Synthesis and magnetic properties of MNb₂O₆ (M = Fe, Co, Ni) nanoparticles

Abstract

Considerable efforts have been exerted on the controllable synthesis of columbite niobate ceramics due to their fascinating properties and applications. Especially, it is still a great challenge to fabricate nanostructures of the niobate series. In this research, FeNb₂O₆, CoNb₂O₆ and NiNb₂O₆ nanoparticles have been successfully prepared via a facile hydrothermal route followed by heat treatment. X-ray powder diffraction patterns show that all the products have the typical orthorhombic columbite structure. The electron microscopy analyses reveal that the obtained nanoparticles have diameters of 50–100 nm. The magnetic property results demonstrate that the magnetically ordered state is hard to observe down to 1.8 K for the FeNb₂O₆ sample, while the magnetic transition temperatures of T_N = 3 K and T_N = 6 K can be obtained for CoNb₂O₆ and NiNb₂O₆ samples, respectively. A weak ferromagnetic moment can be detected below 5 K for both CoNb₂O₆ and NiNb₂O₆ samples. Furthermore, the NiNb₂O₆ sample even exhibits a metamagnetic transition at 1.8 K due to the spin flipping of the ferromagnetic chains.

---

1. Introduction

The binary niobate ceramics, with a general formula ANb₂O₆ where A is a divalent alkaline-earth or transition metal cation with ionic radius less than 1 Å, usually crystallize in the columbite structure. This class of materials has attracted much attention due to their interesting physical properties and extensive applications.^1,2 It is well-known that columbite can be considered as a superstructure of R-PbO₂. Then accordingly, in columbite ANb₂O₆ structure with an orthorhombic symmetry, A and Nb atoms are located at 4c and 8d positions, respectively, and surrounded by six oxygen atoms to form AO₆ and NbO₆ octahedra, which share edges and form independent zigzag chains along the c-axis. Meanwhile, these parallel AO₆ and NbO₆ layers alternate along the a-axis in the sequence A–Nb–Nb–A–Nb–Nb–A.³ The crystal structure of columbite ANb₂O₆ can be depicted as Scheme 1.

Scheme 1 Crystal structure of columbite ANb₂O₆.

Among the columbite niobate family, the ANb₂O₆ (A = Fe, Co, Ni) compounds, in which A refers to the magnetic transition metal, have greatly stimulated considerable research interest because of their prominent magnetic, dielectric, optical and catalytic properties.^1,4–6 In the recent two decades, the low-dimensional magnetic systems have been the subject of intense theoretical and experimental research. Actually, these ANb₂O₆ (A = Fe, Co, Ni) niobates are considered as prototype materials since they exhibit quasi-one-dimensional magnetic characteristics.^7–11 In general, ANb₂O₆ compounds exhibit magnetic ordering at rather low temperatures (typically below 10 K), and all the ordered phases have been found to be antiferromagnetic. Their magnetic structures have been investigated and identified by theoretical calculations and experimental measurements based on both polycrystalline and single-crystal samples. Weitzel first studied the magnetism for FeNb₂O₆ powder by neutron diffraction and reported the antiferromagnetic order at 4.2 K with a collinear spin arrangement.³ Later, according to the magnetization and susceptibility of FeNb₂O₆ single crystal, an antiferromagnetic ordering was observed at about 5.5 K by Yaeger et al.¹² Even later, Heid et al. reinvestigated the magnetism in FeNb₂O₆ through a very detailed research based on both powder sample and single crystal.⁷ The magnetic structure of CoNb₂O₆ powder sample and single crystal was also first studied by Weitzel¹³ and Yaeger¹⁴ group, respectively, and later deeply investigated by Scharf et al.,¹⁵ Heid et al.,^8,9 and Kobayashi et al.^16,17 An intermediate incommensurate magnetic phase was observed below the Néel temperature of 2.95 K. Recently, Sarvezuk and co-workers performed a new investigation of the magnetic structure of CoNb₂O₆, in which the magnetic ordering at 2.5 K was found.¹⁸ As for NiNb₂O₆, it has a similar magnetic structure with FeNb₂O₆ and CoNb₂O₆, which has also been systematically studied by Weitzel et al.,¹⁹ Yaeger et al.,²⁰ and Heid et al.⁷ It exhibits antiferromagnetic order below about 6 K with a canted magnetic structure.

