Over the past two years, astronomers have rewritten the story of our galaxy.
When the Khoisan hunter-gatherers of sub-Saharan Africa gazed upon the meandering trail of stars and dust that split the night sky, they saw the embers of a campfire. Polynesian sailors perceived a cloud-eating shark. The ancient Greeks saw a stream of milk, gala, which would eventually give rise to the modern term “galaxy.”
In the 20th century, astronomers discovered that our silver river is just one piece of a vast island of stars, and they penned their own galactic origin story. In the simplest telling, it held that our Milky Way galaxy came together nearly 14 billion years ago when enormous clouds of gas and dust coalesced under the force of gravity. Over time, two structures emerged: first, a vast spherical “halo,” and later, a dense, bright disk. Billions of years after that, our own solar system spun into being inside this disk, so that when we look out at night, we see spilt milk — an edge-on view of the disk splashed across the sky.
Yet over the past two years, researchers have rewritten nearly every major chapter of the galaxy’s history. What happened? They got better data.
On April 25, 2018, a European spacecraft by the name of Gaia released a staggering quantity of information about the sky. Critically, Gaia’s years-long data set described the detailed motions of roughly 1 billion stars. Previous surveys had mapped the movement of just thousands. The data brought a previously static swath of the galaxy to life. “Gaia started a new revolution,” said Federico Sestito, an astronomer at the Strasbourg Astronomical Observatory in France.
Astronomers raced to download the dynamic star map, and a flurry of discoveries followed. They found that parts of the disk, for example, appeared impossibly ancient. They also found evidence of epic collisions that shaped the Milky Way’s violent youth, as well as new signs that the galaxy continues to churn in an unexpected way.
Taken together, these results have spun a new story about our galaxy’s turbulent past and its ever-evolving future. “Our picture of the Milky Way has changed so quickly,” said Michael Petersen, an astronomer at the University of Edinburgh. “The theme is that the Milky Way is not a static object. Things are changing rapidly everywhere.”
To peer back to the galaxy’s earliest days, astronomers seek stars that were around back then. These stars were fashioned only from hydrogen and helium, the cosmos’s rawest materials. Fortunately, the smaller stars from this early stock are also slow to burn, so many are still shining.
After decades of surveys, researchers had assembled a catalog of 42 such ancients, known as ultra metal-poor stars (to astronomers, any atom bulkier than helium qualifies as metallic). According to the standard story of the Milky Way, these stars should be swarming throughout the halo, the first part of the galaxy to form. By contrast, stars in the disk — which was thought to have taken perhaps an additional billion years to spin itself flat — should be contaminated with heaver elements such as carbon and oxygen.
In late 2017, Sestito set out to study how this metal-poor swarm moves by writing code to analyze the upcoming Gaia results. Perhaps their spherical paths could offer some clues as to how the halo came to be, he thought.
In the days following Gaia’s data release, he extracted the 42 ancient stars from the full data set, then tracked their motions. He found that most were streaming through the halo, as predicted. But some — roughly 1 in 4 — weren’t. Rather, they appeared stuck in the disk, the Milky Way’s youngest region. “What the hell,” Sestito wondered, though he used a different four-letter term. “What’s going on?”
Follow-up research confirmed that the stars really are long-term residents of the disk, and not just tourists passing through. From two recent surveys, Sestito and colleagues amassed a library of roughly 5,000 metal-poor stars. A few hundred of them appear to be permanent denizens of the disk. Another group sifted through about 500 stars identified by another survey, finding that about 1 in 10 of these stars lie flat in circular, sunlike orbits. And a third research group found stars of various metallicities (and therefore various ages) moving in flat disk orbits. “This is something completely new,” said lead author Paola Di Matteo, an astronomer at the Paris Observatory.
How did these anachronisms get there? Sestito speculated that perhaps pockets of pristine gas managed to dodge all the metals expelled from supernovas for eons, then collapsed to form stars that looked deceptively old. Or the disk may have started taking shape when the halo did, nearly 1 billion years ahead of schedule.
To see which was more probable, he connected with Tobias Buck, a researcher at the Leibniz Institute for Astrophysics in Potsdam, Germany, who specializes in crafting digital galaxy simulations. Past efforts had generally produced halos first and disks second, as expected. But these were relatively low-resolution efforts.
Buck increased the crispness of his simulations by about a factor of 10. At that resolution, each run demanded intensive computational resources. Even though he had access to Germany’s Leibniz Supercomputing Center, a single simulation required three months of computing time. He repeated the exercise six times.
Of those six, five produced Milky Way doppelgängers. Two of those featured substantial numbers of metal-poor disk stars.
How did those ancient stars get into the disk? Simply put, they were stellar immigrants. Some of them were born in clouds that predated the Milky Way. Then the clouds just happened to deposit some of their stars into orbits that would eventually form part of the galactic disk. Other stars came from small “dwarf” galaxies that slammed into the Milky Way and aligned with an emerging disk.
The results, which the group published in November, suggest that the classic galaxy formation models were incomplete. Gas clouds do collapse into spherical halos, as expected. But stars arriving at just the right angles can kick-start a disk at the same time. “[Theorists] weren’t wrong,” Buck said. “They were missing part of the picture.”
