A new method to prepare vanadium oxide nano-urchins as a cathode for lithium ion batteries

Abstract

Urchin-like vanadium oxide nanotube clusters, abbreviated to VO_x-NUs, were synthesized using a new method. Vanadium pentoxide, having a layered structure, was modified by lithium fluoride (LiF) and transformed into bi-phase lithium vanadate as the inorganic precursor. Then, VO_x-NUs were prepared by hydrothermal reaction with dodecylamine as a template. This is different from other molecular assembly methods reported. VO_x-NUs-350 nano clusters were obtained by annealing VO_x-NUs at a temperature of 350 °C in air. Both samples were identified as three dimensional urchin-like nano clusters. Based on the characterization data obtained, a formation mechanism was established. By simply varying the LiF stoichiometric ratio, the nano tube density of the VO_x-NUs could be controlled. VO_x-NUs presented a higher initial rate capacity of 400 mA h g⁻¹ and VO_x-NUs-350 maintained a better capacity of 150 mA h g⁻¹ after 50 cycles at a 100 mA g⁻¹ current density between 1.5 and 4 V versus Li/Li⁺.

---

Introduction

Hybrid electric vehicles (HEVs) and electric vehicles (EVs) powered by lithium ion batteries are more and more popular due to their economic and environmental advantages. In particular, EVs that travel remotely become potential substitutes for conventional vehicles. In addition, research of lithium ion battery electrode materials with a large capacity is urgently needed for popular large-screen mobile devices. However, the specific capacity and energy density of current lithium ion batteries are too small to meet the present requirements.

Vanadium oxide has been widely studied due to its excellent electrochemical and catalytic performance. Vanadium pentoxide is considered to have the highest theoretical capacity and be a promising cathode material for high-capacity lithium ion batteries.¹ However, due to irreversibility of the crystalline phases with ion insertion and extraction, poor ionic conductivity and active material solubility in the electrolyte, modified nanostructured vanadium oxide materials^2–4 are used as a new type of electrode material. Nanostructured vanadium oxide can significantly improve these properties because of its higher surface area and shorter Li⁺ ion diffusion distance.^5,6 A variety of one-dimensional nanostructured vanadium oxides have been synthesized.^7–9 However, low-dimensional nano-materials are delivering the same disadvantage of dispersibility (self-aggregation) due to their higher surface energy.^10–12 More orderly three dimensional vanadium oxide materials, such as nano-flowers,^13–16 nano-urchins^17–20 and nano-crusts,²¹ were synthesized by a hydrothermal method or alcothermal method to overcome this problem. By retaining large particles and reducing the contact area between electrodes and electrolytes, less dissolution compared to one-dimensional materials and higher volumetric energy densities are obtained.²² Vanadium oxide nano-urchins have been prepared typically using costly organic vanadium sources²³ or complicated cation-induced inorganic vanadium sources¹⁸ in combination with templates.

In this study, a new approach with low-cost inorganic precursors is proposed. It is different from the molecular assembly methods mentioned previously. Nano-urchins (VO_x-NUs) were synthesized by a hydrothermal method based on a traditional curling mechanism,²⁴ and the nano-tube clusters density was related to the LiF stoichiometric ratio. VO_x-NUs was calcined at 350 °C in air to produce urchin-like V₂O₅ nano crystal clusters abbreviated to VO_x-NUs-350. Morphology, structure and electrochemical characterization of VO_x-NUs and VO_x-NUs-350 were preliminarily determined.

