Dear all,
I am simulating a system with non periodic boundary in z direction. long range interaction section has been defined as follow:
boundary p p f
kspace_style pppm 0.00001
kspace_modify slab nozforce
but I face the error of : “Cannot use nonperiodic boundaries with PPPM”. Any help will be appreciated.
Regards,
Mina
I re-defined kspace_modify again and now I receive this error :
Cannot (yet) use PPPM with triclinic box and slab correction
It seems this feature is not supported yet. So, how can I use non periodic boundary in z direction? Thank you so much.
Regards,
Mina
I re-defined kspace_modify again and now I receive this error :
Cannot (yet) use PPPM with triclinic box and slab correction
It seems this feature is not supported yet. So, how can I use non periodic boundary in z direction? Thank you so much.
adding support for triclinic boxes when using kspace_modify slab is an
open issue: https://github.com/lammps/lammps/issues/410
...and has been open for two years now. as the discussion on github
indicates, this is a non-trivial change.
thus you either need to find/motivate a volunteer, that is willing to
spend the time to implement and test this. or use an appropriate
orthogonal supercell or don't use long-range electrostatics or find a
different MD code, that does support such a system.
axel.
Dear Axel,
Thank you for your reply. Can I enlarge the boundaries in z direction, as the atoms in both sides don not meet each other and use p p p for boundary command and apply vaccum B.C. in this way?
Or use msm kspace style?
Thank you for your advice.
Regards,
Mina
Dear Axel,
what you are asking here (and others before) are questions that you
should discuss with your adviser or supervisor and not for a mailing
list. this is the last time i will respond to something like this. any
future questions of such kind, i will ignore.
Thank you for your reply. Can I enlarge the boundaries in z direction, as the atoms in both sides don not meet each other and use p p p for boundary command and apply vaccum B.C. in this way?
if that is a valid approach is something that you can figure out
yourself. set up a simple test slab system (just a few 10s of point
charges) in an orthogonal box and do single point calculations (with
run 0) and then compare the results (energy and forces) of a fully
periodic system with increasing length in z to a system using the slab
correction in z. do the numbers converge at some point? and if yes,
how much vacuum do you need to add to reduce the interactions between
the periodically replicated slabs until they are small enough, that
you can safely ignore it?
Or use msm kspace style?
why ask? why not make a test? again, you need to figure out which
settings you need to get sufficient accuracy. msm has the advantage to
be applied to a real 2d periodic system, where pppm 2d-periodicity is
done by doing a 3d-calculation with sufficient vaccuum added, followed
by solving the poisson equation to compensate the inter-slab
interactions.
axel.