Rice University Theses and Dissertations

Permanent URI for this collection

https://hdl.handle.net/1911/13110

Rice University makes all graduate theses and dissertations (1916-present) available online at no cost to end users. Occasionally a thesis or dissertation may be be missing from the repository. If you are unable to find a specific dissertation, please let us know and we will attempt to make it available through the repository, provided that the author has not elected for it to be embargoed.

News

Visit the web site for Rice University's Office of Graduate and Postdoctoral Studies for more information about Rice graduate student requirements for theses and dissertations.

Browse

Now showing 1 - 20 of 14665

Continuities and Divergences in Baroque Twelve-Mode Theory
(2025-04-26) Shafner, Noah; Barnett, Gregory R.
Twelve-mode theory, which originated in the theoretical writings of Heinrich Glarean (Dodecachordon, 1547) and Gioseffo Zarlino (Le istitutioni harmoniche, 1558), was widely disseminated in Italian and German theoretical and pedagogical sources until about 1800, which raises the question: what relevance did a theory that originated in plainchant and later adapted to vocal polyphony have for a musical culture increasingly dominated by concertato idioms and the newer genres of opera, oratorio, sonata, and concerto? Any answer to this question is complicated by the fact that twelve-mode theorists rarely addressed the relevance of their theory to newer music in explicit and unequivocal terms. As a result, previous scholarship has offered an array of views on the practical relevance of Baroque-era modal theory, ranging from outright skepticism, on the one hand, to its whole-sale adoption as a means of analyzing tonal style, on the other. In this paper, I argue that twelve-mode theory was never a widespread and coherently disseminated technique; rather, it remained a theoretical construct that, over its long history, underwent diverse, even incompatible interpretations. To illustrate this, I will trace the divergent paths of German and Italian modal theory, examining their distinct technical and stylistic orientations in three chapters. In Chapter 1, I examine two early eighteenth-century theoretical writings by Benedetto Marcello (Lettera famigliare, c. 1716) and Johann Joseph Fux (Gradus ad Parnassum, 1725) to demonstrate that the continued appeal of modal theory in the late Baroque lay in nuances absent from major-minor keys. In Chapter 2, I will compare early seventeenth-century theorists Girolamo Diruta (Seconda parte del Transilvano, 1609) and Seth Calvisius (Exercitationes musicae duae, 1600) along with early nineteenth-century theorists Giacomo Tritto (Scuola di Contrappunto, 1816) and Georg Joseph Vogler (Choral-System, 1800) to demonstrate that German and Italian twelve-mode theories diverged in two ways: first, while Italians largely adhered to traditional modal-polyphonic precepts as a stylistic topos, German theorists sought a more flexible application of the modes as a theory of tonal organization; and second, Italian theoretical sources and modally-designated compositions document a strong association between the modes and fugal technique, whereas German theoretical sources testify to a consistent association between the twelve modes and Lutheran chorales. In Chapter 3, to further substantiate these divergences, I will compare two early eighteenth-century twelve-mode collections by Azzolino Bernardino della Ciaja (Saggi per Organo, 1727) and Georg Philipp Telemann (XX kleine Fugen, 1731). The broader significance of twelve-mode theory’s continued survival lies in the many kinds and purposes of music theory in the early modern era: as I will demonstrate, twelve-mode theorists aimed, not at music analysis in the modern sense, but variously to engage in rational and humanist speculation, demonstrate authority and erudition, provide prescriptions for composers, and negotiate between newly composed and liturgical music.
