Download The formation and habitability of terrestrial planets in the presence of

Icarus 177 (2005) 256–263
www.elsevier.com/locate/icarus
The formation and habitability of terrestrial planets in the presence of
close-in giant planets
Sean N. Raymond a,∗ , Thomas Quinn a , Jonathan I. Lunine b
a Department of Astronomy, University of Washington, Box 351580, Seattle, WA 98195, USA
b Lunar and Planetary Laboratory, The University of Arizona, Tucson, AZ 85287, USA
Received 29 July 2004; revised 7 March 2005
Available online 28 April 2005
Abstract
‘Hot jupiters,’ giant planets with orbits very close to their parent stars, are thought to form farther away and migrate inward via interactions
with a massive gas disk. If a giant planet forms and migrates quickly, the planetesimal population has time to re-generate in the lifetime of
the disk and terrestrial planets may form [P.J. Armitage, A reduced efficiency of terrestrial planet formation following giant planet migration,
Astrophys. J. 582 (2003) L47–L50]. We present results of simulations of terrestrial planet formation in the presence of hot/warm jupiters,
broadly defined as having orbital radii 0.5 AU. We show that terrestrial planets similar to those in the Solar System can form around stars
with hot/warm jupiters, and can have water contents equal to or higher than the Earth’s. For small orbital radii of hot jupiters (e.g., 0.15,
0.25 AU) potentially habitable planets can form, but for semi-major axes of 0.5 AU or greater their formation is suppressed. We show that
the presence of an outer giant planet such as Jupiter does not enhance the water content of the terrestrial planets, but rather decreases their
formation and water delivery timescales. We speculate that asteroid belts may exist interior to the terrestrial planets in systems with close-in
giant planets.
 2005 Elsevier Inc. All rights reserved.
Keywords: Planetary formation; Extrasolar planets; Cosmochemistry; Exobiology
1. Introduction
Roughly one-third of the giant planets discovered to date
outside the Solar System have orbits within 0.5 astronomical
units (AU) of their central stars.1 These close-in giant planets, also known as “hot/warm jupiters,”2 are thought to have
formed farther out and migrated inward via gravitational
torques with a massive gas disk (Lin et al., 1996). If this mi* Corresponding author. Fax: +1 (206) 685 0403.
E-mail addresses: [email protected] (S.N. Raymond),
[email protected] (T. Quinn), [email protected]
(J.I. Lunine).
1 See, e.g., http://www.exoplanets.org.
2 The term “hot jupiter” is generally reserved for planets inside 0.1 AU.
The population of giant planets that we are studying have semimajor axes
0.15 AU a 0.5 AU are better described as “warm jupiters.” For simplicity, we refer to all close-in giant planets as hot jupiters.
0019-1035/$ – see front matter  2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.icarus.2005.03.008
gration occurs within roughly the first million years (Myr)
of the disk lifetime,3 the planetesimal population (the building blocks of terrestrial planets) has time to replenish after
being destroyed during migration. However, if migration occurs later, planetesimals are destroyed without enough time
to re-form, making it impossible for sizable terrestrial planets to form (Armitage, 2003).
In some cases, already-formed terrestrial planets may survive the migration of a giant planet through the terrestrial
zone. The survival rate of terrestrial planets depends on the
rate of migration (faster migration means higher survival
rate) and ranges from 15–40% (Mandell and Sigurdsson,
2003). However, most of the surviving planets are on highly
3 The given value of 1 Myr is a average value, which depends on the dust
diffusion and planetesimal formation parameters—see Armitage (2003) for
more detail.
Forming terrestrial planets with close-in giant planets
eccentric orbits with large semimajor axes. Only a very small
fraction (1–4%) of terrestrial planets survive the migration
event without significant alteration to their orbits.
Recent results show that giant planets can form on very
short timescales via gravitational collapse (Boss, 1997;
Mayer et al., 2002; Rice et al., 2003). New simulations of
the standard, core-accretion scenario (Pollack et al., 1996)
including turbulence (Rice and Armitage, 2003) and migration during formation (Alibert et al., 2004) have shown that
giant planets can form via this mechanism in 1 Myr or less,
in agreement with the observed, 1–10 million year lifetime
of circumstellar disks (Briceño et al., 2001). Observations
of the ∼1 Myr old star Coku Tau 4 with the Spitzer Space
Telescope have revealed an absence of dust inside 10 AU
(Forrest et al., 2004). One explanation is the presence of a
planet orbiting this very young star (Quillen et al., 2004). If
correct, this would be observational evidence for fast giant
planet formation.
