Accepted Manuscript Title: Efficient removal of uranium from aqueous solution by zero-valent iron nanoparticle and its graphene composite Author: Zi-Jie Li Lin Wang Li-Yong Yuan Cheng-Liang Xiao Lei Mei Li-Rong Zheng Jing Zhang Ju-Hua Yang Yu-Liang Zhao Zhen-Tai Zhu Zhi-Fang Chai Wei Qun Shi PII: DOI: Reference:

S0304-3894(15)00116-8 http://dx.doi.org/doi:10.1016/j.jhazmat.2015.02.028 HAZMAT 16599

To appear in:

Journal of Hazardous Materials

Received date: Revised date: Accepted date:

23-10-2014 23-1-2015 9-2-2015

Please cite this article as: Zi-Jie Li, Lin Wang, Li-Yong Yuan, Cheng-Liang Xiao, Lei Mei, Li-Rong Zheng, Jing Zhang, Ju-Hua Yang, Yu-Liang Zhao, Zhen-Tai Zhu, ZhiFang Chai, Wei Qun Shi, Efficient removal of uranium from aqueous solution by zerovalent iron nanoparticle and its graphene composite, Journal of Hazardous Materials http://dx.doi.org/10.1016/j.jhazmat.2015.02.028 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Efficient removal of uranium from aqueous solution by zero-valent iron nanoparticle and its graphene composite Zi-Jie Li a, Lin Wang a, Li-Yong Yuan a, Cheng-Liang Xiao b, Lei Mei a, Li-Rong Zheng c, Jing Zhang c, Ju-Hua Yang a, Yu-Liang Zhao a, Zhen-Tai Zhu d, Zhi-Fang Chai b,*, and Wei Qun Shi a,*

a

Key Laboratory of Nuclear Radiation and Nuclear Energy Technology and Key Laboratory For

Biomedical Effects of Nanomaterials and Nanosafety, Institute of High Energy Physics, Chinese Academy of Sciences, Beijing, 100049, China b

School of Radiological and Interdisciplinary Sciences (RAD-X), and Collaborative Innovation

Center of Radiation Medicine of Jiangsu Higher Education Institutions, Soochow University, Suzhou 215123, China c

Beijing Synchrotron Radiation Facility, Institute of High Energy Physics, Chinese Academy of

Sciences, Beijing, 100049, China d

State Key Laboratory of NBC Protection for Civilian, Beijing, 102205, China

* Corresponding author. Tel.: +86-10-88233968 E-mail: [email protected] (W.Q. Shi) and [email protected] (Z.F. Chai) Highlights

Uranium removal by ZVI-nps: independent of pH, the presence of CO32-, humic acid, or mimic groundwater constituents Rapid removal kinetics and sorption capacity of ZVI-nps is 8173 mg U/g Two reaction mechanisms: sufficient Fe0→reductive precipitation as U3O7; insufficient Fe0→hydrolysis precipitation of U(VI)

Fe/graphene composites: improved kinetics and higher U(VI) reduction ratio Abstract

Zero-valent iron nanoparticle (ZVI-np) and its graphene composites were prepared and applied in the removal of uranium under anoxic conditions. It was found that solutions containing 24 ppm U(VI) could be completely cleaned up by ZVI-nps, regardless of the presence of NaHCO3, humic acid, mimic groundwater constituents or the change of solution pH from 5 to 9, manifesting the promising potential of this reactive material in permeable reactive barrier (PRB) to remediate uranium-contaminated groundwater. In the measurement of maximum sorption capacity, removal efficiency of uranium kept at 100% until C0(U)=643 ppm, and the saturation sorption of 8173 mg U/g ZVI-nps was achieved at C0(U)=714 ppm. In addition, reaction mechanisms were clarified based on the results of SEM, XRD, XANES, and chemical leaching in (NH4)2CO3 solution. Partially reductive precipitation of U(VI) as U3O7 was prevalent when sufficient iron was available; nevertheless, hydrolysis precipitation of U(VI) on surface would be predominant as iron got insufficient, characterized by releases of Fe2+ ions. The dissolution of Fe0 cores was assigned to be the driving force of continuous formation of U(VI) (hydr)oxide. The incorporation of graphene supporting matrix was found to facilitate faster removal rate and higher U(VI) reduction ratio, thus benefitting the long-term immobilization of uranium in geochemical environment.

Keywords: Zero-valent iron nanoparticles; graphene composites; uranium removal; reductive precipitation; hydrolysis precipitation.

1. Introduction

Uranium occurs in nature in primary deposits, U(IV)O2. Currently, uranium is being used as a typical nuclear fuel to enhance electrical production ability. With uranium ore mining, processing, fuel manufacture, spent fuel reprocessing and other related activities, more and more highly mobile U(VI) is released into the environment, making uranium a common contaminant to soils, surface and groundwater [1]. The concentration of uranium in some acid mine water and contaminated areas around waste disposal sites can even attain as high as several tens of ppm [2,3].

On the other hand, the allowed maximum level of uranium for drinking water is only 30 µg/L, recommended by EPA [4]. It is well known that intakes of uranium from food and/or drinking water can lead to internal irradiation and/or chemical toxicity. Long term exposure to uranium may result in cancer, kidney and liver damages, or all.

Permeable reactive barrier (PRB) technology has been successfully utilized to in situ remediate contaminated groundwater. Compared to conventional pump-and-treat, dig-and-treat, and containment technologies, PRB technology has many advantages, by which reactive media filled can adsorb, degrade, and/or precipitate various pollutants as contaminant plumes flow through the subsurface treatment wall. Therefore, it has been regarded to be technically attractive and cost effective [5,6]. Zero-valent iron, metallic iron (Fe0), as the reactive material, has been extensively investigated to remove heavy metals such as Pb2+, Cu2+, AsO43-, CrO42-, Ni2+, Zn2+, Cd2+, Ba2+ [7 and references therein], radionuclides e.g. TcO4- [8] and UO22+ [7,9], and to degrade halogenated-hydrocarbon compounds from contaminated areas. Removal mechanism of uranium by Fe0 was generally described as i) Fe0 and structural Fe(II) ions reduce U(VI) ions to sparingly soluble UO2, precipitating on the surface of iron, ii) physical and/or chemical adsorption of U(VI) on corrosion products of iron [7,10], and iii) probable formation of U(VI) (hydr)oxide precipitate [10].

