7524
Langmuir 2005, 21, 7524-7527
Manipulation of the Magnetic Properties of Magnetite-Silica Nanocomposite Materials by Controlled Stober Synthesis Yuri A. Barnakov, Minghui H. Yu, and Zeev Rosenzweig* Advanced Materials Research Institute and the Department of Chemistry, University of New Orleans, New Orleans, Louisiana 70148 Received April 4, 2005. In Final Form: May 31, 2005 The paper describes the synthesis and characterization of the magnetic properties of magnetite/silica nanocomposites using a modified Stober method. Magnetite nanoparticles averaging 8-10 nm in diameter and stabilized with oleic acid in toluene were used as the magnetic component of the nanocomposites. SQUID magnetic measurements and ferromagnetic resonance spectroscopy measurements were performed at each stage of the synthesis to understand the properties of the formed composites. Changes of blocking temperature in ZFC/FC SQUID curves correlated with corresponding changes of the resonance field in the ferromagnetic spectra of the sample at each stage of formation. The paper concludes that it is possible to manipulate the magnetic properties of silica/magnetite composite materials by controlling their surface properties and silica coating thickness.
Introduction Magnetic core-shell nanostructured materials have attracted increasing attention since they provide an opportunity to study the magnetic properties of magnetic nanoparticles in confined dimensions. Recent studies showed that the separation between magnetic nanoparticles in composite materials and the surface properties of the nanoparticles greatly affect their magnetic properties.1-4 The ability to tune the magnetic properties of magnetic nanocomposite materials is imperative to successful application of magnetic nanoparticles as contrast agents in magnetic resonance imaging (MRI), as carriers for targeted drug delivery, in biomagnetic separations, and in biosensors.5-7 Functional groups on the surface of nanocomposite particles such as carboxyl and amine are often used to conjugate bioactive molecules to the nanocomposite particles. The magnetic component of the nanocomposite particles provides mobility in the presence of a magnetic field. Despite the numerous reports on the preparation of magnetite nanoparticles and their use in biomedical applications, their commercial application is still limited. The main problem is the formation of agglomerates of magnetite nanoparticles in the presence or even the absence of magnetic field. Agglomeration of magnetic nanoparticles results in the appearance of higher coercitivity and remanence than expected from superparamagnetic nanoparticles. Certain biological applications such as magnetic cell separation require fast attraction of magnetic nanoparticles toward a magnet when an * Author to whom correspondence should be addressed. (1) Lu, Y.; Mayers, B. T.; Xia, Y. Nanoletters 2002, 2 (3), 183. (2) Grasset, F.; Labhsetwar, N.; Li, D.; Park, D. C.; Saito, N.; Haneda, H.; Cador, O.; Roisnel, T.; Mornet, S.; Duguet, E.; Portier, J.; Etourneau, J. Langmuir 2002, 18, 8209. (3) Teng, X.; Black, D.; Watkins, N. J.; Gao, Y.; Yang, H. Nanoletters 2003, 3 (2), 261. (4) Lin, J.; Zhou, W.; Kumnhar, A.; Fang, J.; Carpenter, E. E.; O’Connor, C. J. J. Solid State Chem. 2001, 159, 26. (5) Wang, D.; He, J.; Rosenzweig, N.; Rosenzweig, Z. Nanoletters 2004, 4 (3), 409. (6) Mornet, S.; Vasseur, S.; Grasnet, F.; Duguet, E. J. Mater. Chem. 2004, 14, 2161. (7) Weissleider, R. Nat. Rev. Cancer 2003, 2, 1.
