A Low‐Concentration and High Ionic Conductivity Aqueous Electrolyte toward Ultralow‐Temperature Zinc‐Ion Hybrid Capacitors

Aqueous electrochemical energy storage devices have attracted tremendous attention because of its high safety, low cost, and environmental friendliness. However, their low‐temperature operation is plagued by the freeze of electrolytes. Herein, a 3 mol kg−1 Zn(ClO4)2 electrolyte without adding any organic solvents or antifreezing additives is proposed, which yields a high ionic conductivity of 9.4 mS cm−1 even at ultralow temperatures of −60 °C. The strong electrostatic interaction between Zn2+ ion and water molecules and the structure breaking effect of ClO4− ions to destroy the hydrogen bond network between water molecules in Zn(ClO4)2 electrolyte is revealed by spectroscopic characterization and theoretical simulation. This low‐temperature electrolyte renders the zinc‐ion hybrid capacitor to exhibit a high energy density of 40.91 Wh kg−1 at −60 °C and a long‐cycle life (over 200 days) at −30 °C. This study provides a new path to develop low‐concentration antifreezing electrolytes for aqueous electrochemical energy storage devices.

The freezing of aqueous electrolytes at low temperatures arises from the formation of 3D hydrogen bond network. [16][17][18][19] In order to suppress the formation of hydrogen bonds between water molecules, organic solvents are usually introduced to form stronger hydrogen bonds with water molecules, thus destroying the hydrogen bond network between water molecules. [20][21][22] Ethylene glycol is a commonly used antifreeze in industry. When the proportion of ethylene glycol in water reaches 60%, the freezing point of the binary solution can be as low as À50°C. [23] Li et al. introduced ethylene glycol into ZnSO 4 electrolyte to depress the freezing point of the electrolyte. [24] By adjusting the ratio of ethylene glycol to water, the as-prepared hybrid electrolyte not only did not freeze, but also exhibited a superior ionic conductivity of 6.9 mS cm À1 at À40°C, so that the ZIHC using this electrolyte still worked normally at À40°C. Hu et al. adopted the same strategy to realize that the as-prepared ZnSO 4 electrolyte still did not freeze at extremely low temperature of À50°C, and the ZIHC using this electrolyte worked normally in an environment as low as À40°C. [25] Although the addition of organic solvents can significantly expand the low-temperature operation limit of ZIHCs, organic solvents are generally volatile, flammable, and toxic, which would affect the high safety and environmental friendliness of aqueous ZIHCs. More importantly, it would seriously reduce the solubility of salt, resulting in the reduction of ionic conductivity of electrolytes, and thus restricts the lowest operation temperature of ZIHCs. [26] Increasing the concentration of electrolytes can also destroy the original hydrogen bond network between water molecules, and then depress the freezing point of electrolytes. [26,27] Chen et al. found that with the increase of ZnCl 2 electrolyte concentration, the freezing point of the electrolyte decreased gradually due to the gradual destruction of hydrogen bonds. [17] When the concentration of ZnCl 2 electrolyte reached 7.5 mol kg À1 , the freezing point of the electrolyte was reduced to À114°C. When the concentration of ZnCl 2 electrolyte continued to increase (>7.5 mol kg À1 ), the freezing point of the electrolyte increased rather than decreased due to the enhanced interaction between ions. The 7.5 mol kg À1 ZnCl 2 electrolyte presented an ionic conductivity of 1.79 mS cm À1 at À60°C, and the zinc battery using the low-temperature electrolyte showed good performance at À70°C. However, compared with the commonly used low-concentration electrolytes, the high-concentration electrolyte has higher cost. More importantly, the ZnCl 2 electrolyte displays a narrow voltage window (from 0.2 to 1.4 V) in ZICHs, which is not conducive to obtain the high energy density. [28] Therefore, it is very necessary to develop low-concentration and wide voltage window electrolytes to realize the satisfactory low-temperature performance for ZICHs.
