12023 research outputs found
Sort by
BICEP Array Detectors and Instrumentation at 30/40 GHz: Design, Performance, and Deployment to the South Pole for Constraining Primordial Gravitational Waves
Get PDF The discovery of the Cosmic Microwave Background (CMB) in the 1960s has provided strong observational evidence for the Big Bang cosmological model to describe the origin and evolution of the universe. The theory of cosmic inflation was developed in the 1980s to account for the initial density perturbations by a period of exponential expansion in the early Universe to solve the horizon, flatness and monopole problems. Many inflation models predict potentially detectable primordial gravitational-waves (PGWs) background that imprint a B-mode polarization pattern in the CMB. The amplitude of the inflationary B-mode polarization depends on the energy scale of inflation and is parameterized by the tensor-to-scalar ratio r. The detection of a B-mode pattern would open a new window to probe the energy scale at the beginning of time when the universe was a mere fraction of a second old after the Big Bang.
The BICEP/Keck collaboration is building a series of experiments located at the Amundsen-Scott South Pole Station to map the polarization of the CMB at degree angular scales using small-aperture telescopes. Our latest BICEP/Keck publications use data collected through 2018 and report the strongest constraints r0.05 < 0.036 at 95% confidence. The current sensitivity on r is limited by the variance from the gravitational lensing. BICEP/Keck is starting a collaboration with the South Pole Telescope (SPT) team to develop delensing techniques to improve future constraints on r. Characterizing Galactic foregrounds, especially synchrotron emission, remains a priority in order to improve constraints as statistical sensitivity continues to improve. The motivation for this thesis is to develop a highly sensitive receiver at 30 and 40 GHz, at frequencies where the synchrotron foreground dominates. BICEP Array represents the latest phase in the BICEP/Keck experiments, and will map the polarization of the CMB at 30/40, 95, 150, and 220/270 GHz. BICEP Array will search for PGWs with unprecedented sensitivity levels on r by characterizing and removing Galactic synchrotron and dust emission from our maps of the CMB.
My PhD thesis focuses on the technology development for high sensitivity detectors and instrumentation to successfully deploy the first BICEP Array receiver at 30 GHz and 40 GHz to the South Pole in order to constrain the Galactic synchrotron foreground. My dissertation presents the receiver design and performance. I will first explain the engineering design principles, the fabrication and a laboratory demonstration of single-color antenna-coupled Transition Edge Sensor (TES) bolometers. Secondly, I will discuss the design and demonstration of dual-color detectors at 30 and 40 GHz that gain receiver sensitivity by increasing the optical throughput and bandwidth of each pixel. I also developed microstrip diplexer circuits that divide the detector bandwidth into two CMB observing channels. I optimized this approach to design the dual-color bowtie-coupled detector at 90/150 GHz. Thirdly, I will introduce a new wide-band corrugated focal plane module design to minimize the beam mismatch systematic at 30 and 40 GHz bands simultaneously. Our receivers map polarization of the CMB by taking the difference between co-located and orthogonally polarized pair of detectors. Polarized beam difference measurements show a differential beam response due to a shift between the polarization beam centers within a pixel due to an electromagnetic interaction with the focal plane frame. The residual beams leak a temperature to polarization (T-P) in the CMB polarization maps and can produce a false B-mode signal that introduces non-negligible systematic errors for BICEP Array measurements to come with improved sensitivity. The wide-band design reduces this effect and associated systematic errors for 30 and 40 GHz receiver. I also developed a new single-band corrugated focal plane module design for 150 GHz receiver. I performed laboratory measurements of these designs at 30, 40, and 150 GHz to verify the modelled response. The corrugation design will also be extended to the 220/270 GHz receiver. Fourthly, I will show my contributions to the receiver deployment, integration and calibration during the first 2020 observing season. The measurements will include the full optical characterization of the detector camera, in-lab and on-sky sensitivity at the South Pole. I will also describe the tests done to diagnose the challenges during the first season and new upgrades during the second 2022 season to improve the overall sensitivity of the receiver. Improved detector modules have been installed during the 2023 season to further boost the mapping speed for measuring the synchrotron foreground.