Zero-valent iron nanoparticles (ZVI-nps) are believed to have improved performances because of increased specific surface area and more reactive sites on the surface. Additionally, nanoscale metal particles could be applied through direct injection of particle suspensions to contaminated sediments and aquifers instead of constructing metal walls [11]. Bare ZVI-nps tend to agglomerate into larger aggregates due to magnetic properties. Incorporation of nanoparticles into carbon or polymer matrix can prevent the aggregation and their susceptible oxidation, therefore activated carbon-, chitosan-, and bentonite-based hybrid materials have been developed [12,13]. Graphene is a new generation of carbon material and can be regarded as a single layer of graphite. Graphene oxide (GO) can be easily prepared by several classical methods from cheap natural graphite, introducing oxygen-containing functional groups such as carboxyl and hydroxyl groups into carbon sheets. GO has been demonstrated to be a promising adsorbent to remove

heavy metals such as uranium [14,15] from aqueous solution. In this regard, it is expected that the removal performance of ZVI-nps for uranium could be improved by attaching GO sheets. Jabben et al. reported the successful synthesis of nanoscale iron-decorated graphene sheets and their applications in Cr(VI) removal. Higher maximum sorption capacity and higher reduction ratio of Cr(VI) adsorbed were achieved with the composites compared to that in bare ZVI-nps [12]. Moreover, Fe/graphene composites were also used to decolorize methyl blue solution. Higher removal capacities of the composites are due to the increased sorption sites, which originate from the inhibition of the particle aggregation and the reduction of the Fe particle size [16].

In this work, ZVI-np and its graphene composite were prepared successfully and applied in the removal of uranium from aqueous solution in anoxic atmosphere. The influences of solution pH, the presence of NaHCO3, synthetic groundwater constituents, and humic acid on removal efficiencies of uranium were investigated systematically in order to evaluate the efficacy of Fe0 to remediate uranium-contaminated groundwater. Sorption kinetics and capacities were studied as well to clarify reaction mechanisms with the help of various analytical techniques, e.g. SEM, XRD, and XAS, analyzing the reacted adsorbents after sorption processes.

2. Experimental

2.1. Reagents

All common chemicals used in this study were purchased from Aladin (Shanghai, China) and are of analytical grade. A 10 mM uranium stock solution was prepared by dissolving appropriate amount of UO2(NO3)2⋅6H2O (Sigma-Aldrich Co.) in Milli-Q water (18.2 MΩ cm, Millipore Co.). Synthesized groundwater consists of 0.29 mM Ca(NO3)2, 0.31 mM CaBr2, 0.53 mM MgSO4, 0.45 mM Na2SO4, 0.011 mM Na2CO3, 0.60 mM NaHCO3, and 0.43 mM KHCO3 [17].

2.2. Synthesis of ZVI-nps, reduced GO(RGO), and Fe/RGO composites

Fe0 was obtained according to a NaBH4 reduction method [16] under the protection of N2. Single-layered GO sheets, prepared from graphite according to the previously described procedure [14], were used as supporting matrix. In the synthesis of Fe/RGO composites (containing ~50% Fe

by mass), a concentrated FeCl3 solution (724 mg FeCl3⋅6H2O) was dropwise added into 300 mL of 0.5 mg/mL GO solution with ultrasonication. The mixed solution was magnetically stirred overnight, then NaBH4 powder (406 mg) was incrementally added into the mixture to simultaneously reduce Fe3+ and GO to Fe0 and RGO, respectively [12,16], resulting in a black homogenous dispersion. After 30 min shaking, Fe/RGO composites were recovered by vacuum filtration, and the solid was totally washed by deoxygenated ethanol prior to the formation of a liquid meniscus so as to prevent iron rust [18]. Finally, the black solid was vacuum-dried at 70 oC for 4 h, broken up with a spatula, and stored in a common desiccator. The accurate Fe content was determined to be 47.1% by dissolving the composites in 5% HNO3 and measuring Fe2+ quantity. Additional composite materials with ~20% and ~80% Fe loading were also attempted, and the morphology and removal performances for uranium are shown in Figs. S1 and 2 (Supporting Information).

ZVI-nps and RGO were prepared according to the aforementioned method in the absence of GO and FeCl3, respectively. It was noted that newly formed ZVI-nps were prone to aggregating in comparison with the Fe/RGO composites.

2.3. Batch experiments

Anoxic experiments were carried out in an Ar-filled glovebox. A small amount of ingressive O2 from outside was monitored to be around 30 ppm. All aqueous solutions used in sorption experiments were prepared in advance and placed in the anoxic glovebox for at least 24 h to examine the solution stabilities. Next, aqueous solution (50 mL) was transferred into a glass bottle, in which the adsorbent had been weighted and added. Supernatant pH was adjusted by negligible amounts of NaOH and HNO3 solution. The glass bottle was then sealed with a screw cap and the system was vigorously agitated by a magnetic stirrer to maintain the adsorbent suspension. Upon completion, portions of the supernatant were sacrificially filtered by 0.22 µm syringe filters, and the filtrate was acidified by 4% ultrahigh purity HNO3. Finally, concentrations of uranium, iron, and other metals in the solutions were analyzed by inductively coupled plasma optical emission spectrometer (ICP-OES, Horiba JY2000-2, Japan). Sorption capacity (Q) was defined as Q=(C0-Ce)V/m, where C0 and Ce denote the initial and equilibrium concentrations of

U(VI), respectively, in aqueous phase; V and m are the volume of the solution and the dry weight of adsorbents used in the sorption experiments. Experiments were performed in duplicates and the error bars associated with data shown in figures represent the standard deviations of the two runs. Oxic experiments were performed in open laboratory using the screwed glass bottles in consideration that it is more useful for ex situ treatment of contaminated water and/or industrial effluent. Finally, reacted adsorbents were collected and dried in the vacuum oven for characterization analyses.

2.4. Characterizations of as-prepared materials and reacted adsorbents

The microcosmic morphology of adsorbents before and after sorption was observed by scanning electron microscopy (SEM, S-4800, Hitachi) at an accelerating voltage of 10 kV. Nitrogen BET surface analysis was performed using a Micromeritics ASAP 2020. X-ray diffraction (XRD) was carried out on a Bruker D8 Advance X-ray diffraction instrument (Cu Kα, λ=1.5406 Å) with a step size of 0.02o, and the diffraction angel (2θ) from 5o to 90o was scanned. X-ray absorption near edge spectroscopy (XANES) of U LIII-edge was collected at the beamline 1W1B of Beijing Synchrotron Radiation Facility. Detailed experimental parameters and data treatments have been described previously [14].