Besides the magnetic behavior, the microwave dielectric properties of the ANb₂O₆ (A = Fe, Co, Ni) compounds have also drawn great attention for many research groups, such as Pullar,^21–24 Hong,^25,26 and Belous.^27,28 These ceramic materials usually have the high dielectric constant and the low dielectric losses at microwave frequency making them suitable as dielectric resonator and filter for use in the field of mobile and satellite communication. Additionally, these dielectric ceramics have lower sintering temperatures than the perovskites. Arroyo y de Dompablo and co-workers presented a computational study on the stability and electronic properties of the reduced (anion vacancy type) ANb₂O₆ compounds.²⁹ What's more, the electrochemical insertion properties of CoNb₂O₆ and NiNb₂O₆ as cathodes in lithium batteries were also studied.³⁰ Zhou et al. have explored the photoluminescence of NiNb₂O₆ nanoparticles, which show a blue emission at 440 nm due to the distorted edge-shared NbO₆ groups.⁵ The band gaps of NiNb₂O₆ was estimated to be 2.2 eV, which can be served as the solid photocatalyst in water splitting.³¹

There are various methods for the preparation of ANb₂O₆ (A = Fe, Co, Ni) compounds. Among them, the solid state combination has always been the most popular technique for the synthesis of polycrystalline ANb₂O₆ powders using stoichiometric mixtures of Nb₂O₅ and metal elements or oxides at high temperatures.^{3,4,7–11,13,18,19,21–28,30,31} However, multiple heating and regrinding steps are generally essential to overcome the solid state diffusion barrier, and a control of the atmosphere conditions is usually necessary. Additionally, the products always have irregular morphology and large particle size. NiNb₂O₆ nanoparticles have been synthesized by a sol–gel combustion method using citric acid as fuel and nitrates as oxidants at a relatively low temperature compared to the solid-state reaction method.⁵ Ravi group have prepared CoNb₂O₆ and NiNb₂O₆ ceramics by a coprecipitation technique using the mixture of ammonium carbonate and ammonium hydroxide as the precipitants under basic conditions.^32,33 Recently, Ma et al. have fabricated CoNb₂O₆ with rutile structure via a hydrothermal route.³⁴ As for the single-crystalline ANb₂O₆ (A = Fe, Co, Ni), the common growth methods include flux growth,^{12,14–17,20} and floating zone method.²

Motivated by the promising physicochemical properties and broad potential applications of ANb₂O₆ (A = Fe, Co, Ni) materials, it is interesting and challenging to explore a simple synthetic approach to fabricate ANb₂O₆ series, especially their nanostructures. In this research, a convenient and green hydrothermal process followed by heat treatment has been successfully employed for the preparation of ANb₂O₆ (A = Fe, Co, Ni) nanoparticles and the magnetic properties of the products have also been investigated.

2. Experimental

2.1. Chemicals and materials

All the chemicals and reagents were used as received without any further purification. Niobium pentoxide (Nb₂O₅), potassium hydroxide (KOH), hydrated ferrous sulfate (FeSO₄·7H₂O), hydrated cobalt dichloride (CoCl₂·6H₂O), and hydrated nickel chloride (NiCl₂·6H₂O) are of analytical grade purchased from Sinopharm Chemical Reagent Co., Ltd (Shanghai, China).

2.2. Synthesis of ANb₂O₆ (A = Fe, Co, Ni) nanoparticles

All samples were prepared by a hydrothermal method followed with calcination treatment and the synthetic procedure can be described as follows. For the first step, the potassium polyoxoniobate (K₇HNb₆O₁₉·13H₂O) was obtained from the reaction of Nb₂O₅ powder with molten KOH followed with the recrystallization in aqueous solution according to the previous literature.³⁵ Regarding the typical synthesis of ANb₂O₆ (A = Fe, Co, Ni) nanoparticles, 0.6 mmol of FeSO₄·7H₂O (CoCl₂·6H₂O, NiCl₂·6H₂O) and 0.2 mmol of K₇HNb₆O₁₉·13H₂O precursor were respectively dissolved in 10 mL of distilled water. Subsequently, FeSO₄ (CoCl₂, NiCl₂) solution was added dropwise into the polyoxoniobate solution under continuous magnetic stirring and the suspension could be formed. The resulting suspension was then transferred into a stainless steel Teflon-lined autoclave of 50 mL capacity, which was then filled with distilled water up to 80% of the total volume and stirred to be homogeneous. The autoclave was sealed and maintained at 180 °C for 12 h, and then cooled to room temperature naturally. The precipitates were filtered off and washed with absolute ethanol and distilled water several times to remove the soluble impurities, and then dried at 60 °C for 6 h in air. The obtained powders were finally calcined at 800 °C for 2 h in high-purity N₂ atmosphere.

