Since 2016, Breakthrough Listen has used telescopes around the world to listen for possible broadcasts from alien civilizations. The programme has picked up millions of radio blips of unknown origin, nearly all of which could be swiftly classified as coming from radio interference on Earth, from sources such as mobile-phone towers or aircraft radar.

The 2019 signal was different. It was detected by the 64-metre Parkes Murriyang radio telescope in southeastern Australia and came from the direction of Proxima Centauri—the nearest star to the Sun, just 1.3 parsecs (4.2 light years) away. Proxima Centauri is of intense interest to SETI researchers, not just because it is nearby. The star has at least two planets, one of which orbits at the right distance for liquid water to be present on its surface—a prerequisite for life as it exists on Earth. A sibling initiative to Breakthrough Listen, known as Breakthrough Starshot, aims to send a tiny spacecraft to this planet in the future to look for life there.

The mysterious signal was first spotted last year by Shane Smith, an undergraduate student at Hillsdale College in Michigan, who was working as a research intern with Breakthrough Listen. Smith was combing through data that Parkes collected over six days in April and May the previous year. The telescope had been making observations in the direction of Proxima Centauri for 26 hours. It was not hunting specifically for aliens at the time, but was instead monitoring flares on the star’s surface, which could hurt the chances for life to arise on nearby planets.

The data included more than 4 million signals from the vicinity of the star, but Smith noted one signal near 982 megahertz that seemed to originate from the star itself and lasted about 5 hours. “I was excited to find a signal that matched all the criteria I was looking for, but I immediately remained skeptical of it and thought there had to be some simple explanation,” Smith says. “I did not ever think the signal would cause such excitement.”

Smith shared the information with his supervisor Danny Price, who posted it on a Breakthrough Listen Slack channel, and the team started investigating in earnest. “My first thought was that it must be interference, which I guess is a healthy attitude, to be sceptical,” says Price, an astronomer at UC Berkeley and the Breakthrough Listen project scientist in Australia. “But after a while I started thinking, this is exactly the kind of signal we’re looking for.”

The signal, named BLC1 for “Breakthrough Listen candidate 1”, was the first to pass all of the programme’s initial screening tests to rule out obvious sources of interference. “It definitely had me wondering ‘what if?’ for a bit,” says Sheikh.

She, Price and a large group of colleagues began working through possible explanations, from uncatalogued satellites to transmissions from planetary spacecraft. In Australia, the radio-frequency band around 982 megahertz is primarily reserved for aircraft, but the scientists could not identify any aeroplanes that had been in the area and could account for the signal—and certainly not one lasting 5 hours.

In November 2020, and in January and April of this year, the researchers pointed the Parkes telescope at Proxima Centauri to see if they could pick up the signal again. They could not.

Eventually, the team spotted other signals in the original data that looked a lot like the 982-megahertz signal but were at different frequencies. These signals had been tossed out by the team’s automated analysis as being earthly interference. Further analysis showed that BLC1 and these ‘lookalike’ signals were all interference from an unknown source. The signals had modulated and muddied one another, much as a guitar amplifier modulates and distorts a guitar note, which is what made it so difficult to identify BLC1 as interference.

Because the signal didn’t re-appear in the 2020 and 2021 observations, it might have been coming from malfunctioning electronic equipment that got shut down or fixed, says Sheikh. The team suspects the equipment was relatively close to Parkes, perhaps within a few hundred kilometres. The frequency of the signal drifts in a way that is consistent with inexpensive crystal oscillators such as those commonly used in computers, phones and radios, says Dan Werthimer, a SETI astronomer at UC Berkeley who specializes in signal processing.

Working with another student, Sheikh is now using machine-learning algorithms to tease out what frequency the interfering equipment was transmitting at, which might help to track down its source. One lingering mystery is why the signal seemed to appear only when the telescope was pointed at Proxima Centauri. That might just be an unfortunate coincidence, if the cadence of the interference mimicked the cadence with which the telescope was looking at the star.

Radio interference has bedevilled other astronomical searches before, such as when flickering signals picked up at Parkes turned out to be the result of people microwaving their lunches. The famous ‘Wow!’ signal, detected in 1977 by a radio telescope in Ohio, was a powerful blip so intriguing that the observing scientist scribbled “Wow!” in the margins of the computer printout—but its origin could never be traced.

Alien searches have become much more sophisticated since then, Sheikh notes. “Many groups assumed that if you had a detection that only showed up when you were pointed at the source, that was it, break out the champagne, you’re done,” she says. “As technology changes, the way we vet signals also has to change—and that hadn’t come together until BLC1.” One of the Nature Astronomy papers features a detailed checklist to help astronomers determine whether their signal is truly from aliens or not.

“The Universe gives us a haystack,” says Ravi Kopparapu, a planetary scientist at NASA’s Goddard Space Flight Center in Greenbelt, Maryland. “It is our need to find the needle in it, and make sure that it is actually a needle that we found.”