Thesis: Ph. D., Massachusetts Institute of Technology, Department of Chemical Engineering, 2018 / Cataloged from PDF version of thesis. / Includes bibliographical references. / Globular proteins offer powerful solutions for addressing challenges in the fields of medicine, industry, defense, and energy. Enzymes perform reactions with high efficiency and specificity, allowing for minimal generation of undesired side products even while exhibiting rapid turnover-traits difficult to replicate in synthetic catalysts. These targets make proteins attractive tools for immobilization to form functional catalysts and sensors. Nevertheless, there are many challenges in creating these advanced materials. The activity of the protein must be retained, and control over the structure of the material is desirable. Protein-polymer block copolymers offer an attractive solution to these issues. These materials have been shown to selfassemble into ordered nanodomains while retaining protein activity. However, the phase behavior of these materials is not fully understood due to the complex nature of anisotropic interactions between the proteins. / Within this thesis, a method for creating highly-active thin-film catalysts from myoglobin-PNIPAM bioconjugates is established by flow-coating these materials onto solid supports and then cross-linking them with glutaraldehyde. These catalysts exhibit considerable stability and perform reactions 5-10 times more efficiently than catalysts formed using other common immobilization techniques. However, the self-assembly and structural control of this catalyst was observed to be poor, and it was hypothesized that the poor self-assembly relative to mCherry and EGFP systems could be a consequence of the protein shape. In order to probe the effect of protein shape on self-assembly, a panel of mCherry bioconjugates with differing conjugation sites was studied using small-angle x-ray scattering. / The self-assembly behavior of these conjugation site variants was observed to be robust with only minor differences in phase boundaries and observed phases resulting from the changes in conjugation site. However, observed changes in the domain spacing signaled that modifications to conjugation site offer control over protein orientation within the domains. Based on studies showing that polymer chemistry in bioconjugates has a significant effect on self-assembly, an attempt to quantify these protein-polymer interactions was made using contrast-variation small-angle neutron scattering on mCherry and polymer blends. This technique allows for decomposition of the scattering intensity into its component parts corresponding to correlations between the 3 different pairs of the 2 species in the blends. From this analysis, it was determined that the best ordering bioconjugates have primarily repulsive interactions that can be described using a depletion layer model. / Lastly, the effect of protein properties was screened using a large library of bioconjugates made from 11 different proteins. The primary observed trend was that order increases as molecular weight increases, but a narrow region around 28-30 kDa was observed where bioconjugate ordering was significantly enhanced and additional nanostructures were accessible. / by Aaron Huang. / Ph. D. / Ph.D. Massachusetts Institute of Technology, Department of Chemical Engineering

This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections. / Thesis: Ph. D., Massachusetts Institute of Technology, Department of Chemical Engineering, 2019 / Cataloged from student-submitted PDF version of thesis. / Includes bibliographical references. / Two-dimensional (2D) materials, such as, graphene, transition metal dichalcogenides (TMDs) (e.g., molybdenum disulfide (MoS₂)), and hexagonal boron nitride (hBN), have recently received considerable attention, due to their layer-number-dependent optoelectronic, mechanical, and barrier properties. However, physical understanding of the controlled synthesis and interfacial behavior of 2D materials is still lacking. In this thesis: First, I construct a generalized mechanistic model for the growth of TMD monolayers using chemical vapor deposition (CVD). Combining kinetic Monte Carlo (KMC) simulations and a chemical engineering transport model, I am able to predict the experimentally-observed shape and size evolution of the MoS₂ morphology inside a CVD reactor. Second, I address the challenge of solving the Isomer Cataloging Problem (ICP) for lattice nanopores in 2D materials. / Combining electronic structure density functional theory (DFT) calculations, KMC simulations, and chemical graph theory, I generate a catalog of unique, most-probable isomers of 2D lattice nanopores, demonstrating remarkable agreement with experimental microscopy data for nanopores in graphene and hBN. Third, I study the photoluminescent properties of nanoporous defects in hBN by combining my solution to the ICP with extensive hybrid DFT calculations of electronic bandgaps. Doing so, I map the experimentally-observed emission energies to one or more defect shapes in hBN, thereby demonstrating structure-property relationships for defects in hBN, with implications for single-photon emission from hBN devices. Fourth, using molecular dynamics (MD) simulations, I show that electrostatic interactions play a negligible role in determining the contact angle and the friction coefficient of water on the MoS₂ basal plane. / I show that other planes (e.g., the zigzag plane) are polar with respect to interactions with water, thereby illustrating the role of edge effects in MoS₂. Fifth, I combine lattice dynamics calculations with DFT-based MD simulations to develop a force field for hBN for use in mechanical and interfacial applications. The force field predicts the crystal structure, elastic constants, and phonon dispersion relation of hBN with good accuracy, and demonstrates remarkable agreement with the interlayer and water-hBN binding energies predicted by advanced ab initio calculations. Finally, using MD simulations, I study the wetting and frictional properties of hBN by three different liquids of varying degrees of polarity. I infer that electrostatic interactions affect the frictional properties of various liquids in contact with hBN to different extents, and propose the mean-squared total lateral force as a physical metric to rationalize this observation. / This finding implies that liquids with lower wettability can exhibit higher friction on hBN surfaces. In conclusion, the theoretical and simulation methods developed and applied in this thesis should inform the synthesis of 2D materials, and their use in various applications, such as, optoelectronic devices, mechanical composites, and membranes for gas separation and water desalination. / by Ananth Govind Rajan. / Ph. D. / Ph.D. Massachusetts Institute of Technology, Department of Chemical Engineering

