Nanostructured Nb2O5 Polymorphs by Electrospinning for

Dec 16, 2009 - Authors thank Prof. G.V. Subba Rao, Physics dept. NUS for his helpful discussions. This project was partially funded by the Clean Energ...
0 downloads 6 Views 4MB Size
Download PDF
664

J. Phys. Chem. C 2010, 114, 664–671

Nanostructured Nb2O5 Polymorphs by Electrospinning for Rechargeable Lithium Batteries A. Le Viet,†,‡ M. V. Reddy,† R. Jose,‡ B. V. R. Chowdari,* and S. Ramakrishna*,‡ Department of Physics, and Nanoscience and Nanotechnology InitiatiVe, National UniVersity of Singapore, Singapore 117542 ReceiVed: September 14, 2009; ReVised Manuscript ReceiVed: December 1, 2009

Polymorphs of 1D nanostructures of niobium pentoxide (Nb2O5) are synthesized by electrospinning. Pseudohexagonal (H-Nb2O5), orthorhombic (O-Nb2O5), and monoclinic (M-Nb2O5) structures of Nb2O5 are developed in this study by appropriate heat treatment. Morphological, structural, and electrochemical properties of these nanofibrous polymorphs are studied in detail. The H- and O- phases maintain the usual fibrous morphology, whereas the M- phase adopted a distorted nugget structure. These phases are evaluated for their application as cathode for lithium batteries. The M-Nb2O5 exhibits the highest capacity and better capacity retention compared to the other phases. The M-Nb2O5 delivers a specific capacity of 242((3) and 218((3) mAhg-1, cycled at a current of 50 mAg-1 in the voltage range, 1.0-2.6 V versus Li/Li+ at the end of second and 25th cycle, respectively. The electrospun M-Nb2O5 nuggets-based battery performs better than its particle/nanofiber counterpart and could be a cathode material of choice for 2 V due to the commercial viability of the electrospinning process and characteristics of the batteries developed herewith. 1. Introduction Electrospinning as a method for producing nanostructures of advanced materials such as polymers, metal oxides, metals, and so forth is currently gaining immense research interest1-4 not only due to viability in synthesizing 1D nanostructures in a mass scale cost effectively but also due to their interesting physical properties for wide range of applications in regenerative medicine,5-7 photovoltaics,8-12 and filtration.13-15 Electrospinning works under the principle of asymmetric bending of a charged liquid jet while passing through a uniform electric field and is widely used for preparation of continuous polymeric nanofibers.16,17 The synthetic procedure starts by preparing a polymeric solution of optimum viscosity necessary for electrospinning. If the polymeric solution contains a metal ion, then appropriate post annealing of the composite fibers result in desired inorganic 1D nanostructure. A large number of 1D nanostructures of metal oxides have been synthesized by electrospinning, a brief account of which is available in recent reviews.3,18 Niobium Pentoxide (Nb2O5) is an n-type transition metal oxide semiconductor with an oxygen stoichiometry dependent bandgap ranging between 3.2 to 4 eV. Stoichiometric Nb2O5 is an insulator (conductivity, σ ) ∼3 × 10-6 Scm-1) and becomes semiconducting (σ ) ∼3 × 10-3 Scm-1) with decrease in oxygen stoichiometry (Nb2O4.8).19 The Nb2O5 exists in many polymorphic forms; H-Nb2O5 (pseudohexagonal), O-Nb2O5 (orthorhombic), T-Nb2O5 (tetragonal), and M-Nb2O5 (monoclinic) are the most common crystallographic phases.20 Among these phases, the M-phase is thermodynamically more stable, while the H-phase is the least stable one and can be readily transformed into the M-phase by appropriate heat treatment. Owing to its attractive physical properties Nb2O5 has been considered for gas sensors,21 catalysis,22 electrochromics,23 photoelectrodes in dye-sensitized solar cells.24-27 A number of * To whom correspondence should be addressed. E-mail:phychowd@ nus.edu.sg (B.V.R.C.); [email protected] (S.R.). † Department of Physics. ‡ Nanoscience and Nanotechnology Initiative.

