Porous micro-spherical LiFePO₄/CNT nanocomposite for high-performance Li-ion battery cathode material

Abstract

Although many routes have been developed that can efficiently improve the electrochemical performance of LiFePO₄ (LFP) cathodes for Li-ion batteries, few of them meet the urgent industrial requirements of large-scale production, low cost and excellent performance. In this work, using a hydrothermal synthetic process followed by high-temperature lithiation, we are able to design a porous micro-spherical LFP/carbon nanotubes (CNTs) nanocomposite. The efficient combination of the inner CNTs and the outer carbon layer along with the porous structure of the LFP/CNT nanocomposite lead to its excellent electrochemical performance. As a cathode material for Li-ion batteries, the discharge capacity of the nanocomposite can reach 122 mA h g⁻¹ at 20 C. After 500 cycles, a capacity retention of 94.3% can be reached at 1 C. Because of the facile and easily scaled up synthetic method, low-cost raw materials and excellent electrochemical performance, the as-designed LFP/CNT nanocomposite can be expected to be a potential cathode candidate for Li-ion batteries.

---

1 Introduction

By the 21st century, the energy crisis had become one of the world's great challenges. The development of environmentally benign, sustainable, and renewable energy is economically and environmentally critical due to the increasing energy needs of modern society and the emerging ecological concerns. Utilizing renewable energy sources, such as solar power, wind and ocean waves, is an important step towards the worldwide imperative of replacing the inevitably vanishing non-renewable fossil fuels and avoiding the negative effects from the current combustion-based energy (e.g., environmental problems). With the fast development of technology, the conversion of renewable energies to electricity is no longer a challenge. The new urgent challenge is the development of efficient electricity storage devices to allow the widespread use of large electrical appliances such as electric vehicles (EVs), hybrid electric vehicles (HEVs) and so on.

Li-ion batteries (LIBs) are considered as one of the most promising environmentally friendly energy conversion and storage devices because of their unique advantages including high energy density and long cycle life.^1–4 However, more efforts are still needed to upgrade the performance of LIBs for their further applications in EVs and HEVs, which require high capacity, power density and safety. Electrode materials and cathode materials in particular are the determining factors in LIB performance.^5–8 Among reported cathode materials, LiFePO₄ (LFP), which has many appealing features such as long cycle life, high capacity (170 mA h g⁻¹), good safety, low cost, suitable voltage (3.4 V versus Li⁺/Li), environmental friendliness and high thermal stability,^5,9–11 has been considered as a promising cathode candidate for next-generation LIBs. However, the major drawback of olivine-structured LFP is its poor intrinsic electronic and lithium ion conductivities arising from the lack of mixed valency and one-dimensional lithium ion diffusion, which affect its electrochemical performance and particularly its rate capability.^5,9–13

In the past decade, it was found that the electrochemical performance of LFP can be greatly improved by coating with a conductive layer,^14,15 doping with ions^16,17 and reducing the particle size.^18–20

In fact, at present, the biggest obstacle to the commercial application of LFP is not its poor electrochemical performance, but the lack of large-scale synthetic techniques that can produce it with both excellent performance and low cost. In recent years, various synthetic routes were developed to prepare LFP, including mechanical alloying,^21,22 sol–gel methods,^23,24 co-precipitation,²⁵ microwave processes,²⁶ hydrothermal routes,^27,28 emulsion drying synthesis,^29,30 carbothermal reduction,³¹ vapor deposition,³² spray solution techniques,³³ and so on. However, most of the reported methods were either high-cost or impractical due to complicated synthetic procedures that are difficult to expand to large-scale industrialization. Even though solid-state reactions are universally recognized as a useful methodology to prepare LFP, these approaches suffer from polydisperse grain growth due to the high processing temperatures, which greatly decreases the electrochemical performance of LFP. Therefore, it is desirable to develop economic and efficient synthetic routes for the practical application of LFP materials.

