The exoplanet K2-18b has been all over the news lately, with provocative headlines suggesting a life detection because of the possible presence of dimethyl sulfide (DMS), a molecule produced by life on our own planet. Is this a ‘Hycean’ world, covered with oceans under a hydrogen-rich atmosphere? Almost nine times as massive as Earth, K2-18b is certainly noteworthy, but just how likely are these speculations? Centauri Dreams regular Dave Moore has some thoughts on the matter, and as he has done before in deeply researched articles here, he now zeroes in on the evidence and the limitations of the analysis. This is one exoplanet that turns out to be provocative in a number of ways, some of which will move the search for life forward.

by Dave Moore

124 light years away in the constellation of Leo lies an undistinguished M3V red dwarf, K2-18. Two planets are known to orbit this star: K2-18c, a 5.6 Earth mass planet orbiting 6 million miles out, and K2-18b, an 8.6 Earth mass planet orbiting 16 million miles out. The latter planet transits its primary, so from its mass and size (2.6 x Earth’s), we have its density (2.7 g/cm2), which class the planet as a sub-Neptune. The planet’s relatively large radius and its primary’s low luminosity make it a good target to get its atmospheric spectra, but what also makes this planet of special interest to astronomers is that its estimated irradiance of 1368 watts/m2 is almost the same as Earth’s (1380 watts/m2).

Determining an exosolar planet’s atmospheric constituents, even with the help of the James Webb telescope, is no easy matter. For a detectable infrared spectrum, molecules like H2O, CH4, CO2 and CO generally need to have a concentration above 100 ppm. The presence of O3 can function as a stand-in for O2, but molecules such as H2, N2, with no permanent dipole moment, are much harder to detect.

The Hubble telescope got a spectrum of K2-18b in 2019. Water vapor and H2 were detected, and it was assumed to have a deep H2/He/steam atmosphere above a high pressure ice layer over an iron/rocky core, much like Neptune. On September 11 of this year, the results of spectral studies by the James Webb telescope were announced: CH4 and CO2 were found as well as possible traces of DMS (Dimethyl sulfide). No signal of NH3 was found. Nor was there any sign of water vapor. The feature thought to be water vapor turned out to be a methane line of the same frequency.

Figure 1: Spectra of K2-18b obtained by the James Webb telescope

This announcement resulted in considerable excitement and speculation by the popular press. K2-18b was called a Hycean planet. It was speculated that it had an ocean, and the possible presence of DMS was taken as an indication of life because oceanic algae produce this chemical. But that was not what intrigued me. What caught my attention was the seemingly anomalous combination of CH4 and CO2in the planet’s atmosphere. How could a planet have CH4, a highly reduced form of carbon, in equilibrium with CO2, the oxidized form of carbon? A search turned up a paper from February 2021: “Coexistence of CH4, CO2, and H20 in exoplanet atmospheres,” by Woitke, Herbort, Helling, Stüeken, Dominik, Barth and Samra.

The authors’ purpose for this paper was to help with the detection of biosignatures. To quote:

The identification of spectral signatures of biological activity needs to proceed via two steps: first, identify combinations of molecules which cannot co-exist in chemical equilibrium (“non-equilibrium markers”). Second, find biological processes that cause such disequilibria, which cannot be explained by other physical non-equilibrium processes like photo-dissociation. […] The aim of this letter is to propose a robust criterion for step one…

The paper presents an exhaustive study for the lowest energy state (Gibbs free energy) composition of exoplanet atmospheres for all possible abundances of Hydrogen, Carbon, Oxygen, and Nitrogen in chemical equilibrium. To do that, they ran thermodynamic simulations of varying mixtures of the above atoms and looked at the resulting molecular ratios. At low temperatures (T ≤ 600K), they found that the only molecular species you get in any abundance are H2, H20, CH4, NH3, N2, CO2, O2. At higher temperature, the equilibrium shifts towards more H2, and CO begins to appear.

Some examples of their results:

If O > 0.5 x H + 2 x C ––> O2-rich atmosphere, no CH4
If H > 2 x O + 4 x C ––> H2-rich atmosphere, no CO2
If C > 0.25 x H + 0.5 x O ––> Graphite condensation, no H20

They also used the equations to tell what partial pressures of the elemental mixture will produce equal pressures of the various molecules:

If H = 2 x O then the CO2 level will equal CH4
If 12 C = 2 x O + 3 x H then the CO2level will equal H20
If 12 C = 6 x O + H then the H20 level will equal CH4

To summarize, I quote from their abstract:

We propose a classification of exoplanet atmospheres based on their H, C, O, and N element abundances below about 600 K. Chemical equilibrium models were run for all combinations of H, C, O, and N abundances, and three types of solutions were found, which are robust against variations of temperature, pressure, and nitrogen abundance.

Type A atmospheres[which] contain H20, CH4, NH3, and either H2 or N2, but only traces of CO2 and O2.

Type B atmospheres [which] contain O2, H20, CO2, and N2, but only traces of CH4, NH3, and H2.

