How to optimizes the atom positions and box at the same time

eOn code for long time scale dynamics

Moderator: moderators

Post Reply
WHU_hyx
Posts: 3
Joined: Wed Sep 21, 2022 8:40 am

How to optimizes the atom positions and box at the same time

Post by WHU_hyx »

According to the document(https://theory.cm.utexas.edu/eon/optimi ... #optimizer)
i set "opt_method = box" in config.ini, however i got "Unknown optMethod: box" in log file.
And i find that there has no parameter "box" in Optimizer.cpp.

So I'm wondering how to optimizes the atom positions and box at the same time.(my eon version is eon-2470)
graeme
Site Admin
Posts: 2253
Joined: Tue Apr 26, 2005 4:25 am
Contact:

Re: How to optimizes the atom positions and box at the same time

Post by graeme »

This may be a problem with the documentation; we do not have support for cell relaxation in EON.
WHU_hyx
Posts: 3
Joined: Wed Sep 21, 2022 8:40 am

Re: How to optimizes the atom positions and box at the same time

Post by WHU_hyx »

Dear Professor Graeme Henkelman,

Thank you very much for for your reply, but I still have a question---------For a solid-solid phase transition, how can I use EON/TSASE to calculate the phase transition process ?

Specifically, I tried to calculate the phase transition process from Mo A15 to BCC, like this article(doi.org/10.1063/1.4873437). And now I can get the structure of saddle point through the SSD(solid-state dimer) algorithm in TSASE.
But I still don’t know how to get the product from the saddle point, because EON do not have support for cell relaxation, and the SSD method just can calculate the structure of saddle point.Is I missed some important information in the article or in software document?

Any help will be appreciated.
WHU_hyx
Posts: 3
Joined: Wed Sep 21, 2022 8:40 am

Re: How to optimizes the atom positions and box at the same time

Post by WHU_hyx »

graeme wrote: Thu Sep 22, 2022 2:03 pm This may be a problem with the documentation; we do not have support for cell relaxation in EON.
Dear Professor Graeme Henkelman,

Thank you very much for for your reply, but I still have a question---------For a solid-solid phase transition, how can I use EON/TSASE to calculate the phase transition process ?

Specifically, I tried to calculate the phase transition process from Mo A15 to BCC, like this article(doi.org/10.1063/1.4873437). And now I can get the structure of saddle point through the SSD(solid-state dimer) algorithm in TSASE.
But I still don’t know how to get the product from the saddle point, because EON do not have support for cell relaxation, and the SSD method just can calculate the structure of saddle point.Is I missed some important information in the article or in software document?

Any help will be appreciated.
graeme
Site Admin
Posts: 2253
Joined: Tue Apr 26, 2005 4:25 am
Contact:

Re: How to optimizes the atom positions and box at the same time

Post by graeme »

That is a good question/comment. TSASE has code for solid-solid phase transitions, using any ASE calculator including LAMMPS and VASP. For concerted phase transitions, cell changes are essential.

We have also studied solid-solid phase transitions using EON, but in that case, it is entirely modeled by atomic motion within fixed cells.

Using TSASE, you should be able to get the saddle for a reaction mechanism involving cell changes. Simply relaxing from the saddle along the negative mode will give the initial and final state.

If you want more help, send me an email and I will forward it to Penghao Xiao, who is the expert for these calculations.
Post Reply