LiTi2O4 nanocomposite as anode materials for Li-ion secondary batteries

LiTi2O4 nanocomposite as anode materials for Li-ion secondary batteries

Available online at www.sciencedirect.com Journal of Electroanalytical Chemistry Journal of Electroanalytical Chemistry 616 (2008) 7–13 www.elsevier...

735KB Sizes 2 Downloads 28 Views

Available online at www.sciencedirect.com Journal of

Electroanalytical Chemistry Journal of Electroanalytical Chemistry 616 (2008) 7–13 www.elsevier.com/locate/jelechem

Electrochemical properties of Si/LiTi2O4 nanocomposite as anode materials for Li-ion secondary batteries Z.Y. Zeng, J.P. Tu *, X.L. Wang, X.B. Zhao Department of Materials Science and Engineering, Zhejiang University, Hangzhou 310027, China Received 18 September 2007; received in revised form 17 December 2007; accepted 24 December 2007 Available online 11 January 2008

Abstract Si/LiTi2O4 nanocomposite film was synthesized by a sol–gel method in combination with a following heat-treatment process. Through this process, the nanosized Si particles were homogeneously distributed in the porous composite. The electrochemically less active LiTi2O4 working as a buffer matrix successfully prevented Si from cracking/crumbling during the charging/discharging process. The Si/LiTi2O4 nanocomposite exhibited a reversible lithium storage capacity of about 1100 mA h g1 with good cyclability, suggesting its promising nature in anode materials for Li ion batteries. A two-parallel diffusion path model fitted well with the impedance data of porous Si/LiTi2O4 composite at different potentials. The impedance in the low-frequency region was directly related to the diffusion of Li ions, and the Li ion diffusivity was on the order of 1013 cm2 s1 based on the semi-infinite diffusion model. Ó 2008 Elsevier B.V. All rights reserved. Keywords: Si/LiTi2O4; Nanocomposite film; Sol–gel; Anode; Li ion batteries

1. Introduction Silicon is recognized as a potentially high energy per unit volume host material for Li ion battery applications [1]. Attempts at realizing this potential have met with only partial success when nanocomposites of silicon powder and carbon black have been used. The major technical problem associated with the use of silicon/lithium appears to be the mechanical failure brought about by the repeated large volume expansion associated with alloying. Studies have been conducted to overcome this problem by preparation of nanosized Si [2,3], Si based composite anodes [4–10]. A nanocrystalline state only slightly improved the cyclic stability probably due to the reduced density of atoms in a nanosized grain [11]. Recently, a core-shell structure using an inactive/less active phase envelop the active particles has developed. It exhibited an improved electrochemical performances and cyclabilities [12,13]. Nevertheless, this *

Corresponding author. Tel.: +86 571 87952573; fax: +86 571 87952856. E-mail address: [email protected] (J.P. Tu). 0022-0728/$ - see front matter Ó 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.jelechem.2007.12.017

core-shell structure diminished the contact between active materials and electrolyte, which influenced the electrochemical reactive kinetics of active materials with electrolyte. Considering the demerit emerged above, a porous nanocomposite synthesized directly by a surface sol–gel process was investigated [14]. In this nanocomposite electrode, the inactive/less active agents buffered the volume expansion of Si particles during lithiation/delithiation. Furthermore, the active nanoparticles distributed homogeneously in inactive/less active matrix, but were not enveloped by it. Therefore, the porous structure of nanocomposite could facile the contact of active particles with electrolyte. LiTi2O4 has a cubic structure and is known to react with further Li atoms to form zero-strain insertion materials, with reversible Li storage capacity of about 100 mA h g1, and the standard discharge/charge plateau ranging from 1.3 to 1.7 V under a steady state [15–17], which proves a good candidate as a matrix of the Si powder because of its favorable conductivity and structural stability. In this present work, Si/LiTi2O4 nanocomposite film was synthesized by a sol–gel method. Si nanoparticles were dispersed