Experimental

The synthesis was performed following the methods of solid-state reaction and hydrothermal treatment. First, starting materials with different stoichiometric proportions (molar ratios of V₂O₅ and LiF were 1 : 0.3 and 1 : 0.4) were transferred to different porcelain crucibles respectively, and then put in muffle furnaces. The muffle furnaces were heated to 700 °C at a heating rate of 5 °C min⁻¹, and kept at 700 °C for 2–4 h. The melted compounds were left to cool naturally, and then the modified precursors were obtained and denoted by 0.3LiF + VO_x and 0.4LiF + VO_x.²⁵ These as-synthesized precursors were ground into powder by ball milling. Second, the hydrothermal procedure was similar to those reported in literature.^26,27 1 g modified precursors αLiF + VO_x powder and 1.04 g dodecylamine were slowly added in 80 mL deionized water under magnetic stirring for about 30 min, which led to the formation of a yellow suspension liquid. The yellow suspension liquid was vigorously stirred for 12 h and transferred into a 100 mL Teflon lined hydrothermal autoclave with a stainless steel shell. The autoclave was kept at 180 °C in an oven for 7 days. The precipitate obtained was filtered and rinsed repeatedly with ethanol and deionized water and then dried at 70 °C under vacuum for 8 h. The resulting black powders were urchin-like vanadium oxide nanotube clusters (VO_x-NUs). 0.3VO_x-NUs were the product of precursor 0.3LiF + VO_x and 0.4VO_x-NUs was the product of precursor 0.4LiF + VO_x. VO_x-NUs-350 nano crystal clusters were obtained by annealing VO_x-NUs at a temperature of 350 °C for 3 h in air with a heating rate of 5 °C min⁻¹ and are denoted by 0.3VO_x-NUs-350 and 0.4VO_x-NUs-350.

Scanning electron microscopy (SEM: Philips-XL-30FEG) and transmission electron microscopy (TEM: 200 kV JEOL-1230) were introduced to observe the morphologies and characterize the microstructure of samples with very high spatial resolution. X-ray powder diffraction (XRD) patterns were obtained using a RigataD/max-C diffractometer with a CuKα radiation source (λ = 1.5406 Å) in the range of 5–60° for phase identification. FT-IR absorption spectroscopy measurements were carried out using a Bruker-TENSOR27 FTIR spectrometer in the range of 400–4000 cm⁻¹ to characterize surface chemical–physical properties by the KBr disk method. The thermogravimetry (TG) and differential scanning calorimeter (DSC) measurements were performed using a SDT Q600 in the temperature range of 17–650 °C with a heating rate of 10 °C min⁻¹ in air. XPS experiments were conducted using a RBD upgraded PHI-5000C ESCA system (Perkin Elmer) with Mg Kα radiation (hν = 1253.6 eV). Binding energies were calibrated using the containment carbon (C_1s = 284.6 eV). The data analysis was carried out using XPS Peak 4.1 software.

The cathode electrodes for electrochemical characterization were prepared by mixing active materials with carbon black and polyvinylidene fluoride (PVDF) binder (70 : 20 : 10 wt%) in N-methyl-2-pyrrolidone (NMP) solvent. Then, the obtained slurry was uniformly coated on an aluminum foil current collector. The electrodes were dried in vacuum at 120 °C for 8 h and then cut into a circular wafer for testing, with a microporous film (Celgard 2500) as a membrane and lithium metal as the anode and reference electrode in model CR2025 coin cells. The electrolyte was 1 M LiPF₆ dissolved in ethylene carbonate (EC)/dimethyl carbonate (DMC)/diethyl carbonate (DEC) (1 : 1 : 1 in volume). The coin cells were assembled in an argon-filled glove box with moisture content and an oxygen level less than 1 ppm. The electrochemical tests were performed using a LAND cell-testing system and a CHI660C electrochemical workstation at room temperature. Cyclic voltammetry (CV) measurements were performed at a scanning rate of 0.1 mV s⁻¹. Electrochemical impedance spectroscopy (EIS) measurements were carried out in a frequency range between 100 kHz and 0.01 Hz at a discharge state of 2.3 V, with 5 mV as the amplitude of the AC signal. The impedance data were analyzed using Z-view software and fitted based on the simple equivalent circuit. Galvanostatic charge/discharge tests were recorded within a voltage range of 1.5–4.0 V (vs. Li/Li⁺), at a charge/discharge current of 100 mA g⁻¹.