The timescale for the inward migration of a giant planet
depends on the mass of the planet and the mass and viscosity of the gaseous disk, and is typically less than 105 years
for Saturn- to Jupiter-mass planets (Fig. 6 from D’Angelo
et al., 2003; Ward, 1997; Tanaka et al., 2002). Migration
begins immediately after, even during, the formation of the
giant planet (Lufkin et al., 2004). The mechanism by which
migration stops is not well understood, and may involve interactions with magnetic fields (Terquem, 2003) or an evacuated region in the inner disk (Kuchner and Lecar, 2002;
Matsuyama et al., 2003). Many planets may in fact migrate
all the way into the star (Nelson et al., 2000).
Lineweaver (2001) and Lineweaver et al. (2004) implicitly assumed that terrestrial planets cannot coexist with hot
jupiters. However, based on the above arguments, we expect that terrestrial planets can form in a standard, bottom-up
fashion in the presence of a hot jupiter, and survive for the
lifetime of the parent star. If migration takes place late in the
lifetime of the gaseous disk, there remains a small chance
that an already-formed terrestrial planet could survive on a
similar orbit (Mandell and Sigurdsson, 2003).
The character and composition of a system of terrestrial
planets is strongly affected by the total amount of solid material in the protoplanetary disk (Wetherill, 1996; Chambers
and Cassen, 2002; Raymond et al., 2004) and the presence
of one or more giant planets (Chambers and Cassen, 2002;
Levison and Agnor, 2003; Raymond et al., 2004). One theory for how the Earth acquired its water is as follows. During
formation, the proto-Earth accreted bodies which formed in
the outer asteroid belt (between roughly 2.5 and 4 AU), past
the “snow line,” where water could exist as ice in the low
pressure protoplanetary disk (Morbidelli et al., 2000). This
theory is not universally accepted, but does explain both the
total quantity of water on Earth and its isotopic signature
(see Morbidelli et al., 2000, for a discussion).
The habitable zone around a star is defined as the annulus in which the temperature is right for liquid water to exist
on the surface of an Earth-like planet, and is roughly 0.95–
257
1.37 AU in our Solar System (Kasting et al., 1993). A habitable planet not only needs to reside in its star’s habitable
zone, it also needs a substantial water content. The source of
water, however, lies much farther out in the protoplanetary
disk, past the snow line. The formation of a habitable planet
therefore requires significant radial stirring of protoplanets
with different compositions (see Raymond et al., 2004, for a
discussion).
Here we present results of dynamical simulations of terrestrial planet formation in the presence of a close-in giant
planet, both with and without an exterior giant planet. We
include close-in giant planets (hereafter referred to as “hot
jupiters” at all orbital radii) with orbital radii of 0.15, 0.25,
and 0.5 AU, and in some cases outer giant planets at 5.2 AU.
Section 2 outlines our initial conditions and numerical methods. Section 3 presents our results, which are discussed in
Section 4.
2. Method
Our initial conditions are designed to approximate the
state of the protoplanetary disk after the formation of planetary embryos via oligarchic growth (e.g., Kokubo and Ida,
2000), at which time the vast majority of the solid mass is
in the form of ∼1000 km-sized “planetary embryos,” which
are decoupled from the gas disk. A simulation begins with
a disk of 120 to 180 protoplanets which reflects the minimum mass solar nebula model (Hayashi, 1981). Planetary
embryos4 have masses between 0.01 and 0.15 Earth masses,
physical densities of 3 g cm−3 , and are placed from the hot
jupiter out to 5.2 AU. These are randomly spaced by 3–6 mutual Hill radii assuming the surface density of solids scales
with heliocentric distance r as r −1.5 . The surface density is
normalized to 10 g cm−2 at 1 AU, with each disk of embryos containing 6–7 Earth masses of material inside 5 AU.
The discovered giant planets are found to preferentially orbit stars with metallicities higher than the Sun’s (Laws et
al., 2003), indicating that they likely contain a large amount
of solid material with which to build terrestrial planets. Our
chosen value for the surface density is therefore quite low,
and accounts for some depletion during hot jupiter migration. All hot jupiters have masses of 0.5 Jupiter masses and
all outer giant planets are 1 Jupiter mass. All giant planets
are on circular orbits.
We assign protoplanets an initial distribution of water
content which reflects the distribution in chondritic meteorites (see Fig. 2 from Raymond et al., 2004), such that
the inner bodies are dry, past 2 AU planetary embryos contain 0.1% water, and past 2.5 AU embryos contain 5%
4 Unlike Raymond et al. (2004), we only include planetary embryos in
our simulations (no planetesimals). This is to keep the simulations simple,
as computational limitations do not allow us to include a realistic number of
planetesimals. We are in the process of running new simulations to explore
the dynamical effects of smaller bodies on the terrestrial accretion process.
258
S.N. Raymond et al. / Icarus 177 (2005) 256–263
water. Their iron distribution is interpolated between the
content of the planets and chondritic asteroid classes, ignoring the planet Mercury. These range from 0.40 (40%
iron by mass) at 0.2 AU to 0.15 at 5 AU. Each embryo
is given a small initial inclination (1◦ ) and eccentricity
(0.02).
Each simulation is evolved for at least 200 million years
using a hybrid integrator called Mercury (Chambers, 1999),
which evolves the orbits of all bodies and keeps track of
collisions. The hybrid scheme in Mercury uses a symplectic algorithm to evolve orbits of bodies unless they are involved in a close encounter, in which case it switches to
a Bulirsch–Stoer method. Collisions are treated as inelastic
mergers which conserve mass, water and iron content. The
time step in each simulation is chosen to be less than 1/20th
of the orbital period of the innermost body in the simulation, and ranges from 1 day for a hot jupiter at 0.15 AU
to 6 days for a hot jupiter at 0.5 AU. Each simulation
conserved energy to at least one part in 105 , and took between three weeks and three months to complete on a desktop PC.
3. Results
Fig. 1 shows the time evolution of one simulation which
formed a planet in the habitable zone, with a hot jupiter at
0.25 AU and an outer giant planet at 5.2 AU (not shown).
Planetary embryos are dynamically excited by the giant
planets and their mutual gravitation, increasing their eccentricities and causing their orbits to cross. This results in both
accretional impacts and close encounters with giant planets,
which eject roughly half of the terrestrial bodies. By the end
of a simulation only a few terrestrial planets remain. In this
case four terrestrial bodies have formed including two planets inside 2 AU, one of which lies in the habitable zone at
1.06 AU with 1.68 times the mass of Earth with water content higher than the Earth’s value of about 10−3 (which is
uncertain—see Raymond et al., 2004, for a discussion). As
our simulations do not account for water loss during impacts, water content values are upper limits. However, we
do not simulate the secondary delivery of volatiles from farther out in the disk (“late veneer”) which would increase
the water content although likely not by more than 10% if
Fig. 1. Six snapshots in time from a simulation (sim 6—see Fig. 2) with two giant planets (not shown): a 0.5 Jupiter-mass hot jupiter at 0.25 AU and a
Jupiter-mass planet at 5.2 AU, both on circular, coplanar orbits. Each panel plots the eccentricity and semi-major axis of each surviving body in the simulation.
The size of a body is proportional to its mass(1/3) , and the dark region in the center represents the size of its iron core, on the same scale. The color corresponds
to the water mass fraction, which ranges initially from 10−5 to 0.05.
Forming terrestrial planets with close-in giant planets
259
Fig. 2. Final configurations of twelve simulations, with the Solar System (including the largest asteroid, Ceres) shown for scale. The gray circles represent the
giant planets in each simulation and are not to the same scale as the terrestrial bodies. All hot jupiters have masses of 0.5 Jupiter masses and all outer giant
planets are 1 Jupiter mass. The eccentricity of each body is represented by its excursion in heliocentric distance over an orbit. The x axis is on a logarithmic
scale such that a given separation corresponds to a fixed ratio of orbital periods, shown in the scale bar on the top left. The dashed vertical lines represent the
boundaries of the habitable zone (Kasting et al., 1993). Simulations 23 and 24 were run for 200 Myr, simulations 9 and 10 for 500 Myr, and simulations 13
and 14 for 800+ Myr. A comparison shows the long accretion timescales in the outer terrestrial region. Note the presence of protoplanets in 1:1 resonance
with a giant planet in some cases.
it proceeds as in our Solar System (Morbidelli et al., 2000;
Levison et al., 2001).
Fig. 2 shows the final state of twelve simulations (out
of twenty-four), with the Solar System included for scale.
260
S.N. Raymond et al. / Icarus 177 (2005) 256–263
Grey circles represent the positions of giant planets in each
simulation and are not on the same scale as the terrestrial
bodies. The eccentricity of each body is shown beneath it
by its radial excursion over the course of one orbit. Terrestrial planets do form in the habitable zone in the presence
of a hot jupiter, often with substantial water contents. The
possibility of a potentially habitable planet forming depends
on the location of the hot jupiter. In most cases, no planets more massive than 0.2 Earth masses form within a factor
of 3 in period to the hot jupiter, roughly a factor of two in
semi-major axis.
