Premium
Classifications of families of homologous organic compounds based on energy by means of ab initio HF, MP2 and DFT optimizations
Author(s) -
Neugebauer Alexander,
Häfelinger Günter
Publication year - 2002
Publication title -
journal of physical organic chemistry
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.325
H-Index - 66
eISSN - 1099-1395
pISSN - 0894-3230
DOI - 10.1002/poc.535
Subject(s) - homologous series , chemistry , ab initio , molecule , homologous chromosome , basis (linear algebra) , conjugated system , computational chemistry , quantum , quantitative structure–activity relationship , thermodynamics , stereochemistry , crystallography , organic chemistry , physics , quantum mechanics , polymer , mathematics , geometry , biochemistry , gene
A method is suggested for combining the quantum chemical principle of isolated molecules and the systematics of homologous families for different functional groups in traditional organic chemistry. The classification of organic compounds is performed by statistical treatments of fundamental quantum mechanical quantities such as total and partial energies. We used the total energies (computed by DFT BLYP and B3LYP, HF and MP2 methods, which are size‐extensive, applying the 6–31 G** and some other basis sets) of isolated molecules obtained by gradient optimizations to classify homologous series of organic compounds by linear least‐squares regressions versus the total number of electrons. Two numerical parameters, slope ( h ) and intercept ( s ), are obtained characterizing each family of homologous compounds. Families of for example 17 different functional groups of the same CH 2 ‐homologous unit are grouped on parallel lines with the same slope h , differing only in their s parameters. Isomeric compounds differ in their s values if they belong to the same homologous unit. Compounds of other homologous units (conjugated alkenes, cumulenes and benzenoides) are found with different slopes h on regression lines. Energy partitioning leads also to linear regressions of different statistical accuracy for the components which add fully to the observed h and s parameters. The h and s parameters allow the prediction of unknown total energies of homologous compounds from total energies of a subset of smaller compounds which belong to the same homologous family. Likewise, atomization energies can be predicted using the derived h and s parameters. Copyright © 2002 John Wiley & Sons, Ltd.