Multi-Probe Cosmology

Riccardo Murgia
March 2023

written by Ulyana Dupletsa

These notes are based on the course given at the Gran Sasso Science Institute (GSSI) by
Riccardo Murgia in March 2023. The main objective is to present an overall picture of what are
the state-of-the-art challenges in this field, especially for what concerns the cosmological tension
in the measurements of the expansion rate of the Universe. Furthermore, the course focuses on
giving a general overview of what are the main probes that allow us to investigate the Universe.
The major reference on which all the content is based is the Cosmology book by Daniel Baumann.
Contents
1 A Cosmologist’s Starter Kit 3
1.1 FRW Metric and its properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Hubble law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3 Redshift and Distances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3.1 Luminosity distance (standard candles) . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3.2 Angular diameter distance (standard rulers) . . . . . . . . . . . . . . . . . . . . . . 6
1.4 Dynamics of the FRW Universe: the Friedmann Equations . . . . . . . . . . . . . . . . . . 6
1.5 Friedmann equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2 Thermal History of the Universe 10

2.1 H(z) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.2 Species evolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.3 Decoupling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.3.1 Dark matter freeze-out . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

3 The Inhomogeneous Universe: LSS 15

3.1 Structure growth in Newtonian theory: sub-horizon scales . . . . . . . . . . . . . . . . . . 15
3.2 Super-horizon scales . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.3 Statistical properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.4 Matter power spectrum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

4 The Inhomogeneous Universe: CMB 22

4.1 The evolution of the photon-baryon fluid . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
4.2 Cosmic sound waves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
4.3 From primordial sound waves to CMB anisotropies . . . . . . . . . . . . . . . . . . . . . . 24
4.3.1 Baryonic Acoustic Oscillations (BAO) . . . . . . . . . . . . . . . . . . . . . . . . . 25
4.4 The CMB anisotropies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
4.5 CMB angular power spectrum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
4.6 Impact of cosmological parameters on the CMB . . . . . . . . . . . . . . . . . . . . . . . . 28

2
1 A Cosmologist’s Starter Kit
1.1 FRW Metric and its properties
Spacetime metric can be written in the form:

ds2 = gµν dxµ dxν (1)

Assuming spatial homogenuity and isotropy the line element can be written as:

ds2 = −c2 dt2 + a2 (t)dl2 , (2)

where 
dr2  0 flat

dl2 = kr2
+ r2 dΩ2 with k = −1 hyperbolic . (3)
1− R02

+1 spherical


Combining 2 and 3 we can write the full line element:

 
dr2
ds2 = −c2 dt2 + a2 (t)  kr2
+ r2 dΩ2  . (4)
1− R02

We went from 10 independent components of gµν to a single function on time a(t), the scale factor,
and a constant, R0 , the curvature scale today.

• The metric satisfies the following rescaling symmetries:

a → λa
r → r/λ (5)
R0 → R0 /λ

We can exploit this to set the value of the scale factor today at t = t0 to be unity: a(t0 ) ≡ 1.

• Redefining r as:
dr
dχ = q (6)
kr2
1− R02

we can write the line element in a very simple form:

ds2 = −c2 dt2 + a2 (t) dχ2 + Sk2 (χ)dΩ2 =⇒ ds2 = −c2 dt2 + a2 (t)dl2
 
(7)

where χ
 k=0
R0  




χ


Sk2 (χ) ≡ sinh R k = −1 (8)

 0
 
χ


 sin k = +1


R0

• We can write the conformal time η:

dt
dη = (9)
a(t)
and write the line elements as:

3
 ds2 = a2 (η) −c2 dη 2 + dχ2 + Sk2 (χ)dΩ2


(10)

The metric has been now factorized into a static part and a time-dependent conformal scale factor
a(η).

1.2 Hubble law

Starting from the relation between comoving and physical distance

rphysical = a(t)rcomoving (11)

and taking the time derivative:

drphysical da d~rcomoving
= ~rcomoving + a(t)
dt dt dt
ȧ d~rcomoving (12)
= a~rcomoving + a(t)
a dt
= H~rphysical + ~vpeculiar

we end up with the Hubble law:

ȧ
~vgal = H~rphysical + ~vpeculiar with H≡ (13)
a

where H is the Hubble parameter, H~rphysical is known as the Hubble flaw (the velocity of the
galaxy resulting from the expansion of space) and ~vpeculiar is the velocity measured by an observer who
is comoving with the Hubble flaw.

1.3 Redshift and Distances

Given that due to the Universe’s expansion any length scales as the scale factor, we have that energy
scales as:
h
λ ∝ a =⇒ E = ∝ a−1 (14)
λ
This allows us to define the redshift. For t0 > t1 :
a(t0 )
λ0 = λ1 (15)
a(t1 )

so that the wavelength we receive is stretched.

Now consider a galaxy at a distance d. Isotropy allows us to choose coordinates in which the light
emitted by such a galaxy is traveling along the radial direction. Photons move along a null light-like
geodesic, so for a light ray ds2 = 0, which implies:

a2 (η) −c2 dη 2 + dχ2 = 0 =⇒ ±c∆η = ∆χ(η)

 
(16)

where the minus sign is for incoming photons and the plus sign for the outgoing ones. Note the advantage
of working with conformal time: in the χ − η coordinates light rays follow straight lines.
Given the following quantities:

1. emission time from a source η1 (the signal duration interval is short and is δη)

4
 d
2. arrival time at the telescope happens at η0 = η1 + δη = η1 + c

3. the conformal duration of the signal is the same at the source and at the detector, the physical
one instead is given by:
δt1 = a(η1 )δη
(17)
δt0 = a(η0 )δη

We have that the light is emitted with wavelength λ1 = cδt1 and is observed with wavelength λ2 = cδt2 :

λ0 cδt1 a(η0 ) 1
= = =⇒ 1 + z = (18)
λ1 cδt0 a(η1 ) a(t1 )
where a(t0 ) = 1 by definition and where we have used the conventional definition of redshift as the
fractional shift in wavelength:
λ0 − λ1
z≡ (19)
λ1
• If z  1, which is true for nearby sources, we can expand around t = t0 :
1
a(t1 ) = 1 + (t1 − t0 )H0 + · · · =⇒ z = H0 (t0 − t1 ) + (2 + q0 ) H02 (t1 − t0 )2 (20)
2
where  
ä
q0 ≡ − (21)
aH 2 t=t0

is the deceleration parameter. Therefore, at first order:

z = H0 (t0 − t) + · · · =⇒ v ≡ cz ' H0 d (22)

| {z }
d/c

where v is the recession velocity and v = H0 d is known as the Hubble-Lemaitre law.

