Preparation, Characterization, and Electrochemical Performance of

Publication Date (Web): April 14, 2010. Copyright © 2010 American .... Vanadium-based nanostructure materials for secondary lithium battery applicati...
6 downloads 0 Views 3MB Size
Download PDF
J. Phys. Chem. C 2010, 114, 8099–8107

8099

Preparation, Characterization, and Electrochemical Performance of Lithium Trivanadate Rods by a Surfactant-Assisted Polymer Precursor Method for Lithium Batteries A. Sakunthala,†,‡ M. V. Reddy,*,† S. Selvasekarapandian,*,‡,§ B. V. R. Chowdari,*,† and P. Christopher Selvin| Department of Physics, National UniVersity of Singapore, Singapore 117542, DRDO-BU, Centre for Life Sciences, Bharathiar UniVersity, Coimbatore 641046, India, Kalasalingam UniVersity, Krishnankoil, Virudhunagar 626190, Tamil Nadu, India, and NGM College, Pollachi, Tamil Nadu, India ReceiVed: January 20, 2010; ReVised Manuscript ReceiVed: March 30, 2010

Lithium trivanadate (LiV3O8) compound was prepared under different conditions and characterized by the Rietveld-refined X-ray diffraction, scanning electron microscopy (SEM), and Brunauer-Emmett-Teller (BET) surface area and density techniques. The electrochemical performances of the LiV3O8 compounds prepared under different conditions were compared. LiV3O8 rods prepared by the surfactant-assisted polymer precursor method were found to perform well, delivering a discharge capacity of 230 ((5) mA · h/g at the end of the second cycle, with an excellent capacity retention of 99.52% at the end of the 20th cycle, for a current density of 30 mA/g. LiV3O8 rods delivered a discharge capacity of 135 ((5) mA · h/g at the end of 350 cycles, for a current density of 240 mA/g, and reasonably high capacity values were achieved at the different current rates. Impedance spectroscopic studies during the first and eighth cycles at various voltages are analyzed and discussed. 1. Introduction Nanostructured materials play an important role in advancing electrochemical energy storage and conversion technologies such as lithium ion batteries and fuel cells, offering great promise to address rapidly growing environmental concerns and the increasing global demand for energy.1 For example, literature studies show that V2O5 nanostrips obtained by the polyol method have an improved initial discharge capacity owing to the nanostructured morphology of the compound.2 Nanosized and rodlike LiMn2O4 has been found to achieve good electrochemical performance.3 LiFePO4 microstructures with self-assembled nanoplates synthesized by a solvothermal method deliver better cell performance than the corresponding commercial material.4 Li1+xV3O8 has been investigated as a cathode material for the past 20 years. Its crystal structure5 consists of a layered monoclinic structure belonging to the space group P21/m. The structure is made up of (V3O8)(1+x)- layers in the b-c planes, stacked one above the other along the a axis, which, in turn, consists of two basic structural units, VO6 octahedra and VO5 distorted trigonal bipyramids interconnected to each other by corner-sharing oxygen atoms to form V-O layers. Between the V-O layers, there are different octahedral and tetrahedral sites for the lithium ions. Unlike in other layered structures, where the layers are held together only by weak van der Waals forces, in this compound, each layer is held together by Li+ ions at the octahedral sites present in the interlayer. The immobile lithium ions play a pinning role between the layers and keep the layers strongly connected. The excess lithium corresponding to the * Corresponding authors. E-mail: [email protected] (M.V.R.), [email protected] (S.S.), [email protected] (B.V.R.C.). Tel.: 65-651662605 (M.V.R.). Fax: 65-67776126 (M.V.R.). † National University of Singapore. ‡ Bharathiar University. § Kalasalingam University. | NGM College.

amount x is accommodated at the tetrahedral sites between the layers and takes part in charge/discharge processes.6-9 Despite its structural advantages, the electrochemical performance of the above compound was found to be mainly influenced by the synthesis method.10 Synthesis conditions such as temperature and raw materials used and morphological characteristics of the final product such as size, shape, texture (agglomerations), and crystallinity were found to influence its electrochemical performance.11,12 The compound LiV3O8 has many advantages such as low cost, high energy density, and good safety characteristics.10,11 The moderate working voltage (2.5-3.5 vs Li)7 and high-temperature stability of this compound when compared to those of other cathode materials such as LiCoO2 (4.0 V vs Li),13 LiMn2O4 (4.0 V vs Li),1 and LiFePO4 (3.4 V vs Li)1 make it more suitable for use in lithium polymer batteries. Intensive research has been carried out on this cathode material as reported by many authors.10,14,15 The preparation of LiV3O8 by a sol-gel method, other oxides, and electrochemical properties were discussed in detail by Fu et al.16 The LiV3O8 compound has been prepared and studied by many different methods such as a sol-gel process,12,17-19 electrospinning combined with a sol-gel process,20,21 hydrothermal process,22-24 freeze-drying,25 a citric acid and peroxide sol-gel method,26 spray drying,11 a rheological phase reaction method,27 an ultrasonic method,28 a flame pyrolysis method,29 a low-heating solid-state method,30 a microwave sol-gel route,31 a microwave solid-state synthesis,14 a low-temperature reaction route,32 and an ethylenediaminetetraacetic acid (EDTA) sol-gel method.33 In the present work, LiV3O8 nanostructures were prepared by a simple preparation method within a short calcination time of 2 h. The structural and energy storage properties of the material were analyzed, and detailed electrochemical impedance spectroscopy studies were performed.

