Chemical Engineering Journal 220 (2013) 45–52

Contents lists available at SciVerse ScienceDirect

Chemical Engineering Journal

journal homepage: www.elsevier .com/locate /cej

Synthesis and application of magnetic graphene/iron oxides compositefor the removal of U(VI) from aqueous solutions

1385-8947/$ - see front matter � 2013 Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.cej.2013.01.038

⇑ Corresponding author. Tel.: +86 029 82668648.E-mail address: hechaohui@mail.xjtu.edu.cn (C. He).

Pengfei Zong a, Shoufang Wang b, Yaolin Zhao a, Hai Wang a, Hui Pan a, Chaohui He a,⇑a School of Nuclear Science and Technology, Xi’an Jiaotong University, Xi’an 710049, PR Chinab Network Center, English Weekly Co. Ltd, Taiyuan 030024, PR China

h i g h l i g h t s

" A novel Fe3O4/GO was synthesized by chemical co-precipitation method." The Fe3O4/GO showed high sorption capacity towards U(VI) ions." Fe3O4/GO could be easily separated with ease using an external magnet." Fe3O4/GO exhibited high regeneration and repeated reversibility in aqueous solution.

a r t i c l e i n f o

Article history:Received 29 September 2012Received in revised form 9 January 2013Accepted 10 January 2013Available online 24 January 2013

Keywords:Fe3O4/GOU(VI)pHIon strengthTemperature

a b s t r a c t

Graphene has been extensively concerned in multidisciplinary research fields due to its remarkably phys-icochemical characteristics. Herein, magnetic graphene/iron oxides composite (Fe3O4/GO) which wassynthesized from graphene using a chemical reaction approach had been employed as a novel adsorbentfor the preconcentration and solidification of U(VI) ions from aqueous solutions. The sorption behavior ofU(VI) on the surface of Fe3O4/GO was carried out under ambient conditions such as contact time, pH andionic strength according to concentration of CU(VI)initial = 1.12 � 10�4 mol/L. The Langmuir and Freundlichmodels were adopted to simulate sorption isotherms of U(VI) at three different temperatures relying onthe concentration of CU(VI)initial = 2.25 � 10�5 to 2.24 � 10�4 mol/L, the experimental results suggestedthat the sorption reaction was favored at higher temperature. The pH-dependent and ionic strength-inde-pendent U(VI) sorption on Fe3O4/GO demonstrated that the sorption mechanism of U(VI) was inner-sphere surface complexation at low pH values, whereas the removal of U(VI) was achieved by simulta-neous precipitation and inner-sphere surface complexation at high pH values. The maximum sorptioncapacity of U(VI) on Fe3O4/GO at T = 293 K and pH = 5.5 ± 0.1 was about 69.49 mg/g higher than majorityof materials and nanomaterials reported. Magnetic separation has been considered as an effective andquick technique for separating magnetic particles, without filtration and centrifugation. The Fe3O4/GOcan be favorably separated from aqueous solution under an applied magnetic field from large volumesof aqueous solutions. The experimental results show that the Fe3O4/GO is a promising adsorbent forthe removal of radionuclides and heavy metal ions from large volumes of aqueous solution.

� 2013 Elsevier B.V. All rights reserved.

1. Introduction

Contamination of the environment with radionuclides and hea-vy toxic metal ions has been concerned throughout the world dueto application of nuclear weapons, exploiting of nuclear energy,coal combustion, application and production of phosphoric fertil-izer, etc. [1]. Uranium is commonly present in the hexavalent oxi-dation state in wastewater under the aerobic condition [2,3]. Thepredominantly radiological detriment from uranium is alpha

radiation. Uranium is one of toxic radioactivity elements due toits carcinogenic and mutagenic characteristics [4]. Uranium re-leased into the environment can be hazardous to human health,and reach the top of food chain eventually be inhaled by personsleading to detrimental impacts of human health, such as kidneydamage, liver damage and even death [5,6]. Therefore, it is signif-icant to remove uranium from wastewater before it is dischargedinto the environment. Traditional methods have been employedfor the elimination of radionuclides and toxic heavy metal ionssuch as electrodeposition, solvent extraction, coagulation, electrochemical treatment, sorption, membrane processing and reverseosmosis [7–10]. Among these approaches, sorption has been

46 P. Zong et al. / Chemical Engineering Journal 220 (2013) 45–52

widely employed to remove radionuclides and heavy metal ions inindustrial wastewater because it is cost-effective, simple operationand high efficiency. Many adsorbents, i.e., clay materials and car-bon nanotubes have been studied to remove radionuclides andheavy metal ions from wastewater [11–14]. Nevertheless, theseadsorbents cannot provide high removal efficiencies, large ion ex-change properties and high sorption capacities. To overcome thisdifficulty, nanomaterials have been concerned among the scien-tists and researchers because these materials possess high specificsurface areas, plentiful functional groups and abundant sorptionsites on their surfaces. Up to now, many nanomaterials such asactivated carbons, carbon nanotube composites, and even graph-ene [15] have been investigated to decontaminate different inor-ganic/organic pollutions from aqueous solutions, and theexperimental results demonstrate that carbon nanomaterials pos-sess high sorption capacities.

Graphene, which is consisted of extraordinarily hexagonal sp2

carbon network with a two-dimensional honeycomb lattice struc-ture, possesses many amazing characteristics such as high thermalconductivity, large surface area, remarkably optical, electrical andmechanical properties [16]. At present, graphene has been consid-ered as a prospective material for various applications comprisingsupercapacitors, biotechnology, field effect transistors and Li ionbatteries [17,18]. Generally, graphene is easily exfoliated and oxi-dized by strong oxidants leading to the formation of grapheneoxide (GO). The surface of GO can be decorated with biomoleculesand organic molecules by van der Waals forces and p–p stackinginteraction [19,20]. The two dimensional plane structure andone-atom thickness provide GO with large specific surface andhigh aspect ratio. GO, one of the most significant derivatives ofgraphene, has been obtained from the strong oxidation of graphiteusing Hummers method [21]. The functional groups such as car-boxyl, hydroxyl and epoxy are formed because varieties of oxygenatoms are available on the surface sites of GO, which makes GOstrongly hydrophilic [22]. Therefore, the oxygen-containing func-tional groups on the GO having stongly complexation capacitieswith metal ions are appropriate to remove radionuclides and heavymetal ions from wastewater [23]. In recent years, magnetic separa-tion technique has been considered as an effective and quick meth-od for separating magnetic particles. In general, the magneticnanomaterials are usually with high sorption efficiency, sorptioncapacity, and selectivity for radionuclides and heavy metal ions.What’s more, they have intrinsic magnetic properties (Fe3O4),which makes solid–liquid separation favorably and effectively un-der an applied magnetic field without centrifugation or filtration.Fe3O4/GO, which is combined the properties of graphene and ironoxides (Fe3O4), possesses some unusual properties such as largesurface to volume ratio and high conductivity [24]. Therefore,Fe3O4/GO has attracted tremendous attention among researchers,which has shown wide-ranging applications such as immobiliza-tion bioactive substances, arsenic removal, energy storage andenvironmental remediation [25,26]. Considering its amazing char-acteristics and conveniently separated process with an externalmagnetic field, Fe3O4/GO is considered as a potential adsorbentfor the removal of organic/inorganic contaminants, heavy metalions and radionuclides from large volumes of aqueous solutions.

