Removal of uranium(VI) from aqueous solution by magnetic yolk–shell iron oxide@magnesium silicate microspheres

Abstract

Yolk–shell microspheres with magnetic Fe₃O₄ cores and hierarchical magnesium silicate shells (Fe₃O₄@MS) have been successfully synthesized by combining the versatile sol–gel process and hydrothermal reaction. The as-prepared Fe₃O₄@MS microspheres were then assessed as the adsorbent for uranium(VI) removal from water, and could be easily separated by an external magnetic field. Influencing factors to adsorb uranium(VI) were investigated, including pH, ionic strength and coexisted ions, amount of adsorbent and equilibrium time. The results indicated that uranium(VI) adsorption on Fe₃O₄@MS microspheres was strongly dependent on pH and the ionic strength. The maximum adsorption capacity for uranium(VI) was calculated to be 1.51 × 10⁻⁵ mol g⁻¹ based on the Langmuir model and the experimental data fitted the Langmuir model (R² = 0.999) better than the Freundlich model (R² = 0.954). The as-prepared sub-microspheres showed their potential applications as adsorbent for highly efficient removal of heavy metal ions from wastewater.

---

1. Introduction

Environmental contamination with radionuclides and heavy toxic metal ions has been a concern throughout the world due to application of nuclear weapons, exploiting of nuclear energy, coal combustion, phosphoric fertilizer, etc.¹ Uranium is a naturally occurring radioactive element that is an irreplaceable raw material for nuclear energy. However, due to the limited natural resources for uranium,^2,3 numerous works have been conducted to recover uranium from nonconventional resources such as seawater, industrial wastewater, and other waste sources.^4–8 On the other hand, excessive amounts of uranium have entered the environment through activities of nuclear industry. The toxic nature of uranium has been a public health problem for many years.⁹ Therefore, it is urgent to remove uranium from wastewater before it is discharged into the environment. Traditional methods have been employed for the elimination of radionuclides and toxic heavy metal ions such as electrodeposition, solvent extraction, coagulation, electrochemical treatment, adsorption, membrane processing and reverse osmosis.^10–13 Among these approaches, the adsorption process of U(VI) onto various solid materials has been extensively studied, due to its environmentally friendly, cost-effective, simple operation, and highly efficient advantages.¹⁴

Recently, the rapid advance of nanoscience and nanotechnology has brought new opportunities for water treatment. Due to unique physical and chemical properties such as large specific surface areas, high adsorption capacity and fast adsorption rate, nanomaterials have shown their tremendous potential to capture inorganic or organic pollutants from water. Up to now, some nanomaterials have been focused on as adsorbents for heavy metal ions in water, and proven to be promising for environmental remediation.^15–17 Silicate materials have been used as adsorbents by many researchers for removal of some toxic metal ions.^18–23 Nevertheless, due to their easy suspension in water, it is quite difficult to remove these nanomaterials from large volumes of water, which limits their practical application. An effective strategy to solve this problem is to embed magnetic iron oxides to form magnetic nanocomposites, which provides a convenient tool for exploring magnetic separation techniques as a result of their unique magnetic response.^24–28

In the present study, we report the synthesis of Fe₃O₄@MS submicrospheres via a chemical conversion route, using Fe₃O₄@SiO₂ as a chemical template. The as-prepared Fe₃O₄@MS microsphere consists of a magnetite core and a magnesium silicate shell. The as-prepared magnetic adsorbent is applied to remove U(VI) from wastewater by investigating the following parameters, such as initial pH value of the solution, amount of adsorbent, contact time, etc.

2. Materials and methods

2.1. Materials

The chemicals UO₂(NO₃)₂·6H₂O, NaClO₄, HClO₄, HNO₃, NaOH, and Arsenaze-III were purchased with analytical purity. All chemicals were used without any further purification in the experiments. All solutions were prepared with deionized water under ambient conditions.

2.2. Synthesis of the adsorbent

2.2.1. Synthesis of Fe₃O₄ microspheres. Monodispersed Fe₃O₄ microspheres of various average sizes were synthesized according to our previous report. Typically, for the synthesis of 200 nm Fe₃O₄ microspheres, FeCl₃·6H₂O (0.54 g) NaOAc (1.5 g), and sodium acrylate (1.5 g) were dissolved in EG (20 mL) in a beaker. After vigorous stirring for 1 h, the homogeneous solution was transferred to a Teflon-lined stainless-steel autoclave (25 mL volume). Then, it was sealed and heated at 200 °C for 12 h before cooling to room temperature. The obtained Fe₃O₄ microspheres were washed with deionized water, ethanol, and dried in a vacuum oven.

