≡ Menu

Young Planets, Young Stars

We’re going to be bringing both space- and ground-based assets to bear on the detection of rocky planets within the habitable zone in coming years. Cool M-class stars (red dwarfs) stand out in this regard because their habitable zones (in this case defined as where water can exist in liquid form on the surface) are relatively close to the parent star, making for increased likelihood of transits as well as a larger number of them in a given time period. Ground observatories like the Giant Magellan Telescope and the European Extremely Large Telescope should be able to perform spectroscopic studies of M-class planets as well.

So consider this: We have a high probability for planets around these stars (see Ravi Kopparapu’s How Common Are Potential Habitable Worlds in Our Galaxy?), and then factor in what Ramses Ramirez and Lisa Kaltenegger have, the fact that before they reach the main sequence, M-class stars go through a period of ‘infancy’ that can last up to 2.5 billion years. It’s an important period in the life of any planets around them because, as we saw last week in a paper by Rodrigo Luger and Rory Barnes, huge water loss can occur and there is the possibility for runaway greenhouse events that complicate a planet’s habitability.

But we need to keep digging at this issue because early M-dwarf planets are interesting targets for detection. In their new paper, Ramirez and Kaltenegger (Cornell University) point out that in the period before entering the main sequence, these stars have a luminosity much higher relative to their early main sequence values than F or G stars. In fact, an M8 star can be 180 times as bright during its contraction stage as it will be when entering the main sequence. Our G-class Sun, by contrast, was only twice as bright during this period.

What we get is a habitable zone that is located further away from the young star than we would otherwise expect, in a pre-main sequence period that, for some of the smallest stars, can last for over two billion years, meaning there is the potential for life to develop on the surface, perhaps moving below ground as the star’s luminosity gradually declines. The paper computes the luminosities of stars in classes F through M as they go through the contracting phase and onto the main sequence, with clear effects on the orbital distance of the habitable zone around the star.


Image: Changes in the position of the habitable zone over time. Credit: L. Kaltenegger, R. Ramirez/Cornell University.

The coolest of these stars, an M8 dwarf, shows an inner zone of the HZ at 0.16 AU that moves to 0.01 AU at the end of the 2.5 billion year pre-main sequence phase. In the same period, the outer edge of the HZ moves from 0.45 AU to 0.3 AU. Contrast that with our Sun, where the inner HZ edge changes from 1 AU to 0.6 AU at the end of the pre-main sequence stage after a comparatively swift 50 million years. The Sun’s outer HZ edge moves from 2.6 AU to 1.5 AU. Variations in planetary mass appear to have only a small effect on these results.

The potential for water loss on the planets of young M-class stars is enormous, much higher for this than other stellar types. Ramirez and Kaltenegger shows that planets orbiting an M8 star at distances corresponding to the inner and outer main sequence habitable zone could lose up to 3800 and 225 Earth oceans of water respectively. An M5 planet loses 350 and 15 oceans respectively, while an M1 world loses 25 and 0.5 oceans. If we look at K5 and G2-class stars, their planets at the inner and outer habitable zone lose 1.3 and 0.75 Earth oceans respectively. These figures, remember, are for planets that end up being in the habitable zone later in the star’s life, when it is firmly established on the main sequence.

Is this water loss a fatal blow to the planet’s chances for habitability? Ramirez and Kaltenegger think not, a point that is worth emphasizing. From the paper:

Our results… suggest that planets later located in the MS HZ orbiting stars cooler than ~ a K5 may lose most of their initial water endowment. However M-dwarf MS HZ planets could acquire much more water during accretion than did Earth (Hansen, 2014), or water-rich material could be brought in after accretion, through an intense late heavy bombardment (LHB) period similar to that in our own Solar System (Hartmann 2000; Gomes et al. 2005). The second mechanism requires that both the gas disk has dissipated and the runaway greenhouse stage has ended.

The situation of our own Earth is not without interest here. Ramirez commented on the issue in this Cornell news release:

“Our own planet gained additional water after this early runaway phase from a late, heavy bombardment of water-rich asteroids,” says Ramirez. “Planets at a distance corresponding to modern Earth or Venus orbiting these cool stars could be similarly replenished later on.”

We can also factor in possible planetary migration, bringing planets rich in volatiles into the habitable zone after the star has reached the main sequence. Bear in mind that the pre-main sequence period can last 195 million years for an M1 star and up to 2.42 billion for a cool M8, meaning that a planet that has lost its water through this ‘bright young star’ period could regain water and become habitable even before the star reaches the main sequence.