10.1021/jp1005692  2010 American Chemical Society Published on Web 04/14/2010

8100

J. Phys. Chem. C, Vol. 114, No. 17, 2010

2. Experimental Section 2.1. Preparation of Materials. The compound LiV3O8 was prepared under different preparation conditions (with sample codes LiV3O8-I, LiV3O8-II, and LiV3O8-III) as follows: (i) Surfactant-Assisted Polymer Precursor Method (LiV3O8-I). Stoichiometric amounts of lithium acetate (0.54 g, 98% purity, Aldrich) and ammonium metavanadate (1.76 g, 98% purity, Fluka) were mixed with the polymer poly(vinyl pyrrolidone) (PVP K40, 99% purity, Aldrich, molecular weight 40000) dissolved in 100 mL of distilled water, and the mixture was refluxed at 100 °C for 2 h. The total molar ratio metal ions to PVP was maintained at 1:1.5. The resultant solution was evaporated on a hot plate until the volume was reduced by onehalf, and then it was added to 15 mL of the surfactant Igepal (Aldrich) and stirred well. The surfactant-containing solution was transferred in a crucible without any prior drying on a hot plate and heated at a temperature of 550 °C in a Carbolite box furnace for 2 h at a heating rate of 3 °C/min. For comparison, the preparation was repeated, and the resultant solution stirred with Igepal was heat treated at 500 °C for 2 h. To understand the role of the polymer and surfactant, LiV3O8 was prepared using either polymer or surfactant as follows: (ii) LiV3O8-II. Stoichiometric amounts of lithium acetate (0.54 g) and ammonium metavanadate (1.76 g) were mixed with the polymer PVP K40 (Aldrich) dissolved in 100 mL of distilled water and refluxed at 100 °C for 2 h. The resultant solution after reflux was evaporated on a hot plate, and the dry residue was ground and further heat-treated at 550 °C at a heating rate of 3 °C/min for about 2 h in air to obtain the LiV3O8 product. No surfactant was used during this preparation. (iii) LiV3O8-III. The raw materials lithium acetate (0.54 g) and ammonium metavanadate (1.76 g) were dissolved in distilled water (100 mL) in the absence of the polymer and refluxed for 2 h at 100 °C. After reflux, the resultant solution was evaporated to one-half its original volume and then mixed with 15 mL of surfactant, stirred well, and heat-treated in a Carbolite box furnace at a temperature of 550 °C at a heating rate of 3 °C/ min for 2 h in air. 2.2. Characterization Techniques. Thermogravimetric analysis (TGA) and differential thermal analysis (DTA) studies were performed using SDT Q600 TGA/DTA instruments at a heating rate of 20 °C/min in air. The as-prepared samples were characterized by X-ray diffraction (XRD) using a Philips X’PERT MPD unit with Cu KR radiation. The unit-cell lattice parameters were obtained by Rietveld refinement of the powder XRD data using the commercially available software TOPAS, version 2.1. The morphology of the powders was examined by scanning electron microscopy (SEM) measurements using a JEOL JSM-6700F instrument. A Tristar 3000 apparatus (Micromeritics, Norcross, GA) and an AccuPyc 1330 pycnometer (Micromeritics, Norcross, GA) were used to study the Brunauer-Emmett-Teller (BET) surface area and density, respectively, of the compounds. 2.3. Electrode Fabrication and Characterization Techniques. Electrochemical studies were carried out with 2016type coin cells. The electrodes were made of 70:15:15 wt % of active material LiV3O8, conductive carbon black (Super P), and poly(vinylidene fluoride) (PVDF) Kynar 2801 as a binder, respectively. N-Methylpyrrolidone (NMP) was used as the solvent. Lithium foil was used as the counter and reference electrodes. Cyclic voltammetry studies at a scan rate of 0.058 mV/s and galvanostatic cycling studies at different current rates were made in the voltage range of 2.0-4.0 V vs Li, cycled at room temperature. Impedance spectroscopy were carried out with a Solartron Impedance/gain-phase analyzer (model SI 1255)

Sakunthala et al.

Figure 1. Rietveld-refined XRD patterns of (a) (i) LiV3O8-I, (ii) LiV3O8-II, and (iii) LiV3O8-I prepared at 500 °C, 2 h; and (b) LiV3O8-III. (hkl) values are indexed.

coupled to a potentiostat (SI 1268) at room temperature (25 °C). The frequency was varied from 0.18 MHz to 3 mHz with an alternating-current signal amplitude of 10 mV. Nyquist plots (Z′ vs -Z′′) were collected and analyzed using ZPlot and ZView software (version 2.2, Scribner Associates Inc., Southern Pines, NC). Detailed descriptions of coin cell fabrication and instrumentation are given in our previous reports.34,35 3. Results and Discussion 3.1. Structure and Morphology Analysis. The Rietveldrefined XRD patterns of LiV3O8-I and LiV3O8-II are shown in Figure 1a. Only the main (hkl) lines are indexed for clarity. The XRD patterns of LiV3O8 were refined on the basis of reported monoclinic structure. The lattice parameters were found to match with JCPDS card no. 72-1193. (i) LiV3O8-I (550 °C) was found to have a pure phase of layered monoclinic structure without any impurities. The lattice parameter values were found to be a ) 6.663 Å, b ) 3.596 Å, c ) 12.003 Å, and β ) 107.72°. (ii) LiV3O8-II was found to have a layered monoclinic structure with 98.69 wt % LiV3O8 as the main phase and an impurity of 1.31 wt % corresponding to the Li0.3V2O5 phase. The lattice parameter values of the main phase were found to be a ) 6.657 Å, b ) 3.596 Å, c ) 11.998 Å, and β ) 107.73°. (iii) The compound prepared at a temperature of 500 °C for 2 h using the surfactant-assisted polymer precursor method was found to have a high percentage of other active phases. It contained 67 wt % LiV3O8 as the main phase with 28 wt % Li0.3V2O5 and 5 wt % V2O5 as the impurity phases. The lattice

