Preparation of a graphitic N-doped multi-walled carbon nanotube composite for lithium–sulfur batteries with long-life and high specific capacity

Abstract

One of the challenges for lithium–sulfur batteries is a rapid capacity fading owing to the insulating of sulfur and Li₂S₂/Li₂S compounds, the dissolving and consequent shuttling of polysulfide generated as intermediates during charge–discharge processes in the electrolyte. In this work, graphitic N-doped multi-wall carbon nanotube (GN/PNCNTs) composites are synthesized by in situ chemical polymerization and carbonization processes. The nitrogen doping in the GN/PNCNTs composite can effectively enhance chemisorption between sulfur and carbon, which can enable the uniform deposition of discharge products and lead to a high utilization and reversibility of active materials. Because of these technological superiorities, the as-prepared S-GN/PNCNTs cathode with a sulfur content of 60 wt% exhibits high initial specific capacity and excellent cycling stability at up to 600 cycles at 1C. Meanwhile, the rate capacities of the cathode are demonstrated from 0.5C to 6C with a specific capacity of 1178 mA h g⁻¹ (the initial specific capacity) to 586 mA h g⁻¹ (the 60^th cycle).

---

Introduction

The lithium–sulfur (Li–S) battery has been investigated as a promising candidate for the next-generation high energy battery due to its intrinsic high theoretical energy density (2600 W h kg⁻¹), which is 3–5 times higher than that of the lithium ion battery.¹ Nevertheless, the Li–S battery suffers from low utilization of the active material, severe capacity fading and poor cycling stability, which are mainly attributed to three reasons, (i) the dissolution of long-chain lithium polysulfide (PS, Li₂S_n, 4 ≤ n ≤ 8) into the electrolytes causing the redox shuttle effect between the anode and the cathode;² (ii) large volumetric expansion of the elemental sulfur;³ (iii) electronic and ionic insulating of the elemental sulfur and the insoluble Li₂S₂/Li₂S compounds.⁴

A rational strategy to overcome the above challenges is the growth of uniform conducting polymers thin film on the surfaces of carbon materials like multi-walled carbon nanotubes (MWCNTs). In most of these composites, a ultrathin conducting polymer (e.g. polypyrrole) could provide a high specific capacity and enhance carbon materials with a higher electrical conductivity due to the molecular bonds consisted of conjugated double bonds (N=C–C).^5,6 For instance, Kazazi and Vaezi⁷ reported that a polypyrrole (PPy)-coated sulfur/MWCNT cathode material, where a surface coating layer could facilitate electrons transport, restrain the dissolution of polysulfide and alleviate the shuttle effect. However, poor cycling stability and rate capacities are still drawbacks for PPy/S/MWCNTs as cathodes, which is induced by the change in the structural conformation with repeated ion exchange.⁸ Hence, a reasonable strategy is still needed by strong coordination between nitrogen atoms with carbon materials to accommodate the volumetric change.

Currently, many researchers have reported that urea, as a precursor, can not only provide reactive carbon and nitrogen sources to produce graphitic carbon nitride (g-C₃N₄) during pyrolysis, but also produce pores derived from the byproducts gases.^9–11 Owing to its high nitrogen content and facile synthesis procedure, g-C₃N₄ may provide more active reaction sites than other N-doped carbon materials.^12–16 The chemisorption between sulfur and carbon is enhanced by the introduction of pyridine-like defects, which could enable uniform deposition of Li₂S on PPy-coated carbon materials to promote good electrical contact, thus effectively confine the dissolution of polysulfide.¹⁷ Produced small-mesopores on the structures also can effectively confine polysulfide into the pores. In addition, as a graphite-like semiconductor, g-C₃N₄ has many advantages like non-toxicity, chemical and thermal stability.¹¹

Herein, we design a rational pyrolysis process to synthesize GN/PNCNTs composites, which can provide a g-C₃N₄ layer as a two-dimensional (2D) reactive template with high nitrogen content on the surface of PCNTs, as well as enhance the interaction between carbon and sulfur. The inner PCNTs structure not only facilitates electron transport, but also improves a tenacious chemical interaction between the imine group (–N=) of pyrrole and the sulfur. The outer layer g-C₃N₄ provides a nanoscale coating to create pyridine-like defects on the tube's wall, which helps sulfur species (e.g. PS) penetrate into the tube's interior. Our results reveal that the fabricated S-GN/PNCNTs composite with a high content of sulfur (60 wt%) used as a cathode exhibits a long life (exceeding 500 cycles at 1C) with a high retention (from 1068 to 600 mA h g⁻¹ over 500 cycles). In addition, the cells have excellent high-rate performances (0.5–6C) high coulombic efficiency close to 100%.

