≡ Menu

Small Planets, Ancient Star

Finding planets around stars that are two and a half times older than our own Solar System causes a certain frisson. Our star is four and a half billion years old, evidently old enough to produce beings like us, who wonder about other civilizations in the cosmos. Could there be truly ancient civilizations that grew up around stars as old as Kepler-444, a K-class star in the constellation Lyra that is estimated to be fully 11.8 billion years old? It’s a tantalizing speculation, and of course, nothing more than that. But the discovery of planets here still catches the eye.

The just announced discovery and accompanying paper are the work of Tiago Campante (University of Birmingham, UK), who led a large team in the investigation. What we learn is that five planets have been discovered using Kepler data around a star that is 117 light years from Earth. These are not habitable worlds by our standards — all five planets complete their orbits in less than ten days, making them hotter than Mercury.


Image: Kepler-444 is a recently discovered star with at least five Earth-size planets. The system is 11.2 billion years old. Illustration by Tiago Campante/Peter Devine.

Asteroseismology, which measures the oscillations caused by sound waves within the star as shown in minute brightness changes,, was a key part of this work, says Daniel Huber (University of Sydney), a co-author on the paper:

“When asteroseismology emerged about two decades ago we could only use it on the Sun and a few bright stars, but thanks to Kepler we can now apply the technique to literally thousands of stars. Asteroseismology allows us to precisely measure the radius of Kepler-444 and hence the sizes of its planets. For the smallest planet in the Kepler-444 system, which is slightly larger than Mercury, we measured its size with an uncertainty of only 100 km.”

We’ve found planets in low-metallicity environments before, such as the mini-Neptunes around Kapteyn’s Star in the galactic halo. This work takes the low-metallicity planet regime down to the size of terrestrial planets. All five of these planets are below Earth in size, with radius increasing with distance from the star, although three of them — Kepler 444c, Kepler-444d, and Kepler-444e — have similar radii comparable to the size of Mars. Kepler-444f is found to be between Mars and Venus in size. To place the find in perspective, here’s a figure from the paper that relates the Kepler-444 planets to other highly-compact multiple-planet systems.


Image: Semi-major axes of planets belonging to the highly-compact multiple-planet systems Kepler-444, Kepler-11, Kepler-32, Kepler-33, and Kepler-80. Semi-major axes of planets in the Solar System are shown for comparison. The vertical dotted line marks the semi-major axis of Mercury. Symbol size is proportional to planetary radius. Note that all planets in the Kepler-444 system are interior to the orbit of the innermost planet in the Kepler-11 system, the prototype of this class of highly-compact multiple-planet systems. Credit: Campante et al.

But what about that lack of metals we would assume for stars this old? The paper points out that while gas giant planets do seem to form around metal-rich stars, smaller planets (defined here as those with a radius less than four times that of Earth), can form under a wide range of metallicities. From the paper:

This could mean that the process of formation of small, including Earth-size, planets is less selective than that of gas giants, with the former likely starting to form at an earlier epoch in the Universe’s history when metals were far less abundant (Fischer 2012).

What does seem important, the paper argues, are the so-called α-process elements, the α-process being one of the classes of fusion reactions that allows stars to convert helium into higher elements. The α-process elements include carbon, nitrogen, oxygen, silicon and others that are significant in the formation of Earth-like worlds. The paper continues:

In particular, α elements comprise the bulk of the material that constitutes rocky, Earth-size planets (Valencia et al. 2007, 2010). Stars belonging to the thick disk [see diagram below] are overabundant in α elements compared to thin-disk stars in the low-metallicity regime (Reddy et al. 2006), which may explain the greater planet incidence among thick-disk stars for metallicities below half that of the Sun (Adibekyan et al. 2012c). Similarly favorable conditions to planet formation in iron-poor environments seem to be associated with a fraction of the halo stellar population, namely, the so-called high-α stars (Nissen & Schuster 2010). Thus, thick-disk and high-α halo stars were likely hosts to the first Galactic planets.


Image: An edge on view of the Milky Way. Credit: Wikimedia Commons.

It’s striking to consider that when our own planet formed, Kepler-444 and its planetary system were already older than our own planet is today. Kepler-444 appears to be slightly older than Kepler-10, which is known to have two super-Earths in orbit around it. As we find more such worlds, we’re learning that Earth-sized planets may have formed throughout much of the history of the universe.

