Convergence problems with NEB calculations

Message

Kim Do Hwan · #1 Post by **Kim Do Hwan** » Thu Aug 04, 2005 4:24 am

Hi,

I've been trying to perform VASP calculations for my system, especially organic molecules on semiconductor surface.
The calculations are O. K. in case of geometry optimization.
However, I can not find transition state with NEB method.
Each of the images do not converge even after 1000 runs.
I'm afraid I have used inapproprate variables in INCAR options, or that it may be difficult to find transition state for organic molecule on semiconductor system.

I used 3 or 5 images.

My settings for NEB calculations are

EDIFF = 1E-05 for electronic relaxation
ICHAIN = 0
EDIFFG = -0.1
IBRION = 3
POTIM = 0.05
SMASS = 2

Please inform me any method to solve the problems.

Thanks,
Kim, Do Hwan
<span class='smallblacktext'>[ Edited ]</span>

#2 Post by **admin** » Fri Aug 05, 2005 10:36 am

for the VASP releases distributed by us, the INCAR tags to
do an NEB run are
IMAGES =
SPRING = -5

Brane · #3 Post by **Brane** » Wed Aug 10, 2005 12:24 pm

I think he used the improved NEB program provided by Dr. Henkelman. In the improved NEB, ICHAIN=0 is for NEB, and SPRING=-5 is default. Maybe the small POTIM slows the convergence.
Try POTIM = 0.1 and remove SMASS parameter after some steps.

<span class='smallblacktext'>[ Edited Wed Aug 10 2005, 02:34PM ]</span>