Matching Items (79)
Filtering by
- Genre: Doctoral Dissertation
- Creators: Davulcu, Hasan
Description
Cubic boron nitride (c-BN), hexagonal boron nitride (h-BN), and semiconducting diamond all have physical properties that make them ideal materials for applications in high power and high frequency electronics, as well as radiation detectors. However, there is limited research on the unique properties and growth of c-BN or h-BN thin films. This dissertation addresses the deposition of c-BN via plasma enhanced chemical vapor deposition (PECVD) on boron doped diamond substrates. In-Situ X-ray photoelectron spectroscopy (XPS) is used to characterize the thickness and hexagonal to cubic ratio of boron nitride thin films. The effects of hydrogen concentration during the deposition of boron nitride are investigated. The boron nitride deposition rate is found to be dependent on the hydrogen gas flow. The sp2 to sp3 bonding is also found to be dependent on the hydrogen gas flow. Preferential growth of h-BN is observed when an excess of hydrogen is supplied to the reaction, while h-BN growth is suppressed when hydrogen flow is reduced to be the limiting reactant. Reduced hydrogen flow is also observed to promote preferential growth of c-BN.
The hydrogen limited reaction is used to deposit c-BN on single crystal (100) boron-doped diamond substrates. In-situ ultra-violet photoelectron spectroscopy (UPS) and XPS are used to deduce the valence band offset of the diamond/c-BN interface. A valence band offset of -0.3 eV is measured with the diamond VBM above the VBM of c-BN. This value is then discussed in context of previous experimental results and theoretical calculations.
Finally, UPS and XPS are used to characterize the surface states of phosphorus-doped diamond. Variations within the processing parameters for surface preparation and the effects on the electronic surface states are presented and discussed.
ContributorsBrown, Jesse (Author) / Nemanich, Robert J (Thesis advisor) / Alarcon, Ricardo (Committee member) / Lindsay, Stuart (Committee member) / Zaniewski, Anna (Committee member) / Arizona State University (Publisher)
Created2021
Description
Distributed self-assessments and reflections empower learners to take the lead on their knowledge gaining evaluation. Both provide essential elements for practice and self-regulation in learning settings. Nowadays, many sources for practice opportunities are made available to the learners, especially in the Computer Science (CS) and programming domain. They may choose to utilize these opportunities to self-assess their learning progress and practice their skill. My objective in this thesis is to understand to what extent self-assess process can impact novice programmers learning and what advanced learning technologies can I provide to enhance the learner’s outcome and the progress. In this dissertation, I conducted a series of studies to investigate learning analytics and students’ behaviors in working on self-assessments and reflection opportunities. To enable this objective, I designed a personalized learning platform named QuizIT that provides daily quizzes to support learners in the computer science domain. QuizIT adopts an Open Social Student Model (OSSM) that supports personalized learning and serves as a self-assessment system. It aims to ignite self-regulating behavior and engage students in the self-assessment and reflective procedure. I designed and integrated the personalized practice recommender to the platform to investigate the self-assessment process. I also evaluated the self-assessment behavioral trails as a predictor to the students’ performance. The statistical indicators suggested that the distributed reflections were associated with the learner's performance. I proceeded to address whether distributed reflections enable self-regulating behavior and lead to better learning in CS introductory courses. From the student interactions with the system, I found distinct behavioral patterns that showed early signs of the learners' performance trajectory. The utilization of the personalized recommender improved the student’s engagement and performance in the self-assessment procedure. When I focused on enhancing reflections impact during self-assessment sessions through weekly opportunities, the learners in the CS domain showed better self-regulating learning behavior when utilizing those opportunities. The weekly reflections provided by the learners were able to capture more reflective features than the daily opportunities. Overall, this dissertation demonstrates the effectiveness of the learning technologies, including adaptive recommender and reflection, to support novice programming learners and their self-assessing processes.
ContributorsAlzaid, Mohammed (Author) / Hsiao, Ihan (Thesis advisor) / Davulcu, Hasan (Thesis advisor) / VanLehn, Kurt (Committee member) / Nelson, Brian (Committee member) / Bansal, Srividya (Committee member) / Arizona State University (Publisher)
Created2022
Description
With the bloom of machine learning, a massive amount of data has been used in the training process of machine learning. A tremendous amount of this data is user-generated data which allows the machine learning models to produce accurate results and personalized services. Nevertheless, I recognize the importance of preserving the privacy of individuals by protecting their information in the training process. One privacy attack that affects individuals is the private attribute inference attack. The private attribute attack is the process of inferring individuals' information that they do not explicitly reveal, such as age, gender, location, and occupation. The impacts of this go beyond knowing the information as individuals face potential risks. Furthermore, some applications need sensitive data to train the models and predict helpful insights and figuring out how to build privacy-preserving machine learning models will increase the capabilities of these applications.However, improving privacy affects the data utility which leads to a dilemma between privacy and utility. The utility of the data is measured by the quality of the data for different tasks. This trade-off between privacy and utility needs to be maintained to satisfy the privacy requirement and the result quality. To achieve more scalable privacy-preserving machine learning models, I investigate the privacy risks that affect individuals' private information in distributed machine learning. Even though the distributed machine learning has been driven by privacy concerns, privacy issues have been
proposed in the literature which threaten individuals' privacy. In this dissertation, I investigate how to measure and protect individuals' privacy in centralized and distributed machine learning models. First, a privacy-preserving text representation learning is proposed to protect users' privacy that can be revealed from user generated data. Second, a novel privacy-preserving text classification for split learning is presented to improve users' privacy and retain high utility by defending against private attribute inference attacks.
