These lecture notes were prepared for Rutgers Physics 341/342: Principles of Astrophysics

by Prof. Chuck Keeton, and modied by Profs. Saurabh Jha and Eric Gawiser. All rights
reserved. c 2014
Lecture 26: Planet Formation
Now that we have developed a basic idea about how stars form, lets think about how planets
form around stars. The basic idea is that in order to conserve angular momentum some of
the matter settles into a disk around a forming star, rather than falling into the star itself.
Planets form inside this disk. We would like to get an idea of how big the disk is, and to
understand gross features of our Solar System like the fact that the inner planets are small
and rocky while the outer planets are big and gaseous. As we shall see, the classic ideas
about planet formation that seem to describe our own Solar System have been challenged
recently by the discovery of hundreds of planets orbiting other stars.
This discussion draws on Sections 18.2, and 23.2 of Carroll & Ostlie.
I. Simple model of disk size
Suppose the gas cloud that will collapse to form a star initially has size R
0
and is rotating
as a solid body with angular frequency
0
. As it collapses, most of the mass goes into the
central star, but most of the angular momentum goes to form a disk around the star. To
estimate the size of the disk, I will consider a particle that was on the edge of the initial
gas cloud, at the equator; this particle had the highest angular momentum of all particles in
the initial cloud. This particle ought to wind up on the outer edge of the disk, so if we can
gure out the size of its nal orbit then we will know the size of the disk.
This particle had initial angular momentum J
0
= mR
2
0

0
. In the end, it is in a circular
orbit where gravity provides the centripetal force, which means it must have radius R
f
such
that
GM m
R
2
f
=
mv
2
f
R
f
= m
2
f
R
f
where
f
= v
f
/R
f
is its nal angular speed. (In writing the gravitational force as GMm/R
2
f
I have assumed that most of the mass is in the central star, i.e., that the self-gravity of the
disk is not important.) The particles nal angular momentum is J
f
= mR
2
f

f
, so we can
substitute for
f
in the equation above to obtain
GM m
R
2
f
=
J
2
f
mR
3
f
1
which we solve for the radius to nd
R
f
=
J
2
f
GM m
2
Now by conservation of angular momentum we set J
f
= J
0
and nd
R
f
=

2
0
R
4
0
GM
Notice that the particle mass has factored out (as it should). So if we know the initial
properties of the cloud mass, size, and rotation speed we can gure out the size of the
disk that will be produced by the clouds collapse.
Example: Solar System. The protoplanetary disk must have been at least R
f
= 40 AU =
6.0 10
14
cm (the size of Plutos orbit). Assuming the initial cloud had M = M

and
R
0
= 0.05 pc = 1.5 10
17
cm as appropriate for a dense molecular cloud core, we can
estimate that its initial rotation had

0
=

GMR
f
R
4
0

1/2
=

(6.67 10
8
cm
3
g
1
s
2
) (1.99 10
33
g) (6.0 10
14
cm)
(1.5 10
17
cm)
4

1/2
= 1.2 10
14
s
1

0
=
2

0
= 5.3 10
14
s
= 1.7 10
7
yr
II. Temperature Structure of a Protoplanetary Disk
Presumably planets form within the disk around a young star. An important ingredient of
any theory of planet formation is the temperature of the material from which planets form.
We can make a simple model of how temperature varies with position in the protoplanetary
disk, or more generally in any sort of accretion disk. Suppose:
1. The disk is basically in equilibrium; more specically, the amount of mass and the
temperature at a given position are basically independent of time.
2. Mass is owing through the disk onto the star. This seems reasonable for the late stages
of star formation. Let m dm/dt be the mass accretion rate.
3. The disk radiates as a blackbody.
2
Figure 1: From Carroll & Ostlie Fig. 18.7.
4. The total mass in the disk is small compared with the mass of the star, so we can
neglect the self-gravity of the disk.
This model cannot be completely accurate. Stu namely planet formation is presum-
ably happening within the disk. The disk must eventually vanish, at least if we want to
produce something like the Solar System. So the disk cannot really be in equilibrium. Also,
this model doesnt say why mass ows through the disk onto the star, or how much. Still,
as usual in this course we will nd a simplied model to be useful for highlighting some of
the important concepts.
Consider a thin annulus of the disk extending from radius r to r +dr, as shown in Figure 1.
Imagine that a small packet of mass dm moves through the annulus, that changes its gravi-
tational potential energy. For circular orbits (and more rigorously, from the virial theorem),
half of this potential energy goes into the kinetic energy of motion (the mass packet will
orbit faster as it moves closer in), while half is released into the gas to be radiated away.
The energy released is thus
dE =
1
2
dU =
dE
dr
=
1
2
dU
dr
=
1
2
GM dm
r
2
= dE =
GM dm
2r
2
dr
where I have used assumption #4 to write the potential energy as GM dm/r. If this
happens in time dt, with mass accretion rate m = dm/dt the energy release is
dE
dt
=
GM m
2r
2
dr
By assumption #1, this amount of energy must be radiated away to keep the disk in equi-
librium. By assumption #3, the energy is radiated as blackbody radiation. The area of the
3
Figure 2: Model for the temperature structure of the protoplanetary disk from which the
Solar System formed. Note that this heat is generated by material falling towards the
center, not light from the central star. From Carroll & Ostlie Fig. 23.9.
annulus is 2r dr, and there is another factor of 2 because radiation can leave both the top
and bottom of the disk. So
L = 4r T
4
dr
Setting L = dE/dt, we can solve for the temperature:
T =

