Located at a distance of only 4.7 parsecs (15.4 light years) in the constellation Aquarius, Gliese 876 – usually abbreviated GJ 876 – is an ordinary red dwarf of spectral type M4. Roughly one-third the size and mass of our Sun, this star is notable as the first M dwarf to be identified as the host of a planetary system. Since 1998, four planets have been discovered in short- to medium-period orbits around the central star: one Hot Super Earth, two gas giants, and a Warm Neptune (Marcy et al. 1998, Marcy et al. 2001, Rivera et al. 2005, Rivera et al. 2010). Remarkably, the three outer planets exhibit a classic Laplace resonance, meaning that their orbital periods form a ratio of 1:2:4. Such a clockwork relationship has previously been observed only in the satellite system of our own planet Jupiter, among the moons Io, Europa, and Ganymede.

Studies over the past dozen years have helped to clarify the host star’s basic parameters. For the mass of GJ 876, Alexandre Correia and colleagues provide a value of 0.33 MSOL (Correia et al. 2010), almost identical to the 0.32 MSOL assumed by Eugenio Rivera and colleagues (Rivera et al. 2010). The star's luminosity is estimated by Correia’s group as 1.3% Solar, which is typical for M dwarfs in this mass range. The star’s metal content or [Fe/H] has been a matter of disagreement, with estimates ranging from a low of -0.40 to a high of +0.37 (Bailey et al. 2009, Johnson & Apps 2009, Schlaufman & Laughlin 2010). Recent publications show a trend toward higher values, with the latest recalibration favoring a metallicity of +0.23 (Schlaufman & Laughlin 2010). Overall, GJ 876 seems to be one of the more metal-rich M dwarfs known to harbor a planetary system.

The star’s age remains the most difficult parameter to constrain. Correia’s group observe that GJ 876 displays modest magnetic activity and rotates quite slowly, with a period of about 97 days (Rivera et al. 2005) – almost quadruple the Sun’s period of 25 days. Despite these indications of stellar maturity, they provide a minimum age of only 100 million years, and a maximum of 5 billion years (virtually the same age as our Sun), primarily on the basis of the star’s orbital motion around the Galactic Center (Correia et al. 2010). Rivera and colleagues simply observe that the star’s age “exceeds 1 Gyr” (Rivera et al. 2010).

system architecture

The architecture of the GJ 876 system is notable for at least reasons. First, it remains the only M dwarf known to host two gas giants. Although additional two-giant systems are likely to be revealed around other red dwarfs in the near future, given the presence of long-term radial velocity trends for such stars as GJ 317 and GJ 849 (Wright et al. 2009), it is still true that 40% of the known M dwarf systems lack any gas giants whatsoever. This dearth contrasts sharply with the situation for host stars in the mass range of our Sun (0.9-1.1 MSOL), among which 95% harbor at least one gas giant.

Architecture of the GJ 876 system. Colored circles indicate the relative sizes of the 4 planets, assuming the actual masses provided by Rivera et al. 2010, the mass-radius relationships provided by Fortney et al. 2007, and rock/metal cores. Semimajor axes are indicated in astronomical units (AU) on a logarithmic scale. White dots mark the ice line.

Second, GJ 876 is the only known exoplanetary system that contains both a gas giant and a smaller object consistent with the mass range of the ice giants (e.g., Uranus, Neptune, GJ 436 b), such that the more lightweight planet (GJ 876 e) orbits farther from the host star than the heavier planet (GJ 876 b). This distinction makes GJ 876 the only exoplanetary system whose overall architecture mimics the layout of our own Solar System, where we observe three concentric zones: the innermost, occupied by terrestrial planets; the middle zone, occupied by gas giants; and the outer zone, occupied by ice giants.

Third, the system's three outer planets lie in a Laplace resonance, also known as a three-way mean motion resonance, and they are further engaged in a secular resonance. This unusual orbital behavior has the potential to reveal a great deal about the system's evolutionary history. Several objects in the Solar System are also engaged in mean motion and secular resonances, including three of Jupiter’s largest moons (which similarly exhibit a Laplace resonance) and the planet/dwarf planet pair Neptune and Pluto (which orbit in a 3:2 mean motion resonance). In the case of GJ 876, the Laplace resonance means that the outermost planet, “e,” completes one orbit for every two orbits of the third planet, “b,” which in turn completes one orbit for every two orbits of the second planet, “c.” In addition, whenever planet e arrives at its nearest approach to the central star (i.e., periastron), it is also closely aligned with the periastra of planets b and c (Rivera et al. 2010). Various authors have described this behavior in terms of “apsidal alignment” and “apsidal corotation” (Laughlin et al. 2005, Beauge et al. 2008).

