Path to CMBPol Upcoming measurem ments of CMB polarization July 1 - PowerPoint PPT Presentation

Path to CMBPol Upcoming measurem ments of CMB polarization July 1 st , 2009 July 1 st , 2009 A. Ku Kusaka (for QUIET co collaboration) Kavli Institute for Cosmologica cal Physics, University of Chicago 1

� The QUIET Experime ent � Status � Q-band observing � W-band deployment � Analysis Prospects � Analysis status of Q-b band data � Power spectrum prosp spect � Systematic errors 2

Introdu Introdu duction duction 3

The only po post-WMAP HEMT pola HEMT pola larimeter larimeter for CMB. 4

QUIET Phase-I Exp xperiment Summary Frequencies 44 4 / 90 GHz Angular resolutions 28 / 12 2 (FWHM) arcmin at each freq Field centers and sizes 181/-39 39, 78/-39, Ra/Dec (Deg) 12/-48, 8, 341/-36 Size (Deg 2 ) 4x(15 × 1 15) = 900 Telescope type Crossed ed Dragone Polarization Modulations Polarization Modulations Phaseswitch Phaseswitch (4kHz&50Hz), (4kHz&50Hz), Deck, Sky ro rot., Fast scan Detector type HE EMT Bolometer, HEMT etc. Location Chajnantor(A (Atacama),Chile µ K s 1/2 , combined Q Instrument NEQ/U ~70 0 / ~60 and U Observation start date Oct., t., 2008 Planned observing time 150 0 / 100 Elapsed / effective days Projected limit on r 0.5 .5 (?) No foreground assumed

OMT L - R decomposition HEMT Module HEMT Module “Polarimeter On Chip” Key technology for large array (JPL) c.f. CAPMAP polarimeter ~3cm ~30cm

L =E X + iE Y R =E X − iE Y � Q - U simultaneous HEMT Amp. measurement � Demodulation Phaseswitch + 1 ± 1 4kHz, 50Hz 4kHz, 50Hz � 1/ f noise reduction � 1/ f noise reduction � Systematic effect 1 180 ° Coupler cancellation � Double-demodulation Det. Diode D | L ± R | 2 � + Q − Q 90 ° Coupler | L ± iR | 2 � + U − U 7

Platelet Foca ocal Plane Array 2nd Mirror (Re eceiver) � 1.4m primary mirror Electronics � Resolution in FWHM: Box � 13 arcmin (W-band) Primary Mirror � 28 arcmin (Q-band) � 28 arcmin (Q-band) Mount Mount � Inherit CBI mount Platelet Array (W-band) ~40cm Mirrors & Support Structure 8

� Chajnantor Plateau, C , Chile � 17,000’ � Extremely low moistur ure � ~1 hour drive from San an Pedro de Atacama � Year-round access � Observing throughout t the year (day and night) 9

Manchest ester Chicago (KICP) Oslo MPI-Bonn Oxford Fermilab Stanford (KIPAC) (KIPAC) KEK Caltech JPL Columbia Miami Princeton Observa rvational Site Chajnan antor Plateau, Chile 5 countries, 13 5 countries, 13 insti insti stitutes, ~ stitutes, ~35 scientists 35 scientists 10

Observin Observin ving Status ving Status 11

Development Q-band observi rving W-band observing Phase-II 2008, October 2 2009, July Q-band obs. start W W-band obs. start 12

Science data taking start Nov., 2009 End of Q-band obs. Jun., 2009 More than 3000 hours of C f CMB data! CMB observing efficiency cy ~ 60% (before selection) 13

QUIET Polarbear Map precision on 1x1 degree pixel: Patch size: ~15 × 15 deg. 2 Planck: 1 µ K (100 GHz) QUIET Phase-I: ~1 uK (40GHz) Number of patches: 4 QUIET Phase-II: 10 -1 µ K (90 GHz)

15

Scanning g Movie 16

• Number of pixels: 19 • 17 polarimetor pixels • 2 ∆ T pixels Typical noise spectrum of a detector diode • Array sensitivity: ~70 uK* � s • Array sensitivity: ~70 uK* � s • Per module ~300 uK* � s • Knee frequency ~5mHz (measured on the sky) • Scan speed ~2deg./s on sky • Critical � ~ 2 • Beam size: FWHM=28 arcmi in 17

Moon WMAP Galaxy (TT, <100 hrs) QUIET 18

Galaxy (Pol., <100 hrs) sytematic efects not considered yet not considered yet WMAP 19

