Thermometric Soots on Hot Jupiters?

K. Zahnle
NASA Ames Research Center, Moffett Field, CA 94035
arXiv:0911.0728v1 [astro-ph.EP] 4 Nov 2009

Kevin.J.Zahnle@NASA.gov

M. S. Marley
NASA Ames Research Center, Moffett Field, CA 94035

Mark.S.Marley@NASA.gov

and

J. J. Fortney
Department of Astronomy and Astrophysics, University of California - Santa Cruz

Received ; accepted

submitted to Ap. J. Lett.

 –2–

ABSTRACT

We use a 1D thermochemical and photochemical kinetics model to predict

that the stratospheric chemistry of hot Jupiters should change dramatically as
temperature drops from 1200 to 1000 K. At 1200 K methane is too unstable to
reach the stratosphere in significant quantities, while thermal decomposition of
water is a strong source of OH radicals that oxidize any hydrocarbons that do
form to CO and CO2 . At 1000 K methane, although very reactive, survives long
enough to reach the lower stratosphere, and the greater stability of water coupled
with efficient scavenging of OH by H2 raise the effective C/O ratio in the reacting
gases above unity. Reduced products such as ethylene, acetylene, and hydrogen
cyanide become abundant; such conditions favor polymerization and possible
formation of PAHs and soots. Although low temperature is the most important
factor favoring hydrocarbons in hot Jupiters, higher rates of vertical mixing and
generally lower metallicities also favor organic synthesis. The peculiar properties
of HD 189733b compared to other hot Jupiters—a broadband blue haze, no sign
of Na or K, and hints that CO might be underabundant—can be explained by
an organic haze if the planet is cool enough. Whether this interpretation applies
to HD 189733b itself, many organic-rich warm Jupiters are sure to be discovered
in the near future.

Subject headings: planetary systems — stars: individual(HD 189733)

 –3–

1. Introduction

The best studied transiting hot Jupiters are HD 209458b and HD189733b (Fortney et
al 2009). HD 209458b is a typical hot Jupiter with a typical six micron excess (Burrows
et al. 2008). Transit observations show a visible spectrum dominated by the 590 nm
sodium resonance lines (Charbonneau et al 2002), the visibility of which implies a clean,
haze-free atmosphere (Sing et al 2008). HD 189733b is cooler and less typical of known hot
Jupiters (it lacks a six micron excess), although its details are controversial (Fortney et al
2009). Swain et al (2008, 2009) fit transit data between 1.5 and 2.5 µm with CO, CO2 ,
H2 O, and CH4 , all with distinctly subsolar abundances corresponding to a metallicity of
[M/H] ≈ −0.5. They suggest that adding C2 H2 , C2 H4 , or possibly NH3 would improve their
model’s fit. Fortney et al (2009) fit Désert et al’s (2009) dayside spectrum at 4.5 µm with
supersolar CO and CO2 corresponding to 0.5 < [M/H] < 1.5. In visible light HD 189733b
presents a featureless blue haze with no sign of the expected Na and K lines (Pont et al
2008). Lecavalier des Etangs et al (2008) suggest that these lines are obscured by enstatite
grains mixed up to microbar levels.

Here we use the photochemical model that we used to address S chemistry (Zahnle et al
2009) to address the hypothesis that HD 189733b is cool enough to generate a hydrocarbon
haze. In our earlier paper we noted that temperatures below 1200 K bring a fundamental
shift in the kinds of chemistry that take place. The cooler atmospheres are much more
reduced: organic molecules are favored, PAHs (polycyclic aromatic hydrocarbons) and
soots (disorganized agglomerations of bigger PAHs) might form and precipitate, and CO’s
abundance is reduced while the CO2 /CO ratio is increased. The PAHs and soots would be
the source of the haze.
 –4–

2. Previous Models

The possibility that organic hazes might be important in brown dwarfs was first raised
by Griffith et al (1998). Previous photochemical modeling of hot Jupiters addressed the
abundance of photochemical H (Liang et al 2003) and the absence of photochemical smogs
(Liang et al 2004). The latter showed that simple hydrocarbons would not condense to
form photochemical smogs in hot solar composition atmospheres.

