Playing hide-and-seek with colliding dying galaxies

Title: The merger fraction of post-starburst galaxies in UNIONS

Authors: Scott Wilkinson, Sara L. Ellison, Connor Bottrell, Robert W. Bickley, Stephen Gwyn, Jean-Charles Cuillandre, Vivienne Wild

Institution: University of Victoria, Victoria, British Columbia, Canada

Status: Open access on ArXiv.


At first glance, we can split most galaxies into two main types: blue disks with spiral arms and red spheroidal structures – star-forming and quiescent, alive and dead. How and why galaxies die, however, remains one of the biggest questions in astronomy. The best way to understand something’s cause of death, of course, is to watch it die. 

Post-starburst galaxies are the perfect window into this grim part of a galaxy’s life. Post-starburst galaxies (or PSBs, for short) are galaxies that quenched, or stopped forming stars, less than 1 billion years ago following an intense burst of star formation. They have an intermediate structure between disks and spheroids, and are not yet completely red-and-dead. All of this makes PSBs the perfect test subjects for understanding how galaxies die!

So what is killing the PSBs?

The short answer is that we still don’t know. A compelling candidate is a past collision, or merger, with another galaxy – like the ongoing one you can see in Figure 1. A merger can jump-start a burst of star formation, disrupt the disk, and fuel the central supermassive black hole. When two galaxies smash into one another, the remnant also has a spectacularly weird structure, or morphology. This unique structure lets us easily find mergers by looking for clearly disturbed galaxies. Simple, right?

Figure 1. Merging galaxies in the Stephan’s Quintet, as seen by JWST. Mergers can trigger a burst of star formation, and are a compelling formation mechanism for post-starbursts. Also, can you tell I’m excited about the new JWST observations? Image credit: NASA, ESA, CSA, and STScI

Not quite. The post-merger disturbances fade quickly (in a couple of hundreds million years), are difficult to detect in distant galaxies, and depend a lot on how you define “disturbed”. While some PSBs definitely experienced a merger in the past, how many  is still debated – anywhere from 15% to 70%.

Today’s authors delve deeply into the morphology of PSBs, setting out to constrain how many PSBs really are post-mergers, and how many we may have missed. So while the rest of the world marvels at the new images from the James Webb Space Telescope, let’s enjoy the fantastic galaxies observed by the Canada France Hawaii Telescope on Maunakea!

Where’s Waldo the collision?

It is easy to tell if a nearby galaxy recently experienced a merger by its unique tidal tails – you can spot them on the top galaxy in Figure 1. This is more challenging for more distant galaxies, both because they are farther away, making post-merger features difficult to see, and because there are too many to look at by eye. Thus, we turn to automatic ways to detect mergers.

One traditional approach is to measure the asymmetry of a galaxy. Highly asymmetric galaxies are often merger remnants. Asymmetry is easy to measure – you simply flip an image and look at the differences between the two, as shown schematically in Figure 2.

Figure 2. A schematic way to calculate the asymmetry of a galaxy, such as NGC7318 from Stephan’s Quintet. Adapted from Figure 1. 

There are many other ways to measure how weird a galaxy is, and in this work, the authors set out to determine which one is the best. They looked at how well different metrics perform at finding mergers. Some of these metrics are: the asymmetry, shape asymmetry (similar to asymmetry, but calculated only for the shape of the galaxy rather than its light distribution), merger probability given by a neural network, and a visual classification into a merger/non-merger.

One of the difficulties of using an automated metric to find mergers is the threshold above which a galaxy is disturbed enough to be considered a merger. The authors found that using typical thresholds, mergers among PSBs are either as common as among other galaxies (using asymmetry), or 10–15 times more common (using visual and neural network classification). So it really matters which metric you use!

However, when you change the threshold, the picture is different – as you can see on Figure 3. If you set the threshold higher, the number of mergers in PSBs compared to normal galaxies increases significantly. For example, using a high merger probability threshold, PSBs are 100 times more likely to be post-mergers than normal galaxies!

Figure 3. How the fractional excess of galaxy mergers among post-starbursts changes with the threshold chosen to classify an object as a merger for three metrics: asymmetry (left), shape asymmetry (center), and merger probability (right). The common threshold for each parameter is shown as a yellow line. Both shape asymmetry and merger probability show that mergers are much more common among post-starbursts. Adapted from today’s paper – Figures 6, 7, and 8 in Wilkinson et al. (2022).

Choose your telescope wisely

The authors then turned to the Sloan Digital Sky Survey (SDSS) to see whether the image quality makes a difference when looking for mergers. The authors found that almost half of the disturbed PSBs have “regular” shapes as far as SDSS is concerned. So if you want to look for faint tidal tails in galaxies that merged millions of years ago, you better check the deep-sky observations!

Even more remarkably, when the authors compared their results to simulations of merging galaxies, they found that asymmetry and shape asymmetry measurements can only identify 7% and 30% of all post-merger galaxies, respectively. The vast majority of galaxy mergers remain undetected in ground-based imaging! 

So, do PSBs die after a collision?

This question remains hard to answer. Some – 30%, give or take – definitely are. But as the authors of today’s paper showed, finding these post-merger galaxies is challenging, requiring deep-sky imaging and carefully-tailored analysis tools. Many – up to 70%, as simulations suggest – mergers cannot be detected with our current tools, so it is possible that all PSBs had a recent merger.

Deeper, high-resolution imaging is going to be crucial to answer this question, so we better get back to tweets about the new JWST data!

Edited by: Catherine Clark

Featured image credits: The Canada-France-Hawaii telescope atop Mauna Kea by Gordon W. Myers, Detective vector art by publicdomainvectors.org, Artist’s impression of a post-starburst galaxy (ALMA/ESO/NAOJ/NRAO)/S. Dagnello (NRAO/AUI/NSF)), SMACS 0723 Deep Field by The James Webb Space Telescope (NASA, ESA, CSA, and STScI)

About Liza Sazonova

I study how galaxies quench and die, together with the amazing SPOGs team. I look at the structure of galaxies as they are quenching, and see what processes transform them from a disk to a blob. When not coding or looking at galaxies, I like playing board games, climbing rocks, and writing a short novel about Phil the Photon!

Leave a Reply