PuO₂ and Fuel Residues

Of the world’s c. 250 tonnes of separated Pu, >100 tonnes are stored at Sellafield in the UK as PuO2 powder in sealed steel cans. Under certain circumstances, gas generation may occur in these cans, with consequent pressurization. Many routes to gas production have been suggested, several of which involve PuO2/H2O interactions, and all of which are complex, inter-connected and poorly understood.

In light of this, we have recently begun a computational study of the interaction of AnO2 (An = U, Np, Pu) surfaces with water. Typically, such studies are performed using density functional theory (DFT) in conjunction with periodic boundary conditions (PBC). However, it is well known that standard PBC implementations of DFT using generalized gradient approximation (GGA) functionals often fail to reproduce key features of actinide solids, e.g. predicting metallic properties in systems known to be insulating. This failure stems from incorrect description of the strongly correlated 5f electrons, which are overly delocalized by the GGA, and the standard solution to this problem is to correct the GGA functionals with an onsite Coulomb repulsion term known as the Hubbard U.

A more elegant solution is to employ hybrid DFT, in which a certain amount of the exact exchange energy of Hartree Fock theory is incorporated into the Hamiltonian. Such functionals typically produce more localized 5f electrons, and recover insulator behaviour. They are, however, extremely expensive to employ in PBC calculations, and are very rarely used in the calculation of actinide solids. We have therefore sought a computational model which allows the routine use of hybrid DFT in AnO2/water systems, and are developing an approach based on the periodic electrostatic embedded cluster method [1], in which a quantum mechanically treated cluster is embedded in an infinite 1-, 2- or 3-dimensional array of point charges. This approach allows us to treat a cluster of AnO2 and adsorbing water molecules using hybrid DFT (PBE0) whilst the long-range electrostatic interactions with the bulk are modelled via the embedding of charges.

[1] A. M. Burow et al. J. Chem. Phys. 130 174710 (2009)