Relational Computation for Very Large-Scale Machine Learning
(2025-04-25) Tang, Yuxin; Jermaine, Christopher M
In mathematics, a tensor is an algebraic object that describes multilinear rela- tionships among sets of algebraic entities associated with a vector space. From a computational perspective, tensors are commonly represented as multi-dimensional arrays—a format that plays a central role in machine learning. A widely used convention for expressing tensor operations is Einstein summation notation (EinSum), which compactly encodes summation over indexed terms. This notational framework not only streamlines the expression of complex tensor computations but also lends itself to an alternative interpretation: a multi-dimensional array can be viewed as a map- ping from a vector of integers (i.e., a primary key) to a real number. This perspective aligns closely with the classical definition of a relational database relation. As a result, many numerical and machine learning computations in tensor calculus can be reformulated as sequences of joins and aggregations over relational data. Executing these computations within a relational database system offers several key advantages, including automatic parallelization, distribution, and scalability. Moreover, relational databases are particularly effective at handling sparsity, as they are designed to effi- ciently represent and process cases where only a small subset of the possible primary keys actually occur in the relation. In this thesis, I propose an extension to Einstein notation called Upper-Case- Lower-Case Einstein Notation—a simple yet expressive framework for describing tensor programs that interleave operations over sparse (relational) data with efficient kernel calls over dense tensors. This notation enables the concise representation of computations optimized for complex sparsity patterns. To support this notation, I develop a compiler, SparseEinSum, which takes standard EinSum expressions as input, transforms them into extended Upper-Case-Lower-Case Einstein Notation as intermediate representation, and compiles them into tensor-relational algebra. The compiler incorporates sparsity estimation and cost-based schema selection to guide the transformation. The resulting programs can be executed on virtually any relational database system, leveraging arrays to manage dense tensors within a relational execution model. Experiments across tensor computation benchmarks demonstrate that the generated tensor-relational computations offer significant performance improvements. To support automatic differentiation of relational computation compiled from EinSum, I derive key rules that enable automatic differentiation for relational algebra. I introduce functional relational algebra to build functions in the relational domain and define relational analogs of partial derivatives, Jacobians, gradients, and a set of relation-Jacobian product rules for core relational operators, including table scan, selection, aggregation, and join. This functional framework builds the foundation for differentiation in relational algebra. Then, I propose a relational algebra automatic differentiation algorithm using an efficient, correctness-preserving implementation of the relation-Jacobian product. Through extensive experiments, I show that executing machine learning computations on top of a relational engine—augmented with relational algebra automatic differentiation algorithm—can scale efficiently to very large datasets. The resulting system achieves performance competitive with specialized distributed machine learning systems, while retaining the advantages of relational query optimization.
Embargo
Acoustically targeted measurement of transgene expression in the brain
(2025-04-25) Seo, Joon Pyung; Szablowski, Jerzy
Gene expression is a critical component of brain physiology and activity. Brain development, function, and plasticity relies on a regulated process of converting genetic information into functional products. However, monitoring gene expression in the living brain has been a significant challenge. The confined structure of the brain, protected by the cranium and shielded by the blood-brain barrier, has posed difficulty in non-invasive and sensitive measurement of gene expression with specificity. The aim of this thesis is to develop a new paradigm of technology capable of measuring gene expression in the brain non-invasively with cell-type, spatial, and temporal specificity. To achieve this, we combined focused ultrasound liquid biopsy and recovery of engineered protein markers that are designed to be expressed in neurons and exit into the brain’s interstitium. When ultrasound is applied to targeted brain regions, it temporarily opens the blood-brain barrier and releases the interstitial markers into the bloodstream. Once in blood, the markers can be readily detected from blood collection followed by compatible biochemical techniques. We call this Recovery of Markers through InSonation (REMIS). We demonstrated improved recovery of engineered Gaussia luciferase marker, under constitutive promoter, from the brain into the blood in every tested animal. Further, we implemented the markers to measure endogenous neuronal signaling activity by controlling the expression of the marker under a genetic circuit that responds to c-Fos when activated by enhanced neuronal activity. Lastly, we measured enhanced serum level of overexpressed human alpha-synuclein in the engineered Parkinson’s disease model mThy1-aSyn (Line61) mouse strain with REMIS. Overall, our work demonstrates the feasibility of combining engineered gene expression reporters and focused ultrasound liquid biopsy to noninvasively and specifically measure gene expression in the intact brain.
Embargo
Cheap Meat, Plastic Lives: Cattle Imports, Veterinary Care and Shifting Livestock Economies in Turkey
(2025-04-25) Haspolat, Gizem; Gunel, Gokce
This dissertation explores Turkey’s "cheap meat policy" and the traffic in cattle through an ethnographic study of live animal imports. It investigates how cattle are turned into globally traded commodities by focusing on two interconnected dimensions: the sensorial frictions of global live trade and the longer histories of (forced) animal mobilities. The research centers on underexplored sites of the Animal Industrial Complex (Noske 1989), such as importing companies, ports, customs areas, and agricultural expos, where cattle are (re)made into globally tradable commodities. It emphasizes the importance of sensory experience, especially touch and smell, in mediating and at times disrupting the smooth functioning of the traffic in cattle, which situates live imports not just as an economic exchange, but a system that reorganizes and consolidates certain roles ascribed to cattle, veterinarians, importers, and state representatives in their respective ways. By foregrounding the labor, infrastructure, and political strategies required to sustain import-driven livestock economies, the dissertation highlights tensions between economic imperatives and the embodied, sensory presence of living beings in global trade. Rather than approaching “cheap meat” through consumption or as a given category, it focuses on the political-economic processes that situate it as a reflection of claims over welfare and development, which defer conversations around agricultural transformations in Turkey, international regulatory frameworks, and internal conflict and displacement. These longer histories shape the organization of live trade and reveal continuities in global systems of forced mobilities and exploitation, exemplified by live carrier ships functioning as “floating barns.” While these recurrent histories underscore the exhaustive powers of animal capitalism, they also introduce frictions, failures, and resistances that challenge capital’s totalizing grip.
