3.3. Head-tail and other distorted radio structures
There are two general types of radio source structures that occur predominantly in clusters of galaxies. In this section, distortions in the structures of single radio galaxies are discussed. In Section 3.4 cluster sized radio halo sources are reviewed.
A large proportion of relatively isolated radio galaxies have a fairly simple and symmetrical radio structure; for a review of the properties of radio galaxies see Miley (1980). Many of these galaxies have a compact radio source associated with the nucleus of the galaxy. There is also extended radio emission, generally in the form of double radio 'lobes', which lie on either side of the galaxy. These lobes are often of comparable brightness and projected distance from the nucleus, and most importantly, lie on a line through the nucleus of the galaxy. Recently observations have detected 'jets' of radio emission originating at the nucleus and extending out to the radio lobes; in some cases only one jet on one side of the galaxy has been found. The conventional theoretical scenario for the origin and energetics of these radio galaxies contains the following elements (Miley, 1980). First, the ultimate energy source for the radio emission (and all other nonthermal galaxy emission) is thought to be a very compact object in the nucleus of the galaxy. Energy is carried from this nonthermal engine out to the radio lobes by twin 'beams' of plasma; this plasma probably contains a mixture of thermal gas and relativistic nonthermal particles, which cause the beams to be observable in some cases as radio jets. The beams may be more or less continuous, or may consist of blobs of plasma ('plasmoids'). The beams probably move outwards until they encounter a sufficient quantity of intergalactic (or intragalactic) gas, at which point their bulk kinetic energy of motion is converted into thermal energy and into the disordered relativistic motion of particles of nonthermal plasma. This nonthermal plasma produces the emission from the radio lobes.
Radio galaxies in clusters show more complex radio structures, which generally tend to lack the symmetrical, aligned double structure of standard radio galaxies. These range from double lobed radio sources in which the lobes are not aligned with the galaxy nucleus ('bent-doubles' or 'wide-angle-tails') to sources in which all the radio emission lies in a tail on one side of the galaxy, and the galaxy itself forms the head of the tail ('head-tail' or 'narrow-angle-tail' radio galaxies).
The first head-tail (HT) radio galaxies discovered were NGC1265 and IC310 in the Perseus cluster (Ryle and Windram, 1968; see Figure 7), followed by the discovery of head-tail radio galaxies in Coma (Willson, 1970) and the 3C129 cluster (MacDonald et al., 1968). Figure 8 shows a radio map of NGC1265, which is the archetypical head-tail radio galaxy. Some lists of head-tail radio galaxies and other distorted radio sources are those of Rudnick and Owen (1976a, b), Simon (1978, 1979), and Valentijn (1979a), and other observations of head-tail radio galaxies in rich clusters are given in Hill and Longair (1971), Vallee and Wilson (1976), Miley and Harris (1977), Gisler and Miley (1979), Burns and Ulmer (1980), Hintzen and Scott (1980), Bridle and Vallee (1981), Gavazzi et al (1981), Vallee et al. (1981), and Dickey and Salpeter (1984).
Figure 8. A low resolution radio map at a frequency of 5 GHz of the head-tail radio source associated with the galaxy NGC1265 in the Perseus cluster, from Wellington et al. (1973). Contours of constant radio surface brightness are shown superimposed on the optical image of the galaxy.
NGC1265 and IC310, the first two head-tail radio galaxies discovered, are both in the Perseus cluster and have tails that lie on the line from the galaxy to the powerful radio galaxy NGC1275 at the cluster center (Figure 7). Ryle and Windram (1968) suggested that the radio emission from these galaxies was activated by a wind of relativistic particles from NGC1275, which also determined the directions of the tails. However, subsequent head-tail radio galaxies have not been found to show any alignment with the direction to powerful radio galaxies or the cluster center. The accepted explanation of head-tail radio galaxies, originally due to Miley et al. (1972), is that they are conventional radio galaxies moving at a high velocity through a static intracluster gas. The radio emitting beams or plasmoids are decelerated by the ram pressure of the intracluster gas and form a wake behind the galaxy. The high velocity of the galaxy is a result of the gravitational potential of the cluster (that is, the velocity v is comparable to the cluster velocity dispersion). The ram pressure acting on the radio blobs or beams is then
where g is the intracluster gas density (Section 5.9).
