Facile strategy of NCA cation mixing regulation and its effect on electrochemical performance

Abstract

The cation mixing of LiNi_0.8Co_0.15Al_0.05O₂ materials was regulated by a facile strategy via control of oxygen flow rate during sintering. The cation mixing first decrease and then increased with the increasing oxygen flow rates. The effects of cation mixing on the electrochemical performance of LiNi_0.8Co_0.15Al_0.05O₂ materials were studied in detail. Their initial discharge capacities, rate capabilities and cycling retentions (both at room temperature and 55 °C) increase with decreasing cation mixing, and then decrease with increasing cation mixing, while their R_ct (charge transfer resistance) and oxidation/reduction peaks of the CV curves reveal the opposite trend. LiNi_0.8Co_0.15Al_0.05O₂, which possesses the smallest cation mixing, had an initial discharge capacity of 191.3 mA h g⁻¹ with 86.4% coulombic efficiency at 0.1C rate between 2.8 V and 4.3 V (vs. Li/Li⁺), which was maintained at 144.1 mA h g⁻¹ and 139.1 mA h g⁻¹ after 300 cycles at 1C rate at 25 °C and 55 °C, respectively. It is clear that the NCA sample with smaller cation mixing presents better electrochemical performance.

---

Introduction

Since the end of the 20th century, renewable energies and green power storage systems have become the critical issues of contemporary science and technology in attempts to alleviate the worldwide energy crisis while maintaining the sustainable development of our society. The lithium-ion battery (LIB), a new type of detachable energy storage device proposed in the 1990s, has wide applications in smart phones, cameras, tablets, hybrid electric vehicles (HEV) or even electric vehicles (EV). The performance of LIB, including output voltage, specific energy/power density, lifetime and security, is greatly influenced by its electrode materials.^1,2 However, enhancements of energy storage capability achieved in LIB have mainly resulted from optimal design of battery systems rather than materials innovation. Even within the field of materials innovation, much more research and development has been conducted into anode materials than cathode materials. Therefore, cathode materials have become the key factor limiting the overall performance of LIB systems, and, in the past decade, research has turned towards developing alternative cathode materials with higher energy/power density and thermal stability. Among possible candidates, layered Ni-rich lithium transition-metal oxides (Li_zNi_1−xM_xO₂; 1 − x > 0.6; M = Co, Mn, Al, etc.), benefiting from their similar layered structure to LiCoO₂ and higher reversible capacity (>190 mA h g⁻¹), are considered to be reasonable alternatives to dominant cathode material LiCoO₂.^3–9 LiNi_0.8Co_0.15Al_0.05O₂ (NCA), in particular, is regarded as an attractive cathode material for its outstanding electrochemical performance, thermal stability and environmentally friendly features derived from partial co-substitution by Co and Al elements.^10–12 Nevertheless, several major drawbacks of NCA, such as severe capacity fading during the charge–discharge process and thermal runaway attributed to structural instability, have prevented its large-scale use as a cathode material in LIB.

It is generally recognized that there is a close relationship between the drawbacks mentioned above and cation mixing.¹³ The cation mixing, in which transition metal ions (Ni²⁺ most of the time) shift from the transition-metal layer to the lithium layer, because of the similar radius of Li⁺ (0.076 nm) and Ni²⁺ (0.069 nm) as well as the partial Ni-ion reduction (valence from +3 to +2, caused by the nonstoichiometric structure or insufficient oxidation during sintering), is primarily responsible for the capacity fading and structure deterioration of NCA material.^14–18 Therefore, great importance should be placed on suppression of cation mixing during the synthesis process.

Cation mixing of NCA materials can be influenced by sintering temperature, sintering time and atmosphere. However, extensive attempts^19–21 made to modulate sintering temperature and to regulate time have involved high consumption of energy and resulted in low efficiency. There are two potential strategies for atmosphere control incorporating gas partial pressure and flow rates for cation mixing regulation. Precise control of gas partial pressure requires equipment that may raise the cost. Control of gas flow rates is much easier and costs less. It has been proved that an oxygen-rich atmosphere is beneficial for synthesis of layered Ni-rich lithium transition-metal oxide materials.^22–24 Moshtev et al.²⁵ and Nahm²³ successfully synthesized fine crystalline LiNiO₂ under different flow rates of oxygen, but did not do much exploration into any correlation between flow rate and performance. In the present study, a facile strategy is proposed for regulation of cation mixing in NCA materials by control of oxygen flow rates (Fig. 1), and the effect of cation mixing on electrochemical properties is studied in detail.