Embargo
Creation of an Integrated Gas Sensor Platform using 2D Materials
(2025-04-25) Lai, Jiawei; Ajayan, Pulickel
Two-dimensional (2D) materials have shown promising gas sensing capabilities due to their high surface-to-volume ratio and excellent electronic characteristics. However, the weak signals generated by 2D sensors are typically measured using ultra-precise but bulky electrical characterization equipment. Despite their potential, 2D material-based gas sensors have not yet been widely integrated into Internet of Things (IoT) systems for real-time, wireless, and continuous gas monitoring in both urban and industrial settings. This work presents an integrated 2D material-based gas sensing platform that amplifies low-intensity sensing signals on-site and employs Bluetooth 5 technology for long-range, real-time data transmission. It facilitates high-quality data collection from remote 2D gas sensors, addressing modern needs for life safety, smart living, efficient production, and environmental preservation. The thesis introduces 2D material-based gas sensors, details the development of the wireless transmission system, and describes the fabrication processes for 2D sensors and antennas. The creation of this platform lays the foundation for artificial intelligence-assisted, data-driven gas sensing with 2D materials, offering a promising approach to overcoming the selectivity limitations of chemiresistive-type 2D gas sensors, which yields richer insights into sensing behavior and ultimately deepens the understanding of the dynamics of modern human living conditions.
What makes a good leader: Validation of a situational judgment test to assess leader behavioral knowledge
(2025-04-25) Chen, Rebecca; Oswald, Frederick L
Leadership behavior is a critical reflection of one’s leadership capability. Notably, having knowledge of leadership behaviors contributes to actually enacting these behaviors. Thus, the purpose of the current research is to validate a situational judgment test (SJT) measure to assess leadership behavioral knowledge. The first study aims to establish a multidimensional framework of leadership behaviors based on Campbell’s Model of Leader Performance (2012) and leadership inclusion behaviors toward diversity, equity, and inclusion (Shore et al., 2011; Silver et al., 2022). Using this framework, the second study proceeds to validate a SJT measure of leadership behavioral knowledge. Study findings suggest evidence supporting the multidimensional framework of leadership behaviors, in addition to the need for further refinement and validation of the SJT leadership measure. Altogether, the findings contribute to theoretically informing the construct space of effective leadership, in addition to providing practical guidance for developing a leadership assessment to be used for future leader selection, training, and development.
Embargo
Race, Religion, Marriage, and the Making of Black American Muslims
(2025-04-25) Ferguson, Jauhara I; Howard Ecklund, Elaine
Black Americans have the highest rate of religious belief and participation in the United States; yet studies on marriage overlook the impact of religion on the marriage choices of Black Americans. Likewise, extant literature on religion and marriage focuses primarily on Christian communities, ignoring the connections between religion and marriage for Muslims, a significant religious minority among Black Americans. To fill these gaps, I interview 44 never-married Black Muslims living in Houston or Atlanta. I ask: How do religion and race shape marriage processes among Black Muslims? And to what extent do Muslim marriage processes shape Black Muslim identity? I outline the multiple ways that Black Muslims seek out marriage and consider how race, gender, and immigrant status intersect with religion to shape Muslim marriage processes. Although Black Muslims draw on Islamic frameworks to give meaning to marriage, they actively negotiate these frameworks to decide how to navigate romantic relationships as practicing Muslims in the United States. Given the sociohistorical legacies of anti-Black racism in the United States, Black Muslims identify significant challenges to marriage within Black communities and look to Islam as a potential tool to address these broader community concerns. While Islam frames their perceptions of marriage, Black Muslims frequently encounter racialized barriers to marriage within the multiethnic Muslim American community. I consider how Black Muslims navigate these racialized barriers, and I argue that through this process, marriage becomes a site of both religious and racial identity formation. I examine how gender shapes experiences with racialized barriers to marriage in the Muslim American community and illustrate how such experiences can push Black women to reinterpret Islamic frameworks of marriage. Additionally, I analyze the marriage processes of immigrant-origin Black Muslims and show the extent to which family-origin and ethnicity shape marriage opportunities and choice. My findings show how U.S. Black Muslims use Muslim marriage to highlight their place as Muslims in American society, while simultaneously articulating their unique position as Black people. I illustrate how the formation of Black Muslim identity compels Black Muslims to see themselves as distinctive in both Muslim American community and in American national contexts.
Embargo
Development of Extracellular Matrix-Based Colloidal Inks for Cartilage Tissue Engineering
(2025-04-25) Perez, Marissa R; Mikos, Antonios G
Additive manufacturing enables spatial control over bioactive molecules and cells to mimic native tissue architecture, but a lack of bioinks that balance biological relevance and printability limits its potential. Decellularized extracellular matrix retains native biochemical cues but suffers from poor mechanical stability, restricting its use in 3D printing. This work presents the development of composite colloidal inks using methacryloylated decellularized cartilage extracellular matrix nanoparticles blending with gelatin nanoparticles to improve both printability and biofunctionality. The resulting inks are shear-thinning, self healing, and UVcrosslinkable, enabling the fabrication of tunable 3D-printing scaffolds. These scaffolds supported human bone marrow mesenchymal stem cell chondrogenesis, evidenced by enhanced collagen deposition, upregulation of chondrogenic gene expression, and suppression of osteogenic markers expression without exogenous differentiation factors. This study also explored the use of machine learning approaches to predict the print quality of 3D printed poly(propylene fumarate) and to identify relationships between printing parameters and the print quality. Print speed and material composition had the greatest effect on scaffold quality. Additionally, this work examined printing consistency with colloidal inks. Unlike poly(propylene fumarate), the colloidal inks required real-time parameter adjustments to maintain print fidelity, likely due to pressure-induced phase separation. Overall, this research introduces a novel, biologically active, and customizable colloidal ink platform for cartilage tissue engineering and broadens understanding of print behavior in colloidal systems.
Emergence of Ramping Activity in Random Spiking Networks: A Biologically Plausible Model for Learning Timing
(2025-04-25) Feng, Bolu; Alabastri, Alessandro; Shouval, Harel
The ability to perceive and learn timing is crucial for animals to interact with the world. Experiments have shown that during motor planning, certain neurons in multiple brain regions exhibit slow ramping activity that gradually increases until an action begins. This ramping activity is believed to be the neural basis of timing. However, the mechanisms by which these ramping activities emerge from network dynamics and how they are learned through reinforcement learning remain poorly understood. We propose a spiking neural network model and a biologically realistic learning rule to learn this ramping behavior that can last on the order of seconds. Our model consists of a randomly connected recurrent neural network (RNN) that integrates input from a decision-making network. Synaptic connections are trained using an eligibility trace-based learning rule. Using mean-field theory, we analyze the RNN to determine the conditions necessary for it to function effectively as an integrator. This model qualitatively reproduces the ramping patterns observed in mouse brains during decision-making tasks at the single-neuron and population levels. Furthermore, an extended version of our model accounts for additional experimental findings in mouse decision-making tasks.
Embargo
Mechanical Properties of Two-dimensional Nano-composites and Oxides
(2025-04-25) Shin, Bongki; Lou, Jun; Han, Yimo
Two-dimensional (2D) materials have attracted enormous interests owing to their extraordinary properties due to their atomic-level thickness and robust in- plane atomic bonding, positioning them as promising materials for advanced technological applications across electronics, photonics, sensing, energy storage, and structural composites. However, their practical applications have been hindered by intrinsic brittleness, susceptibility to defects, and relatively low fracture toughness. This thesis systematically investigates intrinsic and extrinsic toughening mechanisms, along with anisotropic fracture properties, in selected novel 2D materials and composites to enhance their mechanical robustness and reliability. The intrinsic toughening mechanisms are explored through detailed studies on monolayer amorphous carbon (MAC) nanocomposites, investigating how structural heterogeneities (crystalline and amorphous domain) influence fracture resistance. In-situ scanning electron microscopy (SEM) tensile testing with molecular dynamics (MD) simulations provides comprehensive insights into fracture processes and toughening behaviors. Extrinsic toughening strategies are investigated through two-dimensional covalent organic framework sandwich structures, demonstrating significant improvements in fracture toughness. Additionally, anisotropic fracture behavior is studied in monolayer titania nanosheets, emphasizing the role of crystallographic orientation and defect distributions in determining mechanical performance. Overall, this thesis provides fundamental insights into fracture mechanics and toughening mechanisms in 2D materials, offering practical strategies for designing mechanically robust and reliable nanocomposite systems. The findings not only advance the fundamental understanding of 2D material behavior but also expand their potential applications in next-generation engineering technologies.
Embargo
Non-Equilibrium, Ultra-Fast Heating Techniques for Material Synthesis, PFAS Mineralization and Upcycling.