If a planet forms in the habitable zone with a hot jupiter
at 0.5 AU, it is the innermost terrestrial planet and tends
to be relatively small and dry. This is because the delivery
of a water-rich body from past the snow line to a forming
planet is not a smooth inward diffusion, but rather a stochastic random walk. During this process, the inbound body’s
perihelion distance is likely to come inside the aphelion of
the forming planet. In many cases, it can venture entirely inside the orbit of the forming planet, where it encounters a
strong gravitational perturbation from the hot jupiter. This
usually results in the destruction of the water-rich body via
ejection from the system. Only a fraction of inward-moving,
water-rich bodies can be accreted by the innermost terrestrial
planet due to the dynamical presence of the hot jupiter.
Water-rich planets form readily in the habitable zone with
a hot jupiter at 0.15 or 0.25 AU. As in previous simulations
(e.g., Chambers, 2001; Raymond et al., 2004), the eccentricities of the planets formed are significantly larger than
the time-averaged values for Venus, Earth, and Mars. This is
likely due to the lack of dissipative forces such as dynamical
friction and gas drag in the simulations.
In our Solar System planet formation was suppressed in
the asteroid belt by the gravitational effects of Jupiter. This
is seen in Fig. 2, as no terrestrial planets form within a factor of 3–4 in period to a hot jupiter or an outer gas giant. We
speculate that as this gap is filled with the remnants of terrestrial bodies in our Solar System, systems with hot jupiters
may contain asteroid belts5 interior to the terrestrial planets.
The resolution of current simulations is too low to test this
hypothesis.
Simulations with no giant planet exterior to the terrestrial
region form planets of substantial mass past 2–3 AU on time
scales of hundreds of Myr. Indeed, the systems in Fig. 2 with
no outer giant planet have not yet finished accreting. Simulations 23 and 24 were run for 200 Myr, simulations 9 and
10 for 500 Myr, and simulations 13 and 14 for 800+ Myr.
A comparison between the outer regions of these demonstrates the long formation timescales. An outer gas giant
clears the outer terrestrial region (past 2 AU) of protoplanets quickly, and the water content of the terrestrial planets
5 Note that we use the term ‘asteroid belt’ to refer to any region in which
no planet of significant mass has formed due to the gravitational effects of a
giant planet. This annulus usually extends to a factor of 3–5 in orbital period
from the giant planet (within about a factor of 2–3 in semimajor axis).
that form is about 50% less than in the absence of an outer
giant planet. In all cases, terrestrial planets in the habitable
zone form more quickly in the presence of an outer giant
planet, because more protoplanets are scattered onto orbits
which cross the inner terrestrial region (inside 1.5 AU) by
both the outer giant and, consequently, the protoplanets it
has scattered inward (a process called “secular conduction”
by Levison and Agnor, 2003). A hot jupiter does not significantly accelerate the formation process, as accretion already
proceeds very quickly in the inner disk where dynamical
timescales are short.
The terrestrial planets are delivered water at earlier times
in the presence of an outer gas giant. This suggests that an
outer giant planet’s net effect is to clear material from the
outer terrestrial region (thus forming an asteroid belt) and to
accelerate terrestrial planet formation by exciting the eccentricities of protoplanets (Levison and Agnor, 2003). Its role
in delivering water to the terrestrial planets is not a vital one
in terms of quantity. In fact, an outer giant planet may be
detrimental to the volatile delivery process.
In a typical simulation with a Jupiter-mass outer giant
planet, roughly 0.1 Earth masses (M⊕ ) of water-rich material is scattered inward from past 2.5 AU and ends up in
a massive (> 0.2 M⊕ ) terrestrial planet. This is only about
3% of the 2.5–3 M⊕ in embryos between 2.5 AU and the
giant planet. Simulations with no outer giant planet deliver
0.15–0.25 M⊕ of water-rich material to terrestrial planets of
significant mass inside 1.5 AU.
An outer giant planet quickly excites the eccentricities of
nearby protoplanets, with two possible outcomes: these bodies will either undergo a close approach with the giant planet
and likely be ejected from the system, or will have their orbits removed from the region, inward towards the habitable
zone, possibly delivering volatile-rich material to a forming
planet. If the outer giant planet’s eccentricity is non-zero,
then the probability of a water-rich protoplanet being ejected
from the system increases significantly, thereby decreasing
the water content of the terrestrial planets (Chambers and
Cassen, 2002; Raymond et al., 2004).