Miley (1973) found that the spectral index r and the fractional polarization of the radio emission increased with distance away from the galaxy along the tail. Synchrotron emission energy losses will steepen a radio spectrum, so the spectral variations are consistent with injection of particles at the galaxy. The polarization indicates that the magnetic field is highly ordered and directed along the direction of the tail, probably by the sweeping of the radio emitting plasma behind the galaxy. There are some indications that head-tail radio galaxies are more rapidly moving than typical cluster galaxies (Guindon and Bridle, 1978), but the effect is not very large (Ulrich, 1978), and in any case, only one component of the velocity can be measured. Head-tail radio galaxies are never cD galaxies and are seldom among the most luminous galaxies in a cluster, either at radio or at optical wavelengths (Rudnick and Owen, 1976a, b; Simon, 1978; McHardy, 1979; Valentijn, 1979c). Since cD galaxies are nearly at rest in the cluster potential (Section 2.10.1), they would not be expected to form head-tail radio galaxies.
Of course, the intracluster gas that produces the head-tail radio galaxies also produces X-ray emission; in general, the densities of intracluster gas g derived from X-ray observations are consistent with those needed to give a ram pressure sufficient to produce the observed radio structures (see, for example, Simon, 1979).
The first detailed theoretical work on head-tail radio galaxies was done by Jaffe and Perola (1973). They suggested two models; in the first, blobs of plasma were ejected in opposite directions (taken to be perpendicular to the direction of motion of the galaxy). They gave several arguments against this model. First, the adiabatic expansion of the blobs would produce large losses in their energy; if these losses exceeded losses due to synchrotron emission, the spectrum would not vary as observed. Second, they argued that the magnetic field was too well ordered to be an initially disordered field. Thus they proposed a second model, in which the radio galaxy possessed an extensive magnetosphere, which was swept behind the galaxy. The magnetosphere provided an ordered magnetic field, and they argued that it could confine the adiabatic expansion of the radio emitting blobs. This second argument was shown to be incorrect by Cowie and McKee (1975) and Pacholczyk and Scott (1976). Cowie and McKee showed that the large adiabatic losses found by Jaffe and Perola were due in part to the assumption of high Mach number flow, which is not correct for the temperatures of the intracluster gas derived from X-ray observations (Section 4.3). As a result, subsequent theoretical work has largely been devoted to models for the interaction of free plasmoids (the first Jaffe-Perola model) or twin beams with intracluster gas (Cowie and McKee, 1975; Pacholczyk and Scott, 1976; Begelman et al., 1979; Christiansen et al., 1981).
One important problem with this model is that in many sources the age of the tail (projected length divided by the estimated galaxy velocity) is much longer than the lifetimes of the emitting electrons against synchrotron losses (Wilson and Vallee, 1977). Pacholczyk and Scott (1976) and Christiansen et al. (1981) have argued that these particles are reaccelerated by turbulence in the tails.
Recent high resolution radio observations have detected well defined radio jets leading out from the nucleus of the head-tail radio galaxy to the start of the tails (Owen et al., 1978, 1979; Burns and Owen, 1980). Figure 9 shows these jets in NGC1265. Begelman et al. (1979) presented a twin-jet theory for the formation of HTs. Because the observed jets are narrow and follow well defined curved paths, it is unlikely that they consist of independent plasmoids, because these plasmoids would most likely have different masses and surface areas and would be bent by different amounts by the ram pressure force. The transition from a collimated jet into the swept back tail is fairly abrupt. Jones and Owen (1979) argued that the jets in the inner parts of the galaxies are protected from the ram pressure of the intracluster gas. They suggested that the head-tail radio galaxies contain regions of gas that are bound to the galaxy and too dense to be stripped by ram pressure (Section 5.9). The jets propagate through and are confined by this gas until they reach the intracluster gas and are swept back. However, the jets are curved in this inner region (Figure 9), which is difficult to understand if they propagate through static intragalactic gas (although the curvature could be due to buoyancy force if the intragalactic gas is compressed by ram pressure from the intracluster gas). In any case, Begelman et al. (1979) are able to fit the observed shape of the jets assuming ram pressure from intracluster gas and no intragalactic gas.
Figure 9. A high resolution Very Large Array radio map at a frequency of 4.9 GHz of the head-tail radio galaxy NGC1265, from O'Dea and Owen, 1986. Note the twin radio jets leading from the nucleus of the galaxy out to the radio tails.
In some cases the tails in head-tail radio galaxies are observed to be curved or bent. Jaffe and Perola (1973), Miley and Harris (1977), and Vallee et al. (1979, 1981) suggested that the curvature reflects nonlinear galaxy motion, due to the orbit of the head-tail radio galaxies in the cluster potential, or due to a binary orbit. Jaffe and Perola (1973) also suggested that there might be an intracluster gas wind with shear. Cowie and McKee (1975) suggested that the bending might be due to buoyancy forces; the bendings in a number of head-tail radio galaxies are consistent with this (Gisler and Miley, 1979).