(2025-04-25) Scotland, Phelecia Z; Tour , James; Han , Yimo
The increasing demand for sustainable technologies and materials has led to a critical need for efficient, scalable, and environmentally friendly solutions for material synthesis, environmental remediation, and resource recovery. Among the innovative technologies addressing these challenges, Flash Joule Heating (FJH) has emerged as a versatile and transformative technique. This thesis explores the application of FJH in four significant areas: heteroatom-substituted graphene synthesis, the destruction of per- and polyfluoroalkyl substances (PFAS), the recovery of critical metals from lithium-ion batteries (LIBs) aided by waste PFAS and the synthesis of silicon carbide nanowires using waste glass. Graphene, a two-dimensional carbon-based material, is renowned for its extraordinary properties, including high electrical and thermal conductivity, mechanical strength, and chemical versatility. These properties make graphene a highly sought-after material for applications in energy storage, electronics, and catalysis. The functionality of graphene can be further enhanced by heteroatom substitution, which involves incorporating non-carbon atoms, such as nitrogen, boron, sulfur, and fluorine, into its lattice structure. These heteroatoms modify the electronic and chemical properties of graphene, expanding its range of potential applications. Traditional methods for heteroatom substitution, such as chemical vapor deposition and solvothermal processes, are often time-consuming, resource-intensive, and difficult to scale. In contrast, FJH offers a rapid, energy-efficient, and scalable alternative for producing high-quality, heteroatom-substituted graphene. By subjecting pre-formed graphene or carbon precursors to rapid high-temperature heating in the presence of heteroatom-containing precursors, FJH enables precise control over doping levels and ensures structural integrity, making it a promising method for scalable graphene functionalization. This is covered in chapter one of this thesis. In addition to its role in material synthesis, FJH provides a novel solution to a pressing environmental challenge: the destruction of PFAS, often referred to as "forever chemicals." PFAS are a class of synthetic organofluorine compounds widely used in industrial and consumer applications, including firefighting foams, non-stick coatings, and water-resistant materials. The strong carbon-fluorine bonds in PFAS make them highly resistant to degradation, leading to their accumulation in the environment and posing significant risks to human health and ecosystems. Current remediation techniques, such as adsorption onto granular activated carbon (GAC), capture PFAS but do not degrade them, leaving behind secondary waste. FJH addresses this limitation by degrading PFAS adsorbed onto GAC through high-temperature treatment, breaking the carbon-fluorine bonds and converting PFAS into benign byproducts. This process not only eliminates PFAS but also enables the upcycling of PFAS-contaminated GAC into valuable materials, such as graphene, demonstrating a sustainable approach to waste management. This is covered in chapter two of this thesis. We then demonstrate that FJH offers a sustainable and efficient approach for resource recovery from spent lithium-ion batteries (LIBs), which play a crucial role in modern energy storage systems. LIBs contain valuable metals, such as lithium and cobalt, whose extraction and processing are energy-intensive and environmentally damaging. The growing demand for these metals, driven by the proliferation of electric vehicles and renewable energy technologies, has raised concerns about resource scarcity and the environmental impact of conventional recycling methods. FJH offers a rapid and solvent-free approach to metal recovery, facilitating the fluorination of lithium into lithium fluoride (LiF) and the reduction of cobalt into metallic form. These transformations occur within a few seconds, minimizing energy consumption and environmental impact while enabling the efficient separation of metals for reuse. FJH addresses the challenges associated with LIB recycling as well as waste PFAS degradation. This is covered in chapter three. Finally, we demonstrate a flash process for upcycling waste glass into SiC nanowires within seconds. By introducing fluorine, iron oxide present in the waste glass is activated, catalyzing the formation of one-dimensional (1D) SiC nanowires. The resulting SiC nanowires exhibit superior performance in composite reinforcement compared to conventional SiC powders. Additionally, a life cycle assessment (LCA) and techno-economic analysis (TEA) reveal that our process significantly reduces environmental impact and production costs compared to conventional synthesis methods. This work highlights fluorine as a versatile and cost-effective agent for modulating nanomaterial growth kinetics and tailoring morphology, providing a sustainable and scalable approach for advanced material synthesis and is discussed in chapter 4 of this thesis. This thesis underscores the versatility and scalability of FJH as a platform for material innovation, environmental remediation, and resource recovery. Through its application in heteroatom-doped graphene synthesis, PFAS destruction, metal recovery from LIBs, and synthesis of one-dimension materials, FJH demonstrates its potential to bridge the gap between fundamental research and practical solutions, advancing both sustainability and technological progress
Embargo
Neutron Diffraction Studies of Kagome Metal FeGe
(2025-04-25) Klemm, Mason Lee; Dai, Pengcheng; Yi, Ming
