The phase behavior of block copolymers has been previously studied both in the melt and in liquid solution, however the influence of compressible solvents has not been explored. Here, the effect of carbon dioxide on the phase behavior of block copolymers is investigated. The effect that CO 2 sorption has on the phase behavior of block copolymers depends on the type of transition: upper order-to-disorder transition (UODT) or lower disorder-to-order transition (LDOT). Poly(styrene-block-isoprene) and poly(stryrene-block-hexyl methacrylate) were examined as UODT-type systems. CO2 sorption increases the copolymer miscibility in a manner consistent with liquid diluents. Conversely for poly(styrene- block-n-butyl methacrylate), which exhibits LDOT-type behavior, CO 2 sorption induces microphase separation, in some cases depressing the transition by hundreds of degrees. This behavior is in stark contrast to the effect of liquid diluents on this system, which slightly increase the miscibility of the segments. To explain these differences in phase behavior, a series of light n-alkanes (C1–C4) were used as the solvent in place of CO2. This homologous series can be used to directly observe the effect of polymer phase compressibility on the phase behavior. The enthalpic interactions with the polymer between the series are similar, whereas the effect the sorption of these alkanes have on the compressibility of the polymer phase is highly dependent on chain length for light alkanes. At identical solvent loading, the location of the UODT temperature was invariant with choice of n-alkane, indicating that the phase behavior in this case is dominated by enthalpic effects. However for LDOT-type systems, a systematic effect of chain length on the depression in the transition was found. The compressibility increase of the polymer with fluid sorption is the driving force for the induction of phase separation. One application of CO2 swollen diblock copolymers is the synthesis of mesoporous silica. The synthesis involved the infusion and selective condensation of tetraethylorthosilicate (TEOS) inside a block copolymer film swollen with carbon dioxide (CO2) by an acid catalyzed reaction. The amphiphilic copolymers used for the templating were poly(styrene-block-ethylene oxide). TEOS reacts selectively in the hydrophilic phase to form silica while the styrene phase remains unmodified.

This work is concerned with using solid-fluid equilibrium in separations by crystallization, as well as providing some fundamental understanding of intermolecular forces in solid-fluid equilibrium. The goal is to develop systematic procedures for crystallization-based separations of complex molecules such as chiral molecules and ampholytes, while providing a link between the molecular scale and the plant scale through the use of solid-fluid equilibrium phase diagrams. Major aspects of this project include synthesis of chiral crystallization processes, crystallization of amino acids and ampholytes, crystallization of enantiomers and polymorphs accounting for kinetics and mass transfer effects, and molecular models of solid-fluid equilibrium of single components and mixtures. Process systems engineering for crystallization of these molecules is facilitated by the use of solid-fluid equilibrium phase diagrams. Some of the complexities in the phase diagrams of these molecules, such as polymorphism and compound formation, are related to the phase diagrams of simple molecular models of aromatic hydrocarbons.

The problem of finding the best reactor system configuration has been considered in the chemical engineering literature since the 1940's. The problem was attacked, invariably, by isolating the reactor system and maximizing an objective function which usually was the yield or some linear combination of the reactor inlet and outlet compositions. However, there are numerous cases where it is not possible to uncouple the reactor problem from the rest of the process. The reason for this is that the optimum values of the reactor design variables (e.g., the reactor temperature, the limiting reactant conversion, the molar ratio of the reactants, etc.) are determined by the economic tradeoffs between raw materials costs and recycle costs, where the recycle costs depend on the structure of the flowsheet (i.e., the distillation sequence, the heat exchanger network, etc.). However, in order to design the heat exchanger network and the separation system, one needs to know the reactor exit flows. Thus, there is normally a coupling between the best reactor design and the design of the equipment in the recycle loop. This thesis presents the results of our efforts in providing solutions to the reactor synthesis and design problem considered in the context of the complete chemical process.