In the present work, by modifying our previously published result,³⁴ we are able to synthesize a porous micro-spherical LFP/CNT nanocomposite by employing a facile hydrothermal method followed a lithiation process. When tested as a cathode material for Li-ion batteries, the LFP/CNT nanocomposite electrode shows a high discharge capacity of 122 mA h g⁻¹ at a current rate of 20 C and a capacity retention of 94.3% at 1 C after 500 cycles.

The as-designed LFP/CNT nanocomposite has great application potential as the urgently needed cathode material for Li-ion batteries suitable for EVs and HEVs due to the following reasons. First, the raw materials (Fe(NO₃)₃, NH₄H₂PO₄, LiOH and a small amount of CNTs) are inexpensive. Second, the synthetic method is easily scaled up because it only involves two steps: the hydrothermal fabrication of the FePO₄/CNT microspheres; and the following lithiation process (performed together with the calcinations of the products). The hydrothermal method is now a common industrial approach used to synthesize LiFePO₄, and the following lithiation process is similar to the calcination process, which also involves the hydrothermally synthesized LiFePO₄. Third, as mentioned above, the nanocomposite exhibits excellent electrochemical performance.

2 Experiments

Preparation of materials

Multiwall CNTs (20–30 nm in diameter and 0.5–2 μm in length, Beijing jindao Reagent) were pretreated according to the method reported elsewhere.^35,36 In a typical synthesis of the LFP/CNT nanocomposite, 0.3 g of the as-prepared CNTs was first dispersed in 400 ml purified water and stirring for 10 min. Then, 0.1 mol Fe(NO₃)₃·9H₂O (≥99.5%, Aladdin) was dissolved in the CNT solution with ultrasonic treatment for 10 min. Thereafter, 0.1 mol NH₄H₂PO₄ (≥99.5%, Aladdin) was introduced to the above solution under stirring. After stirring for 30 min, the mixture was transferred into a 400 ml PTFE inner steel autoclave and heated at 180 °C for 10 h. The autoclave was naturally cooled to room temperature. The obtained dark green slurry was washed several times with de-ionized water and dried under vacuum at 80 °C to obtain the FePO₄/CNT powder precursor.

The lithiation proceed as follows. Glucose (1 g) and 0.1 mol LiOH·H₂O were mixed with the as-prepared FePO₄/CNT powder. The obtained mixed powder was ground carefully with a mortar and a pestle and then transferred to a tube furnace, where it was annealed under reducing atmosphere. The tube was purged with 10% hydrogen and 90% argon for 0.5 h before heating began (flow rate: 150 cm³ min⁻¹). The powders were first heated at a rate of 5 °C min⁻¹ to 300 °C and kept for 1 h. Then, the temperature was increased to 700 °C at the same heating rate and held for 10 h. Finally, ∼16 g of LFP/CNT nanocomposite was obtained after naturally cooling to ambient temperature.

Characterization

Characterization details can be found in our previously published paper.³⁴

Electrochemical performance

Electrochemical reactions of samples with lithium were investigated using a simple two-electrode cell. The working electrode consisted of 90 wt% LFP/CNT nanocomposite active material, 5% carbon black as conductive agent, 5 wt% polyvinylidene fluoride (PVDF) as binder, and Al foils as substrate (current collector). An N-methyl pyrrolidinone (NMP) slurry consisting of the above mixture was uniformly coated on an aluminum disk with a diameter of 14 mm. The disk electrodes were dried overnight at 60 °C under vacuum followed by compression at 1.0 × 10⁶ Pa. Coin cells (2016-type) were assembled in an Ar-filled glove box using a polypropylene (PP) micro-porous film as the separator, a solution of 1 M LiPF₆ in ethylene carbonate (EC)/dimethyl carbonate (DMC; 1 : 1, v/v) as electrolyte and metallic lithium foil as the counter electrode. The electrochemical tests were performed on Land CT2001A battery testing systems (Jinnuo Electronics Co. Ltd., China). The charging and discharging tests were performed galvanostatically at various currents, a constant temperature of 25 °C and voltages ranging from 2.0 to 4.2 V. The electronic conductivity was measured at room temperature by the four-probe method. The LFP/CNT powder was dried at 120 °C in vacuum for 10 h and pressed into pellets under 10 MPa at room temperature. Electrical measurements were performed with a Keithley 2400 Digital Source Meter.