Z.Y. Zeng et al. / Journal of Electroanalytical Chemistry 616 (2008) 7–13

homogeneously in the porous LiTi2O4 matrix. The electrochemical properties of the nanocomposite film as anode material for lithium ion batteries were investigated by the galvanostatic charge/discharge method, slow rate cyclic voltammogram and electrochemical impedance spectrum (EIS). 2. Experimental Nanosized Si powder with particle sizes of 30–50 nm was prepared by laser driven silane gas reaction. For sol– gel synthesis, tetrabutyl titanate [Ti(OC4H9)4], triethanolamine (TEA), lithium acetate [LiAc  2H2O] and ethanol were used as the precursor reagents. In a typical synthesis, 3.4 g (0.01 mol) Ti(OC4H9)4 was dissolved in 10 g of ethanol. LiAc  2H2O (0.864 g, 0.008 mol), 1.19 g (0.008 mol) TEA and 1 g poly block copolymer (EO20PO70EO20) were added to the resulting solution with vigorous stirring. During this process, 1.08 g H2O was gradually dropped into this solution and then a crystal clear solution was obtained. After slow hydrolysis for several hours, 1.12 g Si nanoparticles were added and sonicated for 1 h to achieve a uniform sol. At last, 1.16 g poly(vinylpyrrolidone) (PVP) was added as the film-forming assistant. When it became viscous, the sol solution containing Si nanoparticles was spread on a Cu sheet (with 1.0 cm in diameter) by KW4A spin-coating machine. The thickness of smooth film on the Cu substrate could be obtained by controlling the rotation speed and time of the spin-coating process. After aging at 50 °C for 10 h, the obtained gel film was calcined at 600 °C under flowing argon to yield the Si/LiTi2O4 composite. In this work, the composite film contained 57.6 wt.% Si, i.e., the mole ratio of Si and LiTi2O4 was 8:1, and the total mass of the Si/LiTi2O4 composite film was 0.6 mg cm2. The thickness of the composite film was about 600 nm, as determined by an Alpha-stop 200 profilometry. For fabricating the pristine Si anode, a slurry consisting of 80 wt.% nanoscale Si powder as active material, 10 wt.% acetylene black as electronic conductor and 10 wt.% polyvinylidene fluoride (PVDF) dissolved in N-methyl-2pyrrolidone (NMP) as a binder was pasted onto a preweight Cu sheet. The electrodes were dried overnight at 100 °C in vacuum and pressed at 5 MPa for 10 min to enhance the contact between the active material and current collector. Low-angle and wide-angle X-ray powder diffraction (XRD, Rigaku D/max-rA) with Cu Ka radiation ˚ ) was used to analyze the mesostructure and (k = 1.5406 A phase of the calcined product. The morphology and microstructure of the composite film were characterized by a scanning electron microscopy (SEM, SIRION JY/T0101996) and a transmission electron microscopy (TEM, JEM-2010). For electrochemical tests, two-electrode cells were assembled into the coin-type cells (CR 2025) in an argonfilled glove box. The as-prepared Si/LiTi2O4 composite film

on Cu substrate was used as the working electrode and a metallic lithium foil as the counter electrode. A 1M LiPF6 in a mixture of ethylene carbonate (EC) and dimethyl carbonate (DMC) (1:1 volume ration) was used as the electrolyte, and a polypropylene (pp) film (Celgard 2300) as the separator. The galvanostatic charge–discharge tests were conducted on a BS9300 Battery Program-Control Test System at room temperature with the cut-off voltages of 0.02 V and 3 V (vs. Li/Li+) at a specific current density of 160 mA g1 (0.07 C, supposing the theoretical charge/ discharge capacity was about 2230 mA h g1). Cyclic voltammetry (CV) measurements were performed on a CHI640B Electrochemical Workstation with a scan rate of 0.1 mV s1 between 0 and 3.0 V (vs. Li/Li+). The electrochemical impedance spectroscopy (EIS) were potentiostatically measured by applying an ac voltage of 5 mV amplitude over the frequency range 10 kHz–10 mHz after the electrode had attained an equilibrium at each potential. 3. Results and discussion Fig. 1 shows a typical XRD pattern of the Si/LiTi2O4 nanocomposite, prepared by sol–gel method using TEA as a chelating agent, calcined at 600 °C under flowing argon for 3 h. From this pattern, in addition to the sharp peaks of Si powder, some of broaden peaks belonging to the spinels LiTi2O4, are also detected. It is proved that the LiTi2O4 compound is formed but the grain size is small or not well crystallized. Furthermore, some weak peaks attributed to Li2CO3, the main component of SEI film, are also detected. This Li2CO3 should be formed during the aging of the LiTi2O4. First, some of the lithium ions (from the acetate) are not incorporated to the oxide during the sol–gel preparation process, second, CO2 present during the preparation process because of the use of no degasified water. So, they react to form Li2CO3, which is detected by XRD. The morphologies of Si/LiTi2O4 nanocomposite were examined by SEM. As shown in Fig. 2a, the composite are composed of small particles, which are characteristic of sol–gel derived materials comprising spherical nanopar-

