<p> This thesis is based on the studies of cross-linked network polymerization by measurements of dielectric permittivity and a study of the polymers formed by dynamic mechanical analysis. The network polymers are formed by reacting a trifunctional epoxy with three primary amines, namely, aniline, 3-chloroaniline and 4-chloroaniline. The manner in which the dielectric permittivity changes on polymerization was followed from the monomeric to the ultimately polymeric state, and the relaxation time and the distribution of reaction times was followed with reaction time. Isothermal measurements of the heat evolved during the reaction led to a determination of the number of covalent bonds formed at any instant of reaction time, thus the dielectric relaxation time was related to the number of covalent bonds formed during polymerization. The relaxation time increased progressively more rapidly with the increase in the number of covalent bonds formed under isothermal conditions as the reaction progressed towards completion. As the number of covalent bonds increased, a high-frequency, or low-temperature, relaxation process emerged. This was observed for both the time-variant (fixed frequency) and time-invariant (dielectric spectra) dielectric data.</p> <p> The extent of reaction at the gelation time deduced from the irreversible decrease in the dc conductivity on polymerization was 50% lower than anticipated from Flory-Stockmayer theory, but, the extent of reaction at the point of singularity agreed with the theory. As polymerization reactions occurred, both the static and high-frequency dielectric permittivities decreased with time. Data simulated for the dielectric behaviour for a fixed frequency during the course of polymerization were analyzed to confirm that the two analysis procedures (time-invariant and time-variant) yield the same parameters within 2-3%.</p> <p> The three, new, network polymers thus produced had unrelaxed and relaxed moduli that were in the ranges of 1.35-1.51GPa and 6.6-10.6MPa, respectively. GU of the three polymers decreased with increasing temperature and GR increased. The former effect has been attributed to mainly a decrease in the GR of the low-temperature, or high-frequency, β-relaxation process, and the latter to the increase in the temperature, and consequently, the increase in the entropy, as discussed in the entropy theories of the rubber modulus. The spectrum of the shear modulus cannot be superposed by shifting the spectra along the frequency scale alone. The three reasons discussed are; (1) GU and GR change with temperature, therefore the spectra differ in magnitude for each measurement temperature, (2) the presence of a temperature-dependent, frequency-independent background loss changes the magnitude of tan∅ with the temperature, and (3) the influence of secondary-, β-, or Johari-Goldstein relaxations causes deviations in the shape of the spectrum such that they cannot be superposed. It was determined that none of the three effects described contribute significantly to the changes in the isothermal spectra, and hence, the principle of the time-temperature superposition does not apply to the three polymers in this study. The three polymers produced have a mechanical loss peak at room temperature that is attributed to the β-relaxation process and whose magnitude is greater than that of polycarbonates. This suggests that the three new polymers produced here can absorb more energy at room temperature than the polycarbonates.</p> / Thesis / Master of Engineering (MEngr)

High density polyethylene (HDPE) films having simple and well-defined stacked lamellar morphology, either with or without a distinct presence of row-nucleated fibril structures, have been utilized as <i>model</i> materials to carry out investigations on solid state structure-property relationships. Four different subjects that were addressed are: 1) mechanical properties and deformation morphologies, 2) orientation anisotropy of the dynamic mechanical α relaxation, 3) orientation dependence of creep behavior, and 4) crystalline lamellar thickness and its distribution. For the first three topics, appropriate mechanical tests, including tensile (INSTRON), creep (TMA), and dynamic mechanical (DMTA) tests, were performed at <i>different angles with respect to the original machine direction (MD)</i> of the melt extruded films; morphological changes as a result of these mechanical tests were detected by WAXS, SAXS, and TEM. For the forth topic, crystalline lamellar thickness and its distribution were determined by DSC, SAXS, TEM and AFM experiments. In the <i>large strain deformation</i> study (chapter 4.0), samples were stretched at 00°, 45° and 90° angles with respect to the original MD. A distinct orientation dependence of the tensile behavior was observed and <i>correlated</i> to the corresponding deformation modes and morphological changes, namely 1) lamellar separation and fragmentation by chain slip for the 00° stretch, 2) lamellar break-up via chain pull-out for the 90° stretch, and 3) lamellar shear, rotation and break-up through chain slip and/or tilt for the 45° stretch. A strong strengthening effect was observed for samples with row-nucleated fibril structures at the 00° stretch; whereas for the 90° stretch, the presence of such structures significantly limited deformability of the samples. In the <i>dynamic strain mechanical α relaxation</i> study (chapter 5.0), samples were tested at nine different angles with respect to the original MD, and the morphologies of samples <i>before</i> and </i>after</i> the dynamic tests were also investigated. The mechanical dispersions for the 00° and 90° tests were believed to arise essentially from the crystalline phase, and they contain contributions from two earlier recognized sub-relaxations of α_I and α_II. While for the 45° test, in addition to a high temperature α_II relaxation, a interlamellar shear induced low temperature mechanical relaxation was also observed. It is concluded that the low temperature relaxation is related to the characteristics of the interface between the crystalline lamellae and amorphous layers. In the <i>small strain creep</i> study (chapter 6.0), samples were tested at the 00°, 45° and 90° angles at the original MD. Both creep strain and creep rate for samples at the three angles were very different. An Eyring-rate model was utilized to analysis the observed creep behavior, and structural parameters associated with this model, including population of creep sites, activation energy and volume, were obtained by fitting the experimental data to the Eyring-rate equation. It was concluded that the plateau creep rate in these model materials is primarily controlled by the density and physical state of tie-chains in the amorphous phase. For the lamellar thickness and distribution study, DSC, SAXS, TEM and AFM experiments were conducted for samples having a well-defined stacked lamellar morphology. It was found that the most probable lamellar thickness from SAXS and TEM agreed very well; however, these values did not match with those obtained by DSC and AFM. It was pointed out that the use of DSC to determine lamellar thickness and distribution is so sensitive to heating rate and numerical values for the parameters in the Gibbs-Thomson equation that it is not believed to be suitable for quantitative analysis. / Ph. D.

