Expert Voices

Did a Starry 'Mosh Pit' Spawn LIGO’s Gravitational Waves? (Kavli Roundtable)

Globular cluster Messier 54
The globular cluster Messier 54 as seen by the Hubble Space Telescope. Located some 90,000 light-years away, the object might actually be the star-filled core pulled out of a neighboring galaxy by the gravity of our home galaxy, the Milky Way. (Image credit: ESA/Hubble & NASA)

Adam Hadhazy, writer and editor for The Kavli Foundation, contributed this article to's Expert Voices: Op-Ed & Insights.

Much to their surprise, scientists are finding dozens of black holes deep within densely packed collections of stars called globular clusters. Astrophysicists are using a record-breaking computer simulation to learn the clusters’ secrets, including whether they gave rise to recently observed ripples in space-time. 

In 2015, after a century of speculation, the world finally detected the elusive ripples in the universe’s fabric known as gravitational waves. This happened when a wave-hunting experiment called LIGO, which acts like a colossal tuning fork, sensed these waves hurled out from the cataclysmic collision of two massive black holes. (Read more: What is LIGO?)

But where are these collisions occurring? A new paper about LIGO's third gravitational-wave detection, announced June 1, suggests that the black hole smashup might well have been inside of a beautiful object called a globular cluster — a glittering celestial "snow globe" filled with hundreds of thousands of closely packed stars. At their centers, globular clusters are believed to harbor dozens to hundreds of black holes — by far the greatest concentration of these exotic objects found anywhere in the universe. [Hunting Gravitational Waves: The LIGO Laser Interferometer Project in Photos]

Globular clusters could very well be a major source of the gravitational waves scientists are sensing with LIGO. Studying these waves could teach us more about their dense, star cluster origins, and in the process also shed light on the construction of galaxies, the universe's biggest groupings of stars. 

In this simulation, 60 black holes and 500 stars interact with each other at the chaotic core of a globular cluster until two black holes combine to form a black hole binary. Credit: Carl Rodriguez/Northwestern Visualization (Justin Muir, Matt McCrory, Michael Lannum)

The Kavli Foundation spoke with three astrophysicists about the many scientific opportunities globular clusters present for understanding the collisions of black holes as well as the workings of the broader cosmos.  

The participants were:

  • Rainer Spurzem — a professor at the Kavli Institute for Astronomy & Astrophysics at Peking University and the Chinese Academy of Sciences. He specializes in computer simulations of complex astrophysical systems such as galaxies and globular clusters.  
  • Carl Rodriguez — a Pappalardo Postdoctoral Fellow and a postdoctoral scholar at the Massachusetts Institute of Technology (MIT) as well as a member of MIT’s Kavli Institute for Astrophysics and Space Research. His research focuses on dense star clusters, including globular clusters, as well as how black holes form and behave in these crowded systems.
  • Jay Strader — an assistant professor in the department of physics and astronomy at Michigan State University. He conducts searches for black holes in globular clusters.

The following is an edited transcript of their roundtable discussion. The participants have been provided the opportunity to amend or edit their remarks.

The Kavli Foundation: Skygazers have admired the starry brilliance of globular clusters for centuries. Is it surprising that these luminous objects have dark "hearts" full of black holes, as recent computer simulations and observations are bearing out? 

Carl Rodriguez: It’s not surprising from the standpoint that as long as you start off with a large population of stars anywhere in the universe, you will end up with some ultra-bright, massive stars. These rare, monster stars are the ones that gravitationally collapse to form black holes when they die. Because globular clusters are home to so many stars, scientists had long figured some black holes would inevitably be produced in them. Yet until about ten years ago, we had no observational evidence to prove that these hordes of black holes were actually there. 

Rainer Spurzem: It has always been clear that there should be a lot of black holes in globular clusters. But the big question has been, and still is: What happens to them? Do they stay inside globular clusters or get hurled out into space soon after they form? Have we seen evidence for black holes that recently formed or have lingered inside globular clusters for billions of years? The assumption used to be that globular clusters couldn’t retain black holes. But that’s not what computer simulations by me and my colleagues, as well as in work by Carl and others, are now showing. 