• If z > 1 we should carefully treat the notion of distance:

Distances appearing in the metric are not observable, because they do not take into account
that the Universe is expanding and that light takes a finite amount of time to reach us.

1.3.1 Luminosity distance (standard candles)

A way to measure distances is to use the so called standard candles. These are objects of known intrinsic
brightness, so that their observed brightness can be used to determine their distances.
Consider a source at redshift z. The comoving distance to that object is defined as:

dz 0
Z t1 Z z
dt
χ(z) = c =c 0
(23)
t0 a(t) 0 H(z )

The observed isotropic flux emitted by this source, if the space was static, would be:
L
F = (24)
4πχ2
Since we need to take into consideration the expansion of the Universe:

5
 1. 4πχ2 → 4πa2 (t0 )d2M with dM the metric distance
a(t1 ) 1
2. the photon arrival rate is different from the photon emission rate: a(t0 ) = 1+z

1
3. the emitted and observed energies are different by an additional 1+z factor

Therefore:
L L
F = ≡ (25)
4πd2M (1 + z)2 4πd2L

dL (z) = (1 + z)dM (z) (26)

1.3.2 Angular diameter distance (standard rulers)

An alternative way to measure distances is using standard rulers, which are object of known physical
size. The observed angular size of these objects depend then on the distance. The typical size of hot
and cold spots of the CMB can be predicted theoretically and is therefore a standard ruler.
Suppose the transverse physical size of an object is D.In a static Euclidean space then we would expect
its angular size to be:
D
δθ = (27)
χ
where we have assumed δθ  1 (in radians), which is a good assumption for cosmological objects in
general. If we add expansion:
D D
δθ = ≡ (28)
a(t1 )dM dA
Notice that the observed angular size depends on the distance at the time t1 when the light was emitted.

dM
dA = (29)
1+z

Pay attention that luminosity distance and angular diameter distance are not independent:

dL (z) = (1 + z)2 dA (z) (30)

1.4 Dynamics of the FRW Universe: the Friedmann Equations

To study the evolution of the scale factor a(t) we start from Einstein’s equations:

8πG
Gµν = Tµν (31)
c4

where    
T00 T0j scalar (energy density) vector (momentum density)
Tµν = = (32)
Ti0 Tij vector(energy flux) stress tensor
and making the following assumptions:

• homogenuity: T00 = ρ(t)c2

• isotropy: T0j = Tj0

6
 • isotropy around a point ~x: Tij (~x = 0) ∝ δij ∝ gij (~x = 0)
we end up with the perfect fluid equation:
−ρc2
 
 
P P
Tνµ = g µλ Tλν = 
 
 =⇒ Tµν = ρ + 2 Uµ Uν + P gµν (33)
 P  c
P
with U µ = (c, 0, 0, 0) for a comoving observer. Imposing the conservation of the stress-energy tensor
∇µ Tνµ = 0 we end up with the continuity equation:
 
ȧ P
ρ̇ + 3 ρ+ 2 =0 (34)
a c

This equation describes energy conservation in the cosmological context. The usual energy conser-
vation in flat space as derived from Noether’s theorem relies on the symmetry under time translation.
This symmetry is broken for an expanding Universe. The usual notion of energy conservation does not
hold and is replaced by 34.
Most cosmological fluids can be parametrized in terms of a constant equation of state
P
ω≡ (35)
ρc2
so that we can write:
ρ̇ ȧ
= −3(1 + ω) =⇒ ρ = a−3(1+ω) (36)
ρ a

We have:
• matter (non relativistic particles): |P |  ρc2 =⇒ ρ ∝ a−3
• radiation (relativistic particles): ω = 1/3 =⇒ ρ ∝ a−4 (photons besides being diluted V ∝ a−3
get also redshifted ∝ a−1)
• dark energy: P = −ρc2 =⇒ ρ ∝ a0 = const. It could be:
vac = −ρ
– vacuum energy Tµν 2 60
vac c gµν , but ρvac from QFT is 10 ρΛ (the cosmological con-
stant problem)
– cosmological constant:
8πG (Λ) Λc4
Gµν + Λgµν = 4
Tµν =⇒ Tµν =− gµν ≡ −ρΛ c2 gµν (37)
c 8πG
– ωDE ∼ −1: varying equation of state

1.5 Friedmann equations

From Einstein’s equations we can derive the two Friedmann equations are:
 2
ȧ 8πG kc2
= ρ− 2 2
a 3 a R0
  (38)
ä 4πG 3P
=− ρ+ 2
a 3 c

7
In terms of H:
8πG kc2
H2 = ρ− 2 2 (39)
3 a R0
We define the critical energy density as the one needed to have a flat universe:

3H02
ρcrit,0 ≡ (40)
8πG

where the 0 subscript denotes that the quantities are evaluated today at t = t0 .
In this way we can define the abundance parameters as:
ρi,0
Ωi,0 = (41)
ρcrit,0

With this parametrization we can write the Friedmann’s equations as:

H 2 = H02 Ωr a−4 + ΩM a−3 + Ωk a−2 + ΩΛ

(42)

Evaluating both sides at t0 , with a(t0 ) = 1, leads to:

Ωk + Ω M + Ω k + Ω Λ = 1 (43)

One of the central goals in Cosmology is to measure the parameters appearing in 42 and hence determine
the composition of the Universe.

