≡ Menu

Unusual Atmosphere of a ‘Sub-Neptune’

We refine our terminology as we go when a field as new as exoplanetology is in play. Take the case of GJ 3470b. At 12.6 Earth masses, is this a ‘sub-Neptune’ or a ‘super Earth’? Neptune itself is 17 Earth masses, so I’d on balance give the nod to ‘sub-Neptune,’ though categories here get confusing. The planet is 0.031 AU out from its star, a red dwarf half the mass of our Sun. Oddly, it has a hydrogen/helium atmosphere in which heavier elements are all but absent.

We know this because scientists have been able to put data from both the Hubble instrument and Spitzer to work on an analysis of the atmosphere of the planet. This is done through a technique we’ve examined before, transmission spectroscopy, in which astronomers study the absorption of the star’s light as the planet passes across its face (a transit as seen from Earth), and then the loss of reflected planetary light as the planet moves behind the star (this is called a secondary eclipse).

Image: A comparison between transits and secondary eclipses (also sometimes called occultations). In a planetary transit, the planet crosses in front of the star (see lower dip) blocking a fraction of the star’s brightness. In a secondary eclipse, the planet crosses behind the star, blocking the planet’s brightness (see dip in the middle). The latter dip in brightness is fainter due to the faintness of the planet. Credit: astrobites/Josh Winn.

The atmospheric data come from observations of 12 transits and 20 eclipses, giving us a first look at the atmospheric composition of a world like this. GJ 3470b is, to say the least, unusual. Björn Benneke (University of Montreal) is lead author of the paper, now available in Nature Astronomy:

“This is a big discovery from the planet-formation perspective. The planet orbits very close to the star and is far less massive than Jupiter – 318 times Earth’s mass – but has managed to accrete the primordial hydrogen/helium atmosphere that is largely ‘unpolluted’ by heavier elements. We don’t have anything like this in the solar system, and that’s what makes it striking.”

The scientists expected to find heavier elements such as carbon and oxygen, out of which we would detect water vapor and methane. This would point to the Neptune model. But the data threw a curve: What was actually revealed was an atmosphere devoid of heavy elements, to such a degree that Benneke likens it to the hydrogen and helium composition of the Sun. With resemblances to a gas giant atmosphere in being hydrogen-dominated, it is nonetheless one that is depleted in methane.

Here’s how the paper handles the lack of methane, and its bearing on planet formation. One possibility is that there is interior heating that is not being accounted for, but there are others:

Evolution modeling of GJ 3470b indicates that internal heat from formation should have been radiated away within a few Myr, well below the estimated age of the system; however, tidal heating due to forced eccentricity from another unseen planet in the system, similar to the situation with Jupiter’s moon Io could be a possible explanation. The residual non-zero eccentricity of GJ 3470b as independently confirmed by our eclipse observations and radial velocity measurements support this hypothesis. Alternatively, GJ 3470b’s surprising lack of methane could potentially be the results of photochemical depletion due to catalytic destruction of CH4 in deeper atmospheric regions where photolysis of NH3 and H2S release large amounts of atomic hydrogen. The fact that ammonia is also depleted in comparison to expectations based on our chemical-kinetics modeling is consistent with this catalytic-destruction possibility.

Image: This artist’s illustration shows the theoretical internal structure of the exoplanet GJ 3470b. It is unlike any planet found in the Solar System. Weighing in at 12.6 Earth masses the planet is more massive than Earth but less massive than Neptune. Unlike Neptune, which is 4.5 billion kilometers from the Sun, GJ 3470b may have formed very close to its red dwarf star as a dry, rocky object. It then gravitationally pulled in hydrogen and helium gas from a circumstellar disk to build up a thick atmosphere. The disk dissipated many billions of years ago, and the planet stopped growing. The bottom illustration shows the disk as the system may have looked long ago. Observation by NASA’s Hubble and Spitzer space telescopes have chemically analyzed the composition of GJ 3470b’s very clear and deep atmosphere, yielding clues to the planet’s origin. Many planets of this mass exist in our galaxy. Credit: NASA, ESA, and L. Hustak (STScI).

The authors rule out migration of a world that formed beyond the snow line, for this origin would have produced the heavier elements we do not see here. Instead, they believe that GJ 3470b formed where it is today, showing that sub-Neptunes can form with atmospheres that are the result of direct accretion from the protoplanetary disk onto a rocky core. The implication is that we are looking at a planet-forming process that is essentially distinct from more massive planets, one in which the gas envelope is not enriched to any great degree by later collisions.

There is so much to learn about this, indicating just how far we have to go in our understanding of such low-mass, star-hugging planets. The authors point to GJ 3470b as a prime target for the James Webb Space Telescope. Every time I read something like this I think about how much is riding on JWST being launched successfully and can only keep my fingers crossed. For studying atmospheric chemistry is going to require powerful space-based resources as we start delving into atmospheres on worlds this small and aim at rocky worlds that are smaller still.

