The Parkes telescope in Australia detected the first fast radio burst in 2001. Credit: Wayne England

No astronomer had ever seen anything like it. No theorist had predicted it. Yet there it was — a 5-millisecond radio burst that had arrived on 24 August 2001 from an unknown source seemingly billions of light years away.

“It was so bright, we couldn't just dismiss it,” says Duncan Lorimer, who co-discovered the signal1 in 2007 while working on archived data from the Parkes radio telescope in New South Wales, Australia. “But we didn't really know what to do with it.”

Such fleeting radio bursts usually came from pulsars — furiously rotating neutron stars whose radiation sweeps by Earth with the regularity of a lighthouse beam. But Lorimer, an astrophysicist at West Virginia University in Morgantown, saw this object erupt only once, and with more power than any known pulsar.

Fresh confusion over origins of enigmatic radio-wave blasts

He began to realize the significance of the discovery1 only after carefully going over the data with his former adviser, Matthew Bailes, an astrophysicist at Swinburne University of Technology in Melbourne, Australia. If the source really was as far away as it seemed, then for a few milliseconds it had flared with the power of 500 million Suns. “We became convinced it was something quite remarkable,” he says.

But when no more bursts appeared, initial excitement turned to doubt. Radio astronomers have learnt to be sceptical of mysterious spikes in their detectors: the events can all too easily result from mobile-phone signals, stray radar probes, strange weather phenomena and instrumental glitches. Wider acceptance of what is now known as the Lorimer burst came only in the past few years, after observers working at Parkes and other telescopes spotted similar signals. Today, the 2001 event is recognized as the first in a new and exceedingly peculiar class of sources known as fast radio bursts (FRBs) — one of the most perplexing mysteries in astronomy.

Whatever these objects are, recent observations suggest that they are common, with one flashing in the sky as often as every 10 seconds2. Yet they still defy explanation. Theorists have proposed sources such as evaporating black holes, colliding neutron stars and enormous magnetic eruptions. But even the best model fails to account for all the observations, says Edo Berger, an astronomer at Harvard University in Cambridge, Massachusetts, who describes the situation as “a lot of swirling confusion”.

‘Half-pipe’ telescope will probe dark energy in teen Universe

Clarity may come soon, however. Telescopes around the world are being adapted to look for the mysterious bursts. One of them, the Canadian Hydrogen Intensity Mapping Experiment (CHIME) near Penticton in British Columbia, should see as many as a dozen FRBs per day when it comes online by the end of 2017.

“This area is set to explode,” says Bailes.

Curiouser and curiouser

Astronomers might have had more confidence in the Lorimer burst initially had it not been for a discovery in 2010 by Sarah Burke-Spolaor, who was then finishing her astrophysics PhD at Swinburne. Burke-Spolaor, now an astronomer at the US National Radio Astronomy Observatory in Socorro, New Mexico, was trawling through old Parkes data in search of more bursts when she turned up 16 signals that shook everyone's confidence in the original3.

In most ways, these signals looked remarkably similar to the Lorimer event. They, too, showed 'dispersion', meaning that high-frequency waves appeared in the detectors a few hundred milliseconds before the low-frequency ones. This dispersion effect was the most important piece of evidence convincing Lorimer and Bailes that the original burst came from well beyond our Galaxy. Interstellar electrons in clouds of ionized gas are known to interact more with low-frequency waves than with high-frequency ones, which delays the low-frequency waves' arrival at Earth ever so slightly, and stretches the signal (see 'Flight delays'). The delay in the Lorimer burst was so extensive that the wave had to have travelled through a lot of matter — much more than is in our Galaxy.

Credit: Nik Spencer/<i>Nature</i>; Source: Fig. 1 In Keane, E. F. <i>et al</i>. <i>Nature</i> <b>530</b>, 453–456 (2016).

Unfortunately for Lorimer and Bailes' peace of mind, Burke-Spolaor's signals also showed a crucial difference from the original: they seemed to pour in from everywhere, not just from where the telescope was pointing. Dubbed perytons, after a mythical winged creature that casts a human shadow, these bursts could have been caused by lightning, or some human-made source. But they were not extraterrestrial.

Mysterious radiowave blast may have come from starquake

Lorimer decided to postpone his research into FRBs for a while. “I didn't yet have tenure,” he says, “so I had to go back and do more mainstream projects, just to keep my research moving.” Bailes and his team kept going, and upgraded the Parkes detector's time and frequency resolution. In 2013, they turned up four new FRB candidates that resembled the Lorimer burst4. But some outsiders remained sceptical that the signals were really coming from space — not least because all the FRBs thus far had been seen by one team using one telescope. “I was desperate for someone else to find them somewhere else,” says Bailes.