Results and discussion

Fig. 1a and b are SEM and TEM images of 0.3VO_x-NUs, respectively, which have been prepared with as-synthesized 0.3LiF + V₂O₅. As can be seen in Fig. 1a, the 0.3VO_x-NUs show a sea urchin-like three-dimensional configuration with pin-like vanadium oxide nanotubes over it. Large quantities of open-end nanotubes (∼1 μm) are evident. These dense nanotube clusters have diameters in the range of 2–3 μm. The internal structure of nano-urchins can be further seen in the TEM images (Fig. 1b). Expanding from a center point with dense nano-tubes uniformly spread out is shown in the inset of Fig. 1b. Fig. 1b shows one single nano-tube from nano-urchins, which maintain a complete nano-tube structure. Its outside diameter is greater than 60 nm and inside diameter is about 20 nm. The interlamellar distance is clearly 2 nm. The interlamellar distance reflects the two-dimensional structure of the VO_x layers and the 2 nm spacing corresponds to reflection 001.²⁸ Fig. 1c and d show the SEM and TEM of 0.4VO_x-NUs prepared with as-synthesized 0.4LiF + V₂O₅. Compared with 0.3VO_x-NUs, 0.4VO_x-NUs also present a three-dimensional configuration of nano-urchins, as shown in the insets of Fig. 1c and d.

Fig. 1 SEM image (a) and TEM image (b) of 0.3VO_x-NUs. SEM image (c) and TEM image (d) of 0.4VO_x-NUs. SEM image (e) and TEM image (f) of 0.3VO_x-NUs-350. SEM image (g) and TEM image (h) of 0.4VO_x-NUs-350. Insets: low magnification TEM images.

However, the nano-tube volumetric density of 0.4VO_x-NUs was lower than the 0.3VO_x-NUs, as observed in Fig. 1a and c. 0.4VO_x-NUs nano-tubes clusters have larger diameters in the range of 15–20 μm and longer nanotubes (∼10 μm). Fig. 1d displays a single nano-tube of 0.4VO_x-NUs. Where tube walls locally collapse, the outside diameter was less than 60 nm, and inside diameter was about 27 nm. Nano-tube structure remained intact and the interlamellar distance expands to 2.7 nm. It is illustrated that LiF is influential on the three dimensional structure, but also on the structure of a single nano-tube.

Fig. 1e and f show the urchin-like configuration and building nano-blocks of 0.3VO_x-NUs-350. It can be seen that the 0.3VO_x-NUs-350 synthesized by sintering 0.3VO_x-NUs was composed of small interconnected nano-blocks with a size of 50–100 nm. These nano-urchin clusters also have diameters in the range of 2–3 μm with lower volumetric density nano-tubes over them. The same things happen to 0.4VO_x-NUs-350 with larger cluster diameters of 15–20 μm in Fig. 1g and h.

As seen in the FTIR spectrum of Fig. 2, VO_x-NUs and VO_x-NUs-350 present some differences with the characteristic vibration peaks, especially in region characteristic of end group oxygen.²⁹ The absorption peaks at 1465, 2852, and 2922 cm⁻¹ are assigned to the various bending and stretching modes of C–H vibrations in dodecylamine for the VO_x-NUs. These peaks disappear for V₂O₅-NUs-350, indicating the removal of dodecylamine after sintering. Two peaks at 1627 and 3450 cm⁻¹ are attributed to H–O stretching and H–O–H bending vibration modes. For all samples, the characteristic peaks near 1000 cm⁻¹ (999 and 1022 cm⁻¹), ranging from 700 to 900 cm⁻¹ (831 cm⁻¹) and below 700 cm⁻¹ (499, 626 and 630 cm⁻¹) correspond to the stretching vibration of terminal oxygen bonds (V=O), the vibration of doubly coordinated oxygen bonds (O–V–O), and the asymmetric and symmetric stretching vibrations of triply coordinated oxygen bonds in vanadium oxide (3V–O), respectively.^27,30 The shifting of corresponding coordinated oxygen bonds may be caused by the lattice distortion and microscopic stress variation in the coordination geometry. LiF characteristic peaks are not observed in Fig. 2. It is believed that LiF had been removed during the sintering process²⁵ and hydrothermal process. Although nano-urchins do not present additional absorption peaks in contrast to VO_x-NTs,³⁰ LiF does modify precursors and has an impact on the final morphology according to the SEM and TEM analysis mentioned above.

