Exploring metastable states in UO2 using hybrid functionals and dynamical mean field theory

Laura E Ratcliff,Hyowon Park,Peter B Littlewood,Alejandro Lopez-Bezanilla,Luigi Genovese

doi:10.1088/1361-648x/ac3cf1

Abstract

A detailed exploration of the f-atomic orbital occupancy space for UO2 is performed using a first principles approach based on density functional theory (DFT), employing a full hybrid functional within a systematic basis set. Specifically, the PBE0 functional is combined with an occupancy biasing scheme implemented in a wavelet-based algorithm which is adapted to large supercells. The results are compared with previous DFT + U calculations reported in the literature, while dynamical mean field theory is also performed to provide a further base for comparison. This work shows that the computational complexity of the energy landscape of a correlated f-electron oxide is much richer than has previously been demonstrated. The resulting calculations provide evidence of the existence of multiple previously unexplored metastable electronic states of UO2, including those with energies which are lower than previously reported ground states.

Highlights

Decades after efforts were devoted to unveiling the physical properties of UO2, this f-electron oxide remains an active and exciting material of study in condensed matter physics.Uranium is the most studied of the actinide elements, but its successful industrial use as a fissile material has not run in parallel to the refinement of computational tools suitable for computational modeling of its properties.The oxide obtained upon combining U with O in a tetravalent bonding, UO2, is the most widespread uranium mineral
Previous investigations of the ground state of UO2 using hybrid functionals [11, 12, 17,18,19] have been limited to small supercells, while the only detailed exploration of the occupancy space relied on the use of the “exact exchange for correlated electrons” (EECE) approximation to PBE0 [12]
Such an approach resulted in a significantly lower band gap than for previous calculations with full PBE0 – 2.0 eV for the lowest energy state found with EECE vs. 3.1 eV for full PBE0 [17], this discrepancy might be due to other differences in computational parameters, including the employed basis set, use of effective core potentials vs. the projector augmented wave (PAW) approach [20], as well as differences in k-point sampling and initial guess

Summary

Introduction

Decades after efforts were devoted to unveiling the physical properties of UO2, this f-electron oxide remains an active and exciting material of study in condensed matter physics.Uranium is the most studied of the actinide elements, but its successful industrial use as a fissile material has not run in parallel to the refinement of computational tools suitable for computational modeling of its properties.The oxide obtained upon combining U with O in a tetravalent bonding, UO2, is the most widespread uranium mineral. Decades after efforts were devoted to unveiling the physical properties of UO2, this f-electron oxide remains an active and exciting material of study in condensed matter physics. Uranium is the most studied of the actinide elements, but its successful industrial use as a fissile material has not run in parallel to the refinement of computational tools suitable for computational modeling of its properties. The ground state of the tetravalent U ion in UO2 (the oxidation state is +4) exhibits a population of 2 electrons in the seven-fold degenerate 5 f orbitals. As with the other first elements of the actinide series, the strongly correlated U 5 f states are at the boundary between simple localized and itinerant pictures of electrons, with f -orbitals partially occupied with electrons strongly localized close to the atomic core, and yielding a small crystal field splitting. The shielding of the U 5 f orbitals by the s, p, and d-orbitals in UO2 is not as large as in the lanthanide series, which allows valence electrons to interact weakly with the O p-orbitals forming the ligand environment

Methods

Findings

Conclusion