Facile synthesis of epsilon iron oxides via spray drying for millimeter-wave absorption

Abstract

A simple, scalable spray drying method was developed for high-yield epsilon iron oxide (ε-Fe₂O₃) synthesis. The ε-Fe₂O₃ particle size can be tailored by varying the annealing temperature and molar ratio of Fe/Si, producing a high-purity ε-phase. This strategy also enables ferromagnetic resonance tuning, making it potentially usable in millimeter-wave absorbers.

---

High-coercivity oxide magnets have been intensively developed for various applications, such as permanent magnets, information storage systems, electromagnetic interference (EMI) shielding devices, and magnetic recording media.^1–3 In particular, epsilon iron oxide (ε-Fe₂O₃) has garnered considerable attention because of its high coercive field of more than 20 kOe at room temperature, which is 3–4 times higher than those of ferrites and similar to those of rare-earth permanent magnets.^4,5 The large coercive field of the ε-phase is attributed to the orthorhombic structure with a large magnetocrystalline anisotropy (7.7 × 10⁶ erg·cm⁻³), which is in contrast to the other iron oxide polymorphs (α and γ-Fe₂O₃), which have isotropic crystal structures.^4,6 In addition, ε-Fe₂O₃ exhibits electromagnetic (EM) wave absorption at a high frequency (∼182 GHz) because the zero-field ferromagnetic resonance is proportional to its strong magnetic anisotropy field.^7,8 The substitution of nonmagnetic elements with Fe sites enables tunable absorption in the millimeter-wave region (30–300 GHz).^8–10

Thus, several studies have been conducted to produce high-purity ε-Fe₂O₃. The ε-phase can be obtained in nanoparticle (NP) form rather than bulk form because it is an intermediate between γ-Fe₂O₃ and α-Fe₂O₃.^3,4,11 Therefore, a crucial requirement for obtaining single-phase ε-Fe₂O₃ is to control the particle size in a specific range (8–40 nm) that makes the ε-phase stable.⁵ Most commonly, pure ε-Fe₂O₃ NPs are fabricated by the reverse-micelle and sol–gel method,^12–14 rapid sol-gel method,¹⁵ and impregnation of mesoporous templates with Fe salts.^3,4 These approaches prevent the excessive growth of crystallites by using a silica matrix, surfactant, or porous template, resulting in a high-purity ε-phase. However, such methods involve multiple time-consuming steps, produce low yields, and require expensive sacrificial templates, which limit large-scale production and practical applications. Therefore, a facile and scalable technique for producing high-purity ε-Fe₂O₃ should be developed.

Spray drying has been widely used for the industrial manufacturing of dried fine powders for pharmaceuticals, foods, and chemicals.¹⁶ This process involves instant and continuous evaporation of aerosol droplets, enabling the production of powders in a single step without additional washing.¹⁶ Furthermore, this approach has no limit in terms of the combination of multiple compounds in particle form;¹⁶ this makes the synthesis of composites with dispersed elemental Fe and Si possible without the use of templates or surfactants. Therefore, this method has great potential as a high-yield technique for producing pure ε-phase. To the best of our knowledge, there have been no reports on the synthesis of ε-Fe₂O₃via spray drying.

In this study, a simple and scalable method was developed for the synthesis of high-purity ε-Fe₂O₃via spray drying. Single-phase ε-Fe₂O₃ NPs were fabricated by controlling the annealing temperature and precursor molar ratio. The ferromagnetic resonance characteristics of the resulting iron oxide were adjusted in the millimeter-wave region by simple substitution with nonmagnetic metal ions. The results provide a basis for the industrial production of magnetic NPs as millimeter-wave absorbers.

Fig. 1(a) shows a diagram illustrating the spray-assisted synthesis of high-purity ε-Fe₂O₃. We performed the spray drying of a solution mixture of Fe salt (Fe(NO₃)₃·9H₂O) and a silica precursor (tetraethoxysilane, TEOS) in a hot chamber, followed by post-heat treatment and sequential washing. As shown in Fig. 1(b) and (c), the as-spray-dried microspheres exhibit hollow spherical structures induced by drying the ethanolic solvent mixture.¹⁶ Post-annealing of the precursor particles in air induces the formation of iron oxide NPs in the hollow microspheres, as shown in the field-emission TEM images (Fig. 1(d) and (e)). Fe₂O₃ NPs are obtained after subsequent washing to remove the silica matrix (Fig. 1(f) and (g)).

