80264 research outputs found
Sort by
Diamond Surface Engineering and Characterization: Advancing Toward High Quality Epilayer Material for Electronic Applications
No full textDiamond, with its ultra-wide bandgap (~5.5 eV), exceptional thermal conductivity, and high breakdown voltage, is a material that promises advancements in next-generation electronic and quantum applications. Progress in chemical vapor deposition (CVD) techniques have enabled diamond film growth for devices, such as radio frequency (RF) transistors, with enhanced power and frequency capabilities. However, producing large-area, high-quality diamond wafers remains challenging, particularly due to the need for smooth, low-defect substrates and control over defect propagation during homoepitaxial growth.
This dissertation encompasses multiple studies aimed at addressing these limitations through surface engineering, advanced characterization, and materials integration strategies. The first of the two major projects optimized polycrystalline diamond (PCD) substrates using inductively coupled plasma reactive ion etching (ICP-RIE) to achieve smooth, chemically functionalized surfaces with minimal etch pit formation. Among several chemistries, CF4/O2 plasma was identified as particularly effective, producing uniform, fluorinated surfaces with enhanced hydrophobicity, as confirmed by contact angle measurements up to 88.4°. X-ray photoelectron spectroscopy (XPS) and time-of-flight secondary ion mass spectrometry (ToF-SIMS) were used to analyze the depth profile and bonding environments of fluorine-terminated surfaces.
The second major project focused on the characterization of defects in commercially available single-crystal diamond (SCD) substrates. High-resolution X- ray diffraction (HRXRD), including rocking curve (XRC) and reciprocal space map (RSM) analysis, was used to quantify dislocation densities across various substrate types (HPHT, Regular-grade CVD, Electronic-grade CVD, and Heteroepitaxial CVD). Building on diffuse scattering analysis principles previously applied to GaN, these theoretical frameworks were adapted to analyze defect-induced broadening in diamond. A robust and versatile Python-based analysis tool was developed to extract both edge and screw dislocation densities, customizable to the specific scan geometry and reflection type used in the measurement. This tool has also been successfully extended to other wide-bandgap materials, including GaN, allowing for consistent and accurate defect quantification across various crystal systems. Validation through hydrogen plasma etching and scanning electron microscopy (SEM), revealed correlation between etch pit density and the calculated dislocation density values.
Additionally, this dissertation explores two emerging application areas of diamond materials. First, engineered diamond surfaces with controlled wettability and surface charge were used to study bacterial adhesion behavior in Caulobacter crescentus and Escherichia coli, demonstrating selective binding that can pave the
way for spatially controlled biosensing applications. Second, preliminary efforts were made toward optimized strategies of transferring and integrating two- dimensional (2D) materials onto diamond substrates, a critical step in creating hybrid platforms for devices.
Together, these projects present significant advancements in diamond substrate preparation, defect analysis, and interface engineering, laying the groundwork for the next generation of high-performance diamond-based technologies
Sample-to-answer detection of high-risk HPV mRNA expression for early cervical cancer prevention
No full textCervical cancer is preventable if caught and treated early, but due to lack of adequate resources for timely diagnosis and treatment, approximately 350,000 women died from cervical cancer in 2022, 90% of whom lived in low and middle-income countries. mRNA from high-risk human papillomaviruses (hrHPV) is a highly specific biomarker for cervical cancer risk due to its association with infections most likely to progress to cancer. In low-resource settings (LRS), however, where most cervical cancer cases occur, available methods for mRNA detection are inaccessible because they require expensive instrumentation and highly trained users. To address this, in this thesis, I describe the development of a minimally instrumented assay for detecting HPV16, HPV18 and HPV45 mRNA from crude cervicovaginal samples. An extraction-free sample preparation method is developed and optimized in which exfoliated cervical cells are lysed and treated with a DNA-degrading enzyme to remove DNA so only mRNA is available for detection. Following enzyme deactivation, crude lysate is added directly to HPV16, HPV18, and HPV45 reverse transcription recombinase polymerase amplification (RT-RPA) exo assays designed to amplify mRNA encoding the E7 oncoprotein. A β-actin RT-RPA assay is used as a cellular control. The assays are run on a low-cost fluorimeter capable of detecting real-time signal produced by a sequence-specific fluorescent probe for each assay. Prospectively collected