Core–shell composite of wood-derived biochar supported MnO₂ nanosheets for supercapacitor applications

Abstract

Eco-friendly wood-derived biochar (WDB) was used as a substrate material to support sheet-like nano-MnO₂ via an easily-operated in situ redox reaction between the biochar and KMnO₄. WDB was readily obtained by pyrolyzing wood waste of agriculture and industry. The MnO₂/WDB composite displays a core–shell structure and can be utilized as a free-standing and binder-free supercapacitor electrode. The MnO₂/WDB electrode has a moderate specific capacitance of 101 F g⁻¹, an excellent coulombic efficiency of 98–100%, and a good cyclic stability with a capacitance retention of 85.0% after 10 000 cycles, making it useful for supercapacitor applications. Moreover, it is expected that such porous inexpensive WDB can serve as a novel harmless substrate material to combine with other electrochemical active substances for the development of high-performance energy storage devices.

---

1. Introduction

In response to the changing global landscape, energy has become a primary focus of the major world powers and scientific community. There is an increasing importance to develop and refine more efficient energy-conversion and storage devices. One class of such devices, called supercapacitors, have attracted increasing attention due to their ultrahigh power density, fast charge–discharge properties, long lifecycle, excellent reversibility (90–95% or higher), widespread operation temperature, and their ability to bridge the power/energy gap between traditional dielectric capacitors and batteries/fuel cells.^1,2 The most extensively utilized electrode active materials for supercapacitors include carbon materials (typically like activated carbons, carbon nanofibers, carbon nanotubes, and graphene),^1,3,4 conducting polymers (typically like polyaniline, polypyrrole, and polythiophene),⁵ and transition metal oxides/hydroxides (typically like RuO₂, Co₃O₄, NiO, MnO₂, Ni(OH)₂, and Co(OH)₂).⁶ Among these materials, MnO₂ is particularly attractive because of its cost effectiveness, high specific capacitance (the theoretical value reaches up to 1370 F g⁻¹ based on a one-electron redox reaction per manganese atom),⁷ environmental compatibility, and remarkable structural versatility.⁸ Nevertheless, the poor electrical conductivity and redox kinetics of MnO₂ seriously hinder its electrochemical applications.^9,10 To address such intrinsic limitations and maximize utilization of MnO₂ pseudocapacitance, a common strategy is to integrate low-dimensional oxide materials with highly conductive substrates (like carbon materials).^11–13 This strategy creates a hybrid supercapacitor with a large pseudocapacitance in addition to the capacitance provided by the electrical double-layer at the supporting electrode surface.¹⁴ Also, this strategy contributes to acquiring both large active surface area and good electrical connection. Liu et al. coaxially coated ultrathin manganese oxide layers on a vertically aligned carbon nanofiber array via cathodic electrochemical deposition, and the core–shell nanostructure demonstrates high performance in maximum specific capacitance (365 F g⁻¹), specific energy (32.5 W h kg⁻¹) and specific power (6.216 kW kg⁻¹).¹⁴ Recently, Wang et al. conformally coated α-MnO₂ nanowires onto the foamed metal foils supported 3D few-layer graphene/multi-walled carbon nanotube architecture by bath deposition, and the resultant hierarchical hybrid foam shows a high specific capacitance (1108.79 F g⁻¹) and power density (799.84 kW kg⁻¹) and a great capacitance retention (97.94%) after 13 000 charge–discharge cycles.¹⁵ Other typical carbon materials, such as activated carbon,¹⁶ graphene oxide,¹² and carbon foam derived from melamine resin,¹⁷ have also been used to combine with MnO₂ nanostructures with various morphologies for supercapacitor applications.

