[gmx-users] intermolecular energy

Mark Abraham mark.abraham at anu.edu.au
Sat Jun 23 05:15:14 CEST 2007

> The intermolecular energy in the liquid should no be equal to the sum of
> the
> LJ, coulomb(SR, coulb(LR) and Disp Correction? And, that correct,
> shouldn't
> it be equal also to the potential energy minus bond minus angle bending
> energies? What am I missing, that is making me have both different?

Dunno since we can't read your mind about the sort of calculation you
did... What does your logfile have?

> Another question: what exactly is the "coul. recip." standing for? I
> expected it had to be included in the above sums also, but the high values
> it yelds would be nonsense.

It's the reciprocal space sum of the Ewald method you're using. It sounds
like you should read the section in chapter 4(?) of the manual that
describes this. The reason it's large is that it has (roughly) all the
long-range contributions from all of the atoms, and I don't think these
can be sensibly broken down into by-atom components.

So if you're interested in determining intermolecular energies, for some
reason, you should design a simulation protocol that has a chance of doing
this. In a GROMACS context, that probably means GRF electrostatics.


More information about the gromacs.org_gmx-users mailing list