Simple preparation of Cu₆Sn₅/Sn composites as anode materials for lithium-ion batteries

Abstract

Cu₆Sn₅/Sn composites are directly fabricated by a high energy mechanical milling technique and subsequent heat treatment. In particular, the effects of the ratios of Sn to Cu₆Sn₅ (Cu_xSn_y, x = 10 − y, y = 4.5, 7 and 9) on the lithium-ion batteries performances are investigated. The results show that the sample with y = 4.5 is a single-phase Cu₆Sn₅, the sample with y = 7 is slightly Sn rich in the Cu₆Sn₅ (Sn < Cu₆Sn₅), and the sample with y = 9 is excessively Sn rich in the Cu₆Sn₅ (Sn > Cu₆Sn₅). Furthermore, the Cu₆Sn₅/Sn composite has an obvious structure of a core–shell, only when y = 7. As an anode material for lithium-ion batteries, the Cu₆Sn₅/Sn composite with y = 7 exhibits a discharge capacity of 761.6 mA h g⁻¹ after the first cycle, 457.8 mA h g⁻¹ after 20th cycles, and an initial coulombic efficiency of 91.37%, which shows a better electrochemical performance than that of y = 4.5 or 9. In addition, after adding 15 wt% of graphite, the sample with y = 7 maintains a discharge capacity of 605.8 mA h g⁻¹ after 100 repeated cycles, higher than many reported Cu–Sn-based anode materials.

---

1. Introduction

Currently, a mountain of research has been focused on developing alternative anode materials with higher energy density than commercial graphite.^1–3 Among them, tin-based materials have been regarded as one of the most promising alternative anode materials for lithium-ion batteries (LIBs) due to the high theoretical capacity, low potential of lithium-ion intercalation, and low price.^4–6 Unfortunately, the use of pure tin (Sn) as an anode material is limited because of its significant volume change during the lithiation and delithiation processes. This leads to electronic isolation of the lithiated product and poor electrical contact with the current collector. Besides, Li_xSn (x ≤ 4.4) alloys are brittle and pulverized easily by large Li-driven volume variation during the charge and discharge reactions.

In order to overcome the volume fluctuation, a promising approach is to distribute the Sn inside a matrix of Li-inactive phase (denote as M) and form M_xSn_y intermetallic compounds. Various M_xSn_y alloys, such as Cu₆Sn₅, CoSn₂, Ni₃Sn₄ and FeSn_2, etc., have been studied as candidate anode materials for LIBs.^7–11 The results indicate that the M_xSn_y electrode can provide a much more stable cycling performance than pure Sn electrode. In addition, the Cu₆Sn₅ intermetallic compound has been considered as the most promising alternative anode material because it is easy to be synthesized, stable in an air environment, and reversible two-phase lithium insertion/extraction between Cu₆Sn₅ and Li₂CuSn. Furthermore, the Cu₆Sn₅ intermetallic can encase the lithium/tin alloy in a conductive copper matrix. All of the advantages endow the Cu₆Sn₅ intermetallic with good lithium storage performances. For instance, the Cu₆Sn₅ alloy firstly prepared by Kepler et al. in 1999 showed a discharge capacity of ∼200 mA h g⁻¹ after 20 cycles.⁷ Then, He et al. prepared the Cu–Sn alloy by electrodeposition along with a heat-treatment, where the uncoated Cu–Sn alloy delivered a discharge capacity of 428 mA h g⁻¹ after 15 cycles while the Cu-coated one maintained 220 mA h g⁻¹ after 50 cycles.¹² Single-bath, pulsed electrodeposition of Cu–Sn alloy film was performed by Beattie et al.¹³ Their sample exhibited a reversible capacity of ∼200 mA h g⁻¹ after 40 cycles at a current density of 15 mA g⁻¹. By electron-beam deposition, Hu et al. prepared the Sn/Cu₆Sn₅ alloy composite thin film with a discharge capacity of ∼370 mA h g⁻¹ after 50 cycles.¹⁴

