Fast lithium-ion storage of Nb₂O₅ nanocrystals in situ grown on carbon nanotubes for high-performance asymmetric supercapacitors

Abstract

Lithium-ion supercapacitors have been attracting tremendous research interest due to significantly increased energy density. However, the power density and cycling stability are often sacrificed by poor conductivity and sluggish kinetics of transition metal oxide-based electrodes. Herein, we report a composite material based on Nb₂O₅ nanocrystals in situ grown on carbon nanotubes (CNTs) with a unique architecture exhibiting superior rate capability and long cycling stability. Even at 100 C-rate with charge/discharge time of 12 s, the composite still possesses a capacitance of ∼160 C g⁻¹. A prototype of an asymmetric lithium-ion supercapacitor composed of an activated carbon cathode and a CNT–Nb₂O₅ composite anode shows significantly improved energy density of ∼50 W h kg⁻¹ with a power of 86.46 W kg⁻¹. Such composite materials with unique architecture hold great promise for future practical applications.

---

1 Introduction

The increasing concerns over the rapid depletion of fossil oil and sustainability have attracted a tremendous amount of research interest to energy storage systems such as lithium-ion batteries (LIBs) and supercapacitors.^1–4 Compared with LIBs, supercapacitors have a number of properties including much higher power and long cycling life, which can complement or replace batteries.^5–8 However, typical electrochemical double-layer capacitors based on porous carbon materials with high surface area possess a capacitance of around 150 F g⁻¹, by storing charge at the interface of electrode and electrolyte.^9,10 The resulting energy densities are generally a thousand times lower than those of LIBs. By comparison, pseudocapacitors based on inorganic non-carbon based materials store much higher charge (>1000 F g⁻¹) through reversible faradaic reactions.¹¹ Thus, seeking novel pseudocapacitive materials with high capacitance is one of the key research themes in this field.

So far, transition metal oxides have been extensively studied as a class of pseudocapacitive materials due to their different and high oxidation states, such as RuO₂,^12,13 MnO₂,^14,15 V₂O₅ (ref. 16) and Li₄Ti₅O₁₂.^17,18 Among various candidates, niobium pentoxide (Nb₂O₅) have been attracting tremendous research attention owing to its specific capacity of ∼200 mA h g⁻¹ and a comparatively low average operating potential (1.0–1.5 V vs. Li/Li⁺).^19–21 Recently, it is reported that orthorhombic Nb₂O₅ exhibits an intercalation pseudocapacitance so that the charge storage is not limited by solid-state diffusion.^22,23 However, similar to most transition metal oxides, Nb₂O₅ is known to suffer from intrinsically poor conductivity (∼3 × 10⁻⁶ S cm⁻¹ at room temperature),²⁴ which leads to electrodes with inferior power.

So far, many strategies are applied to overcome this barrier.^25–29 For example, utilizing a large amount of conductive agent (e.g. carbon black) can significantly increase electrode rate capability but sacrifice the overall capacitance. Although nanostructured materials such as mesoporous Nb₂O₅ thin film endow the electrodes with impressive rate performance,^25,30 the low mass loading of the active material often leads to inferior energy density. It is well accepted that efficiently reducing ions diffusion and electrons migration paths and enhancing electronic conductivity are the key factors to achieving very high power rates.^31–33 Therefore, developing new materials with nanosized electroactive materials and conductive networks is highly demanded to improve the slow kinetics of Nb₂O₅-based electrodes, which is still a barrier to its practical application.

Recently, a new class of pseudocapacitive nanocomposite materials was developed for asymmetric supercapacitors by physically mixing Nb₂O₅ nanocrystals and carbon nanotubes (CNTs) followed by post-heat-treatment.³⁴ In this work, we report a facile solvo-thermal synthesis of CNT–Nb₂O₅ nanocomposites with superior rate capability, as well as long term cycling stability. As shown in Scheme 1, Nb₂O₅ nanocrystals (NCs) in situ grow on CNTs scaffold, forming intimate contact between the two components, while the further assembling results in a robust marimo-ball structure. This unique architecture offers several major features:^35,36 (i) nanocrystals with low dimensions efficiently shortens lithium-ion diffusion length, leading to better utilization of Nb₂O₅; (ii) interconnected pores created by the CNTs network enables both effective electrolyte transport and fast active material accessibility; (iii) the in situ growth ensures the intimate contacts and robust structure, endowing the electrode with high power density and cycling stability. High-performance prototype of asymmetric supercapacitor with the composite material-based anode and activated carbon-based cathode is fabricated, exhibiting both high energy and power density.

