The POLARBEAR experiment probing the cosmic microwave background polarization
Davide Poletti (APC, Paris) on behalf of the POLARBEAR Collaboration
14 Dec 2015 ! 28th Texas Symposium on Relativistic Astrophysics POLARBEAR Collaboration UC Berkeley UC San Diego KEK McGill University SISSA Brian Barch Chris Aleman Yoshiki Akiba Matt Dobbs Carlo Baccigalupi Yuji Chinone Kam Arnold Takaho Hamada Adam Gilbert Giulio Fabbian Ari Cukierman Matt Atlas Masaya Hasegawa Josh Montgomery Giuseppe Puglisi Tijmen de Haan Darcy Barron Kaori Hattori Graeme Smecher Josquin Errard Tucker Elleflot Masashi Hazumi JAXA Neil Goeckner-Wald George Fuller Yuki Inoue Dalhousie Tomotake Matsumura John Groh Logan Howe Haruki Nishino Scott Chapman UC Irvine Grantland Hall Jon Kaufman Yuuko Segawa Colin Ross Chang Feng Charles Hill Kavon Kazemzadeh Jun-ichi Suzuki Kaja Rotermund William Holzapfel Brian Keating Osamu Tajima Alexei Tikhomirov National Institute Yasuto Hori David Leon Satoru Takakura for Fusion Science Oliver Jeong Lindsay Lowry Sayuri Takatori Lawrence Suguru Takada Adrian Lee Frederick Matsuda Takayuki Tomaru Berkeley NL Mike Myers Martin Navaroli Julian Borrill Reijo Keskitalo Cardiff University Chris Raum Hans Paar Laboratoire Peter Ade Paul Richards Gabriel Rebeiz Astroparticule & Theodore Kisner Blake Sherwin Praween Siritanasak Cosmologie Akito Kusaka NASA Goddard Ian Shirley Nathan Stebor Maude Le Jeune Eric Linder Nathan Miller Julien Peloton Bryan Steinbach Brandon Wilson Kavli IPMU Aritoki Suzuki Amit Yadav Davide Poletti Princeton Radek Stompor Takuro Fujino Zigmund Kermish Nathan Whitehorn Alex Zahn Fumiya Irie Oliver Zahn Nobuhiko Katayama CU Boulder Argonne NL Católica (PUC) Kuniyoshi Mizukami Imperial College Nils Halverson Amy Bender David Boettger Anne Ducout Greg Jaehnig Tetsu Yamashita Rolando Dunner Stephen Feeney David Schenck U. Melbourne Andrew Jaffe Christian Reichardt And many more in years past ✦ B-modes science and measurements
✦ The POLARBEAR experiment
✦ First season’s results
✦ POLARBEAR 2 and Simons array
Davide Poletti - POLARBEAR collaboration 3 The Cosmic Microwave Background
The CMB: • picture of the 380,000 years old universe • information on the inflationary phase Adapted from NASA/WMAP
Temperature anisotropies: • ΛCDM confirmed • Cosmological information exploited
Planck Collaboration
Davide Poletti - POLARBEAR collaboration 4 The CMB polarization
Polarization field ➱ Q, U Stokes parameters ! y Q < 0 ! U > 0 x ! ! Q > 0 ! U < 0 ! ➱ Decomposed in E and B modes
Davide Poletti - POLARBEAR collaboration 5 Primordial B-modes
Perturbations at the last scattering surface: • Scalar ➱ E only (to linear order) • Tensor ➱ E and B
