[gmx-users] How DispCorr influsnces pressure

[Michael Shirts]

> Dispcorr is only for homogeneous liquids. It should not be used for
> membranes.

More precisely -- the dispersion correction is an analytical
correction to both the pressure and energy that is rigorously correct
in the limit of the radial distribution function g(r)=1 outside the
van der Waals cutoff.  For a lipid bilayer, this not a good
approximation, since the system is not isotropic.

Since there is not "right" way to do membrane systems currently, the
best bet is to go with the same cutoff treatment the lipids were
parameterized with.

DispCorr = Ener is probably still a good idea, since g(r)=1 isn't
perfect, but it's better than completely ignoring the long range
energy.  DispCorr = Ener will not change the configurations sampled,
so any phase properties, pressures, etc. will be unaffected, but the
results will be somewhat less cutoff dependent.

See:
http://pubs.acs.org/doi/abs/10.1021/jp0735987
for some additional information.

Members of the Gromacs team are working behind the scenes for various
possible solutions to the problem for nonisotropic systems.  Don't
expect anything until 5.0, though.

Best,
Michael