3 Results and discussion

The synthetic process of the LFP/CNT nanocomposite is schematically illustrated in Scheme 1. As mentioned in the Experimental section, the CNTs were pretreated before use. The pretreatment is known to create oxygen-containing surface functional groups on the chemically inert surfaces of the CNTs, thus facilitating the absorption of Fe³⁺ cations.³⁷ When Fe(NO₃)₃ solution is mixed with the CNTs, Fe³⁺ would be selectively bound to the oxygenated groups via electrostatic force. After pretreatment, NH₄H₂PO₄ was introduced into the solution. During the subsequent hydrothermal reaction, FePO₄ nanocrystals formed and assembled into microspheres with simultaneous embedding of the CNTs; the driving force was the reduction in the surface tensions of the highly dispersed FePO₄ nanoparticles.³⁴ After the high-temperature lithiation process, the FePO₄/CNT precursor was transformed into the LFP/CNT nanocomposite.

Scheme 1 Schematic illustration of the synthetic procedure for LFP/CNT.

Fig. 1a shows an SEM image of FePO₄/CNTs, from which it can be found that the as-prepared product is spherical in morphology. Most of the monodisperse spheres have particle diameters of ∼3 μm, with only a few spheres having diameters of 1–2 μm. The SEM results indicate that even in a 400 ml autoclave, the FePO₄/CNT nanocomposite product still retains its uniform micro-spherical morphology. The SEM results indicate that even in large-scale production, the FePO₄/CNT nanocomposite product can still retain its uniform micro-spherical morphology. ³⁴ No obvious CNTs are attached on the surface of the FePO₄ microspheres or partially exposed, suggesting that they may be embedded in the FePO₄ microspheres due to their short lengths (0.5–2 μm). Fig. 1b shows several unbroken and broken FePO₄/CNT microspheres, from which it can be observed that the surfaces of the microspheres are smooth, no obvious mesopores exist. Fig. 1c is a magnified image of the broken FePO₄/CNT microspheres. We can clearly see that CNTs with diameters of 20–30 nm are wrapped in the FePO₄ microspheres. Fig. 1d shows the XRD patterns of FePO₄/CNT, from which it can be seen that well-crystallized, high-purity FePO₄ was synthesized (matches with JCPDS card no. 29-715). No CNT-related peaks can be seen because: (1) the content of CNTs in the composite is low (1.9% in theory); and (2) the CNTs may be completely embedded in the FePO₄ microspheres.

Fig. 1 SEM images (a–c) of the FePO₄/CNT nanocomposite at different magnifications and (d) XRD pattern of FePO₄/CNT.

EDS elemental mappings of FePO₄/CNT microspheres were performed in order to further confirm that the CNTs were successfully embedded in the FePO₄ microspheres. Fig. 2a is a typical SEM image of three FePO₄/CNT microspheres, and Fig. 2b–e show their corresponding Fe, P, C and O element EDS mappings. The uniform distributions of C, Fe, O and P confirm the existence and homogeneous distribution of CNTs in the FePO₄ microspheres. Furthermore, we compared the FePO₄/CNT microsphere solution before and after hydrothermal reaction. We can clearly see that before reaction, the solution appears black due to the existence of the CNTs. In contrast, the solution after reaction appears dark green, which is consistent with the color of pure FePO₄. This dramatic change in solution color before and after the hydrothermal reaction indicates that the CNTs were successfully embedded in the FePO₄ microspheres. Due to the excellent electronic conductivity of CNTs, a three-dimensional (3D) conductive network of the FePO₄ microsphere is constructed.

