LAMMPS WWW Site - LAMMPS Documentation - LAMMPS Mailing List Archives
[lammps-users] How to estimate time constant for thermostat (Berendsen, Langevin)?
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[lammps-users] How to estimate time constant for thermostat (Berendsen, Langevin)?


From: Christophe Ortiz <christophe.ortiz@...4080...>
Date: Thu, 20 Jul 2017 13:20:31 +0000

Dear LAMMPS users,

As you know, with LAMMPS we can impose that some group of atoms to be at a certain temperature with a fix. For instance with a Berendsen or a Langevin thermostat:

fix            thermostat thermal_atoms langevin ${temp_thermo} ${temp_thermo} 0.1 ${seed3}

fix             thermostat thermal_atoms temp/berendsen ${temp_thermo} ${temp_thermo} 0.1

In the case of collision cascades in a bulk, this is done to mimic the heat conduction with the bulk.
However, the rate at which this occurs ("...how rapidly the temperature is relaxed.") depends on the time constant (damping parameter).
If one wants to simulate in a realistic manner the way heat is transferred to the bulk, one needs thus a realistic value of the damping parameter. Indeed, it is not the same to use a time constant of 1.0 than 0.1. If the damping parameter is too small, it is likely that heat will be extracted too quickly and will freeze the cascade, which would affect the results.

I was thus wondering if there is a way to estimate the time constant corresponding to a given material (Fe for instance).

Any idea? Does it make sense?

With best regards,
Christophe