carbon nanocomposites for high-power lithium ion batteries

Electrochimica Acta 168 (2015) 59–68

Contents lists available at ScienceDirect

Electrochimica Acta journal homepage: www.elsevier.com/locate/electacta

Organophosphonic acid as precursor to prepare LiFePO4/carbon nanocomposites for high-power lithium ion batteries Ming Chen, Leng-Leng Shao, Hua-Bin Yang, Qian-Yong Zhao, Zhong-Yong Yuan * Key Laboratory of Advanced Energy Materials Chemistry (Ministry of Education), Collaborative Innovation Center of Chemical Science and Engineering (Tianjin), College of Chemistry, Nankai University, Tianjin 300071, China

A R T I C L E I N F O

A B S T R A C T

Article history: Received 9 February 2015 Received in revised form 27 March 2015 Accepted 1 April 2015 Available online 2 April 2015

Amino tris(methylene phosphonic acid) (ATMP) is selected as phosphorus and carbon co-source for the synthesis of uniformly nano-sized LiFePO4/C by a quasi-sol–gel method. This strategy using ATMP instead of conventional NH4H2PO4 supplies two advantages: firstly, ATMP in situ chelates Li+ onto its framework and subsequently binds with FeC2O4 in aqueous solution, forming a molecule-scale homogeneous precursor which can obviously improve the purity of LiFePO4. Secondly, the organic carbon contained in ATMP can form uniformly distributed conductive carbon networks among LiFePO4 particles after calcination, which improves the electrical conductivity. The resultant LiFePO4/C with 1.1 wt.% carbon achieves a higher discharge capacity than those of LiFePO4 and LiFePO4/C prepared with inorganic NH4H2PO4. Moreover, core-shell structured LiFePO4/C nanocomposites are also fabricated by further introducing sucrose into the synthesis system. The high-quality carbon shell effectively hinders the LiFePO4 particle growth and aggregation under high-temperature treatment, which further enhances the electrical conductivity and lithium-ion diffusion, resulting in the improved electrochemical performance with excellent cycle stability (the optimum discharge capacity of 158.6 mAh g
1 at 0.1 C and 138.4 mAh g
1 at 2 C). The high purity, nanosize and core-shell structure of LiFePO4/C composites yielded by the novel synthesis strategy account for their outstanding electrochemical performance in high-power lithium-ion batteries. ã 2015 Elsevier Ltd. All rights reserved.

Keywords: Lithium iron phosphate Organophosphonic acid Sucrose Carbon coating Lithium-ion battery

1. Introduction Lithium ion batteries, as the promising energy storage devices for emerging electric vehicles, hybrid electric vehicles and intermittent renewable energies, have been widely investigated in recent years due to their high energy density, high capacity and long cycle life [1]. As one of the most popular cathode materials in lithium-ion batteries, the olivine-type LiFePO4 has received considerable attention due to its acceptable flat operating voltage (3.4 V vs Li/Li+), large theoretical rate capacity (170 mAh g
1), and good thermal stability, as well as low cost, nontoxicity, and environmental benignity [2,3]. Although LiFePO4 has been applied to practical uses in large-sized Li ion batteries (e.g. power battery), the sluggish lithium-ion diffusion (10
14–10
16 cm2 s
1) and low electronic conductivity (10
9–10
10 S cm
1) still limit its applicable electrochemical performance, which need to be improved [4– 6].

* Corresponding author. E-mail address: [email protected] (Z.-Y. Yuan). http://dx.doi.org/10.1016/j.electacta.2015.04.004 0013-4686/ ã 2015 Elsevier Ltd. All rights reserved.

The slow lithium-ion diffusion in the electrodes is closely related to the particle size and material homogeneity, which can be substantially enhanced by decreasing particle size and improving phase purity based on the proper synthesis route and precursor [7– 9]. So far, various routes including solid-state reaction [10], sol–gel method [11–13], hydro-/solvothermal approach [4,7], and spray pyrolysis technique [14], et al., have been developed for the synthesis of LiFePO4. Among them, the sol–gel method is particularly attractive for preparing high purity and homogeneity of nanoscaled LiFePO4 with controllable particle size and morphology. Nevertheless, the extra addition of complexants such as oxalic acid [11], citric acid [12], ethylene glycol [13] and tartaric acid [15] to obtain gel precursors not only increases the cost, but also causes the synthesis process complex. For the precursor, H3PO4 or ammonium phosphate is normally used as the phosphorus source in the synthesis of LiFePO4. However, the process is not green and suffers from the common problems associated with the use of liquid acids (H3PO4) such as corrosion of equipment. Moreover, ammonium phosphate can easily decompose into toxic ammonia which is a cause of air pollution. To meet the demand of economic synthesis, clean environment and high

60

M. Chen et al. / Electrochimica Acta 168 (2015) 59–68

performance, there is an urgent necessity for development of new and green precursors as phosphorus source and construction of a simple and effective route to synthesize high-quality LiFePO4. Recently, the organophosphonic acid, a kind of industrially used water clean compound, which is low-cost and environmentalfriendly, has been used to fabricate a variety of multifunctional inorganic-organic hybrid materials [16–18]. In the synthesis process, the terminal RPO3 groups of organophosphonic acid can in situ chelate metal ions such as Ti4+, Ni2+, Zn2+. Meanwhile, the uniformly distributed organic linking units in their framework will hinder the aggregation of precursor at space scale and allow the formation of well-dispersed products. In consideration of appropriate phosphorus atoms in the organic framework, the strong interaction with metal ions and the characteristics of solid acid, the organophosphonic acid can be a potential phosphorus source for LiFePO4. In addition, to overcome the poor electrical conductivity, modifications of LiFePO4 by coating with electronically conductive agents and doping with supervalent cations have been widely conducted [3,19–22]. Among them, carbon-coating is proved to be the most effective and facile method to improve power performance of LiFePO4 by the multiple positive effects: I) reduction of Fe3+ to Fe2+, II) minimization of the particle size, III) enhancement of electrical conductivity. Over the past decades, various carbon sources, including sucrose [23–25], glucose [26,27], poly(ethylene glycol) [28], poly(vinyl alcohol) [29], and citric acid [30], were explored to prepare LiFePO4 through the carbon-coating. Undoubtedly, carbon-coated LiFePO4 achieves good specific capacity and rate performance at relatively low carbon content. However, in most cases, the carbon layers are unevenly coated, incompletely wrapped and not intimately contacted to the LiFePO4 particles, which may suffer the low electrical conductivity in the interface of LiFePO4 and electrolyte, and lead to large charge-transfer resistance. It has been demonstrated that the formation of perfect carbon coating is not only affected by the species of the carbon precursor but depended on the homogeneity of the synthesis system [31–33], and thus a uniform reaction environment is particularly proposed in order to achieve the satisfactory carboncoating effect in the case of low carbon loading amount. Herein, we report the use of organophosphonic acid ATMP (amino tris(methylene phosphonic acid)) as co-source of phosphorus and carbon to synthesize the nano-sized and high-purity LiFePO4 particles with low carbon content through a quasi-sol–gel route. The introduction of multifunctional ATMP offers a green, low-cost and complexant-free synthesis route. More importantly, the resultant LiFePO4 exhibited superior electrochemical performance to LiFePO4 and LiFePO4/C derived from inorganic NH4H2PO4. Furthermore, with the assistance of sucrose, the prepared LiFePO4/ C nanocomposites with core-shell structure (LiFePO4 as core and carbon as shell) presented the enhanced performance for highpower lithium-ion batteries. The simple alternative route for preparing high-performance LiFePO4 materials from organophosphonic acid is crucial to the practical application of lithium ion batteries.

