The importance of knowing what type of traffic is flowing through a network is paramount to its success. Traffic shaping, Quality of Service, identifying critical business applications, Intrusion Detection Systems, as well as network administra- tion activities all require the base knowledge of what traffic is flowing over a network before any further steps can be taken. With SSL traffic on the rise due to applica- tions securing or concealing their traffic, the ability to determine what applications are running within a network is getting more and more difficult. Traditional methods of traffic classification through port numbers or deep packet inspection have been deemed inadequate by researchers thus making way for new methods. The purpose of this thesis is to investigate if a machine learning approach can be used with flow features to identify SSL in a given network trace. To this end, different machine learning methods are investigated without the use of port numbers, Internet Protocol addresses, or payload information. Various machine learning models are investigated including AdaBoost, Naive Bayes, RIPPER, and C4.5. The robustness of the results are tested against unseen datasets during training. Moreover, the proposed approach is compared to the Wireshark traffic analysis tool. Results show that the proposed ap- proach is very promising in identifying SSL traffic from a given network trace without using port numbers, Internet protocol addresses, or payload information.

This thesis research focuses on fundamental understanding regarding the morphological transitions of weak polyelectrolyte multilayers (PEMs) formed by the layer-by-layer (LbL) electrostatic assembly of oppositely charged polymers. he first part of this thesis focuses on patterning polyelectrolyte multilayers that are able to undergo transitions from continuous films to porous materials by using hydrogel stamps. The stamping process is able to locally etch and pattern the porous transition in the LbL films by using reactive wet stamping (r-WETS). It was found that r-WETS of PEMs can also enable the modification of chemical functionality. The second part is an investigation about morphological changes of weak polyelectrolyte multilayers assembled with PAH and PAA using r-WETS in which hydrogel stamp material was soaked into various salt solutions and then applied to the LbL films. Also, in this study we presented a novel strategy to create a continuous gradient structure in thickness or porosity along the lateral direction of the thin films using concentration gradient salt stamping. The third part is an investigation regarding the mechanism of the transition from a continuous morphology to a porous morphology within weak polyelectrolyte multilayers. These morphological changes were able to be created by both acidic and basic post-assembly treatments, showing various morphological transitions from the introduction of porosity to the collapse of these porous structures and the eventual dissolution of the films. A similar observation of morphological transitions in weak polyelectrolyte multilayers was obtained by applying an electric field to the films in the fourth part of this thesis. Exposure to an electric field resulted in the creation of a porous structure, which can be ascribed to local changes in pH and subsequent structural rearrangements of the weak polyelectrolyte constituents. The final part of this thesis is to make PEMs into nanostructured matrices for inorganic synthesis. Multilayers possessing ion-exchangeable carboxylic acid groups were used for binding metal catalysts such as platinum (Pt) nanoparticles (NPs) within the film. Therefore, polyelectrolyte multilayers were able to stabilize catalytic Pt NPs in order to increase the useful time of catalyst materials suitable for use in proton exchange membrane fuel cells.

This thesis aims to clarify the sources of, and mechanisms associated with, the generation of ozone maxima over the southern African region. Inasmuch as, tropospheric ozone concentration is a function of both chemistry and meteorology, this thesis concentrates on the role of atmospheric dynamics. Firstly, a statistical analysis of the relationship between total ozone and meteorological parameters revealed a generally weak negative relationship between total ozone and the height of the 500, 300 and 100 hPa geopotential surfaces. The relationship is best expressed by the passage of a mid-latitude cyclone while anticyclonic conditions exhibited a weak relationship. An examination of the spatial distribution of total ozone and potential vorticity (PV), during the passage of westerly troughs, prompted a more thorough investigation of the exchange of ozone between the stratosphere and troposphere. The relationship between tropospheric ozone, and low pressure and anticyclonic systems is investigated further using data obtained during the South African Fire-Atmospheric Research Initiative (SAFARI) conducted in 1992. Ozone concentrations, as expressed by ozonesonde data, reveal different characteristic profiles for the two scenarios. Explanations for the differences observed are sought in the observed circulation patterns during the experiment. Case studies at Okaukuejo (Namibia), Irene (South Africa) and Brazzaville (Congo), which were utilised as ground stations during SAFARI, are presented in an attempt to gain insight into the vertical distribution of ozone over the entire expanse of the study region. The role of convective systems in the generation of short-lived upper tropospheric ozone maxima at tropical latitudes is illustrated while the different vertical ozone signatures, expressed under cyclonic and anticyclonic systems as described earlier, are reconfirmed by the Okaukuejo and Irene data. An attempt is made to investigate dynamic links between the troposphere and stratosphere and the concomitant exchange of ozone during the passage of westerly trough systems. Particularly deep troughs or cut off low pressure systems are identified as important mechanisms in the generation of upper tropospheric ozone maxima. An examination of the vertical distribution of ozone at Irene during the passage of a COL, using data obtained from the SA'ARI 1994 experiment, suggests concurrence with Danielsen's (1968) model of tropopause folding. The intrusion of high PV and dry stratospheric air, coupled with downward flow near the tropopause, in the vicinity of the upper tropospheric disturbance, promotes the transport of ozone-rich air to tropospheric altitudes. The limited availability of data has severely hampered the understanding of tropospheric ozo~e in southern Africa in the past. This study demonstrates the value of daily vertical ozone data, even for very short periods. / Thesis (Ph.D.)-University of Natal, 1995.