Fig. 1 Diagram of the cation mixing regulation strategy.

Experimental

Material preparation

LiNi_0.8Co_0.15Al_0.05O₂ samples were synthesized by traditional solid state reaction with commercial Ni_0.8Co_0.15Al_0.05(OH)₂ precursor and LiOH·H₂O under different atmospheres. First, Ni_0.8Co_0.15Al_0.05(OH)₂ precursor and LiOH·H₂O with a molar ratio of 1 : 1.05 were dispersed into an aqueous solvent and stirred vigorously till the solvent evaporated. The mixed powder was then ground thoroughly in an agate mortar followed by sintering at 500 °C for 6 h and 800 °C for 12 h under flowing oxygen atmospheres, with flow rates of 3 L h⁻¹, 5 L h⁻¹, 10 L h⁻¹, and 20 L h⁻¹, which were marked as H-1, H-2, H-3, and H-4, respectively. The tap densities of the precursor, and prepared materials H-1 and H-2 were 2.04 g cm⁻³, 2.18 g cm⁻³ and 2.20 g cm⁻³.

Physical characterization

The content of lithium ions in the bulk of these samples was detected by inductively coupled plasma atomic emission spectrometry (ICP-AES PerkinElmer, Optima5300DV). X-ray diffraction (XRD) patterns were collected with a D/max-rB detector (Rigaku Corporation) equipped with a Cu target X-ray tube and a graphite bent crystal monochromator. Each XRD measurement was performed every 0.05° over a scattering angle region 2θ between 10 and 120° and the counting time was 10 s. Rietveld refinement was conducted using the GSAS-EXPGUI software package. The morphologies of the precursor and the as-prepared sample were observed using scanning electron microscopy (SEM Hitachi 4500). The pore size distributions were measured at liquid nitrogen temperature using a Micromeritics ASAP2020 system.

Electrochemical evaluation

The electrochemical performances of the LiNi_0.8Co_0.15Al_0.05O₂ samples were evaluated with a CR2025 coin cell assembled with a lithium metal anode in an argon-filled glovebox. Each positive electrode was fabricated by blending the prepared powders with 10 wt% of Super P and 10 wt% of polyvinyl difluoride (PVDF) binder. The mass loading of electrodes was 1.3 mg cm⁻². The cathode electrode and lithium anode electrode were separated by a porous polypropylene film, using 1 mol L⁻¹ LiPF₆ in EC/EMC/DMC (1 : 1 : 1 in volume) solvent as the electrolyte. Their electrochemical tests were carried out between 2.8 V and 4.3 V vs. Li⁺/Li electrode at 25 °C. Their initial charge–discharge profiles were carried out at 0.1C rate (the current density of 1C sets as 180 mA h g⁻¹). Their rate performances were tested at 0.1C, 0.2C, 0.5C, 1.0C, 2.0C, 5.0C and 0.1C in sequence. These tests were conducted on a battery test system (Neware CT 3008, Neware Co. China). Cycle performances were set to be tested between 2.8 V and 4.3 V at 1C rate for 300 cycles at 25 °C and 55 °C, respectively. Cyclic voltammetry (CV) was carried out on an electrochemical workstation (CHI604c, Instruments) between 2.8 and 4.3 V at a scan rate of 0.1 mV s⁻¹. The electrochemical impedance spectroscopy (EIS) measurements were conducted in a frequency range of 100 kHz to 0.01 Hz with an amplitude of 5 mV at 25 °C under a normal pressure.

