≡ Menu

Three Regimes of Planet Formation

On Tuesday I mentioned the work of Lars A. Buchhave, an astronomer at the Harvard-Smithsonian Center for Astrophysics (CfA), in connection with the Kepler-10c discovery. The latter is the so-called ‘mega-Earth’ now found to be seventeen times as massive as our own planet, with a diameter of about 29,000 kilometers. A larger population of solid planets with masses above 10 times that of Earth was suggested in the Kepler-10c paper (see Introducing the ‘Mega-Earth’ for more on this), with reference to Buchhave’s ongoing work.

Let’s take a closer look at what Buchhave is doing, because the intriguing fact is that planets four times the size of Earth and smaller comprise about three-quarters of the planets found by the Kepler mission. How large a role does the metallicity of the host star play in planet formation?

At the ongoing meeting of the American Astronomical Society in Boston, Buchhave explained his research methods, which involve measuring ‘metals’ — in astronomical parlance the elements heavier than hydrogen and helium — in stars hosting exoplanets. His team analyzed more than 2000 high-resolution spectra of Kepler Objects of Interest, yielding the metallicities and other parameters of 405 stars orbited by 600 exoplanet candidates.

A statistical examination of the results followed, with this outcome: Planet-hosting stars fall into three groups that can be defined by their compositions. Buchhave’s team found two dividing lines, one at 1.7 times Earth’s radius, the other at 3.9 times the radius of Earth. The inference is that planets smaller than 1.7 Earth radius are completely rocky, while those above 3.9 Earth radius are most likely gas giants. [Addendum: See kzb’s comment below re a mistake I made in the initial version of this post.]

That interesting region between 1.7 and 3.9 times the size of Earth is where we find the so-called ‘gas dwarfs,’ planets whose cores accreted gas from the protoplanetary disk but failed to grow into gas giants of Jupiter-class or larger. Says Buchhave:

“It seems that there is a ‘sweet spot’ of metallicity to get Earth-size planets, and it’s about the same as the Sun. That makes sense because at lower metallicities you have fewer of the building blocks for planets, and at higher metallicities you tend to make gas giants instead.”

From the paper:

…the observed peak in the metallicity–radius plane at 1.7R suggests that the final
mass and composition of a small exoplanet is controlled by the amount of solid material available in the protoplanetary disk. A higher-metallicity environment promotes a more rapid and effective accretion process, thereby allowing the cores to amass a gaseous envelope before dissipation of the gas. In contrast, lower-metallicity environments may result in the assembly of rocky cores of several Earth masses on timescales greater than that inferred for gas dispersal in protoplanetary disks (<10 Myr), yielding cores without gaseous hydrogen–helium atmospheres.

To produce small, terrestrial worlds, then, stars with metallicities similar to the Sun are favored, while stars with gas dwarfs are likely to be those that are slightly more metal-rich. The stars most likely to produce gas giants contain the most metals, generally fifty percent more than the Sun. The finding is intriguing but comes with the caveat that Kepler is best at finding planets relatively close to their star, and the metallicity thesis needs to be tested over a wider range of orbits.


Image: Three kinds of planetary outcome are suggested by the metallicity of the host star, according to new work by Lars A. Buchhave and team. Credit: David Aguilar, CfA.

Whether or not Buchhave’s work can explain a Kepler-10c will depend upon gathering a larger sample of such worlds to work with. But so far his analysis implies that there is no evident cutoff in size for rocky worlds, the data showing that the mass and radius indicating the transition from rocky to gaseous planets should increase with orbital period:

Although additional data are required to confirm this relationship, the fit is apparently consistent with a critical core mass that increases with orbital period and an atmospheric fraction of 5%… If correct, this predicts the existence of more massive rocky exoplanets at longer orbital periods.

Thus the farther a planet is from its star, the larger it can grow before the accretion of a thick atmosphere turns it into a gas dwarf. When we have the observational tools to examine a wide range of planets in outer-system orbits, we may be able to confirm or refute Buchhave’s suggestion that truly massive ‘super-Earths’ like Kepler-10c may not be uncommon.