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Chemical Engineering, 2019 / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 115-124). / Microbial fermentation is an attractive method for the renewable production of chemicals. Glucaric acid was identified as a "top value added chemical from biomass" by the Department of Energy in 2004, and a biological route for its production from glucose in E. coli was developed in our lab in 2009. Two of the pathway enzymes, myo-inositol phosphate synthase (MIPS) and myoinositol oxygenase (MIOX), appear to control flux. This work addressed several limitations of these reactions. One approach was the relief of reactive oxygen species (ROS) to improve MIOX performance. MIOX converts myo-inositol (MI) to glucuronic acid. Overexpression of native catalase and superoxide dismutases led to significantly higher titers of glucuronic acid from MI. This result corresponded to better maintenance of MIOX activity and expression over the course of the fermentation. A reduction in labile iron levels, which are linked to ROS formation, was also shown to improve glucuronic acid titers. / A second approach was the examination of natural MIPS diversity. MIPS competes with central carbon metabolism for its substrate, glucose-6-phosphate. Thirty-one representative MIPS homologs were selected using a sequence similarity network. Nineteen variants produced detectible myo-inositol (MI) from glucose, and H. contortus MIPS performed equally well or better than the current S. cerevisiae MIPS. Interesting differences in stability were identified between the variants, and further work to explore the network may yield more information about important sequence features. A third approach was the evaluation of screening methods for glucuronic and glucaric acid to support protein engineering. We attempted to extend a previous screen to growth from glucose, but while growth was achieved from MI, low flux appeared to prevent growth from glucose. A previously-developed biosensor based on the regulator CdaR was also tested. / We discovered that the biosensor does not respond to glucaric acid but instead to a downstream metabolite, likely glycerate, and that the biosensor is affected by catabolite repression. While a reliable screen was not realized, our improved understanding of native regulation aids in the identification of alternative strategies. This work overall produced significant improvements in the glucaric acid pathway and helped to identify opportunities for further development. / by Lisa Marie Guay. / Ph. D. / Ph.D. Massachusetts Institute of Technology, Department of Chemical Engineering

