elevated temperature: Experimental and XPS study

    Surface complexation modeling of U(VI) adsorption on granite at ambient/elevated temperature: Experimental and XPS study Qiang Jin, Lin Su, Gilles Montavon, Yufeng Sun, Zongyuan Chen, Zhijun Guo, Wangsuo Wu PII: DOI: Reference:

S0009-2541(16)30165-6 doi: 10.1016/j.chemgeo.2016.04.001 CHEMGE 17895

To appear in:

Chemical Geology

Received date: Revised date: Accepted date:

6 January 2016 31 March 2016 1 April 2016

Please cite this article as: Jin, Qiang, Su, Lin, Montavon, Gilles, Sun, Yufeng, Chen, Zongyuan, Guo, Zhijun, Wu, Wangsuo, Surface complexation modeling of U(VI) adsorption on granite at ambient/elevated temperature: Experimental and XPS study, Chemical Geology (2016), doi: 10.1016/j.chemgeo.2016.04.001

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.

ACCEPTED MANUSCRIPT

Surface Complexation Modeling of U(VI) Adsorption on Granite at

T

Ambient/Elevated Temperature: experimental and XPS study

a

SC R

IP

Qiang Jin a, Lin Su a, Gilles Montavon b, Yufeng Sun a, Zongyuan Chen a,c*, Zhijun Guo a,c*, Wangsuo Wu a,c

Radiochemistry Laboratory, School of Nuclear Science and Technology, Lanzhou University, 730000 Lanzhou,

China b

Laboratoire SUBATECH, UMR 6457 CNRS-IN2P3 / Ecole des Mines de Nantes / PRES UNAM, 4 rue A. Kastler,

c

NU

44307 Nantes Cedex, France

The Key Laboratory of Special Function Materials and Structure Design, Ministry of Education, Lanzhou

MA

University, 730000 Lanzhou, China

Abstract

D

The Beishan granitic formation is being investigated as a potential host rock for a

TE

high-level radioactive waste repository in China. It is important to understand the retention processes, including influential parameters such as the metal ion

CE P

concentration, pH, ionic strength (I) and temperature. The present study deals with U(VI) adsorption on Beishan granite using batch-type experiments in a CO2-free atmosphere. U(VI) adsorption on granite is shown to be insensitive to ionic strength.

AC

Temperature has a positive effect on U(VI) adsorption indicating that surface reactions are endothermic. Combining X-ray photoelectron spectroscopy (XPS) analysis and adsorption data at 25 °C, a Generalized Composite (GC) model with three surface complexes, ≡SOUO2+, ≡SO(UO2)2(OH)2+ and ≡SO(UO2)3(OH)5, was constructed. The experimental data at 40 °C and 60 °C were fitted by the proposed model to obtain the equilibrium constants (K) of the surface reactions at these two temperatures. The enthalpy changes (ΔH) of the surface reactions were evaluated from the equilibrium constants obtained at three temperatures via the van’t Hoff equation. Finally, blind modeling predictions were performed to test the robustness of the proposed model and ΔH. Satisfactory agreement with the literature data confirmed *

Corresponding authors. Tel. +8613919217067; fax: +869318913551 (Z. Chen). Tel.: +869318913278; fax: +869318913551 (Z. Guo). E-mail addresses: [email protected] (Z. Chen), [email protected] (Z. Guo). 1

ACCEPTED MANUSCRIPT that this GC model with ΔH proving a useful tool to predict U(VI) adsorption on granite samples, especially on Beishan granite at ambient/high temperature.

AC

CE P

TE

D

MA

NU

SC R

IP

T

Keywords: U(VI); Granite; Adsorption; XPS; Modeling; Enthalpy changes.

2

ACCEPTED MANUSCRIPT

1. Introduction

T

Due to its high compressive strength, low total porosity, good thermal

IP

conductivity and thermal stability, granite has been considered or chosen as the host

SC R

rock (i.e. geological barrier) for high-level radioactive waste repository in many countries, e.g. the granitic formations in Forsmark (Sweden) and Beishan (China). Therefore, a thorough understanding of radionuclide behavior in granite has become a

NU

pressing need for the safety assessment of these repositories (Keisuke et al., 2013; Holgersson, 2012; Kitamura et al., 1999; Papelis, 2001). Radionuclide retention is

MA

affected by many in-situ factors (e.g. alkalinity, atmospheric composition and so on). Temperature is one such factor, the effect of which differs according to different

D

adsorption mechanisms. These effects on radionuclide retention on granite should be

TE

considered when a safety assessment is performed. Uranium isotopes are important constituents of high-level nuclear waste (HLW)

CE P

due to their long half-lives (238U 4.47×109 y,

235

U 7.04×108 y,

234

U 2.455×105 y)

(Allard et al., 1984). A considerable number of studies have focused on the interactions between uranium and the minerals found in repository barriers

AC

(Aksoyoglu, 1989; Arnold et al., 1998; Chisholm-Brause et al., 2001; Aamrani et al., 2002; Pabalan and Turner, 1996; Schmeide et al., 2014; Sylwester et al., 2000). For pure minerals with a definite composition, like kaolinite, montmorillonite and quartz, the interaction processes have been described quantitatively by modeling (Gao et al., 2010; Pabalan and Turner, 1996; Sylwester et al., 2000; Guo et al., 2009b). However, modeling of complex mineral assemblages (e.g. granitic formations) is still a challenge. In general, two concepts have been proposed for such modeling (Davis et al., 1998, 2004): Component Additivity (CA), in which the surface of a complex mineral assemblage is considered to be a mixture of individual constituents with known surface properties, and Generalized Composite (GC), in which the surface reactions are assumed to take place on a type of “general” surface sites. The CA 3

ACCEPTED MANUSCRIPT model requires many modeling parameters. Nebelung and Brendler (2010) modeled uranium adsorption on granite considering quartz, albite, muscovite, hydroxylapatite and hematite as reactive phases. In total, 11 surface complexation reactions as well as

IP

T

9 surface hydrolysis reactions for these reactive phases were included in the modeling. The reactive phases of mineral assemblages and their percentages are not always

SC R

accessible, which makes it difficult to apply the CA model to other assemblages. The GC model is a useful tool to describe a complex natural system, but it is limited by its site-specific nature. However, Tertre et al. (2008) developed a non-electrostatic GC

NU

model to describe the adsorption of rare earth elements on basaltic rock. This model proved to be fairly robust in describing other aluminosilicate systems, which suggests

MA

that the GC model developed from a material with poor adsorption is not strongly site-specific for other materials with similar adsorption properties. The GC model is

D

also the method of choice when the modelling parameters for individual components

TE

are unavailable or not universal (Chen et al., 2014b). The temperatures in the near-field of repositories vary temporally and spatially

CE P

because of the decay of high-level nuclear waste (Duc et al., 2008; Tertre et al., 2005, 2006). Deep disposal of high-level radioactive waste causes thermal fluxes in the near

AC

surface environment of the waste with the temperature reaching a maximum of 90 °C (Wilson, 2011). Some studies show that the temperature effect on the distribution coefficients of monovalent, divalent and trivalent cations is measurable (Bauer et al., 2005; Chen et al., 2014c; Jin et al., 2014; Tertre et al., 2005, 2006). According to adsorption measurements performed at high temperatures with pure clay minerals, Tertre et al. (Tertre et al., 2005, 2006) concluded that the temperature effect depends on the adsorption mechanism, i.e. surface complexation increases slightly with increasing temperature, whereas the influence on cation exchange is negligible. However, the temperature effect reported for the adsorption of one element on minerals and mineral assemblages is not consistent. A negligible temperature effect was observed in an Eu(III)/smectite adsorption system from 25 to 80 °C (Bauer et al., 2005), whereas adsorption edges of Eu(III) on kaolinite and Na-montmorillonite were 4

