| Annu. Rev. Astron. Astrophys. 1980. 18:
321-361 Copyright © 1980 by . All rights reserved |
As mentioned in the introduction, most active extragalactic objects show optical and infrared polarization of a percent or less. At these low levels, the case for any detected polarizations being intrinsic to the nuclear emission is much weaker than for the blazar class. The polarization may also be due to scattering by dust or free electrons or due to transmission through aligned grains either in the host galaxy or in our own. In the first half of this section, we review the polarimetric observations of Seyfert galaxies, where there is good evidence that the polarization is due to dust. Thus, rather than include the highly polarized Seyferts with sources of intrinsically polarized emission in Sections II and III, we discuss them here. For the majority of QSOs, whose polarizations are quite low, the source of the residual polarization is unknown. We review recent survey work of QSOs in the second half of this section.
SEYFERT GALAXIES Since the most detailed polarimetric observations to date have concentrated on two of the brightest Seyferts, NGC 1068 and NGC 4151, these are reviewed separately below. The first general polarization surveys of Seyferts were undertaken by Dombrovsky, Hagen-Thorn, and other collaborators and are summarized in Dombrovsky & Hagen-Thorn (1968), Dombrovsky et al. (1971), and Babadzhanyants & Hagen-Thorn (1975). This work has also been reviewed by Hagen-Thorn (1974) and in Maza's dissertation (1979). Although most of these observations utilized a rather large aperture (26") which resulted in severe galaxy dilution of the nuclear light, these authors correctly attributed the observed polarization to the nuclear light, and found the strong wavelength dependence of polarization in NGC 1068 and NGC 4151, and detected weak polarization in NGC 3227 and NGC 7469 and several others. Because of the variable polarization observed in NGC 1275 (Section II) and NGC 4151 along with the nonthermal radio emission observed in NGC 1275 and other radio galaxies classified as Seyferts, the polarization was generally attributed to optical synchrotron radiation. However, a more recent polarization survey by Maza, Martin & Angel (1980) and spectropolarimetric work of several groups indicate that this is not the case.
Maza, Martin & Angel's survey of 47 Seyferts found only 8 with polarizations greater than 1.5% in a 4" aperture: 5 Type 1's - Mkn 231, Mkn 486, Mkn 376, NGC 6814, IC 4329A; and 3 Type 2's - NGC 3227, Mkn 348, and Mkn 3. [Note: the polarization in NGC 6814 is probably local interstellar polarization (Maza 1979).] Additional multicolor polarimetry showed that most highly polarized Seyferts have stronger polarization in the blue than in the red with little rotation of position angle with wavelength. Spectropolarimetry of Mkn 376, IC 4329A, Mkn 231, NGC 3227, Mkn 3, and NGC 3516 (Stockman, Angel & Beaver 1976, Thompson et al. 1980) show that the emission lines and continuum in these sources have similar polarizations. Together with the strong wavelength dependence of polarization, this argues for dust scattering (see NGC 1068 below) or in at least one case, IC 4329A, an edge-on spiral, for transmission through aligned grains as the origin of the optical polarization. The case for polarization due to dust is not as strong for those few sources with a polarized continuum but unpolarized emission lines (e.g. NGC 4151 below, and our preliminary results for Mkn 486).
We must emphasize that observations of generally weak polarization in Seyferts and the absence of strong radio emission do not rule out the possibility of nonthermal emission in the nuclei that has been reprocessed inside the emission line region or diluted by thermal emission. Indeed, the high X-ray luminosities associated with most of the bright Seyferts (with the notable exception of NGC 1068) argue for a non-stellar origin for much of the optical infrared emission (Mushotzky et al. 1980, Dower et al. 1980).
NGC 1068 NGC 1068, a Type 2
Seyfert (narrow emission lines, FWOI ~ 1000 km
s-1), has had its optical polarization repeatedly measured by
Walker (1964,
1968),
Dibaj & Shakhovskoy
(1966),
Kruszewski (1968),
and by
Dombrovsky and collaborators (see above). Apart from the early, rather
crude measurements, the optical polarization and luminosity of the
nucleus appear constant. The polarization measured through a
2" aperture increases toward the blue, roughly as
P 
-3
from ~ 0.6% at 0.8 µ to ~ 11% at 0.32 µ. The
position angle increases gradually from
~ 94° to 102° over the same spectral range
(Angel et al. 1976).
