<p>Activated carbon adsorption is the most widely considered process for the removal from drinking water of low level toxic chemicals as well as many other organic substances. The adsorption of these compounds involves complex competitive interactions which are not well understood, however, they are crucial to the design of safe and cost effective adsorbers. To obtain a better understanding of competitive adsorption as observed in water treatment adsorber, at the time of a toxic spill, a three part study was conducted with a representative bisolute mixture. The mixture consisted of a toxic chemical (1,1,2-trichloroethane (TCEA)) and the naturally occurring organic substance from a northern river considered as a single compound and measured as TotaI Organic Carbon.</p> <p>The first part of this study consisted of single solute adsorption and desorption experiments as well as blsolute equiIibrium experiment. The adsorption of TCEA was found to be completely reversible, while that of the background organics was virtually irreversible. The bIsolute experiments showed that the background organics substantIally decreased the adsorption capacIty of TCEA, while the adsorption capacity of the background organics was only marginally reduced by the presence of TCEA. The predictive multisolute Isotherm models in the literature could not describe the experimental data. Two other models, which assume that a fraction of the adsorption does not involve competition, were found to describe the data successfully.</p> <p>The second part of this study investigated the rates of adsorption from one-adsorbate and two-adsorbate soIutions. The data of both single adsorbate kinetic experIments were described very well by the homogeneous surface diffusion model. The adsorption of the background organic was very slow, yet it was not necessary to utilize separate terms for the slow and rapid adsorption mechanisms. All of the adsorption occurred by the slow mechanism. The slow adsorption phenomenon, which is suspected to occur in very small pores, appears to be related to the irreversible adsorption observed during equilibrium studies. The background organics that adsorbed by the slow mechanism were virtually not desorbable, while TCEA which adsorbed completely reversibly, did not show any slow adsorption during the kinetics. As expected from the large difference in the values of the diffusion coefficients, to properly model the bisolute batch kinetic experimental data, diffusional interactions such as cross diffusion terms had to be included.</p> <p>The third part of this investigation was a column study involving six different phases, each with a different feed solution. The results were consistent with the equilibrium predictions. Phase IV showed that an expected and potentially dangerous large scale desorption of previously adsorbed toxic organics, caused by a substantial increase in the feed background organics, did not occur. The elution of TCEA actually decreased, probably due to cross diffusional interactions. The column experiment also confirmed that due to competitive interactions and to the increasing loading of background organics on the carbon with time, the adsorbers became substantially less effective at treating toxic spills as the operation cycle proceeded.</p> / Doctor of Philosophy (PhD)

<p>An integrated methodology has been utilized to analyze the factors that affect the activity of metallocene catalysts and the microstructure of the polymers produced under various reaction conditions for both semi-batch and continuous solution polymerization of ethylene. The highly active metallocene catalyst system Cp₂ZrCl₂/MAO was focused on in order to determine the fundamental kinetic rate constants, and a thorough characterization of the polymer microstructure was carried out using high temperature size exclusion chromatography (SEC), carbon-13 nuclear magnetic resonance spectroscopy (¹³C NMR), differential scanning calorimetry (DSC) and shear thinning behaviour as determined by I₁₀/I₂. In the semi-batch polymerization of ethylene in toluene utilizing Cp₂ZrCl₂/MAO, the shape of the polymerization rate-time curves were found to depend on monomer diffusion rates with low ethylene concentrations and poor sparging/mixing impellers giving steady-states while high ethylene concentrations and superior impellers led to decay-type curves/ Increasing temperature increased the catalyst activity with an activation energy for propagation (Ep) of 28.5 kJ/mol while strongly decreasing PE MWs. Several types of commercial and developmental aluminoxanes were studied which indicated that the structure and type of aluminoxane co-catalyst did not influence the MWD of PE although significantly influenced the catalyst activity. Mixing additional TMA with MAO was found not to alter the MWD. For all conditions studied at temperatures between 20-90 ∘C in semi-batch, the PE produced was found to approximate Flory's most probable distribution for a single site-type catalyst. The rate constants for propagation, β-scission and chain transfer to ethylene at 70∘C were found to be kp = 2.7 x 10⁴(M∙s)⁻¹, ktr.β = 0.27 s⁻¹ and ktr.M= 3.8 (M∙s)⁻¹. Long chain branches per 1000 carbons (λN) up to 1.2 were found and the ratio of melt flow indexes indicating shear thinning, I₁₀/I₂, were found up to 15. A high pressure, high temperature CSTR was utilized for the solution polymerization of ethylene at 1500 psig in toluene using the metallocene catalyst system Cp₂ZrCl₂/MAO/TMA. Increasing Cp₂ZrCl₂ concentration led to a decrease in PE MWs while the catalyst activity increased. With increasing temperature between 140-200∘C, the MWs of PE decreased and polydispersities increased while the catalyst activity decreased with an apparent activation energy of -93 kJ/mol. The deactivation of the catalyst is first-order with rate constant kd = 2.1 x 10⁻³ s⁻¹ while kp= 5 x 10³ (M∙s)⁻¹ and ktr.β 3 s⁻¹ at 140∘C. λN's up to 0.6 were found and I₁₀/I₂ were found up to 6.</p> / Doctor of Philosophy (PhD)

