Adjust cell size

LAMMPS LAMMPS General Discussion
1

Hi
how can i adjust cell sizes to give zero pressure during the cooling process i think fix 1 all npt temp 3000 200 0.1 iso 0 0 1 does not work, here is supplementary infos : systems was melted and equilibrated at 3000 K - EAM potential - timestep 1fs - cooling rate of 10^12 K/s
Thanks in advance

2

Please note that “does not work” is a very unscientific expression and thus difficult to give meaningful advice on. Instead you should explain what you expect to see, what you get instead and why you believe that is incorrect.

3

thanks for replying
i mean pressure never reach 0 bar and huge fluctuation “4000 bar” and even after the cooling process end i used npt to relax the system but the pressure keep fluctuating without ever reaching 0 my objectif is to relax the stress to deform the system afterward and also box/relax command leave my system unchanged

4

That is not the fault of fix npt, but due to a lack of knowledge in statistical mechanics.

Pressure always fluctuates, but the magnitude of the fluctuations depend on:

  • compressibility
  • system size
  • averaging

For solids at the length and time scales for typical MD simulations, pressure fluctuations are always very large. If you increase your system size (try adding “replicate 2 2 2” to your input in the proper place) you should see the fluctuations decrease. And if your system is in equilibrium, you can also average over time (i.e. via fix ave/time).

Finally, since it is easier to expand a system, then to compress it (atoms are in their own ways when the system is shrinking), you are not likely to reach exactly zero pressure, but likely it will be close enough. Ultimately, if the average value is substantially smaller than the fluctuations, it is close enough.

1 Like
5

It sounds like you are trying to simulate the liquid-to-glass transition with a constant cooling rate and constant stress conditions. There is a vast literature on this topic going back decades. There are many technical challenges, and describing them goes far beyond the scope of this forum. A key point, made by Axel already, is that pressure in MD is never zero, but only close to zero on average. Another issue is that for a solid, you can have zero average pressure, but still have very large average shear stresses in the structure. Hence it is a good idea to replace “iso 0 0 1” with “tri 0 0 1”. Finally, if you want to produce a fully relaxed static configuration at a local energy minimum with zero forces and zero stress tensor, then start from a low temperature, low stress configuration, and use a combination of “minimize” and “fix box/relax”

1 Like