shear wave in equilibrium simulations

Dear Lammps users/developers,

I have an issue that I thought you might have encountered before. I simulated a 3D system of particles with binary Lennard-Jones (Kob-Anderson 80:20) potential in glassy/liquid state. The simulations does not involve any loading. So therefore I expect only some fluctuations in displacement of the particles. But when I analysed the displacement profile of my system, I see rather oscillations in a wave form rather than fluctuations (the displacement profile is attached). The system is fairly big (N>1e6). I have used both current version and some older version.

Details of the simulation:

system is equilibrated in liquid state (NVT) T=2.0 (T_g~0.4), brought to the final temperature (NVT) (T=0.6 or T=0.2), equilibrated/aged at the final temperature long enough, then displacement is analysed.

I would really appreciate if anyone can help me to avoid this large (in my view non-physical) fluctuation.

disp-Profile.eps (27.8 KB)

First of all, please always attach your input files, it makes thinking a lot easier.

From the description, this might be an artefact of the thermostat. You could use fix momentum to remove the linear momentum of the system every so many time steps, although that might not be of use if you are actually loading and stuff should move.

Have you tried using fix nve + fix langevin? If it is a thermostat artefact, this combo should give significantly different results.