A.P. Safronov, 1,2, I. V. Beketov, 1,2, O. M. Samatov, 1, G. V. Kurlyandskaya*, 2,3, S. M. Bhagat, 4, A. Larrañaga, 5, I. Orue, 5, R. Andrade, 5
1Institute of Electrophysics, RAS, Amundsen St. 106, 620016, Ekaterinburg, Russia, 2Ural Federal University, Lenin St. 51, 620000, Ekaterinburg, Russia, 3Department of Electricity and Electronics, University of Basque Country UPV-EHU, Box 644, 48080 Bilbao, Spain, 4FMR Group, Department of Physics, University of Maryland, College Park, MD 20742, USA, 5SGIker, University of Basque Country UPV-EHU, Box 644, 48080 Bilbao, Spaingalina@we.lc.ehu.es
Currently, magnetic nanoparticles (MNPs) are attracting a lot of attention because of the possibility of many novel applications, especially in biomedical research.1,2 To some extent the use of MNPs has been limited both by small production yields and by the difficulty of reproducibly manufacturing nanoparticles with specific properties. This lack of scale and reproducibility makes it very difficult to measure all of the critical parameters using MNPs from a single batch.3-6 In many cases, it is desirable to use multiple methods to measure the same property are desired.7 Here we review two preparation techniques that allow significantly enhanced production rates. First, we discuss the Electric Explosion of Wire (EEW) which yields high production rates (0.2 Kg/hr) and enables subsequent processing to obtain de-aggregated spherical MNPs of Fe, Fe45Ni55, Al2O3, TiO2, ZrO2, Fe2O3, Fe3O4, etc.8-10 Second, we used Laser Target Evaporation (LTE), a preparation technique capable of producing other interesting MNPs at the rate of 0.05 kg/h.11-12 The goal is to achieve the highest saturation magnetization and full de-aggregation leading to development of uniformly dispersed waterbased ferrofluids for biomedical applications such as introduction into cells. Among various materials, magnetite invites particular attention because of its biocompatiblity, high magnetization, high Curie point, and superparamagnetic response for fine particles.13
The review will provide a summary of some of the essential features of the EEW and LTE techniques followed by an outline of the subsequent physical and chemical processes used to obtain well-separated MNPs for incorporation into biomedically useful ferrofluids. Full details of our investigations have been published elsewhere.10-12 In addition, we include results of structural characterization, transmission electron microscopy, magnetization, and ferromagnetic resonance measurements. These data lead us to conclude that we have manufactured magnetically uniform, nearly spherical particles of magnetite (EEW) and maghemite (LTE) that can be used to prepare useful ferrofluids.
Figure 1A shows a schematic of the EEW method. Discharge of the high voltage source produces a current pulse of 104 to 106 A/mm2 density which passes through the iron wire. The resulting energetic injection causes the energy density of the wire to considerably exceed the binding energy. The wire boils in a burst and a mixture of overheated vapor and boiling droplets scatter in the ambient gas. The wire explosion products expand in a cylindrical shape with a much larger density than that of the surrounding gas.8,10 While oxidation initially occurs only at the surface of the cylindrically shaped expansion of explosion products, oxygen penetrates inside the cylinder as it expands and promotes the process of burning and evaporation of the metal and the subsequent vapor condensation to yield oxide particles. Because the melting point of the oxides is much higher than that of metals, the size of the oxide particles that condense during oxidation is much smaller than that of the metal droplets as vapor concentration decreases. The size of the oxide MNPs can be adjusted broadly by delaying the burning process. This is due to the decrease in oxygen concentration and the concomitant increase in the speed of the particles. At higher speeds, the product presents a mixture of MNPs and micron-sized residual particles. Since a high proportion of microparticles can be undesirable for biomedical applications, a separation system using filters was created. This is accomplished through the inertial trap, shown in (4) of Figure 1A, located in the gas system of the EEW facility directly after the explosion chamber (1).
