Suppressing self-discharge of Li–B/CoS₂ thermal batteries by using a carbon-coated CoS₂ cathode

Abstract

Thermal batteries with molten salt electrolytes are used for many military applications, primarily as power sources for guided missiles. The Li–B/CoS₂ couple is designed for high-power, high-voltage thermal batteries. However, their capacity and safe properties are influenced by acute self-discharge that results from the dissolved lithium anode in molten salt electrolytes. To solve those problems, in this paper, carbon coated CoS₂ was prepared by pyrolysis reaction of sucrose at 400 °C. The carbon coating as a physical barrier can protect CoS₂ particles from damage by dissolved lithium and reduce the self-discharge reaction. Therefore, both the discharge efficiency and safety of Li–B/CoS₂ thermal batteries are increased remarkably. Discharge results show that the specific capacity of the first discharge plateau of carbon-coated CoS₂ is 243 mA h g⁻¹ which is 50 mA h g⁻¹ higher than that of pristine CoS₂ at a current density of 100 mA cm⁻². The specific capacity of the first discharge plateau at 500 mA cm⁻² for carbon-coated CoS₂ and pristine CoS₂ are 283 mA h g⁻¹ and 258 mA h g⁻¹ respectively. The characterizations by XRD and DSC indicate that the carbonization process has no noticeable influence on the intrinsic crystal structure and thermal stability of pristine CoS₂.

---

Introduction

Thermal batteries are chemical power supplies which use molten salts as electrolytes and operate in the range from 350 °C to 550 °C. They are activated by employing an internal pyrotechnic source to produce the heat and melt salt electrolytes. The excellent ionic conductivity of molten salts means thermal batteries can work at high output power. Thermally batteries are mainly used as power sources for guided weapon systems due to their harsh-environment endurance, fast-activation and long storage times.^1,2 Metal sulfides with outstanding conductivity have been widely applied to lithium ion batteries, supercapacitors or lithium–sulfur batteries.^3,4 One of them is CoS₂ which is becoming an important cathode material for thermal batteries. Compared with other metal sulfides like FeS₂ and NiS₂, it has the lowest solubility in most molten electrolytes and the highest electronic conductivity. Furthermore, its thermal decomposing temperature is up to 650 °C, which is of great significance for the safety properties of thermal batteries.

Lithium alloys have been used as anode materials for thermal batteries, such as Li–Al, Li–Si, and Li–B alloy. To match with the excellent electrochemical properties of CoS₂ cathode, Li–B alloy anode which has good power property for its outstanding Li activity, is considered to be the best choice. Different from conventional LiCl–KCl eutectic electrolyte, LiF–LiCl–LiBr molten salt system has an ionic conductivity up to 6.52 S cm⁻¹.^5,6 Thus the Li–B/LiF–LiCl–LiBr/CoS₂ couple is now the most suitable assembly for the high-power output thermal batteries. However, the free lithium in Li–B alloy is very easy to dissolve into the electrolytes. Then the dissolved lithium diffuses to the cathode and reacts with cathode materials once thermal batteries are activated, which results in serious self-discharge at discharge. Above phenomenon can be related to the rule that alkali-metal had a considerable solubility in molten alkali halide.⁷ The solubility of lithium at LiF–LiCl–LiBr eutectic is between 1 mol% and 2 mol%, which was measured by Watanabe N.⁸ The Li activity of the anode has a dramatic influence on increasing the dissolving process and Li–B alloy has much higher Li activity than Li–Si alloy and Li–Al alloy. Therefore the capacity of Li–B/CoS₂ cells is more greatly influenced by the severe self-discharge especially in the case of long operating life thermal batteries. Self-discharge rate of Li–B/LiF–LiCl–LiBr/CoS₂ cells measured by Gao was up to 20 mA cm⁻²,⁹ which is almost times more than Li–Si/CoS₂ cells. In addition, a drastic exothermic reaction accompanied with self-discharge reactions, may destroy the battery and cause serious safety issues. So far, performances of Li–B/CoS₂ batteries are still limited by self-discharge.

