Dear Prof. Henkelman,
I am new to using NEB calculations. I am trying to study the pathway of diffusion of Na atom from 1 site to the diagonally opposite site. (The original structure is Na16Ge16. I have created 1 Na vacancy and relaxed the structures.)
However the calculation seems to be extremely slow. I had used about 112 cores and the calculation still did not converge despite using NSW = 8000 steps. Could you please help me into understanding why this is happening and what I must do to correct this ? I have attached the files here.
Thank you.
Convergence issue with NEB
Moderator: moderators
Convergence issue with NEB
- Attachments
-
- Neb calculation.zip
- (305.85 MiB) Downloaded 548 times
Re: Convergence issue with NEB
If you look at the initial movie, you can see that there is a problem with the ordering of atoms between the initial and final states. See the attached movie, and how atoms are moving across the cell to take the place of identical atoms. You need to reorder the atoms between the initial states.
- Attachments
-
- movie.gz
- (7.91 KiB) Downloaded 481 times
Re: Convergence issue with NEB
Dear Prof. Henkelman,
Thank you for your reply. I'm sorry but I wasn't able to see the movie.(I used jmol). The atoms were stationary only.
Also could you explain what do you mean by reorder the atoms between the initial states?
Thank you.
Thank you for your reply. I'm sorry but I wasn't able to see the movie.(I used jmol). The atoms were stationary only.
Also could you explain what do you mean by reorder the atoms between the initial states?
Thank you.
Re: Convergence issue with NEB
Yes, so for example, the 27th atom in your 00/POSCAR is a Na atom at:
1.14333094938690 2.24730915067498 0.89735201226803 0 27
which becomes the 28th atom in your 08/POCAR:
1.13079578641343 2.48871545028892 0.79269253138857 0 28
you need to reorder the atoms in one of these files so that their identity is consistent. Otherwise the NEB is dragging atoms across the box (which your calculation is doing) simply in order to reorder them in the way you have demanded.
1.14333094938690 2.24730915067498 0.89735201226803 0 27
which becomes the 28th atom in your 08/POCAR:
1.13079578641343 2.48871545028892 0.79269253138857 0 28
you need to reorder the atoms in one of these files so that their identity is consistent. Otherwise the NEB is dragging atoms across the box (which your calculation is doing) simply in order to reorder them in the way you have demanded.
Re: Convergence issue with NEB
Thank you Dr. Henkelman, I shall try doing that.
Re: Convergence issue with NEB
Dear Dr. Henkelman,
The previous one worked out right, thank you for your help.
But i'm having trouble with another similar calculation. This time i rearranged the poscar files to match the identity of atoms. but the problem is that the calculation is stuck at 15 electronic steps and 1 ionic cycle. It is not moving beyond that. Could you please help with this?
Thank you.
The previous one worked out right, thank you for your help.
But i'm having trouble with another similar calculation. This time i rearranged the poscar files to match the identity of atoms. but the problem is that the calculation is stuck at 15 electronic steps and 1 ionic cycle. It is not moving beyond that. Could you please help with this?
Thank you.
- Attachments
-
- NaSite3toSite2_FINAL.zip
- (425.92 KiB) Downloaded 455 times
Re: Convergence issue with NEB
No, sorry, this is a completely crazy calculation. You need to be able to visualize the initial band. Try using the nebmove.pl script, or just cat all the POSCAR files and then use a program such as the ase-gui in ase or the xyz in tsase to visualize it. You will see immediately how the calculation make no sense.
Re: Convergence issue with NEB
Yes i saw the movie and it didn't make sense. I have corrected it now. Thank you for your advice.