CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
ij
Building galaxies
As we saw above, gravitational scattering plays a very small role in the dynamics of stellar orbits
within galaxies or globular clusters.
We return to our original question, “what is a galaxy?”
In
light of the previous section, it seems better to think of a galaxy as a collisionless fluid of stars, a
superposition of a large number of orbits guided by a single background potential.
How does one put together such a galaxy? Superpose orbits! Schwarzschild’s method (see Figure
1.2) follows these simple steps:
•
choose gravitational potential
•
generate an orbit library
•
populate potential with orbits
•
check for self-consistency
gravitational potential
orbit library
light = DF *footprints
orbit library
self−consistent
Figure 1.2: Schwarzschild’s method for
only if
building galaxies. One adjusts the distri-
light ~ mass
bution function (DF) to match the light
distribution or the gravitational poten-
tial.
For example, imagine building a spherically symmetric galaxy. How would the orbits be dis-
tributed in phase space? Even though they may occupy three-dimensional space quite uniformly,
there is no reason to insist that the orbits should be uniformly distributed in momentum space. A
galaxy made only of circular orbits would appear “hot” in the tangential direction and “cold” in the
radial direction, while a collection of radial orbits (an equally valid approach to a spherical system)
would look hot in the radial direction and cold in the tangential direction (see Figure 1.3).
Figure 1.3: Two schemes for constructing a spher-
ical galaxy: out of circular orbits, left, and out of
(nearly) radial orbits, right.
“Hot” and “cold” can be understood quantitatively by defining a symmetric velocity dispersion
tensor:
ε
2
= (v
i
−
v
i
)(v
j
−
v
j
),
(1.14)
where v
i
is the velocity in the
ˆ
e
i
direction averaged at some point. For a collisionless fluid, we can
diagonalize the velocity dispersion tensor, giving three independent temperatures at each point in
1.2.
BUILDING GALAXIES
7
~
�
1
0.8
0.6
0.4
0.2
0
−0.2
−0.4
−0.6
−0.8
−1
−1
−0.8 −0.6 −0.4 −0.2
0
0.2
0.4
0.6
0.8
1
Figure 1.4: The orbit of a star on a
spherically symmetric logarithmic po-
tential. The position of the star is plot-
ted for the first 100 time steps. By fol-
lowing the star for many orbits we accu-
mulate a probability density “footprint”
that may be used to generate a galaxy
model.
Figure 1.5: Left: a disk in circular or-
bit around a spherical bulge. Right:
a disk in polar orbit around a triaxial
bulge. Such “polar ring” galaxies con
stitute perhaps 0.1% of all galaxies.
space. If the random velocities are greater than (or in the same order as) the ordered velocities, the
fluid is hot ; if the ordered velocities are much greater, the fluid is cold.
Observations (Figure 1.7) show that many galaxies have nearly constant rotational velocities
with v(r)
v
c
over the observable range of radius. These flat rotation curves are a departure from
Keplerian orbital velocities, which would scale as v(r)
r
−
1/2
. Flat rotation curves imply mass
increasing linearly with radius, and so cannot continue without bound.
For a spherically symmetric system, the centripetal acceleration is given by
v
2
=
dΛ
(1.15)
r
dr
For a constant rotation velocity v(r)
=
v
c
, the solution for the potential is
Λ
=
v
2
ln r
c
r
ref
.
(1.16)
Combining with Poisson’s equation
�
2
Λ
=
4ψGδ, we find an expression for the density profile:
v
2
δ(r)
=
c
.
(1.17)
4ψGr
2
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
Telescope
Focal Plane
Collimator
Grism
Camera
Detector
Figure 1.6: A schematic representation of a spectrograph.
Light from the telescope is brought to a focus in the focal
plane. The slit of the spectrograph isolates the light from a
narrow strip of the sky, which diverges and passes through
Spectrometer the collimator. Parallel light rays from the collimator enter
the grism (a grating ruled on a prism) which disperses the
different wavelengths. The camera brings the light to a focus
on the detector.
galaxy
[N II]
H
[N II]
red
[SII]
blue
[SII]
Figure 1.7: Left: The slit of a spectrometer is placed along the major axis of a disk galaxy. The disk
is highly inclined to the line of sight. Stars and gas on the left side of the slit have a line of sight
velocity toward us (relative to the center of the galaxy). Right: The spectrum of the galaxy. The
lines shown are typical of gas photo-ionized by hot stars. The lines are blueshifted on the left side
and redshifted on the right side. This gives the rotation curve of the galaxy. The rotation curve is
flat in the outer parts of the galaxy and roughly linear in the inner parts.
spectrograph slit
1.2.
BUILDING GALAXIES
9
�
v
z /rref
r /rref Figure 1.8: Equipotentials for Mestel’s Disk
Among the obvious shortcomings of this model are the failure of the potential to go to zero at large
radius and the prediction of infinite mass. Notwithstanding, it is still a useful potential for many
practical applications.
Another weakness of this model is that it is limited to spherically symmetric galaxies and we
know from experience that many galaxies are shaped more like flattened disks. A corresponding
model for such a galaxy is called a Mestel disk, which is azimuthally symmetric around the z-axis.
The potential for the Mestel disk is
Λ(r, β)
=
v
2
�
ln r
+
ln 1
+
|
cos β
|
�
,
(1.18)
c
r
ref
2
where β is measured from the positive z-axis. The Laplacian in spherical geometry with azimuthal
symmetry is given by
2 1
σ 2 σ 1 σ σ
�
=
r
2
σrr σr
+
r
2
sin β σβ sin βσβ.
(1.19)
Applying Poisson’s equation to equation (1.18), we find that the Mestel disk has zero density every-
where except β
=
ψ/2, where the density is infinite. The more useful (and in this case finite) quantity
is the surface density, determined by integrating the density over the (infinitesimal) thickness of the
disk:
2
Ψ(r)
=
δ(r, z)dz
=
c
.
(1.20)
2ψGr
Like the spherical model for a flat rotation curve, the disk model also suffers from an infinite total
mass and radius, but it is still useful for many applications.
Observations show that most galaxies have elements of both of the above models, with a roughly
spherical bulge and a flattened disk. The relative weight of these features (the “disk-to-bulge ratio,”
Figure 1.9) is a useful parameter for classifying galaxies, with disk strength correlating positively
with other galaxy properties such as young stellar fraction, hydrogen fraction, CO gas, dust, and the
strength of spiral arms. In most galaxies the spiral arms, which are located within the disk are the
primary location for new star formation. The disk and bulge components of most galaxies appear to
have had rather different histories, a subject to be raised again in the context of galaxy formation.
Soon after the realization that many nebulae were in fact distinct galaxies, astronomers (and
in particular Hubble) began systematically classifying galaxies in the hope that taxonomy would
lead to further understanding. The principal component of Hubble’s classification scheme is strongly
correlated with disk-to-bulge ratio. But Hubble conflated the relative proportions of bulge and disk
1
1
2
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
D/B = 0 D/B > 1 D/B < 1 D/B =
Figure 1.9: Galaxies with varying disk to bulge ratios, seen edge on.
Figure 1.10: Two spiral galaxies, floc-
culent to the left and grand design to
the right. Flocculent systems are more
likely to be classified Sa; grand designs
are more frequently Sc.
1.3
with the prominence of spiral arms within the disk. In Hubble’s scheme most galaxies were classified
as “E”, “Sa”, “Sb” or “Sc.” “E” stood for elliptical with no sign of spiral structure. “S” stood for spiral
structure, with “Sa” galaxies showing the least spiral structure and “Sc” galaxies showing the most,
as shown in Figure 1.10. The sequence from E to Sc is approximately one of increasing disk-to- bulge
ratio. But there are also galaxies with prominent disks and little evidence for spiral structure. Hubble
added the S0 classification, placing it between E and Sa. Modern versions of Hubble’s scheme have
put the S0 systems parallel to the S systems, with S0/a, S0/b and S0/c systems having increasingly
large disk-to-bulge ratios. One might hope that disk-to-bulge ratio would completely replace
Hubble’s system, but while easy to describe, disk-to-bulge ratio can be difficult to measure,
particularly for systems that are face-on. Modern “quantitative” classification schemes often use
central concentration as a proxy for disk-to-bulge ratio and use deviation from bilateral symmetry as
a proxy for spiral structure.
8
1.2.
BUILDING GALAXIES
7
ij
Building galaxies
As we saw above, gravitational scattering plays a very small role in the dynamics of stellar orbits
within galaxies or globular clusters.
We return to our original question, “what is a galaxy?”
In
light of the previous section, it seems better to think of a galaxy as a collisionless fluid of stars, a
superposition of a large number of orbits guided by a single background potential.
How does one put together such a galaxy? Superpose orbits! Schwarzschild’s method (see Figure
1.2) follows these simple steps:
•
choose gravitational potential
•
generate an orbit library
•
populate potential with orbits
•
check for self-consistency
gravitational potential
orbit library
light = DF *footprints
orbit library
self−consistent
Figure 1.2: Schwarzschild’s method for
only if
building galaxies. One adjusts the distri-
light ~ mass
bution function (DF) to match the light
distribution or the gravitational poten-
tial.
For example, imagine building a spherically symmetric galaxy. How would the orbits be dis-
tributed in phase space? Even though they may occupy three-dimensional space quite uniformly,
there is no reason to insist that the orbits should be uniformly distributed in momentum space. A
galaxy made only of circular orbits would appear “hot” in the tangential direction and “cold” in the
radial direction, while a collection of radial orbits (an equally valid approach to a spherical system)
would look hot in the radial direction and cold in the tangential direction (see Figure 1.3).
Figure 1.3: Two schemes for constructing a spher-
ical galaxy: out of circular orbits, left, and out of
(nearly) radial orbits, right.
“Hot” and “cold” can be understood quantitatively by defining a symmetric velocity dispersion
tensor:
ε
2
= (v
i
−
v
i
)(v
j
−
v
j
),
(1.14)
where v
i
is the velocity in the
ˆ
e
i
direction averaged at some point. For a collisionless fluid, we can
diagonalize the velocity dispersion tensor, giving three independent temperatures at each point in
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
~
�
1
0.8
0.6
0.4
0.2
0
−0.2
−0.4
−0.6
−0.8
−1
−1
−0.8 −0.6 −0.4 −0.2
0
0.2
0.4
0.6
0.8
1
Figure 1.4: The orbit of a star on a
spherically symmetric logarithmic po-
tential. The position of the star is plot-
ted for the first 100 time steps. By fol-
lowing the star for many orbits we accu-
mulate a probability density “footprint”
that may be used to generate a galaxy
model.
Figure 1.5: Left: a disk in circular or-
bit around a spherical bulge. Right:
a disk in polar orbit around a triaxial
bulge. Such “polar ring” galaxies con
stitute perhaps 0.1% of all galaxies.
space. If the random velocities are greater than (or in the same order as) the ordered velocities, the
fluid is hot ; if the ordered velocities are much greater, the fluid is cold.
Observations (Figure 1.7) show that many galaxies have nearly constant rotational velocities
with v(r)
v
c
over the observable range of radius. These flat rotation curves are a departure from
Keplerian orbital velocities, which would scale as v(r)
r
−
1/2
. Flat rotation curves imply mass
increasing linearly with radius, and so cannot continue without bound.
For a spherically symmetric system, the centripetal acceleration is given by
v
2
=
dΛ
(1.15)
r
dr
For a constant rotation velocity v(r)
=
v
c
, the solution for the potential is
Λ
=
v
2
ln r
c
r
ref
.
(1.16)
Combining with Poisson’s equation
�
2
Λ
=
4ψGδ, we find an expression for the density profile:
v
2
δ(r)
=
c
.
(1.17)
4ψGr
2
1.2.
BUILDING GALAXIES
9
Telescope
Focal Plane
Collimator
Grism
Camera
Detector
Figure 1.6: A schematic representation of a spectrograph.
Light from the telescope is brought to a focus in the focal
plane. The slit of the spectrograph isolates the light from a
narrow strip of the sky, which diverges and passes through
Spectrometer the collimator. Parallel light rays from the collimator enter
the grism (a grating ruled on a prism) which disperses the
different wavelengths. The camera brings the light to a focus
on the detector.
galaxy
[N II]
H
[N II]
red
[SII]
blue
[SII]
Figure 1.7: Left: The slit of a spectrometer is placed along the major axis of a disk galaxy. The disk
is highly inclined to the line of sight. Stars and gas on the left side of the slit have a line of sight
velocity toward us (relative to the center of the galaxy). Right: The spectrum of the galaxy. The
lines shown are typical of gas photo-ionized by hot stars. The lines are blueshifted on the left side
and redshifted on the right side. This gives the rotation curve of the galaxy. The rotation curve is
flat in the outer parts of the galaxy and roughly linear in the inner parts.
spectrograph slit
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
�
v
z /rref
r /rref Figure 1.8: Equipotentials for Mestel’s Disk
Among the obvious shortcomings of this model are the failure of the potential to go to zero at large
radius and the prediction of infinite mass. Notwithstanding, it is still a useful potential for many
practical applications.
Another weakness of this model is that it is limited to spherically symmetric galaxies and we
know from experience that many galaxies are shaped more like flattened disks. A corresponding
model for such a galaxy is called a Mestel disk, which is azimuthally symmetric around the z-axis.
The potential for the Mestel disk is
Λ(r, β)
=
v
2
�
ln r
+
ln 1
+
|
cos β
|
�
,
(1.18)
c
r
ref
2
where β is measured from the positive z-axis. The Laplacian in spherical geometry with azimuthal
symmetry is given by
2 1
σ 2 σ 1 σ σ
�
=
r
2
σrr σr
+
r
2
sin β σβ sin βσβ.
(1.19)
Applying Poisson’s equation to equation (1.18), we find that the Mestel disk has zero density every-
where except β
=
ψ/2, where the density is infinite. The more useful (and in this case finite) quantity
is the surface density, determined by integrating the density over the (infinitesimal) thickness of the
disk:
2
Ψ(r)
=
δ(r, z)dz
=
c
.
(1.20)
2ψGr
Like the spherical model for a flat rotation curve, the disk model also suffers from an infinite total
mass and radius, but it is still useful for many applications.
Observations show that most galaxies have elements of both of the above models, with a roughly
spherical bulge and a flattened disk. The relative weight of these features (the “disk-to-bulge ratio,”
Figure 1.9) is a useful parameter for classifying galaxies, with disk strength correlating positively
with other galaxy properties such as young stellar fraction, hydrogen fraction, CO gas, dust, and the
strength of spiral arms. In most galaxies the spiral arms, which are located within the disk are the
primary location for new star formation. The disk and bulge components of most galaxies appear to
have had rather different histories, a subject to be raised again in the context of galaxy formation.
Soon after the realization that many nebulae were in fact distinct galaxies, astronomers (and
in particular Hubble) began systematically classifying galaxies in the hope that taxonomy would
lead to further understanding. The principal component of Hubble’s classification scheme is strongly
correlated with disk-to-bulge ratio. But Hubble conflated the relative proportions of bulge and disk
1
1
2
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
D/B = 0 D/B > 1 D/B < 1 D/B =
Figure 1.9: Galaxies with varying disk to bulge ratios, seen edge on.
Figure 1.10: Two spiral galaxies, floc-
culent to the left and grand design to
the right. Flocculent systems are more
likely to be classified Sa; grand designs
are more frequently Sc.
with the prominence of spiral arms within the disk. In Hubble’s scheme most galaxies were classified
as “E”, “Sa”, “Sb” or “Sc.” “E” stood for elliptical with no sign of spiral structure. “S” stood for
spiral structure, with “Sa” galaxies showing the least spiral structure and “Sc” galaxies showing the
most, as shown in Figure 1.10. The sequence from E to Sc is approximately one of increasing disk-to-
bulge ratio. But there are also galaxies with prominent disks and little evidence for spiral structure.
Hubble added the S0 classification, placing it between E and Sa. Modern versions of Hubble’s
scheme have put the S0 systems parallel to the S systems, with S0/a, S0/b and S0/c systems having
increasingly large disk-to-bulge ratios. One might hope that disk-to-bulge ratio would completely
replace Hubble’s system, but while easy to describe, disk-to-bulge ratio can be difficult to measure,
particularly for systems that are face-on. Modern “quantitative” classification schemes often use
central concentration as a proxy for disk-to-bulge ratio and use deviation from bilateral symmetry
as a proxy for spiral structure.
8
1.2.
BUILDING GALAXIES
7
ij
Building galaxies
As we saw above, gravitational scattering plays a very small role in the dynamics of stellar orbits
within galaxies or globular clusters.
We return to our original question, “what is a galaxy?”
In
light of the previous section, it seems better to think of a galaxy as a collisionless fluid of stars, a
superposition of a large number of orbits guided by a single background potential.
How does one put together such a galaxy? Superpose orbits! Schwarzschild’s method (see Figure
1.2) follows these simple steps:
•
choose gravitational potential
•
generate an orbit library
•
populate potential with orbits
•
check for self-consistency
gravitational potential
orbit library
light = DF *footprints
orbit library
self−consistent
Figure 1.2: Schwarzschild’s method for
only if
building galaxies. One adjusts the distri-
light ~ mass
bution function (DF) to match the light
distribution or the gravitational poten-
tial.
For example, imagine building a spherically symmetric galaxy. How would the orbits be dis-
tributed in phase space? Even though they may occupy three-dimensional space quite uniformly,
there is no reason to insist that the orbits should be uniformly distributed in momentum space. A
galaxy made only of circular orbits would appear “hot” in the tangential direction and “cold” in the
radial direction, while a collection of radial orbits (an equally valid approach to a spherical system)
would look hot in the radial direction and cold in the tangential direction (see Figure 1.3).