Various strategies have been employed to address the self-discharge caused by dissolved lithium in Li–B/LiF–LiCl–LiBr/CoS₂ cells. Zeng reported that applying additive of nanometal powder to LiF–LiCl–LiBr electrolytes could effectively suppress the dissolution of lithium anode.¹⁰ Different from Zeng, preventing direct contact between dissolved lithium and active cathode materials through surface modification may be another promising strategy in solving this problem. Similar strategies have been applied to suppress the harmful interaction between electrode materials and electrolytes for lithium ion batteries.^11–13 The alumina coating for lithium cobalt fluorophosphate was reported to prevent metal ion dissolution and improve the cycle performance.¹⁴ A widely recognized interpret is that the alumina coating can protect cathode materials from hazardous reaction products such as HF obtained during the charge–discharge process.^15,16 Petnikota has modified FeO, MnO and Co₂Mo₃O₈ by coating of graphene oxide for anode application in lithium ion batteries. All of those composites exhibit better electrochemical properties than corresponding pristine batteries.^17–19 Another example is that amorphous carbon coating is used to prevent PC-based electrolytes insert into graphite accompanied by graphite exfoliation.^20,21 For researches of thermal batteries that anode is Li–Si alloy, carbon coating to CoS₂ cathode was reported by Xie to promote electrochemical properties by enhancing the electronic conductivity.²²

In this paper, considering the higher self-discharge rate at Li–B/CoS₂ system, the carbon coating of CoS₂ is mainly aimed to block dissolved lithium atom react with CoS₂ cathode. Based on above hypothesis, we reported a facile method through pyrolysis reaction of sucrose, to modify CoS₂ with amorphous carbon coating (this composite of carbon and CoS₂ is referred as CoS₂/C hereafter). It has been proved that CoS₂/C was suitable for thermal batteries and successful in reducing self-discharge rate at Li–B/CoS₂ system. Discharge tests indicated that CoS₂/C released higher capacity than pristine CoS₂. Fig. 1 illustrates how the carbon coating influences self-discharge as a physical barrier.

Fig. 1 The influence of modifying CoS₂ cathode materials by the carbon coating at thermal batteries.

Experimental

Carbon-coated CoS₂ preparation

CoS₂ was powder from State Key Laboratory of Powder Metallurgy which was synthesized by solid-state reaction of cobalt and sulphur powder.⁸ To prepare carbon-coated CoS₂, 1 g sucrose (C₁₂H₂₂O₁₁) was dispersed in 2 mL pure water. Then, 9 g CoS₂ and sucrose solution were ground together for 1 h to make it been mixed evenly. After drying, the mixture was carbonized in the argon atmosphere for 4 h after heating to 400 °C at a rate of 5 °C min⁻¹.^23–25

Single cell preparation

Anodes were Li–B alloys strip which were provided by State Key Laboratory of Powder Metallurgy, and synthesized by smelting metal lithium and boron.²⁶ Mass fraction of lithium for Li–B alloy is 61 wt%. It was punched to disk with a diameter of 17.5 mm. Electrolyte were composed by 50 wt% LiF–LiCl–LiBr eutectic salts (9.2 wt% LiF, 22 wt% LiCl, 68.4 wt% LiBr) and 50 wt% Nano-MgO binder. Cathodes include 80 wt% CoS₂ or CoS₂/C and 20 wt% electrolytes.

The cathode and separator were shaped to stratiform slice together by spreading the corresponding powders in a die, and then suppressing them under a static compaction pressure of 250 MPa. Two types of single cells were prepared to study the electrochemical properties of anode and cathode respectively. Single cells that the anode was superfluous were composed of 0.20 g anode, 0.36 g separator and 0.18 g cathode. While single cells with superfluous cathode were composed of 0.09 g anode, 0.36 g separator and 0.40 g cathode. Single cells with CoS₂/C and pristine CoS₂ cathode were abbreviated as CoS₂/C cell and CoS₂ cell respectively hereafter.

