The
self-assembly of binary polymer-grafted nanoparticles (NPs)
in a selective solvent is investigated using coarse-grained simulations.
Simulations are performed using theoretically informed Langevin dynamics
(TILD), a particle-based method that employs a particle-to-mesh scheme
to efficiently calculate the nonbonded interactions. The particles
are densely grafted with two immiscible polymers, A and B, that are
permanently bound to the NP either at random grafting sites (random-grafted)
or with all the A chains on one hemisphere of the NP and all the B
chains on the other hemisphere (Janus-grafted). For NPs with random
grafting, the polymers phase-separate on the surface of the NP to
form Janus-type structures in dilute solution, even though some of
the chains have to stretch around the particle to form the Janus structure.
When the solvent quality is sufficiently poor for the solvophobic
chains, the binary grafted NPs assemble into various structures, including
double-walled vesicles. In particular, vesicles are formed when the
solvophilic volume fraction is between 0.2 and 0.3, in a similar range
to that required for vesicle formation in diblock copolymers in a
selective solvent. For mixed-grafted NPs, there is considerable variation
in the structure of each individual NP, but nevertheless, these NPs
form ordered vesicles, similar to those formed by Janus-grafted NPs.