Ana Kova for Quanta Magazine

Ana Kova for Quanta Magazine

At a talk last month in Santa Barbara, California, addressing some of the world’s leading astrophysicists, Selma de Mink cut to the chase. “How did they form?” she began.

“They,” as everybody knew, were the two massive black holes that, more than 1 billion years ago and in a remote corner of the cosmos, spiraled together and merged, making waves in the fabric of space and time. These “gravitational waves” rippled outward and, on Sept. 14, 2015, swept past Earth, strumming the ultrasensitive detectors of the Laser Interferometer Gravitational-Wave Observatory (LIGO). LIGO’s discovery, announced in February, triumphantly vindicated Albert Einstein’s 1916 prediction that gravitational waves exist. By tuning in to these tiny tremors in space-time and revealing for the first time the invisible activity of black holes — objects so dense that not even light can escape their gravitational pull — LIGO promised to open a new window on the universe, akin, some said, to when Galileo first pointed a telescope at the sky.

Already, the new gravitational-wave data has shaken up the field of astrophysics. In response, three dozen experts spent two weeks in August sorting through the implications at the Kavli Institute for Theoretical Physics (KITP) in Santa Barbara.

Read the related Abstractions post:
Responding Rapidly to Big Discoveries

Jump-starting the discussions, de Mink, an assistant professor of astrophysics at the University of Amsterdam, explained that of the two — and possibly more — black-hole mergers that LIGO has detected so far, the first and mightiest event, labeled GW150914, presented the biggest puzzle. LIGO was expected to spot pairs of black holes weighing in the neighborhood of 10 times the mass of the sun, but these packed roughly 30 solar masses apiece. “They are there — massive black holes, much more massive than we thought they were,” de Mink said to the room. “So, how did they form?”

The mystery, she explained, is twofold: How did the black holes get so massive, considering that stars, some of which collapse to form black holes, typically blow off most of their mass before they die, and how did they get so close to each other — close enough to merge within the lifetime of the universe? “These are two things that are sort of mutually exclusive,” de Mink said. A pair of stars that are born huge and close together will normally mingle and then merge before ever collapsing into black holes, failing to kick up detectable gravitational waves.

E. Buunk

Selma de Mink of the University of Amsterdam has devised a new theory stating that pairs of black holes close enough to merge come from massive stars whose contents have been mixed until they are homogeneous throughout.

Nailing down the story behind GW150914 “is challenging all our understanding,” said Matteo Cantiello, an astrophysicist at KITP. Experts must retrace the uncertain steps from the moment of the merger back through the death, life and birth of a pair of stars — a sequence that involves much unresolved astrophysics. “This will really reinvigorate certain old questions in our understanding of stars,” said Eliot Quataert, a professor of astronomy at the University of California, Berkeley, and one of the organizers of the KITP program. Understanding LIGO’s data will demand a reckoning of when and why stars go supernova; which ones turn into which kinds of stellar remnants; how stars’ composition, mass and rotation affect their evolution; how their magnetic fields operate; and more.

The work has just begun, but already LIGO’s first few detections have pushed two theories of binary black-hole formation to the front of the pack. Over the two weeks in Santa Barbara, a rivalry heated up between the new “chemically homogeneous” model for the formation of black-hole binaries, proposed by de Mink and colleagues earlier this year, and the classic “common envelope” model espoused by many other experts. Both theories (and a cluster of competitors) might be true somewhere in the cosmos, but probably only one of them accounts for the vast majority of black-hole mergers. “In science,” said Daniel Holz of the University of Chicago, a common-envelope proponent, “there’s usually only one dominant process — for anything.”

Star Stories

NASA, ESA, F. Paresce, R. O’Connell and the Wide Field Camera 3 Science Oversight Committee

The R136 star cluster at the heart of the Tarantula Nebula gives rise to many massive stars, which are thought to be the progenitors of black-hole binaries.