In 2014, his wish was finally granted. A team led by astronomer Laura Spitler at the Max Planck Institute for Radio Astronomy in Bonn, Germany, published their observations of a burst at the Arecibo Observatory in Puerto Rico5. “I was ridiculously overjoyed,” says Bailes.

The Arecibo discovery convinced most people that FRBs were the real deal, says Emily Petroff, who is now an astrophysicist at the Netherlands Institute for Radio Astronomy in Dwingeloo. Yet, as long as the Burke-Spolaor signals went unexplained, they cast a shadow of doubt. “At any talks I would give,” says Petroff, “someone would say, 'But what about perytons?'” So in 2015, while still a graduate student at Swinburne, she led a hunt to track down the source of perytons once and for all.

Sign up - it's free!

First, Petroff and her team used the upgraded Parkes detector to pinpoint when the bursts were happening: at lunchtime. “Immediately I thought, 'This isn't weather',” says Petroff. Then came another peryton at a suspiciously familiar radio frequency, which led the team to run experiments in the staff kitchen. Perytons, they discovered, were the result of scientists opening the microwave oven mid-flow. But the Lorimer event was in the clear: records showed that at the time of the burst, the telescope had been pointed in a direction that would have blocked any microwave signal from the kitchen6.

“So then I worried, maybe they've just got a different brand of microwave at Arecibo,” says Bailes, whose team at Parkes had, by then, racked up 14 separate bursts. He did not relax completely until later in 2015, when a burst was spotted at a third facility — the Green Bank Telescope in West Virginia. That burst had another quality that supported an extraterrestrial origin: its waves were rotated in a spiral pattern — which results from passing through a magnetic field — and were scattered as if they had emerged from a dense medium. “There's no way that's a microwave oven,” Bailes told himself.

Bursts of inspiration

But that still leaves the question of what the FRBs actually are. The extreme brevity of the signal, just 5 milliseconds, implied that the source must be a compact object no more than a few hundred kilometres across — a stellar-mass black hole, perhaps, or a neutron star, the compact core left over by a supernova. And the fact that Earth-based telescopes can detect the FRBs at all means that this compact source somehow puts out an immense amount of energy. But that still leaves a long list of candidates, from merging black holes to flares on magnetars: rare neutron stars with fields hundreds of millions of billions of times stronger than the Sun's.

There's no way that's a microwave oven.

An important clue arrived earlier this year when Spitler's team reported that at least one FRB source repeats: data from Arecibo revealed a flurry of bursts over two months, some spaced just minutes apart7. That behaviour has been confirmed by the Green Bank telescope, which detects signals in a different frequency band8. Until then, each of the observed FRBs had been a one-off event, which hinted at cataclysmic explosions, or collisions in which the sources were destroyed. But a repeating FRB implies the existence of a source that survives the pulse event, says Petroff. And for that reason, she says, “I would assume it would be something to do with a neutron star” — one of the few known objects that can emit a pulse without self-destructing.

Bizarre star could host a neutron star in its core

Spitler agrees. As an example, she points to the Crab nebula: the result of a supernova explosion that was observed from Earth in 1054 and left behind a rapidly spinning pulsar surrounded by glowing gas. The Crab pulsar occasionally releases extremely bright and narrow radio flares, Spitler says. And if this nebula were in a distant galaxy and hugely boosted in energy, its emissions would look like FRBs.

If one source repeats, Spitler says, the simplest interpretation is that they all do, but that other telescopes haven't been sensitive enough — or lucky enough — to see the additional signals. Yet others think that perhaps only some are repeating. “I wouldn't be surprised if we end up with two or three populations,” says Petroff.

A long way home

Another crucial question is how far away the FRBs are. The 20 bursts seen so far seem to be scattered randomly around the sky rather than being concentrated in the plane of the Galaxy, which suggests that their sources lie beyond the borders of the Milky Way.

And yet to Avi Loeb, a physicist at Harvard University, such vast distances imply an implausibly large energy output. “If you want the burst to repeat, you won't be able to destroy the source — therefore, it cannot release too much energy,” he says. “That puts a limit on how far away you can put it.” Perhaps, he says, the FRB sources are neutron stars in our own Galaxy, and the dispersion is mostly the result of still unknown electron clouds that blanket them.