In terms of searching for small, rocky planets with the potential for life, then, we need to keep in mind how much farther out the habitable zone is during a young star’s pre-main sequence phase, for the larger separation between planet and star will make such planets easier to find with next-generation telescopes. This work shows that planets around stars in spectral type K5 and cooler — those that will later be located in the habitable zone during the star’s main sequence phase — can easily receive enough stellar flux to lose much or all of their water. They will need to accrete more water than Earth did or else rely on water delivered through later bombardment. They are prime candidates for remote observation and characterization.

Moreover, we learn here that planets outside the habitable zone during the star’s stay on the main sequence may once have had conditions right for life to develop, with the potential for continuing beneath the surface as the star cools. “In the search for planets like ours out there, we are certainly in for surprises,” says Kaltenegger, and it’s clear that young stars, especially M-dwarfs, offer rich targets for observation as we parse out what makes a planet habitable.

The paper is Ramirez and Kaltenegger, “The Habitable Zones of Pre-Main-Sequence Stars,” in press at Astrophysical Journal Letters (preprint).


Comments on this entry are closed.

  • Ashley Baldwin December 8, 2014, 15:46

    Two almost dimetrically opposed papers in a week. What a strange place the Universe is seemingly turning out to be. In simulation.And exciting. But lest we forget, both these excellently crafted publications are based on simulations rather than real data. When the first RV exoplanet was discovered in the mid nineties , Geoff Marcy’s US team confirmed the existence of the “Hot Jupiter” almost immediately because they had been observing the same system themselves but we’re looking much further out , based around our own system’s gas giant disposition. Subsequently they found lots more. So the real Universe is full of surprises too and that’s why we need to be out there looking out for real with Kepler, with WFIRST, and TESS and PLATO….. Can’t wait for what turns up next !

  • Antonio December 9, 2014, 9:06
  • Harry R Ray December 9, 2014, 10:32

    Looks like Kepler 186f is STILL ok (ie via the results in BOTH papers) but KOI 3138.01 (see my EARLIER comment on the ” detecting super-earth transits with ground based telescopes” post) is NOT a true mars analog after all. I did some checking and found that KOI 3138 is smaller and cooler than Proxima Centauri (PROBABLY a M8) and the planet candidate, if it exists at all (check out the HEC habitable zone candidates table, and CLICK “KOI 3138.01. You will see that the orbital period is VERY POORLY DEFINED,ie 8.7 PLUS OR MINUS 3 days, which indicates a false positive or the MOST EXTREME TTV’s yet) it is completely deviod of any kind of atmosphere. KOI 3138 was RECLASSIFIED from a M! to a M8 a couple of years ago,and its distance shortened to around 225 light years. I was unaware Kepler was even sensitive enough to detect such a cool star that far away. If the entier K2 project is completed without incedent (ie all ten fields) we may see DOZENS of planet candidates around stars smaller and cooler than Proxima Centauri

  • Michael December 9, 2014, 15:40

    Just wondering if Venus type worlds will be numerous around Red dwarfs. Imagine a Venus type world created by the long higher luminous contraction phase and then the carbon dioxide gas liquefying as the temperature drops below the critical point. You could have a deep transparent ocean circulating, maybe even an abode for life.



  • Terraformer December 9, 2014, 16:30


    Hmmm. That, perhaps, opens up the possibility that Venus analogues might be “naturally terraformed”, if later impactors replenish the water. They’d be Venus like for the first billion or so years, but once the insolation drops, the CO2 (and water) would rain out, reacting with the rocks to form carbonates. Though during this time, they’d have a ~7 bars atmosphere of CO2. If they’ve got abundant water though, and free oxygen leftover… you wouldn’t need much to settle such a world, especially if you can mine the surface.

  • Ramses Ramirez December 10, 2014, 12:44

    Hello everyone. Thank you for the interest in our work!
    As many of you have mentioned, a major point in our paper is that the superluminous phase for young M-stars need not spell doom and gloom for potentially habitable planets. There are certainly ways that such worlds can retain their habitability, including late accretion of volatiles or inward planetary migration from far away when the star becomes less bright. Moreover, the Hansen (2014) study suggests that M-dwarf habitable zone (HZ) planets may have up to 250 Earth oceans of water available for accretion. If true, some M-dwarf HZ planets could retain substantial volatile inventories even after the initial superluminous phase.

