≡ Menu

TRAPPIST-1: Planets Likely Rich in Volatiles

Yesterday we saw that, by pushing the Hubble telescope to its limits, we could make a call about three of the TRAPPIST-1 planets — d, e and f — and one possibility for their respective atmospheres. The Hubble data rule out puffy atmospheres rich in hydrogen for these three (TRAPPIST-1 g needs more work before a definitive call can be made there).

This is a useful finding, for hydrogen is a greenhouse gas that can heat planets close to their star beyond our usual norms for habitability. Set out deeper in a stellar system, we can think of Neptune, a gaseous world far different from the kind of rocky, terrestrial-class planets most likely to produce surface water. So on balance, the Hubble work, while not telling us anything more about potential atmospheres in this system, does rule out the Neptune scenario. That leaves open the question of whether future instruments will find more compact atmospheres.

The James Webb Space Telescope should be able to probe these worlds, perhaps revealing heavier gases like methane, carbon dioxide, water and oxygen. Meanwhile, we have another new paper to look at, from lead author Simon Grimm and colleagues, taking another angle on the composition of the TRAPPIST-1 worlds. Grimm (University of Bern Centre for Space and Habitability) and team have produced new mass estimates that allow a more fine-grained appraisal of the planets’ density, which is a step toward characterizing each planet.

Image: This chart shows, on the top row, artist concepts of the seven planets of TRAPPIST-1 with their orbital periods, distances from their star, radii, masses, densities and surface gravity as compared to those of Earth. These numbers are current as of February 2018. On the bottom row, the same numbers are displayed for the bodies of our inner solar system: Mercury, Venus, Earth and Mars. The TRAPPIST-1 planets orbit their star extremely closely, with periods ranging from 1.5 to only about 20 days. This is much shorter than the period of Mercury, which orbits our sun in about 88 days. Credit: NASA/JPL-Caltech.

The work of Grimm and team is yet another illustration of why TRAPPIST-1 is such a remarkable target. A transit can tell us about the radius of the world between us and the star, but we also need mass information to make a call on density. Calling this system “…a fascinating setting to study the formation and evolution of tightly-packed small planet systems…” the paper explains the problem in a nutshell:

While the TRAPPIST-1 planet sizes are all known to better than 5%, their densities suffer from significant uncertainty (between 28 and 95%) because of loose constraints on planet masses. This critically impacts in turn our knowledge of the planetary interiors, formation pathway (Ormel et al. 2017; Unterborn et al. 2017) and long-term stability of the system. So far, most exoplanet masses have been measured using the radial-velocity technique. But because of the TRAPPIST-1 faintness (V=19), precise constraints on Earth-mass planets are beyond the reach of existing spectrographs.

Fortunately, in this system we are dealing with seven tightly packed planets (all within the orbit of Mercury around the Sun). In this resonant chain, more massive planets can perturb the orbits of lighter ones, creating transit timing variations (TTV) that can be modeled to produce mass values for each world. 284 transit timing variations obtained with the SPECULOOS and TRAPPIST instruments between September 17, 2015 and March 27, 2017 were complemented by previously published TRAPPIST data and Spitzer as well as Kepler (K2) observations.

The models employed are anything but simple. In fact, the researchers had to examine 35 different parameters, a problem they tackled with new computer algorithms. Simulating orbits until the calculations agree with observed values for the TRAPPIST-1 transits tightens up our previous mass estimates. The work absorbed a year, says Grimm, who goes on to explain:

“The TRAPPIST-1 planets are so close together that they interfere with each other gravitationally, so the times when they pass in front of the star shift slightly. These shifts depend on the planets’ masses, their distances and other orbital parameters. With a computer model, we simulate the planets’ orbits until the calculated transits agree with the observed values, and hence derive the planetary masses.”

What emerges corroborates what the Hubble data show. Rather than being gaseous worlds, the TRAPPIST-1 planets are primarily made of rock. Moreover, they contain significant amounts of volatiles, probably water, given that water in the form of vapor, liquid or ice is the most abundant source of volatiles for the kind of protoplanetary disk that would have produced this system. In some cases, the water can amount to 5% of the planet’s mass. By contrast, Earth has only about 0.02% water by mass. Some of the TRAPPIST-1 planets could thus have 250 times more water than the Earth’s oceans.

TRAPPIST-1 b and c, the two innermost planets, appear to have rocky cores and thick atmospheres, according to this work, while TRAPPIST-1 d, the lightest of the planets (about 30 percent Earth mass) may have a large atmosphere, an ocean or an ice layer. All three possibilities would account for the volume of volatiles thought to match a planet of this density.