Lithium Trivanadate Rods for Lithium Batteries

J. Phys. Chem. C, Vol. 114, No. 17, 2010 8101

Figure 2. SEM images of (a) LiV3O8 prepared by the surfactant-assisted polymer precursor method at a temperature of 500 °C for 2 h; (b) LiV3O8-I, (c) LiV3O8-II, (d) LiV3O8-III. Scale bar: 1 µm.

parameters of the main phase were found to be a ) 6.633 Å, b ) 3.599 Å, c ) 11.978 Å, and β ) 107.76°. The Rietveld-refined XRD pattern of LiV3O8-III is shown in the Figure 1b. The compound was found to have a very highintensity (100) plane over the other planes. This is due to the preferred orientation of the particles along the (100) plane, leading to a longer diffusion path for the lithium ions.36 The compound was found to have 9.5 wt % Li0.3V2O5 phase as the impurity phase and 91.5 wt % LiV3O8 as the main phase. The lattice parameter values of the main phase were found to be a ) 6.648 Å, b ) 3.596 Å, c ) 11.997 Å, and β ) 107.77°. The XRD results showed 550 °C as the optimum temperature for the formation of the single-phase LiV3O8 compound, irrespective of the preparation conditions. Electrochemical studies were performed only for the compound prepared at 550 °C. A similar type of impurity peak corresponding to Li0.3V2O5 was observed in many reports17,21,28,31,37,38 and considered as an active phase.17 A few reports have shown the presence of other active phases to degrade electrochemical performance. Zhou et al.33 reported the cyclic performance to be degraded by the presence of Li0.3V2O5 and LiV2O5 impurities. A similar point was made by Wu et al.31 A few other reports have shown an improved performance of LiV3O8 in the presence of other active phases. Dubarry et al.15 studied LiV3O8/Li0.3V2O5/C nanocomposites by a carbothermal reduction method with rapid heating and found the capacity retention of LiV3O8 to be improved. Tran et al.11 prepared LiV3O8 through a spray-drying method, where they observed the Li0.3V2O5 phase by using carbon in the preparation process, and found better electrochemical performance at a high current rate when compared to the phase-pure compound. Figure 2a-d shows SEM images of the LiV3O8 products prepared under different conditions. Bundles of a needlelike morphology were observed for the LiV3O8 compound prepared

Figure 3. Schematic diagram of the formation of LiV3O8 rods by the surfactant-assisted polymer precursor method.

by the surfactant-assisted polymer precursor method at a temperature of 500 °C for 2 h (Figure 2a). With an increase in the calcination temperature to 550 °C (LiV3O8-I), the needlelike particles were found to grow in length into LiV3O8 rods with less agglomeration (Figure 2b). The observed nanostructures were due to the combined effect of the polymer PVP and the surfactant Igepal, as described in the schematic diagram shown in Figure 3. The raw materials lithium acetate and ammonium metavanadate become finely dispersed and are readily adsorbed on the polymer chains during reflux. The polymer PVP is decomposed at a higher temperature of around 500 °C and acts as a good capping agent in controlling the growth of the particles. It is well-known that the polymer PVP is capable of readily adsorbing metal ions and metal oxides.39 The addition of the surfactant to the polymer mixture aids in faster growth of the particles through the additional thermal energy created by its decomposition. These combined effects of Igepal and the polymer lead to bundles of nanosized needles. With increasing calcination temperature, the needles were found to grow into LiV3O8 rods.

8102

J. Phys. Chem. C, Vol. 114, No. 17, 2010

Sakunthala et al.

Figure 4. Cyclic voltammetry studies on (a) LiV3O8-I, (b) LiV3O8-II, and (c) LiV3O8-III. Conditions: third cycle; scan rate, 0.058 mV/s; voltage range, 2.0-4.0 V; Li metal used as the counter and reference electrodes; room temperature.

