[lammps-users] Question about Diffuse Boundary Condition

Hello everyone,
I want to calculate the thermal conductivity of thin films by using the Langevin thermostat. At the boundaries, instead of specular reflection (it reverses the normal of velocity vector), I want to use diffuse reflection (similar to real boundaries, it randomly assigns a velocity sampled from a distribution velocity). So, does the Langevin thermostat do the same thing? Or should I still apply a wall condition?
Thank you in advance,
Onurcan

Hello everyone,
I want to calculate the thermal conductivity of thin films by using the Langevin thermostat. At the boundaries, instead of specular reflection (it reverses the normal of velocity vector), I want to use diffuse reflection (similar to real boundaries, it randomly assigns a velocity sampled from a distribution velocity).

boundaries do not interact with atoms. you need to place a wall fix on a boundary to have atoms repelled at a boundary.

So, does the Langevin thermostat do the same thing?

fix langevin does not know anything about boundaries or walls. it always follows the implemented and documented algorithm.

Or should I still apply a wall condition?

there are quite a few different kinds of walls in LAMMPS including a variant that with stochastic reflection.
https://docs.lammps.org/Commands_fix.html

axel.