1. University of Oxford B. R. Johnson 1 Title Slide MAXIPOL: A Bolometric, Balloon-Borne Half-Wave Plate Polarimeter for Measuring the Polarization of the CMB Bradley R. Johnson NSF and PPARC Postdoctoral Fellow University of Oxford The MAXIPOL payload on the launch pad at the NSBF in Ft. Sumner, New Mexico, September 2002.

2. University of Oxford B. R. Johnson 2 MAXIPOL Collaboration Matthew Abroe 1 Peter Ade 3 Jamie Bock 4 Julian Borrill 5,6 Jeff Collins 2 Pedro Ferreira 7 Shaul Hanany 1 Andrew Jaffe 8 Bradley Johnson 7 Terry Jones 1 Adrian Lee 2,9 Tomotake Matsumura 1 Paul Oxley 1 Bahman Rabii 2 Tom Renbarger 1 Paul Richards 2 George Smoot 2,6,9 Radek Stompor 5 Huan Tran 2 Celeste Winant 2 Proty Wu 10 Joe Zuntz 8 1) School of Physics and Astronomy, University of Minnesota, Minneapolis, MN, USA 2) Department of Physics, University of California, Berkeley, CA, USA 3) Department of Physics and Astronomy, Cardiff University, Cardiff, UK 4) Jet Propulsion Laboratory, Pasadena, CA, USA 5) Computational Research Division, Lawrence Berkeley National Lab, Berkeley, CA, USA 6) Space Sciences Laboratory, University of California, Berkeley, CA, USA 7) Astrophysics, University of Oxford, Oxford, UK 8) Astrophysics Group, Blackett Lab, Imperial College, London, UK 9) Physics Division, Lawrence Berkeley National Lab, Berkeley, CA, USA 10) Department of Physics, National Taiwan University, Taipei, Taiwan

3. University of Oxford B. R. Johnson 3 What are the Science Goals of MAXIPOL? 1) Implement half-wave plate (HWP) polarimetry in a CMB experiment. - MAXIPOL is a pathfinder experiment -- HWP polarimeter techniques and data analysis algorithms will be useful for B-mode experiments. - MAXIPOL is a reimplementation of the MAXIMA hardware - MAXIMA reported best receiver NET -- 40  K  sec. - We wanted to develop a polarization modulator -- HWP polarimeter is the best. 2)Attempt detection of E-mode signal. - Even with the high receiver sensitivity, predictions showed unambiguous E-mode detection would be challenging -- long integration time was required. - We collected data and data analysis is now concluding.

4. University of Oxford B. R. Johnson 4 MAXIPOL Instrument Flight proven MAXIMA hardware. Sun shields not illustrated.

5. University of Oxford B. R. Johnson 5 Cross-Section of MAXIPOL Instrument

6. University of Oxford B. R. Johnson 6 Half-Wave Plate (HWP) Polarimeter signal amplitude corresponds to polarization magnitude. signal phase corresponds to polarization orientation. Q,U sky signals in the frequency domain are away from 1/f noise in sidebands of 4f Q,U signal band HWP rotation frequency

7. University of Oxford B. R. Johnson 7 HWP Polarimeter Advantages Proven astronomical technique (see Astronomical Polarimetry by Tinbergen). 4f modulation provides strong rejection of systematic error. I, Q and U maps from ONE detector. Therefore, not susceptible to differential bolometer gain problems. Q, U signal is far from typical 1/ f noise. Therefore, best noise performance is achieved.

8. University of Oxford B. R. Johnson 8 HWP Polarimeter Hardware HWP construction: 3.2 mm thick sapphire anti-reflection coated with Herasil.

