Recovering dynamics and material properties in chemically specific coarse-grain models of polymer melts using a dissipative potential

Coarse-grain (CG) representations of polymeric materials aim to reduce computational effort, enabling simulations to reach pertinent length and time scales, while preserving some features, typically, either of the all-atom (AA) representation ("bottom-up" methods) or material properties ("top-down" methods). Here, we combine a bottom-up CG model with a dissipative potential to design a chemically specific and dynamically correct model. First, we utilize iterative Boltzmann inversion (IBI) to parameterize a conservative potential which recovers AA structure. Second, we introduce Langevin dynamics and parameterize the dissipative potential, the Langevin friction coefficient, to correct for the sped-up dynamics of the IBI-generated force-field.  We recently showed that the friction coefficient of the CG representation can be parameterized to recover AA dynamics for various measures of translational and rotational diffusion [J. Chem. Phys. 154, 084114 (2021)]. Here, we additionally compare to a friction coefficient derived from a material property, viscosity. We show that the viscosity-based friction may be predicted by a simpler dynamics-based friction. We point to possible CG model modifications that yield consistent dynamics and material properties.