I am trying to reproduce the data of stanene using VASP. I am getting the exact result of band gap without SOC, and phonon dispersion spectrum, however my calculated band gap with SOC comes out to be 0.14 eV (indirect), while reported one is 0.1 eV (direct). What can be the possible reason for this?
It appears from exact value of band gap without SOC and phonon dispersion relation that my POSCAR is correct, and i have used 15*15*1 as k-mesh and 500 eV ENCUT for band structure calculations. I did the SOC calculations in three steps, using the following INCAR.
Firstly, i took the relaxed structure and performed the collinear normal VASP calculations using the INCAR file as
ISTART = 0
ICHARGE = 2
ISMEAR = 0
SIGMA = 0.001
NBANDS = 20
POTIM = 0.02
ISPIN = 2
ENCUT = 500
IBRION = 2
NSW = 0
then i took the CONTCAR and CHGCAR of previous step as POSCAR and performed non-collinear self consistent calculations using the following INCAR
ISTART = 0
ICHARGE = 11
ISMEAR = 0
SIGMA = 0.001
NBANDS = 40
POTIM = 0.02
ISPIN = 2
ENCUT = 500
IBRION = 2
NSW = 0
LWAVE = .FALSE.
ISYM = 0
LSORBIT = .TRUE.
SAXIS = 0 0 1
MAGMOM = 2*0 2*0 2*1.1
GGA_COMPAT = .FALSE.
then i took CONTCAR and CHGCAR of previous step and performed non-collinear non self consistent calculations with the KPOINT path using the following INCAR
ISTART = 0
ICHARGE = 11
ISMEAR = 0
SIGMA = 0.001
NBANDS = 40
POTIM = 0.02
ISPIN = 2
ENCUT = 500
IBRION = 2
NSW = 0
LWAVE = .FALSE.
ISYM = 0
LSORBIT = .TRUE.
SAXIS = 0 0 1
MAGMOM = 2*0 2*0 2*1.1
GGA_COMPAT = .FALSE.
Can somebody please tell me that what mistake i have done and what could be the possible reason for getting an indirect band gap?
I would be extremely thankful for the same.
and i have used 15*15*1 as k-mesh
I would try to use k-mesh with some more points along z-axis, to begin. As the accurate reproduction of band structure is required, just one point along z-axis quite probably can be insufficient.
Thank you Sir for your reply. But it is a 2-D structure, can i use more k-points in z-direction?
But it is a 2-D structure, can i use more k-points in z-direction?
Depends on the approach which is used to model 2D system. As it is modeled to be 3D system with quite large distance between layers, for instance, using 1 k-point can result sometimes in some curious situations.
Anyway, the discrepancy 0.1-0.14 eV in band gap obtained within approach implemented in VASP (PAW method, if I am not mistaken) hardly could be considered to be reliable enough, it is within the method inaccuracy.
If you could share you POSCAR file, it may be easier to check if there is a mistake in it first.
For III-V bulk semiconductors, the accurate lattice parameters were important to have a proper order of band transitions. I guess the same should apply to 2D layers. Try PBE+D3.
- Badges
- Science topic