Big Bang Nucleosynthesis

Subir Sarkar

DPG Physics school Astroparticle Physics, Physikzentrum Bad Honnef, 20-25 Sept 2009

The universe is made mainly of hydrogen (~75%) and helium (~25%) + traces of heavier elements

Where did all the elements come from?

Big Bang! Stars/Supernovae!

George Gamow is generally credited with having founded the theory of primordial nucleosynthesis and, as a corollary, predicted the temperature of the relic radiation

The real story is that while Gamow had brilliant ideas, he could not calculate very well, so enlisted the help of a graduate student Ralph Alpher (who worked with Robert Herman)

1) was published on 1 April 1948 including Bethe (who had nothing to do with it) but leaving out Herman because he stubbornly refused to change his name to Delter!

The modern theory of primordial nucleosynthesis is based essentially on this paper which followed the crucial observation by Hayashi (Prog. Theoret. Phys. 5:224,1950) that neutrons and protons were in chemical equilibrium in the hot early universe

Alphers achievement was nally recognized when he was awarded the US National Medal of Science in 2005: "For his unprecedented work in the areas of nucleosynthesis, for the prediction that universe expansion leaves behind background radiation, and for providing the model for the Big Bang theory."

(See lectures by Prof Gerhard Brner)

In the absence of dissipative processes (e.g. phase transitions which generate entropy) the comoving entropy is conserved:

i.e.

The dynamics is governed by the Friedmann equation:

a 2 a

8 GN 3

Integrating this gives the time-temperature relationship:

t (s) = 2.42 g-1/2 (T/MeV)-2

So we can work out when events of physical signicance occurred in our past (according to the Standard Model of particle physics)

To get this right we need to count all the bosons and fermions contributing to the relativistic degrees of freedom and take into account our uncertain knowledge of possible phase transitions

quark-hadron (de) connement transition

See lecture by Prof Paolo de Bernadis

The Cosmic Microwave Background Spectrum

This perfect blackbody is testimony to our hot, dense past and directly demonstrates that the expansion was adiabatic (with negligible energy release) back at least to t ~ 1 day we can go back further to t ~ 1 s by studying element synthesis

Weak interactions and nuclear reactions in expanding, cooling universe

(Hayashi 1950, Alpher, Follin & Herman 1953, Peebles 1966, Wagoner, Fowler & Hoyle 1967) Dramatis personae: Radiation (dominates) Matter baryon-to-photon ratio (only free parameter) Initial conditions: T >> 1 MeV, t << 1 s n-p weak equilibrium: neutron-to-proton ratio: Weak freeze-out: Tf ~ 1 MeV, tf ~ 1 s which xes: Deuterium bottleneck: T ~ 1 0.07 MeV D created by but destroyed by high-E photon tail: so nucleosynthesis halted until:

p +! # n + e+ e 1/ 3 " # 2 ! weak (n * p) ( tuniverse ) Tfreeze-out $ GN / GF %

n +! e # p + e"

n p = e!( mn !m p )/ Tf " 1 6

&

'

Tnuc ~ D/-ln()

Element synthesis: Tnuc ~ 0.07 MeV, tnuc ~ 3 min (meanwhile n/p 1/7 through neutron -decay) essentially all n 4He (YP ~ 25% by mass) + left-over traces of D, 3He, 7Li (with 6Li/7Li ~ 10-5) No heavier nuclei formed in standard, homogeneous hot Big Bang must wait for stars to form after a ~billion years and synthesise all the other nuclei in the universe (s-process, r-process, )

Computer code by Wagoner (1969, 1973) .. updated by Kawano (1992) Coulomb & radiative corrections, heating et cetera (Dicus et al 1982)

Nucleon recoil corrections (Seckel 1993)

Covariance matrix of correlated uncertainties (Fiorentini et al 1998)

Updated nuclear cross-sections (NACRE 2003)

The rst three minutes

 Time < 15 s, Temperature > 3 x 109 K

universe is soup of protons, electrons and other particles so hot that nuclei are blasted apart by high energy photons as soon as they form
Time = 15 s, Temperature = 3 x 109 K

Still too hot for Deuterium to survive Cool enough for Helium to survive, but too few building blocks
Time = 3 min, Temperature = 109 K

Deuterium survives and is quickly fused into He no stable nuclei with 5 or 8 nucleons, and this restricts formation of elements heavier than Helium trace amounts of Lithium are formed
Time = 35 min, Temperature = 3 x 107 K

nucleosynthesis essentially complete Still hot enough to fuse He, but density too low for appreciable fusion Model makes precise predictions about the relative abundances of the light elements 2H, 3He, 4He and 7Li, as a function of the nucleon density

Primodial versus Stellar Nucleosynthesis

Timescale Stellar Nucleosynthesis (SN): billions of years Primordial Nucleosynthesis (PN): minutes Temperature evolution SN: slow increase over time PN: rapid cooling Density SN: 100 g/cm3 PN: 10-5 g/cm3 (like air!) Photon to baryon ratio SN: less than 1 photon per baryon PN: billions of photons per baryon
3He 2H 1H 7Li 6Li

9Be

4He

no stable nuclei!

