Preparation of a magnetic reduced-graphene oxide/tea waste composite for high-efficiency sorption of uranium

The preparation and application of adsorptive materials with low cost and high-efficiency recovery of uranium from nuclear waste is necessary for the development of sustainable, clean energy resources and to avoid nuclear pollution. In this work, the capacity of tea waste and tea waste hybrids as inexpensive sorbents for uranium removal from water solutions was investigated. Composites of graphene oxide (GO) and tea waste (TW) exhibited a promising adsorption performance for uranium from aqueous solutions. The composites GOTW and magnetic rGO/Fe3O4/TW show high adsorption capacities (Qm (TW) = 91.72 mg/g, Qm (GOTW) = 111.61 mg/g and Qm (rGO/Fe3O4/TW) = 104.95 mg/g) and removal rates (~99%) for U(VI). The equilibrium sorption of the adsorbents fitted well to the Langmuir model, and the sorption rate fitted well to a pseudo-second-order kinetic model. The thermodynamic parameters indicated that sorption was spontaneous and favourable. The prepared adsorbents were used for the removal of uranium from real water samples as well. The results revealed that GOTW and rGO/Fe3O4/TW can be used to remediate nuclear industrial effluent as a potential adsorbent.


Results and Discussion
Characterization of the products. The FT-IR spectra of GO, Fe 3 O 4 , TW, GOTW and rGO/Fe 3 O 4 /TW re shown in Fig. 1. The IR spectrum of GO was similar to that of GO in the reference 56 and had characteristic peaks at 3345~3229, 1725, 1618, 1387, 1227 and 1061 cm −1 , corresponding to the stretching vibrations of O-H, carbonyl and carboxyl C=O, aromatic C=C and H 2 O, carboxyl O=C-O, and alkoxy C-O-C stretching vibrational modes, respectively. In the IR spectrum of TW, the broad peaks at 3328 cm −1 could come from the -OH of absorbed H 2 O (3200-3500 cm −1 ) or N-H stretching. The double peaks at 2920 and 2852 cm −1 were attributed to aliphatic carbons. The strong peaks at 1627 and 1026 cm −1 indicated the carbonyl stretching of -COOH groups and stretching vibration of the C-O groups of polysaccharides, respectively. These characteristic peaks of the prepared TW were similar to that of tea wastes reported previously 43 . However, compared to that of GO, double peaks at 2920 and 2852 cm −1 attributed to aliphatic carbons appeared in the IR spectrum of GOTW, and the intensity of the characteristic peaks at 3286 cm −1 ascribed to O-H was significantly lower, which suggests that the composite GOTW was prepared successfully. In the IR spectrum of rGO/Fe 3 O 4 /TW, there were characteristic peaks at 460 cm −1 and 350 cm −1 belonging to the Fe-O stretching vibration 57 , and similar absorption peaks as TW, whereas the characteristic peaks assigned to the stretching vibrations of O-H, carbonyl and carboxyl C=O, aromatic C=C and H 2 O, carboxyl O=C-O, and alkoxy C-O-C disappeared. Moreover, the stretching vibrations of C=C and epoxy C-O at 1550 and 1014 cm −1 of rGO functional groups were present in the IR spectrum of rGO/Fe 3 O 4 /TW 58 . These results implied that GO is reduced to rGO and that, simultaneously, Fe(II) is oxidized to Fe(III). The results showed that the magnetic composite rGO/Fe 3 O 4 /TW was prepared successfully.
The XRD patterns of the crystal phase of the prepared GO (a), Fe 3 O 4 (b), rGO/Fe 3 O 4 /TW (c), GOTW (d) and TW (e) are presented in Fig. 2. No obvious diffraction peaks at 2θ=10° and 42° attributed to the crystal planes of GO were observed, confirming the reduction of GO to rGO. The XRD pattern of TW (Fig. 2e) was consistent with that previously reported for tea 59 . Compared to those of GO and TW, differences in the peak intensities in the XRD pattern of GOTW (Fig. 2d) Fig. 2b. The diffraction peaks of the composite rGO/Fe 3 O 4 /TW (Fig. 2c) are consistent with those of Fe 3 O 4 , but the peak intensities were significantly lower with the addition of rGO and TW.
