Improvements in the Organic-Phase Hydrothermal Synthesis of Monodisperse M_xFe_3–xO₄ (M = Fe, Mg, Zn) Spinel Nanoferrites for Magnetic Fluid Hyperthermia Application

Abstract

In the quest for optimal heat dissipaters for magnetic fluid hyperthermia applications, monodisperse M_xFe_3–xO₄ (M = Fe, Mg, Zn) spinel nanoferrites were successfully synthesized through a modified organic-phase hydrothermal route. The chemical composition effect on the size, crystallinity, saturation magnetization, magnetic anisotropy, and heating potential of prepared nanoferrites were assessed using transmission electron microscopy (TEM), dynamic light scattering, X-ray diffraction (XRD), thermogravimetric analysis (TGA), energy-dispersive X-ray spectroscopy (EDS), atomic absorption spectroscopy (AAS), X-ray photoelectron spectroscopy (XPS), and vibrating sample magnetometer (VSM) techniques. TEM revealed that a particle diameter between 6 and 14 nm could be controlled by varying the surfactant ratio and doping ions. EDS, AAS, XRD, and XPS confirmed the inclusion of Zn and Mg ions in the Fe₃O₄ structure. Magnetization studies via VSM revealed both the superparamagnetic nature of the nanoferrites and the dependence on substitution of the doped ions to the final magnetization. The broader zero-field cooling curve of Zn-doped Fe₃O₄ was related to their large size distribution. Finally, a maximum rising temperature (T_max) of 66 °C was achieved for an aqueous ferrofluid of nondoped Fe₃O₄ nanoparticles after magnetic field activation for 12 min.

1. Introduction

Magnetic fluid hyperthermia (MFH) utilizing iron oxide nanoparticles (Fe₃O₄ NPs) is a quickly evolving technology in medical oncology as clinically it is a minimally invasive cancer therapy. In this scenario, activation of an externally applied alternating magnetic field (AMF) switches the magnetic moments of Fe₃O₄ NPs rapidly. This results in heat dissipation in the tumor zone and subsequent necrosis of the tumor.¹ The heat dissipation potency of a clinical Fe₃O₄ fluid is defined by the specific absorption rate (SAR) or specific loss power. The heating potential of commercial fluid at quantities of ∼112 mgFe/mL within a physiologically safe range of AC magnetic field (f = 100 kHz, and H ∼ 2–18 kA/m) is insufficient for complete elimination of a tumor.² This low heating potential is hypothesized to be due to the intrinsic low saturation magnetization (M_s) value of nanosized Fe₃O₄ particles (∼50–60 emu/g lower than that of bulk Fe₃O₄ ∼85–100 emu/g) and a large surface spin disorder of spherical-shaped Fe₃O₄ NPs currently used for magnetic resonance imaging and MFH applications.³ It has been proposed that compositional tuning through a metal dopant substitution of Fe²⁺ with a M²⁺ cation in the tetrahedral (T_d) or octahedral (O_h) interstitial sites of an Fe₃O₄ crystal lattice is a viable and potential nanoscale engineering strategy to synthesize metal-doped spinel M_xFe_3–xO₄ (M = Fe, Mg, Zn) nanoferrites with enhanced heating potential.^4,5 Of all the proposed classical synthetic protocols thus far, thermal decomposition of metal acetylacetonate precursors M(acac)₃ in the presence of surfactants and capping/reducing agents at elevated temperature (>300 °C) has manifested itself as an efficient route to ensure the synthesis of an arsenal of high-quality and monodisperse hydrophobic single domain Fe₃O₄ and metal-doped spinel M_xFe_3–xO₄ nanoferrites.^6,7 This method provides excellent control over the nucleation and growth steps as well as the mean size, shape, and composition. Consequently, various M_xFe_3–xO₄ nanoferrites have been routinely prepared for potential technological applications.^8,9 In spite of these advantages, the thermal decomposition method does face some impediments. These include the high toxicity and cost of the iron precursors and reagents, the utilization of flammable organic solvents at high temperatures (320 °C), a requirement to use an inert atmosphere (N₂) during the reaction, and additional steps to transfer the resulting nanocrystals to the aqueous phase.^10,11 Hence, there is a need for the community to explore more simple and economically robust protocols to synthesize highly monodispersed water-soluble nanocrystals in mass quantities. As a cost-effective and environmentally friendly synthetic approach under mild conditions, the hydrothermal route shows promise, as it encourages the growth of single crystals with definite sizes, shapes, and narrow polydispersity in an autoclave at temperatures (in the range from 130 to 250°C) and vapor pressure (ranging from 0.3 to 4 MPa) through a one-step simple facile reaction scheme.^12,13 Metal nanoferrites of various shapes and size configurations have been synthesized by the hydrothermal synthesis route via either aqueous phase or organic-phase reactions of metal precursors with or without surfactants for potential technological applications.¹⁴ Typically, the decomposition of iron precursors in an aqueous solution with or without surfactants results in aggregated clusters of larger size distribution. For example, a polyethylene glycol (PEG)–Fe₃O₄ nanodisc,¹⁵ ethylene glycol (EG)–Fe₃O₄ nanoflowers,¹⁰ dodecyltrimethylammonium bromide–Fe₃O₄ NPs,¹⁶ Fe₃O₄/C core–shell nanorings with EG/PEG,¹⁷ and PEG–Mn_0.2Ni_0.8Fe₂O₄¹⁸ have all been synthesized utilizing this method. In contrast to the hydrolytic synthesis of nanoferrites, only a limited number of reports have highlighted the organic-phase hydrothermal synthesis of nanocrystals. For example, Tian et al. fabricated ultrasmall (4–6 nm) monodispersed Fe₃O₄ via the hydrothermal method, with Fe(acac)₃ as the iron source, n-octanol as the solvent, and n-octylamine as the reductant.¹⁹ In another example, Dendrinou-Samara et al. have synthesized nanoferrites NiFe₂O₄,²⁰ CoFe₂O₄,²¹ and MnFe₂O₄²² with the surfactant oleylamine (OAm) acting as a stabilizing agent and solvent simultaneously. They observed that the prepared nanocrystals still presented some degree of aggregation. In this contribution, we examine the extent to which a hybrid approach can be applied where the desired properties of thermal decomposition method can be achieved using the more convenient mild hydrothermal route. We demonstrate this by the successful synthesis of highly monodispersed M_xFe_3–xO₄ (M = Fe, Mg, Zn) nanocrystals of mixed morphologies for use in MFH applications. In addition, the effect of transition-metal doping into the crystal lattice was tested with the objective of improving crystallinity and magnetization of bare Fe₃O₄ NPs.

