Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Trends in the valence band electronic structures of mixed uranium oxides

Trends in the valence band electronic structures of mixed uranium oxides Trends in valence band electronic structure of mixed uranium oxides a,b *c,d e f Kristina O. Kvashnina* , Piotr M. Kowalski , Sergei M. Butorin , Gregory Leinders , Janne f g c,d f Pakarinen , Rene Bes , Haijian Li , Marc Verwerft a. Rossendorf Beamline at ESRF − The European Synchrotron, CS40220, 38043 Grenoble Cedex 9, France b. Helmholtz Zentrum Dresden-Rossendorf (HZDR), Institute of Resource Ecology, P.O. Box 510119, 01314 Dresden, Germany c. Address here. Institute of Energy and Climate Research, IEK-6, Nuclear Waste Management and Reactor Safety, Forschungszentrum Jülich GmbH, Wilhelm - Johnen-Strasse, 52428 Jülich, Germany d. JARA High-Performance Computing, Schinkelstraße 2, 52062 Aachen, Germany e. Molecular and Condensed Matter Physics, Department of Physics and Astronomy, Uppsala University, SE-751 20 Uppsala, Sweden f. Belgian Nuclear Research Centre (SCK·CEN), Institute for Nuclear Materials Science, Boeretang 200, B-2400 Mol, Belgium g. Department of Applied Physics, Aalto University, P.O. Box 14100, FI-00076 Aalto, Finland * These authors contributed equally. Valence band electronic structure of mixed uranium oxides (UO , U O , U O , U O , -UO ) has been studied by the resonant inelastic 4 9 3 7 3 8 3 X-ray scattering (RIXS) technique at the U M edge and by computational methods. We show here that the RIXS technique and recorded U 5f - O 2p charge transfer excitations can be used to proof the validity of theoretical approximations. Structural, electronic and chemical properties of uranium (U) oxides vary strongly upon a transformation from the fluorite- type UO structures to the layered structure of the higher U 1–13 oxides (U O and above) . The mechanism of the expansion 3 8 14–16 of the fluorite structure is reasonably straightforward , however the role of oxygen (O) atoms in these structural changes remains less clear. We performed the state-of-art valence band RIXS experiment at the U M edge for a number of binary U oxides – UO , U O U O , U O , -UO – in order to 2 4 9, 3 7 3 8 3 clearly identify the mechanism causing the electronic structure modification upon oxidation of UO Valence band RIXS data at the U M edge (~3550 eV) have been previously reported for UO , UO , UF , UO (NO ) *6(H O) 2 3 4 2 3 2 2 17–20 and several U intermetallic systems , and have been proved to be sensitive to the structural environment of U atom and its ligands. Actually, the valence band RIXS data include the elastic and inelastic scattering profiles with an energy resolution of Fig. 1. Valence band RIXS spectra of UO , U O , U O , U O and 2 4 9 3 7 3 8 ~1eV and provide information on the energy difference -UO recorded at the incident photon energy set to the between the valence band states and the unoccupied U 5f maximum of the U M edge. The experimental total energy states. Fig. 1 shows the valence band RIXS spectra of UO , U O , resolution of ~1eV was achieved by employing Johann-type X- 2 4 9 U O , U O and -UO , recorded at the Beamline ID26 of The ray emission spectrometer with five spherically bend Si(220) 3 7 3 8 3 European Synchrotron (ESRF) , (see ESI). The lowest energy crystal analyzers with 1m bending radius (see ESI for details). feature at ~20 eV is attributed to the U 6p -3d 3/2 5/2 17,18 transitions . The process involves first the excitation of an electron from the U 3d core level (at the U M edge) to the 5/2 5 unoccupied U 5f state and then the U 3d core hole is filled by constant increase of the energy separation between the elastic 5/2 an electron from the occupied U 6p state. The inelastic and inelastic scattering contributions in the spectra through the 3/2 scattering profile at energies transfer ~5-10 eV, reported in Fig.1 series of these mixed valence oxides. However, measured valence band RIXS data for the U oxides (Fig.1) show that the mechanism of the electronic structure modification during the transformation of UO into the mixed oxides is however more complicated. Charge transfer also takes place between U sites 2- and additionally incorporated O ions in binary oxides. As a result, the modification of the U-ligand bonding induces a change in the U oxidation state. In addition to that, the position and distribution of valence band states near the Fermi level changes significantly on a scale of several eV. To gain better understanding and to clarify the mechanism of charge transfer excitations and electronic structure modifications we performed three types of calculations with methods of computational quantum chemistry, which have been used to evaluate the properties of U oxides materials 23–30 previously . Computation of mixed U oxides is a challenging task because of a strongly correlated and localized character of the f 23,30 electrons . The commonly used density functional (DFT) methods often fail even on the qualitative level, for instance predicting a metallic state for UO , which is in reality a Mott insulator with a wide band gap of 2.1 eV . To correct for this, the DFT+U method is often used when an on-site Coulomb interaction is modelled by an additional term in the Hamiltonian, whose strength depends on the Hubbard U 23,32,33 parameter . This parameter in calculations for uranium oxides is usually taken as U=4.5 eV (with additional J parameter 30,33 of 0.54eV) or an effective parameter, U =U-J, is applied . eff These values come from the measurements of the correlation 34,35 energy performed on UO . This approximation is the most common approach in the calculations of the electronic structure 10,19,23,25,29,36,37 of U systems . 