3. Results and discussion

3.1. Characterizations of as-prepared adsorbent materials

SEM and XRD results of ZVI-nps, RGO, and Fe/RGO composites (47.1% Fe0, below omitted) are shown in Fig. 1. ZVI-nps present a spherical shape, aggregating into chain-like structures mainly due to magnetic property, and RGO presents a layered morphology. In the composites, more discrete iron nanoparticles disperse on RGO sheets, indicating an effective inhibition of particle aggregation. The particle size ranges from 30 to 100 nm. As for the XRD pattern of ZVI-nps, a dominant diffraction peak located at 44.7-44.8o is the characteristic of poorly crystalline/amorphous metallic iron [19,20]. In the case of RGO, a characteristic diffraction peak of GO at 10.9o remains, besides a broad diffraction peak (002) of graphite at ~23o and a sharp

peak at 44o, suggesting incomplete reduction of GO [21]. XRD results of Fe/RGO composites reveal the presence of Fe0 and that the orderly layered stacking of RGO sheets can be prevented effectively. BET specific surface areas of ZVI-nps, RGO, and Fe/RGO were determined to be 9.7, 167.8, and 218.9 m2/g, respectively, highlighting the advantage of the composites.

It’s known that in anoxic atmosphere, Fe0 corrodes according to the equation of Fe0 + 2H2O→ Fe2+ + H2(g) + 2OH-. The pH-dependent iron dissolution was studied in the current sorption system and found that at initial pH (pH0) larger than 3.6, the amount of Fe2+ released was less than 0.5% (Fig. S3, Supporting Information).

SEM

a

b

200 nm

c

200 nm

200 nm

XRD

a c b 0 Fe (PDF 06-0696) 15

30

45

60

75

90

2θ (degree)

Fig.1. SEM photographs and XRD patterns of freshly prepared ZVI-nps (a), RGO (b), and Fe/RGO composites (c).

3.2. Removal kinetics of uranium by ZVI-nps and Fe/RGO composites

One practical requirement for reactive materials filled into PRB is that the kinetics of immobilization reactions must be fast. A slow reaction rate would require a prohibitively thick barrier to extend the contaminant’s residence time [6]. Removal rates of uranium by ZVI-nps and Fe/RGO composites with the same iron dosage were therefore examined systematically, and residual uranium (%) in solution as a function of contact time was plotted in Fig. 2a. It can be seen that at C0(U)=24 ppm, the concentration of uranium remaining in solution decreased rapidly, and was below the detection limit of the instrument within 20 min. At C0(U)=333 ppm, it took almost

40 min for the complete removal of uranium by ZVI-nps. Such rapid kinetics suggests that even in column or in situ barriers, the primary reduction process occurs upon first contact of the adsorbents with U(VI) species. To distinguish the kinetics of bare Fe0 and the composites, sorption systems were shaken gently by hand. It was found that for ZVI-nps and Fe/RGO, it required 60 and 40 min, respectively, to clean up the 24 ppm uranium solution, showing that the removal rate by Fe/RGO is faster than that by ZVI-nps. This could be attributed to the following aspects: i) the increased surface area and reaction sites of Fe0 after combination with graphene; ii) residual oxygen-containing functional groups on the surface of RGO, which can capture uranium rapidly; and iii) the improved suspension stability of the composites.

Recorded pH and Fe2+ concentrations in solutions during sorption processes were presented in Fig. 2b and c, respectively. Upon the treatment of 24 ppm uranium solution with either ZVI-nps or Fe/RGO, solution pH increased to 9.3, and faster reaction rate can lead to faster pH increases. In contrast, the treatment of 333 ppm uranium solution resulted in a final pH value (pHf) of 6.0. Iron dissolution was negligible all along at C0(U)=24 ppm, but Fe2+ concentration in the solution containing ZVI-nps and 333 ppm uranium approached 44.4 ppm and maintained stable over the remainder of the sorption period.

In oxic atmosphere, uranium removal rates by ZVI-nps were much slower than those of corresponding anoxic groups, and there was a re-dissolution of already adsorbed uranium. Dissolved oxygen has been reported to be able to re-oxidize bioreduced U(IV) to the soluble U(VI) species within several hours to days [22 and references therein], and was likely responsible for the re-dissolution here. Iron corrosion in the presence of O2 was written as 2Fe0 + 2H2O + O2 → 2Fe2+ + 4OH- and Fe2+ + 5/2H2O + 1/4O2 → Fe(OH)3(s) + 2H+. At both C0(U)=24 and 333 ppm, iron dissolution arose, followed by gradual decreases in Fe2+ concentration. Crane et al. [20] reported a similar tendency and assigned the fall after a rise to the formation of Fe(OH)3 with the slow ingress of atmospheric oxygen.

Additionally, the removal rate of uranium by ZVI-nps and Fe/RGO composites from 24 ppm uranium-containing synthetic groundwater (pH0 7.9) was also examined and found to be very fast as well (Fig. S4, Supporting Information).

Residual uranium (%)

a 100 75

24-U-Fe 24-U-Fe/RGO 333-U-Fe 333-U-Fe-oxic

24-U-Fe-hand 24-U-Fe/RGO-hand 24-U-Fe-oxic

50 25 0

pH

10

b

8

2+

Fe release (ppm)

6 4 60

c

40 20 0 0

100

200

300

400

1000 1500

Contact time (min)

Fig.2. (a) removal kinetics of uranium by ZVI-nps and Fe/RGO in pure Milli-Q water under anoxic and oxic conditions; corresponding pH (b) and Fe2+ concentrations in aqueous phase (c) as a function of contact time. mZVI-nps=4.0 mg, mFe/RGO=8.5 mg, V=50 mL, C0(U)=24 or 333 ppm, and pH0=5.0.