ACCEPTED MANUSCRIPT found to shift toward lower pH values with a temperature increase from 20 to 150 °C (Tertre et al., 2006). A negligible temperature effect was also reported for Eu(III) adsorption on granite (Jin et al., 2014). These discrepancies indicate that it is difficult

IP

T

to evaluate the temperature effect on radionuclide adsorption in complex mineral

SC R

assemblages by the temperature effects measured on individual minerals. In the present study, we aimed to provide U(VI) adsorption data and the corresponding model for the safety assessment of a Chinese HLW repository that may

NU

be built in the Beishan granitic formation. Adsorption data and adsorption models of U(VI) on Beishan granite are rare and, to the best of our knowledge, even fewer

MA

studies take into account the qualitative and quantitative description of temperature effects on the U(VI)/granite system (Fan et al., 2014; Wei, 2012). For these purposes, experiments were carried out as a function of important parameters affecting U(VI)

D

adsorption (pH, ionic strength, concentration, temperature). A GC model was

TE

developed based on the adsorption data and spectroscopic information obtained from X-ray photoelectron spectroscopy (XPS) analyses. The parameters for the temperature

CE P

effect, e.g. enthalpy change (ΔH), were calculated from the surface complexation reaction constant (K) at three temperatures (25, 40 and 60 °C) via the van’t Hoff

AC

equation. Finally, the robustness of this GC model was tested with the literature data to see whether it could describe U(VI) adsorption on granite at ambient and high temperature.

2. Materials and methods 2.1 Chemicals U(VI) stock solution was made from UO2(NO3)2·6H2O (A.R. grade). All other chemicals used were of analytical grade. All solutions and suspensions were prepared with deionized water (18 MΩ·cm-1).

2.2 Granite preparation The granite sample was taken from borehole BS10 in the Beishan area (Gansu 5

ACCEPTED MANUSCRIPT Province,

northwest

China).

The

detailed

conditioning

procedures

and

characterization of granite were described in a previous study (Guo et al., 2011b). Generally, Beishan granite was crushed, sieved (200 mesh) and dispersed in 0.1

IP

T

mol/L NaCl solution (40 g/L). The pH of the suspension was adjusted to ～4.0 with HCl solution and kept for 24 h under shaking to remove carbonates. After

SC R

centrifuging and removing supernatant, the residue was washed with ethanol and finally dried at 45 oC. The N2-BET specific surface area was 4.1 m2/g. Scanning electron microscopy (SEM) was used to observe the morphological features of the

NU

granite sample. As shown in Fig. 1, the granite particles appeared to have different morphologies and sizes. An attempt was made to identify the chemical constituents in

MA

the sample by X-ray fluorescence (XRF). The analytical results were: SiO2 68.59 %, Al2O3 16.70 %, Na2O 8.95 %, K2O 3.82 %, Fe2O3 0.67 %, CaO 0.53 %, MgO 0.46 %.

D

Further investigation of the X-ray powder diffraction (XRD) pattern of the granite

TE

sample showed that quartz, biotite and albite were the main constituent minerals, which was similar to the previously reported microscopic observations of granite (Jan

AC

CE P

et al., 2008; Tsai et al., 2009).

Fig. 1. SEM micrograph of the granite sample. (A) × 500, (B) × 5000. The scale bars are 10 um in (A) and 1 um in (B), respectively.

2.3 Batch experiments Batch-type adsorption experiments with initial concentrations ranging from 5.0 ×10-6 mol/L to 3.0×10-4 mol/L were carried out in a glove box in a nitrogen atmosphere. The appropriate amount of granite sample was dispersed in NaCl 6

ACCEPTED MANUSCRIPT solution (0.01 or 0.1 mol/L) contained in a polyethylene tube at a solid-to-liquid ratio of 10 g/L. The pH of the granite suspensions was adjusted with small amounts of NaOH or HCl to obtain the desired pH values. Then, uranium stock solutions were

IP

T

added to the suspensions. The contact time (under shaking) will be discussed below. The pH of the supernatant was measured at the experimental temperature by a pH

SC R

meter (Metrohm 781 pH/ion meter No. 6.0234.100), which was first calibrated with three standard buffers at the same temperature. After the pH measurement, solid and liquid phases were separated by centrifugation at 18,000 g for 30 min. For the

NU

adsorption experiments at elevated temperature, a water bath was used to control the temperature. The concentration of U(VI) in the supernatant was analyzed by

MA

spectrophotometry at a wavelength of 652 nm using U(VI) and Arsenazo-III complex following the methodology published in Tao et al. (2000) .

D

The amount of U(VI) adsorbed on granite (q, mol/g) and the adsorption

TE

percentage of U(VI) were calculated as:

CE P

(1)

(2)

AC

where C0 is the initial concentration, Ceq (mol/L) is the aqueous concentration at equilibrium, V (L) is the volume of the aqueous solution and m (g) is the mass of the granite. All the experimental data were the average of duplicate or triplicate experiments and the relative error of the experimental data was less than 5 %.

2.4 XPS study The XPS samples were prepared to have a sufficient amount of U(VI) on granite (U(VI) loading [predicted by the model] = 1-6×10-6 mol/g). After a contact time similar to that fixed for the adsorption experiments, the solids were filtered, washed and stored as wet pastes. XPS data were obtained with an ESCALabAB210 surface microanalysis system 7

ACCEPTED MANUSCRIPT (VG Scientific). The photon source was an Al monochromatic X-ray source emitting an incident X-ray beam at 1360 eV with a FWHM (full-width half-maximum) of 0.50 eV. The sample, fixed on a metallic plate, was analyzed in a chamber under a 5×10-9

IP

T

Pa vacuum. The photopeaks were recorded at a constant pass energy of 30 eV. A low-resolution survey spectrum over a binding energy range of 0-1380 eV was

SC R

acquired for each scan. High-resolution spectra of each detected element were obtained for up to 12 scans each. All spectra were fitted by one Gaussian-Lorentzian function after subtraction of the background (Shirley baseline). The system was well

NU

calibrated in energy, i.e. the energy of the oxygen 1S spectrum (532.35 ± 0.03 eV) was in agreement with the reference data of 532.5 eV (Zhang et al., 1999). U 4f width

MA

was determined by analyzing reference uranium samples UO2(NO3)2·6H2O, and a FWHM of 2.40 eV was obtained. Therefore, the FWHM was kept constant at 2.40 eV

D

for all the fittings of the peaks corresponding to uranium surface species. Note that

TE

two surface species can be discriminated when the difference between their respective binding energies is more than 0.2 eV (Kowal-Fouchard et al., 2004; Teterin et al.,

CE P

1999).

2.5 Modeling

The nonlinear least square optimization program PHREEQC was used in

AC

modeling calculations (Parkhurst et al., 1999). Activity coefficients of aqueous species were calculated by the Davies equation. U(VI) thermodynamic data used in modeling (listed in Table 1 ) were taken from the NEA (Nuclear Energy Agency) database (Grenthe et al., 1992), except for those of UO2(OH)2(aq) which were taken from the Nagra/PSI Chemical Thermodynamic Data Base (Hummel et al., 2002); in the NEA database, only the maximum limit of the equilibrium constant was given for this species. The standard enthalpy change of some U(VI) hydrolysis reactions, unavailable in both the NEA and the Nagra/PSI databases, were assumed to be 0. A model with no electrostatic term was used to describe the adsorption measurements. The surface site capacity is estimated using the GC concept (Guo et al., 2011b). A stepwise approach was adopted to get a best fit to the adsorption data, which is 8

ACCEPTED MANUSCRIPT presented in more detail below in the section 3.2.2.

T

Table 1. Thermodynamic data for aqueous of U(VI) used in modeling (Grenthe et al., 1992;

IP

Hummel et al., 2002).