Results and discussion

Ni/Li cation mixing of the material can deteriorate its electrochemical properties because of blockage of the Li-ion transportation channel.^26,27 To investigate the extent of Ni/Li cation disorder, XRD tests were conducted at room temperature and fitted by Rietveld refinement. Fig. 2 shows the X-ray diffraction patterns of the prepared LiNi_0.8Co_0.15Al_0.05O₂ sintered at different oxygen flow rates. All prepared samples were classified as having hexagonal α-NaFeO₂ structure with a space group of R3m, and no extra diffraction peaks were observed about the related secondary phases or impurities. Clear splitting of the (006)/(102) and (108)/(110) doublets is known to occur in formation of a well-ordered layered structure.²⁸ Sintering atmosphere does not change the crystal structure according to the XRD patterns. The crystallinity of the materials improved with increasing amounts of oxygen, as concluded from the sharper and stronger peaks in the patterns. Rietveld refinements were performed to refine the crystal structures of four LiNi_0.8Co_0.15Al_0.05O₂ samples with the structural model of (Li_1−zNi_z)^3a[Li_zNi_0.8−zCo_0.15Al_0.05]^3b{O₂}^6c.²⁹ The Li atoms occupy 3a sites, the Ni, Mn(Al) and Co atoms are randomly located at 3b sites, and oxygen atoms hold 6c sites. Since Ni²⁺ has an approximate radius with Li⁺, small quantities of Ni may take the place of Li on 3a sites. Simultaneously, the same number of displaced Li atoms should occupy the 3b Ni sites.^30,31 The calculated lattice constants as a hexagonal setting and unit cell volume of LiNi_0.8Co_0.15Al_0.05O₂ are summarized in Table 1. The reasonably small weighted profile R factor, R_wp, demonstrates that the proposed model is correct. Additionally, the refinement results of calculated patterns are consistent with that observed as shown in Fig. 2, therefore the refinement is acceptable. On the other hand, the c/a value of the unit octahedron MO₆ (M: Ni, Co, Al) can also express the degree of crystal distortion, which affects the cation ordering.^32–34 All c/a values of four samples were in the range of 4.95–4.96, which is expected for a well-ordered layer of two dimensional structure.³⁵ Sample H-2 obtains the maximal c/a value, implying that it has the smallest degree of trigonal distortion. The degree of NCA cation mixing was analyzed by the Ni²⁺ ion amount in the Li layer based on the refinement results shown in Table 1. When the flow rate of oxygen is increased, the Ni²⁺ amount in the Li layer is first decreased and then increased. The site exchange amount of Ni²⁺ with Li⁺ were found to be 2.25%, 0.21%, 1.45% and 1.45% for H-1, H-2, H-3 and H-4, respectively, which indicates that sample H-2 obtains the lowest amount of cation mixing. The Rietveld analysis suggests that the cation mixing degree of Li/Ni could be regulated by control of oxygen flow rates.

Fig. 2 Rietveld refined XRD data for LNCA samples (a) H-1 (b) H-2 (c) H-3 (d) H-4.

Table 1 Rietveld refined structure data for LNCA samples

Sample	Atom	Site	x	y	z	Occupancy	Rp	Rwp	χ^2
H-1	Li1	3a	0	0	1/2	0.9775	0.0112	0.0203	5.683
	Ni1	3a	0	0	1/2	0.0225
	Li2	3b	0	0	0	0.0225
	Ni2	3b	0	0	0	0.9775	a = b = 2.870454(Å)
	Co1	3b	0	0	0	0.1500	c = 14.217508(Å)
	Al1	3b	0	0	0	0.0500	c/a = 4.9531
	O1	6c	0	0	0.260(11)	1.0000
H-2	Li1	3a	0	0	1/2	0.9979	0.0116	0.0221	6.881
	Ni1	3a	0	0	1/2	0.0021
	Li2	3b	0	0	0	0.0021
	Ni2	3b	0	0	0	0.9979	a = b = 2.868271(Å)
	Co1	3b	0	0	0	0.1500	c = 14.210093(Å)
	Al1	3b	0	0	0	0.0500	c/a = 4.9542
	O1	6c	0	0	0.261(31)	1.0000
H-3	Li1	3a	0	0	1/2	0.9855	0.0111	0.0222	6.985
	Ni1	3a	0	0	1/2	0.0145
	Li2	3b	0	0	0	0.0145
	Ni2	3b	0	0	0	0.9855	a = b = 2.869653(Å)
	Co1	3b	0	0	0	0.1500	c = 14.211464(Å)
	Al1	3b	0	0	0	0.0500	c/a = 4.9523
	O1	6c	0	0	0.259(74)	1.0000
H-4	Li1	3a	0	0	1/2	0.9855	0.0118	0.0221	6.805
	Ni1	3a	0	0	1/2	0.0145
	Li2	3b	0	0	0	0.0145
	Ni2	3b	0	0	0	0.9855	a = b = 2.870022(Å)
	Co1	3b	0	0	0	0.1500	c = 14.209685(Å)
	Al1	3b	0	0	0	0.0500	c/a = 4.9511
	O1	6c	0	0	0.259(58)	1.0000