Materials characterization

Microscopic morphology studies were carried by the scanning electron microscope (LEO 1530 Gemini, Zeiss, Germany) at 15 kV. Crystalline structures of samples were characterized by X-ray diffractometer (D/max2550PC, Rigaku, Japan) at a rate of 8° min⁻¹ from 10° (2θ) to 90° (2θ) with Cu Kα radiation. Differential scanning calorimetry (DSC) was conducted by a thermal analyzer system (Hengjiu, Beijing) with a heating rate of 10 °C min⁻¹ and a 50 ml min⁻¹ Ar flow rate. The carbon coating was observed by transmission electron microscopy (Tecnai G2, FEI, USA). Raman spectroscopies were completed by (T64000, Horiba, Japan) and 514 nm light was used for excitation with an intensity of 20 mW. Particle size distribution was measured by laser particle size analysis (Mastersizer3000, Malvern, England).

Electrochemical measurements

Anode, separation-cathode disk were fabricated to single cell. The test was conducted at 520 °C with a fluctuation of 5 °C. Single cells were discharged at constant current and pulse current respectively. In pulse discharge, the current densities were ranged from 100 mA cm⁻² to 600 mA cm⁻² for 2 s every 20 s to study total polarization and power properties. Discharge data were obtained by electronic load (ITECH, IT8511, USA). Total polarization was calculated by following formula.²⁷

R_total = ΔU/ΔI

here R_total denotes the total polarization of a single cell. ΔU denotes the voltage drop caused by pulse current. ΔI denotes the difference between pulse current and steady current.

Results and discussion

X-ray diffraction patterns of both CoS₂/C and pristine CoS₂ are shown in Fig. 2a. All peaks of pristine CoS₂ and CoS₂/C can be indexed to a pure-phase, and no apparent differences appear between CoS₂ and CoS₂/C patterns, except the whole diffraction intensity. It indicates that the coating process has no affection on the crystal structure of pristine CoS₂. Besides, no peaks involved to carbon are observed in diffraction patterns of CoS₂/C. Fig. 2b is the results of XRD analysis of Li–B alloy anode. It shows that Li–B alloy was mainly made up of free lithium metal and LiB compound. According to past studies, no free lithium metal was appeared at Li–Si or Li–Al alloy anode. It is the free lithium metal that makes the Li–B alloy has much more activity than that of Li–Si or Li–Al alloy.

Fig. 2 XRD patterns of CoS₂/C and pristine CoS₂ (a). XRD patterns of Li–B alloy anode (b).

Fig. 3a and b display SEM images of pristine CoS₂ and CoS₂/C respectively. Both pristine CoS₂ and CoS₂/C show aggregation which is sticked by tiny spherical particles. There are gaps and holes between tiny spherical particles, which can provide the passageway for molten electrolytes at the process of discharge. Compared Fig. 3a and b, CoS₂/C has bigger granularity than pristine CoS₂. This phenomenon can be explained by bonding of sucrose. Since the melting point of sucrose is 186 °C,²⁸ sucrose will be heated to the liquid state before the carbonized stage and CoS₂ particles may be bonded by viscous molten sucrose. The changes of granularity observed by SEM are also identical to the consequence of particle size analysis. Median particle diameter (D50) of pristine CoS₂ is increased from 19.7 μm to 26.8 μm after coating. Fig. 3c and d show the TEM micrographs of CoS₂/C particles. There is a clear boundary between the core of CoS₂ and the bright translucent shell materials of the carbon. The thickness is estimated to be 6 nm. The combination between CoS₂ and carbon is proved to be complete, only in this way can carbon shell protect CoS₂ particles effectively.

Fig. 3 SEM images of pristine CoS₂ (a) and carbon-coated CoS₂ (b). TEM images of carbon-coated CoS₂ (c, d).

