Planets orbiting two stars have been found, but not all that many of them. We’re talking here about a planet that orbits both stars of a close binary system, and thus far, although we’ve confirmed over 6,000 exoplanets, we’ve only found 14 of them in this configuration. Circumbinary planets are odd enough to make us question what it is we don’t know about their formation and evolution that accounts for this. Now a paper from researchers at UC-Berkeley and the American University of Beirut probes a mechanism Einstein would love.
At play here are relativistic effects, having to do with the fact that, as Einstein explained, intense gravitational fields have detectable effects upon the stars’ orbits. This is hardly news, as it was the precession of Mercury in the sky that General Relativity first predicted. The planet’s orbit could be seen to precess (shift) by 43 arcseconds per century more than was expected by Newtonian mechanics. Einstein showed in 1915 that spacetime curvature could account for this, and calculated the exact 43 arcsecond shift astronomers observed.
What we see in close binary systems is that if we diagram the elliptical orbit usually found in such systems, the line connecting the closest approach (periastron) and farthest point in the orbit (apoapsis) gradually rotates. The term for this is apsidal precession. This precession – rotation of the orbital axis – is coupled with tidal interactions between the two stars, which make their own contribution to the effect. Close binary orbits, then, should be seen as shifting over time, partly as a consequence of General Relativity.
The researchers calculate that as the precession rate of the stars increases, that of a planet orbiting both stars slows. The planet’s perturbation can be accounted for by Newtonian mechanics, and its lessening precession is the result of tidal effects gradually shrinking the orbit of the two binary stars. But note this: When the two precession rates match, or come into resonance, the planet experiences serious consequences. Mohammad Farhat, (UC Berkeley) and first author of the paper, phrases the matter this way:
“Two things can happen: Either the planet gets very, very close to the binary, suffering tidal disruption or being engulfed by one of the stars, or its orbit gets significantly perturbed by the binary to be eventually ejected from the system. In both cases, you get rid of the planet.”
Image: An artist’s depiction of a planet orbiting a binary star. Here, the stars have radically different masses and as they orbit one another, they tug the planet in a way that makes the planet’s orbit slowly rotate or precess. Based on dynamic modeling, general relativistic effects make the orbit of the binary also precess. Over time, the precession rates change and, if they sync, the planet’s orbit becomes wildly eccentric. This causes the planet to either get expelled from the system or engulfed by one of the stars. Credit: NASA GSFC.
Does this mean that circumbinary planets are rare, or does it imply that most of them are probably in outer orbits and hard to find by our current methods? Ejection from the system seems the most likely outcome, but who knows? The researchers make three points about this. Quoting the paper:
(i) Systems that result in tight binaries (period ≤ 7.45 days, that of Kepler-47) via orbital decay are more likely than not deprived of a companion planet: the resonance-driven growth of the planet’s eccentricity typically drives it into the throes of its host’s driven instabilities, leading to ejection or engulfment by that host.
(ii) Planetary survivors of the sweeping resonance mostly reside far from their host and are therefore less likely to have their transits detected. Should eccentric survivors nevertheless be detected, they are expected to bear the signature of resonant capture into apse alignment with the binary.
(iii) The process appears robust to the modeling of the initial binary separation, with three out of four planets around tight binaries experiencing disruption…
What we wind up with here is that circumbinary planets are hard to find, but the greatest scarcity is going to be circumbinaries around binary systems whose orbital period is seven days or less. The researchers note that 12 of the 14 known circumbinary planets are close to but not within what they describe as the ‘instability zone,’ where these effects would be the strongest. Indeed, the combination of general relativistic effects and tidal interactions is calculated here to disrupt planets around tight binaries about 80 percent of the time. Most of the planets thus disrupted would most likely be destroyed in the process.
The paper is Farhat & Touma, “Capture into Apsidal Resonance and the Decimation of Planets around Inspiraling Binaries,” Astrophysical Journal Letters Vol. 995, No. 1 (8 December 2025), L23. Full text.
If this hypothesis is correct, would we see many more protoplanetary disks in circumbinary orbits of the binary stars, before the effect occurs after planetary formation (at least in the close orbit where the effect would later impact the planets that formed)?
Very interesting. I haven’t thought of star systems with radically different masses. Some planets might be ejected or broken apart being to close to the star according the idea in this paper. I am wondering where the life belt is in a double star system. If the double star system has stars of similar mass we could have a Tatooine if there is no orbital precession resonance between the star system and planets.
‘”…star systems of radically different masses”.
This opens the door to a whole new set of problems. Not only is the gravitational environment about a close binary unstable in the long term, but there is another factor with astrobiological implications. Stars of highly disparate masses will evolve differently. In short, the evolution of most main sequence stars is roughly the same, but the time frame is a function of the initial mass. The more massive a star is, the brighter and hotter it is, the faster it evolves, and the less time it stays on the main sequence. When we study a binary system, we see only a snapshot of the mutual evolution, we see both stars as they appear NOW. If one was much more massive than the other, it might have already evolved off the main sequence, become a red giant, gone through several planetary nebula outbursts, and eventually settled down to a long, faint old age as a white dwarf. If it was heavier than Chandrasekhar’s Limit (roughly two solar masses), it might have gone supernova, or be getting ready to do so. And if it was close enough to its companion and matter transfer occurred there’s no telling how the system can wind up!
