Optimal Descriptor Based QSPR Models for Catalytic Activity of Propylene Polymerization

Abstract

A heterogeneous Ziegler–Natta (ZN) catalyst is an important catalyst in the field of the polypropylene polymerization industry. The role of electron donors has been crucial in the ZN catalyzed polypropylene polymerization process. In this article, quasi-SMILES-based QSPR models are elaborated for the prediction of catalytic activities. The representations of the molecular structure by quasi-simplified molecular input line entry system were the basis to build the desired QSPR model. These models were developed by means of the Monte Carlo optimization involving the available methods classic scheme (CS), balance of correlations (BC) and balance of correlation with ideal slopes (BCIS). The best QSPR model showed r2 = 0.813 (for external validation set), rm2 (avg)=0.73 and ∆rm2= 0.03.