Orbital Realignment of Hot Jupiters

Advised by Prof. Dan Huber and Prof. Sam Grunblatt

Stellar obliquity—the angle between the spin axis of a star and the vector normal to its planets' orbital plane—provides a crucial opportunity to trace the dynamical history of planetary systems. Observations of hot Jupiters around main-sequence stars have shown that stellar obliquity has a strong dependence on the star's effective temperature, with hot hosts displaying a wide range of obliquities, and cool hosts being preferentially aligned. The critical temperature is approximately at the Kraft break, around 6250 K, corresponding to the divide between stars with primarily radiative envelopes (≳6250 K) and stars with thick convective envelopes (≲6250 K). I identified a population of planetary systems which had not been examined in detail and offered a unique opportunity to investigate the source of these distinct distributions: hot Jupiters orbiting subgiants which were hotter than the Kraft break on the main sequence and have since cooled, crossing the Kraft break.

I proposed and executed observations with the 10-meter Keck-1 telescope to measure stellar obliquities for five subgiants hosting hot Jupiters identified by the GTG survey. The sky-projected stellar obliquity can be constrained by taking spectroscopic observations during a planet's transit to observe the Rossiter-McLaughlin (RM) effect. As the planet occults the rotating stellar surface, it imprints a trend in the RV signal.

These systems, which were part of a hot (>6200 K), typically misaligned main sequence population, were observed to be aligned after cooling and gaining deep convective envelopes, which are thought to dampen obliquities through tides. I produced stellar evolution models using the MESA code for these host stars to trace the evolution of stellar surface convective zones, and placed an upper limit of ~500 Myr on the timescale for realignment. I have obtained additional stellar obliquity measurements to improve this constraint and shed light on dynamical evolution.

• The Paper

Detecting Planets Around Evolved Stars with TESS

Advised by Prof. Dan Huber and Prof. Sam Grunblatt

The explosion of exoplanet detection in the last decade has enabled astronomers to study the underlying characteristics of planet populations. Planets orbiting main-sequence stars have been characterized in detail, however planets orbiting evolved stars remain mysterious due to the small sample of observed transiting planets around post-main-sequence stars.

I am conducting a search for planets around evolved stars in the TESS Full Frame Images (FFIs). I wrote a pipeline to quickly remove the scattered light from TESS which uses linear regression to isolate and remove the bright background from light curves. After de-trending the light curve, my pipeline performs a Box Least Squares (BLS) search for transiting planet signals. We selected a sample of subgiant and red giant stars, and have generated and searched over 1,000,000 light curves. Our survey has identified and confirmed seventeen new planets which were not identified by the NASA SPOC or MIT QLP pipelines, and has led to seven publications (see below) with multiple upcoming papers in prep.

• The Code
• The Papers

Constraining Weakened Magnetic Braking with Asteroseismic Rotation Rates

Advised by Prof. Jen van Saders

As stars age, they gradually lose angular momentum through stellar winds, and the rate at which they rotates decreases. Gyrochronology relates the rotation period of a star to its age. The spin-age relationship has been well characterized for young stars using nearby clusers, but remains unreliable for older stars. A population of old stars with faster than expected rotation prompted the theory of "Weakened Magnetic Braking," a stage of stellar evolution during which the star undergoes a magnetic transition that allows it to maintain rapid rotation.

I created stellar evolution models with MESA and trained an Artificial Neural Network (ANN) to emulate stellar evolution codes. With a new sample of benchmark stars with rotation periods measured from asteroseismology, I tested gyrochronology relations and found that the WMB scenario best describes the observed rotation rates. Sun-like stars appear to deviate from standard spindown slightly earlier than the age of the sun. I also found that while WMB provides the best fit to the data, it cannot completely describe the rotation behavior beyond solar age.

• The Paper

Creating High-precision K2 Light Curves

Advised by Prof. Rory Barnes and Dr. Rodrigo Luger

When the Kepler space telescope lost two of its four reaction wheels, the spacecraft lost the fine pointing necessary for high precision photometry. Engineers were able to reorient the telescope to observe along the ecliptic plane, held stable by the photon pressure from the solar wind, and the mission entered a second phase called K2. However, the spacecraft maintained a slow drift and was re-oriented every 6 hours by a stabilizing thruster fire, causing targets to traverse accross the CCD detector. Because of subtle variations in sensitivity between pixels, as well as within each pixel, K2 light curves have periodic systematic noise.

In order to better understand the characteristics of K2 noise and test noise removal methods, I developed a Python package called scope (Simulated CCD Observations for Photometric Experimentation) to simulate the pixel sensitivty variation of the Kepler detector and motion of targets, ultimately reproducing the noise properties of K2 light curves. Scope was used to test the performance of the everest pipeline, particularly in the case of high motion. The results of the paper linked below influenced the decision to continue observation and upload larger aperture masks for K2 campaign 17.

• The Code
• The Paper
• Examples

lightkurve

A friendly package for Kepler & TESS time series analysis

I am a core developer of the lightkurve package. The goal of lightkurve is to lower the bar for astronomical time series analysis by providing clear and easy to use tools to the astronomy community. Lightkurve is a great software package for introducing undergraduate and graduate students to the basic principles of exoplanet or binary system detection and characterization, stellar variability, or transient detection in space telescope photometry.

Adorable image created by Christina Hedges.

eleanor

Simple light curve extraction from TESS FFIs

I am also a member of the core development team for the eleanorTESS pipeline. This pipeline provides systematics-corrected light curves for hundreds of thousands of targets in the TESS FFIs, and enables users to produce custom systematics-corrected light curves for any target with a given TIC ID, Gaia source ID, or RA and dec coordinates.

scope

Simulated CCD Observations for Photometric Experimentation

I am the lead developer for the scope simulation tool. This package was used to test the performance of the everest pipeline, particularly in the case of high motion. The results of the paper linked below influenced the decision to continue observation and upload larger aperture masks for K2 campaign 17.

everest

Pixel Level Decorrelation of K2 Light Curves

I contributed to the development of the everest K2 pipeline. This pipeline was developed by Luger et al. (2016, 2018) to remove systematic noise from K2 light curves. Using a method called Pixel Level Decorrelation (PLD) based on the work of Deming et al. (2015), we achieved Kepler-like precision in de-trended K2 light curves down to 15th magnitude.