The paper is Campante et al., “An ancient extrasolar system with five sub-Earth-size planets,” Astrophysical Journal (abstract / preprint). This Iowa State news release is helpful.


Comments on this entry are closed.

  • Andrew LePage January 28, 2015, 11:14

    This is yet another example of an ancient star possessing planets. Last year the discovery of a pair of planets orbiting the nearby Kapteyn’s Star (which has an age over ten billion years) was announced with one of those planets believed by some to be potentially habitable. While this habitability claim is overstated (something I will discuss in an upcoming article in Centauri Dreams), it is an interesting system nonetheless which, along with the newly discovered system orbiting Kepler 444, should tell us much about the planet formation process.


  • Joy January 28, 2015, 16:08

    To put a 3rd layer on the “metallicity” story, the elements heavier than iron have a different origin.
    1) “The α-process elements include carbon, nitrogen, oxygen, silicon …” These are the elements that Sol will release into a planetary nebula when it dies.
    2) Iron and similar metals lighter than iron are released in great quantity by ordinary supernovae but not by planetary nebulae
    3) Metals heavier than iron are 99% from neutron star collisions: http://www.spacedaily.com/reports/Stardust_on_ocean_floor_shows_gold_and_uranium_alchemy_in_stars_is_much_less_frequent_than_expected_999.html

    What this means is that there are 3 classes of rocky planets:
    1) planets with rocks, water and potentially biology, but no basis for early industrial development (very poor in copper, tin, and iron). While aluminum, magnesium, titanium, and silicon are available there is no clear pathway for a stone aged culture to start using these elements in a sophisticated way.
    2) planets with α-process elements plus adequate iron ores to start an iron age, but relatively lacking in the elements heavier than iron. So maybe an iron age is possible if the inhabitants can get there without a copper/bronze age kickstart. But getting started in electronics would be very difficult if not impossible without abundant copper ores and precious metals. And forget nuclear power without fissionables.
    3) Truly Earth like planets with gold nuggets lying around and uranium ores.

    The Drake equation implications are obvious

  • Ashley Baldwin January 28, 2015, 16:52

    The underlying implication to this is the use of Asteroseismology to constrain these planets in detail. The University of Birmingham is a centre of excellence of this having pioneered early work. As the number of exoplanets increases into the tens of thousands over coming decades Asteroseismology will be at the centre along with spectroscopy of characterising them. Asteroseismology is essentially light conversion into sound or vibration and this being conveyed from the stellar depths to surface for interpretation in the same way as seismic waves from the Earths core. Not straight forward but there is a couple of good reviews on arxiv. TESS, the next exciting exoplanet finder ,lacks the ability to conduct asteroseismology though for a small budget its photometry will still produce a prodigious return through photometry and clever spectroscopy synergy with JWST . Especially for its longer period and potentially more interesting planets . It isn’t just about M dwarfs by any means and could theoretically locate HBZ planets around even K3 stars. As has been shown with Kepler 444, Kepler possesses generous astroseismology capacity though the PLATO mission of next decade is far more than a simple transit photometry mission with a dedicated large capacity for asteroseismolgy on top of discovery . Just at the right time coming on the back of TESS and Gaia and in tandem with WFIRST. Numbers are fine, but characterisation via spectroscopy and asteroseismolgy is vital for building up the picture of planet formation and variation to really dig into Drake. It going to be a long time , if ever, before we see more than point source planet images but the technologies we do have can go a long way and as is always the case , the more light we get via bigger telescopes , the more we can tell. The next Decadel will be coming around soon and will hopefully have exoplanets centrally placed if it is to get the large TPF class telescopes we all crave.

  • Alex Tolley January 28, 2015, 17:16

    @joy – it may be worse than that – as life on Earth uses elements heavier than Fe for a number of biological processes. Most notably Cu and Zn.

    So both categories of planets ( 1 and 2) might not even have life if terrestrial life requiring certain transition metals is indicative.

  • Geoffrey Hillend January 28, 2015, 17:52

    Humans and life need iron. If there is not enough of that then no higher forms of life can evolve.

  • Abelard Lindsey January 29, 2015, 1:03

    We know elements heavier than Fe are necessary for known life, even prokaryotic bacteria. Zn and Cu are well-known as supplements. There are others. So yes, this implies that terrestrial planets that lack these elements will not have any life on them at all. These would be sterile planets.

    I think sterile “Earth-like” planets would have CO2 rich atmospheres with no O2 at all. Biological processes are necessary to maintain atmospheric O2.