In this work, we proposed a low-concentration and high ionic conductivity Zn(ClO 4 ) 2 aqueous electrolyte without adding any organic solvents or antifreezing additives (Table S1, Supporting Information). The 3 mol kg À1 (denoted as m) Zn(ClO 4 ) 2 electrolyte still exhibited liquid state at extremely low temperature of À65°C, which originates from the strong electrostatic interaction between Zn 2þ ion and water molecules and the structure breaking effect of ClO 4 À ions to destroy the hydrogen bond network between water molecules, thus inhibiting the formation of ice crystals. The low freezing point of the electrolyte endows a high ionic conductivity of 9.4 mS m À1 at ultralow temperature of À60°C. As a result, the

Results and Discussion
Ionic conductivity plays a key role in the low-temperature performance of electrochemical energy storage devices. [14] According to Arrhenius formula or Vogel-Tammann-Fulcher (VTF) empirical equation, the ionic conductivity of electrolytes would decrease with the decrease in temperature. [29] However, if the electrolyte is frozen, the ionic conductivity of frozen electrolyte would drop sharply. By contrast, if the electrolyte can remain liquid state at low temperatures, the ionic conductivity may be reduced slowly following the Arrhenius formula or VTF empirical equation. Thus, we begin our investigation by observing the state of Zn(ClO 4 ) 2 solutions with various concentrations at low temperatures. As shown in Figure 1a, 3 m Zn(ClO 4 ) 2 solution still keeps liquid state at À60°C, which demonstrates that it has a lower freezing point than À60°C, while 1 m Zn(ClO 4 ) 2 solution becomes solid state at À20°C, and 2 m Zn(ClO 4 ) 2 solution and 4 m Zn(ClO 4 ) 2 solution become solid state at À40°C. To further verify the state of various concentration Zn(ClO 4 ) 2 solutions, differential scanning calorimetry (DSC) test was carried out. The DSC curves of four Zn(ClO 4 ) 2 solutions with various concentrations are displayed in Figure 1b. 1 m Zn(ClO 4 ) 2 solution exhibits a peak at À30°C, which is lower than the temperature that the solution becomes solid state as shown in Figure 1a (À20°C). The temperature difference may be ascribed to the supercooling. [30] There is an exothermic peak for 2 m Zn(ClO 4 ) 2 solution at À38°C, indicating that it freezes around À38°C, which is consistent with the phenomenon observed in Figure 1a. Amazingly, the exothermic peak of 3 m Zn(ClO 4 ) 2 solution does not appear from 10 to À65°C, implying that it still does not freeze at À65°C. As the lowest operating temperature of the differential scanning calorimeter used to characterize the freezing point can only reach À65°C, we cannot accurately measure the freezing point of 3 m Zn(ClO 4 ) 2 solution in this work. For 4 m Zn(ClO 4 ) 2 solution, a wide peak appears from 10 to À65°C. Through the above analysis, it can be found that 3 m Zn(ClO 4 ) 2 solution has the lowest freezing point at low temperatures, which signifies that the 3 m Zn(ClO 4 ) 2 solution may have the highest ionic conductivity. To verify the point, we measured the ionic conductivity of the four solutions at various temperatures ( Figure 1c and S1, Supporting Information). The result demonstrates that 3 m Zn(ClO 4 ) 2 solution maintains the highest ionic conductivity at any temperature, and it displays an ionic conductivity of 9.4 mS cm À1 even at À60°C. It should be noted that with the decrease of temperature, the ionic conductivity of 3 m Zn(ClO 4 ) 2 solution also decreased. However, compared with frozen 1 m Zn(ClO 4 ) 2 solution, the phenomenon of ionic conductivity of 3 m Zn(ClO 4 ) 2 solution dropping sharply caused by freezing does not appear, which indicates that suppressing the freezing of electrolyte can effectively slow down the decrease of ionic conductivity.
