Session D18: Undergrad Research IV

Nuclear Binding Energy Calculation: An Interactive Web Platform Integrating Mass Defect and Liquid Drop Models

We have developed an interactive website designed to calculate nuclear binding energy employing two distinct methods: the mass defect approach and the liquid drop model. Accessible at http://Binding-Energy.github.io, the site features three main pages. The first is a search interface for locating specific isotopes. The second, a calculation page, computes and displays the binding energy of the chosen isotope. Lastly, a unique page displays when users input properties of a non-existent isotope. The process initiates when users input proton and neutron numbers (Z and N). The site executes a GET request to our custom API, incorporating these Z and N values. Our API sifts through International Atomic Energy Agency (IAEA) data to match these values, returning the isotope's JSON object if found. This data enables the website to perform and exhibit calculations. Conversely, if no matching isotope exists, the API signals this, and the site informs the user accordingly. Development of this website utilized two open-source JavaScript frameworks: React JS for the front end, integrating JavaScript, HTML, and CSS for a more streamlined web design, and Express JS for the back end, crafting RESTful APIs. Our Express JS implementation connects the front end to the IAEA's nuclear data, allowing us to preprocess the data, converting from CSV (comma-separated values) to JSON (JavaScript object notation) for better manageability with our extensive dataset (over 192,833 lines of data). This preprocessing also includes data filtering based on website requests.   

Monte Carlo Modeling of sub-keV Backgrounds in Superconducting Tunnel Junctions from Gamma-Ray Interactions for SALER@FRIB

The new Superconducting Array for Low-Energy Radiation (SALER) experiment at FRIB aims to directly embed short-lived isotopes in superconducting tunnel junction (STJ) sensors to measure nuclear recoil energies from weak decay as a search for BSM physics. At these eV-scale energies, small energy depositions from simultaneous higher-energy events (since as gamma rays) generate backgrounds which are important to understand. In this talk, I will present the first Monte Carlo modeling using the test case of 137Cs decays where a 662 keV gamma-ray is emitted and deposits small energies in the silicon substrate of the STJ as a critical first step towards background characterization for SALER@FRIB.   

Improving CUORE Energy Reconstruction Using Machine Learning

CUORE (Cryogenic Underground Observatory for Rare Events), located in the Laboratori Nazionali del Gran Sasso in Italy, is an experiment designed to search for neutrinoless double beta decay (0νββ) in ¹³⁰Te. Neutrinoless double beta decay is a theoretically predicted radioactive decay process that, if observed, would determine the Majorana nature of the neutrino. This would indicate that lepton number is not conserved and could help explain the matter antimatter asymmetry in the known universe. I present an analysis on the applications of different machine learning algorithms, including PCA (Principal Component Analysis) and linear regression, on improving energy reconstruction for CUORE. Energy resolution is of great importance for CUORE as it would help better distinguish 0νββ signals from background events. Currently, CUORE uses a software trigger algorithm called an Optimum Filter (OF) trigger for energy reconstruction. However, OF assumes that signal shape is independent of energy which isn’t true in CUORE.  A machine learning aided reconstruction could mitigate the energy dependence of pulse shapes and improve the overall energy resolution of the CUORE detector. I will present the current results of this analysis.    

Understanding Energy Reconstruction in CUORE using Topological Analysis

CUORE (Cryogenic Underground Observatory for Rare Events), located in the Laboratori Nazionali del Gran Sasso in Italy, is an experiment designed to search for neutrinoless double beta decay (0νββ) in 130Te. Neutrinoless double beta decay is a theoretically predicted radioactive decay process that, if observed, would determine the Majorana nature of the neutrino. Experimentally confirming 0νββ would signify that neutrinos are their own antiparticle, allowing for a better understanding of the matter antimatter asymmetry present in the universe. Accurate reconstruction of 0νββ event energy is key to isolating 0νββ from background events. I present an analysis of the energy escape events occurring near (~1 cm) the surface of the TeO2 crystals, operated by CUORE. My work aims to understand and correct biases in the bolometric energy reconstruction and improve the systematic uncertainty on the CUORE energy resolution. Currently, there is little microphysical understanding of the energy biases present within the detectors. The analysis aims to classify energy reconstruction errors in terms of event topology, location, and possible variation of the pulse shape. I will present the current results of this endeavor.   

