Pyrite FeS₂–C composite as a high capacity cathode material of rechargeable lithium batteries

Abstract

Pyrite FeS₂ is a promising cathode material for rechargeable lithium batteries because of its high theoretical capacity (894 mA h g⁻¹), low cost and near-infinite earth abundance. However, the progress in developing viable Li/FeS₂ batteries has been hampered by the poor cyclability of the FeS₂ cathode. Aiming to improve the cyclability of the FeS₂ cathode, we here report a facile method for the synthesis of FeS₂–C composites by a one-pot hydrothermal reaction of FeSO₄ and Na₂S₂ in the presence of carbon black, and examine the effect of composition on the structure of FeS₂–C composites and the cycling performance of Li/FeS₂ cells. It is shown that the added carbon not only surrounds the FeS₂ surface but also penetrates into the entire FeS₂ particle, forming continuously conductive networks throughout the FeS₂ particle. However, introduction of carbon meanwhile increases the particle size of the FeS₂ active material. These two factors lead to an improvement in the rate capability of Li/FeS₂ cells while having little effect on the specific capacity and capacity retention of the FeS₂ cathode. On the other hand, we show that the electrolyte plays an important role in affecting the cyclability of Li/FeS₂ cells, and that the ether- and carbonate-based electrolytes affect the cycling performance of Li/FeS₂ cells in their unique manners.

---

Introduction

Iron pyrite (FeS₂, also referred as to fool's gold) is an earth abundant and environmentally benign natural material that has long been used as the cathode material in commercial primary Li/FeS₂ batteries, and it is shown that Li/FeS₂ batteries have remarkable power performance and significantly longer life than other alkaline batteries.¹ Primary Li/FeS₂ batteries are based on a four-electron overall reduction of FeS₂ to metallic Fe and Li₂S with respect to a theoretical specific capacity of 894 mA h g⁻¹. Typically, Li/FeS₂ batteries have an about 1.8 V open-circuit voltage and operates at a 1.5 V averaged voltage under normal conditions. Previous effort on the primary Li/FeS₂ batteries was focused on the enhancement of FeS₂ active material utilization (i.e., accessible capacity) and rate capability by reducing the size of FeS₂ particles.^2,3 In recent years, however, ever-increasing demand for energy storage has driven the research toward rechargeable batteries due to the low cost and near-infinite abundance of natural FeS₂ material. Up to date, a number of publications have dealt with the redox mechanism of FeS₂ at room temperature^4–8 and reported the modification of FeS₂ materials.^9–14 By the nature of conversion-type redox, FeS₂ is suitable for the cathode material of not only Li and Li-ion batteries but also Na and Na-ion batteries.^13,15–18 Additionally, FeS₂ has also been investigated as the anode material of Li-ion and Na-ion batteries.^19–25 However, progress in developing rechargeable batteries has been hampered by the poor cyclability of FeS₂ material because of the substantially different electrochemical process of FeS₂ in the primary and rechargeable batteries. We have recognized that once being fully discharged, Li/FeS₂ batteries become a combination of a discharged Li/FeS cell and a discharged Li/S cell with the electrochemical processes as described by eqn (1) and (2), respectively.²⁶

Fe + Li₂S − 2e⁻ ↔ FeS + 2Li⁺ (1)

nLi₂S − (2n − 2)e⁻ ↔ Li₂S_n + (2n − 2)Li⁺ (2)

The strongly chemical interaction of sulfur ions in FeS and Li₂S_n results in the formation of a FeS⋯S_nLi₂ complex, which in some cases rearranges to produce FeS₂ as expressed by a net reaction of eqn (3).

FeS + Li₂S_n → FeS₂ + Li₂S_n−1 (3)

Owing to the formation of soluble lithium polysulfide (Li₂S_n, n > 2), Li/FeS₂ batteries are shown to share the same problem as those present in Li/S batteries, such as the loss of sulfur active material, redox shuttle of the dissolved Li₂S_n, and corrosion of Li anode.

