[lammps-users] Doubts about intermolecular potential parameters (lj/cut/coul/long)

EmersonPL · December 7, 2020, 5:51am

Dear colleagues,

I am trying to understand how the intermolecular potential is calculated by the Lammps to obtain intermolecular potential parameters properly from quantum mechanics calculations (to use them later in Lammps simulations, i mean).

I intend to use pair_style lj/cut/coul/long, whose equations are described below:

My question is: what would be the distances between atoms (i.e., “r”) to be considered in Lammps intermolecular potential energy calculations (i.e., both Lennard-Jones and Coulomb potentials)?

For example, in the pair of methane molecules below (where C is the carbon, H is the hydrogen and numbers are IDs):

1 - Would Lammps consider “ij” interactions of only different molecules when calculating the intermolecular potential (i.e. Lennard-Jones and Coulomb)? (e.g., C1C2; C1H5; C1H6; C1H7; C1H8; C2H1; C2H2; C2H3; C2H4; H1H5; H1H6; H1H7; H1H8; H2H5; H2H6; H2H7; H2H8; H3H5; H3H4; H3H7; H4H8).

2 - Or would Lammps consider all possible “ij” interactions? (e.g., C1C2; C1H1; C1H2; C1H3; C1H4; C1H5; C1H6; C1H7; C1H8; C2H1; C2H2; C2H3; C2H4; C2H5; C2H6; C2H7; C2H8; H2H1; H3H1; H4H1; H5H1; H6H1; H7H1; H8H1; H3H2; H4H2; H5H2; H6H2; H7H2; H8H2; H4H3; H5H3; H6H3; H7H3; H8H3; H5H4; H6H4; H7H4; H8H4; H6H5; H7H5; H8H5; H7H6; H8H6; H8H7).

If “1” is true, for a situation where I have 5 atoms connected in sequence in the same molecule (e.g. 1-2-3-4-5), would I have to compute the interactions of atoms 1-5 in the same molecule?

Sorry in advance if my question was too “basic” or if it is out of the scope of this mailing list.

Thanks for a while.

Best regards,

Emerson Parazzi Lyra

PhD candidate

School of Chemical Engineering

University of Campinas
500 Albert Einstein Ave, Campinas, SP, 13083-852, Brazil

Email: [email protected]

Orcid ID: https://orcid.org/0000-0002-7969-3764

akohlmey · December 7, 2020, 6:38am

which intra-molecular non-bonded interactions are included depends on the exact bond topology and the factors set for lennard-jones and coulomb in the special_bonds command.
since those factors depend on the chosen force fields, you would be best off to look up the details for the corresponding force field in the publication(s) describing said force field.
you should be doing this anyway if you are looking to create parameters that should be compatible with a given force field, as those have very specific rules and algorithms how the individual bonded and non-bonded parameters are determined and fitted.

please note that quantum chemical calculations in general are most suitable for computing parameters for bonded interactions given the usual procedure of (mostly) excluding such interactions from the non-bonded interactions. while molecule-solvent interactions are far more difficult to derive and accurate determination of dispersion dominated non-bonded interactions can be very challenging (DFT for example can sinificantly underestimate those). since you are using hydrocarbons as example, you may want to specifically look at the publication(s) for the OPLS/AA or the TraPPE family of force fields for inspiration of how that procedure may be implemented.

Axel.

EmersonPL · December 7, 2020, 8:31am

Thank you very much for your explanation Dr. Axel Kohlmeyer!