Detecting the B-mode Polarisation of the CMB with Clover

Abstract

We describe the objectives, design and predicted performance of Clover, which is a ground-based experiment to measure the faint ``B-mode'' polarisation pattern in the cosmic microwave background (CMB). To achieve this goal, clover will make polarimetric observations of approximately 1000 deg^2 of the sky in spectral bands centred on 97, 150 and 225 GHz. The observations will be made with a two-mirror compact range antenna fed by profiled corrugated horns. The telescope beam sizes for each band are 7.5, 5.5 and 5.5 arcmin, respectively. The polarisation of the sky will be measured with a rotating half-wave plate and stationary analyser, which will be an orthomode transducer. The sky coverage combined with the angular resolution will allow us to measure the angular power spectra between 20 < l < 1000. Each frequency band will employ 192 single polarisation, photon noise limited TES bolometers cooled to 100 mK. The background-limited sensitivity of these detector arrays will allow us to constrain the tensor-to-scalar ratio to 0.026 at 3sigma, assuming any polarised foreground signals can be subtracted with minimal degradation to the 150 GHz sensitivity. Systematic errors will be mitigated by modulating the polarisation of the sky signals with the rotating half-wave plate, fast azimuth scans and periodic telescope rotations about its boresight. The three spectral bands will be divided into two separate but nearly identical instruments - one for 97 GHz and another for 150 and 225 GHz. The two instruments will be sited on identical three-axis mounts in the Atacama Desert in Chile near Pampa la Bola. Observations are expected to begin in late 2009.

Full text

arXiv:0805.3690v2 [astro-ph] 3 Jun 2008
Detecting the B-mode Polarisation of the CMB with C
ℓ
over
C. E. North
1
, B. R. Johnson
1
, P. A. R. Ade
2
, M. D. Au dley
3
, C. Baines
4
, R. A. Battye
4
, M. L. Brown
3
,
P. Cabella
5
, P. G. Calisse
2
, A. D. Challinor
6,7
, W. D. Duncan
8
, P. G. Ferreira
1
, W. K. Gear
2
, D. Glowacka
3
,
D. J. Goldie
3
, P. K. Grimes
2
, M. Halpern
9
, V. Haynes
4
, G. C. Hilton
8
, K. D. Irwin
8
, M. E. Jones
1
, A. N. Lasenby
3
,
P. J. Leahy
4
, J. Leech
1
, B. Maffei
4
, P. Mauskopf
2
, S. J. Melhuish
4
, D. ODea
3
, S. M. Parsley
2
, L. Piccirillo
4
,
G. Pisano
4
, C. D. Reintsema
8
, G. Savini
2
, R. Sudiwala
2
, D. Sutton
1
, A. C. Taylor
1
, G. Teleberg
2
, D. Titterington
3
,
V. Tsaneva
3
, C. Tucker
2
, R. Watson
4
, S. Withington
3
, G. Yassin
1
, J. Zhang
2
1
Astrophysics, University of Oxford, Oxford, UK
2
School of Physics and Astronomy, Cardiff University, UK
3
Cavendish Laboratory, University of Cambridge, Cambridge, UK
4
School of Physics and Astronomy, University of Manchester, UK
5
Dipartimento di Fisica, Univ ersit`a di Roma Tor Vergata, Italy
6
Institute of Astronomy, University of Cambridge, UK
7
DAMTP, University of Cambridge, UK
8
National Institute of Standards and Tec hnology, USA
9
University of British Columbia, Canada
We describe the objectives, design and predicted performance of C
ℓ
over, which is a ground-based
experiment to measure the faint “B-mode” polarisation pattern in the cosmic microwave background
(CMB). To achieve this goal, C
ℓ
over will make pol arimetric observations of approximately 1000 deg
2
of the sky in spectral bands centred on 97, 150 and 225 GHz. The observations will be made with
a two-mirror compact range antenna fed by profiled corrugated horns. The telescope beam si zes for
each band are 7.5, 5.5 and 5.5 arcmin, respectively. The polarisation of the sky will be measured with
a rotating half-wave plate and stationary analyser, which will be an orthomode transducer. The sky
coverage combined with the angular resolution wi ll allow us to measure the angular power spectra
between 20 < ℓ < 1000. Each frequency band will employ 192 single polarisation, photon noise limited
TES bolometers cooled to 100 mK. The background-limited sensitivity of these detector arrays will
allow us to constrain the tensor-to-scalar ratio to 0.026 at 3σ, assuming any polarised foreground
signals can be subtracted with minimal degradation to the 150 GHz sensitivity. Systematic errors will
be mitigated by modulating the polarisation of the sky signals with the rotating half-wave plate, fast
azimuth scans and periodic telescope rotations about its boresight. The three spectral bands will be
divided into two separate but nearly identical instruments — one for 97 GHz and another for 150 and
225 GHz. The two instruments will be sited on identical three-axis mounts in the Atacama Desert in
Chile near Pampa la Bola. Observations are expected to begin in late 2009.
1 Introduction
The currently favoured cosmological model predicts that gravity waves produced during a period of cosmo-
logical inflation should have imprinted a faint primordial “B-mode” polarisation pattern in the CMB
1,2
.
The amplitude of the gravity-wave signal is related to the expansion rate, and hence energy scale
3
, during
inflation and is therefore not yet predicted by fundamental theory. By precisely characterising the polar-
ization of the CMB, it should be possible to put strong constraints on the various theories of inflation. A
discovery of the primordia l B-mode signal wo uld be a significant breakthrough for both astrophysics and
particle physics, since we would then be probing physics at the 10
16
GeV scale.
To date, the characterisation of the CMB temperature anisotropy has given the prevailing ΛCDM cos-
mological model a s trong footing
4,5
. This foo ting was further strengthened when the brighter “E-mode”
polarisation signal, which the prevailing model predicted, was recently measured
6,7
. The anticipated B-mode
signal has not yet been discovered because the current generation of instrumentation is not sensitive enough
to detect it.
In this paper, we describe C
ℓ
over , which is a next-generation CMB ex periment that will have sufficient
sensitivity to measure the T and E-mode signals and, by searching for a cosmological B-mode signal, measure
or constrain the tensor-to-scalar ratio
a
down to r = 0.026. This is a factor of 20 lower than the current upper
limit from CMB measurements alone
8
. Therefore, the core astrophysical go als of C
ℓ
over are the following:
a
The tensor-to-scalar ration, r, is proportional to the energy scale of inflation to the fourth power
3
.

