≡ Menu

Kepler 107: Collision of Worlds

It seems increasingly clear that the factors that govern what kind of a planet emerges where in a given stellar system are numerous and not always well understood. Beyond the snowline, planets draw themselves together from the ice and other volatiles available in these cold regions, so that we wind up with low-density gas or ice-giants in the outer parts of a stellar system. Sometimes. Rocky worlds are made of silicates and iron, elements that, unlike ice, can withstand the much warmer temperatures inside the snowline. But consider:

While we now have 2,000 confirmed exoplanets smaller than three Earth radii, the spread in their densities is all over the map. We’re finding that other processes must be in play, and at no insubstantial level. Low-density giant planets can turn up orbiting close to their stars. Planets not so dissimilar from Earth in terms of their radius may be found with strikingly different densities in the same system, and at no great distance from each other.

Which takes us to a new paper from Aldo S. Bonomo and Mario Damasso (Istituto Nazionale Di Astrofisica), working with an international team including astrophysicist Li Zeng (CfA). The collaboration has produced a new paper in Nature Astronomy that uses the planetary system around the star Kepler-107 to probe another possible formative influence: planetary collisions. Kepler-107 may be flagging a process that occurs in many young systems.

Image: The figure shows one frame from the middle of a hydrodynamical simulation of a high-speed head-on collision between two 10 Earth-mass planets. The temperature range of the material is represented by four colors grey, orange, yellow and red, where grey is the coolest and red is the hottest. Such collisions eject a large amount of the silicate mantle material leaving a high-iron content, high-density remnant planet similar to the observed characteristics of Kepler-107c. Credit: Zoe Leinhardt and Thomas Denman, University of Bristol.

Let’s take a deeper look at this curious system. The two innermost planets at Kepler-107 have radii that are nearly identical — 1.536 and 1.597 Earth-radii, respectively. They both orbit close to the host, a G2 star in Cygnus of about 1.25 solar masses, with orbital periods of 3.18 and 4.90 days. The scientists used the HARPS-N spectrograph at the Telescopio Nazionale Galileo in La Palma to determine the planets’ masses, and because they were working with known radii (thanks to Kepler’s observations of these transiting worlds), they were able to determine their densities. And now things get interesting.

For the innermost planet shows a density of 5.3 grams per cubic centimeter, while the second world comes in at 12.65 grams per cubic centimeter. The inner world, Kepler-107b, is thus about the same density as the Earth (5.5 grams per cubic centimeter), while Kepler-107c shows a much higher number (for comparison, water’s density is 1 gram per cubic centimeter).

The outer of the two is the denser world and by more than twice the inner world’s value, and given the proximity of their orbits, coming up with stellar radiation effects that could have caused mass loss is difficult to do, for such radiation should have affected both in the same way. The remaining strong possibility is that a collision between planets played a role in this system.

“This is one out of many interesting exoplanet systems that the Kepler space telescope has discovered and characterized,” says Li Zeng. “This discovery has confirmed earlier theoretical work suggesting that giant impact between planets has played a role during planet formation. The TESS mission is expected to find more of such examples.”

And this from the paper, which notes the possibility of extreme X-ray and ultraviolet flux in the young system, but dismisses it as operational here, at least to explain the discrepancy:

This imbalance cannot be explained by the stellar XUV irradiation, which would conversely make the more-irradiated and less-massive planet Kepler-107b denser than Kepler-107c. Instead, the dissimilar densities are consistent with a giant impact event on Kepler-107c that would have stripped off part of its silicate mantle.

Image: The video shows a hydrodynamical simulation of a high-speed head-on collision between two 10 Earth-mass planets. The temperature range of the material is represented by four colors grey, orange, yellow and red, where grey is the coolest and red is the hottest. Watch Video. Credit: Zoe Leinhardt and Thomas Denman, University of Bristol.

We can produce Kepler-107c, then, by a collision that results in a high-density remnant. We have apparent evidence for collisions even in our own Solar System. The composition of Mercury, with a dense metallic core and thin crust, may point to this; only Earth is more dense in our system. So too could the emergence of Earth’s moon through a planetary-sized impactor striking our planet. We can also see in the obliquity of Uranus — the planet’s axis of rotation is skewed by 98 degrees from what we would expect — the possibility that, as advanced in some theories, a large object struck the planet long ago.