La majorité des adhésifs utilisés dans l’industrie marine sont des polymères avec un comportement mécanique qui est fortement influencé par les conditions environnementales (vieillissement hydrique ou température). Par conséquent, il est très important pour les ingénieurs travaillant dans des bureaux d’études d’être capable de prendre en compte ces effets lors des différentes étapes de développement et conception des assemblages collés.Le présent travail propose une méthode d’analyse de l’influence du vieillissement hydrique sur le comportement mécanique d’un adhésif structural époxy dans un assemblage collé. Tout d’abord, un modèle viscoélastique-viscoplastique a été développé pour caractériser la réponse mécanique de l’adhésif dans un joint de colle. Pour cela, le dispositif expérimentalArcan a été utilisé. Le modèle est identifié en utilisant la méthode d’identification inverse et les échantillons sont testés à l’état non-vieilli (pas de vieillissement hydrique). Les résultats obtenus après la démarche d’identification sont utilisés pour prédire le comportement mécanique d’éprouvettes massiques.Dans un deuxième temps, afin de diminuer les temps de saturation d’échantillons, l’évolution des propriétés mécaniques de l’adhésif est analysée sous différentes conditions de vieillissement hydrique (immersion dans l’eau de mer et en humidité relative contrôlée) grâce à des essais sur éprouvettes massiques.Les résultats obtenus seront utilisés pour identifier l’évolution de chaque paramètre du modèle proposé, en fonction de la quantité d’eau absorbée. En parallèle, un modèle de diffusion a été développé pour caractériser le gradient de teneur en eau des joints de colle. Les deux approches sont ensuite combinées pour modéliser les profils d’eau pour différents temps de vieillissement et prédire l’évolution des propriétés mécaniques du joint de colle après le vieillissement. Finalement, pour valider la méthode proposée, la prédiction du modèle est comparée avec des essais réalisés sur assemblages collés vieillis. / Most of the adhesives used in the marine industry are polymers with a mechanical behaviour which is strongly influenced by environmental conditions (water activity or temperature). Therefore, it is important for engineers and designers to be able to consider these effects during the different stages of development and manufacturing of a bonded structure.The present work presents a method for analyzing the influence of water ageing on the behaviour of an epoxy adhesive in an adhesively bonded assembly.First, a viscoelastic-viscoplastic model is developed to characterise the mechanical response of the adhesive at initial state in a bonded joint using the modified Arcan device. The model is identified using the inverse identification method and the considered samples are tested at an unaged stage (no water activity). The results obtained after the identification process are used to predict the bulk behaviour of the adhesive. A comparison between numerical results and experimental tests realised on bulk specimens is then made in order to validate this first approach.In a second phase, in order to decrease the times for samples saturation, the evolution of the mechanical properties of the adhesive in bulk form is tested under different water ageing conditions (immersion in seawater and different relative humidity). The obtained results allowed to identify the evolution of the model parameters as a function of water content. In parallel, a diffusion model was developed to characterise the water ingress in the bonded joint. These two approaches are then combined to model the water profiles and to consider the evolution of mechanical properties of a water aged adhesively bonded assembly, for different immersion times. Finally, to validate the framework, the prediction is compared with experimental tests performed on aged specimens.

No description available.