Tailoring the Magnetic and Hyperthermic Properties of Biphase Iron Oxide Nanocubes through Post-Annealing

Abstract

Tailoring the magnetic properties of iron oxide nanosystems is essential to expanding their biomedical applications. In this study, 34 nm iron oxide nanocubes with two phases consisting of Fe₃O₄ and α-Fe₂O₃ were annealed for 2 h in the presence of O₂, N₂, He, and Ar to tune the respective phase volume fractions and control their magnetic properties. X-ray diffraction and magnetic measurements were carried out post-treatment to evaluate changes in the treated samples compared to the as-prepared samples, showing an enhancement of the α-Fe₂O₃ phase in the samples annealed with O₂ while the others indicated a Fe₃O₄ enhancement. Furthermore, the latter samples indicated enhancements in crystallinity and saturation magnetization, while coercivity enhancements were the most significant in samples annealed with O₂, resulting in the highest specific absorption rates (of up to 1000 W/g) in all the applied fields of 800, 600, and 400 Oe in agar during magnetic hyperthermia measurements. The general enhancement of the specific absorption rate post-annealing underscores the importance of the annealing atmosphere in the enhancement of the magnetic and structural properties of nanostructures.

1. Introduction

The recent ability to fabricate magnetic structures of nanometer size, known as nanomagnetism, has allowed for the study of unique magnetic phenomena that are unobservable at larger length scales [1,2,3,4]. It has been shown that by utilizing nanostructures of varying sizes and shapes, their applicability in fields ranging from biomedicine to memory devices can be improved [2,3,5,6,7,8,9]. Iron oxide, one of the most historically studied magnetic systems, has been found to be a unique platform for observing the interplay between different thermodynamically stable phases and their respective magnetic orders, which has allowed for broad applications [3,10,11,12,13]. This composition, already approved by the Food and Drug Administration (FDA), is observed in many devices and has shown potential to be tuned within its iron oxide composition to multiple magnetic phases with distinct features; thus, the phase-tunability of iron oxide phases such as magnetite (Fe₃O₄), hematite (α-Fe₂O₃), maghemite (γ-Fe₂O₃), and wüstite (FeO) are currently being examined [14,15]. The phase-tunability with size of iron oxide nanostructures has shed light on the stability of certain phases in a variety of conditions, such as the FeO phase, which possesses higher stability in smaller structures measuring ~30 nm [16]. As an extension of an initial study of how the stability of iron oxide phases changes with size, different types of phase tunability in iron oxide nanostructures need to be explored. Prior research has shown that functionalizing iron oxide nanostructures for hyperthermia applications requires the tuning of their shape and size [6,17]. Of the various shapes that have previously been studied, nanocube structures measuring ~30 nm have been shown to possess elevated potential compared to other structures such as spheres [6]. This is due to their enhanced anisotropy, which results in higher specific absorption rates (SARs) and higher contrasts in magnetic resonance imaging (MRI) in comparison to its spherical counterparts [5,6,17,18]. Nature itself has shown the importance of contemplating iron oxide nanocubes with structures such as magnetosomes which are found within magnetotactic bacteria that possess the capability to act as compasses by utilizing the Earth’s magnetic field to navigate, opening up its use in further applications spanning from drug delivery to nanoelectronics [19,20,21].

In this study, we explored the effects of different annealing atmospheres to manipulate the phase volume and hence magnetic properties of biphase iron oxide nanocubes for biomedical applications. Nanocubes consisting of Fe₃O₄ and α-Fe₂O₃ with no observable exchange bias or superparamagnetic behavior were annealed in the presence of O₂, N₂, He, and Ar to observe compositional and magnetic changes. Herein, O₂ is a reactive gas that can directly interact with iron oxide structures, while N₂ is a comparatively less reactive gas that can facilitate reaction-stimulating environments. The latter two gases are well-known inert gases, with He being the lightest. Furthermore, the as-prepared and treated nanocubes were evaluated for their effectiveness in magnetic hyperthermia therapy, which showed that the coercivity enhancement plays a major role in tuning hyperthermia efficiency in comparison to the saturation magnetization enhancement.