Wood, a naturally grown composite material of complex hierarchical cellular structure, has been always considered as an important natural renewable resource in the process of human being's development. Wood-derived biochar (WDB) is a carbonaceous solid residue produced from the thermal treatment of wood via oxygen-limited pyrolysis.¹⁸ Compared with the biochar derived from other bioresources, WDB is dominant because of its large specific surface area¹⁹ (as high as 683 m² g⁻¹) and abundant pore structure facilitating the fast penetration of electrolytes to allow rapid electron-transfer for charge storage and delivery.²⁰ Moreover, WDB is more readily available because of its fast simple preparation technique and cheap feedstock from wood waste of agriculture and industry, as compared to some aforementioned carbon materials. It is found that pure WDB supercapacitors (non-faradaic electric double-layer capacitance) display poor electrochemical activity, and the corresponding specific capacitances are only several tens of farad per gram in general.^19,21 When integrated with transition metal oxides²² (like Cu₂O and CuO) or conducting polymers²³ (like polyaniline) with faradaic pseudo-capacitance, the hybrid supercapacitors show significantly improved electrochemical properties. So far, only several literatures studied WDB-based hybrid supercapacitors;^19,21–24 to the best of our knowledge, there is still no report on the hybrid WDB/MnO₂ electrode. We now wish to report this electrode and its electrochemical properties.

MnO₂ can be synthesized using various techniques, such as thermal decomposition,²⁵ electrochemical deposition,²⁶ hydrothermal method,²⁷ and sol–gel method.²⁸ In the present work, we adopted a simple cost-effective chemical reduction method to convert potassium permanganate (KMnO₄) into MnO₂ by directly using the WDB as reducing agent, as illustrated in Fig. 1. This method helps to deposit MnO₂ nanosheets onto the surface of WDB and generate a core–shell structure. This MnO₂/WDB composite is expected to be used as a free-standing and binder-free electrode, which is beneficial to enhance rate performance and stability and reduce interface resistance.²⁹ The MnO₂/WDB composite was characterized by scanning electron microscopy (SEM), transmission electron microscopy (TEM), high-resolution TEM (HRTEM), selected area electron diffraction (SAED), energy dispersive X-ray spectroscopy (EDX), X-ray photoelectron spectroscopy (XPS), and X-ray diffraction (XRD). The electrochemical properties were studied through cyclic voltammograms (CV), galvanostatic charge–discharge (GCD), and electrochemical impedance spectroscopy (EIS) tests in a three-electrode configuration in 1 M Na₂SO₄ aqueous electrolyte.

Fig. 1 Schematic illustration of the synthetic procedures for MnO₂/WDB.

2. Experimental

2.1. Materials

Potassium permanganate (KMnO₄) was supplied by Kemiou Chemical Reagent Co., Ltd. (Tianjin, China) and used as received. Wood waste of agriculture and industry was collected and cut into slices with a thickness of 1 mm. These wood slices were ultrasonically rinsed with distilled water for 30 min and dried at 60 °C for 24 h in a vacuum.

2.2. Synthesis of wood-derived biochar (WDB)

The biochar was synthesized by transferring the wood slice into a tubular furnace for pyrolysis under the protection of nitrogen. In a typical process, the wood slice was heated to 500 °C at a heating rate of 5 °C min⁻¹, and this temperature was maintained for 1 h; then the sample was heated to 1000 °C at 5 °C min⁻¹ and held at this temperature for 2 h to allow for complete pyrolysis. Thereafter, the temperature decreased to 500 °C at 5 °C min⁻¹ and finally decreased naturally to the room temperature.

2.3. Synthesis of MnO₂/wood-derived biochar (MnO₂/WDB) composite

The as-prepared WDB (ca. 0.07 g) was firstly immersed in the aqueous solution of KMnO₄ (50 mL). The weight ratio of biochar to KMnO₄ was set as 4 : 1, which has been confirmed elsewhere to contribute to acquiring favorable electrochemical property for the composites consisting of MnO₂ and carbon.³⁰ The beaker containing the above mixture was then transferred into an oven and covered with a glass culture dish. The mixture was heated at 60 °C for 12 h. After the heating, the sample was rinsed with a great deal of distilled water and finally dried at 60 °C for 24 h in a vacuum.