Next, the mechanism of low freezing point of 3 m Zn(ClO 4 ) 2 solution was investigated. The formation of ice originates from the gradual enhancement of hydrogen bond interaction between water molecules, forming an orderly 3D hydrogen bond network. [16] In order to explore the mechanism of low freezing point of 3 m Zn(ClO 4 ) 2 solution, the hydrogen bond interaction between water molecules was studied by Fourier transform infrared spectroscopy (FTIR) (Figure 1d and S2, Supporting Information). Figure 1d presents the O-H stretching vibration of Zn(ClO 4 ) 2 solutions with various concentrations and pure water. For pure water, there is a wide peak at %3353 cm À1 . When salt is dissolved in water, two shoulders at %3552 and %3222 cm À1 are observed except for the main peak at %3353 cm À1 , which are attributed to the water molecules with nonhydrogen bonds or weak hydrogen bonds and the water molecules with strong hydrogen bonds, respectively. [31] And the high wavenumber shoulder at %3552 cm À1 gradually increases while the low wavenumber shoulder at %3222 cm À1 gradually decreases with gradually increased solution concentration. Moreover, the O-H bending bands shift from 1634 to 1625 cm À1 with the increase in solution concentration ( Figure S2, Supporting Information), which also signifies that the strong hydrogen bonds decrease and the weak hydrogen bonds increase. [32] These results demonstrate that with the gradual increase of Zn(ClO 4 ) 2 solution concentration, the hydrogen bond interaction between water molecules is gradually destroyed.
Raman spectroscopy was used to further reveal the relationship between hydrogen bond and Zn(ClO 4 ) 2 solution concentration. As shown in Figure 1e, pure water presents a broad peak from 3000 to 3800 cm À1 , which can be deconvolved into three subpeaks. The three subpeaks represent strong hydrogen bond, weak hydrogen bond, and nonhydrogen bond from left to right. [17] When Zn(ClO 4 ) 2 salt is dissolved in water, 1 m Zn(ClO 4 ) 2 solution maintains the characteristic peak of pure water (%3445 cm À1 ), while a wide peak appears at %3559 cm À1 . The deconvolution analysis exhibits that the strong hydrogen bond components of 1 m Zn(ClO 4 ) 2 solution are reduced and the nonhydrogen bond components are increased compared with pure water. When the concentration of Zn(ClO 4 ) 2 solution increases to 2 m, the characteristic peak (%3445 cm À1 ) of pure water is divided into two peaks (%3446 and %3554 cm À1 ). It is found that compared with 1 m Zn(ClO 4 ) 2 solution, the strong hydrogen bond component and weak hydrogen bond component of 2 m Zn(ClO 4 ) 2 solution are further reduced, and the nonhydrogen bond components are further increased. When the concentration of Zn(ClO 4 ) 2 solution reaches 3 m, the peak at  %3446 cm À1 disappears, and only one characteristic peak locates at %3554 cm À1 . In 3 m Zn(ClO 4 ) 2 solution, nonhydrogen bond becomes the main component, which demonstrates that Zn(ClO 4 ) 2 salt destroys the hydrogen bond interaction between water molecules. [33] Therefore, 3 m Zn(ClO 4 ) 2 solution is not capable of forming an orderly 3D hydrogen bond network, thus suppressing the formation of ice. Continuing to increase the concentration of Zn(ClO 4 ) 2 solution to 4 m, it basically maintains the characteristic peak of 3 m Zn(ClO 4 ) 2 solution, and the number of hydrogen bonds in the 4 m Zn(ClO 4 ) 2 solution is slightly lower than that of 3 m Zn(ClO 4 ) 2 solution. It should be noted that although the 4 m Zn(ClO 4 ) 2 solution has fewer hydrogen bonds than that of 3 m Zn(ClO 4 ) 2 solution at room temperature, it is easy to salt precipitate at low temperatures because of its high concentration. [20] And the precipitated salt would become nucleation sites, which would help the formation of ice crystals. [30] Therefore, the 4 m Zn(ClO 4 ) 2 solution shows a higher freezing point than that of 3 m Zn(ClO 4 ) 2 solution.
