Metal oxide nanomaterials offering many promising applications in different areas of physics, chemistry, biology, medicine and metal science have received considerable attention [1-4]. V2O5 is an important multifunctional transition metal oxide owing to its remarkable structural, electrical, chemical, optical, electrochemical and thermochromic properties, along with low cost, bio-compability and easy synthesis making it attractive for many applications. It can be used as chemical gas and moisture sensors [5,6], catalysts , solar cells , super capacitors , cathode material in rechargeable Li batteries , optoelectronic switches , electrochromic and thermochromic devices [12,13].
Bulk V2O5 is a diamagnetic semiconductor with the band gap energy of Eg = 2.2 – 2.3 eV . It has several polymorphs including the orthorombic (Pmmn) α-V2O5 , tetragonal or monoclinic (P21/m) β-V2O5 [16,17] and orthorombic (Pnma) γ-V2O5 . The former is the most common phase built up from the V5+O5 square pyramids linked in the layers by sharing their edges and corners. The other two phases can be obtained from α-V2O5 one under special conditions of high temperature or pressure . In the report of Filonenko et al. , the β-V2O5 phase was attained by a high pressure-high temperature method and characterized also as a layered structure of edge-sharing V5+O6 octahedra chains along b axis. β-V2O5 is metastable and thermogravimetric studies showed its transformation to α-V2O5 phase upon heating to 370-380 °C under ambient pressure . γ-V2O5 phase is also metastable and has been reported to transform easily to α-V2O5 through heating at temperatures higher than 340 °C . However, the phase transitions are significantly affected by the sample preparation conditions.
In recent years, the advances in synthesis of V2O5 nanostructures have drawn more attention to them. To date, various methods have been employed for their preparation, such as hydrothermal , solvothermal , sol-gel [22,23], spray pyrolysis , precipitation , mechanical alloying , pulsed laser deposition , electron beam evaporation  and chemical vapor deposition . Among them, sol-gel may be the most favorable method for preparation of highly pure and well-dispersed homogeneous nanocrystals. It involves mainly hydrolysis of molecular precursor in aqueous solution (sol) resulting in a loose gel with liquid filled pores, comprising either discrete particles or continuous polymer networks, and then removal of the liquid and condensation of the product via heat treatment. There are a lot of reports on the synthesis of different V2O5 nanostructures through sol-gel based techniques. However, controlling the growth process to obtain nanocrystals of narrow-sized distribution with uniform morphology is challenging.
In this investigation, we synthesized V2O5 nanoparticles through a rather facile sol-gel route in gelatin medium for the first time. Here, gelatin, a natural polymer capable of being readily dissolved in hot water, contains functional groups that can bind to metal ions and acts as a suitable polymerization agent and owing to its expansion upon heating, it also serves as a terminator for nanoparticles growth hindering the particles to come together. The sol-gel process was followed by calcination at different temperatures of 400, 500 and 600 °C. To explore the optimum calcination temperature, the prepared samples were studied for their structural, morphological, optical and magnetic properties.
MATERIALS AND METHODS
To synthesize about 2 g of V2O5 nanoparticles via a sol-gel route in a gelatin medium, 4 g of gelatin from bovine skin (glatin type B from Sigma-Aldrich) as a stabilizer was firstly dissolved in 120 ml of deionized water in a water bath of 40 °C on a hot plate and stirred for about 30 min to achieve a clear solution. Meanwhile, about 3.5 g of vanadium(III) chloride (VCl3 from Merck with purity of better than 99%) was dissolved in minimum amount of deionized water and stirred for about 30 min. Then, the vanadium solution was gradually added to the gelatin solution while being stirred. The temperature of the water bath raised to 80 °C and kept fixed for about 6 h evaporating the excess solvent until a viscous brownish gel was obtained. The porous xerogel was formed by drying the gel in a oven of 90 °C for 2 h followed by cooling to room temperature. Then, it was powderized using an agate mortar and calcined in air atmosphere at different temperatures of 400, 500 and 600 °C for 10 h using an alumina crucible in a box furnace to remove the gelatin material. Afterwards, it was allowed to be cooled naturally to room temperature to obtain the final nanopowders of green color.
