29 June 2006 Bernard's Cosmic Stories 1
1 st results from the QUaD CMB polarisation experiment
Michael Brown (University of Edinburgh)
Polarisation of the CMB
Temperature
Polarisation Temperature
Polarisation Matrix: P = Q + U
29 June 2006 Bernard's Cosmic Stories 3
The E/B Decomposition
Can decompose Q & U into:
E-modes (even-parity):
(or grad)
B-modes (odd-parity):
(or curl)
Density perturbations produce only E-modes.
Gravitational waves produce both E & B-modes.
B E
B E
Cold Spot: Hot Spot:
The CMB Power Spectra
Have 4 possible spectra: TT, TE, EE, BB (TB = EB = 0).
Primary effects Secondary effects
Reionisation
Gravitational Lensing Diffusion
Damping Sachs-Wolfe
Acoustic Oscillations
Gravitational Wave `Bump’
QUaD l-range
29 June 2006 Bernard's Cosmic Stories 5
The QUaD experiment
QUaD is the QUEST telescope installed on the DASI mount at the South Pole.
2.6 m primary mirror.
31 pixel polarisation sensitive bolometer (PSB) camera.
100GHz and 150GHz observing frequencies.
Secondary supported on foam cone.
DASI infrastructure re-used for QUaD.
Extended ground shield.
QUaD people
Cardiff: Peter Ade, Walter Gear, Simon Melhuish, Angiola Orland, Lucio Piccirillo, Nutan Rajguru, Mike Zemcov.
Caltech: Andrew Lange, Jamie Bock, John Kovac, Ken Ganga (Paris).
Chicago: John Carlstrom, Tom Culverhouse, Robert Friedmann, Eric Leitch (JPL), Clem Pryke, Robert Schwarz (South Pole).
Edinburgh: Michael Brown, Patricia Castro, Andy Taylor
Maynooth: Gary Cahill, Anthony Murphy, Fabio Noviello, Creidhe O’Sullivan.
Stanford: Melanie Bowden, Sarah Church, Jamie Hinderks, Ben Rusholme, Keith Thompson, Ed Wu.
29 June 2006 Bernard's Cosmic Stories 7
QUaD in extended ground shield:
Focal plane:
12 feeds @ 100GHz (6 arcmin), 19 feeds at 150GHz (4 arcmin).
29 June 2006 Bernard's Cosmic Stories 9
1 st season’s observations
Each hour of observations is split between a lead and trail field – separated by 0.5 hrs (7.5°) in RA – exact same scan pattern with respect to ground.
Two 8-hour CMB runs/day: 2
ndrun repeats same scan
pattern as 1
stwith telescope rotated 60° about line of sight axis (deck angle rotation).
Relative calibration from source (RCW38) + “el-nods”
(small el scan to inject atmospheric ramp).
99 days of CMB data taken in 1
stseason covering a 10°×6°
patch of the B03 (low-foreground) deep field region.
Beams measured from RCW38
Used to construct a T-dependent
beam model for each detector.
29 June 2006 Bernard's Cosmic Stories 11
1 st year T maps:
100 GHz: 150 GHz:
. Inverse-variance weighted maps.
3
rdorder polynomial removed from each az-scan.
1 st year Q/U maps at 150GHz:
Smoothed at scale ~5 arcmin in attempt to bring out structure.
Q: U:
29 June 2006 Bernard's Cosmic Stories 13
T, Q & U jacknife maps:
100 GHz T
150 GHz U 150 GHz Q
150 GHz T
Field differencing
Difference lead & trail fields to remove possible ground signal (sensitivity hit: S/N ↓ by √2).
100 GHz T 150 GHz T
150 GHz Q 150 GHz U
29 June 2006 Bernard's Cosmic Stories 15
Simulations:
Using Monte-Carlo based power spectrum estimator so need to simulate the experiment extremely accurately:
Measure auto- and cross-spectra of time-ordered data (TOD) for each pair of PSBs. Using these, inject correlated noise into simulations in fourier space.
Add a CMB signal convolved with a temperature-dependent beam model measured for each bolometer.
Process simulated TOD in exact same way as the real data.
Simulations → noise bias, beam/filtering transfer functions, errors & covariances.
' '
' 1
' ~ ~ ) /
( l l l
ll C N F
M
C l
Simulated maps:
100 GHz T
150 GHz U 150 GHz Q
150 GHz T
29 June 2006 Bernard's Cosmic Stories 17
Jacknife power spectra – 150 GHz real /sims
Weiner filtered E and B maps
29 June 2006 Bernard's Cosmic Stories 19
