b Physics and Chemistry Department, Technical University of Cluj-Napoca, ..... [3] C. Murugesan, M. Perumal, G. Chandrasekaran, Structural, dielectric and ... [11] A. Goldman, Modern Ferrite Technology, Second edition, Springer, 2006, ...
Nanocrystalline/nanosized iron ferrite-Fe3O4 obtained by a combined route ceramicmechanical milling. Effect of prolonged milling time on the chemical composition, formation of phases and powder characteristics
T.F. Marinca, H.F. Chicinaş, B.V. Neamţu, I. Chicinaş, O. Isnard, F. Popa, P. Păşcuţă DOI: 10.1016/j.apt.2016.05.022
Please cite this article as: Please cite this article in press as: T.F. Marinca et al., Nanocrystalline/nanosized Fe3O4 obtained by a combined route ceramic-mechanical milling. Effect of milling on the chemical composition, formation of phases and powder characteristics, Advanced Powder Technology 27 (2016) 1588-1596
Nanocrystalline/nanosized iron ferrite-Fe3O4 obtained by a combined route ceramic-mechanical milling. Effect of prolonged milling time on the chemical composition, formation of phases and powder characteristics T.F. Marincaa, H.F. Chicinaşa, B.V. Neamţua, I. Chicinaşa, O. Isnardc, F. Popaa, P. Păşcuţăb a
Materials Science and Engineering Department, Technical University of Cluj-Napoca, 103-105, Muncii Avenue, 400641 Cluj-Napoca, Romania b Physics and Chemistry Department, Technical University of Cluj-Napoca, 103-105, Muncii Avenue, 400641 Cluj-Napoca, Romania c Institut Néel, CNRS/Université Joseph Fourier, BP166, 38042 Grenoble, Cédex 9, France Abstract Iron ferrite powder (FeFe2O4) has been synthesized via a combined route ceramic method and mechanical milling, starting from stoichiometric mixture of the easily accessible Fe and Fe2O3 precursors. In the first step, the iron ferrite has been obtained in polycrystalline state by heat treatment of the precursors. In the second step, the as obtained Fe3O4 has been processed in a high energy planetary ball mill in order to attain nanocrystalline nanosized particles. The nanocrystalline/nanosized powder is obtained after only 5 minutes of milling. In order to characterize the powders, X-ray diffraction (XRD), magnetic measurements M=f(H), scanning electron microscopy (SEM), Fourier transform infrared spectroscopy (FTIR), laser particle size analyzer and differential scanning calorimetry (DSC) measurements were performed. X-ray diffraction patterns reveal iron contamination for milling times higher than 30 minutes. The FTIR measurements shown the contamination begins at much shorter milling duration than the X-ray diffraction indicated. The mean crystallite size after 4 hours of mechanical milling is 16 nm, while the lattice strain is almost null at this milling duration. The measurements indicate the presence of a small amount of superparamagnetic particles. Introduction Soft magnetic ferrites are generally materials with AB2O4 chemical formula, which crystalize in spinel structure [1-4]. This structure contains eight formula units, leading to 56 ions per unit cell. The oxygen atoms are arranged face-centred cubic while the metallic ions occupy the interstices between them. This spaces, between oxygen atoms, can be tetrahedral sites (A) or octahedral sites (B). Generally, A can be a bivalent metal such as Zn, Mg, Mn, Cu and Fe, or a group of metallic elements with 2+ as total valence [5-8]. If A is a group of metallic elements the ferrite is considered as mixed ferrite. Their spinel structure has a great importance on the electric and magnetic properties [2-4, 6-9]. For the spinel mineral, the 2+ ions are situated in A sites while the 3+ ions are situated in the B sites. This is known as normal spinel structure [10]. The inverse spinel structure has the divalent ions situated in B sites, and the trivalent ions equally distributed between the A and B sites. The inverse spinel structure is also found for the iron ferrite. Iron ferrite
has the Fe3+ ions distributed evenly between the two sub-lattices, octahedral and tetrahedral sites, while the Fe2+ ions are distributed only in octahedral sites [10]. Neither of these two sites, tetrahedral or octahedral, are fully occupied. The origin of the magnetic properties of the iron ferrite comes from the ion distribution in their structure [4,7]. The magnetic moments of the Fe3+ ions are annulling each other, leading to a resulting magnetisation given by the Fe2+ ions. The value of the resulting magnetisation of FeFe2O4 is 4μB/molecule [10,11]. Ferrites are ceramic materials, and as a result they present higher electrical resistivity than metals, and their alloys, making them appropriate for high frequency uses [12-14]. The classical route for obtaining ferrites is the ceramic method [3,10,15-17]. This consists in homogenisation, pressing and sintering of a stoichiometric mixture