<p>In the hydrogenolysis of hydrocarbons, iron is unique among the transition metals since it breaks molecules extensively and produces large quantities of methane even at low conversion. Hydrogenolysis of propane, isopentane and n-hexadecane were studied at 325 and 355ºC over reduced Fe and carbided Fe catalysts prepared from a commercial NH₃ synthesis catalyst. Reaction network analysis applied to the selectivity data yielded a quantitative estimate of the relative importance, for an adsorbed species, of desorbing or undergoing C-C bond splitting; this method indicated that product desorption is the slow step in the reaction mechanism. Power rate expressions were obtained for hydrogenolysis of propane and isopentane at 325ºC. Used catalysts, characterized by Mössbauer spectroscopy, contained α-Fe at high feed ratios of H₂ to hydrocarbon, and Fe₃C, at low ratios. The formation of bulk iron carbides led to changes in selectivity, yielding less methane and more intermediate hydrocarbons. In hydrogenolysis of n-hexadecane, the analysis of the C₇-C₁₅ products by gas chromatography-mass spectrometry showed the presence of alkylbenzenes, alkenes and branched alkanes, which account altogether for as much as 20 mole % of the heavier hydrocarbons. The data also suggest that the breaking of paraffins on iron possibly occurs by rapid demethylation. Hydrogenolysis of hydrocarbons is completely inhibited when carbon monoxide is added to the feed, even at concentration as low as 4 mole %; CO is presumably more strongly adsorbed on the catalyst than hydrocarbons. The products obtained were similar to those found in the Fischer-Tropsch synthesis. Traces of water in the feed reduce the rate of hydrogenolysis. An increase in the K₂O content of Fe catalyst decreases its activity but does not affect its selectivity. A new method of deriving selectivity equations from reaction networks is presented. Network analysis is shown to apply to irreversible reactions for hydrogenolysis and isomerization.</p> / Doctor of Philosophy (PhD)

<p>This thesis reports on the synthesis, purification and anionic polymerization of several derivatives of styrene sulfonic acid and on the block copolymerization of propyl-p-styrene sulfonate with isoprene and styrene. The domain morphology of polyisoprene:polypropyl-p-styrene sulfonate block copolymers and a one to one blend of the corresponding homopolymers are defined, and blood platelet reactivity with the block copolymers, the respective homopolymers and the blend is investigated.</p> <p>p-Styrene sulfonamide terminates the anionic initiator (sodium naphthalene) probably by amidolysis. Methyl- and isopropyl-p-styrene sulfonates polymerize but proper purification is extremely difficult. Propyl-p-styrene sulfonate polymerizes, but propagation is halted without termination of the anions after only a relatively low molecular weight polymer is formed. Possible reasons for the formation of these apparently dormant anions is considered in some detail.</p> <p>The sulfonate-styrene-sulfonate and a broad range of sulfonate-isoprene-sulfonate triblock copolymers were prepared with a range of phase morphology features with dimensions of the order of 100Å. They and the respective homopolymers are all found to induce moderate and indistinguishable in vitro platelet responses as determined by radioactive labeling of platelets and by electron microscopy. In contrast, however, the blend of homopolymers which has phase morphology features with dimensions of the order of magnitude of platelet dimensions is shown to induce nonrandom clustering of adherent platelets. It is possible that morphologies with these dimensions may influence platelet thrombus propagation over the surface.</p>

<p>The population balance model of Kiparissides (1978) for the continuous emulsion polymerization of vinyl acetate, has been extended to predict molecular weight moments and has been corrected for induction time and particle shrinkage due to density changes. The developed model was successfully able to simulate experimental results from Kiparissides (1978) and Greene et al.(1976) for the continuous emulsion polymerization of vinyl acetate. Application of the model to the batch data of Keung (1974) had reasonable access as well. The model was extended to the continuous emulsion polymerization of styrene which follows different nucleation kinetics and was able to predict the average conversions and particle diameters for the data of Brooks et al.(1978). Little success was achieved in predicting the small experimentally observed oscillations for the styrene system.</p> <p>To eliminate the sustained property oscillations in the vinyl acetate system Linear-Quadratic stochastic optimal control theory was applied to the model. Due to extreme non-linearities inherent in the system, this approach was shown to be inadequate. Instead, the reaction system was redesigned to include a small continuous seeding reactor with monomer and water bypass. Model predictions indicated that the redesigned system eliminated the oscillations. Experimental testing of the redesigned system verified that a dramatic improvement in stability was possible. The redesigned reactor configuration was also shown to be more flexible in controlling particle sizes and conversion through use of the bypass.</p> <p>A method for selecting the optimum sensors was developed. It was shown that the measurement combination providing the most information would be conversion (as currently available from an on-line density meter) and some measure of weight average molecular weight.</p> / Doctor of Philosophy (PhD)

