strange band occupation

Message

giacomo giorgi · #1 Post by **giacomo giorgi** » Thu Mar 19, 2009 11:55 am

Dear all,
I am experiencing something that sounds strange (wrong, I guess).

I have a 8 atom cell of GaAs. I replace 1 Ga and 1 As with two Ge atoms and I do spinpolarized calc (ISPIN=2) keeping in mind that in my system there are 82 electrons.

The following is the INCAR

SYSTEM=GaAs
ISTART=1
ISMEAR=0
PREC=Med
ADDGRID=.True.
ISIF=3
NSW=300
IBRION=1
EDIFF=1E-5
EDIFFG=-0.01
GGA=91
ISPIN=2
ENCUT=287.594
ENAUG=531.356

I do band calculation with this INCAR
ISPIN=2
NUPDOWN=0
ISTART=1
LREAL=TRUE
PREC=Med
ICHARG=11
EDIFFG=-0.05
EDIFF=1E-5
GGA=91
ENCUT=287.594
ENAUG=531.356
NWRITE = 2
INIWAV = 1 electr: 0-lowe 1-rand 2-diag
NELMDL = -7
NELM = 90
NELMIN = 10
WEIMIN = 0
LMAXMIX = 6
ALGO =F
IALGO = 48
NSIM = 1 NSIM bands are optimized simultaneously (IALGO=48 only)
LCHARG = .TRUE.
LWAVE = .TRUE.
LELF = .TRUE.
IBRION = -1
ISIF = 3
ISYM = 1
POTIM = 0.1
LORBIT = 12
ISMEAR = 1 (-1-Fermi, 1-Methfessel/Paxton)
SIGMA = 0.02 broadening in eV
NPAR = 1
LPLANE= .TRUE.

and when I check my OUTCAR I see that:

E-fermi : 2.7302 XC(G=0): -9.4599 alpha+bet : -8.3138
add alpha+bet to get absolut eigen values

spin component 1

k-point 1 : 0.0500 0.0500 0.0500
band No. band energies occupation
1 -22.2938 1.00000
......
......
.....
38 0.3681 1.00000
39 2.3121 1.00000
40 2.6090 1.00000
41 2.7235 0.54911
42 2.7235 0.54896
43 3.6999 0.00000
44 3.6999 0.00000
....
....

So it seems that one electron, even if ISPIN=2, is shared by two bands. (Same behavior of course for beta)
Where am I wrong?
thanks,
G

#2 Post by **admin** » Fri Mar 20, 2009 3:34 pm

the two states are degenerate in energy. If this degeneracy is not lifted, the electron has to be 'shared'

giacomo giorgi · #3 Post by **giacomo giorgi** » Sun Mar 22, 2009 4:51 am

Dear admin,
thanks a lot. But let me harass you once more. Technically, you mean lifting degeneracy by imposing fermi temperature = 0 (ISMEAR=-1
SIGMA=0.0)?
Because I also did it but results were not encouraging (tremendously slow convergence).

Still many thanks,
G

#4 Post by **admin** » Mon Mar 23, 2009 10:21 am

The degeneracy of the states most probably is due to the symmetry of your cell. If the space group of the cell allows for degenerate states, you will not get rid of them by decreasing SIGMA because some of the eigenstates of your system simply are degenerate (let me remind you of the d eg-t2gsplitting or the 3-fold degeneracy of p-states of cubic systems, for example).
Please check for the symmetry information as written in OUTCAR.