Brownian Dynamics in oscilatory shear

LAMMPS LAMMPS General Discussion
1

Dear LAMMPS users,

I am trying to shear Brownian particles interacting via WCA potential by applying Brownian Dynamics with implicit solvent using a combination of this commands

pair_style lj/cut 1.122462048
fix brownian …
fix deform wiggle … remap x

In order to calculate G’ and G’’ and study the rheology of the system.

In the past, I have already done simulations with DPD + fix deform remap v for a different project with good results, but I don’t know if I am doing things right in this case for Brownian Dynamics, as I am not very clear if LAMMPS is doing what I want or if I should use another combination of commands like fix/brownian + fix/addforce to apply the shear flow.

In my simulations I want to study different range of density values. Even if my system is a fluid I use “remap x” since the velocity in Brownian dynamics is not well defined and I think it would be a mistake to use “remap v”, however I am not very sure about this point, could you please confirm that this makes sense? Also, at high density when my system behaves like a solid (and according to the “deform” command guide) it is advisable to use “remap x” for these cases.

In other similar studies I have seen that the community uses their own codes with periodic Lees and Edwards conditions. In another answer in this forum they said these conditions are not implemented in LAMMPS but there is a combination of equivalent commands, but using nvt thermostat (Lees Edwards boundary conditions for shear flow). Since I use fix/brownian and fix/deform, is this a good combination to get these boundary conditions in LAMMPS?

Any help would be appreciated
Thanks
J

2

The post you reference is twelve years old. A lot has changed since then!

fix deform with changing box angles over time is equivalent to Lees-Edwards boundary conditions. What fix nvt/sllod tries to do is subtract the streaming velocity from temperature calculations for proper thermostatting. Note that fix nvt/sllod assumes a homogeneous streaming velocity profile through the box – if you suspect that won’t be the case you’ll need compte temp/profile as a “profile-unbiased thermostat” to avoid artefacts like “string phases”.

1 Like
3

@srtee Thank you very much for your reply

I have been doing some test and checking the source code, but I don’t quite understand completly, because in my case when I use “remap x” there is no contribution in the speed of the particles, only in their position (or at least that’s what I understood from the documentation of “fix deform”), so it shouldn’t be necessary to apply any “compute temp/deform” or “compute temp/profile” to my system to unbias velocities, right?

Thank you very much for your help

1 Like