The complications don’t end there. With Gaia, astronomers have found direct evidence of cataclysmic collisions. Astronomers assumed that the Milky Way had a hectic youth, but Helmer Koppelman, an astronomer now at the Institute for Advanced Study in Princeton, New Jersey, used the Gaia data to help pinpoint specific debris from one of the largest mergers.
Gaia’s 2018 data release fell on a Wednesday, and the mad rush to download the catalog froze its website, Koppelman recalled. He processed the data on Thursday, and by Friday he knew he was on to something big. In every direction, he saw a huge number of halo stars ping-ponging back and forth in the center of the Milky Way in the same peculiar way — a clue that they had come from a single dwarf galaxy. Koppelman and his colleagues had a brief paper ready by Sunday and followed it up with a more detailed analysis that June.
The galactic wreckage was everywhere. Perhaps half of all the stars in the inner 60,000 light-years of the halo (which extends hundreds of thousands of light-years in every direction) came from this lone collision, which may have boosted the young Milky Way’s mass by as much as 10%. “This is a game changer for me,” Koppelman said. “I expected many different smaller objects.”
The group named the incoming galaxy Gaia-Enceladus, after the Greek goddess Gaia — one of the primordial deities — and her Titan son Enceladus. Another team at the University of Cambridge independently discovered the galaxy around the same time, dubbing it the Sausage for its appearance in certain orbital charts.
When the Milky Way and Gaia-Enceladus collided, perhaps 10 billion years ago, the Milky Way’s delicate disk may have suffered widespread damage. Astronomers debate why our galactic disk seems to have two parts: a thin disk, and a thicker one where stars bungee up and down while orbiting the galactic center. Research led by Di Matteo now suggests that Gaia-Enceladus exploded much of the disk, puffing it up during the collision. “The first ancient disk formed pretty fast, and then we think Gaia-Enceladus kind of destroyed it,” Koppelman said.
Hints of additional mergers have been spotted in bundles of stars known as globular clusters. Diederik Kruijssen, an astronomer at Heidelberg University in Germany, used galaxy simulations to train a neural network to scrutinize globular clusters. He had it study their ages, makeup, and orbits. From that data, the neural network could reconstruct the collisions that assembled the galaxies. Then he set it loose on data from the real Milky Way. The program reconstructed known events such as Gaia-Enceladus, as well as an older, more significant merger that the group has dubbed Kraken.
In August, Kruijssen’s group published a merger lineage of the Milky Way and the dwarf galaxies that formed it. They also predicted the existence of 10 additional past collisions that they’re hoping will be confirmed with independent observations. “We haven’t found the other 10 yet,” Kruijssen said, “but we will.”
All these mergers have led some astronomers to suggest that the halo may be made almost exclusively of immigrant stars. Models from the 1960s and ’70s predicted that most Milky Way halo stars should have formed in place. But as more and more stars have been identified as galactic interlopers, astronomers may not need to assume that many, if any, stars are natives, said Di Matteo.
The Milky Way has enjoyed a relatively quiet history in recent eons, but newcomers continue to stream in. Stargazers in the Southern Hemisphere can spot with the naked eye a pair of dwarf galaxies called the Large and Small Magellanic Clouds. Astronomers long believed the pair to be our steadfast orbiting companions, like moons of the Milky Way.
Then a series of Hubble Space Telescope observations between 2006 and 2013 found that they were more like incoming meteorites. Nitya Kallivayalil, an astronomer at the University of Virginia, clocked the clouds as coming in hot at about 330 kilometers per second — nearly twice as fast as had been predicted.
The Large and Small Magellanic Clouds rise over Mount Bromo, an active volcano in Bromo Semeru Tengger National Park in Java, Indonesia.
When a team led by Jorge Peñarrubia, an astronomer at the Royal Observatory of Edinburgh, crunched the numbers a few years later, they concluded that the speedy clouds must be extremely hefty — perhaps 10 times bulkier than previously thought.
“It’s been surprise after surprise,” Peñarrubia said.
Various groups have predicted that the unexpectedly beefy dwarfs might be dragging parts of the Milky Way around, and this year Peñarrubia teamed up with Petersen to find proof.
The problem with looking for galaxy-wide motion is that the Milky Way is a raging blizzard of stars, with astronomers looking outward from one of the snowflakes. So Peñarrubia and Petersen spent most of lockdown figuring out how to neutralize the motions of the Earth and the sun, and how to average out the motion of halo stars so that the halo’s outer fringe could serve as a stationary backdrop.
When they calibrated the data in this way, they found that the Earth, the sun, and the rest of the disk in which they sit are lurching in one direction — not toward the Large Magellanic Cloud’s current position, but toward its position around a billion years ago (the galaxy is a lumbering beast with slow reflexes, Petersen explained). They recently detailed their findings in Nature Astronomy.
The sliding of the disk against the halo undermines a fundamental assumption: that the Milky Way is an object in balance. It may spin and slip through space, but most astronomers assumed that after billions of years, the mature disk and the halo had settled into a stable configuration.