|Annu. Rev. Astron. Astrophys. 1977. 15:
Copyright © 1977 by Annual Reviews. All rights reserved
It is the continuum source that provides the extraordinary luminosity of Seyfert galaxies. The continuous spectra do not look like the composite stellar spectra that characterize other galactic nuclei. At least three different radiation mechanisms - stars, nonthermal radiation such as synchrotron radiation, and reradiation from heated dust - have to be combined in various ways to explain Seyfert continua (e.g. Neugebauer et al. 1976). Absorption lines that presumably arise in stars are seen in the spectra of some Seyferts, even bright Sy 1 (Osterbrock et al. 1976). However, the energy distribution in the continua cannot be explained with starlight. Most Seyfert nuclei have power-law spectra, f -, so most of the radiation is attributed to a nonthermal source. This seems to be so for all Sy 1. For the Sy 2, there are indications that the continua may be due primarily to stars heavily reddened by dust, although this is not conclusive. Certainly, a larger fraction of stellar radiation is present in Sy 2 spectra than in Sy 1 (Osterbrock 1976, Neugebauer et al. 1976).
The first explicit demonstration of a Seyfert power-law spectrum was by Oke & Sargent (1968). In the visible spectrum, the many Seyferts observed since usually have indices 1.0 < < 2.0, although can be as large as 2.7 (Oke 1972, Shields et al. 1972, Osterbrock 1976, Osterbrock et al. 1976, Neugebauer et al. 1976). As infrared observations have accumulated, it has become clear that most Seyferts are strong infrared sources that can be accounted for by extensions of the power-law spectra (Penston et al. 1974, Neugebauer et al. 1976, Stein & Weedman 1976). The infrared continua are important to observe in detail because they contain the bulk of the bolometric luminosity (Rieke & Low 1972). However, there is at present a highly inhomogeneous collection of infrared data for Seyfert galaxies. In some cases, the inference has been made that the existence of strong infrared radiation means that objects are astrophysically similar. This is not the case, and it is important to learn how to distinguish between nonthermal infrared radiation and dust reradiation. This is significant even when comparing Sy 1 and Sy 2.
The infrared continuum of NGC 1068 indicates the difficulties in interpretation. It is the only Seyfert galaxy whose infrared spectrum has been followed until turnover, at about 100 µ (Telesco et al. 1976). This turnover mimics synchrotron self-absorption so the entire spectrum might be attributed to a nonthermal source on the basis of this observation. In fact, however, the infrared spectrum is attributed to dust reradiation (Jones & Stein 1975), primarily because the infrared nucleus is resolved (Becklin et al. 1973). The detection of polarization in the optical continuum was initially taken as evidence that the source was nonthermal (Visvanathan & Oke 1968), but Jones and Stein pointed out that dust would also polarize the continuum. Very convincing evidence that dust really does account for the polarization is given by Angel et al. (1976), who found that the continuum is circularly polarized. This would be very unlikely for a nonthermal source but common for dust scattering. Furthermore, they also found the emission lines to be polarized, which is strong evidence that both lines and continuum are strongly affected by dust. We therefore have several independent evidences that the nucleus of NGC 1068 is embedded in dust. A similar conclusion applies to other Sy 2, based on less comprehensive data (Neugebauer et al. 1976). The difference between this and the circumstances deduced for Sy 1 is an important demonstration of the necessity to distinguish between at least two classes of Seyferts.
It is known that continuum radiation exists in Seyfert galaxies out to 10 µ and, in a few cases, to 35 µ (Rieke & Low 1975a). The only Seyfert for which the long wavelength cutoff has been observed is NGC 1068, whose spectrum turns over at about 100 µ. This is deduced from broad-band measures centered at 93 µ and 140 µ (Telesco et al. 1976). Seyfert galaxies are not generally strong radio sources (Kellermann 1972), as indicated by the fact that less than 10% of them are 3C objects. Sramek & Tovmassian (1975) surveyed 506 Markarian galaxies at 6 cm, of which 51 were Seyferts. Only 14% of the Seyferts were detected to a flux limit of 30 mJy. They did find that virtually all detections were of Sy 2, which was the only correlation found between optical and radio properties. A more sensitive survey by de Bruyn & Wilson (1976) with the Westerbork telescope detected 50% of the 43 Seyferts they observed, to a limit of 4 mJy. They confirm that Sy 2 are generally the stronger radio emitters but note that no extended double radio sources are associated with Seyfert galaxies. Another correlation including various Seyferts has been reported between the infrared and 21 cm continuum radiation (Van der Kruit 1971, Rieke & Low 1972). It is not obvious that there is astrophysical meaning to this because such diverse objects as Sy 1, Sy 2, narrow-line galaxies like NGC 4385 and 7714, and weird M82 are all included. Since there is substantial evidence that the infrared radiation from some of these objects is nonthermal, but from others is dust reradiation, it does not now seem appropriate to conclude that there is necessarily a physical relation between the infrared and radio spectra.
