Research

My research sits at the boundary between astrophysics and cosmology, as well as the boundary between observations and theory. Please do contact me at acliu at physics dot mcgill dot ca if you're interesting in being a part of this exciting journey!

Surveying our Universe with neutral hydrogen

The vast majority of our Universe's volume remains unexplored. The image on the left shows the extent of our observable Universe, drawn to scale with us at the center of the circle looking radially outwards. Shown in color are the portions that have been systematically surveyed, typically using the Cosmic Microwave Background (greenish outer shell), galaxy surveys (white cone-shaped regions), and quasar surveys (turquoise dots). However, most of our Universe remains unexplored. In 21cm cosmology, we aim to change this by mapping out the distribution of hydrogen in our cosmos. The potential reach of this technique is vast---mouseover the image to see in orange how much of the observable Universe we can "see" by looking at neutral hydrogen.
Although such an ambitious program will take decades to complete using multiple instruments, there are plenty of questions for us to tackle now. For one, what does one do with such a vast map of our Universe? In a sense, our Universe provides us with the ultimate data science problem. With datasets from 21cm cosmology projected to contain a thousand times the information content of previous cosmological probes, how do we squeeze every bit of information out of our surveys? For instance, can we combine 21cm surveys with other cosmological probes to understand fundamental properties of our Universe like the masses of the neutrinos? (Spoiler: we can!)

Understanding the first stars and galaxies using HERA


A prime application of 21cm mapping is to use it to make direct observations of our Universe during Cosmic Dawn---the as-yet unexplored portion of our cosmic timeline when the first generation of stars and galaxies were formed. I am a founding member of the Hydrogen Epoch of Reionization Array (HERA) collaboration. HERA is a low-frequency radio interferometer being constructed in the South African Karoo desert.

With HERA, we aim to make direct observations of the intergalactic medium during cosmic dawn. This will enable us to understand key properties of the first galaxies. For instance, were first-generation galaxies similar to the galaxies that we see today, in terms of their masses and luminosities in UV and X-rays? Or were they substantially different? How did these galaxies form? And when they formed, how did they affect their surroundings? These questions all feed into the larger goal of understanding galaxy formation. I work towards this goal by establishing the connections between observations, data analysis, and theory within HERA.

A data-driven approach to understanding the sky

Perhaps the most fundamental question of any survey of the sky is "What does the sky look like in all directions at all wavelengths? " Unfortunately, empirical data regarding this question is substantially incomplete in radio wavelengths. One approach to this is to fill in any gaps in the data using theoretical models. I take a complementary data-driven approach. Using Bayesian statistics and machine learning techniques, I ask the question "What is our best-guess map of the sky given all available datasets? "

To answer this question, I currently lead the development of an extended Global Sky Model (eGSM). The eGSM unlocks the rich science opportunities that come from having accurate maps of the sky. For instance, I am eager to use these maps to forecast the level of contamination inherent in extragalactic radio observations, or to explore star formation processes in our Milky Way galaxy, to list a couple of examples.