On the other hand, large volume expansion (up to 255% according to the difference in molar volumes between FeS₂ and final products) accompanying with the conversion of FeS₂ to Fe and Li₂S has been identified to be the other factor to cause the performance fading of Li/FeS₂ batteries as the volume change results in ineffective contact of electrode components and inhomogeneous redistribution of electrolyte.²² In view of material, a number of materials have been attempted to improve the cyclability of FeS₂ cathode, including FeS₂–C composites for accommodation of volume expansion,^{19,20,27–31} micrometer- or nanometer-sized materials for high utilization (specific capacity),^13,14 novel electrolytes for excellent compatibility with the highly reactive polysulfide,^{14,17,18,32,33} ionic liquids for reduced solubility of polysulfide,³⁴ and solid state electrolytes for complete elimination of polysulfide crossover.^35,36 Of them, the FeS₂–C composites are of particular interest because of their advantages in simple process for synthesis, improved electric contact of FeS₂ particles with conductive carbon for high power applications, and porous structure for buffering of incurred volume expansion.

In this work we developed a facile process for the synthesis of FeS₂–C composites by a sucrose-assisted hydrothermal reaction of FeSO₄ and Li₂S₂ in the presence of high porous carbon black. Herein, we report that the content of carbon significantly affects the structure of FeS₂–C composites and the rate capability of Li/FeS₂ cells while having little impact on the specific capacity and capacity retention, and that the ether- and carbonate-based electrolytes affect the cycling performance of Li/FeS₂ cells in their unique manners.

Experimental

Preparation of materials

FeS₂–C composite was prepared by a one-pot hydrothermal reaction of FeSO₄ and Na₂S₂ in the presence of carbon black as the carbon source. In a typical process, 2.50 g (9 mmol) FeSO₄·7H₂O and 2.70 g sucrose were dissolved into 10 mL deionized water in a beaker; 2.27 g (9.45 mmol) Na₂S·9H₂O and 0.288 g (9 mmol) elemental sulfur were weighed into the other beaker containing 20 mL deionized water and magnetically stirred to form a yellow-brown Na₂S₂ solution, followed by adding 0.19 g Ketjenblack EC-300JD carbon black and stirring to obtain a homogenous carbon suspension. Preliminary FeS₂–C composite particles were obtained by adding dropwisely FeSO₄ solution into the resultant carbon suspension and stirring vigorously, and then was transferred into a 45 mL Teflon-lined stainless steel autoclave and heated at 180 °C for 18 h. After naturally cooling down to room temperature, the precipitate was collected by vacuum filtration, rinsed three times with deionized water, and finally dried under vacuum at 80 °C for overnight. The chemical composition and crystal structure of final product were characterized by elemental analysis, X-ray diffraction (XRD) and Raman spectroscopy, and the morphology was analyzed by Brunauer–Emmett–Teller (BET) analysis, scanning electron microscope (SEM), and scanning transmission electron microscopy (STEM). Detailed descriptions about the structural characterizations and morphological analyses are referred to the ESI.†

Electrode preparation and cell assembly

Resultant FeS₂–C composite powder was coated onto a carbon-coated aluminium foil at a weight ratio of 80% FeS₂–C composite, 15% Super-P carbon, and 5% binder by using poly(acrylonitrile-methyl methacrylate) (ANMMA, AN/MMA = 94 : 6, MW = 100 000, Polysciences, Inc.) as the binder and N-methyl pyrrolidinone as the solvent. Using the same procedure, a FeS₂ electrode was prepared as the control in a composition by weight of 75% FeS₂, 10% Ketjenblack EC-300JD carbon black, 10% Super-P carbon, and 5% binder. On average, all electrodes had a loading of 2 mg FeS₂ per cm². The electrode was punched into 1.27 cm² circular discs and dried at 80 °C under vacuum for 16 h. Either an ether-based solution consisting of a 1.0 M lithium bis(trifluoromethanesulfonyl)imide (LiTFSI) dissolved in a 1 : 1 (wt) mixture of dimethyl ether (DME) and 1,3-dioxolane (DOL) or a carbonate-based solution consisting of a 1.0 M LiPF₆ dissolved in a 3 : 7 (wt) mixture of ethylene carbon (EC) and ethylmethyl carbonate (EMC) was used as the electrolyte. Using a Celgard 3410 membrane as the separator, 2032-size coin cell was assembled and filled with 18 μL electrolyte.

