The Nature and Origins of Sub-Neptune Size Planets

Abstract

Planets intermediate in size between the Earth and Neptune, and orbiting closer to their host stars than Mercury does the Sun, are the most common type of planet revealed by exoplanet surveys over the last quarter century. Results from NASA's Kepler mission have revealed a bimodality in the radius distribution of these objects, with a relative underabundance of planets between 1.5 and 2.0 $R_{\oplus }$ . This bimodality suggests that sub-Neptunes are mostly rocky planets that were born with primary atmospheres a few percent by mass accreted from the protoplanetary nebula. Planets above the radius gap were able to retain their atmospheres ("gas-rich super-Earths"), while planets below the radius gap lost their atmospheres and are stripped cores ("true super-Earths"). The mechanism that drives atmospheric loss for these planets remains an outstanding question, with photoevaporation and core-powered mass loss being the prime candidates. As with the mass-loss mechanism, there are two contenders for the origins of the solids in sub-Neptune planets: the migration model involves the growth and migration of embryos from beyond the ice line, while the drift model involves inward-drifting pebbles that coagulate to form planets close-in. Atmospheric studies have the potential to break degeneracies in interior structure models and place additional constraints on the origins of these planets. However, most atmospheric characterization efforts have been confounded by aerosols. Observations with upcoming facilities are expected to finally reveal the atmospheric compositions of these worlds, which are arguably the first fundamentally new type of planetary object identified from the study of exoplanets.

Figure 1

Radii and orbital period for transiting planet candidates detected by Kepler from Thompson et al. (2018). The color of the points (the “Score”) is related to the estimated likelihood of being a true planet, where larger values indicate a higher likelihood. Most of the planet candidates that Kepler discovered are smaller than Neptune and are likely to be real planets. The area in the bottom right of the figure is mostly empty due to selection effects. See Winn (2018) for more on selection biases of exoplanet surveys.

Figure 2

Two views of Kepler's planet radius gap for sub‐Neptune size planets. In both cases the slope of the radius gap to smaller instellations and larger orbital periods matches the expectations of atmospheric mass loss being the key discriminant between the larger and smaller populations. Left: Planet occurrence as a function of planet size and instellation with host star radii derived from spectroscopy and distances. Two peaks in the distribution centered at 2.4 and 1.3 R _⊕ are visible in the data. The two lines represent expectations for models of the formation and evolution of these objects from Lopez and Fortney (2013). Figure taken from Fulton et al. (2017). Right: Precisely measured radii for individual planets from stellar characterization via asteroseismology. The lines present the best fit to the radius gap. This figure was originally presented as Figure 5 in Van Eylen et al. (2018).

Figure 3

Schematic cartoon of the expected outflow structure in a unified picture for hydrodynamic mass loss from close in exoplanets (top) and the three (continuously connected) expected mass‐loss regimes (bottom). The top panel shows the three layers to the planetary atmosphere. The bound atmosphere (yellow region) is where the hydrodynamic outflow is sub‐dominant. The region heated by cooling radiation from the planetary interior (red photons) and the stellar bolometric luminosity (green photons) has an intermediate temperature (blue/green region). Finally, the region heated by stellar XUV irradiation (blue photons) is a few thousand Kelvin or more (orange region). The mass‐loss regimes are shown from left to right as a function of increasing XUV luminosity (or decreasing cooling radiation). Core‐powered mass loss occurs when the sonic surface sits interior to the penetration of XUV photons, which thus do not affect the outflow (i). When the sonic point occurs in the XUV heated region, but the cooling/bolometric heated region is not thin, photoevaporation is enhanced due to the larger sub‐tended absorption area of the planet to XUV photons (ii), and finally when the cooling/bolmetric region is thin mass loss behaves as ‘classic’ photoevaporation (iii). Only scenarios (i) and (iii) have been calculated, and only for each in isolation. XUV, extreme ultraviolet.

Figure 4

The observed radius distribution of Kepler planets with orbital periods <100 days is shown as the gray histogram. The radius distributions predicted by the photoevaporation for different solid core compositions are shown as the colored lines. Lower density cores predict the radius gap to appear at higher radii. The observed radius distribution implies cores have densities consistent with an Earth‐like rock‐iron mixture (i.e., a model intermediate between the red and yellow models). More sophisticated models tightly constrain the silicate‐to‐iron ratio to be ∼3:1, that is consistent with Earth's composition (J. G. Rogers & Owen, 2020). Figure from Owen and Wu (2017).

Figure 5

Radius versus mass from a uniform analysis of small, highly irradiated planets. These planets should not have substantial gaseous envelopes, thus removing a degree of freedom from interior structure models. The data are tightly clustered around the Earth‐like composition line, suggesting a common composition for rocky planets. The higher density outlier K2‐229b could have a higher iron fraction from collisions (although note that current models struggle to create very iron‐rich planets from collisions; Scora et al., 2020), while the low density outlier 55 Cnc e could be the rare small planet with a significant volatile content or no core. Figure from Dai et al. (2019).

Figure 6

Orbital architectures of seven selected sub‐Neptune systems. Each system contains at least one sub‐Neptune planet with $2R_{\oplus }<R_p<4R_{\oplus }$, and systems are ordered vertically by the median planet size, from the largest (top) to smallest (bottom). The size of each planet is proportional to its measured physical size. The x axis is logarithmic such that the distance between neighboring planets in a given system is a measure of their orbital period ratio. For scale, the Mercury‐Venus orbital period ratio is 2.55 and the Earth‐Venus ratio is 1.63. Roughly half of Sun‐like stars have rich systems of close‐in sub‐Neptune size planets like those shown here.

Figure 7

Conceptual formation pathways for close‐in low‐mass planets.

Figure 8

Left: The transmission spectrum of the sub‐Neptune size planet K2‐18b (black points with errors) compared to models (solid lines and shaded confidence intervals). The increase in transit depth detected near 1.4 μm is identified as absorption due to water vapor. Right: Derived constraints on the cloud top pressure and the water vapor abundance from a retrieval analysis of the transmission spectrum. The solid black lines represent the 1 and 2σ confidence regions. Despite the degeneracy between the cloud top pressure and the water vapor abundance it is clear that the planet has a hydrogen‐dominated atmosphere due to the detectability of the spectral features. Future observations with the James Webb Space Telescope will be able to precisely constrain the atmospheric compositions of this and other similar planets and thus provide important constraints on their formation. Figures taken from Benneke et al. (2019).