Rodriguez: That’s right. In earlier models, the idea was that the black holes would essentially fall out of solution, like how heavy dust particles in the atmosphere slowly settle onto the Earth’s surface. But that’s not what we see. [Images: Black Holes of the Universe]

In the more recent computer simulations that my team and Rainer’s team have been doing independently, there’s a lot more mixing, and the black holes don’t force each other out of the globular cluster. Earlier models failed to capture that black holes have a wide range of masses, whereas we deal with black holes having as little as three times, and all the way up to 30, 50, or even 80 times the mass of the sun. So now there’s a smooth continuum. With that, we don’t see this sort of oil-and-water separation of the stars and the black holes in a globular cluster anymore.  

Jay Strader: What is great is that you’ve got these two separate lines of inquiry, of theory and experimental observations. Rainer and Carl are creating computer simulations based on theoretical models, while my team is gathering observational evidence for the existence of black holes in globular clusters. At the moment, we’re reaching the same conclusions, that globular clusters can keep their black holes. And if that’s the case, then globular clusters could be where black holes often collide and create gravitational waves. 

An artist's conception of black holes feeding on matter from companion stars and sending out bright jets into the space within a globular cluster. (Image credit: Benjamin de Bivort; Strader, et al.; NRAO/AUI/NSF)

TKF: Rainer, the simulation you just mentioned you are working on is the most realistic simulation of a globular cluster to date. Because it suggests that the gravitational waves we have been detecting with LIGO came from two black holes within a globular cluster, do you think the mystery of their origin is now solved?

Rainer Spurzem’s research on dense stellar systems relies on state-of-the-art supercomputing systems. (Image credit: Rainer Spurzem)

Spurzem: The simulation we ran is called the DRAGON Simulation Project. It’s an incredibly sophisticated simulation of a globular cluster, created by tracking a million digital stars as they interacted over 12 billion years — nearly the age of the universe. It does a great job of simulating globular clusters, and using it, we did find some merging black holes that more or less agree with the gravitational-wave events LIGO has detected. But we can’t be absolutely sure yet. 

Something we’re missing is a real census of all of the globular clusters in our cosmic neighborhood. We need to know how many are out there to see if their number correlates to the expected rate of gravitational wave events that we expect LIGO will end up detecting per year. The more globular clusters there are, the likelier it is that the first gravitational waves LIGO ever detected were from within one. We have a lot more work to do beyond DRAGON in figuring out the origin of LIGO’s detected gravitational waves. 

TKF: Carl, you’re investigating how globular clusters probably act like "factories" for making pairs of black holes. What have you learned?

Rodriguez: Globular clusters have a significantly fewer stars and black holes than you’d find in a whole galaxy. For example, the Milky Way has 200 billion or so stars, while globular clusters have maybe only a million stars. But when it comes to making pairs of black holes, it’s the density of stars that matters, and that’s where globular clusters have galaxies beat. Whereas the Milky Way is about 100,000 light-years in diameter, give or take, a globular cluster is only around 10 light-years in diameter.

This extreme density of stars allows for dynamical processes you don’t see other places in the universe. Black holes can come close enough to one another that they’ll undergo gravitational interactions and form a pair, known as a binary. You can see that in galaxies, too, but it’s incredibly rare, whereas it happens all the time in globular clusters.

TKF: You’ve also mapped out a way for astrophysicists to figure out the source of a gravitational wave. How does this work? 

Rodriguez: This goes back to my PhD thesis, where I managed to swing the term "black hole mosh pit" into an earlier paper to describe what this cosmic environment would be like. I was pleased with myself for that. [Laughter] 

Carl Rodriguez uses computer modeling for his studies examining how gravitational waves can help explain star formation, especially of massive stars, in cluster environments. (Image credit: Carl Rodriguez)

Black holes rotate along some central axis. When they orbit each other in a binary system, each of the black holes will have its own spin orientation. These spin orientations can be aligned with the black holes’ orbit, like two tops spinning toward each other on a flat table. Or, the spin orientations can differ and be misaligned, like tops spinning whichever way in three-dimensional space. LIGO can actually detect when merging black holes have these sorts of misalignments. And in the newest, third detection of gravitational waves to date, it looks like there's significant evidence that the black holes are not aligned with the orbit. If they are spinning at all, it seems to suggest that they're somewhat anti-aligned with the orbit, just as you would expect if they came from a cluster. So astrophysicists will use this information to figure out whether the merger occurred in a globular cluster or somewhere else less chaotic out in the Milky Way. 