30,38 Fig. 2. Comparison of measured valence band RIXS spectra of Recently, Beridze and Kowalski performed systematic UO , U O , U O and -UO and results of calculations using a 2 3 7 3 8 3 tests of the performance of the DFT+U method with the variety of methods, denoted as standard approach with U Hubbard U parameter derived ab initio using the linear =4.5eV and J=0.54eV (M1), atomic orbitals (M2), Wannier response method of Cococcioni and Gironcoli for the Functions (M3). calculations of actinide bearing molecular and solid compounds, including U oxides. They have shown that the Hubbard U has been attributed to charge transfer process from the parameter values strongly depend on the oxidation state of U, 18–20 occupied O 2p states to the unoccupied U 5f states . The being largest for U(VI) (~3 eV) and smallest for U(IV) (~2 eV). In energy separation between elastic and inelastic scattering follow up studies a similar trend has been shown for other contributions to the spectra depends of the energy difference actinides . Here we test the performance of this methodology between the occupied O 2p states and unoccupied U 5f states. for the computation of the electronic density of states (DOS) Our recent studies of the evolution of the U chemical state that are used for the construction of theoretical RIXS data. in a series of U oxides confirms a changeover of the oxidation In most of the DFT+U implementations the shape of the states U(IV) - U(V)- U(VI) through the charge compensation orbitals of interest (f orbitals in the case of U) have to be 5,19,22 mechanisms . The established formal oxidation states for U provided in order to estimate the occupation of these orbitals in mixed U oxides are included in brackets: UO (IV), U O (IV-V), 2 4 9 29,33 and compute the Hubbard energy correction term . These U O (IV-V), U O (V-VI), -UO (VI). Moreover the exact 3 7 3 8 3 orbitals are usually represented by the atomic orbitals quantitative analysis has been performed and showed the computed for atoms or ions and thus not necessarily adequately presence of 50% and 50% of U(IV) and U(V), respectively, in represent the shape of the orbital in a solid. In recent studies U O ; 33% and 67% of U(IV) and U(V) in U O and 67% and 33% 4 9 3 7 the maximally localized Wannier functions have been applied in of U(V) and U(VI) in U O . 3 8 32,38–40 computation of electronic structure of solids . In this Based on these findings the process of the electron transfer contribution we will test this approach in order to check if the from the O 2p orbitals to the unfiled U 5f shell should show the Wannier functions-based representation of f orbitals can result COMMUNICATION in any significant improvement in the description of the DOS The issue of the DFT or DFT+U predicted band gaps has been functions and RIXS data. discussed previously . One interesting aspect in the case of UO , In order to reproduce the experimentally detected charge which is often used as a model system, is that for the Mott insulator transfer excitations we made simplified calculations by inserting the the band gap value should be well approximated by the value of the calculated partial U 5f and O 2p DOSs into the Kramers-Heisenberg Hubbard U parameter. The measured band gap for UO is 2.1 eV, 36,41 equation . This approach provides straight forward information which is close to the Hubbard U parameter value predicted by the about the validity or accuracy of DOSs calculated using a variety of linear response method (Table 1). The problem as outlined by methods. It describes a correlation between occupied and Breridze and Kowalski is that the atomic orbitals used to represent unoccupied states under assumption that the hybridization between the f orbitals in solids are not a good representation resulting in a U 5f and O 2p states takes place. In that case, the energy difference significant and unrealistic fractional occupation of the unoccupied f between the maxima of the occupied O 2p DOS and unoccupied U 5f levels (up to 0.3 for UO ). In order to remove this obstacle we applied DOS will define the energy transfer values for the observed RIXS the Wannier representations of the f orbitals for the DFT+U transitions. calculations, which resulted in more realistic, close to zero The partial DOSs