The paper gives us some of the details about Kepler 107:

…the difference in density of the two inner planets can be explained by a giant impact on Kepler-107 c that removed part of its mantle, significantly reducing its fraction of silicates with respect to an Earth-like composition. The radius and mass of Kepler-107 c, indeed, lie on the empirically derived collisional mantle stripping curve for differentiated rocky/iron planets… Smoothed particle hydrodynamics simulations show that a head-on high-speed giant impact between two ~10M⊕ exoplanets in the disruption regime would result in a planet-like Kepler-107 c with approximately the same mass and interior composition… Such an impact may destabilize the current resonant configuration of Kepler-107 and thus it likely occurred before the system reached resonance. Multiple less-energetic collisions may also lead to a similar outcome.

Another possibility discussed by the authors: Planet c may have formed closer to the parent star and later crossed Kepler-107b in its orbit. But the authors note that to dampen the orbital eccentricities this would have produced, this scenario is unlikely to have had time to operate.

The paper is Bonomo & Zeng, “A giant impact as the likely origin of different twins in the Kepler-107 exoplanet system,” Nature Astronomy 04 February 2019 (abstract).


Comments on this entry are closed.

  • Thomas R Mazanec February 20, 2019, 15:26

    The temperature range of the material is represented by four colors grey, orange, yellow and red, where grey is the coolest and red is the hottest.

    The “sparks looked yellow at first to me in the video and went to red. That also makes more sense, as the color-temperature of material would go grey-red-orange-yellow, from cold to hot. Could the ordering in the caption be a misprint?

  • Geoffrey Hillend February 20, 2019, 18:55

    You have to look carefully to see the red which appear when the moment the two planets collide. The temperature color code of the video appears to be correct.

  • Robin Datta February 21, 2019, 2:18

    It would seem that we are witness to a transient moment of remarkable tranquillity in our neck of the universal woods. Perhaps it is a consideration in Fermi’s paradox.

  • Bill February 21, 2019, 8:29

    Under what circumstances would there be a head on collision between planets instead of an overtaking collision or merger between planets?

    • Gregor February 21, 2019, 12:17

      I wondered about that too.

    • Bruce D. Mayfield February 21, 2019, 13:25

      Bill and Gregor, a head on collision doesn’t require the planets original orbits to be non co-planer, if that makes sense. Consider the case of two very closely orbiting planets in which their orbits are in approximately the same plane around their star. The inner planet will be moving faster and will overtake the outer planet. As the distance is closed between them the gravity between the two planets will speed up the inner and retard the outer. They might collide squarely, but it would just be a case of the inner ramming into the trailing hemisphere of the outer while their both moving in the same general direction about their star.

      Admittedly this type of collision would be far less violent than if their orbits were in widely differing planes.

    • Ron S. February 21, 2019, 13:57

      It depends on how you define “head on”. All it has to means is that the velocity vector of one body points to the center of the other body (and vice versa). There does not need to be a reference to a third body, including the star. That is, the orbit of one does not need to be retrograde.

  • Andrei February 21, 2019, 10:17

    12,65 g/cm3 is indeed a curiously high density.
    Lead is 11,3 g/cm3 mentioned as comparison.
    Even though there will be compression in the center, this will only be a smaller part of the planet volume.
    So if the value is correct this planet need to consist almost entirely of the heavier metals we know of.
    I am certain they have done their calculations right, and the planet have indeed been created by a collision.
    Even so, from the discussions in other parts of this website, most agree that we should be on the lookout for things that are unusual – and in that respect I think this planet qualify.

