Synthesis and Electrochemical Properties of Nanostructured LiCoO2

Sep 7, 2005 - Possible Li ordering at x = 1/4 and 3/4 for this type of material is described in terms of a [2 x 2] superlattice in a triangular lattic...
17 downloads 17 Views 229KB Size
Download PDF
J. Phys. Chem. B 2005, 109, 17901-17906

17901

Synthesis and Electrochemical Properties of Nanostructured LiCoO2 Fibers as Cathode Materials for Lithium-Ion Batteries Yuanxiang Gu, Dairong Chen,* and Xiuling Jiao Department of Chemistry, Shandong UniVersity, Jinan 250100, The People’s Republic of China ReceiVed: April 27, 2005; In Final Form: August 4, 2005

Nanostructured LiCoO2 fibers were prepared by the sol-gel related electrospinning technique using metal acetate and citric acid as starting materials. The transformation from the xerogel fibers to the LiCoO2 fibers and the nanostructure of LiCoO2 fibers have been investigated in detail. The LiCoO2 fibers with 500 nm to 2 µm in diameter were composed of polycrystalline nanoparticles in sizes of 20-35 nm. Cyclic voltammetry and charge-discharge experiments were applied to characterize the electrochemical properties of the fibers as cathode materials for lithium-ion batteries. The cyclic voltammogram curves indicated faster diffusion and migration of Li+ cations in the nanostructured LiCoO2 fiber electrode. In the first charge-discharge process, the LiCoO2 fibers showed the initial charge and discharge capacities of 216 and 182 (mA‚h)/g, respectively. After the 20th cycle, the discharge capacity decreased to 123 (mA‚h)/g. The X-ray diffraction and highresolution transmission electron microscopy analyses indicated that the large loss of capacity of fiber electrode during the charge-discharge process might mainly result from the dissolution of cobalt and lithium cations escaping from LiCoO2 to form the crystalline Li2CO3 and CoF2 impurities.

1. Introduction Many investigations on microscale lithium-ion batteries have been developed for their various applications in fields related to microsystems such as microsensors, micromechanics, and microelectronics.1 To fabricate the microscale lithium-ion batteries with a high-energy density, fabrication techniques for microscale anodes, cathodes, and electrolytes are needed. In past research there are many reports on the electrospinning polymer fibers for polymer electrolyte and the carbon or metal oxide fibers for anode materials, which exhibit superior electrochemical properties compared with the powder materials.2,3 Among the lithium battery materials, LiCoO2 is the most attractive material due to the best performance in terms of high specific energy density and excellent cycle life, so that it has already been commercialized as a cathode material.4 Numerous investigations into the synthesis and electrochemical properties of LiCoO2 materials have been made for increasing its performance,5 and most of the works focused on the preparation of LiCoO2 particles with different sizes and shapes by use of the conventional solid-state reactions6 and solution-assisted methods such as sol-gel,7 hydrothermal,8 and emulsion drying,9 etc.10 Aside from this, the LiCoO2 films as the cathode materials have been synthesized for the construction of the microscale lithiumion batteries.11 Although the fiber anode materials reveal excellent physical and electrochemical properties for highperformance lithium-ion batteries,2,3 the synthesis and characterization of LiCoO2 fibers as cathode materials, which might exhibit superior electrochemical properties compared with the powders or film materials, have not been reported to the best of our knowledge. As to lithium-ion batteries, it is believed that the limitation in the rate capabilities is caused by the slow solidstate diffusion of Li+ cations within the electrode materials.12 * Corresponding author. Telephone: +86-531-88364280. Fax: +86531-88364281. E-mail: [email protected].