metal oxide nanostructures as an electrode for lithium batteries have been reviewed recently.28 Previously, the Nb2O5 material as a cathode (∼2.0 V) for lithium battery has been studied by different groups.29,30 Kodama et al.29 studied in situ XRD and X-ray absorption fine structure (XAS) during lithium intercalation/deintercalation in Nb2O5 submicrometer-sized particles. They found that T-Nb2O5 and M-Nb2O5 exhibit comparable specific capacity; however, the T-Nb2O5 showed a lower capacity fading during cycling. Recently, Wei et al.31 identified the importance of 1D nanostructures of Nb2O5 for lithium batteries. They showed that the lithium batteries fabricated using O-Nb2O5 nanobelts have high reversible charge/discharge capacity, high rate capability, and excellent cycling stability. The purpose of the present article is to compare the electrochemical properties and lithium cycling of various 1D Nb2O5 polymorphs obtained through a scalable nanofabrication process, that is, electrospinning. We note that there are several reports on the synthesis of Nb2O5 nanofibers by electrospinning.32,33 To the best of our knowledge, no further effort has been made to utilize the electrospun Nb2O5 nanofibers for any device applications. 2. Experimental Section 2.1. Synthesis of Nb2O5 Polymorphs by Electrospinning. The Nb2O5 nanofibers were prepared by combining the sol-gel chemistry and electrospinning technique using the reported procedure with modifications.10,33 The solution for electrospinning was prepared from polyvinylpyrrolidone (PVP; Mw ) 1 300 000, Sigma-Aldrich), ethanol, niobium ethoxide (99.95% trace metals basis, Sigma Aldrich), and acetic acid. Niobium ethoxide (0.5 g) and PVP (0.3 g) were added to a solution of ethanol (3.5 mL) and acetic acid (1 mL). The solution was stirred in an airtight bottle for 24 h to obtain a clear solution. The electrospinning of Nb2O5 was performed in a commercial electrospinning instrument (NANON, MECC, Japan) with an electric field of 3 × 105 Vm-1 and at a flow rate of 2 mlh-1. The composite fibers containing Nb ions in PVP were collected either on a rotating drum or on flat surface wrapped with an

10.1021/jp9088589  2010 American Chemical Society Published on Web 12/16/2009

Nanostructured Nb2O5 Polymorphs aluminum foil. Decomposition of the composite fiber and crystallization behavior were studied using a simultaneous differential thermal and thermogravimetric analyzer (Simultaneous DTA-TGA, SDT-2960, TA Instruments), with a heating rate of 10 °Cmn-1, from room temperature to 1000 °C. The composite fibers were annealed at 500, 800, 1000, and 1100 °C (1 h and 11 h) for 1 h in air in a carbolyte box furnace. The annealed fibers obtained were used for further characterization and fabrication of lithium batteries. 2.2. Characterization of Nb2O5 Polymorphs. Crystal structures of the annealed Nb2O5 fibers were studied by X-ray and electron diffraction techniques. The X-ray diffraction (XRD) patterns were recorded by X-ray diffractometer (Philips, X’PERT MPD, Cu KR radiation). Lattice parameters and average particle size were calculated using TOPAS software by fitting the observed XRD patterns to the respective crystal structure. Morphologies of the Nb2O5 nanofibers were examined by field emission scanning electron microscope (FE-SEM, JEOL JSM5600LV). The Brunauer, Emmett, and Teller (BET) surface area of the fibers was measured by a surface area analyzer (Micromeritics Tristar 3000). The fiber densities were determined by using a pycnometer (AccuPyc 1330, Micromeritics). X-ray photoelectron spectra (XPS) of the Nb2O5 nanofibers were obtained using a VG Scientific ESCA MK II spectrometer with monochromatic Mg KR radiation (1253.6 eV). The survey spectra were recorded in the range 0-1099 eV with constantpass energy of 50 eV; the high-resolution spectra were recorded with a smaller constant pass energy of 20 eV. Charge referencing was carried out against adventitious carbon C, assuming its binding energy at 284.6 eV. Analysis of the XPS spectra was done using XPS Peak-fit software. A Shirley-type background was subtracted from the recorded spectra and curve fitting was carried out with a Gaussian-Lorentzian (ratio 60:40) curve. The derived binding energies (BE) have accuracy better than ( 0.1 eV. 2.3. Electrode Fabrication and Characterization Techniques. Electrodes for the electrochemical studies were prepared by mixing Nb2O5 nanofibers with carbon black and polyvinylidene fluoride copolymer (binder, Kynar 2801) with a weight ratio of 65% Nb2O5 (500, 800 °C):20%:15% and 70% M-Nb2O5(1000, 1100 °C):15%:15%. N-methyl-2pyrrolidinone was used as the solvent to disperse the nanofibers, carbon black, and the binder. The doctor blade technique was used to deposit a ∼20 µm thick layer of the prepared slurry on an etched copper foil. Circular electrode (area ∼2 cm2) was cut from the coated copper foil. Some electrodes were further heated at 220 °C in Argon (Ar) gas for 6 h to improve the contact between Cu-current collector, conducting carbon, active material, and PVDF binder. Electrode heat treatment plays an important role on lithium battery performances.34 Glass microfiber filter (Whatman) membrane was used as separator. The counter and reference electrode was a 2 cm2 circular piece of lithium metal. Cointype test cells (CR2016) were fabricated in an Ar- filled glovebox. Cyclic voltammetry and discharge-charge cycling were carried out using Mac-pile II (Biologic, France), bitrode battery tester (Model SCN, USA), respectively. More details of cell fabrication and instrumentation are described elsewhere.35,36 3. Results and Discussion 3.1. Characterization. The thermal analysis of the as-spun composite polymeric fibers showed usual decomposition and formation of inorganic phase that are typical to these types of structures (Figure 1). The polymeric fiber containing the niobium

