ACTIVE GALAXIES, SEYFERT TYPE GAIL A. REICHERT Seyfert galaxies are galaxies having bright compact nuclei and exhibiting in their spectra broad high-excitation emission lines arising from a wide range of ionization states. First systematically studied by Carl K. Seyfert in 1943, they are one of several classes of galaxies that show signs of intense and violent activity within their nuclei. DISCOVERY Examples of Seyfert galaxies have been known since the early 1900s. However, it was not until the middle 1960s, when it was realized that these galaxies might be the lower luminosity cousins of the mysterious quasars, that Seyfert galaxies began to attract widespread interest. At about the same time, the pioneering work by B.E. Markarian and collaborators greatly expanded the list of known Seyfert galaxies, and showed that nonthermal activity in galaxies is far from uncommon. Searches for Seyfert galaxies usually rely on the fact that the continuum emission from the active nucleus differs markedly from the thermal emission seen from stars. Instead, the nuclear continuum extends to far higher (and lower) photon energies, and is often characterized as a power law. Markarian and his collaborators searched for galaxies with compact nuclei that were comparatively bright in the ultraviolet. Subsequent spectroscopy showed that about 10% of these "UV excess" galaxies were Seyfert galaxies. The search criteria now also include unusual radio, infrared, and x-ray continuum properties as compared to normal galaxies, as well as strong emission lines. Each type of survey yields its own characteristic sample of Seyfert galaxies. Today, more than 600 seyfert galaxies have been identified, and more continue to be found. Seyfert galaxies form a continuous sequence with quasistellar objects (QSOs) in their properties, and share many of their characteristics with other types of active galaxies. Indeed, all active galaxies may actually be the same kind of object, involving essentially the same physical phenomena but perhaps under different external and/or observational conditions. The term active galactic nuclei (or AGN) is commonly used to mean all active galaxies, including QSOs. Although the discussion here will concern only Seyfert galaxies per se, much of it will also apply for AGN in general. SEYFERT GALAXY TYPES Seyfert galaxies are separated into two basic types, according to the relative widths of emission lines arising from permitted transitions in hydrogen, helium, and so forth, and those arising from forbidden transitions. In Seyfert type l (hereafter abbreviated Sy I) galaxies, the permitted lines are broad, typically with full widths at half maximum (FWHM) of several thousand to ten thousand Kilometers per second, whereas the forbidden lines are much narrower, with typical FWHM of several hundred Kilometers per second. In Seyfert type 2 (hereafter Sy 2) galaxies, the permitted and forbidden emission lines have comparable widths, similar to the forbidden line widths in Sy 1 galaxies. The original scheme has since been expanded to distinguish Seyferts whose permitted lines show both narrow and broad components as type I.N, with N ranging from 1 to 9.In these objects, the narrow components of the profiles resemble the lines in Sy 2 galaxies whereas the broad components resemble the lines in Sy 1s. The number N increases with the relative strengths of the narrow components. For example, Sy 1.5 galaxies clearly show both narrow and broad components, whereas in Sy 1.9s the broad compo- nents are quite weak. It is not clear whether the Seyfert types 1.N are subclasses of the first Seyfert type or separate classes intermedi- ate between Seyfert types 1 and 2. A few objects have also been observed to change in their Seyfert types. Finally, a class of narrow line Sy 1 galaxies has been identified. These objects have permitted lines that are not much broader than the forbidden lines (FWHM- 1000-2000kmS)and appear to be Sy 1s at the low end of the distribution in velocity width. COMPARISON OF PROPERTIES BY TYPE Is the separation into Seyfert types physically meaningful? The first step in answering this question is to compare other observable properties, such as radio brightness and size, nuclear luminosity, continuum spectrum, and so forth, and determine how these vary with type. EMISSION LINE SPECTRA Overall, the emission line spectra of the various Seyfert types are very similar. They show the same atomic transitions and similar ratios of the emission line fluxes. However, there are subtle differences. For example, the Sy 1.8s and 1.9s resemble Sy 2s in the ratios of their narrow line fluxes, whereas the Sy 1.5s are more like Sy 1s. Sy 1.8s and 1.9s also tend to show large broad Hx to broad Hb flux ratios, much greater than the ratios predicted by normal recombination theory. The large ratios indicate that dust may play an important role in reddening the broad line spectra of these objects. In contrast, the broad line ratios for Sy 1s and 1.5s tend to show little if any reddening. Sy 1 galaxies also tend to show higher ionization transitions than most Sy 2s. Ionized species as high as [Fe x]^6375 are seen in Sy 1s. Broad Fe ** emission from many blended multiplets appears to be universally present in Seyferts with broad line components (including the narrow line Sy 1s), but so far has been observed in only a handful of Sy 2s. Broad Fe ** emission is also