    • Bruce D. Mayfield February 22, 2019, 15:37

      Good points re the surprising density of the more massive planet Andrei.
      I looked up the metalicity of this system’s star, 0.321 dex, which means that this system has about 2.09 times the iron content of ours. I agree that this dense planet must contain an abundance of the denser elements. Maybe instead of a nickel-iron core it has an osmium-iridium core. Note this re Os and Ir:
      “Osmium has a blue-gray tint and is the densest stable element, approximately twice the density of lead[3] and slightly denser than iridium.[6] Calculations of density from the X-ray diffraction data may produce the most reliable data for these elements, giving a value of 22.587±0.009 g/cm3 for osmium, slightly denser than the 22.562±0.009 g/cm3 of iridium; both metals are nearly 23 times as dense as water.[7]”

      • Andrei February 23, 2019, 10:19

        Yes there would be an abundance of heavy elements that we consider to be rare. And this make me a bit skeptic of this result of a planet 1,81 times Earth radius with such a very high density.
        The high metallicity you mention is unusual in itself I thought the planet would be very volcanic from radioactive elements. Osmium and Iridium could be present also of course.
        I looked at the possble gravity and a basic calculation without compression gave me 40,77 ms2 or 4,17 G at the surface.
        The planet is not newly formed as the star is ~4,31 Gy but if the collision have been recent it could have added to the volcanism and so distilled the planet to become such an unusual world.

        • Bruce D. Mayfield February 23, 2019, 19:00

          You doubt the density figure for Kepler 107c? This figure looks to be on very solid ground, given what’s known; radii of both star and planet from the Kepler mission, and both star and planet’s masses from radial velocity findings. Density = Mass/Volume.

          Given also the system’s high metallicity, the 1.25 solar massed (1.48 solar radii, 5854 K temp) star and the close proximity to the star, the high density of Kepler 107c seems well within the realm of possibility. What’s stranger is the much lower density interior planet. My explanation for it would be that it has some kind of a heat inflated exotic atmosphere that can resist destruction in such a hot environment.

          • Andrei February 24, 2019, 8:52

            Many claims on the characteristics of planets found have often been retracted, with that in mind I do caution with skepticism to the numbers here.
            It it turn out to be correct it would be an extraordinary planet though, the heaviest terrestrial planet known, a lot of the volume of elements from the lower part of the periodic table and then a very interesting world.

  • djlactin February 21, 2019, 10:24

    A head-on collision would require a body in retrograde motion (I presume that neither impactor was an interstellar interloper). How is this possible in accretion-based scenarios of planetary formation?

  • Michael C. Fidler February 21, 2019, 10:34

    What about the merger of a double planet?

  • wdk February 21, 2019, 13:50

    Well, here’s a conjecture. Let’s say that at one time the outer, heavier planet was inside of the lighter inner planet and experienced a history
    in which most of the outer layers were peeled off either by intense heat or collisions. Then there was a close pass with the other planet and they
    exchanged relative orbital positions?

    3 and 4 day periods around a G star more massive than sol is quite a hot bath. Mercury, relative to other rocky planets is stripped to an iron core, but as noted above, what circumstances would leave a core with the density of lead, mercury or other less common elements?

    Could the denser planet have been more like Neptune or Jupiter in earlier times and have been stripped of its more volatile components?
    And then perhaps the less dense object was a satellite that managed
    to escape into “star-centric” orbit when the mass reached a lower level?

    If I were to accept my own hypothesis or hypotheses, I would still have to get around the high density – even in a stripped core. Does any planetary formation theory, even for Jupiter scale bodies, theorize a core that dense? We talk of metallic or atomic hydrogen in the depths of
    Jupiter or Saturn, but could there be any other elements shifting to exotic states at high pressure and density – such that you get overall densities that high in a 3-Earth mass body.

    Double the density suggests a cube root of 1/2 difference factor in diameter: about 0.7937 if this derivation is based on light curves in transit. What should we suppose the error bar would be on the two
    planet induced luminosity dips?

  • Michael February 21, 2019, 14:20

    More likely a Thia type impact that formed our moon, a lagrange point or near semi stable orbit.

  • Geoffrey Hillend February 21, 2019, 16:51

    I am guessing that these super Earths needed the momentum and kinetic energy of a head on collision to put them into separate orbits instead of orbiting around each other on a center of mass like our Earth Moon system?

    • Bruce D. Mayfield February 22, 2019, 11:07

      A “head on”, or perhaps less confusingly termed central collision might be the easiest to model, but in the real cosmos such collisions would be just one of a wide range of possibilities. I wonder what the probabilities are among such. As two massive bodies approach one another both trajectories would bend as they fall into each others’ gravity wells. Since the increasing as the inverse square of the distance force vectors would each point at the other planet’s center, central and near central collisions might be a common trend, but that must be weighed against conservation of the bodies previous momentums.