Tensor perturbation after inflation: • r Energy scale of inflation ( ~1016 GeV for r~0.1) • nt Consistency relation (r = - 8 nt)
Davide Poletti - POLARBEAR collaboration 6 Lensing B-Modes
ESA/Planck
Acquaviva and Baccigalupi (2006) D D (nˆ)= 2 dD s (Dnˆ,D) DDs Different dark! Z energy scenarios d = r
E(l; l0)= [E(l0) cos 2' B(l0)sin2' ][l (l l0)] (l l0) l0l l0l ·
B(l; l0)= [E(l0)sin2' + B(l0) cos 2' ][l (l l0)] (l l0) l0l l0l ·
Davide Poletti - POLARBEAR collaboration 7 Lensing B-Modes
ESA/Planck
Acquaviva and Baccigalupi (2006) D D (nˆ)= 2 dD s (Dnˆ,D) DDs Different dark! Z energy scenarios d = r
E(l; l0)= [E(l0) cos 2' B(l0)sin2' ][l (l l0)] (l l0) l0l l0l ·
B(l; l0)= [E(l0)sin2' + B(l0) cos 2' ][l (l l0)] (l l0) l0l l0l ·
Davide Poletti - POLARBEAR collaboration 8 Measurements of the BB power spectrum
Mar 2014 2015
102 DASI WMAP-9yr CBI QUaD MAXIPOL QUIET-Q BOOMERanG QUIET-W CAPMAP BICEP1-3yr 1 ) 10 2 K µ ) 2 K )(
! 0
10 ) ( (2 / /(2 BB l BB ` -1 C 10 l(l+1)C + 1) ⇥ ( ⇥ 10-2
r=0.2 r=0.025 10-3 10 100 1000 Multipole Moment, ell Davide Poletti - POLARBEAR collaboration 9 Measurements of the BB power spectrum
Mar 2014 2015
POLARBEAR 10 Mar 102 DASI QUaD CBI QUIET-Q MAXIPOL QUIET-W BOOMERanG BICEP1-3yr CAPMAP POLARBEAR 1 WMAP-9yr ) 10 2 K µ ) 2 K )(
! 0
10 ) ( (2 / /(2 BB l BB ` -1 C 10 l(l+1)C
+ 1) POLARBEAR ⇥ ( ⇥ 10-2
r=0.2 r=0.025 10-3 10 100 1000 Multipole Moment, ell Davide Poletti - POLARBEAR collaboration 10 Measurements of the BB power spectrum
Mar 2014 2015
BICEP2 17 Mar 102 DASI QUaD CBI QUIET-Q MAXIPOL QUIET-W BOOMERanG BICEP1-3yr CAPMAP BICEP2-3yr 1 WMAP-9yr POLARBEAR ) 10 2 K µ ) 2 K )(
! 0
10 ) ( (2 / /(2 BB l BB ` -1 C 10 l(l+1)C
+ 1) BICEP2 POLARBEAR ⇥ ( ⇥ 10-2
r=0.2 r=0.025 10-3 10 100 1000 Multipole Moment, ell Davide Poletti - POLARBEAR collaboration 11 Measurements of the BB power spectrum
Mar 2014 2015
ACTPol 21 May 102 DASI QUIET-Q CBI QUIET-W MAXIPOL BICEP1-3yr BOOMERanG ACTPol 1 CAPMAP BICEP2-3yr 10 WMAP-9yr POLARBEAR ACTPol ) QUaD 2 K µ ) 0 2 10 K )( ! ) ( (2
/ -1 /(2 10 BB l BB `
C BICEP2 POLARBEAR -2
l(l+1)C 10 r=0.20 + 1) ⇥ ( ⇥ 10-3 r=0.025
10-4 10 100 1000 Multipole Moment, ell Davide Poletti - POLARBEAR collaboration 12 Measurements of the BB power spectrum
Mar 2014 2015
BICEP2 - Keck Array - Planck 2 Feb 102 DASI QUIET-Q CBI QUIET-W MAXIPOL BICEP1-3yr BOOMERanG ACTPol 1 CAPMAP BKP 10 WMAP-9yr POLARBEAR ACTPol ) QUaD 2 K µ ) 0 2 10 K )( ! ) ( (2
/ -1 /(2 10 BB l BB `
C POLARBEAR -2 BKP
l(l+1)C 10 + 1) ⇥ ( ⇥ 10-3
10-4 10 100 1000 Multipole Moment, ell Davide Poletti - POLARBEAR collaboration 13 Measurements of the BB power spectrum
Mar 2014 2015
