Anastasios (Andy) Tzanidakis was combing through old telescope data from 2020 when he found an otherwise boring star acting very strangely. The star, named Gaia20ehk, was about 11,000 light-years from Earth near the constellation Pupis . It was a stable “main sequence” star, much like our sun, which meant that it should emit steady, predictable light. Yet this star began to flicker wildly.
“The star’s light output was nice and flat, but starting in 2016 it had these three dips in brightness. And then, right around 2021, it went completely bonkers,” said Tzanidakis, a doctoral candidate in astronomy at the University of Washington. “I can’t emphasize enough that stars like our sun don’t do that. So when we saw this one, we were like ‘Hello, what’s going on here?’”
The cause of the flickering had nothing to do with the star itself: Huge quantities of rocks and dust — seemingly from out of nowhere — were passing in front of the distant star as the material orbited the system, patchily dimming the light that reached Earth. The likely source of all that debris was even more remarkable: a catastrophic collision between two planets.
“It’s incredible that various telescopes caught this impact in real time,” Tzanidakis said. “There are only a few other planetary collisions of any kind on record, and none that bear so many similarities to the impact that created the Earth and moon. If we can observe more moments like this elsewhere in the galaxy, it will teach us lots about the formation of our world.”
The analysis of the star was published March 11 in The Astrophysical Journal Letters.
Planets form when gravity forces together matter — dust, gas, ice or rocky debris, for example — orbiting a new star. Early solar systems are chaotic — planets routinely collide and explode or go flying off into outer space. Through this process, and over perhaps 100 million years, solar systems like ours winnow their planets down and settle into an equilibrium.
As common as these collisions probably are, observing one in a distant solar system requires patience and luck. The orbits of the planets must take them directly between us and their star, so that the resulting debris obscures some of the star’s light. The telltale flicker then takes years to play out.
“Andy's unique work leverages decades of data to find things that are happening slowly — astronomy stories that play out over the course of a decade,” said senior author James Davenport , a UW assistant research professor of astronomy. “Not many researchers are looking for phenomena in this way, which means that all kinds of discoveries are potentially up for grabs.”
Tzanidakis, the study’s lead author, studies extreme variability in stars over time. His previous work at the UW identified a system with a binary star and a large dust cloud that caused a seven-year eclipse .
The behavior of Gaia20ehk, however, posed a new mystery. The star’s particular fluctuation — short dips in brightness and then chaos — had never before been observed. The team was stumped, until Davenport suggested that they use data from a different telescope to look for infrared light rather than visible light.
“The infrared light curve was the complete opposite of the visible light,” Tzanidakis said. “As the visible light began to flicker and dim, the infrared light spiked. Which could mean that the material blocking the star is hot — so hot that it’s glowing in the infrared.”
A cataclysmic collision between planets would certainly produce enough heat to explain the infrared energy. What’s more, the right kind of collision could also explain those initial dips in light.
“That could be caused by the two planets spiraling closer and closer to each other,” Tzanidakis said. “At first, they had a series of grazing impacts, which wouldn’t produce a lot of infrared energy. Then, they had their big catastrophic collision, and the infrared really ramped up.”
There are also clues that the collision resembles the one that created the Earth and moon about four and half billion years ago . The dust cloud is orbiting Gaia20ehk at roughly one astronomical unit, the same distance from the sun to the Earth. At that distance, the material could eventually cool down enough to solidify into something similar to our Earth-moon system. Scientists like Tzanidakis and Davenport can’t know for sure until the dust settles — literally — in the system. That could take a few years, or a few million.
In the meantime, their discovery is a call to action to find more collisions. The powerful Simonyi Survey Telescope at the NSF–DOE Vera C. Rubin Observatory will be well suited to the task when it begins its Legacy Survey of Space and Time later this year; some back-of-the-napkin math by Davenport suggests that Rubin could find 100 new impacts over the next 10 years. That could ultimately help narrow the search for habitable worlds outside our solar system.
“How rare is the event that created the Earth and moon? That question is fundamental to astrobiology,” Davenport said. “It seems like the moon is one of the magical ingredients that makes the Earth a good place for life. It can help shield Earth from some asteroids, it produces ocean tides and weather that allow chemistry and biology to mix globally, and it may even play a role in driving tectonic plate activity. Right now, we don’t know how common these dynamics are. But if we catch more of these collisions, we’ll start to figure it out.”
For more information, contact Tzanidakis at atzanida@uw.edu and Davenport at jrad@uw.edu .
This research was funded by Breakthrough Initiatives.
The Astrophysical Journal Letters
Gaia-GIC-1: An Evolving Catastrophic Planetesimal Collision Candidate
11-Mar-2026