of oxides. In order to obtain nanocrystalline, nanosized or nanostructured ferrites several techniques are used: mechanosynthesis [1,9,15-17], sol-gel method [18-20], co-precipitation [21-23], and other methods. This class of soft magnetic materials are important due to their technical applications, such as antennas miniaturisation, nanofluids used in hyperthermia or microinductor applications [24-26]. Mechanosynthesis is a technique well suited for obtaining nanostructured/nanocrystalline materials. This state is achieved through the result of flattening, fracturing and cold-welding processes that occur during ball milling [27]. Mechanical milling of polycrystalline ceramic ferrites leads to the enhancement of their magnetic properties. It is well known that the magnetic permeability and the coercive field are dependent of the mean crystallite size, and the nanocrystalline state leads to better soft magnetic properties [28]. This paper presents results concerning the synthesis and characterisation of the nanocrystalline/nanosized Fe3O4, prepared by a combined method which associates the ceramic method and mechanical milling. The characterisation of the Fe3O4 was done based on structural, magnetic, thermal stability, morphologic investigations. Experimental The initial equimolar mixture contained high purity Fe carbonyl (>99.5%) and Fe2O3 (>98%) (Alfa Aesar) powders, which have been used for the synthesis of Fe3O4. The Fe carbonyl and Fe2O3 powders have homogenised. The homogenous stoichiometric mixture was heat treated at 870 °C for 4 hours in argon atmosphere in order to obtain magnetite (Fe3O4). The as obtained Fe3O4 powders were processed in a high energy planetary ball-mill (Fritsch, Pulverisette 4) up to 240 minutes. The milling conditions were: ball diameter of 14 mm and a ball to powder ratio (BPR) of 20:1, vial rotation speed, ω=800 rpm, disc rotational speed, Ω=-400 rpm. The milling was done under argon atmosphere. The vials, and the milling balls were from tempered steel. For the structural investigation the X-ray diffraction technique has been used. The X-ray diffraction patterns have been recorded in the angular range of 2theta=20 – 110 degrees. The CoKα radiation was used (λ=1.7903) from an Inel Equinox 3000 diffractometer. For the calculus of the mean crystallite size and lattice strain the Williamson-Hall method was used. The particle size distribution and the d10, d50, d90 parameters of the particles were determined using a Laser Particle Size Analyzer, Fritsch Analysette 22 – Nanotec. The used measuring range was 10 nm – 100 μm.
The morphology of the samples was investigated by Jeol-JSM 5600 LV scanning electron microscope (SEM), using x10000 and x70000 magnifications. The Fourier transform infrared absorption spectra of the samples was determined using Jasco FTIR 6200 spectrometer. The scanned wavenumber range was 400-4000 cm-1. The measurements were done using KBr pellets technique. Before pressing, the samples were crushed and ground in an agate mortar, in order to obtain particles of micrometric sizes for a good quality spectra. All the measurements were done using the same KBr/sample weight ratio. By a Setaram Labsys apparatus, differential scanning calorimetry (DSC) analyses were recorded. The used reference was high purity alumina. The heating/cooling rate was 10 °C/min. All of the measurements were done in argon atmosphere in order to avoid the oxidation of the samples. Extraction sample method in a continuous magnetic field up to 8 T was used to record M=f(H) curves at 300 K. Saturation magnetisation was calculated from the derived M=f(1/H 2) curves. The spontaneous magnetisation has been obtained by extrapolation to zero field of the linear variation of magnetization obtained in magnetic field higher than 6 T. Results and discussion In figure 1 are presented the X-ray diffraction patterns of the starting materials, Fe carbonyl and hematite - Fe2O3, and for the Fe3O4 sample obtained by heat treatment at 870 °C for 4 h in argon atmosphere. The hkl indices are marked in the figure for all three samples according to JCPDS files no. 06-0696 for Fe, 33-0664 for Fe2O3 and 19-0629 for Fe3O4. As can be seen after the heat treatment, in the diffraction pattern of the heat treated sample are observed only the Bragg reflection of the iron ferrite (JCPDS file no. 19-0629. Thus indicating a complete reaction between the iron and hematite and the formation of a single phase. The reaction that occurs during heat treatment can be written as:
4 Fe2O3 Fe 3 Fe3O4
(1)
In the limit of detection of the X-ray diffraction technique the as obtained iron ferrite is pure. The simple ceramic synthesis route and easily accessible precursors that have been used are considered to be appropriate in order to obtain a single Fe3O4 phase.
