Please Note: the e-mail address(es) and any external links in this paper were correct when it was written in 1995, but may no longer be valid.
Astronomical Observatory, Jagiellonian University, ul. Orla 171, 30-244 Krakow, POLAND
The large scale structures of extended quasars and radio galaxies can be used as a test for radio galaxy-quasar unification schemes. If both these categories of radio sources are intrinsically the same type of objects, but only appear different to an observer due to the various viewing directions, according the hypothesis of Barthel (1989), then their radio structures are expected to evolve with redshift in the same way. This suggestion was put forward by Gopal-Krishna & Kulkarni (1992) on the grounds of radio linear sizes attained by extended quasars and powerful radio galaxies. However, Chyzy & Zieba (1993) recently came, on the basis of 152 radio galaxies and 173 quasars, to a quite contrary view indicating differences not only in the cosmological evolution of radio galaxy and quasar linear sizes, but also in their size dependence on radio luminosity. A conclusion supporting this point of view was also reported by Singal (1993b) who examined a large sample of 789 sources. Singal (1993a) also showed that observed relative numbers and linear sizes of radio galaxies and quasars are inconsistent with unification even invoking a cosmic evolution in the opening angle of obscuring torus.
In this contribution we present further investigation of cosmic evolution of radio structures based upon the other geometrical parameters which describe the observed structures. Apart from the simplest linear size parameter it is possible to determine two independent parameters assessing the asymmetry of the structure: the arm length ratio , defined as the ratio of the distances of hot spots from the core; and the misalignment , which measures the apparent bending, and is defined as the ratio of the displacement of the core from the source axis to the linear size (see also Figure 1 in Zieba & Chyzy (1991)).
The asymmetry parameters and can potentially be a powerful tool in the consistency test for the orientation based unification scheme as, according to it, their evolutionary patterns should be the same for radio galaxies and quasars. Contrary to the linear size, they are not sensitive to the simple homological rescaling of the whole structure and hence to the age or expansion velocity of the structure. In that case, possibly revealed differences in asymmetry evolution of radio galaxies and quasars might give evidence in favor of even deeper physical differences between these two AGN types of sources. We performed a quantitative comparison of radio galaxy and quasar apparent asymmetry, evaluating the dependence of the , and parameters on redshift and spectral radio luminosity at 1.4 GHz.
The observational base for our discussion comprises two samples which are described in detail in our earlier papers (Zieba & Chyzy (1991),Chyzy & Zieba (1993)). The radio galaxy sample contains 152 triple, edge-brightened FRII powerful objects, carefully selected from the GB/GB2 complete sample Machalski & Maslowski (1982) and the 3CR sample Laing et al. (1983) with all possible high-redshift galaxies included. The main contributions to the quasar sample, which containing 173 objects, come from the lists of Barthel et al. (1988), Hintzen et al. (1983), Miley & Hartsuijker (1978), the 3CR sample Laing et al. (1983) and GB/GB2 sources (Machalski & Maslowski (1982)). To avoid undesirable bias all possible subgalactic compact steep spectrum sources (linear size less than 10 kpc) had been extracted from the final sample.
Arm lengths ratio, misalignment and linear size were calculated from the positions of the hot spots and the central component, which are usually found in publications, or were estimated directly from maps. The radio galaxy sample spans the redshift range and the luminosity range and the quasar sample spans, respectively, and .
In order to derive the evolutionary behavior of the asymmetry of radio structures we estimated the dependence of the median values of and parameter on redshift and radio luminosity in the form: and respectively. The special method was applied to eliminate the influence of the redshift-luminosity correlation (Zieba & Chyzy (1991)). In this method, the dependence of a median value of an asymmetry parameter on redshift was fitted in subsequent steps, simultaneously with rescaling the parameter to the chosen luminosity using the updated values of the relation found in the previous step. The best solutions we have found for the fitted and when this approach was used on our quasar and radio galaxy samples are listed in Table 1 together with 95% confidence intervals and analogous values estimated for linear sizes. All numbers presented in Table 1 were calculated in a uniform manner using and planes for rescaling of the source parameters to a fixed radio luminosity and redshift .
Table 1: The best-fit model parameters of asymmetry and linear size evolution of radio galaxies and quasars and 95% confidence interval in parenthesis.