2. Results and Discussion

2.1. Structural and Compositional Studies

M_xFe_3–xO₄ (M = Fe, Mg, Zn) spinel nanoferrites were synthesized by the decomposition of the metal precursors in a high boiling point solvent, octadecene (ODC), under hydrothermal conditions. In a systematic fashion, the molar ratio of oleic acid (OA) to OAm, amount of tri-n-octylphosphine oxide (TOPO), and reaction time were investigated. Feedback based on size and shape of the particles was obtained using transmission electron microscopy (TEM). In terms of the molar ratios of OA to OAm, individual utilization of each surfactant with 0.5 mmol of TOPO at 240 °C for 120 min resulted in aggregated clusters; therefore, a strategy utilizing both surfactants was adopted, and it was found that an equal amount of OA to OAm of 1:1 minimized the aggregation to some extent. Eventually, the OA to OAm of 1:4 and 1:5 resulted in particles of great uniformity and size distribution (Figure S1 in Supporting Information). Keeping the other conditions constant, the influence of TOPO variation (0.1, 0.3 and 0.5 mmol) was monitored on the monodispersity and size of the Fe₃O₄ NPs. TOPO typically functions as a particle stabilizer where the phosphine oxide periodically binds to the Fe₃O₄ NPs (at the Fe atom), controlling the growth rate of particles under the synthetic conditions employed.²³ Eventually, NPs with appropriate uniformity and shape were obtained utilizing an optimized TOPO concentration of 0.5 mmol (Figure S2). The effect of time on the morphological evolution of the formation of Fe₃O₄ NPs was also recorded at different reaction times of 30, 60, 90, and 120 minutes with a molar ratio of OA to OAm of 1:4 and TOPO concentration of 0.5 mmol at 240 °C for 120 minutes. An aliquot was withdrawn from the reaction medium after the above indicated time periods of the reaction had elapsed and was imaged by TEM. As depicted in Figure S3, the particles start growing and forming well-shaped particles as the time is extended. It is postulated that the particles are generated by the coalescence and reshaping of small particles where two or more particles merge during the reaction to form single daughter-shaped NPs. In summary, a molar ratio of OA to OAm of between 1:4 and 1:5, a concentration of 0.5 mmol TOPO, and a reaction time of 120 min at 240 °C were found to be the optimal conditions to synthesize a library of monodisperse NPs with narrow size distributions. The Fe1 NPs with an OA to OAm ratio of 1:4 clearly show great uniformity and size distribution with an average size of 10.3 ± 2.8 nm (Figure 1a,b,e). The TEM image reveals the assembly of NPs into a close-packed arrangement without any interfacial contact and aggregation. Chemisorption of surfactants can potentially control the nucleation growth and negate the aggregation; nevertheless, the hydrophobic interaction between the tail groups of surfactants adsorbed on the NPs encourages the interdigitating of NPs near to each other.²⁴ Interestingly, a further increase of OA to OAm to a 1:5 ratio resulted in smaller sized NPs (Fe2) with an average size of 6.1 ± 1.1 nm (Figure 1c,d,f). It has been shown previously that increasing the relative amount of the OAm surfactant results in smaller nonuniform particles, no doubt caused by the extra amount of OAm, which adsorbs onto the surface of nuclei, preventing crystal growth, which favors the generation of small units.²⁵

Figure 1.

TEM images of the synthesized Fe1 (a,b) and Fe2 (c,d) NPs at different magnifications of 100 and 25 nm (TOPO 0.5 mmol, time 120 min at 240 °C). Insets: size distribution of Fe1 (e) and Fe2 (f) with mean size and standard deviation value (σ).

Encouraged by the initial results, we extended this methodology to synthesize cation-substituted magnetite-doped M_xFe_3–xO₄ (M = Zn, Mg) NPs.

First, zinc was introduced into the Fe₃O₄ lattice in order to obtain uniform monodispersed NPs with increased magnetization. It has been shown by Cheon et al.²⁶ that (Zn_xFe_1–x) Fe₂O₄ (x = 0, 0.1, 0.2, 0.3, 0.4, and 0.8) NPs with a single crystallinity phase and size monodispersity could be synthesized from the thermal decomposition of iron and zinc precursors in the presence of surfactants. Their prepared NPs exhibited high and tunable nanomagnetism with a maximum M_s value of 161 emu/g for the formulated (Zn_0.4Fe_0.6)Fe₂O₄ nanoparticle.²⁶ Our initial attempt began with a dopant of 0.4 mmol zinc utilizing a convenient hydrothermal method. As evidenced by the subsequent TEM images, we were able to synthesize high-quality particles comparable to that of the prepared particles by Cheon et al.

It can be seen from the corresponding TEM images of ZnFe1 NPs (OA to OAm of 1:4) (Figure 2a,b,e) and ZnFe2 NPs (OA to OAm of 1:5) (Figure 2c,d,f) that a variety of morphologies are present, including spherical, triangular, cubic, and octahedral shapes with average particle sizes of 14.1 ± 4.1 and 12 ± 7 nm for ZnFe1 and ZnFe2, respectively. The shape evolution process has been delineated in terms of surface free energy. The small particles are unstable because of their high surface free energy; therefore, in the process of growth, unconsolidated primary small particles reconstitute into different geometries with more stable structures and less surface free energy.²⁷ In the case of ZnFe2 NPs, an interesting bimodal distribution of small predominantly spherical NPs and larger particles from the growth of the small NPs can be seen, which confirms the growth, prohibiting the effect of OAm as discussed for Fe1. Notably, the average size of Zn-doped NPs is higher than that of bare Fe₃O₄ NPs which could be due to the substitution of larger radius Zn atoms instead of Fe atoms in the iron oxide structure.^28,29

Figure 2.

TEM images of the synthesized ZnFe1 (a,b) and ZnFe2 (c,d) NPs at different magnifications of 100 and 25 nm (TOPO 0.5 mmol, time 120 min at 240 °C). Insets: size distribution of ZnFe1 (e) and ZnFe2 (f) with mean size and standard deviation value (σ).