dried at room temperature overnight and transferred to a quartz boat. Precalcination was carried out at 350  C for 3 h in a tubular furnace under nitrogen protection. After cooling to room temperature, the resulting products were further ground and calcinated at 700  C for 3 h with a heating rate of 2  C/min under nitrogen protection. The prepared composite was labeled as LiFePO4/AS-n, where A and S is the first letter of ATMP and sucrose, respectively; and n (n = 0, 3, 6, 9, 12) denotes the amount of sucrose, corresponding to 0, 0.3, 0.6, 0.9 or 1.2 g, respectively. For comparison, we also synthesized the sample LiFePO4/S-m (m = 0, 3 or 6, standing for the added sucrose with the amount of 0, 0.3 or 0.6 g, respectively) by the same procedure except using NH4H2PO4 as phosphorus source. 2.2. Characterization Scanning electron microscopy (SEM) equipped with an energy dispersive X-ray detector (EDX) was carried out on a JEOL JSM7500F microscope at 5 kV. Transmission electron microscopy (TEM) was performed on a JEOL JEM 2010F at 200 kV. All samples subjected to TEM measurements were ultrasonically dispersed in ethanol and drop-casted onto copper grids covered with carbon film. X-ray diffraction (XRD) patterns were recorded on a Bruker D8 Focus diffractometer with Cu Ka radiation (l = 1.5406 Å) operated at 40 kV and 40 mA. The diffraction data were collected over the 2u angle of 10 to 60 at a scan rate of 3 /min. The lattice parameters were refined by using Rietica software. Thermogravimetric analysis (TGA) of the samples were conducted on a TA SDT Q600 analyzer in air or nitrogen atmosphere with a heating rate of 5 K/min. N2 adsorption–desorption isotherms were recorded on a Quantachrome NOVA 2000e sorption analyzer at liquid nitrogen temperature (77 K). Magnetization studies were performed at room temperature using a LDJ 9600 VSM magnetometer. 2.3. Electrochemical measurement Electrochemical measurements were performed by using a 2032-type coin cell. The working electrode (ca. 3 mg, 0.075 mm) was made by cathode composite LiFePO4/C, carbon black (Super-P) and polyvinylidene fluoride with a weight ratio of 80: 10: 10. The counter electrode was a disk of lithium metal foil (13 mm in diameter, 0.5 mm thick). The electrolyte was a 1 M solution of LiPF6 in 1:1 (v/v) ethylene carbonate/dimethyl carbonate. Celgard 2400 membrane was used as a separator. Galvanostatical charge and discharge measurements were conducted on the electrochemical test instrument (CT2001A, Land Electronic Co., Ltd. Wuhan, China) between 2.5 V and 4.2 V versus Li/Li+ at room temperature (25  C). Cyclic voltammetry (CV) was measured by using an IM6 instrument at a scanning rate of 0.1 mV s
1 between 2.5 and 4.4 V. Electrochemical impedance spectroscopy measurements were carried out at a discharge state (Voc: ca. 2.5 V) with a sinusoidal signal of 5 mV over a frequency range of 100 kHz to 100 mHz. 3. Results and discussion

2. Experimental section 3.1. Use of ATMP to prepare LiFePO4/C nanocomposite 2.1. Material preparation In a typical synthesis of LiFePO4/C nanocomposites, stoichiometric amounts of FeC2O4  2H2O, LiOH  H2O, ATMP and sucrose with a weight ratio of 7.2: 1.7: 4.2: x (x = 0, 0.3, 0.6, 0.9 or 1.2, in gram) were added to a corundum mortar. The mixture turned into slurry after 6 ml of deionized water was added. Then, the colloidal compounds were milled continually to form a yellow paste, following the evaporation of water. The obtained yellow paste was

By the quasi-sol–gel method, the homogeneous mixing of FeC2O4, LiOH and ATMP and the following two-step calcination process render the formation of LiFePO4/C nanocomposite (LiFePO4/AS-0). Organophosphonic acid ATMP is used as a complexing agent to chelate Li+, which would further reacted with Fe(OH)2 that arisen from the hydrolysis of FeC2O4  2H2O in a basic system at the initial heating stage, giving the precursor of LiFePO4/C. The organic compositions in the precursor would be

M. Chen et al. / Electrochimica Acta 168 (2015) 59–68

LiFePO4/AS-0

W e i g h t p e r c e n ta g e (% )

o

192 C

90

DSC

20

0

80

60 50

-20

TG

70

-40

o

500 C

0

200

400 o Temperature ( C)

Heat flow (mW)

100

600

-60

Fig. 1. TG-DSC curves of the precursor of LiFePO4/AS-0 under nitrogen.

(a)

800 º C

Intensity (a.u.)

750 º C

700 ºC

650 º C 550 ºC

10

20

30

40

50

60

2 Theta (deg.)

(b)

LiFePO4/AS-12 LiFePO4/AS-9 LiFePO4/AS-6

Intensity (a.u.)

