Further improvements in a model for 5d electrons in heavy rare earth metals and compounds

Abstract

A model for 5d electrons in the heavy rare earth metals which takes into account 4f-5d interactions in more detail than most theories is improved in four ways. Firstly, a simpler and more accurate method is found for obtaining a relation between the rectangular equivalents of true 5d subband widths in the model and effective subband widths for calculation of subband occupation numbers when d-d interactions are neglected. Secondly, the effects of d-d Coulomb interactions on the choice of best local basis states to use in the model are considered. Thirdly, more reliable values of Slater parameters are used for the (4f) n 5d(6s) 2 configurations of terbium and dysprosium. Fourthly, a more transparent procedure is found for obtaining approximate solutions of a set of coupled equations determining subband occupation numbers and other parameters. These changed procedures bring about only small alterations in predictions of the model for most properties. Large 5d electric quadrupole and octupole moments are still predicted for rare earths other than gadolinium. However, because of the new method of calculating the best local states, 4f magnetic moments are not reduced so much below their free-ion values as in the 1975 version of the model. Brief comments are made on the applicability of a similar model to the RAl 2 compounds (R ≡ rare earth).