Development and testing of a protocol for computational prediction of 1H and 13C NMR chemical shifts and thermochemistry and reaction analysis of benzyne formation and trapping

Abstract

Elucidating structures of novel compounds and investigation of new reactions are two tasks that experimental organic chemists address on a frequent basis. The pursuit of these objectives can be rigorous and time-consuming. Of the methods employed in elucidating the structure of novel compounds, nuclear magnetic resonance (NMR) is by far the most widely applied. Investigation into new reactions may require any number of techniques to understand the reaction scope, kinetics, optimal conditions, mechanisms, etc. In both cases, the use of computational methods is well-suited to augment the experimentalist's data to guide and understand the system being investigated. A protocol for facilitating computational prediction of NMR chemical shifts was developed. Application to a set of natural products previously evaluated against computed NMR shifts, showed improved accuracy, through analysis of the corrected mean-absolute error (CMAE). The protocol was further employed successfully to aid in analysis of experimental spectra for compounds synthesized by collaborators where multiple diastereomers were possible. Graphing templates were also created to allow for rapid inspection of possible structures without more in-depth statistical analysis. Thermodynamic and mechanistic analysis on the formation and reaction of benzyne was also performed. Thermodynamic restrictions on the ring-size of fused benzynocycloalkanes were investigated. Additionally, analysis of the energetics and transition state geometries for small-molecule trapping (both intra and intermolecular) of benzyne are discussed.