Re: fixed atoms / pressure

From: Peter Freddolino (petefred_at_ks.uiuc.edu)
Date: Sun Mar 16 2008 - 09:19:53 CDT

Hi Arturas,
I don't have enough experience working with systems like the one you're
describing to give you too much advice, but my thoughts on this, for
what they're worth, are:
I think for your case flexiblecell and useconstantratio is the way to
go; this is also the way people would generally do systems with
membranes, and allows you to better account for the anisotropic pressure
response of your system Also, why would you want to use fixed particles
in this case? As noted in the namd manual, using a large number of fixed
particles with pressure control is generally not recommended. If you do
know exactly how big your x-y dimensions are, and your quartz slab is
relatively small compared to the water in the z direction, then you
should be ok use constantarea and flexiblecell if you fix the quartz.
However, like I said, I don't have experience with this type of system,
so these are just my instincts on the problem.
Best,
Peter

Arturas Ziemys wrote:
> Hi, Peter,
>
> My mistake about units - forgot: it is particles per cubic Angstrom.
>
> As far as I am dealing with anisotropic (quartz) system first time, I
> want to find/get more information tunning MD parameters. The whole
> point in my question was to get more insight is there any caveats
> using flexible cell, constant area, and significant number of frozen
> particles.
>
> I figured out about my pressure and density - it is consequence of my
> system structure. However I still not 100% if I should use frozen
> particles (minimized; forces ON) with constant_area/flexible_cell. Or,
> let's say, is it reasonable to use frozen particles of my quartz slab
> with flexible cell and without constant area - I'm not sure how NAMD
> would deal with that.
>
> best,
> Arturas
>
>
>
> Peter Freddolino wrote:
>> Hi Arturas,
>> what units are you using for water density here? Also, is the
>> PRESSURE value you're reporting averaged over the entire run? Are you
>> sure you ran long enough for it to be equilibrated?
>>
>> Using constant pressure with a lot of fixed atoms (eg, the whole
>> quartz slab) is probably not a good idea in general... if you must,
>> you might want to try with fixedAtomForces on, but you're probably
>> better off applying constraints to a few atoms.
>>
>> Also, how large is your water layer above and below the quartz slab?
>> And how are you calculating the density of water in your
>> (heterogeneous) system?
>>
>> Best,
>> Peter
>>
>> Arturas Ziemys wrote:
>>> Hi,
>>>
>>> I need some insights and comments.
>>>
>>> MY SYSTEM: quartz slab (~15 A wide; patched to mimic infinite layer)
>>> on XY plane and TIP3P water on both sides of Z. useFexibleCell is
>>> ON. useConstantArea is ON.
>>>
>>> PROBLEMS: 1) I get higher water density (~0.037) than TIP3P
>>> "standart" (~0.033-0.034) at NPT/1bar. 2) PRESSURE fluctuates around
>>> very low values (~ -1000 bars).
>>>
>>> However, my quartz is minimized and fixed after. I get higher water
>>> density using system with with fixed and not fixed quartz slab.
>>>
>>> QUESTIONS: How necessary to use useFexibleCell ON and
>>> useConstantArea ON in my case ?
>>>
>>> Best
>>> Arturas
>>

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:49:20 CST