×
Department Library
Malissa Alley (Senior Thesis, April 2021, Advisor: Scott Sommerfeldt )

Abstract

A mylar tent was placed over a sound source to assist in sound power measurements made with a scanning laser vibrometer. Mylar was used because of its use in measuring small fan noise in ISO standard 10302 and it was believed to be acoustically transparent. However, there were differences between sound power measurements with and without the mylar tent, meaning the mylar was likely interfering with the measurements. As a result, effects within the tent are more complicated than simply transmitting sound through the mylar at normal incidence; sound waves will be incident on the mylar tent’s sides at many random angles of incidence. To better understand how acoustically transparent mylar is, we developed a function for the power transmission coefficient, TΠ, based on angle of incidence and frequency. This function was numerically integrated to get the power transmission coefficient at random incidence. Experiments using a mylar duct verified these theoretical results and data was analyzed as an insertion loss problem. At higher frequencies the sound power transmission coefficient significantly drops off, meaning the mylar isn’t as acoustically transparent as thought. We have shown mylar causes the discrepancies in the mylar tent sound power measurements.

Nicholas Atkinson (Senior Thesis, April 2021, Advisor: Justin Peatross )

Abstract

Fundamental, second, and third harmonic nonlinear Thomson scattering emitted by electrons from a laser focus with $> 10^{18} \frac{W}{cm^2}$ has been measured at BYU. The strength of emission patterns indicated the intensity of the laser used to drive the electrons. In principle, protons should behave identically to electrons if they are driven with a greater laser intensity, in proportion to the square of the larger mass, namely $1833^2$ times higher intensity. We explore through simulations whether protons ionized from hydrogen can be used to benchmark laser intensities in excess of $10^{24} \frac{W}{cm^2}$, which may be achieved in the next decade. A prepulse can be added to a laser to remove electrons from the focal region so that they do not contribute to the signal from the protons.

Lucas Barnes (Senior Thesis, April 2021, Advisor: Brian Anderson )

Abstract

Time reversal (TR) is a method of focusing wave energy at a point in space. The optimization of a TR demonstration is described that is designed to knock over only one selected LEGO minifigure among other minifigures by focusing vibrations within an aluminum plate at the target minifigure. The goal is to achieve high repeatability of the demonstration along with reducing the cost of the demonstration. By comparing the motion of the minifigure and the plate directly beneath its feet, it is determined that a major factor inhibiting the repeatability of the demonstration is that smaller vibrations leading up to the focal event cause the minifigure to bounce repeatedly, losing contact with the plate and ending up in the air during the main focal event intended to knock over the minifigure. The effects of amplitude, frequency, TR technique, plate thickness, and vibration sensor type are explored to determine their impact on the repeatability of the demonstration. A description is given of the implementation of the demonstration for a museum exhibit. This demonstration illustrates the power of TR focusing and the principles learned by optimizing this demonstration can be applied to other real-world applications.

Katherine Burke (Senior Thesis, April 2021, Advisor: )

Abstract

Molten salts are used in nuclear reactors as a substitute for water in the cooling process. Using salt increases the safety of a reactor due to simplifying the overall process for energy production while adding ways to safely turn the system off without having to rely on additional equipment. Due to using molten salts inside of a nuclear reactor, learning about the densities of different salts is useful because I can use other salts to find ways to be more efficient with fuel usage inside reactors. I tested the density of Fluoride-Lithium-Beryllium (FLiBe) salt using the Archimedean Dual-Sinker method and compared the data to Dr. Stanley Candor’s work to measure the accuracy of this method. I was able to get data that was within 5\% of his work which proves that this method can work to measure densities accurately. This result allows me to improve my method and use the Archimedean Dual-Sinker method to test out salt densities of lesser known salts in future work.

RJ Cass (Senior Thesis, April 2021, Advisor: Richard Vanfleet )

Abstract

Carbon Nanotube Templated Microfabrication (CNT-M) processes use nanotube forest growth from a 2D pattern to form 3D structures. The resulting structure is then infiltrated with a second material such as carbon or tungsten to form the final device. Materials properties of these structures with different infiltration materials and varying degrees of infiltration are of interest. We have used force-displacement data (in fixed-free and 3-point bending configurations) to determine the Young’s modulus of tungsten infiltrated CNT structures (W-CNT’s). Via this method we found the Young’s modulus of (W-CNT’s) to be 8.7 ± 5.2 GPa. However, in the case of the tungsten infiltration processes and typical test beams (~250 µm in width), the infiltration was not sufficiently uniform for high confidence results. Smaller beams (< 50 µm width) are difficult to handle using the conventional 3-point bending processes. We report resonance frequency testing, using a Laser Doppler Vibrometer (LDV), of thin CNT-M cantilevers to find the Young’s modulus of these CNT structures when infiltrated with tungsten. We found the Young’s modulus of W-CNT’s developed using our lab’s process to be 8.9 ± 4.1 GPa.