have been calculated by three approaches (see occupations of the unoccupied f orbitals. The RIXS profiles resulting ESI). First we have applied the LDA+U approach assuming the from the later calculations are also plotted in Fig. 2. These represent standard values used in calculations for U oxides (U=4.5eV and the best match to the measured RIXS profiles with good prediction J=0.54eV) (denoted as M1). In addition we computed the Hubbard of the position of the U 5f - O 2p charge transfer and an overall much U parameters values using the linear response method of Cococcioni better match to the observed shape of the U 5f – O 2p charge transfer and Gironcoli . Here, for the calculations of the Hubbard correction excitations. we represent f orbitals for projection of occupations by the atomic orbitals (M2) and the Wannier functions (M3). Table 1: The computed Hubbard U parameter values in eV Computation of the electronic structure of U oxides can often U(IV) U(V) U(VI) lead to the convergence to a metastate instead of ground state . In UO 1.7 order to obtain the correct electronic structure of the considered U O 2.1 2.1 3 7 oxides, for an initial electronic state we computed the expected U O 2.0 2.2 3 8 charges of the different U atoms in the mixed U oxides using bond -UO 2.5 valence sum (BVS) method . The BVS of U atoms for the considered Table 2: Predicted and measured band gaps (in eV) of the mixed oxides: U O (P4 /n) , U O (Amm2), UO (P1211), and UO (Fm-3m) 3 7 2 3 8 3 2 U oxides. were calculated and analyzed applying the following formula: M1 M2 M3 Exp 𝑉 = ∑ exp[(𝑅 − 𝑑 ) /𝑏 ] 𝑗 𝑗 (LDA + U) (Atomic) (Wannier) Here the bond valence parameter Rj and constant b are taken UO 2.4 0.3 0.7 2.1 from Ref. V and dj are the corresponding valence and bond lengths U O 1.4 0.6 0.9 1.6 3 7 for each phase. The BVS results are given in ESI, where the U average U O 1.7 0.9 1.3 1.8 3 8 5,22 charge is in approximate agreement with our previous findings . -UO 2.2 2.0 2.7 2.2 The Hubbard U parameter values computed with the linear Table 3: The computed and measured energy difference response for the considered U oxides are given in Table 1. In general, between U 5f and O 2p states. The energies are given in eV as in our previous studies , the value is smaller than 4.5 eV, but M1 M2 M3 Exp strong dependence on the oxidation state is observed. The largest (LDA + U) (Atomic) (Wannier) value was obtained for UO (U(VI)) and the smallest for UO (U(IV)). 3 2 UO 6.5 6.3 7.0 7.5 Fig. 2 shows a comparison of measured and calculated RIXS U O 5.8 5.6 6.5 5.5 3 7 spectra of UO , U O , U O and -UO at the maximum of the U M 2 3 7 3 8 3 5 U O 5.2 5.3 6.1 6.0 3 8 edge, using three approximations (M1, M2 and M3). The elastic -UO 5.2 5.5 6.5 6.2 36 3 scattering contribution has been added to the calculated RIXS spectra to facilitate a comparison with experimental data. The We present here the electronic structure studies of several U standard approach (M1) does not give an ideal match to the mixed oxides. The RIXS spectra shown in Fig. 1 and Fig. 2 indicate the experimental RIXS data (inaccurate energy difference between experimentally obtained energy between the occupied O 2p and elastic and inelastic theoretical RIXS profiles). The calculations with unoccupied U 5f states. In Table 3 we report the difference between the Hubbard U parameter derived ab initio (M2) are even more the average energy of these states as integrated from the computed deviating from the experiment as a result of smaller Hubbard U DOS functions (see ESI) using the three theoretical methods. The best values than the standard one (4.5 eV). The predicted band gaps (see 29,31 match is obtained with the method M3 (with Wannier functions). Table 2) are also smaller than the measured ones . On the other Here the experimentally observed trend is clearly reproduced with hand, the band gaps predicted by M1 method are in qualitative the largest differences for UO and the smallest for U O . These 2 3 8 agreement with the measurements. It is important to note that U 5f differences result from the decrease of the energy of the O p states - O 2p charge transfer excitations, recorded by RIXS in this case, show (with respect to the Fermi level) with increasing the oxidation state the energy difference between two electronic levels (empty U 5f and (due to the stronger electrons binding) and associated decrease of occupied O 2p) and do not directly related to the band gap values the energy of the unoccupied 5f states. The later effect results from obtained by other experimental methods (like optical spectroscopy). higher Hubbard U parameter values (strength of the on-site Coulomb 574–577. repulsion) for higher oxidation states of U (Table1). These interesting 18 S. M. Butorin, J. Electron Spectros. Relat. Phenomena, results indicate that with such a state-of-art experimental method – 2000, 110–111, 213–233. valence band RIXS - one can improve the theoretical prediction of the 19 K. O. Kvashnina, Y. O. Kvashnin and S. M. Butorin, J. electronic structure of actinide contained materials. Electron Spectros. Relat. Phenomena, 2014, 194, 27–36. 