In the absence of an outer giant planet, little water-rich
material is ever ejected from the system. However, with no
strong perturber it is difficult for the eccentricities of waterrich bodies to grow large enough to interact with the inner
terrestrial region. The delivery of a water-rich embryo to a
planet in the habitable zone therefore takes a long time, and
its inward orbital change is driven by many close encounters
with other protoplanets. But since water-rich bodies are not
being ejected, and the formation timescale of massive terrestrial planets is very long in the outer terrestrial region, more
water-rich protoplanets are delivered to the terrestrial planets
in the absence of a Jupiter-mass outer giant planet.
The optimal scenario for water delivery requires a balance between the high ejection rate of an outer giant planet
and the difficulty in eccentricity pumping in the absence of
an outer giant. This may be accomplished with a less massive giant planet and a higher outer surface density of solids.
Forming terrestrial planets with close-in giant planets
Indeed, Chambers (2003) found that a lower-mass Jupiter
and Saturn result in more water-rich terrestrial planets, and
Raymond et al. (2004) showed that a higher outer density of
material forms more water-rich planets. Some of the most
water-rich planets formed in Raymond et al. (2004) were
in simulations with a less massive outer giant (1/3MJupiter )
and a high outer surface density (simulations 24 and 25 from
Raymond et al. (2004)—see Table 2 and Fig. 8).
The amount of material ejected from the system is a function of the number and configuration of giant planets. An
outer giant planet ejects approximately one half of the total
terrestrial mass in the system, while a hot jupiter can remove
up to one-third of the total mass. If we naïvely assume that
a (zero-eccentricity) gas giant planet ejects all solid material within a factor of three in orbital period (roughly a factor
of two in semimajor axis both inward and outward), then we
can calculate the total amount of material ejected as a function of orbital radius r. In the case of a minimum-mass solar
nebula-like model with surface density profile Σ ∝ r −3/2 ,
the total mass ejected scales with the orbital radius of the
√
giant planet, rg , as Mejec ∝ rg . In the more general case,
with a surface density profile Σ ∝ r −α , and assuming that
all material within a factor F in orbital period of the giant is ejected, Mejec ∝ [(rg F 2/3 )2−α − (rg F −2/3 )2−α ]. The
value of F depends on the mass and eccentricity of the giant planet, which may also scale with rg and α. Indeed, the
“Hill scaling” law predicts that, for an r −3/2 surface density
profile, the mass of a giant planet increases with heliocentric distance r as r −3/4 (Lissauer, 1995). This scaling is also
reflected in our initial distribution of planetary embryos [see
Section 2.1 of Raymond et al. (2004) for details].
An outer giant planet therefore ejects significantly more
material from the system than an inner one, simply because
there is more mass in its “clearing zone,” the region from
which it ejects or destroys most terrestrial bodies. In the case
of a disk of solid material between 0.25 and 5.2 AU, an outer
giant planet at 5.2 AU ejects 40% of the total solid material and a hot jupiter at 0.25 AU ejects 12%. A hot jupiter
at 0.5 AU ejects 27%, including material both inside and
outside its orbit. These values are less than the amount of
material ejected in our simulations, implying that the clearing zone of a giant planet extends beyond a factor of three
in relative orbital period. In simulations with a hot jupiter at
0.5 AU and an outer giant planet, the final terrestrial planets
comprise only one quarter of the initial mass. These planets are systematically depleted in iron because the iron-rich
protoplanets closest to the Sun have been largely removed
by the hot jupiter.
In simulations with a hot jupiter at 0.15 or 0.25 AU, the
iron content of the formed terrestrial planets is very close
to that of Venus, Earth, and Mars. This is not surprising,
since they form largely from embryos in the terrestrial region
whose iron contents were initially linked to those of Venus,
Earth, and Mars. Unlike simulations with a hot jupiter at
0.5 AU, planets formed in these simulations have no mechanism to significantly enhance or deplete the local iron con-
261
tent. However, the discovered extra-solar planets past 0.1 AU
have eccentricities ranging from 0.03 to 0.93. An eccentric
hot jupiter ventures closer to the terrestrial region in its orbit than a circular one, and the secular perturbations felt by
nearby planetary embryos are much stronger (Levison and
Agnor, 2003; Raymond et al., 2004). A highly eccentric hot
jupiter increases the relative velocities of embryos to such a
point that collisions are likely to result in disruption rather
than accretion (seen in simulations of Epsilon Eridani by
Thébault et al., 2002), thereby suppressing the formation of
terrestrial planets. A moderately eccentric hot jupiter has a
wider clearing zone than a circular hot jupiter and therefore
ejects a larger fraction of iron-rich embryos, and forms irondepleted terrestrial planets.
