Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Chemical Engineering, 1989. / Includes bibliographical references (leaves 285-298). / by John Grant Aunins. / Ph.D.

Thesis (S.M.)--Massachusetts Institute of Technology, Dept. of Chemical Engineering, 2001. / Includes bibliographical references (leaves 72-73). / This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections. / A repetitive sequence of quiescent fluid layers of differing viscosities through which small spherical Brownian particles move is analyzed so as to illustrate in a simple context how the theory of macro transport processes, a generalization of Taylor dispersion theory, may be employed to rigorously analyze spatially periodic micropatterned chromatographic separation devices for circumstances in which the solute species to be separated are animated by the action of species-specific external forces oriented asymmetrically relative to the body-fixed pattern. In the generic "vector" separation scheme, illustrated by our elementary example, the different species undergoing separation move, on average, in different directions relative to pattern-fixed axes, whence their chromatographic sorting is effected according to their different mean angular trajectories through the device. This scheme differs fundamentally from traditional "scalar" chromatographic separation schemes, wherein all species move on average parallel to the animating force (including circumstances in which they are passively entrained in a unidirectional solvent flow) and hence for which the sorting is effected by the relative speeds of the several species through the chromatographic column. Vector chromatography is quantified by two global "macrotransport coefficients," namely the solute mobility dyadic M* (representing the tensor proportionality coefficient between the mean solute velocity vector U* and the external force vector F acting upon the solute molecules) and the dispersivity dyadic D* (resulting from the deviation of the instantaneous position of the particle from its mean position based upon its mean velocity vector). In the present example these coefficients are studied parametrically as functions of: (i) the orientation of the external force relative to the symmetry axis of the fluid layers; (ii) the local viscosity distribution within a layer; (iii) the vector particle Peclet number (constructed from the vector force, the length of the viscosity period, and the Boltzmann factor kT); and (iv) the thermodynamic interphase solute partition distribution coefficient between the two fluid layers comprising a unit cell. / by Kevin David Dorfman. / S.M.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Chemical Engineering, 2011. / Cataloged from PDF version of thesis. / Includes bibliographical references (p. 182-188). / The steam methane reforming (SMR) reaction is the primary industrial means for producing hydrogen gas. As such, it is a critical support process for applications including petrochemical processing and ammonia synthesis. In addition, SMR could be an important component of future energy infrastructures as a means for producing hydrogen as an energy carrier for applications including fuel cells in automobiles and direct combustion for electricity generation. Nickel is the preferred SMR catalyst; however, the efficiency of SMR over nickel can be severely hindered by carbon formation, which leads to the deactivation or even destruction of the catalyst particles. Thus, there is significant interest in catalysts that inhibit carbon formation yet retain activity to SMR. In order to develop improved catalysts for SMR, a thorough understanding of the processes occurring on the nickel surface is needed. In this thesis, computational heterogeneous catalysis is applied to investigate steam methane reforming over nickel (Ni) and silver-alloyed nickel (Ni/Ag) catalysts. Electronic structure calculations using density functional theory (DFT) are employed to develop thermochemical landscapes describing the relative stabilities of SMR intermediates on the catalyst surfaces. In addition, DFT calculations are used to obtain kinetic parameters that describe elementary surface reactions taking place during SMR. A detailed statistical thermodynamics framework is developed to allow for the calculation of enthalpies, entropies, and free energies of the surface species at the temperatures and pressures relevant to industrial SMR. The data from the DFT calculations are used to build detailed ab inito microkinetic models of SMR over the multi-faceted nickel catalyst. The resulting microkinetic models are used to provide insight into the processes occurring on the catalyst surface through identifying the most important intermediate species and reactions occurring on the catalyst. The effects of alloying the nickel catalyst with silver are predicted through modeling the dissociative methane adsorption reaction on multiple facets of the Ni/Ag surface with varying concentrations of silver. In addition, DFT calculations are used to investigate carbon formation on the Ni and Ni/Ag catalyst surfaces, including relative stabilities of various carbon-containing intermediates and the effects of alloying the nickel surface with silver on carbon formation. / by Donnie Wayne Blaylock. / Ph.D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Chemical Engineering, 1996. / Includes bibliographical references (leaves 146-153). / by Colleen A. Vandevoorde. / Ph.D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Chemical Engineering, 2010. / Cataloged from PDF version of thesis. / Includes bibliographical references. / This thesis presents one of the first demonstrations of continuous crystallization in microfluidic devices, and illustrates their use for various applications related to crystallization of small organic molecules. Crystallization is an important process in a number of industries, including specialty chemicals, food, cosmetics, nutraceuticals and, most importantly, pharmaceuticals. Most small molecule pharmaceuticals are isolated in crystalline form, and more than ninety percent of all pharmaceutical products are formulated in particulate, mainly crystalline form. However, crystallization is not a completely understood process. The sensitivity of the process to synthesis conditions gives rise to serious reproducibility issues. The traditional batch crystallizers suffer from non-uniform process conditions across the reactor, and chaotic, poorly controlled mixing of the reagents, often resulting in polydisperse crystal size distribution and impure polymorphs. This makes it difficult to obtain reliable information on the process kinetics that can be used for scale-up, as well as to study the fundamentals of the process. Microfluidic systems offer a unique toolset for crystallization because of well-defined laminar flow profiles, enhanced heat and mass transfer, better control over the contact mode of the reagents, and optical access for in situ characterization. The better control over the synthesis conditions gives rise to the potential for controlling the crystal size, as well as the polymorphic form. In addition, low consumption of reagents makes it an attractive research tool for expensive pharmaceutical compounds. Some of the advantages of microfluidics have been demonstrated for crystallization in micro-batches, but so far not in continuous devices. Continuous crystallization is difficult to achieve in microchannels as uncontrolled nucleation, crystal growth, agglomeration and sedimentation of crystals easily clog the small channels. The interaction of crystals with channel walls may also contribute to channel plugging in these devices. This thesis has developed microfluidic devices for continuous crystallization of small organic molecules for the first time. We have decoupled nucleation and growth, the two key steps of crystallization, using reaction engineering principles, and have developed two separate continuous devices, one for each of these two processes. We have used seeded crystallization and reactor design to achieve controlled growth, as well as to suppress secondary nucleation, agglomeration and sedimentation of crystals. In addition, we have eliminated any significant interaction of crystals with channel walls by controlling the properties of channel surfaces. We have also integrated microscopy and spectroscopy tools with the device for in-situ characterization of crystal size and polymorphic form. We have illustrated the use of these devices to extract growth kinetics data for crystals of various shapes, including high aspect ratio systems such as that with acicular or plate-like habits. The reproducibility and control in our devices have allowed us to elucidate the growth mechanism and fundamentals of the growth process for difficult crystal systems. In addition, we have demonstrated that continuous microfluidic devices offer a unique advantage over the current state-of-the art technology to measure the size, size distribution and growth kinetics of high aspect ratio crystal systems more accurately. Moreover, we have demonstrated the use of microfluidic devices for understanding crystal habit modification in the presence of impurities. We take advantage of the high spatiotemporal resolution of microfluidic devices to study the evolution of crystal habit over time, and to obtain information on the kinetics of habit modification in the presence of different impurities. We have developed an understanding of the habit modification mechanism for alpha glycine in the presence of alpha amino acids. Such information may not only provide insights into impurity-crystal interactions, but also serve as a powerful tool for the design of impurities that can be deliberately added to improve the crystallization process. Furthermore, we have designed and developed a second microfluidic device for continuous supercritical crystallization for the first time. Using supercritical fluid as an antisolvent, we have demonstrated continuous spontaneous nucleation of acetaminophen. We have shown the ability to produce micron-sized crystals, which may be useful for increasing the bioavailability of drugs with lower solubility, as well as for inhalable and highly potent drugs with stringent size requirements. The developed platform can also be used as a high-throughput device for safely screening crystallization conditions in the supercritical domain. We have demonstrated such use by screening the effects of pressure and various solvents on the habit, size and polymorphic form of acetaminophen crystals. / by Mahmooda Sultana. / Ph.D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Chemical Engineering, 2012. / Cataloged from PDF version of thesis. / Includes bibliographical references (p. 191-201). / Plants have evolved highly sophisticated mechanisms of self-repair to regenerate proteins that become photo-damaged over time. Key to this self-repair process is the reversible self-assembly of protein complexes, which is characterized by the molecular recognition of parts, kinetic trapping of meta-stable thermodynamic states, and chemical signaling to switch between states. In this thesis, we mimic such regenerative mechanisms in an effort to develop biological light-harvesting devices with prolonged lifetimes. We demonstrate the first synthetic photoelectrochemical cell capable of mimicking key aspects of the self-repair process. Surfactant addition and removal was used to signal between the disassembly and re-assembly of a photoactive complex demonstrating photo-conversion efficiencies of 40%. These dynamic complexes consist of lipid bilayer disks housing photoactive reaction centers (RCs) that align along the length of a single-walled carbon nanotube (SWNT). Application of a regeneration cycle that reversibly signals between the assembled and disassembled states extends the lifetime of the photoelectrochemical cell indefinitely and increases cell efficiency by over 300% over 168 hours. We modeled the kinetic and thermodynamic forces that drive the reversible self-assembly, and we fit this model to spectrofluorimetric measurements that monitor complex formation. The bestfit rate constants for lipid bilayer and bilayer-nanotube complex formation are 79 mM-Is'Iand 5.4x 10 mM-1 s- 1, respectively. We find that these reactions do not occur under diffusioncontrolled conditions, and the phase diagram predicts a locally optimal surfactant removal rate of 8 x 10-4 s-1. This model was subsequently fit to cyclic complex assembly and disassembly measurements, demonstrating that the forces modeled in this study may form the basis for synthetic and natural photoactive complexes capable of dynamic component repair. In an effort to extend our scope to study natural regeneration mechanisms, we established a platform for quantifying reactive oxygen species (ROS) generation in isolated chloroplasts capable of autonomous regeneration. ROS generation from illuminated chloroplasts from S. oleracea was examined in the presence of dextran-wrapped nanoceria (dNC), cerium ions (Ce3 ), fullerenol, and DNA-wrapped SWNTs. ROS concentrations were evaluated using the oxidative dyes, 2',7'- dichlorodihydrofluorescein diacetate (H2DCF-DA) and 2,3-bis(2-methoxy-4-nitro-5- sulfophenyl)-2H-tetrazolium-5-carboxanilide sodium salt (XTT). Chloroplast photoactivity was monitored throughout the illumination period using chloroplast fluorescence and the artificial, photosynthetic electron accepting dye, dichloroindophenol (DCPIP). The results of this study indicate that dNC offers a promising mechanism for effective ROS scavenging whilst preserving chloroplast photoactivity at concentrations below 5 [tM. We have also established several platforms for studying the glucose production of isolated chloroplasts for biofuel cell applications. We developed an algorithm to quantify single-molecule efflux measurements from individual, photoactive chloroplasts. Near-infrared fluorescing SWNTs have been used in previous studies to report single-molecule binding events via stochastic fluctuations in fluorescence. In this thesis, we develop and compare several algorithms for extracting concentration-dependent rates from the stochastic fluctuations. Overall, the birthand- death model most accurately predicts the rate constants, whereas the moment analysis is more accurate at large forward rates (>10-3 s-1). Glucose efflux from chloroplasts was characterized using a glucose oxidase assay, high-pressure liquid chromatography (HPLC), and a biofuel cell. Calculated export rates of 1.9 and 6 tmol/(mg chlorophyll hr) were measured using the HPLC and fuel cell, respectively. Maximum power densities of 110 pW/cm 2 were achieved with alginate encapsulated chloroplasts. In the presence of regenerative materials, such as dNC, this biofuel cell setup provides a promising platform for demonstrating a biological lightharvesting construct capable of autonomous regeneration / by Ardemis A. Boghossian. / Ph.D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Chemical Engineering, 1989. / Includes bibliographical references (v. 2, leaves 400-417). / by Beth H. Junker. / Ph.D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Chemical Engineering, February 2005. / Includes bibliographical references. / Miniaturized chemical-systems are expected to have advantages of handling, portability, cost, speed, reproducibility and safety. Control of fluid path in small channels between processes in a chemical/biological network is crucial for connecting process elements. We show complete separation of individual phases (phase routing) from two-phase gas-liquid and liquid-liquid (aqueous-organic) mixtures on microscale. To provide for robust interfacing of operations in a network, we demonstrate this ability over a wide range of two-phase flow conditions, including transient ones. Enabled by the technique for complete separation of individual phases from two-phase mixtures, we show mixing of liquids by introduction of a passive gas-phase and demonstrate integration of mixing, reaction and phase separation on a single platform. Additionally, we use the principles developed for phase routing to design microfluidic valves that do not rely on elastic deformation of material. Such valves can be used in a variety of chemical environments, where polymer-based deformable materials would fail. / (cont.) We show a concept for realization of logic-gates on microscale using appropriate connections for these valves, paving the way for design of automation and computational control directly into microfluidic analysis without use of electronics. Further, we use the phase separation concept for sampling liquid from gas-liquid and liquid-liquid mixtures. Such sampling ability, when coupled with a suitable analysis system, can be used for retrieving process information (example mass-transfer coefficients, chemical kinetics) from multiphase-processes. We provide evidence of this through estimation of mass-transfer coefficients in a model oxygen-water system and show at least an order-of-magnitude improvement over macroscale systems. Controlled definition of fluid path enabled by laminar flow on microscale is used in a large number of applications. We examine the role of gravity in determining flow path of fluids in a microchannel. We demonstrate density-gradient-driven flows leading to complete reorientation of fluids in the gravitational field. / (cont.) We provide estimates of the time and velocity scales for different parameter ranges through two-dimensional and three-dimensional finite-element models, in agreement with experimental observations. We believe this thesis addresses a number of both: system and fundamental issues, advancing applications and understanding of microfluidic networks. / by Manish Jhunjhunwala. / Ph.D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Chemical Engineering, 2008. / Includes bibliographical references. / Successful administration of chemotherapeutic agents for cancer treatment requires a balance between the efficacy and the safety of the drug. This often limits physicians to a very narrow therapeutic window. To avoid the harmful side-effects, chemotherapeutic agents may be administered at a suboptimal dose. This is not only a less effective treatment, but can lead to the development of drug resistance by cancerous cells. The therapeutic window can be increased through targeted, stimuli-responsive delivery, which increases the drug concentration at the diseased site, and releases or activates the drug only when it reaches the target. Cancer is a highly variable disease occurring in many organs. There is a need for delivery systems that are easily adaptable for a number of targets in different forms of cancers, and that can accommodate various cytotoxic drugs. The motivation of this project was to develop flexible synthesis procedures for the targeted delivery of chemotherapeutic agents. In this work, we have synthesized and tested three drug delivery systems. The first system is a dextran-based polymer conjugate designed to preferentially deliver doxorubicin to hepatocytes. Doxorubicin has been conjugated to dextran of different molecular weights, with varying degrees of galactose substitution. The degree of doxorubicin substitution was maintained by performing the conjugation of doxorubicin and galactose in two sequential steps. The synthesis scheme was simple, efficient and easily adaptable to other therapeutic agents and targeting moieties with free amine groups. In cell culture studies on target hepatocytes, the dextran-doxorubicin-galactose (DDG) conjugates showed lower toxicity compared to doxorubicin, increased toxicity with higher molecular weight polymers, and greater toxicity with higher degree of galactose substitution. / (cont.) Experiments in the control cell lines showed increased toxicity for higher molecular weight polymers; however, there was no effect due to the presence of galactose. At diameters of 15-40 nm, the polymer conjugates were too large to enter the cell nuclei in large quantities; however, a sufficient amount of doxorubicin entered the nuclei to cause cell death. The higher molecular weight polymers were more effective as they had a higher chain loading of doxorubicin. In spite of significant uptake of the targeted conjugates, the cytotoxicity of the first system was limited since the doxorubicin remained attached to the polymer. For the second system, pH-sensitive dextran-doxorubicin conjugates of different molecular weights were synthesized. The doxorubicin was attached to the dextran backbone through a hydrazone bond. These polymer conjugates were stable at a physiological pH of 7.4, but released over 70% of the attached doxorubicin within 24 h at a pH of 5.0. The rate of release was found to be faster for the lower molecular weight polymers. In cell culture studies, the conjugates showed significant cytotoxicity. The effect of lower chain loading of doxorubicin for the lower molecular weight polymers was offset by the rapid initial release; these polymers showed slightly greater toxicity. Live confocal microscopy indicated that the conjugates were internalized by cells within minutes after incubation. Since release of doxorubicin from the conjugates was much slower than cellular trafficking, it is possible that the conjugates went through multiple endocytosis and exocytosis cycles before the doxorubicin was released. Doxorubicin from the dextran-hydrazone-doxorubicin (DHD) conjugates was found to localize almost exclusively in the nuclei of cells. / (cont.) Since doxorubicin attached to dextran with a stable bond showed limited localization in the nuclei, this indicated that doxorubicin from the acid-labile conjugates was released after internalization by cells. The cytotoxicity of the DHD conjugates was significantly greater than the stable DDG conjugates due to the release of doxorubicin inside cells. In the third system, the targeting and pH-sensitivity functionalities were combined by expressing galactose on an amphiphilic, temperature- and pH-sensitive copolymer of Nisopropylacrylamide (NIPAAm), N,N-dimethylacrylamide (DMAAm) and 10-undecenoic acid (UA). The polymer self-assembled in aqueous medium, and was used to encapsulate paclitaxel. Various synthesis parameters were adjusted to yield polymers that achieved high drug loading and rapid release in a temperature- and pH-responsive manner. The appropriate lower critical solution temperature (LCST) was obtained by adjusting the content of DMAAm and UA to change the hydrophilicity of the polymer. The hydrophilicity of UA was dependent on pH and thus, made the polymer pH-sensitive. Galactose was attached to the end-group of the copolymer to target it to hepatocytes. The drug loading, particle size and release rate were affected by the polymer molecular weight. Paclitaxel was encapsulated in particles that released nearly 100% of the drug within 24 h at a pH of 5.0 at 370C. In the target hepatocyte cell line, the stimuli-responsive, galactose-expressing particles were significantly more toxic than the non-stimuli-responsive as well as the non-targeted particles. In the control cell line, the presence of galactose did not have any effect on cytotoxicity. In summary, we have synthesized three targeted drug delivery systems. The DDG conjugates successfully targeted hepatocytes by expressing galactose. The DHD conjugates would be retained in tumors due to the enhanced permeability and retention effect, and release the drug at the target site. / (cont.) The galactose-targeted paclitaxel-loaded particles synthesized from the temperature- and pH-sensitive polymer achieved a remarkable increase in toxicity in the target cell line, while maintaining base toxicity in the control cell line. Both the amount of drug delivered and the rate of release were found to be important in the efficacy of the drug delivery vehicles. / by Noreen Tasneem Zaman. / Ph.D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Chemical Engineering, 2003. / Page 126 blank. / Includes bibliographical references (p. 113-114). / This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections. / Dialysis-related amyloidosis (DRA) is a frequent complication of end-stage renal disease that has been associated with the accumulation of 2-microglobulin (2m). Excluding transplantation, existing kidney replacement technologies are believed to remove insufficient quantities of P2m for the prevention of DRA, as they are non-specific and based on size-exclusion. A proposed DRA therapy is to use immunoadsorptive particles within an extracorporeal Vortex Flow Plasmapheretic Reactor (VFPR) to specifically remove 2m from blood. The compartmental design of the VFPR allows for the use of small adsorbent particles (100 m) that possess inherent mass-transfer advantages over the larger ones (>400 gIm) that are required for safe contact with whole blood for this application. Demonstrating the efficacy of this technology as a therapy for DRA would support its tailored application for treating other pathologies that are caused by circulating compounds such as sepsis, liver failure, autoimmune disease, drug overdoses, and genetic disorders. Whole anti-P2m antibodies (BBM.1) were immobilized onto agarose beads and used within a VFPR to remove donor baseline and defined quantities of recombinant 32m from whole human blood, in vitro. A dynamic immunoadsorption model was developed for the VFPR that was based upon the independent characterization of the mass-transfer processes within the VFPR and the thermodynamics of the immunoadsorbent. The experimentally-observed and model-predicted dynamics of 32m clearance from the blood indicate that the process controlling the rate of P2m removal was the hemofiltration rate (50 mL-plasma/min), which was on the order of the reported supply rate of 2m into the vasculature (70 mL-plasma/min). / (cont.) Single-chain variable region (scFv) antibody fragments offer several potential advantages over whole antibodies due to their size and genetic definition, as well as their amenability for microbial expression and in vitro evolution. Hence, a BBM.1 scFv was expressed by a yeast display vector and its affinity was quantified with a fluorescence-activated cell sorter (KD = 0.008 +/-proposed therapy to treat and/or prevent DRA. / by Eric A. Grovender. / Ph.D.