Circumbinary Planet formation: Here be high-speed impacts

[A shorter-form version of this article I wrote has been published on The Conversation UK.]
The Conversation
Last week was a hot week for those interested in circumbinary planets and how they form. Firstly, Kostov et al. discovered Kepler-413b, another Neptune-size circumbinary planet; this one has some interesting dynamical considerations owing to its inclination. Secondly, Lines et al. published a new N-body simulation of the formation environment close to the Kepler-34 binary.

In this post, I will explain in layman’s terms how we think planet formation around binary stars works.


The giant planet Kepler-34b orbits round two stars. Now that’s just greedy. David A. Aguilar

It is quite remarkable to remember that, until less than two decades ago, scientists had no concrete examples of planets orbiting other stars like the Sun. The Solar System was not only the archetype of planetary system, but the only planetary system astronomers knew of.

Since the discovery of a giant planet around the Sun-like star 51 Peg in 1995, the pace of discovery of exoplanets (planets orbiting other stars) has not slowed down. Quite the contrary: the trickle of planet discoveries became a steady stream in the 2000’s, as technology improved and astronomers became more efficient at homing in on stars likely to have planets orbiting them. In 2009, the Kepler space telescope was launched with the express goal of simultaneously surveying hundreds of thousands of stars for the presence of exoplanets.

The advent of Kepler transformed the steady stream into a veritable deluge of planets and planetary systems. Today, more than a thousand exoplanets are known, with thousands of planet candidates waiting for further verification and vetting. Although this quickly expanding “census” of exoplanets has injected new life to the field of planet formation (the study of how planets are born), it has also brought forth many new questions.

A big leap

Our best understanding of how planets form is inextricably linked with our understanding of how stars form. Many different lines of strong evidence point to planets forming inside a thin, gaseous disk surrounding nascent stars. Within this disk, solid particles (evocatively named “dust”) collide and progressively grow to asteroid-sized bodies. These bodies, called planetesimals, are the essential “bricks” of planet formation. Further collisions among planetesimals build protoplanets — rocky, Earth-sized bodies.

Farther out from the central star, water and other compounds “freeze out” and become part of the solid component. At and beyond this location (the “ice line”), protoplanets can grow even larger and amass thick, massive atmospheres. This sharp divide between small, Earth-sized planets close to the central star (Mercury through Mars) and massive giant planets further out (Jupiter through Neptune) is easily recognized in the Solar System, where the ice line is just inside Jupiter’s orbit.

For this theory to work, it demands an incredible feat: through collisions and gravity alone, it requires growth from microscopic dust particles more than a hundred times smaller than a grain of sand, all the way to Jupiter-sized objects. To put things into perspective, it is the same jump in size between a single atom and an average sized human! This is a very delicate process, involving many physical mechanisms, some of which are still poorly understood today.

Double trouble

One of the sticking points is the stage in which planetesimals collide. Planetesimals need to collide surprisingly gingerly in order to accrete; smash them too fast, and they will “break” into smaller rocks. Regions with high-speed collisions become essentially sterile for planet formation, as no further growth happens under normal circumstances.

This is why the recent discovery of circumbinary planets had astronomy theorists raise an eyebrow (or two). Circumbinary planets are planets that orbit a binary star. Such stars are bound together by gravity into an often-tight orbital dance. Kepler-16 ABb, the first circumbinary planet discovered in 2011 by the Kepler spacecraft, orbits both stars A and B (hence the AB monicker; the lowercase b simply means “the first planet discovered in the system”). The recent discovery of Kepler-413b by the team led by Veselin B. Kostov adds another member to the (so far) very exclusive circumbinary club: seven planets have been discovered orbiting binary stars, all giant planets with sizes between Neptune and Saturn.

Everyone likes to bring up Star Wars' Tatooine, but I like Doctor Who more -- so I will show Gallifrey (the home world of the Time Lords) instead. BBC, Peter McKinstry
Everyone likes to bring up Star Wars’ Tatooine, but I like Doctor Who more — so I will show Gallifrey (the home world of the Time Lords) instead. BBC, Peter McKinstry

That these planets, once the purview of sci-fi fare such as “Star Wars” or “Doctor Who”, were detected is alone quite a testament to the power of Kepler and its skilled team: astronomers had to meticulously analyze small variations in brightness of the two stars, caused both by mutual eclipses (each star passing in front of each other) and planetary eclipses (the planet passing in front of one star or the other).

Their very existence, however, is also a testament to the surprising resilience of planet formation. As mentioned before, our models indicate that planetesimals will be destroyed if their impact speeds are too high. But this should be exactly the case around binary stars! Binary stars should gravitationally perturb planetesimals, just like Jupiter perturbs asteroids and comets in the Solar System. These perturbations fling planetesimals into orbits that, when crossing other planetesimals’ orbits, assure high-speed impacts and mutual destruction. Only far away from the central binary, where perturbations are weaker, it is expected that collision speeds would become low enough to resume planetary building.

All things considered, this should be an easily overcome hurdle: as explained in the previous section, far enough from the central star (or stars) is where giant planets form — and all circumbinary planets found so far are giant planets! Is this a spectacular confirmation of theory?

Too close for comfort?

Not so fast, unfortunately: all circumbinary planets discovered so far are also orbiting very close to their parent binary. So close, in fact, that if they were any closer to the binary, their orbit would be destabilized to the point of ejection from the system or collision with one of the two stars. These planets are essentially flirting with disaster, their orbital velocity closely balancing the gravitational attraction of the binary in a dangerous tug of war. Inside their orbit, in the “instability region”, no planet could survive for long.

Reconciling these two apparently incompatible findings (giant planets in a location where they should have never formed) required invoking an old idea: that of planetary migration.

The very first planet discovered, 51 Peg b, was a giant planet orbiting its parent star at a very small distance — smaller than even Mercury. Such a planet could have never formed so close to the star, as the high temperatures would sublimate rocks and ices. Our theory would then predict that there would not be enough bricks (planetesimals) to build a giant planet. Theorists quickly understood what had occurred early on in this system’s history: 51 Peg likely formed further away from the star, and subsequently interacted with the disk in which it was born in such a way to be pushed further in. This process is called “migration”.

Migration could be at play with circumbinary planets as well. Computer models have shown that a giant planet formed far from the central binary will tend to migrate and move inwards. Encouragingly, the migrating planets do not move all the way to the instability region. Rather, they tend to stop at a specific distance, which matches well with their current location.

One last finding of computer models match another observed property of circumbinary planets. While we admittedly discovered only a handful of circumbinary planets, it is tantalizing that all of them are quite a bit smaller than Jupiter. It is surprising as the bigger a planet is, the easier it is to detect: therefore, we should have discovered a few Jupiter-sized circumbinary planets by now. Computer models explain this last piece of the puzzle: migrating Jupiter-sized circumbinary planets end up strongly interacting with the central binary and are subsequently flung out from the system. We would not observe such planets today simply because they did not survive their turbulent beginnings.

Although there are still many details to be worked out, this theoretical framework appears to be in good concordance with Kepler’s discoveries. However, it is possible, even expected, that further planetary discoveries might surprise us once more.