2.4. Characterization

SEM observations were performed with a Hitachi S4800 SEM equipped with an EDX detector for element analysis. TEM and HRTEM observations and SAED were performed with a FEI, Tecnai G2 F20 TEM with a field-emission gun operating at 200 kV. XPS was carried out using a Thermo Escalab 250Xi XPS spectrometer equipped with a dual X-ray source using Al-Kα. Deconvolution of the overlapping peaks was performed using a mixed Gaussian-Lorentzian fitting program (Origin 8.5, Originlab Corporation). XRD spectroscopy was implemented on a Bruker D8 Advance TXS XRD instrument with Cu Kα (target) radiation (λ = 1.5418 Å) at a scan rate (2θ) of 4° min⁻¹ and a scan range from 5 to 80°.

2.5. Electrochemical measurements

The electrochemical measurements were carried out on a CS350 electrochemical workstation (Wuhan CorrTest Instruments Co., Ltd., China) at room temperature in a three-electrode setup. The MnO₂/WDB composite (or WDB) directly served as the working electrode, and the area exposed to the electrolyte was about 0.5 cm². The mass of the exposed electrodes was around 2.48 mg for MnO₂/WDB and 1.88 mg for WDB, respectively. An Ag/AgCl electrode and a Pt wire electrode served as reference and counter electrodes, respectively. The electrolyte was 1 M Na₂SO₄ solution. CV curves were measured over the potential window from 0 to 0.8 V at different scan rates of 5, 10, 20 and 50 mV s⁻¹. GCD curves were measured in the potential range of 0–0.8 V at different current densities of 0.05, 0.1, 0.2, 0.5, 1, 2, 5 and 10 A g⁻¹. EIS measurements were carried out in the frequency range from 10⁵ to 0.01 Hz with alternate current amplitude of 5 mV.

3. Results and discussion

3.1. Morphology observations and elemental analysis

Morphological characteristics of WDB before and after the deposition of MnO₂ were studied by SEM observations. As shown in Fig. 2a, WDB still maintained the feature structures of pristine wood after the pyrolysis treatment, and a porous structure with several pits whose diameters range from 2 to 4 μm can be observed. After WDB was dipped in the aqueous solution of KMnO₄, a redox reaction took place between KMnO₄ and WDB, which can be expressed as follows:³¹

4MnO₄⁻ + 3C + H₂O → 4MnO₂ + CO₃²⁻ + 2HCO₃⁻ (1)

Fig. 2 (a) SEM image of WDB. (b) Low- and (c) high-magnification SEM images of MnO₂/WDB. (d) Cross-section SEM image of MnO₂/WDB. (e) Magnified image of the green-squared marker region in (d), and the inset is detail with enlarged scale of green marker in (e). (f) EDX patterns of WDB and MnO₂/WDB, and the insets show the corresponding elemental maps. (g) TEM and (h) HRTEM images of MnO₂/WDB, and the inset in (h) presents the corresponding SAED pattern.

This reaction results in the formation of MnO₂ on WDB, as indicated in Fig. 2b. It is clear that the initial smooth surface of WDB substrate was covered with an ocean of nano-MnO₂. According to a higher-magnification SEM image (160 000×) in Fig. 2c, the porous and cross-linked MnO₂ nanosheets are only a few nanometers in thickness and have numerous wrinkles and ripples. Such porous and ultrathin nanosheets can effectively improve the surface/interface area of MnO₂ nanocrystals and are expected to facilitate electrolyte diffusion among interspaces of MnO₂ nanosheets.¹⁷ Fig. 2d shows the cross-section SEM image of MnO₂/WDB. The 3D honeycomb porous structure of the tracheid cell wall, which belongs to the structural characteristics of wood, can be clearly identified. The details on the surface features of MnO₂/WDB (the green-squared marker region in Fig. 2d) are presented in Fig. 2e. Similar to the results of Fig. 2b, the surface of WDB was encapsulated with dense MnO₂ layers (inset in Fig. 2e), which suggests that the material belongs to core–shell composite, i.e., the biochar serves as the core part, and the MnO₂ layers act as the shell part.