GJ 876 was the first exoplanetary system in which a mean motion resonance was demonstrated, and it remains the best-studied example of this behavior outside our Solar System. The dynamic relationship shared by its planets could not exist if they simply formed at their present locations, without any shared orbital evolution (Lee & Peale 2002). This system therefore furnishes excellent evidence for the hypothesis of Type II migration, which argues that gas giant planets are subject to incremental displacement from larger to smaller semimajor axes through interactions with the primordial gas nebula.

The system’s resonant architecture also implies that the participating planets perturb each other’s motion so strongly that their orbits change rapidly over time (Laughlin & Chambers 2001). Continuing observations therefore result in ever more precise determinations of key system parameters. Now that investigations of GJ 876 have extended over two decades, using increasingly sensitive instrumentation, our understanding of this system has improved to the point where several teams have tried to measure the inclination of the planetary orbits against the plane of the sky. Nevertheless, agreement on the correct angle remains elusive, with estimates ranging from 48 degrees to 59 degrees (Rivera et al. 2005, Bean et al. 2009, Correia et al. 2010, Rivera et al. 2010). Once this value is conclusively established, astronomers will be able to calculate the true masses of the system’s planets, as they can already in those rare systems where we observe planets transiting across the face of their host stars.

four planets

The four detected planets of GJ 876 orbit close to the central star in a region that extends from about 0.02 AU outward to 0.33 AU – a span of about 46 million kilometers (29 million miles), smaller than the average separation of Mercury from the Sun. All four planets most likely originated at greater distances and then migrated inward to their present locations. As is customary for extrasolar planets, their alphabetic designations (b, c, d, e) indicate the order of their discovery, rather than their relative distance from the host star.

In 2010, two major studies – led respectively by Alexandre Correia and Eugenio Rivera – presented new values for many of the system’s basic parameters. While the two studies overlap considerably, they diverge on details, and neither analysis can be regarded as definitive. Their principal disagreement relates to the orbital inclinations proposed for the system’s most massive planets. Correia’s group defines this angle as about 48 degrees, whereas Rivera’s group finds 59 degrees. Accordingly, values from both studies are cited below.

1. The innermost planet, GJ 876 d, has a semimajor axis of only 0.021 AU and an orbital period of less than two days. Its minimum mass of about 6 Mea falls in the range of the Super Earths (2-10 Mea). In a system with so many firsts, planet “d” represents the earliest detection of such a low-mass object. Assuming that it orbits in the same plane as the two outer planets, its actual mass will be 6.83 Mea (Rivera et al. 2010). One likely outcome of its proximity to the host star may be a rotational period that has become tidally locked, so that one hemisphere always faces the star and the other lies in perpetual shadow.
   
   Given its star-grazing orbit, planet d’s effective temperature is expected to fall between 430 and 650 K. That would place its likely surface temperature in the range of our own Mercury and Venus, depending on planet d’s internal structure, atmospheric characteristics (if any), and true rotational period. Mercury, the innermost planet of the Solar System, is an airless sphere of metal and rock with a mean surface temperature of 445 K. Venus, although almost twice as far from the Sun, has a still higher surface temperature of 735 K on account of its dense atmosphere and runaway greenhouse effect.
   
   Although most planets detected so close to their central stars have circular orbits as a result of stellar tides, GJ 876 d has a notable eccentricity, estimated by Correia as 0.14 and by Rivera as 0.21. This lopsided motion may imply that the planet is subject to heating through gravitational stress, over and above its irradiation by the host star. In addition, such eccentricity opens the possibility that the planet’s rotation has escaped tidal locking and achieved a spin-orbit resonance like Mercury, with rotates 3 times for every 2 orbits of the Sun (Correia & Laskar 2004).
   
   The regime of Super Earths, of which GJ 876 d is the classic instance, includes at least two basic planetary types: objects composed almost entirely of rock and metal, like Venus and Earth, and objects with a rock/metal core and an extensive mantle of high-pressure phases of ice, like scaled-up versions of Ganymede and Titan. To date, transit studies have identified one likely example of each type among the known Hot Super Earths: the rocky planet CoRoT-7b (Queloz et al. 2009) and the icy planet GJ 1214 b (Charbonneau et al. 2009). Although numerous investigations have attempted to assign GJ 876 d to one class or the other (e.g., Lecavelier des Etangs 2006, Valencia et al. 2007), no conclusive constraints have yet been established.
   
   
2. The second planet, GJ 876 c, has a semimajor axis of 0.13 AU and an orbital period of about 30 days. Its orbital eccentricity of about 0.26 is unremarkable among exoplanets, but still higher than that of its other planetary companions. At its widest separation from the host star (i.e., apastron), planet c is more than 70% farther away than at periastron, suggesting notable temperature variations over the course of a single orbit.
   