Type C atmospheres [which] contain H20, CO2, CH4, and N2, but only traces of NH3, H2, and O2

Type A atmospheres are found in the giant planets of our outer solar system. Type B atmospheres occur in our inner solar system. Earth, Venus and Mars fall under this classification, but we don’t see any planets with Type C atmospheres.

Below is a series of charts showing the results for each of the six main molecular species over a range of mixtures.

Figure 2: The vertical axis is the ratio of Hydrogen to Oxygen, starting at 100% Hydrogen at the bottom and running to 100% Oxygen at the top. The horizontal axis shows the proportion of Carbon in the total mixture (The ratio runs up to 35%.) Molecular concentrations are in chemical equilibrium as a function of Hydrogen, Carbon, and Oxygen element abundances, calculated for T = 400 K and p = 1 bar. The blank regions are concentrations of < 10−4.

The central grey triangle marks the region in which H20, CH4, and CO2 can coexist in chemical equilibrium. The thin grey lines bisecting the triangle indicate where two of the constituents are at an equal concentration. These lines are hard to discern unless you can magnify the original image. For H20 and CO2 at equal concentration, it’s the dashed line (the near vertical line running upwards from 0.2 on the horizontal scale.) For CO2 and CH4, it’s the horizontal line. And for H20 and CH4, it’s the dotted line swooping upwards toward the top right-hand corner.)

The color bars at the right-hand side of the charts are both a color representation of the concentration and show the proportion of Nitrogen tied up as N2, i.e. that which is not NH3. Not surprisingly, the more Hydrogen there is in the mix, the higher the proportion of NH3 there is.

Other Results from the Paper

In the area around the stoichiometric ratio for water you get maximum H20 production and supersaturation occurs. Clouds form and the water rains out. Therefore, you cannot get an atmosphere with very high concentrations of water vapor unless the temperature is over 650°K, the critical point of water. Precipitation results in the atmospheric composition moving out of the area that gives CO2/CH4 mixtures.

Atmospheres with high carbon concentrations and having Hydrogen and Oxygen near their stoichiometric ratio have most of the atmospheric constituents tied up as water, so at a certain point carbon forms neither CO2 nor CH4 but rains out as soot. This, however, only precludes mixtures in the very right hand side of the CO2/CH4 Triangle.

Full-equilibrium condensation models show that the outgassing from warm rock, such as mid-oceanic ridge basalt can naturally produce Type C atmospheres.

Thoughts and Speculations

i) While it is difficult to argue with the man who coined the term, I still think Madhusudhan’s description of K2-18b as Hycean is too broad. Watching Madhusudhan in a Youtube interview, he refers to his paper “Habitability and Biosignatures of Hycean Worlds,’ which suggests that ocean covered planets under a Hydrogen atmosphere can exists within a zone that reaches into a level of irradiance slightly greater than Earth’s; however, he doesn’t mention the work by Lous et al in their paper, “Potential long-term habitable conditions on planets with primordial H–He atmospheres,” that showed that inside irradiance levels equivalent to 2 au from our Sun or greater, the Hydrogen atmosphere required to maintain Earthlike temperatures and not cook it is so thin that it is lost quickly over geological timescales. (You can see this in more detail in my article Super Earths/Hycean Worlds.) I would therefore define a Hycean planet as a rocky world with a radius up to 1.8 x Earth’s outside the irradiance equivalent of 2 au from our sun. K2-18b, being both larger than this and less dense than a rocky world, would fall, in my mind, firmly into the category of sub-Neptune.

ii) Another way of thinking of Type A, Type B and Type C atmospheres is to denote them as Hydrogen dominated, Oxygen dominated and Carbon dominated. Carbon dominated atmospheres may have by far the bulk of their constituents being Hydrogen and Oxygen; but because the enthalpy of the Hydrogen-Oxygen reaction is so much greater than the other reactions, when Hydrogen and Oxygen are close to their stoichiometric ratio, they preferentially remove themselves from the mix leaving Carbon as the dominant constituent. There is no Nitrogen dominated atmosphere because for most of its range Nitrogen sticks to itself forming N2 and is inert.

iii) The lack of H20 spectral lines is puzzling. Madhusudhan in his interview suggests that the spectra was a shot of the high-dry stratosphere. To cross-check the plausibility of this, I looked up the physical data on DMS. Dimethyl Sulfide vaporizes at 37°C and freezes at -98°C, which is lower than CO2’s freezing point. It also has a much higher vapor pressure than water at below freezing temperatures, so this does not contradict the assumption.

iv) I’m surprised this paper is not more widely known as not only does it provide a powerful tool for the analysis of exosolar planets’ atmospheric spectra, but it can also point to other aspects of a planet.

After the Hubble results came out in 2017, papers were published to model the formation of K2-18b, and while a range of possibilities could match the planet’s characteristics, they all came from the assumption that the planet began via the formation of a rocky/iron core followed by the gas accretion of large amounts of H2, Helium, and H20. According to the coexistence paper though, you cannot have large amounts of H2 and get a CO2/CH4 mix with no NH3. So to arrive at this state, this planet must never have had much gas accretion in the first place, or lost large amounts of Hydrogen after it formed. This latter scenario would require the planet to gain a Hydrogen envelope while at less than full mass in a hot nebula and then at full mass, in a cooler environment, lose most of its Hydrogen.