Fig. 2 (a) SEM image of FePO₄/CNT microspheres, (b–e) their corresponding EDS maps of Fe, P, C, and O elements, and (f) FePO₄/CNT microsphere solution before and after hydrothermal reaction.

After the high-temperature chemical lithiation process, we obtained the carbon-coated LFP/CNT nanocomposite. Fig. 3a shows a TEM image of an individual LFP/CNT microsphere, from which the primary LFP nanocrystal size can be estimated to be in the range of 50–200 nm. High-resolution transmission electron microscopy (HRTEM) was performed to reveal the detailed microstructure of the carbon coated surface of the LFP/CNT microsphere. From Fig. 3b, we can clearly see that a carbon layer with a thickness of around 3.5 nm is present on the surfaces of the primary LFP nanocrystals, indicating that the decomposed carbon of glucose was successfully coated on the surfaces of the LFP microspheres during the high-temperature lithiation process. The carbon layer can further enhance the electronic conductivity of the LFP/CNT microspheres. The d spacing of 0.35 nm matches with the (111) plane of LFP.

Fig. 3 (a) TEM and (b) HRTEM images of an individual LFP/CNT microsphere.

The morphology of the LFP/CNT nanocomposite is shown in Fig. 4a–c. Fig. 4a is the overview of the composite; it can be seen that after the high-temperature lithiation process, most of the LFP/GNs still retain the spherical morphology of their precursor. The regular micro-spherical morphology ensures that the LFP/CNT nanocomposite has a high tap density, because micro-sized spherical particles can avoid large amounts of vacant space between the particles and bad powder fluidity, whereas irregular particles cannot.³⁴ As determined by the tap density meter, the nanocomposite exhibits a high tap density of up to 1.54 g cm⁻³. In fact, the tap density of LFP is an important factor that has a great influence on its final energy density. Fig. 4b is a close-up view of several LFP/CNT microspheres; we find that the diameters of the LFP/CNT microspheres (∼4 μm) are larger than those of their precursors (∼3 μm). This may be caused by the lithiation and surface carbon coating. In addition, the surfaces of the microspheres after the lithiation process become rougher compared to the precursor. Fig. 4c is a close-up SEM image of an individual LFP/CNT microsphere, from which two factors are verified: first, the microspheres consist of LFP nanocrystals with particle sizes around 100 nm; and second, mesopores are abundant on the surface of the microsphere. The crystal structure of the nanocomposite was investigated by XRD. As shown in Fig. 4d, all the diffraction peaks of the nanocomposite are well matched with orthorhombic phase LFP (space group Pnma, JCPDS card no.: 81-1173), indicating that high-purity LFP was formed. The average particle size determined by peak broadening (by taking the average value of the major diffraction peaks, (011), (111) and (211)) is 98.3 nm, which is in agreement with the TEM and SEM analyses.

Fig. 4 (a) Low- and (b) high-magnification SEM images of the LFP/CNT microspheres, (c) SEM image of an individual LFP/CNT microsphere and (d) XRD pattern of the LFP/CNT microspheres product.

The typical hysteresis observed in the nitrogen adsorption/desorption isotherms in Fig. 5 reveals the porous characteristic of this intriguing LFP/CNT nanocomposite with abundant mesopores. The BET surface area is 58.46 m² g⁻¹; this high surface area is probably derived from the mesoporous structure and highly dispersed LiFePO₄ microspheres (Fig. 4). The sufficient porosity in the microspheres may be attributed to the high-temperature lithiation process and the evaporation of hydration water in the FePO₄ microspheres. The Barrett–Joyner–Halenda (BJH) pore-size distribution, shown in the inset of Fig. 5, indicates that the LFP/CNT has an average pore diameter of 10–50 nm. Significant porosity in cathode materials can facilitate the access and accommodation of electrolyte and shorten the diffusion length of lithium ions to achieve high power density in electrode materials.³⁴

Fig. 5 Nitrogen adsorption/desorption isotherms of the LFP/GNs. Inset: the pore size distribution plot calculated by the BJH formula in the desorption branch isotherm.

