Evidence for a Heisenberg Type of Exchange Interaction in Nickel and the NiCu Alloy System

The Curie-temperature and average-magnetic-moment data for the NiCu alloy system are considered in terms of a generalized Heisenberg exchange model in which, as pointed out by Van Vleck, only $3d^{10}$ and $3d^9$ states are involved. On this model the magnetic energy per spin should be a linear function of the average moment per atom, and the data bear out this relation, when a small correction is made for fluctuation effects in the magnetically dilute alloys. It is difficult to see how the data can be fitted on the basis of the commonly made assumption that the ferromagnetic energy arises primarily from the Hund's-rule coupling of two electrons simultaneously occupying orbitals on the same atom, $3d^8$ states.