Fig. 4a exhibits the Raman spectroscopy of CoS₂/C. The peaks at 1590 cm⁻¹ and 1358 cm⁻¹ are commonly referred as G peak and D peak.²⁹ G peak is corresponding to a Raman active E_2g mode of two-dimensional graphite layer, while D peak is attributed to a zone boundary phonon activated by disordered graphite or glass carbon.³⁰ The ratio of intensity between D peak and G peak (I_D/I_G) is 0.65. A general expression that gives the crystallite size (La) of graphite from the intensity ratio I_D/I_G is given by

where E_l (2.41 eV) is the excitation laser energy used in the Raman experiment.³¹ The La of the carbon coating calculated by this formula is 25.5 nm. It indicates that the carbon coating has considerable graphitic form and desirable electronic conductivity, which is necessary for the high power property at thermal batteries. In addition, no peak related to the C=C fundamental stretching vibration of S₂C=CS₂ configuration is observed at around 1445 cm⁻¹. It reveals that no polymer is produced among interaction of CoS₂ and sucrose.

Fig. 4 Raman spectroscopy of CoS₂/C (a). Differential scanning calorimeter of pristine CoS₂ and CoS₂/C (b).

Fig. 4b gives results of differential scanning calorimeter of pristine CoS₂ and CoS₂/C. Good thermal stability of cathode materials is a significant requirement at thermal batteries. Sulfide cathode materials with bad thermal stability may decompose and produce sulfur steam because the instantaneous high temperature at the beginning of activation,³² which leads fatal safe issues and loss of capacity. Only one endothermic peak related to decomposition of CoS₂ is observed in both curves. The endothermic peak of pristine CoS₂ is 724.3 °C, and that of CoS₂/C is 723.1 °C. In addition, approximate endothermic onset temperature that sample become decompose are pointed at two curves. DSC date can be inferred that the carbon coating has no apparent influence on the thermal stability of pristine CoS₂.

According to the past studies, the discharge reactions of CoS₂ included three steps which are expressed by following chemical reaction equations.^1,33

(1 − x)CoS₂ + (2 − 4x)e⁻ → Co_(1−x)S + (1 − 2x)S⁻² (1)

Co₉S₈ + 16e⁻ → 9Co⁰ + 8S⁻² (3)

Correspondences of reaction equations are three plateau voltages. V. A. Bryukin inferred that x was ranged from 0.110 to 0.124.³⁴ calculated by the theories of V. A. Bryukin, the specific capacity of first discharge plateau is 374–382 mA h g⁻¹, that of second and third discharge plateau are 102–110 mA h g⁻¹ and 348 mA h g⁻¹ respectively.

Once a thermal battery is activated at the high temperature, self-discharge caused by dissolved lithium anode is remarkable. Thus, on the open circuit, the electromotive force (EMF) of the single cell will be decreased as time goes on due to active electrode materials are lost gradually by self-discharge reaction. Fig. 5a clarifies the relation between EMF and time of the battery cells fabricated with CoS₂ and CoS₂/C cathodes on the open circuit at 520 °C (single cells are designed with superfluous anode materials). EMF-time curves in the Fig. 5a reveal that active cathode materials are exhausted at last. There is no doubt that EMF of CoS₂/C cell is decreased more slowly than that of pristine CoS₂ cell. It can be concluded that CoS₂/C cell has lower self-discharge rate than pristine CoS₂ cell. Multiple EMF plateaus in the Fig. 5a represent phase transition of cathode materials. According to discharge process of CoS₂, three open-circuit plateau voltages ordered from the most to the least are related to phases of CoS₂, Co_1−xS, Co₉S₈ respectively. Fig. 5b shows differential capacity plots of cells. Three peaks located at 1.93 V, 1.80 V, 1.57 V in differential capacity plots are also very consistent with open circuit voltage curves. The EMF of the reaction from CoS₂ to Co_1−xS has been depressed by carbon coating.

Fig. 5 The voltage of single cells as function of time on open circuit at 520 °C (a). Differential capacity plots of CoS₂/C cell and pristine CoS₂ (b).