The paper is Buchhave et al., “Three regimes of extrasolar planets inferred from host star metallicities,” Nature 509 (29 May 2014), pp. 593-595 (abstract / preprint).


Comments on this entry are closed.

  • Michael June 5, 2014, 13:21

    The idea that higher ‘metal’ stars indicates a predisposition towards large gaseous worlds may need to take into account that these large massed planets in lower orbital periods could have pushed smaller rocky (high metal) planets into the star polluting their atmospheres skewing the results. I wonder if they took that into account. There would need to be a large spectrum of planets in a spread of orbital periods to tease out the effect, I am not sure 405 samples would be enough. The stars convection process and age would also need to be taken into account.

  • kzb June 5, 2014, 13:22

    Paul, have you put mass instead of radius in the fourth paragraph?

  • RobFlores June 5, 2014, 14:02

    The blind spot that Kepler spacecraft had, is exactly the information we
    need to come to more general conclusions about planet formation.
    Remind me if someone knows, would a 2 times more sensitive camera/telescope have saved they day or, it this method just too inherently noisy for what we are looking for, Terrestrial planets around sun like stars.?

  • william June 5, 2014, 14:14

    Scientists Say ‘Printing’ Humans May Be The Best Way To Conquer Space

    From pizza to prostheses, 3-D printers are being used to whip up all sorts of things. And now scientists are talking about “printing” out batches of people to colonize outer space.

    Sounds wacky, but these guys are serious.

    “Our best bet for space exploration could be printing humans, organically, on another planet,” Adam Steltzner, lead engineer on NASA’s Curiosity rover mission, said at a futurist conference held this month in Washington, D.C.

    After all, scientists including Stephen Hawking believe our very survival depends on “escaping our fragile planet” and colonizing other planets. Of course, landing humans on other planets is no simple task. A short hop to nearby Mars could take up to 300 days and cost over $6 billion. Once we got there, if we were to hit the red planet’s atmosphere at the same speed that the Curiosity rover did, our retinas would detach from our eyeballs. Yikes.

    Instead, why not just seed the galaxy with tiny organisms designed to recreate our species? Here’s how that might work.

    Scientists already know that microbes can survive long stints in space. In fact, some scientists theorize that alien microbes hitched a ride aboard comets or meteorites and brought life to Earth.

    Based on that idea, some biologists believe it’s possible to send bacteria to terraform a planet — make its environment hospitable for human life. The bacteria would also be encoded with human DNA.

    AOL Ad

    “It’s sort of like an iPod that you send to another planet. And the bacteria can store information very densely,” Gary Ruvkun, a biologist at Harvard University, told Motherboard.

    Ruvkun said he believes we’ll have the technology to store the human genome in bacteria within a decade or two. The trickier part is programming instructions into the bacteria that will tell them what to do once they reach their destination.

    “If we could also send along assembly instructions, for the bacteria to produce an array of descendent organisms that assemble the genome segments over some time period into a human, it is a way to ‘print’ humans remotely,” Ruvkun told The Huffington Post in an email.

    Just as the human egg cell is programmed by our DNA to divide, replicate and develop into a human, so bacteria could be programmed by our DNA to do the same thing, Ruvkun explained.

    “This is far beyond our ability to program bacteria,” he added. “Now. But 1,000 years from now, we will be able to do it. One thousand years is a blink in a 4 billion year timescape.

    Just a blink. And humans are relatively young in the vast timeline of our universe. If you let your mind run wild, you might even wonder whether we are the product of tiny bacteria someone else programmed to colonize Earth, Ruvkun said.

    “Perhaps Earth was terraformed in this way,” Ruvkun said in the email. “More likely, we are a big mistake and the cute little puppy dogs that should have dominated the Earth have been trumped by a glitch called humans.”

  • Leon Froess June 5, 2014, 14:19

    Maybe this is a question Paul can find someone to answer.

    I have wondered at what surface gravity our chemical rockets would not have enough propulsive force to launch a payload into orbit? A mega earth planet would have a high surface gravity and probably a thicker atmosphere as well which means a probe would probably have a higher minimum orbit distance which would take even more energy to launch a satellite into orbit.

