force and potential about pair_style lj/cut + shift

Dear LAMMPS users,

I understand that for the commands pair_style lj/cut plus a pair_modify shift, the Lennard-Jones potential is truncated and shifted up to 0 at the truncated distance. My questions is that since the force is the differential of the potential, what happens to the force when the potential has been cut as the potential at the cutoff point is not differentiable? If the LAMMPS is smoothing the potential at the cutoff point, or doing something else?

Many thanks,
Sunnia

Dear LAMMPS users,

I understand that for the commands pair_style lj/cut plus a pair_modify
shift, the Lennard-Jones potential is truncated and shifted up to 0 at the
truncated distance. My questions is that since the force is the
differential of the potential, what happens to the force when the potential
has been cut as the potential at the cutoff point is not differentiable? If
the LAMMPS is smoothing the potential at the cutoff point, or doing
something else?

‚Äčno. nothing.

the shifting only removes the discontinuity in the energy at the cutoff
distance and thus only matters for minimization in the first place.
the force does indeed have a discontinuity, but since this coincides with
the cutoff, there is no formal issue.

now, there are several lj pair styles in LAMMPS that implement different
kinds of switching functions that change the potential so that also the
force goes smoothly to zero.

axel.

Thanks for your answering - it helps a lot!

Sunnia