≡ Menu

Desert Planets Around Carbon-Rich Stars?

Habitable zones are always controversial. Bring up the classic definition of a zone where liquid water can exist on the surface and you run into queries about places like Europa, far outside the HZ in those terms but perhaps capable of supporting life beneath the ice. For that matter, exotic forms of life cannot be ruled out in settings like Titan, though they would be nothing like what we’re familiar with on Earth. Nonetheless, refining our methods to look for life on planets with liquid water is a rational way to proceed and we’re developing the needed tools.

As we wait for those tools to be funded and built — projects like Terrestrial Planet Finder are on indefinite hold — we can continue to develop our theoretical models for living planets. On that score, Torrence Johnson (JPL) and colleagues have had interesting things to say lately. Johnson spoke at the American Astronomical Society Division of Planetary Sciences meeting in Denver earlier this month, addressing the question of planets rich in carbon. You would think that plentiful carbon, given its importance in living systems, would be a plus, but it turns out that having too much carbon is a serious problem even for planets in clement orbits like the Earth’s.

Here life has yet another ‘filter’ to get through. For the Earth is made up largely of silicates rather than carbon because the Sun is relatively carbon-poor. We assume that stars with much higher carbon content than the Sun will spawn planets so infused with carbon that they may have layers of diamond. What they apparently won’t have, according to Johnson’s models, is water. Much of the available oxygen in such a system would go into making carbon monoxide, with little oxygen left for making water ice. That means no icy asteroids delivering water to planetary surfaces.

Comets and asteroids are thought to have delivered the bulk of Earth’s water billions of years ago, moving into the inner system from beyond the ‘snow line’ where ice was plentiful. Carbon-rich planetary systems will simply lack this resource. Says Johnson: “There’s no snow beyond the snow line… The building blocks that went into making our oceans are the icy asteroids and comets. If we keep track of these building blocks, we find that planets around carbon-rich stars come up dry.”


Image: This artist’s concept illustrates the fate of two different planets: the one on the left is similar to Earth, made up largely of silicate-based rocks with oceans coating its surface. The one on the right is rich in carbon — and dry. Chances are low that life as we know it, which requires liquid water, would thrive under such barren conditions. Credit: NASA/JPL.

This NASA news release has more, and the paper on this work fully develops the importance of the carbon-to-oxygen ratio, a result not only of elements from the Big Bang but also the hydrogen, helium, nitrogen, silicon, carbon and oxygen inherited from earlier generations of stars. The paper notes:

Planetary systems around stars with compositions that differ from the Sun’s will have planetesimals – formed beyond the snow line – which have a wide range of silicate and metal, carbon and ice proportions. The fraction of silicate plus metal in extrasolar planetesimals should depend strongly on the C/O ratio in a given circumstellar nebula, which controls the abundance of water ice, clathrate hydrates, and stochiometric hydrates in the condensed solids. Other volatile ices are less strongly affected by the host star compositions in our study, although C-bearing ices such as CO (ice and clathrate), CO2 and CH3OH will be more abundant with increasing C/O. The planetesimal population of our own outer solar system, as incompletely known as it is, represents only one trajectory through the planetesimal composition space defined by the possible range of C/O and metallicities seen in other stars.

One day, then, we may find an Earth-mass planet in a star’s habitable zone that, because of its star’s composition, is highly unlikely to have anything alive on it. We’re not yet able to make the kind of spectroscopic observations that would flag life’s presence in a planetary atmosphere, but the paper adds that we can follow up these studies in the near-term by using instruments like the James Webb Space Telescope to study the planet-forming zones of young solar systems and the effect of primordial elements through heavy element enrichment in extrasolar giant planets.

The paper is Johnson et al., “Planetesimal Compositions in Exoplanet Systems,” The Astrophysical Journal Volume 757, No. 2 (2012), p. 192 ff (abstract)


Comments on this entry are closed.

  • Mundus Gubernavi October 29, 2013, 16:41

    If comets and asteroids from the frost line delivered the “bulk” of Earth’s water, then to me it makes the most sense to begin searching in that neighborhood for possible signs of life. Ceres seems like a good place to look. Or maybe we can treat the asteroids like gremlins and toss some water on them and see if anything happens. lol

    Question, did Mars also receive its water from asteroids?

  • Kamal Ali October 29, 2013, 20:08

    Yes, Mars also received its water from the asteroids and comets , same as Earth.

