My Community

Dear Admin
I am trying to do a spin polarized DFT+U calculations for determining the bulk band structure of a FCC system.

What i did is:

1- Did a SC calculation with this INCAR and 6x6x6 kpoints

SYSTEM = Bulk band
ISTART = 0
ICHARGE= 2
ISMEAR = 0
SIGMA = 0.05
NBANDS = 70
EDIFF = 1.0E-7
LWAVE = .FALSE.
LCHARG = .TRUE.
ENCUT = 500
IBRION = -1
NSW = 0
ALGO = fast
NPAR = 1
ISPIN = 2
MAGMOM = 2*2.0 2*-2.0 8*0.0
LDAU = .TRUE.
LDAUTYPE = 2
LDAUL = 3 -1
LDAUU = 4.5 0
LDAUJ = 0.51 0
LMAXMIX = 6
LASPH = .TRUE.

2- Then did a NSC calculation changing the ICHARG=11 in the above INCAR and save the CHGCAR file from the SC step. Also changed the kpoints to the FCC high symmetry points in line mode.

The problem is this;
When i use a conventional unitcell with 12 atoms the result is metallic. Actually there is a band gap just above the E-f but E-f is still touching one of the bands.

When i changed the unitcell to the tetragonal one with 6 atoms (of course with the appropriate changes of atom numbers in magmom tag and kpoints of tetragonal cell) the result is insulator. But both of them are still the same structure, just unitcell definitions change. (If i turn the conventional unitcell 45degrees i have the tetragonal one with different number of atoms)

Could you please tell me if i am doing a mistake in the calculation procedure? I am not sure that if i should use the magmom tag in SC calculation.

Many thanks

What values of MAGMOM did you use for the 12 atom unit cell? It appears your tetragonal MAGMOM values are going to give you an anti-ferro-magnetically ordered system with the ordering along the (100) axis. Do your 12-atom unit cell MAGMOM values give you the same magnetic ordering along the same axis? If not, that could be the cause of your metal-to-insulator transition.

My system is AB2 . In 12 atom unitcel, there is 4 A atoms and the magmom for first two atom in the fist layer is 2 and -2 for the second layer A atoms. In the tetragonal cell there is two A atoms so one of them is up spin and the other is down. O atoms in both calculations are unspinned.

The magnetic ordering in the z direction is same for both geometries because you get the tetragonal cell when you rotate the 12atom cell around the z axis. So just x and y vectors change.