Re: [AMBER] Maximum input value for the isothermal compressibility in AMBER

From: David A Case <david.case.rutgers.edu>
Date: Thu, 25 Feb 2016 09:06:39 -0500

On Thu, Feb 25, 2016, BLEY Michael wrote:
>
> I just started a project on determining the equation of state for both
> liquid and gaseous water by molecular dynamics with the POL3 water
> model. For the liquid state it was obviously no big deal to determine
> the relation between pressure and density at room temperature, but for
> the gaseous state I have some problems with the upper limit of the
> isothermal compressibility in AMBER. The highest value which seem to be
> accepted as valid input is 9.999 * 10^-3 bar^-1.
> Is it somehow possible to use higher values for the COMP-input parameter
> in AMBER or is there no limitation for this value at all?
> It is the first time that I am using the AMBER mailing list, so let me
> please know if some important or data is missing. Thank you very much
> for your efforts.

The value of compressibility is only used as a way to convert units in
the Berendsen barostat: it is not related in any real way to isothermal
compressibility. So (in principle) you can just use the default (liquid)
value.

Having said that, I'm not sure you can effectively use Amber to look at
non-ideal gases (but it might be fun to try!). Maintaining the desired
temperature and pressure is tricky in a gas. I think you would have to
use a Langevin thermostat and the Monte Carlo barostat, and be prepared
for lots of experimentation. Your density will be a thousand times
smaller than what the code was designed for, so there may be lots of
hidden gotchas. I'd certainly recommend starting with a simpler water
model (such as TIP3P) first, just to reduce the number of sources of
error.

My guess is that at best, you will get back the ideal gas limit for the
equation of state, but I might be wrong. Sounds like an interesting
educational exercise. Maybe someone else on the list has actually tried this
and can chime in here.

...dac


_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Feb 25 2016 - 06:30:04 PST
Custom Search