Physics, 1936-2026

The quantum bootstrap aims to extract spectral data from quantum systems without ever constructing wavefunctions. By imposing consistency conditions on expectation values, including Hilbert-space positivity, the eigenstate equation, and symmetry constraints, the bootstrap asks what spectra and correlators could be consistent with the general rules of quantum mechanics that any physical state must obey. In this thesis, we examine how far this framework can be extended - from simple quantum mechanical systems to fermionic lattice field theories. We first reproduce and extend the anharmonic-oscillator bootstrap of Han, Hartnoll, and Kruthoff, demonstrating that mixed-correlator constraints involving both position and momentum tighten the feasible regions around the ground and first excited states. We then construct a Level 2 bootstrap for the staggered-fermion Thirring model, exploiting charge and momentum symmetries to decompose the moment matrix into ten symmetry blocks. At L = 4, m = g = a = 1, the bootstrap recovers the exact-diagonalization ground state energy to ten decimal places, the sharpest numerical benchmark in the thesis. We then extend the framework to L = 6, 8, tracking exact diagonalization across the gapless regime at m = 0, and compare the finite-size behaviour with expectations of conformal field theory (CFT). We observe that the free theory matches the expected Casimir-plus-winding scaling, while the interacting g = 1 bootstrap data agree closely with exact diagonalization but are not yet at large enough system size for a clean extraction of asymptotic CFT data.

This thesis describes the design, fabrication, and testing of a persistent current switch (PCS) and superconducting joints for the Princeton Axion Search. A novel planar switch geometry is proposed and validated, offering improved thermal contact and reduced inductance without epoxy impregnation. The final switch iteration demonstrated a normal-state resistance of 7.5 Ω, an upper limit on superconducting-state resistance at 8.2 nΩ, and a critical current of 370 A below 5.5 K. The switch was driven normal in under six seconds using only 10% of available heater power, validating its quench protection capabilities. A PID-controlled metal bath was designed and constructed for fabricating superconducting joints. A constructed joint sustained 520 A of current at 3.08 K, and a persistent current utilizing the switch, two joints, and a mini-coil showed no measurable decay across three days, setting an upper limit on the joint resistance of 1.90 × 10−13 Ω.

Scanning probe microscopes are key in 2D materials research. One challenge in using them is that depth sensing cannot be done in a fully decoupled way using the same instrument used for measurement, and improper depth sensing that results in a crash can damage both tip and sample. For scanning SQUID-on-tip microscopes, a method was devised to use the phase of oscillation of quartz tuning forks in contact with the tip to sense variations in tip-sample interaction corresponding with depth. A low-noise amplification system is necessary to read out the small changes in the phase. In this paper, we describe the theory of tuning forks, amplifier noise, the system we use to characterize amplifiers and tuning forks. We conclude with potential upgrades to the system, regarding both lower-noise amplifiers and different geometries of the tip and tuning fork which may be more sensitive to depth.

