YSP Alumni
Alumni Directory
2025 YSP Program Coordinators:
Coordinators: Dmitra Mukasa & Ahmed Othman & Victoria Berry
Participating Labs
| Lab | YSP Students | Title | Abstract |
|---|---|---|---|
| Faculty: Alam, Md Noor E Mentors: Tianyu Yang | Daniel Golmohammadi, Gabriella Jewitt | Cocaine as a Secondary Substance: Effects on Rehab Completion in Opioid Users Final Presentation Final Poster | The opioid epidemic has been a longstanding issue in the United States where almost 1.5 million individuals were diagnosed with opioid use disorder (OUD) in 2022 alone. Opioid use disorder patients are also vulnerable to using other substances, which may create additional challenges in treating their OUD conditions. Therefore, it is necessary that researchers identify challenges that impact OUD treatment rates so that medical professionals can formulate effective treatment strategies for their patients. |
| Faculty: Alshawabkeh, Akram Mentors: Amanda Thomas, Muhammad Fahad Ehsan, Nima Sakhaee, Marieh Arekhi | Shiza Hussain, Kian Nhuch | Regeneration of Activated Carbon Cathodes for Sustainable Water Treatment Final Presentation Final Poster | Water contamination is a major global challenge, particularly affecting underprivileged communities. In these communities, such as Puerto Rico, affordable and sustainable methods of water purification are a necessity, sparking the need for research surrounding the role that activated carbon can play in more efficient water filtration. This research aims to enhance pollutant (methylene blue) adsorption via optimizing the pore size of the Granular Activated Carbon (GAC) cathode. Additionally, the research conducted searched for optimal regeneration periods for carbon cathodes in order to refine sustainability of the product. |
| Faculty: Amirabadi, Mahshid Mentors: Karen Abbaskhanian | Frances Corcell, Joshua Manoj | DC-DC Converter for Solar Powered Phone Chargers Final Presentation Final Poster | The process of converting sunlight directly into electrical power has been widely researched for its applications across various fields. Our research is specifically catered to garnering a more thorough understanding of how photovoltaic cells (solar panels) can contribute to the renewable energy movement as researchers work towards improving their efficiency. The electrical current produced in a photovoltaic (PV) cell is created when sunlight excites and frees up electrons that transfer energy through electron-hole pairs across the cell. However, due to the variable sunlight absorbed by a PV cell, a voltage regulator is necessary to stabilize the output voltage such that it can be used to charge a cell phone. Our research centers individually around investigating and creating different forms of voltage regulators and examining their efficiency while taking into account our desired resistance and power dissipation levels. Throughout this program, we simulated both switching and linear regulators and eventually developed three different variations of a series transistor voltage regulator. These designs aim to support more stable and efficient energy use from solar panels in small-scale applications. |
| Faculty: Caparco, Adam Mentors: Julia Merlin | Charmaine Cao, Cooper Su | Nanoparticles for Sustained Agrochemical Release Final Presentation: Figma | PDF Final Poster | Pesticides are currently the primary solution towards eliminating phytoparasitic nematodes, agriculturally harmful roundworms. Excess treatment creates unfavorable side effects though, such as toxic run off and pesticide poisoning. Previous studies have shown that Tobacco Mild Green Mosaic Virus (TMGMV), a nanoparticle, encapsulates and protects RNA from soil before entering nematodes. Another method of encapsulation and delivery is through ZRC10ZR and EGFPZE protein supraparticles. The purpose of this study is to compare how well two different encapsulation methods flow through soil, as nematodes often reside deep in the ground. Escherichia Coli (E. Coli) was transformed to produce EGFPZE and ZRC10ZR. After protein extraction and purification, supraparticles were formed and their soil mobility was compared to TMGMV nanoparticles. Particles were assessed for their mobility, and TMGMV has a better elution profile than water. |
