dihedral computation

Hello,

I am using reax potentials. I am trying to calculate all the dihedral angles and dihedral types for my system and print them using the following commands:

compute myDihedral all dihedral/local phi
compute myDihedralType all property/local dtype

Vivek,
Yes, that is because you are using reax where no explicit dihedral
interactions are defined.
The property/local compute doesn't work for that case either.
Carlos

Hello,

I am using reax potentials. I am trying to calculate all the dihedral angles
and dihedral types for my system and print them using the following
commands:

compute myDihedral all dihedral/local phi
compute myDihedralType all property/local dtype
#
fix prntDihedral all print 1 c_myDihedral file dihedral.dat
screen no
fix prntDiType all print 1 c_myDihedralType file
dihedralType.dat screen no

For the first compute statement I get the following error:

ERROR: Compute dihedral/local used when dihedrals are not allowed
(compute_dihedral_local.cpp:41)

While for the second compute statement I get the following error:

ERROR: Compute property/local for property that isn't allocated
(compute_property_local.cpp:174)

Is it because of reax field where don't have to specify dihedrals ? or the
portion of my input script that I have written above is incorrect ? I would
appreciate if someone can tell me how to calculate each dihedral angle and
type using reax potentials.

there *are* no dihedral types with reax (and neither are bond or angle types).
that is the point of this force field.

if you don't define a dihedral, you cannot display it with the
standard computes in lammps. those require that you have explicit
dihedrals defined, like in a traditional classical force field.

axel.

Thanks ! So, I guess the best way for me would be to write coordinates (preferably in xyz format perhaps) every few timesteps and then calculate dihedral angles for each of those steps. Wonder if someone is aware of a tool which does that ?

Thanks again,

Vivek

Thanks ! So, I guess the best way for me would be to write coordinates
(preferably in xyz format perhaps) every few timesteps and then calculate
dihedral angles for each of those steps. Wonder if someone is aware of a
tool which does that ?

VMD. reads native LAMMPS dump files, too.

another potential option would be to define the dihedrals that you
would be interested in in LAMMPS, but set the force constant to zero
and disable exclusions with special_bonds lj/coul 1.0 1.0 1.0

it all depends on what you need the information for and how easy it is
to narrow down the list to make the analysis simple and efficient.

axel.

You can output the "bonds" that the ReaxFF infers in a current
snapshot via the fix reax/bonds command. From the bonds, you
can post-process to infer angles, dihedrals, etc.

Steve