The technologies developed for BICEP Array feed into capabilities for the upcoming CMB-S4 program. For example, I used similar methods to design a diplexer for a CMB-S4 dual-color feedhorn-coupled detector design at 90/150 GHz. I will also detail my work on the cryogenic implementation and test of an Adiabatic Demagnetization Refrigerator suitable for demonstrating 100 mK CMB-S4 detector arrays in a prototype 95/150 GHz telescope planned to observe on the BICEP Array.</p
Essays in Health Economics
Get PDFThis thesis consists of three health economics papers, two studying the effectiveness of policy interventions on the opioid epidemic, and one on the effects of air pollution on school absences. The first two chapters were coauthored with Shiyu Zhang, a former Caltech graduate student.
The first chapter examines the market for prescription opioids following the OxyContin Reformulation, an event that made OxyContin harder to misuse. Using detailed prescription opioid sales data from 2006 to 2014, we show that event did not reduce overdose deaths but led individuals to switch to generic oxycodone as a substitute for OxyContin.
The second chapter examines geographic spillover effects from state prescription drug monitoring programs (PDMPs). We show that these policies reduce prescription opioid sales and opioid overdose deaths in the state they are enacted in. However, because they only track opioids sold locally, these programs induce individuals to drive across state lines to purchase opioids and avoid these regulations.
The final chapter examines the effects of air pollution on NYC school absences using daily changes in wind direction. I show that PM2.5 and Ozone concentrations are strongly influenced by wind patterns, and exposure to these two pollutants causes detectable increases in absences over the following two days. Reductions in PM2.5 pollution over time have prevented approximately 381,000 absences annually in NYC which increases school funding by $19 million.</p
Ultrasound Controlled Drug Delivery by Acoustically Switchable Hydrogels
Get PDFNot only is ultrasound widely used as a diagnostic imaging modality, it can also be focused into deep tissues to perform non-invasive actuation of cells, implants and delivery vehicles and other biological targets. With the addition of gas vesicles (GV), generic hydrogel materials gain the ability to communicate with ultrasound, equipping them with in vivo tracking, targeting and actuation capabilities to safely transport biomolecular cargo. This is possible as GVs function simultaneously as ultrasound contrast agents and steric blockers that can be "erased" by an increase in ultrasound pressure to trigger a rapid outflow diffusion of the payload from within the material. We evaluate this concept through in vitro measurements of ultrasound-modulated diffusion and drug release and targeted in vivo release in the lower gastrointestinal tract. Then we demonstrate the use of orally administered hydrogel particles to deliver etanercept in the duodenum to treat gastrointestinal inflammation in a rat model of colitis. Finally, we explore new directions and applications of GV-hydrogel systems, showcasing their potential for deployment in a wide range of biomedical applications.</p
Optogenetic Approaches for Determining the Temporal Role of Morphogen Inputs on Target Gene Expression
Get PDFThe Dorsal transcription factor and morphogen is important for patterning the Dorsal- Ventral axis of Drosophila melanogaster and while it has been extensively studied, the temporal dynamics of Dorsal are not well understood. There are many processes that contribute to Dorsal nuclear concentration levels, including Toll signaling and Cactus degradation, interactions with other proteins, shuttling of Dorsal to the ventral side, DNA binding, and nuclear spacing. Dorsal nuclear levels are known to activate or repress target gene expression in a concentration or threshold dependent manner. To test how Dorsal dynamics and changes to the Dorsal gradient over time affect target gene expression, we added two optogenetic tags to Dorsal at the endogenous locus to control Dorsal nuclear levels: Blue Light Inducible Degradation (BLID) and Light Inducible Nuclear Export System (LEXY). We found that upon degradation of Dorsal using blue light and BLID that a downstream ratchet was able to maintain the expression of high threshold target genes. Using blue light and LEXY to export Dorsal, we identified an important window where Dorsal activity is required to allow activation of high threshold target genes at later stages. In comparing BLID and LEXY in conjunction with mutations to a nuclear export sequence, we also identified how rapid nuclear import and export of Dorsal is sufficient for low threshold target gene expression but actively disrupts high threshold target gene expression. We conclude that not only are final concentration levels, but also the dynamics leading to those levels are important for proper gene expression
Low-Rank Matrix Recovery: Manifold Geometry and Global Convergence
Get PDFLow-rank matrix recovery problems are prevalent in modern data science, machine learning, and artificial intelligence, and the low-rank property of matrices is widely exploited to extract the hidden low-complexity structure in massive datasets. Compared with Burer-Monteiro factorization in the Euclidean space, using the low-rank matrix manifold has its unique advantages, as it eliminates duplicated spurious points and reduces the polynomial order of the objective function. Yet a few fundamental questions have remained unanswered until recently. We highlight two problems here in particular, which are the global geometry of the manifold and the global convergence guarantee.
