AST5220 Recombination

Transcription

1 AST5220 Recombination Hans Kristian Eriksen February 16th, 2010

2 Overview Some important points: Photons scatter on free electrons In the early universe, the temperature was very hot Too hot for neutral hydrogen to exist; electrons and protons were ripped apart ultraviolet photons Consequently, all electrons in the early universe were free And therefore the universe was opaque It was impossible to see more than ~1 meter ahead, because light was scattered from a different direction Just like walking in a fog But today the universe is transparent Something must have happened at some point Question: How did the universe become transparent?

3 Cosmic timeline

4 Zoom-in on recomination epoch

5 Why is this epoch important to us? Recall our main goal: We want to predict the CMB power spectrum given cosmological parameters The CMB photons we observe today are those that decoupled during recombination The fluctuations we see are mainly those that existed at the time of recombination We therefore need to know when recombination happened 1 how rapidly it happened Power spectrum (10 3 µk 2 ) Multipole moment, l

6 The optical depth To describe recombination quantitatively we introduce the concept of optical depth, τ Imagine you are looking at a light source through a medium that absorbes light The full intensity of the light source is called I 0 The intensity you observe at position x is then I(x)=I 0 e τ(x) where τ is called the optical depth The further away you are, the higher τ is The critical position is where τ ~ 1 If τ >> 1, you don t see anything If τ << 1, the medium doesn t do anything anyway

7 The cosmological optical depth In cosmology, the main source of absorption is free electrons through Thompson scattering The optical depth of Thompson scattering is given by where τ(η)= η0 η n e σ T adη η is conformal time n e = electron density σ T = Thompson cross section ( probability of single scattering?) Our main job: Compute τ(η)

8 How do we compute τ? τ(η)= η0 η n e σ T adη The Thompson cross section is given by particle physics: σ T = 8π 3 ( α m e c ) cm 2 a is the scale factor, and a one-to to-one one function of the conformal time The difficult one is the electron density, n e...

9 Finding the electron density (1) Assume that there are no helium or heavier elements in the universe, only electrons and protons That is, there are four constituents in the universe: Protons, p Electrons, e Neutral hydrogen, H Photons, γ These interact with each other through a two-particle process p+e H+γ We need to find n e, n p, n H and n γ

10 Finding the electron density (2) And from the last lecture, we know how to do this! Recall first the Boltzmann equation for a two-particle process in an expanding universe: a 3d(n ea 3 ) d 3 p 1 d 3 p 2 d 3 p 3 d 3 p 4 = dt (2π ) 3 2E 1 (2π ) 3 2E 2 (2π ) 3 2E 3 (2π ) 3 2E 4 After defining n i =e µ i c2 kt d 3 p E i (2π) 3e kt i = this could (for thermal equlibrium) be written as a 3d(n 1a 3 ) dt (2π) 4 δ (3) (p1+p2 p3 p4)δ(e 1 +E 2 E 3 E 4 ) M 2 [f 3 f 4 f 1 f 2 ] = 1 n(0) 2 σv { n 3 n 4 3 n(0) 4 d 3 p E i (2π) 3e kt n 1n 2 1 n(0) 2 }

11 Finding the electron density (3) Recall also that if the reaction rate is much larger than the Hubble expansion rate, then this implies n 3 n 4 3 n(0) 4 = n 1n 2 1 n(0) 2 For our proton-electron electron-hydrogen-photon plasma, this therefore reads n e n p e p = n Hn γ H n(0) γ This is the Saha equation on a symbolic form, and now we have to write this out in full. To the blackboard...

12 The Saha equation So, the Saha equation reads X 2 e 1 X e 1 n b ( me T b 2π 2 ) 3/2 e ǫ 0/T b This is a good approximation as long as the system is in strong thermodynamic equilibrium But when the temperature and density fall, the system eventually goes out of thermodynamic equilibrium

13 The general equation When that happens, one has to go back to the full Boltzmann equation If the only involved particles and states were photons, electrons s and hydrogen, this would (as usual) look like this: a 3d(n ea 3 ) dt { = e n(0) p σv However, in real life this quickly becomes messy because of the atomic physics involved: First, e + p H s1 + γ doesn t t count the released photon will immediately ionize another nearby hydrogen atom, unless the distance between atoms is very large The most important contribution is e + p H s2 + γ This leads to the Peebles equation For high precision, helium, litium and excited states must be included This is what Recfast does, and it is messy n H H n en p e p }

14 The Peebles equation a 3d(n ea 3 ) dt = e n(0) p σv Let n e = n b X e. Then the left-hand side of the Boltzmann equation becomes a 3d(n ea 3 ) dt The right-hand hand side becomes e p σv { n H H n en p e p { n H H =a 3d(X e(n b a 3 )) dt } =n b σv { (1 X e ) n en p e p } =n b dx e dt ( ) 3/2 } me T e ǫ/kt X 2 2π en 2 b

15 Cancelling terms, the equation then reads dx e dt The Peebles equation = σv { (1 X e ) ( me T 2π 2 ) 3/2 e ǫ/kt X 2 en b } With the correct constants, this leads to the Peebles equation You will solve this numerically in the second milestone of the projectp The full set of constants are given in the project description

16 Numerical solution Figure 1 of Callin (2005)

17 Optical depth and the visibility function We now have everything we need to compute the optical depth τ(η)= η0 since we finally know n e = X e n b η n e σ T adη However, we will also need the so-called visibility function g(η)= dτ dη e τ(η) This is simply the probability for a given observed photon to have scattered at conformal time η

18 Optical depth and the visibility function Solid line = Visibility function z ~ 1100

19 Summary We need to know the recombination history of the universe in order to predict the CMB spectrum Specifically, we need the optical depth, τ,, and the visibility function, g To find these, we must solve the Boltzmann equation for the electron number density in the early universe At early times, use the Saha approximation At late times, use the Peebles equation In order to do this job properly, one should really take into account count heavier elements (helium, litium etc.) and many excitation states s of each atom Not trivial, and fortunately somebody else has spent a lot of time on this, and written Recfast We will be happy with the simpler approximations above in this coursec Recombination (and reionization) is messy...