2.3. Sample characterization

The X-ray powder diffraction (XRD) patterns were recorded on a D8 Focus diffractometer with Cu-Kα radiation (λ = 1.5406 Å) (Bruker, Germany). The transmission electron microscopy (TEM), high resolution transmission electron microscopy (HRTEM) images, selected area electron diffraction (SAED) patterns and energy dispersive X-ray spectroscopy (EDS) spectra were taken from a JEM-2010 transmission electron microscope (JEOL, Japan) with an accelerating voltage of 200 kV. For both TEM and HRTEM tests, after ultrasonic agitation, one or more drops of the ethanol solution containing the synthesized compounds were deposited onto the amorphous carbon film supported on a copper grid and allowed to dry at room temperature in air. The scanning electron microscopy (SEM) measurements were preformed on a JSM-6701F (JEOL, Japan) field-emission scanning electron microscope. The thermogravimetric (TG) and differential thermal analysis (DTA) curves were recorded with a Pyris Diamond TG-DTA thermal analyzer (Perkin-Elmer, USA) in the range of temperature from 20 to 1000 °C at a heating rate of 5 °C min⁻¹ under the flow of nitrogen. The magnetic properties data were collected on a Magnetic Property Measurement System (MPMS), SQUID-VSM (superconducting quantum interference device-vibrating sample magnetometer) (Quantum Design, USA). The prepared powdered samples were put into a gelatin capsule. The temperature dependence of the magnetization was measured under both zero-field-cooled (ZFC) and field-cooled (FC) modes in the temperature range of 1.8–300 K with an applied magnetic field of 100 Oe. For the detailed procedures, in the ZFC measurements, as the sample was cooled to 1.8 K in a zero magnetic field, an applied field was then introduced and the magnetization was recorded in a warming cycle. Then the FC measurements were conducted in a cooling cycle with an applied magnetic field. The isothermal magnetization measurements were carried out in a magnetic field that varied between +6 T and −6 T at different temperatures (1.8, 5, and 10 K). Before each run, for the sake of demagnetization, the sample was heated to room temperature and then cooled to the test temperature in a zero field.

3. Results and discussion

3.1. Phase analysis

The phase purity of the annealed samples is studied by the XRD method. As shown in Fig. 1, it can be seen that all the three phases have very similar XRD patterns, which means that the three different samples possess the same crystal structure. The XRD patterns shown in Fig. 1a–c can be indexed to the orthorhombic columbite structure of FeNb₂O₆ (JCPDS Card Files, no. 72-0483), CoNb₂O₆ (no. 72-0482), and NiNb₂O₆ (no. 72-0481), respectively. According to the diffraction positions, it is clear that the reflection peaks shift to the higher diffraction angle region from Fig. 1a (FeNb₂O₆) to Fig. 1c (NiNb₂O₆), associated with the decreasing lattice constants corresponding to the decreasing ionic radius for Fe²⁺ (0.74 Å), Co²⁺ (0.72 Å), and Ni²⁺ (0.69 Å). This result is in good agreement with the reported data, in which, for example, the diffraction position of the (311) plane for FeNb₂O₆, CoNb₂O₆, and NiNb₂O₆ is 30.173°, 30.273°, and 30.498°, respectively. It further confirms that the three samples consist of columbite ANb₂O₆ (A = Fe, Co, Ni) phases. No characteristic reflection peaks derived from other contaminants such as niobium or metal oxides, and other structure phases of metal niobates can be detected, which indicates that the level of impurities in the sample is lower than the resolution limit of the XRD instrument. According to the Scherrer equation, the average grain size of the obtained FeNb₂O₆, CoNb₂O₆, and NiNb₂O₆ samples can be estimated from the characteristic (311) peak to be 65.2, 58.3, and 44.6 nm, respectively.

Fig. 1 XRD patterns of the prepared columbite niobate ceramics via a general hydrothermal approach at 180 °C followed with annealing at 800 °C for (a) FeNb₂O₆, (b) CoNb₂O₆, and (c) NiNb₂O₆.

3.2. Morphology and structure analysis

The morphology of the products was examined by electron microscopy images. Fig. 2 shows the TEM images of the hydrothermal products without annealing. It is clear that all the samples are nanoparticles with the size below 100 nm. Among them, the mean diameter of CoNb₂O₆ and NiNb₂O₆ is about 50 nm, while the size of FeNb₂O₆ is slightly larger.

Fig. 2 TEM images of the hydrothermal products without annealing for (a) FeNb₂O₆, (b) CoNb₂O₆, and (c) NiNb₂O₆, and the corresponding XRD patterns for the hydrothermal products of (d) FeNb₂O₆, (e) CoNb₂O₆, and (f) NiNb₂O₆ without annealing.

It should be mentioned that all these three hydrothermal products without annealing are amorphous according to the XRD results (as shown in Fig. 2d–f). To understand the thermal behaviour of them, TG-DTA curves were investigated as displayed in Fig. 3. The TG-DTA curves of the three samples have a similar profile. From the TG curves (solid line), it can be observed that there are two steps of weight loss. The first weight loss below 300 °C is due to the evaporation of physically adsorbed water on the surface of the samples. The second weight loss of about 3% occurs in the range of 300–600 °C. If the samples are speculated as the amorphous hydrates, then the second weight loss may be ascribed to the dehydration of coordinated water. After 600 °C, then no obvious weight loss can be detected. As for the DTA curves (dash line), the initial endothermic peak is definitely resulted from the evaporation of water. However, the second endothermic peak corresponding to the dehydration process is absent probably due to the limited sensitivity of DTA. Additionally, at about 650–700 °C, a broad exothermic peak also can be observed, which implies the initiation of crystallization.