Measurements of surface ozone in Durban have been undertaken since September 1994 until August 1995 using a Dasibi ozone monitor model 1108. There is a seasonal variation in surface ozone with a small amplitude of 6.5 ppb. The mean maximum of 12.5 ppb is experienced in August and the minimum in December (6 ppb). The late winter peak and the sharp rise from June to July is possibly related to the greater accumulation of ozone at the surface due to the poor dispersion, higher frequency of inversions and lower mixing depth. The study also focuses on the relationship between synoptic weather systems and daily ozone concentrations in an attempt to establish if any relationship exists. The analysis suggests that there is an increase in ozone concentration during the pre-frontal and established high pressure systems and a reduction in ozone under postfrontal conditions. The mean diurnal variation in surface ozone depicts a maximum between solar noon and 14:00, which is typical of an urban-industrial environment, in which ozone precursor gases have built up during the morning. However, a secondary peak in the early morning during winter and autumn was more difficult to explain and is thought to be due to the transport of ozone from the interior in the mountain-plain wind systems. Comparison of ozone values at Durban with results available from the other locations in South Africa such as Cape Point and the Eastern Transvaal Highveld has been undertaken. Ozone values at Durban are lower than Cape Point, which is a representative of a background monitoring station. It ls hypothesised that those low ozone concentrations may be due to high NOx which in turn depletes ozone. / Thesis (M.Sc.)-University of Natal, 1995.

While considerable progress has been made towards understanding the effect that membrane-based layer-by-layer (LbL) immobilizations have on the activity and stability of enzymatic catalysis, detailed work is required in order to fundamentally quantify and optimize the functionalization and operating conditions that define these properties. This work aims to probe deeper into the nature of transport mechanisms by use of pressure-induced, flow-driven enzymatic catalysis of LbL-functionalized hydrophilized poly(vinyldiene) (PVDF)-poly(acrylic acid) (PAA)-poly(allylamine hydrochloride) (PAH)-glucose oxidase (GOx) membranes. These membranes were coupled in a sealed series following cellulose acetate (CA) membranes for the elimination of product accumulation within the feed-side solution during operation. At pH = 6 and T = 21oC, the enzymatic catalysis of LbL-immobilized GOx from Aspergillus niger performed remarkably well in comparison to the homogeneous-phase catalysis within an analogous aqueous solution. On average, the enzymatic turnover was 0.0123 and 0.0076 mmol/(mg-GOx)(min) for the homogeneous-phase catalysis and the LbL-immobilized catalysis, respectively. Multiple consecutive permeations resulted in replicable observed kinetic results with R2 > 0.95. Permeations taking place over the course of a three week trial period resulted in a retention of >90% normalized activity when membranes were removed when not in use and stored at -20oC, whereas the homogenous-phase kinetics dropped below 90% normalized activity in under one day.