J. Phys. Chem. C, Vol. 114, No. 1, 2010 665

Figure 1. Differential thermal analysis (DTA)/Thermogravimetric analysis (TGA) of Nb2O5 nanofibers, from room temperature to 1000 °C, with a heating rate of 10 °Cmn-1.

ions showed an endothermic peak in the differential thermal analysis (DTA) curve and a weight loss (∼10%) in the thermogravimetric analysis (TGA) curve at ∼47 °C. Although the bending instabilities during the electrospinning increase the jet path length enormously16 through which passage the solvent evaporates and the fibers solidify, a small amount of solvent is expected to adsorb the surface because of the small fiber diameter. The observed weight loss and endothermic events result from the liberation of surface adsorbed ethanol used during electrospinning. The major exothermic peak at ∼351 °C leading to a 31% weight loss was observed due to the decomposition of the polymer. Crystallization of H-Nb2O5 phase and/or complete decomposition of polymer was observed at above ∼453 °C. No clear crystallization peaks were observed for the orthorhombic and monoclinic phases within the limits of the present experiment. The as-spun polymeric samples were annealed at 500, 800, 1000, and 1100 °C to synthesize various polymorphs of Nb2O5. Figure 2 shows SEM images displaying the morphologies of the as-spun polymeric and the annealed metal oxide fibers. The average diameter of the as-spun polymeric fibers was ∼300 nm, which upon annealing up to 800 °C reduced to 160 nm. Formation of Nb2O5 fibers from the composite fibers should involve at least three processes: evaporation of the polymer (PVP), nucleation and growth of Nb2O5 nanocrystals, and directional mass transport of Nb2O5 nanocrystals to form continuous nanofibers. The lowering of fiber diameter on annealing could be partially due to the removal of the polymer (PVP) and partially due to mass transport. The samples heated up to 800 °C maintained the conventional electrospun fiber morphology. However, the fibers heated at temperature g1000 °C showed distortion, adopting a nuggetlike morphology (part d of Figure 2) with a higher size (∼500 nm). The observed collapse of the fiber morphology and adoption of the nugget structure could be due to partial melting of Nb2O5 at temperature above 1000 °C. Parts a, d, and g of Figure 3 show the bright field TEM image of H-, O-, and M-Nb2O5 phases, respectively. As observed in the SEM pictures, the M-Nb2O5 phase had the shape of distorted nuggets, whereas the H-Nb2O5 maintained the usual electrospun fiber morphology. The single crystalline nature of the M-Nb2O5 was evident from the TEM pictures. The crystal structures of the various phases were confirmed by high-resolution transmission electron microscopy. Parts c, f, and i of Figure 3 show the selected area electron diffraction (SAED) patterns of the H-, O-, M-Nb2O5 respectively and parts b, e, and h of Figure 3 are the corresponding high-resolution lattice images. All of these phases gave spotty patterns characteristic of single crystals; however, the sharpness of the spot increased in the order H-Nb2O5,

666

J. Phys. Chem. C, Vol. 114, No. 1, 2010

Viet et al.

Figure 2. SEM images of Nb2O5 nanofiber (a) before annealing, annealed in air for 1 h at (b) 500 °C, (c) 800 °C, and (d) 1100 °C. Bar scale 2 µm.

