Interns and Projects: Summer 2019
 Candidate SAO Summer Intern Program Projects, 2019

| Colloquium schedule | Aug 8 Symposium Program | AAS Abstracts |

1 Jéa Adams (Amherst College)

ADVISOR: Dr. Nimesh Patel (R&G)
MENTOR: Dr. Qizhou Zhang (R&G)

PROJECT TITLE: Triggered star-formation in the globules of IC 1396

Abstract: The bright-rimmed globules at interfaces of ionized hydrogen (HII regions) and dense molecular clouds have long been studied as potential sites of star-formation. These edges are compressed to densities high enough to trigger star formation, due to shock fronts caused by UV radiation from the OB stars in the HII region. But direct evidence of this triggering effect on star-formation is difficult to establish observationally, and obtaining a fuller understanding is challenging because it requires observations at multiple wavelengths, on large angular scales, and which also have the high angular resolution characteristic of molecular line emission from high density tracers, as well as continuum emission from dust. We have recently published our results on a newly discovered protostar using Herschel and IRAM 30m observations of the IC 1396 globule A. (Sicilia-Aguilar et al., 2019, A&A 622,A118). We have also obtained SMA observations of this protostar, as well as another new source in the same globule. In this project, we will analyze the dust continuum emission in this source, and CO 2-1 emission from the bipolar outflow, to further improve our understanding of the phenomenon of triggered star-formation.

2 Amanda Ash (University of North Georgia)

ADVISOR: Dr. Stephanie Douglas (SSP; NSF Fellow)
MENTOR: Amber Medina (SSP)

PROJECT TITLE: Magnetic activity variability in the Praesepe open cluster

Abstract: Stellar rotation and magnetic activity operate in a feedback loop where a rotationally powered dynamo produces magnetic fields, and then stellar winds carry mass and angular momentum away, slowing the star and weakening the dynamo. To measure rotation, we look for periodic brightening and dimming caused by starspots rotating into and out of view. To trace the magnetic field, we use indirect measurements like H-alpha and X-ray emission. Most surveys that compare rotation and activity don't account for potential activity variability, however, even though it's been shown that a single star's H-alpha measurement can vary significantly. Last winter, I carried out a survey of stars in the Praesepe cluster (M44) that tracked H-alpha simultaneously with photometric variability, in an attempt to correlate H-alpha active regions with starspots. The student will measure H-alpha emission from reduced MDM data and correlate H-alpha time-series with light curves from K2.

3 Zoe de Beurs (University of Texas at Austin)

ADVISOR: Dr. Saeqa (Saku) Vrtilek (HEA)
MENTOR: Dr. Nazma Islam (HEA)

PROJECT TITLE: Classifying X-Ray Binaries Using Machine Learning Methods

Abstract: Consisting of compact objects that accrete material from orbiting secondary stars, X-ray binaries have been observed for more than half a century. However, there is still no straightforward means to determine the nature of the compact object: a neutron star or a black hole. In this project, we compare three machine learning-based classification methods (Bayesian Gaussian Processes, K-Nearest Neighbors, and Support Vector Machines) to develop a tool for classifying the compact objects in X-ray binary systems.

Each machine learning method uses spatial patterns common to systems of the same type in 3-D Color-Color-Intensity diagrams. We test a Bayesian Gaussian Process model that has been used to classify RXTE/ASM sources using data from the more sensitive MAXI/GSC. Using the MAXI/GSC data, we reproduce the result that the model can accurately classify well-known X-ray binaries but sometimes mis-classifies non-pulsing neutron star systems containing "bursters" as black holes when they lie close to the boundary between black holes and neutron stars. We find that K-Nearest Neighbors and Support Vector Machines on average predict the correct classification with greater probability and speed than the Bayesian Gaussian Process. Overall, all three methods have high predictive accuracy. It is expected that the work will result in co-authorship of a refereed publication.

An oral presentation describing this work was given in November in Houston at the Gulf Coast Undergraduate Research Symposium, and won the prize for "Best Astrophysics Talk."

