Re: [ESPResSo] Self-avoiding random walk SAW

[Axel Arnold]

On Monday 02 August 2010 19:15:47 Mikheil Azatov wrote:
> Thanks, this explains a lot! I was doing it that way because i wanted all
> my polymers to stay inside the simulation box and couldn't figure out
> another way to do that beside putting each separately closer to the center
> of the box. When I do the way you suggested some of the polymers are close
> to the borders of the simulation box immediately leave after i start
> simulation which makes it hard to watch them in vmd.
>
> I tried using constraints to solve this but i couldn't figure out how they
> work. There are no samples with them and there's not much i the manual.

First of all, constraints keep the particles in a certain space, and that is 
something completely different than a periodic simulation box. Therefore, just 
for visualization inserting walls is not a good idea. To visualize polymers in 
VMD, take a look at the pbctools (http://www.espresso-
pp.de/projects/pbctools/). They allow different methods of clever folding so 
that the effects of the periodic boundary conditions are not as obvious.

> I set up the constraints like this on the borders of the simulation box.
> Then I define LJ interaction between particles and constrains and then I
> set the polymers.
> constraint wall normal 0 0 1 dist 0 type 3
> constraint wall normal 0 0 -1 dist 100 type 3
> constraint wall normal 0 1 0 dist 0 type 3
> constraint wall normal 0 -1 0 dist 100 type 3
> constraint wall normal 1 0 0 dist 0 type 3
> constraint wall normal -1 0 0 dist 100 type 3

> After I run the simulation I get errors like this :
> background_errors 0 {061 wall constraint 1 violated by particle 4} {061
> wall constraint 3 violated by particle 4}...
> What am I doing wrong ?

1. As the User's guide says, the distance is along the normal from 0. 
Therefore, the walls with negative coordinates should also have a negative 
distance, so for example

constraint wall normal -1 0 0 dist -100 type 3

2. If you don't switch it off, Espresso uses periodic boundary conditions. If 
your box is between 0 and 100, then the folded particle coordinates are always 
between 0 and 100, so your walls don't do much. However, written out to VMD 
are unfolded coordinates, which still can be far outside. The constraints 
option for the polymer command is only checks that a polymer does not cross a 
constraint, therefore doesn't help either, since the folded coordinates cannot 
cross your constraints.

Either, you switch off pbc via "setmd periodic 0 0 0", or you put the walls a 
bit inside the box, so, at distances 10 and 90, for example. The distance has 
to be large enough such that a bond cannot cross the gap (in folded 
coordinates), that is, bigger than 8 in your case.

Alternatively, you can place the polymer manually. Creating a simple random 
walk is not very difficult, and you can check for each bead that it does not 
leave whatever area you want to constrain your particles to. After that, you 
slowly switch on the chain-chain LJ interaction and the bonds.

> Also if the constraints are set up correctly can I use them to put all the
> polymers inside a smaller simulation box. I mean for example my box_l is
> 100. And I want to put all the particles inside the small box of size 50 in
> the middle of the actual simulation box. Could I use constraints for that ?

Yes, see above.

Many regards,
Axel

-- 
JP Dr. Axel Arnold Tel: +49 711 685 67609
ICP, Universität Stuttgart      Email: address@hidden
Pfaffenwaldring 27
70569 Stuttgart, Germany