9. University of Oxford B. R. Johnson 9 MAXIPOL Payload on Launch Pad sun shield sun-facing surfaces painted white terrestrial emission baffle balloon light from the sky

10. University of Oxford B. R. Johnson 10 HWP synchronous instrumental signals ie. instrumental polarization Signals in the Time Domain time ordered data is comprised of four primary signals these signals must be rejected note: i is time index modulated sky signals cosmic rays, etc. Reference: B. R. Johnson, J. S. Collins Ph.D. theses

11. University of Oxford B. R. Johnson 11 Signals in the Frequency Domain power spectrum of TOD with h i and g i removed

12. University of Oxford B. R. Johnson 12 Signals in the Frequency Domain telescope scan frequency power spectrum of TOD with h i and g i removed HWP rotation frequency frequency domain representation of the beam

13. University of Oxford B. R. Johnson 13 Signals in the Frequency Domain telescope scan frequency power spectrum of TOD with h i and g i removed HWP rotation frequency frequency domain representation of the beam Q and U signal bands T signal band

14. University of Oxford B. R. Johnson 14 Q,U Demodulation HWP polarimeter works. Nominal noise level is recovered in the Q,U signal band. Maps from demodulated data show no systematic error.

15. University of Oxford B. R. Johnson 15 Polarimeter Characterization

16. University of Oxford B. R. Johnson 16 Polarimeter Characterization receiver cross-polarization and HWP encoder offset modulation efficiency similar measurement limits instrumental polarization to < 1%

17. University of Oxford B. R. Johnson 17 Polarimeter Characterization 0.920.930.90 0.93 0.94 0.950.930.92 Results agree with predicted performance assuming P in = 0.97 and normal incidence for radiation Modulation Efficiency Across the Array error in measured modulation efficiency = 1% for all array elements

18. University of Oxford B. R. Johnson 18 MAXIPOL-1 Flight 26 hour flight from Ft. Sumner, New Mexico, USA in May 2003. Four regions of the sky were observed. Jupiter was mapped twice. - beam shape - calibration CMB dipole was scanned.

19. University of Oxford B. R. Johnson 19 Primary MAXIPOL-1 Scan Regions SFD Dust Map Extrapolated to 140 GHz Primary Dust scan: 1.9 hours centered on Polaris: Dust signal = 38 ± 5  K Primary CMB scan: 7.6 hours centered on Beta Ursae Minoris. Expected dust signal = 4.4 ± 0.7  K MAXIPOL Pointing Illustrated. Schlegel, Finkbeiner and Davis. 1998. ApJ, 500:525-553.

20. University of Oxford B. R. Johnson 20 Intensity Calibration from Jupiter beam map from one 140 GHz photometer Primary calibration from Jupiter observations. Map and best-fit Gaussian contours overplotted (1, 10, 50 and 90%) Some photometer beams were not sampled in every pixel Final calibration results from maximum-likelihood analysis

21. University of Oxford B. R. Johnson 21 Data Analysis Data analysis is near completion. Two analysis algorithms are being used -- MADCAP and frequentist approach. Both methods are producing consistent results. Q,U maps are Gaussian, contain no detectable systematic error.

22. University of Oxford B. R. Johnson 22 Conclusion MAXIPOL demonstrated HWP polarimetry works for CMB experiments to the sensitivity limit of the instrument. Data analysis is near completion. Results papers are currently being written. Papers will include Q, U maps and power spectrum estimates. Experiences from MAXIPOL are advising future B-mode HWP experiments -- EBEX, Polarbear. MAXIPOL-0 Launch

23. University of Oxford B. R. Johnson 23 EBEX expected EBEX performance NASA funded LDB balloon experiment achromatic HWP polarimeter 1476 detectors sensitivity: 0.7  K per pixel Q,U

24. University of Oxford B. R. Johnson 24 END

25. University of Oxford B. R. Johnson 25 Wrong Way

26. University of Oxford B. R. Johnson 26 Instrumental Signals 16 minutes of MAXIPOL TOD HWP synchronous instrumental signal largest contributor is instrumental polarization same data plotted vs. HWP angle

27. University of Oxford B. R. Johnson 27 Instrumental Signal Removal telescope scan frequency power spectrum of TOD differential transmission HWP rotation frequency domain representation of the beam instrumental polarization