Characterization of V-doped SnO2 nanoparticles at ambient and high pressures

Nanoparticles of V-doped SnO2 with stoichiometry Sn1−xO2Vx (x = 0.05, 0.075, 0.125) have been synthesized by a co-precipitation method. Their structural, vibrational, and nuclear properties have been characterized by x-ray Diffraction, Transmission Electron Microscopy, Energy Dispersive x-ray Spectroscopy, Raman Spectroscopy, and Mössbauer Spectroscopy (with 119Sn probe) at ambient pressure. We also performed high-pressure synchrotron x-ray diffraction experiments. The structural behaviour was studied up to ∼10 GPa under quasi-hydrostatic conditions. It has been found that tin dioxide nanoparticles with V are more compressible than un-doped tin dioxide nanoparticles.


Introduction
SnO 2 is a wide band-gap semiconductor with many spintronic applications when doped with transition metals [1]. SnO 2 nanoparticles (NP) are particularly interesting due to they have better performance on solar cells compared to their bulk counterparts [2]. High pressure (HP) experiments are becoming more and more extending over nanomaterials because of the new and potentially interesting properties discovered [3][4][5]. For instance, it has been shown that oxides nanoparticles (NP) could present different transition pressures [6][7][8][9] and different phase transformation sequences [6], presenting on many studies also different compressibilities compared to their bulk counterparts [8][9][10]. Nevertheless, there are some cases where the same behaviour under high pressure was observed for both, nanostructures as well as bulk materials [11].
Regarding HP experiments with doped samples, it has been found that high pressure could be helpful in order to increase the doping concentration. For instance, when doping with Eu 3+ in the LaVO 4 monazite nanocrystals [12]. On the other hand, in Ni-doped TiO 2 NP, it has been found that the transition pressure decreases with the increase of Ni-concentration [13]; also it has been found, in Zn-doped magnetite nanoparticles, that the bulk modulus of magnetite decreases with the increase of Zn-concentration [14].
High pressure x-ray diffraction experiments have been performed on bulk SnO 2 [15] and on pure and 10% Fe-doped SnO 2 nanoparticles [16], finding that the Fe-doped sample presented a bigger bulk modulus than the un-doped one.
In this work, we study the ambient-pressure and high-pressure behaviour of vanadium doped SnO 2 nanoparticles by measuring in-situ x-ray diffraction, using samples with different concentration of V doping. We also characterize the physical properties at ambient pressure by means of Transmission Electron Microscopy, Mössbauer and Raman spectroscopy. The obtained results allow an accurate characterization of the vanadium doping influence in SnO 2 ; in particular, the influence in the SnO 2 compressibility.

Experimental
Vanadium doped tin dioxide nanoparticles with stoichiometry Sn (1−x) V x O 2 (x=0.05, 0.075, 0.125) were prepared by the wet chemical co-precipitation method following the procedure described in [17]. In order to implement the V-doping we have used vanadium chloride (VCl 3 ) from Sigma-Aldrich.
The samples were characterized at ambient-pressure (AP) by means of powder x-ray diffraction (XRD), 119 Sn Mössbauer Spectroscopy (MS), Transmission Electron Microscopy (TEM), Energy Dispersive x-ray Spectroscopy (EDS), and Raman Spectroscopy. The characteristics of the measurements by XRD and MS at ambient pressure are the same that in article of [17].
The TEM micrographs were taken using a TALOS F200X scanning/transmission electron microscope equipped with four x-ray spectrometer detector for EDS measurements. The equipment was used for determining the presence of Vanadium with Energy Dispersive x-ray Spectrometry (EDS) technique.
Raman experiments were carried out in backscattering geometry with a Jobin-Yvon single spectrometer equipped with an edge filter and a thermoelectric-cooled multichannel CCD detector. Measurements with a spectral resolution of 1 cm −1 were performed using the 514.5 nm line of an Ar laser. The laser power was kept below 20 mW to avoid sample heating.
High Pressure (HP) powder diffraction experiments of these samples were carried out at the XDS beam-line of Laboratório Nacional de Luz Síncrotron (LNLS), Campinas, Brazil. The details of the HP powder diffraction experiment are similar to the ones in the experimental section of [14]. The maximum applied pressure (P max ) was different for each sample Sn

Results and discussion
3.1. Ambient pressure characterization 3.1.1. X-ray diffraction at ambient pressure The experimental XRD pattern can be seen in figure 1. These acquired patterns at AP were analyzed by Rietveld refinement using EXPGUI [18] graphic interface of the GSAS [19] analysis program.
All the samples present the peaks of cassiterite SnO 2 , which is isomorphic to rutile (space group P4 2 /mnm), this crystalline structure is graphically represented in figure 2. Tin dioxide is tetragonal (a=b=4.7374(1) Å, c=3.1864(2) Å [20]). The unit cell contains 2 Sn atoms at positions (0, 0, 0) and (½, ½, ½) and 4 O atoms at ±(u, u, 0) and ±(½+u, ½−u, ½) being u an internal parameter that defines the postion of oxygen in the lattice (u=0.3056 at [21]). The first neighbours of the tin atoms are six oxygen atoms, four of them in the equatorial plane, and the rest two in the apical direction.
The structural properties extracted by the Rietveld refinement procedure are shown in figure 3; while the crsytallite size obtained by the analysis is reported in table 1. One of the results of the structural characterization with x-ray diffraction is the observation of the decrease in interatomic distance between Sn and O atoms with the