The morphologies of GO, TW and GOTW were observed by SEM. The morphology of rGO/Fe 3 O 4 /TW was analysed by AFM because of its magnetic properties. Figure 3a shows that GO exhibited a wrinkled lamellar www.nature.com/scientificreports www.nature.com/scientificreports/ structure, and the sheets of GO stacked together due to strong inter-planar interactions. As shown in Fig. 3b, an irregular layered structure with a smooth surface without pores was observed in the SEM image of TW. However, the surface of GOTW (Fig. 3c) presented many tiny pores which could be propitious to adsorb heavy metal ions 60 . In the AFM image of rGO/Fe 3 O 4 /TW (Fig. 3d) we observed that many Fe 3 O 4 particles were attached to the surface of GOTW. Moreover, as seen in the insets of Fig. 3, a significant difference in the macroscopic morphology of the prepared sorbents was observed.
Magnetic analysis of Fe 3 O 4 and rGO/Fe 3 O 4 /TW using a vibrating sample magnetometer (VSM) is shown in Fig. 4. The magnetization saturation values for Fe 3 O 4 and rGO/Fe 3 O 4 /TW were 100 and 10 emu/g, respectively. A nonlinear, reversible magnetization curve with no hysteresis exhibited characteristic super-paramagnetic behaviour. The reduced saturation magnetization was mainly due to the presence of diamagnetic GOTW surrounding    www.nature.com/scientificreports www.nature.com/scientificreports/ XPS peaks of TW and rGO/Fe 3 O 4 /TW were deconvoluted (Fig. 5c,d). Figure 5c shows the C1s spectra with three deconvoluted peaks at 283.8, 285.1 and 287.5 eV associated with C-C, C-O and C=O bonds, respectively. Figure 5(d) shows the O1s spectrum, and the three peaks at 529.6, 531.2 and 532.7 eV belong to Fe-O, C=O and C-O, respectively. The fit results are presented in Table 1. The oxygen-containing groups C-O and C=O were significantly more abundant after combination, implying that TW was successfully loaded onto the surface of GO.
Based on the results described above, a possible reaction mechanism for the composite rGO/Fe 3 O 4 /TW is illustrated in Fig. 6. First, GO combines with TW to form GOTW, and then Fe 2+ ions react with GOTW by a redox reaction to form rGOTW as shown in Eq. (2) in which GO is as an oxidant. Because the hydrolysis of Fe 3+ results in the formation of Fe(OOH), as shown in Eq. (3), the composite rGO/Fe 3 O 4 /TW was obtained after the addition of ammonia solution, as shown in Eq. (4).
Influence of pH on adsorption. The solution pH affects the speciation of uranium in solution and significantly influences the uranium adsorption process. The effect of pH on the adsorption of uranium (VI) by the synthesized adsorbents is presented in Fig. 7. The results showed a substantial impact of pH on uranium adsorption. . The competition between H 3 O + and UO 2 2+ for binding sites on the adsorbent surface results in a low sorption efficiency 61 . At pH = 5.0~7.0, the prominent species of uranium in the solution are UO 2 + , UO 2 (OH) + , (UO 2 ) 2 (OH) 2 2+ , (UO 2 ) 3 (OH) 5 + and (UO 2 ) 4 (OH) 7 + 62 . As seen in Fig. 7, the removal rate of uranium significantly increases at pH > 5.0, which can be attributed to electrostatic interactions of these complex uranium ions with negatively charged groups on the surface of TW, GOTW and rGO/Fe 3 O 4 /TW. The sorption behaviour of UO 2 2+ on GOTW and rGO/Fe 3 O 4 /TW at pH 2.0~7.0 is similar to that reported by Wang et al. 13 . Additionally, the sorption efficiency of uranium by GOTW and rGO/Fe 3 O 4 /TW was higher than that for TW, indicating that the composite of GO and TW had more efficient adsorption of uranium. Consequently, the optimum pH for U(VI) adsorption by TW, GOTW and rGO/Fe 3 O 4 /TW was 5.0.

Influence of contact time and kinetic study.
