Electronic structure of f(1) lanthanide and actinide complexes .2. Non-relativistic and relativistic calculations of the ground state electronic structures and optical transition energies of [Ce(eta-C5H5)(3)], [Th(eta-C5H5)(3)] and [Pa(eta-C8H8)(2)](2)

Non-relativistic and relativistic discrete variational-X alpha calculations have been performed on [Ce(eta-C5H5)(3)], [Th(eta-C5H5)(3)] and [Pa(eta-C8H8)(2)]. Metal-ligand covalent interactions in [Ce(eta-C5H5)(3)] and [Th(eta-C5H5)(3)] are dominated by metal d-orbital participation in the (eta-C5H5) pi(2)-based 2e and 3e molecular orbitals, and an f-orbital contribution to the 1a(2) level. The non-relativistic calculations predict formal ground configurations of 4f(1) and 5f(1) for [Ce(eta-C5H5)(3)] and [Th(eta-C5H5)(3)] respectively, while the destabilisation of the nf and slight stabilisation of the (n + 1)d sigma level on the incorporation of relativistic effects result in a 6d(1) electronic configuration for [Th(eta-C5H5)(3)]. [Pa(eta-C8H8)(2)] is found to have a 5f(1) ground configuration in both non-relativistic and relativistic calculations, and both the 6d and 5f metal orbitals participate significantly in metal-ligand covalent bonding. The 4f-based molecular orbitals of [Ce(eta-C5H5)(3)] are found to be little altered from those of free Ce(III), and show a three-below-four (j = 5/2 below j = 7/2) spin-orbit coupling pattern. The greater radial extension of the 5f atomic orbitals and their closer energy match with vacant carbocyclic ring levels destroys this splitting pattern in [Th(eta-C5H5)(3)] and [Pa(eta-C8H8)(2)]. The electronic absorption spectra of all three molecules have been calculated using the transition state method. In [Ce(eta-C5H5)(3)] and [Pa(eta-C5H5)(3)], the transitions are principally f --> f in nature, although a low-lying f --> d sigma shift is predicted in both cases. In [Ce(eta-C5H5)(3)], the calculated f --> d sigma transition is in good agreement with that found experimentally for [Ce(eta-C5H3(SiMe(3))(2))(3)]. The calculated spectrum of [Th(eta-C5H5)(3)] consists of a series of d --> f transitions, in agreement with the intense peaks observed in the experimental spectrum of [Th(eta-C5H3(SiMe(3))(2))(3)]. Comparison of the calculated transitions of [Pa(eta-C8H8)(2)] with the experimental spectrum of [Pa(eta-C8H4(CH3)(4))(2)] suggests that the latter is due to charge transfer transitions and that the f --> d sigma shift has not yet been experimentally observed.