The story of GW150914 almost certainly starts with massive stars — those that are at least eight times as heavy as the sun and which, though rare, play a starring role in galaxies. Massive stars are the ones that explode as supernovas, spewing matter into space to be recycled as new stars; only their cores then collapse into black holes and neutron stars, which drive exotic and influential phenomena such as gamma-ray bursts, pulsars and X-ray binaries. De Mink and collaborators showed in 2012 that most known massive stars live in binary systems. Binary massive stars, in her telling, “dance” and “kiss” and suck each other’s hydrogen fuel “like vampires,” depending on the circumstances. But which circumstances lead them to shrink down to points that recede behind veils of darkness, and then collide?

The conventional common-envelope story, developed over decades starting with the 1970s work of the Soviet scientists Aleksandr Tutukov and Lev Yungelson, tells of a pair of massive stars that are born in a wide orbit. As the first star runs out of fuel in its core, its outer layers of hydrogen puff up, forming a “red supergiant.” Much of this hydrogen gas gets sucked away by the second star, vampire-style, and the core of the first star eventually collapses into a black hole. The interaction draws the pair closer, so that when the second star puffs up into a supergiant, it engulfs the two of them in a common envelope. The companions sink ever closer as they wade through the hydrogen gas. Eventually, the envelope is lost to space, and the core of the second star, like the first, collapses into a black hole. The two black holes are close enough to someday merge.

Because the stars shed so much mass, this model is expected to yield pairs of black holes on the lighter side, weighing in the ballpark of 10 solar masses. LIGO’s second signal, from the merger of eight- and 14-solar-mass black holes, is a home run for the model. But some experts say that the first event, GW150914, is a stretch.

In a June paper in Nature, Holz and collaborators Krzysztof Belczynski, Tomasz Bulik and Richard O’Shaughnessy argued that common envelopes can theoretically produce mergers of 30-solar-mass black holes if the progenitor stars weigh something like 90 solar masses and contain almost no metal (which accelerates mass loss). Such heavy binary systems are likely to be relatively rare in the universe, raising doubts in some minds about whether LIGO would have observed such an outlier so soon. In Santa Barbara, scientists agreed that if LIGO detects many very heavy mergers relative to lighter ones, this will weaken the case for the common-envelope scenario.

Lucy Reading-Ikkanda for Quanta Magazine

This weakness of the conventional theory has created an opening for new ideas. One such idea began brewing in 2014, when de Mink and Ilya Mandel, an astrophysicist at the University of Birmingham and a member of the LIGO collaboration, realized that a type of binary-star system that de Mink has studied for years might be just the ticket to forming massive binary black holes.

The chemically homogeneous model begins with a pair of massive stars that are rotating around each other extremely rapidly and so close together that they become “tidally locked,” like tango dancers. In tango, “you are extremely close, so your bodies face each other all the time,” said de Mink, a dancer herself. “And that means you are spinning around each other, but it also forces you to spin around your own axis as well.” This spinning stirs the stars, making them hot and homogeneous throughout. And this process might allow the stars to undergo fusion throughout their whole interiors, rather than just their cores, until both stars use up all their fuel. Because the stars never expand, they do not intermingle or shed mass. Instead, each collapses wholesale under its own weight into a massive black hole. The black holes dance for a few billion years, gradually spiraling closer and closer until, in a space-time-buckling split second, they coalesce.

De Mink and Mandel made their case for the chemically homogeneous model in a paper posted online in January. Another paper proposing the same idea, by researchers at the University of Bonn led by the graduate student Pablo Marchant, appeared days later. When LIGO announced the detection of GW150914 the following month, the chemically homogeneous theory shot to prominence. “What I’m discussing was a pretty crazy story up to the moment that it made, very nicely, black holes of the right mass,” de Mink said.

However, aside from some provisional evidence, the existence of stirred stars is speculative. And some experts question the model’s efficacy. Simulations suggest that the chemically homogeneous model struggles to explain smaller black-hole binaries like those in LIGO’s second signal. Worse, doubt has arisen as to how well the theory really accounts for GW150914, which is supposed to be its main success story. “It’s a very elegant model,” Holz said. “It’s very compelling. The problem is that it doesn’t seem to fully work.”

