From Boomerang to BB-Pol

Mem. S.A.It. Vol. 79, 799 c SAIt 2008 Memorie della

From BOOMERanG to B-B-Pol

Balloon-borne observations of cosmic microwave background polarization

P.de Bernardis1,2, M. Calvo1,2, P.Fiadino1,2, C. Giordano1,2, S. Masi1,2, F. Piacentini1,2, and G. Polenta1,2

1 Dipartimento di Fisica, Universita` La Sapienza – P.le A. Moro 2, I-00185 Roma, Italy 2 INFN Sezione di Roma 1, e-mail: [email protected]

Abstract. Balloon experiments have played and will play a pivotal role in the study of the Cosmic Microwave Background (CMB), the fossil radiation coming from the early universe. The BOOMERanG experiment has measured the anisotropy (1998-2002) and the E- mode polarization of the CMB (2003-2006), while a new ﬂight of BOOMERanG is planned to measure the polarization generated by interstellar dust at high galactic latitudes. This in- formation is necessary in view of future satellite experiments aimed at a measurement of the B-mode polarization in the CMB. The B-B-Pol experiment is aimed at measuring the polarization of the CMB at large angular scales, using long duration balloon ﬂights during the polar night. This experiment will validate several of the experimental methods to be used in the future B-Pol satellite mission.

Key words. Stars: abundances – Cosmology: Cosmic Microwave Background – Cosmology: observations – Stratospheric Balloons

1. Introduction observed is not due to instrumental artifacts (Masi et al. 2006). Moreover, it has exactly the spectrum of CMB anisotropy, which means Observations of the Cosmic Microwave that it is not due to foreground emission. So Background have provided detailed views of we can conclude that everything we see in the Universe at the recombination epoch, at these maps results from structures present redshift ∼ 1100. In the CMB anisotropy maps in the Early Universe, and from effects on from BOOMERanG ( Masi et al. 2006), due to light propagation along the path from that the extreme sensitivity of cryogenic bolome- early epoch to our telescopes. The image of ters operated at balloon altitude, a noise of the CMB thus depends on the physics of the ∼ 20µK for each 3.5’ pixel has been achieved. very early universe (initial conditions), on the The consistency of the maps from independent physics of the Universe during the plasma era experiments (for example BOOMERanG and (photons-baryons acoustic oscillations), and WMAP), working at different frequencies and on the geometry of light propagation, i.e. on with very different measurement methods, the geometry of the universe at large scales. is the best evidence that the faint structure 800 P. de Bernardis et al.: from BOOMERanG to B-B-Pol

