The setting of twin suns on some distant world has become an iconic image of science fiction, and easily imagined as an indelibly alien and distant experience. So it's not surprising that detections of some of the more extreme examples of exoplanets in binary star systems have hit the headlines. Most recently, NASA's Kepler mission has spotted the sure-fire signature of two planets transiting across the faces of two stars in the system Kepler-47. The twist is that the stars orbit about each other within the orbits of the planets.
In this system a Sun-sized star and a star about a third of the size waltz around their mutual center-of-mass every 7.45 days, while the two detected planets orbit much further out, taking about 50 and 303 days to complete a circuit. While this is not the first so-called circumbinary planetary system (5 other ones with planetary transits were known before this, and we've even found planet-sized objects orbiting a pulsar and white dwarf pair), it is notable because it includes the smallest known planet in such a system (the inner Kepler-47b at about 3 times the radius of the Earth) and the outer planet is in the nominal 'habitable zone'. It also helps establish the fact that bona-fide circumbinary planetary systems can indeed exist.
This is a non-trivial thing to be able to say. A pair of whirling stars at the center of a system present a tricky dynamical landscape for both the formation of planets and their long-term orbital stability. The situation is somewhat easier in wider binary configurations, where stars may orbit tens or hundreds of astronomical units apart, and indeed we have found exoplanets in such places - hugging one or the other stellar parent where the dynamical 'weather' is calmer. The Kepler-47 result suggests that planets can still form out of a disk of gas and dust surrounding both stars, although figuring out exactly how that works is going to keep astronomers busy for many years to come.
But given the hoopla made about 'Tatooine' systems and the like, (sadly the planet in Kepler-47's habitable zone is perhaps more of a Neptune than a super-Earth) it's worth pausing to ask just how many exoplanets are really out there in either circumbinary orbits or with more distant sister stars in their skies.
The most honest answer is that we really don't know for sure yet, but we do know some interesting things about stellar multiplicity that give us some clues. Years ago the received wisdom was that as many as 2/3rds of all stars were a part of multiple stellar systems, gravitational doubles, triples, etc. But more recent work has suggested that this is not the case, or at least it depends on the type of stars. Work by Lada in 2006 revisited this issue and found that for Sun-like stars (known as G-dwarfs) the multiplicity rate is indeed about 60%, but for lower-mass stars (M-dwarfs) this drops significantly and only about 25% have orbiting companions. These lower mass stars are actually far more numerous than stars like ours, over 75% of all stars are less than half as massive as the Sun.
These statistics are fascinating for understanding the pathways by which nature produces stellar systems, but the upshot for planets is that of all main-sequence (hydrogen burning) stars in our galaxy, irrespective of size, 70% end up single. Of those that are not single it seems that a relatively small percentage (perhaps less than 10%) are in close-orbiting pairs that would make suitable homes for circumbinary planets will Hollywood-worthy sunsets. So whichever way you cut it, the fraction of planets with more than one bright sun in their skies is almost certainly a small minority.
We might have never thought otherwise, and this in no way diminishes the beauty of these systems, but in science it's always good to check the numbers.