Zahnle et al (2009) focused on sulfur chemistry between 1200 and 2000 K. Sulfur
is central, as H2 S is the most reactive molecule in these atmospheres, and its thermal
decomposition is the biggest source of free radicals and H in particular. Our chief results
were that (i) HS and S2 should be abundant at the millibar level, and (ii) CO2 is a good
indicator of metallicity in this temperature range. We found that most incident light from
250 nm to 450 nm is absorbed by stratospheric S2 and HS. The resultant stratospheric
heating can be considerable in planets that are not hot enough for TiO and VO to evaporate
in significant quantities. CO2 ’s sensitivity to metallicity had been pointed out before
(Fegley and Lodders 2002, Visscher et al 2006). We showed that this result applies when
kinetics are included and that CO2 is insensitive to T in the range 1200 < T < 2000 K.

3. Model

We use a standard 1D kinetics code to simulate atmospheric chemistry of hot giant

planets. Volume mixing ratios fi of species i are obtained by solving

∂fi ∂φi
N = Pi − Li Nfi − (1)
∂t ∂z

and
m g m g  ∂fi
a i
φi = bia fi − − (bia + Kzz N) (2)
kT kT ∂z
 –5–

where N is the total number density (cm−3 ); Pi − Li Nfi represent chemical production and
loss terms, respectively; φi is the upward flux; bia , the binary diffusion coefficient between
i and the background atmosphere a, describes true molecular diffusion; ma and mi are
the molecular masses of a and i; and Kzz —the eddy diffusion coefficient—parameterizes
vertical mixing as diffusion. In practice Eqns 1 and 2 are solved as a system of second order
partial differential equations for fi (z, t). Steady state solutions are found by integrating
the equations through time (typically 109 years) using a fully implicit backward-difference
method. The code is briefly described in Zahnle et al (2009) and more completely described
as applied to ancient Earth (Zahnle et al 2007) and Mars (Zahnle et al 2008).

A difference between this study and our previous study is that we include
molecular diffusion. We have implemented molecular diffusion through H2 by setting
bia = 6 × 1019 (T /1400)0.75 cm−1 s−1 — appropriate for CO — for all the heavy species.

We solve 542 chemical reactions for 51 chemical species. New species added since
Zahnle et al (2009) are H2 COH and CH3 OH (methanol); these take part in pathways that
can hydrogenate CO to CH4 . In practice these pathways were not found to be important
under any conditions encountered in this study. The hydrocarbon chemistry is truncated at
C2 Hm . Thus when conditions favor polymerization carbon pools in C2 Hm , because longer
carbon chains are not allowed.

Another limitation is that we have excluded reactions such as C2 H4 + OH + M →

adduct + M. The precise nature of the adduct is generally unknown. We have instead
included only the competing abstraction path C2 H4 + OH → C2 H3 + H2 O. This may be an
acceptable approximation at high temperatures and low pressures where the abstraction
path is favored over the (3-body) addition. But by excluding reactions in which OH
directly attacks a carbon atom in molecules like C2 H4 , we underestimate how well oxidation
competes with polymerization. Consequently our models overestimate the stability of
 –6–

organic molecules.

We assume isothermal atmospheres with constant Kzz . This choice facilitates presenting
the results of parameter studies, as T and Kzz are the most important parameters. The
background atmosphere is assumed 84% H2 and 16% He. C, O, N, and S are assumed
present in solar proportions and scaled as a group according to metallicity. Metallicity
is varied from −0.7 ≤ [M/H] ≤ 1.7. Surface gravity is g = 20 m/s2 . Surface gravity is
important to what a planet looks like, but it has little effect on the chemistry (Zahnle et al
2009). Incident UV radiation is 100× greater than at Earth and the solar zenith angle is
θ = 30◦ unless otherwise noted. In one case we consider θ = 80◦ and θ = 89◦ , which might
be pertinent to planets observed in transit.

