[gmx-users] (no subject)

[Jason de Joannis]

Overuse of constraints may by a little unrealistic. Why not include
an external field that represents the effect of two hard walls?
The same short-ranged (1/r^9) potential would be felt by all atoms.
Then put a vacuum layer above the top wall to screen out the
electrostatic interactions with z-dimensional images. You should
be able to run semi-isotropic pressure coupling provided the
vacuum width remains large enough. That way you can develop a
surface tension - lipid area isotherm. Unfortunately Gromacs does not
have an external field option, but it is easy to code.

 /Jason

> Date: Wed, 4 May 2005 18:20:59 -0400 (EDT)
> From: "David L. Bostick" <dbostick at physics.unc.edu>
> Subject: Re: [gmx-users] pbc = xy
> To: Discussion list for GROMACS users <gmx-users at gromacs.org>
> Message-ID:
>  <Pine.GSO.4.58.0505041807470.16505 at tiberius.physics.unc.edu>
> Content-Type: TEXT/PLAIN; charset=US-ASCII
>
>
> If this is a monolayer such as those used in, for example, dipole potential
> measurement experiments, then it will be natural to have a water-vacuum
> interface. If you wish to mimick an experimental setup, will the lipid
> tails be attached to a substrate? If so, you may want to tether them with
> position restraints.
>
> Typically, in order to simulate such interfaces, you will need to extend
> the box along the dimension perpendicular to the monolayer (usually
> the z-) plane by approximately 3 times the system size in this dimension.
> You can then use PME with a correction term, so called "Ewald 3dc" which
> will give you correct electrostatics for slab (2D) boundary conditions.
> This correction is built into gromacs, just use the option
>
> ewald_geometry = 3dc
>
> It should work unless the code was taken out. Oh yes... you will need to
> use the NVT ensemble so the simulation cell does not collapse, so be sure
> the monolayer setup is good ... with the correct area per
> surfactant molecule. You will see one or two water molecules escape from
> the liquid phase, but don't worry... if you are near ~300K, cohesive forces
> will hold everything together.
>
> For an example, see the following reference:
>
> J. Phys. Chem. B. 2000, 104, 5302-5308
>
> Good luck,
> David
>
> > On Wed, 2005-05-04 at 11:18 -0230, Jennifer Rendell wrote:
> > >
> > > I am modelling a system of a lipid monolayer and water (on the headgroup
> > > side). I am considering ways to prevent the water from jumping, via
> > > periodic boundary conditions, over to the chain side of the layer.
> > >
> > > I have had no success using position restraints on a separately
> > > defined layer of water near that box side. The water molecules still pop
> > > over to the opposite side.
> > >
> > > How difficult is it to modify the gromacs code to keep periodic boundary
> > > conditions in the x and y directions, but not have it in the z direction.
> > >
> > > Where would I start this kind of modification, if it is feasible?
> >
> > A simple alternative is to add a vacuum layer by extending the box by
> > twice its length.
> > >
> > > Thank you, Jennifer
> > >