[gmx-users] Pressure coupling and membrane-type simulations

[Peter C. Lai]

On 2012-02-21 03:41:07PM -0500, Justin A. Lemkul wrote:
> 
> 
> Andrew DeYoung wrote:
> > Hi,
> > 
> > I am interested in doing a membrane-type simulation, in which I have
> > all-atom membrane "walls" parallel to xy plane, at z = -z_0 and z = +z_0
> > (where z_0 is a constant).  I would like to run an NPT simulation at 1 atm.
> > 
> > 
> > What type of pressure coupling should I use?  Isotropic pressure coupling
> > requires only one input value (either compressibility or reference pressure
> > ref_p), whereas I think that semiisotropic, anisotropic, and surface-tension
> > pressure coupling require specification of both the compressibility and
> > ref_p (tensors).
> > 
> > Clearly, I should not use isotropic pressure coupling, because clearly my
> > system is not isotropic.  However, what if I do not know and cannot find in
> > the literature the compressibility of the liquid system that I am placing
> > between the membrane "walls"?  If I do not know the compressibility very
> > accurately or at all, then it seems that I cannot use semiisotropic,
> > anisotropic, or surface-tension pressure coupling.  
> > 
> > If you have time, I would like to ask an additional question.  Now suppose I
> > know the compressibility of the liquid between the membrane.  Now what
> > pressure coupling type should I use; should I use semiisotropic,
> > anisotropic, or surface-tension pressure coupling?  Both semiisotropic and
> > surface-tension look reasonable.  In the manual
> > (http://manual.gromacs.org/current/online/mdp_opt.html#pc), semiisotropic
> > pressure coupling is useful for systems that are isotropic in x and y, but
> > different in z (which is the situation I have here).  Surface-tension also
> > looks like it describes a similar situation, but it requires the
> > specification of the surface tension of the liquid, which I do not know.  
> > 
> > I am sorry that my questions are quite vague.  If you have time, do you have
> > any general thoughts?  Or can you please recommend any papers that would
> > help me understand and choose between the pressure coupling types?
> > 
> 
> I doubt I can really answer much of this, but isn't the most pressing 
> consideration the walls themselves?  If they're reasonably rigid, the 
> compressibility of the fluid layer is largely irrelevant, isn't it?  Then again, 
> if the walls are not rigid, then they become a liability under pressure coupling 
> as they may buckle.  Seems to me this should be the principal concern.
> 
> -Justin
> 

Another consideration is that the documentation states when using built-in
implicit walls at z=0 and z=zbox to use anistropic pressure coupling with the 
x/y compressibility set to 0 "otherwise the surface area will change".
Maybe start with that and go from there?

-- 
==================================================================
Peter C. Lai			| University of Alabama-Birmingham
Programmer/Analyst		| KAUL 752A
Genetics, Div. of Research	| 705 South 20th Street
pcl at uab.edu			| Birmingham AL 35294-4461
(205) 690-0808			|
==================================================================