Professional Documents
Culture Documents
Astronomy
Astronomy
3. The local universe is expanding uniformly and isotropically according to Hubble’s law. We
assume that peculiar velocities can be described local gravitational effects.
4. The Hubble time, H−1, the ages of the oldest stars, and the radioactive dating of terrestrial
elements all approximately agree, i.e. of the order of a few billion years in each case.
7. The observed abundance of H, He and Li (and their isotopes) agree with the predictions of
nucleosynthetic reactions occurring within the early universe.
8. The present day universe is structured (galaxies, clusters and superclusters) - note that this
point does not contradict point (5) above.
9. Dynamical, X-ray and lensing studies imply that most of the matter in the universe is dark.
10. The CMB displays temperature fluctuations of order (∆T/T )rms ~10−5 on scales of order
one degree.
11. Observations of supernovae type Ia (SNe Ia) and the CMB indicate that the universe is
spatially flat (see later) and is expanding at an accelerating rate (see much later). Note that
though many observational tests agree with SNe Ia + CMB results, the SNe Ia + CMB results
are the simplest and best understood.
1 A MATHEMATICAL DESCRIPTION OF THE UNIVERSE 2
2. The cosmological principle (an extension of the Copernican principle) assumes that we occupy
no special place in the universe.
3. The combination of observed isotropy with the cosmological principle implies that the universe
is homogeneous.
Depending upon the exact definition of gµν, the diagonal elements are g00, g11, g22, g33 = +1,−1, −1, −
1 for
Euclidean or flat space (Index 1 refers to time and indices 2, 3 and 4 refer to space). Non–diagonal
elements are zero in this case.
One of the early goals of theoretical cosmology was to develop a spacetime metric or line element that
would satisfy the aforementioned cosmological principle, i.e. homogeneity and isotropy. Robertson
1 A MATHEMATICAL DESCRIPTION OF THE UNIVERSE 3
and Walker (1935) demonstrated that in any homogeneous and isotropic cosmological model based
upon Riemannian geometry, the most general line element is
#
ds = c2 dt2 — a(t) " dr2 2 + r 2 (dθ2 + sin2
2
2 1 − kr θ dφ2) , (3)
Figure 1: The geometry of a two dimensional surface considered embedded within a three dimen-
sional space.
The measured circumference Cmeas = 2πR sin θ and θ = D/R. Rearranging this yields
So the geometry measured on the two dimensional surface is modified by a term depending upon the
inverse square of the effective radius of curvature. The circumference of a circle drawn on a surface
of positive curvature is less than the circumference expected based upon Euclidean geometry (with
the opposite being true for a surface of negative curvature). As the radius of curvature tends to
infinity (i.e. flat space), the properties of a circle drawn upon the surface revert to the Euclidean
case.
as shown in Figure 2a. For an example of a line element describing a curved 2D surface we again
consider the surface of a sphere of radius R (Figure 2b). Here K = 1/R2 and the distance between
1 A MATHEMATICAL DESCRIPTION OF THE UNIVERSE 5
However, as the expansion of the universe will affect all of its geometric properties, including cur-
vature, we define the time dependent curvature in terms of a time independent curvature constant,
i.e.
k
K(t) ≡ 2 . (13)
a (t)
Substituting for r and K in Equation 11 we obtain
2 2
!2
2
(ds) = (c — a (t) √ dω + (ω
2 2
+ (ω sin θ dφ) , (14)
dt) 1 − kω2 dθ)
which is the Robertson-Walker line element presented in Equation 3 (note the slight change in
variable representation).
1 A MATHEMATICAL DESCRIPTION OF THE UNIVERSE 7
1920: The Shapley–Curtis debate. Shapley argued correctly that the sky distribution of globular
clusters indicated that the Sun was located far from the centre of the Galaxy. However, he
incorrectly believed (using an erroneous result from van Maanen) that the spiral nebulae
displayed observable rotational proper motions and were thus located within the (albeit
very large) Galaxy. Curtis argued that the spiral nebulae were extra-galactic, mainly by
assuming that novae observed in the nebulae were similar to novae observed in our galaxy
and then following distance dimming arguments. However, he mistrusted Shapley’s
Cepheid-based distance to the globular clusters and favoured a smaller, Kapteyn Galaxy
(based upon star counts) with the Sun close to the centre.
