While the discussion in Sec. 4.1 has concentrated on clusters of galaxies, other objects also contain extended atmospheres of hot gas and may be sources of detectable Sunyaev-Zel'dovich effects. One possibility is in superclusters, large-scale groups of clusters, where small enhancements of the baryon density over the mean cosmological baryon density (which is well constrained by nucleosynthesis arguments; Walker et al. 1991; Smith et al. 1993) are expected, but where the path lengths may be long so that a significant Sunyaev-Zel'dovich effect builds up. Supercluster atmospheres may originate in left-over baryonic matter that did not collapse into clusters of galaxies after a phase of inefficient cluster formation, and could be partially enriched with heavy metals through mass loss associated with early massive star formation or stripping from merging clusters and protoclusters. A measurement of the mass and extent of supercluster gas would be a useful indication of the processes involved in structure formation.
Most work on supercluster gas has been conducted through X-ray searches. Persic et al. (1988, 1990) searched for X-ray emission from superclusters in the HEAO-1 A2 data, finding no evidence for emission from the gas. Day et al. (1991) searched for intra-supercluster gas in the Shapley supercluster using GINGA scans, and were able to set strong limits on the X-ray emission. More recently, Bardelli et al. (1996) have used ROSAT PSPC data to claim that there is some diffuse X-ray emission in the Shapley supercluster between two of its component clusters.
The thermal Sunyaev-Zel'dovich effect provides another potential probe for intrasupercluster gas. Since this effect is proportional to the line of sight integral of ne, it should be a more sensitive probe than the X-ray emission for studying the diffuse gas expected in superclusters. The angular scales of the well-known superclusters are large (degrees), so that the COBE DMR database is the best source of information on their Sunyaev-Zel'dovich effects: ground-based work is always on too small an angular scale, and the balloon searches do not cover such a large fraction of the sky at present.
Indeed, Hogan (1992) suggested that much of the anisotropy in the CMBR detected by the COBE DMR might be produced by local superclusters. This has been tested by Boughn & Jahoda (1993), who found no sign of the anticorrelation of the HEAO-1 A2 and COBE DMR sky maps that would be expected from such a mechanism and concluded that the COBE DMR signal was not produced by supercluster Sunyaev-Zel'dovich effects.
Limits to the average Sunyaev-Zel'dovich effects from clusters of galaxies (and their associated superclusters) were derived by Banday et al. (1996) through a cross-correlation analysis of the COBE DMR 4-year data with catalogues of clusters of galaxies. The result, that the average Sunyaev-Zel'dovich effect is less than 8 µK (95 per cent confidence limit at 7° angular scale), suggests that these Sunyaev-Zel'dovich effects are not strong. However, although population studies of this type show that average superclusters do not contain atmospheres with significant gas pressures, the COBE DMR database can also be searched for indications of a non-cosmological signal towards particular superclusters of galaxies.
Banday et al.
(1996)
were able to set limits
TRJ(7°
50
µK for the Sunyaev-Zel'dovich effects
towards the well-known Virgo,
Coma,
Hercules, and
Hydra clusters.
Since the largest (in angular-size) of these clusters have low
X-ray luminosity, and the highest X-ray luminosity object (Coma) is
strongly beam diluted, it is not surprising that no signals were
found. However, equivalent results for superclusters should set
interesting new limits on their gas contents.
The most prominent supercluster near us (at z < 0.1) is the Shapley supercluster, which consists of many Abell and other clusters centered on Abell 3558, lies at a distance 140 h100-1 Mpc, and has a core radius of about 20 h100-1 Mpc. The estimated overdensity of the Shapley supercluster is the largest known on such a scale, and this supercluster may be the largest gravitationally-bound structure in the observable Universe (Raychaudhury et al. 1991; Fabian 1991). Searches for gas in the supercluster, conducted by Day et al. (1991) and others, have not led to any convincing detection of such gas. A rough scaling argument suggests that the peak Sunyaev-Zel'dovich effect to be expected from a supercluster of scale R is about
For the Shapley supercluster, and a gas temperature of a few
keV, the
Day et al. (1991)
limit on the X-ray surface brightness
corresponds to an Sunyaev-Zel'dovich effect of about -20
µK.
Molnar & Birkinshaw
(1998a)
have used the COBE DMR 4-year
database to set a limit of about -100 µK on the thermal
Sunyaev-Zel'dovich
effect. This result is an improvement on the constraint of Day
et al. only if the atmosphere is hot, with
kBTe
Superclusters are sufficiently massive objects that they also produce
CMBR anisotropies through their distortion of the Hubble flow
(Rees & Sciama 1968;
Dyer 1976;
Nottale 1984)
as well as any Sunyaev-Zel'dovich effects that
they produce. A supercluster of mass M and radius R will cause a
Rees-Sciama effect of order
which is of the same order as the Sunyaev-Zel'dovich effect (75),
but with a different spectrum (that of primordial anisotropies) and
angular structure. Since the intrinsic anisotropies in the CMBR are
expected to be larger than these supercluster-generated effects, it is
unlikely that even statistical information about
15 keV. However,
improvements in the microwave background
data from the next generation of satellites will achieve a factor 10
or more improvement in sensitivity to the Sunyaev-Zel'dovich effect, and will
strengthen the limits on the mass of gas in the supercluster at all
likely gas temperatures.
TRS
can be obtained, but the next
generation of microwave background satellites should be able to use
Sunyaev-Zel'dovich effect data to constrain supercluster
properties. No useful limits on the mass of superclusters (or the
Shapley supercluster in particular) are obtained
using the COBE DMR 4-year data to search for Rees-Sciama effects
(Molnar & Birkinshaw
1998a).