Page 1 of 1

Spin Orbit Coupling in VASP

Posted: Mon Aug 10, 2009 1:03 pm
by slgupta
Respected All,
I want to study spin orbit coupling using VASP. I have read the corresponding chapters in the manual and I have also gone through the discussion available on the platform of this forum. I am using PAW_GGA potentials. There are certain issues those led me to confusion.

1. In the manual to study the spin orbit coupling, we have to set

LSORBIT =.True.
SAXIS = s_x s_y s_z
MAGMOM = mi_x mi_y mi_z where 'i' corresponds to atomic spices in the system

This led me to a DOS having only one componet irrespective of ISPIN = 1 or 2. Why is it so? If we want to get information about spin up and spin down DOS, Where can we get?

2. During the discussionsn over Spin orbit coupling in this forum, some of the respected members has suggested to use LORBIT tag. I have done the calculations using LORBIT = 12. This time I have got spin up and spin down DOS for ISPIN = 2. But this was different from what I have got from step 1. Now, which one is correct?

I request to respected members to answer my quries and help me out.

Thanks in advance.
-S.L. Gupta
IIT KANPUR, INDIA

Spin Orbit Coupling in VASP

Posted: Mon Oct 12, 2009 11:49 am
by admin
if you set SOC on, non-collinear magnetism is assumed automatically. It makes no sense to speak about spin- up and down only then, the spinor has 4 components, the first block of DOSCAR only gives the total DOS.
HOWEVER, you will find that local densities of states (LORBIT=11 or 12) are given in 4 components for each (l,m):
total(l,m) DOS, and the magnetization along x,y and z (for each l,m) respectively.