aligned dipoles

Good morning to all,
I have a quick question regarding a stockmayer system with aligned dipoles, any help will be appreciated.

I need to simulate a system of dipolar lennard jones particles with aligned dipoles, i.e. the dipoles don't rotate, are always aligned in the z direction. Is there any other way to do this, to disable the rortation of dipoles?

the second question: the dipolar torques will be calculated every timestep which is not needed as they are not rotating, is there a way to disable the calculation of the torqes in both r and k space for the dipolar forces.

Thanks a lot for your time
Manuel Valera

If you don’t integrate the dipoles, they won’t change
orientation. E.g. just use fix nve on the dipolar particles
instead of fix nve/sphere dipole.

2nd Q: if you don’t want to compute the torques at all,
it’s not clear to me why you are defining dipoles in
the first place.

Steve

Thanks a lot for your reply.

If you don’t integrate the dipoles, they won’t change
orientation. E.g. just use fix nve on the dipolar particles
instead of fix nve/sphere dipole.

Thanks, this should work.

2nd Q: if you don’t want to compute the torques at all,
it’s not clear to me why you are defining dipoles in
the first place.

In the system we are simulating, colloids are immerse in an external electric field along the z axis, the induced dipoles on the particles will never rotate, therefore there is no need to calculate torques. The effective potential would be:

u(rij) ~ 1/rij^3*(1-cos(theta)^2)

where theta is the angle between the z axis and the vector from particle i to j (rij). I could implement this potential in lammps and use it as a pair potential with a cut and I will not have to use dipoles if I go this route. However I was hoping to be able to use a k-solver to take care of long range nature of 1/r^3, but I believe ( and correct me if I am wrong) there is not k-solver for dipoles implemented in lammps yet?

Thanks again for your time

Manuel

yes, there is no long-range dipole solver yet -
we have a prototype of one that uses Ewald,
but it has not yet been fully debugged.

Steve