On one hand, it is well known that the Cu content in the Cu–Sn alloy is increased; the specific capacity is sacrificed for capacity retention. Hence, the additive proportion of Sn is the key factor that affects the performance of Cu–Sn alloy for LIBs, but few studies have explored the effects of ratios of Sn to Cu₆Sn₅ on the performances for LIBs. On the other hand, the mechanical milling technique method has many unique advantages in comparison with other routes like the solution method. Firstly, the finished product yield close to almost 100% can be obtained when used the mechanical milling technique, while the solution method usually gains a low value. Secondly, the component of the materials can be easily controlled by adding the raw materials at a designed ratio before milling. Moreover, compared with the solution method, the mechanical milling technique is cost-effective and connecting with the actual production process. Therefore, in this work, we directly fabricate the Cu₆Sn₅/Sn composites with different ratios of Cu₆Sn₅ to Sn by the high energy mechanical milling technique. Then, the electrochemical lithium storage performances of the Cu₆Sn₅/Sn composites as anode materials for LIBs are investigated. To further enhance the cycling performances, the obtained Cu₆Sn₅/Sn composite with y = 7 is reprocessed by adding graphite into. Our synthetic route described herein is a simple, low-cost, and large-scale method to produce the high-performance Sn anode material, presenting a promising prospect for practical application.

2. Experimental

The pure metal powders Cu and Sn (>99.99%) were mixed by the atom ratio of Cu : Sn = x : y, where x = 10 − y, y = 4.5, 7 and 9. Then, the Cu₆Sn₅/Sn composites (Cu_xSn_y) were prepared by the high energy mechanical ball milling technique (SPEX8000D), with the total time of 400 min and an argon atmosphere. In order to avoid the super-heating which cause by the continuous working, the rest of 20 min was set every internal of 40 min. After mechanical milling, a heat treatment with 150 °C for 400 min was performed on the mixture powders, whose aim was to achieve a phase of alloy. The X-ray diffraction (XRD) was used for making the contrast of the different phase of the samples, scanning electron microscope (SEM) was used for observing micro-morphology, high-resolution transmission electron microscope (HRTEM) and energy-dispersive X-ray (EDX) were used for observing the Cu₆Sn₅/Sn core–shell structure.

The working electrode was prepared by compressing a mixture of the active material (Cu_xSn_y), acetylene black, and polyvinylidene fluoride (PVDF) with a weight ratio of 80 : 10 : 10 (mixed with a suitable N-methyl-pyrrolidone solvent) onto a copper foil (12 mm in diameter) and drying in a vacuum oven at 60 °C for 12 h. Then, the cells were assembled in a glovebox filled with highly pure argon gas. The electrochemical measurements were carried out by using CR 2025 coin-type cells with pure lithium as the counter electrode, a Celgard 2400 membrane as the separator, and 1 M LiPF₆ dissolved in ethylene carbonate (EC) and dimethyl carbonate (DMC) with a volume ratio of 1 : 1 as the electrolyte. The cyclic voltammogram was conducted on a BioLogic VMP3 electrochemical workstation. The charge and discharge performances were tested between 0.01 V and 2.0 V using a programmable battery testing system (LAND CT2001A) at room temperature.

3. Results and discussion

Fig. 1 shows the XRD patterns of the Cu_xSn_y (x = 10 − y, y = 4.5, 7 and 9) alloys. The peaks at 30.2°, 35.2°, 43.2° and 43.5° are indexed to single-phase Cu₆Sn₅, when y = 4.5. However, the peaks are indexed to both Sn and Cu₆Sn₅ phases, when y = 7 or 9. Notedly, the characteristic peaks of Cu₆Sn₅ are weaker than that of Sn, which shows the sample with y = 9 has excessively Sn rich in the Cu₆Sn₅ (Sn > Cu₆Sn₅). However, as a vivid contrast, the characteristic peaks of Cu₆Sn₅ are rather strong than that of Sn, suggesting the sample with y = 7 with slightly Sn rich in the Cu₆Sn₅ (Sn < Cu₆Sn₅). The peaks at 29.9° and 33.3° were indexed to SnO phase, which may be formed in the process of high energy mechanical milling. In addition, the sample with y = 7 prepared in different ball-milling times shows time-dependent phase evolution. As shown in Fig. S1 (ESI†), the intensity of Cu₆Sn₅ phase increases as the prolonging of ball milling time within 400 min. However, although the time increased to 600 min, it is nearly the same as the ball milling time with 400 min. It can be seen that the phase composition of the sample is steady after the sufficient ball milling times. Hence, we design the samples with limited ball milling time of 400 min.

