The reactions of Ti(NtBu)(Me3[9]aneN 3)Me+ (1) with a series of polar and nonpolar unsaturated organic substrates give different kinds of products, resulting either from cycloaddition to the Ti=N bond (PhCCH) or from insertion into the Ti-Me bond (PhCCPh, iPrNCNiPr, ethylene polymerization). DFT(B3PW91) calculations of the reaction pathways for cycloaddition and insertion on small model systems illustrate the basic electronic properties in terms of reaction site selectivity. In all cases the cycloaddition is shown to be the preferred pathway kinetically. ONIOM(B3PW91:HF) calculations on the experimental systems allow the evaluation of the influence of the steric bulk of the ligand set on the outcome of the reaction. Introduction of the actual ligands results essentially in a destabilization of the cycloaddition pathway and leads both to a cancellation of the kinetic preference for cycloaddition and to a reduced thermodynamic stability of the cycloaddition product. Only for the terminal alkyne PhCCH does cycloaddition remain the preferred pamway. For all the other substrates, insertion and cycloaddition are kinetically competitive, but insertion is thermodynamically much favored. In the case of Ti(N tBu)(Me3[9]aneN3)Cl+ (2), where insertion is not possible, the calculations show that cycloaddition is endothermic with respect to coordination, explaining the experimental results (coordination for iPrNCNiPr and absence of reaction for PhCCPh and C2H4). Only in the case of PhCCH is cycloaddition favored with 2. ONIOM calculations on a Zr analogue of 1 show that, for the larger metal, the steric bulk does not destabilize the cycloaddition pathway enough to make insertion competitive. © 2008 American Chemical Society.
