I’m simulating a nanoparticle (sphere out of a fcc latticle) as a rigid body. How can I make the nanoparticle interaction with its surrounding happens from the surface of the particle, basically I want a potential i.e E(r-d/2) where r is the distance between the center of the nanoparticle and d is the nanoparticle diameter. I was thinking of using wall/region but I don’t think that’s applicable as the nanoparticle is bouncing arounrd, right? I appreciate any pointers.
Kasra.
I'm simulating a nanoparticle (sphere out of a fcc latticle) as a rigid
body. How can I make the nanoparticle interaction with its surrounding
happens from the surface of the particle, basically I want a potential i.e
E(r-d/2) where r is the distance between the center of the nanoparticle and
d is the nanoparticle diameter. I was thinking of using wall/region but I
don't think that's applicable as the nanoparticle is bouncing arounrd,
right? I appreciate any pointers.
what is the point of building a rigid body from individual atoms, if
you won't use that property in your simulation. if you simply want
your particle to interact with its environment, you could just use a
colloid pair style and drop the rigid body.representation.
axel.
Axel,
Thanks for your quick answer. How about if I want the particle interacts with its surrounding via WCA potential (LJ126 with rc = 2^1/6)? Then you think using the rigid body is pointless as the cutoff is so small that it doesn’t penetrate deep into nanoparticle (nanoparticle diameter is 4sigma)? I’m just interested in hydrodynamic effects on the nanoparticle and not its deformation though.
Kasra.
Axel,
Thanks for your quick answer. How about if I want the particle interacts
with its surrounding via WCA potential (LJ126 with rc = 2^1/6)?
that is an atomic potential. just define it and be done.
Then you
think using the rigid body is pointless as the cutoff is so small that it
doesn't penetrate deep into nanoparticle (nanoparticle diameter is 4sigma)?
no. but you were asking for an averaged potential where the
nanoparticle turns into an ideal sphere and then there is no reason to
model it from individual atoms unless you include a friction term
orthogonal to the purely distance dependent model.
I'm just interested in hydrodynamic effects on the nanoparticle and not its
deformation though.
your question was asking for something different than what you want to
model. you can't expect a correct answer to the wrong question.
axel.
Axel,
Maybe my question was confusing because I wanted to reproduce a simulation in the paper “Molecular simulations of lubrication and solvation forces” Challa and Swol. In which they create a particle out of a fcc lattice and they set the mass of this particle to infinity to not to be affected by its surrounding. That’s why I made the particle to be rigid and then they say that the particle and the solvent interact via a 12-6 LJ (WCA) as E(r-d/2) in which r is the distance to the center of the particle and d is its diameter, “which indicates that the potential is acting from the surface of the sphere”. That’s why I asked the question in my first post. I hope I could make it clearer now. Now I like to know your opinion about it, do you think colloid is the better option in this case or its basically the same?
Kasra.
Axel,
Maybe my question was confusing because I wanted to reproduce a simulation
in the paper "Molecular simulations of lubrication and solvation forces"
Challa and Swol. In which they create a particle out of a fcc lattice and
they set the mass of this particle to infinity to not to be affected by its
surrounding. That's why I made the particle to be rigid and then they say
that the particle and the solvent interact via a 12-6 LJ (WCA) as E(r-d/2)
in which r is the distance to the center of the particle and d is its
diameter, "which indicates that the potential is acting from the surface of
the sphere". That's why I asked the question in my first post. I hope I
could make it clearer now. Now I like to know your opinion about it, do you
think colloid is the better option in this case or its basically the same?
either the paper is describing an inconsistent model or you are not
representing it correctly.
if you have a purely distance dependent model you don't need to
represent it by individual particles,.but if the individual particles
have a 12-6 LJ interaction, then integrating over it would lead to
different powers. the colloid pair style describes exactly such a
particle and its interactions with a 12-6 LJ fluid. see:
http://lammps.sandia.gov/doc/pair_colloid.html
otherwise to have a static non-moving (nano) particle in a fluid, just
don't include it in time integration and set the WCA type LJ
interaction for the particle solvent interaction. this will be an
explicit representation of the colloid particle. to save computational
effort, you can use neigh_modify exclude to skip over computing the
interactions inside the nano particle.
this is all straightforward to do. i am not going to say which is
"better" because that all depends on what you want to learn from it.
but doing some kind of weird hybrid like you describe, is definitely
inconsistent and thus not as good.
axel.
Axel,
Thanks for your thorough answer. You’re right I could just exclude the particle from time integration but that’d be a non-moving particle as you mentioned. I’m using “fix rigid” mainly because of the capability that it has in constraining translation and rotation degrees of freedom. Basically I can position the particle at a distance from the lj93 wall and lock its vertical position from the wall and sample the force on the particle while it’s free to move around in that plane and rotate, however, I don’t how to do this if I just not include the nanoparticle in the time integration?
Kasra.
Axel,
Thanks for your thorough answer. You're right I could just exclude the
particle from time integration but that'd be a non-moving particle as you
mentioned. I'm using "fix rigid" mainly because of the capability that it
has in constraining translation and rotation degrees of freedom. Basically I
can position the particle at a distance from the lj93 wall and lock its
vertical position from the wall and sample the force on the particle while
it's free to move around in that plane and rotate, however, I don't how to
do this if I just not include the nanoparticle in the time integration?
nobody will be able to give proper advice, if you keep changing your
story *every* time you reply. as far as i am concerned, this was one
time too many.
have a nice day,
axel.
Well I’m not changing the story, I was asking to know how I can get a similar capabilities as in “fix rigid” force and torque keyword if I just exclude nanoparticle from time integration or use colloid style.
Kasra.
Well I'm not changing the story,
just re-read the thread.
first you ask about assigning a continuum potential to a particles
made from individual atoms.
then you say you want to reproduce a calculation where a nano particle
was built from atoms with infinite mass (i.e. cannot move).
then you say you don't care about the nano particle only about the fluid
now you say you *do* care about the nano particle and want it to move,
and what it to experience torque, which is impossible with a distance
only potential.
now you say that this is not contradicting.
I was asking to know how I can get a
similar capabilities as in "fix rigid" force and torque keyword if I just
exclude nanoparticle from time integration or use colloid style.
that is a pretty dumb question. of course, you cannot. how should it
be possible?
good luck,
axel.
Fix rigid is a time integrator (for whatever particles are in the rigid bodies).
So if you can do what you want with fix rigid, by using the force and
torque flags, then that’s all you need.
Steve