Inquiry on LAMMPS special_bond and pair_style

LAMMPS LAMMPS Mailing List Mirror
1

Hello All,

This is Masato Koizumi, and I am currently learning how to use LAMMPS.

I was curious if I could ask a question on two commands in LAMMPS: special_bonds and pair_style hybrid/overlay.

Observing the LAMMPS manual, I was curious if the command, “special_bonds lj/coul 0 1 1” means that there are two interactions acting between each pair of atoms: Lennard-Jones and Coloumb. Lennard-Jones is the pairwise interaction and Coulomb is the bond interaction. I was curious if the 0 1 1 means that each pairs of atoms consist either one, two or three bonds. 0 means that for one-bond pair, the LJ pair potential is turned off. Likewise, the 1’s mean that for two-bond and three-bond pairs, the LJ pair potential is turned on.

Lastly, for pair_style hybrid/overlay, I was curious if this command simply means that the potential energy between each pairs of atoms is the summation of the pair potentials. That is, pair_style hybrid/overlay lj/cut 2.5 coul/long 2.0 is defined as the summation of Lennard-Jones and Coulomb potential acting between each pairs of atoms.

Thank you.

Masato

2

Hello All,

This is Masato Koizumi, and I am currently learning how to use LAMMPS.

I was curious if I could ask a question on two commands in LAMMPS:
special_bonds and pair_style hybrid/overlay.

Observing the LAMMPS manual, I was curious if the command, "special_bonds
lj/coul 0 1 1" means that there are two interactions acting between each
pair of atoms: Lennard-Jones and Coloumb. Lennard-Jones is the pairwise
interaction and Coulomb is the bond interaction. I was curious if the 0 1 1
means that each pairs of atoms consist either one, two or three bonds. 0
means that for one-bond pair, the LJ pair potential is turned off.
Likewise, the 1's mean that for two-bond and three-bond pairs, the LJ pair
potential is turned on.

​this sounds a bit confused, but may be right. you may want to read up on
what is often called "exclusions" in your preferred MD textbook.
here is my way to describe it:

if you have a system, where you have non-bonded interactions, e.g. via pair
style lj/cut/coul/cut​, *and* bonded interactions e.g. via bond style
harmonic and angle style harmonic and dihedral style harmonic, then the
special_bonds scaling factors are applied to all 1-2, 1-3, and 1-4 pairs
connected by bonds. for the default setting (0/0/0), that means when two
atoms are connected with a bond (i.e. a 1-2 pair), the non-bonded
interaction is turned off, if there are two bonds to the same central atom
(a-b-c), then the 1-3 interaction (a-c) is scaled by the 1-3 factor. if
there is also a bond between a-c, then the 1-2 factor is applied. same
rules apply for 1-4 pairs and the 1-4 special bonds factor.

Lastly, for pair_style hybrid/overlay, I was curious if this command
simply means that the potential energy between each pairs of atoms is the
summation of the pair potentials. That is, pair_style hybrid/overlay
lj/cut 2.5 coul/long 2.0 is defined as the summation of Lennard-Jones and
Coulomb potential acting between each pairs of atoms.

​yes, this would be the same as lj/cut/coul/long 2.5 2.0. please make a
test and see for yourself.

axel.​

3

Dear Dr. Axel Kohlmeyer,

Hello Dr. Kohlmeyer. Thank you for your email.

I created a LAMMPS input script that applies tensile and compressive stress on the simulation box that I created using “fix deform” command.

I was curious if there is a way to perform cyclic loading without repeating the “fix deform” command. That is, I was curious if there is a way to perform cyclic loading (stretching and compressing) without changing the number of timesteps and the strain rate. Thank you.

Sincerely,

Masato Koizumi

4

Dear Dr. Axel Kohlmeyer,

Hello Dr. Kohlmeyer. Thank you for your email.

I created a LAMMPS input script that applies tensile and compressive stress
on the simulation box that I created using "fix deform" command.

I was curious if there is a way to perform cyclic loading without repeating
the "fix deform" command. That is, I was curious if there is a way to
perform cyclic loading (stretching and compressing) without changing the
number of timesteps and the strain rate. Thank you.

please re-read the fix deform command documentation and pay special
attention to any mentioning of variables.

axel.