Experimental

Synthesis of PPy/MWCNTs (PCNTs) precursor

MWCNTs (20 mg) were firstly dispersed in a mixed solution containing 95 mL deionized water and 8.6 mL hydrochloric acid (36–38 wt%) under ultrasonic for 4 h. Then 20 mg cetyl trimethyl ammonium bromide (CTAB) was added to the above solution under stirring for 10 min. 0.2 mL pyrrole monomer (PM) was added dropwise and stirred for 30 min, subsequently the mixture was placed in the ice-water bath. After that, a solution mixed with 0.68 g ammonium persulphate and 5 mL hydrochloric acid solution (1 mol L⁻¹ HCl) was added dropwise under stirring for 12 h in the ice-water bath. Finally, the product was obtained by freeze-drying after washed with deionized water.

Synthesis of S-g-C₃N₄/PPy/MWCNTs (S-GN/PNCNTs) composite

The suitable amount urea and the above products on the weight ratio of 10 : 1 were mixed under freeze-drying for 48 h. The obtained mixture was transferred into a tubular furnace for pyrolysis under nitrogen flow. The temperature was first raised to 600 °C at 1 °C min⁻¹ and kept for 1 hour, then raised to 800 °C and kept for another 1 hour. The obtained precursor g-C₃N₄/PPy/MWCNTs (GN/PNCNTs) were mixed with sulfur of 4 : 6, then sealed in a teflon-lined stainless-steel autoclave (N₂ atmosphere) and heated to 155 °C for 12 h and 300 °C for 2 h.

Preparation of the S-GN/PNCNTs cathode

The cathode slurry was made by mixing 70% S-GN/PNCNTs composite, 20% carbon black and 10% polyvinylidene fluoride (PVDF) with 1-methyl-2-pyrrolidinone (NMP), and then coated evenly onto carbon cloth (HCP330) where the content of active material sulfur was around 0.7 mg cm⁻². A solution of 1 M lithium bis(trifluoromethane sulfonyl)imide (LITFSI) and 0.1 M lithium nitrate (LiNO₃) dissolved in mixed solvent of dimethoxyethane (DME) and 1,3-dioxolane (DOL) with a volume ratio of 1 : 1 was employed as electrolyte. The coin cells (CR-2025) contained Celgard 2300 separator and lithium foils.

Characterizations

The morphologies of PCNTs, GN/PNCNTs and S-GN/PNCNTs composites were characterized by Hitachi S-4800 field emission scanning electron microscopy. The morphologies and energy-dispersive spectroscopy (EDS) were analyzed by a FEI Tecnai G2 S-Twin instrument. Thermogravimetric (TGA) curve was carried out on a STA 449 °C Jupiter (NETZSCH) thermogravimetry analyzer from room temperature to 800 °C under Ar₂ atmosphere with a heating rate of 10 °C min⁻¹. The X-ray diffraction (XRD) patterns of the samples were determined by a Bruker D8 Focus power X-ray diffractometer using Cu Kα radiation. N₂ adsorption–desorption isotherm was performed on a Micromeritics ASAP 2020 instrument. The surface area measurement was performed according to the Brunauer–Emmett–Teller (BET) method and the pore width distribution was obtained by means of the Barrett–Joyner–Halenda (BJH) method. X-ray photoelectron spectroscopy (XPS) characterizations were carried out in an ESCALAB 250 instrument with 150 W Al Kα probe beam.