Si LiTi2O4 Li2CO3

Intensity

8

10

20

30

40

50

60

70

80

2θ (Degree)

Fig. 1. XRD pattern of Si/LiTi2O4 nanocomposite.

Z.Y. Zeng et al. / Journal of Electroanalytical Chemistry 616 (2008) 7–13

9

Fig. 2. (a) SEM micrograph of the Si/LiTi2O4 composite (Si:LiTi2O4 = 8:1) obtained by a sol–gel method. The elemental Si (b) and Ti (c) maps. (d) Magnified SEM image, insert a low-angle XRD pattern of the Si/LiTi2O4 composite.

ticles. EDX spot elemental analysis was conducted to examine the distribution of the species within the composite film. Based on the EDX element maps, as shown in Fig. 2b and c, both Si and Ti elements are distributed homogeneously within the composite film, indicating the homogeneous distribution of Si nanoparticles in the LiTi2O4 matrix. From a magnified SEM image shown in Fig. 2d, it can be seen that the composite is porous with some Si nanoparticles (show white and raised shape) dispersed in it. Three marks signed in 1, 2, 3 show the diameter of the pore endowed with 14.9, 13.9 and 13.9 nm, respectively. This mesostructured composite that is prepared using EO20PO70EO20 as the structure-directing agent is also conformed by a low-angle XRD pattern that is inserted in Fig. 2d. From this pattern, it can be seen that the calcinated LiTi2O4 mesostructure shows two diffraction ˚ . These diffraction peaks can peaks with d = 113 and 56.5 A be indexed as the (1 0 0) and (2 0 0) reflections from twodimensional (2D) hexagonal mesostructures with lattice ˚ . The formation of the porous strucconstant a0 = 130.5 A ture results from the removing of residual polymers (EO20PO70EO20) during the calcining process. Fig. 3a shows the cycling performances of the Si/ LiTi2O4 nanocomposite film with cut-off voltages from 0.02 to 3 V (vs. Li/Li+) at a current density of 160 mA g1 (0.07 C). The discharge capacity is 1508 mA h g1 for the first cycle, but it drops rapidly to 1127.9 mA h g1 in the

second cycle, i.e., the initial irreversible discharge capacity of Si/LiTi2O4 nanocomposite is 380.1 mA h g1 (25.2%). In addition to the insulting solid electrolyte interphase (SEI) film, the initial irreversible capacity of Si/LiTi2O4 composite can also be ascribed to a few oxidized films of Si nanoparticles formed during the preparation process [18–20]. In which case, Li+ reacted with SiO during the first insertion process and produced some nanosized clusters with silicon distributed in Li2O matrix. These nanoclusters became electrochemically inactive because of poor charge-transferring environment and poor cohesion between silicon and Li2O [20]. From Fig. 3, it can be seen that the coulombic efficiency of the first cycle is not high enough but increases rapidly. It needs few cycles to establish electrochemically stabled electrode for Li+ insertion/extraction. From the 2nd to 50th cycle, both the discharge and charge capacities keep a steady level of about 1100 mA h g1, the coulombic efficiencies achieve as high as 94% and upwards, the average discharge capacity fading from the 2nd to 50th cycle is 1.7 mA h g1 (0.15%), indicating excellent cyclic ability and good electrochemical performance. Such promoted cyclability is attributed to the effect of less active LiTi2O4. As a less active matrix in the composite film, the LiTi2O4 could effectively alleviate the damage to electrode structure caused by the volume change of Li–Si alloy during lithiation/delithiation. Li–Si alloy was surrounded by porous LiTi2O4 and restricted in an enough space for its