  • torque_xtr January 29, 2015, 7:10

    The need for the third level of metallicity is an obstacle only for human history. Just imagine some sort of Pandora- or Leonis-like bio-civilization jumping all the way to info-age by using some planet-wide mycelium brain of neural network! Then they discover all the durable forms of carbon and boron and all the Al-Mg-Si-O-tech (Al-wires are only 1,5 times less conductive than copper ones) and high-energy redox based on H-C-N-O, and leap into space without even making a single iron blade – some 7 billion years ago.
    There is a great need for magnesium to silicon elements to make an earthlike planet. A world composed of only first two period elements would be entirely different and probably a sphere with a very thick CO atmosphere, a mantle of H2O and a core of high-density carbon dioxide, with proportions depending on C/O and with overall density of dense neptunes and light oceanworlds. A solid world with a thin atmosphere would be possible only with extreme oxygen deficiency, and Titan-like chemistry, or with entire-H2O composition…

  • Anthony Mugan January 29, 2015, 7:37

    A very interesting discussion.
    Would I be right in thinking that the earliest stars formed very shortly after the Big Bang…I think the record is around an age of 13.7 bya for a star still in existence. Supernova would therefore have been occurring for over 2 billion years by the time this system formed.

    What I am not at all clear about is how both the total amounts and relative proportions of the heavier elements would have evolved over time, but if I’ve understood it correctly there would have been some of the full range of elements around by the time this star was forming.

    Another possible aspect is the more compact nature of the universe in those very early times. I’m not sure if that might be a factor influencing the density of various elements (as gravitational effects would presumably be the key factor at any one point in space). It still seems astonishing to me that planets could form that early, but there we are!
    The question as to if life etc might have possible that early is a question that deserve careful thought therefore.

  • Ronald January 29, 2015, 13:20

    Fascinating. A few thoughts, also in response to others here;

    “In particular, α elements comprise the bulk of the material that constitutes rocky, Earth-size planets (…), which may explain the greater planet incidence among thick-disk stars for metallicities below half that of the Sun”.
    ??? I would think that iron and magnesium (and to a lesser extent calcium, nickel and aluminum) are also crucial for rocky planet formation.
    Does this imply that these thick disk (and halo) rocky planets are metal-poor or that there was sufficient metals that early already (as A. Mugan suggests)?

    ref. Tolley, Hillend, Lindsey: life needs iron as a trace element, so even if those planets are very iron deficient, but have traces of it, life could still arise and develop.

    It seems now that a few (3?) different planetary configurations (and formation histories) can give rise to terrestrial planets, even in the HZ:

    1) Our type: medium-high metallicity, open system with a gas giant on the outside and terrestrials on the inside; this type starts as a high mass protoplanetary dust disk (proplyd), but because a gas giant is formed which sucks up most material, it results in small planets on the inside.

    2) Medium metallicity: compact system of medium-sized planets (Neptunes, mini-Neptunes), probably the most common system (see in image above Kepler-11, 32, 33, 80, and as more nearby examples 61 Virginis, Nu2 Lupi). The really interesting and relevant question is, whether the density/mass of the proplyd has diminished so much toward the outside (i.e. increasing AU) that terrestrial planets may still have formed in the HZ.

    3) Low metallicity: (compact or open) system of only small planets (Kepler-444, and more nearby examples: Tau Ceti, 82 Eridani). If the series of small planets extends far enough toward the outside, there may be terrestrial planets in the HZ. How common? And what composition(s)?

    Types 2 and 3 may be the reason for terrestrial planets to be very common, the relevant question remaining how common in the HZ?

    (Very high metallicity probably results in just 1 or 2 hot giants, extremely low metallicity in no planets at all).

  • Ronald January 29, 2015, 13:23

    And just in, slightly off-topic, but making habitable terrestrial planets maybe even more common:
    Some potentially habitable planets began as gaseous, Neptune-like worlds,

  • Volucris January 29, 2015, 19:44

    Great summary Joy!

    Although I agree with torque_xtr in that there are ways to overcome elemental deficiencies, whether or not an aspiring civilization or life form could stumble upon them is another question..

    I would think that life could do with a lot less heavy trace elements, this could affect ultimate efficiency of a lot of enzymes and proteins, but I find it hard to believe that it would be impossible to do without them.