Electrochemical measurements

The galvanostatic charge/discharge profiles were recorded between 1.7 and 2.8 V (vs. Li/Li⁺) using LAND CT2001A multi-channel battery testing system at room temperature. The cyclic voltammetry (CV) experiment was performed in the range of 1.7–2.8 V at a sweep rate of 0.1 mV s⁻¹ and the electrochemical impedance spectroscopy (EIS) measurements were conducted over frequency ranges from 700 kHz to 100 mHz on a BioLogic VMP3 station. The specific capacity was calculated based on active sulfur materials with 1675 mA g⁻¹.

Results and discussion

Scheme 1 depicts the synthesis process of the S-GN/PNCNTs composites. Firstly, PPy is coated on the surface of MWCNTs to form the precursor PCNTs by chemical polymerization of pyrrole monomer. Secondly, urea (as carbon and nitride sources) and composite PCNTs are mixed together and sintered. At the first stage of 600 °C, g-C₃N₄ is manufactured accompanying with by-product gases to produce pores. With rising to a higher pyrolytic temperature at the second stage, the carbonyl groups on urea react with amino groups on pyrrole to form the g-C₃N₄ layer and chemically grafted on the structure of PCNTs. Finally, the as-prepared composite GN/PNCNTs and sulfur are mixed and heated to 155 °C (N₂ atmosphere) for 12 h.

Scheme 1 The scheme of the formation process of the S-GN/PNCNTs composite.

The scanning electron microscope (SEM) image and transmission electron microscope (TEM) image in Fig. 1a and b show that the PCNTs are uniform nanotube structure with outer diameter of around 100 nm and inner channel of 10 nm (Fig. 1b). After the pyrolysis process, the surfaces have significant changes with an amorphous graphite-like layer on the similar tube morphology (Fig. 1c). The TEM image (Fig. 1d) of the GN/PNCNTs composite can be seen the existence of tube structure, and further insight into the high-resolution TEM (HRTEM) image clearly exhibits the amorphous coating layer with 5–10 nm and the interlayer spacing of 0.34 nm.¹⁸ The EDS maps of the GN/PNCNTs are performed in Fig. S4† and the C/N stoichiometric ratio on the surface of the GN/PNCNTs is about 1.09. Due to the high carbon content of the MWCNTs composite, the C/N stoichiometric ratio is higher than the theoretical data 0.75 of g-C₃N₄. In Fig. 1e and f, the TEM images present that the GN/PNCNTs can be used as N-doped carbon host to encapsulate sulfur and sulfur is distributed into the tubes without any agglomeration. The element mapping images of S-GN/PNCNTs composite further reveal that the sulfur is distributed homogeneously (Fig. 1g).

Fig. 1 SEM and TEM images of (a and b) PCNTs, (c and d) GN/PNCNTs and (e and f) S-GN/PNCNTs composites, EDS maps of the S-GN/PNCNTs composite.

As depicted in Fig. 2a, the GN/PNCNTs obtains a high BET specific surface area of 211.56 m² g⁻¹ and a total pore volume of 0.885 cm³ g⁻¹ (p/p₀ = 0.9734), which is larger than those of the commercial MWCNTs (<200 m² g⁻¹). According to the theoretical analysis, this total pore volume can host as much as 65 wt% sulfur (ρ_s = 2.07 g cm⁻³). Fig. 2b presents the pore size distribution of the GN/PNCNTs composite, and the ratio of mesopores structure turns out to be around 100%, which can effectively hamper the dissolution of polysulfide and sulfur for a longer cycling stability.¹⁹ Furthermore, the average pore width of the GN/PNCNTs composite is 10.1 nm according to the BJH method. After the sulfur impregnated into the host, the S-GN/PNCNTs composite still obtained a BET specific surface area of approximately of 20 m² g⁻¹ and a pore volume of 0.01 cm³ g⁻¹, which indicated 60 wt% sulfur into the GN/PNCNTs composite could be achieved (Fig. S1†). The large surface areas endow the composite with large electrode–electrolyte contact area and more lithium storage sites to relieve the stress caused by volume change during the charge–discharge process.

Fig. 2 (a and b) The N₂ adsorption–desorption curves and the pore width distribution of the GN/PNCNTs composite, (c) the TGA curves of the S-GN/PNCNTs composite, (d) the XRD patterns of the PCNTs, GN/PNCNTs and S-GN/PNCNTs composites.