SUMMARY AND KEY POINTS

A Universe that is both homogeneous and isotropic is described by the Robertson-Walker met-
ric:  
dr 2
ds2 = −c2 dt2 + a2 (t)  + r2 dθ2 + sin2 θdφ2 


kr2
1 − R2
0

where k = −1, 0, +1 for, respectively, hyperbolic, flat and spherical space with radius of curvature
today R0 .
Light from distant galaxies experiences the expansion of the Universe and as a consequence the
wavelength is stretched according to the redshift value:

λobs − λem a(tobs ) 1

z≡ = − 1 =⇒ z + 1 =
λobs a(tem ) a

The velocity of galaxies, in turn, undergoes both the effects of peculiar motion and the expansion
of the Universe:
ȧ
~vgal = H~rphys + ~vpec with H ≡
a

8
The evolution of the scale factor a is determined by the Friedmann equations:
 2
ȧ 8πG kc2
= ρ− 2 2
a 3 a R0
 
ä 4πG 3P
=− ρ+ 2
a 3 c

where ρ represents the total energy density of the Universe. Requiring a flat Universe k = 0 we
can write the critical energy density:

3H02
ρcrit ≡
8πG
Each component of the Universe satisfies the continuity equation:
ρ̇ ȧ
= −3(1 + ω)
ρ a

with ω = 0 for baryons and dark matter (non-relativistic matter), ω = 1/3 for radiation (relativistic
matter) and ω = −1 for dark energy. Since densities can be written as ρ ∝ a−3(1+ω) , we can write
the two Friedmann equations as one equation:

H 2 = H02 Ωr a−4 + ΩM a−3 + Ωk a−2 + ΩΛ

with H0 the Hubble constant and Ωi ≡ ρi /ρcrit the dimensionless density parameters today.

9
2 Thermal History of the Universe
2.1 H(z)
Rewriting 42 in terms of redshift, we get:
q
(1+z)4 + ΩM (1 + z)3 +  (1+z)2 + ΩΛ

 

H(z) = H0  Ωr Ωk (44)

where Ωr is negligible (BAO). Then we have from CMB observations Ωr  1, ΩM ∼ 0.3 (this includes
both baryons ad cold dark matter) and ΩΛ ∼ 0.7. Generally we can write:

1 p
H(z) ∝ p =⇒ H(z) = H0 ΩM (1 + z)3 + ΩΛ (45)
ρ(z)

On the other side, ΩΛ starts dominating after z = 2. Studying H(z) (see Eq. 45) allows us to determine
whether the dark energy is really constant.

Figure 2.1: Evolution of the energy densities in the Universe. We can distinguish three epochs: radiation
domination (RD), matter domination (MD) and dark energy domination (DED). Figure taken from
Cosmology by Baumann (Fig. 2.10).

The interplay between different components allows us to write the thermal history of the Universe.
The CMB spectrum is a strong observational evidence that the early Universe was in a state of thermal
equilibrium.

10
 PLASMA TRANSPARENT

RD MD DE
10−43 s 10−35 s 10−11 s 10−5 s 3 min 105 yrs 3 · 105 yrs 7 · 108 yrs 3 · 109 yrs 6 · 109 yrs
t

∼ 3000 ∼ 1100 ∼ 10 ∼2 <1

z

∼ 1 eV ∼ 0.1 eV
T
QG GUT EW quark/hadron BBN M-R eq CMB galaxies stars peak M-DE eq

2.2 Species evolution

The early Universe was a hot gas of weakly interacting particles. The description of such a system is
better done using statistical mechanics.
We start from the probability distribution function of the momentum, f (p, t), which can either follow
Fermi-Dirac or Bose-Einstein statistics. The following quantities relate the microscopic properties to
the macroscopic ones, respectively, the number density n(T ), the energy density ρ(T ) and the pressure
P (T ): Z
n(T ) ∝ d3 pf (p, t)
Z
ρ(T ) ∝ d3 pf (p, t)E(p) (46)
p2
Z
P (T ) ∝ d3 pf (p, t)
3E(p)
p
with E(P ) = p2 + m2 . Each particle species has its own distribution function f (p, t), number and
energy densities and pressure. Species that are in thermal equilibrium share the same temperature T .
What we need to compare is the interaction rate Γ with the expansion rate of the Universe H. At
early times, during radiation domination era, we have that Γ  H: particles are in thermal equilibrium
and relativistic (the temperature is much larger than the particle mass):

T  m =⇒ ρ ∝ a−4 , n ∝ T 3 =⇒ ρ ∝ T 4 ∝ (1 + z)4 (47)

In detail:

π2
g∗ T 4
ρ= energy density
30
ρ
P = pressure (48)
3
ρ+P 2π 2 5 3
S= = g T entropy
T 45 ∗

In the non relativistic limit (when the temperature drops below the particle mass), during MD era, we
have:
mT 3/2 −m/T
 
T  m =⇒ n ∝ e (49)
2π

11
 3
ρ ∼ mn + nT
2 (50)
P = nT

Moreover, when the Universe is radiation dominated, we can write the Friedmann’s equation as H 2 ∝ T 4 ,
which gives us the relation between temperature and H(z). As a rule of thumb we can write the following
expression:
 1/2
−1 1/2 T 1s
T ∝a ; a∝t ; ∝ (51)
1Mev t

2.3 Decoupling
We saw that when the temperature drops below a particle’s mass, its number density exponentially
decays (remember the exp (−m/T ) factor). In order to survive till the present time, these particles must
drop out of equilibrium much before the m/T term becomes too large. This decoupling occurs when the
interaction rate of particles becomes much smaller than the expansion rate of the Universe.
While Γ  H, the abundance of a certain species tracks its equilibrium value. Once Γ becomes smaller
than H, the species drops out of equilibrium and decouples from the plasma. To follow the evolution
of each species we need to solve the Boltzmann equation:
d log N1 Γ
= − [. . . ] (52)
d log a H
When a species decouples in the non relativistic regime, it does not interact any more and its abundance
is frozen.

Figure 2.2: Scheme of how particle freeze-out happens. When the interaction rate drops below the ex-
pansion rate, the particles freeze out and maintain a constant relic density. Figure taken from Cosmology
by Baumann (Fig. 3.9).

12
2.3.1 Dark matter freeze-out
Let us apply the Boltzmann equation to dark matter. This can also be a model of how dark matter was
produced in the early Universe. Consider a heavy fermion x that can annihilate with its anti-particle x̄
to produce two light particles. The particles x, x̄ may be dark matter, whereas l could be particles of
the Standard Model.

|x {z
+ x̄} ↔ + ¯l
l|{z} (53)
heavy particle/antiparticle (DM) light products (SM particles)

Then we make the following simplifying assumptions:

• The light particles l are tightly coupled to the rest of the plasma so that they maintain their
equilibrium densities

• We neglect any initial asymmetry between x and x̄ number densities

• We assume there is no other particle annihilation during the freeze-out of the x species. This allows
us to assume T ∝ a−1 .

