Black Holes or Boson Stars: The Mystery of GW190521

A long time ago, in a galaxy far, far, away, two massive objects merged to form a black hole after a collision so powerful that ripples in the fabric of space and time sent a shudder through the Universe that we could detect here on Earth. 

But in a plot twist that would make Hollywood proud, researchers are now suggesting that instead of the collision having been between two black holes as first thought, it was actually the merger of some rather more exotic objects - boson stars, with the data being presented in a new article published today in the peer-reviewed scientific journal Physical Review Letters.

If you’ve been keeping track of the incredible discoveries that have been made over the last few years courtesy of gravitational-wave astronomy, you may well remember GW190521. Announced in September of 2020, it was thought to be a signal consistent with the formation of an intermediate-mass black hole, the first one ever observed, after the collision and merger of two stellar-mass black holes. 

But this interpretation posed an enormous challenge to our understanding of stellar evolution, mostly because the heavier of the two colliding progenitor black holes simply should not have existed.

The initial interpretation of the event posited that the two black holes, 66 times and 85 times the mass of the Sun, collided and merged to form another black hole with a mass 142 times larger than the mass of the Sun. For those keeping up with the math, the remaining 9 solar masses were radiated away as energy in the form of gravitational waves.

Since no black holes had ever previously been observed in the 100 to 100,000 solar mass range, the mass range of intermediate-mass black holes, the detection was one of the most remarkable finds in gravitational-wave astronomy. But the excitement it generated in the astronomical community was somewhat tempered by the fact that one of the progenitor black holes existed in what is known as the pair-instability mass gap.

To understand why this is problematic, let’s take a quick detour into some stellar astrophysics.

Astronomers have known for a while that black holes come in a variety of different sizes, from the ones comparable in scale to the size of a star to the supermassive ones that inhabit the centres of galaxies. There is also speculation, courtesy of work done by Stephen Hawking, that there may have been other black holes that formed soon after the Big Bang with masses as low as a fraction of a grain of sand.

Indeed, scientists have pretty conclusive evidence for the existence of both stellar-mass and supermassive black holes, including an incredible photograph from 2019 of the supermassive black hole at the centre of galaxy M87, and several previous gravitational-wave signals confirming the existence of stellar-mass black holes.

We’re not entirely certain about how supermassive black holes form yet but are pretty confident that the stellar-mass ones are left-overs from the internal collapse of stars that are at least 8 times as massive as the Sun. Those events, known as supernovae, blast a good portion of the stars into space, and we end up with black holes that are somewhat less massive than the stars that preceded them.

That means that the bigger the star, the bigger will be the resulting black hole. Sounds logical, but there is an upper limit.

Above a certain mass, the temperature inside the star gets so hot that the photons that have been holding it up through radiation pressure and preventing gravitational collapse spontaneously convert into matter-antimatter pairs. 

Energy is diverted away from holding up the star, and it collapses further, heating up in the process. What is now a violent runaway fusion reaction in the core of the star finally triggers an explosion that completely obliterates it, leaving nothing behind.

According to our best stellar models, any star massive enough to be capable of forming an 85 solar-mass black hole would have to have died in this way. Which means that there would not have been an 85 solar-mass black hole left at the end; no remnant at all would have remained.

That is not to say that a black hole of this mass would completely defy the laws of physics. Current leading theories suggest that it might be possible for a black hole like this to have been built up over time through mergers of smaller black holes. And there are also some other hypothesised astrophysical processes that could be responsible.

The absence of black holes of more than about 65 solar masses, though, seems to be a result of pair-instability. At the upper end of the scale, for stars more than about 250 times the mass of the Sun, a different mechanism comes into play that produces a hypernova which is followed by the collapse of the star into a black hole. But if there are stars in the Universe that are this big, we haven’t seen them yet.

All of which is to say that either we need to seriously revise our models of the interiors of stars, or we should be looking for a different explanation for the signal known as GW190521.

Gravitational waves were suggested by Einstein in 1915 to be a natural outcome of his general theory of relativity. What we perceive to be gravity is really a distortion in space-time caused by the matter in it. In other words, matter tells space how to curve.

When two massive objects, like a pair of black holes, spiral in towards each other and collide, the distortions in space-time are sent out across the cosmos as waves radiating from the location of the merger event. Travelling at the speed of light they eventually pass over the Earth where some of the largest scientific instruments ever constructed measure these ripples with such precision that we can localise the position of the collision and determine both the size and nature of the progenitor objects pre-merger, and the resultant mass and spin of the post-merger remnant.

How is this possible? Theoretical models have been developed that provide scientists with template signals that would be produced by the different types and configurations of sources. When a gravitational wave is detected, computer algorithms go to work exploring the range of possibilities, until they settle on a most likely match for the signal received.

But the conclusions are only as good as the models we use to perform the analysis. If incorrect assumptions are made when generating template signals, the recovered progenitor objects might not be as closely matched to the gravitational-wave signal as we expect.

That is just a long way of saying that science is not yet perfect and there is still a lot we don’t know. And right on cue, enter the boson star.

When we talk about stars we mostly mean the bright ones that shine at us in the night sky, the ones that are like our Sun only far more distant. But there are other stars that most of us wouldn’t recognise as such even if they were loitering somewhere within our own Solar system.

There are some stars that just do not look at all like stars. Rather than being composed nearly entirely of hydrogen and helium like other stars, they consist of matter that has no electromagnetic signature. These stars are hypothetical because we are still refining the techniques that we use to find them, but they are consistent with the laws of physics, nonetheless.