Bae et al.² synthesized magnesium-doped iron oxide NPs with different Mg²⁺-ion concentrations using the thermal decomposition method. The resulting particle achieved a high temperature value of 180 °C at a biologically safe magnetic induction field with a Mg concentration of 0.13 mmol (Mg_0.13–Fe₃O₄) using a MFH protocol. We therefore doped 0.13 mmol Mg into Fe₃O₄ NPs with our developed hydrothermal route in order to achieve monodisperse NPs with higher magnetization and heating efficiency. The resulting particles show uniform distribution with low interparticle distance without aggregation. From their corresponding TEM images, MgFe1 NPs (OA to OAm of 1:4) (Figure 3a,b,e) and MgFe2 NPs (OA to OAm of 1:5) (Figure 3c,d,f) present excellent monodispersity with average particle sizes of 6.8 ± 1.9 and 6.3 ± 1.5 nm, respectively.

Figure 3.

TEM images of the synthesized MgFe1 (a,b) and MgFe2 (c,d) NPs at different magnifications of 100 and 25 nm (TOPO 0.5 mmol, time 120 min at 240 °C). Insets: size distribution of MgFe1 (e) and MgFe2 (f) with mean size and standard deviation value (σ).

The size distribution of nanoferrites was also measured using dynamic light scattering (DLS). All of the nanoferrites physically clumped together to form bigger size nanocrystals of up to a hundred nanometers because of van der Waals and magnetic dipole–dipole attractions.³⁰ The volume particle size distribution and polydispersity index (PdI) values of the M_xFe_3–xO₄ (M = Fe, Mg, Zn) nanoferrites are presented in Figures S4–S9. Fe1 exhibited the narrowest relative size distribution (d_h ∼ 125 nm and PdI = 0.16) which could lead to a high magnetic and heating response.^31,32 Overall, the Mg- and Zn-doped Fe₃O₄ NPs form larger aggregates than that of the undoped variants (d_h ∼ 214.7 nm and 189.9 for ZnFe2 and MgFe2, respectively). This is an undesirable characteristic for their biomedical utilization, as the large clusters are believed to be removed from the bloodstream more easily and can also cause embolism.^30,33

The X-ray diffraction (XRD) patterns, average crystallite size, and unit cell parameters [d-spacing of lattice planes (hkl), and lattice constant (a) values] of nanoferrites extracted utilizing Debye–Scherrer and Bragg equations are shown in Figure 4 and Table 1.

Figure 4.

(a) Powder XRD patterns and the (b) highlighted (311) diffraction peak of M_xFe_3–xO₄ (M = Fe, Mg, Zn) nanoferrites.

Table 1. Calculated Values of Size and Lattice Parameters for M_xFe_3–xO₄ (M = Fe, Mg, Zn) Spinel Nanoferrites.

nanoferrite	size (nm) TEM	crystallite size (nm) XRD	position of 311 peak in degree (θ)	d, lattice spacing(Å)	a, lattice constant (Å)	organic content from TGA (%)
Fe1	10.3 ± 2.8	4.8 ± 0.89	35.64	2.5170	8.3481	46.39
Fe2	6.1 ± 1.1	3.2 ± 1.5	35.62	2.5184	8.3527	68.29
MgFe1	6.8 ± 1.9	4.6 ± 1.1	35.64	2.5170	8.3481	56.04
MgFe2	6.3 ± 1.5	3.1 ± 1.2	35.62	2.5184	8.3527	52.93
ZnFe1	14.1 ± 4.1	7.4 ± 0.87	35.52	2.5253	8.3754	58.02
ZnFe2	11.1 ± 6.4	6.2 ± 1.3	35.4	2.5336	8.4029	61.35

The XRD patterns of the Fe₃O₄ NPs present the crystallographic planes of (111), (220), (311), (400), (422), (511), and (440) corresponding to the diffraction peaks at 2θ of 18.50, 30.1, 35.6, 43.1, 53.2, 57.2, and 63°, respectively (Figure 4a). The positions and relative intensities of all diffraction peaks indexed well to the standard cubic spinel structure (JCPDS no. 71-1232).^34,35 Lattice parameters were calculated for all M_xFe_3–xO₄ (M = Fe, Mg, Zn) nanoferrites by consideration of the strong Bragg reflection peak (Miller index 311) (Table 1). The average crystallite size estimated from the Debye–Scherrer equation was 4.8 ± 0.89 and 3.2 ± 1.5 for Fe1 and Fe2, respectively. The achieved lattice parameters for Fe1 and Fe2 NPs were 8.34 and 8.35 Å, respectively, matching well with a = 8.35 Å for nanosized Fe₃O₄ NPs.

The Mg_0.13Fe₃O₄ nanoferrites exhibited similar values for the diffraction peaks (2θ) corresponding to the (111), (220), (311), (400), (422), (511), and (440) crystal planes, which confirm the formation of a pure MgFe₂O₄ phase with JCPDS (card no. 36-0398).³⁶ It is assumed that the replacement of Fe²⁺ ions (atomic radius = 0.64 Å) with Mg²⁺ ions of similar atomic radius (0.65 Å) may not induce any peak shifts.³⁷ For instance, the lattice parameter for MgFe2 was 2.5184 Å, similar to that of 2.5184 Å for Fe2 (Figure 4b). The average crystallite sizes estimated from the Debye–Scherrer equation were 4.6 ± 1.1 and 3.1 ± 1.2 for MgFe1 and MgFe2, respectively.³⁸ Similar to the above Mg_0.13Fe₃O₄ nanoferrites, diffraction peaks (2θ) corresponding to the (111), (220), (311), (400), (422), (511), and (440) crystal planes were also observed in the XRD pattern of Zn_0.4Fe₃O₄, confirming the formation of a pure zinc nanoferrite (JCPDS card no. 89-1397).³⁸