Figure 1.3: Two schemes for constructing a spher-
ical galaxy: out of circular orbits, left, and out of
(nearly) radial orbits, right.
“Hot” and “cold” can be understood quantitatively by defining a symmetric velocity dispersion
tensor:
ε
2
= (v
i
−
v
i
)(v
j
−
v
j
),
(1.14)
where v
i
is the velocity in the
ˆ
e
i
direction averaged at some point. For a collisionless fluid, we can
diagonalize the velocity dispersion tensor, giving three independent temperatures at each point in
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
~
�
1
0.8
0.6
0.4
0.2
0
−0.2
−0.4
−0.6
−0.8
−1
−1
−0.8 −0.6 −0.4 −0.2
0
0.2
0.4
0.6
0.8
1
Figure 1.4: The orbit of a star on a
spherically symmetric logarithmic po-
tential. The position of the star is plot-
ted for the first 100 time steps. By fol-
lowing the star for many orbits we accu-
mulate a probability density “footprint”
that may be used to generate a galaxy
model.
Figure 1.5: Left: a disk in circular or-
bit around a spherical bulge. Right:
a disk in polar orbit around a triaxial
bulge. Such “polar ring” galaxies con
stitute perhaps 0.1% of all galaxies.
space. If the random velocities are greater than (or in the same order as) the ordered velocities, the
fluid is hot ; if the ordered velocities are much greater, the fluid is cold.
Observations (Figure 1.7) show that many galaxies have nearly constant rotational velocities
with v(r)
v
c
over the observable range of radius. These flat rotation curves are a departure from
Keplerian orbital velocities, which would scale as v(r)
r
−
1/2
. Flat rotation curves imply mass
increasing linearly with radius, and so cannot continue without bound.
For a spherically symmetric system, the centripetal acceleration is given by
v
2
=
dΛ
(1.15)
r
dr
For a constant rotation velocity v(r)
=
v
c
, the solution for the potential is
Λ
=
v
2
ln r
c
r
ref
.
(1.16)
Combining with Poisson’s equation
�
2
Λ
=
4ψGδ, we find an expression for the density profile:
v
2
δ(r)
=
c
.
(1.17)
4ψGr
2
1.2.
BUILDING GALAXIES
9
Telescope
Focal Plane
Collimator
Grism
Camera
Detector
Figure 1.6: A schematic representation of a spectrograph.
Light from the telescope is brought to a focus in the focal
plane. The slit of the spectrograph isolates the light from a
narrow strip of the sky, which diverges and passes through
Spectrometer the collimator. Parallel light rays from the collimator enter
the grism (a grating ruled on a prism) which disperses the
different wavelengths. The camera brings the light to a focus
on the detector.
galaxy
[N II]
H
[N II]
red
[SII]
blue
[SII]
Figure 1.7: Left: The slit of a spectrometer is placed along the major axis of a disk galaxy. The disk
is highly inclined to the line of sight. Stars and gas on the left side of the slit have a line of sight
velocity toward us (relative to the center of the galaxy). Right: The spectrum of the galaxy. The
lines shown are typical of gas photo-ionized by hot stars. The lines are blueshifted on the left side
and redshifted on the right side. This gives the rotation curve of the galaxy. The rotation curve is
flat in the outer parts of the galaxy and roughly linear in the inner parts.
spectrograph slit
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
�
v
z /rref
r /rref Figure 1.8: Equipotentials for Mestel’s Disk
Among the obvious shortcomings of this model are the failure of the potential to go to zero at large
radius and the prediction of infinite mass. Notwithstanding, it is still a useful potential for many
practical applications.
Another weakness of this model is that it is limited to spherically symmetric galaxies and we
know from experience that many galaxies are shaped more like flattened disks. A corresponding
model for such a galaxy is called a Mestel disk, which is azimuthally symmetric around the z-axis.
The potential for the Mestel disk is
Λ(r, β)
=
v
2
�
ln r
+
ln 1
+
|
cos β
|
�
,
(1.18)
c
r
ref
2
where β is measured from the positive z-axis. The Laplacian in spherical geometry with azimuthal
symmetry is given by
2 1
σ 2 σ 1 σ σ
�
=
r
2
σrr σr
+
r
2
sin β σβ sin βσβ.
(1.19)
Applying Poisson’s equation to equation (1.18), we find that the Mestel disk has zero density every-
where except β
=
ψ/2, where the density is infinite. The more useful (and in this case finite) quantity
is the surface density, determined by integrating the density over the (infinitesimal) thickness of the
disk:
2
Ψ(r)
=
δ(r, z)dz
=
c
.
(1.20)
2ψGr
Like the spherical model for a flat rotation curve, the disk model also suffers from an infinite total
mass and radius, but it is still useful for many applications.
Observations show that most galaxies have elements of both of the above models, with a roughly
spherical bulge and a flattened disk. The relative weight of these features (the “disk-to-bulge ratio,”
Figure 1.9) is a useful parameter for classifying galaxies, with disk strength correlating positively
with other galaxy properties such as young stellar fraction, hydrogen fraction, CO gas, dust, and the
strength of spiral arms. In most galaxies the spiral arms, which are located within the disk are the
primary location for new star formation. The disk and bulge components of most galaxies appear to
have had rather different histories, a subject to be raised again in the context of galaxy formation.
Soon after the realization that many nebulae were in fact distinct galaxies, astronomers (and
in particular Hubble) began systematically classifying galaxies in the hope that taxonomy would
lead to further understanding. The principal component of Hubble’s classification scheme is strongly
correlated with disk-to-bulge ratio. But Hubble conflated the relative proportions of bulge and disk
1
1
2
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
D/B = 0 D/B > 1 D/B < 1 D/B =
Figure 1.9: Galaxies with varying disk to bulge ratios, seen edge on.
Figure 1.10: Two spiral galaxies, floc-
culent to the left and grand design to
the right. Flocculent systems are more
likely to be classified Sa; grand designs
are more frequently Sc.
1.4
with the prominence of spiral arms within the disk. In Hubble’s scheme most galaxies were classified
as “E”, “Sa”, “Sb” or “Sc.” “E” stood for elliptical with no sign of spiral structure. “S” stood for spiral
structure, with “Sa” galaxies showing the least spiral structure and “Sc” galaxies showing the most,
as shown in Figure 1.10. The sequence from E to Sc is approximately one of increasing disk-to- bulge
ratio. But there are also galaxies with prominent disks and little evidence for spiral structure. Hubble
added the S0 classification, placing it between E and Sa. Modern versions of Hubble’s scheme have
put the S0 systems parallel to the S systems, with S0/a, S0/b and S0/c systems having increasingly
large disk-to-bulge ratios. One might hope that disk-to-bulge ratio would completely replace
Hubble’s system, but while easy to describe, disk-to-bulge ratio can be difficult to measure,
particularly for systems that are face-on. Modern “quantitative” classification schemes often use
central concentration as a proxy for disk-to-bulge ratio and use deviation from bilateral symmetry as
a proxy for spiral structure.
8
1.2.
BUILDING GALAXIES
7
ij
Building galaxies
As we saw above, gravitational scattering plays a very small role in the dynamics of stellar orbits
within galaxies or globular clusters.
We return to our original question, “what is a galaxy?”
In
light of the previous section, it seems better to think of a galaxy as a collisionless fluid of stars, a
superposition of a large number of orbits guided by a single background potential.
How does one put together such a galaxy? Superpose orbits! Schwarzschild’s method (see Figure
1.2) follows these simple steps:
•
choose gravitational potential
•
generate an orbit library
•
populate potential with orbits
•
check for self-consistency
gravitational potential
orbit library
light = DF *footprints
orbit library
self−consistent
Figure 1.2: Schwarzschild’s method for
only if
building galaxies. One adjusts the distri-
light ~ mass
bution function (DF) to match the light
distribution or the gravitational poten-
tial.
For example, imagine building a spherically symmetric galaxy. How would the orbits be dis-
tributed in phase space? Even though they may occupy three-dimensional space quite uniformly,
there is no reason to insist that the orbits should be uniformly distributed in momentum space. A
galaxy made only of circular orbits would appear “hot” in the tangential direction and “cold” in the
radial direction, while a collection of radial orbits (an equally valid approach to a spherical system)
would look hot in the radial direction and cold in the tangential direction (see Figure 1.3).
Figure 1.3: Two schemes for constructing a spher-
ical galaxy: out of circular orbits, left, and out of
(nearly) radial orbits, right.
“Hot” and “cold” can be understood quantitatively by defining a symmetric velocity dispersion
tensor:
ε
2
= (v
i
−
v
i
)(v
j
−
v
j
),
(1.14)
where v
i
is the velocity in the
ˆ
e
i
direction averaged at some point. For a collisionless fluid, we can
diagonalize the velocity dispersion tensor, giving three independent temperatures at each point in
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
~
�
1
0.8
0.6
0.4
0.2
0
−0.2
−0.4
−0.6
−0.8
−1
−1
−0.8 −0.6 −0.4 −0.2
0
0.2
0.4
0.6
0.8
1
Figure 1.4: The orbit of a star on a
spherically symmetric logarithmic po-
tential. The position of the star is plot-
ted for the first 100 time steps. By fol-
lowing the star for many orbits we accu-
mulate a probability density “footprint”
that may be used to generate a galaxy
model.
Figure 1.5: Left: a disk in circular or-
bit around a spherical bulge. Right:
a disk in polar orbit around a triaxial
bulge. Such “polar ring” galaxies con
stitute perhaps 0.1% of all galaxies.
space. If the random velocities are greater than (or in the same order as) the ordered velocities, the
fluid is hot ; if the ordered velocities are much greater, the fluid is cold.
Observations (Figure 1.7) show that many galaxies have nearly constant rotational velocities
with v(r)
v
c
over the observable range of radius. These flat rotation curves are a departure from
Keplerian orbital velocities, which would scale as v(r)
r
−
1/2
. Flat rotation curves imply mass
increasing linearly with radius, and so cannot continue without bound.
For a spherically symmetric system, the centripetal acceleration is given by
v
2
=
dΛ
(1.15)
r
dr
For a constant rotation velocity v(r)
=
v
c
, the solution for the potential is
Λ
=
v
2
ln r
c
r
ref
.
(1.16)
Combining with Poisson’s equation
�
2
Λ
=
4ψGδ, we find an expression for the density profile:
v
2
δ(r)
=
c
.
(1.17)
4ψGr
2
1.2.
BUILDING GALAXIES
9
Telescope
Focal Plane
Collimator
Grism
Camera
Detector
Figure 1.6: A schematic representation of a spectrograph.
Light from the telescope is brought to a focus in the focal
plane. The slit of the spectrograph isolates the light from a
narrow strip of the sky, which diverges and passes through
Spectrometer the collimator. Parallel light rays from the collimator enter
the grism (a grating ruled on a prism) which disperses the
different wavelengths. The camera brings the light to a focus
on the detector.
galaxy
[N II]
H
[N II]
red
[SII]
blue
[SII]
Figure 1.7: Left: The slit of a spectrometer is placed along the major axis of a disk galaxy. The disk
is highly inclined to the line of sight. Stars and gas on the left side of the slit have a line of sight
velocity toward us (relative to the center of the galaxy). Right: The spectrum of the galaxy. The
lines shown are typical of gas photo-ionized by hot stars. The lines are blueshifted on the left side
and redshifted on the right side. This gives the rotation curve of the galaxy. The rotation curve is
flat in the outer parts of the galaxy and roughly linear in the inner parts.
spectrograph slit
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
�
v
z /rref
r /rref Figure 1.8: Equipotentials for Mestel’s Disk
Among the obvious shortcomings of this model are the failure of the potential to go to zero at large
radius and the prediction of infinite mass. Notwithstanding, it is still a useful potential for many
practical applications.
Another weakness of this model is that it is limited to spherically symmetric galaxies and we
know from experience that many galaxies are shaped more like flattened disks. A corresponding
model for such a galaxy is called a Mestel disk, which is azimuthally symmetric around the z-axis.
The potential for the Mestel disk is
Λ(r, β)
=
v
2
�
ln r
+
ln 1
+
|
cos β
|
�
,
(1.18)
c
r
ref
2
where β is measured from the positive z-axis. The Laplacian in spherical geometry with azimuthal
symmetry is given by
2 1
σ 2 σ 1 σ σ
�
=
r
2
σrr σr
+
r
2
sin β σβ sin βσβ.
(1.19)
Applying Poisson’s equation to equation (1.18), we find that the Mestel disk has zero density every-
where except β
=
ψ/2, where the density is infinite. The more useful (and in this case finite) quantity
is the surface density, determined by integrating the density over the (infinitesimal) thickness of the
disk:
2
Ψ(r)
=
δ(r, z)dz
=
c
.
(1.20)
2ψGr
Like the spherical model for a flat rotation curve, the disk model also suffers from an infinite total
mass and radius, but it is still useful for many applications.
Observations show that most galaxies have elements of both of the above models, with a roughly
spherical bulge and a flattened disk. The relative weight of these features (the “disk-to-bulge ratio,”
Figure 1.9) is a useful parameter for classifying galaxies, with disk strength correlating positively
with other galaxy properties such as young stellar fraction, hydrogen fraction, CO gas, dust, and the
strength of spiral arms. In most galaxies the spiral arms, which are located within the disk are the
primary location for new star formation. The disk and bulge components of most galaxies appear to
have had rather different histories, a subject to be raised again in the context of galaxy formation.
Soon after the realization that many nebulae were in fact distinct galaxies, astronomers (and
in particular Hubble) began systematically classifying galaxies in the hope that taxonomy would
lead to further understanding. The principal component of Hubble’s classification scheme is strongly
correlated with disk-to-bulge ratio. But Hubble conflated the relative proportions of bulge and disk
1
1
2
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
D/B = 0 D/B > 1 D/B < 1 D/B =
Figure 1.9: Galaxies with varying disk to bulge ratios, seen edge on.
Figure 1.10: Two spiral galaxies, floc-
culent to the left and grand design to
the right. Flocculent systems are more
likely to be classified Sa; grand designs
are more frequently Sc.
with the prominence of spiral arms within the disk. In Hubble’s scheme most galaxies were classified
as “E”, “Sa”, “Sb” or “Sc.” “E” stood for elliptical with no sign of spiral structure. “S” stood for
spiral structure, with “Sa” galaxies showing the least spiral structure and “Sc” galaxies showing the
most, as shown in Figure 1.10. The sequence from E to Sc is approximately one of increasing disk-to-
bulge ratio. But there are also galaxies with prominent disks and little evidence for spiral structure.
Hubble added the S0 classification, placing it between E and Sa. Modern versions of Hubble’s
scheme have put the S0 systems parallel to the S systems, with S0/a, S0/b and S0/c systems having
increasingly large disk-to-bulge ratios. One might hope that disk-to-bulge ratio would completely
replace Hubble’s system, but while easy to describe, disk-to-bulge ratio can be difficult to measure,
particularly for systems that are face-on. Modern “quantitative” classification schemes often use
central concentration as a proxy for disk-to-bulge ratio and use deviation from bilateral symmetry
as a proxy for spiral structure.
8
1.2.
BUILDING GALAXIES
7
ij
Building galaxies
As we saw above, gravitational scattering plays a very small role in the dynamics of stellar orbits
within galaxies or globular clusters.
We return to our original question, “what is a galaxy?”
In
light of the previous section, it seems better to think of a galaxy as a collisionless fluid of stars, a
superposition of a large number of orbits guided by a single background potential.
How does one put together such a galaxy? Superpose orbits! Schwarzschild’s method (see Figure
1.2) follows these simple steps:
•
choose gravitational potential
•
generate an orbit library
•
populate potential with orbits
•
check for self-consistency
gravitational potential
orbit library
light = DF *footprints
orbit library
self−consistent
Figure 1.2: Schwarzschild’s method for
only if
building galaxies. One adjusts the distri-
light ~ mass
bution function (DF) to match the light
distribution or the gravitational poten-
tial.
For example, imagine building a spherically symmetric galaxy. How would the orbits be dis-
tributed in phase space? Even though they may occupy three-dimensional space quite uniformly,
there is no reason to insist that the orbits should be uniformly distributed in momentum space. A
galaxy made only of circular orbits would appear “hot” in the tangential direction and “cold” in the
radial direction, while a collection of radial orbits (an equally valid approach to a spherical system)
would look hot in the radial direction and cold in the tangential direction (see Figure 1.3).
Figure 1.3: Two schemes for constructing a spher-
ical galaxy: out of circular orbits, left, and out of
(nearly) radial orbits, right.
“Hot” and “cold” can be understood quantitatively by defining a symmetric velocity dispersion
tensor:
ε
2
= (v
i
−
v
i
)(v
j
−
v
j
),
(1.14)
where v
i
is the velocity in the
ˆ
e
i
direction averaged at some point. For a collisionless fluid, we can
diagonalize the velocity dispersion tensor, giving three independent temperatures at each point in
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
~
�
1
0.8
0.6
0.4
0.2
0
−0.2
−0.4
−0.6
−0.8
−1
−1
−0.8 −0.6 −0.4 −0.2
0
0.2
0.4
0.6
0.8
1
Figure 1.4: The orbit of a star on a
spherically symmetric logarithmic po-
tential. The position of the star is plot-
ted for the first 100 time steps. By fol-
lowing the star for many orbits we accu-
mulate a probability density “footprint”
that may be used to generate a galaxy
model.
Figure 1.5: Left: a disk in circular or-
bit around a spherical bulge. Right:
a disk in polar orbit around a triaxial
bulge. Such “polar ring” galaxies con
stitute perhaps 0.1% of all galaxies.
space. If the random velocities are greater than (or in the same order as) the ordered velocities, the
fluid is hot ; if the ordered velocities are much greater, the fluid is cold.