O-Nb2O5, and M-Nb2O5 due to their respective higher calcination temperatures. This enhanced crystallinity was well reflected in the extended periodicity observed in the lattice images. Sharp grain boundaries were observed for O-Nb2O5 and M-Nb2O5, whereas no clear boundaries were observed for H-Nb2O5. The particle sizes calculated from the lattice images were similar to that determined from the XRD patterns. Figure 4 shows the X-ray diffraction (XRD) patterns of the Nb2O5 fibers annealed at 500-1100 °C for 1 h. All of the peaks in the XRD pattern of fibers annealed at 500 and 800 °C are indexed for the pseudohexagonal phase (H-Nb2O5) and orthorhombic (O-Nb2O5) phases, respectively. The lattice parameters of H-Nb2O5 are a ) 3.600(2) Å, c ) 3.919(3) Å (space group: P6/mmm) and those of O-Nb2O5 are a ) 6.144(2) Å, b ) 29.194(3) Å, c ) 3.940(4) Å (space group: Pbam). The XRD patterns of the fibers annealed at 1000 and 1100 °C show monoclinic (M-Nb2O5) structure. The M-Nb2O5 heated at 1100 °C shows a small shift in the peak positions as well as lowering of relative intensity of (1 05j) plane compared to that in the 1000 °C heated sample (part b of Figure 4). This shift could arise from incomplete phase formation or due to the presence of Nb interstitials in the low-temperature phase. The Nb2O5 nanofibers were further annealed at 1100 °C for 11 h in air to study the effect of crystallinity and particle size on the electrochemical cycling behavior. A Rietveld refinement of XRD pattern of the

fibers heated at 1100 °C for 11 h is in part c of Figure 4, the refined lattice parameters are: a ) 21.163(2) Å, b ) 3.824(3) Å, c ) 19.355(6) Å. The obtained lattice parameter values of H-, O-, and M-Nb2O5 (1100 °C) in this study are in good agreement with the reported values.20 The average crystallite size (Lorentzian) obtained using the TOPAS software for H-Nb2O5, O-Nb2O5, M-Nb2O5 (1000 °C, 1 h), M-Nb2O5 (1100 °C, 1 h), M-Nb2O5 (1100 °C, 11 h) are 20, 42, 53, and 160 nm, respectively. BET surface areas were 55 ((0.2), 8 ((0.05), and 1.3((0.02) m2g-1 for H-Nb2O5 (500 °C, 1 h), O-Nb2O5 (800 °C, 1 h), and M-Nb2O5 (1100 °C, 1 h), respectively. Observed decrease in the BET surface area could be from higher annealing temperatures for synthesizing the respective polymorphs. The H-Nb2O5 (500 °C) fibers have smaller pores within the grains, which were minimized and/or underwent surface smoothening upon annealing at 800 °C due to grain growth. Consequently, the surface area resulting from the gas adsorption was minimized. Similar decrease in BET surface area with increase in annealing temperature has been reported for a number of ceramic oxides.35,36 Valence states of the metal ions of the Nb2O5 polymorphs have been studied by X-ray photoelectron spectroscopy (XPS). XPS is a well-adopted nondestructive technique for the evaluation of valence states of the metal/nonmetal ions and is

Nanostructured Nb2O5 Polymorphs

J. Phys. Chem. C, Vol. 114, No. 1, 2010 667

Figure 3. Bright field TEM, high-resolution TEM, and SAED patterns of (a, b, c) H-Nb2O5 (500 °C, 1 h), (d, e, f) O-Nb2O5 (800 °C, 1 h), (g, h, i) M-Nb2O5 (1100 °C, 1 h), and (j) M-Nb2O5 (1100 °C, 11 h).

extensively used in the characterization of lithium battery materials and other oxides.37,38 The core-level XPS spectra of H-Nb2O5 and M-Nb2O5 (1100 °C for 1 h) are shown in Figure 5. The Nb-3d level binding energies (BE) of H-Nb2O5 (500 °C) were 209.26 and 206.51 eV, which are assigned to Nd3d3/2 and Nd3d5/2 core levels of Nb5+, respectively. The O1s level BE of oxygen were 529.43 and 531.83 eV, which are assigned to the oxygen (O 2-) anion in Nb-oxides and surface oxygen, respectively. The BE of Nd3d3/2, Nd3d5/2, and O1s in M-Nb2O5 were 209.64, 206.90, 529.88, and 531.06 eV, respectively. The BEs of Nb5+ ions are in close agreement with reported values.38

Slight differences in the BEs of Nb core levels were observed, which were thought to arise from the differences in the crystal structure. 3.2. Electrochemical Studies. Cyclic Voltammetry is a well accepted technique to study the redox couple and structural transformations during Li-intercalation/deintercalation reactions of cathode and anode electrode materials.31,35-37 Figure 6 shows cyclic voltammograms (CV) of Nb2O5 fibers annealed in the range 500-1100 °C. The CV studies were carried out in the range 1.0-2.6 V, with a scan rate of 0.058 mVs-1 and at room temperature. The Nb2O5 cells showed an open circuit voltage

668

J. Phys. Chem. C, Vol. 114, No. 1, 2010

Viet et al.

