Experimental conditions for room temperature ferromagnetism in Fe-doped SnO2 via mechanochemical milling and thermal treatment

Identifying optimal experimental conditions, preferably through a simple and cost-effective method, for the fabrication of oxide-diluted magnetic semiconductors, such as Fe-doped SnO2, holds great significance in the quest for spintronic materials operating at room temperature (RT). While mechanochemical milling is a well-established technique meeting these requirements, its numerous milling variables necessitate careful consideration of restricted experimental conditions. In this study, we present some experimental mechanochemical milling conditions to prepare impurity-free iron-doped tin dioxide nanoparticles exhibiting RT ferromagnetic signal. To achieve this, we investigated the effects of milling time, the choice of the starting Sn reactant, and iron concentration on the purity of Sn1−xFexO2 (x = 0, 0.03, and 0.05) nanopowders obtained through mechanochemical milling followed by thermal treatment. Characterization through XRD, XANES, and EXAFS at the Fe K-edge, RT Raman spectroscopy, 119Sn and 57Fe Mössbauer spectroscopies, and magnetic measurements was conducted. Among the experimental techniques, micro-Raman spectroscopy proved the most effective in detecting the formation of hematite as an impurity phase. Our results indicate that extending the milling time to 12 h, as opposed to 3 h, employing anhydrous SnCl2, instead of SnCl2·2H2O and using the low iron concentration of x = 0.03, results in proper conditions for producing impurity-free samples with a robust RT ferromagnetic signal. The oxidation states for iron and tin ions were determined to be 3+ and 4+, respectively, with both occupying octahedral sites, suggesting iron’s replacement of tin. Our findings propose that both the bound magnetic polaron and RKKY models offer potential explanations for the origin of the ferromagnetic signal observed at room temperature in Sn0.97Fe0.03O2 sample milled for 12 h.

Semiconductor doping is not a trivial task although several synthetic methods to produce them have been explored.To produce nanosized SnO 2 , techniques such as spray pyrolysis, pulsed laser ablation, solid-state reaction, co-precipitation, hydrothermal, solvothermal, sol-gel method, and mechanochemical processing have been explored, among others [20].Each method has its own advantages and disadvantages, such as producing impurity phases or causing magnetic ion clustering.Therefore, finding optimal experimental conditions for successful preparation has become a significant challenge in this research area.For instance, the properties of Fedoped SnO 2 are highly dependent on the synthesis method and the precursors used during preparation, which in turn influence the physicochemical, crystallographic, and magnetic characteristics of the final product.
Let's review recent studies from the literature that highlight the ongoing interest in exploring new methods to synthesize these materials and to understand the origin of ferromagnetism.Ahmed et al [21], investigated RTFM in Nd-doped SnO 2 nanostructures prepared by sol-gel method, concluding that defects and oxygen vacancies play a crucial role in the FM signal.The origin of RTFM was explained through the bound magnetic polaron (BMP) model.Jiang et al [22], used the precipitation method to prepare Ni-doped SnO 2 , reporting that the interaction between Ni 2+ ions and oxygen vacancies leads to ferromagnetic coupling, again explained by the BMP model.Kaviya Pandimeena et al [23], observed magnetic switching from paramagnetism to soft ferromagnetism at room temperature.Narzary et al [24], synthesized Sn 0.94−y Ag 0.06 Mg y O 2 by solid-state reaction method and reported that Ag-Mg co-doping induces room temperature ferromagnetism with ultralow coercivity in SnO 2 .Supin et al [25] explored the impact of Co, Fe, and Ni doping on the structural, optical, and magnetic properties of SnO 2 nanoparticles synthesized by the sol-gel method.They found that while pure SnO 2 was diamagnetic, doped SnO 2 nanoparticles exhibited ferromagnetism and paramagnetism at 5 K, and these results were also consistent with the bound polaron model.Chetri and Shukla [26], studied Fe-doped SnO 2 nanoparticles prepared with the sol-gel method and noted that all samples, including those with pure SnO 2 , showed ferromagnetism.Their findings, based on density functional theory calculations, indicated that the absence of oxygen vacancies might promote antiferromagnetic interactions among Fe ions.Pandimeena et al [27], found that Fe x/2 Ni x/2 Sn 1−x O 2 with x = 0.06 and 0.09 prepared by one-step co-precipitation method exhibited soft ferromagnetism and that a crucial factor to maximize this behaviour is the presence of pure polar covalent apical and equatorial Fe-O, Ni-O and Sn-O bonding in absence of interstitial charge accumulation.Vizhi and Rajan [28], studied Co-Fe co-doped SnO 2 nanoparticles synthesized by the sol-gel method, finding that the ferromagnetic signal was influenced by oxygen vacancies, electronic and structural transformations, and surface properties.