At the upper boundary we set a zero flux lid at ∼ 1µbar. For the lower boundary we
assume thermochemical equilibrium mixing ratios of H, H2 O, CO, CH4 , NH3 , N2 , and H2 S
(Lodders and Fegley 2002, Visscher et al 2006). Other species are assumed to vanish at the
lower boundary. This ensures that photochemical products flow into the deep atmosphere,
where they are recycled.

The lower boundary pressure was immaterial to our previous study in which
temperatures were very high and the relevant chemical reactions very fast. But at 1000
K and below, the carbon and nitrogen chemistries become sensitive to the choice of lower
boundary conditions. The reason for the sensitivity is that, even at 100 bars, gas phase
chemistry provides no effective means of converting CO to CH4 and N2 to NH3 , and is only
beginning to convert excess H to H2 . In practice it is likely that conversion of CO to CH4
and N2 to NH3 take place on grains, in particular Fe-Ni grains if present. Here we consider
lower boundary pressures of 1 and 100 bars.
 –7–

4. Results

Figure 1 illustrates compositions of representative 1200 K and 1000 K atmospheres

for modest Kzz = 106 cm2 /s at solar and planetary metallicities ([M/H] = 0 and 0.7,
respectively). Sulfur chemistry is included but not shown. The overall structure of the
atmosphere resembles a poorly functioning gas stove, with methane introduced at the
bottom and flowing upward, where it is either oxidized or made into bigger organic
molecules. At 1200 K, methane is quickly and cleanly oxidized to CO, but at 1000 K
methane combustion generates ethylene (C2 H4 ) and HCN, to the point where ethylene
becomes more abundant than methane. This does not mean that C2 H4 actually would be
more abundant than CH4 (our model stops at C2 Hn ), but it does mean that polymerization
is a major methane reaction pathway, and it suggests that hydrocarbons bigger than C2 H4
are likely to form.

The factors behind the temperature dependence seen in Figure 1 are the different
thermal stabilities of methane and water and the high abundances of H and H2 . Both H2 O
and CH4 are attacked by H:
H + CH4 → H2 + CH3 R1

H + H2 O → H2 + OH. R2

R2 is significantly endothermic, while R1 is neutral. Both R1 and R2 also have significant

activation energies, but because R2 is endothermic water becomes unreactive at a higher
temperature than methane, while the reverse of R2, the reaction of OH with H2 , remains
quite fast at 1000 K. At 1200 K water is reactive enough that OH is abundant and the
gas is generally oxidizing. But as the temperature drops to 1000 K, water becomes more
stable and OH becomes scarce, while CH4 becomes stable enough to be abundant. As a
consequence the effective C/O ratio in the reacting gases is high and the products are
hydrocarbons.
 –8–

Mismatches between the kinetics model and thermochemical equilibrium at the lower
boundary cause flows into or out of the lower atmosphere. The most important mismatches
stem from kinetic inhibition of the 3-body reactions that hydrogenate CO to CH4 and N2
to NH3 , and the 3-body reactions that reconstitute H2 from H. The latter results in a high
disequilibrium abundance of H. Very high abundances of H are a prominent feature of all
the atmospheres in Figures 1 and 2. The overabundance of H pushes R1 farther to the right
than it would be in thermochemical equilibrium. Analogous downstream reactions, such as
CH3 + H → H2 + CH2 , are also pushed to the right, with the result that reactions between
radicals to form secondary products, such as CH3 + CH2 → C2 H4 + H, are much more
important than they would be in equilibrium.

Figure 1 also shows the influence of the lower boundary. Placing the lower boundary
at 1 bar puts more methane and less ammonia at 1 bar than when the boundary is at 100
bars. At 1200 K and below, ammonia rises faster than it is lost to N2 . The equilibrium
abundance of NH3 is higher at high pressure, hence the deeper lower boundary puts more
NH3 in the stratosphere. The artificial methane at 1 bar creates a sharp gradient and a big
upward flux that results in more ethylene. As both the 1 bar and 100 bar lower boundaries
are arbitrary in this study it is not useful to take this any further.