Transmission electron microscopy (TEM) images (Figure 1B) show that MNPs collected from the cyclone (shown as (5) in Figure 1A) are spherical. These so-called “cyclone MNPs” contain a coarser fraction and the presence of larger MNPs compared to samples collected from the filter (shown as (6) in Figure 1A). Both cyclone and filter samples were separated by centrifuging after which they were found to be almost identical with average mean diameters dw of approximately 10 nm.
Figure 1.A) Installation for EEW—method to fabricate large batches of MNPs: 1–explosion chamber; 2–wire feeding mechanism; 3–high-voltage source; 4–inertial trap; 5–cyclone; 6–electric filter; 7–gas mixture deposit; 8–gas circulation fan. B) TEM image of EEW NMPs from the cyclone.
Figure 2 shows the X-ray diffraction (XRD) results of MNP studies performed using a DISCOVER D8 (Bruker) diffractometer operating with Cu-Kα radiation (λ = 1.5418 A). Quantitative analysis was performed using TOPAS-3 software. The average size of coherent diffraction domains was estimated using the Scherrer approach.14 Figures 2A and 2B show the XRD spectrum of MNPs collected from the filter before and after separation by centrifuging. The typical cell parameter for magnetite was a = 0.8396 nm. Slightly lower values of the cell parameters (a = 0.8390 and a = 0.8368 nm for MNPs from cyclone and filter before separation, respectively) indicated that the structure can be described as “defective spinel.” Although particles smaller than 30 nm predominated, a substantial number of particles with diameters above the superparamagnetic13 limit were still present. This minor fraction of large particles is responsible for the ferromagnetism; therefore, the separation of large particles is a requirement for the creation of nanofluids. Conventionally, this is performed using a liquid suspension of powder by sedimentation, which is based on Stokes’s law. Fractionization is the result of the more rapid sedimentation of the large particles compared to the small ones. Successful fractioning can be performed only when the particles are dispersed in liquid and move separately from one another. This necessary condition of de-aggregation is difficult to maintain. Aggregation during redispersion in liquid suspension is one of the main problems in processing air-dry nanoparticles.15 EEW iron oxide MNPs do not form stable suspensions in pure water even after ultrasonic treatment, i.e., as-prepared air-dry MNPs are strongly aggregated and cannot be disassembled by shaking the water molecules. To address this, additional repulsive forces between MNPs must be introduced to overcome their mutual interactions. We performed specific studies on the factors controlling the colloidal stability of EEW MNPs in water suspensions. For this purpose, sodium citrate was selected as the electrostatic dispersant and stabilizer. We used zeta-potential measurements to determine the optimal concentration of electrostatic stabilizer, sodium citrate, and optimal pH to achieve stable EEW Fe3O4-water suspensions.16
Next, the magnetic properties of the MNPs were examined. M(H) magnetization curves and magnetization as a function of temperature in zero-field-cooling (ZFC) and field-cooling (FC) processes were measured in the temperature range 5 ≤ T ≤ 300 K. Information about the size and anisotropy of the EEW Fe3O4 MNPs was obtained by the fitting of ZFC curves and assuming a coherent rotation of the magnetic moments inside each nanoparticle during the relaxation processes (Langevin formalism).17 A good fit was obtained, assuming a log-normal size distribution for the MNPs such that each nanoparticle has an average uniaxial anisotropy (K ~ 1.2 × 106 erg/cm3) and no interactions between particles. The absence of hysteresis of M(H) curves at room temperature and the shape of the hysteresis loops is typical of superparamagnetic systems (SPM). In SPM systems, at low temperatures the magnetization process in an increasing external field is associated with the rotation of the magnetic moment of each MNP toward the direction of the magnetic field. This process is energetically costly and results in high coercivity and remanence. With the increase of temperature, thermal fluctuations give rise to an SPM state, characterized by zero coercivity and a non-saturating behavior.
