adding bonds during simulation

Dear Axel,
  
  Your reply seems to mean I have to modify some codes in LAMMPS to achieve the dynamical bond creation process? Is there is available command in LAMMPS to get it?

Best wishes,
Yangpeng

在 2012-1-17,下午2:44, Axel Kohlmeyer 写道:

2012/1/17 Yangpeng Ou <[email protected]...>:

Dear Axel,

       Your reply seems to mean I have to modify some codes in LAMMPS to achieve the dynamical bond creation process? Is there is available command in LAMMPS to get it?

none that i can think of.
your request is too specific and a bit unusual.

there may be multiple ways to
implement a scheme like this. my first choice
would be to implement a custom pair style
and add this to the other interactions via a
hybrid/overlay. thus in some way the inverse
of what bond/quartic does.

the custom pair style would maintain a list
of bond partners and then augment the regular
non-bonded interaction between the "bonded"
atoms so that it represents the bonded forces/energies
and not the non-bonded interaction. this would
require passing around the extra information
as atoms migrate between processors, but there
are other styles that do this and can be used as
a template.

axel.

Dear Axel,

My understanding of your method is to mimic bond interaction by own defined pair style. I forgot to mention another constrain in my model. After a bond is created between a free monomer and a tri-fuctional linker, the bond will live and do not break forever. It looks this scheme need much work to let it run. I have seen a paper, they discussed this scheme. However, They used COGNAC (http://octa.jp/OCTA/cognac.html) to finish the project.

Best wishes,
Yangpeng

在 2012-1-17,下午2:54, Axel Kohlmeyer 写道:

2012/1/17 Yangpeng Ou <[email protected]...>:

Dear Axel,
My understanding of your method is to mimic bond interaction by own defined
pair style. I forgot to mention another constrain in my model. After a bond
is created between a free monomer and a tri-fuctional linker, the bond will

that is how i understood your model. this would work with the scheme
that i have been proposing. i have done something similar in the past
as a correction that adds a many-body term to a regular pairwise potential
to reduce over coordination.

live and do not break forever. It looks this scheme need much work to let
it run. I have seen a paper, they discussed this scheme. However, They used

in order to maintain bonds once they are formed, that information
would have to be stored by the pair style so you would know about
open valences.

the major part of the effort is to understand how to maintain
and pass around this kind of information internally. in some
cases, one would use a specific fix to handle that kind of
data. this is for example done by the peridynamics pair styles,
where you have the situation that you have non-bonded interactions
that - once they are broken - are gone forever.

the major difference in your case is that your bonding
will be on top of the regular interactions, which is not
different in peridynamics.

another issue to worry about is to make sure that
the switch between non-bonded and bonded is smooth
and has no discontinuities.

cheers,
     axel.

Dear Axel,
  
  Thanks for your reply. I need some google search to understand your proposal.

Best wishes,
Yangpeng

在 2012-1-17,下午3:26, Axel Kohlmeyer 写道: