Polyhedron 33 (2012) 33–40

Contents lists available at SciVerse ScienceDirect

Polyhedron

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

X-ray absorption fine structure spectroscopy study of Eu(III) sorption productsonto amorphous silica and c-alumina: Effect of pH and substrate

Sumit Kumar a, Aishwarya S. Kar a, Bhupendra S. Tomar a,⇑, Dibyendu Bhattacharyya b

a Radioanalytical Chemistry Division, Bhabha Atomic Research Centre, Mumbai 400085, Indiab Applied Spectroscopy Division, Bhabha Atomic Research Centre, Mumbai 400085, India

a r t i c l e i n f o a b s t r a c t

Article history:Received 19 July 2011Accepted 2 November 2011Available online 29 November 2011

Keywords:LanthanidesTrivalent actinidesSilicaAluminaSpeciationSorptionEXAFS

0277-5387/$ - see front matter � 2011 Elsevier Ltd. Adoi:10.1016/j.poly.2011.11.009

⇑ Corresponding author. Tel.: +91 22 2559 5006; faE-mail address: bstomar@barc.gov.in (B.S. Tomar).

A molecular level understanding of metal ions sorption onto solids is necessary for modelling the sorptionprocess in a predictive manner and thus for planning of the safe disposal of nuclear wastes. In this study,we have used X-ray absorption fine structure spectroscopy (XAFS) in conjunction with batch sorption andsolubility studies to investigate the effect of pH and substrate on the formation and structure of Eu(III)sorption products on two environmentally relevant mineral oxides, namely amorphous silica and c-alu-mina. 0.1 mM Eu solution was contacted with silica and alumina over pH 4–8 in an ionic medium of 0.1 MNaClO4. Batch sorption data indicates a stronger sorption capacity of alumina in comparison to silica. Sil-ica solubility is orders of magnitude higher than that of alumina over pH 3–8. The pH (6–8) and metal ionconcentration in the XAFS samples corresponds to the undersaturated to oversaturated state with respectto Eu(OH)3 precipitation. Modelling of the EXAFS spectra indicates: (1) the formation of a Eu sorptionproduct at pH �6 in the form of a small atomic cluster wherein Eu binds to oxygen atoms on both solidsin a monodentate corner sharing and edge sharing manner, (2) alumina prefers an edge sharing modewith increasing pH over silica, (3) both solids form a surface precipitate containing Eu at higher pH val-ues, with the sorption product being richer in Eu content at the silica surface, and (4) the appearance of aEu–Eu neighbour distance of �3.5 and 4.0 Å for alumina and silica, respectively. The Eu–Eu distances forsorption samples, compares with the 3.67 and 4.09 Å Eu–Eu distances in Eu(OH)3, and thus substantiatethe edge and corner sharing preference for alumina and silica, respectively. Though Eu forms similarsorption products on silica and alumina over pH 6–8, there is a distinct difference in their formation pat-tern at the two surfaces.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction

A major factor controlling the migration of contaminants inaquatic systems is their sorption onto the solids present in theenvironment, either in the form of a fixed rock structure or colloi-dal sized particles [1,2]. Retention and thereby partitioning of themetal ions in these sorption processes proceeds through mecha-nisms involving adsorption, formation of a metal ion precipitateon the solid surface or a metal ion co-precipitate with the dissolvedpart of the solids [3–5]. As more sorption data is being generated inmacroscopic observations, the requirement of linking this informa-tion with molecular-level sorption mechanisms intensifies so thatpredictive models for the sorption processes as a function of vari-ous other system variables can be developed. These studies assumespecial significance for actinides having low solubility, e.g., Pu, Am,etc., which may enter into water bodies as a result of different nu-clear activities, such as disposal of nuclear high level waste in deep

ll rights reserved.

x: +91 22 2550 5151.

underground repositories, accidental release of actinides fromoperating power plants, underground nuclear tests and re-entryof nuclear powered satellites into earth, and can be a matter ofconcern for human health [6–9].

Surface Complexation Modelling, a mechanistic descriptiondeveloped to explain metal ion sorption at a solid–water interface,envisages a sorption reaction similar to aqueous complexationreactions with solid surface sites, for example a silanol („SiOH)site on a silica surface [10]. Assuming Langmuir isotherm behav-iour, it reproduces the sorption behaviour at lower metal ion con-centrations, that is at a lower surface coverage. At a higher surfacecoverage surface precipitation becomes the dominant mechanismand the surface speciation varies from pure metal hydroxides toa solid solution comprising of a metal ion and the central atomof the solid substrate [11]. According to Dzombak and Morel[10], surface precipitation should be considered once the dissolvedsorbate concentration exceeds one-tenth of its solubility or onehalf of the total surface site concentration.