observed in other classes of broad line AGN. There is some evidence that broad line radio galaxies tend to show weaker Fe ** emission, but some of this may be due to increased line blending. The forbidden lines have different velocity widths, such that lines of higher critical density (i.e., density at which the level collisionally de-excites) and/or ionization energy tend to be broader. Line width appears to correlate with ionization energy in Sy 1s and with critical density in Sy 2s, although there are many exceptions. LUMINOSITY OF THE ACTIVE NUCLEUS The nuclei of Sy 1s and 1.5s tend to be more luminous than Sy 2 nuclei. At the highest luminosities there are no Sy 2s. Practically all QSOs are broad line objects. OPTICAL/ULTRAVIOLET CONTINUUM OPTICAL/ULTRAVIOLTET CONTINUUM The presence of a featureless, nonthermal optical/UV continuum is a defining characteristic of Seyfert galaxies. However, the nonthermal continuum is much weaker in Sy 2 galaxies than in Sy 1s. Sy 2s therefore tend to have redder colors than Sy 1s, and so are often missed from surveys selecting for blue colors or ultraviolet excess. HOST GALAXY When the host galaxies of Seyfert nuclei can be classified, they are nearly always spiral or "barred" spiral galaxies. Many Seyfert nuclei are found in interacting galaxies and/or in galaxies with peculiar morphology. In contrast, strong radio galaxies are nearly all ellipticals. RADIO BRIGHTNESS AND SIZE Seyfert galaxies tend to be faint radio sources, both fainter and less extended than the sources in radio galaxies. Sy 2 galaxies tend to be brighter radio sources, relative to their optical and ultraviolet emission, than Sy 1s. Sy 2s also tend to be more extended radio sources than Sy 1s. Sy 1.5s appear to be intermediate between Sy 1s and Sy 2s in both brightness and extent. The Seyfert radio sources that are spatially resolved are nearly always linear in structure. A small number of Seyferts have been observed with high enough spatial resolution that the emission from the narrow lines can be mapped. In these cases, there is a very strong tendency for the radio and narrow line emission to be coaligned. INFRARED EMISSION Seyfert galaxies are typically strong infrared emitters, but their infrared properties remain poorly understood.Most presently available data have insufficient spectral and spatial resolution to separate the nonthermal emission from the active nucleus from other sources, for example, thermal photospheric emission from stars in the host galaxy and thermal emission from dust heated either by the active nucleus or by the intense bursts of star formation often found in in active galaxies. Present data indicate that the infrared-visible continuum spectra of Sy 1s are generally dominated by nonthermal emission from the active nucleus, whereas the infrared spectra of Sy 2s may be dominated by thermal emission from extended regions of hot dust. Seyfert galaxies found in infrared surveys tend to be more heavily reddened than those found by UV excess. Dust opacities are many times lower in the infrared than in the visible or ultraviolet, so that infrared emission is far less affected by reddening. Many more previously unknown Sy 2s (which tend to be more heavily reddened than Sy 1s) have been discovered via infrared surveys than Sy 1s. Hence the "true" proportion of Sy 2s to broad line Seyferts is roughly 4:1, far higher than the 1:3 ratio found for UV excess surveys. Infrared surveys selected on the basis of "warm" (i.e., color temperatures of approximately hundreds of degrees kelvin) infrared colors also tend to be biased in favor of narrow line Sy 1s, which generally have warmer infrared colors than most Sy 1s. X-RAY EMISSION Active galaxies are almost always strong X-ray emitters relative to their optical brightness, and X-ray surveys have led to the identification of many previously unknown AGN. However, almost all of the Seyferts detected in X-rays show broad components to their emission lines at some level (although in some cases the broad components are extremely weak, weaker than those found in Sy 1.9s). Bona fide Sy 2s tend to be much weaker X-ray emitters than broad line Seyferts, typically by factors of 30- ^100. However, the X-ray spectra of the few that have been measured do not differ appreciably from the spectra of broad line Seyferts. Although X-ray spectra do not appear to differ systematically by Seyfert type, the spectra of low luminosity objects tend to show the effects of photoelectric absorption by large amounts of intervening material. In some cases the absorbing material appears to cover only a fraction of the X-ray emitter. The spectra for higher luminosity objects rarely show such absorption. There is some evidence that suggests that the probability of intervening absorption increases as luminosity decreases, rather than the covered fraction or the intervening column density. OPTICAL POLARIZATION OPTICAL POLORIZATION Optical polariration observations provide some of the clearest evidence linking Sy 1s and 2s. Only a small fraction of Seyfert galaxies emit polarized light, typically at the level of a few percent of the total emission. Polarized Sy 2s tend to have relatively higher polarization that is usually independent of wavelength, indicating that it is caused by electron scattering.Sy 1s generally have lower polarization due to a variety of causes. In general, polarization from an