Fig. 1 The XRD patterns of the Cu_xSn_y (x = 10 − y, y = 4.5, 7 and 9) alloys. (a) The sample with y = 4.5, (b) the sample with y = 7, and (c) the sample with y = 9.

Fig. 2 presents the SEM and HRTEM images of the Cu_xSn_y (x = 10 − y, y = 4.5, 7 and 9) alloys. It could be seen that the particles distribution with sizes of several decades' nanometers. There is not obviously difference of the primary particles size in the obtained samples. To further confirm the structure characterization of the as-synthesized Cu_xSn_y alloys, the HRTEM images are obtained (Fig. 2d–f). It is observed that a Sn layer is deposited onto the Cu₆Sn₅ particle when y = 7. However, no core–shell particles are founded when y = 4.5 or 9. The lattice fringes can be clearly seen in the HRTEM images, in which the 0.209 and 0.296 nm can be assigned to the (110) and (101) interplane spacing of Cu₆Sn₅, and the 0.292 and 0.279 nm can be assigned to the (200) and (101) interplane spacing of Sn.

Fig. 2 The SEM and HRTEM images of the Cu_xSn_y (x = 10 − y, y = 4.5, 7 and 9) alloys. (a) The SEM image of sample with y = 4.5, (b) the SEM image of sample with y = 7, (c) the SEM image of sample with y = 9, (d) the HRTEM image of sample with y = 4.5, (e) the HRTEM image of sample with y = 7, and (f) the HRTEM image of sample with y = 9.

In order to provide further insight into the structure information on the characterization of core–shell, an EDX line-scan analysis of Cu and Sn along the scan-path is shown in Fig. 3. It indicates that there is no clear difference of the Sn content along the scan-path. However, an obviously changes of the Cu content along the scan-path is observed when y = 7. A lesser Cu content in the two sides of the scan-path and a higher Cu content in the center of the scan-path are obtained, indicating a core–shell structure possessed by the sample with y = 7. According to the contents, the components of core and shell are distinguished as Cu₆Sn₅ and Sn, respectively. The Sn layers are homogeneously deposited onto the Cu₆Sn₅ particles, indicating that Cu₆Sn₅ can serve as a stable support for Sn alloy, which are originated from the slightly Sn rich in the Cu₆Sn₅ (Sn < Cu₆Sn₅).

Fig. 3 The EDX line-scanning profile of Cu and Sn along the scan-path for (a) the sample with y = 4.5, (b) the sample with y = 7, and (c) the sample with y = 9.

To simply understand the formation process of the core–shell structure, a schematic diagram is given as shown in Fig. S2 (ESI†). The power derived from the ball milling is the driving force for the formation of Sn shell over Cu₆Sn₅. With y = 7, due to the thin Sn film over Cu₆Sn₅ core, this core–shell structure can be maintained, although the following ball milling continued. However, with y = 9, because of the superfluous Sn film over Cu₆Sn₅ core, this core–shell structure can be turned into numerous small core–shell, namely, numerous Cu₆Sn₅ core diffused into the Sn matrix.

Fig. 4 shows the CV curves in the initial three cycles of Cu_xSn_y (x = 10 − y, y = 4.5, 7 and 9) alloys electrodes with potential sweep rate of 0.2 mV s⁻¹. A cathodic current peak is appeared around 0.8 V in the first cycle when y = 4.5, which corresponds to the decomposition of solvent and the formation of solid electrolyte interphase (SEI) film on the electrode surface (according with the result of reported ref. 15). The peak does not appear in second and third cycles, reflecting the stable property of SEI film formed at first cycle. While the potential lower than 0.6 V, the increase of cathodic current is corresponded to the intercalation reaction of lithium in Cu₆Sn₅ to form Li₂CuSn and Li_xSn alloys (as described in eqn (1) and (2)).

10Li⁺ + 10e⁻ + Cu₆Sn₅ → 5Li₂CuSn + Cu (1)

(x − 10)Li⁺ + (x − 10)e⁻ + 5Li₂CuSn → Li_xSn + 5Cu (10 < x ≤ 22) (2)

Fig. 4 The CV curves in the initial three cycles of Cu_xSn_y (x = 10 − y, y = 4.5, 7 and 9) alloys electrode with potential sweep rate of 0.2 mV s⁻¹. (a) The sample with y = 4.5, (b) the sample with y = 7, and (c) the sample with y = 9.