2. Materials and Methods

The synthesis of iron oxide nanocubes (NCs) was carried out following the procedures put forward by Nemati et al. [6]. The chemicals Fe(III)-acetylacetonate (Fe(acac)₃, ≥99.9%), oleylamine (OA, 70%), oleic acid (OY, 90%), benzyl ether (BE, 98%), and 1,2-hexadecanediol (HDD, 90%) (Sigma-Aldrich, St. Louis, MO, USA) were uniformly mixed using a magnetic stirrer in a three-necked round-bottomed flask with a jacketed heating mantle. The side openings were dedicated to a thermometer and continuous N₂ flow, which was initially carried out with the condenser detached at the first stage to facilitate degassing for 30 min at 110 °C to ensure that moisture and air were removed from the flask. The condenser was then fitted with cold water circulation to ensure maximum retention of the reactants throughout the process. The nucleation stage was then initiated at 200 °C and lasted for 120 min, with the final reflux stage carried out for 45 min, after which the solution was allowed to reach room temperature in the absence of a heat source. The obtained turbid, reddish solution was then cleaned a minimum of two times with ethanol and a small amount of hexane. The cleaning was carried out via back-to-back sonication for ~5 min and by centrifuging the solution at 5000 revolutions per minute for 5 min; after each time, the supernatant was disposed. The obtained product was then allowed to dry down to a fine powder which was evaluated for its structural and morphological characterization using a FEI Morgagni 268 transmission electron microscope (TEM) (FEI, Hillsboro, OR, USA) operating at 60 kV; this was followed by a compositional evaluation via diffractometry using a Bruker AXS D8 X-ray diffractometer (XRD) (Bruker, Madison, WI, USA) functioning in Bragg–Brentano geometry at a Cu Kα wavelength. The obtained samples, which conformed to the initial structural and compositional evaluation, were then separated, with some kept as controls while the others were heated in the presence of multiple gases at 300 °C for 2 h in a ceramic combustion boat placed inside a tube furnace. The gases oxygen (O₂), nitrogen (N₂), helium (He), and argon (Ar) were of ultra-high purity grade and were obtained from nexAir. The treated samples were then evaluated again for their structural and morphological characterization using the TEM and XRD. All magnetic measurements were performed in a DynaCool Physical Property Measurement System (PPMS) (Quantum Design, San Diego, CA, USA), utilizing the vibrating sample magnetometer (VSM) option. Magnetic hyperthermia measurements were performed using a 4.2 kW Ambrell EasyHeat Li3542 system with varying magnetic fields (0–800 Oe) at a constant 310 kHz frequency. The samples were measured at 20 °C as the starting temperature for a period of 300 s and with 1 mg/mL of nanoparticles in a 2% by-weight agar solution prepared using deionized water.

3. Results and Discussion

3.1. Structural Characterization

The structural and morphological properties of the iron oxide nanocubes (NCs) were evaluated using ImageJ 1.53k software using the images obtained from the transmission electron microscope (TEM), which yielded a size distribution of nanocubes at 34 ± 4 nm. Furthermore, the size distribution remained statistically similar in all annealed samples. The lattice parameter “a”, extracted from an X-ray diffractometry (XRD) peak analysis, was incorporated into

Table 1. Magnetic and lattice parameters and SAR values for all samples tested at 300 K.

