Catalytic Performances of Carbon-Based Tetrel Bond Catalysis by Density Functional Theory and Machine Learning.

Abstract

The non-covalent interaction catalysis has been widely developed and applied in organocatalysis for its green and economical characteristics. In recent years, the carbon-based tetrel bonds have been successfully applied in organocatalysis. In this work, the structure-property relationship of carbon-based tetrel bond catalysts is established by utilizing density functional theory (DFT) and machine learning (ML). Taking the Michael addition reaction as an example, the reaction mechanism and new insight into the catalytic active site are introduced based on the DFT-calculated results. The bowl-like 1,1-dicyano-3,3-dicarbonylcyclopropane (DCDC) unit based σ-hole is much more important than the 1,1,2,2-tetracyanocyclopropane (TCCP) unit in tetrel bond catalysts. Introducing electron-withdrawing groups into catalysts significantly enhances the catalytic activity and provides a new strategy for designing efficient catalysts. Furthermore, the construction and evaluation of ML models demonstrate their potential in predicting the catalyst performance, offering a new protocol for fast prediction of the catalytic performance of tetrel bond catalysts.