Molecular field analysis in half-titanocene complexes: computational study towards data-driven in silico optimization of single-site olefin polymerization catalysts

Abstract

We performed molecular field analysis using computed data of half-titanocene-catalyzed olefin polymerization. The activation energies of ethylene insertion, propylene insertion, and the energy differences between ethylene insertion and β-hydrogen transfer calculated with DFT methods were employed as target variables for regression analysis. Molecular fields (voxel data) calculated from corresponding transition-state structures were used as descriptors. The structural information visualized based on the molecular field-based regression analysis provided a catalyst design guideline. A phosphinimide catalyst designed following the guideline showed enhanced computed free energy values. According to a previous report, the designed catalyst exhibited higher activity and polymer molecular weight in ethylene polymerization compared to a high performance phosphinimide catalyst in the training samples.