The Dark Ages of the

Universe
Simon Glover
 Lecture 1:
Cosmological preliminaries
The standard model of cosmology
• The current standard model of cosmology is based on
two fundamental assumptions:

- On large scales, the Universe is isotropic

- Our location is not special (the Copernican principle)

• Together, these assumptions imply that on large scales,

the Universe is homogeneous

• We usually supplement these with the further assumption

that general relativity (GR) is the correct theory of gravity
Credit: NASA
Credit: NASA
 c. 1800 Mpc

Depth of slice c. 150 Mpc; colours

Credit: D. Eisenstein/SDSS-III collaboration
correspond to depth in slice
 Gonçalves et al (2018, MNRAS, 481, 5270)
D2 is the fractal dimension;
for homogeneity, expect D2 = 3
• Key lesson from GR: space and time are not distinct

• We live in four-dimensional spacetime

• Geometry of this spacetime is described by a metric

tensor gμν

• Locally, can choose coordinates so that spacetime is flat

• Globally, this is not a priori possible. However,

homogeneity and isotropy imply a simple form for the
metric tensor:
• This metric tensor is known as the Friedmann-Robertson-
Walker (FRW) metric

• (r, φ, θ) are our usual spherical polar coordinates

• a(t) is a scale factor that depends on time but not location

• fK describes the curvature of the Universe:

• Suitable rescaling of our radial coordinate allows us to
write this in a simpler form:

• Homogeneity + isotropy allow only three possible global

curvatures: positive (K > 0), flat (K = 0), or negative (K < 0)
Credit: Wikipedia
• Homogeneity implies that the scale factor a must be
independent of location, but does not prevent it
depending on time

• a(t) increasing with time corresponds to expansion

• a(t) decreasing with time corresponds to contraction

• Observations show that we live in an expanding Universe

• In an expanding Universe, light coming to us from distant
sources is redshifted, i.e. shifted to lower frequencies

• Consider a source comoving with the expansion that

emits a beam of light at time te that is observed at time to

• For simplicity, assume the observer is also comoving and

is located at the origin of the coordinate system

• In this case, the FRW metric reduces to:

• We can write the coordinate distance between the source
and the observer as:

• Taking the time derivative and rearranging yields:

• This in turn implies that:

• A time interval of dte in the source frame is multiplied by a
factor of a0/ae in the observer’s frame

• Expansion implies a0 > ae, so we get time dilation

• If we apply this to the period of our EM wave, then we see

that this must lengthen

• Therefore, the frequency of light must decrease: the light

is redshifted by a factor a0/ae = 1 + z
• So far, we have not used any results from GR

• The FRW metric follows from our assumptions of

homogeneity and isotropy and holds for any metric theory
of gravity

• Redshift is a direct consequence of the FRW metric

• So where does GR come in? We need it to derive the

equations describing how the scale factor evolves with t
• Here, ! is the energy density of matter & radiation, and p
is its corresponding pressure

• Λ is a quantity known as the cosmological constant

• These equations are known as the Friedmann equations.

• For non-relativistic matter (“dust”), p is negligible
compared to the energy density and hence:

• For relativistic matter & radiation (usually just referred to

as “radiation”), we have p = !c2 / 3 and:

• Here !0 and !r,0 are the present-day energy densities of

dust and radiation, and we have chosen units so that
a0=1
• Let us now introduce several important dimensionless
parameters:

Hubble parameter

Critical density

Dimensionless density parameters

• We can rewrite the first of the Friedmann equations in
terms of these parameters as:

• Here and above, H0 is the value of the Hubble parameter

at z=0 (also referred to as the “Hubble constant”)

• From this, we see that the term in brackets must equal 1

at z=0 (i.e. when a=1)

• Sign of ΩK (and hence sign of K) depends on the sum of

the other three dimensionless density parameters

• Implies that the curvature of the Universe depends on the

value of these parameters!
Carroll, Press & Turner (1992, ARA&A, 30, 499)
• Density parameters not only determine curvature but also
fate of Universe: eternal expansion or eventual collapse?

• Observations suggest we live in a spatially flat Universe

with ΩΛ ≈ 0.7, Ωm ≈ 0.3, Ωr currently negligible

• Friedmann equations show that relative importance of

these terms changes with redshift

• In this course, we are mostly interested in the Universe

between z = 6 and z = 1000, where Ωm dominates
Thermal evolution of the Universe
• How does the temperature of the Universe change as it
expands?

