nano patterns

I don’t need to model any specific physical process. Nothing like lithography or embossing…

I was trying to use these processes (to actually implement them) in lammps to get a nano pattern. So for example I would have some diamond mold pressed into my glass to indent it and create the patterns. It didn’t work out as it should. It’s not as fast as it should be either.

Now I’m looking at another more commonly used way of producing nanopatterns from a bulk structure and that is simply deleting atoms. In this way I would delete atoms from the box to get the pattern that I want. (This pattern can be a cactus, a cylinder, etc)

My problem here is that the material I am working with is amorphous. That means there is a chance I end up with the chopped up polymer structures (if I am deleting atoms from an amophous polymer cell, and these are usually unphysical) or a charged structure (if let’s say I end up deleting more Si than O from my glass).

My point is while the deletion is taking place in VMD or lammps itself, there seems to be no straightforward way to guard agains these problems (also described below).

So my question has to do with the quality of the way I am setting up the simulation.H

Is it acceptable to keep the distorted structures if I choose to freeze them for the other simulations?

It really is not a lammps question, but I’m flying solo on this one. So the smallest threads form the community are a big help.


If you want to create some complicated structure as input
to an MD simulation (e.g for LAMMPS), then I would
write your own program to put atoms exactly where you want.
Just have that tool write out a data file for LAMMPS to input,
and you can do whatever you want. So long as the restulting
structure doesn't have overlapping atoms, you should be
able to minimize its energy and start doing MD with no problem.