Scheme 1 Schematic of in situ growing Nb₂O₅ nanocrystals on CNTs scaffold after post sintering process.

2 Experimental

2.1 Synthesis of CNT–Nb₂O₅ nanocomposites

Multi-wall carbon nanotubes were treated with acid by creating functional groups (e.g. carboxylic groups) through a similar method previously reported³⁷ A facile solvo-thermal synthesis was applied to produce CNT–Nb₂O₅ composites. Briefly, functionalized CNTs (75 mg) were homogeneously dispersed in ethanol (10 mL) by ultrasonicating for 30 min followed by stirring overnight. Niobium pentachloride (0.325 g) was dissolved in ethanol (4 mL) before mixed with above CNT dispersion under stirring for 30 min. The mixture was delivered to a 25 mL Teflon-lined autoclave and cooked for 12 h at 200 °C. The resulted dark precipitates were collected by centrifuge, and further rinsed by ethanol and dried at 80 °C for 12 h, and then sintered at 600 °C for 4 h under N₂.

2.2 Material characterization

X-ray diffraction (XRD) measurements were conducted on an Inel XRG 3000 X-ray diffractometer (monochromatic Cu Kα X-rays). Nitrogen adsorption/desorption isotherms were collected on Micromeritics ASAP 2020 analyzer at 77 K. The Brunauer–Emmett–Teller (BET) method is applied to calculated the surface areas, while the Barrett–Joyner–Halenda (BJH) model is applied to determine the pore size distributions. Scanning electron microscopic (SEM) experiments were performed on a LEO FESEM-1530; while transmission electron microscopic (TEM), high-resolution TEM and EDX mapping images were obtained from a JEOL 2010F TEM/STEM field emission microscope in Canadian Center for Electron Microscopy (CCEM, McMaster University, Canada). X-ray photoelectron spectroscopy (XPS) data was collected from an AXIS Ultra DLD (Kratos Analytical Co. Ltd.).

2.3 Electrode fabrication

To prepare the slurry for electrode fabrication, the CNT–Nb₂O₅ nanocomposite powder, carbon black and polyvinylidenedifluoride (PVDF) binder were homogeneously dispersed in N-methylpyrrolidinone (NMP) in a mass ratio of 86 : 4 : 10. The homogenous slurry was drop-casted on copper foil current collector and dried at 100 °C for 2 h. The electrodes were pressed at 2.0 MPa before further dried at 100 °C overnight under vacuum. The mass loading of active materials was adjusted between 1.5 to 2.0 mg cm⁻² on each electrode.

2.4 Electrochemical measurements

The electrodes were assembled into 2032-type coin cells in an argon-filled glovebox for electrochemical testings. In each coin cell, a piece of glass microfiber filter (Whatman, Grade GF/C) and a piece of lithium foil were used as separator and counter electrode, respectively, and 1 M LiPF₆ in a 1 : 1 (v/v) mixture of ethylene carbonate (EC) and dimethyl carbonate (DMC) was used as electrolyte. Cyclic voltammetric (CV) curves were collected on a VSP3 potentiostat/galvanostat (Bio-Logic LLC, Knoxville, TN) using cutoff voltages of 3.0 and 1.0 V (vs. Li/Li⁺) at different scan rates. The galvanostatic charge/discharge measurements were performed on NEWARE BTS-CT3008 (Shenzhen Neware Technology, Co. Ltd., Shenzhen, China) at different current densities. The total charge or the specific capacitance of the composite electrodes is calculated based on the total mass of the composite materials. Electrochemical impedance spectroscopy measurement was conducted on a Princeton Applied Rsearch VersaSTAT MC potentiostat. The Nyquist plots were recorded potentiostatically by applying an AC voltage of 10 mV amplitude in the frequency range of 0.01 to 100 kHz. All electrochemical measurements were carried out at room temperature.

