Figure 1. A stellar nursery: IC 348, the nearest rich star cluster still embedded in its primordial cloud of gaseous hydrogen and dust. Located in the constellation Perseus at a distance of 320 parsecs (1040 light years), IC 348 occupies one end of an extensive complex of molecular hydrogen clouds that includes NGC 1333. At an age of 2 to 3 million years, IC 348 contains about 300 infant stars, ranging from dim red dwarfs to hot blue stars of spectral class B. Credit: NASA/Lucas Cieza
Back when the only planets ever discovered were the ones in our Solar System, astronomers thought they knew how planetary systems formed. Then came the news about extrasolar planets – first a trickle of oddballs and outliers, eventually a flood of rank and file candidates from the Kepler Mission and its successors. Theoreticians are still struggling to produce models that can explain the full range of system architectures and planetary types implied by observations.
The most widely accepted theory of planet formation is core accretion, which descends from the so-called nebular hypothesis proposed in the 18th century by Immanuel Kant and Pierre-Simon de Laplace. In different languages, these two philosophers argued that the young Sun was surrounded by a thin cloud of dust grains, which coagulated in vortices to form larger units called planetesimals (Planetesimale, planétésimaux), which in turn collided to form the known planets.
The 21st century variation on this hypothesis begins with a similar circumsolar nebula, now specified as a blend of hydrogen and helium (H/He) supporting a sprinkling of dust. The ambient cloud is typically called a protoplanetary disk or primordial nebula or some variation on those terms. Figure 2 is a photograph of such a structure at an approximate age of 1 million years.
Inside these clouds, sticky collisions among dust grains, pebbles, and planetesimals form larger objects known as protoplanets (or planetary cores or planetary embryos). These objects have masses ranging from Moon-like to Mars-like. Under appropriate conditions, protoplanets can continue to accrete mass by colliding and merging with other protoplanets. Beyond some threshold in the range of 1 to 5 Earth masses (Mea), a growing planet will acquire a H/He envelope, forming an object like Uranus or the puffy Super Earths discovered by Kepler. If an object continues stockpiling gas until its envelope rivals the mass of its solid core, runaway accretion will ensue, resulting in a massive gas-dominated planet like Jupiter.
Figure 2. The protoplanetary nebula around HL Tauri
This photograph by the Atacama Large Millimeter/submillimeter Array (ALMA) captures a series of bright rings and dark gaps in the protoplanetary disk surrounding HL Tauri, a newborn Sun-like star located at a distance of about 140 parsecs (456 light years). Various sources provide radii in the range of 90-120 astronomical units (AU) for the disk, which is about 1 million years old. These values are larger by a factor of 3 to 4 than the radius proposed for our Sun’s protoplanetary disk.
Most contemporary research on planet formation is based on the accretion scenario, which is currently available in two basic models. In one, disk-driven migration is a fundamental process governing system evolution. Pebbles, protoplanets, and planets change their radial position in the nebula over time, migrating either inward or outward by interacting with the flow of gases, and in turn interacting with each other, to produce planetary systems. The other approach favors in situ formation, arguing that growing planetary cores stay where they are and simply accrete material from their immediate surroundings. Subsequent posts in this series will explore each model.
the extrasolar bestiary
We know that planets come in a limited number of species, as determined by their mass and composition (Figure 3). The easiest to discover are the high-mass or gas giant planets, which are objects more massive than about 50 Mea with bulk compositions that are more than 50% H/He. Our Solar System has two beauties: Jupiter and Saturn.
Then come the low-mass planets, which can be divided into two sub-populations. Gas dwarfs are generally at least 2 Mea but less than 50 Mea. Their defining characteristic is a bulk composition that is less than 50% but at least 0.1% H/He. Terrestrial planets are devoid of gaseous H/He and generally less massive than 10 Mea. Our Solar System has two gas dwarfs (Uranus and Neptune) and four terrestrial planets (Mercury, Venus, Earth, and Mars).
