Annu. Rev. Astron. Astrophys. 1980. 18: 321-361 Copyright © 1980 by Annual Reviews. All rights reserved

VI. RELATIVISTIC JETS AND OPTICAL POLARIZATION

We have reviewed in Sections III and V the suggestion by Blandford & Rees (1978) that in AO 0235+164 and similar objects relativistic jets may be responsible for beamed polarized optical radiation, as well as the very high radio brightness temperatures and apparent superluminal expansion. If the superluminal sources are beamed with cone angle of ~ 1 / , and 5 is required for the more more rapidly separating doubles, then for every source beamed toward us there must be 100 sources at the same distance with misdirected beams. Analyzing the source counts for QSOs, Scheuer & Readhead (1979) identify these two classes as radio-loud flat-spectrum quasars, and the much larger number of radio-quiet objects. The radio emission from double-lobe or extended steep-spectrum sources is not beamed, so the statistics of occurrence of central flat-spectrum sources in such sources can be used to measure independently the degree of beaming. Scheuer & Readhead find 2 by this method.

In this section we consider the arguments that can be made for the rapidly variable, highly polarized optical emission characteristic of blazars being produced in relativistic jets. The situation is closely related to but somewhat different from the radio emission. As we have seen, the ratio of polarized optical flux to high-frequency radio emission varies by over a thousand from flat-spectrum QSOs with no polarized optical emission, through violent QSOs like 3C 345, to weak radio sources like Mkn 421.

From the objects listed in Table 1 we can distinguish two types that must be inherently different, irrespective of any possible beaming. These are the strong emission line objects, and the generally weaker BL Lac (weak-lined or line-free) types. The first group contains the violent variable QSOs that are apparently strongly beamed, 3C 345 with superluminal expansion v = 5c and 3C 279 with v 10c (Cotton et al. 1979). There are also another dozen or so strong-lined sources in Table 1 for which there is no direct evidence of relativistic beaming, but which have similar luminosity and variability of polarized radiation. The space density out to z = 0.7 of these objects is ~ 1/Gpc³, allowing for incomplete coverage in the south. If we assume that these sources are all beamed towards us with s of ~ 5 and corresponding (half) cone angles of 0.2, then the space density of misdirected sources would be ~ 40/Gpc³. This is close to the density of 140/Gpc³ found for local QSOs down to the lowest luminosity class of L_opt = 10⁴⁶ ergs s^-1 considered by Schmidt (1978). Thus we reach a conclusion similar to that of Scheuer & Readhead. If relativistic beaming is responsible for the strong polarization seen in some QSOs, then most QSOs, including radio-quiet ones, must be emitting misdirected beams of polarized light.

If this picture is correct, then there are some interesting consequences. All these essentially identical strong line objects would be emitting most of their energy in the optical-infrared spectrum as isotropic unpolarized emission with the characteristic complex spectral structure, very possibly from thermal emission in an accretion disk (Shields 1978). The energy in the beam, since it does not swamp the emission lines even when directed toward us, is a small fraction of the QSOs energy budget when Doppler enhancement is accounted for. It may be significant that the maximum luminosity of extended radio sources, ~ 10⁴⁶ ergs s^-1 (Miley 1980), is roughly equal to the power that would be required in a beamed component of a bright blazar. This supports the idea that the same jet is responsible for both phenomena. The scenario of a polarized, low-luminosity source that is relativistically beamed is quite different from that envisioned by Stockman (1978). If the polarized emission were isotropic, then at its brightest it would have a total luminosity equal to that of the brightest known QSOs (~ 10⁴⁸ ergs s^-1). Unpolarized objects could then be explained as objects whose light was produced initially in the same way but was then depolarized and stabilized by scattering.

The situation for the BL Lac objects is not clear. The many similarities in the observed properties suggest we are seeing something closely related to the polarized quasars but without the strong emission lines and associated unpolarized continuum. In terms of the relativistic beam models, it may be that in the strongest BL Lac's, like AO 0235 + 164 or OJ 287, we are viewing directly into a similar beam, and the weaker ones are the same thing viewed from an oblique angle. Alternatively, or in addition, we could be viewing a different population of weaker objects head on.

Direct evidence that our view is sometimes oblique is provided by the two objects that lie at the center of double-lobe radio sources, 1400+162 and 3C 390.3. It is hard to guess what the angle between the lobe axis and the line of sight might be, but at least in 3C 390.3, which has an unusually small ratio of the lobe size to separation (Harris 1972), it would seem that it cannot be small. The fact that in both these objects the polarization is stable in angle and is aligned with the radio axis suggests that the property of stable position angle in many of the fainter BL Lac objects (Section III) may be related to their being viewed obliquely.

In a relativistic jet theory the misdirected jets that we argue must also be present in radio-quiet quasars could be emitting at the same strength of radio emission and polarized optical radiation that we see in BL Lac objects, and would pass unnoticed. The luminosity of the weaker BL Lacs of 10³¹ ergs s^-1 Hz^-1 would be detected at less than 1 mJy for z > 0.3, and the optical polarization could easily be diluted to less than 1% by the unpolarized continuum associated with the emission lines. It is tempting to speculate that the weakly aligned polarization found in double-lobe radio quasars could be of this type. If this were correct it would imply that the optical flux is not extremely dependent on angle, and that relatively low values of (< 2) are typical.

A case that at least some BL Lac objects are viewed at small angle to the jet axis can be made if we identify the steep-spectrum extended component in some objects as radio doubles seen end on (M.J. Rees, private communication). This idea can be developed as follows. Three of the objects in Table 1 at z < 0.1 show such a component to their radio emission, namely 3C 84, 3C 371, and PKS 0521-36. A lower limit to the typical angle of view of these objects can be derived if we assume they emit the steep-spectrum component isotropically. The space density of all radio sources of similar luminosity is some 30 times greater than that of these objects. Thus if all radio galaxies would show polarized nuclei when viewed end on, then we deduce the typical angle of view for the three objects is ~ 0.25. Larger angles will be needed if only a subset of radio galaxies are involved. If the strong extended component of PKS 0521-36 and 3C 371 were shown to be a halo like 3C 84 and not two lobes closely spaced by foreshortening, this would be strong evidence that these objects were being viewed at close to random angles, since halo sources are rare. On the other hand, if these two sources showed core-jet structure it could indicate that even the low-frequency component is relativistically beamed. Such beaming is required for the strong, steep-spectrum emission of the classic OVV quasars in the statistical analysis of Scheuer & Readhead (1979).

One prediction made by relativistic beam models is that the fluctuation time scale for relativistic jets seen head on will be more rapid than when viewed obliquely. The fact that the time scale of one day for the extremely bright source B2 1308+326 is not markedly different from much fainter sources is given as an argument against relativistic beams by Moore et al. (1980). However, we note that rather modest values of of 2 or 3 give extremely large intensity enhancement in the forward over oblique directions, while causing modest changes in time scale. Given our present knowledge of variability, such changes could easily have escaped notice. In fact B2 1308+326 has as large an amplitude variation from day to day (allowing a factor two for cosmological redshift) as any known object. A good test of these ideas would be a comparison of the time scales for changes in the polarization in the most luminous sources with the time scales in sources where a restricted range of position angle or the presence of a double radio source indicates an oblique view. In general, these latter sources are not as well studied as the wildly varying ones.