[lammps-users] Inquiry – LAMMPS Mailing List Mirror

Hi All,

Sorry if this question might be trivial: I am wondering to know whether it is possible to define a bond between the atoms (let’s say atoms on a graphene layer) near the simulation box surface and those same imaginary atoms from the adjacent imaginary simulation boxes, in periodic boundary conditions. In this case, an infinite simulated model would be created.

Is periodic boundary conditions (p p p) already taking care of it?

No.

Explicit bonds are between different atom IDs, so you would need to have a bond between atom IDs at different “sides” of the box.
LAMMPS will print a warning about inconsistent image flags in this case, but this cannot be avoided. However when LAMMPS processes the bond forces it looks at the closest pair and then those bonds have normal length.
The situation is different with manybody force fields (Tersoff, Stillinger-Weber, AIREBO, ReaxFF, etc.). In these cases bonds are implicit and they will look at the closest atoms through periodic boundaries automatically. In fact, you must not define explicit bonds in these cases.

Axel.

Read more here: Source link