2.5 Prototype asymmetric supercapacitor fabrication and testing

The asymmetric device was assembled using CNT–Nb₂O₅ nanocomposite based anode and activated carbon (AC) based cathode. The AC-based cathodes were fabricated by mixing AC, PVDF binder, and carbon black in the mass ratio of 8 : 1 : 1 in NMP before coated on aluminium foil. The ratio of CNT–Nb₂O₅ nanocomposite and activated carbon was controlled to be around 1 : 3.5 by mass. Asymmetric device was cycled at constant current between 3.0 and 0.5 V. The power density and energy density were calculated according to the total mass of anode and cathode materials, and a packaging factor of 0.4 was also applied. The energy density was determined by the equation E = 1/2CV², where C is the total capacitance and V is the maximum cell voltage, and power density was calculated using P = E/t, where P is the power density, t is the discharging time.

3 Results and discussion

The structure of CNT–Nb₂O₅ nanocomposites can be revealed by scanning electron microscopic (SEM) and transmission electron microscopic (TEM). Fig. 1A presents a representative SEM image of the as-synthesized CNT–Nb₂O₅ composites. Spherical morphology with marimo-like structure can be observed for most of the composite materials with sizes ranging from hundreds of nanometers to several microns. These micrometer-sized spheres facilitate the current electrode fabrication based on conventional slurry-coating process, which is compatible with industrial battery manufacturing. High-magnification SEM image in Fig. 1B clearly reveals the detailed structure of the nanocomposites, where CNTs build a network spreading through the whole composites. Fig. 1C presents a representative TEM image of a single composite sphere with the size of around 2 microns. The Nb₂O₅ NCs and CNTs are firmly combined, where no obvious separated and isolated NCs or CNTs can be observed, indicating a strong bonding between the two components. Moreover, the composite sphere also exhibits a hierarchically porous structure created from the entangled CNTs networks. The successful in situ growth of Nb₂O₅ nanocrystals on CNTs was verified by high-magnification and high-resolution TEM in Fig. 1D and E. Nb₂O₅ nanocrystals with average size of ∼10 nm grow on CNTs directly and uniformly (Fig. S-1 in the ESI†), with a lattice parameter of 0.39 nm, corresponding to the (001) plane of orthorhombic Nb₂O₅. Fig. 1F shows EDX mapping images of the composite material. The well dispersed element Nb and O further confirms the uniform coating of Nb₂O₅ nanocrystals on the surface of CNTs and the intimate contact. Therefore, the low dimension of active Nb₂O₅ nanocrystals can efficiently shorten the ion-diffusion length, while the continuous and conductive scaffold and the intimate contacts between the two components can facilitate electron transportation. Therefore, such a unique architecture is highly desired for the electrodes with excellent lithium-storage ability.

Fig. 1 Representative (A) low- and (B) high-magnification SEM images of as-synthesized CNT–Nb₂O₅ nanocomposites with spherical shape; (C) TEM image of a single sphere of CNT–Nb₂O₅ nanocomposites; and (D) TEM image of part of a single CNT with in situ grown Nb₂O₅ nanocrystals; (E) high-resolution TEM image of part of the composite; and (F) EDX mapping of elements C, O and Nb.