In this work, suitable precursor powders are milled to form a nanocomposite structure of the starting materials, which react during milling or subsequent heat treatment.In this case, the required reaction can occur between solid-solid or solid-gas reactants.Although it is well known that the oxidation of SnCl 2 , leads to the in situ formation of SnO 2 nanoparticles with a perfect single crystalline structure [29].This investigation is presented to understand the effects of 3 and 12 h of milling times, the type of starting tin reactive and the Fe concentration on the degree of purity and on the physical properties of Sn 1−x Fe x O 2 (x = 0, 0.03 and 0.05) nanopowders obtained by mechanochemical alloying followed by thermal treatment.On the other hand, annealing temperatures above 500 °C result in a significant degree of nanocrystalline growth and the presence of a normal tetragonal phase of SnO 2 (Cassiterite).One additional step associated with such synthesis is also a washing procedure with distilled water of the salt matrix after annealing treatment [30].The main purpose of this investigation is to find proper milling conditions to produce impurity-free Sn 1−x Fe x O 2 nanoparticles exhibiting room temperature ferromagnetism and to elucidate the source of this magnetic behaviour.

Experimental
Samples of Sn 1−x Fe x O 2 with x = 0, 0.03 and 0.05, were prepared by mechanochemical milling processing and thermal treatment.The starting reactants were SnCl 2 , FeCl 3 , Na 2 CO 3 , and NaCl was added as diluent.To obtain the Fe-doped SnO, the precursors were mixed taking into account the different concentrations of Fe and the following mechanochemical reaction: The sample with x = 0.05 was also prepared from SnCl 2 •2H 2 O instead of SnCl 2 as the tin precursor, in order to compare the influence of water of hydration.Samples were milled for 3 and 12 h under atmospheric conditions in a Planetary ball mill Fritsch Pulverisette 5, with two Cr-based stainless-steel jars of 250 ml of volume and balls of the same material with 12 mm in diameter.The rotation velocity of the disc was fixed to 250 rpm, and the ball to powder ratio (BPR) considered was 20:1.Under these conditions, the cumulative kinetic energy transferred to the powders by the balls during milling is below the order of ∼ 10 8 J kg −1 [31].In our previous reports, we guaranteed that under these milling conditions, contamination came from the milling jars and balls is discarded [32].The as-milled mixtures were subsequently annealed at 600 °C for 3 h in air atmosphere to obtain Fe-doped SnO 2 , as shown in: The products were finally washed with double deionized water to remove the NaCl, until a negative test with AgNO 3 3 M.The samples were characterized by different techniques.x-ray diffraction (XRD) patterns were obtained with a Cu (Kα) =1.5406 Å radiation in Bragg-Brentano geometry; data were collected in 2θ range from 20°to 80°with a 0.02°scanning step and a counting time of 3 s per step.The x-ray data were analysed using the MAUD program, this software can analyse diffraction and other spectroscopic data to get more info on materials, including phases content and crystal structures, microstructural characteristics like size and strains, crystallographic texture and planar defects, by fitting all the XRD patterns using a unified model in a Rietveld like refinement and a Fourier transform analysis [33].In this process, important parameters such as unit cell, average crystallite size, texture and relative volume fraction, among others were refined.The average crystallite size 〈D〉 and the texture were assumed to be isotropic and arbitrary for all phases.Room temperature x-ray Absorption Near Edge Structure (XANES) and Extended x-ray Absorption Fine Structure (EXAFS) spectra at the Fe K-edge (7112 eV) were recorded in transmission mode at the x-ray absorption spectroscopy (XAS) beamline of the LNLS (Laboratório Nacional de Luz Síncrotron) in Campinas, Brazil.The spectra analysis was performed by pre-edge background subtraction followed by a normalization procedure considering the extended region.The fine structure oscillations χ(k) of each spectrum in the extended region were isolated using the ATHENA program and Fourier transformed (FT) over a specific k range [34].Micro-Raman measurements were performed on a Horiba Jobin Yvon LabRam HR system with the 633 nm line of a He-Ne laser.The spectra were fitted to Lorentzian functions with Origin software (2021). 57Fe and 119 Sn Mössbauer spectra were collected at RT by using a 57 Co and Ba 119 SnO 3 sources, respectively.The technique was carried out with a time-mode spectrometer working in the transmission geometry using a constant acceleration drive with a sinusoidal reference signal.The velocity calibration was performed by α-Fe measurements and isomer shift (δ = 0) of BaSnO 3 .All spectra were analysed using the NORMOS program.The hysteresis loops were measured at RT by using a VSM module of a SQUID (superconducting quantum interference device) magnetometer, up to an applied field of 3.0 T.

Results and discussions
Figure 1 shows the XRD patterns for SnO 2 obtained from SnCl 2, Sn 0.97 Fe 0.03 O 2 for 3 h and Sn 0.95 Fe 0.05 O 2 for 3 and 12 h.Additionally, the sample Sn 0.95 Fe 0.05 O 2 from SnCl 2 •2H 2 O for 3h is displayed.The patterns were adequately fitted by introducing the cassiterite phase (tetragonal rutile structure) of SnO 2 [35].Table 1, summarizes: lattice parameters, unit cell volume, crystallite sizes 〈D〉 and SnO 2 phase abundance, derived from the analysis.For undoped SnO 2 , the presence of any contamination coming from jars and balls used in the milling was not observed, corroborating that these milling conditions are suitable for obtaining samples without impurities or foreign phases.Besides, the orthorhombic phase of SnO 2 was not detected, unlike Galatsis et al [30], which reported the presence of the two polymorphic phases of SnO 2 under similar conditions synthesis.Additionally, the SnO phase was not observed in the XRD patterns [36].Now, hematite and/or Sn doped α-Fe 2 O 3 [9,37,38], as an impurity phase was detected for Sn 1−x Fe x O 2 (x = 0.05) for both milling times, 3 h and 12 h obtained from both tin precursors.The insets from figure 1 show the plots of the patterns for the expanded region (30-38°in 2θ range) to demonstrate the presence or absence of hematite α-Fe 2 O 3 (H).
The XRD patterns of the samples with x = 0.03 (not shown here) not revealed the presence of α-Fe 2 O 3 , as demonstrated below by Raman, 57 Fe Mössbauer and XAS techniques.Additionally, it is evident, from figure 1 and table 1, the higher presence of α-Fe 2 O 3 for the sample prepared from SnCl 2 •2H 2 O than those prepared from SnCl 2 .Interestingly, the 〈D〉 for the samples milled for 12 h are higher than those milled for 3 h.This effect could be associated that to a low milling time predominates the reduction of the size and exists a critical milling time where begins the agglomeration of the particles, which increases the 〈D〉.The grain size of materials decreases with milling time and reaches a saturation level when a balance is established between the fracturing and coldwelding events.This minimal grain size is different depending on the material and milling conditions as reported by Suryanarayana et al [39].