But others suggest that such a dense cloud in the Galaxy should be visible in other wavelengths. At the California Institute of Technology (Caltech) in Pasadena, astrophysicist Shri Kulkarni has scoured data from several telescopes for a galactic source and turned up nothing9. Kulkarni, who directs Caltech's optical observatories, initially argued for galactic FRBs, and even made a US$1,000 bet on it with astronomer Paul Groot of Radboud University Nijmegen in the Netherlands. Now, he finds the evidence for extragalactic FRBs to be overwhelming, and has agreed to settle the bet — grudgingly. “I think I will pay him in $1 bills,” he says.

The black-hole collision that reshaped physics

Still, Kulkarni hasn't ruled out the possibility that the FRB sources lie in galaxies that are perhaps a billion light years away, rather than many billions. Such a distance would still require at least some of the signal dispersion to come from electron clouds in the host galaxy, he says. But closer FRBs would not have to be so energetic. “It takes them from being amazingly exotic, to just exotic,” he says.

The answer could mean a great deal to observers. If the FRB signals have travelled through local plasma clouds, they could give weather reports from neighbouring galaxies. But if they are truly cosmological — coming from halfway across the visible Universe — they could solve a long-standing cosmic mystery.

For decades, astronomers have known from observations of the early Universe that the cosmos should contain more everyday matter — the kind made up of electrons, protons and neutrons — than exists in the visible stars and galaxies. They suspect that it lies in the cold intergalactic medium, where it is effectively invisible. But now, for the first time, the dispersion of the FRB signals could enable them to measure the medium's density in any given direction. “Then, we have essentially a surgical device to do intergalactic tomography,” says Kulkarni.

Rapid-fire detection

First, however, astronomers have to find a lot more FRBs and pin down their locations. “Until then, we are stumbling in the dark,” says Berger.

Mysterious radio burst pinpointed in distant galaxy

One way to accomplish that is to extract the FRBs from radio-telescope data in real time, so that scientists at other observatories can observe the bursts in multiple wavelengths. Since last year, the Parkes team has been doing this by boosting the observatory's in-house computing power, and scientists at Arecibo hope to follow suit this year. In February, the strategy seemed to be paying off when an independent team followed up within two hours of an FRB's detection at Parkes. The team tentatively pinpointed the burst to a specific galaxy almost 6 billion light years away. Further observations cast doubt on that interpretation. But even so, says Lorimer, the method is sound and may pay off in the future.

Others observers are putting their hopes in new telescopes. In 2014, astrophysicist Victoria Kaspi at McGill University in Montreal, Canada, submitted a proposal to adapt CHIME, which was originally designed to map the expansion of the Universe in its early years. “It became clear very quickly that it would be a fantastic FRB instrument,” says Kaspi. Although dish telescopes such as Arecibo can be highly sensitive, they observe only a single, tiny patch of sky at a time. CHIME, by contrast, consists of four 100-metre-long half-pipes dotted with antennas that can monitor much bigger stretches of sky in long lines. After undergoing testing and debugging, CHIME should see its first FRBs sometime next year, says Kaspi, ultimately finding more than a dozen per day.

In Hoskinstown, Australia, meanwhile, Bailes and his colleagues are refurbishing the 1960s-vintage Molonglo Observatory Synthesis Telescope, turning it into an FRB observatory with a single half-pipe 16 times longer than CHIME's, although one-quarter as wide. The team has already found three as-yet-unpublished FRBs with the facility working at only about 20% of its final sensitivity, says Bailes.

The origins of space and time

Another strategy for locating the FRB sources is to work with existing facilities such as the Very Large Array: an 'interferometer' that uses the time difference between signals from 27 radio telescopes spaced across 36 kilometres of grassland near Socorro, New Mexico, to create a single, high-resolution image. Sometime in the next year or so, says Lorimer, the array could detect an FRB and locate its home galaxy. “Ultimately, that could settle a lot of arguments and bets,” he says.

Kulkarni, meanwhile, is leading two projects. The first uses ten 5-metre-wide dishes in an array that can see and locate only super-bright FRBs, but that makes up for its low sensitivity by peering at a huge swathe of sky. The second takes the principle to the extreme, using 2 antennas spaced at observatories 450 kilometres apart that will see only the very brightest FRBs, but that are able to examine half the sky at once. That would enable it to catch the rare FRBs that presumably exist within our own Galaxy, but whose extreme brightness existing telescopes are not designed to see. “Most facilities would just discount it as interference,” says Kulkarni.

If FRBs do turn out to come from cosmological distances, says Loeb, their identification would be a major breakthrough, potentially unravelling a new class of source that could be used to probe the Universe's missing matter. But then, he says, FRBs could also be something that no one has thought of yet: “Nature is much more imaginative than we are.”