[lammps-users] Mapping LJ unit and Langevin parameters to a real system

Dear Lammps Community,

I would like to map the LJ and Langevin parameters to a real. By using a=100 nm, T=298 K, and m=18.54*10^4 Da, I get epsilon=4.11 zJ and tau=27.34 ns. However, I do not know how to appropriately mapping friction coefficient. Based on Lammps’s documentation and dimension analysis, for damp=10 and a dt=0.005, I get the following results:

dt = 0.00527.35 ns = 0.136 ns,
damping constant = 1/(damping factor
tau) = 3.6510^6 s^-1,
friction coefficient = mass
damping constant = 1.226 fkg / s,
diffusion coefficient = epsilon/friction coefficient =3.657 mm^2/s,
viscosity = 1.19 nKg/(m.s) — Based on Stokes-Einstein formula for a sphere,

all measured in SI units.

Are my conversion formula for damping constant correct and thus the subsequent results correct?

I also have one further question about the meaning of damping factor: If we set damp=10, do we expect to see diffusive behaviour approximately in time steps larger than damp/dt? In my example above, time steps >> 10/0.005 (or time steps >> 2000)?

Thank you in advance for your response.
Kind regards,
Amir

It is unlikely that somebody has the time to proofread your computations.
The examples/UNITS folder has some examples for liquid argon on how to do the same simulation in different units.
For understanding the physical interpretation of the damping factor, you need to read the relevant published literature, and if that is not sufficient consult with somebody that performs research in that area.

Dear Axel,

Thanks for your response. My question was actually about the order of magnitude of Langevin-related parameters - Sorry for asking it vaguely. I got some ideas from "1990 - Kremer K Grest GS - Dynamics of entangled linear polymer melts A molecular-dynamics simulation” article.

Kind regards,
Amir