Experimental studies on metal ion sorption establish the depen-dence of sorption on environmental variables such as pH, Eh, ionic

34 S. Kumar et al. / Polyhedron 33 (2012) 33–40

strength, temperature and presence of ligands [12–14]. Thesechemical conditions change the speciation of the metal ion andthus affect its sorption on mineral surfaces. Sverjensky reviewedthe Eu redox transformation under varying temperature and pres-sure in water ranging from surface water to reducing mediumwater, and observed significant Eu(II/III) oxidation state fraction-ation at higher temperatures and in the presence of complexing li-gands such as carbonate, sulfate and chloride [15]. In the sorptionof Eu(III) on c-alumina over different pH (4–8) under varying ionicstrength (0.001–0.1 M NaClO4) and metal ion concentration (10�9–10�5 mol/l) Rabung et al. observed the formation of an inner spherecomplex of Eu(III) on the alumina surface [16]. Further, the exis-tence of three different surface complexes, (Al–O–Cm(OH)x)(2–x)

(H2O)5–x (x = 0–2) over the studied pH was established by a TimeResolved Fluorescence Spectroscopy (TRFS) investigation of Cm(III)sorption on c-alumina under similar experimental conditions [17].In the potentiometric and modelling study of neodymium sorptionat a rutile (TiO2)–water interface [18], Ridley et al. observed a sys-tematic increase in the Nd(III) sorption with increasing tempera-ture. The experimental results were rationalized by considering aunique ‘‘tetradentate’’ configuration of Nd(III) on the rutile surface,wherein the sorbing cation bonds directly with two adjacent ‘‘ter-minal’’ and two adjacent ‘‘bridging’’ surface oxygen atoms. In themodelling study of the sorption of rare earth ions on several oxides[19], Piasecki and Sverjensky observed the proportion of mono-and tetradentate surface species varying as function of pH, ionicstrength, surface coverage and type of oxide. The results infer aweaker monodentate (relative to the tetradentate complex) com-plex of trivalent lanthanides on rutile compared with hematite orsilica. The dependence of sorption characteristics on the systemvariables results from the interfacial chemical reaction associatedwith sorption in the electric field at the oxide–water interface,which, in turn, depends on the metal ion speciation on the surface.

Direct evidence of surface speciation of metal ions on solid sur-faces has been obtained from synchrotron based X-ray absorptionfine structure (XAFS) spectroscopic studies [20–22]. Analysis ofthe X-ray absorption spectrum (extended X-ray absorption finestructure, EXAFS, and X-ray absorption near edge structure,XANES) provides structural information, including the identifica-tion, number and bond lengths of the neighbouring atoms aroundthe probe atom, and thus enables the structure of the surface spe-cies to be delineated [23]. Very few literature studies have probedtrivalent lanthanides/actinides speciation on silica and aluminasurfaces using XAFS. In an EXAFS study of Nd and Lu sorptionon silica [24], no effect of pH on the surface speciation was foundin the pH range 6–9, with the formation of an edge sharing sur-face complex of Nd/Lu on the silica tetrahedra. In another EXAFSstudy for the mechanism of Eu retention on calcium silicate hy-drates, even in the metal concentration and pH range amenableto precipitation there was no evidence of Eu(OH)3(s) on the solidsurface [25]. Fourier transforms of EXAFS spectra of Eu for bothadsorption and precipitation samples was found to exhibit com-parable structural features, pointing to similar crystallographicenvironments [25]. An EXAFS investigation of Nd- and Lu-sorbedon layered silicate, hectorite ([Mg2.77Li0.66]Si4O10(OH)2), however,shows distinct spectral dissimilarities when compared with thespectral features observed for silica [24]. In a Cm(III) sorptionstudy on c-alumina [17], an EXAFS investigation was carried outusing Gd and Lu as probe atoms, wherein the authors concludedthat a mononuclear monodentate speciation existed on thesurface, though shells beyond the first neighbour could not beexplicitly delineated in this study. Dardenne et al. [26] observedmonodentate Lu sorbed species on hematite (iron oxide/hydrox-ides are isomorphous to alumina) at pH 8, while edge sharingbidentate binding of Lu was obtained on amorphous 2-linehydrous ferric oxide above pH 5.5.

Silica (SiO2) and alumina (Al2O3) are two ubiquitous oxide min-erals in the geological environment. Apart from their geologicalimportance, these solids present two structurally different sub-strates; Si atoms are tetrahedrally surrounded by oxygen atomsin silica, whereas Al is present in both the tetrahedral as well asoctahedral sites of the close-packed arrangement of oxygen atomsin alumina. Surface site silanol („SiOH) in silica is, therefore, at-tached to a tetrahedron and the aluminol („AlOH) site in aluminais attached either to a tetrahedron or an octahedron, which maycause a significant change in the surface reactivity of these two sol-ids towards metal ions. Silica and alumina are substrates with dif-ferent dielectric constants [19] which causes a difference in thesolvation environment of the two solids. Piasecki and Sverjensky[19] explained the difference in binding characteristics of trivalentlanthanides on various mineral oxides as a result of varying solva-tion characteristics of the solids. Investigation of Am(III)/Eu(III)binding to the aluminol („AlOH) site is also important for under-standing the metal ion speciation on clay surfaces, in view of thefinding that metal ion complexation to aluminosilicate clays hap-pens primarily at the aluminol site [13].

In this work, we have examined Eu(III) sorption onto amorphoussilica (am-SiO2) and c-alumina (c-Al2O3) in near neutral pH condi-tions (6–8) with X-ray absorption fine structure spectroscopy.Eu(III) has been chosen considering it as an analogue for variouslong-lived trivalent lanthanides and actinides [12]. The absorptionspectra were collected for samples equilibrated with Eu(III) solu-tions under- and over-saturated with respect to solid Eu-hydroxidewith the aim of elucidating the role of pH and mineral substrate oninfluencing the local atomic coordination and binding of trivalentlanthanides/actinides to naturally occurring mineral oxides.