Electrochemical measurements

Using a Solartron SI 1287 Electrochemical Interface and a Solartron SI 1260 Impedance/Gain-Phase Analyzer, cyclic voltammetry (CV) was scanned at 0.1 mV s⁻¹ between 1.0 V and 3.0 V by starting from the cell's open-circuit potential, and ac-impedance spectrum was collected at charged state over a frequency range from 100 kHz to 0.01 Hz with a 10 mV oscillation. The cell was galvanostatically cycled between 1 V and 2.6 V on a Maccor Series 4000 cycler, and was charged at 0.5 mA cm⁻² for all rate tests. The specific capacity was expressed by referring as to the mass of FeS₂ active material.

Results and discussion

Three FeS₂ samples having carbon content of 1.9%, 13.4% and 18.7% C, respectively, were prepared and coded as FeS₂, FeS₂–C-I, and FeS₂–C-II, of which FeS₂ was prepared as the control in the absence of carbon back. The carbon detected in FeS₂ is due to the carbon precursor coated on the surface of particles, which has been proven to be formed by the hydrothermal decomposition of sucrose.³⁷ The SEM images of carbon black and these three samples are displayed in Fig. 1. It can be seen that carbon black shows a porous amorphous matrix and FeS₂ consists of microcubes in size less than 1 μm (see Fig. 1a and b). In FeS₂–C composites (Fig. 1c and d), the FeS₂ particles are randomly embedded within the amorphous carbon matrix, and interestingly the FeS₂ particles are found to change in shape into microspheres and increase in size with the content of carbon. In particular, Fig. 1d shows that FeS₂–C-II microspheres have an averaged diameter of 2 μm. This is because carbon particles serve as the seed for nucleation of FeS₂ precipitation and the FeS₂ crystals preferentially grow from the carbon surface.

Fig. 1 SEM images of (a) carbon black, (b) FeS₂, (c) FeS₂–C-I, and (d) FeS₂–C-II.

Fig. 2a shows the XRD patterns of three FeS₂ samples, which well match with the standard pyrite FeS₂ (PDF card #9000594) that has a space group of Pa3, in which the Fe atoms are octahedrally coordinated by six S atoms while the S atoms are tetrahedrally coordinated to three Fe atoms and one S atom, as illustrated by Fig. 2b. The Raman spectra of carbon black, FeS₂ and FeS₂–C-II are compared in Fig. 2c. Obviously, FeS₂–C-II combines the characteristics of FeS₂ and carbon black, including two relatively strong peaks at 363 and 375 cm⁻¹ and a very weak peak at 433 cm⁻¹ for the S₂ vibration (E_g), S–S in-phase stretch (A_g), and coupled vibration and stretch (T_g) modes of pyrite FeS₂,^38,39 and two strong peaks at 1349 and 1593 cm⁻¹ for the disorder-induced D-band and graphite structure-derived G-band of carbon materials.⁴⁰ There is an additional broad peak at 1100–1680 cm⁻¹ in the control FeS₂, which is due to the carbon precursor coated on the surface of FeS₂ particles as suggested by the 1.9% C in FeS₂. Fig. 2d shows the SEM image of a typical FeS₂–C-II microsphere, indicating that the surface of FeS₂ microsphere in FeS₂–C-II is composed of tightly assembled nanorods with a diameter of about 40 nanometers and that amorphous carbon randomly surrounds around the FeS₂ microsphere.