Observations (Figure 1.7) show that many galaxies have nearly constant rotational velocities
with v(r)
v
c
over the observable range of radius. These flat rotation curves are a departure from
Keplerian orbital velocities, which would scale as v(r)
r
−
1/2
. Flat rotation curves imply mass
increasing linearly with radius, and so cannot continue without bound.
For a spherically symmetric system, the centripetal acceleration is given by
v
2
=
dΛ
(1.15)
r
dr
For a constant rotation velocity v(r)
=
v
c
, the solution for the potential is
Λ
=
v
2
ln r
c
r
ref
.
(1.16)
Combining with Poisson’s equation
�
2
Λ
=
4ψGδ, we find an expression for the density profile:
v
2
δ(r)
=
c
.
(1.17)
4ψGr
2
1.2.
BUILDING GALAXIES
9
Telescope
Focal Plane
Collimator
Grism
Camera
Detector
Figure 1.6: A schematic representation of a spectrograph.
Light from the telescope is brought to a focus in the focal
plane. The slit of the spectrograph isolates the light from a
narrow strip of the sky, which diverges and passes through
Spectrometer the collimator. Parallel light rays from the collimator enter
the grism (a grating ruled on a prism) which disperses the
different wavelengths. The camera brings the light to a focus
on the detector.
galaxy
[N II]
H
[N II]
red
[SII]
blue
[SII]
Figure 1.7: Left: The slit of a spectrometer is placed along the major axis of a disk galaxy. The disk
is highly inclined to the line of sight. Stars and gas on the left side of the slit have a line of sight
velocity toward us (relative to the center of the galaxy). Right: The spectrum of the galaxy. The
lines shown are typical of gas photo-ionized by hot stars. The lines are blueshifted on the left side
and redshifted on the right side. This gives the rotation curve of the galaxy. The rotation curve is
flat in the outer parts of the galaxy and roughly linear in the inner parts.
spectrograph slit
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
�
v
z /rref
r /rref Figure 1.8: Equipotentials for Mestel’s Disk
Among the obvious shortcomings of this model are the failure of the potential to go to zero at large
radius and the prediction of infinite mass. Notwithstanding, it is still a useful potential for many
practical applications.
Another weakness of this model is that it is limited to spherically symmetric galaxies and we
know from experience that many galaxies are shaped more like flattened disks. A corresponding
model for such a galaxy is called a Mestel disk, which is azimuthally symmetric around the z-axis.
The potential for the Mestel disk is
Λ(r, β)
=
v
2
�
ln r
+
ln 1
+
|
cos β
|
�
,
(1.18)
c
r
ref
2
where β is measured from the positive z-axis. The Laplacian in spherical geometry with azimuthal
symmetry is given by
2 1
σ 2 σ 1 σ σ
�
=
r
2
σrr σr
+
r
2
sin β σβ sin βσβ.
(1.19)
Applying Poisson’s equation to equation (1.18), we find that the Mestel disk has zero density every-
where except β
=
ψ/2, where the density is infinite. The more useful (and in this case finite) quantity
is the surface density, determined by integrating the density over the (infinitesimal) thickness of the
disk:
2
Ψ(r)
=
δ(r, z)dz
=
c
.
(1.20)
2ψGr
Like the spherical model for a flat rotation curve, the disk model also suffers from an infinite total
mass and radius, but it is still useful for many applications.
Observations show that most galaxies have elements of both of the above models, with a roughly
spherical bulge and a flattened disk. The relative weight of these features (the “disk-to-bulge ratio,”
Figure 1.9) is a useful parameter for classifying galaxies, with disk strength correlating positively
with other galaxy properties such as young stellar fraction, hydrogen fraction, CO gas, dust, and the
strength of spiral arms. In most galaxies the spiral arms, which are located within the disk are the
primary location for new star formation. The disk and bulge components of most galaxies appear to
have had rather different histories, a subject to be raised again in the context of galaxy formation.
Soon after the realization that many nebulae were in fact distinct galaxies, astronomers (and
in particular Hubble) began systematically classifying galaxies in the hope that taxonomy would
lead to further understanding. The principal component of Hubble’s classification scheme is strongly
correlated with disk-to-bulge ratio. But Hubble conflated the relative proportions of bulge and disk
1
1
2
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
D/B = 0 D/B > 1 D/B < 1 D/B =
Figure 1.9: Galaxies with varying disk to bulge ratios, seen edge on.
Figure 1.10: Two spiral galaxies, floc-
culent to the left and grand design to
the right. Flocculent systems are more
likely to be classified Sa; grand designs
are more frequently Sc.
1.5
with the prominence of spiral arms within the disk. In Hubble’s scheme most galaxies were classified
as “E”, “Sa”, “Sb” or “Sc.” “E” stood for elliptical with no sign of spiral structure. “S” stood for spiral
structure, with “Sa” galaxies showing the least spiral structure and “Sc” galaxies showing the most,
as shown in Figure 1.10. The sequence from E to Sc is approximately one of increasing disk-to- bulge
ratio. But there are also galaxies with prominent disks and little evidence for spiral structure. Hubble
added the S0 classification, placing it between E and Sa. Modern versions of Hubble’s scheme have
put the S0 systems parallel to the S systems, with S0/a, S0/b and S0/c systems having increasingly
large disk-to-bulge ratios. One might hope that disk-to-bulge ratio would completely replace
Hubble’s system, but while easy to describe, disk-to-bulge ratio can be difficult to measure,
particularly for systems that are face-on. Modern “quantitative” classification schemes often use
central concentration as a proxy for disk-to-bulge ratio and use deviation from bilateral symmetry as
a proxy for spiral structure.
8
1.2.
BUILDING GALAXIES
7
ij
Building galaxies
As we saw above, gravitational scattering plays a very small role in the dynamics of stellar orbits
within galaxies or globular clusters.
We return to our original question, “what is a galaxy?”
In
light of the previous section, it seems better to think of a galaxy as a collisionless fluid of stars, a
superposition of a large number of orbits guided by a single background potential.
How does one put together such a galaxy? Superpose orbits! Schwarzschild’s method (see Figure
1.2) follows these simple steps:
•
choose gravitational potential
•
generate an orbit library
•
populate potential with orbits
•
check for self-consistency
gravitational potential
orbit library
light = DF *footprints
orbit library
self−consistent
Figure 1.2: Schwarzschild’s method for
only if
building galaxies. One adjusts the distri-
light ~ mass
bution function (DF) to match the light
distribution or the gravitational poten-
tial.
For example, imagine building a spherically symmetric galaxy. How would the orbits be dis-
tributed in phase space? Even though they may occupy three-dimensional space quite uniformly,
there is no reason to insist that the orbits should be uniformly distributed in momentum space. A
galaxy made only of circular orbits would appear “hot” in the tangential direction and “cold” in the
radial direction, while a collection of radial orbits (an equally valid approach to a spherical system)
would look hot in the radial direction and cold in the tangential direction (see Figure 1.3).
Figure 1.3: Two schemes for constructing a spher-
ical galaxy: out of circular orbits, left, and out of
(nearly) radial orbits, right.
“Hot” and “cold” can be understood quantitatively by defining a symmetric velocity dispersion
tensor:
ε
2
= (v
i
−
v
i
)(v
j
−
v
j
),
(1.14)
where v
i
is the velocity in the
ˆ
e
i
direction averaged at some point. For a collisionless fluid, we can
diagonalize the velocity dispersion tensor, giving three independent temperatures at each point in
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
~
�
1
0.8
0.6
0.4
0.2
0
−0.2
−0.4
−0.6
−0.8
−1
−1
−0.8 −0.6 −0.4 −0.2
0
0.2
0.4
0.6
0.8
1
Figure 1.4: The orbit of a star on a
spherically symmetric logarithmic po-
tential. The position of the star is plot-
ted for the first 100 time steps. By fol-
lowing the star for many orbits we accu-
mulate a probability density “footprint”
that may be used to generate a galaxy
model.
Figure 1.5: Left: a disk in circular or-
bit around a spherical bulge. Right:
a disk in polar orbit around a triaxial
bulge. Such “polar ring” galaxies con
stitute perhaps 0.1% of all galaxies.
space. If the random velocities are greater than (or in the same order as) the ordered velocities, the
fluid is hot ; if the ordered velocities are much greater, the fluid is cold.
Observations (Figure 1.7) show that many galaxies have nearly constant rotational velocities
with v(r)
v
c
over the observable range of radius. These flat rotation curves are a departure from
Keplerian orbital velocities, which would scale as v(r)
r
−
1/2
. Flat rotation curves imply mass
increasing linearly with radius, and so cannot continue without bound.
For a spherically symmetric system, the centripetal acceleration is given by
v
2
=
dΛ
(1.15)
r
dr
For a constant rotation velocity v(r)
=
v
c
, the solution for the potential is
Λ
=
v
2
ln r
c
r
ref
.
(1.16)
Combining with Poisson’s equation
�
2
Λ
=
4ψGδ, we find an expression for the density profile:
v
2
δ(r)
=
c
.
(1.17)
4ψGr
2
1.2.
BUILDING GALAXIES
9
Telescope
Focal Plane
Collimator
Grism
Camera
Detector
Figure 1.6: A schematic representation of a spectrograph.
Light from the telescope is brought to a focus in the focal
plane. The slit of the spectrograph isolates the light from a
narrow strip of the sky, which diverges and passes through
Spectrometer the collimator. Parallel light rays from the collimator enter
the grism (a grating ruled on a prism) which disperses the
different wavelengths. The camera brings the light to a focus
on the detector.
galaxy
[N II]
H
[N II]
red
[SII]
blue
[SII]
Figure 1.7: Left: The slit of a spectrometer is placed along the major axis of a disk galaxy. The disk
is highly inclined to the line of sight. Stars and gas on the left side of the slit have a line of sight
velocity toward us (relative to the center of the galaxy). Right: The spectrum of the galaxy. The
lines shown are typical of gas photo-ionized by hot stars. The lines are blueshifted on the left side
and redshifted on the right side. This gives the rotation curve of the galaxy. The rotation curve is
flat in the outer parts of the galaxy and roughly linear in the inner parts.
spectrograph slit
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
�
v
z /rref
r /rref Figure 1.8: Equipotentials for Mestel’s Disk
Among the obvious shortcomings of this model are the failure of the potential to go to zero at large
radius and the prediction of infinite mass. Notwithstanding, it is still a useful potential for many
practical applications.
Another weakness of this model is that it is limited to spherically symmetric galaxies and we
know from experience that many galaxies are shaped more like flattened disks. A corresponding
model for such a galaxy is called a Mestel disk, which is azimuthally symmetric around the z-axis.
The potential for the Mestel disk is
Λ(r, β)
=
v
2
�
ln r
+
ln 1
+
|
cos β
|
�
,
(1.18)
c
r
ref
2
where β is measured from the positive z-axis. The Laplacian in spherical geometry with azimuthal
symmetry is given by
2 1
σ 2 σ 1 σ σ
�
=
r
2
σrr σr
+
r
2
sin β σβ sin βσβ.
(1.19)
Applying Poisson’s equation to equation (1.18), we find that the Mestel disk has zero density every-
where except β
=
ψ/2, where the density is infinite. The more useful (and in this case finite) quantity
is the surface density, determined by integrating the density over the (infinitesimal) thickness of the
disk:
2
Ψ(r)
=
δ(r, z)dz
=
c
.
(1.20)
2ψGr
Like the spherical model for a flat rotation curve, the disk model also suffers from an infinite total
mass and radius, but it is still useful for many applications.
Observations show that most galaxies have elements of both of the above models, with a roughly
spherical bulge and a flattened disk. The relative weight of these features (the “disk-to-bulge ratio,”
Figure 1.9) is a useful parameter for classifying galaxies, with disk strength correlating positively
with other galaxy properties such as young stellar fraction, hydrogen fraction, CO gas, dust, and the
strength of spiral arms. In most galaxies the spiral arms, which are located within the disk are the
primary location for new star formation. The disk and bulge components of most galaxies appear to
have had rather different histories, a subject to be raised again in the context of galaxy formation.
Soon after the realization that many nebulae were in fact distinct galaxies, astronomers (and
in particular Hubble) began systematically classifying galaxies in the hope that taxonomy would
lead to further understanding. The principal component of Hubble’s classification scheme is strongly
correlated with disk-to-bulge ratio. But Hubble conflated the relative proportions of bulge and disk
1
1
2
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
D/B = 0 D/B > 1 D/B < 1 D/B =
Figure 1.9: Galaxies with varying disk to bulge ratios, seen edge on.
Figure 1.10: Two spiral galaxies, floc-
culent to the left and grand design to
the right. Flocculent systems are more
likely to be classified Sa; grand designs
are more frequently Sc.
with the prominence of spiral arms within the disk. In Hubble’s scheme most galaxies were classified
as “E”, “Sa”, “Sb” or “Sc.” “E” stood for elliptical with no sign of spiral structure. “S” stood for
spiral structure, with “Sa” galaxies showing the least spiral structure and “Sc” galaxies showing the
most, as shown in Figure 1.10. The sequence from E to Sc is approximately one of increasing disk-to-
bulge ratio. But there are also galaxies with prominent disks and little evidence for spiral structure.
Hubble added the S0 classification, placing it between E and Sa. Modern versions of Hubble’s
scheme have put the S0 systems parallel to the S systems, with S0/a, S0/b and S0/c systems having
increasingly large disk-to-bulge ratios. One might hope that disk-to-bulge ratio would completely
replace Hubble’s system, but while easy to describe, disk-to-bulge ratio can be difficult to measure,
particularly for systems that are face-on. Modern “quantitative” classification schemes often use
central concentration as a proxy for disk-to-bulge ratio and use deviation from bilateral symmetry
as a proxy for spiral structure.
8
1.2.
BUILDING GALAXIES
7
ij
Building galaxies
As we saw above, gravitational scattering plays a very small role in the dynamics of stellar orbits
within galaxies or globular clusters.
We return to our original question, “what is a galaxy?”
In
light of the previous section, it seems better to think of a galaxy as a collisionless fluid of stars, a
superposition of a large number of orbits guided by a single background potential.
How does one put together such a galaxy? Superpose orbits! Schwarzschild’s method (see Figure
1.2) follows these simple steps:
•
choose gravitational potential
•
generate an orbit library
•
populate potential with orbits
•
check for self-consistency
gravitational potential
orbit library
light = DF *footprints
orbit library
self−consistent
Figure 1.2: Schwarzschild’s method for
only if
building galaxies. One adjusts the distri-
light ~ mass
bution function (DF) to match the light
distribution or the gravitational poten-
tial.
For example, imagine building a spherically symmetric galaxy. How would the orbits be dis-
tributed in phase space? Even though they may occupy three-dimensional space quite uniformly,
there is no reason to insist that the orbits should be uniformly distributed in momentum space. A
galaxy made only of circular orbits would appear “hot” in the tangential direction and “cold” in the
radial direction, while a collection of radial orbits (an equally valid approach to a spherical system)
would look hot in the radial direction and cold in the tangential direction (see Figure 1.3).
Figure 1.3: Two schemes for constructing a spher-
ical galaxy: out of circular orbits, left, and out of
(nearly) radial orbits, right.
“Hot” and “cold” can be understood quantitatively by defining a symmetric velocity dispersion
tensor:
ε
2
= (v
i
−
v
i
)(v
j
−
v
j
),
(1.14)
where v
i
is the velocity in the
ˆ
e
i
direction averaged at some point. For a collisionless fluid, we can
diagonalize the velocity dispersion tensor, giving three independent temperatures at each point in
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
~
�
1
0.8
0.6
0.4
0.2
0
−0.2
−0.4
−0.6
−0.8
−1
−1
−0.8 −0.6 −0.4 −0.2
0
0.2
0.4
0.6
0.8
1
Figure 1.4: The orbit of a star on a
spherically symmetric logarithmic po-
tential. The position of the star is plot-
ted for the first 100 time steps. By fol-
lowing the star for many orbits we accu-
mulate a probability density “footprint”
that may be used to generate a galaxy
model.
Figure 1.5: Left: a disk in circular or-
bit around a spherical bulge. Right:
a disk in polar orbit around a triaxial
bulge. Such “polar ring” galaxies con
stitute perhaps 0.1% of all galaxies.
space. If the random velocities are greater than (or in the same order as) the ordered velocities, the
fluid is hot ; if the ordered velocities are much greater, the fluid is cold.
Observations (Figure 1.7) show that many galaxies have nearly constant rotational velocities
with v(r)
v
c
over the observable range of radius. These flat rotation curves are a departure from
Keplerian orbital velocities, which would scale as v(r)
r
−
1/2
. Flat rotation curves imply mass
increasing linearly with radius, and so cannot continue without bound.
For a spherically symmetric system, the centripetal acceleration is given by
v
2
=
dΛ
(1.15)
r
dr
For a constant rotation velocity v(r)
=
v
c
, the solution for the potential is
Λ
=
v
2
ln r
c
r
ref
.
(1.16)
Combining with Poisson’s equation
�
2
Λ
=
4ψGδ, we find an expression for the density profile:
v
2
δ(r)
=
c
.
(1.17)
4ψGr
2
1.2.
BUILDING GALAXIES
9
Telescope
Focal Plane
Collimator
Grism
Camera
Detector
Figure 1.6: A schematic representation of a spectrograph.
