Fractional occupancies in the Ni free atom

[Free Electron]

Dear all,

Following the guidelines given in the Vasp guide http://cms.mpi.univie.ac.at/vasp/vasp/D ... atoms.html, I have tried to calculate the total energy of the Ni free atom. My input files are as below:

POSCAR -----------
1
13.0 00.0 00.0
00.0 14.0 00.0
00.0 00.0 15.0
1
direct
0.0 0.0 0.0

KPOINTS ----------
1
rec
0 0 0 1

INCAR ------------
ISYM = 0
ISPIN = 2
ISMEAR = 0
SIGMA = 0.002
AMIX = 0.2
BMIX = 0.0001
ICHARG = 1
ENCUT=600
EDIFF=0.00000001
ALGO = A ; LSUBROT = .FALSE. ; NELM = 500 ; TIME = 0.2

And I used the pbe_paw POTCAR.

I executed VASP three times and the total energy was converged, but I ended up with fractional occupancies:

spin component 1

k-point 1 : 0.0000 0.0000 0.0000
band No. band energies occupation
1 -6.3009 1.00000
2 -5.6743 1.00000
3 -5.5825 1.00000
4 -5.3029 1.00000
5 -5.3028 1.00000
6 -4.9296 1.00000
7 -1.0534 0.00000
.
.
.
spin component 2

k-point 1 : 0.0000 0.0000 0.0000
band No. band energies occupation
1 -4.7026 1.00000
2 -4.4841 1.00000
3 -4.4840 1.00000
4 -4.1190 0.70933
5 -4.1174 0.29067
6 -3.9445 0.00000
8 -0.2898 0.00000

Any suggestions to solve this problem?

Moreover, in the POTCAR file, the valence electron configuration (VRHFIN) of Ni is not given:

PAW_PBE Ni 06Sep2000
10.0000000000000000
parameters from PSCTR are:
VRHFIN =Ni:
LEXCH = PE
EATOM = 1077.9025 eV, 79.2236 Ry

Why is it like that for Ni? And what is the configuration?

Thank you in advance ..

[alex]

Hi,

have you considered to freeze the spin state with NUPDOWN ?

Cheers,

alex

[Free Electron]

Dear Alex,
Thank you very much!

I used (NUPDOWN = 2) in the INCAR but I got exactly the same result!

I think NUPDOWN controls the "difference" between number of electrons in up and down spin component. This was controlled properly in my case, despite the existence of fractional occupancies!

When I used (NUPDOWN = 1), I got fractional occupancies in both up and down spin components, while the "difference" was exactly 1 !

Please, let me know if I misunderstand something ..

Cheers

[kelum]

Hi,
I think setting SIGMA=0 could help, or setting the occupancies manually with
FERWE= 6*1 N*0 (N is the NBANDS-NOCCUPIED and NOCCUPIED in your case is 6)
FERDO=4*1 (N+2)*0
either of this might work

Best
jik