The porous structure of the materials is characterized by analyzing the nitrogen adsorption–desorption isotherms (Fig. 2A). A Brunauer–Emmett–Teller (BET) surface area was calculated to be 122 m² g⁻¹ for CNT–Nb₂O₅ composites, while the pore size distribution is concentrated below 10 nm. In contrast, without the CNTs scaffold, pure Nb₂O₅ synthesized at the same condition gives a BET surface area of only 51 m² g⁻¹, which is close to that of the bulk material with average size of microns. Such a phenomenon suggests that the growth or aggregation of NCs can be effectively prevented by the existence of CNTs, which holds great importance in preserving the short ion-transport length. Furthermore, the porous feature and the high surface area endow the electrodes with efficient ion-transport and abundant active sites, respectively. Fig. 2B compares the X-ray diffraction (XRD) patterns of CNT–Nb₂O₅ composite before and after sintering. After sintering process, the Nb₂O₅ NCs in CNT–Nb₂O₅ composites possess an orthorhombic crystal structure (JCPDS card no. 27-1003),^38,39 which is the same as the pure Nb₂O₅ crystals synthesized in identical conditions without using CNTs (Fig. S-2 in the ESI†). The average domain size of Nb₂O₅ nanocrystals in the composites is around 9.82 nm calculated according to Scherrer's equation, which aligns well with the value from TEM observation. Fig. 2C shows the X-ray photoelectron spectroscopic (XPS) results of CNT–Nb₂O₅ composite. Two well-defined peaks in the core level spectra at 209.7 and 206.9 eV respectively with an energy separation of 2.8 eV correspond to the Nb 3d of Nb₂O₅, which is consistent with previous reports, thus further confirms the formation of Nb₂O₅.^40–42 Fig. 2D presents the thermogravimetric analysis (TGA) results of pure CNTs and CNT–Nb₂O₅ composites. At temperature below around 500 °C, the weight loss of CNTs can be negligible, while a steep weight loss can be observed between 550 to 700 °C for both pure Nb₂O₅ and the composites due to the consumption of CNT. The composites consist of 77 wt% of Nb₂O₅ and 24 wt% of CNTs based on the TGA results. The mass ratio of Nb₂O₅ and CNTs can be easily adjusted to synthesize a class of composites with tunable properties. In this paper, we chose the composite material with 77 wt% of Nb₂O₅ and 23 wt% of CNTs to demonstrate the outstanding performance of in situ growth of Nb₂O₅ on CNTs composites.

Fig. 2 (A) Nitrogen adsorption–desorption isotherms of CNT–Nb₂O₅ nanocomposites and pure Nb₂O₅ (inset: pore-size distribution of CNT–Nb₂O₅ nanocomposites and pure Nb₂O₅); (B) XRD patterns of CNT–Nb₂O₅ nanocomposites before and after sintering process; (C) XPS core-level spectra of Nb 3d_3/2 and Nb 3d_1/2 peaks (inset: XPS survey spectrum of CNT–Nb₂O₅ nanocomposites); (D) TGA curves of CNTs and CNT–Nb₂O₅ nanocomposites.

The electrochemical property of the CNT–Nb₂O₅ composites was analyzed by cyclic voltammetry (CV) using coin cells. Generally, the equation Nb₂O₅ + xLi⁺ + xe⁻ ↔ Li_xNb₂O₅ is used to describe the Li⁺ electrochemically intercalating and de-intercalating into and from Nb₂O₅, where x (0 < x < 2) represents the mole fraction of the intercalated Li⁺. The typical cyclic voltammogram curves of the CNT–Nb₂O₅ composite at a scan rate of 0.5 mV s⁻¹ is presented in Fig. 3A. A couple of symmetric redox peaks can be found at ∼1.5 V, which is consistent with Nb₂O₅-based materials,^21,43 corresponding to the redox conversion between Nb⁵⁺ and Nb⁴⁺ by two Li⁺. Moreover, the composite exhibits broader peaks, implying electrochemical reactions with broader distribution of energy,²⁴ which is commonly observed in nanostructured materials. The total charge stored on the electrode can be calculated from the area within the CV curve. Charge storage of CNT–Nb₂O₅ composite was compared with pure Nb₂O₅ synthesized at identical condition at scan rate ranging from 1 to 50 mV s⁻¹ (corresponding to charge/discharge times from 2000 to 40 s, respectively) (Fig. S-3 in the ESI†). As shown in Fig. 3B, obviously, the CNTs/Nb₂O₅ composite can store much more charge than the pure Nb₂O₅. At the same scan rate of 50 mV s⁻¹, the pure Nb₂O₅ only stores 31% (125 C g⁻¹, 62.5 F g⁻¹) of the total charge obtained at the scan rate of 1 mV s⁻¹, while the CNTs–Nb₂O₅ composite stores substantially more charge of 305 C g⁻¹ (152.5 F g⁻¹), which is ∼55% of the value at 1 mV s⁻¹.