The nanostructured fibers might be used to achieve a fast solidstate diffusion due to the short diffusion distance of Li+ cations. In the present work, the preparation of nanostructured LiCoO2 fibers with 500 nm to 2 µm in diameter by the sol-gel assisted electrospinning technique is introduced. Furthermore, the electrochemical properties of the LiCoO2 fibers have been investigated. The nanostructured electrodes offer a higher initial discharge capacity of 182 (mA‚h)/g compared with ca. 140 (mA‚ h)/g of conventional powder and film electrodes, which might provide a new model system for the construction of the microscale lithium-ion batteries. The goal of this work is to develop the LiCoO2 fiber cathode for the fabrication of the microscale lithium-ion batteries and to increase the performance of cathode materials through adjusting the structures and morphologies, although much effort has been made on the substitution of Co3+ cations with other metal ions or surface modification to improve the performance of LiCoO2.13,14 2. Experimental Section Synthesis. All the reagents were analytical grade. In a typical experiment, 30.3 mmol (3.092 g) of lithium acetate (Li(CH3COO)‚H2O) and 30.0 mmol (7.472 g) of cobalt acetate (Co(CH3COO)2‚4H2O) were dissolved in 90.0 mL of distilled water to give a pink solution, and 36.0 mmol (7.565 g) of citric acid (C6H8O7‚H2O) was added into 30.0 mL of distilled water. The two solutions were mixed under stirring and then aged at 70 °C to turn into a viscous transparent sol. The rheological curve of the sol exhibited its Newton-fluid nature. The precursor sol was further aged at room temperature until its viscosity increased to 2.0 Pa‚s. The as-obtained spinnable sol was loaded into a Teflon bushing equipped with a spinneret made of stainless steel (its aperture, 500 µm), which was connected to a high-voltage supply (BGG-200 kV/20 mA). The voltage was 25 kV, the distance between the spinneret and stainless steel collector was

10.1021/jp0521813 CCC: $30.25 © 2005 American Chemical Society Published on Web 09/07/2005

17902 J. Phys. Chem. B, Vol. 109, No. 38, 2005 30 cm, and the N2 pressure of Teflon bushing was 0.3 MPa. The gel fibers were deposited on the collector, which were dried at 70 °C for 1 h to remove the volatile components. The obtained xerogel fibers were first heated at a rate of 0.5 °C‚min-1 from room temperature to 400 °C, kept at 400 °C for 2 h, and then heated at 700 °C for 8 h at the heating rate of 2.0 °C‚min-1 from 400 to 700 °C. The calcinations were carried out in air. As a comparison, the LiCoO2 powders were also prepared under the same conditions besides the electrospinning process. Characterization. The thermal decomposition behavior of the xerogel fibers was examined by thermogravimetric analysis (TGA) in the O2 flow from 40 to 600 °C with the heating rate of 10 °C‚min-1. An X-ray diffractometer (XRD) with a graphite monochromator and Cu KR radiation (λ ) 0.154 18 nm) was applied to identify the phase of fibers, while the voltage and electric current were held at 40 kV and 20 mA (2θ ) 10-80°). The morphology and microstructure of fibers were characterized by scanning electron microscopy (SEM, Hitachi S-520), transmission electron microscopy (TEM, Model H-800), and highresolution transmission electron microscope (HRTEM, GEOL2010) with an accelerating voltage of 200 kV. For the TEM or HRTEM measurement, the LiCoO2 fibers were sliced off by a diamond knife and attached on the copper or carbon grid for further analysis. The infrared (IR) spectra were recorded on a Nicolet 5DX-FTIR spectrometer using the KBr pellet method in the range of 400-4000 cm-1. Raman spectra were collected at room temperature using a Jobin-Yvon ISA U1000 Raman spectrometer in a quasi-backscattering configuration, the wavelength of Ar laser being 514.5 nm. The chemical composition of the inorganic fibers was analyzed on an inductively coupled plasma atomic emission spectroscopy (ICP-AES, PE Instruments ICP-OES Optima 2000 DV). The BET surface area was measured on an SSA-3500 micromeritics using nitrogen adsorption-desorption method. Electrochemical Measurements. Electrochemical properties of LiCoO2 fibers were studied by assembling a laboratory cell. A working electrode was prepared by loading fibers into a bag made up of Ni net connected with a 0.2 mm Ni wire as downlead and pressed by the pressure of 10 MPa. The weight of active materials was determined by weighing the Ni net before and after pressing the fibers. The powder materials as a comparison were treated in the same way as the fibers. The lithium foil was pressed onto the Ni net used as the reference and counter electrodes. The electrolyte was 1.0 mol‚dm-3 LiPF6 solution in ethylene carbonate (EC) and diethylene carbonate (DEC) with the volume ratio of 1:1, in which the content of water and HF was less than 20 and 22 ppm, respectively. All the manipulations were performed at room temperature in a glovebox filled with purified argon gas. Cycle voltammetry measurement was made on a CHI660 electrochemical workstation in a Teflon-made cell. The cell was electrochemically cycled 20 times between 3.0 and 4.3 V with a scanning rate of 0.1 mV‚s-1. Charge/discharge characteristics of the cells were recorded using the LAND cell-testing system at the current density of 20 mA/g between 2.6 and 4.3 V. 3. Results and Discussion Fabrication of LiCoO2 Fibers. Electrospinning has been extensively investigated for generating the ultrathin polymer and inorganic fibers, which involves the stretching of a sol by electrostatic force and the falling of gel fibers onto the collector.15 In this process, a sol held by its surface tension at the end of a capillary tube was subjected to an electric field and the charge was induced on the liquid surface by the applied

Gu et al.