We present a new approach to solid solution thermodynamics using the concept of cell theory. We looked particularly at the influence of packing effects on freezing. To isolate this contribution we studied substitutionally ordered and substitutionally disordered binary hard sphere solid solutions. The absence of long range interactions makes the influence of inhomegeneities in composition and molecular sizes readily measurable. Such mixtures not only exhibit a freezing transition but the shape of the pressure-composition phase diagram depends drastically on the size difference between particles. Our predictions agree well with available Monte Carlo simulation results. In the case of substitutionally disordered binary hard sphere mixtures our predictions also agree with the Hume-Rothery rule which states that a substitutionally disordered binary alloy cannot exist for size differences greater than 15%. This approach is also capable of predicting the formation of some compounds such as AB (NaCl structure) $AB\sb2$ $(AlB\sb2$ structure) and $AB\sb{13}$ $(NaZn\sb{13}$ structure) as well as their domain of stability in terms of molecular size differences. The similarity between the solid-fluid phase diagrams we obtained in the case of binary hard sphere mixtures and that of binary organic systems is a sign that packing effects may actually play a much more important role in the freezing of real mixtures than what was previously believed. To investigate this question and broaden the applicability of our approach we added attractive forces in the form of a 12-6 Lennard-Jones potential to our model. Predictions for the pure component were in good agreement with published simulation results, especially when correlations between the motions of the particles were considered. In the binary case such correlations were not included for simplicity. Preliminary results obtained at constant pressure indicate that the size ratio of the particles dictates the overall shape of the solid-fluid phase diagram while the effect of attractive interactions is to position the diagram in terms of temperature. In addition, the equilibrium lines are relatively insensitive to pressure. A comparison with experimental data for methane and rare gas mixtures was done and yielded good qualitative agreement.

The shape of a crystalline organic solid has a major impact on its downstream processing and on its end-product quality, issues that are becoming increasingly important in the specialty chemical industry. In particular, shape affects the key washing and filtering steps in solution crystallization, and it determines solids agglomeration and dissolution characteristics. Traditional models for predicting crystal shape are based exclusively on the internal crystal structure. They are able to predict sublimation-grown crystals, but are not able to account for the effects of the environment (i.e., solvent and impurities) that dominate solution-grown shapes. Detailed kinetic theories of crystal growth—the spiral, two-dimensional nucleation, and rough growth mechanisms—have also been developed, but have not been widely employed. They require as input certain face-specific and solvent-specific properties that are generally unknown. This work explores all of the traditional methods and detailed kinetic theories in an attempt to develop a new method for predicting crystal shape. The detailed kinetic theories are simplified, yielding relative face growth rate expressions that depend primarily on kink and edge free energies—two microscopic properties of crystal faces. A method for estimating these properties is also proposed. It depends on the solvent's surface free energy, which is often known, and the crystal's internal energy, which can be determined from simple molecular mechanics calculations. The method is used to predict the shape of adipic acid grown from water, ibuprofen grown from polar and non-polar solvents, and biphenyl grown from toluene. An extension to this technique is developed in order to estimate hydrogen bonding between solvents and crystals: it is applied to succinic acid grown from water and isopropanol, and to paracetamol grown from water and acetone. This approach appears to be the first practical technique that can successfully predict solution-grown organic crystal shapes.