We expect that the characteristic scale of the ing electrons. Where there is a velocity gra- acoustic horizon is imprinted in this image. dient, a quadrupole anisotropy in the radia- Since this linear scale can be computed from tion is generated, and the scattered radiation first principles, we can carry out an angular is polarized. So the measurement of the polar- size test on the image of the CMB, thus ization of the CMB probes the velocity field deriving the density parameter. This and other present at recombination. This effect, due to cosmological parameters are best estimated scalar fluctuations, produces a non-rotational by means of a power spectrum analysis of polarization field (called E-modes of CMB po- the maps of the CMB (de Bernardis et al. larization). The E-mode polarization spectrum 2002). This method has been so successful EE (and its correlation with the anisotropy that nowadays we have a minimalist model TE) can be computed very accurately from fitting the angular power spectrum of the CMB the density fluctuations inferred by the mea- and a number of other cosmological obser- sured anisotropy spectra. A signal of the or- vations (the abundances of primordial light der of a few µK rms is expected for an exper- elements, the recession of galaxies, the fluxes iment with angular resolution better than 1o. of high redshift SN1a candles, the large-scale The power spectrum of this signal has max- distribution of galaxies and the distribution of ima where there are minima in the anisotropy Ly-α clouds...). The model has only 6 param- power spectrum, just because in a density os- eters (the Hubble constant Ho, the density of cillation there is maximum velocity when the baryons Ωb, the density of matter - including density is equal to the average, while there is dark matter - Ωm, the density of Dark Energy zero velocity when the density fluctuation is ΩΛ, the spectral index of the power spectrum maximum. The second source of quadrupole of primordial density perturbations ns, and anisotropy at recombination is the presence of the mass power spectrum normalization σ8), long wavelength gravitational waves. If, as ex- and works remarkably well from an empiric pected in the inflationary scenario, a stochas- point of view (see e.g. Tegmark et al. 2004). tic background of gravitational waves is gener- There are, however, three main enigmas in the ated in the very early universe, then we should current scenario: the nature of dark matter, the be able to see an additional component of the nature of dark energy, and the origin of the polarization field. The amplitude of this com- universe itself and of its structure. In the fol- ponent is very small (100 nK or lower, de- lowing we will focus on the third one, and will pending on the energy scale of inflation), but show how balloon-borne missions can help this has also a curl component (B-mode, BB). in solving the enigma of the origin of cosmic Using proper analysis methods, the B-mode in- structures, through detailed observations of the flationary component can thus be disentangled polarization of the CMB. from the dominant E-mode, opening a unique window to probe the very early universe and the physics of extremely high energies (around 2. CMB Polarization measurements 1016 GeV) [see e.g. (Dodelson 2003)]. CMB photons are Thomson-scattered by elec- BOOMERanG was the first instrument to trons at recombination. Linear polarization in produce images of the CMB with resolution the scattered radiation is obtained if there is and sensitivity good enough to show the sub- a quadrupole anisotropy in the incoming, un- degree structure of acoustic horizons at recom- polarized photons (Rees 1968). This can be bination. We published multi-frequency maps produced in two ways. The first, unavoidable of 3% of the sky, with 10’ FWHM resolution source of quadrupole anisotropy of the pho- (de Bernardis et al. 2000). From these, accu- ton field is the density fluctuation field present rate angular power spectra of the CMB (l=50- at recombination. Density fluctuations induce 1000) were computed (Netterfield et al. 2002; peculiar velocities in the primeval plasma. For de Bernardis et al. 2002; Ruhl et al. 2003), this reason a given electron receives redshifted showing the presence of three acoustic peaks or blueshifted radiation from the surround- at l = 210, 540, 845, due to the acoustic os- P. de Bernardis et al.: from BOOMERanG to B-B-Pol 801

Fig. 1. Launch of the BOOMERanG experiment in Antarctica (Jan.6th,2003) cillations of the photon-baryon plasma. From bility to the measurement (Hanany et al. 2000; these spectra, cosmological parameters were Lee et al. 2001). derived (Lange et al. 2001; de Bernardis et al. The BOOMERanG payload has been mod- 2002; Ruhl et al. 2003). We found that the ge- ified in 2002 to make it sensitive to the polar- ometry of the universe is very nearly flat ( ization. Polarization sensitive bolometers have Ωo = 1.03 ± 0.04 ); that the fluctuations de- been custom developed (Jones et al. 2003), rive from a primordial density power spectrum and an improved attitude reconstruction sys- nearly scale invariant (ns = 1.02 ± 0.07): two tem based on a steerable star-camera was used. predictions of the inflationary scenario fully All the details of this instrument, flown in 2003 confirmed. Moreover, we found that the den- (see fig.1), and of the produced maps are re- 2 sity of baryons is Ωbh = 0.023 ± 0.003, ported in(Masi et al. 2006) . In (Jones et al. perfectly consistent with the predictions of 2006; Piacentini et al. 2006; Montroy et al. the Big Bang Nucleosynthesis scenario tested 2006) we reported the measured TT, TE and from the measurement of primordial abun- EE spectra. While both TE and EE are de- dance of light elements. The same density tected with high significance, BB is not de- of baryons is required in the nucleosynthe- tected, and the precision of the other spectra sis process, in the first three minutes after the is not sufficient yet to constrain the cosmologi- Big Bang, and 380000 years later, during the cal parameters better than anisotropy measure- acoustic oscillations of the primeval plasma. ments (Mac Tavish et al. 2006). This situation This fact strongly strengthens the general au- is common to all the CMB polarization mea- toconsistency of our early universe model. surements to date (see e.g. (Readhead et al. We also investigated the Gaussianity of the 2004; Leitch et al. 2005; Barkats et al. 2005; CMB maps by means of Minkowski function- Page et al. 2006; Ade et al. 2007)). A notice- als, Bispectrum, Trispectrum (Polenta et al. able exception is the measurement of the op- 2002; De Troia et al. 2003, 2007). In paral- tical depth of reionization, estimated from the lel, using the ancillary measurements at 410 large-scale polarization of the CMB better than GHz, we have shown that the contamina- from anisotropy measurements (Page et al. tion form interstellar dust is less than 1% of 2006). CMB Power Spectrum at 150 GHz (Masi et al. Considerable effort is being spent by CMB 2001). The dual experiment MAXIMA, flown experimentalists to prepare high accuracy and from Palestine (TX), produced almost at the sensitive experiments to detect the rotational same time as the first BOOMERanG results component of CMB polarization. There are a CMB map statistically consistent with the three big issues to be solved for this measure- BOOMERanG one. The power spectra ob- ment: tained from the two experiments were also sta- • Detector noise : the signal we are look- tistically consistent, giving even higher relia- ing for is in the 1-100 nK range. Successful 802 P. de Bernardis et al.: from BOOMERanG to B-B-Pol