The two-dimensional kagome lattice is the operative characteristic in a large class of materials that exhibit a wide range exotic magnetic and electronic phenomena purely as a consequence of the unique corner-sharing triangle geometry and associated symmetries of the kagome lattice. Some kagome materials possess multiple degrees of freedom, enabling the study of the interplay between competing orders like superconductivity, magnetism, charge density wave (CDW), and topological phases among others. This thesis centers on neutron scattering experiments of the antiferromagnetically ordered (AFM) kagome metal FeGe that was recently found to host CDW order — a first in a magnetically ordered kagome material. A variety of supplemental experimental techniques were performed to aid in the characterization of various properties in FeGe including Raman spectroscopy, transport, resonant inelastic x-ray scattering (RIXS), scanning tunneling electron microscopy (STEM), muon spin resonance, and angle-resolved photoemission spectroscopy (ARPES). We additionally utilize a post-growth annealing process to tune samples from long-range CDW order to no CDW order repeatedly, acting as a powerful tuning parameter in our studies. Using elastic and inelastic neutron scattering, we uncover two competing magnetic orders at low temperatures, one A-type AFM and one screw-like order, previously believed to be a single double-cone order below TCanting. Additionally, the low temperature screw-like magnetic order is suppressed and enhanced in tandem with CDW order during annealing while the A-type order only exhibits a small enhancement at TCDW in samples with long-range CDW order. Our transport measurements reveal an order of magnitude enhancement of the anomalous Hall effect (AHE) in samples with long-range CDW order and the complete absence of AHE in samples with no CDW order. We find the AHE in FeGe is of intrinsic origin stemming from a large Berry curvature and mirrors the magnitude of the giant AHE in the related superconducting kagome family AV3Sb5 (A= K, Rb, Cs). Through a combination of Raman, ARPES, and neutron Larmor diffraction, we identify lattice instabilities and electronic band shifts at TCanting and TCDW in long-range CDW ordered samples suggesting a strong coupling between spin, lattice, and electronic degrees of freedom in the system. Lastly, we describe a microscopic mechanism for the suppression and enhancement of CDW via annealing through neutron scattering and STEM imaging. We find annealing at high temperatures creates Ge vacancies that inhibit the formation of CDW when spread uniformly throughout the sample. Upon annealing at lower temperatures, the vacancies coalesce to form large stacking faults favorable to CDW formation.
Ground-State Squeezing and Chiral Photonic Crystal Cavities in Ultrastrong Light-Matter Coupling
(2025-04-25) Muralidhar Kulkarni, Kiran; Kono, Junichiro
This thesis investigates quantum light–matter interactions in photonic-crystal cavities, focusing on two complementary projects. The first project develops a theoretical framework for Landau polaritons in terahertz cavities by computing ground-state current-current correlations in a two-dimensional electron gas under ultrastrong coupling (USC). We show that these correlations reveal an intrinsically squeezed ground state—an effect absent in conventional linear spectroscopy. The second project designs, fabricates, and characterizes a one-dimensional chiral photonic-crystal cavity that breaks time-reversal symmetry. The cavity consists of a silicon layer sandwiched between lightly doped indium antimonide (InSb) wafers, exploiting InSb’s low carrier mass and magnetoplasma nonreciprocity to support a single circularly polarized mode at 0.67 THz under a 0.3 T magnetic field, achieving a quality factor above 200. Systematic experiments—varying temperature, magnetic field, and polarization—together with simulations, confirm the cavity’s nonreciprocal behavior and robust mode confinement. Altogether, these studies deepen our understanding of USC-induced quantum effects in photonic cavities and introduce a versatile platform that enables precise manipulation of material properties by breaking key symmetries.
Understanding and Suppressing Degradation in Mixed-Dimensional Halide Perovskites
(2025-04-25) Metcalf, Isaac W; Mohite, Aditya; Ajayan, Pulickel
A new theory of 2D and 3D-2D perovskite degradation in humid air is demonstrated, introducing the novel concept of an A-site cation – spacer cation – PbI2 – H2O quaternary phase diagram. The degradation pathway of a given perovskite is determined by the path charted across the phase diagram as water is introduced to the three-component (n=1) or four-component (n>1) system, often moving through several multi-phase regions that include hydrate phases as nodes. Because of the tendency of organic cations to escape the sample through solvation with the ambient humid air, the effective composition of the system tends towards the PbI2 side of the phase diagram and away from the organic side over time, while also increasing in H2O content. With this degradation picture in mind, a new set of design principles can be conceptualized focusing on suppressing the formation of hydrates and other “avoidable” degradation phases besides PbI2. This new conceptual approach has been explored using the series of linear alkylammonium 2D perovskite spacer cations from methylammonium to octylammonium, and the linear alkyldiammonium spacer cations from butyldiammonium to decyldiammonium, both of which form series of quasi-1D ribbon-like hydrate phases. The new conceptual framework can explain the lower moisture stability of Dion-Jacobson phase 2D perovskites, and also explains the various pathways seen for degradation of higher-n-value 2D and mixed 3D-2D perovskites in humid air. I exploit the quantized ribbon length of the hydrate structure to engineer spacer cations so as not to fit within the organic site of any ribbon-like hydrate, reducing the number of “avoidable” phases in the phase diagram and as a result improving the moisture stability.