The structural vibration damping behaviour of plates and beams can be improved by the application of viscoelastic passive damping materials. Unconstrained layer damping treatments applied to metal plate systems were studied experimentally. Design and modelling of novel fibre reinforced constrained layer damping materials was performed, and implementation of these composite damping materials into laminated composite sandwich constructions commonly used as structural elements within large composite marine vessels was explored. These studies established effective methods for examining, designing and applying damping materials to metal and composite marine structures. Two test fixtures were designed and constructed to facilitate testing of viscoelastic material damping properties to ISO 6721-3 and ASTM E756. Values of material damping made in accordance with ASTM E756 over a range of temperatures were compared to values produced by a Dynamic Mechanical Analyser (DMA). Glass transition temperatures and peak damping values were found to agree well, although results deviated significantly at temperatures above the glass transition temperature. The relative influence of damping layer thickness, ambient temperature, edge conditions, plate dimensions and substrate material on the system damping performance of metal plates treated with an unconstrained viscoelastic layer was investigated experimentally. This investigation found that substrate material had the greatest influence on system damping performance, followed by damping layer thickness and plate size. Plate edge conditions were found to have little influence on the measured system damping performance. These results were dependent on the values of each variable used in the study. Modal damping behaviour of a novel fibre reinforced composite constrained layer damping material was investigated using finite element analysis and experimental methods. The material consisted of two carbon fibre reinforced polymer (CFRP) layers surrounding a viscoelastic core. Opposing complex sinusoidal fibre patterns in the CFRP face sheets were used to achieve stress-coupling by way of orthotropic anisotopy about the core. A finite element model was developed in MATLAB to determine the modal damping, displacement, stress, and strain behaviour of these complex patterned fibre constrained layer damping (CPF-CLD) materials. This model was validated using experimental results produced by modal damping measurements on CPF-CLD beam test specimens. Studies of multiple fibre pattern arrangements found that fibre pattern properties and the resulting localised material property distributions influenced modal damping performance. Inclusion of CPF-CLD materials in laminated composite sandwich geometries commonly used in marine hull and bulkhead constructions was studied experimentally. Composite sandwich beam test specimens were fabricated using materials and techniques frequently used in industry. It was found that the greatest increases in modal damping performance were achieved when the CPF-CLD materials were applied to bulkhead geometries, and were inserted within the sandwich structure, rather than being attached to the surface.

The goal of the proposed study is to understand the morphology, physical, and responsive properties of synthetic polymer and biopolymer layer-by-layer (LbL) arrays using the inkjet printing and stamping technique, in order to develop patterned encapsulated thin films for controlled release and biosensor applications. In this study, we propose facile fabrication processes of hydrogen-bonded and electrostatic LbL microscopic dot arrays with encapsulated target organic and cell compounds. We study encapsulation with the controllable release and diffusion properties ofpoly(vinylpyrrolidone) (PVPON), poly(methacrylic acid) (PMAA), silk-polylysine, silk-polyglutamic acid, pure silk films, and E-coli cells from the multi-printing process. Specifically, we investigate the effect of thickness, the number of bilayers, and the hydrophobicity of substrates on the properties of inkjet/stamping multilayer films such as structural stability, responsiveness, encapsulation efficiency, and biosensing properties. We suggest that a more thorough understanding of the LbL assembly using inkjet printing and stamping techniques can lead to the development of encapsulation technology with no limitations on either the concentration of loading, or the chemical and physical properties of the encapsulated materials. In addition, this study offers new encapsulation concepts with simple, cost effective, highly scalable, living cell-friendly, and controllable patterning properties.

Layer-by-Layer (LbL) deposition of electrochemically active materials on porous carbon electrodes is a proven method to leverage both electrochemical double-layer capacitance and pseudocapacitance for charge storage on the same electrode. LbL coatings are held together by electrostatic attraction between adjacent layers of oppositely charged molecules. Previous studies have used Keggin polyoxometalates to great effect as the anionic layer in LbL electrode modification, but little effort has been devoted to cationic material selection and LbL process optimization. This work investigated alternatives to the conventional, electrochemically inert polydiallyldimethylammonium (PDDA) cation. The use of fuchsin molecular cations in LbL deposition improved the specific energy and specific power of modified electrodes. Fuchsin cation also rendered the environmentally harmful oxidative surface activation step unnecessary for LbL deposition. Process parameters were optimized for MWCNT/Fuchsin/POM samples, and post-LbL electrochemical polymerization was found to further improve the performance of these electrodes.