Even for those few Seyferts like 3C 120 that are strong radio sources, the radio continuum lies substantially below the extrapolation of the infrared spectrum at 10 µ (Shields et al. 1972). Somewhere - at millimeter, submillimeter, or far-infrared wavelengths - the spectra of all Seyferts must turn over. The point at which they do so is critical for determining their bolometric luminosities. For example, estimated bolometric luminosities for Sy 1 would differ by factors of two to four depending on whether the turnover is put at 3.5 µ or 350 µ (Stein & Weedman 1976). (For the power-law index = 1, the same bolometric luminosity is contained within each decade of frequency, so a spectrum from 350 Å to 350 µ would be twice as luminous as one from 350 Å to 3.5 µ).
As discussed above, the infrared continuum in Sy 1 is attributed to nonthermal radiation while that in Sy 2 probably contains substantial dust reradiation. There must therefore be an extremely luminous ultraviolet source in Sy 2 such as NGC 1068 to explain the production of the emission lines and the infrared reradiation (e.g. Neugebauer et al. 1976, Jones & Stein 1975, Shields & Oke 1975a). Whether this source is thermal or nonthermal is not known, but Sy 2 can be interpreted consistently by invoking large numbers of hot stars as the radiation source (Adams & Weedman 1975, Angel et al. 1976, Harwit & Pacini 1975). ther than NGC 1068, the Seyfert that shows the strongest evidence for dust reradiation is the pathological Sy 1, Markarian 231. In addition to having a very steep infrared spectrum, Mkn 231 has the silicate signature at 10 µ that characterizes dust radiation for infrared sources (Allen 1976 and private communications therein). The initial distinction of Mkn 231 was because of the extremely strong absorption lines of NaI and CaII in its spectrum (Arakelian et al. 1971, Adams & Weedman 1972, Adams 1972a). These lines have components and are blueshifted relative to the emission line spectrum, so they are attributed to interstellar absorption within the nucleus. There are also absorption lines attributed to early type stars at the emission-line redshift. The presence of extreme interstellar absorption, a steep Balmer decrement, and unusual continuum colors are all evidence that this galaxy is heavily dust reddened. A thorough study of Mkn 231 has been prepared by Boksenberg et al. (1977). The empirical demonstration that structured absorption lines, blueshifted relative to the galaxy, can arise in an Sy 1 is an important analogy for the absorption line QSOs. However, Boksenberg et al. (1977) emphasize that the absorption lines are broader and shallower than in most QSOs. Mkn 231 has also been reported to be the most luminous galaxy in the universe (Rieke & Low 1972). While it is of great interest to understand how Mkn 231 achieved such a distinction, it must be realized that it also has unique spectroscopic properties so that an explanation of it may not be generally applicable to other Seyferts.
Empirical evidence therefore exists that Seyfert galaxies have a luminous continuum from the far infrared down to the ionizing ultraviolet. How much further does it go? The only Seyferts so far detected as X-ray sources are NGC 1275 and 4151 (Gursky et al. 1971) and possibly NGC 3783 (Cooke et al. 1976). NGC 1275 is intrinsically much more luminous but, as remarked previously, is atypical of Seyferts and more, nearly resembles the giant elliptical radio galaxies. Other Seyferts have been observed but not detected (Ulmer & Murray 1976). Baity et al. (1975) detected hard X-rays from NGC 4151 and concluded that the X-ray luminosity was comparable to that in all other parts of the spectrum. These X-rays are too strong to arise from an extrapolation of the power-law spectrum in the optical, but they could arise from a similar nonthermal mechanism such as synchrotron radiation. Other suggested alternatives are that the X-rays are Compton scattered optical photons or are thermal emission from a very hot gas (Baity et al. 1975), but the soft X-ray data imply that the X-rays are not thermal and probably arise in the same volume as the optical nonthermal spectrum (Ives et al. 1976). Baity et al. emphasize that NGC 4151 could not be typical of Seyfert galaxies or else the contribution from Seyferts alone would give a diffuse X-ray background 100 times greater than observed. On the basis of this, it seems that Seyferts are not in general strong X-ray sources.