[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [ESPResSo] Question about bond-angle interaction

From: Olaf Lenz
Subject: Re: [ESPResSo] Question about bond-angle interaction
Date: Thu, 18 Oct 2007 08:56:49 +0200
User-agent: Mozilla/5.0 (X11; U; Linux i686 (x86_64); en-US; rv: Gecko/20070728 Thunderbird/ Mnenhy/

Hash: SHA1

Hi Lorenzo,

Lorenzo Isella wrote:
> My problem is to have these particles "sticking" when they collide. This
> is essential for me, since I want to study the complicated, absolutely
> non-spherical, aggregates they give rise to.

Only now I understand, that by "sticking" you mean, that the particles
should also keep a given angle towards other particles in the same
compound. So far, I supposed you wanted particles that are just strongly
bound to each other.

> inter 4 angle 1000.0 3.14
> part 0  bond 4 0 0
> Leads to some strange errors. Actually, the number after part should
> probably be the individual number stuck on each particles and not the
> type but then how can I say that I want all the particles to interact
> according to that potential?

A bond is always between individual particles, and not between particle
types. So you will have to give the individual particle numbers.

I think the best solution for your problem goes along Burkhard's
suggestions: as soon as particles collide, you will have to create a
bond between them make them stick. Instead of using a FENE bond, it is
probably best to use a rigid bond, i.e. a bond that has a fixed distance.

Unfortunately, it will be necessary to check for a particle collision on
 the Tcl level after every few integration steps, which might slow down
the whole simulation quiet a bit, but it should be doable.

To keep the angle between three particles fixed, I see two possibilities:
1. You could indeed use a bond-angle potential, as you have tried. Note
however, that the potentials always have a certain flexibility. You have
tried to reduce this flexibility by using a very large interaction
parameter. This, however, will require a very small timestep, otherwise
you might get broken bonds. What you would really need, is a rigid angle
potential, i.e. a potential with a given, fixed angle, that cannot be
changed. Such a potential is not yet implemented in ESPResSo, but it
could probably be done. Note however, that this is not as trivial as it
might sound.
2. A better solution might be to simply use two rigid bonds to bind the
new sticking particle to two of the old particles. However, I'm not
completely sure, how the integrator handles this case. Just test it, it
might just work.

I hope this helps.
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Mozilla -


reply via email to

[Prev in Thread] Current Thread [Next in Thread]