In this work, we prepared Fe3O4/GO from natural flake graphiteand employed it as an adsorbent to remove U(VI) ions. The con-crete objectives of this work are: (1) to characterize Fe3O4/GOusing X-ray diffraction (XRD), fourier transform infrared spectros-copy (FTIR), scanning electron microscopy (SEM), potentiometricacid–base titrations, magnetization curves and transmission elec-tron microscopy (TEM); (2) to investigate the sorption kineticsand to analyze the experimental data with a pseudo-second-orderequation; (3) to study the influence of different experimental con-ditions on U(VI) sorption behaviors, such as, pH and ionic strength

based on the concentration of CU(VI)initial = 1.12 � 10�4 mol/L usingbatch technique; (4) to simulate sorption isotherms of U(VI) atthree different temperatures of 293, 318, 343 K based on the con-centration of CU(VI)initial = 2.25 � 10�5 to 2.24 � 10�4 mol/L byadopting Langmuir and Freundlich models, respectively; and (5)to presume the sorption mechanism of U(VI) sorption on the sur-face of Fe3O4/GO and to estimate the possible application ofFe3O4/GO in wastewater disposal.

2. Experimental

2.1. Materials

Graphite powder (99.92% purity, average diameter of 25 mm,Qingdao Tianhe Graphite Co. Ltd., China), 98% H2SO4, KMnO4,ammonia solution (28 wt%), FeCl2�4H2O, FeCl3�6H2O, H2O2

(30 wt%) and all other chemicals were purchased in analytical pur-ity from Sinopharm Chemical Reagent Co. Ltd.

2.2. The synthesis of GO

The sample of GO was fabricated by a modified Hummers andOffeman’s approach from natural flake graphite [27]. Typically,3.0 g of NaNO3 and 4.0 g of graphite were put into 500 mL ofH2SO4 solution under potent stirring and ice-water bath condi-tions, and then 18.0 g of KMnO4 was gradually added over approx-imately 2.5 h. The suspension was persistently churned for 5 daysbelow 25 �C. Subsequently, 560 mL of 5 wt% H2SO4 was dropwiseadded into the suspension over around 2.5 h with stirring atT = 98 �C, and then the mixtures were continuously stirred for2.5 h. When the temperature was decreased to 55 �C, 12 mL ofH2O2 (about 30 wt%) was dropwise added into the suspension byliquid-transfering gun, and then the mixtures were continuouslystirred for 2.5 h at room temperature. After centrifugation at19,000 rpm for 20 min, the solid and liquid phases were separatedby using potent stirring and bath ultrasonication for 20 min at thepower of 140 W. The ultrasonication and centrifugation were cir-cularly utilized for several times, and then the sample was washedwith distilled water until the filtrated solution became neutral. Thefinal product was parched in a vacuum oven at room temperature,thus the sample of GO was obtained.

2.3. The synthesis of Fe3O4/GO

The sample of Fe3O4/GO was prepared as follows: 0.50 g GO wasadded into a 300 mL flask of 1.56 g FeCl3�6H2O and 0.58 g FeCl2-

�4H2O in an oil bath according to the molar ratio of Fe2+:Fe3+ = 1:2under the protection of nitrogen condition for 8 h at 90 �C. Theammonia solution (28 wt%) was gradually dropped into the mix-ture solution until iron oxides particles were present. After theaddition of ammonia solution, the pH of the final solution was ad-justed to 10.5, and the reaction was permitted to maintain for 8 h.Finally, the black colored solution was filtered and rinsed with dis-tilled water until the pH of the filtrated solution reduced to 7.0, andthen dried in vacuum oven at 70 �C for 24 h. The desired productwas considered as Fe3O4/GO.

2.4. Characterization

The XRD pattern was measured on a scintag XDS-2000 diffrac-tometer with a Cu Ka source (k = 1.54178 Å). The measurementswere conducted in the 2h range of 10–70� with a scanning velocityof 2�/min. The sample of Fe3O4/GO was characterized employingFTIR spectra (Perkin Elmer spectrum 100, America) with KBr pel-lets. The spectral resolution was set to 1 cm�1, and then 150 scans

Fig. 1. Characterization of Fe3O4/GO: XRD patterns (A), FTIR spectrum (B), acid–base titration curve (C) and SEM (D).

P. Zong et al. / Chemical Engineering Journal 220 (2013) 45–52 47

were collected for each spectrum. The potentiometric acid–basetitration of Fe3O4/GO was performed employing DL50 AutomaticTitrator under argon conditions in 0.01 mol/L KNO3 backgroundelectrolyte. The potentiometric acid–base titrations curve wasshown in Fig. 1C. Briefly, 0.05 g Fe3O4/GO was added into0.01 mol/L KNO3 background electrolyte at room temperature,and purged with argon gas for 2 h to eliminate atmospheric CO2

(g). The initial pH of suspension was adjusted to pH 3.0 by adding0.01 mol/L of HNO3 under vigorous stirring for 2 h, and thenammonia was slowly dropped into the mixture solution, the pHof the final solution was adjusted to 11.0. The data which set ofpH versus the net consumption of H+ or OH� were determined tocalculate intrinsic acidity constant (pK int

a ). The pK inta demonstrated

dissociation of surface hydroxyl groups at zero surface charge.Although the parameters could not be measured directly, theycould be obtained from conditional acidity constant by combiningacid–base titration data or using extrapolation to a situation ofzero surface charge. The SEM image of Fe3O4/GO was performedon a JEOL JSM-6330F. The TEM micrograph of the sample was ob-tained with JEOL 2010 microscope. In addition, magnetic curve ofthe sample was performed on a MPMS-XL SQUID magnetometer.

