Characteristics of chemical bonding of pentavalent uranium in La-doped UO₂

Abstract

The effect of La doping on the electronic structure of U in UO₂ was studied using an advanced technique, namely, X-ray absorption spectroscopy (XAS) in the high-energy-resolution fluorescence-detection (HERFD) mode, at the U 3d_3/2 (M₄) edge. Thanks to a significant reduction of the core–hole lifetime broadening and distinct chemical shifts of the HERFD-XAS lines, the U(V) formation as a result of La doping was identified. The isolated contribution of U(V) in the M₄ HERFD-XAS spectrum reveals the so-called charge-transfer satellites due to the U 5f–O 2p hybridization. The analysis of the experimental data within the framework of the Anderson impurity model (AIM) indicates a significant change in the characteristics and degree of covalency for the chemical bonding in the U(V) subsystem of UO₂ as compared to undoped UO₂, which is a Mott–Hubbard system. The results are also supported by AIM calculations of X-ray photoelectron and optical absorption data.

---

The change of charge on uranium atoms, the dopant position within the crystal lattice, and the possible crystallographic re-arrangement of the oxygen sublattice are fundamental questions (see e.g.ref. 1 and 2) accompanying the doping of UO₂ that help to better understand the observed properties of resulting compounds. These questions are also important for the nuclear industry since UO₂ is used as a nuclear fuel. Some of the created fission products are dissolved in the UO₂ matrix and thereby affect the structural, thermomechanical, thermodynamic and other properties of the nuclear fuel. In the case of UO₂ doped with trivalent rare-earths (REs), it is generally expected that there will be an appearance of the U(V) fraction or/and formation of O vacancies as charge-compensation mechanisms.³ However, the situation appears to be more complex since both O vacancies and oxygen excess are found in U_1−xRE_xO_2±δ for the range of the RE doping⁴ and the phase separation between U_1−xRE_xO_2−δ and U_1−xRE_xO₂ is also suggested for doped systems.^5,6

The charge compensation mechanism due to the change of the oxidation state of U atoms in UO₂ upon RE insertion also presents an open question. Besides the U(V) appearance in such a case of doping, the U(VI) contribution has been also found from the analysis of the magnetic susceptibility² and X-ray photoelectron spectroscopy (XPS) data.^7–10 However, the results obtained by different research groups do not agree with each other. Herrero et al.⁸ have found dominant U(V) with a small U(VI) admixture in hyperstoichiometric U_1−xLa_xO_2+δ for the 0.3 < x < 0.5 range and have only detected the existence of the U(IV) + U(VI) combination with a dominant U(VI) contribution in hypostoichiometric U_1−xLa_xO_2−δ for x > 0.5. On the other hand, the U(V) content has been reported by Venkata Krishnan et al.⁹ to be significantly larger than that of U(VI) in hypostoichiometric U_1−xLa_xO_2−δ.

The consequences of the U(V) appearance in the RE-doped UO₂ lattice in terms of the local symmetry and local structure changes are also not clear. While, in Am(III)-doped UO₂, which is claimed¹¹ to behave more like trivalent lanthanide-doped UO₂ (once the Am/(U + Am) ratio is ≥40%), the U(V) appearance is accompanied by a formation of the U₆O₁₂ cuboctahedral-type clusters as detected by extended X-ray absorption fine structure and Raman spectroscopies;¹¹ the Raman measurements on Nd-doped⁶ and Gd, Dy-doped¹⁰ UO₂ have not confirmed that. Another important consequence is a change in the chemical bonding that affects the physical and chemical properties of the doped compound. While it has been shown that stoichiometric UO₂ is a Mott–Hubbard system¹² with a high degree of 5f localization, for which even the description of relatively delocalized U 6d¹³ and O 2p¹⁴ states requires taking into account the 5f–5f Coulomb interaction, there are no spectroscopic data available for RE-doped UO₂ that would help to assess the situation.