Figure 1. IR spectra of the xerogel fibers (a) and the fibers calcined at 300 (b) and 400 °C (c) for 8 h in air.

voltage, causing a force directly opposite to the surface tension. As the intensity of the electric field was high enough, a charged sol jet was ejected from the spinneret to form the gel fibers falling on the collector. The diameter and the structure of the gel fibers were influenced by electrospinning parameters and characteristics of the spinnable sol. For example, the strength of the electrical field played an important role in determining the structure and the diameter of gel fibers. In general, the diameter of gel fibers decreased as the electric field was enhanced. The rheological property of the spinnable sol was also crucial to the diameter of the final fibers. The gel fibers became thinner with the viscosity decreasing as the electrospinning parameters were kept constant. In the present work, the effects of the electrospinning parameters and the rheological characteristics of the sol on the structure and the morphology of the gel fibers have not been further discussed. In the precursor solution, the citric acid, acetic acid, and water could coordinate to Co(II) to form the stable Co(II) species, which competed with each other. As heated at 70 °C, the solution would be gradually polymerized to form the sol with part of the acetic acid and the water slowly evaporating. In the process of the electrospinning and the following drying, the volatile components would rapidly vaporize, and the gel fibers would be transformed to the xerogel fibers. When the xerogel fibers were heated in air, they would decompose to form the inorganic fibers. The TGA curve (Supporting Information, Figure S1) shows that the weight loss from 40 to 200 °C, which is attributed to the departure of the acetic acid and the water, is ca. 8.42%. The weight loss of 60.81% from 200 to 330 °C is due to the decomposition of the organics. There would be no further weight loss when the temperature was higher than 330 °C, and the remains were ca. 30.77%. According to the above results, if the acetic acid and the water in the xerogel fibers did not coordinate to metal ions and could be removed completely, the LiCoO2 content would be 33.60% (30.77/(60.81 + 30.77)). Whereas based on the calculation according to dosage in the typical experiment, the LiCoO2 content in the xerogel fibers was 33.35%, considering only citrate coordinated to Co(II). These results were approximately equal within the analysis error, which indicated that the citrate but not the acetate coordinated to metal ions in the xerogel fibers. The citric acid could coordinate to Co(II) by different modes, including chelate, monodentate, and bridge-coordinate modes, which could be distinguished from the IR spectra.16 The IR spectrum of the xerogel fibers (Figure 1) shows the bands around 1592 cm-1 (νas(COO)) and 1412 cm-1 (νs(COO)), which are attributed to the asymmetric and symmetric vibrations of coordinated carboxyls, respectively.17 The absorbance at 1713 cm-1 corresponds to Vas(CdO) and the band at 1264 cm-1 is assigned to the formation of a metal hydroxo complex (δ(MOH) bending

Nanostructured LiCoO2 Fibers as Cathode Materials

J. Phys. Chem. B, Vol. 109, No. 38, 2005 17903

Figure 3. SEM (a) and HRTEM (b) images of LiCoO2 fibers. The insets of a and b are the high-magnified SEM and TEM images of LiCoO2 fibers, respectively. Figure 2. XRD patterns of the xerogel fibers calcined at different temperatures.

