Magnetic iron oxide nanoparticles (MNPs) have been researched extensively due to their promising applications for biomedicine such as hyperthermia^[1,2], magnetic resonance imaging^[3-4] and biolabeling^[5]. Magnetic characteristics is highly significant for successful application of MNPs in biomedicine/biotechnology^[6]. Recent studies suggest that a metal (Mn, Zn) dopant substitution strategy can achieve high magnetic performance^[7,8]. The Zn, Mn doping level, a key factor, can be regulated by simply adjusting the original molar ratio of metal precursors. Moreover, when the ratio of doping Zn and Mn is 2:3, saturation magnetization (M_s) can reach its maximum^[9].

For decades, a variety of preparation methods of MNPs achieved in hydrolytic phase have been developed, including hydrothermal reaction^[10], Sol-Gel process^[11] and coprecipitation method^[12]. Magnetic iron oxide nanoparticles can also be obtained by organic-phase thermal decomposition of metal precursor, for example, decomposition of Fe(acac)₂ followed by oxidation to Fe₃O₄ in the presence of oleylamine and oleic acid^[6,13]. Recent studies showed that organic-phase thermal decomposition methods could overcome the shortcomings of traditional hydrolytic one: it allows more strict control on particles uniformity and structure^[14,15].

Size-dependent biomagnetic applications are the most remarkable properties of MNPs^[16,17]. Different biological applications have specific requirements on particle size. For example, in the separation of cells, the size of MNPs should be 8 to 10 nm^[17,18], but magnetothermal therapy requires larger nanoparticles, resulting from their higher specific absorption rate^[19,20]. On the other hand, the size variation also affects the coercivity (H_c) and saturation magnetization (M_s) of the magnetic nanoparticles. However, it remains a great challenge to control MNPs size due to the lack of sufficient understanding of formation mechanism of magnetic particles referring to nucleation and growth process^[21]. Therefore, it is necessary to study the effects of related parameters on particle size so that we can synthesize magnetic particle with tunable size. In this study, we present a one-step preparation of 5-20 nm Zn, Mn doped Fe₃O₄ (ZnMn-Fe₃O₄) nanoparticles, which can adapt to broader biomagnetic applications.

Material Oleic acid (OAc) and Oleylamin (OAm) were purchased from Sigma-Aldrich. Other chemicals were purchased from Shanghai Aladdin Biochemical Technology Co. Ltd.

Preparation of 5 nm ZnMn-Fe₃O₄ nanoparticles Oleylamine (3 mmol), oleic acid (3 mmol), Fe(acac)₃ (1 mmol), Zn(acac)₂ (0.2 mmol) , Mn(acac)₂ (0.3 mmol) and 1,2-hexadecanediol (5 mmol) were placed in a 50 mL three-neck round-bottom flask in the presence of 15 mL benzyl ether. The mixture was heated to 200 ℃ for 30 min and then to 300 ℃ for 1 h under argon atmosphere. After cooling to room temperature, excess ethanol was added to the solution, then a black powder precipitated and was collected by centrifugation. The magnetic nanoparticles were redispersed in cyclohexane.

Preparation of 10 nm ZnMn-Fe₃O₄ nanoparticles Oleylamine (3 mmol), oleic acid (3 mmol), Fe(acac)₃ (1 mmol), Zn(acac)₂ (0.2 mmol) and Mn(acac)₂ (0.3 mmol) were placed in a 50 mL three-neck round-bottom flask in the presence of 15 mL benzyl ether. The mixture was heated to 200 ℃ for 30 min and then to 300 ℃ for 1 h under argon atmosphere. After cooling to room temperature, excess ethanol was added to the solution, then a black powder precipitated and was collected by centrifugation. The magnetic nanoparticles were redispersed in cyclohexane.

Preparation of 15 nm ZnMn-Fe₃O₄ nanoparticles Oleylamine (3 mmol), oleic acid (3 mmol), Fe(acac)₃ (1 mmol), ZnCl₂ (0.2 mmol) and MnCl₂ (0.3 mmol) were placed in a 50 mL three-neck round-bottom flask in the presence of 15 mL benzyl ether. The mixture was heated to 200 ℃ for 30 min and then to 300 ℃ for 1 h under the atmosphere of argon. After cooling to room temperature, excess ethanol was added to the solution. The black precipitate was collected by centrifugation and washed with ethanol prior to redispersion in cyclohexane for further use. Similarly, by prolonging reaction time to 1.5 or 2 h, 20 nm ZnMn-Fe₃O₄ nanoparticles can be prepared.

Characterization Transmission electron microscopy (TEM) was performed on JEM-2100F instruments. The XRD studies were conducted via a Rigaku D/MAX-2250V diffractometer with a Cu Kα radiation source (40 kV, 120 Ma). Magnetic measurements were performed on a Vibrating Sample Magnetometer (VSM). FT-IR spectra were recorded by an IRPRESTIGE-21 spectrometer (Shimadzu) using KBr pellets. Size distribution were measured on Nano-Zetasizer (Malvern Instruments Ltd.).