patient cervicovaginal samples are also tested using these methods, and characterized for their nucleic acid and blood content, which informs further optimization and refinement of the sample preparation method. With traditional RNA extraction, all assays have a clinically relevant limit of detection of 100 mRNA copies per reaction with a linear correlation between starting quantity and time to amplification, and all were highly specific. The extraction-free sample preparation method successfully degrades cellular DNA, and all assays detect their specific mRNA target in a clinically-relevant range of cell concentrations. Testing of patient cervicovaginal samples showed agreement between RT-qPCR gold standard and RT-RPA results. In summary, by coupling extraction-free RNA sample preparation with isothermal amplification, this assay has the potential to bring more specific cervical cancer detection to LRS. Future work will focus on further testing clinical samples, simplifying the assay workflow, modifying reagents for stability during transportation, and conducting a pilot study testing patient cervicovaginal samples in an LRS
An Adaptive Surrogate Model Refinement (ASMR) Framework for Simulation and Optimization of Dynamical Systems
Full text linkThis thesis develops a framework for solving simulation and optimization problems governed by dynamic equations with expensive black-box functions using adaptively refined surrogate models. This adaptive surrogate-based approach makes dynamic optimization problems with expensive black-box models tractable while appropriately managing solution quality through iterative improvement of the surrogates. The surrogate models are computed from samples of the expensive black-box functions, and it is assumed that pointwise estimates for the error between the sample-based surrogate models and the black-box functions are available. Unlike existing approaches, the approach in this thesis does not require parametrized surrogate models and is purely deterministic. The sensitivity of the trajectory with respect to the surrogate models is computed by performing a parameter-free sensitivity analysis with respect to perturbations of a function. This sensitivity information is combined with error estimates for the sample-based surrogate models to determine new points at which to sample the expensive black-box functions to generate improved surrogate models as needed. It is shown that interpolation in reproducing kernel Hilbert spaces is one way to construct surrogate models and obtain error estimates that satisfy the requirements of the proposed adaptive surrogate-based approach. The proposed framework is used to simulate and optimize the trajectory of a notional hypersonic vehicle whose lift, drag, and moment coefficients are given by black-box functions. Numerical results indicate that the proposed model refinement approach significantly enhances solution quality at a low computational cost and reliably predicts the improvement in solution quality when the model is refined
Single- and Multi-Sensory Haptic Feedback for Prosthetics Applications
No full textThe lack of sensory feedback associated with limb loss is a major impediment to the control, embodiment, and overall benefit of prosthetic devices. Haptic devices, which provide touch sensations such as vibration or squeeze, can help bridge the gap between user and device. Haptic devices come in many forms and configurations, from single vibrotactors to complex multi-sensory devices. Engineers must consider key issues of device design when approaching these systems, primary among them being the user's perceptual accuracy of cues conveyed by the device and the practical benefit the device provides to the user. As a first step towards answering these questions, this thesis presents a detailed review of multi-sensory device elements and design factors. Using this information, two specific aspects of feedback perceptual accuracy and benefit are investigated. In the first, a lower-limb vibrotactile feedback system is used to determine that less tactor overlap produces more distinguishable feedback for sequential tactile signals. In the second, a virtual prosthesis simulator is used to determine that, in the absence of vision, single- and multi-sensory feedback have a benefit to user performance of a common assessment task. This thesis details the engineering methodology and scientific experiments used in answering these questions, and in doing so seeks to illuminate some of the possible benefits and best practices to be utilized by engineers in designing feedback devices for prosthetics applications
Towards Efficient Knowledge Graph Generation From Textbooks: A Dual Framework Approach
Full text linkThe recent proliferation of LLMs necessitates a strategy for addressing these models' deleterious shortcomings: hallucination, and lack of explainability. Knowledge graphs (KGs) have gained attention as a potential solution to these problems, as they can serve as a traceable, factual database for LLMs; however, constructing high-quality KGs efficiently remains a challenge.
