While doing a VASP calculation, I'm getting a total magnetic moment of +1 mu_B in relaxation (due to one magnetic impurity atom). When I do SCF with double the k-mesh compared to relaxation (without including spin-orbit coupling), I don't get any magnetic moment in system. However, If I switch on spin-orbit coupling, I still get small magnetic moment of 0.33 /mu_B. Can anybody please comment why it is so strange with magnetic moment?
I would suggest you check the convergence with k-points during scf calculations. What is the k-mesh you are using for relaxation and scf calculations?
I think you'd check the KPOINTS as suggested by Manchanda's.
Though I don't know what the impurity atom is, the results you presented seems not reliable, thus I won't comment on the magnetic moments. However, for a structure with good convergence, the magnetic moment obtained from the last relaxation calculation should be similar to that in SCF calculation.
you can also check your pseudopotentials ( more valence electron is better), second when you perform spin-orbit coupling calculation in VASP you need to turn non co-linear magnetism switch, which demand some caution.
Also it is may be useful to check your result with another code,just to make sure you are incorporating your input parameter properly,
Thanks you all. I really appreciate. It was indeed due to the k-mesh problem. In fact I did relaxation again with a dense k-mesh and magnetism vanished there. Therefore, in the subsequent calculations (scf), I didn't face any problem, nor in spin-orbit calculation.
What about the magnetic moment obtained in band structure calculation? I am finding it to be almost double the one obtained in SCF calculations.
Should we keep the MAGMOM flag in band structure calculations as the Magnetic moment obtained in the self-consistent step? Kindly suggest me the same.
Thanks in advance!
Dear Shahid Sattar ,
Thank you for your reply.
I am attaching the INCAR of my scf and non-scf steps, with the magnetic moment output, please have a look and let me know where i am doing wrong.
INCAR-scf
# Basic setup:
SYSTEM = Relax
ISTART = 0
ICHARGE = 2
ISMEAR = 0
SIGMA= 0.001
EDIFF = 0.0000001
#EDIFFG = -0.005
PREC = Accurate
LWAVE = .FALSE.
LCHARGE = .TRUE.
LVTOT = .FALSE.
ENCUT = 550
LREAL = Auto
AMIN = 0.01
ISIF = 2
NPAR = 10
IBRION = -1
NSW = 0
ISPIN = 2
LORBIT = 11
MAGMOM = 8*0.0 8*0.0 8*0.0 2.0 9*0.0 2.0 56*0.0 2*2.0
INCAR-nscf
# Basic setup:
SYSTEM = Relax
ISTART = 1
ICHARGE = 11
ISMEAR = 0
SIGMA= 0.001
EDIFF = 0.0000001
#EDIFFG = -0.005
PREC = Accurate
LWAVE = .FALSE.
LCHARGE = .FALSE.
LVTOT = .FALSE.
ENCUT = 550
LORBIT = 11
LREAL = Auto
AMIN = 0.01
ISIF = 2
NPAR = 10
IALGO = 48
IBRION = -1
NSW = 0
ISPIN = 2
LORBIT = 11
MAGMOM = 8*0.0 8*0.0 8*0.0 0.27 9*0.0 0.22 56*0.0 2*4.0
I am just copying CHGCAR from the scf step to the non-scf, and I replaced the magmom flag as the values i obtained in the OUTCAR of scf. If i keep it as the same tag as i used in scf step, then again magnetic moment value is drastically different.
Kindly help me with the same.
Thanks and regards,
Payal
- Badges
- Science topic
- Similar topics
- Physics
- Condensed Matter Physics