mode).18

On the basis of the frequencies of νas(COO) and νs(COO) and the separation ∆ν ) 180 cm-1, it is concluded that the citric acid coordinated to Co(II) by monodentate mode.19 After the xerogel fibers were heat-treated at 300 °C for 8 h, the vibrations at ca. 1713 and 1264 cm-1 obviously weakened, and those assigned to the coordinated carboxyls hardly changed, indicating that the uncoordinated carboxyls were decomposed at a relatively lower temperature. When the temperature was increased to 400 °C, the vibrations assigned to organics almost disappeared. The bands between 590 and 560 cm-1 were the characteristic vibrations of Co-O.20 The IR analyses were in agreement with the TGA results, indicating the carboxyls of the citrate but not the carboxyls of the acetic acid coordinated to Co(II) in the xerogel fibers. It also revealed that the Co(II) cations could bind the citric acid into a polynuclear complex as a chainlike or linear-type polymer, leading to the formation of a spinnable sol. The IR spectrum of the xerogel fibers after being heat-treated at 300 °C indicates that the organics have not been completely decomposed, which agrees well with the TGA results. When the heating temperature was increased to 400 °C, the absorbance of Co-O bond stretching increased and the corresponding absorbance of the organics disappeared, which indicated the nature of the inorganic compound. The XRD patterns of the fibers at different heating temperatures were applied to track the formation of LiCoO2 fibers. The XRD patterns shown in Figure 2 indicate that the xerogel fibers are amorphous and the crystalline LiCoO2 is formed at 300 °C. With the raise of the calcined temperature, the crystallinity of the fibers increased and the crystal structure was in good agreement with the hexagonal LiCoO2 with a layered structure (JCPDS file No.44-145). After the xerogel fibers were calcined at 400 °C for 8 h, the phase-pure LiCoO2 fibers were obtained, but the inorganic fibers had poor crystallinity. To obtain the perfect crystal structure of the flexile LiCoO2 fibers, it was necessary to heat the xerogel fibers at 700 °C for 8 h. The corresponding calculation from the XRD pattern gave the unit cell parameters of a ) 2.819 Å, c ) 14.059 Å, and a/c ) 0.2005. On the basis of the Scherrer equation, the particle size of LiCoO2 nanocrystals in the fibers was ca. 28 nm from the fwhm (full width at half-maximum) of (003) reflection. The Raman spectrum (Supporting Information, Figure S2) exhibits two bands at 485 and 595 cm-1, indicating that the LiCoO2 has a layered structure with the space group R3m and the cobalt is exactly trivalent.21 Table 1 shows the chemical composition of the LiCoO2 fibers after being calcined at different temperatures for 8 h. The molar ratio of lithium and cobalt in the fibers decreases with the increasing of the calcined temperature due to the evaporation of lithium, but is very close to 1.

TABLE 1: Chemical Compositions of the Fibers after Being Calcined at Different Temperatures element content (wt %) temp (°C)

Li

Co

C

molar ratio of Li to Co

600 650 700

7.02 7.06 6.91

59.27 60.11 59.44

0.45

1.005 0.998 0.995

The morphology of the fibers and their evolution with the increasing of the calcined temperature were observed from the SEM images. The xerogel fibers were continuous, which overlapped each other and looked like a piece of nonwoven cloth. The SEM images show that the xerogel fibers with smooth surface and circinal cross-section have 1-5 µm in diameter. After they were calcined, the xerogel fibers would asymmetrically shrink to form the elliptical LiCoO2 fibers, whose diameter ranged from 500 nm to 2 µm (Figure 3a). The high-magnified SEM image (inset in Figure 3a) indicates that the fibers have a dense and slippery outer shell. The TEM image (inset in Figure 3b) indicates that the LiCoO2 fibers are composed of uniform nanoparticles with the size of 20-35 nm. It can also be seen from the TEM image that the inner parts of the LiCoO2 fibers have a few of nanopores between the particles although the outer shell of the fibers is dense. The HRTEM image (Figure 3b) shows that the particles have a smooth surface and orderly basal planes. The crystallographic planes between adjacent particles match very well, and no amorphous impurities are observed from the HRTEM images. Electrochemical Studies of LiCoO2 Fibers. The cycle voltammogram curves of the fiber electrode (Figure 4a) indicate that the main lithium intercalation and deintercalation peaks appear below 4 V in the first cycle, which correspond to the first-order phase transition between two hexagonal phases. Above 4 V, there are two small redox peaks, which are caused by the order-disorder phase transition.22 Moreover, the voltammetric peak separation (DEpk) value of the main redox peaks is small, which indicates the high reversibility and fast Li+ cations diffusion. In the second cycle, the obvious decrease of the main redox peaks and the disappearance of the two small redox peaks are indicative of some degradation of crystal or interface structures of LiCoO2, which leads to an obvious capacity loss. It can be seen from the CV curves of the powder electrode there is only one redox peak above 4.0 V due to being overlapped or covered with the strong and broad redox peaks (Figure 4b). The powder electrode has a larger peak separation between the anodic and cathodic peaks compared with the fiber electrode, suggesting that the diffusion and migration of Li+ cations are slower in the LiCoO2 powder electrode than in the nanostructured LiCoO2 fiber electrode.