Sample	HC (Oe)	MS (emu/g)	SAR in Agar @ 800 Oe (W/g)	Lattice Parameter (a) of Fe3O4 (Å)
30 nm Nanocubes [6]	33.7	67.5	~540	Not Applicable
AP	57.5	67.6	544	8.27
AO	109	61.2	1001	8.32
AN	67.1	74.3	692.8	8.37
AH	64.8	77.6	624.4	8.40
AA	80	73.8	712.3	8.37

for the primary phase Fe₃O₄, illustrating the evolution of the crystal parameters [22,23]. The as-prepared (AP) sample in the inset of Figure 1a displays a uniform distribution of nanocubes, showing the presence of ferrimagnetic (FiM), magnetite (Fe₃O₄), and antiferromagnetic (AFM) hematite (α-Fe₂O₃), as seen in the XRD patterns. When the AP nanocubes were annealed in O₂, the appearance of distinctive (113) and (024) peaks indicated an increase in the volume fraction of the α-Fe₂O₃ phase. The exposure to N₂ removed a significant amount of the α-Fe₂O₃ volume fraction; this change signified a change in the oxidation number from 3 to 8/3 as the α-Fe₂O₃ reduced to Fe₃O₄ [24]. This reduction in the oxidation number confirms the ability of N₂ to act as a reducing agent at high temperatures and pressures [24]. The reaction, which has been reported for hematite nanowires when annealed for 1 h at 350 °C, seemed to be initiated at 300 °C in our iron oxide nanocubes when they were annealed for 2 h [25]. Here, neither inert gas showed a significant phase change from the AP nanocubes. Though significant morphological changes were absent in all samples post-annealing, a chain formation was observed in the annealed samples which is understood to be due to dipole–dipole interactions between the NCs with the annihilation of the surfactants [26]. This chain formation can be advantageous for hyperthermia treatments as it tends to increase the specific absorption rate (SAR) values with the enhancement of the anisotropic interactions [20,21,27,28,29].

3.2. Magnetic Properties

Temperature-dependent magnetization, M(T), measurements were carried out between 10 and 300 K following the zero-field cooled (ZFC), field-cooled cooling (FCC), and field-cooled warming (FCW) protocols in the presence of a 500 Oe magnetic field. The Verwey transition (VT), a first-order metal–insulator transition in which the crystalline phase changes from a high-temperature cubic structure to low-temperature monoclinic structure indicative of the Fe₃O₄ phase, is observed in all samples except the sample annealed in O₂ (AO) at temperatures of 109 K ≤ T_V ≤ 121 K [30,31,32]. The sharpness of the transition which translates the crystallinity of the Fe₃O₄ phase can be observed to increase with annealing in the presence of N₂ (AN), He (AH), and Ar (AA), with the exception of the AO sample. The AO sample is understood to be oxidized in multiple stages simultaneous with the final products of α-Fe₂O₃, following the reaction [33,34].

The inside of the structure is expected to be oxidized to a γ-Fe₂O₃-like (or γ-Fe₂O₃) phase from Fe₃O₄ which results in the disappearance of the VT [33,35,36,37,38]. The absence of the VT observed in prior annealing on iron oxide nanorods by Attanayake et al. supports the oxidation to a γ-Fe₂O₃-like phase [35,39] The AN, AH, and AA, which indicated the reduction in the α-Fe₂O₃ phase indicated by the XRD data and the absence of the Morin transition (MT), are observed to show the transformation at comparative temperatures in the presence of various mixtures of gases, in our case, N₂, He, and Ar [40,41]. Relative to the AP sample, the MT feature became less pronounced in the annealed samples due to the increased crystallinity of the Fe₃O₄ phase. Furthermore, the AA indicates a slightly higher decrease in α-Fe₂O₃ which can be observed in the absence of (018) in Figure 1e and the sharpness of the VT in Figure 2d compared to AH.

The first-order magnetic transition, MT, commonly known as the temperature-driven spin–flop transition, is associated with the presence of the α-Fe₂O₃ phase which occurs as the spin alignment changes from perpendicular to the c-axis above the TM to parallel to the c-axis below the TM; this is observable only in the AP sample [42,43,44,45,46,47].