2. Experimental

2.1. Materials and characterization

Silica and alumina were purchased from M/s. Aerosil Ltd. andM/s. Degussa India Ltd., respectively, and were used as received.They were characterized for surface area by BET analysis and crys-talline phase by X-ray diffraction (XRD) analysis. The europiumused in the study was purchased from M/s. Indian Rare Earth IndiaLtd. in its oxide form. It was dissolved in conc. HNO3 under heatingand the dried residue was dissolved in 0.1 M HNO3. The Eu concen-tration in the stock solution was determined by complexometrictitration using EDTA [27]. Silica and alumina suspensions werecharacterized for Zeta potential using a Malvern Zeta Sizer nano-ZS. To study the solubility of these solids under different pH condi-tions, the same amount (3 g/l) of the two solids were equilibratedwith 0.1 N NaClO4 solution at different pH values for 24 h and sub-sequently the suspensions were centrifuged at 16500 rpm for 1 h.The required volume of the supernatants was aliquoted and as-sayed for the amount of dissolved silicon or aluminum using ICP-AES. The pH in the experiment was measured using a combinationglass electrode based pH meter (M/s. Lab India Pvt. Ltd., India).

2.2. Sorption study

One hundred micromolar solution of Eu(III) containing 154Euradiotracer was equilibrated with 1 g/l silica and alumina suspen-sions for 24 h at 25 (±1) �C. A kinetic study (at pH 5), similar to thesorption study, was carried out to find the time required for attain-ment of equilibrium in the sorption process. The study showed24 h as sufficient for attainment of the equilibrium sorption. Sus-pensions were made in the near neutral pH range (pH 6–8) rele-vant to environmental aquatic chemistry. Before equilibrationwith the metal ions, the solids were suspended in 0.1 M NaClO4

S. Kumar et al. / Polyhedron 33 (2012) 33–40 35

and left overnight for the attainment of the equilibrium of the sur-face sites with the aqueous medium. After addition of the metalion, the pH was adjusted with the addition of HClO4 or NaOHand the equilibration was carried out in a wrist action shaker. Afterequilibration, the pH of the suspensions were measured and thesuspensions were centrifuged at 16500 rpm for 1 h. One millilitreof the supernatant was counted for 154Eu radioactivity in a NaI(Tl)scintillation detector coupled to a 2048 multichannel analyzer. Theactivity of 154Eu added to the suspensions, and that left in thesupernatant after equilibration, were used to obtain the amountof Eu sorbed on the solids. Further, the amount sorbed was normal-ized with the surface area of the solid particles to obtain the sorp-tion density, denoted by the symbol C in units of lmol/m2.

-20

0

20

40

Alumina Silica

Zeta

pot

entia

l (m

V)

2.3. XAFS experiment and spectrum analysis

The XAFS sample preparation was similar to the sorption exper-iment, with the only exception of a larger (250 ml) suspension vol-ume for the XAFS samples. After filtering the solids from thesuspensions, they were vacuum dried in an oven at temperaturebelow 40 �C. A calculated amount of the dried powder was mixedwith polyvinyl pyrolidone to make a pellet, which gave an edgestep P0.1 in the Eu–LIII X-ray absorption spectrum.

XAFS data was collected using the XAFS beamline of an Elettrasynchrotron facility with the storage ring operating at 2.0 GeV and300 mA. Measurements were carried out in the transmission modeusing three ionization chambers, one each for initial X-ray intensity(I0), transmitted intensity (I) through the sample and intensity (Im)after the monitor foil (Fe) placed downstream of the second ioniza-tion chamber for simultaneous energy calibration. Harmonics of thebeam were rejected by detuning the second monochromator crystalby 45%. The spectra were collected over the energy range 6700–7600 eV, focusing on the Eu–LIII edge energy (6977 eV), with a pre-edge energy step of 5 eV, 0.20 eV in the edge region and thereaftera constant k step of 0.03 Å�1 in the extended XAFS region. Dataacquisition was limited to 7600 eV due to the Eu–LII edge just abovethis energy. Typically three scans were taken for each sample andthe data could be obtained in the photoelectron momentum spaceextending up to 10 Å�1. To aid in the spectrum analysis, the XAFSspectra for three Eu reference samples, (1) hydrated Eu ions, Eu(aq),(2) europium hydroxide, Eu(OH)3, and (3) europium oxide, Eu2O3,were also acquired.

The Eu(LIII) edge X-ray absorption spectra of the reference com-pounds and the samples were analyzed by the software packageIFEFFIT. A data reduction step was carried out using the standardprocedure of background removal and normalization in Athena,while phase-shift and amplitude functions were obtained usingthe theoretical multiple scattering calculations of FEFF version6.01. A final fitting of the spectra was done on the normalized,background subtracted v(R) data using all shells simultaneouslyin Artemis after fitting the spectra shell-wise either in R space orafter back transforming to k space. For the standards, Eu2O3 andEu(OH)3, non-linear least square fitting was performed usingtheir crystallographic data, while the paths used in the fitting ofthe sorption samples were created using dysprosium disilicate(Dy2Si2O7) [28], with Dy replaced by Eu and europium aluminumgarnet (Eu3Al5O12) [29]. The triclinic type B structure of Dy2Si2O7

is adopted by the disilicates of neighbouring rare earth ions [28].