2.5. Batch sorption experiments

All the sorption experiments of U(VI) were performed in poly-ethylene test tubes by using batch technique under ambient condi-tions. The U(VI) concentration was prepared to be 1.12 � 10�4 mol/L by dissolving uranyl nitrate hexahydrate (UO2(NO3)2�6H2O) withdistilled water. The stock suspensions of Fe3O4/GO and KNO3 solu-tion were pre-equilibrated for 1 day and then U(VI) stock solutionwas added in polyethylene centrifuge tubes to reach the desiredconcentrations of different components. The pH values of the

suspensions were adjusted to be in the range of 2.0–12.0 by addingneglected amount of 0.1 or 0.01 mol/L HNO3 or NaOH. The suspen-sions were oscillated for 1 day, and then solid phase was separatedfrom the solution by a permanent magnet. The preliminary exper-iments demonstrated that 24 h was enough for the suspension toreach sorption equilibrium. Subsequently, an aliquot of the suspen-sion (2.0 mL) was removed by liquid transferring gun. The sorptionof U(VI) which was negligible on the tube wall was based on theexperiment of U(VI) sorption in the absence of Fe3O4/GO. The con-centration of U(VI) ions was determined by spectrophotometryusing U(VI)–Chlorophosphonazo(III) complex at wavelength670 nm. The amount of U(VI) adsorbed on the surface of Fe3O4/GO was calculated from difference between initial and equilibriumconcentration. The special expressions were as follows,

Sorption ð%Þ ¼ C0 � Ceq

C0� 100 ð1Þ

qe ¼C0 � Ceq

m� V ð2Þ

where C0 and Ceq are the liquid-phase concentrations of U(VI) at ini-tial and equilibrium time, respectively; m is the mass of Fe3O4/GO;V is the volume of the suspension and qe is the amount of U(VI) ad-sorbed on Fe3O4/GO.

3. Results and discussion

3.1. Characterization of Fe3O4/GO

The XRD pattern of the Fe3O4/GO structure is shown in Fig. 1A.Our experimental results are consistent with the literature [28] re-ported by Liu et al. The diffraction peaks marked to (220), (311),

Fig. 3. Influence of contact time on U(VI) sorption onto Fe3O4/GO and the pseudo-second-order rate equation fit (inset). T = 293 K, pH = 5.5 ± 0.1, CU(VI)initial = 1.12 -� 10�4 mol/L, m/V = 0.3 g/L, I = 0.01 mol/L KNO3.

48 P. Zong et al. / Chemical Engineering Journal 220 (2013) 45–52

(400), (422), (511), and (440) planes emerging at 2h = 30.15�,36.27�, 43.32�, 53.89�, 57.13�, and 62.29�, respectively, are inaccordance with the standard XRD data for the cubic phase Fe3O4

with a face-centered cubic structure. The broad diffraction bandssuggest that the nanoparticles possess small sizes. Fig. 1B showsthat many oxygen-containing functional groups on the surfacesof Fe3O4/GO are described by FTIR spectroscopy. The two charac-teristic bands at 1730 cm�1 and 1620 cm�1 are contributed to theC@O, C@C group, respectively. The peaks at 1220 cm�1 and1100 cm�1 are corresponding to CAO group. The intense FeAOsorption band at 586 cm�1 corroborates that Fe3O4/GO compositeincludes magnetite. The potentiometric acid–base titration dataof Fe3O4/GO are shown in Fig. 1C. TOTH, which is total concentra-tion of consumed protons in the titration process [29], is calculatedfrom the following equation:

TOTH ¼ �ðVb � Veb1ÞV0 þ Vb

� Cb ð3Þ

where V0 represents initial volume of the suspension, Vb is the vol-ume of ammonia used in titration process at each point, Veb1 is thevolume of ammonia used in titration at Gran point to zero on theacidic side, Cb is the concentration of ammonia solution. FromFig. 1C, it is concluded that the point of zero charge (pHpzc) valueis determined to be 4.56 by using a computer controlled PC-titra-tion system (DL50 Automatic Titrator, Mettler Toledo). Huang [30]reported the pHpzc was isoelectric point, which was determinedby some electrokinetic method. In addition, Babic et al. [31] re-ported that the pHpzc was also obtained using batch equilibrium ap-proach and the experimental results indicated that pHpzc wasindependent of the concentration of ionic strength solutions. There-fore, one could obtain that the surface of Fe3O4/GO was positivecharge at pH < pHpzc, and was negative charge at pH > pHpzc. Theparticular SEM image of the Fe3O4/GO is displayed in Fig. 1D. Itcan be seen that Fe3O4 sheets are the matrix of GO nanoparticlesand then some wrinkles are observed on the surface of the Fe3O4/GO, which could be significant for loading magnetic nanoparticlesand preventing aggregation of GO. The TEM micrograph of the sam-ple is shown in Fig. 2A. It can be obtained that Fe3O4 sheet are dis-orderly distributed on the surface of GO nanoparticles, meanwhile,the average size of the Fe3O4 nanoparticles is about 20 nm. Asshown in Fig. 2B, the magnetic property of the Fe3O4/GO is investi-gated by measuring the magnetization curve at room temperature.The saturation magnetization (rs) of the Fe3O4/GO sample is31.2 emu/g (magnetic field ± 20 kOe), indicating that the Fe3O4/GOhas a high magnetism.

Fig. 2. The TEM micrograph (A) and magne

3.2. Time dependent sorption

Fig. 3 shows the sorption of U(VI) on Fe3O4/GO as a function ofcontact time at pH 5.5. The removal proportion of U(VI) increasessharply during the first 2 h, and then the sorption process attainsequilibrium after 4 h. The rapidly initial removal percentage maybe assigned to the sharp diffusion of U(VI) to the external surfacesof Fe3O4/GO from the solution. With the binding sites being filledgradually, the sorption became slow and the mechanism of kineticschanged to be totally dependent on the rate at which the sealedU(VI) ions were transformed from the bulk phase to the bindingsites. Therefore, the slow diffusion will bring about a laggard in-crease of the removal proportion at later stages [32]. As a whole,the removal process of U(VI) on Fe3O4/GO is very fast and main-tains an invariable level after a time period of 4 h. As mentionedabove, to reach and guarantee the completely experimental equi-librium in our following tests, the oscillating time is determinedto 24 h [33].

