Our understanding of the origins and distribution of planets like the Earth has been revolutionized in the last decade or so by the discovery of over 300 planets orbiting other stars. These planetary systems differ in fundamental ways from our own solar system, and thus show that the diversity of possible planetary systems is much greater than most astronomers had suspected.

One of the most important of these differences is in the shapes of the orbits. The orbits of most planets in the solar system are very nearly circular; for example, the distance of the Earth from the Sun varies by only 3% as it travels around its orbit. Mathematically, the shape of an orbit is described by its eccentricity, which ranges from zero for a circular orbit such as the Earth's to nearly unity for extremely elongated orbits like those of some comets.

One of the most remarkable properties of the planets that have been discovered around other stars is that their orbits are much more elongated than the orbits of planets in the solar system. Figure 1 shows the orbits of two of these planets compared to Earth and Venus; the planet in green is GJ 317b, which has a typical eccentricity, and the planet in light blue, HD 80606b, has the largest known eccentricity of any planet. These elongated shapes are surprising since planets are believed to have condensed out of a circular disk of gas and dust orbiting the host star (a concept originally suggested by the Prussian philosopher Immanuel Kant in 1755), and so the orbit of a planet should be circular, like that of the gas from which it formed.

Researchers Scott Tremaine (MPA and Institute for Advanced Study) and Mario Juric (Princeton University) explored the possibility that planetary orbits acquire their shapes after the process of planet formation is complete and the gas disk has dissipated. In this hypothesis, planets form on circular orbits but then the small, regular gravitational tugs that they exert on one another as they travel around their orbits gradually make the orbits more and more elongated. In some cases this process leads to collisions of planets, ejection of planets into interstellar space, or incineration of planets by the star, while in others some or all of the planets survive but on orbits with substantial eccentricities.

To test this hypothesis, Juric and Tremaine constructed several thousand models of planetary systems, containing multiple planets on nearly circular orbits around their host stars, and followed the evolution of these orbits numerically for 100 million years. They found that in a remarkably large fraction of these model systems, the distribution of orbit shapes or eccentricities of the surviving planets matched the distribution of eccentricities in the known extrasolar planetary systems (see Figure 2). Other properties of the model systems, such as the spacings of the planets, also matched the observations well.

These findings imply that, although giant planets like Jupiter and most of the known extrasolar planets were formed within the first million years after the birth of their host stars, some of the properties of planetary systems are determined much later, in a long, slow phase of dynamical evolution lasting about 100 times longer. Their research also suggests that the orbits of the planets in the solar system are nearly circular because the planets were too widely spaced, or of too low mass, to excite significant eccentricities. They do not address the tantalizing question of why humans find themselves in such an unusual planetary system. One obvious speculation is that climatic conditions are inhospitable to life if the planetary orbits are too eccentric.

Scott Tremaine

Publications

M. Juric and S. Tremaine, "Dynamical origin of extrasolar planet eccentricity distribution ", 2008, Astrophysical Journal (to be published)