  • Rob Henry October 29, 2013, 20:40

    The logic here seems rather convoluted, as rareEarth arguments often are. Just the previous article on Centauri Dreams placed both Super-Earths and Mini-Neptunes as more common than Earths. The usual problem that rareEarthers have with those heavier planets is too much molecular hydrogen in the atmosphere. For high C:O ratio protoplanetary disks the condensing cometary ices would be largely carbon monoxide, and that is what would be delivered to the inner carbon planets here, not water. Now I wonder what happens if you mix CO and H2 at Earth like temperatures for a few million years?

  • Eniac October 29, 2013, 20:55

    I don’t think this problem can be evaluated by looking at C/O ratios alone. Si is just as important. If there were more carbon, more of it would probably be bound in SiC and diamond, perhaps leaving more oxygen to form water, rather than less.

    It would be more pertinent to look at the (C+Si)/O ratio to give a lower bound as to how much oxygen would be available for water. It is a lower bound because of SiC, which would be expected to form more readily if the C/Si ratio is closer to 1/2. Also, this assumes the worst case that hydrogen is only the third choice for oxygen to bind to, after silicon and carbon, which I am not at all sure of. Is there a chemist in the house?

    Something tells me that the limiting factor for water on planets or asteroids is the abundance of hydrogen, not oxygen. This would mean that the C/O ratio would not make any difference in how much water there is. What would change instead would be the ratio between SiC and SiO2. A crust richer in silicon carbide and diamond and poorer in silicates, but with just as much water. A hard, abrasive world, if you will, but not a desert.

  • Tarmen October 30, 2013, 11:11

    One can only imagine the varied landscapes of these borderline earth-like worlds, whether desert sculptures or diamond mountains gleaming in the sun. Or towering waterfalls and stormy oceans. Maybe silicon forests and silicon critters. Wilderness worlds worth a look, even if unsuitable for our kind.

  • Mike Lockmoore October 30, 2013, 12:27

    Mmmmm – dessert planets, full of carbon-rich materials like C6H12O6! Oh wait, did you day _desert_ planets? Drat.

  • andy October 30, 2013, 14:22

    One question is whether the C/O ratios are as high as has been claimed. There seem to be a few suggestions that the values have been overestimated, e.g. Nissen (2013) and Teske et al. (arXiv 2013).

  • Rob Henry October 30, 2013, 18:34

    Eniac, your comment about silicon surprises me. I would agree with the thrust of it, but for one thing: is silicon ever that high in ratio to carbon and oxygen?? When I look at Wikipedia’s mass fraction in the Milky Way, and divide by Earth standard molecular weights, I find Si sixteen times less common than C.

    Two more quibbles I have
    1) unlike C:O there is a problem in how you measure the stoichiometry of its effectiveness. As oxygen levels drop below critical levels, it doesn’t just form hygroscopic SiC it also forms metallic Si which, if memory serves, becomes sequestered in planetary cores just as iron does if you have too much of it. Silicon nitride should also be common but how active is the geology on these planets, and how much would the free water be expected to cycle through it where one silicone atom can take out two oxygen?
    2) Are you sure Mg is not more important here. It is very reactive, and will not end up in the metallic portion those planetary cores, and its molecular abundance in the Milky Way is 95% of silicon.

  • stephen October 30, 2013, 19:56

    Wikipedia has an article on hypothetical types of biochemistry. Sulfuric acid is one solvent cited, with Alkenes replacing carbonyls. I don’t remember who suggested silicon based oxygen breathers in oceans of sulfuric acid. Isaac Asimov wrote a lot about alternative biochemistries, including carbon monoxide.

  • Ron S October 31, 2013, 13:13

    Mike: “Mmmmm – dessert planets, full of carbon-rich materials like C6H12O6!”

    Unfortunately there is a book in existence (and on my bookshelf) that deals with this very subject. The book is National Lampoon’s “Doon” by Ellis Weiner, a parody of “Dune”. Get the joke? Well, the joke runs for over 200 pages.

    Beware the giant pretzels!

  • Eniac November 1, 2013, 23:15

    Rob: My comments on silicon were driven by the fact that it is the second most abundant element in the Earth’s crust (after oxygen). Carbon is comparatively rare, by two orders of magnitude or so. Since we are talking planets, crustal abundance might be more relevant than galactic abundance.

    You are right that the metallic form of silicon would make available more oxygen to silicon rich planets (like the Earth) as compared to silicon poor planets. However, metallic silicon is not found in significant amounts on Earth, so this does not seem to be a factor. I also think it is not plausible for there to be metallic silicon as long as there is enough oxygen to bind it. The same probably goes for diamond. However, silicon carbide does actually occur naturally, and would likely much more common if there were more carbon around.

