Slide 1

First results from QUIET Osamu Tajima (KEK) The QUIET Collaboration 1 Slide 2 B-modes have NOT been observed yet ! Direct limits : r < 0.7 (ground experiment) Indirect limits: r < 0.2 Contribution from Gravitational lensing Primordial B-modes B-modes power Angular Scale Large scale Small scale Multipole l (=180 o / ) QUIET aims to detect B-modes from ground ! 2 Slide 3 The QUIET Collaboration 5 countries, 14 institutes, ~35 scientists Chajnantor Plateau (5,080m) Chile Atacama Desert Worlds best site for observation frequencies of QUIET ! 3 Slide 4 Observation Patches 4 CMB patches were chosen (~3% of full sky) Observing them DEEPLY(Galaxy observation when CMB patches are not visible) Map precision on 1x1: ~1K (7.5 months at 43GHz) 4 ~20 o Slide 5 CMB QUIET Telescope Receiver ( detector array inside) CMB 5 ~30cm 90 detectors array for 95 GHz Slide 6 QUIET observation time at Chajnantor, 5,080m 19 detectors at 43GHz array sensitivity 69uKs 1/2 90 detectors at 95GHz array sensitivity ~70uKs 1/2 ~30cm 7.5 months 1.5 years > 11,000 H 6 Slide 7 Whats important towards B-mode detection ? B-mode ~ 1/100 of E-modes x100 better sensitivity than past experiments Detector array: High sensitive instrument limitation of single detector sensitivity Several hundreds ~ thousand detectors several (past) ~100 (Now) ~1000 (Future) Good systematic error control for instrument Understanding of Foregrounds QUIET : intermediate stage (2008-2010) - Observation with 90 (19) detectors at 95GHz (43GHz) - One of the best B-modes search to date - Proof of technology for future QUIET : intermediate stage (2008-2010) - Observation with 90 (19) detectors at 95GHz (43GHz) - One of the best B-modes search to date - Proof of technology for future 7 Slide 8 Foregrounds and observation bands B-mode (as QUIET-1 limits) QUIET Other experiments using bolometer 43 GHz 95 GHz QUIET 43GHz data is very important to understand the contribution of Synchrotron emission 8 Slide 9 Impact of systematic error Have to minimize spurious polarization < 1% Have to achieve < 2 o precision Temperature anisotropy E-modes lensing B-modes r = 0.10 r = 0.01 In case of 1% precision of calibrations spurious pol. 1% of I to Q/U 2 o for pol. angle Multipole l (=180 o / ) l(l+1)C l /2 (uK 2 ) 9 Slide 10 QUIET polarization detector array CMB Polarization Sensor Module Septum Polarizer 3cm 90 detector array for 95 GHz Array sensitivity ~70 uKs 1/2 Robust detector against to the systematic biases 10 Slide 11 Septum Polarizer (OMT) x y Input Output Input Output L = E X iE Y R = E X iE Y R L 11 Slide 12 Polarization Sensor Module L R QQ UU UU 11 11 QQ GAGA GBGB Septum polarizer HEMT amp. Phase switch modulation at 4kHz & 50Hz 180 Coupler (1) 90 Coupler (i) W-band module Antenna to pick up L, R 12 Slide 13 Polarization Sensor Module L R QQ UU UU 11 11 QQ GAGA GBGB Septum polarizer HEMT amp. Phase switch modulation at 4kHz & 50Hz 180 Coupler (1) 90 Coupler (i) Simultaneous measurement of Stokes Q and U Polarization (Q, U) G A x G B Strong immunity from systematic bias NO spurious polarization, NO polarization angle rotation, i.e. Q/U rotation, even though there is gain fluctuation QUIET detector is extremely stable for the polarization response 13 Slide 14 I Q/U Leakage Caused by cross talk in septum polarizer NO time variation because it caused by waveguides components CMB Polarization Sensor Module Septum Polarizer II Variation of atmosphere thickness Elevation nods QQ Spurious polarization Instrumental spurious polarization 43GHz receiver I Q : 1.0% I U : 0.2% (precision 0.1%) average 0.6% 95GHz receiver I Q : < 0.5% I U : < 0.5% ~~~~ 14 Slide 15 Rotate parallactic angle with keeping the line of sight Q ~4 min scan time for each Q U Calibration for Polarization (43GHz receiver) T Q(U) cos(2( - )) absolute = 1.7 Catalog uncertainty for polarization angle 1.5 at 43GHz (WMAP) 0.2 at 95GHz (IRAM) Taurus QUIET telescope Crab nebula (TauA) by Y. Chinone 15 Slide 16 Whats important towards B-mode detection ? B-mode ~ 1/100 of E-modes x100 better sensitivity than past experiments Detector array: High sensitive instrument limitation of single detector sensitivity Several hundreds ~ thousand detectors several (past) ~100 (Now) ~1000 (Future) Good systematic error control for instrument Understanding of Foregrounds Robust coherent detector Calibration, Scan strategy Analysis method Intermediate stage 43GHz receiver for Synchrotron emission Verified with first results 16 Slide 17 First results from QUIET with 43GHz Receiver 17 Slide 18 End-Analysis Strategy Data Selection Filter / Map Making Power Spectra Cosmological Parameters Validation Tests 18 Slide 19 Validation Tests End-Analysis Strategy Data Selection Filter / Map Making Blind Analysis Framework Power Spectra Cosmological Parameters Systematic Error Check Calibrations Box Open Un-blinding the results - after passing validation tests - after confirmation of syst. errors 19 Slide 20 Data Selection - way to control hidden systematic bias - Contaminated Clean Data Set Selection Criteria Good weather Extremely bad weather Time-ordered-data for polarization response To determine the selection criteria, we need the way to evaluate such hidden bias 0.1 mK 80 mK 20 Slide 21 (S + N 1 ) (S + N 2 ) Way to evaluate the hidden bias in data without looking at the results Analysis Validation : Null Tests MC (N 1 N 2 ) MC Same CMB signal but different noise, contaminations Q U diodes diff. We performed null tests with various subdivisions (42 different ways). - weather condition - cryostat temperature - We determined selection criteria with feed-back from null tests 69.4% for 43 GHz detector array 21 Slide 22 Evaluation of Null Spectra 22 Significant non-null bias (20% of statistical error) w/ Cross-correlation w/o Cross-correlation Auto-correlation There is significant bias even if the criteria are tighten (auto-correlation) It indicates that faint contamination was always exists in the data Need the way to drop such effect with keeping the CMB signal = C l / l Bias estimator : MC w/o any contamination Slide 23 Cross-correlation Maps of different time periods S l + N 1 l S l + N 2 l S l + N 3 l + Cross-correlations with all the combinations Technique to eliminate the noise and remaining contamination CMB signals (S l ) are the same and correlations do not vanish, while noise terms (N l ) have no correlations = 0. 23 Slide 24 There were far-sidelobes during 43GHz observation season 43GHz observation 95GHz observation Contaminations by far-sidelobes were always existed, e.g. picking up ground structure