17904 J. Phys. Chem. B, Vol. 109, No. 38, 2005

Gu et al.

Figure 4. Cyclic voltammogram curves of the LiCoO2 fibers (a) and powders (b).

Figure 5. First five charge-discharge curves of LiCoO2 fibers (a) and the cycle life of the LiCoO2 fiber and powder electrodes (b) at the current density of 20 (mA‚h)/g.

Galvanostatic charging and discharging curves (Figure 5a) display the first five discharge-charge curves of the fiber electrode. The potential plateau around 3.9-4.0 V in the charge-discharge profiles is the typical property of the layered LiCoO2 phase, which corresponds well to the reversible twophase reaction in the cyclic voltammogram. In the first cycle, the electrode material shows 182 (mA‚h)/g discharge and 216 (mA‚h)/g charge capacity, respectively. The discharge and charge capacity reduces rapidly to 173 and 193 (mA‚h)/g, respectively, in the second cycle. During the subsequent cycles, the loss of the irreversible capacity will gradually reduce. After 20 cycles, the nanostructured electrode can keep its discharge capacity at 123 (mA‚h)/g and 32% of the initial capability is lost (Figure 5b). As a comparison, LiCoO2 powders were prepared under the same conditions including starting materials and treated process. Comparing with the fibers, the powder electrode shows a relatively lower initial discharge capacity of ca. 132 (mA‚h)/g, but the capacity still remains at 108 (mA‚ h)/g after 20 cycles and only 18% of the capacity is lost, which is in agreement with the reported results. XRD and HRTEM techniques were applied to the study of the evolution of the fibers during the charge-discharge process. Figure 3b shows the crystallographic planes between the adjacent particles match well to reduce the surface energy, so that the Li+-Li+ coulomb repulsion is lower at the surface of nanoparticles, enhancing the local Li capacity because there are no neighboring Li+ cations outside the paticles.23 The nanostructure of the LiCoO2 fibers can reduce the resistance of Li+ cation diffusion during the deintercalation of Li+ cations and lead to a fast Li+ cation diffusion.12b This special nanostructure

might result from the sol characteristics, in which the sol particles should be linear polymers and would easily be extended in a one-dimensional direction during electrospinning. It was just the nanostructure that enabled the fibers to have much larger charge and discharge capacities than the powder materials. On the other hand, the surface area measurements gave the BET surface areas of 18.04 and 3.14 m2/g for the prepared LiCoO2 fiber and powder, respectively. The large surface area and the short distance of Li+ diffusion in the nanostrctured fibers were also responsible for the high capacities.24 The large surface area resulted in the delay of concentration polarization, causing the hysteresis of a rise/drop in battery voltage before the maximum capacity was utilized.25 After five cycles, the XRD pattern given in Figure 6a shows that a few of crystalline CoF2 (JCPDS file No.24-329) came into being, whose content increased with the proceeding of charge-discharge. After 20 cycles, the crystalline Li2CO3 (JCPDS file No.22-1141) was found from the XRD pattern. After five cycles, the particle surface and the boundary between the nanoparticles became rough and obvious defects appeared on the particle surface while a few impurities were produced near the LiCoO2 particles (Figure 7a). With the proceeding of the electrochemical cycles, more and more impurities were formed. When the LiCoO2 fibers after 20 cycles were ultrasonicly treated, the sediment, which was the mixture of crystalline CoF2 and Li2CO3, would be desquamated from the fibers and aggregated (inset of Figure 7b). During the charge-discharge process, the Co(III) cations in LiCoO2 could attack and oxidize the carbon groups of the electrolyte molecules due to its acidic/nucleophilic properties and were reduced to Co(II) cations,26 which were combined with

Nanostructured LiCoO2 Fibers as Cathode Materials

J. Phys. Chem. B, Vol. 109, No. 38, 2005 17905

Figure 6. XRD patterns of LiCoO2 fibers (a) and powders (b) after the charge-discharge cycles.

Figure 7. HRTEM images of the LiCoO2 fibers after the 5th cycle (a) and the 20th cycle (b).

