g) calculated from the measurement, i is the voltammetric current, m is the mass of the active material (mg), ν is the scan rate (mV/s), V f and V i are the bounds of the potential window. The C cv of the WO 3 electrode was 240, 180, 150, 120, 98, and 84 F/g at scan rates of 5, 10, 20, 30, 40, and 50 mV −1 , respectively. Under the same conditions, the CV curves for SnO 2 displayed quasi-rectangular shapes, with redox peaks at 0.54-0.69 V (oxidative) and 0.12-0.37 V (reductive), which indicated the more pseudocapacitive nature of the SnO 2 electrode (Figure 5b). As the scan rate was increased, the CV curves for SnO 2 exhibited more rectangular shapes, and the difference between the redox peaks increased, which suggested relatively slow redox reactions occurring on the surfaces of the SnO 2 nanorods. The specific capacitance values for the SnO 2 electrodes were calculated from the CV curves and were 140, 120, 90, 60, 50, and 45 F/g at scan rates of 5, 10, 20, 30, 40, and 50 mV/s, respectively, as shown in (Figure 5e). The CV curves for the WO 3 /SnO 2 nanocomposite showed no significant redox peak (Figure 5c), which suggested the complementary work of the two components and a synergistic effect arising from interactions between WO 3 and SnO 2 (see Figure 2 and Table 1). This WO 3 /SnO 2 nanocomposite electrode generated greater areas under the CV curves, which indicated a higher specific capacitance than the single-component electrodes (WO 3 and SnO 2 ) (Figure 5e). The specific capacitances of the WO 3 /SnO 2 nanocomposite were calculated as 530, 410, 280, 210, 160, and 150 Fg −1 at scan rates of 5, 10, 20, 30, 40, and 50 mV s −1 , respectively. Comparing the CV curves and the specific capacitances of the three electrodes, the WO 3 /SnO 2 nanocomposite electrode had a higher specific capacitance, especially at low scan rates (Figure 5d,e). This enhanced performance for the nanocomposite electrode could be explained by activation of the SnO 2 via the synergistic effect and the smaller crystals of SnO 2 formed on WO 3 (Figure 4). The small SnO 2 crystals could have enhanced the surface area of the electrode and shortened the electron pathway from the surface of SnO 2 to WO 3 [55]. Figure 6a illustrates the GCD curves of the WO 3 electrode in the potential window-0.2 to 0.7 V vs. Ag/AgCl at various current densities. The nonlinear nature of the charge/discharge processes confirmed the pseudocapacitive nature and the high activity at low potential (<0.1 V). The specific capacitance was calculated from the GCD curves and Equation (9) [9,53,54]: where C G is the gravimetric capacitance (F/g) from the GCD measurement, I g is the given current, dt is the discharging time (s), and ∆V(t) is the potential window as a function of t. The C G of WO 3 electrode was calculated to be 120 Fg −1 at a current density of 1.25 Ag −1 . The SnO 2 electrode exhibited two steps in the charge process and one step in the discharge process (Figure 6b). At 1.25 Ag −1 , the specific capacitance of the SnO 2 electrode was 54 Fg −1 . Figure 6c shows the GCD curves for the WO 3 /SnO 2 nanocomposite electrode at different current densities. The shapes of the GCD curves were more triangular, which indicated that the nanocomposite electrode exhibited better pseudocapacitive nature than the others, as was also suggested by the CV measurements ( Figure 6). The specific capacitance of the WO 3 /SnO 2 nanocomposite was determined to be 180, 120, 80, and 68 Fg −1 at the various current densities 1.25, 1.88, 2.82, and 5.00 Ag −1 , respectively. Figure 6d provides a comparison of the GCD curves for the SnO 2 , WO 3 , and WO 3 /SnO 2 electrodes at the same current density of 1.25 Ag −1 . Figure 6e shows the EIS results for the WO 3 , SnO 2 , and WO 3 /SnO 2 nanocomposite electrodes, and the magnified EIS data for the electrode are shown in (Figure 6f). The charge transfer resistance values (Rct) calculated for the WO 3 , SnO 2 , and WO 3 /SnO 2 nanocomposite electrodes were 2.12 Ω, 1.94 Ω, and 1.72 Ω, respectively. The lower Rct of the WO 3 /SnO 2 nanocomposite electrode indicated rapid charge transfer in the nanocomposite. The resistance of the electrode-electrolyte interference (Rs), which corresponds to the x-axis intercept in the high-frequency region, was calculated to be 2.37 Ω, 2.32 Ω, and 1.87 Ω for the WO 3 , SnO 2 , and WO 3 /SnO 2 nanocomposite electrodes, respectively. The low Rs suggested that the WO 3 /SnO 2 nanocomposite contacted the electrolytes with large surface areas because of its small grain sizes (Figure 4).

