Ab Initio X-ray Absorption Spectroscopy Study of the Solvation

ARTICLE pubs.acs.org/JPCA

Ab Initio X-ray Absorption Spectroscopy Study of the Solvation Structure of Th(IV), U(IV), and Np(IV) in Aqueous Solution Jes
us Chaboy*,†,§ and Sofía Díaz-Moreno‡ †

Instituto de Ciencia de Materiales de Arag
on, Consejo Superior de Investigaciones Científicas, CSIC-Universidad de Zaragoza, 50009 Zaragoza, Spain ‡ Diamond Light Source Ltd., Harwell Science and Innovation Campus, Diamond House, Chilton, Didcot, Oxfordshire, OX11 0DE, U.K. § Departamento de Física de la Materia Condensada, Universidad de Zaragoza, 50009 Zaragoza, Spain

bS Supporting Information ABSTRACT: The coordination structures of U(IV), Np(IV), and Th(IV) in aqueous solution have been determined by studying the X-ray absorption near edge structure (XANES) of the actinide (An) L3-edge absorption spectra. The high sensitivity of XANES to the bonding geometry provides an unambiguous determination of the coordination polyhedron. On the basis of the comparison of ab initio computations with the experimental data we conclude that the hydration sphere of the three An(IV) aqua-ions studied is best modeled by 9 water molecules forming a tricapped trigonal prism.

1. INTRODUCTION In recent years we have witnessed an increased level of research into actinide speciation. The driving force for this activity is mainly the determination of the environmental impact of the geological disposal of radioactive wastes.1 The determination of the coordination properties of the actinide ions in solution is a basic step for any safety assessment of a proposed repository, as this characteristic is closely related to the ion’s chemical properties, e.g., redox processes, ligand exchange reactions, transport properties, reaction mechanisms, hydrolysis, etc. In particular, the study of the hydration structure of the actinide ion is essential since water is the most widely available solvent in nature. The hydration structure of actinide ions has been studied by a broad range of techniques. X-ray and neutron scattering and extended X-ray absorption fine structure spectroscopy (EXAFS) in particular have been widely used. However, the results obtained are often controversial. Although some consensus exists regarding the interatomic distances between the actinide ion and their first neighboring atoms, there is significant dispersion in the data relating to the ion coordination numbers (CNs). The case of the hydration structure of the U(IV) ion is a clear illustration of this controversy. While X-ray scattering studies2,3 point to a polyhedron formed by eight water molecules around the central cation as the hydration structure, earlier EXAFS investigations indicate a coordination number of 10.8 ( 0.5.4 More recent EXAFS studies suggest that the U(IV) coordination number is 8.7,5 a value closer to that reported by X-ray scattering. In addition, Ikeda et al.,6 using the same technique, report a mixed coordination number around the U(IV) ion, with some absorbers surrounded by nine water molecules and some by ten, r 2011 American Chemical Society

providing an averaged coordination number of 9.5. Theoretical calculations performed in the same system have also added to the controversy. On the basis of the bond valence sum calculations6,7 it has been suggested that the obtained distance of RU(IV)-O(H2O) = 2.40 Å is statistically appropriate for a CN of 8.88, close to 9, in agreement with relativistic density functional theory (DFT) results for U(IV) in aqueous solutions.9 Similar dispersion of results exists in the literature concerning the coordination of Th(IV)2-4,10-12 and Np(IV)1,13-18 in aqueous solutions. It has been postulated that this uncertainty in the determined CN is associated with the intrinsic limitations of the techniques employed. Indeed, both the EXAFS and X-ray scattering methods can give the interatomic distances of the dominating interactions with high accuracy, in particular the first-shell metal-oxygen bond distances. However, their accuracy in the direct determination of the coordination number is fairly low. In the case of EXAFS the coordination number is directly proportional to the amplitude reduction factor, S20, and also strongly correlated to the Debye-Waller factor. As a consequence, EXAFS always suffers from relatively large errors when determining the coordination numbers, ranging from 10% to 25%.15,19 In addition, several works have pointed out the existence of doubleelectron excitations in the L3-edge X-ray absorption spectra of actinides which may affect the structural determination by EXAFS.17,20 This problem has been largely discussed in the case of the absorption at the L-edges of the lanthanides21,22 or at the K-edge of their isoelectronic yttrium atom.23 Received: November 4, 2010 Revised: December 20, 2010 Published: February 28, 2011 2345