<p>In the plastics industry, twin screw extruders are widely used for melting, dispersing and homogenizing polymers. There are a diversity of designs employed throughout the polymer industry, each one having different operating principles and applications. Among the different arrangements of twin screw systems, the intermeshing co-rotating configuration has been found to be one of the most efficient mixers and it is one of the most commonly used pieces of equipment among the continuous mixers due to its self wiping properties. The problem of mixing of polymers involves aspects of fluid dynamics and rheology. Mixing is usually obtained through a combination of mechanical motion of the mixing device and the resulting deformation induced in the flowing material. The quantitative description of the flow patterns is now feasible even in the most complicated geometries through the development of computational fluid dynamics (CFD) tools and the continuous increase in computer resources at lower costs. Intermeshing co-rotating twin screw extruders (ICRTSE) are usually built in a modular fashion to meet the diversity of tasks performed by this type of machine. There are two main types of elements; full flight conveying elements and kneading block mixing elements. The kneading blocks have been the focus of attention for the theoretical analysis of flow due to their significant contribution to the mixing performance of the extruder and the fact that kneading blocks normally work under a fully filled channel condition, which is one of the fundamental assumptions in CFD simulations. The objective of this thesis is to understand the flow mechanisms in the kneading disc section of co-rotating twin screw extruders. This is done by means of the 3D numerical simulation of the flow process within the complex geometry involving intricate passages and continuously moving surfaces. A quasi-steady state finite element model was developed assuming isothermal, non-Newtonian flow. The intricate geometry of the kneading disc section required the development of a tailored finite element mesh generator. An analysis based on particle trajectories, calculated from the obtained velocity field, was carried out to study the effect of geometry on the mixing performance. The approach used for the initial location of the particle tracers was to cover the entire cross section of the kneading blocks. The problem of particles leaving the flow field due to time discretization was addressed by determining the locations where particles crossed the solid surfaces of the discs and reincorporating them into the flow field. The calculation of particle trajectories and deformation history give valuable information about the rapid fluctuations in shear stress experienced by different particles within the flow field. A rigorous examination of the model results was carried out. Comparisons of the 3D model against experimental pressure data in the radial and axial directions are presented. The simulation results were also compared against experimental results of velocity obtained via particle image velocimetry. Results confirmed the ability of the model to predict the flow behaviour. It was determined that inlet and outlet boundary conditions play a significant role in the development of flow patterns in the kneading disc section. The assumption of isothermal flow introduced limitations in the predictions made by the model. Future work should include the addition of the energy equation to the model.</p> / Doctor of Philosophy (PhD)

<p>The removal of fulvic acid, the most soluble fraction of humic matter, with aluminum and polyacrylamide was investigated. The experimental conditions were similar to those usually encountered in practice. The pH ranged from 6 to 8, the aluminum dosage from 10ˉ⁵M to 0.5 x 10ˉ³ M, and the polymer concentration from 0 to 1 mg/L. The fulvic acid concentration was constant at 10 mg/L.</p> <p>The removal of fulvic acid with aluminum was found to depend on the particle size of the aluminum-fulvic acid flocs. This was demonstrated by both the increased removals when filters with a pore size of 0.10 μm were used (as opposed to filters with a pore size of 0.45 μm) and the beneficial effect of the presence of nonionic polyacrylamide. The nonionic polymer promoted both the removal after settling and the removal after filtration. The latter was attributed to the flocculation of micro-colloidal particles which in the absence of the polymer passed through the pores of the filter.</p> <p>Based on a three-level factorial experimental design an empirical statistical model of the removal of fulvic acid was developed. The model was based on the consideration that for a narrow operating region, the response surface may be approximated with a quadratic mathematical expression. By applying this technique the effect of the pH, aluminum dosage and polymer dosage were quantitatively modelled. A polymer dosage of 0.1 mg/L was found to be optimal because it gave the minimum cost of chemicals, for the desired high degree of removal.</p> <p>Finally, the most significant factors that affected the removal of fulvic acid with aluminum, in addition to the pH, the aluminum dosage and the polymer dosage, were the calcium present in the water and the interaction of calcium with the pH. The effect of calcium was larger at pH 8 than at pH 7. The effect of the polymer properties was not as large as the effect of the calcium, with the effect of the polymer molar mass being more important than the effect of the degree of hydrolysis of the polyacrylamide.</p> / Doctor of Philosophy (PhD)