We ran three simulations for one billion years or more to
test the long term stability of terrestrial planets in the presence of hot jupiters. The short dynamical timescales in the
inner disk result in a fast clearing of unstable objects, so a
longer integration produces no change. The asteroid belt is
slowly cleared by an outer giant planet, but all planets which
are well separated from a giant planet (by a factor of 3–4 or
more in orbital period) are stable for long timescales.
We ran two simulations under the assumption that the hot
jupiter’s migration took place later in the lifetime of the protoplanetary disk, thereby depleting the planetesimal reservoir (Armitage, 2003). The surface density of solid material
was reduced by a factor of five to 2 g cm−2 at 1 AU in these
simulations, which included a hot jupiter at 0.25 AU and an
outer giant planet. After 200 Myr of evolution, these systems
formed no planets more massive than 0.16 Earth masses and
left a large number of small bodies in the terrestrial region,
reminiscent of a large asteroid belt.
4. Discussion
All the simulations presented here contain giant planets on circular orbits with fixed masses. The observed hot
jupiters inside 0.1 AU (51 Pegasi-type hot jupiters) tend to
have circular orbits due to tidal interactions with the central
star. More distant giant planets can have a large range in eccentricity and mass. The effects of these parameters can be
extrapolated using previous results. An eccentric outer giant planet preferentially ejects water-rich material from the
planetary system rather than scattering it inward, which results in dry terrestrial planets (Chambers and Cassen, 2002;
Raymond et al., 2004) with large eccentricities, located far
from the giant planet. Similarly, an eccentric hot jupiter
preferentially ejects iron-rich material. A more massive giant planet or a higher surface density of solid material results in a smaller number of more massive terrestrial planets
(Wetherill, 1996; Chambers and Cassen, 2002; Raymond et
al., 2004). We apply this to known planetary system 55 Cancri (Marcy et al., 2002), which contains two hot jupiters at
0.115 and 0.241 AU, an exterior giant planet at 5.9 AU,
and a recently discovered, roughly Neptune-mass planet at
262
S.N. Raymond et al. / Icarus 177 (2005) 256–263
0.038 AU (McArthur et al., 2004). The two hot jupiters are
close to the 3:1 resonance and the less massive, outer hot
jupiter has an eccentricity of 0.33. The outer giant planet’s eccentricity is 0.16 and its minimum mass (M sin(i)) is
about four Jupiter masses. Applying our previous arguments,
we expect a small number of terrestrial planets to form in 55
Cancri far away from the hot jupiters as well as from the
outer giant. The high eccentricities should strongly deplete
the solid material, resulting in low-mass planets. Simulations (not including the newly discovered, ∼ Neptune-mass
planet) have shown this to be the case, with at most two terrestrial planets forming in 55 Cancri, with masses no greater
than 0.6 Earth masses (Raymond and Barnes, 2005).
We have argued that terrestrial planets can form in the
presence of hot jupiters. We have shown that potentially
habitable planets with orbits in the habitable zone and substantial water contents can form in such conditions. The
obliquity of such planets would likely be stable over long
timescales (Atobe et al., 2004). We hypothesize that asteroid belts may exist between the terrestrial planets and
hot jupiters. Based on this and previous work it is possible to predict the character of the terrestrial planets around a
star, from observables such as the orbit and mass of a giant planet and the metallicity of the star. Our predictions
will be testable in the near future with upcoming space missions such as Kepler6 and COROT,7 that will detect giant and
(hopefully) terrestrial planets around other stars. Longerterm missions like Terrestrial Planet Finder8 and Darwin9
hope to obtain spectra of terrestrial planets and search for
signs of water and life. Although these missions will surely
focus on finding planetary systems like our own, we suggest
that stars with hot jupiters may be a good place to look for
extra-solar terrestrial planets.
This result can also be applied to constrain the location of the Galactic Habitable Zone (Gonzalez et al., 2001;
Lineweaver et al., 2004). This is defined as the region in the
galaxy in which various factors conspire to make the area
suitable for life (e.g., the average metallicity of stars, the
rate of supernovae, time needed for life to evolve). In particular, Lineweaver et al. (2004) assume (from Lineweaver,
2001) that the probability of a star to host a habitable planet
drops precipitously if its metallicity is higher than 0.2–0.3
dex (solar metallicity is defined to be 0.0). This is based on
the fact that higher metallicity stars are more likely to have
hot jupiters (Laws et al., 2003), and the assumption that any
migration event would preclude the formation of terrestrial
planets in the system. Our result, that potentially habitable
planets can exist around stars with hot jupiters, effectively
widens the Galactic Habitable Zone to include regions at
small galactocentric distances and recent times (“too metal
rich” regions in Figs. 3 and 4 of Lineweaver et al., 2004).