SPTPol 8 Mar 102 DASI QUIET-Q CBI QUIET-W MAXIPOL BICEP1-3yr BOOMERanG ACTPol 1 CAPMAP BKP 10 WMAP-9yr SPTpol ACTPol ) QUaD POLARBEAR 2 K µ ) 0 2 10 K )( ! ) ( (2
/ -1 /(2 10 SPTpol BB l BB `
C POLARBEAR -2 BKP
l(l+1)C 10 + 1) ⇥
( r=0.09 ⇥ 10-3
10-4 10 100 1000 Multipole Moment, ell Davide Poletti - POLARBEAR collaboration 14 Measurements of the BB power spectrum
Mar 2014 2015
BICEP2 / Keck 30 Oct 102 DASI QUIET-Q CBI QUIET-W MAXIPOL BICEP1-3yr BOOMERanG ACTPol 1 CAPMAP BK14 10 WMAP-9yr SPTpol ) QUaD POLARBEAR 2
K ACTPol µ ) 0 2 10 K )( ! ) ( (2
/ -1 /(2 10 SPTpol BB l BB `
C POLARBEAR -2 BK14
l(l+1)C 10 + 1) ⇥ (
⇥ r=0.07 10-3
10-4 10 100 1000 Multipole Moment, ell Davide Poletti - POLARBEAR collaboration 15 Measurements of the BB power spectrum
Mar 2014 2015
BICEP2 / Keck 30 Oct 102 DASI QUIET-Q CBI QUIET-W MAXIPOL BICEP1-3yr BOOMERanG ACTPol 1 CAPMAP BK14 10 WMAP-9yr SPTpol ) QUaD POLARBEAR 2
K ACTPol µ ) 0 2 10 K )( ! ) ( (2
/ -1 /(2 10 SPTpol BB l BB `
Plus measurementsC through! POLARBEAR cross-correlation!-2 ! BK14 l(l+1)C 10 • SPTPol+ 1) ! ⇥ (
• POLARBEAR⇥ ! r=0.07 -3 • ACTPol! 10 • Planck 10-4 10 100 1000 Multipole Moment, ell Davide Poletti - POLARBEAR collaboration 16 ✦ B-modes science and measurements
✦ The POLARBEAR experiment
✦ First season’s results
✦ POLARBEAR 2 and Simons array
Davide Poletti - POLARBEAR collaboration 17 POLARBEAR experiment
• CMB B-modes dedicated experiment • Atacama desert (~5200 m altitude) ‣ Access to 80% of the sky ‣ Dry atmosphere ‣ Targeting both primordial and lensing B-modes
Crab Nebula Planck 857GHz (TauA) FIRST SEASON polarization angles calibrator • Period: May 2012 to June 2013
PB1-RA23 HA • Target: PB1-RA12 HA Overlap w/ QUIET, deep integration of Overlap w/ Herschel Herschel Atlas PB1-RA4p5 3 patches 5 deg x 5 deg Overlap w/ QUIET, BOSS
Davide Poletti - POLARBEAR collaboration 18 ⇣.✏⇧⌃⌦↵ ↵⌅ ↵ ⇤ ⌥ ↵⌃↵ ⇥ ⌃ ⌃ 197 IA: azimuth encoder zero-point
IE: elevation encoder zero-point
TF: telescope flexure
that describes the mechanical non-idealities affecting the telescope pointing. The parameters are reconstructed through a linear model assuming similar to the standard mapmaking equation
az d = As + n (9.1) el ⇥ where the vector s contains the parameter to be estimated and the noise term is such that its noise correlation is the identity matrix
nT n = 1.(9.2)
This is equivalent to the assumption that the source does not move during Instrumentalthe design raster scan. Details on the role of each parameter and on the pointing matrix A are discussed in more detail in Errard (2012). The rms residual with between the telescope pointing after the pointing model correction is applied has an amplitude of 17arcsec in both azimuth and elevation, lead- ing to a total rms error of 25arcsec.