Figure 2.XRD patterns for EEW MNPs extracted from the filter before A) and after B) separation by centrifugation. Points–experimental data; lines–fitting results. Bragg positions are shown by the colored bars: red for hematite and blue for magnetite.
LTE MNPs are prepared using a device that includes a Ytterbium (Yb) fiber laser with a wavelength of 1.07 μm (Figure 3A). To do this, a 65 mm diameter, 20 mm thick pellet target is pressed from commercially available coarse micron-sized iron oxide powder with a specific surface area of 0.5 to 1 m2/g. The pellet target is annealed at 1,000 °C for 1 hr in air prior to MNP fabrication by mounting it as part of the driving mechanism inside the evaporation chamber (shown as (1) in Figure 3). Both rotation and horizontal movement of the pellet were controlled. An optical system with 200 mm focal length and 0.45 mm diameter of the focal spot were used for focusing the laser beam onto the surface of the target. The driving mechanism provides a 20 cm/s beam scan rate on the target, ensuring uniform wear of the surface. A pulsed laser with a frequency of 5 kHz and pulse duration of 60 μs favors the formation of fine MNPs with narrow particle size distribution. The working gas mixture of N2 and O2 is introduced into the evaporation chamber by a fan. Vapors of iron oxide that are driven out of the focal spot, condenses into MNPs, and are further carried out by the working gas into the cyclone (shown as (4) in Figure 3) and the fine filter (shown as (5) in Figure 3). The cyclone collected the coarse fraction of MNPs while the filter collects the fine fraction. TEM studies (Figures 3B and 4A) confirmed that the LTE MNPs are spherical in shape. Particle size distribution was found to be log-normal in the case of LTE MNPs with a mean diameter of 9.2 nm and dispersion of 0.368. MNPs fabricated by physical condensation in the vapor are found to aggregate. The proper choice of LTE parameters avoids the coalescence of the liquid droplets and prevents formation of non-spherical coarse agglomerates. However, due to the collision of physical surface forces, solid particles are inevitably aggregated in the gas phase. As a result, the iron oxide LTE MNPs contained a fraction of physical aggregates. To address this, the suspension was deaggregated by exhaustive ultrasonic treatment to achieve a constant average hydrodynamic diameter, followed by centrifuging. Two types of de-aggregated suspensions were obtained: MNPs-I (a water-based suspension without electrostatic stabilizer) and MNPs-II (a water-based suspension with sodium citrate as an electrostatic stabilizer, as shown in Figure 4B). The stoichiometric ratio Fe2+/Fe3+ in the MNPs is determined by redox potentiometric titration with potassium dichromate using an TitroLine automatic titrator (Schott Instruments). Titration is performed under an argon atmosphere to prevent oxidation of the Fe2+ in the air. The lattice constant of the crystalline phase is substantially smaller than that of stoichiometric magnetite, but larger than the lattice constant of γ-Fe2O3.
While a detailed description of the analysis of LTE MNPs magnetization behavior can be found elsewhere,12 it is instructive to examine a few parameters. For example, the temperature dependence of the spontaneous magnetization in MNPs, Ms(T), is due to thermal magnon excitation and can be described by the modified Bloch law:18 where M(0) is the zero temperature magnetization, B is the Bloch factor and α=3/2 for the material. The factor α is derived by fitting experimental values; these are strongly size dependent and approach the bulk limit for large powders with diameters on the order of hundreds of nm. In contrast to the bulk, the value of α in MNPs is > 3/2, increasing as the size of the MNP decreases due to the finite-size effect. Using the best fit for M(T), the value α = 2.27 } 0.03 is obtained for LTE MNPs. This value is consistent with the data for ferrite particles obtained by other techniques.18
Figure 3.A) Installation for laser target evaporation method for fabrication of the iron oxide MNPs: 1–evaporation chamber; 2–target; 3–laser; 4–cyclone; 5–electric filter; 6–gas mixture deposit; 7–gas circulation fan. B) TEM image of LTE NMPs from the filter.
