how to change the Rcutoff for long range interaction in reax/c

Dear lammps user's:
i would use the ReaxFF potential in my simulation for this reason i
would use the reax or the reax/c package implemented in lammps
software.
i would change the Rcutoff for long range interaction (coulombic)
which has the stantard value of 10A° but i have read the manual and i
don't find any idea expect the use of K_SPACE package.
so if any one have an idea please inform me.
best regards

Aidan would know if there is a setting in the ReaxFF potential
file for the Coulomb cutoff. I imagine that changing it is
a bad idea, since the potential is likely parameterized for a specific
cutoff.

ReaxFF does not do long range Coulombics, e.g. PPPM. It does
Coulomb interactions with a cutoff. That is inherent to the force
field. It does do charge equilibration which is effectively a long-range
operation.

Steve

The Coulomb cutoff distance is stored in the variable named swb in both
codes. The value is normally read in from the ffield file. For pair_style
reax/c you also must set the cutoff in the fix qeq/reax command. In my
experience, the value of 10 A is always used. Steve is right about it
being a bad idea to change the cutoff to a non-standard value. LAMMPS will
work fine, but changing the cutoff may produce unphysical (or unchemical)
interactions between atoms. More is not always better.

Aidan

thank you for the reply but what happen if the simulation box lenght
is less to 2*Rcutoff which means (20 A°). so in these case there will
be also unphysical results since any atom will interact with the same
atom in the box many times due to image position and periodic boundary
conditions.
best regards

thank you for the reply but what happen if the simulation box lenght
is less to 2*Rcutoff which means (20 A°). so in these case there will
be also unphysical results since any atom will interact with the same
atom in the box many times due to image position and periodic boundary
conditions.

why should that be unphysical?

unless there are any special restrictions that are due to
the reax implementation (i.e. have a close look at the
documentation), this will work just fine.

including those periodic images is exactly the same as
having a larger box made from replicated unit cells.

axel.

I don't understand why you would decide to try to change the cutoff to
handle this issue, especially since I think it is not actually an issue
(see below), and you can easily verify that yourself by running tests on
different box sizes.

In general, LAMMPS handles the issue of multiple image atoms correctly. I
believe this is now also true for pair_style reax and reax/c. The only
exception might be bonded interactions between periodic images of the same
atoms (can only occur for very small boxes).

Aidan