Measurements of the product of the integrated bandwidth and fractional Array performance throughput, ⇥ ⇤ were also made from the beam maps observation previ- ously discussed, as well as from earlier test of the receiver in the lab and from elevation nods of the telescope. The fractional throughput is a mea- sure of the percentage of the power seen by a detector from a source at the input of the receiver compared to what would be seen if the detector had
perfect efficiency to2.5 Meters that same source. The overall loss is due to expected efficiencies of the detectors and optical componentsPrimary: along the 3.5m optical path. Laboratory measurements of ⇥ ⇤ are consistent with a ⇥ = 37% given the measured 37GHz integrated bandwidth. Measurements of ⇥ ⇤ made in the field are consistent with these values. The design bolometer noise equivalent temperatures (NET) of table 9.1 are slightly degraded due to bolometer saturation powers and atmospheric con- ditions. Once the relative gain calibration for a pair of detector is available we can compute the timestream noise in polarization and temperature for the Q or U stokes parameters in individual pixels. Since the two orthogonal antennas in a focal plane pixelHuan (from now Tran on referredTelescope to as top (t) and bottom (b) ) measure two orthogonal polarization, their timestream model is
t Hex Module d (t)=gtop [I(nˆ (t)) + Q(nˆ (t)) cos(2⌅(t)) + U(nˆ (t)) sin(2⌅(t))] b d (t)=gbot [I(nˆ (t)) Q(nˆ (t)) cos(2⌅(t)) U(nˆ (t)) sin(2⌅(t))] (9.3) 6mm lenslet Fourier transforming the semi-differenceMicrostrip of the top and bottom detectors Antenna allows us to see how well the unpolarizedFilter atmosphere is suppressed at low frequencies. The fknee of this 1/f frequency noise gives us an estimates 1274 bolometers @ 150 GHz of the NET sensitivity of theTES detectors in polarization assuming the atmo- sphere is unpolarized andbolometer the contribution of the residual atmosphere to the Cooled to 250 mK measured timestream difference1 mm is low. Fig 9.10 shows the sum and differ- 8cm ence of the top and bottom timestreams power spectrum where we see can
Davide Poletti - POLARBEAR collaboration 19 Instrument characterization
• Ground based and astrophysical calibrators ‣ Beam: Jupiter ‣ Calibration of the detectors: Saturn ‣ Polarization angle: Tau A • 3.5 arcmin beam FWHM • Ellipticity < 5%, differential ellipticity 1% • Array NET 23 µKps
Davide Poletti - POLARBEAR collaboration 20 Map and power spectrum
Planck SMICA map
POLARBEAR preliminary Filtered mapmaking:! • 1 1 1 Unbiased mapmaking: | | ˆs =(A>N A) A>Fd • ˆs =(A FA) A Fd • Very fast! • Fast, iterative • Accurate, explicit estimator estimator • Flat-sky MASTER pseudo ! power spectrum estimation • Curved-sky, pure, pseudo power spectrum with daily cross-spectra estimator
Cross-check and validation
21
8 ✦ B-modes science and measurements
✦ The POLARBEAR experiment
✦ First season’s results
✦ POLARBEAR 2 and Simons array
Davide Poletti - POLARBEAR collaboration 22 Results: lensing from polarization alone
Polarization! lensing! d estimation dd Measurement C`
Phys. Rev. Lett. 112, 131302 (2014) Editors’ Suggestion
4.2σ B-modes evidence
Davide Poletti - POLARBEAR collaboration 23 Results: cross-correlation with CIB 3
Polarization! Cosmic Infrared Eq. 5). The maximum ℓ of included temperature modes is the only difference between the lensing maps used in Measurement Background this work and those in [12]. The final normalization is obtained in a two step pro- cess, as in [12]. A first-order approximation for the Hand et al. 2013 normalization is computed from the data power spec- trum, with an additional, small correction factor (of or- 1 1 Tracer of = d = 2 der 10%) applied from Monte Carlo simulations, which 2r· 2r density field are designed to match both the signal and noise prop- erties of the ACT data. Finally, we obtain a simulated mean field map ⟨κˆ⟩ from 480 Monte Carlo realizations of Estimator of κ reconstructed CMB lensing convergence maps and sub- CIB map from! tract this mean field from the reconstructed ACT lensing κ from POLARBEAR gal maps. The simulated mean field is non-zero due to noise X Hershel FIG. 1. The lensing kernel W (solid) for the CS82 red- polarization maps shift distributionPhys. Rev. of Lett. source 112, galaxies 131302 (as given (2014) in Eq. 6) and and finite-map effects giving rise to a small (∼5%) ar- normalizedEditors’ to a unitSuggestion maximum. For comparison, the kernel tificial lensing signal, which must be subtracted. Note for CMB lensing (Eq. 3) is shown as dashed, also normalized that this set of 480 Monte Carlo realizations is also used to a unit maximum. to estimate error bars on the final cross power spectrum measurement, as described in section V.
