Published online 19 November 1998 | Nature | doi:10.1038/news981119-1


Playing pool with planets

Hardly a month goes by without the detection of planets orbiting stars other than our Sun. A source of surprise and puzzlement is how, very often, these extrasolar planetary systems differ markedly from our own Solar System.

Some of this exoticism may come from dramatic, catastrophic games of planetary pool, played out in the early history of a planet's life. Close encounters between a planetary system and a nearby binary star system have potential for severe orbital disruption.

In a report in Astrophysical Journal Letters, Gregory Laughlin of the University of California, Berkeley, and Fred C. Adams of the University of Michigan, Ann Arbor, Michigan, look at the likelihood of this kind of disruption in planetary systems forming in densely populated 'open clusters'. The researchers show that planets in this situation run a risk approaching 50 per cent of severe orbital disruption during their lifetimes - their circular orbits may be squeezed into elliptical shapes that take them far away or dangerously close to their parent star; their parent star may find itself becoming a binary; of the planet may be ejected from the system completely, projected into interstellar space.

All of which makes one grateful for the peace of our own Solar System, in which planets revolve around the Sun in orderly, near-circular orbits. Small planets, such as the Earth and Mars, orbit relatively close to the Sun. Larger planets, such as Jupiter and Saturn, revolve at more remote removes. (Distant Pluto, which has a highly eccentric - that is, elliptical - orbit, is something of an oddity, but may also be thought of as a kind of giant asteroid than a proper planet.)

This is consistent with the idea that the planets condensed from a gaseous 'nebula', the thickest parts of which at a distance from the young Sun of between five and 10 astronomical units (AU, where one AU equals the mean Earth-Sun distance of about 93 million miles) - distances corresponding to the orbital radii of the giant planets. Jupiter, for example, orbits at 5.2 AU, Saturn at 9.5 AU.

But ideas about planetary formation were challenged in 1995 by the very first extrasolar planet found, 51 Pegasi B - a body around half the size of Jupiter, but orbiting its star, 51 Pegasi, at a distance far closer than the torrid planet Mercury is to our own Sun. In other words, here was a giant planet so close to its star as to be almost grazing it. Several other so-called 'epistellar Jovians' have since been found. In addition, several planetary systems have been found in which the planets are large, and have highly eccentric orbits - and there is even one observation of what is claimed to be a planet in the course of ejection from a system. What is going on?

Much progress has been made in explaining unusual planetary orbits, such as that of 51 Pegasi B, in terms of tidal interactions between planets, and between planets and the stellar nebula, as planets form. The idea is that large, Jupiter-style planets emerge relatively far from the star, but in unstable orbits - stability is achieved by spiralling closer to the star. But this cannot explain everything.

Laughlin and Adams suggest that gravitational disruption in densely packed star clusters may explain much more. They cite as an example the Trapezium, an 'open cluster' in Orion, in which the density of population is around 200 per cubic light-year. This contrasts with our own relatively sparse part of the galactic neighbourhood, in which the population density per cubic light-year is precisely 1 - our own Sun, and that's that.

In a dense region, stars often approach one another close enough - of the order of 200 AU - to perturb the orbits of any planets they might have. But how often might this occur? Laughlin and Adams's computer simulations take an imaginary star cluster with a density of around 40 stars per cubic light-year, in which each star moves at around a kilometre per second. They home in on an imaginary star in this system, a star that has a Jupiter-sized planet moving in a circular orbit of 5 AU. They then ran 40,000 simulations in which this system made a close approach to a binary star system. The characteristics of the binary star system, and the trajectory of the approach, could be varied from simulation to simulation. The researchers then tabulated the results of each run.

Their results show that a planet orbiting a star in this situation has a 13% chance of suffering severe orbital disruption from an approaching binary in any given 100-million-year interval. As the initial planetary orbit was circular, the result of most encounters was to make the orbit more elliptical, and the notional planet was carried further away, or drawn towards the star. Within the body of 40,000 trials, there were 995 escapes, showing that a planet stands a 5% chance of being shot out of the system altogether, as a result of the interaction.

The results cannot explain everything. For example, an estimate that up to 2% of all star-systems contain an 'epistellar Jovian' represents a proportion too great to be explained by planetary pool-playing - tidal mechanisms within each system must also be at work.

So much for simulations - is there any evidence that the cosmic pool-player has been at work? Of the 40,000 trials, 148 events showed 'captures', in which the planet jumped between stars, or in which the imaginary solar system itself captured one of the two approaching stars. Laughlin and Adams look at a binary system called 16 Cygni, in which the two stars are separated by 835 AU. One of them, 16 Cygni B, has a Jupiter-like planet with a highly eccentric orbit. Spectroscopic analysis of both stars suggest that they are unlikely to have formed together, from the same nebula - they had very different origins, and may have come together as a result of the kind of encounter described in Laughlin and Adams's celestial pool hall.