Phosphate group functionalized magnetic metal–organic framework nanocomposite for highly efficient removal of U(VI) from aqueous solution

The phosphate group functionalized metal-organic frameworks (MOFs) as the adsorbent for removal of U(VI) from aqueous solution still suffer from low adsorption efficiency, due to the low grafting rate of groups into the skeleton structure. Herein, a novel phosphate group functionalized metal–organic framework nanoparticles (denoted as Fe3O4@SiO2@UiO-66-TPP NPs) designed and prepared by the chelation between Zr and phytic acid, showing fast adsorption rate and outstanding selectivity in aqueous media including 10 coexisting ions. The Fe3O4@SiO2@UiO-66-TPP was properly characterized by TEM, FT-IR, BET, VSM and Zeta potential measurement. The removal performance of Fe3O4@SiO2@UiO-66-TPP for U(VI) was investigated systematically using batch experiments under different conditions, including solution pH, incubation time, temperature and initial U(VI) concentration. The adsorption kinetics, isotherm, selectivity studies revealed that Fe3O4@SiO2@UiO-66-TPP NPs possess fast adsorption rates (approximately 15 min to reach equilibrium), high adsorption capacities (307.8 mg/g) and outstanding selectivity (Su = 94.4%) towards U(VI), which in terms of performance are much better than most of the other magnetic adsorbents. Furthermore, the adsorbent could be reused for U(VI) removal without obvious loss of adsorption capacity after five consecutive cycles. The research work provides a novel strategy to assemble phosphate group-functionalized MOFs.


Preparation of silica layer coated magnetic nanoparticles (Fe 3 O 4 @SiO 2 ). Naked Fe 3 O 4 NPs were
prepared by a the facile solvothermal method. Typically, 2.70 g FeCl 3 ·6H 2 O, 7.71 g ammonium acetate, 0.8 g sodium citrate were dissolved in 140 mL EG solution under sonication to form a homogenous solution. Then the mixture was heated at 200 °C for 16 h after transferring to a Teflon-lined stainless-steel autoclave (200 mL). Under an external magnetic field, the prepared Fe 3 O 4 NPs were separated from the reaction solvent, washed with water and EtOH for several times in turn, and dried in vacuum at 40 °C. 300 mg Fe 3 O 4 NPs were dispersed in the mixed solution of 0.75 mL ammonium hydroxide, 12 mL water and 46 mL EtOH under sonication, and then 0.9 mL TEOS in 3 mL EtOH was added dropwise to the above mixed solution. The mixture was stirred magnetically for 12 h at room temperature. Under an external magnetic field, the prepared Fe 3 O 4 @SiO 2 NPs were separated from the reaction solvent and washed with water and EtOH for several times in turn, and dried in vacuum at 40 °C. Preparation of phosphate group functionalized magnetic metal-organic frameworks nanoparticles (Fe 3 O 4 @SiO 2 @UiO 2 -66-TPP). 15 mg phytic acid sodium salt was dissolved in the mixed solution of 4 mL water and 20 mL CH 3 COOH (2%, v/v), and then 40 mg Fe 3 O 4 @SiO 2 @UiO-66-NH 2 was dispersed in the above solution under sonication. After pH value of the reaction system was adjusted to 5, the mixture was heated at 50 °C for 30 min, and then kept at 115 °C for 12 h. Under an external magnetic field, the prepared

Characterizations
The synthesized nanoparticles were systematically characterized through diverse techniques including transmission electron microscopy (TEM), Fourier transform infrared spectroscopy (FT-IR) and vibration sample magnetometer (VSM). The specific surface area, total pore volume, pore size and the Zeta potentials were also measured.Transmission electron microscopy (TEM) and composition mapping images were taken with a JEM-2100 (Japan) transmission electron microscope. Fourier transform infrared (FT-IR) spectra (4000-400 cm −1 ) were obtained on a BRUKER TENSOR 27 (Germany) Fourier transform infrared spectrophotometer in KBr pellets. The N 2 adsorption-desorption isotherms were analyzed on a Micromeritics ASAP (USA) 2010 apparatus. The magnetic properties were characterized using a LDJ9600-1 (USA) vibrating sample magnetometer (VSM). The zeta potentials under different pH conditions were measured with a Brookhaven ZetaPALS (USA) analyzer at room temperature.

