Effective Group Potentials. 2. Extraction and Transferability for Chemical Groups Involved in Covalent or Donor−Acceptor Bonds

The effective group potential (EGP) methodology developed in the first article is supported by the idea that some chemical properties of a molecule depend only on a few nuclei and electrons. This technique, which allows one to reduce the number of electrons and nuclei explicitly implied in an ab initio calculation, can be applied to systems which can be separated into an active part and some spectator groups. Chemical groups involved in covalent or donor−acceptor types of bonding have been studied. For example, the silyl group, SiH<INF>3</INF>, is replaced by a silicon pseudoatom with only one active electron, and the associated EGP is designed for properly taking into account the electronic effects of the whole fragment on the neighboring chemical group. Three other molecular groups have been replaced by an EGP, namely PH<INF>3</INF>, NH<INF>3</INF>, and C<INF>5</INF>H<INF>5</INF>. The latter EGP is designed for a suitable description of the interaction of the π system of the cyclopentadienyl molecule with a metallic atom. The transferability of the EGPs and their usefulness for theoretical calculations on realistic cases are also discussed.