2.2.2. Synthesis of Fe₃O₄@SiO₂ microspheres with core–shell nanostructure. 300 mg of the prepared Fe₃O₄ nanoparticles were dispersed in a mixture of 180 mL of ethanol and 50 mL of deionized water by ultrasonication for 10 min. Then under continuous mechanical stirring, 24 mL of ammonia solution (25%) and 1.8 mL of TEOS were added by dropwise to the mixture. The reaction was allowed to proceed at 40 °C for 6 h. The resulting products were collected and washed, and then refluxed under continuous mechanical stirring in 120 mL ethanol at 80 °C for 12 h. The solid products were collected using an external magnetic field, rinsed with deionized water, ethanol and dried in a vacuum oven at 60 °C for 4 h.

2.2.3. Synthesis of yolk-like Fe₃O₄@MS microspheres. 50 mg Fe₃O₄@SiO₂ particles, 1.35 g urea and 0.13 g Mg(NO₃)₂·6H₂O were dispersed in a mixture solution of 40 mL deionized water and 20 mL absolute ethanol. The dispersed solution was heated up to 190 °C for 24 h. The synthesized Fe₃O₄@MS were separated by centrifugation, rinsed with deionized water and absolute ethanol three times, and dried under vacuum at 80 °C for 10 h.

2.3. Characterization

The X-ray diffraction (XRD) patterns were recorded in a reflection mode (Cu Kα radiation, λ = 1.54 Å) on a Scintag XDS-2000 diffractometer. A field emission scanning electron microscope (FE-SEM, Sirion200, FEI Corp., Holland) and transmission electron microscopy (TEM, JEM-2011, JEOL, Japan) were used to determine the morphologies and microstructures. Magnetic measurements were conducted with a MPMS9XL SQUID magnetometer.

2.4. Batch adsorption experiment

The adsorption of U(VI) was investigated by using batch adsorption experiments in 15 mL polyethylene centrifuge tubes at T = 25 ± 1 °C in the presence of 0.01 mol L⁻¹ NaClO₄. NaClO₄ was usually chosen as background electrolyte, due to the ClO₄⁻ noncomplexing behavior with metal ions and numerous sorbent surfaces. The stock suspensions of sorbent and NaClO₄ were pre-equilibrated for 24 h in the dark, and then HA/FA or U(VI) stock solution was added to achieve the desired concentration of the different components, and finally HClO₄ or NaOH was added to adjust the pH. The test tubes were shaken for 2 days to reach equilibrium (preliminary experiments found that this was adequate for the suspension to reach equilibrium). After shaking for 24 h, the solid phase was separated from the liquid phase by centrifugation at 18 000 rpm for 60 min, and then the supernatant was filtered using 0.45 μm membrane filters. The concentration of U(VI) was analyzed using Arsenazo III Spectrophotometric Method at wavelength of 650 nm. The quenching effect due to the presence of organic adsorbents, which was <4%, was considered in the calculations. All experimental data were the average of duplicate determinations, and the average uncertainties were <5%.

The adsorption percentage (%) of U(VI) was calculated from the difference between the initial concentration (C₀) and the final concentration of U(VI) in the supernatant (C_e):

The concentration of U(VI) adsorbed on the solid phase (q_e) was calculated from the initial concentration (C₀), the final concentration (C_e), the volume of the suspension (V) and the mass of the sorbent (m_sorbent):

3. Results and discussion

3.1. Synthesis and characterization of Fe₃O₄@MS yolk–shell microspheres

The synthesis procedure consists of three main steps, as illustrated in Scheme 1. In step 1 the Fe₃O₄ microspheres is synthesized. Step 2 involves the uniform coating of magnetic Fe₃O₄ microspheres with a layer of SiO₂ to produce magnetic Fe₃O₄/SiO₂ core–shell particles. This silica coating step by a modified Stöber's process is highly reproducible and the silica coating acts as not only the template but also the starting material for the core–shell. Step 3 is a hydrothermal process, the urea was decomposed under hydrothermal conditions to form a stable alkaline solution. Then, the SiO₂ layer was slowly dissolved to form the silicate anion, which would preferentially react with Mg²⁺ around the magnetic Fe₃O₄/SiO₂ and produce magnesium silicate deposited on the surface of the magnetic Fe₃O₄/SiO₂. Thereafter, the dissolution/diffusion process may lead to a net material flux across the SiO₂ interface owing to the preferred outward elemental diffusion. Finally, the hierarchical core–shell magnetic Fe₃O₄@MS are achieved.