TRAPPIST-1 e turns out to be somewhat denser than the Earth, suggestive of a dense iron core and, perhaps, the absence of a thick atmosphere, ocean or ice layer. In terms of insolation from the central star, as well as size and density, this is the planet most like the Earth. The question of why it seems to have a rockier composition than any of its companions remains unresolved.

As to the outer worlds, TRAPPIST-1 f, g and h are distant enough for ice to be frozen on their surfaces, and according to Grimm’s team, are unlikely to have any more than thin atmospheres.

Image: This graph presents known properties of the seven TRAPPIST-1 exoplanets (labeled b through h), showing how they stack up to the inner rocky worlds in our own solar system. The horizontal axis shows the level of illumination that each planet receives from its host star. TRAPPIST-1 is a mere 9 percent the mass of our Sun, and its temperature is much cooler. But because the TRAPPIST-1 planets orbit so closely to their star, they receive comparable levels of light and heat to Earth and its neighboring planets. The vertical axis shows the densities of the planets. Density, calculated based on a planet’s mass and volume, is the first important step in understanding a planet’s composition. The plot shows that the TRAPPIST-1 planet densities range from being similar to Earth and Venus at the upper end, down to values comparable to Mars at the lower end. Credit: NASA/JPL-Caltech.

Among the most interesting things about TRAPPIST-1 is the history of its planetary system, which the paper addresses this way:

The resonant structure of the TRAPPIST-1 system (Luger et al. 2017) is a telltale sign of orbital migration (Terquem & Papaloizou 2007; Ogihara & Ida 2009). The fact that all seven planets form a single resonant chain indicates that the entire system migrated in concert (Cossou et al. 2014; Izidoro et al. 2017). Indeed, orbital solutions generated by disk-driven migration have been shown to be more stable than other solutions (Tamayo et al. 2017b). Whereas most resonant systems are likely to be unstable ((Izidoro et al. 2017; Matsumoto et al. 2012)), the TRAPPIST-1 can be interpreted as a system that underwent a relatively slow migration creating a long-lived resonant system.

We will need the James Webb Space Telescope to move to greater certainty on the question of whether atmospheres actually exist here and what they are made of. I notice that the robotic SAINT-EX Observatory is under construction in Mexico, with the goal of searching for terrestrial planets around cool stars like TRAPPIST-1 (it will also provide ground support for the European Space Agency’s CHEOPS mission). Demory and team hope to apply the computer code they used in the TRAPPIST-1 work on systems detected by SAINT-EX, which should begin operations this year.

The paper is Grimm et al., “The nature of the TRAPPIST-1 exoplanets,” in press at Astronomy & Astrophysics (preprint).


{ 29 comments… add one }
  • Brett February 7, 2018, 17:21

    That’s great news for TRAPPIST-1-e. Earth-size, slightly greater than Earth-density, Earth-level solar insolation – assuming the potentially nasty stellar properties of TRAPPIST-1 didn’t ruin the planet, it could be earth-like for real.

  • Andrei February 7, 2018, 17:39

    Surpriseing but pleasing result, these Trappist planets should be water covered then. This fits with models, and of a type that could be moderated with water to be life bearing.

  • Geoffrey Hillend February 7, 2018, 18:11

    I thought that the Trappist-1 explanets couldn’t have migrated from too far since they are not gases worlds like Neptune but maybe one can’t have a proto Neptune form too close to a proto red dwarf since it’s larger mass and gravity might grab most of the gas? I can’t wait for better spectra on this one.

    • Nicky February 9, 2018, 13:12

      They definitely experienced early migration, but their cores masses are too small to capture a large amount of gas.

  • Geoffrey Hillend February 7, 2018, 18:15

    In a smaller system like a red dwarf star maybe there is just less gas so no gas giants could form in Trappist 1 so there is no reason why it’s exoplanets could not have migrated.

  • Martin February 7, 2018, 20:14

    So much for all the models in which the atmosphere(s) of the planets would have been blown (eroded) away.

