2022 Projects
UW Physics REU 2022 Project List
Projects are offered from the following physics subfields:
Cosmology and astrophysics
Elementary particle physics
Nuclear physics and astrophysics
Atomic physics
Physics education
Condensed matter and nanostructure physics
Numerical modeling and simulations
Biophysics
Additional projects may be added to this list. If you have a special interest not represented in the list below, feel free to contact either Subhadeep Gupta or Gray Rybka for help. They may be able to design new projects that align with your interests.
Experimental Projects
Controlling electron-electron interactions in layered quantum materials
Arthur Barnard
Quantum materials formed by the stacking of atomically thin 2-dimensional crystals are host to strongly correlated electron behavior—electrons in these planar materials couple to each other via electronic and magnetic interactions, causing sometimes surprising physics to emerge. Magic-angle twisted bilayer graphene (MATBG) is a particularly exciting system of study, as it exhibits potentially unconventional superconductivity and topological conduction in a manner controlled by the twist-angle between two individual layers of hexagonally arranged carbon atoms. At present day, we don’t fully understand much of the observed emergent states because there are many parallel ways the electrons can couple.
In this project, the REU student will help in the endeavor of better understanding the observed superconductivity in MATBG by developing a tool to control how electrons experience electrostatic forces from their neighboring electrons. The student will develop a means of fabricating flat micron-scale conductive probes to controllably place in the vicinity of a MATBG sample. When the probe is particularly close, it will act to reduce the electrostatic repulsion among electrons in MATBG. By controlling the proximity of this probe, we will smoothly control the degree of "electrostatic screening" and help elucidate the way that electrons pair to form a superconducting state in MATBG. Beyond this particular application, the techniques developed during this REU project will have broad applications to a variety of other quantum materials.
Quantum Computing with Trapped Ions
Boris Blinov
In the trapped ion quantum computing lab at the University of Washington we experimentally investigate the techniques for building a conceptually new type of computational device. A quantum computer will be extremely fast at solving some important computational problems, such as the factoring and the database search. While days of practical quantum computing may be quite far in the future, we are developing the main building blocks of such a device – the quantum bits ("qubits"), the basic logic operations, the qubit readout... The physical implementation of the qubit in our lab is the hyperfine spin of a single, trapped barium ion. A student in this REU project will participate in experiments with laser-cooled, RF-trapped single ions, will help develop techniques for single- and multi-qubit manipulation via microwave-induced hyperfine transitions and ultrafast laser-driven excitations. They will gain valuable hands-on experience with lasers and optics, RF and digital electronics, and ultrahigh vacuum technology.
Deep learning techniques for tau calibration with the ATLAS detector at the Large Hadron Collider
Quentin Buat
The tau is a heavy cousin of the electron and one of the heaviest particles describing matter in the Standard Model of Particle Physics. As such, many predicted new particles are favored to couple to the tau, starting with the recently discovered Higgs boson. The large mass of the tau is also a curse as it means it will decay almost instantly after being produced. This leaves us with the task of inferring its presence, in most cases, through the detection of a set of pions, the most commonly produced particle at the Large Hadron Collider (LHC).
In 2013, the reconstruction of tau pioneered the usage of Boosted Decision Trees in the ATLAS experiment that was critical to establish the coupling of the Higgs boson to the tau lepton [1]. The adoption of recurrent neural networks led to a significant boost in the performance of tau detection [2]. More recently convoluted neural network and graph networks have shown promising results in other LHC-related applications and could improve the performance of the tau detection even further
In its first 10 years of operation, the LHC only produced 1/10th of the total dataset it is expected to produce. With a major upgrade of the accelerator planned in a few years, aggregate data rates will exceed 1 petabytes per second. Neural-network-based algorithms enabled by modern computing architecture could offer a unique opportunity to efficiently cope with such a gigantic amount of data.
The goal of the project is to improve the determination of the energy scale of hadronic decays of the tau leptons. Using a mixture density network architecture, we can combine measurements from the calorimeters and the tracking system of the ATLAS detector and also infer the energy resolution. You will explore this network architecture on ATLAS simulation and determine if it can replace the current approach.