10

Z.Y. Zeng et al. / Journal of Electroanalytical Chemistry 616 (2008) 7–13 0.16

2400 1.0

0.6

1200

0.4 discharge charge efficiency

600

0

0

10

Current (mA)

0.8

Efficiency (100 %)

-1

Capacity (mAh g )

0.08

1800

0.00

-0.08 Si/LiTi2O4

-0.16

Si

0.2

20

30

40

-0.24

0.0 50

0.0

0.5

2.0

2.5

3.0

Potential (V vs. Li/Li )

Fig. 4. Cycle voltammetry of Si and Si/LiTi2O4 nanocomposite between 0 and 3.0 V (vs. Li/Li+) at a scan rate of 0.1 mV s1.

3.5 Si/LiTi2O4

3.0

Si 2.5

Voltage (V)

1.5

+

Cyclic number

2.0 1.5 1.0 0.5 0.0

1.0

0

700

1400

2100

2800

3500

-1

Capacity (mAh g )

Fig. 3. (a) The galvanostatic charge/discharge capacity and coulombic efficiency of Si/LiTi2O4 nanocomposite as a function of cycle number. (b) The first galvanostatic charge/discharge profiles of Si and Si/LiTi2O4 composite.

expansion/contraction, which could prevent the exfoliation and the contact loss of active Si. On the other hand, lithium ions in electrolyte could facilely get across these pores to reach the surface of active Si, i.e., the contact area between electrode and electrolyte was enlarged. As a result, the transfer distance of lithium ions was abbreviated and lithium ions could react efficiently with Si during discharge– charge, which guaranteed good electrochemical performance of the composite film anode. Fig. 3b compares the first galvanostatic charge/discharge profiles of Si and Si/ LiTi2O4 composite. In the discharge state, there are three discharge plateaus centred at 1.7, 1.25, 0.05 V for the Si/ LiTi2O4 composite whereas there is only one discharge plateau centred at 0.05 V for the Si anode. In the charge state, the charge plateau of the Si/LiTi2O4 composite that is attributed to the reaction of Li with Si is centred at 0.5 V, but for the pristine Si anode, the charge plateau centred at 0.35 V, indicating the higher polarization of the Si/ LiTi2O4 composite than the pristine Si nanoparticles. Fig. 4 shows typical cyclic voltammograms of the Si and Si/LiTi2O4 nanocomposite film between 0 and 3.0 V (vs. Li/Li+) at a scan rate of 0.1 mV s1. For the bare nano-

sized Si anode, only one cathodic peak located at 0.05 V and a broad peak centred at 0.5 V are observed. Combine with the peaks attributed to Li1+xTi2xO4 [17,21], we can get the voltammetric behavior of the Si/LiTi2O4 composite. For this composite, five cathodic peaks (Li insertion state) centred at 1.73, 1.56, 1.30, 0.82 and 0.30 V, and five anodic peaks (Li extraction state) located at 0.35, 0.53, 1.65, 2.10 and 2.50 V were observed. One cathodic peak located at 0.30 V and two anodic peaks located at 0.35 and 0.53 V were attributed to the reversible reaction of Li with Si, as indicated in the following equation: Lix Si $ xLiþ þ e þ Si

½7

ð1Þ

Then, the cathodic peaks located at around 1.73, 1.56 and 1.30 V, and the corresponding anodic peaks around 2.10 and 1.65 V, were attributed to the reaction of Li with spinel Li1+xTi2xO4 (0 6 x 6 1/3) and few anatase, which occurred as follows: Liþ þ e þ LiTi2 O4 $ Li2 Ti2 O4 E ¼ 1:30=1:65 V ½15

