[Wien] Choice of magnetization axis in the case with spin-orbit coupling

[Zhu, Jianxin]

Dear Peter and Wien2k users,

For the spin-orbit case, when we run initso_lapw, we are prompted to
choose the moment direction

---->Please select the direction of the moment ( h k l )

I notice that the choices of e.g.,  (0  0  1) and (0  0  -1) give slight
different values.
In particular, when I choose (0  0  1), I obtain from LDA+U+SO

:MMINT: MAGNETIC MOMENT IN INTERSTITIAL =   -0.32640
:MMI001: MAGNETIC MOMENT IN SPHERE   1    =   -4.71648

:ORB001:  ORBITAL MOMENT:  0.00000  0.00000  4.59939 PROJECTION ON M
4.59939

while with (0 0 -1), I obtain from LDA+U+SO with the rest of conditions
the same

:MMINT: MAGNETIC MOMENT IN INTERSTITIAL =   -0.34861
:MMI001: MAGNETIC MOMENT IN SPHERE   1    =   -4.72342


:ORB001:  ORBITAL MOMENT:  0.00000  0.00000  4.52838 PROJECTION ON M
-4.52838


Should I always use positive values for (h k l)?
Or should I choose (0 0 1) when I prepare the spin moment to be up, that
is, during the stage of "instgen_lapw -ask"
select spinpolarization up, dn or non-magnetic ( u, d, n )
u

while (0 0 -1) when I prepare the spin moment to be down, that is, during
the stage of "instgen_lapw -ask"
select spinpolarization up, dn or non-magnetic ( u, d, n )
d


I appreciate if you can share your thought.



Thanks, 

Jianxin