<p>A general purpose finite element program has been developed for numerical modelling of flow of polymer melts through processing equipment. Several problems have been solved for two-dimensional creeping flow. These include flow in the gap between calender rolls, driven cavity flows and planar entry and exit flows with free surfaces (extrudate swell) in slit dies. Newtonian, power-law and viscoelastic models have been used under isothermal and nonisothermal conditions. Slip at the boundary walls has also been incorporated. An "upwind" scheme for triangular elements has been used to stabilize the solution when convection is considered in the nonisothermal analysis. The free surfaces are determined through an iterative technique that allows no flow normal to the surface.</p> <p>The finite element analysis of calendering shows that the power-law model describes adequately the behaviour of the polymer melt and the results are in good agreement with experimental data. The shape and location of the free surface of the melt bank is also determined an compares well with photographs available in the literature. Entry and exit flow calculations in slit dies and the determination of extrudate swell are in good agreement with theoretical and experimental studies for inelastic fluids. Visco-elastic calculations performed using the Criminale-Ericksen-Filbey (CEF) constitutive equation diverged for high elasticity levels (Deborah number above 1.0). A simple constitutive equation based on empirical relations for polymer melts was found to enjoy a wider convergence range for elasticity levels encountered in practical situations (De < 5). It predicts several phenomena of polymer melts such as vortex size and intensity, entrance and exit pressure losses and extrudate swell in agreement with experimental findings. It is proposed that such empirical relations can be used along with the finite element method to predict successfully the flow of polymer melts in various polymer processes.</p> / Doctor of Philosophy (PhD)

<p>The objectives of this research were to prepare model long chain branched polyolefin samples with controlled molecular weights and chain structural properties, and to elucidate the effects of long chain branching on the rheological and mechanical properties of polyethylene and polypropylene using a variety of synthesis and characterization techniques. Significant long chain branch (LCB) formation was found by ¹³C NMR measurements in polyethylene (PE) samples produced using the homogeneous catalyst bis(cyclopentadienyl) zirconium dichloride in a semi-batch slurry polymerization. Enhanced levels of LCB were attributed to the in-situ reaction of ethylene macromonomer and the encapsulation of active centres by precipitated polymer chains. PE samples with controlled levels of LCB were synthesized using a two-stage polymerization process. Narrowly dispersed poly( ethylene-co-propylene) (EPR) macromonomer was first produced in a continuous stirred tank reactor (CSTR). The macromonomer was subsequently copolymerized with ethylene in a semi-batch reactor using the constrained geometry catalyst (CGC-Ti). The two-stage reaction process was also employed to synthesize polypropylene (PP) with EPR long chain branches using the catalyst system rac-dimethylsilylenebis(2-ethylbenz[e]indenyl)zirconium dichloride / MMAO. It was found that the stoichiometry of the macromonomer could be used to efficiently control the long chain branch frequency (LCBF) of the copolymers without greatly influencing the copolymer molecular weight. It was also possible to control the long chain branch length by varying the molecular weight of the macromonomer. The rheological properties of various long chain branched polyolefins were measured. These polymers exhibited significantly higher zero shear viscosities (110) and displayed greater shear thinning than linear polymers with similar molecular weights. It was observed that the branch MN had to be greater than 7000 g/mol in order to form sufficient entanglements to significantly influence the rheological responses. The 110 was found to be a sensitive indicator of branch frequency and branch length. Increasing the branch MN led to enhanced values of 110, improved shear-thinning, and elevated flow activation energies. Long chain branching played a significant role in the dynamic mechanical behaviour of polyolefins. Increasing the frequency of branching increased the stiffness of polyethylene, as reflected by the storage modulus. Long chain branching also served to enhance the damping or energy dissipation of PE, shown by increased values of the loss modulus. The dynamic mechanical behaviour of polypropylenes with EPR long chain branches was also characterized. There appeared to be a critical EPR branch length at a MN of approximately 6000 g/mol. When the branch length was below this critical MN, a homogenous system resulted. If this critical MN was exceeded, a two-phase system developed, with fine rubbery domains dispersed in a PP matrix. When a two-phase system developed, it enhanced the loss modulus of the copolymer in a manner similar to impact modified PP.</p> / Doctor of Philosophy (PhD)