ContributorsAlnasser, Walaa (Author) / Liu, Huan (Thesis advisor) / Davulcu, Hasan (Committee member) / Shu, Kai (Committee member) / Bao, Tiffany (Committee member) / Arizona State University (Publisher)
Created2022
Description
Social media platforms provide a rich environment for analyzing user behavior. Recently, deep learning-based methods have been a mainstream approach for social media analysis models involving complex patterns. However, these methods are susceptible to biases in the training data, such as participation inequality. Basically, a mere 1% of users generate the majority of the content on social networking sites, while the remaining users, though engaged to varying degrees, tend to be less active in content creation and largely silent. These silent users consume and listen to information that is propagated on the platform.However, their voice, attitude, and interests are not reflected in the online content, making the decision of the current methods predisposed towards the opinion of the active users. So models can mistake the loudest users for the majority. To make the silent majority heard is to reveal the true landscape of the platform. In this dissertation, to compensate for this bias in the data, which is related to user-level data scarcity, I introduce three pieces of research work. Two of these proposed solutions deal with the data on hand while the other tries to augment the current data. Specifically, the first proposed approach modifies the weight of users' activity/interaction in the input space, while the second approach involves re-weighting the loss based on the users' activity levels during the downstream task training. Lastly, the third approach uses large language models (LLMs) and learns the user's writing behavior to expand the current data. In other words, by utilizing LLMs as a sophisticated knowledge base, this method aims to augment the silent user's data.
ContributorsKarami, Mansooreh (Author) / Liu, Huan (Thesis advisor) / Sen, Arunabha (Committee member) / Davulcu, Hasan (Committee member) / Mancenido, Michelle V. (Committee member) / Arizona State University (Publisher)
Created2023
Description
The problem of monitoring complex networks for the detection of anomalous behavior is well known. Sensors are usually deployed for the purpose of monitoring these networks for anomalies and Sensor Placement Optimization (SPO) is the problem of determining where these sensors should be placed (deployed) in the network. Prior works have utilized the well known Set Cover formulation in order to determine the locations where sensors should be placed in the network, so that anomalies can be effectively detected. However, such works cannot be utilized to address the problem when the objective is to not only detect the presence of anomalies, but also to detect (distinguish) the source(s) of the detected anomalies, i.e., uniquely monitoring the network. In this dissertation, I attempt to fill in this gap by utilizing the mathematical concept of Identifying Codes and illustrating how it not only can overcome the aforementioned limitation, but also it, and its variants, can be utilized to monitor complex networks modeled from multiple domains. Over the course of this dissertation, I make key contributions which further enhance the efficacy and applicability of Identifying Codes as a monitoring strategy. First, I show how Identifying Codes are superior to not only the Set Cover formulation but also standard graph centrality metrics, for the purpose of uniquely monitoring complex networks. Second, I study novel problems such as the budget constrained Identifying Code, scalable Identifying Code, robust Identifying Code etc., and present algorithms and results for the respective problems. Third, I present useful Identifying Code results for restricted graph classes such as Unit Interval Bigraphs and Unit Disc Bigraphs. Finally, I show the universality of Identifying Codes by applying it to multiple domains.