The introduction of Zn into the Fe₃O₄ structure resulted in sharper peaks, indicating an improvement in crystallinity. This was accompanied by the (311) crystal plane of the Zn-doped Fe₃O₄ shifting to slightly smaller angles (from 2θ = 35.64° for Fe1 to 35.52° for ZnFe1) (Figure 4b). The average crystallite sizes estimated from the Debye–Scherrer equation were 7.4 ± 0.87 and 6.2 ± 1.3 for ZnFe1 and ZnFe2, respectively. Furthermore, the corresponding lattice constant (a) and d-spacing of lattice planes (d) increased from 8.3481 to 8.3754 Å and 2.5170 to 2.5253 Å, respectively. Substitution of larger Zn²⁺ ions (atomic radius of 0.74 Å) with an Fe X+ ion [Fe²⁺ ion (atomic radius = 0.64 Å) and Fe³⁺ ion (atomic radius = 0.49 Å)] in an Fe₃O₄ crystal is thought to cause the expansion in length of the a-axis of the crystals unit cell. This has been observed previously.^39,40 Overall, there is a discrepancy in size measurements from XRD, TEM, and DLS techniques. The hydrodynamic size (d_h) obtained from DLS and mean size (D_h) obtained from TEM are larger than that of the crystallite size (D) measured by XRD. This could be ascribed to the different medium in which NPs were measured. In the case of DLS, solution aggregation may be responsible for the enhanced sizes,⁴¹ whereas nanocrystal stacking may account for the source of error in TEM imaging.

In order to determine the thermal stability and organic fraction of the nanoferrites, thermogravimetric analysis (TGA) was performed under a N₂ flow with a heating rate of 10 °C/min from room temperature up to 800 °C. The mass percentage of the residue reflects the fraction of inorganic cores inside the nanoferrites. As can be seen from the thermograms, Fe1, Fe2, MgFe1, and MgFe2 nanoferrites exhibit similar profiles with four stages of mass loss (Figure 5). The mass loss includes below 200 °C (evaporation of water or organic solvents),⁴² 200–300 °C (decomposition of free surfactants/capping molecules adsorbed on the surface),⁴³ 300–450 °C (decomposition of directly attached surfactants),⁴⁴ and 650–700 °C corresponding to the reduction of the inorganic core under an inert atmosphere.⁴⁵ For the Zn-doped Fe₃O₄ NPs, the mass loss for inorganic core reduction was a two-stage plateau between 600–650 and 700–750 °C. Similar results have also been observed for OA-capped zinc nanoferrites.⁴⁶

Figure 5.

(a,b) TGA curves of M_xFe_3–xO₄ (M = Fe, Mg, Zn) nanoferrites.

Importantly, the content of organic molecules on the nanoferrite surface decreases with increased magnetic core size (Table 1). This can be ascribed to the smaller surface-to-volume ratio and subsequent lower active binding sites on the surface of bigger cores, which discourage the adsorption of surfactants.⁴⁷

The elemental composition and atomic percentage (at. %) of monodispersed M_xFe_3–xO₄ (M = Fe, Mg, Zn) nanoferrites were qualitatively evaluated with energy-dispersive X-ray spectroscopy (EDS) analysis. The EDS spectra were visualized through a line scan of several randomly selected areas under different magnifications for each nanoferrite. All nanoferrite NPs presented very strong signals corresponding to C atoms which were ascribed to the carbon chains of the surfactants/capping agents on the NP surface and the coated carbon to increase the conductivity of samples before imaging.⁴⁸ The characteristic peaks for Fe and O were detected in the EDS spectra of Fe1 and Fe2 NPs. The EDS spectra of MgFe1 and MgFe2 exhibit the signals for Fe, Mg, and O atoms. In the same fashion, the representative signals for Fe, Zn, and O atoms were detected in the EDS spectra of ZnFe1 and ZnFe2. It should be noted that two EDS spectra of similar NPs exhibit very similar atomic percentages (at. %). For instance, the at. % of the Mg atoms were 0.50 and 0.48% for MgFe1 and MgFe2, respectively. A similar trend was obtained for Fe and Zn (Figures S10–S12). This confirms the chemical uniformity and homogeneity of the composition of the nanoferrites. The intensity of the Fe signal decreases in the case of Zn- and Mg-doped NPs. This decrement in intensity is more pronounced for Zn, which is in good agreement with the initial molar ratio of metal precursors undertaken for the synthesis of nanoferrites with 1/0.4 mmol for Fe/Zn in contrast to 1/0.13 mmol for Fe/Mg-doped Fe₃O₄ NPs. These results are an indication of utility of the substitutional feature of the hydrothermal method for the doped Fe₃O₄ NPs. Kolen’ko et al. observed the same trend by doping Zn in the Fe₃O₄ NPs.⁴⁹ The elemental atomic ratios of Mg, Zn, and Fe and actual chemical composition of M_xFe_3–xO₄ nanoferrites were quantified by atomic absorption spectroscopy (AAS) following the procedure as reported by Pellegrino et al.³ (Table 2). Taking into account the AAS results, the elemental atomic ratios and the chemical compositions from theoretical assumptions deviates from experimental values so that all doped nanoferrites except ZnFe1 present compositional deficiency. This deviation from theoretical stoichiometry has also been observed in Co- and Mn-doped Fe₃O₄ NPs.^50,51 Different thermal stabilities of ions (Fe³⁺, Co²⁺, Mn²⁺) are supposed to be an influential contributor. Additionally, the excessive portion of the surfactant might decrease the decomposition temperature of the metal complex and affect the growth mechanism. In addition, the reductive nature of the reaction medium may encourage the partial reduction of iron(III) to iron(II) which would compete with the doped ions in the growth process.⁵¹

Table 2. Molar Ratios and Chemical Formula for M_xFe_3–xO₄ (M = Mg, Zn) Nanoferrites Found by AAS.

	molar ratio		chemical formula
nanoferrites	theoretical	experimental	theoretical	experimental
MgFe1	0.13:1 Mg/Fe	0.13:1.8 Mg/Fe	Mg0.35Fe2.65O4	Mg0.2Fe2.8O4
MgFe2	0.13:1 Mg/Fe	0.06:1.7 Mg/Fe	Mg0.35Fe2.65O4	Mg0.1Fe2.9O4
ZnFe1	0.4:1 Zn/Fe	0.48:1.2 Zn/Fe	Zn0.85Fe2.15O4	Zn0.85Fe2.15O4
ZnFe2	0.4:1 Zn/Fe	0.45:0.82 Zn/Fe	Zn0.85Fe2.15O4	Zn0.77Fe2.23O4

X-ray photoelectron spectroscopy (XPS) was conducted to gain insight into the chemical compositions and metal (M²⁺, Fe³⁺) valance states of M_xFe_3–xO₄ (M = Fe, Mg, Zn) nanoferrites. Wide-scan survey spectra of the nanoferrites revealed photoelectron lines related to C 1s, O 1s, Fe 2p, Zn 2p, Mn 2p, and Mg 1s, as illustrated in Figures 6 and S13–S17. The binding energy scale was calibrated utilizing the C 1s signal (originating from adventitious hydrocarbon) at 285 eV as an energy reference. For each nanoferrite, the high-resolution narrow-scan XPS spectra of the related elements were also recorded.

