Exoplanets in neighboring orbits have radically different sizes, masses

The Solar System is clearly divided: rocky terrestrial planets close in to the Sun, gaseous Jovian planets farther out, icy Kuiper Belt objects more distant still. However, exoplanetary systems—planets orbiting other stars—commonly violate those divisions. A whole class of exoplanets known as "hot Jupiters" are large planets with orbits smaller than Mercury's, indicating that planet formation may not follow the same rules in all cases.

As described by Joshua A. Carter et al. in Science, a newly discovered system known as Kepler-36 is even stranger. The star hosts two planets with radically different densities in very similar orbits. One planet is roughly 4.5 times more massive than Earth, indicating it is probably a rocky "super Earth," while the second planet is about 8 times more massive than Earth and roughly Neptune-sized, meaning it is likely gaseous. The difference in density is about 8-fold, indicating the two worlds must have formed in different regions of the star system, yet their orbits differ by only 10 percent. In other words, they are in closer proximity than any two planets in the Solar System, but have a larger difference in density. This challenges both naive planetary formation models that demand strict separation in planet types, and naive migration models in which planets form in one place and drift to another due to tidal forces in the protoplanetary disk.

As the name of the system suggests, Kepler-36 is one of the star systems observed by the Kepler observatory. This orbiting telescope locates exoplanets via transits: when a planet briefly eclipses its host star, in much the same way as the famous Venus transit on June 5 and 6, 2012. The amount of light that is blocked is very small in these cases, but it's enough to be measured, even for relatively small worlds. The shape of the light curve—the dip in received light from the star as the planet passes in front of it—reveals the size of the planet, while the time between the eclipses reveals the size of its orbit.

As the authors of the present study pointed out, the initial analysis of the Kepler data overlooked the Kepler-36 system. This is because the automated search algorithm was designed to seek out perfectly regular transits: eclipses that occur at equally spaced intervals. However, the two planets in the system—designated Kepler-36b and Kepler-36c, while the host star is Kepler-36a—orbit closely enough together that they slow and speed each other up as they pass. This makes the times between transits variable, so it was only when researchers adjusted the algorithm to allow for more sophisticated searches that they turned up the planets.

Even with the improved algorithm, Kepler-36b was hard to spot: it only blocks about 1 percent of the light that Kepler-36c does. However, the astronomers were able to identify both as distinct bodies orbiting the same star, partly due to their mutual interaction: when one planet was slow to arrive to its transit, the other was fast, indicating they were pulling on each other gravitationally. (A similar effect occurs in the Solar System, which is how astronomers in the 19th century discovered Neptune: they predicted where it must be based on its interaction with Uranus.)

Upon analyzing the data, the researchers determined that Kepler-36b orbits at a distance of 0.115 astronomical units (AU), while Kepler-36c is 0.128 AU from the host star. (1 AU is the average distance from Earth to the Sun; Mercury orbits at 0.30 AU, so both of these exoplanets are much closer to Kepler-36a than any planet in the Solar System.) Based on the amount of light they block, Kepler-36b is about 1.5 times Earth's radius in size, but Kepler-36c is 3.7 times larger—comparable to Neptune.

The host star is about the same mass as the Sun, but about 1.6 times the Sun's radius. These results are based on the analysis of the star's spectrum and astroseismology: the variations in light due to turbulence in the Kepler-36a atmosphere. This indicates the star is Sun-like, but older, nearing the end of its life cycle. Knowing the mass and size of the star in turn allowed the researchers to measure the masses of the exoplanets: Kepler-36b being between 4.18 and 4.78 times more massive than Earth, and Kepler-36c having between 7.62 and 8.68 times Earth's mass.

With both the masses and radii, the densities of the planets are easily calculated. Kepler-36b is very dense: about 7.5 g/cm³ (7.5 times the density of water), compared to Earth's density of 5.5 g/cm³. Thus, Kepler-36b must be rocky in composition, though it is much larger and denser than any terrestrial planet in the Solar System, so it is not particularly Earth-like. On the other hand, Kepler-36c's density is about 0.89 g/cm³, meaning it would float on water; this density is greater than Saturn's, but less than any other Solar System planet.

However, it's the contrast in density—a factor of 8—compared to the similarity in orbits that makes the Kepler-36 system odd. No other known star system has such a Laurel-and-Hardy combination: an extreme mass density ratio in planets so close to each other. According to current understanding, planets form from a protoplanetary disk of matter surrounding the young star. Friction and tidal forces within the disk can slow the motion of massive planets in the outer regions, causing them to fall inward toward the star. Whether the Kepler-36 system is consistent with such a model remains to be seen; it doesn't necessarily contradict the standard model, but the authors admitted they do not know whether the system fits or not. It is also possible Kepler-36b may have begun as a gaseous planet, but lost much of its mass through bombardment by the stellar wind (particles streaming out from the host star). This means it could have begun life farther out, similarly to Kepler-36c.

Since the star system is older than the Solar System, the conditions under which it formed are long gone. However, as with the discovery of hot Jupiters, Kepler-36 indicates how diverse planetary systems are, and how naive models for planet formation may not be sufficient to explain this diversity.

Science, 2012. DOI: 10.1126/science.1223269  (About DOIs).