The magnetization vs. applied field M(H) measurements shown in Figure 3a were carried out at room temperature with a maximum 10 kOe applied field for all samples. The magnetization at the maximum value of the applied magnetic field (M_S) decreased when moving from AP to AO but improved in all other samples as the FiM Fe₃O₄ phase (Figure 3b). Though the M_S value decreased in the AO sample, the value did not significantly decrease, as observed with the sole α-Fe₂O₃-carrying nanostructures which have less than 10 emu/g in most cases since the sample undergoes a two-step oxidation, as explained by Zheng et al., which leaves a significant volume fraction of the FiM Fe₃O₄ phase intact [33,48,49]. Relative to bulk Fe₃O₄ with an M_S of 92 emu/g, the decrease in the M_S observed for its nanoparticle counterpart is commonly attributed to a reduction in long-range magnetic ordering and the enhancement of surface spin disorder resulting from the reduction in particle size to the nanoscale [50]. The samples did not portray superparamagnetic features, similar to the observations made by Attanayake et al., but displayed an inversely proportional relationship between the M_S and coercivity (H_C) which can be related to the Stoner–Wohlfarth theory, H_K = (2K_Ani)/(μ₀M_S), which shows that the magnetocrystalline anisotropy of a single-domain particle is expected to show an inversely proportional relationship between M_S and H_K; thus, we see a change in H_C in Figure 3b [16,51,52]. The general increment in H_C in all the annealed samples may be due to the magnetic hardening of the nanostructures with annealing; this is because the magnetic phases, especially the FM/FiM phases, become more stable, with annealing making them more resistant to demagnetization; additionally, defect formation with annealing can lead to magnetic pinning, a well-known defect-engineering mechanism [53,54]. An H_C of above 100 Oe was only registered in AO, understandably due to the combination of the effects: magnetic hardening, defect formation, and increased magnetic anisotropy [5,54,55]. The tunability of the AP sample in terms of the M_S and H_C with varying gas types opens up the possibility for nanostructures to be explored for the effectiveness of magnetic hyperthermia measurements with enhanced H_C and M_S separately, further tuned in accordance to varying applications such as magnetic particle imaging and magnetic tracers [53].

Low-temperature M(H) measurements at 10 K (Figure 4) showed higher M_S values compared to the 300 K typical for nanostructures as the superficial spins tend to become better aligned with the applied magnetic field at lower temperatures as the thermal energy is lower; this has been observed in smaller nanostructures with high surface-to-volume ratios, which follows Kneller’s law that predicts an inversely proportional relationship between H_C and T [56,57]. The ZFC and FC protocols applied on the nanostructures did not yield an exchange bias effect at the FiM Fe₃O₄ and AFM α-Fe₂O₃ interfaces, which is understood to be due to negligible interfacial interactions due to frozen spins [16,36]. The pinning effect is observed in all samples and is the smallest in the AO sample (Figure 4b), which may be due to the formation of a γ-Fe₂O₃-like phase similar to the observations made with iron oxide nanorods annealed with oxygen [36].

The H_C with and without the application of the 10 kOe field shows the separation with the observed biphases (Figure 5a). The ZFC protocol which lets the spins be randomly oriented while being cooled down to a low temperature will possess a higher H_C as the antiferromagnetic exchange coupling will pin the ferrimagnetic moments at the interface, but with the application of the FC protocol, the ferrimagnetic phase will align, minimizing the interfacial pinning with the assistance of the cubic structure. The AO sample is the only sample that does not show a significant difference between the said two protocols with its dominant γ-Fe₂O₃-like phase which does not possess regular biphasic behavior in magnetic measurements. A similar observation is made in the magnetization curves in Figure 5b with AO.

