Effect of a SiO2 coating on the magnetic properties of ...

Home

Search

Collections

Journals

About

Contact us

My IOPscience

Effect of a SiO2 coating on the magnetic properties of Fe3O4 nanoparticles

This article has been downloaded from IOPscience. Please scroll down to see the full text article. 2012 J. Phys.: Condens. Matter 24 266007 (http://iopscience.iop.org/0953-8984/24/26/266007) View the table of contents for this issue, or go to the journal homepage for more

Download details: IP Address: 130.206.165.223 The article was downloaded on 18/06/2012 at 16:22

Please note that terms and conditions apply.

IOP PUBLISHING

JOURNAL OF PHYSICS: CONDENSED MATTER

J. Phys.: Condens. Matter 24 (2012) 266007 (6pp)

doi:10.1088/0953-8984/24/26/266007

Effect of a SiO2 coating on the magnetic properties of Fe3O4 nanoparticles S Larumbe, C G´omez-Polo, J I P´erez-Landaz´abal and J M Pastor Departamento de Fisica, Universidad P´ublica de Navarra, Campus de Arrosad´ıa, 31006 Pamplona, Spain E-mail: [email protected]

Received 16 December 2011, in final form 26 April 2012 Published 14 June 2012 Online at stacks.iop.org/JPhysCM/24/266007 Abstract In this work the effect of a SiO2 coating on the magnetic properties of Fe3 O4 nanoparticles obtained by the sol–gel method is analyzed. Two sets of samples were prepared: Fe3 O4 nanoparticles and Fe3 O4 @SiO2 core–shell composites. The samples display the characteristic spinel structure associated with the magnetite Fe3 O4 phase, with the majority of grain sizes around 5–10 nm. At room temperature the nanoparticles show the characteristic superparamagnetic behavior with mean blocking temperatures around 160 and 120 K for Fe3 O4 and Fe3 O4 @SiO2 , respectively. The main effect of the SiO2 coating is reflected in the temperature dependence of the high field magnetization (µ0 H = 6 T), i.e. deviations from the Bloch law at low temperatures (T < 20 K). Such deviations, enhanced by the introduction of the SiO2 coating, are associated with the occurrence of surface spin disordered effects. The induction heating effects (magnetic hyperthermia) are analyzed under the application of an AC magnetic field. Maximum specific absorption rate (SAR) values around 1.5 W g−1 were achieved for the Fe3 O4 nanoparticles. A significant decrease (around 26%) is found in the SAR values of the SiO2 coated nanocomposite. The different heating response is analyzed in terms of the decrease of the effective nanoparticle magnetization in the Fe3 O4 @SiO2 core–shell composites at room temperature. (Some figures may appear in colour only in the online journal)

1. Introduction

biocompatible organic or inorganic coatings are employed (polymers, dextran, chitosan or silica [2, 4, 5]). Silica coating is relatively easy to prepare through a sol–gel method using metal alkoxides as precursors. This synthesis procedure is able to synthesize in one step the magnetite nanoparticles with the silica coating and optimize the size of nanoparticles and their structure through the synthesis conditions. However, the immersion of the magnetic nanoparticles within the silica matrix is usually associated with a decrease in the net magnetization [6, 7]. This effect is correlated with the occurrence of surface spin disorder induced by the silica coating. Such a core–shell structure modifies the magnetic response of the nanoparticles and determines their magnetic characteristics. In this work, the effect of the silica coating on the magnetic properties of magnetite nanoparticles is studied. Two samples of Fe3 O4 and Fe3 O4 @SiO2 were synthesized by a sol–gel method. The magnetic properties of both samples

Ferrite nanoparticles have been extensively explored for different applications, for example in magnetic hyperthermia due to their biocompatibility and for their special magnetic properties as nanosized systems. Different synthesis methods are employed in the preparation of magnetite nanoparticles, including chemical methods (sol–gel, coprecipitation, hydrothermal. . . ) or physical methods (laser ablation, mechanical ball milling, sputtering) [1–3]. For biomedical applications, the nanoparticle systems should be biocompatible and easily eliminated by the organism. In this sense, it is of great importance to prevent their agglomeration by reducing the particle diameter to the superparamagnetic range [1]. Hence, two aspects are relevant: the size of the magnetic nanoparticles and the magnetic interaction between them. Furthermore, to combine the biocompatible characteristics and reduce particle agglomeration, different 0953-8984/12/266007+06$33.00

1

c 2012 IOP Publishing Ltd Printed in the UK & the USA

J. Phys.: Condens. Matter 24 (2012) 266007

S Larumbe et al

were analyzed and their self-heating features (magnetic hyperthermia) were evaluated in terms of the values of the specific absorption rate (SAR).

