As a graduate student at the Harvard-Smithsonian Center for Astrophysics, I worked with Drs. David Charbonneau and Jennifer Winters to study low-mass M dwarfs, the smallest (and most common) type of star. Broadly, my research explores how these stars differ from our Sun and what those differences mean for the planets that orbit them.
Stars rotate rapidly in their youth, spinning down to longer rotation periods over time as they shed angular momentum. For stars like the Sun, this spindown is gradual: the Sun has a rotation period of about 27 days after billions of years of angular momentum loss. Mid-to-late M dwarfs behave differently. Many of them rotate very quickly, with rotation periods shorter than 2 days. Others rotate very slowly, with periods longer than 100 days. There are very few with intermediate periods, implying that the transition between these modes must occur quickly, not gradually like the spindown of the Sun.
It is hard to measure ages for M dwarfs, and so it is unclear whether all mid-to-late M dwarfs make this transition at a similar age, or if it differs from star to star. The distinction is important for planetary habitability because stellar rotation correlates with stellar activity. If all mid-to-late M dwarfs are active for billions of years, their planets have been bombarded by high-energy environments for a very long time, making it difficult for those planets to retain atmospheres.
In this work, I explored this question by studying M dwarfs in wide binaries. Stars in binaries formed together and therefore share the same age. I found that rapidly rotating M dwarfs spin down gradually within the rapidly rotating mode, before abruptly making the transition to the slowly rotating mode at ages of a few billion years. I also identified a handful of stars that had made the transition to the slowly rotating mode at ages younger than a billion years. Some mid-to-late M dwarfs therefore become inactive much earlier than others -- which may bode well for the atmospheres of their planets.
While working on the previous project, I discovered a strange system. G 68-34 appeared to be an early-type M dwarf in a wide binary with a white dwarf. By studying the white dwarf, I inferred that G 68-34 was at least five billion years old. However, it was very rapidly rotating, with a rotation period of less than a day. Based on its unusual properties, I thought that G 68-34 was likely a close binary, with interacting components maintaining its rapid rotation even at its advanced age.
To confirm this hypothesis, I gathered high-resolution spectra of the star using the Tillinghast Reflector Echelle Spectrograph (TRES). With these data, I was able to observe two sets of spectral lines, indicating that the star was indeed a binary. After analyzing new photometric observations from the Transiting Exoplanet Survey Satellite (TESS), I found that the pair also eclipse. Eclipsing binaries are exciting systems because they allow the masses and radii of both stars to be precisely measured, providing an observational test of theoretical stellar evolution models.
Ultimately, I found that G 68-34 was a roughly equal-mass binary consisting of two fully convective M dwarfs. The system is spin-orbit coupled, meaning that the rotation period of the stars is the same as their orbital period due to interactions between the stars. As I had originally suspected, close binarity has allowed activity and rapid rotation to persist to older ages than would otherwise be possible.
Jupiter had a big dynamical influence on the evolution of the inner solar system, shaping the Earth into the world we know and love. If Jupiter didn't exist, the Earth may have been very different with respect to its water content, size, atmosphere, and even habitability. In exosolar systems around Sun-like stars, planets like Jupiter are reasonably common -- by "planets like Jupiter", I mean cold giant worlds beyond their star's snow line, where water exists as a solid and can be incorporated into the formation of planets. However, giant planets are predicted to be rare around M dwarfs, and especially around low-mass M dwarfs. If this is the case, their terrestrial worlds (which are common) will have evolved in an entirely different dynamical environment to the Earth.
In this work, I studied a volume-complete sample of mid-to-late (or specifically, 10-30% the mass of the Sun) M dwarfs within 15pc. Our team (led by Jen Winters) has been surveying these stars with high-resolution spectrographs since 2016. We have gathered four spectra of each star, and I used these to search for doppler shifts indicative of the presence of a planet. After omitting active stars and close binaries, which are more likely to generate false positive detections, my sample totalled 200 stars. Ultimately, I did not detect any giant planets, despite a strong sensitivity to Jupiter-mass planets beyond the snow line.
Our result supports the canonical picture of giant planet formation, which predicts that such planets are rare around low-mass M dwarfs. This means that the terrestrial planets of these tiny stars -- notably, planets that are top-priority targets for atmospheric characterization with the James Webb Space Telescope -- evolved in a very different dynamical environment to our Earth, with all sorts of implications for their evolution, composition, and habitability. For more details, check out the press release, press coverage, and astrobite.