1925: Using Cepheid variable stars Hubble demonstrates that M31 and M33 lie at 285 kpc from
the Galaxy and that NGC 6822 lies at > 214 kpc (note that these distance estimates were
much too low due to a problem with Cepheid brightness calibration). All appear to be
extragalactic, isolated systems.
1929: Using Cepheids to calibrate a secondary distance indicator (the brightest star in a galaxy),
Hubble extended his distance–recession velocity sample and reported the relation v = H d
where H ≈ 500 kms−1 Mpc−1 (the value for H is so high compared to the modern value due
to the aforementioned problem with Cepheid calibration – if d is biased low, H will be biased
high).
v = dl
l(t)
H(t) = a˙
=
dt (16)
a
a˙
a
A similar result is obtained working straight from the RW line element: one notes that kr2 =
a2r2/R2 and assumes thatqthe quantity ar/R 1. The radial proper distance is related to the
metric distance by dl(t) = −(ds) ≈ a(t) dr. The velocity of the object is then
2
dl a˙ a˙
v= ≈ a˙ r = l ⇒ H(t) = . (17)
dt a a
The assumption of ar/R 1 simply tells us that Hubble’s law is only valid on scales much less than
the global curvature constant. The conclusion from Milne’s analysis and the analysis of the RW line
element is that the apparent recession of galaxies with us located at the origin is naturally explained
as a universal expansion of all test particles (galaxies) within a homogeneous and isotropic universe.
The geometric model of the universe receives its first ‘tick’.
Consider a spherical volume of the universe of radius l and mass M . We further consider the
dynamical behaviour of a test particle (a single galaxy if you like) of mass m located on the surface
of this spherical shell. Birkhoff’s theorem states that the mass within the sphere will act upon
the test particle as if the entire mass M were concentrated at the centre of the sphere. From the
Newtonian equation of motion of the test particle we therefore obtain
d 2l GMm
m =− . (18)
dt2 l2
Multiplying the equation by l˙ generates
d l˙ 2 GM
d = . (19)
l
dt 2 dt
Integrating yields
l˙2 GM
= E, (20)
−l
2
where the integration constant has units of energy. Note: this equation may be interpreted as a
basic energy relation of the form, Kinetic + Potential = Total. The RW line element indicates that
l(t) = l0 a(t), where l0 is independent of time and a(t) is the time–varying, universal scale factor.
Therefore, we may write
!
l˙ 2 G 4πl3ρ(t)
= E
− l 3
2
a˙ 2 4πGρ(t)a2(t) E
— = l02 ,
2 3
2 8πG 2E
a˙ — 3 ρ(t) a2(t) = . (21)
l02
The result is one form of Friedmann’s equation and it has three general solutions:
E < 0 : The potential term is proportional to 1/a(t) and always dominates (see Equation 20). Hence
a(t) cannot increase without limit, instead it must reach some maximum value (at which point
a¨ < 0) and decrease.
E = 0 : a(t) increases throughout time, tending toward (but never reaching) an asymptotic maxi-
mum scale as t → ∞. The critical density value corresponding to this case is
3
2 3H2(t)
ρc(t) = a˙ = . (22)
8πG 8πG
a
The current value of the Hubble parameter is H0 = H(t0) = 70 ± 7 kms−1 Mpc−1 and the
corresponding value of the critical density is ρc,0 = (9.2 ± 1.8) × 10−27 kg m−3 = (1.4 ±
1 A MATHEMATICAL DESCRIPTION OF THE UNIVERSE 10
0.3) × 1011 M⊙ Mpc−3. Current estimates of the space density of galaxies are of order 1.4 ×
10−2 Mpc−3. If one assumes that the mass of a typical galaxy is 1011M⊙, then one concludes
that visible galaxies contribute about 1% of the value of ρc,0.
E > 0 : a(t) increases monotonically for all time. The universe expands forever.