Fig. 3 (A) Cyclic voltammograms of the CNT–Nb₂O₅ composites electrode at a scan rate of 0.5 mV s⁻¹ with a cutting off voltage of 3.0 to 1.0 V vs. Li/Li⁺ at room temperature; (B) comparison of the rate capability of the CNT–Nb₂O₅ composite and pure Nb₂O₅ nanocrystals based electrodes; (C) analysis of the total charge stored on the electrode based on CNT–Nb₂O₅ composite; (D) galvanostatic charge/discharge profiles of CNT–Nb₂O₅ composite electrodes at various C-rates between 3.0 and 1.0 V; (E) corresponding rate performance of the composite electrode; (F) corresponding discharge capacity dependence on C-rates.

In order to investigate how the in situ growth of Nb₂O₅ onto conductive CNTs results in superior rate performance, Trasatti's method was applied to further analyze the contribution of the total stored charge.⁴⁴ As can be observed, the total stored charge q is strongly dependent on the scan rate (v), during which Li⁺ slowly diffuses into the “inner” region from surface. According to the diffusion law, q has a linear relation to v^−1/2, which can be expressed by q(v) = q₀ + constant × (v^−1/2). Obviously, constant × (v^−1/2) is the diffusion-controlled charge, while q₀ is independent on scan rate, which represents the charge stored on the accessible outer surface, and can be calculated by plotting the v^−1/2 with total stored charge q. Fig. 3C displays the total stored charge on the electrodes at 1 and 20 mV s⁻¹, respectively. The composites possess not only more total charge but also significantly higher outer surface charge than the pure Nb₂O₅ (Fig. S-4 in the ESI†). This result quantitatively indicates that building networks of CNTs with Nb₂O₅ nanocrystals allows more accessible surface sites of Nb₂O₅, which aligns well with the observed enhanced electron conductivity and rates performance.³⁴

The excellent rate performance of CNT–Nb₂O₅ composite can also be observed from galvanostatic charge/discharge curves. Fig. 3D shows charge and discharge profiles at various C-rates (1 C = 200 mA g_composite⁻¹). In good alignment with other Nb₂O₅-based electrode, the typical sloped feature can be found, indicating the nature of intercalation-type redox reaction. In another word, the lithium intercalation remains a single-phase process after repeated cycling.³⁴ The initial discharge capacity at 0.25 C is 232 mA h g⁻¹ (corresponding to ∼418 F g⁻¹ or ∼835 C g⁻¹), corresponding to ∼2.32 Li⁺ intercalated into the composite. The CNT–Nb₂O₅ composite delivers a reversible capacity of 175 and 112 mA h g⁻¹ (corresponding to 315 and 202 F g⁻¹ or 630 and 403 C g⁻¹) at a rate of 1 C and 10 C, respectively. Interestingly, the composite still possesses a high capacity even at an extremely high rate. As shown in Fig. 3E and F, the composite still possesses a specific capacity of ∼65 mA h g⁻¹ (∼117 F g⁻¹ or ∼234 C g⁻¹) at a rate of 50 C (corresponding to charge/discharge in 23 s). Even at 100 C (charge/discharge in 12 s), a capacity of ∼44 mA h g⁻¹ (∼79 F g⁻¹ or ∼160 C g⁻¹) still remains.

The superior rate capability CNT–Nb₂O₅ is mainly attributed to the unique composite architecture, where Nb₂O₅ nanocrystals in situ grow on the highly conductive CNTs networks. On one hand, the well-distributed Nb₂O₅ nanocrystals and porous CNTs networks create a large surface area, allowing fast accessibility of the active materials by electrolyte, while the low dimension of Nb₂O₅ materials efficiently shorten the Li⁺ diffusion length. On the other hand, the in situ growth ensures the intimate contact of Nb₂O₅ with highly conductive CNTs, which provides effective electron pathways. By comparison, physically mixed Nb₂O₅ nanocrystals and CNTs without any post-treatment leads to severe aggregation of the nanocrystals and poor contact between the active material with CNTs networks, showing an inferior rate capability (Fig. 3E).

