My Community

I need a good zoom in of the valence band splitting of Silicon around the Gamma point when spin-orbit coupling is taken into consideration. Here is what I am doing:

The INCAR for the self consistent run is

PREC=high
ISIF=3
ISMEAR=-5
IBRION=-1
NSW=5
NBANDS=5

The KPOINTS is

8x8x8
0
mp
12 12 12
0 0 0

And then I run a non self consistent run with INCAR

PREC=high
ISMEAR=1
ISPIN=2
SAXIS=0 0 1
MAGMOM=0 0 1 0 0 1
ICHARG=11

And KPOINTS

k-points along high symmetry lines
200 ! 10 intersections
Line-mode
cart
.1 .1 .1 ! W
0 0 0 ! gamma
0 0 0 ! gamma
.1 0 0 ! X

I need to see a .044eV split off between one of the original triply degenerate bands and the other two stay degenerate at the top of the valence band but I get 3 non degenerate band values around Gamma neither of which are split by~.044eV.

What do you suggest?

I've never done any spin-orbit calculations myself but anyhow here are my thoughts.

If you want to account for the spin-orbit coupling, you'll have to set LSORBIT = .TRUE. , which you havn't done as far as I can see (if the INCAR you have posted is the entire file). Please have a look in the manual

http://cms.mpi.univie.ac.at/vasp/vasp/node159.html

Note that ISPIN = 2 doesn't switch on the spin-orbit coupling. (It determines whether or not you allow for a spin-polarized density in the collinear case.)

EDIT: Also, since LSORBIT = .TRUE. automatically will set LNONCOLLINEAR = .True. you'll have to use a VASP binary compiled without the flags -DNGXhalf and -DNGZhalf.

Cheers,
/Dan



<span class='smallblacktext'>[ Edited Thu May 28 2009, 10:50AM ]</span>

Thanks Dan

I do have a VASP with those flags removed, I did what you said and now I get

Error reading item 'ICHARG' from file INCAR.
Error code was IERR=5 ... . Found N= 0 data.

What do you think?