Light from the telescope is brought to a focus in the focal
plane. The slit of the spectrograph isolates the light from a
narrow strip of the sky, which diverges and passes through
Spectrometer the collimator. Parallel light rays from the collimator enter
the grism (a grating ruled on a prism) which disperses the
different wavelengths. The camera brings the light to a focus
on the detector.
galaxy
[N II]
H
[N II]
red
[SII]
blue
[SII]
Figure 1.7: Left: The slit of a spectrometer is placed along the major axis of a disk galaxy. The disk
is highly inclined to the line of sight. Stars and gas on the left side of the slit have a line of sight
velocity toward us (relative to the center of the galaxy). Right: The spectrum of the galaxy. The
lines shown are typical of gas photo-ionized by hot stars. The lines are blueshifted on the left side
and redshifted on the right side. This gives the rotation curve of the galaxy. The rotation curve is
flat in the outer parts of the galaxy and roughly linear in the inner parts.
spectrograph slit
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
�
v
z /rref
r /rref Figure 1.8: Equipotentials for Mestel’s Disk
Among the obvious shortcomings of this model are the failure of the potential to go to zero at large
radius and the prediction of infinite mass. Notwithstanding, it is still a useful potential for many
practical applications.
Another weakness of this model is that it is limited to spherically symmetric galaxies and we
know from experience that many galaxies are shaped more like flattened disks. A corresponding
model for such a galaxy is called a Mestel disk, which is azimuthally symmetric around the z-axis.
The potential for the Mestel disk is
Λ(r, β)
=
v
2
�
ln r
+
ln 1
+
|
cos β
|
�
,
(1.18)
c
r
ref
2
where β is measured from the positive z-axis. The Laplacian in spherical geometry with azimuthal
symmetry is given by
2 1
σ 2 σ 1 σ σ
�
=
r
2
σrr σr
+
r
2
sin β σβ sin βσβ.
(1.19)
Applying Poisson’s equation to equation (1.18), we find that the Mestel disk has zero density every-
where except β
=
ψ/2, where the density is infinite. The more useful (and in this case finite) quantity
is the surface density, determined by integrating the density over the (infinitesimal) thickness of the
disk:
2
Ψ(r)
=
δ(r, z)dz
=
c
.
(1.20)
2ψGr
Like the spherical model for a flat rotation curve, the disk model also suffers from an infinite total
mass and radius, but it is still useful for many applications.
Observations show that most galaxies have elements of both of the above models, with a roughly
spherical bulge and a flattened disk. The relative weight of these features (the “disk-to-bulge ratio,”
Figure 1.9) is a useful parameter for classifying galaxies, with disk strength correlating positively
with other galaxy properties such as young stellar fraction, hydrogen fraction, CO gas, dust, and the
strength of spiral arms. In most galaxies the spiral arms, which are located within the disk are the
primary location for new star formation. The disk and bulge components of most galaxies appear to
have had rather different histories, a subject to be raised again in the context of galaxy formation.
Soon after the realization that many nebulae were in fact distinct galaxies, astronomers (and
in particular Hubble) began systematically classifying galaxies in the hope that taxonomy would
lead to further understanding. The principal component of Hubble’s classification scheme is strongly
correlated with disk-to-bulge ratio. But Hubble conflated the relative proportions of bulge and disk
1
1
2
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
D/B = 0 D/B > 1 D/B < 1 D/B =
Figure 1.9: Galaxies with varying disk to bulge ratios, seen edge on.
Figure 1.10: Two spiral galaxies, floc-
culent to the left and grand design to
the right. Flocculent systems are more
likely to be classified Sa; grand designs
are more frequently Sc.
1.6
with the prominence of spiral arms within the disk. In Hubble’s scheme most galaxies were classified
as “E”, “Sa”, “Sb” or “Sc.” “E” stood for elliptical with no sign of spiral structure. “S” stood for spiral
structure, with “Sa” galaxies showing the least spiral structure and “Sc” galaxies showing the most,
as shown in Figure 1.10. The sequence from E to Sc is approximately one of increasing disk-to- bulge
ratio. But there are also galaxies with prominent disks and little evidence for spiral structure. Hubble
added the S0 classification, placing it between E and Sa. Modern versions of Hubble’s scheme have
put the S0 systems parallel to the S systems, with S0/a, S0/b and S0/c systems having increasingly
large disk-to-bulge ratios. One might hope that disk-to-bulge ratio would completely replace
Hubble’s system, but while easy to describe, disk-to-bulge ratio can be difficult to measure,
particularly for systems that are face-on. Modern “quantitative” classification schemes often use
central concentration as a proxy for disk-to-bulge ratio and use deviation from bilateral symmetry as
a proxy for spiral structure.
8
1.2.
BUILDING GALAXIES
7
ij
Building galaxies
As we saw above, gravitational scattering plays a very small role in the dynamics of stellar orbits
within galaxies or globular clusters.
We return to our original question, “what is a galaxy?”
In
light of the previous section, it seems better to think of a galaxy as a collisionless fluid of stars, a
superposition of a large number of orbits guided by a single background potential.
How does one put together such a galaxy? Superpose orbits! Schwarzschild’s method (see Figure
1.2) follows these simple steps:
•
choose gravitational potential
•
generate an orbit library
•
populate potential with orbits
•
check for self-consistency
gravitational potential
orbit library
light = DF *footprints
orbit library
self−consistent
Figure 1.2: Schwarzschild’s method for
only if
building galaxies. One adjusts the distri-
light ~ mass
bution function (DF) to match the light
distribution or the gravitational poten-
tial.
For example, imagine building a spherically symmetric galaxy. How would the orbits be dis-
tributed in phase space? Even though they may occupy three-dimensional space quite uniformly,
there is no reason to insist that the orbits should be uniformly distributed in momentum space. A
galaxy made only of circular orbits would appear “hot” in the tangential direction and “cold” in the
radial direction, while a collection of radial orbits (an equally valid approach to a spherical system)
would look hot in the radial direction and cold in the tangential direction (see Figure 1.3).
Figure 1.3: Two schemes for constructing a spher-
ical galaxy: out of circular orbits, left, and out of
(nearly) radial orbits, right.
“Hot” and “cold” can be understood quantitatively by defining a symmetric velocity dispersion
tensor:
ε
2
= (v
i
−
v
i
)(v
j
−
v
j
),
(1.14)
where v
i
is the velocity in the
ˆ
e
i
direction averaged at some point. For a collisionless fluid, we can
diagonalize the velocity dispersion tensor, giving three independent temperatures at each point in
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
~
�
1
0.8
0.6
0.4
0.2
0
−0.2
−0.4
−0.6
−0.8
−1
−1
−0.8 −0.6 −0.4 −0.2
0
0.2
0.4
0.6
0.8
1
Figure 1.4: The orbit of a star on a
spherically symmetric logarithmic po-
tential. The position of the star is plot-
ted for the first 100 time steps. By fol-
lowing the star for many orbits we accu-
mulate a probability density “footprint”
that may be used to generate a galaxy
model.
Figure 1.5: Left: a disk in circular or-
bit around a spherical bulge. Right:
a disk in polar orbit around a triaxial
bulge. Such “polar ring” galaxies con
stitute perhaps 0.1% of all galaxies.
space. If the random velocities are greater than (or in the same order as) the ordered velocities, the
fluid is hot ; if the ordered velocities are much greater, the fluid is cold.
Observations (Figure 1.7) show that many galaxies have nearly constant rotational velocities
with v(r)
v
c
over the observable range of radius. These flat rotation curves are a departure from
Keplerian orbital velocities, which would scale as v(r)
r
−
1/2
. Flat rotation curves imply mass
increasing linearly with radius, and so cannot continue without bound.
For a spherically symmetric system, the centripetal acceleration is given by
v
2
=
dΛ
(1.15)
r
dr
For a constant rotation velocity v(r)
=
v
c
, the solution for the potential is
Λ
=
v
2
ln r
c
r
ref
.
(1.16)
Combining with Poisson’s equation
�
2
Λ
=
4ψGδ, we find an expression for the density profile:
v
2
δ(r)
=
c
.
(1.17)
4ψGr
2
1.2.
BUILDING GALAXIES
9
Telescope
Focal Plane
Collimator
Grism
Camera
Detector
Figure 1.6: A schematic representation of a spectrograph.
Light from the telescope is brought to a focus in the focal
plane. The slit of the spectrograph isolates the light from a
narrow strip of the sky, which diverges and passes through
Spectrometer the collimator. Parallel light rays from the collimator enter
the grism (a grating ruled on a prism) which disperses the
different wavelengths. The camera brings the light to a focus
on the detector.
galaxy
[N II]
H
[N II]
red
[SII]
blue
[SII]
Figure 1.7: Left: The slit of a spectrometer is placed along the major axis of a disk galaxy. The disk
is highly inclined to the line of sight. Stars and gas on the left side of the slit have a line of sight
velocity toward us (relative to the center of the galaxy). Right: The spectrum of the galaxy. The
lines shown are typical of gas photo-ionized by hot stars. The lines are blueshifted on the left side
and redshifted on the right side. This gives the rotation curve of the galaxy. The rotation curve is
flat in the outer parts of the galaxy and roughly linear in the inner parts.
spectrograph slit
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
�
v
z /rref
r /rref Figure 1.8: Equipotentials for Mestel’s Disk
Among the obvious shortcomings of this model are the failure of the potential to go to zero at large
radius and the prediction of infinite mass. Notwithstanding, it is still a useful potential for many
practical applications.
Another weakness of this model is that it is limited to spherically symmetric galaxies and we
know from experience that many galaxies are shaped more like flattened disks. A corresponding
model for such a galaxy is called a Mestel disk, which is azimuthally symmetric around the z-axis.
The potential for the Mestel disk is
Λ(r, β)
=
v
2
�
ln r
+
ln 1
+
|
cos β
|
�
,
(1.18)
c
r
ref
2
where β is measured from the positive z-axis. The Laplacian in spherical geometry with azimuthal
symmetry is given by
2 1
σ 2 σ 1 σ σ
�
=
r
2
σrr σr
+
r
2
sin β σβ sin βσβ.
(1.19)
Applying Poisson’s equation to equation (1.18), we find that the Mestel disk has zero density every-
where except β
=
ψ/2, where the density is infinite. The more useful (and in this case finite) quantity
is the surface density, determined by integrating the density over the (infinitesimal) thickness of the
disk:
2
Ψ(r)
=
δ(r, z)dz
=
c
.
(1.20)
2ψGr
Like the spherical model for a flat rotation curve, the disk model also suffers from an infinite total
mass and radius, but it is still useful for many applications.
Observations show that most galaxies have elements of both of the above models, with a roughly
spherical bulge and a flattened disk. The relative weight of these features (the “disk-to-bulge ratio,”
Figure 1.9) is a useful parameter for classifying galaxies, with disk strength correlating positively
with other galaxy properties such as young stellar fraction, hydrogen fraction, CO gas, dust, and the
strength of spiral arms. In most galaxies the spiral arms, which are located within the disk are the
primary location for new star formation. The disk and bulge components of most galaxies appear to
have had rather different histories, a subject to be raised again in the context of galaxy formation.
Soon after the realization that many nebulae were in fact distinct galaxies, astronomers (and
in particular Hubble) began systematically classifying galaxies in the hope that taxonomy would
lead to further understanding. The principal component of Hubble’s classification scheme is strongly
correlated with disk-to-bulge ratio. But Hubble conflated the relative proportions of bulge and disk
1
1
2
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
D/B = 0 D/B > 1 D/B < 1 D/B =
Figure 1.9: Galaxies with varying disk to bulge ratios, seen edge on.
Figure 1.10: Two spiral galaxies, floc-
culent to the left and grand design to
the right. Flocculent systems are more
likely to be classified Sa; grand designs
are more frequently Sc.
with the prominence of spiral arms within the disk. In Hubble’s scheme most galaxies were classified
as “E”, “Sa”, “Sb” or “Sc.” “E” stood for elliptical with no sign of spiral structure. “S” stood for
spiral structure, with “Sa” galaxies showing the least spiral structure and “Sc” galaxies showing the
most, as shown in Figure 1.10. The sequence from E to Sc is approximately one of increasing disk-to-
bulge ratio. But there are also galaxies with prominent disks and little evidence for spiral structure.
Hubble added the S0 classification, placing it between E and Sa. Modern versions of Hubble’s
scheme have put the S0 systems parallel to the S systems, with S0/a, S0/b and S0/c systems having
increasingly large disk-to-bulge ratios. One might hope that disk-to-bulge ratio would completely
replace Hubble’s system, but while easy to describe, disk-to-bulge ratio can be difficult to measure,
particularly for systems that are face-on. Modern “quantitative” classification schemes often use
central concentration as a proxy for disk-to-bulge ratio and use deviation from bilateral symmetry
as a proxy for spiral structure.
8
1.2.
BUILDING GALAXIES
7
ij
Building galaxies
As we saw above, gravitational scattering plays a very small role in the dynamics of stellar orbits
within galaxies or globular clusters.
We return to our original question, “what is a galaxy?”
In
light of the previous section, it seems better to think of a galaxy as a collisionless fluid of stars, a
superposition of a large number of orbits guided by a single background potential.
How does one put together such a galaxy? Superpose orbits! Schwarzschild’s method (see Figure
1.2) follows these simple steps:
•
choose gravitational potential
•
generate an orbit library
•
populate potential with orbits
•
check for self-consistency
gravitational potential
orbit library
light = DF *footprints
orbit library
self−consistent
Figure 1.2: Schwarzschild’s method for
only if
building galaxies. One adjusts the distri-
light ~ mass
bution function (DF) to match the light
distribution or the gravitational poten-
tial.
For example, imagine building a spherically symmetric galaxy. How would the orbits be dis-
tributed in phase space? Even though they may occupy three-dimensional space quite uniformly,
there is no reason to insist that the orbits should be uniformly distributed in momentum space. A
galaxy made only of circular orbits would appear “hot” in the tangential direction and “cold” in the
radial direction, while a collection of radial orbits (an equally valid approach to a spherical system)
would look hot in the radial direction and cold in the tangential direction (see Figure 1.3).
Figure 1.3: Two schemes for constructing a spher-
ical galaxy: out of circular orbits, left, and out of
(nearly) radial orbits, right.
“Hot” and “cold” can be understood quantitatively by defining a symmetric velocity dispersion
tensor:
ε
2
= (v
i
−
v
i
)(v
j
−
v
j
),
(1.14)
where v
i
is the velocity in the
ˆ
e
i
direction averaged at some point. For a collisionless fluid, we can
diagonalize the velocity dispersion tensor, giving three independent temperatures at each point in
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
~
�
1
0.8
0.6
0.4
0.2
0
−0.2
−0.4
−0.6
−0.8
−1
−1
−0.8 −0.6 −0.4 −0.2
0
0.2
0.4
0.6
0.8
1
Figure 1.4: The orbit of a star on a
spherically symmetric logarithmic po-
tential. The position of the star is plot-
ted for the first 100 time steps. By fol-
lowing the star for many orbits we accu-
mulate a probability density “footprint”
that may be used to generate a galaxy
model.
Figure 1.5: Left: a disk in circular or-
bit around a spherical bulge. Right:
a disk in polar orbit around a triaxial
bulge. Such “polar ring” galaxies con
stitute perhaps 0.1% of all galaxies.
space. If the random velocities are greater than (or in the same order as) the ordered velocities, the
fluid is hot ; if the ordered velocities are much greater, the fluid is cold.
Observations (Figure 1.7) show that many galaxies have nearly constant rotational velocities
with v(r)
v
c
over the observable range of radius. These flat rotation curves are a departure from
Keplerian orbital velocities, which would scale as v(r)
r
−
1/2
. Flat rotation curves imply mass
increasing linearly with radius, and so cannot continue without bound.
For a spherically symmetric system, the centripetal acceleration is given by
v
2
=
dΛ
(1.15)
r
dr
For a constant rotation velocity v(r)
=
v
c
, the solution for the potential is
Λ
=
v
2
ln r
c
r
ref
.
(1.16)
Combining with Poisson’s equation
�
2
Λ
=
4ψGδ, we find an expression for the density profile:
v
2
δ(r)
=
c
.
(1.17)
4ψGr
2
1.2.
BUILDING GALAXIES
9
Telescope
Focal Plane
Collimator
Grism
Camera
Detector
Figure 1.6: A schematic representation of a spectrograph.
Light from the telescope is brought to a focus in the focal
plane. The slit of the spectrograph isolates the light from a
narrow strip of the sky, which diverges and passes through
Spectrometer the collimator. Parallel light rays from the collimator enter
the grism (a grating ruled on a prism) which disperses the
different wavelengths. The camera brings the light to a focus
on the detector.
galaxy
[N II]
H
[N II]
red
[SII]
blue
[SII]
Figure 1.7: Left: The slit of a spectrometer is placed along the major axis of a disk galaxy. The disk
is highly inclined to the line of sight. Stars and gas on the left side of the slit have a line of sight
velocity toward us (relative to the center of the galaxy). Right: The spectrum of the galaxy. The
lines shown are typical of gas photo-ionized by hot stars. The lines are blueshifted on the left side
and redshifted on the right side. This gives the rotation curve of the galaxy. The rotation curve is
flat in the outer parts of the galaxy and roughly linear in the inner parts.
spectrograph slit
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
�
v
z /rref
r /rref Figure 1.8: Equipotentials for Mestel’s Disk
Among the obvious shortcomings of this model are the failure of the potential to go to zero at large
radius and the prediction of infinite mass. Notwithstanding, it is still a useful potential for many
practical applications.
Another weakness of this model is that it is limited to spherically symmetric galaxies and we
know from experience that many galaxies are shaped more like flattened disks. A corresponding
model for such a galaxy is called a Mestel disk, which is azimuthally symmetric around the z-axis.
