Long ago, around an otherwise unremarkable faraway star, two infant planets had an extraordinarily bad day. The two collided in a giant impact that brought both to a violent end. Where once these worlds had twirled, the cataclysm left behind only a diminished molten lump and a churning 10-million-kilometer-wide cloud of incandescent vapor and pulverized debris that should eventually condense into a new, second-generation planet.

Despite sounding like the climax of a Hollywood space opera, astronomers may have recently witnessed such an apocalyptic event. They detailed their findings on October 11 in Nature.

The saga began in late 2021 when Matthew Kenworthy, an astronomer at the Leiden Observatory in the Netherlands, and a co-lead author of the paper, responded to an alert about the sudden, near-total dimming of a sunlike star some 1,800 light-years from Earth. The initial dimming data and alert alike came from the All-Sky Automated Survey for Supernovae (ASAS-SN) project, a globe-spanning network of 24 small telescopes. Kenworthy was interested in the star, now called ASASSN-21qj, because such extreme dimming events could be caused by giant exoplanetary ring systems—one of his scientific specialties.

In this case, however, something even stranger was in store. After Kenworthy posted about the discovery on social media, Arttu Sainio, an amateur astronomer and an eventual co-author of the study, replied to say that the star had also exhibited a sharp brightening about two and a half years earlier, as seen in public data Sainio had examined from NASA’s infrared space telescope NEOWISE. The revelation stirred Kenworthy because sunlike stars only rarely show such a sudden infrared brightening or strong optical dimming. For one star to display both so close together in time seemed unlikely to be a coincidence. Within days, Kenworthy had pieced together an impact-related explanation and began expanding his search through additional historical and ongoing real-time datasets to shore up the hypothesis.

Using archival infrared observations from NEOWISE, as well as optical data from the Las Cumbres Observatory Global Telescope, a network of 25 modest telescopes, Kenworthy tracked the star’s wavelength-dependent changes in brightness. He found that the infrared flaring corresponded to a heat emission of 1,000 kelvins—hot enough to melt aluminum—and that it was consistent with a source around the star that was some 750 times the size of Earth.

The roughly 900-day delay between the star’s infrared outburst and its subsequent optical dimming (which concluded in late September 2022) strengthened Kenworthy’s conviction that the explanation was a planet-planet impact. Dust from the impact would gradually expand along an orbital path, forming a shroud that drifted across the star’s face as seen from Earth. Another data point in the scenario’s favor was ASASSN-21qj’s age, which Kenworthy and his co-authors estimated to be some 300 million years—young enough for the star to still be in a phase of rowdy, unsettled youth, when giant impacts are more common.

“It’s counterintuitive,” Kenworthy says, explaining why a once-hidden planetary mass of material could suddenly become visible. “You can have a big rock sitting next to a star, and we’ll never see it because it’s got a tiny surface area. If you grind it into sand, the surface area grows tremendously, and we can see that.”

In principle, the collision-causing culprit could have even been a rogue planet from interstellar space that plowed into an unlucky world that orbited ASASSN-21qj. “But that’s a unicorn,” Kenworthy says. “It’s more likely to be a collision of two planets already in the system.”

Although planet-vaporizing collisions may seem like science fiction, for proof that they occur, one need look no further than Earth’s moon, which was likely born from a Mars-sized impactor striking our world billions of years ago. Further afield, one leading theory to account for certain quirks of our solar system’s architecture posits that early shifts in the orbits of Jupiter and Saturn sparked brutal cascades of collisions between nascent protoplanets. More remote and circumstantial evidence abounds from studies of other planetary systems, but until now, astronomers had never seemingly seen one happen before their very eyes.

Working backward, Kenworthy and several of his colleagues used the time difference between ASASSN-21qj’s brightening and dimming to surmise that the initial collision occurred so far from the star that the two planets were likely ice giants akin to our own Uranus and Neptune, bulked up from vast quantities of frozen water and other volatile compounds. This dovetailed with ASASSN-21qj’s long dimming, which, in a giant-impact scenario, would demand a dust cloud from the near-total vaporization of two similarly hefty worlds.