Figure 2(a), shows the XANES spectra, at Fe K-edge, for samples with x = 0.03, milled for 3 and 12 h obtained from SnCl 2 along with the reference spectra of Fe, FeO (wustite), α-Fe 2 O 3 , (hematite) and Fe 3 O 4 (magnetite).Clearly, the near-edge structure ruled out the presence of α-Fe, FeO or Fe 3 O 4 .From figure 2(a), also it is observed that the spectra of Fe doped SnO 2 samples are similar to each other and display three main peaks identified in the XANES.The small pre-edge peak A, the dominant main peak B and the shoulder C at the postedge region.From the crystal structure of rutile SnO 2 and α-Fe 2 O 3 , both Sn and Fe atoms have six nearest oxygen atoms, but the configuration of oxygen atoms is different.The inset of figure 2(a) shows that the pre-edge peak of the Fe doped samples is qualitatively too low in comparison to the pre-edge one in the hematite (7115 eV), with Oh symmetry, which can be associated with higher distortion of the octahedron.The peaks B and C in XANES spectra suggest 3+ oxidation state for the Fe ions in the doped samples.Although electronic energies of Fe in our samples are similar to α-Fe 2 O 3 , these peaks are broader and it is observed a splitting in peak crest (B), which could indicate the presence of nonequivalent chemical environments for Fe ions in the crystalline structure of SnO 2 , or the presence of disorder in the material and/or nanosize structures [40], different to this one of Fe ions in the hematite, where the main peak lies between the two peaks observed in the Fe doped samples.Further, the peak broadening is probably related to iron distribution among locally disordered sites whose energy transition varies according to the average bond length as reported by Rodriguez Torres et al for Fe-doped TiO 2 nanoparticles [41].A typical Fourier transformed (FT) Fe K-edge EXAFS spectrum of samples with x = 0.03 is compared in figure 2(b) with α-Fe 2 O 3 powder.From the analysis, we present several points as follows.First, the region between 1 and 2 Å in the Fe doped samples is different from the reference sample and the major peak position in the sample milled for 3 h is shifted to a lower interatomic distance compared with this of the hematite, and the intensity of peak increases with increasing milling time and is shifted to lower values.Second, the intensity of the peaks between 2 and 4 Å, corresponding to the second and the third coordination layers in the doped samples are substantially lower in comparison to the α-Fe 2 O 3 and additionally, appear two contributions, which indicates the existence of tetragonal SnO 2 , consistent with the reported by Bilovol et al [42].These points suggest that the local structure of Fe atoms is completely different from that of the hematite, which indicates that Fe atoms have been incorporated into the SnO 2 host lattice successfully replacing Sn atoms.Additionally, a third point is that in the sample milled for 3 h, appears a shoulder at 1.66 Å and is transformed into a broad band in the sample milled for 12 h.For α-Fe 2 O 3 , the first peak (RFe-O) is located at 1.45 Å, whereas the second peak (RFe-Fe) is approximately 2.47 Å.Then, in the Fe doped SnO 2 samples with x = 0.03, we consider that the main peak located at 1.32 and 1.22 Å milled for 3 and 12 h, respectively, are attributed to Fe-O-Sn in an environment six coordinated and the difference with the peak position of the α-Fe 2 O 3 can be attributed to the difference of scattering phase shift and scattering amplitude between Fe and Sn.Now, the shoulder located at 1.66 Å could be probably associated to Fe 3+ ions located in pseudo-octahedral sites, mostly on the surface of the grains, where the presence of oxygen vacancies and/or defects is higher, which increase with increasing the milling time.On the other hand, in studies of Fe doped TiO 2 and Fe doped SnO 2 samples with x = 0.03, Fourier transformed at Fe K-edge EXAFS have clearly shown the presence of α-Fe 2 O 3 , a fact that is not observed in our Fe doped samples [41].