2 theta (degree)
Figure 1. X-ray diffraction patterns of the starting materials, Fe carbonyl and Fe2O3, and for the Fe3O4 sample obtained by heat treatment at 870 °C for 4 h in argon atmosphere. The patterns have been vertically shifted for clarity.
Figure 2 presents X-ray diffraction patterns of the as obtained Fe3O4 sample (marked as 0 h MM), and for the Fe3O4 milled samples (1, 5, 15, 30, 60, 120, 180, 240 minutes). For milling times up to 15 minutes, in the X-ray patterns can be noticed only the peaks characteristic for the cubic spinel structure of the magnetite. In the X-ray pattern corresponding to the sample milled for 30 minutes a new peak at the Bragg diffraction angle at about 2theta=52,35 ° is found. This peak’s position matches with the iron’s most intense peak position according to JCPDS file no. 06-0696. This is the result of iron contamination during mechanical milling. The contamination with chemical elements that are in the composition of the materials from which are fabricated the vials and balls used for milling, can be encountered during high energy mechanical milling. Increasing the milling time up to 60 minutes leads to formation of a new phase, FeO – wüstite (JCPDS file no. 06-0615). This new phase is assumed to result from the reaction between the Fe resulted from contamination and ferrite. The most intense Bragg peak of the wüstite phase (200) can be observed at 2 theta = 50.5°, where it can be remarked as a shoulder of the Fe3O4 peak (400). In the figure 3 are presented the fit and deconvolution of the experimental X-ray diffraction patterns for the samples milled for 30 and 240 minutes. The reaction of the iron with the iron ferrite lead to the reduction of Fe3O4 and the chemical reaction that occurs can be written as:
Fe3O4 Fe 4 FeO
(2)
The formation of the FeO is most likely to appear if we take into account the fact that this oxide is the only one that possess just Fe2+ ions from the three simple (that have only Fe and O elements) oxides of iron: FeO (only Fe2+ ions), Fe3O4 ( both Fe2+ and Fe3+ ions) and Fe2O3 (only Fe3+ ions). The reduction of ferrites with the iron provided by contamination has been reported in the literature
[29] in the case of zinc ferrite - ZnFe2O4. Increasing the milling time leads to the preservation of the existing phases (Fe, FeO and Fe3O4) up to the final milling time (240 minutes). In the case of the samples milled for 120, 180 and 240 minutes in the diffraction patterns, are visible three Bragg peaks of FeO (the most intense ones). The structural disorder and distorted structure are generally induced by the mechanical milling processes. A cation inversion among tetrahedral and octahedral sites can occur. The contamination increases with milling time, and as a consequence the amount of the FeO increases, due to the continuous input of the iron from the milling bodies. This is proved by the increasing area of the FeO (200) peak in comparison with the areas of the Fe (110) and Fe 3O4 (311) peaks. The evolution of the ratio of the FeO (200) peak area, Fe (110) peak area and Fe3O4 (311) peak area/sum of areas of FeO (200), Fe (110) and Fe3O4 (311) peaks versus milling time is shown in the figure 4. It is well known that the amount of phase is related to the peak’s area relative to the peak’s area of the other phases presented in material. It can be observed that the area of FeO (200) peak increases continuously during increasing the milling time from 30 up to 240 minutes. In the same time, the area of the iron peak is almost constant and the area of ferrite peaks decreases. This proves the increase of the FeO amount and the decrease of the Fe3O4 amount.
2 theta (degree)
Figure 2. X-ray diffraction patterns of the as obtained Fe3O4 sample (marked as 0 h MM), and for the Fe3O4 milled samples (1, 5, 15, 30, 60, 120, 180, 240 minutes). The patterns have been vertically shifted for clarity.
47
48
49
50
51
2 theta (degree)
52
53
54
47
48
49
50
51
52
53
54
2 theta (degree)
a) b) Figure 3. Fit and deconvolution of the experimental X-ray diffraction patterns recorded for the: (a) sample milled for 30 minutes and b) sample milled for 240 minutes. In the figures are indicated: (1) experimental XRD points; (2) the best fit curve of the XRD experimental points; (3) peaks position, (4) difference between the experimental points and the corresponding fit. The corresponding diffraction lines are marked for each phase in the top part of the figures.
Figure 4. The evolution of the ratio of the FeO (200) peak area-noted as AFeO (200), Fe (110) peak area-noted as AFe (110) and Fe3O4 (311) peak area - noted as AFe3O4 (311)/sum of areas of FeO (200), Fe (110) and Fe3O4 (311) peaks - noted as AFeO (200)+AFe (110)+AFe3O4 (311) as a function of milling time.
a)
Milling time (minutes)
b) Figure 5. a) Williamson-Hall plots for all the milled Fe3O4 samples (1, 5, 15, 30, 60, 120, 180, 240 minutes) and b) evolution of the Fe3O4 mean crystallite size and lattice strain for milling times up to 240 minutes.
Another structural change that is indicated by the X-ray diffraction patterns is the peak broadening. This is due to the decrease of the crystallite size and the increase of the internal stress caused by severe deformation that occurs during mechanical milling procedure. In figure 2 are presented the Williamson-Hall plots (β·cosθ vs 4·sinθ) for all the milled samples and the evolution of the mean crystallite size and lattice strain versus milling time for the Fe3O4, which have been derived from the plots using the Williamson-Hall method. In the Williamson-Hall plots can be noticed the increase of the β·cosθ product and the change of the plots slope. This is translated in the crystallite decrease and the lattice strain change. Indeed, the crystallite size is decreasing down to 15 nm in the case of the sample milled for 240 minutes. The mean crystallite size variation presents two stages. In the first stage, the crystallites are reduced very quickly by the milling process. After only 5 minutes of milling the mean crystallite size is at about 126 nm. In the second stage a saturation of the crystallite reduction can be noticed from 60 minutes up to 240 minutes of milling. The lattice strain has a different evolution. It presents a large increase up to 30 minutes of milling. Prolonged milling times lead to the decrease of the lattice strain. The milling process induces stresses into the structure and stresses cumulate upon increasing the milling time. A part of this internal stresses represents the lattice strain. The decrease of the lattice strain after 30 minutes of milling can be explained by the formation of the new phase FeO. The energy stocked into the structure is assuming to be consumed during the reaction of the iron with the ferrite. On the other hand, the structure can release the accumulated strain also during the crystallites reduction. Although, it can be noticed that we have a limitation of the crystallites reduction but, in the same time a large decrease of the lattice strain. This suggests that the formation of the FeO has a much larger effect on the lattice strain as compared to the crystallites decrease and our assumption is that the energy stocked as lattice strain is in major part consumed during the reaction of wüstite formation. In figure 6 are presented the particle size distribution of the magnetite powder un-milled and milled 1, 5, 15, 30, 60, 120, 180 and 240 minutes. SS
1 min MM
Particle size ( μm)
Particle size ( μm)
5 min MM
15 min MM
Particle size (μm)
Particle size (μm)
30 min MM
15 min MM
Particle size (μm)
Particle size (μm)
120 min MM
60 min MM
Particle size (μm)
Particle size (μm)
180 min MM
240 min MM
Particle size (μm)
Particle size (μm)
Figure 6. Laser particle size distribution for the Fe3O4 samples milled for 0, 1, 5, 15, 30, 60, 120, 180 and 240 minutes. The Fe3O4 sample obtained by ceramic route reveals, as expected, a bimodal distribution of the size classes. This distribution is the result of different granulation of the starting powders. The hematite powder (Fe2O3) is very fine, below 4 μm, and the iron carbonyl powder is comprised in the 6 – 9 μm interval. A milling time of 1 minute increases the size of particles, due to the cold welding of the particles. The sample milled for 5 minutes presents a consistent decrease of the particle size and the particles becomes more homogeneous in size. This sample presents a Gaussian particles size distribution. The particles with sizes of hundreds of nanometers are found in the samples milled for up to 15 minutes. Also, in the case of the sample milled for 1 and 5 minutes, one can observe the fact that in the material are existing particles with the dimension less than 100 nm. In the case of this samples it can be observed that the particle size distribution is ending suddenly. This indicates that the apparatus does not distinguish between the different particles classes that have the dimension less than 100 nm. The milling time up to 15 minutes leads to the formation of nanoparticles and larger particles with nanocrystalline structure. One can remark that during milling the nanoparticles could weld one to each other, resulting in larger particles. During the measurement, the particles remain welded, so the ultrasonic bath is not sufficient for separating them. Somehow, this type of investigation overestimates the particles size. For the sample milled
30 minutes, the particle size distribution is multimodal. This is assigned to the presence of iron particles coming from contamination, and the formation of the new wüstite phase. Further increase of the milling time up to 240 minutes reveals once again a bimodal particle size distribution. The particle size distribution for the samples milled for 60, 120, 180 and 240 minutes are quite similar. The particles size tends to become homogenous although in material are coexisting three phases.