20 K. O. Kvashnina, H. C. Walker, N. Magnani, G. H. Lander and R. Caciuffo, Phys. Rev. B, 2017, 95, 245103. Notes and references P.M.K. and J.L. thank the JARA-HPC 21 C. Gauthier, V. A. Solé, R. Signorato, J. Goulon and E. for giving time on the supercomputing resources awarded through Moguiline, J. Synchrotron Radiat., 1999, 6, 164–6. JARA-HPC Partition. S.M.B. acknowledges support from the Swedish 22 K. O. Kvashnina, S. M. Butorin, P. Martin and P. Glatzel, Research Council (research grant 2017-06465) and K.O.K Phys. Rev. Lett., 2013, 111, 253002. acknowledges support from ERC (research grant 759696). 23 X.-D. Wen, R. L. Martin, T. M. Henderson and G. E. Scuseria, 1 G. C. Allen and N. R. Holmes, J. Nucl. Mater., 1995, 223, Chem. Rev., 2013, 113, 1063–1096. 231–237. 24 X.-D. Wen, R. L. Martin, G. E. Scuseria, S. P. Rudin, E. R. 2 J. M. Flitcroft, M. Molinari, N. A. Brincat, N. R. Williams, M. Batista and A. K. Burrell, J. Phys. Condens. Matter, 2013, T. Storr, G. C. Allen and S. C. Parker, J. Mater. Chem. A, 25, 25501. 2018, 6, 11362–11369. 25 S. O. Odoh and G. Schreckenbach, J. Phys. Chem. A, 2010, 3 J. Kretzschmar, T. Haubitz, R. Hübner, S. Weiss, R. Husar, V. 114, 1957–1963. Brendler and T. Stumpf, Chem. Commun., , 26 X.-D. Wen, R. L. Martin, L. E. Roy, G. E. Scuseria, S. P. Rudin, DOI:10.1039/C8CC02070A. E. R. Batista, T. M. McCleskey, B. L. Scott, E. Bauer, J. J. 4 T. Gouder, R. Eloirdi and R. Caciuffo, Sci. Rep., 2018, 8, Joyce and T. Durakiewicz, J. Chem. Phys., 2012, 137, 5 G. Leinders, R. Bes, J. Pakarinen, K. Kvashnina and M. 27 B. Dorado, B. Amadon, M. Freyss and M. Bertolus, Phys. Verwerft, Inorg. Chem., 2017, 56, 6784–6787. Rev. B, 2009, 79, 235125. 6 G. Leinders, J. Pakarinen, R. Delville, T. Cardinaels, K. 28 S. L. Dudarev, D. N. Manh and A. P. Sutton, Philos. Mag. Binnemans and M. Verwerft, Inorg. Chem., 2016, 55, 3915– Part B, 1997, 75, 613–628. 29 H. He, D. A. Andersson, D. D. Allred and K. D. Rector, J. 7 J. M. Elorrieta, L. J. Bonales, N. Rodríguez-Villagra, V. G. Phys. Chem. C, 2013, 117, 16540–16551. Baonza and J. Cobos, Phys. Chem. Chem. Phys., 2016, 18, 30 G. Beridze and P. M. Kowalski, J. Phys. Chem. A, 2014, 118, 28209–28216. 11797–11810. 8 M. Caisso, S. Picart, R. C. Belin, F. Lebreton, P. M. Martin, K. 31 J. Schoenes, J. Appl. Phys., 1978, 49, 1463–1465. Dardenne, J. Rothe, D. R. Neuville, T. Delahaye and A. 32 B. Himmetoglu, A. Floris, S. de Gironcoli and M. Cococcioni, Ayral, Dalt. Trans., 2015, 44, 6391–6399. Int. J. Quantum Chem., 2014, 114, 14–49. 9 V. Kapaklis, G. K. Pálsson, J. Vegelius, M. M. Haverhals, P. T. 33 M. Cococcioni and S. de Gironcoli, Phys. Rev. B, 2005, 71, Korelis, S. M. Butorin, A. Modin, M. Kavčič, M. Zitnik, K. Bučar, K. O. Kvashnina and B. Hjörvarsson, J. Phys. 34 A. Kotani and T. Yamazaki, Prog. Theor. Phys. Suppl., 1992, Condens. Matter, 2012, 24, 495402. 108, 117–131. 10 D. A. Andersson, G. Baldinozzi, L. Desgranges, D. R. 35 Y. Baer and J. Schoenes, Solid State Commun., 1980, 33, Conradson and S. D. Conradson, Inorg. Chem., 2013, 52, 885–888. 2769–78. 36 K. O. Kvashnina, Y. O. Kvashnin, J. R. Vegelius, A. Bosak, P. 11 S. D. Conradson, D. Manara, F. Wastin, D. L. Clark, G. H. M. Martin and S. M. Butorin, Anal. Chem., 2015, 87, 8772– Lander, L. A. Morales, J. Rebizant and V. V Rondinella, Inorg. Chem., 2004, 43, 6922–35. 37 Y. Yun, J. Rusz, M.-T. Suzuki and P. Oppeneer, Phys. Rev. B, 12 R. Zhao, L. Wang, Z.-J. Gu, L.-Y. Yuan, C.-L. Xiao, Y.-L. Zhao, 2011, 83, 1–10. Z.-F. Chai and W.-Q. Shi, CrystEngComm, 2014, 16, 2645. 38 G. Beridze, A. Birnie, S. Koniski, Y. Ji and P. M. Kowalski, 13 G. Leinders, R. Delville, J. Pakarinen, T. Cardinaels, K. Prog. Nucl. Energy, 2016, 92, 142–146. Binnemans and M. Verwerft, Inorg. Chem., 2016, 55, 9923– 39 J. Ma and L.-W. Wang, Sci. Rep., 2016, 6, 24924. 40 D. Novoselov, D. M. Korotin and V. I. Anisimov, J. Phys. 14 L. Desgranges, G. Baldinozzi, G. Rousseau, J.-C. Nièpce and Condens. Matter, 2015, 27, 325602. G. Calvarin, Inorg. Chem., 2009, 48, 7585–92. 41 J. Jiménez-Mier, J. van Ek, D. L. Ederer, T. A. Callcott, J. J. 15 L. Desgranges, G. Baldinozzi, D. Simeone and H. E. Fischer, Jia, J. Carlisle, L. Terminello, A. Asfaw and R. C. Perera, Inorg. Chem., 2016, 55, 7485–7491. Phys. Rev. B, 1999, 59, 2649–2658. 16 L. Desgranges, Y. Ma, P. Garcia, G. Baldinozzi, D. Siméone 42 P. C. Burns, R. C. Ewing and F. C. Hawthorne, Can. Mineral., and H. E. Fischer, Inorg. Chem., 2017, 56, 321–326. 1997, 35, 1551–1570. 17 S. Butorin, D. Mancini, J.-H. Guo, N. Wassdahl, J. Nordgren, 43 W. . Zachariasen, J. Less Common Met., 1978, 62, 1–7. M. Nakazawa, S. Tanaka, T. Uozumi, A. Kotani, Y. Ma, K. Myano, B. Karlin and D. Shuh, Phys. Rev. Lett., 1996, 77, http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Physics arXiv (Cornell University)

Trends in the valence band electronic structures of mixed uranium oxides

Download PDF
4 pages

Loading next page...