Fig. 2 FT-IR spectra of 0.3VO_x-NUs (#1), 0.4VO_x-NUs (#2), 0.3VO_x-NUs-350 (#3) and 0.4VO_x-NUs-350 (#4).

Fig. 3 gives the TG-DSC curves of VO_x-NUs in air. The weight loss of 10% from the room temperature to about 200 °C is assigned to the thermal evaporation of absorbed water molecules. The thorough oxidative decomposition of organic templates for VO_x-NUs was completed at 450 °C, causing a total of 51% weight loss which will describe the approximate content of the organic templates of the VO_x-NUs. The exothermic peaks at 240 °C can be ascribed to the decomposition of dodecylamine. Two exothermic peaks at 398 °C and 480 °C can mainly be ascribed to the recrystallization of the VO_x layers. The organic templates will mostly be removed at 350 °C and recrystallization will not change the 3D nanostructures.

Fig. 3 TG-DSC curves of αVO_x-NUs (α = 0.3, 0.4) in air.

The XPS technique was employed to identify the elemental composition of αLiF + VO_x (α = 0.3, 0.4) and analyze the content of the vanadium element with different valence states. The XPS general spectrum of αLiF + VO_x is given in Fig. 4a and c. The F_1s peak at about 690 eV originates from LiF is too weak to identify.

Fig. 4 XPS general spectrum and V_2p region of 0.3LiF + VO_x (a and b) and 0.4LiF + VO_x (c and d).

Fig. 4b and d present the V_2p region of the XPS spectrum, which includes the V_2p3/2 and V_2p1/2 peaks; the V_2p3/2 peak can be divided into two peaks at binding energies of 517.3 and 516.1 eV, which are ascribed to V⁵⁺ and V⁴⁺, respectively. As shown in Fig. 4b and d, XPS Peak 4.1 software was applied to fit the experimental data and ascertain the ratio of V⁵⁺/V⁴⁺ in the αLiF + VO_x through the calculation of V⁵⁺ peak area/V⁴⁺ peak area. The fitted result illustrates that the ratio of V⁵⁺/V⁴⁺ in 0.3LiF + VO_x is 0.29, that is, V⁵⁺ and V⁴⁺ account for 22.7% and 77.3%. The ratio of V⁵⁺/V⁴⁺ in 0.4LiF + VO_x is 1.02, that is, V⁵⁺ and V⁴⁺ account for 50.5% and 49.5%, respectively.

Fluorine is believed not to insert into the vanadium oxide layers due to the reaction of fluorine and silicon oxide, which is the main component of the crucible. The reaction has been described in detail in another study.²⁵