    • J. Jason Wentworth February 8, 2018, 1:42

      In size, mass, luminosity, and even distance, that abstemious little star and its close-knit family are reminiscent of Starholme, the fictional–but possible–planet of a very faint red dwarf 52 light-years away. In Arthur C. Clarke’s classic novel “The Fountains of Paradise,” the Starholmers’ long-lived, Bracewell-type “fly-through” interstellar messenger probe–called Starglider–passed through our solar system in 2069. Also:

      Like small dogs, Shetland ponies, and miniature horses and donkeys, little stars live longer–*much* longer–than their larger cousins. For space colonization purposes, with no need to settle the local planets (although it’s an option, if suitable ones are available), but parking large “slow-boat” starships in habitable zone orbits, and building new colonies from the local asteroids’ and comets’ resources, red dwarfs are–for all practical purposes–eternal sources of light, energy for agriculture & industry, and warmth. They’re cozy little celestial campfires, and they are quite possibly conscious, as Dr. Gregory Matloff has hypothesized on “Centauri Dreams” and elsewhere (see: https://www.centauri-dreams.org/2012/06/13/star-consciousness-an-alternative-to-dark-matter/ & https://www.centauri-dreams.org/2015/09/18/greg-matloff-conscious-stars-revisited/ ), and:

      TRAPPIST-1 is probably not far from the red dwarf/brown dwarf boundary, being an ultra-cool M8V – M8.2V star only slightly larger (although more massive, of course) than Jupiter. While flare activity has been observed (see: http://en.wikipedia.org/wiki/TRAPPIST-1 ), TRAPPIST-1’s very low mass, even for a red dwarf (just ~0.089 or so of the Sun’s), may keep the flares fairly mild, so that even weak, atmosphere-shielding planetary magnetospheres can “shrug them off.” Even if our nearest neighbor, little Proxima Centauri, has no habitable zone planets that still retain atmospheres, O’Neill- or McKendree Cylinder-type (see: http://www.orionsarm.com/eg-article/48473a892041c ) space colony vessels arriving there would find no shortages of energy or of raw materials for building new ones, which–thanks to solar concentrator mirrors–need not all occupy the relatively narrow habitable zone.

    • Nicky February 9, 2018, 19:42

      This volatile-rich result does not contradict the current atmospheric escape models. Current models predict only a few Earth oceans of volatile would be lost, and the volatile masses on TRAPPIST-1 planets are orders of magnitude higher, which are impossible to be completely evaporated with such low insolation.

  • Ramses Ramirez February 8, 2018, 2:55

    This is a most interesting, and surprising, analysis (if it turns out to be accurate). There are definitely uncertainties with the assumptions used for the interior models, which would affect the density measurements. Setting that aside, however, assuming that several of these worlds are indeed water-rich then they must have accreted many hundreds, if not thousands, of Earth-oceans during the early stages of planetary accretion. They would have to be true water worlds (with high pressure ices on the sea floor) with enough water to offset the tens to hundreds of Earth-oceans they would have had to lose during the superluminous pre-main-sequence phase. If they had migrated in a bit later, that would help. They would unlikely have continents unless one of them is close to losing most of its initial water inventory. Under such circumstances, the open question is whether such ocean worlds could be habitable or not.

  • Harry R Ray February 8, 2018, 10:48

    It really excites me that SPECULOOS is now operational, and that SAINT-EX will come on line very soon! Boy do I have a mission for them! Kipping et al recently proposed that 5 mile high MOUNTAINS could be detected on a Mars-sized planet in a one day orbit around a white dwarf star! If you could scale things up a bit, substituting TRAPPIST-1d for Mars, TRAPPIST-1 for a white dwarf star, and 50 mile high exo-hypercanes for 5 mile high mountains, ASSUMING THAT TWO such exo-hypercanes are PERMANENT FEATURES(like Jupiter’s Great Red Spot) at the two equatorial regions of TRAPPIST-1d’s terminator, they, too; should be detectable!!! Assuming that a FIRM detection could be made in the next year or so, JWST could be scooped for this planet. If they DO exist, it would mean the following: ONE, TRAPPIST-1d’s atmosphere would have to be TRANSPARENT down to the surface, which would have to be a DEEP GLOBAL OCEAN. TWO: The temperature of this ocean at the terminator would have to be AT LEAST 120 degrees farenheit. THREE: The ocean would be BOILING at the stellar point, producing a HUGE PLUME OF STEAM, which, together with the two exo-hypercanes, would mean that the permanent day side of the planet would be completely covered in either hazes or clouds. Should JWST then CONFIRM these parameters, it would be an excellent CALIBRATION TEST for what it could do for the other, more interesting planets like TRAPPIST-1e

    • ljk February 8, 2018, 17:06

      Harry Ray, you always have fascinating and out-of-the-box insights on these matters. Do you also send your ideas to the parties in question who can actually do something about them? If so, what is your success rate?

