| Described here are chiral Ln3+ complexes of Schiff-base ligands and their use as catalysts for the kinetic resolution of the 4-nitrophenyl esters of N-Boc glutamine and phenylalanine through an enantioselective methanolysis/ethanolysis reaction. Catalysts were screened in various solvents and temperatures to optimize the selectivity value k2L/k2 D or k2D/k2L. The greatest selectivity at 25°C determined was 7.1, however, reducing reaction temperature increases selectivity and after kinetic resolution at -40°C a product mixture with >99% e.e. with ~80% of the desired product remaining.;The detailed mechanism of base-promoted hydrolysis and methanolysis of O,O'-dimethyl O-aryl phosphorothioates and phosphates was examined computationally. Generally, with increased leaving group acidity, the reactions showed a tendency toward being concerted (a single transition state) and while substrates with leaving groups with lower acidity tend to involve 5-coordinate phosphorane intermediates. Bronsted beta lg values are also computed, and found to be similar to experimentally determined values.;The mechanism of palladacycle-promoted phosphorothioate methanolysis is also modeled computationally, showing a change in rate determining step from substrate binding ( pssKa of leaving group phenol 13). The calculations also predict the existence of a stable metal-bound 5-coordinate thiophosphorane intermediate which has been tentatively verified experimentally by others. The preliminary results of a computational investigation into the mechanism for the methanolytic cleavage of a series of N-methyl N-aryl thiobenzamides catalyzed by a similar palladacycle complex is also presented, showing metal-assisted nucleophilic attack and leaving group departure.;A computational method is presented for the prediction of the Bronsted parameter betaeq, for a phosphorothioyl group transfer in methanol by comparing relative free energy values of starting materials and phenolate products for a series of related substrates. Direct comparison of starting material and anionic leaving group free energies results in a poor approximation of betaeq, however given that the beta eq value for the equilibrium of phenol and phenoxide is -1, the comparison of computed free energies of neutral starting material with neutral phenol product provides betaeq approximations that are satisfactorily close to the experimentally derived values. |