The potential for the Mestel disk is
Λ(r, β)
=
v
2
�
ln r
+
ln 1
+
|
cos β
|
�
,
(1.18)
c
r
ref
2
where β is measured from the positive z-axis. The Laplacian in spherical geometry with azimuthal
symmetry is given by
2 1
σ 2 σ 1 σ σ
�
=
r
2
σrr σr
+
r
2
sin β σβ sin βσβ.
(1.19)
Applying Poisson’s equation to equation (1.18), we find that the Mestel disk has zero density every-
where except β
=
ψ/2, where the density is infinite. The more useful (and in this case finite) quantity
is the surface density, determined by integrating the density over the (infinitesimal) thickness of the
disk:
2
Ψ(r)
=
δ(r, z)dz
=
c
.
(1.20)
2ψGr
Like the spherical model for a flat rotation curve, the disk model also suffers from an infinite total
mass and radius, but it is still useful for many applications.
Observations show that most galaxies have elements of both of the above models, with a roughly
spherical bulge and a flattened disk. The relative weight of these features (the “disk-to-bulge ratio,”
Figure 1.9) is a useful parameter for classifying galaxies, with disk strength correlating positively
with other galaxy properties such as young stellar fraction, hydrogen fraction, CO gas, dust, and the
strength of spiral arms. In most galaxies the spiral arms, which are located within the disk are the
primary location for new star formation. The disk and bulge components of most galaxies appear to
have had rather different histories, a subject to be raised again in the context of galaxy formation.
Soon after the realization that many nebulae were in fact distinct galaxies, astronomers (and
in particular Hubble) began systematically classifying galaxies in the hope that taxonomy would
lead to further understanding. The principal component of Hubble’s classification scheme is strongly
correlated with disk-to-bulge ratio. But Hubble conflated the relative proportions of bulge and disk
1
1
2
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
D/B = 0 D/B > 1 D/B < 1 D/B =
Figure 1.9: Galaxies with varying disk to bulge ratios, seen edge on.
Figure 1.10: Two spiral galaxies, floc-
culent to the left and grand design to
the right. Flocculent systems are more
likely to be classified Sa; grand designs
are more frequently Sc.
1.7
with the prominence of spiral arms within the disk. In Hubble’s scheme most galaxies were classified
as “E”, “Sa”, “Sb” or “Sc.” “E” stood for elliptical with no sign of spiral structure. “S” stood for spiral
structure, with “Sa” galaxies showing the least spiral structure and “Sc” galaxies showing the most,
as shown in Figure 1.10. The sequence from E to Sc is approximately one of increasing disk-to- bulge
ratio. But there are also galaxies with prominent disks and little evidence for spiral structure. Hubble
added the S0 classification, placing it between E and Sa. Modern versions of Hubble’s scheme have
put the S0 systems parallel to the S systems, with S0/a, S0/b and S0/c systems having increasingly
large disk-to-bulge ratios. One might hope that disk-to-bulge ratio would completely replace
Hubble’s system, but while easy to describe, disk-to-bulge ratio can be difficult to measure,
particularly for systems that are face-on. Modern “quantitative” classification schemes often use
central concentration as a proxy for disk-to-bulge ratio and use deviation from bilateral symmetry as
a proxy for spiral structure.
8
1.2.
BUILDING GALAXIES
7
ij
Building galaxies
As we saw above, gravitational scattering plays a very small role in the dynamics of stellar orbits
within galaxies or globular clusters.
We return to our original question, “what is a galaxy?”
In
light of the previous section, it seems better to think of a galaxy as a collisionless fluid of stars, a
superposition of a large number of orbits guided by a single background potential.
How does one put together such a galaxy? Superpose orbits! Schwarzschild’s method (see Figure
1.2) follows these simple steps:
•
choose gravitational potential
•
generate an orbit library
•
populate potential with orbits
•
check for self-consistency
gravitational potential
orbit library
light = DF *footprints
orbit library
self−consistent
Figure 1.2: Schwarzschild’s method for
only if
building galaxies. One adjusts the distri-
light ~ mass
bution function (DF) to match the light
distribution or the gravitational poten-
tial.
For example, imagine building a spherically symmetric galaxy. How would the orbits be dis-
tributed in phase space? Even though they may occupy three-dimensional space quite uniformly,
there is no reason to insist that the orbits should be uniformly distributed in momentum space. A
galaxy made only of circular orbits would appear “hot” in the tangential direction and “cold” in the
radial direction, while a collection of radial orbits (an equally valid approach to a spherical system)
would look hot in the radial direction and cold in the tangential direction (see Figure 1.3).
Figure 1.3: Two schemes for constructing a spher-
ical galaxy: out of circular orbits, left, and out of
(nearly) radial orbits, right.
“Hot” and “cold” can be understood quantitatively by defining a symmetric velocity dispersion
tensor:
ε
2
= (v
i
−
v
i
)(v
j
−
v
j
),
(1.14)
where v
i
is the velocity in the
ˆ
e
i
direction averaged at some point. For a collisionless fluid, we can
diagonalize the velocity dispersion tensor, giving three independent temperatures at each point in
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
~
�
1
0.8
0.6
0.4
0.2
0
−0.2
−0.4
−0.6
−0.8
−1
−1
−0.8 −0.6 −0.4 −0.2
0
0.2
0.4
0.6
0.8
1
Figure 1.4: The orbit of a star on a
spherically symmetric logarithmic po-
tential. The position of the star is plot-
ted for the first 100 time steps. By fol-
lowing the star for many orbits we accu-
mulate a probability density “footprint”
that may be used to generate a galaxy
model.
Figure 1.5: Left: a disk in circular or-
bit around a spherical bulge. Right:
a disk in polar orbit around a triaxial
bulge. Such “polar ring” galaxies con
stitute perhaps 0.1% of all galaxies.
space. If the random velocities are greater than (or in the same order as) the ordered velocities, the
fluid is hot ; if the ordered velocities are much greater, the fluid is cold.
Observations (Figure 1.7) show that many galaxies have nearly constant rotational velocities
with v(r)
v
c
over the observable range of radius. These flat rotation curves are a departure from
Keplerian orbital velocities, which would scale as v(r)
r
−
1/2
. Flat rotation curves imply mass
increasing linearly with radius, and so cannot continue without bound.
For a spherically symmetric system, the centripetal acceleration is given by
v
2
=
dΛ
(1.15)
r
dr
For a constant rotation velocity v(r)
=
v
c
, the solution for the potential is
Λ
=
v
2
ln r
c
r
ref
.
(1.16)
Combining with Poisson’s equation
�
2
Λ
=
4ψGδ, we find an expression for the density profile:
v
2
δ(r)
=
c
.
(1.17)
4ψGr
2
1.2.
BUILDING GALAXIES
9
Telescope
Focal Plane
Collimator
Grism
Camera
Detector
Figure 1.6: A schematic representation of a spectrograph.
Light from the telescope is brought to a focus in the focal
plane. The slit of the spectrograph isolates the light from a
narrow strip of the sky, which diverges and passes through
Spectrometer the collimator. Parallel light rays from the collimator enter
the grism (a grating ruled on a prism) which disperses the
different wavelengths. The camera brings the light to a focus
on the detector.
galaxy
[N II]
H
[N II]
red
[SII]
blue
[SII]
Figure 1.7: Left: The slit of a spectrometer is placed along the major axis of a disk galaxy. The disk
is highly inclined to the line of sight. Stars and gas on the left side of the slit have a line of sight
velocity toward us (relative to the center of the galaxy). Right: The spectrum of the galaxy. The
lines shown are typical of gas photo-ionized by hot stars. The lines are blueshifted on the left side
and redshifted on the right side. This gives the rotation curve of the galaxy. The rotation curve is
flat in the outer parts of the galaxy and roughly linear in the inner parts.
spectrograph slit
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
�
v
z /rref
r /rref Figure 1.8: Equipotentials for Mestel’s Disk
Among the obvious shortcomings of this model are the failure of the potential to go to zero at large
radius and the prediction of infinite mass. Notwithstanding, it is still a useful potential for many
practical applications.
Another weakness of this model is that it is limited to spherically symmetric galaxies and we
know from experience that many galaxies are shaped more like flattened disks. A corresponding
model for such a galaxy is called a Mestel disk, which is azimuthally symmetric around the z-axis.
The potential for the Mestel disk is
Λ(r, β)
=
v
2
�
ln r
+
ln 1
+
|
cos β
|
�
,
(1.18)
c
r
ref
2
where β is measured from the positive z-axis. The Laplacian in spherical geometry with azimuthal
symmetry is given by
2 1
σ 2 σ 1 σ σ
�
=
r
2
σrr σr
+
r
2
sin β σβ sin βσβ.
(1.19)
Applying Poisson’s equation to equation (1.18), we find that the Mestel disk has zero density every-
where except β
=
ψ/2, where the density is infinite. The more useful (and in this case finite) quantity
is the surface density, determined by integrating the density over the (infinitesimal) thickness of the
disk:
2
Ψ(r)
=
δ(r, z)dz
=
c
.
(1.20)
2ψGr
Like the spherical model for a flat rotation curve, the disk model also suffers from an infinite total
mass and radius, but it is still useful for many applications.
Observations show that most galaxies have elements of both of the above models, with a roughly
spherical bulge and a flattened disk. The relative weight of these features (the “disk-to-bulge ratio,”
Figure 1.9) is a useful parameter for classifying galaxies, with disk strength correlating positively
with other galaxy properties such as young stellar fraction, hydrogen fraction, CO gas, dust, and the
strength of spiral arms. In most galaxies the spiral arms, which are located within the disk are the
primary location for new star formation. The disk and bulge components of most galaxies appear to
have had rather different histories, a subject to be raised again in the context of galaxy formation.
Soon after the realization that many nebulae were in fact distinct galaxies, astronomers (and
in particular Hubble) began systematically classifying galaxies in the hope that taxonomy would
lead to further understanding. The principal component of Hubble’s classification scheme is strongly
correlated with disk-to-bulge ratio. But Hubble conflated the relative proportions of bulge and disk
1
1
2
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
D/B = 0 D/B > 1 D/B < 1 D/B =
Figure 1.9: Galaxies with varying disk to bulge ratios, seen edge on.
Figure 1.10: Two spiral galaxies, floc-
culent to the left and grand design to
the right. Flocculent systems are more
likely to be classified Sa; grand designs
are more frequently Sc.
with the prominence of spiral arms within the disk. In Hubble’s scheme most galaxies were classified
as “E”, “Sa”, “Sb” or “Sc.” “E” stood for elliptical with no sign of spiral structure. “S” stood for
spiral structure, with “Sa” galaxies showing the least spiral structure and “Sc” galaxies showing the
most, as shown in Figure 1.10. The sequence from E to Sc is approximately one of increasing disk-to-
bulge ratio. But there are also galaxies with prominent disks and little evidence for spiral structure.
Hubble added the S0 classification, placing it between E and Sa. Modern versions of Hubble’s
scheme have put the S0 systems parallel to the S systems, with S0/a, S0/b and S0/c systems having
increasingly large disk-to-bulge ratios. One might hope that disk-to-bulge ratio would completely
replace Hubble’s system, but while easy to describe, disk-to-bulge ratio can be difficult to measure,
particularly for systems that are face-on. Modern “quantitative” classification schemes often use
central concentration as a proxy for disk-to-bulge ratio and use deviation from bilateral symmetry
as a proxy for spiral structure.
8
1.2.
BUILDING GALAXIES
7
ij
Building galaxies
As we saw above, gravitational scattering plays a very small role in the dynamics of stellar orbits
within galaxies or globular clusters.
We return to our original question, “what is a galaxy?”
In
light of the previous section, it seems better to think of a galaxy as a collisionless fluid of stars, a
superposition of a large number of orbits guided by a single background potential.
How does one put together such a galaxy? Superpose orbits! Schwarzschild’s method (see Figure
1.2) follows these simple steps:
•
choose gravitational potential
•
generate an orbit library
•
populate potential with orbits
•
check for self-consistency
gravitational potential
orbit library
light = DF *footprints
orbit library
self−consistent
Figure 1.2: Schwarzschild’s method for
only if
building galaxies. One adjusts the distri-
light ~ mass
bution function (DF) to match the light
distribution or the gravitational poten-
tial.
For example, imagine building a spherically symmetric galaxy. How would the orbits be dis-
tributed in phase space? Even though they may occupy three-dimensional space quite uniformly,
there is no reason to insist that the orbits should be uniformly distributed in momentum space. A
galaxy made only of circular orbits would appear “hot” in the tangential direction and “cold” in the
radial direction, while a collection of radial orbits (an equally valid approach to a spherical system)
would look hot in the radial direction and cold in the tangential direction (see Figure 1.3).
Figure 1.3: Two schemes for constructing a spher-
ical galaxy: out of circular orbits, left, and out of
(nearly) radial orbits, right.
“Hot” and “cold” can be understood quantitatively by defining a symmetric velocity dispersion
tensor:
ε
2
= (v
i
−
v
i
)(v
j
−
v
j
),
(1.14)
where v
i
is the velocity in the
ˆ
e
i
direction averaged at some point. For a collisionless fluid, we can
diagonalize the velocity dispersion tensor, giving three independent temperatures at each point in
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
~
�
1
0.8
0.6
0.4
0.2
0
−0.2
−0.4
−0.6
−0.8
−1
−1
−0.8 −0.6 −0.4 −0.2
0
0.2
0.4
0.6
0.8
1
Figure 1.4: The orbit of a star on a
spherically symmetric logarithmic po-
tential. The position of the star is plot-
ted for the first 100 time steps. By fol-
lowing the star for many orbits we accu-
mulate a probability density “footprint”
that may be used to generate a galaxy
model.
Figure 1.5: Left: a disk in circular or-
bit around a spherical bulge. Right:
a disk in polar orbit around a triaxial
bulge. Such “polar ring” galaxies con
stitute perhaps 0.1% of all galaxies.
space. If the random velocities are greater than (or in the same order as) the ordered velocities, the
fluid is hot ; if the ordered velocities are much greater, the fluid is cold.
Observations (Figure 1.7) show that many galaxies have nearly constant rotational velocities
with v(r)
v
c
over the observable range of radius. These flat rotation curves are a departure from
Keplerian orbital velocities, which would scale as v(r)
r
−
1/2
. Flat rotation curves imply mass
increasing linearly with radius, and so cannot continue without bound.
For a spherically symmetric system, the centripetal acceleration is given by
v
2
=
dΛ
(1.15)
r
dr
For a constant rotation velocity v(r)
=
v
c
, the solution for the potential is
Λ
=
v
2
ln r
c
r
ref
.
(1.16)
Combining with Poisson’s equation
�
2
Λ
=
4ψGδ, we find an expression for the density profile:
v
2
δ(r)
=
c
.
(1.17)
4ψGr
2
1.2.
BUILDING GALAXIES
9
Telescope
Focal Plane
Collimator
Grism
Camera
Detector
Figure 1.6: A schematic representation of a spectrograph.
Light from the telescope is brought to a focus in the focal
plane. The slit of the spectrograph isolates the light from a
narrow strip of the sky, which diverges and passes through
Spectrometer the collimator. Parallel light rays from the collimator enter
the grism (a grating ruled on a prism) which disperses the
different wavelengths. The camera brings the light to a focus
on the detector.
galaxy
[N II]
H
[N II]
red
[SII]
blue
[SII]
Figure 1.7: Left: The slit of a spectrometer is placed along the major axis of a disk galaxy. The disk
is highly inclined to the line of sight. Stars and gas on the left side of the slit have a line of sight
velocity toward us (relative to the center of the galaxy). Right: The spectrum of the galaxy. The
lines shown are typical of gas photo-ionized by hot stars. The lines are blueshifted on the left side
and redshifted on the right side. This gives the rotation curve of the galaxy. The rotation curve is
flat in the outer parts of the galaxy and roughly linear in the inner parts.
spectrograph slit
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
�
v
z /rref
r /rref Figure 1.8: Equipotentials for Mestel’s Disk
Among the obvious shortcomings of this model are the failure of the potential to go to zero at large
radius and the prediction of infinite mass. Notwithstanding, it is still a useful potential for many
practical applications.
Another weakness of this model is that it is limited to spherically symmetric galaxies and we
know from experience that many galaxies are shaped more like flattened disks. A corresponding
model for such a galaxy is called a Mestel disk, which is azimuthally symmetric around the z-axis.
The potential for the Mestel disk is
Λ(r, β)
=
v
2
�
ln r
+
ln 1
+
|
cos β
|
�
,
(1.18)
c
r
ref
2
where β is measured from the positive z-axis. The Laplacian in spherical geometry with azimuthal
symmetry is given by
2 1
σ 2 σ 1 σ σ
�
=
r
2
σrr σr
+
r
2
sin β σβ sin βσβ.
(1.19)
Applying Poisson’s equation to equation (1.18), we find that the Mestel disk has zero density every-
where except β
=
ψ/2, where the density is infinite. The more useful (and in this case finite) quantity
is the surface density, determined by integrating the density over the (infinitesimal) thickness of the
disk:
2
Ψ(r)
=
δ(r, z)dz
=
c
.
(1.20)
2ψGr
Like the spherical model for a flat rotation curve, the disk model also suffers from an infinite total
mass and radius, but it is still useful for many applications.
Observations show that most galaxies have elements of both of the above models, with a roughly
spherical bulge and a flattened disk. The relative weight of these features (the “disk-to-bulge ratio,”
Figure 1.9) is a useful parameter for classifying galaxies, with disk strength correlating positively
with other galaxy properties such as young stellar fraction, hydrogen fraction, CO gas, dust, and the
strength of spiral arms. In most galaxies the spiral arms, which are located within the disk are the
primary location for new star formation. The disk and bulge components of most galaxies appear to
have had rather different histories, a subject to be raised again in the context of galaxy formation.