dx.doi.org/10.1021/jp1105482 | J. Phys. Chem. A 2011, 115, 2345–2349

The Journal of Physical Chemistry A The precise determination of the coordination polyhedra around the actinide ions in aqueous solution is as challenging as the determination of the solvation structure of the yttrium(III) cation in different solvents including water. In the yttrium(III) case, the authors undertook an earlier study of the X-ray absorption near edge structure (XANES) region of the spectra providing an unambiguous answer to this problem.23,24 The study demonstrated that XANES is extremely sensitive to the stereochemical details of the absorbing site, i.e., overall symmetry, interatomic distances, and bond angles, offering better capabilities than a direct EXAFS analysis for the determination of the coordination polyhedron around the absorbing atom. Consequently, we have applied a similar methodology to the study of the An L3-edge XANES spectra of An(IV) ions in aqueous solution. The calculations have been carried out for eight different coordination polyhedra involving coordination numbers ranging from 8 to 10. The results of the calculations have been compared to the experimental data previously reported in the literature. Our results indicate that the three cations studied, U(IV), Np(IV), and Th(IV), present a very similar hydration sphere, and this is best modeled by nine water molecules forming a tricapped trigonal prism.

2. COMPUTATIONAL METHODS The computation of the An L3-edge XANES spectra was carried out by using the multiple-scattering code CONTINUUM.25 A complete discussion of the procedure can be found elsewhere.26 The potential for the different atomic clusters was approximated by a set of spherically averaged muffin-tin (MT) potentials built by following the standard Mattheis’ prescription. The muffin-tin radii were determined following the Norman criterion and by using a 1% overlap factor. Special attention has been given to the choice of the exchange and correlation parts of the final state potential.27 We have tested the performance of the energy dependent Hedin-Lundqvist (HL) and Dirac-Hara (DH) exchange and correlation potentials (ECP), and have also made computations by using only the real part of the HL ECP. As shown in Figure S1 of the Supporting Information, the best agreement between the experimental and theoretical XANES spectra is obtained by using the DH potential. The DH calculations reproduce all the spectral features, their relative energy separation, as well as their relative intensity ratio, while in the case of the HL ECP the calculated absorption maxima underestimate the experimental spectra. We have found that single-channel multiple-scattering calculations are able to reproduce the experimental L3-edge spectra of An(IV) in aqueous solution in contrast to other cases of aqua ions in which two absorption channels are needed to account for a proper description of the final state during the photoabsorption process.28 It should be highlighted that no free parameter has been used in these calculations. The theoretically calculated spectra have been directly compared to the experimental XANES, i.e., no fitting procedure has been used. The assessment of the quality of the theoretical computations is based on the correct reproduction of the shape and energy position of the different spectral features as well as on their relative energy separation and the intensity ratio. In all the cases, the theoretical spectra have been convoluted with a Lorentzian line shape function to account for the core-hole lifetime and the experimental resolution. The Γ used in the

ARTICLE

Figure 1. Comparison of the An L3-edge XANES spectra of Th(IV) (red, —), U(IV) (blue, O), and Np(IV) (b) adapted from ref 17.

Table 1. Coordination Numbers and Interatomic Distances of the Different Coordination Polyhedra9 Used for the Computation of the An L3-edge XANES Spectra model

Ncoord

rU-O

symmetry

M1

8

2
2.38

D2h

6
2.56 M2

8

8
2.48

D4h

M3

8

8
2.47

D4d

M4

8

4
2.43

D2d

9

4
2.47 1
2.42

C2v

M5

4
2.52 4
2.56 M6

9

6
2.47

D3h

3
2.54 M7

10

2
2.52

C2v

8
2.56 M8

10

2
2.53 8
2.75

D2d

computations is 7.1, 7.4, and 7.6 eV29 for the cases of Th, U, and Np, respectively.