SC R

Reactions

AC

CE P

TE

D

MA

NU

q)

log K

Δ r H mθ （KJ/mol）

0.17 -1.10 -5.20 -12.0 -19.2 -33.0 -2.70 -5.62 -11.9 -15.55 -31.0 -21.9 9.68 16.94 21.60 54.00 -19.01 -17.5 -72.5

8.0 15 58 20 ─ ─ ─ 54 ─ 105 ─ ─ 5.0 18.5 -39.2 -62.7 ─ ─ ─

3.1 Experimental results 3.1.1 Adsorption kinetics and effect of solid-to-liquid (m/V) ratio To establish the equilibrium time for adsorption and to know the kinetics of the adsorption process, U(VI) adsorption on granite was investigated as a function of contact time (Fig. 2). The adsorption of U(VI) by granite is rapid and reaches equilibrium within 5 hours. In practice, a period of 72 h was used as the contact time

for the following experiments to ensure that adsorption equilibrium was reached. The rapid initial uptake of U(VI) is consistent with the results of U(VI) adsorption on 9

ACCEPTED MANUSCRIPT silica (Guo et al., 2009b). By assuming that U(VI) adsorption is dependent on the amount, a pseudo-second order rate equation (Ho and Mckay, 1999) is applied to fit the kinetics of U(VI) adsorption. The t/qt (h·g/mmol) versus t (h) (Fig. 2 ) indicates

IP

T

that the kinetics of U(VI) adsorption can be well fitted by the linear form of the rate

SC R

equation as follows:

(3)

where k (mmol/(g·h)) is the rate constant of adsorption, qe (mmol/g) the equilibrium

-3

MA

1.2x10

D

-3

8.0x10

-4

-4

AC

4.0x10

1.0x10

5

8.0x10

4

6.0x10

4

4.0x10

4

2.0x10

4

t/qt (h·g/mmol)

TE

1.6x10

CE P

qt (mmol/g)

NU

adsorption capacity, and qt (mmol/g) the amount of U(VI) adsorbed at time t (h).

0.0

y = 876.2x + 520.1 2 R = 0.9977

0

20

40

60

80

100

120

t (h)

0

20

40

60

80

100

120

t (h) Fig. 2. Adsorption kinetics of U(VI) on granite at CU(VI) = 2.17 × 10-5 mol/L with the conditions of pH = 5.65 ± 0.10, m/V = 10 g/L, T = 25 ± 2 °C, I = 0.01 mol/L NaCl.

The influence of the solid-to-liquid ratio (m/V, g/L) on U(VI) adsorption on granite is shown in Fig. 3. As expected, a higher m/V is desirable for U(VI) adsorption within the range of 1 – 30 g/L. This can be explained by the fact that more adsorption sites are available when the granite content increases. The distribution coefficient (Kd, mL/g) is also plotted as a function of m/V in Fig. 3. Kd can be 10

ACCEPTED MANUSCRIPT calculated from the concentration of U(VI) in suspension (C0) and in the supermatant liquid (Ceq) according to equation (4):

IP

T

(4)

SC R

where V (L) is the volume of aqueous solution and m (g) is the mass of the granite. It was found that Kd is almost independent of the mass of granite, which justifies the assumption of an adsorption process. The result was also similar for the adsorption of

MA

100

D TE

60

CE P

Adsorption %

80

10

4

10

3

10

2

10

1

10 35

0

40

AC

20

: Adsorption %; , : Kd.

0

0

5

10

15

20

25

30

Kd (mL/g)

NU

U(VI) on silica (Guo et al., 2009b; Zhang and Tao, 2002).

m/V (g/L) Fig. 3. Effect of m/V on U(VI) adsorption on granite at CU(VI) = 5.21 × 10-5 mol/L with the conditions of pH = 5.40 ± 0.10, T = 25 ± 2 °C, I = 0.01 mol/L NaCl. The points are experimental data and the line correspond to modeling results as indicated below.

3.1.2 Effect of pH, ionic strength and U(VI) concentration pH is an important parameter for the adsorption behavior of radionuclides. It not only affects the degree of protonation and deprotonation of granite surface hydroxyl groups, but also determines the distribution of the radionuclide species in solution 11

ACCEPTED MANUSCRIPT (Guo et al., 2011b; Yang et al., 2010). In practice, adsorption edges (adsorption percentage vs. pH) are commonly used to constrain surface complexation models (SCMs). In the present paper, two adsorption edges of U(VI) at different ionic

IP

T

strengths (0.01 and 0.10 mol/L) were collected (Fig. 4). The pH edge presents two distinct parts: (i) for pH up to 7, the adsorption of U(VI) increases greatly with the

SC R

increase in pH, and (ii) the adsorption of U(VI) decreases slightly when the pH is higher than 7. The pH edges at different ionic strengths are very close, indicating that U(VI) adsorption is insensitive to ionic strength. A weak dependence of adsorption on

NU

ionic strength was also reported for U(VI) adsorption on olivine ( Aamrani et al., 2002)

MA

and quartz (Nair et al., 2014).

Different adsorption mechanisms (including ion-exchange, inner-sphere and outer-sphere surface complexation) can be distinguished by studying the effects of

D

ionic strength on anion/cation adsorption (Aamrani et al., 2002; Sun et al., 2012;

TE

Hayes et al., 1988). Inner-sphere surface complexation is insensitive to ionic strength, while ion-exchange and outer-sphere surface complexation reactions are sensitive to

CE P

ionic strength. In this study, the lack of strong ionic strength dependence was regarded as an indication that inner-sphere surface complexation is the dominant mechanism U(VI)

adsorption

AC

for

12

on

granite.

ACCEPTED MANUSCRIPT

60

C

40 B

NU

A

20

3

4

5

6

pH

7

8

9

10

D

0 2

SC R

IP

80

T

: I = 0.01 M; : I = 0.10 M.

, ,

MA

Adsorption (%)

100

CE P

TE

Fig. 4. Adsorption edge of U(VI) on granite at different ionic strengths (I) with the conditions of CU(VI) = 5.21×10-5 mol/L, m/V = 10 g/L, T = 25 ± 2 °C. The points are experimental data and the lines are modeling results. The dashed lines illustrate the modeled contributions of different surface species to U(VI) adsorption at I = 0.01 mol/L: A (≡SOUO2+), B (≡SO(UO2)2(OH)2+) and C (≡SO(UO2)3(OH)5).

In addition to pH adsorption edges, adsorption isotherms are also commonly

AC

measured and used for constructing SCMs. Fig. 5 shows the adsorption of U(VI) on granite as a function of U(VI) equilibrium concentration at different pH values (i.e. ~ 4.40 and 6.60). The adsorbed amount of U(VI) at pH 6.60 ± 0.10 is higher than that at 4.40 ± 0.10, which is consistent with the results for the pH adsorption edges (Fig. 4).

13

ACCEPTED MANUSCRIPT

10

-4

IP

-2

SC R

10

T

0

NU

q (mmol/g)

10

, ,

-6

10

-7

10

-6

MA

10

10

-5

: pH = 4.40 ± 0.10; : pH = 6.60 ± 0.10.

10

-4

10

-3

D

Ceq (mol/L)

TE

Fig. 5. Adsorption isotherm of U(VI) at pH = 4.40 ± 0.10 and 6.60 ± 0.10 with the conditions of m/V = 10 g/L, T = 25 ± 2 °C, I = 0.01 mol/L NaCl. The points are experimental data and the lines are modeling results.