Recently, an advanced technique, namely X-ray absorption spectroscopy (XAS) in the high energy resolution fluorescence detection (HERFD) mode, which improves the experimental resolution due to a significant reduction in the core–hole lifetime broadening of the recorded spectra, has been extensively used in the study of molecular systems and solids.^15–17 For actinides in particular, it has been shown¹⁸ that HERFD-XAS at the actinide 3d_3/2 edge can easily distinguish between U(IV), U(V) and U(VI) oxidation states. Furthermore, this advanced technique has allowed us to resolve both the crystal-field splitting of the 5f shell and the so-called charge-transfer satellites in the HERFD-XAS spectra of Th (ref. 19) and U (ref. 20) compounds. Analysis of these satellites provides information about the characteristics of the chemical bonding and the degree of hybridization of the actinide states with the ligand states.

In this communication, we employ the advanced HERFD-XAS technique to study the effect of La doping on UO₂. By isolating the U(V) contribution in the spectra and by analyzing this contribution in the framework of the Anderson impurity model²¹ (AIM), we find a significant change in the characteristics of the chemical bonding in the U(V) subsystem of UO₂, which must affect the physical and chemical properties of the compound.

U_1−xLa_xO_2±δ (x = 0.06, 0.11, 0.22, and 0.40) samples have been prepared via gel-supported precipitation, also referred to as sol–gel external gelation.²² Uranyl nitrate and ¹³⁹La nitrate solutions were mixed to achieve a molar stoichiometry of x = La/(U + La), as defined above. An organic thickener (Methocel, Dow Chemicals) was added to increase the viscosity of the solutions. These were then dropped into an ammonia bath where the droplet to particle conversion took place due to hydroxide precipitation of the metals inside the polymer backbone of the droplet. The resulting beads were washed, dried, and then calcined at 873 K in air (2 h) to remove organics and then at 973 K in Ar/H₂ (2 h) to ensure uranium reduction. The samples were then enriched at about 30% in ¹⁷O by heating in ¹⁷O enriched gas at 1073 K (24 h) and then sintered for 4 hours at 1923 K under Ar/H₂. The resulting samples represented solid solutions with the single-(FCC)-phase fluorite structure as confirmed by X-ray diffraction analysis. The β-UO₃ sample was obtained by calcining uranyl nitrate in air at 723–773 K over several weeks as described in ref. 23.

The measurements in the energy range of the U 3d X-ray absorption edge were performed at beamline ID26²⁴ of the European Synchrotron in Grenoble. The incident energies were selected using the 〈111〉 reflection from a double Si-crystal monochromator. The XAS data were measured in the HERFD mode using an X-ray emission spectrometer.¹⁷ The U HERFD spectra at the M₄ (3d_3/2 → 5f_5/2,7p transitions) edge were obtained by recording the outgoing photons with an energy of ∼3337 eV as a function of the incident energy. This emission energy which corresponds to the maximum of the U Mβ (4f_5/2 → 3d_3/2 transitions) X-ray emission line was selected using the 〈220〉 reflection of five spherically bent Si crystal analyzers (with 1 m bending radius) aligned at 75° Bragg angle. The directions of the incident and emitted photons were 45° to the surface of the sample. The spectral intensity was normalized to the incident flux. A combined (incident convoluted with emitted) energy resolution of 0.4 eV was obtained as determined by measuring the full width at half maximum (FWHM) of the elastic peak.

The Anderson impurity model²¹ was used for the calculations which included the 5f and core (3d or 4f) states on a single actinide ion and a filled ligand 2p band. The calculations were performed in a manner described in ref. 25 and 26.

The XAS and core-level X-ray photoelectron spectroscopy (XPS) data were calculated using the following equations:

andwhere |g〉 and |m〉 are the ground and XAS final states of the spectroscopic process with energies E_g and E_m, respectively. D is the operator for the optical dipole transition with the incident photon energy represented by ω and lifetime broadening Γ_m of the final state in terms of the half width at half maximum (HWHM). |f〉 is the XPS final state with energy E_f, E_B is the binding energy, and a_c is the annihilation operator of a core electron.

The required Slater integrals F^k, G^k and R^k (ref. 27), spin–orbit coupling constants ζ and matrix elements were obtained using the TT-MULTIPLETS package which combines Cowan's atomic multiplet program²⁸ (based on the Hartree–Fock method with relativistic corrections) and Butler's point-group program,²⁹ which were modified by Thole,³⁰ as well as the charge-transfer program written by Thole and Ogasawara.