Batch experiments.
To evaluate the performance of the magnetic adsorbents for the removal of U(VI) from aqueous solution, batch sorption experiments were carried out in 45 mL polyethylene tubes. Briefly, 10 mg magnetic adsorbents were dispersed in 25 mL uranium solution or multi-ion solution at a certain pH value adjusted by adding negligible volume of dilute HCl or NaOH solution. After being incubated for a given time at a certain temperature, the magnetic nanocomposites loaded with U(VI) were separated under an external magnetic field. The concentration of UO 2 2+ in the supernatant was calculated by determining absorption of its complex with arsenazo(III) at 656 nm using an ultraviolet-visible spectrometer (SHIMADZU UV-1750). The concentration of metal ions including U(VI) was analyzed by inductively coupled plasma mass spectrometry (ICP-MS), when investigating the effect of coexisting ions and ionic strength on adsorption capacity of the magnetic adsorbent for U(VI). The adsorption capacity (q e (mg/g)) and adsorption percentage (% adsorption) were calculated according to the following two equations: where C 0 and C e (mg/L) are the initial and equilibrium concentrations of UO 2 2+ in the solution, respectively; V (L) is the volume of solution; and m (g) is the weight of the magnetic adsorbents.
Uranium-selectivity (S u ) to reflect the level of adsorption selectivity of the magnetic adsorbent towards U(VI), which was calculated according to the following equation: where q e(U) and q e(tol) are the U(VI) sorption capacity and all metal ions including U(VI) for the magnetic adsorbent, respectively.
The desorption study was carried out by using three kinds of acid solution (0.01 M H 2 SO 4 , HNO 3 , and HCl). 10 mg Fe 3 O 4 @SiO 2 @UiO-66-TPP NPs were shaken with 25 mL uranium solutions at pH 5.0 under ambient temperature (298 K) for 1 h. Under an external magnetic field, uranium loaded Fe 3 O 4 @SiO 2 @UiO-66-TPP NPs were separated and washed with water. Finally, the U(VI) captured by Fe 3 O 4 @SiO 2 @UiO-66-TPP NPs was released by the above-mentioned acid solutions (25 mL) at 298 K for 1 h. The nanoparticles were separated, and the U(VI) was analyzed by ICP-MS.

Results and discussions
Characterization of nanoparticles. The assembly process of phosphate group functionalized magnetic metal-organic frameworks nanoparticles (Fe 3 O 4 @SiO 2 @UiO 2 -66-TPP) is illustrated in Fig. 1. Firstly, magnetic iron oxide nanoparticles are coated with a silica layer via sol-gel emulsion method. The surface hydroxyl groups coordinate with Zr 4+ , which assists UiO 2 -66-NH 2 to settle on the surface of the nanoparticles. Finally, the nanoparticles were dotted with phytic acid through intermolecular hydrogen bonding. To investigate whether Fe 3 O 4 @SiO 2 @UiO 2 -66-TPP shown in Fig. 1 proceed successfully and the properties of the nanoparticles, TEM and composition mapping images, FT-IR, N 2 adsorption-desorption, VSM and Zeta potential were analyzed in this work.
The size and morphology of the as-prepared nanoparticles were examined by TEM. As shown in Fig. 2a, the diameter of magnetic iron oxide was 150-270 nm, which was wrapped by a 30-nm-thick silicon layer. The SiO 2 layer was coated with MOF (UiO-66-NH 2 ) layer. Zr element are uniformly dispersed on the nanoparticles. After the modification of phytic acid, the phosphate group was successfully modified on the surface of the material (Fig. 2b). In addition, Fig. 2c showed that P, Zr, and U elements are evenly distributed on nanoparticles after the  Fig. 3a. In the spectrum of Fe 3 O 4 @SiO 2 , the peak located at 590 cm −1 was assigned to the stretching vibration of Fe-O bond, the peak at 1086 cm −1 was ascribed to the Si-O-Si vibration, and the peak at 1628 cm −1 and the broad peak centered at 3431 cm −1 were ascribed to the stretching vibration of C=O bond and the stretching vibration of O-H and/or N-H bond, respectively. Compared to the above spectrum, there are some new peaks in the spectrum of Fe 3 O 4 @SiO 2 @UiO-66-NH 2 . The additional peak at 1428 cm −1 was assigned to the symmetrical stretching vibration of C-O bond, the peaked at 1506 cm −1 was assigned to the stretching vibration of C=C bond, and the peak at 1572 cm −1 was assigned to the deformation vibration of N-H bond. These bonds all come from the organic monomer (2-aminpterephthalic acid) that constructed the MOF layer. In the spectrum of Fe 3 O 4 @SiO 2 @UiO-66-TPP, the peak at 1055 cm −1 was ascribed to the stretching vibration of P=O bond.