• The answer depends on whether we’re considering

radiation or non-relativistic matter

• For non-relativistic matter, expansion implies that particle

velocities evolve in the absence of collisions as

• Now consider what this does to a Maxwell-Boltzmann

velocity distribution:
• If the velocity changes from v to v/a, then v2 changes to
v2/a2 and d3v changes to d3v/a3

• If the temperature changes from T to T/a2 for the same

change of scale factor, the velocity distribution retains the
same form, i.e. it remains a thermal distribution

• Therefore, for non-relativistic matter, the temperature

scales with redshift as:
• We can derive a similar result for non-relativistic gas by
starting with the ideal gas law (see notes)

• We can also use a similar approach to show that for

radiation, an initial black-body spectrum retains its shape
but changes its temperature by a factor:

• Note that this scaling also follows in a very simple fashion

from the Stefan-Boltzmann law. For a black-body radiation
∝ ∝
field, urad T4 and in an expanding Universe urad (1+z)4

• ∝ ∝
Therefore, T4 (1+z)4 and hence T (1+z)
• Considered in isolation, therefore, the gas temperature
evolves more rapidly with time than the radiation
temperature

• However, in reality, we cannot consider the two

components in isolation, as they can interact

• For our present purposes, the most important interaction

is Compton scattering
 Credit: Wikipedia

• Compton scattering: inelastic scattering of photons by

electrons

• If Trad > Tgas, then on average energy flows from

photons to electrons
• Energy transfer rate per unit volume depends on radiation
energy density, electron number density

• This yields a rate per unit volume scaling as (1+z)7 !

• Rate per particle scales as (1+z)4 (i.e. proportional to

radiation energy density)

• Compton scattering is therefore very effective at

transferring energy from radiation to matter at high z, but
becomes much less effective at low z

• In practice, Tgas tracks Trad very closely for z > 200, but
temperatures become decoupled at lower z
Pfenniger & Puy (2003, A&A, 398, 447)
• In this course, we mostly care about the Universe at
redshifts z < 2000, where T is relatively low

• However, one bit of physics occurring in the early, hot

Universe is important for us: primordial nucleosynthesis

• In the early Universe, protons and neutrons first form

once the temperature drops to kT ~ 1 GeV

• Initially, the proton-neutron ratio is kept in thermal

equilibrium by efficient weak interactions

• Once kT ~ 800 keV, these reactions “freeze out”, fixing the

ratio at:
• At later times, the neutron/proton ratio steadily decreases
owing to beta decay

• When the temperature drops to kT ~ 80 keV, formation of first

heavy nuclei becomes possible — prior to this, any D or 3He
formed is immediately photodisintegrated

• At this point, most available neutrons consumed by nuclear

reactions leading to production of 4He. Trace amounts remain
in D and 3He or are used to form Li

• Formation of elements heavier than Li blocked by lack of

stable nuclei with weights A = 5, A = 8: process gets “stuck”
at 4He

• In stars, can get past this point via triple-alpha process, but
cosmic density too small, timescale too short to allow this
• Outcome is gas which is ~75% hydrogen and ~25%
helium by weight, with small D and 3He fractions and a
tiny Li fraction

• Precise numbers depend on temperature at which it first

becomes possible for D to survive photodisintegration by
the radiation component

• This depends on the baryon-to-photon ratio and hence

on the baryon density parameter Ωb

• For currently favoured ΛCDM model, we get a helium

mass fraction Y = 0.247, a D/H ratio of 2.6 ⨉10-5 and a
lithium fraction of 4 x 10-10
Coc et al (2004, ApJ, 600, 544)
Summary
• We live in an expanding Universe that is homogeneous
and isotropic on large scales and described by the FRW
metric

• The Universe appears to be spatially flat and its energy

density at the present time is dominated by a
cosmological constant

• At the redshifts of interest for this course, the energy

density was dominated by the rest-mass energy of non-
relativistic matter

• The Universe was much hotter in the past than it is now

• Consequences of early high temperatures:

- Primordial nucleosynthesis: Universe consists of a mix

of H, He (mass fractions ~75%, 25%), with small traces
of D and Li

- At z >> 1000, the matter content is highly ionized

• Topic for next time: how does the Universe recombine

from this initially highly ionized state?