Besides the significant improvement of rate performance, the CNT–Nb₂O₅ also exhibits impressive cycling stability. The long-term cycling performance of the composite is shown in Fig. 4A. A specific capacity of ∼80 mA h g⁻¹ remains (corresponding to ∼288 C g⁻¹) after 1000 cycles at a current of 10 C-rate, corresponding to more than 80% of its initial value. The very slow capacity decay could be attributed to the gradual loss of active materials and irreversible reactions. By comparison, physically mixed Nb₂O₅ and CNTs without any treatment shows a poor cycling stability. At the same 10 C-rate, the mixture electrode shows a dramatic capacity deterioration after only 200 cycles. The excellent durability of the electrodes mainly lies on the robust architecture, which is confirmed by electrochemical impedance spectroscopy (EIS) analysis. Fig. 4B shows the Nyquist plots of the composite electrode at different cycling status, where a single semicircle with a small intercept at high frequency range can be observed, implying a small ohmic and charge-transfer resistance. At low-frequency region, the Warburg tail possesses a moderately high slope, indicating the fast lithium-ion diffusion. After 100 cycles at high rate (10 C-rate), similar features can still remain, confirming the robust electrode structure, though a slight increase of charge-transfer and diffusion resistance can be observed from the low-frequency region. Nevertheless, the electrode preserves a small overall resistance due to the robust CNTs scaffold, which matches its excellent long-term cycling stability.

Fig. 4 (A) Comparison of long-term cycling performance of the CNT–Nb₂O₅ composites and physically mixed Nb₂O₅ and CNTs at 10 C-rate; (B) Nyquist plots of the CNT–Nb₂O₅ nanocomposites electrode at 1st, 5th and 100th discharge/charge cycles over the frequency range from 100 kHz to 0.01 Hz. Inset shows enlarged spectra at high-frequency range.

To further explore the feasibility to practically utilize CNT–Nb₂O₅ composite materials in the application of asymmetric supercapacitors, a full device was fabricated with composite material based anode and activated carbon based cathode. Based on the capacitance of each electrode as well as operation window, a mass ratio of cathode and anode of around 3.5 : 1 and a working window of 3.0–0.5 V were applied. Fig. 5A shows the representative galvanostatic charge/discharge profiles of the asymmetric supercapacitor. A symmetric triangle shape can be obtained at various current densities with negligible IR drop, indicating a typical capacitive response of supercapacitors. To better demonstrate the device performance, Ragone plot in Fig. 5B shows energy and power of CNT–Nb₂O₅ based asymmetric supercapacitor derived from constant current charge/discharge profiles. A high energy density of 47.17 W h kg⁻¹ can be reached, with a power density of 86.46 W kg⁻¹. Even at a high power of 6753.54 W kg⁻¹ (corresponding to a discharging time of 7.6 s), the asymmetric supercapacitor still possesses the energy of 14.77 W h kg⁻¹, indicating the high-power nature of supercapacitors can be well maintained. This type of asymmetric device exhibits significantly improved energy density compared with commonly used AC-based symmetric supercapacitors.⁴⁵ Moreover, such performance is even better than that of those asymmetric devices based on Nb₂O₅ materials involving either complex synthesis⁴⁶ or low electrode mass loading.³⁴

Fig. 5 (A) Galvanostatic charge/discharge profiles of asymmetric supercapacitor with CNT–Nb₂O₅ composite anode and AC cathode at various current densities between 3.0 and 0.5 V; (B) corresponding Ragone plot of the asymmetric supercapacitor and the comparison with activated carbon-based symmetric supercapacitor,^34,45 physically mixed CNTs/Nb₂O₅ nanocrystals-AC asymmetric supercapacitor,³⁴ and Nb₂O₅ nanowires-AC asymmetric supercapacitor⁴⁶ reported in recent literatures. Generally, the mass of the electrode materials was estimated to be 40% of the total device weight, therefore a packaging factor of 0.4 was applied in this work.⁶

4 Conclusions

We have successfully designed and developed high-performance electrode based on pseudocapacitive material of Nb₂O₅ nanocrystals in situ grown on CNTs for supercapacitor applications. The composites show excellent rate capability and long cycling stability, which is mainly attributed to the unique composite architecture where low dimensional Nb₂O₅ nanocrystals have intimate contact with highly conductive and robust CNTs scaffold. The prototype of asymmetric supercapacitor has also been assembled using CNT–Nb₂O₅ composite-based anode and AC-based cathode, and shows impressing energy and power performance. It is expected that such a facile synthesis of the composite materials as well as the fabrication of asymmetric device hold great promise in developing practical next-generation high performance asymmetric supercapacitors.