Figure 4. (a) XRD pattern of Nb2O5 nanofibers sintered at (a) 500-1100 °C for 1 h, (b) a magnified XRD pattern of M-Nb2O5 sintered at 1000 and 1100 °C, and (c) Rietveld refined XRD pattern of Nb2O5 (prepared at 1100 °C for 11 h in air). Symbols represent experimental data and continuous line represents simulated data. Difference curve and miller indices (hkl) lines are shown.

Figure 5. XPS core-level spectra of Nb2O5 fibers sintered at (a) 500 °C (H-Nb2O5) and (b) 1100 °C (M-Nb2O5).

of 2.8((0.1) V. The CV of H-Nb2O5 fibers showed (part a of Figure 6) cathodic and anodic peaks at 1.84 V, which indicates that the Li ion intercalation-deintercalation is one-step process with a broad distribution of energy, or it may be two steps with overlapping energies. On the other hand, O-Nb2O5 fibers show cathodic peaks at 1.82 and 1.62 V and anodic peak at 1.82 V (part b of Figure 6). Interestingly, current peaks of O-Nb2O5 are slightly higher compared to H-Nb2O5, which is directly related to the specific capacity and was clearly revealed in the galavanostatic cycling to be discussed later.

The CVs of the M-Nb2O5 (1000 and 1100 °C) of first, second, and fifth cycle are shown in parts c and d of Figure 6. The first cathodic scan (1000 °C, part c of Figure 6) showed major peak at ∼1.5 V and a minor peaks at ∼1.12, ∼1.04 V. First anodic scan showed major peaks at ∼1.27 and ∼1.73 V and minor peaks at ∼1.4 and ∼2.1 V. Second and subsequent cycles showed reversible peaks (cathodic/anodic) at 1.62/1.73 V and 2.10/2.0 V. The first two cycles are the formation cycles; during these cycles good electrical contact between active material, binder, and conducting carbon and current collector

Nanostructured Nb2O5 Polymorphs

J. Phys. Chem. C, Vol. 114, No. 1, 2010 669

Figure 6. Cyclic voltammograms of Nb2O5 nanofibers sintered at (a) 500 °C, (b) 800 °C, (c) 1000 °C, and (d) 1100 °C. V ) 1.0-2.6 V, Scan rate, 0.058 mVs-1. Li-metal anode was the counter and reference electrode, CV was recorded at room temperature.

Figure 7. Galvanostatic charge-discharge of Nb2O5 nanofibers sintered at (a) 500 °C, (b) 800 °C, (c) 1000 °C, and (d) 1100 °C for 1 h. The numbers indicate cycle number. Voltage range, 1.0-2.6 V versus Li, at a current rate of 50 mAg-1. Geometrical area of the electrode was 2.0 cm2. Cycling studies was carried at room temperature.

is established. The CVs of M-Nb2O5 (1100 °C) (part d of Figure 6) shows clear reversible cathodic and anodic peaks at 1.62 and 1.74 V, respectively. The difference in the CV between the M-Nb2O5 (1000 °C) and M-Nb2O5 (1100 °C) could come from the difference in their crystallinity. Sharp reversible peaks in the CVs observed for M-Nb2O5 (1100 °C) indicate good lithium intercalation/deintercalation. The peak currents were increased from H-Nb2O5 to M-Nb2O5; which indicate that M-Nb2O5 should display higher specific capacity compared to the other polymorphs. This observation was clearly revealed in the galavanostatic cycling performance, that is the initial capacity and the capacity retention to be discussed later. The main cathodic and anodic peaks correspond to insertion/extraction of Li to/from the Nb2O5 lattice, the reaction during discharge scan can be represented as Nb2O5 + xLi+ + xe- f LixNb2O5

(x ) 1, 2).29 The cathodic/anodic peaks correspond to redox couples of Nb5+/4+ and Nb 4+/3+.31 The CV results demonstrated that the cathodic/anodic peak potentials are highly sensitive to morphology and crystal structure. Figure 7 shows results of galvanostatic discharge-charge cycling performed at a current rate of 50 mAg-1 in the voltage range 1.0 - 2.6 V at room temperature of H-Nb2O5, O-Nb2O5, and M-Nb2O5 phases. The first discharge cycle of H-Nb2O5, and O-Nb2O5 (parts a and b of Figure 7) have a similar line profile indicative of a single phase reaction. The differential capacity plots of all the three polymorphs are obtained from the 2nd and 10th galvanostatic discharge-charge cycles (Figure 7), and they are shown in Figure 8. The potentials obtained from the differential capacity plots were similar to CVs of the respective phases (Figure 6). Broad peaks were observed in

670

J. Phys. Chem. C, Vol. 114, No. 1, 2010

Viet et al.