  • Dave Moore December 10, 2014, 15:57

    There is another point to consider. Curiosity’s latest findings on Mars show it was much warmer and wetter than models indicate. This may mean the outer edge of the Habitable Zone is further out than we estimate.

  • Galacsi December 10, 2014, 18:08

    How these planets which are tidally locked can loose all their water ? As I see it when atmospheric pressure go down water will condense on the dark side. First as a liquid then as ice.And stay there. After all there is probably some ice on the moon and on Mercury in the polar regions. This is the same but at a much greater scale.

  • Michael December 11, 2014, 4:21

    @Galacsi December 10, 2014 at 18:08

    ‘How these planets which are tidally locked can loose all their water ? As I see it when atmospheric pressure go down water will condense on the dark side. First as a liquid then as ice.And stay there. After all there is probably some ice on the moon and on Mercury in the polar regions. This is the same but at a much greater scale.’

    These planets will be further out and therefore less likely to be tidally locked. There will always be some degassing of the mantle and cometery replenishment but not alot I would think. I like the idea of a transparent CO2 ocean but it is unlikely to be as the winds would more than likely turn up the ground and form a lot of dust. If we look at Venus and apply it’s atmosphere to this situation the water condensed out would be less than 3cm of coverage of the planets surface.

  • Ramses Ramirez December 11, 2014, 15:01

    @ Dave Moore

    You raise an interesting point that is directly related to another of my big research interests, the climate of early Mars. The recent press release from Gale Crater does seem to suggest that early Mars may have been a warm and wet planet for some period of time (at least tens of millions of years, if not for longer). Their inference that an ocean was probably needed to supply the lake, if true, would really strengthen the warm early Mars hypothesis. In my opinion, it is hard to envision how such an ocean would have existed at all if mean surface temperatures were many 10s of degrees below freezing!

    With respect to your comment, in my habitable zone work (for main-sequence stars) with Penn State collaborators, our empirical outer edge limit assumes that early Mars was warm and wet. And if true, as Dave says, would push our computed present day outer edge limit from ~1.67 Au to at least 1.77 AU.

    Earlier this year, I published a Nature Geosc. paper where we argued that a CO2 greenhouse, supplemented by hydrogen, could have provided the required warming for early Mars. We are working now on how to test this hypothesis..

  • Galacsi December 11, 2014, 16:58


    Sorry , but I must say but you did not adress my objection at all.

  • Rob Henry December 11, 2014, 19:43

    Galacsi, you are simultaneously overrating and underrating the effect of tidal lock. To a first approximation it goes like this. In a thin atmosphere like Earth’s, everything freezes out on the far side INCLUDING the atmosphere itself. Else the atmosphere is thick enough to efficiently transfer heat to the far side as per Venus and the temperature becomes very even over the entire surface.

  • Dave Moore December 12, 2014, 1:57

    @Ramses Ramirez

    Why Hydrogen as the greenhouse gas? Why not Methane? Methane seems more likely as a major constituent of the early Martian atmosphere than Hydrogen.

    I find details of Martian geology the most fascinating discovery in our solar system at the moment. Gale crater, in particular, is the story of a dying planet–Barsoom–a record of events in our solar system from a time when all we have on Earth is the odd Zircon to go on.

    The periodicity of the formations fascinates me. Are they the result of Mars cycling in and out of a snowball condition? Or are they due to changes in obliquity and orbital eccentricity? Did the great bombardment cause abrupt, permanent changes in Mars’s atmosphere, or did it thin gradually?

    One day we will find out. I’m hoping it is sooner rather than later.

  • Ramses Ramirez December 12, 2014, 10:43

    @ Dave Moore

    Why hydrogen? The best reason is that is one of the few secondary greenhouse gas that has been shown to work! :)

    Although the state of the martian mantle during the purported warm/wet period (3.8 – 3.6 Ga) is unknown, it is generally assumed that the mantle was highly reduced (i.e., O2-poor). In that case, one would expect volcanoes to belch hydrogen out in prodigious amounts. Hydrogen is a symmetric molecule, and by itself, is normally a poor greenhouse gas. However, if broadened by a background gas (e.g. CO2 or N2) , H2 starts absorbing in wavelength regions where CO2 and H2O do not, providing a lot of additional warming. We show that early Mars could have been warmed with 5 – 20% H2.

    We tried methane as well. Unfortunately, the major problem with CH4 is that it likes to absorb in the upper atmosphere, creating large stratospheric inversions that counteract act the greenhouse warming near the surface. We could not get more than 1K or so of additional warming with CH4 at reasonable concentrations.

