month goes by without a new revelation - Specialized Research AI at MIT

month goes by without a new revelation
of some sort or another. Already well
over a hundred extrasolar planet candi-
dates have been announced, and the
pace of discovery promises to quicken
as additional ground-based search pro-
grams swing into action. Meanwhile a
number of powerful space-based obser-
vatories speci½cally designed to search
for and characterize planets as small in
mass as Earth are being planned for the
next two decades.

These advances have fueled, in turn,
furious theoretical work on the forma-
tion and migration mechanisms of plan-
ets inside and outside our solar system.
All the extrasolar planets discovered to
date appear to be gas giant planets, simi-
lar to Jupiter and Saturn, and the theory
of gas giant planet formation is in flux as
a result.

The amazingly short period of the ½rst

extrasolar, Jupiter-mass planet discov-
ered brought the possibility of planet
migration to the attention of theorists.
The extrasolar planet orbits its host
star, 51 Pegasi, in a mere 4.23 days, com-
pared to Jupiter’s leisurely 11.9-year orbit
around the Sun–and, according to Kep-
ler’s third law, it orbits 51 Pegasi about a
hundred times closer than Jupiter orbits
the Sun. The formation of a Jupiter-mass
planet so close to its parent star appears
to be dif½cult, if not impossible, so theo-
rists such as myself have hypothesized
that some giant planets must form at
larger distances and then migrate in-
ward to their ½nal orbital distances.

There are two very different ideas for
how gas giant planets might form. Most
astronomers favor the conventional the-
ory of core accretion, where a solid core
forms ½rst and then accretes a gaseous
envelope. In 1997 I proposed a very dif-
ferent mechanism, based on the hypoth-
esis that a protoplanetary disk was likely
to pass through a phase of marginal

Alan P. Boss

on the search for
extrasolar planets

Astrobiology, the search for life’s ori-
gins and its existence elsewhere in the
universe, used to seem like a visionary
dream. But in recent years, it has become
a true science, thanks in part to new de-
velopments in the search for Earth-like
planets outside our own solar system.
A new era in scienti½c discovery was
initiated in 1995 with the announcement,
by the Swiss astronomers Michel Mayor
and Didier Queloz, of the ½rst Jupiter-
mass companion to a Sun-like star.
Progress on the detection of planets out-
side our solar system has occurred at a
breathtaking pace ever since; scarcely a

Alan P. Boss, a Fellow of the American Academy
since 2003, has been a research staff member at
the Carnegie Institution of Washington since
1981. Internationally known for his theoretical
work on the formation of stellar and planetary
systems, he advises nasaon the search for extra-
solar planets. He is the author of “Looking for
Earths: The Race to Find New Solar Systems”
(1998).

116

Dædalus Summer 2004

D
o
w
n
o
a
d
e
d

f
r
o
m
h

t
t

:
/
/

d
i
r
e
c
t
.

i
t
.

e
d
u
d
a
e
d
a
r
t
i
c
e
–
p
d

f
/

1
3
3
3
1
1
6
1
8
3
1
5
6
7
0
0
1
1
5
2
6
0
4
1
5
0
4
5
9
7
p
d

b
y
g
u
e
s
t

o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3

The
search for
extrasolar
planets

gravitational instability, where random
density perturbations could lead rapidly
to the growth of self-gravitating clumps
of gas and dust in the disk that might
survive to form giant planets. The two
competing theories have very different
implications for the formation environ-
ment of the solar system, and hence for
the frequency of planetary systems simi-
lar to our own, for the number of habit-
able planets that may orbit nearby stars,
and for our chances of ½nding another
Earth-like planet outside our own solar
system.

The theory of core accretion supposes

the collisional accumulation of solid
bodies, the process that is universally
accepted as the formation mechanism
of the terrestrial planets. Collisional
accumulation simply means that when
a swarm of particles is in orbit around
a star, random collisions between these
particles may lead to their sticking to-
gether to form a larger body, if they hit
each other gently enough. This accumu-
lation is thought to proceed through suc-
cessively larger bodies–starting with
submicron-sized dust grains, inherited
from previous generations of stars, that
stick together by intermolecular forces
when they collide; to meter-sized boul-
ders; on up to kilometer-sized planetesi-
mals (comets), where self-gravity begins
to become important; to lunar-sized
planetary embryos; and ½nally to Earth-
sized planets. The core accretion theory
envisions this process as occurring in
both the inner and outer regions of a
star’s planet-forming, rotationally flat-
tened disk of gas and dust.