Effects of Redox-Active Electrolyte
Electrochemical analyses of the WO 3 , SnO 2 , and WO 3 /SnO 2 electrodes were performed in the presence of the redox additive 0.2 M K 3 Fe (CN) 6 in the 1 M Na 2 SO 4 aqueous electrolyte. First, CV measurements were conducted with the carbon cloth, WO 3 , SnO 2 , and WO 3 /SnO 2 nanocomposite electrodes (Figure 7a). current densities 1.25, 1.88, 2.82, and 5.00 Ag , respectively. Figure 6d provide ison of the GCD curves for the SnO2, WO3, and WO3/SnO2 electrodes at the s density of 1.25 Ag −1 . Figure 6e shows the EIS results for the WO3, SnO2, and nanocomposite electrodes, and the magnified EIS data for the electrode are sh ure 6f). The charge transfer resistance values (Rct) calculated for the WO3 WO3/SnO2 nanocomposite electrodes were 2.12 Ω, 1.94 Ω, and 1.72 Ω, respe lower Rct of the WO3/SnO2 nanocomposite electrode indicated rapid charge tra nanocomposite. The resistance of the electrode-electrolyte interference (Rs), w sponds to the x-axis intercept in the high-frequency region, was calculated t 2.32 Ω, and 1.87 Ω for the WO3, SnO2, and WO3/SnO2 nanocomposite electro tively. The low Rs suggested that the WO3/SnO2 nanocomposite contacted the with large surface areas because of its small grain sizes (Figure 4).

Effects of Redox-Active Electrolyte
Electrochemical analyses of the WO3, SnO2, and WO3/SnO2 electrode formed in the presence of the redox additive 0.2 M K3Fe (CN)6 in the 1 M Na2S electrolyte. First, CV measurements were conducted with the carbon cloth, and WO3/SnO2 nanocomposite electrodes (Figure 7a).  6 at scan rate of 30 mV/s; CV curves of (c) WO 3 electrode, (d) SnO 2 electrode, and (e) WO 3 /SnO 2 nanocomposite electrode with K 3 Fe(CN) 6 at various scan rates; (f) specific capacitance of various electrodes with/without K 3 Fe(CN) 6 at various scan rates; (g) double logarithmic plots of scan rate vs. peak current for WO 3 , SnO 2 , and the WO 3 /SnO 2 nanocomposite; (h) cycling stability and coulombic efficiency versus cycle number for the WO 3 /SnO 2 nanocomposite electrode with K 3 Fe(CN) 6 ; (i) EIS analyses of WO 3 , SnO 2 , and WO 3 /SnO 2 nanocomposite electrodes with K 3 Fe(CN) 6 over the frequency range 100 kHz to 100 mHz.
