[lammps-users] Implement A and B style lennard-jones potential

Dear Lammps users,

Is it possible to have a A and B style LJ potential added in the LAMMPS package? I am using a forcefield with direct A and B (negative). My programming ability is not enough to implement it in the source code. If you experts can implement it in the Lammps, this pair style would be a really helpful function.

Kind regards,

The two different representations can usually be converted from one to the other.
For cases that cannot, you can still use pair style table or consider lennard/mdf ( https://docs.lammps.org/pair_mdf.html )


automatic mixing requires a representation in (or conversion to) epsilon/sigma format as the mixing rules apply to epsilon and sigma, not A and B.
thus any other implementation of an A-B style Lennard-Jones potential cannot not have automatic mixing either.

if you have only a tiny distance between inner and outer cutoff for the tapering function there should be no significant difference. besides, if the cutoff is so small that the tapering function makes a significant difference, then you introduce a significant error into your simulation anyway.

Thanks Axel. I am aware of the lennard/mdf and table method. However, they are not as powerful as the epsilon and sigma method. They need to mix manually and lennard/mdf is with a tapering function. I’ll try them first to see the results.


Thank you, Axel. I’ll try it.


Do you mean pair_style lennard/mdf (https://lammps.sandia.gov/doc/pair_mdf.html) ?