In the innermost region of the disk
out of which our sun and solar system
formed, collisional accumulation leads
over the course of several tens of mil-
lions of years to the formation of Earth-
sized rocky planets. In the outer region
of the disk, beyond the asteroid belt, the

same process is thought to lead to the
formation of solid cores, equal in mass
to roughly ten Earths, which may then
acquire massive gaseous envelopes from
the disk gas. These cores are said to form
through runaway accretion, where the
largest bodies grow the fastest because
their self-gravity increases their colli-
sional cross-sections; two bodies that
would otherwise miss each other will
hit because their mutual gravitational
attraction deflects their orbits toward
each other.

At an early phase, the accretion of disk

gas falling onto the protoplanet causes
an atmosphere to form on its growing
core. As the protoplanet continues to
grow by accreting disk gas and solid
planetesimals, its atmosphere eventually
can no longer be supported in hydrostat-
ic equilibrium, and so it contracts. This
contraction culminates in a brief period
of atmospheric collapse, during which
the protoplanet gains the bulk of its ½nal
mass.

At Jupiter’s distance from the Sun, the

timescale for the entire core accretion
process is estimated to be on the order of
several million years or more. Estimates
of the lifetimes of planet-forming disks
range from a few million years in quies-
cent regions of star formation, like the
Taurus molecular cloud, to well under a
million years in regions where the most
massive stars form, such as the Orion
Nebula cluster.

If there were only one solar system to
explain, core accretion might be an at-
tractive theory, because there are proba-
bly some disks that last long enough for
this process to form gas giant planets.
But unless the timescale for core accre-
tion is signi½cantly shorter than the pre-
vailing estimates suggest, the theory
seems unable to account for the ob-
served abundance of gas giants else-
where.

Dædalus Summer 2004

117

D
o
w
n
o
a
d
e
d

f
r
o
m
h

t
t

:
/
/

d
i
r
e
c
t
.

i
t
.

e
d
u
d
a
e
d
a
r
t
i
c
e
–
p
d

f
/

1
3
3
3
1
1
6
1
8
3
1
5
6
7
0
0
1
1
5
2
6
0
4
1
5
0
4
5
9
7
p
d

b
y
g
u
e
s
t

o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3

Note by
Alan P.
Boss

Eliminating this timescale problem is
one of the main attractions of my alter-
native to the theory of core accretion–
the theory of disk instability. The disk
instability theory envisions a rapid pro-
cess somewhat the opposite of core ac-
cretion. In disk instability, a clump of
disk gas and dust forms ½rst, and then
the dust grains settle to the center of that
clump to form a solid core. This mecha-
nism requires a marginally gravitational-
ly unstable disk, a disk cool enough to
be on the verge of breaking up into self-
gravitating spiral arms and clumps. (The
inner regions of the disk gas rotate fast-
er than the outer regions, shearing the
growing clumps of gas into spiral arms.)
While early theoretical models of mar-
ginally gravitationally unstable disks
suggested that such disks would only
form spiral arms, more recent computer
models have indicated that self-gravitat-
ing clumps may form within the spiral
arms and survive their subsequent orbit-
al evolution to form gaseous protoplan-
ets. A planet-forming disk with a mass
about one tenth of the Sun’s (which is
the sort of disk mass that core accretion
models typically assume) will be mar-
ginally gravitationally unstable, provid-
ed that the disk temperatures in the gi-
ant planet region are on the order of
50º K or less. Observations of planet-
forming disks and of delicate molecular
species in long-period comets seem to
support such temperatures.

As a clump forms within the spiral
arms, the dust grains within it begin to
sediment down toward this center of the
burgeoning protoplanet. This process is
hastened by the coagulation of the dust
grains as they migrate to the center. As a
result, disk instability may be capable of
forming a self-gravitating protoplanet
within a time period as short as about a
thousand years. For a Jupiter-mass (318
Earth masses) protoplanet containing

the solar abundance (2 percent by mass)
of elements heavier than hydrogen or
helium, this central core could be as
massive as 6 Earth masses.