GM m
8 r
3

1/4
The important scaling is
T r
3/4
In our simple model, the temperature decreases as radius increases in the disk due to a
combination of weaker gravity and larger area for radiation to escape from.
The proportionality constant depends on the mass accretion rate m. What do we do about
that? You could try to get it from models of the formation of the Sun. However, it is simpler
to gure that somewhere just inside the orbit of Jupiter the temperature became cool enough
for water to condense. This leads to the pictures shown in Figures 2 and 3.
4
Figure 3: Sketch of the protoplanetary disk. From Carroll & Ostlie Fig. 23.10.
III. Classic Picture of Planet Formation
The general picture is that within the protoplanetary disk the matter was not uniformly
distributed. There were places where the density was higher than average, which meant that
there was a little more gravity
1
that pulled in the nearby material and made the density
get even higher. This runaway process created clumps of matter that eventually grew into
planetesimals. Then the planetesimals started colliding with one another and growing into
bona de planets.
Observational support for this picture:
Asteroids: Are they left over planetesimals?
Moons: Small moons might be planetesimals that got captured by a planet rather
than colliding with it. Large moons (like our own) were probably created by a violent
collision between a planet and a big planetesimal.
Rotation: Most of the planets rotate in the same direction as they orbit the Sun, which
makes sense. However, Venus rotates backwards, and Uranus is tipped on its side.
These unusual rotations could be explained as the results of violent collisions.
Are there other aspects of the Solar System that we can explain? For example, why are the
inner planets small and rocky while the outer planets are big and gaseous?
A natural explanation comes from the temperature structure of the protoplanetary disk.
From our specic model above, and from general reasoning, we imagine that the inner
regions of the disk were hot. Light atoms and molecules like hydrogen and water were
therefore unable to condense and join the reservoir of material out of which the planets
1
We neglected the self gravity of the disk in making our model of the temperature structure above. But
we cannot neglect it when considering how planets form!
5
would form. That left only rocky and metallic material available which is why the inner
planets are rocky.
Beyond some threshold radius (called the snow line), water was able to condense into
the disk. Whats more, the water vapor that was pushed out of the hot inner regions all
condensed around the snow line. This meant there was a lot more mass from which to build
planetesimals. As the planetesimals and young planets grew, they developed strong enough
gravity to capture light gases like hydrogen and helium. There was a lot of mass in these
two gases, so the planets were able to become quite massive (for planets, that is). Thats
why the outer planets are big and gaseous (and largely made of hydrogen).
This picture seems to explain the main features of our Solar System very nicely. That
is reassuring. Then again, maybe its not a surprise: theories of planet formation have
presumably been tuned to agree with what we observe. To test the theories we really need
to consider other planetary systems. This has become possible only in the last decade
thanks to the discovery of extrasolar planets. The new planetary systems have provided
quite a shock. . .
IV. Extrasolar Planets
The rst planets around other (normal) stars were discovered in 1995. Since then, there has
been an explosion in the number of extrasolar planets discovered.
2
Here is the inventory of
planets around normal stars, as of April 21, 2011:
Doppler planets: 499 planets around 419 stars; 49 multiple-planet systems
Transiting planets: 127 planets and over 1200 candidates from Kepler
See http://exoplanet.eu
The big surprise is that many of the planets are massive like Jupiter, but very close to the
star. To take one example, the system HD 209458 is notable because the planet passes in
front of the star and we observe a weak eclipse. The depth of the eclipse allows us to measure
the size of the planet. Combining the mass and size, we can determine the planets density:
Mass (M
J
) Radius (R
J
) Density (g cm
3
)
Earth 0.0031 0.089 5.5
Jupiter 1 1 1.2
HD 209458 0.69 1.3 0.4
2
In fact, I have been involved with the discovery of a few of them, like the planet orbiting the star
Coronae Borealis; see http://arxiv.org/abs/astro-ph/9704248.
6
These numbers conrm that the planet around HD 209458 is a gas giant. (Note that the
planet is less dense than Jupiter; it is presumably pued up by the heat of the star.) Anyway,
here is how the planets orbit compares with Mercury:
Mass (M
J
) Distance from star (AU)
Mercury 0.00017 0.39
HD 209458 0.69 0.045
(See Figure 4 for more examples.)
I should comment on the sample of exoplanets. When people search for planets by looking
for stars that wobble, they are most sensitive to big planets close to the star. It is much
harder to detect planets that are smaller and/or farther from the star. So just because we
have not seen so many of those does not mean they dont exist. Nevertheless, we can say for
certain that big planets close to stars do exist and are quite common.
V. Exoplanets: Challenging Theories of Planet Formation
So whats the big deal? The big planets close to stars are colorfully referred to as hot
Jupiters, and that term tells the story. According to our ideas of planet formation, there
is simply no way that a massive gas giant could form so close to the star. So it is assumed
that the planet actually formed out beyond the snow line, in the vicinity where Jupiter lies
in our Solar System, then traveled in to be close to the star. This raises some important
questions:
How did the planet migrate?
Why did it stop so close to the star? Why didnt it fall in?
Why didnt Jupiter migrate?
Were there other (Earth-like) planets in exoplanetary system? What happened to
them when the big guy migrated?
What does this imply for the possibility of habitable planets and extraterrestrial life?
There are ideas about how to answer these questions, but as far as I am aware there is
not yet a compelling theory that explains both hot Jupiter planetary systems and our own
Solar System. It just goes to show that there are still important discoveries to make, and
important work to do to understand how we came to be here!
7
6
Figure 4: Histograms of extrasolar planet masses (top) and orbit sizes (middle), and their
correlation (bottom). From http://exoplanet.eu/catalog-RV.php
8