To explore the existing forms of Zn 2þ ions and ClO 4 À ions in the solution, the Cl-O stretching vibration in Raman spectrum was analyzed. As displayed in Figure 1f, when salt is dissolved in water, the Cl-O stretching vibration peak exhibits a redshift from 945 to 942 cm À1 , which is attributed to the dissociation of Zn(ClO 4 ) 2 salt dissolved in water. For Zn(ClO 4 ) 2 salt, it exists in the form of ion pair, and the ion pair would be dissociated into Zn 2þ cation and ClO 4 À anion when it is dissolved in water. [34] For ClO 4 À ions in the solutions, four species including free ClO 4 À anions, solvent separated ion pairs M þ -solvent-ClO 4 À (M stands for metal), contact ion pairs M þ ClO 4 À , and multiple ion aggregates [M þ ClO 4 À ] n from low wavenumber to high wavenumber have been identified. [33,35] It is found that Zn 2þ ions and ClO 4 À ions mainly exist in the form of solvent separated ion pair [Zn 2þ (H 2 O) n (ClO 4 À ) 2 ] in 1 m Zn(ClO 4 ) 2 solution. [33] With the gradual increase of solution concentration, the form of solvent separated ion pair is kept, and its intensity increases slightly with the increase of solution concentration ( Figure S3, Supporting Information). The deconvolution analysis results also confirm that solvent separated ion pair is the main existing form for Zn 2þ ions and ClO 4 À ions ( Figure 1f ). Therefore, in 3 m Zn(ClO 4 ) 2 solution, Zn 2þ ions and ClO 4 À ions also exist mainly in the form of solvent separated ion pair.
Molecular dynamics (MD) simulation was conducted to further elucidate the mechanism of low freezing point of 3 m Zn(ClO 4 ) 2 solution ( Figure 2). Figure 2a and S4, Supporting Information, exhibit the MD simulation snapshots of Zn(ClO 4 ) 2 solution with different concentrations, and they reveal that the ionic solvation cluster of Zn 2þ ion coordinated with water molecules and ClO 4 À ions exists in the solutions. To understand the effect of Zn(ClO 4 ) 2 salt on the solution structure, the radial distribution function (RDF), and coordination numbers for Zn 2þ ion coordinating with water molecules (denoted as Zn-O*), Zn 2þ ion coordinating with ClO 4 À ions (denoted as Zn-O═) was calculated. In terms of Zn-O*, the first solvation shell of Zn 2þ ion is surrounded by water molecules, and the frequency of water molecules within the Zn 2þ ion first solvation shell gradually increases with the increase in solution concentration (Figure 2b). The calculated coordination number result shows that there are six water molecules in the first solvated shell of Zn 2þ ion, which is consistent with the reported literature. [36,37] And the coordination number of water molecules does not change with the increase of solution concentration ( Figure 2c, Table S2, Supporting Information). By contrast, the frequency of water molecules in the second solvated shell of Zn 2þ ion is significantly lower than that in the first solvated shell, and the frequency of water molecules gradually decreases with the increase of solution concentration. Combined with the RDF of Zn-O═, we find that ClO 4 À ions do not exist in the first solvated shell of Zn 2þ ion, and the second solvated shell of Zn 2þ ion is mainly occupied by ClO 4 À ions (Figure 2e). And with the increase of solution concentration, the frequency of ClO 4 À ions within the second solvated shell of Zn 2þ ion gradually increases, and the coordination number of Zn-O═ gradually increases with the increase of solution concentration (Figure 2f and Table S2, Supporting Information), which explains why the frequency and coordination number of water molecules gradually decrease with the increase of solution concentration within the second solvated shell of Zn 2þ ion. The above analysis demonstrates that Zn 2þ ions and ClO 4 À ions mainly form solvent separated ion pair of solution structure in the solution ( Figure S5, Supporting Information), which is in line with the Raman spectrum results. To verify that the existence of such solution structure arises from the strong electrostatic interaction between Zn 2þ ion and water molecules, we calculated the binding energy to estimate the interaction between them ( Figure S6, Supporting Information). The computational results demonstrate that water molecules have a stronger interaction with Zn 2þ ion than that of between water molecules. It is the strong interactions