4 Frederick Dauphin (Carnegie Mellon University)

ADVISOR: Dr. Griffin Hosseinzadeh (OIR)
MENTOR: Dr. Edo Berger (OIR)

PROJECT TITLE: Selecting Superluminous Supernovae from Transient Surveys with Machine Learning

Abstract: Superluminous supernovae are a rare type of stellar explosion whose power source is poorly understood. In order to characterize the observational properties of the class and compare them to models, we need to significantly increase the sample size of well observed events. Today's transient surveys (e.g., the Zwicky Transient Facility) produce hundreds of supernova candidates per night, but only a small fraction of these can be classified spectroscopically. This means we need a way to predict which candidates will be superluminous supernovae based only on information available at discovery: how bright they are and where they happened in the sky. Previous work has shown that superluminous supernovae preferentially occur in dwarf galaxies, whereas other types of supernovae are more likely to be found in more massive galaxies. Although the correlation is not perfect, this type of contextual information can be used to sift through the stream of transient discoveries to produce a much smaller sample of events for spectroscopic classification that are highly likely to be superluminous supernovae. We propose to use supervised machine learning to carry out this filtering, which most astronomers currently do by hand. The intern would explore various machine learning algorithms and train them on the set of all previously classified supernovae. The resulting classifier would then be used on new discoveries to determine which candidates we should observe spectroscopically. If this method is more efficient than human filtering, it will have wide-reaching implications for the Large Synoptic Survey Telescope, which will discover transients at an even more unmanageable rate.

5 John Della Costa (University of Florida)

ADVISOR: Dr. Howard Smith (OIR)
MENTOR: Dr. Matthew Ashby (OIR)

PROJECT TITLE: Colliding Galaxies, Star Formation, and Black Holes

Abstract: Collisions between galaxies are ubiquitous, and drive massive bursts of star formation and the growth of supermassive black holes / active nuclei (AGN). Over cosmic time, these cataclysms help to build up the populations of stars and AGN seen today, and during their most active phases they light up the galaxies, turning them in to ultra- or even hyper luminous monsters that are detectable across billions of parsecs and deep into the cosmic past.

We have a ongoing project that has compiled the largest, most reliable, and most complete spectral energy distributions (SED) of hundreds of mergers and luminous galaxies using newly reprocessed photometry from space missions and ground programs extending the ultraviolet to the far infrared (about 33 wavelength bands). Using new modeling codes, we have extracted from these SEDs dozens of key physical parameters from star formation rates, AGN fraction, dust temperature, stellar mass, to the dust mass and character; we also have complimentary spectroscopic datasets. This unique sample has enough objects to offer excellent statistical analyses, we believe for the first time. The results so far have spotted important trends - including for example (1) the star formation rate versus the AGN fraction, and
(2) the mass of the supermassive black hole mass versus the stellar mass of the system.
Results have also identified outlier objects whose behaviors offer clues to the inner physical processes at work.

Significantly, we have extensive computational simulations of merging galaxies that can be used to test the modeling. First comparisons confirm the reliability of the modeling codes, but much more remains to be done. Key issues include the role of AGN feedback to stimulate or suppress star formation, the nature of the ionized interstellar medium, and the existence of a previously underestimated (but massive) cold dust component. Another key question is whether or not luminous galaxies in the early universe (whose morphology is not known) follow the same star formation processes we find in the local universe. This REU internship will emphasize work in all these areas. The intern will simultaneously gain familiarity with multiple datasets and with computational modeling. The student will work with our team to prepare one or more papers for publication about black holes, star formation, and luminous galaxies. Some emphasis will be given to improving the SED modeling with modules to enable more accurate retrieval of key physics.

An oral presentation describing this work was given in November in Houston at the Gulf Coast Undergraduate Research Symposium.

6 Hannah Gulick (University of Iowa)

ADVISOR: Dr. Sarah Sadavoy (R&G; Hubble Fellow)
MENTOR: Dr. Luca Matra (R&G)

PROJECT TITLE: ALMA Observations of Protostellar Disks

Abstract: Planets form in dense, flat disks around young stars. The seeds for planet formation are small particles of dust within these disks that must grow from micron sizes to pebbles and boulders. The timescales for this formation process are not well understood, and recent theoretical models suggest that a good metric to study the onset of dust grain growth is through polarized scattering. In this project, we explore why some disks around young stars show evidence of polarized dust scattering ( and hence grain growth) whereas other disks do not show this detection. One explanation could be that the masses and densities of these disks are too low for substantial dust grain growth at early stages. Another explanation could be that the undetected disks have poor geometry to detect dust scattering. Therefore, creating a model which explains the structure and geometry of protostellar disks is critical to use polarization to study how disks form planets at early (< 1 Myr) stages.

