|Annu. Rev. Astron. Astrophys. 1977. 15:
Copyright © 1977 by . All rights reserved
Within the past five years, extraordinary progress has been made in the accumulation of spectrophotometric data for Seyfert galaxies. This is because of the multichannel spectrum scanners now available for most large telescopes. Technical details of these instruments can be found in some of the references included in Table 1, primarily the work involving Osterbrock, Oke, or Boksenberg. Earlier studies of bright Seyferts have been repeated with higher accuracy, and precision studies of faint spectral features as well as extensive observations of previously unstudied objects have begun. Perhaps a measure of the progress in our knowledge of Seyfert galaxies is the fact that it is no longer feasible to discuss each one individually in a review. Consequently, recent spectrophotometric studies of the various Seyferts are referenced in Table 1. The review that follows is a synthesis of the generalities that can be deduced from this data.
As the sample of Seyfert galaxies grew, it became obvious that all such objects by no means had identical spectra. The Seyferts were therefore subclassified strictly on the basis of their spectroscopic properties in an attempt to identify those subgroups with similar physical properties. This, of course, is the motivation for any astronomical classification scheme. Because Seyfert galaxies are characterized by strong and broad emission lines, the nature of these lines was used as the classification criterion. A simple division into classes 1 and 2 (Sy 1 and Sy 2), which depended only on relative emission-line widths, was used by Weedman (1970, 1973) and Khachikian & Weedman (1971, 1974). Illustrations of spectra and line profiles are given in these references. The Sy 1 have broad hydrogen lines but narrower forbidden lines, such as the classical Seyfert NGC 5548. Other permitted lines are also broad when the Balmer lines are broad though they do not necessarily have the same profiles (Boksenberg et al. 1975a, Osterbrock 1976, Osterbrock et al. 1976). The only such lines observed are HeI 5876, He II 4686, occasional blends of FeII lines, and OI 8446. These features are so weak relative to the Balmer lines that they have not been measured in all Sy 1. It is probably appropriate to infer that the other permitted lines arise in much the same volume as do the Balmer lines. Comprehensive spectrophotometric data on 40 Sy 1 galaxies are given by Osterbrock (1977b). This includes relative line intensities as well as line profile widths.
The Sy 2 are like NGC 1068, having hydrogen and forbidden lines of the same width, up to about 103 km sec-1. The classification into Sy 1 or Sy 2 is most easily made just by comparing H with an adjacent [OIII] line. The simplest explanation for the spectroscopic difference between Sy 1 and Sy 2 is that the broad Balmer lines characterizing Sy 1 arise in a different part of the nucleus than do the narrower forbidden lines, whereas all lines arise together in Sy 2. There is little ambiguity as to which galaxies are Sy 1, where the Balmer lines often have full widths of 104 km sec-1. The problem in deciding whether to call a galaxy a Seyfert is in defining the line width at which an Sy 2 is distinguished from other emission-line galaxies. Those few objects classified Sy 2 whose line profiles have been measured show full widths at half maximum intensity greater than 500 km sec-1 (Weedman 1970), whereas such widths for representative narrow-line galaxies (from the same study) did not exceed 200 km sec-1. The ambiguity between Sy 2 and narrow-emission-line galaxies is most serious in low-resolution spectroscopic surveys. For this reason, there are probably a number of Sy 2 in the Markarian and Tololo surveys that have not yet been observed with sufficient resolution to be classified and so are not included in Table 1. To date, the most detailed study of Sy 2 galaxies is by Koski (1976).
After the profile classification was adopted, other properties of the nuclei were found to correlate well with the Sy 1 and Sy 2 classes. A useful secondary classification is the Balmer-line to forbidden-line intensity ratio, especially H: [OIII]. The forbidden lines are unusually strong in Sy 2, with an average H: 5007 ratio of 0.1, whereas H and 5007 typically have the same total flux for Sy 1 (Adams & Weedman 1975). For the narrow-emission-line galaxies, H is also comparable to 5007 in most cases. Correlations with UBV colors also exist because the continuous spectra for Sy 1, Sy 2, and narrow-line galaxies are different. It is then possible to use UBV colors as a secondary classification (Markarian 1973). Color-color diagrams showing results for Seyfert galaxies along with comparative colors of narrow-line galaxies are summarized in Weedman (1973) and Adams & Weedman (1975). Extensive UBV observations of narrow-line galaxies have been made by Huchra (1976), and these can be compared to the Seyfert colors in Table 1. The UBV colors provide reasonable distinctions in most cases among Sy 1, Sy 2, and narrow-line galaxies even if high-resolution line profiles are not available.
It is the Sy 1 that have power-law continuous spectra and the resulting ultraviolet excesses like QSOs. This correlation between the presence of very broad Balmer emission lines and a nonthermal continuum is one of the more interesting results established. Searle & Sargent (1968) pointed out the similar equivalent widths of the broad Balmer lines for Seyferts then known and showed that these were too small for the continuum to be emission from the same gas producing the Balmer lines. Adams & Weedman (1975) showed quantitatively that Balmer-line luminosities and continuum luminosities scale together for Sy 1 over six magnitudes of luminosity. They concluded that the Balmer lines therefore arise primarily from ionization by a radiative source that also produces the visible continuum, in accordance with models such as those of Williams & Weymann (1968). Extensions of the observed power-law continua into the ionizing ultraviolet generally provide enough photons to account for the Balmer-line strengths (Osterbrock 1971, 1976, Stein & Weedman 1976). The nature of the continuous spectra in Sy 2 is not as clear because they are affected significantly by starlight. The probable presence of heavy dust reddening and a possible nonthermal contribution means that the nature of Sy 2 continua is not as well understood as Sy 1.
If cosmological redshifts are assumed, the observed forbidden-line luminosities of Sy 1 and Sy 2 range over comparable values, whereas the Balmer lines can be up to ten times more luminous in Sy 1 (Adams & Weedman 1975). Osterbrock & Koski (1976) have emphasized that the broad Balmer-line wings provide the distinction between Sy 1 and Sy 2. The Balmer-line cores in Sy 1 have similar profiles and intensity ratios relative to forbidden lines as in the Sy 2. So, observationally at least, an Sy 2 could be transformed to an Sy 1 by adding a strong power-law continuum accompanied by the broad Balmer-line wings. Conversely, removing these would change Sy 1 to Sy 2. As Osterbrock & Koski put it, "Galaxies in which the broad Balmer line component is relatively strong are classified as Sy 1, while galaxies in which it is invisible are classified as Sy 2....Physically, different Seyfert galaxies must have different relative amounts of the material and physical conditions that emit these two components." They suggest that the Sy 1 have a higher proportion of very dense gas in which the broad wings arise. A similar conclusion was stated by Neugebauer et al. (1976). However, the ease with which an Sy 2 could be converted to an Sy 1 is not necessarily evidence that there is an evolutionary relation between them. They may be totally unrelated phenomena. One particularly troubling aspect is the implication that Sy 2 may be heavily obscured by dust while Sy 1 are not. In this case, the intrinsic Balmer-line luminosities of Sy 2 would be comparable to Sy 1, and the forbidden lines typically ten times more luminous (Adams & Weedman 1975). Under such circumstances, addition of more Balmer emission would not make the intrinsic properties of Sy 2 match Sy 1.