| Faculty: Davidow, Juliet Mentors: Brianna Aubrey, Erica Niemiec | Helena Brain, Sheila Kastrati | Developmental Changes in Tissue Iron to Examine Dopamine's Role in Adolescent Reinforcement Learning and Memory Final Presentation Final Poster | Adolescence is a time of increased independence, leading to changes in decision making, learning, and memory. Cognitive changes in adolescence are supported by brain development in earlier-developing reward-sensitive regions like the striatum which promote risk-taking and later-developing prefrontal cortical regions which support executive control [1]. These differences make adolescents especially sensitive to reward. To understand and study reward processing in the adolescent brain, we used magnetic resonance imaging (MRI) to measure striatal tissue iron in a developmental sample (N = 105, range 7-24 years, Mean Age = 18.71 years, NFemale = 57). Tissue iron accumulation acts as a proxy for dopamine synthesis, allowing us to study the relationship between dopamine, reinforcement learning, and memory [2]. MRI quality control (QC) is necessary to ensure that data acquisition and preprocessing steps were performed correctly, improving data reliability and signal to noise ratio. Specifically, we checked for alignment between expected and available data, correct segmentation of gray and white matter, accuracy of mapping from subject to standard space, presence of artifacts, and excessive motion. The goal of our research is to explore developmental changes in tissue iron and dopamine’s relationship with reinforcement learning and memory. We hope to gain a better understanding of how adolescents learn and take risks and how dopamine signalling changes across development. This can help inform education methods, improve understanding of peer interaction and identity formation, and mitigate dangerous risky behaviours, such as substance use. |
| Faculty: Erb, Randall Mentors: Echo St. Germain | Noah Dollard, Thanai Papageorgiou | Iron Fuel Cycle for Carbon-Free Coal Replacement Final Presentation Final Poster | The iron fuel cycle is being explored as an energy option for a sustainable future based on carbon free energy generation andstorage. For iron to be a successful alternative more research into the variables which effect iron ignition need be explored. To improve the understanding of how oxidation can change the ignition properties of iron: ✓ Cross sections of samples were prepared for imaging by sanding and polishing pre-oxidized iron powder ✓ The minimum amount of time needed for each step while still achieving usable samples was investigated ✓ Powders collected from various furnace environments were analyzed using ImageJ to display changes in morphology ✓ Python code was written to correct thermocouple temperature readings, considering radiation for more accurate results ✓ Operating procedures for data analysis and sample preparation were created to the increase efficiency of future student training This analysis of over 50 different samples aids in expanding researchers understanding of the sensitivity of iron as a fuel due to the formation of a solid oxide layer. |
| Faculty: Gallaway, Joshua Mentors: Yogeshwaran Agilan | Calli Allaire, Vincent Willits | Mitigating Capacity Loss of Copper Oxide Cathodes in Rechargeable Batteries Final Presentation Final Poster | Developing safe and cost-effective batteries will improve energy storage and capacity for intermittent energy sources. Alkaline Zn-CuO batteries are promising candidates for improving safety and reducing cost. These batteries have a high theoretical capacity and are based on earth-abundant, inexpensive materials. Importantly, it is compatible with aqueous electrolytes, which are non-flammable and offer higher ionic conductivities. However, CuO cathodes were found to have active material dissolution and migration from the electrode into the separator. So, we began to study how to mitigate this dissolution and migration during cycling. In terms of altering electrolytic concentration, the 25% KOH had lower capacity loss compared to 35%, with both showing better results than the 15% KOH, and supporting electrolyte was beneficial to battery capacity. The addition of a graphite layer, however, was not beneficial to the battery, showing similar or worse capacity degradation. Finally, we used a rotating ring-disk electrode (RRDE) to analyze when the dissolved species appears, and learned that it appears during the charge process. |
