M.

Pettini: Introduction to Cosmology — Lecture 8

PRIMORDIAL NUCLEOSYNTHESIS

Our discussion at the end of the previous lecture concentrated on the rela-
tivistic components of the Universe, photons and leptons. Baryons (which
are non-relativistic) did not figure because they make a trifling contribu-
tion to the energy density. However, from t ∼ 1 s a number of nuclear
reactions involving baryons took place. The end result of these reactions
was to lock up most of the free neutrons into 4 He nuclei and to create trace
amounts of D, 3 He, 7 Li and 7 Be. This is Primordial Nucleosynthesis, also
referred to as Big Bang Nucleosynthesis (BBN).

8.1 The Neutron-to-Proton Ratio

Before neutrino decoupling at around 1 MeV, neutrons and protons are kept
in mutual thermal equilibrium through charged-current weak interactions:
n + e+ ←→ p + ν̄e
p + e− ←→ n + νe (8.1)
n ←→ p + e− + ν̄e
While equilibrium persists, the relative number densities of neutrons and
protons are given by a Boltzmann factor based on their mass difference:
∆m c2

n 1.5
= exp − = exp − (8.2)
p eq kT T10
where
∆m = (mn − mp ) = 1.29 MeV = 1.5 × 1010 K (8.3)
and T10 is the temperature in units of 1010 K.

As discussed in Lecture 7.1.2, this equilibrium will be maintained so long as

the timescale for the weak interactions is short compared with the timescale
of the cosmic expansion (which increases the mean distance between parti-
cles). In 7.2.1 we saw that the ratio of the two rates varies approximately
as: 3
Γ T
' . (8.4)
H 1.6 × 1010 K
1
This steep dependence on temperature can be appreciated as follows. We
have already encountered many times the expression for the Hubble pa-
rameter as a function of redshift:

H(z)
q
= Ωm,0 (1 + z)3 + Ωrad,0 (1 + z)4 + Ωk,0 (1 + z)2 + ΩΛ,0 (8.5)
H0
from which we see that in the radiation dominated era the expansion rate
is proportional to the square root of the energy density in radiation, which
in turn is proportional to T 4 (eq. 7.28). Thus, H ∝ T 2 .

On the other hand, the rate per neutron of the reactions at eq. 8.1 is
proportional to (i) the number of electron neutrinos, n(νe , ν̄e ), and (ii) the
weak interaction cross-section hσi. In turn, hσi is proportional to (i) T 2 ,
−1
and (ii) the reciprocal of the neutron half-life for free decay, τ1/2 , which
measures the intrinsic strength of the interaction. Since n(νe , ν̄e ) ∝ T 3 ,
Γ ∝ T 5 , and hence Γ/H ∝ T 3 .

As T decreases, there comes a point at

kTd ≈ 0.8MeV = (mn − mp ) − me

where the weak interaction rate falls rather suddenly below the expansion
rate and the ratio n/p is frozen (apart from free decay and some residual
weak interactions). Thus, at t ' 2.6 s, Td ' 0.8 MeV (the neutron freeze-
out), the ratio n/p is:
∆m c2

n 1.3
= exp − = exp − = 0.20 = 1 : 5 (8.6)
p kTd 0.8

The precise value of the decoupling temperature Td depends on the two

physical constants Nν and τ1/2 :
1/2
11 7
Td3 ∝ τ1/2 + Nν ; (8.7)
4 8
implying that larger values of either of these constants lead to higher Td ,
higher n/p, and therefore a higher primordial abundance of 4 He, given
that essentially all of the available neutrons are later incorporated into
4
He nuclei, as we shall see in a moment.

2
Figure 8.1: Dependence of the primordial abundance of 4 He by mass, Yp , on the value
of the neutron mean life, τn (s), from the error propagation calculations by Cyburt et al.
(2015).