partly decomposed during the precalcination, and the subsequent calcination at higher temperature causes the substantial pyrolyzation of the precursors with total breakage of C-P bonds, and the crystallization of LiFePO4 particles, consequently resulting in the formation of LiFePO4/C nanocomposite with uniform carbon networking inter-LiFePO4 particles. To synthesize high-quality LiFePO4/AS-0 nanocomposite, the calcination temperature of the precursor is a crucial factor which needs preferential consideration [33]. The optimum precalcination temperature was investigated by TG-DSC analysis. Fig. 1 shows the TG-DSC curves of the precursor of LiFePO4/AS-0. The initial weight loss before 192  C accompanied with an obvious endothermic peak is attributed to the evaporation of adsorbed and crystal water. A broad exothermic peak over the temperature range of 200 to 500  C is observed in the DSC curve, accompanying with the main weight loss, which corresponds to the decomposition of the organic component in ATMP and H2C2O4 (arisen from the hydrolysis of FeC2O4  2H2O), as well as the formation enthalpy of LiFePO4. However, no weight loss can be found after 550  C. Thus, the precalcination temperature of the precursor is determined to be in the range of 200 to 550  C, optimally at 350  C under nitrogen for 3 h, and the subsequent calcination at elevated temperature (550–800  C) is needed to obtain the good crystallization of LiFePO4 nanoparticles. Fig. 2a presents the XRD patterns of LiFePO4/AS-0 that were crystallized at different temperatures. The crystallinity of the samples below 700  C is weak and strengthened with the increase of temperature, suggesting that the crystallization temperature should be no less than 700  C. No obvious change in crystallinity can be observed when the temperature further increased to 800  C. Ensuring the similar crystallinity in the range of 700–800  C, the relative low reaction temperature avoids the growth of the crystal particles and the occurrence of the impurities such as Fe2O3 and Li3Fe2(PO4)3 [33,34]. In addition, the pyrolytic carbon from ATMP still possesses good electrical conductivity at 700  C [27]. Therefore, the crystallization temperature of 700  C is very critical in synthesizing high-quality LiFePO4/AS-0 nanocomposite. For comparison, the LiFePO4 and LiFePO4/C were also prepared by using inorganic NH4H2PO4 and sucrose as phosphorus and carbon sources, respectively. The XRD patterns of LiFePO4/AS-0 and LiFePO4/S-m (m = 0, 3 or 6) prepared at 700  C are compared in Fig. 2b. High crystallinity and pure phase of LiFePO4 can be obviously observed for all the samples. No Li3PO4, Li4P2O7 and FePO4 impurity phases were detected [35–38]. All Bragg peaks for LiFePO4/AS-0 and LiFePO4/S-m can be indexed to the olivine orthorhombic structure (space group Pmnb, triphylite). There are also no obvious carbon diffraction peaks in LiFePO4/AS-0 and LiFePO4/S-3 composites due to the low content and amorphous

61

LiFePO4/AS-3 LiFePO4/AS-0 LiFePO4/S-6 LiFePO4/S-3 (020)

10

(011) (111) (120)

20

(121)

(131) (031)

(211)

LiFePO4/S-0

30 40 2 Theta (deg.)

50

(222)

60

Fig. 2. XRD patterns of the prepared LiFePO4/AS-0 at different crystallization temperatures (a); and XRD patterns of LiFePO4/AS and LiFePO4/S prepared at 700  C (b).

state. The average crystallite diameter (D) was calculated by the Scherrer’s equation (D = 0.89k/bcosu) from the integral breadth b of four strong and well-resolved reflection peaks corresponding to [0 11], [111], [1 2 1] and [1 3 1] crystallographic directions [35,39]. The calculated crystallite size of LiFePO4/AS-0 is 58  5 nm, while that of LiFePO4/S-0, LiFePO4/S-3 and LiFePO4/S-6 is 150  5,100  5 and 85  5 nm, respectively. The carbon contents in LiFePO4/AS-0, LiFePO4/S-3 and LiFePO4/ S-6 composites were calculated from the thermogravimetric analysis (Fig. S1, Supporting Information), and the obtained carbon content of LiFePO4/AS-0 is 1.1 wt.%, similar to 1.2 wt.% carbon of LiFePO4/S-3, while LiFePO4/S-6 contains 2.5 wt.% carbon (Table 1). Derived from nitrogen sorption isotherms (Fig. S2), the BET surface areas of LiFePO4/AS-0 and LiFePO4/S-3 are 14.4 and 11.0 m2 g
1, accompanied with the mean pore sizes of 2.9 and 3.8 nm, respectively, revealing a negligible porous structure. Thus, the samples LiFePO4/AS-0 and LiFePO4/S-3 having similar carbon content and surface area can be used for comparison of the effect of the preparation method. Magnetic measurements were further carried out to investigate the purity of the prepared samples for their high sensitivity on ferro- or ferrimagnetic impurities [40–42]. The magnetization curves in Fig. 3 show that the saturation magnetization (Ms) is 8.538 and 2.708 emu/g for LiFePO4/S-0 and LiFePO4/S-3, respectively. The high Ms of LiFePO4/S-0 and LiFePO4/S-3 with two

62

M. Chen et al. / Electrochimica Acta 168 (2015) 59–68

Table 1 The structural and electrochemical properties of LiFePO4/AS and LiFePO4/S composites. Sample

Lattice parameters a [Å]

b [Å]

c [Å]

LiFePO4/S-0 LiFePO4/S-3 LiFePO4/S-6 LiFePO4/AS-0 LiFePO4/AS-3 LiFePO4/AS-6 LiFePO4/AS-9 LiFePO4/AS-12

6.0175 6.0142 6.0145 6.0166 6.0153 6.0074 6.0076 6.0017

10.3283 10.3275 10.3219 10.3264 10.3253 10.3249 10.3175 10.3162

4.6943 4.6874 4.6843 4.6872 4.6795 4.6767 4.6658 4.665

BET [m2 g
1]

ID/IG

Ms [emu/g]

Rs [V]

Rct [V]

C (0.1C)[mAh g
1]

Carbon [wt.%]

8.0 11.0 15.6 14.4 21.3 28.6 30.0 35.0

– – – 2.69 3.05 2.92 2.93 2.94

8.538 2.708 0.784 0.326 – 0.313 – –

6.0 4.5 3.7 3.9 3.5 3.4 3.0 3.1

250 125.7 61.5 76.8 48.3 34.2 25.1 21.6

90.2 113.3 137.1 132.5 145.4 158.6 149.7 142.1

0 1.2 2.5 1.1 2.3 3.5 4.6 5.7

apparent hysteresis loops indicates the presence of ferric compounds impurities [41]. Even 2.5 wt.% carbon involved in LiFePO4/ S-6, 0.784 emu/g of Ms is still obtained. In contrast, a pair of parallel lines in the magnetization curve of LiFePO4/AS-0 is observed with Ms of only 0.326 emu/g, suggesting low magnetization and high purity. Compared to those of LiFePO4/S-0 and LiFePO4/S-3, the high purity of LiFePO4/AS-0 nanocomposite might benefit from the positive effect of ATMP. Primarily, as a kind of excellent complexing agent, ATMP can effectively chelate Li+ in aqueous solution, and then forms the homogeneously molecule-scale LiFePO4/C precursor with FeC2O4  2H2O, which avoids the poor homogeneity of the reagents in the solid-state synthesis process. Secondly, the organic carbon contained in ATMP can form highly dispersed carbon networks among LiFePO4 particles, avoiding the oxidation of Fe2+ and simultaneously reducing the small amount of impurity Fe3+ in the reagents to Fe2+. Thus the purity of LiFePO4/AS-0 can be obviously improved. SEM and TEM images of LiFePO4/AS-0 are shown in Fig. 4 to understand its morphology and microstructure. The particles of LiFePO4/AS-0 are relatively uniform and granular with the sizes of 600–700 nm, much smaller than those of LiFePO4/S-0 (6–15 mm, Fig. S3A) and LiFePO4/S-3 (2–5 mm, Fig. S3B). In the TEM image shown in Fig. 4B, LiFePO4/AS-0 exhibits the uniform and smallsized (65 nm) nanoparticles with well-distributed fiber-like carbon networks among the particles, illustrating the important roles of ATMP in controlling particle sizes and the carbon webs in inhibiting particles growth. While, the crystal particles of LiFePO4/S-3 (Fig. S4A-D) exhibit various sizes (tens of nanometers to several hundreds of nanometers) and tend to agglomeration, with an uneven carbon layer. Carbon structure of the LiFePO4/AS-0 was studied by Raman spectroscopy. Two broad peaks in the range of 1500–1650 cm
1