      • Harry R Ray February 9, 2018, 10:29

        Only if the principal investigators have websites upon which ANYONE can post comments(LIKE THIS ONE!). Otherwise, I try to encourage third party attempts to get the principal investigators to do a GUEST POST on THIS website(i.e. David Kipping on exomountain detection would be FASCINATING!). I also assume that I am not ALONE in realizing the potential for success in a certain directed set of observations, and any principal investigator in his or her right mind would love to get ironclad details on TRAPPIST-1d before JWST does.

  • Geoffrey Hillend February 8, 2018, 16:56

    A thick atmosphere might block some x ray radiation, but ozone is needed for the ultra violet radiation, so colonists might need plastic domes for the agriculture like on Mars if there is no ozone in the Trappist-1 exoplanet atmospheres. Atmospheres can be partially replenished through volcanism.

  • andy February 9, 2018, 16:50

    Can you still get exposed continents on a volatile-rich planet by having the volatiles stored in the rocky mantle? As far as I’m aware, there are estimated to be a few oceans’ worth of water in the Earth’s deep interior, so what we see at the surface is (fortunately for us land-living creatures) only a fraction of our planet’s volatile inventory. Presumably there’s some point where too much water will be outgassed and the planet ends up with a global ocean/ice layer/supercritical envelope but where is that?

    Then again, I’d guess that allowing for mantle storage of water would change the estimated volatile fractions for the TRAPPIST-1 planets relative to the three-layer models.

    • Nicky February 10, 2018, 2:03

      The planet mantle has the potential to store tens of Earth ocean amount of water and the core can also store tens of Earth ocean’s worth of hydrogen. It indeed might apply to the planets with water mass fraction less than 0.5%. As long as plate tectonics operate on the planet, water degassing through magmatism and regassing through plate subduction can balance the land-ocean distribution of the planet. But it is less likely to happen on TRAPPIST-1 planets except for e and f, because they all have water mass fractions too high to store in the interiors.

  • John walker February 10, 2018, 8:05

    In terms of ecosystem building in this system, simulations have shown that compact orbital grouping will cause permanent axial instabilities through recoprocal gravitational influences. The Trappist system is an ideal case for this effect. The constant ecosystem variance will be a hindrance to the evolution of complex life which needs longer periods of stability.

  • Ronald February 11, 2018, 5:10

    And now of course the news about the nearest transiting planets around a K6 star, and the question: are they rocky planets, i.e. true super-earths, or gas dwarfs?

  • Ramses Ramirez February 11, 2018, 22:14

    The more I think about it, the more I believe either these results (possibly the density measurements) and/or interpretations are incorrect. The authors predict that the innermost planet TRAPPIST-1 b should have a thick atmosphere or ocean/ice (possibly water-rich). But (unlike d-h) the innermost planets (b and c) have NEVER been in the HZ. They have been in a full runaway greenhouse state for the entire lifetime of star, which is ~5 billion years (give or take). Even had “b” accreted a *thousand* Earth-oceans of water, virtually *all* of it would have been lost over this time. The really low density they obtain for “b”is hard to explain unless much of that water is locked up within the interior and is not in the atmosphere or on the surface. Something is strange here and it will require follow-up atmospheric observations to figure out where the issue is.

    • Harry R Ray February 12, 2018, 10:30

      Dr Ramirez: One possible way to find out would be using the STIS instrument on the HST to observe transits of TRAPPIST-1b. If you get very lucky, you might see a “comet tail” a la HD209458 or Gliese 436. If the water vapor dominated atmosphere Gillon et al propose is correct, the most likely result would be a marked difference in the radius vis a vis Spitzer and Kepler observations. If the radius is the same for all three bandwidths, either your “hydrated minerals” or my “thick helium atmosphere”(see comment above)scenario is probably the right one. Oh! And by the way, don’t limit the STIS observations to jist TRAPPIST-1b! Possible variations in the radii of the other planets could indicate what type of atmosphere they have(i.e. transparent in one wavelength and opaque in another).

      • Harry R Ray February 12, 2018, 10:37

        Oops. My “thick helium atmosphere” scenario does not appear in the comments section on this post. It appears in the comments section of a previous post: Probing The TRAPPIST-1 Planet Atmospheres.

    • Nicky February 12, 2018, 22:21

      You have overestimated the efficiency in water-loss. In fact, TRAPPIST-1b is receiving less than 4 times of radiation that Earth receives, which is rather cool compared to other close-in exoplanets. Water is resistant to photoevaporation by experiencing much less thermal evolution and having high mean-molecular weight. Super-Earth with as little as 10% of water won’t be desiccated even under 1000 times of solar insolation after 5 Gyr (Lopez, 2017). The ultraviolet observation from Hubble Telescope indeed reveals that the water-loss on TRAPPIST-1b won’t be greater than 100 Earth oceans even after 8 Gyr (Bourrier et al., 2017).
      The physics behind TTV measurements is well understood, and many of them are mutually proven by RV and TTV.

