Metrical Oxidation States of 1,4-Diazadiene-Derived Ligands

Abstract

The conventional method of assigning formal oxidation states (FOSs) to metals and ligands is an important tool for understanding and predicting the chemical reactivity, in particular, in catalysis research. For complexes containing redox-noninnocent ligands, the oxidation state of the ligand can be ambiguous (i.e., their spectroscopic oxidation state can differ from the FOS) and thus frustrates the assignment of the oxidation state of the metal. A quantitative correlation between the empirical metric data of redox-active ligands and their oxidation states using a metrical oxidation state (MOS) model has been developed for catecholate- and amidophenoxide-derived ligands by Brown. In the present work, we present a MOS model for 1,4-diazabutadiene (DADⁿ) ligands. This model is based on a similar approach as reported by Brown, correlating the intra-ligand bond lengths of the DADⁿ moiety in a quantitative manner with the MOS using geometrical information from X-ray structures in the Cambridge Crystallographic Data Center (CCDC) database. However, an accurate determination of the MOS of these ligands turned out to be dependent on the coordination mode of the DAD^2- moiety, which can adopt both a planar κ²-N₂-geometry and a η⁴-N₂C₂ π-coordination mode in (transition) metal complexes in its doubly reduced, dianionic enediamide oxidation state. A reliable MOS model was developed taking the intrinsic differences in intra-ligand bond distances between these coordination modes of the DAD^2- ligand into account. Three different models were defined and tested using different geometric parameters (C═C → M distance, M-N-C angle, and M-N-C-C torsion angle) to describe the C═C backbone coordination with the metal in the η⁴-N₂-C₂ π-coordination mode of the DAD^2- ligand. Statistical analysis revealed that the C═C → M distance best describes the η⁴-N₂-C₂ coordination mode using a cutoff value of 2.46 Å for π-coordination. The developed MOS model was used to validate the oxidation state assignment of elements not contained within the training set (Sr, Yb, and Ho), thus demonstrating the applicability of the MOS model to a wide range of complexes. Chromium complexes with complex electronic structures were also shown to be accurately described by MOS analysis. Furthermore, it is shown that a combination of MOS analysis and FOD calculations provides an inexpensive method to gain insight into the electronic structure of singlet spin state (S = 0) [M(trop₂dad)] transition-metal complexes showing (potential) singlet biradical character.

Figure 1

(a) Accessible oxidation states of the diazabutadiene ligand framework. Neutral diimine (left), one-electron reduced semi-iminato (middle), and fully reduced enediamide (right) forms. (b) κ²-N₂ (left) and η⁴-N₂C₂ (right) binding modes of the fully reduced enediamide form. R = H, CH₃.

Figure 2

Metal complexes included in the training data set containing 1,4-diazabutadiene ligands highlighted in blue.

Figure 3

DAD bond distances as a function of reported DAD ligand oxidation states. Exact values of the C–C and C–N bond lengths are shown as a function of literature-reported ligand oxidation state.

Figure 4

Distribution of calculated MOSs for the initial data set without separate treatment of the enamide κ²-N₂-DAD^2– and η⁴-N₂C₂-DAD^2– binding modes as (a) histogram and (b) box plot (MOS = 0: range between +0.3 and −0.4; MOS = −1: range between −0.6 and −1.3; MOS = −2: a broad range between −1.5 and −2.6, split into two smaller distributions; one distribution between −1.5 and −2.0 and another between −2.0 and −2.6).

Figure 5

(a) Distribution of calculated MOSs using the C=C → M distance as a weight factor to describe enamide κ²-N₂-DAD^2– and η⁴-N₂C₂-DAD^2– binding modes. (b) Fitted distributions of the C=C → M distance in η⁴-N₂C₂-coordinated and κ²-N₂-coordinated structures.

Figure 6

Sr, Ho, and Yb complexes containing the DADⁿ moiety.^, For the holmium complex, the siloxane cluster is not shown (dipp = 2,6-diisopropylphenyl; thf = tetrahydrofuran). The calculated MOSs are listed next to the ligand, with the model error in brackets. Crystallographic inequivalent entries were calculated separately, and their MOS values were averaged and the error in the model was pooled.

Figure 7

(a) Homoleptic chromium(II) DAD complex [Cr(^dippDAD)₂][Li(thf)₄] best described as the ligand oxidation state DAD^1.5–. (b) Homoleptic Th(IV) DAD complex [Th(^MesDAD)₂I] best described as having one semi-iminato and one enediamide ligand.