A Tatooine Family

Astronomers have discovered two exoplanets orbiting two stars, both answering and raising questions about how planets form.

The vast majority of stars in the Milky Way don’t go it alone — most have a companion, or even two or three. It stands to reason that life outside the solar system might have a sky less like our own and more like Tatooine, the Star Wars planet with a double sunset.

A Tatooine family
An artist's illustration shows a possible view of the Kepler 47 system, where two planets — one super-Earth and one Uranus-sized planet — orbit two stars — one Sun-like and the other a red dwarf.
NASA / JPL-Caltech / T. Pyle
Many binary systems host exoplanets, but astronomers still don’t have a good understanding of how planets could form in stable orbits around multiple stars. The Kepler spacecraft has been looking out for such planets, watching for the dip in brightness as a planet crosses in front of one of its stars. Kepler has already found four circumbinary planets, each one alone in its orbit around two stars.

Now Kepler has another so-called Tatooine planet to add to the growing list, but this one’s less lonely. Observations reported by Jerome Orosz (San Diego State University) and his colleagues in Science this week and announced at the International Astronomical Union meeting in Beijing, China show not just one but two planets in stable orbits around a binary star system called Kepler-47.

The inner planet, possibly a rocky one about three times the size of Earth, whizzes around its host stars roughly every 50 days. The outer planet is probably a gas giant slightly larger than Uranus and takes just over 300 days to complete an orbit. Ironically, it’s the gas giant, not the super-Earth, that lies in the stars’ “habitable zone,” the region where an Earth-like planet could have liquid water on its surface.

The larger, brighter star in the system is similar to the Sun in mass, girth, and temperature; its companion is 176 times fainter and one-third the size. Kepler detected both planets while monitoring light from the brighter, Sun-like star. Remember the transit of Venus? For several hours on June 5-6th, Venus’s black disk blocked 0.1% of the Sun’s light. Kepler sees something similar as the planets block 0.08% and 0.2% of the starlight, respectively.

Kepler has been watching long enough to capture 18 transits of the inner planet and 3 transits of the outer, making the detection secure even though the Sun-like star is plagued by starspots. Orosz and his colleagues analyzed the observations and determined that both planets likely orbit both stars, not just the bigger one.

Kepler-16 system showing the two stars and planet b.
An artist's conception of the first Tatooine-like planet discovered by Kepler, known as Kepler-16b. The planet orbits a close binary pair - an orange and a red dwarf.
NASA/JPL-Caltech/R. Hurt (SSC)
The planetary duo are only the 5th and 6th planets discovered via the transit method to orbit both stars in a binary system, and the Kepler-47 system is the first observed to host more than one planet in this configuration. With a growing sample of circumbinary planets, astronomers can start narrowing down how planets could form in such exotic systems.

In the standard picture of planet formation, dust grains only a few microns across fly thick and furious around a newly formed star, smacking into one another again and again until they start to stick together. Eventually, enough grains glom together to form so-called planetesimals a few miles wide, and those planetesimals collide to form still-larger planet-sized objects.

But in binary systems, the gravitational tugs from the two stars speed up the planetesimals so that they stop glomming together and instead start smashing apart. “The disk would basically grind itself down without managing to form planets,” says Greg Laughlin (University of California, Santa Cruz), who was not involved in the study. “Clearly that didn’t happen in Kepler-47!”

Computer simulations suggest one possibility to get around the binary problem, says Stefano Meschiari (University of Texas, Austin). If the planets form further away from their hosts, they’ll be less affected by the stars’ gravitational tugs. Then, after forming in a stable environment, the planets migrate inward to their current positions via interactions with the leftover material in the circumbinary disk that formed them.

Interestingly, all the circumbinary planets discovered so far live close to the “instability limit.” Inside this imaginary circle, the stars’ tugs become so strong that they can eject the planet from the system altogether. So the question is, why did the planets migrate inward, only to stop just outside the instability limit?

One answer might come from the planet-forming disk itself. The two central stars will naturally carve a hole around themselves that’s roughly the size of the instability region, explains Meschiari. So if planets migrate by interacting with the disk, they’ll naturally stop where the disk stops.

Then again, another possibility is that many planets form and migrate inward, and only a few survive the plunge. The rest might be ejected to roam the galaxy alone.

Laughlin suggests a third possibility: the planets didn’t migrate at all, but instead formed right where they are. In that case, the standard theory of planet formation via collisions large and small wouldn’t apply. Perhaps, he suggests, a gravitational instability caused parts of the gas disk to collapse into planets.

The authors also note that all of the six confirmed transiting circumbinary planets are Saturn-size and smaller — not one is Jupiter-size, the easiest class of planet to detect. So however these planets were created, it looks like the process doesn’t favor larger planets orbiting close to the stars.

Fortunately, Kepler has four more years to continue the search for Tatooines and unlock their secrets of planet formation.

6 thoughts on “A Tatooine Family

  1. Rod

    I prefer option 3, in situ formation which suggest abrupt and rapid formation. Computer models do not start with tiny dust grains in the disk and grow from this initial stage into planets (dust grains bounce and dissipate quickly). The models start with much larger objects near 1 km in diameter and grow in the simulations from this stage.

  2. Robert

    Kepler-47c may be to big to develop life as we know it, but if it has a moon big enough to retain an atmosphere, could it have life on it? Imagine the view : twin sunsets with a huge blue/green planet rising on the opposite side WOW! Not so far, far away from the Star Wars galaxy I may add…

  3. Bruce

    Nice comments. Rod, I agree that in situ formation makes more sense in Kepler-47 and other circumbinary systems, and if it’s required for planetary formation in cases of close binary stars then why shouldn’t it also happen in single star systems like ours? Anyway these planets of binary systems findings are showing that planets might be even more common than many have thought. Robert is right to bring up the moon possibility for Kepler-47(AB)c, which orbits close to the inner, toasty edge of the system’s HZ. The discovery of one or more large moons orbiting “c” would be a great finding. The polar areas of such hypothetical worlds could be quite appealing, temperature wise. I see that the stellar metallicity of this system is listed as 0.0000 [Fe/H]. Does this mean that Kepler-47 has exactly the same metal content as the Sun? If so this system might be even more interesting. Too bad it’s 3,878 light years away.

  4. Robert

    "Kepler has already found four circumbinary planets, each one alone in its orbit around two stars."

    Alone as far as we know! They may have small (earth size) planets around them that haven’t been detected yet…

  5. Robert

    "Kepler has already found four circumbinary planets, each one alone in its orbit around two stars."

    Alone as far as we know! They may have small (earth size) planets around them that haven’t been detected yet…

  6. Bruce

    That’s very true Robert. Small planet transit signal is hard to detect with flaring, spot ridden stars, and this noise has to be even harder to see though in multiple star systems, so small planets are very likely to be eluding detection I would think. And of course with the low odds to begin with of orbital alignment even producing transits, there will naturally be a large number of planet containing systems for every one that the Kepler team finds. Has anyone come across an estimate of the planetary transit detectability rate for Kepler stars? For example, if only, say, 2% of planetary systems would produce transits, then there would be about 50 undetectable systems for every one that Kepler can find.