Convergence problem of Cl-NEB calculation

Vasp transition state theory tools

Moderator: moderators

Post Reply
lanzi
Posts: 2
Joined: Mon Jun 21, 2010 9:10 am

Convergence problem of Cl-NEB calculation

Post by lanzi »

Dear all,
I've been suffering from the convergence problem with Cl-NEB calculation just for a simple system of CO dissociation on Pt(111) surface. I have read some discussions in this issue, but still I can not figure out what is going on with my calculation. It runs for hundreds of ionic steps without force convergence. Here I paste my INCAR for the second step (there's no other difference for the first step except POTIM=0.01):
SYSTEM=Pt-CO
ISTART=0;ICHARG=2
ENCUT= 400
IBRION=3
NSW=500
POTIM=0.05
GGA=PE
IALGO = 48

NSIM =4
LPLANE = .TRUE.
NPAR = 4
ICHAIN=0
IMAGES=8
SPRING=-5.0
LCLIMB=.TRUE.
LTANGENTOLD=.FLASE.
LDNEB=.FALSE.

NELM=200
EDIFFG=-5E-02
LCHARG=.FALSE.
LWAVE=.FALSE.
LREAL=.TRUE.
LORBIT=11
ISMEAR=1;SIGMA= 0.28
By the way, I would also like to ask why the script command of nebef.pl doesn't work for my job. The column which is supposed to show the force shows all zero. So, I check the force from "TOTAL-FORCE (eV/Angst)" in OUTCAR. Is there any problem with this?
Thanks in advance for all help and suggestion.

Jinlan
graeme
Site Admin
Posts: 2291
Joined: Tue Apr 26, 2005 4:25 am
Contact:

Re: Convergence problem of Cl-NEB calculation

Post by graeme »

Do you have our code linked into your vasp executable? If you can't see the forces with the nebef.pl script, you are probably not using our NEB routines.
lanzi
Posts: 2
Joined: Mon Jun 21, 2010 9:10 am

Re: Convergence problem of Cl-NEB calculation

Post by lanzi »

Dear graeme,
Thank you so much for your response. Yes, I also have realized this problem and found the way to confirm after I submitted the question here. So, the VASP from my former group member which is supposed to have the link to VTST code doesn't have it any more! So, I am recompiling my VASP with VTST code embeded and hope the Cl-NEB method will solve my convergence problem.
Thanks again for all your work.
Jinlan
Post Reply