Theoretical Approaches to the Vibrational Spectra of Polyatomic Molecules

Abstract

Computational chemistry can make useful contributions to other branches of chemistry by supplementing infrared and Raman spectroscopy as a source of information about the mechanical flexibility of molecules. A major problem in doing so is the difficulty in getting sufficient accuracy and reliability in theoretical prediction of vibrational force fields, anharmonicity, and spectral frequencies. Two methods for solving this problem are outlined.

First, advantage can be taken of the remarkable uniformity in the computational error for a given type of vibration (e.g., C—H bond stretch, heavy atom bend, etc.) in different molecular environments. Similarly, the contribution of anharmonicity for a given type of motion is highly transferable. This permits development of a small set of scaling factors which can be used as transferable corrections between known and unknown molecules.

Alternatively, a very high-level quantum chemical computation of the vibrational potential energy hypersurface followed by a variational solution of the vibrational Schrödinger equation can yield highly accurate anharmonic frequency predictions for small molecules. Advantages and limitations of the two procedures are discussed.