The role of the ligand field was also evaluated using the Condon-Slater parameters for interelectronic repulsion at CAS(7,7)SCF level though the so-called ab initio ligand field theory (AILFT). As was pointed out recently by Neese et al. , NEVPT2 results are not relevant to the bonding analysis and frequently introduce large errors.6 As shown in Table 3 for both molecules, the ratio Fk/Fkfree-ion(k =2, 4, 6) < 1 with variations over 3%, which implies a nephelauxetic effect responsible for the reduction of the electron-electron repulsion due to the strong interaction of the 5f orbitals with the carbonate ligands. LFDFT yields similar parameters for the free-ion (Table 3), however they differ significantly when the molecular structure is considered, i.e. Fk and ζ  are reduced compared to AILFT ; with reductions ca. 44% with respect to the free-ion. This shows a more effective penetration of the ligand’s electron cloud into that of the metal compared to AILFT6. The difference could reside in the localized description of f-electrons under HF and post-HF methods compared to the well-known overdelocalization of DFT; though, this could be argued against when using hybrid functionals as in this case. Carnall and coworkers have obtained these parameters from BkF4experimental spectra.10,11 These parameters are larger than those obtained by LFDFT, but in agreement with the nephelauxetic series. This could prove the effect of covalency on the reduction of the electrostatic repulsion in actinides.
The analysis of the optimized molecular orbitals shows an important bonding interaction between 5f and ligands orbitals. The active behavior of the ligand orbitals was corroborated through analysis of the occupation numbers after a careful inclusion of the symmetry adapted orbitals.
Table 3. One-electron Slater-Condon Fk  (k = 2, 4, 6) and effective spin-orbit coupling parameters