8.2 Deuterium Formation

After neutron freeze-out the only reaction that appreciably changes the
number of neutrons is neutron decay:

n −→ p + e− + ν̄e

with a mean life τn = 880.3 ± 1.1 .1

Without further reactions to preserve neutrons within stable nuclei, the

Universe today would be essentially pure hydrogen. The reaction that
preserves neutrons is deuteron formation (deuteron is the nucleus of deu-
terium, the simplest isotope of hydrogen):

p + n ←→ d + γ

This reaction is exothermic, with an energy difference of 2.225 MeV (the

binding energy of the D nucleus, ED ). Since the strong interaction is
at play here, the formation of D proceeds very efficiently. However, the
temperature at t ∼ a few seconds is not much lower than 2.2 MeV and,
given that photons are a billion times more numerous than baryons (see
eq. 7.6), there are sufficient photons with energies Eγ > ED in the Wien
tail of the blackbody distribution to instantly destroy newly formed D
1
In general, for radioactive decays N (t) = N0 e−λt , where the decay constant λ ≡ 1/τ , and the half-life
t1/2 = τ ln 2.

3
by photo-dissociation. We thus have to wait until the temperature has
fallen sufficiently for a substantial concentration of D to build up before
primordial nucleosynthesis can get going in earnest. This is sometimes
referred to as the Deuterium bottleneck.

Consideration of the relevant balance equations shows that the formation

rate of D exceeds its photo-dissociation rate at TD ' 8 × 108 K. This is
attained at time t ∼ 300 s, at which point the ratio n/p has fallen from the
value at eq. 8.6 by a factor exp[−300/(880.3)] = 0.71 to n/p = 0.14 or 1:7.

This ratio immediately gives us an approximate estimate of the primordial

helium mass fraction, Yp , because virtually all neutrons surviving after
300 s are later incorporated into 4 He. Thus:
n
4 n2 2n 2p
Yp = = = = 0.25 (8.8)
p + n p + n 1 + np

which corresponds to a ratio 4 He/H = 1/12 by number (since a 4 He nucleus

weighs four times as much as H).

It is a coincidence that the neutron mean life is so comparable to the

time it takes for nuclei to form; if it were much shorter, all the neutrons
would have decayed before primordial nucleosynthesis had had a chance to
proceed and only hydrogen would remain.

8.3 Nuclear Reactions

Once the d/p ratio has built up to ∼ 10−5 , further reactions proceed to
synthesize helium and other light nuclei. The most important reactions
are collected in Figure 8.2.

As already mentioned, 4 He soaks up virtually all of the neutrons available.

After the formation of 4 He, traces amounts D and 3 He survive because
nuclear reactions are frozen out by low density and temperature before
their destruction is complete. Still smaller traces of 7 Li and 7 Be (which
later decays to 7 Li) are formed and survive. The outcome of primordial
nucleosynthesis is calculated by numerically following the series of relevant
nuclear reactions shown in Figure 8.2; Figure 8.3 shows the progress of
BBN with time.

4
Figure 8.2: The network of reactions for Big Bang nucleosynthesis (from Nollett & Burles
2000). Key: p(n, γ)d ≡ p + n → d + γ.

BBN stops at 7 Li because no stable nucleus of mass number 5 or 8 exists

and thus no new nuclei can be formed in collisions of two He nuclei or a
proton with a He nucleus. Collisions between three nuclei are far too rare
to contribute. The important point here that how far the BBN nuclear
reactions in Figure 8.2 proceed before they are frozen out by low density
and temperature depends on the parameter η ≡ nb /nγ which we have
already encountered in Lecture 7. As explained there, η is related to the

Figure 8.3: Fractional abundances of the light elements produced in BBN as a function
of time and temperature. Note that the minus sign in the exponent of the units on the
y-axis appears to have been lost in the reproduction of this figure.