Given these assumptions the Boltzmann equation for the evolution of the x particle can be written as:

1 d(nx a3 )
= hσvi n2x − (neq 2
 
3 x ) (54)
a dt
Changing variable to Yx ≡ Tnx3 ∝ Nx , which is the number of particles in comoving volume, we end up
with the Riccati equation:

dYx λi 
= − 2 Yx2 − (Yeq )2

(55)
dx x

where
Mx dx
x≡ ; = Hx
T dt
(56)
Γ(Mx ) M 3 hσvi
λ= = x
H(Mx ) H(Mx )
Note that x is a convenient measure of time since the interesting moment is when T ∼ Mx .
The Riccati equation can be solved numerically. Solving this equation for the decoupling of the weakly
interacting particles, we obtain:
10−8 Gev
Ωx ∼ (57)
hσvi
Observe that the observed dark matter density today depends inversely on the annihilation cross-section.
This allows us to estimate the order of magnitude of the cross section, hσvi ∼ 10−8 Gev, which leads
to a temperature corresponding to the temperature of electro-weak transition. The fact that a thermal
relic with a cross-section characteristic of the weak interaction gives the right order of magnitude for the
observed dark matter abundance today is known as the WIMP miracle.

13
SUMMARY AND KEY POINTS
We have sketched the essential points of the thermal history of the Universe for which the interplay
between different components of the energy density has the major role. At early times, the rate of
particle interaction Γ is much larger than the expansion rate of the Universe H. Therefore,
all the particles are in equilibrium at a common temperature T . The energy density is dominated
by relativistic species.
To study non-equilibrium effects we use the Boltzmann equation. As long as the interaction
rate is much larger than the expansion rate, the particle abundance tracks its equilibrium value.
Once the interaction rate drops below the expansion rate, the particles drop out of thermal equi-
librium and decouple from the thermal bath. Integrating the Boltzmann equation allows us to
follow the non-equilibrium evolution of particle species: dark matter freeze-out, BBN and
recombination

14
3 The Inhomogeneous Universe: LSS
To understand the large scale structure (LSS) of the Universe we need inhomogeneities and we need to
track their evolution. When perturbations are small we can treat them as linear and use Newtonian
gravity. Then we need to go in full GR.
For most of the history of the Universe, the growth of structure was dictated by the evolution of dark
matter. Even though, we do not know yet the nature of dark matter, we only need two parameters, its
equation of state and a sound speed, to capture its long-wavelength effects.

Figure 3.1: Evolution of the primordial curvature perturbations. Figure taken from Cosmology by
Baumann (Fig. 6.2).

3.1 Structure growth in Newtonian theory: sub-horizon scales

We have a primordial spectrum of perturbations. How do we pass from this to what we actually observe?
For all the sub-horizon scales, we can use Newtonian theory and do a numerical treatment, since we
work with non relativistic fluids. We have the following equations:
Pertubation Theory Equations

∂ρ
1. Continuity Equation: ∂t + ∇(ρ~u) = 0

2. Euler Equation: d~
u
dt + (~u∇)~u = − ∇ρ
ρ − ∇φ

3. Poisson Equation: ∇2 φ = 4πGρ

These equations are meant for a static Universe. Adding expansion we can write the small perturbations
around the background solution as:
~r = a(t)~x
ρ = ρ̄(1 + δ)
~u = Ha~x + ~v (58)
P = P̄ + δP
φ = φ̄ + δφ
As long as δ  1 we can linearize the fluid equations. Moreover, assuming that each perturbation evolves

15
independently, we obtain the linear perturbation equations:
1
1. =⇒ δ = − ∇~v
a
1 1 (59)
2. =⇒ ~v + H~v = − ∇δP − ∇δφ
aρ̄ a
2 2
3. =⇒ ∇ δφ = 4πGa ρ̄δ

We can combine everything in one equation and obtain the time evolution of a linear density contrast:
 2 
• In coordinate space: δ̈ + 2H δ̇ − acs2 ∇2 + 4πGρ̄(t) δ = 0
 
k2
• In Fourier space: δ̈(k, t) + 2H δ̇(k, t) + c2s a2
− kJ2 δ(k, t) = 0

where kJ is the wave number corresponding to the Jeans scale:

p
4πGρ̄(t)
kJ ≡ (60)
cs (t)

We have two extreme regions:

• At small scales: ka  kJ . Here the pressure term dominates and we get the damped harmonic os-
cillator equation, the damping term being the expansion rate of the Universe H. We get oscillatory
solutions with a decreasing amplitude.

• At large scales: k
a  kJ , the pressure term is negligible and we get the following equation:

δ̈ + 2H δ̇ − 4πGρ̄δ = 0 (61)

As we said, this treatment applies to non-relativistic fluids. Two important examples of these are cold
dark matter and baryons after decoupling.

• For baryons the sound speed and the Jeans scale are strongly dependent on time. Before recom-
bination, baryons are strongly coupled to photons in a single photon-baryon fluid. The Jeans scale
is of order of the Hubble radius, so that there is no growth of baryonic fluctuations on sub-horizon
scales. After recombination, the sound speed of baryons suddenly drops and its fluctuations start
to grow.

• For cold dark matter the sound speed is negligible at all times, so that sub-horizon modes can
start to grow earlier. The evolution of the growth depends on the evolution of the background
density.

So we have that only fluctuations on scales larger than the Jeans scale can grow (due to the so called
Jeans instability). Solving the above equations in the three eras leads us to the following results where
we call density contrast the quantity δ ≡ δρρ̄ :

• RD: δ ∼ ln(a), which means suppressed growth (Meszaros effect)

• MD: δ ∼ a, which means linear growth

• DED: δ ∼ const, which means no growth at all.

