The Effects and Role of Polyvinylpyrrolidone on the Size and Phase Composition of Iron Oxide Nanoparticles Prepared by a Modified Sol-Gel Method

Fe2O3 nanoparticles (as maghemite and hematite mixtures) were prepared using adapted sol-gel method from a polyvinylpyrrolidone (PVP) aqueous solution in various Fe : PVP monomer ratios. Analysis of X-ray diffraction was obtained to evaluate the crystalline state, average crystallite size, and composition of iron oxide phases.Themorphology and average nanoparticles size were evaluated by electronic transmission microscopy. Magnetic properties of iron oxide were analyzed at low and room temperatures. Optical characteristics were evaluated by UV-Vis photoacoustic spectroscopy and the Mössbauer spectrum was obtained in order to evaluate subtle changes in the nuclear environment of the iron atoms.


Introduction
The last two decades have seen a significant growth in the study of nanomaterials such as iron oxide nanoparticles.The interest in research related to iron oxides nanoparticles, like magnetite (Fe 3 O 4 ), maghemite (-Fe 2 O 3 ), and hematite (-Fe 2 O 3 ) has been expanded in significate manner at present as a result of their characteristics as magnetism and optical activity and low cost and to their potential applications as catalysts and adsorbent for natural and synthetic micropollutants [1,2], in Li batteries [3], and in some medical and pharmaceutical uses [4][5][6].Moreover, they have attracted much attention because of their surface effect (large surface-area-to-volume ratio) that leads to novel physical and chemical properties and, consequently, the synthesis of nanostructured iron oxides particles has been studied [7].
Various methods have been reported for preparing nanoparticles, for example, sol-gel methods [8,9], ultrasonicassisted methods [10], thermal and hydrothermal pyrolysis [11], combustion [12], and microemulsion [13].However, these methods present some disadvantages such as the control of pH, moisture, and temperature of the medium, as well as the long time required to obtain the final materials and the use of expensive equipment and organic solvents, which are harmful to the environment [14][15][16].To conquer the disadvantages related to previously cited methods, polymeric capping substances are used to protect the nanoparticles and avoid the aggregation.The characteristics of iron oxide can be changed by regulating the diameter and crystallization form.So, to control the rise of nanostructured Fe 2 O 3 , polymeric molecules like polyvinyl alcohol (PVA) and polyvinylpyrrolidone (PVP) are included in the time of the obtaining procedure to cover the nanoparticles [4,[17][18][19][20][21][22].

Synthesis.
Nanostructured Fe 2 O 3 was synthesized using an adapted sol-gel synthesis [4,17].PVP (10% w/v) in aqueous solution and saturated ferric nitrate solutions were separately prepared and then mixed at Fe 3+ : PVP monomeric unit ratios of 1 : 6, 1 : 12, 1 : 18, and 1 : 24 and the samples were, respectively, named as P6, P12, P18, and P24.The mixtures were stirred for 120 min at 25 ∘ C. Afterwards, they were submitted to warm with hard hustle until wholly dehydration.The organic matter was burned at 423 K and nanoparticles were obtained upon calcination at 620 K under air atmosphere during 240 min.

Analysis.
The nanostructured Fe 2 O 3 were evaluated using infrared spectroscopy (FTIR) with an FTIR-BOMEM-100 Spectrometer (KBr pellets 1/100 w/w).The X-ray measurements were obtained by a Siemens D-5000 powder diffractometer using monochromated Cu-K radiation ( = 1.54056Å).The diffractograms were obtained in 10 ∘ to 90 ∘ range.The crystallite diameter was estimated by Scherrer's equation: in which  represents wavelength,  represents full width at half maximum (FWHM) of (104) peak, and cos  represents Bragg angle.The Rietveld refinement has been made, by using the FullProf Suite software, in order to analyze the Xray spectra, in which the following parameters were refined: zero point, scale factor, background, unit cell, half-width, pseudo-Voigt, and asymmetry for the peak shape and thermal isotropic and position factors [23].
Structures of obtained iron oxides were evaluated using a 120 kV JEOL JEM-1400 transmission electron microscopy (TEM), in which samples were put on a thin carbon film Cu grid (200 mesh) (CF200-Cu, EMS).
The photoacoustic spectra were performed with a monochromatic light emitted by a 1000 W xenon arc lamp, Oriel Corporation 68820, with a monochromator (model 77250, Oriel Instruments, mechanical chopper, Stanford Research Systems SR40).
The low temperature (5-300 K) magnetic measurements were performed in a Quantum Design MPMS XL-7 SQUID, whereas the room temperature measurements were performed in a Lakeshore vibrating sample magnetometer (VSM-2T) equipped with an argon flow furnace (300-1200 K).
Mössbauer analysis was obtained using spectrometer with constant acceleration, 57 Co (Rh) source of 25 mCi of nominal starting activity.Mössbauer measurements at low temperature were realized using a helium/nitrogen flow cryostat.Spectra were evaluated by nonlinear least-square method using Lorentzian line shapes.Hyperfine magnetic field distribution ( hf ) was applied as histograms in spectral evaluation.The isomer shift () data are related to -Fe.

