[lammps-users] crosslinked polymer and thermodynamic integration

LAMMPS LAMMPS Mailing List Mirror
1

Hello,

I have two questions about LAMMPS.

  1. I’m wondering if LAMMPS can be used to simulate crosslinked polymer where the atoms in the main simulation cell make bonds to their periodic images? This is not normally supported by the architecture of MD softwares such as CHARMM, AMBER, etc.

  2. Is it possible to implement thermodynamic integration on LAMMPS where the nonbond interaction of a specific pair (AA, AB, BB) is scaled with a constant which maybe different for different pairs?

Thank you very much for your help,
Surasak

2

Answers Below.

  1. Yes. See bond/create command
    2 Yes. Do you want to scale the interaction during the simulation?
3

Hi,

Thank you very much for your help.

  1. Yes. The interaction should be scaled during the simulation. Is it possible?

Thank you,
Surasak

4

Comments below.

Steve

Hello,

I have two questions about LAMMPS.

1) I'm wondering if LAMMPS can be used to simulate crosslinked polymer where
the atoms in the main simulation cell make bonds to their periodic images?
This is not normally supported by the architecture of MD softwares such as
CHARMM, AMBER, etc.

If you mean can you simulate a polymer chain that winds in and out of the
periodic box, then yes.

If you mean you want to create bonds over time, e.g. in a dynamic cross-linking
model, then you could use the fix bond/create command. PBC are not an issue.

2) Is it possible to implement thermodynamic integration on LAMMPS where the
nonbond interaction of a specific pair (AA, AB, BB) is scaled with a
constant which maybe different for different pairs?

If you mean define different epsilon/sigma for different pairs, then this is
done thru atom type pairs in LAMMPS. See the pair_coeff command for
pair_style lj/cut.

if you mean make epsilon/sigma time-varying, like in a TI calculation, then
currently you could only do this by putting a run and pair_coeff command in
a loop, and incrementing the pair coeffs every so many steps.

We are about to release a TI package that will let you specify some potentials
with a lamda param, that can vary dynamically from 0 to 1, which is used
for some kinds of TI, but you'll have to see if it does what you want.

5

Thank you very much for your email.

To clarify on crosslinked polymers, some atoms in the crosslinked polymers also make bonds to atoms in their periodic images and as a result make an infinite network. This cannot be done in CHARMM for example because it requires atoms to make bonds with only other atoms in the main cell to be able to define a topology and other relevant parameters. In this case (CHARMM), one cannot really define a topology for a polymer because the structure is infinite and is not contained in just one main simulation cell.

Surasak

6

I don't know how CHARMM handles this, but in LAMMPS there is no
conceptual problem with creating bonds to periodic image atoms (that
are really on the other side of the box). Either initially (e.g. in your input
data file), or adding them over time via fix bond/create.

Steve

7

Thank you very much for your email. This is very helpful.

Surasak