unresolved source indicates intrinsic asymmetry. The distribution in polarization angle also appears to be bimodal.Sy 2s show polarization that is perpendicular to the radio structure, whereas Sy 1s show polarization that is parallel. Seyferts classified 1.N are polarimetrically similar to Sy 1s, as are low polarization QSOs. Optical spectra in polarized light have been obtained for a few of the brightest Sy 2 galaxies, for example, NGC 1068. The polarized spectra of several show broad components in the permitted lines, Fe ** emission, and featureless nonthermal continua. These Sy 2 galaxies must therefore contain "hidden" Sy 1 nuclei.Except for light scattered and polarized by external regions of hot electrons, the nuclei must be totally blocked from our view. VARIABILITY VARIABILITY Active galaxies can vary substantially, both in continuum output and in the fluxes of the broad emission lines. Continuum variability on time scales as short as 10 min has been reported, whereas broad line fluxes typically vary on longer time scales of ^1 week to a few months. Narrow line fluxes have been observed to vary in only one object, over a period of a few years. The relative levels of broad to narrow line components can change dramatically. Some objects have been observed to change in Seyfert classification from Sy 1s to 1.9s or even 2's, and and vice versa. There are suggestions that the continua of lower luminosity objects may vary on faster time scales than those of higher luminosity objects, but the issue is far from settled. SPACE DENSITIES The true relative space densities of the various Seyfert types are difficult to determine. UV-excess and x-ray surveys tend to be biased against Sy 2s in favor of broad line Sys, whereas infrared surveys tend to be biased in favor of Sy 2s and narrow line Sy 1s. Sy 1.9s can also be misclassified as Sy 2s if the spectra are of low signal/noise so that the broad components are not detected. Unbiased emission line surveys suggest that there are roughly as many Sy 1.8s and 1.9a as there are Sy 1s and 1.5s, and eight times as many Sy 2s. many Sy 2s. PHYSICAL INTERPRETATIONS OF THE DIFFERENCES BETWEEN SEYFERT TYPES What causes the observed differences between the various Seyfert types? The general picture is far from clear, because our understanding of the physical nature(s) of AGN remains limited. Several possibilities have been proposed: 1.The various types of Seyfert galaxies may actually be fundamentally different kinds of objects. This picture is unlikely because there are many basic similarities, not only between the different Seyfert types, but also between Seyfert galaxies and other classes of AGN. 2. The different Seyfert types may actually be intrinsically the same kind of object, viewed under different external and/or observational conditions. For example, Sy 2s may simply lack the denser, higher velocity regions of gas responsible for the broad emission line components seen in Sy 1s to 1.9s. Alternatively, the broad emission line (BLR) and continuum regions may also be present in Sy 2S, but obscured from the line of sight. Some Sy 2s are known to show otherwise "hidden" Sy 1 nuclei in polarized light and would appear as Sy 1s if viewed from a different direction. Obscuration (and reddening by associated dust) may also explain the differences in x-ray emission between Sy 1s and 2s, the wide range in infrared continuum shapes observed for Sy 1s, and the elongated shapes of the narrow emission line regions in Sy 2s. The problems with this hypothesis are the differences in Sy 1 and 2 radio properties, the presence of weak featureless continua (which should be blocked if the BLA is blocked) in Sy 2s, and the existence of Seyferts that appear to change in Seyfert type without accompanying changes in reddening or obscuration. 3.The various Seyfert types may represent different evolutionary stages in the lifetime of a single kind of object. For example, Sy 1s might evolve into Sy 2s via triggering of bursts of star formation that are left behind when the type 1 nucleus switches off. Alternatively, Sy 2 nuclei might be triggered by interactions between galaxies and might evolve to Sy 1s by expelling the dust from the BLR. The problems with these scenarios are again the differences in radio properties and the fact that some objects can change Seyfert type on short time scales of months. The relative numbers of Sy 2s versus Sy 1s also suggest that the active nuclei must spend most of the time switched off. If this were the case, however, we might expect the narrow line regions in Sy 1s and 2a to show very different excitation properties, which are not observed. Additional Reading Osterbrock, D.E.(1989).Astrophysics of Gaseous Nebulae and Active Galactic Nuclei. University Science Bcoks, Mill Valley,CA. Shipman, H.L.(1980).Black Holes, Quasars, and the Universe, 2nd ed. Houghton MiffIin, Boston, MA. Weedman, D.W.(1977).Seyfert galaxies. Annual Reviews of astronomy and Astrophysics 1569. Weedman, D.W.(1986).Ouasar Astronomy. Cambridge University press, Cambridge. Weymann, R.J.(1969). Seyfert galaxies Scientific Anierican, 220 (No. 1) 28. See also Active GaIaxies and Quasistellar Objects, Emission Line Regions, Active Galaxies and Quasistellar Objects, Infrared Emission and Dust; Active Galaxies and Quasistellar Objects, Interrelations of Various types;active Galaxies and Quasistellar objects, x-ray emission.