Gay-Berne potential without attraction

Hi. I’m quite new to this Lammps community.

I’m looking for a pair-style potential to simulate hard core volume interaction between elliptical particles. and I think GB with correct cutoff will be the one. like in Phys. Rev. E 101, 022602 (2020) - From scalar to polar active matter: Connecting simulations with mean-field theory

and also, Hoomd-blue gives repulsive-only option

but to me it seems like current version of Lammps only provides isotropic cutoff length.

can you recommend other pair-style potential of teach me how to modify the cutoff length to be anisotropic?

Thank you

You can also try the resquared potential for ellipsoidal particles.

For the cutoff, you have to chose a distance that is at least bigger than the biggest axis of the ellipsoids in your system, otherwise there will be significant numerical instabilities when two particles approach from the direction of the longest axis. Given that the particles could be randomly oriented, I don’t see how an anisotropic cutoff radius would benefit your simulation (besides messing up with the creation of the neighbour list).

Thank you for your answer.

but as I know, RE-squared potential also have attracting tail, which I don’t want

also thank you for your comment on ‘cutoff radius’. I think it can be confusion from what I said.
to me, cutoff radius is the threshold for computing potential, and there should be some radius larger than this to make neighbor list

If you want a purely repulsive potential, then you could use the Buckingham pair_style buck with the attractive coefficient C set to zero.

I don’t think Buckingham potential can simulate elliptic object which is anisotropic

The general rule with LAMMPS is the following: all available functionality is documented.
This means, if you are looking for functionality that is not documented, then it is not available directly. You will have to either emulate it with available functionality or implement it yourself through modifying the source code.

To the best of my knowledge, there is neither an anisotropic cutoff nor a repulsive-only GB potential available.

You are right, my bad! Then the easiest option is to fit a reference potential with either a GB or RSQ potential: probably you will end up with a small well and the resulting potential will be dominated by the repulsive part. It’s easy and, even if it is not exactly what you want, it may be a good enough approximation (unless you are interested in the behaviour at very low temperature).

Thank you for your comment.
I’ve read the source code : pair_gayberne.cpp
and I think I can add this functionality by removing all the force/torque and setting the potential to zero for varrho>2^(-1/6).

I think I can implement it myself for my own use (like only removing force/torque as I don’t care about potential for my case) but it would be beneficial to add this functionality officially.
How can I request such option to the community?

You can submit a feature request issue on GitHub.

But please keep the following in mind. People using aspherical particles are a small subset of the LAMMPS community, and people wanting to run such interactions with repulsive-only interactions are a subset of those. In my 15+ years of following LAMMPS development, I do not recall anybody asking about this before. So what you are asking for is not something that many people are waiting for and will be of use to a significant fraction of the LAMMPS user community. And since there is no shortage of requests and most of the core LAMMPS developers have their own preferred projects and features on top of that, your request will have very little chance to be picked up, unless there is somebody else that feels your need and has the time, experience, and desire to implement and contribute it.

The chance to have the feature, you are looking at, integrated into LAMMPS would be much larger, if you would implement it and submit a pull request. If the change you would implement is done cleanly and well enough, and is properly documented, then there is a good chance that it will be integrated. Most of the current LAMMPS code came into existence this way. In fact, the aspherical particle pair styles are contributed code themselves.

For more info please see 3.2. Submitting new features for inclusion in LAMMPS — LAMMPS documentation and 3.4. LAMMPS programming style — LAMMPS documentation