Fig. 3 TG (solid line) and DTA (dash line) curves of the hydrothermal products of FeNb₂O₆, CoNb₂O₆, and (c) NiNb₂O₆ without annealing.

Fig. 4a and b respectively presents the SEM and TEM image of the FeNb₂O₆ sample after annealing at 800 °C in N₂ atmosphere. From the SEM image, it can be seen that the sample consists of large-scale quasi-round nanoparticles with diameters of about 100 nm. The corresponding size distribution graph (inset of Fig. 4a) reveals that most nanoparticles have the sizes no more than 100 nm. However, there exists an obvious agglomeration of these nanoparticles to form the larger particles probably owing to the sintering effect during annealing. To further explore the microstructure of the obtained FeNb₂O₆ sample, the HRTEM and SAED analyses were undertaken from a single constituent nanoparticle. As shown in Fig. 4c, the HRTEM image shows clearly resolved two-dimensional atomic lattice fringes, suggesting a good crystallinity of these nanoparticles. The observed interplanar spacings can be measured to be 0.3 nm and 0.48 nm with a separation angle of about 70°, which match well with the (311) and (1̄01) planes of orthorhombic columbite structure of FeNb₂O₆ phase.³⁶ Fig. 4d presents the corresponding SAED patterns of the FeNb₂O₆ nanoparticle, which also confirms the good single-crystalline structure. These results further demonstrate that the prepared product consists of FeNb₂O₆ pure phase. The EDS spectrum taken from the nanoparticles is displayed in Fig. 4e, which shows the presence of Fe, Nb, O, Cu, C and Cr. The element Cu and C is derived from the copper grid and carbon film, respectively, while the element Cr should be originated from the TEM sample holder. The EDS spectrum shows that the atomic ratio of Fe : Nb : O is 1 : 2.03 : 8.16. The ratio of Fe to Nb is very close to 1 : 2 and the excessive O may be from the adsorbed O₂ and/or H₂O. So, it also confirms the formation of FeNb₂O₆ phase.

Fig. 4 (a) SEM, (b) TEM, (c) HRTEM images, (d) SAED patterns, and (e) EDX spectrum of the FeNb₂O₆ sample prepared by the hydrothermal method at 180 °C for 12 h followed with annealing at 800 °C for 2 h. The inset shows the size distribution histogram of the nanoparticles.

The SEM and TEM images of the CoNb₂O₆ sample after annealing are shown in Fig. 5a and b. It is also made up of round-like nanoparticles in the size range 50–100 nm, which is a little smaller than that of FeNb₂O₆ nanoparticles. In the HRTEM image as presented in Fig. 5c, the interplanar spacings can be measured to be 0.29 nm and 0.26 nm with a separation angle of about 75°, which can be indexed to the (311) and (2̄20) planes of columbite CoNb₂O₆ phase.³⁶ The corresponding SAED patterns are displayed in Fig. 5d, which also reveal the good single-crystalline nature of these CoNb₂O₆ nanoparticles. The EDS spectrum (Fig. 5e) indicates the presence of Co, Nb, O, Cu, C and Cr with the atomic ratio of Co : Nb : O = 1 : 2.1 : 7.28. All these results confirm the formation of columbite CoNb₂O₆ phase.

Fig. 5 (a) SEM, (b) TEM, (c) HRTEM images, (d) SAED patterns, and (e) EDX spectrum of the CoNb₂O₆ sample prepared by the hydrothermal method at 180 °C for 12 h followed with annealing at 800 °C for 2 h. The inset presents the size distribution histogram of the nanoparticles.

As to the NiNb₂O₆ sample, according to its SEM and TEM images presented in Fig. 6a and b, it is interesting that the further smaller nanoparticles with an average diameter of about 50 nm are obtained. From the TEM image, it can be observed that most nanoparticles have the diameter of about 50 nm and below, but there are still few particles with very large size. As displayed in Fig. 6c, the HRTEM image of the individual NiNb₂O₆ nanoparticle also reveals a good crystallinity, in which the observed interplanar spacings can be measured to be 0.29 nm and 0.23 nm with a separation angle of about 65°, which are reasonably consistent with the (311) and (11̄2) planes of NiNb₂O₆ columbite structure.³⁶ Accordingly, its SAED patterns are shown in Fig. 6d revealing the single-crystalline characteristic of the NiNb₂O₆ nanoparticles. The elemental analysis based on EDS spectrum presented in Fig. 6e demonstrates that the sample is composed of Ni, Nb, and O elements (Cu and C signals arise from the TEM grid, and Cr signal from the TEM sample holder), and the atomic ratio of Ni : Nb : O is 1 : 2.02 : 7.15, indicating the production of NiNb₂O₆ phase.

