[gmx-users] Re: cut offs

[Michael Shirts]

Adding a dispersion correction corrects for the truncation of the r^-6 term in
the LJ term in a "mean field" way.  Read over the manual for more details.
It's the DispCorr keyword in an .mdp file.  This should make the
density/energy of a box of water almost independent of the VdW cutoff distance
chosen, as long as it's over 0.8 nm or so.

Right now, the dispersion correction is only calculated correctly for abrupt
cutoffs.  I have given code to Erik Lindahl to calculate the dispersion
correction correctly for Shifted and Switched VdW cutoffs as well --
presumably that will be in later versions of Gromacs.

Cheers,
Michael


> I currently have both cutoffs LJ and Couloumb cutoffs set 1.5 nm. The energy
> of the water in 10% more stable and more dense with a 1.5 nm LJ cutoff then
> a .8 nm LJ cutoff. This causes the solvation free energies to increase
> dramatically about 50 % in some cases. Any advice on how to hack gromacs So
> that I can have a different cuttoffs for the solute-h2o than h20-h20.

Cheers,
Michael