The powder XRD pattern of Fe 3 O 4 @SiO 2 @UiO-66-TPP NPs is given in Fig. 3b. UiO-66-NH 2 was assembled sucessfully on the surface of The N 2 adsorption-desorption isotherms were employed to characterize the specific surface areas of Fe 3 O 4 @ SiO 2 @UiO 2 -66-NH 2 and Fe 3 O 4 @SiO 2 @UiO 2 -66-TPP (Fig. 3c). The N 2 sorption isotherms showed typical type I curves, suggesting a mesoporous structure. The BET surface areas of the as-synthesized nanoparticles were calculated to be 515.3 and 600.9 m 2 /g, respectively. On account of the presence of the Fe 3 O 4 @SiO 2 core, the BET surface areas of the as-synthesized nanoparticles is lower than that for UiO-66-NH 2 . The high specific area of Fe 3 O 4 @SiO 2 @UiO 2 -66-TPP NPs was beneficial for the effective removal of U(VI) from solution.
The magnetic property of Fe 3 O 4 @SiO 2 @UiO 2 -66-TPP NPs was characterized by VSM. As shown in Fig. 3d, the nanoparticles were superparamagnetic, the saturation magnetization (M s ) value was 10.15 emu g −1 , and the magnetic coercivity (Hc) value was 3.67 Oe (Fig. 3 inset). The Fe 3 O 4 @SiO 2 @UiO 2 -66-TPP NPs can be separated and redispersed effectively in aqueous solution with/without an external magnetic field, which is beneficial to its application in U(VI) removal from aqueous solution.
The surface charge properties of Fe 3 O 4 @SiO 2 @UiO-66-NH 2 and Fe 3 O 4 @SiO 2 @UiO-66-TPP NPs were evaluated through the measurement of Zeta potentials over pH range 1.5-5.5. As shown in Fig. 4, the surface charges of Fe 3 O 4 @SiO 2 @UiO-66-NH 2 NPs changed little over a large pH region, while the surface charge of Fe 3 O 4 @SiO 2 @ UiO-66-TPP NPs changed greatly, which is due to the decoration of phosphate groups on the nanoparticles. The isoelectric point for Fe 3 O 4 @SiO 2 @UiO-66-TPP NPs was about 3.1. When the pH value was higher than 3.1, the surface of the nanoparticles was negative, and the Zeta potential was − 23.3 mV at the pH 5.0. The high negative charge is beneficial for dispersion of Fe 3 O 4 @SiO 2 @UiO-66-TPP NPs, and diffusion of U(VI) ions towards the surface of the magnetic adsorbents.