external field is applied. To meet this requirement, the magnetic nanoparticles should be superparamagnetic and characterized with the lowest possible blocking temperature. In addition, the particles must be characterized with a high magnetic moment and no cytotoxicity.8 This can be realized by encapsulating the magnetic nanoparticles in an inert shell, which inherently modifies their surface properties and prevents direct contact between the particles. The theoretical analysis of the properties that are described in this paper is largely based on the assumption that when isolated by a shell, the physical properties of magnetic nanocomposite particles could be described by the Neel-Brown theory for single-domain superparamagnetic particles.9 There have been numerous reports that described the formation of magnetite/silica composite nanoparticles. Haddad et al. prepared 20-30 nm silica-coated magnetite nanoparticles with a shell thickness of 5-7 nm. The resulting particles showed superparamagnetic properties with a decrease in the saturation magnetization of about 40%.10 Giersig et al. assembled silica-coated magnetite nanoparticles on a glass slide using a layer-by-layer selfassembly technique.11 Mikhaylova et al. immobilized bovine serum albumin (BSA) on silica-coated magnetite nanoparticles and applied these particles for cellular measurements.12 In recent studies, Xia and co-workers described the synthesis of silica nanoparticles that were loaded with magnetite nanocrystals.1,13 They were able to control the size of the resulting particles by adjusting the solvent hydrophilicity with smaller particles formed in solvent mixtures of higher hydrophilicity. The resulting particles maintained their superparamagnetic properties, which was attributed to the separation of individual (8) Shen, T.; Weissleder, R.; Papisov, M.; Bogdanov, A.; Brady, T. J. Magn. Reson. Med. 1994, 31, 599. (9) Neel, L. Ann. Geophys. 1949, 5, 99. (10) Haddad, P. S.; Duarte, E. L.; Baptista, M. S.; Goya, G. F.; Leite, C. A. P.; Itri, R. Prog. Colloid Polym. Sci. 2004, 128, 232. (11) Giersig, M. A.; Kotov, N. A.; Liz-Marzan, L. M. Langmuir 1998, 14, 6430. (12) Mikhaylova, M.; Kim, D. K.; Berry, C. C.; Zagorodni, A.; Toprak, M.; Curtis, A. S. G.; Muhammed, M. Chem. Mater. 2004, 16, 2344. (13) Im, S. H.; Herricks, T.; Xia, Y. Chem. Phys. Lett. 2005, 401, 19.
10.1021/la0508893 CCC: $30.25 © 2005 American Chemical Society Published on Web 06/30/2005
Magnetite-Silica Nanocomposite Materials
Langmuir, Vol. 21, No. 16, 2005 7525
Figure 1. An SEM image of the magnetite/silica composite particles averaging 250 nm in diameter (a); TEM images of magnetite/ silica composite particles: (b) at low (120 kV) and (c) higher (300 kV) magnification showing the presence of magnetite agglomerates of 30-40 nm in diameter in the silica composite particles (b) and that the magnetite nanocryystals are separated (c).
magnetite nanocrystals when embedded in the silica particles by the oleic acid capping ligands. The objective of our study was to investigate in more detail the effect of embedding magnetite nanocrystals averaging 10 nm in diameter in silica particles that are 20 times larger on the magnetic properties of the formed composite material. The importance of the study arises from the common, although somewhat undesirable, observation that magnetic nanoparticles agglomerate more often than none in solution, and dry powder forms. Agglomeration of these particles often leads to interparticle interactions that diminish their superparamagnetic properties. The underlying hypothesis of our studies was that forming a silica shell around individual magnetite particles would overcome this problem even if the silicacoated particles still agglomerate. To test this hypothesis, we formed aggregates of silica-coated magnetite particles and evaluated their magnetic properties. The paper describes a simple route for the preparation of clusters of magnetite nanocrystals in silica particles and their characterization using scanning electron microscopy (SEM), transmission electron microscopy (TEM), ferromagnetic resonance spectroscopy (FMR), and SQUID magnetometry. Experimental Section Materials and Reagents. Toluene, amino-propyltrimetoxysilane (99.75%) (APTMS), ethanol (97%), methanol (99%), ammonia (27%), and tetraethylortosilicate sodium (99%) (TEOS) were purchased from Aldrich. Deionized water with 18.2 ohm resistivity was used to prepare the solutions. A ferrofluid (EMG 304) consisting of magnetite nanoparticles averaging 8-10 nm in diameter, coated with oleic acid, and suspended in toluene was used to form the magnetic component of the nanocomposites. Synthesis of Magnetite/Silica Nanocomposite Particles. The synthesis of silica/magnetic nanocomposite structures was realized in three stages: First, 20 mL of methanol was added to 100 µL of magnetic nanoparticle suspension to wash the particles off the surfactant molecules. The magnetic nanoparticles precipitated overnight. The brownish precipitate was collected by slow-speed centrifugation and was dried in air. The magnetite nanoparticle powder was redispersed in 3 mL ethanol in the presence of 50 µL of APTMS to form a stable silica-coated magnetite particle suspension. Ammonia, which is often used to catalyze the formation of silica coatings around nanoparticles, was not used in this reaction to limit the coating thickness of APTM on the surface. APTMS was preferred over the commonly used TEOS for the same reason. The well-known Stober method was used to form the larger silica/magnetite composites.14 A TEOS solution was prepared by adding 0.5 mL TEOS to 1.5 mL deionized water. Then, a second solution was prepared by adding 3 mL ammonia to a mixture of 10 mL ethanol and 1 mL of water. The TEOS and ammonia solutions were added to a 3 mL solution of APTMS coated magnetite nanoparticles under rapid stirring at (14) Sto¨ber, W.; Fink, A.; Bohn, E. J. Colloid Interface Sci. 1968, 26, 62.
800 rpm at room temperature. The mixture was left for 2 days at room temperature. A brownish-opaque suspension was formed, and the silica/magnetic nanocomposite was isolated by magnetic separation using a 500 Oe permanent magnet. The formed powder was washed multiple times with ethanol and was dried in air. Characterization Methods. Scanning electron microscopy (SEM) (JEOL 5410 XL20) and transmission electron microscopy (TEM) JEOL JEM 2010 operated at 200 kV were used to characterize the structural properties of the magnetite/silica nanocomposites. The chemical composition of the particles was determined by X-ray energy-disperse spectroscopy (EDS). Ferromagnetic resonance (FMR) measurements of the nanocomposites were carried out using an X-band (9.83 GHz) Bruker EMX EPR spectrometer. A SQUID magnetometer (Quantum Design, MPMS-5S) with a 5.5 T magnet, a temperature range between 1.7 and 400 K, and with AC and DC magnetic field capabilities were used to obtain zero-field-cooled\field-cooled (ZFC\FC) thermal magnetization curves.
Results and Discussion Previously, there have been many theoretical and experimental studies that aimed to better understand the relations between the magnetic properties of individual particles and the magnetic properties of magnetic nanoparticle assemblies.9,15-17 However, the organic ligands commonly used to stabilize nanoparticles in solution were ineffective in preventing contact between the magnetic nanoparticles when in a powder form. The magnetite/ silica composite particles present a unique system that enables studying the interactions between magnetic nanoparticles in confined dimensions since the rigid silica layer prevents direct contact between the magnetite particles quite effectively.18 As mentioned in the Experimental Section, magnetite (Fe3O4) nanoparticles were coated with a silica layer by incubating magnetic nanoparticles that were washed off stabilizing surfactant molecules in an ethanol solution containing an excess of APTMS. Electronic microscopy images of the magnetite/ silica composite nanoparticles are shown in Figure 1. Figure 1a is an SEM image of the composite particles. It shows that the average size of the composite particles is about 250 nm and as expected they aggregate when dried. Figure 1b and 1c is low- and high-magnification TEM images of the magnetite-silica composite particles formed using the Stober method.13 Figure 1b shows agglomerates of 30-40 nm in diameter in a large number of composite particles. In Figure 1c it can be seen that the magnetite nanocrystals are separated from each other by the thin silica layer first formed when coating the magnetite particles with APTMS. Energy-disperse spectroscopy (15) Shtrikmann, S.; Wohlfarth, E. P. Phys. Lett. A 1981, 85, 467. (16) Dormann, J. L.; Bessais, L.; Fiorani, D. J. Phys. C: Solid State Phys. 1988, 21, 2015. (17) Morup, S.; Tronc, E. Phys. Rev. Lett. 1994, 72, 3278. (18) Batlle, X.; Labarta, A. Top. Rev. 2002, 35, R15.
