Upcoming Doctoral Dissertations

"The goal of this project was to determine if an industrial waste containing a high concentration of nitrate could be discharged to a municipal wastewater treatment facility for removal of the nitrate. The hypothesis was that this high-nitrate industrial waste could be denitrified in a municipal water resource recovery facility (WRRF) using the influent chemical oxygen demand (COD) as the carbon source, as well as endogenous carbon. " Read more...

"Nitrogen is a fundamental component of biomolecules essential for the structure and function of living organisms. Nitrogen-containing species play critical roles in various scientific, biological, pharmacological, and environmental contexts. Understanding the diverse functions and behaviors of these compounds is crucial in advancing fields ranging from biomedicine to environmental science. This work includes four detailed quantum chemical studies that collectively contribute to our understanding of metal-mediated reactions and interactions involving essential nitrogen containing species" Read more...

"Lithium niobate stands as a cornerstone material in the optics and photonics landscape, distinguished by its superlative attributes in both linear and nonlinear optics. Its wide transparency window, relatively high quadratic nonlinear coefficient, and strong electro-optic response make it an exceptional candidate for various optical applications. However, the development of thin film lithium niobate (TFLN) as a nanophotonic platform has not been as
smooth as its counterparts like silicon nitride or other silicon family materials. This slower pace is primarily attributed to the challenges associated with its material processing and nanofabrication. Over the past eight years, our laboratory has contributed to breakthroughs in TFLN nanofabrication techniques. We have successfully fabricated nanophotonic TFLN waveguides characterized by ultra-low propagation losses, consistently measured at less than 0.5 dB/cm." Read more...

"The emergence of next-generation communications and computing paradigms such as edge computing has made distributed machine learning technologies like federated learning increasingly popular in emerging applications. Unlike centralized learning, which trains models at the central server, federated learning enhances data privacy by allowing participants (i.e., users) to train models locally and only send the trained model parameters or gradients to the remote server without revealing their private datasets. However, recent research has shown that attackers can infer training examples held by users by disclosing local model parameters or gradient updates. To protect local models against disclosure, only the user should know their own local model, while the global model will be disclosed to all users. This variant of federated learning, which protects users' private data and models, is known as PrivacyPreserving Federated Learning (PPFL). However, privacy comes at a cost, and two major challenges in PPFL are efficiency and security. In this research, the author approaching the PPFL in following aspects: (1) Enable backdoor attack detectability in PPFL; (2) Improving computational and communicational efficiency in PPFL and its variant: Federated Distillation; (3) Exploring the vulnerability of PPFL trained black-box model" Read more...

"Despite continuing technological advances in neural interfaces, challenges with the tissue reaction to chronic brain implants remain. The foreign body response of the brain to an inserted electrode induces sustained inflammation due to the persistent presence of the device and promotes local neurodegeneration. The use of brain stimulation to minimize inflammation is a burgeoning field, and so far, most interventions have been non-invasive. Here, we investigated an invasive method of inflammation-minimizing stimulation for application in individuals with neural electrode implants, e.g., deep brain stimulation and cortical neural prosthetics. We hypothesized that invasive stimulation at the gamma frequency band acutely increases microglial recruitment and chronically reduces astrogliosis at the tissue-electrode interface. " Read more...

"Novel quantum materials combined with controlled light-matter interaction are promising to enable implementation of desired tasks in quantum information processing and quantum sensing directly on-chip. Specifically, 2D transition metal dichalcogenides (TMDC) like WSe2 stand out as a host for 0D quantum emitters (QE), which can be deterministically produced through site-controlled stressor substrate engineering. To effectively control the light-matter interaction, we have developed angular-resolved detection techniques to experimentally determine the dipole orientation of QE inform of intralayer and interlayer excitons in WSe2. With the dipole orientation knowledge gained, we successfully designed and fabricated plasmonic nanocavity devices coupled with interlayer QE, demonstrating brightening of optically forbidden transitions that are further enhanced through the Purcell effect." Read more...