Fig. 2 Structural characterizations of FeS₂–C composite. (a) XRD pattern, (b) schematic representation of a pyrite FeS₂ unit cell, (c) Raman spectrum, and (d) SEM image of a FeS₂–C-II composite particle.

Chemical composition of FeS₂–C-II composite is further analyzed by acquiring an X-ray energy dispersive spectroscopy (XEDS) spectrum image in the scanning transmission electron microscopy (STEM) and analyzing the data with an automated eXpert spectral image analysis (AXSIA) software.^41,42 The results are summarized in Fig. 3, in which Fig. 3a shows the cross section of a FeS₂–C-II microsphere prepared by ultramicrotomy using high angle annular dark field (HAADF) imaging and Fig. 3b shows three main component phases by color of the interested cross section. The green, red, and blue colors in Fig. 3b are identified to be the carbon, FeS₂ with a small amount of carbon, and FeS₂ phase, respectively, by the XEDS component spectra as indicated in Fig. 3c. The carbon (green color) surrounding around the FeS₂–C-II particle arises from the epoxy resin that was used to hold the particle in the STEM experiment, whereas those distributing over the entire particle, as indicated by the red color, are the added carbon black that connects together to form continuously conductive networks. The SEM images (Fig. 1c and d, and 2d) and STEM image (red color in Fig. 3b) distinctly demonstrate that the added carbon black not only surrounds around the surface of FeS₂ microspheres but also penetrates throughout the particle. The formed conductive carbon networks benefit to the rate capability of FeS₂ cathode.

Fig. 3 STEM X-ray microanalysis of the cross-section of a FeS₂–C-II particle. (a) HAADF image, (b) map of three main component phases, and (c) EDS component spectra of each main phase in (b) where the green, red and blues colors represent the carbon, FeS₂ with a small amount of carbon, and FeS₂ phase, respectively.

Fig. 4 shows the basic electrochemical properties of FeS₂ cathode measured in an ether-based electrolyte and a carbonate-based electrolyte, respectively. In both types of electrolytes, FeS₂ presents a major cathodic current peak starting at ∼1.5 V vs. Li/Li⁺ in the first CV scanning (see Fig. 4a and c) and a voltage plateau at ∼1.5 V in the first discharge (see Fig. 4b and d). These results agree with a number of previous publications, and are attributed to two equal-capacity solid–solid phase transitions as described by eqn (4) and (5).^3–6

FeS₂ + 2Li → Li₂FeS₂ (4)

Li₂FeS₂ + 2Li → Fe + Li₂S (5)

Fig. 4 Cyclic voltammogram and voltage profile of the first two cycles of Li/FeS₂ cells with different types of electrolytes. (a) and (b): 1.0 M LiSO₃CF₃ 1 : 1 DME/DOL; (c) and (d): 1.0 M LiPF₆ 3 : 7 EC/EMC.

However, these two electrochemical processes cannot be visually distinguished in the CV and discharging voltage curve because of their close operation voltages. In recharge, the Li/FeS₂ cell behaves as a combination of a discharged Li/FeS cell and a discharged Li/S cell. Consequently, the cell shows two distinct electrochemical processes relating to eqn (1) and (2), respectively, as indicated by two anodic current peaks at 1.96 V and 2.60 V in CV and two voltage plateaus at 1.86 V and 2.45 V in recharge. Significant differences in electrochemical behavior between two types of electrolytes are observed from the 1^st recharge and 2^nd discharge. In the 1^st recharge, the ether-based electrolyte cell suffers slight redox shuttle at 2.5 V (see Fig. 4b); in the 2^nd discharge, the ether-based electrolyte cell has much lower currents in the 2.0 V cathodic peak (Fig. 4a vs. c) and larger capacity loss from the 1^st to 2^nd discharge (Fig. 4b vs. d). All these differences consistently suggest the dissolution of Li₂S_n and resulting loss of sulfur active material in the ether-based electrolyte, namely the same problems as those present in the Li/S batteries. It is well known that carbonate-based electrolytes are chemically incompatible with Li/S batteries because of the nucleophilic reaction between the carbonate solvents and Li₂S_n.^43–45 Herein, the ability of FeS₂ cathode operating in a carbonate-based electrolyte is attributed to the formation of a complex between the solid FeS and the Li₂S_n dissolved in electrolyte, i.e. the chemical adsorption, which greatly reduces the reactivity of Li₂S_n.