<p>The counter-diffusion of a series of low moIecuIar weight, permanent gas sorbate pairs on synthetic zeolites of various pore-geometries and dimension, was investigated in a semi-automatic flow system. Results of the binary diffusion studies were compared to the single component kinetics of the same species to determine the effect of the adsorbate/desorbate interactions as well as the relationship between the zeolitic pore dimension and molecular size and the respective contributions to the binary diffusion rates.</p> <p>Equilibrium studies of the individual sorbate molecules were carried out in a conventional BET volumetric apparatus. The experimental isotherms were correlated to a two-dimensional analog of a three-dimensional virial equation with good results.</p> <p>Existing diffusion models applicable to the binary sorption system were investigated. Also, a novel kinetic model for diffusion in zeolites based on the principle of random walk has been developed. Qualitatively, the model produces many of the pertinent features observed in the experimental studies.</p> / Doctor of Philosophy (PhD)

<p>The high temperature gaseous desulphurization characteristics of calcined bastnaesite concentrates and cerium oxides have been investigated. The former is a solid solution of rare earth oxyfluorides (RE = La, Nd, Pr) and CeO₂, and has a fluorite type structure. The presence of dissolved CeO₂ in these concentrates is believed to be responsible for the reduction, desulfurization and regeneration characteristics of this mineral.</p> <p>The chemistry of the desulfurization and regeneration process has been elucidated using a fixed bed of cerium oxide at high temperatures (800-1000°C). Coal gases, similar in composition to those generated by a Koppers-Totzek gasification unit, were desulfurized from 1.0% H₂S by volume to less than 4 ngS/mL (3 ppm) in a laboratory scale reactor. This represents a desulfurization efficiency of 99.98%. Desulfurization takes place through a non-catalytic gas-solid reaction where the reactant is in the form of a non-stoichiometric oxide, CeO₂_n.</p> <p>2CeO₂_n(s) + (1-2x)H₂(g) = Ce₂O₂S(s) + 2(1-x)H₂O(g)</p> <p>Adsorption of sulfur is facilitated by the presence of oxygen vacancies in the reduced sorbents. An interactive computing system was used in conjunction with the Ce-O-S phase stability diagram to predict the compositions of reactants and products in terms of the sulfur and oxygen potentials of the gas phase.</p> <p>Regeneration of spent sorbents occurs readily in air, with the evolution of sulfur dioxide:</p> <p>Ce₂O₂S(s) + O₂(g) = 2CeO₂(s) + SO₂(g)</p> <p>Desulfurization efficiencies of cerium oxide compared favourably with those of similar high temperature technologies using iron oxide, limestone or dolomite.</p> / Doctor of Philosophy (PhD)

<p>The purpose of this research was to determine the factors controlling oxygen depletion in the water column (more specifically, the hypolimnion).</p> <p>Five factors were hypothesized to be important:</p> <p>1. Soluble organic carbon concentration</p> <p>2. Temperature</p> <p>3. Nutrient concentration</p> <p>4. Initial oxygen concentration</p> <p>5. Algal respiration</p> <p>These factors were related by a two-level Factorial Experimental Design. Two separate but related Designs were used to "block" and "average" the effects, allowing for differentiation between main effects (for example, temperature) and two-factor interactions (for example, the combined effect of soluble organic carbon concentration and temperature).</p> <p>Four bodies of water (Hamilton Harbour, Guelph reservoir, Red Chalk Lake, and Chub Lake) were tested. The results show that in all cases, except Chub Lake, temperature is the most important factor controlling oxygen depletion. Soluble organic carbon concentration is the most important variable in Chub Lake. The second most important variables are: soluble organic carbon concentration in Hamilton Harbour and Red Chalk Lake, temperature in Chub Lake and the initial oxygen concentration in the Guelph Reservoir. The remaining factors have varying degrees of importance depending on the water body. This suggests that biodegradation is not limited by nutrients and that algal respiration is a small component of oxygen consumption.</p> <p>In addition to these main studies, Temperature and Oxygen Kinetics were also investigated. The Temperature Kinetic experiments showed that higher temperatures produce higher oxygen uptake rates. The Oxygen Kinetic experiments showed no trends except when Hamilton Harbour water was tested. In Hamilton Harbour, higher initial oxygen concentrations produced higher oxygen uptake rates.</p> <p>It is proposed that the Experimental Design represents a bioassay on oxygen consumption. It represents quantification of biodegradation under standard conditions. The relative magnitudes of oxygen consumption under these standard conditions correlate with empirical ideas of degradability of organics in the different water bodies. More work is required to refine the scaling of factors selected, seasonal variability, etc.</p> / Master of Engineering (ME)