Fig. 1 (a) Schematic illustration of the preparation of ε-Fe₂O₃ NPs via spray drying. (b) and (c) Scanning electron microscopy (SEM) images of as-spray-dried precursor particles with a precursor molar ratio (Fe/Si) of 0.4 : 1. (d) and (e) Field-emission transmission electron microscopy (FE-TEM) images of ε-Fe₂O₃ NPs embedded in SiO₂ particles after annealing at 1180 °C for 4 h. (f) and (g) TEM images of the corresponding ε-Fe₂O₃ NPs after SiO₂ removal.

Since the ε-phase is stable only in certain size regimes,⁴ the annealing temperature and ratio between the Fe and silica precursor were varied to produce pure ε-Fe₂O₃. Fig. 2(a) and Fig. S1(a) (ESI†) provide representative X-ray diffraction (XRD) patterns and phase fractions of the obtained powders (Fe₂O₃@SiO₂). For the sample annealed in air at 930 °C for 4 h, the obtained pattern indicates only the γ-phase of Fe₂O₃ (ICDD No. 00-025-1402). As the temperature increases to 1050 °C, the intensity of the (122) peak of the ε-phase (ICDD No. 00-016-0653) significantly increases, and the (311) peak of the γ phase begins to decrease. A pure ε-Fe₂O₃ phase is observed at 1180 °C in Fig. S1(a) (ESI†). However, the powders treated at 1240 °C contain ε-Fe₂O₃ (59.4 wt%) and α-Fe₂O₃ (40.6 wt%) (ICDD No. 00-33-0664), which implies that the ε-phase partially transforms into α-Fe₂O₃ when the heating temperature increases. In addition, the obtained particle size significantly increases from 7.76 ± 1.7 nm to 56.77 ± 29 nm as the temperature increases (Fig. S2(a), ESI†). These results demonstrate that the Fe₂O₃ synthesized in this study displays a γ → ε → α phase transformation pathway as the heating temperature and particle size increase, which is analogous to the results of a previous study.¹¹

Fig. 2 XRD patterns of ε-Fe₂O₃ NPs@SiO₂ (a) annealed in air at various temperatures for 4 h using precursor particles with a precursor molar ratio (Fe/Si) of 0.4 : 1 and (b) prepared using various precursor molar ratios (Fe/Si) under the optimized annealing conditions (1180 °C, 4 h) and (c) and (d) magnetic hysteresis loops of ε-Fe₂O₃ NPs in (a) and (b) after SiO₂ removal.

We changed the Fe/Si molar ratio of the precursor under optimized heat-treatment conditions (1180 °C, 4 h). Fig. 2(b) and Fig. S1(b) (ESI†) demonstrate that powders with an Fe/Si molar ratio of ∼0.15 comprise 14.0 wt% γ-Fe₂O₃ and 86.0 wt% ε-Fe₂O₃. As the ratio increases to 0.4, ε-Fe₂O₃ with a phase fraction of ∼100 wt% is obtained. Above 0.6, sharp increases in the intensities of the (104) and (110) peaks of α-Fe₂O₃ and a decrease in the intensity of the (122) peak of ε-Fe₂O₃ are observed. Fig. S2(b) (ESI†) also demonstrates the dependence of the particle size on the Fe/Si molar ratio. As the Fe/Si molar ratio increases, the particle size increases from 20.41 ± 3.8 nm to 57.54 ± 28 nm. The standard deviation of the size also increases for molar ratios greater than 0.60, owing to further formation of α-Fe₂O₃. The particle sizes obtained for high-purity ε-Fe₂O₃ (avg. 34.63 ± 11 nm) are within the size range of 8–40 nm that is desired to form a single ε-phase, as reported previously.⁴ Based on the results, the optimized synthetic conditions (Fe/Si molar ratio = 0.4, annealing temperature = 1180 °C) were fixed to obtain a high-purity ε-phase, which was quantitatively analyzed using ⁵⁷Fe Mössbauer spectra. The four sextets for Fe sites corresponding to ε-Fe₂O₃ primarily exhibit one doublet for α-Fe₂O₃ (Fig. S3, ESI†).⁶ The relative areas of ε-Fe₂O₃ and α-Fe₂O₃, which were obtained from the fitted Mössbauer parameters, were calculated to be 92.7% and 7.3%, respectively (Table S1, ESI†).

