new finds about "Broken bonds or deformation of molecules in long run using ML_FF after refit"

Message

suojiang_zhang1 · #1 Post by **suojiang_zhang1** » Tue Jul 11, 2023 6:32 am

Hi，
following my previous post titled "Broken bonds or deformation of molecules in long run using ML_FF after refit"
firstly, many thanks for your suggestions.

If I build a supercell via expanding CONTCAR at xyz directions, it is nice to run the system using ML_FF, there no molecules were found to deformed or bond-broken.
But if I build a larger box from scratch including same molecules, some molecules were found to deform or bond-broken during the run MD using ML_FF.

I dont understand the physics.

#2 Post by **jonathan_lahnsteiner2** » Tue Jul 11, 2023 5:57 pm

Dear suojiang_zhang1,

Could you please supply your input files such that I can rerun your calculation.
How to upload files is described in the vasp forum guidelines.
https://www.vasp.at/forum/viewtopic.php?f=4&t=17928
I would need the POTCAR, KPOINTS, ML_FF, INCAR files and the POSCAR file of the calculation which goes wrong.
Without the files it is not possible to tell where the issue in your computation is originating from.

All the best

Jonathan

My Community

new finds about "Broken bonds or deformation of molecules in long run using ML_FF after refit"

new finds about "Broken bonds or deformation of molecules in long run using ML_FF after refit"

Re: new finds about "Broken bonds or deformation of molecules in long run using ML_FF after refit"