7526
Langmuir, Vol. 21, No. 16, 2005
Barnakov et al.
Figure 2. FMR spectra of (a) the original ferrofluid in a dry form; (b) ethanol-treated ferrofluid; and (c) silica-coated magnetite nanoparticles.
(EDS) analysis of the formed magnetite/silica composites showed the presence of 12.8% Fe, 43% Si, and 44% oxygen. FMR has been previously used to study spin relaxation mechanisms in ferromagnetic systems.19 FMR resonance lines are broadened and shifted with increasing interactions between ferromagnetic nanoparticles. Figure 2 shows FMR spectra of (a) the original ferrofluid in a dry form, the magnetic nanoparticles are coated with oleic acid; (b) the magnetite particles following the addition of ethanol to the ferrofluid, the nanoparticles precipitate out of the toluene-ethanol mixture since the oleic acid ligands are desorbed in this mixture; and (c) the magnetite/silica composite particles. It can be seen that the resonance peaks shift in response to a change in their capping ligand and environment. Removing the oleic acid ligands induces a resonance shift toward lower magnetic field. This may indicate strong coupling interactions between the particles that are no longer separated by the capping ligands. These particle aggregates show ferromagnetic behavior. Encapsulation of the magnetite particles in silica induces a strong resonance peak shift toward a higher magnetic field. This indicates an increasing separation distance between the magnetite nanoparticles in the magnetite/silica composites, which decrease coupling interactions between the magnetic particles when encapsulated in these silica composites. Another interesting observation is the dependence of the FMR line width, ∆H, on the physical state of the particles. Line widths of FMR spectral peaks are affected by size, shape, composition, porosity in polycrystalline samples, and the distribution of magnetic anisotropy energy. If the magnetic moment Ms of a given sample is known, it is possible to calculate the percent of free volume, defined as porosity P, of polycrystalline powders of magnetic nanoparticles on the basis of the observed ∆H using the Schlo¨mann equation.20
∆H ∼ 1.5(4πMs)[(P/(1 + P)]
(1)
In our experiments, the values of Ms were obtained from magnetic hysteresis loops (data not shown) and were corrected to the mass of Fe used to prepare the particles. To determine whether the porosity of the samples is a main cause for peak broadening, we carried out annealing experiments in which the samples of magnetic nanopar(19) Vonsovskii, S. V. Ferromagnetic Resonance; Pergamon Press: 1966. (20) Schlo¨man, E. Raytheon Technol. Rep. 1956, R-15.
Figure 3. SQUID ZFC/FC curves of (a) the original ferrofluid in a dry form; (b) ethanol-treated ferrofluid; and (c) silica-coated magnetite nanoparticles.
ticles were annealed at 600 °C for 2 h. Our results showed that the line width of all samples did not change significantly because of annealing. Since annealing is known to decrease the porosity of polycrystalline powders, these results indicated that other factors affect the FMR line widths to a higher degree than the porosity of the samples. Since the size, shape, and composition of magnetic materials in our samples were similar and the porosity did not seem to affect the line width, we concluded that the distribution of magnetic anisotropy energy of the samples is the major factor that induces FMR peak broadening in our samples. The distribution of magnetic anisotropy energy, Ka, is defined as the difference between magnetic energies along the easy and hard crystallographic axes in magnetic
Magnetite-Silica Nanocomposite Materials
Langmuir, Vol. 21, No. 16, 2005 7527
materials. In magnetic nanocrystals that are characterized by high surface/bulk ratio, the magnetic anisotropy is typically higher than in bulk materials because of an increase in spin disorder on the particle surface. In particle systems that are characterized by strong coupling interactions, it is possible to estimate the magnetic anisotropy field (Ha) on the basis of the following expression:19
∆H ∼ HaHa/4πMs
(2)
The distribution of magnetic anisotropy energy (Ka) in a given sample is proportional to Ha and could be derived using the following expression:
Ka ) HaMs/2
(3)