Effect of solution pH. The solution pH is an important parameter that affects the adsorption performance of adsorbents towards U(VI), because it closely related to the speciation of U(VI), the surface charges and binding sites of adsorbents. Herein, the sorption studies of U(VI) on magnetic adsorbents (Fe 3 O 4 @SiO 2 @UiO-66-NH 2 ,Fe 3 O 4 @SiO 2 @UiO-66-TPP) were carried out over pH range 1.5-5.5. As shown in Fig. 5a, the adsorption capacity of Fe 3 O 4 @SiO 2 @UiO-66-TPP NPs increased rapidly with increasing pH value up to 5.0 for U(VI), followed by a decline at pH 5.5. The Fe 3 O 4 @SiO 2 @UiO-66-TPP NPs reached maximum adsorption capacity at pH 5.0, and the q e values was 247.7 mg/g. Meanwhile, the adsorption capacity of Fe 3 O 4 @SiO 2 @UiO-66-NH 2 for U(VI) had no obvious change with the increase of pH, and the maximum adsorption capacity was 32.51 mg/g at pH 5.0.  www.nature.com/scientificreports/ Uranium mainly exists in the form of UO 2 2+ in aqueous solution at or below pH 4.0 68 . When pH < 3.1, the phosphate groups of Fe 3 O 4 @SiO 2 @UiO-66-TPP NPs were positive, and intense competition between H + and UO 2 2+ for bindng sites resulted in lower adsorption capacity of the adsorbent. As pH increased, the deprotonation of phosphate groups was promoted, and U(VI) still existed in the form of positive ions, therefore the high and rapid adsorption efficiency could be attributed to the strong electrostatic interaction and chelation between phosphate groups and U(VI). At pH 5.5, the pH environment of solution is not conducive to the chelation between phosphate groups and U(VI), which resulted in the decrease of the adsorption efficiency of the adsorbent for U(VI). When pH is 6, schoepite precipitation (UO 3 ·2H 2 O) appeared in the solution, so further adsorption experiments were conducted at pH 5.0.  The pseudo-first-order, pseudo-second-order, mixed 1,2 order, intraparticle diffusion and Avrami kinetic models were used to further explore the interaction mechanism (Fig. 6).
Pseudo-first-order equation:  where q e and q t (mg/g) are the sorption amounts of U(VI) on Fe 3 O 4 @SiO 2 @UiO-66-TPP at equilibrium and at time t (min), respectively; k 1 (1/min) and k 2 (1/min) represent the rate constants of the pseudo-first-order model   www.nature.com/scientificreports/ and pseudo-second-order model, respectively; k (mg/(g min)) is the adsorption rate constant, f 2 (dimensionless) and K ip (mg/(g min 0.5 )) are the coefficient of mixed 1,2 order and intraparticle diffusion, respectively; c ip (mg/g) is the intraparticle diffusion constant; k av (1/min) is the Avrami rate constant; n av (dimensionless) is the Avrami component. Table 1 listed the fitting parameters calculated from these kinetic models (Fig. 6). The pseudo-first order, pseudo-second order, mixed 1, 2-order and Avrami models with a high correlation coefficient (R 2 > 0.98) described the adsorption process of U(VI) better than the intrsparticle difusion kinetic model (R 2 = 0.44).

Effect of initial U(VI) concentration and isotherm studies. The adsorption behavior of U(VI)
on Fe 3 O 4 @SiO 2 @UiO-66-TPP NPs was investigated by changing the initial U(VI) concentration from 1.6 to 280 mg/L. As shown in Fig. 7a, the adsorption amounts increased quickly with the enhancement of the initial U(VI) concentration from 1.6 to 78.5 mg/L. Then the adsorption reached the maximum capacity when the initial U(VI) concentration was above 200 mg/L, which is ascribed to the saturation of phosphate group binding sites embellished on Fe 3 O 4 @SiO 2 @UiO-66-TPP NPs. The most widespread isotherm models (Langmuir, Freundlich, Temkin, Sips, Toth and Langmuir-Freundlich,) were applied to describe the adsorption behavior (Fig. 7a).