Jared Davidson (Senior Thesis, April 2021, Advisor: Eric Hintz )

Abstract

The δ Scuti variable star GW Draconis was found to be variable by the Hipparcos satellite in 1997. Since that time of discovery, it has only been examined a few times. Its high declination and short period make it a valuable target for variable star research, allowing for observations over a larger portion of the year that can capture multiple periods. We took observations from 2000 to 2020 and reprocessed this data to corroborate findings from previous papers. From 37 nights of photometric data, we determined GW Dra to have a period of 0.126186 days. Others have reported periods of 0.2545 and 0.1262 d. Frequency analysis yielded several unique frequencies and some subharmonics, but no harmonics of orders higher than the fundamental. Spectroscopic analysis shows GW Dra is an A8 or A9 star and does not indicate the presence of a stellar companion. From our data it is clear that GW Dra is multiperiodic.

Ethan Edwards (Senior Thesis, April 2021, Advisor: Steve Turley )

Abstract

When the wavelength of light is comparable to the length scale of a surface’s features, physical and geometrical optics approximations of reflectance fail. Since even the smoothest surfaces have nanometer-scale defects, finding direct methods for evaluating optical performance is critical in the extreme ultraviolet. One direct method for calculating reflectance uses the electric field integral equation (EFIE). In this project I use the EFIE to study how defects affect the far-field reflection of monochromatic plane-wave light from a one-dimensional blazed grating. Three defects were studied: uncorrelated and correlated ruling errors, and surface roughness. I examined effects on the resolving power (R) and efficiency (E) of a first-order diffraction peak. Considering 0.2 wavelengths of RMS error in each case, I found that roughness decreased R by 0.4% and E by 94.5%, uncorrelated ruling errors decreased R by 0.0% and E by 1.3%, and correlated ruling errors decreased R by 4.4% and E by 32.0%.

Ethan Fletcher (Senior Thesis, April 2021, Advisor: Benjamin Frandsen )

Abstract

V2O3 is an important material for the study of Mott insulators and other strongly correlated electron systems. Despite decades of research, a complete understanding of the metal-insulator transition in V2O3 has not been conclusively established. Here, we present comprehensive atomic and magnetic pair distribution function (PDF) analyses of V2O3 using both x-ray and neutron total scattering measurements, shedding new light on the mechanism of the transition from the point of view of short-range structural and magnetic correlations on both sides of the transition. We observe an abrupt structural transition with no hint of short-range monoclinic distortions above the transition temperature. This lack of structural fluctuations above the transition contrasts with the known presence of magnetic fluctuations in the high-temperature state, suggesting that the lattice degree of freedom plays a secondary role behind the spin degree of freedom in the transition mechanism.

Gabriel Fronk (Senior Thesis, April 2021, Advisor: Traci Neilsen )

Abstract

Humans can tell direction of sound sources by comparing the pressure signal received at each ear. Similarly, in an underwater environment two hydrophones can be used to determine directionality of sound. Acoustic vector intensity, the metric used to determine directionality and sound level of sources, varies depending on assumptions made about the free-field propagation of sound and reflections present in the environment. An environment in which sound propagates freely without reflections is said to be anechoic. To ensure correct interpretation of intensity measurements made in our water tank, we first characterized the reflections in the tank. With this knowledge, we have made the first acoustic intensity estimates using this two-microphone approach, or pressure gradient method, to determine directionality and sound intensity in our lab’s water tank. These results will provide a good foundation for future intensity measurements done in our lab.

Kenan Fronk (Capstone, April 2021, Advisor: David Allred )

Abstract

NASA is preparing to send new telescopes into space with the capacity to see into the far ultraviolet (UV) spectrum. Many materials lose their reflectance the farther into the UV that they go, but Al is a prime candidate because of its good reflectance in the far ultraviolet. As such, mirrors with Al bases and protective layers are being researched as candidates for the thin film mirrors needed on future telescopes. Without a protective layer, Al oxidizes and loses much of its far-ultraviolet reflectance. Prior to sending telescopes into space, many components are placed in storage for extended periods of time. While in storage, thin films may degrade depending on the temperature or humidity of the environment. We stored multiple Al coated with 30 nm AlF3 bilayer mirrors in a 327 K oven in dry air (276 K dew point) to simulate a hot storage room. The Al layers for all samples are 20 nm. Using spectroscopic ellipsometry over the 190 to 1700 nm range to periodically measure samples, we found that there was no significant change in the ~30 nm AlF3 capping layer over a period of 2500 hours.