    Could an intellegent species be “planet locked” because chemical rockets could never achieve orbit due to the high surface gravity of the planet they evolved on?

  • Paul Gilster June 5, 2014, 16:23

    kzb writes:

    Paul, have you put mass instead of radius in the fourth paragraph?

    Glad you caught that! Asleep at the switch… I’ve made the correction.

  • Chris June 5, 2014, 17:02

    Question – Could an intellegent species be “planet locked” because chemical rockets could never achieve orbit due to the high surface gravity of the planet they evolved on?

    Answer – Interesting question. I think it is easy to fall into the mindset that chemical propulsion is the only means of attaining orbit (see startram, space elevator, scramjet and fusion propulsion). If a civilization on one of these worlds with a larger gravitational well has attained a similar level of technological sophistication as ours, they would currently be “planet-locked”. However, this is no different than our civilization 75-100 years ago when low performance solid rocket propulsion was the only option. Indeed, given our future propulsion options to LEO, we may evolve beyond chemical options. My point is that such “planet-locked” problems are overcome by technology in the relative blink of cosmic time.

  • Dave Moore June 5, 2014, 17:23

    The most intriguing relationship to me is that the further a planet forms from its star the less likely it is to have a massive atmosphere. Intuitively, you would assume that the closer to a star a planet is the less atmosphere it would have as the more likely the planet would be to loose its atmosphere through thermal escape.

    I can’t remember offhand the relationship between planetary formation speed and orbital radius, but I know that it drops off markedly at least by the power of two with distance.

    What this relationship implies is that the speed of formation determines whether a super Earth becomes a rocky giant or a gas dwarf. If it coalesces before the gaseous nebula disperses then it becomes a gas dwarf. If the nebula disperses before the planetoids can coalesce into a planet then you get a rocky giant.

    If a planetary nebula is carbon rich, then the surplus carbon will condense out of the inner nebula as graphite. Graphite’s high melting point will mean that there will be solids in the nebula almost all the way down to the nascent star’s surface. This will result in the rapid formation of close planets, which will become gas dwarfs. (55 Cancri is the poster child of this.)

    All other things being equal, in an oxygen rich nebula (like our solar nebula) will have no solids until about 0.3 au out, so any planets that do form will tend to coalesce after the Hydrogen nebula has dispersed, so you will tend to get rocky giants.

  • Alex Tolley June 5, 2014, 18:14

    The bacteria would also be encoded with human DNA.

    I doubt it. Human DNA is huge in comparison to bacterial DNA, and it has to fit into a much smaller bacterial cell.
    e.g. bacteria genomes ~ 1E4 base pairs.
    human genome 3E9 bp.

    So you are talking 100,000x larger. Then bacteria have to replicate it for each division. The energy and resources alone would slow division to a crawl. Bacterial replication is not as accurate as eukaryote replication, so that could be a problem too.

    It seems to me that someone thought that because the TAQ polymerase reaction will replicate eukaryote DNA, that bacteria can be induced to replicate whole genomes internally.

  • Brian Swiderski June 6, 2014, 1:16

    Given the amazing visual diversity of solid objects in our own solar system alone, I can’t wait for the day when we get some decent imagery of a super- or mega-Earth.

  • andy June 6, 2014, 2:15

    So terrestrial planet formation works best around low-metallicity stars, and Kapteyn’s star strongly suggests that low-mass HZ planets were formed early on, a couple of billion years after the Big Bang. So there goes one possible resolution of the Fermi paradox: that the conditions for forming suitable planets only arose in the recent past.

    Discounting hypotheses such as cosmic conspiracies that are enforced over unfeasibly large volumes of space to keep everyone quiet, we seem to have left either that technological civilisation is a vanishingly rare outcome of the evolution of life, or that technological civilisation tends to be short-lived and usually never gets beyond the home planet.

  • tesh June 6, 2014, 3:13

    @ william June 5, 2014 at 14:14

    Call it vanity on my behalf but the journey to new worlds should be made by humans as we are now and not by some digitised version or a billion year “evolved” version of our genetic material. All of this 3-D printing, nano-tech and wishing for wormhole tech to suddenly materialise are just either buzz words for grant grabbing or waiting for a knight in shining armour to whisk us off our feet and carry us to the aforementioned new worlds.