      The collision that lead to the Earth-Moon system has been best modeled as an off-center impact. Indeed, if that impact had been central would we even have two bodies now, or would the debris mainly just fell down to form a somewhat larger Earth?

      • wdk February 22, 2019, 16:45

        This is all still a new situation to confront, but as time goes by, I’m more inclined to a “flip” situation rather than a collision. Probably back in the files of Centauri Dreams there are some articles about
        Solar System models developed by the Nice Group in France about a decade ago. Gas giants in the early solar system were modeled. In some cases Neptune size bodies escaped into interstellar space – and in others the giant planets exchanged places out from the sun, the result of orbital element drifts and close passage.

        So, I’m inclined to argue for a flip of the two worlds, with the denser originating closer to the Kepler 107 and the less dense flipped inside.
        But how the denser of the two can be as dense as described,…?
        With this hypothesis, would anticipate a density difference a few
        grams per cubic centimeter closer – and still end up with very “refractory” material. To explain the factor of 2, somebody else will have to step up to the podium.

  • Bruce D. Mayfield February 22, 2019, 19:19

    That’s a good alternative possibility wdk. (The Nice model you referred to had Jupiter and Saturn swapping places, thereby stirring things up by causing the Late Heavy Bombardment.)

    So let’s assume that such a planetary “flip” occurred here and that these planets accreted close to the star. It would have been hellaciously hot that close to a G star during its T-Tauri stage. I note that Os, in addition to having the highest density, also has one of the highest melting points. This environment seems like the kind of blast furnace that could have smelted the planet down and stripped off the lighter elements.

    The problem with this notion is the low density planet. But, perhaps it has an atmosphere which has been inflated by its proximity to the star.

  • wdk February 23, 2019, 15:23

    On reflection, the first exoplanet discovery was made by Aleksander Wolszczan in 1992, around pulsar PSR 1257. I haven’t heard much about
    successive neutron or pulsar planetary systems, but I would imagine that if anything were left of these objects, they would be of a highly refractory nature as well. Maybe with similar densities. Since our data about their existence is primarily from the radio spectrum and doppler effects, there probably hasn’t been a specific value for densities of the planets involved. Well, all the osmium you want – for the taking. For someone who has everything: a huge ballpoint pen.

  • Geoffrey Hillend February 25, 2019, 2:01

    If the Moon hit the Earth head on then we would not have the angular momentum to increase the rotation. The Earth might not have a fast rotation and magnetic field as a result.

    • Alex Tolley February 25, 2019, 12:20

      That is a hypothesis that might be testable quite soon when telescopes start to gain low-resolution images of the surfaces of exoplanets. As most are likely to be moonless, we should see very slow rotations of such worlds where tidal locking should be absent.

      However, it seems to me that Mars represents a counterfactual. It has a rotation similar to Earth, yet presumably did not have a collision. I also wonder about the gas giants too as they seem to rotate quite quickly.

      • Bruce D. Mayfield February 25, 2019, 14:23

        Conservation of angular momentum as bodies form up from many collisions, along with tidal exchanges of energy are sufficient to explain all the various rotations of our system’s planets. The piece of giant molecular cloud that collapsed into the Sol system must have possessed a great deal of angular momentum.

        In a system like ours with high angular momentum and widely spread planets rapidly spinning planets with spins generally in the same plain as the orbit would be common, but with exceptions due to the randomness of collisions. Sometimes a planet might get hit in such a way that its rotation is cancelled out or twisted.

        Solar tides on Mars are too weak to slow it down even over billions of years. Tidal locking only happens if objects are fairly near each other.

        • Alex Tolley February 25, 2019, 15:53

          Rereading your comment I see that you don’t actually say that the collision with Earth was responsible for our rotation, only that the collision could not be head-on which could have slowed our rotation. My apologies, as my interpretation from my original reading of it, was that you were making another “Rare Earth” argument. It was that that interpretation that I was trying to counter.

          • Bruce D. Mayfield February 25, 2019, 22:15

            I still think that the Rare Earth hypotheses has some strong points, but not dogmatically so. I recommend keeping an open mind (to a reasonable degree, don’t fall for everything of course). Since I incourage open mindedness to others, I need to practice it myself.