Single-molecule junctions represent the ultimate limit of electronic miniaturization, where quantum mechanical effects and molecular orbital symmetry directly determine device behavior. This thesis investigates the transport properties of CO–FePhthalocyanine molecular junctions from a theoretical and computational perspective, establishing a quantitative framework for understanding and predicting rectification in transition-metal phthalocyanine systems. Using a Landauer transport model parameterized by density functional theory values and scanning tunneling spectroscopy measurements from the literature, a standard approach in molecular electronics that combines physically grounded parameter estimates with an interpretable trans port framework, we compute the current–voltage characteristics of CO–FePc and bare FePc junctions across a systematic variation of orbital energy, coupling asymmetry, tip–molecule distance, temperature, Fe charge state, and CO orientation. The central finding is that CO enhances the rectification ratio R = |I(−V )|/|I(+V )| by a factor of 3.10 relative to bare FePc, producing R = 9.18 at reference conditions of |V | = 1.0V and T = 50K, through two independent mechanisms. First, the σ-donation interaction shifts the Fedz2 orbital from −0.40eV to −0.15eV relative to the electrode Fermi level, enhancing the low-bias conductance G by a factor of 6.71, compared to an experimentally observed enhancement factor of approximately 3. Second, the permanent CO dipole moment couples to the junction electric field via a linear Stark shift, generating a bias-dependent asymmetry parameter α = αgeom + αdip, where αgeom reflects the asymmetric tip–substrate coupling and αdip reflects the CO dipole Stark shift that drives the transmission resonance preferentially through the Fermi window under negative tip bias. Reversing the CO orientation from carbon end-down to oxygen-end-down reverses the sign of αdip, switching the junction from a robust forward rectifier (R = 9.18) to a weak reverse rectifier (R = 0.945) while leaving G unchanged, a consequence of the orientation-independence of σ-donation. The two orientations are spectroscopically distinguishable by the sign of their differential conductance dI/dV peak voltage: −0.155V for C-down and +0.210V for O-down. The rectification ratio degrades by only 19% from 4K to 300K, reaching R = 7.43 at room temperature, a consequence of the Γ-broadened transport regime in which thermal smearing is a perturbation on the dominant coupling broadening Γ = ΓL+ΓR, where ΓL and ΓR are the tip and substrate coupling strengths, respectively. A Marcus theory model for Fe(II)/Fe(III) charge-state switching predicts bistable hysteretic I–V characteristics with a threshold bias sweep rate of u∗ ≈ 0.08V/s, establishing that the junction functions simultaneously as a molecular diode and a memory element. A rectification phase diagram computed across the full (α, ΓL/ΓR) parameter space demonstrates that the sign of α is the primary determinant of rectification direction, and that CO orientation control moves the operating point across the R = 1 flip boundary without any chemical modification of the molecule or electrode. These results establish CO–FePhthalocyanine as a field-configurable molecular device whose rectification magnitude and direction are continuously tunable through junction geometry, axial ligand chemistry, and atomic-scale CO manipulation.

One of the major challenges of plasma physics is the difficulty of identifying instabilities, which is the first step in learning to control or eliminate them. This thesis applies several machine learning techniques to identify instabilities in labeled datasets of fast camera and interferometer data from the PFRC-2 device (Princeton Field-Reversed-Configuration- second generation). The machine learning techniques used include a convolutional neural network and a triplet network. A combined final classification model outputs correlation values for each measuring instrument to an instability. This thesis provides a proof of concept for similar frameworks for lightweight, fast analysis of data to be developed across experimental physics.

When Myxococcus xanthus colonies are grown on hydrogels which have been compressed, the colony grows anisotropically, elongated perpendicular to the direction of compression. This phenomenon, known as mechanotaxis, was first reported in 1942 and since has been experimentally confirmed at the colony scale, but its underlying biophysical mechanism remains unclear. This thesis seeks to understand the cell--substrate interactions that give rise to mechanotaxis. We develop an improved hydrogel compression device which allows for controlled uniaxial compression while minimizing various confounding factors. Through a series of experiments using this device, we show that colony anisotropy arises from a motility driven single-cell reorientation effect significantly modulated by gel hydration. In sufficiently dry and strongly compressed substrates, we then show that mechanotaxis can occur in cells with either pilus- or gliding-based motility, indicating that the phenomenon is not exclusive to gliding motility as previously reported. Furthermore, we reproduce mechanotaxis on both agarose and PAA hydrogels, implying that the underlying mechanism depends on compression-induced anisotropies common to both substrates. Birefringence and bulk stiffness measurements suggest that anisotropic elasticity is a likely driver of the phenomenon. Finally, we discuss possible mechanisms of cell-substrate interaction that may explain our experimental findings. Together, these results provide a deeper understanding of the biophysical mechanism behind mechanotaxis in surface-motile bacteria on compressed substrates.

DNA-tethered metallic nanoparticles are a promising class of building blocks for nanocrystals. Due to their unique emergent optical and electronic properties these nanostructures have been the focus of much research in recent years. To thoroughly understand the properties of these nanostructures—which are constantly in motion—we must study their dynamics. Observations of their motion, however, require these structures to be confined to two dimensions. At this scale, confinement is very perturbative to these artificial structures. In this thesis, we investigate a simple form of confinement we call a fluid sandwich that is, in principle, less invasive to nanostructure dynamics. We begin by reviewing the existing methods of 2D confinement with an eye towards their potential perturbative effects. We, then, discuss the trials to create this configuration, summarizing the core challenges we encounter. Finally, we synthesize the lessons from our work and suggest a promising path forward to create a fluid sandwich.

