[lammps-users] same group-->two regions --> two potentials

Hi all,

do you know if/how it is possible to use two different potentials for the same group in two different regions?

I have some atoms (let's say C). In a certain region (region1) they interact with a Fe cluster and I want the C-C interaction to be Teresoff. In another region (region2) however, where there is no cluster I want them to follow a simpler Lennard Jones.
Is there a simple way to obtain it? I cannot use different C groups since are the same atoms that going from region2 go to region1 must change potential.

Thanks

Alessio

If you make the C atoms in the 2 regions of different atom
types (both C, just different types), then you can use pair hybrid
to assign different potentials to different types.

If you are saying you want a C atom to interact in one geometric
region with one potential, then if it diffuses somewhere else,
to interact with a different potential, then LAMMPS won't do
that. You'd have to write a pair potential yourself that checked
for it. Seems like an odd model.

Steve

If you are saying you want a C atom to interact in one geometric
region with one potential, then if it diffuses somewhere else,
to interact with a different potential, then LAMMPS won't do
that. You'd have to write a pair potential yourself that checked
for it. Seems like an odd model.

i'd say it is worse than odd. it is almost guaranteed to give you wrong
results and very unphysical effects at the boundaries (not to mention
no energy conservation). i've been involved in several projects that
do QM/MM simulations (which is a very similar problem due to having
a QM region embedded in MM atoms), and handling the boundaries in a
consistent and energy conserving way is the by far biggest problem.
so not only would a method be needed to compute the interactions in
two different ways, but also a proper way to (smoothly) switch from
one to the other without imposing too many artefacts.

cheers,
  axel.