View article

With the results of a multitude of numerical tests of the Hohenberg-Kohn-Sham local spin density (LSD) approximation (Hohenberg et al., 1964, Kohn et al., 1965) in hand, there is a growing consensus amongst researchers concerning the long term viability of the LSD approximation over other ab initio techniques for applications to large molecules. Coupled with the ever increasing numerical data base of results is an enhanced formal understanding of the relative attributes and shortcomings of the existing approximations to the universal functional. The consensus that has emerged is that the existing LSD approximation is expected to give very accurate information about ground state molecular geometries (eg bond lengths and angles), molecular vibrational phenomena, and qualitatively correct energetics. While improvements of LSD based molecular energetics is currently a field of active research, there is …