WMAP final nine-year results David Larson Johns Hopkins University June 10, 2013 OIST, Okinawa, Japan Monday, June 10, 2013
WMAP Science Team Chuck Bennett (PI) Chris Barnes Steve Meyer Rachel Bean Mike Nolta Olivier Doré Nils Odegard Jo Dunkley Lyman Page Ben Gold Hiranya Peiris Mike Greason Kendrick Smith Mark Halpern David Spergel Bob Hill Greg Tucker Gary Hinshaw Licia Verde Norm Jarosik Janet Weiland Al Kogut Dave Wilkinson Eiichiro Komatsu Ed Wollack David Larson Ned Wright Michele Limon 2 Monday, June 10, 2013
WMAP took data for nine years Launch: June 30, 2001 Data taking from L2: August 10, 2001 - August 10, 2010 3 Monday, June 10, 2013
WMAP is differential to minimize systematic errors, with 10 differencing assemblies K1 23 GHz Ka1 33 GHz Q1, Q2 41 GHz V1, V2 61 GHz W1, W2, W3, W4 94 GHz 4 map.gsfc.nasa.gov Monday, June 10, 2013
WMAP frequencies were chosen to surround maximum CMB to foreground ratio Lines correspond to foregrounds outside KQ85 and KQ75 masks KQ85 KQ75 5 Bennett et al. 2013 Monday, June 10, 2013
WMAP scans feedhorn pairs across the sky Bennett et al, 2003 This cross-linked pattern is for one hour of scanning 6 Monday, June 10, 2013
Q band (41 GHz) 7 Monday, June 10, 2013
Q band, with 3.3 mK dipole 8 Monday, June 10, 2013
Q band, with 3.3 mK dipole, masked 9 Monday, June 10, 2013
WMAP 1-hour scan Dipole due to WMAP ʼ s motion around SSB (270 μ K amplitude) not shown 10 Monday, June 10, 2013
WMAP 1-hour scan Dipole due to WMAP ʼ s motion around SSB (270 μ K amplitude) not shown 11 Monday, June 10, 2013
WMAP 1-hour scan Dipole due to WMAP ʼ s motion around SSB (270 μ K amplitude) not shown 12 Monday, June 10, 2013
WMAP 1-hour scan Dipole due to WMAP ʼ s motion around SSB (270 μ K amplitude) not shown 13 Monday, June 10, 2013
Calibration (gain and baseline) iteratively fit using WMAP-velocity modulation of the dipole gain data dipole anisotropy baseline intensity polarization Each of 10 DAs is independently calibrated Final absolute calibration accuracy is 0.2% 14 Hinshaw et al. (2009), Bennett et al. (2013) Monday, June 10, 2013
WMAP achieved 98.4% observing efficiency over the nine-year survey Data is dropped during infrequent (~3x per year) L2 station keeping maneuvers 15 Monday, June 10, 2013
Beam estimation Based on 17 seasons of Nyquist sampled Jupiter observations, and physical mirror model fit to this data (two ~50 day seasons every ~400 days) High S/N region: binned data are used Low S/N region (< 0.25% of integrated response): model is used Full far-sidelobe treatment is included in data analysis, based on ground measurements, early flight moon observations, and analytic models Sidelobe effects are iteratively updated in the gain/baseline calibration: dipole: Δ T vi = Δ T v,main,i + Δ T v,side,i anisotropy: Δ T ai = Δ T a,main,i + Δ T a,side,i 16 Monday, June 10, 2013
New in WMAP9: an additional mapmaking process deconvolves beam asymmetry This is Tau A in K-band, a worst-case scenario Residual maps after subtracting circularly symmetric profiles Residuals are much smaller after symmetrization Bennett et al, 2013 17 Monday, June 10, 2013
K band, 23 GHz, temperature 18 Monday, June 10, 2013
Ka band, 33 GHz, temperature 19 Monday, June 10, 2013
Q band, 41 GHz, temperature 20 Monday, June 10, 2013
V band, 61 GHz, temperature 21 Monday, June 10, 2013
W band, 94 GHz, temperature 22 Monday, June 10, 2013
K band, 23 GHz, polarization 23 Monday, June 10, 2013
Ka band, 33 GHz, polarization 24 Monday, June 10, 2013
Q band, 41 GHz, polarization 25 Monday, June 10, 2013
V band, 61 GHz, polarization 26 Monday, June 10, 2013
W band, 94 GHz, polarization 27 Monday, June 10, 2013
Foreground results 28 Monday, June 10, 2013
WMAP band differences show foregrounds Bennett et al. 2013 Red = W - V is mostly dust Green = (K - Ka) - 1.7 (Q - W) is synchrotron and spinning dust Blue = Q - W is mostly free-free 29 Monday, June 10, 2013
The ILC removes most foregrounds ± 200 μ K Bennett et al. 2013 The ILC provides a region-by-region “Internal Linear Combination” of 5 band maps that minimizes the variance of each region. It keeps the CMB and some small linear combination of foregrounds that partially cancels the CMB fluctuations. 30 Monday, June 10, 2013