To address these challenges, this thesis proposes Words2Wisdom, a logic-informed, LLM-based framework for generating quality KGs from textbooks. Words2Wisdom creates expressive KGs by leveraging the structure of propositional logic, and ensures accurate fact representation, demonstrating knowledge validity (precision) greater than 95% when using the GPT-4o model in a few-shot environment. Our results suggest targeted fine-tuning and model specialization can further enhance KG quality.
Furthermore, this thesis examines whether LLMs are able to assess the quality of KGs. We introduce Libra, a framework establishing a novel KG evaluation protocol for validating KGs against textbook sources. Preliminary results show high observed agreement between Libra and human experts, suggesting that KG construction and evaluation and can indeed be effectively automated, paving the way for future research on the role of LLMs in hallucination mitigation
Secure Circuits: Efficient hardware countermeasures against physical side-channel attacks
No full textThe adoption of Artificial Intelligence (AI) and IoT devices has seen unprecedented growth in recent times. AI engines demand more processing capabilities while simultaneously handling sensitive user information and proprietary IP. Concurrently, IoT devices generate vast amounts of sensitive data, necessitating robust security measures to safeguard against breaches and ensure data privacy. Hence, there is a definitive need for faster and more secure systems. We present two key implementations that address the problem. Firstly, we introduce MBSNTT (Multi-Bit Serial Number Theoretic Transform Accelerator) to accelerate the NTT operation in Homomorphic Encryption. In this implementation, we apply processing-in-memory techniques to the NTT operation, thereby achieving high parallelism. In the second design, namely HDCIM (Hybrid Security-based Digital Compute in Memory Accelerator for Protected Inference), we propose hybrid security by applying mathematical masking techniques to neural network operations and closing the security gaps with low-overhead physical security measures
Defaunation Increases Clustering and Fine-Scale Spatial Genetic Structure in a Small-Seeded Palm Despite Remaining Small-Bodied Frugivores
No full textAnthropogenic pressures such as hunting are increasingly driving the localised functional extinctions of large- and medium-sized wildlife in tropical forests, a phenomenon broadly termed ‘defaunation’. Concurrently in these areas, smaller-bodied species benefit from factors such as competitive release and increase in numbers. This transformation of the wildlife community can impact species interactions and ecosystem services such as seed dispersal and seed-mediated geneflow with far-reaching consequences. Evidence for negative genetic effects following defaunation is well-documented in large-seeded plants that require large frugivores for long-distance seed dispersal. However, how defaunation affects plants with small or medium-small seeds (< 1.5 cm), which tend to be consumed and dispersed by frugivorous mutualists of a range of body sizes and responses to anthropogenic threats, is not well understood. To better understand defaunation's impacts on tropical plant communities, we investigated spatial and genetic patterns in a hyperabundant medium-to-small-seeded palm, Euterpe precatoria in three sites with different defaunation levels. Results indicate that defaunation is associated with higher fine-scale spatial genetic structure among seedlings and increased spatial clustering within seedling cohorts and between seedlings and conspecific adults, as well as a reduction in nearest-neighbour distances between seedlings and conspecific adults. There were no clear effects on inbreeding or genetic diversity. However, we caution these trends may indicate that defaunation reduces seed dispersal services for species previously presumed to be robust to deleterious effects of losing large frugivores by virtue of having smaller seeds and broad suites of dispersal agents, and negative downstream effects on genetic diversity could occur
Hybrid Protein-Polymer Materials and their Applications
No full textThe conjugation of biomacromolecules, such as proteins, to polymeric materials has many applications. These applications are as varied as the formation of protein−polymer conjugates used in therapeutic treatments to applications in sensors, biocatalysts, and tools for separation of biomolecules. The diverse range of hybrid materials available necessitates a diverse range of corresponding methodologies to support their construction. Among the key focuses of this thesis includes methodologies for the development of protein−polymer biomaterials and their wide-ranging applications.
