[lammps-users] NPT in a vacuum for a crystal surface with xy periodicity

Berk,

I am not sure why you would want to perform an NPT simulation with a
vacuum termination. If you want to use NVT, then I recommend using
"kspace_modify slab" option. This creates a vacuum region but still
allows you to use 3D periodic boundary conditions for the long-range
interactions. Here is the documentation:
http://lammps.sandia.gov/doc/kspace_modify.html

Regards,
Jeff Greathouse

Thanks to Jeffery, Zhimin, and Steve for their help. Jeff's suggestion
should circumvent the problem. In fact, we want to perform an NPT
simulation to thermally expand the crystal.