The default direction of spin in collinear calculations.

Message

Neutrino · #1 Post by **Neutrino** » Wed Jan 22, 2014 4:23 am

I compiled parallel VASP with the tag NGZhalf and expected the default spin in collinear calculations to be aligned along z-axis but I keep obtaining it aligned along x-axis. Is there a way to get aligned along z-axis?
Thank you in advance.

#2 Post by **admin** » Tue Feb 04, 2014 6:55 pm

the default direction for spins is z if you set ISPIN=2 (collinear FM or AFM spin alignment). Please don't be confused by the line
magnetization (x)
in OUTCAR when listing the local magnetic moments (this is the default 'header' of this OUTCAR section for spin-polatized runs).
any other spin directions can only be obtained if you do non-collinear calculations.

Neutrino · #3 Post by **Neutrino** » Thu Feb 13, 2014 12:54 am

Thank you for your reply, admin.
However, there is one test (Collinear) I performed which indicates that spin is actually in x. The test is on UO2 in fluorite structure. The experimental magnetic structure is AFM in <111>. However, in a previous Phys. Rev. B, it was shown that a collinear calculation with AFM aligned in [001] will lead to shortening the [001] vector compared to [010] and [100].
I did this test myself and I get the [100] shortened instead of [001]. I appreciate your input regarding this. Thank you!

Neutrino · #4 Post by **Neutrino** » Thu Feb 13, 2014 1:37 am

Regarding the paper I mentioned, it is Phys. Rev. B 82, 035114 (2010) and the authors used VASP. The paragraph of interest is on p4 and it the last one before section B.