Raman spectroscopy is a powerful tool used to investigate the changes in crystallinity, disorder, surface and bulk atoms and size effects in nanometric crystallites [43][44][45][46].For SnO 2 the 6 unit cell atoms give a total of 18 vibrational modes in the first Brillouin zone.The representation of these modes using the Koster notation is given by: 1A 1g + 1A 2g + 2A 2u +1B 1g + 1B 2g + 2B 1u + 1E g +4E u .Figure 3 shows the micro-Raman spectra of all samples fitted to Lorentzian functions.Each band is characterized by its frequency (see table 2), intensity, and Full Width at Half Maximum (FWHM).The micro-Raman spectrum of pure SnO 2 exhibits bands located about 471 cm −1 (Eg), 629 cm −1 (A 1g ), and 772 cm −1 (B 2g ), which are assigned to bulk vibrational modes in agreement with those reported for a rutile SnO 2 single crystal [44].The bands at 712, 517, 437, 234 and 302 cm −1 are ascribed to A 2u (νLO), A 2u (νTO), E u( ν2LO), E u (νTO) and E u (νLO) surface vibrational modes.Additionally, the two peaks located at 351 and 586 cm −1 , have not been observed in the Raman spectra of single-crystal and polycrystalline SnO 2 , but are only observed in nanometer sized SnO 2 [45].The micro-Raman spectrum of the Fe-doped samples x = 0.05 for 3 h and 12 h milled samples and this prepared from SnCl 2 •2H 2 O show two peaks at around 225 and 496 cm −1 (A1g modes), as well as three peaks at 293, 409 and 610 cm −1 (Eg modes), which are associated with the presence of hematite [46].Micro-Raman spectroscopy has demonstrated a higher sensibility to the presence of minor amounts of hematite as impurity, in contrast with XRD technique for our samples, in according to reported by Beltran et al [47].The effect of the doping concentration on the intensity, area, FWHM, and position of the principal band of the bulk SnO 2 , located at around 629 cm −1 (A 1g ), for the samples milled during 3 h (not shown here), revealed there is a minimum in the intensities and areas at x = 0.05, which corresponds also with a maximum for the FWHM and the Raman shift.These results could be related to critical concentration at about 5 at% Fe in SnO 2 for which the hematite formation occurs.These results are in good agreement with the relative abundances found by both XRD and 57 Fe Mössbauer spectrometry (see table 4 below).The observed intensity variation can be associated with changes in structural disorder and in particle size in the samples [48].These variations are also observed in the FWHM.It is interesting to notice that the bulk A 1g and the surface E u vibrational modes are clearly visible in all Micro-Raman spectra.In contrast, most surface and nanometric sized vibrational modes disappeared for those samples that evidenced the presence of hematite.Another important fact is that for a given iron concentration, the bulk A 1g Raman band shift to higher wave numbers for those samples milled for 12 h in comparison to those milled for 3 h.Clearly this is an effect of the milling process.
Figure 4 shows the room temperature 119 Sn Mössbauer spectra for the Sn 1−x Fe x O 2 (x = 0.00, 0.03 and 0.05) samples obtained for both milling times.The spectrum of the undoped sample was fitted using one only doublet D1, and the iron doped samples were analysed including two doublets D1 and D2.Both doublets are associated with the presence of Sn 4+ , and the presence of Sn 2+ was not detected.The Mössbauer parameters are shown in   [49][50][51].Now, the parameters of doublet D2 in the iron-doped samples (δ ~0.01 mm s −1 and Δ between 0.63 and 1.00 mm s −1 ) can be linked to a more distorted tin (Sn) microenvironment.From table 3, it is evidenced that the relative area of component D2 increases with a rise in iron concentration, indicating that the presence of iron in the SnO 2 structure distorts the Sn sites.These defects are probably associated with oxygen vacancies, primarily on the surface of the grains [49].
Figure 5 shows the room temperature 57 Fe Mössbauer spectra of the Sn 1−x Fe x O 2 (x = 0.03 and 0.05) samples prepared from SnCl 2 and SnCl 2 •2H 2 O, milled during 3 and 12 h, respectively.The spectra were properly fitted by introducing two doublets and one sextet.The magnetic sextet is associated with the presence of α-Fe 2 O 3 .The use of a second doublet is justified by the presence of a small shoulder located at the left sided peak of the spectra [7].The Sn 1−x Fe x O 2 with x = 0.03 samples, milled for 3 and 12 h present only paramagnetic doublets [52].The derived parameters are reported in table 4.These parameters suggest that all the iron ions are in high spin 3+, replacing Sn in the SnO 2 structure in good agreement with XRD and micro-Raman results.The doublets with the lowest quadrupole splitting can be attributed to Fe ions in less distorted sites, whereas the doublets with the largest quadrupole splitting values to more distorted sites.Since the surface sites are likely to be more distorted than the core ones, these results may also imply that these Fe3+ with higher quadrupole splitting can be located at the surface in agreement with the observations above mentioned in XAS analysis.Another plausible interpretation, initially proposed for Fe doped TiO 2 samples, can be used in our samples too [41,53].For each octahedral site, there are four basal and two apical possible sites for an oxygen vacancy.D1 corresponds to Fe with the vacancy in a basal position and D2 with the vacancy in an apical position.Because the expected 2:1 area relation between the two quadrupolar doublets is not observed, we could speculate that this is due to an additional inhomogeneity in the iron sites due to the difference in the synthesis method used by Rodriguez et al [41].