CE P

3.1.3 Effect of temperature

The temperature in the near-field of a repository varies temporally and spatially

AC

because of the decay of high-level nuclear waste. Therefore, it is necessary to evaluate the temperature effect on radionuclide adsorption on barriers of repository (e.g. granite). It is described quantitatively by ΔH of surface reactions involving radionuclide adsorption (Duc et al., 2008; Tertre et al., 2005, 2006). Fig. 6 shows the adsorption edges for U(VI) on granite at three temperatures (25, 40 and 60 °C). The aqueous U(VI) speciation changes a bit with the temperature increase (Fig. 7). Temperature appears to have a significant effect on U(VI) adsorption, i.e. the edge generally shifts to a lower pH when the temperature increases. The increase in U(VI) adsorption with increasing temperature indicates that it is endothermic. Similar observations were reported for Eu(III) adsorption on clay minerals from 20 to 150 °C (Tertre et al., 2006). Angove et al. (1998) also found that adsorption of Cd(II) and Co(II) on a kaolinite surface was temperature-dependent, and thermodynamic 14

ACCEPTED MANUSCRIPT parameters estimated from both the Langmuir and surface complexation models showed that adsorption of Cd(II) and Co(II) was endothermic. U(VI) adsorption at elevated temperature decreases more abruptly than adsorption at ambient temperature

IP

T

when pH>7. Further study is required to determine the mechanism responsible for this

SC R

decrease.

NU

80

MA

60

, , ,

D

40 20 0

3

CE P

2

TE

Adsorption (%)

100

4

5

6

7

o

: T = 25 ± 2 C; o : T = 40 ± 2 C; o : T = 60 ± 2 C;

8

9

10

pH

AC

Fig. 6. U(VI) adsorption as a function of pH at different temperatures with the conditions of CU(VI) = 5.21×10-5 mol/L, m/V = 10 g/L, I = 0.01 mol/L NaCl. The points show the experimental data and the lines represent the results calculated by the proposed model with the equilibrium constants listed in Table 3.

15

ACCEPTED MANUSCRIPT +

90

-

(UO ) (OH)5

A

2 3

2+

2

(UO ) (OH)7

UO

2 3

2

60

UO (OH)3

-

2+

+

(UO ) (OH)2

UO ) (OH)7

2 2

30

2 4

+ 2

B

90

+ 2+

UO

(UO ) (OH)5

2

2

2 3

2+

30

(UO ) (OH)2 2 2

0

C +

2+

UO OH

UO

2

2

60

2+

(UO ) (OH)2

30

2 2

0

2

3

4

2 3

-

UO (OH)3

+

5

6

-

UO (OH)3 2

(UO ) (OH)5

NU

90

-

(UO ) (OH)7

SC R

60

+

UO OH

IP

T

0

MA

U(VI) species (%)

UO OH

2

2 3

-

(UO ) (OH)7 2 3

7

8

9

10

D

pH

TE

Fig. 7. Speciation of 5.21×10-5 mol/L U(VI) in 0.01 mol/L NaCl solutions at 25 oC (A), 40 oC (B) and 60 oC (C).

CE P

3.2 Surface complexation modeling of U(VI) adsorption at ambient temperature Because granite consists of multiple minerals, a precise description of such a complex adsorption system remains a challenge. The Generalized Composite (GC)

AC

approach is one method of choice, which has been proven to be effective for such a system (Chen et al., 2014b; Davis et al., 2004; Tertre et al., 2008). In the GC approach, a rock is assumed to have ‘‘general’’ surface sites on which surface reactions take place. In addition, the GC approach ignores protonation and deprotonation reactions and the electrostatic effect, which makes it a simple model with few adjustable parameters. Our previous studies indicated that the adsorption of Eu(III), Am(III), Se(IV), Ni(II) and Cu(II) on granite could be described using the GC approach (Chen et al., 2013; Guo et al., 2011a, 2011b; Jin et al., 2014) and thus it was applied in this study. 3.2.1 Estimation of site capacity

16

ACCEPTED MANUSCRIPT The site capacity, i.e. the density of reactive sorption sites at the mineral surface, is a key parameter that must be determined before constructing a surface complexation model. The “general” surface site capacity of a mineral assemblage can

IP

T

be obtained by fitting adsorption isotherms using the Langmuir equation (Tertre et al., 2008), the assumptions of which are: (i) surface sites and adsorbate occur in a 1:1

SC R

ratio response; and (ii) adsorption free energy is independent of surface electrostatic interactions. Tertre et al., (2008) deduced the site capacity of surface complexation on basaltic rock by comparing measured CEC and fitted site densities (by the Langmuir

NU

equation) of Eu(III) and phosphate. They attributed the difference between CEC and fitted density to the site capacity of cation exchange because ionic strength strongly

MA

influences Eu(III) adsorption. However, this is not the case in our study in which no effect of ionic strength was observed (Fig. 4). Thus, we can be sure that it is

D

unnecessary to consider cation exchange sites on granite for U(VI) adsorption, and

TE

only one type of “general” site (≡SOH) is considered. In order to obtain a reliable site density, two adsorption isotherms for both U(VI)

CE P

and phosphate were measured at pH 4.40 and pH 3.30, respectively (Fig. 8). The U(VI) isotherm at pH 4.40 was selected to estimate site density because the dominant

AC

species in solution is UO22+ under these conditions and free from precipitation even at the highest U(VI) concentration (Fig. 7). To assess the reliability of the site capacity thus determined, complementary experiments were carried out with phosphate. Previous studies have shown that phosphate adsorption is not influenced by cation exchange and phosphate forms monodentate inner-sphere complexes on a solid/liquid interface at low pH (Arai and Sparks, 2001; Chen et al., 2014a). The Langmuir model is expressed in equation (5): (5)

where q and Ceq refer to the concentration of solid-phase (mol/g) and liquid-phase (mol/L) after U(VI)/phosphate adsorption equilibrium. Smax is the maximum

17

ACCEPTED MANUSCRIPT adsorption capacity (mol/g), KL is the Langmuir equation constant (L/mol). The parameters are fitted by matching the Langmuir equation to the experimental

T

data, using a least squares analysis method (Fig. 8). Site densities of 2.4×10-6 mol/g

IP

and 2.8×10-6 mol/g were obtained for U(VI) and phosphate, respectively (Table 2).

SC R

As expected, the site density determined for phosphate is close to the value for U(VI) at pH = 4.40. The closeness of the fitted site values for U(VI) and phosphate isotherms may indicate that the reactive sites on the general surface are the same in

NU

both cases and our experiments yield a relatively good estimate of their densities. Therefore, we can assume the site distribution of the general surface is 2.4×10-6 mol/g

MA

(5.85×10-7 mol/m2) for granite in the following surface complexation modeling. This site capacity is similar to quartz (3×10-7 mol/m2) (Nair et al., 2014) but lower than

TE

D

silica (2.6×10−6 mol/m2) (Guo et al., 2009b).

-2

-2

10

CE P

10

q (mmol/g)

-3

10

B

AC

q (mmol/g)

A

-4

10

-3

10

pH = 4.40 ± 0.10

pH = 3.30 ± 0.10

-5

10

-4

-6

10

-5

10

-4

10

-3

10

Ceq (mol/L)

10

-5

10

-4

10

-3

10

Ceq (mol/L)

Fig. 8. Experimental adsorption data as a function of aqueous equilibrium concentration with the corresponding Langmuir isotherm. (A) U(VI) at pH = 4.40 ± 0.10; (B) Phosphate at pH = 3.30 ± 0.10.