3. RESULTS AND DISCUSSION The experimental L3-edge absorption spectra of U(IV), Np(IV), and Th(IV) in aqueous solution, adapted from refs 15,17, and 30, are reported in Figure 1. The three spectra shown in the figure are very similar. Indeed, all of them show the same spectral shape with an intense white-line and a broad energy dip centered at ∼24 eV. This is followed by a broad positive maximum at ∼40 eV above the edge. Although on aligning the spectra to the energy of the maximum of the white line slight differences can be appreciated, the similarity of the overall spectral shape strongly suggests that the three aqua-ions present the same coordination geometry. This result disagrees with previous works in which the hydration structure found for Th(IV) is proposed to be different from those of U(IV) and Np(IV).30 However, it should be noted that these conclusions were based on a fingerprint analysis of the 2346

dx.doi.org/10.1021/jp1105482 |J. Phys. Chem. A 2011, 115, 2345–2349

The Journal of Physical Chemistry A

ARTICLE

Figure 2. Comparison of the experimental U L3-edge XANES spectrum of U(IV) in aqueous solution [adapted from refs 17 (b) and 30 (—)] and the theoretical spectra calculated for four different 8-coordinated models: M1 (red, O), M2 (green, —), M3 (blue, 3 3 3 ), and M4 (black, - 3 -). The theoretical spectra have been aligned to the maximum of the white line and vertically shifted for the sake of clearness. (See text for details.)

EXAFS spectra. The proposed smaller coordination numbers reported for U(IV) when compared with Th(IV) were solely based on a qualitative comparison of their EXAFS amplitudes. However, it is well-known that the amplitude of the EXAFS signals is not determined only by the coordination numbers, and as previously discussed, other factors are also known to affect it. If the CNs of the An(IV) aqua-ions are different (8-, 9- or 10coordinated), different coordination polyhedra should be expected, involving the modification of both the interatomic distances and bond angles between the ions and the hydrating water molecules. Accordingly, the occurrence of different coordination polyhedra should be directly reflected in the spectral shape of the XANES. Here we compare the experimental An L3-edge XANES spectra to ab initio theoretical computations performed for different coordination models. Following the work by Tsushima and Yang9 we have considered the following 8-coordinated models: a bicapped hexagonal prism (M1), regular cube (M2), square antiprism (M3), and distorted cube (M4). In addition, computations have been performed for two 9-fold polyhedra, a capped square antiprism (M5) and a tricapped trigonal prism (M6), and for two 10-coordinated models, a bicapped square antiprism (M7) and a bicapped regular cube (M8). For the sake of simplicity we have maintained the same notation and the same interatomic distances (see Table 1) as in ref 9. The results of the computations performed on U(IV) 8-coordinated models are reported in Figure 2. Both the M2 and M4 models, based, respectively, on a regular and distorted cube of oxygen atoms surrounding the uranium center, yield a spectral shape that does not resemble the experimental spectrum. In contrast, the computation performed for the bicapped hexagonal prism (M1) shows a XANES profile similar to the experimental spectrum. However, the relative energy separation between the different spectral features does not match the experimental data. In particular, the energies of both the minimum of the negative dip and the maximum of the broad resonance, fall short of the experimental values. In contrast, the computation performed for the square antiprism (M3) shows a very good agreement with the

experimental data, in both the shape and relative energies of the different spectral features. Similar computations have been performed for the proposed 9- and 10-coordinated polyhedra. As shown in Figure 3, the theoretical spectra corresponding to the 10-fold M7 and M8 coordination models yield spectral features appearing at lower energies than experimentally observed. In addition, the spectral shape of the broad positive oscillation centered at ∼40 eV above the edge is not reproduced by the computations that contrastingly yield an asymmetric peak. In the case of the 9-coordination models the situation improves and the theoretical spectra provide a more accurate reproduction of the experimental data. This is particularly true in the case of the tricapped trigonal prism, (M6) model, for which the relative energy separation of the main XANES structures is better reproduced. These results show that the best reproduction of the experimental spectra is obtained by using both an 8-coordinated square antiprism (M3) and a 9-fold tricapped trigonal prism (M6). A detailed comparison of the results obtained by using these hydration models is given in Figure 4. Although the agreement between the experimental and the theoretical spectra is remarkable in both cases, the computation performed by using a 9-fold tricapped trigonal prism shows the best reproduction of the experimental data, in agreement with the DFT results by Tsushima et al.9 Identical results have been obtained for the two other cations modeled, Th(IV) and Np(IV). For the sake of completion we report in Figure 5 the comparison between the experimental Th L3-edge XANES spectrum of Th(IV) in aqueous solution and the theoretical spectra calculated for the M3 and M6 coordination models. A similar comparison in the case of the Np L3-edge is shown in Figure 6. We have also included in these cases the comparison with the theoretical computation performed for the distorted cube, M4, model as the DFT study by Tsushima et al.9 indicated that for neptunium M4 was found to be slightly more stable than M6. As in the case of U(IV) both M3 and M6 models yield a good reproduction of the experimental data although the spectral features computed for the 8-coordinated M3 model appear shifted to higher energies than experimentally 2347