As for the global geometry, we point out that there exist some spurious critical points on the boundary of the low-rank matrix manifold Mᵣ, which have rank smaller than r but can serve as limit points of iterative sequences in the manifold Mᵣ. For the least squares loss function, the spurious critical points are rank-deficient matrices that capture part of the eigen spaces of the ground truth. Unlike classical strict saddle points, their Riemannian gradient is singular and their Riemannian Hessian is unbounded.
We show that randomly initialized Riemannian gradient descent almost surely escapes some of the spurious critical points. To prove this result, we first establish the asymptotic escape of classical strict saddle sets consisting of non-isolated strict critical submanifolds on Riemannian manifolds. We then use a dynamical low-rank approximation to parameterize the manifold Mᵣ and map the spurious critical points to strict critical submanifolds in the classical sense in the parameterized domain, which leads to the desired result. Our result is the first to partially overcome the nonclosedness of the low-rank matrix manifold without altering the vanilla gradient descent algorithm. Numerical experiments are provided to support our theoretical findings.
As for the global convergence guarantee, we point out that earlier approaches to many of the low-rank recovery problems only imply a geometric convergence rate toward a second-order stationary point. This is in contrast to the numerical evidence, which suggests a nearly linear convergence rate starting from a global random initialization. To establish the nearly linear convergence guarantee, we propose a unified framework for a class of low-rank matrix recovery problems including matrix sensing, matrix completion, and phase retrieval. All of them can be considered as random sensing problems of low-rank matrices with a linear measurement operator from some random ensembles. These problems share similar population loss functions that are either least squares or its variant.
We show that under some assumptions, for the population loss function, the Riemannian gradient descent starting from a random initialization with high probability converges to the ground truth in a nearly linear convergence rate, i.e., it takes O(log 1/ϵ + log n) iterations to reach an ϵ-accurate solution. The key to establishing a nearly optimal convergence guarantee is closely intertwined with the analysis of the spurious critical points S_# on Mᵣ. Outside the local neighborhoods of spurious critical points, we use the fundamental convergence tool by the Łojasiewicz inequality to derive a linear convergence rate. In the spurious regions in the neighborhood of spurious critical points, the Riemannian gradient becomes degenerate and the Łojasiewicz inequality could fail. By tracking the dynamics of the trajectory in three stages, we are able to show that with high probability, Riemannian gradient descent escapes the spurious regions in a small number of steps.