Specifically, the kinetics of the sorption reaction is analyzedemploying the pseudo-second-order rate equation [34,35], to sim-ulate the kinetics of U(VI) sorption behavior onto Fe3O4/GO:

tqt¼ 1

2Kq2e

þ 1qe

t ð4Þ

tization curve (B) of Fe3O4/GO sample.

Fig. 4. Effect of pH on the sorption of U(VI) on the surface of Fe3O4/GO. T = 293 K,CU(VI)initial = 1.12 � 10�4 mol/L, m/V = 0.3 g/L, I = 0.01 mol/L KNO3.

Fig. 5. Relative proportion of aqueous U(VI) species as a function of pH.

P. Zong et al. / Chemical Engineering Journal 220 (2013) 45–52 49

where K (g/(mg h)) is the pseudo-second-order rate coefficient ofU(VI) sorption; qe (mg/g) is the equilibrium sorption ability; andqt (mg/g) is the amount of U(VI) adsorbed on Fe3O4/GO at time t(h). The linear curve of t/qt vs. t (inset in Fig. 3) was obtained accord-ing to Eq. (4). The correlated parameter (R2 = 0.9985) and interceptof 0.083 of the linear curve demonstrated that the sorption processcould be described well by the pseudo-second-order kinetic model.

3.3. Influence of pH and ionic strength

Fig. 4 shows the U(VI) sorption on Fe3O4/GO as a function of pH in0.001, 0.01 and 0.1 mol/L KNO3 solutions, respectively. As can beseen from Fig. 4, the pH of aqueous solution has a great impact onU(VI) sorption on the surface of Fe3O4/GO. The removal percentageof U(VI) increases gradually at pH 2.0–4.0, increases quickly at pH4.0–7.0, then decreases abruptly at pH > 7.0. The increase of U(VI)sorption on Fe3O4/GO with increasing solution pH values may beattributed to the surface charges and dissociation of functionalgroups of Fe3O4/GO as well as the distributed species of U(VI) in solu-tion. From Fig. 1C, one can see that the pHpzc value of Fe3O4/GO isabout 4.56 from the acid–base titration. The surface of Fe3O4/GO be-comes positively charged owing to the protonation reaction(i.e.,„SOH + H+ ? „SOH2

+) at pH < pHpzc. Therefore, the low removalefficiency is contributed to the electrostatic repulsion betweenU(VI) and the edge function groups with positive charge (SOH2

+)on the surface of Fe3O4/GO. Whereas at pH > pHpzc, the surface ofFe3O4/GO becomes negatively charged owing to deprotonation reac-tion (i.e., „SOH ? „SO� + H+) and electrostatic attraction increasesbetween U(VI) ions and negative charge (SO�) groups on Fe3O4/GO,which leads to the increase of U(VI) sorption proportion [36]. More-over, the changing values of pH in solution are shown in Table 1. Un-der acidic conditions, the solution pH values increase a little due toprotonation reaction on the surfaces of Fe3O4/GO, while under alka-line conditions, the solution pH values decrease because of deproto-nation reaction on Fe3O4/GO. Furthermore, as more surfacefunctional groups of Fe3O4/GO are dissociated at high pH, more sorp-tion sites are favorable to restrict U(VI) ions.

The feature of U(VI) complex that dominates at a specific solutionpH may play an significant role in the removal efficiency of Fe3O4/GO

Table 1The initial pH and final pH in sorption system of U(VI) onto Fe3O4/GO.

pHinitial 2.18 3.02 3.91 4.73 5.29 5.99pHfinal 2.36 3.14 4.03 4.82 5.42 6.08

towards U(VI). In order to reasonably explain U(VI) sorption behav-ior, the relatively distributed proportion of U(VI) species is illus-trated in Fig. 5 according to the hydrolysis constants in previousliterature [37]. In our sorption experiments, no attempt is made toexclude air. All experiments are conducted at room temperature.The U(VI) concentration is prepared to be 1.12 � 10�4 mol/L by dis-solving uranyl nitrate hexahydrat (UO2(NO3)2�6H2O) with distilledwater. The conditions indicating in Fig. 5 are consistent with ourexperimental conditions. One can see that U(VI) exists in the formsof UO2þ

2 , ðUO2Þ3ðOHÞþ5 , UO2OH+, UO2CO03, ðUO2ÞðCO3Þ4�3 and

ðUO2ÞðCO3Þ2�2 at different pH values. At pH < 4.0, the dominant spe-cie is UO2þ

2 and then the removal of U(VI) is mainly dominatedthrough sorption reaction. In the range of pH 4.0–7.0, the removalproportion of U(VI) on Fe3O4/GO increases sharply. The prominentU(VI) species are UO2OH+, ðUO2Þ3ðOHÞþ5 and UO2CO0

3 at pH 4.0–7.0,therefore the removal process of U(VI) is achieved through simulta-neous precipitation of UO2CO0

3 as well as sorption of cation uranylions of UO2OH+ and ðUO2Þ3ðOHÞþ5 . Above pH > 7.0, the decrease ofU(VI) sorption on Fe3O4/GO can be attributed to negative charge ura-nyl ions species of ðUO2ÞðCO3Þ4�3 and ðUO2ÞðCO3Þ2�2 , because thesenegatively charged species are very difficult to be adsorbed ontothe negatively charged surface of Fe3O4/GO via electrostatic repul-sion. Besides, the iron dissolution experiment of the Fe3O4/GO as afunction of pH is conducted under ambient environment, and theexperimental results are shown in Fig. 6. It can be seen from Fig. 6that little amount of iron ions can be dissolved from Fe3O4/GO intoaqueous solution under alkaline conditions, therefore, the precipi-tates of Fe(OH)3 cannot be formed from the dissolution of Fe3O4 par-ticles, and thus cannot affects the removal of uranium.

