Is this Antiferro- or Ferrim-magnetic soloution-Again

[vaspman]

Hello! this is my second time to ask this problem
I am trying to do some test calculations on DMS. I have one question about the Antiferromagnetic (AFM) calculations. When I use MAGMOM = 1 -1 X*, where X are some non magnetic atoms in the super cell , the total magnetic moment of the unit cell is not zero but it should be zero in principle. My first question is, why the net magnetic moment of the unt cell is not zero when don’t used NUPDOWN=0 , does it mean that AFM converged to ferrimagnetic (FIM) state?
Then I read the Forum and I found that we can also use NUPDOWN=0 and in this case the total magnetic moment of the unit cell is indeed zero. I don’t want to impose further constrain on the direction of the magnetic moment i.e. want to use collinear magnetism only. When we use NUPDOWN=0 , this means we are forcefully impose that the solution should me AFM but in some cases the AFM converges to FIM ( some DMS shows such FIM behavior where AFM converges to FIM during SCF calculations) and FIM is the ground sate. So my question is if some material has FIM behavior and when we impose AFM constrain with NUPDOWN=0 then it means we are bypassing the FIM behavior. Am I right?
Please explain me.
Thank you

[admin]

please first of all check if the calculation really is converged (rms(c) in OSZICAR should be about 1e-3 or less). If it is, compare the total energies of the FIM output and one with enforced AFM (NUPDOWN=0, if the calculation does not converge to the AFM result spontaneously). The ground state is the one with the lower total energy.

[ Edited Mon Oct 08 2007, 09:32AM ]