Using data obtained through a large aperture (~ 10") Visvanathan & Oke (1968) and Kruszewski (1968) found the polarized flux to be essentially constant with wavelength, suggesting a flat, nonthermal component to the nuclear light. However, spectropolarimetric observations by Angel et al. (1976) showed the polarization of the permitted emission lines to be similar to that of the surrounding continuum. This evidence, together with the strong wavelength dependence and the observed large ellipticity, strongly point to dust scattering as the polarizing agent (see Section V). Infrared observations by Knacke & Capps (1974) and Lebofsky, Rieke & Kemp (1978) indicate strong wavelength dependence on degree and position angle of polarization in the spectral range 1.2-10 µ which may be due to an added nonthermal component in the region 1-5 µ or a complex cloud geometry around the nucleus (Jones et al. 1977, Elvius 1978). While NGC 1068 is the best-studied Seyfert, Maza (1979) points out that in visible light its core luminosity is very weak relative to the host spiral galaxy. If it were at redshifts typical of the majority of known Seyferts, its polarization would be practically undetectable, being less than 0.5% with the 4" aperture used in the survey.
NGC 4151 The spectrum of NGC 4151 shows strong
narrow and broad emission line
components and is generally classified as a Type 1 Seyfert (FWOI
5000 km s-1). Unlike NGC 1068, the nuclear luminosity and
degree of
polarization of NGC 4151 does vary on a time scale of years, though
with little change in the position angle
(Dombrovsky &
Hagen-Thorn 1968,
Babadzhanyants, Hagen-Thorn
& Lyutyi 1972,
and Kruszewski 1977).
The degree of polarization is small, ~ 1%, and is roughly
independent of wavelength into the near infrared, where it falls to ~
0.1% at 2.2 µ
(Kemp et al. 1977).
Spectropolarimetric observations by
Thompson et al. 1979
and Schmidt & Miller
(1979)
show the
polarization of both the broad and narrow emission line components are
much weaker than that of the surrounding continuum. These observations
are consistent either with nonthermal continuum diluted by thermal
emission and starlight or with scattering within the broad-line
emission region.
While the weak but variable polarized component of the optical continuum in NGC 4151 resembles that seen in the blazars, it should be noted that the radio emission is far weaker than any object in Table 1. In addition, the radio source is not compact but is extended over ~ 300 pc, with a steep spectrum. De Bruyn & Willis (1974) observed a flux of 0.135 Jy at 5 GHz with a spectral index of 0.74.
QSOs Polarimetrists who first searched quasi-stellar sources for the high polarizations expected from synchrotron radiation were rewarded with the strong and rapidly variable polarization of the optically violent variables 3C 345, 3C 446, 3C 454.3, 3C 279 (Section II). Other QSOs such as 3C 273 (e.g. Whiteoak 1966, and Liller 1969) had disappointingly low polarizations. More extensive polarimetric surveys by Appenzeller & Hiltner (1967) and Visvanathan (1968) failed to discover any additional strongly polarized QSOs. The sensitivity of these surveys was such, however, that polarizations of a few percent could not be unambiguously detected.
Recently,
Stockman & Angel
(1978)
have reported preliminary results
of a large polarimetric survey of bright QSOs,
V 
17 (see also
Stockman 1978).
They find that ~ 90% of this sample have low
polarizations, < 2%. Much of the average polarization, ~ 0.6%, is due
to local interstellar polarization. With the exception of the unusual
object PHL 5200 (Section II), none of the
radio-quiet QSOs are
strongly polarized. Since the vast majority of QSOs are believed to be
radio quiet, the blazar type must be extremely rare. Indeed, the gap
in polarimetric properties between the few strongly polarized ones and
the remaining QSOs (even radio emitters) suggests that the blazar QSOs
are a qualitatively distinct subclass of objects
(Sections II and V).
The origin of the small polarizations seen in the majority of QSOs
is unknown. Multicolor observations indicate that, while there is a
statistical tendency for the polarizations to increase to the blue,
the optical polarization is essentially wavelength independent
(Moore, Stockman & Angel
1980).
The weak polarizations and their lack of
marked wavelength dependence suggests that, unlike Seyfert nuclei,
QSOs have little surrounding dust
(
<< 1).
Stockman, Angel & Miley (1979) found that in those QSOs with double radio lobes the position angle of polarization was roughly aligned with the radio structure. Thus, unlike the most variable blazars, the position angle must be relatively constant for time scales of 106-107 yrs.
As with the Seyfert galaxies, the lack of strong polarization does not preclude a luminous nonthermal central source. Tananbaum et al. (1979) find many QSOs with an X-ray luminosity comparable to their optical-infrared luminosity, thus indicating a powerful, nonstellar engine. However, the very weak polarizations do argue for efficient reprocessing of any nonthermal optical emission or for dilution by an unpolarized, isotropic source which is probably thermal (Section V). More detailed study of the low polarization sources (wide baseline multicolor and spectropolarimetric observations; continued monitoring) will give valuable information concerning the origin of the polarization and should be capable of distinguishing between reprocessing and thermal dilution models.