It can be also seen from Fig. 4, the impact of ionic strength onU(VI) sorption on the surface of Fe3O4/GO in the wide pH rangeis considerably negligible. The ionic strength can affect interfacepotential and electrical diffused double layer thickness, thus fur-ther affect the binding species of the adsorbed U(VI) ions. Basedon triple-layer model, a probable removal mechanism of U(VI) withFe3O4/GO can be interpreted as follows: (i) If U(VI) ions react anal-ogously to a background electrolyte with the surface of Fe3O4/GO,then ion-pairs may be generated at the b-plane, where the sorptionreaction is nonspecific and the reaction product is an outer-spheresurface complex; (ii) If the sorption process of U(VI) is visualized asa especial reaction, the sorption reaction could be considered as an

6.49 7.37 8.33 9.22 10.69 11.666.57 7.14 7.98 8.85 10.13 11.29

Fig. 6. The results of iron dissolution experiment of the Fe3O4/GO composites as afunction of pH. Fig. 7. Sorption isotherms of U(VI) onto Fe3O4/GO at three different temperatures.

pH = 5.5 ± 0.1, m/V = 0.3 g/L, I = 0.01 mol/L KNO3, CU(VI)initial = 2.25 � 10�5 to2.24 � 10�4 mol/L. Symbols indicate experimental data, solid lines show the modelfitting of Langmuir equation and dash lines represent the model fitting ofFreundlich equation.

Table 2The relative parameters for Langmuir and Freundlich isotherm models of U(VI)sorption on Fe3O4/GO at different temperatures.

Sorption isotherm parameters

T(K)

Langmuir Freundlich

qmax (mol/g) b (L/mol) CC(R2)

kF (mol1�n Ln/g)

n CC(R2)

293 2.92 � 10�4 3.22 � 104 0.992 1.05 � 10�2 0.417 0.960318 3.98 � 10�4 2.98 � 104 0.994 2.31 � 10�2 0.468 0.967343 4.42 � 10�4 5.62 � 104 0.991 1.46 � 10�2 0.390 0.957

50 P. Zong et al. / Chemical Engineering Journal 220 (2013) 45–52

inner-sphere surface complex. Uranium ions can react with 1 and2 mol of active surface hydroxyl on Fe3O4/GO to form monodentateand bidentate inner-sphere complexes, respectively [38,39]. Thebackground electrolyte ions are placed at the same plane as theouter-sphere surface complexes, hence, outer-sphere surface com-plexes are more susceptible to ionic strength than inner-spheresurface complexes. Hayes and Leckie [40] determined that the im-pact of background electrolyte on sorption reaction could be usedto predict sorption reaction. b-plane sorption mechanism can besupposed to take place when the background electrolyte ions eas-ily effect the sorption reaction; however, o-plane sorption behaviormay happen. Therefore, the ionic strength independent sorptionbehavior at the whole pH range suggests that U(VI) ions take partin an o-plane complex reaction by the formation of inner-spheresurface complexation. Based on the theory mentioned above, thepH-dependent and ionic strength-independent U(VI) sorption onthe surface of Fe3O4/GO demonstrate that the sorption mechanismof U(VI) is inner-sphere surface complexation at low pH, whereasthe removal of U(VI) is achieved by coinstantaneous precipitationand inner-sphere surface complexation at high pH values.

3.4. Sorption isotherms

Generally speaking, sorption isotherms can provide some signif-icant information in optimizing the application of adsorbents.Descriptions on the interaction between sorption capacity andbond energy, adsorbents and sorbates, can be determined from iso-therm equilibrium models. The sorption isotherms of U(VI) on thesurface of Fe3O4/GO at 293, 318, and 343 K is shown in Fig. 7. It isobvious that the sorption isotherm is the highest at 343 K and isthe lowest at 293 K, which illustrates that higher temperature isfavorable for U(VI) sorption on Fe3O4/GO. There are several factorsto interpret the phenomenon: (i) the increase in temperature maycorrespondingly improve the activity and proportion of U(VI) ionsin aqueous solution, the potential charge of Fe3O4/GO and the rela-tionship of Fe3O4/GO toward U(VI) ions [41]; (ii) changes in Fe3O4/GO pore sizes as well as the increase of binding sites quantities be-cause of the destroying of some internal bonds at higher tempera-tures; and (iii) the diffusion rate of U(VI) into Fe3O4/GO pores maystrengthen with increasing temperature [42].

For sorption isotherm fitting, both the Langmuir and Freundlichmodels are commonly employed to describe the short-term, mono-component sorption characteristics and mutual interaction be-tween the amount of U(VI) adsorbed on the surface of Fe3O4/GOand the concentration of U(VI) remained in aqueous solution.

Langmuir model : qe ¼b � qmaxCeq

1þ bCeqð5Þ

Freundlich model : qe ¼ kF � Cneq ð6Þ

where qmax (mol/g) and b (L/mol) are the most significant parame-ters of Langmuir model related with sorption capacity and sorptionenergy, respectively; qe (mol/g) is amount of U(VI) adsorbed on perweight unit of Fe3O4/GO after equilibrium; kF (mol1�n Ln/g) and ncoefficients of Freundlich model are related to the sorption capacityand the intensity of dependence of sorption with equilibrium con-centration, respectively; and Ceq (mol/L) is the equilibrium concen-tration of U(VI) remained in aqueous solution.

The simulated experimental results are also demonstrated inFig. 7 using Langmuir and Freundlich models. The correlative coeffi-cients obtained from the fitting sorption isotherm models are tabu-lated in Table 2. From Fig. 7, one can also see that the two modelssimulating the sorption isotherms have no obvious discrimination,but it can be deduced that Langmuir model fitted the sorption datais better than Freundlich model from R2 parameters (see Table 2).The sorption data of U(VI) based on Langmuir isotherm illustratethat the surface of Fe3O4/GO is uniform sorption activity, thus ad-sorbed U(VI) ions do not compete with each other. The phenomenonalso demonstrates that sorption process is mainly predominated bychemoisorption [43]. Besides, Fe3O4/GO processes finite sorptioncapacity and specific surface area, the sorption of U(VI) could be bet-ter simulated by Langmuir than Freundlich model because an expo-nentially increasing is supposed in the Freundlich model. A largevalue of b illustrates strong binding of U(VI) ions on Fe3O4/GO. Thelarge kF coefficient of Freundlich model suggests that Fe3O4/GOhas a strong sorption property toward U(VI) ions. The parameter of

Table 3Comparison of U(II) sorption capacities of Fe3O4/GO with other adsorbents.

Sorbents Solution chemistryconditions

qmax

(mg/g)Refs.

