Effect of Ti Atoms on Néel Relaxation Mechanism at Magnetic Heating Performance of Iron Oxide Nanoparticles

Abstract

The study was based on understanding the relationship between titanium (Ti) doping amount and magnetic heating performance of magnetite (Fe₃O₄). Superparamagnetic nanosized Ti-doped magnetite ((Fe_1−x,Ti_x)₃O₄; x = 0.02, 0.03 and 0.05) particles were synthesized by sol-gel technique. In addition to (Fe_1−x,Ti_x)₃O₄ nanoparticles, SiO₂ coated (Fe_1−x,Ti_x)₃O₄ nanoparticles were produced as core-shell structures to understand the effects of silica coating on the magnetic properties of nanoparticles. Moreover, the magnetic properties were associated with the Néel relaxation mechanism due to the magnetic heating ability of single-domain state nanoparticles. In terms of results, it was observed that the induced RF magnetic field for SiO₂ coated (Fe_0.97,Ti_0.03)₃O₄ nanoparticles caused an increase in temperature difference (ΔT), which reached up to 22 °C in 10 min. The ΔT values of SiO₂ coated (Fe_0.97,Ti_0.03)₃O₄ nanoparticles were very close to the values of uncoated Fe₃O₄ nanoparticles.

1. Introduction

The crystal structure of magnetite (Fe₃O₄) has a spinel cubic structure with Fd-3m space group [1,2]. Fe₃O₄ crystal structures obtain tetrahedral and octahedral sublattices, occupied by Fe²⁺ and Fe³⁺ cations coordinated with 32 oxygen atoms [1,2]. Ferrimagnetic properties are governed by the coupling of cation spins in octahedral and tetrahedral sites [1,2]. Properties, such as low toxicity, suitable magnetic properties, and easy fabrication, make the ferrite particles suitable for hyperthermia usages [3]. Each magnetic domain in a magnetic nanoparticle is oriented along the direction of the externally applied magnetic field. Also, especially for single-domain magnetic nanoparticles, all particles can be oriented in the direction of the magnetic field (Brownian motion), and the magnetic domains inside a particle can rotate without rotating the magnetic particle (Neel motion) [4]. In both cases, due to the desire to be in line with the external magnetic field, alternating magnetic fields initiate heating in and around single-domain magnetic particles.

During the last few decades, magnetic hyperthermia has become a widely researched topic (especially for cancer treatments) [5,6,7,8,9,10,11]. In recent technological applications, ferrite nanoparticles have emerged as the most popular candidate for magnetic hyperthermia applications. The heating performance of magnetic nanoparticles is investigated mainly depending on size and magnetic domain number in the particle [12]. The heating performance of multi-domain magnetic nanoparticles is related to energy losses from magnetic hysteresis due to the domain wall motions and eddy currents induced in magnetic grains [12,13]. Unlike multidomain nanosized particles, the magneto-heating performance of single magnetic domain nanosized particles is mainly correlated with magnetic anisotropy, inter/intra particles interactions and homogeny distributed magnetic nanoparticles [14]. Magnetic relaxation is generally the dominant mechanism in magnetic anisotropy and thus in magneto heating performance of nanoparticles. The magneto-heating performance of single magnetic domain nanosized particles is associated with two mechanisms, Néel and Brownian relaxations [15,16,17]. Brownian motion is created by mechanical fluctuation that performing by an entire nanoparticle rotating its own axis. On the other hand, the Néel relaxation mechanism is independent from the rotational motion of particles. The internal flip of spins with respect to the crystalline lattice creates magnetic heating associated with Néel relaxation [18,19,20]. Both relaxation mechanisms cause an increase in the temperature of magnetic nanoparticles. Even though magneto-heating can be observed with temperature increase, the magneto-heating performance of magnetic nanoparticles can be measured by specific absorption rate (SAR) values. SAR values can be defined by absorbed/converted magnetic energy into thermal energy and in technological applications (especially in cancer treatments) mainly modified properties of magnetic nanoparticles [3]. For magnetic hyperthermia applications, Néel relaxation has certain advantages due to its high SAR values [3]. On the other hand, Brownian relaxation is largely dependent on the viscosity of the medium surrounded by particles, and in cancer treatment SAR value is not enough for magneto-heating processes [3]. Liquid environmental conditions, such as different viscosities of the medium, agglomeration of nanoparticles within different cells, or fixation of nanoparticles in cell membranes (or extracellular tissue), weaken or inhibit the mechanical rotation of magnetic nanoparticles [21]. During the cancer treatments, due to the lack of movement or fluctuation of magnetic nanoparticles inside cancer cells Néel relaxation became the dominant factor for magnetic heating in compare to Brownian movement [3,20,21].

