Exoplanet Chemistry

157
8
Exoplanet Chemistry
Katharina Lodders
8.1
A Goodly Gallery of Planets
The current state of observations on the chemical nature of exoplanets is comparable

to that of the early days of solar-system planetary science, when not much more
than the orbits and densities of planets were known. Information about exoplanet
chemical compositions from direct observations is still limited. However, chemical
and physical observations and models for planets in our solar system developed
since the 1950s should provide the tools to understand other planets, if models for
solar-system planets are indeed correct. As such, the observational data to come in
the following years undoubtedly will test the fine-tuned planetary formation and
evolution models that seem to work well for understanding the planets in our solar
system.
There are four broad basic types in our gallery of planets in the solar system
and very likely in most of the other planetary systems. The reason for this is the
first-order similarity of the relative elemental abundances of most stars in the
galactic disk and the compounds that the elements can form. Planets originate
from the materials in a stellar accretion disk that had a composition similar to
that of the final star itself. This limits the possible chemical nature of the building
blocks of planets, despite the wide range of possible physical properties in the
different accretion disks around other stars. Different condensed phases in an
accretion disk are stable within different temperature limits; refractory phases are
stable up to high temperatures, and volatile phases are only stable at relatively
low temperatures. Thus, the formation or thermal destruction of solids at different
temperatures can lead to a fractionation of refractory elements in the solids from
the more volatile elements that are in the gas.
Figure 8.1 shows how the four types can be categorized by their chemical nature.
Terrestrial-like planets, defined in the following, are composed mainly of rocky
substances, and are represented in our solar system by Mercury, Venus, Earth, Mars,
Ceres and, although not planets, by a large number of asteroids. Terrestrial planets
can have atmospheres which are, however, insignificant to the total planetary mass
budget. The second type consists of planets largely made of rocky and icy materials
Formation and Evolution of Exoplanets. Edited by Rory Barnes

Copyright  2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
ISBN: 978-3-527-40896-2
158 8 Exoplanet Chemistry
300
1.0
15
250
0.002
Earth masses of phase
200
10
150
0.5
0.001
100
5
50
0.0 0.000 0 0
M V E M P U N J S
Rocky Rocky Rocky Rocky
(metallic + icy + icy + icy
+ silicates) + H2 /He + H2 /He
(a) (b) (c) (<50%) (d) (>50%)
Figure 8.1 The four chemical types of plan- (b) Rock plus ice (Plutonian) planets primar-
ets represented by solar-system objects. The ily consist of rocky and icy solids. (c) Giant
shading indicates the contributions (in Earth (Neptunian) planets consist mainly of rocky
masses) of different phases: black = metallic, and icy substances and less than half is He
light gray = silicate, dark gray = various ices, and H2 . (d) Gas-giant (Jovian) planets are
second light gray on top of ices is H2 + He dominated in mass by He and H2 . Data
gas. (a) Rocky (terrestrial) planets mainly from Lodders and Fegley [1].
consist of silicates and metallic FeNi(S).
(e.g., H2 O, CH4 , N2 ices), which may have tenuous atmospheres composed of icy
evaporates. In the solar system, such Plutonian planets, as one would call them,
are only represented by IAU-defined dwarf planets such as Pluto, Eris, Haumea,
and Makemake, but most objects beyond Neptune should qualify for this chemical
distinction. Some of the asteroids in the outer asteroid belt may also belong to
the rocky–icy objects. Giant (Neptunian) planets consist mainly of compounds
derived from rocky and icy substances in the accretion disk and contributions from
He and H2 to their total masses are considerable but remain below 50%. Uranus
and Neptune could only accrete about 10–20% of their total mass as H and He
gas, which makes them ‘‘gas-poor’’ giant planets, the third type. The fourth type,
the Jovian gas-rich giant planets in which H and He dominate, is represented by
Jupiter and Saturn, and they have the closest similarity in elemental composition
to their host star, the Sun.
Silicates and metallic phases define the rocky portion of a planet, however, their
proportions are reasonably well known only for the terrestrial planets. For all other
objects in Figure 8.1, a silicate/metallic mass ratio according to solar elemental
abundances and geochemical considerations was applied to the ‘‘rocky’’ portion.
For the giant planets, similar considerations were used to divide the heavy element
fractions into ‘‘metallic, silicate, and icy’’ components.
Most known exoplanets have masses between that of Saturn and 10 times that
of Jupiter (1MJup = ∼318 Earth masses, ME ); but smaller Neptune size objects
(∼0.05MJup = ∼17ME ), and objects of a few Earth masses (1ME = 0.003MJup ) are
8.2 Elemental Ingredients of Planets 159
also known (see the frequent exoplanets updates on Jean Schneider’s webpage
at http://exoplanet.eu). A few companions to stars reach masses above ∼13MJup ,
where theoretical models predict that interior deuterium burning is possible. This
mass limit is usually taken to divide ‘‘planets’’ from ‘‘brown dwarfs’’ that are also
observed to orbit stars. However, this definition may be arbitrary (see Chapter 7).
To make this point, consider the object HD41004B b, which orbits its early M dwarf
host star at 0.0177 AU with an orbital period of ∼1.3 days. This is one of the closest
orbits of a nonstellar object around a star currently known. However, HD41004B b
has a mass of 18.4MJup –clearly above the D-burning mass limit, and is not really
counted as a planet.
The atmospheric gas and cloud chemistry in brown dwarfs is not much different
from that in gas-giant exoplanets if these objects have similar effective temperatures
(see below). The major difference would be the absence of deuterated gases such as
CH3 D, HDO, NH2 D, and HD in brown dwarfs that masquerade as huge planets.
If observationally feasible, the presence or absence of D-bearing gases may be a
means to distinguish large planets from brown dwarfs, but more importantly, it
would finally provide observational measures of the minimum mass required for
D-burning.
There is no other practical means to distinguish these objects other than by mass.
Differences in possible formation modes could also distinguish these larger objects.
Gas-giant planets and brown dwarfs can both form by gravitational instabilities
in a protostellar accretion disk (Chapter 5). Another prominent formation model
for giant planets is ‘‘core accretion’’ (Chapter 5). This is accretion of larger
protocores and subsequent gravitational capture of large quantities of gas (see
below), but this mechanism may not work for making massive brown dwarfs in
a timely fashion in a short-lived accretion disk. In principle, the different modes
of formation can imprint different chemical compositions onto a gas-giant planet
(see below). However, even the interpretation of the observed chemistry of the
well-studied gas-giant planets Jupiter and Saturn is not unique. Currently, a large
number of researchers seem to favor the core-accretion model for Jupiter’s and
Saturn’s origins, but the case is not clear-cut. Hence, the much more limited
chemical information that is available so far for gas-giant planets outside the solar
system cannot be used to distinguish planets and brown dwarfs by their formation
mode.
8.2
Elemental Ingredients of Planets
The densities of exoplanets can be found if their radii and masses can be determined.
Radial velocity measurements provide masses (or upper mass limits M sin i) for
exoplanets, and radius determinations are available for planets observable during
transits of their host star (see Chapter 1). Once planet densities are available,
planetary overall compositions may be estimated from the chemical and physical
properties of plausible components. Several models for the evolution of gas-giant
exoplanets and their interiors and spectra have been described by, e.g., Burrows
et al. [2–4], Charbonneau et al. [5], Chabrier et al. [6], Fortney et al. [7–9], Guillot
[10], Guillot et al. [11, 12], Sudarsky et al. [13]. A wide range of mass–radius relations
for terrestrial to gas-giant planets is discussed by Fortney et al. [14]. There are also
several different structure types of terrestrial-like planets that have been modeled,
based on different components that may result from solar-composition element
mixtures, e.g., Fortney et al. [14], Seager et al. [15].
Ultimately, it is the abundance of the chemical elements and the physical prop-
erties of their compounds that govern the constitution of a planet. As already
mentioned, most stars in the solar neighborhood have relative elemental abun-
dances similar to that of the Sun. Hence, the compounds and their relative amounts
available in the accretion disks around other stars should be similar to what was
present in the solar planetary accretion disk (i.e., the solar nebula). There are, of
course, limitations to this first-order approach because the distribution of these
compounds in a disk is expected to vary as physical disk properties vary. Planets
not only develop around forming G2V dwarf stars like the Sun but are also found
around lower mass F, K, and M dwarfs, and around stars with up to ∼4.5 solar
masses that have evolved into G and K giants. Detections of planets around very
massive and hotter dwarf stars are not yet reported. Planet formation around early
type (O, B, and A-type) dwarfs may be difficult to achieve as estimated planetary
formation and evolution times of a few to tens of million years may coincide with
stellar lifetimes, and early mass loss from massive stars may interfere with planet
formation in an accretion disk.
The estimated minimum mass of the solar nebula is about 0.01–0.02Msun ,
depending on the solar abundance compilation used and the adopted heavy
element fractions for the giant planets. Accretion rates through the disk onto
the young Sun are estimated as ∼10−5 to 10−7 Msun per year, depending on the
accretion stage. However, accretion disks around other stars may have had total
masses different from that of the solar nebula, accretion rates different from
that of their parental molecular clouds, and thus different thermal gradients
and evolution timescales of their disks. Disk lifetimes are important because the
formation of gas-giant planets requires that H2 and He-rich gas is accreted before
the disk dissipates. Meteoritic evidence from short-lived radioactive nuclides and
observations of accretion disks in young stellar system suggest that disk lifetimes
are at least two to three million years, with an upper limit of about 10 million
years, which is then the maximum time available for gas-giant planet formation.
On the other hand, the disk lifetimes and the presence of nebula gas may not
be as critical for the formation of terrestrial-type planets, which can form as long
as rocky planetesimals are abundantly available. The formation of the moon by
a giant impact occurred after core formation on the Earth was complete, which
was within 30 million years after the oldest known components in meteorites had
been processed in solar nebula gas ∼4600 million years ago [16]. The craters on
Moon, Mars, and Mercury are witness to the late bombardment period which lasted
until ∼3900 million years ago. Clearly, formation of the Moon and late planetary
8.2 Elemental Ingredients of Planets 161
bombardments are the larger finalizing events in terrestrial planet accretion, but
their timing is outside any reasonable lifetimes of the solar nebula.
Depending on the location of a star in the galactic disk and on the timing of star
formation, a star and its accretion disk may also have had metallicity different from
that of the Sun (i.e., elements heavier than He have different abundances relative
to H). This additional factor influences the outcome of a planet’s final constitution.
The metallicity of a star is important because it is observed that metal-rich stars
are more likely to harbor planets than stars of lower metallicity [17]. Most of the
known exoplanets are likely to be gas giants, but one should expect that there is
a similar preference for the occurrence of rocky planets around metal-rich stars.
The formation of planets is plausibly fostered by a metal-rich disk because a higher
mass density of the ‘‘metallic’’ elements (in the astronomer’s sense) provides more
rocky and icy building blocks, which should increase the probability of planet
accretion.
8.2.1
Diagnostics from Elemental Abundance Fractionations
The elemental composition of a gas-giant planet compared to the composition of

