Session Z19: Dynamics and Transport in Polymers

Withdrawn

The use of atomistic simulations is currently being used in addressing questions of mechanism in many materials, such as self-assembly phenomena in polymers, or host-guest binding applied for drug design, among others. When combining simulations with advance sampling methods that are able to bias the potential energy surface based on preset slow collective variables the simulation improves its efficiency, and this opens the possibility to treat more realistic complex systems. In this work, we will follow the analysis of the free energies governing the interactions of complex systems by employing the Artificial Neural Network sampling method developed by Hythem and Whitmer (JCP 2018). The discussion will highlight the configurational sampling using atomistic simulations of a polymer chain model, and its interaction with different small-weight molecules, representing with their interaction the plasticization process. In particular, we focus on conformational and hydrogen bond-structure changes induced in globules of a polymer chain by the plasticizer molecules, with the hypothesis that hydrogen bonding plays a role in the incorporation into polymer materials, and thus in the observed mechanical properties. The findings derived from this system showcase physical features relevant to the design of tailored materials, and the methods developed are intended to be part of a robust framework applicable to an assortment of experimental works useful for industrial proposes.   

Validation and refinement of unified analytic model for flexible, semiflexible and stiff polymer melt entanglement

Two of us recently proposed [R. S. Hoy and M. Kröger, Phys. Rev. Lett. 124, 147801 (2020)] analytic expressions for the reduced entanglement length L_e/l_K, tube diameter a/l_K, and plateau modulus Gl_K³/k_BT of polymer melts that unified previously-incompatible scaling theories for flexible, semiflexible and stiff polymers. In this talk, we stress-test these expressions by checking whether they remain valid when a different topological analysis (TA) method and different Ne-estimators are employed, and by comparing them to the available experimental data on flexible, semiflexible and stiff polymer melts. We find that when we correct for all known sources of systematic error in N_e-estimation (e.g. chain-end effects and the non-Gaussian statistics of primitive paths) and reduce our statistical uncertainties by sampling large numbers of model well-entangled melts, we obtain simplified versions of our original expressions that (i) quantitatively match bead-spring simulation data over the entire range of chain stiffnesses for which melts remain isotropic, and (ii) semiquantitatively match all available experimental data.   

Segmental dynamics of normal and pre-oriented PMMA glasses in the strain-hardening regime

Strain-hardening is a feature of polymer glasses during large deformation, which helps to stabilize the glasses against breakage. Little is known about the segmental dynamics during strain hardening, and such data is important for building a molecular-level theory of this process. Here, using a photobleaching technique, we measure the segmental dynamics of normal and pre-oriented PMMA glasses during deformation into strain-hardening regime at temperatures from T_g-30K to T_g-20K, with local strain rates from 3×10^-5 s^-1 to 9×10^-5 s^-1. The pre-oriented glasses were prepared by stretching above Tg followed by rapid cooling.  For both normal and pre-oriented glasses, we find that the segmental relaxation time follows a power law relationship with the strain rate in the strain-hardening regime with an exponent near 0.85. For the normal sample, the relationship between segmental relaxation time and strain rate is the same as in the early  flow regime. At the same engineering strain, the segmental dynamics of the normal and pre-oriented samples are almost the same, with the pre-oriented sample having dynamics about 15% faster.   

Role of solvent in enhancement of thermal conductivity of epoxy/graphene nanocomposites.

Uniform dispersion of graphene nanoparticles into epoxy is critical for achieving high thermal conductivity epoxy-graphene nanocomposites. Uniform dispersion can reduce gap between graphene nanoparticles increasing the potential of direct contact between them, establishing thermal percolation, leading to higher thermal conductivity of the composite. Organic solvents typically lead to efficient dispersion of graphene into the epoxy matrix. In this study, we compare the effect of two organic solvents (dimethylformamide (DMF) and acetone) in terms of their efficiency in dispersing graphene into the epoxy matrix and their effect on enhancing thermal conductivity of the composite. While the effect of solvents on mechanical properties of polymer-graphene nanocomposites has been studied, an understanding of their effect on thermal conductivity is lacking. In this study we find that polymer-graphene composites made with DMF show 44% higher thermal conductivity than those made using acetone. Laser scanning confocal microscopy (LSCM) imaging reveals improved dispersion of graphene-nanoplatelets in samples prepared using DMF compared to acetone. These results provide new avenues to achieve higher thermal conductivity graphene-epoxy composites.   

Role of Water Molecules on Vehicular Transport Mechanism in Polynorbornene-based Anion Exchange Membranes

Ion conducting polymer electrolytes are at the heart of fuel cells and water electrolyzers. The rise in anion exchange membranes (AEMs) research is attributed to the viability of using platinum-group-metal free electrocatalysts. However, the complex role of water molecules in facilitating ion transport is still in debate. Here, we investigated polynorbornene-based AEMs as our model polymers due to their high alkaline stabilities and outstanding fuel cell performance. We employ atomistic molecular dynamics (MD) simulations to decouple site-hopping and the vehicular mechanism by monitoring bromide ion transport. The conductivity-temperature relationship at different relative humidity (RH) conditions follows an Arrhenius behavior. We find that with increasing RH, the activation energy drops dramatically at around 55% RH and then remains constant. In addition to the common site-hopping mechanism, we use a quantitative model to show that the formation of water percolation can significantly change ion solvation and facilitate vehicular transport. To enable fast vehicular transport, both percolated water networks and dynamic water molecules are needed. We expect these new findings will provide useful guidance for the design of novel polymer membranes.   