ð2Þ

þ

Li þ e þ Li½Li1=3 Ti5=3 O4 $ Li2 ½Li1=3 Ti5=3 O4 E ¼ 1:56=1:65 V

½16; 22

Liþ þ e þ TiO2 ðanatase; sÞ $ Lix TiO2 ðsÞ E ¼ 1:73=2:10 V ½23

ð3Þ ð4Þ

Besides, there was a cathodic peak located at 0.82 V yet, which might be attributed to the reaction of Li+ with not well crystallized titanium-based oxides [24]. And an anodic peak located at 2.50 V should be attributed to the decomposition of part of the Li2O [25,26]. According to Poizot’s research [27], the transition metal oxides, with their nanometric characters, could decrease the binding energy of Li2O tremendously, as a result, Li2O should be easy to decompose. Fig. 5 shows the XRD pattern of the Si/LiTi2O4 composite at discharge state (Li ion insertion). From the XRD pattern, it can be seen that while Li ions were being inserted, the Bragg diffraction peaks of Si decreased with

Z.Y. Zeng et al. / Journal of Electroanalytical Chemistry 616 (2008) 7–13

11

1500 Li2Ti2O4

TiO1.8

Li12Si7

1200

-ZIM (Ω)

Intensity

Si

900

600 0.35 V 1.68 V 2.10 V 2.52 V

300 10

20

30

40

50

60

70

80

0

2θ (degree)

0

300

600

900

1200

1500

80

100

ZRE (Ω)

Fig. 5. XRD pattern of Si/LiTi2O4 composite after electrochemical Li insertion. 75 0.35 1.68 2.10 2.52

60

-ZIM (Ω)

Li12Si7 peaks appeared, which proved that Li ions reacted with Si to form the Li12Si7 phase [28]. Besides, there are Li2Ti2O4 and TiO1.8 peaks appeared in the XRD pattern, indicating that the reaction of Li with Li1+xTi2xO4 occurred, which consist with the results obtained from the cyclic voltammograms. To investigate in more detail the electrochemical behavior of the Si/LiTi2O4 composite film, impedance spectrum measurements were conducted. Fig. 6 shows the variation of Nyquist plots of the Si/LiTi2O4 nanocomposite with decreasing the potential from 2.52 V to 0.35 V in the frequency range of 10 kHz to 10 mHz. The overall shape of Nyquist plot was two depressed semicircles and a capacitive component. Qualitatively, all the spectra in Fig. 6a have similar features: a large flat semicircle related to the high to medium frequencies, a small semicircle related to the low frequencies, and a sloping line of a changing slope, which becomes very steep at the lowest frequencies (the millihertz region). Considering that there are some differences between the time constants for diffusion and for double-layer charging that the impedance lies on the real-axis for the intermediate frequencies. In other word, there is a shoulder on the low-frequency end of the RC loop [29]. And as the potentials of the composite varies from 0.35 V to 2.52 V, at intermediate frequencies, the aparting distances of this ‘‘shoulder” from the real axis increases, as seen in the magnified graph in Fig. 6b. The interpret of this interesting phenomenon is that the time constant for the film capacitance in parallel with the faradaic impedance is not completely separated from that for particle diffusion. According to the previously published two-parallel-diffusion-path model [30], the Si/LiTi2O4 composite can be simplified as a porous electrode in which there are two different spherical particle size distributions (radius Rs,1 and Rs,2). Combining the impedance models of Meyers et al. [29], Barsoukov et al. [31] and Levi et al. [30,32], we present the schematic diagram of this model in Fig. 7. From Fig. 7a, there are several processes which occur: ion conduction in solution, diffusion of ions in solution to the interface, migration of Li ions through the surface films,

V V V V

45

30

15

0

0

20

40

60

ZRE (Ω)

Fig. 6. Impedance spectra of Si/LiTi2O4 nanocomposite at the potentials from 2.52 V to 0.01 V, electrode area 1 cm2, and frequency range 10 kHz to 10 mHz.