TGA curve (Fig. 2c) further confirms the weight changes after sulfur impregnated into the GN/PNCNTs composite. The weight loss before 300 °C is considered to be the evaporation of sulfur. Thus the sulfur content of the S-GN/PNCNTs sample can be calculated as approximately 60 wt%, which is consistent with the experiment result.

XRD patterns of the PCNTs, GN/PNCNTs and S-GN/PNCNTs composites are presented in Fig. 2d. The peaks around 25° of the PCNTs, GN/PNCNTs composites present that the existence of the amorphous carbon. One peak around 26.5° of the GN/PNCNTs composite, indexed as g-(002) peak, corresponds to the interlayer graphitic packing motif of aromatic segments and be consistent with reported literature.^20,21 It is worth noting that the weakening and shift of this peak is partly caused by carbon species of MWCNTs.²² The pattern of S-GN/PNCNTs composite exhibits sharp diffraction peaks matched well with the standard diffraction lines of sulfur (JCPDS: 08-0247), which can be indexed to the orthorhombic phase.²³

Fig. 3a shows the XPS survey spectrums of PCNTs, GN/PNCNTs and S-GN/PNCNTs composites. It can be seen three peaks around 284.5 ± 0.3, 400.0 ± 0.3 and 533 ± 0.3 eV, corresponding to C 1s, N 1s and O 1s, respectively. The peak centering at 284.5 ± 0.3 eV is corresponding to the graphitic carbon, and the other two types of peaks are identified as sp² C bonded with N or O (C–N/C–O) at 286.3 ± 0.3 eV, and aromatic C bonded with N or O (C=N/C=O) at 288.2 ± 0.3 eV (Fig. 3b), which provide evidence for the presence of different chemical states of carbon in the S-GN/PNCNTs composite.^24,25 The peaks in Fig. 3c centering at 163.5 ± 0.3 and 164.5 ± 0.3 eV correspond to S 2p_3/2 and S 2p_1/2 components, and the peak around 167.5 ± 0.3 eV is attributed to the oxidized sulfur form of the C–SO_x–C (x = 2–4) bond.²⁶ Fig. 3d–f show the nitrogen environment of the PCNTs, GN/PNCNTs, S-GN/PNCNTs composites, for the bonding states of nitrogen, four main types of nitrogen peaks are identified and correspond to pyridinic N at 398.2 ± 0.3 eV, pyrrolic N at 400.2 ± 0.3 eV, quaternary N at 401.3 ± 0.3 eV and pyridine oxide at 402.7 ± 0.3 eV.²⁷ And the ratios of pyridinic N, pyrrolic N, quaternary N and pyridine oxide for high resolution N 1s peaks are summarized in Table 1. For the nitrogen bonding state of PCNTs composite, the pyrrolic N in a C₅ ring possesses the largest concentration, which is ascribed to the PPy chains on MWCNTs.²⁸ However, after the pyrolysis, the heterogeneity of GN/PNCNTs composite is higher, and pyridinic N and quaternary N predominate. As shown in Scheme 2, pyridinic N (C=N–C) with two sp² hybridized C neighbors replaces a carbon atom in the C₆ ring.²⁹ Quaternary N with three sp² carbon neighbors replaces a carbon atom and has two valence electrons.¹¹ Hence, the excess electrons contributed by the nitrogen atoms can produce more N-doping effect (compared to one electron per carbon atom and the just pyrrolic N doping).^30,31 After introducing sulfur species, the ratios of the four N species change obviously, verifying the interaction between sulfur and g-C₃N₄.

Fig. 3 (a) XPS spectrums of the PCNTs, GN/PNCNTs and S-GN/PNCNTs composites, (b and c) C 1s, S 1s spectrums of the S-GN/PNCNTs composite, respectively, (d–f) high resolution N 1s peaks of the PCNTs, GN/PNCNTs and S-GN/PNCNTs composites.