5
Figure 8.4: The abundances of light elements relative to H predicted by BBN calculations
are shown as functions of η ≡ nb /nγ (bottom x-axis) and Ωb,0 h2 (top x-axis). It is
4
customary
FIG. 1.toPrimordial
express the primordial
abundances of theabundance
light nuclidesofas He by mass
a function (Yp ), baryon
of cosmic and those of the
content, as other
elements by number. The widths of the curves reflect the uncertainties in the predictions,
predicted by SBBN (“Schramm plot”). These results assume Nν = 3 and the current measurement
propagated from uncertainties in the experimental values of the relevant nuclear reaction
rates. of(Figure
the neutron lifetime τn =from
reproduced 880.3Cyburt
± 1.1 s. Curve
et al.widths
2015). show 1 − σ errors.

"−1.60# $0.395! "0.95#

10
!
D 0.41
−5 10 η Nν GN τn
$
cosmic densityHof baryons
= 2.579×10 in
6.10
units of
3.0
the Ωb,0 h2 via the
critical
GN,0
density,
880.3s
temperature of the cosmic
× [p(n, γ)d]−0.19microwave
[d(d, n)3 He] background
−0.53
[d(d, p)t]−0.47 today, Tγ,0 :
× [d(p, γ)3 He]
−0.31 3
[ He(n, p)t]
0.023
[3 He(d, p)4 He]
3
−0.012
(14)
2.7255 K
η10 = 273.3 Ωb,0 h2 (8.9)
3
He
!
10
10 η
"−0.59#
Nν
$0.14!
GN
T"0.34#
γ,0 τn
$0.15
= 9.996×10−6
H 6.10 3.0
880.3s GN,0
where η10 is 1010 η. Figure 0.088
8.4 3illustrates
dependence
0.21
on η and Ωb,0 h2
the
× [p(n, γ)d] [d(d, n) He] [d(d, p)t]−0.27
of the relative abundances
0.38
of the light
−0.17
elements−0.76
produced −0.009
in BBN.
× [d(p, γ)3 He] [3 He(n, p)t] [3 He(d, p)4He] [t(d, n)4 He] (15)
It is worthwhile considering the
"2.11#following
! aspects
"−0.73# of Figure 8.4:
7
1010 η
! −0.284
Li
$
Nν
$
GN τn 0.43
= 4.648×10−10
H 6.10 3.0 GN,0 880.3s
• Yp has a very minor dependence on η because, as we have explained,
essentially all of the neutrons20end up in 4 He. For larger baryon-to-
photon ratios, the balance between D formation and photo-dissociation
moves to slightly larger values of TD (i.e. the deuterium bottleneck
does not last as long; see section 8.2). The earlier D can form, the
fewer neutrons have decayed, and hence Yp is slightly larger.
• In a similar vein, the higher η, the more efficient is the conversion of D
into 4 He which proceeds via two-body reactions with p, n, D, and 3 He

6
(see Figure 8.2). This leads to the steep inverse dependence of (D/H)p
on η in Figure 8.4. 3 He declines more gently because this nucleus is
more robust.
• 7 Li has a bimodal behaviour because it can be produced via two chan-
nels, labelled 10 and 7 in Figure 8.2. Reaction 7 is favoured at low
baryon densities, while 10 takes over at high values of η.

8.4 Measures of Primordial Abundances of the Light

Elements

The remarkable implication of Figure 8.4 is that, by measuring the primor-

dial abundance of one of the light elements created in BBN, it is possible
to literally ‘weigh’ the Universe, at least in baryons. Measurement of two
or more elements would in addition allow us to test the theory, since the
values of η indicated by any two elements should agree. In this section, we
shall consider the experimental verification of BBN predictions.

Before we start, we should make clear a few key facts. The relative abun-
dances of the light elements created in BBN remained unaltered for the
first ∼ 200 Myr.2 However, once the first stars formed at z ∼ 20, the
BBN abundances began to be changed by the process of galactic chemical
evolution. All of the elements of the periodic table heavier than Boron
were synthesised in the interiors of stars and some fraction of these was
returned to the interstellar medium to be incorporated in successive stellar
generations. Thus, as time progressed, the ‘metal’ content of the Universe
steadily increased (astronomers use the terms ‘metals’ and ‘metallicity’ to
refer to all elements with atomic number 6 or greater—that is C and heavier
elements—that were created exclusively by stellar nucleosynthesis).

Since we cannot view the Universe directly at the early times when BBN
took place or soon after, the determination of the primordial abundances
of the light elements is based on identifying astrophysical environments—
whether galaxies, gas clouds or stars—that have undergone minimum en-
richment by stellar nucleosynthesis. Determination of 4 He/H, D/H, 7 Li/H
in such ‘low metallicity’ environments can then lead to these elements’ pri-
2
With the exception of 7 Be which swiftly decays to 7 Li, but only once the Universe has cooled suffi-
ciently for the 7 Be nucleon to capture an electron.

7
Figure 8.5: Left: Near-IR image of the central part of the Orion nebula, obtained with
the Very Large Telescope of the European Southern Observatory. The nebula is diffuse
gas ionised by hot stars. Right: Typical emission line spectrum of an H ii region such as
the Orion Nebula. The most important spectral features are labelled.

mordial abundances, sometimes with appropriate corrections/extrapolations

to zero metallicity. Note that we have not included 3 He in the above list.
This isotope has so far been detected only in the the Milky Way interstellar
medium, which is far from being a pristine environment, and correcting for
its post-BBN history is too uncertain for 3 He to be a useful probe of BBN.

We now discuss the most recent determinations of Yp , (D/H)p , and (7 Li/H)p .

For practical reasons, each of these ratios is most easily accessed in a dif-
ferent astrophysical environment.

4
8.4.1 Helium

The abundance of 4 He has been measured for decades from the emission line
spectra of H ii regions, volumes of the interstellar medium ionised by the
ultraviolet radiation of massive, short-lived stars. Our closest H ii region
is the Orion nebula, visible with the naked eye from northern latitudes in
winter months (see Figure 8.5).
4
He is created by the fusion of 4 H nuclei in the hot cores of stars; thus
its abundance increases over the primordial value (which should provide a
universal ‘floor’) as interstellar gas is cycled through successive episodes of
star formation.

In the local Universe, there is a well-established relationship between the

mass of a galaxy and its metallicity: the least massive galaxies are also the
ones with the lowest fraction of metals, either because they have not turned
a significant fraction of their gas into stars, or because they cannot retain

8
Figure 8.6: The abundance of 4 He by mass in the lowest metallicity dwarf galaxies known,
plotted as a function of the O abundance. For comparison, in the Sun, in the Orion
nebula, and generally in the Milky Way galaxy (O/H) ' 5 × 10−4 ; thus, the galaxies
with the lowest O abundance in this figure have (O/H) ' 0.04 (O/H) . The authors of
this study (Aver et al.2015) find a shallow gradient in the regression of Y with (O/H):
d(Y )/d(O/H) = 79 ± 43 and deduce Yp = 0.245 ± 0.004. However, a zero gradient would
be an almost equally good fit to the data. This illustrates one of the uncertainties in the
determination of the primordial value of Y .

the products of stellar nucleosynthesis in their shallow potential wells (or

both). Thus, efforts to determine Yp have focussed on measuring the ratio
4
He/H from the emission lines of H ii regions in the lowest mass (dwarf),
most-metal-poor galaxies known.

As can be seen from Figure 8.6, the current best estimate Yp = 0.245±0.004
is in excellent agreement with BBN predictions (eq. 8.8). This is strong
empirical evidence that the basic idea of a hot Big Bang is correct.3

It can also be appreciated from Figure 8.4 that Yp is not a sensitive mea-
sure of η and Ωb,0 : Yp increases by only ∼ 15% from Yp = 0.220 to 0.252 as
η varies by one order of magnitude, from η = 10−10 to 10−9 . It can be diffi-
cult to achieve precisions of a few percent in astrophysical measurements,
usually because all the sources of systematic error are hard to identify and
correct for. The determination of the He abundance in H ii regions is a case
in point, with many factors contributing uncertainties of a few percent. The
extrapolation from the values of Y measured in metal-poor galaxies to the
3
As a historical note, the fact that there is a floor to the He abundance at Yp ' 0.25 was beginning to
be realised at the end of the 1940s and it led to the conjecture that (some) chemical elements may have
been synthesises in the early Universe.