Fig. 5a shows the effect of electrolytes on the capacity retention of Li/FeS₂ cells with three different FeS₂ cathodes. With both types of electrolytes, the cells delivers a 810–900 mA h g⁻¹ capacity, equaling to a 90–100% of the theoretical capacity, in the first discharge. It is indicated by comparing Fig. 4b and d that from the 1^st discharge to 2^nd discharge, the ether-based electrolyte cell loses more capacities than the carbonate-based counterpart. This is because Li₂S_n has much higher solubility in the ether solvents than in the carbonate solvents, resulting in larger loss of sulfur active material. However, the ether-based electrolyte cells remain better capacity retention, and particularly two types of cells reach similar capacities after ∼50 cycles (Fig. 5a). The faster capacity fading rate of the carbonate-based electrolyte cells suggests that slow reactions between the carbonate solvents and Li₂S_n are still present although the reactivity of Li₂S_n has been greatly weakened by the chemical adsorption of FeS. In order to further understand the effect of electrolytes, the voltage curves of the 46^th cycle at which two Li/FeS₂ cells with the ether- and carbonate-based electrolytes had the same capacity (440 mA h g⁻¹) are plotted and compared with those of the 2^nd cycle in Fig. 5b. It can be clearly observed that the plateau feature entirely disappeared from the discharge voltage profile of the ether-based electrolyte cell, suggesting that the Li anode had been severely polarized by the insoluble reaction products (mainly Li₂S and Li₂S₂) of metallic Li and dissolved Li₂S_n. The above results reveal that the cycling performance of Li/FeS₂ cells is affected by the ether- and carbonate-based electrolytes in their unique manners, i.e., by the dissolution of Li₂S_n in the ether-based electrolytes and by the nucleophilic reaction between the solvent and Li₂S_n in the carbonate-based electrolytes.

Fig. 5 Effect of electrolytes on the cycling performance of Li/FeS₂ cells. (a) Capacity retention, and (b) voltage curve in the 2^nd and 46^th cycles.

Additionally, it can be seen from Fig. 5a that the composition of FeS₂–C composites affects the cycling performance of Li/FeS₂ cells differently in the ether- and carbonate-based electrolytes. With an increase in the carbon content of the FeS₂–C composites, the specific capacity of FeS₂ slightly increases in the ether-based electrolyte while remaining unchanged in the carbonate-based electrolyte. This is because in the ether-based electrolyte, the highly porous carbon helps to trap the dissolved Li₂S_n from diffusing out of the cathode by physical absorption, whereas in the carbonate-based electrolyte, this effect is masked by the low solubility of Li₂S_n. In both types of electrolytes, however, the composition of the FeS₂–C composite seems not to affect the capacity retention of Li/FeS₂ cells as suggested by the fact that all these cells with the same electrolyte experience the similar capacity fading rate. Insignificant effect of the carbon content on the capacity retention of the FeS₂ cathode is because the introduction of carbon black increases the size of FeS₂ microspheres as indicated in Fig. 1, which consequently reduces the utilization of FeS₂ active materials.

Fig. 6 indicates the effect of charging cutoff voltage on the cycling performance of Li/FeS₂–C-II cells. When cutting off at 2.3 V, namely right before the 2.5 V voltage plateau appears, the cell has lower capacity but leads to much stable capacity retention. The similar phenomenon has also been observed from Li/FeS cells with an ether-based electrolyte.³⁷ This is because no long-chain (soluble) Li₂S_n can be formed at or lower than 2.3 V. Therefore, the improved capacity retention by the 2.3 V charging cutoff voltage can be attributed to the reduced reactivity of the short-chain Li₂S_n for the carbonate-based electrolyte cell and to the low solubility of the short-chain Li₂S_n for the ether-based electrolyte cell.

