Premium
Electron density redistribution in the stabilization of a molecular stacking complex: The nature and correction of basis set superposition errors
Author(s) -
Osman Roman,
Topiol Sid,
Weinstein Harel
Publication year - 1981
Publication title -
journal of computational chemistry
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.907
H-Index - 188
eISSN - 1096-987X
pISSN - 0192-8651
DOI - 10.1002/jcc.540020113
Subject(s) - counterpoise , basis set , stacking , chemistry , valence (chemistry) , ab initio , atomic physics , redistribution (election) , molecular physics , computational chemistry , density functional theory , physics , organic chemistry , politics , political science , law
The effect of the basis set superposition error ( BSSE ) on the calculated electronic structure of a molecular stacking complex is analyzed with the counterpoise correction method. The complex between para ‐hydroxyaniline (PHA) and formamidinium cation (FAM) is calculated ab initio with the STO ‐3G and Whitman–Hornback minimal bases, and with two split valence basis sets: 4‐31G and STO ‐3G(D). When the counterpoise correction is applied, the charge redistribution in the PHA/FAM complex calculated with all four basis sets suggests that the complex is electrostatic in nature and that the main polarization is from the PHA toward FAM. The FAM cation is polarized away from the intermolecular region, thus causing further increase in the electrostatic interaction. This picture is not evident with the STO ‐3G related bases if the counterpoise correction is not applied. Thus, the BSSE in the charge redistribution is shown to be particularly large in the STO ‐3G basis and in the diffuse, split valence, STO ‐3G(D). Both these basis sets suffer from an inappropriate description of the core region. Where there is an improved description of the electron density in the core region, as in the calculations with the energy‐optimized Whitman–Hornback basis and with the 4‐31G basis, the counterpoise correction has only a very small effect on the charge redistribution. After the counterpoise correction is applied, the two minimal basis sets yield nearly identical charge redistribution results, as do the two split valence bases. It is therefore suggested that basis sets used in the calculation of molecular complexes might be classifiable according to criteria such as the degree of contraction and the quality of the description of angular polarization. This would help in the comparative evaluation of results obtained with different bases since minimal basis sets (or split valence bases) would become directly identifiable as groups yielding qualitatively similar results.