Figure 1: Theoretical angular power spectra for the temperature, E-mode and B -mode signals assuming the ΛCDM cosmological
model
9
and r = 0.1. The black, B-mode curve comprises a primordial inflationary gravity wave signal and a predicted non-
primordial B-mode for eground signal produced by the gr avitational lensing of partially polarised CMB radiation as it passed
through large-scale structures. The expected performance of C
ℓ
over is overplotted as binned points, with error bars taking
into account the atmospheric and instrumental noise for 150 GHz and the sample variance for the C
ℓ
over survey area. For
comparison, we plot the power spectra of the forecasted foreground signals in the C
ℓ
over observation regions before cleaning,
based on polarisation observations and models of unpolarised emission
10,11,12
. These foreground signals include polar ised
thermal emiss ion f rom aligned dust grains and polarised synchrotron emission. The dust and s ynchrotron signals are plotted
at 225 and 97 GHz, respectively, to show the worst case scenario. Both signals should be smaller at 150 GHz.
to determine the energy scale of inflation, to improve constraints on a collection of co smological parameters,
to measure a non-primordial gravitaional lensing-induced B-mode signal, and to precisely characterise any
polarised Galactic signals such as synchrotron radiation and polarise d emission from dust. An ove rview of
C
ℓ
over is g iven in Section 2, while instrument a nd obse rvation details are given in Sections 3 & 4.
2 The C
ℓ
over Experiment
During two years of operation, C
ℓ
over will make polarimetric observations of approximately 1000 deg
2
of the
sky from Pampa la Bola in the Atacama Desert, Chile, which is within the ALMA science preserve. This
location was selected because the millimetre-wave transmittance of the atmosphere at the site is amo ng the
best in the world for ground-based observatories, primarily due to the high altitude of 4.9 km. The polari-
metric observations will be ma de in spectral bands centred on 97, 1 50 and 225 GHz with two independent
instruments, one for 97 GHz and another for 150 and 225 GHz. The three frequency bands, primarily set
by atmospheric transmission windows, are well matched to the 2.7 K CMB blackbody spectrum and are
exp ected to provide sufficient leverage for the spectral removal of the anticipated astrophysical foreground
signals. The baseline design of the experiment calls for the use of 5 76 superco nducting transition edge sensors
(TES), cooled to 100 mK with a pulse tube cooler, a He-7 refrigerator and a dilution refrigerator
13
. The
large number of detector outputs will be multiplexed with a time-domain cryogenic readout. The receivers
containing the detectors will be mounted at the focal planes of low cross-polarisation, off-axis compact
range antenna reflecting telescopes. The 97 GHz telescope will provide 7.5 arcmin beams on the sky, while
the 150/2 25 GHz telescope will provide 5.5 arcmin beams. The expected instrument NET ≃ 18 µK
√
sec at
150 GHz gives a predicted map sensitivity of around 1.7 µK per 5.5 arcmin re solution element for the Q and U
Stokes parameters after a two-year observing campaign. The beam s izes are well matched to the primordial