---

It is known that LiOH evaporation occurs from 450 °C during the synthesis process, which causes deviation of Li content between the final product and the initial molar ratio. Small quantities of excess lithium (Li/[Ni + Co + Mn] = 1.05–1.10) were designed to compensate for such evaporation,³⁶ resulting in stoichiometric composition in the final products. The oxygen flow rate has an effect on the Li content. The ICP-AES analysis results of as-prepared LNCA (Table 2) revealed that the Li/M (M = Ni + Co + Al) first increases to the 1.05 of H-2 and then decreases. The oxygen suppresses decomposition of LiOH,³⁶ but the oxygen flow takes away the lithium vapor around the reaction region. Thus, at flow rate up to 5 L h⁻¹ Li/M increases because of the suppression. However, the Li/M decreases when the flow rate increases beyond this as the lithium vapor is taken away. Meanwhile, the oxygen content inside varies with lithium content according to electrical neutrality. It can be concluded that oxygen flow rates regulate cation mixing by lithium^37–39 and oxygen regulation.

Table 2 ICP analysis data of lithium content in the as-prepared LNCA samples

Sample	Li/M (molar ratio)	Molecular formula
H-1	0.986	Li0.986Ni0.8Co0.15Al0.05O1.998
H-2	1.050	Li1.050Ni0.8Co0.15Al0.05O2.025
H-3	1.025	Li1.025Ni0.8Co0.15Al0.05O2.0125
H-4	1.008	Li1.008Ni0.8Co0.15Al0.05O2.004

---

SEM images of precursor and as-prepared LNCA samples before and after cycles are shown in Fig. 3. These images show that both the precursor and LNCA present microsized round-shaped secondary particles with a diameter of about 9 μm. After cycles there are micro-cracks in the particles, which were mainly caused by internal stress.⁵ Fewer micro-cracks could be observed in H-2 because of less cation mixing. Fig. 4 shows the corresponding EDX images of LNCA. The major elements (Ni, Co, Al, O) are uniformly distributed through the whole cross-section, and the particles inside also develop densely. Samples with different degrees of cation mixing have similar morphology and uniform element distribution.

Fig. 3 SEM images of the precursor and as-prepared LNCA before and after cycles.

Fig. 4 EDX mapping spectra of Ni, Co, Al and O element distribution in the cross-section of LNCA.

As seen from Fig. 5, there are small differences between the four samples and the pore volumes are less than 0.035 cm³ g⁻¹ because of the dense particles. The volume of pores around 7 nm could be attributed to the piled pores produced by the redundant lithium. Therefore, the amount of redundant lithium first decreases with the flow rates and then increases. H-2 has the least amount of redundant lithium.

Fig. 5 Pore-size distribution of NCA powders.

Electrochemical characterizations

The phase transition of materials during electrochemical reactions is usually studied by cyclic voltammetry (CV) curve analysis. As shown in Fig. 6, CV tests with a slow scan speed were performed to examine the effect of cation mixing on the phase transformations during the charge–discharge process. During the reduction process, the phase changes from hexagonal structure (H1) to monoclinic structure (M) at the first peak. Subsequently, it transforms to the second hexagonal phase (H2) and a third hexagonal phase (H3) at the second and third reduction peaks.^40,41 Meanwhile, the corresponding oxidation peaks indicate the opposite phase transformation. All of the four as-prepared samples revealed three pairs of phase transition peaks. H-2 exhibited the highest oxidation and reduction peaks, implying smaller cation mixing in favor of higher electrochemical activity.