TKF: Why do we care whether the waves were the result of a merger of two black holes within a cluster or not?

Rodriguez: Being able to discriminate these two populations of binary black holes — those inside and outside of globular clusters — would be very, very interesting and really help unlock the potential of gravitational-wave astrophysics. If we discover, for instance, that most of these black hole binaries are coming from dense star clusters, then we could map that back onto the evolution of these clusters and get a better idea about their formation and behavior. 

On the other hand, if we can discriminate two separate populations and show that, say, half of the binary black holes are coming from clusters and half are coming from binary stars outside clusters, then we could start to also answer questions about the physics of producing the universe’s most massive stars.

TKF: What else makes globular clusters so intriguing?

Strader: By learning more about what keeps globular clusters together, we can learn about the varying conditions under which stars are born. Globular clusters can form with a million stars near each other. The collective gravity of that whole stellar ensemble can help the stars stay together for many billions of years. That’s very different from most stars in the universe, which we think formed in groups of maybe a hundred or a thousand stars. Generally, these star clusters are held together very tenuously by gravity, so they quickly disperse over time. 

We can also learn about the evolution of the broader universe. Early in its history, the universe was a very dense place, which we think was quite conducive to forming massive clusters, including globular clusters. We can test that idea by examining the ages and lifetimes of globular clusters.

Spurzem: Globular clusters offer fascinating tests of celestial mechanics, of how hundreds of thousands or even millions of stars interact. They are also among the oldest objects you’ll find in our galaxy as well as others, so they can offer a unique window into understanding galaxy formation. 

TKF: Speaking of galaxy formation, Jay, you are part of the SLUGGS survey, which uses ground telescope-based telescopes to study this very topic. How are globular clusters useful for tracing the development of galaxies over the history of the universe?  

Jay Strader's work delves into globular cluster systems for insights into the formation of galaxies. (Image credit: Jay Strader)

Strader: Globular clusters are relatively bright objects. That makes them easier to observe and study at greater distances than other groups of stars. This is important, because as you move from our Milky Way out to nearby galaxies and then into more distant galaxies, often it becomes impossible to study their individual stars to learn anything about their characteristics, such as their chemical composition and age. But you can still do those studies on these galaxies’ globular clusters. So globular clusters allow you to push these sort of looking-back-in-time, "fossil" studies of how galaxies formed out to substantially larger distances, letting us learn about much more about galaxies in the universe. 

Another way that globular clusters inform us about galaxy formation actually has a connection to black holes. Recently, we’ve discovered that a lot of objects we thought were the most massive globular clusters don’t appear to be globular clusters at all. Instead, they seem to be the remnants of more massive galaxies that were torn apart by another galaxy, leaving behind only the original galaxy’s core. Some of these cores are loaded with stars, so they look like little, starry nuggets — very similar to a globular cluster. But they have just a single, supermassive black hole in their center, like we have in the center of our own Milky Way galaxy. We can study these remnants to understand supermassive black holes and how galaxies interact over the lifetime of the universe.

Spurzem: There’s an object called Omega Centauri in our own galaxy that might be one of these remnants of a more massive galaxy. Omega Centauri has long been considered the Milky Way’s largest globular cluster and it contains 10 million stars.  

Strader: That’s right. It’s also almost certainly the case that the globular cluster Messier 54 near the Milky Way is actually the nucleus of a galactic neighbor, the Sagittarius Dwarf Elliptical Galaxy. There are a few other clusters that are candidates as former galaxies that were eaten by the Milky Way.

TKF: Returning to the DRAGON simulation, Rainer, what did you and your colleagues hope to learn about globular clusters, and did you turn up any surprises? 

Spurzem: Surprisingly, there were few surprises! That showed us we are on the right track in broadly understanding globular clusters. The DRAGON simulation was successful in fitting with the observations of real globular clusters, taken by the Hubble Space Telescope, for example. The simulated cluster also retains dozens of black holes.  