Acknowledgements

This work was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC), the University of Waterloo, and the Waterloo Institute for Nanotechnology. The authors also thank Canadian Center for Electron Microscopy (CCEM, McMaster University, Canada) for TEM characterization.

Notes and references

1. B. E. Conway, Electrochemical Supercapacitors: Scientific Fundamentals and Technological Applications, Kluwer Academic/Plenum Publishers, New York, 1999, ch. 1, p. 29.
2. P. Simon and Y. Gogotsi, Nat. Mater., 2008, 7, 845–854.
3. Y.-M. Chiang, Science, 2010, 330, 1485–1486.
4. K. Kang, Y. S. Meng, J. Breger, C. P. Grey and G. Ceder, Science, 2006, 311, 977–980.
5. K. Naoi, W. Naoi, S. Aoyagi, J.-i. Miyamoto and T. Kamino, Acc. Chem. Res., 2012, 46, 1075–1083.
6. S. W. Lee, N. Yabuuchi, B. M. Gallant, S. Chen, B.-S. Kim, P. T. Hammond and Y. Shao-Horn, Nat. Nanotechnol., 2010, 5, 531–537.
7. A. S. Arico, P. Bruce, B. Scrosati, J.-M. Tarascon and W. van Schalkwijk, Nat. Mater., 2005, 4, 366–377.
8. V. Chabot, D. Higgins, A. Yu, X. Xiao, Z. Chen and J. Zhang, Energy Environ. Sci., 2014, 7, 1564–1596.
9. L. L. Zhang and X. S. Zhao, Chem. Soc. Rev., 2009, 38, 2520–2531.
10. F. M. Hassan, V. Chabot, J. Li, B. K. Kim, L. Ricardez-Sandoval and A. Yu, J. Mater. Chem. A, 2013, 1, 2904–2912.
11. V. Augustyn, P. Simon and B. Dunn, Energy Environ. Sci., 2014, 7, 1597–1614.
12. J. P. Zheng and T. R. Jow, J. Power Sources, 1996, 62, 155–159.
13. Z.-S. Wu, D.-W. Wang, W. Ren, J. Zhao, G. Zhou, F. Li and H.-M. Cheng, Adv. Funct. Mater., 2010, 20, 3595–3602.
14. G. Yu, L. Hu, N. Liu, H. Wang, M. Vosgueritchian, Y. Yang, Y. Cui and Z. Bao, Nano Lett., 2011, 11, 4438–4442.
15. Z.-S. Wu, W. Ren, D.-W. Wang, F. Li, B. Liu and H.-M. Cheng, ACS Nano, 2010, 4, 5835–5842.
16. Z. Chen, V. Augustyn, J. Wen, Y. Zhang, M. Shen, B. Dunn and Y. Lu, Adv. Mater., 2011, 23, 791–795.
17. H.-G. Jung, S.-T. Myung, C. S. Yoon, S.-B. Son, K. H. Oh, K. Amine, B. Scrosati and Y.-K. Sun, Energy Environ. Sci., 2011, 4, 1345–1351.
18. L. Shen, H. Li, E. Uchaker, X. Zhang and G. Cao, Nano Lett., 2012, 12, 5673–5678.
19. J.-T. Han, D.-Q. Liu, S.-H. Song, Y. Kim and J. B. Goodenough, Chem. Mater., 2009, 21, 4753–4755.
20. J.-T. Han, Y.-H. Huang and J. B. Goodenough, Chem. Mater., 2011, 23, 2027–2029.
21. M. Wei, K. Wei, M. Ichihara and H. Zhou, Electrochem. Commun., 2008, 10, 980–983.
22. V. Augustyn, J. Come, M. A. Lowe, J. W. Kim, P.-L. Taberna, S. H. Tolbert, H. D. Abruna, P. Simon and B. Dunn, Nat. Mater., 2013, 12, 518–522.