Fig. 6 Cyclic voltammetry curves of LNCA. (0.1 mV s⁻¹ between 2.8 V and 4.3 V).

EIS spectra of as-prepared samples were measured after cell activation discharged to 2.8 V (as shown in Fig. 7). The equivalent circuit simulating the EIS spectra is exhibited at the top of Fig. 7. The fitting outcomes of R_s, R_SEI, and R_ct for samples H-1, H-2, H-3 and H-4 are shown in Table 3. Generally, the EIS spectra consist of high-frequency semi-circles and inclined lines which represent different physical significances.^16,37,42 According to the equivalent circuit model in Fig. 7, the high-frequency arc represents the impedance of Li-ion migration through the SEI film fitted by R_SEI and CPE1, and the medium-frequency one refers to the impedance derived from bulk electrode and electrode/electrolyte interface fitted by R_ct and CPE2. Usually, R_s means the ohmic resistance including the electrolyte, electrode substrate metal, electrode leads, terminals, etc. Here the equivalent circuit parallel constant-phase elements (CPE1 and CPE2) are substituted for pure capacitors.⁴³ However, there are no obvious sloping lines in the low-frequency region. The resistance of charge transfer (R_ct) is too big to overlap the Warburg impedance because of its relatively low response frequency. The fitting results reveal that R_ct and R_SEI vary almost linearly with cation mixing, which indicates that the kinetics of charge transportation could be seriously reduced by Ni/Li cation mixing and surface decomposition caused by oxygen deficiency. The smallest cation mixing produced the lowest charge transfer resistance (R_ct) and solid electrolyte interface layer resistance (R_SEI).

Fig. 7 Electrochemical impedance spectroscopy of LNCA.

Table 3 The fitting outcomes of R_s, R_SEI, and R_ct

Sample	Rs (Ω)	RSEI (Ω)	Rct (Ω)	Chi-squared
H-1	4.99	46	1231	0.0017112
H-2	4.878	30.2	752.4	0.0029456
H-3	4.796	157.3	1094	0.00086847
H-4	6.231	70	2751	0.0017567

---

The initial charge–discharge profiles of the four samples are shown in Fig. 8. The initial discharge capacity and efficiency of H-1, H-2, H-3 and H-4 are 182.2, 191.3, 186.1 and 176.9 mA h g⁻¹, 83.9%, 86.4%, 84.5% and 82.9%, respectively. The discharge capacity and efficiency first increase, and then decrease with increasing oxygen flow rate. There is a slight peak at the beginning of charge, which could be attributed to electrolyte polarization.⁴⁴ It is commonly considered that SEI film built with the released lithium ions forms the cathode materials during the first discharge process.⁴⁵ Generally, the initial discharge capacity and efficiency are related to the specific surface area (depending on particle morphology and size),⁴⁶ disorder degree of the material under the same cell preparation, and testing process. Li/Ni cation mixing, which is caused by blockage of lithium insertion, plays a pernicious role in electrochemical performance of nickel-rich cathodes.³⁹ It seems that H-2 with the smallest cation mixing exhibits the highest initial discharge capacity and efficiency.

Fig. 8 Initial charge–discharge curves of LNCA. (0.1C at 25 °C).

Fig. 9 displays the cycling performance of the LiNi_0.8Co_0.15Al_0.05O₂ at 25 °C and 55 °C. The discharge specific capacities of H-1, H-2, H-3 and H-4 remained at 144.1, 142.9, 140.0 and 138.1 mA h g⁻¹ after 300 cycles, with corresponding capacity retention rates of 88.2%, 90.2%, 88.6% and 88.0%. Capacity retention rates at 55 °C are a little lower than that at 25 °C, which is caused by thermal decomposition of cathode material and electrolyte. H-1, H-2, H-3 and H-4 maintained capacities of 113.2, 139.1, 85.4 and 97.9 mA h g⁻¹ with retention rates of 63.7%, 76.6%, 44.9% and 55.3%, respectively, at 55 °C. For all the samples, the capacity was a little higher at 55 °C than that at 25 °C, which is mainly attributed to the increased electrochemical activity at elevated temperatures. The phenomenon is usually observed in Ni-rich cathodes materials.^47,48 Although its discharge capacity is not the highest, H-2 clearly presents the highest retention rate. The improved interlayer structural stability can be attributed to the minimal cation mixing reflected by the smallest Li/Ni occupancy value, as shown in Table 1.