HF including free HF in the electrolyte and those coming from the decomposition of LiPF6 to form the crystalline CoF2. On the other hand, the Li+ cations left the lattice due to their large concentration gradient in and out of the LiCoO2 particles and were combined with the Co(III)-oxidized electrolyte molecules to form RCO2OLi, which further decomposed into Li2CO3. The detailed formation processes of CoF2 and Li2CO3 are listed as follows. -

+

2EC + 2e +2Li f (CH2OCO2Li)2V + CH2dCH2 v

27

2EC + 2e- +2Li+ f LiOCO2(CH2)4OCO2Li V DEC + e- + Li+ f CH3CH2OCO2Li + CH3CH2• 28 DEC + e- + Li+ f CH3CH2OCO• + CH3CH2OLi V ROLi + CO2 f ROCO2Li V 29 2ROCO2Li + H2O f Li2CO3 V + 2ROH + CO2 v 2Li+ + 2CO2 + 2e- f Li2CO3 V + CO v LiPF6 f LiF V + PF5 30 PF5 + H2O f 2HF + PF3O 2ROCO2 + 2LiCo3+O2 + 4HF f 2ROCO2Li + 2CoF2 V + O2 v + 2H2O 26 The lithium escape and dissolution of cobalt are more severe in nanostructured fibers because more Li+ cations deintercalate and exist in the electrolyte. It can be concluded that these reactions are the most important factor that influences the apparent capacity loss, and the anisotropic expansion and contraction during cycling can also result in the capacity

fading.31 On the other hand, a larger amount of lithium interaction can trigger a phase transition and structure change. Therefore the electrochemical Li-driven irreversible structural, morphological, and textural changes in the nanostructured LiCoO2 fibers will also lead to capacity fading.32 Compared experiments showed that the powder obtained at the same conditions had a relatively larger particle size. The random arrangement of the particles and the larger particle size resulted in a relatively smaller initial capacity compared with the fiber electrode. The XRD analyses indicate that there are no obvious peaks of any impurities in the XRD patterns of the powder electrode after the 5th cycle, and the peak intensity relevant to Li2CO3 and CoF2 respectively is obviously lower than that of the fiber electrode after the 20th cycle (Figure 6b). Therefore, it is concluded that the dissolution of Co(II) and Li+ cations escaping from the microelectrode may not be so severe as from the LiCoO2 nanostructured fibers. Thus a smaller capacity and a smaller capacity loss are obtained in the powder electrode. The high capacity and rapid capacity loss may be the character of the nanostructured electrode materials.33 4. Conclusions Nanostructured LiCoO2 fibers with diameters of 500 nm to 2 µm have been synthesized by the electrospinning combined sol-gel method. Investigations of the X-ray diffraction and the Raman spectra demonstrated that the LiCoO2 fibers had the layered structure with the space group R3m. Crystallographic planes between the adjacent particles in LiCoO2 fibers matched well to reduce the surface energy, which made for more Li+ cations insertion and faster solid-state diffusion, greatly increasing the performance of batteries. The LiCoO2 fibers exhibited a high initial charge and discharge capacity of 216 and 182 (mA‚h)/g, respectively. However, more Li+ cation deintercalation to the electrolyte led to the formation of Li2CO3 and the dissolution of LiCoO2 at the particle surface resulted in the formation of crystalline CoF2, which brought out a rapid capacity decay in nanostructure fiber materials. Acknowledgment. The authors acknowledge the financial support of the Doctoral Fund (Grant 040422065) of China. Supporting Information Available: Figures showing a Raman spectrum of LiCoO2 fibers and a TGA curve of xerogel fibers. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) (a) Balkanshi, M. Sol. Energy Mater. Sol. Cells 2000, 62, 21-35. (b) Kim, M. K.; Park, K. S.; Son, J. T.; Kim, J. G.; Chung, H. T.; Kim, H.