Figure 7. Evolution of the d10, d50, d90 parameters of the Fe3O4 as a function of milling times. The evolution of the d10, d50, d90 parameters is presented in figure 7. In the first stage of milling, up to 5 minutes, all the three parameters are decreasing. The decrease signifies the predominance of the fragmentation processes. This is due to the fact that the particles got shallow welded during the heat treatment, and the bonds are breaking easy. The second stage, from 5 minutes up to 60 minutes, reveals an increase of the d10, d50, d90 parameters. This increase is the effect of the preponderance of cold welding processes that occur during mechanical milling. In the final stage of milling, from 60 minutes up to 240 minutes, the dn parameters have variations that indicate a quasi-equilibrium between the cold welding and the fracturing processes. The d50 is increasing but in the same time the d90 is decreasing. The d10 parameter is almost constant from 30 minutes of milling up to 240 minutes of milling, indicating a constant amount of fine particles. The supposition that in the final stage of milling is a quasi-equilibrium between the cold welding and the fracturing processes, suggested by the evolutions of dn parameters, is sustained also by the Xray diffraction, which doesn’t indicate any phase change in the 60-240 minutes range. For milling times up to 60 minutes, this processes can be influenced by the changes that take place, such as iron contamination and the formation of the FeO phase. The assumption that the bigger particles have nanoparticles attached on their surface, and also the smaller one, are in fact agglomerations of nanoparticles is confirmed by the scanning electron microscopy images presented in figure 8. For each sample there are two images, at x10,000 and x70,000 magnifications.
0 min MM
5 min MM
30 min MM
60 min MM
240 min MM .
SEM Figure 8. images of the Fe3O4 samples mechanically milled for 0, 5, 30, 60, 240 minutes at x10000 and x70000 magnifications.
The SEM images reveal polyhedral irregular particles shape for the sample obtained by ceramic route. It can be also observed that there are no individual particles with the dimension of tens of micrometers as the particle size distribution indicated. The larger particles are formed by smaller particles that are welded to each other. Also the SEM images, reveal the presence of nanosized particles, and also larger particles up to 10 micrometers. This results are confirming the laser particle size analysis. Mechanical milling leads to the decrease of the particles size, but on the other hand it also leads to formation of bigger cold welded particles, as can be seen in the SEM images of the mechanically milled Fe3O4. For the samples milled for longer times, it is noticeable that the particles are in fact gatherings of nanosized particles. The milled Fe3O4 powder is formed from nanoparticles and bigger nanocrystalline Fe3O4 particles. The FTIR spectra of the Fe3O4 milled for 0, 1, 5, 15, 30, 60, 120, 180, 240 minutes are presented in figure 9. In the spectra of the un-milled sample are observed the vibrational bands characteristic for the Fe3O4 spinel structure at 573 and 386 cm-1. The vibrational band from 573 cm-1 assigned to the Fe-O group stretching, with the iron cations in tetrahedral sites. The other vibrational band, from 386 cm-1, is characteristic for the Fe-O group stretching, with the iron cations in octahedral sites. Besides these two vibrational bands, characteristic for iron ferrites, there are other three vibrational bands that are characteristic to H2O (O-H group bending and symmetrical stretching at 1636 and 3438 cm-1 respectively) and CO2 (C-O group asymmetrical stretching at 2367 cm-1). The presence of H2O and CO2 is due to the powder manipulation, which has been done in air and that led to the absorbance on the powder surface of atmospheric water and carbon dioxide. After only 5 minutes of milling in the spectra can be observed a new vibrational band at about 410 cm-1. This band is attributed to the Fe-O group stretching mode of wüstite. Thus indicating that the germs of this iron oxide are forming much earlier than it was proved by X-ray diffraction. This vibrational band is found also for all the samples milled more than 5 minutes. The intensity of this vibrational band increases with higher milling times, revealing an increase of the wüstite phase amount in the material as a result of the iron contamination. The FTIR investigation revealed the fact that, a part of the iron ferrite is reduced at very shorter duration of high energy mechanical milling and consequent it can be assumed that the contamination with iron occurs in the incipient stage of milling. The Fe-O group stretching mode of wüstite band is more and more prominent upon increasing the milling time indicating the increase of the amount of this phase. This comes to confirm the results obtained by X-ray diffraction investigated.