18

ACCEPTED MANUSCRIPT Table 2. Experimental conditions and Langmuir parameters (maximal site densities, adsorption constants) for U(VI) and PO43- adsorption. Langmuir parameters pH ± 0.10

Ionic strength (NaCl) in mol/L

T

Adsorbate

4.40

0.01

Phosphate

3.30

0.01

2.4×10-6

SC R

U(VI)

IP

Smax in mol/g

2.8×10-6

KL in L/mol 2.2×106 1.6×106

3.2.2 Surface complexation model construction: Combining U(VI) adsorption data

NU

and XPS study

MA

As mentioned above, because a strong ionic strength effect was not observed for U(VI) adsorption on granite, the inner-sphere surface complex is the only one to be considered in the modeling. Although it was assumed that only one type of “general”

D

surface site is available for surface complexation reactions, the number of reactions,

TE

i.e. the number of surface species still needs to be constrained. The number of

CE P

reactions differs in the one-site surface complexation models. Guo et al. (2009c) interpreted U(VI) adsorption on γ-alumina using two surface complexation reactions. However, more studies considered three reactions to describe U(VI) adsorption (Guo

AC

et al., 2009a, 2009b;Yang et al., 2010). The surface complexation reactions can be more constrained, especially by XPS spectra analysis as performed here (Fig. 9). To assess the change in species distribution, samples A, B and C were prepared at pH = 4.50, pH = 5.20, and pH = 6.30, respectively. The spectrum corresponding to the lowest pH value (Fig. 9A) is well-fitted with two components located at 377.90 eV and 380.58 eV, whereas for the other two samples three binding energies (377.90 eV, 380.58 eV, and 381.36 eV) are needed to account for the total spectrum. It is impossible to achieve a good fit for samples B and C considering only two components. The differences between the three components are higher than 0.2 eV, indicating that three different surface species are included in samples B and C. We can conclude from the deconvolution of these three

19

ACCEPTED MANUSCRIPT samples that: (i) U(VI) forms two different surface species at low pH; (ii) another surface species occurs as the pH increases and (iii) the proportion of each surface species changes according to the pH change (Table 3). The information provided by

IP

T

the XPS study will be taken into account in the following surface complexation

SC R

model.

According to U(VI) speciation in the aqueous phase (Fig. 7A), the first adsorbate considered is free UO22+, which is the dominant species up to pH below 5 in the

NU

aqueous phase. The corresponding surface complexation reaction on ‘‘general’’

MA

surface sites can be described as:

(6)

It was found that reaction (6) alone could not explain the adsorption edges,

D

especially at pH higher than 4.50, which is consistent with XPS analysis in which two

TE

surface species were predicted at pH 4.50. More U(VI) species in the aqueous phase must be considered as potential adsorbates. In the U(VI) concentration range of this

CE P

study, (UO2)2(OH)22+, (UO2)3(OH)5+ and (UO2)3(OH)7- are the dominant U(VI) species in the pH range of 5 to 9 (Fig. 7A), Thus, besides reaction (6), three surface

AC

complexation reactions on “general” surface sites could be written, respectively, as: (7) (8) (9) In fitting exercises, we took into account all of the experimental data, including adsorption under different m/V (Fig. 3), pH edges under different ionic strengths (Fig. 4) and adsorption isotherms at different pH (Fig. 5). The adsorption model may be inconsistent with the XPS analysis if the four surface reactions above are considered. In agreement with this statement, fitting exercises suggested that reaction (9) was not important for all of the experimental data. A relatively simple model with three surface reactions could successfully explain the experimental data of U(VI) 20

ACCEPTED MANUSCRIPT adsorption. The surface complex reactions and the corresponding equilibrium constants, resulting from the best fit to the experimental data, are listed in Table 3, and

T

detailed deconvolutions of XPS spectra are listed in Table 4.

IP

According to the distribution of surface species predicted by modeling (Fig. 4),

SC R

≡SOUO2+ and ≡SO(UO2)2(OH)2+ are the main surface complexes at pH < 5.0, and ≡SO(UO2)3(OH)5 occurs from pH = 5.0 and becomes dominant when pH > 5.5 under the experimental conditions. Therefore, the peak (381.36 eV) included only in sample

NU

B and sample C is assigned to ≡SO(UO2)3(OH)5. The proportion of surface species ≡SOUO2+ decreases from pH 4.5 to pH 7. A similar trend is observed for peak (377.90

MA

eV) in XPS analysis so this is assigned to ≡SOUO2+. The last unassigned peak (380.58 eV) also has a consistent trend with ≡SO(UO2)2(OH)2+, the proportion of which is higher than ≡SOUO2+ but less than ≡SO(UO2)3(OH)5 at pH > 6. The good

D

agreement between the model prediction and XPS analysis suggests that our model is

TE

realistic.

CE P

It should be noted that our model underestimates experimental data for U(VI) aqueous concentration lower than 10-5 mol/L at equilibrium (Fig.5). This may be because of our assumption that granite includes only one type of site. In addition to

AC

the main sites predicted by Langmuir fitting, granite may have some sites with low capacity and high affinity, like the “strong site” in a previously published GC model for sediments (Davis et al., 2004). Because the underestimation is very limited and the basic idea of this study is to describe uranyl adsorption using as few parameters as possible, neglecting these possible “strong sites” is deemed acceptable.

21

IP

T

ACCEPTED MANUSCRIPT

SC R

Table 3. The modeling parameters for U(VI) adsorption on granite. Description of Beishan granite 4.1 m2/g

General surface sites:

SOH

5.85×10-7 mol/m2

NU

Specific surface area:

MA

Surface reactions and corresponding equilibrium constants

25 ± 2 °C

ΔH

logK 40 ± 2 °C

60 ± 2 °C

（KJ/mol）

-0.1

0.2

0.5

32

-5.4

-4.5

-3.4

108

-16.5

-15.4

-13.8

147

AC

CE P

TE

D

U(VI) surface complexation reactions

Fig. 9. U 4f7/2 spectra adsorption on granite at CU(VI) = 1.0×10-4 mol/L, m/V = 10 g/L, I = 0.01 mol/L NaCl. (A) pH = 4.50; (B) pH = 5.20; (C) pH = 6.30. Experimental data are discrete points and calculated curves are solid black lines. 22

IP

T

ACCEPTED MANUSCRIPT

Table 4. Spectroscopic characteristics of U(VI) adsorption on granite and spectral deconvolution.

2.40

NU

2.40

SC R

≡SOUO2+/ ≡SO(UO2)2(OH)2+

≡SOUO2+/ ≡SO(UO2)2(OH)2+/≡SO(UO2)3(OH)5

≡SOUO2+/ ≡SO(UO2)2(OH)2+/≡SO(UO2)3(OH)5

(eV, relative % content) 377.90/380.58 (71.5/28.5) 377.90/380.58/381.36 (36.2/39.4/24.4) 377.90/380.58/381.36 (26.6/33.5/39.9)

D

6.30

2.40

Deconvolution

3.2.3 Surface complexation model application: U(VI) adsorption at high temperature

TE

C

5.20

Predicted species

and determination of enthalpies (ΔH)

CE P

B

4.50

FWHM (eV)

Thermodynamic parameters were estimated from the temperature dependence of the surface equilibrium constants following the van’t Hoff equation:

AC

A

pH

MA

Samples

(10)

where K is the equilibrium constant of the surface reaction at absolute temperature T and R is the gas constant. Provided enthalpy and entropy are constant across the temperature range studied, a plot of log K versus 1/T is linear with a slope of ΔH/2.303R, and the enthalpy of adsorption can be calculated from surface complexation model parameters. The surface equilibrium constants at 25, 40 and 60 °C were calculated to assess ΔH for each surface reaction in the model. K at 40 and 60 °C were obtained by fitting experimental data at these two temperatures with the same model described above. It should be noted that the prerequisite information for these fittings is that all ΔH of the 23

ACCEPTED MANUSCRIPT related reactions, both in the aqueous phase and the surface, are known. However, as the ΔH of some U(VI) hydrolysis reactions are not available in either the NEA or the Nagra/PSI databases, these ΔH were assumed to be 0 in calculations, the applicability

IP

T

of which has been proven by Duc et al. (2008) and Yang et al. (2010). The obtained K values are listed in Table 3 and the calculated results are represented by lines in Fig.

SC R

10. It was found that all K values of the reactions in Eqs. (6)-(8) increased with

NU

increasing temperature.