Fig. 2. Concept diagram of the High Frequency Instrument for B-B-Pol . In the box, a front view of the detector array, consisting of 37 detectors per channel, is shown. The size of the array does not require a very large cryostat. polarization surveys to date have reported ferent levels of noise has been estimated in measurements of signals of the order of 2 several papers (see e.g. (Pagano et al. 2007)). µK rms. Since cryogenic bolometers are al- However, the real difficulty is to solve simulta- ready close to be limited by photon noise neously also the other issues listed below. of the CMB itself, the 100x improvement in • Systematic Effects : this is a real nightmare. sensitivity can only be obtained increasing The signal to be detected is so much smaller the number of detectors, i.e. by developing than detector noise that it is very difficult to de- large-format arrays of microwave detectors. tect low level systematics. Current polarization Different technological solutions are compet- experiments have managed to keep the sys- ing: Transition Edge Sensor bolometers with tematic effects below about 0.1 µK. Here we SQUID readout are the most advanced, and im- want to improve this by a factor around 100. portant scientific results are already being pro- The only way to improve here is to experi- duced (Halverson et al. 2008). Cold Electron ment. New polarization modulators and opti- Bolometers and Kinetic Inductance Detectors cal components must be characterized with un- are also improving at high speed. Coherent de- precedented accuracy. Enough to keep the ex- tectors also have been compactly packaged to perimentalists busy for years with laboratory build large focal planes. The likelihood to de- developments and ground-based and balloon- tect the primordial gravitational waves for dif- borne experiments. P. de Bernardis et al.: from BOOMERanG to B-B-Pol 803

