Re: Bond Breaking in Polymer Networks - United Atom Forcefield

From: Axel Kohlmeyer (akohlmey_at_gmail.com)
Date: Tue Nov 13 2012 - 10:53:11 CST

norman,

On Tue, Nov 13, 2012 at 5:40 PM, Norman Geist <
norman.geist_at_uni-greifswald.de> wrote:

> This is not a NAMD question but a charmm/amber question. The breaking of
> bonds is not possible in both of them, and
>

not really.

> usually not in general for classical forcefields, but I heard of one that
> is supposed to be able to.
>

breaking, creating and swapping bonds is quite a normal process
in the simulation of (model) polymers with classical potentials.
those steps are often treated like monte carlo events. the
LAMMPS (classical) MD code, has such a feature, for example.

however, it is quite difficult to do this well (the swap between the
disabling the bonded potential and enabling the corresponding
non-bonded interaction, that was previously excluded must not
lead to a (significant) discontinuity in the forces).
things get even more complicated, when angles and dihedrals
are involved. since updating those interactions is not as simple,
then just the bonds alone. particularly, when trying to avoid those
pesky discontinuities in the forces that would easily to a divergence
in integration of the equation of motions. also, to do this in a
parallel code has its problems. often it would be easiest to
stop the MD, update the topology file from an external program
or script and then continue the MD.

the LAMMPS code i mentioned above, only supports bond breaking
or creating or swapping for models that have only bonds. including
angles and dihedrals would required significant additional programming
effort. people have asked for it repeatedly, but to the best of my
knowledge, none of them has yet stepped up to implement it.

****
>
> ** **
>
> What you are maybe interested in is QM and QM/MM hybrid solutions.
>

those, and even reactive force fields like REAXX are
computational overkill and not easily applicable to most
of these cases.

cheers,
     axel.

> ****
>
> ** **
>
> Norman Geist.****
>
> ** **
>
> *Von:* owner-namd-l_at_ks.uiuc.edu [mailto:owner-namd-l_at_ks.uiuc.edu] *Im
> Auftrag von *Stefanos Anogiannakis
> *Gesendet:* Dienstag, 13. November 2012 11:15
> *An:* namd-l_at_ks.uiuc.edu
> *Betreff:* namd-l: Bond Breaking in Polymer Networks - United Atom
> Forcefield****
>
> ** **
>
> Dear all,****
>
> ** **
>
> I'm interested in performing MD simulation of polymer networks under
> tension,****
>
> where some chains could break as elongation proceeds.****
>
> ** **
>
> I use a United Atom (UA) forcefield. ****
>
> The bonded interactions of my model include:****
>
> (i) a bond stretching potential****
>
> (ii) an angle potential (bond bending)****
>
> (iii) a dihedral (torsional) potential.****
>
> ** **
>
> The nonbonded, pairwise interactions are based on a 6-12 Lennard-Jones
> (LJ) potential,****
>
> which acts between all possible pairs (inter- and intra-chain), ****
>
> except from intra-chain pairs that are separated by 1, 2, or 3 sequential
> bonds,****
>
> (the so called (1-5) intra-chain LJ potential).****
>
> ** **
>
> I would like to ask you if NAMD is able to break bonds 'on the fly', as
> simulation proceeds,****
>
> according to a distance criterion for the bond length.****
>
> ** **
>
> In terms of potentials, a bond scission event means that ****
>
> ** **
>
> (a) For the UAs making this bond, the bond potential will not be evaluated
> (or will be turned off) after breakage ****
>
> (b) The angle and dihedral potentials related with this bond will cease to
> exist (or be evaluated)****
>
> (c) A LJ pairwise interaction between the UAs of the broken bond will
> start to apply.****
>
> ** **
>
> Thank you very much for your help,****
>
> Stafanos****
>
> ** **
>

-- 
Dr. Axel Kohlmeyer  akohlmey_at_gmail.com  http://goo.gl/1wk0
International Centre for Theoretical Physics, Trieste. Italy.

This archive was generated by hypermail 2.1.6 : Mon Dec 31 2012 - 23:22:15 CST