Bernardis (Sapienza University of Rome) LSPE Collaboration ... · P. de Bernardis (Sapienza...
Transcript of Bernardis (Sapienza University of Rome) LSPE Collaboration ... · P. de Bernardis (Sapienza...
P. de Bernardis (Sapienza University of Rome) for the LSPE CollaborationSCAR AAA Meeting
26/07/2013 Certosa di Pontignano (Siena, Italy)
The LSPE collaboration
The LSPE collaboration: G. Amicoa, P. Battagliab, E. Battistellia, A. Baùc, P. de Bernardisa, M. Bersanellib, A. Boscalerid, F. Cavaliereb, A. Coppolecchiaa, A. Cruciania, F. Cuttaiae, A. D’ Addabboa,
G. D'Alessandroa, S. De Gregoria, F. Del Tortoc, M. De Petrisa, L. Fiorineschif, C. Franceschetb, E. Franceschif, M. Gervasic, D. Goldieg, A. Gregorioh,p, V. Haynesi, N. Krachmalnicoffc , L. Lamagnaa, B. Maffeii,
D. Mainoc, S. Masia, A. Mennellac, Ng Ming Wahi, G. Morgantee, F. Natia, L. Paganoa, A. Passerinic, O. Peverinil, F. Piacentinia, L. Piccirilloi, G. Pisanoi, S. Ricciardie, P. Rissonef, G. Romeom, M. Salatinoa, M. Sandrie, A.Schillacia, L. Stringhettin, A. Tartaric, R. Tasconel, L. Terenzie, M. Tomasin, E. Tommasio,
F. Villae, G. Vironel, S. Withingtong, A. Zaccheip, M. Zannonic
a Dipartimento di Fisica, Sapienza Università di Roma, P.le A. Moro 2, 00185 Roma, Italyb Dipartimento di Fisica, Università di Milano, Via Caloria 16. 20133 Milano, Italyc Dipartimento di Fisica, Università di Milano Bicocca, Piazza della Scienza 3, 20126 Milano, Italyd IFAC‐CNR Via Madonna del Piano, 10 50019 Sesto Fiorentino (FI), Italye IASF‐INAF Via Gobetti 101, 40129 Bologna, Italyf Dip. Meccanica e Tecnologie Industriali, Univ. di Firenze Via S. Marta, 3, 50139 Firenze, ItalygCavendish Laboratory, University of Cambridge, JJ Thomson Avenue, Cambridge CB3 0HE, UKh Physics Department, University of Trieste via A. Valerio 2, 34127 Trieste, Italyi Jodrell Bank Centre for Astrophysics, University of Manchester, Macclesfield, SK11 9DL, UK l IEIIT‐CNR, Corso Duca degli Abruzzi 24, 10129, Torino, ItalymIstituto Nazionale di Geofisica e Vulcanologia, Via di Vigna Murata, 605, 00143 Roma, Italyn IASF‐INAF, Via Bassini 15, 20133 Milano, Italyo Agenzia Spaziale Italiana, Viale Liegi 26, 00198 Roma, Italyp OAT‐INAF, Via G.B. Tiepolo 11, 34143 Trieste, Italy
• The Large‐Scale Polarization Explorer is– a spinning stratospheric balloon payload
– flying long‐duration, in the polar night
– aiming at CMB polarization at large angular scales
– using polarization modulators to achieve high stability
• Frequency coverage: 40 – 250 GHz (5 channels)
• Two instruments: SWIPE (here) and STRIP (hear Mario Zannoni)
• Angular resolution: 1.5 – 2.3 deg FWHM
• Sky coverage: 20‐25% of the sky per flight
• Combined sensitivity: 10 μK arcmin per flight
In a nutshell :
Scientific Target• Cosmic Microwave Background photons undergo repeated
Thomson scatterings in the early universe, until the universecools down enough to become neutral.
• Small inhomogeneities of the primeval plasma at the epoch of the last scattering (13.7Gy ago, 380ky after the big bang) induce a small degree of linear polarization in the CMB (E‐modes).
• Gravitational waves produced during inflation, a split‐secondafter the big‐bang, also induce linear polarization in the CMB (both E‐modes and B‐modes)
• Measuring CMB polarization with high precision one can studythe very early universe and the inflation process, happening at energies (>1016 GeV) which cannot be reproduced in the laboratory.