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Chemical Engineering, 2018 / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 132-143). / Since its discovery by Banting and Best, the administration of exogenous insulin to control blood glucose levels has been a primary method of treatment for severe cases of diabetes mellitus. Several decades of insulin engineering and development has led to the clinical introduction of two broadly classified categories of protein therapies: prandial and basal insulins. Although these developments have had profound effects on disease management with respect to insulin-dependent diabetes, the overall strategy for development has historically been restricted to rational design criteria based on static pharmacodynamic profiles, profiles that are inherently naive to physiological changes in the diabetic patient. As a result, stringent patient-dependent regimens are the standard of care with regard to glycemic monitoring and management. / When coupled with issues such as patient noncompliance, severe hypoglycemia, as well as the adverse health effects that result from chronic mismanagement of hyperglycemia, it is obvious that there are still major hurdles that must be overcome to properly treat the disease. Herein, we introduce innovative strategies aimed towards the advancement of novel insulin bioconjugate design and development for enhanced long-term efficacy and glucose-responsive activity. First, we develop a class of unimolecular, glucose-responsive insulin conjugates towards the design of a state-responsive, patient-dependent therapy. / Optimization of this system resulted in the identification of a lead candidate, lns-PL-4FPBA, capable of (1) glucose-mediated changes in solubility for long-term sequestration and intelligent depot formation, (2) superior safety in comparison to clinically used long-acting insulins, and (3) glucose-responsive pharmacokinetic and pharmacodynamic activity in both healthy and diabetic animal models. Next, we pioneer the first design and synthesis strategy of a novel class of sugar-responsive insulin conjugates, with the ultimate goal of developing an insulin bioconjugate capable of sugar-mediated receptor binding interactions. In this effort, we created dynamically cyclized insulin conjugates that were found to exhibit enhanced chemical stability and superior thermal stability relative to the native protein, as well as sugar-mediated destabilization, suggesting the potential to exploit the insulin receptor binding mechanism. / Lastly, we focus on improving basal activity of the insulin protein by utilizing a novel class of zwitterionic insulin polymer conjugates towards the generation of ultra long-acting insulin therapies. The resulting library is demonstrated to afford equivalent biological potency relative to native human insulin, augmented thermal and chemical stability capable of withstanding thermal aggregation for over 80 days, as well as ultra long-acting basal insulin potential. / by Abel Bryan Cortinas. / Ph. D. / Ph.D. Massachusetts Institute of Technology, Department of Chemical Engineering

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Chemical Engineering, 2018 / Cataloged from PDF version of thesis. / Includes bibliographical references. / Modern optoelectronic devices such as photovoltaics and LEDs operate based on transport of nanoscale energy carriers. Promising material for these devices are assemblies of nanometer-sized semiconductors, including quantum dots (QD) and organic conjugated polymers. Unlike crystalline semiconductors that are homogenous in energy and space, there exist static and dynamical heterogeneity in nanoscale semiconductors. To control energy transport and improve their device efficiencies, this thesis presents nonequilibrium energy transport models to understand the effect of nanoscale heterogeneity on material-wide optoelectronic properties with emphasis on excitons. At first, continuum-level analytical theories and finite element simulations are employed to derive exciton distributions in semiconductor films. These models are applied to transient photoluminescence interfacial quenching experiments of CdSe QD thin film interface to measure exciton diffusion length. / A linear elasticity theory is constructed to understand the effect of surface ligands on low-frequency vibrations of colloidal QDs. This theory combined with Raman spectroscopy allows measurement of elastic properties of surface bound ligands that are otherwise challenging to probe. Next, a nonequilibrium exciton dynamics model developed at the coarse-grained level is used to investigate energy transport in disordered QD solids. Through kinetic Monte Carlo and chemical master equation simulations combined with time-resolved spectroscopy techniques, this model reveals that static energetic disorder causes ensemble-averaged exciton diffusivity to decrease over time such that the net diffusivity is reduced relative to the ordered case. A subsequent model based on resonance energy transfer theory discovers scenarios in which there can be disorder-enhanced incoherent energy transport. Such enhancements can be important in processes that are sensitive to molecular-scale fluctuations. / Finally, the role of dynamical disorder due to the environment in the case of organic conjugated polymers in solution is interrogated: the interplay between thermal fluctuations and excited state forces drives exciton migration along the polymer backbone. Simulation results are verified with anisotropy decay measurements of poly(3-hexylthiophene). To simulate exciton transport in large systems like long-chain conjugated polymers, a constrained adiabatic dynamics method is developed. Application of this method highlights that failure to preserve wavefunction symmetry while preventing trivial unavoided problem in an adiabatic dynamics simulation can create unphysical electronic dynamics. / by Elizabeth Moon Young Lee. / Ph. D. / Ph.D. Massachusetts Institute of Technology, Department of Chemical Engineering

