The idea that the composition of a star and its rocky planets are connected is a natural one. Both classes of object accrete material within a surrounding gas and dust environment, and thus we would expect a link between the two. Testing the hypothesis, researchers from three institutions — the Instituto de Astrofísica e Ciências do Espaço (Portugal), the NCCR PlanetS project at the University of Bern, and the University of Zürich — have confirmed the concept while fine-tuning the details. After all, we still have to explain iron-rich Mercury as an outlier in our own Solar System.

Image: Mercury has an average density of 5430 kilograms per cubic meter, which is second only to Earth among all the planets. It is estimated that the planet Mercury, like Earth, has a ferrous core with a size equivalent to two-thirds to three-fourths that of the planet’s overall radius. The core is believed to be composed of an iron-nickel alloy covered by a mantle and surface crust. Credit: NASA.

Starlight contains the spectroscopic signature of the star’s composition, but because we have directly imaged few planets, the composition of rocky planets has to be inferred by examining their mass and radius. A significant factor in this study is what is known as the Bern Model of Planet Formation and Evolution, which covers quite a bit of ground, from processes in the protoplanetary disk, accretion models of a planet’s growing core, and the eventual gravitational interactions of young planets. The authors apply the model in estimating the iron mass fraction of rocky exoplanets.

Christoph Mordasini (University of Bern), a co-author of the paper on this work, comments on the method:

“…since stars and rocky planets are quite different in nature, the comparison of their composition is not straightforward. Instead, we compared the composition of the planets with a theoretical, cooled-down version of their star. While most of the star’s material – mainly hydrogen and helium – remains as a gas when it cools, a tiny fraction condenses, consisting of rock-forming material such as iron and silicate.”

The researchers, led by Vardan Adibekyan (Instituto de Astrofísica e Ciências do Espaço), chose the planets for their study from an initial cut of 364 worlds orbiting F, G and K-class stars. They then narrowed the list to 56 planets with the highest precision in mass and radius, excluding planets whose masses had been determined by transit-timing variations because these results can differ from mass determined by radial velocity methods. They then whittled their list down to 22 potentially rocky planets with radii less than twice that of Earth in 21 stellar systems.

While the analysis confirms that the composition of terrestrial-class worlds is linked to the composition of the host star, the abundance of planetary iron can be higher than what is found in the star. The correlation exists but not precisely in a 1:1 ratio. The implication: Planets in formation may shed lighter materials while leaving dense iron behind. The paper identifies five planets (K2-38 b, K2-106 b, K2-229 b, Kepler-107 c, and Kepler406 b) with a higher iron content than the rest, all seemingly higher-mass analogs of Mercury as planets with Earth-like composition but higher mass.

The likely formation and evolution of these ‘super-Mercuries’ demands investigation, and early system collisions alone may not suffice: From the paper:

The five super-Mercuries we identify have a wide range of masses, unlike the concentration around ~5 M? predicted by simulations of giant impacts. We suggest that a giant impact alone is not responsible for the high density of super-Mercuries. Planet formation simulations that incorporate collisions are unable to produce the highest-density super-Mercuries.

If not collisions, then what? All five of the super-Mercuries found in the study orbit stars with high iron abundance, which the authors consider a proxy for the overall content of heavy elements in stars:

The first trend may suggest that the mechanism responsible for the overabundance of iron in these planets is related to the composition of the protoplanetary disk. The second trend could imply a more efficient planet formation, leading to a formation of multiple planets and resulting in frequent collisions. We suggest that both iron enrichment and collisional mantle stripping may need to be invoked to produce an iron enrichment in the general planet population and explain the presence of super-Mercuries.

The findings of the paper regarding the correlation between planet and star in terms of iron abundance remain significant even if the five super-Mercuries are removed from the sample. Thus the iron mass fraction, computed for planets through their mass and radius, suggests that distinct populations of super-Earths and planets like Mercury can occur, their composition reflecting factors involved in their formation. But the broader picture is that given that density is but one clue to composition, if the host star’s composition is a reliable marker we are justified in making inferences about the makeup of its planets.

The paper is Adibekyan et al., “A compositional link between rocky exoplanets and their host stars,” Science Vol 374, Issue 6565 (15 October 2021), pp. 330-332. Abstract.