22.02.2022
A new simulation aims to determine whether the standard view of dark matter can explain how unique our galaxy’s neighborhood is.
IN THE 1980S and ’90s, when Carlos Frenk worked on some of the first theories of cold dark matter—“cold” refers to the invisible particles’ relatively slow speed—he thought the idea wouldn’t last very long. He and his colleagues had already tested a theory of faster-moving “hot” dark matter, the possibility that it is made of particles like neutrinos, and quickly ruled it out. Instead, the theory of cold dark matter became astrophysicists’ “standard model” for two decades, a mantle it still carries.
Now Frenk’s trying to poke holes in his cold dark-matter theory again. With a new simulation, he hopes to tackle open questions which may or may not be answered in the theory’s favor. “That's how science works. One of my ambitions today is to shoot down the theory I’ve worked on,” says Frenk, an astrophysicist at Durham University in the United Kingdom.
Frenk and his colleagues at Durham and in Helsinki, Finland, just completed the first part of a computer simulation of the dark-matter universe; it’s dubbed the Simulations Beyond the Local Universe project, or SIBELIUS, after the Finnish composer. The project was led by Stuart McAlpine and Till Sawala, both of whom previously conducted research with Frenk at Durham. Theirs isn’t just any dark-matter simulation, but one with galaxies modeled in it, providing a detailed, three-dimensional picture of what our galaxy and our corner of the universe likely looks like—if the standard view of cold dark matter is right. They published their new research this month.
“This is the first attempt to simulate our patch of the universe, with all the structures we know and love, including the Coma cluster and the Virgo cluster,” Frenk says, referring to large conglomerations of galaxies. Those kinds of cosmic landmarks, which lie tens of millions of light years from Earth or even farther away, might matter for understanding the assembly and evolution of our own galaxy over billions of years. They might also affect physicists’ perspective on how fast the universe is expanding. Frenk and his team hope their simulation will be a useful tool for addressing such weighty questions. And if it can’t answer them, it could mean that current dark-matter theories have problems.
Past efforts by theorists, including by Frenk himself, have either simulated a huge piece of the universe that only resembles the real one in a statistical sense, getting the number of galaxies and galaxy clusters about right, or they’ve zoomed in and focused only on our own Milky Way. But there’s plenty to learn from our galaxy’s surroundings too. Astronomers have thoroughly mapped out our local region, spotting dozens of small and faint “satellite” galaxies, like the Large Magellanic Cloud, which orbit the Milky Way similar to the way the moon orbits the Earth. For decades, if not longer, they’ve also charted galaxy clusters and other objects beyond the neighborhood. (The French astronomer Charles Messier first discovered the Virgo cluster in 1781, in the constellation of the same name.)
SIBELIUS is more complex, because it builds on these impressive observations of our cosmic neighborhood and it actually tries to reproduce, to some extent, that local geography. The SIBELIUS simulation box is a big one, meant to resemble a 3D space that’s 3.3 billion light-years on a side. By design, in this virtual cosmos, we’re the center of the universe—the Milky Way resides in the middle, along with the neighboring Andromeda galaxy.
SIBELIUS is something called a “constrained realization,” meaning that simulations of these and other local galaxies must closely match what’s known about them in the real universe. By mapping them in a broader context, the team wants to see whether this region is representative of the entire universe, or rather atypical. Atypical might mean that there are many more—or fewer—galaxies in the surrounding environs than the expected average.
Most physicists believe that huge yet hidden webs of dark matter hold galactic structures together. In some spots in the SIBELIUS box, there’s a little more dark matter than in others. Here, dark matter starts clumping together, and then those clumps grow. Frenk and his colleagues model how galaxies build up and grow within those clumps, and then they compare what happens in this simulation to what’s known about the real world.
Mike Boylan-Kolchin, an astrophysicist at the University of Texas at Austin whose research involves simulations of dark matter and galaxies, likens the situation to someone counting present-day urban metropolises and then developing a more nuanced picture that includes their interconnected histories and the roads that link them. “It’s like, if you know the number of big cities in the US, that’s fine. But if you start to know where they are in relation to one another and their geography, then you can understand more about the history and how they formed,” he says. And in terms of our galaxy’s cosmic history, he says, we want to know how dark matter and other galaxies beyond the Milky Way’s borders shaped its past. “Does it matter that we have a certain distribution of galaxies around us? How rare are certain attributes of the Milky Way, and how much of that is related to the larger-scale environment?” he asks. “All of those questions I think you can only really answer with a simulation of the kind these people are producing.”
