Electron correlation effects on uranium isotope fractionation in U(VI)-U(VI) and U(IV)-U(VI) equilibrium isotopic...

Uranium isotope fractionation has been extensively investigated in the fields of nuclear engineering and geochemical studies, yet the underlying mechanisms remain unclear. This study assessed isotope fractionations in U(VI)-U(VI) and U(IV)-U(VI) systems by employing various relativistic electron correlation methods to explore the effect of electron correlation and to realize accurate calculations of isotope fractionation coefficients (ε). The nuclear volume term, ln K_nv, the major term in ε, was estimated using the exact two-component relativistic Hamiltonian in conjunction with either HF, DFT(B3LYP), MP2, CCSD, CCSD(T), FSCCSD, CASPT2, or RASPT2 approaches for small molecular models with high symmetry. In contrast, chemical species studied in prior experimental work had moderate sizes and were asymmetrical. In such cases, electron correlation calculations other than DFT calculations were not possible and so only the HF and B3LYP approaches were employed. For closed-shell U(VI)-U(VI) systems, the MP2, CCSD and CCSD(T) methods yielded similar ln K_nv values that were intermediate between those for the HF and B3LYP methods. Comparisons with experimental results for U(VI)-U(VI) systems showed that the B3LYP calculations gave results closer to the experimental data than the HF calculations. Because of the open-shell structure of U(IV), multireference methods involving the FSCCSD, CASPT2 and RASPT2 techniques were used for U(IV)-U(VI) systems, but these calculations exhibited instability. The average-of-configuration HF method showed better agreement with the experimental ε values for U(IV)-U(VI) systems than the B3LYP method. Overall, electron correlation improved the description of ε for the U(VI)-U(VI) systems but challenges remain with regard to open-shell U(IV) calculations because an energy accuracy of 10^-6-10^-7E_h is required for ln K_nv calculations.