Probing Inflation and Reionization with Large-Scale CMB Polarization Vinícius Miranda Department of Physics and Astronomy University of Pennsylvania Berkeley, December 4 2017 - Vinícius Miranda
Probing Inflation and Reionization with Large-Scale CMB Polarization B-MODE from Space - Berkeley Vinicius Miranda, Postdoctoral Researcher, University of Pennsylvania Part I - The Epoch Of Reionization First stars: source of ionizing radiation Difficult to model: radiative transfer + big volume One of the least understood aspects of cosmology December 4, 2017
Probing Inflation and Reionization with Large-Scale CMB Polarization B-MODE from Space - Berkeley Vinicius Miranda, Postdoctoral Researcher, University of Pennsylvania How Can We Probe The EOR? 21 cm Cosmic Microwave Background Measure more than total optical depth Better than Cosmic Variance Limitations: Cosmic Variance Atmosphere limits redshift range December 4, 2017
Probing Inflation and Reionization with Large-Scale CMB Polarization B-MODE from Space - Berkeley Vinicius Miranda, Postdoctoral Researcher, University of Pennsylvania Hypothesis: The Instantaneous Reionization Model Higher optical depth implies transition at higher redshift Helium dz (1 + z ) 2 z 2 ∫ τ ( z 1 , z 2 ) ∝ x e ( z ) H ( z ) / H 0 z 1 The only free parameter given tau is the transition width. December 4, 2017
Probing Inflation and Reionization with Large-Scale CMB Polarization B-MODE from Space - Berkeley Vinicius Miranda, Postdoctoral Researcher, University of Pennsylvania Hypothesis: The Instantaneous Reionization Model Low and high redshift behavior are linked together higher redshift => higher l December 4, 2017
Probing Inflation and Reionization with Large-Scale CMB Polarization B-MODE from Space - Berkeley Vinicius Miranda, Postdoctoral Researcher, University of Pennsylvania Beyond IRM: Principal Components Analysis 4 Eigenvectors covariance 3 2 a = 1 a = 3 a = 4 a = 5 ∑ S a ( z ) 1 x e ( z ) = x e fid ( z ) + m a S a ( z ) 0 − 1 a − 2 a = 2 0 . 3 Lower PCs: better constrained a = 1 0 . 2 τ ( z, z max ) 0 . 1 a = 2 a=1 ~ average optical depth a = 3 a = 4 a = 5 0 . 0 a=2 ~ Difference high-low z − 0 . 1 10 15 20 25 30 z December 4, 2017
Probing Inflation and Reionization with Large-Scale CMB Polarization B-MODE from Space - Berkeley Vinicius Miranda, Postdoctoral Researcher, University of Pennsylvania PCA: Completeness 5 PCs are complete! 0 . 08 tanh ML 0 . 07 Complete in polarization tanh ML PC PC ML 0 . 06 / 2 π [ µ K 2 ] error < cosmic variance 0 . 05 0 . 04 l ( l + 1) C EE l 0 . 03 0 . 02 0 . 01 0 . 00 5 10 15 20 25 30 35 40 l December 4, 2017
Probing Inflation and Reionization with Large-Scale CMB Polarization B-MODE from Space - Berkeley Vinicius Miranda, Postdoctoral Researcher, University of Pennsylvania PCA Results On The EOR V . Miranda Wayne Hu Chen He @ JPL @ Job market Has tenure December 4, 2017
Probing Inflation and Reionization with Large-Scale CMB Polarization B-MODE from Space - Berkeley Vinicius Miranda, Postdoctoral Researcher, University of Pennsylvania PCA Results On The EOR 0 . 3 Instantaneous reionization unphysical 0 . 2 Is IRM favored by the data? 0 . 1 ∑ x e ( z ) = x e fid ( z ) + 0 . 0 m a S a ( z ) m 2 a − 0 . 1 − 0 . 2 σ ( m 1 ) < σ ( m 2 ) < σ ( m 3 ) < ... − 0 . 3 Gaussian − 0 . 4 − 0 . 1 0 . 0 0 . 1 0 . 2 m 1 December 4, 2017
Probing Inflation and Reionization with Large-Scale CMB Polarization B-MODE from Space - Berkeley Vinicius Miranda, Postdoctoral Researcher, University of Pennsylvania PCA Results On The EOR 0 . 14 tanh ML PC 0 . 12 Does the IRM spuriously PC mean 0 . 10 PC 68 , 95% CL 0 . 08 disfavor high redshift τ ( z, z max ) 0 . 06 sources? 0 . 04 0 . 02 0 . 00 − 0 . 02 Planck 2015 τ PC ( z = 15, z max ) = 0.033 − 0 . 04 0 5 10 15 20 25 30 z December 4, 2017
Probing Inflation and Reionization with Large-Scale CMB Polarization B-MODE from Space - Berkeley Vinicius Miranda, Postdoctoral Researcher, University of Pennsylvania PCA Results On The EOR 0 . 08 What the polarization tanh ML 0 . 07 tanh ML PC spectrum look like? PC ML 0 . 06 / 2 π [ µ K 2 ] 0 . 05 0 . 04 l ( l + 1) C EE l Presence of high redshift 0 . 03 0 . 02 sources does NOT imply 0 . 01 unreasonable tau 0 . 00 5 10 15 20 25 30 35 40 l December 4, 2017
Probing Inflation and Reionization with Large-Scale CMB Polarization B-MODE from Space - Berkeley Vinicius Miranda, Postdoctoral Researcher, University of Pennsylvania PCA: Completeness Forward Modeling Only Can models with metal-free stars be the source of the high redshift signal? V . Miranda Wayne Hu Chen He Adam Lidz December 4, 2017