These three cases are shown in Figure 4. Returning to the form of Friedmann’s equation, one may
2
a˙ 8πG kc2
a — ρ(t) = − . (24)
3 a2
Note that this form of the equation was that obtained by Friedmann (and re–discovered by Lemaitre)
by inserting the corresponding elements of the metric tensor applicable to the RW line element into
the EFQ. The parameter k takes one of three values
E < 0 ⇒ k = +1
E=0 ⇒ k=0
E > 0 ⇒ k = −1. (25)
1 A MATHEMATICAL DESCRIPTION OF THE UNIVERSE 11
We have now obtained the complete link between the time dependence of the universal scale factor
a(t) and the curvature constant k first encountered in the RW line element. Proceeding further,
one may redefine the Friedmann equation in terms of a new dimensionless density parameter
ρ(t)
Ω(t) = . (26)
ρc(t)
If we divide the Friedmann equation (Equation 24) by H(t)2 and use the above identity, we obtain
kc2
1 − Ω(t) = − 2 2 , or
Ha
kc2
= Ω(t) − 1. (27)
H2a2
This expression has the following consequences
k = +1 ⇒ Ω(t) > 1
k = 0 ⇒ Ω(t) = 1 for all times,
k = −1 ⇒ Ω(t) < 1 (28)
where
Therefore, the total matter/energy content of the universe determines the overall spatial geometry,
the time variation of a(t) and the ultimate fate of the universe2 Importantly, these equations indicate
that the universe has a well–defined characteristic matter density that marks the limit of each of
the above cases. Therefore, the determination of the total matter content of the universe became
an immediate challenge for early observational cosmologists. However, to determine this critical
density, one requires H(t) or H0 – the present day value of the Hubble parameter.
why should it be anywhere close to 1 at the present epoch?”. If Ω /= 1, why do we appear to be living in a special
epoch?
1 A MATHEMATICAL DESCRIPTION OF THE UNIVERSE 12
energy density urel to the universal mass-energy budget. The effective matter density of relativistic
particles is ρrel = urel/c2. In addition, observations of SNe Ia indicate that the expansion of the
universe is accelerating. One explanation of this effect may be the presence of a “Cosmological
Constant” referred to as Λ. Although these considerations will be described in later chapters, we
note here the modified form of the Friedmann equation required to incorporate these effects into
our model of the universe:
2
a˙ 8πG 1 2 kc2
— [ρm(t) + ρrel(t)] − Λc =− 2 . (30)
a 3 3 a
1.7 The Fluid and Acceleration equations
At present we have one equation – the Friedmann equation – and two unknowns, a(t) and ρ(t).
The Friedmann equation can be thought of as an expression of Newtonian energy conservation for
a test mass in an expanding/contracting universe. Another approach is to consider a
thermodynamic expression of energy conservation. From the first law of thermodynamics we
have
dQ = dE + PdV, (31)
where dQ is the heat flow into or out of a region, dE is the change in internal energy, P is the
pressure and dV is the change in volume of the region. In a homogeneous universe we must have
dQ = 0, i.e. no bulk flow of heat. Processes for which dQ = 0 are known as adiabatic. Therefore,
applying the first law of thermodynamics to an expanding universe, we may write
E˙ + P V˙
= 0. (32)
We return to our previous example of an expanding spherical region of the universe where the
physical radius l(t) = l0 a(t). The volume of this region is
4π 3 3
V (t) = l a (t), (33)
0
3
and the rate of change of volume is
4π a˙
V˙ = l3 (3a2 a˙ ) = V 3 . (34)
3 0 a
The internal energy of the sphere is
a˙
E˙ = V ρ˙c2 + V˙ ρc2 = V ρ˙c2 + 3ρc2 a . (36)
1 A MATHEMATICAL DESCRIPTION OF THE UNIVERSE 13
Therefore, combining equations 32, 34 and 36 we can re-write the first law of thermodynamics in
the expanding region of the universe as
a˙ a˙
V ρ˙c2 + 3 ρc2 + 3 P = 0, or
a a
dρ a˙ 2
+ 3 (ρ + P/c ) = 0. (37)
dt a
This is the Fluid equation and is the second key equation describing the expansion of the universe.
