My Community

Dear All,

I am trying to run vdW-DFT using vasp5.2.12. I used PAW-PBE potentials with spin or non-spin polarization. I was surprised that I couldn't find any keywords in the OUTCAR file related to the vdW interaction part. It looked same with the conventional VASP calculation using vasp.4.6 without vdW.The INCAR file is attached below:

SYSTEM = surface

LPLANE = .TRUE.
NPAR = 4

Startparameter for this run:
PREC = Medium
EDIFF = 1e-4 high convergence not necessary?
EDIFFG = -1e-2
ISPIN = 2 spin polarized calculation?
IALGO = 48 pre-conditioned conjugate gradient
# ALGO = Fast
# VOSKOWN = 1
# NELMDL = 6
# LORBIT = 11
NELMIN = 4 the minimum number of electronic interations
# ISYM = 1 symmetry off
# AMIX = 0.2
# BMIX = 0.0001
ISTART = 0
# ICHARG = 1
ENCUT = 600
LDIPOL = T
IDIPOL = 3

LUSE_VDW=.TRUE.
AGGAC = 0.00000
GGA= MK
PARAM1= 0.1234
PARAM2= 1.0000

LREAL = On real-space projection

OUTPUT CONTROL
LCHARG = .FALSE.
LWAVE = .FALSE. the i/o cost is not worth it.
LVTOT = .FALSE.

IONIC RELAXATION
NBLOCK = 1
NSW = 200 number of steps for IOM
IBRION = 2 standard ab initio MD (set to 1-2 for geo opt)
POTIM = 0.2

DOS RELATED (disabled)
ISMEAR = 0 occupation forced by INCAR(Ask section 7.30)
SIGMA = 0.05

Any comments? Thanks!

Also, I was confused that when I tried other exchange functionals for vdW-DFT, the total energy of the whole systems differs much with each other and also with that using conventional DFT (no vdW interaction included). I know there must be some difference, but shouldn't be so large (several tens of eV). Any one can help me about this? Thanks!

<span class='smallblacktext'>[ Edited Sun Dec 04 2011, 10:21AM ]</span>

Hi,

in the current version no parameters are written to OUTCAR,
they should be present in vasprun.xml.
The energy will change with different exchange functionals and
for large system it can be several tens of eVs easily.
Also these are not the absolute energies but relative to POTCAR reference
so the changes just seem large.

jik