8
One of the wonderful things about the night sky is that it’s available to everyone. At least, those who aren’t plagued by bad weather. Anyone with clear skies can head outside, observe the sky with or without a telescope and run through the scientific method to try and explain the observations they have made. Advancements in technology have also made doing these observations far easier, from night-sky apps that tell you exactly what you’re looking at, to telescopes and cameras that allow astrophotographers to capture images from their back gardens that would have been the dream of early twentieth-century astrophysicists. One thing technology has given us is the ability to ‘see’ without light. In a whole new way.
The majority of stars like our Sun aren’t found alone. Our Sun is quite rare in that respect – more than 50 per cent of Sun-like stars are found orbiting another star. The two stars will orbit around a common centre of mass. If the two stars are exactly the same mass, the centre of mass will be perfectly in the middle and the two will orbit like two friends who spin around holding hands, perfectly equidistant from each other, following the same orbit. But if one star is heavier than the other, then the centre of mass will be offset. Picture the two stars on a seesaw: if one star is heavier than the other, you’d have to move the pivot point from the middle to a point closer to the heavier star to get the seesaw to balance perfectly. That point is the centre of mass that they orbit around, meaning the heavier star traces a smaller orbit at a slower speed, and the smaller star a much longer one at a faster speed.
Diagrams of two stars of the same mass (left) and different masses (right) orbiting the centre of mass, marked by the crosses between them.
Two stars orbiting each other is known as a binary system, but you can throw in another star orbiting those two to give a tertiary system, or have two pairs orbiting the centre of mass between them in a quadruple system. The largest number of stars we’ve ever found in one star system (at least at time of writing) is a whopping seven. There are two seven-star systems that we know of; Nu Scorpii and AR Cassiopeiae.62 AR Cassiopeiae is a binary system, orbiting a binary system, orbiting a triple system. Nu Scorpii is slightly simpler, with a triple system orbiting a quadruple system.
The more massive the star, the more likely it is to be found in a multi-star system with companions. In the case of very small red dwarf stars (which are very low mass and faint but make up about 85 per cent of all the stars in the Milky Way63), only 25 per cent have a companion star, but that increases to more than 80 per cent for the most massive stars, which will collapse into black holes at the end of their lives. To form a massive star, you need lots of gas in one place, and so the majority of these form in big clusters of stars from a single giant gas cloud. So many stars in an astronomically speaking ‘small’ space increases the likelihood of massive stars ending up in multi-star systems.
The most massive stars also run out fuel quicker; as we learnt earlier, they live fast and die young. Because they’re so massive, the crush of gravity inwards is huge, so they have to burn more fuel to counteract it and therefore run out a lot quicker. While the Sun will live for about 10 billion years (it’s currently middle aged at around 4.5 billion years old), the most massive stars live for 100,000 years if they’re lucky; burning the brightest for the shortest of astronomical times. This means that more often than not you end up with a black hole (or neutron star or white dwarf) in orbit around a normal star which is still going to keep happily fusing hydrogen to make helium for many millions, if not billions, of years. This is what happened to our old friend Cygnus X-1 – the first ever candidate black hole from the previous chapters – and countless other systems.
The set-up of seven-star system Nu Scorpii, with stars shown by the filled circles and their orbits shown by the rings. A tertiary and quadruple system orbit each other. The tertiary system is one star orbiting a binary system of two stars. The quadruple system is one star orbiting another star which orbits a binary system of two stars.
We spot these binary systems containing black holes all across the Milky Way thanks to their X-ray light. But what if the second star in that system is also a massive star, which goes supernova and becomes a black hole? Then you’d end up with two black holes orbiting each other. The scale of the gravitational forces involved in this would be unfathomable. The stable orbit that the two stars would have been in during their lives would be completely disrupted by the two supernovae. A supernova throws the majority of the outer layers of the star, and therefore the majority of its mass, out into space, leaving only the core of the star to collapse into a black hole. So a new centre of mass would be needed to balance the system.
Since the second star going supernova throws off mass and results in a less massive black hole, that would mean the two black holes have to get closer together to find a new centre of mass. But there’s no way two black holes will reach a stable orbit when they are that close together. What happens is that they eventually end up spiralling ever closer together over millions of years to an inevitable end in the most monumental collision you’ve ever seen. Or technically, not seen. Once both of the objects in the binary system are black holes, there’s no material left to steal from a normal star to form an accretion disk which glows in X-rays. The whole system becomes completely invisible to us, at least until the very last moment.