Probing Inflation and Reionization with Large-Scale CMB Polarization B-MODE from Space - Berkeley Vinicius Miranda, Postdoctoral Researcher, University of Pennsylvania Source Of High Redshift Ionization: Pop-III? Pop-II We solve the following ODE Pop-III Pop-III, self-regulated Tanh d 〈 x i 〉 ) − 〈 x i 〉 ( = d dt ζ II f c , II + ζ III f c , III z end = 6 dt t rec ( z ) Efficiency 〈 x T 〉 ≠ 〈 x II 〉 + 〈 x III 〉 Baseline model (Pop-II stars) December 4, 2017
Probing Inflation and Reionization with Large-Scale CMB Polarization B-MODE from Space - Berkeley Vinicius Miranda, Postdoctoral Researcher, University of Pennsylvania Source Of High Redshift Ionization: Pop-III? Pop-II PC Projection Pop-III Pop-III, self-regulated Tanh P15 68%, 95% CL m a = 1 30 ∫ zS a ( z ) 〈 x e 〉 ( z ) − 〈 x e ⎡ fix 〉 ( z ) ⎤ d ⎣ ⎦ 24 6 Then evaluate optical depth ∑ τ ( z 1 , z 2 ) = τ fid ( z 1 , z 2 ) + m a τ a ( z 1 , z 2 ) December 4, 2017
Probing Inflation and Reionization with Large-Scale CMB Polarization B-MODE from Space - Berkeley Vinicius Miranda, Postdoctoral Researcher, University of Pennsylvania Source Of High Redshift Ionization: Pop-III? Pop-II Get polarization spectrum Pop-III Pop-III, self-regulated Tanh Tanh 68%, 95% CV Shades = CV assuming instantaneous ionization is the correct model* *Tricky type of question December 4, 2017
Probing Inflation and Reionization with Large-Scale CMB Polarization B-MODE from Space - Berkeley Vinicius Miranda, Postdoctoral Researcher, University of Pennsylvania Make Results Useful to Everyone: PCA Fast Likelihood tanh N ∑ w i K f ( m − m i ) PC (data | m ) = L tanh L PC Gaussian i = 1 P ( τ | data) Y our Favorite model MCMC models 0 . 03 0 . 06 0 . 09 0 . 12 τ (tanh model) December 4, 2017
Probing Inflation and Reionization with Large-Scale CMB Polarization B-MODE from Space - Berkeley Vinicius Miranda, Postdoctoral Researcher, University of Pennsylvania Make Results Useful to Everyone: PCA Fast Likelihood tanh N ∑ w i K f ( m − m i ) PC (data | m ) = L tanh L PC Gaussian i = 1 P ( τ | data) Chain multiplicities Kernel (Gaussian) 0 . 03 0 . 06 0 . 09 0 . 12 τ (tanh model) December 4, 2017
Probing Inflation and Reionization with Large-Scale CMB Polarization B-MODE from Space - Berkeley Vinicius Miranda, Postdoctoral Researcher, University of Pennsylvania Make Results Useful to Everyone: PCA Fast Likelihood tanh N ∑ w i K f ( m − m i ) PC (data | m ) = L tanh L PC Gaussian i = 1 Good: fast (no CAMB) P ( τ | data) Bad: needed ~10x more chain points than normal convergence 0 . 03 0 . 06 0 . 09 0 . 12 τ (tanh model) December 4, 2017
Probing Inflation and Reionization with Large-Scale CMB Polarization B-MODE from Space - Berkeley Vinicius Miranda, Postdoctoral Researcher, University of Pennsylvania What Our Results Means for 21-cm, neutrinos, CMB-S4… τ PC ( z = 15, z max ) = 0.033 • CMB-S4: neutrino mass constraints with lensing • Optical Depth is one of the highest sources of error • 21-cm claims they can measure tau better than CV • This claim will fail if our result is not due to systematics December 4, 2017
Probing Inflation and Reionization with Large-Scale CMB Polarization B-MODE from Space - Berkeley Vinicius Miranda, Postdoctoral Researcher, University of Pennsylvania Part II - Inflationary Features Georges Obied Cora Dvorkin V . Miranda Wayne Hu Chen He December 4, 2017
Probing Inflation and Reionization with Large-Scale CMB Polarization B-MODE from Space - Berkeley Vinicius Miranda, Postdoctoral Researcher, University of Pennsylvania The Generalized Slow-Roll Approximation 2 = I 0 ( k ) + ln[1 + I 1 ln Δ R 2 ( k )] ∫ I j ( k ) ∝ d ln s W j ( ks ) ′ G (ln s ) Single kernel encompasses power spectrum observables Wayne Hu Cora Dvorkin December 4, 2017
Probing Inflation and Reionization with Large-Scale CMB Polarization B-MODE from Space - Berkeley Vinicius Miranda, Postdoctoral Researcher, University of Pennsylvania Kernel Expansion: Non-Parametric Spline Basis ∑ G '(ln s ) = (1 − n s ) + B i (ln s ) w i i SB: more efficient than PCAs for localized features V . Miranda C. Dvorkin W. Hu December 4, 2017
Probing Inflation and Reionization with Large-Scale CMB Polarization B-MODE from Space - Berkeley Vinicius Miranda, Postdoctoral Researcher, University of Pennsylvania WMAP/Planck l~20 Features on Temperature Spectrum 1 . 5 1 . 0 0 . 5 / σ ℓ 0 . 0 ∆ C TT ℓ − 0 . 5 − 1 . 0 − 1 . 5 Λ CDM TT ( ℓ < 1000) 5 SB TT − 2 ∆ ln L TT 0 − 5 − 10 − 15 10 1 10 2 10 3 ℓ December 4, 2017
Recommend
More recommend