The upshot is that if we see a star with a more massive close companion (even if it appears small and faint to us now), it may have already gone through a history that was unlikely to result in a stable life-friendly environment. It is unlikely that life arising in either star would survive the stellar evolutionary consequences of the life cycle of the massy companion.
Add to this the companion star pulling material off from the red giant star. This must change the luminosity of the stars, especially the companion star. There are many beautiful paintings of such binaries, such as this one by David Hardy.
How do all these many phenomena affect the needed stability of the climate on the circumbinary planet?
It is useful to constantly remind oneself when speculating about these binaries that the radius, brightness and surface temperature of a star change throughout its lifetime, and although related to the initial mass and age, do so only very indirectly.
So, for example, a faint, tiny, very hot white dwarf and a bright, enormous and very cool red giant may be companions, and they must be of the same age because they are members of a binary system, but without knowledge of their orbital mechanics you have no way of knowing their masses.
We tend to visualize these systems as a small star orbiting a big one, but it is mass, not radius, that determines how the stars orbit a their common center of gravity. The Tattoine scenario implies two similar sunlike stars in close orbit, not a likely circumstance.
This was a challenging entry to think about – and it has mostly been by skirting around GR. Bu there goes.
With 6000 exoplanets found, is identification of 14 orbiting close binaries necessarily a low score?
Maybe this is a little afield, but I note that the two Dog Stars, Sirius and Procyon, are binaries orbiting with white dwarfs. Two rather bright and young main sequence stars, right? After all they are brighter than the sun but destined to have much shorter lifetimes. Yet their binary partners are white dwarfs.
My simple minded explanation for these two paradoxical systems is that Procyon A and Sirius A are the results of mergers of closely orbiting lesser mass main sequence stars. With a third star the three body system might have allowed for dissipative mechanisms to cause an inward spiral, but that might be waving hands. This might be a difficult hat trick to pull according to stellar evolutionary theory,… It might have been rough on any neighborhood planets too.
Close binaries are not necessarily all alike. One could have more mass than the other and then their orbits would be inversely proportional and decaying as well due to tidal dissipation. That could cause planets to draw away without invoking
GR.
But otherwise looking at the two dog stars from a stellar evolutionary standpoint feels much akin to a cat trying to avoid staring into a mirror.
Looking for clues about planets and stellar binaries from another angle, turned to the remarkable set of binary hierarchies: Castor Aa & Ab, Ba & Bb, Ca & Cb. Castor is Alpha Geminorum and Pollux being Beta… Beta Geminorum, Pollux not really part of the aforesaid hierarchy but so often next mentioned, might have a planet. But I see no mention of planets in the Castor hierarchy of embedded binaries. A lot of hierarchy displayed ,nonetheless. But if planets were ejected “here”, then maybe they would be captured “there”?
Depending on the binary system, luminosity can vary considerably. For close binaries, transit observation just might be a challenge due to glare associated with the two stellar components, or whether an exoplanet passed in front of both or only one and how far removed it has to be for stable celestial dynamics. The shorter the period, the faster an exoplanet’s orbit can be confirmed. Third body detection from doppler shift, well that might have difficulties too. Some binaries might be near circular, but others not.
It would seem after reviewing the possibilities, it could be that in isolation, GR could result in some exoplanet casualties or losses, but I suspect a lot of other environmental elements muddy the picture.
A white dwarf in orbit around a main sequence star, even a fairly young one, may be the result of matter transfer between the pair at some time in the past. But it could also be a result of the more massive star having rushed though its evolution quickly. Matter transfer would only be an issue if the binary members were fairly close together to begin with. For the formation of contact binaries or type I supernovae the pair would have to be a relatively close pair.
This raises other issues. Our transit and tidal techniques for planetary detection work much better in close planetary orbits that are more likely to be disrupted in a close binary. There is also a selection effect at work here, binaries are much easier to detect than planets. Planets are easiest to find when they orbit close to their primary (or primary pair) and when the plane of the system allows transit observations to be possible. Identifying stars as binaries is much easier than finding planetary companions, because neither the distance between the two stars nor the geometry of the orbital plane/celestial line of sight is an issue.
You’re absolutely right; “… a lot of other environmental elements muddy the picture.” But we must always keep in mind that the evolution of star is highly dependent on its initial mass. A 20 solar mass giant (aside from being highly unusual, a very rare bird) may only have a life span of a few million years before it evolves off the main sequence. It will take Sol a few Billion years to do that. As for your 0.1 solar mass red dwarf, the universe just isn’t old enough yet for any of them, even the earliest to form after the Big Bang, to have fully evolved.
For SETI purposes, or for astrobiological speculations, close (< several AU) binaries are probably not a good place to search for traces of extrasolar life. Unless both stars are of roughly the same mass, the elder would have evolved to the point where the entire habitable zone of the system would be sterilized and all volatiles boiled off from around the survivor. IOW, BOTH stars would need sufficient time for life to arise on just ONE of them.