3 4 5 6 7 8 9 10 11

-40

pH

Fig. 1. Zeta potential plot for silica and alumina in 0.01 M NaClO4 over pH 3–10.

3. Results and discussion

3.1. Substrates characterization

The XRD pattern for silica lacks any peak-like feature that canbe attributed to the amorphous nature of the particles. Alumina

shows characteristic peaks for a c crystallographic phase in itsXRD pattern. The specific surface areas obtained by the BET meth-od (N2 absorption) were found to be 180 ± 5 and 202 ± 4 m2/g forsilica and alumina, respectively. Zeta potential measurements(Fig. 1) show that silica is negatively charged above pH 3, indicat-ing its point of zero charge (PZC) to be less than pH 2. The aluminaPZC was found to be �8.9 (Fig. 1). In the neutral pH range silica isthus negatively charged while alumina is positively charged. If onlythe electrostatic interaction is assumed to be responsible for metalion sorption, a negatively charged silica surface should favour thesorption process. The solubility of silica in 0.1 M NaClO4 was foundto be significantly higher than that of alumina (Fig. 2). At lower pHthe solubility of silica is an order of magnitude higher than that ofalumina while in the pH range 6–8 it is more than two orders high-er. Modelling of the solubility behaviour of silica and alumina [30]indicates that the neutral species Si(OH)4(aq) and Al(OH)3(aq) gov-ern the solubility in the neutral pH range.

3.2. Batch sorption experiment

Fig. 3 shows the data on the sorption of Eu(III) onto silica andalumina as a function of pH. Eu(III) sorption on silica and aluminarises sharply over the pH 6–7 range, starting from �5% at pH 4 toquantitative sorption at pH 7. The sorption profile is sharper in thecase of alumina in comparison to silica. This agrees with the liter-ature report suggesting a higher sorption ability of alumina [21].The sorption process may follow different mechanisms rangingfrom adsorption to surface precipitation. A preliminary assessmentabout the possible mode of Eu(III) binding on silica and aluminaindicates Eu binding on silica/alumina in a monodentate and/orbidentate manner in case of surface adsorption [31,32], while theprecipitation of Eu with hydroxyl ions, dissolved silicates and alu-minates may occur in the case of surface precipitation. Fig. 4 showsthe plot of log([Eu]disso) versus pH for the prepared sorption sam-ples for the XAFS study along with the calculated solubility curvefor Eu(OH)3. The dotted line represents the initial Eu(III) concentra-tion (10�4 M) and the symbols are for the final concentration ofEu(III) left in the supernatant after equilibration of the XAFS sorp-tion samples. Table 1 lists the Eu sorption density on silica and alu-mina under the studied pH conditions. A europium hydroxidesolubility calculation was carried out at a fixed ionic strength of0.1 M NaClO4 using the formation constant given by Yun et al.[33] for different hydroxide species. Eu(III) solubility in the case

4 6 8 10 12 14-12

-8

-4

0

4

8

SilicaAlumina

log[Eu(III) disso.]

pH

Fig. 4. Calculated Eu(OH)3 solubility curve as a function of pH. The different speciescontributing to the solubility of europium hydroxide include, Eu3+, Eu(OH)2+,Eu(OH)2

+, Eu(OH)3(aq) and Eu(OH)4�.

Table 1Eu(III) sorption on silica and alumina in XAFS samples.

pH % Sorption C(Eu sorbed) (lmol/m2)

am-Silica 6.08 26.4 0.1447.05 91.2 0.5068.02 96.4 0.536

c-Alumina 6.18 42.8 0.2087.15 95.7 0.4728.17 97.4 0.480

3 4 5 6 7 81E-6

1E-5

1E-4

1E-3

Alumina Silica

Al/S

i Con

c.(M

)

pH

Fig. 2. Solubility of silica and alumina measured over pH 3–8, (3 g/L) in 0.1 MNaClO4 medium.

4 5 6 7 8 90

10

20

30

40

50

60

70

80

90

100

AluminaSilica

% S

orpt

ion

pH

Fig. 3. Eu(III) sorption onto am-SiO2 and c-Al2O3 as a function of pH.

36 S. Kumar et al. / Polyhedron 33 (2012) 33–40

of both solids follows the same trend. At pH �6 it is in an under-saturated state, falling two orders of magnitude below the solubil-ity limit. It falls just below the solubility limit at pH �7, while it isin an over-saturated state at pH �8. However, the possibility ofprecipitation of Eu-silicates or Eu-aluminates in the sorption sam-ples cannot be ruled out, as they have limited solubility [34] andsilicate formation has been observed in various sorption studieson silica at a higher surface coverage [35].

3.3. XAFS spectroscopy

3.3.1. XANES spectroscopyXANES spectra of all the reference solids and sorption samples

fall at the same positions (Fig. 5). The position of the X-ray absorp-tion edges at �6982 eV compares well with the edge for trivalentEu in solids and it is located at significantly higher energy thanthe maximum of X-ray absorption for divalent Eu (�6973 eV)[36]. This indicates the existence of the +3 oxidation state for Euin all the sorption and reference samples.

