Annu. Rev. Astron. Astrophys. 1980. 18: 489-535
Copyright © 1980 by . All rights reserved

Next Contents Previous


The degree of polarization of the CBR is another attribute that can be directly measured. Polarization anisotropies may be produced by the same sources generating intensity anisotropies and could survive, if the intergalactic medium does not randomize the polarization through a dispersive Faraday rotation (Rees 1968).

Penzias & Wilson (1965) noted that over a substantial portion of the sky, the background was less than 10% linearly polarized. Nanos (1973) carried out the first systematic search for linear polarization. The instrument was a single beam X band (3.2 cm) polarization sensitive radiometer aimed at the zenith. The output signal of the radiometer was the difference in intensity of two orthogonally polarized components. By rotating the entire apparatus about the optic axis, two Stokes parameters (Q or S1 and U or S2) as well as the instrumental polarization asymmetries were measured. As the same beam includes both polarization states, the atmospheric fluctuations cancel in the measurement, providing only that the polarization is switched at a rate fast compared to the time scale of the atmospheric fluctuations. The observation covered a celestial circle of fixed declination, 40°, sampled every beam width, 12°, in right ascension. Each sample consists of two differences in intensity of polarization, one difference being between the North-South plane and the East-West plane and another pair at ± 45°. The rms noise per point is close to receiver noise and corresponds leq 0.03% polarization. No statistically significant average, 24 hour or 12 hour periodic components were observed in the data at the same level of sensitivity.

A similar experiment, carried out at 9 mm (thetaB = 7°) (Lubin & Smoot 1979) with increased sky coverage, delta = 38°, 53°, and 63°, and a factor of about 2 improvement in sensitivity, produced a null result as well.

Caderni et al. (1978) have set upper limits on the polarization in the 3 to 20 cm-1 band with a balloon-borne polarimeter. The instrument consisted of a telescope with a choice of fields of view, 1/2° or 15°, coupled to a bolometer through a rotating linear polarizer. Such systems generally have intrinsic wavelength-dependent polarization anisotropies which must be calibrated by viewing an unpolarized source with a known spectrum. The atmospheric emission, by way of the instrument polarization anisotropy, produced a large systematic polarization offset which, although measured by secant scanning, still limited the observation. The observations, extending over a patch of sky 30° × 60° in the vicinity of the galactic center, set limits of ~ 1% to ~ 0.1% on the degree of polarization on angular scales of 1.5° to 40°. The limits could be unproved if the experiment were done by rotating the entire apparatus rather than the polarizer as is done in the microwave experiments.

No experiments have measured the third Stokes parameter, V or S3, to establish the degree of circular polarization of the CBR. At present all the polarization measurements are limited by instrument-derived, random noise rather than by polarization anisotropies in the atmosphere or from galactic sources. It seems, therefore, that with improved instrumentation a 10 to 100 times better limit could be set before the intervention of local polarization anisotropies. The open question is on what angular scales is the search most interesting?

Next Contents Previous