Fig. 9 Cycling performance at 25 °C and 55 °C.

Rate capabilities of the electrodes are shown in Fig. 10. All the discharge capacities of the samples decreased with increased C-rates. However, those of the H-2 samples decreased slowly, reflecting better rate capabilities for various C-rates. The discharge capacities of 0.1C, 0.2C, 0.5C, 1C, 2C and 5C for H-2 are 200.3, 189.7, 177.6, 163.4, 154.2 and 142.0 mA h g⁻¹, respectively. The capacities at 5 C-rate of all electrodes were similar, which can be ascribed to slow diffusion of electrolyte. The excellent rate capability of H-2 can be explained by less cation mixing which produces smaller internal resistance.

Fig. 10 Rate performance of LNCA.

Conclusions

High-performance LiNi_0.8Co_0.15Al_0.05O₂ as a cathode material was synthesized by a solid state reaction under oxygen. Regulation of NCA cation mixing was realized by control of oxygen flow rates. The cation mixing first decreased and then increased with increasing flow rates. LiNi_0.8Co_0.15Al_0.05O₂, which has the smallest cation mixing oxygen flow rate, had an initial discharge capacity of 191.3 mA h g⁻¹ with 86.4% coulombic efficiency at 0.1C rate between 2.8 V and 4.3 V (vs. Li/Li⁺), which was maintained at 142.9 mA h g⁻¹ and 139.1 mA h g⁻¹ after 300 cycles at 1C rate at 25 °C and 55 °C. Cation mixing has a significant effect on the electrochemical performance of NCA.

Acknowledgements

We acknowledge the National Natural Science Foundation of China (Grant No. 21273058), China postdoctoral science foundation (Grant No. 2012M520731 and 2014M70350), Heilongjiang postdoctoral financial assistance (LBH-Z12089), Harbin technological achievements transformation projects (2016DB4AG023) for their financial support.