XRD patterns are performed to determine the crystalline phases in Fig. 5. The black line (Fig. 5a) is the XRD pattern of sample 0.3LiF + V₂O₅. The red line (Fig. 5b) is for 0.4LiF + V₂O₅. It is thought that peaks marked by ○ ascribed to LiV₃O₈ phase (JCPDS card: no. 35-0437), and other strong peaks are due to the Li_0.3V₂O₅ phase (JCPDS card: 18-0755), marked by Δ. It can be seen that the LiV₃O₈ phase in 0.4LiF + V₂O₅ is stronger than 0.3LiF + V₂O₅.²⁵ The blue line (Fig. 5c) and cyan line (Fig. 5d) are the XRD patterns of the 0.3VO_x-NUs and 0.4VO_x-NUs, respectively. VO_x-NUs samples can be indexed on the basis of the orientation (002), (003), (110), (200), (210), and (310).^19,20 The Bragg reflections indicate that VO_x-NUs are lamellar intercalated products of amine and vanadium oxide. The peaks (002) and (003) with higher intensity at the small angle region corroborate the interlayer distance and can be used to calculate the distance between VO_x layers.²⁷ According to Bragg's law, the distances are 2.8 nm for 0.3VO_x-NUs and 3 nm for 0.4VO_x-NUs, which are close to the results of the TEM observation. The characteristic reflection peaks (110), (200), (210) and (310) indicate high structural order of vanadate layers, but also intercalated amine groups between these layers.^20,31 Remarkably, the positions of the hk0 reflections are the same for all the VO_x-NUs samples, indicating that the contents of the wall structure are not affected by LiF. However, the (00l) peaks of 0.4VO_x-NUs significantly move to lower degree values than those of 0.3VO_x-NUs. The shifts mean that the distance between the layers in 0.3VO_x-NUs is smaller than that in 0.4VO_x-NUs. Prelithiation of V₂O₅ makes more lithium ions insert into the vanadium oxide layers, which can increase the layer spacing and is consistent with the TEM images. VO_x-NUs-350 samples display a series of characteristic diffraction peaks (200), (001), (101), (110), and (111), (020), which can be indexed to the orthogonal V₂O₅ phase (JCPDS no. 41-1426, space group: Pmmn, a = 11.516 Å, b = 3.566 Å, c = 4.372 Å), and no LiF peaks were detected, indicating the high purity of V₂O₅ nano-crystals.^12,27 By calculation based on Scherrer formula, the average grain size of V₂O₅ nano-crystals is ∼40 nm, which is in agreement with the TEM and SEM observations.

Fig. 5 XRD patterns of 0.3LiF + V₂O₅ (a), 0.4LiF + V₂O₅ (b), 0.3VO_x-NUs (c), 0.4VO_x-NUs (d), 0.3VO_x-NUs-350 (e), 0.4VO_x-NUs-350 (f) and the orthogonal V₂O₅ standard peaks (JCPDS no. 41-1426, space group: Pmmn, a = 11.516 Å, b = 3.566 Å, c = 4.372 Å).

In Fig. 6, the formation of nano-urchins is simulated. The as-prepared precursors were double-crystal materials known from the XRD in ref. 25. In addition, these precursors are bulk materials in above SEM-1 of Fig. 6. Grain boundaries exist between different crystal substances. With the insertion of the template, the layered materials produce split lines due to unbalanced stress, which are schematically shown in Fig. 6-1. In the hydrothermal process (Fig. 6-2), the fracture layers center on one point and partly curl as VO_x-NTs,⁹ as shown in SEM-2 of Fig. 6. This image is a sectional view of semi-finished VO_x-NUs internal structure and nano-urchins forming state has been shown clearly. Basically, it can be speculated that the forming process of nano-tubes began from the outside with a common center. Curling of many super imposed layers eventually forms nano-urchins (Fig. 6-3). These correspond to the SEM and TEM images.

Fig. 6 Schematic diagram showing the formation of VO_x-NUs.

Cyclic voltammetry was employed to evaluate the samples' electrochemical properties. Fig. 7 shows the initial three cycles under a low scan rate of 0.1 mV s⁻¹ between 1.5 and 4 V. At the first cycling of 0.3VO_x-NUs, two anodic peaks at 2.8 V and 3.1 V were observed. One cathodic peak at 2.2 V related to the lithium ion extraction is observed. At the second and third cycle, the cathodic peaks moved to 2.25 V and 2.3 V, respectively. For 0.3VO_x-NUs-350, one anodic peak at 2.5 V is observed and moves to 2.7 V for the following cycles. The cathodic peak is maintained at 2.25 V, indicating a good cycling ability. 0.4VO_x-NUs has three anodic peaks at 2.5 V, 3.25 V and 3.75 V at the first cycle, which become two peaks at 2.75 V and 3.4 V at the second cycle and disappears on the third cycle. It does not show a distinct cathodic peak. 0.4VO_x-NUs-350 revealed cycling ability as good as 0.3VO_x-NUs-350. One anodic peak at 2.5 V was observed that moved to 2.7 V for the subsequent cycles. The cathodic peak was maintained at 2.25 V.