3.3.2. EXAFS spectroscopyEXAFS spectra of the reference samples (Fig. 6a) show increas-

ing structural complexity from Eu(aq) to Eu2O3. The Eu(aq) spec-trum has a monotonically decreasing beat pattern beyondk = 3.0 �1, characteristic of the presence of a single neighbouringshell. The oscillating patterns in the k-space spectra of the twoother reference samples clearly indicate the presence of manyshells. Table 2 lists the required shells to model the EXAFS spectraof these reference samples along with the quantitative estimatesfor the parameters of each shell, found in the fitting of the Fouriertransform of the EXAFS spectra (Fig. 6b).

3.4. EXAFS spectra of Eu(III) sorbed onto silica and alumina

The EXAFS spectra of Eu(III) sorbed onto silica and alumina(Fig. 7a) appear similar in beat pattern to the spectra of Eu(aq)and Eu(OH)3, though they are not exactly the same. Below6.3 Å�1, at which the maximum in the oscillation pattern was ob-served, all the sorption samples, Eu(aq) and Eu(OH)3 have similarfeatures, exhibiting Eu–O as the shell responsible for this feature.Oscillations in the 9–10 Å�1 range for the sorption samples appearat higher k than for the case of Eu(aq), and are different from thatfor Eu(OH)3 which shows a new oscillation at �9 Å�1. The differ-ence in the EXAFS spectra of the sorption samples from that ofEu(OH)3, in the k range 9–10 A�1 suggests the absence of surfaceprecipitation of Eu(III) in the former at all pH values, despite theover-saturation conditions at pH 8.

Fig. 7a and b show the EXAFS data and their Fourier transforma-tion into R-space. The solid lines in Fig. 7b represent the modelled

6960 6980 7000 7020 70400

1

2

3

4

5

6

7

8

(8)(9)

(7)(6)

(5)

(4)

(3)

(2)

Nor

mal

ised

χ( E

)

Energy (eV)

(1)

Fig. 5. XANES spectra of the Eu(III) reference and sorption samples (1) Eu (aqua),(2) Eu(OH)3 (solid), (3) Eu2O3 (solid),(4) Eu-silica-pH 6.08, (5) Eu-silica-pH 8.02, (6)Eu-silica-pH 7.05, (7) Eu-alumina-pH 6.18, (8) Eu-alumina-pH 7.10, (9) Eu-alumina-pH 8.17.

S. Kumar et al. / Polyhedron 33 (2012) 33–40 37

data. Fig. 7c represents the back-Fourier transformed data ofFig. 7b over the R + DR window 1.2–4.4 Å. The first shell in thespectra, appearing at R + DR � 1.8 Å, compares well with theEu–O shell of the Eu-reference samples, confirming Eu-O as thefirst shell of the Eu-sorption product on the surfaces of silica andalumina. The next prominent peak in the sorption samples is at�2.8 Å (phase uncorrected), which is absent in Eu(aq), and thismatches with a peak in the Eu(OH)3 and Eu2O3 spectra. This peakis due to the Eu–Eu shell in the reference samples, but it was fittedonly with the Eu–Si/Al shells in the sorption samples (Table 3).Similarly the peak in the sorption samples at �3.7 Å, thoughmatching with Eu(OH)3, could be fitted only considering theEu–Si/Al shells.

2 4 6 8 10 12

-5

0

5

k3 *χ(

k)

K (A0)-1

-5

0

5

Eu2O3

Eu(OH)3

-5

0

5 Euaq

(a)

Fig. 6. (a) k3 Weighted Eu–LIII EXAFS spectra of Eu(III) reference samples. (b) Experimentreference samples.

3.4.1. First shell fittingThe FT peak at 1.8 Å for all the sorption samples was modelled

by assuming a single oxygen shell in R-space over 1.2–2.5 Å. Thefits yielded a NO value of 7.10 ± 0.37, and REu–O distances2.41 ± 0.01 Å (Table 3). The REu–O distance is longer thanREu–O = 2.34 Å for hexadentate Eu in Eu2O3, and is marginally short-er than REu–O = 2.42 Å for 9-cordinated Eu in water (Eu(aq)). Thefirst shell coordination for lanthanides decreases from 9 to 8 aswe move from La to Lu, and in a recent study it has been shownthat Eu(III) actually exists as a mixture of both 8- and 9-fold coor-dinated in Eu(III) aqua ions [37]. Taking into consideration theuncertainty associated with the analysis for the first shell (10%for the N value), this means Eu is hepta to octa coordinated inthe sorption products. As expected, REu–O = 2.41 Å and the NO valuecompares well with the coordination of Eu substituted on acalcium site in calcium silicate hydrate [25] and Gd sorbed on c-alumina [17]. The similarity of this shell for both the substrateand all the pH values indicates the similarity of the primary inter-action mechanism of Eu in these two substrates.

3.4.2. Higher shell fitting (Eu–Si/Al)FT features over 2.5–4.2 Å indicate mainly two peaks. However,

back transformation of higher pH spectra presents a beat patternmaximizing at three k values (�3, 6.5, and 9 �1) with the beat pat-tern intensity at �9 �1 substantially subdued in the case of alu-mina compared to silica. This suggests a complex structuralarrangement in the sorption products. Due to this complexity,the approach used here was to propose different structurally rea-sonable silicate/aluminate clusters around the sorbed Eu as the ini-tial fit model and then to adjust the parameters to get a goodmatch with the back transformed k data. Evaluation of Eu(OH)3

as the fitting model produced no sensible sets of parameters forthe best fit case. Inclusion of a Eu shell was found necessary forbetter modelling of the EXAFS spectra at higher pH values, espe-cially to model the small peak-like feature at �3.5 Å and the peakat �4 Å.

