Observations of 21-cm line radiation allow the exploration of cosmological questions because the processes leading to either emission or absorption are generally well understood. The redshifts measured with commonly available HI line spectrometers are among the most accurate. While galaxies of high optical surface brightness dominate volumes of high galaxy density, the low-surface-brightness objects found in the regimes of groups and supercluster peripheries are usually HI rich and thus are easily studied by 21-cm techniques. Furthermore, since the HI signature of an HI disk is distinctive, the observation of the characteristic two-horned 21-cm profile can testify to the presence of a gas-rich disk structure even in objects not visible optically. The 21-cm line thus serves not only as the complement of observations made at other wavelengths but can, in some cases, provide information not obtainable any other way.
In this section, we discuss further the use of 21-cm line studies to address questions of cosmological relevance. The use of the velocity width-magnitude relation to measure deviations from the Hubble flow has already been discussed in Section 12.4.2. In conjunction with measures made at other wavelengths, 21-cm line observations are critical to our understanding of the structure and evolution of the universe.
12.7.1. 21-cm Redshift Surveys
In contrast with studies at optical wavelengths, HI redshift surveys are especially applicable to objects that cluster least and are therefore useful for studying the large-scale structure of the galaxy distribution. The spectacular surge of interest in the study of the large-scale structure of the universe that has occurred in the last decade has stimulated the undertaking of massive surveys to determine the radial velocities of galaxies at progressively fainter magnitude levels. While the emphasis at the beginning of the 1980s was mainly on the rough mapping and the determination of the scale of the inhomogeneities in the distribution of luminous matter, attention has more recently shifted to the topology of the galaxian distribution and its segregation properties. Entering an area traditionally the monopoly of optical instruments, large-aperture radio telescopes have become important as redshift machines. In this section, we will review the technical aspects of redshift surveys, and in the next one, some of their results will be presented.
Following Table 12.1, we can assume that the mean expectation value for the HI mass of a spiral galaxy is on the order of 3 × 109 M (for h = l, as will be assumed throughout this section). Via Equation (12.3), that mass translates into a flux integral of 1.3 × 104d-2 Jy km s-1, where d is the distance to the galaxy in Mpc. For a most probable inclination of the disk, an observed velocity width of about 400 km s-1 should be expected; if we express the distance in terms of a radial velocity v3 (in 1000 km s-1 units), the average flux density of the galaxy over the line profile will be 320v3-2 mJy. For comparison, a typical HI observation of 5 minutes on source plus 5 minutes off source (for subtraction of sky and instrumental effects) yields an rms noise of about 1 mJy, at 20-km s-1 resolution, with the Arecibo 305-m telescope (using a GaAs FET receiver and the current line feed system, which give a system temperature of 35 to 65 K, depending on the zenith distance of the source). Because the signal will be spread over many channels, an average signal-to-noise ratio of 2 will be more than sufficient to ensure detection. In other words, a "typical" spiral can be detected in one "on-off" pair, at Arecibo, out to a distance corresponding to + 15,000 km s-1. Such a radial velocity is larger than the characteristic redshift "depth" of most galaxy catalogues, such as the Uppsala General Catalogue (Nilson 1973) or the Catalogue of Galaxies and Clusters of Galaxies (Zwicky et al. 1961-68), which, combined, provide listings for about 30,000 galaxies in the northern hemisphere. Improvements in receiver technology and other instrumental upgrades promise to expand the number of accessible sources by one order of magnitude in the next few years.
Unlike continuum surveys with single-dish radio telescopes, 21-cm redshift surveys are not confusion limited. Adding the third dimension provided by the radial velocity guarantees that the vast majority of sources will be separable, even in high-density environments such as clusters. Confusion - or the inability to separate line blends - only occurs in very tight pairs or groups with small velocity dispersions and affects less than 1% of all surveyed sources.
Currently, other instrumental limitations, besides sensitivity, limit the scope of 21-cm redshift surveys. Standard autocorrelation spectrometers allow instantaneous coverage of velocity windows on the order of +8000 km s-1. Thus, in spite of the fact that angular size is, to the first order, a good distance indicator, the search for the 21-cm line emission of a given galaxy may take several observations in contiguous velocity ranges. Currently, effective redshift searches of 21-cm emission are limited to velocities below + 25,000 km s-1. Man-made interference constitutes a severe handicap at some frequencies of interest for the redshifted 21-cm line, and special precautions, such as lateral shielding of the receiving feed sytem or software excision of interfering signals, are becoming necessary.
The first extensive 21-cm survey of cosmological significance was conducted by Fisher and Tully (1981), on a sample of approximately 2000 galaxies in the Local Supercluster. Currently, a sample about twice as large is nearing completion in the Pisces-Perseus supercluster region (Giovanelli and Haynes 1985b), and somewhat smaller efforts are under way in the Cancer and Hercules regions (Bicay and Giovanelli 1986). Figure 12.9 displays partial results of the Pisces-Perseus survey, including the mapping of a long filamentary structure which probably extends well beyond the boundaries of the surveyed region. The Local Supercluster is also being studied to fainter levels, with current attention concentrating on the characteristics of the dwarf galaxy population (Hoffman et al. 1987).
Figure 12.9. The upper panel displays the density of galaxies brighter than m = 15.7 per unit solid angle, as contours of different shade intensity. The lower panel illustrates the radial velocity distribution, as a function of Right Ascension, of the galaxies which are contained within the jagged contour outlined in the upper panel, emphasizing the existence of a three-dimensional structure which coincides with the enhancement in the surface density visible in the upper panel (From Giovanelli et al. 1986).