Fig. 7 Cyclic voltammogram of samples at a scan rate of 0.1 mV s⁻¹ between 1.5 and 4.0 V.

Electrochemical impedance spectroscopy (EIS) was introduced to evaluate the electrochemical mechanism of these samples. The typical Nyquist plots at a discharge state of 2.3 V are presented in Fig. 8a and b. The plots show depressed semicircles in the high-frequency region and slope lines in the low-frequency region. Generally, the diameter of the semicircle in the high-frequency region is associated with the charge transfer reaction at the electrolyte/electrode interface. The smaller the diameter the semicircle is, the smaller the charge transfer resistance will be. The slope line in the low-frequency region is related to the Warburg impedance associated with the ion diffusion process in the electrode materials.^27,30 As can be seen from the insets of Fig. 8a and b, the semicircle diameters of VO_x-NUs-350 are smaller than those of VO_x-NUs in the high-frequency region, indicating that VO_x-NUs-350 samples had a lower charge transfer resistance compared to the VO_x-NUs samples.

Fig. 8 Nyquist plots (a) for 0.3VO_x-NUs and 0.4VO_x-NUs and (b) for 0.3VO_x-NUs-350 and 0.4VO_x-NUs-350 with 5 mV amplitude of AC signal at the state of charge of 2.3 V (insets: equivalent circuit used for fitting the Nyquist plots and Nyquist plots magnification of the high frequency region).

The two simple equivalent circuits, shown in the inset of Fig. 8a and b, were designed to analyze the impedance spectra of the samples. In the circuit, R_e presents the electrolyte resistance, R_ct stands for the charge transfer resistance, CPE is the double layer capacitance, and W is the Warburg impedance. The electrochemical parameters of the samples were fitted by using Z-view software, and a good agreement between the experimental results and the parameters obtained from the equivalent circuit can be seen from the EIS spectra. The fitted electrochemical parameters are listed in Table 1. As can be seen from Table 1, the charge transfer resistances (R_ct) for 0.3VO_x-NUs and 0.4VO_x-NUs at an open circuit potential (R_ct1 + R_ct2) are 223.1 ohm and 232.74 ohm, respectively. The charge transfer resistances (R_ct) for 0.3VO_x-NUs-350 and 0.4VO_x-NUs-350 are 163 ohm and 163.2 ohm, respectively. It is clear that the R_ct values of VO_x-NUs-350 samples are considerably smaller than that of VO_x-NUs samples, indicating that charge transfer is easier for VO_x-NUs-350 samples than VO_x-NUs.

Table 1 The fitted electrolyte resistance (R_e), charge transfer resistance (R_ct) for the two samples at the state of charge (SOC) of 2.3 V

Samples		0.3VOx-NUs	0.3VOx-NUs-350	0.4VOx-NUs	0.4VOx-NUs-350
Open circuit potential	Rct1/Ω	123	163	88.44	163.2
	Rct2/Ω	100.1		144.3

---

The Bode plots in Fig. 9 can be used to estimate the effectiveness of lithium ion diffusion in the electrode materials. According to previous studies,^25,32 the lithium ion diffusion is related to the phase angle in the low-frequency region. The smaller the phase angle is, the faster the lithium ions diffusion is. As shown in Fig. 9, the phase angles of the VO_x-NUs-350 samples are considerably smaller than that of VO_x-NUs samples in the low-frequency region, indicating that VO_x-NUs-350 samples have a more rapid lithium ions diffusion speed than that of the VO_x-NUs samples. The improved electrochemical kinetics may be attributed to the removal of the organic template according to the FT-IR in Fig. 2.

Fig. 9 Bode plots of samples at an open circuit potential of 2.3 V.

