varieties of extrasolar hosts

Planet formation is common around a wide variety of stars. Cumming and colleagues (2008) estimate that at least 17% of Sun-like stars have gas giant planets orbiting within 20 AU (the approximate distance of Uranus from our Sun). Greaves and colleagues (2007) predicted that at least 14% of all stars have giant planets, while an additional 16% are accompanied by dusty debris disks and smaller planets. The actual numbers may turn out to be even higher, given the observation that at least 80% of young stars are surrounded by the clouds of gas and dust (protoplanetary disks) whose evolution gives rise to planetary systems (Haisch et al. 2001).

As it is, planets have been detected or strongly suspected around stars of all the commonly occurring spectral classes in the Solar neighborhood (A through M). They have been discovered orbiting single, binary, and triple stars. They have been found orbiting young stars with extensive debris disks (Epsilon Eridani) as well as around highly evolved stars that have left the main sequence and expanded into subgiants and giants (Kappa Coronae Borealis, Pollux). They have been detected around white dwarfs (GD 66) and their binary companions (Gliese 86/HD 13445), proving that planetary systems can survive the explosive stellar collapse that creates such objects. They have even managed to form out of supernova debris in orbit around neutron stars (PSR 1257+12), which result from explosions so powerful and bright that they can be seen millions of light years away.

metallicity and mass

Two important statistical regularities have been observed within the sample of detected systems. First, stellar metallicity is the best predictor of a star's likelihood of hosting massive planets, particularly those in close orbits of a few days to a few months. Stellar metallicity is defined as the proportion of heavy elements (especially iron) to hydrogen and helium in a star’s chemical composition. Metallicity is expressed in logarithmic form and written [Fe/H]. Zero represents the metallicity of the Sun; positive numbers indicate higher metallicity, negative numbers indicate lower.

Surveys of F, G, and K stars within our Local Bubble suggest that about 25% of stars at the high end of metallicity have gas giant planets, compared to only about 3% of stars at the low end (Marcy et al. 2005). Paradoxically, the Sun's metallicity falls near the middle of this range, demonstrating that a modest degree of metallic enhancement is no obstacle to the formation of complex planetary systems.

The second significant regularity in planet occurrence is stellar mass. Johnson and colleagues (2007a, 2007b) find that giant planets become increasingly more likely as star mass increases from the range typical for M dwarfs to that of bright A-type stars. Further, planets of more massive stars are typically larger than those of lightweight stars. These observational trends are supported by the analytic studies of Kennedy & Kenyon, who conclude, “The probability that a given star has at least one gas giant increases linearly with stellar mass from 0.4 MSOL to 3 MSOL” (Kennedy & Kenyon 2008).

These trends indicate that the architecture of planetary systems is at least partly a function of the host star's spectral class (see Comparative Sizes of Nearby Stars and Architecture of Planetary Systems).

system architecture and spectral type

Hot white stars of spectral class A span the mass range from about 1.6 to 3 times Solar (Johnson et al. 2007a). Although no planetary companions have yet been confirmed for any A stars on the main sequence, such nearby systems as Vega, Fomalhaut, and Beta Pictoris have yielded data strongly suggesting the presence of exoplanets (see The Vega Trio). In addition, increasing numbers of planets are being detected in orbit around red giants like Epsilon Tauri, which represent later phases in the evolution of A stars (Sato et al. 2007, Johnson et al. 2007a).

Because these stars are hot and massive, their ice lines – the distance at which volatiles such as water will freeze – are considerably wider than those around stars like our Sun. A value of 8 AU or more has been suggested for the radius of the ice line around a young Vega-like star (Kennedy & Bromley 2007). While the formation of gas giants is encouraged just outside the ice line of most stars (Ida & Lin 2004), such planets are likely to form even inside the ice lines of A-type stars (Kretke et al. 2008), given the mass and pressure gradient of their protoplanetary disks. In addition, many such stars preserve debris belts for hundreds of millions of years – a significant fraction of their main sequence lifetimes of 600 million to 1 billion years. Although these brief lifespans probably rule out the development of complex organisms on any potential planets, A star systems will still contain substantial habitable zones – that is, orbital space in which a body of water on the surface of a terrestrial-mass planet could persist in liquid form. It remains uncertain, nevertheless, whether rocky planets can form or maintain stable orbits in these systems, unless perhaps as satellites of massive gas giants.

Yellow and orange stars of spectral classes F, G, and K are cooler, smaller, and longer-lived, with temperatures decreasing and longevity increasing at lower masses. Upsilon Andromedae, our Sun, and HD 128311 are typical representatives of each class. Together they comprise the population of so-called “Sun-like” stars. Their mass ranges overlap considerably. Spectral type F is associated with masses between 1.1 and 1.5 Solar; spectral type G with masses between 0.85 and 1.2 Solar; and spectral type K with masses between 0.6 and 0.95 Solar. The ice lines around these stars are much tighter than those of A-type stars – at about 3 AU for a G star like the Sun – so their terrestrial planet-forming regions are correspondingly smaller. The midpoint of their habitable zones ranges from 0.7 AU, for a K star like HD 128311, to about 2 AU, for a late F star like Upsilon Andromedae. Because this population of stars is the most intensively studied, most observational data and theoretical modeling pertain to their planetary systems.

Red dwarf stars of spectral class M constitute a distinct low-mass regime of their own (see Planets of Red Stars). These are the commonest and most long-lived spectral type both in the Local Bubble and in the entire Milky Way – such that all red stars born in the primordial galaxy, barring random cataclysms, are still burning. Cool dim stars of this type have only recently become the target of radial velocity searches, and preliminary results suggest that giant planets, at least, are less likely to exist in these systems than around more massive stars (Endl & Cochran 2006). Nevertheless, M stars constitute at least 75% of all stars in the galaxy (Tarter et al. 2007), so that their planetary companions may outnumber those of all other spectral types. Indeed, large-sample gravitational lensing observations of the Galactic Bulge discover planets around them more often than around other types of stars (Sahu et al. 2006). Because the largest M stars on the main sequence have only 50% of the Sun’s mass, their ice lines are correspondingly tighter, at about 0.4 AU (Ida & Lin 2005), while their habitable zones center in the range of 0.1-0.2 AU (Tarter et al. 2007, Jones et al. 2006).

Last updated February 2008