Imaging the SZ Effect in Galaxy Clusters

Holzapfel Group | ACBAR | APEX | BICEP | BIMA Anisotropy | CBI | DASI | SZ Clusters | South Pole 'scope | UCB Physics
SZ Imaging Home


Instrument Description

Observations and Results

Project Team



Observations and Results

BIMA Observations

All SZ observations were made using the BIMA array in Hat Creek, CA, and the OVRO array in Big Pine, CA. Nine 6.1 m BIMA telescopes were equipped for operation at 28.5 GHz, providing a 6.6' FWHM field of view. Six 10.4 M OVRO telescopes were equipped for operation at 28.5 GHz, providing a 4.0' FWHM field of view. All field observations were bracketed by observations of bright radio point sources in order to track the phase fluctuations caused by the instrument and the atmosphere.

Treatment of Radio Point Sources

Emission from radio point sources at 28.5 GHz will contaminate SZ observations. To discriminate between the extended emission from a cluster of galaxies and the point source emission from a radio source, a hybrid array is used for the observations. The majority of the telescopes are arranged in a compact configuration. The short baselines formed by a pair of telescopes in the compact configuration will be sensitive to all structure on the sky that is smaller than ~2' in angular diameter. The remaining telescopes are placed at distances greater than ~100 ft from the close packed array. These telescopes provide long baselines which are no longer sensitive to the extended emission from the galaxy cluster. The long baselines provide information about the point sources that can be used to eliminate the effects of radio sources in the cluster observations. Although the entire SZ analysis is performed directly on the visibilities, it is informative to image the effects of point source contamination and removal. Below is a sequence of steps that demonstrates this process.

The first image is derived from only the short baselines of the array. These baselines are sensitive to both radio point sources and galaxy clusters. Note that the positive(red) emission from the radio source dominates the signal. There is a cluster of galaxies in this image, but it is impossible to discriminate between the cluster and the point source using only short baselines at 28.5 GHz.

The second image is made from only the long baselines in the array. The point source that appears on center is much smaller in angular size than it was in the first image. In fact, the noise features appear much finer in the second image than they did in the first. This change in structure between the first and second images results from the choice of baselines used in imaging the data. The short baselines from the first image are sensitive to large scale structure, the long baselines in the second image provide a much finer resolution.

The point source at the center of the image is modeled from the data in the second frame. The modeled source is subtracted directly from the visibility data. The result is an image made from the long baselines that is just instrumental noise. There is no cluster of galaxies apparent in this image because the extended emission is washed out by the high resolution of the long baselines. The long baselines are simply not sensitive to large scale structure.

The fourth image in the series is derived from the short baselines again, but this time the point source that was contaminating the data has been removed. Compare this image to the first one. The structure in the image has changed. Where there was originally just a bright radio point source, there is now a bright cluster of galaxies which appears as a decrement (deep blue/black) in the image. This data in the first image was sensitive to this cluster, but the resolution was too coarse to differentiate between the cluster and the point source.

Because the u-v coverage of an interferometer has a non-uniform sampling, the fourier transform to the image plane creates an image of the sky convolved with the a 'dirty' beam. This dirty beam creates the ringing that is apparent in the image of the cluster in the 4th panel. The rings that form around the cluster are residual sidelobes from the dirty beam. To remove this effect, the dirty beam is deconvolved from the image resulting in a model for the signal on the sky. The image at the bottom is the final result from the cleaning process.


Over 40 clusters have been observed with the BIMA and OVRO arrays. The figure above shows images of 12 of those clusters. The scale in this figure is inverted so that a decrement corresponding to the SZ signal is represented by the red end of the color scale.

SZ observations have been combined with X-ray observations to derive electron temperatures and electron densities associated with a cluster. With knowledge of these quantities, it is possible to model the absolute size of the cluster. From geometrical arguments, it is then easy to determine the distance to the cluster. These measurements have been used to constrain the Hubble constant and cluster gas-mass fractions. BIMA and OVRO observations yield a value of 60 +/- 3 km s-1 Mpc-1 for the Hubble constant and a baryon density of ΩBh2 = 0.019 +/- 0.0012. For a more detailed account of the analysis, see this paper or this paper. The figure below shows the distance to the cluster derived from the observations vs. redshift observed from emission lines in the X-ray observations or optical observations. The colored lines represent theoretical models assuming a given cosmology while the points and corresponding error bars represent the values derived for each cluster in the sample.

Copyright Holzapfel Group, 2002 Page last modified December 16, 2002