5

logK = ﹣1697×1/T+6

MA

2

-

≡SOUO2

-5

R = 0.9948 logK = ﹣5672×1/T+14 2

R = 0.9999

+

-10

logK = ﹣7677×1/T+9

≡SO(UO2)3(OH)5

CE P

-15

D

≡SO(UO2)2(OH)2

TE

log K

0

2

R = 0.9938

AC

-20 -3 -3 -3 -3 -3 -3 2.9x10 3.0x10 3.1x10 3.2x10 3.3x10 3.4x10 -1

1/T (K )

Fig. 10. The dependence of equilibrium constants (K) of U(VI) surface complexation reactions on temperature according to the van’t Hoff equation.

According to the van’t Hoff equation, logK values were plotted against 1/T, and the plots for reactions in Eqs. (6)–(8) are illustrated in Fig. 10. Good correlation coefficients in all cases may imply that the model assumptions are to some extent reasonable. The ΔH of reactions in Eqs. (6)–(8) were calculated from the slopes of these plots. Positive ΔH values mean that the overall process of U(VI) adsorption is endothermic. The proposed GC model and ΔH of surface reactions enable the adsorption behavior of U(VI) on granite at other temperatures to be predicted.

24

ACCEPTED MANUSCRIPT 3.3 Blind predictions for the literature data To test the robustness of the proposed model and ΔH values, literature data at

IP

considering the experimental conditions used to obtain them.

T

ambient/elevated temperature were summarized and predicted by our model

SC R

3.3.1 Literature data at ambient temperature

The data from four different studies were extracted and compared to predictions

NU

from our model (Fig. 11). The granites in these studies originate from different areas (Table 5). The data on Forsmark granite (Sweden) were not included because the Kd was given in a wide range (5.0×10–4 − 1.2×10–1 m3/kg at pH 7-9) (Crawford et al.,

MA

2006). The capacity of the total “general” surface sites ≡SOH was re-calculated according to their specific surface area (SSA) in the modeling. The presence of CO2

TE

D

and the composition of the background solution were also taken into account. Table 5. Summary of predictions for the literature data. SSA

U(VI) Concentration

NaHCO3/Ca(HCO3)2 in

Prediction by

(oC)

(m2/g)

(mol/L)

suspention

our model

20-50

1

1×10-5

Not mentioned

Acceptable

Raw sample

25-65

2.5

5×10-6 - 1×10-4

Not mentioned

Acceptable

Treated by acid

25

0.35

1×10-9 - 1×10-3

Not mentioned

Underestimated

Raw sample

23

0.18

1.1×10-7-2.8×10-7

1-46.2 mg/L HCO3-

Underestimated

treatment

Treated by acid

AC

Gyeonggi-do granite (Korea) (Keum et al., 2002) Beishan 03 granite (China) (Fan et al., 2014) Eibenstock granite (Germany) (Nebelung and Brendler, 2010) Grey granite (Canda) (Ticknor, 1994)

Temperature

CE P

Sample Granite sample

Our model can give acceptable predictions for the data for Gyeonggi-do granite (Fig. 11A, Keum et al., 2002) and Beishan 03 granite (different borehole BS03 from ours, ) (Fig. 11B, Fan et al., 2014). The data for Beishan 03 granite were collected

25

ACCEPTED MANUSCRIPT using a raw sample, which means that calcite was included. Because the concentration of carbonates was not mentioned in this study, we assumed that the system was equilibrated with CO2 in atmosphere. The relatively good agreement may indicate that,

IP

T

under the conditions of this study, the carbonate effect is limited.

SC R

Our model clearly underestimated adsorption on Eibenstock granite (Fig. 11C, Nebelung and Brendler, 2010). The Eibenstock granite was treated with acid, which suggests the underestimation cannot be attributed to the effects of carbonate or natural

NU

organic matter (NOM). It is important to note that the specific surface area (SSA) of this granite sample is much lower than our sample (Table 5). Comparing with

MA

Gyeonggi-do granite and Beishan 03 granite, this underestimation may indicate that the site capacity of our model is applicable to granite with a similar SSA, but not to granite with a very small SSA. As mentioned above, the site capacities were

D

re-calculated according to the SSA. A low site capacity was estimated for granite with

TE

a small SSA, and this causes insufficient sites for U(VI) in the modeling. A much better fit (dashed line in Fig. 11C) was obtained when site capacity was not

CE P

re-calculated. Therefore, we speculated that part of the surface, i.e. some sites, although not exposed in BET measurement, become available when contacting the

AC

aqueous phase. The site capacity of our model was obtained from adsorption data and Langmuir fitting, which to some extent is independent of measured SSA. Therefore, whether this site capacity can be converted according to the SSA difference needs to be confirmed with further study. Our model was used to predict the data obtained from system with a considerable amount of carbonates (Fig. 11D, Ticknor, 1994), and an obvious underestimation was observed. The grey granite sample used in this study has a small SSA (0.18 m2/g). However, this underestimation is not due to the issue of site capacity because underestimation is still obvious even using the original site capacity (predicted data was not shown). In addition to HCO3‾, it was observed that Ca2+ also greatly influences the prediction. Underestimation compared to the experimental data became higher when considering Ca2+ in calculation (Fig. 11D). The carbonate effect is 26

ACCEPTED MANUSCRIPT complicated: in addition to forming a ternary surface complex (Catalano et al., 2005), a slower uptake by incorporation may occur with time on calcite (Lakshtanov and Stipp, 2007). The U(VI) retention in such systems does not include only adsorption,

CE P

TE

D

MA

NU

SC R

IP

T

and was not specifically studied in this study.

AC

Fig. 11. Comparison of the data published by other groups with those calculated using the model proposed at ambient temperature; lines are the calculated results from the proposed model; symbols are the data published by other groups: (A) Gyeonggi-do granite, CU(VI) = 1.0×10-5 mol/L I = 0.01 mol /L NaClO4 , m/V = 250 g/L (Keum et al., 2002); (B) Beishan 03 granite, pH = 5.60 ± 0.20, I = 0.01 mol /L NaCl, m/V = 10 g/L (Fan et al., 2014); (C) Eibenstock granite, pH = 5.00, m/V = 10 g/L, I = 0.1 mol/L NaClO4 (Nebelung and Brendler, 2010); (D) Grey granite, SWG A-D represents four different synthetic groundwaters with different solution compositions (Ticknor, 1994).

3.3.2 Literature data at elevated temperature Our model and obtained ΔH were also applied to predict adsorption at elevated temperature. Fig. 12A (Keum et al., 2002) presents the adsorption edge of U(VI) on Gyeonggi-do granite at 35 °C and 50 °C. It was found that U(VI) adsorption increased with increasing temperature in the pH range of 4-9. Because the cation exchange 27

ACCEPTED MANUSCRIPT reaction was shown to be independent of temperature, this result suggests that the exchange reaction is not important for U(VI) adsorption on granite, which is consistent with the assumptions of our model and to some extent justifies our model.

IP

T

By re-calculating the surface reaction constants according to the ΔH we proposed, our

SC R

model can satisfactorily reproduce these experimental data.

Similarly, data on Beishan granite at elevated temperature (Fan et al., 2014) can be even better reproduced by our model (Fig. 12B). The good agreement with the

NU

experimental data at both ambient and elevated temperature suggests that our model is

MA

a generic one for predicting U(VI) adsorption on Beishan granite.

100

CE P

60

q (mmol/g)

TE

80

Adsorption (%)

B

D

A

-2

10

40

AC

20

， ，

： T = 35 oC； ： T = 50 oC.

-3

10

, ,

-4

10

: T = 45 oC; : T = 65 oC.

0

2

4

6

8

10

pH

-7

10

-6

10

-5

10

-4

10

-3

10

Ceq (mol/L)

Fig. 12. Comparison of the data published by other groups with those calculated using the model proposed at elevated temperature; lines are the results calculated by the proposed model; symbols are the data published by other groups: (A) Gyeonggi-do granite, CU(VI) = 1.0×10-5 mol/L I = 0.01 mol /L NaClO4 , m/V = 250 g/L (Keum et al., 2002); (B) Beishan 03 granite, pH = 5.60 ± 0.20, I = 0.01 mol /L NaCl, m/V = 10 g/L (Fan et al., 2014).