Thermogravimetric (TG) measurement was used to estimate the carbon content (CNTs and the glucose-decomposed carbon) in the LFP/CNT sample. Fig. 6 gives the TG curves of the LFP/CNT nanocomposite tested in oxygen. It should be mentioned that in the temperature range of 250–500 °C, olivine LFP is oxidized to Li₃Fe₂(PO₄)₃ and Fe₂O₃, corresponding to a theoretical weight gain of 5.07%.^11,34 For the LFP/CNT, at 750 °C, the oxidations of both LFP and carbon are complete, and the weight of the sample remains constant. Thus, by taking into account the weight gain of pure LFP (5.07%) and the actual weight gain of the LFP/C composite (2.31%) during TG measurement, the total mass ratio of carbon in the composite can be calculated to be 2.63% as follows:³⁴

C_weight% = 1 − (1 + 2.31%)/(1 + 5.07%) = 2.63%

Fig. 6 TGA patterns of the LFP/GNs heated in oxygen from 50 to 850 °C.

In general, the optimal carbon content in LFP/C is less than 5 wt% (ref. 38) because too much carbon would greatly reduce the final tap density of LFP due to the low density of carbon, while insufficient carbon would not guarantee the good conductivity of the material. This means that the determined carbon content of 2.63 wt% would contribute to the high tap density of the LFP/CNT nanocomposite. Based on the total carbon loading of 2.63%, which includes both CNT and amorphous carbon, and considering the nearly 100% yield of the LiFePO₄ product, the CNT and amorphous carbon loadings can be generally calculated as 1.88% and 0.75%, respectively.

The introduction of CNTs plus the efficient carbon coating dramatically enhanced the electrical conductivity of the nanocomposite to 10.5 S cm⁻¹, which is 10 orders of magnitude higher than that of pure LFP (10⁻⁹ S cm⁻¹).³⁸ The greatly improved electrical conductivity of LFP also implies a well-constructed conducting network. Compared with the LFP/graphene nanocomposite we synthesized previously (the total carbon weight of the two samples are comparable: 2.63% vs. 2.64%),³⁴ the even higher electrical conductivity can be attributed to the following two factors: first, the higher efficiency of the combination of the inner CNTs and the outer carbon layer compared to the single graphene layer; and second, the higher Li-ion transport efficiency of the LFP/CNT nanocomposite compared to the LFP/graphene nanocomposite because Li-ions cannot pass through the carbon atomic arrays in two-dimensional sheets of graphene.³⁹

The ionic conductivity of the as-prepared electrode was also tested using the Randles–Sevcik equation; a detailed introduction can be found in the ESI (Fig. S1 and S2, ESI†). The obtained Li-ion diffusion coefficient is 1.12 × 10⁻¹⁰ cm² s⁻¹, which is 3–4 orders higher than that of the bare LiFePO₄ electrode (10⁻¹³ to 10⁻¹⁴ cm² s⁻¹). Therefore, the prepared ionic conductivity of the LiFePO₄/CNT electrode is greatly improved. With the improvements in both the electronic and ionic conductivities of the electrode, we expect the LiFePO₄/CNT cathode to exhibit excellent electrochemical performance.