    As for technology, it makes for an interesting speculation! Al is a great conductor, not that much worse than copper, and also makes for a good structural metal. And I suppose Si and other light elements (for dopants) are enough to make semiconductors and for example ICs and solar panels. But the big snag is if there is no Fe, Ni or heavier friends there are no ferromagnets! And no ferromagnets means that it’s practically impossible to construct efficient transformers, electric motors or generators (except with superconductors, you would still have f. ex. magnesiumdiboride, but thats quite a learning curve right there..). That only leaves solar panels and chemical means for electricity generation.

  • torque_xtr January 30, 2015, 9:23

    Quick googling showed there are some compounds of 2nd and 3rd period elements like C-doped AlN, which show ferromagnetic properties, some even above room temperature [something like http://arxiv.org/pdf/0709.2059v1.pdf%5D, and some organic-based or even purely organic ferromagnets are also known. They all are cryogenic, but I believe there is lot to be found, although not possibly as good as classical ferrites. Ferromagnetism arises as the property of unpaired electrons, and organic chemistry has the potential to get them into needed structure even without d- and f-orbitals. So maybe there would be some materials for cores plus some weird air-core topologies, like that. Also, to make elements in the Mg – Si range, supernovae are needed, so there would be iron as well, the cutoff would be after iron. And iron admixes well into magnesium and aluminium miderals – so these early universe guys would have iron/ferrite cores as well, though not so early in the development, and tens of times more expensive comparing to other circuit components, if they go this way :-)

  • Rob Henry January 31, 2015, 6:57

    Does anyone know the expected C:O on these planets. If it is much higher than 0.8 Si metal would be common. Also I would guess that boron would be relatively enriched in the lower metal component. Boron is the only element even more versatile than carbon,, such that it is easy to imagine it helping life’s cause in other ways.

  • Eniac February 1, 2015, 23:26


    Note that all planets in the Kepler-444 system are interior to the orbit of the innermost planet in the Kepler-11 system, the prototype of this class of highly-compact multiple-planet systems.

    I believe this might be misleading. Jupiter-like planets in Jupiter-like orbits would simply not show up in Kepler data, and neither would Earth-like planets in Earth-like orbit. These systems, then, have not been shown to be really different from ours, except they have a few more close-in planets compared to our own, lone, mercury. I am not sure, but I can imagine that even that is a consequence of selection bias more than anything else.

  • Eniac February 1, 2015, 23:37

    Abelard Lindsey:

    We know elements heavier than Fe are necessary for known life, even prokaryotic bacteria. Zn and Cu are well-known as supplements. There are others. So yes, this implies that terrestrial planets that lack these elements will not have any life on them at all. These would be sterile planets.

    I think this is a false conclusion. Just because these elements are utilized in known life does not mean that they could not be easily replaced, if necessary.

    We see many instances in biology where the same thing is accomplished with radically different means. For example, arthropod blood uses copper to bind oxygen for more efficient circulation, vertebrates use iron. Who is to say that there are not other means to the same end, even some that do not require a heavy element at all?

  • Mark Zambelli February 5, 2015, 12:57

    @Anthony Mugan
    I think you’re correct to doubt the young Universe’s smaller volume having much of an effect on mixing. The expansion of the Universe is only happening in the voids away from matter… there is enough local gravity within a single galaxy to halt any expansion there. The earliest stars formed in proto-galaxies (ie volumes with enough matter collapsing to seed galaxy formation) so any metal enrichment would occur within the confines of these early galaxies. The number of early galaxies was greater and in a smaller volume, yes, but the expansion was much faster and inter-galactic distances would still be huge.

    Three methods for dispersing metals on galactic scales come to mind… 1~ mergers/close-encounters (tidal and stellar streamers; any material pulled away that doesn’t end up bound into the merged galaxy may not get recycled into a new galaxy at a later time being lost to the hydrogen clouds in the voids). 2~ AGN (feeding super-massive black holes create jets that can drill out into the intergalactic medium carrying metals away), and 3~ Massive starburst activity can push away huge volumes of gas due to numerous SN shockfronts. The bubbles that M82 is blowing is a case in point). Mixing on larger scales will be locally confined to the groups of galaxies that adorn the filaments as they slide down them to meet their respective clusters/superclusters.

    Even after just 2Gyrs the number of supernovae cycles within any given galaxy must’ve been staggering to allow for so much mixing within adjoining parts of molecular clouds /nebulae that triggers even more. It seems as though far more enrichment and mixing was occurring within each galaxy rather than material being dispersed throughout the early Universe as a whole during those cosy times.