A scheme for calculating nuclear magnetic shieldings using the infinite-order Foldy-Wouthuysen (FW) transformation proposed by Barysz and Sadlej (BS) is presented. The nuclear magnetic shieldings of hydrogen halides are calculated by three variant BS schemes; a double finite perturbation method for the external magnetic flux density (B 0 ) and the nuclear magnetic dipole moment ( M ) (BS/FPT-2), a single finite perturbation method for B 0 with analytical differentiation of energy with respect to ( M ) (BS/FPT-1), and an approximate analytical differentiation method with respect to both B 0 and M (BS/CHF). Although the BS/FPT-2 method is exact theoretically, the actual computation for heavy nuclei includes large error due to reduction of the number of significant figures. The BS/FPT-1 and BS/CHF approaches, on the other hand, yield reasonable values for all of the shieldings. Although several results could not be obtained by the BS/FPT-2 method, no serious contradictions were recognized among these three results. From a comparison of our results with values in the literature, our shieldings for the halogen nuclei are lower than those determined by the four-component relativistic random phase approximation (4-RPA), but the reason for this is not obvious.
IntroductionFully relativistic treatments [1, 2], based on the fourcomponent spinor form of the one-electron wave functions, are still very costly, primarily due to the cost for a proper description of the small two-component spinor. In most relativistic quantum chemistry problems, however, explicit consideration of the negative energy states or the electron-positron pair creation processes is unnecessary. For example, the information included in the four-component Dirac equation [3] is excessive for the majority of problems encountered in relativistic quantum chemistry. Thus, the investigation and development of more appropriate two-component methods is particularly attractive. Many methods have been proposed to reduce the four-component Dirac formalism to computationally much simpler two-component schemes. Two groups of approximate two-component methods have been successful thus far; methods based on the Douglas-Kroll-Hess (DKH) approach [4][5][6][7][8] and those based on the regular Hamiltonian approximation (RA) [9][10][11][12][13][14][15][16].A powerful approach to decoupling positive energy states (electronic states) and negative energy states