2. Experimental procedure All reagents supplied by Sigma Aldrich were of analytical grade and were used without any further purification. Ferrite nanoparticles were synthesized via an autocombustion sol–gel method. Citric acid was used as a fuel to promote the crystallization of the spinel through its combustion and ferric nitrate was employed as the chemical precursor salt. After the hydrolysis of the nitrate in distilled water with a hydrolysis rate of 130, citric acid with a molar ratio of 1:2 and 1:1 (ferric nitrate:citric acid) for Fe3 O4 and Fe3 O4 @SiO2 nanoparticles, respectively, was added as fuel. These molar ratios were found to be the optimum values to obtain a single spinel phase. Afterwards, an ethanolic solution of the silica precursor, TEOS (tetraethyl orthosilicate), was dropped into the dark brown solution to give the SiO2 composite. Taking into account the relative molar concentrations (ferric nitrate:TEOS) employed during the process, the SiO2 content was approximately 23% by weight with respect to the magnetite weight. The reaction was performed in an acid medium (pH ≈ 1) to reduce the nanoparticle diameter (decrease in the condensation rate). The calcination temperature leading to the decomposition of the organic matrix was determined by thermogravimetric analysis (HI-RES 2950 TA Instruments). The dark brown as-synthesized gel was calcined in an inert atmosphere to prevent the oxidation of the sample to maghemite or hematite. The structure of the calcined samples was analyzed by x-ray diffraction (Siemens D-5000) with Cu Kα ˚ and the mean diameter was estimated radiation (1.5418 A) through the Debye–Scherrer equation. Transmission electron microscopy (TEM; JEM 2100 HT at the ICTS Centro Nacional de Microscop´ıa Electr´onica UCM, Spain) confirms the nanometric size of the calcined nanoparticles. Magnetic measurements were carried out with a SQUID magnetometer (Quantum Design MPMS XL7). The induction heating curves under an AC field (magnetic hyperthermia) were measured with a home-made set up, composed of a water refrigerated coil with six turns (N = 100 turns m−1 ) connected to a 2 kW RF power amplifier (Electronic and Innovation, model 1240L). The temperature rise of the nanoparticles (powder) was measured with a fiber optic thermometer (Neooptix, model T1) under an AC magnetic field (amplitude 170–340 Oe, frequency 340 kHz).

Figure 1. TG thermograms (—) and derivative curves (◦) for (a) Fe3 O4 and (b) Fe3 O4 @SiO2 gels.

the derivative curves a deeper analysis about the effect of the silica matrix in the decomposition temperatures can be performed. Both samples display a peak below 200 ◦ C related to the evaporation of solvents. The second peaks observed at 238 ◦ C and 218 ◦ C for Fe3 O4 and Fe3 O4 @SiO2 , respectively, correspond to the oxidation of citric acid by nitrates, whose heat combustion leads to the crystallization of the spinel phase [8, 9]. The third peak at 373 ◦ C and 334 ◦ C (Fe3 O4 and Fe3 O4 @SiO2 , respectively) is due to the decomposition of the excess of citric acid added to the initial solution, taking into account that the molar ratio between the citric and iron salts was 2:1. Finally, the peak at 635 ◦ C only observed in the precursor gel of the Fe3 O4 sample would correspond to the dehydroxylation of the OH groups located at the surface of the nanoparticles [10]. In the Fe3 O4 @SiO2 nanoparticles, the covalent interaction between the Si atoms and the magnetite surface decreases the number of surface hydroxyl groups at the surface, giving rise to a decrease in the final decomposition temperature at which all the hydroxyl groups disappear [10]. Thus, according to the present thermogravimetric characterization (see figure 1), optimum calcination temperatures (complete decomposition of the organic matrix) of 650 ◦ C and 400 ◦ C are obtained for Fe3 O4 and Fe3 O4 @SiO2 , respectively. Accordingly, the as-synthesized gels were calcined at the indicated temperatures for 2 h. In order to prevent the