  • Eniac November 1, 2013, 23:31

    It seems to me that the reason there is so little carbon on Earth is that it less readily condenses into dust, since CO2 is the most volatile of CO2, H2O, and SiO2. To the extent that rocky planets are made from dust, not gas, they would be expected to be poor in carbon.

    A higher C:O ratio would then mean less material for planets, but it is not clear to me how the ice/rock ratio would be affected, if at all. You could argue that ice would form more readily than rock when there is less oxygen, since it requires only one oxygen atom per molecule, instead of two. This would mean a higher ice/rock ratio in systems with high C:O, opposite to what is proposed in the article.

    Silicon carbide water-worlds, then?

  • Rob Henry November 2, 2013, 20:50

    Eniac, to my eyes you seem to phrase the situation in a novel manner. To me, an easier (more correct?) way to oversimplify is as follows…At C:O 1 it starts with the same problem for oxygen.

    Other points I wish to add are
    1) looking at the Earth’s crust is a good start but it could be misleading due to the degree that it has been reworked. An example is your statement that
    “My comments on silicon were driven by the fact that it is the second most abundant element in the Earth’s crust (after oxygen)”
    But if you look at what is known of the chemistry of our mantel, one of the first things that pops up is that it must be magnesium rich and silicon poor. I often wonder whether the best description of the ‘rock’ portion of Earth is magnesium oxide with a bit of silicate added, or the other way around?!
    2) yes that metallic silicon will still be a rare form at C:O of 0.8, you would probably have to get a ratio more like 1, but, if you go high enough, eventually plenty of it will begin to precipitate (at least in some older models, and I doubt that has changed)
    3) I am unsure if the snowline for CO2 is ever anything like as important as that for CO. It doesn’t help that CO2 should form much later than CO, and from it.

  • Rob Henry November 2, 2013, 20:55

    oops,the above first paragraph has misprinted. It should read.

    Eniac, to my eyes you seem to phrase the situation in a novel manner. To me, an easier (more correct?) way to oversimplify is as follows…At C:O 1 it starts with the same problem for oxygen.

  • Rob Henry November 2, 2013, 21:00

    Its done it again. I will have to rephrase the bit after those dots

    At C to O below 0.8 the protoplanetary disk starts with neigh all its carbon locked up in CO, and it must work its subsequent chemistry hard to release any carbon, and if C to O is above 1 it starts with the same problem for oxygen.

  • Eniac November 3, 2013, 14:51


    At C:O 1 it starts with the same problem for oxygen.

    Well, yes and no. H and Si (or other metals) compete with C for oxygen, though, and only the products of the latter reactions are going to end up in asteroids or planets. My point is that this may not have been duly considered. What determines how wet or dry a planet is is the ratio of H2O/SiO2 (ice/rock), which should be independent of C, in first order.

    I often wonder whether the best description of the ‘rock’ portion of Earth is magnesium oxide with a bit of silicate added

    All abundance tables that I have seen show Mg at much less than Si, especially in the crust. Even if Mg were more common, would it not support my arguments just the same, if lumped together with or considered instead of Si?

    As for the CO/CO2 snow lines, it seems that CO is even more volatile than CO2, so, again, my argument seems to remain intact after duly considering yours.

  • Eniac November 3, 2013, 15:02

    As an aside: All this, to me, poses the question “Where did the carbon go?”. Most reasonably, CO and CO2 being gases, it went into the gas giants, were, being heavier than H and He, it must have sunk towards the interior. So, perhaps the notion of a giant diamond at the cores of the larger planets is not as misplaced as it seems at first…

  • Rob Henry November 4, 2013, 16:43

    Eniac, surely the SiO2/H2O (putting aside Mg that I am still placing as equally important) is irrelevant when at a C:O around 1 there is far more silicon nitride and carbide than oxides precipitated from the protoplanetary disk. Many primordial silicates have hydroxide ions in them, and some bound water of hydration, so that you can squeeze water out of them at high temperatures and pressures. By contrast, those carbides (and nitrides?) are hygroscopic and react strongly with water over geological time. A carbon planet’s crust will be able to absorb ocean loads of water, and its mantel will release none.

    PS Eniac when you asked where did all the carbon go? The part I always wonder is what proportion of it ended in comets?

  • Rob Henry November 4, 2013, 18:23

    Eniac, I have thought further about your retention of high hopes for the importance of the CO2 snowline v CO snowline. The H2O snowline should form around 4AU, but not the CO line until 40AU. My very first post on this thread assumed that, in the absence of water, the critical snowline should be for CO, but now I wonder if 40AU is just too far out for comet and giant planet formation. Unfortunately, no one seems interested in calculating just where CO2 snowline lies, but at a guess from the relative mp’s it should be well less than 20AU. Perhaps it is important after all (in the case of carbon planets)

  • Eniac November 4, 2013, 23:05


    Eniac, surely the SiO2/H2O (putting aside Mg that I am still placing as equally important) is irrelevant when at a C:O around 1 there is far more silicon nitride and carbide than oxides precipitated from the protoplanetary disk.