It is therefore possible to obtain information about the magnetic anisotropy from ∆H values obtained in FMR spectra and Ms values obtained in magnetic hysteresis loops. In our measurements, the highest value of Ha of about 9000 Oe was found for magnetite/silica composite particles. This value is about 20 times higher than the value of Ha for bulk magnetite which is estimated as 500 Oe.21 Ha values for the unmodified ferrofluid and the ethanol-treated ferrofluid were lower. The decrease in Ha values correlated well with the values of Ms. The Ka value for the magnetite/silica composite particles was about 1.8 × 105 J/m3 which was similar to a Ka value previously reported by Markovich et al.22 for near isolated particles that were dispersed in frozen dodecane. To further support our interpretation of the FMR spectral data, we carried out SQUID zero-field-cooled/ field-cooled (ZFC/FC) measurements of our samples in the form of powders. Figure 3 describes the ZFC/FC curves of (a) the original ferrofluid, (b) the ethanol treated ferrofluid, and (c) the magnetite/silica composite material. The weak temperature dependence of the FC curves in all samples indicates strong coupling interactions between the particles that result in maintaining constant magnetic moment over almost the entire temperature range. The coupling interactions are weaker in the magnetite/silica composite particles at high temperature. The ZFC curves show that the blocking temperature Tb depends largely on the physical state of the material. Tb increases by 125 K when the ethanol-treated magnetite particles (5A) are encapsulated in silica (5B), which decreases the coupling interactions between the particles because of the rigidity of the silica shell formed around them. According to Neel’s theory of superparamagnetism,9 the blocking temperature (21) Bickford, L. R., Jr.; et al. Proc. Inst. Electr. Eng. (London) 1957, 104B Suppl. No. 5, 238. (22) Poddar, P.; Telem-Sharif, T.; Fried, T.; Markovich, G. Phys. Rev. B 2002, 66, 060403.
Tb of a single domain particle is related to the particle volume (V) and the anisotropy energy Ka and can be described as follows:
25κBTb ) KaV
(4)
On the basis of this expression, the calculated Ka value for our magnetite/silica composites was about 1.6 × 105 J/m3, which deviated by only 10% from the value of 1.8 × 105 J/m3 that was obtained in our FMR measurements of the same system. This implies that the magnetite nanocrystals in our magnetite/silica composite materials could be described as superparamagnetic. The deviations between the Ka values obtained by FMR and the calculated values from eq 4 increased significantly up to 100% when the capping ligands of the particles are removed. The large deviations further supported our observations that decreasing the separation distance between the particles leads to a loss of their superparamagnetic properties because of aggregation. Conclusions The paper shows that a modified Sto¨ber method could be used to form magnetite/silica composite particles of about 250 nm in diameter in which magnetite nanocrystals are separated by a thin rigid silica shell. Ferromagnetic resonance spectroscopy and SQUID magnetometry measurements show that the magnetic nanocrystals maintain their superparamagnetic properties in the silica composites. On the other hand, ferrofluids in which magnetite nanocrystals are capped with oleic acid or treated with ethanol to remove these capping ligands tend to aggregate because of strong coupling interactions, particularly in a powder form. This leads to a significant degradation of their superparamagnetic properties. The study implies that proper encapsulation of magnetic nanoparticles is imperative to maintaining superparamagnetism and that it is possible to form large nanocomposite magnetic materials if a sufficiently thick silica shell separates between the individual magnetite nanoparticles. The study also shows that when combined with SQUID ZFC/FC measurements it is possible to utilize FMR spectroscopy to monitor the degree of aggregation in magnetic nanoparticle samples. Acknowledgment. The authors would like to thank Dr. Yonut Dumitru of UNO/AMRI for FMR measurements and Prof. Kevin Stokes of UNO/AMRI and Prof. Alex Burin of the Chemistry Department of Tulane University for valuable discussions. This work was supported by DoD/ DARPA Grant HR0011-04-C-0068 and NSF Grants CHE0134027 and NIRT-0103587. LA0508893