its host star may reveal more about its mode of formation, and again the relative
amounts of the heavy elements are of particular interest here. A gas-giant planet
that is essentially identical in elemental composition to its host star is likely to
have formed through gravitational instability in the protostellar accretion disk
([18, 19]; Chapter 4). On the other hand, in the core-accretion model for giant
planet formation, a solid protocore of ∼5–10ME has to form first, which is large
enough to gravitationally capture surrounding nebular H2 and He gas ([20–22],
see also [23], and Chapter 5). This leads to a gas-giant planet rich in elements
heavier than He and H when compared to the composition of the primary star if
formation of the initial core required accumulation of solids from wider regions
in the disk but gas was only accreted locally. (A similar result is obtained when
incomplete accretion of He and H2 fails to add to the rocky core abundances to
get back abundances of the host star.) This model also leads to the notion that all
giant planets formed by this mechanism should have ‘‘cores,’’ or more precisely,
the equivalent mass of heavy elements distributed somewhere in their interior
[24].
In addition to various accretion scenarios that separate solid components from
the He–H2 gas, chemical fractionations among heavy elements themselves in the
disk can cause nonsolar abundance ratios of the heavy elements found in planets.
For example, solar composition has a mass ratio of C/Si ∼ 3, whereas the entire
Earth has a ratio of ∼0.0003. This is a consequence of the different volatilities of
the chemical compounds formed by C and Si. The most likely place of fractionation
of C and Si was in the solar nebula, and the Earth could not accrete much C and
other volatile-element-bearing compounds to begin with.
8.3
Planetary Building Blocks
The basic results of chemistry in a H- and He-rich gas, such as the solar nebula,
remain applicable for systems that evolve with solar or close-to-solar overall
elemental compositions, if temperatures and densities (or total pressures) are
not dramatically different. It is well known from condensation calculations that
over a wide range of total pressures (i.e., P < 10−3 bars), the chemistry in a
very H2 and He-rich, solar-like gas leads to quite similar minerals and thermal
stability sequences ([25–27] and references therein). For example, usually the most
thermally stable solid containing most of the Fe is an FeNi metal alloy; for Mg,
it is the Mg-silicate forsterite (Mg2 SiO4 ), and for sulfur, it is troilite, FeS. Hence,
the available planetary building blocks in different disks should be comprised of
quite similar oxides, silicates, metals, sulfides, and ices. Aside from the thermal
stabilities, the quantities of the building blocks are limited by the abundances of
the elements that make them. This simplifies the modeling of possible exoplanet
compositions, even if the exact conditions in a former accretion disk and the
distribution of the solids are not known.
The radial distribution of the rocky, icy, and gaseous ingredients for the formation
of planets is determined by the temperature and total pressure gradients in accretion
disks. Therefore, even in the absence of any possible turbulent redistribution of
constituents, the available proportions of the solid planetary building blocks vary
with radial location, which will affect the locations, masses, types, and number of
resulting planets.
Three principal types of compounds, loosely referred to as rock, ice, and gas, are
available in accretion disks, and can wind up in planets. When discussing planetary
compositions, the terminology rock, ice, and gas is often kept, but it may relate to
the nature of the phases that have accreted, and not necessarily to the nature of
phases that occurs in a planet today. The relative proportions of gas, rock, and ice
in the accretion disk determine a planet’s possible mass but the equations of state
determine the planet’s size and which stable phases will be present.
The gas component is mainly H2 and He and includes all the remaining mass
that is left after rocky and icy condensates are removed. Depending on temperature
and the details of the accretion disk chemistry of ices, inert gases (noble gases, N2 )
can contribute a tiny mass to the gas. Since H and He make ∼98.5% of all mass in
solar-system material, these elements must make up the majority of mass in any
gas-giant planet (this assumes that there were no changes in the H- and He-content
from postaccretion processing of a giant planet that has migrated too close to its
host star).
The rocky and icy components (noble gases inclusive) make up ∼1.5% of the
total mass. Figure 8.2 shows the mass percentages in a solar elemental composition
system for various rocky and icy substances as a function of temperature. The three
panels correspond to different assumptions of thermodynamic equilibrium, which
mainly affect the amount and type of phases present at low temperatures.
8.3 Planetary Building Blocks 163
1.5
Approximate rock and ice masses
-- CH4
in a planetary accretion disk
of solar composition
-- NH3.H2O assuming equilibrium
1.0
Mass percent
CH4.(6-7)H2O
H2O
0.5
Talc, Refractory
hydrous silicates oxides
Silicates
Fe – S |
Magnetite Metal |
0.0
0 200 400 600 800 1000 1200 1400 1600
(a) Approximate temperature (K) at 10−5 bar
1.5
Approximate rock and ice masses

-- N2 in a planetary accretion disk
of solar composition
1.0 -- CO assuming nonequilibrium at low T
Mass percent
-- CO. 6H2O
H 2O
0.5
Silicates Refractory
oxides
Fe– S |
Magnetite Metal |
0.0
0 200 400 600 800 1000 1200 1400 1600
(b) Approximate temperature (K) at 10−5 bar
1.5
-- N2
-- CO Approximate rock and ice masses
in a planetary accretion disk
-- CO. 6H2O of solar composition
1.0
Mass percent
assuming non-equilibrium at low T

H2O and half of all C in stable organics
C as organics
0.5
Silicates Refractory
oxides
Fe – S |
Magnetite Metal |
0.0
0 200 400 600 800 1000 1200 1400 1600
(c) Approximate temperature (K) at 10−5 bar
The high-temperature rock includes elements like Si, Mg, Ca, Al, Ti, etc., which
form oxides and silicate minerals. Here, FeNi metal alloys and iron sulfide also
belong to the ‘‘rock’’ in the environs of planetary accretion disks. Formulas and
names of major minerals commonly encountered in planetary science and some
mineral systematics are listed in Table 8.1. A good working definition of ‘‘rock’’ is a
component containing the approximate elemental abundances and consisting of the
mineral phases that are observed in undifferentiated meteorites called chondrites.
All rocky compounds are relatively refractory and require high temperatures for
evaporation or they readily condense out of a gas of solar-system composition at
relatively high temperatures (see below for condensation). Since rocky material
is the last material to evaporate or the first to condense, it is always present
among possible solids in an accretion disk. Planets primarily made of such rocky
materials – the terrestrial(-like) planets – characterize the chemical makeup, but
not necessarily the size of such planets. The solar-system abundance of the elements
limits the amount of rocky material to ∼0.5% of all mass in an accretion disk.
At temperatures below ∼600 K, gas–solid reactions become sluggish and equilib-
rium may not be reached. As discussed in detail by Fegley [28], under equilibrium
conditions, one may expect the formation of sulfide and magnetite from metal,
and hydrated silicates from pyroxenes and olivine. However, considering reaction
kinetics, only sulfide formation is quantitative within the lifetimes of solar neb-
ula, whereas magnetite formation is a borderline case. The hydration of silicates,
usually associated with hydrothermal, high-pressure environs, in the ‘‘vacuum’’
of accretion disks can be ruled out since this requires reaction times that are
many orders of magnitude above estimated lifetimes of these disks. Therefore,
the nonequilibrium cases in Figure 8.2 do not show hydrous silicates as possible
planet-building phases.
Icy materials become stable below ∼200 K, and encompass substances containing
C, N, and O. The solids that more or less quantitatively sequester C, N, and O
only become stable at low temperatures in low-pressure solar-composition systems.
Hence, these low-temperature phases are collectively called ‘‘ice.’’ Water ice is the
most important ice because O is the third most abundant element in the solar
system, and water ice is the most refractory ice among these ices. (Oxygen is also
present in rocky materials but the other elements in rock are only abundant enough
to bind ∼20% of all available solar-system oxygen).
In addition to water ice, other possible abundant ices are solid methane, CH4 ;
methane clathrate CH4 · nH2 O; carbon monoxide, CO; carbon monoxide clathrate,
CO·nH2 O; carbon dioxide CO2 ; carbon dioxide clathrate, CO2 · nH2 O; ammonia,
NH3 , ammonia monohydrate NH3 ·H2 O; molecular nitrogen ice, N2 ; and nitrogen
clathrate, N2 · nH2 O. In the clathrate hydrates, with the generic formula X·nH2 O,
←−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−
Figure 8.2 The distribution of solid phases as a function of temperature at 10−5 bar total pres-
sure. The amounts shown are mass percent for solar-system composition. (a) Thermodynamic
equilibrium case, which is limited by slow gas–gas and gas–solid reaction kinetics at low
temperatures. (b and c) Possible (and nonunique) low-temperature nonequilibrium phases and
distributions.
Table 8.1 Typical minerals made of more abundant elements.
Mineral group End-member mineral Ideal formula
Olivine (Mg,Fe)2 SiO4 Forsterite Mg2 SiO4