16
3.2 Super-horizon scales
All of the results we just showed apply only to perturbations that are within the Hubble radius. To
describe super-horizon perturbations we need to apply the full relativistic perturbation theory. We get
that the super-horizon evolution of the density contrast is given by::
(
a2 RD
δ∼ (62)
a MD

We can observe that:

• Modes entering the horizon during MD era grow with same behavior (∝ a) both for sub and super
horizon

• In RD era, sub and super horizon modes behave very differently:

sub-horizon : δ ∝ log a
(63)
super-horizon : δ ∝ a2

Since the moment of horizon entering depend on the wavenumber of the mode k = aH, we have
a k−dependent growth of perturbations. This effect is described using the so called transfer
function T (k).
D+ (ti ) δ(k, t)
T (k) ≡ (64)
D+ (t) δ(k, ti )
where D+ is the scale-independent growth factor.

If denote by keq ≡ (aH)eq the wavenumber of the mode that entered the horizon at the time of matter-
radiation equality, this scale separates modes that entered the horizon in the radiation era (k > keq )
from those that entered in the matter domination era (k < keq ). We can write the following transfer
function: 

 1 k < keq
2 
T (k) ' (65)
 
keq keq
 ln k > keq
k k


3.3 Statistical properties

LSS in the Universe is not distributed randomly, but has interesting correlations between spatially
separated points. By definition, the mean value of density perturbations is zero hδi = 0. The first
statistical non trivial measure of the density field is, at some fixed time t, the 2-point correlation function:

ξ(|x − x0 |, t) ≡ hδ(x, t)δ(x0 , t)i (66)

In Fourier space:
hδ(k, t)δ ∗ (k 0 , t)i = (2π)3 δD (k − k 0 )P (k, t) (67)
where P (k, t) is the power spectrum. We have that ξ(r) is easy to observe, whereas the power spectrum
is easy to predict.
The linear evolution of Fourier modes we described earlier can be also written in terms of its power
spectrum:
D2 (t)
P (k, t) = T 2 (k) + P (k, ti ) (68)
D+ (ti )

17
The primordial power spectrum can also be written in the following form (Harrison-Zel’dovich spec-
trum):
P (k, ti ) = Ak n (69)
where A and n are constants. n is called the spectral index. Harrison and Zel’dovich, before inflation
was introduced, argued that the initial perturbations of our Universe were likely to have a power law
with spectral index n = 1. This value is special as it means scale invariance (k-independence).
From observations from the CMB, instead, we have:

n = 0.9667 ± 0.0040 (70)

This fits perfectly to what is predicted from inflation. We expect, in fact, fluctuations that are close
to scale invariance because the inflationary dynamics is approximately time-translation invariant. Since
inflation has to end, we must have a small time-dependence in the evolution. This translates into a small
deviation from a perfectly scale-invariant spectrum.

3.4 Matter power spectrum

Combining Eq. 65 and Eq. 69 we can predict the late-time linear matter power spectrum:

kn

 k < keq
P (k, t) ∝
 
k (71)
k n−4 ln k > keq
keq

Therefore, for a scale-invariant initial spectrum we expect the matter power spectrum to scale as k on
large scales and as k −3 on small scales.
Taking into account relativistic perturbation theory, the non-linear clustering, at a certain time η0 after
the MR equality, reads as:

k −3


 k  k0

 k k0 < k < keq
P (k, η0 ) =  2 (72)
 k
k −3 ln k > keq


keq

All sub-horizon modes grow as a and the amount by which the modes will have grown depends on when
they entered the horizon.

• For k > keq modes that entered the horizon in the radiation era all evolve in the same way
during the matter era, so the spectrum is uniformly boosted and no additional scale dependence
is generated for these scales.

• Modes with k < k0 are still outside the horizon and still today are unobservable

• On sub-horizon scales, for k0 < k < keq , we get, as expected, the same results of the Newtonian
approach

18
Figure 3.2: Measurements of the linear matter power spectrum. Figure taken from Cosmology by
Baumann (Fig. 5.2).

More about the 2-point galaxy correlation function

If we take the correlation function of galaxies, it tells us the probability of finding a galaxy at a
given distance. The 2-point galaxy correlation can be also written in physical space as:

DD(r)∆r
ξ(r) = −1 (73)
RR(r)∆r

where DD(r)∆r is the number of galaxy pairs within r ± ∆r/2 and RR(r)∆r is what one would
expect of galaxies were randomly distributed in space.
It assumes the following values:

> 0 positive correlation

= 0 uncorrelated (74)

< 0 anti correlation


Typically we measure the correlation function in redshift space since it is easier observationally.
Because of peculiar velocities ξ(z) 6= ξ(r). this is known as redshift space distorsion (RSD).
We find that:  γ
r
ξ(r) = (75)
r0
where r0 = 5.05 Mpc/h and γ ' 1.68. Galaxies are:

• strongly clustered on scales < 5 Mpc/h

• weakly clustered on scales > 10 Mpc/h

19
SUMMARY AND KEY POINTS
The growth of structure in Newtonian theory is valid for non-relativistic fluids (such as cold
dark matter and baryons after decoupling) and on sub-horizon scales. For relativistic fluids and
on super-horizon scales a full relativistic treatment is necessary.
In Newtonian theory we start from
∂ρ
1. Continuity Equation: ∂t + ∇(ρ~u) = 0

2. Euler Equation: d~
u
dt + (~u∇)~u = − ∇ρ
ρ − ∇φ

3. Poisson Equation: ∇2 φ = 4πGρ

we incorporate the expansion of the Universe, then we write all the quantities as fluctuations
around the homogeneous background solution and finally expand the above equations to linear
order in these fluctuations.
The linearized equations can be combined into a single equation for the evolution of the density
contrast:  2 
cs 2
δ̈ + 2H δ̇ − ∇ + 4πGρ̄(t) δ=0
a2
• On large scales pressure dominates and we get oscillatory solutions with a frequency set by
the speed of sound.

• On small scales the pressure is negligible and gravity dominates. The resulting growth of
instabilities is called the Jeans instability.