We used the AIM approach for the calculations of the spectra because the ab-initio methods based on density functional theory, which are successful in many cases,^15–17 still have difficulties in the proper description of the transitions involving the 5f electrons³¹ when the on-site 5f–5f Coulomb interaction is strong.

The U M₄ HERFD-XAS spectra of UO₂, U_1−xLa_xO₂ (x = 0.06, 0.11, 0.22, and 0.4) and β-UO₃ are displayed in Fig. 1. The spectra are normalized to the maximum. A significant chemical shift between the U M₄ lines of UO₂ and UO₃ is observed since they contain nominal U(IV) and U(VI), respectively. Similar chemical shifts between the M₄ HERFD-XAS spectra of the U(IV) and U(VI) species have been reported earlier.^{18,20,31–33} Upon La doping, the shoulder on the high-energy side of the U M₄ main line of UO₂ starts to grow and evolves into a dominant peak at ∼3726.4 eV for U_0.6La_0.4O₂. The energy of this peak falls in between those of the M₄ main lines of U(IV) and U(VI) substances. As in the case of U₄O₉¹⁸ and NaUO₃,²⁰ this is an indication of U(V) formation. It is clear from Fig. 1 that the U(V) content increases with x in U_1−xLa_xO₂.

Fig. 1 High-resolution XAS spectra at the U M₄ edge of UO₂,³² U_1−xLa_xO₂ and β-UO₃. The spectra are normalized to the maximum.

To analyze the newly appeared ∼3726.4 eV structure, we subtracted the U(IV) contribution from the U M₄ HERFD-XAS spectrum of U_0.6La_0.4O₂ using the U M₄ HERFD-XAS spectrum of undoped UO₂ from Fig. 1. Based on the charge-compensation considerations for the stoichiometric O content, the U(IV) contribution was taken as one-third of the total U M₄ spectral weight. The resulting spectrum is shown in Fig. 2. It looks similar to the U M₄ HERFD-XAS spectrum of NaUO₃ (ref. 20), thus supporting its own U(V) origin. However, the U cation in NaUO₃ is six-fold coordinated while the eight-fold coordination is expected for the U(V) atoms in the UO₂ lattice. To compare with experimental data, we calculated the M₄ XAS spectrum of U(V) in the UO₂ lattice using the AIM and eqn (1). Earlier we have shown²⁰ that there is no significant difference between calculated conventional XAS (with artificially reduced broadening) and HERFD-XAS spectra of U(V), therefore we limit the calculations to the conventional XAS spectrum to simplify the computational framework.

Fig. 2 High-resolution U M₄ XAS spectrum of U_0.6La_0.4O₂ after subtraction of the U(IV) contribution compared to the calculated XAS spectrum at the M₄ edge of U(V) in the UO₂ lattice.

In the AIM calculations, the ground (final) state of the system was described using a linear combination of the 5f¹ and (3d⁹5f² and ) configurations, where stands for an electronic hole in the O 2p band. The Slater integrals were reduced to 70% of their Hartree–Fock values to account for the configuration interaction effect.³⁴ To describe the crystal-field interaction for the U(V) atom with the eight-fold coordination, the values of Wybourne's crystal-field parameters³⁵ for the 5f shell in the cubic (O_h) symmetry were set to B₀⁴ = −0.93 eV and B₀⁶ = 0.35 eV, which are similar to those derived from the inelastic neutron scattering measurements³⁶ on UO₂.

In the limit of the U 5f–O 2p hybridization term V → 0, the difference between the configuration averaged energies for the ground state can be written as (where Δ ≡ ε_f − ε_n, with ε_n corresponding to the center of the O 2p band) which is the so-called charge-transfer energy (ε_f and ε_n are one-electron energies of the U 5f and O 2p levels). For the final state, this difference is , where U_ff denotes the 5f–5f Coulomb interaction and U_fc is the 3d core–hole potential acting on the 5f electron. Treated as parameters, the Δ, U_ff and U_fc values were taken to be 3.5, 3.5, and 6.0 eV, respectively. The hybridization term (hopping matrix element) between 5f¹ and configurations in the ground state was taken to be V_g = 1.2 eV. To account for the electronic-configuration dependence of V, this value was scaled down to 75% (0.75V_g = V_m) to describe the mixing between 3d⁹5f² and configurations in the final state.^25,37 The 3d_3/2 core–hole-lifetime Lorentzian broadening (Γ_m) was taken to be 0.3 eV with an additional Gaussian convolution to account for the instrumental resolution.