In order to further validate above deductions, X-ray diffraction was applied for analyzing the composition of the cathode part which has experienced an open circuit losses. To prepare appropriate samples for XRD, CoS₂/C cell and pristine CoS₂ cell were laid on open circuit for 40 min at 520 °C respectively. Then, cathode parts were rinsed with distilled water to remove electrolytes which will produce obvious interference signal. Fig. 6 displays XRD results of remainder active cathode materials which are obtained during open circuit. Diffraction spectrums show that the major phases of two samples are CoS₂ and Co_1−xS. Peaks located at 32.39° (P1) and 35.18° (P2) are contributed to CoS₂ and Co_1−xS respectively. The mass ratio of CoS₂ and Co_1−xS on the samples is directly proportional to intensity ratio of P1 and P2 (P1/P2), and the P1/P2 of CoS₂/C cell is 0.83, while for pristine CoS₂ is 0.14. Thus we can get the conclusion that there are less lost active materials on CoS₂/C cell. Those results are consistent with the analysis of the EMF curve in Fig. 5a. It should be noted that obvious peaks related to Co₉S₈ and Co₃S₄ appear at two patterns. Co₉S₈ phase may exist because of the non-uniform distribution of self-discharge reaction. A small amount of Co₃S₄ can be related to thermal decomposition of CoS₂.³⁵

Fig. 6 XRD pattern of cathode electrode materials after 40 min on open circuit losses at 520 °C.

Fig. 7 reports discharge curve of single cell that Li–B alloy anode is superfluous at 100 mA cm⁻² (a) and 500 mA cm⁻² (b). It can be used to study electrochemical properties of cathode materials. Because only the capacity of first discharge plateau for CoS₂ is utilized at practice application, specific capacity of first discharge plateau is regarded as standard to compare discharge capability of CoS₂ cathode. Fig. 7a reveals that the specific capacity of first discharge plateau for CoS₂/C is 243 mA h g⁻¹, which is 50 mA h g⁻¹ higher than of pristine CoS₂, and accounts for 64.9% of the theoretical capacity of CoS₂ (374 mA h g⁻¹). In Fig. 7b, the specific capacity for first discharge plateau of CoS₂/C and pristine CoS₂ are 283 mA h g⁻¹ and 258 mA h g⁻¹, respectively. CoS₂/C cells perform more flat and long first voltage plateau than pristine CoS₂ cell. The carbon coating has a significant effect in increasing specific capacity whether at 500 mA cm⁻² or 100 mA cm⁻². However, due to the loss of capacity caused by self-discharge will increase with discharge time, and completing discharge takes a longer time at 100 mA cm⁻² than at 500 mA cm⁻², CoS₂/C with lower self-discharge represents a more obvious advantage when single cells are discharged at the low current rate of 100 mA cm⁻².

Fig. 7 Specific discharge capacity of cobalt disulfide cathode, data of pristine CoS₂ cell or CoS₂/C cell at 100 mA cm⁻² (a) and 500 mA cm⁻² (b).

Sustained self-discharge reactions will consume lithium from anode. Thus self-discharge rate at cathode has a great influence on the specific capacity of Li–B alloy anode. To compare the specific capacity of Li–B alloy anodes which are matched with different cathodes, single cells that cathode is superfluous are discharged at 100 mA cm⁻² (Fig. 8a) and 500 mA cm⁻² (Fig. 8b). In the Fig. 8a, as a result of using CoS₂/C cathode, the specific capacity of Li–B anode is increased from 805 mA h g⁻¹ to 885 mA h g⁻¹ at first plateau. In the Fig. 8b, an increase of 51 mA h g⁻¹ also appears on first discharge plateau of Li–B alloy anode. There is no doubt that CoS₂/C cathode improves discharge efficient of Li–B alloy anode. The consequences of Fig. 8 provide further evidence that carbon coating on the CoS₂ has evident effect in depressing self-discharge to the Li–B/LiF–LiCl–LiBr/CoS₂ cells. Li–B alloy is made up of lithium and LiB compound, which results in first and second discharge plateaus respectively.³⁶ Besides, only lithium metal can dissolve into electrolytes and cause self-discharge. Therefore, CoS₂/C cathode only has positive influence on the first discharge plateau of Li–B alloy anode.