| Faculty: Hajal, Cynthia Mentors: Guadalupe Garcia, Yeqing Ni | Clayton Pierce, Xyden Procaccianti | Modeling Drug Delivery to Gliomas Final Presentation Final Poster | Gliomas are a type of brain tumor; a subset of gliomas called Glioblastoma (GBM) is a more severe type. The Blood Brain Barrier (BBB) restricts access of foreign species entering the human brain. This poses a challenge when trying to treat gliomas, as the drugs used to treat the tumors have trouble permeating through the BBB. This research aims to understand how that stiffness affects tumor growth and drug response. There are two components to the research, one which involves the physical experimentation, while the latter component consists of simulating the expected results of the experiment via COMSOL. The physical experiments consist of making microfluidic devices and seeding them with tumor cells in hydrogels of varying stiffness. While differences were not statistically significant, we did observe a trend suggesting that cells in softer hydrogels exhibited greater growth as a responsive to chemotherapy compared to those in stiffer environments. Prior to this physical experiment, the microvascular network was simulated in COMSOL using different physics systems incorporated in the software to account for permeability of cell tissue and vascular fluid velocity. In the simulation the velocity of the blood was in the physiological range, that result being 400-1000 μm/s. The concentration of the medicine in the tumor mass was higher approaching the edge of the mass and was lower when approaching the center of the mass. |
| Faculty: Kaeli, David Mentors: Zlatan Feric, Mouad Tiahi | Aodhfionn Downs, Wenxi Wang | Linking PROTECT Research Papers to Data Dictionary Variables Using Natural Language Processing Final Presentation Final Poster | Large Language Models (LLMs) are powerful tools for processing and extracting information. The PROTECT Center has been studying preterm birth rates in Puerto Rico for over a decade, publishing more than 300 research papers and collecting raw data on over 2000 women. Papers written by PROTECT researchers utilize information from the dataset, but does not cite which variables they used in the data dictionary. Our project employs LLMs to link the papers and the data dictionary variables, thereby enhancing reproducible research. Our baseline approach involved giving each paper to GPT-4 with the data dictionary and asking for the variables, but this led to context window overflow, hallucinations, and irrelevant results. To address the issues, we developed a customized Retrieval Augmented Generation (RAG) pipeline that uses an LLM to identify variables in the papers without the data dictionary and matches the LLM’s output to the dictionary variables using cosine similarity on the embeddings. We used the LLM-as-a-judge method to evaluate the second step of our preprocessing. |
| Faculty: Kim, Hyeon Yu | Bilal Shaikh, Alla Othman | The Development of Heart Disease Models for Better Patient Understanding Final Presentation Final Poster | This research topic allows for deeper understanding in the topic of cardiovascular health. The specific disease researched in this lab is Hypertrophic Cardiomyopathy. Knowledge on this disease allows for early detection, better treatment, improved outcomes, and a deeper understanding in cardiovascular pathology. |
| Faculty: Koppes, Abigail Mentors: Katherine Nilov, Eric Johnson, Ricardo Fernandez | Kailyn Love, Justin Zou | Analyzing Bubble Traps in Organ-on-a-Chip Systems to Improve Efficiency and Reproducibility for Drug Screening Final Presentation Final Poster | Organs-on-Chips (OoCs), miniature organ systems modeled on microfluidic devices, are used to emulate real tissue and cell environments, allowing for the study of human physiology without the use of live animals or human subjects (1). Currently, a significant concern in OoC systems is the presence of inevitable bubble formation that obstructs flow and increases variability in data collection. To reduce variability, bubble traps can be implemented into the OoCs to capture or divert large bubbles that affect the flow of the media, and therefore, cell interactions (2). It was hypothesized that the use of lower-pressure channels and a double row of pillars can significantly reduce bubbles throughout the cell flow channels of OoC systems, offering valuable insights into alternative strategies for bubble mitigation. Using fluid mechanics principles and bubble-trapping architecture, engineered bubble traps diffuse smaller bubbles and block larger bubbles from disrupting the cell’s environment. These designs could improve reproducibility in high-throughput testing, allowing OoC systems to have a quicker turnaround in the clinical phase, accelerating drug research and the study of human systems. |
