Lindsay Roy

Quantum chemistry approaches to actinide electronic structure

The corrosion reactions of uranium and transuranic metals represent a challenge for both experimentalists and theorists because of their chemical complexity, while presenting a serious concern for environmental preservation. The long-term storage of radioactive waste (containing mainly uranium and plutonium compounds) requires a comprehensive knowledge of the possible oxidation reactions that could occur. For example, plutonium crystallizes in seven distinct phases, is highly reactive, and can have five different oxidation states when forming compounds. The surface of metallic plutonium easily oxidizes to PuO₂ when exposed to air and moisture, while Pu₂O₃ plays an important role in oxidation kinetics.

Given the incomplete understanding of the plutonium-oxidation chemistry and the myriad problems associated with experimental studies on plutonium oxides, theory can provide fundamental insight into the electronic structure and properties of these systems.

Unfortunately, theoretical studies on these materials are quite difficult. The 5f electrons can either be localized or contribute to bonding, and their relativistic effects and electron–electron correlations are very important factors in deciding the degree of localization. In particular, actinide oxides (AnO₂) exhibit amazingly complex behavior despite having a simple binary formula because they straddle the metallic, ionic, and covalent bonding descriptions used by chemists and physicists. Because of this, previous theoretical studies on these materials using local spin-density approximation (LSDA) and generalized gradient approximations (GGAs) of density functional theory fail to accurately describe the bonding and electronic structure of these compounds.

Recently, Richard Martin of Los Alamos’ Theoretical Chemistry and Molecular Physics Group, Gustavo Scuseria of Rice University, and their coworkers have shown that a third generation of functionals, the hybrid density functional theory approximation that combines the exact, non-local, Hartree-Fock exchange interaction with the traditional local (LDA) or semi-local (GGA) exchange and correlation interactions, has been able to correctly predict a magnetic ground state, an insulating gap, and the lattice constants for the uranium oxide (UO₂) series.

Continuing the research, we have shown that the 5f orbitals are localized in the band structure of UO₂ but that there is small dispersion of approximately 300 millielectronvolts (meV). One might expect later members of the series to show even more localized character because the f-orbital radial extent decreases moving across the row. Interestingly, a 5f–O₂p orbital energy degeneracy occurs, which leads to significant orbital mixing and covalency in the intermediate region (PuO₂–CmO₂), and ground state in curium oxide (CmO₂) prefers a half-filled f ⁷ subshell (Cm³⁺) instead of the expected f ⁶ configuration (Cm⁴⁺). Preliminary results of the band structure of PuO₂ reveal that there is some amount of mixing between the 5f and 2p bands. It will be interesting to see how the f–p mixing changes in CmO₂, where the oxygen 2p atoms are donating spin density to the 5f to stabilize a half-filled subshell.

One caveat to previous calculations is the omission of spin-orbit coupling, or the interaction of the electron spin magnetic moment with the magnetic moment due to the orbital motion of the electron, which is significant for heavy atoms. Calculations of UO₂ with spin-orbit coupling show that the band gap decreases by 0.06 eV, but that spin-orbit coupling is a minor perturbation to the calculations. We are currently working on including spin-orbit coupling effects to the AnO₂ series.

Calculations of the uranium oxide (UO₂) band structure (large image) were performed on an antiferro-magnetic tetragonal cell (inset). The red arrows in the inset represent the direction of the 5f² electrons on each uranium atom. The oxygen p bands are well separated from the uranium f bands, and the occupied 5f bands show some dispersion and have a bandwidth of approximately 300 millielectronvolts.

Next: Manufacturing of Mixed-Oxide Fuels

Lindsay Roy works in the Theoretical Division’s Theoretical Chemistry and Molecular Physics Group (T-12). She received her doctorate from Texas A&M University under the direction of Timothy Hughbanks, where her research focused on developing theoretical models for understanding the magnetic interactions in lanthanide-based molecules, clusters, and solids. She began her postdoctoral appointment in November 2006. Her mentor is Richard Martin of T-12.