Scheme 1

Fig. 1A and B show TEM and FESEM images of the Fe₃O₄ particles, which possess spherical shapes and an average diameter of 200–300 nm. It can be clearly seen in the FESEM image that the Fe₃O₄ particles are composed of small primary nanocrystals with a very rough surface. Fig. 1C shows the FESEM image of the obtained Fe₃O₄@SiO₂ core–shell microspheres. Due to the deposition and growth of the silica layer, Fe₃O₄@SiO₂ microspheres exhibit a more regular spherical shape with smooth surface, as compared with the Fe₃O₄ particles, The Stöber method is applied to coat the Fe₃O₄ particles with a silica layer of 40–50 nm in thickness (Fig. 1D). After the hydrothermal reaction, FESEM was applied to observe the morphology of the Fe₃O₄@MS, as shown in Fig. 1E, which exhibit an urchin-like shape with an average diameter of ca. 400 nm, and consist of aligned needle-like nanosize primary particles. From a broken microsphere, a unique yolk–shell structure with an interior core, an outer shell, and void space in between can be observed (Fig. 1E). TEM further confirms the synthesized microspheres with a typical yolk–shell structure. It can be clearly seen in Fig. 1F that the microspheres are composed of a dark particle individually encapsulated in ultrafine nanoneedle-assembled shells. The average size of the microspheres is approximately 400 nm, and the shell thickness is about 90 nm. To further investigate their microstructure, elemental mapping is employed to investigate the elemental distributions in the unique yolk–shell structure, as depicted in Fig. 2. The Fe element stays in the core region, and the Mg and Si elements are detected in the shell region, while the O element can be observed in both regions. EDS analysis further indicates strong signals from Fe, O, Si, and Mg elements in the unique yolk–shell structure.

Fig. 1 FESEM and TEM images of the Fe₃O₄ microspheres (A and B), Fe₃O₄@SiO₂ (C and D) and Fe₃O₄@MS yolk–shell microspheres.

Fig. 2 The elemental mapping shows homogenous dispersion of Fe, Mg, Si and O element in the Fe₃O₄@MS yolk-cell microspheres (A) and the EDS analysis (B).

The crystallographic structure and phase purity of the synthesized products are identified by XRD. Fig. 3A shows an XRD pattern of the Fe₃O₄, Fe₃O₄@SiO₂ and Fe₃O₄@MS. The well-defined diffraction peaks (black curve) at 2θ values of 30.1, 35.4, 37.1, 43.1, 53.4, 56.9, and 62.5° can be indexed to the (220), (311), (222), (400), (422), (511) and (440) planes of the cubic inverse spinel structure of magnetite (JCPDS card no. 19-0629). The disappearance of these unique peaks indicates the successful coating of SiO₂ layer on the Fe₃O₄ (red curve), as well as the Fe₃O₄@MS yolk–shell microspheres (blue curve). Fig. 3B illustrates the FT-IR spectra of Fe₃O₄, Fe₃O₄@SiO₂ and Fe₃O₄@MS, respectively. For pure magnetite, the vibrational band at around 586 cm⁻¹ is related to the ν(Fe–O) lattice vibration. The silica coated magnetite sample shows a characteristic absorption band at 1099 cm⁻¹ and some weak absorption bands at 804 and 950 cm⁻¹, corresponding to the stretching vibrations of ν(Si–O–Si), ν(Si–OH) and ν(Si–O–Fe), respectively.³⁵ These results indicate that SiO₂ is immobilized on the surfaces of Fe₃O₄ microspheres. The FT-IR spectrum of Fe₃O₄@MS is similar to that of Fe₃O₄@SiO₂ except for the lower wavenumber shifts of some typical absorption bands, due to the changes in the microenvironment of these groups. The magnetization property of Fe₃O₄@MS was investigated at room temperature by measuring the magnetization curve, as shown in Fig. 3C. The saturation magnetization (M_s) of Fe₃O₄@MS is 49.1 emu g⁻¹ (magnetic field ±20 kOe), indicating that Fe₃O₄@MS have high magnetism, which endows Fe₃O₄@MS with an easy separation property by the external magnetic field from large volumes of aqueous solutions in real applications (Fig. 3D). In addition, the specific surface area of Fe₃O₄@MS was ∼358.5 m² g⁻¹.

Fig. 3 XRD patterns (A) and FT-IR spectra (B) of Fe₃O₄, Fe₃O₄@SiO₂ and Fe₃O₄@MS microspheres, magnetization curve of Fe₃O₄@MS microspheres (C), photo of magnetic separation (D).

3.2. Adsorption properties

3.2.1. Effect of the contact time. The effect of the contact time versus the adsorption capacity of U(VI) is screened under the following conditions: 1.0 g L⁻¹ adsorbent amount, pH 5.5 ± 0.1, 25 ± 1 °C and 2.0 × 10⁻⁵ mol L⁻¹ U(VI). The contact time varies from 0.2 to 25 h, as depicted in Fig. 4. The adsorption capacity of U(VI) onto Fe₃O₄@MS composite increases as the contact times increases and reaches adsorption equilibrium within 5 h. Kinetics of adsorption of U(VI) consisted of two phases: an initial rapid phase, where adsorption was fast, which was contributed significantly to equilibrium uptake, and a slower second phase whose contribution to the total metal adsorption was relatively small. The first phase is interpreted to be the instantaneous adsorption stage or external surface adsorption. The second phase is interpreted to be the gradual adsorption stage where intraparticle diffusion controls the adsorption rate until finally the metal uptake reaches equilibrium.²⁹ After 5 h, the adsorption capacity of U(VI) stays constant with no obvious changes. Therefore, the adsorption equilibrium time considered for the further work is set to be 5 h.