 
/lp/arxiv-cornell-university/trends-in-the-valence-band-electronic-structures-of-mixed-uranium-QkYF00A0Sr

References (44)

ISSN
1359-7345
eISSN
ARCH-3341
DOI
10.1039/c8cc05464a
Publisher site
See Article on Publisher Site

Abstract

Trends in valence band electronic structure of mixed uranium oxides a,b *c,d e f Kristina O. Kvashnina* , Piotr M. Kowalski , Sergei M. Butorin , Gregory Leinders , Janne f g c,d f Pakarinen , Rene Bes , Haijian Li , Marc Verwerft a. Rossendorf Beamline at ESRF − The European Synchrotron, CS40220, 38043 Grenoble Cedex 9, France b. Helmholtz Zentrum Dresden-Rossendorf (HZDR), Institute of Resource Ecology, P.O. Box 510119, 01314 Dresden, Germany c. Address here. Institute of Energy and Climate Research, IEK-6, Nuclear Waste Management and Reactor Safety, Forschungszentrum Jülich GmbH, Wilhelm - Johnen-Strasse, 52428 Jülich, Germany d. JARA High-Performance Computing, Schinkelstraße 2, 52062 Aachen, Germany e. Molecular and Condensed Matter Physics, Department of Physics and Astronomy, Uppsala University, SE-751 20 Uppsala, Sweden f. Belgian Nuclear Research Centre (SCK·CEN), Institute for Nuclear Materials Science, Boeretang 200, B-2400 Mol, Belgium g. Department of Applied Physics, Aalto University, P.O. Box 14100, FI-00076 Aalto, Finland * These authors contributed equally. Valence band electronic structure of mixed uranium oxides (UO , U O , U O , U O , -UO ) has been studied by the resonant inelastic 4 9 3 7 3 8 3 X-ray scattering (RIXS) technique at the U M edge and by computational methods. We show here that the RIXS technique and recorded U 5f - O 2p charge transfer excitations can be used to proof the validity of theoretical approximations. Structural, electronic and chemical properties of uranium (U) oxides vary strongly upon a transformation from the fluorite- type UO structures to the layered structure of the higher U 1–13 oxides (U O and above) . The mechanism of the expansion 3 8 14–16 of the fluorite structure is reasonably straightforward , however the role of oxygen (O) atoms in these structural changes remains less clear. We performed the state-of-art valence band RIXS experiment at the U M edge for a number of binary U oxides – UO , U O U O , U O , -UO – in order to 2 4 9, 3 7 3 8 3 clearly identify the mechanism causing the electronic structure modification upon oxidation of UO Valence band RIXS data at the U M edge (~3550 eV) have been previously reported for UO , UO , UF , UO (NO ) *6(H O) 2 3 4 2 3 2 2 17–20 and several U intermetallic systems , and have been proved to be sensitive to the structural environment of U atom and its ligands. Actually, the valence band RIXS data include the elastic and inelastic scattering profiles with an energy resolution of Fig. 1. Valence band RIXS spectra of UO , U O , U O , U O and 2 4 9 3 7 3 8 ~1eV and provide information on the energy difference -UO recorded at the incident photon energy set to the between the valence band states and the unoccupied U 5f maximum of the U M edge. The experimental total energy states. Fig. 1 shows the valence band RIXS spectra of UO , U O , resolution of ~1eV was achieved by employing Johann-type X- 2 4 9 U O , U O and -UO , recorded at the Beamline ID26 of The ray emission spectrometer with five spherically bend Si(220) 3 7 3 8 3 European Synchrotron (ESRF) , (see ESI). The lowest energy crystal analyzers with 1m bending radius (see ESI for details). feature at ~20 eV is attributed to the U 6p -3d 3/2 5/2 17,18 transitions . The process involves first the excitation of an electron from the U 3d core level (at the U M edge) to the 5/2 5 unoccupied U 5f state and then the U 3d core hole is filled by constant increase of the energy separation between the elastic 5/2 an electron from the occupied U 6p state. The inelastic and inelastic scattering contributions in the spectra through the 3/2 scattering profile at energies transfer ~5-10 eV, reported in Fig.1 series of these mixed valence oxides. However, measured valence band RIXS data for the U oxides (Fig.1) show that the mechanism of the electronic structure modification during the transformation of UO into the mixed oxides is however more complicated. Charge transfer also takes place between U sites 2- and additionally incorporated O ions in binary oxides. As a result, the modification of the U-ligand bonding induces a change in the U oxidation state. In addition to that, the position and distribution of valence band states near the Fermi level changes significantly on a scale of several eV. To gain better understanding and to clarify the mechanism of charge transfer excitations and electronic structure modifications we performed three types of calculations with methods of computational quantum chemistry, which have been used to evaluate the properties of U oxides materials 23–30 previously . Computation of mixed U oxides is a challenging task because of a strongly correlated and localized character of the f 23,30 electrons . The commonly used density functional (DFT) methods often fail even on the qualitative level, for instance predicting a metallic state for UO , which is in reality a Mott insulator with a wide band gap of 2.1 eV . To correct for this, the DFT+U method is often used when an on-site Coulomb interaction is modelled by an additional term in the Hamiltonian, whose strength depends on the Hubbard U 23,32,33 parameter . This parameter in calculations for uranium oxides is usually taken as U=4.5 eV (with additional J parameter 30,33 of 0.54eV) or an effective parameter, U =U-J, is applied . eff These values come from the measurements of the correlation 34,35 energy performed on UO . This approximation is the most common approach in the calculations of the electronic structure 10,19,23,25,29,36,37 of U systems . 30,38 Fig. 2. Comparison of measured valence band RIXS spectra of Recently, Beridze and Kowalski performed systematic UO , U O , U O and -UO and results of calculations using a 2 3 7 3 8 3 tests of the performance of the DFT+U method with the variety of methods, denoted as standard approach with U Hubbard U parameter derived ab initio using the linear =4.5eV and J=0.54eV (M1), atomic orbitals (M2), Wannier response method of Cococcioni and Gironcoli for the Functions (M3). calculations of actinide bearing molecular and solid compounds, including U oxides. They have shown that the Hubbard U has been attributed to charge transfer process from the parameter values strongly depend on the oxidation state of U, 18–20 occupied O 2p states to the unoccupied U 5f states . The being largest for U(VI) (~3 eV) and smallest for U(IV) (~2 eV). In energy separation between elastic and inelastic scattering follow up studies a similar trend has been shown for other contributions to the spectra depends of the energy difference actinides . Here we test the performance of this methodology between the occupied O 2p states and unoccupied U 5f states. for the computation of the electronic density of states (DOS) Our recent studies of the evolution of the U chemical state that are used for the construction of theoretical RIXS data. in a series of U oxides confirms a changeover of the oxidation In most of the DFT+U implementations the shape of the states U(IV) - U(V)- U(VI) through the charge compensation orbitals of interest (f orbitals in the case of U) have to be 5,19,22 mechanisms . The established formal oxidation states for U provided in order to estimate the occupation of these orbitals in mixed U oxides are included in brackets: UO (IV), U O (IV-V), 2 4 9 29,33 and compute the Hubbard energy correction term . These U O (IV-V), U O (V-VI), -UO (VI). Moreover the exact 3 7 3 8 3 orbitals are usually represented by the atomic orbitals quantitative analysis has been performed and showed the computed for atoms or ions and thus not necessarily adequately presence of 50% and 50% of U(IV) and U(V), respectively, in represent the shape of the orbital in a solid. In recent studies U O ; 33% and 67% of U(IV) and U(V) in U O and 67% and 33% 4 9 3 7 the maximally localized Wannier functions have been applied in of U(V) and U(VI) in U O . 3 8 32,38–40 computation of electronic structure of solids . In this Based on these findings the process of the electron transfer contribution we will test this approach in order to check if the from the O 2p orbitals to the unfiled U 5f shell should show the Wannier functions-based representation of f orbitals can result COMMUNICATION in any significant improvement in the description of the DOS The issue of the DFT or DFT+U predicted band gaps has been functions and RIXS data. discussed previously . One interesting aspect in the case of UO , In order to reproduce the experimentally detected charge which is often used as a model system, is that for the Mott insulator transfer excitations we made simplified calculations by inserting the the band gap value should be well approximated by the value of the calculated partial U 5f and O 2p DOSs into the Kramers-Heisenberg Hubbard U parameter. The measured band gap for UO is 2.1 eV, 36,41 equation . This approach provides straight forward information which is close to the Hubbard U parameter value predicted by the about the validity or accuracy of DOSs calculated using a variety of linear response method (Table 1). The problem as outlined by methods. It describes a correlation between occupied and Breridze and Kowalski is that the atomic orbitals used to represent unoccupied states under assumption that the hybridization between the f orbitals in solids are not a good representation resulting in a U 5f and O 2p states takes place. In that case, the energy difference significant and unrealistic fractional occupation of the unoccupied f between the maxima of the occupied O 2p DOS and unoccupied U 5f levels (up to 0.3 for UO ). In order to remove this obstacle we applied DOS will define the energy transfer values for the observed RIXS the Wannier representations of the f orbitals for the DFT+U transitions. calculations, which