Please use this identifier to cite or link to this item:
|Title:||Molecular dynamics investigation of viscoelastic properties of polymer melts|
|Abstract:||Polymers and polymer mixtures play such important roles in our lives that it is hard to imagine life without them. Although a lot of progress has been made in the past few decades in our understanding of polymer dynamics and rheology using theoretical, computational, and experimental approaches, there are significant gaps in what is still left to be done. For example, polydispersity is the norm in industrially-produced polymers. However, a lot of the theories that have been propounded are predicated on monodisperse polymers. Most recently, molecular simulations have significantly improved our understanding of polymer dynamics by either validating theories or offering novel insights into their dynamics at the micro-structure level of detail. This thesis also uses molecular simulation to explore three different classes of polymer systems – monodisperse, polydisperse, and polymer-additive mixtures. Two of the most significant theories of polymer dynamics are the Rouse model describing the dynamics of short and unentangled chains, and the tube and reptation model, which describes the motion of long, entangled chains. The reptation model predicts different dynamical regimes that are marked by distinct time scales and further introduced the concept of an entanglement length Ne – which signifies the length scale at which topological interactions between a test chain and surrounding matrix chains become significant. These distinct time scales have been investigated by various researchers, but the reported values vary with different groups. Here, we devised a protocol for the accurate determination of these time scales. We also calculated Ne using these time scales and compared our results with those reported in the literature. Furthermore, using Rouse mode analysis – a technique that resolves the coupled motion of monomers into distinct and uncoupled modes, we showed that the method can be used to determine Ne, with the values obtained closer to that using monomer displacement at longer time scales. The computational calculation of the linear viscoelastic (LVE) properties of polymers is very expensive due to the long relaxation times associated with polymers. Using three different methods – Equilibrium molecular dynamics (EMD) via the Green-Kubo relation, non-equilibrium molecular dynamics (NEMD), and corrected Rouse Mode Analysis (cRMA), we determined the LVE properties of polymer melts for both unentangled and entangled chains and compared the uncertainty associated with each of them. Specifically, we demonstrate that the cRMA method, although applicable to only the shorter chains, yielded the lowest uncertainty. Compared to earlier reported results, we also show that, using shorter computational runs, the NEMD gives an acceptable level of accuracy in the calculation of the LVE properties. Beyond methodology development, this thesis also studied the viscoelasticity of realistic polymer systems of practical interest, including polydisperse polymers and polymer-additive mixtures. We idealized a polydisperse system using a bidisperse model wherein we combined longer chains with shorter chains at different concentration levels. By investigating the individual motion of the different chains, we showed that the dynamics of the longer chains are sped up by the shorter chains, whereas the longer chains impeded the dynamics of the shorter chains, although the nature of the dynamics of the shorter chains was not altered. We also showed that the shorter chains reduce the extent of the entanglement effect of the longer chains using an RMA approach. By using the individual LVE profiles of the long and short chains, we tested a semi-empirical mixing rule for predicting the stress relaxation modulus G(t) of the bulk mixture. We found that a simple mixing rule works well when the shorter chains are the majority while the double reptation model – that assumes a simultaneous relaxation of both the test chains and surrounding matrix chains, does a better job of the prediction when the longer chains are the majority. We further explored the effect of molecular structure of single-bead additives on the thermo-physical properties, such as the glass transition temperature and Young’s modulus on the polymer-additive mixture. By varying the dimensions of the molecular additive over a range of concentration of the molecular additives, we found that smaller-sized additives are better able to reduce the glass transition temperature Tg and increase the Young’s modulus Y of the mixture due to the improved packing efficiency. On the other hand, larger-sized particles are only marginally able to reduce the Tg. The LVE properties, specifically the zero-shear viscosity of the mixtures showed an opposite trend to the Y, where once again the smaller-sized particles better reduce the zero-shear viscosity. This essentially shows the decorrelation of traditional plasticization markers. A reduction in one property does not imply a reduction in another property and this varies with the dimensions and concentration of the additives. We finally describe our various attempts at developing a multi-bead plasticizer model. For the models we have tried, we tuned the chain lengths of the beads and their interaction with the polymer. Detailed micro-structure analysis and viscoelasticity calculations reveal that they are either incompatible with the polymer, resulting in phase separation or only marginally compatible over a very limited range.|
|Appears in Collections:||Open Access Dissertations and Theses|
Files in This Item:
|Adeyemi_Oluseye_B_2022April_PhD.pdf||14.67 MB||Adobe PDF||View/Open|
Items in MacSphere are protected by copyright, with all rights reserved, unless otherwise indicated.