Fig. 6 (a) SEM, (b) TEM, (c) HRTEM images, (d) SAED patterns, and (e) EDX spectrum of the NiNb₂O₆ sample prepared by the hydrothermal method at 180 °C for 12 h followed with annealing at 800 °C for 2 h. The inset is the size distribution histogram of the nanoparticles.

3.3. Magnetic properties

The magnetic properties of the prepared ANb₂O₆ (A = Fe, Co, Ni) samples are characterized by the temperature and field dependence of magnetization. As shown in Fig. 7a, the temperature dependence of the molar susceptibility (χ) for the FeNb₂O₆ nanoparticles was measured during ZFC and FC processes from 1.8 to 300 K with an applied magnetic field of 100 Oe. It can be seen that the ZFC and FC curves are nearly superimposable at all test temperatures. With decreasing the temperature, however, the ZFC and FC curves maintain overlapped and rise sharply as below about 50 K. Actually, the two curves may be not overlapped at very low temperatures. For easy identification, the ZFC and FC curves at a temperature region lower than 5 K are enlarged for a close-up view and shown in the left inset panel. A faint bifurcation between the ZFC and FC curves can be detected. Previously, it has been well-reported that columbite FeNb₂O₆ exhibits antiferromagnetism at a very low temperature.^3,7,12 Therefore, the subtle separation between ZFC and FC magnetization curves at about 2 K probably indicates the magnetic transition. Because 1.8 K is the limiting temperature of the experiment setup, the typical feature (for example the hump) of the antiferromagnetic ordering in ZFC and FC curves cannot be completely observed.^10,12 The right inset panel of Fig. 7a shows the reciprocal of the molar magnetic susceptibility as a function of temperature for the synthesized FeNb₂O₆ nanoparticles sample. As expected, an obvious linear section at a very broad temperature region can be found in the χ⁻¹–T curve, which indicates that the product follows Curie–Weiss law in the paramagnetic state. The data can be then linearly fitted by a least-squares method to the equation χ = C/(T − Θ), where C is the Curie constant related to the effective magnetic moment (μ_eff), T is the absolute temperature and Θ is the Weiss temperature. The Curie constant can be calculated from the slope of the fitted line to be C = 3.64, while the Weiss temperature should be obtained by extrapolating the fitted line to χ⁻¹ = 0 and intercepting the T axis to be Θ = −255.82 K. The negative Θ value suggests the predominance of antiferromagnetic exchange interactions in the sample.³⁷ The value of the effective magnetic moment can be calculated to be μ_eff = 5.42μ_B, which is very close to the theoretical moment (5.4μ_B) for Fe²⁺ by considering both spin and orbital magnetism contribution.³⁸

Fig. 7 (a) The temperature dependence of the molar susceptibility curves (χ–T) of the prepared FeNb₂O₆ nanoparticles in ZFC and FC processes measured at an applied magnetic field of 100 Oe. The left inset shows the enlarged view of the low temperature region. The right inset presents the corresponding temperature dependence of the reciprocal molar magnetic susceptibility curves (χ⁻¹–T). (b) Magnetization hysteresis loops measured at 1.8, 5 and 10 K in ZFC mode for the synthesized FeNb₂O₆ nanoparticles. The inset is the corresponding zoom view of the hysteresis loops at low magnetic fields.

To further investigate the magnetic properties of the synthesized FeNb₂O₆ nanoparticles, the magnetization as a function of applied magnetic field was measured at different temperatures (1.8, 5, and 10 K) displayed in Fig. 7b. For the curve measured at 10 K, it can be found that the magnetization linearly increases with the applied field, signifying the typical paramagnetic phase, while the curve at 5 K slightly deviates from the linearity. When the measurement temperature is decreased to 1.8 K, the hysteresis loop tends to be an S-shaped line rather than the straight one, implying a possible magnetic order. Since there is still no opening of the magnetization hysteresis loop (as shown in the inset), i.e. no coercivity and remanence, it means that no ferromagnetic moment occurs and accordingly the system is approaching a magnetically ordered state at a still lower temperature, but also may be caused by the short-range correlations in a disordered state due to suppression of long-range order by frustration.¹⁰

Fig. 8a shows the thermal variation of the magnetic susceptibility for obtained CoNb₂O₆ nanoparticles. Both ZFC and FC curves have nearly the same shape. In the temperature region above 15 K, a very small magnetization associated with the existence of paramagnetic and spin-disorder states can be observed. After then, the magnetization steeply increases. At the very low temperature region, a close-up view of the ZFC and FC curves are displayed in the left inset panel. It should be noted that the ZFC and FC curves will reach a maximum and then decrease giving rise to a hump of magnetization at about 3 K suggesting a magnetic transition from paramagnetic to antiferromagnetic behavior, which should be denoted as the Néel temperature (T_N = 3 K) of the prepared CoNb₂O₆ nanoparticles, consistently with what reported in literatures.^13–18 The corresponding χ⁻¹–T curves of the CoNb₂O₆ sample are presented in the right inset panel of Fig. 8a. Based on the Curie–Weiss law in the paramagnetic state, the Curie constant of C = 2.93 and the Weiss temperature of Θ = −2.72 K can be obtained, further confirming the antiferromagnetic exchange in the product. The effective magnetic moment can be calculated to be μ_eff = 4.86μ_B, which is in good accordance with the theoretical moment (4.8μ_B) for Co²⁺ based on both spin and orbital contribution.

