time step for reax

LAMMPS LAMMPS Mailing List Mirror
1

Dear Lammps Users & Developers

I am using the Reax/C potential for a simple tensile simulation.

I found the time step for Reax potential in some input script and literature smaller than 1 femto second (~0.5 fms).

I used the time step = 1 femto second and it seems there is no any visual problem and the temperature fluctuation is very reasonable, and the results seems not bad.

but I am worry because of qeq, I must use time step smaller than 1 fms that have very computational time cost. I tried to understand is it necessary for very small time step and why, but I cant find any valuable reason. please advise me

Thanks in advance

Hadi

2

I guess you need to improve your searching skills. The question of how to choose a suitable time step has been discussed on this very list many times. It also is a fundamental topic in any text book on MD. In fact, if you cannot find an answer to this, you should not be running MD.

Axel.

3

Dear Axel

Thanks for your attention. I searched in mailing list but couldn’t valuable topic for suitable timestep and reax.
I think because of qeq in reax potential there is a limitation to choose timestep. I study some documentation from van Duin

and the effect of some parameters like charge equilibrium tolerance but cant clear notation for time step. but I need some advise for timestep in reax potential. if it is possible please give me a reference that describe suitable timestep for reaxff potential

Thanks

Hadi

4

Dear Axel
Thanks for your attention. I searched in mailing list but couldn't valuable
topic for suitable timestep and reax.
I think because of qeq in reax potential there is a limitation to choose
timestep. I study some documentation from van Duin
and the effect of some parameters like charge equilibrium tolerance but cant
clear notation for time step. but I need some advise for timestep in reax
potential. if it is possible please give me a reference that describe
suitable timestep for reaxff potential

you *still* need to improve your searching and problem solving skills.
you have looked for the wrong answer. the procedure of finding a
suitable time step in an MD run is completely independent of using
reaxff or charge equilibration. true, they have an effect, but so do
other force field components. as i said, this is standard text book
stuff. obviously, this is the place where you have not looked.

axel.