The as-synthesized V2O5 particles were characterized for structural, morphological, optical and magnetic properties to find out the optimum calcination temperature. Their structural characteristics and phase formation were evaluated by analyzing the X-ray diffraction (XRD) patterns recorded by a D8-Advance Bruker diffractometer using Cu-Kα radiation (λ = 1.54056 Å) in the angular range of 2θ = 10 – 80 ° in a continuous scan mode with a step size of 0.04 °. The XRD profiles were refined by CELREF software working on the basis of the least-square method.
The shape and size of the synthesized particles were studied by transmission electron microscopy (TEM) of the suspensions of the involved powders dispersed in ethanol and deposited over some carbon coated Cu grids using a Philips CM120 TEM system.
For optical characterization, to avoid difficulties in obtaining the UV-Vis absorption spectra of the dispersed samples, the diffuse reflectance spectra (DRS) of dry powders were studied. To do so, the DRS of dry-pressed disks with thickness of less than 1.5 mm from the synthesized powders and the standard material of BaSO4 with the ratio of 1:10 were recorded in the wavelength range of 250 – 1100 nm using an AvaSpec-2048-TEC Avantes spectrometer with AvaLight DH-S deuterium-halogen light source.
The magnetic behavior of the samples were examined through the room temperature magnetization loops measured by a homemade vibrating sample magnetometer (VSM, Meghnatis Daghigh Kavir Kashan Co., Iran) under the applied magnetic fields up to ±14 kOe.
RESULTS AND DISCUSSION
Structural and morphological analysis
The XRD patterns of the V2O5 powders with the calcination temperatures (Tc) of 400, 500 and 600 °C are shown in Fig. 1. All the samples are crystalline; this was expected since it has been reported that the crystallization of V2O5 starts upon heat treatment at 300 °C . The sample calcined at 400 °C exhibits predominately the diffraction peaks belonging to α-V2O5 orthorombic phase with Pmmn space group (JCPDS file No. 41-1426) along with some additional sharp peaks at about 12.4, 18.9, 27.9 and 28.8 ° corresponding to (200), (221), (240) and (241) reflections of β-V2O5 tetragonal phase (JCPDS file No. 45-1074), respectively. There are also a few peaks of negligible intensity from V3O7 monoclinic minor phase with C2/c space group (JCPDS file No. 27-0940). The coexistense of different vanadium oxides minor phases has been reported for other attempts to synthesize vanadium pentaoxide [30-32], as well.
On increasing the calcination temperature, the intensity of the reflections belonging to β-V2O5 and V3O7 crystalline phases increases and moreover, two additional peaks at 17.8 and 45.8 ° emerge which can be indexed as (002) and (005) characteristic reflections of V6O13 monoclinic phase with P21/a space group (JCPDS file No. 25-1251). This shows that the heat treatment not only controls the crystallization process, but also results in emergence of more oxygen-rich phases. The calcination at Tcal = 400 °C provides V and O ions enough mobility to form the crystalline α-V2O5. At higher temperatures, as the ions attain more energy, some of α-V2O5 unit cells are distorted to form the β-V2O5 ones.
The enlarged view of the region around (001) peak (Fig. 2a) reveals a slight shift to higher angles on increasing the calcination temperature. In other words, the interplanar spacing shrinks with the calcination temperature increment. The refined lattice parameters of α-V2O5 unit cell in the synthesized particles are given in Table 1. The lattice contraction on Tcal increment observed also for some other oxide nanoparticles under heat treatment [33-36] can be associated to the decreased density of oxygen vacancies and the consequent local lattice disorders whose probable result may be a shrinkage in the unit cell and/or the presence of the incompletely coordinated ions at the surface of the nanoparticles. The surface ions having dangling bonds form a parallel array of repulsive electric dipoles. These repulsive interdipolar forces decrease on account of the adsorbed oxygen ions at higher temperatures and hence the attractive electrostatic interactions between vanadium and oxygen ions increase resulting in a lattice contraction .
