Annu. Rev. Astron. Astrophys. 1992. 30: 653-703
Copyright © 1993 by . All rights reserved

Next Contents Previous

3.2.3 ANISOTROPY OBSERVATIONS AT THE SOUTH POLE Lubin and collaborators have carried out a number of very sensitive measurements of the isotropy of the microwave background radiation (Lubin et al 1990, Meinhold & Lubin 1991, Gaier et al 1992) using a 1 m diameter off-axis parabolic mirror with a sinusoidally nutating subreflector. At 91 GHz the resulting switching pattern approximates a square wave chop of 1° at 8 Hz and their beam is well approximated by a Gaussian of dispersion sigma = 13', and beamwidth thetaFWHM approx 30'. They used this system first on a balloon to obtain useful estimates of the Galactic dust emission (see later), and then at the South Pole. They have had two successful observing runs at the South Pole - from November 1988 to January 1989, and from December 1990 to January 1991. We shall refer to these as the first and second observing runs, respectively.

In the first observing run (Meinhold & Lubin 1991) they used an SIS receiver, operating at 91 GHz with a bandwidth of 550 MHz, and in the second observing run (Gaier et al 1992) they used a transistor amplifier operating at 25-35 GHz. This takes advantage of an important new development of the last five years: the advent of very low noise ``HEMT'' transistor amplifiers [see e.g. Das (1987), Mishra et al (1988), Chao et al (1989, 1990) and Tan et al (1991)] which has revolutionized receivers at frequencies between 10 GHz and 50 GHz by providing receiver noise temperatures at most 50% higher than maser amplifiers over bandwidths one to two orders of magnitude larger, thus yielding sensitivities 2-7 times better than the best high frequency maser amplifiers available in the late 1980's.

In the first run ten fields separated by an angle on the sky equal to the effective beamthrow of 1° were observed at constant declination in the single switching mode. Thus the telescope was pointed at nine positions. It took 15 minutes for the consecutive observations of the ten fields. The mean levels for the nine DeltaT's were around a few mK, and there were slow drifts of less than 1 mK per hour. For each 15 minute scan the mean level and the drift were subtracted from the data and consecutive scans were then averaged. With the scanning strategy adopted it would be possible, in principle, to probe angular scales from 13' to ~ 5°, however the subtraction of the means and drifts eliminates the information on angular scales larger than approximately 1°. The results for the first observing run are shown in Figure 1d. This data set has a chi2 of 6.9 for 7 degrees of freedom. Meinhold and Lubin carried out Monte Carlo tests simulating the drifts in the mean level, changing linear drifts with angle, and Gaussian instrument noise and in each case they recover the shape of the data set shown in Figure 1d. Thus the data correction procedure i was not responsible for the characteristics of the data set. They then assumed a Gaussian spectrum for the distribution of sky fluctuations and derived the 95% confidence upper limit given in Table 3.

These results are particularly important in placing limits on many models of galaxy formation, for which the predicted values of DeltaT/T peak in the range 10' to ~ 1°. Comparing these results with those of Readhead et al (1989) at the same power in the likelihood ratio test, Vittorio & Muciaccia (1991) have calculated that the latter results place slightly more stringent restrictions on cold dark matter cosmologies. On the other hand Bond and Myers (1991a, b) have used a Bayesian analysis which indicates that the former results place slightly more stringent constraints on cold dark matter cosmologies for the case of standard recombination, and considerably more stringent constraints for an Omegab = 0.1 universe with no recombination. These authors and Bond et al (1991) have also used the combined South Pole and NCP data sets to show that hot dark matter models are convincingly ruled out, and to place interesting constraints on cold dark matter and isocurvature baryon models with power law density perturbation spectra having indices of n = -1, n = -0.5, and n = 0.

The most important point to emphasize here is that two completely independent programs using different instruments at different frequencies place very similar constraints on the cold dark matter model, and that modest improvements in the sensitivity of the observations would either detect intrinsic anisotropy or conflict with cold dark matter models.

In the second observing run a similar observing strategy was adopted, and the observations were pushed to considerably higher sensitivities. The observations were made in four independent frequency channels each 2.5 GHz wide, covering the range 25-35 GHz. The beamwidth was 1°.5 and the chop was 3° on the sky. The data have not yet been fully analyzed, but the preliminary results are encouraging. Figure 1e shows the results for the 32.5-35 GHz channel. If, as seems likely the clear evidence of ``signals'' in the lower frequency channels (not shown) is ascribable to Galactic synchrotron emission which is not evident in the highest frequency channel shown here, then the data from the highest frequency channel will provide considerably more stringent limits on intrinsic anisotropy than the observations discussed above. These multi-channel data also illustrate the problem which now confronts observers in this field, and which we have already seen in the cases of the VLA and OVRO observations, namely that the sensitivities have reached levels at which other sources of cosmic radiation are a significant source of confusing signals which have to be subtracted from the data sets in order to study the intrinsic anisotropy of the microwave background radiation per se.

Next Contents Previous