Kristen Funk (Senior Thesis, April 2021, Advisor: Michael Ware )

Abstract

Yang-Mills theory is a non-abelian gauge theory based on a special unitary group that seeks to describe the behavior of various elementary particles. The behavior of gluons in the absence of quarks is described when the underlying Lie group is the special unitary group of degree 3 (SU(3)). Additionally, SU(3) Yang-Mills theory is foundational to quantum chromodynamics (QCD). Experiment and computer simulation suggest that a mass-gap should exist in the solution of quantum Yang-Mills equations, but this property is not understood from an analytical perspective. We present a simple method for quantizing SU(3) Yang-Mills theory based on methods used in the quantization of the electromagnetic field. However, we conclude that this method is insufficient for further study of gluon self-interaction or the mass-gap as the resulting Hamiltonian is extraordinarily complicated.

Ethan Jacob Gibson (Senior Thesis, April 2021, Advisor: Benjamin Frandsen )

Abstract

Quantum electrodynamics (QED) is a quantum field theory for electromagnetic interactions. By quantizing the electric field, we can analyze the properties of the electromagnetic interactions between quantum systems. Historically, QED has been able to predict the behavior of quantum particles with unprecedented success. However, the success of QED has relied upon perturbation theory to approximate results to a given order of precision. Physicists have largely abandoned analytical, closed form solutions to QED due to the problematic mathematical divergences in attempts to resolve systems non-perturbatively. By using the dressed vacuum as a basis, we may be able to renormalize systems which possess Hamiltonians of similar attributes. We use analytical tactics to avoid the divergence, and form a basis upon which the Hamiltonian can be normalized for the dressed vacuum, giving hope that we can obtain an analytical, closed-form solution for some systems in QED. By doing so, we accomplish an analytical, closed-form type of solution to problems within QED.

Kennedy Gifford (Capstone, April 2021, Advisor: Mark Transtrum )

Abstract

Streptococcus pneumoniae causes over 150,000 cases of pneumonia annually in the United States alone. We present a meta-analysis of publicly available raw sequence data representing host transcriptomes before and during pneumococcus infection and carriage. We divide studies into infection and carriage samples to further investigate the differences of these models. Using computational methods, we identify the differentially expressed genes and intracellular signaling pathways that change in human and mouse cells during infection and carriage with this bacteria to test if a general infection or carriage model in mice could adequately represent these model states in humans. We found no overlapping significant signaling pathways between the mouse and human studies in either model, indicating that the mouse infection model is not specific enough to direct therapeutics for human infection. These results also suggest that there is no clear and general connection between host infection and carriage models of pneumococcus between mouse and human samples appearing in transcriptomics. Our findings are relevant to understanding the underlying mechanism of how this pathogen causes disease and how we can better combat its effects through the development of improved prophylactics and/or therapeutics.

Abigail Graham (Senior Thesis, April 2021, Advisor: Darin Ragozzine )

Abstract

During its prime mission, the Kepler Space Telescope found over 700 systems with more than one transiting planet. These multiple planet systems (multis) are the most information rich and dynamically interesting of all exoplanets. We picked 46 multis where one planet was experiencing transit timing variations (TTVs) not obviously caused by the other known planets. TTVs are caused by interactions between planets and therefore can provide evidence of additional, hidden planets in these systems. We first tried to determine if the TTVs could be reasonably explained by the known planets. We then projected six possible hidden planets for each system and performed the same analysis on the hidden planet in the strongest resonance with the TTV-experiencing planet that was estimated to be stable. Five of our systems have good fits with the known planets, 39 have good fits with the hidden planet added, and two require more work to find a satisfactory answer. This work significantly improves our understanding of the architectures of some of the most interesting multis from Kepler.

Jarrod Hansen (Senior Thesis, May 2021, Advisor: Denise Stephens )

Abstract

Historically Brown Dwarf Binaries have been discovered utilizing large amounts of time on ground or space-based telescopes to determine the orbits of closely spaced objects. This makes these discoveries hard to accomplish as you have to spend several nights of observing on a single object in the hope that it is a binary system. We examine utilizing distance measurements from GAIA and UKIRT to trim down target lists and identify the best binary candidates. We examine a sample of known binaries to determine how effective our method is at identifying known systems. We determine that luminosity measurements using distances are a good indicator for binarity.