charge transfer, solid state diffusion, and finally, accumulation, which is a capacitive-type behavior, and there is a phase-transfer reaction if in case several phases are present. One point to emphasize is that because of the porous characteristic of the composite electrode, the distributed resistance representing electronic and ionic resistance of the layer on active material results in ‘‘stretching” of impedance along the X-axis, and thus makes the RC loop squashed [32,33]. Hence, an equivalent circuit of two-parallel diffusion paths which are suitable for describing these processes is presented in Fig. 7b. The corresponding impedance equations are presented below. The impedance of an individual particle with ion inserts/ deinserts, Zpart,i can be presented as follows: Z part;i ¼

Rct;i þ

Rpart;i Y s;i

1 þ jxC dl;i ðRct;i þ

Rpart;i Þ Y s;i

þ

Rsl;i 1 þ jxRsl;i C sl;i

ð5Þ

with the finite-space diffusion resistive element, Rpart,i, of the follow form: Rpart;i ¼

R2s;i si ¼ 3D0 C part;i 3C part;i

ð6Þ

12

Z.Y. Zeng et al. / Journal of Electroanalytical Chemistry 616 (2008) 7–13

(1/Ys,i) is a spherical analogue of the characteristic function of the angular frequency (x) in the expression for a linear finite-space Warburg element: pffiffiffiffiffiffiffiffiffi 1 tanhð jxsi Þ pffiffiffiffiffiffiffiffiffi ð7Þ ¼ pffiffiffiffiffiffiffiffiffi Y s;i ð jxsi Þ  tanhð jxsi Þ Here, Cpart,i stands for the limiting low-frequency capacitance of a spherical particle, Rs,i is the diameter of the spherical particle, D0 is the chemical diffusion coefficient, si designates the related diffusion time constant, and x is defined as the angular frequency of the alternative current. Subscript i denotes particles with different radius, e.g., either 1 or 2 (Fig. 7). The total admittance of an electrode comprising a mixture of two types of particles in terms of different size, 1/ Zmix, is regarded as an averaged sum of the individual admittances 1/Zpart,1 and 1/Zpart,2 1 h1 ð1  h1 Þ ¼ þ Z mix Z part;1 Z part;2

ð8Þ

where h1 is the fraction of the total capacity due to a contribution of the ‘‘small” particles. The distributed impedance of the porous electrode, Zporous, related to the impedance of the mixed particle electrode as follows from the model of Meyers et al. [29]

Fig. 7. (a) Schematic presentation of kinetic steps involved in Li intercalation into Si/LiTi2O4 composite; (b) Equivalent circuit analogue for a two-parallel diffusion path model accounting for the ion migration through the surface layer covering each particle (high-frequency RslkCsl semicircle, with Rsl and Csl standing for the resistance of the migration and capacity of the layer, respectively), double-layer charging and slow interfacial charge transfer (middle-frequency RctkCdl semicircle, with Rct and Cdl designating the related charge-transfer resistance and a doublelayer capacitance) and a finite-diffusion Warburg element, FSW. Indices 1 and 2 refer to path (branch) 1 and 2, respectively.

Z porous ¼

  2 þ ðrj þ jrÞ cosh v L 1þ v sinh v jþr

ð9Þ

with a complex parameter v of the form V ¼ Lð

j þ r 1=2 a 1=2 Þ ð Þ jr Z mix

ð10Þ

where k and r are defined as the specific conductivities of the electrolyte solution which surrounds the particles and the solid particles, respectively, L is the thickness of the porous electrode, a denotes the interfacial surface area of the particles per unit volume of electrode and Zmix is obtained from Eq. (8). From the impedance fitting processes achieved by us, this two-parallel diffusion path model fitted profoundly well with the experimental data of the impedances for the Si/LiTi2O4 electrode that measured at different potentials, revealing the correctness and applicability of this two-parallel diffusion path model. Thus, based on the simulation, when the semi-infinite diffusion conditions are applied, the variation of Warburg slope r with the open-circuit cell voltage (or the stage of charge) E is r ¼ ðdE=dX ÞV m =nFAð2DÞ