The LFP/CNT nanocomposite electrode was cycled at different rates from 1 C to 20 C in the potential range between 2.0 and 4.2 V vs. Li⁺/Li. At 1 C, the discharge curve is characterized by a plateau at ∼3.4 V, which corresponds to the reaction LiFePO₄ → Li⁺ + FePO₄ + e⁻. The electrode capacity is 152 mA h g⁻¹, which is 89.4% of the theoretical capacity. At higher current rates, the capacity decreases but remains as high as 140 mA h g⁻¹ at 5 C, 131 mA h g⁻¹ at 10 C, and 122 mA h g⁻¹ at the high current of 20 C (Fig. 7). In fact, the value of 120 mA h g⁻¹ at 20 C is not very outstanding compared to previously reported results.¹² We want to emphasize that higher carbon loading in the electrode is known to correspond to batter rate performance (for the LiFePO₄ electrode) but a lower volumetric energy density. For our electrode, it is surprising that the LFP/CNT nanocomposite can exhibit such a high rate capability with such a low carbon loading of 2.63% compared to the reported values.¹² The high rate performance is attributed to the presence of the CNTs, which efficiently connect the nanocrystals inside the LFP microspheres to the effective carbon layer coated on the surfaces of the microspheres, dramatically enhancing the electrical conductivity of the LFP/CNT nanocomposite. We performed further experiments to confirm this conclusion. An LFP/amorphous carbon (LFP/AC) electrode (prepared in the absence of CNTs) and an LFP/bare CNT electrode (prepared without the carbon-coating process) with the same carbons loading as our LFP/CNT electrode were prepared for comparison. As shown in Fig. 7b and c, at a low current rate of 1 C, the LFP/AC electrode displayed charge/discharge capacities of 108.1 and 104.9 mA h g⁻¹, while those of the LFP/bare CNT electrode were 135.1 and 133.2 mA h g⁻¹. At a higher current rate of 5 C, the LFP/AC electrode showed charge/discharge capacities of 90.1 and 84.3 mA h g⁻¹, while those of the LFP/bare CNT electrode were 111.3 and 106.2 mA h g⁻¹. The two electrodes exhibit much worse electrochemical performances than the LFP/CNT electrode. More details regarding the cyclability and rate capability of LFP/amorphous carbon can be found in our previously published work.³⁴ For comparison, we also performed an electrochemical impedance test with LFP/AC and LFP/bare CNT electrodes at the end of 10 cycles at a rate of 1 C. According to the Nyquist plots of the three electrodes (Fig. 7d), the LFP/CNT electrode has a charge-transfer resistance value of 24 Ω, which is smaller than those of the LFP/AC (82 Ω) and LFP/bare CNT (112 Ω) electrodes. This indicates that the charge-transfer resistance in the LFP/CNT electrode is lower compared to the other electrodes. In addition, the high surface area of the nanocomposite, the existence of abundant mesopores and the tubular structures of the CNTs greatly facilitate the transport of Li ions across LFP nanoparticles, both inside and outside of the microspheres. In our tests, the LFP/CNT nanocomposite was able to support current densities as high as 3400 mA g⁻¹.

Fig. 7 (a–c) Charge/discharge profiles of the LFP/CNT nanocomposite, LFP/AC, and LFP/bare CNT at different rates, respectively. (d) Impedance spectra of the above three electrodes at a rate of 1 C at the end of 10 cycles.

Fig. 8 shows the capacities of the LFP/CNT electrode cycled at different rates for five cycles at each rate. The cell capacity remains stable, even at rates as high as 20 C, at which the cell is fully charged or discharged within 3 min. It is worth noting that when the current density is again reduced to 1 C, the capacity returns to its original value, indicating that this electrode material has high recovery ability. The good recovery ability is attributed to the embedded CNTs, which improve the mechanical stability of the nanocomposite, and to the interconnected mesopores system, which facilitates mass transport. Both of these characteristics endow the LFP/CNT nanocomposite with the ability to withstand the stresses caused by the fast phase change from LFP to FePO₄ and by the infiltration of electrolyte into the nanocomposite. The results indicate that our LFP/CNT nanocomposite is tolerant of high charge/discharge currents, which is a desirable characteristic required for high-power applications such as EVs and HEVs.