3.3. Influences of several environmental factors on uranium removal by ZVI-nps

It’s well known that CO32- occurs ubiquitously in natural or contaminated groundwater [23], and it can form stable complexes with uranium especially under neutral-basic conditions. In real groundwater, ternary calcium-uranyl-carbonato complexes, i.e. Ca2UO2(CO3)3 and CaUO2(CO3)32-, are prevalent, which are less efficient electron acceptors than uranium hydroxyl species [23,24]. Humic acid, a representative of natural organic matters in groundwater, possesses organic functional groups such as carbonyl, carboxylate, phenol, and hydroxyl groups, which are capable of capturing uranium. Recently, there have been several reports about reduction ability of humic acid [25,26]. Therefore, the effectiveness of ZVI-nps for the removal of uranium was investigated in 1.0 mM NaHCO3, synthetic groundwater, and 10 ppm humic acid matrices, and typical pH

values (5∼9) of groundwater were selected to study pH influences.

Sorption capacities of ZVI-nps for uranium were plotted in Fig. 3 as a function of pH0. In pure Milli-Q water, uranium sorption on ZVI-nps was recorded to be ∼300 mg U/g, corresponding to 100% removal of uranium, independent of solution pH0, except of the extremely acidic condition (pH0=2.1). Under that condition, 92.9% iron dissolved and little uranium was removed, in contrast at pH0=4.0-9.0, Fe2+ releases into solution were negligible. It was also found that the presence of NaHCO3, synthetic groundwater constituents, and humic acid could not affect the efficient removal of uranium from water. Yan et al. [24] once reported the inhibitive effect of higher pH, the presence of HCO3-, and the utilization of the synthetic groundwater on the removal rate of uranium by ZVI-nps. It is easily understood because the formation of either carbonate or calcium-uranyl-cabonato complexes can stabilize U(VI) in aqueous phase, therefore suppressing U(VI) sorption and reduction. However, complete removal of uranium could be expected from a thermodynamics point of view since continuous liberation of UO22+ from U(VI) complexes occurred during the treatments, e.g. Fe0(s) + UO2(CO3)34- + 6H+ → UO2(s) + Fe2+ + 3CO2(g) + 3H2O [20]. Accordingly, it is concluded that ZVI-nps are promising candidates for the effective remediation of uranium-contaminated groundwater. Additionally, the uranium-containing humic acid solution was stable over the timescale considered, suggesting that humic acid was incapable of reducing UO22+ under current experimental conditions. Czerwinski et al. also reported that no reduction of U(VI) occurred up to 40 days in the presence of aquatic humic acid [27].

It is interesting to note that unlike the MilliQ-water and humic acid systems, the Fe0-treated NaHCO3 solution and synthetic groundwater presented blue-green, the characteristic color of green rust. Green rust is a layered Fe(II)/Fe(III) hydroxide with hydrated anions such as CO32- and SO42- located in interlayers. A green rust mineral morphology, pseudo-hexagonal form, was then observed in SEM of the reacted ZVI-nps (Fig. S5, Supporting Information). Actually, green rust has been reported to be the corrosion product of iron in an in situ remediation of radionuclide-contaminated groundwater [5].

100

300 250

150 100

MilliQ-water NaHCO3

75

Mimic groundwater Humic acid 2+ Fe release in water (right axis)

50

25

50

2+

200

Fe release (%)

Sorption capacity (mg U/g)

350

0

0 2

3

4

5

6

7

8

9

10

pH0

Fig.3. Left axis: influence of solution pH on the uranium removal by ZVI-nps in MilliQ-water, 1.0 mM NaHCO3, mimic groundwater, and 10 ppm humic acid solutions, respectively; right axis: influence of solution pH on the Fe2+ release in the MilliQ-water system. mZVI-nps=4.0 mg, V=50 mL, C0(U)=24 ppm, and contact time=24 h.

3.4. Sorption capacities of ZVI-nps, RGO, and Fe/RGO composites

Maximum sorption capacities of ZVI-nps, RGO, and Fe/RGO composites for uranium were determined in Milli-Q water containing 0∼714 ppm uranium (pH0=5.0). The results were plotted in Fig. 4a as a function of initial uranium concentration. It was found that: i) with the increase in C0(U), removal efficiencies of uranium by ZVI-nps kept at 100% until C0(U)=643 ppm, and decreased to 87% at C0(U)=714 ppm, corresponding to the maximum sorption capacity of 8173 mg U/g; ii) the sorption capacity of RGO increased gradually with the increase in C0(U), attaining a saturation of 341 mg U/g, which is far smaller than that of ZVI-nps, but comparable to the previously reported value of this material [14]; iii) the sorption tendency of uranium on Fe/RGO was similar to that on ZVI-nps, achieving an equilibrium of 4174 mg U/g. This value is reasonable in light of the Fe0 percentage in composites. In the case of 100 mg ZVI-nps, complete removal of uranium was achieved even at C0(U)=714 ppm. This agrees with the description that metallic iron has infinite sorption capacity for uranium so long as Fe0 is sufficient in the system, maintaining a favorable reducing condition. Gu et al. also observed this ‘infinite’ removal efficiency: interactions between UO22+ (0-1.8×104 ppm) and Fe0 (2.0×104 mg/L) resulted in 100% removal of uranium from solution (pH0=5.0). The solution pH increased from 5 to 10 in less than 30 min [28].

After the treatment with ZVI-nps, pHf increased to 9.2 and 5.4 at C0(U)=24 and 714 ppm, respectively, as labeled in the figure. Figure 4b gives Fe2+ releases from the iron-bearing materials as a function of C0(U). In the case of 4.0 mg ZVI-nps, Fe2+ releases were not observed at C0(U)=0-71 ppm; with the continuous increase in C0(U), Fe0 started to dissolve and Fe2+ concentration increased, the recorded maximum dissolution was 58.8% of iron. The case of Fe/RGO is similar to that of ZVI-nps, and a more apparent dissolution stop (61.0% of iron) was concurrent with the saturation sorption of Fe/RGO for uranium. As for 100 mg ZVI-nps, pHf values were recorded to be larger than 9.0 and Fe2+ releases were all negligible.

Under oxic conditions, uranium removal by ZVI-nps was largely decreased with the maximum

Sorption capacity (mg U/g)

sorption capacity of 1354 mg U/g. Fe2+ releases were inhibited as well.

8000

a

pHf=6.6, 5.4

Fe Fe-100mg Fe/RGO RGO Fe-Oxic

6000

87% removal

70% removal

4000

2000

pHf=9.2

pHf=5.3, 5.4, 5.4

0

60

Fe Fe-100mg Fe/RGO Fe-Oxic

40 30 20

2+

Fe release (ppm)

b 50

10 0 0

100

200 300

400

500

600 700

800

C0(U) / ppm

Fig.4. (a) sorption capacities of ZVI-nps, Fe/RGO, and RGO for uranium as a function of C0(U); (b) corresponding Fe2+ concentrations in solutions. m=4.0 or 100 mg (indicated in Figure), V=50 mL, pH0=5.0, contact time=24 h.

