Generate a LAMMPS data file with bonds, angles, and dihedrals crossing the unit cell borders

Dear all,

I would like to do an MD simulation of a H-ZSM-5 zeolite using the Hill-Sauer force field. H-ZSM-5 is a periodic structure and the Hill-Sauer force field includes energy contributions from bonds, angles, and dihedrals. Consequently, it is necessary to write a LAMMPS data file that includes bonds, angles and dihedrals that cross the unit cell borders. How can this be done?

I tried to use topotools to solve this problem, but I discovered that bonds, angles, and dihedrals that are crossing the unit cell borders were not included. Can this be done with topotools, or are there other tools available that can solve the problem?

Regards,
Torstein

the topotools discussion is best done on the vmd mailing list. only so much, topotools is dependent on VMD’s internal mechanism for detecting bonds, which does not include periodic boundaries. inferring angles and dihedrals should work, though.

there are different ways to address this. you may look for other tools that can detect and assign bonds with periodic boundaries and output a file containing that bond information and then continue processing with topotools. you could use VMD/Tcl scripting to detect the missing bonds and add them.
you can also try using the fix bond/create command in LAMMPS after reservice sufficient space for added bond/angle/dihedral types and storage of the bonds/angles/dihedrals and special neighbor lists. this will likely need multiple calls to “find” all bonds and only works well for rather simple topologies and force field type assignments. if that is not sufficient, you can also use LAMMPS (which does consider PBC) to build the topology itself but then using VMD scripting to assign the types afterwards.

there are other pre-processing and visualization tools, that may be able to help in the process or scripting tools/programming environments, but i am not familiar with them. https://lammps.sandia.gov/prepost.html

Axel.

Hi Torsten,

the topotools discussion is best done on the vmd mailing list. only so much, topotools is dependent on VMD’s internal mechanism for detecting bonds, which does not include periodic boundaries. inferring angles and dihedrals should work, though.

there are different ways to address this. you may look for other tools that can detect and assign bonds with periodic boundaries and output a file containing that bond information and then continue processing with topotools. you could use VMD/Tcl scripting to detect the missing bonds and add them.

I recently used this particular approach using Axel’s TopoTools in a VMD script (Tcl language), to build polymer crystals of different sizes and/or connectivity maps. Adding bonds across boundaries via topo addbond, and then using again TopoTools to add the missing angles and dihedral entries that depended on them, works quite well. This was a CHARMM-style topology, so it was better done in VMD, where you can select atoms using their type’s name.

In other cases, the LAMMPS approach that he describes below may be more suitable.