Figure 9. FTIR spectra of the Fe3O4 milled for 0, 1, 5, 15, 30, 60, 120, 180, 240 minutes. Measured interval 400-4000 cm-1. The heating/cooling DSC curves in the temperature range of 150-700°C, for the Fe3O4 samples milled up to 240 de minutes, are presented in figure 10. For the sample obtained by ceramic route, on the heating curve an endothermic event is noticeable at about 326 °C. This endothermic event is assigned to the superficial oxidation of the Fe3O4. This oxidation leads to a partial transformation of the Fe3O4 phase in γ-Fe2O3, which is isomorph from the crystallographic point of view. This oxidation seems to take place at higher temperatures upon increasing the milling time and it becomes less prominent. This behaviour can be influenced by the appearance of the two new phases, Fe and FeO, upon increasing the milling time and structural changes such as crystallite reduction and internal stresses. It can be assumed that due to the small oxygen content first is forming oxides less rich in oxygen, FeO (iron oxidation) and Fe3O4 (FeO oxidation) and therefore only at a slightly higher temperature the oxidation of Fe3O4 take place. For high milling times (after 60 minutes) this endothermic event it is not visible anymore. It is masked by the exothermic phenomenon assigned to the stresses release which becomes more intense while it becomes less prominent. Also, for longer milling times, on the heating curve a broad exothermic peak is observed in the temperature range of 150 °C up to 360-370 °C. This thermal event is characteristic for milled samples and represents the release of internal stresses induced by mechanical milling. This peak overlaps the superficial oxidation peak previously mentioned for the samples milled for prolonged durations. For every sample around 570 °C, an endothermic event can be noticed. This peak is attributed to the ferrimagnetic/paramagnetic transition and represents the Néel temperature. In confirmation of this assignment, on the cooling curve there is an equivalent exothermic event at about the same temperature. This temperature has a decreasing tendency upon increasing the milling time, and this evolution is presented in figure 11. This decrease is influenced by several factors. Among those factors structural disorder, distorted structure and impurities determined lower Néel temperature. All these three factors have been highlighted by X-ray diffraction and FTIR spectroscopy. The decrease of the Néel temperature
according to DSC investigation is from about 570 ºC for the un-milled sample up to 555 ºC in the case of the sample milled for 240 minutes. As compared to the literature, we have obtained a smaller Néel temperature even for the un-milled sample (570 ºC compared to 585 ºC). This is attributed to the method of determination and different routes of synthesis.
a)
b)
Figure 10. DSC curves for the Fe3O4 samples milled for 0, 1, 5, 15, 30, 60, 120, 180 and 240: a) heating curves and b) cooling curves.
Figure 11. Evolution of Néel temperature for the Fe3O4 versus milling time. The Néel temperature has been derived from the DSC measurements.
Figure 12 presents first magnetisation curves registered at 300 K, M=f(1/H2) plots and evolution of spontaneous and saturation magnetisation as a function of milling time of Fe3O4 samples. The first magnetisation curve of the un-milled magnetite is characteristic for the ferrimagnetic material.
The ferrites are in generally ferrimagnetic and so the as obtained iron ferrite is considered as being ferrimagnetic as well. It can be observed that by increasing the milling time the magnetisation is decreasing. The ceramic sample, and the samples milled up to 5 minutes have relatively the same behaviour. The samples milled for more than 5 minutes reveal a more prominent decrease. Also, the samples milled for duration of at least 30 minutes have the tendency to saturate more difficult. This type of behaviour is typical for the ferrites synthetized by mechanosynthesis [4].The diminution of particle size during mechanical milling is leading to an increase of the surface area of the powder. It must be taken into account also, that the larger particles are formed by coldwelded fine particles. Generally large surface area of the particles is leading to the spin canted effect. This effect explains the difficulty to saturate of the samples milled for prolonged periods. The magnetic susceptibility is defined as the ratio between the magnetisation and the magnetic field that is applied. The decrease of the magnetic susceptibility and the presence of very small particles suggests the presence of some superparamagnetic particles. Nanometric particles have been identified by SEM investigation and have been also highlighted by laser particles size analysis. In the high magnetic fields region the M=f(1/H2) plots resulting from the first magnetisation curves reveal a good linearity for all the Fe3O4 samples. The saturation magnetisation calculated from this plots decrease drastically up to the final milling time, from about 92.2 up to 69.3 A·m2/kg. This decrease is assigned to the structural changes and changes of the composition which is caused by the iron contamination highlighted earlier by FTIR spectroscopy and X-ray diffraction. Indeed, if we made a simple calculus on the theoretical magnetic moments involved in the reaction 2, it results that by the reaction of one Fe atom with one Fe3O4 molecule a drop of magnetisation with 6.2 µB occurs. Although, before reaction the amount of Fe provided by contamination should lead to a small increase of the magnetisation. This increase is probably counterbalanced by the reaction of the iron with the ferrite in the incipient stage of milling. In the calculus a net magnetic moment of 4 µB for Fe3O4, 2.2 µB for Fe and 0 µB for FeO have been taken into account. The spontaneous magnetisation versus milling time has also similar behaviour. The decrease of spontaneous magnetisation is more accentuated as compared to the saturation magnetisation. The increase of the difference between saturation and spontaneous magnetisation can be explained by the decrease of the magnetic susceptibility of the material. The response of the magnetic material to the applied field is influenced by the distorted structure and the canted spins.
a)
b)
c)
Figure 12. a) First magnetisation curves registered at 300 K, b) M=f(1/H2) plots and c) evolution of spontaneous and saturation magnetisation as a function of milling time of Fe3O4 samples.