In the positively going sweep, the current peaks appear around 0.6 and 0.83 V, which reflect the deintercalation reaction of lithium from Li_xSn (Li_xSn₅ → Li₂CuSn) and Li₂CuSn (Li₂CuSn → Cu₆Sn₅) alloys. When y = 7, the CV profile in the first cycle shows a new reduction peak around 0.28 V, derived from lithium intercalation processes of Li_xSn, as described in eqn (3).

xLi⁺ + xe⁻ + Sn → Li_xSn (x ≤ 22) (3)

It indicates that the Sn and Cu₆Sn₅ react continuously with lithium. Upon charging, four anodic peaks located at 0.55, 0.60, 0.74 and 0.82 V are observed which can be ascribed to the dealloying of the lithium, as Li_xSn to Sn (0.55 and 0.74 V), Li_xSn to Li₂CuSn (0.6 V), and Li₂CuSn to Cu₆Sn₅ (0.82 V). An almost identical CV profile in Fig. 4b is founded when y = 9, except a cathodic current peak appears at 1.32 V which corresponds to the intercalation reaction of lithium in SnO to form Li₂O. The nearly same CV profile could be attributed to the same composition of Cu₆Sn₅ and Sn alloys with y = 7 and 9. The shapes of CV profile in Fig. 4 are almost the same since the second cycle, suggesting the high reversibility of the following lithiation and delithiation processes.

In addition, the lithiation mechanism of the sample with y = 7 has been further proved by ex situ XRD measurements. As presented in Fig. S3 (ESI†), the sample with y = 7 has the characteristic peaks of Cu₆Sn₅ phase and pure Sn phase before discharge. After discharge to 0.4 V, the Li₂CuSn phase arises with strong intensity. Moreover, it is replaced by the Li_xSn phase once discharge to 0.01 V. The phase evolution of the sample with y = 7 in different lithiation stages measured by ex situ XRD is well confirmed the CV results. Notedly, the presence of unreacted Cu₆Sn₅ and pure Sn phase, although low in intensity, is likely due to the pulverization of the active material during lithium insertion/extraction that results in material detachment, a phenomenon commonly observed for conversion-type SnSb materials.¹⁶

Fig. 5 presents the initial two charge/discharge curves of Cu_xSn_y (x = 10 − y, y = 4.5, 7 and 9) alloys electrodes at the current density of 100 mA g⁻¹. In the first charge process, a potential plateau appeared around 0.9 V could be attributed to the decomposition of electrolyte solution. In first cycle, the discharge and charge capacities of the sample with y = 4.5 are 826.9 and 597.0 mA h g⁻¹ with coulombic efficiency of 72.2%, respectively. We observe that the discharge and charge capacities are 1223.0 and 693.7 mA h g⁻¹, respectively, when y = 7. Its coulombic efficiency is 56.7%. The increase of charge/discharge capacity is corresponding to the existence of slightly Sn rich in the Cu₆Sn₅. The decrease of coulombic efficiency may be corresponding to the reaction of Sn decomposition and formation of SEI film on the surface of electrode. However, we observe the discharge and charge capacities are 894.3 and 575.0 mA h g⁻¹, respectively, when y = 9. The decrease of charge/discharge capacity could be corresponding to the lower electrical conductivity of Sn than that of Cu₆Sn₅ (Cu₆Sn₅ can encase the lithium/tin alloy in a conductive copper matrix¹⁷). The higher coulombic efficiency of 64.3% than that of the sample with y = 7 could be attributed to the reaction of Cu₆Sn₅ decomposition and formation of part SEI film on the surface of electrode. Furthermore, five potential plateaus appear around 0.36, 0.52, 0.65, 0.9 and 1.35 V, corresponding to the decomposition of Sn, Li₂CuSn, SnO and the formation of SEI films (agree with the result of ref. 18). During second cycle, there are no obvious changes of potential plateaus in Fig. 5. However, the charge/discharge capacity is lower than that in first cycle, because of the SEI films formation in the first cycle. The charge/discharge capacity decreases as the SEI film increases. The coulombic efficiency is higher than that in first cycle, because of the high reversibility of the following lithiation/delithiation process after first cycle.

