too high LJ potential cause to weird fluctuation of polymer in NVT run.

Hello,

I have an in. file which runs in two parts. one without any interaction of atoms. to put them in a random position. and 2nd part, continue with existing positions (random initial condition) and all interactions switched on. ( the difference in time step is to notice the moment of switching on the potential).
with very high epsilon for lj/cut I see the end atoms interact but middle atoms are almost frozen until they are pulled and stretched by two ends!

Q1. I’m expecting very strong attraction of interacting atoms. why I don’t see that?
Q2. Why inner atoms get frozen after a while and then start moving when are pulled?
(probably both problems have same cause)

cheers,

Hamed.

data_long.dat (49.8 KB)

in_long.dat (833 Bytes)

Hello,

I have an in. file which runs in two parts. one without any interaction of
atoms. to put them in a random position. and 2nd part, continue with
existing positions (random initial condition) and all interactions switched
on. ( the difference in time step is to notice the moment of switching on
the potential).
with very high epsilon for lj/cut I see the end atoms interact but middle
atoms are almost frozen until they are pulled and stretched by two ends!

Q1. I'm expecting very strong attraction of interacting atoms. why I don't
see that?

it is because of your choice of cutoff parameter.
your non-bonded interactions are cut off after 2.5 sigma,
but you also have bonded interactions with bonds and angles
and the default exclusions. atoms that are are not excluded
from the non-bonded interactions are far beyond the cutoff,
so it doesn't matter how large their epsilon is.

Q2. Why inner atoms get frozen after a while and then start moving when are
pulled?
(probably both problems have same cause)

atoms follow the model you impose on them.
it is *your* model, so it is *your* job to understand it.
LAMMPS is just doing what you tell it to do
and apparently it is not what you expect, but
unless you have prove that LAMMPS is not
doing what it is programmed to do, then it
is of no concern for anybody but you.

axel.