κCMB 2.5σ B-modes evidence and W is z 4.0∼ 0σ.9, polarized illustrating lensing that the cross power spectrum is sensitive to the amplitude of structure at in- B. CS82 Lensing Data termediate redshifts. 1. Data III. CMB AND GALAXY LENSING DATA The Canada-France-Hawaii Telescope Stripe 82 Survey ′ A. ACT CMB Lensing Data is an i -band survey of the so-called Stripe 82 region of sky along the celestial equator [41]. The survey was de- signed with the goal of covering a large fraction of Stripe ACT is a 6-meter telescope located in the Atacama 82 with high quality i′-band imaging suitable for weak Davide Poletti - POLARBEAR collaboration desert in Chile [36–38]. The CMB temperature maps24 lensing measurements. With this goal in mind, the CS82 used in this work are made from observations taken dur- survey was conducted under excellent seeing conditions: ing 2008 - 2010 in the 148 GHz frequency channel and the Point Spread Function (PSF) for CS82 varies between have been calibrated to 2% accuracy as in [39]. The maps 0.4′′ and 0.8′′ over the entire survey with a median see- are centered on the celestial equator with a width of 3 ing of 0.6′′. In total, CS82 comprises 173 MegaCam i′- degrees in declination and 108 degrees in right ascension band images, with each image roughly one square degree and are identical to those used in [12]. in area with a pixel size of 0.187 arcseconds. The area The lensing convergence fields are reconstructed from covered by the survey is 160 degrees2 (129.2 degrees2 af- the CMB temperature maps using the minimum variance ter masking out bright stars and other artifacts). The quadratic estimator of [40] following the procedure used completeness magnitude is i′ ∼ 24.1 (AB magnitude, 5σ in [27]. The lensing deflection induces correlations in the in a 2′′ aperture). Image processing is largely based on Fourier modes of the previously uncorrelated, unlensed the procedures presented in [42, 43]. Weak lensing shear CMB. The lensing convergence is estimated from these catalogs were constructed using the state-of-the-art weak Fourier correlations with a quadratic estimator: lensing pipeline developed by the CFHTLenS collabora- tion which employs the lensfit shape measurement algo- rithm [44, 45]. We refer to these publications for more L L 2l L L l κˆ( )=N( ) d f( ,l)T (l)T ( − ), (5) in-depth details on the shear measurement pipeline. ! Following [44] and [45], source galaxies are selected to where l and L are Fourier space coordinates, N is the have w>0 and FITSCLASS = 0. Here, w represents an normalization function, T is the temperature field, and inverse variance weight accorded to each source galaxy by f is a weighting function that maximizes the signal-to- lensfit, and FITSCLASS is a flag to remove stars but also noise ratio of the reconstructed convergence (see [40] for to select galaxies with well-measured shapes (see details details). In the lensing reconstruction, we filter out tem- in [44]). After these cuts, the CS82 source galaxy den- perature modes with a low signal-to-noise ratio, specif- sity is 15.8 galaxies arcmin−2 and the effective weighted ically those modes below ℓ =500andaboveℓ =4000. galaxy number density (see equation 1 in [45]) is 12.3 This filtering does not prevent the measurement of low- galaxies arcmin−2. Note that these numbers do not in- ℓ lensing modes, as the lensing signal at a given scale ℓ clude any cuts on photometric redshift quality since for is obtained from temperature modes separated by ℓ (see the purposes of this paper, we only need to know the Draft version March 11, 2014 Preprint typeset using LATEXstyleemulateapjv.12/16/11