Fig. 6 Influence of charging cutoff voltage on capacity retention of Li/FeS₂ cells.

Fig. 7a compares the rate capability of two cells using FeS₂ and FeS₂–C-II, respectively, as the cathode material and a carbonate-based electrolyte. In order to access to the full capacity, all charging processes were performed at 0.5 mA cm⁻². It was observed that all cycles had near 100% coulombic efficiencies although the cell was charged and discharged at different current densities. It can be seen from Fig. 7a that two cells have the similar capacities when the current density is lower than 1.0 mA cm⁻², and that the cell with FeS₂–C-II cathode outperforms once the current density exceeds 1.0 mA cm⁻². Fig. 7b shows the discharge voltage curves of the cell at different current densities. As normal, the voltage and capacity are declined with an increase in the discharging current density. The former is due to the IR drop in which the R includes both of the ohmic resistance and faradic resistance, and the latter relates to the increased ionic concentration polarization as well as the limited electrode reaction kinetics. As indicated by the voltage curves at 2.5 and 3.0 mA cm⁻² in Fig. 7b, the lower voltage plateau cannot reach the end before the voltage declines to the cutoff voltage (1.0 V) because of the significant IR drop.

Fig. 7 Rate capability of Li/FeS₂ cells with a 1.0 M LiPF₆ 3 : 7 EC/EMC electrolyte. (a) Comparison of FeS₂ and FeS₂–C-II cathodes, and (b) discharging voltage curves of a Li/FeS₂–C-II cell at different current densities.

Impedance analysis was used to understand the effect of carbon content on the rate capability of the FeS₂–C cathode. Fig. 8 compares the impedance spectra of two cells with the FeS₂ and FeS₂–C-II cathodes, respectively, which were collected at fully charged state by charging the cell to 2.6 V and keeping it at 2.6 V until the current density declines to 0.1 mA cm⁻². As normal, the impedance spectra of both cells show two overlapped semicircles followed by a straight slopping line in the low frequency end, which briefly reflect the ohmic resistance (R_sl) of the electrolyte–electrode interface and the faradic charge-transfer resistance (R_ct) of the electrode reaction. The R_sl and R_ct are two important elements of the impedance spectrum of a solid electrode, and can be fit using an equivalent circuit as shown by the inset in Fig. 8.^46,47 It can be seen from Fig. 8 that these two cells have almost the same bulk resistance (R_b), however, the one with the FeS₂–C-II cathode has lower R_sl and R_ct. This observation verifies that the FeS₂–C composite forms higher conductive electrolyte–electrode interface and offers faster electrode reaction kinetics, providing an excellent endorsement for the better rate capability observed from Fig. 7a.

Fig. 8 Impedance spectra of two Li/FeS₂ cells with FeS₂ and FeS₂–C-II cathodes, respectively, collected at charged state, in which the inset shows an equivalent circuit suitable for fitting of the impedance spectra.

Conclusions

In summary, we developed a facile process for the synthesis of FeS₂–C composites by a one-pot hydrothermal reaction of FeSO₄ and Na₂S₂ in the presence of carbon. A wide range of resources for commercial carbon materials provide great flexibility for architectural design and property modification of the FeS₂–C composite materials. The results of the present work show that the cyclability of FeS₂ cathode is affected by the FeS₂–C composite in multiple aspects, including (1) the particle size of FeS₂ active material, which is found to increase with the carbon content in the composite, (2) the physical absorption of carbon surface to Li₂S_n, which helps to trap the dissolved Li₂S_n from diffusing out of the cathode, and (3) the formation of conductive carbon networks, which increases rate capability of the battery. In addition, the type of electrolytes is shown to play an important role in affecting the cyclability of FeS₂ cathode. Briefly, the cyclability of the Li/FeS₂ batteries is affected by the ether-based electrolytes through the dissolution of lithium polysulfide and by the carbonate-based electrolyte through the nucleophilic reaction between the carbonate solvent and polysulfide anions.