between Zn 2þ ion and dipolar water molecules that significantly reduce the aggregation of water molecules and weaken internal hydrogen bonds, and the average hydrogen bond number decreases with the increase of solution concentration (Figure 2d). And the second solvated shell of Zn 2þ ion formed by ClO 4 -ions would block the interaction between the water molecules in the first solvated shell and the water molecules outside the solvated shell, which attributes to structure breaking effect of ClO 4 À ions on the hydrogen bond network of water molecules. [31,38] The destruction of hydrogen bond between water molecules inhibits ice crystallization in 3 m Zn(ClO 4 ) 2 solution. Therefore, the liquid state enables 3 m Zn(ClO 4 ) 2 solution to maintain excellent ionic conductivity at À60°C.
The high ionic conductivity of electrolytes at low temperatures would help electrochemical energy storage devices obtain outstanding low-temperature performance. To examine the hypothesis, we first used the 3 m Zn(ClO 4 ) 2 solution as electrolyte for symmetric supercapacitors based on ions adsorption/desorption behavior, which was assembled with commercial activated carbon as electrode material. The electrochemical performance of supercapacitors at various operating temperatures was carried out by cyclic voltammetry (CV) and galvanostatic charging/discharging (GCD). Figure 3a exhibits the CV curves of supercapacitors at various operating temperatures. The CV curves are almost rectangular in the operating temperature from 20 to À50°C. With the decrease of operating temperature, the area of CV curves only decreases slightly, which signifies that capacitance decay of the supercapacitor at low temperatures is not serious. According to the specific capacitance calculated from the CV curves, the supercapacitor obtains a specific capacitance of 30.0 F g À1 at room temperature (Table S3, Supporting Information). When the operating temperature drops from 20 to À50°C, the specific capacitance of the supercapacitor still maintains 77.8% of the room temperature capacitance (Table S3, Supporting Information). The results of GCD curves also confirm that the specific capacitance of supercapacitor does not decline much when the operating temperature decreases from 20 to À50°C (Figure 3b). Figure 3c and S7-S11, Supporting Information, present the rate capability of the supercapacitor at various operating temperatures. It can be observed that the rate capability of the supercapacitor increases first and then decreases (Table S3, Supporting Information). At  www.advancedsciencenews.com www.small-structures.com the operating temperature of 20°C, the rate capability of supercapacitor is 88.16%. When the operating temperature drops to À20°C, the rate capability of the supercapacitor reaches a maximum value of 91.42%, which probably results from reduced side effects and parasitic capacitance at low current density at low temperature. [22] Then, the rate capability of the supercapacitor decreases gradually with the decrease of operating temperature. At the operating temperature of À50°C, the rate capability of the supercapacitor is reduced to 85.14%, which is still a satisfactory result and better than our previous work. [20] In order to explore the mechanism for excellent performance of the supercapacitor even at the operating temperature of À50°C, the electrochemical impedance spectroscopy (EIS) of the supercapacitor at various operating temperatures was measured (Figure 3d). It is obvious that the internal resistance and charge transfer resistance of the supercapacitor are still small when the operating temperature drops to -50°C (Figure 3d,e, Table S3, Supporting Information), which demonstrates that the ions still have fast kinetics even at À50°C. To further investigate the relationship between ion kinetics and operating temperature, the relaxation time constant (τ 0 ) is derived from the plots of impedance phase angles (Figure 3f ). Shorter relaxation time represents faster ion transport kinetics. [39] Notably, a very small τ 0 of 3.16 s is presented at À50°C, which indicates that the ions still have a very fast migration capacity even at such low temperature. These results confirm that fast ions transport in the 3 m Zn(ClO 4 ) 2 electrolyte endows the device with excellent charging/discharging capability at low temperatures.