Fig. 5 The initial two charge/discharge curves of Cu_xSn_y (x = 10 − y, y = 4.5, 7 and 9) alloys electrode at the current density of 100 mA g⁻¹. (a) The sample with y = 4.5, (b) the sample with y = 7, and (c) the sample with y = 9.

Fig. 6 shows the cycling behavior of Cu_xSn_y (x = 10 − y, y = 4.5, 7 and 9) alloys electrodes. The discharge capacity is remained to 600.7 mA h g⁻¹ after first cycle and 261.8 mA h g⁻¹ after 20th cycles, when y = 4.5. The coulombic efficiency is in the region of 94.63–94.65% in whole process except in the first cycle. However, the discharge capacities and coulombic efficiency are 761.6 mA h g⁻¹ after first cycle, 457.8 mA h g⁻¹ after 20th cycles and 90.63–91.37%, when y = 7. The higher discharge capacity of the sample with y = 7 than that with y = 4.5 is corresponded to the existence of slightly Sn rich in the Cu₆Sn₅ because the Sn has a higher theoretical capacity than that of Cu₆Sn₅. The decrease of coulombic efficiency could be attributed to the Sn decomposition and formation of SEI film on the surface of electrode. However, we observe the discharge capacities and coulombic efficiency are 557.6 mA h g⁻¹ after first cycle, 151.8 mA h g⁻¹ after 20th cycles and 88.81–93.02%, when y = 9. The decreased discharge capacity and coulombic efficiency could be attributed to the existence of excessively Sn rich in the Cu₆Sn₅ (Sn > Cu₆Sn₅). This leads the fact that the gigantic volume changes of Sn during cycling cause the breakdown of Sn particles and the loss of contact between each particle. Obviously, the discharge capacity of the sample with y = 7 increased 74.87% (from 261.8 mA h g⁻¹ to 457.8 mA h g⁻¹) and 201.58% (from 151.8 mA h g⁻¹ to 457.8 mA h g⁻¹), compared with the sample with y = 4.5 and 9, respectively. It indicates that slightly Sn rich in the Cu₆Sn₅ (Sn < Cu₆Sn₅) and Cu₆Sn₅–Sn core–shell structure can display an enhanced electrochemical performance.

Fig. 6 The cycling behavior of Cu_xSn_y (x = 10 − y, y = 4.5, 7 and 9) alloys electrodes. (a) The sample with y = 4.5, (b) the sample with y = 7, and (c) the sample with y = 9.

Along with the galvanostatic cycling behavior, the rate discharge abilities of the Cu_xSn_y (x = 10 − y, y = 4.5, 7 and 9) alloys electrodes are shown in the Fig. 7. It can be seen that the sample with y = 7 shows an enhanced rate capability compared with the other ones with y = 4.5 or 9. The average discharge capacities of 800.5, 700.0, 599.6 and 429.4 mA h g⁻¹ at current densities of 50, 100, 200 and 500 mA g⁻¹ are delivered, separately. Importantly, the discharge specific capacity resumes to 500 mA h g⁻¹ and maintains at the value of 492.3 mA h g⁻¹ after 25 cycles when the current density is lowered to 50 mA g⁻¹. For comparison, after 25 cycles at variational current density, the samples with y = 9 and 4.5 only maintain the discharge specific capacities of 296.1 and 218.2 mA h g⁻¹, respectively.

Fig. 7 The rate capacity of Cu_xSn_y (x = 10 − y, y = 4.5, 7 and 9) alloys electrodes. (a) The sample with y = 4.5, (b) the sample with y = 7, and (c) the sample with y = 9.

For supporting the possible reason for the difference performance of Cu₆Sn₅/Sn samples, the morphology of electrodes after 20 charge–discharge cycles are given. As shown in Fig. S4 (ESI†), the particle sizes of the sample with y = 9 is slightly lager than that of the samples with y = 7 or 4.5, which suggest that the secondary particle aggregation obviously occurs on the sample with y = 9. In addition, the thicknesses of the electrodes before and after 20 cycles are measured and compared. As exhibited in Table 1, the volume expansion ratio of the sample with y = 9 is 350%, which is evidently lager than that of the samples with y = 7 or 4.5. Both results indicate the drastic capacity fading in the sample with y = 9, namely, the samples with y = 7 or 4.5 with enhanced capacity retention rate. Notedly, the sample with y = 7 has no obviously increased volume expansion ratio in comparison with the sample with y = 4.5. Moreover, plus with the higher tin content, the sample with y = 7 can maintain higher specific capacity compared with the sample with y = 4.5.

