The Early Universe

In the first 3 minutes of the universe, it was full of energetic

gamma rays. When the temperature of the radiation field is high
enough, the gamma-rays will annihilate and produce particles
and antiparticles
Proton-antiproton pairs can be produced when the energy of the
background radiation field is high enough:
 kT > mpc2 or T >> 1013 K
 This is the case during the first 10-4 seconds!
 After that the universe cooled, and most of the protons and
antiprotons annihilated to produce a pair of gamma rays
But wait, we’re here! For some (unknown) reason, there were
slightly more protons than antiprotons in the early universe.
109 + 1 protons for every 109 antiprotons. This means there
was matter left over to create galaxies (and stars, and planets,
and us!)
The photons from this annihilation are what we see in the
cosmic microwave background.
The Early Universe

Note that the early universe is radiation dominated. The energy density
of photons decreases as R-4 while the energy density of matter
decreases as R-3 where R is the scale factor. This is because the
photon’s energy is decreasing due to the redshifting by an extra factor
of 1/R (alternatively the temperature is dropping.)
This means the expansion rate of the early universe is different, R ~ t1/2
As the expansion proceeds the matter density drops more slowly than
the radiation density, eventually the universe becomes matter
dominated. This happens when the universe cools to 4000K at about
10,000 years.
When the universe is radiation dominated, any density perturbations
didn’t grow. Between matter-radiation equality and recombination
(T=380,000 years), only perturbations in non-baryonic dark matter
grow. Baryonic matter interacts with photons and is supported against
collapse. After recombination, baryonic matter can also fall into the
dark matter potential wells. This extra growth in density perturbations
for non-baryonic dark matter means a lower initial density perturbation
is needed to produce the structure we see today.
Time Temperature What’s Happening?
t<10-10s T>1015K ???

10-10<t<10-4s 1015>T>1012K Free electrons, quarks,

photons, neutrinos, strong
interactions
10-4<t<1 s 1012>T>1010K Free electrons, protons,
neutrons, photons,
neutrinos, strong
interactions
1<t<1012s 1010>T>10000K Nucleosynthesis begins --
atomic nuclei, free
electrons, photons,
neutrinos
1012<t<1013s 10000K>T>300 Universe becomes matter
0K dominated

1013s<t<t0 3000K>T>3K Atoms formed from nuclei

and free electrons, CMB
Timeline (in temperature units) of symmetry
breaking in the early universe
Temperature versus Time Density versus Time

17,000 years old 17,000 years old

100,000 Kelvin 10 trillion atoms per m3
Temperature [Kelvin]

Density [logarithmic]
13 billion years old 13 billion years old
3 Kelvin 0.1 atoms per m3

Time [Gyr] Time [Gyr]

The Early Universe

Electrons are 2000 times less massive than protons, with rest-
energies of 0.5 MeV.
So the radiation still produces pairs of electrons and positrons
(antielectrons), they can annihilate to produce electron neutrinos
(νe and antineutrinos νe). Similarly neutrons and protons were
being created and destroyed. Pair production also occurs.
 e- + e+ ←→ νe + νe
 e- + p ←→ n + νe , νe + p ←→ n + e+
 n ←→ p + e- + νe
 e- + e+ ←→ γ + γ
This occurs until the temperature drops to T~1010 K, t ~ 1 sec
In equilibrium there will slightly more protons than neutrons
since the neutron mass is slightly (1.293 MeV) larger
The Early Universe

We can calculate the equilibrium ratio of neutrons to protons via

the Boltzmann equation,
 n/p = exp{-(mp – mn)c2/kT}
 At T~ 1012 K, n/p = 0.985
Neutrinos are only weakly interacting particles, we need a high
density of neutrinos for them to interact with neutrons and
protons and keep them in equilibrium. By the time the universe
has cooled to T~1010 K, the density is low enough that neutrinos
are highly unlikely to interact with protons or neutrons. At the
same time, the energy level drops below the electron-positron
pair production energy level, so no new neutrinos are produced.
Thus the n/p ratio is “frozen” at the value it had at T= 1010 K ,
n/p = 0.223. For every 1000 protons, there are 223 neutrons.
The Early Universe