Soon after the realization that many nebulae were in fact distinct galaxies, astronomers (and
in particular Hubble) began systematically classifying galaxies in the hope that taxonomy would
lead to further understanding. The principal component of Hubble’s classification scheme is strongly
correlated with disk-to-bulge ratio. But Hubble conflated the relative proportions of bulge and disk
1
1
2
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
D/B = 0 D/B > 1 D/B < 1 D/B =
Figure 1.9: Galaxies with varying disk to bulge ratios, seen edge on.
Figure 1.10: Two spiral galaxies, floc-
culent to the left and grand design to
the right. Flocculent systems are more
likely to be classified Sa; grand designs
are more frequently Sc.
1.8
with the prominence of spiral arms within the disk. In Hubble’s scheme most galaxies were classified
as “E”, “Sa”, “Sb” or “Sc.” “E” stood for elliptical with no sign of spiral structure. “S” stood for spiral
structure, with “Sa” galaxies showing the least spiral structure and “Sc” galaxies showing the most,
as shown in Figure 1.10. The sequence from E to Sc is approximately one of increasing disk-to- bulge
ratio. But there are also galaxies with prominent disks and little evidence for spiral structure. Hubble
added the S0 classification, placing it between E and Sa. Modern versions of Hubble’s scheme have
put the S0 systems parallel to the S systems, with S0/a, S0/b and S0/c systems having increasingly
large disk-to-bulge ratios. One might hope that disk-to-bulge ratio would completely replace
Hubble’s system, but while easy to describe, disk-to-bulge ratio can be difficult to measure,
particularly for systems that are face-on. Modern “quantitative” classification schemes often use
central concentration as a proxy for disk-to-bulge ratio and use deviation from bilateral symmetry as
a proxy for spiral structure.
8
1.2.
BUILDING GALAXIES
7
ij
Building galaxies
As we saw above, gravitational scattering plays a very small role in the dynamics of stellar orbits
within galaxies or globular clusters.
We return to our original question, “what is a galaxy?”
In
light of the previous section, it seems better to think of a galaxy as a collisionless fluid of stars, a
superposition of a large number of orbits guided by a single background potential.
How does one put together such a galaxy? Superpose orbits! Schwarzschild’s method (see Figure
1.2) follows these simple steps:
•
choose gravitational potential
•
generate an orbit library
•
populate potential with orbits
•
check for self-consistency
gravitational potential
orbit library
light = DF *footprints
orbit library
self−consistent
Figure 1.2: Schwarzschild’s method for
only if
building galaxies. One adjusts the distri-
light ~ mass
bution function (DF) to match the light
distribution or the gravitational poten-
tial.
For example, imagine building a spherically symmetric galaxy. How would the orbits be dis-
tributed in phase space? Even though they may occupy three-dimensional space quite uniformly,
there is no reason to insist that the orbits should be uniformly distributed in momentum space. A
galaxy made only of circular orbits would appear “hot” in the tangential direction and “cold” in the
radial direction, while a collection of radial orbits (an equally valid approach to a spherical system)
would look hot in the radial direction and cold in the tangential direction (see Figure 1.3).
Figure 1.3: Two schemes for constructing a spher-
ical galaxy: out of circular orbits, left, and out of
(nearly) radial orbits, right.
“Hot” and “cold” can be understood quantitatively by defining a symmetric velocity dispersion
tensor:
ε
2
= (v
i
−
v
i
)(v
j
−
v
j
),
(1.14)
where v
i
is the velocity in the
ˆ
e
i
direction averaged at some point. For a collisionless fluid, we can
diagonalize the velocity dispersion tensor, giving three independent temperatures at each point in
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
~
�
1
0.8
0.6
0.4
0.2
0
−0.2
−0.4
−0.6
−0.8
−1
−1
−0.8 −0.6 −0.4 −0.2
0
0.2
0.4
0.6
0.8
1
Figure 1.4: The orbit of a star on a
spherically symmetric logarithmic po-
tential. The position of the star is plot-
ted for the first 100 time steps. By fol-
lowing the star for many orbits we accu-
mulate a probability density “footprint”
that may be used to generate a galaxy
model.
Figure 1.5: Left: a disk in circular or-
bit around a spherical bulge. Right:
a disk in polar orbit around a triaxial
bulge. Such “polar ring” galaxies con
stitute perhaps 0.1% of all galaxies.
space. If the random velocities are greater than (or in the same order as) the ordered velocities, the
fluid is hot ; if the ordered velocities are much greater, the fluid is cold.
Observations (Figure 1.7) show that many galaxies have nearly constant rotational velocities
with v(r)
v
c
over the observable range of radius. These flat rotation curves are a departure from
Keplerian orbital velocities, which would scale as v(r)
r
−
1/2
. Flat rotation curves imply mass
increasing linearly with radius, and so cannot continue without bound.
For a spherically symmetric system, the centripetal acceleration is given by
v
2
=
dΛ
(1.15)
r
dr
For a constant rotation velocity v(r)
=
v
c
, the solution for the potential is
Λ
=
v
2
ln r
c
r
ref
.
(1.16)
Combining with Poisson’s equation
�
2
Λ
=
4ψGδ, we find an expression for the density profile:
v
2
δ(r)
=
c
.
(1.17)
4ψGr
2
1.2.
BUILDING GALAXIES
9
Telescope
Focal Plane
Collimator
Grism
Camera
Detector
Figure 1.6: A schematic representation of a spectrograph.
Light from the telescope is brought to a focus in the focal
plane. The slit of the spectrograph isolates the light from a
narrow strip of the sky, which diverges and passes through
Spectrometer the collimator. Parallel light rays from the collimator enter
the grism (a grating ruled on a prism) which disperses the
different wavelengths. The camera brings the light to a focus
on the detector.
galaxy
[N II]
H
[N II]
red
[SII]
blue
[SII]
Figure 1.7: Left: The slit of a spectrometer is placed along the major axis of a disk galaxy. The disk
is highly inclined to the line of sight. Stars and gas on the left side of the slit have a line of sight
velocity toward us (relative to the center of the galaxy). Right: The spectrum of the galaxy. The
lines shown are typical of gas photo-ionized by hot stars. The lines are blueshifted on the left side
and redshifted on the right side. This gives the rotation curve of the galaxy. The rotation curve is
flat in the outer parts of the galaxy and roughly linear in the inner parts.
spectrograph slit
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
�
v
z /rref
r /rref Figure 1.8: Equipotentials for Mestel’s Disk
Among the obvious shortcomings of this model are the failure of the potential to go to zero at large
radius and the prediction of infinite mass. Notwithstanding, it is still a useful potential for many
practical applications.
Another weakness of this model is that it is limited to spherically symmetric galaxies and we
know from experience that many galaxies are shaped more like flattened disks. A corresponding
model for such a galaxy is called a Mestel disk, which is azimuthally symmetric around the z-axis.
The potential for the Mestel disk is
Λ(r, β)
=
v
2
�
ln r
+
ln 1
+
|
cos β
|
�
,
(1.18)
c
r
ref
2
where β is measured from the positive z-axis. The Laplacian in spherical geometry with azimuthal
symmetry is given by
2 1
σ 2 σ 1 σ σ
�
=
r
2
σrr σr
+
r
2
sin β σβ sin βσβ.
(1.19)
Applying Poisson’s equation to equation (1.18), we find that the Mestel disk has zero density every-
where except β
=
ψ/2, where the density is infinite. The more useful (and in this case finite) quantity
is the surface density, determined by integrating the density over the (infinitesimal) thickness of the
disk:
2
Ψ(r)
=
δ(r, z)dz
=
c
.
(1.20)
2ψGr
Like the spherical model for a flat rotation curve, the disk model also suffers from an infinite total
mass and radius, but it is still useful for many applications.
Observations show that most galaxies have elements of both of the above models, with a roughly
spherical bulge and a flattened disk. The relative weight of these features (the “disk-to-bulge ratio,”
Figure 1.9) is a useful parameter for classifying galaxies, with disk strength correlating positively
with other galaxy properties such as young stellar fraction, hydrogen fraction, CO gas, dust, and the
strength of spiral arms. In most galaxies the spiral arms, which are located within the disk are the
primary location for new star formation. The disk and bulge components of most galaxies appear to
have had rather different histories, a subject to be raised again in the context of galaxy formation.
Soon after the realization that many nebulae were in fact distinct galaxies, astronomers (and
in particular Hubble) began systematically classifying galaxies in the hope that taxonomy would
lead to further understanding. The principal component of Hubble’s classification scheme is strongly
correlated with disk-to-bulge ratio. But Hubble conflated the relative proportions of bulge and disk
1
1
2
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
D/B = 0 D/B > 1 D/B < 1 D/B =
Figure 1.9: Galaxies with varying disk to bulge ratios, seen edge on.
Figure 1.10: Two spiral galaxies, floc-
culent to the left and grand design to
the right. Flocculent systems are more
likely to be classified Sa; grand designs
are more frequently Sc.
with the prominence of spiral arms within the disk. In Hubble’s scheme most galaxies were classified
as “E”, “Sa”, “Sb” or “Sc.” “E” stood for elliptical with no sign of spiral structure. “S” stood for
spiral structure, with “Sa” galaxies showing the least spiral structure and “Sc” galaxies showing the
most, as shown in Figure 1.10. The sequence from E to Sc is approximately one of increasing disk-to-
bulge ratio. But there are also galaxies with prominent disks and little evidence for spiral structure.
Hubble added the S0 classification, placing it between E and Sa. Modern versions of Hubble’s
scheme have put the S0 systems parallel to the S systems, with S0/a, S0/b and S0/c systems having
increasingly large disk-to-bulge ratios. One might hope that disk-to-bulge ratio would completely
replace Hubble’s system, but while easy to describe, disk-to-bulge ratio can be difficult to measure,
particularly for systems that are face-on. Modern “quantitative” classification schemes often use
central concentration as a proxy for disk-to-bulge ratio and use deviation from bilateral symmetry
as a proxy for spiral structure.
8
1.2.
BUILDING GALAXIES
7
ij
Building galaxies
As we saw above, gravitational scattering plays a very small role in the dynamics of stellar orbits
within galaxies or globular clusters.
We return to our original question, “what is a galaxy?”
In
light of the previous section, it seems better to think of a galaxy as a collisionless fluid of stars, a
superposition of a large number of orbits guided by a single background potential.
How does one put together such a galaxy? Superpose orbits! Schwarzschild’s method (see Figure
1.2) follows these simple steps:
•
choose gravitational potential
•
generate an orbit library
•
populate potential with orbits
•
check for self-consistency
gravitational potential
orbit library
light = DF *footprints
orbit library
self−consistent
Figure 1.2: Schwarzschild’s method for
only if
building galaxies. One adjusts the distri-
light ~ mass
bution function (DF) to match the light
distribution or the gravitational poten-
tial.
For example, imagine building a spherically symmetric galaxy. How would the orbits be dis-
tributed in phase space? Even though they may occupy three-dimensional space quite uniformly,
there is no reason to insist that the orbits should be uniformly distributed in momentum space. A
galaxy made only of circular orbits would appear “hot” in the tangential direction and “cold” in the
radial direction, while a collection of radial orbits (an equally valid approach to a spherical system)
would look hot in the radial direction and cold in the tangential direction (see Figure 1.3).
Figure 1.3: Two schemes for constructing a spher-
ical galaxy: out of circular orbits, left, and out of
(nearly) radial orbits, right.
“Hot” and “cold” can be understood quantitatively by defining a symmetric velocity dispersion
tensor:
ε
2
= (v
i
−
v
i
)(v
j
−
v
j
),
(1.14)
where v
i
is the velocity in the
ˆ
e
i
direction averaged at some point. For a collisionless fluid, we can
diagonalize the velocity dispersion tensor, giving three independent temperatures at each point in
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
~
�
1
0.8
0.6
0.4
0.2
0
−0.2
−0.4
−0.6
−0.8
−1
−1
−0.8 −0.6 −0.4 −0.2
0
0.2
0.4
0.6
0.8
1
Figure 1.4: The orbit of a star on a
spherically symmetric logarithmic po-
tential. The position of the star is plot-
ted for the first 100 time steps. By fol-
lowing the star for many orbits we accu-
mulate a probability density “footprint”
that may be used to generate a galaxy
model.
Figure 1.5: Left: a disk in circular or-
bit around a spherical bulge. Right:
a disk in polar orbit around a triaxial
bulge. Such “polar ring” galaxies con
stitute perhaps 0.1% of all galaxies.
space. If the random velocities are greater than (or in the same order as) the ordered velocities, the
fluid is hot ; if the ordered velocities are much greater, the fluid is cold.
Observations (Figure 1.7) show that many galaxies have nearly constant rotational velocities
with v(r)
v
c
over the observable range of radius. These flat rotation curves are a departure from
Keplerian orbital velocities, which would scale as v(r)
r
−
1/2
. Flat rotation curves imply mass
increasing linearly with radius, and so cannot continue without bound.
For a spherically symmetric system, the centripetal acceleration is given by
v
2
=
dΛ
(1.15)
r
dr
For a constant rotation velocity v(r)
=
v
c
, the solution for the potential is
Λ
=
v
2
ln r
c
r
ref
.
(1.16)
Combining with Poisson’s equation
�
2
Λ
=
4ψGδ, we find an expression for the density profile:
v
2
δ(r)
=
c
.
(1.17)
4ψGr
2
1.2.
BUILDING GALAXIES
9
Telescope
Focal Plane
Collimator
Grism
Camera
Detector
Figure 1.6: A schematic representation of a spectrograph.
Light from the telescope is brought to a focus in the focal
plane. The slit of the spectrograph isolates the light from a
narrow strip of the sky, which diverges and passes through
Spectrometer the collimator. Parallel light rays from the collimator enter
the grism (a grating ruled on a prism) which disperses the
different wavelengths. The camera brings the light to a focus
on the detector.
galaxy
[N II]
H
[N II]
red
[SII]
blue
[SII]
Figure 1.7: Left: The slit of a spectrometer is placed along the major axis of a disk galaxy. The disk
is highly inclined to the line of sight. Stars and gas on the left side of the slit have a line of sight
velocity toward us (relative to the center of the galaxy). Right: The spectrum of the galaxy. The
lines shown are typical of gas photo-ionized by hot stars. The lines are blueshifted on the left side
and redshifted on the right side. This gives the rotation curve of the galaxy. The rotation curve is
flat in the outer parts of the galaxy and roughly linear in the inner parts.
spectrograph slit
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
�
v
z /rref
r /rref Figure 1.8: Equipotentials for Mestel’s Disk
Among the obvious shortcomings of this model are the failure of the potential to go to zero at large
radius and the prediction of infinite mass. Notwithstanding, it is still a useful potential for many
practical applications.
Another weakness of this model is that it is limited to spherically symmetric galaxies and we
know from experience that many galaxies are shaped more like flattened disks. A corresponding
model for such a galaxy is called a Mestel disk, which is azimuthally symmetric around the z-axis.
The potential for the Mestel disk is
Λ(r, β)
=
v
2
�
ln r
+
ln 1
+
|
cos β
|
�
,
(1.18)
c
r
ref
2
where β is measured from the positive z-axis. The Laplacian in spherical geometry with azimuthal
symmetry is given by
2 1
σ 2 σ 1 σ σ
�
=
r
2
σrr σr
+
r
2
sin β σβ sin βσβ.
(1.19)
Applying Poisson’s equation to equation (1.18), we find that the Mestel disk has zero density every-
where except β
=
ψ/2, where the density is infinite. The more useful (and in this case finite) quantity
is the surface density, determined by integrating the density over the (infinitesimal) thickness of the
disk:
2
Ψ(r)
=
δ(r, z)dz
=
c
.
(1.20)
2ψGr
Like the spherical model for a flat rotation curve, the disk model also suffers from an infinite total
mass and radius, but it is still useful for many applications.
Observations show that most galaxies have elements of both of the above models, with a roughly
spherical bulge and a flattened disk. The relative weight of these features (the “disk-to-bulge ratio,”
Figure 1.9) is a useful parameter for classifying galaxies, with disk strength correlating positively
with other galaxy properties such as young stellar fraction, hydrogen fraction, CO gas, dust, and the
strength of spiral arms. In most galaxies the spiral arms, which are located within the disk are the
primary location for new star formation. The disk and bulge components of most galaxies appear to
have had rather different histories, a subject to be raised again in the context of galaxy formation.
Soon after the realization that many nebulae were in fact distinct galaxies, astronomers (and
in particular Hubble) began systematically classifying galaxies in the hope that taxonomy would
lead to further understanding. The principal component of Hubble’s classification scheme is strongly
correlated with disk-to-bulge ratio. But Hubble conflated the relative proportions of bulge and disk
1
1
2
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
D/B = 0 D/B > 1 D/B < 1 D/B =
Figure 1.9: Galaxies with varying disk to bulge ratios, seen edge on.
Figure 1.10: Two spiral galaxies, floc-
culent to the left and grand design to
the right. Flocculent systems are more
likely to be classified Sa; grand designs
are more frequently Sc.
