Vol 440|2 March 2006|doi:10.

1038/nature04497

LETTERS
Episodic outgassing as the origin of atmospheric
methane on Titan
Gabriel Tobie1,2, Jonathan I. Lunine2,3 & Christophe Sotin1

Saturn’s largest satellite, Titan, has a massive nitrogen atmosphere
containing up to 5 per cent methane near its surface. Photo-
chemistry in the stratosphere would remove the present-day
atmospheric methane in a few tens of millions of years1. Before
the Cassini-Huygens mission arrived at Saturn, widespread liquid
methane or mixed hydrocarbon seas hundreds of metres in
thickness were proposed as reservoirs from which methane
could be resupplied to the atmosphere over geologic time2. Titan
fly-by observations3–5 and ground-based observations6 rule out the
presence of extensive bodies of liquid hydrocarbons at present,
which means that methane must be derived from another source
over Titan’s history. Here we show that episodic outgassing of
methane stored as clathrate hydrates within an icy shell above an
ammonia-enriched water ocean is the most likely explanation for
Titan’s atmospheric methane. The other possible explanations all
fail because they cannot explain the absence of surface liquid
reservoirs and/or the low dissipative state of the interior. On the
basis of our models, we predict that future fly-bys should reveal
the existence of both a subsurface water ocean and a rocky core,
and should detect more cryovolcanic edifices.
The presence of methane within the atmosphere and the interior
is primarily determined by the nature of the materials from which
Titan was built, and the bulk density of Titan indicates that the
body is roughly 0.5–0.7 by mass silicate (for a silicate density of
3,000–4,000 kg m23) and the remainder is mostly water ice7,8. Even
though external sources of methane derived from cometary impact
may have been significant in a far past, they cannot explain its current
abundance9. Furthermore, direct sampling of Titan’s atmosphere by
the mass spectrometers on board the Orbiter10 and Probe11 provide
circumstantial evidence that ammonia (NH3) is the primordial
source of Titan’s atmospheric molecular nitrogen9. For ammonia
to be the source of the nitrogen, the hydrogen must have been
removed by atmospheric photochemistry, charged particle chemistry
or impacts, so that part of the ammonia brought by the accreted
materials should have been devolatilized during or soon after Titan’s
formation. This is consistent with previous formation models12–14,
and a consistent post-accretional internal structure comprises an
inner homogeneous core made of a mixture of rock and icy materials,
overlain by a silicate layer and a water layer enriched with liquid
ammonia12 (see Fig. 2c).
At the end of accretion, only the inner undifferentiated portion of
Titan’s interior was able to hold methane. As most of the region
outward of this proto-core was warm liquid water, in which methane
has very low solubility, potentially very large amounts of methane
ended up in the primitive atmosphere and on the surface. Methane in
the inner core could have been incorporated there during accretion,
reflecting its presence in the planetesimal building blocks of Titan, or
it might have come from the conversion of CO2 through ‘serpenti-
nization’ in the silicate-rich core in a way similar to the process
1
Laboratoire de Planétologie et Géodynamique, UMR-CNRS 6112, Université de Nantes, 44322 Nantes cedex 03, France. 2Lunar and Planetary Laboratory, University of Arizona,
Tucson, Arizona 85721, USA. 3Instituto di Fisica dello Spazio Interplanetario, INAF, Via del Fosso del Cavaliere 100, 00133 Roma, Italia.
involved in the Earth’s hydrothermal systems15. Whatever its origin,
laboratory experiments16,17 show that methane under the temperature–
pressure conditions of Titan’s cold, undifferentiated proto-core
would rapidly combine with water molecules to form methane
clathrate hydrate17. Owing to its low density (see Supplementary
Information 1.1), any methane clathrate released during the overturn
that formed a discrete rock core12 would ascend to the top of the
outer liquid layer without being dissociated because of its high
stability18 (see Supplementary Information 1.2).
