Gaussian chain

Dear Sir,

I use the tools of lammps named chain.f to generate the long chain system as coarse-grain polymer.
But I’m not sure the structure is gaussian chain or not.

no - if you read the header of the file you will see that it
gives you a very simple, unequilibrated chain with potentially
many overlaps.


Dear Sir,

Then how to use lammps to generate the Gaussian chain.
Any idea or suggestion are really appreicated.

Dear Sir,

   Then how to use lammps to generate the Gaussian chain.
   Any idea or suggestion are really appreicated.

LAMMPS internally only has the ability to generate
atomic lattices. for everything, you have to use or
write an external tool that generates the suitable
start configuration in the format required by the
read_data command. in other words, you have to
write a suitable script or program yourself.


Any simple program which generate random polymer coordinates is very
likely to occasionally place two atoms close to each other (or place 3
atoms in a collinear arrangement). This is not a problem.

Once you have a chain polymer prepared for LAMMPS, you can minimize
it, and run a simulation at NVT conditions, In theory if you simulate
the polymer for "long enough", the polymer will relax to a
conformation which is consistent with the conditions you want to
study. If you have a long polymer, or a compact polymer, you may need
to use enhanced sampling techniques like the "temper" command (or
allowing atoms to pass through each other), to accelerate the
equilibration process. I don't know of a simpler way to generate
random (but physically reasonable) polymer conformations.

   Try running the simulation at NVT conditions (without temper) and
see what happens. (If you want a random-coil, it may help if you turn
off all pair attractive interactions between beads in the chain.)

Good luck


P.S. The "temper" command is explained (briefly) here: