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.

Go to: Front page, Participants, Group Photograph, Preface, Contributions, Acknowledgements

The quasar fraction in a sample of giant radio sources

Garret Cotter

Mullard Radio Astronomy Observatory, Cavendish Laboratory, Madingley Road, Cambridge CB3 0HE, UNITED KINGDOM


A classic test of the standard orientation-dependent unified schemes has been the assessment of the quasar fraction in complete radio samples. Using a sample of giant radiosources (projected size ) out to redshifts of , I demonstrate that the fraction of broad line objects is smaller than that obtained by comparison of quasars and galaxies in the 3CR sample. I stress that to make more accurate predictions about proposed torus opening angles, one must carefully bin objects by size and apply a knowledge of the distribution of in the population.


1. Introduction

Giant radio sources () have traditionally been of great interest for several reasons. Their large angular sizes give excellent opportunities for the study of source physics, and as they lie at the extreme size limit of the radio source population, they can be used to investigate and constrain models of source lifetime evolution and cosmic population evolution. Taking candidates from a sample of possible distant giant radio sources (Dingley (1992)), we have obtained spectrophotometric measurements on the William Herschel Telescope and have identified 16 giant radio sources at redshifts between and (Cotter, Rawlings & Saunders (1995)).

This sample of giants provides an interesting new position from which to discuss unification schemes for radio loud quasars and powerful radio galaxies (Barthel (1989), Antonucci (1993)). These schemes propose that both types of object are intrinsically similar but present different appearances to observers viewing them from different directions. An obscuring torus hides the nuclear broad line and non-thermal continuum emission from a large fraction of the sky; the opening angle of this torus determines the relative numbers of galaxies and quasars seen by one observer. By comparing the numbers and projected sizes of galaxies and quasars in the 3CR sample, Barthel (1989) concluded that the opening angle of the torus is approximately . There has since been much debate on this issue, centered on the objectivity of the samples used and the difficulty in determining whether or not some objects have broad emission lines. Aside from these points, the calculation of opening angles by this method can suffer from an important bias: for objects of any particular projected size, one selects quasars which are intrinsically larger than galaxies. If one does not have a sample which adequately represents the entire range of in the population, the results obtained by simply comparing the numbers and relative projected sizes of quasars and galaxies may be inaccurate.

2. The sample

The 7C giants sample is drawn from a selection of 35 radio sources from the 7C survey (Dingley (1992), McGilchrist et al. (1990)). The selection criteria were that the flux density at 151 MHz should lie between 0.4 and 1.0 Jy and the angular size should lie between 1.5 and 3.0 arcmin. Of these objects, 30 were identified to a limiting magnitude of using the University of Hawaii 2.2m telescope (Dingley (1992)) and the University of Texas 2.7m telescope (Cotter, Rawlings & Saunders (1995)). Using the 4.2m William Herschel Telescope we obtained spectrophotometric measurements of the identified objects and a blind spectrum (Rawlings, Eales & Warren (1990)) of one of the unidentified objects. The giants sample now contains 16 objects with spectroscopic redshifts and is complete in the magnitude range , corresponding to a redshift range of . Objects brighter than this are to close to be giants and there are several objects fainter than which remain unidentified.

Figure: Spectra of a selection of objects from the 7C giants sample, obtained with the WHT. Wavelength is shown in units of Ångstrom and flux in units of .

Some typical spectra of broad and narrow emission line objects in the giants sample are show in Figure 1. The distribution of the 7C giants on the radio luminosity-linear size () diagram is shown in Figure 2. Radio galaxies from the revised 3C ``LRL'' sample of Laing, Riley & Longair (1983) are shown for comparison.

Figure 2: The distribution in the linear size-radio luminosity plane of the sources in the 7C giants sample (diamonds) and the 3C LRL sample (crosses)

3. Unified schemes

In the redshift range through which the giants sample is complete, there are two objects w broad emission lines and and 13 radio galaxies showing only narrow emission lines. These relative numbers are immediately consistent with the unified schemes-since the largest objects seen will mostly lie in the plane of the sky, we expect them to be predominantly radio galaxies.

It is notable that the 7C objects fill a region of space which is relatively empty in the LRL sample. This ``hole'' identified by Saunders (1982) is thought to be caused by sources dropping out of the 3C sample as they fall in during their own lifetimes (Cotter, Rawlings & Saunders (1995)). The implication is the even LRL does not fully sample the entire range of , so one must try to reconstruct a more complete picture of the distribution by adding in other surveys.

With this new sample one may make a comparison with Barthel's result by combining a selected number of the giants with an equivalent selection of 3C sources. This gives a value of the quasar fraction appropriate to all radio sources in one particular radio luminosity bin at one particular cosmic epoch. I choose the 7C giants in the redshift range , the range used by Barthel, in which the giants sample is complete. These redshifts limits correspond to a range of radio luminosity . In this range the 7C giants sample contains 2 quasars and 8 galaxies of . Objects of this radio luminosity are detected in the LRL sample from zero redshift out to . In the LRL sample, 16 galaxies and 9 quasars fall into the same radio luminosity bin but have linear sizes outside the range of the 7C objects. However we cannot simply take the objects from 7C and the objects from 3C and add them together to define a new sample. The 3C objects sample a co-moving volume of and the 7C objects sample a volume of , so the contributions from each sample must be weighted accordingly. Further, a correction must be made for cosmic evolution. In effect, the size of the sample of 3C objects must be back-evolved to correct for the change in co-moving density which we know has taken place. I use the current best estimate of the radio luminosity function (Dunlop & Peacock (1990)) to make this cosmic evolution correction. The Dunlop and Peacock radio luminosity function estimates that the number density of objects in this radio power range decreased by a factor of approximately four between the high redshift bin and the low redshift bin . The number of 3C objects must be divided by 3.4 because it samples a larger volume, and multiplied by 4 because it samples a region of redshift where the source density is known to be less that the high redshift bin of the 7C giants. Making these corrections, the combined sample contains effectively 12.6 quasars and 27.2 galaxies. These numbers give a torus opening angle of , in agreement with Barthel's estimate for sources of higher , but using objects of similar radio power to those sometimes used to ``disprove'' unification schemes.

The problem with this analysis is that the radio luminosity function contains no information about the cosmic evolution of in the population. When turning the clock back for the 3C sources, I have to assume that their distribution does not change. So although the above estimate of the opening angle includes sources at large that were missed in 3C, there is no indication whether or not the relative sizes of the quasars and galaxies involved are consistent with this fraction. The object of the exercise is not to determine conclusively the torus opening angle; rather to demonstrate the need for careful selection of samples at different flux levels which span as large a range of as possible.

4. Conclusions

This sample provides clear supporting evidence that powerful radio galaxies and radio loud quasars are indeed intrinsically similar objects. Simple techniques of counting the total numbers of quasars and galaxies in entire samples can be influenced by various biases. To determine the opening angle of the torus, and to investigate how this may change with radio source power, one must calculate the quasar/galaxy ratio at different points in the diagram. One may then attempt to convolve a proposed opening angle with the true distribution in the population, for comparison with the observed quasar/galaxy ratio as a function of at any particular radio luminosity.


YERAC 94 Account
Wed Feb 22 16:49:34 GMT 1995