Discharge capacities versus cycle numbers of VO_x-NUs and VO_x-NUs-350 are plotted in Fig. 10 at a constant current density of 100 mA g⁻¹ between 1.5 and 4.0 V vs. Li/Li⁺. 0.3VO_x-NUs and 0.4VO_x-NUs delivered initial specific capacities of 396.4 mA h g⁻¹ and 351.9 mA h g⁻¹, respectively. After 20 cycles, capacities of 40 and 50 mA h g⁻¹ remained. The initial capacity values for 0.3VO_x-NUs-350 and 0.4VO_x-NUs-350 was 366.3 and 363.5 mA h g⁻¹, respectively. After 50 cycles, both faded to about 150 mA h g⁻¹. The rapid capacity decay of VO_x-NUs may be caused by the decomposition of organic dodecylamine and its severe contamination of the electrolyte. It can be concluded that nano urchin-like VO_x-NUs-350, which is composed of interconnected V₂O₅ nanocrystals and free of organic templates, possesses a high specific capacity and superior cycling performance.

Fig. 10 Galvanostatic cycles of 0.3VO_x-NUs, 0.4VO_x-NUs, 0.3VO_x-NUs-350 and 0.4VO_x-NUs-350 at the current density of 100 mA g⁻¹ between 1.5 and 4 V.

Conclusions

Two species of nano-urchins of vanadium oxide were prepared by a new method involving bi-phase lithium vanadate materials and a hydrothermal method with dodecylamine. SEM and TEM showed the structure and configuration of nano-urchins. XRD and FT-IR analysis confirmed that components of nano-urchins are nano-tubes, which V₂O₅ nanocrystals substitute for after sintering. Both the VO_x-NUs and VO_x-NUs-350 samples exhibited an excellent initial discharge specific capacity of about 400 mA h g⁻¹ at a 100 mA g⁻¹ current density between 1.5 and 4 V versus Li/Li⁺. However, the VO_x-NUs-350 samples possessed outstanding cycling performance and a better capacity of 150 mA h g⁻¹ after 50 cycles due to the removal of organic template.

Acknowledgements

The authors gratefully acknowledge the financial support from the Shanghai Committee of Science and Technology (11nm0501300 and 13JC1408700), the National Natural Science Foundation of China (51472182, 51272179 and 51102183).

