DevEX - reference for building teams, processes, and platforms
Cosmology with the 21cm line
1. Cosmology with the 21cm line
Gianni Bernardi
SKA SA
(RARG: O. Smirnov, G. Bernardi, T. Grobler, C. Tasse)
Collaborators: (LOFAR-EoR, MWA-EoR, PAPER?)
AIMS, August 14th 2013
2. The 21cm line is ideal to study the first billion years
Dark Ages: no structures were
formed, primordial
fluctuations are imprinted in
the HI gas
Cosmic Dawn: first luminous
structures (Pop III stars?
Micro quasars?) are formed in
the dark matter halos
Reionization (EoR): luminous
structures (galaxies, AGNs) re-
ionize the IGM
5. Observational specs
for 21cm line experiments:
Frequency coverage:
30-200 MHz (6 < z < 35)
Angular resolution:
fluctuations 5 < θ < 30 arcmin you need a radio interferometer
imaging up to < 1 arcmin you need a radio interferometer
Sensitivity:
mK sensitivity is required to constrain most of the HI models
(The VLA @ 74 MHz has an rms sensitivity of 26 K (1 hour))
Challenges:
- correction of ionospheric distortions
- calibration of time and frequency variable telescope response (beam)
- subtraction of bright foregrounds (and their coupling with the instrumental response)
21. We live in the era of exploration: current and future 21 cm
experiments
GMRT
LOFAR
PAPER
MWA
HERA - SKA
22. GB et al. 2009
~2.3 arcmin resolution
frequency: ~150 MHz
peak flux ~ 2.8 Jy
conversion factor:
1mJy/beam=4 K
noise: 0.75 mJy/beam
The key point to detect the 21cm signal is how well
foregrounds can be removed! What do foregrounds look like?
23. GB et al. 2009
Statistical properties of foregrounds
* 2noise
180
Y
l b
C X l X l b
N N
Power law behavior with best fit
amplitude
A400= (0.0019 0.0003) K2,
and best fit slope βI = -2.2 0.3:
diffuse Galactic emission
Flat power spectrum:
residual point sources
Power law behavior with best fit
amplitude A700= (90 7) (mK)2
and best fit slope βIp = -1.52 0.16:
diffuse polarized Galactic
emission
25. Do we know how to subtract foregrounds? How well?
• Subtraction of Galactic diffuse emission and extragalactic radio sources:
they are supposed to have smooth spectra compared to the 21 cm signal;
Bowman, Morales & Hewitt 2009
26. Do we know how to subtract foregrounds? How well?
• Subtraction of Galactic diffuse emission and extragalactic radio sources:
they are supposed to have smooth spectra compared to the 21 cm signal;
EoR + FG + noise
Eor + noise
EoR ~ 5 mK
FG ~ 2 K
noise ~ 50 mK
How well does it work on data?
Jelic, .., GB, et al. 2008
27. • An interferometer never samples all the Fourier modes PSF sidelobes corruption
(k┴,k║);
• Instrumental frequency response corrupts the foreground frequency smoothness (k║);
• Telescope beams change with frequency and pointing direction (dipoles do not track
the sky) and they can be different from each other (k┴,k║);
• The ionosphere is no longer transparent (time and frequency dependent distortion &
refraction) (k┴,k║);
• RFI corrupts the sky signal (mostly ,k║);
• Real foreground polarized signal can leak into total intensity due to polarized beams
a/o imperfect polarization calibration (k║);
The point is that instrument calibration and foregrounds are
coupled foreground properties are corrupted by the
instrumental response
32. Giant Metrewave Radio Telescope (GMRT)
• Large collecting area;
• Clever calibration strategy
(pulsar on–off);
• Stable and known beams;
• Small field of view;
• Severe RFI problems;
Paciga et al. 2011, 2013
33. (Low Frequency ARray) LOFAR
• Largest collecting area;
• Complex elements (levels of
dipole clustering) element
beams are inherently different
from each;
• Small field of view;
• Active RFI environment
(mitigated by high time and
frequency resolution);
35. Murchison Widefield Array (MWA)
• Centrally condensed core to
maximize power spectrum
sensitivity for the EoR (but
smaller collecting area);
• Large field of view;
• Minimum RFI contamination;
• Analog signal paths;
courtesy A. Offringa
36. Upper limits on the EoR at z~8.5 from the 32T prototype
Dillon, …, GB, et al. 2013
Δk < 0.26 K @ 95% c.l.
37. Precision Array to Probe the Epoch of Reionization (PAPER)
• Maximum redundant configuration
(baselines length are equal to each
other as much as
possible), optimized for EoR
power spectrum measurement;
• The simplest design (beam
stability, smoothness, minimal
ionospheric impact);
• Very large field of view;
• Minimum RFI contamination;
• Analog signal paths;
• Smallest collecting area;
38. Upper limits on the EoR from PAPER 32
WSRT (GB et al., 2010)
Courtesy J. Pober
39. What have we learned about reionization from current 21cm
measurements?
Xi ~ 0.5
The IGM must have been heated by X-rays (MXRBs a/o quasars)
40. Hydrogen Epoch of Reionization Array:
HERA-576
http://reionization.org
41. Conclusions
• The redshifted 21cm line promises to be a fantastic probe of the high-z Universe;
• Steady progress towards the first detection of HI at z > 6;
• Many challenges still to be overcome (calibration, foreground subtraction) –
development required!;
• The detection will open up the field for a characterization of the EoR and Dark Ages;
• SKA low and HERA looking ahead;
• Observations of the global sky signal represent a way to probe the cosmic dawn at z
~ 25-30;
• HI intensity mapping! BAOs at 0.1 < z < 6
Notas do Editor
PSA-64, 32 antennas in a maximum-redundancy configuration, 30m between rows, 4m spacing within rows