EDX analysis was used to obtain additional information about the elemental compositions. For WDB, only C, O and Au elements were detected (Fig. 2f). The Au element originates from the coating layer used for electric conduction during SEM observation. The low oxygen content of 5.3 wt% is attributed to the pyrolysis treatment that damaged the oxygen-containing groups. In contrast to WDB, apart from the common C, O and Au elements, the Mn and K elements were also detected for MnO₂/WDB. The strong Mn signals are assigned to the sheet-like nano-MnO₂ and indicate a high content (82.4 wt%) on the surface of MnO₂/WDB. The K signal might be derived from the KMnO₄ since there is always a possibility of K⁺ co-existing in the MnO₂.^30,32 In addition, the increased mass ratio of oxygen to carbon from 0.06 (WDB) to 0.76 (MnO₂/WDB) is associated with the generated core−shell structure, i.e., WDB coated with abundant MnO₂ nanosheets. The TEM image of MnO₂/WDB is presented in Fig. 2g, in which the wrinkled MnO₂ nanosheets are supported on WDB and no conspicuous agglomeration appeared. The HRTEM image of MnO₂/WDB reveals the presence of birnessite-type MnO₂ as shown in Fig. 2h; i.e., lattice fringes with spacings of around 0.257 and 0.213 nm agree well with the (201̄) and (112̄) lattice spacings. The corresponding SAED pattern exhibits a polycrystalline feature (inset in Fig. 2h) owing to the random orientation of different MnO₂ nanocrystals. Furthermore, the diffraction rings can be indexed to the (201), (201̄), and (203) planes of birnessite-type MnO₂.

3.2. Crystal structure and chemical compositions

The crystal structures of WDB and MnO₂/WDB were characterized by XRD analysis. As shown in Fig. 3, WDB shows two broad diffraction peaks centered at 22.0° and 43.9°, which are ascribed to the (002) and (100) planes of amorphous carbon originated from the pyrolysis of wood.³³ For MnO₂/WDB, in addition to the peaks from WDB substrate, several new peaks located at around 12.4°, 36.4° and 65.6° can be observed, which correspond to the (001), (110) and (312̄) planes of birnessite-type MnO₂ crystalline phase (JCPDS no. 42-1317). The results are in accordance with the HRTEM and SAED analysis.

Fig. 3 XRD patterns of WDB and MnO₂/WDB, respectively. The bottom line is the standard JCPDS card no. 42-1317 for birnessite-type MnO₂.

XPS analysis was carried out to investigate the oxidation state of manganese and the chemical bonding states of MnO₂/WDB. Fig. 4a shows the XPS survey spectrum of MnO₂/WDB. The XPS signals from elements C, O and Mn can be seen, agreeing well with the results of EDX analysis. The high-resolution Mn 2p spectrum is presented in Fig. 4b, where two strong peaks at 642.7 and 654.4 eV can be clearly identified, corresponding to the binding energy of Mn 2p_3/2 and Mn 2p_1/2,³⁴ respectively. Moreover, the spinning energy separation of 11.7 eV between the Mn 2p_3/2 and Mn 2p_1/2 peaks is consistent with previously reported data in the spectrum of MnO₂,^30,35 suggesting that the predominant oxidation state is +4. The high-resolution O 1s spectrum displays two peaks at 530.2 and 532.1 eV (Fig. 4c), which are assigned to Mn–O–Mn and Mn–O–H,³⁶ respectively.

Fig. 4 XPS spectra of MnO₂/WDB: (a) wide scan survey spectrum; the core-level XPS signals of (b) Mn 2p and (c) O 1s.

3.3. Electrochemical properties

The electrochemical behaviors of WDB and MnO₂/WDB electrodes were investigated by CV, GCD, and EIS in a three-electrode system with 1 M Na₂SO₄ as the aqueous electrolyte. CV response of both WDB and MnO₂/WDB electrodes carried out at the scan rate of 10 mV s⁻¹ in the potential range of 0–0.8 V is shown in Fig. 5a. Obviously, the current density and the integration area of CV curve obtained from MnO₂/WDB electrode are much larger than those of WDB, indicating that the capacitance of the hybrid electrode is primarily from the pseudocapacitance of MnO₂ nanosheets. It is observed in Fig. 5b that the CV curves of WDB exhibit nearly symmetrical rectangular shape at scan rates of 5, 10, 20 and 50 mV s⁻¹, indicating the low equivalent series resistance and the fast diffusion of electrolyte ions into the electrode.³⁷ WDB electrode (carbon as its main component) stores charges through electric double-layer capacitance, and thus expresses in the form of a rectangular shape of the voltammetry characteristics.³⁸ For MnO₂/WDB electrode (see Fig. 5c), a pair of weak redox peaks can be observed in the CV curves due to the reaction (MnO₂ + Na⁺ + e⁻ = MnOONa) of MnO₂ in the hybrid electrode. These CV curves are nearly rectangular, reflecting the good capacitive performance and excellent reversibility of MnO₂/WDB electrode. However, at the high scan rates (e.g., 50 mV s⁻¹), the CV profile deviates from the rectangular shape, which might be attributed to the polarization of the desolvation process of the hydrated sodium ions and the relatively low conductivity of MnO₂ in the hybrid electrode.³⁹

