Exploring Dark Matter Interactions: Extending the LUX-ZEPLIN Search to Non-Standard WIMP-Nucleon Interactions, Anomaly Finding Techniques, and Liquid Xenon Detector Characterization

Abstract

Dark matter is a prime mystery in modern physics. Independent measurements point to a quarter of the universe's mass-energy (and 85% by mass alone) being made up of a cold, collisionless species of matter. Its ubiquity in the universe, confirmed through gravitational phenomena and cosmological evidence, belies the fact that dark matter particles have not been observed to interact directly with regular atoms. A leading candidate for dark matter is the Weakly Interacting Massive Particle (WIMP), which has a mass between a proton and a complex protein, and interacts with regular matter with a strength similar to the electroweak force (mediated by the W and Z particles). A series of experiments have unsuccessfully searched for evidence of galactic WIMPs passing through Earth, culminating in the LUX-ZEPLIN (LZ) experiment's result in July 2022 that placed the most stringent upper bounds on the WIMP-nucleon interaction strength.

This thesis extends LZ's first result by expanding the energy window of the search (by a factor of 7.5) to be sensitive to non-standard WIMP-nucleon interactions. No evidence for WIMPs is found in this new analysis region, but the strengths of 54 distinct WIMP-nucleon interactions, generated by an Effective Field Theory (EFT), are constrained by the analysis. The EFT is a theoretical tool that enumerates all of the possible interactions distinguishable at the energy scale of the WIMP-nucleon scatter; this LZ EFT result provides the tightest constraints for nearly all of these interactions.

Three other topics are presented in addition to the LZ EFT analysis. First, a study to characterize and remove data that mimic non-standard WIMPs is described, enabling the extension of the energy window to where such signals are expected. Second, the question of anomalous experimental data caused by noise, detector and signal processing effects is addressed, leading to the development of an interpretable general-purpose anomaly finding tool using techniques in machine learning. Applications of the anomaly finder on simulated and real LZ data are also shown. Third, this thesis presents the conceptualization and partial results of an experimental study with the Michigan Xenon (MiX) detector to precisely measure the electron and photon production yields of liquid xenon, the detector medium employed in LZ and other WIMP detectors.