The combustion nature of the polymer around 500 °C is supported by thermal studies of the precursor or dry residue corresponding to LiV3O8-II, which was collected before calcination. The results are discussed in the Supporting Information (SI 1). No thermal studies on the precursor related to LiV3O8-I were performed. This is due to the experimental limitations, where the refluxed solution after the addition of surfactant was as such calcinated without any prior drying on the hot plate. It is well understood that the decomposition of the surfactant occurs at a relatively lower temperature compared to polymer PVP. The specific roles of the polymer PVP and the surfactant Igepal on morphology are clearly seen in SEM studies of compounds LiV3O8-II and LiV3O8-III (Figure 2c,d). The particles were found to be highly agglomerated with no proper crystal shape in the case of LiV3O8-II (prepared using polymer in the absence of surfactant) (Figure 2c). The compound LiV3O8-III (prepared using the surfactant Igepal in the absence of polymer) was found to form larger, agglomerated, micrometersized rods (Figure 2d). The rodlike particles obtained within a short reaction time showed faster growth of the particles in the surfactant medium. The larger particle size is due to uncontrolled growth in the absence of the polymer. Better electrochemical performance could be expected for LiV3O8-I (Figure 2b), because the size and aggregation of the grains were found to play a major role in the electrochemical properties. Literature studies showed the shape of the LiV3O8 compound to be a main factor influencing the cyclability.40 Experimental densities of LiV3O8-I, LiV3O8-II, and LiV3O8-III were found to be 3.5566, 3.5862, and 3.6828 g/cm3, respectively. The errors between experimental and calculated densities were within 5-8% for all of the compounds. The BET surface areas of LiV3O8-I, LiV3O8-II, and LiV3O8-III were found to be 0.85, 0.53, and 0.84 m2/g, respectively. 3.2. Electrochemical Studies. 3.2.1. Cyclic Voltammetry. Cyclic voltammetry results for the compounds LiV3O8-I, LiV3O8-II, and LiV3O8-III are shown in Figure 4a-c, cycled in the range of 2.0-4.0 V at a scan rate of 0.058 mV/s at room temperature. For clarity, only the third cycle is shown in each case. During the cathodic scan, splitting of many peaks was

Figure 5. Galvanostatic discharge/charge cycles for compounds (a) LiV3O8-I, (b) LiV3O8-II, and (c) LiV3O8-III. Conditions: current density, 120 mA/g; voltage range, 2.0-4.0 V; Li metal used as the counter and reference electrodes; room temperature.

observed because of the different lithium sites with energy differences for holding the lithium ions.28,41 The peaks around

Lithium Trivanadate Rods for Lithium Batteries

J. Phys. Chem. C, Vol. 114, No. 17, 2010 8103

Figure 6. Comparison of differential capacity vs voltage for compounds (a,b) LiV3O8-I, (c,d) LiV3O8-II, and (e) LiV3O8-III. Numbers indicate cycle numbers. Conditions: current density, 120 mA/g; voltage range, 2.0-4.0 V.

3.5, 2.8, 2.7, and 2.3-2.4 V are due to the single-phase reaction with lithium ions occupying empty tetrahedral sites,40,41 and the peak at ∼2.5 V is related to the filling of octahedral sites formed upon intercalation during which a two-phase transition from Li3V3O8 to Li4V3O8 takes place.18,41 The last step at ∼2.35 V is attributed to the slower kinetic insertion process where the single-phase transition corresponding to the Li4V3O8 phase takes place and all of the lithium ions are octahedrally coordinated.42,43 Both of the peaks at 2.5 and 2.3 V were considered as primary reasons for the origin of capacity fading during the cycling of LiV3O8.40 The ∼2.35 V peak is also correlated with the reactivity between the electrolyte and the Li4+xV3O8 phase, leading to passive film formation or material dissolution, which, in turn, leads to capacity fading.42 A cathodic peak around 2.6 V and an anodic/cathodic peak around 3.4 V were noticed only in the case of compounds LiV3O8-II and LiV3O8-III. These two peaks belong to the active phase Li0.3V2O5 as observed in the XRD analysis of these compounds. 3.2.2. GalWanostatic Cycling. The galvanostatic discharge/ charge curves of the compounds LiV3O8-I, LiV3O8-II, and LiV3O8-III at different cycle numbers for a current density of 120 mA/g are shown in Figure 5a-c. A flat plateau region around 2.8 V was observed for all of the compounds. LiV3O8-III was found to have an explicit two-plateau region around 2.5 and 2.6 V. The plateau at 2.6 V belongs to the other active phase Li0.3V2O5 observed in the XRD pattern at a

significant level. The multiple plateau regions are well reflected in the differential capacity plots (Figure 6a-e). The differential capacity plots were found analogous to the cyclic voltammetry results. (i) The cathodic peak at ∼2.5 V corresponding to the two-phase transition region was absent during the first discharge cycle for LiV3O8-I. This is due to the easy intercalation of lithium ions owing to its morphology. The appearance of the 2.5 V peak on further cycling indicates the changes in structure by means of electrochemical grinding. For compounds LiV3O8-II and LiV3O8-III, the two-phase transition corresponding to the 2.5 V peak was found to appear during the first discharge cycle. (ii) The cathodic peak at ∼2.6 V was observed from the second cycle onward for compounds LiV3O8-II and LiV3O8-III. This peak is due to the presence of another active phase corresponding to Li0.3V2O5, as discussed in relation to cyclic voltammetry. The influence of the Li0.3V2O5 phase was found to be suppressed with increasing cycle number, which is reflected by the disappearance of the 2.6 V peak with increasing number of cycles for both compounds LiV3O8-II and LiV3O8-III. The peak at ∼2.6 V was not found to appear for LiV3O8-I in any of the cycles, indicating the pure phase of the compound. With increasing cycle number, a very good reversible cathodic/anodic peak around 2.7 V with a high peak current was observed for LiV3O8-I, when compared to the other two compounds.

8104

J. Phys. Chem. C, Vol. 114, No. 17, 2010

Sakunthala et al.

Figure 7. Plots of capacity vs cycle number for (a) LiV3O8-I at current densities of (*) 30 and (]) 120 mA/g, (O) LiV3O8-II at 120 mA/g, and (3) LiV3O8-III at 120 mA/g and (b) (x) LiV3O8-I at a current density of 240 mA/g. Voltage range: 2.0-4.0 V.