While the carbon cloth was almost inactive, the CV curves for the metal oxide electrodes exhibited quasi-rectangular shapes, with a pair of redox peaks from [Fe (CN) 6 ] 3− /[Fe (CN) 6 ] 4− at 0.35-0.43 V and 0.18-0.26 V. The WO 3 /SnO 2 nanocomposite electrode had the largest area at a scan rate of 50 mV/s, which indicated the highest specific capacitance. The WO 3 /SnO 2 nanocomposite electrode also resulted in higher intensities of redox peaks than the other electrodes, which indicated that this electrode exhibited the highest activity for the redox reactions of [Fe(CN) 6 ] 3− /[Fe(CN) 6 ] 4− . Moreover, considering the high peak intensities, the difference between the oxidation and reduction potentials was smaller than those of the other electrodes. The reduction peak for [Fe(CN) 6 ] 3− with the WO 3 /SnO 2 nanocomposite electrode was at the highest potential among the electrodes, while the oxidation potential for [Fe(CN) 6 ] 4− with the WO 3 /SnO 2 nanocomposite electrode was higher than the others. These behaviors suggested that the WO 3 /SnO 2 nanocomposite catalyzed the redox reactions of [Fe(CN) 6 ] 3− /[Fe(CN) 6 ] 4− , which is preferable for energy storage because of the low overpotential. The catalytic activity of the WO 3 /SnO 2 nanocomposite could be due to interactions between WO 3 and SnO 2 , as suggested by the XPS measurements ( Figure 2 and Table 1). Therefore, the CV curves generated with two sets of conditions were compared, i.e., with K 3 Fe (CN) 6 and without K 3 Fe (CN) 6 (Figure 7b). With K 3 Fe (CN) 6 , the areas of the CV curves were larger, which indicated that the redox additive increased the charge storage capacity. Using the CV curves generated at different scan rates (Figure 7c-e), the specific capacitance of the electrode in the redox-active electrolyte was calculated. At scan rates of 5, 10, 20, 30, 40, and 50 mV/s, the respective specific capacitance values were 440, 360, 280, 240, 190, and 150 F/g for the WO 3 electrode; 310, 220, 150, 130, 120, and 110 F/g for the SnO 2 electrode; and 640, 520, 340, 330, 310, and 290 F/g for the WO 3 /SnO 2 nanocomposite electrode. Moreover, the addition of K 3 Fe(CN) 6 resulted in more pseudocapacitive (more rectangular CV curves) for the WO 3 and WO 3 /SnO 2 nanocomposite electrodes. In details, weak and broad reduction bands newly appeared at −0.1-0.0 V for the WO 3 electrode and 0.0-0.1 V for the WO 3 /SnO 2 electrode in the presence of K 3 Fe(CN) 6 (Figure 7b,c,e). The pseudocapacitive behavior suggested that K 3 Fe(CN) 6 provided the additional redox reactions of the [Fe(CN) 6 ] 3− /[Fe(CN) 6 ] 4− couple to increase the specific capacitance of these electrodes [56,57], and also promoted the redox reactions of the WO 3 component. That is, these metal oxides (WO 3  The kinetics of the energy storage process were analyzed from the perspectives of peak current and scan rate (Figure 7g). In general, the relationship between the peak current and the scan rate can be described as in (Equation (10)) [58,59]: where I p is the peak oxidation current (mA), ν is the scan rate (mV/s), and both a and b are adjustable parameters. The b value is obtained from the slope of a plot of log I p versus log ν. When the charge storage process is a surface-controlled capacitive process (ion adsorption/desorption), the value of b is 1. On the other hand, if diffusion is the rate-controlling process, b is 0.5 [60]. With and without the K 3 Fe(CN) 6 electrolyte, the b values of the WO 3 /SnO 2 nanocomposite electrode were 0.55 and 0.71, respectively. These b values suggested that addition of the K 3 Fe(CN) 6 electrolyte favored diffusion control of the energy storage process because the approach of the redox-active species enabled charge transfer at the electrode surface. Moreover, the b values for the WO 3 and SnO 2 electrodes were 0.62 and 0.75, respectively. These b values were higher than that of the WO 3 /SnO 2 nanocomposite electrode, which indicated that these single-component electrodes were less affected by diffusion of the redox-active species. The GCD measurements made in the presence of the K 3 Fe(CN) 6 electrolyte were not adequate to calculate the specific capacitance values of the synthesized electrodes because the discharge process took longer than the charge time (see Figure S1 in Supporting Information). This extension of the discharge process could be due to reduction of [Fe(CN) 6 ] 3− to [Fe(CN) 6 ] 4− , which could be used for the redox-flow battery system. Therefore, GCD measurements were conducted for stability testing (Figure 7h). The coulombic efficiency of the electrode was calculated with Equation (11): where η is the coulombic efficiency, t d is the discharging time, and t c is the charging time (s) [9]. After 2000 GCD cycles at a current density of 2.5 A/g in the presence of K 3 Fe(CN) 6   The WO 3 /SnO 2 nanocomposite electrode exhibited a lower Rct than the other electrodes in the electrolyte systems with/without K 3 Fe(CN) 6 . For all electrodes used in this study, K 3 Fe(CN) 6 drastically decreased R ct , which indicated that K 3 Fe(CN) 6 facilitated charge transfer between the active materials. That is, K 3 Fe(CN) 6 activated the electrical behaviors of the metal oxides used in this study. The R s values also showed improvement, except for that of the WO 3 electrode, which was attributed to an additional redox reaction occurring between the electrode surface and the electrolyte, in addition to the increased ion concentration of the electrolyte solution.