    In reality we are about as far from printing a human as we are from having WARP drive. It is just a lot (A LOT) of hot air.

  • kzb June 6, 2014, 7:45

    I must admit I find this a difficult paper to understand ! Can someone explain in layman’s terms what exactly is plotted in the top panel of figure 1 ?

    Also is this a fair point: they are doing statistical analysis on what is an observationally biased sample. It’s an incomplete sample and, worse, it’s incomplete in a systematic way ?

    Next point, as Michael says, we had a paper the other week about planets falling into the star and affecting the most common metallicity indicator, [Fe/H]. So the measured metallicity can be a RESULT of planet formation not the cause. Does that throw a spanner in the works of studies like this?

  • ljk June 6, 2014, 23:43

    June 06, 2014

    “Intelligent Life May Be in Its Early Stage in the Universe” (Today’s Most Popular)

    THe 200 billion galaxies in our oberservable Cosmos show a clear potential to continue on as we see them today for hundreds of billions of years, if not much longer. Because planets and life are so young in the Universe, says Harvard’s Dimitar Sasselov, perhaps “the human species are not late comers to the party. We may be among the early ones.”

    That may explain why we see no evidence of “them” and may go a long way to explaining the famous Fermi Paradox, which asks if there’s advanced intelligent life in the Universe, where are they? Why haven’t we discovered any evidence of their existence?

    The story of the Universe according to Sasselov in is new study, The Life of Super-Earths, looks like this: generations of stars made enough iron and oxygen, silicon and carbon, and all the other elements from the original hydrogen and helium about 13 billion years ago to be able to form the Earth we live on and the planets the Kepler Mission is discovering today.

    Full article here:


    To quote:

    Sasselov concludes that the statistical argument for Fermi’s Paradox “holds true only if the timescale for the emergence of life is much shorter than the age of the universe, but not so if the two are comparable.” The future for life in the Universe looks excellent, says Sasselov.

    Planets may be just a tiny fraction of the Universe because of their small size, but there are so many of them that the probability of life grows exponentially. The Universe is passing through the stelliferous era –its peak of star formation–but appears to be still peaking in its formation of planets. There are more stars in the Universe than there are grains of sand on Earth and there are an equal number of planets.

  • Adam June 7, 2014, 0:16

    Alex Tolley – the smallest free-living bacteria, the aptly named Pelagibacter ubique, has a genome that’s 1,308,759 bp long. Parasitic mycobacteria have genomes as small as 582,970 bp. That’s as small as a self-replicator goes. Smaller exo-/endo-parasitic bacterial cells exist with smaller genomes, but nothing below ~150 Kbp, and none known that can live without their hosts – i.e. they’re on their way to becoming endo-somes like mitochondria, which have lost most of their genes to their ‘host’.

    At the other end of the scale, for bacteria, there are some Monster cells – the Epulopiscium fishelsoni symbiote carries 200,000 copies of its genome and thus has many times the amount of DNA in human cells. Thus giga-Genomes in bacteria aren’t unthinkable, though unless it’s expressed and/or functionally vital, DNA replication’s natural noisiness will mutate the genomes into new genes or pseudogenes before they can ‘evolve’ into higher lifeforms to be expressed.

  • Rob Henry June 7, 2014, 18:10

    @Adam@Alex Tolley, you said much of what I was going to, but I want to add that those bacteria can hold onto that ‘noise’ far longer if they don’t occupy the niches of easily cultured bacteria (that might include virtually every well know & studied bacteria, but it is only 0.1% of their species), and thus are less subject to purifying selection.

    @andy, you seem to have left from your analysis the probability of abiogenesis itself. The idea that life on Earth evolves to the conditions has been debated among biologists for two and a half millennia, but Darwin added a new twist. To him there could be no new trait through evolution, save by modifying a preexisting one. It was a revolution, but left one problem – the origin of the first organism. Much worse still, of all life’s functions and properties, non-trivial selfreplication is the only one that humankind can’t even design a model for (at least not in this world. We have found proofs that there design is possible within mathematical constructs that were themselves designed to allow this to be easier).

