Dark matter—the unseen 80 percent of the universe’s mass—doesn’t emit, absorb or reflect light. Astronomers know it exists only because it interacts with our slice of the ordinary universe through gravity. Hence the hunt for this missing mass has focused on so-called WIMPs—Weakly Interacting Massive Particles—which interact with each other as infrequently as they interact with normal matter.

Quanta Magazine


Original story reprinted with permission from Quanta Magazine, an editorially independent division of the Simons Foundation whose mission is to enhance public understanding of science by covering research developments and trends in mathematics and the physical and life sciences

Physicists have reasons to look for alternatives to WIMPs. For two decades, astronomers have found less dark matter at the centers of galaxies than what WIMP models suggest they should. The discrepancy is even worse at the cores of the universe’s tiny dwarf galaxies, which have few ordinary stars but lots of dark matter.

About four years ago, James Bullock, a professor of physics and astronomy at the University of California, Irvine, began to wonder whether the standard view of dark matter was failing important empirical tests. “This was the point where I really started thinking hard about alternatives,” he said.

Bullock thinks that dark matter might instead be complex, something that interacts with itself strongly in the way that ordinary matter interacts with itself to form intricate structures like atoms and atomic elements. Such a self-interacting dark matter, Bullock suspects, could exist in a “dark sector,” somewhat parallel to our own light sector, but detectable only through the way it affects gravity.

He and his colleagues have created numerical simulations that predict what the universe would look like if dark matter feels strong interactions. They expected to see the model fail. Instead, they found that it was consistent with what astronomers observe.

Quanta Magazine spoke with Bullock about complex dark matter, how this mysterious mass might behave, and the best places in the universe to find it. An edited and condensed version of the interview follows.

QUANTA MAGAZINE: What do we know about dark matter?

JAMES BULLOCK: We are confident that it’s there, that it has mass, and that it tugs on itself and on other things via gravity. That’s about it. While dark matter has a gravitational tug, it doesn’t interact with normal matter—the stuff that makes up you and me—in a very intense way. It doesn’t shine. It’s invisible. It’s transparent. It doesn’t glow when it gets hot. Unfortunately, those are the ways astronomers usually study the universe; we usually follow the light.

So we don’t know what it’s made of?

We’ve come to understand that we can describe the world that we experience by the Standard Model of particle physics. We think of the particles that make up you and me as being broken down into constituent things, like quarks, and those quarks combine into neutrons and protons. There is a complicated dance that allows these particles to interact in certain ways. It gives rise to the periodic table of elements and all of the vast complexity we see around us. Just 20 percent of the mass of the universe is all of this complexity.

On the other hand, dark matter makes up something like 80 percent of the mass. First-guess models for what it is suggests that it is one particle that doesn’t really interact with much of anything—WIMPs. These are collisionless, meaning when two dark matter particles come at each other they basically go through each other.

Another possibility is this 80 percent of the universe is also complex. Maybe there’s something interesting going on in what’s called the dark sector. We know that whatever ties us to the dark matter is pretty weak or else we would have already seen it. This observation has led to the belief that all the interactions that could be going on with dark matter are weak. But there’s another possibility: When dark matter particles see themselves, there are complex and potentially very strong interactions. There even could be dark atoms and dark photons.

Those two worlds—this dark sector and our own sector—only communicate by gravity and perhaps other weak processes, which haven’t yet been seen.

How can you probe this dark sector if you can’t interact with it?

Now what we’re talking about doing is not just looking at the gross properties of the dark matter but the very makeup of the dark matter, too. The most obvious place to see those effects is where dark matter is bunched up. We believe the centers of galaxies and galaxy clusters are densest. And so by studying the behavior of dark matter by indirect methods—basically by the dynamics of stars and gas and galaxies in galaxy clusters—we can start to understand how dark matter is distributed in space. To start to discriminate between models, we can compare differences in dark matter’s spatial clumpiness in simulations, for example, and then look for those differences in data.

What does the data say?

In models using cold, collisionless dark matter—WIMPs—the dark matter is very dense at the middle of galaxies. It appears that those predicted densities are much higher than what’s observed.

What might be going on is that something a little more complex is happening in the dark sector, and that complexity is causing these slight disagreements between theory and observation at places where the dark matter is really clumped or starts congregating, like in the centers of galaxies or the centers of galaxy clusters.

I’m interested in running cosmological simulations of how the universe should evolve from the very beginning until now. I look at what happens, when I run those simulations forward, if I allow cold dark matter to occasionally collide and exchange energy. The simulations start with a small, almost-smooth primordial universe and end with beautiful agreement with large-scale structure—galaxies stretched out across the universe in the way we observe them. But the hearts of galaxies are less dense in dark matter in my simulations than they are in simulations where the dark matter is cold and collisionless.

How long have researchers known about these disagreements between the models and the data?

We’ve known that there’s a bit of a problem at the centers of galaxies for about 20 years. At first it was thought maybe we’re interpreting the data wrong. And now the question comes down to: Does galaxy formation eject dark matter somehow, or do we need to modify our understanding of dark matter?

Why did you start looking into self-interacting dark matter?