17906 J. Phys. Chem. B, Vol. 109, No. 38, 2005 G. Solid State Ionics 2002, 152-153, 267-272. (c) Rho, Y. H.; Kanamura, K.; Umegaki, T. J. Electrochem. Soc. 2003, 150 (1), A107-111. (2) (a) Tatsumi, K.; Kawamura, T.; Higuchi, S.; Hosotubo, T.; Nakajima, H.; Sawada, Y. J. Power Sources 1997, 68, 263-266. (b) Yoon, S. H.; Park, C. W.; Yang, H.; Korai, Y.; Mochida, I.; Baker, R. T. K.; Rodriguez, N. M. Carbon 2004, 42, 21-32. (c) Kim, C.; Yang, K. S.; Lee, W. J. Electrochem. Solid-State Lett. 2004, 7 (11), A397-399. (3) Choi, S. W.; Jo, S. M.; Lee, W. S.; Kim, Y. R. AdV. Mater. 2003, 15 (23), 2027-2031. (4) (a) Sun, Y. K. J. Mater. Sci. 1996, 3617-3621. (b) Tarascon, J. M.; Armand, M. Nature 2001, 414, 359-367. (5) (a) Li, L.; Meyer, W. H.; Wegner, G.; Wohlfahrt-Mehrens, M. AdV. Mater. 2005, 17 (8), 984-988. (b) Chen, Z.; Dahn, J. R. Electrochem. SolidState Lett. 2004, 7 (1), A11-A14. (c) Liu, J.; Wen, Z.; Gu, Z.; Wu, M.; Lin, Z. J. Electrochem. Soc. 2002, 149 (11), A1405-A1408. (6) (a) Reimers, J. N.; Dahn, J. R. J. Electrochem. Soc. 1992, 139, 2091-2097. (b) Ohzuku, T.; Ueda, A. J. Electrochem. Soc. 1994, 141 (11), 2972-2977. (7) (a) Sun, Y. K.; Oh, I. H.; Hong, S. A. J. Mater. Sci. 1996, 31, 3617-3621. (b) Yoon, W. S.; Kim, K. B. J. Power Sources 1999, 81-82, 517-523. (8) Tabuchi, M.; Ado, K.; Kobayashi, H.; Sakaebe, H.; Kageyama, H.; Masquelier, C.; Yonemura, M.; Hirano, A. J. Mater. Chem. 1999, 9, 199204. (9) Han, C. H.; Yeh, P. Y. J. Mater. Chem. 2000, 10, 599-601. (10) Ying, J.; Jiang, C.; Wan, C. J. Power Sources 2004, 129, 264269. (11) (a) Song, S. W.; Fujita, H.; Yoshimura, M. AdV. Mater. 2002, 14 (4), 268-271. (b) Park, Y. J.; Kim, J. G.; Kim, M. K.; Chung, H. T.; Kim, H. G. Solid State Ionics 2000, 130, 203-214. (c) Mantoux, A.; Groult, H.; Balnois, E.; Doppelt, P.; Gueroudji, L. J. Electrochem. Soc. 2004, 151 (3), A368-A373. (d) Rho, Y. H.; Kanamura, K. J. Electrochem. Soc. 2004, 151 (1), A106-A110. (12) (a) Li, N.; Patrissi, C. J.; Che, G.; Martin, C. R. J. Electrochem. Soc. 2000, 147 (6), 2044-2049. (b) Bueno, P. R.; Leite, E. R. J. Phys. Chem. B 2003, 107, 8868-8877. (13) (a) Antolini, E.; Giorgi, L.; Carewska, M. J. Mater. Sci. Lett. 1999, 18, 325-327. (b) Zou, M.; Yoshio, M.; Gopukumar, S.; Yamaki, J. Electrochem. Solid-State Lett. 2004, 7 (7), A176-A179. (14) (a) Wang, Z.; Wu, C.; Liu, L.; Wu, F.; Chen, L.; Huang, X. J. Electrochem. Soc. 2002, 149 (4), A466-A471. (b) Kim, Y. J.; Kim, T. J.; Shin, J. W.; Park, B.; Cho, J. J. Electrochem. Soc. 2002, 149 (10), A1337A1341. (15) (a) Li, D.; Xia, Y. Nano Lett. 2003, 3 (8), 1167-1171. (b) Madhugiri, S.; Sun, B.; Smirniotis, P. G.; Ferraris, J. P.; Balkus, K. J., Jr. Microporsous Mesoporous Mater. 2004, 69, 77-83. (16) (a) Djordjevic, C.; Lee, M.; Sinn, E. Inorg. Chem. 1989, 28, 719723. (b) Deacon, G. B.; Phillips, R. J. Coord. Chem. ReV. 1980, 33, 227251.

