CMB Polarization and Cosmology Wayne Hu KIPAC, May 2004
Outline Reionization and its Applications Dark Energy The Quadrupole Gravitational Waves Acoustic Polarization and Initial Power Gravitational Lensing as Signal and Contaminant Recent Polarization Collaborators: Christopher Gordon Matt Hedman Gil Holder Manoj Kaplinghat Takemi Okamoto Kendrick Smith
100 Polarized Issues 10 reionization ΘE (µk) 1 EE BB 0.1 0.01 Hu & Dodelson (2002) gravitational waves 10 100 1000 l (multipole) gravitational lensing
Polarization from Thomson Scattering Quadrupole anisotropies scatter into linear polarization Hu & White (1997) aligned with cold lobe
Whence Polarization Anisotropy? Observed photons scatter into the line of sight Polarization arises from the projection of the quadrupole on the transverse plane Hu & White (1997)
Polarization Multipoles Mathematically pattern is described by the tensor (spin-2) spherical harmonics [eigenfunctions of Laplacian on trace-free 2 tensor] Correspondence with scalar spherical harmonics established via Clebsch-Gordan coefficients (spin x orbital) Amplitude of the coefficients in the spherical harmonic expansion are the multipole moments; averaged square is the power Seljak & Zaldarriaga (1997); Kamionkowski, Kosowsky & Stebbins (1997) E-spin harmonic l=2, m=0
Reionization
Temperature Inhomogeneity Temperature inhomogeneity reflects initial density perturbation on large scales Consider a single Fourier moment:
Locally Transparent Presently, the matter density is so low that a typical CMB photon will not scatter in a Hubble time (~age of universe) transparent observer recombination
Reversed Expansion Free electron density in an ionized medium increases as scale factor a -3 ; when the universe was a tenth of its current size CMB photons have a finite (~10%) chance to scatter recombination rescattering
Polarization Anisotropy Electron sees the temperature anisotropy on its recombination surface and scatters it into a polarization recombination polarization
Temperature Correlation Pattern correlated with the temperature anisotropy that generates it; here an m=0 quadrupole
WMAP Correlation Measured correlation indicates the universe remained at least partially ionized to a surprisingly large redshift or early time (z>10) (l+1)c l/2π (µk 2 ) 3 2 1 0 Reionization TE Cross Power Spectrum Kogut et al (2003) -1 0 10 40 100 200 400 800 1400 Multipole moment (l)
Why Care? Early ionization is puzzling if due to ionizing radiation from normal stars; may indicate more exotic physics is involved Reionization screens temperature anisotropy on small scales making the true amplitude of initial fluctuations larger by e τ Measuring the growth of fluctuations is one of the best ways of determining the neutrino masses and the dark energy Offers an opportunity to study the origin of the low multipole statistical anomalies Presents a second, and statistically cleaner, window on gravitational waves from the early universe
Ionization History
Polarization Power Spectrum Most of the information on ionization history is in the polarization (auto) power spectrum - two models with same optical depth but different ionization fraction partial ionization step 500µK' l(l+1)c l EE /2π 10-13 50µK' 10-14 10 100 Kaplinghat et al (2002) [figure: Hu & Holder (2003)] l
Principal Components Information on the ionization history is contained in ~5 numbers - essentially coefficients of first few Fourier modes 0.5 δx δx 0-0.5 0.5 0 (a) Best -0.5 (b) Worst Hu & Holder (2003) 5 10 z 15 20
Representation in Modes Reproduces the power spectrum and net optical depth (actual τ=0.1375 vs 0.1377); indicates whether multiple physical mechanisms suggested 0.8 l(l+1)c l EE /2π Hu & Holder (2003) 0.4 10-13 0 10-14 x 10 15 20 25 true sum modes fiducial z 10 100 l
Total Optical Depth Optical depth measurement unbiased Ultimate errors set by cosmic variance here 0.01 Equivalently 1% determination of initial amplitude for dark energy 10 2 10 1 1 10-1 10-2 10-3 σ µ σ τ (cumul.) prior prior 10-4 10-5 τ µ 5 10 15 Hu & Holder (2003) mode µ
Gravitational Waves
Gravitational Waves Inflation predicts near scale invariant spectrum of gravitational waves Amplitude proportional to the square of the Ei=V 1/4 energy scale If inflation is associated with the grand unification Ei~10 16 GeV and potentially observable transverse-traceless distortion
Gravitational Wave Pattern Projection of the quadrupole anisotropy gives polarization pattern Transverse polarization of gravitational waves breaks azimuthal symmetry density perturbation gravitational wave
Electric & Magnetic Polarization (a.k.a. gradient & curl) Alignment of principal vs polarization axes (curvature matrix vs polarization direction) E B Kamionkowski, Kosowsky, Stebbins (1997) Zaldarriaga & Seljak (1997)
The B-Bump Rescattering of gravitational wave anisotropy generates the B-bump Potentially the most sensitive probe of inflationary energy scale 10 P (µk) 1 EE BB maximum amplitude! reionization B-bump recombination B-peak lensing contaminant 10 100 1000 l
T/S, Inflation and the B-Bump B-bump up to 20x more sensitive to T/S In combination with recombination peak, constrain spectrum T/S 10-1 10-2 τ=0.17 peak 10 idealized Planck 16 10-3 10-4 10-5 bump V 1/4 (GeV) 10-6 lensing contaminated lensing partially removed 10 15 1 10 Hu (2001) Knox & Song (2002); Cooray et al (2002) noise (µk-arcmin) further: Hirata & Seljak (2003)