Conclusions Iron ferrite, magnetite, was successfully obtained in nanocrystalline state by a combined method which consists in heat treatment, for the reaction of the accessible precursors (iron and hematite), and mechanical milling, for achieving nanocrystalline state and nanosized particles. This synthesizing method was chosen due to the low cost of the starting mixture’s high purity powders. The level of impurity of the magnetite powder obtained from the precursors is below the limit of the X-ray diffraction. The characterisation of the nanocrystalline/nanosized revealed the fact that small milling times are favourable for the formation of the magnetite, Fe3O4 in nanocrystalline/nanosized state. Longer milling times lead to powder contamination with iron, and also the reaction between the magnetite and the iron which leads to the formation of an undesirable wüstite phase. Iron is the main constituent of the material from which the milling bodies were manufactured. The FTIR spectroscopy showed the iron contamination is starting at shorter milling times than the one identified from the X-ray diffraction. For duration up to 4 hours, the mean crystallite size is about 15 nm, while the lattice strain is almost null.
Longer milling times lead to a lower magnetisation value of the material due to the structural disorder, internal stresses and the spin canted effect. The magnetic behaviour of the material, and the small particle size classes indicate the presence of superparamagnetic particles. Acknowledgement This paper was supported by the Post-Doctoral Programme POSDRU/159/1.5/S/137516, project co-funded from European Social Fund through the Human Resources Sectorial Operational Program 2007-2013.
References [1] T.F. Marinca, I. Chicinas, O. Isnard, Structural and magnetic properties of the copper ferrite obtained by reactive milling and heat treatment, Ceramics International, 39 (2013) 4179– 4186. [2] A. Milutinovic, Z. Lazarevic, C. Jovalekic, I. Kuryliszyn-Kudelska, M. Romcevic, S. Kostic, N. Romcevic, The cation inversion and magnetisation in nanopowder zinc ferrite, Materials Research Bulletin 48 (2013) 4759–4768. [3] C. Murugesan, M. Perumal, G. Chandrasekaran, Structural, dielectric and magnetic properties of cobalt ferrite prepared using auto combustion and ceramic route obtained by soft mechanochemical processing, Physica B 448 (2014) 53–56. [4] M. Siddique, N.M. Butt, Effect of particle size on degree of inversion in ferrites investigated by Mossbauer spectroscopy, Physica B 405 (2010) 4211–4215 . [5] S. Choudhury, M. Sinha, H. Dutta, M.K. Mandal, S.K. Pradhan, A.K. Meikap, Activation behavior and dielectric relaxation of nanocrystalline zinc ferrite, Materials Research Bulletin 60 (2014) 446–452. [6] E. Pervaiz, I.H. Gul, High frequency AC response, DC resistivity and magnetic studies of holmium substituted Ni-ferrite: A novel electromagnetic material, Journal of Magnetism and Magnetic Materials 349 (2014) 27-34. [7] W. Wang, C. Zang, Q. Jiao, Synthesis, structure and electromagnetic properties of Mn-Zn ferrite by sol-gel combustion technique, Journal of Magnetism and Materials 349 (2014) 116120. [8] V.Babayan, N.E. Kazantseva, I. Sapurina, R. Moucka, J. Stejskal, P. Saha, Increasing the high-frequency magnetic permeability of MnZn ferrite in polyaniline composites by incorporating silver, Journal of Magnetism and Magnetic Materials, 333 (2013) 30-38. [9] T.F. Marinca, I. Chicinas, O. Isnard, Synthesis, structural and magnetic characterization of nanocrystalline CuFe2O4 as obtained by a combined method reactive milling, heat treatment and ball milling, Ceramics International 38 (2012) 1951-1957. [10] B.D. Cullity, C.D. Graham, Introduction to magnetic materials, second edition, IEEE Press&Wiley, 2009, New Jersey, USA, ISBN: 978-0-471-47741-9. [11] A. Goldman, Modern Ferrite Technology, Second edition, Springer, 2006, Pittsburgh, USA, ISBN: 0-387-28151-74.