Figure 2 illustrates the influence of different rates of vertical transport in 1000 K

atmospheres. The overall trend is that faster vertical transport lifts CH4 to higher altitudes,
and thus raises the altitudes where C2 H2 , C2 H4 , and HCN form. Faster transport increases
the efficiency of hydrocarbon formation vis a vis CO formation, as also seen in the CO
profiles, and it changes the mix of products to favor C2 H2 and HCN over C2 H4 .

The influence of Kzz on carbon chemistry is viewed from a different angle in Figure 3.
Here we show how carbon is allocated at the altitude that is most favorable to hydrocarbon
production. We divide carbon into 3 categories. The primary category is CH4 , the source
 –9–

molecule. The other two are oxidized species containing C-O bonds (CO, CO2 ) and the
more reduced products that do not contain C-O bonds (C2 H2 , C2 H4 , HCN, CS2 , etc).
Call these xCH4 , xCO , and xC2 , respectively. Total carbon is ΣC = xCH4 + xCO + xC2 . We
define the most favorable altitude pf for organic synthesis to be where xC2 /ΣC reaches its
maximum value. Whilst every C-O bond is doomed to end in the irreversible production of
CO, it is possible that much of the C in xC2 can end in larger molecules such as PAHs or in
the irreversible production of refractory soots.

Figures 4 and 5 generalize these results to different atmospheres. Figure 4 expresses

model results as pie charts on a T -Kzz grid. The relative amounts of carbon in different
species is presented at the altitude pf . It is an optimistic figure, in the sense that it
shows the most favorable conditions for carbon polymerization in each model. The abrupt
chemical transition at ∼1100 K and the more modest influence of Kzz are readily seen.

Figure 5 addresses metallicity and some other parameters. The pies show a clear
negative dependence on metallicity. This is caused by H2 , which removes OH. Of course H2
is abundant in any hot Jupiter, but it is by definition even more abundant when metallicity
is low. Another way to view this is that “metals” are on balance oxidizing because the C/O
ratio is less than one. Adding “metals” creates more oxidized conditions, other things equal.

Figure 5 also addresses the influence of the lower boundary and UV insolation, albeit
sparsely. Raising the lower boundary to 1 bar generates more hydrocarbons because it
creates a bigger flux of CH4 into the atmosphere. UV photolysis inhibits hydrocarbon
formation. This occurs because C2 H2 and C2 H4 are subject to photolysis, and H2 O
photolysis is a source of OH. Evidently soot’s precursors grow better in the dark.

Ethylene, acetylene, and hydrogen cyanide are first generation products of methane.
Benzene is at best a third generation product, and PAHs like coronene (C24 H12 ) and ovalene
(C32 H14 ) are probably two or three generations further evolved. Each step of the way some
 – 10 –

carbon is lost to CO. If it takes n generations to make something resembling a soot, we

might expect the efficiency of soot formation to go like xnC2 . This is greater than 10% for
xC2 > 0.8 and n = 10, so it seems that atmospheres where xC2 > 0.8 have potential to
develop significant organic hazes.

PAHs interact strongly with visible light (Lou Allamandola, pers. comm.). Coronene is
golden yellow, ovalene is red, and still bigger ones tend to black. Thus PAHs by themselves
may suffice to explain high altitude haziness in hot jupiters; it may not be necessary that
they agglomerate into grains, albeit grains are likely to be more refractory.

5. Discussion

The chemistry that we have modeled is unremarkable. Most of the important reactions
involving small organic molecules and radicals have been studied both experimentally and
theoretically. Although uncertainties are considerable, none of the important reactions,
apart from some involving sulfur, have had to be invented. Thus the prediction that
methane can react to make ethylene rather than CO seems secure. The presumption that
this will lead to bigger hydrocarbons than ethylene seems equally secure. Whether this
process continues up to PAHs and soot is less secure.

Our results appear to disagree with what Liang et al (2004) found previously.
However, there are enough differences between their simulations and ours that the apparent
disagreement is misleading. The higher hydrocarbons that form in our models are primarily
products of disequilibrium thermochemistry, not photochemistry. They form because
recombination of H to H2 is kinetically inhibited and consequently the atomic hydrogen
density is much higher than it would be in equilibrium. Liang et al have less H than we
do because they did not include sulfur; we found (Zahnle et al 2009) that the reaction
 – 11 –

H2 S + H ↔ HS + H2 is generally the most frequent reaction in the atmosphere by a wide

margin and is the leading source and sink of H.

