The Circumgalactic Medium (CGM) is an extended structure surrounding galaxies, populated with hot diffuse gas and cold dense clouds of gas. The CGM acts as an intermediary between gas within galaxies (Interstellar Medium, or ISM) and gas between galaxies (Intergalactic Medium, or IGM). The mass of a given galaxy's ISM at the time of its formation is much too small to fuel star formation over galactic timescales, so gas must be accreted from the CGM to continue fueling star formation, and insights into the properties of this gas could improve our understanding of galaxy evolution.  During galaxy formation, gas heats up as it falls into dark matter halos from the cosmic web, forming the hot CGM. This gas then cools radiatively (through metal lines and Bremsstrahlung emission in the UV and X-Ray). The time required for this gas to cool and collapse depends on the properties observed from emission—hotter gas has more energy to radiate, high density or metal-rich gas cools more efficiently. If the cooling time is short, heating is required to prevent the gas from collapsing onto the galaxy; one such mechanism is the galactic winds driven by supermassive black holes and supernovae. The details of these heating processes are still uncertain and are a topic of active research in the field. For example, recent observations from the X-Ray telescope eROSITA have been shown to conflict with numerical hydrodynamic simulations of galaxy formation.  In our work we construct analytical models of the CGM with parameters such as density and temperature distribution and spatial extent. We calculate the gas mass and luminosity of the CGM to constrain the reasonable values of the model parameters. We will then use detailed spectral models of gas emission to compare our calculated luminosities to those observed and relate them to heating mechanisms.

Massive stars place powerful constraints on stellar evolution and exhibit a wide range of exotic evolutionary phases. They play a crucial role in regulating their environments, driving the chemical evolution of host galaxies, and establishing energy equilibrium through feedback processes. Stellar variability, notably, acts as a profound probe into the poorly-constrained physics of massive star evolution, illuminating intrinsic properties such as surface gravity. Drawing upon the collective insights from past literature on the dynamics of stellar rotation and surface gravity, this project delves into the correlation between variability metrics from the Zwicky Transient Facility (ZTF) and surface gravity measurements from the Apache Point Observatory Galactic Evolution Experiment (APOGEE) data. By merging these datasets, our aim is to use modern statistical methods to decode the relationship between observed variability and surface gravity in massive stars, shedding light on their rotational behaviors and structural changes over time. This endeavor not only seeks to deepen our understanding of stellar mechanisms but also to improve the precision in classifying stellar masses by utilizing variability as a key diagnostic tool. We endeavor to connect empirical observations with theoretical frameworks, paving the way for future advancements in our comprehension of stellar dynamics and their cosmic significance. Our results will place key constraints on the study of massive stars.

Modern astronomical survey telescopes, like the Vera C. Rubin Observatory and Extremely Large Telescope, are projected to produce terabytes of data each observing night, raising the need for efficient machine learning algorithms to flag astronomical data for further study. One such algorithm is the Random Forest (RF) algorithm, which has previously been demonstrated to process some 2.36 million different galaxy spectra data (in 12 hours over 128 CPUs) for any potentially unique or even undiscovered phenomena. However, this previous demonstration used spectra taken from the combined light of a given galaxy. RF algorithms are ensembles of structures called “decision trees,” which categorize data points with value-comparison questions. This method can be extended to create a metric to calculate how unusual a data point is compared to other points in the dataset. Our project will extend this RF algorithm for the Mapping of Nearby GAlaxies survey (MaNGA), which accounts for the fact that spectra vary with regions of a galaxy. We explore the behavior of the RF algorithms when accounting for these spatially-varying features. Our methods include the generation of synthetic data to train random forest algorithms, RF model hyperparameter searches, and comparison of models. Furthermore, our project compares the similarity of our results to the results from Baron and Poznanski (2016), which previously applied the RF algorithms to the 2.36 million spectra data. We present the conclusions from the RF algorithms and whether prevalent emission lines are flagged by the algorithms, such as hydrogen-alpha and O-III lines. Our project also discusses the features characteristic of outlier galactic region spectra data. Successful implementation of the RF algorithms to process data pipelines from upcoming large surveys has the potential to accelerate the rate of astronomical discoveries to unprecedented levels.

