Abstract
Cobalt ferrite nanoparticles (CoFe2O4 NPs) were synthesized by coprecipitation followed by treatments with diluted nitric acid and sodium citrate. Transmission electron microscope (TEM) and photon correlation spectroscopy (PCS) characterization showed that the size distributions of these nanoparticles were monodisperse and that no aggregation occurred. This colloid showed a long-term stability. Through adjustment of the concentrations of reactants and reaction temperature, the size of the NPs can be tuned from 6 to 80 nm. The size-control mechanism is explained by a nucleation-growth model, where the local concentration of monomers is assumed to decide the size of nuclei, and reaction temperatures influence the growth of nuclei. Magnetization and relaxivity measurements showed that the NPs revealed size-dependent magnetization and relaxivity properties, which are explained via a “dead magnetic layer” theory where reductions of saturation magnetization () and are assumed to be caused by the demagnetization of surface spins.
1. Introduction
The development of uniform magnetic nanoparticles (MNPs) has been intensively pursued for their scientific and technological importance [1–3]. The synthesis of MNPs with average sizes from 2 to 50 nm is of significant importance because of their applications in several fields, especially in biomedicine for magnetic resonance imaging (MRI) [4, 5], cell labelling [6, 7], and drug delivery [8–11]. Of special interest are their magnetic properties in which the differences between a massive or bulk material and a nanoscaled one are especially pronounced. The magnetic properties are particularly sensitive to the particle size, which is determined by the finite size effects (related to the reduced number of spins cooperatively linked within the particle), and by surface effects (greater as the particle size decreases) [12–15]. The water solubility of MNPs is necessary for medical applications, and their aggregation, caused by the huge specific surface area and magnetic interactions, must be avoided. It is thus necessary to adopt methods to stabilize the MNPs, either by using surfactants or by changing their surface potential.
Oleic acid and oleylamine are the surfactants most used for the synthesis of MNPs in organic solvent [16–19] and citric acid for water phase synthesis [20–23], as these capping agents tend to be absorbed on the particular high-energy facets. Their overall specific surface energy is more or less reduced, so the tendency towards aggregation is decreased. The ratio between growth rates in different directions can be varied by special adsorption of organic surfactants onto particular crystallographic facets specially inhibiting the growth and also the possible aggregation. On the other hand, the stability of MNPs can also be achieved by changing their surface potential [24–27], by introducing an electric repulsive force between NPs.
The general chemical formula of spinel ferrite is MFe2O4 (M = Fe, Mn, Co, Mg, Zn, Ni, etc.). It has a face-centered cubic (fcc) structure with a large unit cell including eight formula units. There are two kinds of lattices for cation occupancy: tetrahedral and octahedral, which are denoted as A and B sites, respectively (see Figure 1). Spinels with only ions occupying A or B sites are called direct (or called normal) or inverse, respectively [28].
Among MFe2O4 NPs, cobalt ferrite (CoFe2O4) recently attracted interest due to its higher magnetization and proton relaxivity than magnetite (Fe3O4) which is mostly used as MRI contrast agent [29]. CoFe2O4 has a partially inverse spinel structure with the formula FO4, where the first and second square brackets indicate A and B sites, respectively. The ratio, , has been found to vary from to for extremes-quenched and slowly cooled samples [30, 31], respectively. For both ionic distributions, a magnetic moment of more than 3 per unit chemical formula or 71.5 Am2 kg−1 is anticipated assuming 5 for and 3 for ions (as ions are assumed to be 4 , the total moment of Fe3O4 and -Fe2O3 per formula unit is anticipated to be 4 and 2.5 , resp., and the experimental values vary with different synthetic methods [32, 33]). The observed values, 3.4 for extremes-quenched samples and 3.9 for slowly cooled samples, respectively, confirmed this expectation. There have been many reports on synthesis of CoFe2O4 NPs by different methods. Davies et al. have reported the preparation of CoFe2O4 NPs with sizes of about 3 nm by coprecipitation method [34], which exhibited multiaxial anisotropy. The value of the magnetic anisotropy constant calculated from measurements of the decay of remanence was similar to that reported for bulk cobalt ferrite. Moume and coworkers have prepared CoFe2O4 NPs with sizes of about 5 nm by the inverse micellar process [35].