Thus, high-purity ε-Fe₂O₃ could be synthesized via an efficient aerosol-assisted method. To prepare the precursor solution, TEOS was mixed in an ethanolic mixture containing iron nitrate, which was hydrolyzed into silicic acid in a low-pH solution.¹⁷ Rapid evaporation of the clear solution during spraying induced the instant formation of microparticles consisting of elemental Fe and Si. In the annealing step, the polymerization of hydrolyzed TEOS occurs first, creating an amorphous silica network in the microspheres.¹⁸ Therefore, it was speculated that the Fe salt embedded in the densified silica underwent spatial confinement during heating. The excessive growth of iron oxide in the sphere could be restricted, causing the formation of the ε-phase. These results show that spray drying can be an efficient means of producing high-purity ε-Fe₂O₃ without the use of surfactants or sacrificial templates in a single step. Commonly, the duration of ε-Fe₂O₃ synthesis depends on the preparation of the composite of Fe salt and silica matrix. In particular, the time for hydrolysis and condensation of TEOS determines the time to form a silica matrix, which has taken 2–20 h in previous studies.^9,15 In contrast, spray drying involves the immediate evaporation of droplets combined with the continuous feeding of a precursor solution, which enables the reduction of the synthetic time with high yields. Furthermore, compared to the previously reported wet reaction conducted in a batch reactor,^9,14 this aerosol-assisted method enables continuous production of dried composite particles without washing or post-drying. Thus, this strategy provides an efficient and scalable manufacturing means of synthesizing ε-Fe₂O₃.

To elucidate the effects of the annealing temperature and precursor molar ratio on the magnetic properties, a magnetic property measurement system (MPMS) analysis of the obtained NPs after the removal of the silica matrix was performed. The M–H curves of the samples and corresponding coercivity (H_c), remanence (M_r), and maximum magnetization under the maximum applied field (M_5.0T) are presented in Fig. 2(c) and (d) and Table S2 (ESI†). The samples treated at 930 °C exhibit soft magnetic behavior of γ-Fe₂O₃ NPs with an average diameter of 7.76 ± 1.7 nm.¹⁹ As the heating temperature increases to 1050 °C, a distinctive kink in the magnetic hysteresis loop (Fig. 2(c)) appears because of the coexistence of the ε- and γ-phases. However, the oxide particles annealed at 1180 °C exhibit the typical magnetic behavior of ε-Fe₂O₃ with an exceedingly high H_c of more than 22 kOe. Above 1180 °C, M_5.0T decreases with nearly stable H_c, which is related to the formation of the α phase.

The resulting powders prepared using various precursor molar ratios exhibit magnetic behavior depending on γ → ε → α phase transformation (Fig. 2(d)). The kink in the M–H curve observed at an Fe/Si ratio of 0.15 results from the soft magnetic γ phase with a fraction of ∼14 wt% (Fig. S1(b), ESI†). As the Fe/Si ratio increases to 0.4, the kink disappears and hard magnetic behavior is observed with increasing H_c (over 22 kOe) and decreasing M_5.0T, resulting from complete phase transformation from the γ phase to the ε-phase. As the molar ratio increases from 0.40 to 0.80, hard magnetic behavior is observed with decreasing M_5.0T owing to α-phase development. Based on the results, under the same heating conditions, control of the Fe/Si ratio enables modulation of the purity and magnetic properties of ε-Fe₂O₃, which originates from the size-driven phase transition of the ε-phase.^4,15 The resulting ε-Fe₂O₃ prepared under the optimized conditions shows similar magnetic properties (H_c = 22.4 kOe and M_5.0T = 14.3 emu g⁻¹) to those reported for ε-Fe₂O₃ (Table S3, ESI†),⁹ which indicates that this synthetic method can reproduce the excellent magnetic properties of ε-Fe₂O₃.

