Dear Lammps users,
I ran a simulation study of NAS (sodium alminosilicate glass) structure using reaxff. The structure has successfully equilibrated, but when I tried to heat it from 300K to 4000K @…8150…/ps in npt ensemble, I got an error of loss of atoms at 3000K. Any suggestions are highly appreciated.
I have attached the log files and the input files.
Thanks.
log.lammps (367 KB)
NASinput.in (2.19 KB)
Dear Lammps users,
I ran a simulation study of NAS (sodium alminosilicate glass) structure using reaxff. The structure has successfully equilibrated, but when I tried to heat it from 300K to 4000K @10K/ps in npt ensemble, I got an error of loss of atoms at 3000K. Any suggestions are highly appreciated.
I have attached the log files and the input files.
at higher temperature, atoms move faster, thus you may need a shorter time step.
1 fs is quite large for reaxff anyway.
apart from that, i don't think it is a good idea to heat and cool down
with fix npt. it is very easy for a system to expand, but very
difficult and time consuming to properly shrink back. so you should do
your heat/cool cycle with fix nvt and only at the very end relax with
fix npt after you have relaxed with fix nvt first.
axel.
Dear Axel,
Thanks for your quick response. Initially, I tried with a heating/cooling cycle in nvt ensemble, but the structure created a hole at the end of the simulation. It is informed to me that such hole is created because I ran in nvt ensemble (as the volume of liquid is more than the volume of solid).
This is almost certainly a problem with the timestep. Take the last restart file and do some experiments with NVE dynamics with different step sizes. Look at energy conservation and overall stability.
Aidan
Thanks, Aidan for the information.