Fig. 4 Adsorption of U(VI) on Fe₃O₄@MS as a function of contact time. T = 25 ± 1 °C, pH = 5.5 ± 0.1, C_U(VI)initial = 2.0 × 10⁻⁵ mol L⁻¹, m/V = 1.0 g L⁻¹, I = 0.01 M NaClO₄.

3.2.2. Effect of the solid content. The effect of Fe₃O₄@MS content on U(VI) adsorption is shown in Fig. 5. It can be clearly seen that the concentration of U(VI) (q_e) adsorbed on Fe₃O₄@MS decreases with the Fe₃O₄@MS sorbent content increasing. This phenomenon can be attributed to the following factors: (1) the higher amount of Fe₃O₄@MS effectively reduces unsaturation of the adsorption sites and, correspondingly, the number of such sites per unit mass comes down, resulting in comparatively less adsorption at higher Fe₃O₄@MS amount; (2) when the Fe₃O₄@MS content is low, the U(VI) can easily access the adsorption sites and thus results high q_e values. With the rise in Fe₃O₄@MS content, the corresponding increase in adsorption per unit mass is less because of lower adsorption capacity utilization of the Fe₃O₄@MS, which may be ascribed to overcrowding of particles that may be termed as a kind of solid concentration effect; (3) higher Fe₃O₄@MS content creates particle aggregation, resulting in a decrease of the total surface area and an increase in the diffusional path length which contributes to the decrease in the adsorption amount.³⁰ However, the extent of the adsorption percent (%) of U(VI) increases with increasing Fe₃O₄@MS content. This is to be expected because, for a fixed initial U(VI) concentration, increasing Fe₃O₄@MS content can provide more adsorption sites and thereby increases the adsorption of U(VI).

Fig. 5 Adsorption of U(VI) on Fe₃O₄@MS as a function of the solid content. T = 25 ± 1 °C, pH = 5.5 ± 0.1, C_U(VI)initial = 2.0 × 10⁻⁵ mol L⁻¹, I = 0.01 M NaClO₄.

3.2.3. Effect of pH. In order to properly explain U(VI) adsorption behavior and mechanism, we calculated the aqueous species of U(VI) as a function of pH in the absence of sorbent according to the thermodynamic data listed in Table 1.³¹ The distribution of aqueous U(VI) species in water solution at U(VI) concentrations (2.00 × 10⁻⁵ mol L⁻¹) is presented in Fig. 6. The distribution of U(VI) species showed a dependency on pH values. Free uranyl ion (UO₂²⁺) was the dominant species at pH < 5, and then U(VI) hydrolysis complexes, and multinuclear hydroxide complexes were the dominant species in U(VI) aqueous solution. Furthermore, U(VI) species are positively charged in the pH range of 5–8, and negatively charged in the pH range above 8. In the pH range of 5–8, U(VI) species change gradually from multinuclear hydroxide complexes to hydroxide complexes with the decrease of U(VI) total concentration.³¹

Table 1 Aqueous complexation reactions of U(VI)

Reactions	log K (I = 0)
UO2^2+ + H2O = UO2(OH)^+ + H^+	−5.25
UO2^2+ + 2H2O = UO2(OH)2^0 + 2H^+	−12.15
UO2^2+ + 3H2O = UO2(OH)3^− + 3H^+	−20.25
UO2^2+ + 4H2O = UO2(OH)4^2− + 4H^+	−32.4
2UO2^2+ + H2O = (UO2)2(OH)^3+ + H^+	−2.70
2UO2^2+ + 2H2O = (UO2)2(OH)2^2+ + 2H^+	−5.62
3UO2^2+ + 5H2O = (UO2)3(OH)5^+ + 5H^+	−15.55
3UO2^2+ + 7H2O = (UO2)3(OH)7^− + 7H^+	−32.20
4UO2^2+ + 7H2O = (UO2)4(OH)7^+ + 7H^+	−21.90

---

Fig. 6 Distribution of aqueous U(VI) species as a function of the pH values.