Figure 8. Differential capacity versus voltage plots extracted from the second charge discharge galvanostatic cycles of Nb2O5 fibers sintered for 1 h in air at (a) 500 °C, (b) 800 °C, (c) 1000 °C, and (d) 1100 °C.

Figure 9. Capacity versus cycle number plots of (a) bare H-, O-, and M-Nb2O5, current rate: 50 mAg-1; (b) M-Nb2O5 heat-treated electrode at 220 °C at 6 h in Ar, current rate: 50 and 400 mAg-1. Voltage range: 1.0-2.6 V versus Li/Li+. Li-metal as counter and reference electrodes. Cycling studies were carried at room temperature.

differential capacity plots for H- and O-Nb2O5 and sharp peaks were observed for M-Nb2O5 phases. The specific capacities during second discharge cycle for H-Nb2O5, O-Nb2O5, M-Nb2O5 (1000 °C), and M-Nb2O5 (1100 °C) were 152, 189, 208, and 242 ((3) mAhg-1, respectively. These capacity values correspond to 1.5 to 2.4 mols of lithium per mole of Nb2O5 inserted during discharge cycle. Part a of Figure 9 shows the capacity versus cycle number plots of Nb2O5 fibers/nuggets. The M-Nb2O5 (1100 °C) delivered highest and stable capacity compared to the others. The capacity delivered by the fibers heated at M-Nb2O5 (1000 °C) was slightly higher than that of the O-Nb2O5 phase but

was lower than that of M-Nb2O5 (1100 °C). The corresponding capacity fading with respect to the cycling of M-Nb2O5 (1000 °C) was also higher than M-Nb2O5 (1100 °C). The M-Nb2O5 (1000 °C) showed high capacity fading (∼22%) between the 2nd to 25th cycles; corresponding fading for the 1100 °C phase was 10%. Kodama et al.29 compared the cycling performance of different Nb2O5 polymorphs prepared by heat treatment of commercially available fine powder of Nb2O5. They reported similar initial capacity for O-Nb2O5 and M-Nb2O5 to that of the present results; however, capacity retention differed in nanofibers from their particulate counterparts. In their cycling studies, Kodama et al.29 reported a capacity fading of ∼6% and ∼11% between the second and 19th cycle for O-Nb2O5 and M-Nb2O5 respectively. That is, the capacity fading of M-Nb2O5 is double that of the O-Nb2O5 phase. In the current study using Nb2O5 nanostructures, the capacity fading observed for M-Nb2O5 (∼15%) was half that of O-Nb2O5 (∼30%) between the 2nd and 42nd cycles. We note that the cycling performance can only be qualitatively compared due to the different cycling conditions (operating voltage, current density) and preparation of the electrode (composition, surface area, heat treatment). The particulate analogue showed high initial specific capacity in M-Nb2O5 but with inferior capacity retention compared to those of O-Nb2O5.29 The electrospun nanostructures demonstrate high initial capacity as well as superior capacity retention in M-Nb2O5 compared to those of O-Nb2O5 (Figure 9). The O-Nb2O5 and M-Nb2O5 nanofibers-based cells were cycled in the 1.2 - 3.0 V range with a current density of 150 mA/g to allow the above direct quantitative comparison29 and the work has been published elsewhere. To study the effect of particle size (average crystallite size obtained from XRD data) on the cycling performances, the M-Nb2O5 heated for 1 h (∼53 nm) and 11 h (∼160 nm) were used as cathode and performed galvanostatic chargedischarge cycling. Corresponding TEM images are given in parts g and j of Figure 3. The M-Nb2O5 heated for 1 h duration delivered a reversible capacity of 242 ((3) mAhg-1