• The signal from primordial B‐modes is extremely small, < 0.1 μKrms and is mainly at large angular scales.
primevalfireball
structures inthe making
with CMB datawe can study
all these phases
B-mod
epo
lariza
tion,
n s
Prim
ary
Anis
otro
py,
E-m
ode
SZ
effe
ct,
lens
ing
Scientific Target• Power spectra (variance versus angular frequency) for the CMB
Anisotropy:Measured very well
Gradient polatization:Measured
Curl polarization:Lensing partrecently detectedby SPT
Primordial part stillto be detectedBB
Scientific Target• Gradient (E‐mode) polarization vs Curl (B‐mode) polarization
E‐modes as detected by Planck B‐modes
Scientific TargetFigure 1. The bottom red line represents the contribution from each multipole to the total mean square fluctuation of the tensor component of CMB polarization (B‐modes, assuming a tensor to scalar ratio r = 1). The bump at small multipoles is due to photons last scattered during reionization, while the second bump is due to photons from z=1100. The thin blue line is the cumulative of the B‐modes, i.e. the variance measured by an experiment sensitive from multipole 2 to a given multipole l. The top blue thick line represents the beam function B2l for an experiment with a 1.5o FWHM Gaussian beam: despite of the coarse angular resolution such an experiment collects most of the polarization signal from B‐modes.
1.5° beam
Primordial B‐modes
cumulative
Instrument configuration• Small signal at large angular scales requires :
– 1.5° : Small telescope aperture sufficient (λ/D) – Wide frequency coverage (foregrounds) : two instruments
– Large sky coverage (cosmic variance) : spinning payload – LDB night flight
– Polarization modulator (to beat systematic effects) : HWP
– Large mapping speed (to reach target survey senstivity)Two ways:• Traditional (EBEX, SPIDER …) : Large number of single‐mode detectors
• SWIPE‐LSPE : Reasonable number of multi‐moded detectors
SWIPESTRIP
Instrument configuration• Small signal at large angular scales requires :
– 1.5° : Small telescope aperture sufficient (λ/D) – Wide frequency coverage (foregrounds) : two instruments
– Large sky coverage (cosmic variance) : spinning payload – LDB night flight
– Polarization modulator (to beat systematic effects) : HWP
– Large mapping speed (to reach target survey senstivity)Two ways:• Traditional (EBEX, SPIDER …) : Large number of single‐mode detectors
• SWIPE‐LSPE : Reasonable number of multi‐moded detectors
Batteries + Electronics(430 Kg)
Star Sensor(40 Kg)
STRIPPolarimeter
(350 Kg)
Al frame(500 Kg)
Azimuth Pivot(100 Kg)
Sketch of the LSPE experiment. The height of the gondola for this configuration is about 4.5m.
STRIPPolarimeter
(350 Kg)Spin3 rpm
Attitude Control & Reconstruction
Fast Star SensorNati et al. Rev.Sci.Inst.74, 9, 2003
Actuators: Azimuth pivot with torque motors
and flywheels;Linear elevation actuators
Processor: PC104 with ADC in / PWM out
H-bridges for motors
Attitude sensors: Star sensor
Laser GyroscopesElevation Encoders
TMTC
• Attitude Control System (ACS): composed of a set of actuators, a processor and a set of attitude sensors.
• Rotates the payload at 3 rpm and controls the elevation of the instrument, to scan the CMB sky and the calibration sources
• Acquires attitude data precise enough to allow sub‐arcmin reconstruction of the pointing for data analysis.
• Is derived directly by systems we have developed for previous balloon missions, like BOOMERanG and Archeops.
• The experiment is flown (2015) as a stratospheric balloon payload during the polar night, in a long duration flight launched from Longyearbyen (Svalbard). See fig.3.
• In this way it can access most of the northern sky in a single flight,
– without contamination from the sun in the sidelobes
– within a cold and very stableenvironment
– Accumulating more than 10 days of integration at float (38 km altitude).
The mission
Figure 3. Bottom: Ground path of a small pathfinder test flight performed in January 2011, in the middle of the polar night. The eastward trajectory is evident.. Top: Launch of a heavy-lift balloon from the Longyearbyen airport (SvalbardIslands, latitude 78oN).
