Inquiry on RDF: Selecting Group-ID

Dear Dr. Axel Kohlmeyer,

Hello Dr. Kohlmeyer. I deeply appreciate your many advices on my inquiries to LAMMPS.

I would like to ask a question on compute rdf in LAMMPS.

I have a system 100 atoms in my simulation box where I have 40 Type 1 atoms and 60 Type 2 atoms interacting via LJ potential.

I wanted to see how the Type 2 atoms are distributed with respect to Type 1 atoms. I named by Type 1 atoms as “PN” and my Type 2 atoms as “SN.” Therefore, I included the following command:

compute myRDF all rdf 200 1 2 cutoff 5.0

According to the documentation, if I specified “all” as the group-ID, then the radial distribution function will be normalized with the bulk density of the entire system.

However, when I specified the “PN” as the group-ID, I was curious if the bulk density will only consist of the Type 2 atoms as if the following:

compute myRDF PN rdf 200 1 2 cutoff 5.0

I compared my simulation result between using these two commands, and I obtained exact same RDF for the two cases, one that considers all atoms and other that considers only Type 1 atoms.

I was curious if there are ways to have the radial distribution function only compute the bulk density of the specific types of atoms desired. Thank you.

Sincerely,

Masato Koizumi

compute 1 fluid rdf 500 1 1 1 2 2 1 2 2

Dear Dr. Axel Kohlmeyer,

Hello Dr. Kohlmeyer. I apologize that my previous message was confusing.

I would like to ask if the group-ID in the compute rdf command has any contributions to the radial distribution calculations.

That is, since radial distribution function is local density divided by the particle (bulk) number density, and if I specified:

compute myRDF all rdf 200 1 2 cutoff 5.0

I understand that the local density will be divided by the particle density of the entire system for “all” atom types available in the system.

Meanwhile, if I specified:

compute myRDF SN rdf 200 1 2 cutoff 5.0,

and SN refers to only the Type 2 atoms contained in my entire system, I was curious if the local density will be divided by the particle density of ONLY SN atoms, NOT the entire system consisting all atoms types.

I tried these two cases, and I obtained the exact same RDF curves. I was curious if there are anyways that I can have control of dividing the local density with respect to the specific particle density of the system based on the type of atoms I specified.

Thank you so much for your time.

Sincerely,

Masato Koizumi