[lammps-users] random force term in fix_langevin

LAMMPS LAMMPS Mailing List Mirror
1

Dear Steve and all,
I recently try to use Langevin thermostat for the NEMD simulation. However, when I read the script, I didn’t understand why the prefactor of the random force term is 24.0 in

gfactor2[i] = sqrt(atom->mass[i]) *
sqrt(24.0*force->boltz/t_period/update->dt/force->mvv2e) /
force->ftm2v;

or

gamma2 = sqrt(rmass[i]) * sqrt(24.0*boltz/t_period/dt/mvv2e) / ftm2v;

According to all the literatures I have found, should this prefactor be 2.0? In the pair_dpd or pair_dpd_tstat, they have similar random force term, but for dpd, the prefactor is 2.0 like

sigma[i][j] = sigma[j][i] = sqrt(2.0boltztemperature*gamma[i][j]);

Best wishes and happy holiday,
Frank

2

The 24 has to do with using a uniform random #, instead
of a Gaussian random #. I think the Dunweg ref on the
fix langevin doc page discusses this.

Steve