Another effect of the calcination temperature is also revealed from the enlarged peaks view in Fig. 2a as a reduction in the width of the diffraction peaks, i.e. a growth in the crystallite size. The Williamson-Hall (W-H) equation was used to determine both the crystallite size (D) and lattice microstrain (ε) of the synthesized particles based on the the full width at half maximum (FWHM) of the peaks:
where, θ is the diffraction angle of a main reflection, β is FWHM of that peak and λ is the X-ray wavelength, 1.54056 Å. The W-H plots (β cos θ against 4 sin θ ) for the synthesized samples are presented in Fig. 2b. First of all, they exhibit a positive slope indicating the presence of uniform tensile strains in the involved particles. The values of crystallite (grain) size and lattice microstrain obtained from the intercept and slope of the linearly fitted W-H plots, respectively, are listed in Table 1. The synthesized particles are found to be nanocrystalline whose average size increases from 22 to 29 nm upon raising the calcination temperature from 400 to 600 °C. This can be due to presence of oxygen vacancies and the consequent local lattice disorders, again, whose reduction upon increasing the calcination temperature results in growth of the ordered domains. Another reason for the grain growth observed upon increasing Tcal has been proposed by Fang et al. : at high temperatures, the ions attain the sufficient energy to diffuse and reside at the suitable lattice sites resulting in larger grains with less surface energy.
The dislocation density being a measure of the defects in the synthesized crystalline particles was evaluated by Williamson and Smallman’s relation: δ = 1/D2 . The calculated δ values tabulated in Table 1, as well as the ε values are consistent with the crystallite size variations in the involved nanoparticles. The calcination at elevated temperatures results in less microstrain and dislocation and consequently larger crystallites.
The morphology of the synthesized particles was studied by transmission electron microscopy. The TEM micrographs of the samples with different calcination temperatures in Fig. 3 represent that all the samples consist of individual spherical-like and oval-like nanoparticles of approximately uniform size. The effect of raising Tcal is revealed as the increasing number of the oval-shaped particles, probably due to the existence of a mixed phase, as well as a slight increment in their mean size. Regarding Fig. 3, the mean particle size of the samples with Tcal of 400, 500 and 600 ° are found to be about 34, 36.5 and 40 nm respectively, i.e. more than the crystallite sizes calculated from the XRD peaks width. This indicates the polycrystalline nature of the synthesized nanoparticles. The nanoparticles size growth with the calcination temperature is in good agreement with other reports (e.g. ) and was expected since on temperature increment, the tendency toward the minimum surface energy gives rise to larger particles with the less surface to volume ratio.
The room temperature UV-Vis diffuse reflectance spectra (DRS) of the involved nanoparticles in the wavelength range of 250 - 1100 nm are presented in the inset of Fig. 4a. To use the diffuse reflected data, they were recalculated by Kubelka-Munk (K-M) equation :
Here, K and S are the K-M absorption and scattering coefficients representing the portions of light absorbed and scattered per unit vertical lengths, respectively; They have no direct meaning on their own. R is the ratio of the sample’s reflectane to that of the standard (here, BaSO4) and F(R) is the K-M or re-emission function representing the optical absorption behavior. The DRS data of the synthesized particles after K-M treatment are shown in Fig. 4a. As seen, the high values at lower incident photon energies (hν) decrease and then increase for higher energies, very different from the absorption behavior of thin films (e.g. ). The high absorbance (α) at lower energy side of the spectrum hardly originates from the electronic transitions; This should be due to the scattering phenomenon in the involved samples consisting of some nanoparticles with rough surfaces.
The obtained spectra reveal two absorption bands around 1.31 and 2.35 eV almost independent of the calcination temperature. The observed peaks in the absorption spectra do not correspond to the optical band gap energy of the samples. However, occurrence of a sudden raise in the absorbance at a particular incident energy reflects the existence of an optical bang gap in the involved particles. Considering a parabolic band structure, the optical band gap energy (Eg) and its type can be estimated using Tauc’s relation :, where c is a constant and index n, depending on the nature of the involving electronic transition, is 2 and 1/2 for the direct and indirect allowed transitions, respectively. Assuming that the samples scatter in a perfectly diffuse manner, the K-M absorption coefficient (K) equals to 2α, so considering S independent of energy , one has:
The graphs of (Fhν)2 and (Fhν)1/2 versus hν were plotted and it was found that the former has the best line fitting revealing that the synthesized nanoparticles are the direct bandgap materials. The direct band gap values were deduced by extrapolating the linear portion of the curves to the energy axis (F = 0), as shown in Fig. 4b. Despite the differences in the phases formed in the particles calcined at 400 to 600 °C and their mean size (D’), their optical band gap energies show only a slight difference from 2.92 to 2.77 eV (Table 1). The obtained Eg values are comparable to but more than that reported for the bulk V2O5 (2.3 eV, ) due to the quantum confinement effect in the synthesized nanoparticles. The decrease observed in Eg values of the involved nanoperticles with the calcination temperature or in other words with the particle size growth is also consistent with the quantum size effect.