Fig. 7 shows the adsorption of U(VI) onto Fe₃O₄@MS as a function of pH in 0.001, 0.01 and 0.1 M NaClO₄ solutions, respectively. The adsorption of U(VI) onto Fe₃O₄@MS increases abruptly at pH 3–6, and then decreases sharply at pH > 6. Similar adsorption behaviors of UO₂²⁺ on montmorillonite were reported by Hsyun et al.³² and Kowal-Fouchard et al.,³³ which was contributed to the different U(VI) species in solution at different pH values and at different U(VI) concentrations. The adsorption property of U(VI) as a function of pH value corresponded to the change of U(VI) species with varying of pH values. As can be seen from Fig. 6, U(VI) mainly exists as UO₂²⁺ at pH < 5, and then mainly exists as U(VI) hydrolysis complexes and multinuclear hydroxide complexes. From Fig. 6, (UO₂)₃(OH)₅⁺ and (UO₂)₄(OH)₇⁺ are the dominant species in 2.00 × 10⁻⁵ mol L⁻¹ U(VI) solution at pH range of 5–8. The relative proportion of (UO₂)₃(OH)₅⁺ specie decreases with increasing pH at pH range of 5–8, whereas the relative proportion of (UO₂)₃(OH)₇⁻ increases with increasing pH at pH range of 7–10. Therefore, the result that the increase adsorption of UO₂²⁺ on Fe₃O₄@MS with increasing pH at pH < 6 can be attributed to the species of UO₂²⁺ and (UO₂)₃(OH)₅⁺, whereas the decrease adsorption of UO₂²⁺ with increasing pH at high pH values is due to the decrease of the relative proportion of (UO₂)₃(OH)₅⁺ specie and the increase of the relative proportion of (UO₂)₃(OH)₇⁻ species.

Fig. 7 Effect of pH and the ionic strength on the adsorption of U(VI) on Fe₃O₄@MS, T = 25 ± 1 °C, C_U(VI)initial = 2.00 × 10⁻⁵ mol L⁻¹, m/V = 1.0 g L⁻¹.

3.2.4. Adsorption isotherms and thermodynamic studies. The adsorption isotherms for U(VI) onto Fe₃O₄@MS at 298 and 318 K are shown in Fig. 8. The adsorption isotherm at T = 318 K is higher than that at T = 338 K, indicating an endothermic process for U(VI) adsorption onto Fe₃O₄@MS, which can be explained with the following factors. With increasing temperature, the diffusion rate of U(VI) into the Fe₃O₄@MS pores will be accelerated, which results in a higher adsorption isothermal.³⁴ At elevated temperatures, the pore sizes of Fe₃O₄@MS, as well as the adsorption sites will be enlarged, due to breaking of some internal bonds near Fe₃O₄@MS surface edge, resulting in increment of the activity of U(VI) ions in solution, the affinity of U(VI) to the surface, and the charge and the potential of Fe₃O₄@MS surface.

Fig. 8 Adsorption isotherms of U(VI) adsorption onto Fe₃O₄@MS at two different temperatures, I = 0.01 mol L⁻¹ NaClO₄, pH = 5.5 ± 0.1, m/V = 1.0 g L⁻¹.

In order to get a better understanding of the adsorption mechanism, herein, the Langmuir and Freundlich isotherm equations are conducted to simulate the adsorption isotherms and to establish the relationship between the amount of U(VI) adsorbed on Fe₃O₄@MS and the concentration of U(VI) remained in the solution.

The Langmuir model assumes that adsorption occurs in a monolayer or that adsorption may only occur at a fixed number of localized sites on the surface with all adsorption sites identical and energetically equivalent. The form of the Langmuir isotherm can be represented by the following equation:³⁵

Eqn (3) can be expressed in a linear form:

where C_e is the equilibrium concentration of U(VI) in the supernatant after centrifugation (mol L⁻¹); q_e is the amount of U(VI) adsorbed on Fe₃O₄@MS after equilibrium (mol g⁻¹); q_max is the maximum amount of U(VI) at complete monolayer coverage (mol g⁻¹), and b (L mol⁻¹) is a constant that relates to the heat of adsorption.

The Freundlich expression is an exponential equation with the assumption that as the adsorbate concentration increases so too does the concentration of sorbate on the heterogeneous adsorbent surface. This model allows several kinds of adsorption sites on the solid surface and represents properly the adsorption data at low and intermediate concentrations on heterogeneous surfaces.³⁵

The model has the following form:

q_e = K_FC_eⁿ (5)

Eqn (5) can be expressed in linear form:

log q_e = log K_F + n logC_e (6)

where K_F (mol¹⁻ⁿ Lⁿ g⁻¹) represents the adsorption capacity when metal ion equilibrium concentration equals to 1, and n represents the degree of dependence of adsorption at equilibrium concentration.

The experimental data of U(VI) adsorption onto Fe₃O₄@MS are regressively fitted with the Langmuir and Freundlich models (Fig. 9). The relative parameters calculated from the two models are listed in Table 2. The correlation coefficients for both models are very close to 1. The fact that the Langmuir model fits the experimental data very well shows an almost complete monolayer coverage of the Fe₃O₄@MS particles. Moreover, Fe₃O₄@MS has a limited adsorption capacity, thus the adsorption could be better described by the Langmuir model rather than by the Freundlich model. In the Freundlich model, the value of n is from unity, which indicates that a nonlinear adsorption takes place on the heterogeneous surfaces.