3.5. Characterizations of adsorbents after uranium sorption SEM and XRD. SEM photographs and XRD patterns of reacted adsorbents are shown in Figs. 5 and 6, respectively. A blank experiment without uranium addition yielded blank ZVI-nps, which appearance (Fig. 5a) is not significantly different from the freshly prepared one and XRD analyses revealed that metallic iron is the predominant phase still. After the treatment of 24 ppm uranium solution, bare ZVI-nps (Fig. 5b) and the nanoparticles decorated on RGO sheets (Fig. 5c) grew larger and surfaces became rougher, which could be attributed to a combination precipitation of corrosion products of iron and U(IV) (hydr)oxide on iron surface [10]. Magnetite (Fe3O4) and/or maghemite (γ-Fe2O3) was identified as the predominant corrosion product of iron by XRD. XRD technique is unable to distinguish Fe3O4 and γ-Fe2O3 because structural Fe(II) ions located at octahedral sites of the oxide can be reversely oxidized and reduced in the same crystalline structure. During uranium removal, Fe(II)/Fe(III) mixture oxide was generally reported as the oxidative product of iron [10,29,30]. The presence of Fe(III) ions confirmed that structural Fe(II) ions can reduce U(VI) further [29,31,32]. The reflection peak of residual Fe0 is obvious in XRD of treated ZVI-nps, while it is very weak for Fe/RGO composites, implying that Fe0 incorporated into the composites is more reactive. Unfortunately, there is no uranium-related diffraction peak, probably due to small amounts of deposited uranium phase. At C0(U)=238 ppm, the presence of U3O7 (UO2.3) in reacted ZVI-nps was verified by the XRD analysis, illustrating a partial reduction of U(VI) removed by ZVI-nps. Iron oxidation to Fe3O4/γ-Fe2O3 was accompanied. The partially reductive precipitation of U(VI) on the surface of iron has been addressed: Scott et al. [31 and references therein] identified uranium species associated with mild steel surface as U4O7 based on XPS results; Riba et al. [10] observed a broad peak at 2θ=28o in XRD, which was assigned to UO2 although U(VI) was found invariably from XPS. After the treatment of 333 ppm uranium solution, not only iron particles but also RGO sheets (Figs. 5d and e) were covered with unknown substance. The SEM coupled with energy dispersive X-ray (EDX) analyses (Figs. 5g and h) revealed the presence of great amounts of uranium. The treated ZVI-nps and Fe/RGO composites are somewhat XRD amorphous, but it seems like that

Fe/RGO is rich in U3O7. This might be reasonable in consideration of the increased reactivity of iron nanoparticles decorated on RGO sheets. After the treatment of 714 ppm uranium solution, the particle morphology of ZVI-nps disappeared (Fig. 5f) and the XRD reflection peaks are more flattened. The SEM microphotograph of reacted ZVI-nps obtained in air (Fig. 5i) displays irregularly spherical particles and thin plates, XRD analyses demonstrate the concurrent occurrence of U3O7 and Fe3O4/γ-Fe2O3. Chemical reduction of U(VI) by Fe0 has been reported to be possible in both anoxic and oxic atmosphere, but promoted at a low oxygen level [20]. If appreciable oxygen is available, uranium removal would be dominated by the sorption of U(VI) on corrosion products of iron [29]. Further analysis suggested the coexistence of FeOOH, its thin plate form has already been observed in SEM. FeOOH was considered to be generated from Fe(OH)3 dehydration and the higher proportion of Fe(III) ions in an oxic system was attributed to greater amounts of dissolved O2.

a

b

c

100 nm

100 nm

100 nm

d

e

f

100 nm

100 nm

100 nm

g

h

i

100 nm keV

keV

Fig.5. SEM photographs of ZVI-nps prepared in a blank experiment (a), ZVI-nps (b) and Fe/RGO (c) treated at C0(U)=24 ppm, ZVI-nps (d) and Fe/RGO (e) treated at C0(U)=333 ppm, ZVI-nps (f) treated at C0(U)=714 ppm, and reacted ZVI-nps obtained under oxic conditions (i); EDX spectra (g and h) for d and e, respectively. mZVI-nps=4.0 mg, mFe/RGO=8.5 mg, V=50 mL, and pH0=5.0.

Blank 24-U-Fe 24-U-Fe/RGO

238-U-Fe 333-U-Fe 333-U-Fe/RGO 714-U-Fe Fe-Oxic U3O7 (PDF 42-1215) Maghemite (PDF 39-1346) FeOOH (PDF 70-0714) 20

30

40

50

60

70

80

90

2θ (degree) Fig.6. XRD patterns of the reacted adsorbents. XANES analyses. XANES spectra provide an important fingerprinting method for judging oxidation states of investigated metal ions by comparing with known and standard samples [33]. The energy of absorption edge increases with increasing average oxidation state of an absorbing atom, and in uranium spectra, a resonance shoulder at ~17180 eV is assigned to multi-scattering of axial oxygen atoms in the uranyl moiety [34,35]. In this work, XANES were performed on the above reacted adsorbents and the results were plotted in Fig. 7. It can be seen that the near edge absorption curves locate in the right side between those of U(IV)O2 and U(VI)O2(OH)2, illustrating a mixture of U(IV) and U(VI) in reacted adsorbents. For 24 ppm uranium-treated ZVI-nps, the energy of absorption edge and the absence of the resonance peak suggests that U(IV) is the predominant oxidation state. Absorption edges of 24, 238, 333, and 714 ppm uranium-treated ZVI-nps are arranged from low to high energy, and resonance features of O=U=O are increasingly apparent in the same order, which implies that the U(VI) proportion of uranium removed by iron increases with the increase in C0(U). The absorption curve of 714 ppm uranium-treated ZVI-nps closely follows that of U(VI) precipitate. Additionally, the absorption edge energy of treated Fe/RGO at C0(U)=333 ppm is a little lower than that of the corresponding

ZVI-nps, and the Oax peak is more inhibited, confirming the higher U(VI) reduction ratio in Fe/RGO composites.

