Files
Abstract
I have applied density functional theory (DFT) to various problems in organic and
inorganic chemistry. The use of various DFT levels has rigorously been tested to assess the reliability of these methods to accurately compute structures and energies. The B3LYP DFT theory with a reasonably large basis sets is sufficiently accurate. In few cases, however, B3LYP leads to serious errors in predicting the ground state structures. Work employing the B3LYP method to investigate the reaction of phosphaethene with 1,3-dienes revealed that the structures of the reactants and products and their energies were sufficiently accurate to explain the experimental regioselectivites. Additionally the B3LYP method also accurately reproduce the 1H NMR chemical shifts. In the study regarding the stability of [4n]annulenes (antiaromatic hydrocarbons), the use of an appropriate isodesmic equation can give B3LYP stabilization energies which are in excellent agreement with those
given by other methods. In these cases, results from B3LYP-DFT were sufficiently accurate and reliable. Computationally challenging for all the DFT methods was the prediction of structures and energies of [4n+2]annulenes. The widely accepted X-ray and B3LYP indicated CC bond equalized structures of [14]-, [18]-, and [22]annulenes are found to be incorrect since the computed proton chemical shifts using these geometries gave 1Hs which are in gross disagreement with the experimental values. Instead lower symmetry structures with CC bond alternation are given by KMLYP and BHLYP DFT methods. These lower symmetry structures are found to be more stable, as also confirmed by the CCSD(T) method, and give proton chemical shifts in excellent agreement with experimental values. Our application of the B3LYP level to various problems indicates that this method, in general, is reliable for computing accurate structures and energies. However, in a few cases caution is indicated when applying the B3LYP method, this study suggests that the KMLYP is capable of predicting structures more reliably than the B3LYP method. We also propose that data sets containing aromatic annulenes (larger than benzene) could serve as the basis for
developing even better hybridizations of density functional and other methods.