1=2

ð11Þ

where Vm is the molar volume of the Si/LiTi2O4 composite (15.47 cm3 mol1), the dE/dX is the gradient of open circuit potential vs. the composition x, n is the number of electrons transferred and F is the Faraday constant, and A is the active surface area. By observing the scanning electron microscopy (SEM) image, we selected two characteristic radius of the spherical particles, Rs,i, were 20 nm and 100 nm, respectively. Assuming that the radius of the particles located in the same layer are the same, the composite film with geometric surface area of 1 cm2 and thickness of 600 nm has two layers that are composed of the large particles and five layers that are composed of small particles. Then the total surface area of the active porous electrode composite is 21.98 cm2. The chemical diffusion coefficients at different potentials are calculated using Eq. (11) and presented in Table 1. From this table, the value of DLi is on the order of 1013 cm2 s1 and it decreases with a much decrease of the electrode potential. These calculated diffusion coefficients are quite approach with previous reported results [34–36], as all the diffusion coefficients of silicon based anodes calculated by them are estimated as 1013 cm2 s1, which are larger than that of Li in the crystalline Si (1014) [37]. Since the Li concentration increases with Table 1 The chemical diffusion coefficient DLi in porous Si/LiTi2O4 composite at different potentials Voltage (V)

DLi (cm2 s1)

2.52 2.10 1.68 0.35

4.2 E-13 7.31 E-13 4.03 E-12 5.45 E-13

Z.Y. Zeng et al. / Journal of Electroanalytical Chemistry 616 (2008) 7–13

decreasing electrode potential, therefore, it can be concluded that the lithium diffusion coefficient is affected by the Li concentration in the Si/LiTi2O4 composite, which is agreement with the previous result [38]. 4. Conclusions The Si/LiTi2O4 nanocomposite film was prepared by a surface sol–gel process and Si nanoparticles were dispersed homogeneously in the LiTi2O4 matrix. The porous structure of the Si/LiTi2O4 nanocomposite film could offer convenient channels and also provide a buffering interspace to alleviate the volume expansion during the intercalating and deintercalating of Li-ions with Si nanoparticles. The Si/ LiTi2O4 nanocomposite showed the stable cycling performance with a reversible capacity of about 1100 mA h g1 with average discharge capacity fading of 1.7 mA h g1 (0.15%) from the 2nd to 50th cycle, suggesting its promising nature in anode materials for lithium ion batteries. A two-parallel diffusion path model fitted well with the impedance data of Si/LiTi2O4 composite at different potentials. Impedance in the low-frequency region was directly related to the diffusion of Li ion, and the Li ion diffusivity was on the order of 1013 cm2 s1 based the semi-infinite diffusion model. References [1] R.A. Sharma, R.N. Seefurth, J. Electrochem. Soc. 123 (1976) 1763. [2] M. Holzapfel, H. Buqa, W. Scheifele, P. Novak, F.M. Petrat, Chem. Commun. 12 (2005) 1556. [3] S.Y. Chew, Z.P. Guo, J.Z. Wang, J. Chen, P. Munroe, S.H. Ng, L. Zhao, H.K. Liu, Electrochem. Commun. 9 (2007) 941–946. [4] T. Moritaz, N. Takami, J. Electrochem. Soc. 153 (2006) A425. [5] Y. Zhang, Z.W. Fu, Q.Z. Qin, Electrochem. Commun. 6 (2004) 484. [6] Y. Liu, K. Hanai, T. Matsumura, N. Imanishi, A. Hirano, Y. Takeda, Electrochem. Solid State Lett. 7 (2004) A492. [7] G.X. Wang, J.H. Ahn, J. Yao, Electrochem. Commun. 6 (2004) 689. [8] I. Kim, G.E. Blomgren, P.N. Kumta, J. Power Sources 130 (2004) 275. [9] P. Patel, I. Kim, P.N. Kumta, Mater. Sci. Eng. B 116 (2005) 347. [10] Z.P. Guo, Z.W. Zhao, H.K. Liu, S.X. Dou, J. Power Sources 146 (2005) 190. [11] K.C. Suryanarayana, Prog. Mater. Sci. 46 (2001) 1.