Figure 7a shows plots of capacity vs cycle number for LiV3O8-I at current densities of 30 and 120 mA/g and for LiV3O8-II and LiV3O8-III at a current density of 120 mA/g in the voltage range 2.0-4.0 V vs Li. Figure 7b shows a similar plot for LiV3O8-I at a higher current density of 240 mA/g. For the current density of 120 mA/g, LiV3O8-I was found to have an initial discharge capacity of 164 mA · h/g, and during the second cycle, its capacity increased to 182 mA · h/g. The discharge capacity then steadily increased and reached a maximum value of 191 mA · h/g at the 25th cycle. LiV3O8-I maintained a discharge capacity value of 180 mA · h/g at the end of the 60th cycle, which corresponds to 1.94 mol of lithium per mole of LiV3O8. The capacity retention at the end of the 60th cycle from the maximum was found to be 94%. The overall capacity retention after 60 cycles calculated from the second cycle was 99%. For LiV3O8-II, the initial and second discharge capacity values were the same, 161 mA · h/g (for 120 mA/g). The discharge capacity slowly decreased until the 19th cycle (158 mA · h/g), corresponding to a 1.86% capacity loss, after which it started to increase and reached a maximum of 172 mA · h/g at the 33rd cycle. Then, the discharge capacity stabilized, reaching a value of 163 mA · h/g at the end of the 60th cycle. The capacity retention from the maximum was found to be 95%. The overall capacity retention after 60 cycles from the second cycle was 100%.

The preliminary electrochemical cycling studies of LiV3O8-III (Figures 5c and 7a), with the larger particle size, revealed a lower discharge capacity of 114 mA · h/g during the initial cycle and a stable discharge capacity of 130 mA · h/g at the end of the 15th cycle. The lower capacity value is due to the larger particles, which lead to longer diffusion paths or kinetic limitations of the material. At current densities of 30 and 240 mA/g, the initial discharge capacities of LiV3O8-I were found to be 206 and 120 mA · h/ g, respectively. At the second cycle, the capacity values were found to increase to 230 and 151 mA · h/g, respectively. At the lower current rate (30 mA/g), the discharge capacity was found to be stabilized from the second cycle onward, with a capacity retention of 99.5% from the second to the 16th cycle and 96% from the second to the 40th cycle. For the higher current density of 240 mA/g, the maximum discharge capacity of 160 mA · h/g was attained at the 33rd cycle, and the capacity retention at the end of the 60th cycle was 94%. The overall capacity retention after 60 cycles from the second cycle was 100%, and after 350 cycles, it still remained at 88%, calculated from the second cycle (Figure 7b). The electrochemical performance of LiV3O8-I in the present work was compared with literature reports, as summarized in Table 1. Although the capacity retention of the LiV3O8polypyrrole composite prepared by a low-temperature solution route was found to be good, the discharge capacity value was found to be very low for the applied current density (40 mA/ g).44 LiV3O8 prepared by the spray-drying technique was found to have a high discharge capacity value with capacity retention of 84% in the 2.0-3.3 V range, but it was reported that, when the same sample was cycled between 2.0-3.7 V, the capacity fading was >2% per cycle.11 For the LiV3O8 nanorods prepared by a hydrothermal method followed by solid-state reaction, which could not be treated as either an anode or a cathode material, the higher discharge capacity values were due to a lower cutoff discharge voltage of 1.5 V. Moreover, no welldefined plateau region was observed during charge/discharge except for the first cycle, as reflected in the cyclic voltammetery results.42 In the present work, good discharge capacity values were obtained at a lower current rate, and moderate values were obtained at a high current rate because of kinetic limitations. The cyclic stability of LiV3O8-I was found to be better in the voltage range of 2.0-4.0 V vs Li, with a good flat plateau region around 2.8 V, corresponding to the working voltage. For LiV3O8-I, at current densities of 30, 120, and 240 mA/g, the amounts of lithium inserted during the second discharge were 2.46, 1.96, and 1.63 mol of Li, respectively, with Coulombic efficiencies of around 94-96% (plots are shown in the Supporting Information, SI 2). In comparison, the electrochemical performance of LiV3O8-I in our present study showed the best

TABLE 1: Literature Reports on LiV3O8 Prepared by Different Methods and Its Electrochemical Properties synthesis methodref hydrothermal24 sol-gel45 sol-gel46 flame pyrolysis29 solution route44 spray drying11 hydrothermal42 present work: LiV3O8-I

composition LiV3O8 Li1.2V3O8 Li0.96Ag0.04V3O8 LiV3O8 LiV3O8-polypyrrole (PPy) composite Li1.1V3O8 LiV3O8 LiV3O8

temperature (°C)/time (h)

particle morphology

current density (mA/g)

capacity (mA · h/g) (cycle number)

capacity retention (%)

400/12 350/10 300/16 350/30 min 480/12

rods; diameter 200-500 nm, length 2-5 µm agglomerated nanoparticles, 50 nm amorphous PPy and LiV3O8 particles

120 60 150 100 40

212.8 (1)-152.1 (18) 281 (2)-200 (40) 328 (1)-252.7 (50) 271 (1)-180 (50) 184 (1)-183 (100)

72 71 77 66 99

320/1 450/10 550/2

spherical aggregates consisting of nanorods single-crystalline nanorods rods; diameter < 1 µm

116 100 30 120 240

260 (2)-220 (60) stabilized at 236 (100) 230 (2)-219 (40) 182 (2)-180 (60) 152 (2)-152 (60) 152 (2)-135 (350)

84 96 96 99 100 89

Lithium Trivanadate Rods for Lithium Batteries

Figure 8. Discharge capacity of the cell LiV3O8-I vs Li cycled at different current density values (i.e., 30, 120, 180, 240, 300, 360, 420, 540, 620, and 740 mA/g). The current rate increased after completion of four discharge/charge cycles. Each cycle was charged at 10 mA/g, after which the system was allowed to relax for 4 h and was then subjected to a discharge cycle at various current densities.

