The latest data from the crippled Kepler space telescope point to three distinct molds of exoplanets — rocky planets, gas dwarfs, and ice or gas giants — distinguishable based on the level of heavy elements in their host star’s atmosphere.

This artist’s conception shows a terrestrial exoplanet, a gas giant and a mid-sized gas dwarf. Image credit: J. Jauch.
This artist’s conception shows a terrestrial exoplanet, a gas giant and a mid-sized gas dwarf.
J. Jauch.

Astronomers using archival data from NASA’s crippled Kepler space telescope have found that the presence of three types of exoplanets — rocky planets, gas dwarfs, and ice or gas giants — around a star depends highly on the host star’s metallicity.

Lars Buchhave (Harvard-Smithsonian Center for Astrophysics and University of Copenhagen, Denmark) and colleagues published their findings in the May 29th Nature and announced the results June 2nd at the summer meeting of the American Astronomical Society in Boston.

When exoplanet discoveries first started trickling in, the majority of known worlds were hot Jupiters, sizzling gas giants circling close to their host stars. Once the exoplanet count passed a few dozen in the late 1990s a new relationship became clear: hot Jupiters more often circled stars with high levels of metals — the astronomer’s slang term for elements heavier than hydrogen and helium.

This wasn’t a huge surprise. The chemical fingerprints of those stars point back to the chemical makeup of the ancient disk of dust and gas that formed both the star and its orbiting hot Jupiter. In this disk, grains glom together until they ultimately form planets. Metals within the disk should quicken this process, allowing planets to build up cores and large gaseous envelopes before the disk dissipates.

And so the question arises: do small planets follow the same trend?

Until now the answer was no: small planets circle stars spanning a wide range of metallicities. But with very little data, the absence of a trend for small planets wasn’t set in stone.

So Buchhave and colleagues looked at a much larger dataset, by taking follow-up spectra of 405 stars orbited by 600 small exoplanet candidates. “We wanted to see if these host stars differed significantly from each other when we looked at [small] planets of different sizes,” says Buchhave.

Without any analysis there’s still a wide range of metallicities. “It’s not possible to pick out one system and from its metallicity predict what type of planets are orbiting its star,” says Bucchave.

But when Bucchave and colleagues took a statistical look, they were surprised to see two clear dividing lines, one at a radius 1.7 times Earth’s and the other at a radius 3.9 times Earth’s. The divisions separate the host stars into three metallicity groups, suggesting that three different populations of exoplanets exist, each roughly tied to different metallicities.

“When you examine metallicity on a large statistically significant sample, it is possible to see features in the metallicities of the host stars,” says Bucchave. “This indicates that although metallicity is only one of the many factors affecting planet formation, it does seem to play an important role in the outcome of planet formation.”

Despite the scatter, Bucchave and colleagues find the three metallicity regions are distinct from one another by 4.5σ. The rule of thumb in physics is that 1σ could be due to random chance or scatter, while 3σ counts as a true observation. In order to verify the Higgs Boson, particle physicists had to make a 5σ detection — so 4.5σ is pretty good.

The two clear transitions hint at changes in composition between these three planet populations.

In our own solar system we only have two distinct populations: terrestrial planets and gas or ice giants. But Kepler has discovered an abundance of exoplanets in the middle, including the so-called superEarths. “So we really want to know about these planets,” says Bucchave. “Are they rocky planets with a thin compact atmosphere like the Earth? Or are they rocky cores with some sort of extended hydrogen-helium envelope where there’s really no surface?”

For planets smaller than 1.7 times the size of Earth, the host star’s metallicity tends to be roughly solar. These planets are the terrestrial sort we’re familiar with: solid bodies made of rocky material, perhaps with an iron core and with only a compact atmosphere, says coauthor David Latham (Harvard-Smithsonian Center for Astrophysics).

At a slightly higher metallicity, stars are more likely to have planets between 1.7 and 3.9 times the size of Earth — dubbed gas dwarfs by the team — that have rocky cores and thick atmospheres of hydrogen and helium, of which there are no solar system equivalents.

Above 3.9 times the size of Earth are familiar gas and ice giants, which are more common around stars with an even higher metallicity (about 1.5 times solar), in agreement with previous studies.

The results suggest that there’s a metallicity sweet spot for terrestrial planets to form, roughly the same as the Sun’s. That bodes well for forming Earth-like planets around Sun-like stars.

Although Buchhave and colleagues successfully show three distinct exoplanet populations based purely on metallicity, they emphasize that multiple factors affect planet formation.

"Clearly the data are telling us that metallicity matters, but how exactly is still not entirely clear," says exoplanet expert Jarrett Johnson (Los Alamos National Laboratory). "It's now a challenge to piece together a complete, coherent picture of planet formation that explains the results of this work, among others."

The next step will be to use the upcoming Transiting Exoplanet Survey Satellite (TESS) to probe nearer and brighter transiting planets, allowing follow-up measurements and confirmation of these metallicity-versus-size results, says Latham.

L. Buchhave et al. “Three Regimes of Extrasolar Planet Radius Inferred from Host Star Metallicites,” Nature, May 28, 2014


Image of Anthony Barreiro

Anthony Barreiro

June 3, 2014 at 7:51 pm

This is an interesting finding. My background is in the social sciences, so it makes me happy to read about inferential population statistics being used to analyze exoplanets! We're starting to have a context for understanding our own solar system, both the ways it is similar to others and the ways it may be unique.

I'm curious if there are any ideas why our particular solar system has both terrestrial planets and gas giants. Among known planetary systems, how many systems have one, two, or all three kinds of planets?

You must be logged in to post a comment.

You must be logged in to post a comment.