DRAGON is an endpoint of a very long struggle to simulate a million moving, interacting bodies. Back in 1990, scientists thought it could be done with powerful computers in 10 years. It was not so easy! Even after we developed the computing codes, we still needed a machine that could run them. 

Going back nearly a decade, my colleagues and I began using a supercomputer at the National Astronomical Observatories of the Chinese Academy of Sciences, which I had helped acquire and is one of the reasons I first came to China. We collaborated with researchers at the Max Planck Computing and Data Facility at the Max Planck Institute for Astrophysics in Germany. This Chinese-European collaboration enabled the DRAGON simulation project, and a student of mine, Long Wang, spearheaded the effort. We finally reached our goal of running the simulation.  

TKF: Jay, instead of a supercomputer, you’re studying these black holes using radio data collected by the Very Large Array, a ground-based telescope. How do these instruments let you "see" and probe black holes?

Strader: Well, with a black hole, it’s right there in the name — you can’t see it. So what you’re looking for is indirect evidence. That might be a companion star or gas that surrounds the black hole and is heated up by it. 

These signatures don’t tell you whether an object is a black hole or something else called a neutron star — the dense remnants of stars that weren’t quite massive enough to produce a black hole when they died. So my work with the Very Large Array, which is this big radio telescope in New Mexico—you saw it in the movie "Contact" —takes advantage of the fact that black holes emit radio jets that are substantially brighter than any jets that come off of neutron stars. That’s what pinpoints candidate black holes for us. 

Then we look with the Hubble Space Telescope, which has the sharpest vision, allowing us to associate those radio jets with individual stars. Once we’ve done that, we can say if that star might be in a binary system with a black hole. Finally, we study that star to see if its motion provides dynamical evidence that the object it’s paired with is actually a black hole. [The Hubble Space Telescope: A 25th Anniversary Photo Celebration]

TKF What are some of the other major open questions about globular clusters and their black holes, and how might we answer them?

Strader: We’re finding candidate black holes inside of the Milky Way’s globular clusters, and we’re currently working on confirming them. A big question I have is, what is the relationships between the objects we’re finding and the total black hole population in these clusters? How many are we not finding? Most black holes inside of a globular cluster are probably not going to be observable, so how do we estimate the overall rate at which black holes form pairs that might be detectable by LIGO?

Spurzem: We’ve recently learned that globular clusters can host multiple generations of stars, meaning that not all of the cluster’s stars were born at the same time. That goes against our standard models, which say the stars in globular clusters should all be about the same age. 

The other question is rotation in globular clusters. It’s a mystery, because when we look at globular clusters now, all of them are spherical. But when they formed, they must have had a natural rotation, and you would not expect them to ball up into such tight, round groups of stars. Some process allowed them to lose the rotation and form spheres. That’s been very much neglected.

Rodriguez: In certain, ancient globular clusters that have not dynamically evolved much with other clusters or galaxies, you should theoretically get very, very massive black holes, as large as 50 times the mass of the sun or even bigger. I think it would be very interesting to see if we could catch a glimpse of one of those monsters inside one of these unevolved clusters, because we have never indirectly seen a black hole like that. 

Spurzem: Something that is most fascinating is that in the intergalactic space between groups of galaxies, there can be tens of thousands of globular clusters that are just freely floating there. If this is common, it would mean there may be many, many more globular clusters around than we had previously thought.

— Adam Hadhazy, Spring 2017

Follow all of the Expert Voices issues and debates — and become part of the discussion — on Facebook, Twitter and Google+. The views expressed are those of the author and do not necessarily reflect the views of the publisher. This version of the article was published on

Join our Space Forums to keep talking space on the latest missions, night sky and more! And if you have a news tip, correction or comment, let us know at:

Adam Hadhazy
Contributing Writer

Adam Hadhazy is a contributing writer for Live Science and He often writes about physics, psychology, animal behavior and story topics in general that explore the blurring line between today's science fiction and tomorrow's science fact. Adam has a Master of Arts degree from the Arthur L. Carter Journalism Institute at New York University and a Bachelor of Arts degree from Boston College. When not squeezing in reruns of Star Trek, Adam likes hurling a Frisbee or dining on spicy food. You can check out more of his work at