Neutrons by themselves (not in atomic nuclei) are unstable to

beta decay
 n → p + e- + νe
 The exponential decay time for this is 886 seconds.
 This decay destroys about 25% of the neutrons, before we can the
neutrons can combine with the protons.
When the temperature has dropped to 109 K (t=230s), neutrons
can combine with protons to form deuterium. (Recall that
deuterium is “heavy hydrogen” with a nucleus containing both a
neutron and a proton.)
 n+p →D+γ
 At this point n/p has dropped to ~0.14, so there are 140 neutrons
for every 1000 protons. The excess protons (which are converted
become the nuclei of hydrogen atoms) account for about 75% of
the total mass.
Big Bang Nucleosynthesis

Once we have deuterium, and the universe has cooled a bit

(T<109 K), we are ready to create some slightly heavier
elements (helium)
 D + D → 3H + 1H
 3H + D → 4He + n

 D + D → 3He + n
 3He + D → 4He + 1H

 These are not the same reactions as in stars (the pp chain)!

All of the neutrons are converted into helium. This means that
the we will end up with 140/2 = 70 He nuclei for every 1000
protons. 140 of those protons will end up in He, the rest (860)
are left to form the nuclei of Hydrogen atoms.
So the mass fraction of helium (Y) in the early universe is:
 Y = 4(70)/[860+4(70)] =25%
 This is very close to observed primordial helium abundance of 22-
24% which is one of confirmations of the Big Bang model!
Big Bang Nucleosynthesis

Because all the neutrons are tied up in helium, the abundance of

helium is insensitive to the matter density of the universe.
In contrast, the other elements produces in the early universe,
D, 3He, and 7Li are dependent on the amount of baryonic matter
in the universe. (We also see some beryllium.)
 4He + 3H ←→ 7Li + γ
 These elements are much less abundant than helium. The universe
expanded to rapidly to build up heavier elements.
 Note that there are no stable mass-5 nuclides, and combining
helium and tritium to get lithium requires overcoming coulomb
repulsion. So almost all of the neutrons end up in helium instead.
 There is another gap at mass-8, so BBN ends with lithium!
 At t~1000s (20 minutes), BBN ends when the temperature drops
below 3 x 108 K and the density becomes too low for fusion. There
is another “freeze-out”, as no new nuclei are created and none are
destroyed. From then until recombination (T=3000K, t=380,000
years) nothing much happens (yawn!).
Stable mass gaps in the periodic table

9Be

7Li

6Li

4He

3He

2H No stable nuclei

1H

The lack of stable elements with masses 5 and 8

make it more difficult for cosmic nucleosynthesis
to progress beyond Lithium and even Helium.
1%

Mass fraction of nuclei vs time (and temperature)

Big Bang Nucleosynthesis

The BBN model makes detailed predictions of the abundances of

light elements
These are generally given as a function of η which is baryon to
photon ratio, nn/nγ. This number is pretty small, so we usually
define it in units of 1010. Also it is directly related to Ωb, the
baryon density relative to the critical density of the universe.
Thus, η10 = 1010(nn/nγ) = 274 Ωbh2
As the universe evolves this ratio is preserved, so that what we
observe today should reflect the conditions in the early universe.
If we can observe PRIMORDIAL abundances of these elements
we can:
 Test the big bang model!
 Measure the baryon density of the universe!
SBBN predicted primordial abundances of
D, 3He, 7Li, and the helium mass fraction Y
as a function of nucleon abundance
Deuterium

Any deuterium that is incorporated into stars is quickly fused into

helium and destroyed.
But there is also no astrophysical locations where deuterium will
be created in large amounts after BBN.
Thus, we need to measure deuterium in an area that has
undergone very little stellar processing
This can be done by measuring deuterium absorption in QSO
absorption line systems. The intervening systems are high
redshift (and low metallicity). This has been done for ~6 QSO
absorption line systems. (Requires high resolution spectra from
10m-class telescopes, this is hard!)
One also has to separate out H absorption from D absorption
and account for any velocity effects.
There is a lot of dispersion in the results,
D/H = (2.6 +/- 0.4) x 10-5
Measuring D/H in
QSO absorption lines

Tytler & Burles

Deuterium abundance (D/H) vs metallicity (X where
X is generally Si) from QSO absorption line systems
Helium-3

Interstellar 3He that is incorporated into stars is burned to 4He in

the hot stellar interiors, but preserved in cooler, outer stellar
layers.
3He is also created in hydrogen fusion in low-mass stars.