This thesis explores new and interesting palladium-based materials using a recently-developed method of chemical reactions between two-dimensional van der Waals materials (termed van der Waals nanoreactors) to search for new palladium-based superconductors. Deposited palladium source and various other two-dimensional materials were stacked and prepared for chemical reactions under boron nitride encapsulation with the intention to find a new crystal of the form La-Pd-O. We find that these van der Waals chemical reactions are successful at growing new crystals that are partially composed of palladium and oxygen, but full elemental analysis and transport tests were inconclusive. This thesis reviews the relevant experimental procedures of fabrication and testing, and discusses the details of resultant crystal growths and measurements.

Neutron-star masses encode the physics of core-collapse supernovae and the sub- sequent binary evolution that may alter a neutron star after birth. In this work, I construct a theoretical neutron-star gravitational birth-mass function using core- collapse supernova simulation results and compare it to the observed neutron-star mass distribution. The model combines a progenitor-to-compactness mapping from two-dimensional FORNAX simulations with a compactness-to-gravitational-birth- mass mapping motivated by late-time three-dimensional FORNAX simulations. This produces a composite relation MZAMS → ξ1.75 → Mg,birth, which is then weighted by a Salpeter initial mass function to estimate the relative contribution of different progenitor intervals to the neutron-star birth distribution. I further examine the effects of removing black-hole-forming candidates and applying a simple period-dependent accretion correction to observed recycled systems. The resulting theoretical birth function is strongly concentrated in the low-mass neutron-star regime, with its largest contributions coming from the lowest-mass core- collapse progenitors, as expected from the steep Salpeter weighting. The model re- produces the qualitative structure of the low-mass observed populations, especially systems expected to remain close to their birth masses, but it underproduces the high-mass tail seen in some X-ray binaries, redbacks, and black widows. This dis- crepancy suggests that the observed high-mass regime cannot be interpreted as a direct, unmodified birth distribution. Instead, it likely reflects a combination of post-birth binary accretion, selection effects, black-hole formation boundaries, and potentially distinct evolutionary channels. These results support a picture in which the neutron-star birth-mass function is intrinsically weighted toward low masses, while the present-day observed distribution is broadened by subsequent binary evo- lution and compact-remnant survival effects.

Harnessing the energy of alpha particles produced in fusion reactions is desirable for building economically feasible fusion power plants. A key consideration for maximizing alpha particle energy extraction and removal from the machine is how the alpha particle distribution is affected by diffusive processes such as wave-induced diffusion and Coulomb collisions. The evolution of the alpha particle distribution function subject to diffusive processes can be modeled using the Fokker-Planck equation. In transforming the Fokker-Planck equation into the equivalent Langevin formulation for use in Monte Carlo simulations, we find a geometric effective drift arising from non-uniformity of the measure. We study the effects of this geometric correction for a collision operator of alpha particles colliding with a deuterium-tritium plasma. For this process, the geometric correction enforces the expected isotropic Maxwellian distribution. In a magnetic mirror fusion device, we demonstrate that neglecting this geometric correction leads to overestimating losses while alpha particles initially slow down and then underestimating the total confinement time. This treatment can be extended to other diffusive processes and particle species for applications such as alpha channeling, direct energy conversion, and impurity transport.

The formation, dynamics, composition, and dissolution of biomolecular condensates are often regulated by post-translational modifications such as phosphorylation. Here, I introduce Mpipi-P, a coarse-grained model for simulating phosphorylated proteins and phosphorylation reactions in biomolecular condensates. Mpipi-P is the first coarse-grained model to enable explicit representation of phosphorylation reactions and to capture the kinetics of these reactions. Mpipi-P is able to correctly predict the sequence-specific effects of phosphorylation on biomolecular condensate formation and can demonstrate that phosphorylation inhibits FUS condensate formation in a dose-dependent manner. Additionally, Mpipi-P is used to show that FUS phosphorylation can contribute to but is unlikely to be the primary driver of DNA damage repair-associated condensate dissolution. These findings demonstrate that Mpipi-P can provide mechanistic insight into how phosphorylation modulates biomolecular condensates. The Mpipi-P reaction framework can be extended to study other phosphorylation-regulated condensates and, more broadly, enzymatically active condensates in general.