It is well known that ε-Fe₂O₃ with a large coercive field absorbs high-frequency EM waves because the zero-field ferromagnetic resonance is dependent on the magnetic anisotropy field.^7,8 In particular, nonmagnetic substitution of Fe enables the adjustment of the coercivities of ε-Fe₂O₃,^20,21 resulting in tunable ferromagnetic resonance from 30 GHz to 230 GHz.^9,10 Thus, to provide a proof-of-concept of the millimeter-wave absorber, we prepared Ga-substituted ε-Fe₂O₃ (ε-Ga_xFe_2−xO₃, 0.24 ≤ x ≤ 0.48) via a spray drying method. All magnetic oxides formed a high-purity ε-phase, as shown in the XRD patterns in Fig. S4 (ESI†). The magnetic properties of the ε-Ga_xFe_2−xO₃ powders are presented in Fig. 3(a) and Table S4 (ESI†) for x = 0.24, 0.32, 0.42, and 0.48. H_c decreases from 22.3 kOe (x = 0) to 5.4 kOe (x = 0.48) as x increases. In addition, M_5.0T increases from 14.2 emu g⁻¹ (x = 0) to 26.3 emu g⁻¹ (x = 0.42) and then decreases to 23.5 emu g⁻¹ (x = 0.48). These results are in accordance with those of a previous study (Table S3, ESI†), showing that substituting ε-Fe₂O₃ with nonmagnetic ions adjusts the magnetic properties.^8,9 The absorbing composite films were prepared by mixing ε-Fe₂O₃ NPs and thermoplastic urethane (TPU) at a ε-Ga_xFe_2−xO₃ : TPU weight ratio of 8 : 2 (Fig. S5; see ESI†). EM absorption spectra and permeability values in the range 60–110 GHz were measured for the films using a vector network analyzer. Strong absorption of the ε-Ga_xFe_2−xO₃ composite films is observed in Fig. 3(b). The resonant frequencies (f_r) of the magnetic composites shift from 105 GHz to 71 GHz with increasing Ga substitution (105 GHz (x = 0.24), 93 GHz (x = 0.32), 82 GHz (x = 0.42), and 71 GHz (x = 0.48)). These findings are in accordance with those of previous studies, demonstrating that nonmagnetic substitution changes the magnetocrystalline anisotropy, resulting in tunable ferromagnetic resonance characteristics.^7–10 The corresponding permeability and permittivity values of the ε-Ga_xFe_2−xO₃ composite film are presented in Fig. 3(c) and Fig. S6 (ESI†).

Fig. 3 (a) Magnetic hysteresis loops of ε-Ga_xFe_2−xO₃ (0.24 ≤ x ≤ 0.48) powders. (b) Millimeter-wave absorption spectra normalized by film thickness. (c) Real (μ′) and (d) imaginary (μ′′) parts of magnetic permeability obtained from ε-Ga_xFe_2−xO₃-based TPU composite films. The EM absorption properties were measured under zero magnetic field at room temperature. In the x = 0.24 sample, the absorption peak exceeds the measurement range.

A sudden decrease in the real permeability (μ′) and spike in the imaginary permeability (μ′) are observed. μ′′ exhibits the maximum value at each resonant frequency, which is attributed to natural resonance.²⁰ The magnetic loss tangent describes the power lost versus the power stored (tan δ_m = μ′′/μ′), which implies energy dissipation in each composite film.²² The loss tangent increases from 0 to approximately 0.2 in the ε-Fe₂O₃ films (Fig. 3(d)), which means that high absorption of EM waves occurs at each f_r. Therefore, magnetic iron oxide synthesized by spray drying has great potential as a millimeter-wave absorber for use in automotive radar systems, military applications, and next-generation wireless communication.

In this study, a simple and scalable method of synthesizing high-purity (92.7%) ε-Fe₂O₃ NPs with a large coercive field (∼22 kOe) was successfully developed. This synthesis approach based on rapid spray drying of the precursor solution enables the spatial confinement of Fe salts in silica microspheres, which restricts excessive crystal growth during the annealing step. The particle size required to form the ε-phase can be easily tuned by controlling the precursor molar ratio and annealing temperature. Thus, high-purity ε-Fe₂O₃ particles can be obtained via a simple, fast, and scalable process without utilizing surfactants or porous templates. Additionally, the use of a simple and continuous process to produce precursor powders can considerably reduce the duration of the synthetic process. Furthermore, the ability to absorb millimeter waves was demonstrated using the ε-Fe₂O₃ composite film. This synthetic strategy is expected to pave the way for industrial applications of ε-Fe₂O₃ such as permanent magnets, magnetic recording media, information storage, and millimeter-wave absorbers for use in high-speed wireless communication.

This research was supported financially by the Fundamental Research Program (PNK 8310) of the Korea Institute of Materials Science and the National Research Foundation of Korea (#NRF-2020M3H4A2084780).