Table 1 The percentage of core level peaks for the high resolution N 1s peaks of the PCNTs, GN/PNCNTs and S-GN/PNCNTs composites

Samples	Pyridinic N (%)	Pyrrolic N (%)	Quaternary N (%)	Pyridine oxide (%)
	398.2 ± 0.3 eV	400.2 ± 0.3 eV	401.2 ± 0.3 eV	402.7 ± 0.3 eV
PCNTs	—	80.4	19.6	—
GN/PNCNTs	38.2	8.4	43.2	10.2
S-GN/PNCNTs	21.2	—	54.6	24.2

---

Scheme 2 (a) Structure of graphitic g-C₃N₄, (b) scheme structure of GN/PNCNTs composite.

Fig. 4 shows CV curves of the S-GN/PNCNTs and S-PCNTs cathodes at the second cycle between 1.7 V and 2.8 V at the scan rate of 0.1 mV s⁻¹. The CV curve of the S-GN/PNCNTs cathode presents two reduction peaks and one oxidation peak.³² The first reduction peak at ∼2.35 V is the element sulfur (S₈) reduction to the long-chain polysulfide (S_n²⁻, 4 ≤ n ≤ 8). The second reduction peak at ∼2.05 V indicates the further reduction of long-chain polysulfide (S_n²⁻, 4 ≤ n ≤ 8) to the S₂²⁻/S²⁻. During the anodic sweeping, the oxidation peak at ∼2.38 V indicates the reversible transformation from S₂²⁻/S²⁻ to the element sulfur (S₈).³³ This reaction process is consistent with the conventional redox reaction between S₈ and Li as its CV curve indicated.³⁴ As a comparison, the curve of the S-PCNTs cathode shows a significant different behavior. The voltage gap between reduction and oxidation peaks of the S-GN/PNCNTs cathode remains smaller than that of the S-PCNTs cathode, demonstrating the fast kinetics in the surface oxidation reaction, and thus a high redox electrochemical reversibility within this electrode.

Fig. 4 The 2nd CV curves between 1.7 and 2.8 V at the scan rate of 0.1 mV s⁻¹.

Fig. 5a and b show the discharge/charge curves of the initial and 200^th cycle of the S-GN/PNCNTs, and S-PCNTs cathodes at the current density of 0.5C. In the initial cycle, their discharge specific capacities are 1153 mA h g⁻¹ (with a coulombic efficiency of about 100%), 1254 mA h g⁻¹, respectively. For S-PCNTs cathode, a sufficient contact of the active material with the electrolyte greatly contributes to a higher capacity. After 200 cycles, the capacity of S-PCNTs cathodes fades fast and has an average capacity decay of about 0.294% per cycle, while the S-GN/PNCNTs cathode just has a capacity decay of 0.241% per cycle. Compared to the S-PCNTs cathode, the longer voltage plateau and smaller voltage polarizations (ΔE) of the S-GN/PNCNTs cathode after 200 cycles, are consistent with a reduced impendence in Fig. 7. In addition, the cycling performances of the S-GN/PNCNTs and S-PCNTs cathodes at the current density of 0.5C (Fig. 5c) further demonstrate the relatively higher stability of the S-GN/PNCNTs cathode. The great rate capability of the S-GN/PNCNTs cathode is exhibited in Fig. 5d. The reversible capacity is found to be 1178 mA h g⁻¹ in the initial cycle at 0.5C. Further cycling at 1C, 2C, 4C, 6C keeps a reversible capacity of 686 mA h g⁻¹, 632 mA h g⁻¹, 603 mA h g⁻¹, 586 mA h g⁻¹, separately. When the rate switches to 1C and 0.2C, the capacities are largely recovered to 686 mA h g⁻¹ and 732 mA h g⁻¹, respectively. The S-GN/PNCNTs cathode still has a good cycling stability after 600 cycles at 1C and has just an average capacity decay of about 0.08% per cycle with coulombic efficiency close to 100% during cycling (Fig. 6).

Fig. 5 (a and b) The initial and 200^th discharge/charge curves, (c) the cycling performance of the S-PCNTs and S-GN/PNCNTs electrodes at 0.5C, respectively, (d) the rate performance and coulombic efficiency of the S-GN/PNCNTs electrode.