Subsequent to core overturn, methane outgassing would have
occurred only when the conditions required to dissociate methane
clathrate were reached within the icy mantle of Titan’s interior. We
use a coupled thermal-orbital model19 to determine the conditions
and timing of clathrate dissociation. Although some of the latest
interior models18,20 indicate that methane clathrate dissociation
within Titan’s interior is very difficult, it must be noted that those
models included neither the coupling between Titan’s interior and its
orbit nor the low thermal conductivity17 and high viscosity21 of
methane clathrate as compared with water ice. Here we show that the
incorporation of those effects strongly influences the cooling rate of
Titan’s interior and leads to episodic outgassings of methane. Our
Figure 1 | Global heat power provided by the rocky core. Three different
mineral compositions, reflecting different hydration states, have been
used to estimate the possible evolution of the silicate core over a reasonable
range: 1, r sil ¼ 3,000 kg m23, H 0 ¼ 3 £ 10211 W kg21, T i ¼ 1,400 K;
2, r sil ¼ 3,300 kg m23, H 0 ¼ 4 £ 10211 W kg21, T i ¼ 1,500 K;
3, r sil ¼ 4,000 kg m23, H 0 ¼ 5 £ 10211 W kg21, T i ¼ 1,600 K, where r sil is
the silicate density, H 0 is the initial radiogenic heating rate and T i is the
interior temperature required to trigger thermal convection in the silicate
core. Evolution of the silicate core is characterized by a diffusive stage
following the differentiation and a convective stage once the interior is
sufficiently hot. Depending on the composition of the silicate core, we
estimate an onset time of convection of between 1.8 and 2.8 Gyr (see
Supplementary Information 2.1).

61
© 2006 Nature Publishing Group
LETTERS
model includes heat transfer through the outer layer, clathrate
dissociation and crystallization of the liquid layer, thermal evolution
of the silicate core, and tidal dissipation and its effect on orbital
eccentricity decay (see Supplementary Information 2).
Titan shows an orbit with an eccentricity of 3%, which cannot be
explained by orbital resonance. Titan’s current free eccentricity is
probably the fossil of a higher primordial eccentricity19. We therefore
consider a wide range of initial eccentricity values (5–30%), and we
retain only models that reproduce the current value. We also explore
different interior compositions with varying rheological assumptions
and different concentrations of radiogenic elements. Figure 1 shows a
reasonable range of global heat power coming out of the silicate core,
assuming silicate minerals with different compositions.
A typical example of interior evolution and methane outgassing is
shown in Fig. 2. Three major episodes of outgassing are predicted and
result from, first, internal differentiation (silicate core formation);
second, onset of convection in the silicate core; and third, late
subsequent cooling and crystallization of the outer layer. The first
episode follows the overturn of the undifferentiated proto-core with
the silicate carapace and the simultaneous formation of the methane
clathrate layer above the ocean. The clathrate accumulation near the
surface results in an increase in the ocean temperature, owing to the
low conductivity and the high viscosity of the methane clathrate, and
hence release of methane. The second episode occurs when the
Figure 2 | The evolution of the interior and the outgassing rate of methane
over Titan’s history. In the simulation shown, we use composition 2 in
Fig. 1 for evolution of the silicate core and we assume an ammonia fraction
of 5% and a value of 1014 Pa s for the viscosity of ice close to its melting point.
Furthermore, the initial eccentricity is set to 13.5% to fit the current value.
We start the evolution at time, t ¼ 0.5 Gyr once the formation of the silicate
core is achieved12 and once the energy contribution from meteoritic
bombardment can be safely neglected. The methane release rate, normalized
by the current loss rate (1.25 £ 1010 cm22 s21; a), is controlled by the
interior evolution (c) and especially the outer region evolution (b) and is
characterized by three major outgassing episodes. The amount of methane
outgassed during the two first episodes is determined mainly by the initial
amount of methane in the proto-core, the time and duration of the core
overturn (,0.2–1 Gyr), and the efficiency of convection in the silicate core.
62
© 2006 Nature Publishing Group
NATURE|Vol 440|2 March 2006
silicate core is sufficiently hot for thermal convection to initiate.
The heat flux at the surface of the core increases, which leads to a
thinner outer clathrate-rich layer. Methane accumulates at the base of
the outer layer because the ocean is saturated. The methane can
ascend along buoyancy-driven cracks12. Afterwards, the silicate core
starts cooling down. The third and last episode is a natural conse-
quence of both the interior cooling and the crystallization of the
liquid layer. Once the ice I layer that crystallizes below the clathrate
layer becomes unstable against thermal convection (at 3.9 gigayears
(Gyr) in Fig. 2), hot thermal plumes, favoured by tidal dissipation22,
penetrate the clathrate layer and induce clathrate dissociation. As
shown by Figs 2a and 3, the final outgassing episode is more than
enough to maintain the atmospheric abundance at or above the
current level of 5% (ref. 11), and indeed to saturate the atmosphere
and hence soak the surface in extensive methane rainfall—evidence of
which is seen in the data4,23. Our calculations show that the final
episode must have started relatively late (after 3.5 Gyr), to explain the
high value of orbital eccentricity19 and to induce a sufficiently high
outgassing rate (Fig. 3).