Table 1 The thickness of the electrodes before and after 20 cycles

Samples	Before cycles	After 20th cycles	Volume expansion ratio
Cu6Sn5 (y = 4.5)	10 μm	27 μm	270%
Cu3Sn7 (y = 7)	10 μm	28 μm	280%
Cu1Sn9 (y = 9)	10 μm	35 μm	350%

---

Combined with the above results, it can be seen that the sample with y = 7 exhibits some enhanced lithium storage performances in comparison with the samples with y = 9 or 4.5 and other reported Cu–Sn-based alloy anodes as listed in Table 2. However, the cycling performance of the sample with y = 7 is also not excellent, although with some enhancements. For achieving the comprehensive results, we further improve the lithium storage performance of the sample with y = 7 by introducing the C additive based on our former obtained recipes. The Cu₆Sn₅/Sn/C composite is prepared by milling the obtained Cu₆Sn₅/Sn alloy with y = 7 and commercial graphite at a weight ratio of 85 : 15. As shown in Fig. 8, the obtained Cu₆Sn₅/Sn/C composite can maintain a discharge capacity of 605.8 mA h g⁻¹ even after 100 repeated cycles at the current density of 100 mA g⁻¹, which is larger than that of many reported Cu–Sn-based anodes such as the listed ones in Table 2. In consideration of the comparable electrochemical performance, the obtained Cu₆Sn₅/Sn/C composite is more suitable as an anode material for LIBs.

Table 2 Comparison of the electrochemical performances of the as-prepared Cu₆Sn₅/Sn and Cu₆Sn₅/Sn/C composite with previously reported Sn-based anode materials for LIBs

Materials	Current density	Cycle numbers	Reversible capacity (mA h g^−1)	References
Cu6Sn4	16.7 mA g^−1	20	∼200	7
Cu coated Cu–Sn alloy	0.3 mA cm^−2	50	220	12
Uncoated Cu–Sn alloy	0.3 mA cm^−2	15	428	12
Cu3.83Sn	15 mA g^−1	40	∼200	13
Sn/Cu6Sn5 composite	0.6 mA cm^−2	50	∼370	14
Porous Cu6Sn5	0.5 mA cm^−2	30	420	19
Cu6Sn5 alloy power	0.1C	20	313	20
Microspherical Cu6Sn5	50 mA g^−1	20	407	21
3D Sn–Cu alloy	100 mA g^−1	50	454	22
Cu–Sn alloy	0.25 mA cm^−2	50	413.5	23
Cu–Sn thin film	50 mA g^−1	100	300	24
Sn–Cu/MWCNT composite	150 mA g^−1	30	439	25
3D Cu–Sn nanostructure	0.3C	50	590	26
Sn–Co alloy film	250 mA g^−1	75	600	9
3D Ni–Sn alloys	0.1 mA cm^−2	200	500.5	10
Sn–Fe–C	100 mA g^−1	100	494	11
Cu3Sn7 (y = 7)	100 mA g^−1	20	457.8	This work
Cu6Sn5/Sn/C	100 mA g^−1	100	605.8	This work

---

Fig. 8 The lithium storage performances of Cu₆Sn₅/Sn/C composite electrode. (a) Voltage vs. specific capacity curve, (b) specific capacity vs. cycle number curve with corresponding coulombic efficiency.

4. Conclusions

A simple and low-cost method using to prepare Cu₆Sn₅/Sn composite is developed. The structure and morphology of the obtained materials are investigated. It is found that the peaks of the obtained materials are mainly consistent with Cu₆Sn₅ and Sn phases. Furthermore, the Cu₆Sn₅/Sn composite with a core–shell structure displays an outstanding electrochemical performance, when used as the anode material for lithium ion batteries. Our work gives an effective solution to solve the gigantic volume change of Sn during cycling. Given its unique advantages, we believe this advanced method may have good potential for practical application in high energy lithium ion battery areas.