Results and Discussion
Figure 1 shows the FTIR spectra of obtained iron oxides.These spectra show resembling appearance indicating that the chemical composition of P6, P12, P18, and P24 is similar.The IR bands at 3400 cm −1 and 1600 cm −1 can be assigned to OH-stretching (] OH) and HOH-bending ( OH) vibrational bands due to the adsorbed water in the sample [24,25].The quantity of adsorbed water increases from P6 to P24, along with the increase of the amount of PVP, probably due to the higher amount of carbon present in the samples after the calcination.Carbonaceous residual tends to be good  adsorbent materials, presenting high amount of adsorbed humidity.The bands in the range of 750-400 cm −1 are related to Fe-O lattice vibration.Hematite phase presents bands at 620, 540, and 470 cm −1 and maghemite phase presents bands at 700, 660-630, 590, and 430 cm −1 , but the positions of the bands are sensitive to the particles size and shape [26,27].It can be observed that the shape of the bands at 540 cm −1 (related to hematite phase) presents some differences with P24.This behavior can be attributed to the different phase composition of this sample, indicating higher maghemite contents than the other samples.
The structure, as well as the average nanoparticle size, was analyzed in all samples using X-ray diffraction (Figure 2).The Fe 2 O 3 phases were estimated using quantitative Rietveld analysis based on the structure data presented in previous studies [28,29].The diffraction patterns are in perfect agreement with the standard JCPDS 33-0664 (hematite), with a minority phase of maghemite JCPDS 39-1346.Crystallite diameters and composition analysis (obtained by XRD) are presented in Table 1.
It is observed that by changing the molar ratio of the Fe 3+ : PVP monomer from 1 : 6 to 1 : 24, a decrease in the peaks intensity of hematite phase (104) and an increase in the peaks corresponding to the maghemite phase (311) : Proposed mechanism of interactions between PVP and Fe 3+ ions in the formation of iron oxide nanoparticles, adapted from Naseri et al. [35].are found; however, for all samples, the majority phase observed is hematite.Previous study using PVA on iron oxide nanoparticles synthesis presents distinct behavior, in which the majority phase observed was maghemite [21].These different behaviors can be attributed to the distinct molecular structure of used polymers (PVP and PVA), since PVA monomers present relative simple chemical structure, with a pendant hydroxyl group, whereas PVP presents a pyrrolidine ring attached to the carbon skeleton.The pyrrolidine ring contains carbonyl groups and a nitrogen heteroatom, beyond the large size of this ligand, compared to hydroxyl group of PVA [30].Probably, these structural and chemical differences between PVP and PVA influence the different obtained iron oxide majority phases, causing growth of iron oxide crystals in different preferential planes [31,32].Also, it is known that high temperatures lead to the transformation of maghemite to hematite [26].As PVP enthalpy of combustion (133.01 kJ mol −1 ) is more than 10 times higher than PVA enthalpy of combustion (10.95 kJ mol −1 ) [33] it can be assumed that the largest energy release in combustion of PVP causes this transformation, tending to form higher hematite phase amount.Although the maghemite peaks are not visible in the P6 and P12 sample diffractograms, the Rietveld analysis confirms the existence of a small fraction of this phase.The nanoparticles average size shows a decrease as the amount of PVP polymer in relation to iron ions is increased (data showed in Table 1), corroborating the results obtained with PVA polymer [21].
The role of PVP in the reactive liquid medium is to afford a polymeric network to prevent the cation mobility permitting local stoichiometry and preventing precipitation of undesirable crystalline forms.Interactions between PVP cap molecules and iron cations are shown in the scheme presented in Figure 3. Fe (III) cations are linked by the strong ionic linkages formed by Fe (III) ions and PVP amide group.
The polymer performs like a support for iron nitrate by means of steric/electrostatic stabilization with pyrrolidine rings (amide functions) and methyl groups.The PVP can be broken into small molecules, producing shorter polymer groups coating the Fe (III) cations, when those are linked.The Fe (III) ions are regularly distributed inside the networks created by the polymeric chains.The performance ends by loss of water.PVP role is not limited to liquid medium.The polymer may too influence the formation of the nanoparticles nuclei in the thermal degradation.
In agreement with results obtained in previous studies using PVP as a capping agent, but using a different thermal treatment for the synthesis of metal oxides, it can be observed that the PVP amount directly influences the obtained particle size and crystalline phases [20,34,35].During the decomposition of this metal ion with PVP capping, initial iron oxides forms are synthetized inside porous carbonaceous matrix.Then, these initial oxides produce the requested phase of nanostructured Fe 2 O 3 .Carbonaceous material degradation eliminates gases (such as CO, CO 2 , NO  , and water vapor) that maintain the combustion energy loss and prevent the aggregation of small particles during the nanoparticles synthesis process [36,37].
In order to evaluate the diameter and morphology of the obtained nanoparticles, the samples were observed using transmission electron microscopy (TEM).
TEM micrographs of P6, P12, P18, and P24 samples are shown in Figure 4.The obtained nanoparticles present sizes that range from 25 to 45 nm (corroborating with average size of 42 nm estimated by XRD analysis).These particles present a hexagonal tendency with clear facets.Figure 4(e) shows TEM micrograph of a unique iron oxide crystallite (sample P24), clearly indicating the hexagonal morphology of hematite nanoparticles, the majority phase obtained.The observed morphology is common for this crystalline phase of iron oxide [38,39].
Hematite and maghemite are n-type semiconductors, and intending to study the optical properties (such as absorption and optical band gap energy) of the obtained samples, the photoacoustic UV-Vis spectrum for P12 sample (Figure 5) was measured.Photoacoustic is a photothermal phenomenon used to evaluate the thermooptical characteristics of materials, enlightening a sample with a modulated light beam and measuring the sound effect induced in consequence of temperature change in a solid.The main advantage of this method is that it allows the study of weakly absorbing or opaque samples, which cannot be evaluated by UV-Vis spectroscopy standard method [40,41].
A broad absorption was registered near the 280-550 nm range of the spectrum.It is the consequence of different electronic transitions: Fe(III) crystal transitions, magnetically coupled Fe(III) interactions, and oxygen-iron charge transfer excitations from O(2p) nonbonding valence bands to the Fe(3d) ligand field orbitals.The charge transfer transitions involving Fe III -O are mainly responsible for absorption of visible light.The electronic transitions 4 A 1 20 ← 6 A, dd electron pair transition, and 4 T 2 ← 6 A 1 present the respective band position range in 420-440 nm, 520-570 nm, and 670-700 nm for hematite and 420-430 nm, 490-500 nm, and 650-700 nm for maghemite [26].
Based on the photoacoustic spectrum data, the band gap energies were calculated, using Tauc's equation [42]: in which  is the absorption coefficient,  is a constant related to the effective mass of the electrons and holes,  is 0.5 for allowed direct transition and 2 for allowed indirect transitions, and   is the energy gap.The band gap values were determined from the curve extrapolation, exhibiting an indirect transition at nearly 1.44 eV and direct transition at 1.73 eV.Higher values of direct band gap (2.1 eV) have already been reported previously for bulk or nanoparticle of hematite/maghemite [43,44].It is possible that the energy gap is influenced by the obtained iron oxide nanoparticles size as well as the presence of two different crystalline phases [45].The indirect transition has been identified as a spinforbidden Fe 3+ 3d → 3d excitation while the direct transition corresponds to the O 2− 2p → Fe 3+ 3d charge transfer [46,47].
The magnetization curves for all samples (Figure 6) were measured as a function of temperature (5-300 K), for an applied magnetic field of 0.05 T, using zero-field-cooling (ZFC) and field-cooled-warming (FCW) procedures.
In the ZFC measurements, the samples were cooled down, from 300 K to 5 K, without the presence of an applied magnetic field.After reach 5 K, the magnetic field was applied and the magnetization was measured during the temperature run.In the FCW measurements, the samples were cooled down to 5 K with the same magnetic field applied during the ZFC.After reaching 5 K, the magnetization was acquired during the temperature run.
All the ZFC-FCW curves (Figure 6) show irreversibility below  irr (temperature below which the ZFC and FCW curves are separated) and these temperatures were found to be very close to room temperature (∼297 K) for the all samples.There are broad peaks below 273 K in the ZFC curves for P18 and P24 samples, with a maximum at  peak ∼ 233 K and 246 K, respectively, that is characteristic of the unblocking step of superparamagnetic nanoparticles [21,[48][49][50].
However, for P6 and P12 samples, that present almost the total composition formed by hematite phase, a transition can be observed in both ZFC and FCW curves, due to Morin transition.The magnetic moments of hematite are nearly perpendicular to the rhombohedral axis with a slight inclining of the magnetic sublattices, resulting in a small net magnetic moment.Throughout freezing, particles suffer a first-order spin reorientation transition.The moments are oriented antiferromagnetically with the rhombohedral axis, and the net magnetic moment reaches zero.For oxides in massive state, Morin transition temperature,   , is nearly 260 K, but, in nanostructured materials,   varies according to crystallite diameter [51,52].For P6 and P12, with 40 and