resulted in more realistic, close to zero The partial DOSs have been calculated by three approaches (see occupations of the unoccupied f orbitals. The RIXS profiles resulting ESI). First we have applied the LDA+U approach assuming the from the later calculations are also plotted in Fig. 2. These represent standard values used in calculations for U oxides (U=4.5eV and the best match to the measured RIXS profiles with good prediction J=0.54eV) (denoted as M1). In addition we computed the Hubbard of the position of the U 5f - O 2p charge transfer and an overall much U parameters values using the linear response method of Cococcioni better match to the observed shape of the U 5f – O 2p charge transfer and Gironcoli . Here, for the calculations of the Hubbard correction excitations. we represent f orbitals for projection of occupations by the atomic orbitals (M2) and the Wannier functions (M3). Table 1: The computed Hubbard U parameter values in eV Computation of the electronic structure of U oxides can often U(IV) U(V) U(VI) lead to the convergence to a metastate instead of ground state . In UO 1.7 order to obtain the correct electronic structure of the considered U O 2.1 2.1 3 7 oxides, for an initial electronic state we computed the expected U O 2.0 2.2 3 8 charges of the different U atoms in the mixed U oxides using bond -UO 2.5 valence sum (BVS) method . The BVS of U atoms for the considered Table 2: Predicted and measured band gaps (in eV) of the mixed oxides: U O (P4 /n) , U O (Amm2), UO (P1211), and UO (Fm-3m) 3 7 2 3 8 3 2 U oxides. were calculated and analyzed applying the following formula: M1 M2 M3 Exp 𝑉 = ∑ exp[(𝑅 − 𝑑 ) /𝑏 ] 𝑗 𝑗 (LDA + U) (Atomic) (Wannier) Here the bond valence parameter Rj and constant b are taken UO 2.4 0.3 0.7 2.1 from Ref. V and dj are the corresponding valence and bond lengths U O 1.4 0.6 0.9 1.6 3 7 for each phase. The BVS results are given in ESI, where the U average U O 1.7 0.9 1.3 1.8 3 8 5,22 charge is in approximate agreement with our previous findings . -UO 2.2 2.0 2.7 2.2 The Hubbard U parameter values computed with the linear Table 3: The computed and measured energy difference response for the considered U oxides are given in Table 1. In general, between U 5f and O 2p states. The energies are given in eV as in our previous studies , the value is smaller than 4.5 eV, but M1 M2 M3 Exp strong dependence on the oxidation state is observed. The largest (LDA + U) (Atomic) (Wannier) value was obtained for UO (U(VI)) and the smallest for UO (U(IV)). 3 2 UO 6.5 6.3 7.0 7.5 Fig. 2 shows a comparison of measured and calculated RIXS U O 5.8 5.6 6.5 5.5 3 7 spectra of UO , U O , U O and -UO at the maximum of the U M 2 3 7 3 8 3 5 U O 5.2 5.3 6.1 6.0 3 8 edge, using three approximations (M1, M2 and M3). The elastic -UO 5.2 5.5 6.5 6.2 36 3 scattering contribution has been added to the calculated RIXS spectra to facilitate a comparison with experimental data. The We present here the electronic structure studies of several U standard approach (M1) does not give an ideal match to the mixed oxides. The RIXS spectra shown in Fig. 1 and Fig. 2 indicate the experimental RIXS data (inaccurate energy difference between experimentally obtained energy between the occupied O 2p and elastic and inelastic theoretical RIXS profiles). The calculations with unoccupied U 5f states. In Table 3 we report the difference between the Hubbard U parameter derived ab initio (M2) are even more the average energy of these states as integrated from the computed deviating from the experiment as a result of smaller Hubbard U DOS functions (see ESI) using the three theoretical methods. The best values than the standard one (4.5 eV). The predicted band gaps (see 29,31 match is obtained with the method M3 (with Wannier functions). Table 2) are also smaller than the measured ones . On the other Here the experimentally observed trend is clearly reproduced with hand, the band gaps predicted by M1 method are in qualitative the largest differences for UO and the smallest for U O . These 2 3 8 agreement with the measurements. It is important to note that U 5f differences result from the decrease of the energy of the O p states - O 2p charge transfer excitations, recorded by RIXS in this case, show (with respect to the Fermi level) with increasing the oxidation state the energy difference between two electronic levels (empty U 5f and (due to the stronger electrons binding) and associated decrease of occupied O 2p) and do not directly related to the band gap values the energy of the unoccupied 5f states. The later effect results from obtained by other experimental methods (like optical spectroscopy). higher Hubbard U parameter values (strength of the on-site Coulomb 574–577. repulsion) for higher oxidation states of U (Table1). These interesting 18 S. M. Butorin, J. Electron Spectros. Relat. Phenomena, results indicate that with such a state-of-art experimental method – 2000, 110–111, 213–233. valence band RIXS - one can improve the theoretical prediction of the 19 K. O. Kvashnina, Y. O. Kvashnin and S. M. Butorin, J. electronic structure of actinide contained materials. Electron Spectros. Relat. Phenomena, 2014, 194, 27–36. 