Reductive effect As shown in Fig. 1(a), monodispersed ZnMn-Fe₃O₄ nanoparticles about 5 nm were fabricated by decomposition of Fe(acac)₂, Mn(acac)₂ and Zn(acac)₂ in the presence of oleylamine, oleic acid and 1,2-hexadecanediol. Under the same experimental conditions, if 1,2-hexadecanediol was not added, the size of particles increased from 5 nm to 10 nm (Fig. 1(b)), which were further proved by the alteration of size distribution (Fig. 1(c, d)). Organic-phase thermal decomposition in fact is a process that activates metal atoms to nucleate and grow^[21,22], so the size of particles is mainly dependent on the nucleation and growth rate of metal atoms. Herein, the addition of strong reductive 1,2-hexadecanediol contributed to facile reduction of M(acac)₂ to M⁰ (M=Zn, Mn), leading to faster consumption of metal acetylactone and nucleation of particles, as a result, smaller nanoparticles were formed.

Metal precursor effect Metal precursors play an important role in size regulation of magnetic nanoparticles. The size of nanoparticles synthesized by metal chloride as precursor is about 15 nm with narrow size distribution (Fig. 2(a, b)). The high resolution TEM (Fig. 2(c)) and selected area electron diffraction (SAED) pattern (Fig. 2(d)) suggest that the particles have high-quality crystallinity. In contrast, using acetylacetone as metal precursor, the size of obtained nanoparticles is about 10 nm (Fig. 1(b)). According to chemical bond theory, metal chloride is more stable than metal acetylacetone, thus it requires higher decomposition temperature. Therefore, the addition of metal chloride may lead to slower nucleation rate, more metal precursor could deposit surrounding the nuclei formed in mixture and resulting in larger nanoparticles.

Reflux time effect It is also important to study the effect of reaction time on the size of nanoparticles. In this work, we found that larger particles could be obtained by longer reflux time. As shown in Fig. 3(a), when reflux time was postponed from 1 h to 1.5 h under the same reaction condition, the particles size increased from 15 nm (Fig. 2(a)) to 20 nm (Fig. 3(a)), which were further proved by the alteration of size distribution (Fig. 3(c)). However, further prolongation of the reflux time to 2 h resulted in no distinct increase in the size of nanoparticles (Fig. 3(b, d)), which indicates that the nanocrystals will stop growing when the reaction reaches homeostasis. In the meantime, it seems to produce new nanocrystal with longer reflux time (Fig. 3(b)).

The XRD pattern (Fig. 4(a)) shows a single-phase spinel structure of 5-20 nm ZnMn-Fe₃O₄ nanoparticles and the relative intensity and position of all diffraction peaks/rings accord well with previously reported Zn, Mn doped Fe₃O₄ powder^[23], which indicates that crystal structure has little relevance with particle size. Two characteristic stretching were recorded in FT-IR spectrum at 2920 and 2850 cm^-1, 1630, and 1410 cm^-1 corresponding to H-C-H, C-O stretching, respectively (Fig. 4(b)), which reveals that particles are coated by oleylamine and oleic acid. In this case, in addition to acting as reductive, oleylamine and oleic acid also work as surfactant. The organic surfactants can be removed by 2,3-dimercaptosuccinic acid (DMSA) to yield water-soluble nanoparticles^[20]. The Zn, Mn doping levels were measured by energy dispersive X-ray spectroscopy (EDS) and the results indicate that 5-20 nm particles have similar element composition (Fig. 4(c)).

Size effect on magnetic characteristics Magnetic characteristics of MNPs with different size were measured by VSM at 300 K, and the results are shown in Fig. 5(a). The hysteresis curve of MNPs in 5-15 nm exhibits superparamagnetism without remnant magnetization. However, 20 nm particles are ferromagnetic with a coercivity of 150 Oe (1 Oe≈79.6 A/m). Saturation magnetization (M_s) value of MNPs gradually increased as the size of MNPs increased from 5 to 20 nm, and reached the maximum of 77.7 emu/g (emu/g=4π×10^-7 Wb∙m/kg). As Fig. 5(b) shown, M_s and d^-1 (the reciprocal of the nanoparticles size) display an approximate linear relationship, which is consistent with previous reports^[24,25].

We have successfully synthesized 5-20 nm ZnMn-Fe₃O₄ by changing metal precursors, varying reflux time and adding 1,2-hexadecanediol. These ferrite nanoparticles with different particle sizes could adapt to the needs of extensive biomagnetic applications. Smaller sized nanoparticles could be obtained by introducing extra reducing power 1,2-hexadecanediol, and we can prepare larger particles by using more stable metal precursor and extending reflux time. These seem to indicate that powerful reducing agent and metal precursor with low decomposition temperature will yield smaller particles. Meanwhile, the magnetic characteristics of ZnMn-Fe₃O₄ are closely related to particle size. Understanding how the different parameters impact on particle size is extremely critical to develop a more predictive and controllable way to synthesize the MNPs.