After addressing the two problems of global geometry and global convergence guarantee, we use two applications to demonstrate the broad applicability of our analytical tools. The first is the robust principal component analysis problem on the manifold Mᵣ with the Riemannian subgradient method. The second application is the convergence rate analysis of the Sobolev gradient descent method for the nonlinear Gross-Pitaevskii eigenvalue problem on the infinite dimensional sphere manifold. These two examples demonstrate that the analysis of manifold first-order algorithms can be extended beyond the previous framework, to nonsmooth functions and subgradient methods, and to infinite dimensional Hilbert manifolds. This exemplifies that the insights gained and tools developed for the low-rank matrix manifold Mᵣ can be extended to broader scientific and technological fields.</p
Inorganic Phototropism: Emergent Properties Directing Growth of Mesostructured Semiconductors
No full textNature exhibits emergent growth phenomena where in neighboring features result in ensemble effects that direct the overall growth morphologies. Plants, such as palm trees, display phototropism where-in the crown grows toward the time weighted average position of the sun to optimize solar collection. A methodology, known as inorganic phototropic growth, utilizes a similar mechanism with the incident illumination during electrochemical deposition directing the growth of mesostructured semiconductors. This photoelectrochemical deposition process, generates highly anisotropic, periodic lamellar features resulting in the capability to fabricate nanostructured features over macroscopic areas. The process is lithography-free and uses no templates or directing agents of any kind and relies solely on the incident illumination to direct semiconductor growth. In this thesis, the nanophotonic phenomena and emergent synergistic absorption that drives the inorganic phototropic growth process was investigated using unconstrained and confined susbtrates. Additionally, the impact of inclined, off-normal incident illumination on the evolution of structure morphology was investigated for patterned and isotropic substrates revealing the mechanism behind the non-monotonic relationship between incident angle and observed out-of-plane orientation for unconstrained substrates.</p
Insights into the Sources of Atmospheric Aerosols and Greenhouse Gases in California
Get PDFA substantial fraction of atmospheric science research is motivated by uncertainties in the sources of urban particulate matter and greenhouse gases. Such a focus is justified, as particulate matter exposure is responsible for up to nine million annual premature deaths globally, while climate change is rapidly altering ecosystems across the world. Recognizing the urgency of these interrelated problems, regulatory agencies in the U.S. and elsewhere have sought to limit emissions contributing to air quality degradation and global warming.
In this dissertation, we use a combination of ambient measurements, statistical models, and computational models to identify the sources of urban particulate matter and methane in multiple locations in California. In Los Angeles, we investigated the effects of reductions in mobile source pollutant emissions (i.e., on-road and off-road vehicles) on ambient aerosol concentrations. Mobile sources have historically accounted for the dominant fraction of urban particulate matter in Los Angeles, but despite notable reductions in their emissions over the last decade, ambient aerosol concentrations have not declined appreciably. Measurements using an Aerosol Mass Spectrometer demonstrate the complex interplay of direct (i.e., intended) and indirect effects of simultaneous reductions in organic aerosol (OA) precursor and nitrogen oxide emissions from these sources. Mobile sources are found to account for a modest and declining fraction of the total aerosol burden, while the contributions of non-traditional sources such as volatile chemical products (e.g., paints and coatings, cleaning products, adhesives and sealants) have increased. Simulations of organic and inorganic aerosol formation informed by in-situ measurements are developed to identify possible targets of future regulatory efforts.
In the San Joaquin Valley, we used airborne measurements of methane fluxes to evaluate dairy emissions inventories used by state regulatory agencies for policy development. Dairy operations currently account for nearly half of the state’s methane emissions, and recent legislation has mandated a 40% reduction in emissions by 2030. Observed methane fluxes align well with emission inventory predictions and demonstrate the utility of airborne flux measurements to track emission reduction progress in the future. Factor analysis of a combined dataset of greenhouse gas and volatile organic compound concentrations indicates dairy operations account for ~65% of total methane emissions in the southern San Joaquin Valley, with the remainder attributed to fugitive oil and gas emissions.</p
Distributed and Localized Model Predictive Control
Get PDFThe increasing presence of large-scale distributed systems highlights the need for scalable control strategies where only local communication is required. Moreover, in safety-critical systems it is imperative that such control strategies handle constraints in the presence of disturbances and enjoy theoretical and performance guarantees. In response to this need, we present the Distributed and Localized Model Predictive Control (DLMPC) algorithm for large-scale linear systems. DLMPC is a distributed closed-loop model predictive control (MPC) scheme wherein only local state and model information needs to be exchanged between subsystems for the computation and implementation of control actions. The resulting distributed algorithms tackle various types of additive disturbances and enjoy recursive feasibility and asymptotic stability guarantees that introduce minimal conservatism and can be computed in an offline fashion without adding to the computational burden. We also provide analysis and guarantees on the global performance of DLMPC, and demonstrate that in cases where the underlying topology of the system is sparse (as is the case in most large-scale networks), the inclusion of local communication constraints does not result in a suboptimal solution. Moreover, we show that when no noise is present, this algorithm can be extended to the purely data-driven case where all previous guarantees hold and the need for a model is fully replaced by past-trajectory data. We show that the amount of data needed for our synthesis problem is independent of the size of the global system. Lastly, we explore the potential of DLMPC for hardware accelerated implementation in GPU by exploiting the fact that the structure of the DLMPC problem captures some of the limitations of GPU computations. In all algorithmic and theoretical results presented in this thesis, only local information exchange is necessary, and computational complexity is independent of the global system size. DLMPC is the first MPC algorithm that allows for the scalable, efficient and data-driven computation and implementation of distributed closed-loop control policies and enjoys theoretical guarantees
Synthetic Circuits for Multicellular Spatial Patterning
No full textSelf-organized spatial periodic patterning mechanisms are responsible for the generation of repetitive structures, such as digits, vertebrae, and teeth, during multicellular development. Adopting a synthetic biology approach, we aim to unravel the core principles of multicellular spatial patterning by designing and reconstituting it in tissue-cultured cell lines.