Raman spectroscopy at ambient pressure
Cassiterite SnO 2 is isomorphic to rutile. It belongs to space group P4 2 /mnm. According to group theory analysis it has four Raman active modes which can be assigned as B1g, Eg, A1g, and B2g [21]. From these modes only three can be detected in our set-up. The B1g mode with a wavenumber of 87 cm −1 is suppressed by the edge filter of the Raman set-up. In table 2 we present the frequencies of the measured Raman modes from un-doped and V-doped SnO 2 . The frequencies of the un-doped sample agree well with the literature [22]. As it can be seen from the table, the frequencies of the modes decrease with the increase of the V content of the samples. At same time, from XRD patterns, we have concluded that the unit-cell volume also decreases, which is equivalent to an increase of pressure. Since the frequencies modes of Raman spectra are known to harden under compression [23], due to the decrease of volume and the Sn-O bond distance, the results of Raman are then apparently inconsistent with the structural data extracted from XRD. The only explanation to the decrease of the Raman frequencies (despite the volume decrease) is that the restoration force associated with the Sn-O bond decreases with the incorporation of vanadium [24]. One possibility for the reduction of the restoration forces associated to Sn-O bond is the lengthening of the interatomic Sn-O distance (in spite of the volume decrease). However, such hypothesis is not supported by the results reported in figure 3. Therefore, assuming a rigid ion model, the most  reasonable cause of the phonon weakening should be related with the smaller ionic radius of V regarding the radius of Sn [21]. As a first approximation, the substitution of Sn by V can be considered as a 'negative' chemical pressure, which could overcome the 'positive' pressure induced by the volume decrease, leading to the observed Raman frequency decrease [25]. Further studies will be needed to confirm this hypothesis.

Mössbauer spectroscopy at ambient pressure
Mössbauer spectroscopy (MS) with radioactive 119 Sn source was performed with all samples at ambient pressure and ambient temperature. In figure 5 we show the Mössbauer absorption spectrum of one of the V-doped SnO 2 samples. All spectra exhibit a doublet which corresponds to a quadrupolar hyperfine interaction of Sn atoms. The hyperfine parameters adjusted to the doublet for each sample can be seen in figure 6. The hyperfine parameters of the pure SnO 2 (without V) are present here as the reference ones, being the value of isomer shift (IS) equal to −0.0187 mm s −1 and quadrupole splitting (QS) equal to 0.52 mm s −1 . The data are in agreement with those reported in literature where the same material as a radioactive source was used [26]. As it is well known, the isomer shift is related to electron density at the nucleus of Sn. The electron density ρ v (0) at the 119 Sn nucleus arises, mainly, from valence 5s electrons. By varying Sn-O distance, IS value should be altered. Accepting the model of the ion-covalent Sn-O bonds [27], which is in agreement with the more general model for Sn-X ones [28], where X: S, Se, Te, for Sn 4+ in octahedral environment, the valence s electrons in SnO 2 are screened by p electrons, causing a displacement of the s electrons from the nucleus. As a result, longer is Sn-O bond, larger is IS value and vice versa.  We would like to remember that in SnO 2 , the Sn atoms are surrounded by six oxygen atoms forming a slightly distorted octahedron SnO 6 . There are two apical Sn-O bond lengths of nearly 2.05 Å and four basal Sn-O bond lengths of nearly 2.06 Å. The mentioned slight distortion of the SnO 6 octahedron is due mainly to the distinct angles between the bonds in the basal plane (nearly 102°and 78°). As a result of this distortion and due to the electrical field gradient at the tin nucleus, a quadrupole interaction appears giving rise to the quadrupole splitting in a Mössbauer spectrum of SnO 2 .
As it can be seen from figure 6, by adding vanadium to SnO 2 , the IS value is affected demonstrating than Sn atom is sensitive to the presence of vanadium in the samples. The trend of the IS behaviour from Mössbauer fits agrees very well with the one for Sn-O distance (average) extracted from Rietveld refinement (figure 2). For pure SnO 2 , where the averaged Sn-O bond is about 2.058 Å, the value of IS is minimum, −0.018 mm s −1 . Doping SnO 2 firstly with 5 mol%. of V leads to a longer Sn-O bond, while the IS value is increased as expected (−0.008 mm s −1 ). When adding more vanadium to SnO 2 , the IS, as a consequence, is progressively decreasing due to the shortening of Sn-O bonds. It is worth to note that the shortening of bonds as it was extracted from Rietveld refinement is averaged value of four basal and two apical bond lengths. As it can be noted from figure 3, the trend for apical and basal bond lengths is opposite. While the apical bonds lengths become shorter, the basal lengths are larger as a result of the presence of vanadium in the SnO 2 . Similar effect on Sn-O bond lengths was reported for tin dioxide doped with 19 mol%. of vanadium [29].
On the other hand, the presence of V in the samples alters the QS values of the Sn atom. For SnO 2 doped with 5 mol%. of V, the value of QS grew up to 0.62 mm s −1 indicating a higher degree of distortion of Sn environment in comparison to the undoped sample. Since the lattice parameters of the SnO 2 lattice were progressively decreasing with V content, we assume that V atom is incorporated to SnO 2 matrix. Moreover, it is very likely that its incorporation is substitutional for Sn as it was previously observed, for instance, for Fe-doped SnO 2 [30]. The second nearest neighbours (SNN) of the central Sn atom are two Sn atoms placed symmetrically respect to the plane of four basal oxygen at nearly 3.18 Å. Due to a probable scenario, a vanadium atom substitutes for one of these SNN progressively. Since the ionic radii of Sn and V are unequal, the substitution of V for Sn causes some degree of distortion of SnO 6 octahedron. Probably, this phenomenon is more effective for the case when the vanadium content in SnO 2 is 7.5 mol%. (the QS reaches a maximum value for this content of V). When SnO 2 is doped with 12.5 mol%. of V, the environment of Sn seems to be more symmetrical than for the latter case (QS is near 0.55 mm s −1 ). It could be thought that for a relatively high doping level, V begins to substitute also the second tin atom from the SNN situated symmetrically respect to its pair. In such a case, the symmetry around the central Sn atom should be higher.

