Modeling actinides with electronic structure theories
is challenging
because these systems present a strong ligand field and metal–ligand
covalency. We systematically investigate the effectiveness of pair-density
functional theory (PDFT) for the calculation of vertical excitation
energies in An(III), [AnIIICl6]3–, and [AnVIO2]2+ (An = U, Np, Pu,
and Am). We compare the performance of PDFT, hybrid PDFT, and multistate
PDFT with traditional active-space methods followed by perturbation
theory, like multistate CASPT2, and with experimental data. Overall,
multistate PDFT gives quantitative agreement with multistate CASPT2
at a significantly reduced computational cost.