The CCD module for DAMIC-M
Alvaro Chavarria
DAMIC-M is the next step in the DAMIC dark matter search program: a large array of charge-coupled devices (CCDs) capable of detecting extremely small energy depositions that ionize as little as two electron-hole pairs in the CCD silicon target. DAMIC-M will provide unprecedented sensitivity to low-mass dark matter particle candidates in the galactic halo that may interact with ordinary atoms in the target. The building block of the array is the CCD module, which consists of four CCDs integrated on a carrier wafer. The goal of this project is to demonstrate satisfactory mechanical, thermal and electrical performance of the CCD module by testing prototype modules in the laboratory.
Two-dimensional nanodevices
David Cobden
In our group we investigate new physics in electronic nanodevices made from combinations of two-dimensional van der Waals materials like graphene. Phenomena under study include topological phases, superconductivity, ferroelectricity, and charge and heat pumping. In this project the student will learn to make their own 2D devices with a particular physics goal in mind, and then carry out a range of measurements them which may include low temperature transport in high magnetic fields, scanning probe techniques, and optical spectroscopy.
Nuclear Spin Coupling to Nitrogen Vacancy Centers
Kai-Mei Fu
Single defects in crystals are promising candidates to host small quantum registers that can be connected by photonic links. The UW Quantum Technologies Teaching and Testbed laboratory (QT3) is building a small quantum testbed based on the NV center in diamond. One REU project is to implement a method to determine the coupling of a single defect in diamond to its nearest nuclear spins. This information will help us choose the particular defect that will be used for this testbed.
Using microwaves for nuclear physics
Alejandro Garcia
We are carrying out experiments to search for new physics by doing precision measurements of beta spectra.
We focus on the nuclear beta decays of 6He and 19Ne, produced in our local accelerator. The beta energies are determined using cyclotron radiation emission spectroscopy (CRES). The latter determines the beta energy by measuring the frequency of the cyclotron radiation emitted by the beta in a magnetic field. Typically, for B=1 Tesla and beta energies of approx. 1 MeV one gets frequencies of approx. 20 GHz. This student would get familiar with radiofrequency and nuclear techniques, such as, optimal use of low-noise RF amplifiers, tuning beams through our local accelerator, using scintillator detectors read by SiPM's (Silicon Photo Multipliers).
Radiofrequency Detectors for Particle Physics
Gray Rybka
Advances in ultra low noise microwave electronics have opened the door to a new generation of detectors with extreme sensitivity to very low energy signals. This project will be related to developing detectors that have applications to astroparticle physics: axion dark matter and neutrino mass measurements.
The Physics of Million-Year Old Concrete
Jerry Seidler
The long-term storage of spent nuclear fuel raises a grand challenge: How can the fuel be safely stored for about 1 Million years, after which its hazards to the biosphere will be small? The best strategies involve multiple layers of containment within deep geological repositories. In this project, we are studying a range of concrete material for use in such repositories, including especially samples that have undergone accelerated aging in environments that should elicit likely chemical and structural failure modes that could occur during the needed 1 Million year storage time. We are developing new experimental methods to understand the atomic and electronic structure of these materials. Students involved in this project will work closely with my team on x-ray spectroscopic study of the changing electronic structure of metal species in cements and also on surface electric potential mapping to watch the consequences of internal corrosion. The students will learn about these methods, their full range of applications and data analysis, and the many challenges that must be overcome for the safe disposal of nuclear fuels.
Optoelectronics of 2D Semiconductor Heterostructures
Xiaodong Xu
Heterostructures of 3D semiconductors are a central component of condensed matter physics and modern solid state technologies (such as diode lasers and high-speed transistors). The recent discovery of monolayer semiconductors offers exciting opportunities to engineer analogous 2D heterostructures for exploring a wide range of new properties and functionalities. In this project, the REU student will be involved in the investigation optoelectronic properties of the 2D heterostructures. The student will learn how to obtain monolayer semiconductors, heterostructure device fabrication and characterization, and be involved in optical measurements (such as polarization and time-resolved photoluminescence, Raman spectroscopy, second harmonic generation etc.).