and 1250–1500 cm
1 are observed in the Raman spectra (Fig. S5), which can be assigned to the graphite band (G-band) and the disorder-induced phonon mode (D-band), respectively. Their relative intensity ratio of D/G (ID/IG) depends on the type of graphitic materials and reflects the graphitization of carbon materials [43]. The ID/IG of LiFePO4/AS-0 is 2.69, implying that organic carbon in the ATMP network is pyrolyzed to form partially graphitized carbon at 700  C, which is expected to enhance the electrical conductivity and electrochemical performance of LiFePO4/AS-0 [44,45]. To investigate the advantages of organophosphonic acid ATMP over NH4H2PO4 in the synthesis of high-performance LiFePO4/C composites, the electrochemical behaviors of LiFePO4/AS-0 and LiFePO4/S-m (m = 0, 3, 6) were tested. Electrochemical performance of all the samples was examined by charge–discharge tests, EIS and cyclic voltammetry. The cell was cycled between 2.5 and 4.2 V, and charged at 0.1 C firstly. The detailed specific capacities for all the samples were summarized in Table 1. It can be seen that each sample displays a pair of very flat charge and discharge plateaus

8

LiFePO4/S-0

6

LiFePO4/S-3

4 2

LiFePO4/S-6 LiFePO4/AS-0 LiFePO4/AS-6

0 -2 Magnetization (emu/g)

Magnetization (emu/g)

10

-4 -6 -8

0.00

-0.02 -500

-10 -6000

0.02

-250

0

250

500

Applied field (Oe)

-4000

-2000

0

2000

4000

6000

Applied field (Oe) Fig. 3. Magnetization curves of LiFePO4/AS-n (n = 0, 6) and LiFePO4/S-m (m = 0, 3, 6) at room temperature.

Fig. 4. (A) SEM and (B) TEM images of the LiFePO4/AS-0 sample.

M. Chen et al. / Electrochimica Acta 168 (2015) 59–68

around 3.5–3.4 V (Fig. 5a), indicating the occurrence of two-phase transformation between LiFePO4 and FePO4 due to the lithium extraction and insertion reactions [46]. The pure LiFePO4 (i.e. LiFePO4/S-0) corresponding to the shortest length of the discharge plateau exhibits the lowest specific capacity of 90.2 mAh g
1 at 0.1 C, while the specific capacity of LiFePO4/S-3 reaches 113.3 mAh g
1. However, the specific capacity of 132.5 mAh g
1 is obtained for LiFePO4/AS-0, superior to both LiFePO4/S-0 and LiFePO4/S-3, and comparable to LiFePO4/S-6 (137.1 mAh g
1) that possesses

4.4

(a)

4.2

LiFePO4/S-0 LiFePO4/S-3

+

Voltage (V vs Li /Li)

4.0

LiFePO4/S-6

3.8

LiFePO4/AS-0

3.6 3.4 3.2 3.0 2.8 2.6 2.4 0

20

40

60

80

100

120

140

63

relatively more carbon amount. The variation of specific capacity among the samples can be attributed to the difference in phosphorus source. In the traditional synthesis strategy, NH4H2PO4 just acts as a reactant to provide P element for chemical composition of LiFePO4/S-0. Its function is quite simple and common, whereas ATMP not only supplies phosphorus source but provides carbon source. Furthermore, the uniformly distributed carbon networks in LiFePO4/AS-0 are more effective in reducing particle sizes and improving electrical conductivity in comparison with LiFePO4/S-3 with uneven carbon distribution, thus leading to the enhanced specific capacity. EIS measurements were performed to figure out the effect of carbon networks on the performance of LiFePO4/AS-0. For the formation of stable SEI film and the percolation of electrolyte through electrode particles, all the cells were performed galvanostatical cycles for three times before impedance measurements. Fig. 6a shows Nyquist plots of LiFePO4/AS-0 and LiFePO4/S-m (m = 0, 3, 6) electrodes. In a typical Nyquist plot, the intercept at the Zreal axis in the high frequency corresponds to the series resistance (Rs), which represents the total resistance of the electrolyte, separator, and electrical contacts. The depressed semicircle in the high-frequency range is related to the charge transfer resistance (Rct) originating from the lithium-ion insertion and deinsertion electrochemical reaction and corresponding double-layer capacitance (CPE). The inclined line in the low-frequency region represents the Warburg impedance (Zw) assigned to the lithium-

Specific capacity (mAh/g)

4.4

(b)

+

Voltage (V vs Li /Li)

4.2

LiFePO4/AS-0 LiFePO4/AS-3

4.0

LiFePO4/AS-6 LiFePO4/AS-9

3.8

LiFePO4/AS-12

3.6 3.4 3.2 3.0 2.8 2.6 2.4

0

20

40

60

80

100

120

140

160

180

Specific capacity (mAh/g)

4.4 4.2

(c)

4.0

Voltage (V)

3.8 3.6 3.4 3.2 3.0

10 C

5C

2C

0.1 C

100

120

140

160

2.8 2.6 2.4 0

20

40

60

80

180

Capacity (mAh/g) Fig. 5. Voltage profiles for the (a) LiFePO4/AS-0, LiFePO4/S-m (m = 0, 3, 6) at 0.1C, (b) LiFePO4/AS-n (n = 0, 3, 6, 9, 12) at 0.1C and (c) LiFePO4/AS-6 at different C-rates.

Fig. 6. The Nyquist plots and equivalent circuit of (a) LiFePO4/AS-0, LiFePO4/S-m (m = 0, 3, 6) and (b) LiFePO4/AS-n (n = 0, 3, 6, 9, 12) electrodes at the end of the 3th cycles at a 0.1 C rate. The inset in (b) is the relationship between Z0 and v
1/2 of LiFePO4/AS in the low frequency region.