The study is based on the evaluation of the magneto heating performance of individual, non-interacting, and monodisperse particles in highly viscous environments confirming the Néel relaxation mechanism. The magneto heating mechanism is investigated for superparamagnetic Fe₃O₄ nanoparticles doped with Ti atoms. The magneto heating performance of magnetic nanoparticles is not sufficient for cancer treatments, and therefore the surface of magnetic nanoparticles must be functionalized by ligands [22,23]. The general way to improve the surface modification and functionalization ability of magnetic nanoparticles is to coat each nanoparticle with a SiO₂ shell [23]. Optical transparency, highly biocompatibility, biodegradability, and manufacturability with porous surfaces make SiO₂ very useful as a biomaterial [24]. The study also includes understanding the effect of SiO₂ on the heating performance of (Fe_1−x,Ti_x)₃O₄ nanoparticles. Although coating with SiO₂ has no effect on crystal structures, the SiO₂ thickness has caused a remarkable decrease in internal magnetization value of Fe₃O₄ [25]. One of the consequences of the reduction of internal magnetization can be observed as a decrease in the magneto heating performance of magnetic nanoparticles [26,27]. On the other hand, SiO₂ coating reduces interparticle interactions and inhibits aggregation, thereby increasing the heating capabilities of magnetic nanoparticles in AC fields [28,29,30]. Therefore, the magneto heating performance of magnetic nanoparticles is correlated with interacting or non-interacting particles [26,27,28,29,30].

The magnetic nanoparticles such as Fe₃O₄ seem as efficient nano-heaters in biomedical applications [31,32,33]. Surface modified (Fe_1-x,Ti_x)₃O₄ nanoparticles are expected to be potential bio-materials which suitable for loading anticancer drugs and heating under both excitations, RF magnetic field, and UV radiation, which utilize the magnetic nanoparticles useful at clinical hyperthermia applications.

2. Experimental

Magnetite nanoparticles were prepared via the co-precipitation method. Ferrous chloride tetrahydrate (FeCl₂·4H₂O) and ferric chloride (FeCl₃) were used as iron precursors containing different valance states. On the other hand, Tetraisopropil Ortotinatate (TIPO) was used as a Titanium source. Natrium hydroxide (NaOH: 25% by weight) in H₂O and hydrochloric acid in water (HCl) were used as the precipitating agent. To prevent the agglomeration of magnetic nanoparticles, nanoparticles were coated with oleic acid at the end of the process. Ethanol and DI water were used to remove excessive coating agent. The schematic diagram of the procedure was demonstrated in Figure 1. The synthesis was performed on magnetic stirrer at 90 °C. Firstly, ferrous and ferric chloride iron salts were dissolved in water with HCl under Argon gas flow. Secondly, after half an hour, TIPO and NaOH were dropped into the solution. Oleic acid was added to the solution as the last step.

The coated nanoparticles were washed with DI water and ethanol to remove chloride ions and excessive coating materials. Furthermore, the same procedure was performed for the synthesis of the pure magnetite nanoparticles.

The nanoparticles were also coated with SiO₂ by base-catalyzed silica formation from tetraethylorthosilicate (TEOS) in a water-in-oil microemulsion technique, which was mentioned in previous study [34]. The resulting mixture was vigorously stirred for 24 h.

The crystal structures of samples were investigated by x-ray powder diffractometer (XRD), employing a MiniFlex model XRD, produced by Rigaku Corporation (Tokyo, Japan). The XRD patterns were taken under Cu K_α radiation (1.5406 Å) in 2θ range from 10° to 90°. A JEM-2010F high-resolution transmission electron microscope (HR-TEM) (JEOL Ltd., Tokyo, Japan) was used to picture the structural morphologies of nanoparticles. Dc magnetization (σ (H)) measurements were performed at 300 K temperatures in a magnetic field range of ± 2 T.