Effect of Permanent Crosslinks and Temperatures on Network Relaxation and Penetrant Diffusion in Tight Polymer Networks

Selective separation of mixtures of organic molecules is particularly difficult when the components have similar molecular weight and intermolecular interactions. Tight flexible polymer networks are promising candidates for selective transport because of their strong selectivity to molecular size and shape and lower energy demand. However, there remains a need for a molecular-level understanding of network-based selectivity to design materials for separation applications. We performed Molecular Dynamics simulation of dilute penetrant diffusion in dense polymer networks with varying crosslink density. We found that glass transition temperatures (T_g) for networks shows a nearly linear relation with different crosslink densities, which compared well with measurements from experimental collaborators. To understand how segmental dynamics changes with crosslink densities and temperatures, we calculated the network relaxation time and found that the relaxation time dramatically increases as temperature gets close to T_g. We calculated the diffusion coefficient of penetrants in networks with varying crosslink density and temperature. The dependence between diffusion coefficient and temperature or size ratio of penetrant to network mesh size qualitatively agree with experimentally determined diffusivity of fluorescent dyes.   

Effect of Morphology and Segmental Dynamics on Ion Transport in Polymerized Lyotropic Liquid Crystals Containing Ionic Liquid

We investigate how the morphology and segmental dynamics affect the ion transport in polymerized lyotropic liquid crystals (polyLLCs) containing 1-butyl-3-methylimidazolium tetrafluoroborate ionic liquid (IL). We demonstrate that grain size and chain density at the interface are the two important factors that affect ion conduction in polyLLCs. The polyLLC with large grain size (70 nm) showed significant reduction in ion conductivity (one order of magnitude) compared to its homopolymer/IL mixture counterpart. However, the polyLLC with small grain size (20 nm) has little difference in ion conductivity compared to the corresponding homopolymer/IL mixture. It is observed that decreasing the chain density enhances the interaction of IL with polymer chains and consequently slows the relaxation of polymer chains. In addition, comparing the dynamics of polymer chains in mixtures of homopolymer/IL and templated LLC mesophases shows that the confinement in LLC structures prolongs the relaxation of polymer chains.   

Free Volume Enhances Catalyst Diffusion in Reactive Polymer Resins

Quantitative reaction-diffusion models are critical to the design of high-resolution lithographic processes based on acid-catalyzed deprotection of glassy polymer resins. Prior works show that catalyst diffusion is enhanced by the deprotection reaction; however, the underlying mechanisms remain unclear. Literature suggests that the generation of excess free volume due to the fast removal of reaction byproducts could be central to the apparent reaction-diffusion rates. We examine the validity of this mechanism in a model polymer using experiments and atomistic simulations. Extent of reaction was examined through infrared absorbance spectroscopy, and densification of reacting films was probed through spectroscopic ellipsometry. Comparison of these data indicates that most reaction occurs prior to full relaxation of the polymer. Atomistic simulations follow the effects of deprotection and subsequent volume relaxation on catalyst mobility. We establish that catalyst diffusion is enhanced in recently reacted regions and slowly relaxes to that observed in the absence of reaction. Estimated relaxation rates of excess free volume are similar to those of catalyst mobility, suggesting that excess free volume generation could impact deprotection kinetics.   

Diffusion of water in aqueous poly(ethylene oxide) at molecular to macroscopic length scales

Poly(ethylene oxide) (PEO) is a ubiquitously used water-soluble polymer with applications from drug delivery to water treatment membranes. Despite widespread use of PEO, models for water diffusion near the PEO surface and in the bulk lack molecular detail. We measure bulk diffusion coefficients of water in PEO solutions from 0 to 90 wt% PEO using Pulsed-field gradient NMR and local water diffusion within 1 nm of spin-labeled PEO by Overhauser Dynamic Nuclear Polarization (ODNP). We find that the bulk water diffusion coefficient scales with PEO concentration as described by a free volume model that accounts for interstitial space. Molecular dynamics simulations confirm that slower translational water dynamics correlates with enhanced tetrahedral water structure. ODNP experiments reveal that water diffusivity is slower near PEO surfaces than in bulk, but converges at the overlap concentration. PEO is not simply an obstruction to water transport but integrates into the water network to modify the structure and dynamics of the hydration shell.   

Roles of mesh size, segmental dynamic, glass transition temperature, and molecular structure on penetrant diffusion in dense vitrimers

We have systematically investigated how dynamic crosslinks can facilitate the transport of polyaromatic dye in the n-butyl acrylate networks. Two different boronic ester crosslinks with different bond exchange rates were compared to study the effect of bond exchange on penetrant diffusion. Diffusion of a large anisotropic fluorescent molecule was measured by fluorescence recovery after photobleaching (FRAP). The diffusion coefficients of the dye were normalized by mesh size and Tg/T, and in both cases diffusion is enhanced in the dynamic networks. Finally, we investigated the selective diffusion of two different dyes. These two penetrants have similar structure, but one of them could participate in reversible dynamic bonding with the dynamic crosslinks and the other could not. When the mesh size was smaller than the dye size, diffusion was abruptly restricted for the non-reactive dye, while the reactive dye was able to continue diffusion with a faster rate. Our results show that dynamic crosslinks can enhance penetrant diffusion and provide a route towards selective transport in polymer membranes.