All Spun Up

Along with the masses of the colliding black holes, LIGO’s gravitational-wave signals also reveal whether the black holes were spinning. At first, researchers paid less attention to the spin measurement, in part because gravitational waves only register spin if black holes are spinning around the same axis that they orbit each other around, saying nothing about spin in other directions. However, in a May paper, researchers at the Institute for Advanced Study in Princeton, N.J., and the Hebrew University of Jerusalem argued that the kind of spin that LIGO measures is exactly the kind black holes would be expected to have if they formed via the chemically homogeneous channel. (Tango dancers spin and orbit each other in the same direction.) And yet, the 30-solar-mass black holes in GW150914 were measured to have very low spin, if any, seemingly striking a blow against the tango scenario.

Courtesy of Daniel Holz

Daniel Holz of the University of Chicago works on the classic common-envelope explanation for the formation of black-hole binaries.

“Is spin a problem for the chemically homogeneous channel?” Sterl Phinney, a professor of astrophysics at the California Institute of Technology, prompted the Santa Barbara group one afternoon. After some debate, the scientists agreed that the answer was yes.

However, mere days later, de Mink, Marchant, and Cantiello found a possible way out for the theory. Cantiello, who has recently made strides in studying stellar magnetic fields, realized that the tangoing stars in the chemically homogeneous channel are essentially spinning balls of charge that would have powerful magnetic fields, and these magnetic fields are likely to cause the star’s outer layers to stream into strong poles. In the same way that a spinning figure skater slows down when she extends her arms, these poles would act like brakes, gradually reducing the stars’ spin. The trio has since been working to see if their simulations bear out this picture. Quataert called the idea “plausible but perhaps a little weaselly.”

Lucy Reading-Ikkanda for Quanta Magazine; Source: LIGO

On the last day of the program, setting the stage for an eventful autumn as LIGO comes back online with higher sensitivity and more gravitational-wave signals roll in, the scientists signed “Phinney’s Declaration,” a list of concrete statements about what their various theories predict. “Though all models for black hole binaries may be created equal (except those inferior ones proposed by our competitors),” begins the declaration, drafted by Phinney, “we hope that observational data will soon make them decidedly unequal.”

As the data pile up, an underdog theory of black-hole binary formation could conceivably gain traction — for instance, the notion that binaries form through dynamical interactions inside dense star-forming regions called “globular clusters.” LIGO’s first run suggested that black-hole mergers are more common than the globular-cluster model predicts. But perhaps the experiment just got lucky last time and the estimated merger rate will drop.

Adding to the mix, a group of cosmologists recently theorized that GW150914 might have come from the merger of primordial black holes, which were never stars to begin with but rather formed shortly after the Big Bang from the collapse of energetic patches of space-time. Intriguingly, the researchers argued in a recent paper in Physical Review Letters that such 30-solar-mass primordial black holes could comprise some or all of the missing “dark matter” that pervades the cosmos. There’s a way of testing the idea against astrophysical signals called fast radio bursts.

It’s perhaps too soon to dwell on such an enticing possibility; astrophysicists point out that it would require suspiciously good luck for black holes from the Big Bang to happen to merge at just the right time for us to detect them, 13.8 billion years later. This is another example of the new logic that researchers must confront at the dawn of gravitational-wave astronomy. “We’re at a really fun stage,” de Mink said. “This is the first time we’re thinking in these pictures.”

This article was reprinted on

View Reader Comments (16)

Leave a Comment

Reader CommentsLeave a Comment

  • Thanks for covering this story, Quanta. I really appreciate your very careful efforts to cover the complex discussions going on.

    It would like to stress that there are many more people behind the ideas and models presented here. The days where a single person came up with a new theory seem from the past. Science is the results of teamwork and interaction between many people these days. The list of about thousand co-authors of the discovery and analysis papers certainly show that. All credit should firstly and foremost to this team (of which I am no part).