Cast your mind all the way back to Chapter 3 and Einstein’s theory of general relativity: mass curves spacetime. Massive objects, like black holes, have the most effect, curving spacetime to its extremes. As they spiral in towards each other, two black holes in a binary system accelerate along their orbits, constantly changing the curvature of spacetime around them. Changing the curvature of spacetime to the extreme, on a regular basis, takes a phenomenal amount of energy; energy that comes from the black holes themselves. It’s like a shock to the system for space itself; it can’t contain that much energy in such a small area so the energy disperses, rippling away across space like a shockwave.
If we go back to our analogy of massive objects as basketballs on a trampoline, imagine bouncing two very heavy basketballs off the trampoline in a steady rhythm. The surface of the trampoline doesn’t stay flat – it constantly bounces up and down as it absorbs energy from the bouncing basketballs, which ripples away across the surface of the trampoline. This is how we can imagine space, curved to the extreme and back by two black holes orbiting each other. The energy involved, that can’t be contained in the area, is rippled away through space, like the ripples on the surface of a pond, as something known as a ‘gravitational wave’. A wave through space itself, changing the curvature of space as it goes, fuelled by the energy injected by two orbiting black holes.
Einstein predicted the existence of gravitational waves way back in 1915 when he first published general relativity (although he didn’t predict they’d be produced by black holes, but still by very dense, compact objects), but their existence would have to wait to be proven, at least indirectly, for another fifty-nine years. In 1974, two American astrophysicists, Joseph Taylor and Russell Hulse (both at the University of Massachusetts Amherst; Taylor was a professor and Hulse his PhD student), discovered the very first binary pulsar system, dubbed PSR B1913+16 (although it’s now known as the Hulse–Taylor binary). This system consists of two neutron stars in orbit around each other, formed after the massive stars that preceded them had gone supernova.
Hulse and Taylor were using the Arecibo telescope at the time: a huge 305-metre radio dish in Puerto Rico, most famous outside astronomy research circles for starring in the 1995 James Bond film GoldenEye and the 1997 film Contact starring Jodie Foster.64 Hulse and Taylor were caught up in the pulsar frenzy that had taken the early 1970s by storm, after Jocelyn Bell Burnell’s discovery in 1967. At first they thought they had detected a normal pulsar, which pulsed with radio waves every 59 milliseconds (i.e. it rotates on its axis 17 times per second).
But as they continued to observe their newly discovered pulsar, they noticed something strange. The pulses weren’t exactly 59 milliseconds apart – every time they took a measurement the time between pulses would be slightly longer or slightly shorter. That was weird: pulsars are some of the most precise clocks in the Universe; their period (the time between pulses) shouldn’t change. When they plotted out the times that they were measuring, they got a wave shape: a sine curve. The variations in the time between pulses came back to the same measurements: every 7¾ hours. This was so regular that they could predict what the time between the pulses would be based on how long ago their last measurement was.
Hulse and Taylor realised that this could be explained if the pulsar was in orbit around another star,65 with shorter measurements between pulses recorded as the pulsar moved towards Earth along its orbit, and longer measurements as it moved away from Earth along its orbit, repeating every 7¾ hours. This was the first time a pulsar had ever been discovered in a binary system like this, and so for the next six years, it was studied in excruciating detail until another curious property was spotted: the 7¾-hour orbit of the two stars was slowly decreasing. The orbits of the two stars were decaying; they were losing energy as they spiralled closer together. This energy was being lost to space itself and rippled away as gravitational waves.
It was Taylor, along with Lee Fowler and Australian astronomer Peter McCulloch, who published the results to the world in 1979, confirming that the decay in the orbits was exactly as Einstein had predicted (at least within the uncertainties of our knowledge on the distance of the pulsar from Earth), and not what other alternate theories of gravity being debated at the time had predicted.
This was the first indirect evidence for gravitational waves. Both Taylor and Hulse won the Nobel Prize in Physics in 1993 for their discovery of PSR B1913+16 which, according to the prize citation, was ‘a discovery that has opened up new possibilities for the study of gravitation’.66 Taylor, Hulse, Fowler and McCulloch weren’t the first people to contemplate the existence of gravitational waves though; with the technological advancements in all other areas of astronomy charging forward post-Second World War, there were some who set their sights on actually detecting gravitational waves here on Earth. This reached a fever pitch in the 1970s, after a false claim of a detection of gravitational waves in 1969 by Joseph Weber, an engineer at the University of Maryland.
