The red dwarf star Gliese 581, usually abbreviated GJ 581, has repeatedly made headlines as the host of potentially Earthlike planets orbiting in the vicinity of the system's proposed habitable zone. Even if we set aside the question of habitability, this system remains intriguing as the home of a family of lightweight planets ranging (roughly) from Earth-mass to Neptune-mass, without any gas giants in evidence.

Other key factors enhance the system's interest. Not only is GJ 581 one of the nearest stars known to host planets; it is a typical M dwarf, of the type that comprises about 75% of all stars in the Milky Way Galaxy. Thus we can expect many exoplanetary systems, near and far, to resemble GJ 581.

The primary is an ordinary M3 star located at a distance of only 6.27 parsecs (20 light years) in the constellation Libra. Its mass is about 31% Solar, and its diameter is about 29% Solar, less than three times the size of Jupiter. With bolometric correction, the star’s luminosity is only 0.013 Solar, a typical value for M dwarfs. Its age is estimated at 2-4 billion years, compared to 4.6 billion for the Solar System (values after Bonfils et al. 2005, Udry et al. 2007, EPE).

As with most M dwarfs, the metallicity of GJ 581 is a matter of controversy. Recent publications have proposed values ranging from -0.33 (significantly less than the Sun’s and much lower than most planet-bearing stars) to -0.10 (closer to the average of field stars in the Solar neighborhood) (Bonfils et al. 2005, Bean et al 2006, Bailey et al. 2009, Johnson & Apps 2009). The correct value, whatever it may be, carries important implications for our understanding of planet formation around M dwarfs generally, as well as for the specific case of GJ 581.

system architecture

GJ 581 presents an ensemble of four low-mass planets orbiting within 0.3 AU of the central star, whose combined minimum mass is about 30 times Earth. Given the small size and undistinguished metallicity of the host star, it is improbable that these objects formed in their present locations. A sufficient quantity of solid materials was simply not available in the inner reaches of the system's primordial nebula.

According to the theory of planet formation by accretion, a key determinant of the evolution and resulting architecture of a planetary system is the location of its ice line. Because icy particles are abundant outside this boundary, the likelihood of planet formation is at its maximum. For M dwarfs like GJ 581, however, calculating the location of the ice line is a complex problem, since the luminosity (and thus the temperature) of such stars changes dramatically during the early stages of system evolution (Kennedy et al. 2006). Ida and Lin (2008) suggest a radius of about 0.4 AU for the ice line of a star like GJ 581. This value implies that the system's principal nursery for planet formation lay beyond 1 AU (see Kennedy et al. 2006).

Architecture of the GJ 581 system. Colored circles indicate the relative sizes of the 4 planets, assuming the minimum masses provided by Mayor et al. 2009 and the mass-radius relationships provided by Fortney et al. 2007. Semimajor axes are indicated in astronomical units (AU) on a logarithmic scale. White dots mark the ice line.

The previous considerations suggest that the three outer planets (b, c, and d) formed well outside the ice line and then spiraled inward to their present orbits through a variety of possible mechanisms. On the other hand, the innermost planet (e) may have assembled as a consequence of the shepherding of rocky planetesimals within the orbit of planet b as it swept through the inner regions of the nebula.

four planets

Unless otherwise indicated, all planet parameters derive from Mayor et al. 2009.

The inner planet, GJ 581 e, is a terrestrial-mass object at the low end of the range for Super Earths. Its minimum mass is estimated at 1.94 MEA, while its maximum mass, determined on the basis of dynamical considerations, is only 3.1 MEA. Planet e currently holds the title of the least massive exoplanet detected by the radial velocity method. With a semimajor axis of 0.03 AU and an orbital period of 3.1 days, the planet's surface temperature must exceed Mercury's. Michel Mayor and colleagues conclude that planet e "is almost certainly rocky," and that its temperature is too high to permit the survival of a significant atmosphere. Given its proximity to the host star, the planet's rotation must be tidally locked, with one hemisphere always in light and the opposite always in shadow. Such conditions may cause portions of its daylight surface to be molten, with a "magma pond" occupying the hottest point (Ganesan et al. 2008).

The second planet, GJ 581 b, is a Hot Neptune. In fact, its minimum mass of 15.7 MEA and maximum mass of about 30 MEA make it a near-twin to the original cold Neptune. However, its orbital period and semimajor axis are only 5.4 days and 0.041 AU, respectively. Having assembled in the region beyond 1 AU around a cool, metal-poor star, GJ 581 b probably consists largely of ices, with a rocky core and a substantial hydrogen atmosphere. Such is the inferred composition of the Hot Neptune that transits GJ 436, an M dwarf very similar to GJ 581 (Gillon et al. 2007). Because GJ 581 b is hotter than Venus, its ices must subsist in a high-pressure, superheated environment unlike anywhere in the Solar System.

Two more planets travel on wider orbits that are notably more eccentric than those of planets e and b. Both are in the mass range of GJ 876 d, making them Super Earths.

GJ 581 c has a minimum mass of 5.36 MEA (maximum 10 MEA), a period of 12.9 days, and a semimajor axis of 0.07 AU. Its orbital eccentricity is calculated as 0.17, a little less than Mercury's. Like its inner companion, planet c must have a substantial quantity of ice in its bulk composition, similarly pressurized and superheated.