Bryce Hedelius (Senior Thesis, June 2021, Advisor: Dennis Della Corte )

Abstract

Density-functional theory (DFT) is the gold standard for quantum chemistry calculations but is prohibitively expensive to use in molecular dynamics simulation. Deep learning has the potential to speed up these calculations by five orders of magnitude. Atomistic systems may be represented as point clouds so SE(3) equivarient neural networks (graph networks that preserve translational and rotational symmetry) provide the appropriate architecture. I present several SE(3)-Transformer models designed to predict DFT energy and forces. Given a point cloud of atoms, the network predicts the overall energy and the net force on each atom. Unlike other SE(3) networks, the SE(3)-Transformer architecture has attention weights that add degrees of freedom in the angular direction. I demonstrate that the network can learn inter-atomic energy and forces.

Kate Hendrickson (Senior Thesis, April 2021, Advisor: Darin Ragozzine )

Abstract

During the Kepler Space Telescope's 9-year mission, it discovered over 2,300 planets around other stars but could not document information about the angle between the orbital planes of planets in the same stellar system. As scientists have examined the Kepler data, these angles between exoplanetary orbits, termed mutual inclinations, have been largely overlooked because they typically cannot be fully inferred, even though they provide valuable insight into the formation and evolution of planetary systems. Mutual inclinations for the entire Kepler population have been estimated by a variety of researchers, but there are still many questions about whether or how the mutual inclination distribution depends on the system architecture. We are exploring what mutual inclination information can be derived from light curves of individual Kepler systems of multiple transiting planets. The strongest information comes from the ~26 systems with clear Transit Duration Variations (TDVs)–variations in the length of time a planet passes in front of its star–which likely come as the result of detectable nodal precession due to mutual gravitational interactions (Kane et al. 2019). Our photodynamical model, PhoDyMM, has been used to explicitly study mutual inclinations on unusual systems before, but most light curves are studied by fixing all longitudes of the ascending nodes to zero by default. Our project uses synthetic light curves of systems we create with known solutions to determine the accuracy and precision of our methods in determining mutual inclinations. Using synthetic light curves, we assess PhoDyMM’s ability to correctly infer mutual inclinations (and other parameters) under a variety of model assumptions and find it to be efficient and accurate. We will also present investigations of the Kepler-18 system.

Scott Johnstun (Senior Thesis, April 2021, Advisor: Jean-Francois Van Huele )

Abstract

The phase estimation algorithm is a powerful quantum algorithm with applications in cryptography, number theory, and simulation of quantum systems. We use this algorithm to simulate the time evolution of system of two spin-1/2 particles evolving under a Heisenberg Hamiltonian. The simulation is performed on both classical simulations of quantum computers and real quantum computers. We also introduce three optimizations to the algorithm: circular, iterative, and Bayesian. We apply these optimizations to our simulations and investigate how the performance improves. We find that the circular and Bayesian optimizations exhibit the best performance in noiseless simulations and the iterative and Bayesian optimizations exhibit the best performance on quantum computers which are subject to noise. We also discuss the paradigms of iterative and update-based algorithms, which are attributes of these optimizations that can improve quantum algorithms generally.

Braedon Jones (Senior Thesis, April 2021, Advisor: Mark Transtrum )

Abstract

Superconducting resonance cavities are used in particle accelerators to accelerate beams of charged particles to near light speed. The fundamental limit to performance in these cavities is the maximum induced magnetic field that the superconductors can expel due to the Meissner effect. Traditionally, cavities have been made from Niobium; however, current technology has nearly reached the theoretical limit of performance for Niobium-based cavities. To overcome these limitations, Nb3Sn is being explored as a potential next-generation material. In actual development of Nb3Sn cavities, material defects arise that may limit performance. Time-dependent Ginzburg-Landau simulate deficiencies to explore if they cause detrimental effects to cavity performance. This research focuses on small ‘island’ regions containing deficits of Sn. These islands have been observed below the surface in real Nb3Sn cavities after fabrication. This research shows that these islands may affect performance if they are near the surface but become irrelevant when they are located more than a penetration depth below the interface.

Jenny Kang (Capstone, April 2021, Advisor: David Allred )

Abstract

A revolutionary solution for colorblindness was devised by Dr. Don McPherson who invented color corrective glasses for the colorblind to see colors in 2012. We are hoping to create greater empathy for colorblind individuals and to assist in creating a more colorblind-accessible environment. Hence, converse to the colorblind corrective glasses, in this project we assumed a person with healthy color vision would perceive the world as a red-green colorblind person if only 540 nm to 570 nm of wavelengths were let through by the dichroic filters. We report on the creation of googles that causes people wearing them to lose red-green discrimination. In short, our goal is to allow non-colorblind individuals to experience red/green colorblindness. Seventeen volunteers completed Ishihara colorblind tests, sixteen of them with healthy color vision were diagnosed as severely red/green colorblind while taking the test with colorblind goggles on.