<p>This thesis investigates the building and application of principal components analysis (PCA) and projection to latent structures (PLS) models when some objects in the data set have missing measurements. Score calculation is treated first, followed by model application to prediction and monitoring. Model building is explored in the final part of the thesis. The first problem treated in this work is that of estimating scores from an existing PCA or PLS model when new observation vectors are incomplete. Several methods for estimating scores from data with missing measurements are presented and analysed, including a novel method involving data replacement by the conditional mean. Expressions are developed for the error in the scores calculated by each method and the factors that lead to error are drawn from these expressions. The error analysis is illustrated using simulated data sets designed to highlight problem situations. A larger industrial data set and a simulated process data set are also used to compare the approaches. In general, all the methods perform reasonably well with moderate amounts of missing data (up to 20% of the measurements). However, in extreme cases where critical combinations of measurements are missing, the novel method is generally superior to the other approaches. Uncertainty intervals arising from missing measurements are then developed for the squared prediction error (SPE), Hotelling T2, PLS predictions and their contributions. These uncertainty intervals provide performance measures and diagnostics when measurements are missing in process monitoring and prediction. The uncertainty intervals derived agree well with the values calculated from the complete objects and give valuable information about the true level of knowledge about the process. The uncertainty in the contributions is used to correctly diagnose which missing measurement in a set gives the greatest reduction in uncertainty when it is recovered. Insight gained in the analysis of score estimation and model application is applied to model building with the nonlinear iterative partial least squares (NlPALS), maximum likelihood principal components analysis (MLPCA), expectation maximisation (EM) and iterative replacement algorithms. Challenges unique to model building and factors that lead to error in individual steps of each model building algorithm are examined. Recommendations are made to improve the quality of the models obtained, and a procedure is proposed to screen data for objects and variables with missing measurements that would have an adverse effect on a model built using missing measurements.</p> / Doctor of Philosophy (PhD)

<p>Dynamic models for a plasticating extruder melt pressure are formed from steady state, step test and time series analysis data. The extruder pressure response to steps in screw speed is modelled by a first order transfer function with a time constant of approximately 0.2 s. The noise associated with the extruder pressure consists of a pressure surging at 0.125 Hz frequency and a longer term drift associated with the die temperature controller. This drift has a period of approximately 3.0 min. The pressure also has a signal noise at the same frequency as the screw rotation caused by the screw flight passing the tip of the pressure transducer.</p> <p>The digital PI and the self-tuning regulator control algorithms are used to control the extruder. These controllers are implemented in combination with various filters which remove the flight noise. The most successful controller/filter combination is a digital PI controller used with a filter constructed from 2 poles and 2 zeroes.</p> <p>The dynamic response of a disturbance caused by changing the input polymer quality is also examined. The use of Box and Jenkins seasonal time series models as bandstop filters is studied and it is concluded that they have little practicality as filters. Another topic covered by the thesis is the effect that including filters in the forward transfer function part of a process has on the self-tuning regulator algorithm.</p> / Master of Engineering (ME)

<p>The McMaster Pore-filled (MacPF) membranes are a new class of nanofiltration membranes prepared by co-polymerizing a polyelectrolyte (PEL) gel into a microporous substrate. The PEL gives unique characteristics to the MacPF membranes that make them suitable for water softening and fractionation. The main objective of this investigation is to develop a procedure that can be used to design MacPF membranes for specific applications from theory a priori. This procedure involves the combination of two mechanistic models. The first model (gel model) predicts the physicochemical structure (i.e., membrane parameters) of the pore-filling gel (Mika and Childs, 2001). The second model (transport model) can be used in combination with the parameters predicted by the gel model to predict the pressure-driven transport of the MacPF membranes. The transport model is developed in this dissertation. The design procedure and mathematical models outlined in here would be the first time a membrane design is attempted using only theoretical arguments. The membrane transport model is a pseudo 2-dimensional transport model that consists of the extended Nemst-Planck equation (viscous flow and frictional model), a modified Poisson Boltzmann equation, and hydrodynamic calculations to relate the frictional and steric forces to the membrane structure. The model has three fitting parameters that describe four structural properties of the membrane:pore radius, pure water permeability, surface charge density and the ratio of effective membrane thickness to water content. The model successfully predicts and explains the performance of the MacPF membranes as well as one commercial thin-film composite membrane tested simultaneously. The model parameters estimated here are statistically significant and are comparable to previous results (Garcia-Aleman, 1998). A comparison of the performance of the commercial and MacPF membrane was performed using experimental data and the transport model. This analysis shows that the rejection and transport mechanisms are the same in the commercial and MacPF membranes. Convection, diffusion and electromigration are the main mechanisms of solute transport. However, their contribution is different for each type of membrane. Solute rejection in NF membranes is determined primarily by steric and electrostatic effects. For the MacPF membranes separation and selectivity are primarily determined by the latter. The contribution of steric effects to rejection is significantly smaller compared to commercial membranes. Additionally, the membrane parameters were estimated using both the gel and transport models. At this point, the results obtained separately from the two models show similar parameter estimates and separation. However, additional work has to be done to improve the agreement between the models before the procedure outline here can be used to truly tailor-make the MacPF membranes.</p> / Doctor of Philosophy (PhD)