Astronomers have naturally focused their telescopes on the part of the universe closest to us, since those stars and galaxies can be examined in the greatest detail. But astrophysicists have sometimes struggled to square the population of our own galactic neighborhood with dark-matter theories. For example, earlier models predicted more neighboring galaxies than have actually been spotted in the real universe, an issue dubbed the “missing satellites” problem.
Big clumps of dark matter should have enough gravitational pull to bring in the gas that builds up into stars and, later, galaxies. But another problem is that some simulations end up producing big, orbiting dark matter clumps, which look like the ones that shouldhost satellite galaxies—but they don’t seem to have any real-universe counterparts. This is called the “too-big-to-fail” problem, since huge blobs of dark matter are thought to be too massive to fail to form galaxies within them.
A third challenge comes from the fact that the satellite galaxies swirling around the Milky Way and Andromeda seem to be orbiting in a plane, rather than spread out all around—something dark-matter physicists hadn’t predicted.
There are also cosmological problems that Frenk and his colleagues want to address. Astronomers using nearby supernova explosions and other local phenomena to measure how fast the universe is currently expanding get different answers than those probing the early universe. If dark-matter models are right, there has to be a way to resolve the troubling and persistent discrepancy between past and current observations.
But simulations like SIBELIUS might help. It might turn out that where a galaxy lives on the cosmic web of dark matter really does make a difference for measurements of the universe’s expansion rate. What if the Milky Way lies sort of in a “hole” in the web—if it’s more like a rural area between dark matter metropolises? If our part of the universe isn’t actually representative, then our local measurements of how fast the universe is blowing outward might be a little biased.
The Milky Way might happen to be situated in a fairly dense region of dark matter or in a sparse one, says Priyamvada Natarajan, a Yale University astrophysicist and dark-matter expert. “What is cool about this simulation is that they can address: How typical or unusual is our local volume? How rare is the distribution of matter that we see around us? Are we on a mountain or are we in a valley?” she says.
When comparing galaxies observed with telescopes to what’s seen in simulations, it’s necessary to compare apples to apples, says Jenny Sorce, an astrophysicist at the Institut d’Astrophysique Spatiale in Orsay, France, who helped design a similar kind of simulation, called CLONE, focused on galaxies in the Virgo cluster. “It’s not like you can compare one type of cluster with another one if they don’t share the same history or the same environment,” she says.
Frenk and his team did plenty of initial tests with their own computers at low resolution. But time on supercomputers, like on telescopes, is limited. They only had a single chance to run their full simulation, which took millions of hours of computing time on thousands of computer cores. But based on their simulation’s results, they find that the Milky Way’s neighborhood indeed seems atypical: We live in a cosmic region with fewer than average galaxies, but there are also more big galaxy clusters than on average. It’s like living in a low-elevation city, like Los Angeles, that nonetheless has mountain ranges in the distance.
If the Milky Way is indeed an oddball, it might help explain some dark matter mysteries, Frenk and Boylan-Kolchin speculate. If we're in a sparse part of the universe, that might explain why local measurements of the expansion rate are different than one would expect based on measurements of the faraway universe.
And if our galaxy is in the middle of an atypical neighborhood, that might explain why the satellites are in an unusual configuration—maybe they were pulled into the Milky Way's orbit in a particular way.
In other words, if the Milky Way’s neighborhood is indeed unusual, it means the cold dark-matter theory will survive these challenges—for now.
The jury’s still out. And there’s plenty of room for improvement with the SIBELIUS simulation. It would be an even better resource if their galaxy formation model incorporated fluid dynamics to follow the gas clouds that form new stars and make galaxies grow, Sorce says. That way, galaxies would emerge more naturally within dark matter clumps, which could prove helpful for investigating the more subtle dark-matter problems. Frenk and his team plan to do exactly that, though it will take much more supercomputer time.
In the meantime, Frenk will keep using these simulations to explore challenges to the still-favored cold dark-matter model. “If it’s wrong,” he says, “I want to be the one who proves it wrong.”
Quelle: WIRED