Nanostructured Nb2O5 Polymorphs at the end of second cycle with 10% capacity loss between the second and the 25th cycle. On the other hand, 11 h annealed M-Nb2O5 fiber delivered a lower capacity of 210 ((3) mAhg-1 at the end of the second cycle with an increased capacity fading of ∼15%. Li-intercation/deintercalation is expected to be more facile with particle of smaller size compared to the bigger ones, which explains the better cycling performance of the M-Nb2O5 annealed for 1 h (part g of Figure 3) compared to the M-Nb2O5 annealed for 11 h (part j of Figure 3). The rate capability of heat-treated M-Nb2O5 was studied in the same voltage range 1.0-2.6 V with current densities of 50 and 400 mAg-1. Corresponding voltage versus capacity plots are shown in the Supporting Information. At a current density of 50 mAg-1, the M-Nb2O5 (1 h) showed a capacity of 227 ((3) mAhg-1 at the end of the second cycle, with low capacity fading ∼6% at the end of the 60th cycle. At higher current rate of 400 mAg-1, the sample delivered a second cycle capacity of 228 ((3) mAhg-1 and 4% capacity fading cycled between 10 to 100 cycles (part b of Figure 9). The heat-treated M-Nb2O5 showed negligible capacity fading during cycling when compared to electrode without heat treatment even under adverse cycling conditions of high current. 4. Conclusions In conclusion, Nb2O5 fibers in different polymorphic forms were synthesized by commercially scalable nanofabrication technique, that is, electrospinning. The pseudohexagonal (HNb2O5), orthorhombic (O-Nb2O5), and monoclinic (MNb2O5) phases were prepared by annealing as-spun composite fibers at 500 °C, 800 °C, and g1000 °C. The H-Nb2O5 and O-Nb2O5 phases had an average diameter ∼160 nm. The M-Nb2O5 adopted submicrometer sized “near 1D” nuggetlike morphology. The lattice parameters of M-Nb2O5 obtained by annealing the samples at 1000 and 1100 °C differed from one another; the 1100 °C phase has reported lattice parameters. All of the fibers were highly crystalline; particles composing the fibers were increased in size with increase in annealing temperature. Correspondingly the specific surface area of the fibers decreased. The polymorphs of electrospun Nb2O5 fibers were evaluated for their application as a cathode material for lithium batteries. All of the electrospun fibers polymorphs showed good cycling stability; however, the specific capacities differed considerably depending on the crystal structure. The M-Nb2O5 showed highest capacity (∼242 ((3) mAhg-1) among all the phases considered in the present study. The M-Nb2O5 phase also showed the lowest capacity fading among all of the phases. It was also demonstrated from the specific capacity and cycling stability of M-Nb2O5 annealed at different temperatures that incomplete phase formation or highly defective crystalline structure adversely affects lithium intercalation/deintercalation behavior both in terms of specific capacity and cycling stability. The observed stability is thought to arise from the peculiar porous structure of electrospun nanofibers. The results of the present study indicate that electrospun M-Nb2O5 nuggets could be an excellent candidate for rechargeable 2 V lithium batteries in view of the commercial viability of the electrospinning process and performance of the batteries developed herewith. Acknowledgment. Authors thank Prof. G.V. Subba Rao, Physics dept. NUS for his helpful discussions. This project was partially funded by the Clean Energy Program Office, Singapore.