Gu et al. (17) Zhecheva, E.; Toyanova, R.; Gorova, M.; Alcantara, R.; Motales, J.; Tirado, J. L. Chem. Mater. 1996, 8, 1429-1440. (18) Gajbhiye, N. S.; Bhattacharya, U.; Darshane, V. S. Thermochim. Acta 1995, 264, 219-230. (19) (a) Battaglia, L. P.; Bonamartini C. A.; Marcotrigiano, G.; Menabue, L.; Pellacani, G. C. J. Am. Chem. Soc. 1980, 102, 2, 2663-2669. (b) Alcok, N. W.; Tracy, V. M.; Waddington, T. C. J. Chem. Soc., Dalton Trans. 1976, 2243-2249. (20) Julien, C.; El-Farh, L.; Rangan, S.; Massot, M. J. Sol-Gel Sci. Technol. 1999, 15, 63-72. (21) Huang, W.; Frech, R. Solid State Ionics 1996, 86-88, 395-400. (22) Dokko, K.; Horikoshi, S.; Itoh, T.; Nishizawa, M.; Mohamedi, M.; Uchida, I. J. Power Source 2000, 90, 109-115. (23) (a) Ye, S. H.; Gao, X. P.; Wu, F.; Song, D. Y. Electrochem. Acta 2004, 49 (9-10), 1623-1628. (b) Wagemaker, M.; Kearley, G. J.; Van Well, A. A.; Mutka, H.; Mulder, F. M. J. Am. Chem. Soc. 2003, 125, 840848. (24) (a) Takayuki, D.; Yasutoshi, I.; Takeshi, A.; Zempachi, O. Chem. Mater. 2005, 17, 1580-1582. (b) Li, N.; Martin, C. R. J. Electrochem. Soc. 2001, 148 (2), A164-A170. (c) Zhou, Y.; Li, H. J. Mater. Chem. 2002, 12, 681-686. (25) (a) Patrissi, C. J.; Martin, C. R. J. Electrochem. Soc. 1999, 146, 3176-3180. (b) Zhou, Y.; Cao, L.; Zhang, F.; He, B.; Li, H. J. Electrochem. Soc. 2003, 150 (9), A1246-A1249. (c) Yan, H.; Sokolov, S.; Lytle, T. C.; Stein, A.; Zhang, F.; Smyrl, W. H. J. Electrochem. Soc. 2003, 150 (8), A1102-A1107. (26) Wang, Z.; Huang, X.; Chen, L. J. Electrochem. Soc. 2004, 151 (10), A1641-A1652. (27) Aurbach, D.; Markovsky, B.; Shechter, A.; Ein-Eli, Y.; Cohen, H. J. Electrochem. Soc. 1996, 143 (12), 3809-3820. (28) Ein-Eli, Y.; Thomas, S. R.; Koch, V.; Aurbach, D.; Markovsky, B. and Schechter, A. J. Electrochem. Soc. 1996, 143 (12), L273-L277. (29) (a) Levi, E.; Schechter, A.; Moshkovich, M.; Cohen, Y. J. Power Source 1999, 81-82, 95-111. (b) Aurbach, D.; Ein-Eli, Y.; Chusid, O.; Cameli, Y.; Babai, M.; Yamin, H. J. Electerochem. Soc. 1994, 141 (3), 603-611. (30) Heider, U.; Oesten, R.; Jungnitz, M. J. Power Source 1999, 8182, 119-122. (31) (a) Cho, J.; Kim, Y. J.; Park, B. Chem. Mater. 2000, 12, 37883791. (b) Cho, J.; Kim, Y. J.; Kim, T.-J.; Park, B. Angew. Chem., Int. Ed. Engl. 2001, 40, 3367-3369. (32) (a) Liu, L.; Wang, Z.; Li, H.; Chen, L.; Huang, X. Solid State Ionics 2002, 152-153, 341-346. (b) Yoon, W. S.; Lee, K. K.; Kim, K. B. J. Electrochem. Soc. 2002, 149 (2), A146-A151. (33) (a) Gao, X. P.; Bao, J. L.; Pan, G. L.; Zhu, H. Y.; Huang, P. X.; Wu, F.; Song, D. Y. J. Phys. Chem. B 2004, 108, 5547-5551; (b) Wang, Y.; Takahashi, K.; Shang, H.; Cao, G. J. Phys. Chem. B 2005, 109, 30853088. (c) Wang, Y., Lee, J. Y. J. Phys. Chem. B. 2004, 108, 17832-17837.