Compute diffusion in a complex geometry

Hi all,

I am willing to compute mean square displacement (MSD) in a lets say complex geometry,namely a CNT junction composed of for example (10,10) and a (6,6). In first place I thought to put PBC at the two sides but it is not correct for many reasons. Then I thought to immerse it in a box but then again create a region with lammps commands but then again it would be hard to create the region accurately. Any suggestions ?

Regards
Jim

I don't understand your question. Presumably you
are going to create the atom coords yourself and input
them via a data file. The only remaining Q is what
boundary conditions to use via the boundary command.
It's options are documented. It's up to you to figure
out which ones match your model ...

Steve