µ(E)

1.5

1.0 UO2 24-U-Fe 333-U-Fe 714-U-Fe

0.5

UO2(OH)2 238-U-Fe 333-U-Fe/RGO Fe-Oxic

0.0 17140

17160

17180

17200

17220

E (eV)

Fig.7. XANES of the reacted adsorbents. Desorption experiments. Based on the references [35,36], (NH4)2CO3 (0.1 M) was adopted to desorb U(VI) ions associated with iron oxides and to dissolve U(VI) minerals such as metaschoepite, therefore quantifying U(VI)/U(IV) proportion. From Table 1, it can be seen that at C0(U)=24 ppm, very small amounts of removed uranium on ZVI-nps and Fe/RGO could be leached by the CO32- solution, confirming the U(IV) valence predominant in the adsorbents. At C0(U)=333 ppm, U(VI) ratio increased largely to 34.0% in reacted ZVI-nps; the supporting material, RGO and higher pH0 enhanced U(VI) reduction ratio to some extents. At C0(U)=714 ppm, nearly half of uranium removed by ZVI-nps was ready to wash out. The maximum sorption capacity of synthetic magnetite nanoparticles was only 27 mg U/g [37] and Gu et al. [28] reported that the percentage of U(VI) associated with iron corrosion products was in the range of 2.77-3.91%. It is therefore not reasonable to assign such high proportion of U(VI) to mere U(VI) adsorption on iron oxides, hydrolysis precipitation of U(VI) on iron surface was thus brought forward.

3.6.

3.7.

3.8. Reaction mechanisms

Important references studying removal mechanisms of Fe0 for uranium are summarized in Table S1 (Supporting Information), and chemical reactions written below are supposed to occur during the treatment in anoxic atmosphere,

2>Fe-OH + UO22+ → [(>Fe-O)2UO2]ads + 2H+

(1)

UO22+ + Fe0 + 2H2O → U(IV)(s)+ Fe2+ + 4OH-

(2)

UO22+ + 2Fe(II)(structural) + 2H2O → U(IV)(s) + 2Fe(III)(structural) + 4OH-

(3)

Fe0 + 3Fe2O3(s) + H2O → 2Fe3O4(s) + Fe2+ + 2OH-

(4)

UO22+ + 2OH- → UO2(OH)2(s)

(5)

Reductive precipitation of U(VI) on iron surface. According to reaction 1, the oxide layers of Fe0 particles contain hydroxyl groups, which are capable of adsorbing UO22+ via proton exchanges. Adsorbed U(VI) ions will be reduced to the sparely soluble U(IV) precipitate by Fe0 cores (reaction 2). The predominant corrosion product of iron has been identified as Fe3O4/γ-Fe2O3, revealing that structural Fe(II) ions can reduce U(VI) to U(IV) further (reaction 3). Reaction 4 presents the process that a Fe0 core transfers electrons to the surface Fe(III) oxide, leading to the re-generation of Fe(II) ions. As a result of these reactions, solution pH largely increases, leading to a total precipitation of released Fe2+ as Fe(OH)2 (Ksp(Fe(OH)2)=8.0×10-16 [38]) [5]. Hydrolysis precipitation of U(VI). Hydrolysis precipitation of U(VI) on iron surface is described subsequently, characterized by an inhibition of the pH increase and releases of significant amounts of Fe2+ ions. The continuous precipitation of U(IV) and iron (hydr)oxides on surface inevitably reduces iron reactivity and limits UO22+ access to Fe0/Fe3O4(and/or γ-Fe2O3) surfaces. At the high pH values, UO22+ ions would precipitate out from solution according to reaction 5, resulting in acidification of the solution, which could initiate the dissolution of Fe0 cores along surface cracks. The generated OH- ions precipitate UO22+ as UO2(OH)2 continuously (Ksp(UO2(OH)2) =3.5×10-23 [39]), accompanied with Fe2+ releases. Since UO22+ reduction by aqueous Fe2+ ions is a rather slow process [32], Fe0 plays a precipitant role at this time virtually. When Fe0 cores are depleted, the hydrolysis reaction of UO22+ will stop, and the so-called maximum sorption of Fe0 for uranium will be established, just like the data shown in Fig. 4. Riba et al. confirmed the occurrence of metaschoepite (UO3⋅2H2O) in treated ZVI-nps with 850 ppm

uranium solution (pH=5.0-6.0) by XRD, and observed a simultaneous increase in Fe2+ concentration [10]. Zhang et al. [19] reported a similar phenomenon when using ZVI-nps to sequester Pb2+. Lead removal was well correlated with a rise of Fe2+ concentration in solution. Over 90% of deposited lead was found to exist as Pb(II) ions rather than Pb0 from XPS analyses, and Fe0 cores were depleted, leaving behind empty iron oxide shells observed by TEM. The shell constituent of Fe(OH)3 was finally assigned to dissolve Fe0, leading to the formation of Pb(OH)2 precipitate.

Under oxic conditions, more Fe3+ is present and it’s known that Fe3+ ions more tend to precipitate out (Ksp(Fe(OH)3)=4.0×10-38 [40]), therefore UO22+ hydrolysis could be prevented effectively. This viewpoint can be supported by the characterization results of reacted ZVI-nps obtained in air (Sec. 3.5).

Influence of supporting materials and higher pH0 on the pathway of uranium immobilization. At C0(U)=333 ppm, Fe2+ releases accounted for 33.9% and 14.2% of iron in ZVI-nps and Fe/RGO composites, respectively with pHf values of 6.3 and 6.5. In view of the above discussed correlation between the speciation of adsorbed uranium and pH changes, Fe2+ releases, U(IV) proportion among uranium adsorbed on Fe/RGO is expectedly higher than that on bare ZVI-nps, which has been affirmed by XRD, XANES, and the desorption experiments. Recently, Sheng et al. [13] observed that the reduction of U(VI) to U(IV) was enhanced by using Na-bentonite as the support of ZVI-nps. Higher reduction ratio (60%) of Cr(VI) was achieved in Fe/RGO compared to the 45% value in bare ZVI-nps. The authors held that iron decorated graphene composites have smaller size of iron nanoparticles and higher surface area, which could increase catalytic and adsorption sites of Cr(VI) [12]. These results prove that introducing a supporting material can not only enhance sorption capacities, improve removal rate but also increase reductive transformation of redox active heavy metal ions into less toxic precipitates. After the treatment of 333 ppm uranium solution (pH0=6.5), the percentage of iron dissolution decreased to 8.7% of ZVI-nps and pHf was 6.7. Higher reduction of U(VI) to U(IV) was demonstrated as well in the desorption experiments. pH0 higher than 6.8 was reported to be a benefit for the reductive precipitation of U(VI) [7,24]. Riba et al. [10] addressed that after 12 min

reaction, 30% more UO2 was formed on iron surface at pH0 6.5 and 7.0 in comparison with pH0 5.0, 5.5, and 6.0 based on XPS analyses. The authors attributed this to the proportion of ≡(FeIIIOFeIIOH)0, a more efficient electron donor over the other surface species of ≡(FeIIIOFeII)+, increasing with alkalinity [41].