GJ 581 d is more remote and only slightly heavier, with a minimum mass of about 7.09 MEA (maximum 14 MEA), a semimajor axis of 0.22 AU (about half the distance of Mercury from the Sun), and a period of almost 67 days. Planet d's orbital eccentricity of 0.38 is the highest yet to be proposed for an M dwarf planet, indicating a periastron of 0.14 AU and an apastron of 0.30 AU. This large variation in orbital separation implies large temperature changes over the course of a single orbit, so that planet d may have pronounced seasons. During its brief summer, this planet might be warm enough for water to rain out of its turbulent skies.

The planet's high eccentricity suggests another possibility. Although close-in planets are usually regarded as tidally locked, with one hemisphere always facing the parent star and the other always in shadow, our own Mercury escaped this outcome despite its proximity to a star three times heavier than GJ 581. Mercury has achieved a stable "spin-orbit resonance," such that it rotates three times for every two orbits around the Sun. Because this orbital behavior correlates with Mercury's eccentricity, a similar resonance might have evolved in GJ 581 d, making it still more favorable to the emergence of Earthlike conditions.

Finally, a large body of theoretical work argues that elliptical orbits originate in episodes of planet-planet scattering. We can surmise that the GJ 581 system originally contained an additional object less massive than planet d, which approached its sibling close enough to be swung into a much wider orbit, or even to be ejected from the system. Having won the contest, planet d settled into its current state of eccentricity (see Ford et al. 2005).

questions of habitability

Much speculation has centered on the question of whether either of the Super Earths, GJ 581 c or GJ 581 d, might be habitable. While the precise delineation of habitable zones remains an art rather than a science, the original discovery team for planet c argued that it probably orbits at the inner edge of this region. If it sustains a relatively transparent atmosphere, with no extreme greenhouse effect, it might be covered by a global ocean. When the planet's discovery was announced in 2007, astronomer Xavier Delfosse was widely quoted in the media (CNN, Time Magazine, Agence France-Presse) for observing, "On the treasure map of the universe, one would be tempted to mark this planet with an X."

Subsequent studies have been less optimistic. Franck Selsis and colleagues conducted a detailed exploration of the system's potential habitability, foregrounding the complexity of the problem and proposing alternative models for the atmospheres of planets c and d (Selsis et al. 2007a). They offered a range of boundaries for the habitable zone, among which the most conservative were 0.08 AU to 0.2 AU, and the most generous were 0.06 AU to 0.3 AU. W. von Bloh and colleagues performed a competing analysis, favoring larger radii. Their most generous boundaries for the habitable zone were about 0.13 AU to 0.26 AU (von Bloh et al. 2007).

Both groups concluded that GJ 581 c is too hot to sustain liquid water, while the cooler fourth planet (d) makes a much better candidate.

The composition of these two planets remains unknown in the absence of any observational techniques to shed light on their diameters or densities. Nevertheless, current theory indicates that protoplanetary disks – as well as the planets that evolve within them – share fundamental characteristics with their host stars (Greaves et al. 2007, Raymond et al. 2007, Johnson et al. 2007). Given a low-mass, low-metallicity star like GJ 581, we can expect a similarly low-mass, low-metal protoplanetary disk. Such an environment is not conducive to the formation of rocky planets as massive as GJ 581 c and d (Lissauer et al. 2007, Raymond et al. 2007, Thommes et al. 2008).

It seems likely, therefore, that these two planets consist of about 50% water ice and 50% metals and silicates, which is the suggested ratio for icy Super Earths (Leger et al. 2004, Selsis et al. 2007b). They may have formed more or less at their present locations as a result of the passage of planet b (the Hot Neptune) through the primordial disk, because this migration would have stirred up icy planetesimals and made them available for accretion on small orbits (see Mandell et al. 2007). However, a more likely alternative is that both assembled near the system's ice line, just like their larger companion, and then traveled to their present orbits by Type I migration (Selsis et al. 2007b, Kennedy & Kenyon 2008), perhaps undergoing an episode of planet-planet scattering in the process.

If both objects are indeed icy, planet c may be a "Steam Planet," with a scalding atmosphere of pure water vapor in direct contact with a layer of high-pressure, superheated ice. Planet d might be an "Ocean Planet," cool enough to sustain a global ocean 100 kilometers deep above its icy mantle.

additional planets?

In announcing the discovery of GJ 581 e, Michel Mayor and colleagues proposed the exciting possibility that an unconfirmed fifth planet might orbit in the zone of stability between planets c and d. This hypothetical planet f would have a maximum mass about twice that of Earth and an orbital period of about 25 days. As Mayor's team pointed out, "Such a planet would be Earth-type and in the middle of the habitable zone" (Mayor et al. 2009). Other astronomers have been quick to take up this notion, albeit informally, with Greg Laughlin musing, "Gliese 581 f seems like such a made-to-order confection that it’s simply got to be there. Which is a flimsy argument, I admit." (See Laughlin's blog entry, "Bode's Law," at http://oklo.org/).

Last update January 2010