Our numerical results provide an explanation for the appearance of
flow features and volcanic edifices, as suggested by both infrared3 and
radar4 observations, and circumstantial evidence of shallow subsurface
reservoirs of methane4,11,23. Although methane outgassing could have
occurred in the past (before 2.5 Gyr), only recent outgassing can
The third and last episode is triggered by upwelling icy plumes, whose
typical temperature (,250 K) is close to the dissociation temperature of
methane clathrate hydrate at a depth of 1 km. The maximum outgassing rate
is expected just after the onset of convection (,4 Gyr), when the surface heat
flow is maximal. Then, the methane release rate is expected to diminish as
tidal dissipation decreases owing to eccentricity decay and as the outer layer
and the conductive lid thicken. Panel a shows a representative methane
outgassing rate from 3.9 to 4.55 Gyr, assuming the decomposition of
10–50% of the clathrate mass remaining in the crust before the onset of
convection. Arrows indicate the possible onset time of outgassing obtained
for varying internal composition and rheology. We note that, for the large
range of parameter values explored, the three episodes of outgassing may
differ only in duration and vigour.
NATURE|Vol 440|2 March 2006
Figure 3 | Thickness of the available methane clathrate reservoir when the
final episode initiates. The beginning of the third episode is related to the
onset time (t onset) of thermal convection in the outer layer, and depends
mainly on the ammonia fraction x NH3 and the viscosity of ice I near its
melting point h m. Filled symbols, x NH3 ¼ 5%; open symbols, x NH3 ¼ 0%.
Labels 1, 2 and 3 refer to the three different schemes for evolution of
the silicate core shown in Fig. 1. The two dotted lines represent the
reservoir thickness required to counterbalance the methane loss rate
(1.25 £ 1010 cm22 s21), assuming that 10 or 50% of the clathrate reservoir
has been outgassed since the onset of the last episode. We show that the
thickness of the available reservoir (b clath) is determined mainly by the
evolution of the silicate core and that the duration and vigour of the last
episode depends on the ammonia fraction x NH3 and the ice viscosity at the
melting point h m. Even though a third outgassing episode starting as early as
3–3.5 Gyr is possible for some internal parameters, it requires either initial
eccentricities .20%, which is difficult to explain in view of the proximity of
Hyperion19, or a high-viscosity value (5 £ 1014 Pa s) and an ammonia
fraction of 5%, which leads to low surface heat flow (,10 mWm22) and
hence a very low outgassing rate. Nevertheless, our calculations show that
models with some fraction of ammonia and a relatively low concentration of
radiogenic elements in the silicate (core schemes 1 and 2) lead to both a late
onset (t . 3.8 Gyr) and a high outgassing rate. Such models require lower
initial eccentricity values (,10–15%) and retain larger masses of clathrate
reservoir through Titan’s history.
explain the presence of methane in the current atmosphere in the
absence of very massive, and hence surface-exposed, reservoirs to
buffer the loss of atmospheric methane in a short time (,107 yr) as
compared with the time elapsed since outgassing episodes 1 and 2.
The third outgassing episode would be due to thermal destabilization
of the clathrate layer to form a methane-filled chamber, which would
be outgassed periodically or catastrophically through eruptive pro-
cesses. Our finding is fully consistent with the carbon isotopic ratio
13 12
C/ C in hydrocarbon molecules close to the terrestrial value11,24.
This ratio reflects a bulk carbon inventory that did not participate in
the massive escape phase of the ancient post-accretional atmosphere,
in contrast to nitrogen, whose isotopic ratio 15N/14N is enhanced and
thus suggests escape11,24.