13

[12] T. Zhang, J. Gao, H.P. Zhang, L.C. Yang, Y.P. Wu, H.Q. Wu, Electrochem. Commun. 9 (2007) 886. [13] L.J. Fu, H. Liu, H.P. Zhang, C. Li, T. Zhang, Y.P. Wu, R. Holze, H.Q. Wu, Electrochem. Commun. 8 (2006) 1. [14] D.Q. Shi, J.P. Tu, Y.F. Yuan, H.M. Wu, Y. Li, X.B. Zhao, Electrochem. Commun. 8 (2006) 1610. [15] K.M. Colbow, J.R. Dahn, R.R. Haering, J. Power Sources 26 (1989) 397. [16] T. Ohzuku, A. Ueda, N. Yamamoto, J. Electrochem. Soc. 142 (1995) 1431. [17] Petr Krtil, Dina Fattakhova, J. Electrochem. Soc. 148 (2001) A1045. [18] M. Uematsu, M. Gunji, M. Tsuchiya, K.M. Itoh, Thin Solid Films 515 (2007) 6596. [19] A. Hemeryck, N. Richard, A. Este‘ve, M.D. Rouhani, Surf. Sci. 601 (2007) 2339. [20] X.L. Yang, Z.Y. Wen, X.X. Xu, B. Lin, S.H. Huang, J. Power Sources 164 (2007) 880. [21] T. Doi, Y. Iriyama, T. Abe, Z. Ogumi, Chem. Mater. 17 (2005) 1580. [22] L. Kavan, J. Prochazka, T.M. Spitler, M. Kalbac, M. Zukalova, T. Drezen, M. Gratzel, J. Electrochem. Soc. 150 (2003) A1000. [23] L. Kavan, M. Gratzel, J. Rathousky, A. Zukal, J. Electrochem. Soc. 143 (1996) 394. [24] C.M. Shen, X.G. Zhang, Y.K. Zhou, H.L. Li, Mater. Chem. Phys. 78 (2002) 437. [25] Y.M. Kang, M.S. Song, J.H. Kim, H.S. Kim, M.S. Park, J.Y. Lee, H.K. Liu, S.X. Dou, Electrochem. Acta 50 (2005) 3667. [26] Y.M. Kang, K.T. Kim, J.H. Kim, H.S. Kim, P.S. Lee, J.Y. Lee, H.K. Liu, S.X. Dou, J. Power Sources 133 (2004) 252. [27] P. Poizot, S. Laruelle, S. Grugeon, L. Dupont, J.M. Tarascon, Nature 407 (2000) 496. [28] W.J. Weydanz, M. Wohlfahrt-Mehrens, R.A. Huggins, J. Power Sources 81–82 (1999) 237. [29] J.P. Meyers, M. Doyle, R.M. Darling, J. Newman, J. Electrochem. Soc. 147 (2000) 2930. [30] M.D. Levi, C. Wang, D. Aurbach, J. Electroanal. Chem. 561 (2004) 1. [31] E. Barsoukov, D.H. Kim, H.S. Lee, H. Lee, M. Yakovleva, Y. Gao, J.F. Engel, Solid State Ionics 161 (2003) 19. [32] M.D. Levi, D. Aurbach, J. Phys. Chem. B 108 (2004) 11693. [33] Evgenij Barsoukov, J. Ross Macdonald, Impedance Spectroscopy Theory, Experiment, and Applications, Wiley-Interscience, Hoboken, NJ, 2005. [34] T.L. Kulova, A.M. Skundin, Yu.V. Pleskov, E.I. Terukov, O.I. Kon’kov, J. Electroanal. Chem. 600 (2007) 217. [35] Hi Xia, Songbai Tang, Li Lu, Mater. Res. Bull. 42 (2007) 1301. [36] K. Yoshimura, J. Suzuki, K. Sekine, T. Takamura, J. Power Sources 146 (2005) 445. [37] M. Green, E. Fielder, B. Scrosati, M. Wachtler, J.S. Moreno, Electrochem. Solid-State Lett. 6 (2003) A75. [38] Z.H. Yang, Y. Feng, Z.F. Li, S.B. Sang, Y.H. Zhou, L.H. Zeng, J. Electroanal. Chem. 580 (2005) 340.