References

1. M. Armand and J. M. Tarascon, Nature, 2008, 451, 652.
2. B. Dunn and J. M. Tarascon, Science, 2011, 334, 928.
3. C. H. Chen, J. Liu and M. E. Stoll, et al., J. Power Sources, 2004, 128, 278.
4. K. Kleiner, D. Dixon and P. Jakes, et al., J. Power Sources, 2015, 273, 70.
5. W. Liu, P. Oh and X. Liu, et al., Angew. Chem., Int. Ed., 2015, 54, 4440.
6. T. Ohzuku, A. Ueda and M. Kouguchi, J. Electrochem. Soc., 1995, 142, 4033.
7. H. Arai, M. Tsuda and K. Saito, et al., Electrochim. Acta, 2002, 47, 2697.
8. A. M. Kannan and A. Manthiram, J. Electrochem. Soc., 2002, 150, A349.
9. Z. Liu, H. Zhen and Y. Kim, et al., J. Power Sources, 2011, 196, 10201.
10. S. Hwang, W. Chang and S. M. Kim, et al., Chem. Mater., 2014, 26, 1084.
11. M. Jo, M. Noh and P. Oh, et al., Adv. Energy Mater., 2014, 4, 13072.
12. C. H. Chen, J. Liu and M. E. Stoll, et al., J. Power Sources, 2004, 128, 278.
13. N. S. Choi, Z. Chen and S. A. Freunberger, et al., Angew. Chem., Int. Ed., 2012, 51, 9994.
14. A. Rougier, I. Saadoune and P. Gravereau, et al., Solid State Ionics, 1996, 90, 83.
15. B. J. Neudecker, R. A. Zuhr and J. D. Robertson, et al., J. Electrochem. Soc., 1998, 145, 4160.
16. W. Li, J. N. Reimers and J. R. Dahn, Solid State Ionics, 1993, 67, 123.
17. T. Ohzuku, A. Ueda and M. Nagayama, J. Electrochem. Soc., 1993, 140, 1862.
18. K. K. Lee and K. B. Kim, J. Electrochem. Soc., 2000, 147, 1709.
19. S. Cui, Y. Wei and T. Liu, et al., Adv. Energy Mater., 2016, 6, 1501309.
20. T. J. Park, J. B. Lim and J. T. Son, Bull. Korean Chem. Soc., 2014, 35, 1671.
21. S. Xia, Y. Zhang and P. Dong, et al., Eur. Phys. J.: Appl. Phys., 2013, 65, 152.
22. P. Kalyani and N. Kalaiselvi, Sci. Technol. Adv. Mater., 2005, 6, 689.
23. Y. S. Lee, Y. K. Sun and K. S. Nahm, Solid State Ionics, 1999, 118, 159.
24. M. S. Idris and R. A. M. Osman, Adv. Mater. Res., 2013, 795, 464.
25. R. V. Moshtev, P. Zlatilova and V. Manev, et al., J. Power Sources, 1995, 54, 329.
26. D. D. Macneil, Z. H. Lu and J. R. Dahn, J. Electrochem. Soc., 2002, 149, A1332.
27. H. Yu, Energy Environ. Sci., 2013, 7, 1068.
28. Y. Makimura, T. Sasaki and T. Nonaka, et al., J. Mater. Chem. A, 2016, 4, 8350.
29. J. Z. Kong, H. F. Zhai and C. Ren, et al., J. Electrochem. Soc., 2014, 18, 181.
30. Z. H. Lu, L. Y. Beaulieu and R. A. Donaberger, et al., J. Electrochem. Soc., 2001, 149, A778.
31. G. H. Kim, S. T. Myung, H. S. Kim and Y. K. Sun, Electrochim. Acta, 2006, 51, 2447.
32. T. Ohzuku and Y. Makimura, Res. Chem. Intermed., 2006, 32, 507.
33. A. Rougier, P. Gravereau and C. J. Delmas, J. Electrochem. Soc., 1996, 143, 1168.
34. J. K. Ngala, N. A. Chernova and M. S. Whittingham, J. Mater. Chem., 2004, 14, 214.
35. J. Cho, H. S. Jung and Y. C. Park, et al., J. Electrochem. Soc., 1999, 147, 15.
36. E. Mccalla, G. H. Carey and J. R. Dahn, Solid State Ionics, 2012, 219, 11.
37. F. Wu, J. Tian and Y. Su, et al., ACS Appl. Mater. Interfaces, 2015, 7, 7702.
38. Y. Makimura, T. Sasaki and T. Nonaka, et al., J. Mater. Chem. A, 2016, 4, 8350.
39. Y. Bi, W. Yang and R. Du, et al., J. Power Sources, 2015, 283, 211.
40. W. Li, J. N. Reimers and J. R. Dahn, Solid State Ionics, 1993, 67, 123.
41. H. B. Xie, K. Du and G. R. Hu, et al., J. Mater. Chem. A, 2015, 3, 20236.
42. Y. Zhang, Z. B. Wang and J. Lei, et al., Ceram. Int., 2015, 41, 9069.
43. K. Du, J. Huang and Y. Cao, et al., J. Alloys Compd., 2013, 574, 377.
44. G. Zhu, K. Wen and W. Lv, et al., J. Power Sources, 2015, 300, 29.
45. M. Winter and R. J. Brodd, Chem. Rev., 2004, 10, 4245.
46. Z. Wang, L. Heng and W. Jian, et al., RSC Adv., 2016, 6, 32365.
47. T. Sasaki, T. Nonaka and H. Oka, et al., J. Electrochem. Soc., 2009, 156, A289.
48. O. Sha, Z. Tang and S. Wang, et al., Electrochim. Acta, 2012, 77, 250.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra20146f

---