Fig. 8 (a) The χ–T curves of the prepared CoNb₂O₆ nanoparticles in ZFC and FC processes measured at an applied magnetic field of 100 Oe. The left inset shows the close-up view of T_N = 3 K. The right inset presents the corresponding χ⁻¹–T curves. (b) Magnetization hysteresis loops measured at 1.8, 5 and 10 K in ZFC mode for the CoNb₂O₆ nanoparticles. The inset displays the corresponding zoom view of the hysteresis loops at low magnetic fields.

The isothermal magnetization curves of CoNb₂O₆ nanoparticles measured at 1.8, 5, and 10 K are gathered in Fig. 8b. All the three M–H curves have an S-shaped profile. However, actually, the curves at 1.8 and 5 K show the opening hysteresis loops according to the corresponding zoom view at low magnetic fields as shown in the inset, which confirms the antiferromagnetic structure and the presence of a ferromagnetic moment. The coercivity (H_C) is very small and can be determined as about 35 and 18 Oe for the hysteresis loop at 1.8 and 5 K, respectively. The magnetization increases steeply at low field region and rises relatively slowly above 10 kOe, but is not completely saturated even up to 60 kOe, further suggesting the predominant antiferromagnetic ordering. For the M–H curve at 10 K, the absence of opening in hysteresis loop, i.e. the zero coercivity and remanence, suggests the presence of superparamagnetic state. Accordingly, considering the finite size of the constituent nanoparticles, the possibility of superparamagnetic blocking may be expected in the sample.

The χ–T curves under both ZFC and FC conditions of the synthesized NiNb₂O₆ nanoparticles are shown in Fig. 9a. Similar with FeNb₂O₆ and CoNb₂O₆, the ZFC and FC curves undergo almost the same behavior and show a small magnetization due to the paramagnetic state in the high temperature region. With decreasing the temperature, the ZFC and FC curves maintain overlapped and rise sharply until a maximum magnetization in ZFC curve. Subsequently, the ZFC curve decreases suddenly causing a peak of magnetization at 6 K, while the FC curve reaches a maximum value at about 5.5 K. The separation between ZFC and FC magnetization curves at 6 K indicates a characteristic irreversibility, corresponding to the transition from paramagnetic to antiferromagnetic order. Therefore, the peak at 6 K in ZFC curve can be characterized as the Néel temperature (T_N = 6 K) for the synthesized NiNb₂O₆ sample. This result is consistent with the values reported by Heid et al.⁷ and Yaeger et al.²⁰ The susceptibility data also can be fitted by a least-squares method to the Curie–Weiss equation as given in the right inset panel of Fig. 9a. Accordingly, it is paramagnetic and obeys the Curie–Weiss law in the high temperature region above 25 K. As deduced from the fitting results, the value of C and Θ for NiNb₂O₆ sample is 1.31 and −42.78, respectively. The calculated effective magnetic moment is μ_eff = 3.25μ_B, which is quite close to the theoretical value (3.2μ_B) for Ni²⁺ considering both spin and orbital contribution.

Fig. 9 (a) The χ–T curves of the prepared NiNb₂O₆ nanoparticles in ZFC and FC processes measured at an applied magnetic field of 100 Oe. The left inset shows the close-up view of T_N = 6 K. The right inset presents the corresponding χ⁻¹–T curves. (b) The overview of magnetization hysteresis loops measured at 1.8, 5 and 10 K in ZFC mode for the NiNb₂O₆ nanoparticles. The close-up view of magnetization hysteresis loops measured at (c) 1.8 K, and (d) 5 and 10 K. The insets display the corresponding zoom view at low magnetic fields.