Langmuir isotherm equation: where q e (mg/g) is the equilibrium adsorption capacity (mg/g), q m and q MLF (mg/g) are the maximum adsorption capacities, K L is the adsorption equilibrium constant (L/mg), K F (mg/g)/(mg/L) is the Freundlich parameter, n is the Freundlich intensity parameter, C e (mg/L) is the equilibrium concentration, b T and A T (L/g) represent the constant and equiliburium constant of Temkin isotherm model, respectively, R (8.314 J/(mol K)) is the universal gas constant, T is the absolute temperature at 298 K, n and K LF are the isotherm constants, MLF represent the heterogeneous parameter. The Langmuir and Freundlich isotherm models describes a monolayer and multilayer adsorption of targeted molecules on adsrobent surfaces [69][70][71] . Temkin model incorporates a linear variation of the adsorption enthalpy, which is an extension of Langmuir model 72 . Toth model is also an extension of traditional Langmuir model, which takes into account of heterogeneity and non-uniformity of the binding sites on the adsorbent surface 73 . Langmuir-Freundlich model represents a combination of Langmuir and Freundlich isotherm models, at low concentration it reduces to Freundlich isotherm, it predicts a Langmuir monolayer adsorption 74 . The model isotherm parameters obtained from fitting curves (Fig. 7a) were summarized in Table 2. Obviously, the adsorption process followed Langmuir-Freundlich and Langmuir isotherm models due to high correlation coefficients (R 2 ). The maximum adsorption capacities calculated from Langmuir-Freundlich and Langmuir isotherm models were very close (310.1 and 308.9 mg/g, respectively) to practical adsorption amount (307.8 mg/g).
To investigate the effect of temperature on the U(VI) adsorption capacity of the Fe 3 O 4 @SiO 2 @UiO-66-TPP NPs, the adsorption isotherms of U(VI) on Fe 3 O 4 @SiO 2 @UiO-66-TPP NPs were conducted at three temperatures (298, 313, 328 K) at pH 5.0 (Fig. 7b). The thermodynamic parameters, including ∆S 0 (entropy change), ∆H 0 (enthalpy change) and ∆G 0 (Gibbs free energy change), were calculated by Van't Hoff equation and Gibb's free energy function, to reveal whether the adsorption process was endothermic and exothermic, spontaneous or nonspontaneous.
Van't Hoff equation: Gibb's free energy function: where K d is the equilibrium constant at different temperature, R is the gas constant (8.314 J/(mol K)), ∆S 0 (J (mol K)), ∆H 0 (kJ/mol), ∆G 0 (kJ/mol) are the entropy, enthalpy and Gibbs free energy change, respectively. www.nature.com/scientificreports/ The thermodynamic parameters were listed in Table 3. The positive ∆H 0 value proved that the adsorption of U(VI) on Fe 3 O 4 @SiO 2 @UiO-66-TPP was endothermic. The positive ∆S 0 value revealed an increase in the randomness at the solid-liquid interface during the adsorption process. The negative ∆G 0 value indicated that the adsorption process was spontaneous and Fe 3 O 4 @SiO 2 @UiO-66-TPP owned high affinity toward U(VI) in aqueous solution. ∆G 0 values decreased with temperature increasing, indicating that higher temperature, the higher the spontaneous trend of spontaneous adsorption of U(VI) on Fe 3 O 4 @SiO 2 @UiO-66-TPP.
Effect of ionic strength and coexisting ions. Nuclear industrial wastewater and seawater contain many kinds of ions, which may affect the mutual interaction between Fe 3 O 4 @SiO 2 @UiO-66-TPP and U(VI). Therefore, the effects of ionic strength and coexisting ions on the adsorption selectivity and capacity of Fe 3 O 4 @ SiO 2 @UiO-66-TPP for U(VI) should be further investigated. Figure 8 shows the influence of ionic strength on the adsorption capacity of Fe 3 O 4 @SiO 2 @UiO-66-TPP. As shown in Fig. 8a, the adsorption capacity decreased  www.nature.com/scientificreports/ slightly by about 18% in the NaCl concentration range of 0.1-0.5 mol/L, which may be attributed to the decrease of ion transfer rate and the interference of electrostatic interaction caused by higher NaCl concentration. As shown in Fig. 8b, the adsorption capacity of Fe 3 O 4 @SiO 2 @UiO-66-TPP for U(VI) is significantly higher than that of other coexisting ions. The uranium-selectivity (S u ) of Fe 3 O 4 @SiO 2 @UiO-66-TPP was 94.4% calculated according to the Eq. (3), exhibiting an excellent selectivity of the magnetic adsorbent for U(VI). The mutual interaction of Fe 3 O 4 @SiO 2 @UiO-66-TPP and U(VI) depends on the chelation of U(VI) with phosphate groups grafted on the magnetic adsorbent. Phosphate group is more inclinded to coordinate with actinides than other metal ions in aqueous solutions, leading to the excellent selectivity of Fe 3 O 4 @SiO 2 @UiO-66-TPP for U(VI). Meanwhile, the phosphate group endowed Fe 3 O 4 @SiO 2 @UiO-66-TPP with high adsorption capacity (320.3 mg/g) and fast adsorption rate (15 min). The adsorption selectivity, capacity and rate of Fe 3 O 4 @SiO 2 @UiO-66-TPP are higher than other magnetic adsorbents listed in Table 4.