The first chapter is a review of methods of site-selective protein conjugation with polymers via naturally encoded sequences. This review covers a variety of methodologies for protein−polymer conjugation moving from non-specific methods to more sophisticated, site-selective methods. The second chapter will review the structure and applications of protein-biomaterials, such as those conjugated to a nano-object or immobilized to a solid substrate. The second chapter will also cover the enhanced properties of novel materials at the interphase between nano, surface, and biological chemistry.
The development of a boronic acid resin for the selective immobilization of canonically encoded (pyroglutamate-histidine-tagged) proteins is covered in the third chapter. The fourth chapter demonstrates a unique application of these protein−polymer biomaterials as a template in the synthesis of fluorescent copper nanoclusters. The fifth chapter will focus on efforts towards the controlled release of boronic acid-based therapeutics by tailoring boronate ester hydrolysis kinetics. Finally, the sixth chapter will showcase the antibacterial activity of capacitively coupled plasma from laser-induced graphene and the experiments elucidating the molecular mediator of bacterial cell death
Houston and Harris County Disaster Preparedness and Attitudes Leading Up to the 2024 Hurricane Season
No full textInformed by the existing literature, this research study will examine the current levels of both perceived and actual disaster preparedness among Houston and Harris County residents, as well as contextual and socio-cognitive predictors of each. The findings can be used to inform actions aimed at improving disaster preparedness and response efforts in the Greater Houston area
Engineering Glucose Enzymes Through Domain Insertion for Adaptive Bioelectronic Sensors
No full textBiosensors are essential in diagnostics, monitoring, and therapeutics. A major example is the glucometer, which effectively utilizes glucose-oxidizing enzymes to generate accurate electrical signals that report blood sugar levels. This bioelectrochemical sensor’s affordability, manufacturability, and suitability for patient-side use make glucose enzymes highly appealing for broader sensing applications. Although existing studies have explored mutagenizing glucose redox enzymes to enhance their stability and activity, significant obstacles remain in repurposing these enzymes to detect other biomarkers. These challenges stem from an incomplete understanding of glucose enzyme design and the limited effectiveness of current protein engineering approaches.
This thesis addresses these challenges by using pyrroloquinoline quinone glucose dehydrogenase (PQQ-GDH) as a robust platform for glucose-dependent oxidoreductase applications. Through comprehensive methods developed to identify structural elements crucial to its function, this work demonstrates the repurposing of PQQ-GDH to produce electrochemical output for non-glucose analytes. Additionally, a high-throughput screening system is introduced to accelerate the development of a broad range of bioelectronic sensors. In Chapter 2, we integrate small peptide sequences into PQQ-GDH to investigate the structure-sequence-function relationships at various structural levels. In Chapter 3, we engineer PQQ-GDH conformational switches to create electronic sensors capable of detecting cancer therapeutics in blood samples, pushing the boundaries of traditional glucose sensing. In Chapter 4, we establish a high-throughput selection system for glucose enzyme variants by manipulating glucose metabolism and NADPH regeneration in E. coli through targeted knockouts.
Our research explores multiple strategies for functionalizing PQQ-GDH to enhance bioelectronic diagnostics. These findings provide critical insights into how the structure and sequence of PQQ-GDH influence its function—particularly at the active site and dimerization interface, which are essential for enzyme activity and stability. When integrated onto electrode interfaces, our functionalized PQQ-GDH variants demonstrate a significant electrical response to the cancer therapeutic 4-hydroxytamoxifen in blood. This advancement lays a solid foundation for real-time, point-of-care diagnostics in therapeutic monitoring. Additionally, our innovative growth complementation assay enriches enzyme variants in direct proportion to their activity levels, establishing a novel selection method for variants that exhibit superior performance. These contributions advance biosensing technologies and significantly expand the application scope of bioelectrochemical systems. We are paving the way for reliable point-of-care diagnostic devices and therapeutic monitoring platforms that promise to transform future healthcare solution