To further demonstrate the feasibility of 3 m Zn(ClO 4 ) 2 electrolyte in the ZIHC, we employed the zinc plate as an anode to assemble the ZIHC. Previous studies have revealed that the energy storage mechanism of ZIHC is the hybrid energy storage mechanism of Zn 2þ ions plating/stripping on the Zn anode and the adsorption/desorption of ClO 4 À ions and Zn 2þ ions on the activated carbon cathode. [40] Specifically, during charging process, Zn 2þ ions are deposited on the anode, and Zn 2þ ions are desorbed and ClO 4 À ions are adsorbed on the cathode. During the discharging process, the reverse process would occur. Therefore, before testing the electrochemical performance of the ZIHC, the effect of 3 m Zn(ClO 4 ) 2 electrolyte on the Zn electrode was first explored. The Zn stripping/plating coulombic efficiency on Cu foils and cycle life of Zn-Zn symmetric cells in 3 m Zn(ClO 4 ) 2 electrolyte at room temperature were tested. After the initial stabilization cycles, the coulombic efficiency achieves an average value of 94% ( Figure S12, Supporting Information). Moreover, the Zn-Zn symmetric cells exhibit superior cycling stability, and the voltage profiles were nearly unchanged for at least 270 h ( Figure S13, Supporting Information), further demonstrating the high efficiency of Zn electrode in 3 m Zn(ClO 4 ) 2 electrolyte.
The electrochemical performance of the ZIHC at various operating temperatures was performed by CV and GCD tests. Figure 4a displays the CV curves of the ZIHC with the scan rate of 5 mV s À1 at various operating temperatures. The nonideal rectangle shown by these CV curves confirms the hybrid energy storage mechanism of ZIHCs. [41] A pair of redox peaks is shown at around 0.9 V, which are attributed to the Zn 2þ deposition/stripping on the surface of Zn anode. [42] The normal CV curve obtained at À60°C demonstrates that the ZIHC can work normally at À60°C. The GCD curve of the ZIHC at current density of 0.5 A g À1 also identifies that the ZIHC can work normally at -60°C (Figure 4b). The excellent ion adsorption/desorption behavior on the activated carbon cathode and fast ion transport in the electrolytes enables the ZIHCs to obtain remarkable charge/discharge capability at various operating temperatures. It should be noted that our ZIHC achieves a lower operating temperature than other reported works, [24,43] and it can work normally at a higher current density than our previous work (Figure 4c). [10] The specific capacitance of ZIHC at various operating temperatures was calculated. At room temperature, the ZIHC obtained a specific capacitance of 152.09 F g À1 . When the operating temperature dropped to À60°C, the ZIHC can still obtain a specific capacitance of 102.27 F g À1 , which is 67.24% of the room temperature capacitance. The voltage window of the ZIHC is 0.1-1.7 V, which is higher than that of the supercapacitor. Meanwhile, as the specific capacitance of the ZIHC also increases significantly, the energy density has been greatly improved. At room temperature, the energy density of ZIHC reaches 60.83 Wh kg À1 . Even if the operating temperature reduced to À60°C, the ZIHC still has a high energy density of 40.91 Wh kg À1 , which is higher than other previous reported work. [24] Also, the cycling stability of the ZIHC with 3 m Zn(ClO 4 ) 2 electrolyte is evaluated. Figure 4d depicts the cyclic curve of the ZIHC with the current density of 1 A g À1 at room temperature (20°C). It can be found that the ZIHC continuously operates over 30 000 cycles (more than 180 days) without noticeable capacitance decay. When the ZIHC was placed in a constant environment of À30°C, it exhibits ultralong cycling stability over 40 000 cycles (continuous