B-mode signal, which dominates the non-primordial B-mode signal below ℓ ≃ 100 for r & 0.01, and will
allow very good sensitivity to the non-primordial B-mode signal from gravitational lensing (see Figure 1).
To mitigate the effect of 1/f noise from detector drifts and atmospheric fluctuatio ns, to reject systematic
errors and to achieve the best noise performance, C
ℓ
over will use rapid polarisation modulation. This modu-
lation is produced by a rotating achromatic half-wave pla te and fixed orthomode transducers, which are the
polarisation analysers. Because the B-mode signal is faint, the instrument design outlined above has been
advised by detailed cos mological simulations to e nsure our results should not be contaminated by spurious
instrumental signals
14
. The expected perfo rmance of C
ℓ
over is shown in Figure 1, and a model of one of the
C
ℓ
over instruments is shown in Figure 2.
3 Instrument Description
Both the C
ℓ
over ins truments will use a compact ra nge antenna (CRA), which is composed of a parabolic
primary mirror and hyperbolic secondary mirror. Both mirrors are off axis. This optical design gives excellent
cross-polarisation performance and low aberrations across a large, flat focal plane
15
. For all focal plane
elements, the Str e hl ratio is greater than 0.95, and the cross polarisation — including the cross polarisation
of the horn — is better than −38 dB. The projected diameters of the primary mirrors are 1.8 and 1.5 m for
the 97 and 15 0/225 GHz instruments respectively. The telescope mirrors will be surrounded by a co-moving
baffle lined with millimetre-wave absorber, which will prevent signals in the far side lobes of the telescope
from being modulated as the telescope sc ans.
Each array element in the foca l plane comprises a profiled corrugated horn, an ortho mode transducer
(OMT) and two (TES) detectors. The TES detectors are photon noise limited Mo/Cu super c onducting
bolometers with transition temperatures around 19 0 mK (430 mK for 225 GHz) and time constants around
0.5 ms
16,17
. In the 97 GHz focal plane the OMT is composed of an electroformed turnstile junction with
a circular waveguide input and two rectangular waveguide o utputs
18
. Each output waveguide terminates
in a microstrip-coupled TE S via a finline transition
19
. At 150 and 225 GHz each horn couples to a planar
OMT composed of four rectangular probes in a cylindrical waveguide
20
. Outputs from opposing pairs of
probes are combined in a microstrip circuit, which terminates in the TES; this design combines the OMT
and detectors on a single chip. The focal plane elements are grouped into eight-element modules, and each
module is read o ut with a SQUID multiplexer
21,22,23
.
A single achromatic half-wave plate (AHWP) will b e mounted approximately 100 mm in front of each
fo c al plane
24,25
. The AHWP is composed of three A-cut sapphire discs at 97 GHz and five discs at
150/225 GHz. The diameter of each disc is approximately 300 mm, and the disc thicknesses are 4.65 mm and
2.43 mm at 97 and 150 /225 GHz respectively. The 97 GHz AWHP central disc has a birefringent axis oriented
at 59
◦
relative to the outer discs, while the second to fifth crystals in the 150/225 GHz stack are rotated by
29
◦
, 95
◦
, 29
◦
and 0
◦
respectively with r e spect to the first crystal in the stack. To improve transmission and
to minimise spurious polarisation from differential reflection, the AHWP will have a three-layer, broadband
anti-reflection coating on the external faces of the stack.
The cryostat and optics are mounted to a c ommon optical assembly, which sits on top of a three-axis
mount as shown in Figure 2. T he third axis, rotation around the boresight, is necessary for identifying and
suppressing systematic errors from instrumental polarisation and for increasing cross-linking.
4 Observation Strategy
C
ℓ
over will observe a total of 1000 deg
2
of the sky. This coverage is divided into the four regions shown
in Figure 3, each roughly 20
◦
in diameter. Two of the fields are in the southern sk y and two lie along
the equator. To c ontrol the contribution from the atmosphere each telescope will scan at approximately
0.5 deg/sec in azimuth at a constant elevation. Every few hours, the elevation angle of the telescope will be
re-pointed to allow for field tracking. During all observations, the AHWP will be rotated continuously at
approximately 5 Hz and the telescope will be rotated around its boresight a xis periodically. Observations at
97 GHz are expected to begin in late 2009, with 150/225 GHz observations expected from mid-20 10.

