DOSCAR empty?

Message

rpnabar · #1 Post by **rpnabar** » Fri Jul 14, 2006 6:06 pm

Is there a tag that I need to turn on to obtain the data in DOSCAR? Right now my jobs run fine but my DOSCAR is always empty except for the first few lines.

My INCAR looks like this:

ISTART = 0
ICHARG = 2

LWAVE = .FALSE.
LCHARG = .FALSE.

Electronic Relaxation 1
ENMAX = 396 eV
NELM = 50 max number of electronic steps
EDIFF = 2E-05
EDIFFG = -.03

Ionic Relaxation
NSW = 50 max number of geometry steps
IBRION = 2 ionic relax: 0-MD 1-quasi-New 2-CG
ISIF = 0 (0:force=y stress=n ions=y shape=n volume=n,
POTIM = .5

DOS related values:
ISMEAR = 1
SIGMA = .2 broadening in eV

Electronic Relaxation
IALGO = 48
LREAL = .FALSE. real-space projection

Veronika · #2 Post by **Veronika** » Mon Jul 17, 2006 10:08 am

Set LORBIT!
(see also http://cms.mpi.univie.ac.at/vasp/vasp/node113.html)

#3 Post by **admin** » Mon Jul 17, 2006 12:14 pm

LORBIT only determines whether the LOCAL PARTIAL DOS (LDOS) is written onto DOSCAR. the total DOS should be written after each NBLOCK*KBLOCK steps, or, in any case, quite at the end of a VASP run. please check whether your job really finished without errors or whether it stopped before (eg for some timeout).

rpnabar · #4 Post by **rpnabar** » Mon Jul 17, 2006 12:58 pm

Yes, the jobs seem to have terminated naturally; I do have the line that says:
reached required accuracy - stopping structural energy minimisation
LOOP+: VPU time10820.45: CPU time10823.47

followed by the "General timing and accounting informations for this job"

I was expecting, just like head-admin said, that I have the total DOS information in my DOSCAR. Its reproducible and each time I get just a 5 line DOSCAR that looks somethig like this:
******************
56 56 0 0
0.1428571E+03 0.2000000E-08 0.2000000E-08 0.2000000E-08 0.5000000E-15
1.000000000000000E-004
CAR
for sample slab (R.S)
*****************************
The last line being a comment from my INCAR.

I'm totally puzzled; willing to try any sugestions to debug this problem!

des · #5 Post by **des** » Mon Jul 17, 2006 4:17 pm

It seems the CHARG is not written in ur output files. Try using
LWAVE = .TRUE.
LCHARG = .TRUE.
and
LORBIT = 10 or 11

I think this will give you the DOS

rpnabar · #6 Post by **rpnabar** » Mon Jul 17, 2006 5:47 pm

I cannot be using LORBIT = 10 or 11 since I am using ultrasoft psps; LORBIT=10/11 works only with PAW right?

But thanks for the other sugesstion des. I set LCHARG and LWAVE to true and then I get in data the CHGCAR and CHG file.

I followed this by doing a static calculation NSW=0; IBRION=-1; Now I do get data in the DOSCAR.

Is VASP set up so that DOSCAR is written only after static runs?

#7 Post by **admin** » Tue Jul 18, 2006 11:07 am

1) DOSCAR
Please check the value of KBLOCK in your actual run (unless it is explicitely set in INCAR, it is set to NSW in the code) ---> if your system needs less than NSW steps to reach ionic convergence, DOSCAR is not written (please remind that the DOS should be calculated in a static run with an increased k-mesh density after ionic convergence is reached, hence it is not reasonable to set NSW>1 in such a case)
If you want to have DOSCAR written during the ioic convergence procedure though, please give KBLOCK explicitely in INCAR.
2) LORBIT with US-PPs
if you use US-PPs, you can set LORBIT=1, and give the RWIGS
for each atom type in INCAR explicitely.

rpnabar · #8 Post by **rpnabar** » Tue Jul 18, 2006 9:00 pm

Thanks head admin. I noticed that the POTCAR contains values for RWIGS; if I set LORBIT=1 will it default to that? Also why are there two values in the POTCAR? Pt_GGA for example has:

RWIGS = 2.750; RWIGS = 1.455 wigner-seitz radius

They seem pretty different too? They should be close to approximately its covalent radius?

#9 Post by **admin** » Wed Jul 19, 2006 9:26 am

RWIGS is given in atomic units (2.75) and in Ã… (1.455) in POTCAR