Liang et al (2003, 2004) attempt to use realistic approximations to T (z) and Kzz (z),
we have not. This is a feature of our study — we emphasized the survey of parameter
space, once we identified T and Kzz as the two key parameters — but we are aware that
the isothermal assumption risks missing important physics. For Kzz , Liang et al use
Kzz ∝ N −0.6 . Their assumption gives Kzz = 2.4 × 107 cm2 /s at 1 bar, 1.5 × 109 cm2 /s at
1 mbar, and 1.0 × 1011 cm2 /s at 1 µbar (all at 1200 K). This range is similar to our high
Kzz cases. For T (z), Liang et al use three specific profiles derived from detailed radiative
transfer model results. All three are very hot at 1 bar (the coolest is 1800 K). These
temperatures are much higher than anything we address here. In our model methane is
immediately destroyed if T > 1200 K. We agree with Liang et al that very hot planets
won’t have hydrocarbon hazes.

As a test we computed chemistry using the nominal p-T profile from Fortney et al’s
(2009) most recent model of HD 189733b. Their model is hot (T ≈ 1500 K) below 3
bars and cool (T ≈ 900 K) above 10 mbars, with a smooth transition in between. We
set Kzz = 108. We find very little methane above the lower boundary and no significant
polymerization. We then reduced the lower temperature to 1200 K, and then to 1100 K.
At 1200 K, NH3 is abundant but hydrocarbons are not, while at 1100 K we see rampant
hydrocarbon polymerization. These results are quite consistent with Figure 4.

In passing, we note that solar C/O is an assumption. Snowline models of giant planet
formation can give lower C/O, while the less well known tar-line model (Lodders 2004)
would predict higher C/O. High C/O can change the chemistry dramatically; among other
things, soots would be favored at higher temperatures. The possibility that HD 189733b
has high C/O should be in play.
 – 12 –

6. Conclusions

We show that vertical mixing in hot Jupiters with temperatures below 1100 K can
provide environments suitable for turning methane into ethylene, acetylene, and hydrogen
cyanide. We speculate that such atmospheres would be conducive to generating PAHs and
perhaps soots. The presence of abundant PAHs or soots at high altitudes could account for
the up-to-now unique properties of HD 189733b, if HD 189733b proves to be cool enough.
This would make HD 189733b the first of a new class of warm Jupiters. Whether our
thermometric interpretation of HD 189733b actually applies to HD 189733b (recent models
suggest that the planet is too hot if it has solar C/O) is not the main point. The main point
is that many organic-rich hazy warm Jupiters are sure to be discovered in the near future.

7. Acknowledgements

KJZ thanks NASA’s Exobiology Program for support.

 – 13 –

REFERENCES

Burrows, A., Budaj, J., & Hubeny, I. 2008, ApJ, 678, 1436-1457 “Theoretical Spectra and Light
Curves of Close-in Extrasolar Giant Planets and Comparison with Data”

Charbonneau D, Brown TM, Noyes RW, & Gilliland RL 2002. ApJ, 568, 377-384.

Désert J.-M., Lecavelier des Etangs A., Hébrard G., Sing DK, Ehrenreich D, Ferlet R, &
Vidal-Madjar A. 2009, Astrophys. J. 699, 478-485.

Fortney JJ, Shabram M, Showman AP, Lian Y, Freedman RS, Marley MS, and Lewis NK (2009).
Transmission Spectra of Three-dimensional Hot Jupiter Model Atmospheres. Astrophys. J.

Griffith CA, Yelle RV, Marley MS 1998. “The dusty atmosphere of the brown dwarf Gliese 229B”
Science 282, 2063-2067.

Lecavelier des Etangs A, Pont F, Vidal-Madjar A, & Sing D. (2008) “Rayleigh scattering in the
transit spectrum of HD 189733b” Astron. Astrophys. 481, L83-L86.

