[lammps-users] displace_atoms Bug?


I have noticed some peculiarities with the "displace_atoms" command when
trying to move a group of atoms. I have attached an input script that
demonstrates the issue I am seeing.

In the script I create a block of atoms (block_1), then displace block_1
in the z-direction by two lattice units.

I then try to create a second block of atoms in the position where
block_1 used to be. However, lammps fails to create any atoms.

Interestingly enough, if I displace block_1 by only 1 lattice unit, I
get the creation of 5 overlapped atoms (14 should be created with FCC).
Also, if I create a single atom at 0, 0, 0 before the displacement of
the block (but keeping this single atom fixed), lammps will successfully
create the second block of atoms.

It seems like some sort of reference frame is being moved when the atoms
are displaced, or am I missing something?

I am running version 15 Jan 2010.


displace.test (735 Bytes)

You're using a non-periodic box and shrink-wrap boundary conditions,
so I imagine what is happening when you move all the atoms, is
that the box is re-wrapped around the atoms in their new position.
Create_atoms won't let you create atoms outside the current box,
so it does nothing. What you want is to fix the box size and move
the atoms within it (without changing the box), to leave space
for your subsequent insertions.