The reaction-diffusion mechanism, as an established paradigm, has successfully elucidated and forecasted pattern formation across varying scales and species. However, the potential for reconstituting synthetic reaction-diffusion patterns using unconventional reaction-diffusion elements within mammalian cell cultures has been insufficiently explored, thus leaving a gap in our comprehension of how spatial periodic patterns could be generated.
The simplest reaction-diffusion systems are thought to necessitate a minimum of two morphogens to generate periodic patterns. In contrast, with the help of mathematical modeling, we illustrate that a simpler circuit, comprising only a single diffusible morphogen, can adequately produce long-range, spatially periodic patterns. These patterns propagate outward from transient initiating perturbations and remain stable after the disturbance is removed. Moreover, introducing an additional bistable intracellular feedback or operation on a growing cell lattice can enhance the robustness of the patterning against noise.
Concurrently, we reconstruct the Turing pattern in mammalian cell culture utilizing a bottom-up approach. We construct a synthetic circuit based on two signaling pathways. After validation of each circuit component, we exhibit the spatial pattern formation driven by a synthetic reaction-diffusion circuit within the mammalian cell line. This adaptable circuit facilitates us to adjust circuit parameters or implement various boundary conditions, thereby revealing the impact of these alterations on patterning dynamics.
Collectively, these findings lay the groundwork for the engineering of pattern formation in the nascent field of synthetic developmental biology.</p
Neurotechnology for Multiplexed Interrogation of Brain Circuits and Synaptic Activity
Get PDFThis thesis describes the development of neural technologies for 1) multiplexed brain circuit electrophysiology (ephys) recordings and control of activity in optogenetic mice lines with concurrent recording paired with two-photon imaging and 2) multiplexed measurements of synaptic release events in microfluidic platforms. The first part of this thesis describes efforts to provide deterministic correlation of excited neuron action potential with resulting ephys recordings in vivo. This consisted of technological development of novel, high density multisite silicon probes for electrophysiology recordings in vivo. The probes consist of four columns of electrodes densely packed at the shank tip. This density of electrode arrays allowed for higher resolution isolation of more distinct waveforms than previous ephys probes and benchmarking measurements to triangulate the locations of emitting neurons. These measurements help benchmark the ability of existing silicon extracellular probes to capture surrounding extracellular activity. When combined with two-photon imaging, we can simultaneously record ephys activity, image the probe and surrounding brain, quantify brain damage during probe implantation, and control neural activity using optogenetic mouse lines.
The second project described development of a microfluidic platform to monitor synaptic release of the neurotransmitter glutamate. Microfluidic devices were used to isolate synaptic processes expressing synaptic reporters and provide targeted recording of glutamate activity across the synapse. Synaptic glutamate release was monitored with a two part genetically encoded fluorescent reporter that detects glutamate released at the synapse, called split-iGluSnFR, developed in Professor Lin Tian’s lab at UC Davis. We designed new microfluidic devices to better isolate neuron processes with split-iGluSnFR and be compatible with existing fluorescent complementary metal–oxide–semiconductor (CMOS) contact imagers. Using computational fluid dynamic simulations, we demonstrate efficient perfusion in the device. The form factor of this new device is designed to be compatible with CMOS contact imagers, and that when combined will help us achieve our ultimate goal to monitor the kinetics of simultaneous synaptic release events modulated by perfused neuromodulating drugs.</p