   Planet c’s mass clearly falls within the gas giant range. Correia's group calculates its true mass at 0.83 MJUP, while Rivera’s group argues for 0.71 MJUP. In either case, planet c is intermediate in mass between Saturn and Jupiter, and significantly below the median for the full population of extrasolar gas giants (~1.7 MJUP). Given its proximity to the central star, this second planet almost certainly rotates more slowly than the Solar System’s gas giants, both of which have rotational periods of about 10 hours. However, planet c’s eccentric orbit and mean motion resonance with planet d may have prevented it from spinning down to a tidal lock. Conceivably, it may retain a rotational period of a few days to a few weeks.
   
   
3. The third planet, GJ 876 b, has a semimajor axis of 0.208 AU, an orbital period of about 61 days, and a small eccentricity of 0.03. However, considering the large eccentricity of its inner companion, planets b's periastron (0.202 AU) is less than 6 million km (3.6 million mi) wider than planet c’s apastron (0.163 AU). By comparison, Earth's periastron is about 42 million km (26 million miles) wider than that of Venus.
   
   Planet b’s minimum mass has been defined as 1.93 MJUP. Correia's group calculates its actual mass at 2.64 MJUP, while Rivera's group proposes 2.28 MJUP. It remains unknown whether the planet’s rotation has become tidally locked, given its likely orbital history and present Laplace resonance. Nevertheless, tidal locking is the default assumption for all short-period exoplanets.
   
   
4. The fourth planet, GJ 876 e, was announced by Rivera and colleagues, who define its actual mass as approximately 14.6 Mea, making it comparable to Uranus (14.5 Mea). At a semimajor axis of 0.33 AU, this likely ice giant completes an orbit in 124 days, with a minimal eccentricity of 0.06 (a bit higher than that of Uranus itself). Conjecture regarding potential rings and moons, as well as any constraints on the planet’s rotational period, will have to wait for future studies of the system’s dynamical evolution.

The orbits of the two giant planets, b and c, approximately coincide with the system's habitable zone (Rivera & Haghighipour 2007), indicating that both may be cool enough to support water clouds in their dense atmospheres. The gravitational influence of the host star might have prevented either planet from maintaining a system of satellites or rings. However, one study suggests that GJ 876 b might sustain moons even as massive as Earth (Barnes & O'Brien 2002). Such heavy companions would be conceivable only if they were captured by the host planet, since theories of satellite formation place strict constraints on the maximum size of co-formed moons (Canup & Ward 2006). Objects in the mass range of the Galilean satellites of Jupiter (perhaps up to the mass of Mercury) may be more likely -- if indeed any moons at all are possible.

Two studies over the past several years have identified a region of dynamic stability just inside the orbit of GJ 876 c, where a terrestrial-mass planet might survive after being captured in a 2:1 mean motion resonance with planet c and a 4:1 resonance with planet b (Laughlin et al. 2005, Correia et al. 2010). With a proposed semimajor axis of 0.08 AU and an orbital period of 15 days, this hypothetical planet might be responsible for the unexpected eccentricity of GJ 876 d through gravitational perturbations of the innermost planet's orbit (Correia et al. 2010). Correia's group notes that such a planet would be detectable by a dedicated search using the most sensitive of currently available technologies.

from dusty clouds to resonant giants

The aggregate mass of the system’s four planets (about 3.5 MJUP) substantially exceeds that of the orbiting bodies of the Solar System (less than 1.5 MJUP), even if we include Pluto and all 150+ moons. In addition, the two most massive planets around GJ 876 show clear evidence of having reached their current orbits through Type II migration. These clues indicate that the early environment of the GJ 876 system was highly "planetic" – that is, unusually favorable to planet formation.

The star’s enhanced metallicity most likely resulted in a generous endowment of free-floating solids in the primordial nebula, which in turn was probably inordinately massive, since heavyweight dust clouds are predicted to yield heavyweight planets (Thommes et al. 2008). As Greg Laughlin and colleagues concluded, "Standard core accretion theory predicts that systems such as GJ 876 are drawn from the extreme high-mass end of the circumstellar disk mass distribution, and will thus be intrinsically rare" among M dwarf systems (Laughlin et al. 2004).

The three resonant giants probably formed beyond the system ice line, perhaps in the vicinity of 1 AU. The first to assemble may have acted as a stimulus to the others, and all three most likely underwent some scenario featuring convergent migration in order to arrive at their present orbits (e.g., Lee & Thommes 2009). The innermost planet most likely accreted as a consequence of the inward migration of planet c. As it swept through the inner nebula, this giant's gravitational perturbations forced rocky planetesimals in the inner disk to clump inside its orbit's shrinking radius, where they underwent ejections, collisions, and accretion (Fogg & Nelson 2005, Raymond et al. 2006b, Mandell et al. 2007). The ultimate result was the Hot Super Earth that we now detect.

Last update June 2010