Despite their diminutive size and difficulty of observation in other star systems, most humans are most interested in the terrestrials. The reason is simple. As far as anyone knows, life – or at least life that we would recognize as such – can evolve only on like Earth.
Figure 3. Planets across five orders of magnitude in mass
Planets are shown at their relative sizes. All have well-measured masses and radii. The numbers above the red line indicate the objects’ approximate mass in Earth units (Mea). In more exact terms, Mars is 0.11 Mea, Uranus is 14.5 Mea, and Saturn is 95 Mea. The numbers below the red line indicate the objects’ mass in Jupiter units or Mj. The image of WASP-10b is an artist’s impression; the other images are photographs or photo composites.
Data from the most successful search methods – radial velocity and transit observations – demonstrate that low-mass planets are the most numerous species orbiting within a few astronomical units of main sequence stars (astronomical unit = AU; 1 AU = distance between Earth and Sun). They occur preferentially in multiplanet systems and are observed in a variety of architectures. Most interesting are the high-multiplicity systems, meaning those with at least three planets. A representative sample of high-multiplicity architectures centered on Sun-like stars appears in Figure 4.
Figure 4. Distribution of mass in selected systems with at least 3 planets
Star masses are indicated in Solar units. Planet masses are indicated in Earth units
Within the high-multiplicity sample, we now know well over a hundred examples of compact low-mass systems consisting of three to six low-mass planets orbiting within 1 AU of stars with spectral types ranging from M to F (e.g., Kepler-20, HD 69830, and Kepler-62 in Figure 4). We even know of several compact mixed-mass systems that include gas giants orbiting alongside two or more low-mass planets (e.g., Kepler-90, WASP-47, Kepler-48, and HD 10180 in Figure 4). Both types of systems contain a much higher concentration of mass on short-period orbits than does our Solar System. Next to these robust ensembles, the inner Solar System seems anemic, with just four pint-size spheres and a swarm of rocky debris inside 5 AU, collectively totaling only 2 Mea.
In fact, after two decades in which the extrasolar census grew from a few dozen planets to a few thousand, we can confidently assert that our system is odd. Although many systems support a clutch of small planets on short-period orbits, and many others have gas giant planets orbiting outside 1 AU, relatively few support both configurations. Virtually none of these bear a significant resemblance to the Solar System. In a recent study, Morbidelli & Raymond (2016) estimated that only 1% of G-type stars like our Sun are accompanied by a gas giant on a circular orbit outside 1 AU. Planets like our Jupiter might be as rare as Hot Jupiters.
Another remarkable feature of our system is the hierarchical distribution of planetary mass, such that each planet exceeds the sum of the masses of all smaller planets. It’s unclear how common this architecture might be, since relatively few systems with four or more planets have well-constrained masses. Although mass hierarchies analogous to the Solar System are attested for 55 Cancri and Gliese 876, many other systems have flatter mass distributions. Among them, Figure 4 depicts Kepler-20, HD 10180, and Mu Arae. Another well-known example is Kepler-11. Compact multiplanet systems in particular appear to favor collections of similar-mass planets over hierarchical arrangements.
How do all these varied system architectures arise? And almost as important, why should we care?
Before addressing the difficult question of how, here are some reasons why:
- Existing data on exoplanetary systems are sparse, incomplete, and likely to remain so for decades. If we had a working model of the conditions and processes that produce the full range of system architectures, we could use available data to predict or rule out unseen companions of known objects.
- Knowing how and especially where any given planet formed provides an invaluable clue to its likely composition. Planets assembled from material available near the central star will be rocky, while planets formed from material with a more distant origin will contain a significant fraction of volatiles.
- Composition is a critical factor in habitability, since a planet’s structure and bulk constituents are associated with atmospheric properties, surface temperatures, and the potential for open bodies of water.
In sum: knowing how and where planets of various types form will help us predict the distribution of Earth-like planets in the Solar neighborhood and throughout our Galaxy.
(This discussion will continue with Protoplanetary Disks and In Situ Formation and conclude with Accretion with Migration in Radially Structured Disks.)