Specific technical problems for a winter polar flight:
– Thermal management. The temperature of the stratosphere during the polar night is around ‐80oC, and there is no solar radiation available to warm‐up the payload. • Electronic systems must be thermally insulated from the environment to achieve self‐heating conditions
• Vacuum seals must be manufactured in indium or in special elastometers• Mechanical actuators play and lubrication must be specially designed for low temperature and low pressure
– Power supply for the experiment and the telemetry: 700 W for 15 days.• Electrical energy storage close to 1 GJ : lithium batteries• Located inside the same insulated box containing the powered electronics, to reach a temperature above 0oC and maintain most of their nominal capacity.
– Data management: LSPE will produce a raw data rate of about 400kbps • Entirely stored on‐board on solid state disks. • Selected data and essential housekeeping / flight info transmitted though the Iridium network
• Line‐of‐sight telemetry at full rate available during the first day of the flight• Data‐dumps when the system flies over selected locations hosting dedicated receiving stations.
• Spin rate : 3 rpm• Latitude: +78° N• Longitude : variable• Elevation range : 30°‐40°• 23% of sky outside WMAP mask
The STRIP Instrument
0.62.2Improvement factor wrt Planck-LFI
6.681.783.773.953.27
Delta Q(U) per 1.5° pixel (micro-K)
104331531731471-s sensitivity (μK×s1/2)
1.1131.8221.3821.8042.068Tsky (antenna) (K)
87< 1Telesc. / window (K)
412033.217.011.3Tnoise (K)
749632N_horn
16.27.712.04.14.5Bandwidth (GHz)
0.4670.467303030Obs Time (months)
1818100100100Sky coverage (%)
0.51.00.220.470.55Resolution (deg)
9043704430Frequency (GHz)
STRIPPLANCK LFI
• An array of coherent polarimeters for accurate measurement of the low‐frequency polarized emission, dominated by Galactic synchrotron. • Its design is described in detail in a companion poster by Bersanelli et al. • In the Table the performance of STRIP is compared to Planck‐LFI.
Hear Mario ZANNONI in a while
The SWIPE Instrument
4.93.22.7Improvement factor wrt Planck-HFI
0.170.130.10--2.60.80.40.3Delta Q(U) (uK)on SWIPE beams
225415660--10152033Integration/beam (s)
1.91.81.9////140453125Channel NET (uK s^1/2)
1108680008888N_det (polarized)
252525333333333333Bandwidth (%)
0.4670.4670.467303030303030Obs Time (months)
202020100100100100100100Sky coverage (%)
7489110555679FWHM Resolution (arcmin)
24514595857545353217143100Frequency (GHz)
SWIPE PLANCK – HFI
• An array of high-throughput bolometric polarimeters to measure accurately the linear polarization of the CMB and of the galactic dust foreground.• Design described in detail in a companion poster by de Bernardis et al. • In the Table the performance of SWIPE is compared to Planck-HFI.
foam window
HWP
HDP Lens
Array 1
Array 2
3He Fridge4He
Instrument SketchBoresight
elevation 20 o-60 o
Foam window
Thermal filters
HWP
UHDPE lens f/1.88, 470 mm Ø+ 440 mm Ø cold stop
L 4He tank (250 liters)
Wire grid polarizer
Reflectedarray
Transmittedarray
3He sorption refrigerator
145 GHz 90 GHz90 GHz
FocalPlaneMap
12o
20o
scan
245 GHz
In second cryostatscan
Compensation on scansNon linearities
TES Correlated thermal drift
AC bias / T stabilization1/f noise
Calibration on anisotropyGain stability
Calibration on anisotropyGain uncertaintyDetectors
Not polarized / orthogonal detectorsPendulation and atmospheric emission
ACS SensorsPointing errorPointing
Large shields, cold stopSidelobes pickup of Earth and Balloon
Large shields, cold stopSidelobes pickup of sky signal
Reduced in multimode system; lab. and flight calibrationMain beam ellipticity
Laboratory calibration / observation of planetsMain beam uncertaintyOptics
Scan with HWP steady, thermal link HWP – cryogenThermal fluctuations of HWP
Lab Calibration / Moon / CrabAbsolute polar angles calibration
Lab. CalibrationCross polar leakage
Scan with HWP steadySlant incidence of rays on HWP
Scan with HWP steady, Spectral bandwidth optimizationDifferential phase shift by HWP
Scan with HWP steady, antireflection coating on HWP Differential reflection of HWP
Scan with HWP steadyDifferential transmission of HWP
Low temperature WG, antireflection coating on HWPWire grid emission, reflected by HWP
Scan with HWP steady, Low temperature HWPHWP emission Polarization
MitigationSystematic effect
Instrument configuration• Small signal at large angular scales requires :
– 1.5° : Small telescope aperture sufficient (λ/D) – Wide frequency coverage (foregrounds) : two instruments
– Large sky coverage (cosmic variance) : spinning payload – LDB night flight
– Large mapping speed (to reach target survey senstivity)Two ways:• Traditional (EBEX, SPIDER …) : Large number of single‐mode detectors
• SWIPE‐LSPE : Reasonable number of multi‐moded detectors
– Polarization modulator (to beat systematic effects) : HWP
High mapping speed
• Two ways:• Traditional (EBEX, SPIDER …) : Large number of single‐mode detectors
• SWIPE‐LSPE : Reasonable number of multi‐moded detectors
• Example: Nm = 25 modes. With respect to a single mode detector: – Increase by a factor 5 the SNR.