operation more than 200 days) with the current density of 1 A g À1 , which is significantly superior to previous reported ZIHC devices (Figure 4e). [24,43] After 40 000 cycles, the capacitance of ZIHC does not decline, indicating that it has very excellent cycle performance. To further investigate the effect of 3 m Zn(ClO 4 ) 2 electrolyte on cycling stability of the ZIHC, the scanning electron microscopy (SEM) and X-ray diffraction (XRD) tests were carried out. Dendrite-free Zn on the surface of Zn anode is observed from SEM images at 20 and À30°C ( Figure S14, Supporting Information), which would contribute the ZIHC to achieving excellent cycle stability. [44] XRD analysis suggests that although the by-product zinc hydroxide is generated after 500 cycles at 20°C, there is almost no by-product at À30°C ( Figure S15, Supporting Information), which is ascribed to the reduced side reactions at low temperatures. [40,45] These results demonstrate that the 3 m Zn(ClO 4 ) 2 electrolyte can stabilize the operation of ZICHs.

Conclusion
In this work, a low-concentration, cost-effective, and anti-freezing Zn(ClO 4 ) 2 electrolyte without any antifreeze additive is proposed. The 3 m Zn(ClO 4 ) 2 electrolyte exhibits a high ionic conductivity of 9.4 mS cm À1 even at an ultralow temperature of À60°C. Experimental analyses together with theoretical calculations reveal that the unique solvation structure of Zn 2þ ion in the electrolyte effectively destroys the hydrogen bond network between water molecules, thus providing an electrolyte with a lower freezing point. As a proof-of-concept, the 3 m Zn(ClO 4 ) 2 electrolyte enables ZIHCs to exhibit superior low-temperature performance even at À60°C. A superb energy density of 40.91 Wh kg À1 at À60°C and a remarkable long-term cycling stability over 40 000 cycles (more than 200 days) at À30°C are obtained. This work not only proposes an antifreezing electrolyte with low-concentration and high ionic conductivity for low-temperature ZIHCs, but also offers new insight into the solvation structure of Zn 2þ ion in the electrolytes, which provides new ideas in designing other low-temperature electrolytes.

Experimental Section
Preparation of the Aqueous Electrolytes: Zn(ClO 4 ) 2 •6H 2 O purchased from Alfa was dissolved in deionized water to prepare different concentration aqueous electrolytes. The bound water in inorganic salts should be considered when preparing the aqueous electrolytes. The content of added water is equal to the total amount of 1 kg water minus the corresponding bound water content.
Assembly of Symmetric Supercapacitors: Commercial activated carbon (YP-50F, Kuraray) was used as the electrode material of supercapacitors, and stainless steel mesh was used as the collector. Activated carbon, acetylene black, and polytetrafluoroethylene binder were mixed in the ratio of 8:1:1, and the slurry was scraped onto the collector. The loading amount of active materials was 1 mg cm À2 . Glass fiber (GF/D, Whatman) was acted as separator, and 3 m Zn(ClO 4 ) 2 solution was used as electrolyte to assemble the supercapacitor in air.
Assembly of ZIHCs: Commercial activated carbon (YP-50F, Kuraray) and commercial zinc plate (thickness of 0.2 mm) were used as cathode material and anode material of ZIHCs, respectively. After zinc plate was cut to 1 cm Â 1 cm, it was directly used as the anode. The cathode adopted stainless steel mesh as the current collector. And the activated carbon, acetylene black, and polytetrafluoroethylene binder were mixed in the ratio of 8:1:1, and the slurry was scraped onto the current collector. The glass fiber (GF D, Whatman) and 3 m Zn(ClO 4 ) 2 solution were used as separator and electrolyte, respectively.