1.9
with the prominence of spiral arms within the disk. In Hubble’s scheme most galaxies were classified
as “E”, “Sa”, “Sb” or “Sc.” “E” stood for elliptical with no sign of spiral structure. “S” stood for spiral
structure, with “Sa” galaxies showing the least spiral structure and “Sc” galaxies showing the most,
as shown in Figure 1.10. The sequence from E to Sc is approximately one of increasing disk-to- bulge
ratio. But there are also galaxies with prominent disks and little evidence for spiral structure. Hubble
added the S0 classification, placing it between E and Sa. Modern versions of Hubble’s scheme have
put the S0 systems parallel to the S systems, with S0/a, S0/b and S0/c systems having increasingly
large disk-to-bulge ratios. One might hope that disk-to-bulge ratio would completely replace
Hubble’s system, but while easy to describe, disk-to-bulge ratio can be difficult to measure,
particularly for systems that are face-on. Modern “quantitative” classification schemes often use
central concentration as a proxy for disk-to-bulge ratio and use deviation from bilateral symmetry as
a proxy for spiral structure.
8
1.2.
BUILDING GALAXIES
7
ij
Building galaxies
As we saw above, gravitational scattering plays a very small role in the dynamics of stellar orbits
within galaxies or globular clusters.
We return to our original question, “what is a galaxy?”
In
light of the previous section, it seems better to think of a galaxy as a collisionless fluid of stars, a
superposition of a large number of orbits guided by a single background potential.
How does one put together such a galaxy? Superpose orbits! Schwarzschild’s method (see Figure
1.2) follows these simple steps:
•
choose gravitational potential
•
generate an orbit library
•
populate potential with orbits
•
check for self-consistency
gravitational potential
orbit library
light = DF *footprints
orbit library
self−consistent
Figure 1.2: Schwarzschild’s method for
only if
building galaxies. One adjusts the distri-
light ~ mass
bution function (DF) to match the light
distribution or the gravitational poten-
tial.
For example, imagine building a spherically symmetric galaxy. How would the orbits be dis-
tributed in phase space? Even though they may occupy three-dimensional space quite uniformly,
there is no reason to insist that the orbits should be uniformly distributed in momentum space. A
galaxy made only of circular orbits would appear “hot” in the tangential direction and “cold” in the
radial direction, while a collection of radial orbits (an equally valid approach to a spherical system)
would look hot in the radial direction and cold in the tangential direction (see Figure 1.3).
Figure 1.3: Two schemes for constructing a spher-
ical galaxy: out of circular orbits, left, and out of
(nearly) radial orbits, right.
“Hot” and “cold” can be understood quantitatively by defining a symmetric velocity dispersion
tensor:
ε
2
= (v
i
−
v
i
)(v
j
−
v
j
),
(1.14)
where v
i
is the velocity in the
ˆ
e
i
direction averaged at some point. For a collisionless fluid, we can
diagonalize the velocity dispersion tensor, giving three independent temperatures at each point in
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
~
�
1
0.8
0.6
0.4
0.2
0
−0.2
−0.4
−0.6
−0.8
−1
−1
−0.8 −0.6 −0.4 −0.2
0
0.2
0.4
0.6
0.8
1
Figure 1.4: The orbit of a star on a
spherically symmetric logarithmic po-
tential. The position of the star is plot-
ted for the first 100 time steps. By fol-
lowing the star for many orbits we accu-
mulate a probability density “footprint”
that may be used to generate a galaxy
model.
Figure 1.5: Left: a disk in circular or-
bit around a spherical bulge. Right:
a disk in polar orbit around a triaxial
bulge. Such “polar ring” galaxies con
stitute perhaps 0.1% of all galaxies.
space. If the random velocities are greater than (or in the same order as) the ordered velocities, the
fluid is hot ; if the ordered velocities are much greater, the fluid is cold.
Observations (Figure 1.7) show that many galaxies have nearly constant rotational velocities
with v(r)
v
c
over the observable range of radius. These flat rotation curves are a departure from
Keplerian orbital velocities, which would scale as v(r)
r
−
1/2
. Flat rotation curves imply mass
increasing linearly with radius, and so cannot continue without bound.
For a spherically symmetric system, the centripetal acceleration is given by
v
2
=
dΛ
(1.15)
r
dr
For a constant rotation velocity v(r)
=
v
c
, the solution for the potential is
Λ
=
v
2
ln r
c
r
ref
.
(1.16)
Combining with Poisson’s equation
�
2
Λ
=
4ψGδ, we find an expression for the density profile:
v
2
δ(r)
=
c
.
(1.17)
4ψGr
2
1.2.
BUILDING GALAXIES
9
Telescope
Focal Plane
Collimator
Grism
Camera
Detector
Figure 1.6: A schematic representation of a spectrograph.
Light from the telescope is brought to a focus in the focal
plane. The slit of the spectrograph isolates the light from a
narrow strip of the sky, which diverges and passes through
Spectrometer the collimator. Parallel light rays from the collimator enter
the grism (a grating ruled on a prism) which disperses the
different wavelengths. The camera brings the light to a focus
on the detector.
galaxy
[N II]
H
[N II]
red
[SII]
blue
[SII]
Figure 1.7: Left: The slit of a spectrometer is placed along the major axis of a disk galaxy. The disk
is highly inclined to the line of sight. Stars and gas on the left side of the slit have a line of sight
velocity toward us (relative to the center of the galaxy). Right: The spectrum of the galaxy. The
lines shown are typical of gas photo-ionized by hot stars. The lines are blueshifted on the left side
and redshifted on the right side. This gives the rotation curve of the galaxy. The rotation curve is
flat in the outer parts of the galaxy and roughly linear in the inner parts.
spectrograph slit
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
�
v
z /rref
r /rref Figure 1.8: Equipotentials for Mestel’s Disk
Among the obvious shortcomings of this model are the failure of the potential to go to zero at large
radius and the prediction of infinite mass. Notwithstanding, it is still a useful potential for many
practical applications.
Another weakness of this model is that it is limited to spherically symmetric galaxies and we
know from experience that many galaxies are shaped more like flattened disks. A corresponding
model for such a galaxy is called a Mestel disk, which is azimuthally symmetric around the z-axis.
The potential for the Mestel disk is
Λ(r, β)
=
v
2
�
ln r
+
ln 1
+
|
cos β
|
�
,
(1.18)
c
r
ref
2
where β is measured from the positive z-axis. The Laplacian in spherical geometry with azimuthal
symmetry is given by
2 1
σ 2 σ 1 σ σ
�
=
r
2
σrr σr
+
r
2
sin β σβ sin βσβ.
(1.19)
Applying Poisson’s equation to equation (1.18), we find that the Mestel disk has zero density every-
where except β
=
ψ/2, where the density is infinite. The more useful (and in this case finite) quantity
is the surface density, determined by integrating the density over the (infinitesimal) thickness of the
disk:
2
Ψ(r)
=
δ(r, z)dz
=
c
.
(1.20)
2ψGr
Like the spherical model for a flat rotation curve, the disk model also suffers from an infinite total
mass and radius, but it is still useful for many applications.
Observations show that most galaxies have elements of both of the above models, with a roughly
spherical bulge and a flattened disk. The relative weight of these features (the “disk-to-bulge ratio,”
Figure 1.9) is a useful parameter for classifying galaxies, with disk strength correlating positively
with other galaxy properties such as young stellar fraction, hydrogen fraction, CO gas, dust, and the
strength of spiral arms. In most galaxies the spiral arms, which are located within the disk are the
primary location for new star formation. The disk and bulge components of most galaxies appear to
have had rather different histories, a subject to be raised again in the context of galaxy formation.
Soon after the realization that many nebulae were in fact distinct galaxies, astronomers (and
in particular Hubble) began systematically classifying galaxies in the hope that taxonomy would
lead to further understanding. The principal component of Hubble’s classification scheme is strongly
correlated with disk-to-bulge ratio. But Hubble conflated the relative proportions of bulge and disk
1
1
2
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
D/B = 0 D/B > 1 D/B < 1 D/B =
Figure 1.9: Galaxies with varying disk to bulge ratios, seen edge on.
Figure 1.10: Two spiral galaxies, floc-
culent to the left and grand design to
the right. Flocculent systems are more
likely to be classified Sa; grand designs
are more frequently Sc.
with the prominence of spiral arms within the disk. In Hubble’s scheme most galaxies were classified
as “E”, “Sa”, “Sb” or “Sc.” “E” stood for elliptical with no sign of spiral structure. “S” stood for
spiral structure, with “Sa” galaxies showing the least spiral structure and “Sc” galaxies showing the
most, as shown in Figure 1.10. The sequence from E to Sc is approximately one of increasing disk-to-
bulge ratio. But there are also galaxies with prominent disks and little evidence for spiral structure.
Hubble added the S0 classification, placing it between E and Sa. Modern versions of Hubble’s
scheme have put the S0 systems parallel to the S systems, with S0/a, S0/b and S0/c systems having
increasingly large disk-to-bulge ratios. One might hope that disk-to-bulge ratio would completely
replace Hubble’s system, but while easy to describe, disk-to-bulge ratio can be difficult to measure,
particularly for systems that are face-on. Modern “quantitative” classification schemes often use
central concentration as a proxy for disk-to-bulge ratio and use deviation from bilateral symmetry
as a proxy for spiral structure.
8
1.2.
BUILDING GALAXIES
7
ij
Building galaxies
As we saw above, gravitational scattering plays a very small role in the dynamics of stellar orbits
within galaxies or globular clusters.
We return to our original question, “what is a galaxy?”
In
light of the previous section, it seems better to think of a galaxy as a collisionless fluid of stars, a
superposition of a large number of orbits guided by a single background potential.
How does one put together such a galaxy? Superpose orbits! Schwarzschild’s method (see Figure
1.2) follows these simple steps:
•
choose gravitational potential
•
generate an orbit library
•
populate potential with orbits
•
check for self-consistency
gravitational potential
orbit library
light = DF *footprints
orbit library
self−consistent
Figure 1.2: Schwarzschild’s method for
only if
building galaxies. One adjusts the distri-
light ~ mass
bution function (DF) to match the light
distribution or the gravitational poten-
tial.
For example, imagine building a spherically symmetric galaxy. How would the orbits be dis-
tributed in phase space? Even though they may occupy three-dimensional space quite uniformly,
there is no reason to insist that the orbits should be uniformly distributed in momentum space. A
galaxy made only of circular orbits would appear “hot” in the tangential direction and “cold” in the
radial direction, while a collection of radial orbits (an equally valid approach to a spherical system)
would look hot in the radial direction and cold in the tangential direction (see Figure 1.3).
Figure 1.3: Two schemes for constructing a spher-
ical galaxy: out of circular orbits, left, and out of
(nearly) radial orbits, right.
“Hot” and “cold” can be understood quantitatively by defining a symmetric velocity dispersion
tensor:
ε
2
= (v
i
−
v
i
)(v
j
−
v
j
),
(1.14)
where v
i
is the velocity in the
ˆ
e
i
direction averaged at some point. For a collisionless fluid, we can
diagonalize the velocity dispersion tensor, giving three independent temperatures at each point in
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
~
�
1
0.8
0.6
0.4
0.2
0
−0.2
−0.4
−0.6
−0.8
−1
−1
−0.8 −0.6 −0.4 −0.2
0
0.2
0.4
0.6
0.8
1
Figure 1.4: The orbit of a star on a
spherically symmetric logarithmic po-
tential. The position of the star is plot-
ted for the first 100 time steps. By fol-
lowing the star for many orbits we accu-
mulate a probability density “footprint”
that may be used to generate a galaxy
model.
Figure 1.5: Left: a disk in circular or-
bit around a spherical bulge. Right:
a disk in polar orbit around a triaxial
bulge. Such “polar ring” galaxies con
stitute perhaps 0.1% of all galaxies.
space. If the random velocities are greater than (or in the same order as) the ordered velocities, the
fluid is hot ; if the ordered velocities are much greater, the fluid is cold.
Observations (Figure 1.7) show that many galaxies have nearly constant rotational velocities
with v(r)
v
c
over the observable range of radius. These flat rotation curves are a departure from
Keplerian orbital velocities, which would scale as v(r)
r
−
1/2
. Flat rotation curves imply mass
increasing linearly with radius, and so cannot continue without bound.
For a spherically symmetric system, the centripetal acceleration is given by
v
2
=
dΛ
(1.15)
r
dr
For a constant rotation velocity v(r)
=
v
c
, the solution for the potential is
Λ
=
v
2
ln r
c
r
ref
.
(1.16)
Combining with Poisson’s equation
�
2
Λ
=
4ψGδ, we find an expression for the density profile:
v
2
δ(r)
=
c
.
(1.17)
4ψGr
2
1.2.
BUILDING GALAXIES
9
Telescope
Focal Plane
Collimator
Grism
Camera
Detector
Figure 1.6: A schematic representation of a spectrograph.
Light from the telescope is brought to a focus in the focal
plane. The slit of the spectrograph isolates the light from a
narrow strip of the sky, which diverges and passes through
Spectrometer the collimator. Parallel light rays from the collimator enter
the grism (a grating ruled on a prism) which disperses the
different wavelengths. The camera brings the light to a focus
on the detector.
galaxy
[N II]
H
[N II]
red
[SII]
blue
[SII]
Figure 1.7: Left: The slit of a spectrometer is placed along the major axis of a disk galaxy. The disk
is highly inclined to the line of sight. Stars and gas on the left side of the slit have a line of sight
velocity toward us (relative to the center of the galaxy). Right: The spectrum of the galaxy. The
lines shown are typical of gas photo-ionized by hot stars. The lines are blueshifted on the left side
and redshifted on the right side. This gives the rotation curve of the galaxy. The rotation curve is
flat in the outer parts of the galaxy and roughly linear in the inner parts.
spectrograph slit
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
�
v
z /rref
r /rref Figure 1.8: Equipotentials for Mestel’s Disk
Among the obvious shortcomings of this model are the failure of the potential to go to zero at large
radius and the prediction of infinite mass. Notwithstanding, it is still a useful potential for many
practical applications.
Another weakness of this model is that it is limited to spherically symmetric galaxies and we
know from experience that many galaxies are shaped more like flattened disks. A corresponding
model for such a galaxy is called a Mestel disk, which is azimuthally symmetric around the z-axis.
The potential for the Mestel disk is
Λ(r, β)
=
v
2
�
ln r
+
ln 1
+
|
cos β
|
�
,
(1.18)
c
r
ref
2
where β is measured from the positive z-axis. The Laplacian in spherical geometry with azimuthal
symmetry is given by
2 1
σ 2 σ 1 σ σ
�
=
r
2
σrr σr
+
r
2
sin β σβ sin βσβ.
(1.19)
Applying Poisson’s equation to equation (1.18), we find that the Mestel disk has zero density every-
where except β
=
ψ/2, where the density is infinite. The more useful (and in this case finite) quantity
is the surface density, determined by integrating the density over the (infinitesimal) thickness of the
disk:
2
Ψ(r)
=
δ(r, z)dz
=
c
.
(1.20)
2ψGr
Like the spherical model for a flat rotation curve, the disk model also suffers from an infinite total
mass and radius, but it is still useful for many applications.
Observations show that most galaxies have elements of both of the above models, with a roughly
spherical bulge and a flattened disk. The relative weight of these features (the “disk-to-bulge ratio,”
Figure 1.9) is a useful parameter for classifying galaxies, with disk strength correlating positively
with other galaxy properties such as young stellar fraction, hydrogen fraction, CO gas, dust, and the
strength of spiral arms. In most galaxies the spiral arms, which are located within the disk are the
primary location for new star formation. The disk and bulge components of most galaxies appear to
have had rather different histories, a subject to be raised again in the context of galaxy formation.
Soon after the realization that many nebulae were in fact distinct galaxies, astronomers (and
in particular Hubble) began systematically classifying galaxies in the hope that taxonomy would
lead to further understanding. The principal component of Hubble’s classification scheme is strongly
correlated with disk-to-bulge ratio. But Hubble conflated the relative proportions of bulge and disk
1
1
2
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
D/B = 0 D/B > 1 D/B < 1 D/B =
Figure 1.9: Galaxies with varying disk to bulge ratios, seen edge on.
Figure 1.10: Two spiral galaxies, floc-
culent to the left and grand design to
the right. Flocculent systems are more
likely to be classified Sa; grand designs
are more frequently Sc.
1.10
with the prominence of spiral arms within the disk. In Hubble’s scheme most galaxies were classified
as “E”, “Sa”, “Sb” or “Sc.” “E” stood for elliptical with no sign of spiral structure. “S” stood for spiral
structure, with “Sa” galaxies showing the least spiral structure and “Sc” galaxies showing the most,
as shown in Figure 1.10. The sequence from E to Sc is approximately one of increasing disk-to- bulge
ratio. But there are also galaxies with prominent disks and little evidence for spiral structure. Hubble
added the S0 classification, placing it between E and Sa. Modern versions of Hubble’s scheme have
put the S0 systems parallel to the S systems, with S0/a, S0/b and S0/c systems having increasingly
large disk-to-bulge ratios. One might hope that disk-to-bulge ratio would completely replace
Hubble’s system, but while easy to describe, disk-to-bulge ratio can be difficult to measure,
particularly for systems that are face-on. Modern “quantitative” classification schemes often use
central concentration as a proxy for disk-to-bulge ratio and use deviation from bilateral symmetry as
a proxy for spiral structure.
8
1.2.
BUILDING GALAXIES
7
ij
Building galaxies
As we saw above, gravitational scattering plays a very small role in the dynamics of stellar orbits
within galaxies or globular clusters.
We return to our original question, “what is a galaxy?”
In
light of the previous section, it seems better to think of a galaxy as a collisionless fluid of stars, a
superposition of a large number of orbits guided by a single background potential.