3.3. Magnetic Hyperthermia

Figure 6a–c depicts the usage of samples in hyperthermia measurements, namely, heating T(t) curves, at varying fields for a 300 s window, which is a relatively short period that has been understood to be comparatively more effective in conjunction with radiation therapy at moderately lethal temperatures which are higher than the general hyperthermia window of 40–44 °C [58,59,60,61]. Additionally, longer treatment periods are also associated with patient fatigue, adjacent healthy tissue destruction, and changes in body condition which can lead to complex reactions, adversely affecting the health and well-being of the patient [62]. Nanoparticle introduction was kept at 1 mg/mL due to the tendency of instinctive reactions to lead to adverse conditions irrespective of compatibility [63,64]. In the in vitro study, to ensure the performance of the nanoparticles, a 2% by weight agar, a medium denser than water, was used to ensure an environment that was on par with or closer to the environment inside the body [7,65]. The SAR value of the heating curves was calculated using the initial slope of the heating curve, using the following equation;

$$SAR=\frac{\Delta T}{\Delta t}\frac{C_P}{\phi }$$

(1)

The method devised assumes a homogenous sample temperature and negligible heat loss at certain time intervals at the start of the alternating magnetic field [66,67]. Here, ΔT/Δt gives the rate of the change in temperature, and C_P indicates the specific heat capacity of the liquid medium. Here, the agar, at 2% by weight, is assumed to have not significantly changed its specific heat capacity, which is thus 4.186 J/(g °C); the heat capacity of water is used. The symbol φ indicates the mass of the magnetic material per unit mass of liquid, which essentially gives the concentration used in the measurements.

It can be seen in Figure 6d that all the samples in general showed significant improvements in heating potential and SAR values compared to the AP sample, while the AO sample showed an exceptional enhancement. The general enhancement of heating capabilities can be due to defect enhancement and increased grain size, which lead, in turn, to increased anisotropy [53,68]. The AO sample showed the highest SAR value in all the applied magnetic fields, with the highest of 1001 W/g at 800 Oe (Figure 6d and Table 1), which is almost four times higher than that reported in FeO/Fe₃O₄ nanocubes synthesized by Khurshid et al. [18]. Further, when compared with FiM iron oxide nanocubes of a similar size synthesized by Nemati et al. which showed almost the same magnetic properties compared to the AP sample (refer Table 1), annealing was observed to have significantly increased the SAR in all the samples [6]. This suggests and re-affirms the already understood phenomena of SAR enhancement with an increasing M_S and H_C [69,70,71]. In our study, 34 nm nanocubes were used, eliminating variability in size and shape parameters. The adjusted annealing environments between the AO and AN samples clearly show that the M_S and H_C played a pivotal role in altering the SAR, but the enhanced H_C in the AO sample is observed to lead to a higher SAR value compared to the improved M_S in the AN sample. Table 1 summarizes the changes in the H_C, M_S, SAR, and the lattice parameter—and in all the samples—and highlights the exceptional SAR value observed in the AO sample which possessed the highest H_C, underscoring the general relation between H_C and SAR which supersedes the effect of the M_S. It is also worth mentioning that the lattice parameter (a) of sample AH, which shows the highest M_S of 77.6 emu/g, matches the value of Fe₃O₄ (8.40 Å) [23,72,73]. This highlights a correlation between the structure and magnetism of the Fe₃O₄ phase.

4. Conclusions

The phase-tunability of 34 nm iron oxide nanocubes with different types of gases enabled the enhancement of different fundamental magnetic properties, such as coercivity, saturation magnetization, and crystallinity. The enhancement of the specific absorption rate in all the samples with annealing, along with the general increment in coercivity, reflected how magnetic anisotropy can affect magnetic hyperthermia efficiency. Comparing the annealing-induced changes in coercivity and saturation magnetization, which are well-known parameters in tuning the specific absorption rate, it is observed that the former plays a larger role as the values in all the three tested applied magnetic fields were significantly higher in the AO sample with enhanced H_C. Additionally, we were able to establish a relation between the crystallinity of the structure and the enhancement of the saturation magnetization via the lattice parameter, indicating changes in the crystal structure. At low temperatures, we also observed the pinning effect in the samples.