For dark matter instabilities, during RD, the expansion of space counteracts the gravitational
instability and the growth of the perturbations is only logarithmic. During MD, instead, the
perturbations start to grow linearly with the scale factor.
Observations measure the spatial correlations of the LSS. In linear theory, these correlations
are easiest to predict in Fourier space, since the different Fourier modes evolve independently. The
simplest correlation function in Fourier space is the power spectrum:

hδ(k, t)δ ∗ (k 0 , t)i = (2π)3 δD (k − k 0 )P (k, t)

For a Gaussian random field, the power spectrum contains all the statistical information of the
density field. The late-time power spectrum can be written as:
2 (t)
D+
P (k, t) = T 2 (k) P (k, ti )
D+ (ti )

A nontrivial transfer function arises because of the suppressed growth of sub-horizon modes during
the radiation era. The asymptotic scalings of the transfer function are:


 1 k < keq
2 
T (k) '
 
keq keq
 ln k > keq
k k


The primordial power spectrum is:

P (k, ti ) = Ak n

20
with n ∼ 1 for the Harrison-Zel’dovich spectrum.
The matter power spectrum at late times, instead, is predicted using the full relativistic theory:

k −3


 k  k0

 k k0 < k < keq
P (k, η0 ) =   2
 k
k −3 ln k > keq


keq

21
4 The Inhomogeneous Universe: CMB
4.1 The evolution of the photon-baryon fluid

on
-horiz n
er o
2 sup -horiz
a sub
(
∝ a
δ DM

δDM ∝ log a

matter/radiation recombination
equality CMB decoupling

Figure 4.1: Evolution of densities contrasts for a representative Fourier mode. Initially the fluctuations
of all components are of equal size and the baryons are tightly coupled to the photons. After decoupling
the baryons lose the pressure support of the photons and fall into the gravitational potential wells created
by dark matter. Note that around the time of decoupling the fluid approximation breaks down and the
Boltzmann equations have to be solved numerically for the evolution of each species. Figure taken from
Cosmology by Baumann (Fig. 6.8).

The analytic treatment we have seen so far, based on fluid approximation, breaks down around the time
of CMB decoupling, η∗ . The evolution equation of each species should be solved numerically (like with
CLASS code). In the figure above Fig. 4.1 we have the summary of the evolution of overdensities as a
function of conformal time. Note that it is crucial that the clustering of dark matter started before and
then aided the growth of baryon perturbations. Without this assisted growth baryons would not have
been able to cluster fast enough to explain the formation of galaxies.

Let us now study the evolution of photons and baryons before the decoupling. During matter-radiation
equality, the photon-baryon fluid is tightly coupled (tight coupling approximation) and they are

22
treated as a single photon-baryon fluid wuth velocity ~vb = ~vγ . What we obtain is the equation of a
forced harmonic oscillator:
00 R0 0 1 4 00 4R0 0
δγ + δγ − ∇2 δγ = ∇2 Ψ +4φ + φ (76)
|1 +
{zR} 3(1 + R) |3 {z } 1+R
pressure gravity

where R is the fractional contribution of baryons defined as:

(
3 ρ̄b 2 R  1 RD
R≡ ∝ Ωb h a(t) =⇒ (77)
4 ρ̄γ R ' 1 MD
This parameter is small at early times, but grows linearly with a(t) and becomes of order ∼ 1 around
the time of recombination.
Equation 76 is the fundamental equation describing the evolution of perturbations in the photon-baryon
fluid. From the pressure term we can read the sound speed of the photon-baryon fluid:
1
c2s ≡ (78)
3(1 + R)

4.2 Cosmic sound waves

For simplicity, if we ignore gravity terms in Eq.76 and consider the solutions to the homogeneous equation:

00 R0 0 1
δγ + δ − ∇2 δγ = 0 (79)
1 + R γ 3(1 + R)
we can highlight two timescales: the expansion time and the oscillation time. On small scales, for k  Hc
the oscillation time-scale is much smaller than the expansion rate. We can neglect gravity terms and
adopt the WKB approximation to expand R(t) and get the following solution:
cos(krs ) sin(krs )
δγ (η, k) = C(k) 1/4
+ D(k) (80)
(1 + r) (1 + R)1/4
where Z η
0
rs (η) ≡ cs (η )dη 0 (81)
0
is the sound horizon at time η.
• During CMB, at the time of photon decoupling η∗ :
rs (η∗ ) ' 145 Mpc (82)

• During BAO:
rs (ηdrag ) ∼ 147 Mpc (83)

The C(k) and D(k) functions are fixed by the super-horizon initial conditions.
At late times (high k values) oscillations get damped (see Fig. 4.2) for 2 reasons:
1. Diffusion (silk) damping: photon diffusion from hot (high density) regions to cold (low density)
regions erase temperature differences on small scales (the typical scales are at ksilk ∼ 7 Mpc and
the corresponding multipole is ls ∼ 1300)
2. Landau damping: recombination is not instantaneous; fluctuations that are smaller than the
last scattering surface are somehow averaged, so that their amplitude is reduced

23
Figure 4.2: Evolution of density of photons (gray) and the velocity of baryons (black ) for k = 1 hMpc−1 .
Figure taken from Cosmology by Baumann (Fig. 6.9).

4.3 From primordial sound waves to CMB anisotropies

Fluctuations of different wavelength are captured (at decoupling): when we are looking the CMB we
are seeing a snapshot of the primordial sound waves at the moment of photon decoupling. We see
from Eq. 80 that the oscillation frequency of the photon-baryon system depends on the wavenumber
k, meaning that different modes decouple at different moments in their evolution and therefore have
different amplitudes at last scattering (see Fig. 4.3).

Figure 4.3: Illustration of the origin of peaks in the CMB power spectrum. On the left we see how
different wavelengths are captured at different moments in their evolution and therefore have different
amplitudes at decoupling. Since the square of the amplitude determines the power at a given length
scale, waves that are captured at an extremum (A or C) produce the peaks in the CMB spectrum, while
waves that are captured with zero amplitude (B) produce the troughs. Figure taken from Cosmology
by Baumann (Fig. 6.10).

