Fragmentation into Atoms

For its fragments, the CLOGP method uses a basis set of "isolating" carbon atoms (aliphatic and aromatic taking different values), attached hydrogen atoms, and a variety of "polar" fragments, again having different values for an attachment to an aromatic or aliphatic structure. Of course, fragmentation of a molecule is somewhat arbitrary, and there are advantages and disadvantages of any fragmentation scheme. Fragments larger than a single atom can be selected, so that significant electronic interactions are contained within one fragment, and this is perceived as the main advantage of using fragments larger than single atoms. The advantage of using an atomic fragmentation approach is that ambiguities are avoided, but a disadvantage is that a very large number of atom types are needed to describe a reasonable range of molecules, unless atomic charges are calculated to distinguish between various electronic forms of the same, or similarly hybridised, atom. In 1984, Broto et al. [34], in 1986 Ghose and Crip-pen [35], and in 1989 Viswanadhan et al. [36] implemented atomic level schemes of additivity. Ghose and coworkers extended and refined their atom values in 1987, and suggested that fewer atom types may be needed, provided that a separate parameter for atomic charge were to be included [37]. They suggested that use of a molecular orbital, CNDO/2 calculation, may be appropriate.

Atom level fragment schemes work well in many instances, but a common shortcoming, pointed out by Leo, is the failure to deal with long-range interactions such as found in p-nitrophenol [32],

0 0

Post a comment