Fig. 9 Langmuir (A) and Freundlich (B) isotherms for U(VI) adsorption onto Fe₃O₄@MS at two different temperatures, I = 0.01 mol L⁻¹ NaClO₄, pH = 5.5 ± 0.1, m/V = 1.0 g L⁻¹.

Table 2 The parameters for Langmuir and Freundlich adsorption isotherms of U(VI) onto Fe₃O₄@MS at different temperatures

T (K)	Langmuir			Freundlich
	qmax (mol g^−1)	b (L mol^−1)	R	KF (mol^1−n L^n g^−1)	n	R
298	1.15 × 10^−3	1.94 × 10^3	0.999	6.14 × 10^−3	0.089	0.954
318	1.51 × 10^−3	2.89 × 10^3	0.998	5.87 × 10^−3	0.077	0.940

---

The thermodynamic parameters (ΔH⁰, ΔS⁰, and ΔG⁰) for U(VI) adsorption onto Fe₃O₄@MS can be determined from the temperature dependence. Free energy change (ΔG⁰) is calculated from the relationship:

ΔG⁰ = −RT ln K⁰ (7)

where K⁰ is the adsorption equilibrium constant. The values of ln K⁰ are obtained by plotting ln K_d versus C_e (Fig. 10) and extrapolating C_e to zero.³⁶ Its intercept with the vertical axis gives the value of ln K⁰. The standard entropy change (ΔS⁰) is calculated using the equation:³⁶

Fig. 10 Linear plots of ln K_d versus q_e for U(VI) adsorption onto Fe₃O₄@MS at two different temperatures, I = 0.01 mol L⁻¹ NaClO₄, pH = 5.5 ± 0.1, m/V = 1.0 g L⁻¹.

The average standard enthalpy change (ΔH⁰) is then calculated from the relationship:

ΔH⁰ = ΔG⁰+TΔS⁰ (9)

The values obtained from eqn (7) to (9) are tabulated in Tables 3 and 4. The determination of thermodynamic parameters provides an insight into the mechanism concerning the sorptive interaction of U(VI) with Fe₃O₄@MS. The positive value of ΔH⁰ indicates that the adsorption is an endothermic process. One possible interpretation of endothermicity of the enthalpy of adsorption was that U(VI) was well solvated in water. In order for these ions to adsorb, they were to some extent denuded of their hydration sheath, and this dehydration process of ions needed energy. It was assumed that this energy of dehydration exceeded the exothermicity of the ions attaching to the surface. The removal of water molecules from ions was essentially an endothermic process, and it appeared that the endothermicity of the desolvation process exceeded that of the enthalpy of adsorption to a considerable extent. Thereby, the adsorption process is favored at higher temperature. The results also revealed that the enthalpy change (ΔH⁰) and entropy change (ΔS⁰) are positive indicating the endothermic nature of the adsorption process and the increase of randomness at the solid–liquid interface during the adsorption process, while the negative free energy change (ΔG⁰) suggests that the adsorption process is spontaneous. Furthermore, the value of ΔG⁰ becomes more negative with increasing temperature, indicating higher adsorption efficiency at higher temperature.

Table 3 Constants of linear fit of ln K_d vs. q_e (ln K_d = A + Bq_e) for U(VI) adsorption onto Fe₃O₄@MS

T (K)	A	B	R
298	18.59	−4.98 × 10^2	0.960
318	20.95	−5.49 × 10^2	0.951

---

Table 4 Values of the thermodynamic parameters for the adsorption of U(VI) onto Fe₃O₄@MS

T (K)	ΔG^0 (kJ mol^−1)	ΔH^0 (kJ mol^−1)	ΔS^0 (J mol^−1 K^−1)
298	−16.92	18.64	121.3
318	−19.26	18.73	121.3

---

As for the adsorption mechanism of Fe₃O₄@MS microspheres, it could be demonstrated as below, based on the unique hierarchically core–shell structure. In this study, the unique hierarchical core–shell structure may be divided into three layers: core layer, hollow layer, hierarchically shell layer, which rendered fast adsorption kinetics and high adsorption capacity. As we know, the concentration gradients in some specific region are the power source of diffusion in solution. So we speculate that there would be an adsorbate-consumption layer formed inside the hollow layer. In this region, U(VI) ions would be adsorbed either onto the exterior hierarchical layer (MS nanosheets) or onto the interior layer (Fe₃O₄ core), leading to the formation of a consumption layer of U(VI). In fact, the exterior hierarchical structure itself is conducive to enhanced adsorption. That is, the concentration of U(VI) in the shell layer is very high while that in the core is very low. In the hollow layer, the concentration of U(VI) is very unstably based on the diffusion theory, leading to form the consumption layer. Therefore, the consumption layer was the key as a relatively blank solution layer to sustainable power for adsorbing U(VI) from the surrounding solution (core layer and shell layer). Finally, the high adsorption efficiency was obtained.²⁸

We compared the q_m for Fe₃O₄@MS with those reported previously using different absorbents (Table 5). It can be seen that the adsorbent had the best maximum adsorption capacity of all magnetic absorbents.