    One thing to consider is that the periodicity of formation of features on the martian surface does not necessarily imply transient warming events in a cold climate. On Earth, terrains evolve over time too. Rivers meander and break, forming new pathways. Lakes dry up and new ones appear in other places. But Earth is clearly a warm and wet world. All of the mechanisms you mentioned have been invoked to explain how the fluvial features on Mars (including the vast ancient valley networks and open-lake basins) could have been formed in a predominantly cold climate, but all of them underestimate the water amounts required to form them. In my opinion, a warm and wet early Mars (irrespective of what combination of gases does the trick) offers a more natural explanation, explaining what is seen on the surface. So, I would say that early Mars was warm and wet for some time (~100 Myr or so) before gradually becoming drier and colder as time went on. We see geologic evidence for this as the geomorphology of the fluvial features change and decrease in size over time.

  • Ramses Ramirez December 12, 2014, 10:51

    Another point to consider (as mentioned in our paper) is that the superluminous phase in mid- to late-M dwarf systems may be sufficiently intense to trigger water loss in some water-rich planetesimals located beyond the ice line. Thus, if the planetesimals are dry to begin with, other concerns (like where the moisture goes in a tidally-locked planet located in the main-sequence HZ) become less relevant. In this case, planetesimals located farther out (or a late heavy bombardment-style event) may be able to replenish these planets.

    Previous papers (e.g., Lissauer 2007 and Raymond et al., 2007) have also argued that the tightly-packed nature of M-dwarf systems leads to more energetic collisions with these planetesimals, leading to atmospheric erosion. If the atmosphere becomes too thin through this, it doesn’t bode well for a planet either (whether tidally-locked or not).

  • Michael December 13, 2014, 12:13

    @Ramses Ramirez December 12, 2014 at 10:51

    ‘Another point to consider (as mentioned in our paper) is that the superluminous phase in mid- to late-M dwarf systems may be sufficiently intense to trigger water loss in some water-rich planetesimals located beyond the ice line.’

    Once a dusty crust is formed on a comet it will have a protective effect on the ices remaining below leaving a substantial amount remaining. If there is a mechanism for bringing the ‘belt’ inwards such as a large planet then there is a better chance of replenishment of water. These worlds will most likely remain dry or just waiting for collisions, but having said that their stars live a long time, maybe a 100 times that of ours leaving plenty to accumulate water.

  • Rob Henry December 14, 2014, 17:17

    Here’s what I can’t understand Ramses Ramirez. Let’s forget all non thermal loss mechanisms, and stick the bare minimum – Jeans loss. From this alone, all H2 would be lost from a 273K atmosphere in much much less that 100 million years if the escape velocity is less than 9.2 km/s, yet for Mars it is only 5.0. Your high proportion H2 atmosphere must come close to allowing hydrodynamic loss, and that would have consequence for all light gasses of Mars, not just H2 (moreover, its loss would be insanely rapid).

  • Ramses Ramirez December 14, 2014, 23:41

    @ Rob Henry

    Yes, great point! You are right that H2 escape would be insanely rapid and hydrodynamic. At these concentrations (5 – 20% H2), the fastest it would go would be at the diffusion limit. This is what we assumed in that paper, in lieu of a sophisticated 3-D hydrodynamic escape early Mars model. Moreover, we think the early martian mantle was rather reduced (orders of magnitude more so than the Earth, as suggested by the SNCs and meteorites ALH84001 and “Black Beauty”), so assuming diffusion-limited escape and a heat flux per unit area similar to that of Earth’s, hydrogen fluxes coming from volcanoes would be high enough to build up ~2 – 3 % H2 or so. We can then argue for higher concentrations than these by pointing out that spherical geometry (in a proper 2-D or 3-D model) and magnetic fields would likely slow down escape below the diffusion limit, allowing H2 to build up. However, this is an excellent area for future work.

  • Rob Henry December 19, 2014, 19:16

    @Ramses Ramirez, your reply set me thinking. You have obviously done a lot of work on this, and you have reminded me of the part of today’s atmospheric models of which I have the greatest doubts. I know from theory that water clouds don’t nucleate at nearly the rate observed, but that other factors, such as cosmic rays, meteoric dust, and dimethyl sulphide are needed to solve that gap. I believe that the problem has never been solved in detail, and so the observed values must have been assumed to be the natural ones. All the above change with time and life, thus it is entirely possible that water vapour on early Mars produced less cooling, and a greater greenhouse effect than today’s models suggest.