20 K. O. Kvashnina, H. C. Walker, N. Magnani, G. H. Lander and R. Caciuffo, Phys. Rev. B, 2017, 95, 245103. Notes and references P.M.K. and J.L. thank the JARA-HPC 21 C. Gauthier, V. A. Solé, R. Signorato, J. Goulon and E. for giving time on the supercomputing resources awarded through Moguiline, J. Synchrotron Radiat., 1999, 6, 164–6. JARA-HPC Partition. S.M.B. acknowledges support from the Swedish 22 K. O. Kvashnina, S. M. Butorin, P. Martin and P. Glatzel, Research Council (research grant 2017-06465) and K.O.K Phys. Rev. Lett., 2013, 111, 253002. acknowledges support from ERC (research grant 759696). 23 X.-D. Wen, R. L. Martin, T. M. Henderson and G. E. Scuseria, 1 G. C. Allen and N. R. Holmes, J. Nucl. Mater., 1995, 223, Chem. Rev., 2013, 113, 1063–1096. 231–237. 24 X.-D. Wen, R. L. Martin, G. E. Scuseria, S. P. Rudin, E. R. 2 J. M. Flitcroft, M. Molinari, N. A. Brincat, N. R. Williams, M. Batista and A. K. Burrell, J. Phys. Condens. Matter, 2013, T. Storr, G. C. Allen and S. C. Parker, J. Mater. Chem. A, 25, 25501. 2018, 6, 11362–11369. 25 S. O. Odoh and G. Schreckenbach, J. Phys. Chem. A, 2010, 3 J. Kretzschmar, T. Haubitz, R. Hübner, S. Weiss, R. Husar, V. 114, 1957–1963. Brendler and T. Stumpf, Chem. Commun., , 26 X.-D. Wen, R. L. Martin, L. E. Roy, G. E. Scuseria, S. P. Rudin, DOI:10.1039/C8CC02070A. E. R. Batista, T. M. McCleskey, B. L. Scott, E. Bauer, J. J. 4 T. Gouder, R. Eloirdi and R. Caciuffo, Sci. Rep., 2018, 8, Joyce and T. Durakiewicz, J. Chem. Phys., 2012, 137, 5 G. Leinders, R. Bes, J. Pakarinen, K. Kvashnina and M. 27 B. Dorado, B. Amadon, M. Freyss and M. Bertolus, Phys. Verwerft, Inorg. Chem., 2017, 56, 6784–6787. Rev. B, 2009, 79, 235125. 6 G. Leinders, J. Pakarinen, R. Delville, T. Cardinaels, K. 28 S. L. Dudarev, D. N. Manh and A. P. Sutton, Philos. Mag. Binnemans and M. Verwerft, Inorg. Chem., 2016, 55, 3915– Part B, 1997, 75, 613–628. 29 H. He, D. A. Andersson, D. D. Allred and K. D. Rector, J. 7 J. M. Elorrieta, L. J. Bonales, N. Rodríguez-Villagra, V. G. Phys. Chem. C, 2013, 117, 16540–16551. Baonza and J. Cobos, Phys. Chem. Chem. Phys., 2016, 18, 30 G. Beridze and P. M. Kowalski, J. Phys. Chem. A, 2014, 118, 28209–28216. 11797–11810. 8 M. Caisso, S. Picart, R. C. Belin, F. Lebreton, P. M. Martin, K. 31 J. Schoenes, J. Appl. Phys., 1978, 49, 1463–1465. Dardenne, J. Rothe, D. R. Neuville, T. Delahaye and A. 32 B. Himmetoglu, A. Floris, S. de Gironcoli and M. Cococcioni, Ayral, Dalt. Trans., 2015, 44, 6391–6399. Int. J. Quantum Chem., 2014, 114, 14–49. 9 V. Kapaklis, G. K. Pálsson, J. Vegelius, M. M. Haverhals, P. T. 33 M. Cococcioni and S. de Gironcoli, Phys. Rev. B, 2005, 71, Korelis, S. M. Butorin, A. Modin, M. Kavčič, M. Zitnik, K. Bučar, K. O. Kvashnina and B. Hjörvarsson, J. Phys. 34 A. Kotani and T. Yamazaki, Prog. Theor. Phys. Suppl., 1992, Condens. Matter, 2012, 24, 495402. 108, 117–131. 10 D. A. Andersson, G. Baldinozzi, L. Desgranges, D. R. 35 Y. Baer and J. Schoenes, Solid State Commun., 1980, 33, Conradson and S. D. Conradson, Inorg. Chem., 2013, 52, 885–888. 2769–78. 36 K. O. Kvashnina, Y. O. Kvashnin, J. R. Vegelius, A. Bosak, P. 11 S. D. Conradson, D. Manara, F. Wastin, D. L. Clark, G. H. M. Martin and S. M. Butorin, Anal. Chem., 2015, 87, 8772– Lander, L. A. Morales, J. Rebizant and V. V Rondinella, Inorg. Chem., 2004, 43, 6922–35. 37 Y. Yun, J. Rusz, M.-T. Suzuki and P. Oppeneer, Phys. Rev. B, 12 R. Zhao, L. Wang, Z.-J. Gu, L.-Y. Yuan, C.-L. Xiao, Y.-L. Zhao, 2011, 83, 1–10. Z.-F. Chai and W.-Q. Shi, CrystEngComm, 2014, 16, 2645. 38 G. Beridze, A. Birnie, S. Koniski, Y. Ji and P. M. Kowalski, 13 G. Leinders, R. Delville, J. Pakarinen, T. Cardinaels, K. Prog. Nucl. Energy, 2016, 92, 142–146. Binnemans and M. Verwerft, Inorg. Chem., 2016, 55, 9923– 39 J. Ma and L.-W. Wang, Sci. Rep., 2016, 6, 24924. 40 D. Novoselov, D. M. Korotin and V. I. Anisimov, J. Phys. 14 L. Desgranges, G. Baldinozzi, G. Rousseau, J.-C. Nièpce and Condens. Matter, 2015, 27, 325602. G. Calvarin, Inorg. Chem., 2009, 48, 7585–92. 41 J. Jiménez-Mier, J. van Ek, D. L. Ederer, T. A. Callcott, J. J. 15 L. Desgranges, G. Baldinozzi, D. Simeone and H. E. Fischer, Jia, J. Carlisle, L. Terminello, A. Asfaw and R. C. Perera, Inorg. Chem., 2016, 55, 7485–7491. Phys. Rev. B, 1999, 59, 2649–2658. 16 L. Desgranges, Y. Ma, P. Garcia, G. Baldinozzi, D. Siméone 42 P. C. Burns, R. C. Ewing and F. C. Hawthorne, Can. Mineral., and H. E. Fischer, Inorg. Chem., 2017, 56, 321–326. 1997, 35, 1551–1570. 17 S. Butorin, D. Mancini, J.-H. Guo, N. Wassdahl, J. Nordgren, 43 W. . Zachariasen, J. Less Common Met., 1978, 62, 1–7. M. Nakazawa, S. Tanaka, T. Uozumi, A. Kotani, Y. Ma, K. Myano, B. Karlin and D. Shuh, Phys. Rev. Lett., 1996, 77,

Journal

Physics – arXiv (Cornell University)

Published: Oct 16, 2020

There are no references for this article.