To our knowledge, very little work has been done on size-controllable synthesis of CoFe2O4 NPs. Recently, it has been reported that the magnetic and relaxivity properties are dependent on the size of NPs [4]. Here CoFe2O4 NPs were obtained by coprecipitation, and the size-controlling factors were studied during the synthesis process, as well as the stabilizing effects of both citrates as the capping agents and dilute nitric acid solution to change the surface potential of NPs.
2. Materials and Methods
2.1. Materials
Distilled water was used to prepare all aqueous solutions. Ferric chloride hexahydrate (FeCl36H2O), cobalt chloride hexahydrate (CoCl26H2O), sodium citrate, 1, 10-phenanthrolin (99%), and hydroxylamine hydrochloride (NH2OHHCl) were purchased from Sigma Chemical (Bornem, Belguim). All other reagents and solvents were purchased from Aldrich Chemical (Bornem, Belguim) and were of the highest grade commercially available.
2.2. Methods
1 mL of 20 M NH4OH was added dropwise to the solution mixture of 20 mL of 2 M FeCl36H2O and 1 M CoCl26H2O in 0.4 M HCl (denoted as (Co+Fe)) with constant stirring, at 80°C. The precipitation occurred immediately, turning the color of the reaction suspension into dark brown. During the precipitation, the reaction mixture was vigorously stirred for 2 h at desired temperature. The resulting precipitates were separated by a permanent magnetic field and washed with distilled water and the separation/washing processes were repeated 5 times until the washed water was free of chloride ions. The washed samples were dispersed in 20 mL 0.02 M HNO3 solution and stirred for 1 h at 80°C, followed by adding 0.8 g sodium citrate and stirring for another 4 h. Centrifugal filter units (Millipore, 10k MWCO) were used to purify the samples and remove the unabsorbed and citric ions.
2.3. Instrumentation
TEM images were collected on a Tecnai 10 (FEI, Hillsboro, USA) with operating voltage at 200 kV. IR spectra were obtained on a Spectrum 100 Series (Perkin Elmer, Massachusetts, USA) spectrometer in ATR mode. The hydrodynamic size and surface potential of CoFe2O4 NPs were measured on Zetasizer Nano Series Zen 3600 (Malvern, Worcestershire, United Kingdom), into which 1 mL sample solution of about 2 mM was put in the cell and measured at 37°C.
As there were peak overlaps between Fe and Co absorption for inductively coupled plasma (ICP) measurements, the concentration of Fe was confirmed by spectrometry, using 1,10-phenanthrolin as the chromogenic agent, and the concentration of Co was obtained through the external standard method by measuring the values, for which the detailed process is described in supplementary in supplementary material available online at http://dx.doi.org/10.1155/2013/462540.
The stability of the NPs was evaluated using a Turbiscan optical analyzer (Formulaction SA, L’Union, France), an innovative analytical instrument able to determine the long-term stability of colloidal systems. 15 mL sample colloid was placed into a cylindrical glass tube and submitted to Turbiscan Lab Expert stability analysis. The measurements were carried out at °C for 24 h at intervals of 1 h.
To study the magnetic properties of the NPs obtained, magnetic hysteresis curves were obtained by VSM-NUVO (Molspin Ltd., Newcastle-on-Tyne, England), were measured on MiniSpec mq-20 (20 MHz) and mq-60 (60 MHz) (Bruker, Ettlingen, Germany), and nuclear magnetic relaxation dispersion (NMRD) profiles were performed on Spinmaster-FFC 2000 relaxometer (Stelar SRT, Mede, Italy).
3. Results and Discussion
3.1. Preparation and Size Controlling of CoFe2O4 NPs
The reaction to synthesize CoFe2O4 NPs can be described by where the variable stands for the relative amount of NH4OH. The total concentration of Co2+ and Fe3+ is denoted by Co2+ + Fe3+, and the concentration of OH− is denoted by [OH−]. By adjusting , the ratio Co2+ + Fe3+ : OH− (denoted by , where ) varied from 0.00625 to 0.1. The average diameter of CoFe2O4 NPs can be tuned from 13 to 38 nm, by varying the reaction parameters, such as feeding order, reactant concentrations and pH value. Figure 2 shows the TEM pictures of CoFe2O4 NPs after treatment with diluted nitric acid and sodium citrate (denoted as CoFe2O4//citrate with the hydrodynamic diameter (measured by PCS, denoted as ) of 28.35 nm). The TEM picture agrees with monodisperse particles as confirmed by the corresponding statistics graphs using PCS (see Figures S3 and S4).
