Empty DOSCAR & EIGENVAL files for GW0 Calculations?

Message

Mahlaga · #1 Post by **Mahlaga** » Mon Feb 22, 2010 1:08 pm

Dear VASP users,

I am running a GW0 calculation for ZnO system, but to my suprise, The DOSCAR, EIGENVAL, and CHGCAR Files are EMPTY at the end of the calculation.
The first step (generating WAVEDER) runs well. After the first I copy WAVEDER and the basic 4 files into a new directory for the GW0 calculation. My INCAR file for the second step is as follows:
SYSTEM = ZnO
NBANDS = 96
ISMEAR = 0
SIGMA = 0.05
LOPTICS = .TRUE.
ENCUT = 500
ALGO = GW0
NOMEGA = 60
NELM = 4
LSPECTRAL = .TRUE.

My POSCAR file is as follows
SYSTEM = ZnO
NBANDS = 96
ISMEAR = 0
SIGMA = 0.05
LOPTICS = .TRUE.
ENCUT = 500
ALGO = GW0
NOMEGA = 60
NELM = 4
LSPECTRAL = .TRUE.

and the KPOINTS file is
Automatic mesh
0
Gamma
6 6 6

Does anyone know what might be causing this problem?
I also need to plot the GW0 bandstructure. Is the bandstructure calculated the same way as in DFT? That is, after GW0 step 2, do I have to run a non-selfconsistent calculation with ICHARG = 11?

Thank you in advance.

Mahlaga

Mahlaga · #2 Post by **Mahlaga** » Mon Feb 22, 2010 1:14 pm

My mistake: I pasted INCAR instead of POSCAR in my thread. My POSCAR file is as follows
ZnO :rs(B1)
4.34000000000000
0.0000000000000000 0.5000000000000000 0.5000000000000000
0.5000000000000000 0.0000000000000000 0.5000000000000000
0.5000000000000000 0.5000000000000000 0.0000000000000000
Zn O
1 1
Direct
0.0000000000000000 0.0000000000000000 0.0000000000000000
0.5000000000000000 0.5000000000000000 0.5000000000000000

0.00000000E+00 0.00000000E+00 0.00000000E+00
0.00000000E+00 0.00000000E+00 0.00000000E+00

giacomo giorgi · #3 Post by **giacomo giorgi** » Tue Feb 23, 2010 5:03 am

I do not think you need to do a non-selfconsistent. What you need to set is

ENCUTGW

since ENCUTGW represents the energy cutoff for the response function and its default is ENCUT, I think your calculation risks to crash by setting ENCUTGW=500....
Try starting with ENCUTGW=90....

I hope it helps,
G

Mahlaga · #4 Post by **Mahlaga** » Tue Feb 23, 2010 10:15 am

Thank you Giorgi. I will try this and see how it goes.

About BandStructure:
I need the bandstructure along specific high symmetry lines. So if a non-selfconsistent calculation is not necessary, then this means I have to enter the high symmetry points explicitly in the KPOINTS file, or use the line mode as described in the manual. Am I right?