Natural clinoptilolite zeolite pH = 5.0, T = 298 K 2.88 [1]Multi-walled carbon

nanotubespH = 5.0, T = 298 K 26.18 [45]

Oxidized multi-walledcarbon nanotubes

pH = 5.0, T = 298 K 33.32 [46]

Chitosan grafted MWCNTs pH = 5.0, T = 298 K 39.2 [47]Magnetic Fe3O4@SiO2 pH = 6.0, T = 298 K 52 [48]Modified carbon CMK-5 pH = 4.0, T = 283 K 65.4 [49]Fe3O4/GO pH = 5.5, T = 293 K 69.49 Present

studyGraphene oxide nanosheets pH = 5.0 T = 293 K 97.5 [50]

Fig. 8. Influence of different percentages of iron oxide and GO on U(VI) sorption.T = 293 K, pH = 5.5 ± 0.1, CU(VI)initial = 1.12 � 10�4 mol/L, m/V = 0.3 g/L, I = 0.01 mol/LKNO3.

Fig. 9. Recycled efficiency of Fe3O4/GO in the removal of U(VI). T = 293 K,pH = 5.5 ± 0.1, CU(VI)initial = 1.12 � 10�4 mol/L, m/V = 0.3 g/L, I = 0.01 mol/L KNO3.

P. Zong et al. / Chemical Engineering Journal 220 (2013) 45–52 51

n is less one, explaining that a nonlinear sorption of U(VI) occurs onthe surface of Fe3O4/GO.

3.5. Comparison of U(VI) sorption capacity with other adsorbents

In order to evaluate the potential application prospect of Fe3O4/GO, qmax, a Langmuir coefficient that has been used in the presentstudy for comparing the sorption capacity of Fe3O4/GO with otheradsorbents reported in the literatures (see Table 3). This paperemphasizes the material of Fe3O4/GO as a novel adsorbent in envi-ronmental remediation. Although a direct comparison of Fe3O4/GOwith other adsorbents is very difficult owing to different experi-mental conditions adopted, it is concluded that U(VI) sorptioncapacity of Fe3O4/GO is higher than that of natural clinoptilolitezeolite, multi-walled carbon nanotubes, chitosan grafted MWCNTs,magnetic Fe3O4@SiO2 and modified carbon CMK-5, but lower thanthat of graphene oxide nanosheets. It is necessary to point out thatthe iron oxide nanoparticles presented in the surface of Fe3O4/GOdecrease surface charge sorption area of GO, further leading to lesssorption capacity than GO. Fe3O4/GO-bound U(VI) can be favorablyand quickly separated from a solution by using the external mag-netic field. The magnetic composite of Fe3O4/GO would be a pro-spective adsorbent for the real-life preconcentration of heavymetal ions and radionuclides from large volumes of aqueous solu-tions in the removal of contamination if the Fe3O4/GO can be syn-thesized at a low price and on a large scale in the near future.

3.6. Influence of iron oxide content

To determine optimal proportion of GO and iron oxide on U(VI)sorption performance, Fe3O4/GO composite with different iron oxidecontents of 0%, 20%, 40%, 60%, 80% and 100%, respectively, were syn-thesized and used to remove U(VI) from aqueous solutions. As can beseen from Fig. 8, it is obvious that the GO has the highest sorptionperformance. Although the iron oxide can be also used as adsorbentdue to its limited sorption capacity in the removal of U(VI) [44], theexperimental results demonstrate that the sorption capacity ofFe3O4/GO decreases with increasing iron oxide content. It is worthnoting that GO has higher sorption capacity than iron oxide, whereasit is difficult to separate GO from aqueous solutions in the removal ofcontamination. The magnetic property of Fe3O4/GO can convince so-lid–liquid separation favorably and effectively under an appliedmagnetic field from large volumes of aqueous solutions in real work.Taking the two aspects involved above, one can obtain that theFe3O4/GO composite with 20% iron oxide is better than others inthe removal of U(VI) ions from aqueous solutions.

3.7. Regeneration and reversibility

In order to evaluate further the application prospect of Fe3O4/GO in removal of U(VI) from wastewater in real work, the repeated

practicability of Fe3O4/GO for U(VI) sorption was also studied viamany cycles of sorption/desorption. At the moment, the reclama-tion and utilization of Fe3O4/GO in the decontamination of U(VI)were tested. We found that the recycling percentage was alsoeffective for at least six times as demonstrated in Fig. 9 to give asatisfied experimental result even in the sixth round. The resultsof regeneration and reversibility investigation demonstrate thatthe Fe3O4/GO has a potential application prospect for the precon-centration of U(VI) from large volumes of aqueous solutions inthe removal of contamination.

4. Conclusion

In this study, the Fe3O4/GO composite can be synthesized forthe removal of uranium ions from aqueous solutions. The sampleof Fe3O4/GO was characterized by using XRD and FTIR to determineits chemical constituents and functional groups. The Fe3O4/GOcomposite had high magnetism from the magnetization curve ofthe sample. The pH-dependent and ionic strength-independentU(VI) sorption on the surface of Fe3O4/GO demonstrated that thesorption mechanism of U(VI) was inner-sphere surface complexa-tion at low pH, whereas the removal of U(VI) was achieved bysimultaneous precipitation and inner-sphere surface complexationat high pH values. Fe3O4/GO-bound U(VI) can be favorably and

52 P. Zong et al. / Chemical Engineering Journal 220 (2013) 45–52

quickly separated under an applied magnetic field from aqueoussolutions in real work. The Fe3O4/GO composite is a prospectiveadsorbent for preconcentration and solidification of radionuc-lides/heavy metal ions from large volumes of aqueous solutionsin the removal of contamination. With the development of tech-nology, the graphene will be produced at a low price and on a largescale and then graphene composites will be also employed for realwork in the future. This paper opens up promising application ofFe3O4/GO in environmental contamination remediation as well asin other research fields for instance electronic devices.

Acknowledgment

Financial support from the National Natural Science Foundationof China (11275147) is acknowledged.

References

[1] L.M. Camacho, S.G. Deng, R.R. Parra, Uranium removal from groundwater bynatural clinoptilolite zeolite: effects of pH and initial feed concentration, J.Hazard. Mater. 175 (2010) 393–398.

