10 CHAPTER 1.

GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA

1.3 Extragalactic empiricism

Before we can discuss galaxies quantitatively, we must first introduce some new terms. The specific
~ is a measure of energy flux incident on a unit surface area per unit solid angle and
intensity Iν (θ)
frequency (erg/s/cm2 /sterad/Hz). Then the specific flux on a given surface element is the specific
intensity integrated over all solid angle:
Z
fν = Iν dΩ. (1.21)

For the purpose of the present discussion, the units of specific flux are ergs/s/cm2 /Hz. The specific
luminosity Lν at a given frequency ν of an object at a distance D is then
Lν = 4πD2 fν . (1.22)

(Continued on next page.)

1.3. EXTRAGALACTIC EMPIRICISM 11




 



















































Figure 1.11: The observed velocities of nearby





  galaxies are proportional to their distances, per-



 



 


 


 


 


 


 


 


 


 
  mitting distances to be estimated from Doppler




 

 

 

 

 

 

 

  shifts.















Astronomers often use Pogson quantified magnitudes, a logarithmic system to describe the incident
flux in a certain frequency bandpass. There are many different variants, but the one that will appeal
most to physicists is the “absolute” system with

mAB ≡ −2.5 log fν − 48.60 . (1.23)

The zero points of the various magnitude systems are arbitrary 2 , but they all try, to some extent,
to preserve Ptolemy’s scheme of 2000 years ago. The surface brightness µ (in dimensionless units of
magnitude per square arcsecond) is then

µ ≡ −2.5 log Iν − 48.60 − 2.5 log(206265)2 . (1.24)

Observers will immediately recognize the last quantity in parentheses as the number of arcseconds in
a radian (and the number of astronomical units in a parsec). As a matter of convention, the visible
and near infrared portions of the electromagnetic spectrum are split up into a number spectral bands
that have their origin in recipes for the glass that was used (by Harold Johnson) to construct filters.
In order of increasing wavelength they are called U, B, V, R, I, J, H, and K where the first five letters
stand for “ultraviolet”, “blue”, “visual”, “red”, and “infrared”. When giving the luminosity or
brightness of a source, one typically specifies a wavelength. For example, the background magnitude
of the night sky in the B filter is roughly µB ∼ 22.05 Bmag /arcsec2 , corresponding to a luminosity
surface density of 100L¯ /pc2 .
Empirically, it is found that many galaxies obey simple “laws” that give the surface brightness
as a function of radius from the galactic center. The best-known of these empirical relations are the
Freeman disk and the deVaucouleurs spheroid :
 ³ ´
 I0 exp rr
 : Freeman disk
· s
I(r) =  ³ ´1/4 ¸ (1.25)
 I0 exp −7.67 rr : deVaucouleurs spheroid.
e

For disks, the central intensity is such that at r = 0, the surface brightness is µ 0 ∼ 21.5 Bmag /arcsec2 .
For spheroids, the effective radius, re , is defined so that half the light lies interior to it. At r = re
the surface brightness is typically µe ∼ 22.5 Bmag /arcsec2 . These are both very close to the surface
brightness of the background night sky, which is made up primarily of zodiacal light and terrestrial
airglow. This makes it relatively difficult to detect low surface brightness (LSB) galaxies, and those
given to fretting (or quibbling) might worry that we were missing a large fraction of the universe
2
This particular definition gives a magnitude for the star Vega of mAB ' 0 at 5556Å
12 CHAPTER 1. GALAXIES: DYNAMICS, POTENTIAL THEORY, AND EQUILIBRIA


TU LM KJ NO

DE FG IH
?=< > BC@A
76 45 ;: 89
 +* ,- ./ 3201 Figure 1.12: A schematic plot of redshift versus distance for
log
 ! () "# &'
Doppler
     a sample of galaxies. The slope is unity, implying a linear
velocity V XV
XW relation. Notice that a few points lie far from the mean. Real
 RS
   PQ data is never pretty.
   $%

log distance

in failing to find such galaxies. Recent work by Blanton et al. using the Sloan Digital Sky Survey
indicate that at most ∼ 5% of the total luminosity density of the Universe might be lurking in galaxies
too faint to detect against the surface brightness of the night sky.
In addition to these scaling laws that describe the surface brightness of a single galaxy, there
are also a number of relationships that describe how galaxies of different luminosities are distributed
throughout the Universe. Unfortunately, intrinsic luminosities cannot be determined without knowing
the distance to a galaxy. Distances are difficult to measure. This would be a serious problem were
it not for the Hubble relation, which tells us (to first order) that distance is proportional to Doppler
shift, D = v/H0 (see Figure 1.12). The Hubble constant, H0 , was poorly known until recently.
In the absence of a definitive measure of H0 , it was absorbed into many results, giving an overall
map of the universe with a sliding scale. An example is Schechter’s Law, an empirical expression for
the “luminosity function” [# galaxies/unit volume with luminosity between L and L + dL],
µ ¶α µ ¶
? L L
φ(L)dL = φ exp − ? d(L/L? ). (1.26)
L? L

Schechter’s law works best for galaxies in the range 0.003 < L/L? < 3 where L? (' 1010 L¯ h−2 ) is
determined from a fit giving m?B = −19.75 + 5 log h, φ? = 2.0 × 10−2 h3 Mpc−3 , and α = −1.09. In
each of these parameters, h is the dimensionless expansion factor of the Universe as a fraction of the
canonical value for the Hubble constant3 :
H0 v/D
h≡ = . (1.27)
100 km/s/Mpc 100 km/s/Mpc

The luminosity function φ(L) has a form often seen in physical processes, a power law dependence
with an exponential cutoff, shown schematically in Figure 1.13. Although φ(L) diverges for faint
galaxies (L → 0), the integrated luminosity is finite and actually analytic:
Z
φ(L)LdL = Γ(α + 2)φ? L? = (1 + α)!φ? L? . (1.28)

Given the difficulty of measuring absolute distances and magnitudes of galaxies, it is useful to
establish that one or another observable property of galaxies is correlated (the more strongly the
better) with the intrinsic luminosity of a galaxy. The Tully-Fisher Law for spiral galaxies is one such
3
The best measurements today give h ' 0.7, or equivalently, H0 ' (14Gyr)−1
1.4. POTENTIAL THEORY 13

log # Figure 1.13: A schematic representation of the lu-

per unit minosity function for galaxies, cutting off sharply
volume
at the bright end and diverging slowly at the faint
end.
absolute magnitude

empirical relation, between the circular velocity vc and the total luminosity, L, of a galaxy, with
µ ¶1/4
L
vc ' 250 km/s ? (1.29)
L
and the Faber-Jackson Law gives a similar relationship for the line-of-sight velocity dispersion σ los
for elliptical galaxie, :
µ ¶1/4
L
σlos ∼ 220 km/s ? . (1.30)
L
The fundamental plane relation for ellipticals is an extension of the Faber-Jackson relation to include
a dependence upon surface brightness. It is usually expressed in terms of deVaucouleurs’ effective
radius re rather than luminosity,
re ∝ σe1.49±0.05 Ie−0.75±0.01 . (1.31)
Here effective radius, re , depends upon distance, but the velocity dispersion and intensity do not.
Another two techinques for estimating galaxy distances involve the luminosity functions for glob-
ular clusters and planetary nebulae within galaxies. Yet another technique (the surface brightness
fluctuation method) involves measuring the graininess of the observed surface brightnesses of ellipti-
cal galaxies due to the finite numbers of stars in each square arcsecond. The best technique requires
patience and luck – observing a supernova within a galaxy.