Another common morphological type of cluster radio source is wide-angle-tails' or WATs. These resemble classical double radio sources in which the two radio lobes and/or jets are not oppositely aligned, but lie at an angle. Figure 10 shows the radio image of the WAT 1919+479, which is in a Zwicky cluster. Some lists of these sources include those of Owen and Rudnick (1976b), Simon (1978), Guindon and Bridle (1978), Valentijn and Perola (1978), Valentijn (1979b, c), and van Breugel (1980). WATs are generally associated with optically dominant cluster galaxies, often cD galaxies (Owen and Rudnick, 1976b; Simon, 1978; Valentijn, 1979c), and tend to be more luminous radio sources than head-tail radio galaxies. They tend to be slower moving galaxies than those containing head-tail radio galaxies (Guindon and Bridle, 1978). A particularly fascinating case is the radio source 3C75 in the cluster A400, which has two intertwined pairs of jets which merge into a WAT (Owen et al., 1985).
A major theoretical question about WATs is whether they are the result of the same physical process (ram pressure by intracluster gas) as head-tail radio galaxies. Owen and Rudnick (1976b) suggested that ram pressure was the mechanism behind WATs; since the WATs occur in slow moving (cD) galaxies, the ram pressure is less, and the radio structure is bent at a smaller angle. Another possible reason that WATs might be less bent is that they are associated with stronger radio sources. The plasmoids or beams associated with these sources may have more momentum than those in weaker radio sources, and may be harder to deflect. Valentijn and Perola (1978) and Valentijn (1979b) suggested that WATs were similar to HTs, except that the ejection angle of the beams or plasmoids was not nearly perpendicular to the velocity of the galaxy. This would not, by itself, explain why WATs occur in galaxies that are more prominent optically and in radio emission. Valentijn (1979b) also invokes the smaller galaxy velocity and more rigid beams discussed above, and suggests that these brighter galaxies may have more intragalactic gas, which shields the beams from the intracluster gas.
Burns (1981) and Burns et al. (1982) suggest that the WATs are bent mainly by buoyancy forces and not ram pressure forces. Of course, if the galaxy associated with the WAT were a cD galaxy located at the center of the cluster potential and if the intracluster gas density were spherically symmetric about this center, no bending due to buoyancy could occur. Burns et al. (1982) argue that the cD associated with a WAT may not be exactly at the cluster center, and that accretion by the slowly moving cD may produce an aspherical density distribution in the intracluster gas.
Sparke (1983) argues that WATs and other distorted radio sources associated with cDs are indications that the clusters are in the process of collapsing. She notes that many of the WATs are associated with irregular clusters with clumpy X-ray emission and low X-ray temperatures (the irregular X-ray clusters of Table 3 below).
|No X-ray Dominant Galaxy (nXD)||X-ray Dominant Galaxy (XD)|
|Low Lx 1044 erg/s||Low Lx 1044 erg/s|
|Cool gas Tg = 1-4 keV||Cool gas Tg = 1-4 keV|
|X-ray emission around many galaxies||Central galaxy X-ray halo|
|Irregular||Irregular X-ray distribution||Irregular X-ray distribution|
|High spiral fraction > 40%||High spiral fraction > 40%|
|Low central galaxy density||Low central galaxy density|
|Prototype: A1367||Prototype: Virgo/M87|
|High Lx 1044 erg/s||High Lx 3 × 1044 erg/s|
|Hot gas Tg 6 keV||Hot gas Tg 6 keV|
|No cooling flow||Cooling flow onto central galaxy|
|Regular||Smooth X-ray distribution||Compact, smooth X-ray distribution|
|Low spiral fraction 20%||Low spiral fraction 20%|
|High central galaxy density||High central galaxy density|
|Prototype: Coma (A1656)||Prototype: Perseus (A426)|
The association between these distorted radio morphologies and clusters of galaxies containing intracluster gas has been used to detect previously unobserved clusters and cluster X-ray sources. Burns and Owen (1979) found a number of distorted radio sources associated with poor Zwicky clusters, and suggested they might be X-ray sources. This conjecture was confirmed by Holman and McKee (1981). Fomalont and Bridle (1978) discovered a number of WATs in groups of galaxies.
Hintzen and Scott (1978) suggested that quasars with distorted radio structure were likely to be members of clusters of galaxies, and that radio observations of quasars could be used to detect high redshift clusters of galaxies and cluster X-ray sources. In general, optical studies have not found that quasars are associated with rich clusters, and any method that selected quasars in clusters or any type of high redshift clusters would be very useful. This method has in fact been used to detect clusters around several quasars (Hintzen et al., 1981; Harris et al., 1983a) and to provide a list of other candidates (Hintzen et al., 1983).
Additional evidence for the dynamical effect of the hot intracluster gas on radio galaxies comes from distortions in the structure of the radio source produced if the radio galaxy is moving relative to the intracluster medium, as discussed below.