Neutrinos are abundant fundamental particles that rarely interact with matter, making them extremely difficult to detect and study. Yet we hope to study them because they can yield insights into the origin of matter and the evolution of the universe. Despite the elusive nature of neutrinos, nuclear reactors, which produce an immense flux of reactor antineutrinos as a byproduct of beta decay, enabled the first detection of neutrinos and have been used to safely monitor nuclear reactor facilities. However, current neutrino detectors lack the capability to measure most neutrinos from nuclear activities, and the multi-ton weight of existing detectors constrains future scalability and portability. In this thesis, we investigate the initial design and simulation of devices for a new approach: the quantum sensing of neutrinos with superfluid helium. Adopting techniques from dark matter experiments, we utilize CaWO$4$ as a target crystal, coated in a thin film of superfluid helium, to study the quantum evaporation of helium atoms induced by quasiparticle excitations. We develop a toy model to approximate the quantum evaporation efficiency, yielding simulated results consistent with previous calculations used in dark matter experiments. Additionally, we prepared and designed hardware and apparatuses for the ongoing construction of the quantum sensor, including a high-current printed circuit board, gold-plating procedures, and thin-film cesium depositors. Our work lays the foundation for realizing the first stage of the quantum sensor, upon which future continuations will build to realize initial experiments.

The formalization of renormalization in quantum field theory was a conceptual milestone that enabled the precise calculation of higher-order perturbative processes. It was later revealed that the underlying combinatorics of this procedure are governed by the algebraic structure of a Hopf algebra. This senior thesis aims to extend this algebraic framework to processes in string theory. Unlike in QFT, in string theory, ultraviolet (UV) divergences arise from geometric degenerations at the boundaries of the worldsheet moduli space, specifically when closed loops shrink to zero length. These short-distance UV divergences can be equivalently interpreted via open/closed string duality as an infrared (IR) phenomenon caused by a closed string propagating over a long distance and going on-shell. Our goal is to not just better understand these divergences, but also to better understand how Hopf algebras are somehow universal in describing renormalization procedures.

Two-dimensional transition metal dichalcogenides (TMDs) exhibit favorably unique optical and electronic properties that make them promising building blocks for engineered heterostructures. Nanoscrolls, formed by rolling monolayer TMD flakes, are a simple and scalable way to create multi-layer structures. However, their formation process is not fully understood, and their resulting properties remain challenging to control. In this work, we develop an experimental framework for producing WS_2 nanoscrolls with improved yield and reproducibility. By systematically comparing flakes that roll with those that do not, we identify key factors, including flake size, shape, and experimental conditions, that govern the rolling process, and propose a practical rubric for selecting favorable flake candidates. We then use Raman and photoluminescence spectroscopy techniques to characterize the resulting structures and verify the presence of interlayer coupling. By comparing nanoscrolls formed under different fabrication conditions, we show how optical and electronic responses vary, and identify trends that link fabrication parameters to material behavior. Together, these results demonstrate that nanoscroll formation can be used not only to produce multi-layer structures, but also to begin controlling their properties. This thesis provides a foundation for using TMD nanoscrolls as a scalable and reliable method for engineering higher-order heterostructures.

The goal of many modern cosmic microwave background (CMB) experiments is to probe the physics of the early universe. In particular, many experiments aim to measure or constrain the tensor-to-scalar ratio, r, whose detection would provide strong evidence for primordial gravitational waves generated during inflation. SPIDER is a balloon-borne CMB experiment designed to search for this signal through observations of large-scale polarization anisotropies. An important stage of the analysis is the use of null tests to assess the self consistency of the data and diagnose potential systematic contamination. In this thesis, a preliminary version of the SPIDER-2 null-test pipeline is implemented and applied to three data splits: a left/right scan split, an inner/outer focal-plane radius split, and a checkerboard split. The X1 150 GHz data pass all null tests. The 90 GHz FPUs, X2 and X6, also pass all null tests, with the exception of the X6 inner/outer focal-plane split, which exhibits a statistically significant failure driven by excess power at low multipoles. Map-level analysis reveals a periodic systematic feature in the X6 data, particularly in the Q polarization maps. While the physical origin of this feature remains unclear, it is localized to a subset of detectors and to specific time intervals, providing a clear direction for future timestream-level investigation. The 280 GHz FPUs (Y3–Y5) pass the left/right null test but fail both the inner/outer and checkerboard splits. These failures are also driven by excess low multipole power in the null spectra. In contrast to the X6 inner/outer radius split failure, no clear corresponding features are identified in map space, suggesting a more diffuse or complex systematic origin. These results demonstrate the importance of null tests as a diagnostic tool in CMB data analysis and provide guidance for a continued investigation of systematic effects in SPIDER-2.