Fig. 5 (a) CV curves of WDB and MnO₂/WDB electrodes at the scan rate of 10 mV s⁻¹. CV curves of (b) WDB and (c) MnO₂/WDB electrodes at various scan rates.

Fig. 6a and b show the charge–discharge characteristics of WDB and MnO₂/WDB electrodes between 0 and 0.8 V at various current densities of 0.05–10 A g⁻¹. During the charging and discharging steps, the charge curves of MnO₂/WDB electrode are almost symmetric to their corresponding discharge counterparts, revealing a good capacitive behavior and a highly reversible faradic reaction between Na⁺ and MnO₂, i.e., MnO₂ + Na⁺ + e⁻ ↔ MnOONa.^7,40 In addition, comparing the discharge curves of WDB and MnO₂/WDB electrodes, MnO₂/WDB electrode shows much longer discharge time, which is consistent with specific capacitance behavior since discharge time is directly proportional to the specific capacitance of electrode. The specific capacitance (C_m) was calculated from the GCD curves according to the following equation:⁴¹

where C_m is the specific capacitance (F g⁻¹), I is the charge–discharge current (A), Δt is the discharge time (s), ΔV is the potential window (V), and m is the electrode mass (g). The specific capacitance was plotted versus discharge current in Fig. 6c. Because of the deposition of MnO₂, MnO₂/WDB electrode has the highest specific capacitance of 101 F g⁻¹ at a current density of 0.05 A g⁻¹, which is about five times of that of WDB (19 F g⁻¹). The enhancement of capacitance is ascribed to the pseudocapacitance redox reactions of MnO₂. Nevertheless, the specific capacitance of 101 F g⁻¹ in this work is relatively lower, as compared with that of some recently reported composites consisting of MnO₂ and various carbon-based materials, such as carbon fiber cloth (684 F g⁻¹),⁴² graphene (328 F g⁻¹),⁴³ and carbon papers (307 F g⁻¹).³⁰ The specific capacitance of MnO₂/WDB electrode is expected to be further improved by the following modifications:

Fig. 6 GCD curves of (a) WDB and (b) MnO₂/WDB electrodes at various current densities. The insets show the enlarged images of the GCD curves at the high current densities. (c) Specific capacitance of WDB and MnO₂/WDB electrodes as a function of discharge current.

(1) A special oxidation of carbon for increasing the surface functionality (through chemical treatment,⁴⁴ electrochemical polarization,⁴⁵ plasma treatment⁴⁶).

(2) Integrating the MnO₂/WDB electrode with some other types of electrochemical active substances (e.g., conducting polymers).^5,47,48

(3) Further inserting electroactive particles of transition metals oxides (e.g., RuO₂, TiO₂, Cr₂O₃, and Co₃O₄) into the carbon material (namely WDB).⁶.

(4) Minimizing the thickness of WDB substrate for reducing the weight of hybrid electrode.

