The curing process of epoxy and amine mixtures was investigated from the viewpoint of molecular size of reactants via all-atom molecular dynamics simulation. Of four mixtures composed of either smaller or larger epoxy and amine molecules, the system of both smaller epoxy and amine reached the highest conversion at a given time, whereas the conversion for the combination of both larger epoxy and amine was the lowest. These can be explained in terms of their diffusion coefficients. In addition, the size effect on the reaction kinetics was more striking for epoxy than amine in the curing process. This was because epoxy reacted only once but amine did twice. When the primary amine reacted with epoxy, the resultant secondary amine was incorporated into the three-dimensional network. Hence, even if the primary amine had a higher mobility, the movement of the part after the reaction slowed down, making the further reaction to be the tertiary amine slow. Finally, it was shown that although the smaller epoxy reacted faster than the larger one in the mixed system of larger/smaller epoxy and amine, there was no difference in the reaction progress depending on the molecular size of amine.

In polymer melts, the interaction between segments are considered to be screened and the ideal Gaussian chain statistics is recovered. The experimental fact that linear viscoelasticity of unentangled polymers can be well described by the Rouse model is naively considered as due to this screening effect. Although various theoretical models are based on the screening effect and the screening effect is believed to be reasonable, the screening effect cannot be fully justified on a solid theoretical basis. In this work, we study the screening effect by utilizing a simple dumbbell type model. We perform simulations for dumbbell systems in which particles interact via the Gaussian soft-core potential. We show that, if the density of dumbbells is high, the Gaussian soft-core interaction is actually screened and the static structures are well described by the ideal model without the Gaussian soft-core interaction. We also show that the relaxation moduli of interacting dumbbell systems approximately coincide to those of the non-interacting dumbbell systems. In the low density systems, we observe the deviations from the ideal non-interacting systems. For example, the relaxation moduli become relatively broad. However,the relaxation moduli of such systems can be decomposed into the relaxation modes by the Gaussian soft-core interaction and the bond. The bond relaxation mode can be successfully described by a single Maxwell relaxation with effective relaxation strength and time. Our results support a naive use of the Rouse model to analyze unentangled polymer melts.

Primitive chain network simulations for randomly cross-linked slip-link networks were performed. For the percolated networks, the stress-strain relationship was compared to the theories by Ball et al. [Polymer, 22, 1010 (1981)] and Rubinstein and Panyukov [Macromolecules, 35, 6670 (2002)]. The simulation results are reasonably reproduced by both theories when the contributions from cross-links and slip-links are determined irrespective of the network structure. The fraction of slip-links contained in the network thus obtained is model dependent, and it does not coincide with the number of active slip-links in the actual networks. This discrepancy between the theories and the simulation results is due to the network inhomogeneity.

We evaluated three slip-spring dissipative particle dynamics models implemented in the general-purpose coarse-grained molecular dynamics program COGNAC. One fixed number of the slip-spring model proposed by Langoloth and co-workers and the two grand canonical models proposed by Unemaya and Masubuchi, and by Ramírez-Hernández and co-workers were implemented based on their original papers. Each implementation was validated by reproducing the original results. The details behind all the settings of the various parameters such as fictitious chemical potentials used in the grand canonical methods were studied. A linear relationship between the slip-spring density and exp(ν/k_BT) was confirmed for the Uneyama and Masubuchi algorithm, where ν is a fictitious chemical potential. Moreover, the coefficient for our implementation was determined to enable quantitative adjustments to the slip-spring density for each target application. With the slip-spring model being a promising model in the study of rheological properties of entangled polymer melts and solutions, this work contributes by advancing the application of the slip-spring model both in academia and industry.

We have investigated the flow properties of monodisperse and bidisperse entangled linear polymer melts in a contraction-expansion channel. Unlike conventional macroscopic approaches, where the transport equations are combined with a phenomenological constitutive equation for the stress originating from the polymer dynamics, we use a multiscale simulation (MSS) method, in which a macroscopic simulation model for solving the momentum balance equation is connected to a molecular-based mesoscopic model that can describe the entangled polymer dynamics. In the MSS method, we combine the smoothed particle hydrodynamics method with the dual slip-link model. In this study, we newly investigate the flow properties of a bidisperse entangled linear polymer melt since the polymer melts used in most industrial processes are not monodisperse. Compared with constitutive equations, the slip-link model employed in our MSS is effective at addressing the rheological properties of bidisperse entangled polymer melts. We especially focus on the flow properties of an entangled polymer melt containing a relatively small number of longer polymer chains. The MSS approach enables the relations between the macroscopic complex flows and microscopic states of bidisperse entangled polymer chains to be investigated. As a result of the MSSs, we observe that the flow rate for the bidisperse melt induced by an externally imposed pressure difference is clearly smaller than that for a monodisperse melt because of the high stress originating from the long chains.

This review focuses on the use of Bayesian Data Analysis and Machine Learning Techniques to study and analyze flow problems typical to polymer melt systems. We present a brief summary of Bayesian probability theory, and show how it can be used to solve the parameter estimation and model selection problems, for cases when the model(s) are known. For the more complex non-parametric regression problem, in which the functional form of the model is not known, we show how Machine-Learning (through Gaussian Processes) can be used to learn arbitrary functions from data. In particular, we show examples for solving steady-state flow problem as well as learning the constitutive relations of polymer flows with memory.

We investigated the strong thermal adhesive properties between thermoplastic polyamide elastomers (PAE) and other polymers (thermoplastic polyurethanes (TPU) and cross-linked polybutadienes (cBR)). The interfacial tensions between PAE and TPU or BR were evaluated from the dispersed particle size and rheological characteristics of slow relaxation mechanism for PAE/TPU and PAE/BR blends prepared by melt-blending. Although the values of interfacial tensions by both evaluation methods did not agree very well, it was found that the interfacial tensions of both polymer systems were considerably small. The good thermal adhesive properties between PAE and TPU or PAE and cBR are thought to be due to their low interfacial tensions and thick interfacial thickness of these polymer systems. However, in order to obtain sufficient adhesive strength for polymer systems with small interfacial tension, it is important to set molding conditions so that the interfacial thickness becomes sufficiently thick.

In this study, we investigate the sessile drop oscillation of aqueous PEO solutions after the transition from visco-capillary to elasto-capillary regime in the capillary thinning process of low-viscosity elastic fluids. Among various factors that affect oscillation, we focus on the effect of operating conditions that induce capillary thinning. The lifting velocity to lift the upper plate and the plate diameter to load the sample are selectively controlled. The generated oscillation is quantitatively analyzed in terms of oscillation frequency and decay ratio. The oscillation frequency and oscillation decay ratio are found to be independent of the lifting velocity. The oscillation frequency of all samples is about 140 Hz independent of the lifting velocity, and the decay ratio is not affected by the lifting velocity. On the contrary, the plate diameter affects the oscillation. The oscillation frequency increases to 210 Hz and the decay ratio shows a larger value (close to one) when a smaller plate is used. This study opens a way to improve the measurement of extensional properties which suffers from oscillation.