Journal of Nanoparticle Research

, Volume 13, Issue 1, pp 213–220

Magnetite hollow spheres: solution synthesis, phase formation and magnetic property


  • Qian Sun
    • Key Laboratory of Micro-Nano Measurement-Manipulation and Physics (Ministry of Education), Department of PhysicsBeijing University of Aeronautics and Astronautics
  • Zheng Ren
    • Key Laboratory of Micro-Nano Measurement-Manipulation and Physics (Ministry of Education), Department of PhysicsBeijing University of Aeronautics and Astronautics
    • Key Laboratory of Micro-Nano Measurement-Manipulation and Physics (Ministry of Education), Department of PhysicsBeijing University of Aeronautics and Astronautics
  • Weimeng Chen
    • Department of PhysicsPeking University
    • Department of PhysicsPeking University
Research Paper

DOI: 10.1007/s11051-010-0020-5

Cite this article as:
Sun, Q., Ren, Z., Wang, R. et al. J Nanopart Res (2011) 13: 213. doi:10.1007/s11051-010-0020-5


Polycrystalline magnetite hollow spheres with diameter of about 200 nm and shell thickness of 30–60 nm were prepared via a facile solution route. For the reaction, ethylene glycol (EG) served as the reducing agent and soldium acetate played the role of precipitator. In addition, polyvinylpyrrolidone (PVP) served as a surface stabilizer. The morphologies and structures were characterized by scanning electron microscopy, transmission electron microscopy and X-ray diffraction. The intermediate products at different stages were also studied to shed light on the evolution of phase formation. It revealed that the hollow structure formed via self-assembly of nanocrystallites (about 15 nm) using sodium acetate as mild precipitator. Evidences further pointed out that the Ostwald ripening process well explained the growth mechanism of the hollow structure. Magnetization measurements showed that the coercivity of magnetite hollow spheres at low temperature is about 200 Oe and the saturation magnetization is about 83 emu g−1, roughly 85% that of the bulk phase, close to the value of its solid counterpart. In addition, a freezing transition was observed at 25 K.


Magnetite hollow spheresPhase formationMagnetic property


As a useful functional material, magnetite (Fe3O4) has wide application potentials in diverse areas such as ferrofluid (Raj et al. 1995; Butter et al. 2002; Hong et al. 2006), magnetic data storage materials (Zeng et al. 2002), devices (Dadarlat et al. 2008; Ge et al. 2008; Zhang et al. 2008) and environment protection materials (Oliveira et al. 2004). Moreover, due to its excellent biocompatibility, magnetite is also an ideal candidate for biomedical applications (Doyle et al. 2002; Hung et al. 2007; Zhang et al. 2009).

In the past few years, motivated by the belief that novel properties of nanomaterials can be gained by rationally tuning their size, shape, as well as the way they are fabricated (Puntes et al. 2001; Sun and Xia 2002; Wang et al. 2005; Franger et al. 2007; Guo et al. 2007; Wang et al. 2007), magnetite nanostructures such as nanoparticles, nanowires, and other novel nanostructures have been synthesized via various approaches (Ling et al. 2004; Zhong et al. 2006; Kovalenko et al. 2007; Aldea et al. 2009; Zhang and Zhang 2009). In particular, nano-sized hollow magnetite has attracted great interest because of its properties such as low density, selective permeability, and large specific area, etc. (Li et al. 2006; Cong et al. 2008; Deng et al. 2008; Huang et al. 2009). However, among all these previous works, expensive complex polymers, toxic chemical reagents, or sacrificial templates were needed to obtain magnetite hollow nanostructures. Besides, these methods still face shortcomings including low productivity and rigorous reaction conditions. Developing simple and efficient route to synthesize magnetite hollow structures remains as a challenge in wet chemical synthesis.

Hence, in this work, a facile solvothermal method was developed to synthesize submicron-sized magnetite hollow spheres. The novelty of this work could be characterized by its convenience and environmental friendliness. Besides, the phase formation mechanism was investigated. It was found that the presence of sodium acetate was crucial to the formation of pure magnetite phase and its concentration was important to the morphology evolutions from solid spheres to their hollow counterparts or to the self-assembly chains of solid particles.