The presence of a significant amount in the atmosphere of the 40K
decay daughter 40Ar (refs 10, 11) (10% of the maximum amount that
could possibly be outgassed assuming a chondritic composition for
the rock core) can be understood in the framework of our model. The
interior was warm enough in the past (before 2.5 Gyr) for the water
ocean to be enriched in potassium through interactions with the
rocky core25. Argon could then have ended up in the atmosphere if
potassium-enriched water cryomagmas were directly deposited on
the surface25 or, more likely, if it was outgassed simultaneously with
methane. Although argon incorporates in the clathrate hydrate to a
lesser extent than methane26,27, mobilization of argon contained in
the water ocean through methane bubble formation (occurring
during episodes 1 and 2), as well as successive destabilization of
methane clathrate containing traces of argon, may progressively
enrich the atmosphere in argon. The presence of argon is thus a
strong indicator of past and recent internal activities.
© 2006 Nature Publishing Group
LETTERS
Future fly-by observations will provide pertinent tests of our ideas.
If a subsurface ammonia–water ocean and a rocky core are present, as
predicted by our model, Titan’s tidal secondary potential Love
number k 2 should be higher than 0.3 (ref. 28) and its moment of
inertia factor C/MR 2 should be lower than 0.33; these quantities
should be measured by the Radio Science Subsystem later in the
mission29. Our model predicts that a major outgassing event
occurred about 2 billion years ago, followed by cessation of methane
injection and associated cryovolcanic activity until a smaller episode
in the recent past. This has two testable consequences.
First, the cratering record should show a surface age neither
significantly more nor significantly less than 2 billion years. Since
the two impact craters found so far4 imply a slightly smaller age if
no further such structures exist, we predict that several more large
(50–100 km) impact craters—intact or partially eroded/buried—will
be found by the Orbiter. Moreover, depletion of methane before the
final outgassing episode may have cooled and thinned the atmos-
phere30, augmenting the number of smaller impact craters relative to
what is expected with the screening effects of today’s atmosphere.
Second, the two first episodes would have been widespread and able
to put a large amount of liquid methane on the surface for extended
periods of time, whereas the recent one would be more localized and
controlled by the activity of hot ice plumes. Evidence for modifi-
cation of large areas of the surface by the action of liquid methane
seems to be present in the radar data in the form of widespread fluvial
channels and relatively featureless (sediment-filled?) basins4 ;
additional remote sensing observations and crater-density statistics
should reveal geological differences between terrains formed during
the ancient widespread episodes and recently modified terrains
associated with cryovolcanic edifices.
Received 22 June; accepted 28 November 2005.
1. Yung, Y. L., Allen, M. A. & Pinto, J. P. Photochemistry of the atmosphere of
Titan: comparison between model and observations. Astrophys. J. Suppl. Ser. 55,
465–-506 (1984).
2. Lunine, J. I., Stevenson, D. J. & Yung, Y. L. Ethane ocean on Titan. Science 222,
1229–-1230 (1983).
3. Sotin, C. et al. Release of volatiles from a possible cryovolcano from near-
infrared imaging of Titan. Nature 435, 786–-789 (2005).
4. Elachi, C. et al. Cassini radar views the surface of Titan. Science 308, 970–-974
(2005).
5. Porco, C. et al. Imaging of Titan from the Cassini spacecraft. Nature 434,
159–-168 (2005).
6. West, R. A., Brown, M. E., Salinas, S. V., Bouchez, A. H. & Roe, H. G. No oceans
on Titan from the absence of a near-infrared specular reflection. Nature 436,
670–-672 (2005).
7. Stevenson, D. J. in Proceedings Symposium on Titan, Toulouse, France, 9–-12
September 1991 29–-33 (European Space Agency, Noordwijk, The Netherlands,
1992).
8. Grasset, O., Sotin, C. & Deschamps, F. On the internal structure and dynamics
of Titan. Planet. Space Sci. 48, 617–-636 (2000).
9. Owen, T. On the origin of Titan’s atmosphere. Planet. Space Sci. 48, 747–-752
(2000).
10. Waite, H. et al. Ion-neutral mass spectrometer results from the first flyby of
Titan. Science 308, 982–-986 (2005).
11. Niemann, H. et al. The abundances of constituents of Titan’s atmosphere
from the GCMS instrument on the Huygens probe. Nature 438, 779–-784
(2005).
12. Lunine, J. I. & Stevenson, D. J. Clathrate and ammonia hydrates at high
pressure—application to the origin of methane on Titan. Icarus 70, 61–-77
(1987).
13. Kuramoto, K. & Matsui, T. Formation of a hot proto-atmosphere on the
accreting giant icy satellite: Implications for the origin and evolution of
Titan, Ganymede, and Callisto. J. Geophys. Res. 99, 21183–-21200 (1994).