An estimate of the CMB-foreground covariance error in the ILC This error is small for a large fraction of the sky Color bar goes up to 66 μ K = 1 σ of CMB fluctuations 31 Monday, June 10, 2013
Detailed foreground models Frequencies fit: 0.408, 23, 33, 41, 61, 94 GHz χ 2 > 10 Bennett et al. 2013 32 Monday, June 10, 2013
Model 9 uses Strong et al. 2011 synchrotron model with -0.5 < Δβ < 0.5 408 MHz WMAP This is a physical model β (A step function works equally well for us) 33 Monday, June 10, 2013
Model 9 fits each pixel independently Strong et al. β synchrotron free-free β dust = 1.8 spinning dust 6 bands (Haslam+WMAP) 7 degrees of freedom Bennett et al. 2013 34 @ 40 GHz Monday, June 10, 2013
Degeneracies in the fit parameters make the model susceptible to noise For Model 9, therm Bennett et al. 2013 dust SD SD peak peak SD SD free-free free-free β synch β synch Since fits are independent between pixels, foreground reconstruction shows noise synch 35 Monday, June 10, 2013
Cosmological Results 36 Monday, June 10, 2013
New in WMAP9: an optimal C -1 estimator for the TT power spectrum, instead of MASTER C -1 MASTER 7-17% reduction in σ 2 (C l ), depending on l Hinshaw et al. 2013 C -1 algorithm: Bennett et al. 2013 Smith et al. 2007 Bond et al. 1998 Tegmark et al. 1997 37 Monday, June 10, 2013
Optimal C -1 estimator reduces error bars 7-17% reduction in σ 2 (C l ), depending on l Most improvement is around l = 600 where S ~ N Bennett et al. 2013 38 Monday, June 10, 2013
Nine years of WMAP data reduce the CMB-only allowed parameter space by a factor of 68,000 Comparison of sqrt(det(cov)) of a Markov chain using likelihood from “Last stand before WMAP” Wang et al, 2003 with WMAP9-only parameter chain Ω b h 2 = 0.02264 ± 0.00050 Ω c h 2 = 0.1138 ± 0.0045 Ω Λ = 0.721 ± 0.025 10 9 Δ 2R = 2.41 ± 0.10 n s = 0.972 ± 0.013 τ = 0.089 ± 0.014 39 Bennett et al. 2013 Monday, June 10, 2013
TE spectrum is well fit with the same Λ CDM model Reionization Super-horizon adiabatic modes 40 Bennett et al. 2013 Monday, June 10, 2013
EE spectrum: determines BB spectrum reionization optical depth Bennett et al. 2013 41 Monday, June 10, 2013
WMAP measures an angle: location of first peak Ratio of first two peak heights gives the physical baryon density: Ω b h 2 = 0.02264 ± 0.00050 Photon to baryon ratio modifies the speed of sound: c s = c / [3(1 + 3 Ω b /4 Ω r )] 1/2 WMAP measures Baryons, photons, and FRW give time to last scattering. this angle Integrating sound speed over this time gives the horizon. on the sky If the universe is flat (Euclidean), this lets us measure the distance to last scattering, which gives us the Hubble constant: H 0 = 70.0 ± 2.2 km s -1 Mpc -1 (and d A above) If we measure the expansion of the universe (H 0 = 73.8 ± 2.4), we find it is flat WMAP9+eCMB+BAO+H 0 : 42 Monday, June 10, 2013
WMAP9 has a tight curvature-H 0 degeneracy: one determines the other axis 43 Bennett et al. 2013 Monday, June 10, 2013
44 Monday, June 10, 2013
WMAP limits tensors mostly through TT r = 10, τ = 0.050 r = 1.2, τ = 0.075 r = 0.2, τ = 0.080 5-year WMAP data, Komatsu et al, 2009 WMAP9 only: r < 0.38 (95% CL) 45 Monday, June 10, 2013
WMAP limits tensors mostly through TT r < 0.17 (95% CL), n s = 0.970 ± 0.011 WMAP9+eCMB r < 0.13 (95% CL), n s = 0.9636 ± 0.0084 WMAP9+eCMB+BAO+H 0 Hinshaw et al, 2013 46 Monday, June 10, 2013
WMAP does not detect bispectrum non-Gaussianity: f NL Komatsu 2010 f NLlocal = 37.2 ± 19.9 (Bennett et al, 2013) Expected to be 0.015, from n s measurements Can be created from a local perturbation: Weights squeezed triangles positively: k 1 = k 2 >> k 3 . f NLeq = 51 ± 136 Can be created by inflation models with non-canonical kinetic terms (i.e. Dirac-Born-Infeld) Weights equilateral triangles positively: k 1 = k 2 = k 3 . f NLorth = -245 ± 100 Can be created by a linear combination of higher + derivative scalar-field interaction terms F NLorth mostly orthogonal to F NLlocal and F NLeq . - Weights equilateral triangles positively, and squeezed and folded triangles negatively - These are not all possible ways to search for a bispectrum 47 Monday, June 10, 2013
Recommend
More recommend