9
primordial value adds a further layer of uncertainty (see Figure 8.6). For
all of these reasons, even the most careful analyses return an error which
is still frustratingly larger than required to pin down Ωb,0 . The conclusion
by Aver et al. (2015) that Yp = 0.245 ± 0.004 translates to Ωb,0 h2 values
between ∼ 8×10−3 and ∼ 2.7×10−2 , or 0.018 ≤ Ωb,0 ≤ 0.059 for h = 0.675
(from now on we shall quote values of Ωb,0 which assume h = 0.675). Note
that even the upper limit immediately tells us that baryons only account
for a few percent of the critical density.

8.4.2 Deuterium

While less abundant than 4 He by two-to-four orders of magnitude, D is

a far more sensitive ‘baryometer ’, given its steep inverse dependence on η
(Figure 8.4). Unlike He, the fragile D is burnt in the late stages of star
formation; thus, over the cosmic ages the ratio D/H has decreased from
its primordial value—a process referred to as the ‘astration’ of deuterium.
Consequently, measurements of D/H in the interstellar gas of the Milky
Way provide an upper limit to Ωb,0 .

(D/H)p is deduced from the analysis of absorption lines in pockets of in-

terstellar gas at high redshift which has undergone minimum processing
through stars, as indicated by its low metallicity (Zgas < 1/100Z ). As we
shall see in subsequent lectures, the spectra of high redshift sources (mostly
quasars) show a multitude of absorption lines at different redshifts, pro-
duced by intervening hydrogen gas in galaxies and between galaxies. Each
of these H i absorption lines has a component due to D, blueshifted by
82 km s−1 as a result of the slightly higher energy levels of the D atom
compared to H.

Normally, the D i component cannot be separated from the stronger H i

absorption, or is too weak to be detected at all, but in a few, rare cases
where all the required conditions are met,4 it has proved possible to mea-
sure (D/H)p with high precision (see Figure 8.7). The value (D/H)p =
(2.527 ± 0.030) × 10−5 reported by Cooke et al. (2017) implies Ωb,0 =
(4.91 ± 0.11) × 10−2 (both random and systematic errors are included).
This value is consistent with, and much more precise than, the value of
4
These include low velocity dispersion, high column density, low metallicity and convenient redshift
of the absorbing gas, and high luminosity of the background quasar.

10
Figure 8.7: Left: The D i component is clearly resolved at v = −82 km s−1 in the transi-
tion from the n = 1 to the n = 7 energy level of H i seen in absorption against the quasar
J1419+0829. Right: The seven high-precision determinations of (D/H) from quasar ab-
sorption line spectroscopy. The dark and light red bands show the 1 and 2σ confidence
limits on (D/H)p . Figures adapted from Cooke et al. (2014, 2017).

Ωb,0 deduced from Yp (section 8.4.1).

7
8.4.3 Lithium

7
Li is most easily observed in the atmospheres of cool stars, via a pair of
weak absorption lines of Li i at λλ6707.76, 6707.91. The post-BBN evolu-
tion of the Li abundance is complex; 7 Li is both synthesised and destroyed
during the lifetime of stars, and in addition can be produced by the inter-
action of high energy cosmic rays with atoms in the Galactic ISM. For this
reason, astronomers have looked preferentially at the oldest stars, which
are located in the halo of the Milky Way, with the hope of finding a floor
in the values of 7 Li/H indicative of (7 Li/H)p .