Liang M-C, Parkinson CD, Lee AYT, Yung YL, and Seager S, (2003) “Source of atomic hydrogen
in the atmosphere of HD 209458b” Astrophys. J. 596, L247L250.

Liang M-C, Seager S, Parkinson CD, Lee AYT, and Yung YL (2004) “On the insignificance of
photochemical hydrocarbon aerosols in the atmospheres of close-in extrasolar giant planets”
Astrophys. J. 605, L61L64.

Lodders K. 2002. “Atmospheric Chemistry in Giant Planets, Brown Dwarfs, and Low-Mass Dwarf
Stars. I. Carbon, Nitrogen, and Oxygen” Icarus 155, 393-424.

Lodders K. 2004. “Jupiter Formed with More Tar than Ice” Astrophys. J. 611, 587-597.

Pont F, Knutson H, Gilliland RL, Moutou C, & Charbonneau D. 2008. “Detection of atmospheric
haze on an extrasolar planet: the 0.55-1.05 µm transmission spectrum of HD 189733b with
the Hubble Space Telescope” Mon. Not. Roy. Astron. Soc. 385, 109-118.
 – 14 –

Showman, A. P., Fortney, J. J., Lian, Y., Marley, M. S., Freedman, R. S., Knutson, H. A., &
Charbonneau, D. 2009 Astrophys. J. 699, 564-584.

Sing, D.K., Vidal-Madjar A., Désert, J.-M., Lecavelier des Etangs, A., and Ballester, G. (2008).
Astrophys. J. 686, 658-666.

Swain, M. R., Vasisht, G., & Tinetti, G. 2008. Nature 452, 329-331.

Swain, M. R., Vasisht, G., Tinetti, G., Bouwman, J., Chen, P., Yung, Y., Deming, D., & Deroo,
P. 2009, Astrophys. J. Lett. 690, L114-L117.

Visscher C, Lodders K, and Fegley B. (2006). Astrophys. J. 648, 1181-1195.

Zahnle KJ, Claire MW, Catling DC (2006). Geobiology 4, 271-282.

Zahnle KJ, Haberle RM, Catling DC, Kasting JF (2008). J. Geophys. Res. 113, E11004,
doi:10.1029/2008JE003160.

Zahnle KJ, Marley MS, Freedman RS, Lodders K, and Fortney JJ (2009). Astrophys. J. 701,
L20-L24.

This manuscript was prepared with the AAS LATEX macros v5.2.
 – 15 –

-5
10

-4
10

N OH H
-3 H N
10
Pressure (bars)

OH 2 2
CO CO
2 2
-2
10

-1 NH
10 CO 3 CO
NH
3
0
10 HCN
H O H O
2 HCN 2
1
10 CH
C H 4 T=1200 K C H CH T=1200 K
2 4 2 4 4
[M/H]=0 [M/H]=0.7
2
HCN
10
-5
10

-4
10 CO N H CO N H
2 2 2 2
CO
-3
10
Pressure (bars)

NH NH
3 3
CO
-2
10

-1
10

0
10 C H
2 2 H O H O
HCN 2 C H 2
1
2 2 HCN
10 C H T=1000 K
C H T=1000 K 2 4
2 4 CH [M/H]=0
CH [M/H]=0.7
4 4
2
10
-9 -8 -7 -6 -5 -4 -3 -2 -1 -8 -7 -6 -5 -4 -3 -2 -1
10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10

Volume Mixing Ratio Volume Mixing Ratio

Fig. 1.— Vertical profiles of selected species in 1200 K and 1000 K isothermal atmospheres
at solar ([M/H] = 0) and planetary ([M/H] = 0.7) metallicities. Solid curves assume ther-
mochemical equilibrium boundary conditions at 100 bars; open symbols, the same at 1 bar.
Fixed model parameters are Kzz = 106 cm2 /s, I = 100, and g = 20 m/s2 . At 1200 K
methane is quickly oxidized to CO and CO2 . At 1000 K, there is significant disequilibrium
production of organic molecules, especially of ethylene. The overall picture strongly suggests
that bigger hydrocarbons that lead to soots or PAHs should form.
 – 16 –

