Extragalactic radio sources: What are we trying to unify ?

Powerful extragalactic radio sources harbour an active galactic nucleus (AGN) which, due to some process associated with the central massive black hole, produces powerful jets of highly-collimated relativistic plasma (Scheuer 1974). The attendant radio emission is non-thermal, having an extent independent of the host galaxy. In these powerful sources the radio-emitting regions can extend out to hundreds of kpc or even Mpc beyond the host galaxy, although a notable exception is shown by the peaked spectrum sources which have galactic-size radio structures (tens of kpcs). In contrast, very low-radio power sources (normal and `starburst' galaxies) have diffuse, thermal radio emission associated with stellar activity which is limited in extent by the optical host galaxy.

Extragalactic radio source morphologies vary enormously. Historically, a first-order classification describes them as extended or compact sources: extended sources are those with large-scale emission structures whilst compact sources lack such features, usually being characterised by a dominant unresolved radio core. This division between extended and compact sources nearly correlates with that between steep and flat-spectrum respectively, describing the observed shape of the radio spectrum. However, it is now realised that both classifications (compact/extended and steep/flat) have major limitations in describing the known radio source types, namely that (i) hybrid sources exist, (ii) these characteristics are a function of observing frequency, (iii) peaked spectrum sources (CSS and GPS) directly refute compact=flat, extended=steep dogma and (iv) that these classifications are affected by projection effects.

At low radio frequencies (i.e. 400 MHz) the radio morphologies of the double-lobed extended sources can be described by the scheme of Fanaroff & Riley (1974). This scheme measures the distance between the central maxima of radio source compared to its overall size. When the regions of maximum brightness are separated by more than 0.5 times the overall source size the source is classified as an FRII (as the source is edge-brightened). Where the regions of maximum brightness lie within this limit the source is an FRI (centrally-concentrated). The FR-class division very nearly correlates with radio power such that the highest-radio-power sources are predominantly FRII-type whilst those of lower-radio-power are usually FRI-type. According to the FR classification all quasars² are FRIIs, whereas BL Lac sources have been observed with both FRI and FRII morphologies (Dallacasa et al. 1997).

Since the adoption of Fanaroff & Riley's morphological scheme, much effort has been expended on understanding the physical origin of the two FR classes. Fundamentally this origin is attributed to the following two factors:

1.

`Nature': Differing Central Engines.

This hypothesis suggests that the two FR classes have different central accretion processes and in turn these account for the differences in large-scale radio structure associated with each FR type (Baum, Zirbel & O'Dea 1995). Whilst FRIIs have near-Eddington-luminosity accretion rates, it has been hypothesized that FRIs undergo sub-Eddington accretion due to an advection-dominated accretion flow (ADAF). A simple ADAF model is strongly-supported by observations of the hard X-ray spectra from FRIs (Reynolds et al. 1996; Di Matteo & Fabian 1997), although not by their radio/sub-mm spectral characteristics (Di Matteo et al. 1998).

2.

`Nurture': Environmental Differences.

This hypothesis suggests that the FRI/II division is a product of the effects of the environment in which the host galaxy exists. In this case both FRIs and FRIIs have similar accretion processes producing relativistic radio jets, with FRI radio jets decelerating over shorter distances than FRII radio jets due to entrainment of a dense surrounding medium (Bicknell 1996). This hypothesis requires that FRIs exist in regions of high IGM density and indeed there is accumulated evidence that they do (Longair & Seldner 1979; Lilly & Prestage 1987; Prestage & Peacock 1988; Yates, Miller & Peacock 1989; Hill & Lilly 1991; Zirbel 1997). However, as `classical' FRIIs are also found in clusters (e.g. 3C 34 Best, Longair & Rottgering (1996)) this cannot be the entire story.

It is almost certain that both factors contribute to the FRI/II division, although which is the dominant factor is yet to be determined. Further clues come from the observation that FRIs have optically-brighter host galaxies (for the same intrinsic radio power) compared to FRIIs (Ledlow & Owen 1996).

Further complicating any simple classification of powerful radio sources are the peaked-spectrum radio sources - compact steep-spectrum (CSS) and giga-hertz peaked spectrum (GPS) sources. These are intriniscally powerful, yet galactic-sized radio sources which exhibit a turnover in their radio spectrum around 1 GHz. Both CSS and GPS sources come in radio galaxy and quasar flavours. Their radio morphologies suggest that they are compact versions of the `classical' FRIIs, although why they are so small is not yet established: It is hypothesized that these are either young FRIIs or FRIIs trapped in a dense environment (Fanti & Fanti 1990; O'Dea, Baum & Stanghellini 1991; Fanti & Fanti 1994). Given their FRII morphologies these peaked-spectrum sources are treated as part of the FRII population in the discussion which follows.

© Copyright Astronomical Society of Australia 1997