4. Conclusions Adsorption of U(VI) on Beishan granite increases with increasing pH and is

28

ACCEPTED MANUSCRIPT insensitive to ionic strength. Temperature has a positive effect on U(VI) adsorption. Considering the complexity of granite composition, we proposed a generalized composite approach to describe U(VI) adsorption quantitatively. The site density was

IP

T

obtained by fitting adsorption isotherms of both U(VI) and phosphate using the Langmuir equation. Based on the analysis of XPS measurements and adsorption data

SC R

at 25 °C, a generalized composite model considering the formation of only three surface complexes, ≡SOUO2+, ≡SO(UO2)2(OH)2+ and ≡SO(UO2)3(OH)5, was constructed. In order to evaluate the temperature effect quantitatively, ΔH of the

NU

surface reactions were calculated from the K obtained at three temperatures via the van’t Hoff equation. Finally, the proposed model with ΔH was applied to reproduce

MA

the literature data. Our model with ΔH can give a satisfactory prediction for granite samples with a SSA close to the granite used in this study. However, it

D

underestimated adsorption on granite samples with a really small SSA. This highlights

TE

the importance of understanding whether we can convert site capacity in the generalized model just by considering the SSA difference. Further study is needed to

CE P

explain this. For the system with considerable amounts of HCO3‾ and Ca2+, U(VI) retention, including both adsorption and incorporation processes, cannot be explained just by the present model. More parameters and corresponding experiments are

AC

needed. Overall, our GC model with ΔH appears to be a useful tool to predict U(VI) adsorption on different granite samples at ambient/high temperature, especially on Beishan granite, which is being considered the host rock for HLW repository in China.

Acknowledgements Financial support from the Special Foundation for High Level Waste Disposal, China (no. [2012] 494), the National Natural Science Foundation of China (Grant Nos. J1210001) and the Fundamental Research Funds for the Central Universities (lzujbky-2015-270 and lzujbky-2015-238) is gratefully acknowledged.

29

ACCEPTED MANUSCRIPT References

AC

CE P

TE

D

MA

NU

SC R

IP

T

Aamrani, F.Z.E., Duro, L., Pablo, J.D., Bruno, J., 2002. Experimental study and modeling of the sorption of uranium(VI) onto olivine-rock. Appl. Geochem. 17 (17), 399-408. Aksoyoglu, S., 1989. Sorption of U(VI) on granite. J. Radioanal. Nucl. Chem. 134 (2), 393-403. Allard, B., Olofsson, U., Torstenfelt, B., 1984. Environmental Actinide Chemistry. Inorg. Chim. Acta 94 (4), 205-221. Angove, M.J., Johnson, B.B., Wells, J.D., 1998. The Influence of Temperature on the Adsorption of Cadmium(II ) and Cobalt( II ) on Kaolinite. J. Colloid Inter. Sci. 204 (1), 93-103. Arai, Y., Sparks, D.L., 2001. ATR–FTIR Spectroscopic Investigation on Phosphate Adsorption Mechanisms at the Ferrihydrite–Water Interface. J. Colloid Inter. Sci. 241 (2), 317-326. Arnold, T., Zorn, T., Bernhard, G., Nitsche, H., 1998. Sorption of uranium(VI) onto phyllite. Chem. Geol.151 (1–4), 129-141. Bauer, A., Rabung, T., Claret, F., Schäfer, T., Buckau, G., Fanghänel, T., 2005. Influence of temperature on sorption of europium onto smectite: The role of organic contaminants. Appl. Caly Sci. 30 (1), 1-10. Catalano, J. G., Brown Jr., G. E., 2005. Uranyl adsorption onto montmorillonite:evaluation of binding sites and carbonate complexation. Geochim. Cosmochim. Acta 69 (12), 2995-3005. Chen, Z.Y, Jin, Q., Guo, Z.J., Montavon, G., Wu, W.S., 2014a. Surface complexation modeling of Eu(III) and phosphate on Na-bentonite: Binary and ternary adsorption systems. Chem. Eng. J. 256 (8): 61-68. Chen, Z.Y., Montavona, G., Guo, Z., Wang, X.K., Razafindratsima S., Robinet, J.C., Landesmana. C., 2014b. Approaches to surface complexation modeling of Ni(II) on Callovo-Oxfordian clayrock. Appl. Clay Sci. 101, 369-380. Chen, Z.Y., Montavona, G., Ribet, S., Guo, Z.J., Robinet , J.C., David, K., Tournassat, C., Grambowa, B., Landesmana, C. 2014c. Key factors to understand in-situ behavior of Cs in Callovo–Oxfordian clay-rock (France). Chem. Geol. 387, 47-58., Chen, ZY., Zhang.R., Yang, X.L., Wu, W.S., Guo. Z.J., Liu, C.l., 2013. Adsorption of Co(II) and Ni(II) on Beishan Granite: Surface Complexation Model and Linear Free Energy Relationship. Acta Phys. Chim. Sin. 29 (9), 2019-2026. Chisholm-Brause, C.J., Berg, J.M., Matzner, R.A., Morris, D.E., 2001. Uranium(VI) Sorption Complexes on Montmorillonite as a Function of Solution Chemistry. J. Colloid Inter. Sci. 233 (1), 38-49. Crawford, J., Neretnieks, I., Malmström, M., 2006. Data and uncertainty assessment for radionuclide Kd partitioning coefficients in granitic rock for use in SR-Can calculations. Sweden, SKB R-06-75, Swedish Nuclear Fuel and Waste Management Co. Davis, J.A., Coston, J.A., Kent, D.B., Fuller, C.C., 1998. Application of the surface complexation concept to complex mineral assemblages. Environ. Sci. Technol. 32 (19), 2820-2828. Davis, J.A., Meece, D.E., Kohler, M., Curtis, G.P., 2004. Approaches to surface complexation modeling of Uranium(VI) adsorption on aquifer sediments. Geochim. Cosmochim. Acta 68 (18), 3621-3641. Duc, M., Carteret, C., Thomas, F., Gaboriaud, F., 2008. Temperature effect on the acid-base behaviour of Na-montmorillonite. J. Colloid Inter. Sci. 327 (2), 472-476. Fan, Q.H., Hao, L.M., Wang, C.L., Liu, C.L., Wu, W.S., 2014. The adsorption behavior of U(VI) 30