J. Phys. Chem. C, Vol. 114, No. 1, 2010 671 Supporting Information Available: Figure of galvanostatic charge-discharge cycling of heat-treated M-Nb2O5 cycled at current rate of 50 and 400 mAg-1. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Greiner, A.; Wendorff, J. H. Angew. Chem., Int. Ed. 2007, 46, 5670. (2) Thavasi, V.; Singh, G.; Ramakrishna, S. En. EnV. Sci. 2008, 1, 205. (3) Ramaseshan, R.; Sundarrajan, S.; Jose, R.; Ramakrishna, S. J. Appl. Phys. 2007, 102, 111101. (4) Li, D.; Xia, Y. N. AdV. Mater. 2004, 16, 1151. (5) Thomas, V.; Dean, D. R.; Vohra, Y. K. Curr. Nanosci. 2006, 2, 155. (6) Yang, F.; Murugan, R.; Wang, S.; Ramakrishna, S. Biomaterials 2005, 26, 2603. (7) Li, W. J.; Laurencin, C. T.; Caterson, E. J.; Tuan, R. S.; Ko, F. K. J. Biomed. Mater. Res. 2002, 60, 613. (8) Jose, R.; Kumar, A.; Thavasi, V.; Ramakrishna, S. Nanotechnol. 2008, 19, 424004. (9) Jose, R.; Kumar, A.; Thavasi, V.; Fujihara, K.; Uchida, S.; Ramakrishna, S. Appl. Phy. Lett. 2008, 93, 023125. (10) Fujihara, K.; Kumar, A.; Jose, R.; Ramakrishna, S.; Uchida, S. Nanotechnol. 2007, 18, 365709. (11) Onozuka, K.; Ding, B.; Tsuge, Y.; Naka, T.; Yamazaki, M.; Sugi, S.; Ohno, S.; Yoshikawa, M.; Shiratori, S. Nanotechnol. 2006, 17, 1026. (12) Song, M. Y.; Ahn, Y. R.; Jo, S. M.; Kim, D. Y.; Ahn, J. P. Appl. Phy. Lett. 2005, 87, 113113. (13) Barhate, R. S.; Ramakrishna, S. J. Membr. Sci. 2007, 296, 1. (14) Gopal, R.; Kaur, S.; Ma, Z. W.; Chan, C.; Ramakrishna, S.; Matsuura, T. J. Membr. Sci. 2006, 281, 581. (15) Barhate, R. S.; Loong, C. K.; Ramakrishna, S. J. Membr. Sci. 2006, 283, 209. (16) Reneker, D. H.; Yarin, A. L. Polymer 2008, 49, 2387. (17) Reneker, D. H.; Yarin, A. L.; Fong, H.; Koombhongse, S. J. Appl. Phys. 2000, 87, 4531. (18) Sigmund, W.; Yuh, J.; Park, H.; Maneeratana, V.; Pyrgiotakis, G.; Daga, A.; Taylor, J.; Nino, J. C. J. Am. Ceram. Soc. 2006, 89, 395. (19) Aegerter, M. A. SOLMAT 2001, 68, 401. (20) Scha¨fer, H.; Gruehn, R.; Schulte, F. Angew. Chem., Int. Ed 1966, 5, 40. (21) Wang, Y.-D.; Yang, L.-F.; Zhou, Z.-L.; Li, Y.-F.; Wu, X.-H. Mater. Lett. 2001, 49, 277. (22) Carniti, P.; Gervasini, A.; Marzo, M. J. Phys. Chem. C 2008, 112, 14064. (23) Mujawar, S. H.; Inamdar, A. I.; Patil, S. B.; Patil, P. S. Solid State Ionics 2006, 177, 3333. (24) Jose, R.; Thavasi, V.; Ramakrishna, S. J. Am. Ceram. Soc. 2009, 92, 289. (25) Ahn, K. S.; Kang, M. S.; Lee, J. K.; Shin, B. C.; Lee, J. W. Appl. Phys. Lett. 2006, 89. (26) Chen, S. G.; Chappel, S.; Diamant, Y.; Zaban, A. Chem. Mater. 2001, 13, 4629. (27) Eguchi, K.; Koga, H.; Sekizawa, K.; Sasaki, K. J. Ceram. Soc. Jpn. 2000, 108, 1067. (28) Bruce, P. G.; Scrosati, B.; Tarascon, J.-M. Angew. Chem., Int. Ed 2008, 47, 2930. (29) Kodama, R.; Terada, Y.; Nakai, I.; Komaba, S.; Kumagai, N. J. Electrochem. Soc. 2006, 153, A583. (30) Kumagai, N.; Koishikawa, Y.; Komaba, S.; Koshiba, N. J. Electrochem. Soc. 1999, 146, 3203. (31) Wei, M. D.; Wei, K. W.; Ichihara, M.; Zhou, H. S. Electrochem. Commun. 2008, 10, 980. (32) Macias, M.; Chacko, A.; Ferraris, J. P.; Balkus, K. J. Microporous Mesoporous Mater. 2005, 86, 1. (33) Viswanathamurthi, P.; Bhattarai, N.; Kim, H. Y.; Lee, D. R.; Kim, S. R.; Morris, M. A. Chem. Phys. Lett. 2003, 374, 79. (34) Li, J.; Dahn, H. M.; Krause, L. J.; Le, D.-B.; Dahn, J. R. J. Electrochem. Soc. 2008, 155, A812. (35) Reddy, M. V.; Madhavi, S.; SubbaRao, G. V.; Chowdari, B. V. R. J. Power Sources 2006, 162, 1312. (36) Reddy, M. V.; SubbaRao, G. V.; Chowdari, B. V. R. J. Phys. Chem. C 2007, 111, 11712. (37) Tan, K. S.; Reddy, M. V.; SubbaRao, G. V.; Chowdari, B. V. R. J. Power Sources 2005, 141, 129. (38) Ozer, N.; Barreto, T.; Buyuklimanli, T.; Lampert, C. M. SOLMAT 1995, 36, 433.

JP9088589