[gmx-users] pressure for ambient water

[Justin A. Lemkul]

Amit Choubey wrote:
> 
> 
> On Sun, May 16, 2010 at 7:41 PM, Justin A. Lemkul <jalemkul at vt.edu 
> <mailto:jalemkul at vt.edu>> wrote:
> 
> 
> 
>     Amit Choubey wrote:
> 
> 
> 
>         On Sun, May 16, 2010 at 5:15 PM, Justin A. Lemkul
>         <jalemkul at vt.edu <mailto:jalemkul at vt.edu>
>         <mailto:jalemkul at vt.edu <mailto:jalemkul at vt.edu>>> wrote:
> 
> 
> 
>            Amit Choubey wrote:
> 
>                Hi Justin,
> 
>                Since the density (1 gm/cc) and T (300 K) correspond to
>         ambient
>                condition, should not the equation of state dictate a
>         pressure
>                around 1 atm?
> 
> 
>            If the equation of state involves temperature and pressure, yes.
> 
>         So if you fix the density and temperature shouldnt you land up
>         with right pressure which we know should be 1 atm.
>          
> 
> 
>     You are not fixing (or conserving) the temperature in an NVE
>     ensemble.  That would be an NVT ensemble, employing a thermostat.  
> 
> 
> I did do NVT first and then for sampling i removed the thermostat. Also 
> as you mentioned there was not much difference b/n pressure values 
> during NVT or NVE . The pressure value is as high as 1000 bar in both cases.
> 
>  
> 
>     Since instantaneous pressure is calculated (in part) from the
>     kinetic energy, and since the kinetic energy is not guaranteed to be
>     conserved, the pressure term will also fluctuate accordingly.
> 
> 
> This is correct. The pressure thus fluctuates about 20 % of the above 
> mentioned value.
> 
> 
>     http://www.gromacs.org/Documentation/Terminology/Pressure
> 
>     The internal energy of the system is constant in an NVE ensemble,
>     the other terms may fluctuate as necessary such that all microstates
>     occur with the same probability and the energy surface remains flat. 
> 
> 
> I agree
>  
> 
>      Also recall that an NVE ensemble represents a thermodynamically
>     isolated system, not conducting heat or engaging in work with the
>     surrounding system. 
> 
> 
> true
>  
> 
>      So any concept of external pressure and equilibrating the pressure
>     is irrelevant.
> 
> 
> I am not trying to equilibriate the pressure. I am trying to measure the 
> pressure. I also know that at 300 K and 1gm/cc the P should be 1 bar.
> 
> Also, theoretically NVE or NVT are no different than NPT as far as 
> measuring observables is concerned. Hence i was thinking that if you 
> have the right volume density and temperature shouldnt you have the 
> right pressure.

I think the disconnect is arising because you're expecting a model of water to 
behave almost like an ideal gas.  The SPC water model, under NPT conditions of 
300 K and 1 bar, does not give the experimental density of water; it is actually 
somewhat less than 1 g/mL.  So constraining the system to fit some pre-conceived 
notion of the volume to force the density to be right conflicts with the 
properties of the water model itself.  SPC wants to be at a lower density, 
you're forcing it to be at a higher density, all while fixing the volume of the 
simulation cell.  Sounds to me like a recipe for high pressure, since SPC wants 
to expand but you won't let it.

So the starting configuration, assembled with the right density, has not 
properly equilibrated under NPT conditions, yet you are expecting it to do so 
when applying NVE conditions.  I don't know that you'll ever be able to satisfy 
all of these requirements simultaneously unless you can come up with a better 
water model that replicates both ideal and real behavior :)

-Justin

-- 
========================================

Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
MILES-IGERT Trainee
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================