Thanks to recent developments in hardware and software, quantum chemical methods are increasingly used for interpreting the complex mechanisms underlying polymerization reaction by homogeneous catalysis. Unfortunately, the dimensions of even the smallest realistic models are too large to permit the use of state-of-the-art composite wave function methods. Under these circumstances, density functional theory still offers the best compromise between cost and accuracy. However, comprehensive benchmarks of different functionals are not yet available for this important research field. The main aim of the present paper is to fill this gap by performing an unbiased comparison of several density functionals and continuum solvent models for the stereo-control in the propylene polymerization on prototypical catalysts inducing different reaction mechanisms. While it was not possible to define a unique computational protocol providing the best results in all the situations, the B3PW91 functional in conjunction with D3 empirical dispersions and the solvent model density solvent model performs remarkably well for three out of the four investigated catalysts. Under such circumstances, it is recommended to compare the results delivered by different models when approaching additional classes of catalysts.

中文翻译：

---

预测过渡金属催化剂上的丙烯立体选择性：密度泛函理论的一项艰巨任务

由于硬件和软件的最新发展，量子化学方法越来越多地用于解释均相催化聚合反应的复杂机制。不幸的是，即使是最小的现实模型的尺寸也太大，无法使用最先进的复合波函数方法。在这种情况下，密度泛函理论仍然提供了成本和精度之间的最佳折衷方案。然而，这个重要的研究领域还没有不同功能的综合基准。本文的主要目的是通过对几种密度泛函和连续溶剂模型进行无偏比较来填补这一空白，以在诱导不同反应机制的原型催化剂上进行丙烯聚合的立体控制。虽然不可能定义在所有情况下提供最佳结果的独特计算协议，但 B3PW91 功能与 D3 经验分散体和溶剂模型密度溶剂模型相结合，对于所研究的四种催化剂中的三种催化剂表现得非常好。在这种情况下，建议在接近其他类别的催化剂时比较不同模型提供的结果。