Notes and references

1. D. Linden and T. B. Reddy, Handbook of Batteries, McGraw-Hill, Inc., 3rd edn, 2002, ch. 14.
2. Y. Shao-Horn, S. Osmialowski and Q. C. Horn, J. Electrochem. Soc., 2002, 149, A1499–A1502.
3. Y. Shao-Horn, S. Osmialowski and Q. C. Horn, J. Electrochem. Soc., 2002, 149, A1547–A1555.
4. R. Fong, J. R. Dahn and C. H. W. Jones, J. Electrochem. Soc., 1989, 136, 3206–3210.
5. R. Brec, E. Prouzet and G. Ouvrard, J. Power Sources, 1989, 26, 325–332.
6. D. A. Totir, I. T. Bae, Y. N. Hu, M. R. Antonio, M. A. Stan and D. A. Scherson, J. Phys. Chem. B, 1997, 101, 9751–9756.
7. E. Strauss, D. Golodnitsky and E. Peled, Electrochim. Acta, 2000, 45, 1519–1525.
8. L. A. Montoro, J. M. Rosolen, J. H. Shin and S. Passerini, Electrochim. Acta, 2004, 49, 3419–3427.
9. E. Peled, D. Golodnitsky, E. Strauss, J. Lang and Y. Lavi, Electrochim. Acta, 1998, 43, 1593–1599.
10. D. Zhang, J. P. Tu, J. Y. Xiang, Y. Q. Qiao, X. H. Xia, X. L. Wang and C. D. Gu, Electrochim. Acta, 2011, 56, 9980–9985.
11. T. Takeuchi, H. Kageyama, K. Nakanishi, Y. Inada, M. Katayama, T. Ohta, H. Senoh, H. Sakaebe, T. Sakai, K. Tatsumi and H. Kobayashi, J. Electrochem. Soc., 2012, 159, A75–A84.
12. L. Li, M. Caban-Acevedo, S. N. Girard and S. Jin, Nanoscale, 2014, 6, 2112–2118.
13. M. Walter, T. Zund and M. V. Kovalenko, Nanoscale, 2015, 7, 9158–9163.
14. Z. Hu, K. Zhang, Z. Zhu, Z. Tao and J. Chen, J. Mater. Chem. A, 2015, 3, 12898–12904.
15. T. B. Kim, J. W. Choi, H. S. Ryu, G. B. Cho, K. W. Kim, J. H. Ahn, K. K. Cho and H. J. Ahn, J. Power Sources, 2007, 174, 1275–1278.
16. A. Kitajou, J. Yamaguchi, S. Hara and S. Okada, J. Power Sources, 2014, 247, 391–395.
17. Y. Zhu, L. Suo, T. Gao, X. Fan, F. Han and C. Wang, Electrochem. Commun., 2015, 54, 18–22.
18. Z. Hu, Z. Zhu, F. Cheng, K. Zhang, J. Wang, C. Chen and J. Chen, Energy Environ. Sci., 2015, 8, 1309–1316.
19. B. Wu, H. Song, J. Zhou and X. Chen, Chem. Commun., 2011, 47, 8653–8655.
20. D. Zhang, Y. J. Mai, J. Y. Xiang, X. H. Xia, Y. Q. Qiao and J. P. Tu, J. Power Sources, 2012, 217, 229–235.
21. D. Zhang, G. Wu, J. Xiang, J. Jin, Y. Cai and G. Li, Mater. Sci. Eng., B, 2013, 178, 483–488.
22. J. Xia, J. Jiao, B. Dai, W. Qiu, S. He, W. Qiu, P. Shen and L. Chen, RSC Adv., 2013, 3, 6132–6140.
23. W. Qiu, J. Xia, H. Zhong, S. He, S. Lai and L. Chen, Electrochim. Acta, 2014, 137, 197–205.
24. X. Wen, X. Wei, L. Yang and P. K. Shen, J. Mater. Chem. A, 2015, 3, 2090–2096.