Furthermore, the (super)-exchange interaction³⁸ and the linear polarization of the incident X-ray beam were taken into account in the calculations of the U(V) M₄ spectrum; this spectrum was calculated at room temperature.

The results of the calculations are shown in Fig. 2, together with the U M₄ spectrum of U_0.6La_0.4O₂ after subtraction of the U(IV) contribution. The experimental and calculated spectra are in rather good agreement, thus indicating that most of the U(V) atoms are in the cubic-symmetry fluorite-type environment and coordinated by eight O atoms. The structure at ∼8 eV above the main line in the experimental spectrum is assigned by the calculations to the U 5f → O 2p charge-transfer satellite as a result of the U 5f–O 2p hybridization. This satellite depends on both V and Δ values.

Our estimations of the AIM parameters for the U(V) subsystem from the analysis of the HERFD-XAS data at the U 3d edge are also supported by the calculations of the U 4f XPS spectrum (Fig. 3) of La-doped UO₂ that contains 100% U(V) as found in ref. 9. The U_0.4La_0.6O_2−δ sample also had the single-(FCC)-phase fluorite crystal structure. Despite that the sample was hypostoichiometric, the U(V) cations are eight-fold coordinated because the O vacancies are expected to be located around the La sites.

Fig. 3 Calculated U 4f XPS spectrum of U(V) in the UO₂ lattice using the Anderson impurity model. The spectrum is compared with the experimental data from ref. 9. The dashed curves represent the photoelectron background.

The experimental data were well reproduced (Fig. 3) using the same set of AIM parameters apart from that for U_fc which is expected to be smaller for the 4f level. In this case, the U_fc value was set to 5.0 eV, which is similar to that used in the AIM calculations in ref. 39. In our calculations of the U(V) 4f XPS spectrum, the final state configurations with the 4f core hole had the same number of 5f electrons as the corresponding ground state configurations due to an excited electron leaving the system. The calculated spectrum reproduces the experimental data in terms of the energy separation (∼8 eV) and relative intensity of the charge-transfer satellites at binding energies of ∼388.4 and ∼399.4 eV for the main U 4f_7/2 and 4f_5/2 lines, respectively.

Table 1 compares the lowest excited states of the U(V) 5f multiplet obtained in the AIM calculations with those measured using diffuse reflectance spectroscopy (DRS), which is an analogue of optical absorption spectroscopy.^8,40 The DRS spectra of La-doped UO₂ with a dominant U(V) contribution show an intense peak at ∼6400 cm⁻¹ and a broad structure at ∼12 000 cm⁻¹, which are associated with the electronic transitions within the U(V) 5f shell (also see the DRS spectrum of U(V) in stabilized ZrO₂ [ref. 41]). Since the DRS spectra were measured in the 4000–40 000 cm⁻¹ range, the first calculated transition (Γ₈ → Γ₇) in Table 1 is out of the measured range. Our calculations suggest that the ∼6400 cm⁻¹ peak consists of two transitions Γ₈ → Γ₈′ and Γ₈ → Γ₆, thus explaining why this peak is so intense. The broad structure at ∼12 000 cm⁻¹ can be assigned to the Γ₈ → Γ₇′ transition.