Combining the Friedmann and Fluid equations generates the Acceleration equation providing
the rate of change of universal expansion. Multiplying the Friedmann equation by a2 generates
8πG
a˙ 2 = ρ a2— kc2. (38)
3
Taking the time derivative of this equation we have
8πG
2a˙ a¨ = (ρ˙a2 + 2ρa˙a). (39)
3
Dividing by 2a˙ a
!
gives a¨ 4πG dρ a
= + 2ρ . (40)
a 3 dt
Substituting the Fluid equation into this term yields
a˙
a¨ 4πG 3P
a =− 3 ρ + c2 . (41)
We see that a universe consisting of material with a positive energy density, i.e. ρc2 > 0, has the
effect of slowing down the expansion with time. Universal components such as baryons and photons
each exert positive pressure (the result of either random kinetic motion or radiation pressure). The
effect of such pressure terms is also to slow down expansion. However, should a universal component
contribute a negative pressure such than P < —ρc2/3, the net effect will be to increase the rate of
expansion with time. Dark energy (see Lecture 8) is an example of such a universal component.
for dilute cosmological “gases”, i.e. a gas of matter or photon particles. Consider a low density
gas of non-relativistic massive particles – such as galaxies. Non-relativistic means that the random
motions of the particles are small compared to the speed of light. Such a gas obeys the perfect gas
law ρ
P = kT, (43)
µ
where µ is the mean mass of the particles. For a non-relativistic gas, the temperature T is related
to the root mean square thermal velocity ⟨v2⟩ by
ρ ∝ a−3 or
3
ρ a0
ρ0 = a . (46)
Note that this equation is really only telling us that the matter component of the universe is
conserved. One may then re-write the Friedmann equation in the form
8
a a˙ 2 = πGρ a3
3
a a˙ 2 = constant
a1/2 da = A dt
2
a3/2 = A t + C (C = 0 as a → 0 as t → 0)
3
or a ∝ t2/3. (47)
The assumption that the constant of integration is zero seems innocuous. However, it gave
Lemaitre considerable trouble. With early estimates of the Hubble parameter biased high, this in
turn implied a troublingly low value for the age of the universe when compared to the estimated
ages of the stars. Lemaitre attempted to reconcile these two opposing age estimates by setting C
> 0 and thus allowing the universe to exist in a static state for some arbitrary period of time prior
to expanding. With hindsight this appears to be (and indeed is) a fudge – but one which at the
time was forced upon Lemaitre by observations.
Note that H = a˙ /a, therefore
H(t) = a˙ /a = (2/3t−1/3)/t2/3, (48)
which can be rearranged as
2 1
t0 = . (49)
3 H0
A further case considers an open universe (ρ ρc), Λ = 0 scenario, i.e.
2
a˙ kc2
a = 0−
a2
a˙ 2 = A (50)
1 A MATHEMATICAL DESCRIPTION OF THE UNIVERSE 16
√
which leads to a = A t or, in terms of H one may write
H = a˙ /a = 1/t
This is the Hubble time – the maximum time elapsed since a = 0 for a Λ = 0 universe. Note that
the actual age of the universe in this case is a little bit less than 1/H0. The degree of inaccuracy in
using a tangent to the current expansion curve will lie in the extent of the deviation of ρ from
zero. Therefore, consideration of the Friedmann and Fluid equations for several basic models
generates a characteristic age of the universe in terms of H0.
3. The fact that a(t) is increasing with time implies that it was smaller in the past. We therefore
expect the pressure term (radiation from CMB and neutrinos together with random matter
velocities) to have been more important in the past – see Lectures 3 and 4.
1 A MATHEMATICAL DESCRIPTION OF THE UNIVERSE 17
1.11 Asides
1.11.1 The nature of redshift within the cosmological model
Redshift was originally defined in purely observational terms. However, cosmological redshift
arises naturally from a consideration of the RW line element. Consider the light path of a photon
travelling through the universe along a radial path (radial null geodesic), i.e. ds2 = dθ = dφ = 0,
" #
2 2
c dt dr2
2
= a (t) . (52)
1 − kr2
For a light pulse emitted at te and observed at to by an observer at a distance re, one may write
∫ ∫
to c dt re
= √ ≡ f (r ). (53)
dr a(t) 1 − kr2
te 0 e
Note, that f (re) is a fixed (or comoving) distance. If one has a good enough understanding of a(t)
and k one can determine the relation between distance and time in the universe. However, one
may consider a second pulse of light emitted and observed at te + ∆te and to + ∆to respectively, i.e.