<p>Mass transfer performance in cocurrent extraction has been studied in a 5.08 cm I.D. Karr reciprocating plate column.</p> <p>A model for cocurrent mass transfer in a reciprocating plate column, which uses some of the same concepts that have proven useful in the development of a countercurrent mass transfer model, was experimentally confirmed. The indicator colour change method, which employs an instantaneous chemical reaction in the presence of an indicator, was developed to determine the mass transfer rate in cocurrent extraction and has proven very useful in avoiding the additional mass transfer which is always encountered in the sampling and analysis of exit streams. The local mass transfer rate in the column was also investigated by this method.</p> <p>The time-averaged pressure drop for single phase (water) flow was measured in the column and has been compared with a quasi-steady model.</p> / Master of Engineering (ME)

<p>The goal of this research work is to examine the capability of alternative biochemical models for explaining quantitatively the rapid depletion of hypolimnetic dissolved oxygen in Hamilton Harbour during summer stratification. The technique used in the construction of mechanistic mass-balance models in which physical transport mechanisms and biochemical reactions are the building blocks. Three energy cources for oxygen consumption (land-based inputs of BOD and ammonia in situ formation of organics) are developed into four biochemical models. Each model is functionably controlled by land-based inputs of alternative materials (BOD, ammonia, phosphorus, or phosphorus plus nitrogen). Each model is potentially capable of explaining the oxygen deficient. Limnological information collected on Hamilton Harbour in 1975 to 1977 serve as the observational data base for discriminating the merits of each of these oxygen models.</p> <p>The BOD oxygen model was found to explain only 5% of the total dissolved oxygen depletion in the hypolimnion of Hamilton Harbour. The nitrogen-oxygen model and the phosphorus oxygen model explained 40% and 20% respectively. The phytoplankton-oxygen model using oxygen sinks of plankton decay, nitrification and sediment oxygen demand explains the dissolved oxygen (DO) decline in a reasonable fashion and is considered to be verified for predictive purpose. Analysis of water quality management alternatives suggest that a significant improvement will occur ( a minimum DO of 3.5 mg/ℓ) in the long term upon significant reduction in phosphorus and ammonia discharges. An immediate improvement will not be observed due to the high rate of sediment oxygen demand.</p> <p>Further work to improve the model should be directed in three main areas. Estimates of vertical exchange across the thermocline and Lake-Harbor exchange flows on a weekly rather than monthly or seasonal basis would significantly improve the agreement between model predictions and observations. A sediment model is required capable of predicting the rate of change of sediment oxygen demand due to control of land-based inputs. Winter-time data is required for checking and improving the phytoplankton-oxygen model. Good agreement between predictions for and observations in chlorophyll a cannot be expected until the effects of variations in scales of turbulence upon phytoplankton growth are expected.</p> / Master of Engineering (ME)

<p>In Part I of this dissertation an experimental investigation of the molecular weight characterization of dextrans and non-ionic polyacrylamides by aqueous SEC with a DRI/LALLSP detector system is reported. Methodology for the determination of the molecular weight calibration curve and peak broadening parameters σ² (variance of a Gaussian instrumental spreading function) across the chromatogram have been developed. A blend of dextran standards and a blend of non-ionic polyacrylamide standards permit, with one injection, measurement of the molecular weight calibration curve over a wide range of retention volumes, respectively. Measurements with salt-free water as mobile phase have confirmed that some dextran chains may have a negative charge.</p> <p>In Part II of this dissertation aqueous solution polymerization of acrylamide at high monomer concentrations were done using potassium persulphate as initiator in the temperature range 20-70°C, to provide rate and molecular weight data under diffusion-controlled termination. Molecular weight development was followed by measurement of weight-average molecular weight with a low angle laser light scattering photometer (Chromatix KMX6). A kinetic model which accounts for diffusion-controlled termination was developed and used to identify strategies to maximize productivity and molecular weight.</p> <p>In Part III of this dissertation the solution copolymerization of acrylamide (AM) and acrylic acid (AA) was carried out with K₂S₂O₈ initiator at 40°C and 60°C. A new model for the rate of copolymerization of AM and AA in water as solvent was proposed. The differential rate equations for copolymerization derived from the model were transformed into differential equations for the homopolymerization. These differential equations have been solved on a digital computer to simulate or predict the conversions and molecular weight distribution as a function of time.</p> / Doctor of Philosophy (PhD)