The best fit model was found with two Si/Al shells at �3.22 and4.01 Å and an Eu backscattering shell at 4.00 Å for silica (3.45–3.53 Å for alumina) at higher pH values (7–8) (Table 3). NSi forthe Si shell at �3.22 Å decreases with pH, while NAl for the Al shellat 3.22 Å increases with pH. Considering the Eu–Al/Si distances inall possible coordination geometries on the alumina/silica surface(see Appendix), the �3.22 Å shell relates to Eu bonded to Si or Al

0

4

8

|χ( R

)|

Euaq

0

4

8Eu(OH)3

1 2 3 4 5 60

4

8Eu2O3

R (A0)

(b)

al (open symbol) and simulated (solid line) moduli of the Fourier transforms for the

k3 .χ

(k)

k (A0)-1

1

2

3

4

5

6

(a)

2 3 4 5 6 7 8 9 10 11 1 2 3 4 5 60

5

R (A0)

|χ(R

)|

(b)

0

5

0

5

0

20

5

0

5

42 3 5 6 7 8 9 10 11 12

-505

K (A0)-1

-505

-505

-303

-505

-505(C)

Fig. 7. (a) k3 weighted Eu–LIII EXAFS spectra of Eu(III) sorption samples. Numbers 1–3 stand for Eu sorption on alumina at pH 6–8, respectively and spectra numbered 4–6 arefor Eu sorbed onto silica at pH 6–8. (b) FT magnitude for these EXAFS spectra. The black open symbol is for the experimental data while the solid line is the fitting line. (C)Experimental (open symbol) and modelled (solid line) Fourier filtered k3v(k) contributions for the next-neighbours backscattering shells in R + DR range 1.2–4.2 Å.

Table 2Structural parameters from EXAFS analysis of Eu(III) reference solids.

Reference Path R N r2 DE (eV) Rf

Eu(aq) Eu–O 2.420 (0.002) 9.0a 0.008(0.001) 0.9 0.03

Eu(OH)3(s) Eu–O 2.432(0.007) 7.7(0.6) 0.010(0.001) 3.1 0.01Eu–Eu 3.672(0.004) 1.7(0.6) 0.007(0.001) 3.4Eu–O–O 3.918(0.004) 30.6(0.6) 0.013(0.008)Eu–O 3.944(0.003) 2.5(0.6) 0.013(0.008)Eu–Eu 4.090(0.006) 5.1(0.6) 0.014(0.002) 0.40Eu–Eu–O 3.993(0.007) 10.3(0.6) 0.010(0.005)Eu–O 4.114(0.007) 5.1(0.6) 0.010(0.005)

Eu2O3(s) Eu–O 2.340(0.014) 6.0a 0.011(0.001) 0.72 0.11Eu–Eu 3.601(0.018) 6.0(0.1) 0.009 0.71Eu–Eu 4.071(0.001) 6.0(0.1) 0.010(0.005)Eu–O 4.235(0.004) 6.0(0.1) 0.010(0.005)Eu–Eu–O 4.334(0.001) 12.0(0.1) 0.010(0.005)

R, N, r2, DE and Rf are the symbols used for the parameters near-neighbour distance, number of near-neighbour atoms, variable making Debye–waller factor, energy shift andresidual factor (¼

Pkðk

3vexp � k3vcalcÞ=P

kðk3vexpÞ), respectively, in the EXAFS equation. The residual factor measures the quality of the fitted model with respect to the

experimental data. Data in parentheses show the estimated standard deviation on the main value.a Values are fixed during the fitting.

38 S. Kumar et al. / Polyhedron 33 (2012) 33–40

Table 3Results of the quantitative analysis of EXAFS spectra of Eu(III) sorption samples.

Sample/pH Eu–O Eu–Si/Al Eu–Eu Eu–Si/Al

N R r2 N R r2 N R r2 N R r2

Silica/6.08 7.14(0.10)

2.41(0.03)

0.008(0.005)

2.22(0.10)

3.24(0.04)

0.011(0.007)

4.05(0.01)

4.01(0.04)

0.007(0.001)

Silica/7.05 6.86(0.13)

2.41(0.03)

0.011(0.002)

2.05(0.05)

3.27(0.02)

0.008(0.010)

1.55(0.13)

4.02(0.11)

0.02(0.008)

3.75(0.05)

4.01(0.01)

0.014(0.003)

Silica/8.02 6.85(0.5)

2.40(0.005)

0.009(0.001)

1.22(0.1)

3.19(0.02)

0.008(0.004)

3.66(0.02)

4.00(0.04)

0.015(0.008)

2.18(0.13)

4.02(0.02)

0.007(0.003)

Alumina/6.18 7.69(0.6)

2.39(0.030)

0.010(0.005)

2.11(0.08)

3.23(0.07)

0.015(0.009)

3.4(0.04)

4.07(0.04)

0.006(0.001)

Alumina/7.15 7.32(0.14)

2.41(0.01)

0.009(0.001)

2.42(0.36)

3.22(0.08)

0.018(0.008)

0.9(0.03)

3.45a 0.018(0.01)

3.2(0.07)

4.03(0.02)

0.011(0.003)

Alumina/8.17 6.71(0.2)

2.41(0.10)

0.009(0.001)

3.63(0.09)

3.34(0.11)

0.017(0.002)

2.5(0.2)

3.53(0.04)

0.018(0.008)

2.37(0.1)

4.02(0.08)

0.009(0.001)

Values given in the parentheses, below the main value, are the estimated standard deviation on the main value obtained in the least square fitting routine.a Values kept fixed in the fitting.