We use WL 17, a protostar with a known disk, to create a robust model. WL 17 has physical properties similar to disks that are well detected in dust polarization from dust self-scattering, however, recent polarization observations of this disk find it has very low polarization fractions that are within the ALMA instrument sensitivity limit. We fit WL 17's radial intensity profile with three simple functions: a Gaussian disk with a cavity, an asymmetric disk with a cavity, and a Nuker profile. We use our best-fit parameters to calculate the spectral index, optical depth, and geometric properties of WL 17 to determine why WL 17 is undetected in polarization and if disk properties or disk geometry affects the detection rate of dust scattering.

7 Caleb Harada (University of Maryland, College Park)

ADVISOR: Dr. Antonija Oklopcic (TA)
MENTOR: Dr. Ana Bonaca (TA, OIR)

PROJECT TITLE: Atmospheric escape in exoplanets

Abstract: A significant fraction of exoplanets discovered to date orbit their host stars at much closer separations than any of the Solar System planets. These close-in exoplanets are subject to intense stellar radiation, which can have dramatic effects on their atmospheres. The upper layers of a planetary atmosphere can get heated by stellar X-ray and UV radiation to temperatures of several thousand degrees, creating pressure gradients that drive a supersonic planetary outflow. Recently, the helium line at 1083 nm has been identified as an excellent probe of extended exoplanet atmospheres which contain evidence of atmospheric escape. The goal of this theoretical project is to establish what information about the uppermost layers of planetary atmospheres can be extracted from observations, particularly from transit light curves in the helium line at 1083 nm. The student will use a set of pre-generated models of escaping exoplanet atmospheres to produce synthetic transit light curves and investigate how the shape of the transit light curve depends on various properties of the planet and/or its host star.

Note: aspects of this project have been accepted for publication in the AJ.

8 Maryam Hussaini (University of Texas at Austin)

ADVISOR: Dr. Andra Stroe (HEA/OIR; Clay Fellow)
MENTOR: Dr. Grant Tremblay (HEAD)

PROJECT TITLE: Unraveling the physics of shock cluster galaxies

Abstract: Nearby clusters usually contain many old galaxies, with characteristic red colors indicating they stopped forming new stars long ago. Unlike peaceful nearby clusters, distant galaxy clusters undergo frequent cosmic collisions with other clusters. These collisions were commonplace in the early Universe, but astronomers recently realized they are just as prevalent in the nearby Universe. Studying far away galaxy clusters is difficult, but we can use nearby colliding clusters as probes of the early Universe. Cluster collisions drive dangerous cosmic `weather': giant shock waves and turbulence which travel through galaxies like enormous tsunamis and tornadoes. As the shock passes through them, the typical old, dormant cluster galaxies can actually be awoken to start forming new stars. Capitalizing on unique optical data from the Gemini telescope in Hawaii, the student will be the first to unveil the spatially resolved physics of 'shocked' cluster galaxies. The student will measure the star-formation rate, the presence of black hole activity and the nature and strength of the ionizing radiation and how these vary within the galaxies. The student will learn how to reduce and analyze integral-field unit spectroscopic data and will become familiar with concepts of cluster formation, galaxy evolution, star formation and emission line physics.

For either project led by Dr. Stroe, the student should have some knowledge of the Linux operating system, and with scripting and programming in general. Some knowledge of Python in particular would be useful but is not required.

An oral presentation describing this work was given in November in Houston at the Gulf Coast Undergraduate Research Symposium.