Figure 4.A) High resolution TEM image of iron oxide LTE MNPs. B) LTE MNP-based aqueous suspension with sodium citrate as an electrostatic stabilizer.
For practical purposes, both the saturation magnetization of a ferrofluid at room temperature as well as the de-aggregation state stability with respect to application of high magnetic field, are very important parameters. Figure 5 shows an example of M(H) loop LTE MNPs-I waterbased suspension with a maximum obtained concentration of 5 wt. % without an electrostatic stabilizer.
Ferromagnetic resonance (FMR) is a very informative technique for determining the sphericity of powder particles. This is because the field location of the line is strongly affected by deviations from spherical shapes. Accordingly, using techniques that we have developed over the years,4,9,12,19 we measured FMR at room temperature and 8.5 GHz for all of our powders, both as-prepared as well as when diluted with non-magnetic talc. A typical signal is shown in Figure 5B. The following features are notable:
Figure 5.A) Hysteresis loop of LTE MNPs-I water-based suspension without electrostatic stabilizer; inset shows the same data recalculated for the unit mass of the iron oxide. B) Microwave losses at f = 8.85 GHz as a function of the external magnetic field for as-prepared LTE MNPs and dried MNPs corresponding to MNPs-II water-based suspension with electrostatic stabilizer and MNPs-I water-based suspension without electrostatic stabilizer.
As a final step, the biocompatibility of fabricated LTE MNPs-I and MNPs-II ferrofluids was tested. Two non-pathogenic types of yeasts were used: Exophiala nigrum (black yeasts) and its mutant strain (red yeasts) were selected for study of cytotoxicity and the possibility of MNPs accumulation in a living system (Figure 6). Two yeast strains were originally isolated from the Baikal and the Schumak river. These yeasts are interesting model systems as they play an important role in keeping equilibrium in the Baikal ecosystem. Liquid and gelatinous agar nutrient media were used for the cultivation of yeasts with controlled quantities of the ferrofluids. Magnetic measurements performed on dry yeast samples revealed more effective MNPs absorption from liquid nutrient than from gelatinized nutrient. Total reflection X-ray fluorescence (TXRF) using a Nanohunter spectrometer by Rigaku analyzed iron concentrations: yeasts grown in liquid medium had higher Fe concentrations (about 1,000 ppm for black yeasts) than yeasts grown in gelatinous agar medium (120 ppm for black yeasts). Cell morphology was studied by optical, scanning (SEM), and transmission electron microscopies (TEM). In all cases for the selected conditions, no significant alterations of cell morphology were observed (Figure 6).
Figure 6.SEM microscopy of non-pathogenic strains: A) Exophiala nigrum (black yeasts) and B) its mutant strain (red yeasts).
Ferrofluids possess the unique combination of fluidity and capability to be controlled by the application of an external magnetic field. Biomedical applications demand a large amount of de-aggregated magnetic nanoparticles in the form of water-based ferrofluids. High-power physical evaporation methods (EEW and LTE) are excellent candidates to ensure fabrication of fit-for-purpose ferrofluids. Recently, we have shown the possibility of measuring the concentration dependence of an LTE MNPs water-based suspension with an electrostatic stabilizer using a thin film giant magneto-impedance sensor.20 In another case, the sensitivity of giant magneto-impedance sensitive elements was improved by covering the surface with polymer composite-containing EEP Ni water-based MNPs.21 These are interesting directions to follow.
This work was supported by CRDF and UB RAS grant 16991 and SAIOTEKREMASEN grants. Selected studies were performed at SGIker services of UPV-EHU. We thank A.I. Medvedev, A.V. Bagazeev, A.M. Murzakaev, S.V. Komogortcev, J.P. Novoselova, N.S. Kulesh, T.P. Denisova and K.V. Zarubina for special support.