| Summary: | Quantum chemistry deals, inter alia, with the determination of NMR parameters for prediction or explanation of physical effects. Most methods use non-relativistic approaches, which lead to errors. Those errors occur not just for heavy elements, but also for the NMR parameters of light nuclei close to heavy elements. The influence of such effects is called, e.g. SO-HALA effect. In this thesis, NMR parameters of light nuclei in the presence of heavy nuclei are calculated. Different quantum chemistry packages, approaches and implementations are used and compared. For testing, benchmark molecules are assembled from the p-block elements. Central point of this work is the validation of available ORCA algorithms. ORCA[1] is a modern quantum chemistry package and developed especially for spectroscopic applications. It is found that the ZORA implementation in ORCA shows the same results as the ZORA implementation in NWChem. A pilot implementation RIZORA provides the same results while eliminating the grid dependency of ZORA. However, a a-posteriori spin-orbit (SO) approach in ORCA does not give the same results as two- or four-component calculations in DIRAC. Since spin-orbit coupling is essential for the description of the SO-HALA effect, the use of the BSS Hamiltonian in DIRAC currently offers a good solution. In this work, NMR shielding constants of bismuth-aryl compounds with halides are investigated, because NMR experimental spectra showed an Inverse Halogen Dependence (IHD) of these compounds. This means that the chemical shift of the ortho proton to the bismuth atom increases with increasing nuclear charge of the halides. This effect has not yet been explained or simulated in the literature. It is surprisingly found that the IHD is not caused by a HALA effect, while NMR parameters are drastically effected by geometry changes.
|
|---|