Electrochemical Performance of Symmetric Systems
For practical application of the WO 3 /SnO 2 nanocomposite electrode, a symmetric supercapacitor system with identical electrodes was constructed with an aqueous electrolyte containing 1 M NaSO 4 and 0.02 M K 3 Fe(CN) 6 . The total mass of the WO 3 /SnO 2 nanocomposite on both electrodes was~3.4 mg (1.7 mg for each). To confirm the optimal voltage window of the device, CV curves were measured over different potential ranges (from 1.0 to 1.8 V) at a scan rate of 50 mVs −1 , as shown in (Figure 8a). The CV profile showed that the working potential was extended to 1.8 V, and water hydrolysis occurred at 1.9 V. As the potential range was increased, the shapes of the CV curves changed from quasi-rectangular (up to 1.4 V) to rectangular with redox peaks (higher than 1.6 V). This change indicated that a potential range larger than 1.6 V converted the pseudocapacitive mechanism to an energy storage mechanism, including the redox reactions of [Fe(CN) 6  the working potential was extended to 1.8 V, and water hydrolysis occurred at 1.9 V. As the potential range was increased, the shapes of the CV curves changed from quasi-rectangular (up to 1.4 V) to rectangular with redox peaks (higher than 1.6 V). This change indicated that a potential range larger than 1.6 V converted the pseudocapacitive mechanism to an energy storage mechanism, including the redox reactions of  The CV curves of the symmetric WO3/SnO2 nanocomposite electrode systems with and without K3Fe(CN)6 showed nearly rectangular shapes, which indicated pseudocapacitive behavior, even at a high scan rate of 50 mV/s (Figure 8b,c). Comparing the two systems, the CV profile of the WO3/SnO2 nanocomposite electrode with K3Fe(CN)6 had a greater area than that without K3Fe(CN)6, as expected from the half-cell experiment above (Figure 7b). GCD measurements of the symmetric WO3/SnO2 nanocomposite electrodes with and without K3Fe (CN)6 were also performed with various current densities ( Figure  8e,f), and the GCD curves generated at 1.25 A/g were compared (Figure 8g). The IR drop in the initial phase of discharge was decreased in the presence of K3Fe(CN)6 (dropped from 1.8 V to ~1.0 V without K3Fe(CN)6 and from 1.8 V to ~1.3 V with K3Fe(CN)6). The specific capacitance values were determined at a current density of 1.25 A/g and were 285 F/g with K3Fe(CN)6 and 186 F/g without K3Fe(CN)6. The coulombic efficiencies were low in these cases, which could be explained by a minor water hydrolysis and the diffusion of [Fe(CN)6] 3− /[Fe (CN)6] 4− ions. At higher potential than 1.6 V, the edge of water electrolysis started (Figure 8a), and the minor gas generation resulted in the energy loss. Using K3Fe(CN)6, the reaction products ([Fe(CN)6] 3− and [Fe(CN)6] 4− , mutually) diffused into the whole electrolyte solution in the cell, which decreased the availability of these ions for the discharge process [28]. The CV curves of the symmetric WO 3 /SnO 2 nanocomposite electrode systems with and without K 3 Fe(CN) 6 showed nearly rectangular shapes, which indicated pseudocapacitive behavior, even at a high scan rate of 50 mV/s (Figure 8b,c). Comparing the two systems, the CV profile of the WO 3 /SnO 2 nanocomposite electrode with K 3 Fe(CN) 6 had a greater area than that without K 3 Fe(CN) 6 , as expected from the half-cell experiment above (Figure 7b). GCD measurements of the symmetric WO 3 /SnO 2 nanocomposite electrodes with and without K 3 Fe (CN) 6 were also performed with various current densities (Figure 8e,f), and the GCD curves generated at 1.25 A/g were compared (Figure 8g). The IR drop in the initial phase of discharge was decreased in the presence of K 3 Fe(CN) 6 (dropped from 1.8 V to~1.0 V without K 