The three ‘planets’ were discovered by the Kepler space telescope, but now – years later – they turn out to be no planets at all.

Instead, they’re small stars, researchers write in this week The Astronomical Journal† In the sheet they also point out a fourth planet that may not be a planet at all.

New measurements

The researchers base their conclusion on new measurements of the size of some putative planets. Those measurements indicate that three of the four planets are too large to be classed as planets and should instead be labeled as small stars. The fourth specimen is on the border between planet and star in terms of mass and is given the benefit of the doubt for the time being. But it is quite possible that it will also lose its planet status in the future.

Suspicious objects

“Most exoplanets are about the size of Jupiter or much smaller,” explains researcher Prajwal Niraula. “Twice the size of Jupiter is already suspect.” If an object is much larger than that, in most cases it cannot be a planet. And that’s what researchers have now established for three objects. In concrete terms, we are therefore three exoplanets poorer: Kepler-854b, Kepler-840b and Kepler-699b will in any case go up in smoke as a planet. And then rise from the ashes like a small star.

Kepler-747b meanwhile keeps the stamp ‘planet’, but gets the stamp ‘suspicious’ with it. New measurements indicate that this planet is about 1.8 times larger than our own Jupiter. This allows it to compete with the largest planets whose existence has now been conclusively established. It is therefore not impossible that Kepler-747b really is a planet, but it is also not obvious, given the great distance from the parent star. For the same money, it’s just a small star. But that is not hard to do at the moment.


Funnily enough, the astronomers who are now stripping three Kepler planets of their planetary status weren’t out to fish for false positives from the Kepler space telescope’s rich database. Instead, they were actually looking for systems with clear signs of tidal deformation. “If you have two objects that are close together, the gravitational pull of one will cause the other to become ovoid or ellipsoidal,” Niraula explains. For example, a parent star can use its gravitational pull to give a nearby planet rugby ball-like proportions. And the degree of deformation can tell more about both the mass of the planet and that of the parent star.

On the hunt for distorted planets in the Kepler database, however, the scientists stumbled upon something strange in the Kepler-854b system. “Suddenly we had a system where we saw a huge ellipsoidal signal and immediately knew it couldn’t come from a planet,” said study researcher Avi Shporer. “Then we thought: this is not right.”

Zoom in

The researchers therefore decided to zoom in more closely on both the star and the alleged planet. Kepler had inferred the existence of this alleged planet in 2016 – as always – from dips in the brightness of the parent star (see box).

Kepler space telescope detected planets using the transit method. The telescope stared at stars for a long time, hoping to witness their brightness decreasing with some regularity. Indeed, such a regular decrease or dip in the brightness of the star’s light may indicate the presence of a planet, which – as it orbits the star – occasionally interposes between Kepler and the parent star, creating a ( small) part of the star light.

However, the dip in starlight can’t just reveal a planet orbiting a star. It can also provide more insight into the size of that planet. For example, the rate at which the star’s brightness decreases tells more about how the planet’s size compares to the size of its parent star. And so if you know how big the star is, you can also see how big the planet should be from the degree to which the brightness of that star decreases. In 2016, researchers estimated the size of Kepler-854b based on the observed dip and an estimate of the size of the parent star. And they determined that the object must be a planet.


But now we are more than 5 years later and researchers are much better able to determine the size of a star – and therefore also the size of the planet that orbits it. It’s all thanks to Gaia, an observatory that has been mapping the properties of millions of stars in the Milky Way since 2013. In 2016, Gaia had not yet collected data on Kepler-854 and, based on the information available, researchers estimated the size of Kepler-854 and subsequently Kepler-854b. But now it is 2022 and, thanks to Gaia, researchers have much more accurate measurements of Kepler-854. And researchers have now caught them. And that has far-reaching implications for Kepler-854b. Based on the new stellar measurements, the supposed planet must be no less than three times the size of Jupiter. “It’s impossible for the universe to make a planet of that size,” Shporer said. “That just doesn’t exist.”

Impossible Planets
Shporer is very clear above: a planet three times larger than Jupiter does not exist. Are there really no known larger planets? We checked it out and soon came across HD 100546b, which takes the cake with an estimated size of about 7 times the radius of Jupiter. How can the existence of this huge planet be explained? We asked Shporer. “We can determine the size (radius) of planets in two ways,” he explains first. “Through empirical evidence and theoretical understanding. Empirical evidence includes, for example, the hundreds of planets that we have seen moving in front of their stars and whose masses we were able to measure directly. Theoretical concept, on the other hand, is an understanding of the physical processes that determine the radius of the planet, such as the gravity that makes the planet shrink and the pressure that particles in the planet exert on each other and cause the planet to expand. Based on the latter, we know that planets can reach a maximum of twice the radius of Jupiter. There may be some strange exceptions: planets that are getting bigger, for example because they are very young and so are in fact still in their birth process and therefore still both collecting gas and contracting. But those planets are not only very rare, they are also very difficult to measure accurately.” Wherever it is difficult to measure is the planet HD 100546b, which we just mentioned and which would therefore be almost seven times larger than Jupiter. “This is a planet that has been directly observed, so the radius has been estimated using theoretical models, looking in particular at the planet’s brightness. Therefore, no direct measurements were performed. The margin of error is therefore large; nearly 3 times the radius of Jupiter. In addition, it is suspected that the estimated radius is formed by both the planet itself and the gas disk it is still accumulating. That makes it extra difficult to estimate the radius of HD 100546b.” So the planet could be an exception that proves the rule, but it could just as well be much smaller than we think.

A few more planets unmasked

The unmasking of Kepler-854b naturally raised the question of whether there were more false positives hidden in the Kepler data. The researchers took a closer look at another 2,000 planets discovered by Kepler. Using the Gaia data, they determined the size of each of these planets, based on much more accurate measurements of their parent star. And so a few more planets were found whose size was so underestimated that they could not possibly be planets. “We now have three objects that are not planets and the fourth is probably not a planet,” Niraula concluded.

Better understanding

According to the researchers, it is unlikely that many more planets will die in a similar way in the future. They point out that the corrections come from a better understanding of the parent stars. That understanding is constantly improving, but with Gaia a very big step has been taken. It doesn’t mean we know everything now, but it does mean that future corrections are likely to be much smaller and won’t directly lead to planet collapse.

That we now have to say goodbye to some planets may seem sour. But we have to see that differently, says Shporer. “Actually, our study makes the current list of planets more complete. People rely on that list when they study the planet’s population as a whole. If such a list contains some intruders, your results may be incorrect. So it is important that the list is correct.”