    The article does partially misplace credit in a few places, in some cases in my favor. The original paper on homogenous evolution in tight binaries had several coauthors that contributed, which include M. Cantiello, N. Langer, O.R. Pols, I. Brott, and S.-Ch Yoon. This paper does not mention binary black holes nor gravitational waves. The idea of homogeneous evolution as a pathway to form binary black holes was pointed out in two papers that came out this year, nearly simultaneously. One was lead by P. Marchant and one was co-lead by Mandel and myself.

    Also the fact that "most known massive stars live in binary systems" had been pointed out by many before 2012. Moreover, this particular paper was in fact lead by H. Sana.

  • This is a fascinating text, I am not a scientists, but love everything about physics and captures my imagination, hope to see news about this amazing topic (black holes are exiting things).

  • Quick question; I looked but couldn't find the obvious explanation: these these were two supermassive galactic cores that had been stripped (or eaten) all their star and then merged. Why is normal BH production in two galaxies that then merged not discussed but complicated models of native binary star system two-BH formation is? The galactic center model would go a long way to explain how they have spin and reach such large sizes.

  • What's missing from this article is even a modicum of scientific perspective. The discussion is entirely about theoretical mathematical models that have little or no real correspondence to physical reality. Set aside for the moment the dubious LIGO claims of a gravitational wave detection (, the modern theory of black holes is only a mathematical model.

    This theoretical black hole model as currently presented is a physical impossibility. It is a physical impossibility because, according to the model, one of its defining features is a singularity and singularities are physically impossible states. That is to say that singularities cannot and do not exist in the physical reality we actually observe and measure.

    Therefore it is not reasonable to claim that black holes as currently defined by theory actually exist in physical reality. Theorists of the cosmological scale, to the extent that their models do not resemble physical reality are only doing mathematics, not science.

    Science is limited in its scope to that which can be observed and measured (by our instruments as well as our senses). It is that empirical requirement that has allowed science to stand out in its effectiveness against all other philosophical disciplines, including mathematics. To abandon empiricism is to abandon science.

  • Mr. Rapanault, You might benefit (and your opinion might be changed) by reading "Boltzmann's Atom" by David Lindley. Boltzmann's scientific critic of that time was Ernst Mach, a scientist and philosopher who took the same perspective you do. Boltzmann used just the mathematical and theoretical perspectives you denounce. Boltzmann is remembered as a great scientist who preceded and influenced Einstein — while Mach is generally only remembered for Mach numbers (as in Mach 1, Mach 2, etc) . Mach is still considered important — but not great.

  • "and singularities are physically impossible states. "

    The north and south poles are singularities yet they exists. They points to a deficiency in the stereo-graphic mapping system which means we should search for and use a new mapping system that doesn't have singularities, like hexagonal/pentagonal tile mapping.

    Similarly, "abnormally large collections of mass", call them black holes, exist and the singularity points to a deficiency in our current mapping system, i.e., physics.

    Hence, the quest to understand black holes is the quest to understand a better mapping system for the universe, a better physics.

  • Mr Conway, You misunderstand my critique of the article. I was not denouncing mathematical and theoretical perspectives at all. I was denouncing the absence of a scientific perspective both in the article and in the broader realm of modern cosmology.

    Perspective can only be had from multiple viewpoints. You can have no perspective without, at minimum, binocular vision. With regard to science, it might be helpful to think of science as the apex of an equilateral tetrahedron (three-sided pyramid). The base triangle of that structure consists of three vertices, theory, mathematics, and empiricism (observation and measurement). So a proper scientific perspective incorporates and balances the theoretical, the mathematical, and the empirical viewpoints.

    My argument then is: it is precisely that integration of all three necessary elements that is absent both in the article and in modern cosmology generally, and it is specifically the absence of the empirical element that is most often the problem. In this context then, black hole theory is a vivid illustration of the lack of scientific perspective in areas that are nominally "scientific".