The lack of stable elements with masses 5 and 8 make it hard for BBN (2-body processes, short time-scale) to synthesise elements beyond helium

this can be happen only in stars, on a longer timescale

The neutron lifetime normalises the weak interaction rate: n = 885.7 0.8 s (a recent measurement is 6.5 lower not included by the PDG in the average)

Uncertainties in synthesized abundances are correlated estimate using Monte Carlo methods
(Krauss & Romanelli 1988; Smith, Kawano & Malaney 1993; Krauss & Kernan 1994; Cyburt, Fields & Olive 2004)

Courtesey: Keith Olive

Linear propagation of errors covariance matrix (in agreement with Monte Carlo results)

Fiorentini, Lisi, Sarkar & Villante (1998)

BBN Predictions
line widths theoretical uncertainties (neutron lifetime, nuclear cross sections)

Courtesey: Keith Olive

Nucleosynthesis without a computer

source

sink

but general solution is:

If

then abundances approach equilibrium values

Freeze-out occurs when:

Examine reaction network to identify the largest source and sink terms
... analytic solution

obtain D, 3He and 7Li to within a factor of ~2 of exact numerical solution, and 4He to within a few %

Dimopoulos, Esmailzadeh, Hall & Starkman (1988)

 can use this formalism to determine joint dependence of abundances on expansion rate as well as baryon-to-photon ratio and so:

can therefore employ simple 2 statistics to determine best-t values and uncertainties (faster than Monte Carlo + Maximum Likelihood)

Lisi, Sarkar & Villante (2000)

Inferring primordial abundances

Courtesey: Keith Olive

For a quantity of such fundamental cosmological importance, relatively little effort has been spent on measuring the primordial helium abundance

Courtesey: Keith Olive

This is the value (and uncertainty) presently recommended by the PDG

Olive & Skillman (2004)

... is easily destroyed in stars

Look in Quasar Absorption Systems - low density clouds of gas seen in absorption along the lines of sight to distant quasars (when universe was only ~10% of its present age) The difference between H and D nuclei causes a small change in the energies of electron transitions, shifting their absorption lines apart and enabling D/H to be measured

Primordial deuterium?

But: Hard to nd clean systems Do not resolve clouds Dispersion/systematics?

W. M. Keck Observatory
Spectra with the necessary resolution for such distant objects can be obtained with 10m-class telescopes this has revolutionised the determination of the primordial D abundance

The observed scatter is not consistent with uctuations about an average value!

Courtesey: Keith Olive

Primordial Lithium?
Observe in primitive (Pop II) stars: (most abundant isotope is 7Li) - Li-Fe correlation mild evolution - Transition from low mass/surface temp stars (core well mixed by convection) to higher mass/temp stars (mixing of core is not efcient)
Courtesey: Keith Olive

Plateau at low Fe (high T) constant abundance at early epochs so infer observed 7Li plateau is primordial (Spite & Spite 1982)

Inferred primordial abundances

4He

observed in extragalactic HII regions:

2H

observed in quasar absorption systems (and ISM):

7Li

observed in atmospheres of dwarf halo stars:

(3He can be both created & destroyed in stars so primordial abundance cannot be reliably estimated)
Systematic errors have been re-evaluated based on scatter in data
for details see Review of Particle Physics (Fields & Sarkar, Phys. Lett. 667, 1, 2008)

allowing for large systematic uncertainties in the inferred elemental abundances

!BBN

BBN versus CMB

is in agreement with CMB

Conrms and sharpens the case for (two kinds of) dark matter Baryonic Dark Matter: warm-hot IGM, Ly- , X-ray gas + Non-baryonic dark matter: neutralino? axion?

Particle data Group: Fields & Sarkar (2008)

Another argument comes from considerations of structure formation in the universe

Perturbations in metric (generated during ination) induce perturbations in photons and (dark) matter

These perturbations begin to grow through gravitational instability after matter domination

Before recombination, the primordial uctuations just excite sound waves in the plasma, but can start growing already in the sea of collisionless dark matter
Courtesey David Spergel

These sound waves leave an imprint on the last scattering surface of the CMB as the universe turns neutral and transparent sensitive to the baryon/CDM densities

For a statistically isotropic gaussian random eld, the angular power spectrum can be constructed by decomposing in spherical harmonics:

Hu, Sugiyama, Silk [astro-ph/9604166]

The Cosmic Microwave Background

!T !Bh provide independent measure of
2

Acoustic oscillations in (coupled) photon-baryon uids imprint o features at small angles (< 1 ) in angular power spectrum Detailed peak positions, heights, sensitive to cosmological parameters e.g. 2nd/1st peak baryon density

WMAP-5 best-t:
Bond & Efstathiou (1984) Dodelson & Hu (2003)

 in more detail
Predict BBN abundances with WMAP determination of CMB(blue) compare with observations (yellow) D agreement excellent, 4He also OK But 7Li is discrepant

- systematic errors in observations? - theoretical uncertainties? - new physics (e.g. decaying relic particles)? this has additional motivation from the observation that 6Li has also been observed with an abundance > 104 times higher than expected!
Cyburt, Fields & Olive (2008)