Flash Joule Heating for Materials Production
(2025-04-25) Eddy, Lucas; Tour, James M
Flash Joule heating has been widely used as an ultrafast, scalable, and versatile synthesis method, most prominently in the synthesis of flash graphene and other carbon materials. Whereas most chemical synthesis methods transfer heat through a medium into which most heat is lost, flash Joule heating reactions utilize the target feedstock itself as the heating medium, enabling near optimal heating efficiency and consequently extremely high heating rates. Herein, I present an overview of the use of flash Joule heating for materials production, including graphene, graphite, carbon nanotubes, doped graphene, silicon carbide, and p-block metal dichalcogenides. I present different engineering and reaction techniques to facilitate the kilogram-scale production of these materials while performing life cycle assessments and technoeconomic analyses of these processes. I further highlight the impact that the passage of electrical current through the reactant feedstock has on the mechanics of the flash Joule heating technique, finding that this phenomenon can reduce reaction activation energy. I finally discuss the historical foundations of graphene production and provide evidence that Thomas Edison may have synthesized graphene as early as 1879.
Fine-Grained Paging Mechanism for Offloading-Reloading Tensor for LLM
(2025-04-25) Yao, Jiawen; Jermaine, Christopher M.
The rapid growth in the size and complexity of large language models has imposed severe challenges on memory management, particularly when these models are deployed on GPUs with limited memory. This thesis introduces a fine-grained paging mechanism that dynamically offloads and reloads tensors at the granularity of individual operations, thereby mitigating out-of-memory (OOM) issues during the inference and prefill phase of transformer-based models. Instead of traditional static, layer-based offloading methods, the proposed approach uses compile-time, simulationbased memory allocation to optimize GPU memory usage, making runtime possible under severe memory constraints. This work is based off of the Einsummable system, a framework that represents tensor computations using Einstein summation notation. Einsummable transforms high-level mathematical specifications into an optimized execution pipeline through a series of intermediate representations, notably the taskgraph and the memgraph. The taskgraph captures the data dependencies and operational flow of tensor computations, while the memgraph extends this representation by incorporating detailed memory location information and managing offload-reload operations. The transformation from taskgraph to memgraph is achieved through a simulated execution process — the core of this thesis — that relies on two key components: an allocation horizon, which pre-allocates memory for future operations, and an execution horizon, which tracks the simulated execution progress of the computation. A key contribution of this thesis is the design and implementation of specialized memory allocation routines—simMalloc, simMallocForceReld, and simMallocOffld. These routines not only allocate memory for tensor outputs but also manage dependencies by inserting offload and reload nodes into the memgraph whenever GPU memory resources depletes. By leveraging full knowledge of the simulated execution order, our offload-reload heuristic selects tensors for offloading based on their computed reuse distance, thereby deferring memory transfers until they are most convenient. This future-aware strategy mitigates the frequency and impact of memory transfers compared to reactive approaches, enabling a finer control over GPU memory usage. Extensive experimental evaluations were conducted using two configurations of NVIDIA GPUs — Tesla P100 and V100 — to benchmark the performance of the proposed system against state-of-the-art techniques such as ZeRO-Inference. The evaluation focused on the prefill stage of inference in LLaMA models with 7B and 65B parameters, a phase known to be particularly memory-bound. The results demonstrate that the fine-grained paging mechanism supports a broader range of configurations, successfully executing inference tasks across varying batch sizes and sequence lengths. While the finer granularity of tensor-level management introduces some communication overhead due to more frequent offloading and reloading, the overall improvements in memory utilization and reduction in OOM errors outweigh these costs. In summary, this thesis makes a contribution to the field of deep learning by addressing the critical challenge of GPU memory constraints through a fine-grained paging mechanism. Future work will explore further optimizations to reduce communication overhead, overall computation latency, and GPU RAM utilization.
Stochastic Assignment with Expiration
(2025-04-25) Shapoval, Boris Alexandrovich; Perez-Salazar, Sebastian
This thesis introduces a capacitated online stochastic bipartite matching problem, where offline nodes may be matched multiple times and expire at unknown stochastic times. This problem is PSPACE hard; thus we first focus on the subproblem where each offline node can be matched at most once and aim to develop algorithms that achieve large expected overall values from the matchings. A decision maker (DM) must balance obtaining a matching reward now and keeping enough possibilities for the future with possible expirations. Since this problem is intractable, we first provide a compact linear program (LP) formulation that upper bounds the expected value of an optimal algorithm. Based on this LP, we design a polynomial-time algorithm that guarantees an expected value of at least a $1 - 1/e$ fraction of the optimal expected value. We demonstrate the tightness of our LP-based analysis by providing tight integrality gaps as well as worst-case instances. Returning to the capacitated problem, we provide another LP relaxation. We generalize our previous algorithms to evaluate their numerical performance on the harder, capacitated problem. We observe that some natural ideas do not generalize, while others seem to remain competitive.