dx.doi.org/10.1021/jp1105482 |J. Phys. Chem. A 2011, 115, 2345–2349

The Journal of Physical Chemistry A

ARTICLE

Figure 3. Comparison of the experimental U L3-edge XANES spectrum of U(IV) in aqueous solution [adapted from refs 17 (b) and 30 (—)] and the theoretical spectra calculated for two different 9-coordinated models, M5 (red, O) and M6 (green, —), and two 10-coordinated models, M7 (blue, 3 3 3 ) and M8 (black, - 3 -). The theoretical spectra have been aligned to the maximum of the white line and vertically shifted for the sake of clearness. (See text for details.)

Figure 4. Comparison of the experimental U L3-edge XANES spectrum of U(IV) in aqueous solution (b) and the theoretical spectra calculated for the 8-coordinated square antiprism, M3 (blue, 3 3 3 ), and for the 9-fold tricapped trigonal prism, M6 (red, —). The theoretical spectra have been aligned to the maximum of the white line and vertically shifted for the sake of clearness. (See text for details.)

observed. By contrast, the computation performed for the distorted cube model yields a spectral shape that does not resemble the experimental spectrum. It is worth noting that all the computations have been performed by using the structural models obtained for [U(H2O)N]4þ clusters optimized in the gas phase by using relativistic density functional theory calculations.9 In all cases the An-O distances obtained by Tsushima et al.9 are longer by 0.06-0.2 Å than those determined from EXAFS analysis. In order to verify the validity of our results we have additionally performed ab initio calculations for the U(IV) aqua-ion using the same coordination polyhedra but adopting the average interatomic distances (2.41 Å) obtained from EXAFS (see Table 2 in the Supporting Information). The results of these computations are included in the Supporting Information (Figures S2-S4)

Figure 5. Comparison of the experimental Th L3-edge XANES spectrum of Th(IV) in aqueous solution (b) and the result of the computations performed for different coordination models: the 8-fold square antiprism (M3, blue, 3 3 3 ), the 9-fold tricapped trigonal prism (M6, red, —), and the distorted cube (M4, green, ---).

where it can be seen that the spectral shape of the theoretical XANES spectra is not affected by modifying the interatomic distances. Consequently, the above conclusions remain unchanged; the best reproduction of the experimental data is obtained by using a 9-fold coordination model of 9 water molecules forming a tricapped trigonal prism.

4. SUMMARY AND CONCLUSIONS In this work we have studied the coordination structure of U(IV), Np(IV), and Th(IV) in aqueous solution by means of XANES at the L3 absorption edge. The high sensitivity of XANES to the bonding geometry allows us to discriminate between the range of coordination polyhedra involving similar coordination numbers but different bond angles. The comparison between the experimental data and the theoretical 2348

dx.doi.org/10.1021/jp1105482 |J. Phys. Chem. A 2011, 115, 2345–2349

The Journal of Physical Chemistry A

Figure 6. Comparison of the experimental Np L3-edge XANES spectrum of Np(IV) in aqueous solution (b) and the result of the computations performed for different coordination models: the 8-fold square antiprism (M3, blue, 3 3 3 ), the 9-fold tricapped trigonal prism (M6, red, —), and the distorted cube (M4, green, ---).

calculations performed in this work point toward the three cases for the solvation sphere of the An(IV) ions being best modeled by 9 water molecules forming a tricapped trigonal prism.