• Polarized Foregrounds. Our galaxy emits po- 3. B-B-Pol larized radiation: elongated interstellar grains, aligned by the Galactic magnetic field, pro- A set of preliminary balloon flights (named duce thermal emission with a few percent po- B-B-Pol, for Balloon B-Pol) will be used to larization degree(Ponthieu et al. 2005); radia- develop and validate the involved technolo- tion emitted by the electrons of the interstellar gies, and, most important, to detect system- medium spiralizing in the Galactic magnetic atic effects related to the measurement, vali- field produce highly polarized synchrotron ra- dating mitigation strategies. The scientific tar- diation. A first survey of this emission has been get of B-B-Pol is to explore the low-` region carried out by the WMAP satellite (Page et al. of the CMB polarization angular power spec- 2006). The level of total polarized emission trum, ` ∼< 50 − 100. This region is very diffi- is up to 100 times larger than the goal po- cult to explore from ground due to the vari- larization signal from primordial gravitational ability of the atmosphere at large scales, espe- waves. It is thus necessary to characterize the cially at high frequencies (ν ∼> 100GHz). The polarized foreground at the 1% level. final measurement has to be done from space, but a balloon measurement can already pro- The EBEX balloon-borne instrument duce significant results. The first issue is sensi- (Oxley et al. 2004) aims at detecting the tivity. With current technology, a long-duration B-modes by means of a high resolution (< 8’) balloon flight will last from 2 to 4 weeks. Even telescope and large-format arrays of bolome- longer flights have been carried out by NASA ters working at 150, 250, 350, 450 GHz. The in Antarctica, during the Antarctic summer. polarization is modulated by an achromatic However, a large-scale measurements needs to rotating half wave plate: a technique pioneered cover a very large fraction of the sky, which by the MAXIPOL experiment (Johnson et al. means a polar night flight. Balloons have been 2006). The instrument is to be flown on a flown during the Arctic polar night (for ex- long-duration balloon in Antarctica. ample CNES has flown Archeops for a 2- SPIDER (Mac Tavish et al. 2006; days flight), so the technology is mature to at- Crill et al. 2 008) is devoted to large-scale tempt long duration flights in the polar night. measurements of CMB polarization, obtained The Italian Space Agency with Andoya Rocket with cryogenic bolometers covering the spec- Range is developing the Svalbard site for polar tral range from 80 to 270 GHz. A large fraction flights (see the contribution of S. Peterzen in of the sky is observed during an ultra-long these proceedings), and a test polar night flight duration balloon flight. The instrument is will be tried soon. This will give us a definitive a precursor of the EPIC CMB Polarization confirmation of the feasibility of long duration satellite, studied in the framework of the balloon flights around the north pole during the NASA Beyond Einstein program (Bock et al. polar night. 2006). Assuming that polar night flights are fea- On the European side, in the framework of sible, we have designed B-B-Pol as a spinner, the Cosmic Vision ESA program, the B-Pol rotating at constant speed in azimuth. If the de- satellite has been proposed (B-Pol 2007). B- tector beams are oriented at an elevation of 30 Pol is based on a set of refractive telescopes to 45 degrees, a large fraction of the northern (one per band, from 45 to 350 GHz), each with hemisphere can be scanned every day of the its own wave-plate polarization modulator and mission, in an extremely cold and stable envi- its bolometric detector array. Though B-Pol ronment. was not selected for the first round of Cosmic For the high frequency instrument, we have Vision missions, the importance of the CMB considered a multi-pixel polarimeter arranged polarization science and the necessity to carry as in fig.2 out all the technological developments for this The incoming radiation is modulated by experiment have been stressed by ESA, in view the assembly of a rotating HWP and a wire of the next round of calls. grid polarizer. There is no other optical com- 804 P. de Bernardis et al.: from BOOMERanG to B-B-Pol

Fig. 3. Sample sky coverage in equatorial coordinates achievable in a single 14 days ﬂight during the polar winter. The region observed by the experiment is the lighter horizontal band at declinations between 23o to 77o. The background map is extrapolated from the IRAS 100 µ survey, and represents the brightness of interstellar dust.