A MEASUREMENT OF THE COSMIC MICROWAVE BACKGROUND B-MODE POLARIZATION POWER SPECTRUM AT SUB-DEGREE SCALES WITH POLARBEAR The Polarbear Collaboration: P.A.R. Ade29, Y. Akiba32,A.E.Anthony2,5, K. Arnold14,M.Atlas14,D. Barron14,D.Boettger14,J.Borrill3,31,S.Chapman9,Y.Chinone17,13, M. Dobbs25,T.Elleflot14,J.Errard31,3, G. Fabbian1,18,C.Feng14,D.Flanigan13,10, A. Gilbert25,W.Grainger28,N.W.Halverson2,5,15,M.Hasegawa17,32, K. Hattori17,M.Hazumi17,32,20,W.L.Holzapfel13,Y.Hori17,J.Howard13,16,P.Hyland24,Y.Inoue32,G.C. Jaehnig2,15,A.H.Jaffe11, B. Keating14, Z. Kermish12, R. Keskitalo3, T. Kisner3,31,M.LeJeune1,A.T.Lee13,27, E.M. Leitch4,19,E.Linder27,M.Lungu13,8,F.Matsuda14,T.Matsumura17,X.Meng13,N.J.Miller22,H.Morii17, S. Moyerman14,M.J.Myers13,M.Navaroli14,H.Nishino20,H.Paar14,J.Peloton1,D.Poletti1,E.Quealy13,26, G. Rebeiz6, C.L. Reichardt13, P.L. Richards13,C.Ross9,I.Schanning14,D.E.Schenck2,5,B.D.Sherwin13,21,A. Shimizu32,C.Shimmin13,7,M.Shimon30,14,P.Siritanasak14,G.Smecher33,H.Spieler27, N. Stebor14,B. Steinbach13, R. Stompor1,A.Suzuki13,S.Takakura23,17,T.Tomaru17,B.Wilson14,A.Yadav14,O.Zahn27 Draft version March 11, 2014
ABSTRACT We report a measurement of the B-mode polarization power spectrum in the cosmic microwave background (CMB) using the Polarbear experiment in Chile. The faint B-mode polarization signa- ture carries information about the Universe’s entireResults: history of gravitational BB spectrum structure formation, andmeasurement the cosmic inflation that may have occurred in the very early Universe. Our measurement covers the 2 angular multipole range 500 < ⇥ < 2100 and is based on observations of 30 deg with 3.5⇥ resolution 97.5% c.l. B-modes at 150 GHz. On these angular scales, gravitationalFirst direct lensing evidence of the CMB of lensing by intervening B-modes structure in the Universe is expected to be the dominant• source of B-mode polarization. Including both system- evidence atic and statistical uncertainties, the hypothesis of no B-mode polarization power from gravitational lensing is rejected at 97.5% confidence – the equivalentAmplitude of of 2.0 lensing for a compared normal distribution. to ΛCDM The band Astrophysical J. 794, 171 (2014) powers are consistent with the standard cosmological• model. Fitting a single lensing amplitude pa- +0.04 rameter ABB to the measured band powers, ABB =1.12 0.61(stat) 0.10(sys) 0.07(multi), where ! ± ± ABB = 1 is the fiducial wmap-9 CDM value. In this expression, “stat” refers to20 the statistical uncertainty, “sys” to the systematic uncertainty associated with possible biases from the instrument and astrophysical foregrounds, and “multi” to the calibration uncertainties that have a multiplicative Negligible contamination e⇥ect on the measured amplitude ABB. • ) 0.8 2
Subject headings: Cosmic Microwave Backgroundfrom astrophysical foregrounds K
µ 0.6 ) 2 )( K !
