Hi all,
I am trying to use LAMMPS with water model TIP5P and six-site. But currently LAMMPS does not support this two models. I have an idea of how to use these models for water. I am just wondering if it is applicable for LAMMPS codes.
Since six-site water model has three negative charge on dummy atoms, I am thinking of creating a normal molecule with six sites. Three dummy sites in this molecule have a mass of 1.0e-10 percent of a hydrogen, which can be regarded as massless. The other three sites are normal atoms, like O and H. So LAMMPS can treat this six-site molecule as normal and calculate LJ and Coulombic energy between molecules.
Is there any suggestion of this problem? Any inputs are more than welcome. Thanks very much!
Regards,
Joy Yan
Hi all,
I am trying to use LAMMPS with water model TIP5P and six-site. But currently
LAMMPS does not support this two models. I have an idea of how to use these
that is not entirely correct. you don't have any specialized styles
in LAMMPS that are dedicated to these potentials. that there are
special tip4p styles is merely rooted in the fact, that you can
treat tip4p as a 3-site model and thus save a lot of compute time.
with 5 or more site water models, that is - of course - not possible.
models for water. I am just wondering if it is applicable for LAMMPS codes.
why not?
Since six-site water model has three negative charge on dummy atoms, I am
thinking of creating a normal molecule with six sites. Three dummy sites in
this molecule have a mass of 1.0e-10 percent of a hydrogen, which can be
regarded as massless. The other three sites are normal atoms, like O and H.
So LAMMPS can treat this six-site molecule as normal and calculate LJ and
Coulombic energy between molecules.
Is there any suggestion of this problem? Any inputs are more than welcome.
i don't quite get what you are asking here. first you say you have
an idea, but then you only talk about masses. where is the idea?
and where is the problem? and what exactly is the kind of input
that you are looking for?
axel.
If you want to treat each water molecule with its 6 sites
and input them directly in a data file, then you can just
do it, and integrate each as a rigid body, via fix rigid.
Steve
Thanks for the reply!
That’s what I gonna do. I am just wondering if lammps could work with very small stoms.
I will treat the bond between Oxygen and massless charge as normal bonds so lammps could work with the six-site model.
Thanks for the reply!
That's what I gonna do. I am just wondering if lammps could work with very
small stoms.
I will treat the bond between Oxygen and massless charge as normal bonds so
lammps could work with the six-site model.
you don't need any bonds if you are using
a rigid integrator, and you cannot use shake
and an atomistic integrator without a proper mass.
all you need is a proper initial geometry (with all sites)
and assigning proper molecule ids to the molecules.
then you can use a rigid (fix rigid, fix rigid/nve or fix rigid/nvt)
integrator to have it integrate the equations of motion
and set the molecule flag. you probably also want to
use "neigh_modify exclude molecule" to not compute
the intramolecular forces without need.
btw: that works for 3-site and 4-site models, too, and
i suggest you start with that as a benchmark test (since
you can use fix shake on them).
axel.