Abstract
The Dirac vector model, as considered by Van Vleck, has been applied to calculate the energy values of configurations of the type dkms and dkmsns, in cases where the dk core conforms to Russell-Saunder's coupling. The spin-orbit energy is not assumed small compared to the electrostatic energy coupling dk to ms or dk to ns. Johnson's formulas for the four vector problem have been used to set up a general secular determinant for configurations of the type dkmsns. The special cases where dkmsns may be treated as a three vector problem or as a case of Russell-Saunder's coupling are also considered. The following multiplets have been examined: Y I, 4d(D2)5s6s; Cu I, 3d9(D2)4s5s; Zr I, 4d2(F3)5s6s; Ni I, 3d8(F3)4s5s; Co I, 3d7(F4)4s5s; Fe I, 3d6(D5)4s5s; and Ti I, 3d2(F3)4s6s. The exchange integrals are considered as parameters and chosen to give the best fit with experimental data. The agreement with experiment for dkmsns is better, as it ought to be, with the four vector problem (Ld, Sd, sm, sn) than with the three vector problem (Ld, Sd+m, sn) which in turn agrees better with experiment than does the two vector problem (Ld, Sd+m+n), the case resulting from the assumption of Russell-Saunder's coupling. Configurations dkms (other than Houston's singlet-triplet case) have been examined as a three vector problem (Ld, Sd, sm). Some examples treated are: Ti II, 3d2(F3)5s; Ni II, 3d8(F3)5s; Co I, 3d8(F3)5s; Co I, 3d8(F3)4s; Pd II, 4d8(F3)6s; Fe I, 3d7(F4)5s; Co II, 3d7(F4)5s.