High pressure x-ray diffraction
In figure 7 we present the x-ray diffraction patterns of all the samples measured at different pressures. The sample Sn 0.925 V 0.075 O 2 was measured in DAC with lower angular aperture, the integrated 1D pattern shows data up to lower 2θ than the other samples. All patterns at all pressures present the diffraction peaks of the tetragonal rutile structure of SnO 2 (space group P4 2 /mnm), discarding any phase transition in any of the samples at the measured pressures. All the Bragg peaks of the samples shift to the right (to higher 2θ position) with pressure as it is expected due to the decrease of the lattice constants.
The analysis of the Bragg peaks positions by the Rietveld refinement done by using the MAUD program allowed determining the lattice constants for each sample at each applied pressure. The obtained unit-cell parameters for each sample at all pressure measured are displayed in figure 8.
A linear regression has been fitted to the data of unit-cell parameters a and c of the tetragonal rutile structure in order to obtain the linear compressibility of the samples. The results are displayed in table 3 together with their reference.
As it can be seen from precedent table, it is clear that as the Vanadium content in tin dioxide increases, the linear compressibility of both axes a and c of the tetragonal rutile structure also increases.
We have checked that, with exception of few values, normalized pressure (F) is independent of the Eulerian strain (f) [31]. This indicates that the experimental data are adequately described by a second-order EOS; i.e. the pressure derivative of the bulk modulus (B 0´) is equal to 4. The resulting values of bulk modulus (B 0 ) for our V-doped Tin dioxide nanoparticles are collected in table 4, together with the results for undoped (SnO 2 ) Tin dioxide sample analyzed at [16].  As it can it can be observed in table 4, the incorporation of Vanadium atoms into SnO 2 lattice clearly influences its structure as well as its elastic properties: the bulk modulus decreases with increasing Vanadium content, indicating a more compressible material (as it could be foreseen from the linear compressibilities).We can only compare this result with our work on Fe-doped SnO 2 nanoparticles at ref [16]. In the mentioned study the Fe-doped (10 mol%) SnO 2 sample resulted to have a bigger bulk modulus (218 GPa) than undoped one; in contrast to our current work on V-doped SnO 2 . This is a clear demonstration that how varying doping elements it is possible to tune the compressibility of SnO 2 matrix.

Conclusions
In this work, we report a room-temperature powder study on V-doped Tin dioxide nanoparticles (stoichiometry Sn (1−x) V x O 2 (x=0.05, 0.075, 0.125)) at ambient pressure and also under compression using synchrotron radiation. The samples were synthesized by a co-precipitation technique and, previously to the HP experiments, characterized at ambient pressure showing that the V-doping content affects the structure and the hyperfine parameters retrieved by Mössbauer spectroscopy.
We determined the effect of pressure in the tetragonal rutile structure and a pressure-volume EoS for their different V-doped concentrations. None of the samples undergoes a phase transition at the pressures measured. It has been found than V-doped SnO 2 nanoparticles are more compressible than pure SnO 2 nanoparticles. Further, the volume compressibility increases with Vanadium concentration, as found from the adjusted bulk modulus values from second order Birch-Murnaghan equation of state to the unit-cell volume versus applied pressure data.