Based on the obtained Eg values, the radii of the synthesized nanoparticles (R) were additionally determined using Brus equation based on the effective mass approximation :
where Egb is the optical band gap energy of the bulk material, μ is the combined effective mass of electrons and holes (= me mh/(me + mh), being equal to 1.99 mo for V2O5, using me = 3.0 mo and mh = 5.9 mo ; mo is the free electron mass), eo is the permittivity of free space, and ε is the dielectric constant (ε = 5.9 for α-V2O5 ). Considering the spherical-like particles, their size (Do = 2R) determined by this way are in the range of 35 – 40 nm (Table 1). They are almost the same as the mean particle sizes obtained from TEM images and a bit larger than the mean crystallite sizes obtained from the XRD lines broadenings.
The room temperature magnetization (M) of V2O5 nanoparticles freshly calcined at 400 °C versus the applied magnetic fields (H) up to ±14 kOe is shown in Fig. 5. The magnetic loop with a small remanence (0.0056 emu/g) and coercivity (135 Oe) demonstrates a weak ferromagnetic behavior for the involved nanoparticles (see lower inset in Fig. 5). It is notable that the XRD patterns reveal no detectable sign of any magnetic impurity phase. V2O5 whose V5+ ions (3d0) have no unpaired electrons does not possess intrinsic magnetic moments as well, but a direct connection has been experimentally and theoretically established between the oxygen vacancies and ferromagnetism in this material [46,47]. By density functional theory (DFT) calculations, Xioa et al.  predicted that V2O5−δ would be a diluted ferromagnet with magnetic moment of about 2 μB/vacancy for δ < 0.13 and 0.19 < δ < 0.45.
As seen, the involved nanoparticles are not saturated even under the maximum applied magnetic field of 14 kOe. The saturation magnetization (Ms) of the synthesized nanoparticles was obtained considering the law of approach to magnetic saturation for high magnetic field region :
M(H,T) = Ms(T) (1 – a/H – b/H2) + χhf H (5)
Here, the fitting coefficients a and b are related to the contributions of local inhomogeneities and the magnetocrystalline anisotropy, respectively, and the last term is the forced magnetization. As the forced magnetization is negligible below Curie temperature and so at room temperature, one can estimate the Ms value through ignoring the last term and just fitting M(1/H) data by a second order polynomial function, as shown in the upper inset of Fig. 5. Accordingly, the room temperature Ms value of the involved V2O5 nanoparticles was found to be 0.139 emu/g being small as compared with some other reported values for nanoparticles of this compound with similar size (2.05 emu/g, ), possibly due to the less oxygen vacancies present in the involved sample.
CONCLUSION The V2O5 nanoparticles with different particle sizes ranging from about 34 to 40 nm were synthesized via a simple sol-gel route in gelatin medium followed by calcination at different temperatures of 400, 500 and 600 °C. The structural characterization shows that the synthesized particles consist mainly of α-V2O5 orthorombic crystallites of 22 – 29 nm and the amount of other vanadium oxides minor phases, namely tetragonal β-V2O5, increases with the calcination temperature. The calcination at elevated temperatures results also in the lattice contraction; this can be due to the decreased density of oxygen vacancies. The optical studies indicate that the involved V2O5 nanoparticles are direct band gap semiconductors whose band gap value decreases with calcination temperature from 2.92 to 2.77 eV due to the size effects. The size of the synthesized nanoparticles estimated based on the optical results are in very good agreement with the mean particle sizes obtained by electron microscopy (35 – 40 nm). The V2O5 nanoparticles calcined at 400 °C showed a weak ferromagnetism at room temperature with saturation magnetization of 0.139 emu/g owing probably to the small amount of oxygen vacancies present in the sample.
CONFLICT OF INTEREST
The authors declare that there is no conflict of interests regarding the publication of this manuscript.