Fayalite Fe2 SiO4
Pyroxene (Mg,Fe,Ca)SiO3 Enstatite MgSiO3
Ferrosilite FeSiO3
Wollastonite CaSiO3
Feldspar Anorthite CaAl2 Si2 O8
Albite NaAlSi3 O8
Orthoclase KalSi3 O8
Metal alloys Iron–nickel FeNi
Sulfides Troilite FeS
Pyrrotite Fe1−x S
Oxides Magnetite Fe3 O4
Hydrous silicates Talc Mg3 (Si4 O10 )(OH)2
n = 5–7 is the number of water molecules that characterize a unit structure into
which another molecule ‘‘X’’ can be retained as a ‘‘guest’’, which is why clathrates
are also called cage compounds.
One may add solid organic substances to the group of ‘‘ices.’’ High-polymer
organics can be stable up to temperatures of 400–600 K [29], which makes them
more refractory than water ice.
Gas reactions involving very stable molecules such as CO and N2 are unlikely
to reach thermodynamic equilibrium (see [30, 31]). This influences the type of
low-temperature C and N gases, which, in turn, determines the type and amount
of the ices that can form. In addition, the available proportions of the different ices
depend on the relative thermal stability of these ices.
The top graph in Figure 8.2 is the equilibrium case where all CO converts to CH4
gas, which then condenses as methane clathrate and CH4 ice. Both phases form
because the amount of water is insufficient to clathrate all CH4 gas as methane
clathrate (solar abundances give C : O about 1 : 2 but a ratio of 1 : (6–7) is needed
for methane clathrate). The thermodynamically stable low-temperature gas for
nitrogen is ammonia gas, which condenses as ammonia hydrate, NH3 ·H2 O.
The formation of methane ice(s) requires that the very stable CO gas, dominant
at high temperatures (e.g., >1500 K at 10−5 bars), was efficiently converted into
methane gas at low temperatures, which is unfeasible within typical lifetimes of
the disks (see [30]). With inefficient conversion, CO gas remains, and CO clathrates
and CO ice condense (Figure 8.2b). Some CO (several percent) is also converted
to CO2 since this reaction is kinetically favored and can proceed down to lower
temperatures (800–1000 K) at low total pressures. Thus, smaller amounts of CO2 -
bearing ices (equivalent to ∼10% of all O) can be present [30, 31], which, for clarity,
are not shown in the nonequilibrium case panels in Figure 8.2. Slow gas-phase
kinetics also prevents the conversion of N2 to NH3 and N2 gas remains. This leads
to N2 -clathrate ice (see [30, 31]).
Another possibility is that interstellar low-temperature organic substances re-

mained unprocessed or that such substances formed in the outer planetary disks
through catalyzed reactions such as in the Fischer–Tropsch-type process, where
Fe-metal acts as a catalyst to convert CO and H2 to hydrocarbons. This reduces the
amounts of other C-bearing gases and ices.
To first approximation, the exact C-, N-, and O-bearing phases are irrelevant
for mass balance as long as all C, N, and nonsilicate O are condensed as ices.
Condensation of all C, N, and nonsilicate O into ices makes about ∼1% of all
material in a solar-composition system. Water ice alone makes up to ∼0.6%,
depending on whether all remaining O after rock condensation is in H2 O gas or
remains in CO gas. However, full retention of the icy phases requires temperatures
less than ∼30 K, and one can expect that the most volatile ices such as CO, CH4 ,
and N2 ice are easily evaporated during formation and evolution of the accretion
disk.
Planets may form by accreting varying amounts of icy and rocky materials but
this does not mean that rocks and ices remain in a state of ‘‘ice’’ or ‘‘rock’’ as
one typically associates with it. Occasionally, it is convenient to collectively equate
the amounts of C, N, and O as ‘‘ice’’ in planetary materials, which reflects that
these elements may have been brought to the planet in low-temperature phases.
Similarly, ‘‘rock’’ is a convenient summary of all other condensed phases other than
‘‘ice’’. However, temperature and pressure conditions during planetary accretion
and subsequent planetary evolution can be quite different from those under which
these phases are stable in a low-temperature, low-pressure accretion disk.
Only small planetary objects such as Kuiper belt objects (dormant comets?) in
the outer solar system are most likely to have preserved their originally accreted
icy phases. However, photochemical reactions are known to produce chemical and
physical changes on surfaces of ‘‘icy’’ objects. On the other hand, icy phases and
rock can thermally decompose during the larger scale differentiation of a planet,
or react with other substances once they accreted into larger planetary objects (e.g.,
reactions of rocky materials with H2 upon accretion into a gas-giant atmosphere).
During accretion of smaller kilometer-size objects to Mars- and Earth-sized
objects, enough gravitational energy is released to allow melting and (partial)
evaporation of accreting solids, even if about half the gravitational energy is lost
through radiation. The point is that phases present in the accreting bodies may break
down, and the elements from such phases react and redistribute into phases that are
stable under conditions of the planetary environment. The separation of ‘‘rocky’’
elements into the silicate and metal portions in the Earth and other terrestrial
planets is a well-known example. Another result of planetary differentiation is the
redistribution of rocky elements within the Earth’s silicate portion into a crust, the
upper mantle of mainly olivine, and a lower mantle containing larger amounts
of a high-pressure modification of (Mg,Fe)SiO3 called silicate perovskite (not to
be confused with the Ca–Ti mineral perovskite which has a similar structure at
normal pressure). By analogy to meteorites, the prime candidates for constituents
of the metallic cores in terrestrial planets are the ‘‘rocky’’ elements Fe, Ni, and S,
which transform into high-pressure metal and sulfide phases in planetary interiors,
again phases in states quite different from those encountered in the low-pressure
mineralogy of chondritic meteorites.
Another example for the likely differentiation processes is the accretion of rock
and ice to form protocores of 5–20 Earth masses, in the core-accretion model for
Jupiter and other gas-giant (exo)planets. These protocores may differentiate in a
manner similar to the terrestrial planets. If heated by gravitational energy release
or radioactive decay, icy materials begin to outgas and to form an atmosphere. With
estimated formation times of up to about eight million years for Jupiter [20], at
least some partial outgassing can be expected before rapid H2 and He gas capture
occurs. In the absence of larger amounts of H and He, such an atmosphere should
contain the same gases that made the initially accreted ices, adjusted for local
temperature and pressure equilibria. Thus, an icy + rocky protocore may show
an atmosphere made of H2 O, CO, CO2 , N2 , and possibly some traces of methane
and ammonia before evolving into a gas-giant planet. If so, hot differentiated
and outgassed protocores may be detectable since these gases have characteristic
absorption bands.
In the disk instability model, an extended protoplanet (approximately few 100
Jupiter radii) may develop a rocky core. Recently, Helled and Schubert [32] addressed
the survival of grains and the sedimentation efficiency in the protoplanetary
atmospheres of gas-giant planets that form through disk instability. They find that
protoplanets of five Jupiter masses or larger cannot form silicate cores because
temperatures become too high. Smaller protoplanets allow grain settling and silicate
‘‘core’’ formation if contraction rates are low and low internal temperatures (<1300
K) are maintained. However, the temperatures reached during accretion are still too
high to allow the accretion of ices to the protoplanet. Thus, the protocore formation
process fractionates refractory rock from more volatile elements (C, N, O) that form
ices.
There is, however, no evidence that such ‘‘cores’’ formed during disk instability
or in the core-accretion model would remain intact as cores in giant planets,
presumably still composed of the heavy elements originally present in the initial
core structures. The destruction of rocky and icy protocores during the gradual
growth of a giant planet cannot be excluded. On the other hand, the interior
structures (moments of inertia) and thermal evolution models of Jupiter and
Saturn require that denser cores are present today [24, 33]. The amount of heavy
elements in Jupiter may be up to 42ME , depending on the mass distribution of these
elements between the gaseous envelope and the core. Jupiter’s interior structure
can be consistently modeled not only without a heavy element core but also with a
core of up to 10ME . Saturn’s structure seem to require core masses made of heavy
elements between 10 and 25ME , where up to 10ME of heavy elements may be
located in the envelope. One larger uncertainty in these interior models is that the
equations of state for H and He at the extreme temperatures and pressures within
Saturn, Jupiter, and the more massive giant exoplanets are not yet well known.
Thus, whether cores are remnants of the original heavy element protocores that
never eroded, or are results of interior phase separations is not fully clear (see also
[11, 34]).
The destruction of rock and ice certainly happens during the later stages of
accretion when smaller planetesimals plunge into a hot and H2 -rich atmosphere
of a more mature gas-giant planet. For example, accreting water ice evaporates as
the growing planet is gravitationally heated. Then water vapor becomes a more
prominent constituent of the gaseous atmosphere. Another example is solid organic
(C-bearing) compounds that react with H2 to form methane, a major component
in cool gas-giant planets and also found in gas-giant exoplanets. Similarly, iron
sulfide contained in rocky accreting material can easily break down and react with
H2 to H2 S at temperatures above 600 K. Hydrogen sulfide remains as a gas in a
hot giant planet atmosphere (below the NH4 SH cloud base at ∼200 K) although it
may also be photolyzed by UV light (λ < 280 nm) at higher altitudes.
Overall, the content of heavy elements in the observable atmospheres of gas-giant
planets is from three principal sources:
1) The first source consists of elements present in gaseous form in the planetary
accretion disk. In the case of the core-accretion model, the stage of gravitational
H2 and He capture may also include volatile elements in noncondensed phases
CO, N2 , CH4 , etc., depending on temperature in the protoplanetary accretion
disk. In the disk instability model of giant planet formation, all elements
accrete in the same proportions from the accretion disk as present in the host
star.
2) If giant planets form through core accretion, the relative heavy element to
hydrogen ratio in the outer atmosphere may be higher than that in the host
star. There are two possibilities for this. First, the rock and ice protocores erode
during accretion of H2 -rich gas and subsequent planetary evolution. However,
transport and integration of material from eroding cores into the overall planet
may become limited through the metallic and liquid H2 /He layers if the planet
is large, and if the planet is not fully convective during its entire evolution.
The second possibility is inefficient H2 and He accretion, which also causes a
higher metal to hydrogen ratio.
3) The third source of rocky and icy elements in outer planetary atmospheres
comes from the late accretion of rocky and icy planetesimals such as asteroidal
or cometary bodies. These contributions affect elemental abundances in giant
planets irrespective of their formation mode (core accretion or disk instability
mechanism). In the solar system, the time of heavy bombardment of the inner
planets lasted until about ∼3.9 billion years ago, so accretion of larger rocky
and icy objects by the giant planets may have proceeded for ∼0.7 billion years
after the start of solar-system formation. However, given the mass currently
observed in asteroids (less than one Earth mass) and in Kuiper Belt objects
(around one Earth mass), the absolute amount contributed to giant planets
during late-stage evolution may only have been in the Earth-mass range, and
may not be a dominant source for most heavy elements in the observable
atmospheres of Jupiter and the other outer planets.
8.4 Similarities between Gas-Giant Planets and Brown Dwarfs 169
8.4
Similarities between Gas-Giant Planets and Brown Dwarfs
The densities for the known exoplanets suggest that planets with Jupiter masses and
above are H2 - and He-rich gas-giant planets. A first-order assumption for modeling
is that such planets have similar elemental compositions as their host stars, that is,
near-solar or somewhat enriched in elements heavier than helium. This assumes
that there were no physical fractionations of the icy and rocky phases through radial
redistributions so that the formation location of a gas-giant planet indeed has the
same abundances of the elements as the host star. It further assumes that there
were no chemical gas–solid fractionations and no fractionations between the solid
icy and rocky phases during planetary accretion or by the differentiation process in
the planet’s interior. A planet’s mass and the overall elemental composition are then
taken as inputs to model the planet’s interior density and temperature structure.
The temperature–pressure structure of the outer atmosphere is influenced by
the opacity of the gases and condensed phases that are present, and therefore
chemistry becomes important. The chemistry is also dependent on T and P, hence
the construction of models for the pressure–temperature (P–T) structure of a
planet is an iterative process [35–38]. The atmospheric P–T structure also depends
on the age of the object and its radiative cooling rates. Yet another variable is the
distance of a planet to its host star and the radiation that it receives from it.
The outer atmospheric chemistry of most gas-giant planets should be dominated
by molecular and atomic chemistry so that they resemble either cool methane
T dwarfs, hotter brown dwarfs of type L, or even late M dwarfs. A major factor
for the astonishing range in the chemical variations is the different distances of
gas-giant exoplanets to their host stars (see below). The type of chemistry causes
differences in the optical, far-red and near-infrared spectra, and guidelines of what
we can expect for exoplanet spectra are already available from the many spectra
of ultracool stars and brown dwarfs. The appearance and disappearance of certain
molecular bands and atomic lines in optical and near IR spectra are applied in the
classifications of hot L and cool T dwarfs, which as ‘‘brown dwarfs’’ are the links
between real dwarf stars of type M and planets like Jupiter [39]. The atmospheric
chemistry that characterizes M stars (strong absorption bands of gaseous TiO, VO,
CO), L dwarfs (strong alkali lines, strong CO, H2 O, FeH, CrH bands), and T dwarfs
(CH4 , H2 O bands) should be quite similar to what could be seen for exoplanets of
different ages and locations from their primaries. The following compares brown
dwarf and exoplanet atmospheres.
The P–T structure for Jupiter, two T dwarfs (Gl570D, Gl229B), two generic L
dwarfs, two M dwarfs, and M giant stars are shown in Figure 8.3 to give some
orientation of the T and P range that we are concerned with (data from [40, 41]).
Note that the P and T axes are reversed so that T and P increase toward the origin
(‘‘into the object’’). The T and L dwarfs have masses above ∼13MJup , the minimum
mass required for D-burning, but are below ∼60–70MJup needed for H-burning.
The T dwarf Gl229B has an effective temperature Teff ∼ 960 K and a mass of
∼58MJup . It is a binary companion to the M dwarf star Gl229A at a separation of
Jupiter dT Gl570D 800 K,1000 m s−2

