[gmx-users] kappa in Generalized reaction field

[Berk Hess]

>From: Tomoshi Kameda <kameda-tomoshi at aist.go.jp>
>Reply-To: Discussion list for GROMACS users <gmx-users at gromacs.org>
>To: gmx-users at gromacs.org
>Subject: [gmx-users] kappa in Generalized reaction field
>Date: Tue, 16 May 2006 14:35:31 +0900
>
> > >
> > >Do we want to keep this code at all?
> >
> > Good question.
> > I think simulating ionic solutions with reaction-field causes large
> > artifacts,
> > unless the concentration is very high and the cut-off quite long.
> > PME is a much better choice.
>
>Really?
>Recently, I read a paper, which shows the GRF and PME yielded results
>consistent with experiment by using GROMACS ......
>
>http://pubs.acs.org/cgi-bin/abstract.cgi/jpcbfk/2005/109/i45/abs/jp051325a.html

I must say that I disagree with this conclusion.

Everybody would probably agree that plain cut-offs produce horrible 
artifacts.
This paper shows that the native state is sampled in the same way by
PME and RF. But the sampling of non-native conformations is very different
for PME and RF. Therefore I would conclude that RF is not a good method
for such a system.

BTW in this case, or in general for simulations in water, RF of GRF makes
no difference as any dielectric constant larger than roughly 20 equals
infinity.

Berk.