Boson stars are just one in this family of objects, known as exotic stars, and, as their name suggests, they are composed almost entirely of bosons. And what is a boson? It is one of the two types of fundamental particles, the one that carries forces. The other, the fermion, is what makes up ‘normal’ stars, and all the other matter that we see.

For the technically minded, it is relevant to mention here that there are a variety of bosons, and that boson stars are sometimes categorised to reflect the type of boson that they are made of. 

As an example, Proca stars are vector boson stars, meaning that their constituent bosons have a spin of 1. They are also a bit unique amongst boson stars, because the stars themselves can spin without being disrupted.

And what would a boson star look like? They would be completely dissimilar to any other star that we know of. For example, they would most likely be shaped like an enormous donut because of the centrifugal forces acting on the bosonic matter, and, bizarrely, they would be transparent; any matter absorbed by them would be visible at their centres.

But if boson stars do exist, they might provide the evidence we need for a long sought-after dark matter particle. That’s because the said particle, the axion, is a boson. And we’ve been searching (unsuccessfully) for axions in numerous experiments on Earth for decades.

That brings us back to the research at hand. After the initial excitement of the first-ever observation of an intermediate-mass black hole, it was quickly realised that the very existence of such an object was not consistent with any of our stellar models. Perhaps it was itself a product of previous, smaller, black hole collisions, or maybe there was something else at play.

Thus, the challenge for scientists was to come up with a theory that could explain the presence of the intermediate-mass black hole progenitor of GW190521, while still being consistent with the original signal. And by assuming that it was caused by merging boson stars, rather than black holes, an international team led by OzGrav alumni and astronomer at the University of Santiago de Compostela Dr Juan Calderón Bustillo and Dr Nicolás Sanchis-Gual of the University of Lisbon might have been able to do just that.

Apart from the problems associated with the pair-instability mass gap, any potential hypothesis needed to explain something a bit unusual about the GW190521 signal. Normally gravitational waves that originate in merging binary systems oscillate at higher and higher frequencies as the two progenitors spiral in towards each other. But for GW190521, the inspiral signal before the merger was barely detectable.

An extremely abbreviated inspiral could perhaps be explained if two black holes collided head-on rather than by circling into each other, and so that is the first thing that Dr Bustillo and Dr Sanchis-Gual’s team looked at. What they found didn’t help much.

“We first tried to fit the data to head-on collisions of black holes, but these happen to produce a final black hole whose spin is too low to reproduce the GW190521 signal. The reason is that the lack of an inspiral diminishes a lot of the spin of the final black hole, and the individual spins of the black holes, which also contribute to the spin of the final one, are bounded by a limit called the Kerr limit,” said Dr Bustillo.

That is when the team started looking at boson stars, or Proca stars to be exact. Then the pieces started to fall into place.

They compared the GW190521 signal to computer simulations of Proca star mergers and found that statistically they were a considerably better fit to the data than when it was assumed that the progenitors were black holes. Dr Bustillo explained the appeal of their hypothesis, and what it implied about the black hole that was produced in the merger.

“First, we would not be talking about colliding black holes anymore, which eliminates the issue of dealing with a forbidden black hole. Second, because boson star mergers are much weaker, we infer a much closer distance than the one estimated by LIGO and Virgo.”

“This leads to a much larger mass for the final black hole, of about 250 solar masses, so the fact that we have witnessed the formation of an intermediate-mass black hole remains true.”

This is an exciting result in itself, as the final black hole formed by the merger in this case would have to be about 62% larger than previously thought. And rather than the signal originating from a point that is now some 17-billion light-years from us, it would have been just a bit over 1.8-billion light-years distant.

(Oh, and if you are wondering how we could receive a signal from 17-billion light-years away in a universe that is less than 14-billion years old, it is because of the expansion of space. The size of the sphere of space that we can currently observe is actually more than 93-billion light-years in diameter.)

Now is a good time to be reminded that the GW190521 gravitational-wave signal was not even one-tenth of a second long. Being able to ascertain so much information from such little data is incredible, but Dr Bustillo did caution against assuming that this was the final word on the matter.

“Of course there are potentially many ways in which this event may be explained, as this is an event for which we have very little information about what produced the final black hole we observe. The best we can say right now is that the data tells us that a collision of Proca stars is approximately 8 times more likely than the black hole collision scenario.”

And what of the implications of discovering the first boson stars?

“That would be dramatic. Boson stars and their building blocks – the ultralight bosons – are one of the most solid candidates for forming what we know as dark matter.”

“We know dark matter exists because while we cannot see it, we see the gravitational effects it produces. We estimate that it forms 27% of the whole content of the Universe, while what we know as ‘ordinary matter’ is only around 3%, the rest being known as dark energy.”

“If our result is further confirmed by future observations, it would represent the first actual measurement of the particle responsible for dark matter.”

There’s a lot to take in here… A hypothetical star possibly revealed in its physical form in the depths of space, and a tiny particle that could account for over a quarter of the Universe. And the science that has given us this new view of the cosmos is really only just getting started.

“Gravitational-wave astronomy is still very much in its infancy,” said Dr Rory Smith, an astronomer from Monash University in Melbourne and one of the collaborators in this research.

“However, the fact that we are already able to start drawing connections between gravitational-wave observations and fundamental particle physics is a remarkable sign of how powerful this new field is. Even if future observations rule out boson stars as real astronomical objects, we should expect many new and exciting discoveries in the future.”

Boson stars today, tomorrow… who knows? But I can’t wait to find out.