ContributorsBasu, Kaustav (Author) / Sen, Arunabha (Thesis advisor) / Davulcu, Hasan (Committee member) / Liu, Huan (Committee member) / Xue, Guoliang (Committee member) / Arizona State University (Publisher)
Created2022
Description
Deep neural networks have been shown to be vulnerable to adversarial attacks. Typical attack strategies alter authentic data subtly so as to obtain adversarial samples that resemble the original but otherwise would cause a network's misbehavior such as a high misclassification rate. Various attack approaches have been reported, with some showing state-of-the-art performance in attacking certain networks. In the meanwhile, many defense mechanisms have been proposed in the literature, some of which are quite effective for guarding against typical attacks. Yet, most of these attacks fail when the targeted network modifies its architecture or uses another set of parameters and vice versa. Moreover, the emerging of more advanced deep neural networks, such as generative adversarial networks (GANs), has made the situation more complicated and the game between the attack and defense is continuing. This dissertation aims at exploring the venerability of the deep neural networks by investigating the mechanisms behind the success/failure of the existing attack and defense approaches. Therefore, several deep learning-based approaches have been proposed to study the problem from different perspectives. First, I developed an adversarial attack approach by exploring the unlearned region of a typical deep neural network which is often over-parameterized. Second, I proposed an end-to-end learning framework to analyze the images generated by different GAN models. Third, I developed a defense mechanism that can secure the deep neural network against adversarial attacks with a defense layer consisting of a set of orthogonal kernels. Substantial experiments are conducted to unveil the potential factors that contribute to attack/defense effectiveness. This dissertation also concludes with a discussion of possible future works of achieving a robust deep neural network.
ContributorsDing, Yuzhen (Author) / Li, Baoxin (Thesis advisor) / Davulcu, Hasan (Committee member) / Venkateswara, Hemanth Kumar Demakethepalli (Committee member) / Yang, Yezhou (Committee member) / Arizona State University (Publisher)
Created2022
Description
Diamond as a wide-bandgap (WBG) semiconductor material has distinct advantages for power electronics applications over Si and other WBG materials due to its high critical electric field (> 10 MV/cm), high electron and hole mobility (??=4500 cm2/V-s, ??=3800 cm2/V-s), high thermal conductivity (~22 W/cm-K) and large bandgap (5.47 eV). Owing to its remarkable properties, the application space of WBG materials has widened into areas requiring very high current, operating voltage and temperature. Remarkable progress has been made in demonstrating high breakdown voltage (>10 kV), ultra-high current density (> 100 kA/cm2) and ultra-high temperature (~1000oC) diamond devices, giving further evidence of diamond’s huge potential. However, despite the great success, fabricated diamond devices have not yet delivered diamond’s true potential. Some of the main reasons are high dopant activation energies, substantial bulk defect and trap densities, high contact resistance, and high leakage currents. A lack of complete understanding of the diamond specific device physics also impedes the progress in correct design approaches. The main three research focuses of this work are high power, high frequency and high temperature. Through the design, fabrication, testing, analysis and modeling of diamond p-i-n and Schottky diodes a milestone in diamond research is achieved and gain important theoretical understanding. In particular, a record highest current density in diamond diodes of ~116 kA/cm2 is demonstrated, RF characterization of diamond diodes is performed from 0.1 GHz to 25 GHz and diamond diodes are successfully tested in extreme environments of 500oC and ~93 bar of CO2 pressure. Theoretical models are constructed analytically and inii
Silvaco ATLAS including incomplete ionization and hopping mobility to explain space charge limited current phenomenon, effects of traps and Mott-Gurney dominated diode ???. A new interpretation of the Baliga figure of merit for WBG materials is also formulated and a new cubic relationship between ??? and breakdown voltage is established. Through Silvaco ATLAS modeling, predictions on the power limitation of diamond diodes in receiver-protector circuits is made and a range of self-heating effects is established. Poole-Frenkel emission and hopping conduction models are also utilized to analyze high temperature (500oC) leakage behavior of diamond diodes. Finally, diamond JFET simulations are performed and designs are proposed for high temperature – extreme environment applications.
ContributorsSurdi, Harshad (Author) / Goodnick, Stephen M (Thesis advisor) / Nemanich, Robert J (Committee member) / Thornton, Trevor J (Committee member) / Lyons, James R (Committee member) / Arizona State University (Publisher)
Created2022
Description
In this dissertation, atomic layer processing and surface characterization techniques were used to investigate surface conditions of wide band gap materials, gallium nitride (GaN) and gallium oxide (Ga2O3). These studies largely focused on mitigation and removal of defect formation induced by ions used in conventional plasma-based dry etching techniques. Band bending measured by x-ray photoelectron spectroscopy (XPS) was used to characterize charge compensation at the surface of GaN (0001) and determine densities of charged surface states produced by dry etching. Mitigation and removal of these dry-etch induced defects was investigated by varying inductively coupled plasma (ICP) etching conditions, performing thermal and plasma-based treatments, and development of a novel low-damage, self-limiting atomic layer etching (ALE) process to remove damaged material. Atomic layer deposition (ALD) and ALE techniques were developed for Ga2O3 using trimethylgallium (TMG). Ga2O3 was deposited by ALD on Si using TMG and O2 plasma with a growth rate of 1.0 ± 0.1 Å/cycle. Ga2O3 films were then etched using HF and TMG using a fully thermal ALE process with an etch rate of 0.9 ± Å/cycle. O2 plasma oxidation of GaN for surface conversion to Ga2O3 was investigated as a pathway for ALE of GaN using HF and TMG. This process was characterized using XPS, in situ multi-wavelength ellipsometry, and transmission electron microscopy. This study indicated that the etch rate was lower than anticipated, which was attributed to crystallinity of the converted surface oxide on GaN (0001).
