Science lab spotlight

Simulating galactic formation

MIT’s Caterpillar Project models the formation of galaxies like our own Milky Way

9448 caterpillar1
The Caterpillar Project visualizes galaxies through beautiful simulations.
Courtesy of Kaley Brauer
9449 caterpillar3
The Caterpillar Project's galactic simulations contribute to our understanding of the galaxy and its formation.
Courtesy of Kaley Brauer

For as much as astrophysicists know about the cosmos, the formation of galaxies — even the formation of our own galaxy, the Milky Way — is still an open question. This is the question that the Caterpillar Project, led in part by MIT physics faculty Anna Frebel and Mark Vogelsberger, hopes to answer. The Caterpillar Project is a suite of simulations of galaxies which, given a set of initial conditions from astronomical data (like the cosmic microwave background radiation), can give a statistical probe into the formation of thirty-five galaxies like the Milky Way.

One of the easiest ways to understand how galaxies formed is through “galactic archaeology”: looking at the oldest, faintest galaxies in the universe. These are known as ultra-faint dwarf galaxies (UFD). “It’s similar to traditional archaeology that uses fossils to study previous societies and learn how they evolved,” says PhD student Kaley Brauer, who focuses on the Caterpillar Project’s simulations of stellar halos. “We look at old stars with different kinematic and chemical signatures and by studying them, see if we can determine how they got there and how our galaxy became as it is today.” 

UFDs are the fossils of the universe, inactive since the reionization of the universe billions of years ago when stars released their energy, ionizing hydrogen in the universe. Because of this, they allow researchers to look into the past; the UFDs which survive today have looked the same for most of the universe’s history. 

Astrophysicists have learned that larger galaxies like our own are formed from the merging of smaller galaxies like UFDs. The way this happens is by “hierarchical formation,” where gravitational interactions cause smaller galaxies to collapse into each other and merge into larger ones. This fact implies that, to understand the formation of larger galaxies, the first step is to understand the formation of the smaller ones.

The Caterpillar Project is made to simulate the formation of a large number of Milky Way-like galaxies at a high resolution from a statistical standpoint. The Caterpillar Project aims to understand galaxy formations by using dark-matter-only simulations. That is, the simulations ignore interactions between baryonic matter — only around ten percent of the universe which consists of matter like quarks, electrons, protons, neutrons, etc. — and only focuses on the effects due to dark matter. 

To include baryonic physics would be quite computationally difficult, but since baryonic matter makes up so little of the universe relative to the dark matter content, baryonic interactions can be ignored to a very good approximation.

Historically, dark-matter simulations, like the Millennium Simulation Project out of the Max Planck Institute for Astrophysics, have shown the power of dark-matter-only simulations, demonstrating that this simple approximation — excluding baryonic physics — opens a lot of computational doors for researchers. 

Other galaxy simulations, like the FIRE and Latte simulations (from a collaboration with Northwestern, Caltech, UC San Diego, and UC Berkeley) or the Illustris simulation (created by MIT’s own Mark Vogelsberger), have been able to include both baryonic interactions and dark matter interactions, but these simulations typically have a lower resolution compared to dark-matter-only simulations like the Caterpillar Project. 

The high-resolution, high-statistics approach from the simulation side is advantageous when it comes to comparing with actual data. “It just so happens that in the last few years, and in the few years coming up, astronomers are collecting an absolutely astronomical amount of data on stars in our galaxy,” Brauer says. “Their kinematics, their spectra, and so now this is a big data/statistical problem.” 

The Caterpillar Project’s statistically-driven simulations are well-suited for comparing to this kind of data. Brauer hopes to reach new insights through these comparisons, either affirming or contradictory: “It will be useful to see both similarities that we find between the simulated halos and the Milky Way, and if any differences arise, why they appear to be different.”

However, the assumption that there is no baryonic matter in the universe is simply wrong. Researchers have to patch up the simplifications using various correction methods; for example, one can relate the number of stars to the amount of dark matter in the area. But with only a few corrections, the predictions made by the Caterpillar Project have proved very useful. For instance, by studying the stellar halo of galaxies simulated in the Caterpillar Project — stars and other matter that accrete on the outskirts of a galaxy — one can directly compare detailed information like chemical compositions, stellar spectra, and the kinematics of the simulated halos to actual data. 

The group has found recent success in simulating the rapid neutron capture process (R-process), the process by which heavy elements are formed. When they produced stars in the early period of the universe’s existence, UFDs likely produced stars with heavy elements like gold or platinum, which use the R-process. This has the potential to provide a chemical signature tag for galaxies which use this process, which could give a recipe for the formation of the Milky Way.

The Caterpillar Project’s biggest upcoming challenge is tracking the step-by-step formation of stars and supernovae to track exactly where the elements go to form the galaxies. This is called “galactic chemical evolution,” and the group hopes to finish this within the next year or so. Of course, the biggest step in improving the simulations is to finally include baryonic physics to fully simulate how our universe works. This requires not only computational power, but also using clever approximations to make these simulations feasible.

The members of the Caterpillar Project feel confident in its potential. “Among the people in the world who could do this work, we may be the ones that are the most prepared to do it,” says Brauer, who thinks that the group is close to “being able to say how many small galaxies merged to form the Milky Way with any level of confidence.” All eyes are on the Caterpillar Project as a leader in the world of galactic simulations.