[2] M. Majdan, S. Pikus, A. Gajowiak, A. Gładysz-Płaska, H. Krzy _zanowska, J. _Zuk,M. Bujack, Characterization of uranium(VI) sorption by organobentonite, Appl.Surf. Sci. 256 (2010) 5416–5421.

[3] H. Shahandeh, L.R. Hossner, Role of soil properties in phytoaccumulation ofuranium, Water, Air, Soil Pollut. 141 (2002) 165–180.

[4] G.H. Wang, J.S. Liu, X.G. Wang, Z.Y. Xie, N.S. Deng, Adsorption of uranium(VI)from aqueous solution onto cross-linked chitosan, J. Hazard. Mater. 168 (2009)1053–1058.

[5] S.B. Xie, J. Yang, C. Chen, X.J. Zhang, Q.L. Wang, C. Zhang, Study on biosorptionkinetics and thermodynamics of uranium by citrobacter freundii, J. Environ.Radiact. 99 (2008) 126–133.

[6] K.G. Orloff, K. Mistry, P. Charp, S. Metcalf, R. Marino, T. Shelly, E. Melaro, A.M.Donohoe, R. Jones, Human exposure to uranium ground water, Environ. Res. 94(2004) 319–326.

[7] P. Thakur, P.N. Pathak, G.R. Choppin, Complexation thermodynamics and theformation of the binary and the ternary complexes of tetravalent plutoniumwith carboxylate and aminocarboxylate ligands in aqueous solution of highionic strength, Inorg. Chim. Acta 362 (2009) 179–184.

[8] S.B. Yang, J. Hu, C.L. Chen, D.D. Shao, X.K. Wang, Mutual effect of Pb(II) andhumic acid adsorption onto multiwalled carbon nanotubes/poly(acrylamide)composites from aqueous solution, Environ. Sci. Technol. 45 (2011) 3621–3627.

[9] X.K. Wang, C.L. Chen, J.Z. Du, X.L. Tan, D. Xu, S.M. Yu, Effect of pH and agingtime on the kinetic dissociation of 243Am(III) from humic acid coated c-Al2O3:a chelating resin exchange study, Environ. Sci. Technol. 39 (2005) 7084–7088.

[10] S.T. Yang, G.D. Sheng, X.L. Tan, J. Hu, J.Z. Du, G. Montavon, X.K. Wang,Determination of Ni(II) uptake mechanisms on mordenite surfaces: acombined macroscopic and microscopic approach, Geochim. Cosmochim.Acta 75 (2011) 6520–6534.

[11] L.J. Groven, J.A. Puszynski, Combustion synthesis and characterization of nickelaluminide–carbon nanotube composites, Chem. Eng. J. 183 (2012) 515–525.

[12] S.T. Yang, G.D. Sheng, Z.Q. Guo, X.L. Tan, J.Z. Xu, X.K. Wang, Investigation ofradionuclide 63Ni(II) sequestration mechanisms on mordenite by batch andEXAFS spectroscopy study, Sci. China Chem. 55 (2012) 632–642.

[13] C.Y. Lu, C.T. Liu, F.S. Su, Sorption kinetics, thermodynamics and competition ofNi2+ from aqueous solutions onto surface oxidized carbon nanotubes,Desalination 249 (2009) 18–23.

[14] C.L. Chen, X.K. Wang, M. Nagatsu, Europium adsorption on multiwal carbonnanotube/iron oxide magnetic composite in the presence of polyacrylic acid,Environ. Sci. Technol. 43 (2009) 2362–2367.

[15] P. Khanra, T. Kuila, N.H. Kim, S.H. Bae, D.S. Yu, J.H. Lee, Simultaneous bio-functionalization and reduction of graphene oxide by baker’s yeast, Chem. Eng.J. 183 (2012) 526–533.

[16] A.A. Balandin, S. Ghosh, W.Z. Bao, I. Calizo, D. Teweldebrhan, F. Miao, C.N. Lau,Superior thermal conductivity of single-layer graphene, Nano Lett. 8 (2008)902–907.

[17] M.D. Stoller, S.J. Park, Y.W. Zhu, J.H. An, R.S. Ruoff, Graphene-basedultracapacitors, Nano Lett. 8 (2008) 3498–3502.

[18] R. Muszynski, B. Seger, P.V. Kamat, Decorating graphene sheets with goldnanoparticles, J. Phys. Chem. C 112 (2008) 5263–5266.

[19] N. Behabtu, J.R. Lomeda, M.J. Green, A.L. Higginbotham, A. Sinitskii, D.V.Kosynkin, D. Tsentalovich, A.N.G. Parra-Vasquez, J. Schmidt, E. Kesselman, Y.Cohen, Y. Talmon, J.M. Tour, M. Pasquali, Spontaneous high-concentrationdispersions and liquid crystals of graphene, Nat. Nanotechnol. 5 (2010) 406–411.

[20] X.Y. Qi, K.Y. Pu, H. Li, X.Z. Zhou, S.X. Wu, Q.L. Fan, B. Liu, F. Boey, W. Huang, H.Zhang, Amphiphilic graphene composites, Angew. Chem. Int. Ed. 49 (2010)9426–9429.

[21] H.P. Cong, J.J. He, Y. Lu, S.H. Yu, Water soluble magnetic functionalized reducedgraphene oxide sheets: in situ synthesis and magnetic resonance imagingapplications, Small 6 (2010) 169–173.

[22] F. Barroso-Bujans, S. Cerveny, R. Verdejo, J.J.D. Val, J.M. Alberdi, A. Alegría, J.Colmenero, Permanent adsorption of organic solvents in graphite oxide and itseffect on the thermal exfoliation, Carbon 48 (2010) 1079–1087.

[23] G.X. Zhao, J.X. Li, X.M. Ren, C.L. Chen, X.K. Wang, Few-layered graphene oxidenanosheets as superior sorbents for heavy metal ion pollution management,Environ. Sci. Technol. 45 (2011) 10454–10462.

[24] X.Y. Li, X. Wang, S.Y. Song, D.P. Liu, H.J. Zhang, Selectively deposited noblemetal nanoparticles on Fe3O4/graphene composites: stable, recyclable andmagnetically separable catalysts, Chem. Eur. J. 18 (2012) 7601–7607.

[25] Y.P. He, Q.L. Sheng, J.B. Zheng, M.Z. Wang, B. Liu, Magnetite grapheme for thedirect electrochemistry of hemoglobin and its biosensing application,Electrochim. Acta 56 (2011) 2471–2476.