3 Fe(CN) 6 and from 1.8 V to~1.3 V with K 3 Fe(CN) 6 ). The specific capacitance values were determined at a current density of 1.25 A/g and were 285 F/g with K 3 Fe(CN) 6 and 186 F/g without K 3 Fe(CN) 6 . The coulombic efficiencies were low in these cases, which could be explained by a minor water hydrolysis and the diffusion of [Fe(CN) 6 ] 3− /[Fe (CN) 6 ] 4− ions. At higher potential than 1.6 V, the edge of water electrolysis started (Figure 8a), and the minor gas generation resulted in the energy loss. Using K 3 Fe(CN) 6 , the reaction products ([Fe(CN) 6 ] 3− and [Fe(CN) 6 ] 4− , mutually) diffused into the whole electrolyte solution in the cell, which decreased the availability of these ions for the discharge process [28].
The specific energy density (SE) and the specific power density (SP) of the WO 3 /SnO 2 symmetric supercapacitor system with/without redox activity were important parameters for determining the feasibility of the device application. S E and S P were calculated using Equations (12) and (13) [53,54]: where M is the total mass of both the positive and negative electrodes, I is the current, and V is the potential as a function of discharge time (t d ). The Ragone diagram for the energy density vs. power density of the WO 3 /SnO 2 nanocomposite electrodes with/without additive is shown in (Figure 8h). The WO 3 /SnO 2 symmetric supercapacitor system with the K 3 Fe(CN) 6 electrolyte exhibited the highest S E of 64 Whkg −1 at a S P of 542 Wkg −1 , whereas without K 3 Fe(CN) 6 , the S E was 35 Whkg −1 for a S P of 468 Wkg −1 (in both cases, the current density was 1.25 A/g). Compared with previous reports, the symmetric device with the WO 3 /SnO 2 nanocomposite electrodes exhibited excellent electrochemical performance in the Na 2 SO 4 /K 3 Fe(CN) 6 electrolyte (Table 3). In the symmetric system, the active materials simultaneously worked as positive (anode) and negative (cathode) electrodes, and the GCD curves confirmed the reversible balanced redox reactions occurring at the anode and cathode [61,62]. Although the WO 3 /SnO 2 ratio and the concentration of K 3 Fe(CN) 6 should be optimized in future studies, the approach in this study would be useful for designing metal oxide nanocomposites for supercapacitors.

Discussion
In this study, microspheres of WO 3 , nanorods of SnO 2 , and WO 3 /SnO 2 nanocomposite thin films were synthesized via chemical bath deposition, electrodeposition, and calcination. The hexagonal phase of WO 3 and the tetragonal phase of SnO 2 were observed in both the single-component and nanocomposite materials. The average WO 3 crystal size in the nanocomposite was smaller than that in single-component WO 3 . An interaction between WO 3 and the SnO 2 in the nanocomposite was also suggested by the shifts in binding energies measured by XPS. SEM observations indicated the mesoporous surface structures of these nanomaterials, which are preferable for supercapacitor applications.
The CV measurements revealed that the specific capacitance of the WO 3 /SnO 2 nanocomposite electrode reached 530 Fg −1 without K 3 Fe(CN) 6 and 640 Fg −1 with K 3 Fe(CN) 6 at a low scan rate of 5 mV/s. The improved capacitance in the presence of K 3 Fe(CN) 6 was explained by the redox reactions of [Fe(CN) 6 ] 3− /[Fe(CN) 6 ] 4− ions, in addition to the redox reactions of the WO 3 /SnO 2 nanocomposite. Moreover, the WO 3 /SnO 2 nanocomposite and K 3 Fe(CN) 6 could mutually facilitate their redox reactions. The improved specific capacitance was also observed for single-component electrodes (both of WO 3 and SnO 2 ). Therefore, this approach can be used for other electrodes using these metal oxides, although the reaction efficiency could be differed by the electrode species.