Axial mole-fraction profiles were measured in a low-pressure reactive flow of propylene/chlorine and a low-pressure, fuel-rich ethylene flame doped with allene. The purpose was to generate data and test models for improving allyl chloride production and pollutant-related C3 flame chemistry. Molecular-beam mass spectrometry was the principal analytical technique. The propylene/chlorine system had feed conditions of 71.2% propylene and 28.8% chlorine, 76.00 ± 0.01 Torr, and 17.5 cm/s burner-surface gas velocity (298 K). Because propylene and chlorine could pre-react, a novel multidiffusion burner was developed. Mole fraction profiles were mapped for seven stable species. Temperature measurements were made using a K-type thermocouple, and the constant flow cross-section was determined visually. By modeling as a plug-flow reactor with literature rate constants and data and rate constants determined here, a self-consistent reaction mechanism was constructed and used to predict concentration profiles for unmeasured species. Predicted profiles were consistent with measured data. Thus, by also accounting for pressure effects, the new model provides a sound basis for modeling the industrial process. The allene-doped ethylene flame had a fuel equivalence ratio of 1.9 and feed gas composition of 0.5% allene, 18.9% C2H4, 30.9% O 2, and 49.7% Ar. It was operated at 20.00 ± 0.01 Torr with a burner surface gas velocity at 298 K of 62.5 cm/s. Mole fraction profiles were measured for 41 stable and radical species. Data from this allene-doped ethylene flame were compared to the data of Bhargava's (1997) nearly identical fuel-rich undoped ethylene flame. Addition of allene enhanced the production of phenyl and benzene, supporting the arguments that C3 species play a very important role in the formation of phenyl and benzene. Using the data and reaction path analysis, Bhargava's (1997) reaction set was improved. New rate constants were determined, incorrect reactions were removed, and new chemistry was added, improving many of the model predictions. Modeling suggested that the major reaction responsible for increased production of phenyl and benzene was 2C3H3 = phenyl +H. Identification and analysis of an important error in Bhargava's reaction set suggest that a reactive boundary condition at the burner surface may be necessary for improved modeling of this flame.

The mechanism of molecular mass transport processes such as solubilization and Ostwald ripening in surfactant stabilized oil-in-water emulsions and the effect on their bulk physicochemical properties was investigated. Solubilization was slightly influenced by the initial emulsion droplet size. The smaller the droplets were, the faster the oil molecules were incorporated into surfactant micelles. During solubilization, the droplet size of the emulsion increased. This increase was more pronounced for emulsions with low oil droplet concentrations and small droplet sizes. Solubilization was influenced by the nature of the dispersed phase. Small molecular weight n-hydrocarbons were rapidly solubilized. Triglycerides could not be incorporated in micelles due to their large molecular weight. The nature and concentration of surfactant micelles; also influenced solubilization kinetics. Surfactant micelles with HLB numbers close to seven incorporated more oil more rapidly than surfactants with larger HLB numbers. The growth of oil droplets in surfactant stabilized emulsions during solubilization experiments was due to Ostwald ripening. Ostwald ripening depended on the nature of the oil droplets. Low molecular weight hydrocarbons had a higher solubility in the aqueous phase and therefore aged more quickly than high molecular weight compounds. Ostwald ripening was accelerated by the presence of surfactant micelles that acted as carriers between emulsion droplets. A larger number of surfactant micelles therefore resulted in increased Ostwald ripening rates. Surfactants that were highly surface-active and consequently had a higher diffusion coefficient accelerated Ostwald ripening strongly. Both Ostwald ripening and solubilization were accelerated in emulsions were surfactant micelles were present. In comparison, solubilization proceeded faster than Ostwald ripening. Finally, the effect of mass transport processes emulsion rheology and color was investigated. Ostwald ripening influenced the rheology of emulsions. Emulsions made with n-hexadecane exhibited a rapid solid-liquid. like transition whereas the rheological properties of n-octadecane emulsion droplets did not change significantly during the course of the experiment. Ostwald ripening also influenced the color of emulsions that contained a red dye in the aqueous phase. Aging experiments conducted on n-hexadecane emulsion droplets did show significant color changes with time whereas n-octadecane emulsion droplets retained their color during the course of the experiment.