4. Conclusions

ZVI-nps are very effective reactive materials for PRB technique to remediate uranium-contaminated subsurface area. Rapid reductive precipitation of U(VI) to the sparingly soluble U(IV) species is responsible for uranium immobilization. However, corrosion products of iron and uranium precipitate formed during water treatment decrease the reactivity of iron surface gradually. The role of Fe0 changes from a reductant to a precipitant, leading to a formation of U(VI) (hydr)oxide and a release of Fe2+ ions. This is not favorable for long-term immobilization of uranium in geochemical environment. Composite materials such as Fe/RGO could decrease the particle size of iron nanoparticles and prevent their aggregation, thus largely increasing specific surface area for reaction with uranium. Iron is therefore more fully utilized and U(VI) reduction ratio is enhanced. Additionally, neutral-basic groundwater (pH0≥6.5) was considered to benefit the reduction pathway of U(VI) over hydrolysis precipitation. In extremely acidic solution such as uranium-containing acid mine water, iron nanoparticles decorated on RGO sheets are much easier to be dissolved into solution, decreasing the effectiveness of uranium removal (Fig. S6, Supporting Information). Therefore, ZVI-nps are relatively more appropriate to such conditions.

Acknowledgements

This work was supported by the Natural Science Foundation of China (Grants 11205169, 91326202, 11275219, 21261140335, 11105162 and 21101157) and the "Strategic Priority Research program" of the Chinese Academy of Sciences (Grants.XDA030104). The financial support from the State Key Laboratory of NBC Protection for Civilian (No.SKLNB2014-12) is also acknowledged.

Appendix A. Supporting Information

Supporting Information associated with this article can be found in the online version.

References

[1] W.Q. Shi, L.Y. Yuan, Z.J. Li, J.H. Lan, Y.L. Zhao, Z.F. Chai, Nanomaterials and nanotechnologies in nuclear energy chemistry, Radiochim. Acta 100 (2012) 727-736.

[2] http://public.ornl.gov/nabirfrc/dataarea123.cfm; accessed August 6, 2002.

[3] S. Klimkova, M. Cernik, L. Lacinova, J. Filip, D. Jancik, R. Zboril, Zero-valent iron nanoparticles in treatment of acid mine water from in situ uranium leaching, Chemosphere 82 (2011) 1178-1184.

[4] http://www.epa.gov/rpdweb00/radionuclides/uranium.html#epadoing; accessed August 6, 2014.

[5] Y. Roh, S.Y. Lee, M.P. Elless, Characterization of corrosion products in the permeable reactive barriers, Environ. Geo. 40 (2000) 184-194.

[6] K.J. Cantrell, D.I. Kaplan, T.W. Wietsma, Zero-valent iron for the in situ remediation of selected metals in groundwater, J. Hazard. Mater. 42 (1995) 201-212.

[7] M. Dickinson, T.B. Scott, The application of zero-valent iron nanoparticles for the remediation of a uranium-contaminated waste effluent, J. Hazard. Mater. 178 (2010) 171-179.

[8] J.G. Darab, A.B. Amonette, D.S.D. Burke, R.D. Orr, S.M. Ponder, B. Schrick, T.E. Mallouk, W.W. Lukens, D.L. Caulder, D.K. Shuh, Removal of pertechnetate from simulated nuclear waste streams using supported zero-valent iron, Chem. Mater. 19 (2007) 5703-5713.

[9] T.B. Scott, I.C. Popescu, R.A. Crane, C. Noubactep, Nano-scale metallic iron for the treatment of solutions containing multiple inorganic contaminants, J. Hazard. Mater. 186 (2011) 280-287.

[10] O. Riba, T.B. Scott, K.V. Ragnarsdottir, G.C. Allen, Reaction mechanism of uranyl in the

presence of zero-valent iron nanoparticles, Geochim. Cosmochim. Acta 72 (2008) 4047-4057.

[11] C.B. Wang, W.X. Zhang, Synthesizing nanoscale iron particles for rapid and complete dechlorination of TCE and PCBs, Environ. Sci. Tech. 31 (1997) 2154-2156.

[12] H. Jabeen, V. Chandra, S. Jung, J.W. Lee, K.S. Kim, S.B. Kim, Enhanced Cr(VI) removal using iron nanoparticle decorated graphene, Nanoscale, 3 (2011) 3583-3585.

[13] G. Sheng, X. Shao, Y. Li, J. Li, H. Dong, W. Cheng, X. Gao, Y. Huang, Enhanced removal of U(VI) by nanoscale zerovalent iron supported on Na-bentonite and an investigation of mechanism, J. Phys. Chem. A 118 (2014) 2952-2958.

[14] Z.J. Li, F. Chen, L.Y. Yuan, Y.L. Liu, Y.L. Zhao, Z.F. Chai, W.Q. Shi, Uranium adsorption on graphene oxide nanosheets from aqueous solutions, Chem. Eng. J. 210 (2012) 539-546.

[15] Z.Q. Bai, Z.J. Li, C.Z. Wang, L.Y. Yuan, Z.R. Liu, J. Zhang, L.R. Zheng, Y.L. Zhao, Z.F. Chai, W.Q. Shi, Interactions between Th(IV) and graphene oxide: Experimental and density functional theoretical investigations, RSC Adv. 4 (2014) 3340-3347.

[16] J. Guo, R. Wang, W.W. Tjiu, J. Pan, T. Liu, Synthesis of Fe nanoparticles@graphene composites for environmental applications, J. Hazard. Mater. 225-226 (2012) 63-73.

[17] C. Liu, J.M. Zachara, N. P. Qafoku, Z. Wang, Scale-dependent desorption of uranium from contaminated subsurface sediments, Water Resource Res. 44 (2008) W08413.