S. Kumar et al. / Polyhedron 33 (2012) 33–40 39

in an edge sharing manner and the �4 Å shell corresponds to cor-ner shared binding of Eu to Si/Al polyhedra. The relative trend forthe number of coordinating Si/Al atoms at the �3.22 Å shell thussuggests that while the corner sharing mode of interaction is morepreferred for silica, Eu binds to the alumina surface predominantlyin an edge sharing manner. A similar conclusion was observedin the XAFS study of Cu(II) sorption on amorphous silica andc-alumina at a lower surface coverage [38].

3.4.3. Higher shell fitting (Eu–Eu)For the Eu shell, NEu increases with pH for both silica and alu-

mina, with the enhancement being more significant for silica thanfor alumina. The higher NEu value for silica can be attributed to itsweaker sorption capacity (compared to alumina) and higher disso-lution tendency, which causes surface precipitation of Eu with thedissolved silica.

The high NSi/Al value at the 4 Å shell for the sorption samples atpH 6 indicates the formation of some small multinuclear atomicclusters. These atomic clusters consist of Eu bound in a monoden-tate manner to many oxygen atoms present at the corner of silicaor alumina polyhedra. At higher pH values, many such atomic clus-ters may come together forming Eu–Eu bonds, thereby resulting ina bigger multinuclear complex.

Multinuclear sorption product formation on oxide surfaces, athigher surface coverage, has been observed as the norm ratherthan the exception [3–5,11,22]. Eu forming multinuclear surfacespecies on silica surfaces at higher surface coverages has been re-ported in studies involving batch sorption [39] and fluorescenceinvestigations [40]. Very few studies have examined the sorptionproducts of trivalent lanthanides/actinides on alumina surfaces athigher surface coverages. At lower surface coverages, the rare earthions (Lu and Gd) exhibit monodentate binding on c-alumina [17],while fluorescence and EXAFS based characterization of sorbedlutetium species onto synthetic hydrous ferric oxide at pH P5.5indicate an edge shared binding of the surface complex formation.Fe-oxides/hydroxides are isomorphous to aluminium oxides/hydroxides and sorption products ranging from surface adsorptionto surface precipitation have been observed with increasing Zn(II)sorption density on hematite [3].

4. Conclusion

The present study investigates the Eu sorption product on silicaand alumina surfaces under varying pH conditions, amenable tounder-saturated to over-saturated with respect to Eu(OH)3 forma-tion. The surface charge characteristics and solubility of the twosolids are quite different. XANES spectra indicate the presence of

Eu in the +3 oxidation state for all sorption products. Quantitativeresults for the modelling of the EXAFS data indicate a similar bind-ing geometry of Eu on both amorphous silica and c-alumina sur-faces. Eu binds in an edge as well as a corner sharing mode forboth the solids, though the formation pattern of these two typesof binding modes differs. The silica surface prefers a corner sharingbinding mechanism, whilst edge sharing binding is more prefera-ble in the sorption product on the alumina surface. At pH �6, Euforms sorption species involving corner sharing and edge sharingmonodentate binding, whereas EXAFS spectra at higher pH valuesindicate the formation of atomic clusters containing Eu. At higherpH values, the number of coordinating Eu atoms is more for silicathan alumina, indicating the effect of the sorption capacity and dis-solution characteristics on the sorption product. Our results dem-onstrate that variations in the characteristics of the solids affectsthe mode of Eu sorption, and pH plays an important role in the for-mation of different sorption products.

Acknowledgements

Authors acknowledge the Elettra Synchrotron Facility and Inter-national Centre for theoretical Physics (ICTP) for beam time alloca-tion and a travel grant, under the ICTP-Elettra Users Programme(Proposal No. 20090024). We thank Drs. Giuliana Aquilanti andAndrea Cognini for support during the XAFS measurements.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.poly.2011.11.009.

References

[1] A.P. Novikov, S.N. Kalmykov, S. Utsunomiya, R.C. Ewing, F. Horreard, A.Merkulov, S.B. Clark, V.V. Tkachev, B.F. Myasoedov, Science 314 (2006) 638.

[2] DOE, US Department of Energy report on the workshop: Basic Research NeedsFor geoscience: Facilitating 21st Century Energy System, 2007. Available from:<http://www.sc.doe.gov/bes/reports/list.html> (accessed 12 July 2011).

[3] J. Ha, T.P. Trainor, F. Farges, G.E. Brown Jr., Langmuir 25 (2009) 5574.[4] P.T. Trainor, G.E. Brown Jr., G.A. Parks, J. Colloid Interface Sci. 231 (2000) 359.[5] C. Charlet, A.A. Manceau, J. Colloid Interface Sci. 148 (1992) 443.[6] A.B. Kersting, D.W. Efurd, D.L. Finnegan, D.J. Rokop, D.K. Smith, J.L. Thompson,

Nature 397 (1999) 56.[7] J.I. Kim, Nucl. Eng. Technol. 38 (2006) 459.[8] R.C. Ewing, Elements (2006) 331.[9] J.S. Contardi, D.R. Turner, T.M. Ahn, J. Contam. Hydrol. 47 (2001) 323.