Fig. 6 The cycling performance and coulombic efficiency of the S-GN/PNCNTs electrode at 1C.

Fig. 7 (a and b) The EIS spectrums of the S-PCNTs and S-GN/PNCNTs electrodes of the fresh cells and the 200^th cycle cells, respectively.

To further investigate the electrochemical reaction mechanism, the EIS results of the S-GN/PNCNTs and S-PCNTs cathodes at fully charged state for the initial and 200^th cycle are presented in Fig. 7. The resistance of both two cells decreases after 200 cycles, which is attributed to active material occupying a more electrochemically favorable position. Hence, a closer contact and better coverage of active material on the composite can be achieved.³⁵ The charge transfer resistance (R_ct, the intersection of the semicircle on the real axis) of the S-GN/PNCNTs cathode remains much lower than that of the S-PCNTs cathode after cycling, which may be due to slower accumulation of Li₂S on the interface and be consistent with the slight decrease in the capacity during cycling.³⁶

The splendid performances of the S-GN/PNCNTs cathode can be attributed as follows: firstly, the PCNTs provide electronic channels and a specific surface area to host sulfur; secondly, the coated PPy on PCNTs can enhance the transportation of the electrons, which can lower the resistance of the electrode to some extent, and the high nitrogen content can provide more defects of the structure to enhance the interaction between carbon and sulfur.³⁷ Last but not the least, the coated graphitic N provide more reactive sites and bring uniform distribution of discharge products to accommodate the volume variations to prevent the loss of the active materials.²⁰ All these effects can retard the dissolution of polysulfide to the electrolyte, decrease the shuttle effect, and enhance the cycling stability and specific capacity.

Conclusions

In summary, the GN/PNCNTs composite is synthesized via the pyrolysis process, of which g-C₃N₄ improves heterogeneity of nitrogen and creates more carbon vacancies on the tubes. Liberated electrons of the nitrogen atoms can contribute to the system (compared to one electron per carbon atom). Hence, the GN/PNCNTs composite can enhance the interaction between sulfur species and carbon structures. Used as a sulfur host for the Li–S battery, the S-GN/PNCNTs cathode shows an excellent cycling stability and a great rate performance.

Acknowledgements

This work is supported by the Creative Research Groups of the National Natural Science Foundation of China (Grant No. 21221061).