[26] A.H. Lu, E.L. Salabas, F. Schuth, Magnetic nanoparticles: synthesis, protection,functionalization, and application, Angew. Chem. Int. Ed. 46 (2007) 1222–1244.

[27] W.S. Hummers, R.E. Offeman, Preparation of graphitic oxide, J. Am. Chem. Soc.80 (1958) 1339–1341.

[28] M.C. Liu, C.L. Chen, J. Hu, X.L. Wu, X.K. Wang, Synthesis of Magnetite/GrapheneOxide Composite and Application for Cobalt(II) Removal, J. Phys. Chem. C 115(2011) 25234–25240.

[29] X.F. Shuai, G. Zinati, Proton charge and adsorption of humic acid andphosphate on goethite, Soil Sci. Soc. Am. J. 73 (2009) 2013–2020.

[30] C.P. Huang, Chemical interactions between inorganics and activated carbon,in: P.N. Cheremisinoff, F. Ellerbusch (Eds.), Carbon Adsorption Handbook, AnnArbor Science Publishers, Ann Arbor, 1978.

[31] B.M. Babic, S.K. Milonjic, M.J. Polovinna, B.V. Kaludierovic, Point of zero chargeand intrinsic equilibrium constants of activated carbon cloth, Carbon 37(1999) 477–481.

[32] M.H. Al-Qunaibit, W.K. Mekhemer, A.A. Zaghloul, The sorption of Cu(II) ions onbentonite-a kinetic study, J. Colloid Interf. Sci. 283 (2005) 316–321.

[33] Q.H. Fan, P. Li, Y.F. Chen, W.S. Wu, Preparation and application of attapulgite/iron oxide magnetic composites for the removal of U(VI) from aqueoussolution, J. Hazard. Mater. 192 (2011) 1851–1859.

[34] X.L. Tan, X.K. Wang, H. Geckeis, T.H. Rabung, Sorption of Eu(III) on humic acidor fulvic acid bound to alumina studied by SEM-EDS, XPS, TRLFS and batchtechniques, Environ. Sci. Technol. 42 (2008) 6532–6537.

[35] Z.Q. Guo, Y. Li, S.W. Zhang, H.H. Niu, Z.S. Chen, J.Z. Xu, Enhanced sorption ofradiocobalt from water by Bi(III) modified montmorillonite: a novel adsorbent,J. Hazard. Mater. 192 (2011) 168–175.

[36] G.D. Sheng, S.T. Yang, J. Sheng, D.D. Zhao, X.K. Wang, Influence of solutionchemistry on the removal of Ni(II) from aqueous solution to titanatenanotubes, Chem. Eng. J. 168 (2011) 178–182.

[37] D.L. Guerra, V.L. Leidens, R.R. Viana, C. Airoldi, Amazon kaolinite functionalizedwith diethylenetriamine moieties for U(VI) removal: thermodynamic ofcation-basic interactions, J. Hazard. Mater. 180 (2010) 683–692.

[38] C.H. Wu, C.F. Lin, P.Y. Horng, Adsorption of copper and lead ions ontoregenerated sludge from a water treatment plant, J. Environ. Sci. Health. A 39(2004) 237–252.

[39] C.H. Wu, Studies of the equilibrium and thermodynamics of the adsorption ofCu2+ onto as-produced and modified carbon nanotubes, J. Colloid Interf. Sci.311 (2007) 338–346.

[40] K.F. Hayes, J.O. Leckie, Modeling ionic strength effects on cation adsorption athydrous oxide/solution interfaces, J. Colloid Interf. Sci. 115 (1987) 564–572.

[41] F. Partey, D. Norman, S. Ndur, R. Nartey, Arsenic sorption onto laterite ironconcretions: temperature effect, J. Colloid Interf. Sci. 321 (2008) 493–500.

[42] H. Genç-Fuhrman, J.C. Tjell, D. McConchie, Adsorption of arsenic from waterusing activated neutralized red mud, Environ. Sci. Technol. 38 (2004) 2428–2434.

[43] Y.T. Zhou, H.L. Nie, C. Branford-White, Z.Y. He, L.M. Zhu, Removal of Cu2+ fromaqueous solution by chitosan-coated magnetic nanoparticles modified with a-ketoglutaric acid, J. Colloid Interf. Sci. 330 (2009) 29–37.

[44] L. Cláudio, B. Stopa, M. Yamaura, Uranium removal by chitosan impregnatedwith magnetite nanoparticles: adsorption and desorption, Int. J. Nucle. 5(2010) 283–289.

[45] D.D. Shao, Z.Q. Jiang, X.K. Wang, J.X. Li, Y.D. Meng, Plasma induced graftingcarboxymethyl cellulose on multiwalled carbon nanotubes for the removal ofUO2

2+ from aqueous solution, J. Phys. Chem. B 113 (2009) 860–864.[46] Y.B. Sun, S.T. Yang, G.D. Sheng, Z.Q. Guo, X.K. Wang, The removal of U(VI) from

aqueous solution by oxidized multiwalled carbon nanotubes, J. Environ.Radioact. 105 (2012) 40–47.

[47] D.D. Shao, J. Hu, X.K. Wang, Plasma induced grafting multiwalled carbonnanotube with chitosan and its application for removal of UO2þ

2 , Cu2+, and Pb2+

from aqueous solutions, Plasma Process. Polym. 7 (2010) 977–985.[48] F.L. Fan, Z. Qin, J. Bai, W.D. Rong, F.Y. Fan, W. Tian, X.L. Wu, L. Zhao, Rapid

removal of uranium from aqueous solutions using magnetic Fe3O4@SiO2

composite particles, J. Environ. Radioact. 106 (2012) 40–46.[49] G. Tian, J. Geng, Y. Jin, C. Wang, S. Li, Z. Chen, H. Wang, Y. Zhao, S. Li, Sorption of

uranium(VI) using oxime-grafted ordered mesoporous carbon CMK-5, J.Hazard. Mater. 190 (2011) 442–450.

[50] G.X. Zhao, T. Wen, X. Yang, S.B. Yang, J.L. Liao, J. Hu, D.D. Shao, X.K. Wang,Preconcentration of U(VI) ions on few-layered graphene oxide nanosheetsfrom aqueous solutions, Dalton T. 41 (2012) 6182–6188.