Infusion is the process defined as filling in the pores of starch-based food materials by partially wetting liquid foods so as to obtain a relatively higher calorie food product in a lower bulk volume. The main objective of this research was to investigate the effects of the physical properties of the liquid (surface tension, angle of contact, and viscosity) and the pore structure of the starch-based solid (porosity, and pore size distribution) on the infusion process. Chessmen Butter Cookies (Pepperidge Farms) and Wonder White Sandwich Bread (ITT Continental Baking Co.) were selected as the starch-based porous solid food materials. Bread is a highly expanded product (porosity = 0.92), while cookies are relatively dense products (porosity = 0.58). Safflower oil (surface tension = 32dyn/cm; viscosity = 47centipoises) and silicone oil (surface tension = 20dyn/cm) of the same viscosity were selected to study the effect of surface tension. Silicone oils of 47 and 490 centipoises viscosity were chosen to study the effect of viscosity. Chocolate syrup was selected to assess the effect of nonNewtonian behavior. Both bread and cookies contain pores larger than 200$\mu$m pores which do not offer resistance to flow. Pore diameters measured from SEM micrographs proved the existence of pores with diameters up to 650$\mu$m for bread. The smallest pore size for both bread and cookies was found as 6$\mu$m by mercury porosimetry. Therefore, in order to achieve uniform infusion of liquid food materials containing particulate solids, it is necessary to use starch-based porous food materials with a controlled minimum pore size which is greater than the particle size of the particulate solids integrated with the infusing liquid. Infusion of oils into starch-based food materials is a very efficient process which occurs between vacuum and 200,000dyn/cm$\sp2$, since surface tension of oil-based liquid foods are low. Moreover, the process is aided by the oil absorption capacity of starch-based foods indicated by partial wetting behavior. Contact angles of oils on bread and cookies were found as 45 degrees (partially wetting), while the contact angle of chocolate syrup was measured as 65 degrees on cookies (partially wetting) and 126 degrees on bread (nonwetting).

High molecular weight (HMW) hydrocarbons (defined herein as C35+ compounds) are difficult to characterise by conventional analytical methods. Very few studies have reported precise and reproducible quantification of such compounds in fossil fuels. Nonetheless, such components have important effects on the physical and biological fate of fossil fuels in the geosphere. For example, the phase behaviour of waxy gas condensates is significantly affected by the varying proportions of HMW compounds. Similarly HMW compounds are amongst the most resistant petroleum components to biodegradation. The current study reports the development of reproducible quantitative high temperature capillary gas chromatography (HTCGC) methods for studying both these aspects of the chemistry of HMW hydrocarbons. In addition those hydrocarbons which remain unresolved when analysed by gas chromatography (so called unresolved complex mixtures UCMs) are also studied. UCMs may account for a large portion of the hydrocarbons in many fossil fuels yet very little is known about their composition. Knowledge of these compounds may be important in enhancing the prediction of phase behaviour. Oxidative degradation and GC-MS is used to elucidate the types of structures present within the UCM. The concentrations of C3S4h. ydrocarbons in two unusually waxy gas condensates from high temperature wells in the North Sea were determined by HTCGC. The whole C, 5+ fraction comprised about 20% of the total hydrocarbons and consisted of compounds with carbon numbers extending up to and beyond Coo. By paying particular attention to sample dissolution and injection, good reproducibility and precision were obtained. For example, for authentic n-C, to n-C60 alkanes a relative standard deviation of under 5% for manual injection, linear response factors (1.01 Cm to 0.99 C6), and a linear calibration for 5 ng to 25 ng on-column were found. Limits of detection are reported for the first time for HMW n-alkanes. The limits were found to be as low as 0.8 ng for Cto to 1.87 ng for C60. Tristearin is proposed as a suitable HTCGC internal standard for quantification since the FID response factor (1.1) was close to that of the HMW n-alkanes and response was linear. Importantly, when co-injected with the two waxy North Sea condensates, tristearin was adequately separated from the closest eluting alkanes, n-C59 and n-C60 under normal operating conditions. Qualitative characterisation of the HMW compounds in the waxy gas condensates and in synthetic wax blends (polywax 1000) using HTCGC-EI MS and HTCGC-CI MS produced molecular ions or pseudo molecular ions for n-alkanes up to n- C6o. The spectra of some HMW compounds contained fragment ions characteristic of branched compounds but detailed characterisation was very limited. This study has also shown, for the first time, the significance of the unresolved complex mixture in gas condensatesU. CM hydrocarbonsa ccountedf or over 20% of the total hydrocarbons in a waxy North sea condensateT. he condensatew as first distilled and the distillate UCMs isolated. Thesew ere found to be between 64 to 97 % unresolved after molecular sieving (5A) and urea adduction. The UCMs were oxidised using CrO3/AcOHw hich produced5 -12% C02, and 55-83% dichloromethane-solublep roducts. Thus 65-94% of the original UCMs were accounted for as oxidation products. The remainder were thought to be water soluble acids which could not be determined in the presence of the AcOH reagent. Of the recovered oxidised products, 27- 81 % were resolved and these comprised mainly n-monocarboxylic acids (19-48 %). The average chain length was found to be C12 indicating the average length of alkyl groups. Branched acids, ketones, ketoacids, ndicarboxylic acids, branched dicarboxylic acids, lactones, isoprenoid acids, alkylcyclohexane carboxylic acids and toluic acids accounted for the majority of the remaining resolved products. The distillate UCMs all showed variations in amountso f productsb ut not in composition. Retro-structurala nalysis suggestedth at the UCM in the gasc ondensatew as mainly aliphatic and branched.T he numbero f isomerso f simple brancheda lkaneso ver the UCM molecular weight range (determined by cryoscopy) was calculated to be over 15000. Overall, oxidation provided structural information for about half of the UCM. HTCGC was also used to measure the biodegradability of HMW alkanes in a waxy Indonesian oil. Traditional alkane isolation techniques (TLC and CC) discriminated against HMW compounds above C40 whereas adsorption onto alumina in a warm cyclohexane slurry provided an aliphatic fraction still rich in HMW compounds and suitable as a biodegradation substrate. A waxy Indonesian oil was subjected to 136 day biodegradation by Pseudomonas fluorescens. Extraction efficiencies of over 90 % (RSD <5 %) were obtained for n-alkanes up to C6o using continuous liquid-liquid extraction. Over 80 % of the oil aliphatic fraction was degraded within 14 days. After 136 days only 14% of the original aliphatic fraction remained, yet surprisingly no decreases in the concentrations of compounds above C45 were observed. However, the use of a rapid screening biodegradation method proved conclusively that Pseudomonasfluorescens was capable of utilising n-alkanes up to C60 once the bacteria had acclimated to the HMW alkanes. This is the first report of bacterial utilisation of an n-alkane as large as C.

Several extensions to the phase behavior of binary polymer blends were investigated. Complicated and intricate phase diagrams were found in ternary block copolymer/homopolymer blends as microphase separation competes with macrophase separation. Composition fluctuations were accounted for in the Hartree approximation and the phase diagrams are an extension to mean-field theory. The Lifshitz point, a mean-field multicritical point at the boundary between macrophase separation and microphase separation, is destroyed by the inclusion of composition fluctuations. The kinetics of phase separation in ternary block copolymer/homopolymer mixtures was modeled using the time-dependent Landau-Ginzburg (TDGL) theory calculating the expansion coefficients from molecular parameters. The expulsion of diblock copolymers from the interior of developing domains and subsequent accumulation at the domain interfaces was observed. The simulations suggest that for higher block copolymer content saturation of the interface with block copolymer is responsible for the formation of separate copolymer-rich domains at later stages. Since the diblock copolymer acts as a compatibilizer between the two homopolymers, phase separation kinetics are slowed down. In order to study the viscoelastic effect of polymers, phase contrast optical microscopy and time-resolved small-angle light scattering was used to study the relaxation behavior of a high molecular weight polybutadiene/polyisoprene. Upon the cessation of steady-state shear anisotropic morphologies relaxed by a variety of ways depending on the previously applied shear rate. Application of sufficiently high shear rates resulted in the fastest relaxation of the anisotropy. After passing through a dynamic scaling regime, the largest domains were formed. The TDGL formalism with a convective flow was used to simulate such relaxation behavior. By incorporating only surface tension some of the basic features of the relaxation behavior were recovered. Finally the phase separation in a binary blend in the vicinity of a patterned surface was studies using the TDGL equation. For a short range potential long-range “checkerboard-like” oscillations normal to the surface were observed as transient states, the extension of which strongly depended on the noise. Spinodal decomposition is observed for films as thin as one spinodal wavelength. For thinner films the surface imposes its periodicity on the blend if the surface interaction is sufficiently strong.