Journal of Nanomaterials
Relative intensity (a.u.) 300 400 500 600 800 200 700 Wavelength (nm)  43 nm, respectively,   are nearly 232 K for FCW curves and 242 K for ZFC curves for both samples.The magnetization versus magnetic field curves, performed at room temperature, is shown in Figure 7.
The maximum applied field of 2 T was sufficient to saturate all the samples.As can be seen in Figure 7, the highest magnetization saturation (33.3 emu/g) was achieved by the sample P24 (72 wt% hematite and 28 wt% maghemite), followed by 12.4 emu/g of the sample P18 (88 wt% hematite and 12 wt% maghemite) and 9.8 emu/g and 2.2 emu/g, respectively, by the samples P12 and P6.Although sample P6 presents more maghemite amount (5%) then sample P12 (2%), the latter presents smaller particles, and possibly, due to this, a higher magnetization saturation than the first.This can be explained by the presence of defects in the hematite nanoparticles caused by oxygen vacancies near the surface [53].These defects could destroy the antiferromagnetic superexchange interaction of the Fe 3+ -O 2− -Fe 3+ leading to a ferromagnetic ordering [54].
Figure 8 shows the room temperature transmission of Mössbauer spectra of the samples P6, P12, P18, and P24.Table 2 gives the results of the spectra fitting.
At room temperature, hematite spectrum consists of a sextet with a Δ of −0.20 mm s −1 , a  hf of 51.8 T, and  of 0.38 mm s −1 [26,55].Only one crystalline structure site is present for Fe(III) in hexagonal hematite [56].Maghemite presents a spinel lattice, in which Fe(III) sites can be represented by Fe(III)|◻Fe(III)|O 3 , tetrahedral sites are  represented by Fe(III), and octahedral sites by |Fe(III)| in a cubic spinel structure.The symbol ◻ represents vacancies in maghemite.The spectrum of maghemite is composed of two sextets, hardly distinguishable at room temperature, due to the fact that hyperfine parameters for the two sites are very similar in magnitude.Commonly occurring effects such as small particle size distributions tend to cause line broadening and/or a line-shape asymmetry, and, in many cases, the room temperature spectrum has to be fitted with a distribution of unresolved A-and B-site hyperfine fields [57].This is the case here, where the maghemite phase was fitted as distribution M. In surveying the literature for maghemite, it was noticed that an inconceivably wide scatter for the  exists, with the values at room temperature ranging from 0.04 mm s −1 up to 0.30 mm s −1 for A-site and from 0.04 mm s −1 up to 0.40 mm s −1 for B-site and  hf values for the two sites were 50.0 T [26,57].The spectra for all samples were adjusted by introducing one sextet relative to hematite and, for P18 and P24, also a distribution relative to maghemite.It can be observed that the Mössbauer spectra of hematite and maghemite overlap at room temperature, so it is difficult to quantify the relative proportions of these phases in the samples.Due to this, there may be slight difference between the proportions of crystalline phases (-Fe 2 O 3 and -Fe 2 O 3 ) determined by Mössbauer spectroscopy and Rietveld refinement.

