Dear all,

I want to calculate the msd of Ga and As adatoms on the GaAs surface. I just put 10 Ga adatoms radomly on the surface and use the "compute msd“ to obtain the msd, run for 100ps, using tersoff potential. However the result of msd is like a line. I saw the frames by VMD, it seems that Ga adatoms form bond with the surface. So I am wondering whether we can calculate the msd when the Ga adatoms form bond with the surface and how to solve it to obtain the proper msd for Ga adatoms?

Best

XJ

If you're saying the atoms bond to the surface and don't

diffuse, then the MSD will be a horizontal line. If you

want the atoms to diffuse, then you need to change

your model, not the MSD calculation.

Steve

Many thanks for your reply. I used the tersoff potential, if the distance larger than 2.7A ,then no interaction between Ga adatom and the surface atoms, if smaller than 2.7A, then bond will be formed. how can I change the model?

And I found that at first 2100steps, the msd is a slip line, after 2100 steps, msd is a horizontal line, can I use this first part as the correct data as the diffusion coefficient ?

Best

XJ

I doubt it. For a diffusing system, over long timescales,

the MSD should be a noisy line with positive slope. For

less than 2100 steps, I imagine it is just the motion of

atoms around their lattice sites.

Steve

Ok, I see, than hoe to change the model? I will be appreciate it if you can give me further suggestions.

Changing the potential?but I want to study the diffusion behavior under tersoff potential.or it only can be studied by DFT? Thank you again for discussion to make the question clearer.

Best

XJ

在 2012-07-26 23:16:48，"Steve Plimpton" <[email protected]> 写道：

I send you the msd data, there seems a a noisy line with positive slope when I simulate 50ns , but still feel some strange.

I thought a model that fix the z position of Ga adatoms to prevent them form bond with surface but still have some VDW interactions with surface and try to see the results.

However, I am still a little bit puzzled about how to figure out the diffusion behavior of Ga adatoms along z direction as in new model, Ga adatoms only allow to move in xy plane.

Many thanks for discussion,

Best

XJ

msd_TopGa.dat (1.88 MB)

I suggest you find a paper in the literature that models surface

diffusion of yours (or a similar system), and reproduce the

calculation in LAMMPS. That will get you started.

Steve

Tersoff potential responded to your setup in a way it should respond.

If the potential energy is decreased when adatoms bond to the surface,

then without strong enough driving force, why would the adatoms

diffuse?

Even if you fix z positions of the adatoms, they are just going to

bond or settle to a local minimum on the x-y plane. You need to read

relavant papers and properly drive these adatoms.

Best,

Ray

Many thanks, I have found the reference for simulating Ga diffusion behavior on GaAs surface.

But no details for how to add strong enough driving force. Shall I use the langevin dynamics to simulate and obtain enough driving force?

Best

XJ

Hi XJ,

I am not sure how to model surface diffusion with MD, but I am sure

there are quite a few publications in this regard. You can also look

at simulation works using SW and BOP, not just Tersoff. Or you can

contact the authors of the publications for more details.

Also, are you sure MD is the best choice of studying diffusion?

Usually MC is a better choice.

Best,

Ray