25. H. Xue, D. Y. W. Yu, J. Qing, X. Yang, J. Xu, Z. Li, M. Sun, W. Kang, Y. Tang and C. S. Lee, J. Mater. Chem. A, 2015, 3, 7945–7949.
26. S. S. Zhang, J. Mater. Chem. A, 2015, 3, 7689–7694.
27. L. Liu, Z. Yuan, C. Qiu and J. Liu, Solid State Ionics, 2013, 241, 25–29.
28. J. Liu, Y. Wen, Y. Wang, P. A. van Aken, J. Maier and Y. Yu, Adv. Mater., 2014, 26, 6025–6030.
29. S. B. Son, T. A. Yersak, D. M. Piper, S. C. Kim, C. S. Kang, J. S. Cho, S. S. Suh, Y. U. Kim, K. H. Oh and S. H. Lee, Adv. Energy Mater., 2014, 4, 1300961.
30. T. S. Yoder, M. Tussing, J. E. Cloud and Y. Yang, J. Power Sources, 2015, 274, 685–692.
31. T. S. Yoder, M. Tussing, J. E. Cloud and Y. Yang, ECS Meeting Abstracts, 2014, vol. 2, No. 257.
32. L. A. Montoro and J. M. Rosolen, Solid State Ionics, 2003, 159, 233–240.
33. J. W. Choi, G. Cheruvally, H. J. Ahn, K. W. Kim and J. H. Ahn, J. Power Sources, 2006, 163, 158–165.
34. T. Evans, D. M. Piper, S. C. Kim, S. S. Han, V. Bhat, K. H. Oh and S. H. Lee, Adv. Mater., 2014, 26, 7386–7392.
35. T. A. Yersak, H. A. Macpherson, S. C. Kim, V. D. Le, C. S. Kang, S. B. Son, Y. H. Kim, J. E. Trevey, K. H. Oh, C. Stoldt and S. H. Lee, Adv. Energy Mater., 2013, 3, 120–127.
36. V. Pelé, F. Flamary, L. Bourgeois, B. Pecquenard and F. le Cras, Electrochem. Commun., 2015, 51, 81–84.
37. D. T. Tran and S. S. Zhang, J. Mater. Chem. A, 2015, 3, 12240–12246.
38. H. Vogt, T. Chattopadhyay and H. J. Stolz, J. Phys. Chem. Solids, 1983, 44, 869–873.
39. A. K. Kleppe and A. P. Jephcoat, Mineral. Mag., 2004, 68, 433–441.
40. M. A. Pimenta, G. Dresselhaus, M. S. Dresselhaus, L. G. Cancado, A. Jorio and R. Saito, Phys. Chem. Chem. Phys., 2007, 9, 1276–1290.
41. P. G. Kotula, M. R. Keenan and J. R. Michael, Microsc. Microanal., 2003, 9, 1–17.
42. M. R. Keenan and P. G. Kotula, Surf. Interface Anal., 2004, 36, 203–212.
43. J. Gao, M. A. Lowe, Y. Kiya and H. D. Abruna, J. Phys. Chem. C, 2011, 115, 25132–25137.
44. T. Yim, M. S. Park, J. S. Yu, K. J. Kim, K. Y. Im, J. H. Kim, G. Jeong, Y. N. Jo, S. G. Woo, K. S. Kang, I. Lee and Y. J. Kim, Electrochim. Acta, 2013, 107, 454–460.
45. S. S. Zhang, J. Power Sources, 2013, 231, 153–162.
46. S. S. Zhang, K. Xu and T. Jow, J. Electrochem. Soc., 2002, 149, A1521–A1526.
47. S. S. Zhang, K. Xu and T. R. Jow, Electrochim. Acta, 2002, 48, 241–246.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra18895d

---