J. Phys. Chem. C, Vol. 114, No. 17, 2010 8105 results. Therefore, the LiV3O8 nanostructure prepared within a very short-duration (i.e., 2-h) calcination using surfactant and polymer in the preparation step was found to have good electrochemical performance. Figure 8 shows the observed discharge capacities at different current rates. The second discharge capacity for the current density of 30 mA/g was found to be 220 mA · h/g. Discharge capacity values of 201, 189, 176, 172, 164, 142, 113, 96, and 76 mA · h/g were observed for current densities of 120, 180, 240, 300, 360, 420, 540, 620, and 740 mA/g respectively. Slight differences in the capacity values when compared to long-termcycled electrodes were due to the slow current charging step and to the 4-h relaxation. 3.3. Electrochemical Impedance Spectroscopy (EIS) Studies. EIS is a well-established technique for studying the electrode kinetics of cathode and anode materials because of its inherently nondestructive nature and its ability to distinguish various phenomena at different time scales by frequency modulation

Figure 9. Three-dimensional Nyquist plots (Z′ vs -Z′′) for LiV3O8-I: (a) first discharge cycle (3.5-2.0 V), (b) eighth discharge cycle (2.0-4.0 V) (c) first charge cycle (2.4-4.0 V), and (d) eighth charge cycle (2.4-4.0 V). Symbols represent the experimental data, and continuous lines represent fitted data. (e) Equivalent circuit used for fitting the impedance spectra consisting of Ri and Ri//CPEi combinations (Rb and CPEb were included only for the spectra corresponding to the discharged state at 2.8 V, during the first cycle). In some spectra, Rsf+ct and CPEsf+dl elements were used as discussed in the text.

8106

J. Phys. Chem. C, Vol. 114, No. 17, 2010

Sakunthala et al.

TABLE 2: Impedance Values during First and Eighth Discharge-Charge Cyclesa,b first cycle discharge

eighth cycle charge

discharge

charge

voltage

Rsf+ct

CPEsf+dl

Rsf

Rct

CPEsf

CPEdl

Rsf

Rct

CPEsf

CPEdl

Rsf

Rct

CPEsf

CPEdl

4.0 3.8 3.6 3.4 3.2 3.0 2.8

193 197 199 207 67c 170 143 143 141

20 21 26 29 17d 26 38 35 42

11 14 13 13 14 10 5

54 40 44 48 51 55 60

232 121 225 257 253 462 105

67 36 50 54 59 70 88

12 11 10 10 5 4

69 95 97 83 109 139

699 708 555 567 119 12

69 91 88 80 41 50

12 5 2 1 2 1 4

60 180 169 187 197 203 167

459 49 11 4 5 2 131

65 44 48 52 50 56 39

5 5

159 152

34 16

47 56

5 3 -

197 177 -

24 1 -

75 103 -

2.6 2.4 2.2 2.0

-

116e 144e

-

-

48f 41f

-

131e 166e

100f 120f

a Resistance values in ohms, and capacitance values in farads. b Error values: Rsf, (3 Ω; Rct, (5 Ω; Rsf+ct, (5 Ω; CPEsf, (3 µF; CPEdl, (3 µF; CPEsf+dl, (5 µF. c Bulk resistance, Rb, in ohms (error: Rsf, (3 Ω). d Bulk capacitance, CPEb, in millifarads (error: CPEb, (3 mF). e Rsf+ct. f CPEsf+ct.

in conjunction with electrode kinetics.35,47-51 In the present work, electrochemical impedance spectra were recorded at room temperature on the cells with LiV3O8-I vs Li during the first and eighth discharge-charge cycles. The data were collected at selected voltages in the range of 2.0-4.0 V at a current density of 30 mA/g. The cells were kept under equilibrium for 3 h before measurements. Figure 9a-d shows three-dimensional Nyquist plots for the first and eighth discharge-charge cycles. The impedance spectra were fitted using the reported equivalent electrical circuit35 shown in Figure 9e. The impedance spectra corresponding to the first cycle in the charged state (4.0-2.8 V) and the spectra corresponding to the eighth cycle in both the charged and discharged states (4.0-2.4 V) were fitted using two equivalent circuit models corresponding to surface-film and charge-transfer resistance, that is, using Rsf, CPEsf and Rct, CPEdl, respectively. For all other impedance spectra, the surface-film and the charge-transfer resistances are indistinguishable, and therefore, they were fitted using one equivalent circuit model with Rsf+ct and CPEsf+dl elements. Rsf+ct is the combined impedance of surface film and charge transfer, and CPEsf+dl is the corresponding constant phase element. For the impedance spectra of the fresh cell, the high- to medium-frequency region corresponds to the surface film (Rsf), followed by a line inclined at an angle of 45° attributed to Warburg resistance. This indicates a surface-film resistance (Rsf) value of 128 Ω with a corresponding constant phase element [i.e., surface-film capacitance (CPEsf)] value of 23 µF. During the initial discharge cycle, the width of the semicircle varies, indicating an increase in impedance at 3.4 V compared to the open-circuit voltage (OCV), and remains more or less the same until 3.0 V. From 2.6 to 2.0 V, a decrease in the impedance value was observed which corresponds to the combination of surface-film and charge transfer resistance. The corresponding impedance values are given in Table 2. The impedance spectra at the discharge voltage of 2.8 V during the first cycle showed a small semicircle in the lowfrequency region corresponding to the bulk resistance (Rb) of the electrode, arising from the electronic conductivity of the active material and ionic conductivity in the pores of the composite electrode.35,52 In all cases, the straight line inclined at an angle of 45° corresponds to the finite-length Warburg resistance, related to solid-state diffusion. The line following the Warburg resistance at very low frequency corresponds to the intercalation capacitance (Cint). The resistance due to the

