Ground-state actinide chemistry with scalar-relativistic multiconfiguration pair-density functional theory

Abstract

We have examined the performance of Multiconfiguration Pair-Density Functional Theory (MC-PDFT) for computing the ground-state properties of actinide species. Specifically, we focused on the properties of UN2 and various actinyl species. The properties obtained with MC-PDFT at the scalar-relativistic level are compared to Kohn-Sham DFT (KS-DFT); complete active space self-consistent field theory, CASSCF; coupled-cluster theory, CCSD(T) and CCSDT; as well as multireference perturbation theory (CASPT2). We examine the degree to which MC-PDFT improves over KS-DFT and CASSCF while aligning with CASPT2, CCSD(T), and CCSDT. All properties that we considered were for the CASPT2 electronic ground states. For structural parameters, MC-PDFT confers very little advantage over KS-DFT, especially the B3LYP density functional. For NpO23+, MC-PDFT and local KS-DFT functionals excessively favor the bent structure, whereas CCSDT and CASPT2 predict the bent and linear structures as isoenergetic. For this special case, hybrid KS-DFT functionals like PBE0 and B3LYP provide results closer to CASPT2 and CCSDT than MC-PDFT. On a more positive note, MC-PDFT is very close to CASPT2 and CCSD(T) for the redox potentials, energetics of redox chemical reactions, as well as ligand-binding energies. These are encouraging results since MC-PDFT is more affordable. The best MC-PDFT functional is ft-PBE. Our findings suggest that MC-PDFT can be used to study systems and excited states with larger strong electron correlation effects than were considered here. However, for the systems and properties considered here, KS-DFT functionals do well, justifying their usage as the bulwark of computational actinyl chemistry over the last two to three decades.