One of the Most Earthlike Planets Ever Found May Not Exist

All sorts of excitement accompanied astronomers' discovery of Gliese 581g in 2010—the alien world looked like Earth in both size and temperature, and thus seemed potentially hospitable to life.

But according to a paper published Thursday in the journal Science, that excitement was misplaced. (Related: "First Truly Habitable Planet Discovered, Experts Say.")

"Gliese 581g doesn't exist," said lead author Paul Robertson of Penn State University in State College, Pennsylvania. Neither, he said, does another planet in the same solar system, known as Gliese 581d, announced in 2009—less clearly hospitable to life, but still once seen by some astronomers as a possible place to find aliens.

Initial Discovery

The original evidence for both worlds' existence came from measurements of its home star, Gliese 581—a dim red dwarf, about a third as massive as the sun, that resides about 22 light-years away from our solar system. (Related: "Land on 'Goldilocks' Planet for Sale on Ebay.")

Most exoplanets are too close to their stars to be seen directly with telescopes, so astronomers find them with indirect clues. In the case of Gliese 581g, they watched for subtle wobbles caused by the gravity of an orbiting planet tugging back and forth on the star in a regular pattern.

That's what Paul Butler of the Carnegie Institution for Science, in Washington, D.C., and Steven Vogt of the University of California, Santa Cruz, thought they'd observed when they announced the discovery of Gliese 581g.

The time it took the "planet" to complete one orbit (37 days) told them how far it was from the star. In the case of this cool star, that was "just at the right distance to have liquid water on its surface," Butler said at the time. The strength of the tugging, meanwhile, told them the planet was about three times as massive as Earth.

Rising Doubt

But even at the time, other astronomers questioned whether Gliese 581g was really there. A star's wobbles are measured by looking at its spectrum—its light, smeared out to form a sort of rainbow. The wobbles are so tiny, however, that it takes some statistical analysis to find a back-and-forth pattern.

Critics such as exoplanet expert Eric Ford, then at the University of Florida and now at Penn State, said that Butler's and Vogt's analysis was unconvincing, arguing that the pattern wasn't even clearly there.

Robertson and his colleagues, however, did find a pattern: "There is a real, physical signal," he said. The bad news: "It's just that it's coming from the star itself, not from the gravity of planets d and g."

Starry Eyes

What's happening, they say, is that magnetic disturbances on Gliese 581's surface—starspots—are altering the star's spectrum in such a way that it mimics the motion induced by a planet.

The star itself rotates once every 130 days, carrying the starspots with it; the disputed planets appeared to have periods of almost exactly one half and one fourth of the 130-day period. When the scientists corrected for the starspot signal, both planets disappeared.

"This analysis demonstrates pretty convincingly that these signatures are more due to stellar activity than to the existence of planets," said Ford, who wasn't involved in this research. (Butler declined to comment on the new result, and Vogt did not respond to an emailed request for comment.)

Silver Lining

At the same time, subtracting the starspot signal actually made the evidence for three other worlds in the Gliese 581 system—planets b, c, and e, all of which are too hot to be habitable—even stronger.

"It's unfortunate that the other planets don't exist," said co-author Suvrath Mahadevan, also at Penn State. "But the important takeaway is that stellar activity is an important source of contamination, and that we can [now] take it into account."

It's very encouraging, agreed Robertson, that "we can now take out the stellar influence and reveal the planet's existence."

Follow Michael D. Lemonick on Twitter.