[gmx-users] compressing the system

[Justin A. Lemkul]

Moeed wrote:
> 
> Hello,
> 
> My efforts for compressing a polymer chain all failed. Even when I am 
> compresing very slowly with tau-p=10 and ref_p of 1 bar simulation 
> crashes when box size is close to chain size.. I tried different 
> pressures of 1 10 30 and tau_p of 1 2 5 10 and cobination of these...Has 
> anyone idea on suitable parameters ?
> 
> ;        Pressure coupling
> Pcoupl              =  Parrinello-Rahman
> Pcoupltype          =  semiisotropic   
> tau_p               =             1         1       
> compressibility     =  4.5e-5 4.5e-5           ;
> ref_p               =  1.0                1.0           
> 

I have no real experience with polymers, but these settings bring several 
questions to my mind:

1. Why are you using Parrinello-Rahman for a system that is not yet 
equilibrated?  The fluctuations in pressure for such a system could be huge.

2. Why are you using semi-isotropic pressure coupling?  This setting is more 
suited to interfacial or membrane systems.  For a box of PE, which should end up 
likely as an entangled systems eventually, what purpose does it serve?  I would 
think isotropic coupling would be more appropriate.

Please also remind me (and everyone on the list) what's in your system?  You've 
been asking essentially the same question for what seems like weeks and I can't 
exactly remember the details.  Was it 8 linear PE molecules in a rectangular 
configuration?  If so, what makes you think that such an arbitrary configuration 
will even result in a reasonable system?  The problem is the linearity, I would 
assume.  The box is collapsing in before the chains can re-orient themselves in 
any way.  I would think you could generate a far better starting structure by:

1. Simulating a single PE chain under NVT conditions (in a large box) such that 
the chain can re-orient and collapse on itself somewhat.

2. Constructing a new system from this collapsed molecule, and then worry about 
how to get a reasonable density.

-Justin

> Thanks,
> 

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

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

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