Using Convolutional Neural Networks for Event Pileup Discrimination in CUPID

Neutrinoless double beta decay (0νββ) is a proposed radioactive decay process that could prove the neutrino’s Majorana nature and demonstrate the violation of lepton number conservation. The observation of 0vββ would provide a powerful hint to the origin of the matter-antimatter asymmetry in the Universe. Amongst the next generation of 0vββ experiments, CUPID (CUORE Upgrade with Particle IDentification) will search for this process in the isotope ¹⁰⁰Mo. Since cryogenic calorimeters behave as signal integrators, two events that occur close enough in time result in a total signal that is close to the sum of the two single events. This family of events, called “pileups”, constitutes one of the major sources of background in the region of interest for the CUPID experiment. The pileup of two neutrino double beta decays (2vββ) is expected to be the dominant contribution. Deep learning algorithms trained to identify these anomalies in time series can provide a major improvement to the standard analysis methods. I present an analysis of the efficacy of convolutional neural networks (CNN), a deep-learning architecture typically used for image recognition and processing, as a method of identifying and discriminating pileup events in time series data.   

Measurements of the 134Xe Neutron Capture Cross-section Between 0.43 and 5.5 MeV

In the study of rare event physics, such as neutrinoless double beta decay, it is important to understand potential background events. Events such as cosmic rays and stray alpha particles can lead to neutron-induced events even in experiments deep underground. Experiments that study neutrinoless double beta decay of ¹³⁶Xe use material enriched in ¹³⁶Xe. However, a substantial fraction of the enriched material is ¹³⁴Xe. One neutron-induced event of interest is neutron capture on ¹³⁴Xe, which can emit gamma rays that have the potential to Compton scatter into the Q-value region of interest for double-beta decay of ¹³⁶Xe. In this study, we investigate neutron capture on ¹³⁴Xe by looking for gamma rays emitted from both the de-excitation of resultant metastable excited states of ¹³⁵Xe, and the subsequent decay to ¹³⁵Cs. Highly enriched xenon gas was irradiated in the neutron beam at Triangle Universities Nuclear Laboratory, and the resultant decays were counted in the low-background counting facility located in the Duke Physics Building. We will report our results for the neutron capture cross-section for incident neutron energies of 0.43, 0.8, 1.5, 2.0, 4.2, and 5.5 MeV.   

Measuring Intrinsic Correlated Avalanches: SiPM Testing in Liquid Xenon

The nEXO experiment aims to search for neutrinoless double beta decay of Xe-136 with a half-life sensitivity >10^28 years. The nEXO detector is a 5-tonne, single-phase liquid xenon time projection chamber (LXe TPC) enriched to 90% in the mass 136 isotope. nEXO collects both ionization charge and scintillation light and uses ~ 4.6 m^2 of vacuum ultraviolet (VUV) silicon photomultipliers (SiPMs) to detect 178 nm scintillation photons. The intrinsic properties of the SiPMs, such as charge gain, breakdown voltage, correlated avalanche (CA) probability, and photon detection efficiency (PDE) play an important role in achieving the best possible energy resolution and need to be measured in the 167K LXe environment. This talk will report the measurements of the gain, breakdown voltage, and CA probability in a kg-scale LXe setup at UMass Amherst. A method using a pulsed, UV LED in low illumination mode to extract the single photoelectron (SPE) signal beneath the otherwise overwhelming scintillation glow in the cell due to radioactivity and cosmic rays is presented. The results are compared with measurements of CA probability taken in vacuum and gaseous nitrogen, in which scintillation is absent and systematic effects are quantified. A comparison across various cryogenic temperatures and media will also be presented.