The disk instability mechanism re-
quires the presence of a strong flux of
ultraviolet (uv) light to explain the for-
mation of ice giant planets like Uranus
and Neptune. (Strong fluxes of uv light
occur in regions of high-mass star for-
mation, such as the Orion and Carina
Nebulas.) Intense uv light can heat up
and photoevaporate the disk gas outside
a critical radius from the parent star;
for a solar-mass star, the critical radius,
which depends on the mass of the par-
ent star, is roughly equivalent to Sat-
urn’s. Giant gaseous protoplanets that
form from clumps outside this critical
orbital radius–stripped by the uv light
of the bulk of their gaseous envelopes,
reduced down to their solid rock and ice
cores, with only thin veneers of remain-
ing gas–will be turned into ice giants.
Protoplanets inside this critical radius,
meanwhile, will be largely unaffected by
the uv light. This scenario explains the
bulk compositions of Jupiter, Uranus,
and Neptune, as well as Saturn’s reten-
tion of most of its once much larger gas-
eous envelope. The formation of habit-
able planets in the inner region of the
planet-forming disk would proceed
more or less unfazed by this searing
experience in the disk’s outer region.
Earth-like planets are thought to be
able to form with just about equal proba-
bility whether the gas giant planets form
quickly (as in disk instability) or slowly
(as in core accretion). In either case, the
collisional growth of Earth-sized planets
on orbits similar to Earth’s requires tens
of millions of years to run to comple-
tion, so the events in the ½rst few thou-
sand or million years are not necessarily
critical to the formation of such planets.
Furthermore, in either case, habitable

118

Dædalus Summer 2004

D
o
w
n
o
a
d
e
d

f
r
o
m
h

t
t

:
/
/

d
i
r
e
c
t
.

i
t
.

e
d
u
d
a
e
d
a
r
t
i
c
e
–
p
d

f
/

1
3
3
3
1
1
6
1
8
3
1
5
6
7
0
0
1
1
5
2
6
0
4
1
5
0
4
5
9
7
p
d

b
y
g
u
e
s
t

o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3

planets should be able to form along
with the gas giants. Gas giant planets are
important to have around, because they
shield the habitable planets from con-
stant bombardment by residual icy plan-
etesimals that might otherwise frustrate
the origin and evolution of life.

This helps explain the importance of
recent theoretical developments in the
emerging ½eld of astrobiology. Estimates
of the number of technological civiliza-
tions in our galaxy are commonly based
on the equation ½rst presented by Frank
Drake in 1961. Two of the many factors
in the Drake equation are fp and ne–the
fraction of stars with planetary systems
(presumably similar to the solar system,
the only known example in 1961) and the
number of habitable planets per plane-
tary system, respectively.

If my heretical theory of disk instabili-
ty is correct, then fp can be considerably
larger than conventional wisdom holds.
Conventional wisdom would seem to
limit the formation of solar-like plane-
tary systems to stars formed in regions
of low-mass star formation like Taurus.
Core accretion presumably could not
form gas giant planets in a region like
Orion, because the lifetimes of the Ori-
on disks are even shorter than those in
Taurus. But if the heretical approach is
correct, then solar-like planetary sys-
tems can form essentially everywhere
Sun-like stars form–even in Orion.
Roughly 90 percent of stars are thought
to form in regions of high-mass star for-
mation like Orion. This implies a differ-
ence in fp as large as about a factor of ten
between the orthodoxy and the heresy.
Knowing the prevalence of habitable
planets in our region of the galaxy is im-
portant for our search for other Earths. If
Earths are rare, telescopes built to detect
them will need to be designed different-
ly. nasa is now in the process of design-
ing several such telescopes, called the

The
search for
extrasolar
planets

Terrestrial Planet Finders, intended for
launch around 2015 and 2020. If all goes
well with this extremely ambitious, dif-
½cult project, in a little over a decade we
will know if we have any neighboring
planets that are capable of supporting
life–or, indeed, are actually supporting
life right now. I, for one, hope that we
heretics are right–that the prevalence
of life elsewhere in the universe could be
much greater than the conventional the-
ory predicts.

D
o
w
n
o
a
d
e
d

f
r
o
m
h

t
t

:
/
/

d
i
r
e
c
t
.

i
t
.

e
d
u
d
a
e
d
a
r
t
i
c
e
–
p
d

f
/

1
3
3
3
1
1
6
1
8
3
1
5
6
7
0
0
1
1
5
2
6
0
4
1
5
0
4
5
9
7
p
d

b
y
g
u
e
s
t

o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3

Dædalus Summer 2004

119
Download pdf