Observations of the SZ with the CBI2

1
Observations of the SZ with the CBI2
James Allison Oxford University Angela
Taylor Tony Readhead Steve Myers Mike
Jones Tim Pearson Brian Mason Steve
Rawlings Clive Dickinson (NRAO) (Oxford
University) Joey Richards Rodrigo Reeves Peter
Wilkinson Larry Weintraub Ricardo Bustos Scott
Kay Kieran Cleary Nolberto Oyarce Richard
Davis Martin Shepherd (Astro-Norte/U de
Concepcion) Rod Davies (Caltech/JPL) Dick
Bond (Manchester University) Simon Cassasus Jon
Sievers Glenn White (U de Chile) (CITA) (RAL)

2
The Cosmic Background Imager

• CBI1 (1999 2005)
• Interferometer located at the Chajnantor
  Observatory, Antofagasta, Chile
• 13 Element co-mounted array of 90 cm dishes
• Ka-Band (26-36GHz)
• Key Science results
• CMB Power spectrum CBI Excess
• CMB EE Polarisation
• Pointed SZ cluster observations
• Galactic observations

3
CBI2 The Upgrade

• Antenna upgrade
• 13 x 1.4m dishes
• Low Spill-over design
• 28.2 arcminute field-of-view
• at 31 GHz
• New Array
• Improved sensitivity on long baselines
• Good for high-l CMB
• Less contamination to SZ from low-l CMB
• Corresponds to angular scales between 5 and 25
  arcminutes, similar to the virial size of a
  cluster at redshifts 0.1-0.3

4
CBI2 The Science

• Observations from 2006-2008
• Key Science goals
• Follow up observations of the CMB power spectrum
  excess at high-l values
• SZ scaling relations from observations of known
  X-ray samples taken from the REFLEX survey
• Study of SZ cluster physics out to the virial
  radius
• The spectral SZ from joint observations with
  SuZIE
• Observations of galactic sources e.g. Planetary
  Nebulae, Supernova Remnants, HII Regions

5
The Sunyaev-Zeldovich Effect

• Brief over-view
• The result of inverse-Compton scattering of CMB
  photons off of the hot intra-cluster medium
• The net effect is increase in photon energy,
  leading to a shift to higher frequencies
• At 31 GHz is seen as a decrement in measured CMB
  brightness in the direction of a cluster
• Proportional to the integral of the line-of-sight
  pressure of the ICM (y) multiplied by a frequency
  dependant co-factor (f)

6
The Sunyaev-Zeldovich Effect
7
The X-ray Cluster Samples

• The clusters were chosen from known X-ray samples
  to cover a range of redshifts and luminosities
• They will be used to calibrate the scaling
  relations between SZ and X-ray properties and
  investigate any evolution with redshift
• In total 40 clusters have been observed with
  CBI2, some of which are shown in the maps to the
  right
• (collaborators S Kay, H Boehringer, G Pratt, E
  Pointecouteau)

8
REXCESS Sample, z0.15(Boehringer et al. 2007)

• Flux limited sample of 33 X-ray clusters
• CBI2 has observed the brightest 1/3 of the
  sample
• Lx range 3 - 14 x1044 erg s-1 (in the 0.1 2
  keV band)

9
REFLEX-DXL Sample, z0.3(Zhang et al. 2006)

• Volume limited sample of brightest, distant
  clusters in REFLEX
• CBI2 observed all but 2 out 13 of the sample
• Lx range 6.5 - 21 x1044 erg s-1 (0.1 2 keV)

10
XMM High-redshift sample, z0.5(M. Arnaud PI)

• Complete X-ray flux limited sample
• Bright, distant clusters
• Lx range 20 55 x 1044 erg s-1 (0.1 2 keV)
• Completes redshift range from 0.15 to 0.5

11
Joint SuZIE sample (Benson et al. 2004)

• CBI2 has observed the SuZIE sample from Benson et
  al. 2004, in collaboration with Sarah Church
  (Stanford)
• Combined analysis of 30 GHz CBI2 and 150, 220,
  270 and 350 GHz SuZIE data
• Allows improved SZ spectra and Cluster peculiar
  velocities

12
Interesting SZ candidates

• In addition to the above samples, CBI1 and CBI2
  have observed interesting individual SZ
  candidates
• These include
• The Shapley supercluster
• XMMU J2235.2577 (z1.393) high redshift SZ
  candidate
• 2df 1435008, a possible lensed quasar pair
• Data from CBI2 will significantly reduce the
  uncertainty due to the CMB on the longer baselines

