A fresh analysis of data from gravitational microlensing surveys for planets orbiting stars other than the Sun finds that gas-giant planets similar to Jupiter are more common than previously thought.
The Sun's planetary configuration seems ideally suited to the emergence of advanced life. Small, rocky planets lie in the pleasantly warm region close to the Sun. The giant, gas-rich planets orbit far enough away to avoid disturbing the rocky planets, while their gravitational perturbations help scatter incoming comets that represent a collision hazard. Is the Solar System typical or is this arrangement unusual, perhaps even unique? Writing in The Astrophysical Journal, Gould et al.1 present an analysis of extrasolar planetary discoveries that brings us a step closer to answering this question.
So far, most extrasolar planets have been found by one of two methods. The Doppler velocity technique measures small shifts in the radial velocity of a star caused by the gravitational tug of an orbiting planet. The transit technique detects periodic decreases in a star's brightness when a planet passes in front of it. Both methods yield a population biased towards massive planets orbiting close to their star. Analysis of data from a large Doppler velocity survey2 suggests that about 10% of solar-mass stars possess a planet at least as massive as Saturn, orbiting within 3 astronomical units (AU) of the star (1 AU is the average distance from Earth to the Sun).
Gould et al.1 analysed data obtained using a different technique, microlensing. Here, a relatively faint star passing in front of a distant bright star acts as a gravitational lens, focusing light from the distant object, magnifying it and causing it to brighten and fade with a characteristic 'light curve' over a period of weeks (Fig. 1a). If the nearer star possesses a planet, it too acts as a lens, altering the light curve accordingly (Fig. 1b). This alteration can be large, even for a low-mass planet, but the deviation lasts for only a matter of hours, so finding and characterizing a planet requires continual monitoring of an ongoing microlensing event.
Resources are limited, so many events are observed only sporadically, biasing the distribution of planets that are found as a result. Gould and colleagues argue that rare, very-high-magnification events receive sufficient attention to provide an essentially unbiased sample. Out of 13 such events between 2005 and 2008, five resulted in planetary detections, including one two-planet system — yielding six planets in all. Using these data, the authors have estimated the abundance of planets as a function of mass and orbital separation from the host star.
The small sample size means that the numbers may change somewhat in light of future discoveries. However, two conclusions seem to be robust. The first is that many more stars have giant planets than was previously believed. Microlensing is sensitive to planets that lie at a particular separation from a star when projected on the sky. This separation is about 2.5 AU for typical lensing stars that have masses roughly half that of the Sun. Planets that lie far from this separation, especially low-mass planets, will not be seen. Despite this limitation, more than a third of the stars in the sample of Gould et al.1 have planets, a fraction that is likely to be an underestimate, yet one that is already several times larger than the fraction found by Doppler velocity surveys2.
Five of the six microlensing planets have masses between 50 and 300 times that of Earth. These objects are likely to be 'gas giants' similar to Jupiter and Saturn, consisting of a massive hydrogen–helium-rich envelope surrounding a small solid core. Theorists have long predicted3 that gas giants form in the cool, outer regions of a young star's protoplanetary disk, beyond the 'snow line' — the distance at which water ice begins to condense. For solar-mass stars, the snow line probably lies several astronomical units from the star4, which means that almost all the planets found by the Doppler velocity and transit techniques actually lie inside — not outside — the snow line.
This apparent contradiction can be resolved if some planets migrate inwards after they form, owing to tidal interactions with the protoplanetary disk or gravitational interactions with other planets. Estimating the fraction of planets that migrate significantly is not easy from a theoretical point of view. The microlensing planets are particularly useful in this regard because they probably orbit outside their stars' snow line. The high fraction of stars with microlensing planets compared with the results of the Doppler surveys suggests either that most giant planets do not migrate very far, or that most migrating planets do not survive subsequently.
The second conclusion to be drawn from Gould and colleagues' study1 concerns the planetary mass distribution. All six microlensing planets are less massive than Jupiter, even though high-mass planets can be detected in more circumstances than can low-mass objects. A similar mass distribution has been found2 for close-in planets by the Doppler velocity technique. However, in the latter case, it is unclear whether low-mass planets are more abundant than massive objects or simply more likely to migrate inwards. The microlensing results show that the mass distribution of newly formed gas giants is truly skewed towards low-mass objects. One of the six microlensing planets has a mass of roughly 13 Earth masses, comparable to the 'ice giants' Uranus and Neptune. A strong bias operates against finding such an object, so ice giants may be even more common than gas giants.
The ubiquity of low-mass gas giants poses a problem for theorists. The favoured 'core-accretion' model3 postulates that these planets form in a two-step process. Initially, dust grains in a protoplanetary disk aggregate into a solid core several times more massive than Earth. The core then accretes gas from the surrounding disk at an accelerating rate until the supply is shut off. Recent numerical simulations5 have found that core accretion can produce gas-giant planets in a wide variety of protoplanetary disks within a typical disk lifetime, and that these planets will have core masses similar to that inferred for Jupiter6. In this respect, theory and observation seem to agree. Unfortunately, however, simulations7 also predict that cores generally continue to accrete gas until the object is several times more massive than the microlensing planets or, indeed, Saturn. Clearly, there is more work to do.
The abundance of small, rocky planets at pleasantly warm distances from a star is still unconstrained by microlensing and other search techniques. Thus, we do not know whether Earth is one of many such planets or a rare gem. However, the new microlensing data suggest that the Sun's giant planets are in good company.
About this article
Reviews of Modern Physics (2011)