The Exoplanet That Wasn’t There
Exoplanet astronomy—the search for planets orbiting other stars—has two major goals. The first is to describe planets in our galaxy in all their weirdness and wonder. The second is to find planets as close as possible to Earth in size, composition, and position in their star system. Where we run into trouble is that the smaller the planet, the harder it is to find—and the more difficult it is to be confident we’ve detected a planet as opposed to something else.
That’s the case for a planet candidate known as Gliese 581g, described in 2010. It was an exciting discovery. Gliese 581g seemed to be the first rocky exoplanet to orbit its host star at the right distance for liquid water on its surface.
Now the verdict is in: Gliese 581g is an ex-planet, not an exoplanet. A new study, published in the journal Science, showed that what seemed to be the sign of a planet was more likely to be from the star’s “weather”—the same sort of magnetic fluctuations that cause prominences and sunspots on the Sun. Not only that, but a second planet in the same system, Gliese 581d, probably doesn’t exist either, for the same reasons.
Your first response may be to say, “Stupid scientists! How could they get this so wrong?!” (You aren’t a very nice person, you know.) But that’s a mistake. The story of Gliese 581g highlights how hard exoplanet-hunting is, and how science at its best is self-correcting. All results have to be taken as tentative at first; if they survive under scrutiny or (ideally) replication by other researchers, we can trust them. It’s often messy or slow and can lead to ego clashes, but it’s how science works.
The star system Gliese 581 is fairly close to Earth—only about 20.5 light-years away—though the star is too faint to see without a telescope. Astronomers found the presence of several exoplanets in the system. Following the convention, the star itself is known as Gliese 581a, while the first exoplanet is designated “b,” the second “c,” and so forth.
Those planets were found through their gravitational pull on the star. On each orbit, the planets pull the star slightly, which can be detected by the Doppler effect. Obviously this method is most useful for very massive planets orbiting close to the host star. However, once one big planet has been spotted, it can be possible to see the much smaller influence of lighter planets as an added disturbance. That’s like weighing your cat on a bathroom scale: the cat herself is too light to register, but if you pick her up, you can subtract your weight from the total number of both of you together.
But of course, you know your cat exists (especially if, like my cat, she tries to claw you when you pull this sort of stunt), but you don’t know in advance that any exoplanets are orbiting a star. To make things worse, variations in the star’s light can look like the Doppler nudging of an exoplanet, at least at first glance. (Another popular method for planet hunting uses “transits,” or the slight dimming in a star’s light we observe when a planet eclipses it.)
Back to Gliese 581. Astronomers clearly identified planet candidates “b” and “c,” with “d,” “e,” “f,” and “g” following. (At the time, “e” was the lowest-mass exoplanet yet seen.) Gliese 581a is an “M dwarf” or “red dwarf” star, much less massive and hot than the Sun. That means most of these planets orbit closer than Mercury does to the Sun.
Of those, “g” seemed the most exciting. At an estimated mass three times Earth’s, it would have a solid surface. It also orbited in its star’s “habitable zone,” the region where liquid water could conceivably exist on the planet’s surface. Admittedly, it was much bigger than Earth, and much closer to the star than Earth is.
But M dwarfs are messy stars. Their surfaces are a froth of magnetic storms, proportionally more violent than the worst weather on the Sun. Storms on the Sun can knock out satellites or interfere with electrical power generation on Earth; we can only imagine what similar weather might do to a planet orbiting an M dwarf at half the distance of Mercury’s orbit.
Almost from the start, however, some other astronomers questioned if the Gliese 581g actually existed, but the data were ambiguous either way. In the usual weasel language: more observations were needed.
So, when the authors of the Science paper monitored Gliese 581a in a certain type of light emitted by hydrogen, they determined it had natural fluctuations that mimicked the Doppler effect caused by the planet candidate Gliese 581d. That means “d” probably doesn’t exist, but even worse: the detection of planet “g” had depended on that measurement. Without “d,” no “g,” and no rocky planet in the star’s habitable zone. When the astronomers corrected for the star’s weather, the evidence for both planets vanished.
It’s sad in a way to say goodbye to the first rocky exoplanet in a habitable zone. However, since 2010 researchers have found a number of other worlds unplagued by the same sorts of uncertainties. Let Gliese 581g, the exoplanet that wasn’t, be a reminder that science works.