Figure 2: A model of one of the two C
ℓ
over instruments. The three-axis mount also allows the rotation of the entire optical
assembly around the telescope boresight. The mirrors are held inside a co-moving baffle (translucent in the figure f or clarity).
The co-moving baffle is lined with absorber to reduce the effect of side lobes. A counter balance, which is not shown here, will
be mounted on the boresight axis on the opposite side of the elevation axis from the telescope. All of the instrument hardware
shown here is either built or under construction.

Figure 3: The four observation regions selected for C
ℓ
over, centred on 0430−40, 2230−45, 0900+00 and 1400+00, plotted in
equatorial coordinates over a model of the magnitude of the polarised Galactic emission at 225 GHz. The field locations were
chosen based on models of the Galactic emission
11,12
, and are optimally distributed in RA. These fields are intended to overlap
nicely with the survey areas of other B-mode experiments. Also plotted are the zenith declination (dashed) from Pampa la
Bola (22
◦
28
′
S, 67
◦
42
′
W) and the ecliptic plane (dotted). Grey ar eas never rise above 45
◦
from the C
ℓ
over site, which means
they are not useful or accessible.
Acknowledgement
C
ℓ
over is funded by the Science and Technology Facilities Council. CEN acknowledges an STFC studentship.
BRJ acknowledges an STFC postdoctoral fellowship, and an NSF IRFP fellowship. Some of the re sults in
this paper have been derived using the HEALPix
26
package.
References
1. Hu, W., Hedman, M. M., and Zaldarri aga, M. (2003) Phys. Rev. D 67(4), 043004.
2. Seljak, U. and Zaldarriaga, M. (1997) Phys. Rev. Lett. 78(11), 2054.
3. Knox, L. and Song, Y.-S. (2002) Phys. Rev. Lett. 89(1), 011303.
4. Spergel, D. N. et al. (2007) ApJS 170, 377.
5. Hinshaw, G. et al. (2008) ArXiv e-prints 0803.0732.
6. Ade, P. et al. (2008) ApJ 674, 22.
7. Bischoff, C. et al. (2008) ArXiv e-prints 0802.0888.
8. Dunkley, J. et al. (2008) ArXiv e-prints 0803.0586.
9. Komatsu, E. et al. (2008) ArXiv e-prints 0803.0547.
10. Page, L. et al. (2007) ApJS 170, 335.
11. Giardino, G. et al. (2002) A&A 387, 82.
12. Finkbeiner, D. P. , Davis, M., and Schlegel, D . J. (1999) ApJ 524, 867.
13. Teleberg, G., Chase, S. T., and Piccirillo, L. (2008) J. Low Temp. Phys. 151(3-4), 669.
14. O’Dea, D., Challinor, A., and Johnson, B. R. (2007) MNRAS 376(4), 1767.
15. Johnson, B. R. et al. (2008) MNRAS , in preparation.
16. Audley, M. D. et al. (2008) Proceedings of the SPIE 7020, in press.
17. Audley, M. D. et al. in Proceedings of the 19th Int. Symp. on Space THz Tech., (2008). in press.
18. Pisano, G. et al . (2007) IEEE Microwave Compon. Lett. 17(4), 286.
19. North, C. E. et al. (2008) MNRAS , in preparation.
20. Grimes, P. K. et al. (2007) Elect ronic Letters 43(21), 1146.
21. Battistelli, E. S. et al. (2008) J. Low Temp. Phys. 151, 908.
22. Reintsema, C. D. et al. (2003) Rev. Sci. Instrum. 74(10), 4500.
23. de Korte, P. A. J. et al. (2003) Rev. Sci. Instrum. 74(8), 3807.
24. Pisano, G. et al . (2006) App. Opt. 45(27), 6982.
25. Savini, G., Pisano, G., and Ade, P. A. R. (2006) App. Opt. 45(35), 8907.
26. G´orski, K. M. et al. (2005) ApJ 622, 759.