[gmx-users] NPT compress Box using negative Pressure

[Elton Carvalho]

On Wed, Nov 25, 2015 at 1:49 AM, Christian Bope Domilongo
<christianbope at aims.ac.za> wrote:
>
> Dear all,
>
> Need help on how to set negative pressure in order to compress the Box.
> I have used negative pressure (-0.2) but the box is still expending.
>
> ; Pressure coupling
> Pcoupl                   = parrinello-rahman
> Pcoupltype               = semiisotropic
> nstpcouple               = -1
> ; Time constant (ps), compressibility (1/bar) and reference P (bar)
> tau-p                    = 1.0 1.0
> compressibility          = 0 4.5e-5
> ref-p                    = 1.0 -0.2
>
> Regards,
> Bope

Dear Bope,

I would expect that *positive* pressures would compress the box. If
you set the reference temperature to a negative value, the barostat
would tend to increase the box volume such that the pressue measured
in the box drops to the negative reference value. Try a positive
reference pressure.

-0.2 is a really small pressure, much smaller in absolute value than
the typical fluctuations in NPT. Is this value sensible? What
pressures would you expect experimentally to comprepress your box?

How dense is your box? Do you have a dense liquid or a low density gas? Vacuum?

Are you trying to compress your box (in the z-direction) so the box
height is adequate in order to reproduce some density? In this case
you are running a pre-equilibration and maybe the berendsen barostat
would give you better results (converge faster, flucutuate less). All
depends on what you are trying to accomplish.

Greetings from a cloudy yet hot Natal,
-- 
Elton Carvalho