Fig. 9b shows the magnetization vs. magnetic field plots for prepared NiNb₂O₆ nanoparticles measured at 1.8, 5, and 10 K. In all the three magnetization curves, no perfect saturation can be established even at a magnetic field of 6 T, corresponding to the antiferromagnetic ordering. However, it can be observed that in the M–H curve at the lower temperature, the magnetization exhibits a much slower linear increase with the applied magnetic field after 4 T. It is interesting that the magnetization curve at 1.8 K, as shown in Fig. 9c, not only shows the open hysteresis loop with a coercivity of about 200 Oe, but also exhibits a metamagnetic transition with a distinct change of slope at about 9 kOe, which is greatly similar to the results reported earlier by Heid et al.⁸ and Sarvezuk et al.^10,11 Such a transition has been interpreted as the result from the spin flipping of the ferromagnetic chains with increasing the applied magnetic field, which are kept in an overall antiferromagnetic structure by very weak interchain interactions.^10,11 Fig. 9d presents the close-up view of the hysteresis loops measured at 5 and 10 K. The evident opening can be observed in both curves indicating the existence of ferromagnetic moment, and the coercivity is about 125 and 85 Oe for the hysteresis loop at 5 and 10 K, respectively. This weak ferromagnetic component should arise from spin canting of the atomic magnetic moments and uncompensated spin at the surfaces of the NiNb₂O₆ nanoparticles.^7,19,20 As Néel suggested, the magnetic moments tilt toward one another and the uncompensated surface spins possess a lower coordination number. Then a net magnetic moment can appear in antiferromagnetic nanoparticles owing to surface disorder and spin canting.³⁹ Generally, the smaller size of the nanoparticles, the larger ratio of surface to volume, and then the more significant effects will be exerted on the magnetic behavior. To facilitate comparison, the magnetic data including the observed and reported Néel temperatures (T_N), the observed, reported and theoretical effective magnetic moments (μ_B), and the coercivity fields (H_C) at different temperatures of these prepared niobates are summarized in Table 1.

Table 1 Summary of the magnetic data of the prepared niobates samples

	TN (K) (observed)	TN (K) (reported)	μeff(μB) (observed)	μeff(μB) (reported)	μeff(μB) (theoretical)	HC (Oe) (1.8 K)	HC (Oe) (5 K)	HC (Oe) (10 K)
FeNb2O6	<2	4.2 (ref. 3)	5.42	5.42 (ref. 7 and 8)	5.4	—	—	—
CoNb2O6	3	2.95 (ref. 5)	4.86	4.40 (ref. 7)	4.8	35	18	—
NiNb2O6	6	6 (ref. 17)	3.25	3.30 (ref. 8)	3.2	200	125	85

---

4. Conclusions

In summary, a facile hydrothermal process followed by heat treatment has been successfully employed to fabricate nanoparticles of FeNb₂O₆, CoNb₂O₆ and NiNb₂O₆ ceramics. The magnetic properties of the products have been investigated. In the high temperature region, all the samples exhibit paramagnetic state and obey the Curie–Weiss law. When at a very low temperature, the magnetically ordered state still cannot be observed down to 1.8 K for FeNb₂O₆ sample, while the magnetic transition temperatures of T_N = 3 K and T_N = 6 K are obtained for CoNb₂O₆ and NiNb₂O₆ samples, respectively. Based on the magnetization versus temperature results, the ferromagnetic moment can be detected with the coercivity of about 35 and 18 Oe for CoNb₂O₆ sample at 1.8 and 5 K, respectively, and 200, 125 and 85 Oe for NiNb₂O₆ sample at 1.8, 5, and 10 K, respectively. Additionally, the prepared NiNb₂O₆ nanoparticles exhibit an interesting metamagnetic transition at about 9 kOe due to the spin flipping of the ferromagnetic chains. In the long term, this synthetic approach should be expected to be extendable for the general synthesis of a series of columbite niobate ceramics and their solid solutions.

Acknowledgements

Financial supports by National Natural Science Foundation of China (21461014, 21001062), the Natural Science Foundation of Jiangxi Province (20132BAB216016), and the National High Technology Research and Development Program of China (863 Program, 2014AA020539) are gratefully acknowledged.

