(159l) Multiscale Modeling of Polystyrene in the Solution and in the Melt

To investigate the behaviors of
polymers on the local scale, atomistic simulations are required, and it would
be highly desirable to have a multiscale model which can include all relevant
interactions. Polymer dynamics of longer chains or higher concentrations cases
can only be captured by a mesoscale model. We investigate the polystyrene
dynamical behaviors on multiple length scales and different environments.

Solvents significantly alter
polymer properties and can lead to favorable or unfavorable changes. We
investigated the liquid structure and dynamics of a mixed solvent of
cyclohexane and N, N-dimethylformamide molecules in the neighborhood of atactic
polystyrene oligomers¹. We found consecutive side rings to be
perpendicular at local scales and to parallel each other due to the quadrupolar
interaction at slightly larger scales. Polystyrene side rings are the primary
targets for the solvents with a strong preference to cyclohexane over N,
N-dimethylformamide. Increasing temperatures speeds up the movement of all the
small molecules and increasing concentrations slows down the reorientation
process.

Some phenomena, for example, entanglement or phase
separation behaviors, occur on length scales not accessible by a fully
atomistic simulation. We build a mesoscale polystyrene model based on the
atomistic simulation of pure 48 chains of polystyrene at chain length of 15
monomers². We group the 16 atoms of a monomer together as a
superatom and its center has been placed on the carbons connecting the backbone
with the side ring³. ?Iterative Boltzmann Inversion is used to
reproduce the structure by means of radial distribution functions. Simulation
results provide an extensive description of the transition from Rouse motion to
reptation behaviors. Our analysis provides persistent evidence for the
conclusion that the entanglement length of this coarse-grained PS model is
about 85 monomers, which is not too far below the experimental value of about
130 monomers.

We elucidate the possibility to
optimize a coarse-grained model for a blend of cis-polyisoprene and polystyrene
as a test case⁴. Experimentally, this system is known to be miscible
at short chain lengths, and demixing is observed at longer chain lengths. The
generalization of Iterative Boltzmann Inversion to binary blends requires
optimizing three potentials from the corresponding distribution functions:
polyisoprene-polyisoprene, polystyren-polystyrene, and
polyisoprene-polystyrene. The iteration of blends is technically more demanding
in that mapping takes into account the effects of three sets of highly
correlated potentials. It is effective to start with those potentials that are
lease affected by changes in the others. We started the iteration of
polyisoprene-polyisoprene and polystyrene-polystyrene pairs first as they are
relatively independent of each other compared to the degree to which they
interdepend on the polyisoprene-polystyrene potential. Using our mesoscale
model, we can embark now on a study of the phase behavior of the blend. The
phase separation sets in at around seven monomers. Various morphologies of
lamella, cylinder, and sphere are observed, which are associated with
equi-mass, medium unbalanced, and extreme concentration ratios⁵.

Reference:

1. Qi Sun and Roland Faller J Phys Chem B 109(33),
   15714, 2005.
2. Qi Sun and Roland Faller Macromolecules 39,
   812, 2006.
3. Qi Sun and Roland Faller Comp & Chem Eng. 29,
   2380, 2005.
4. Qi Sun and Roland Faller J Chem Theor Comp, 2006.
5. Qi Sun and R. Faller J Am Chem Soc in preparation.