Spencer King (Senior Thesis, April 2021, Advisor: John Colton )

Abstract

Zinc oxide is a semiconductor with a wide direct band gap and optical properties advantageous for three-photon absorption. Apoferritin is a hollow protein previously used to achieve uniform growth of nanoparticles of other materials. I demonstrate that zinc oxide nanoparticles can be synthesized within ferritin using a Slow Chemical Reaction (SCRY) scheme and that native impurities can be utilized to exhibit strong visible luminescence during the three-photon absorption process. My results were analyzed and verified through TEM imaging, spectrophotometry, ICP-MS, and various other methods.

Sam Liechty (Senior Thesis, April 2021, Advisor: John Ellsworth )

Abstract

Nuclear fusion is being studied extensively. Low energy experiments may provide an answer to discrepancies between theoretical and observed values of fusion rates, particularly in detecting catalyzing electrons. To detect them, an accelerator beam line and detector must be designed with code to analyze the results. Since this process (d,p) involves inverse kinematics, an accelerator beam must be designed to work with such collisions. Various ion optics were modeled using Mathematica computing software including double quadrupoles, analyzing magnets, and einzel lenses. I created a functional accelerator beam line simulation and am ready to use the simulation to show what design is necessary for our laboratory. In addition, theoretical work concerning the focusing power of double quadrupoles and the electrostatic potential of a uniformly charged cylinder was explored which resulted in reducing the traditional two-dimensional integral to a one-dimensional integral.

Carsen Lindorff (Capstone, April 2021, Advisor: Nathan Powers )

Abstract

Eric Manner (Senior Thesis, April 2021, Advisor: Mark Transtrum )

Abstract

Time delays are an inherent part of real-world systems. Besides the resulting slowing of the system, these time delays alter the dynamics and often cause destabilization. It has been shown that a system that possesses the property of intrinsic stability (a stronger form of global stability) maintains its stability when subject to any time-varying time delays, e.g., constant, periodic, stochastic, etc. Here, we begin to examine the effects and uses of adding stochastic time varying time delays to certain gradient-based optimizers. These optimizers include the well-known gradient descent and the Adam Optimizer, where the latter is commonly used in neural networks for deep learning. We show that time delays in the Adam optimizer can significantly improve the optimizer's performance on certain objective functions. We also explore the conditions under which gradient descent is intrinsically stable. Finally, to cover a wider range of loss functions, we investigate a new property of gradient descent, termed almost intrinsic stability which describes the systems' ability to get arbitrarily close to being intrinsically stable. We then use this definition to numerically examine conditions under which an almost intrinsically stable, and hence the gradient descent system, can maintain its stability when exposed to stochastic time delays bounded by a given maximum delay.

Daniel McPherson (Senior Thesis, April 2021, Advisor: Karine Chesnel )

Abstract

Magnetic nanoparticles have a wide range of applications, from the engineering to the medical field. Understanding the properties of assemblies of magnetic nanoparticles is necessary to enhance the effectiveness of their use. A key feature of self-assemblies of magnetite nanoparticles is their superparamagnetic behavior. My current project focuses on understanding the dynamics of magnetic fluctuations in magnetite (Fe3O4) nanoparticles. The method of study is using x-ray resonant magnetic scattering (XRMS). We collected our XRMS data at SSRL. We study variations in the XRMS signal caused by assemblies of magnetite nanoparticles. A certain amount of data refinement is necessary to isolate the coherent x-ray scattering signal, also known as magnetic speckle. Upon isolation of coherent signal, speckle patterns collected throughout 1000 second periods are cross-correlated in order to study the magnetic fluctuation dynamics. We compare two-time cross-correlation maps collected at different temperatures throughout the superparamagnetic blocking transition.

Andrew Jared Miller (Masters Thesis, April 2021, Advisor: Scott Sommerfeldt )

Abstract

[Abstract]

Christena Oldham (Senior Thesis, April 2021, Advisor: Robert Davis )

Abstract

Sepsis (bacterial infection in the blood) currently requires 24 hrs. to detect the bacteria strain. However, sepsis can prove fatal if correct treatment is not administered sooner. Our solution is to use magnetic nanoparticle (MNP) extraction to complete the detection process within 3 hrs. Successful extraction requires chemically coupling MNPs with molecules that have an affinity to bacteria, such as bis(dipicolylamine) with Zinc (bis-Zn-DPA). With modified MNPs we have extracted bacteria with, on average, 100% accuracy. However, these results are consistent regardless of having the DPA molecule attached to the nanoparticles. Further research will clarify the current results. We continue to improve MNP synthesis and bacteria capture with the goal to implement our technique in clinical settings.