The adsorption dynamics data 63,64 were analysed based on Eqs (5) and (6). Figure 8(a) presents the time-dependent adsorption over a contact time ranging from 1 to 120 min of U(VI) by TW, GOTW and rGO/Fe 3 O 4 /TW. As seen in Fig. 8(a), the removal rate of the prepared adsorbents exceeded 96% within 1 min, and the adsorption equilibrium time was 60 min. Adsorption kinetic data of the pseudo-first-order and pseudo-second-order model at different temperature are given in Table 2. The correlation coefficient of pseudo-second-order model (R 2 = 0.9999 and 1.0000) was superior to the pseudo-first-order model, which indicated that the adsorption of U(VI) onto TW, GOTW and rGO/Fe 3 O 4 /TW fitted the pseudo-second-order model better. The fit results demonstrated that a chemical reaction played a significant role in the rate-controlling steps. The surface functional groups of the prepared adsorbents might form strong electrostatic and chemical interactions with U(VI) ions 65 . Moreover, the kinetic model fits results and parameters of rGO/Fe 3 O 4 /TW at different temperature are shown in Fig. 8(b) and Table 2, respectively. The result shows that temperature has no evident effect on the adsorption rate of rGO/Fe 3 O 4 /TW.
where k 1 (min −1 ) is the Lagergren rate constant of adsorption, and k 2 (g/(mg·min)) is the rate constant of pseudo-second-order adsorption. www.nature.com/scientificreports www.nature.com/scientificreports/ Adsorption isotherm. The Langmuir and Freundlich isotherm models are expressed in Eqs (7) and (8), respectively 66 . The fit results of Langmuir and Freundlich isotherm models of GO, TW, GOTW and rGO/Fe 3 O 4 / TW at 298 K are shown in Fig. 9(a). The Langmuir and Freundlich isotherm models of rGO/Fe 3 O 4 /TW at different temperature are shown in Fig. 9(b). The isotherm parameters calculated from fitting processes at different temperature are listed in Table 3. The Langmuir equation of the adsorbents fitted the experimental data well, with a higher correlation coefficient (R2).    where Q e (mg/g) is the equilibrium adsorption capacity, C e (mg/L) is the uranium concentration at equilibrium, Q m (mg/g) is the maximum adsorption capacity, K L (L/mg) and K F (mg 1−n L n /g) are the Langmuir and Freundlich constants, respectively, and n is the Freundlich adsorption exponent.
Moreover, the Dubinin-Radushkevich (D-R) isotherm is applied to estimate the U(VI) adsorption behaviour (chemical or physical) onto the sorbent. The D-R equation 67 is written as follow:  Table 2. Parameters of pseudo-first-order and pseudo-second-order kinetic models for U adsorption by TW, GOTW and rGO/Fe 3 O 4 /TW at different temperature.   www.nature.com/scientificreports www.nature.com/scientificreports/ where β (mol 2 /J 2 ) is activity coefficient depending on the mean free energy of adsorption and ε is the Polanyi potential (J/mol). R and T represent the gas constant (8.314 J/mol K) and absolute temperature (K), respectively. The D-R isotherm fit parameters of GO, TW, GOTW and rGO/Fe 3 O 4 /TW at different temperature are listed in Table 3. The D-R isotherm usually used the mean free energy E (kJ/mol) to assess the type of adsorption mechanism. The E value can be calculated according to Eq. (11). A value higher than 8 kJ/mol is considered to indicate chemical adsorption while it is less than 8 kJ/mol for the physical adsorption 68 . Seen from Table 3 A comparison of the Q m of different adsorbents is presented in Table 4. Table 4 shows that the Q m of the prepared rGO/Fe 3 O 4 /TW indicated a promising adsorbent for the treatment of uranium-bearing wastewater.
