atoms overlapping with LJ/cut

LAMMPS LAMMPS Mailing List Mirror
1

Hello there,

i’ve got a pretty strange problem with lammps.
Sometimes while modeling a single polymer chain
with lj/cut parameters sigma=1, epsilon=1
with timestep 0.2fs
equlibrium bond length 0.8

two atoms (at the distance 2 beads along the chain,these atoms are not bonded)
overlap completely (e.g. distance between them -> 0 )
and move together all the time then.

so the question is - how is that possible, and how LAMMPS works when potential energy is close to infinity?

thank you
alex

2

Hello there,

i've got a pretty strange problem with lammps.
Sometimes while modeling a single polymer chain
with lj/cut parameters sigma=1, epsilon=1
with timestep 0.2fs
equlibrium bond length 0.8

two atoms (at the distance 2 beads along the chain,these atoms are not
bonded)
overlap completely (e.g. distance between them -> 0 )
and move together all the time then.

so the question is - how is that possible, and how LAMMPS works when
potential energy is close to infinity?

this is possible, when your exclusion settings is not suitable for your model.

if you have only bonds, but no angles or dihedrals that would keep
atoms apart, which are separated by one or two atoms, you have to
change from the default settings to exclude only immediately bound
atoms. see the special_bonds command for details.

also, your sigma and epsilon indicate you may be using reduced units,
but then your time step would be very large.

axel.