64

M. Chen et al. / Electrochimica Acta 168 (2015) 59–68

ion diffusion in the LiFePO4 particles [47,48]. Simplified equivalent circuit model (Fig. 6a(inset)) was constructed to analyze the Nyquist plots. The experimental data well fits with the equivalent circuit and the detailed parameters of the equivalent circuit for all the samples were obtained from computer simulations by using the ZMan software. The EIS data in Table 1 shows that the Rs of LiFePO4/S-m decreases gradually from 6.0 to 3.7 V with the increase of carbon content. While, LiFePO4/AS-0 presents the Rs of 3.9 V, which is much lower than that of LiFePO4/S-3 (4.5 V), and comparable with that of LiFePO4/S-6. As the same addition of conductive carbon black (Super-P) and electrolyte, the decline of Rs originated from the improved electrical conductivity provided by intimate networking of carbon with LiFePO4 particles. Moreover, the Rct of LiFePO4/AS-0 is 76.8 V, much lower than 250.0 V of LiFePO4/S-0 and 125.7 V of LiFePO4/S-3. Theoretically, LiFePO4/S3 with 1.2 wt.% carbon which is a little higher than that of LiFePO4/ AS-0 (1.1 wt.% carbon), would achieve the smaller or at least the equal Rct to that of LiFePO4/AS-0. The opposite results indicate that apart from carbon content, the carbon distribution and even carbon structure may play important roles on the electrochemical reactivity [49–51]. The well-developed conductive carbon networks in LiFePO4/AS-0 particles are more efficient in improving electrical conductivity and electrochemical reactivity, in comparison with that of LiFePO4/S-0 and LiFePO4/S-3. Fig. 7a shows the cyclic voltammetry curves of LiFePO4/AS0 and LiFePO4/S-3 conducted at a scanning rate of 0.1 mV s
1. The

0.8

(a)

LiFePO4/S-3

Current (mA)

LiFePO4/S-6 LiFePO4/AS-0

0.4

0.0

Δ Ep = 0.28 V Δ Ep = 0.33 V

-0.4

Δ Ep = 0.25 V -0.8

2.8

3.0

3.2

3.4

3.6

3.8

4.0

4.2

+

Voltage (V vs. Li/Li )

1.5

(b)

LiFePO4/AS-0 LiFePO4/AS-3

1.0 Current (mA)

LiFePO4/AS-6 LiFePO4/AS-9

0.5

LiFePO4/AS-12

0.0

ΔEp = 0.20 V ΔEp = 0.24 V

-0.5

potential difference (DEp) between the anodic and cathodic peaks is a key parameter to estimate the polarization levels of LiFePO4 [52]. It can be observed that the DEp of LiFePO4/AS-0 is 0.28 V, smaller than that of LiFePO4/S-3 (0.33 V), indicating a lower polarization behavior and a more reversible redox reaction due to the higher electrical conductivity. This difference of LiFePO4/AS0 and LiFePO4/S-3 in electrochemical properties further confirms the advantage of the ATMP used in the synthesis of LiFePO4/C nanocomposites. 3.2. Sucrose assisted synthesis of core-shell structured LiFePO4/C nanocomposites Although the 1.1 wt.% carbon from ATMP is of great importance in improving electrochemical performance of LiFePO4/AS-0, the specific capacity of LiFePO4/AS-0 is not yet high enough for practical application. In order to further enhance the electrochemical behaviors, sucrose was introduced as the second carbon source in the fabrication of LiFePO4/AS-n nanocomposites. Fig. 2b shows the XRD patterns of LiFePO4/AS-n prepared with different sucrose amount. The XRD patterns of LiFePO4/AS-n display high crystallinity and pure phase, consistent with the result of magnetic measurements. The calculated crystallite size of LiFePO4/AS-n (n = 3, 6, 9, 12) based on Scherrer’s equation is 55, 48, 45 and 41  5 nm respectively. The detailed refined lattice parameters for all the samples calculated from the XRD patterns based on the Pmnb space group are listed in Table 1. Interestingly, the values of the lattice parameters of a,b and c were shrunk on increasing carbon content in LiFePO4/AS-n, suggesting the carbon coating slightly altered the crystallinity of the samples [38]. The actual carbon content for the LiFePO4/AS-n calculated from TG curves is listed in Table 1. Particularly, there is about 3.5 wt.% carbon in LiFePO4/AS-6, rendering it a tap density of 1.25 g cm
3. Nevertheless, no obvious changes on the carbon structure can be found in Raman spectra along with the increment of carbon content (Fig. S5). Fig. 8 displayed the SEM and TEM images of LiFePO4/AS-6. SEM image of LiFePO4/AS-6 shows the average particle size was around 500 nm with narrow size distribution, obviously smaller than that of LiFePO4/S-6 (Fig. S3C) with the particle sizes in the scope of hundreds of nanometers to several micrometers. Moreover, compared to that of LiFePO4/AS-0, the particles of LiFePO4/AS6 get smaller and more homogeneous, suggesting that the carbon coating layer arising from sucrose can further inhibit particles conglomerating and reduce particle sizes. As TEM images show, with the assistance of sucrose, a newly approximate 4 nm thick amorphous carbon layer is coated evenly on the surface of LiFePO4/AS-6 particles (Fig. 8B–D), forming a typical core-shell structure. The average crystal size of LiFePO4/AS6 is about 50 nm, which is smaller than 65 nm of LiFePO4/AS-0 (Fig. 4B), further confirming that the carbon shell effectively reduced the crystal sizes. In addition, the carbon content will affect the carbon thickness [53]. Through comparison of Fig. S6A-D, it is found that when the carbon content increases from 2.3 wt.% of LiFePO4/AS-3 to 5.7 wt.% of LiFePO4/AS-12, the corresponding carbon coating layer gets thicker from 3 nm to 9 nm. Moreover, the high-resolution TEM image of LiFePO4/AS-6 (Fig. 8D and F) indicates the presence of regular crystalline lattice and amorphous carbon at the edge of the lattice. The measured lattice pitch d

ð011Þ

-1.0

; dð120Þ and interfacial angle of ð011Þ and (1 2 0) for LiFePO4/

AS-6 are about 0.43, 0.39 nm and 72.5 , respectively. The selected 3.0

3.2

3.4

3.6

3.8

4.0

+

Voltage (V vs. Li/Li ) Fig. 7. Cyclic voltammetry profiles of (a) LiFePO4/AS-0, LiFePO4/S-3 and LiFePO4/S6, and (b) LiFePO4/AS-n (n = 0, 3, 6, 9, 12).

area electron diffraction pattern (Fig. 8E) from ½211 of the LiFePO4/ AS-6 bulk region exhibits a regular and clear parallelogram diffraction spot array, reflecting an ordered orthorhombic structure of LiFePO4. The indexes of crystal planes and d spacing of

M. Chen et al. / Electrochimica Acta 168 (2015) 59–68

65

Fig. 8. (A) SEM and (B–D, F) TEM images of the LiFePO4/AS-6 sample, and SAED patterns of the LiFePO4/AS-6 composite bulk region (E) and particle boundary (D (inset)).

carbon content delivers the highest specific capacity of 158 mAh g
1 at 0.1 C, indicating 3.5 wt.% carbon is the optimal one. In order to further investigate the electrochemical behaviors of LiFePO4/AS-6, the cycle performance at different C-rates were measured (Fig. 5c). The charge–discharge curves show flat voltage plateaus in the range of 3.5–3.2 V. The first specific discharge capacity of LiFePO4/AS-6 is 138.4, 123.2, 102.7 mAh g
1 at 2, 5 and 10C, respectively. Moreover, LiFePO4/AS-6 can maintain a stable 180