ContributorsHatch, Kevin Andrew (Author) / Nemanich, Robert J (Thesis advisor) / Ponce, Fernando A (Committee member) / Smith, David J (Committee member) / Zhao, Yuji (Committee member) / Arizona State University (Publisher)
Created2021
Description
In this dissertation, the surface interactions of fluorine were studied during atomic layer deposition (ALD) and atomic layer etching (ALE) of wide band gap materials. To enable this research two high vacuum reactors were designed and constructed for thermal and plasma enhanced ALD and ALE, and they were equipped for in-situ process monitoring. Fluorine surface interactions were first studied in a comparison of thermal and plasma enhanced ALD (TALD and PEALD) of AlF3 thin films prepared using hydrogen fluoride (HF), trimethylaluminum (TMA), and H2-plasma. The ALD AlF3 films were compared ¬in-situ using ellipsometry and X-ray photoelectron spectroscopy (XPS). Ellipsometry showed a growth rate of 1.1 Å/ cycle and 0.7 Å/ cycle, at 100°C, for the TALD and PEALD AlF3 processes, respectively. XPS indicated the presence of Al-rich clusters within the PEALD film. The formation of the Al-rich clusters is thought to originate during the H2-plasma step of the PEALD process. The Al-rich clusters were not detected in the TALD AlF3 films. This study provided valuable insight on the role of fluorine in an ALD process.
Reactive ion etching is a common dry chemical etch process for fabricating GaN devices. However, the use of ions can induce various defects, which can degrade device performance. The development of low-damage post etch processes are essential for mitigating plasma induced damage. As such, two multistep ALE methods were implemented for GaN based on oxidation, fluorination, and ligand exchange. First, GaN surfaces were oxidized using either water vapor or O2-plasma exposures to produce a thin oxide layer. The oxide layer was addressed using alternating exposures of HF and TMG, which etch Ga2O3 films. Each ALE process was characterized using in-situ using ellipsometry and XPS and ex-situ transmission electron microscopy (TEM). XPS indicated F and O impurities remained on the etched surfaces. Ellipsometry and TEM showed a slight reduction in thickness. The very low ALE rate was interpreted as the inability of the Ga2O3 ALE process to fluorinate the ordered surface oxide on GaN (0001).
Overall, these results indicate HF is effective for the ALD of metal fluorides and the ALE of metal oxides.
ContributorsMessina, Daniel C (Author) / Nemanich, Robert J (Thesis advisor) / Goodnick, Stephen (Committee member) / Ponce, Fernando A (Committee member) / Smith, David (Committee member) / Arizona State University (Publisher)
Created2021
Description
With the exponential growth of multi-modal data in the field of computer vision, the ability to do inference effectively among multiple modalities—such as visual, textual, and auditory data—shows significant opportunities. The rapid development of cross-modal applications such as retrieval and association is primarily attributed to their ability to bridge the gap between different modalities of data. However, the current mainstream cross-modal methods always heavily rely on the availability of fully annotated paired data, presenting a significant challenge due to the scarcity of precisely matched datasets in real-world scenarios. In response to this bottleneck, several sophisticated deep learning algorithms are designed to substantially improve the inference capabilities across a broad spectrum of cross-modal applications. This dissertation introduces novel deep learning algorithms aimed at enhancing inference capabilities in cross-modal applications, which take four primary aspects. Firstly, it introduces the algorithm for image retrieval by learning hashing codes. This algorithm only utilizes the other modality data in weakly supervised tags format rather than the supervised label. Secondly, it designs a novel framework for learning the joint embeddings of images and texts for the cross-modal retrieval tasks. It efficiently learns the binary codes from the continuous CLIP feature space and can even deliver competitive performance compared with the results from non-hashing methods. Thirdly, it conducts a method to learn the fragment-level embeddings that capture fine-grained cross-modal association in images and texts. This method uses the fragment proposals in an unsupervised manner. Lastly, this dissertation also outlines the algorithm to enhance the mask-text association ability of pre-trained semantic segmentation models with zero examples provided. Extensive future plans to further improve this algorithm for semantic segmentation tasks will be discussed.
ContributorsZhuo, Yaoxin (Author) / Li, Baoxin (Thesis advisor) / Wu, Teresa (Committee member) / Davulcu, Hasan (Committee member) / Yang, Yezhou (Committee member) / Arizona State University (Publisher)
Created2024