6 http://www.kepler.arc.nasa.gov.
7 http://www.astrsp-mrs.fr/projets/corot.
8 http://planetquest.jpl.nasa.gov/TPF.
9 http://ast.star.rl.ac.uk/darwin.
Acknowledgments
We thank referees Shigeru Ida and Luke Dones for insightful comments. S.R. thanks Lucio Mayer, Chris Laws,
and Graeme Lufkin for helpful discussions, and congratulates the Boston Red Sox on winning the World Series. This work was funded by NASA Astrobiology Institute and NASA Planetary Atmospheres. These simulations were run under Condor, which is publicly available
at http://www.cs.wisc.edu/condor.
References
Alibert, Y., Mordasini, C., Benz, W., 2004. Migration and giant planet formation. Astron. Astrophys. 417, L25–L28.
Armitage, P.J., 2003. A reduced efficiency of terrestrial planet formation
following giant planet migration. Astrophys. J. 582, L47–L50.
Atobe, K., Ida, S., Ito, T., 2004. Obliquity variations of terrestrial planets in
habitable zones. Icarus 168, 223–236.
Boss, A.P., 1997. Giant planet formation by gravitational instability. Science 276, 1836.
Briceño, C., Vivas, A.K., Calvet, N., Hartmann, L., Pacheco, R., Herrera,
D., Romero, L., Berlind, P., Sanchez, G., Snyder, J.A., Andrews, P.,
2001. The CIDA-QUEST large-scale survey of Orion OB1: Evidence
for rapid disk dissipation in a dispersed stellar population. Science 291,
93–97.
Chambers, J.E., 1999. A hybrid symplectic integrator that permits close
encounters between massive bodies. Mon. Not. R. Astron. Soc. 304,
793–799.
Chambers, J.E., 2001. Making more terrestrial planets. Icarus 152, 205–
224.
Chambers, J.E., Cassen, P., 2002. The effects of nebula surface density profile and giant-planet eccentricities on planetary accretion in the inner
Solar System. Meteorit. Planet. Sci. 37, 1523–1540.
Chambers, J.E., 2003. The formation of life-sustaining planets in extrasolar
planetary systems. Lunar and Planetary Institute Conference Abstracts,
34, 2000.
D’Angelo, G., Kley, W., Henning, T., 2003. Orbital migration and mass
accretion of protoplanets in three-dimensional global computations with
nested grids. Astrophys. J. 586, 540–561.
Forrest, W.J., 20 colleagues, 2004. Mid-infrared spectroscopy of disks
around classical T Tauri stars. Astrophys. J. Suppl. Ser. 154, 443–447.
Gonzalez, G., Brownlee, D., Ward, P., 2001. The Galactic habitable zone:
Galactic chemical evolution. Icarus 152, 185–200.
Hayashi, C., 1981. Structure of the solar nebula, growth and decay of magnetic fields and effects of magnetic and turbulent viscosities on the
nebula. Prog. Theor. Phys. Suppl. 70, 35–53.
Kasting, J.F., Whitmire, D.P., Reynolds, R.T., 1993. Habitable zones around
main sequence stars. Icarus 101, 108–128.
Kokubo, E., Ida, S., 2000. Formation of protoplanets from planetesimals in
the solar nebula. Icarus 143, 15–27.
Kuchner, M.J., Lecar, M., 2002. Halting planet migration in the evacuated
centers of protoplanetary disks. Astrophys. J. 574, L87–L89.
Laws, C., Gonzalez, G., Walker, K.M., Tyagi, S., Dodsworth, J., Snider,
K., Suntzeff, N.B., 2003. Parent stars of extrasolar planets. VII. New
abundance analyses of 30 systems. Astron. J. 125, 2664–2677.
Levison, H.F., Dones, L., Chapman, C.R., Stern, S.A., Duncan, M.J.,
Zahnle, K., 2001. Could the lunar “Late Heavy Bombardment” have
been triggered by the formation of Uranus and Neptune? Icarus 151,
286–306.
Forming terrestrial planets with close-in giant planets
Levison, H.F., Agnor, C., 2003. The role of giant planets in terrestrial planet
formation. Astron. J. 125, 2692–2713.