The world largest HEMT MT array polarimeter 20

Q-band Rx taken apart (Jun. 16 th ) W-band Rx landed on mount (Jun. 19 th ) 21

22

� Number of pixels: 90 � 84 polarization pixels � 6 ∆ T pixels 0 µ K* � s (lab. measurement) � Array sensitivity: ~60 � Knee frequency: ~100 00 mHz (lab. measurement) � Critical � ~ 40 (?) � Beam FWHM = 12 ar arcmin 23

Analysis prospe Analysis prospe ects, Systematics ects, Systematics 24

Polarization Noise source Ta au A Moon once / 1.5 hours once / 2d 2days, 1 pix once / 7days Rel. gain, angles Abs. gain, a , angle, Beam Rel. gain, angles Gain stability Temperature Sky-dip Jupiter once / 1.5 hour once / 7days Rel. gain, stability ∆ T gain, beam + Supplemental measurement nts (e.g., polarization wire grid) 25

Moon Pol. angle ~ ± 2˚, Rel. ga gain Pol. gain ~ ± 10% Pol. gain ~ ± 10% Skydip Gain stability ~7%, leaka kage ~0.2% Pol. angle ~ ± 2˚ TauA Jupiter, Venus, RCW38 ∆ T T gain ~ ± 5% 26

Q-band W-band (already collected) (expected) 27

� Fake signal source � Instrumental I � Q / U � Gain Fluctuations � Noise misestimate Avo voidable by cross-correlation � E � B mixing source � mixing source � � Polarization Angle � Optics cross-polarization � Q/U gain mismatch � TOD filtering Sp pecific to pseudo- C � analysis � Geometry/weighting � Overall signal size � Overall gain calibration � Beam size calibration � Pointing error 28

CMB Signal Observed S d Sky After Filtering Detector Noise Contamina nation 1/ f noise is removed by detector 1/ 1/ f noise by high-pass filtering CMB power extraction 29

� Instrumental I � Q / U � See the next slide � Gain fluctuations � Even 20% fluctuation i n is negligible for phase-I � Noise misestimate � Can be avoided by cro ross-correlation technique � Accurate estimate is n necessary for likelihood analysis � QUIET will have both s h schemes with independent pipelines 30

ses I � Q/U leakage • Imperfection of OMT cause • Q-band modules typically h y have 1% of leakages • This leads to negligible bia ias for phase-I • Further cancelation is expe xpected owing to deck rotation Q-band (19 horns) W-band (91horns) Simulation assumi ming 1% leakage 31

� Polarization Angle, Op Optics cross-polarization � Only affects by order o r of ∆θ 2 � Pol. Angle calibration i n in phase-I is better than 5˚. � Q/U gain mismatch (S (See next slide) � TOD filtering � No acceleration � No o leakage � Acceleration leads to l leakage, but negligible � Geometry/weighting ( (See next next slide) 32

� ~20% gain mismatch is negligible for phase-I. � Gain fluctuation of � Gain fluctuation of QUIET-type polarimeter is common for Q and U at the first order � Relative gain between Q and U is stable. � For phase-II, this is just a Fake B-mode due to Q/U gain mismatch calibration issue over the season. 33

Mixing due to pixelization Mixing due to finite patch Mixing due to patch irregularity A = 1/3 Pseudo-C l estimator w/o leakage by K. M. Smith can be used. 34

Current Performance Likely Improvements (noise, duty cycle, 1/f) No foreground assumed ∆ r 0.018 0.005 10 σ lensing 35 σ 35

� Leaks � ∆ T � Pol.: should be < e <<0.1% � OMT improvement � Cancelation thanks to to deck rotation � Precise monitoring � E � B: angle error << <2˚ � Dedicated calibrator � Regulation improveme ent � Both are manageable. le. � Multiplicative errors 36

� Unique RF technology � Different (perhaps better) syst systematics � Q & U measured simultaneo eously at each pixel � Modulation: � 4kHz switching in the module � 4kHz switching in the module les les � Q to U to -Q every 6 hours (fr s (from sky rotation) � ~6% of the sky mapped 100 0 times deeper than Planck � Sensitivity to r = 0.01 � Based on current performan ance (noise, observing eff.) � Foreground treatment: Allian ance with POLARBEAR & ABS � Identical Patches � Frequencies straddle WMAP P “sweet spot” 37

� QUIET � HEMT array polarimeter er receiver: a unique choice of technology wer up to �� ~1000 wer up to �� ~1000 � Mapping CMB pol. powe � Mapping CMB pol. powe � Phase-I � Q-band receiver has alre lready collected 3000 hours of data. Analysis in progress. ess. � W-band receiver has just ust been deployed. � Phase-II � Factor ~20 scale-up of p f phase-I � Technology, Site, Obser serving: All proven at phase-I 38