Table 1 Lowest states of the U 5f multiplet (in meV) calculated for the cubic crystal-field environment (eight-fold coordination of U(V) cation) within the Anderson impurity model and comparison with the results obtained using diffuse reflectance spectroscopy^8,40

State	Experiment	Calculations
Γ 8		0
Γ 7		175
Γ 8′	794	954
Γ 6	794	958
Γ 7′	∼1488	1378

---

As a whole, the results of our calculations indicate a very significant change in the nature of the chemical bonding in the formed U(V) subsystem upon RE-doping of UO₂ as compared to undoped UO₂. Based on the ratio between Δ and U_ff (ref. 42), the locally formed U(V) environment is characterized as a charge-transfer subsystem in contrast to the Mott–Hubbard system of UO₂. The contributions of the 5f¹ and configurations in the ground state of the U(V) subsystem were estimated to be 65% and 35%, respectively, thus resulting in the 5f occupancy of n_f = 1.35 electrons. In comparison with nominal n_f = 1 for U(V), extra n_f arises from the U 5f–O 2p hybridization. This extra n_f amounts to 0.35 electrons and is higher than that in undoped UO₂ (0.24 electrons) estimated from similar AIM calculations.^39,43

References

1. (a) G. C. Allen, P. A. Tempest and J. W. Tyler, Nature, 1982, 295, 48–49; (b) S. D. Conradson, S. M. Gilbertson, S. L. Daifuku, J. A. Kehl, T. Durakiewicz, D. A. Andersson, A. R. Bishop, D. D. Byler, P. Maldonado, P. M. Oppeneer, J. A. Valdez, M. L. Neidig and G. Rodriguez, Sci. Rep., 2015, 5, 15278.
2. Y. Hinatsu, J. Alloys Compd., 1993, 201, 139–144.
3. (a) C. R. A. Catlow, J. Chem. Soc., Faraday Trans. 2, 1987, 83, 1065–1072; (b) J. M. Solomon, V. Alexandrov, B. Sadigh, A. Navrotsky and M. Asta, Acta Mater., 2014, 78, 282–289.
4. (a) T. B. Lindemer and J. Brynestad, J. Am. Ceram. Soc., 1986, 69, 867–876; (b) K. Park and D. R. Olander, J. Nucl. Mater., 1992, 187, 89–96.
5. T. Fujino, K. Ouchi, Y. Mozumi, R. Ueda and H. Tagawa, J. Nucl. Mater., 1990, 174, 92–101.
6. L. Desgranges, Y. Pontillon, P. Matheron, M. Marcet, P. Simon, G. Guimbretiere and F. Porcher, Inorg. Chem., 2012, 51, 9147–9149.
7. C. Keller and C. K. Jørgensen, Chem. Phys. Lett., 1975, 32, 397–400.
8. P. Herrero, P. Garcia-Chain and R. M. Rojas, J. Solid State Chem., 1990, 87, 331–343.
9. R. Venkata Krishnan, V. K. Mittal and R. Babu, J. Alloys Compd., 2011, 509, 3229–3237.
10. M. Razdan and D. W. Shoesmith, J. Electrochem. Soc., 2014, 161, H105–H113.
11. F. Lebreton, D. Horlait, R. Caraballo, P. M. Martin, A. C. Scheinost, A. Rossberg, C. Jégou and T. Delahaye, Inorg. Chem., 2015, 54, 9749–9760.
12. (a) S. L. Dudarev, D. Nguyen Manh and A. P. Sutton, Philos. Mag. B, 1997, 75, 613–628; (b) Y. Baer and J. Schoenes, Solid State Commun., 1980, 33, 885–888.
13. K. O. Kvashnina, Y. O. Kvashnin, J. R. Vegelius, A. Bosak, M. Martin and S. M. Butorin, Anal. Chem., 2015, 87, 8772–8780.
14. A. Modin, M.-T. Suzuki, J. Vegelius, Y. Yun, D. K. Shuh, L. Werme, J. Nordgren, P. M. Oppeneer and S. M. Butorin, J. Phys.: Condens. Matter, 2015, 27, 315503.