Weber had a large cylinder of aluminium that he claimed rang like a gong when impacted with a gravitational wave. Weber’s supposed detections made no scientific sense, and were discredited by many leading astrophysicists at the time. But what his false claims did was spur on others to redouble the search and build their own gravitational wave detectors. The discovery of the orbital decay in PSR B1913+16 only added fuel to the fire. But how do you actually detect a gravitational wave?
Gravitational waves stretch and squash space itself as they move through it. So the distances between objects in space get shorter and longer as a wave passes by. If you can measure the distance between objects changing, then you can detect the presence of a gravitational wave. You need to do this very precisely though; the method of choice usually being to use a laser. A laser is a source of light of just one specific wavelength (and therefore the same colour, hence why you have the choice of a green or red laser pointer) that is emitted in the same direction to give a very tight beam. This means you can point it in one direction and know the majority of light will go in that direction, unlike a light bulb which emits light willy-nilly in all directions.
That means if you shine a laser at a mirror, the majority of the light will make it to the mirror and then reflect back off it, so you can still detect the same laser beam where it was first emitted (don’t actually try this with a mirror at home folks; lasers can blind). Knowing the speed of light, you can then work out the round-trip distance the laser travelled thanks to the age-old classic equation: distance = speed × time. So there you have it, an accurate way to measure the distance between objects67 (the laser and the mirror) to check if gravitational waves are passing by, squashing and stretching the distance between them.
The problem with gravitational waves, which Einstein himself pointed out, is that their effect (how much they squash and stretch) is absolutely tiny. We’re talking a change in the distance between two objects of smaller than the diameter of a proton: less than 0.0000000000000001 m. Measuring anything with that kind of precision, even with lasers, is a tall order. Instead, during the 1960s and 1970s (there’s no real consensus on who had the idea ‘first’), astrophysicists realised they could use a trick of physics to be able to measure with such precision; again due to the nature of lasers.
Constructive (top) and deconstructive (bottom) interference between waves that are in phase and out of phase.
The light given off by lasers is all the same. The peaks and troughs of every wave are lined up; physicists call that being in phase (like being in sync with someone). If you add a second laser into the mix, you can position it so that the waves in the two lasers are also in phase, so that when they meet at a detector the waves add together and you detect something twice as bright. We say that the two waves interfered with each other constructively. Or, you can position the second laser so that the waves are out of phase, misaligned so that they cancel each other out completely and no light is detected. In this case, we say the two light beams interfered with each other destructively. This is exactly how noise-cancelling headphones work – recording sound waves arriving at the headphones, and playing the inverted out-of-phase sound wave into your ear at the same time so the two cancel out with destructive interference.
So, one of the best methods to detect gravitational waves is to use this physics trick of interfering waves. Detectors are built in an L-shape, with two lasers firing at 90 degrees to each other and reflecting off mirrors to bring them back together, so they’re perfectly misaligned and they cancel each other out with destructive interference. There’s a detector perfectly positioned to record the combination of the two beams – if everything is normal, the distances between the lasers and the mirrors stay the same and the detector does not record any light. But if the distance between one of those laser-mirror pairs changes due to a passing gravitational wave, the phase difference between the two lasers changes and the detector will record a detection of some of the laser light. Depending on how bright the point is, between no detection to twice the brightness of a single laser, you can tell how much the two waves are out of phase as a fraction of the wavelength of the light emitted by the laser. This method (called interferometry, because you use the interference of the two waves) is how you can measure the tiny changes in distance between objects caused by gravitational waves, even down to less than the size of a proton.68
It was American physicist Robert L. Forward who built the very first prototype gravitational wave detector using interferometry of lasers in 1971. Each L-shape was 8.5 metres long and the detector was left for 150 hours to record any gravitational waves, but was unsuccessful (it also didn’t agree with any of Weber’s gravitational-wave ‘gong’ detectors). It was American astrophysicist Rainer Weiss from the Massachusetts Institute of Technology (MIT) who pointed out that a much bigger distance than 8.5 m between the laser and the mirror would be needed to detect gravitational waves; in the early 1970s he calculated that to detect gravitational waves from the Crab Pulsar (formed in the supernova observed by Chinese astronomers in 1054) you would need 1 km (0.62 miles) between the laser and mirror. He even went as far as to suggest building an interferometer in space.69
It was in the summer of 1975 that Rainer Weiss met with his old friend Kip Thorne, an American theoretical physicist working at the California Institute of Technology (Caltech), known for his work on black holes and general relativity.70 The two were attending a conference in Washington on cosmology and relativity and, according to Weiss, stayed up the entire night before the conference discussing the big unknowns in gravity research, before together deciding that their focus in the future should be on gravitational waves. In order to seriously tackle the problem they knew they’d need two things: 1) a whole lot of funding, and 2) an experimental physicist (Weiss and Thorne were both very much immersed in the theory of gravitational waves, and not so much in the engineering or experiment design beyond prototypes).
