Twist and Torque

Hello,

I would like to simulate a chain of beads that has not only a potential depending on the angle formed by each triplet, but also dependent on the twist that each bead represents
with respect to its direct (bonded) neighbour.

It seems to me that lubricate goes in that direction, but it specifies this interaction for each pair of particles - and it does not take into the account the difference of rotation between two particles, but only the difference in angular velocity. In addition, I want this kind of interaction only for particles that share a bond.

Is that feasible?

Thanks in advance!

Sure, if you write a pair style to implement the model.

Steve

Or maybe it could be done with a “bond” or “angle” style?

For example, I use this angle style to control the orientation of spherical particles in relation to bonded neighbors (which sounds somewhat similar to what you want to do):

http://lammps.sandia.gov/doc/angle_dipole.html

Best,
Mario

Thanks for posting the link to angle_style dipole. I've been
wondering how to do this with ellipsoids and dipoles for a while.
This looks very useful.

Question: Is there a similar command for ellipsoids?

(Note: Ellipsoids can be more complicated because there is is often an
additional angle to worry about. ..."phi" in the attached diagram.)
Unfortunately I don't use dipoles or ellipsoids yet in my own work.
(I won't be submitting a paper using ellipsoids or dipoles anytime
soon.) But it seems like it would also be very useful to others, for
coarse-graining various polymers, etc..

Thanks
Andrew

I ask because I wrote a molecule builder (moltemplate.org) and I've
been wondering if I modify it to support for dipoles and ellipsoids.
(It does not handle these atom types yet.) I did not know about the
"dipole" angle_style. I will add support for dipoles. (ETA a few
weeks, or sooner if there is demand for it.) I'm wondering whether I
should add support for ellipsoids...

Cheers!
Andrew

bonded_ellipsoids.svg

bonded_ellipsoids.png

My apologies. I realize my previous post is (probably) not related to
Björn's question.

   Getting back to Björn's question... If I understand his post
correctly, Björn wants a 4-body interaction between bonded atoms,
whose energy is not a simple sum of 3-body "angle" energies (1-2-3 &
2-3-4) and 4-body "dihedral" energy. Is this correct?

interactions1-4.png

Thanks for posting the link to angle_style dipole. I’ve been
wondering how to do this with ellipsoids and dipoles for a while.
This looks very useful.

Question: Is there a similar command for ellipsoids?

I don’t think so, but it may be easy to derive one from angle_style dipole. Specifically, the ellipsoid axis could take the place of the point dipole vector. This would allow control of “theta” in your sketch.
Something else would be needed to restrain “phi”…

Best,
Mario

Thanks. I know somebody who hacked this feature (bonded ellipsoids)
into a customized version of LAMMPS, but I don't think he did it in a
way which would be usable by others, and I don't think he's interested
in posting what he did. If I get a chance to communicate with him
again, I'll ask him.
Cheers and thanks for the reply.

Andrew