3. Results and discussion Firstly, in order to estimate the optimum calcination conditions associated with the decomposition temperature of the solvents and the organic matrix, the thermograms (TG) of the initial gels were analyzed employing heating and Ar flow rates of 20 ◦ C min−1 and 60 ml min−1 , respectively. Figure 1 shows the TG (weight loss) and the derivative curves of the gels for both nanoparticle systems. Through 2

J. Phys.: Condens. Matter 24 (2012) 266007

S Larumbe et al

At room temperature both samples display the characteristic anhysteretic behavior of superparamagnetic nanoparticles. The decrease of temperature promotes a noticeable increase in the coercivity, HC , as a result of the blocking of the nanoparticle magnetization (see figure 4(b)). Figure 5 displays the temperature dependence of HC for both nanoparticles. The coercive field displays the characteristic temperature decay of superparamagnetic systems. In fact, a mean estimation of the blocking temperature, TB , can be performed through the fitting of the temperature dependence of HC to the T 1/2 law of uniaxial non-interacting single domain particles [12]: "  1/2 # T HC = HK 1 − (1) TB

Figure 2. X-ray diffraction patterns for the calcined samples.

where HK = 2Keff /MS , with Keff the effective anisotropy constant and MS the saturation magnetization. The solid lines in figure 5 represent the data fitting according equation (1) with the following fitting parameters: (i) Fe3 O4 , HK = (670 ± 50) Oe, TB = (160 ± 20 K); (ii) Fe3 O4 @SiO2 , HK = (530 ± 40) Oe, TB = (120 ± 10 K). The slight increase in TB for the Fe3 O4 sample would correspond to the higher value of the mean nanoparticle size with respect to the Fe3 O4 @SiO2 sample. However, a clear dispersion in the coercivity fitting for both nanoparticles for measuring temperatures above the estimated TB should be noticed. This behavior could be explained in terms of the contribution of the nanoparticles with higher mean size. However, it should be kept in mind that equation (3) is strictly valid for uniaxial non-interacting single domain particles. Thus, the occurrence of magnetic (dipolar) interactions would also contribute to the detected behavior. In order to analyze this effect in further detail, the zero field cool–field cooled (ZFC–FC) magnetization curves were analyzed under an applied magnetic field of 5 mT. As figure 6 shows, the ZFC–FC magnetization curves do not display the characteristic features of a defined superparamagnetic blocking temperature. The irreversible behavior found for temperatures above the estimated TB should be associated with the magnetization contribution of the larger blocked nanoparticles. In fact, some features of the Verwey transition are detected around TV ≈ 120 K in the Fe3 O4 sample with higher mean nanoparticle size [13]. Additionally, the occurrence of strong interactions between the calcined nanoparticles would also contribute to the anomalous shape of the ZFC–FC curves. However, the main effect of the silica coating is detected in the temperature dependence of the high field magnetization (measured at µ0 H = 6 T). As figure 7 shows, at low temperatures the Bloch law, MS (T) = MS (0)(1 − BT α ), with MS (0) the saturation magnetization at 0 K and B the Bloch constant, is not followed and the experimental data can be suitably fitted through the introduction of an additional exponential term [14, 15]:

oxidation of the sample to maghemite or hematite an inert atmosphere was employed during the calcination procedure. Figure 2 shows the x-ray diffraction patterns of the calcined Fe3 O4 and Fe3 O4 @SiO2 nanoparticles. A single spinel phase (PDF card number 01-089-0691 from the PDFWIN database), with lattice parameters close to the ˚ is observed in both reported bulk magnetite (a = 8.39 A) ˚ for Fe3 O4 calcined nanoparticles: a = 8.39 and 8.40 A and Fe3 O4 @SiO2 , respectively. Lattice parameters were estimated using the Bragg law reflection peaks. A linear decrease in the lattice parameter of the spinel is reported with the ˚ (magnetite: oxygen vacancies (δ; Fe3(1−δ) O4−δ ) from 8.39 A ˚ Fe3 O4 ) to 8.35 A; maghemite: Fe2 O3 ) [11]. This result confirms the Fe3 O4 composition of the calcined nanoparticles and therefore the effectiveness of the heating procedure in an inert atmosphere to prevent the oxidation of the samples. On the other hand, the large width of the diffraction peaks indicates the nanometric size of the samples. Mean values of crystallite size, d, of 12 and 11 nm were obtained for Fe3 O4 and Fe3 O4 @SiO2 , respectively, employing the kλ Debye–Scherrer equation, d = β cos θ , where k = 0.9, λ is the ˚ β is full width at wavelength of the Cu Kα line (1.5418 A), half maximum and θ is the diffraction angle corrected with the instrumental width (βinst ); β = βexp − βinst ). In order to check the morphology and mean crystallite size, the calcined nanoparticles were analyzed by TEM. Figure 3 shows the TEM micrographs for the Fe3 O4 and Fe3 O4 @SiO2 nanoparticles (figures 3(a) and (c), respectively). As can be seen, low particle size dispersion is detected with mean sizes close to the estimated values by x-ray diffractometry (d = 5 and 7.5 nm for Fe3 O4 and Fe3 O4 –SiO2 , respectively). Thus, the close match between the estimated x-ray and TEM sizes indicates the single crystalline structure of the nanoparticles. Energy dispersive x-ray analysis (EDX) confirms the presence of the SiO2 in the Fe3 O4 @SiO2 nanoparticles (see figure 3(d)). With respect to the magnetic behavior of the samples, the hysteresis loops were analyzed in the temperature range from 5 to 300 K using a maximum applied magnetic field, µ0 H = 6 T. As an example, figures 4(a) and (b) display the hysteresis loops of the calcined nanoparticles at high (300 K) and low (5 K) temperatures, respectively.

( −T T )

MS (T) = MS (0)[(1 − BT α ) + A0 e

f

].

(2)

This deviation is explained by the presence of spin surface disordered effects and the occurrence in the nanoparticles of a core–shell structure. Thus, the spins 3

J. Phys.: Condens. Matter 24 (2012) 266007

S Larumbe et al

Figure 3. TEM images for (a) Fe3 O4 and (c) Fe3 O4 @SiO2 nanoparticles. EDX analyses are shown in (b) and (d) images.

Figure 5. Coercive field, HC , versus temperature, T.

at the surface will display a disordered state mainly due to the broken exchange bonds at the nanoparticle surface (shell) area [16–20]. Accordingly, the constant A0 in equation (2) would represent a measurement of the fraction of the disordered surface and Tf the characteristic freezing temperature below which the deviations are observed. Table 1 displays the obtained fitting parameters for both samples. It should be remarked that the estimated saturation magnetization MS (0) is below the reported value in bulk magnetite (MS (0) = 92 emu g−1 ) [21]. Such a decrease is inherent in nanoparticle systems and is correlated to the existence of spin disordered effects [17]. Moreover, the best fitting is obtained in both samples for α = 2,

Figure 4. Hysteresis loops (M–H) at (a) 300 K and (b) 5 K for both nanoparticles.

4

J. Phys.: Condens. Matter 24 (2012) 266007

S Larumbe et al

Figure 8. Temperature rise, 1T, versus time, t, under an AC magnetic field (initial temperature 18 ◦ C).

Figure 6. ZFC–FC magnetization (M) curves (applied magnetic field 5 mT).

and subtracting the 23% weight of the SiO2 coating, the saturation magnetization in this sample would reach a value of 48 emu g−1 . This value is still far below the estimated magnetization in the Fe3 O4 system (MS (0) = 72 emu g−1 ; see table 1). With respect to the induction heating effects (magnetic hyperthermia), figure 8 shows the temperature rise, 1T, versus time, t, in both nanoparticle systems under the application of an AC magnetic field (amplitude, HAC = 340 Oe and frequency, f = 340 kHz). The samples were in powder form (negligible Brownian contribution) and the 1T versus t curves were registered five times in order to estimate the mean value and its dispersion. The SAR was calculated through the initial slope of the heating curve:

Figure 7. Temperature (T) dependence of the high field magnetization M (applied magnetic field 6 T).