Table 5 Comparison of the U(VI) sorption capacity of Fe₃O₄@MS composites with other adsorbents

Sorbents	Experimental conditions	qmax (mg g^−1)	Reference
Hematite	pH = 5.5, T = 298 K, I = 0.01 M NaClO4	5.6	37
Nanoporous zirconium phosphate	pH = 7.5, T = 295 K, I = 0.01 M NaNO3	3.3	38
Multiwalled carbon nanotubes	pH = 5.0, T = 298 K, I = 0.01 M NaClO4	26.2	39
Oxidized multiwalled carbon nanotubes	pH = 5.0, T = 298 K, I = 0.01 M NaClO4	33.3	39
Modified carbon CMK-5	pH = 4.0, T = 298 K	62	40
Graphene/iron oxides composites	pH = 5.5, T = 293 K, I = 0.01 M KNO3	69.5	41
Graphene oxide nanosheets	pH = 5.0, T = 293 K, I = 0.01 M NaClO4	97.5	42
Amidoximated hydrogel	pH = 3.0, T = 298 K	39.5	43
Quercetin modified Fe3O4 nanoparticles	pH = 3.7, T = 298 K	12.3	44
Amidoximated magnetite/graphene oxide composites	pH = 5.0, T = 298 K, I = 0.01 M NaClO4	284.9	16
Fe3O4@MS	pH = 5.5, T = 298 K, I = 0.01 M NaClO4	242.5	This study

---

4. Conclusions

The batch technique is used to study the adsorption of U(VI) from aqueous solutions onto Fe₃O₄@MS as a function of various influencing factors such as contact time, pH, ionic strength and temperature under ambient conditions. The obtained results indicate that the adsorption efficiency increases with increasing pH values at pH < 6.0, and then decreases with increasing pH values at pH > 6.0. The adsorption of U(VI) is dependent on the ionic strength at low pH values, and independent of the ionic strength at high pH values. The thermodynamic analysis derived from temperature dependent adsorption isotherms suggests that the adsorption process of U(VI) onto Fe₃O₄@MS is spontaneous and endothermic. Besides, the adsorption of U(VI) onto Fe₃O₄@MS is dominated by ion exchange or the outer-sphere surface complex at low pH values, and by the inner-sphere surface complex at high pH values. Considering the accessibility and low cost of Fe₃O₄@MS, Fe₃O₄@MS have great potential applications for the cost-effective disposal of U(VI)-contaminated wastewaters. More investigation on the adsorption property of Fe₃O₄@MS towards various environmental pollutants is ongoing in our laboratory in order to have a deeper understanding of high-efficiency materials for wastewater disposal.

Acknowledgements

Financial support from the National Natural Science Foundation of China (21207136, 21007074, 2013GB110000) is acknowledged.