Notes and references

1. R. C. Pullar, J. Am. Ceram. Soc., 2009, 92, 563–577.
2. D. Prabhakaran, F. R. Wondre and A. T. Boothroyd, J. Cryst. Growth, 2003, 250, 72–76.
3. H. Weitzel, Z. Anorg. Allg. Chem., 1971, 380, 119–127.
4. C. M. Morris, R. Valdés Aguilar, A. Ghosh, S. M. Koohpayeh, J. Krizan, R. J. Cava, O. Tchernyshyov, T. M. McQueen and N. P. Armitage, Phys. Rev. Lett., 2014, 112, 137403.
5. Y. Y. Zhou, M. K. Lü, Z. F. Qiu, A. Y. Zhang, Q. Ma, H. P. Zhang and Z. S. Yang, Mater. Sci. Eng., B, 2007, 140, 128–131.
6. Y. G. Su, X. Xin, Y. F. Wang, T. T. Wang and X. J. Wang, Chem. Commun., 2014, 50, 4200–4202.
7. C. Heid, H. Weitzel, F. Bourdarot, R. Calemczuk, T. Vogt and H. Fuess, J. Phys.: Condens. Matter, 1996, 8, 10609–10625.
8. C. Heid, H. Weitzel, P. Burlet, M. Bonnet, W. Gonschorek, T. Vogt, J. Norwig and H. Fuess, J. Magn. Magn. Mater., 1995, 151, 123–131.
9. C. Heid, H. Weitzel, P. Burlet, M. Winkelmann, H. Ehrenberg and H. Fuess, Phys. B, 1997, 234–236, 574–575.
10. P. W. C. Sarvezuk, E. J. Kinast, C. V. Colin, M. A. Gusmão, J. B. M. da Cunha and O. Isnard, Phys. Rev. B: Condens. Matter Mater. Phys., 2011, 83, 174412.
11. P. W. C. Sarvezuk, M. A. Gusmão, J. B. M. da Cunha and O. Isnard, Phys. Rev. B: Condens. Matter Mater. Phys., 2012, 86, 054435.
12. I. Yaeger, A. H. Morrish, B. M. Wanklyn and B. J. Garrard, Phys. Rev. B: Solid State, 1977, 16, 2289–2299.
13. H. Weitzel and S. Klein, Solid State Commun., 1973, 12, 113–116.
14. I. Maartense, I. Yaeger and B. M. Wanklyn, Solid State Commun., 1977, 21, 93–96.
15. W. Scharf and H. Weitzel, J. Magn. Magn. Mater., 1979, 13, 121–124.
16. S. Kobayashi, S. Mitsuda, M. Ishikawa, K. Miyatani and K. Kohn, Phys. Rev. B: Condens. Matter Mater. Phys., 1999, 60, 3331–3345.
17. S. Kobayashi, S. Mitsuda and K. Prokes, Phys. Rev. B: Condens. Matter Mater. Phys., 2000, 63, 024415.
18. P. W. C. Sarvezuk, E. J. Kinast, C. V. Colin, M. A. Gusmão, J. B. M. da Cunha and O. Isnard, J. Appl. Phys., 2011, 109, 07E160.
19. H. Weitzel, Acta Crystallogr., Sect. A: Found. Crystallogr., 1976, 32, 592–597.
20. I. Yaeger, A. H. Morrish and B. M. Wanklyn, Phys. Rev. B: Solid State, 1977, 15, 1465–1476.
21. R. C. Pullar, J. D. Breeze and N. McN. Alford, Key Eng. Mater., 2002, 224–226, 1–4.
22. R. C. Pullar, K. Okeneme and N. McN. Alford, J. Eur. Ceram. Soc., 2003, 23, 2479–2483.
23. R. C. Pullar, C. Vaughan and N. McN. Alford, J. Phys. D: Appl. Phys., 2004, 37, 348–352.
24. R. C. Pullar, J. D. Breeze and N. McN. Alford, J. Am. Ceram. Soc., 2005, 88, 2466–2471.
25. H. J. Lee, I. T. Kim and K. S. Hong, Jpn. J. Appl. Phys., 1997, 36, L1318–L1320.
26. H. J. Lee, K. S. Hong, S. J. Kim and I. T. Kim, Mater. Res. Bull., 1997, 32, 847–855.
27. A. Belous, O. Ovchar, B. Jancar and J. Bezjak, J. Eur. Ceram. Soc., 2007, 27, 2933–2936.
28. A. G. Belous, O. V. Ovchar, A. V. Kramarenko, D. O. Mishchuk, B. Jancar, J. Bezjak and D. Suvorov, Inorg. Mater., 2006, 42, 1369–1373.
29. M. E. Arroyo y de Dompablo, Y. L. Lee and D. Morgan, Chem. Mater., 2010, 22, 906–913.
30. A. Martínez-de la Cruz, N. L. Alcaraz, A. F. Fuentes and L. M. Torres-Martínez, J. Power Sources, 1999, 81–82, 255–258.
31. J. H. Ye, Z. G Zou and A. Matsushita, Int. J. Hydrogen Energy, 2003, 28, 651–655.
32. I. S. Mulla, N. Natarajan, A. B. Gaikwad, V. Samuel, U. N. Guptha and V. Ravi, Mater. Lett., 2007, 61, 2127–2129.
33. V. Samuel, A. B. Gaikwad, A. D. Jadhav, N. Natarajan and V. Ravi, Mater. Lett., 2007, 61, 2354–2355.
34. R. S. Ma, F. Cao, J. Wang, G. R. Zhu and G. S. Pang, Mater. Lett., 2011, 65, 2880–2882.
35. C. M. Flynn Jr and G. D. Stucky, Inorg. Chem., 1969, 8, 178–180.
36. H. Weitzel, Z. Kristallogr. – Cryst. Mater., 1976, 144, 238–258.
37. N. A. Spaldin, Magnetic Materials: Fundamentals and Applications, Cambridge University Press, Cambridge, UK, 2nd edn, 2011, p. 101.
38. N. A. Spaldin, Magnetic Materials: Fundamentals and Applications, Cambridge University Press, Cambridge, UK, 2nd edn, 2011, p. 55.
39. L. Néel, in Low Temperature Physics, ed. C. Dewitt, B. Dreyfus and P. G. de Gennes, Gordon and Breach, New York, 1962, pp. 413–440.

---