By comparing the values of the relative area of the sextet component (hematite) for all samples, we can notice two interesting observations: (i) the presence of hematite increases with increasing milling time, (ii) hematite is low for the sample prepared from SnCl 2 in comparison to the sample prepared from SnCl 2 •2H 2 O.These results suggest restricted experimental conditions to produce impurity free Fe doped SnO 2 samples by using the mechanochemical milling technique.
Figure 6 shows the RT magnetizations, M(emu/g), versus magnetic field H(Oe) curves for a) Sn 1−x Fe x O 2 (x = 0.03) obtained from SnCl 2 for both milling times, and b) Sn 1−x Fe x O 2 (x = 0.05) obtained from SnCl 2 and SnCl 2 •2H 2 O for 3 h samples.Undoped SnO 2 (not shown here) and Sn 0.97 Fe 0.03 O 2 milled for 3 h, consist of a linear paramagnetic component.And the other samples consist of both, a linear and magnetic hysteresis loop component.For the samples doped with x = 0.05 (see figure 6(b))) the shape of the loop is associated with the hematite presence, according to XRD, XANES, Raman and 57 Fe Mössbauer techniques.The enhanced antiferromagnetic interaction between neighbouring Fe-Fe ions suppressed the ferromagnetism at higher doping concentrations of Fe, according to Fe-doped ZnO results obtained by a coprecipitation method [54].On the other hand, for the sample with 12 h of milling time and x = 0.03, the magnetic measurements show a paramagnetic behaviour and the coexistence of weak ferromagnetism without the presence of α-Fe 2 O 3 .This fact confirms that low Fe doping concentrations and higher milling time could induce intrinsic ferromagnetic behaviour at room temperature.
In order to separate out the paramagnetic (PM) and ferromagnetic (FM) phases we have fitted the experimental data curves using the function [4]: 8 The first term is the usual function customarily used to fit the ferromagnetic phase hysteresis curves while the second term accounts for the paramagnetic component, with χ as the magnetic susceptibility.The quantities M S , H C and M R give the saturation magnetization of the ferromagnetic component, the coercivity of the hysteresis loop, and the remanent magnetization, respectively.Figure 7 displays the experimental curve, and the ferromagnetic and paramagnetic contributions for the Sn 1−x Fe x O 2 (x = 0.03) sample obtained from SnCl 2 for 12 h of milling time.The absence of a sextet, as evidenced by room temperature 57 Fe Mössbauer spectroscopy, suggests that there is no magnetic coupling between the Fe3+ ions.The saturation magnetization and coercivity estimated from the hysteresis loop turn out to approach 1.2 × 10 -2 emu g −1 and 150 Oe, respectively.According to previous experimental and theoretical results, the values close to those of 8.3 × 10 −5 emu g −1 and 100 Oe reported by Mishra et al [55] for Fe-doped SnO 2 (x = 0.01, 0.03 and 0.05), obtained via the co-precipitation method; and to those reported by Cabrera et al [56], where coercive field values of 30-70 Oe are reported for samples obtained from Fe-doped SnO 2 (x = 0.05 and 0.  originate partially due to the balance of charge in the sample when it is doped with Fe.Following the aforementioned, both BMP and RKKY models could potentially offer explanations for room-temperature ferromagnetism (RTFM) in one way or another.