| Faculty: McCleary, Jacqueline Mentors: Sayan Saha | Marcus Michaud, Leila Ohashi | Systematics Mitigation in SuperBIT Weak Lensing Data Final Presentation Final Poster | About 95% of the universe is composed of invisible components, roughly 69% dark energy and 26% dark matter, while only ~5% is ordinary (baryonic) matter, which we can observe through electromagnetic radiation. Despite being a small fraction, this visible matter can help us infer properties of the dark components. The Super Pressure Balloon-Borne Imaging Telescope (SuperBIT) targeted ~35 merging galaxy clusters, ideal sites to study the dark matter distribution in a non-equilibrium environment. Our goal was to improve the accuracy of results extracted from SuperBIT images through careful handling of systematics. In this project, we cleaned up noise and diffraction patterns that caused spurious detections near bright stars and image edges, and differentiated between background and cluster member galaxies. This all works towards the greater goal of more completely understanding the interplay between baryonic and dark matter. |
| Faculty: Onabajo, Marvin Mentors: Thomas Gourousis, Minghan Liu, Junyi Yang | Aryaa Mutha, Eric Nie | Filtering Temperature Sensor Outputs for Power Dissipation Monitoring Final Presentation Final Poster | With the development of the Internet of Things (IoT), Hardware Trojans (HT) pose threats to modern integrated circuits and electronic systems, compromising performance, causing system failure, or leaking sensitive information. Recently, a non-invasive anomaly identification approach and machine learning algorithm for HT detection using analog on-chip differential temperature sensors has been proposed. However, electronic noise can obscure the output signals from the temperature sensing front end, thus reducing the detection accuracy. In this project, we designed and experimentally validated an analog system that behaves as a low-pass filter to improve signal clarity. This system consists of an operational amplifier buffer, continuous time integrator with reset and inverting amplifier. |
| Faculty: Ramezani, Alireza Mentors: Kaushik Venkatesh | Trinh Huynh, Ethan Men | Design and manufacturing solutions for robustness of Crater Observing Bio-inspired Rolling Articulator (COBRA) Final Presentation Final Poster | COBRA is designed to traverse craters on the moon, through multiple mediums such as tumbling in a rolling configuration and slithering. Traveling through rough terrain in space, dust is very prone to interfere with electronics and cause short circuiting and overheating. Especially in the head module, which connects to the tail to form the rolling configuration. In our project we ensure that COBRA is structurally robust through design, and that wire management and access to electronics are protected and convenient. All to ensure that the head module of COBRA is sealed, protected, and easy to assemble. These components work together to ensure that the crucial electronics and components are protected, discrete, and accessible; for safe and reliable function. |
| Faculty: Zhang, Xufeng Mentors: Yu Jiang | Esther Ji, Aryan Sharma | Integrated Magnetic Transducer for Advanced Magnetic Sensors Final Presentation Final Poster | Magnons are quasiparticles that represent collective spin excitations in magnetic materials. Cavity magnonics, in particular, is an important hybrid magnonic platform for coherent interactions between magnons, photons, and phonons within resonant structures, which enables fast information exchange across physical systems. These interactions have the potential to allow for energy-efficient signal processing, quantum transduction, and hybrid computing technologies. Our research focuses on developing hybrid magnonic devices that leverage light-matter interactions in the slow-wave regime, where electromagnetic (EM) waves are significantly slowed down compared to conventional EM waves to enhance interaction strength. In this work, we introduce a novel waveguide structure consisting of an array of coupled rectangular split-ring resonators (RSRR), allowing for compact designs and improved field confinement while maintaining slow-wave capabilities. By leveraging the unique properties of slow-wave hybrid systems, our new structure promises great potential for both fundamental research and practical applications. Additionally, the concept of this device can be extended to other systems, such as optomagnonics (involving optical photons rather than microwave photons) and magnomechanics (involving mechanical vibrations), paving the way for coherent information science. |
| Faculty: Zhao, Qing Mentors: Colin Gallagher, Connor Fawcett | Rohan Harrison, Lydia Van Voorhis | Stability Assessment of Graphene-Based Single-Atom Electrocatalysts for CO2 Reduction Final Presentation Final Poster | Unsustainable CO2 emissions pose several ecological and humanitarian challenges. Electrochemical CO2 reduction (CO2R) has been proposed as a solution. Traditional CO2R catalysts suffer from several practical limitations. Graphene-based catalysts featuring isolated metal centers, namely single-atom catalysts, provide a platform for new CO2R catalysts that can outperform traditional cathode materials. |