One part of the puzzle was still missing, however. The inferred temperatures didn’t match up because a full head-on collision should have produced a much hotter outburst of 2,000 to 3,000 kelvins. Kenworthy realized that the evidence required a special type of planetary collision that would have led to a strange cosmic doughnut known as a synestia. “If [two planets] hit each other slightly off center, then spin around, and the collision is extreme enough, it smears out into this kind of red blood cell shape of dust called a synestia,” he says.

With a synestia-shaped system—and some water vapor from the icy planets to help the cooling process—an outburst matching the measured 1,000 kelvins was possible.

If it is validated by further observations, the result will mark a first. “When you think of how long it takes to grow a planet, giant impacts are relatively short events,” says Sarah Stewart, a planetary scientist and synestia expert at the University of California, Davis, who was not involved in the study. “You have to be lucky to see one.”

Based on these parameters, Simon Lock, a planetary scientist and at the University of Bristol in England and a co-lead author of the paper, constructed a timeline for the collision. A mere hour postimpact, the synestia took shape. The outpouring of energy from the collision heated up the dust, causing the extra glow of light NEOWISE first observed in 2019. Fourteen hours later, hardly any signs remained of the icy planets, only two molten metal-rich cores. By the third day, the two cores combined into a single white-hot core—the makings of a future planet.

“In millions of years presumably the material will condense down into a new planet,” Kenworthy says. “Some of the stuff may ultimately form moons.”

Some of the dust cloud still gravitationally clings to that core, but the rest swept out in its orbit of ASASSN-21qj until, years later, it extensively eclipsed the star as seen from Earth.

Researchers are cautiously optimistic that this picture will hold up and suggest further studies of the celestial catastrophe will feed into a better overall understanding of how planetary systems form and evolve. “We don’t know how many giant impacts occur, and we don’t know a lot about how the bodies cool and recover from those impacts,” Stewart says. “So seeing one is actually a pretty great perspective.”

But not everyone is onboard. Buried in the pile of less titillating astronomical results, another paper preceded Kenworthy’s: in August Jonathan Marshall, a research scholar at the Academia Sinica Institute of Astronomy and Astrophysics in Taiwan, and his eight co-authors showed in the the Astrophysical Journal how the measurements could instead be explained in terms of disintegrating interstellar comet fragments.

Marshall points out that star-grazing comets are even more common than planet-planet collisions, arguably making them a more likely explanation. Additionally, infrared spectrum measurements hinted that the chemical makeup of the dust more closely matched a comet than a planet. One final point of contention is the age of ASASSN-21qj. The two teams each used a different method to arrive at vastly different ages. Marshall’s approach suggests the star to be about five billion years old, more than 15 times Kenworthy’s estimate. That more advanced age would presumably correspond with a more sedate phase of the planetary system’s existence.

“Whichever age is more accurate, it is interesting to note that this star is relatively old to be undergoing such an event,” Marshall says, noting that theory and observation alike suggest giant impacts are most likely to occur in the tumultuous environs of very young planetary systems. “There’s nothing to say that it isn’t a planet-planet collision—but it’s important to consider all possibilities.”

To decide between the two theories, more data will be needed from more powerful observatories such as NASA’s James Webb Space Telescope or the European Southern Observatory’s ground-based Extremely Large Telescope, which is now under construction in Chile and due to debut toward the end of the decade.

“We always learn something new every time there is a new piece of data or model,” says Kate Su, an astronomer at the University of Arizona, who reviewed the October Nature study but was not directly involved with the work or the August Astrophysical Journal paper. “We thought the planetary arrangement of the solar system was universal until we discovered the first few exoplanets that were so different from ours. We will learn even more from oddballs like ASASSN-21qj.”