Hayden Oliver (Senior Thesis, April 2021, Advisor: Gus Hart )

Abstract

It is more important today than ever before to accelerate the discovery of new revolutionary materials. Advances in computing, transportation, spaceflight, and all other technology sectors are happening faster than ever before, and demand is on track to surpass supply for energy, food, and other resources if big advancements are not made in the next 50 years. New computational methods have been developed to assist in that effort with great success. A Moment Tensor Potential (MTP) is a machine-learned interatomic potential that can predict material properties at a fraction of the cost of density functional theory while maintaining comparable accuracy. This is shown proved useful in creating phase diagrams for ternary alloys. In this work, MTP has been used to identify a new stable intermetallic phase in the Co-Al-W high-temperature superalloy system which may provide improvements in efficiency and performance from jet turbine engines, nuclear powerplants, and find itself useful in other high-temperature industrial uses.

Logan Page (Senior Thesis, April 2021, Advisor: David Allred )

Abstract

Uranium, a nuclear fuel source, can oxidize and degrade in reactor conditions. Previous studies have shown oxidation resistance in a uranium-niobium alloy. The nature of the oxides that form on U-6Nb after long exposure to air was explored using neutron diffraction at Los Alamos National Laboratory. We deposited thin films of uranium-niobium alloys for oxidation studies. We used ellipsometry to quantify the oxide growth over time as a function of niobium content. We found that the oxide thickness increases linearly with the logarithmic of time. This study also supports the hypothesis that uranium and niobium oxides form a protective passivation layer on a uranium alloy, preventing oxidation and extending the life of the fuel.

Chanhyun Pak (Senior Thesis, April 2021, Advisor: Jean-Francois Van Huele )

Abstract

Quantum cloning, the process of duplicating information contained in an arbitrary quantum state is limited by the no-cloning theorem [Wootters, W., Zurek, W. A single quantum cannot be cloned. Nature 299, 802–803 (1982)]. We introduce the concept of InterDimensional quantum Cloning (IDC) and investigate its usefulness. Fidelity measures the effectiveness of imperfect cloning. Fidelity depends on the dimension of the space in which the information is stored. We use fidelity to measure the effectiveness of the IDC scheme. IDC aims to achieve higher cloning fidelity by navigating between various dimensions. We present a couple of IDC implementations for different dimensions. We identify cases where IDC may have the advantage over other quantum cloning methods.

Hannah Pfost (Senior Thesis, April 2021, Advisor: Robert Davis )

Abstract

While the fields of physics and international development may seem disparate, the insights gained from studying each one of them can improve understanding of the other. Here, I demonstrate that concept as applied to computational optics and historical memory. Thus, the purpose of this project is threefold: (1) to computationally model light transport through tissue, and use that model to inform choices about a physical system; (2) to determine the types of historical memory recommended in the final reports of truth commissions; and (3) to give evidence for the usefulness of human-centered design in both areas. To model light transport, I used a Monte Carlo simulation of light at 1602 nm in tissue. I found that properly focusing the beam of light in a tissue-spectrometer system resulted in a fractional increase of 8.044×10^-1 in the arterial signal-to-shot-noise ratio, with a fractional error of 1.807×10^-2. To investigate truth commissions, I classified coded data from 15 of the national-level final reports studied by the Global Truth Commission Index according to types of historical memory, divided into 58 distinct variables. I found that the most commonly cited form of historical memory, building a monument, was recommended in the final reports of 10 of the commissions studied; 7 commissions recommended public sensitization/awareness programs; and 7 mentioned creating a holiday or a day of remembrance.

Joshua Rasband (Senior Thesis, April 2021, Advisor: Mark Transtrum )

Abstract

Predictive models are key to understanding the behavior of physical systems. Effective models can facilitate understanding of a complicated system. Ineffective models may have a large number of parameters, leading to the phenomenon of sloppiness, characterized by large uncertainties in estimating parameter values from data. Sloppiness has previously been observed in many fields, including power systems, chemical kinetics, and systems biology. We observe that the Hodgkin- Huxley model, a canonical model of the action potential in the giant squid axon, is a sloppy model. We describe the Manifold Boundary Approximation Method (MBAM), a technique for general model reduction. We use MBAM to construct minimal versions of the Hodgkin-Huxley model of the action potential for two example behaviors. These minimal models can better inform large- scale simulation of neurons in addition to lending important insight into biologically conserved characteristics of the neuron.

Emma Rasmussen (Senior Thesis, April 2021, Advisor: John Ellsworth )

Abstract

Laboratory nuclear astrophysics reactions typical of stellar energies are becoming of great interest, specifically the second step of the p-p chain which shows the release of a gamma from a proton-deuteron fusion. It is hypothesized that instead of a gamma, an electron can be released. We wish to validate an observation of this catalyzing electron from this p-chain type reaction. To accomplish this, I investigate extending the capabilities of a typical silicon ∆E/E charged particle detector to include electrons by adding a plastic calorimeter. Efforts to construct and model the detector using Geant4 simulation software are reported here.

