The bottom-up prediction of the properties of polymeric materials based on molecular dynamics simulation is a major challenge in soft matter physics. Coarse-grained (CG) models are often employed to access greater spatiotemporal scales required for many applications, but these models normally experience significantly altered thermodynamics and highly accelerated dynamics due to the reduced number of degrees of freedom upon coarse-graining. While CG models can be calibrated to meet certain properties at particular state points, there is unfortunately no temperature transferable and chemically specific coarse-graining method that allows for modeling of polymer dynamics over a wide temperature range. Here, we pragmatically address this problem by “correcting” for deviations in activation free energies that occur upon coarse-graining the dynamics of a model polymeric material (polystyrene). In particular, we propose a new strategy based on concepts drawn from the Adam−Gibbs (AG) theory of glass formation. Namely we renormalize the cohesive interaction strength and effective interaction length-scale parameters to modify the activation free energy. We show that this energy-renormalization method for CG modeling allows accurate prediction of atomistic dynamics over the Arrhenius regime, the non-Arrhenius regime of glass formation, and even the non-equilibrium glassy regime, thus allowing for the predictive modeling of dynamic properties of polymer over the entire range of glass formation. Our work provides a practical scheme for establishing temperature transferable coarse-grained models for predicting and designing the properties of polymeric materials.

The dynamical characteristics of ring and linear polyethylene (PE) molecules in the melt have been studied by employing atomistic molecular dynamics simulations for linear PEs with carbon atom numbers N up to 500 and rings with N up to 1500. The single-chain dynamic structure factors S(q,t) from entangled linear PE melt chains, which show strong deviations from the Rouse predictions, exhibit quantitative agreement with experimental results. Ring PE melt chains also show a transition from the Rouse-type to entangled dynamics, as indicated by the characteristics of S(q,t) and mean-square monomer displacements g 1 (t). For entangled ring PE melts, we observe g 1 (t) ∼ t 0.35 and the chain-length dependence of diffusion coefficients D N µ N -1.9 , very similar to entangled linear chains. Moreover, the diffusion coefficients D N remain larger for the entangled rings than the corresponding entangled linear chains, due to about a 3-fold larger chain length for entanglement. Since rings do not reptate, our results point toward other important dynamical modes, based on mutual relaxations of neighboring chains, for entangled polymers in general.

The mass scaling of the self-diffusion coefficient D of polymers in the liquid state, D ∼ M(β), is one of the most basic characteristics of these complex fluids. Although traditional theories such as the Rouse and reptation models of unentangled and entangled polymer melts, respectively, predict that β is constant, this exponent for alkanes has been estimated experimentally to vary from -1.8 to -2.7 upon cooling. Significantly, β changes with temperature T under conditions where the chains are not entangled and at temperatures far above the glass transition temperature Tg where dynamic heterogeneity does not complicate the description of the liquid dynamics. Based on atomistic molecular dynamics simulations on unentangled linear alkanes in the melt, we find that the variation of β with T can be directly attributed to the dependence of the enthalpy ΔHa and entropy ΔSa of activation on the number of alkane backbone carbon atoms, n. In addition, we find a sharp change in the melt dynamics near a "critical" chain length, n ≈ 17. A close examination of this phenomenon indicates that a "buckling transition" from rod-like to coiled chain configurations occurs at this characteristic chain length and distinct entropy-enthalpy compensation relations, ΔSa ∝ ΔHa, hold on either side of this polymer conformational transition. We conclude that the activation free energy parameters exert a significant influence on the dynamics of polymer melts that is not anticipated by either the Rouse and reptation models. In addition to changes of ΔHa and ΔSa with M, we expect changes in these free energy parameters to be crucial for understanding the dynamics of polymer blends, nanocomposites, and confined polymers because of changes of the fluid free energy by interfacial interactions and geometrical confinement.

Using atomistic molecular dynamic simulations we study the transitions between solid herringbone and liquid crystalline hexagonal mesophases of discotic liquid crystals formed by hexabenzocoronene derivatives. Combining a united atom representation for the side chains with the fully atomistic description of the core, we study the effect of side chain substitution on the transition temperatures as well as molecular ordering in the mesophases. Our study rationalizes the differences in charge carrier mobilities in the herringbone and hexagonal mesophases, which is predominantly due to the better rotational register of the neighboring molecules.

Flory theorem," stating that polymers in the melt are nearly "ideal," [28] in the specific sense that they should have similar excluded volume (EV) interactions as polymers at their theta point, i.e., ν = 1/2, simply does not hold for ring melts. Instead, ν for ring melts has often been reported to be near 2/5 [11,12,21,24] which is consistent with the theoretical estimate of Cates and Deutsch. [29] Recent computational studies [16][17][18]30,31] have suggested that very long rings in the melt should be in a collapsed globular state for which ν equals 1/3. On the other hand, Lang et al. [32] have argued for a value of ν = 3/8 for high mass rings, an exponent close, but not equal to the value for collapsed chains. Despite the uncertainty in the exact exponent value, it seems safe to conclude that ν is significantly less than 1/2 so that the conformation of melt rings does not conform to random walk (RW) chains with screened excluded volume interactions.The absence of screening of excluded volume in rings implies that the addition of solvents and linear polymers to ring melts should lead to different effects than adding these molecular species to linear polymer chain melts. This phenomenon has been considered in previous work, [33][34][35][36][37][38][39][40][41][42][43][44][45][46][47][48] but the nature of this swelling process is not yet entirely clear, and the present work is devoted to further elucidating this problem from a new perspective not relying on the reptation model. There have admittedly been numerous arguments, [29,33,34] experiments, [33][34][35][36][37][38][39][40][41] and simulations [42][43][44][45][46][47][48] pertaining to the configurational properties and dynamics of rings diluted by linear chains. Originally, Cates and Deutsch [29] suggested that the topologically induced EV interactions in the ring melt are relieved upon diluting the rings by linear chains so that the rings in the melt should become "ideal" upon dilution by linear chains. Recent simulations [44][45][46][47][48] and experiments [40,41] on ring-linear polymer blends have reported that rings increase their size upon the dilution and their conformational structure resembles ideal rings in a theta solvent, clearly supporting a change in the EV interactions in ring melts upon adding linear chains. Moreover, the rate of diffusion of tracer ring polymers in the melt was found to be greatly slowed by adding linear chains, an effect interpreted as arising from the ring-linear polymer threadings. [33,[35][36][37][38][39]43,48] Although there have been many attempts at quantifying this physically appealing "threading effect," [7,[35][36][37][38][43][44][45][46]48] the validity of the threading model has not yet been established. In this communication, we consider an alternative explanation of the slowing down of Ring-Linear Chain BlendsAtomistic molecular dynamics simulations of ring-linear polyethylene blends are employed to understand the relationship between chain conformational structure and the melt dynamics of these blends. As o...