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.

中文翻译：

---

半二茂钛配合物的分子场分析：数据驱动的单中心烯烃聚合催化剂的计算机优化计算研究

我们使用半茂钛催化烯烃聚合的计算数据进行了分子场分析。采用DFT方法计算的乙烯插入、丙烯插入的活化能以及乙烯插入与β-氢转移之间的能量差作为回归分析的目标变量。从相应的过渡态结构计算出的分子场（体素数据）被用作描述符。基于分子场回归分析的可视化结构信息提供了催化剂设计指南。按照指南设计的磷酰亚胺催化剂显示出增强的计算自由能值。根据之前的报告，与训练样品中的高性能膦酰亚胺催化剂相比，设计的催化剂在乙烯聚合中表现出更高的活性和聚合物分子量。