Page 1 of 1

NEB

Posted: Wed Oct 19, 2011 1:16 am
by apple
hi,
I am doing a NEB calculation for an adatom diffusing between two stable sites A and B on a metal surface, with A higher in energy than B. Starting with A, the energies for the 5 images leading to B are all decreasing. Does it mean that NEB did not find a transition state? Should there be one if the two sites are energetically stable? Could I say that going from A to B there is not barrier (which makes me wonder why A was found to be a stable site), and going from B to A the barrier is at least the difference in the energies between A and B? How should I interpret this result? Thanks.

NEB

Posted: Wed Oct 26, 2011 12:46 am
by imcnab
This does mean that NEB didn't find a transition state.
You don't give many details, but it might be worth looking at the Austin groups website, and using Climbing Image NEB (if you are not already). In principle, if there is a transition state, this will find it.

Check out:
http://theory.cm.utexas.edu/vtsttools/

If you decide to use these wonderful add-ons, you will need to re-build vasp.

The quick answer is YES, there is no transition state found. Why did VASP find a stable site A? Without you giving details of your INCAR, for both calculations, it isn't possible to say.

NEB

Posted: Wed Oct 26, 2011 8:34 pm
by apple
hi,
thanks for replying to my post.
Below is my INCAR file for the optimizations:
===============================
SYSTEM = metal1
NWRITE = 1
ISTART = 0


ENCUT = 400.0 eV
PREC = HIGH
EDIFF = 1E-05
EDIFFG = -1E-03

ISYM = 0
LCORR = T
IBRION = 1
NSW = 200
ISIF = 2
ISMEAR = 0; SIGMA = .1
IALGO = 38
LDIAG = T
LREAL = F

NBANDS = 160
===================================


And the NEB INCAR file:
====================================
SYSTEM = metal1
NWRITE = 1
ISTART = 0

ENCUT = 400.0 eV
PREC = HIGH
EDIFF = 1E-05
EDIFFG = -1E-03


ISYM = 1
LCORR = T
IBRION = 1
NSW = 200
ISIF = 2

ISMEAR = 0; SIGMA = .01
IALGO = 38
LDIAG = T
LREAL = F

NBANDS = 160

NPAR = 2
IMAGES = 5
SPRING = -5
LCLIMB = T
=================================

NEB

Posted: Sat Oct 29, 2011 5:36 am
by Neutrino
While I do not know why you obtained this strange result, I find using Gaussian smearing with a metal not very suitable. Also, I wouldn'r recommend ISYM=1 with NEB calculations.

My suggestion is to do a quick scanning for the potential energy surface on the metal surface. Construct a grid of calculations such that your adatom scan the metal surface. This should give you hints about stable sites and possible diffusion pathways. Keep in mind that the pathway can be curved or my be by exchange mechanism.