0.4
) ( Negligible contamination • (2 / 1 AstroParticule et Cosmologie, Univ Paris Diderot,from systematic effects /(2 0.2 BB l BB CNRS/IN2P3, CEA/Irfu, Obs de Paris, Sorbonne Paris 17 High Energy Accelerator Research Organization (KEK), Cit´e,France Tsukuba, Ibaraki 305-0801, Japan C 0 2 Center for Astrophysics and Space Astronomy, University 18 International School for Advancedl(l+1)C Studies (SISSA), Trieste -0.2 of Colorado, Boulder, CO 80309, USA 34014, Italy + 1) ⇥
3 Computational Cosmology Center, Lawrence Berkeley 19 Kavli Institute for Cosmological( Physics, University of ⇥ -0.4 National Laboratory, Berkeley, CA 94720, USA Chicago, Chicago, IL 60637, USA 4 Department of Astronomy and Astrophysics, University of 20 Kavli Institute for the Physics and Mathematics of the Chicago, Chicago, IL 60637, USA Universe (WPI), Todai Institutes for Advanced 0 Study, 500 The 1000 1500 2000 2500 5 Department of Astrophysical and Planetary Sciences, University of Tokyo, Kashiwa, Chiba 277-8583, Japan Multipole Moment, ell 21 University of Colorado, Boulder, CO 80309, USA Miller Institute for Basic Research inBB Science, University 2 Fig. 12.— Binned C spectrum measured using data from all three patches ( 30 deg ). A theoretical wmap-9 ⇥CDM high-resolution 6 BB ⇥ C spectrum with ABB=1isshown.Theuncertaintyshownforthebandpowersisthediagonalofthebandpowercovariancematrix, Department of Electrical and Computer Engineering, Uni- of California, Berkeley, CA 94720,including USA beam covariance. versity of California, San Diego, CA 92093, USA 22 Observational Cosmology Laboratory, Code 665, NASA Davide Poletti - POLARBEAR collaboration process to measure the upper bound. The multiplicative25 7 TABLE 8 Department of Physics and Astronomy, University of Goddard Space Flight Center, Greenbelt,Reported Polarbear MD band 20771, powers and USA the diagonal uncertainties are the quadrature sum of all the multi- 23 elements of their covariance matrix plicative uncertainties discussed in Section 7. The measurement rejects the hypothesis of no CBB California, Irvine, CA 92697-4575, USA Osaka University, Toyonaka, OsakaCentral ⇥⇥( 560-0043,⇥ +1)CBB/2 [µK2] Japan ⇥ (⇥ +1)CBB/2 [µK2] 8 24 { } from lensing with a confidence of 97.5%. This is calcu- Department of Physics and Astronomy, University of Physics Department, Austin College,700 Sherman, 0.093 TX 0. 75090,056 lated using the bias-subtracted band powers described 1100 0.149 0.117 1500 0.317 0.236 above (the most conservative values to use for rejecting Pennsylvania, Philadelphia, PA 19104-6396, USA USA 1900 0.487 0.482 this null hypothesis), and integrating the likelihood of 9 25 ABB> 0. This significance is the equivalent of 2.0 for a Department of Physics and Atmospheric Science, Dalhousie Physics Department, McGill University, Montreal, QC normal distribution. University, Halifax, NS, B3H 4R2, Canada H3A 0G4, Canada trum; including statistical uncertainty and beam covari- 10 Department of Physics, Columbia University, New York, 26 Physics Department, Napaance, Valley this PTE isCollege, 42%. Table 8 enumerates Napa, the CA band 9. SUMMARY & DISCUSSION powers reported here. We have reported a measurement of the CMB’s B- We fit the band powers to a CDM cosmological BB NY 10027, USA 94558, USA mode angular power spectrum, C , over the multipole model with a single ABB amplitude parameter. We find range 500 < ⇥ < 2100. This measurement is enabled by 11 27 +0.04 Department of Physics, Imperial College London, London Physics Division, Lawrence BerkeleyABB =1.12 0 National.61(stat) 0.10(sys) Laboratory,0.07(multi), where the unprecedented combination of high angular resolu- ± ± ABB =1isdefinedbythewmap-9 CDM spectrum. tion (3.5⇥) and low noise that characterizes the Polar- SW7 2AZ, United Kingdom Berkeley, CA 94720, USA To calculate the lower bound on the additive uncertain- bear CMB polarization observations. 12 28 ties on this number, we linearly add, in each band, the To validate the Polarbear measurement of this faint Department of Physics, Princeton University, Princeton, Rutherford Appleton Laboratory,upper bound STFC, band powers of Swindon, all the additive systematic SN2 signal, we performed extensive tests for systematic er- e⇥ects discussed in Section 7, and the uncertainty in the rors. We evaluated nine null tests and estimated twelve NJ 08544, USA 1SZ, United Kingdom removal of E to B leakage. We then subtract this possi- sources of instrumental contamination using a detailed 13 Department of Physics, University of California, Berkeley, 29 School of Physics and Astronomy,ble bias from the measured Cardi band powers,University, and calculate instrument model, and found that all the systematic un- ABB.ThisproducesalowerABB, and sets the lower certainties were small compared to the statistical uncer- CA 94720, USA Cardi CF10 3XQ, United Kingdombound of the additive uncertainty. We then repeat the tainty in the measurement. To motivate comprehensive 14 Department of Physics, University of California, San Diego, 30 School of Physics and Astronomy, Tel Aviv University, Tel CA 92093-0424, USA Aviv 69978, Israel 15 Department of Physics, University of Colorado, Boulder, 31 Space Sciences Laboratory, University of California, Berke- CO 80309, USA ley, CA 94720, USA 16 Department of Physics, University of Oxford, Oxford OX1 32 The Graduate University for Advanced Studies, Hayama, 2JD, United Kingdom Miura District, Kanagawa 240-0115, Japan Results: cosmic birefringence / primordial magnetic fields
Birefringent universe, polarization angle rotation by α LAST SCATTERING SURFACE
α
Polarization! Estimation! ↵↵ Measurement α field C`
Constraint on cosmic birefringence and primordial magnetic fields < 93 nG (95% c.l.)
