Hi everyone,

I build a layer structure consisting of two parts, in which part 1 is fixed and part 2 moves towards it along z direction. The distance between two parts is 10Å.

NVT dynamics simulations are conducted for this layer structure under 333.15K for 300ps and 1000ps. Finally, part 2 is in contact with part 1. Then I analyze mean square displacement (MSD) for part 2 based on trajectories.

MSD results are shown in the following figures. Since part 2 diffuses along z direction, only total MSD and zz component are provided here.

 For 300ps simulation time: 

For 1000ps simulation time:

From these figures, I find three phenomena: 

(1) trajectories with same simulation time have different MSD results. 

(2) trajectories with different simulation time also have different MSD tendencies.

(3) I thought that because part 1 is fixed, zz MSD of part 2 will increase at first and after contact of two parts, its increasing speed will gradually decrease and finally zz MSD of part 2 will keep constant. But from those figures, zz MSD still increases rapidly even after a long simulation time. 

I don't know the reasons for these phenomena and I also don't know which trajectory and what simulation time are proper for diffusion coefficient calculation of part 2. Can anyone give me some advice? Thank you very much.