Using the symmetric configurations of WO 3 /SnO 2 nanocomposite electrodes, the addition of K 3 Fe(CN) 6 to the 1 M Na 2 SO 4 electrolyte generated a high energy density and a power density of 64 Whkg −1 at 542 Wkg −1 , respectively. The capacitive retention efficiencies for the WO 3 /SnO 2 and WO 3 electrodes with the K 3 Fe(CN) 6 electrolyte was 94.7% and 90.4%, respectively, as indicated by GCD curves after 2000 cycles at 2.5 Ag −1 . The higher cycling stability and the excellent coulombic efficiency (99.9%) of the WO 3 /SnO 2 nanocomposite electrode relative to those of the pure WO 3 electrode were due to the high catalytic activity of the WO 3 /SnO 2 nanocomposite for the redox reactions of [Fe(CN) 6 ] 3− /[Fe(CN) 6 ] 4− ions. As a result, the binder-free WO 3 /SnO 2 nanocomposite with a redox-active electrolyte constituted a promising system for pseudocapacitive energy storage devices. Although the optimal configuration of the WO 3 /SnO 2 nanocomposite and the electrolyte system should be studied further in the future, this approach can also be used with other metal oxide nanocomposites in reactive electrolytes.

Chemical Bath Deposition of WO 3 Nanosphere Thin Films
WO 3 nanospheres were deposited on a carbon cloth substrate by the CBD method. First, a carbon cloth substrate (1 × 1 cm 2 ) was treated with 1 M HNO 3 , acetone, ethanol, and water sequentially with ultrasonication for 20 min and then dried overnight in a vacuum. In a typical preparation of WO 3 nanospheres, Na 2 WO 4 ·H 2 O (0.82 g) was dissolved in 50 mL of water. After stirring for 20 min, 1 M HCl was added dropwise to produce a pH of 1.2 ± 1.0. After stirring for 30 min, (NH 4 ) 2 SO 4 (1.32 g) was added to the above solution.
After stirring for 10 min, 40 mL of the above mixture was transferred to a 100 mL beaker, and the carbon cloth substrate was immersed in it. Then, the beaker was heated on a hotplate at 80 • C for 4 h. The carbon cloth was removed from the bath, rinsed several times with water, and dried overnight at room temperature. Finally, the film was calcined at 400 • C for 2 h. The mass of WO 3 loaded onto the carbon cloth substrate was measured by subtracting the weight of the original carbon cloth substrate from the weight of the product and determined to be 1.2 mg.

Electrochemical Deposition of SnO 2 Nanorods
SnO 2 nanorods were synthesized through electrochemical deposition. First, a carbon cloth substrate with an area of 1 × 1 cm 2 was treated with 1 M HNO 3 , acetone, ethanol, and water through ultrasonication and sequentially dried overnight at ambient temperature. Next, an aqueous solution of SnCl 2 (0.112 g) was prepared in 50 mL of water, and 2.6 mL of HCl was added. After stirring for 30 min, 25 mL of the solution was transferred to a 100 mL beaker, and a three-electrode system was set up: carbon cloth was used as the working electrode (WE), Ag/AgCl in 3 M Cl was used as the reference electrode (RE), and a platinum wire was used as the counter electrode (CE). Then, in a solution of SnCl 2 with HCl, potentiostatic electrodeposition was performed at an applied potential of −0.9 V for 10 min using a galvanostat/potentiostat (Hokuto-Denko, Model HA-151B) at ambient temperature. After deposition, the WE was rinsed with water and dried overnight at ambient temperature. Finally, the deposited material on the carbon cloth was calcined at 400 • C for 2 h. The deposited mass was measured by subtracting the weight of the original carbon cloth substrate from the weight of the product and determined to be 0.4 mg.