    Modern black hole theory is widely believed to represent real physical objects that exist in the cosmos. And yet according to the theory itself, black holes contain a singularity. A singularity is, however, nothing but a name given to an invalid mathematical operation (division by zero) that is both theoretically and empirically meaningless.

    This division by zero is a consequence of taking a mathematical model of gravitation that can be said to work on one scale, the stellar, and simplistically extrapolating it down to a scale on which it yields, mathematical, theoretical and physical nonsense. And therein lies the entire basis for modern black hole theory.

    And on the basis of that dubious theory, black holes are presented, both in this article and more broadly in all of modern cosmology as if they were real physical objects. However, scientifically speaking, there is no evidence from the realms of mathematics, physical theory, or empiricism to support the idea that such physical objects can and do exist. How is it that an invalid mathematical operation lies at the heart of a commonly accepted physical theory, one that also lacks theoretical and empirical support? What has become of the scientific perspective?

    I'll take a look at the book you recommend, it sounds interesting. For a historical illustration of how empiricism, theory and mathematics work together to produce sound science I would highly recommend the recent "Faraday, Maxwell, and the Electromagnetic Field" by Nancy Forbes and Basil Mahon.

  • In the classical theory, the "singularity" is hidden behind an horizon and is unobservable – and there are hypothesis (no naked singularities) that this is always the case. So the classical theory is consistent, and there are observations of velocity dependence near the horizon that support classical black hole theory. As for the singularity which can only be observed if you fall through the horizon, that issue would need to be resolved by a quantum theory of gravity which will probably do away with the singularity at the origin. This all leads back to all the firewall controversy and you may want to look at the many papers on that subject. But classical black hole theory is consistent and has some if not overwhelming experimental support.

  • Its been my thinking that these intermediate sized black holes were created through a direct collapse mechanism very early in the formation of the universe. Once formed they began to consolidate, many becoming / contributing their mass to the black holes in the cores of galaxies immersed in their halos of black hole siblings.

  • A few of the authors of arXiv:1207.6397 (cited in this article) have this month uploaded another paper,, that cites the former. With a naïve perspective one may wonder if the latter has relevance to the above topic.

  • Regarding the protestations and overly rigorous and zealous wish for application of scientific principles by Bud Rapanault – "… missing from this article is even a modicum of scientific perspective." I believe that he also misses many well know scientific facts:
    1. His lack of agreement or understanding of the meaning of a "Singularity"; e.g., his (inappropriate [?]) definition of "And yet according to the theory itself, black holes contain a singularity. A singularity is, however, nothing but a name given to an invalid mathematical operation (division by zero) that is both theoretically and empirically meaningless." [This statement is fundamentally flawed.) A black hole certainly does not go to zero dimension. It has dimension and occupies space/time no matter how distorted. Just as the problems that arise when dealing with problems and conundrums when dealing with "Infinity" and/or division by zero". They are humanly derived numbering concepts for the counting and the accounting of things, and for human amusement, advancement and edification of mathematicians. By my definition the problem with division with zero arises because of the inappropriate application of our humanly derived numbering/counting system to scientific scales in both the micro and the comic arenas. In the grand scheme of things there has never been and will never be any such thing a "absolutely" nothing, or zero. There has always been something, even before the Big Bang – I suggest that it was two equal and opposite polarities each of enormous energy of "One Universal Unit-less unit of One (1)". Therefore, for scientific applications, zero does not ever exist in real terms, just as infinity can never exist. Energy decreases by coalescing into matter. Matter and Space/Time will never exceed Universal Units of One. Everything is bounded. The model of this universe is limited, bounded and flat.
    2. Regarding the following statement by Bud Rapanault – "… there is no evidence from the realms of mathematics, physical theory, or empiricism to support the idea that such physical objects can and do exist. How is it that an invalid mathematical operation lies at the heart of a commonly accepted physical theory, one that also lacks theoretical and empirical support? What has become of the scientific perspective?" You arrogance in pontificating, negating and presuming to obliterate all of the scientific work by innumerable world experts in this vast fied is astounding and extreme to say the least. a.) What is the basis of your believe that an invalid mathematical operation lies "… at the heart of a commonly accepted physical theory"?
    b.) Your presumptuous statement "… there is no evidence from the realms of mathematics, physical theory, or empiricism to support the idea that such physical objects can and do exist." You fail to take into account that for years, and even decades, there have been and are many ways to 'see and measure' empirical evidence indirectly – such a lensing of light which is a valid form of first hand empirical measurement.
    c.) I believe that your overall problem and difficulty is one of excessive hubris and ignorance of specific definitions, lack of understanding of mathematical models and special scientific methods being used in the study of this subject.