electrolyte and the cell components was less than 5 ((2) Ω, and the Cint value was ∼0.1 F. The CPE values corresponding to the surface film and the double layer were in the range of microfarads, and that of the bulk was in the millifarad range. The fitted impedance values obtained are shown in Table 2. It was observed that the impedance during discharge was much reduced for the eighth cycle when compared to that of the first cycle. However, the charge spectra show an increase in impedance, and the reason for this behavior is not well understood at present. In the discharged state, for voltages less than 2.8 V, a decreasing trend of impedance values was observed for the first cycle. In contrast, an increasing trend was observed for the eighth cycle, which is due to the two-phase reaction formed after the first cycle as discussed in relation to the cyclic voltammetery and galvanostatic studies. 4. Conclusions The surfactant-assisted polymer precursor method was found to be simple and suitable for the preparation of phase-pure LiV3O8 rods. The resulting materials were characterized by a variety of physical and electroanalytical techniques. Galvanostatic and cyclic voltammetry studies showed average discharge-charge voltages of 2.8 V vs Li. Galvanostatic cycling studies of LiV3O8-I showed a stable capacity of 135 mA · h/g at the end of the 350th cycle at a current rate of 240 mA/g with negligible capacity fading. Acknowledgment. The authors thank Prof. G. V. Subba Rao, Department of Physics, NUS, for helpful discussions. A.S. thanks the Defence Research and Development Organisation (DRDO), India, for the grant of a Senior Research Fellowship. M.V.R. and B.V.R.C. thank the National Research Foundation (NRF), Singapore. Supporting Information Available: Thermal analysis of the precursor corresponding to LiV3O8-II prepared by the polymer precursor method and plots of Coulombic efficiency with cycle number for LiV3O8. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Manthiram, A.; Vadivel Murugan, A.; Sarkar, A.; Muraliganth, T. Energy EnViron. Sci. 2008, 1, 621.

Lithium Trivanadate Rods for Lithium Batteries (2) Ragupathy, P.; Shivakumara, S.; Vasan, H. N.; Munichandraiah, N. J. Phys. Chem. C 2008, 112, 16700. (3) Shaju, K. M.; Bruce, P. G. Chem. Mater. 2008, 20, 5557. (4) Yang, H.; Wu, X. L.; Cao, M. H.; Guo, Y. G. J. Phys. Chem. C 2009, 113, 3345. (5) Wadsley, A. D. Acta Crystallogr. 1957, 10, 261. (6) kawakita, J.; Mori, H.; Miura, T.; Kishi, T. Solid State Ionics 2000, 131, 229. (7) Zhang, X.; Frech, R. Electrochim. Acta 1998, 43, 861. (8) kawakita, J.; Katayama, Y.; Miura, T.; Kishi, T. Solid State Ionics 1998, 107, 145. (9) Kawakita, J.; Miura, T.; Kishi, T. Solid State Ionics 1999, 120, 109. (10) Kannan, A. M.; Manthiram, A. J. Power Sources 2006, 159, 1405. (11) Tran, N.; Bramnik, K. G.; Hibst, H.; Prolss, J.; Mronga, N.; Holzapfel, M.; Scheifele, W.; Novak, P. J. Electrochem. Soc. 2008, 155, A384. (12) Jouanneau, S.; Verbaere, A.; Lascaud, S.; Guyomard, D. Solid State Ionics 2006, 177, 311. (13) Tan, K. S.; Reddy, M. V.; Subba Rao, G. V.; Chowdari, B. V. R. J. Power Sources 2005, 147, 241. (14) Yang, G.; Wang, G.; Hou, W. H. J. Phys. Chem. B 2005, 109, 11186. (15) Dubarry, M.; Gaubicher, J.; Moreau, P.; Guyomard, D. J. Electrochem. Soc. 2006, 153, A295. (16) Fu, L. J.; Liu, H.; Li, C.; Wu, Y. P.; Rahm, E.; Holze, R.; Wu, H. Q. Prog. Mater. Sci. 2005, 50, 881. (17) Deptula, A.; Dubarry, M.; Noret, A.; Gaubicher, J.; Olczak, T.; Lada, W.; Guyomard, D. Electrochem. Solid State Lett. 2006, 9, A16. (18) Jouanneau, S.; La Salle, A. L. G.; Verbaere, A.; Deschamps, M.; Lascaud, S.; Guyomard, D. J. Mater. Chem. 2003, 13, 921. (19) Cao, X. Y.; Xie, L. L.; Zhan, H.; Zhou, Y. H. Mater. Res. Bull. 2009, 44, 472. (20) Gu, Y. X.; Chen, D. R.; Jiao, X. L.; Liu, F. F. J. Mater. Chem. 2006, 16, 4361. (21) Lee, K. P.; Manesh, K. M.; Kim, K. S.; Gopalan, A. I. J. Nanosci. Nanotechnol. 2008, 8, 1. (22) Liu, H. W.; Yang, H. M.; Huang, T. Mater. Sci. Eng. B 2007, 143, 60. (23) Xu, H. Y.; Wang, H.; Song, Z. Q.; Wang, Y. W.; Yan, H.; Yoshimura, M. Electrochim. Acta 2004, 49, 349. (24) Xu, J.; Zhang, H. L.; Zhang, T.; Pan, Q. Y.; Gui, Y. H. J. Alloys Compd. 2009, 467, 327. (25) Brylev, O. A.; Shlyakhtin, O. A.; Egorov, A. V.; Tretyakov, Y. D. J. Power Sources 2007, 164, 868. (26) Feng, J.; Liu, X.; Zhang, X. M.; Jiang, J. Z.; Zhao, J.; Wang, M. J. Electrochem. Soc. 2009, 156, A768. (27) Liu, Q. Y.; Liu, H. W.; Zhou, X. W.; Cong, C. J.; Zhang, K. L. Solid State Ionics 2005, 176, 1549. (28) Liu, Y. M.; Zhou, X. C.; Guo, Y. L. J. Power Sources 2008, 184, 303.