1T SARFe3 O4 = cFe3 O4 P 1t ci mi 1T SARFe3 O4 @SiO2 = i mFe3 O4 1t

Table 1. Parameters obtained from the fitting to the modified Bloch’s law (see equation (2)). Sample Fe3 O4 Fe3 O4 @SiO2

MS (0) (emu g−1 ) 72 37

B×106 (K−2 ) 1.1 3.2

Tf (K)

A0 3.3 × 10 0.116

−3

(3)

with ci the heat capacity of each component (magnetite 0.937 J g−1 K−1 [23]; silica 0.713 J g−1 K−1 [24]), 1T/1t the initial slope of the heating curve and mFe3 O4 the mass of magnetite in the samples [25, 26]. For the Fe3 O4 @SiO2 nanoparticles the average heat capacity is calculated taking into account the relative mass of silica in the sample (23%). Thus, SAR values of 1.5 ± 0.1 and 1.08 ± 0.04 W g−1 were obtained, respectively, for Fe3 O4 and Fe3 O4 @SiO2 nanoparticles. However, the spin disorder surface effects and the fact that not all the Fe3 O4 nanoparticles mass contribute to the magnetic heating around room temperature should be taken into account. In fact, considering the high field magnetization values at 300 K (65 and 32 emu g−1 for Fe3 O4 and Fe3 O4 @SiO2 nanoparticles, respectively) and the mass correction of the silica in the Fe3 O4 @SiO2 nanoparticles, just a 67% of the mass of Fe3 O4 nanoparticles in this silica coated system would be magnetically active for the heating process at room temperature. If this mass correction is introduced in the SAR estimation through equation (3), values around 1.6 W g−1 are obtained for the Fe3 O4 @SiO2 sample.

8 15

indicating a faster decrease of magnetization with temperature than in the bulk state. This increase in the α exponent has been previously reported in other ferrite nanoparticles and is associated with finite size effects and a lack of magnetic coordination at the surface [14, 22]. With respect to the disordered surface contribution, as figure 7 shows, the silica coating gives rise to a clear enhancement of the low temperature magnetization deviations. This enhancement is directly reflected in the highest value of the A0 parameter for the magnetite nanoparticles dispersed in the silica matrix (see table 1) [6, 7]. Besides this, the reduction in magnetization with respect to the bulk value is also enhanced with the introduction of the silica coating. The surface spin disorder contribution is enhanced for thicker silica shells, giving rise to an increase in the effective anisotropy constant of the nanoparticles [6, 7]. Taking the estimated value of MS (0) 5

J. Phys.: Condens. Matter 24 (2012) 266007

S Larumbe et al

field indicates the main contribution of the N´eel relaxation to the heating process.

Acknowledgment This work was has been performed within the framework of the project MET-NANOEFA17/08 (POCTEFA).