Fig. 8 Specific discharge capacity of Li–B alloy anode, data of pristine CoS₂ cell or CoS₂/C cell at 100 mA cm⁻² (a) and 500 mA cm⁻² (b).

In Fig. 7a and 8a, the first plateau voltage of CoS₂/C cell is equal to pristine CoS₂ cell at 100 mA cm⁻². In Fig. 7b and 8b, compared with pristine CoS₂ cell, however, a decrease about 0.4 V to the plateau voltage appears at CoS₂/C cell when the current density is increased to 500 mA cm⁻². Apparently, the carbon coating also has blocked diffusion of lithium ion, which produces additional concentration polarization for single cells at 500 mA cm⁻². The unique porous structure of amorphous carbon provide channels for the diffusion of lithium ion, which ensures enough lithium ion will diffuse to reaction interface in time at the low current density, but not at the high current density of 500 mA cm⁻². To know the detail of polarization increased by the carbon coating and estimate properties of single cell at transitory high current, single cells were discharged at pulse loading mode. Fig. 9a depicts pulse discharge curve. The total polarization was calculated by previous formula. Fig. 9b compares the difference of total polarization between CoS₂ cell and CoS₂/C cell. The resistances of both cells are increased with discharge time. Though CoS₂/C cell shows higher total polarization than CoS₂ cell especially when the voltage is higher than 1.65 V. CoS₂/C cell also exhibits excellent performances at pulse current, which meets the requirement of delivering electricity with high power for thermal batteries.

Fig. 9 Discharge performance at a background current of 100 mA cm⁻² and pulse current of 600 mA cm⁻² for 2 s every 20 s (a). Comparison of total polarization between CoS₂ cell and CoS₂/C cell (b).

Conclusions

In summary, to solve self-discharge problem at Li–B/LiF–LiCl–LiBr/CoS₂ system, we have modified CoS₂ with carbon coating by facile pyrolysis reaction of sucrose. Compare to the CoS₂/Li–B couple, thermal batteries fabricated with CoS₂/C and Li–B alloy can deliver much higher specific capacity. The enhanced discharge efficiency can be explained by the mechanisms that the carbon coating can prevent CoS₂ from directly exposing to the electrolytes and minimize the self-discharge reactions between the cathode and dissolved lithium. Besides, the carbonized process has no obvious influences on the thermal stability of pristine CoS₂. It has been proved that CoS₂/C was efficient and safe cathode materials. CoS₂/C has great merits and value particularly at the long-life thermal batteries for the low self-discharge rate. The method of modifying cathode materials with carbon may be applied to other cathode materials such as FeS₂ and NiCl for thermal batteries, all of those studies are underway.

Conflicts of interest

There are no conflicts of interest to declare.

Acknowledgements

This research work was financially supported by the National Natural Science Foundation of China (51771236), the Innovation-Driven Project of Central South University (No. 2017CX002).

