The constrained geometry alkoxide [Ti(III)(η5-C5Me5){η5-C5Me4CH2CMe2O-κO}] (2) and titanocene alcoholates [Cp*2TiOR] for R = 1-adamantyl (3) and pentafluorophenyl (4) were prepared by reacting acetone and the respective alcohols with the singly tucked-in titanocene [Ti(III)(η5-C5Me5)(η5:η1-C5Me4CH2)] (1) in order to add the simplest member to constrained geometry alkoxides 5 and 6 and to extend the series of [Cp*2TiOR] (R = Me (8), tBu (7), Ph (9), and H (10)) compounds for one of the most bulky and the most electron attracting substituent. The titanocene alkoxides with the alkoxy carbon atom tethered to one of the cyclopentadienyl ligands showing the crystallographic Ti−O−C angle close to 133° are a suitable group of compounds to be compared with the [Cp*2TiOR] compounds, where the Ti−O−C angle is about 174° for 3 and 7 and 180° for 4. In spite of these Ti−O−C angle differences, the wavelength of the first electronic transition λexp(1a1 → b2), supposed to indicate the extent of the oxygen π-electron donation into the Ti−O bond (as found by Andersen et al. J. Am. Chem. Soc. 1996, 118, 1719−1728), was slightly higher for 2 (1390 nm) compared to 3 (1335 nm) or 7 (1300 nm). Since the Ti−O−C carbon atom has a similar environment in all these compounds (three methylenes for 3, three methyl groups for 7, and one tethered methylene and two methyl groups for 2), the batochromic shift of λexp(1a1 → b2) for 2 could be attributed to a smaller extent of the oxygen π-electron donation effect. These differences in λexp are, however, very small and hardly can be correlated with the crystallographic Ti−O bond lengths, whose differences of max. 0.02 Å can be attributed also to the unequal steric congestion in 2, 3, and 7. On the other hand, the electron-attracting C6F5 group in 4 decreased the oxygen π-electron donation considerably, as indicated by the large bathochromic shift of λexp(1a1 → b2) to 1950 nm and the discernible elongation of d(Ti−O)exp and shortening of d(O−C)exp distances. These observations are in full agreement with our previous experimental and theoretical study (Varga et al. Organometallics 2009, 28, 1748) concluding that an increasing polarity of the Ti−O bond decreases the oxygen π-donation. The time-dependent DFT calculations of λcalc(1a1 → b2) could follow the trend λexp(1a1 → b2) (4 ≫ 9 > 10 > 3 > 7 > 8), although with the theoretical values biased systematically toward higher energies by about 200 nm. This shift has been ascribed to the insufficient treatment of electron correlation by DFT, resulting in the overestimation of interelectron repulsion. Complete active space SCF studies for [Cp*2TiOMe] (8) involving both the ground and the first excited state gave an excellent agreement of λcalc(1a1 → b2) with λexp(1a1 → b2) and proved no involvement of other orbitals in this transition. The DFT study on [Cp*2TiOH] (10) covering the range of linear-to-perpendicular Ti−O−H angles established the energy minimum for the Ti−O−H angle at 118.4° and the OH group situated in the Cg,Ti,Cg plane, in excellent agreement with the single-crystal model. The difference in the overall delocalization energies Edeloc(O→Ti) for the bent and the linear arrangement was, however, small in comparison with variations of Edeloc(O→Ti) in [Cp*2TiOR] compounds induced by the nature of substituents R.
Read full abstract