Notes and references

1. M. S. Whittingham, Chem. Rev., 2004, 104, 4271–4301.
2. X. L. Jia, L. Q. Zhang, R. F. Zhang, Y. F. Lu and F. Wei, RSC Adv., 2014, 4, 21018–21022.
3. Z. L. Li, Q. Y. Zhu, S. N. Huang, S. S. Jiang, S. Lu, W. Chen and G. S. Zakharova, RSC Adv., 2014, 4, 46624–46630.
4. B. Saravanakumar, K. K. Purushothaman and G. Muralidharan, RSC Adv., 2014, 4, 37437–37445.
5. R. S. Devan, R. A. Patil, J. H. Lin and Y. R. Ma, Adv. Funct. Mater., 2012, 22, 3326–3370.
6. M. Shahid, D. S. Rhen, I. Shakir, S. P. Patole, J. B. Yoo, S. J. Yang and D. J. Kang, Mater. Lett., 2010, 64, 2458–2461.
7. J. Livage, Materials, 2010, 3, 4175–4195.
8. F. Sediri, F. Touati and N. Gharbi, Mater. Lett., 2007, 61, 1946–1950.
9. W. Chen, J. F. Peng, L. Q. Mai, Q. Y. Zhu and Q. Xu, Mater. Lett., 2004, 58, 2275–2278.
10. Y. G. Guo, J. S. Hu and L. J. Wan, Adv. Mater., 2008, 20, 2878–2887.
11. L. Q. Mai, L. Xu, C. H. Han, X. Xu, Y. Z. Luo, S. Y. Zhao and Y. L. Zhao, Nano Lett., 2010, 10, 4750–4755.
12. A. Q. Pan, H. B. Wu, L. Yu, T. Zhu and X. W. Lou, ACS Appl. Mater. Interfaces, 2012, 4, 3874–3879.
13. H. Yu, X. Rui, H. Tan, J. Chen, X. Huang, C. Xu, W. Liu, D. Y. W. Yu, H. H. Hng, H. E. Hoster and Q. Yan, Nanoscale, 2013, 5, 4937–4943.
14. Y. X. Tang, X. H. Rui, Y. Y. Zhang, T. M. Lim, Z. L. Dong, H. H. Hng, X. D. Chen, Q. Y. Yan and Z. Chen, J. Mater. Chem. A, 2013, 1, 82–88.
15. H. H. Yin, K. Yu, Z. L. Zhang, M. Zeng, L. Lou and Z. Q. Zhu, Electroanalysis, 2011, 23, 1752–1758.
16. H. Fei, Y. Lin and M. Wei, J. Colloid Interface Sci., 2014, 425, 1–4.
17. M. Roppolo, C. B. Jacobs, S. Upreti, N. A. Chernova and M. S. Whittingham, J. Mater. Sci., 2008, 43, 4742–4748.
18. H. L. Fei, Z. R. Shen, J. G. Wang, H. J. Zhou, D. T. Ding and T. H. Chen, Electrochem. Commun., 2008, 10, 1541–1544.
19. C. O'Dwyer, V. Lavayen, S. B. Newcomb, E. Benavente, M. A. Santa Ana, G. Gonzalez and C. M. S. Torres, Electrochem. Solid-State Lett., 2007, 10, A111–A114.
20. C. O'Dwyer, D. Navas, V. Lavayen, E. Benavente, M. A. Santa Ana, G. Gonzalez, S. B. Newcomb and C. M. S. Torres, Chem. Mater., 2006, 18, 3016–3022.
21. Q. Song, H. C. Pang, W. T. Gong, G. L. Ning, S. Gao, X. L. Dong, C. J. Liu, J. Y. Tian and Y. Lin, RSC Adv., 2015, 5, 4256–4260.
22. A. S. Arico, P. Bruce, B. Scrosati, J. M. Tarascon and W. Van Schalkwijk, Nat. Mater., 2005, 4, 366–377.
23. V. Lavayen, C. O'Dwyer, M. A. Santa Ana, S. B. Newcomb, E. Benavente, G. Gonzalez and C. M. S. TorreS, Phys. Status Solidi B, 2006, 243, 3285–3289.
24. M. Niederberger, H. J. Muhr, F. Krumeich, F. Bieri, D. Gunther and R. Nesper, Chem. Mater., 2000, 12, 1995–2000.
25. J. Wang, G. Gao, X. Zhou, J. Wu, H. Yang, Q. Li and G. Wu, J. Solid State Electrochem., 2014, 18, 2459–2467.
26. C. J. Cui, G. M. Wu, J. Shen, B. Zhou, Z. H. Zhang, H. Y. Yang and S. F. She, Electrochim. Acta, 2010, 55, 2536–2541.
27. X. Zhou, C. Cui, G. Wu, H. Yang, J. Wu, J. Wang and G. Gao, J. Power Sources, 2013, 238, 95–102.
28. F. Krumeich, H. J. Muhr, M. Niederberger, F. Bieri, B. Schnyder and R. Nesper, J. Am. Chem. Soc., 1999, 121, 8324–8331.
29. A. Surca and B. Orel, Electrochim. Acta, 1999, 44, 3051–3057.
30. C.-J. Cui, G.-M. Wu, H.-Y. Yang, S.-F. She, J. Shen, B. Zhou and Z.-H. Zhang, Electrochim. Acta, 2010, 55, 8870–8875.
31. M. Worle, F. Krumeich, F. Bieri, H. J. Muhr and R. Nesper, Z. Anorg. Allg. Chem., 2002, 628, 2778–2784.
32. C. Cui, G. Wu, H. Yang, S. She, J. Shen, B. Zhou and Z. Zhang, Solid State Commun., 2010, 150, 1807–1811.

---