Vibration frequancy

Message

simoneca · #1 Post by **simoneca** » Tue Nov 06, 2007 10:10 am

Hi,
I am trying to calculate vibrational frequencies for a single atom adsorbed on a surface. I'm new in VASP community hence I made a mistake somewhere..... javascript:addtext('

')
I used the following INCAR file:

INCAR for frequencies calculation.

! Electronic relaxation
ALGO = Fast ! Algorithm for electronic relaxation
NELMIN = 8 ! Minimum # of electronic steps
EDIFF = 1E-6 ! Accuracy for electronic groundstate
ENCUT = 500 ! Cut-off energy for plane wave expansion
PREC = Accurate ! Normal/Accurate
LREAL = Auto ! Projection in reciprocal space?
ISMEAR = 0 ! Smearing of partial occupancies. k-points >2: -5; else 0
SIGMA = 0.2 ! Smearing width
ISPIN = 2 ! Spin polarization?

! Ionic relaxation
NSW = 1 ! One-step just to build Hessian matrix
IBRION = 5 ! Finite difference scheme for derivatives
POTIM = 0.02 ! Step for ionic movements

All ions are free to move along x,y,z.
What I got is an OUTCAR file with the electronic optimization and then

Finite differences:
Step POTIM = 2.000000000000000E-002
Degrees of freedom DOF = 27
LATTYP: Found a hexagonal cell.
ALAT = 4.9120000000
C/A-ratio = 4.0716612378

Lattice vectors: .......

Than the generation of the automatic k point mesh and nothing more. VASP
stops here, apparently without moving any atom for the hessian calculation using the finite differece. What I did wrong?
Thanks in advice for any precious suggestion.

Franky · #2 Post by **Franky** » Tue Nov 06, 2007 11:55 am

Please, read the manual carefully.
Some of your parameters dont make sense.

Use for example:
NSW = 5000
IBRION = 5
NFREE = 2
ISIF = 0
POTIM = 0.02
NELM = 80
NELMIN = 3
EDIFF = 1E-8
LREAL = .FALSE.
PREC = High
...etc...

Do you have a pre-converged configuration?
Do you really need spinpolarization?
Check ISMEAR carefully! Better ISMEAR=1 plus approp. SIGMA.

simoneca · #3 Post by **simoneca** » Tue Nov 06, 2007 1:10 pm

Thanks Franky.
I have read the manual again and I think the problem is in NSW.
Tell me if I am right...
When NSW is 1 then the default value for IBRION sholud be -1, so no ionc movement. I will try your settings.

The parameters I mentioned before are taken from the post http://cms.mpi.univie.ac.at/vasp-forum/ ... .php?4.138
where the user "Baki" used for the frequencies
NSW=1
IBRION=5
POTIM=0.02

Yes, I have a preconverged configuration. I have used both ISMEAR 0, and -5 but I'll try with 1. Thanks again.

#4 Post by **admin** » Tue Nov 06, 2007 1:52 pm

what is written in the job.stdout file, is there any error message which gives a hint why vasp stopped? Your input should work properly, this cannot be the reason.

Franky · #5 Post by **Franky** » Thu Nov 08, 2007 9:16 am

Just in case:
I absolutely didnt meen to sound rude, which I did.
Sorry, for that! I was in a bit of a hurry, which isnt a good excuse.

Looks like I have to read the manual again. NSW isnt the problem.
Thanks for making me aware of the thread concerning the ISMEAR issue. I wasnt aware of that and have to check my results.