J. Phys. Chem. C, Vol. 114, No. 17, 2010 8107 (29) Patey, T. J.; Ng, S. H.; Buchel, R.; Tran, N.; Krumeich, F.; Wang, J.; Liu, H. K.; Novak, P. Electrochem. Solid State Lett. 2008, 11, A46. (30) Yang, H.; Li, J.; Zhang, X. G.; Jin, Y. L. J. Mater. Process. Technol. 2008, 207, 265. (31) Wu, F.; Wang, L.; Wu, C.; Bai, Y.; Wang, F. Mater. Chem. Phys. 2009, 115, 707. (32) Liu, L.; Jiao, L. F.; Sun, J. L.; Zhang, Y. H.; Zhao, M.; Yuan, H. T.; Wang, Y. M. J. Alloys Compd. 2009, 471, 352. (33) Zhou, Y.; Yue, H. F.; Zhang, X. Y.; Deng, X. Y. Solid State Ionics 2008, 179, 1763. (34) Reddy, M. V.; Sundara Manoharan, S.; John, J.; Singh, B.; Subba Rao, G. V.; Chowdari, B. V. R. J. Electrochem. Soc. 2009, 156, A652. (35) Reddy, M. V.; Subba Rao, G. V.; Chowdari, B. V. R. J. Phys. Chem. C 2007, 111, 11712. (36) West, K.; Zachau-Christiansen, B.; Skaarup, S.; Saidi, Y.; Barker, J.; Olsen, I. I.; Pynenburg, R.; Koksbang, R. J. Electrochem. Soc. 1996, 143, 820. (37) Liu, X. H.; Wang, J. Q.; Zhang, J. Y.; Yang, S. R. J. Mater. Sci. 2007, 42, 867. (38) Liu, Y. M.; Zhou, X. C.; Guo, Y. L. Electrochim. Acta 2009, 54, 3184. (39) Christina, G.; Dembski, S.; Hofmann, A.; Ruhl, E. Langmuir 2006, 22, 5604. (40) Jouanneau, S.; La Salle, A. L. G.; Verbaere, A.; Guyomard, D. J. Electrochem. Soc. 2005, 152, A1660. (41) Bonino, F.; Panero, S.; Pasquali, M.; Pistoia, G. J. Power Sources 1995, 56, 193. (42) Liu, H. M.; Wang, Y. G.; Wang, K. X.; Wang, Y. R.; Zhou, H. S. J. Power Sources 2009, 192, 668. (43) Benedek, R.; Thackeray, M. M.; Yang, L. H. Phys. ReV. B 1999, 60, 6335. (44) Chew, S. Y.; Feng, C. Q.; Ng, S. H.; Wang, J. Z.; Guo, Z. P.; Liu, H. K. J. Electrochem. Soc. 2007, 154, A633. (45) Xie, J. G.; Li, J. X.; Zhan, H.; Zhou, Y. H. Mater. Lett. 2003, 57, 2682. (46) Sun, J. L.; Jiao, L. F.; Yuan, H. T.; Liu, L.; Wei, X.; Miao, Y. L.; Yang, L.; Wang, Y. M. J. Alloys Compd. 2009, 472, 363. (47) Shaju, K. M.; Subba Rao, G. V.; Chowdari, B. V. R. J. Electrochem. Soc. 2004, 151, A1324. (48) Reddy, M. V.; Yu, T.; Sow, C. H.; Shen, Z. X.; Lim, C. T.; Subba Rao, G. V.; Chowdari, B. V. R. AdV. Funct. Mater. 2007, 17, 2792. (49) Das, B.; Reddy, M. V.; Krishnamoorthi, C.; Tripathy, S.; Mahendiran, R.; Subba Rao, G. V.; Chowdari, B. V. R. Electrochim. Acta 2009, 54, 3360. (50) Levi, M. D.; Aurbach, D. J. Phys. Chem. B 2004, 108, 11693. (51) Nobili, F.; Tossici, R.; Marassi, R.; Croce, F.; Scrosati, B. J. Phys. Chem. B 2002, 106, 3909. (52) Reddy, M. V.; Madhavi, S.; Subba Rao, G. V.; Chowdari, B. V. R. J. Power Sources 2006, 162, 1312.

JP1005692