Fig. 8 Rate capacities of the LFP/CNT nanocomposite.

The cycling performance of secondary batteries is extremely important in EVs and HEVs because it directly determines their working life. Considering the practical use situation, the 1 C rate test is the most sensible. Thus, we carried out long-term cycling tests of our LFP/CNT electrode at a rate of 1 C. As shown in Fig. 9, the LFP/CNT electrode can deliver an initial discharge capacity of 152 mA h g⁻¹. After 500 cycles, a capacity of 143 mA h g⁻¹ is retained, corresponding to 94.3% of its initial capacity. This excellent long cycling performance makes our LFP/CNT nanocomposite suitable for use in EVs and HEVs. The CNTs, which have good mechanical flexibility, can serve as an elastic buffer to relieve the strain during the Li ion insertion/de-insertion process, which has a beneficial effect on the cycling performance of the nanocomposite.

Fig. 9 Cycling performance of the LFP/CNT nanocomposite.

4 Conclusions

In conclusion, we have successfully designed an LFP/CNT cathode for Li-ion batteries. The as-prepared composite material exhibits excellent electrochemical performance: the LFP/CNT electrode can deliver a discharge capacity as high as 122 mA h g⁻¹ at a high current rate of 20 C; the capacity can recovery to its original value even from 20 C to 1 C; and after 500 cycles, the capacity retention can reach up to 94.3%. The efficient combination of the inner CNTs and the outer carbon layer as well as the porous structure, which dramatically enhanced the electronic and ionic conductivities, contribute to the excellent electrochemical performance of the LFP/CNT nanocomposite. With the simple procedure and low-cost materials, this nanocomposite with a high electrochemical performance has great potential for industrial applications.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (11079002, 21475085, 21271125/B010601) and the Innovation Scientists and Technicians Troop Construction Projects of Henan Province, Program for Innovative Research Team in Science and Technology in University of Henan Province (2012TRTSTHN018).