While I neglected active stars from the previous project, we nonetheless gathered observations of them. Of the low-mass M dwarfs in the volume-complete sample without close binaries, 123 are active (or equivalently, 38%) -- this number is so high because low-mass M dwarfs remain active for billions of years, making these active stars a significant stellar demographic. Moreover, M dwarfs that are inactive today were once these active stars, and their planets formed and evolved during this lengthy phase of activity. Understanding M-dwarf activity is therefore important for understanding the evolution of their planets.
For this project, I analyzed our multi-epoch spectroscopic data alongside complementary photometry from TESS and the ground-based MEarth array. With both spectroscopy and photometry, I was able to measure activity levels, rotation periods, rotational broadening, inclinations, and radial velocity variability, allowing me to investigate the properties of active M dwarfs at the population level.
I found that three quarters of active, low-mass M dwarfs rotated with periods shorter than two days. Stars that rotated with slightly longer rotation periods tended to have more modest activity levels, suggesting that activity tempers over time as these stars gradually spin down within the rapidly rotating mode. Interestingly, stars within the gap between the rapidly and slowly rotating modes had enhanced activity levels. This may reflect a changing magnetic field structure as these stars quickly shed angular momentum and abruptly spin down.
Blurb to come. In the meantime, read the press release!
The Kuiper belt -- an area of the solar system past Neptune, stretching from 30 to 50 AU -- is densely populated with small objects. There are an estimated tens of billions to trillions of objects in this belt that are larger than a kilometre, but only around a thousand have been discovered. All of these detected Kuiper belt objects (KBOs) are larger than 15km, big icy bodies like the dwarf planet Pluto. Most objects in the Kuiper belt are far smaller, but kilometre-sized bodies are too small to be seen directly by telescopes.
If you are capable of imaging very dim objects at a very fast rate, these small KBOs can be detected by looking for the diffraction effects caused when they pass in front of distant stars. Fresnel diffraction causes a characteristic pattern in the light curve of the star that varies depending on KBO size, observing wavelength, the star's angular diameter, and so on. These events last for only a fraction of a second, but modern electron multiplying charge-coupled devices (EMCCDs) are capable of imaging these occultations.
At Western University, I developed a detection pipeline for Colibri, a new telescope array dedicated to the search for KBOs via this occultation method. With over 6TB of data imaged per night, the pipeline must employ autonomous real-time detection, erasing most collected data and only retaining information on candidate occultation events. The pipeline is primarily written in Python, incorporating functionality from Astropy, Source Extractor, Joblib, and Astrometry.net.
My work on the Colibri pipeline is published in the Publications of the Astronomical Society of the Pacific. This project was supervised by Stan Metchev at Western University.
Due to the finite speed of light, looking at distant objects means looking back in time. Look far enough away and you can find protoclusters: the precursors to the galaxy clusters we see today.
As a byproduct of a South Pole Telescope survey to study the fine structure of the cosmic microwave background radiation, a number of dusty, star-forming galaxies were identified at high redshift. Using data from ALMA, APEX, Spitzer, and Herschel, I analyzed the regions surrounding these galaxies, searching for other sources at similar redshift, quantifying their properties, and determining whether the regions were truly protoclusters. I also designed spectroscopic masks for follow-up observations with Gemini and ran simulations to predict the results of future surveys.
I am a co-author of a paper published in Nature, with further results from this work published in MNRAS and ApJ. This project was supervised by Scott Chapman at Dalhousie University.
Many exoplanets have been detected by the transit method: when these planets pass in front of their star, they create a detectable dip in the star's light. For large, hot exoplanets (hot Jupiters) we can also detect a dip during secondary eclipse, when the planet passes behind the star. Measurements of the secondary eclipse at infrared wavelengths provide information on the planet's thermal emission, which can be converted into measurements of brightness temperature. My project involved using Gaussian process regression, a machine learning technique, to estimate the overall effective temperatures of hot Jupiters from sparse brightness temperature data.
A paper presenting our results has been published by the Monthly Notices of the Royal Astronomical Society. This project was supervised by Nick Cowan at McGill University.
The Cosmological Advanced Survey Telescope for Optical and ultraviolet Research (CASTOR) is a proposed Canada-led space telescope that would perform wide-field, high-resolution imaging in three ultraviolet and blue-optical bands. For my fourth-year research project, I simulated CASTOR observations of the COSMOS field, estimating the telescope's limiting magnitudes and sensitivity to galaxies transitioning between the blue and red sequences as a result of quenching.
I am a co-author of the science report on CASTOR presented to the Canadian Space Agency (internal document), with an abridged version published as a Canadian Long Range Plan for Astronomy and Astrophysics White Paper (LRP2020). This project was supervised by Michael Balogh at the University of Waterloo.