23. J. W. Kim, V. Augustyn and B. Dunn, Adv. Energy Mater., 2012, 2, 141–148.
24. A. L. Viet, M. V. Reddy, R. Jose, B. V. R. Chowdari and S. Ramakrishna, J. Phys. Chem. C, 2009, 114, 664–671.
25. E. Lim, H. Kim, C. Jo, J. Chun, K. Ku, S. Kim, H. I. Lee, I.-S. Nam, S. Yoon, K. Kang and J. Lee, ACS Nano, 2014, 8, 8968–8978.
26. L. Kong, C. Zhang, S. Zhang, J. Wang, R. Cai, C. Lv, W. Qiao, L. Ling and D. Long, J. Mater. Chem. A, 2014, 2, 17962–17970.
27. G. Li, X. Wang and X. Ma, J. Mater. Chem. A, 2013, 1, 12409–12412.
28. G. Li, X. Wang, Z. Chen, X. Ma and Y. Lu, Electrochim. Acta, 2013, 102, 351–357.
29. G. Li, X. Wang and X. Ma, J. Energy Chem., 2013, 22, 357–362.
30. K. Brezesinski, J. Wang, J. Haetge, C. Reitz, S. O. Steinmueller, S. H. Tolbert, B. M. Smarsly, B. Dunn and T. Brezesinski, J. Am. Chem. Soc., 2010, 132, 6982–6990.
31. Z.-S. Wu, W. Ren, L. Xu, F. Li and H.-M. Cheng, ACS Nano, 2011, 5, 5463–5471.
32. E. Frackowiak and F. Beguin, Carbon, 2001, 39, 937–950.
33. C. Zhang, R. Maloney, M. R. Lukatskaya, M. Beidaghi, B. Dyatkin, E. Perre, D. Long, W. Qiao, B. Dunn and Y. Gogotsi, J. Power Sources, 2015, 274, 121–129.
34. X. Wang, G. Li, Z. Chen, V. Augustyn, X. Ma, G. Wang, B. Dunn and Y. Lu, Adv. Energy Mater., 2011, 1, 1089–1093.
35. Z. Chen, Y. Yuan, H. Zhou, X. Wang, Z. Gan, F. Wang and Y. Lu, Adv. Mater., 2014, 26, 339–345.
36. Z. Chen, D. Zhang, X. Wang, X. Jia, F. Wei, H. Li and Y. Lu, Adv. Mater., 2012, 24, 2030–2036.
37. C. Gao, C. D. Vo, Y. Z. Jin, W. Li and S. P. Armes, Macromolecules, 2005, 38, 8634–8648.
38. J. He, Y. Hu, Z. Wang, W. Lu, S. Yang, G. Wu, Y. Wang, S. Wang, H. Gu and J. Wang, J. Mater. Chem. C, 2014, 2, 8185–8190.
39. G. Park, N. Gunawardhana, C. Lee, S.-M. Lee, Y.-S. Lee and M. Yoshio, J. Power Sources, 2013, 236, 145–150.
40. G. E. McGuire, G. K. Schweitzer and T. A. Carlson, Inorg. Chem., 1973, 12, 2450–2453.
41. Z. Weibin, W. Weidong, W. Xueming, C. Xinlu, Y. Dawei, S. Changle, P. Liping, W. Yuying and B. Li, Surf. Interface Anal., 2013, 45, 1206–1210.
42. S. Qi, L. Fei, R. Zuo, Y. Wang and Y. Wu, J. Mater. Chem. A, 2014, 2, 8190–8195.
43. L. Li, J. Deng, R. Yu, J. Chen, X. Wang and X. Xing, Inorg. Chem., 2010, 49, 1397–1403.
44. S. Ardizzone, G. Fregonara and S. Trasatti, Electrochim. Acta, 1990, 35, 263–267.
45. G. B. Appetecchi and P. P. Prosini, J. Power Sources, 2005, 146, 793–797.
46. X. Wang, C. Yan, J. Yan, A. Sumboja and P. S. Lee, Nano Energy, 2015, 11, 765–772.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra05140a

---