The striking result of the comparison of the fitted parameters is the stronger evolution of the asymmetry and for radio galaxies than for quasars, in concordance with the faster decrease of radio galaxy linear sizes reported by Chyzy & Zieba (1993). Furthermore, quasars, contrary to the radio galaxies, show rather weak dependence on radio luminosity (small values) in common for all the discussed geometrical parameters. The differences between parameters found for radio galaxies and quasars are statistically significant (at least at 95% of confidence level).
Figure 1: Estimated cosmological evolution for geometrical parameters of the radio galaxy () and quasar () structures
The possible causes underlying for the evolution of geometrical parameters seems to be similar for radio galaxies and quasars. For both types of objects with increasing redshift there is a fast decrease in their sizes, increase in bending and slower increase in arm asymmetry: see Figure 1.
Extragalactic radio sources are observed as (elongated) structures being rather far from spherical symmetry, so knowledge of their orientation to our line of sight is an essential part of understanding their intrinsic structure and possible strong selection effects. Properly speaking, in the simple unified scheme proposed by Barthel (1989), the differences between quasars and radio galaxies are simply treated as a result of a strong orientation effect.
However, the differences found at a high statistical level between quasars and radio galaxies in cosmological evolution of structure parameters seems to contradict the unification scenario based entirely on the viewing angle. The intrinsic structure of extragalactic radio sources must then play an important role in the distinction between these two classes of objects. Of course the line of sight angle should be taken into account, as it is a factor which influences the observed structures of physically different sources. As radio morphology is closely connected with the state of galactic surroundings we may conjecture that the evolutionary state of galactic environment associated with quasars and radio galaxies slightly differs even if their deep interiors are identical.
Another example supporting this point of view is the comparison of liner sizes of nearby () radio sources. In our data sample of complete 3CR and GB/GB2 surveys, contrary to the unification scheme prediction, the linear sizes of radio galaxies (the median ) are less at of significance level in Student's test than quasars sizes (with the median ). A similar tendency was spotted by Singal (1993b) in a large data set. The higher linear sizes of nearby quasars cannot be attributed to their possible higher redshift in comparison to radio galaxies since if the cosmological evolution of linear sizes is homogeneous the nearby quasars sizes would be even underestimated.
Recently, there has been proposed a hypothesis (Antonucci (1993a),Antonucci (1993b)) that the problem of radio source sizes can be caused by separate group of FRI dull galaxies - with weak emission lines and continuum - which do not participate in the unification with quasars. However, the additional estimation of the model for our radio galaxies but without the weakest objects () has not disclosed any statistical difference in the description. Only when the weakest and intermediate power radio galaxies (, roughly corresponding to ) had been excluded was the description of galaxy linear sizes different. In that case the model parameters were estimated at unsatisfactory confidence levels that may be attributed either to low density of objects in space, or to higher variety of morphological properties of objects. However, the change in the description of radio galaxy structures cannot be caused by suggested FRI galaxies because they are less powerful objects with different type of structure than those extracted in the last evaluation.
Nevertheless, the interesting possibility is that a part of low power radio galaxies of FRII type, which cover at least the redshift range up to , do not posses the broad line region (BLR) and hence do not participate in the unification with quasars. This would explain the observed evolution of relative numbers of quasars to radio galaxies. Supposing that the absent-BLR objects have relatively smaller and more asymmetric structures than the remaining galaxies they would also explain relatively low median linear sizes of all nearby galaxies, and would also account for the correlation of asymmetry with radio power (in nearby galaxies those which are less powerful are also more asymmetric and smaller in size, which more than likely arises from the interaction of expanding structures with the surrounding medium).
In this context it is intriguing that the close radio galaxy Cygnus A does not show BLR even in its polarized flux spectrum (Antonucci (1993b)). Further spectropolarimetric observation of all nearby radio galaxies should reveal whether the subgroup of FRII objects without BLR really exists and hence whether this supplementation of simple Barthel's unification scheme is correct.
The investigation of radio structures of quasars and radio galaxies shows that the cosmological evolution of geometrical properties of these two AGN types are different. As well as linear size, the arm asymmetry and bending evolve more strongly with epoch for radio galaxies and their dependence on radio luminosity is also stronger for radio galaxies than for quasars.
These findings seem to contradict the pure unification scenario based entirely on the viewing angle and may reveal a slightly different state of environmental conditions established during the evolution of these objects.
The other attractive possibility explaining the considered data is to admit the existence of subpopulation of moderate redshift FRII galaxies without BLR and with slightly smaller and more asymmetric structures than the remaining part of the observed radio galaxies.