References

1. Y. Z. Zhang, K. E. Sun and P. Chen, Open Mater. Sci. J., 2011, 5, 217.
2. Y. V. Mikhaylik and J. R. Akridge, J. Electrochem. Soc., 2004, 151, A1969.
3. H. Chen, C. Wang, W. Dong, W. Lu, Z. Du and L. Chen, Nano Lett., 2015, 15, 798.
4. R. Demir-Cakan, M. Morcrette, Gangulibabu, A. Guéguen, R. Dedryvère and J.-M. Tarascon, Energy Environ. Sci., 2013, 6, 176.
5. X. Liang, M. G. Zhang, M. R. Kaiser, X. W. Gao, K. Konstantinov, R. Tandiono, Z. X. Wang, H. K. Liu, S. X. Dou and J. Z. Wang, Nano Energy, 2015, 11, 587.
6. L. Nyholm, G. Nystrom, A. Mihranyan and M. Stromme, Adv. Mater., 2011, 23, 3751.
7. M. Kazazi, M. R. Vaezi and A. Kazemzadeh, Ionics, 2014, 20, 635.
8. G. Wang, L. Zhang and J. Zhang, Chem. Soc. Rev., 2012, 41, 797.
9. M. Tahir, N. Mahmood, J. Zhu, A. Mahmood, F. K. Butt, S. Rizwan, I. Aslam, M. Tanveer, F. Idrees, I. Shakir, C. Cao and Y. Hou, Sci. Rep., 2015, 5, 12389.
10. W. B. Luo, S. L. Chou, J. Z. Wang, Y. C. Zhai and H. K. Liu, Small, 2015, 11, 2817.
11. H. Pan, J. Phys. Chem. C, 2014, 118, 9318.
12. Y. Gong, M. Li and Y. Wang, ChemSusChem, 2015, 8, 931.
13. J. Liu, T. Zhang, Z. Wang, G. Dawson and W. Chen, J. Mater. Chem., 2011, 21, 14398.
14. Y. Zhang, J. Liu, G. Wu and W. Chen, Nanoscale, 2012, 4, 5300–5303.
15. H. Yu, L. Shang, T. Bian, R. Shi, G. I. Waterhouse, Y. Zhao, C. Zhou, L. Z. Wu, C. H. Tung and T. Zhang, Adv. Mater., 2016, 398.
16. F. Dong, L. Wu, Y. Sun, M. Fu, Z. Wu and S. C. Lee, J. Mater. Chem., 2011, 21, 15171.
17. J. Song, T. Xu, M. L. Gordin, P. Zhu, D. Lv, Y.-B. Jiang, Y. Chen, Y. Duan and D. Wang, Adv. Funct. Mater., 2014, 24, 1243.
18. J. W. Liu, X. Y. Chen, W. C. Yu, X. M. Liu and Y. T. Qian, J. Am. Chem. Soc., 2003, 125, 8088.
19. H. Wei, Y. Wang, J. Guo, N. Z. Shen, D. Jiang, X. Zhang, X. Yan, J. Zhu, Q. Wang, L. Shao, H. Lin, S. Wei and Z. Guo, J. Mater. Chem. A, 2015, 3, 469.
20. J. Yi, K. Liao, C. Zhang, T. Zhang, F. Li and H. Zhou, ACS Appl. Mater. Interfaces, 2015, 7, 10823.
21. J. Zhang, M. Zhang, L. Lin and X. Wang, Angew. Chem., Int. Ed., 2015, 54, 6297.
22. K. Ariga, T. Mori, T. Nakanishi, S. Hishita, D. Golberg, Y. Bando and A. Vinu, Adv. Mater., 2005, 17, 1648.
23. F. F. Zhang, G. Huang, X. X. Wang, Y. L. Qin, X. C. Du, D. M. Yin, F. Liang and L. M. Wang, Chem. –Eur. J., 2014, 20, 17523.
24. X. Y. Zhou, F. Chen and J. Yang, J. Energy Chem., 2015, 24, 448.
25. H. Long du, M. G. Jeong, Y. S. Lee, W. Choi, J. K. Lee, I. H. Oh and H. G. Jung, ACS Appl. Mater. Interfaces, 2015, 7, 10250.
26. Q. Li, Z. Zhang, Z. Guo, K. Zhang, Y. Lai and J. Li, J. Power Sources, 2015, 274, 338.
27. Q. Guo, Y. Xie, X. Wang, S. Zhang, T. Hou and S. Lv, Chem. Commun., 2004, 26,  10.1039/b311390f.
28. Y. Wu, C. Guo, N. Li, L. Ji, Y. Li, Y. Tu and X. Yang, Electrochim. Acta, 2014, 146, 386.
29. X.-J. Wang, X. Tian, F.-T. Li, J. Zhao, Y.-P. Li, R.-H. Liu and Y.-J. Hao, Dalton Trans., 2015, 1.
30. Y. Qiu, W. Li, W. Zhao, G. Li, Y. Hou, M. Liu, L. Zhou, F. Ye, H. Li, Z. Wei, S. Yang, W. Duan, Y. Ye, J. Guo and Y. Zhang, Nano Lett., 2014, 14, 4821–4827.
31. Q. Pang and L. F. Nazar, ACS Nano, 2016, 10, 4111–4118.
32. X. Ji and L. F. Nazar, J. Mater. Chem., 2010, 20, 9821.
33. S. S. Zhang, J. Power Sources, 2013, 231, 153.
34. R. Chen, T. Zhao and F. Wu, Chem. Commun., 2015, 51, 18.
35. S.-H. Chung and A. Manthiram, Electrochim. Acta, 2013, 107, 569.
36. J. Yang, J. Xie, X. Y. Zhou, Y. L. Zou, J. J. Tang, S. C. Wang, F. Chen and L. Y. Wang, J. Phys. Chem. C, 2014, 118, 1800.
37. S. H. Chung and A. Manthiram, Adv. Mater., 2014, 26, 1360.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra11898d

---