    Surely, if a lot of the rock were to be free of oxgen (as in Si/Mg nitrides and carbides), that would leave more oxygen to form water? Doubly so in the case of carbides, which would free additional oxygen by removing the carbon: Each SiC means 4 more H2O when forming instead of SiO2 and CO2

    The part I always wonder is what proportion of it ended in comets?

    Good idea. Any comets which formed outside of the CO/CO2 snowlines would presumably contain substantial amounts of CO2. That would be all of them, I suppose, if they come from the Oort cloud?

  • Eniac November 4, 2013, 23:25

    Rob: Note that a C:O of 1 is not much more than double that of the lower end of the range (0.4, I believe). The increased binding of oxygen by the extra carbon could easily be counteracted by 1) increased formation of carbides, and 2) decreased formation of rock vs. ice. In any case, I think it is clear that C:O alone is not enough data, you have to know the amount of Si/Mg, too.

  • andy November 5, 2013, 16:25

    There has definitely been some study of the impact of the Mg/Si ratio on terrestrial planets formed with varying C/O abundances, e.g. Carter-Bond et al. (2012) – see also the references/citations.

  • Rob Henry November 5, 2013, 17:18

    Eniac, you still don’t seem to be acknowledging how much CO rules. The reasoning is somewhat like this.
    1) A molecular cloud fragments and collapses. A piece of it forms a protostar
    2) The material around that protostar remains dust and gas, and plays little part in planet formation.
    3) The protostar is spinning too fast and its disk spreads out. The portions of this that provide the material to form planets are so hot that the ONLY (or first forming) stable molecule is CO. The kinetics at that temperature allow that initial reaction to almost go to completion.
    4) As the disk cools, the equilibrium for other molecules becomes favoured, but the kinetics of CO reactions are so slow at these densities that even at 1000K, it would take three or four years to near the new equilibrium.
    5) the disk cools way to quick to allow this – at least apparently, judged by the result of these simulations.
    6) particularly hard to form is water, and with such high C:O, neigh every drop must come from Fischer-Tropsch. This is very slow and inefficient, and provides more tar than retained H2O.

  • Eniac November 5, 2013, 21:11

    Rob: It looks like you know a lot more about astrochemistry than I do. Why is it hard for water to form, when H is the most common of elements, and O the most common of the heavier ones? Did you really mean 3 or 4 years, not thousands or millions? That seems awfully fast to me. Where can one read about all this?

  • stephen November 6, 2013, 0:39

    The wikipedia article on the Antennae Galaxies indicates that they have a much higher abundance of several elements–presumably because the two galaxies colliding induced more star formation. So unusual biochemistries might be more abundant. What percentage of the universe’s galaxies are colliding?

  • Rob Henry November 6, 2013, 16:59

    Eniac asks “Where can one read about all this?” and I would like to say a modern review such as

    The problem is that these assume so much and explain so little, so why not go back to the popular article in which the phenomenon was first explained. Its author tried to term the water shortage in high C:O disks ‘carbonosis’ but that never caught on.

    Carbonosis: Organic Dessication and the Fermi Paradox
    By Stephen Gillett
    The March 1993 issue of Analog

    And, yes I meant three our four years for that CO reaction. 1000K is critical temperature below which it would take your thousands of years, and above which CO is too favoured. Apparently CO falls through it too fast because the temperature gradients are extreme from protoplanetary disk surface to its central plain portions and the mixing rate too high. And I know… that sounds the wrong result to me too, but it is what seems to persistently come out of those simulations.

  • stephen November 7, 2013, 9:47

    “Noble gases can also form endohedral fullerene compounds where the noble gas atom is trapped inside a fullerene molecule.”

    I realize xenon is about as abundant as tin, but the thought of xenon-based life seems really cool. No, life wouldn’t be xenon-based, but xenon might be involved.

    “Chemical compounds consisting of noble gases combined with hydrocarbon molecules – a feat previously thought to be unattainable – have been created as the result of the work of researchers at the Hebrew University of Jerusalem.”

    Not to be confused with carbohydrates, but…

    Also, google “X-based life” for almost any element and you’ll get hits. In the Voyager episode, “Hope and Fear”, a xenon-based life-form was mentioned. But scientific accuracy was not a priority for those writers.

    Wikipedia also mentions the possibility of helium hydride.