(a)
(b)
During the process of drying samples on carbon-coated copper grids, several NPs were adsorbed onto each other, which look like “aggregations” in Figure 2. To clearly prove the absence of aggregations, sizes of CoFe2O4 NPs measured by different methods are compared in Table 1. Although various methods (VSM, TEM, NMRD, and PCS) are available, the accuracy of the results for NPs sizes is not always clear because of possible systematic errors of the analysis. Each technique has its own advantage and disadvantage. TEM is certainly the most direct method, providing real images of the particles, which can be considered as an essential tool to get an impression of the homogeneity of given samples, but the lack of contrast or overlap of particles complicates the TEM image analysis. The technologies of VSM and PCS are both based on indirect calculations. The calculation of NPs sizes by VSM is based on the Langevin curve fitting as follows: where is the magnetization of the suspension at a field , is the magnetization at saturation, and is the Langevin function (3) where and are the Boltzmann constant and the absolute temperature, respectively, and the fitting process will be described subsequently [4]. PCS is based on a time-dependent fluctuation in the scattering intensity, as the source light hits small particles and the light scatters in all directions (Rayleigh scattering) so long as the particles are small compared to the wavelength (below 250 nm). The size value obtained by PCS is the hydrodynamic diameter of a sphere, which may be larger than that from TEM due to the presence of surfactants or coating molecules. As shown in Table 1, the diameter values obtained by VSM, TEM, and PCS are denoted by , , and , respectively. for each sample is much bigger than , and , because of the absorbed citrate layer.
It has been found that the values of some parameters experimented will affect the particle size and morphology [29, 38]. The following parameters were varied: the value of , the reaction temperature, and the pH (see Figure 3). It was found that the NPs size is very sensitive to the ratio with adjustment of [Co2+ + Fe3+], as shown in Figure 3(a). To systemically investigate the influence of on the size, [Co2+ + Fe3+] or [OH−] was changed for different feeding orders: adding NH4OH into (Co + Fe) or adding (Co + Fe) into NH4OH, as shown in Figures 3(a) and 3(b), respectively. Increasing induced completely opposite effects on the final sizes of CoFe2O4 NPs depending on the feeding orders. When NH4OH is added into (Co + Fe) (see Figure 3(a)), the increase of caused by varying [Co2+ + Fe3+] (the line with triangle symbols) results in a linear increase of the hydrodynamic size of the NPs from 15 to 80 nm, whereas varying [OH−] (the line with star symbols) does not influence the NPs sizes (the size is about 22 nm, regardless of the value of ). On the other hand, when the (Co + Fe) mixture was added into the NH4OH solution (see Figure 3(b)), NPs sizes decreased from 22 to 7 nm as [OH−] decreased (increasing ), but remained constant (about 17 nm) when [Co2+ + Fe3+] was varied. Such a phenomenon had never been previously reported and can be explained as described hereafter.
(a) Adding NH4OH into (Co+Fe) of pH = 2
(b) Adding (Co+Fe) of pH = 2 into NH4OH
(c) Adding NH4OH into (1 M Co+2M Fe)
(d) Adding (1 M Co+2M Fe) into NH4OH
It is well known that the formation of NPs can be divided into two-step processes: nucleation initiated by a sudden increase of monomer concentration up to super-saturation levels and subsequent growth of nuclei with progressive consumption of monomers in reaction medium (see Figure S5) [39]. The final geometrical properties of NPs are determined by several parameters during the nucleation and growth processes. During the coprecipitation of Co2+ and Fe3+ into CoFe2O4, when the drop of added solution (denoted by SolA, as the green solution in a pipette in Figure S5) comes into contact with the initial solution (denoted by SolB, as the yellow solution in a beaker in Figure S5), the reaction of (1) is initiated immediately to form CoFe2O4 monomers. As a consequence, the local concentration of monomers suddenly reaches saturation and then the super-saturation state, followed by the nucleation and growth processes. In such a classical ionic reaction, SolA is at a high concentration (and small volume) to make sure that the total volume of the reaction system does not change too much. So the amount and concentration of monomers formed mainly depend on SolB which is at a lower concentration.
