[lammps-users] airebo, hits with wall, fix deposit

LAMMPS LAMMPS Mailing List Mirror
1

Hi all,

I have collected a few questions:

  1. I have found that the airebo potential is only working with atom_style atomic, while other styles esp. bond gives wrong or no results. Maybe you could specify this in the manual more precise or change it in the code.

  2. I want to calculate the hits of atoms with a wall (e.g a CNT): Is there a fast and simple way to do this with lammps? I thought about compute coord/atom or postprocessing atom coordinates.

  3. I would like to add molecules (e.g. methane, however not as a united atom) occasionally to my simulation space. Would it be possible to change the fix deposit command for doing this, or is there any other way?

Thanks for your help

Thomas

2

I have collected a few questions:

1) I have found that the airebo potential is only working with
atom_style atomic, while other styles esp. bond gives wrong or no results.
Maybe you could specify this in the manual more precise or change it in the
code.

I just ran one of my AIREBO test cases with atom-style atomic and
atom_style bond and they gave identical answers. I had to change
the data file to include the molecule ID that bond expects.

2) I want to calculate the hits of atoms with a wall (e.g a CNT): Is
there a fast and simple way to do this with lammps? I thought about compute
coord/atom or postprocessing atom coordinates.

What defines a "hit"? What defines the "wall". If you can specify
that you could write a fix that monitors and detects hits.

3) I would like to add molecules (e.g. methane, however not as a
united atom) occasionally to my simulation space. Would it be possible to
change the fix deposit command for doing this, or is there any other way?

Anything is possible. Adding an entire molecule is hard for several reasons.
You need to find space for it, you need to create all it's topology
etc. Fix deposit
could be enhanced to do something like that, but it would be non-trivial.

3

Dear Steve,

Here I give you my two input files for an example for diverging results. I create the input files with Topotools from vmd and they consist of a bunch of methanes.

About definition of hit and wall:
I would define a hit of an atom with the wall, using compute coord/atom. By setting a cutoff value to appr. 0.5nm, a hit of a single atom could be specified for coordination numbers originating from near "wall" atoms of greater than 1.

Best Regards

Thomas

1) I have found that the airebo potential is only working with
atom_style atomic, while other styles esp. bond gives wrong or no results.
Maybe you could specify this in the manual more precise or change it in the
code.

I just ran one of my AIREBO test cases with atom-style atomic and
atom_style bond and they gave identical answers. I had to change
the data file to include the molecule ID that bond expects.

2) I want to calculate the hits of atoms with a wall (e.g a CNT): Is
there a fast and simple way to do this with lammps? I thought about compute
coord/atom or postprocessing atom coordinates.

What defines a "hit"? What defines the "wall". If you can specify
that you could write a fix that monitors and detects hits.

3) I would like to add molecules (e.g. methane, however not as a
united atom) occasionally to my simulation space. Would it be possible to
change the fix deposit command for doing this, or is there any other way?

Anything is possible. Adding an entire molecule is hard for several reasons.
You need to find space for it, you need to create all it's topology
etc. Fix deposit
could be enhanced to do something like that, but it would be non-trivial.

in_atomic.txt (421 Bytes)

run_atomic.data (2.9 KB)

run_bond.data (4.19 KB)

in_bond.txt (410 Bytes)

4

I didn't know you were actually defining bonds in
your data file (why would you do that for AIREBO?).
In which case, you need to use

special_bonds lj/coul 1 1 1

else you will exclude neighbor pairs from the AIREBO
neighbor list. It needs to see all the atom pairs.

Also the periodic box sizes are different in your 2 data
files. If you make them the same, the the 2 sets of
input/data files give identical answers.

Steve