The Millisecond Flash That Crosses the Universe

Somewhere in the sky, a flash of radio energy lasts less than a thousandth of a second. In that sliver of time it pours out roughly as much energy as the Sun radiates across three full days. By the time the pulse sweeps across Earth it has been traveling for billions of years, smeared and stretched by everything it passed through. A radio dish records it. And then it is gone, usually never to repeat, leaving astronomers to argue over a question that took sixteen years to answer: what could possibly produce such a thing?

The burst that should not have been found

In 2007, Duncan Lorimer, a radio astronomer at West Virginia University, handed a student named David Narkevic a routine assignment. Go through six-year-old survey data from the Parkes radio telescope in Australia, recorded in 2001, and look for anything unusual. Buried in those archives was a single spike so bright and so brief that it had been overlooked entirely.

The pulse lasted under five milliseconds. It carried a peak flux of about 30 janskys, an enormous figure for a fraction of a second of radio noise. Most striking was its dispersion measure: roughly 375 in the standard units astronomers use, far higher than anything our own galaxy could explain. Lorimer and his colleagues published the find in Science under a deliberately careful title, "A bright millisecond radio burst of extragalactic origin." The word that mattered was extragalactic. Whatever produced this had done so hundreds of millions of light-years away.

The reaction was skepticism. A single event, found in archival data, sitting near a busy frequency band, looked suspiciously like interference. For years the Lorimer Burst stood almost alone, an anomaly that might be an instrument artifact, a lightning strike near the dish, or a microwave oven door opening in the observatory kitchen. That last possibility was not a joke. Parkes really did record a class of false signals later traced to its own staff break room.

A single flash, six years old, found by a student in a data archive. It would take a decade of new instruments to prove it was real, and another to prove what made it.

How a flash becomes a ruler

The reason astronomers could say anything at all about a millisecond pulse comes down to a quirk of physics called dispersion. Radio waves of different frequencies do not travel at quite the same speed when they pass through the thin plasma of intergalactic space. Higher frequencies arrive first, lower frequencies trail behind, and the longer the journey, the wider the gap between them.

That delay is measurable. It is called the dispersion measure, and it counts the total number of free electrons the signal crossed on its way to Earth. A burst from inside the Milky Way carries a small dispersion measure. A burst from across the cosmos carries a large one. The Lorimer Burst's value was far too high to come from our galaxy alone, which is precisely why the team called it extragalactic.

This turned a nuisance into a tool. The dispersion measure became a rough cosmic ruler, a way to estimate how far a burst had traveled before anyone could see its host galaxy. It also turned fast radio bursts, or FRBs, into something more than a curiosity. If they came from across the universe in their thousands, they could be used to weigh the diffuse matter scattered between galaxies, the so-called missing baryons that ordinary surveys struggle to detect. A flash too brief to understand had become a probe of the cosmos itself.

From a single burst to a flood

For most of a decade, fast radio bursts trickled in one or two at a time. The bottleneck was the telescopes. A dish like Parkes sees only a narrow patch of sky, and FRBs do not announce themselves in advance. Catching them was a matter of luck.

That changed with CHIME, the Canadian Hydrogen Intensity Mapping Experiment. CHIME has no moving parts. It is a set of four stationary half-pipe reflectors in British Columbia that stare straight up and let the rotation of the Earth sweep the entire overhead sky past them every day. Built to map hydrogen across the universe, it turned out to be a nearly perfect fast radio burst machine, watching a huge swath of sky at the low radio frequencies between 400 and 800 megahertz.

The numbers tell the story. The first CHIME/FRB catalog, covering July 2018 to July 2019, reported 536 bursts, including 62 from eighteen sources that had been seen to repeat. The second catalog, released in early 2026 and covering observations through September 2023, reported 4,539 bursts from 3,641 distinct sources, with 981 of those bursts coming from 83 known repeaters. In a few years the field went from a handful of disputed events to a population large enough to do statistics on.

CHIME does not chase fast radio bursts. It lies on its back, lets the planet turn, and catches whatever the sky throws past.

Repeaters and one-offs

As the catalog grew, a division appeared. Most fast radio bursts flash once and are never heard from again. A minority repeat, sometimes erratically, sometimes in clustered storms of activity separated by long silences. The first repeating source, FRB 121102, fired again and again from a dwarf galaxy roughly three billion light-years away, which allowed astronomers to pin down its position and study its surroundings in a way a one-off burst never permits.

The split raised an obvious question. Are repeaters and non-repeaters two different kinds of object, or the same kind caught in different moods? A source that repeats rarely could easily masquerade as a one-off if no one happened to be watching during its quiet years. The statistical properties of the two groups differ, with repeaters tending toward longer, narrower-band pulses, but no one has cleanly separated them into distinct physical families. It remains one of the field's genuinely open questions.

