Dear experts,
I am a newbie to NEB calculation and VASP. I am trying to study decomposition of HNNOH molecule to nitrogen (N2) and water (H2O). Although my images look correct with no overlapping of atoms as such, but still during the neb calculations my images split weirdly into individual atoms. I am not sure what's wrong with my calculation.
Procedure I followed -
1) obtain optimized structure of first and last point.
2) create 5 images for CI-NEB using linear interpolation
3) Run CI-NEB using VASP
I have attached my INCAR and images trajectory at the start of the simulation and in between the simulation.
I look forward to some help and suggestions on this.
Regards,
Raghav
Hello Raghav,
I looked through your final.xyz and saw that the H2O is already formed by frame 2. It looks to me that you are being too aggresive with your inputs and your search for the solution. A few things to try:
1. Make sure that there is no obvious 'atom overlaps' in your linearly interpolated inputs. When atoms overlap they cause huge potential energy rise and VASP will be rattled by this.
2. Try to do NEB without the Climbing Image first. With decreasing EDIFFG. Start with 0.1 eV/Angs, then 0.05, then 0.02 eV/Angs; and then do 0.02 eV/Angs again with Climbing Image switched on to find the transition state.
Hi Kelvin,
Thanks a lot for the answer. You are correct.
I am not sure how I missed it, but after having a closer look at my images I realized there was indeed an overlap b/w my nitrogen atoms and that lead to images blowing up. The problem solved after I fixed my images.
Best,
Raghav
Kelvin Anggara Raghav Saxena I have 5 images and the first 3 have converged really well but the last are not. The funny thing is the structures are way different from the first three. In fact some of the atoms have moved from the surface. Please any reason why? And what better way can I fix this. Thank you.
Hi Dennis,
Hard to comment on this without further information. VASP sometimes could converge to different solutions if you change your initial and final states; and/or change the number of images between initial and final states. Perhaps you could share some files here?
- Badges
- Science topic
- Similar topics
- Comment