125 K
25 m s−2 dT Gl229B
960/1000
500
dL 1500,3000
dL 1800,1000
1000
Temperature (K)
dM 2500,1000
1500
m s−2
10
Giants, 3000 K
2000 00 1
,10
00 0.1
30
dM 0.01
2500
3 2 1 0 −1 −2 −3 −4 −5 −6
Log (total pressure (bar))
Figure 8.3 Pressure and temperature structures in Jupiter,

brown dwarfs, and cool dwarf and giant stars. Numbers
after labels indicate effective temperature (K) and surface
gravity (in m s−2 ). Data from Marley et al. [40], and Marley
and Saumon [41], except for late type stars from Gustafsson
et al. [44].
∼44 AU, so one could think of this brown dwarf as a very massive planet. The
T dwarf Gl570D has similar properties Teff = 800 K and 34MJup [42]. The P–T
models for the T dwarfs have a surface gravity of ∼1000 m s−2 (Jupiter’s is ∼25
m s−2 ) and the radii of T and L dwarfs are comparable to that of Jupiter and gas-
giant exoplanets. Models for L dwarfs with relatively similar surface gravities and
effective temperatures of 1500 and 1800 K are included for comparison. Figure 8.3
shows that objects with higher effective temperatures have higher temperatures at
the same total pressure level (e.g., at 1 bar: Jupiter 165 K, Gl229B ∼750 K, L dwarf
∼1900 K). The age of an object is also an issue because without an internal heat
source cooling occurs over time. Thus, objects of similar mass may show different
spectral characteristics if they differ in age. The L dwarf model is for an object of
∼33MJup and has Teff = 1800 K, whereas the methane T dwarf Gl570D with about
the same mass (34MJup ) only has Teff = 800 K and different spectral characteristics.
Thus, if not externally heated, the effective temperature of objects of similar mass
depends on age (see, e.g., [3, 43] for P–T evolution of brown dwarfs with time).
Since gas-giant planets, like brown dwarfs after D-burning, lack strong internal
energy sources and are of similar overall elemental composition, the chemistry of
their outer atmospheres should change similarly with age, if they are not heated by
radiation of their hot stars. Figure 8.4 shows characteristic P–T profiles for Jupiter
and Jupiter-mass exoplanets at different separations from their primary star.
The model atmospheres of gas-giant exoplanets with about the same gravity
as Jupiter illustrate how the atmospheric P–T structure changes with a gas-giant
planet’s position from its primary. For comparison, the brown dwarf P–T models
from the previous figure are shown using thin lines. The P–T structures for
young (like Jupiter 4.6 billion years ago) planets are shown at the top and plotted
parallel to the Jupiter profile [9]. These Jupiter-mass planets have an intrinsic Teff
of ∼100 K if they would not receive any flux from the host star. However, when
0
2 AU
Jupiter 1 “Hot”
125 K 0.5 jupiter
25 m s−2 0.3
g = 25 m s−2
0.2
500
AU “Close-in”
dL 0.08 planet
0.06
0.05 g = 15 m2 s−1
1000
Temperature (K)
dT
1500
0.04 AU “Close-in”
planet
g = 15 m2 s−1
2000 0.03 AU
dL tres-3
28.4 m s−2
dM
dM
2500 Giants
3 2 1 0 −1 −2 −3 −4 −5 −6
Figure 8.4 Pressure–Temperature structures of Jupiter-mass

planets at different distances (indicated by AU) from their
primary stars (planet data kindly provided by J. Fortney; see
also [8, 9]). The P–T structures of some of the more mas-
sive objects from the previous figure are shown with thin
lines for comparison.
exposed to radiation from a star like the Sun, the P–T profiles change as shown,
from top to bottom, for planetary orbital distances of 2, 1, 0.5, 0.3, and 0.2 AU,
respectively. These P–T profiles become similar to that of T dwarfs, as seen from
the coinciding profile of the T dwarf Gl 229B and that of a Jupiter-mass planet
at 0.3 AU. However, because Jupiter-mass objects have lower surface gravities
than brown dwarfs do, they will have lower pressures in their photospheres if
they have the same effective temperatures. Conversely, the depth level from the
top of the atmosphere at which the same pressure and temperature is reached in
a giant planet is larger than in brown dwarfs because a larger overlying mass is
required to build up this pressure. Therefore, there is more opacity and the effective
temperature of such giant planets will be lower; e.g., compare the T dwarf Gl580D
(Teff ∼ 800 K) and the hot Jupiter at 0.5 AU with Teff ∼ 370 K.
The second set of profiles in Figure 8.4 is for planets at smaller orbital distances
(0.05–0.08 AU). These profiles (provided by J. Fortney) are for a somewhat lower
gravity of 15 m s−2 , corresponding to a Saturn-like planet (∼10 m s−2 ) instead
of Jupiter (∼25 m s−2 ). However, the change in gravity does not change the P–T
structures as much as does the closer proximity to the primary (taken to be a
Sun-like star). The upper parts of these atmospheric P–T profiles resemble those of
L dwarfs as radiative heating from the primary leads to stronger adiabatic expansion
of the atmospheres (less dense; shift to lower total pressure) when compared to the
planets located at 0.2 AU and beyond. Many known exoplanets are extremely close
to their host stars with orbits of semimajor axis <0.1 AU and with orbital periods
of a few days. The outer atmospheres of these heated ‘‘Pegasi planets’’ (after the
prototype 51 Peg b), ‘‘Roaster Planets’’ or ‘‘Hot-Jupiters’’ can reach temperatures
exceeding 1500 K. Without absorbed flux from the primary, such planets may only
have effective temperatures of a few hundred kelvin (depending on age). However,
because of the high temperatures, the ongoing chemistry should be comparable to
that in hot brown dwarfs (L dwarfs), hence spectra of such exoplanets should be
comparable to ‘‘free-floating’’ brown dwarfs of type L. TrES-1b and HD 209458 b
are well-studied, transiting, hot, Pegasi-type exoplanets with outer atmospheres
heated to temperatures that correspond to those of brown dwarfs near the L to T
dwarf transition (Teff ∼ 1200 to ∼1400 K).
The two planetary thermal profiles at the highest temperatures are for planets
at 0.04 and 0.03 AU, the latter a characteristic of the planet Tres-3 b. Exoplanets
that are in extremely close orbits with orbital periods of a few hours may develop
atmospheric thermal inversions [8, 37, 45] where high temperatures and tenuous
pressures that are comparable to regimes in giant and supergiant star atmospheres;
and hence the comparison to these much more massive objects here.
8.4.1
Chemistry in Gas-Giant Planets
A useful start of the chemistry discussion is with the more abundant reactive
elements (e.g., C, N, O) because their chemistry influences that of less-abundant
elements. For example, many elements present as monatomic gases can form
1 AU
r
pite
Ju
500 CH4 (g)
O
=C
4
CH 0.06 AU
Temperature (K)
1000
1500
0.04 AU
2000
CO (g)
0.03 AU
2500
3 2 1 0 −1 −2 −3 −4 −5 −6
Figure 8.5 Carbon chemistry in low-mass objects is domi-

nated by the gases CO (at high T and low P) and CH4 (at
low T and high P). The solid line shows where CO and CH4
have equal abundances. The dotted lines are P–T profiles
for Jupiter and Jupiter-mass objects at different separations
(see previous figure).
oxides in cooling gases, depending on the availability of oxygen. In cooler planets,

