A theoretical investigation of the bonding of early transition metals to tellurium

Non-relativistic and relativistic discrete variational Xα calculations have been performed on the model complexes [W(PH3)4E2](E = O, S or Te). The results are consistent with the formal description of the compounds as eighteen-electron tungsten(IV) d2 systems. Metal–chalcogen σ bonding becomes increasingly covalent as the chalcogen is altered from O to Te, while the almost equal tungsten and chalcogen contributions to the π levels remain more constant. Ligand-based spin–orbit coupling effects are substantial in [W(PH3)4Te2] and smaller for [W(PH3)4S2] and [W(PH3)4O2]. Mixing of the σ/π character of the non-relativistic 3a1(W–Te σ) and 3e (Te 5pπ lone pair) molecular orbitals of [W(PH3)4Te2] occurs in the relativistic calculations. A similar effect is seen in the 1e (W–O π) and 2a1(W–O σ) levels of [W(PH3)4O2]. The theoretical results are in good agreement with the experimental electronic absorption spectra of [W(PMe3)4E2](E = S, Se or Te). The bonding in the model complex [Zr(TeSiH3)4] is discussed in conjunction with the closely related ZrI4. The separation of the metal–ligand σ and π levels is significantly greater in the tellurium compound. Previous predictions for the spin–orbit splitting of the t2(σ) and t2(π) subshells of Group IVA tetrahalides are reproduced more closely in [Zr(TeSiH3)4] than in ZrI4, in which there is substantial t2(σ)–t2(π) mixing. Attempts to assign the valence photoelectron spectra of MX4(M = Ti, Zr or Hf; X = Br or I) in terms of the earlier theoretical model are considered to be inappropriate in light of the lack of t2(σ)–t2(π) separation. Comparison of the metal–tellurium bonding in [W(PH3)4Te2] with that in [Zr(TeSiH3)4] suggests that the former is significantly more covalent.