Characterizations: DSC1 (Mettler Toledo) was employed to characterize the freezing point of solutions under a nitrogen atmosphere with a cooling rate of 5°C min À1 . Raman spectra were acquired on Horiba LabRAM HR Evolution with 532 nm excitation. FTIR spectra were collected by PerkinElmer Spectrum Two infrared spectrometer.
The ionic conductivity was measured by EIS. The testing device consisted of two 1 cmÂ1 cm platinum plates. The ionic conductivity of solutions was calculated by the following formula [46] σ ¼ l RS (1) Figure 4. a) CV curves of the ZIHC with the scan rate of 5 mV s À1 at various temperatures. b) GCD of the ZIHC with the current density of 0.5 A g À1 at various temperatures. c) Performance comparison of the ZIHCs. Cycling performance of ZIHC at d) room temperature and e) À30°C with the current density of 1 A g À1 .
www.advancedsciencenews.com www.small-structures.com where σ is the ionic conductivity, l is the distance between two platinum plates, S is the area of platinum plate, and R is the resistance of solution.
The resistance can be obtained by the intersection of electrochemical impedance spectrum and x-axis.
The electrochemical tests, such as CV, GCD and EIS, were conducted by an electrochemical workstation (CHI 660E, Chenhua). The cycling stability test was carried out by a LAND system (CTA2001A) in an ultralow-temperature freezer (BC/BD-102DNE, AUCAMA). The specific capacitance of the device is calculated by equation [39] C ¼ 1 2ΔVvm where C is the specific capacitance, i is the current, ΔV is the difference of voltage window, v is the scan rate, m is the mass of electrode material, V 1 and V 2 are the initial and vertex potentials, respectively. Low-temperature electrochemical performance tests were carried out in a temperature and humidity test chamber (Guangdong Kowin Testing Equipment Co., Ltd., China). Before low-temperature measurements, the device should be kept at each test temperature for at least 2 h to ensure the temperature of device reaches steady state.
Computational Details: MD simulations were performed with open-source software Gromacs. [47] The simulation systems consist of 2220 water molecules, 40/80/120/160 Zn 2þ ions and 80/160/240/320 ClO 4 À ions, corresponding to solution concentrations of 1, 2, 3, and 4 m, respectively. SPC/E model was employed for water. [48] Parameters proposed by Merz et al. were employed for Zn 2þ ion. [49] General Amber force field with the RESP atomic charges was used to model ClO 4 À ion. [50] Van der Waals interactions were represented by the Lennard Jones (LJ) potential with a cutoff value of 1.2 nm, with Lorentz-Berthelot combining rules to derive parameters between different atoms. Electrostatic interactions were calculated with the particle-mesh Ewald method, with the short-range part truncated at 1.2 nm. Energy minimization was first conducted for initial configurations. Subsequently, 1 ns equilibration at 298.15 K and 1 bar was conducted, and the final conformations were used as the starting points of 10 ns MD runs carried out under NPT ensemble (298.15 K/1 bar, with V-rescale thermostat and Parrinello-Rahman barostat, respectively). The leap-frog algorithm with a time step of 1 fs was used to integrate the equations of motion. 3D periodic boundary conditions were applied in all simulations.
Vienna Ab Initio Simulation Package (VASP) code was adopted to calculate the binding energies for Zn(ClO 4 ) 2 -H 2 O and H 2 O-H 2 O. [51] Generalized-gradient approximation with the PBE functional was used to describe the exchange correlation interaction. The interaction between ions and electrons was described by the projector augmented wave (PAW) method. The cutoff energy of the plane wave basis set was 520 eV, and the energy and force convergence criteria were 10 À5 eV and 0.01 eV Å À1 , respectively. In this work, a supercell of 10 Å Â 10 Å Â 10 Å and Γ-centered 5 Â 5 Â 5 k-point mesh was used for geometry optimization and static calculations.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.