Unclear how much of this is returned into the interstellar

medium versus being consumed in post-main sequence
evolution.
Thus any determination of primordial 3He abundance is model
dependent. Can be used to provide a consistency check on
other measurements.
Singly ionized 3He has been observed in emission in a handful of
Galactic HII regions, 3He/H = (1.1 +/- 0.2) x 10-5
Helium-4

4He is the second most abundant element in the universe. It is

created in main sequence stars raising its abundance over its
primordial value.
We need to observe 4He from recombination lines in extremely
low-metallicity regions. These conditions can be found in
extragalactic HII regions in low-metallicity galaxies.
Expect a plateau in helium abundance. We have measured Y
(helium mass fraction) in galaxies ranging from ~1/2 to 1/40
solar. We can either measure the bound from the lowest
metallicity measurement OR extrapolate the trend to zero
metallicity.
Note there are statistical uncertainties in these measurements
related to corrections for collisional excitation of neutral helium,
assumed temperature and ionization, temperature fluctuations,
etc.
Find Y = 0.234 +/- 0.003, 0.244 +/- 0.002, OR the compromise
mean with the larger error bar: 0.238 +/- 0.005.
Helium mass fraction Y vs O/H for low-metallicity
extragalactic HII regions
Helium mass fraction Y vs O/H for low-metallicity
extragalactic HII regions. Same data, different interpretation!
Lithium-7

7Li is fragile, it burns in stars at a relatively low temperature.

Thus the majority of any interstellar 7Li is cycled through stars
and destroyed.
However it is also difficult for stars to create new 7Li or to return
any newly synthesized 7Li to the ISM before it is destroyed by
nuclear burning.
We expect a plateau in 7Li in low metallicity environments.
7Li is observed in absorption in the atmospheres of cool, metal-

poor, Population II halo stars.

There are uncertainties due to not only observational
uncertainties, but model dependencies in stellar atmospheres.
Find [Li] = 12 + log(Li/H) = 2.2 +/- 0.1, but values have been
quoted from 2.1 to 2.4…
1012 (Li/H)

Compilation of Lithium abundance data vs. [Fe/H]

from stellar observations.
What does it mean?

We can use the observed deuterium abundance to fix the

photon/baryon ratio. This gives η10 = 6.1 +/- 0.7 which
corresponds to Ωbh2 = 0.022 +/- 0.003.
This is in AMAZING agreement with the WMAP estimate of Ωbh2
= 0.0223 +/- 0.0009. Big bang nucleosynthesis works!!
To first order, this baryon density is consistent with the observed
abundances of 3He, 4He, and 7Li. But there are some
discrepancies with 4He and 7Li.
BBN prediction for 3He fraction is (1.0 +/- 0.1) x 10-5,
measurement is (1.1 +/- 0.1) x 10-5. This is good.
BBN prediction for Y (4He mass fraction) is 0.248 +/- 0.001. It is
tightly constrained. But the measurement is 0.238 +/- 0.005.
This is bad …
BBN prediction for 7Li is [Li] = 2.65 +/- 0.11. The observation is
[Li] = 2.2 +/- 0.1. This is also bad …
SBBN predicted primordial abundances of
D, 3He, 7Li, and the helium mass fraction Y vs.
baryon density
SBBN predicted primordial mass fraction of 4He (Y) vs. D/H.
Point is measurement of primoridal abundances
SBBN predicted primordial Li/H vs. D/H.
Points are measurements of primoridal abundances
What does it mean?

Can we save BBN?

Maybe this is a sign of new physics –
 Y is dependent on the expansion rate. If the universe expands
slower, the predicted Y will go down.
 WMAP data also are sensitive to early universe expansion rate and
agree with BBN, but data are still consistent with standard BBN.
 What would cause this is a mystery …
 Another possibility is neutrino asymmetry, (number of electron
neutrinos to electron antineutrinos) which regulates the pre-BBN
neutron/proton ratio. Again data are not inconsistent with
symmetry but skew towards slightly more neutrinos than
antineutrinos. This would be due to a non-zero chemical potential
for neutrino-antineutrino annihilation.
Note that neither of these solves the lithium discrepancy! This
may be due a poor understanding of the evolution of very old,
metal-poor halo stars. Perhaps lithium is being depleted in the
surfaces of these stars.
SBBN predicted primordial mass fraction of 4He (Y) vs. D/H plus
predictions of faster and slower early universe expansion rates.
Point is measurement of primordial abundances.