[12] S. Wang, Y. Chiang, Y. Hsu, C. Chen, Effects of additives on the loss characteristics of MnZn ferrite, Journal of Magnetism and Magnetic Materials 365 (2014) 119-125. [13] K. Gheisari, S. Javadpour, H. Shokrollahi, B. Hashemi, Magnetic losses of the soft magnetic composites consisting of iron and Ni-Zn ferrite, Journal of Magnetism and Mangetic Materials 320 (2008) 1544-1548. [14] R. Ali, M. A. Khan, A. Mahmood, A. H. Chughtai, A. Sultan, M. Shahid, M. Ishaq, M. F. Warsi, Structural, magnetic and dielectric behaviour of Mg1-xCaxNiyFe2-yO4 nano-ferrites synthesized by the micro-emulsion method, Ceramics International 40 (2014) 3841–3846. [15] A. Arab, M.R. Mardaneh, M.H. Yousefi, Investigation of magnetic properties of MnZnsubstituted strontium ferrite nanopowders prepared via conventional ceramic technique followed by a high energy ball milling, Journal of Magnetism and Magnetic Materials 374 (2015) 80–84. [16] A.S. Ponce, E.F. Chagas, R.J. Prado, C.H.M. Fernandes, A.J. Terezo, E. Baggio-Saitovitch, High coercivity induced by mechanical milling in cobalt ferrite powders, Journal of Magnetism and Magnetic Materials 344 (2013) 182–187. [17] Y. Cedeno-Mattei, O. Perales-Perez, O.N.C. Uwakweh, Effect of high-energy ball milling time on structural and magnetic properties of nanocrystalline cobalt ferrite powders, Journal of Magnetism and Magnetic Materials 341 (2013) 17–24. [18] S. Wu, A. Sun, W. Xu, Q. Zhang, F. Zhai, P. Logan, A. Volinsky, Iron-based soft magnetic composites with Mn–Zn ferrite nanoparticles coating obtained by sol–gel method, Journal of Magnetism and Magnetic Materials 324 (2012) 3899–3905. [19] M. Sajjia, M. Oubaha, M. Hasanuzzaman, A.G. Olabi, Developments of cobalt ferrite nanoparticles prepared by the sol–gel process, Ceramics International 40 (2014) 1147–1154. [20] S.A. Seyyed Ebrahimi, S.M. Masoudpanah, H. Amiri, M. Yousefzadeh, Magnetic properties of MnZn ferrite nanoparticles obtained by SHS and sol-gel autocombustion techniques, Ceramics International 40 (2014) 6713-6718. [21] A. Xia, C. Jin, D. Du, G. Zhu, Comparative study of structural and magnetic properties of NiZnCu ferrite powders prepared via chemical coprecipitation method with different coprecipitators, Journal of Magnetism and Magnetic Materials 323 (2011) 1682–1685. [22] S.M. Olhero, D. Soma, V.S. Amaral, T.W. Button, F.J. Alves, J.M.F. Ferreira, Coprecipitation of a Ni–Zn ferrite precursor powder: Effects of heat treatment conditions and deagglomeration on the structure and magnetic properties, Journal of the European Ceramic Society 32 (2012) 2469–2476. [23] Y. Zhang, Z. Yang, D. Yin, Y. Liu, C. Fei, R. Xiong, J. Shi, G. Yan, Composition and magnetic properties of cobalt ferrite nano-particles prepared by the co-precipitation method. [24] I. Sharifi, H. Shokrollahi, S. Amiri, Ferrite-based magnetic nanofluids used in hyperthermia applications, Journal of Magnetism and Magnetic Materials 324 (2012) 903–915. [25] N. Varalaxmi, K.V. Sivakumar, Structural and dielectric studies of magnesium substituted NiCuZn ferrites for microinductor applications, Materials Science and Engineering B 184 (2014) 88–97. [26] J.L. Mattei, E. Le Guen, A. Chevalier, A.C. Tarot, Experimental determination of magnetocrystalline anisotropy constants and saturation magnetostriction constants of NiZn and NiZnCo ferrites intended to be used for antennas miniaturisation, Journal of Magnetism and Magnetic Materials 374 (2015) 762–768.
[27] C. Suryanarayana, Mechanical Alloying and Milling, Marcel Dekker, New York, U.S.A., 2004, ISBN: 0-8247-4103-X. [28] G. Herzer, Modern soft magnets: Amorphous and nanocrystalline materials, Acta Materalia 61 (2013) 718-734. [29] T. Verdier, V. Nachbaur, M. Jean, Mechanosynthesis of zinc ferrite in harened steel vials: Influence of ZnO on the appearance of Fe(II), Journal of Solid State Chemistry 178 (2005) 3243–3250.
