ORGN 12 |
Semi-empirical quantitative structure selectivity relationships: Accurate and precise enantiomeric excess predictions for chiral catalysts
| James Ianni, Venkatachalam Annamalai, Puay Wah Phuan, Manoranjan Panda, and Marisa C Kozlowski. Department of Chemistry, University of Pennsylvania, 231 S 34th St, Philadelphia, PA 19104-6323 |
| Semi-empirical quantum mechanical quantitative structure selectivity (QSSR) relationships have been developed for accurate and precise enantiomeric excess predictions of chiral catalysts using the program QMQSAR. In particular, the structures of various beta-amino alcohol catalysts have been correlated to their enantioselectivities in the asymmetric addition of diethylzinc to benzaldehyde. Beginning with a set of known catalysts (training set), the QMQSAR program generates multilinear regression models correlating calculated variables to the known enantioselectivities. The variables are the interaction energies calculated between a carbon 2s orbital and the training set's molecular orbitals. These calculated variables are searched via a simulated annealing approach to create the multilinear regression models. For a leave-out-two cross-validation of an 18 beta-aminoalcohol training set, an overall r2 of 0.85 was obtained, indicating a highly predictive model. With this QMQSAR method the selectivities of new catalysts were calculated. Subsequent chemical synthesis and analysis of the new catalysts indicated that the model was very useful and easily distinguished catalysts of low (33% ee), moderate (59% ee), and high (98% ee) selectivity. Also, the first reported case of a QSSR extrapolation from a training set containing only low enantiomeric excess catalysts to a prediction set of high enantiomeric excess catalysts will be presented. 
|
| |
|
|