0.1 C

160 -1

Discharge capacity (mAhg )

crystal lattice calculated from the SAED pattern fit well with the XRD and HRTEM results, which further confirms the good singlecrystalline characteristic of the sample [54]. In addition, The SAED pattern of LiFePO4/AS-6 (Fig. 8D(inset)) particle boundary with a hollow ring pattern indicates the amorphous phase of the carbon layer, which agrees with the Raman spectra in Fig. S5. For BET surface area and pore size distribution of LiFePO4/AS-n (n = 3, 6, 9, 12), it reveals that the surface area increased a little from 21.3 to 35.0 m2 g
1 with the increase of carbon content, indicating its limited effects on electrochemical performance. Nevertheless, the larger surface area and more uniform pores for the LiFePO4/ASn (n = 3, 6, 9, 12) were observed than those of LiFePO4/S-6 due to the improved homogeneity. The electrochemical properties of carbon-coating LiFePO4/AS-n were firstly examined by charge–discharge tests. For the LiFePO4/ AS-n (n = 0, 3, 6, 9, 12), it is interesting that the specific capacity is getting bigger when the amount of added sucrose increased to 0.6 g, and then smaller on further increasing the amount of sucrose (Fig. 5b). This trend indicates that LiFePO4/AS with higher amount of carbon may have better electrochemical performance than the pristine LiFePO4/AS-0 due to the enhanced surface electrical conductivity originating from the carbon shell, but the excessive carbon formed a thick coating layer on the surface of LiFePO4 particles, which is unfavorable for Li+ percolating through electrolyte to crystal lattice, causing the specific capacity degraded [53]. Thus, the carbon content should be controlled at a moderate level to obtain the optimized carbon coating layer for the good balance between conductivity and Li+ diffusion. The chargedischarge tests demonstrate that LiFePO4/AS-6 with 3.5 wt.%

LiFePO4/AS-6

2C

140

LiFePO4/S-6

5C

120

10 C

100 80 60 40 20 0

0

10

20

30

40

50

60

70

80

90 100 110

Cycle number Fig. 9. Cycle performance of cells with LiFePO4/AS-6 and LiFePO4/S-6 cathodes at 0.1, 2, 5, and 10 C rates.

66

M. Chen et al. / Electrochimica Acta 168 (2015) 59–68

160

(a) 5 C

-1

Specific capacity (mAhg )

140 120 100 80 60 40 20 0

0

100

200

300

400

500

Cycle number

120

(b) 10 C -1

Specifc capacity (mAhg )

100 80 60 40 20 0

0

200

400

600

800

Cycle number Fig. 10. Cycle performance of the LiFePO4/AS-6 at discharging rate of 5 C for 450 cycles (a); and 10 C for 800 cycles (b).

cycle performance at different discharge rates (Fig. 9), whereas the discharge capacity of LiFePO4/S-6 deteriorates dramatically after several cycles at each rate. For the long-term cycle performance, LiFePO4/AS-6 (Fig. 10) still exhibits the specific capacity of ca. 120.0 and 96. 9 mAh g
1 after 450 circles at 5 C and 800 circles at 10 C, respectively. The good electrochemical performance is attributed to the high purity, perfect carbon coating and narrow nanosize distribution. As shown in Scheme 1, upon synthesis process, the ATMP functions as a chelating agent to form a C–P–O–Li space structure before sintering. Meanwhile, the organophosphonic acid ATMP dissolved in water serves as a binder to form a colloid with FeC2O4 and sucrose. The homogeneously reactive precursor with the element distribution of Li: Fe: P near to 1:1:1 not only avoids the occurrence of impurities during the sintering process, but yields the uniform LiFePO4/C particles. Besides, ATMP combining with sucrose forms well-distributed carbon networks intra- and interparticles after carbonization so as to improve the electrical conductivity [25]. Furthermore, the perfect carbon coating from sucrose can also inhibit particles growth and contribute to the preparation of nano-sized LiFePO4/AS-6 particles. Finally, the nano-sized particles with narrow size distribution increases the utilization of active material and shorten the Li+ diffusion pathway, resulting in the favorable performance of Li+ intercalation and deintercalation reaction [55]. On the contrary, the LiFePO4/S6 prepared by NH4H2PO4 and sucrose (Fig. 9) exhibits a specific discharge capacity of 137.1 mAh g
1 at 0.1 C and 95.5 mAh g
1 at 2 C, respectively, which is much lower than those of LiFePO4/AS6 and even smaller to LiFePO4/AS-3 (145.4 mAh g
1 at 0.1 C). It can be attributed to that the solid-state mixed compounds cannot form a very homogeneously reactive precursor without complex procedures (e.g. ball milling) [56]. The prepared LiFePO4/S-6 is more likely to be non-stoichiometric in elements composition, thus affecting the purity of the products. The phenomenon can be confirmed from the EDX spectrum (Fig. S7), in which the surface elements Fe–P–O (Li could not be detected by EDX detector) composition ratio of LiFePO4/AS-6 is 1.00:1.01:4.02, otherwise that of LiFePO4/S-6 is 1.00:1.03:4.25. Another disadvantage is that sucrose cannot form an effective and uniform carbon networks

Scheme 1. The preparation processes of nano-sized LiFePO4/AS particles with a complete carbon coating shell and its electron and Li+ transfer pathway.

M. Chen et al. / Electrochimica Acta 168 (2015) 59–68

between the unevenly solid state mixtures, simultaneously leading to the big and irregular particles in LiFePO4/S-6, which consists well with the above SEM and TEM analysis. Fig. 6b shows that the values of Rct for LiFePO4/AS-n (n = 3, 6, 9, 12) composites are getting smaller with the increase of carbon amount, indicating the trend that carbon coated LiFePO4/AS-n may have better electrochemical performance with the increment of carbon content. But the Galvanostatic charge-discharge tests demonstrated that LiFePO4/AS-6 has the best electrochemical performance. This contradiction may be due to that the carbon coating layer of LiFePO4/AS-6 composite is thinner than those of LiFePO4/AS-9 and LiFePO4/AS-12 composites, as supported by the TEM images (Fig. S6), which facilitates the lithium ion diffusion and results in the optimum electrochemical performance. Importantly, LiFePO4/AS-6 exhibited a lower Rct (34.2 V) than that of LiFePO4/AS-3(48.3 V) and much lower Rct than that of LiFePO4/S-6 (61.5 V, Fig. 6a), demonstrating a higher electrochemical activity. However, the excess carbon results in the decrease of specific capacities in LiFePO4/AS-9 and LiFePO4/AS-12 due to the inferior penetration of electrolyte in and out of the LiFePO4 particles and the slow lithium ion diffusion, although lower Rct was achieved [53]. To further confirm the effect of carbon-coating on the lithium ion diffusion of LiFePO4/AS composites, the lithium ion diffusion coefficients are calculated according to Eq. (1) [57]: D¼

R2 T 2 2A2 n4 F 4 C 2

s2

(1)

where R is the gas constant, T is the absolute temperature, A is the surface area of the cathode (geometric area 1.33 cm2), n is the number of electrons per molecule during oxidization, F is the Faraday constant, C is the concentration of lithium ion (7.69  10
3 mol cm
3 in this study), s is the Warburg factor associated with Z0 . Z 0 ¼ RD þ RL þ sv
1=2