-6
10
5 7
-5 K =10 K =10
10 zz zz

CO N CO
0.0001 OH 2 2
Pressure (bars)

H H
OH CO N
0.001 2 2

0.01 HCN
CO C H
0.1 NH 2 6
3
C H
2 2
C H HCN
1 2 2

C H
10 C H 2 4
2 6 C H CH H O H O
2 4 4 2 NH CH 2
3 4
100-6
10
11
9 OH CO HCN K =10
-5 K =10 OH CO N CO 2 zz
10 zz 2 2

H
0.0001
Pressure (bars)

H
0.001

0.01

0.1 C H HCN N
2 6 C H 2
2 6 C H
2 2
1 C H
2 2

10
C H NH CH H O CO C H NH CH H O
2 4 3 4 2 2 4 3 4 2

100
-9 -8 -7 -6 -5 -4 -3 -2 -1 -8 -7 -6 -5 -4 -3 -2 -1
10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10
Volume Mixing Ratio Volume Mixing Ratio

Fig. 2.— Vertical profiles of selected species in isothermal 1000 K atmospheres and [M/H] =
0.7 as a function of vertical mixing. Strong vertical mixing (high Kzz ) favors disequilibrium
production of C2 H4 , HCN, and especially C2 H2 . Acetylene (C2 H2 ) — the canonical soot
precursor — becomes abundant at 1 mbar with Kzz > 107 cm2 /s. In general high altitudes
are oxidizing (high OH) from H2 O photolysis. The horizontal gray bars denote the most
favorable altitudes for hydrocarbon formation pf as defined in the text.
 – 17 –

1 0.0001
Fractions of oxidized, reduced,
f

Most favorable altitude for

and unreacted carbon at p

reduced carbon, p [bars]

CO + CO ! C2Hn, HCN etc
2
0.8 0.001

0.6 p 0.01
f

0.4 0.1

f
0.2 1

CH
4
0 10
5 6 7 8 9 10 11
10 10 10 10 10 10 10
2
Vertical Mixing K [cm /s]
zz

Fig. 3.— Methane’s fates at the most favorable altitude pf as a function of vertical mixing
Kzz . Fixed model parameters are T = 1000 K, [M/H] = 0.7, I = 100, p0 = 100 bars, and
g = 20 m/s2 . Irreversible oxidization (xCO ) is denoted “CO and CO2 ”, interesting products
(xC2 ) are denoted “ΣC2 Hn , HCN etc”. High Kzz favors organic molecules. At 1000 K, soot
formation seems reasonably likely for Kzz > 108 cm2 /s. The right hand axis shows that pf is
a monotonically increasing function of Kzz .
 – 18 –

Fig. 4.— Generalization of Figure 3 to other temperatures. Each pie-chart shows how carbon
is allocated at pf . The pies are placed on a T -Kzz grid. To first approximation higher values
of Kzz can also be read as higher altitudes. Temperature is the most important factor; high
Kzz favors organic synthesis. Very high values of Kzz ≥ 1010 cm2 /s mix CH4 to altitudes
that are more favorable to C2 H2 , the traditional soot precursor. Fixed model parameters
are [M/H] = 0.7, I = 100, p0 = 100 bars, g = 20 m/s2 , and θ = 30◦ incidence angle for UV
radiation.
 – 19 –

Fig. 5.— Generalization of Figure 3 to other metallicities. The pie-charts have the same
meanings as in Figure 4. Fixed model parameters (unlabeled pies) are T = 1000 K, I = 100,
p0 = 100 bars, g = 20 m/s2 , and θ = 30◦ . Note the inverse correlation between hydrocarbons
and metallicity; note also that low metallicity favors C2 H2 and HCN over C2 H4 . Alternative
pie a sets I = 1 to simulate feeble stellar UV. Alternatives b and c set θ = 80◦ and θ = 89◦
to simulate the more grazing incidence angles that may be pertinent to transit observations.
The d pies illustrate the modest consequences of raising the lower boundary of the model
from 100 bars to 1 bar.