24
Evaluating the solution at decoupling, η = η∗ , gives the k−dependent amplitude of the fluctuations at
last scattering. The k-dependent amplitude of δγ (η) at η = η∗ is fixed by:
nπ
kn = (84)
rs (η∗ )
which gives us the position of the peaks. These are the same peaks we observe in the matter power
spectrum. However, there is no one-to-one correspondence between k and r spaces.
The power spectrum measures |δγ |2 so that waves captured in their extremes correspond to peaks in the
spectrum (see A ad C in Fig. 4.3). Whereas waves with zero amplitude correspond to troughs (see B in
Fig. 4.3). We have that:
• The first peak corresponds to the sound horizon

• Odd peaks correspond to compressions in the plasma

• Even peaks correspond to rarefactions in the plasma

Note that the odd peak is always higher than the subsequent even one due to baryons falling in the
compressed regions of the plasma (following the DM potential wells).
The positions of the peaks are determined by a combination of the sound horizon at decoupling and the
angular distance to the surface of last scattering. The heights of the peaks carry information about the
amount of dark matter and baryons in the Universe. The size of the damping on small scales depends
on the density of neutrinos. This is how the measurements of the CMB anisotropies allow precise
measurements of the key cosmological parameters.

4.3.1 Baryonic Acoustic Oscillations (BAO)

The same oscillations that are observed in the CMB anisotropies are imprinted in the distribution of
baryonic matter. They leave an imprint in the clustering of galaxies. This happens because photons and
baryons oscillate together before decoupling. Like photons, the different Fourier modes of the baryons
decouple at different phases in their evolution. The initial conditions for the gravitational growth of
the baryons include the oscillations of the photon-baryon fluid. This oscillatory feature gets transferred
to the gravitational potential and the matter fluctuations, but the effect is smaller than in the CMB.
Since correlations in the galaxy distribution are inherited from correlations in the matter density, these
baryonic acoustic oscillations (BAO) are imprinted in the galaxy power spectrum.
The moment when the baryons are released from the drag of the photons is known as the drag epoch.
Although photon and baryon decoupling are related they are not the same. Baryons decouple slightly
later than photons at zdrag ∼ 1020 (compared to z∗ ∼ 1090). This is because there are much more
photons than baryons in the Universe, so that photons stop feeling the effect of the baryons before
the baryons stop experiencing the influence of the photons. After the drag epoch, baryons behave as
pressureless matter.
We measure BAO both in galaxy clustering and in the L-α forest.

4.4 The CMB anisotropies

What makes the CMB such a powerful cosmological probe is the fact that the fluctuations were captured
when they were still small and therefore they are accurately described by linear perturbation theory.
It is useful to separate the CMB spectrum (see Fig. 4.4) into three different regions:
• The large-angle correlations are sourced by fluctuations that were still super-horizon at recom-
bination (therefore they did not evolve before CMB decoupling and are a direct probe of inflation).

25
 • Perturbations with shorter wavelengths entered the horizon before recombination. Inside the hori-
zon, the perturbations in the tightly coupled photon-baryon fluid propagate as sound waves
supported by the large photon pressure. The oscillation frequency of these waves is a function of
their wavelength and so different modes are captured at different moments in their evolution when
the CMB was released at photon decoupling. This is the origin of the oscillatory pattern seen in
the angular power spectrum.

• On scales smaller than the mean free pat of photons, the random diffusion of the photons can erase
the density contrast in the plasma leading to a damping of the wave amplitudes. This suppresses
the amplitude of the CMB power spectrum on small angular scales.

large-angle sound diffusion

correlations waves damping

Figure 4.4: Planck measurements of the angular power spectrum of CMB temperature fluctuations.
Figure taken from Cosmology by Baumann (Fig. 6.11).

4.5 CMB angular power spectrum

The largest anisotropy in the CMB is a temperature dipole due to the motion of the Solar System with
respect to the rest frame of the CMB. Pay particular attention that the underlying assumption is
that we remove the dipole mode due to kinematics, but we are not sure that it is all due to kinematics.

26
 δT (n̂)
p~
δT (n̂)
θ T = n̂ · ~v = v cos θ
~v
n̂
moving
observer Surface of last scattering

Here n̂ is the unit vector pointing at the observer’s line of sight (which is opposite to the momen-
tum p~ of the incoming radiation). ~v is the observer’s velocity, i.e. the orbital velocity of the solar system
plus the galaxy velocity with respect to the center of mass of the local group plus the velocity of the
local group with respect to the CMB rest frame.
The problem is the discrepancy between the CMB estimation of the kinematic dipole and that from
quasars (∼ 5σ tension) and from SnIa (∼ 2 − 3σ tension).

We can start from the primordial power spectrum (the same one of the matter distribution) and then
deduce the one at later times. What we obtain is:
δT (n̂)
T (n̂) ≡ T̄0 (1 + θ(n̂) with θ(n̂) = (85)
T̄0
In this case the 2-point correlation function depends on the angle θ:
l
X X
θ(n̂) = alm Ylm (n̂) (86)
l=2 m=−l

with alm the multiple moments and Ylm the spherical harmonics. Averaging over the whole Universe:

C(θ) = hθ(n̂)θ∗ (n̂0 )i (87)

Exploiting spherical harmonics properties we end up with:

Z +1
Cl = 2π d cos(θ)C(θ)Pl (cos θ) (88)
−1

with Pl being the Legendre polinomiums.We also find a nice relation between T0 and Cl :

l(l + 1)
∆2T ≡ Cl T02 (89)
2π
Typically this is what is plotted as the CMB spectrum. For a fixed l, we have (2l + 1) different alm s, i.e.
(2l + 1) independent estimates of the true Cl s. Assuming we have extracted all the multipole moments
alm s we can estimate the true Cl s as:
q
hCl − Cl2 i
r
1 X
2 ∆Cl 2
Ĉl ≡ |alm | and = = (90)
2l + 1 m Cl Cl 2l + l

27
 ∆Cl
where Cl is the intrinsic error on the Cl due to the limited data sample and it is called cosmic variance.

