Dear Professor,
I have tried to use your solid state NEB to find minimum energy path for the phase transition from MoS2 2H phase to 1T phase recently, in the first 90 steps, the force vibrates a lot, one image even exceeded 9eV/A, the energies on the image are zigzag along the reaction coordinate, rather than just one peak. So I stoped the job. I had read a paper from your college published on Nature Communication where the author also used your ssneb method to calculate the same system, viz bulk MoS2, but they got the perfect reaction path in there supporting material.
This is the article's link:
https://www.nature.com/articles/ncomms9311
I had uploaded my work, could you please take a look and try to find a way to make it converge if you have time.
Sincerely
Thank you!
SSNEB failed
Moderator: moderators
SSNEB failed
- Attachments
-
- ssneb.tgz
- (30.02 MiB) Downloaded 1045 times
Re: SSNEB failed
Here are a few suggestions:
First, make sure that your endpoints are minimized.
Second, you need to think about a sensible mapping between the order of atoms in your initial and final structures. The ordering of the atoms determines which atom in the initial structure moves to which final state geometry. In you case, the S atoms were moving too far, over Mo atoms, so that it is unlikely to be an elementary process. As you found, such paths can break up into a series of steps.
In the attached calculation, I have used a smaller unit cell and reset the mapping between atoms in the initial and final state. This is not fully converged, but I think it is sufficient to get you on the right track.
First, make sure that your endpoints are minimized.
Second, you need to think about a sensible mapping between the order of atoms in your initial and final structures. The ordering of the atoms determines which atom in the initial structure moves to which final state geometry. In you case, the S atoms were moving too far, over Mo atoms, so that it is unlikely to be an elementary process. As you found, such paths can break up into a series of steps.
In the attached calculation, I have used a smaller unit cell and reset the mapping between atoms in the initial and final state. This is not fully converged, but I think it is sufficient to get you on the right track.
- Attachments
-
- ssneb.tar.gz
- (7.89 MiB) Downloaded 1085 times
Re: SSNEB failed
Thank for your valuable suggestions and your time. I will keep that in mind.