– 70 uK sqrt(s) ‐> 12 uK sqrt(s) per detector
– Decrease by a factor 5 the angular resolution
– 15’ FWHM becomes 1.3° FWHM
mNW ∝
mNNET ∝ mN∝θmNNS ∝/2λΩ
≈ANm
SWIPE• The Short Wavelength Instrument for the Polarization Explorer • Uses overmoded bolometers, trading angular resolution for sensitivity• Sensitivity of photon-noise limited bolometers vs # of modes:
Number of modes actually coupling to the bolometer absorber
Light collectors and detectors
Figure 5. Far field calculation of the angular efficiency of a SWIPE Winston horn coupling 17 modes at 145GHz. The lower solid and dashed lines represent two orthogonal cross‐sections for linearly polarized photons. For reference, the solid top line is the beam for unpolarized radiation. In the inset, 2D map of the same beam.
• Baseline detectors: spider‐web bolometers with Transition Edge Sensors (TESs) thermistors:– extreme electro‐thermal feedback, reasonable time
constants, despite of the large absorber area necessary to couple to a multi‐mode beam.
– Rejection of primary cosmic rays present in the stratosphere and potentially very dangerous for CMB measurements.
• Baseline light collector: parabolic Winston horns– very high coupling efficiency in the main beam – high suppression of stray radiation, through the
incoherent superposition of the propagated modes at large angles.
– HFSS + custom‐developed polarized far field calculation code: results in Fig. 5
– Beam ellipticity is very small, a highly desirable feature for this instrument, and will be mitigated even more coupling the horn to the telescope and its Lyot‐stop.
– SWIPE will feature 80 horns at 95 GHz (each coupled to 23 modes), 86 horns at 145 GHz (35 modes), 110 horns at 245 GHz (64 modes). This setup will provide a performance comparable to that of competing instruments based on thousands of single‐mode detectors.
Instrument configuration• Small signal at large angular scales requires :
– 1.5° : Small telescope aperture sufficient (λ/D) – Wide frequency coverage (foregrounds) : two instruments
– Large sky coverage (cosmic variance) : spinning payload – LDB night flight
– Large mapping speed (to reach target survey senstivity)Two ways:• Traditional (EBEX, SPIDER …) : Large number of single‐mode detectors
• SWIPE‐LSPE : Reasonable number of multi‐moded detectors
– Polarization modulator (to beat systematic effects) : HWP
Polarization Modulator• The 50 cm Ø rotating HWP modulator
is the first optical element encounteredby the incoming radiation. This allowsus to relax the specifications for the polarization purity of the telescope and of the radiation collectors.
• The HWP is made of anisotropic metallic grids behaving differently along two orthogonal axes. Cascading many grids with specific geometries it is possible to produce a flat differential phase‐shift of 180° across wide bandwidths. Across a 30% bandwidth the measured cross‐polarisationresulted to be below ‐25dB5.
• The HWP is kept at <4K to limitsystematic effects related tounbalanced emission. The cryogenic rotator is similar to the one we haveproduced for the PILOT experiment6.
20 cm diameter prototype of a mesh‐HWP for the W‐band (left), and a close‐up of the grids (right).
The cryogenic polarization modulator of the PILOT experiment. It works with a power dissipation of a few mW.
LSPE HWP
LSPE HWP
LSPE HWP
• SWIPE is a sky‐scanning Stokes polarimeter, using a rotating half‐wave plate (HWP) and a steady polarizer in front of the detectors to modulate the linear polarization component of the incoming brightness.
• The HWP is rotated by a given angular step (say 22.5°) after a number of full revolutions of the gondola (say every 3 minutes, i.e. after 9 revolutions), and kept steady otherwise.
