Re: high pressure

From: Giacomo Fiorin (giacomo.fiorin_at_gmail.com)
Date: Thu Mar 11 2010 - 13:00:56 CST

Hi Tillmann, as Dimitar suggested, you can keep the NPT ensemble, by using
harmonic position restraints. If you also set their force constants to
reproduce the expected fluctuations in the gold lattice, it actually
improves the physical description a whole lot.

In general, and not only regarding temperature and pressure coupling, it's
always better to use harmonic restraints as opposed to fixed atoms, unless
you have a particular reason to keep those positions constant up to the last
digit.

Giacomo

---- ----
Giacomo Fiorin
  ICMS - Institute for Computational Molecular Science
    Temple University
    1900 N 12 th Street, Philadelphia, PA 19122
work phone: (+1)-215-204-4216
mobile: (+1)-267-324-7676
mail: giacomo.fiorin_at_gmail.com
---- ----

On Thu, Mar 11, 2010 at 11:29 AM, Tillmann Utesch <
tillmann.utesch_at_mail.tu-berlin.de> wrote:

> Hi Dimitar,
>
> Thanks for the quick reply.
>
> I used a fixed gold layer, because I don't have bonding parameters for gold
> atoms. I have heard of problems of fixed atoms and an NPT enesemble
> and tried to solve the discrepancy by excluding the surface from pressure
> calculations (without changing the pressure problem).
> The observed pressure fluctuations after a 2 ns long run are around 2050
> bar (+- 150 bar). If I decrease the charge density, the average value moves
> to lower pressure,
> so I think it is a special problem with high immobile charge densities.
> In all simulations, I set the LangevinPistonTarget to 1 athmosphere, but I
> did not reach a fluctuation around this target.
>
> Another point is that introduction of ions into the system leads to
> pressure fluctuations around the desired value of 1 athmosphere after
> equilibration of the ions.
>
> I will try a NVT ensemble now and hope it will solve the problem.
>
> Best regards,
> Tillmann
>
>
>
> ----- Original Message ----- From: "Dimitar V Pachov" <
> dpachov_at_brandeis.edu>
> To: "tillmann utesch" <tillmann.utesch_at_mail.tu-berlin.de>
> Cc: <namd-l_at_ks.uiuc.edu>
> Sent: Thursday, March 11, 2010 4:09 PM
> Subject: Re: namd-l: high pressure
>
>
>
>
>> ----- "tillmann utesch" <tillmann.utesch_at_mail.tu-berlin.de> wrote:
>>
>> | From: "tillmann utesch" <tillmann.utesch_at_mail.tu-berlin.de>
>> | To: namd-l_at_ks.uiuc.edu
>> | Sent: Thursday, March 11, 2010 7:01:06 AM GMT -05:00 US/Canada Eastern
>> | Subject: namd-l: high pressure
>> |
>> | Hi all,
>> |
>> | I run a NPT simulation of an enzyme on a SAM coated gold surface.
>> | During the simulation all gold atoms were fixed to their positions.
>>
>> It is not a good idea at all to run NPT with any positional constraints.
>> Pressure is controlled by volume scaling. Fixed coordinates introduce
>> artifacts and certain frequency wave propagation/fluctuations that are not
>> desired. You cannot say that your system samples appropriately the
>> equilibrium NPT ensemble in that case.
>>
>>
>> | The total system is neutrally charged, but the head-groups of the
>> | monolayers are positively and the gold atoms are negatively charged.
>> |
>> | So the charges are not able (gold) or limited (SAM) in moving.
>> | The simulations are very stable, but the the pressure is fluctuating
>> |
>> | around a very high value.
>>
>>
>> What is "very high"? How large are the fluctuations? Instantaneous
>> pressure tends to fluctuate a lot, but your average should be around the
>> externally specified value.
>>
>>
>> | For the pressure control I used the Langevin
>> |
>> | piston method. Forces between fixed atoms were ignored and for the
>> | electrostatic interactions I used the PME summation.
>> | Exclusion of gold and SAM atoms of the pressure calculations did not
>> |
>> | change the very high pressure in the system.
>> | What shall I do now? Is the high pressure a result of the high charge
>> |
>> | density in the SAM and the gold? How can a modify my system to get
>> | normal pressure values?
>>
>> What is "normal"?
>> I do not know why you want to fix the gold atoms, but it would be much
>> better (but not prefect) to put (small) positional harmonic restraints which
>> will allow at least some spacial degrees of freedom. Or you could try NVT.
>>
>> Dimitar
>>
>> |
>> | Thanks,
>> | Tillmann
>> |
>> | -- | Tillmann Utesch
>> | Institut für Chemie, Max-Volmer-Laboratorium
>> | TU Berlin, PC 14
>> | Straße des 17. Juni 135
>> | D-10623 Berlin
>> | Tel. +49-(0)30-314-26389
>>
>> --
>> ========================================================
>> Dimitar V Pachov
>>
>> PhD Candidate
>> Physics & Biochemistry Department Phone: (781) 736-2326
>> Brandeis University, MS 057 Email: dpachov_at_brandeis.edu
>> ========================================================
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:53:53 CST