<p>On-line Multivariate Image Analysis (MIA) and Multivariate Image Regression (MIR) methods are developed for purposes of on-line monitoring and feedback control of industrial processes that are equipped with vision systems. The thesis progresses via three main investigative studies through applications of the proposed methods in the steel manufacturing and forest products industries. These studies are concerned with (i) vision based automatic grading of softwood lumber; (ii) empirical modeling of pulp and paper characteristics using multi-spectral imaging sensors; and (iii) texture based classification of steel surface samples with image texture analysis. The first industrial application study addresses the problem of automatic quality grading (classification) of sawn softwood lumber based on visually identifying the severity and distribution of common defects. An extended MIA approach for on-line monitoring of true color (RGB) image representations of lumber boards is proposed, which provides both qualitative and quantitative measures of lumber defects. The proposed approach involves developing a robust MIA model on typical defects commonly found in lumber. These defects are then monitored using the MIA model on lumber boards being imaged by an on-line RGB imaging sensor. The Near-Infrared (NIR) wavelength region (900 nm - 1700 nm) of the electromagnetic spectrum is also investigated for lumber defect analysis using MIA of multi-spectral NIR images. Advantages and shortcomings of using NIR imaging spectroscopy versus RGB cameras for lumber grading are highlighted. The second industrial application involves empirical model based prediction of the properties of finished dry pulp sheets and the classification of paper samples having different compositions. In the pulp study a novel MIR technique extracts relevant feature information from multi-spectral images of the samples acquired through NIR imaging spectroscopy, and uses Partial Least Squares (PLS) regression to relate the extracted NIR feature space to the corresponding (non-image) quality data measured via laboratory analysis. The proposed MIR scheme is successfully used to monitor pulp quality variations in an at-line mode on an industrial pulp process during several grade changes. In the paper classification problem the feature space, extracted from NIR spectroscopic images, is further interrogated using Principal Component Analysis (PCA) to classify the finished samples based on their chemical ingredient information. The third industrial application addresses the problem of classifying steel sheet samples based on their overall surface roughness characteristics. A novel MIA based image texture analysis technique has been proposed, which extracts textural features from grayscale, color, or multi-spectral images in the latent variable space of PCA. The proposed method enables interactive texture analysis of individual images using visual MIA tools. Furthermore, a MIA model can be developed to monitor textural features from various images for the purpose of image classification. The scheme is illustrated on a set of steel surface images with varying degrees of roughness characteristics. Image classification achieved by the proposed technique is compared with that obtained by other classical multivariate statistical methods and conventional texture analysis approaches.</p> / Doctor of Philosophy (PhD)

<p>The synthesis of alcohols from carbon monoxide and hydrogen on a commercial Cu/ZnO catalyst, promoted with K₂CO₃ for the synthesis of higher alcohols, particularly 2-methyl-1-propanol, has been investigated.</p> <p>Discrimination of rate equations for the methanol synthesis, derived from published proposals for the reaction mechanism, was achieved using sequential experimental design. The optimum equation assumed the hydrogenation of a CH₂O surface intermediate as the rate determining step, consistent with proposals from adsorption and isotope studies. Rate equations derived from mechanisms, assuming the hydrogenation of hydroxycarbene or methoxy intermediates as the rate determining step, failed to describe the measured kinetics. The temperature dependence of the parameters for the optimum equation was also determined.</p> <p>The optimum promoter concentration for the Cu/ZnO catalyst was 0.5% K₂CO₃ by weight. The promotion effect was related to the CO and CO₂ chemisorption uptakes relative to the N₂ monolayer volume, the maximum effect occurring when these were equal. The intrinsic activity of the Cu/ZnO catalyst permitted less severe operating conditions than has previously been reported for the 2-methyl-1-propanol synthesis. The H₂/CO feed ratio was important in determining the higher alcohol selectivity; increasing CO and decreasing H₂ favoured higher alcohol production. At low H₂/CO ratios, the increased butanol selectivity with increased conversions were explained by a decreased hydrogenation of the methanol precursor and increased concentration of the higher alcohol precursors. The reaction temperature was limited by low butanol selectivities below 260°C and catalyst instability above 300°C.</p> <p>A chain growth scheme which quantitatively described the observed alcohol distribution for a range of operating conditions, is also presented. Growth was assumed to occur at the α- or β- carbon atoms of the adsorbed precursor. The rate of growth varied as the CO partial pressure and the rate of termination as the H₂ partial pressure. The assumed growth rules are consistent with increased alcohol yields observed with the addition of alcohols to the feed gas. • .' " increased alcohols . growth ~ied as. the CO par~~~ of terminat~on as the H2 partial growth rules are consistent with II with the adAition of</p> / Doctor of Philosophy (PhD)