In a typical process, 1.68 mmol ferric chloride (FeCl3·6H2O) and 3.6 g polyvinylpyrrolidone (PVP) were dissolved in 70 mL ethylene glycol (EG) to form clear “initial solution”. It was followed by the addition of 16.8 mmol sodium acetate (NaAc) under magnetic stirring. The mixture was then transferred to an 80 mL Teflon-lined autoclave and maintained at 200 °C for 16 h. After the autoclave was cooled to room temperature, the black products were obtained by centrifuging and washed with ethanol and water for several times. The powders were dried at 60 °C for 6 h for further characterization. All the reagents were of analytical purity and used without further purification. The process is summarized in a flowchart in Fig. 1.
Fig. 1

Flowchart of the typical synthesis process

The X-ray diffraction (XRD) patterns of the as-prepared products were recorded on an analytical X’Pert Pro MPD X-ray diffractometer equipped with Cu-Kα radiation (λ = 0.15409 nm). The morphologies and structures were characterized by a Hitachi S-4800 scanning electron microscope (SEM) with a cold field emission gun and by a JEOL JEM-2100F transmission electron microscope (TEM) and high-resolution TEM (HRTEM) operated at 200 kV. The specimens for the SEM studies were prepared by dispersing the powder samples on the silicon substrates, while that for TEM and HRTEM investigations were obtained by dispersing of the as-prepared products in ethanol solution and then drops of the suspension were placed on copper grids and dried in air. Magnetic properties of the powder sample, ~5.06 mg, were measured by a superconducting quantum interference device (SQUID) magnetometer (Quantum Design).


The XRD spectrum for the powder sample of the hollow spheres is shown in Fig. 2. All of the diffraction peaks match well with the Bragg reflections of the standard face-centered cubic (f.c.c.) Fe3O4 structure (JCPDF #19-0629). The average crystallite size is calculated to be ~10 nm by Scherrer’s equation.
Fig. 2

XRD patterns for the as-synthesized magnetite (Fe3O4) hollow spheres

A broad view of SEM image shows that the as-synthesized Fe3O4 are composed of hollow spheres many of which are with openings, as shown in Fig. 3a. Size distribution is given inset and the statistical average diameter is about 190 nm. The shell thickness is estimated to be 30–60 nm from figures. A SEM image at higher magnification is presented for a typical hollow sphere with an opening, as shown in Fig. 3b. It reveals that the hollow sphere is made up of nanoparticles, about 10 to 15 nm in diameter. The size of the nanocrystallites on the outer surface appears to be larger than inner. TEM investigations give a deeper insight into the fine structures and the hollow nature of the Fe3O4 spheres, as shown in Fig. 3c. The strong contrast between the relatively darker edges and the brighter center of each sphere provides a further evidence for their hollow features. A HRTEM image taken from the shell region discloses perfect lattice fringes of the selected region, which demonstrates good crystallization of the nanocrystallites. The lattice fringes can be indexed to magnetite along [125] zone axis. The grain boundaries of the nanocrystallites are clearly visible, as marked by the dashed curve for one of them in Fig. 3d.
Fig. 3

Morphology investigations by SEM, TEM and HRTEM. a An overview of the hollow spheres by SEM. Inset shows the histogram of size distribution. b A magnification SEM image for one of the hollow spheres shown in a. c Bright field TEM image of the hollow spheres. d HRTEM image for one typical hollow sphere

Formation of magnetite hollow spheres

To understand the phase formation of magnetite and the mechanism for the morphology evolution from solid particles to hollow spheres, experiments of time evolutions were carried out by the addition of low concentration of NaAc (0.24 M), which is 16.8 mmol in 70 mL of the “initial solution in procedure 2 of the synthesis process illustrated in Fig. 1. Furthermore, reactions without NaAc and with high concentration of NaAc (0.72 M), about 50.4 mmol in 70 mL of the “initial solution”, were also conducted as controlled experiments to clarify the phase formation mechanism of magnetite hollow spheres.