Synthesis of a WO 3 /SnO 2 Nanocomposite Electrode
WO 3 /SnO 2 nanocomposite thin films were synthesized through CBD and the electrochemical deposition method. First, the carbon cloth substrate of area (1 × 1 cm 2 ) was sequentially treated by ultrasonication in 1 M HNO 3 , acetone, ethanol, and ultrapure water and then dried overnight at ambient temperature. WO 3 microspheres were deposited on the carbon cloth by CBD, and then the carbon cloth/WO 3 was immersed in the SnCl 2 /HCl solution (25 mL, 0.112 g, and 0.65 M) in a 100 mL beaker [63]. Then, the three-electrode system was set up: the WE were carbon cloth supported by WO 3 nanospheres, the RE was Ag/AgCl in 3 M Cl, and the CE was a platinum wire. Then, potentiostatic electrodeposition was performed at an applied potential of −0.9 V for 10 min using a galvanostat/potentiostat (Hokuto-Denko, Model HA-151B) at ambient temperature. After deposition, the WE were rinsed with water and then dried overnight at ambient temperature. Finally, the deposited material on the carbon cloth was calcined at 400 • C for 2 h on a hot plate. The deposited mass was measured by subtracting the weight of the original carbon cloth substrate from the weight of the product and was determined to be 1.6 mg.

Electrochemical Analyses
All electrochemical analyses were performed using a standard three-electrode system in an electrochemical cell connected to an electrochemical workstation (ZAHNER mass system, model xpot-26366, Germany). An Ag/AgCl electrode in 3 M KCl was used as the RE; a Pt wire as CE; and the WO 3 , SnO 2 , and WO 3 /SnO 2 nanocomposites formed on carbon cloth were used as WEs. An aqueous solution of 1 M Na 2 SO 4 with/without 0.02 M K 3 Fe (CN) 6 was used as the electrolyte for all electrochemical measurements. The electrochemical performance of WO 3 and SnO 2 , as well as the WO 3 /SnO 2 nanocomposite electrodes, was studied using cyclic voltammetry (CV) over the range −0.2 V to 0.7 V vs. Ag/AgCl at different scan rates and galvanostatic charge-discharge (GCD) at different current densities. Electrochemical impendence spectroscopy (EIS) was performed with an amplitude of 10 mV over the frequency range 100 kHz to 100 mHz in an open circuit and described with a Nyquist plot.

Conclusions
In this work, we developed a preparation method of WO 3 /SnO 2 nanocomposite on carbon cloth for supercapacitor electrodes. WO 3 nanospheres were first deposited on a carbon cloth substrate by a chemical bath deposition method, and then a SnO 2 layer was formed by an electrochemical deposition and calcination. The WO 3 /SnO 2 nanocomposite on carbon cloth exhibited a higher specific capacitance (530 F/g at 5 mV/s scan rate) than that of single-component electrodes (240 F/g for WO 3 and 140 F/g for SnO 2 ) in an aqueous electrolyte of 1M Na 2 SO 4 . The addition of 0.02 M K 3 Fe(CN) 6 redoxactive electrolyte into the 1M Na 2 SO 4 electrolyte further improved the charge storage performance of electrodes (440 F/g for WO 3 , 310 F/g for SnO 2 , and 640 F/g for WO 3 /SnO 2 nanocomposite). These results show the synergistic effects of WO 3 /SnO 2 nanocomposites and the advantages of using redox-active electrolytes for supercapacitor systems. Using a symmetric configuration of WO 3 /SnO 2 nanocomposite electrodes, the electrolyte solution of 0.02 MK 3 Fe(CN) 6 and 1 M Na 2 SO 4 electrolyte demonstrated a high energy density and a power density of 64 Whkg −1 at 542 Wkg −1 , respectively. The capacitive retention efficiency for the WO 3 /SnO 2 and WO 3 electrodes with the K 3 Fe(CN) 6 electrolyte was 94.7% after 2000 cycles at 2.5 Ag −1 . Thus, the WO 3 /SnO 2 nanocomposite electrode grown on carbon cloth is a promising candidates as a binder-free electrode for the high performance of a supercapacitor device, and the supercapacitor system using redox-active K 3 Fe(CN) 6 electrolyte was proposed. These strategies developed in this study can also be applied to other metal oxide nanocomposites to improve the performance of supercapacitors.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.