Measuring CMB Polarization from the Garden State with CAPMAP Colin Bischoff The Path to CMBPol Workshop July 1, 2009
Overview of CAPMAP Experiment Collaboration between U. of Chicago, Princeton, U. of Miami, (C. Bischoff, L. Hyatt, J. McMahon, G. Nixon, D. Sambtleben, and JPL K. Smith, K. Vanderlinde, D. Barkats, P . Farese, T. Gaier, J. Gunderson, M. Hedman, S. Staggs, B. Winstein) Three observing seasons from 2003-2005 at the Crawford Hill 7-meter antenna in Holmdel, New Jersey . Third season consisted of 1658 hours from December 2004 through April 2005. Sixteen HEMT-based correlation polarimeters: 12 operating at 90 GHz (3.3' FWHM) 4 operating at 40 GHz (6.5' FWHM) E-mode polarization measured in 7 bins from 200 < < 3000 with total detection significance of B-mode polarization 95% confidence upper limit of
CAPMAP Ring Scan Circular scan about the NCP with 0.725 radius and 21 second period Good noise uniformity and complete parallactic angle coverage every 12 hours 2 107 d0 response [mV] 1 106 0 -1 105 0 20 40 60 time [seconds] Ring scan introduces a large -2 signal in total power data -2 -1 0 1 2
Removal of Atmospheric Modes Five parameter fit to each 21 second orbit
Removal of Atmospheric Modes Five parameter fit to each 21 second orbit TP 100K 500mK 0 0 0 Pol <150mK <5mK 0 0 0
Removal of Atmospheric Modes Five parameter fit to each 21 second orbit Bad weather was identified using the goodness of fit for the five parameter model to total power data on two time scales
Removal of Atmospheric Modes Five parameter fit to each 21 second orbit 400 After removing the model, number of cycles 300 polarization data is consistent with a 200 distribution, indicating that no additional 100 filtering is necessary . 0 0 10 20 χ 2
Null T est Suite Thirty-six different null maps were constructed from different divisions of the data and analyzed. Results of the nulltest suite were used to tune weather cuts, gain models, etc, for a "blind" analysis Q-Band EE Q-Band BB W-Band EE W-Band BB Flat Bandpower [ µ K 2 ] Flat Bandpower [ µ K 2 ] Flat Bandpower [ µ K 2 ] Flat Bandpower [ µ K 2 ] Category Null Test -50 0 50 -50 0 50 -20 0 20 40 -20 0 20 40 Pe rio dI v sII Pe rio dI v sIII Pe rio dII v sIII Time D a y v s N ig h t Divisions M o rn in g v s Ev e n in g Qu a d ra t u re 0 v s 1 + 2 + 3 1 v s 0 + 2 + 3 Cryostat Divisions 2 v s 0 + 1 + 3 3 v s 0 + 1 + 2 S0 v s S1 + S2 Spectral S1 v s S0 + S2 Divisions S2 v s S0 + S1 IF W a rm v s Co ld House- LN A W a rm v s Co ld keeping Divisions M irro r W a rm v s Co ld 1 /f N o ise Gro u n d -sy n ch ro n o u s Performance St ru ct u re Divisions I-> Q Le a k a g e
Null T est Suite Full null test results show no evidence of contamination = 77.8 for 72 degrees of freedom (P .T.E. = 30%) Not included in the null test suite is the null map made from differencing the 40 GHz and 90 GHz data, which provides evidence that foregrounds are not contributing significant power. 30 Number of Null Tests 25 20 15 10 5 0 0 2 4 6 8 10 χ 2 to Zero
Results from T wo Independent Pipelines Separate analyses produced consistent results using complementary methods. 60 EE BB l ( l +1)C l / (2 π ) [ µ K 2 ] 40 20 0 -20 -40 0 1000 2000 3000 1000 2000 3000 l l Systematic errors are estimated from many full season simulations and found to be sub-dominant.
150 MAXIPOL DASI CBI BOOMERa nG 100 WMAP QUAD 1)C l / 2 π [ µ K 2 ] CAPMAP 05 W-Ba nd CAPMAP 08 Q+ W 50 0 l(l+ -50 3 10 0 500 1000 1500 2000 2500 l 2 10 1)C l / (2 π ) [ µ K 2 ] 10 1 MAXIPOL 06 -1 DASI 05 10 l(l+ CBI 05 BOOMERa nG 02 QUa D 07 -2 10 WMAP 06 CAPMAP 08 Q+ W -3 10 0 500 1000 1500 2000 2500 l
Recommend
More recommend