Phys. Rev. D 92, 123509 (2015) Editors’ Suggestion
Davide Poletti - POLARBEAR collaboration 26 Control of the systematics
• Blind policy: assess data selection and quality without looking at the scientific products • Null-test (Jackknife test): temporal scan properties (azimuthal direction, elevation) weather calibration properties detector selection bright sources distance. ➱Compatible with uniform distribution ➱No significant outlier (Probability to exceed)
Davide Poletti - POLARBEAR collaboration 27 Control of the systematics
• Systematics pipeline: systematic injection and propagation through the whole science pipeline ➱ Residual systematics are negligible
Statistical uncertainty
Most notably:
B-mode expected Boresight and differential signal pointing Combined systematics Differential beam size Polarization angle Differential beam ellipticity HWP dependent gain
Davide Poletti - POLARBEAR collaboration 28 Situation after season I
• Season I successfully probed small scales
• Systematics: under control
• Sensitivity: close to lensing B-modes level
‣ Analysis of the season II ongoing
• BICEP2 and Planck: foregrounds dominate large scales
Davide Poletti - POLARBEAR collaboration 29 ✦ B-modes science and measurements
✦ The POLARBEAR experiment
✦ First season’s results
✦ POLARBEAR 2 and Simons Array
Davide Poletti - POLARBEAR collaboration 30 The future: POLARBEAR 2 and Simons array
2016: POLARBEAR 2 new telescope and receiver ‣ 7,588 detectors ‣ Multichroic pixels (95/150 GHz)
2017: Simons Array new telescopes, 2 new PB2-like receivers ‣ 22,764 detectors ‣ 95/150/220 GHz channel
Artist Conception 2 95/150&
1x 150/220&
Davide Poletti - POLARBEAR collaboration 31 Simons Array: sensitivity and foreground rejection
Simons Array polarized dust @ 95GHz 90/150/220 GHz ! p=15%, f sky=65% combined with ! Planck and C-Bass r=0.1 3 (r =0.1) = 6 10 3 (4 · 10 ) polarized synchrotron @ 95GHz polarized dust @ 95GHz · (⌃m⌫ ) = 40 meV p=15%, f p=15%, f r=0.01 sky=5% (19 meV) polarized synchrotron @ 95GHz sky=65% Combined with DESI BAO p=15%, f 95% c.l. upper limit on the foreground residual r < 0.07 BK VI (2015) sky=5% Σmν < 0.15 eV Palanque-Delabrouille et al (2015)
*Dust level: Planck Intermediate XXX ➱Constrain inflation, neutrino mass hierarchy, primordial magnetic fields and more...
Davide Poletti - POLARBEAR collaboration 32 Summary
• B-mode era has begun and accuracy is rapidly increasing • POLARBEAR: probing CMB B-Modes from the Atacama desert • SEASON 1: first measurement of lensing B-modes using the CMB alone, validated with the CIB cross-correlation
!
! • SEASON 1I: Analysis ongoing. ! • FUTURE: probing both lensing and primordial B-modes with POLARBEAR 2 and Simons Array. High sensitivity and foreground rejection with multi-frequency coverage.
Davide Poletti - POLARBEAR collaboration 33 Thank you