Desorption analysis. The selection of suitable eluent for desorption of U(VI) from the uranium-loaded Fe 3 O 4 @SiO 2 @UiO-66-TPP NPs may provide better recovery of U(VI). The desorption of U(VI) from Fe 3 O 4 @ SiO 2 @UiO-66-TPP NPs was studied in a batch mode using different eluents (HCl, H 2 SO 4 , HNO 3 ) at a concentration of 0.01 M. As shown in Fig. 9a, H 2 SO 4 was the best eluent with a elution rate of 86.0% due to the extensive protonation on the adsorbent surface, which was further used for five successive adsorption-desorption cycles under identical experimental conditions. As shown in Fig. 9b, the adsorption capacity of U(VI) on Fe 3 O 4 @ SiO 2 @UiO-66-TPP NPs decreased from 248.5 to 217.5 mg/g after five regeneration, which was approximately 12.5% reduction. Moreover, the XRD analysis result indicated that the utilized Fe 3 O 4 @SiO 2 @UiO-66-TPP NPs retained the crystal structure stability after 5 cycles (Fig. 9c). These results indicate the potential reusability of Fe 3 O 4 @SiO 2 @UiO-66-TPP NPs for U(VI) removal from aqueous medium.
Application in adsorption of uranium from pre-treated seawater. From the experimental results obtained, it can be seen that Fe 3 O 4 @SiO 2 @UiO-66-TPP NPs has a high adsorption rate and a high adsorption capacity for U(VI) and the adsorption can reach equilibrium in a very short time. Furthermore, the adsorbent exhibits a very good selectivity for uranium ions in the presence of coexisting ions. In order to evaluate the potential application of Fe 3 O 4 @SiO 2 @UiO-66-TPP NPs for U(VI) removal from seawater, we carried out the experiments on adsorption of U(VI) from the uranium-doped seawater. The natural seawater used in the adsorption experiments came from near-surface seawater from Tianjin, China. The concentrations of uranium ions in the uranium-doped seawater is 100 mg/L, and the pH of the solution was adjusted to 5.0. The adsorption capacity of U(VI) on Fe 3 O 4 @SiO 2 @UiO-66-TPP NPs from the pre-treated seawater was 228.6 mg/g, and the adsorption rates reached 91.4%, which were only a little bit lower than the adsorption capacity (249.3 mg/g) and adsorption efficiency (99.7%) from ultrapure water (Fig. 10). This demonstrates that Fe 3 O 4 @SiO 2 @UiO-66-TPP NPs have great potential application in the removal of U(VI) from radionuclide-polluted seawater.

Conclusions
In summary, a novel phosphate group functionalized magnetic metal-organic framework nanocomposite composed of magnetic Fe 3 O 4 NPs and UiO-66-NH 2 was successfully prepared and characterized by various techniques. The magnetic nanocomposite was used to remove U(VI) from aqueous solution. The nanocomposite was interspersed with phosphate group that forms a stable chelate with U(VI), and the adsorption of Fe 3 O 4 @SiO 2 @ UiO-66-TPP NPs for U(VI) reached equilibrium in 15 min, the maximum adsorption capacity is 307.8 mg/g, and the selectivity (S u ) is 94.4% in aqueous media including 10 coexisting ions. Fe 3 O 4 @SiO 2 @UiO-66-TPP NPs possess high adsorption capacities, outstanding selectivity and excellent recylability towards U(VI), which were www.nature.com/scientificreports/ www.nature.com/scientificreports/ endowed with magnetic separation performance by Fe 3 O 4 cores, but the low adsorption efficiency at pH < 3.1 under anoxic condition limited its practical applications. Neverthless, the research work provides a novel strategy to assemble phosphate group-functionalized MOFs.