9 Kenneth Lin (Univ. of Massachusetts Amherst)

ADVISOR: Dr. Dong-Woo Kim (HEA)
MENTORS: Drs. Nazma Islam (HEA) and Ewan O'Sullivan (HEA)

PROJECT TITLE: The Extended Hot Gaseous Halos of Early-Type Galaxies with the X-Ray Galaxy Atlas

Abstract: Over the past two decades, X-ray observations have transformed our understanding of the galaxy population by allowing us to study their high energy content in exceptional detail. In particular, we are interested in the hot interstellar medium, which is the dominant gas phase in galaxies otherwise lacking a significant gaseous component. This has a large impact in our understanding of their properties (e.g., total mass) and of their formation and evolution, hence on the history of the Universe. For this reason, we have built an X-ray Galaxy atlas consisting of a large number of galaxies using the high-resolution Chandra data and wide-field XMM-Newton data. We have published the first version of the Atlas (see arXiv/1812.02718, GalaxyAtlas ) with the Chandra data alone. To complete our picture of these early-type galaxies, we incorporate XMM-Newton observations to exploit its larger field of view and effective area to study the outer regions of these systems where the gaseous halos extend, using a new pipeline modeled after the Chandra analysis pipeline framework, that takes advantage of our access to the computational power of the Smithsonian high-performance super-cluster, Hydra.

As a preliminary studey of our data products from the completion of the XMM-Newton Galaxy Atlas (NGA) pipeline, we present 2D maps and 3D radial profiles of the hot gas properties (temperature, density, metallicity, projected pressure, entropy, and mass) for two nearby giant ellipticals, NGC1550 and NGC4636, produced from XMM-Newton archival data to uniquely examine the outskirts of the extended hot gaseous halos. We show evidence for asymmetric gas temperature and metal distributions around the galactic cores in 2D surface brighteness, temperature, and for the first time, Fe abundance maps. We further compute the radial cooling-time profiles and the ratios between the cooling and free-fall times of the ambient hot gas to determine the possible astrophysical processes responsible for extended features in their temperature and abundance distributions. Using our spectral maps and radial profiles, we address the possible effects of AGN feedback and tidal interactions on the multiphase interstellar medium.

An oral presentation describing this work was given in November in Houston at the Gulf Coast Undergraduate Research Symposium.

10 Serena Moseley (Carleton College)

ADVISOR: Dr. Chelsea MacLeod (HEA)
MENTOR: Dr. Paul Green (HEA)

PROJECT TITLE: Long-Term Variability of Nearby Quasars

Abstract: The Time Domain Spectroscopic Survey (TDSS) is accumulating approximately 2600 high-S/N spectra of nearby quasars that were observed before by the Sloan Digital Sky Survey about a decade ago. These spectra will be used to study the long-term variability of the broad Balmer emission lines on rest-frame time scales of several years, which corresponds to the dynamical time of the broad-line region. The variability patterns will be used to test models for the dynamics of the gas in the quasar broad-line regions. The student's work will be to examine the variations in the available spectra and identify and classify the variability patterns. The inspection and classification will first be done by eye and then by making measurements of the Balmer emission lines. This work will give the student the opportunity to learn about manipulation of optical spectra, programming, and the basics of the physics of the gas in quasar broad-line regions.

11 Lily Whitler (Arizona State University)

ADVISOR: Dr. Charlotte Mason (OIR; Hubble & CfA Fellow)
MENTOR: Dr. Charlie Conroy (OIR)

PROJECT TITLE: Measuring the Timeline of Cosmic Reionization

Abstract: We cannot see the first galaxies which formed in the universe, but we can observe the impact they had on their surroundings. In the universe's first billion years the majority of hydrogen transitioned from neutral to ionized. This process, `Reionization', was probably caused by photons from the first galaxies, and is an exciting frontier of observational and theoretical astrophysics. The Hubble Space Telescope and biggest telescopes on the ground can observe galaxies at the end of Reionization. By studying these galaxies and how they are affected by Reionization we can learn how Reionization started and about the nature of the very first galaxies.

This project combines theoretical modeling and analysis of observations. The student will use a theoretical model of Reionization which predicts the properties of galaxies at our observational frontiers, using a combination of analytical techniques and large-scale simulations. The student will update the model using Python code to explore how the connections between dark matter and galaxies impact Reionization. The model will then be compared with existing observations of galaxies using statistical techniques to measure the timeline of Reionization. These results will be be pivotal in understanding how Reionization took place, and thus the properties of the first galaxies.

Note: this project has been written up and submitted to MNRAS.


Clay Fellow Warren Brown