The above hypothesis explains why only the concentration of SolB influences the size of CoFe2O4 NPs, no matter whatkind of feeding order is adopted. When adding NH4OH into the (Co + Fe) mixture solution, [OH-] is much larger than [Co2+ + Fe3+], so only the increase of [Co2+ + Fe3+] (SolB) can result in bigger CoFe2O4 NPs, and the change of [OH-] (SolA) does not influence the size of NPs. On the other hand, when adding the (Co + Fe) mixture into NH4OH, the size of NPs decreases with lowering [OH-] (SolB) but remains constant with changing [Co2+ + Fe3+] (SolA). This means that NPs sizes do not dependent on the ratio between [Co2+ + Fe3+] and [OH-] but depend on the concentration of monomers. To confirm this hypothesis, HCl was used to adjust the pH of the Co + Fe mixture, when NH4OH was added to this mixture. The results show that NPs size increases as the pH increases (see Figure 3(c)). The reason should be that the protons in the Co + Fe mixture react with the OH− of the NH4OH solution. Hence the local concentration of NH4OH (SolB) is decreased and smaller NPs are obtained. It had been reported by Lee et al. that the size and structure of nuclei were the key to deciding the final size of obtained NPs [40]. Consequently, a higher local concentration of monomers leads to larger nuclei and thus increases the size of final NPs.
The influence of the reaction temperature was studied and the results are shown in Figure 3(d). The NPs sizes become larger as the temperature increased. This phenomenon matches with literature reports showing that high temperature contributes to the growth of nuclei to obtain large NPs [41].
3.2. Stabilization of CoFe2O4 NPs
Li et al. reported that acidic treatment on CoFe2O4 NPs resulted in a non-crystalline porous wrapping layer on the surface and in the constituting of coated bodies [42]. Here the stability of CoFe2O4 NPs treated with diluted nitric acid (denoted by CoFe2O4/) and NPs only treated by sodium citrate (denoted as CoFe2O4/citrate) were compared after dispersion in water and being left at room temperature. CoFe2O4/citrate suspension was found to deposit in about 0.5 h, while CoFe2O4/ remains stable for more than 2 months, without any sedimentation. To further test the stability of CoFe2O4/, a strong magnetic field was applied (see Figure 4). It was found that CoFe2O4 without any treatment and CoFe2O4/citrate NPs were easily separated by magnetic force, but that CoFe2O4/ suspension kept stable. After nitric acid treatment for 1 h, sodium citrate was added and stirring was continued for another 4 h. The obtained NPs, denoted by CoFe2O4//citrate, are kept stable under magnetic field like CoFe2O4/ (not shown). of CoFe2O4, CoFe2O4/citrate, CoFe2O4/, and CoFe2O4//citrate measured by PCS was 1100, 600, 12 and 13 nm, respectively. Because of their ultrasmall size, CoFe2O4/ and CoFe2O4//citrate NPs have no obvious response to an applied magnetic field and the electrostatic repulsion force also contributes to keeping the suspension stable. Figure S6 gives the IR spectra of CoFe2O4/citrate and CoFe2O4//citrate NPs. The peaks at 3392, 2905, 2850, 1587, and 1379 cm−1 are vibration modes of sym(OH), asy(CH2), sym(CH2), asy(CO2), and sym(CO2), respectively. The spectrum demonstrates the successful attachment of sodium citrate onto the surface of CoFe2O4//citrate because the characteristic absorption peaks of sodium citrate at 3392, 1587, and 1379 cm−1 are strong (the black curve in Figure S6), whereas these peaks are weak for CoFe2O4/citrate (the red curve in Figure S6). Sodium citrate was thus not effectively anchored on the CoFe2O4 NPs surface and there were aggregations as well, according to Figure 4(b). This can be explained by the fact that the acidic treatment induces etching of the bigger particles or aggregations as well as other single smaller particles, thus forming porous viscous layers on their surface, where sodium citrate can be adsorbed.
