| Abstrakt: |
In order to determine the optimal methodology for evaluation of the magnitude of intermolecular charge transfer, several methods have been examined: Mulliken population, natural population analysis, atoms in molecules (AIM) as well as charges from electrostatic potentials using a grid-based method (Chelp and Chelpg) procedures using a series of correlation consistent cc-pVXZ (X = D, T, Q) basis sets within LCAO MO SCF, MP2, DFT, and coupled cluster theory levels. In contrast to previous nonconclusive comparative studies, the present calculations reveal close matching of the recently available experimental data for six Lewis acid-base adducts with theoretical values derived from the Chelpg approach, whereas for the remaining methods relative errors are almost doubled. On the other hand, AIM and Chelpg results display the best linear correlation coefficients with the experimental data. Since reasonably accurate Chelpg results could be already obtained with SCF or DFT B3LYP methods using cc-pVDZ, such an approach opens the way to study intermolecular charge transfer in larger molecular systems. Preliminary results obtained within cc-pVDZ basis set and B3LYP functional for pyridine-SO[SUB3] complex do not exceed relative error limits observed for other smaller complexes. Analysis of corresponding interaction energy components calculated consistently in the dimer basis set indicates significant role of electrostatic, exchange and delocalization contributions, with rather negligible correlation term. In contrast to previous findings, the experimentally observed amount of transferred charge seems not to correlate with any interaction energy term. [ABSTRACT FROM AUTHOR] |