Funding was not easy to come by: there were many problems to overcome, including a whole bunch of technological advancements. One thing that was needed was a way to isolate both the lasers and mirrors from any seismic activity. Although high-magnitude earthquakes that have the power to cause huge destruction are very rare, lower power earthquakes causing minor shakes, barely even noticed by us humans as we go about our daily lives, are very common. According to the Incorporated Research Institute of Seismology (IRIS), there are on average a few hundred earthquakes of less than 2 magnitude on the Richter scale (with about the same power as a lightning bolt hitting the ground) that occur every single day across the globe. Given the level of accuracy and sensitivity required of gravitational wave detectors, they needed to be isolated from this shaking, otherwise all you’ve built is a very expensive earthquake detector.
Similarly, even a heavy truck driving nearby could be enough to shake up the laser and mirror set-up. Burying the detector deep underground fixes the truck problem, but only exacerbates the seismic situation. It was Italian physicist Adalberto Giazotto, working at the University of Pisa, who cracked what to do. He was developing new suspension systems, which he called super-attenuators. He presented his new device at a meeting in Rome in 1985 and pointed out that they would be able to isolate the mirrors from any seismic activity. At the same meeting, a French physicist, Jean-Yves Vinet, who had been working at the Applied Optics Laboratory in Paris (Laboratoire d’Optique Appliquée) presented his work on laser recycling, which allowed you to bump up the power of a laser so that it could still be detected over the large distances needed in gravitational wave detectors.
There was already great interest in Europe in building a gravitational wave interferometer, pioneered by French physicist Alain Brillet, but funding was once again proving to be the biggest barrier. Eventually, both the American and European collaborations were awarded funding (after losing out for many years prior to other projects, such as the Very Large Telescope, VLT, in the Atacama Desert in Chile71). The Caltech–MIT collaboration of Weiss and Thorne was funded by the USA’s National Science Foundation (NSF) in 1988, and was dubbed the Laser Interferometer Gravitational-Wave Observatory (LIGO). The European collaboration of Brillet, Vinet and Giazotto was jointly funded by the French CNRS (Le Centre National de la Recherche Scientifique) in 1993 and in 1994 by the Italian INFN (Istituto Nazionale di Fisica Nucleare), and was dubbed the VIRGO observatory (after the largest nearby galaxy cluster, in the Virgo constellation).
Having solved their funding issues, the problem was now building and actually getting the interferometers working. LIGO suffered immensely in its early years due to disagreements between members on how to both build and manage the project. In 1994, an American experimental physicist called Barry Clark Barish was brought on as director of the collaboration. He was an expert in experimental high-energy physics and crucially had experience managing this new style of big-budget physics project. He redesigned the whole project, deciding that it would be built in two stages: an initial prototype, which could then be improved where needed to increase sensitivity and accuracy in the final stage. Given the intricacies of the interferometer method, this was a very smart move.
Construction of both LIGO and VIRGO prototypes progressed through the 1990s, with problems solved as they reared their heads. VIRGO found a site in Tuscany, Italy. LIGO was to be two separate detectors, one in Livingston, Louisiana and another in Hanford, Washington, both in the USA. Again, this was a smart move; it meant that if the detectors at both sites, separated by around 3,000 kilometres (1,865 miles), reported the exact same detection around 10 milliseconds apart (the travel time of light between the two sites) you could be sure that what you’d detected was a gravitational wave, and not a local disturbance (like a very heavy truck passing overhead). From the delay, you can get a good idea of a direction that the gravitational wave came from in space. Adding a third detector into the mix makes this even more accurate; you can literally triangulate the direction of the gravitational wave. So in 2007, the two separate projects of LIGO and VIRGO joined together to share results and detections.