Theory/Numerical Modeling Projects
Quantum Simulation with Interacting Photons
Arka Majumdar
Understanding correlated many-body effects, including high-temperature superconductivity and fractional quantum Hall physics is not only of fundamental scientific interest but also has the potential for transformational societal impact, with applications in faster, more efficient electronics and topological quantum computers. However, such systems are extremely difficult to study theoretically due to the massive computational resources required. An innovative solution is to build an alternative, well-controlled experimental platform to simulate the performance of these correlated electronic systems. Several physical systems, including cold atoms, ion traps, defect centers, and superconducting circuits, have been considered for quantum simulations. In particular, strongly interacting optical photons provide a unique approach to non-equilibrium many-body quantum simulation due to the ease of adding and destroying photons via external driving and photon loss, as well as measuring multi-photon correlations using readily available single photon detectors. Such interactions, however, require the realization and control of nonlinear optics (NLO) at the few photon level, a daunting task in practice. We have recently developed a platform using solution processed quantum dots coupled to an array of cavities. Unfortunately, in this platform, we cannot reach single photon nonlinearity easily. In the REU project, we plan to model this quantum system (using master equation and quantum trajectory method) and identify the steady-state observables, which exhibit quantum mechanical behavior, even in the absence of single photon nonlinearity. The project also aims to understand the use of this nonlinear coupled cavity array to problems such as quantum machine learning.
Learning the shape of protein universe
Armita Nourmohammad
Proteins play a central role in all parts of biology from immune recognition to brain activity. A key challenge is to predict how protein sequence and structure determines function, such as the protein’s binding affinity to ligands or its enzymatic activity. With the growth of molecular data, machine learning has become a powerful tool in protein science. However, these techniques often generate black-box models, which are powerful but hard to interpret. In this project, students work to develop AI-guided approaches to protein structure-function maps, which respect the physical symmetries in the representation of proteins, and hence, result in more interpretable models of protein micro-environments that could reflect the underlying biophysics.
This project will allow students to work with protein sequence and structural data as well as implementing basic statistical methods for machine learning in Python. This project is computationally intensive and students will learn to work with high performance computing (HPC) to analyze and interpret large data. Background in biology is not necessary but students should have a strong background in statistical mechanics and should be familiar with programming languages such as Python, Julia, or others.
Information processing and control of evolving populations
Armita Nourmohammad
Adaptive Darwinian evolution is an act of information processing: populations sense and measure the state of their environment and adapt by changing their configurations accordingly. Changes of the environment result in an irreversible out-of-equilibrium adaptive evolution, with a constant flow of information. In this project students will explore the fundamental limits in evolution and ecology by combining theoretical approaches grounded in statistical physics and information theory with molecular data. In particular, we aim to develop adaptive optimal control approaches for artificial selection to drive evolving populations towards desired phenotypic targets. The efficacy of such control approaches is limited by our ability to predict evolution — a theoretical limit that we will closely study in this project.
To tackle this problem, students will develop stochastic models for the evolutionary process and apply numerical and analytical techniques to study the underlying dynamics. Background in biology is not necessary but students should have a strong background in statistical mechanics and should be familiar with programming languages such as Python, Julia, or others.
Functional organization of the adaptive immune system
Armita Nourmohammad
The adaptive immune system consists of highly diverse B- and T-cell receptors, which can recognize and specifically react to a multitude of diverse pathogens. Over the past decade, there has been a growth of sequence data on immune receptors, along with more limited information on their interactions with pathogens. Statistical inference and AI-guided approaches to interpret these high-throughput immune repertoire sequences has shed light on the generation of immune receptor diversity and how biophysical features of immune receptors are selected for different function. The goal of this project is to construct data-driven approaches to infer sequence determinants of function for B-cell receptors, as they differentiate to perform distinct functions within an individual.
This project will allow students to work with high-throughput immune receptor repertoire data and will familiarize them with statistical modeling and inverse-inference techniques. This project is computationally intensive and students will learn to work with high performance computing (HPC) to analyze and interpret large data. Background in biology is not necessary but students should have a strong background in statistical mechanics and should be familiar with programming languages such as Python, Julia, or others.