Quadrupole Aside
Low Quadrupole Known since COBE: a ~2σ problem 6000 ΘΘ l(l+1)c l /2π (µk 2 ) 4000 2000 10 l 100
ISW Spatial Modes ISW effect comes from nearby acceleration regime Shorter wavelengths project onto same angle
Quadrupole Origins Transfer function for the quadrupole 0.2 total T E 2 T Θ 2 0-0.2 0.002 0 ISW SW total 10 3 z<1 (a) Temperature -0.002 0.0001 0.001 k (Mpc -1 ) (b) Polarization 0.01 Gordon & Hu (2004)
Lowering the Quadrupole Transfer function for the quadrupole 0.2 T E 2 T Θ 2 0-0.2 0.002 0-0.002 0.0001 fiducial sound speed cut-off 0.001 k (Mpc -1 ) (a) Temperature (b) Polarization 0.01 Gordon & Hu (2004) [Contaldi et al 2003; Hu 1998; Erikson et al 2002; Bean & Dore 2003]
Low Quadrupole Models Distinguished by polarization ΘE l(l+1)c l /2π (µk 2 ) 5 0-5 fiducial sound speed cut-off EE l(l+1)c l /2π (µk 2 ) 0.4 0.3 0.2 0.1 Gordon & Hu (2004) [Dore, Holder & Loeb 2003] 10 l 100
Low Quadrupole Models Models: initial conditions vs. dark energy ΘΘ l(l+1)c l /2π (µk 2 ) 6000 4000 2000 fiducial sound speed cut-off 10 100 l Gordon & Hu (2004) [Contaldi et al 2003; Hu 1998; Erikson et al 2002; Bean & Dore 2003]
Initial Spectrum
Hu & Okamoto (2003) Transfer of Initial Power
Prospects for Initial Conditions Polarization crucial for detailed study of initial conditions, decade in scale of the acoustic peaks can provide exquisite tests of scale free initial conditions Hu & Okamoto (2003) 1 cosmological parameters fixed fractional error 0.1 temperature 0.01 & polarization 0.01 k (Mpc -1 ) 0.1 even with large lever arm other tests must match accuracy Wang et al (1999); Kinney (2001); Miller et al (2002); Tegmark & Zaldarriaga (2002); Bridle et al (2003)
Prospects for Initial Conditions Polarization crucial for detailed study of initial conditions, decade in scale of the acoustic peaks can provide exquisite tests of scale free initial conditions Hu & Okamoto (2003) 1 cosmological parameters marginalized fractional error 0.1 temperature 0.01 & polarization 0.01 k (Mpc -1 ) 0.1 even with large lever arm other tests must match accuracy Wang et al (1999); Kinney (2001); Miller et al (2002); Tegmark & Zaldarriaga (2002); Bridle et al (2003)
100 Temperature and Polarization Spectra 10 reionization ΘE (µk) 1 EE BB 0.1 0.01 gravitational waves 10 100 1000 l (multipole) gravitational lensing
Gravitational Lensing
Gravitational Lensing Gravitational lensing by large scale structure distorts the observed temperature and polarization fields Exaggerated example for the temperature Original Lensed
Polarization Lensing
Polarization Lensing Since E and B denote the relationship between the polarization amplitude and direction, warping due to lensing creates B-modes Original Lensed E Zaldarriaga & Seljak (1998) [figure: Hu & Okamoto (2001)] Lensed B
100 Temperature and Polarization Spectra 10 reionization ΘE (µk) 1 EE BB 0.1 0.01 gravitational waves 10 100 1000 l (multipole) gravitational lensing
Lensing by a Gaussian Random Field Mass distribution at large angles and high redshift in in the linear regime Projected mass distribution (low pass filtered reflecting deflection angles): 1000 sq. deg rms deflection 2.6' deflection coherence 10
Power Spectrum Measurements Lensed field is non-gaussian in that a single degree scale lens controls the polarization at arcminutes Increased variance and covariance implies that 10x as much sky needed compared with Gaussian fields fractional power deviation 0.02 0.01 0-0.01-0.02 realizations Gaussian variance 100 Smith, Hu & Kaplinghat (2004) l 1000
Sample vs Noise Variance Non-Gaussian sample variance doubles total variance at 4µK' for resolved B-modes 10 Variance Degradation 8 6 4 2 20' 10' 3' FWHM=1' 1 P (µk-arcmin) 10
Reconstruction from Polarization Lensing B-modes correlated to the orignal E-modes in a specific way Correlation of E and B allows for a reconstruction of the lens Reference experiment of 4' beam, 1µK' noise and 100 deg 2 Original Mass Map Hu & Okamoto (2001) [iterative improvement Hirata & Seljak (2003)] Reconstructed Mass Map
Matter Power Spectrum Measuring projected matter power spectrum to cosmic variance limit across whole linear regime 0.002< k < 0.2 h/mpc P P deflection power 0.6 ( mtot ev 10 7 10 8 ) Reference 10 100 1000 Hu & Okamoto (2001) [parameter forecasts: Kaplighat et al 2003] L Linear σ(w) 0.06
Matter Power Spectrum Measuring projected matter power spectrum to cosmic variance limit across whole linear regime 0.002< k < 0.2 h/mpc Linear deflection power 10 7 10 8 Reference Planck 10 100 1000 Hu & Okamoto (2001) [non-linear: see Amblard, Vale & White (2004)] L σ(w) 0.06; 0.14
Summary CMB polarization generated by scattering alone and hence provides probes that are well localized in time and space Early reionization provides a new window not only on the first generation of structure but also on gravitational waves, statistical anomalies on large scales, and calibration of fluctuations for dark energy studies Acoustic polarization can provide exceedingly precise measurements of the initial power spectrum and any features that might exist in the decade of the peaks Lensing of the acoustic polarization provides a means of reconstructing the mass distribution and hence constrain the neutrino mass and the dark energy