What both populations share is the energy problem. To produce a burst bright enough to cross billions of light-years and still register on a radio dish, the source has to release a staggering amount of power in a coherent flash, far more than a normal star or even an ordinary neutron star could manage. Whatever made these signals had to be both compact and violently energetic. The list of candidates kept circling back to one kind of object.

The night the answer came from home

On April 28, 2020, a fast radio burst arrived that broke the entire pattern. Its dispersion measure was small. It had not crossed the cosmos. It came from inside the Milky Way, from a known object catalogued as SGR 1935+2154, a magnetar sitting a few tens of thousands of light-years away.

A magnetar is a neutron star with an absurd magnetic field. SGR 1935+2154 spins once every 3.24 seconds and carries a surface field strength around 2.2 times ten to the fourteenth gauss, hundreds of millions of times stronger than the most powerful magnets ever built on Earth. These objects are already known for erratic flares of X-rays and gamma rays. What no one had caught before was one producing a fast radio burst.

Two instruments saw it at once. CHIME recorded the event, designated FRB 200428, as two roughly millisecond-long components separated by about 30 milliseconds. A second, smaller experiment in California called STARE2, the Survey for Transient Astronomical Radio Emission 2, caught the same burst at higher frequencies and measured a tremendous fluence of around 1.5 million jansky-milliseconds, reported by Bochenek and colleagues. Crucially, a fleet of space telescopes recorded a one-second X-ray burst from exactly the same magnetar at exactly the same moment. The radio flash and the X-ray flare were the same event.

It was not as bright as a true extragalactic FRB. Scaled to the distances of cosmic bursts, FRB 200428 would have been on the faint end of the population, perhaps an order of magnitude or two below the typical event. But it closed a gap that had been open since 2007. A magnetar, a known and well-studied kind of object, had just produced something that looked and behaved like a scaled-down fast radio burst. The prime suspect finally had a face.

For thirteen years the source was a category of theory. On one night in April, it was a specific star with a name and an address.

Reading the magnetic field directly

One detection, however convincing, does not settle a science. It showed that magnetars can make bursts like these. It did not prove that most fast radio bursts come from magnetars, nor did it reveal where on or around the star the radio light is actually generated. Those questions moved to the next generation of observations, and the most decisive came from the polarization of the light itself.

In 2024, two companion papers in Nature dissected a single nearby burst called FRB 20221022A, localized to a galaxy about 65 million light-years away. McKinven and colleagues measured how the angle of the burst's polarization rotated as it played out. Over roughly two and a half milliseconds, the polarization angle swung by about 130 degrees, tracing an S-shaped curve. That shape is a signature familiar from pulsars, where it maps the geometry of a rotating magnetic field sweeping past our line of sight.

In the companion study, Nimmo and colleagues used the way the burst twinkled, a scintillation effect caused by intervening plasma, to estimate the size of the region that emitted it. The answer was small, no larger than about ten thousand kilometers across, and likely much smaller. A region that compact, combined with the pulsar-like polarization swing, pointed to a single conclusion: the radio emission was generated inside the magnetosphere of a rotating neutron star, in the violently magnetized region hugging its surface, rather than in a distant shock wave far from the star.

The picture has continued to sharpen. Detailed multi-frequency campaigns on the hyperactive repeater FRB 20201124A, including a study with India's upgraded Giant Metrewave Radio Telescope published in late 2025, found bursts with high circular and linear polarization and complex behavior across frequencies, consistent with emission from multiple active zones inside a magnetar's magnetosphere embedded in a turbulent plasma. The evidence keeps accumulating in the same direction.

What is settled and what is not

The core claim is now firm. Magnetars can produce fast radio bursts, and at least some bursts are generated close to the neutron star, within its magnetosphere. The 2020 galactic event proved the mechanism is possible, and the polarization work proved the emission region is small and magnetically structured. That is a genuine resolution to the central mystery of where these signals are born.

What remains open is whether every fast radio burst follows the same script. The universe produces thousands of these flashes every day, with an enormous spread in brightness, repetition, and environment. Some may come from young magnetars freshly born in supernovae. Others may arise from magnetars in older, quieter settings, or from interactions in binary systems, or by mechanisms that have not yet been written down. The split between repeaters and one-offs is still unexplained. And no one has yet caught a true extragalactic FRB with the same multi-messenger clarity as the galactic one, where a radio flash and an X-ray flare were timed to the same instant from the same star.

So the millisecond flash that crosses the universe is no longer a signal from nowhere. It has a plausible engine, a magnetized stellar corpse a dozen miles wide, and a stage on which the burst is struck, the twisted field lines just above its surface. The mystery has narrowed from "what on Earth could do this" to "which kinds of magnetar, and exactly how." That is what progress looks like in astronomy: a question that once had no answer now has a face, even if it does not yet have a complete biography.

It still lasts less than a thousandth of a second. It still carries a Sun's worth of energy across the dark. But the flash from nowhere now has a name, a kind of star, and a place on that star where the light is struck.