To investigate the long-term stability of the CoFe2O4//citrate colloid, the sample was tested by both its optical transmission and photon backscattering properties. The principle of this measurement is based on the variation of the droplet volume fraction (migration) or mean size (coalescence), which results in the variation of backscattering and transmission signals. The backscattering signal will show change as a function of time if particle migration occurs and is graphically reported in the form of positive (backscattering increase) or negative peaks (backscattering decrease). In fact, the migration of particles from the top to the bottom of a sample leads to a progressive concentration decrease at the top of the sample, which causes an increase in the backscattering signal (positive peak) and a decrease in the intensity of the transmission (positive peak). Such changes will be affected if aggregation of particles occurs [43–45].
The transmission and backscattering profiles of CoFe2O4//citrate colloid ( nm) are shown in Figure S7. The backscattering profile being within the interval ±2% means that no variation of particle size occurs during measurements. Variations of backscattering profile () greater than 10% (either positive or negative value) are representative of an unstable system. In Figure S7, the curves between 5 and 30 mm (bracketed by red dashed lines) represented transmission and backscattering intensions of colloid parts. was found within the interval ±2% during 24 h, which proved that no variation of particle size occurred and this CoFe2O4 colloid was stable. Other CoFe2O4 samples showed similar results (not shown here), confirming that NPs obtained by our strategy are stable and that no aggregation occurs.
To investigate the stability of CoFe2O4 NPs in medium with different pH values, the sizes and zeta potentials of CoFe2O4/ NPs and CoFe2O4//citrate (see Figure 5) were studied. As shown in Figure 5(a), with increasing pH from 3 to 5, the hydrodynamic size of acid treated CoFe2O4 NPs increased from 34 to 90 nm. In acid medium, there is an equilibrium between the adsorption and desorption of ions on the surface of CoFe2O4 NPs. When increasing the pH, lower [] in the medium led to more desorption of ions on the surface of NPs and decreased the electrostatic repulsion, which caused the aggregation and enlarged the size of NPs. It was furthermore confirmed by the study of zeta potential (the red curve in Figure 5(a)) that the surface of acid treated CoFe2O4 NPs was positively charged in acidic medium, and the zeta value decreased from 46 to 36 mV as the pH increased. At , the aggregation was too high and caused deposition of NPs. On the other hand, the size and zeta potential of CoFe2O4//citrate remain constant (about 33 nm and −27 mV, resp., in the pH region of 3–10) (see Figure 5(b)). As discussed above, acidic treatment makes NPs surfaces positively charged and forms porous viscous layers, where citrate can be adsorbed. The adsorbed citrate turns NPs surface from being positively to negatively charged, which has a stable zeta potential over the pH range.
(a)
(b)
3.3. Magnetic Properties of CoFe2O4 NPs
Magnetic measurements were carried out at room temperature on aqueous suspensions of CoFe2O4//citrate NPs with different . The fitting of the experimental magnetization data (see Figure 6) by (2) and (3) allows for the determination of the crystal size and its . The lack of hysteresis (almost zero coercive force, Hc) and remanence at ambient temperatures is characteristic of superparamagnetic materials. The specific magnetization is the magnetic moment per unit mass. In this paper, magnetizations are reported with respect to the total mass of iron and cobalt (i.e., Am2/kg Fe + Co). The results (see Figure 6) indicate that the saturation magnetization, , increases from 28.42 to 45.50 Am2 kg−1, as increases from 12.98 to 37.89 nm. The size-dependent change of has already been observed in some small particle systems [46]. This can be explained by postulating the existence of a “dead magnetic” layer due to the demagnetization of the surface spins when the size of the sample is in nano-region, which causes a reduction of [47]. The ratio of surface to volume is inversely proportional to the particle size. As the particle gets smaller, surface molecules make up greater shares, which leads to more “dead magnetic” components. So as the particle size decreases, the proportion of “dead magnetic” components increases and decreases.