What happens to P (k, t0 )? It evolves and the evolution tells us all the properties of the photon-baryon
fluid:  
δγ  
Ψ free streaming
=⇒ =⇒ δT (n̂)today (91)
projection effects
vb η
∗,REC

Let us now write the evolution equation:

  Z η0
δT δγ 0 0
(n̂) = +Ψ − (n̂vb )η∗ + dη(φ + Ψ ) (92)
T 4 η∗ η
| ∗
| {z }
| {z } Doppler {z }
Sachs-Wolf integrated Sachs-Wolf

• Sachs-Wolf effect: happens at the surface of last scattering. The δγ /4 term tells us the intrinsic
photon density fluctuation. The Ψ term is related to the additional temperature perturbation due
to gravitational redshift of photons (climbing out of the potential well inside the last scattering
surface). The cause are the inhomogeneites in the density field at recombination.
• Doppler effect: due to the continuous scattering between free electrons and photons on the last
scattering surface. This shift leads to an additional temperature fluctuation.
• Integrated Sachs-Wolf effect: this is an additional gravitational redshift due to the fact that
gravitational wells are evolving with time in the DE-dominated era, as well as when radiation was
still non negligible.
During RD the early integrated Sachs-Wolf effect increases the height of the first peak of the spectrum.
During DED, the late integrated Sachs-Wolf effect, instead, has an impact on the large scales leading to
more power at low ls. The Doppler effect reduces the contrast between the valleys and the throats.

4.6 Impact of cosmological parameters on the CMB

In the ΛCDM model we have 6 parameters:
{As , ns , ωb , ωM , ΩΛ , τ } (93)
where the first two parameters represent the amplitude and the slope of the primordial power spectrum
and τ is the optical depth at reionization (a late Universe parameter that can be substituted with zREI ).
Remember then that the ΛCDM model is flat by definition.
Instead of using ΩΛ as a free parameter, we can use h:
r
ωM H0
h= =⇒ h = km
∼ 0.7 (94)
1 − ΩΛ 100 sM pc

Or better we could use the angular size of the sound horizon at recombination θs :
R zCMB
∞ dz √ cs (z)
rs (zCMB ) ρTOT (z)
θs = = R zCMB (95)
dA (zCMB ) 0
√ dz
ρTOT (z)

where the numerator regards pre-recombination physics (ωb , ωM ), while the denominator the post-
recombination one (Ωk , ΩΛ , ΩM , H0 ).
We better use as a free parameter θs because it is very well measured by Planck.
What do we measure with BAO and SnIa?

28
 • With SnIa we measure H(z):
p p
H(z) ∝ ρTOT (z) and H0 ∝ ρTOT (z = 0) (96)

We have that:
s
z
cdz 0
Z
8πG X
dL = (1 + z) −−−→ czH0−1 ' vH0−1 =⇒ H(z) = ρi (z) (97)
0 H(z 0 ) z1 3
i

• With BAO we can measure both the perpendicular and the parallel components of the damping
scale θd :
rs (zDRAG ) k
θd⊥ = ; θd = rs (zDRAG )H(z) (98)
dA (z)
We are measuring the Hubble flow.
Why is it so difficult to solve the tension? We have that:
• At low z (post recombination) we have h ∼ 0.73

• At higher z (pre-recombination) we have h ∼ 0.67

These two values have a 5σ tension. The measure of θs we are getting from CMB is model-independent
and we cannot change it. Moreover, joining CMB+SnIa+BAO everything is over constrained and there
is no way to solve the tension. We could do it if there was a model that predict a reduction of 7 − 10%
of the horizon size. We could play a bit with the radiation content of the early Universe going beyod
ΛCDM: " #
7 4 4/3
 
ρr = 1 + Neff ργ (99)
8 11
Usually Neff = 3 to take into account all the relativistic species at recombination. It could be interesting
to see what value of Neff could help us to obtain an H0 value which matches early and late Universe
measurements. However, Neff 6= 3 is not supported by Planck, which means that we should find a way
to modify rs without modifying Neff (for example, there are ideas for early dark energy). In fact if
Neff > Neff,standard we have an impact on the ratio between the damping scale θD and the horizon scale
θS :
θD 1/2
∝ H∗ (100)
θS
where H∗ is the Hubble parameter at recombination.

We want to understand how the cosmological parameters determine the shape of the power spectrum,
so that we can understand how the CMB measurements constrain these parameters. Therefore, let us
summarize all of the effects:

1. peak location ↔ θS
2. odd/even peak amplitude ↔ ω
3. overall peak amplitude ↔ ωM (shifts of MR equality)
rD∗ ↔ωM ,ωb
4. damping scale θD = dA∗ ↔ΩΛ ,ωM

5. global amplitude ↔ AS
6. global tilt ↔ nS

29
 7. amplitude for l < 10 ↔ τ

8. tilt for l < 20 ↔ ΩΛ (late ISW effect)

As a side note, there is another parameter that has a tension which is σ8 which tells how matter is
clusterized and apparently we obtain a σ8 values which predicts that matter should more clusterized
than what we actually observe.

SUMMARY AND KEY POINTS

Before decoupling, photons and baryons can be treated as a single tightly coupled photon-baryon
fluid. The density contrast of the photon-baryon fluid evolves as:

00 R0 0 1 4 00 4R0 0
δγ + δγ − ∇2 δγ = ∇2 Ψ + 4φ + φ
1+R 3(1 + R) 3 1+R

where this is the equation of a harmonic oscillator with a gravitational driving force. This
oscillator equation plays an important role in the physics of the CMB anisotropies.
We can study the physics of the CMB anisotropies. The important equation is the following:
  Z η0
δT δγ 0 0
(n̂) = +Ψ − (n̂vb )η∗ + dη(φ + Ψ )
T 4 η∗ η
| ∗
| {z }
| {z } Doppler {z }
Sachs-Wolf integrated Sachs-Wolf

which relates the observed temperature fluctuations in a particular direction and the corresponding
fluctuations on the surface of last scattering.
The temperature anisotropies can be written as a sum over spherical harmonics:
l
δT (n̂) X X
= alm Ylm (n̂)
T̄0
l=2 m=−l

At last, we have the relation between the parameters of the ΛCDM model and the shape of the
CMB power spectrum. Three length scales are imprinted in the spectrum:

• The sound horizon at last scattering, on which the peak locations depend

• The damping of the spectrum which is determined by the diffusion scale

• The distance to last scattering on which the angular scales depend

The sound horizon and the diffusion scales are fixed by pre-recombination physics and hence in
the ΛCDM model depend only on ωb and ωm . These parameters also determine the peak heights.
The angular diameter distance to the last scattering is sensitive to the geometry (Ωk , H0 ) and the
energy content (Ωm , ΩΛ ) of the late Universe.

30