Relativistic DFT calculations of magnetic shielding tensors for spin ½ heavy nuclei
University of Delaware
A computational study of magnetic-shielding tensors of spin-½ heavy nuclei in solids has been carried out by employing relativistic DFT and cluster models. The performance of various theoretical treatments and cluster models has been investigated by comparing the agreement between theory and experiment as a metric of the goodness of the calculation. ☐ A significant amount of effort in this study has been dedicated to the development of cluster models for accurate calculation of magnetic-shielding tensors in the solid state. The performance of cluster models of various sizes, symmetries, as well as clusters with different net charges and with preparation by different truncation methods have been studied. The convergence of calculated principal components with cluster size is monitored in benchmark calculations. The results suggest that inclusion of higher coordination shells in the molecular cluster is generally necessary for quantitative predictions of magnetic-shielding tensors. However, it has been found possible to reduce the size of these computationally expensive molecular-cluster calculations with limited effect on the calculated NMR parameters by carefully introducing the frozen core approximation and locally dense basis sets. ☐ For network solids, a new formalism, which employs pseudo-atoms with altered nuclear charges and parameters obtained from bond-valence theory, is proposed for the truncation of clusters. This model has been applied to a large selection of systems with success. The performance of the cluster models in network solids is also compared to models that account for the full translational symmetry of the extended system (periodic boundary conditions with GIPAW) for lighter nuclei such as 29Si and 31P. ☐ The importance of treating a system with the relativistic Hamiltonian for accurate prediction of principal components of the magnetic-shielding tensor of heavy nuclei (207Pb, 199Hg, 125Te and 119Sn) is demonstrated within the cluster approach. The results demonstrate that inclusion of the spin-orbit component in the ZORA Hamiltonian is essential to obtain good agreement with experimental results. It is shown that spin-orbit effects on the principal components are strongly dependent on the oxidation state and coordination geometry about the NMR nuclei. ☐ Finally, the performance of hybrid functionals (B3LYP and PBE0) is examined for the prediction of magnetic-shielding tensors of 207Pb, 125Te and 119Sn. The results show that employing hybrid functionals improves agreement between theory and experiment, compared to GGA functionals. This improvement is more noticeable in the case of 207Pb than it is for 125Te and 119Sn.
Pure sciences , DFT , Magnetic shielding , Nuclear magnetic resonance , Relativistic effects , Spin-orbit coupling