How to prevent the shape of ellipsoids from changing?

Dear LAMMPS Users,

I am performing coarse-grained MD simulations on polymer chains, with Gay-Berne potential for non-bonded interactions in LAMMPS (30 Jul 2016).

All my CG beads are ellipsoids, using atom_style hybrid ellipsoid full. I am running npt/asphere. I am computing the shape and quaternion information of the ellipsoids via compute property/atom. I find that the shape of the ellipsoid is changing (shrinking in my case) and I don’t want that to happen. I want them to move translationally and their orientation can change. How do I prevent the change of shape of the ellipsoids in LAMMPS? It would be great if someone can help me with suggestions.

Thanks in advance for your help,

Sandipan

Dear LAMMPS Users,

I am performing coarse-grained MD simulations on polymer chains, with Gay-Berne potential for non-bonded interactions in LAMMPS (30 Jul 2016).

this is a pretty old version, you may want to try out the most recent one, in case what you see is a bug that has been fixed. it is also possible that you are misinterpreting your output. there is not sufficient information here to determine if either point is true or false.

All my CG beads are ellipsoids, using atom_style hybrid ellipsoid full. I am running npt/asphere. I am computing the shape and quaternion information of the ellipsoids via compute property/atom. I find that the shape of the ellipsoid is changing (shrinking in my case) and I don’t want that to happen. I want them to move translationally and their orientation can change. How do I prevent the change of shape of the ellipsoids in LAMMPS? It would be great if someone can help me with suggestions.

do you see the same kind of “shrink” also with the inputs in the examples/ellipse folders? I don’t see anything unexpected with the current development version.

Axel.

I will soon upgrade to the newer version. However, what I am reporting is not unusual or unexpected. The shape of the ellipsoids change as I haven’t asked lammps to keep them constant. My question is “how to ask lammps to keep the shape of ellipsoids constant?” One possible option is fix rigid, where I’ll have to specify each particle (ellipsoid) as a rigid particle. However, won’t it prevent the particles (ellipsoids) from moving at all? Also, will npt/asphere work with rigid particles (ellipsoids)? I am attaching my input and restart files, in case that helps.

PEEK_100mono_10chain_GBCG.in (1.03 KB)

PEEK_GBCG.restart (398 KB)

Binary restart files are not portable across LAMMPS versions and platforms and thus useless.

There is nothing in LAMMPS that changes the shape of ellipsoids outside the set command. It can be easily confirmed by checking out the example files I referred to or checking out the source code.

You need to provide more convincing and more easily to reproduce evidence that what say is happening does happen and not that you are misinterpreting the simulation data or have set up a model that is different from what you think you have.

Axel

Dear LAMMPS Users,

I am performing coarse-grained MD simulations on polymer chains, with Gay-Berne potential for non-bonded interactions in LAMMPS (30 Jul 2016).

Perhaps, the easiest thing to do is to take an example of a simulation of ellipsoids which is working, and then modify it to suit your needs. If you just want a simple example how to simulate ellipsoids, take a look at the “ASPHERE/poly” example included in the “examples” directory with the LAMMPS distribution. If you’re willing to consider using moltemplate, there’s an example here:
http://moltemplate.org/visual_examples.html#ellipsoids

More general and more useful perhaps, are the instructions how to visualize the results of that particular simulation using OVITO:

https://github.com/jewettaij/moltemplate/blob/master/examples/coarse_grained/MOLC/README_visualization_OVITO_settings.png

Some of the settings in that picture depend on the format of the dump file you created in the input script for your simulation. For that example, we used this command to generate the dump file which OVITO reads:

dump 1 all custom 100 traj.lammpstrj id type x y z &
c_q[1] c_q[2] c_q[3] c_q[4] &
c_shape[1] c_shape[2] c_shape[3] &

angmomx angmomy angmom

Perhaps those settings will help you with your current simulation.

Although this does not address the question you posted, since you are simulating bonded ellipsoids, you might be interested in the new LAMMPS features which are not yet part of the LAMMPS distribution. They are described here:
https://pubs.rsc.org/en/content/articlelanding/2019/cp/c9cp04120f#!divAbstract

Hopefully their code will be included with LAMMPS soon. Meanwhile if you are interested in using their code, it is included in the supplementary files for that paper. The authors of that paper may be willing to send you the latest version of this code.

Andrew