What is the minimum computational temperature in LAMMPS?

LAMMPS LAMMPS General Discussion
1

We know that T=0K is not allowed for calculating the Nose-Hoover thermostat. Explicitly (fix nvt command — LAMMPS documentation):
" For the temp keyword, the final Tstop cannot be 0.0 since it would make the external T = 0.0 at some timestep during the simulation which is not allowed in the Nose/Hoover formulation."

Then, what is the lowest temperature we can perform simulations with LAMMPS using the Nose/Hoover formulation?

2

That depends a bit on multiple factors including your choice of units and size of time step. You need a small number that is large enough so you can still get a stable time integration of the fictitious degrees of freedom for the thermostat coupling. Try 1.0e-10 or even 1.0e-100.

3

Dear Akohlmey,

Thank you for the reply. In my case, T=0.2 K and upper temperatures have been successful. However, T=0.1 K and lower temperatures failed.
units metal
timestep 0.0005

Kind regards,
Rasoul

4

I forgot to mention, that your choice of tdamp in the thermostat also has a significant impact.

That said, running MD at such low temperature is a pretty pointless exercise, since atoms won’t be able to move much due to the low kinetic energy. You could just as well perform a minimization and save some time.

5

That might be a good idea. The minimization itself works well. I tried for T= 1.0e-30 K. Thanks again.