    Don’t believe this ‘just add water’, school of thought for abiogenesis. Its not just another assumption, its one that runs contrary to all we currently know, its sole advantage being that it fits better with Copernican thinking.

  • ljk July 19, 2014, 12:08

    Distant Stellar Atmospheres Shed Light on How Jupiter-like Planets Form

    by SHANNON HALL on JULY 18, 2014

    It’s likely that Jupiter-like planets’ origins root back to either the rapid collapse of a dense cloud or small rocky cores that glom together until the body is massive enough to accrete a gaseous envelope.

    Although these two competing theories are both viable, astronomers have, for the first time, seen the latter “core accretion” theory in action. By studying the exoplanet’s host star they’ve shed light on the composition of the planet’s rocky core.

    “Our results show that the formation of giant planets, as well as terrestrial planets like our own Earth, leaves subtle signatures in stellar atmospheres”, said lead author and PhD student Marcelo Tucci Maia from University of São Paulo, Brazil, in a press release.

    Full article here:


  • ljk July 30, 2014, 17:33

    Early Earth: A Battered, Hellish World with Water Oases for Life

    By Charles Q. Choi, Space.com Contributor | July 30, 2014 01:01 pm ET

    Asteroids and comets that repeatedly smashed into the early Earth covered the planet’s surface with molten rock during its earliest days, but still may have left oases of water that could have supported the evolution of life, scientists say.

    The new study reveals that during the planet’s infancy, the surface of the Earth was a hellish environment, but perhaps not as hellish as often thought, scientists added.

    Earth formed about 4.5 billion years ago. The first 500 million years of its life are known as the Hadean Eon. Although this time amounts to more than 10 percent of Earth’s history, little is known about it, since few rocks are known that are older than 3.8 billion years old.

    Full article here:


    To quote:

    These findings suggest that Earth’s surface was buried over and over again by large volumes of molten rock — enough to cover the surface of the Earth several times. This helps explain why so few rocks survive from the Hadean, the researchers said.

    However, although these findings might suggest that the Hadean was a hellish eon, the researchers found that “there were time gaps between these large collisions,” Marchi said. “Generally speaking, there may have been something on the order of 20 or 30 impactors larger than 200 km (120 miles) across during the 500 million years of the Hadean, so the time between such impactors was relatively long,” Marchi said.

    Any water vaporized near these impacts “would rain down again,” Marchi said, and “there may have been quiet tranquil times between collisions — there could have been liquid water on the surface.”

    The researchers suggested that life emerging during the Hadean was probably resistant to the high temperatures of the time. Marchi and his colleagues detailed their findings in the July 31 issue of the journal Nature.

  • ljk July 30, 2014, 17:38

    How the Moon Got Its Lemon Shape

    By Mike Wall, Senior Writer | July 30, 2014 01:00 pm ET

    Earth’s powerful gravity tugged the moon into its oddball shape long ago, shortly after both bodies formed, a new study suggests.

    Tidal forces exerted during the early days of the solar system can explain most of the moon’s large-scale topography, including its slight lemon shape, reports the study, which was published online today (July 30) in the journal Nature.

    The new findings could help scientists tackle longstanding lunar mysteries, such as why the moon’s near side is dominated by dark volcanic deposits, while the far side is not, researchers said.

    Full article here:


    To quote:

    The new results could have applications beyond the moon, potentially helping researchers better understand any celestial body strongly affected by tidal forces, Garrick-Bethell said.

    “This idea was inspired by Europa,” he said, referring to the huge moon of Jupiter. Europa is similar today to Earth’s moon long ago, he added, in that it harbors a solid shell (of ice in Europa’s case rather than rock) sitting atop an ocean layer (which consists of liquid water rather than magma).

    The study could even yield insights about the evolution of faraway alien planets, Garrick-Bethell said.

    “Tides are so ubiquitous; they’re everywhere across the galaxy,” he said. “So understanding tidal processes is always important.”