The activation in the milling processes involves an increase in the reaction ability of a substance that remains chemically unchanged [62].These processes are dependent on initial contact areas, particle sizes, and factors that influence diffusion rates, such as defect densities, local temperatures, and product morphology.Our results suggest that the increase in the activation rate associated with higher milling times and higher Fe concentrations favours the formation of α-Fe 2 O 3 .In this line of thought, the advantages of the mechanochemical milling and thermal treatment of SnCl 2 can be considered as an eco-friendly and cost-effective method to synthesize singlephase Fe-doped SnO 2 -based semiconductor oxide materials.

Conclusions
Mechanochemical milling of SnCl 2 , FeCl 3 , Na 2 CO 3 , and NaCl, followed by thermal treatment, was investigated using a variety of techniques, including x-ray diffraction with the Rietveld method, XANES, Micro-Raman spectroscopy, 57 Fe and 119 Sn Mössbauer spectroscopies, and magnetic measurements.The x-ray diffraction patterns of the Sn 1−x Fe x O 2 samples revealed peaks consistent with the cassiterite phase of SnO 2 in all samples.The Raman spectra exhibited bands characteristic of SnO 2 , along with the presence of hematite (α-Fe 2 O 3 ) in samples with x = 0.05 after milling for 3 and 12 h.The 57 Fe Mössbauer spectra for the Sn 1−x Fe x O 2 samples showed a combination of magnetic and paramagnetic signals.The paramagnetic signals indicated that Fe 3+ ions were incorporated into the SnO 2 crystallographic structure, while the magnetic sextet suggested the presence of α-Fe 2 O 3 .The 119 Sn Mössbauer spectra for all samples revealed only Sn 4+ in octahedral sites, with doublet D1 corresponding to bulk SnO 2 , and doublet D2 suggesting a highly distorted Sn environment due to neighbouring crystal defects.Analysis of the M versus H curves indicated a mixture of paramagnetic and magnetically ordered contributions, which could not be solely attributed to the formation of α-Fe 2 O 3 .Our findings suggest that by extending the milling duration to 12 h, choosing anhydrous SnCl 2 over SnCl 2 •2H 2 O, and using a low iron concentration of x = 0.03, it is possible to create impurity-free samples exhibiting room-temperature  ferromagnetism (RTFM).The origin of the RTFM signal could be satisfactorily explained by the bound magnetic polaron (BMP) and Ruderman-Kittel-Kasuya-Yosida (RKKY) models.

Table 1 .
Unit cell parameters, a and c, unit cell volume, V, average crystallite sizes, 〈D〉, and relative component abundance, of Sn 1−x Fe x O 2 (x = 0.00, 0.03 and 0.05) samples obtained from SnCl 2 for 3 and 12 h of milling time.x = 0.05 * is for sample prepared from SnCl 2 •2H 2 O. x a = b (Å) c(Å) V (Å

Figure 3 .
Figure 3. Micro-Raman spectra of Sn 1−x Fe x O 2 (x = 0.00, 0.03 and 0.05) samples obtained from SnCl 2 and SnCl 2 •2H 2 O for all milling times.The dots are the experimental data, and the continuous line is the result of the fitting process.
10), obtained by mechanical milling of SnO 2 powders in their rutile and α-Fe 2 O 3 phases.The field cooling (FC) and zero field cooling (ZFC) curves are shown in figure 8.The FC and ZFC curves are separated over a wide range of temperatures, and this separation decreases with

Figure 7 .
Figure 7. Fit of RT hysteresis loop of Sn 0.97 Fe 0.03 O 2 sample obtained from SnCl 2 for 12 h of milling time.

Figure 8 .
Figure 8. Field cooling (FC) and zero field cooling (ZFC) curves taken at 500 Oe of Sn 0.97 Fe 0.03 O 2 sample obtained from SnCl 2 for 12 h of milling time.

Table 3 .
Hyperfine parameters derived from the fit of the 119 Sn Mössbauer spectra of Sn 1−x Fe x O 2 (x = 0.00, 0.03 and 0.05) samples obtained from SnCl 2 for milling times of 3 and 12 h.x = 0.05 * is for sample prepared from SnCl 2 •2H 2 O.Estimated errors are of about ±0.01 mm s −1 for the center shift (δ), quadrupole splitting (Δ) and line widths (Γ), and about ±2% for the relative area (A).