there are several coupled reactions between molecular gases and condensates, but
many of these reactions are regulated by the C and O chemistry.
One major chemical difference between T dwarfs and L dwarfs is in their carbon
and oxygen chemistry, which is the same as for cool (old and distant) and hot
(young/and or heated) Jupiter-mass planets. In T dwarfs and planets like Jupiter,
methane is the major C-bearing gas. In L and M dwarfs and close-in gas-giant
planets, it is CO gas. The chemistry of C and O as functions of T and P are shown
in Figures 8.5 and 8.6, which are in the same format as the previous figures.
The curves indicate where the major C and O gases are equal in abundance and
define ‘‘fields’’ where a certain gas is dominant. Many details of the C, N, and O
distribution and speciation are discussed in Lodders and Fegley [46]. Over a large
P–T range, the equilibrium distribution of carbon is controlled by the reaction
Water ice 1 AU
er
500 pit
Ju
H2O (g)
CO
O=
H2
0.06 AU
Temperature (K)
1000
tion
ensa
tite cond
a
Enst
terite
Fors
1500
0.04 AU
2000
CO (g)
0.03 AU
2500 O (g)
3 2 1 0 −1 −2 −3 −4 −5 −6
Figure 8.6 Oxygen chemistry in low-mass objects is dom-

inated by the gases CO and water. The solid line shows
where CO and water have equal abundances. The dotted line
are P–T profiles for Jupiter and Jupiter-mass objects at dif-
ferent separations (see also Figure 8.5).
between CO and CH4 through the net reaction:
CO + 3 H2 = CH4 + H2 O (8.1)
This reaction, which depends on total P and temperature, describes the equimolar
abundance curve of CO and CH4 that bisects the C distribution diagram (Figure 8.5).
At higher temperatures and lower total pressures, CO is the major C-bearing
gas, and methane abundances gradually decrease when moving away from the
equimolar abundance curve into the CO field. For comparison, P–T profiles of
Jupiter, and Jupiter-like planets at 1, 0.06, and 0.04 AU from Figure 8.4 are overlaid.
The region dominated by CO coincides with the upper atmospheric P–T profiles
of hot Pegasi-type planets (and L and M dwarfs). The major C-bearing gas in
Jupiter’s outer observable atmosphere is CH4 , and moving Jupiter to 1 AU or
0.2 AU would not change this. By moving a planet closer to the primary, the
internal P–T structure shifts closer to the CO=CH4 curve. The CO abundances
increase exponentially as one moves more toward the CO field, where CO reaches
its maximum abundance. The boundary is crossed if a Jupiter-like planet moves to

0.08 AU or closer, and CO becomes the major gas and CH4 abundances continue
to drop toward higher T and lower P.
Objects whose upper atmospheric P–T profiles fall into the CH4 -rich field are
expected to show methane absorption bands and only little CO absorption, if
any at all. However, this only applies if the CO to methane conversion reaches
equilibrium values. This is the same problem as noted before for the C chemistry
in protoplanetary disks. However, due to the higher total pressures in planetary
atmospheres, the CO to methane conversion can proceed down to ∼1000 K,
whereas in the solar nebula, the temperature for freezing the conversion reaction
was about ∼1500 K. At quench-level temperatures in planetary or brown dwarf
atmospheres, the chemical reaction timescale becomes too long when compared
to atmospheric mixing timescales, and the CO to methane conversion will not
proceed to completion. Thus, the amount of CO is frozen-in or quenched, and
CO is overabundant when compared to its expected equilibrium abundance at
low temperatures. The overabundance of CO is observed in Jupiter, Saturn, and
Neptune and is plausibly explained by quenching of the CO to CH4 conversion due
to rapid vertical mixing (see [47] for a discussion). Given the similarity in chemistry,
an overabundance in CO was also predicted for the first discovered methane T
dwarf Gl 229b [48], which was subsequently observed [49]. The overabundance of
CO is now established for several T dwarfs [50]. Hence quenching of the CO to
methane conversion should indeed be a universal phenomenon in atmospheres
with thermal profiles in the CO-rich field at higher temperatures and in the CH4
field at lower temperatures. If so, observations and quantization of CO abundances
in methane-rich objects provide a tracer of the mixing processes in such objects.
The oxygen distribution is shown in Figure 8.6. The major O-bearing gases are
CO and H2 O. Under conditions where CO is the major C-bearing gas, most O is
also in CO but H2 O is always quite close in abundance. With a bulk solar C/O = 0.5
and all carbon tied to CO, oxygen should be equally distributed between CO and
H2 O. However, SiO and other metal oxides reduce the H2 O abundances and the
more stable CO becomes the most abundant gas. The conversion reaction of CO
to methane produces equimolar amounts of H2 O. The H2 O abundance increases
proportionally to the decrease in CO. Therefore, the curve where CO and H2 O are
equal in abundance is at a similar location as the CO=CH4 abundance curve in
Figure 8.5.
Abundances of O-bearing gases are affected by the formation of condensates in
the atmospheres. Forsterite (Mg2 SiO4 ) and enstatite (MgSiO3 ) are the two most
important sinks for oxygen at high temperatures because Mg and Si are the
most abundant rock-forming elements. All high-temperature oxides and silicates
consume about 20% of total oxygen. The remaining oxygen at lower temperatures
is mainly in water vapor, which condenses into a cloud layer around 200–300 K,
depending on details of the P–T structure. It is interesting to note that the P–T
profile for the hot Jupiter (at 1 AU) in Figure 8.6 coincides closely with the
condensation curve for water ice (at solar metallicity) over an extended range.
Such planets would be prime candidates to show spectral signatures of water ice,
whereas in planets like Jupiter, water clouds are buried too deep in the atmosphere
to be visible.
The atmospheric P–T profiles of the more strongly heated planets at 0.04 and
0.06 AU cross the Mg-silicate stability curves several times. This implies different
possible locations within the atmosphere to trap Mg-silicate cloud material from
cooler atmospheric regions. However, the atmosphere of the planet at 0.06 AU
should be free of Si- and Mg-bearing gases at temperatures below ∼1600 K and an
enstatite/forsterite cloud layer is expected between ∼1600 and ∼2000 K. The planet
at 0.04 AU has a more complicated Mg-silicate cloud distribution, and the behavior
of condensates in such strongly heated planet with thermal inversions is not fully
understood; see, e.g., the discussion by Fortney et al. [8] on TiO and removal of
Ti-bearing condensates.
One can look at all elements in the periodic table and check which gases and
condensates are formed at a given temperature and pressure to sort out conditions
where important opacity sources (e.g., atomic lines, TiO, VO, CaH, CO, H2 O,
CH4 , FeH, CrH, . . .) appear, what their abundance are, and when these gases
are removed by formation of other gases and condensates so that their spectral
signatures become lost. This evaluation for all stable elements was done for the
Jovian and Saturnian atmospheric pressure and temperature profiles [47]. It is
an impractical approach to model the chemistry for all of the newly discovered
exoplanets this way. However, these results, listed in abbreviated form in Table 8.2
for the 20 most abundant reactive elements, should be useful to give an indication
of possible major gases and condensates. Where appropriate, differences in the
major gas abundances expected for hot and cooler objects are indicated.
8.4.2
Condensate Clouds
Thermodynamic equilibrium and kinetic calculations are useful to find which ele-
ments and chemical species are present as functions of T, P, and metallicity. The
computation of gas-phase equilibria is straightforward, but condensate formation
in gas-giant planets must consider that initially formed condensates sediment into
cloud layers and do not react with gases at higher (cooler) altitudes. The chemical
condensate treatment requires the distinction of two possible end-member set-
tings: condensate formation in a strongly gravitationally bound atmosphere and
condensate formation in a low-gravity environment such as in protoplanetary disks.
Figure 8.7 illustrates the differences in condensates for some elements. In a bound
atmosphere, condensates forming directly from the gas at high temperatures (‘‘pri-
mary condensates’’) settle due to the influence of gravity and form cloud layers.
These primary condensates do not react with the cooler gas at altitudes above the
condensate clouds so they are out of equilibrium with the gas in the overlying
atmosphere. In contrast, condensates in protoplanetary disks can remain dispersed
in the gas and react with it to form secondary condensates during cooling. For
example, in both cases, Fe-metal is the first Fe-bearing condensate. However, in
planetary atmospheres, Fe3 P and FeS cannot form because Fe is in a deep cloud.
0 0
“Proto planetary disk condensates” “Cloudy atmosphere condensates”
NH4SH
H2O
H2O NH4H2PO4
500 500
iter
FeS
T condensates (K)
T condensates (K)
KCl
Ju p
S
Na in Na 2
50%
1000 Na 2S U 1000
Na in 6A
50% ar 0.0
p
felds
tite
1 AU
Ensta
Fe3P
1500 s 1500
xide tal
Ti-o me es
te l Ca– eri
te Fe i-oxid
stati rite eta rst T
En rste
m
Fo a–
Fo Fe C
t h ite 0.0
4A
es
nebula
or oxid s U
Solar
2000 An – Al- xide 2000
0.
Ca Al-o
Ca-
03
AU
3 2 1 0 −1 −2 −3 −4 3 2 1 0 −1 −2 −3 −4
(a) Log (total pressure (bar)) (b) Log (total pressure (bar))
Figure 8.7 (a) Some of the condensates condensates settle into cloud layers. There
expected in protoplanetary disks. Primary, are no secondary Fe-bearing condensates
high T condensates (e.g., Fe-metal) can con- and different types of condensates appear
tinue to react with gas to form secondary at low T. Note that high T and high P are
condensates (e.g., FeS) at lower tempera- toward the bottom left in the graphs.
tures. (b) Condensates that appear if primary
Then P condenses mainly into NH4 H2 PO4 , and S condenses into sulfides such as
Na2 S, MnS, and mainly into NH4 SH.
The condensate cloud layer concept (often called rainout) was initially developed
for the Jovian planets [51, 52], and it has been found that only models including
cloud sedimentation are capable of matching observed brown dwarf color diagrams
and spectra. Hence what works for Jupiter and brown dwarfs should work for
exoplanets with masses in between.
The cloud masses that will be available depend on the availability of the
constituent elements in the condensates. Table 8.2 shows that the most abundant
rock-forming elements Mg, Si, and Fe produce massive clouds of forsterite, and
enstatite, and liquid iron, respectively. The other less-abundant rock-forming
elements such as Ca, Al, Ti, Mn, Cr, and the alkalis will produce less-massive
clouds. In addition to the more refractory (high thermal stability) condensates,
O, S, N, and C form abundant condensates at temperatures below ∼200 K. Giant
planets with low-temperature upper atmospheres can have massive icy cloud
layers of water (ice or liquid), NH4 SH (s), NH3 ice, and CH4 ice. However, they
will also have the full sequence of deeper seated cloud layers containing the more
refractory elements in the interior.
Once a condensate becomes stable, the abundance of the constituent gases
exponentially decreases with decreasing temperature in the atmosphere. With
Table 8.2 Condensates and gases in solar-composition giant planets.
Abundance Condensate type Major gases

