2.5. Angular fluctuations in the cosmic microwave background

"If no fluctuations can be found, we have no direct evidence at all that galaxies are formed at early epochs through gravitational instability."

G.R. Burbidge, 1980

The search for anisotropy in the cosmic blackbody radiation has frustrated many astronomers since Penzias and Wilson (1965), in their discovery paper, established an isotropy limit of 10%. The motivation for this search has been the notion that the early universe was highly, but not completely, homogeneous. Matter density fluctuations were present that eventually would grow by gravitational instability into galaxy-sized condensations. Searching for the corresponding fluctuations in the background radiation provides a unique probe of the structure of the very early universe.

2.5.1. Dipole anisotropy

The first detection of anisotropy was that of the dipole component, expected on the sole assumption that the blackbody radiation was of extragalactic origin. Independent experiments have confirmed the effect, both in amplitude and direction of the intensity maximum, despite limited sky coverage. An unavoidable problem which arises because of the partial sky coverage is the presence of higher order moments which will bias the various experiments somewhat differently.

Taking a simple average of the Berkeley (Silk and Lubin, 1979), Princeton (Cheng et al. 1979), and Florence (Fabbri et al. 1980) experiments leads to the following parameters for the dipole anisotropy:

and direction of peak intensity RA 1.5^h ± 0.4^m, declination +0.2° ± 7°. If this were entirely due to the motion of the earth with velocity v, relative to the blackbody radiation, one predicts a dipole anisotropy for the temperature angular distribution

(2.35)

Correction for Galactic rotation (v = 225 km s^-1 in the direction = 21.2^h, = + 48°) yields the motion of the Local Group relative to the cosmic background radiation:

(2.36)

On the other hand, measurements of galaxy redshifts in the Local Supercluster have yielded information on the Virgocentric flow. The Local Group is falling towards the Virgo cluster ( = 12.5^h, = + 12.4°) with a velocity estimated at between 180 ± 30 km s^-1 (Yahil et al., 1980) and 470 ( ± 75) km s^-1 (Tonry and Davis, 1981b): the difference is in part due to selection of the galaxy sample and to how the centre of mass of the Virgo cluster is defined. According to the former analysis, the motion of the Local Group relative to the background radiation must be largely determined by the matter distribution outside the Local Supercluster.

However, on the basis of the interpretations of the available redshift data, we cannot exclude the entire effect as being due to the Virgocentric flow. While the apex of the dipole motion is apparently some 45° away from Virgo, the effects of the partial sky coverage indicate that the errors in direction have most likely been underestimated, as is also implied by comparing the different results on the dipole anisotropy. Moreover, non-radial motions are very likely to be produced during the collapse of a complex region like the Virgo supercluster. The simple model of White and Silk (1979) indicates that a modest degree of shear could readily reconcile the radiation anisotropy with the dynamics of the Local Supercluster.

The motivation for believing that the dipole anisotropy should be explained by a simple Virgocentric flow model becomes questionable if we accept the existence of the Rubin-Ford effect. Rubin et al. (1976) found a motion for our galaxy of 600 ± 125 km s^-1 relative to a sample of Sc galaxies at a typical redshift cz = 5000 km s^-1 and in a direction = 2^h, = 53°. Large-scale inhomogeneity, beyond the Virgo supercluster, represents a possible explanation of this effect; such structure would inevitably contribute to the gravitational acceleration of the local rest frame and therefore also affect the dipole anisotropy. More recently, Hart and Davies (1982) found for a sample of Sbc galaxies at a median redshift cz = 3000 km s^-1, intermediate between that of Virgo and the original Rubin-Ford sample, that our galaxy possesses a relative velocity of 440 ( ± 60) km s^-1 in the same direction as the microwave dipole anisotropy. This result sheds some doubt on the reality of the Rubin-Ford effect.