Figure 6.

XPS spectra of the Fe1 nanoferrite (a) survey scan, (b) C 1s, (c) O 1s, and (d) Fe 2p regional scans.

The C 1s XPS spectra of all nanoferrites revealed a single peak centered at 285 eV arising from C–C bonds from organic molecules adsorbed on the surface⁵² and adventitious carbon from exposure of the samples to the air.⁴⁸ The O 1s core-level spectra of all nanoferrites exhibit a main peak at ∼532 eV corresponding to −COO– carboxylate groups⁴⁸ and a second signal centered at ∼530 eV corresponding to the lattice oxygen of metal–oxygen bonds (M–O) of M_xFe_3–xO₄.^53,54

Typically, Fe³⁺ ions at the O_h site of Fe₃O₄ exhibit the spin–orbit split doublet of Fe 2p_1/2 and Fe 2p_3/2 at 724 and 711 eV,⁴⁸ whereas the Fe 2p_3/2 peak at ∼707 and ∼709 eV verifies the existence of Fe in either 0 and +2 oxidation states.⁵⁵ Additionally, γ-Fe₂O₃ will exhibit a small satellite peak at 718–719 eV between the 2p_1/2 and 2p_3/2 peaks, which is ascribed to the Fe 2p XPS spectrum of Fe³⁺ ions.⁵⁶ Taking into account the electronic state of Fe³⁺ and Fe²⁺ ions, the high-resolution regional XPS spectra of Fe 2p were recorded for all nanoferrites, as shown in Figures 6 and S13–S17. Each Fe 2p peak reflects additional satellite peaks at higher binding energies because of the possible excitation of an unpaired electron (from a 3d orbital) to a higher bound energy level (4s orbital line). The fine scan of the Fe 2p region in the spectra for the nondoped Fe₃O₄ NPs (Fe1 and Fe2), reveals two peaks centered at 711.1 eV corresponding to the Fe 2p_3/2 peak of Fe³⁺ ions at the O_h site and 724.5 eV corresponding to the Fe 2p_1/2 peak of Fe³⁺ ions at the T_d site. The observed binding energies match well with literature values for magnetite formation. Notably, the Fe 2p spectra of the Fe2 and MgFe1 NPs do not exhibit any diagnostic peaks, indicating the absence of Fe on their surface.

In the case of Mg-doped Fe₃O₄ (MgFe1 and MgFe2), the absence of peaks in the high-resolution Mg 1s photoelectron spectra indicates the absence of magnesium on the surface. Additionally, the Fe 2p peaks identify an oxidation state of 3+ for iron, which is in line with the XRD results, confirming the formation of magnesium ferrite (MgFe₂O₄).^5700,5800

For the Zn-doped Fe₃O₄ nanoferrites (ZnFe1), the presence of two major peaks at 711 and 724 eV corresponding to Fe 2p_3/2 and Fe 2p_1/2, respectively, rules out the possibility of a 3 + oxidation state for Fe. The Zn 2p core-level XPS spectrum reveals two typical peaks at binding energies of 1021 and 1045 eV corresponding to Zn 2p_3/2 and Zn 2p_1/2 photolines, respectively. The values match well with the reported literature values for a 2+ oxidation state for the Zn ions, verifying the formation of zinc ferrite ZnFe₂O₄^57,58

2.2. Magnetization Studies

To identify the magnetic nature of the synthesized nanoferrites, magnetization measurements were recorded as a function of magnetic field (M–H) and temperature (M–T) with a vibrating sample magnetometer (VSM). Magnetization curves as a function of the applied magnetic field (M–H loops) were collected from −20 to 20 kOe magnetic field strength at room temperature (∼300 K) (Figure 7). From these scans, the related parameters such as remanence (M_r), coercivity (H_c) and normalized remanence (M_r/M_s) values were calculated (Table 3). The prepared NPs exhibit a Langevin-like approach to magnetization saturation, with negligible coercivity and remanence as observed in the hysteresis loops shown in Figure 7 and Table 3. This negligible remanence and coercivity can be ascribed to the particle size distributions.⁵⁹ The inclusion of Mg²⁺ and Zn²⁺ ions in the spinel structure of Fe₃O₄ does not alter its superparamagnetic characteristic which is advantageous for biomedical applications.⁶⁰ Looking into the magnetization values, a reduction in magnitude was observed when the ration of surfactants (OA/OAm) increased from 1:4 to 1:5. OA and OAm were used as surfactants to stabilize the particles and reduce the aggregation through steric repulsion. It has been shown that coating a nonmagnetic layer on their surface can significantly reduce the magnetic moment as a fraction of total mass.⁶¹

Figure 7.

(a) Magnetic hysteresis loops of M_xFe_3–xO₄ (M = Fe, Mg, Zn) nanoferrites at room temperature. (b) Magnified view of the hysteresis loop of nanoferrites in low magnetic field, as indicated by the dashed box in (a). (c) Inset shows the magnetic response of ZnFe2 to an external magnetic field.

Table 3. Parameters of ZFC/FC and M–H Hysteresis Loops for M_xFe_3–xO₄ (M = Fe, Mg, Zn) Nanoferrites.

