[Wien] Orientation of Atomic Spin Moments

[Chiung-Yuan Lin]

Thanks for the comments. Let me back up a little bit from the
ambitious orientation thing. In stead, I would like to study the two
magnetic atoms (the unit cell contains
many other non-magnetic atoms; these 2 magnetic atoms are close to
each other and well separated from the magnetic ones in the next unit
cell) using the Heisenberg-like model: H= J  S1 . S2  ------ So my 2
important questions will be:  1) How do I figure out the atomic spin
Sa? I have guessed of using n(up) - n(dn)
as S but I don't understand why it works this way. Sa*(Sa+1) is the
eigenvalue of the atomic spin operator S^2, where S is the sum of
electron spin (vector) operators.
2) Is there a way to calculate J without going to noncolinear?

Chiung-Yuan