15. (a) A. J. Atkins, M. Bauer and C. R. Jacob, Phys. Chem. Chem. Phys., 2015, 17, 13937–13948; (b) A. J. Atkins, M. Bauer and C. R. Jacob, Phys. Chem. Chem. Phys., 2013, 15, 8095–8105.
16. J. C. Swarbrick, Y. Kvashnin, K. Schulte, K. Seenivasan, C. Lamberti and P. Glatzel, Inorg. Chem., 2010, 49, 8323–8332.
17. P. Glatzel and U. Bergmann, Coord. Chem. Rev., 2005, 249, 65–95.
18. K. O. Kvashnina, S. M. Butorin, P. Martin and P. Glatzel, Phys. Rev. Lett., 2013, 111, 253002.
19. S. M. Butorin, K. O. Kvashnina, J. R. Vegelius, D. Meyer and D. K. Shuh, Proc. Natl. Acad. Sci. U. S. A., 2016, 113, 8093–8097.
20. S. M. Butorin, K. O. Kvashnina, A. L. Smith, K. Popa and P. M. Martin, Chem. – Eur. J., 2016, 22, 9693–9698.
21. P. W. Anderson, Phys. Rev. [Sect.] B, 1961, 124, 41.
22. (a) U. Carvajal-Nunez, D. Prieur, T. Vitova and J. Somers, Inorg. Chem., 2012, 51, 11762–11768; (b) L. Martel, J.-F. Vigier, D. Prieur, S. Nourry, A. Guiot, K. Dardenne, J. Boshoven and J. Somers, J. Phys. Chem. C, 2014, 118, 27640.
23. P. C. Debets, Acta Crystallogr., 1966, 21, 589–593.
24. C. Gauthier, V. A. Sole, R. Signorato, J. Goulon and E. Moguiline, J. Synchrotron Radiat., 1999, 6, 164–166.
25. S. M. Butorin, D. C. Mancini, J.-H. Guo, N. Wassdahl, J. Nordgren, M. Nakazawa, S. Tanaka, T. Uozumi, A. Kotani, Y. Ma, K. E. Myano, B. A. Karlin and D. K. Shuh, Phys. Rev. Lett., 1996, 77, 574–577.
26. M. Nakazawa, H. Ogasawara and A. Kotani, Surf. Rev. Lett., 2002, 9, 1149–1153.
27. J. C. Slater, Phys. Rev., 1929, 34, 1293–1322.
28. R. D. Cowan, Theory of Atomic Structure and Spectra, University of California Press, Berkeley, 1981.
29. P. H. Butler, Point Group Symmetry Applications: Methods and Tables, Plenum Press, New York, 1981.
30. B. T. Thole, G. van der Laan and P. H. Butler, Chem. Phys. Lett., 1988, 149, 295–299.
31. K. O. Kvashnina, Y. O. Kvashnin and S. M. Butorin, J. Electron Spectrosc. Relat. Phenom., 2014, 194, 27–36.
32. A. L. Smith, P. E. Raison, L. Martel, D. Prieur, T. Charpentier, G. Wallez, E. Suard, A. C. Scheinost, C. Hennig, P. Martin, K. O. Kvashnina, A. K. Cheetham and R. J. M. Konings, Inorg. Chem., 2015, 54, 3552–3561.
33. R. Bès, M. Rivenet, P.-L. Solari, K. O. Kvashnina, A. C. Scheinost and P. M. Martin, Inorg. Chem., 2016, 55, 4260–4270.
34. J. Sugar, Phys. Rev. B: Solid State, 1972, 5, 1785–1792.
35. B. G. Wybourne, Spectroscopic Properties of Rare Earths, Wiley, New York, 1963.
36. H. Nakotte, R. Rajaram, S. Kern, R. J. McQueeney, G. H. Lander and R. A. Robinson, J. Phys.: Conf. Ser., 2010, 251, 012002.
37. O. Gunnarsson and O. Jepsen, Phys. Rev. B: Condens. Matter Mater. Phys., 1988, 38, 3568–3571.
38. C. Miyake, O. Kawasaki, K. Gotoh and A. Nakatani, J. Alloys Compd., 1993, 200, 187–190.
39. T. Yamazaki and A. Kotani, J. Phys. Soc. Jpn., 1991, 60, 49–52.
40. S. Kemmler-Sack, Z. Anorg. Allg. Chem., 1968, 363, 295–304.
41. E. R. Vance, Y. Zhan and and Z. Zhang, J. Nucl. Mater., 2010, 400, 8–14.
42. J. Zaanen, G. A. Sawatzky and J. W. Allen, Phys. Rev. Lett., 1985, 55, 418–421.
43. S. M. Butorin, A. Modin, J. R. Vegelius, K. O. Kvashnina and D. K. Shuh, J. Phys. Chem. C , submitted.

---