Magneto-thermal characterization was taken by a homemade setup constructed using the equipment with a frequency of 150 kHz (power generator, thermometer, etc.). Experiments were performed in a custom-made setup with an alcohol thermometer, a covered glass tube, a water-cooled magnetic coil (diameter 50 mm and four turns for 160 Oe), and an AC power generator (Istanbul, Turkey) with a constant frequency of 150 kHz. The SAR measurements were conducted in non-adiabatic conditions as in many publications. The colloidal solution was put into a glass tube and the tube was placed in the coil. The thermometer was directly inserted into the solution. The temperature was measured using the thermometer as a function of time for a duration of 15 min. The filling level of the solution in the tube was adapted to the half-length of the coil to minimize the effects of magnetic field inhomogenity. Between the glass tube and the coils, we used styrofoam as an insulating material.

3. Result and Discussions

The structural analyses were performed employing XRD patterns for SiO₂ coated and uncoated particles as shown in Figure 2a,b respectively. The patterns were in an agreement with Fe₃O₄ diffraction pattern shown in ICDD card (PDF# 74-0748). No contamination or unexpected phase such as TiO₂ based structures, was detected on the XRD patterns. As seen in Figure 2b, even though having high background intensity, originating from the amorphous phase of SiO₂, the Fe₃O₄ patterns were distinctly distinguished at each XRD pattern.

Not observing Ti elements or compounds in the xrd patterns indicated the displacement of Ti atoms inside of Fe₃O₄ lattice. The ionic radius of Ti⁴⁺ is approximately 0.61Å, which is close to the ionic radius of Fe³⁺ (0.64 Å). Thus, Ti⁴⁺ ions are expected to settle instead of by Fe³⁺ ions in the octahedral lattice sites. Due to charge neutrality, a Ti⁴⁺ ion replacement in an octahedral site gives rise to change the valence state of Fe³⁺ ion to Fe²⁺ ion as shown in the chemical equation of (1) [35]. Due to its charge neutrality, the substitution of Ti⁴⁺, a tetravalent positive ion, causes an Fe³⁺ ion to change its valency into Fe²⁺. Chemical Equations (1) and (2) assign the charge neutrality occurring in (Fe_1−x,Ti_x)₃O₄ lattice.

Fe²⁺↔Fe³⁺ + e⁻

(1)

2Fe³⁺→Ti⁴⁺ + Fe²⁺

(2)

Increase in Ti⁴⁺ substitution amount (0.01 ≤ x ≤ 0.025) in (Fe_1−x,Ti_x)₃O₄ lattice cause to formation of Fe²⁺_A(Fe²⁺, Ti⁴⁺)_BO₄ (A, tetrahedral side; B, octahedral side), which inhibits the hopping mechanism between iron ionic states. And thus, the new configuration cause to increase in magnetic anisotropy as mentioned in literature [35,36]. Excessive Ti⁴⁺ substitution (x ≥ 0.025) easily bypasses Fe²⁺ ions, resulting in the formation of differential vacancies [35].

After understanding the crystal structure and possible defects in a lattice, the particle size distributions were investigated employing TEM micrographs. TEM Figures assigned homogeny distributed nanoparticles. In addition, due to covering with oleic acid, no agglomeration across the entire particle distribution was realized as seen in Figure 3.

The size frequencies of particles were calculated by the subprogram of Image-J2. Particles with high differences in particle size were selected in the calculations. As shown on Figure 4, the particles size distributions which were confirmed by ImageJ, were found as 10.3 ± 0.6 nm and 18.7 ± 0.5 nm for uncoated and coated (Fe_0.97,Ti_0.03)₃O₄, respectively. The particles’ size distribution indicated that the particles were in superparamagnetic regions.

The magnetization (σ (H)) measurements at the room temperature were illustrated in the Figure 5 for both samples uncoated and coated (Fe_1−x,Ti_x)₃O₄. As seen from the Figure, the zero remanent magnetization assign the overcoming thermal energy to the magnetic anisotropy energy barrier at all samples. For both coated and uncoated (Fe_1−x,Ti_x)₃O₄ particles, only difference in magnetization curves was the decrease in magnetization value by Ti amount in lattice. In order to understand the magnetic domain states of particles, room temperature magnetic hysteresis curves were obtained as shown in Figure 5.

