GROMACS version: 2021
GROMACS modification: No
Hello,
I wish to calculate the dielectric constant for oleic acid.
I first created a simulation box I equilibrated to reproduce the actual density.
Then I simply used gmx dipoles and got a dielectric constant (1.16) twice less than the actual one (2.13).
Do you have any suggestion?
Cheers
Matteo
It sounds like the electrostatic parameters need to be adjusted.
Thank you for the input.
Then it is just to recalculate atomic charges ab-initio?
You would have to target some kind of meaningful QM data like (scaled) dipole moment, etc. or whatever is appropriate for the force field you’re using to derive the parameters. It will also depend on other aspects of the molecule’s topology. If it’s very flexible, the internal configurations that are sampled matter, so high-quality dihedral fitting is needed. And LJ parameters affect atomic distributions, so it all matters. The charges are the most directly relevant, but like anything in force field parametrization, all terms are linked.
I would think that a dielectric constant of around 2 is nearly always mainly the result of electronic polarization, which is not present in a molecular mechanics force field without polarizability. gmx dipoles computes dipole fluctuations due to fluctuations in atomic coordinates. These fluctuations dominate for any system with a dielectric constant that is a bit larger than 2.