Effect of temperature and adsorption thermodynamics. The enthalpy (ΔH 0 ), entropy (ΔS 0 ) and standard free energy ΔG 0 from 303 K to 333 K in the adsorption process were calculated from the slope and intercept of the linear line of lnK d versus 1/T using Eqs (12)~(14) 15 . The thermodynamic parameters and the plots of lnK d versus 1/T onto TW, GOTW and rGO/Fe 3 O 4 /TW are shown in Table 5 and Fig. 10, respectively. The negative value of ΔH 0 indicated that the adsorption reaction was endothermic. The positive ΔS 0 and negative ΔG 0 sugges that the adsorption process was spontaneous.   www.nature.com/scientificreports www.nature.com/scientificreports/ Real wastewater sample analysis. To evaluate the applicability of the prepared adsorbents in this study for real uranium-bearing wastewater samples, the removal efficiency of U(VI) from four different batches of wastewater samples under the optimum adsorption conditions was investigated. The measured parameters of real nuclear wastewater and adsorption experiment results are presented in Table 6 and Fig. 11, respectively. As shown in Fig. 11, co-existing ions had no effect on the removal efficiency of GOTW and rGO/Fe 3 O 4 /TW for uranium, and they can to be applied in the treatment of uranium-containing nuclear waste effluents as potential adsorbents.   www.nature.com/scientificreports www.nature.com/scientificreports/

Materials and Methods
Materials. Stock solutions of uranium (5~150 mg/L) were prepared by dissolving UO 2 (NO 3 ) 2 ·6H 2 O (Xi'an Dingtian Chemical Reagent Co.) in deionized water (DW) and acidifying with a small amount of concentrated HNO 3 . The green tea used in this work originated from Sichuan Pingwu in China. All reagents used were of analytical grade and were used without further purification. DW was used throughout the experiments. preparation of tea wastes (tW). Green tea was washed with DW several times to remove all dirt substances. It was then boiled in DW at 80 °C for 1 h to remove coloured and soluble components, and then washed with DW until virtually colourless. The obtained TW was dried in an oven at 100 °C for 24 h. Finally, the products were crushed to powder in a pulveriser 72 . preparation of the composite GotW. GO was prepared from natural graphite by the modified Hummers method 73 . GOTW was prepared using GO and TW under ultrasonic treatment. Briefly, the mixture of GO (0.5 g) and TW (0.5 g) was dispersed in 100 ml DW under ultrasonication for 3 h. The obtained GOTW was centrifuged and washed with DW and ethanol. Finally, GOTW was dried at 50  Characterization of the products. Fourier transform infrared (FTIR) spectra of as-prepared adsorbents were obtained using an FTIR spectrometer (Bruker VERTEX 70, Germany). The crystal phases of the samples were characterized by X-ray diffractometer (XRD) (2700 model, China). The surface morphology of the products was determined using a scanning electron microscope (SEM; FEI Helios 600i, USA). The magnetic measurements of Fe 3 O 4 and rGO/Fe 3 O 4 /TW were conducted at 300 k under a varying magnetic field (PPMS-9 ECII, USA Quantum Design Co.). X-ray photoelectron spectroscopy (XPS) was studied using an ESCALAB 250 X-ray photoelectron spectroscopy (Thermo fisher, USA).
Adsorption experiments. The influence of pH, contact time, initial uranium concentration, and temperature on the removal efficiency of uranium were investigated. The solution pH was adjusted using NaOH and HCl. The prepared adsorbent was added to 20 mL U(VI) solution and shaken in a shaker (Kangshi, China). After filtration, residual uranium concentrations were measured by a micro-quantity uranium analyser (MUA model, China). The removal rate R (%) and the adsorption capacity of U(VI) Q (mg/g) were calculated according to Eqs (15) and (16), respectively: where c 0 (mg/L) is the initial uranium (VI) concentration, c t (mg/L) is the uranium concentration at time t, V (L) is the solution volume and W (g) stands for the weight of adsorbent.

Conclusion
In summary, four adsorbents GO, TW, GOTW and rGO/Fe 3 O 4 /TW were fabricated for the adsorption of uranium from aqueous solutions. The adsorption data of U(VI) by TW, GOTW and rGO/Fe 3 O 4 /TW were consistent with the Langmuir isotherm model and pseudo-second-order kinetics. The composites GOTW and rGO/ Fe 3 O 4 /TW exhibited higher adsorption capacities and faster adsorption kinetics than did GO and TW. The results showed that GOTW and rGO/Fe 3 O 4 /TW could be utilized effectively as promising sorbents to removal uranium from real multi-component uranium-containing nuclear waste effluents. Furthermore, due to the advantageous magnetic properties, rGO/Fe 3 O 4 /TW can be easily separated from aqueous solutions, thus enhancing post-treatment efficiency for further practical applications.