(653c) Strong Influence of the Nucleophile on the Rate and Selectivity of 1,2-Epoxyoctane Ring-Opening Catalyzed By Tris(pentafluorophenyl)Borane

Tris(pentafluorophenyl)borane, B(C₆F₅)₃ (or BCF), is a strong Lewis acid with high thermal stability, water tolerance and chemical versatility, and has been utilized as a catalyst or an initiator for a wide variety of reactions. One industrially relevant example is the role of BCF in the catalytic, regioselective ring-opening (RO) polymerization of terminal epoxides, as these reactions are essential for the formation of the polyether polyol intermediates required to produce polyurethanes, polyethylene glycols, and polypropylene oxides. While a wide variety of catalysts have been used in epoxide RO polymerization, they are all either non-selective or preferentially activate the unencumbered carbon atom, leading to the formation of a secondary alcohol product. However, BCF and other aryl boranes can achieve high levels of primary alcohol selectivity for the RO of aliphatic epoxides. In this work, we use density functional theory (DFT) calculations and microkinetic modelling to investigate the mechanisms of BCF-catalyzed 1,2-epoxyoctane RO by primary and secondary alcohol nucleophiles, i.e. 1-propanol and 2-propanol. We develop a reaction network consisting of three major catalytic pathways by which RO could occur: a Lewis acid pathway, a water-mediated pathway, and an alcohol-mediated pathway, where each pathway corresponds to the different BCF adducts formed under reaction conditions. The model developed captures the notable differences in experimental behavior between 1-propanol and 2-propanol and is applicable to a wide range of reaction conditions. For each alcohol nucleophile, the observed sensitivity of regioselectivities and rates to reaction conditions (water concentration, temperature, and extent of conversion) were caused by different speciation of BCF adducts, and different amounts of flux through each pathway. More broadly, we also consider whether machine learning algorithms can be used in the prediction of selectivities and rates for other aryl boranes.