Courtesey: Keith Olive

Ryan, Beers, Olive, Fields & Norris (2000)

Systematic Errors in the inferred Lithium abundance

Observational systematics

Measure Li I absorption line(s) to infer 7Li/H Teff critical (mostly Li II) But required shift in T scale is ~500 K - very unlikely
Melendez & Ramirez (2004); Fields, Olive & Vangioni-Flam (2005)

Astrophysical systematics

Stellar depletion over ~1010 yr if Li burned need to correct Lip upward But no scatter seen around Spite plateau - also 6Li preserved
Ryan et al (2000)

Nuclear Systematics

production channel - 3He (, ) 7Be - normalization error? But same reaction also key for Solar neutrinos standard Solar model OK!
7Li

Cyburt, Fields & Olive (2004)

Recently a primordial plateau in 6Li has indeed been detected with 6Li/7Li ~ 0.1 (cf. standard expectation 6Li/7Li ~ 10-5 )
(Nissen et al 1999; Asplund et al 2001, 2004)

Coupled with the fact that the 7Li abundance is ~3 times smaller than expected, this has refocussed interest on non-standard BBN

However the detection of 6Li is based on ts to the line shape need more data to establish the reality of a 6Li plateau!

Also stars in which 6Li is detected are close to the main-sequence turn-off in the H-R diagram

Lambert (2005)

Does the Lithium anomaly imply new physics?

Jedamzik (2000, 2004)

BBN and decaying particles

Extensions of the Standard Model predict new (typically) unstable particles, which would have been created (thermally) in the early Universe, e.g. TeV mass gravitinos in supergravity

(Weinberg 1982; Khlopov & Linde 1983; Ellis, Nanopoulos & Sarkar 1985; Reno & Seckel 1988)

The high energy photons would have photo-dissociated the synthesized elements severe limits on the decaying particle abundance

Mass x relic abundance (GeV)

This requires that highest temperature reached in our past (after ination) was < 108 GeV - constraint on baryogenesis!

particle lifetime (s)

Cyburt et al 2003

May be possible to solve both lithium problems with relic decaying particle having suitable abundance/lifetime e.g. gluino in split supersymmetry, supersymmetric stau Next-to-LSP (with gravitino LSP),

Bailly, Jedamzik & Moultaka 2008

Gluino in split supersymmetry

If mass scale of SUSY scalar superpartners is raised well above a TeV (to evade various problems with weak scale SUSY breaking), then predict long-lived gluinos

A small number of these would survive annihilation in the early universe and decay during nucleosynthesis stringent bound from overproduction of D + 3He

This would require supersymmetry breaking scale to be < 1010 GeV

Arvinataki, Davis, Graham, Pierce & Walker (2005)

There may also be new charged quasi-stable relic particles in Nature which would form bound states with 4He Although the 4He (D, ) 6Li reaction is normally highly suppressed, this is not so for the bound state

Pospelov (2006) Thus the lithium anomaly may be due to supersymmetric particles (e.g. stau) which catalyse relevant nuclear reactions if so these could be seen soon at the LHC!

There is something fascinating about science. One gets such wholesome returns of conjectures out of such triing investment of fact.
Mark Twain

Courtesey: Keith Olive

Example: Neutrino Counting

Element abundances sensitive to expansion history during BBN observed values constrain relativistic energy density
(Hoyle & Taylor 1964, Peebles 1966; Shvartsman 1969; Steigman, Schramm, & Gunn 1977)

Pre-CMB: 4He as probe, other elements give With from CMB: All abundances can be used 4He still sharpest probe D competitive if measured to 3%

Cyburt, Fields, Olive & Skillman (2005); Lisi, Sarkar & Villante (1999),

N N 3 < 1.5 @ 95% c.l. so singlet neutrino (cf. LSND) is allowed

Example: Fundamental couplings

Note n-p mass difference is sensitive to both em and strong interactions, while freeze-out temp is sensitive to weak interactions and gravity, hence 4He abundance is exponentially sensitive to all coupling strengths Conversely obtain bound of < few % on any additional contribution to energy density driving expansion e.g. rules out of O(H2) always (since this would correspond to a large renormalisation of GN)

In fundamental theories e.g. string theory, the physical constants do vary with time but the BBN constraint says that this must have stopped before t ~ 0.1 s

Courtesey: Keith Olive

Summary
Observational inferences about the primordially synthesised abundances of D, 4He and 7Li presently provide the deepest probe of the Big Bang, based on an established physical theory
The overall concordance between the inferred primordial abundances of D and 4He with the predictions of the standard cosmology requires most of the matter in the universe to be non-baryonic, and enables constraints to be placed on any deviations from the usual expansion history (e.g. new neutrinos or dark energy)

Anomalies in the abundances of 6Li and 7Li have been interpreted as indications for new physics beyond the Standard Model (viz. unstable supersymmetric particles) need better understanding of the astrophysical processing of lithium to investigate this further
Nucleosynthesis marked the beginning of the development of modern cosmology and it is still the nal observational frontier as we look back to the Big Bang!