It is much easier to explain the planet’s characteristics by assuming it formed outside the snowline, never gained much of a gas envelope in the first place and spiraled into its present position. If it was formed from icy bodies like Ganymede and Titan (density ~ 1.9 gm/cc), this would give a good match for its density (2.7 gm/cc) allowing for gravitational contraction. The snow line is also the zone where carbonaceous chondrites form, so this would give the planet a higher carbon content than a pure rocky/iron one.

v) Madhusudhan, again from his interview, seems to think that K2-18b is an ocean planet, but I’m dubious about this for two reasons:

The first is that from the work done on Hycean planets by Lous et al, any depth of atmosphere especially with the potent greenhouse mix of CO2 and CH4 is likely to result in a runaway-greenhouse steam atmosphere inside the classically defined habitable zone (inside 2 au. for our sun).

The planet’s CO2/CH4 mix also points against this. From the paper, if there is a slight excess of Hydrogen over the stoichiometric ratio for water, then condensing H20 out, as either water or high pressure ice, pushes the planet’s atmosphere towards a Type A Hydrogen excess with no CO2 and NH3 lines appearing.

All of this would point towards a planet with a rocky/iron core overlaid by high pressure ice, which would, at about the megabar level, transition to a gas atmosphere composed mainly of super-critical steam. This would make up a significant volume of the planet. At the top of this atmosphere, the water, now in the form of steam, would condense out as virago rain leaving a dry stratosphere consisting mainly of CO2, CH4, H2 and N2.

To test my assumption, I did a rough back of the envelope calculation using online calculators, and looked at the wet adiabatic lapse rate (the rate of increase in temperature when saturated air is compressed) per atm. pressure doubling starting from 1 bar at 20°C. This rate (1.5°C/1000 ft) is considerably less than the rate for dry gases (3°C/1000 ft).

It was all very ad hoc, but the first thing I noted was that for each pressure doubling, the boiling point of water goes up significantly–at 100 bar, water boils at 300°C–until its temperature approaches its critical point (374°C) where it levels off. So the lapse rate increase in temperature chases the boiling point of water as you go deeper and deeper into the atmosphere; however, from my calculations, it catches water’s boiling point at 270°C and 64 bar. The calculations are arbitrary—I was using Earth’s atmospheric composition and gravity–and small changes in the parameters can result in big changes in the crossover point; but what this does point to is that if the planet has an ocean, it could be a rather hot one under a dense atmosphere, and if the atmosphere has any great depth then the ocean is likely to be a supercritical fluid.

Also, for the atmosphere to be thin, the planet’s ratio of CO2, CH4 and H2 must be less than 1/10,000 that of H20, which is not something I regard as likely, given what we know about the outer solar system.

I’ll leave you with a phase diagram of water with (red line) the dry adiabat of Venus moved 25°C cooler to represent a dry Earth and the wet adiabat (blue line) the one I calculated out. It’s also a handy diagram to play with as it gives you an idea of how deep the ocean or critical fluid layer will be at a given temperature before it turns into a layer of high pressure ice.

vi) One final point, and this reinforces the purpose of the paper: that we need to thoroughly understand planetary chemistry to eliminate false bio-markers. DMS is widely touted as a biomarker, but if we look at the most thermodynamically stable forms of sulfur: In a Type A reducing atmosphere, it’s H2S; and in a wet, oxidizing, Type B atmosphere, it’s the Sulfate (SO42-) ion. Unfortunately, the authors of the paper did not extend their thermodynamic analysis to Sulfur, but if we look at DMS’s formula (CH3)2S, it looks an awful lot like a good candidate for the most thermodynamically stable form of Sulfur for a Type C atmosphere, not a biomarker.


Wikipedia: K2-18b

N. Madhusudhan, S. Sarkar, S. Constantinou, M Holmberg, A. Piette, and J. Moses, Carbon-bearing Molecules in a Possible Hycean Atmosphere, Preprint, arXiv: 2309.05566v2, Oct 2023

P. Woitke, O. Herbort, Ch. Helling, E. Stüeken, M. Dominik, P. Barth and D. Samra, Coexistence of CH4, CO2, and H2O in exoplanet atmospheres, Astronomy & Astrophysics, Vol. 646, A43, Feb 2021

N. Madhusudhan, M. Nixon, L. Welbanks, A. Piette and R. Booth, The Interior and Atmosphere of the Habitable-zone Exoplanet K2-18b, The Astrophysical Journal Letters, 891:L7 (6pp), 2020 March 1

Super Earths/Hycean Worlds, Centauri Dreams 11 November, 2022

Youtube interview of Nikku Madhusudhan, Is K2-18b a Hycean Exoworld? on Colin Michael Godier’s Event Horizon