Conflicts of interest

There are no conflicts to declare.

Notes and references

1. A. H. Lu, E. L. Salabas and F. Schüth, Angew. Chem., Int. Ed., 2007, 46, 1222–1244.
2. R. M. Cornell and U. Schwertmann, The Iron Oxides: Structure, Properties, Reactions, Occurences and Uses, John Wiley & Sons, 2003.
3. A. Namai, M. Yoshikiyo, K. Yamada, S. Sakurai, T. Goto, T. Yoshida, T. Miyazaki, M. Nakajima, T. Suemoto, H. Tokoro and S. Ohkoshi, Nat. Commun., 2012, 3, 1–6.
4. S.-i Ohkoshi, A. Namai, K. Imoto, M. Yoshikiyo, W. Tarora, K. Nakagawa, M. Komine, Y. Miyamoto, T. Nasu, S. Oka and H. Tokoro, Sci. Rep., 2015, 5, 14414.
5. J. Jin, S.-i Ohkoshi and K. Hashimoto, Adv. Mater., 2004, 16, 48–51.
6. J. Tucek, R. Zboril, A. Namai and S.-i Ohkoshi, Chem. Mater., 2010, 22, 6483–6505.
7. C. Kittel, Phys. Rev., 1947, 71, 270–271.
8. A. Namai, M. Yoshikiyo, S. Umeda, T. Yoshida, T. Miyazaki, M. Nakajima, K. Yamaguchi, T. Suemoto and S. Okoshi, J. Mater. Chem. C, 2013, 1, 5200–5206.
9. S. Ohkoshi, S. Kuroki, S. Sakurai, K. Marsumoto, K. Sato and S. Sasaki, Angew. Chem., Int. Ed., 2007, 46, 8392–8395.
10. Y. Gu, N. J. O. Silva, M. Yoshikiyo, A. Namai, R. Pinol, G. Maurin-Pasturel, Y. Cui, S. Ohkoshi, A. Millan and A. Martinez, Chem. Commun., 2021, 57, 2285–2288.
11. S. Sakurai, A. Namai, K. Hashimoto and S.-i Ohkoshi, J. Am. Chem. Soc., 2009, 131, 18299–18303.
12. S.-i Ohkoshi and H. Tokoro, Bull. Chem. Soc. Jpn., 2013, 86, 897–907.
13. S. Sakurai, K. Tomita, K. Hashimoto, H. Yashiro and S.-i Ohkoshi, J. Phys. Chem. C, 2008, 112, 20212–20216.
14. J. López-Sánchez, A. Serrano, A. del Campo, M. Abuín, E. Salas-Colera, A. Muñoz-Noval, G. R. Castro, J. de la Figuera, J. F. Marco, P. Marín, N. Carmona and O. Rodríguez de la Fuente, RSC Adv., 2019, 9, 17571.
15. E. Gorbachev, M. Soshnikov, M. Wu, L. Alyabyeva, D. Myakishev, E. Kozlyakova, V. Lebedev, E. Anokhin, B. Gorshunov, O. Brylev, P. Kazin and L. Trusov, J. Mater. Chem. C, 2021, 9, 6713.
16. D. P. Debecker, S. Le Bras, C. Boissière and A. Chaumonnot, Chem. Soc. Rev., 2018, 47, 4112–4155.
17. A. Douy, J. Eur. Ceram. Soc., 2006, 26, 1447–1454.
18. S. Bruni, F. Cariati, M. Casu, A. Lai, A. Musinu, G. Piccaluga and S. Solinas, Nanostruct. Mater., 1999, 11, 573–586.
19. C. de Montferrand, Y. Lalatonne, D. Bonnin, N. Lièvre, M. Lecouvey, P. Monod, V. Russier and L. Motte, Small, 2012, 8, 1945–1956.
20. R. Kinugawa, K. Imoto, Y. Futakawa, S. Shimizu, R. Fujiwara, M. Yoshikiyo, A. Namai and S. Ohkoshi, Adv. Eng. Mater., 2021, 2001473.
21. S. Ohkoshi, K. Imoto, A. Namai, S. Anan, M. Yoshikiyo and H. Tokoro, J. Am. Chem. Soc., 2017, 139(38), 13268–13271.
22. Y.-K. Hong and J. Lee, Solid State Phys., 2013, 64, 237–329.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d2cc03168j

---