This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections. / Thesis: Ph. D., Massachusetts Institute of Technology, Department of Chemical Engineering, 2018 / Cataloged from student-submitted PDF version of thesis. / Includes bibliographical references. / Therapeutic protein products with high solution concentration often possess extremely high viscosity and have difficulties in processing and delivery. It is desirable to predict and control the viscosity of protein solutions based on their interactions at the molecular level. Fundamental understanding on their rheology will greatly facilitate the development and engineering of biopharmaceuticals. In general, viscosity of attractive colloidal dispersions increases with their concentration and attraction strength, and diverges at the gel point. In this thesis, we investigate the mechanism of enhanced viscosity of concentrated protein solutions and colloidal dispersions due to inter-particle attractions. Coarse-grained models of protein solutions and colloidal dispersions are developed. / We improve a previously developed 12-bead model by considering the hydro-dynamic interactions and using the correct forms of screened electrostatic potential and dispersion forces to simulate monoclonal antibody solutions. The model captures anisotropic effects and correctly recovers the solution micro-structures. A random patchy sphere model with controllable surface patchiness is also developed to describe more general colloidal particles with anisotropic interactions. We observe signicant deviations in micro-structure and thermodynamics from isotropic particles at modest particle concentrations. Dynamics and rheology are sensitive to near-field non-central interactions and the resulting rigid constraints. Considering these constraints improves the viscosity prediction of concentrated antibody solutions and explains the diverging viscosity during gelation of attractive colloidal dispersions. / It is also noticed that the rigid constraints in physical gels play a similar role in rheology as the cross-links in chemical gels. We have demonstrated that the rigid constraints, which are seldom accounted for in previous works, are indispensable when computing the stress of a sheared suspension. / Thesis is funded by Genentech Inc. and the MIT Portugal Program / by Gang Wang. / Ph. D. / Ph.D. Massachusetts Institute of Technology, Department of Chemical Engineering

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Chemical Engineering, 2019 / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 113-134). / The focus of this thesis is the development and assessment of techniques for the analysis of electrochemical and transport processes in redox flow batteries (RFBs) at moderate to high active species concentrations under direct current conditions. RFBs hold promise as an energy-intensive storage technology suitable for supporting the integration of intermittent renewable energy sources into the grid, but further improvements in technical performance and reductions in system cost are needed for broad deployment. At their core, all thesis projects are aimed at enabling the development of system descriptors that correlate material properties (e.g., viscosity, conductivity), cell geometry (e.g., flow field design), and operating parameters (e.g., flow rate, current density) to system performance metrics, such as cycle efficiencies and area-specific resistance. / More specifically, the investigation is divided into three primary projects: the development and assessment of a research-scale flow cell; measurements of mass-transfer coefficients; and integration of a polarization model into a standalone application useful for assessing system performance. The differential flow cell is engineered leveraging validation material from industrial collaborators. Not only is the performance is consistent with that of a ten-fold larger cell, but its smaller modular design enables rapid assessment of new chemistries and cell components with minimal materials requirements. Mass-transfer coefficients are then measured using this cell with a well-behaved redox active electrolyte (RAE), in which glucose is added in various amounts to modify the system viscosity with minimal changes to other properties. / The results or methodology developed could be extended other similar RAE systems either as preliminary estimates of mass-transfer performance or as a protocol for carefully evaluating the impact of new system parameters on mass-transfer. Finally, results of this mass-transfer analysis are incorporated into a simple flexible stack model, which can be used to estimate system performance as a function of key materials properties with limited empiricism. / by John Leonard Barton. / Ph. D. / Ph.D. Massachusetts Institute of Technology, Department of Chemical Engineering

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Chemical Engineering, 2018 / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 167-171). / Emulsions are complex fluids with interesting physiochemical properties, which have been widely used in health and personal care, food, coating, and manufacturing, etc.. Rather than considering emulsions as passive materials, they can also be used as active blocks for material preparation and chemical synthesis. This thesis presents a study of mass transport phenomena in two specific types of emulsion systems: microfluidic emulsions and concentrated food emulsions. For microfluidic emulsion systems, the first mass transport phenomenon studied is the exchange of chemicals between microfluidic droplets, which are 10-100 pm in size, and nanodroplets, which are dispersed as a nano- or mini-emulsion. Chemically, thermally, or electrically induced coalescence and micelle activity control the mass exchange between micro- and nano-droplets, leading to applications in reaction engineering and microparticle preparation. / Microdroplets function as micro-reactors that receive chemical from nanodroplets with both the addition rate and dosage well-controlled. The microdroplets could also function as micro-reservoirs that steadily supply chemical to the nanodroplets. For microparticle preparation, microdroplets function as templates to be solidified by reagents carried by the nanodroplets. The second mass transport phenomenon in microfluidic emulsions is the evaporation of droplet solvents or the exchange of solvents between droplets and the continuous phase, which leads to solid precipitation. In particular, this thesis focuses on the formation of drug crystalline particles. A novel solvent/anti-solvent exchange method with a hydrogel binder was developed to prepare highly monodisperse microparticles of either hydrophilic or hydrophobic drugs from microdroplet templates. In addition, we also improved a previously developed spherical crystallization method based on droplet solvent evaporation. / We used the same hydrogel but as a temporary immobilization media to prevent droplet coalescence and to expand the applicable solvent library of this method for industrial applications. For concentrated food emulsions, the mass transport phenomenon studied is the fast removal of the continuous phase and the microencapsulation of lipids into microparticles. With the spray drying technology, "powdered oil" containing up to 55 wt% (dry mass basis) of liquid oil was successfully prepared from concentrated milk protein stabilized emulsions. We discovered that pre-evaporation of raw milk not only offers energy cost savings, but also reduces fat loss. With additional carbohydrates, the surface extractable fat was reduced and powder wettability was improved. This product will serve as the main ingredient of an instant powder ready-to-use therapeutic food for treating child malnutrition in India. / by Tonghan Gu. / Ph. D. / Ph.D. Massachusetts Institute of Technology, Department of Chemical Engineering

