Embedded Files
Asymmetric organocatalysis: Brønsted acid and Lewis acid catalysis
Computational chemistry, machine learning, and automated synthesis for asymmetric catalysis
1. Asymmetric activation of inert functional groups
While a series of hydrocarbons are obtained from crude oil as feedstock chemicals, their direct transformation into valuable materials remains underdeveloped. For instance, alkanes are undeniably a fundamental and ubiquitous class of compounds in chemistry; however, their asymmetric transformation via C–C bond activation has remained challenging due to the absence of heteroatoms, which are essential for facilitating efficient interactions between the substrate and the catalyst. Inspired by the extraordinary ability of enzymes to catalyze the asymmetric functionalization of hydrocarbons, we designed highly acidic and confined catalysts capable of promoting asymmetric "cracking" reactions of simple cyclopropanes, yielding the corresponding alkenes as products in excellent enantioselectivities.
2. Understanding microenvironments
Owing to bulky substituents of our catalysts, their catalytic active sites often located in a narrow pocket; the substrates need to fit into this confined chiral microenvironment to undergo the transformation, and this is the key to achieve high enantioselectivities with simple substrates. Using quantum chemical techniques, we try to understand the origin of enantioselectivities and analyze the nature of the privileged catalysts. Recently, we have found that the key to control the IDPi-catalyzed syn– or anti-selective asymmetric Mukaiyama-aldol reactions is an unconventional C–H hydrogen bonding of the CF2H group, which controls the size of the catalytic cavity. Further evaluations of subtle electronic and steric effects are ongoing.
3. Designing highly enantioselective catalysts
Catalyst optimization processes typically rely on the inductive and qualitative assumptions of chemists based on screening data. Although these “educated guesses” of chemists somewhat work, we aim to go beyond and rationally design more selective catalyst structures by developing methodologies to predict the enantioselectivities of catalysts. Recently, we developed a predictive model using 2D fragment descriptors, which accurately predicted highly enantioselective catalysts from training data including only moderately selective catalysts. Even more exciting chemistry is in progress.
Page updated
Google Sites
Report abuse