The paper is Benneke et al., “A Sub-Neptune Exoplanet with a Low-Metallicity Methane-Depleted Atmosphere and Mie-Scattering Clouds,” published online by Nature Astronomy 1 July, 2019 (preprint).

{ 19 comments… add one }
  • Ron S. July 9, 2019, 15:12

    “GJ 3470b may have formed very close to its red dwarf star as a dry, rocky object. It then gravitationally pulled in hydrogen and helium gas from a circumstellar disk to build up a thick atmosphere.”

    I am curious about this. Since there was enough time for a rocky planet to form, would there have been sufficient volatile gases remaining in the inner disk? Depending on when the protostar begins radiating the inner disk would be cleared of volatiles.

  • Brett July 9, 2019, 15:38

    Sounds about right. If I remember the modeling correctly, planets above 2 times Earth mass can start holding on to hydrogen and helium the more massive they get (and depending on when they form), and once you get above 5-6 Earth masses odds are all of them are holding on to significant hydrogen and helium envelopes.

  • Geoffrey Hillend July 9, 2019, 15:57

    I agree with the idea that exoplanet GJ 3470b did not migrate. I think it formed in situ. The explanation for it’s lack of heavy elements is explained by the idea that it never started with a rocky core, but formed like its star by the collapsing of a gas cloud of hydrogen and helium. This planet might be the smoking gun for the idea that expoplanet gas giants which form near stars don’t have any rocky core. All gas giants might have formed without rocky cores, but maybe the ones which formed behind the snow line do have them. If we see more of these metal poor gas giants near their stars it might support such hypothesis.

  • Geoffrey Hillend July 9, 2019, 16:02

    I was thinking that maybe stars don’t star out with rocky cores and since the gas giant formed at the same time as the star, it would not need a solid surface to attract hydrogen gas since the area around the star and the gas was not yet hot? Gravity can compress and accumulate gas without a solid surface to form a star, but if a star needed to have a solid surface, then we would expect the same for the gas giant, but I don’t think a star does need a solid surface to form.

  • Geoff P July 9, 2019, 21:01

    Maybe it’s neither a super-Earth nor a sub-Neptune – just a very small brown dwarf.

    • Bruce D. Mayfield July 11, 2019, 11:09

      The known mass of approximately 14 Earths is way below that for Brown Dwarfs. The least massive BDs weigh in at a hefty 13 Jupiters. Jupiter out masses Earth by about 318 times!

  • Geoffrey Hillend July 9, 2019, 22:30

    Thinking about this more carefully, I agree with the authors idea that the photochemical depletion of GJ 3470b is the most likely explanation for its lack of heavy elements. I can’t think of a reason why an exoplanet gas giant might not form first with a rocky core near a star. Most likely the solid core formed first and the the lighter element atmosphere was attracted to it. I still have to reject the migration hypothesis as being too improbable.

  • Robin Datta July 10, 2019, 3:07

    Should gravitational effects need “rocky cores”?
    Density of uranium (at room temperature) 19.1 g/cm³
    “The core of the Sun extends from the center to about 20–25% of the solar radius. It has a density of up to 150 g/cm³.” (Wikipedia). And that’s mostly H & He.

  • Michael C. Fidler July 10, 2019, 5:31

    A very old Dyson Sphere.

  • ericSECT July 10, 2019, 6:56

    If they only detect Hydrogen and Helium….. perhaps the observation method is not working as well as expected and they are getting too much light contamination from the primary?

  • wdk July 10, 2019, 14:24

    Looking at the pre-prints, I saw a great deal about the observation campaign with Hubble and Spitzer, plus considerable data analysis.
    I did not catch any estimate of the age of the star. Is it an old system?
    Perhaps the metalicity of the planet reflects something about early planets just as stellar low metal content pegs stars?

    Like others above, I am puzzled by the accretion idea that a circumstellar disk would be primarily be helium and hydrogen and then swept into the gravity well of an already formed rocky or metallic planet. What’s a good name for a self made planet that picked itself up by its boot straps?

  • Geoffrey Hillend July 10, 2019, 17:26

    The core of our Sun or 25 percent of it is the only part where fusion occurs. The temperatures in that region are above 10 million kelvin, the temperature needed in for fusion to occur. Source: Astrophysics is Easy, Inglis, 2015. Fusion does not occur below that temperature. Fusion is caused by the extreme gravity which moves the particles of hydrogen to move and collide at high velocity like 500 kilometers a second which collide together to fuse hydrogen into helium. These collisions are shown in the proton proton chain particles which emit x-rays, positrons, etc which provide a radiation pressure to balance for the force of gravity. It is this force of gravity that gives the Sun’s core a high density. The Sun’s density in the core is roughly 225 billion atmospheres because of the greater mass of the Sun has a greater gravitational pressure at the cores center than Earth. I recall over 90 Jupiter masses is needed for the core temperature to be hot enough to start fusion and core temperature is based on mass and gravity.

