Dear Steve and Paul,
I want to simulate n-alkanes using united atom force field. I am considering alkane molecule
to be a chain of Ch3-ch2-…-ch2-ch3. In the united atom force field, bonds are of fixed length,
and there are harmonic angles, dihedrals etc.
In LAMMPS, this seems possible only by using POEMS (SHAKE or RIGID not suitable). I defined
each bond as a rigid body, however I get an error “ERROR: Rigid body has degenerate moment of inertia”.
The documentation states that each body should have at least 3 non-colinear atoms.
I also tried out the examples uploaded by Paul in the new wiki site:
http://lammps.wetpaint.com/page/LAMMPS+POEMS+input+scripts
but I get the same error.
Alkanes with this force-field have been simulated in LAMMPS
(“Substructured MD using Multibody Dynamics Algorithms” R. M. Mukherjee, Paul Crozier, Steve Plimpton and Kurt Anderson)
Could you please give some inputs as to how this was done?
Looking forward to your inputs.
Thanks and Warm Regards,
Gopal
Dear Steve and Paul,
I want to simulate n-alkanes using united atom force field. I am considering
alkane molecule
to be a chain of Ch3-ch2-.....-ch2-ch3. In the united atom force field,
bonds are of fixed length,
and there are harmonic angles, dihedrals etc.
please explain why the bonds should be rigid?
you could use a harmonic bond potential just as well (or better).
there are suitable parameters (non-bonded and bonded) for this in
the Gromos and the CHARMM force field. i am certain that there are
also additional, more specifically parametrized interaction potentials
for n-alkanes in the literature.
please keep in mind, that LAMMPS is a "simulation engine".
the work of finding the proper choice of potentials is first and foremost
yours, not that of the LAMMPS documentation or the mailing list.
In LAMMPS, this seems possible only by using POEMS (SHAKE or RIGID not
suitable). I defined
each bond as a rigid body, however I get an error "ERROR: Rigid body has
degenerate moment of inertia".
The documentation states that each body should have at least 3 non-colinear
atoms.
yes, of course. otherwise the angular velocity would become infinity.
modeling an n-alkane as a rigid body is a very strange model and
very likely not what you want. there are a lot of e-mails on that subject
in the mailing list archive. a lot of people get confused about what
fix rigid does.
cheers,
axel.
I'm not clear what your model is. Do you mean flexible chains with rigid
bond lengths? I'm not sure that POEMS can do that, or that it was
done in the paper. Paul or Rudra (co-authors) would know better. I would
send them an email if you think that's what it says POEMS did.
Steve