These above modifications will be deeply studied in our future researches. Both electrodes show gradually decreased capacitance with the increase of current density. In contrast to WDB electrode (storing charges through electric double-layer capacitance), MnO₂/WDB electrode shows a more remarkable decrease of specific capacitance as the current density increased. The reduction is believed to be ascribed to the discrepant insertion–deinsertion behaviors of Na⁺ from the electrolyte to MnO₂.^12,27 At a low current density, the diffusion of ions from the electrolyte can gain access to almost all available pores of the electrode; nevertheless, with the increment of current density, the effective interaction between the ions and the electrode was significantly reduced, and thus a remarkable reduction in capacitance appeared. However, MnO₂/WDB electrode still retained 43% (a decrease from 101 F g⁻¹ to 43 F g⁻¹) of its capacitance as the current density increased from 0.05 to 10 A g⁻¹. In addition, the capacitance value is about three times higher than 10 F g⁻¹ of WDB electrode at 10 A g⁻¹. These results confirm the superior capacitive characteristics of MnO₂/WDB electrode.

EIS tests were carried out in a frequency range from 10⁵ to 0.01 Hz to further evaluate the electrochemical behaviors of WDB and MnO₂/WDB electrodes. The EIS data were analyzed using Nyquist plots (see Fig. 7). The Nyquist plots can be divided into three parts: (1) a semicircle in the high-to-medium frequencies region indicating the intersection point at the real impedance axis (Z′) representing the bulk solution resistance (R_s) and the diameter of which represents the charge transport resistance (R_ct); (2) a straight line with a slope of 45° in the medium-frequency range, corresponding to semi-infinite Warburg impedance due to the frequency dependence of ion diffusion/transport in the electrolyte; and (3) a vertical line at the low frequencies, indicating the pure capacitive behavior.^35,49 In the high-frequency region (inset in Fig. 7), the both electrodes have almost identical R_s of ∼2.0 Ω. However, the R_ct value of MnO₂/WDB electrode is obviously higher than that of WDB, due to the poor electrical conductivity of MnO₂ in the hybrid electrode. Such difference in R_ct is consistent with the prediction results of CV analysis. In the medium-frequency region, the projected length of Warburg curve on the Z′ axis characterizes the ion penetration process.⁵⁰ The Warburg length of MnO₂/WDB electrode is shorter than that of WDB, which indicates that the hybrid electrode has a shorter ion-diffusion path. This is possibly because the compact covering layers of MnO₂ nanosheets increases the electrochemically active surface area accessed by electrolyte, thus promoting the electrolyte ions adsorption and shortening the migration pathways of ions during the charge–discharge processes.

Fig. 7 Nyquist plots of WDB and MnO₂/WDB electrodes.

The long–term cycle performances of WDB and MnO₂/WDB electrodes in a potential window of 0–0.8 V were measured by the consecutive GCD tests at the current density of 5 A g⁻¹. In Fig. 8a, WDB electrode demonstrates good long-term cycle stability, retaining about 95.4% of its initial capacitance after 10 000 cycles. Also, the coulombic efficiency calculated from the ratio between charge capacitance and discharge capacitance remained a quite high value of 98%–100% during the whole cycle. In contrast, MnO₂/WDB electrode shows slightly decreased capacitance retention of 85.0% after 10 000 cycles, but still maintained a high coulombic efficiency of more than 98% (Fig. 8b). The differences in electrochemical stability between WDB and MnO₂/WDB electrodes are possibly attributed to the different double-layer and pseudocapacitive contributions.¹⁰ It is well-known that the double-layer capacitance only involves a charge rearrangement, while the pseudocapacitance is related to a chemical reaction. The double-layer capacitors generally have better electrochemical stability but lower specific capacitance as compared with those of pseudocapacitators.⁵¹ In this work, WDB electrode (carbon as its main component), making more double-layer contribution compared to that of the hybrid electrode, has a significantly lower capacitance than the latter; however, its cycle stability was enhanced accordingly. For MnO₂/WDB electrode, the faster decrement in capacitance retention is attributed to the dissolution of active materials and mechanical faults caused by the ion insertion/extraction to the MnO₂ nanosheets.^10,52

Fig. 8 Cycle performance and coulombic efficiency of (a) WDB and (b) MnO₂/WDB electrodes measured at the current density of 5 A g⁻¹.

4. Conclusions

Wood-derived biochar was employed as a novel eco-friendly substrate to support MnO₂ nanosheets via a simple cheap in situ redox reaction between the biochar and KMnO₄. This composite with a core−shell structure can serve as a free-standing and binder-free supercapacitor electrode, which shows a moderate specific capacitance of 101 F g⁻¹ at a current density of 0.05 A g⁻¹, an excellent coulombic efficiency of 98–100%, and a favorable cyclic stability with a capacitance retention of 85.0% after 10 000 cycles. The specific capacitance is expected to be improved by integrating the electrode with some new electrochemical active substances (e.g., conducting polymers), oxidation of carbon for increasing the surface functionality, or minimizing the thickness of WDB substrate, which will be further studied in our future researches. Perhaps wood-derived biochar is not the most suitable substrate material, but we try to give a new potential candidate for the development of high-performance energy storage devices.

Acknowledgements

This study was supported by the National Natural Science Foundation of China (grant no. 31270590 and 31470584) and the Fundamental Research Funds for the Central Universities (grant no. 2572016AB22).

References

1. L. L. Zhang and X. Zhao, Chem. Soc. Rev., 2009, 38, 2520–2531.
2. G. Wang, L. Zhang and J. Zhang, Chem. Soc. Rev., 2012, 41, 797–828.
3. S. He and W. Chen, Nanoscale, 2015, 7, 6957–6990.
4. S. He, H. Hou and W. Chen, J. Power Sources, 2015, 280, 678–686.
5. G. A. Snook, P. Kao and A. S. Best, J. Power Sources, 2011, 196, 1–12.
6. M. Zhi, C. Xiang, J. Li, M. Li and N. Wu, Nanoscale, 2013, 5, 72–88.
7. M. Toupin, T. Brousse and D. Bélanger, Chem. Mater., 2004, 16, 3184–3190.
8. M. Thackeray, M. Rossouw, A. De Kock, A. De la Harpe, R. Gummow, K. Pearce and D. Liles, J. Power Sources, 1993, 43, 289–300.
9. Y. Peng, Z. Chen, J. Wen, Q. Xiao, D. Weng, S. He, H. Geng and Y. Lu, Nano Res., 2011, 4, 216–225.
10. Y. Hou, Y. Cheng, T. Hobson and J. Liu, Nano Lett., 2010, 10, 2727–2733.
11. S. W. Lee, J. Kim, S. Chen, P. T. Hammond and Y. Shao-Horn, ACS Nano, 2010, 4, 3889–3896.
12. S. Chen, J. Zhu, X. Wu, Q. Han and X. Wang, ACS Nano, 2010, 4, 2822–2830.
13. S. He, R. Zhang, C. Zhang, M. Liu, X. Gao, J. Ju, L. Li and W. Chen, J. Power Sources, 2015, 299, 408–416.
14. J. Liu, J. Essner and J. Li, Chem. Mater., 2010, 22, 5022–5030.
15. W. Wang, S. Guo, K. N. Bozhilov, D. Yan, M. Ozkan and C. S. Ozkan, Small, 2013, 9, 3714–3721.
16. V. Khomenko, E. Raymundo-Pinero and F. Béguin, J. Power Sources, 2006, 153, 183–190.
17. S. He and W. Chen, J. Power Sources, 2014, 262, 391–400.
18. Q. Fang, B. Chen, Y. Lin and Y. Guan, Environ. Sci. Technol., 2013, 48, 279–288.
19. E. Taer, M. Deraman, I. A. Talib, A. A. Umar, M. Oyama and R. M. Yunus, Curr. Appl. Phys., 2010, 10, 1071–1075.
20. Y.-Y. Horng, Y.-C. Lu, Y.-K. Hsu, C.-C. Chen, L.-C. Chen and K.-H. Chen, J. Power Sources, 2010, 195, 4418–4422.