Not observing coercivity and remanence values on hysteresis curves at room temperature proved that the particles were in superparamagnetic states [37,38]. Heating performance of an individual, non-interacting, and monodisperse particles have high SAR value due to dominating of Néel relaxation in a highly viscous environment. In the superparamagnetic state at room temperature, the magnetic interaction decreased to the lowest values with the effect of thermal energy. Nanoparticle size, anisotropy and interparticle interaction are the principal factors that influence heat generation [39,40]. Moreover, agglomerations between the magnetic nanoparticles occur due to strong magnetic dipole–dipole interactions between particles. As seen in Figure 5, coating with SiO₂ and doping with Ti atoms reduced the magnetization of the particles. A possible mechanism reducing the magnetization of particles is the number of vacancies in the crystal structure. The vacancies may form non-magnetic regions on the particle surface [41]. In addition, the magnetic anisotropy of the easy-axis is another parameter affecting the magnetization values that can be different either parallel or perpendicular to the easy-axis orientation of the domains [42]. These parameters manage the SAR values of particles.

Being predominant of thermal energy to magnetic energy at room temperature indicated that the nanoparticles were in superparamagnetic region. Since the particles were in superparamagnetic state size, the Néel magnetic relaxation was expected to dominate the magneto heating performance of the particles.

In Figure 6, the heating performance of particles was investigated under an ac magnetic field, approximately 13 kA/m field intensity and frequency of 150 kHz (the biological limits are 5 × 10⁹ A/(m.s). The magneto heating measurements were taken immediately after arranging nanoparticles as magnetic fluids in 1 ml ethanol media. As seen on the Figure, the temperature difference (ΔT) reach up to 30 °C for pure Fe₃O₄ and for the same time interval Ti doping lowered the ΔT value down to 20 °C ((Fe_0.97,Ti_0.03)₃O₄ nanoparticles). For (Fe_0.97,Ti_0.03)₃O₄ nanoparticles, the ΔT value was measured as 22 °C, which was good enough for use in vivo studies.

Then, for each particle, the SAR value was calculated as shown in

Table 1. SAR values of (Fe_1−x,Ti_x)₃O₄ (x = 0.00, 0.02 and 0.03) nanoparticles.

Nanoparticles	SAR (W/g)
Fe3O4	155
x = 0.02	70
x = 0.03	3
SiO2 coated Fe3O4	104
x = 0.02 (SiO2 coated)	34
x = 0.03 (SiO2 coated)	116

. The physical quantity of SAR value was determined by defined as the heat released from colloidal magnetic nanoparticles in unit time by Equation (3) [43].

$$\mathrm{SAR}=\frac{\mathrm{Q}}{\Delta {\mathrm{t} \mathrm{m}}_{\mathrm{mag}}}$$

(3)

where $\mathrm{Q}=\mathrm{mc}\Delta \mathrm{T},{ \mathrm{m}}_{\mathrm{mag}}$ is the mass of magnetic nanoparticle, c is the specific heat of the colloid (only ethanol is taken into account, the contribution of magnetic nanoparticles, oleic acid, and SiO₂ to the specific heat are neglected). The calculations were performed for the heat capacity and the density of ethanol 2.57 kJ/(kgK), 0.789 g/mL, respectively.

The SAR values were illustrated in Table 1. As understood from Table 1, coating with SiO₂ lowered the SAR values of (Fe_1−x,Ti_x)₃O₄ nanoparticles. However, for SiO₂ coated samples increase in Ti⁴⁺ ions amount in lattice caused an increase in SAR value, which getting closer to the value of pure Fe₃O₄ nanoparticles.

4. Conclusions

In the study, homogeny size distributed (Fe_1−x,Ti_x)₃O₄ ferrite nanoparticles in oleic acid and at SiO₂ matrix were synthesized via a chemical route. The particles were obtained as superparamagnetic Fe₃O₄ nanoparticles to dominate the Néel relaxation over Brownian relaxation mechanism. Furthermore, lowering the particle size down to superparamagnetic region, coating with SiO₂ and Ti doping into the lattice was the tuned parameters to produce individual, non-interacting, and monodisperse particles. Then, the heating mechanism of SiO₂ coated Ti doped Fe₃O₄ nanoparticles were only correlated with Ti atoms amount in the lattice. Due to the expected coupling between Ti⁴⁺-Fe²⁺ ions in the octahedral site, the heating performance by Ti doping was lower than pure Fe₃O₄. On the other hand, for SiO₂ coated (Fe_0.97,Ti_0.03)₃O₄ nanoparticles, the increase in the amount of Ti⁴⁺ ions in lattice cause an increase in SAR value (ΔT = 22 °C in 10 min), while decreasing for uncoated nanoparticles. The heating performance of (Fe_0.97,Ti_0.03)₃O₄ nanoparticles coated with SiO₂ was almost close to the heating performance of pure magnetite.