A single polymer electrolyte membrane fuel cell (PEMFC) is comprised of several sub-millimetre thick layers of varying porosity sandwiched together. The thickness of each layer, which typically ranges from 10 to 200μm, is kept small in order to minimize the transport resistance of heat, mass, electrons, and protons, that limit reaction rate. However, the thickness of these materials presents a significant challenge to engineers characterizing the transport properties through them, which is of considerable importance to the development and optimization of fuel cells. The objective of this research is to address the challenges associated with measuring the heat conduction and gas diffusion transport properties of thin porous media used in PEMFCs. An improvement in the accuracy of the guarded heat flow technique for measuring thermal conductivity and the modified Loschmidt Cell technique for measuring gas diffusivity are presented for porous media with a sub-millimetre thickness. The improvement in accuracy is achieved by analyzing parameters in each apparatus that are sensitive to measurement error and have the largest contribution to measurement uncertainty, and then developing ways to minimize the error. The experimental apparatuses are used to investigate the transport properties of the gas diffusion layer (GDL) and the microporous layer (MPL), while the methods would also be useful in the study of the catalyst layer (CL). Gas diffusion through porous media is critical for the high current density operation of a PEMFC, where the electrochemical reaction becomes rate-limited by the diffusive flux of reactants reaching reaction sites. However, geometric models that predict diffusivity of the GDL have been identified as inaccurate in current literature. Experimental results give a better estimate of diffusivity, but published works to date have been limited by high measurement uncertainty. In this thesis, the effective diffusivity of various GDLs are measured using a modified Loschmidt cell and the relative differences between GDLs are explained using scanning electron microscopy and the method of standard porosimetry. The experimental results from this study and others in current literature are used to develop a generalized correlation for predicting diffusivity as a function of porosity in the through-plane direction of a GDL. The thermal conductivity and contact resistance of porous media are important for accurate thermal analysis of a fuel cell, especially at high current densities where the heat flux becomes large. In this thesis, the effective through-plane thermal conductivity and contact resistance of the GDL and MPL are measured. GDL samples with and without a MPL and coated with 30%-wt. PTFE are measured using the guarded steady-state heat flow technique described in the ASTM standard E 1225-04. Thermal contact resistance of the MPL with the iron clamping surface was found to be negligible, owing to the high surface contact area. Thermal conductivity and thickness of the MPL remained constant for compression pressures up to 15bar at 0.30W/m°K and 55μm, respectively. The thermal conductivity of the GDL substrate containing 30%−wt. PTFE varied from 0.30 to 0.56W/m°K as compression was increased from 4 to 15bar. As a result, the GDL contain- ing MPL had a lower effective thermal conductivity at high compression than the GDL without MPL. At low compression, differences were negligible. The constant thickness of the MPL suggests that the porosity, as well as heat and mass transport properties, remain independent of the inhomogeneous compression by the bipolar plate. Despite the low effective thermal conductivity of the MPL, thermal performance of the GDL can be improved by exploiting the excellent surface contact resistance of the MPL while minimizing its thickness.

Recent advances in multilayer self-assembly have enabled the precise construction of nanocomposite ultrathin films on a variety of substrates, from large-area planar surfaces to nanoparticles. As a result, a wide range of physico-chemical properties may be represented by selecting from an array of surface preparations, molecules, assembly conditions, and post-assembly treatments. Such multilayer nanofilm assemblies are particularly attractive for use as specialized membranes for selective transport, which have many applications for separations, sensors, and drug delivery systems. In this work, nanocomposite ultrathin films built with layer-by-layer (LbL) self-assembly methods have been applied to surface modification to control interfacial behavior, including diffusion, anti-fouling, and biomimetic membranes. Transport and interfacial properties of nanocomposite membranes constructed using LbL self-assembly with synthetic and/or bio-polymers were characterized, and permeability values of clinically relevant small molecules through the nanofilms were determined. Correlations between permeability and film properties were also examined. Nanofilm coatings around 100nm thickness decreased diffusion coefficients of glucose up to five orders of magnitude, and were found to greatly affect enzymatic glucose sensor responses. Surface modification on top of the nanofilms with poly(ethylene glycol) provided anti-fouling effects. However, weak-weak polyelectrolyte multilayers (PEMs) should not be used to control transport due to their susceptibility under normal physiological conditions. Natural/biological polymers also provided multilayer film structures at the specific conditions, but their transport-limiting properties were not significant compared to synthetic PEMs. Even when covalently crosslinked, biological PEMs did not reduce the permeability of a small molecule. Finally, the predicting model of projecting analyte permeation through multi-phase nanocomposite films comprised with known diffusion coefficients was theoretically and experimentally evaluated. The modeling was matched reasonably well to experimental data. The outcomes will be the key knowledge or engineering principles to support future efforts in research and development. It is anticipated that the system developed for determining transport properties will provide a general platform for assessing new candidate materials. The theory developed will be useful in estimating transport properties of novel nanocomposite materials that may be interesting in a broad array of chemical and biological systems, from analytical separations to implantable biomedical applications, and will provide useful design rules for materials and fabrication process selection.