[gmx-users] calculating dielectric constant

[Qiao Baofu]

Hi David,
Thanks for your kindly reply.
Because I am simulating a pure fluid system, I used the epsilon-r = 1in the
.mdp file.
_____________________________________________________
coulombtype              = pme
rcoulomb-switch          = 0
rcoulomb                 = 1.1
; Dielectric constant (DC) for cut-off or DC of reaction field
epsilon-r                = 1
_____________________________________________________

 If I don't use the Reaction-Field method, how should I use the -rpsilonRF
option? (I want to use 0. But because 0 means infinity here, I have no idea
how to use it)  I calculated dielectric constant with and without
"-esilonRF" option, please see the following results
_____________________________________________________
g_dipoles -temp 425 -f md2 -s md2 -enx md2.edr
Epsilon = 826.496

g_dipoles -temp 425 -epsilonRF 1 -f md2 -s md2 -enx md2.edr
Epsilon = -2.01094
_____________________________________________________

Neither result are the values what I expect. What wrong?



2006/11/13, David van der Spoel <spoel at xray.bmc.uu.se>:
>
> Qiao Baofu wrote:
> > Hi all,
> >
> > I have some question about calculating dielectric constant:
> > 1. From the manual, it seems that there are two methods to calculate the
> > dielectric constant: a. use g_dipoles -temp 400 -f -s directly, b) use
> > g_dipoles -nonormalize -corr total  -c dipcorr.xvg, then g_dielectric -f
> > dipcorr.xvg -o epsw.xvg. (Please correct me if the commandlines are
> > wrong). Thus I want to know the difference between them.
>
> g_dielectric does another analyis. You want to run g_dipoles either with
> the -enx flag or with a trajectory.
>
> > 2. If I want to calculate the dielectric constant, must I use the
> > Reaction Field in the couloumbtype? Now I have some results from PME.
> > Can I use these results directly to calculate the dielectric constant?
>
> PME is fine. You have to supply the right dielectric constant to the
> program (though the default of 0 == infinity) is what most people use.
>
>
> > Thanks very much in advances!
> >
> >
> >
> > Sincerely yours,
> > Baofu Qiao
> >
> >
> >
> --
> David.
> ________________________________________________________________________
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20061114/7201b298/attachment.html>