[10] D.A. Dzombak, F.M.M. Morel, Surface Complexation Modelling: Hydrous FerricOxide, John Wiley & Sons, Inc. USA, 1990.

[11] S.N. Towle, J.R. Barger, G.E. Brown Jr., G.A. Parks, J. Colloid Interface Sci. 187(1997) 62.

40 S. Kumar et al. / Polyhedron 33 (2012) 33–40

[12] Th. Rubung, Th. Stumpf, H. Geckeis, R. Klenze, J.I. Kim, Radiochim. Acta 88(2000) 711.

[13] E. Tertre, G. Berger, E. Simoni, S. Castet, E. Giffaut, M. Loubet, H. Catalette,Geochim. Cosmochim. Acta 70 (2006) 4563.

[14] C. Alliot, L. Bion, F. Mercier, P. Vitorge, P. Toulhoat, Radiochim. Acta 93 (2005)435.

[15] D.A. Sverjensky, Earth Planet. Sci. Lett. 67 (1984) 70.[16] T. Rabung, T. Stumpf, H. Geckeis, R. Klenze, J.I. Kim, Radiochim. Acta 88 (2000)

711.[17] Th. Rabung, H. Geckeis, X.K. Wang, J. Rothe, M.A. Denecke, R. Klenze, Th.

Fanghänel, Radiochim. Acta 94 (2006) 609.[18] M.K. Ridley, M.L. Machesky, D.J. Wesolowski, D.A. Palmer, Geochim.

Cosmochim. Acta 69 (2005) 63.[19] W. Piasecki, D.A. Sverjensky, Geochim. Cosmochim. Acta 72 (2008) 3964.[20] M.A. Denecke, Coord. Chem. Rev. 250 (2006) 730.[21] A. Kowal-Fouchard, R. Drot, E. Simoni, N. Marmier, F. Fromage, J. Ehrhardt, New

J. Chem. 28 (2004) 864.[22] C.J. Chishlom-Brause, P.A. O’Day, G.E. Brown Jr., G.A. Parks, Nature 348 (1990)

528.[23] G. Bunker, Introduction to XAFS: A Practical Guide to X-ray Absorption

Fine Structure Spectroscopy, Cambridge University Press, Cambridge, UK,2010.

[24] M.L. Schlegel, I. Pointeau, C.E.A. de Saclay, Report available from: <http://ftp.esrf.eu/pub/UserReports/25930_A.pdf> (accessed 12 July 2011).

[25] M.L. Schlegel, I. Pointeau, N. Coreau, P. Reiller, Environ. Sci. Technol. 38 (2004)4423.

[26] K. Dardenne, T. Schäfer, M.A. Denecke, R. Rothe, J.I. Kim, Radiochim. Acta 89(2001) 469.

[27] J.C. Bailar, H.J. Emelius, R. Nyholm, A.F. Trotman-Dickenson, ComprehensiveInorganic Chemistry, vol. 5, Pergamon Prees, Oxford, 1973.

[28] M.E. Fleet, X. Liu, Acta Cryst. B 56 (2000) 940.[29] E. Garskaite, S. Sakirzanovas, A. Kareiva, J. Glaser, H.J. Meyer, Z. Anorg. Allg.

Chem. 633 (2007) 990.[30] W. Stumn, J.J. Morgan, Aquatic Chemistry: Chemical Equilibria and Rates in

Natural Waters, John Wiley & Sons, Inc. USA, 1996.[31] N. Marmeir, A. Delises, F. Fromage, J. Colloid Int. Sci. 212 (1999) 228.[32] Th. Rubung, D. Schild, H. Geckeis, R. Klenze, Th. Fanghänel, J. Phys. Chem. B

1088 (2004) 17160.[33] Jong-II Yun, T. Bundschuh, V. Neck, J.I. Kim, Appl. Spectrosc. 55 (2001) 273.[34] I.A. Bondar, N.A. Toropov, Materials Research Bulletin, vol. 2, Pergamon Press,

Inc. USA, 1967. p. 479.[35] P.A. O’Day, C.J. Chisholm-Brause, S.N. Towle, G.A. Parks, G.E. Brown Jr.,

Geochim. Cosmochim. Acta 60 (1996) 2515.[36] T. Yoshida, T. Tanaka, S. Yoshida, S. Hikita, T. Baba, T. Hinode, Y. Ono, Appl. Surf.

Sci. 156 (2000) 65.[37] B. Marmidee, K. Jahn, F. Ariese, C. Gooijer, M.U. Kumke, J. Phys. Chem. A 114

(2010) 13050.[38] S. Cheah, G.E. Brown Jr., G.A. Parks, J. Colloid Interface Sci. 208 (1998) 110.[39] Y. Takahashi, T. Kimura, Y. Kato, Y. Minai, T. Tominaga, Radiochim. Acta 82

(1998) 227.[40] A.S. Kar, B.S. Tomar, S.V. Godbole, V.K. Manchanda, Colloid Surf. A:

Physicochem. Eng. Aspects 378 (2011) 44.