Coupled length and time scales determine the dynamic behavior of polymers and underlie their unique viscoelastic properties. To resolve the long-time dynamics it is imperative to determine which time and length scales must be correctly modeled. Here we probe the degree of coarse graining required to simultaneously retain significant atomistic details and access large length and time scales. The degree of coarse graining in turn sets the minimum length scale instrumental in defining polymer properties and dynamics. Using linear polyethylene as a model system, we probe how coarse graining scale affects the measured dynamics. Iterative Boltzmann inversion is used to derive coarse-grained potentials with 2-6 methylene groups per coarse-grained bead from a fully atomistic melt simulation. We show that atomistic detail is critical to capturing large scale dynamics. Using these models we simulate polyethylene melts for times over 500 µs to study the viscoelastic properties of well-entangled polymer chains.Polymer properties depend on a wide range of coupled length and time scales, with unique viscoelastic properties stemming from interactions at the atomistic level. The need to probe polymers across time and length scales to capture polymer behavior makes probing dynamics, and particularly computational modeling, inherently challenging. With increasing molecular weight, polymer melts become highly entangled and the longtime diffusive regime becomes computationally inaccessible using atomistic simulations. In these systems the diffusive time scale increases with polymerization number N faster than N 3 , becoming greater than 10 10 times larger than the shortest time scales even for modest molecular weight polymers. While it is clear that the largest lengths scales of polymer dynamics are controlled by entanglements, the shortest time and length scales required to resolve dynamic properties are not obvious. This knowledge is critical for developing models that can transpose atomistic details into the long time scales needed to model long, entangled polymer chains.One path to overcoming this computational challenge is to coarse grain the polymer, reducing the number of degrees of freedom and increasing the fundamental time scale. The effectiveness of this process depends on retaining the smallest length scale essential to capturing the polymer dynamics. The process of coarse graining amounts to combining groups of atoms into pseudoatom beads and determining the bead interaction potentials [1,2]. Simple models like the bead-spring model [3], capture characteristics described by scaling theories, but disregard atomistic details and cannot quantitatively describe properties like structure, local dynamics or densities. Immense efforts have been made to systematically coarse grain polymers and bridge the gap of time and length scales while retaining atomistic characteristics [4]. One critical issue underlying the coarse graining process is the degree to which a polymer can be coarseA single C96H194 PE chain represented with in...

The distinctive viscoelastic behavior of polymers results from a coupled interplay of motion on multiple length and time scales. Capturing the broad time and length scales of polymer motion remains a challenge. Using polyethylene (PE) as a model macromolecule, we construct coarse-grained (CG) models of PE with three to six methyl groups per CG bead and probe two critical aspects of the technique: pressure corrections required after iterative Boltzmann inversion (IBI) to generate CG potentials that match the pressure of reference fully atomistic melt simulations and the transferability of CG potentials across temperatures. While IBI produces nonbonded pair potentials that give excellent agreement between the atomistic and CG pair correlation functions, the resulting pressure for the CG models is large compared with the pressure of the atomistic system. We find that correcting the potential to match the reference pressure leads to nonbonded interactions with much deeper minima and slightly smaller effective bead diameter. However, simulations with potentials generated by IBI and pressure-corrected IBI result in similar mean-square displacements (MSDs) and stress autocorrelation functions G(t) for PE melts. While the time rescaling factor required to match CG and atomistic models is the same for pressure- and non-pressure-corrected CG models, it strongly depends on temperature. Transferability was investigated by comparing the MSDs and stress autocorrelation functions for potentials developed at different temperatures.

No abstract

No abstract