References

1. R. A. Guidotti and P. Masset, J. Power Sources, 2006, 161, 1443–1449.
2. K. Nitta, S. Inazawa, S. Sakai, A. Fukunaga, E. Itani and K. Numata, SEI Tech. Rev., 2013, 76, 33–39.
3. P. Kulkarni, S. K. Nataraj, R. G. Balakrishna, D. H. Nagarajua and M. V. Reddy, J. Mater. Chem. A, 2017, 5, 22040–22094.
4. M. Chen, X. Yin, M. V. Reddy and S. Adams, J. Mater. Chem. A, 2015, 3, 10698–10702.
5. S. Fujiwara, M. Inaba and A. Tasaka, J. Power Sources, 2010, 195, 7691–7700.
6. R. A. Guidotti and P. Masset, J. Power Sources, 2008, 183, 388–398.
7. A. J. Burak and M. F. Simpson, JOM, 2016, 68, 1–7.
8. N. Watanabe, K. Nakanishi and A. Komura, Journal of the Society of Chemical Industry Japan, 1968, 71, 1599–1602.
9. J. X. Gao, Microcosmic characteristics and discharge properties of cobalt sulfide as cathode in thermal batteries, Central South University, 2014.
10. P. Zeng, Dissolution characteristics and discharge properties of LiB alloys as anode in thermal batteries, Central South University, 2016.
11. F. Wu, Q. Xue, L. Li, X. Zhang, Y. Huang and E. Fan, RSC Adv., 2017, 7, 1191–1199.
12. P. R. Kumar, V. Madhusudhanrao and B. Nageswararao, J. Solid State Electrochem., 2016, 20, 1855–1863.
13. M. V. Reddy, G. V. Subba Rao and B. V. Chowdari, Chem. Rev., 2013, 113, 5364–5457.
14. S. Amaresh, K. Karthikeyan and K. J. Kim, RSC Adv., 2014, 4, 23107–23115.
15. M. Seungtaek, I. Kentarou and K. Shinichi, J. Mater. Chem., 2005, 17, 3695–3704.
16. M. Yoshio, H. Wang and K. Fukuda, J. Electrochem. Soc., 2000, 147, 1245–1250.
17. S. Petnikota, S. K. Marka, A. Banerjee, M. V. Reddy, V. V. S. S. Srikanth and B. V. R. Chowdari, J. Power Sources, 2015, 293, 253–263.
18. S. Petnikota, V. V. S. S. Srikanth, P. Nithyadharseni, M. V. Reddy, S. Adams and B. V. R. Chowdari, ACS Sustainable Chem. Eng., 2015, 3, 3205–3213.
19. S. Petnikota, S. K. Marka, V. V. S. S. Srikanth, M. V. Reddy and B. V. R. Chowdari, Electrochim. Acta, 2015, 178, 699–708.
20. L. Wang, Z. Liu, Q. Guo, X. Guo and J. Gu, RSC Adv., 2017, 7, 36735–36743.
21. S. Petnikota, N. K. Rotte, V. V. S. S. Srikanth, B. S. R. Kota, M. V. Reddy and K. P. Loh, J. Solid State Electrochem., 2014, 18, 941–949.
22. S. Xie, Y. Deng and J. Mei, Electrochim. Acta, 2017, 231, 287–293.
23. S. Amaresh, K. Karthikeyan and I. C. Jang, J. Mater. Chem. A, 2014, 2, 11099–11106.
24. J. Dong, D. Li and Z. Peng, J. Solid State Electrochem., 2008, 12, 171–174.
25. J. Wang and X. Sun, Energy Environ. Sci., 2012, 5, 5163–5185.
26. Z. J. Liu, Z. Y. Li, W. Duan, X. H. Qu, B. Y. Huang and S. Q. Zhang, J. Mater. Sci. Technol., 2000, 16, 581–584.
27. S. Xie, Y. Deng, J. Mei, Z. Yang, W. M. Lau and H. Liu, Composites, Part B, 2016, 93, 203–209.
28. M. Hurtta, I. Pitkänen and J. Knuutinen, Carbohydr. Res., 2004, 339, 2267–2273.
29. J. Schwan, S. Ulrich and V. Batori, J. Appl. Phys., 1996, 80, 440–447.
30. A. C. Ferrari and J. Robertson, Phys. Rev. B: Condens. Matter Mater. Phys., 2008, 61, 14095–14107.
31. L. G. Cancado, K. Takai and T. Enoki, Appl. Phys. Lett., 2006, 88, 163106.
32. S. Serge, J. Power Sources, 2005, 142, 361–369.
33. Y. Wang, J. Wu, Y. Tang, X. Lu, C. Yang and M. Qin, ACS Appl. Mater. Interfaces, 2012, 4, 4246–4250.
34. V. A. Bryukvin and L. I. Blokhina, Nickel-Cobalt International 97 Symposium, Canada, August 1997.
35. M. A. Rodriguez, E. N. Coker, J. J. M. Griego, C. D. Mowry, A. S. Pimentel and T. M. Anderson, Powder Diffr., 2016, 31, 90–96.
36. P. Sanchez, C. Belin, G. Crepy and A. D. Guibert, J. Mater. Sci., 1992, 27, 240–246.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c7ra13071f

---