The largest scale structure in the universe creates a cosmic web. Nodes of the web are connected by mega-parsec scale filaments of warm gas, galaxies, and cold dark matter. Cosmic filaments are typically assumed to form and cool uniformly in many cosmological models. Recent works suggest that the internal structure, particularly with regard to the cooling mechanism, may be more complex. Gas clouds larger than a characteristic length appear to shatter as they cool below 106 K, fragmenting into smaller cloudlets. Other factors may also contribute to a more turbulent and irregular structure. The behavior of these filaments has implications for many of the issues at the forefront of astronomy. Cold dense clouds could create Lyman limit systems in the early universe, restricting the the distance that ionizing photons can travel. Further, the structure of these filaments may influence the mass of dark matter and constraints on cosmological models. To study these systems, I am simulating the formation of a single cosmic filament using a cosmological hydrodynamics code. Together with my mentor, I derived functions for the initial displacement and velocities of the particles from theory to create the desired collapse of matter into a filament. I then run the simulations with these initial conditions on a supercomputer, enabling simulations with tens of millions of particles. The simulations have mass resolutions that have not been previously achieved, allowing us to better understand the internal behavior. Preliminary findings indicate that there is indeed a complicated structure within the filaments.

Dark matter is a theoretical form of matter that doesn’t interact with light or conventional matter despite its large expected abundance in our universe. One potential candidate for dark matter, predicted in models, is weakly-interacting long-lived particles (LLP). The ForwArd Search ExpeRiment (FASER), located in the Large Hadron Collider (LHC) at CERN, uses detectors to search for LLP produced in proton-proton collisions. Upon hitting the detectors, particles generate electronic hit signals that are used to reconstruct the decay products of LLPs. Analyzing these tracks may offer insights into the properties and characteristics of LLP. The FASER detector is made of four tracking stations: 1, 2, 3, and interface tracker, each comprising of three layers with eight Semiconductor Tracker modules in each layer. Due to the imprecise installation of these tracking stations, misalignment occurs preventing accurate track reconstruction. To address this issue, I aim to execute an iterative local Chi-square alignment test to determine alignment parameters for each station individually and collectively using previously collected FASER data. I hypothesize that modules will have improved residual values and sensitivity after alignment.

The lithium isotopic composition of metasomatized rocks preserves a history of fluid movement that may be related to large subduction zone earthquakes. High pressure and temperature conditions within subduction zones cause the dehydration of hydrous minerals, and the resulting fluid can raise the pore pressure and trigger slip along the plate interface. The role of fluid movement in subduction zone processes can be better understood by constraining the duration of these events. These relatively fast and periodic fluid increases are recorded by chemical diffusion between the fluid and rock. Fluid containing lithium within a subduction zone can drive the diffusion of lithium into the surrounding rock. Lithium has two stable isotopes that diffuse at different rates, ⁶Li diffusing faster than ⁷Li, creating spatial heterogeneity in the isotopic composition of the reacted rock. Metamorphic rocks from the Western Alps have a reacted rind structure where fluid interaction occurred within an extinct subduction zone. The period of this interaction was examined using lithium isotopes. I prepared these samples for isotopic analysis, first weighing rock powders, then digesting them in acid, and finally separating the lithium with cation exchange columns. Lithium isotopes were measured on a multi-collector inductively coupled mass spectrometer. The spatial distribution of lithium isotopes along the profile from the reaction rind to the unreacted core, together with a thermodynamic model of Li diffusion through the rock, constrain the duration of the fluid interaction and provide insight into the role of fluid in catalyzing slip along the plate interface. I expect the duration of this fluid contact to be short, consistent with pulsed fluid movement within subduction zones. Modern subduction zones, including the Cascadia Subduction Zone that underlies Seattle, pose seismic hazards that can be better understood by examining the relationship between fluid movement and slip in extinct subduction zones.