Dear LAMMPS Users
I have a very naive question regarding the finite-size particles atom_style in LAMMPS. Please excuse for posting such a simple question.
After setting up the correct atom_style and associated property (e.g. mass, diameter/ mass, diameter and quaternion for sphere/ellipsoid particles), would they be treated as “hard sphere/finite-size particles with a impenetrable finite volume” in simulations by default, or that depends on the potential that is assigned to them?
For example, I am using atom_style ellipsoid with assigned diameter, while I am assigning them gayberne potentials, the pairwise distance seems not limited by the assigned diameters but more by the assigned gayberne parameter. I am not sure if I have done the test wrongly.
Thank you very much for your patience!
Best Wishes
Lunna
Dear LAMMPS Users
I have a very naive question regarding the finite-size particles atom_style in LAMMPS. Please excuse for posting such a simple question.
After setting up the correct atom_style and associated property (e.g. mass, diameter/ mass, diameter and quaternion for sphere/ellipsoid particles), would they be treated as “hard sphere/finite-size particles with a impenetrable finite volume” in simulations by default, or that depends on the potential that is assigned to them?
it depends on the potential function in use, i.e. the pair style. modelling hard spheres “exactly” is “hard” (pun intended) for an MD code structured like LAMMPS.
For example, I am using atom_style ellipsoid with assigned diameter, while I am assigning them gayberne potentials, the pairwise distance seems not limited by the assigned diameters but more by the assigned gayberne parameter. I am not sure if I have done the test wrongly.
Gay-Berne is a generalization of Lennard-Jones and thus by definition a “soft” potential. Please note, that while you can make your potentials “hard”, you will also have to adjust (= reduce) the length of the time step accordingly. the closest to hard spheres are the pair styles in the GRANULAR package.
axel.
Dear Axel
Thanks for the information. I understand that hard spheres in MD is difficult, but it resolves my doubt that I will definitely need to use pair_potential to adjust the “apparent size” of finite-size particles as it does not come by default. Many Thanks!
Best Wishes
Lunna
Dear Axel
Thanks for the information. I understand that hard spheres in MD is difficult, but it resolves my doubt that I will definitely need to use pair_potential to adjust the “apparent size” of finite-size particles as it does not come by default. Many Thanks!
one additional comment:
please note, that you have to carefully study the potential function formulation and the “Restrictions” section in the individual pair styles’ documentation to see, which information from the atom style is actually used. for example, if you use pair style colloid, then the particles’ diameters will be taken from the pair_coeff lines, regardless of whether you assign a diameter via atom style sphere or not.
axel.
Pair resquared in the ASPHERE package is not as soft as GayBerne. It goes to infinity
at a finite separation between 2 particles equal
to the sum of their radii. The approach to that
is soft, unlike a “hard sphere” model where
the interaction is zero until it jumps to infinity.
Steve