Mesoscopic simulation of viscoelastic fluid using LAMMPS

Dear LAMMPS users,

I want to model a system as rigid spherical particles disperse in a dilute solution comprised of the polymers and an underlying Newtonian solvent. Our studies should consider the interaction between the polymer chains and particles at nano-scale, as well as the hydrodynamics of the solvent. Other group has used multiparticle collision dynamics (MPCD) to deal with similiar cases as shown in following snapshot. I wonder whether the lammps package can also model this case using fix srd.


Any suggestions will be appreciated.

Best Regards,


Dear Su

   It is not too difficult to create a complex system with polymers
and particles mixed together. I wrote a molecule builder to help with
this process (, but there are several other
good programs which do the same thing at if you are familiar with the
LAMMPS data file format, you can write you own script to build these
files yourself if you prefer. Here are some snapshots of simulations
of cells we've been working on:

--- walls ----

You can create a smooth cylinder barrier to confine the polymers using
"fix wall"

More complex wall shapes can be formed by joining multiple regions
together, or by arranging atoms on the boundary, and immobilizing
them. (Which you can do by omitting them from the group of atoms that
you move using fix nve, fix langeven. That's how I created the blue
ellipsoidal wall in the second image.)

--- disparate sizes ---

  One issue with your simulations is the large difference in particle
sizes in your simulations. Some of your particles are much bigger
than others. Some of my simulations have this issue too. This will
slow your simulations. But it is still possible to run these kinds of
simulations efficiently in LAMMPS if you use the the "multi" options
with the "comm_modify" and "neighbor" commands:

Alternately, you can construct large hollow spheres using many small spheres
(For example, see
To hold the small spheres together, you can use "fix rigid". I am not
sure if this would be faster.

--- hydrodynamics ---

   I was not clear exactly how you wanted to model the solvent. And
even if you had indicated it, I wouldn't have been able to offer much
helpful advice. Getting the dynamics correct is not too important in
the simulations I have run so far, and I never had to worry about
   There are a lot of fixes, packages, and external programs which
couple particles in LAMMPS to a fluid in various ways. I strongly
suspect you will be able to find something which will do what you

I don't think I answered all your questions, but hopefully I convinced
you that you can probably do what you're trying to do in LAMMPS.
Good luck.