How does one put together such a galaxy? Superpose orbits! Schwarzschild’s method (see Figure
1.2) follows these simple steps:
•
choose gravitational potential
•
generate an orbit library
•
populate potential with orbits
•
check for self-consistency
gravitational potential
orbit library
light = DF *footprints
orbit library
self−consistent
Figure 1.2: Schwarzschild’s method for
only if
building galaxies. One adjusts the distri-
light ~ mass
bution function (DF) to match the light
distribution or the gravitational poten-
tial.
For example, imagine building a spherically symmetric galaxy. How would the orbits be dis-
tributed in phase space? Even though they may occupy three-dimensional space quite uniformly,
there is no reason to insist that the orbits should be uniformly distributed in momentum space. A
galaxy made only of circular orbits would appear “hot” in the tangential direction and “cold” in the
radial direction, while a collection of radial orbits (an equally valid approach to a spherical system)
would look hot in the radial direction and cold in the tangential direction (see Figure 1.3).
Figure 1.3: Two schemes for constructing a spher-
ical galaxy: out of circular orbits, left, and out of
(nearly) radial orbits, right.
“Hot” and “cold” can be understood quantitatively by defining a symmetric velocity dispersion
tensor:
ε
2
= (v
i
−
v
i
)(v
j
−
v
j
),
(1.14)
where v
i
is the velocity in the
ˆ
e
i
direction averaged at some point. For a collisionless fluid, we can
diagonalize the velocity dispersion tensor, giving three independent temperatures at each point in
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
~
�
1
0.8
0.6
0.4
0.2
0
−0.2
−0.4
−0.6
−0.8
−1
−1
−0.8 −0.6 −0.4 −0.2
0
0.2
0.4
0.6
0.8
1
Figure 1.4: The orbit of a star on a
spherically symmetric logarithmic po-
tential. The position of the star is plot-
ted for the first 100 time steps. By fol-
lowing the star for many orbits we accu-
mulate a probability density “footprint”
that may be used to generate a galaxy
model.
Figure 1.5: Left: a disk in circular or-
bit around a spherical bulge. Right:
a disk in polar orbit around a triaxial
bulge. Such “polar ring” galaxies con
stitute perhaps 0.1% of all galaxies.
space. If the random velocities are greater than (or in the same order as) the ordered velocities, the
fluid is hot ; if the ordered velocities are much greater, the fluid is cold.
Observations (Figure 1.7) show that many galaxies have nearly constant rotational velocities
with v(r)
v
c
over the observable range of radius. These flat rotation curves are a departure from
Keplerian orbital velocities, which would scale as v(r)
r
−
1/2
. Flat rotation curves imply mass
increasing linearly with radius, and so cannot continue without bound.
For a spherically symmetric system, the centripetal acceleration is given by
v
2
=
dΛ
(1.15)
r
dr
For a constant rotation velocity v(r)
=
v
c
, the solution for the potential is
Λ
=
v
2
ln r
c
r
ref
.
(1.16)
Combining with Poisson’s equation
�
2
Λ
=
4ψGδ, we find an expression for the density profile:
v
2
δ(r)
=
c
.
(1.17)
4ψGr
2
1.2.
BUILDING GALAXIES
9
Telescope
Focal Plane
Collimator
Grism
Camera
Detector
Figure 1.6: A schematic representation of a spectrograph.
Light from the telescope is brought to a focus in the focal
plane. The slit of the spectrograph isolates the light from a
narrow strip of the sky, which diverges and passes through
Spectrometer the collimator. Parallel light rays from the collimator enter
the grism (a grating ruled on a prism) which disperses the
different wavelengths. The camera brings the light to a focus
on the detector.
galaxy
[N II]
H
[N II]
red
[SII]
blue
[SII]
Figure 1.7: Left: The slit of a spectrometer is placed along the major axis of a disk galaxy. The disk
is highly inclined to the line of sight. Stars and gas on the left side of the slit have a line of sight
velocity toward us (relative to the center of the galaxy). Right: The spectrum of the galaxy. The
lines shown are typical of gas photo-ionized by hot stars. The lines are blueshifted on the left side
and redshifted on the right side. This gives the rotation curve of the galaxy. The rotation curve is
flat in the outer parts of the galaxy and roughly linear in the inner parts.
spectrograph slit
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
�
v
z /rref
r /rref Figure 1.8: Equipotentials for Mestel’s Disk
Among the obvious shortcomings of this model are the failure of the potential to go to zero at large
radius and the prediction of infinite mass. Notwithstanding, it is still a useful potential for many
practical applications.
Another weakness of this model is that it is limited to spherically symmetric galaxies and we
know from experience that many galaxies are shaped more like flattened disks. A corresponding
model for such a galaxy is called a Mestel disk, which is azimuthally symmetric around the z-axis.
The potential for the Mestel disk is
Λ(r, β)
=
v
2
�
ln r
+
ln 1
+
|
cos β
|
�
,
(1.18)
c
r
ref
2
where β is measured from the positive z-axis. The Laplacian in spherical geometry with azimuthal
symmetry is given by
2 1
σ 2 σ 1 σ σ
�
=
r
2
σrr σr
+
r
2
sin β σβ sin βσβ.
(1.19)
Applying Poisson’s equation to equation (1.18), we find that the Mestel disk has zero density every-
where except β
=
ψ/2, where the density is infinite. The more useful (and in this case finite) quantity
is the surface density, determined by integrating the density over the (infinitesimal) thickness of the
disk:
2
Ψ(r)
=
δ(r, z)dz
=
c
.
(1.20)
2ψGr
Like the spherical model for a flat rotation curve, the disk model also suffers from an infinite total
mass and radius, but it is still useful for many applications.
Observations show that most galaxies have elements of both of the above models, with a roughly
spherical bulge and a flattened disk. The relative weight of these features (the “disk-to-bulge ratio,”
Figure 1.9) is a useful parameter for classifying galaxies, with disk strength correlating positively
with other galaxy properties such as young stellar fraction, hydrogen fraction, CO gas, dust, and the
strength of spiral arms. In most galaxies the spiral arms, which are located within the disk are the
primary location for new star formation. The disk and bulge components of most galaxies appear to
have had rather different histories, a subject to be raised again in the context of galaxy formation.
Soon after the realization that many nebulae were in fact distinct galaxies, astronomers (and
in particular Hubble) began systematically classifying galaxies in the hope that taxonomy would
lead to further understanding. The principal component of Hubble’s classification scheme is strongly
correlated with disk-to-bulge ratio. But Hubble conflated the relative proportions of bulge and disk
1
1
2
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
D/B = 0 D/B > 1 D/B < 1 D/B =
Figure 1.9: Galaxies with varying disk to bulge ratios, seen edge on.
Figure 1.10: Two spiral galaxies, floc-
culent to the left and grand design to
the right. Flocculent systems are more
likely to be classified Sa; grand designs
are more frequently Sc.
with the prominence of spiral arms within the disk. In Hubble’s scheme most galaxies were classified as “E”,
“Sa”, “Sb” or “Sc.” “E” stood for elliptical with no sign of spiral structure. “S” stood for spiral structure, with
“Sa” galaxies showing the least spiral structure and “Sc” galaxies showing the most, as shown in Figure 1.10.
The sequence from E to Sc is approximately one of increasing disk-to- bulge ratio. But there are also galaxies
with prominent disks and little evidence for spiral structure. Hubble added the S0 classification, placing it
between E and Sa. Modern versions of Hubble’s scheme have put the S0 systems parallel to the S systems,
with S0/a, S0/b and S0/c systems having increasingly large disk-to-bulge ratios. One might hope that disk-to-
bulge ratio would completely replace Hubble’s system, but while easy to describe, disk-to-bulge ratio can be
difficult to measure, particularly for systems that are face-on. Modern “quantitative” classification schemes
often use central concentration as a proxy for disk-to-bulge ratio and use deviation from bilateral symmetry as
a proxy for spiral structure.
8
1.2.
BUILDING GALAXIES
7
ij
Building galaxies
As we saw above, gravitational scattering plays a very small role in the dynamics of stellar orbits
within galaxies or globular clusters.
We return to our original question, “what is a galaxy?”
In
light of the previous section, it seems better to think of a galaxy as a collisionless fluid of stars, a
superposition of a large number of orbits guided by a single background potential.
How does one put together such a galaxy? Superpose orbits! Schwarzschild’s method (see Figure
1.2) follows these simple steps:
•
choose gravitational potential
•
generate an orbit library
•
populate potential with orbits
•
check for self-consistency
gravitational potential
orbit library
light = DF *footprints
orbit library
self−consistent
Figure 1.2: Schwarzschild’s method for
only if
building galaxies. One adjusts the distri-
light ~ mass
bution function (DF) to match the light
distribution or the gravitational poten-
tial.
For example, imagine building a spherically symmetric galaxy. How would the orbits be dis-
tributed in phase space? Even though they may occupy three-dimensional space quite uniformly,
there is no reason to insist that the orbits should be uniformly distributed in momentum space. A
galaxy made only of circular orbits would appear “hot” in the tangential direction and “cold” in the
radial direction, while a collection of radial orbits (an equally valid approach to a spherical system)
would look hot in the radial direction and cold in the tangential direction (see Figure 1.3).
Figure 1.3: Two schemes for constructing a spher-
ical galaxy: out of circular orbits, left, and out of
(nearly) radial orbits, right.
“Hot” and “cold” can be understood quantitatively by defining a symmetric velocity dispersion
tensor:
ε
2
= (v
i
−
v
i
)(v
j
−
v
j
),
(1.14)
where v
i
is the velocity in the
ˆ
e
i
direction averaged at some point. For a collisionless fluid, we can
diagonalize the velocity dispersion tensor, giving three independent temperatures at each point in
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
~
�
1
0.8
0.6
0.4
0.2
0
−0.2
−0.4
−0.6
−0.8
−1
−1
−0.8 −0.6 −0.4 −0.2
0
0.2
0.4
0.6
0.8
1
Figure 1.4: The orbit of a star on a
spherically symmetric logarithmic po-
tential. The position of the star is plot-
ted for the first 100 time steps. By fol-
lowing the star for many orbits we accu-
mulate a probability density “footprint”
that may be used to generate a galaxy
model.
Figure 1.5: Left: a disk in circular or-
bit around a spherical bulge. Right:
a disk in polar orbit around a triaxial
bulge. Such “polar ring” galaxies con
stitute perhaps 0.1% of all galaxies.
space. If the random velocities are greater than (or in the same order as) the ordered velocities, the
fluid is hot ; if the ordered velocities are much greater, the fluid is cold.
Observations (Figure 1.7) show that many galaxies have nearly constant rotational velocities
with v(r)
v
c
over the observable range of radius. These flat rotation curves are a departure from
Keplerian orbital velocities, which would scale as v(r)
r
−
1/2
. Flat rotation curves imply mass
increasing linearly with radius, and so cannot continue without bound.
For a spherically symmetric system, the centripetal acceleration is given by
v
2
=
dΛ
(1.15)
r
dr
For a constant rotation velocity v(r)
=
v
c
, the solution for the potential is
Λ
=
v
2
ln r
c
r
ref
.
(1.16)
Combining with Poisson’s equation
�
2
Λ
=
4ψGδ, we find an expression for the density profile:
v
2
δ(r)
=
c
.
(1.17)
4ψGr
2
1.2.
BUILDING GALAXIES
9
Telescope
Focal Plane
Collimator
Grism
Camera
Detector
Figure 1.6: A schematic representation of a spectrograph.
Light from the telescope is brought to a focus in the focal
plane. The slit of the spectrograph isolates the light from a
narrow strip of the sky, which diverges and passes through
Spectrometer the collimator. Parallel light rays from the collimator enter
the grism (a grating ruled on a prism) which disperses the
different wavelengths. The camera brings the light to a focus
on the detector.
galaxy
[N II]
H
[N II]
red
[SII]
blue
[SII]
Figure 1.7: Left: The slit of a spectrometer is placed along the major axis of a disk galaxy. The disk
is highly inclined to the line of sight. Stars and gas on the left side of the slit have a line of sight
velocity toward us (relative to the center of the galaxy). Right: The spectrum of the galaxy. The
lines shown are typical of gas photo-ionized by hot stars. The lines are blueshifted on the left side
and redshifted on the right side. This gives the rotation curve of the galaxy. The rotation curve is
flat in the outer parts of the galaxy and roughly linear in the inner parts.
spectrograph slit
CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA
�
v
z /rref
r /rref Figure 1.8: Equipotentials for Mestel’s Disk
Among the obvious shortcomings of this model are the failure of the potential to go to zero at large
radius and the prediction of infinite mass. Notwithstanding, it is still a useful potential for many
practical applications.
Another weakness of this model is that it is limited to spherically symmetric galaxies and we
know from experience that many galaxies are shaped more like flattened disks. A corresponding
model for such a galaxy is called a Mestel disk, which is azimuthally symmetric around the z-axis.
The potential for the Mestel disk is
Λ(r, β)
=
v
2
�
ln r
+
ln 1
+
|
cos β
|
�
,
(1.18)
c
r
ref
2
where β is measured from the positive z-axis. The Laplacian in spherical geometry with azimuthal
symmetry is given by
2 1
σ 2 σ 1 σ σ
�
=
r
2
σrr σr
+
r
2
sin β σβ sin βσβ.
(1.19)
Applying Poisson’s equation to equation (1.18), we find that the Mestel disk has zero density every-
where except β
=
ψ/2, where the density is infinite. The more useful (and in this case finite) quantity
is the surface density, determined by integrating the density over the (infinitesimal) thickness of the
disk:
2
Ψ(r)
=
δ(r, z)dz
=
c
.
(1.20)
2ψGr
Like the spherical model for a flat rotation curve, the disk model also suffers from an infinite total
mass and radius, but it is still useful for many applications.
Observations show that most galaxies have elements of both of the above models, with a roughly
spherical bulge and a flattened disk. The relative weight of these features (the “disk-to-bulge ratio,”
Figure 1.9) is a useful parameter for classifying galaxies, with disk strength correlating positively
with other galaxy properties such as young stellar fraction, hydrogen fraction, CO gas, dust, and the
strength of spiral arms. In most galaxies the spiral arms, which are located within the disk are the
primary location for new star formation. The disk and bulge components of most galaxies appear to
have had rather different histories, a subject to be raised again in the context of galaxy formation.
Soon after the realization that many nebulae were in fact distinct galaxies, astronomers (and
in particular Hubble) began systematically classifying galaxies in the hope that taxonomy would
lead to further understanding. The principal component of Hubble’s classification scheme is strongly
correlated with disk-to-bulge ratio. But Hubble conflated the relative proportions of bulge and disk
1
1
2
D/B = 0 D/B > 1 D/B < 1 D/B =
Figure 1.9: Galaxies with varying disk to bulge ratios, seen edge on.
Figure 1.10: Two spiral galaxies, floc-
culent to the left and grand design to
the right. Flocculent systems are more
likely to be classified Sa; grand designs
are more frequently Sc.
1.11
with the prominence of spiral arms within the disk. In Hubble’s scheme most galaxies were classified
as “E”, “Sa”, “Sb” or “Sc.” “E” stood for elliptical with no sign of spiral structure. “S” stood for spiral
structure, with “Sa” galaxies showing the least spiral structure and “Sc” galaxies showing the most,
as shown in Figure 1.10. The sequence from E to Sc is approximately one of increasing disk-to- bulge
ratio. But there are also galaxies with prominent disks and little evidence for spiral structure. Hubble
added the S0 classification, placing it between E and Sa. Modern versions of Hubble’s scheme have
put the S0 systems parallel to the S systems, with S0/a, S0/b and S0/c systems having increasingly
large disk-to-bulge ratios. One might hope that disk-to-bulge ratio would completely replace
Hubble’s system, but while easy to describe, disk-to-bulge ratio can be difficult to measure,
particularly for systems that are face-on. Modern “quantitative” classification schemes often use
central concentration as a proxy for disk-to-bulge ratio and use deviation from bilateral symmetry as
a proxy for spiral structure.
8
1.2.
BUILDING GALAXIES
7
ij
Building galaxies
As we saw above, gravitational scattering plays a very small role in the dynamics of stellar orbits
within galaxies or globular clusters.
We return to our original question, “what is a galaxy?”
In
light of the previous section, it seems better to think of a galaxy as a collisionless fluid of stars, a
superposition of a large number of orbits guided by a single background potential.
How does one put together such a galaxy? Superpose orbits! Schwarzschild’s method (see Figure
1.2) follows these simple steps:
•
choose gravitational potential
•
generate an orbit library
•
populate potential with orbits
•
check for self-consistency
gravitational potential
orbit library
light = DF *footprints
orbit library
self−consistent
Figure 1.2: Schwarzschild’s method for
only if
building galaxies. One adjusts the distri-
light ~ mass
bution function (DF) to match the light
distribution or the gravitational poten-
tial.
For example, imagine building a spherically symmetric galaxy. How would the orbits be dis-
tributed in phase space? Even though they may occupy three-dimensional space quite uniformly,
there is no reason to insist that the orbits should be uniformly distributed in momentum space. A
galaxy made only of circular orbits would appear “hot” in the tangential direction and “cold” in the
radial direction, while a collection of radial orbits (an equally valid approach to a spherical system)
would look hot in the radial direction and cold in the tangential direction (see Figure 1.3).
Figure 1.3: Two schemes for constructing a spher-
ical galaxy: out of circular orbits, left, and out of
(nearly) radial orbits, right.
“Hot” and “cold” can be understood quantitatively by defining a symmetric velocity dispersion
tensor:
ε
2
= (v
i
−
v
i
)(v
j
−
v
j
),
(1.14)
where v
i
is the velocity in the
ˆ
e
i
direction averaged at some point. For a collisionless fluid, we can
diagonalize the velocity dispersion tensor, giving three independent temperatures at each point in