Fe3O4 nanoferrites	Ms (emu/g)	Mr (emu/g)	Hc (Oe)	Ms/Mr	TB	keff (J·m–3) × 104
Fe1	31.6	13	48	2.4	260	15.6
MgFe1	7.5	2.1	37	3.5	95	28.7
ZnFe1	10.7	4.5	14	2.3	50	1.2
Fe2	18.6	6.3	168	2.9	48	13.9
MgFe2	10.7	1.3	109	8.2	120	31.6
ZnFe2	16.1	7.2	43	2.2	128	6.2

The saturation magnetization of NPs can be influenced by their particle size and composition. For the iron oxide NPs (Fe1 and Fe2), there is a correlation between size (TEM) and magnetization values where the magnetization decreases with a decrease in the particle size. This trend has been observed in some other ferrites and is ascribed to the surface spin canting effect.^62,63 Another influential contributor is the control distribution of dopant ions into the Fe₃O₄ unit cell for maximal magnetization.⁶⁴ The magnetization value for Fe1 was 31.6 emu/g (Figure 7, black line). Doping Zn ions (0.4 mmol) into Fe₃O₄ reduces the magnetization to 10.7 emu/g (Figure 7, pink line). Similarly, doping Mg ions (0.13 mmol) resulted in reduction in magnetization to 7.5 emu/g (Figure 7, green line). The magnetization value for Fe2 was measured to be 18.6 emu/g (Figure 7, dark blue line). In a similar fashion, doping Zn ions (0.4 mmol) and Mg ions (0.13 mmol) into Fe₃O₄ reduced the magnetization to 16.2 emu/g (Figure 7, orange line) and 10.6 emu/g (Figure 7, light blue line), respectively. The utilized concentration of 0.4 and 0.13 mmol for Zn and Mg resulted in reduction in magnetic moment values which were different from those reported in the literature.^2,26 These results can be explained by consideration of the Fe₃O₄ NP atomic structure. Fe₃O₄ is also represented as Fe³⁺. Fe₃O₄/Fe²⁺O₄ is an inverse spinel ferrite (AB₂O₄ type) where Fe³⁺ ions occupy the tetrahedral site (T_d) and Fe³⁺/Fe²⁺ ions have occupied octahedral interstices (O_h) in a ratio of 1:1.^4,65 The magnetic moments of the Fe³⁺ ions at the T_d and O_h sites are aligned in opposite directions and cancel each other. Accordingly, the magnetic moment of Fe²⁺ ions (4 μB) in the O_h sites determine the net magnetization of the Fe₃O₄ NPs. Both Zn²⁺ and Mg²⁺ ions are diamagnetic with zero magnetic moment (0 μB). Considering the reduced magnetization, we can safely assume that both Zn²⁺ and Mg²⁺ ions have occupied the T_d sites with the Fe³⁺ ions occupying the O_h sites, leading to antiferromagnetic coupling interactions of Fe³⁺ atoms between these sites which decrease the magnetization moment.⁶⁶

The temperature dependence of magnetization (M–T curves) was plotted following the field cooling (FC) and zero field cooling (ZFC) protocols between 10 and 350 K under a constant magnetic field of 10 Oe. The M–T curves, extracted blocking temperature (T_B) and magnetic crystalline anisotropy energy (K_eff) of all nanoferrites are shown in Figure 8 and Table 3.

Figure 8.

FC/ZFC curves of the M_xFe_3–xO₄ (M = Fe, Mg, Zn) nanoferrites recorded at a constant magnetic field of 10 Oe.

Utilizing Néel law, K_eff = 25k_BT_B/V_m, the magnetic anisotropy constant (K_eff) was calculated for all nanoferrites, where k_B is Boltzmann’s constant (1.3807 × 10^–23 J/K) and V_m is the volume of a single nanocrystal assuming sphere as V_m = πD_m³/6 (D_m is the magnetic core diameter from TEM).^8,67 For all nanoferrites, there is a peak in the ZFC measurements, indicating a blocking behavior, with the blocking temperature T_B well below room temperature (300 K). FC/ZFC curves converged at temperatures higher than the blocking temperatures (T > T_B). These are signs of superparamagnetic-like behavior. ZFC measurements typically exhibit a qualitative indication of the size distribution of the particles and T_B refers to the maximum of the ZFC magnetization plot. For an ideal system of noninteracting monodispersed particles, T_B would be one single temperature. For a collection of particles with narrow size distribution, the ZFC curve would present a sharp peak. However, for a collection of particles with a large size distribution, there would be a distribution of the T_B so that the ZFC curve maxima would be broad.⁶⁸ In our case, the Zn-doped nanoferrites exhibited a broader ZFC curve. This is consistent with the TEM images where the standard deviation (σ) values of the particle size are larger because of the large size distribution.⁶⁹

2.3. Magnetic Hyperthermia Studies

The heating potential of prepared nanoferrites was determined in water through a calorimetric nonadiabatic setup. With consideration toward clinical applications, the product of H and f should be below the threshold limit of H × f = 5 × 10⁹ A m^–1 s^–1. In our experiments, the field configuration was 1 order of magnitude higher than the required safety limit (1.7 × 10¹⁰ A m^–1 s^–1); however, other research groups have also considered the same order of magnitude as the threshold limit for their applications.^30,70

