[lammps-users] mixing long coulombic pair potentials?

Hello all,
I’m having some trouble with something and thought maybe the lammps-users can help: lammps doesn’t allow one to create a hybrid pair potential using two potentials with coul/long. Does anyone know the reason why this is? I’m trying to reproduce the results of a potential model done in cerius2 used on a surface-water interface where, as far as I can tell, they’ve used an ewald sum for the long range forces and short-range interactions are treated with a LJ with a cut-off for the water atoms but buckingham with a cut-off for surface atoms, and between surface atoms and water atoms.
My theoretical (mis?)understanding is that since the short range forces are cut off at a fixed distance, their functional form shouldn’t matter for how the ewald sum is calculated. This is assuming that the atoms have a charge calculated in a similar way, and that the cut-off is the same. Am I wrong? Thank you very much, I greatly appreciate it.


You're right. As long as you use a consistent Coulomb cutoff for the
2 (or more) short-range potentials, you should be able to do
a single Ewald or PPPM with all the charges in the system.
The latest version of LAMMP (hopefully released in a few weeks)
will allow that, but I haven't tested it. You might want to just
comment out the error check you're seeing and try it.