Research in two-dimensional semiconductors is highly promising towards advancements in quantum electronics and has led to many steps towards understanding 2D electron systems and defining the behavior of the quantum phases of solids. Among these materials, Transitional Metal Dichalcogenides (TMDs) are uniquely promising due to their large optical bandgap and host strong electron-electron interactions which make them a good avenue for studying the resulting emergent phases. So far, many interesting electronic states have been observed in various TMD devices; however, the performance of these 2D semiconductor devices and the physics hosted are limited by the quality of contact between the TMD and the external circuitry. Adding a ‘charge transfer’ monolayer of RuCl3 between one TMD, monolayer WSe2, and external contacts has been shown to create a highly transparent contact and allow the measurement of record high hole mobility and observation of Fractional Quantum Hall. This works by utilizing the unique band properties of both materials which allow RuCl3 to donate charges to the TMD. This study seeks to expand these results to study MoS2, a TMD whose valence band has not been accessible through traditional contacts and twisted MoS2 which may host interesting electronic phases since the charge carriers may inhabit different valley states than the monolayer. Learning more about the effectiveness of a charge transfer layer could benefit the development of 2D devices such as field effect transistors and photo detectors as well as

Upcoming astronomical surveys like the Legacy Survey of Space and Time (LSST) will image billions of galaxies to determine properties of dark matter and dark energy, requiring efficient methods to estimate their distances, or redshifts. Traditionally, redshifts are obtained via spectroscopy, which is highly accurate but expensive, and only available for a small fraction of galaxy populations. Photometric redshift methods approximate spectroscopy and are cheaper and available for more representative galaxy populations in the universe, but are less accurate. Machine learning models, the focus of this work, can be trained on galaxy photometry and spectroscopic or photometric ground truth redshift to predict redshift efficiently. To improve model generalizability, we use image data and mix ground truths by either transfer learning, in which a model trained on one dataset is adapted to another, or training on combined datasets. We use two datasets in this work– GalaxiesML, a spectroscopic redshift dataset compiled from HSC PDR2, and TransferZ, a photometric redshift dataset compiled from the COSMOS2020 survey. We train a convolutional neural network model on either TransferZ or GalaxiesML, then refine the model using transfer learning on the other dataset, and compare this to training on a combined dataset. By comparing RMS, bias and outlier metrics, we study how machine learning approaches can enhance generalization of predicting redshifts on a broader population of galaxies, vital for next-generation cosmic surveys.

The study of materials exhibiting quantum behavior on the macroscopic scale, such as superconductivity and charge density waves (CDW), reveals the fundamental physics of how electrons self-organize, how quantum phases interact, and how electronic behaviors emerge from complex many-body interactions. Just as the liquid-to-solid transition of water depends on temperature and pressure, quantum phases are sensitive to temperature, strain, etc. Tuning these parameters reveals correlated phenomena. Scanning tunneling microscopy and spectroscopy (STM/STS) is an atomic-scale imaging technique for probing quantum phase evolution via quantum tunneling between a sharp tip and the sample surface. However, many quantum microscopes operate only at fixed temperatures of 300 K, 77 K, or 4 K, limiting our study of temperature-driven transitions. In this talk, I will present my work designing and fabricating a compact, low-power, cost-efficient sample holder that enables STM measurements on high-temperature superconductors while tuning sample temperature, strain, and current. The heater provides 20 K tunability via resistive Joule heating and a Python-controlled negative-feedback loop. I will also share initial results on using atomic force microscopy in lateral force mode to align superconducting samples prior to STM measurements. By integrating temperature control and precise alignment, we aim to map the superconducting gap and study how strain and in-plane current flow alter the electronic properties of quantum materials.

When generating plasma for scientific studies, such as in the Large Plasma Device (LaPD), a moving probe drive would allow for a probe to move axially through the experimental chamber, allowing for a greater number of data points, as opposed to fixed axial locations that are accessible through ports. This instrumentation raises the challenge of appropriate temperature management. As such, we present an alternative to conventional cooling methods. Considering the conditions and constraints of the LaPD, air cooling is not possible, and similar difficulties exist for water and refrigerant cooling. This is due to the difficulty of installing these in the chamber, particularly because we are attempting to build a moving probe drive on a carriage. This study examines the effectiveness of Peltier coolers with passive radiators in vacuum, such as those present in the LaPD. Different combinations of heaters, coolers, and power were used in a simulation of the instrumentation that will be used in the LaPD. Our study shows that Peltier coolers with passive radiators are effective in vacuum, but that stacking coolers in series make a negligible difference, and that its radiative ability is greatly reduced when the chamber size is reduced.