Measurement Method
Figure 1. Top: Noiseless simulation of the measurement of microwave emission from the CMB, the instrument, and the stratosphere, at 145 GHz (in CMB temperature units) during 4 revolutions of the payload. In this simulation, at the end of each revolution the HWP is stepped by 22o.5. Non‐idealities of the HWP and its thermal emission produce large steps in the detected power. Bottom:the same data once the averages over each revolution and the dipole have been removed. CMB anisotropy is evident. Inset: detail of the signal detected from the same sample sky region for 4 HWP orientations. The small differences are due to the small degree of polarization of the CMB.
Observations and Calibration Plan• For SWIPE, the telescope bore‐sight will be
at a nominal elevation of 40o and can be moved in the range from 20o to 60o , mainly to observe planets for beam calibration and investigate the effect of ground pickup.
• Elevation changes every several hours in a limited range (from 30o to 40o). This, combined with the azimuthal spin, and with Earth rotation, results in a 35% coverage of the sky, of which 25% unmasked. This is suitable for estimation low‐ell multipoles.
• In addition, the rotation of the waveplateresults in a very uniform coverage of the polarimeter angles for all observed pixels.
Sky coverage in a single flight
⎥⎥⎦
⎤
⎢⎢⎣
⎡⎟⎟⎠
⎞⎜⎜⎝
⎛+⎟⎟
⎠
⎞⎜⎜⎝
⎛== ∑∑∑
==
2
1
2
12 2sin2cos1 obsobs
obs
N
ii
N
iipixel
pixelpixel N
WWS αα
Distribution of polarization anglesnon‐uniformity indicator
• SWIPE in flight calibration is based on: (i) absolute calibration of each detector on Planck CMB anisotropies; (ii) Point Spread Function (beam) and time response estimation based on scans over planets, with different scan speed; (iii) polarimeters calibration (efficiency and polarization angle), by observation of the Moon
• The in‐flight polarimeter calibration is used to confirm a more complete ground‐based calibration, obtained illuminating the telescope by means of a polarized source in the far field.
• Scanning strategy: payload spin in azimuth, at 3 rpm (18°/s)
• Coverage of the same sky area by the two instruments
• Elevation changes once a day, at the same time for both instruments
• Specific calibration observations of– Jupiter (to map the main beam, see
figure below, samples = white dots)
– the Crab nebula and the Moon Limb (to calibrate the main axis of the polarimeters)
– the Moon can be used to map sidelobes
Observations and Calibration Plan
24%33%STRIP 30
20%27%STRIP 45
26%35%SWIPE [30‐50]
18%22%SWIPE 45
19%24%SWIPE 35
20%27%SWIPE [40‐50]
23%31%SWIPE [30‐40]
UnmaskedCoverageElevationLSPE coverage for different sets of elevation changes. The first column reports the boresight elevation range in degrees for the two instruments. Second column, the full coverage. Third column, the coverage after masking the galaxy with the WMAP polarization mask.
2.50.80.30.240.0516Uranus
7024971.4-6Saturn
850275100801527Jupiter
185.621.60.300Mars
282322182034Crab
200000070000020000002000003750030Moon
S/N per sampleat 245 GHz
S/N persampleat 145 GHz
S/N per sampleat 95 GHz
S/N per sampleat 90 GHz
S/N per sampleat 44 GHz
Culmination (deg)
Source
Sources culmination angle, and sensitivity for a launch on Jan 1st, 2015. Sampling rate is set at 60 Hz. We assume full Moon, as it is when it is observable by LSPE. The Crab flux is based on the free‐free spectrum reported in Macías‐Pérez, et al. Ap. J., 711, 417 (2010)
Constraining Cosmological Parameters
Constraining Cosmological Parameters
Left: Likelihood for the optical depth to reionization as estimated from Spider3 (dashed line) and SWIPE (continuous line). Right: Same for the tensor to scalar ratio. The two experiments have similar expected performance, but quite different measurement techniques. The estimates have been obtained using the publicly available Markov Chain Monte Carlo package cosmomc as usual in this field. This is based on the direct evaluation of the likelihood in the pixel space, without any power spectrum estimation step, assuming the CMB sky to be gaussian distributed.
Conclusions and Acknowledgments
We gratefully acknowledge support from the Italian Space Agency through
contract I‐022‐11‐0 “LSPE”
PRISM white paperastro-ph/1306.2259
www.prism-mission.org