The efficacy of NPs as contrast agent for MRI is related to their and relaxivity values (Table 2). Relaxivities are reported with respect to the total molarity of iron and cobalt (i.e., s−1 mM−1 Fe + Co). The values of (related to contrast) increase with particle diameters at 20 and 60 MHz. When of CoFe2O4//citrate NPs increases from 12.89 to 37.89 nm, the values of at 20 MHz increase sharply from 16.6 to 75.1 s−1 mM−1. values of NPs with larger than 60 nm cannot be accurately measured because of their aggregation in the presence of a magnetic field. This size dependence is believed to arise from surface spin anisotropy, which is more pronounced for smaller particles due to larger surface area to volume ratios. As the NPs are superparamagnetic, the transverse relaxivity rate is enhanced due to rapid diffusion and exchange of surrounding water molecules. In such a “motion averaging regime” (MAR) [48], it is postulated that the CoFe2O4 contrast agents are homogenously distributed and the diffusion of surrounding water molecules to and from the surface layer of superparamagnetic cores occurs in faster time scales than proton magnetic relaxation.
Similar phenomena were found in the NMRD measurements. Obtained curves give the relaxivity evolution versus the external magnetic field, where the relaxivity is defined as the increase of the relaxation rate of the solvent (H2O) induced by 1 mmol/L of the active iron and cobalt ions. The spin lattice relaxation profiles of the above CoFe2O4 NPs are shown in Figure 7. In general, for magnetic NPs colloids, the shape of the NMRD profile is determined by (i) the size and crystallinity of magnetic cores, as they affect the ; (ii) the accessibility of the cores to the diffusing solvent molecules; and (iii) interactions between cores, as they affect the anisotropy energy [4]. The size effect is known to be a potentially strong contributor in low frequency range in particular. The NMRD data demonstrate that increasing NPs sizes produced systematic changes in the magnetic resonance properties of the colloids, particularly in the low-frequency (sub MHz) part of the profiles, where values of increase as the diameter of NPs increases. All samples show characteristic superparamagnetic behavior, as their Larmor frequencies present continuous decreasing trends in the high-frequency part and the corresponding fitting lines coincide well with the superparamagnetic model, and the magnetization saturation and particle diameters obtained by the fitting, as well as those obtained in the above VSM fitting, are listed in Table 3. As both kinds of fittings are based on the Langevin function, the values of and particle diameters obtained by fittings of NMRD profiles and VSM data are quite similar.
4. Conclusions
In summary, the present work demonstrates a facile approach to the size-controllable preparation of superparamagnetic CoFe2O4 NPs by coprecipitation. Using different feeding orders, it was found that only the low-concentration solution has influence on the size of NPs. By a novel model, it was presumed that local concentration of monomers decided the size of nuclei in the nucleation process, and the temperature influences the growth of nuclei.
In addition to playing a role in stabilizing, acidic treatment also formed porous viscous layers on particle surface, where sodium citrate could be absorbed. The adsorbed citrate turned NPs surface from positive to negative charged, resulting in an NPs colloid solution stable over a wide pH range.
The prepared CoFe2O4 NPs showed size-dependent magnetization and proton relaxivity, which can be explained with a “dead magnetic” layer model. The surface spin anisotropy induces decreases in and for smaller particles. Furthermore, this facile synthetic procedure may be widely applied in size-controllable synthesis of mono- metallic or multi-metallic oxides by using different metallic cations.
Conflict of Interests
The authors declare that there is no conflict of interests.
Acknowledgments
This work was supported by the Fonds de la Recherche Scientifique, the ARC Program 05/10-335 of the French Community of Belgium, and the ENCITE program of the European Community. The support and sponsorship concerted by COST Action D38 “Metal-Based Systems for Molecular Imaging Applications” and the EMIL program are kindly acknowledged. The authors are grateful to Dr. Yves GOSSIUN in UMons for performing VSM analysis and to Dr. David PEREZ-MORGA for the TEM analysis.
Supplementary Materials
The Supporting Materials include the measurement of iron and cobalt concentrations [Fe] and [Co] in the prepared CoFe2O4 nanoparticles, using an external standard method combining colorimetry and relaxivity. Size measurements of CoFe2O4 sample by photon correlation spectroscopy (PCS) are shown, along with the infrared (IR) spectra and transmission and backscattering profiles to prove the stability of obtained colloids. In addition, a scheme of the assumed nucleation-growth model is presented to explain the mechanism of size tuning in our article.