From: Philippe Bopp (
Date: Tue May 01 2018 - 21:38:17 CDT


 <!--#yiv5153370203 P {margin-top:0;margin-bottom:0;}-->>Axel, the  problem I have is that I have 3 different molecules in my simulation. Therefore, if I use the lammps code, I >am not sure of the validity of the result since the molecules are not the same.
 you seem to be confused about diffusion coefficients. The SELF-diffusion coefficient is a single molecule property,
 i.e it 'looks' only at one (type of) molecule, independently of what the neighbors may be. Of course, the value you find will depend on the composition/phase/....of your system, but it is still associated with only one (type of) molecule
The INTERdifusion is a different story, check books on diffusion,e.g. by Jörg Kärger (Leipzig)

>Instead, what I did was use the chunk command to output the COM of the different molecules into a dump file and try >to use topotools to convert the dump file into a xyz file so that VMD can open it. But right now I'm getting this error.>
>readvarxyz: read in 0 frames requiring 0 atoms storage.Type map:Max type counts:0

how does the xyz file look like?

and, as Axel mentioned, you must take care of the box-periodicity when studyingdiffusion.
If your box is constant, (NVsometing) simulation, set the box once and for alleither with a command or in the Extensions-> Analysis-> Radial Pair ...... (top left)
If your box is not constant, (Npsometing) simulation, you can- either use the average box if your system is nicely at equilbrium  which usually introduces only a small error
- or do some tricks, dut diffusion in non-equilibrium systems is not well defined
-then unwrap your trajectory
- if you want self-diffusion coefficients of specific molecules (e.g all A molecules  that are surrounded by B molecules), you can use select commands,
 However you must be careful Either you say: - I want D_s of the whole, unchanged group AB_n , but that may decay too fast, or
 - I am happy if A was surrounded by Bs (n of them, say) at the origin, and I don not   care how the neighborhood of A changes during the diffusion, or
 - you can think of other cases

Philippe A. Bopp
professor of chemistry (retired)

philippebopp  AT
eSSENCE modelling meeting 11-13 June 2018