Si = 106 atoms
H 2.88E+10 Minor fraction in H2 O H2 , (H2 O, CH4 )

O 15 500 000 Oxides, silicates, water H2 O; CO high T
C 7 080 000 CH4 CH4 low T, CO high T
N 2 090 000 NH4 SH, NH3 NH3 low T, N2 high T
Mg 1 020 000 Mg2 SiO4 , MgSiO3 Mg(OH)2 , MgOH, MgH, Mg
Si 1 000 000 Mg2 SiO4 , MgSiO3 SiH4 , SiO, SiS
Fe 838 000 Fe-alloy Fe, Fe(OH)2 , FeH
S 445 000 NH4 SH; minor MnS, ZnS, Na2 S H2 S
Al 84 100 Al2 O3 , Ca–Al-oxides Al2 O, AlH, AlOH, HAlO2
Ca 62 900 Ca–Al-oxides, Ca-titanates Ca, Ca(OH)2 , CaOH, CaH, CaS
Na 57 500 Na2 S Na, NaH, NaCl, NaOH
Ni 47 800 NiFe-alloy Ni, NiH, NiS
Cr 12 900 Cr-metal, Cr2 O3 Cr, CrH, CrS, CrO
Mn 9170 MnS Mn, MnH, MnS
P 8060 NH4 H2 (PO4 )3 , minor Cu3 P PH3 , PH2 , P2 , PO
Cl 5240 NaCl, KCl, NH4 Cl HCl, NaCl, KCl
K 3690 KCl K, KCl, KOH
Ti 2420 Ca-titanates, e.g., CaTiO3 TiO, TiO2 , Ti
Co 2320 FeNi alloy Co, CoH
Zn 1230 ZnS Zn, ZnH, ZnS
F 841 NaF, KF, NH4F HF
Cu 527 Cu3 P Cu, CuH
V 288 V-oxides into Ca–Ti and Ca–Al- VO, VO2 , V
oxides
Ge 121 Ge, minor GeTe GeH4 , GeS, GeSe, GeTe
Se 65.8 PbSe, other selenides H2 Se
Li 55.5 LiF, Li2 S Li, LiF, LiCl, LiOH, LiH
Ga 36.0 GaS GaOH, Ga2 S, GaS
Sc 34.2 Sc2 O3 ScO
Sr 23.6 SrS Sr(OH)2 , SrOH, Sr
B 17.3 H3 BO3 H3 BO3 , NaBO2 , KBO2 , HBO2
Br 11.3 NaBr, KBr, RbBr, NH4 Br HBr, alkali bromides
Zr 11.2 ZrO2 ZrO, ZrO2 , ZrS
Rb 6.57 RbCl, RbBr Rb, RbF, RbCl
the removal of the gases, their opacity also disappears. Therefore, the cloud-
layer structure in the atmosphere becomes traceable through the changes in
gas chemistry. The expected spectral features caused by the removal of gases
into condensate clouds in giant planets should be similar to that known for the
hot-to-cool brown dwarf sequence (see the discussion in [53]).
Extremely heated gas-giant planets will have very strong TiO and VO bands in the
optical, and H2 O, CO, and FeH bands in the near-infrared. Near ∼1800–2200 K,
refractory ceramics such as corundum (Al2 O3 ) or Ca-aluminates (such as hibonite,

CaAl12 O19 ) and Ca-titanates (such as perovskite, CaTiO3 ) become stable. This leads
to a gradual disappearance of TiO and VO bands, since VO follows TiO by forming
solid solutions with the Ti-bearing condensates. Similarly, Ca- and Al-bearing
gases are removed. The exact nature of the refractory condensate is sensitive to
the total pressure [54]. The first Ca and Al condensate at high total pressures in
the deeper interiors of cool Jupiter-like giant planets is melilite (a solid solution
of gehlenite Ca2 Al2 SiO7 and akermanite Ca2 MgSi2 O7 ). However, it is hibonite
below ∼1 bar, and Al2 O3 below 0.01 bar, which is more relevant to Pegasi-type
planets or extremely heated exoplanets. The initial Al-bearing condensate in Pegasi
planets is Ca-poor, which affects Ca gas abundances. The Ca (and Al) chemistry
determines which Ca-titanate forms, which, in turn, is important for the TiO gas
abundances.
Since the temperature of gas-giant planets in close orbits is regulated by the
irradiation from their host stars, one can expect that the closest planets (around the
same type of star) are the most likely to be heated above temperatures necessary for
the condensation of ceramics. In that case, TiO and VO remain, and their opacity
determines whether a Pegasi-type planet can develop a stratospheric temperature
inversion [8, 37, 55]. The photometric data and spectra of HD 209458b [56, 57]
seem to be consistent with the presence of TiO and VO and a low-pressure inverted
temperature structure in this exoplanet.
Molecular bands of CrH and FeH become stronger in cooler atmospheres
but decrease in strength when Fe-metal and, depending on total pressure, Cr
or Cr2 O3 condensation occurs. This can remove all Fe and Cr gases, respec-
tively, around ∼1200–1500 K. The removal of Fe into a cloud layer is also the
reason why H2 S remains in the atmospheric gas at lower temperatures. If Fe
cloud settling did not occur, H2 S would be completely absent from the upper
atmospheres because formation of FeS (troilite) from Fe-metal grains with H2 S
gas starting at ∼700 K consumes all H2 S gas (solar Fe/S ∼ 2). However, forma-
tion of secondary FeS and removal of H2 S at 600–700 K is at odds with the
Galileo probe observations on Jupiter. The Galileo entry probe mass spectrome-
ter (GPMS) detected H2 S at approximately three times the solar S/H2 ratio in
Jupiter’s atmosphere [58] which is only possible if H2 S is not removed. Because
H2 S is observed in the Jovian and Saturnian tropospheres at altitudes below
the NH4 SH cloud condensation level, the condensate formation models must
call for the depletion of Fe-metal into a cloud deep in the Jovian and Saturnian
atmospheres.
In the temperature range of Fe-metal condensation, Mg-silicate (forsterite and
enstatite) cloud formation is expected as well. This will remove all Mg (e.g., Mg,
MgH, MgOH) and Si (e.g., Si, SiO, SiS, SiH4 ) gases from the cooler atmosphere.
As for sulfur, Jupiter and Saturn are test cases for the cloud-layer approach. The
absence of silane (SiH4 ) and the presence of germane (GeH4 ) in Jupiter and Saturn
is due to depletion of refractory Si, but not of volatile Ge, by condensate formation
deep in their atmospheres [47]. Silicon is much more abundant than Ge in a solar-
composition gas (atomic Si/Ge ∼ 8300) but there are only observational upper
limits for SiH4 /H2 ∼ 1 × 10−9 (1 ppb) on Jupiter and Saturn. For comparison, the
protosolar Si/H2 molar ratio is 8.23 × 10−5 , which is ∼82 000 times larger than
the observational upper limit on the silane abundance. In contrast, the observed
GeH4 /H2 is ∼0.7 ppb on Jupiter and ∼0.4 ppb on Saturn [1]. These values are
closer to the protosolar Ge/H2 molar ratio of 9.9 ppb and the difference arises
because not all Ge in Jupiter and Saturn is present as GeH4 [47]. The presence
of Mg-silicate clouds should be testable by searching for SiH4 (which has strong
IR bands) in the atmospheres of methane-rich planets, or for SiO and SiS in hot,
CO-rich planets. These gases should be absent or depleted above the Mg2 SiO4 and
MgSiO3 clouds if the cloud condensation models are correct.
Enstatite is the lowest temperature condensate of the major element (Al, Ca, Ti,
Mg, Si, Fe, Ni) condensates. Thus, all of the major rock-forming elements are out
of the gas above the enstatite cloud. However, lines of monatomic alkali metals
K, Na, Cs, and Rb should remain prominent and persist to lower temperatures
that correspond to hotter methane-rich objects (Figure 8.5). Monatomic K gas is
observed in T dwarfs such as Gl 229B and Gl 570D [2, 42]. Monatomic Na, at less
than the solar Na abundance, is inferred for the exoplanet HD 209458b [59]. The
monatomic gases are dominant until they gradually convert to NaCl, KCl, and other
gases such as oxides, hydroxide, and hydrides, before being removed as sulfide
and halide condensates. Sodium is much less abundant than S, so all Na can be
sequestered into a Na2 S cloud but only 6.5% of all sulfur would be lost, which does
not lead to a significant loss in total H2 S.
The presence of Na and K in cooler objects requires that refractory, rock-forming
elements such as Al, Ca, and Si are depleted by condensate cloud formation deep in
the atmosphere. Otherwise, Na- and K-gases would condense into silicate minerals
such as XAlSi3 O8 (albite (X = Na) and orthoclase (X = K)) at high temperatures,
and Na and K gas would be depleted from the observable atmosphere. However,
if Al, Ca, and Si are in deep condensate clouds, XAlSi3 O8 condensates cannot
form and Na and K stay in the gas. Only model spectra that consider Al, Ca,
and Si removal into deep cloud layers show good agreement with the higher
observed K I line strengths for Gl 570D and Gl 229B [42]. The chemistry of Li
and the other alkalis Rb, and Cs is similar. Instead of being removed from the
gas by reaction with high-temperature feldspar, they only form their own sulfide
(e.g., Na2 S) and halide condensates at temperatures 500 K lower or more than,
e.g., Mg-silicates.
The switch from CO to methane should cause an onset of absorption in the
1.6, 2.2, and 3.3-µm methane bands; similar to what was expected and found
in the L dwarf spectral series [60]. With methane becoming more abundant, the
infrared colors will change because of the strong IR methane absorption bands,
as is known for the change in L and T dwarfs. In cool methane-dominated
objects, the alkali lines have vanished. Methane, water, and ammonia charac-
terize the cooler objects until condensation of liquid water or water ice and
solid NH3 leaves only methane behind in an otherwise He- and H2 -rich atmo-
sphere. Removal of methane from the atmosphere into clouds requires very low
temperatures (<50 K).
8.4.3
The Effects of Varying the C/O Ratio on Gas-Giant Planet Chemistry
8.4.3.1 Gas Chemistry Variations through Changes in C/O Ratio

