W0148

The Definition of a Single Chemical Bond through Structure Correlation with Physical and Computational Data. Ronald F. See, Anthony D. Dutoi, Kevin W. McConnell, Seetharaman Santhanam and William M. Strub, Department of Chemistry, Saint Louis University, St. Louis, MO 63103 USA

The concept of the single bond (and through it, bond order) is key to almost all discussions of structural chemistry, but a quantified, and uniformly applicable, definition of this entity remains elusive. In particular, such a definition would prove very useful in the study of main-group molecular compounds which may be hypervalent. The Bond Valence Model, in the parameter dij, provides such a definition of the single bond, and values have been collected (O'Keeffe, M.; Brese, N.E.J. Am. Chem. Soc. 1991, 113, 3226-9) to allow dij to be calculated formost atom pairs. Unfortunately, their remain many questions as to the suitability of these values for the study of discrete molecular compounds. To answer these questions, the re-evaluation of the O'Keeffe & Brese parameters was undertaken for the main-group, non-metallic elements. This was done through structure-correlation techniques, using structural data for the Cambridge Structure Database and gas electron diffraction, physical data such as IR force constants and bond enthalpies and computational data from ab initio molecular orbital calculations. These results indicate that the O'Keeffe & Brese parameters were generally accurate, except in the case of oxygen, where a longer value is consistently indicated. The value of the constant b, used in the Bond Valence Model, was also re-calculated; a value of ~0.32 was found, closer to the 0.31 suggested by Pauling than 0.37, which has been used in more recent studies. Implications of these findings on the present theoretical conceptions of bonding in main-group elements will also be discussed.