ES+EAM potential

OK, calm down…
I rewrite it clearly.
Firstly, I read a article:
Electrostatic potentials for metal-oxide surfaces and interfaces, Phys. R. B (1994) by F. H. Streitz and J. W. Mintmire
It introduces a potential, ES+EAM, namely, electrostatic plus embedded-atom method.
Actually, it contains two parts, electrostatic potential and embedded-atom potential.
But it can not be achieved by pair_style hybrid/overlay, since I can not find any potential with the same form referred in the article.
Then, I try to program it.
Embedded-atom method is relatively easy to be done.
Electrostatic potential stuck me.
In the formulas of electrostatic potential, an item --1/r_ij in V_ij can not be cancelled and evaluated by standard Ewald summation.
Here is the point. How to get a badly convergent 1/r_ij by Ewald summation.
I wonder whether any code in LAMMPS can carry out this–evaluated 1/r_ij in all range?

PS: I tried some ridiculous ways… and ridiculous answers…

As Axel told you, electrostatics is pair_style coul or coul/long
or lj/cut/coul/long. You can use the "long" ones with a kspace_style
to do long-range electrostatics. And you can use pair_style hybrid
with the eam style to combine them. However, I thought Streitz
and Mintmire was a charge-equilibrated potential, so you cannot
yet do that. Other pair styles in LAMMPS do charge-equilibration
(Reax, COMB), but we don't yet have a stand-alone option for it.

Steve

Meihai3166,

Have a look at pair_style comb, which uses ES+ with Tersoff with QEq.
Function sm_table() of pair_comb.cpp generates a look-up table for the
Streitz-Mintmire ES+ potential.

Although ES+ in COMB is solved with Wolf summation instead of Ewald,
given some similarity between Wolf and Ewald, reading this function
might help you a little.

Ray