The heating profile of samples is presented in Figure 9a. As can be seen from the curve, the temperature increases monotonically and saturates after a certain time. This rapid rise is correlated to Néel relaxation and Brownian rotation of particles under a switching magnetic field.⁵⁹ The temperature saturation is due to temperature loss from the magnetic fluid to the environment.⁷² The nanoferrites differ both in the initial speed of the temperature rise and the time taken to reach the hyperthermic threshold temperature (43 °C). For instance, Fe1 NPs reach the temperature (45 °C) within 3 min, before all other nanoferrites. Additionally, the maximum rising temperature (T_max) was 66 °C for Fe1 NPs after activation of the magnetic field for 12 min compared to 50 °C for ZnFe1. It should be noted that the concentration of nanoferrites in our study (5 mg/mL) is lower than other published accounts and significantly lower than the recognized clinical concentration of ferrofluids at 112 mg Fe/mL by Jordan et al.⁷³⁰⁰ Accordingly, an increased temperature in a shorter time might be feasible with a higher concentration.⁷³ The heat dissipation rate of the ferrofluids (termed as SAR) was quantified by the initial slope method from heating curves for all nanoferrites and is summarized in Table 4. The maximum SAR value was 87 W/g for Fe1 NPs (Figure 9b). Because SAR values are usually obtained under different extrinsic magnetic field parameters (H and f), it is not feasible to directly compare the heating potential of similar ferrofluids in terms of size and chemical composition. Accordingly, SAR values were normalized to the intrinsic loss power (ILP) as an AC field-independent parameter to allow the direct comparison of the heating efficiency of prepared nanoferrites in our work with the available literature data for similar ferrofluids (Figure 9c). The ILP values of the aqueous dispersions of the prepared nanoferrites listed in Table 4 span from 0.02 to 0.06 nH m² kg^–1 which are smaller than reported values in the literature for nanoferrites with a similar average core size and composition. For instance, ILP values of 3.8 and 1.75 nH m² kg^–1 have been reported for Fe₃O₄ NPs.⁷⁴ Kusigerski et al.⁷⁷ achieved an ILP value of 0.57 nH m² kg^–1 for Mg_0.1Fe_2.9O₄ NPs, which is slightly higher than that of the Mg_0.13Fe_2.87O₄ nanoferrites prepared by our group.⁷⁵⁰⁰ Kallumadil et al. investigated the heating potentials of commercially available iron oxide ferrofluids supplied by different companies with different iron contents, nanocrystallinity, and hydrodynamic diameter (d_h). Calorimetric measurements recorded at 900 kHz with a field amplitude of 5.66 kA/m revealed ILP values ranging from 0.15 to 3.1 nH m² kg^–1 (Resovist, ILP ≈ 3.1, d_h: 60 nm, carboxy dextran–Fe₃O₄).⁷⁵⁰¹ Later studies have reported ILP values higher than that of commercial ones. These include ILP of 4.1 nH m² kg^–1 by Thanh et al.⁷⁵ (citric acid–Fe₃O₄, d_h: 141 nm), ILP of 5.6 nH m² kg^–1 by Pellegrino et al.⁷⁶ (Fe₃O₄ nanocubes, d_h: 19 nm), and ILP of 6.1 nH m² kg^–1 by Parkin et al.⁷⁷ (tiopronin–Fe₃O₄, d_h: 135 nm). Several intrinsic features such as particle size and size distribution,⁷⁸ geometry,⁷⁹ chemical composition, magnetocrystalline anisotropy,⁸⁰ saturation magnetization, concentration, agglomeration state,⁸¹ and dipole–dipole interactions^82,83 affect the heating performance of the ferrofluids. In the case of our nanoferrites, Fe1 has the highest heating potency because of lower aggregation (DLS), lower surfactant fraction on the surface (TGA), and a higher magnetization value (VSM). Considering the simultaneous contribution of these factors to the final heating outputs, it is however rather complex to interpret the heating outputs of Mg- and Zn-doped nanoferrites. ZnFe2 has a higher heating efficiency than MgFe2 because of higher magnetization and lower surfactant fraction on the surface. On the other hand, MgFe1 has a higher heating efficiency than ZnFe1 because of lower surfactant fraction on the surface; however, the magnetization is lower. It is worth noting that there is no full consensus on the influence of the aggregation state (dipolar interactions) on heating efficiency of fluids because of controversial results reported in the literature.⁸⁴⁻⁸⁶ It has been shown that the dipolar interactions significantly impair the heat dissipation process because of the disturbed magnetization relaxation time.⁸⁷ Conversely, superior heating performance (5 orders of magnitude) has been reported for Fe₃O₄ clusters compared to randomly distributed NPs.⁸⁸ In the present study, the potential aggregation for Mg- and Zn-doped nanoferrites is larger than that for nondoped Fe₃O₄ NPs (DLS). This has translated to the lower magnetization (VSM) and heating performance of the doped versus undoped nanoferrites. The mechanism by which the heat can be created was explored further. Upon exposure of the NPs to an AMF, heat is dissipated into the medium through a variety of different pathways depending on their magnetic profile and size, either by relaxation loss (Néel or Brown spin relaxations for a superparamagnetic regime size <30 nm) or by hysteresis loss (for a ferromagnetic regime size >30 nm).⁸⁹ In our case, hysteresis loss does not dominate because of the superparamagnetic behavior of our nanoferrites with negligible hysteresis⁹⁰ (Figure 7). Therefore, it appears that magnetic relaxation loss is the greatest contributor to the heat generation. For a single domain noninteracting magnetic particle, the particle can generate heat through two related mechanisms, either Néel or Brown relaxations. Néelian relaxation is the energy loss in the form of heat by an internal rotation of individual magnetic moments within the particle. Brownian relaxation pertains to the energy loss in the form of heat by the physical rotation of the particle itself under AMF.⁹¹ The Brownian (τ_B) and Néel (τ_N) relaxation times of a single particle assuming as sphere can be calculated utilizing the following formulae

1

2

where τ_N = Néel relaxation time, τ₀ = effective relaxation time (∼10^–9 s), K = effective anisotropy constant, D_m is the magnetic core diameter from TEM, k_B = Boltzmann constant (1.38 × 10^–23 J/K), T = the absolute temperature in kelvin, τ_B = Brownian relaxation time, η = dynamic viscosity of the surrounding medium (η = 0.7978 × 10^–3 Pa·s for water), and D_h is the hydrodynamic diameter of the particle from DLS measurements.⁹² Considering that these two mechanisms take place in parallel but independently, the effective relaxation time τ_eff is given by

3

when τ_N ≫ τ_B or τ_N ≪ τ_B, τ_eff is minimized. Utilizing the corresponding eqs 1–3, τ_N, τ_B, and τ_eff were calculated for all nanoferrites (Table 4). An examination of the data reveals that τ_N ≪ τ_B for all nanoferrites except for Fe1 which results in τ_eff ≈ τ_N. This implies that the Néel relaxation is the dominant mechanism whereby rotation of the magnetic moments inside the particles contribute to the heat generation. However, for Fe1, τ_B ≪ τ_N, which suggests that Brownian relaxation is the dominant pathway. This coexistence behavior has been observed before.^93,94

Figure 9.

(a) Heating curves of water-dispersed M_xFe_3–xO₄ (M = Fe, Mg, Zn) nanoferrites at a field amplitude of 114.01 mT. (b) SAR and (c) ILP values obtained from these curves.

Table 4. Relaxation Times and Heating Parameters for M_xFe_3–xO₄ (M = Fe, Mg, Zn) Nanoferrites.

