Geometry prediction of bridged zirconocene dichlorides by quantum chemical methods

The ab initio Hartree–Fock theory has been demonstrated to give accurate geometry predictions for bridged zirconocene dichlorides. Equilibrium geometries of crystallographically characterized bridged zirconocene dichlorides were optimized by Hartree–Fock, MP2, BLYP, and B3LYP methods, with basis sets ranging from 3-21G* to 6-311G**. Selected geometrical parameters were compared with experimental crystal structures. The least expensive HF/3-21G* method proved to be notably accurate. The accuracy of HF/3-21G* was verified by a comprehensive data set of 62 bridged zirconocene dichlorides. Furthermore, experimental corrections were applied to the optimized geometry parameters to eliminate systematic deviations. Corrections resulted in considerably improved accuracy for systematically overestimated metal–ligand distances, with maximum deviation falling from 0.081 to 0.039 Å, and absolute average deviations from 0.048 to 0.012 Å. Ligand–metal–ligand angles were predicted accurately with absolute average deviations of 0.7–1.3°. Zirconium–chlorine distances and chlorine–zirconium–chlorine angles are relatively constant in the studied molecules. Zirconium–cyclopentadienyl distances can be influenced mainly by modifying the ligand structure, whereas cyclopentadienyl–zirconium– cyclopentadienyl angles and cyclopentadienyl–cyclopentadienyl plane angles can be controlled by bridge modifications. The HF/3-21G* method can be applied for the estimation of steric effects in zirconocene catalyzed polymerization reactions, therefore being suitable for the construction of structure–polymerization property correlations.