Abstract

Magnetic nanoparticles, such as Fe3O4, are responsive in a magnetic field. Three different sizes of Fe3O4 nanoparticles (5 nm, 12.5 nm, and 20 nm) were probed by vibrating sample magnetometry (VSM) and muon spin relaxation (μSR). VSM is a bulk magnetic probe, and μSR is a local magnetic probe. Particles of 5 nm, 12.5 nm, and 20 nm average sizes were found to exhibit strong superparamagnetic characteristics and distinct blocking temperatures. The blocked state transition occurred between 3 K and 45 K for the 5 nm particles, between 80 K and 160 K for 12.5 nm particles, and between 150 K and 300 K for the 20 nm particles. Both the VSM and μSR techniques showed spin-flip energy and magnetic anisotropy.

Gabe Richardson (Senior Thesis, April 2021, Advisor: David Allred )

Abstract

Miniature spectrometers are of great interest to NASA as necessary instrumentation is scaled down and optimized for specific space applications. Semiconductor nanocrystals called quantum dots (QD) are being used to create a miniature high-resolution filter-based spectrometer, with the goal of use in space within five years. Computational imaging techniques—such as automated image analysis and mathematical spectrum reconstruction algorithms—are two of the key aspects to making the QD spectrometer a reality. This thesis discusses the process of developing these computational methods, along with the improvements that have occurred from previous work.

Jason Saunders (Senior Thesis, April 2021, Advisor: Jean-Francois Van Huele )

Abstract

Quantum resources, such as entanglement, can decrease the uncertainty of a measurement procedure beyond what is classically possible. This phenomenon is well described for systems with asymptotically many measurement resources by the Quantum Cramer-Rao Bound, but no general description exists for the regime of limited measurement resources. I address this problem by defining a Bayesian quantifier for uncertainty suitable for the regime of limited resources, and by developing a mathematical description for two parameter-estimation procedures: using qubit probes to estimate a rotation angle induced on them, and using a Mach-Zehnder interferometer to estimate a phase shift. I simulate the qubit metrology scheme in the regime of limited resources. I show that, in noiseless systems, entanglement between qubits always decreases the uncertainty of the estimation; however, the quantum advantage decreases as fewer qubits are used in the estimation. I also show that the presence of strong dephasing noise removes the quantum advantage completely, regardless of the number of qubits used.

Nathan Schwartz (Senior Thesis, April 2021, Advisor: John Colton )

Abstract

A microwave cavity, also known as a radio frequency cavity, is a specific type of resonator. Typical cavities consist of a closed metal structure that acts as a truncated waveguide for electromagnetic fields in the microwave spectrum. Technological devices that utilize microwave cavities are widespread and include filters, amplifiers, and oscillators. However, the computational resource cost and time expense required to find solutions to these cavities are prohibitive. Thus, a set of neural networks has been developed to find the frequency and mode characteristics of any given cavity configuration in ~.05 seconds with a mean absolute error of 0.012 for a normalized frequency range of 0.143-1.949, a cumulative mean absolute error of 0.0022 for 200 coefficients, and an accuracy of 95.46% for mode predictions. The creation, refinement, and final outputs of the neural networks are discussed.

Jackson Steele (Senior Thesis, April 2021, Advisor: Mike Joner )

Abstract

We typically measure the redshift for distant galaxies using spectroscopy. Although spectroscopy is highly accurate, it requires the use of large telescopes (the time on which is limited). Photometric redshift, another technique, is highly imprecise but can be done on much smaller telescopes. We have created a new method to measure redshift photometrically that is an order of magnitude more precise than other photometric redshift techniques. To do this, we use three specialized narrow-band filters: two filters with variable (linear) transmission that are sloped oppositely (called 'ramp' filters), and a third 21nm FWHM filter to measure the continuum. These isolate the Hα emission-line for galaxies with 0.01 < z < 0.03. Because the transmission is variable in our ramp filters, the brightness of Hα in each filter is a function of the redshift. We have tested this method observationally with 16 Seyfert galaxies and computationally with 197 emission-line galaxies. We are able to predict the redshift with a standard error of 572 km/s. This error decreases for galaxies with stronger Hα emission, and we find that the error drops to 252 km/s if the Hα equivalent width is over 40 Angstroms.

Benjamin Szamosfalvi (Senior Thesis, April 2021, Advisor: Jean-Francois Van Huele )

Abstract

We verify silicon photonic circuit designs for photon number state tomography. We perform homodyne detection using a 90-degree optical hybrid receiver, then reconstruct the input signal’s photon number statistics using a maximum likelihood Bayesian estimator. We also demonstrate the potential to reconstruct the photon number statistics of an unknown quantum optical state without single photon detectors or knowledge of the phase of the input states. These devices may be useful for chip-scale quantum information processing tasks in communications, sensing, and computing.