Fe3O4 nanoferrites	τB	τN	Τeff	SAR (W/g)	ILP (nH m2 kg–1)
Fe1	2.3 × 10–4	2.6 × 101	2.3 × 10–4	87.8	0.055
MgFe1	2.1 × 10–2	9.6 × 10–5	9.6 × 10–5	46.0	0.028
ZnFe1	3.3 × 10–3	6.4 × 10–8	6.4 × 10–8	36.2	0.022
Fe2	1.2 × 10–5	5.3 × 10–8	5.3 × 10–8	50.2	0.031
MgFe2	9.1 × 10–3	2.3 × 10–5	2.3 × 10–5	60.8	0.038
ZnFe2	6.2 × 10–2	4.1 × 10–5	4.1 × 10–5	69.7	0.043

3. Conclusions

A series of monodispersed M_xFe_3–xO₄ (M = Fe, Mg, Zn) spinel nanoferrites were successfully synthesized using the hydrothermal method. By monitoring the resulting NPs utilizing TEM, a molar ratio of surfactants OA/OAm of 1:4, a reaction time of 120 min, and a temperature of 240 °C were found to be the optimum conditions to create high-quality nanoferrites of different shapes and sizes. The corresponding unit cell parameters were constant after Mg²⁺ ion doping, but the inclusion of Zn²⁺ ions resulted in the expansion of the crystal unit of the pure Fe₃O₄. TGA and XPS results confirmed a high concentration of carbon atoms present on the surface of the nanoferrites. The composition experimental values, as determined from AAS results, are consistent with the EDS elemental results, but these both differ from the chemical compositions obtained from theoretical assumptions such that all doped nanoferrites except for ZnFe1 are compositionally deficient. The magnetization values and heating potential of naked Fe₃O₄ were decreased by the inclusion of Mg²⁺ and Zn²⁺ ions. The content of organic molecules on the nanoferrite surface decreased with increased magnetic core size as evidenced by TGA results. Néel spin relaxation was found to be the dominant mechanism for heat production.

4. Experimental Section

4.1. Materials

Iron(III) acetylacetonate (Fe(acac)₃ ≥ 99.9% trace metals basis), OAm (≥ 70%), OA (90%), TOPO (99%), ODC (90%), tetramethylammoniumhydroxide (20% w/w), and magnesium acetate (Mg(CH₃COO)₂) were purchased from Sigma-Aldrich and used without further purification. Zinc chloride (ZnCl₂) was purchased from Ajax Finechem. Milli-Q water was used after first being filtered through a 0.22 μm pore size hydrophilic filter with a resistivity of 18.2 MΩ cm from Millipore. All other chemicals were of analytical grade and used as received from commercial sources without further purification.

4.2. Synthesis of Monodisperse M_xFe_3–xO₄ (M = Fe, Mg, Zn) Spinel Nanoferrites

Monodispersed Fe₃O₄ NPs at various set surfactant ratios and reaction times were developed by a simple one-step hydrothermal route. In a typical procedure, Fe(acac)₃ (1 mmol), TOPO (0.5 mmol), OA (0.64 mL), and OAm (2.56 mL) were mixed in octadecene (20 mL) under stirring (500 rpm) at 100 °C for 60 min. Afterward, Ar (g) was bubbled into the solution for 2 min to remove the air, and the mixture was then transferred into a 100 mL polytetra-fluoroethylene-lined autoclave. The autoclave was sealed and maintained at 200 °C for 30 min and then heated to 240 °C for 120 min. After this time, the reaction was deemed complete, and the autoclave was left on the bench to cool slowly to RT over 3 h. Upon addition of ethanol (10 mL), black NPs were precipitated from the solution and isolated by centrifugation. Consecutively NPs were dispersed in hexane for further use. Using the methodology of Cheon et al. and Dang et al., a series of monodisperse M_xFe_3–xO₄ (M = Fe, Mg, Zn) nanoferrites were synthesized.^2,26 A summary of the experimental ratios employed is given in Table 5.

Table 5. Synthesis Conditions of M_xFe_3–xO₄ (M = Fe, Mg, Zn) Spinel Nanoferrites.

sample	OA/OAm	OA (mL)	OAm (mL)	Fe(acac)3 (mmol)	ZnCl2 (mmol)	Mg(CH3COO)2 (mmol)	TOPO (mmol)
Fe3O4 (Fe1)	1:4	0.64	2.56	1	0.0	0.0	0.5
Fe3O4 (Fe2)	1:5	0.64	3.2	1	0.0	0.0	0.5
Mg0.13Fe3O4 (MgFe1)	1:4	0.64	2.56	1	0.0	0.13	0.5
Mg0.13Fe3O4 (MgFe2)	1:5	0.64	3.2	1	0.0	0.13	0.5
Zn0.4Fe3O4 (ZnFe1)	1:4	0.64	2.56	1	0.4	0.0	0.5
Zn0.4Fe3O4 (ZnFe2)	1:5	0.64	3.2	1	0.4	0.0	0.5

4.3. Instrumentation and Measurements

Structural characterization, chemical composition, and magnetic features of the newly synthesized nanoferrites were probed with TEM (Tecnai G2 Spirit Bio-TWIN), powder XRD (Rigaku Spider X-ray diffractometer), DLS (Zetasizer; Nano ZS, Malvern), energy-dispersive X-ray connected scanning electron microscopy (SEM–EDX) (FE-SEM FEI Quanta), AAS (AAS-9000 spectrometer, Shimadzu), thermogravimetric analyzer (TA Instruments Q500) XPS (Kratos Axis Ultra^DLD), and VSM (Quantum Design P935A USA, physical property measurement system). The heating potential of nanoferrites was assessed with a commercialized magnetic alternating generator (Ambrell EASYHEAT, 2.4 kW, 196–197 kHz). More details regarding characterization are described in the Supporting Information.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.0c01641.

• More details regarding structural characterization and magnetic hyperthermia measurements of the synthesized M_xFe_3–xO₄ (M = Fe, Mg, Zn) nanoferrites, TEM images of the synthesized Fe₃O₄ NPs at different reaction parameters, volume particle size distribution of M_xFe_3–xO₄ (M = Fe, Mg, Zn) nanoferrites, EDX patterns of Fe1, Fe2, MgFe1, MgFe2, ZnFe1, and ZnFe2 nanoferrites, and XPS spectra of the Fe2, MgFe1, MgFe2, ZnFe1, and ZnFe2 nanoferrites (PDF)

The authors declare no competing financial interest.