    It is true that gravitational effects don’t need rocky cores and the accretion of cosmic dust or metals. Usually in area near the star is supposed to have more heavy elements. Some M dwarf stars have rocky exoplanets which are in orbits close to the M dwarf and I doubt they migrated but probably formed in situ. The age of the star can tell us something since population II stars are metal poor, but GJ 3470 is a population I star at an estimated 2 billion years old. I also wonder if an ambiguous spectral signature is possible.

    • wdk July 11, 2019, 23:16

      That’s an interesting and vivid picture of stellar fusion. Appreciate it.
      However, to clarify, I didn’t mean to suggest that planets were experiencing fusion – and I would adhere to lower limits imposed on hydrogen ( 0.08 solar for suns and, as you say, dozens of Jupiters for deuterium for consideration as a brown dwarf. But what I was wondering was whether an old sun would have a surrounding interstellar medium and accretion disk with low metallicity as well. i suspect that an ancient interstellar medium would be metal depleted – but according to information provided, this star is not necessarily metal lacking. But if there is such a correlation, age and lack of metals, we could imagine that formed worlds would have more H2 and He in their atmospheres, say if an analog to the solar system formed. But then when you go back to very early in the universe history, nearly all H and He, would have to wonder about what a planet would nucleate.

      Best regards,

  • kzb July 11, 2019, 7:22

    Has anyone taken into account the metallicity of the star?

  • Bruce D. Mayfield July 11, 2019, 13:46

    Taking Geoffrey’s statement that GJ 3470 is only 2 billion years old its metallicity should be fairly high kzb. Therefore I think that the expectation that this planet has a large rocky core as shown in the graphic is very well founded, especially considering the planet’s mass.

    I wondered how strong GJ 3470 b’s “surface gravity” (well, at the top of its visible atmosphere) is and it works out to approx 0.91 g. A little less that Earth’s 1 g., but if the great majority of the planet’s mass is made up of rock (a reasonable assumption) then the G force at the bottom of that thick atmosphere could be significantly higher than 1 g. I’m suggesting that heavier gases may have been trapped underneath the H and He.

    Also, GJ3470b’s atmosphere is known to be evaporating away at a high rate, so the topmost layers would have been stripped away. Think Jupiter peeled back to its bulk atmosphere of sunlike H and He.

  • wdk July 12, 2019, 23:05

    Bruce D. Mayfield,
    Interesting idea for a number of reasons. Without having done analysis, I think there are arguments pro and con for this. For one, there is some
    atomic or molecular weight ordering out in the upper regions of Earth’s atmosphere, but at densities that we hardly note that the atmosphere is there. But on the other hand, many atmospheres are convective. And we’ve had discussions earlier on this site on things like carbon monoxide in Jupiter and other jovian atmospheres out of equilibrium for pressure and temperature, suggesting upwelling. Probably confirmed visibly with belt and cloud formations. Typically, evidence of convection would be established in planetary or stellar atmospheres based on Rayleigh or other fluid dynamic parameters: temperatures, molecular weights and scale heights would be in there somewhere.
    Worse, some atmospheres alternate between convective and radiative.
    You could look at the water filled troposphere and dry stratosphere as
    an analog, but it’s hard for me as yet to imagine molecules bound to the lower region vs. hydrogen and helium, unless radiative heat transfer were somehow limited to those species in the upper region. I don’t know the answer, but the heat transfer model might be key.

    • Bruce D. Mayfield July 14, 2019, 14:50

      Thanks for that thoughtful response wdk. You’re very right to expect convection to be very active on GJ3470b. Just looking at Wikipedia’s brief article on this planet, it must have a very wide range of atmospheric temperatures. With its star hugging orbit the sub-stellar point is thought to be at about 1700K, while the average temp is expected to be only about 600K. That much thermal gradient would have to drive a lot of atmospheric mixing, wouldn’t it?

      So try this idea then; The atmosphere is very well mixed over time. Whenever molecules like H2O, CH4, etc. are swept up into the sub stellar region they are broken down by heat and radiation. The heavier elements meanwhile tend to bond chemically with rocks when they are in contact with the surface. Over time the atmo becomes nothing but H and He.

  • kzb July 16, 2019, 7:30

    Bruce: H2O would break down in the upper atmosphere as you state. However it is hydrogen that would escape preferentially over the oxygen. This mechanism would lead to an oxygen rich atmosphere not a hydrogen rich one.
    As to the star metallicity, with a bit of Googling I found [Fe/H] = +0.20. So, yes, the low metallicity planet atmosphere can’t be explained by there being no metals to start with, which is what I was wondering.

    • Bruce D. Mayfield July 16, 2019, 18:58

      kzb, of course H2 will be and is escaping, but there is so much of it that it still remains at ample, even solar abundance levels. What I’m suggesting is that elements from Carbon on up settle (rain?) or bond to the surface, thus resulting in an atmosphere depleted in everything but H and He. With such a varied range of temperatures available molecules heavy enough to settle out of GJ3470b’s atmosphere surely formed and dropped out.

Leave a Comment