CORRELATED AND GAUGE ORIGIN INDEPENDENT CALCULATIONS OF MAGNETIC-PROPERTIES .1. TRIPLY BONDED MOLECULES

Citation
Spa. Sauer et al., CORRELATED AND GAUGE ORIGIN INDEPENDENT CALCULATIONS OF MAGNETIC-PROPERTIES .1. TRIPLY BONDED MOLECULES, Molecular physics, 81(1), 1994, pp. 87-118
Citations number
53
Categorie Soggetti
Physics, Atomic, Molecular & Chemical
Journal title
ISSN journal
00268976
Volume
81
Issue
1
Year of publication
1994
Pages
87 - 118
Database
ISI
SICI code
0026-8976(1994)81:1<87:CAGOIC>2.0.ZU;2-F
Abstract
We have applied a gauge origin invariant method for the calculation of magnetizabilities and nuclear magnetic shielding constants to the tri ply bonded molecules CO, N-2, HCN, CN- and HCCH. In particular, we hav e compared the basis set dependence of the sum-over-states expression for the diamagnetic term with the ground state average value expressio n and with the basis set dependence of the paramagnetic contribution. A systematic procedure for choosing an atomic basis set for the indivi dual atoms was developed and applied to all molecules studied. Inclusi on of p and d functions with large exponents was found important for t he sum-over-states diamagnetic contribution to the nuclear magnetic sh ielding, whereas it was necessary to include diffuse d functions for t he magnetizability. The effect of electron correlation on the diamagne tic and paramagnetic contributions to the magnetic properties was inve stigated within the second order polarization propagator approximation (SOPPA) and various coupled cluster polarization propagator approxima tions (CCDPPA/CCSDPPA). We find that the SOPPA gives a much smaller co rrelation contribution for the heavy atoms in these molecules than com parable MP2 and MC-RPA calculations. A large effect on the inclusion o f coupled cluster single amplitudes suggests that the orbital relaxati on terms might be quite important for magnetic properties of triply bo nded molecules. Correlation effects are larger in small basis sets. Th is leads sometimes (N-2 and CO) to much better agreement between theor y and experiment using incomplete basis sets.