References

1. L. M. Camacho, S. G. Deng and R. R. Parra, J. Hazard. Mater., 2010, 175, 393–398.
2. M. Tamada, N. Seko and F. Yoshii, Radiat. Phys. Chem., 2004, 71, 221–225.
3. S. A. Kumar, N. S. Shenoy, S. Pandey, S. Sounderajan and G. Venkateswaran, Talanta, 2008, 77, 422–426.
4. S. Das, A. K. Pandey, A. Athawale, V. Kumar, Y. K. Bhardwaj, S. Sabharwal and V. K. Manchanda, Desalination, 2008, 232, 243–253.
5. A. Zhang, G. Uchiyama and T. Asakura, React. Funct. Polym., 2005, 63, 143–153.
6. A. Zhang, T. Asakura and G. Uchiyama, React. Funct. Polym., 2003, 57, 67–76.
7. A. Mellah, S. Chegrouche and M. Barkat, J. Colloid Interface Sci., 2006, 296, 434–441.
8. T. Sakaguchi and A. Nakajima, Sep. Sci. Technol., 1986, 21, 519–534.
9. S. Yusan and S. Akyil, J. Hazard. Mater., 2008, 60, 388–395.
10. P. Thakur, P. N. Pathak and G. R. Choppin, Inorg. Chim. Acta, 2009, 362, 179–184.
11. S. B. Yang, J. Hu, C. L. Chen, D. D. Shao and X. K. Wang, Environ. Sci. Technol., 2011, 45, 3621–3627.
12. X. K. Wang, C. L. Chen, J. Z. Du, X. L. Tan, D. Xu and S. M. Yu, Environ. Sci. Technol., 2005, 39, 7084–7088.
13. T. J. Bell and Y. Ikeda, Dalton Trans., 2011, 40, 10125–10130.
14. A. Mellah, A. Silem, A. Boualia and R. Kada, Chem. Eng. Process., 1992, 31, 191–194.
15. W. Chouyyok, Y. Shin, J. Davidson, W. D. Samuels, N. H. Lafemina, R. D. Rutledge and G. E. Fryxell, Environ. Sci. Technol., 2010, 44, 6390–6395.
16. Y. G. Zhao, J. X. Li, S. W. Zhang, H. Chen and D. D. Shao, RSC. Adv., 2013, 3, 18952–18959.
17. G. B. Cai, G. X. Zhao, X. K. Wang and S. H. Yu, J. Phys. Chem. C, 2010, 114, 12948–12954.
18. F. Ciesielczyk and A. Krysztafkiewicz, J. Mater. Sci., 2007, 42, 3831–3840.
19. C. Filip, K. Andrzej and J. Teofil, Appl. Surf. Sci., 2007, 253, 8435–8442.
20. Y. Zhuang, Y. Yang and G. L. Xiang, J. Phys. Chem. C, 2009, 113, 10441–10445.
21. Y. Q. Wang, G. Z. Wang, H. Q. Wang and C. H. Liang, Chem.–Eur. J., 2010, 16, 3497–3503.
22. I. M. El-Naggar and M. M. Abou-Mesalam, J. Hazard. Mater., 2007, 149, 686–692.
23. S. Lazarevic, I. Jankovic-Castvan and D. Jovanovic, Appl. Clay Sci., 2007, 37, 47–57.
24. L. C. A. Oliveira, D. I. Petkowicz, A. Smaniotto and S. B. C. Pergher, Water Res., 2004, 38, 3699–3704.
25. X. W. Liu, Q. Y. Hu, Z. Fang, X. J. Zhang and B. B. Zhang, Langmuir, 2009, 25, 3–8.
26. H. M. Chen, X. H. Lu, C. H. Deng and X. M. Yan, J. Phys. Chem. C, 2009, 113, 21068–21073.
27. S. W. Zhang, J. X. Li, T. Wen, J. Z. Xu and X. K. Wang, RSC Adv., 2013, 3, 2754–2764.
28. S. W. Zhang, W. Q. Xu, M. Y. Zeng, J. Li, J. X. Li, J. Z. Xu and X. K. Wang, J. Mater. Chem. A, 2013, 1, 11691–11697.
29. B. Hu, W. Cheng, H. Zhang and S. Yang, J. Nucl. Mater., 2010, 406, 263–270.
30. K. G. Bhattacharyya and S. S. Gupta, Colloids Surf., A, 2008, 317, 71–79.
31. D. D. Shao, Z. Q. Jiang, X. K. Wang, J. X. Li and Y. D. Meng, J. Phys. Chem. B, 2009, 113, 860–864.
32. S. P. Hsyun, Y. H. Cho, P. S. Hahn and S. J. Kim, J. Radioanal. Nucl. Chem., 2001, 250, 55–62.
33. A. Kowal-Fouchard, R. Drot, E. Simoni and J. J. Ehrhardt, Environ. Sci. Technol., 2004, 38, 1399–1407.
34. C. L. Chen, X. K. Wang and M. Nagatsu, Environ. Sci. Technol., 2009, 43, 2362–2367.
35. J. X. Li, Z. Q. Guo, S. W. Zhang and X. K. Wang, Chem. Eng. J., 2011, 172, 892–897.
36. G. X. Zhao, H. Zhang, Q. H. Fan, X. M. Ren, J. X. Li, Y. Chen and X. K. Wang, J. Hazard. Mater., 2009, 173, 661–668.
37. D. L. Zhao, X. B. Wang, S. T. Yang, Z. Q. Guo and G. D. Sheng, J. Environ. Radioact., 2012, 103, 20–29.
38. W. Um, S. Mattigod, R. J. Serne, G. E. Fryxell, D. H. Kim and L. D. Troyer, Water Res., 2007, 41, 3217–3226.
39. Y. B. Sun, S. T. Yang, G. D. Sheng, Z. Q. Guo and X. K. Wang, J. Environ. Radioact., 2012, 105, 40–47.
40. G. Tian, J. X. Geng, Y. D. Jin, C. L. Wang, S. Q. Li, Z. Chen, H. Wang, Y. S. Zhao and S. J. Li, J. Hazard. Mater., 2011, 190, 442–450.
41. P. F. Zong, S. F. Wang, Y. L. Zhao, H. Wang, H. Pan and C. H. He, Chem. Eng. J., 2013, 220, 45–52.
42. G. X. Zhao, T. Wen, X. Yang, S. B. Yang, J. L. Liao, J. Hu, D. D. Shao and X. K. Wang, Dalton Trans., 2012, 41, 6182–6188.
43. N. Seko, A. Katakai, M. Tamada, T. Sugo and F. Yoshii, Sep. Sci. Technol., 2004, 39, 3753–3767.
44. S. Sadeghi, H. Azhdari, H. Arabi and A. Z. Moghaddam, J. Hazard. Mater., 2012, 215, 208–216.

---