(2)

67

may bring in higher passivation and the lower reactivity. Therefore, thanks to the optimum uniform carbon-coating layer, LiFePO4/AS6 has an improved electrochemical kinetics and proper passivation. Moreover, the smallest DEp suggests the optimum reversible electrochemical reactions of LiFePO4/AS-6, in good agreement with its long cycle life of 800 cycles (Fig. 10b). Additionally, compared to those of the LiFePO4/S-m, anodic and cathodic peak current intensities of the LiFePO4/AS-6 and the LiFePO4/AS-3 are much larger, implying that Li-ions and electrons were participating more actively in redox reactions due to the enhanced electrical and ionic conductivity. The DEp is 0.20 V and 0.24 V for LiFePO4/AS-6 and LiFePO4/AS-3, respectively, smaller than that of LiFePO4/S-6 (0.25 V, Fig. 7a). The smaller DEp of LiFePO4/AS-6 and LiFePO4/ AS-3 composite represents higher electrochemical reactivity and lower series resistance, which is in good agreement with EIS results. 4. Conclusions High-purity and core-shell structured LiFePO4/C nanocomposites were synthesized by a quasi-sol–gel route using organophosphonic acid ATMP as phosphorus and carbon co-precursor, and sucrose as assistant carbon source. The samples LiFePO4/AS0 with 1.1 wt.% carbon and LiFePO4/AS-6 with 3.5 wt.% carbon display an initial specific capacity of 132.5 and 158.6 mAh g
1 at 0.1 C, respectively. Moreover, LiFePO4/AS-6 exhibits excellent cycle stability and rate performance. The outstanding electrochemical performance of the LiFePO4/AS-6 composites is attributed to its high purity, narrow size distribution in nanoscale, and completely uniform carbon coating with appropriate carbon layer thickness. The simple and environmental-friendly synthesis route is very promising to produce LiFePO4/C cathode material for high-power lithium-ion batteries, and can also be extended to the synthesis of other olivine type LiMPO4 (M = Mn, Ni, Co, V) cathode materials. Acknowledgements

Fig. 6b(inset) shows the relationship between Z0 and reciprocal square root of angular frequency (v
1/2) in the low-frequency region, in which the slope of the beeline is the Warburg factor s. Based on Eqs. (1) and (2), the calculated lithium ion diffusion coefficients for LiFePO4/AS-n (n = 0, 3, 6, 9, 12) nanocomposites with different carbon contents are correspondingly 2.36  10
14, 5.41 10
14, 4.43  10
13, 1.06  10
13 and 6.45  10
14 cm2 s
1. The values are comparable to the previous reported data (10
13 to 10
14 cm2 s
1 by sucrose) [57]. This indicates that the diffusion velocity of lithium-ion, which is considered to be the ratedetermine step in the redox process, has been dramatically promoted by the carbon-coating strategy. The enhanced diffusion capability is rooted from the nanosized core-shell architectures which provide shorter diffusion paths as well as more accessible space for Li+ intercalation and de-intercalation. Nevertheless, the values of diffusion coefficients decrease with further increment of carbon for the LiFePO4/AS-9 and LiFePO4/AS-12 after it reaches the highest for LiFePO4/AS-6. It is indicated that though the lithium ion diffusion pathway was extensively shortened as the reduction of crystal particle sizes from the strengthened carbon-coating effect, the thicker carbon-coating layer evolved to be the barrier for Li+ diffusion, thus leading to the slower lithiation and delithiation reaction and deterioration of the electrochemical performance. The CVs of LiFePO4/AS-n (n = 0, 3, 6, 9, 12) in Fig. 7b show a trend that the DEp turns smaller and then bigger with the increasing of carbon content. The same change tendency was also observed for anodic and cathodic peak current intensities. In practice, the current intensity in CVs is related to the passivation level of the electrode surface [58]. The excessive thick carbon coating layer

This work was supported by the National Natural Science Foundation of China (21421001), the Natural Science Foundation of Tianjin (13JCZDJC32000), the 111 project (B12015), and the Key Laboratory of Advanced Catalytic Materials in Zhejiang Normal University (ZJHX201301). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j. electacta.2015.04.004. References [1] N.N. Sinha, C. Shivakumara, N. Munichandraiah, ACS Appl. Mater. Interfaces 2 (2010) 2031–2038. [2] Y. Zhou, J. Wang, Y. Hu, R. O’Hayre, Z. Shao, Chem. Commun. 46 (2010) 7151–7153. [3] W. Kim, W. Ryu, D. Han, S. Lim, J. Eom, H. Kwon, ACS Appl. Mater. Interfaces 6 (2014) 4731–4736. [4] X. Huang, S. Yan, H. Zhao, L. Zhang, R. Guo, C. Chang, X. Kong, H. Han, Mater. Charact. 61 (2010) 720–725. [5] C. Masquelier, L. Croguennec, Chem. Rev. 113 (2013) 6552–6591. [6] L.-X. Yuan, Z.-H. Wang, W.-X. Zhang, X.-L. Hu, J.-T. Chen, Y.-H. Huang, J.B. Goodenough, Energy Environ. Sci. 4 (2011) 269–284. [7] C. Nan, J. Lu, C. Chen, Q. Peng, Y. Li, J. Mater. Chem. 21 (2011) 9994–9996. [8] T. Zhang, D. Xin, Y. Lu, G. Luo, Ind. Eng. Chem Res. 53 (2014) 6723–6729. [9] C. Delacourt, P. Poizot, S. Levasseur, C. Masquelier, Electrochem. Solid-St. 9 (2006) A352–A355. [10] X. Yan, G. Yang, J. Liu, Y. Ge, H. Xie, X. Pan, R. Wang, Electrochim. Acta 54 (2009) 5770–5774. [11] H. Gao, L. Jiao, J. Yang, Z. Qi, Y. Wang, H. Yuan, Electrochim. Acta 97 (2013) 143–149. [12] K.-F. Hsu, S.-Y. Tsay, B.-J. Hwang, J. Mater. Chem. 14 (2004) 2690–2695.

