Msd graph of dpd simulation

LAMMPS LAMMPS General Discussion
1

Hello everyone

I ran the following dpd simulation. If you want to find the diffusivity according to the msd value and obtain the slope value by displaying the graph, it is different from the paper value. Can I get some advice on this?
input_1.txt (1.1 KB)
log.lammps (24.1 KB)

msd graph
msd graph1090×802 34.6 KB

2

I don’t know what the paper value is, but it looks from the figure that your simulation time is too short such that the atoms have not entered the ‘diffusion’ region yet. You can plot log(MSD) versus log(t) and see if the slope is close to 1 (which indicates the diffusion region). You can check https://doi.org/10.1016/j.ijhydene.2022.01.023 (Is hydrogen diffusion in amorphous metals non-Arrhenian?) and references therein for more information.

3

If you find the slope through the msd graph, can you tell me how to get D?

4

Your input has two problems:

  1. You did not “unfix” your nve integrator before fixing your nvt integrator, so after that point your particles are being integrated twice and the trajectories are invalid.
  2. DPD style particles have an internal temperature, but I did not see where you set the internal temperature of the particles you created. So the initial fix nve will have transferred the initial translational thermal energy into internal temperature and left your system at some temperature below 1.0, and the subsequent fix nvt (if it had worked) would have “started cold”.
5

Just use the Einstein relation to get D. Note in the relation, time t is to infinity (or very large), which means that atoms are in the diffusive regime. You can get more info from some MD textbook or paper, such as the references I mentioned.