Acknowledgements

This work is financially supported by the National Science Foundation of China (Project No. 11204102, 20111061, 21373198 and 51475207), the station Foundation of Ministry of Education for New Teachers (Project No. 20120061120041), Open Project of State Key Laboratory of Superhard Materials (Project No. 201410), and the China Postdoctoral Science Foundation (Grant No. 2015m580253).

References

1. Y. Zhang, L. Jiang and C. Wang, Nanoscale, 2015, 7, 11940–11944.
2. Y. Yang, D. Chen, B. Liu and J. Zhao, ACS Appl. Mater. Interfaces, 2015, 7, 7497–7504.
3. L. Wu, X. Hu, J. Qian, F. Pei, F. Wu, R. Mao, X. Ai, H. Yang and Y. Cao, Energy Environ. Sci., 2015, 7, 323–328.
4. L. Sun, X. Wang, R. Susantyoko and Q. Zhang, Carbon, 2015, 82, 282–287.
5. J. Chen, L. Yang, S. Fang, Z. Zhang and S. Hirano, Electrochim. Acta, 2013, 105, 629–634.
6. J. Chen, L. Yang, S. Fang, S. Hirano and K. Tachibana, J. Power Sources, 2012, 199, 341–345.
7. K. Kepler, J. Vaughey and M. Thackeray, J. Power Sources, 1999, 81, 383–387.
8. X. Wang, W. Han, J. Chen and J. Graetz, ACS Appl. Mater. Interfaces, 2010, 2, 1548–1551.
9. F. Ke, L. Huang, H. Wei, J. Cai, X. Fan, F. Yang and S. Sun, J. Power Sources, 2007, 170, 450–455.
10. S. Woo, N. Okada, M. Kotobuki, K. Sasajima, H. Munakata, K. Kajihara and K. Kanamura, Electrochim. Acta, 2010, 55, 8030–8035.
11. S. Yoon, J. Lee, H. Kim, D. Im, S. Doo and H. Sohn, Electrochim. Acta, 2009, 54, 2699–2705.
12. W. Pu, X. He, J. Ren, C. Wan and C. Jiang, Electrochim. Acta, 2005, 50, 4140–4145.
13. S. Beattie and J. Dahn, J. Electrochem. Soc., 2003, 150, A894–A898.
14. R. Hu, M. Zeng and M. Zhu, Electrochim. Acta, 2009, 54, 2843–2850.
15. R. Thomas, K. Rao and G. Rao, Mater. Express, 2014, 4, 65–71.
16. Y. Cheng, Y. Shao, L. Parent, M. Sushko, G. Li, P. Sushko, N. Browning, C. Wang and J. Liu, Adv. Mater., 2015, 27, 6598–6605.
17. X. Fan, X. Tang, D. Ma, P. Bi, A. Jiang, J. Zhu and X. Xu, J. Solid State Electrochem., 2014, 18, 1137–1145.
18. M. Wu, X. Li, Q. Zhou, H. Ming, J. Adkins and J. Zheng, Electrochim. Acta, 2014, 123, 144–150.
19. H. Shin and M. Liu, Adv. Funct. Mater., 2005, 15, 582–586.
20. S. Ju, H. Jang and Y. Kang, J. Power Sources, 2009, 189, 163–168.
21. X. Fan, F. Ke, G. Wei, L. Huang and S. Sun, Electrochem. Solid-State Lett., 2008, 11, A195–A197.
22. X. Fan, Y. Shi, J. Wang, J. Wang, X. Shi, L. Xu, L. Gou and D. Li, Solid State Ionics, 2013, 237, 1–7.
23. H. Zhao, C. Jiang, X. He, J. Ren and C. Wan, Ionics, 2008, 14, 113–120.
24. D. Polat, J. Lu, A. Abouimrane, O. Keles and K. Amine, ACS Appl. Mater. Interfaces, 2014, 6, 10877–10885.
25. M. Uysal, T. Cetinkaya, M. Kartal, A. Alp and H. Akbulut, Thin Solid Films, 2014, 572, 216–223.
26. Q. Deng, Z. Huang, X. Dai, Y. Wang, Z. Li and J. Li, J. Solid State Electrochem., 2015, 19, 1765–1771.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra23788b

---