Embargo
Weathering Inequality: The Racial Stratification of Relocation as Climate Adaptation Following Major Environmental Hazards
(2025-04-25) Priest, Anthony Alexander; Elliott, James R
Climate change is intensifying environmental hazards, exacerbating inequalities, and making relocation a critical, yet potentially unequal, adaptation strategy. Existing research, often limited by aggregate data, overlooks how racial stratification shapes community- and household-level mobility post-disaster. This research addresses this gap by examining nuanced relocation patterns following major hurricanes, using unique longitudinal consumer data combined with high resolution flood and wind hazard metrics. This research involves three distinct analyses. First, examining community relocation rates following Hurricane Harvey's flooding in Harris County, TX reveals persistently higher relocation occurring only in majority Black, Hispanic, or Asian block groups facing extreme flooding. Second, investigating whether post-Harvey movers relocated to areas with lower future flood risk finds that the capacity to reduce risk is significantly stratified by the racial composition of households’ origin communities, even following severe impacts. Third, broadening the focus beyond a single event, an analysis of household relocation and instability following wind impacts from four major hurricanes demonstrates that storm intensity effects are moderated by the racial composition of households’ origin communities, creating divergent and potentially unequal pathways. Collectively, these chapters demonstrate that vulnerability and adaptive capacity – encompassing community-level relocation rates, household ability to reach safer destinations, and patterns of relocating versus instability within individual households – are deeply shaped by racialized spatial inequalities. Findings challenge narratives of relocation as a universally accessible or effective adaptation solution, revealing how racial inequality structures mobility patterns and outcomes across different hazard types and geographic scales, thereby underscoring the urgent need for equitable climate adaptation.
Embargo
Personalizing Assessment of Motor Impairment for Stroke Rehabilitation
(2025-04-25) Rice, Elijah; O'Malley, Marcia K
Motor impairment assessments of stroke are used by therapists to track recovery and prescribe treatment protocols that optimize rehabilitative outcomes. For outpatient-based stroke rehabilitation, lengthy administration times of traditional clinical assessments limit associated benefits and preclude additional therapist-directed rehabilitation that improves outcomes. Robotic and sensor-instrumented systems provide an objective method of assessment that offers additional resolution via measurement of kinematic quantities of movement. Prior research has validated assessment automation with such systems, but has neglected automating assessment using systems that are independently usable by stroke survivors. In this thesis, we analyze the effect of range of motion on a common assessment metric, movement smoothness, which gauges motor coordination as a facet of motor impairment. Based on our findings, we present guidelines for implementation of movement smoothness assessment that preserves task construct validity. A novel device, the FlexWrist, is presented as a usability-focused, glove-based flex sensor system for recording home exercise movement of stroke survivors' hemiparetic wrist and hand.
Components of the Emerton-Gee Moduli Stack of Galois Representations for GL2 and A Graph-Theoretic Approach to Computing Selmer Groups of Elliptic Curves over Q(i)
(2025-04-25) Savoie, Ben; Levin, Brandon
Let K be a finite unramified extension of Q_p with p ≥ 5. In the first part of this thesis, we study the local geometry of the irreducible components in the reduced part of the Emerton–Gee stack for GL_2, which serves as a moduli space for two-dimensional mod p representations of Gal(K/K). We determine precisely which irreducible components are smooth, which are normal, and which have Gorenstein normalizations. We prove that the normalizations of these components admit smooth–local covers by Cohen-Macaulay and resolution-rational varieties, which are generally not Gorenstein. Finally, we determine the singular loci in the components, providing insights which up- date expectations about the conjectural categorical p–adic Langlands correspondence. In the second part of this thesis, we introduce a graph-theoretic algorithm to compute the φ-Selmer group of the elliptic curve E_b : y^2 = x^3 + bx defined over Q(i), where b ∈ Z[i] and φ is a degree 2 isogeny of E_b. We begin by associating a weighted graph G_b to each curve E_b, whose vertices correspond to the odd Gaussian primes dividing b. The weights on the edges connecting these vertices are determined by the quartic residue symbols between these primes. We then establish a bijection between the elements of the φ-Selmer group of E_b and certain partitions of the graph G_b. This correspondence provides a linear-algebraic interpretation of the φ-Selmer group through the Laplacian matrix of G_b. Using our algorithm, we explicitly construct several subfamilies of elliptic curves E_b over Q(i) with trivial Mordell–Weil rank. Furthermore, by combining our method with Tao’s Constellation Theorem for Gaussian primes, we prove the existence of infinitely many elliptic curves E_b over Q(i) with rank exactly 2. Additionally, we show that for each pair of rational twin primes (p, q), the curve E_{pq} considered over Q(i) has rank either 2 or 4, with the rank exactly 2 when p ≡ 5 mod 8. Lastly, we show that for each rational prime of the form p = a^2 + c^4 (of which there are infinitely many), the elliptic curve E_{-p} over Q(i) has rank either 2 or 4, with rank exactly 2 if p ≡ 5 or 9 mod 16.

Browse

Recent Submissions