Fig. 4. Sketch of the BOOMERanG-FG focal plane. The experiment is aimed at measuring the polarization of interstellar dust at high galactic latitudes at 350 GHz, a frequency very close to those used to measure CMB polarization. ponent in front of the system, so that the po- of the instrument by a significant factor at the larization purity depends entirely on the qual- expense of a lower angular resolution, which ity of the HWP and only slightly on the polar- is not needed for the target of this instrument. ization properties of the horns in the detector Worksheet calculations show√ that a sensitivity array. The main feature of the detection sys- of the order of 10 µK/ Hz can be reached tem is the use of bolometers feeded by multi- both at 140 GHz and at 220 GHz for each de- moded horns. With respect to single-mode detector of the array. In fig.3 it is shown a sam- tectors, this approach increases the sensitivity ple sky coverage achievable in a single two P. de Bernardis et al.: from BOOMERanG to B-B-Pol 805 weeks flight during the polar winter. Using this are at the Universities of Manchester and Oxford. simulation, it can be shown that the measure- BOOMERanG-Fg is developed in Italy by the ments of the reionization peak in the polariza- groups at University of Rome La Sapienza, IFAC- tion power spectrum at ` ∼ 6 are limited by CNR Firenze, INGV-Rome. International partners cosmic variance for r ∼ 0.01. are at the Universities of Cardiff and the University In addition to this already important sci- of Chalmers. The developments of BOOMERanG- FG and B-B-Pol are supported by the Italian Space entific target, B-B-Pol will be used to test Agency. key technologies required by ESA to achieve undisputable technology readiness for the new Cosmic Vision Call. As evident from fig.2 the References high frequency instrument polarimeter will use a rotating half wave plate (HWP) / polarizer Ade P., et al., 2007, astro-ph/0705.2359 assembly to modulate the linear polarization Barkats D., et al., 2005, ApJ, 629, L127 from the sky. As in B-Pol, this requires a high- Bock J., et al., 2006, astro-ph/0604101 quality HWP (see e.g. Pisano et al. (2006); B-Pol collaboration: www.b-pol.org Savini et al. (2006)) and a low power, high Crill B., et al., 2008, astro-ph/0807.1548 speed cryogenic rotator (see e.g. Hanany et al. de Bernardis P., et al. 2000, Nature, 404, 955 (2003)). This system is crucial to the success de Bernardis P., et al., 2002, ApJ, 564, 559 of the satellite mission, and will be fully devel- De Troia, G., et al., 2003, MNRAS 343, 284 oped and tested in the balloon B-B-Pol . De Troia, G., et al., 2007, ApJ, 670, L73 Dodelson S., ”Modern Cosmology: Anisotropies and Inhomogeneities in 4. BOOMERanG-FG the Universe” Academic Press 2003 Fowler J.W., et al., 2007, Applied Optics, 46, The new flight of BOOMERanG is devoted to 3444 the measurement of the polarized brightness of Johnson B.R., et al., 2007, ApJ, in press, astro- interstellar dust at high galactic latitudes, in a ph/0611394 band centered at 350 GHz. This frequency has Jones, B., et al., 2002, SPIE, 4855, 227 been selected because it very close to the op- Jones B. et al., 2006, ApJ, 647, 823 timal bands to measure CMB polarization (50- Halverson N.W. et al., 2008, astro- 220 GHz): the extrapolation of the interstellar ph/0807.4208 dust polarization to CMB frequencies is thus Hanany S., et al., 2005, ApJ Lett., 545, L5 more accurate than from high frequency sur- Hanany et al., 2003, IEEE Trans. on Appl. veys. The focal plane of BOOMERanG-FG is Supercond. 13, 2128 sketched in fig.4. The baseline detectors for Kuzmin, L., 2007, EUCAS-07, Brussel, this mission are cold electron bolometers de- Journal of Physics: Conference Series, in veloped in Chalmers (Kuzmin 2007). The rest press of the payload is very similar to the one de- Lange, A., et al., 2001, Phys.Rev. 63, 042001 scribed in Masi et al. (2006). In a long dura- Lee A., et al., 2001, ApJ, 561, L1 tion summer flight from Svalbards, we plan to Leitch E., et al., 2005, ApJ, 624, 10 carry out a deep survey of ∼ 400 square de- MacTavish C. et al., 2006, ApJ, 647, 799 grees region at high galactic latitudes (center MacTavish C., et al., 2006, Proceedings of at RA∼ 16h, dec∼ 40o) and a survey of ∼ 1000 Science PoS(2006)71 square degrees in the galactic plane (center at Masi, S., et al., 2001, ApJ, 553, L93 RA ∼ 20.7h, dec∼ 40o). Masi S., et al., 2006, A&A, 458 , 687 Acknowledgements. B-B-Pol is a nation-wide col- Montroy T. et al., 2006, ApJ, 647, 813 laboration including the CMB groups of Univ. Netterfield C.B., et al. 2002, Ap.J. 571, 604 of Rome La Sapienza, Bologna IASF and IRA, Oxley, P., et al., 2004, Proc.SPIE Opt.Eng. Univ. of Milano, Univ. of Milano Bicocca, IEIIT- 5543, 320 CNR Torino. The pre-phase-A study is funded by Pagano L., et al., 2007, astro-ph/0707.2560 the Italian Space Agency. International partners Page L., et al., 2006, astro-ph/0603451 806 P. de Bernardis et al.: from BOOMERanG to B-B-Pol

Piacentini F. et al., 2006, ApJ, 647, 833 Rees M., 1968 ApJ Lett., 53, L1 Pisano, G., et al., 2006, Appl. Opt. 45, 6982 Ruhl, J., et al., 2003, ApJ, 599, 786-805 Polenta, G., et al., 2002, ApJ Lett., 572, L27 Savini, G., et al., 2006, Applied Optics, 45, Ponthieu N., et al., 2005, Astron.Astrophys. 8907 444, 327 Tegmark M., et al., 2004, Phys.Rev. D69, Readhead A.C.S., et al., 2004, Science, 306, 103501 836