∫ ∫
to+∆toc dt c dt to
te+∆te a(t)− te a(t)
= 0
c ∆to c ∆te
− = 0
a(to) a(te)
c ∆t o
c ∆te = a (t o ) (54)
a(te)
However, c ∆t = λ, and the above analysis is related to the (observationally defined) redshift by
λo − λe = λo − 1 = a(to) − 1. (55)
z=
λe λe a(te)
This expression is normally written as 1 + z = a0/ae. Cosmological redshift describes the relative
expansion of the universal scale factor between the epochs of emission and observation 3. Note:
the observed redshift of an astrophysical source is often a combination of cosmological redshift
and other physical effects, e.g. gravitational redshift, the Sachs–Wolfe effect and peculiar
velocities. Peculiar velocities take alter the redshift via czobs = czcosmo + vpec(1 + zcosmo) and will
discussed further in Lecture 2.
3
Though redshift may also be defined via SR as arising from the recession velocity associated with the Hubble
flow, SR is only valid in locally flat reference frames and describes the case where (effectively) gµν = constant.
1 A MATHEMATICAL DESCRIPTION OF THE UNIVERSE 18
s0 se
a(t0) = a(t )
e
s0 = se a(t )
a(te) = s e (1 + z). (56)
0
c dt = a(t) dr
√
1 − kr2
∫ ∫
t0 c dt rH
=
dr √ , (57)
0 a(t) 1 − kr 0 2
Where t0 indicates the current age of the universe and rH is the horizon distance. For an EdS
universe with k = 0 we may write
∫ t0 c dt
rH =
0
a(t)
∫ t0 c dt t02/3
=
0 a0 t2/3
= 3 c t 0. (58)
1 A MATHEMATICAL DESCRIPTION OF THE UNIVERSE 19
Consideration of the cosmological horizon at the current epoch provides one method to answer
Olbers’ paradox, or “why is the night sky dark?”. The universe may well be infinite in space,
but it is finite in time. The combination of a finite age of the universe with a finite speed of light
ensures that we can only observe photons from a finite region of the universe.
However, computation of the cosmological horizon gives rise to the horizon problem. When we
look at the universe at the cosmological horizon, it is isotropic, i.e. large scale structure and CMB
temperatures are statistically identical even though separated by 180◦ on the sky. However, these
two regions, though in casual contact with us, are themselves causally isolated. How could the
CMB and LSS have developed in exactly the same manner? The answer to the horizon problem
is provided by cosmic inflation that postulates that the universe underwent a rapid phase of
expansion at early times. Regions in the early universe were originally in causal contact and were
subsequently inflated to scales much larger than the horizon during the inflationary epoch.
1 A MATHEMATICAL DESCRIPTION OF THE UNIVERSE 20
2. Universal spacetime is assumed to be isotropic and homogeneous. The most general line ele-
ment describing an isotropic and homogeneous spacetime is the RW line element (also referred
to as a metric). The line element describes the infinitesimal distance between neighbouring
coordinates.
3. The universal scale factor a(t) of the RW line element evolves with time. The relationship
between scale factor and time depends upon the universal mass density, pressure and presence
of a Λ–term. The general form of the Friedmann equation indicates that the universe may
be expanding or contracting. We shall see during our investigation of Λ that deriving a static
universe is difficult.
4. The Hubble relation indicates that the universe is currently expanding. For the simple case
of the EdS universe, the overall geometry (curvature) of the universe, and its eventual fate,
are determined by the total matter density and the current value of the Hubble parameter.
5. Therefore, coordinated observational tests (total matter, oldest age, current expansion rate)
may constrain parameter values that define the cosmological model.
Note that such observations do not tell us whether the cosmological model as we have designed
it is correct (though it is clearly a good approximation to reality), simply which cosmological
model values our universe resembles.