Research
Millisecond pulsarsPulsar timingNanohertz gravitational wavesDynamics
I work with millisecond pulsars, which are pulsars that spin at up to 716 times per second. In particular, I study what we can learn about when their pulses arrive on earth, known as "times of arrival."
Counting every rotation
Pulsar timing accounts for every rotation of a pulsar between observations. By doing so, we can learn about how fast the pulsar is rotating, its position in the sky, the nature of how it orbits another star, and even properties of the Milky Way galaxy. Getting the count right normally takes tightly scheduled monitoring and a good deal of practice. I wrote an algorithm that does the phase connection for binary pulsars on its own, using the statistical F-test and the quadratic behaviour of the reduced χ² near a minimum.
It produced timing solutions for two newly discovered binary pulsars in the globular cluster Terzan 5, from about 70 Green Bank Telescope observations spanning 13 years. A run can take hours, but as we wrote in the paper:
This may seem like a long runtime, but a human would likely take much longer and give up from frustration after many incorrect attempts.
Algorithmic Pulsar Timer for Binaries, 2024
Trojans around pulsars
Jupiter shares its orbit with thousands of asteroids called Trojans, which gather 60 degrees ahead of and behind the planet. Trojans of up to planetary mass have been predicted around other stars, but none have been confirmed. I searched for them around radio pulsars with very low-mass companions, where the times of arrival are measured well enough to pick up the wobble a Trojan would give the pulsar. Eight pulsars from the NANOGrav 15 yr dataset gave upper limits near one Earth mass. Two of them show weak evidence for signals consistent with Trojans of a few Jupiter masses, but those signals most likely have some other source.
The paper offers a few possible reasons for why we didn't find any. This is the one I like best:
Mass transfer and disk dynamics in pulsar binary systems are clearly turbulent, and so, any interloping Trojan may find itself ejected from the system or torn apart.
Searching for Exotrojans in Pulsar Systems, 2026
Free-floating objects
Rogue planets, brown dwarfs and large asteroids that don't orbit any star are expected throughout the Milky Way, and three interstellar objects have already been caught passing through the solar system on hyperbolic orbits. If one passes close enough to a millisecond pulsar, its gravity shifts the times of arrival by a measurable amount. We searched the NANOGrav 15 yr narrowband dataset, 68 pulsars in all, for those encounters. We found no statistically significant events, which lets us set upper limits on how many of these objects there can be nearby. These are the first such limits from pulsar timing data.
Search for Gravitational Scattering of Pulsars by Free-floating Objects
Finding new pulsars
Some of my work is finding new pulsars. A targeted search with the Green Bank Telescope for radio pulsations from a redback candidate turned up the 1.68 ms binary pulsarPSR J2053+6905, and multi-epoch optical photometry and spectroscopy of the same system constrain the two component masses. The CHIME fast radio burst instrument found PSR J2108+5055, a sporadic mode-changing source whose brightest pulse reached a flux density of 3.1 kJy. MeerKAT observations of Terzan 5 produced ten more millisecond pulsars, bringing the count for that cluster to 49.
Redback pulsar PSR J2053+6905An extremely luminous sporadic radio pulsarTen new millisecond pulsars in Terzan 5