Lin, D.N.C., Bodenheimer, P., Richardson, D.C., 1996. Orbital migration of
the planetary companion of 51 Pegasi to its present location. Nature 380,
606–607.
Lineweaver, C.H., 2001. An estimate of the age distribution of terrestrial
planets in the universe: Quantifying metallicity as a selection effect.
Icarus 151, 307–313.
Lineweaver, C.H., Fenner, Y., Gibson, B.K., 2004. The galactic habitable
zone and the age distribution of complex life in the Milky Way. Science 303, 59–62.
Lissauer, J.J., 1995. Urey prize lecture: On the diversity of plausible planetary systems. Icarus 114, 217–236.
Lufkin, G., Quinn, T., Wadsley, J., Stadel, J., Governato, F., 2004. Simulations of gaseous disc-embedded planet interaction. Mon. Not. R. Astron.
Soc. 347, 421–429.
Mandell, A.M., Sigurdsson, S., 2003. Survival of terrestrial planets in the
presence of giant planet migration. Astrophys. J. 591, L111–L114.
Marcy, G.W., Butler, R.P., Fischer, D.A., Laughlin, G., Vogt, S.S., Henry,
G.W., Pourbaix, D., 2002. A planet at 5 AU around 55 Cancri. Astrophys. J. 581, 1375–1388.
Matsuyama, I., Johnstone, D., Murray, N., 2003. Halting planet migration
by photoevaporation from the central source. Astrophys. J. 585, L143–
L146.
Mayer, L., Quinn, T., Wadsley, J., Stadel, J., 2002. Formation of giant planets by fragmentation of protoplanetary disks. Science 298, 1756–1759.
McArthur, B.E., 12 colleagues, 2004. Detection of a Neptune-mass planet
in the ρ 1 Cancri system using the Hobby–Eberly telescope. Astrophys.
J. 614, L81–L84.
Morbidelli, A., Chambers, J., Lunine, J.I., Petit, J.M., Robert, F., Valsecchi,
G.B., Cyr, K.E., 2000. Source regions and timescales for the delivery of
water on Earth. Meteorit. Planet. Sci. 35, 1309–1320.
263
Nelson, R.P., Papaloizou, J.C.B., Masset, F., Kley, W., 2000. The migration
and growth of protoplanets in protostellar discs. Mon. Not. R. Astron.
Soc. 318, 18–36.
Pollack, J.B., Hubickyj, O., Bodenheimer, P., Lissauer, J.J., Podolak, M.,
Greenzweig, Y., 1996. Formation of the giant planets by concurrent accretion of solids and gas. Icarus 124, 62–85.
Quillen, A.C., Blackman, E.G., Frank, A., Varnière, P., 2004. On the planet
and the disk of COKU TAURI/4. Astrophys. J. 612, L137–L140.
Raymond, S.N., Quinn, T., Lunine, J.I., 2004. Making other earths: Dynamical simulations of terrestrial planet formation and water delivery.
Icarus 168, 1–17.
Raymond, S.N., Barnes, R., 2005. Predicting planets in known extra-solar
planetary systems III: Forming terrestrial planets. Astrophys. J. Submitted for publication. astro-ph/0404212.
Rice, W.K.M., Armitage, P.J., 2003. On the formation timescale and core
masses of gas giant planets. Astrophys. J. 598, L55–L58.
Rice, W.K.M., Armitage, P.J., Bonnell, I.A., Bate, M.R., Jeffers, S.V., Vine,
S.G., 2003. Substellar companions and isolated planetary-mass objects
from protostellar disc fragmentation. Mon. Not. R. Astron. Soc. 346,
L36–L40.
Tanaka, H., Takeuchi, T., Ward, W.R., 2002. Three-dimensional interaction between a planet and an isothermal gaseous disk. I. Corotation and
Lindblad torques and planet migration. Astrophys. J. 565, 1257–1274.
Terquem, C.E.J.M.L.J., 2003. Stopping inward planetary migration by a
toroidal magnetic field. Mon. Not. R. Astron. Soc. 341, 1157–1173.
Thébault, P., Marzari, F., Scholl, H., 2002. Terrestrial planet formation in
exoplanetary systems with a giant planet on an external orbit. Astron.
Astrophys. 384, 594–602.
Ward, W.R., 1997. Protoplanet migration by nebula tides. Icarus 126, 261–
281.
Wetherill, G.W., 1996. The formation and habitability of extra-solar planets.
Icarus 119, 219–238.