constraints problem reaxFF force field

Dear all,

I am using the fortran version of reaxFF, not reax/c.

I want to fix a dihedral angle in the system in minimization(including box relaxation), by the following command.

fix d2 all restrain dihedral 4 7 10 1 2000 2000 180

However, I find in the output, only the four atoms related to this restrain “moved”, the rest of the system stayed in the initial position.

Anyone knows what is the problem and how to fix??

Thanks!!!

Rui DONG

NCSU physics Deparment

Dear all,

I am using the fortran version of reaxFF, not reax/c.
I want to fix a dihedral angle in the system in minimization(including box
relaxation), by the following command.

fix d2 all restrain dihedral 4 7 10 1 2000 2000 180

However, I find in the output, only the four atoms related to this restrain
"moved", the rest of the system stayed in the initial position.

Anyone knows what is the problem and how to fix??

no and no, since there is no useful information here to make any kind
of suggestion.

axel.

Sorry if I didn’t make my self clear.

I want to do minimization with constraint using reaxFF forcefield, however only atoms within the specified dihedral constraint can move, the rest of atoms just stay in the initial position.

Here is the input file

Sorry if I didn't make my self clear.
I want to do minimization with constraint using reaxFF forcefield, however
only atoms within the specified dihedral constraint can move, the rest of
atoms just stay in the initial position.

how do you determine this? via visualization?? or comparing actual numbers?

have you considered that the forces due to your restraints may be very
large, while the rest of the forces are small in comparison and thus
the minimizer will only move those atoms significantly?

the problem you describe has to be an issue of your setup and not a
LAMMPS problem.

axel.

Thanks for you reply.

I did both visualization and comparing the actual number. The coordinates for the non-moving atoms change about 10E-6 Angstrom. As I listed in the input file. I used the “hftn” minimization style, so there is no velocity related to the procedure.

To me, it seems lammps doesn’t recognize any bonds except the restrains I specified.

Rui

Thanks for you reply.

I did both visualization and comparing the actual number. The coordinates
for the non-moving atoms change about 10E-6 Angstrom. As I listed in the

so they *do* change.

input file. I used the "hftn" minimization style, so there is no velocity
related to the procedure.

who said anything about velocity?

To me, it seems lammps doesn't recognize any bonds except the restrains I
specified.

this statement is complete and utter nonsense.

axel.

To me, it seems lammps doesn’t recognize any bonds except the restrains I
specified.

One thing to realize is that a ReaxFF model has no permanent
bonds that you specify. Bonds come and go as the atoms
rearrange. So the fix restrain dihedral you specified is the only
bond(s) that are present in your model. I am guessing if
you run w/out the fix restrain, the other atoms in your model
also won’t move then, which means you have some other
kind of problem with your setup. Adding a single fix restrain
dihedral should not damp the motion of your entire system.

Steve