Conclusions
This work performed a research in order to comprehend how PVP affects the synthesis of nanostructured iron oxide by a modified sol-gel method.Succinctly, according to the data obtained by XRD diffractograms, TEM micrographs, FTIR spectra, UV-Vis photoacoustic spectra, Mössbauer spectra, SQUID, and VSM curves, PVP act to control the growth and crystalline phase formation of maghemite and hematite.Since the PVP content is increased, the samples do not show Morin transition present in the samples with lower PVP content and the saturation magnetization increases as the PVP amount increases.It is impossible to define an optimum Fe 3+ : PVP monomer ratio because it depends on the sample application and whether it is more important to achieve a higher magnetization or use the Morin transition.Then, this study shows that it is possible to obtain samples with different behaviors only by changing the PVP content.

Figure 1 :
Figure 1: FTIR spectra for obtained iron oxides showing the main differences in samples composition with the variation of Fe 3+ : PVP monomer ratio.

Figure 2 :
Figure 2: XRD pattern of P6, P12, P18, and P24 samples.The Miller indices of reflection are assigned as * for maghemite and # for hematite.The prevalence of hematite phase was observed for all samples.

Figure 5 :
Figure 5: Photoacoustic UV-Vis spectrum of P12 sample showing a broad absorption in the range of 280-550 nm due to the sample electronic transitions.

Figure 7 :
Figure 7: Room temperature magnetization versus magnetic applied field curves showing the dependence of magnetization saturation with Fe 3+ : PVP monomer ratio.

Table 1 :
Hematite and maghemite amount (in weight%) and hematite nanoparticle average size for the obtained samples by modified sol-gel method using PVP.Sample -Fe 2 O 3 (wt%) -Fe 2 O 3 (wt%)

Table 2 :
Hyperfine parameters and subspectral areas for the obtained iron oxides.