[gmx-users] [Fwd: NPAT ensemble]

[Erik Lindahl]

David van der Spoel wrote:

>-----Forwarded Message-----
>
>From: Markus Müller <markus.mueller at unibas.ch>
>To: David van der Spoel <spoel at xray.bmc.uu.se>
>Subject: NPAT ensemble
>Date: 10 Feb 2003 11:46:53 +0100
>
>Hi David,
>
>I would like to simulate lipids and to conserve fluid phase during the simulation
>
>by applying constant area per lipid.
>
>How can I do this with GROMACS ?
>
>Using semiisotropic pressure coupling with compressibility close to zero in the xy plane ?
>  
>

The compressibility should be about the same as water, i.e. 6.5e-5. 
Unless you want the box to be
deformable (i.e. triclinic instead of rectangular) you should set the 
off-diagonal elements to zero, i.e.

compressibility = 6e-5 6e-5 6e-5 0 0 0


The reference pressure should be one atm in x,y,z directions, and zero 
off-diagonal. Asymptotically,
you should get the same result with anisotropic or semiisotropic 
scaling, but they might have slightly
different properties on short timescales.

Cheers,

Erik