The first paper exploring ideas that the dark matter might be more complex was in Physical Review Letters, April 2000, by David Spergel and Paul Steinhardt. I actually started working on this several years later when I began seeing papers from the particle physics community exploring these ideas. My initial reaction was, that couldn’t be true, because I had this prejudice that things work so well with collisionless dark matter.

In the first set of simulations we ran, we gave dark matter a cross-section with itself. The bigger the cross-section is, the higher the probability that these particles are going to run into one another in any given amount of time. We set the value of the cross-section to something we were convinced would be ruled out [by the data], but when we ran our simulation we found that we couldn’t see any difference between that model and the classic one. And so we thought maybe we don’t know quite as much as we thought we knew.

Then, we dialed it up and looked at a strong interaction similar to if you threw two neutrons together. We saw something that looks really close to observations on large scales but does produce differences in the hearts of galaxies. Rather than the dark matter getting denser and denser as you approach the center of the galaxy, it reached a threshold density.

Could it be that these little discrepancies we’ve been seeing in the observational data are actually a clue that there’s something interesting and fun going on in the dark sector that we weren’t thinking about before?

How have these simulations evolved since the first ones you performed?

We’ve been running very high cross-section values to see when this model starts to break compared to some observations. We’re also focusing energy on including all of the star-formation and galaxy-formation physics in these simulations. The hardest part with these simulations is that the universe isn’t just made of dark matter. There’s all of this other annoying normal stuff that we have to think about, too. Gas that can turn into stars—and some of those stars are going to be so massive that they blow up as supernovae. When they blow up as supernovae, they are effectively jostling the gravitational field around them, and this jostling can potentially move the dark matter around. Is it possible that these discrepancies that we’re seeing in the observed densities of dark matter and the predicted densities of dark matter is because the galaxy-formation process itself is changing things in a way that we don’t understand very well?

Something else that I spend my time on is figuring out the cleanest and clearest cases for determining what comes from the physics of dark matter versus the physics of star formation and galaxy formation. We have to think hard about how clean our cosmological experiments are.

Where is that cleanest cosmological laboratory?

My opinion is that the cleanest sites are the teeniest, tiniest galaxies we know about—dwarf galaxies. They have very few stars but huge amounts of dark matter. In some cases they have 100 times as much dark matter within their visible extent as they have visible matter. (The Milky Way interior to the Sun is about half dark matter and half normal matter.) Dwarf galaxies have so much dark matter compared to their stars, they’re excellent laboratories for dark matter. They’re as clean as we have.

But studying dark matter physics in something that doesn’t give off much light is pretty difficult.

The nice thing about these objects is that a lot of them are really close by. They’re close enough that you can actually measure the velocities of individual stars. That allows you to build as precise a model as you can of the dark matter density at the centers of these galaxies. They’re close enough to study with great precision, but they’re chock full of dark matter so you don’t have to worry as much about what’s going on with the stars.

There have been recent observational studies focusing on galaxy clusters. Are observations and theoretical models starting to move in a similar direction?

Imagine a swarm of bees; a cluster of galaxies is sort of like that. Massive collisions, where two galaxy clusters have come at each other and pass through each other, are one place to look for complex dark matter. If the dark matter is strongly interacting, when those massive clusters come together, the galaxies will keep flying right on through, but the dark matter, because it’s strongly interacting with itself, will sort of bunch up in the middle.

The Bullet Cluster shows the aftermath of a cosmic collision between two galaxy clusters. In this false-color image, the hot gas (pink) slowed down in the collision due to a drag force, while the dark matter (blue) appeared to keep passing through, as one would expect if dark matter is collisionless.The Bullet Cluster shows the aftermath of a cosmic collision between two galaxy clusters. In this false-color image, the hot gas (pink) slowed down in the collision due to a drag force, while the dark matter (blue) appeared to keep passing through, as one would expect if dark matter is collisionless. NASA

In the famous example of the Bullet Cluster, astronomers used the effect of gravitational lensing to look at where the dark matter was. They found that the dark matter has moved right on through along with the galaxies, which is what you’d expect with collisionless dark matter. Because of this result, people said, “Well, there’s no way the dark matter is strongly interacting with itself.”

That was a few years ago and a couple things have happened since then. We’ve realized that a lot of the first-order estimates people have used to determine how much the dark matter ought to drag on itself were overestimated. Also, several other clusters have less-clear results, and in some cases maybe there is more drag than we thought before. Richard Massey’s group found evidence that some kind of dark pressure, ram pressure, is ripping the dark matter out of a galaxy.

We really aren’t at the point yet where I think we’ve done enough, though. We need to invest more effort into simulating the calculations properly with these various classes of dark matter to figure out what it is we know and what it is we don’t know. I think we’ve seen exciting hints, and they motivate us to try to do as well as we can to figure out what they mean.

Original story reprinted with permission from Quanta Magazine, an editorially independent publication of the Simons Foundation whose mission is to enhance public understanding of science by covering research developments and trends in mathematics and the physical and life sciences.

Go Back to Top. Skip To: Start of Article.


Dark Matter May Be More Complex Than Physicists Thought