Astron. Astrophys. 336, L37-L40 (1998)

4. Discussion

From three epoch observations we find strong evidence for an increase in separation of the two outer pc-scale components of 0108+388 (C1 and C7) at a velocity of . These outer components have all the properties expected of hotspots in a CSO (in contrast for instance to being knots in a two sided jet): the two components lie at the extremities of the pc-scale structure, have simple SSA spectra and are connected via jets to a core (Taylor et al. 1996). Given their identification as hotspots the measured separation rate of components C1 and C7 implies that the pc-scale source in 0108+388 is very young, i.e. yrs.

Recent results from other CSOs suggest that they are also very young radio sources (e.g. Fanti et al. 1995, Readhead et al. 1996b, Owsianik & Conway 1998). The evidence therefore suggests that CSOs are not in general `frustrated' slowly growing sources within a dense environment but are instead fairly rapidly () growing sources. There are then several possibilities for the subsequent evolution of CSOs. The simplest possibility is that they evolve via a Compact Steep Spectrum phase into classical double radio sources (FRIs or FRIIs). Given their fairly rapid expansion rate the sources would only be expected to spend a short time in their CSO phase. In order to explain the large fraction of CSOs in flux limited samples there must be a strong negative luminosity evolution with increasing source size (e.g. Readhead et al. 1996b). If such luminosity evolution occurs the `population problem' for CSOs is explained because CSOs then evolve into much weaker, more numerous sources. Such negative luminosity evolution is expected in theory (e.g. Begelman 1996), and evolution of just the amount required is found if we compare the efficiency of radio production in CSOs and classical radio double sources (e.g. Readhead et al. 1996a, Owsianik & Conway 1998).

Another way to explain the large population of CSOs is via recurrent activity in these sources (e.g. O'Dea & Baum 1997, Reynolds & Begelman 1997). In this model the sources are quite old but the activity occurs in short bursts. Every time the activity restarts the jet must propagate again from the nucleus to the kpc working surface. Hence the sources appear as Compact Symmetric Objects (size kpc) for a significant fraction of their lifetime. The source 0108+388 appears to be a good candidate for such recurrent activity, since in this source Baum et al. (1990) have detected weak extended emission ( kpc) to the East of the nucleus.

The fact that we see kpc-scale structure only on one side of 0108+388 might be understood as a natural consequence of intermittent activity in the nucleus (although Baum et al. 1990 discuss other possibilities). If the on-off cycle is short then separate regions of activity (light grey areas in Fig. 3) will propagate out towards the kpc-scale hotspots. Due to light travel time effects we are viewing the far side of the source at an earlier time than the near side. It may be that we are viewing the Eastern, near side kpc-scale hotspot at a time when it is being supplied by jet material and the Western, far side kpc-scale hotspot while unsupplied. Such an unsupplied hotspot fades very fast as electrons diffuse to regions of lower density and magnetic field and radiate inefficiently. Given the size of the Eastern hotspot at the lowest 5 GHz contour (Baum et al. 1990), the fact that the Western hotspot is not detected at this contour and assuming an electron backflow velocity of , we calculate that the Eastern hotspot must have been left unsupplied for at least of yrs, providing a lower limit on the on-off cycle time.

Fig. 3. Diagram illustrating a possible explanation of the extended emission in 0108+388. Light grey areas represent material emitted from the central engine in previous periods of activity which propagates through a low density cocoon. Dark grey areas represent working surfaces against high density circumnuclear and IGM gas from which we get significant radio emission on pc- and kpc-scales respectively.

An approximate upper limit to the cycle time can be set by considering the effects of light travel time and jet propagation. We first note that in the proposed model (see Fig. 3) the kpc-scale hotspots advance with low average speed (v) through the relatively dense IGM or intercluster gas; hence the observed projected distances to the two kpc-scale hotspots are expected to be approximately the same. From the measured distance of the Eastern kpc-scale hotspot from the core and the estimated angle to the line of sight (see Sect 3.2), we can calculate that this hotspot is viewed yrs earlier than the Western kpc-scale hotspot. The regions of jet activity (light grey areas in Fig. 3) are assumed to propagate out at high speed through the low density radio cocoon. Relativistic velocities for this material are consistent with the detection of a possible weak kpc-scale jet feature on the Eastern side of the source (Baum et al. 1990). If the on-off cycle time was long compared to the light travel time difference then we would most likely observe either both kpc-scale hotspots supplied or unsupplied. To have a high probability of seeing only one hotspot supplied we require that the lengths of the periods of activity must be comparable or shorter to the light travel time (see Fig. 3); hence we can estimate an upper limit on the cycle time of about yrs.

We conclude that the pc-scale structure in 0108+388 is very young, but that activity in this source may be intermittent. In this case we are probably viewing 0108+388 just as a new phase of high radio-efficiency activity in the central engine has begun. Our results imply that at least some CSOs are recurrent sources. In contrast other CSOs (e.g. 0710+439 and 2352+495) apparently show no extended emission or signs of recurrent activity, so the CSO population may contain a mixture of intermittent and non-intermittent sources.

© European Southern Observatory (ESO) 1998

Online publication: July 20, 1998
helpdesk.link@springer.de