Notes and references

1. K. S. Kang, Y. S. Meng, J. Bréger, C. P. Grey and G. Ceder, Science, 2006, 311, 977.
2. C. K. Chan, H. L. Peng, G. Liu, K. Mcllwrath, X. F. Zhang, R. A. Huggins and Y. Cui, Nat. Nanotechnol., 2008, 3, 31.
3. V. Etacheri, R. Marom, R. Elazari, G. Salitra and D. Aurbach, Energy Environ. Sci., 2011, 4, 3243–3262.
4. J. F. Liang, W. Wei, D. Zhong, Q. Yang, L. Li and L. Guo, ACS Appl. Mater. Interfaces, 2012, 4, 454.
5. M. S. Whittingham, Chem. Rev., 2004, 104, 4271.
6. F. Cheng, J. Liang, Z. Tao and J. Chen, Adv. Mater., 2011, 23, 1695.
7. W. Wei, Z. Wang, Z. Liu, Y. Liu, L. He, D. Chen, A. Umar, L. Guo and J. Li, J. Power Sources, 2013, 238, 376.
8. D. Chen, W. Wei, R. Wang, X. Lang, Y. Tian and L. Guo, Dalton Trans., 2012, 41, 8822.
9. A. K. Padhi, K. S. Nanjundaswamy and J. B. Googenough, J. Electrochem. Soc., 1997, 144, 1188.
10. Y. K. Zhou, J. Wang, Y. Y. Hu, R. O'Hayre and Z. P. Shao, Chem. Commun., 2010, 46, 7151.
11. W. Wei, D. Chen, R. Wang and L. Guo, Nanotechnology, 2012, 23, 475401.
12. G. Wang, H. Liu, J. Liu, S. Qiao, G. M. Lu, P. Munroe and H. Ahn, Adv. Mater., 2010, 22, 4944.
13. X. L. Wu, L. Y. Jiang, F. F. Cao, Y. G. Guo and L. J. Wan, Adv. Mater., 2009, 21, 2710.
14. N. Ravet, Y. Chouinard, J. F. Magnan, S. Besner, M. Gauthier and M. Armand, J. Power Sources, 2001, 97/98, 503.
15. Y. H. Huang and J. B. Goodenough, Chem. Mater., 2008, 20, 7237.
16. S. Y. Chung, J. T. Bloking and Y. M. Chiang, Nat. Mater., 2002, 1, 123.
17. P. G. Bruce, B. Scrosati and J. M. Tarascon, Angew. Chem., Int. Ed., 2008, 47, 2930.
18. Y. J. Lee, H. Yi, W. Kim, K. Kang, D. S. Yun, M. S. Strano, G. Cender and A. M. Belcher, Science, 2009, 324, 1051.
19. M. H. Lee, J. Y. Kim and H. K. Song, Chem. Commun., 2010, 46, 6795.
20. H. K. Song, K. T. Lee, M. G. Kim, L. F. Nazar and J. Cho, Adv. Funct. Mater., 2010, 20, 3818.
21. S. Franger, F. L. Cras, C. Bourbon and H. Rounault, Electrochem. Solid-State Lett., 2002, 5, A231.
22. C. W. Kim, M. H. Lee, W. T. Jeong and K. S. Lee, J. Power Sources, 2005, 146, 534.
23. J. Yang and J. J. Xu, Electrochem. Solid-State Lett., 2004, 7, A515.
24. G. X. Wang, S. Bewlay, S. A. Needham, H. K. Liu, R. S. Liu, V. A. Drozd, J.-F. Lee and J. M. Chen, J. Electrochem. Soc., 2006, 153, A25.
25. K. S. Park, K. T. Kang, S. B. Lee, G. Y. Kim, Y. J. Park and H. G. Kim, Mater. Res. Bull., 2004, 39, 1803.
26. K. S. Park, J. T. Sun, H. T. Chung, S. J. Kim, C. H. Lee and H. G. Kim, Electrochem. Commun., 2003, 5, 839.
27. S. Yang, P. Y. Zavalij and M. S. Whittingham, Electrochem. Commun., 2001, 3, 505.
28. K. Shiraishi, K. Dokko and K. Kanamura, J. Power Sources, 2005, 146, 555.
29. S. T. Myung, S. Komaba, N. Hirosaki, H. Yashiro and N. Kumagai, Electrochim. Acta, 2004, 49, 4213.
30. T. H. Cho and H. T. Chung, J. Power Sources, 2004, 133, 272.
31. J. Barker, M. Y. Saidi and J. L. Swoyer, J. Electrochem. Soc., 2003, 6, A53.
32. I. Belharouak, C. Johnson and K. Amine, Electrochem. Commun., 2005, 7, 983.
33. K. Konstantinov, S. Bewlay, G. X. Wang, M. Lindsay and J. Z. Wang, Electrochim. Acta, 2004, 50, 421.
34. W. Wei, S. Gao, Z. Yang and L. Guo, RSC Adv., 2014, 4, 56701.
35. H. Fang, S. Zhang, X. Wu, W. Liu, B. Wen, Z. Du and T. Jiang, J. Power Sources, 2013, 235, 95.
36. H. Fang, S. Zhang, T. Jiang, R. Lin and Y. Lin, Electrochim. Acta, 2014, 125, 427.
37. J. Lu, Carbon, 2007, 45, 1599.
38. L. X. Yuan, Z. H. Wang, W. X. Zhang, X. L. Hu, J. T. Chen, Y. H. Huang and J. B. Goodenough, Energy Environ. Sci., 2011, 4, 269.
39. G. Kucinskis, G. Bajars and J. Kleperis, J. Power Sources, 2013, 240, 66.

---

Footnote

† Electronic supplementary information (ESI) available: The calculation of the ionic conductivity of the LiFePO₄/CNT electrode. See DOI: 10.1039/c5ra05988g

---