Indeed, such a plateau was discovered in the early 1980s: stars with iron
abundance −3 ≤ [Fe/H] ≤ −1.55 have (7 Li/H) = (1.6 ± 0.3) × 10−10 . The
lack of significant dispersion about this value led to the suggestion that it
may indeed represent the primordial abundance of 7 Li.

However, this interpretation has now come into question for two reasons:
(1) The plateau does not continue to lower metallicities—stars with [Fe/H] <
−3 have even lower values of 7 Li/H (see Figure 8.8); and (2) The plateau
value, (7 Li/H) = (1.6 ± 0.3) × 10−10 is ∼ 3 times lower than the value pre-
dicted by BBN calculations if Ωb,0 = (4.83 ± 0.10) × 10−2 , as deduced from
deuterium (see Figure 8.9). To date, the ‘Li problem’ remains unsolved.
It is unclear whether its solution will come from a better understanding
5
This is a convenient shorthand notation: [Fe/H] ≡ log (Fe/H)star − log (Fe/H) .

11
of physical processes in the interiors of stars, or a revision of the relevant
network of nuclear reactions.

8.5 Dark Matter

Anticipating some of the material in the next lectures, a completely inde-

pendent determination of Ωb,0 h2 is arrived at from the analysis of the power
spectrum of temperature fluctuations of the cosmic microwave background.
The latest, and most accurate, measure is from the Planck mission of the
European Space Agency which yielded Ωb,0 h2 = (2.226 ± 0.023) × 10−2 , or
Ωb,0 = (4.886±0.050)×10−2 for h = 0.675. As can be seen from Figure 8.9,
the CMB value is in spectacular agreement with that deduced from anal-
ysis of the D i absorption lines, giving us additional confidence in the hot
Big Bang model of the Universe.

Thus, it now seems a secure statement that baryons contribute less than
5% of the critical density to the ‘Cosmic Inventory’ of Table 1.1. This
leaves us with both a non-baryonic and a baryonic dark matter ‘problem’.

By non-baryonic dark matter we mean some component of the Universe

which does not interact with photons, and therefore does not absorb nor
emit light. This component interacts with ordinary matter (baryons) only
through gravity, and indeed its existence is surmised from observations that

Figure 8.8: A collection of 7 Li abundance measurements in metal-poor halo stars, from

Sbordone et al. (2010). The different symbols refer to different data sets brought together
here. Typical errors are indicated by the black cross. The Li abundance scale (y-axis) is
defined as A(Li) ≡ log(Li/H) + 12.

12
Figure 8.9: BBN theory confronts observations 2015 (adapted from Ichimasa et al. 2014).

ultimately involve gravity, such as the rotation curves of spiral galaxies

(Lecture 1.6), the velocity dispersion of galaxies in groups and clusters,
the growth of large scale structure in the distribution of galaxies, and
gravitational lensing— all of these topics will be considered in subsequent
lectures.

Again, it is the analysis of the CMB fluctuations that gives us the most
accurate estimate: Ωm,0 = 0.312 ± 0.009, indicating that baryons account
for only ∼ 15% of the total matter in the Universe. The nature of non-
baryonic dark matter remains a mystery and many searches are underway
in underground laboratories and in particle accelerators to identify the
presumably massive particle(s) responsible.

The term baryonic dark matter, on the other hand, is intended to simply
point to a shortcoming in our accounting of the baryons in the Universe.

13
From consideration of the mass-to-light ratio, it is estimated that stars (and
their remnants) and gas in galaxies contribute a trifling Ωstars,0 ∼ 0.003 to
the critical density. Presumably the remaining ∼ 94% of the baryons are
in gas in the halos of galaxies and in between galaxies (the intergalactic
medium) that it is difficult to detect perhaps because it is of low density
and/or at high temperature. Later on in the course we shall see that indeed
a significant fraction of the baryons is accounted for by the intergalactic
absorption lines found in the spectra of distant quasars at high redshifts.