14. Mousis, O., Gautier, D. & Bockelée-Morvan, D. An evolutionary turbulent
model of Saturn’s subnebula: implications for the origin of the atmosphere of
Titan. Icarus 156, 162–-175 (2002).
15. Lowell, R. P. & Rona, P. A. Seafloor hydrothermal systems driven by the
serpentinization of peridotite. Geophys. Res. Lett. 29, 1531, doi 10.1029/
2001GL014411 (2002).
16. Loveday, J. S. et al. Stable methane hydrate above 2 GPa and the source of
Titan’s atmospheric methane. Nature 410, 661–-663 (2001).
17. Sloan, E. D. Clathrates Hydrates of Natural Gases 2nd edn (Marcel Dekker, New
York, 1998).
63
LETTERS
18. Grasset, O. & Pargamin, J. The ammonia water system at high pressures:
implications for the methane of Titan. Planet. Space Sci. 53, 371–-384 (2005).
19. Tobie, G., Grasset, O., Lunine, J. I., Mocquet, A. & Sotin, C. Titan’s internal
structure inferred from a coupled thermal-orbital model. Icarus 175, 496–-502
(2005).
20. Sohl, F., Hussmann, H., Schwentker, B., Spohn, T. & Lorenz, R. D. Interior
structure models and tidal Love numbers of Titan. J. Geophys. Res. 108, 5130,
doi 10.1029/2003JE002044 (2003).
21. Durham, W. B., Kirby, S. H., Stern, L. A. & Zhang, W. The strength and
rheology of methane clathrate hydrate. J. Geophys. Res. 108, 2182, doi:10.1029/
2002JB001872 (2003).
22. Sotin, C., Head, J. W. & Tobie, G. Europa: tidal heating of upwelling thermal
plumes and the origin of lenticulae and chaos melting. Geophys. Res. Lett. 29,
1233, doi:10.1029/2001GL013844 (2002).
23. Tomasko, M. G. et al. Rain, winds and haze during the Huygens probe’s
descent to Titan’s surface. Nature 438, 765–-778 (2005).
24. Lunine, J. I., Yung, Y. L. & Lorenz, R. D. On the volatile inventory of Titan from
isotopic abundances in nitrogen and methane. Planet. Space Sci. 47, 1291–-1303
(1999).
25. Engel, S., Lunine, J. I. & Norton, D. L. Silicate interactions with ammonia-water
fluids on early Titan. J. Geophys. Res. 99, 3745–-3752 (1994).
26. Dyadin, Y. A., Aladko, E. Y. & Larionov, E. G. Decomposition of methane
hydrates up to 15 kbar. Mendeleev Commun. 7, 34–-35 (1997).
27. Dyadin, Y. A., Larionov, E. G., Mirinski, D. S., Mikina, T. V. & Starostina, L. I.
64
© 2006 Nature Publishing Group
NATURE|Vol 440|2 March 2006
Clathrate formation in the Ar–-H2O system under pressures up to 15000 bar.
Mendeleev Commun. 7, 32–-34 (1997).
28. Castillo, J., Rappaport, J., Moquet, A. & Sotin, C. Clues on Titan’s internal
structure from Cassini-Huygens mission. Lunar Planet. Sci. (LPI contribution no.
1109) 33, 1989 (2002).
29. Rappaport, N., Bertotti, B., Giamperi, G. & Anderson, J. D. Doppler
measurements of the quadrupole moments of Titan. Icarus 126, 313–-323
(1997).
30. Lorenz, R., Mckay, C. P. & Lunine, J. I. Photochemical collapse of Titan’s
atmosphere. Science 275, 642–-646 (1997).
Supplementary Information is linked to the online version of the paper at
www.nature.com/nature.
Acknowledgements The first author benefits from post-doctoral fellowships
from the Lavoisier programme of the ‘Ministères des Affaires étrangères’ and
from the Centre National d’Etudes Spatiales (France). Support for the project
and paper was provided by the Cassini Project and by the INSU ‘Programme
National de Planétologie’.
Author Information Reprints and permissions information is available at
npg.nature.com/reprintsandpermissions. The authors declare no competing
financial interests. Correspondence and requests for materials should be
addressed to G.T. (gabriel.tobie@univ-nantes.fr).