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

SC R

IP

T

on granite. Environ. Sci. Proc. Impacts 16 (3), 534-541. Gao, L., Yang, Z.Q., Shi, K.L., Wang, X.F., Guo, Z.J., Wu, W.S., 2010. U(VI) sorption on kaolinite: effects of pH, U(VI) concentration and oxyanions. J. Radioanal. Nucl. Chem. 284 (3), 519-526. Grenthe, I., Fuger, J., Konings, R., Lemire, R. J., Muller, A. B.,Nguyen-Trung, C., Wanner, H, 1992. Chemical Thermodynamics of Uranium. (Wanner, H., Forest, I., eds.) Nuclear Energy Agency, Organisation for Economic Co-operation and Development, Elsevier. Guo, Z.J., Chen, Z.Y., Wu, W.S., Liu, C.L., Chen, T., Tian, W. Y., Li, C., 2011a. Adsorption of Se(IV) onto Beishan Granite. Acta Phys. Chim. Sin. 27 (9), 2222-2226. Guo, Z.J., Chen, Z.Y., Wu, W.S., Liu, C.L., Chen, T., Tian, W.Y., Li, C., 2011b. The adsorption of Eu(III) on Beishan granite. Sci. Sinic. Chim. 41 (5), 907–913. Guo, Z.J., Li, Y., Wu, W.S., 2009a. Sorption of U(VI) on goethite: effects of pH, ionic strength, phosphate, carbonate and fulvic acid. Appl. Radiat. Isot. 67 (6), 996-1000. Guo, Z.J., Su, H.Y., Wu, W.S., 2009b. Sorption and desorption of uranium(VI) on silica: experimental and modeling studies. Radiochim. Acta 97 (3), 133-140. Guo, Z.J., Yan, C., Xu, J., Wu, W. S., 2009c. Sorption of U(VI) and phosphate on γ-alumina: Binary and ternary sorption systems. Colloid. Surface. A 336 (1-3), 123-129. Hayes, K.F., Papelis, C., Leckie, J.O., 1988. Modeling ionic strength effects on anion adsorption at hydroes oxide/solution interfaces. J. Colloid Inter. Sci. 125 (2), 717-726. Ho Y.S., McKay, G., 1999. Pseudo-second order model for sorption processes. Process Biochem. 34 (5), 451-465. Holgersson, S., 2012. Studies on Batch Sorption Methodologies: Eu Sorption onto Kivetty Granite. Proc. Chem. 7 (12), 629-640. Hummel, W., Berner, U., Curti, E., Pearson, F. J., Thoenen, T., 2002. Nagra/PSI Chemical Thermodynamic Data Base 01/01. Universal Publishers/uPUBLISH.com USA, . Also issued as Nagra Technical Report NTB 02-16, Nagra, Wettingen, Switzerland. Jan, Y.L., Wang, T.H., Li, M.H., Tsai, S.C., Wei, Y.Y., Teng, S.P., 2008. Adsorption of Se species on crushed granite: A direct linkage with its internal iron-related minerals. Appl. Radiat. Isot. 66 (1), 14-23. Jin, Q., Wang, G., Ge, M.T., Chen, Z.Y., Wu, W.S., Guo Z.J., 2014. The adsorption of Eu(III) and Am(III) on Beishan granite: XPS, EPMA, batch and modeling study. Appl. Geochem. 47 (8), 17-24. Keisuke, F., Koushi Maeda, Y.H., Aoi, Y., Tamura, A., Arai, S., Yamamoto, Y., Aosai, D., Mizuno, T., 2013. Sorption of Eu(III) on Ganite: EPMA, LA–ICP–MS, Batch and Modeling Studies. Environ. Sci. Technol. 47 (22), 12811-12818. Keum, D.K., Choi, B.J., Baik, M.H., Hahn, P.S., 2002. Uranium(VI) adsorption and transport in crushed granite. Environ. Eng. Res. 7 (2), 103-111. Kitamura, A., Yamamoto, T., Nishikawa, S., Moriyama, H., 1999. Sorption behavior of Am(III) onto granite. J. Radioanal. Nucl. Chem. 239 (3), 449-453. Kowal-Fouchard, A., Drot, R., Simoni, E., Marmier, N., Fromage, F., Ehrhardt, J.J., 2004. Structural identification of europium(III) adsorption complexes on montmorillonite. New J. Chem. 28 (7), 864-869. Lakshtanov, L.Z., Stipp, S.L.S., 2007. Experimental study of nickel(II) interaction with calcite: 31

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

SC R

IP

T

Adsorption and coprecipitation. Geochim. Cosmochim. Acta 71 (15), 3686-3697. Nair, S., Karimzadeh, L., Merkel, B.J., 2014. Surface complexation modeling of Uranium(VI) sorption on quartz in the presence and absence of alkaline earth metals. Environ. Earth Sci. 71 (4), 1737-1745. Nebelung, C., Brendler, V., 2010. U(VI) sorption on granite: prediction and experiments. Radiochim. Acta 98 (9-11), 621-625. Pabalan, R.T., Turner, D.R., 1996. Uranium(6+) sorption on montmorillonite: Experimental and surface complexation modeling study. Aquat. Geochem. 2 (3), 203-226. Papelis, C., 2001. Cation and anion sorption on granite from the Project Shoal Test Area, near Fallon, Nevada, USA. Adv. Environ. Res. 5 (2), 151-166. Parkhurst, D.L., Appelo, C. A. J., 1999. User’s guide to PHREEQC (Version 2) – a computer program for speciation, batch-reaction, one dimensional transport, and inverse geochemical calculations.U.S.G.S. Water-Resources Report 99-4259. Schmeide, K., Gürtler, S., Müller, K., Steudtner, R., Josephl, C., Bok, F., Brendleret, V., 2014. Interaction of U(VI) with Äspö diorite: A batch and in situ ATR FT-IR sorption study. Appl. Geochem. 49 (49), 116-125. Sun, Y. B., Yang, S.T., Sheng, G.D., Guo, Z.Q., Wang, X.K., 2012. The removal of U(VI) from aqueous solution by oxidized multiwalled carbon nanotubes. J. Environ. Radioactiv. 105 (2), 40-47. Sylwester, E.R., Hudson, E.A., Allen, P.G., 2000. The structure of uranium (VI) sorption complexes on silica, alumina, and montmorillonite. Geochim. Cosmochim. Acta 64 (14), 2431-2438. Tao, Z.Y., Chu, T.W., Du, J.Z., Dai, X.X., Gu, Y.Y., 2000. Effect of fulvic acids on sorption of U(VI), Zn, Yb, I and Se(IV) onto oxides of aluminum, iron and silicon. Appl. Geochem, 15 (2), 133-139. Tertre, E., Berger, G., Castet, S., Loubet, M., Giffaut, E., 2005. Experimental sorption of Ni2+, Cs+ and Ln3+ onto a montmorillonite up to 150°C. Geochim. Cosmochim. Acta 69 (21), 4937-4948. Tertre, E., Berger, G., Simoni, E., Castet, S., Giffaut, E., Loubet, M., Catalette, H., 2006. Europium retention onto clay minerals from 25 to 150°C: Experimental measurements, spectroscopic features and sorption modelling. Geochim. Cosmochim. Acta 70 (18), 4563-4578. Tertre, E., Hofmann, A., Berger, G., 2008. Rare earth element sorption by basaltic rock: experimental data and modeling results using the “Generalised Composite approach”. Geochim. Cosmochim. Acta 72 (4), 1043-1056. Teterin, Y.A. Teterin, A.Y., Lebedev, A.M., Dementjev, A.P., Utkin, I.O., 1999. X-ray photoelectron spectroscopy investigation of the interaction of U (VI) and Fe (III) with natural humic acid in aqueous solutions. J. Prakt. Chem. 341 (8), 773-777. Ticknor, K.V., 1994. Uranium Sorption on Geological Materials. Radiochim. Acta 64 (3-4), 229-236. Tsai, S.C., Wang, T.H., Li, M.H., Wei, Y.Y., Teng, S.P., 2009. Cesium adsorption and distribution onto crushed granite under different physicochemical conditions. J. Hazard. Mater. 161 (2–3), 854-861. Wei, H. G., 2012. Investigation of Sorption and Migration Effect Factors of Uranium on Granites 32

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

SC R

IP

T

in Beishan. Geosci. 26 (4), 823-828. Wilson, J., Savage, D., Bond, A., Watson, S., Pusch, R., Bennett, D, 2011. Bentonite – a review of key properties, processes and issues for consideration in the UK context, Quintessa Limited, QRS-1378ZG-1.1, p.32. Yang, Z.Q., Huang, L., Lu, Y., Guo, Z.J., Montavon, G., Wu., W.S., 2010. Temperature effect on U(VI) sorption onto Na-bentonite. Radiochim. Acta 98 (12), 785-791. Zhang, H.X., Tao, Z.Y., 2002. Sorption of uranyl ions on silica: Effects of contact time,pH, ionic strength, concentration and phosphate. J. Radioanal. Nucl. Chem. 254 (1), 103-107. Zhang, Z., Sun, Y., Zheng, T., Su, Q.D., 1999. Structural study of compartmental complexes of europium and copper. J. Mol. Struct. 478 (1), 23-27.

33