[18] S.M. Ponder, J.G. Darab, T.E. Mallouk, Remediation of Cr(VI) and Pb(II) aqueous solutions using supported nanoscale zero-valent iron, Environ. Sci. Technol. 34 (2000) 2564-2569.

[19] Y. Zhang, Y. Su, X. Zhou, C. Dai, A.A. Keller, A new insight on the core-shell structure of zerovalent iron nanoparticles and its application for Pb(II) sequestration, J. Hazard. Mater. 263 (2013) 685-693.

[20] R.A. Crane, M. Dickinson, I.C. Popescu, T.B. Scott, Magnetite and zero-valent iron

nanoparticles for the remediation of uranium contaminated environmental water, Water Res. 45 (2011) 2931-2942.

[21] W.F. Chen, L.F. Yan, In situ self-assembly of mild chemical reduction graphene for three-dimensional architectures, Nanoscale 3 (2011) 3132-3137.

[22] K. Ulrich, E.S. Ilton, H. Veeramani, J.O. Sharp, R. Bernier-Latmani, E.J. Schofield, J.R. Bargar, D.E. Giammar, Comparative dissolution kinetics of biogenic and chemogenic uraninite under oxidizing conditions in the presence of carbonate, Geochim. Cosmochim. Acta, 73 (2009) 6065-6083.

[23] S.C. Brooks, J.K. Fredrickson, S.L. Carroll, D.W. Kennedy, J.M. Zachara, A.E. Plymale, S.D. Kelly, K.M. Kemner, S. Fendorf, Inhibition of Bacterial U(VI) reduction by calcium, Environ. Sci. Technol. 37 (2003) 1850-1858.

[24] S. Yan, B. Hua, Z. Bao, J. Yang, C. Liu, B. Deng, Uranium(VI) removal by nanoscale zerovalent iron in anoxic batch systems, Environ. Sci. Technol. 44 (2010) 7783-7789.

[25] P. Ivanov, T. Griffiths, N.D. Bryan, G. Bozhikov, S. Dmitriev, The effect of humic acid on uranyl sorption onto bentonite at trace uranium levels, J. Environ. Monit. 14 (2012) 2968-2975.

[26] W. Jiang, Q. Cai, W. Xu, M. Yang, Y. Cai, D.D. Dionysiou, K.E. O’Shea, Cr(VI) adsorption and reduction by humic acid coated on magnetite, Environ. Sci. Technol. 48 (2014) 8078-8085.

[27] K.R. Czerwinski, G. Buckau, F. Scherbaum and J.I. Kim, Complexation of the uranyl ion with aquatic humic acid, Radiochim. Acta 65 (1994) 111-119.

[28] B. Gu, L. Liang, M.J. Dickey, X. Yin, S. Dai, Reductive precipitation of uranium(VI) by zero-valent iron, Environ. Sci. Technol. 32 (1998) 3366-3373.

[29] J.N. Fiedor, W.D. Bostick, R.J. Jarabek, J. Farrell, Understanding the mechanism of uranium removal from groundwater by zero-valent iron using X-ray photoelectron spectroscopy, Environ. Sci. Technol. 32 (1998) 1466-1473.

[30] T.B. Scott, G.C. Allen, P.J. Heard, A.C. Lewis, D. F. Lee, The extraction of uranium from groudwaters on iron surfaces, Proc. R. Soc. A 461 (2005) 1247-1259.

[31] T.B. Scott, G.C. Allen, P.J. Heard, M.G. Randell, Reduction of U(VI) to U(IV) on the surface of magnetite, Geochim. Cosmochim. Acta 69 (2005) 5639-5646.

[32] L. Charlet, E. Liger and P. Gerasimo, Decontamination of TCE- and U-rich waters by granular iron: role of sorbed Fe(II), J. Environ. Eng.124 (1998) 25–30.

[33] W.Q. Shi, L.Y. Yuan, C.Z. Wang, L. Wang, L. Mei, C.L. Xiao, L. Zhang, Z.J. Li, Y.L. Zhao, Z.F. Chai, Exploring actinide materials through synchrotron radiation techniques, Adv. Mater. 26 (2014) 7807-7848.

[34] E.J. O’loughlin, S.D. Kelly, K.M. Kemner, XAFS investigation of the interactions of U(VI) with secondary mineralization products from the bioreduction of Fe(III) oxides, Environ. Sci. Technol. 44 (2010) 1656-1661.

[35] M.C. Duff, J.U. Coughlin, D.B. Hunter, Uranium co-precipitation with iron oxide minerals, Geochim. Cosmochim. Acta 66 (2002) 3533-3547.

[36] C.F.V. Mason, W.R.J.R. Turney, B.M. Thomson, N. Lu, P.A. Longmire, C.J. Chisholm-Brause, Carbonate leaching of uranium from contaminated soils, Environ. Sci. Technol. 31 (1997) 2707-2711.

[37] R. Leal, M. Yamaura, Equilibrium adsorption isotherm of U(VI) at pH 4 and pH 5 onto synthetic magnetite nanoparticles, Int. J. Nucl. Energ. Sci. Technol. 6 (2011) 1-7.

[38] D.L. Leussing, I.M. Kolthoff, The solubility product of ferrous hydroxide and the ionization of aquo-ferrous ion, J. Am. Chem. Soc. 75 (1953) 2476-2479.

[39] K. Fujiwara, H. Yamana, T. Fujii, K. Kawamoto, T. Sasaki, H. Moriyama, Solubility product of hexavalent uranium hydrous oxide, J. Nucl. Sci. Technol. 42 (2005) 289-294.

[40] X. Liu, F.J. Millero, The solubility of iron hydroxide in sodium chloride solutions, Geochim. Cosmochim. Acta 63 (1999) 3487-3497.

[41] E. Liger, L. Charlet, P. Van Cappellen, Surface catalysis of uranium (VI) reduction by iron (II), Geochim. Cosmochim. Acta 63 (1999) 2939–2955.

Table 1. (NH4)2CO3 leaching of the reacted adsorbents (V=10 mL, leaching time=30 min)

24 ppm

24 ppm

333 ppm

333 ppm

333 ppm

714 ppm

U-Fe

U-Fe/RGO

U-Fe

U-Fe/RGO

U-Fe-pH0

U-Fe

6.5 U(VI) leaching

0.7%

0.3%

34.0%

21.1%

30.4%

47.5%