Dear Lammps users,
I have a system which has a certain attribute. The non
bonded interactions are divided between local and
non-local, meaning that a different set of epsilon and
sigma are applied if the distance of the atoms are
more than four bonds appart and another if the
distance is three or four bonds appart.
Is it possible to do that in Lammps?
I have seen the pair_style command which enables you
to choose different cut-offs for different types of
atoms.
Yet, if the type of atom is the same can I use
different set of arguments?
Thank you in advanve,
Claus
Jetzt Mails schnell in einem Vorschaufenster �berfliegen. Dies und viel mehr bietet das neue Yahoo! Mail - www.yahoo.de/mail
Dear Lammps users,
claus,
I have a system which has a certain attribute. The non
bonded interactions are divided between local and
non-local, meaning that a different set of epsilon and
sigma are applied if the distance of the atoms are
more than four bonds appart and another if the
just to reconfirm: do you mean you have differnt interactions
for particles that are excluded (with special bonds) and
regular non-bonded or do the two cutoffs apply to anything
(in that case it would be interesting to know the settings
you use for the special bonds key word)?
in the first case you have to build a special form of the
respective bond, angle, dihedral and improper classes
that include the non-bonded term, in the second case you
have to build a special case for the non-bonded terms.
i've done something similar recently for our CMM coarse
grained model and it is pretty straightforward to do
(LAMMPS is designed to make this easy).
cheers,
axel.
The special_bonds command will let
you apply a scale factor to 1-2, 1-3, 1-4
interactions. These are what I would
call atom pairs that are 1,2,3 bonds apart.
Nothing can be done for atoms 4 bonds
apart (1-5 interactions?). You'd have
to add some reasonably complex code
(see special.cpp) just to find these pairs.
If this is not sufficient for your model, I think
you'd need to add your own pair potential
that read in these additional epsilon/sigma
parameters and applied them as you wish.
See the lj/charmm potentials as a starting
point; they read in extra epsilon/sigma for
1-4 interactions.
Steve
Hello,
the lj/charmm potential, if I understood correctly, is
close to what I'm looking for, yet I see that I can
use different parameters only for 1-4 interactions and
not for 1-5 interactions as well. In case I do insist
to use this potential, do I have to modify the code so
as to have the same set of non bonded parameters both
in 1-4 and 1-5 interactions?
If the answer is affirmative could you tell me which
is the file that I have to modify?
Thank you once again,
Claus
--- Steve Plimpton <[email protected]> schrieb:
LAMMPS doesn't compute or store a set of
interactions that are 1-5. Only 1-2,1-3,1-4.
That is done in special.cpp. So if you want
to find 1-5 interactions, you'd have to extend
special.cpp in a non-trivial way.
Then you'd have to modify/extend one of the pair
classes to do the 1-5 calculation you want.
You'd also have to modify the neighbor list
routines to flag when an interaction is 1-5.
I don't know of any force-field that worries
about 1-5 interactions. So if you're doing
this on your own, you might want to re-think
why you're doing it.
Steve