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Chemical Engineering, 2018 / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 107-113). / Computers are better than ever at extracting information from visual media like images, which are especially powerful in biology. The field of computer vision tries to take advantage of this fact and use computational algorithms to analyze image data and gain higher level understanding. Recent advances in machine learning such as deep learning based architectures have greatly expanded their potential. However, biologists often lack the training or means to use new software or algorithms, leading to slower or less complete results. This thesis focuses on developing different computer vision methods and software implementations for biological applications that are both easy to use and customizable. The first application is cardiomyocytes, which contain sarcomeric qualities that can be quantified with spectral analysis. Next, CellProfiler Analyst, an updated software application for interactive machine learning classification and feature analysis is described along with its use for classifying imaging flow cytometry data. Further software related advances include the first demonstration of a deep learning based model designed to classify biological images with a user-friendly interface. Finally, blood smear images of malaria-infected blood are examined using traditional machine learning based segmentation pipelines and using novel deep learning based object detection models. To entice further development of these types of object detection models, a software package for simpler object detection training and testing called Keras R-CNN is presented. The applications investigated here show how computer vision can be a viable option for biologists who want to take advantage of their image data. / by Jane Yen Hung. / Ph. D. / Ph.D. Massachusetts Institute of Technology, Department of Chemical Engineering

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Chemical Engineering, 2018 / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 113-119). / Crystallization is a separation technique widely used in chemical processes to produce high-purity solid products. The impact of solution chemistry on the kinetics and thermodynamics of crystallization processes is neither well understood nor properly characterized. Therefore, there exists a need for research to develop chemistry that can exploit the effect of impurities, additives and foreign molecules on the chemistry within crystallizing solutions. The use of rational chemical interactions has the potential of enhancing the controllability of crystallization unit operations, providing a new process handle for chemical engineers with which they can create new crystal forms, enhance product purity, improve yields, or inhibit the formation of undesirable crystals. This thesis focuses on the use of small-molecule chemical additives that exhibit selective intermolecular interactions with crystallizing solutes or their impurities. / Within the work reported, there were two major areas of study: purification and nucleation control. Additive-driven solution complexation with impurities was demonstrated to be a powerful tool for enhancing the purity of crystal product, without penalizing the process yield. The technique was implemented for the separation of structural isomers, and tested for the purification of a large pharmaceutical compound with challenging chemical features. The results herein helped elucidate the capabilities of complex-assisted crystallization, and also outline the thermodynamic and chemical limitations of the technique. The second half of the work explored the impact on nucleation rates of dilute impurities that interact with the supersaturated crystallizing solute. For the first time, impurity-driven nucleation inhibition was systematically and quantitatively proven, using high-throughput induction measurements. / The experimental results were used to discern the thermodynamic and kinetic impact of the inhibitor, and to elucidate a potential underlying mechanism for the observed behavior. Data demonstrated that even a weakly-interacting dilute additive can lead to massive nucleation rate depression through a kinetic pathway, most-likely due to the disruption of the ordering of the solute molecules within high-concentration clusters during nucleation. / by Carlos A. Pons-Siepermann. / Ph. D. / Ph.D. Massachusetts Institute of Technology, Department of Chemical Engineering