Error Using quaternion in simulation

Dear users,
I want to simulate a system using atom_style hybrid bond sphere, so that I can account for rotation of the particles in sphere style (as it is a particle with finite radius). In order to describe its orientation I need to use quaternion . But when I use the set command ‘quat’ or ‘quat/random’ for the particles in atom_style sphere I am getting following error;

Cannot set this attribute for this atom style (src/set.cpp:279)

I have the query that I can not use quaternion in atom_style hybrid bond sphere?

How can I solve this problem?

Thanks

Rahul S

Dear users,
   I want to simulate a system using atom_style hybrid bond sphere, so that I can account for rotation of the particles in sphere style (as it is a particle with finite radius). In order to describe its orientation I need to use quaternion . But when I use the set command 'quat' or 'quat/random' for the particles in atom_style sphere I am getting following error;

spheres have no orientation, only angular velocity.

                    Cannot set this attribute for this atom style (src/set.cpp:279)

I have the query that I can not use quaternion in atom_style hybrid bond sphere?

you cannot. very obviously.

How can I solve this problem?

how about spending more time reading the manual instead of writing e-mails?
https://lammps.sandia.gov/doc/atom_style.html

axel.

Quaternions in LAMMPS are used for aspherical particles
which have a unique orientation (the quat) and an
angular momentum vector as well. As Axel said,
finite-size spherical particles only have an angular
velocity vector (omega). However for some aspherical
models (like ellipsoids) there is no problem with using
them to model spheres (e.g. a degenerate ellipsoid).
Then you will have a quat and arg momentum vector
for a sphere.

It is up to you to use (or create your own new) pair style
that computes interactions appropropriate for your
model, i.e. something that induces torque.

Steve

Quaternions in LAMMPS are used for aspherical particles
which have a unique orientation (the quat) and an
angular momentum vector as well. As Axel said,
finite-size spherical particles only have an angular
velocity vector (omega). However for some aspherical
models (like ellipsoids) there is no problem with using
them to model spheres (e.g. a degenerate ellipsoid).
Then you will have a quat and arg momentum vector
for a sphere.

but - like i explained in a different e-mail thread on a similar
topic - to update those orientations, also a different time
integration fix must be used, e.g. fix nve/asphere instead of fix
nve/sphere

axel.

Thanks for your response… I will try with this…

Quaternions in LAMMPS are used for aspherical particles
which have a unique orientation (the quat) and an
angular momentum vector as well. As Axel said,
finite-size spherical particles only have an angular
velocity vector (omega). However for some aspherical
models (like ellipsoids) there is no problem with using
them to model spheres (e.g. a degenerate ellipsoid).
Then you will have a quat and arg momentum vector
for a sphere.

but - like i explained in a different e-mail thread on a similar
topic - to update those orientations, also a different time
integration fix must be used, e.g. fix nve/asphere instead of fix
nve/sphere

axel.

I have gone through the above discussion about the query. I have a fixed frame (which is not moving) and a body frame which is attached with the sphere. So if this sphere rotates then orientation of the body axes will change with respect to the fixed frame. In this case can I use quaternion formulation in LAMMPS?