Jonathan Treter (Senior Thesis, June 2021, Advisor: Mark Transtrum )

Abstract

Neurons are complex physical systems with many interacting components. The foundational model of neural behavior is the Hodgkin-Huxley model. It models the cell membrane as a capacitor and protein ion channels as voltage-dependent resistors. The membrane voltage responds to an applied current and is calculated as a system of differential equations using standard circuit analysis. The Hodgkin-Huxley model involves four dynamical variables and 26 parameters; however, previous work explicitly constructing a reduced-order approximation showed that many of these parameters are irrelevant. A more realistic model from Buchholtz et al. expands on the model of Hodgkin-Huxley and involves 14 dynamical variables and 68 parameters. We implement the Buchholtz model in the Julia programming language and conduct a “sloppy model” analysis of the parameters. We show that this model is sloppy, meaning the importance of parameters used to explain the model behavior is exponentially distributed. Most of this behavior can be explained by a reduced number of combinations of parameters, suggesting that the model can be approximated by a low-order, reduced model. This work lays the foundation for a future parameter reduction analysis to find a simplified version of the Buchholtz model.

Jason Trump (Senior Thesis, April 2021, Advisor: M. Jeannette Lawler )

Abstract

Many planetariums are situated at institutions of higher learning, but there is little documentation about how these facilities are being used. We present an analysis of a survey designed to explore planetarium use in introductory astronomy courses taught to undergraduates. The survey asked about 11 learning objectives, which were chosen through an investigation of online course descriptions at ten universities in the United States. Planetarium users answered questions about what they are teaching, how long they are teaching it, and what media they are using to teach it. We distributed the survey to approximately 289 institutions around the United States which were categorized as institutions of learning in the online Worldwide Planetariums Database. There were 85 responses to the survey with 78 providing enough information to be useful. Results show that planetariums are primarily being used to teach the night sky and that planetarium users prefer to teach through unscripted use rather than scripted shows. We discuss potential implications to content development and further research in instructional methodology.

Savanah Turner (Senior Thesis, April 2021, Advisor: Denise Stephens )

Abstract

While many new brown dwarf observations have been made since the first was discovered in 1995, there is still much that remains unknown about these failed stars. Many stellar formation theories fail to explain the existence of objects with masses as low as brown dwarfs. One key to learning more about the formation of low-mass objects will be determining an accurate binary fraction, helping to either support or reject the different formation models based on their various binary fraction predictions. However, because brown dwarf binaries have small angular separations they can be difficult to visually resolve. A spectral fitting code has been developed here at Brigham Young University that works to identify unresolved brown dwarf binaries by statistically comparing the spectrum of the target object to model brown dwarf spectra. Spectral fitting is the only way to find many binary pairs, and thus is critically important to the efforts to find the binary fraction. I have improved upon this code by adapting it to use real brown dwarf spectral data rather than models in the fitting routine. The new code is cleaner, faster, and more user-friendly. I have verified my results by comparing them to the findings of others in the field. This improved spectral fitting code will enable our team to more efficiently search for unresolved binary brown dwarf systems and learn more about this universe we call home.

Zach Westhoff (Capstone, April 2021, Advisor: Robert Davis )

Abstract

The development and features of a microcontroller-based system for collecting data from a novel infrared spectrometer, intended for eventual use as a noninvasive glucose sensor, are presented. Sample spectra were collected with no incident light, reflected from tinfoil, and transflected from a human subject’s finger. Pulsatile data from a human subject’s finger is shown. System characteristics such as stability over time and temperature dependence were analyzed.

Kinamo Williams (Senior Thesis, April 2021, Advisor: Mark Transtrum )

Abstract

Interatomic models (IMs) are used in molecular modeling to predict material properties of interest. The development of an IM can take several months to years and relies on expert intuition, and yet these potentials are usually only valid for a particular application of interest. Extending existing IMs to new applications is an active area of research. Quantifying the uncertainty of an IM can tell us how much we can trust the predictions it makes. I discuss here two methods for analyzing uncertainty: Fisher Information Matrix (FIM) and Markov Chain Monte Carlo (MCMC). Using MCMC methods, I sample from the posterior distribution of the parameters when trained on data. I demonstrate this method on Lennard-Jones and Morse potentials fit to triclinic crystal configurations from the OpenKIM database. In particular, IMs are often sloppy, i.e., have likelihood surfaces with long, narrow canyons and broad, flat plateaus. I will be comparing the benefits and drawbacks of the two methods.