The results of time evolution experiments are summarized as follows. With the presence of low concentration NaAc, no precipitations were observed when the reaction time was less than 2 h. With reaction time for 3 h, the products were brownish and amorphous-like, as shown in Fig. 4a. Increased reaction time to 5 h, black precipitate was obtained (Fig. 4b). Cluster-like aggregations (~50 nm in diameter) of nanocrystallites were found and most of them are connected. The aggregates become compact solid spheres with the diameter of 120–150 nm as increasing the reaction time to 8 h (Fig. 4c). The solid spheres are uniform in shape and obviously polycrystalline. Finally, with the reaction time of 16 h, the hollow spheres were obtained as shown in Fig. 3. The XRD patterns (Fig. 4d) indicates that the aggregates (shown Fig. 4b) are already formed f.c.c. Fe3O4 nanocrystallites. As raw materials, there was no ferrous ion source, but only ferric salt in our experiment which indicates that reducing reaction might happened during the synthesis process.
Fig. 4

Phase evolution processes towards the formation of hollow spheres with low concentration of NaAc. a A SEM image for the amorphous precursor after reaction for 3 h. b SEM image for the initial stage of magnetite crystallization after reaction for 5 h. c Polycrystalline solid particle formed of magnetite nanocrystallites after reaction for 8 h. d XRD spectrum of the products with reaction time of 5 h

EG serving as a reducing agent is well discussed in the polyol process (Sun and Xia 2002; Zhong et al. 2006; Wang et al. 2008a, b). The ferrous ion might be the result of the reducing reaction between ferric iron and EG. To confirm the reducing property of EG in our experiment, we substituted of EG with water as solution and only Fe2O3 was obtained as final products.

Thus, the reactions for the phase formation of magnetite is expressed as follows,
$$ {\text{CH}}_{ 3} {\text{COO}}^{ - } + {\text{H}}_{ 2} {\text{O}} \to {\text{ CH}}_{ 3} {\text{COOH}} + {\text{OH}}^{ - } $$
$$ 2 {\text{Fe}}^{ 3+ } + {\text{Fe}}^{ 2+ } + 8 {\text{OH}}^{ - } \to {\text{ Fe}}_{ 3} {\text{O}}_{ 4} + 4 {\text{H}}_{ 2} {\text{O}} $$
NaAc acted as a mild precipitator and the weak hydrolyzation of it as described by Eq. 1 would control the release rate of OH, thus the nucleation of Fe3O4 (Eq. 2) could progress in a controlled manner. Without the addition of NaAc, neither magnetite nor iron oxides were obtained even with longer reaction time. Since iron has a strong tendency to coordinate with carboxylate groups (Karmakar and Chakravorty 1996), NaAc in the present experiment may coordinate with iron to form precursors. The overall reaction is thus expressed as follows,
$$ {\text{Fe}}\left( {\text{Ac}} \right)_{ 2} + 2 {\text{Fe}}\left( {\text{Ac}} \right)_{ 3} + 4 {\text{H}}_{ 2} {\text{O}} \to {\text{ Fe}}_{ 3} {\text{O}}_{ 4} + {\text{ 8HAc}} $$
NaAc is a key factor to dictate the concentration of the precursors, Fe(Ac)2 and Fe(Ac)3 in Eq. 3. Therefore, its concentration is critical for the control of the kinetic process of phase formation. By increasing the concentration of NaAc from 0.24 to 0.72 M, well-crystallized, chain-like network composed of Fe3O4 nanoparticles with a diameter of 50 nm are obtained (Fig. 5a, b). The bright field TEM image shown in Fig. 5c reveals that these chains of particles are solid, rather than hollow. A HRTEM image is shown in Fig. 5d. The lattice spacing marked in the Fig. 5d is measured to be about 0.48 nm, corresponding to the interplanar spacing between (111) planes of cubic Fe3O4.
Fig. 5