      • Harry R Ray February 13, 2018, 10:18

        Dr Ramirez is NOT talking about CURRENT conditions. He is talking about conditions in the FIRST ONE BILLION YEARS of the system, when TRAPPIST-1 was a PRE-MAIN SEQUENCE STAR, and was many times more luminous than it is now.

        • Nicky February 13, 2018, 16:20

          I am not talking about the current condition either. I am talking about its whole lifetime. The water-loss rate calculated in the references obviously includes PMS. After the disk dissipates (assuming 10 Myr), 1b experiences no more than 100 times of Earth receives; 50 Myr later, the flux is already down to only 20 times (Baraffe et al., 2015). Water content evolution model on 1b shows no more than 20 Earth oceans would be lost during the first one billion years (Bourrier et al., 2017).

          • Ramses Ramirez February 14, 2018, 4:40

            Nicky, I am not sure where you get that TRAPPIST-1b is “cool” but 4.25 times Earth insolation is a *really* big deal.

            TRAPPIST-1b is *currently* receiving more than twice the radiation that received by Venus, which lost its entire water inventory with much lower fluxes.

            That is present day. It would have been much, much worse a few Gyr ago.

            Bourier et al. assume a much lower (1%) heating efficiency, which is considerably lower than what my study (Ramirez and Kaltenegger 2014) or other similar ones (Luger and Barnes, 2015) have assumed. Even with this low heating efficiency, “b” should have been in a runaway greenhouse state for the entire lifetime of TRAPPIST-1 (~5 Gyr), which is much longer than the 1 Gyr you mention. This is because it was never in the HZ to begin with. It was always in a runaway greenhouse state. It would have easily lost 100 Earth oceans (more at higher heating efficiencies- hundreds or even more), which makes it increasingly unlikely “b” would have ever retained any water. All of this also ignores impactors, that in such small tight orbits, would have further eroded these planetary atmospheres. This also ignores the stellar wind/erosion, which would further deplete these atmospheres. So, the water loss estimates for hydrodynamic escape are actually a lower bound on the nasty things that would have happened to such close-in planets.

            But there are other issues with the Grimm et al. analysis which I have not mentioned. Their chi-squared fits for a few of the planets (f – g) are actually quite bad (~2 or higher), which would affect their mass measurements for the other planets closer in as well. In all, my colleagues and I think it is quite possible that Grimm et al. are underestimating the density of some of these planets, which would push them in a more water-poor regime.

            The only real way to answer this question, however, is through follow-up atmospheric observations. I remain skeptical that these planets are water-rich.

            • Harry R Ray February 14, 2018, 10:52

              Dr Ramirez: Due to the near-failure of the latest Ariane launch, JWST may not launch until late 2019. It would probably be until 2021 before we get even any preliminary JWST data on the TRAPPIST-1 planets’ atmospheres. This opens up a rather large time window to test Gillon et al’s claims simply by monitoring the planets’ transits. What I mean here, is that a long-term ephemeris should be able to be derived based on the mass ranges stated in their paper. Any deviations from this time-frame in the planets’ transit times would support your(and your colleagues)claims.

            • Nicky February 14, 2018, 22:18

              Dr. Ramirez, indeed, Bourier et al. assumed 1% in Astronomy & Astrophysics, 599, L3., but here I was referring to a more recent paper The Astronomical Journal, 154(3), 121. They calculated the efficiencies for each planet and obtained 6%~12%. Plus, I wasn’t not just saying the first one Gyr. In fact, that comment was a reply to Harry Ray who misunderstood your comment. I agree with you that hydrodynamic escape is only one of many factors that drive atmospheric escape. We don’t have much constraint on stellar wind evolution of low-mass stars, but ion escape model shows a rather insignificant loss rate (taking billion years to evaporate an Earth ocean, though PSM might be more active, still…) compared to hydrodynamic escape (Dong et al., 2018).
              A drier regime is possible, but it would contrast current planet formation theory, as the system architecture suggests a water-rich assembly.

              • Ramses Ramirez February 20, 2018, 2:55

                Nicky, an *initial* water-rich assembly is certainly possible , and I would even say that I don’t doubt that part that much. However, the question is whether these planets can hold on to that water or does all of it (or the majority of it) get lost afterwards?
                Observations will shed light on this question.

  • Tom Mazanec February 13, 2018, 12:04

    I count 44 space exoplanet projects and 112 earth exoplanet projects

Leave a Comment