    Notwithstanding all of my comments. You are free to remain in your ignorance and to hold and express opinions regardless of whether they are based on the very methods you so vigorously protest that are "What's missing from this article is even a modicum of scientific perspective." are also sorely missing within your very own comments. Support your own arguments with valid scientific research on the subject.

    Respectfully, Robert Morin.

  • Do black holes have any "function" or are they merely cosmological nuisances?
    The scientific significance of these discoveries is that they raise questions instead of confirming what had been known all along.
    Question: if the collision produced gravity waves what happened to them? Absorbed somewhere? Left our universe?
    Here is an event with which the reality of gravity waves might be established in principle but probably not in practice. A large asteroid hits our Moon. In the process the Moon loses mass. Is that recorded simultaneously on Earth or is there a time lag equal to the distance Moon-Earth divided by c? Similarly: a huge asteroid hits Earth. Earth gains mass. Needs to speed up.

  • Thank you for a fascinating and readable in-depth article. I have a single question on LIGO event detections of the merger type: does the analysis permit us to unambiguously deduce that gravitational influences within the sphere of influence of the merger propagate exactly at c? In other words, can models be created which work when this condition does not hold?

  • The wide spectrum of ideas and opinions presented here—in both the article and the comments—is in itself noteworthy.

    Taking the LIGO observations for face value, many astrophysicists feel compelled to modify old models or invent new ones to be consistent with LIGO's analysis. At the opposite extreme, one of the commenters (Rapanault) is skeptical of the whole enterprise because of its basis in hypothetical objects whose theoretical prediction one of Einstein's collaborators (Peter G. Bergmann) considered an indication that General Relativity “carries within itself the seeds of its own destruction.”

    I suggest two things. First, a middle ground, whereby we reserve judgment on LIGO until its results are corroborated by simultaneous EM-spectrum observations (intense X-Ray and/or gamma ray events) as are expected to be produced in the predictably more common neutron star-binary collisions.

    Second, as we await further data to arrive from existing observatories, we should build an apparatus to investigate an unexplored, though much more accessible domain of gravity, one that has invited probing since 1632. That is when Galileo inquired as to what happens when a test object falls into an <em>ordinary</em> hole through the center of an <em>ordinary</em> body of matter.

    All this discussion and investment in exotic states of matter and extremely HIGH-energy collisions is arguably premature, because we have not yet tested this common prediction of both Newton's and Einstein's theories of gravity. We have not yet confirmed the simple prediction involving <em>collision-free</em> radial motion to the center of a gravitating body under conditions of extremely LOW energy. The apparatus needed to conduct the test may thus be called a Small Low-Energy Non-Collider, which could be designed to work either in an Earth-based laboratory or an orbiting satellite.

    It is common practice to <em>assume</em> that we know the result of Galileo's belated gravity experiment. Were he alive today, however, the Father of Modern Science might not be very impressed with leaving this assumption untested for nearly 400 years. While the contentious issues of astrophysics and high-budget gravity research sort themselves out, why not let's make the much more modest investment to fill a long-standing gap in our empirical knowledge of gravity? Why not let's build and operate humanity's very first Small Low-Energy Non-Collider?

  • Great work! As an interested layman, I've always thought / hoped that we have a 'breathing' universe. Ultimately, the black holes all eat each other until there is only one, which then creates another Big Bang and we do it all over again. So there's always another chance to get it right!

Comments are closed.