The pedal steel guitar (PSG) is a common instrument in American country music, with an inseparable history and present steeped in country, but recently, it has become increasingly adopted in other genres and spaces outside of the American countryside because of the instrument's unique color and capabilities for expression. The aim of this thesis is to analyze the nature of the pedal steel sound and the methods for re-synthesizing it to improve a pedal steel physical model, specifically the one which runs on the embedded hardware of the Electrosteel instrument, to allow for real-time performance. The creation of the Electrosteel by Jeff Snyder, Davis Polito and Matthew Wang, was to allow PSG players to interface and perform with a synthesizer through the means of the pedal steel's distinguishable expression and timbre. In this way, players can not only transfer their techniques to new sonic timbres and textures, but using a realistic physical model, also extend the sound of the PSG to create near-unreproducible sounds otherwise on physical pedal steels.

Through the methods provided in this thesis, an investigation was taken into how the physical model for string propagation works in this low-computation environment to explain what parts from electric guitar modeling is applicable to the physical model of the PSG and what parts of this model affect tone and timbre. This thesis also clarifies some of what can be quantified with the sound of the pedal steel, and how one can improve the physical model further, as well as what improvements were made through this work, all in an effort to enable future live-session usage and replacement for the PSG, particularly in use cases where the portability of the Electrosteel matters (i.e touring).

Through this work, three different methods were used to model the decay of sounds, some of which were taken from measurements of two types of PSG with different characteristics, and other models were investigated based on electric guitar models as well. These models were iterated on and simulate the effects of an electromagnetic pickup based both on the physical PSG's electrical circuits, and from recordings of the PSGs.

Within the Princeton Field-Reversed Configuration (PFRC-2) plasma machine, an odd-parity, rotating magnetic field (RMF) near the ion cyclotron frequency is used to accelerate and heat the plasma to high temperatures. However, due to the non-uniform nature of the PFRC’s internal magnetic field, a sharp cyclotron resonance cannot exist. Recent experiments suggest that a linearly polarized (LP) magnetic field, decomposable into the sum of two counter-rotating fields, heats the plasma to higher energies. The aim of this thesis is to analyze the X-ray emissions of experimental data from an LP field used in the PFRC-2. A derivation of the Hamiltonian for RMF and LP will be used to predict how a plasma particle will be transported. Then, by using the \textit{RMF}, a simulation of a single-particle trajectory simulations for both LP and RMF fields will be produced. The paths of the tracked particle and features like the total energy of the particle over time will be compared. This will then be compared with the Hamiltonian derivation to see if it is properly described, and then an analysis of the practicality of LP in fusion reactors will be conducted.

Vocal registers are important for vocalists across different styles, ranging from contemporary to musical theater to classical singing. Although some researchers believe that clear register distinctions may not exist, it is quite common for singers to learn techniques that help achieve certain sounds when singing in these registers. Thus, we analyze four vocal registers—head/falsetto, mix, mix belt, and chest—using Fourier analysis of audio recordings of trained low-voice and high-voice singers. Low-voices are told to sing in falsetto, while high-voices are told to sing in head voice. Both sing a series of six notes, across vowels “Ah”, “Eh”, and “Ooh”. Through different methods of graphing, results indicated that vocal registers—especially chest and head/falsetto—can be distinguished, most likely due to differences in subglottal pressure. Additionally, graphing amplitudes of harmonics for higher notes above the passaggio, or the break from chest to head/falsetto, revealed a difference between mix and mix belt. We ultimately conclude that different registers within the modal register are distinct, which may be applicable to future works such as creating software that assist singers in improving their vocal technique.