Multiscale modeling of polymeric materials

Polymers are characterized by processes at length and time scales that span orders of magnitude. Molecular dynamics simulations using atomistic models are thus unable to access long-time processes such as diffusion, terminal dynamics and entanglements. This also results in a gap in the length and time scales probed by atomistic and mesoscale models. In this thesis, we develop coarse-grained models that bridge this gap, provide access to simulating entangled chains, and investigate terminal dynamics. These coarse-grained models retain chemical identity to a specific polymer by using information from atomistic models. The main focus is on predicting accurate dynamic evolution of polymers using coarse-grained models. In this work, we develop coarse-grained models for pure polymers and polymer blends.; In the first part of the thesis, we develop a coarse-grained model for polyethylene using information from an underlying atomistic simulation. The dynamics from the coarse-grained simulation are observed to be accelerated by a constant amount, called the speed-up factor. We illustrate that the speed-up factor is predictable and that the dynamics using the speed-up factor are in quantitative agreement with experimental observations. The speed-up factor is also shown be only dependent on temperature. Hence, we use the coarse-grained model to simulate longer polyethylene chains that are entangled. The results from the coarse-grained simulations of entangled chains demonstrate that the model accurately predicts both static and dynamics properties.; Subsequently, we extend the coarse-graining procedure to two more polymers that are widely used: poly(ethylene oxide) and poly(methyl methacrylate). Here also, we illustrate that static and dynamic properties from coarse-grained simulations are in excellent agreement with experiments and atomistic simulations. We demonstrate the applicability of the coarse-grained models to simulate longer and entangled polymers.; Next, we focus our attention on developing a coarse-grained model for poly(ethylene oxide)/poly(methyl methacrylate) blend. We observe distinct speed-up factors in the coarse-grained simulations, one for each component in the blend. Dynamics of each component calculated using their respective speed-up factors are shown to be in good agreement with viscosity and rheological measurements. We also investigate the effect of poly(methyl methacrylate) on the motion of poly(ethylene oxide) by simulating entangled chains.; Finally, we develop a novel adaptive hybrid model that allows using atomistic and coarse-grained polymer chains in the same simulation. This represents one of the first studies to perform multiscale simulation of polymers involving atomistic and coarse-grained scales. The model introduces a hybrid region between atomistic and united atom regions, which allows for a seamless transition from one scale to the other. We demonstrate that the multiscale simulation predicts static properties in excellent agreement with fully coarse-grained or fully atomistic simulations.