questions about ISYM-tag in spin-orbital coupling case

Message

dxs571 · #1 Post by **dxs571** » Sun Dec 04, 2005 2:54 am

When we do non-collinear magnetization calculation, we turn off the symmetry(ISYM=0),but what about in orbit moment
case.Since I don't know very well this kind of problem,any suggestion will be highly appreciated.Thanks :p

#2 Post by **admin** » Mon Dec 05, 2005 9:35 am

If you want to calculate SOC, the non-collinear orientation of the magnetic moments is enabled automatically, hence, also all
synmetry has to be switched off as well. Please set ISYM=-1

dxs571 · #3 Post by **dxs571** » Sun Dec 11, 2005 1:22 pm

According to your suggestion I have done the orbital
moment calculation, but the moment seems so small compared to the correct value.Here is my main procedure:
First do a non-spin self-consistent calculation,I do it with LDA+U(3.0) approach.
Second turn the sym tag off do a spin-orbit coupling
cal and get the orbital moment at the end of OUTCAR but with a high U value (7.0ev).
Here is my INCAR file in the second step
ISPIN = 2
ISYM = 0
ICHARG = 1
SAXIS = 0 0 1
ISMEAR = 0; SIGMA = 0.2
prec=a
ALGO = Fast
LSORBIT = .TRUE.
LORBMOM = .TRUE.
MAGMOM= 0.0 0.0 4.0 141*0.0
NBANDS = 312
AMIX = 0.2
BMIX = 0.00001
AMIX_MAG = 0.8
BMIX_MAG = 0.00001
LORBIT = 11
EDIFF = 1.0E-04

LDAU = .TRUE.
LDAUTYPE = 2
LDAUL = 2 -1 -1
LDAUU = 7.00 0.00 0.00
LDAUJ = 0.90 0.00 0.00
LDAUPRINT = 2
LMAXMIX = 4

Though it's a supercell one (48 atoms),only the first atom is expected to have spin and orbital moment and the orbital moment of the first atom in the z direction amounts to only 0.01.
What's the problem ? That really drives me crazy recently,any suggestion will be appericiated.