From: Axel Kohlmeyer (
Date: Sat Feb 06 2016 - 07:41:31 CST

On Sat, Feb 6, 2016 at 2:23 PM, M K <> wrote:
> Hi,
> I intend to output the final frame of a periodic system to a non-periodic
> form in VMD and then export it as a lammps data file (topo tools) and use it
> in a non-periodic simulation study in lammps.
> I use "pbc wrap" on the last frame, followed by this command:

this is likely the mistake you are making. if you write out the last
frame in *unwrapped* coordinates (or use a data file generated by
then you should have all molecules already joined. if not, you may
have a problem with your periodic calculation already.
to prepare for non-periodic, you should then wrap by *fragment*, and
not use the default of wrapping by atoms.
at this point, no further processing should be needed outside of
enlarging the box in the dimensions that are becoming non-periodic,
since otherwise the atoms outside the box will be ignored (which is
different from reading coordinates for periodic dimensions).


> pbc join fragment -bondlist
> However, there is still one bond overstretched along the cell, which makes
> problem later in the lammps data file output. Also the angles corresponding
> to those two atoms produce errors of lammps at step 0 (lammps user may find
> this useful: Bond atoms 718 733 missing on proc 1 at step 0) (This is
> obvious since there is not periodic boundaries anymore).
> Can you pleas help me with this issue, making a good output useful for the
> non-periodic simulation? Any suggestion/solution is appreciated.
> Bests
> Mah

Dr. Axel Kohlmeyer
College of Science & Technology, Temple University, Philadelphia PA, USA
International Centre for Theoretical Physics, Trieste. Italy.