21. J. Jiang, L. Zhang, X. Wang, N. Holm, K. Rajagopalan, F. Chen and S. Ma, Electrochim. Acta, 2013, 113, 481–489.
22. S. Teng, G. Siegel, M. C. Prestgard, W. Wang and A. Tiwari, Electrochim. Acta, 2015, 161, 343–350.
23. S. Yu, D. Liu, S. Zhao, B. Bao, C. Jin, W. Huang, H. Chen and Z. Shen, RSC Adv., 2015, 5, 30943–30949.
24. E. Taer, M. Deraman, I. A. Talib, A. Awitdrus, S. Hashmi and A. Umar, Int. J. Electrochem. Sci., 2011, 6, 3301–3315.
25. S. H. Kim, S. J. Kim and S. M. Oh, Chem. Mater., 1999, 11, 557–563.
26. S. Chou, F. Cheng and J. Chen, J. Power Sources, 2006, 162, 727–734.
27. V. Subramanian, H. Zhu, R. Vajtai, P. Ajayan and B. Wei, J. Phys. Chem. B, 2005, 109, 20207–20214.
28. X. Wang, X. Wang, W. Huang, P. Sebastian and S. Gamboa, J. Power Sources, 2005, 140, 211–215.
29. W. Li, L. Zeng, Z. Yang, L. Gu, J. Wang, X. Liu, J. Cheng and Y. Yu, Nanoscale, 2014, 6, 693–698.
30. S. He, C. Hu, H. Hou and W. Chen, J. Power Sources, 2014, 246, 754–761.
31. X. Zhao, L. Zhang, S. Murali, M. D. Stoller, Q. Zhang, Y. Zhu and R. S. Ruoff, ACS Nano, 2012, 6, 5404–5412.
32. P. Ragupathy, D. H. Park, G. Campet, H. Vasan, S.-J. Hwang, J.-H. Choy and N. Munichandraiah, J. Phys. Chem. C, 2009, 113, 6303–6309.
33. E. Sajitha, V. Prasad, S. Subramanyam, S. Eto, K. Takai and T. Enoki, Carbon, 2004, 42, 2815–2820.
34. X. Zhao, Y. Du, Y. Li and Q. Zhang, Ceram. Int., 2015, 41, 7402–7410.
35. M. Kim, Y. Hwang and J. Kim, J. Power Sources, 2013, 239, 225–233.
36. Z. Lei, F. Shi and L. Lu, ACS Appl. Mater. Interfaces, 2012, 4, 1058–1064.
37. J. Yan, T. Wei, B. Shao, F. Ma, Z. Fan, M. Zhang, C. Zheng, Y. Shang, W. Qian and F. Wei, Carbon, 2010, 48, 1731–1737.
38. E. Frackowiak and F. Beguin, Carbon, 2001, 39, 937–950.
39. H. Li, Y. He, V. Pavlinek, Q. Cheng, P. Saha and C. Li, J. Mater. Chem. A, 2015, 3, 17165–17171.
40. D. Bélanger, L. Brousse and J. W. Long, Electrochem. Soc. Interface, 2008, 17, 49.
41. Z. Lei, J. Zhang and X. Zhao, J. Mater. Chem., 2012, 22, 153–160.
42. S. He and W. Chen, J. Power Sources, 2015, 294, 150–158.
43. M. Kim, Y. Hwang and J. Kim, J. Power Sources, 2013, 239, 225–233.
44. K. Jurewicz and E. Frackowiak, Mol. Phys. Rep., 2000, 27, 38–45.
45. T. Momma, X. Liu, T. Osaka, Y. Ushio and Y. Sawada, J. Power Sources, 1996, 60, 249–253.
46. K. Okajima, K. Ohta and M. Sudoh, Electrochim. Acta, 2005, 50, 2227–2231.
47. Z. Spitalsky, D. Tasis, K. Papagelis and C. Galiotis, Prog. Polym. Sci., 2010, 35, 357–401.
48. C. Peng, J. Jin and G. Z. Chen, Electrochim. Acta, 2007, 53, 525–537.
49. L. Sun, C. Tian, M. Li, X. Meng, L. Wang, R. Wang, J. Yin and H. Fu, J. Mater. Chem. A, 2013, 1, 6462–6470.
50. Z. Weng, Y. Su, D. W. Wang, F. Li, J. Du and H. M. Cheng, Adv. Energy Mater., 2011, 1, 917–922.
51. M. Winter and R. J. Brodd, Chem. Rev., 2004, 104, 4245–4270.
52. H. B. Wu, J. S. Chen, H. H. Hng and X. W. D. Lou, Nanoscale, 2012, 4, 2526–2542.

---

Footnote

† C. W. and Y. J. contributed equally to the paper.

---