8.6 Beyond the Standard Model?

Big-Bang nucleosynthesis is a topic at the interface between astrophysics

and particle physics. One example of this close relationship is the neutron
mean life, τn , which determines the value of Yp (see Figure 8.1). In the
1980s, the He abundance in metal-poor nebulae was used to place con-
straints on τn , a quantity which is not straightforward to measure in the
laboratory.

We have already seen in the preceding section that the existence of non-
baryonic dark matter implied by the finding that Ωb,0 ' 1/6 Ωm,0 requires
an extension of the standard model of particle physics. If we must have
dark matter, can we also have dark radiation? This term is meant to refer
to any relativistic component of the Universe which is not accounted by
the standard model.

If such a component existed and contributed to the thermal bath at the

time of BBN, it would increase the effective number of degrees of freedom
g∗ in eq. 7.26 that determines the total energy density:

1 4 4 7 7
u = g∗ a T = a T 1 + + Nν (8.10)
2 4 8
where Nν is the number of neutrino families.

The possible existence of an unrecognised relativistic component is often

incorporated into the parameter Nν . We saw in eq. 8.7 that Nν determines
the temperature of neutron freeze-out, and therefore the n/p ratio. It
follows that one can use the observed abundances of the light elements to
place constraints on the value of Nν . Figure 8.10 (left) shows that altering

14
Figure 8.10: Left: BBN predictions for light element abundances for Nν = 4 (red), 3
(green) and 2 (blue) from Cyburt et al. (2015). Right: Joint 1σ (darker colours) and 2σ
(lighter colours) confidence contours for Nν and Ωb,0 obtained from the measured values
of Yp and (D/H)p (from Cooke et al. (2014).

Nν has a large effect on the predicted value of Yp (as expected); the impact
on the abundances of the other light elements is smaller but detectable.

When the most recent empirical determinations of Yp and (D/H)p discussed

above (sections 8.4.1 and 8.4.2) are analysed together, they yield Nν =
2.85 ± 0.28 (Cyburt et al. 2015). The tighter limit, Nν = 2.88 ± 0.16 is
obtained by combining BBN with the CMB. Thus, there is no evidence
at the moment for a significant departure from the standard model value
Nν = 3. For comparison, the width of Z-boson decays in electron-positron
colliders implies Nν = 2.984 ± 0.008 (ALEPH Collaboration 2006).

8.7 Summary

In this lecture we have seen how the simple consideration that the tempera-
ture of the cosmic background radiation increases as T (z) = 2.7255 (1+z) K
leads us to conclude that there must have been an epoch in the early Uni-
verse when, over the course of a few minutes, the lightest elements of the
periodic table were synthesised from protons and neutrons via a chain of

15
nuclear reactions whose rates can be measured in the laboratory today.

BBN makes the first-order prediction that 4 He accounts for approximately

25% of the mass in baryons, as indeed observed in the least polluted en-
vironments where the abundance of He can be measured. Recent mea-
surements of the primordial abundance of deuterium strengthen the case
further. While the ‘Li problem’ remains a thorn in the side, we feel justified
in concluding that BBN is one of the pillars of the hot Big Bang model of
the Universe.

Together with inferences drawn from the analysis of the CMB, it is now
clear that baryons, the ordinary matter of our world, makes up just a little
less than 5% of the cosmic mass-energy budget, and only about 15% of
the cosmic density of matter. This opens the door to an unknown compo-
nent of the Universe, Dark Matter, whose existence is primarily deduced
from the gravity it exerts on the baryons, and in the bending of light rays
in gravitational lensing. We have no idea of the masses or other physi-
cal properties of the particles making up the Dark Matter and how they
might fit in the standard model of particle physics which includes quarks,
lepton and gauge bosons. Interestingly, the existence of an analogous dark
relativistic component seems excluded by: (a) the internal consistency of
measures Yp and (D/H)p , and (b) the good agreement between BBN and
CMB determinations of Ωb,0 .

16