Characterization on the well-crystallized solid magnetite particles after reaction for 16 h with high concentration of NaAc. a An overview by SEM image. b A magnified SEM image of a.c Bright field TEM image. d A HRTEM and corresponding fast Fourier transform (FFT) pattern (inset)

Moreover, the effect of the PVP was also studied. Previous investigations indicated that it served as surfactant-templating in the synthesis of Co and Ni hollow sphere chains (Guo et al. 2007; Wang et al. 2008). However, in this experiment, hollow spheres were also obtained without the addition of PVP but with bigger pores in surface and broader size distribution (shown in Fig. 6). So the role of PVP in our experiment did not serve as soft template. Such phenomenon was also found in the synthesis of Fe3O4 hollow sphere (Liu et al. 2009). Similar experiments for the effect of PEG in the synthesis of self-assembly magnetite hollow spheres have also been reported (Jia and Gao 2008). Based on the discussion above, it is rational to conclude that PVP severed as a surface stabilizer rather than the soft template.
Fig. 6

Morphologies of the magnetite prepared without PVP: SEM images of the products

Discussions on growth mechanism

By the studies presented above, the addition of low concentration of NaAc (0.24 M) is the key to produce the hollow spheres of magnetite. It is noticeable that the diameter of solid particles increases sequentially from about 50 nm (Fig. 4b) to 100–120 nm (Fig. 4c), and eventually reaches 200 nm as diameter of hollow spheres (Fig. 3) as prolonging reaction time from 5, 8 to 16 h with reaction temperature of 200 °C. Due to the sintering of several solid spheres with increase in the time of the reaction, solid aggregation spheres with diameter about 120 nm formed with sintering of several smaller solid spheres (i.e. 50 nm). In addition, for the hollow spheres, the size of nanocrystallites decreases inwardly from outer surface, i.e., from over 15 nm on the outer surface down to about 10 nm in the interior (shown in Fig. 3b). Furthermore, the large nanocrystallites (~15 nm) is packed more densely on the surface than those in the interior (~10 nm), as shown in Fig. 3b. From the energy point of view, it evidences a larger surface cohesive energy, which in turn favors the dissolution of the inner small nanocrystallites. The above discussions suggest that an inside-out mass transportation occurs and that the large nanocrystallites on the outer surface grow at the expense of the smaller one inside. This phenomena was similar to inside-out Ostwald ripening which was first described by Ostwald in 1896 (Ostwald 1900) concerning the growth of large precipitates at the expense of smaller precipitates. Besides, similar phenomena were also observed in other hollow structures of a wide range of materials obtained via the Ostwald ripening (Lou et al. 2006; Teo et al. 2006; Ottaviano et al. 2009).

Based on the experiment results and discussions above, we believe that the formation of the magnetite hollow spheres experienced a two-step process. First is the sintering process, which could be observed as the diameter increase of the spheres as shown in Fig. 4b and c with reaction time increase from 5 to 8 h. During this process, the sintering process is the main reason for the diameter increase from 50 to 120 nm. Second, with reaction time prolonging to 16 h, the magnetite hollow structure with larger nanoparticles on surface and smaller ones inside appeared which can attributed to the inside-out Ostwald ripening. Thus, the main reason for the formation of magnetite hollow spheres was the inside-out Ostwald ripening.

On the contrary, high concentration of NaAc (0.72) led to the formation of magnetite solid spheres without obvious size increase or hollow structure emerges which means the formation mechanism of magnetite hollow spheres and well-crystallized solid particles were different according to the concentration of NaAc.