Variations in the C/O ratio affect the chemical composition of the atmosphere
because the CO, methane, and water equilibria are changed. This affects the
interior evolution and spectra of gas-giant planets. For example, Fortney et al. [36,
61] show that changes in metallicity and C/O ratios from solar values strongly
influence the emergent spectra of brown dwarfs and hot exoplanets. Changes
in C/O mainly affect abundances of gases containing C and/or O. To a first
approximation, abundances of gases such as NH3 and N2 are not affected much
because the most abundant N-bearing gases contain neither C nor O. However,
the increasing abundances of gases such as HCN that contain C are of minor
importance in a solar-composition gas.
Figure 8.8 shows the gas abundances as a function of C/O ratio at 1100 K and
0.01 bar to illustrate trends. At this temperature, C is mainly distributed between
CO and CH4 , and O between CO and H2 O in a solar-composition gas. The gas
abundances are calculated for C/O ratios from solar to unity by increasing the
C abundance (closed symbols in Figure 8.8) or by decreasing O from the system
(open symbols). The partial pressures of CO, CH4 , C2 H2 , HCN, and CN increase,
whereas that of H2 O declines over orders of magnitude at otherwise constant
conditions. The increase in the CO fugacity is relatively modest (a factor of 2),
which reflects increasing the C/O from 0.5 to unity and that most C is in the very
10−5 CO 10−6
N2
10−7
10−6
10−8
10−7
10−9 NH3
10−8
10−10
Partial pressure
CH4 H2O
10−9 10−11
HCN
−10
10
10−12
10−11
C2H2 10−19
10−12
10−13 10−20
CN
−14
10
1100 K −21 1100 K
10
10−15 Ptot = 0.01 bar Ptot = 0.01 bar
10−16 10−22
0.5 0.6 0.7 0.8 0.9 1.0 0.5 0.6 0.7 0.8 0.9 1.0
(a) Bulk C /O ratio (b) Bulk C /O ratio
Figure 8.8 Variation of gas chemistry of carbon and oxygen

(a) and nitrogen (b) as a function of C/O ratio at 1100 K
and 0.01 bar. The C/O ratio was increased from solar by
increasing the total C abundance (closed symbols) or de-
creasing the total oxygen (open symbols). Note the break in
the concentration scale for nitrogen gases.
stable CO molecule. Therefore, changes in CO opacity, important for the 4.5-µm

spectral window, decreases. A particularly strong increase is for C2 H2 and HCN,
which can become new opacity sources in the infrared. On the other hand, the
H2 O fugacity drops by several orders of magnitude if the C/O ratio is increased
from solar to unity because with a relatively high C-content, more O is in CO and
less O is available for water and other O-bearing molecules.
Figure 8.8 also illustrates that it matters how the C/O ratio is altered. An increase
in C/O ratio by O removal yields a larger effect than adding C because the
absolute amount of O-bearing molecules decreases. This is the reason why the
H2 O abundance decreases much more strongly when O is removed. Since water
is a major opacity source in gas-giant exoplanets, substantial decreases in water
abundances should change the spectra in the 8-µm region of these objects [36].
Together with observations of CO, this may constrain planetary C/O ratios, which
has possible implications for planetary formation scenarios. If a gas-giant planet
has a different C/O ratio from its host star, the variation of the C/O ratio must
have to do with planet formation itself, if chemical fractionations (such as water-ice
condensation) in the planet can be ruled out.
8.4.3.2 Possible Scenarios to Alter C/O Ratios during Planetary Formation

There are two possibilities to change the C/O ratio in a giant planet from the
value of ∼0.5 in an otherwise solar-composition gas. For an increase in C/O in
an accretion disk, the plausible mechanisms are to increase the carbon abundance
or decrease the oxygen abundance. For a decrease in C/O, an increase in O is the
most plausible possibility.
Assuming that gas-giant planets form by core accretion, an increase in the C
abundance occurs if abundant interstellar or locally produced organics are incorpo-
rated into the rocky protocore of a giant planet. As discussed earlier, tarry organics
could exist in larger amounts in planetary accretion disks beyond an organic con-
densation/evaporation front (a tarline). This hypothesis is analogous to assuming
that planets like Jupiter grew from a protocore composed of rock and water ice,
at an orbital distance near or beyond the water-ice condensation/evaporation front
(the snowline). Organic substances may be similarly suitable to increase the mass
density of the disk to aid rapid formation of a core and by ‘‘gluing’’ rocky and
carbonaceous substances (see, [62]).
The compositional outcome of these two cases is quite different. If a planet forms
with excess water ice, the resulting C/O ratio is lower than the host star’s C/O ratio.
On the other hand, accretion and growth from a carbonaceous and rocky core leads
to a C/O ratio above the host star’s ratio.
Organics constituting a tarline require higher temperatures to evaporate than
water ice; hence, there can be regions in accretion disks that are rich in carbonaceous
material, but poor in water ice. The stability of the organics would be enhanced if
water was cold-trapped in a more distant region from the central star, because then
water vapor, which is a potential oxidant for the carbonaceous material, is also less
abundant.
8.5 Outlook 183
The cold-trapping of water ice also provides a possibility to raise the C/O
ratio because it removes oxygen in the planet-forming region. Radial diffusive
redistribution of gas and water cold-trapping at and beyond the snowline would
dehumidify the region inward of the snowline [63]. This leads to increased C/O
ratios between the star and the snowline. Even if half of all O is in CO as expected
from the solar C/O = 0.5, removal of other oxygen moves the C/O ratio toward
unity. If planets then acquire water-depleted gas, they may have a higher C/O ratio.
However, if gas and water ice are not physically fractionated before accretion to a
planet, the bulk C/O ratio remains solar.
Thus, a planet with a higher C/O ratio than its host star formed either in a
water-depleted region or in a region that was enriched in carbonaceous material.
Since fast growth of a massive rocky protocore may be required for making gas-
giant planets, the first option, accretion in a water-depleted region could cause a
problem because the surface mass density of condensed rocky mass alone is too
low for fast growth. In this case, the presence of abundant organic materials could
help to increase the mass density. This limits the origin of protocore formation for
gas-giant planets with higher C/O ratios to regions in protoplanetary disks below
∼400 K, where carbonaceous material remains stable.
The other extreme is a gas-giant planet with a C/O ratio smaller than in its
primary star, which is formed by preferential accretion of water ice. Water ice
trapping from the inner planetary system must lead to ice pileup at the snowline
and beyond, which would be an ideal source. Such a pileup of water is probably
recorded by the water-rich planets Uranus and Neptune. If these planets formed
near or beyond the snowline with abundant water ice, water enhancements, and
smaller C/O ratios would result [64]. This formation scenario applies to disk regions
with temperatures below ∼180 K to stabilize water ice.
8.5
Outlook
The chemistry of exoplanets is somewhat difficult to summarize since chemical

information is mainly available in indirect form through density measurements,
and direct spectroscopic observations so far provide limited results. Most known
exoplanets are gas-giant planets, and most modeling is thus available for gas-
giant exoplanets. This will also undoubtedly change in the coming years, when
the upcoming missions begin to discover terrestrial-like planets. However, even
without too many observations of chemistry yet, exoplanet chemistry is predictable:
it is constrained by the abundances of the elements and the physical conditions
in planet formation environments as well as in the planet itself. The observable
chemistry of brown dwarfs, the exoplanets’ larger ‘‘cousins,’’ also has provided
valuable guidelines to what can be expected in gas-giant exoplanet spectra, and
the known chemistry of the terrestrial planets in the solar system will be equally
valuable in interpreting the observations of exoplanets to come.
Acknowledgments
The author thanks Jonathan Fortney and Mark Marley for kindly providing model
atmospheres, and Bruce Fegley for discussions. This work was supported in part
by NSF grants AST 07-07377 and AST 08-07356 and NASA grant NNG06GC26G.
References
1 Lodders, K. and Fegley, B. (1998) The W.B. McKinnon), Cambridge University

