Can you break the oxo-wall? A multiconfigurational perspective

Abstract

The concept of the “oxo-wall” was conceived about 60 years ago by Harry B. Gray, and has been found to be related to the non-existence of high-valent M–oxo species in the +IV oxidation state in a tetragonal geometry beyond group 8 in the periodic table. Several efforts have been made in the past decades to test and find examples that violate this theory. Several claims of violation in the past were attributed to the difference in the geometries/coordination number and, therefore, these are not examples of true violation. In recent years, substantial efforts have been undertaken to synthesise a true Co^IVO species with various ligand architectures. Co^IVO and Co^III–O˙ are electromers and, while they are interchangeably used in the literature; the former violates the oxo-wall while the latter does not. The possibility that these two species could exist in various proportions similar to resonating structures has not been considered in detail in this area. Furthermore, there have been no attempts to quantify such mixing. In this direction, we have employed density functional theory (DFT) and ab initio CASSCF/NEVPT2 methods to quantify the co-existence of Co^IVO and Co^III–O˙ isomeric species. By thoroughly studying six different metal-oxo species, we affirm that the nature of such electromeric mixing is minimal/negligible for Fe^IVO and Mn^IVO species – both are pre-oxo-wall examples. By studying four different ligand architectures with Co–oxo species, our results unveil that the mixing of Co^IVO ↔ Co^III–O˙ is substantial in all geometries, with dominant Co^IVO species favourable for the S = 3/2 intermediate spin state. The percentage of the Co^III–O˙ species is enhanced substantially for the S = 1/2 low-spin state. We have attempted to develop a tool to estimate the percentage of the Co^III–O˙ species using various structural parameters. Among those tested, a linear relationship between % of Co^III–O˙ and a bond length based ratio is found (, where d_(Co–O) and d_(Co–Nax) are the axial Co–O and Co–N_ax bond lengths in Å, respectively). It is found that the higher the R_d, the greater the Co^III–O˙ character will be and the geometrically portable correlation developed offers a way to qualitatively compute the % of Co^III–O˙ character for unknown geometries.