68

M. Chen et al. / Electrochimica Acta 168 (2015) 59–68

[13] J. Yang, J.J. Xu, Electrochem. Solid-St. 7 (2004) A515–A518. [14] M.-R. Yang, T.-H. Teng, S.-H. Wu, J. Power Sources 159 (2006) 307–311. [15] H. Gao, Z. Hu, K. Zhang, F. Cheng, J. Chen, Chem. Commun. 49 (2013) 3040–3042. [16] Y.-P. Zhu, Y.-L. Liu, T.-Z. Ren, Z.-Y. Yuan, Nanoscale 6 (2014) 6627–6636. [17] Y.-P. Zhu, Y.-L. Liu, T.-Z. Ren, Z.-Y. Yuan, RSC Advances 4 (2014) 16018–16021. [18] T.-Y. Ma, X.-Z. Lin, X.-J. Zhang, Z.-Y. Yuan, Nanoscale 3 (2011) 1690–1696. [19] Y. Yang, G.H. Yu, J.J. Cha, H. Wu, M. Vosgueritchian, Y. Yao, Z.N. Bao, Y. Cui, ACS Nano 5 (2011) 9187–9193. [20] C. Lai, Q. Xu, H. Ge, G. Zhou, J. Xie, Solid State Ionics 179 (2008) 1736–1739. [21] C.-Y. Chiang, H.-C. Su, P.-J. Wu, H.-J. Liu, C.-W. Hu, N. Sharma, V.K. Peterson, H.W. Hsieh, Y.-F. Lin, W.-C. Chou, C.-H. Lee, J.-F. Lee, B.-Y. Shew, J. Phys. Chem. C 116 (2012) 24424–24429. [22] P.S. Herle, B. Ellis, N. Coombs, L.F. Nazar, Nat. Mater. 3 (2004) 147–152. [23] M.R. Roberts, A.D. Spong, G. Vitins, J.R. Owen, J. Electrochem. Soc. 154 (2007) A921–A928. [24] J. Ying, M. Lei, C. Jiang, C. Wan, X. He, J. Li, L. Wang, J. Ren, J. Power Sources 158 (2006) 543–549. [25] A.D. Spong, G. Vitins, J.R. Owen, J. Electrochem. Soc. 152 (2005) A2376–A2382. [26] S. Ferrari, R.L. Lavall, D. Capsoni, E. Quartarone, A. Magistris, P. Mustarelli, P. Canton, J. Phys. Chem. C 114 (2010) 12598–12603. [27] X. Zhi, G. Liang, L. Wang, X. Ou, L. Gao, X. Jie, J. Alloy. Compd. 503 (2010) 370–374. [28] Y. Wang, R. Mei, X. Yang, Ceram. Int. 40 (2014) 8439–8444. [29] B. Zhao, Y. Jiang, H. Zhang, H. Tao, M. Zhong, Z. Jiao, J. Power Sources 189 (2009) 462–466. [30] K.-F. Hsu, S.-Y. Tsay, B.-J. Hwang, J. Mater. Chem. 14 (2004) 2690–2695. [31] J.J. Tan, A. Tiwari, Nanosci. Nanotech. Lett. 3 (2011) 487–490. [32] Z.-y. Chen, H.-l. Zhu, S. Ji, R. Fakir, V. Linkov, Solid State Ionics 179 (2008) 1810–1815. [33] A. Yamada, S.C. Chung, K. Hinokuma, J. Electrochem. Soc. 148 (2001) A224–A229. [34] C.M. Julien, K. Zaghib, A. Mauger, M. Massot, A. Ait-Salah, M. Selmane, F. Gendron, J. Appl. Phys. 100 (2006) 063511–063517. [35] G. Arnold, J. Garche, R. Hemmer, S. Ströbele, C. Vogler, M. Wohlfahrt-Mehrens, J. Power Sources 119–121 (2003) 247–251. [36] S. Franger, F. Le Cras, C. Bourbon, H. Rouault, J. Power Sources 119–121 (2003) 252–257.

[37] H.M. Xie, R.S. Wang, J.R. Ying, L.Y. Zhang, A.F. Jalbout, H.Y. Yu, G.L. Yang, X.M. Pan, Z.M. Su, Adv. Mater. 18 (2006) 2609–2613. [38] Y.L. Cao, L.H. Yu, T. Li, X.P. Ai, H.X. Yang, J. Power Sources 172 (2007) 913–918. [39] M. Sanchez, G. Brito, M. Fantini, G. Goya, J. Matos, Solid State Ionics 177 (2006) 497–500. [40] C.M. Julien, A. Ait-Salah, A. Mauger, F. Gendron, Ionics 12 (2006) 21–32. [41] A. Ait Salah, A. Mauger, K. Zaghib, J.B. Goodenough, N. Ravet, M. Gauthier, F. Gendron, C.M. Julien, J. Electrochem. Soc. 153 (2006) A1692–A1701. [42] A. Yamada, Y. Takei, H. Koizumi, N. Sonoyama, Ryoji Kanno, Chem. Mater. 18 (2006) 804–813. [43] S. Yoon, C. Liao, X.-G. Sun, C.A. Bridges, R.R. Unocic, J. Nanda, S. Dai, M.P. Paranthaman, J. Mater. Chem. 22 (2012) 4611–4614. [44] V. Palomares, A. Goñi, I.G.d. Muro, I. de Meatza, M. Bengoechea, I. Cantero, T. Rojo, J. Power Sources 195 (2010) 7661–7668. [45] M.M. Doeff, J.D. Wilcox, R. Kostecki, G. Lau, J. Power Sources 163 (2006) 180–184. [46] Y. Orikasa, T. Maeda, Y. Koyama, H. Murayama, K. Fukuda, H. Tanida, H. Arai, E. Matsubara, Y. Uchimoto, Z. Ogumi, Chem. Mater. 25 (2013) 1032–1039. [47] G. Tan, F. Wu, L. Li, R. Chen, S. Chen, J. Phys. Chem. C 117 (2013) 6013–6021. [48] S. Yang, H. Song, X. Chen, Electrochem. Commun. 8 (2006) 137–142. [49] Y. Wang, Y. Wang, E. Hosono, K. Wang, H. Zhou, Angew. Chem. Int. Ed. 47 (2008) 7461–7465. [50] M.M. Doeff, Y. Hu, F. McLarnon, R. Kostecki, Electrochem. Solid-St. 6 (2003) A207–A209. [51] R. Dominko, M. Gaberš9 cek, J. Drofenik, M. Bele, S. Pejovnik, Electrochem. SolidSt. 4 (2001) A187–A190. [52] Z.G. Lu, M.F. Lo, C.Y. Chung, J. Phys. Chem. C 112 (2008) 7069–7078. [53] R. Dominko, M. Bele, M. Gaberscek, M. Remskar, D. Hanzel, S. Pejovnik, J. Jamnik, J. Electrochem. Soc. 152 (2005) A607–A610. [54] C. Sun, S. Rajasekhara, J.B. Goodenough, F. Zhou, J. Am. Chem. Soc. 133 (2011) 2132–2135. [55] X.-L. Wu, L.-Y. Jiang, F.-F. Cao, Y.-G. Guo, L.-J. Wan, Adv. Mater. 21 (2009) 2710–2714. [56] Y. Zhang, Q.-y. Huo, P.-p. Du, L.-z. Wang, A.-q. Zhang, Y.-h. Song, Y. Lv, G.-y. Li, Synthetic Met. 162 (2012) 1315–1326. [57] J. Li, Q. Qu, L. Zhang, L. Zhang, H. Zheng, J. Alloy. Compd. 579 (2013) 377–383. [58] Y.-D. Cho, G.T.-K. Fey, H.-M. Kao, J. Power Sources 189 (2009) 256–262.