Planetary Scientist’s Companion, Oxford Press, Cambridge, pp. 35.
University Press, New York, pp. 384. 12 Guillot, T., Santos, N.C., Pont, F.,
2 Burrows, A., Hubbard, W.B., Lunine, Iro, N., Melo, C., and Ribas, I. (2006)
J.I., Marley, M.S., and Saumon, D. Astronomy and Astrophysics, 453, L21.
(2000) in Protostars and Planets IV 13 Sudarsky, D., Burrows, A., and Hubeny,
(eds. V. Mannings, A.P. Boss, and S.S I. (2003) The Astrophysical Journal, 588,
Russell), University of Arizona Press, 1121.
Tucson, pp. 1339. 14 Fortney, J.J., Marley, M.S., and Barnes,
3 Burrows, A., Sudarsky, D., and Lunine, J.W. (2007) The Astrophysical Journal,
J.I. (2003) The Astrophysical Journal, 596, 659, 1661.
587–596. 15 Seager, S., Kuchner, M.,
4 Burrows, A., Sudarsky, D., and Hubeny, Hier-Majumder, C.A., and Militzer,
I. (2004) The Astrophysical Journal, 609, B. (2007) The Astrophysical Journal, 669,
407. 1279.
5 Charbonneau, D., Brown, T.M., 16 Kleine, T., Muencker, C., Mezger, K.,
Burrows, A., and Laughlin, G. (2007) and Palme, H. (2002) Nature, 418,
in Protostars and Planets V (eds. B.
952–955.
Reipurth, D. Jewitt, and K. Keil), Uni-
17 Gonzalez, G. (2003) Reviews of Modern
versity of Arizona Press, Tucson, pp.
Physics, 75, 101.
701.
18 Boss, A.P. (1997) Science, 276, 1836.
6 Chabrier, G., Baraffe, I., Selsis, F.,
19 Boss, A.P. (2001) The Astrophysical
Barman, T.S., Hennebelle, P., and
Journal, 563, 367.
Alibert, Y. (2007) in Protostars & Plan-
20 Pollack, J.B., Hubickyj, O.,
ets V (eds. B. Reipurth, D. Jewitt, and
Bodenheimer, P., Lissauer, J.J., Podolak,
K. Keil), University of Arizona Press,
Tucson, pp. 623. M., and Greenzweig, Y. (1996) Icarus,
7 Fortney, J.J., Marley, M.S., Hubickyj, 124, 62.
O., Bodenheimer, P., and Lissauer, J.J. 21 Hubickyj, O., Bodenheimer, P., and
(2004) Astronomische Nachrichten, 326, Lissauer, J.J. (2005) Icarus, 179, 415.
925. 22 Lissauer, J.J. and Stevenson, D.J. (2007)
8 Fortney, J.J., Lodders, K., Marley, M.S., in Protostars and Planets V (eds. B.
and Freedman, R.S. (2008a) The Astro- Reipurth, D. Jewitt, and K. Keil), Uni-
physical Journal, 678, 1419. versity of Arizona Press, Tucson, pp.
9 Fortney, J.J., Marley, M.S., Saumon, D., 591.
and Lodders, K. (2008b) The Astrophysi- 23 Chambers, J.E. (2003) in Treatise on
cal Journal, 683, 1104. Geochemistry, vol. 1 (ed. A.M. Davis),
10 Guillot, T. (2008) Physica Scripta 2008, Elsevier, Amsterdam, pp. 461.
130, 014023. 24 Saumon, D. and Guillot, T. (2004) The
11 Guillot, T., Stevenson, D.J., Hubbard, Astrophysical Journal, 609, 1170.
W.B., and Saumon, D. (2004) in Jupiter: 25 Grossman, L. and Larimer, J.W. (1974)
The Planet, Satellites, and Magnetosphere Review of Geophysics and Space Research,
(eds. F. Bagenal, T.E. Dowling, and 12, 71.
References 185
26 Lewis, J.S. (1974) Annual Review of 43 Burrows, A., Hubbard, W.B., Lunine,
Physical Chemistry, 24, 339. J.I., and Liebert, J. (2001) Reviews of
27 Lodders, K. (2003) The Astrophysical Modern Physics, 73, 719.
Journal, 591, 1220. 44 Gustafsson, B., Edvardsson, B.,
28 Fegley, B. (2000) Space Science Reviews, Eriksson, K., Jørgensen, U.G.,
92, 177. Nordlund, A.A., and Plez, B. (2008)
29 Kouchi, A., Kudo, T., Nakano, H., Astronomy and Astrophysics, 486, 951.
Arakawa, M., Watanabe, N., Sirono, 45 Knutson, J.A., Charbonneau, D.,
S.I., Higa, M., and Maeno, N. (2002) Burrows, A., O’Donovan, F.T., and
The Astrophysical Journal, 667, L121. Mandushev, G. (2009) The Astrophysical
30 Fegley, B. and Prinn, R.G. (1989) in Journal, 691, 866.
Formation and Evolution of Plane- 46 Lodders, K. and Fegley, B. (2002) Icarus,
tary Systems (eds. H. Weaver and L. 155, 393.
Danly), Cambridge University Press, 47 Fegley, B. and Lodders, K. (1994) Icarus,
Cambridge, p. 171. 110, 117.
31 Fegley, B. (1993) in The Chemistry 48 Fegley, B. and Lodders, K. (1996) The
of Life’s Origins, vol. 416 (eds. Astrophysical Journal, 472, L37.
M. Greenberg, C.X. Mendoza-Gomez, 49 Noll, K.S., Geballe, T.R., and Marley,
and V. Pirronello), NATO Advanced M.S. (1997) The Astrophysical Journal,
Science Institute, Series C, Kluwer 489, L87–L91.
Academic Publishers, Dordrecht, p. 75. 50 Saumon, D., Marley, M.S.,
32 Helled, R. and Schubert, G. (2008) Leggett, S.K., Geballe, T.R.,
Icarus, 198, 156. Stephens, D., Golimowski, D.A.,
33 Wuchterl, I.G., Guillot, T., and Lissauer, Cushing, M.C., Fan, X., Rayner, J.T.,
J.J. (2000) in Protostars and Planets IV Lodders, K., and Freedman, R.S. (2007)
(eds. V. Mannings, A.P. Boss and S.S. The Astrophysical Journal, 656, 1136.
Russell), University of Arizona Press, 51 Lewis, J.S. (1969) Icarus, 10, 393.
Tucson, p. 1081. 52 Barshay, S.S. and Lewis, J.S. (1978)
34 Stevenson, D.J. (1985) Icarus, 62, 4. Icarus, 33, 593.
35 Burrows, A., Sudarsky, D., and 53 Lodders, K. and Fegley, B. (2006) in
Hubeny, I. (2006) The Astrophysical Astrophysics Update 2 (ed. J.W. Mason),
Journal, 650, 1140. Springer Verlag, Heidelberg, p. 1.
36 Fortney, J.J., Saumon, D., Marley, M.S., 54 Lodders, K. (2002) The Astrophysical
Lodders, K., and Freedman, R.S. (2006) Journal, 577, 974.
The Astrophysical Journal, 642, 495. 55 Burrows, A., Hubeny, I., Budaj, J.,
37 Hubeny, I., Burrows, A., and Sudarsky, Knutson, H.A., and Charbonneau, D.
D. (2003) The Astrophysical Journal, 594, (2007) The Astrophysical Journal, 668,
1011. L171.
38 Marley, M.S., Seager, S., Saumon, D., 56 Knutson, H.A., Charbonneau, D., Allen,
Lodders, K., Ackerman, A.S., Freedman, L.E., Burrows, A., and Megeath, S.T.
R., and Fan, X. (2002) The Astrophysical (2008) The Astrophysical Journal, 673,
Journal, 568, 335–342. 526.
39 Kirkpatrick, J.D. (2005) Annual Review 57 Désert, J.M., Vidal-Madjar, A.,
of Astronomy and Astrophysics, 43, 195. Lecavelier Des Etangs, A., Sing, D.,
40 Marley, M.S., Saumon, D., Guillot, Ehrenreich, D., Hébrard, G., and
T., Freedman, R.S., Hubbard, W.B., Fegley, B. (1993) in The Chemistry of
Burrows, A., and Lunine, J.J. (1996) Life’s Origins, NATO Advanced Sci-
Science, 272, 1919. ence Institute, Series C, vol. 416 (eds.
41 Marley, M.S. and Saumon, D. (2008) M. Greenberg, C.X. Mendoza-Gomez,
The Astrophysical Journal, 689, 1327. and V. Pirronello), Kluwer Academic
42 Geballe, T.R., Saumon, D., Leggett, Publishers, Dordrecht, p. 75.
S.K., Knapp, G.R., Marley, M.S., and 58 Niemann, H.B., Atreya, S.K., Carignan,
Lodders, K. (2001) The Astrophysical G.R., Donahue, T.M., Haberman, J.A.,
Journal, 556, 373–379. Harpold, D.N., Hartle, R.E.,
Hunten, D.M., Kasprzak, W.T., 61 Fortney, J.J., Marley, M.S., Lodders, K.,
Mahaffy, P.R., Owen, T.C., and Saumon, D., and Freedman, R. (2005)
Way, S.H. (1998) Journal of Geophysical The Astrophysical Journal, 627, L69.
Research, 103, 22831–22845. 62 Lodders, K. (2004) The Astrophysical
59 Charbonneau, D., Brown, T.M., Noyes, Journal, 611, 587.
R.W., and Gilliland, R.L. (2002) The 63 Stevenson, D.J. and Lunine, J.I. (1988)
Astrophysical Journal, 568, 377. Icarus, 75, 146.
60 Noll, K.S., Geballe, T.R., Leggett, S.K., 64 Lodders, K. and Fegley, B. (1994) Icarus,
and Marley, M.S. (2000) The Astrophysi- 112, 368.
cal Journal, 541, L75–L78.

Exoplanet Chemistry

Uploaded by

Copyright:

Available Formats

Exoplanet Chemistry

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Exoplanet Chemistry

Uploaded by

Copyright:

Available Formats

157

The current state of observations on the chemical nature of exoplanets is comparable

Formation and Evolution of Exoplanets. Edited by Rory Barnes

The elemental composition of a gas-giant planet compared to the composition of

Approximate rock and ice masses

assuming non-equilibrium at low T

Table 8.1 Typical minerals made of more abundant elements.

Mineral group End-member mineral Ideal formula

Olivine (Mg,Fe)2 SiO4 Forsterite Mg2 SiO4

Another possibility is that interstellar low-temperature organic substances re-

Jupiter dT Gl570D 800 K,1000 m s−2

Figure 8.3 Pressure and temperature structures in Jupiter,

Figure 8.4 Pressure–Temperature structures of Jupiter-mass

500 CH4 (g)

Figure 8.5 Carbon chemistry in low-mass objects is domi-

oxides in cooling gases, depending on the availability of oxygen. In cooler planets,

Figure 8.6 Oxygen chemistry in low-mass objects is dom-

between CO and CH4 through the net reaction:

its maximum abundance. The boundary is crossed if a Jupiter-like planet moves to

2000 An – Al- xide 2000

Table 8.2 Condensates and gases in solar-composition giant planets.

Abundance Condensate type Major gases

H 2.88E+10 Minor fraction in H2 O H2 , (H2 O, CH4 )

refractory ceramics such as corundum (Al2 O3 ) or Ca-aluminates (such as hibonite,

8.4.3.1 Gas Chemistry Variations through Changes in C/O Ratio

Figure 8.8 Variation of gas chemistry of carbon and oxygen

stable CO molecule. Therefore, changes in CO opacity, important for the 4.5-µm

8.4.3.2 Possible Scenarios to Alter C/O Ratios during Planetary Formation

The chemistry of exoplanets is somewhat difﬁcult to summarize since chemical

1 Lodders, K. and Fegley, B. (1998) The W.B. McKinnon), Cambridge University

You might also like