ARTICLE

(16) Denecke, M. A.; Marquardt, C. M.; Rothe, J.; Dardenne, K.; Jensen, M. P. J. Nucl. Sci. Technol. 2002, No. 3 Suppl., 410. (17) Hennig, C. Phys. Rev B 2007, 75, 035120. (18) Hennig, C.; Ikeda-Ohno, A.; Tsushima, S.; Scheinost, A. C. Inorg. Chem. 2009, 48, 5350. (19) Sandstr€om, M.; Persson, I.; Jalilehvand, F.; Lindquist-Reis, P.; Spangberg, D.; Hermansson, K. J. Synchrotron Radiat. 2001, 8, 657. (20) Hennig, C.; Le Naour, C.; Den Auwer, C. Phys. Rev B 2008, 77, 235102. (21) Chaboy, J.; Marcelli, A.; Tyson, T. A. Phys. Rev. B 1994, 49, 11652. (22) Chaboy, J.; Tyson, T. A. Phys. Rev. B 1994, 49, 5869. (23) Díaz-Moreno, S.; Mu~ noz-P
aez, A.; Chaboy, J. J. Phys. Chem. A 2000, 104, 1278. (24) Díaz-Moreno, S.; Chaboy, J. J. Phys. Chem. B 2009, 113, 3527. (25) Natoli, C. R.; Misemer, D. R.; Doniach, S.; Kutzler, F.W. Phys. Rev. A 1980, 22, 1104. (26) Chaboy, J. J. Synchrotron Radiat. 2009, 16, 533. (27) Chaboy, J.; Maruyama, H.; Kawamura, N. J. Phys.: Condens. Matter 2007, 19, 216214. (28) Chaboy, J.; Mu~ noz-P
aez, A.; Carrera, F.; Merkling, P.; S
anchez Marcos, E. Phys. Rev. B 2005, 71, 134208. (29) Krause, M. O.; Oliver, J. H. J. Phys. Chem. Ref. Data 1979, 8, 329. (30) Ikeda-Ohno, A.; Hennig, C.; Tsushima, S.; Scheinost, A. C.; Bernhard, G.; Yaita, T. Inorg. Chem. 2009, 48, 7201.

’ ASSOCIATED CONTENT

bS

Supporting Information. Additional figures and table. This material is available free of charge via the Internet at http:// pubs.acs.org.

’ ACKNOWLEDGMENT This work was partially supported by the Spanish CICYT MAT2008-06542-C04 grant. The authors would like to thank Daniel T. Bowron for his useful suggestions and fruitful discussions. ’ REFERENCES (1) Denecke, M. A. Coord. Chem. Rev. 2006, 250, 730. (2) Pocev, S.; Johansson, G. Acta Chem. Scand. 1973, 27, 2146. (3) Johansson, G.; Magini, M.; Ohtaki, H. J. Solution Chem. 1991, 20, 775. (4) Moll, H.; Denecke, M. A.; Jalilehvand, F.; Sandstr€om, M.; Grenthe, I. Inorg. Chem. 1999, 38, 1795. (5) Hennig, C.; Tutschku, J.; Rossberg, A.; Bernhard, G.; Scheinost, A. C. Inorg. Chem. 2005, 44, 6655. (6) Ikeda-Ohno, A.; Hennig, C.; Rossberg, A.; Funke, H.; Scheinost, A. C.; Bernhard, G.; Yaita, T. Inorg. Chem. 2008, 47, 8294. (7) Brown, I. D. Chem. Soc. Rev. 1978, 7, 359. (8) Zachariasen, W. H. J. Less-Common Met. 1978, 62, 1. (9) Tsushima, S.; Yang, T. X. Chem. Phys. Lett. 2005, 401, 68. (10) Hennig, C.; Schmeide, K.; Brendler, V.; Moll, H.; Tsushima, S.; Scheinost, A. C. Inorg. Chem. 2007, 46, 5882. (11) Rothe, J.; Denecke, M. A.; Neck, V.; Muller, R.; Kim, J. I. Inorg. Chem. 2002, 41, 249. (12) Fratiello, R. E. L. A.; Schuster, R. E. Inorg. Chem. 1970, 9, 391. (13) Allen, P. G.; Bucher, J. J.; Shuh, D. K.; Edelstein, N. M.; Reich, T. Inorg. Chem. 1997, 36, 4676. (14) Antonio, M. R.; Soderholm, L.; Williams, C. W.; Blaudeau, J. P.; Bursten, B. E. Radiochim. Acta 2001, 89, 17. (15) Ikeda-Ohno, A.; Hennig, C.; Rossberg, A.; Funke, H.; Scheinost, A. C.; Bernhard, G.; Yaita, T. Inorg. Chem. 2008, 47, 8294. 2349

dx.doi.org/10.1021/jp1105482 |J. Phys. Chem. A 2011, 115, 2345–2349