Shapley supercluster CBI-SZ (top) and
X-ray(bottom)
13
Data Reduction

• Calibration
• Primary Calibrator e.g. Jupiter or TauA
• Phase Calibrator near to source
• Internal noise source calibration
• Ground Signal removal
• Low elevation data suffers from spillover of
  ground signal into primary beam
• This is removed from SZ data by subtracting
  nearby reference fields at the same declination
• These reference fields are ideally blank, with
  only a CMB component
• Point source removal
• Observe at 30 GHz positions of known point
  sources (e.g. from NVSS) with ATCA
• Project out point source flux error at model
  fitting stage

14
Initial results and maps

• Sub-sample of 8 clusters from the REFLEX-DXL
  sample
• Only Bullet cluster has had source subtraction
  applied using 30 GHz ATCA fluxes of known SUMSS
  sources
• 30 GHz point source fluxes to be obtained from
  ATCA in September for rest of REFLEX-DXL sample

15
Initial results and maps

• Sub-sample of 8 clusters from the REFLEX-DXL
  sample
• Only Bullet cluster has had source subtraction
  applied using 30 GHz ATCA fluxes of known SUMSS
  sources
• 30 GHz point source fluxes to be obtained from
  ATCA in September for rest of REFLEX-DXL sample

16
Initial results and maps
RXJ0658.5-5556 RXJ0232.2-4420 The Bullet
Cluster
Scale of Chandra image
17
Initial results and maps
RXJ0658.5-5556 RXJ0232.2-4420 The Bullet
Cluster
Blue Thermal noise Red Total uncertainty
(Noise, CMB, Sources)
18
Data Gridding

• The calibrated visibility data is then gridded
  onto a regular grid using MPIGRIDDR (Myers et al.
  2003)
• Resulting output includes gridded visibilities
  and covariance matrices for the thermal noise,
  CMB component (from input power spectrum) and
  possibly any point sources
• Real Imaginary

19
Data Gridding

• The calibrated visibility data is then gridded
  onto a regular grid using MPIGRIDDR (Myers et al.
  2003)
• Resulting output includes gridded visibilities
  and covariance matrices for the thermal noise,
  CMB component (from input power spectrum) and
  possibly any point sources

Real Imaginary
Blue Thermal noise Red Total uncertainty
(Noise, CMB, Sources)
20
Analysis

• Parameterised models of the SZ are then converted
  into mock visibility data and compared with the
  gridded visibilities
• For multivariate gaussian noise the likelihood is
  given by
• Where C is the covariance matrix equal to the sum
  of the thermal noise, CMB and point source
  components
• The posterior is explored for a given SZ model
  and data set using an MCMC engine BAYESYS (John
  Skilling)
• Output includes parameter probability
  distributions, and the evidence (ie integrated
  likelihood) for a given SZ model

21
Initial Results

• RXCJ0658.5-5556 The Bullet Cluster

Lbol 4.87 /- 0.24 x 1045 erg s-1 Y500 4.48
/- 0.51 Mpc2
22
A Physical SZ model

• Isothermal-Beta model is inadequate to describe
  cluster physics out to the virial radius (e.g.
  Zhang et al. 2008, Pratt et al. 2006, Vikhlinnin
  et al. 2006)
• Therefore when combining X-ray and CBI2-SZ data
  we require a more physically based model
• Assume underlying matter distribution (NFW)
• Use entropy model consistent with shock accretion
  and entropy floor from theory and X-ray
  observations (e.g. Ponman et al. 2003)

23
A Physical SZ model

• Assume hydrostatic equilibrium holds for suitably
  relaxed gas on large scales
• So we get the following pressure dependence as a
  function of radius
• And hence SZ temperature as an integral of
  pressure along the line-of-sight
• Input parameters r200, c200 (NFW), n0, T0, rcore
  and alpha (Gas)

24
A Physical SZ model

• Given sensible parameter values for this model,
  the scaled properties for a typical massive
  cluster would be

• Radius r500 1.41Mpc
• Electron Density
• n0 1x10-2cm-2 (Solid)
• Electron Temperature
• T0.2-0.5r500 9.05 keV (Dashed)
• Total Mass
• M500 1.10x1015 Msolar
• (Dash-Dotted)
• fgas
• Scaled to fbaryon (Dotted)

25
A Physical SZ model

• This is consistent with the results of recent
  X-ray temperature profiles
• Solid line Model
• Dashed line Isothermal
• Grey regions
• Scaled X-ray temperatures
• (Zhang et al. 2008,
• Pratt et al. 2006,
• Vikhlinnin et al. 2006)

26
A Physical SZ model

• Fit our model to simulated CBI2 visibility data
  constructed from hydrodynamical/N-body
  simulations (Kay et al. 2005)

Simulated y-map
Simulated CBI2 map
Histogram for Y200 Dashed line Simulation Value
27
Summary

• Can measure SZ out to the largest cluster radii
• CBI2 has observed 40 clusters over a 2 year
  period
• Currently reducing the data to construct complete
  samples for X-ray vs SZ scaling relations
• Investigating a physically based model for
  exploring global cluster properties and profiles
  out to the virial radius