References [1] Tartaj P, Puerto Morales M, Veintemillas-Verdaguer S, Gonz´alez T and Carre˜no C J 2003 J. Phys. D: Appl. Phys. 36 R182 [2] Gupta A K and Gupta M 2005 Biomaterials 26 3995 [3] L´evy M, Wilhelm C, Siaugue J M, Horner O, Bacri J C and Gazeau F 2008 J. Phys.: Condens. Matter 20 204133 [4] Frimpong R A and Hilt J Z 2011 Nanomedicine 5 1401 [5] Thakur M, De K, Giri S, Si S, Kotal A and Mandal T K 2006 J. Phys.: Condens. Matter 18 9093 [6] Coskun M, Korkmaz M, Firat T, Jaffari G H and Shah S I 2010 J. Appl. Phys. 107 09B523 [7] Mitra S, Mandal K and Kumar P A 2006 J. Magn. Magn. Mater. 306 254 [8] Chen D H and He X R 2001 Mater. Res. Bull. 36 1369 [9] Srivastave S, Hubey C and Ogha A K 2009 Mater. Chem. Phys. 118 174 [10] Guo Z, Lei K, Li Y, Ng H W, Prikhodko S and Hahn H T 2008 Compos. Sci. Technol. 68 1513 [11] Belin T, Guigue-Millot N, Caillot T, Aymes D and Niepce J C 2001 J. Solid State Chem. 163 459 [12] Battle X, Garc´ıa del Muro M, Tejada J, Pfeiffer H, G¨oand P and Sin E 1993 J. Appl. Phys. 74 3333 [13] Goya G F, Berqu´o T S, Fonseca F C and Morales M P 2003 J. Appl. Phys. 94 3520 [14] V´azquez-V´azquez C, L´opez-Quintela M A, Buj´an-N´un˜ ez M C and Rivas J 2011 J. Nanopart. Res. 13 1673 [15] Aquino R, Depeyrot J, Sousa M H, Tourinho F A, Dubois E and Perzynski R 2005 Phys. Rev. B 72 184435 [16] Suzuki M, Fullem S I and Suzuki I S 2009 Phys. Rev. B 79 024418 [17] Cotica L F, Santos I A, Girotto E M, Ferri E V and Coelho A A 2010 J. Appl. Phys. 108 064325 [18] Berkowitz A E, Kodama R H, Makhlouf S A, Parker F T, Spada F E, McNiff E J and Foner S Jr 1999 J. Magn. Magn. Mater. 196–197 591 [19] Kodama R H 1999 J. Magn. Magn. Mater. 200 359 [20] Kodama R H, Berkowitz A E, McNiff E J and Foner S Jr 1996 Phys. Rev. Lett. 77 394 [21] Verg´es M A, Costo R, Roca A G, Marco J F, Goya G F, Serna C J and Morales M P 2008 J. Phys. D: Appl. Phys. 41 134003 [22] Fern´andez-Garc´ıa M P, Gorria P, Sevilla M, Fuertes A B, Gren`eche J M and Blanco J A 2011 J. Alloys Compounds 509 S320 [23] Baker I, Zeng Q, Li W and Sullivan C R 2006 J. Appl. Phys. 99 08H106 [24] Inaga S, Oda S and Morinaga K 2001 J. Non-Cryst. Solids 306 42 [25] Hosono T, Takahashi H, Fujita A, Justin J R, Tohji K and Jevadevan B 2009 J. Magn. Magn. Mater. 321 3019 [26] Li Z, Kawashita M, Araki N, Mitsumori M, Hiraoka M and Doi M 2010 Mater. Sci. Eng. C 30 990 [27] Hergt R, Dutz I and R¨oder M 2008 J. Phys.: Condens. Matter 20 385214 [28] Carrey J, Mehdaoui and Respaud M 2011 J. Appl. Phys. 109 083921

Figure 9. Dependence of the SAR on the amplitude of magnetic 2 field. HAC . The line represents the fitting to HAC .

In order to evaluate the main relaxation mechanisms associated with the heating process, the heating curves were determined as a function of the amplitude of the applied AC magnetic field, HAC . Figure 9 displays the estimated SAR values as a function of HAC , where the described mFe3 O4 mass correction was introduced in the Fe3 O4 @SiO2 system. As figure 9 shows, both samples show similar SAR values in the range of the applied HAC field taking into account the mass correction. Moreover, the samples display the characteristic quadratic field dependence of the superparamagnetic nanoparticles (N´eel relaxation) [24, 27]: SAR(HAC , f ) ∝

χ 00 (f )HAC 2 ρ

(4)

with χ 00 the imaginary component of the magnetic susceptibility and ρ the electrical resistivity. Similar quadratic dependences are obtained within the framework of a general hysteresis model, taking into account the linear response theory [28]. It should be noted that in spite of the wide size distribution in both nanoparticle systems, the contribution of the largest nanoparticles to the self-heating effects can be disregarded. In this case, the field dependence of the SAR should depart from the quadratic field dependence experimentally found in the analyzed magnetite nanoparticles.

4. Conclusions The effect of the silica matrix on the magnetic properties and on the induction heating (magnetic hyperthermia) of magnetite nanoparticles was evaluated. At room temperature the samples display the characteristic superparamagnetic behavior of magnetite nanoparticles. Surface spin disorder is evidenced by the deviations in the temperature dependence of the saturation magnetization in the low temperature range. These disordered effects are greatly enhanced with the silica coating of the nanoparticles. As a result of a lower fraction of Fe atoms being magnetically active at room temperature, the Fe3 O4 @SiO2 nanoparticles display lower SAR values with respect to the sample without silica. Moreover, the quadratic dependence of the SAR on the amplitude of the AC magnetic 6