The two different formation mechanisms were illustrated in a schematic diagram presented in Fig. 7 along with time evolution. With higher concentration of NaAc (0.72 M) in the reaction, magnetite nanocrystallites grow at the same rate with continues release of OH- and residual iron ions (stage 1). Due to the high surface energy of nanocrystallites, they tend to aggregated to self-assembly solid spheres and might undergo re-crystallization to form single nanoparticles (50 nm in diameter with reaction time for 16 h as shown in Fig. 5) under solvothermal condition (stage 2). As we known, NaAc was recognized as electrostatic stabilization to prevent the nanoparticles from aggregation. The electrostatic effect decreased with the decrease of the concentration of NaAc in the solution. Then the smaller solid spheres (~50 nm as shown in Fig. 4b) might undergo a sintering process to form larger solid ones (120 nm as shown in Fig. 4c). While due to the low energy of nanoparicles on surface compared to the smaller inner ones, the larger nanoparticles on the surface grew at the expanse of inner ones to form hollow structure (stage 2).
Fig. 7

Schematic illustration for the phase formation and morphological evolutions for both hollow magnetite spheres and solid nanoparticles with different concentration of NaAc

Magnetic properties

Magnetization of the samples has been measured by a SQUID magnetometer (Quantum Design). Zero-field-cooling (ZFC) and field-cooling (FC) M(T) curves were measured with the applied field, Happ = 90 Oe, in the warming process from T = 5–360 K. Before the warming for data collection, the field applied during the process of sample-cooling is 0 Oe for the ZFC measurement, and 10 kOe for the FC measurement. The temperature dependent magnetization, M(T), curves were shown in Fig. 8. In the low temperature region, at T ~ 25 K in both ZFC and FC M(T) curves, shoulder-like features are observed. This behavior is likely attributed to the presence of a surface spin glass (SG) phase with the nanocrystallites forming the hollow spheres. Similar behavior has been found at TF ~ 41 K in Fe3O4 nanoparticles with an average size of 20 nm (Wang et al. 2004). On the other hand, the Verway transition is not observed, which usually appears at 125 K for the bulk and is suppressed to 98 K and 16 K with the nanocrystals of 150 nm and 50 nm in size, respectively. It disappears for even smaller nanocrystals, 20 nm in size. Hence, we do not expect to observe the Verway transition with the present sample. We have also measured the field dependent magnetization M(H) curves at T = 5 and 300 K, as shown in Fig. 9. The inset shows the detail in the low field region. The saturation magnetization MS is determined as ~82.8 emu g−1 (5 K) and 75.5 emu g−1 (300 K), which is slightly smaller than the bulk value. The saturation magnetization for bulk magnetite (Fe3O4) is 98 emu g−1 at 0 K and 92 emu g−1 at room temperature.
Fig. 8

Temperature dependent magnetization, M(T), curves for magnetite hollow spheres from 5 to 360 K. Data were measured in the field of 90 Oe. The shoulder-like structure at T ~ 25 K in both curves is ascribed to the freezing temperature
Fig. 9

Field dependent magnetization, M(H), curves for magnetite hollow spheres at T = 5 and 300 K. Inset shows the detail in the low field region


Polycrystalline Fe3O4 hollow spheres were synthesized by a facile hydrothermal method. PVP was added as a surface stabilizer and both of EG and sodium acetate (NaAc) were essential for the formation of the magnetite hollow spheres. To be specific, EG was reducing agent, and NaAc played the role of precipitator controlling the release rate of OH which is necessary for the formation of the hollow structure. The growth mechanism of the magnetite hollow spheres was suggested by the Ostwald ripening effect, which was verified by the time evolution experiments. Magnetization measurements show that the saturation magnetization, Ms = 82.8 emu g−1 at 5 K is about 85% of the bulk value and the coercivity is 200 Oe. From this point of view, the magnetite hollow spheres are promising for a potential application as carriers for drug targeting.


This work was supported by the National Natural Science Foundation of China (Nos. 50671003, 50971011 and 10874006), Beijing Natural Science Foundation (No. 1102025), the National Basic Research Program of China (Nos. 2009CB939901 and 2010CB934601), the Program for New Century Excellent Talents in University (NCET-06-0175) and Research Fund for the Doctoral Program of Higher Education of China (20091102110038).

Copyright information

© Springer Science+Business Media B.V. 2010