Despite multiple observing runs through the late 2000s, no detections were made. Updates were needed to improve the sensitivity of the detectors and their isolation from seismic activity. These updates were made through the early 2010s and the detectors weren’t switched on again until September 2015. In the days that followed, the detectors remained in ‘engineering mode’, so that tweaks and calibrations could still be made where necessary. It was during this time that an Italian astrophysicist, Marco Drago, a postdoctoral researcher72 working at the Max Planck Institute for Gravitational Physics in Hanover, Germany, received an email from the automated LIGO system that a detection had been made at both the Livingstone and Hanford detectors.
The two detections were identical and looked the right shape to be a gravitational wave ripple from two black holes, but recorded at slightly different times at each detector with milliseconds difference. It could only be one of two things, either 1) a real gravitational wave, or 2) a fake model signal, artificially ‘injected’ into the system to check that all the procedures of detection were working properly. However, LIGO was still in engineering mode, meaning there was no way to inject fake signals yet. Drago knew this detection had to be real, but asked his colleague, another postdoctoral researcher called Andrew Lundgren, to double check. They called both Livingstone and Hanford to check whether there was anything unusual to report, but there wasn’t. An hour after receiving the first email, Drago then sent an email to the entire LIGO collaboration asking if there was any way a spurious signal could be generated at both detectors, but got no reply. In the following days, senior LIGO members confirmed to the collaboration that there had been no fake signals injected. Within two days of being switched on after its upgrade, LIGO had finally achieved what Weiss and Thorne discussed that night forty years earlier.
This discovery was perhaps the worst kept secret in the history of astronomy. The LIGO collaboration is so large that word eventually got out. I was doing my PhD at the University of Oxford at the time, and it felt like within a few weeks everyone in the astronomy community was abuzz with the news that LIGO had detected something. No one quite knew what had been detected until the news was officially announced at a press conference six months later, in February 2016. The entire collaboration had spent that time confirming that the signal had not been caused by a glitch in the detectors, earthquake or even spurious light sources. The signal, dubbed with the highly poetic name GW150914 (after the date it was detected on) was the first ever direct detection of gravitational waves, and its shape matched the predictions of Einstein’s theory of general relativity for the spiral inwards and merger of a pair of black holes.
Not only was this another win for general relativity, but it was the first time in the history of humankind that we had observed the Universe with something other than light. We could ‘see’ in a whole new way. But it wasn’t the visual graphs of the signal that captured the attention of the public; instead it was the sound produced when the signals were converted to frequencies in the human hearing range that delighted people worldwide. It’s akin to the sound when you close your mouth around your index finger and push your finger into the inner part of your cheek and release it. Pop! It’s perhaps one of my favourite parts of this whole story, that a cheeky pop can represent the most devastating of collisions between two of the Universe’s biggest mysteries.
This discovery of gravitational waves was rewarded in 2017 with a Nobel Prize in Physics. The prize was split between Rainer Weiss, Kip Thorne and the savvy director of the LIGO team, Barry Barish. With the sheer size of the LIGO–VIRGO collaboration, and many other physics experiments worldwide, a prize awarded to just three people doesn’t quite summarise the sheer scale of human effort that went into that one single discovery. There are over 1,200 people working in the LIGO collaboration alone.
This discovery of gravitational waves confirmed the existence of binary black hole systems – where two massive stars had previously lived, died and gone supernova – that astronomers had long suspected but had been unable to detect. It wasn’t long before yet more detections were made, with another popping up in December 2015 before the first was even announced, and a total of fifty detections by October 2020. From black hole binary mergers, to black hole–neutron star mergers, and neutron star binary mergers. The neutron star binary mergers are often the most exciting detections, as we also detect a flash of light from these before their combined mass collapses into a black hole. This can give us an accurate distance to the pair along with a more accurate estimate of the Tolman–Oppenheimer–Volkoff limit for the maximum mass of a neutron star (or the minimum mass of a black hole).
We can’t know just how many doors of discovery this detection will open in the future, but what we do know is that it has completely and irrevocably changed the entire field of astronomy. In the same way that we were previously limited to only what our own eyes could reveal before telescopes gave us a way to view the whole spectrum of light, now gravitational wave detectors have given us a whole new way to see.