ACTIVE GALAXIES AND QUASISTELLAR OBJECTS, INFRARED EMISSION AND DUST GEORGE H. RIEKE "Normal" galaxies have fairly similar populations of stars in their nuclei, which produce similar nuclear spectra and colors. Some galaxies roughly 1-2% of large spirals and ellipticals-harbor a powerful compact nuclear source that emits copiously outside the spectral region occupied by normal stellar radiation (e.g., in the radio, infrared, ultraviolet, and x-ray). Historically, these active galactic nuclei and QSOs have been discovered by detecting excess radiation in the blue and ultraviolet above that to be expected from "normal" galaxies. However, an abnormal excess in the infrared is at least as ubiquitous among these sources as blue or ultraviolet one; they are described as blue-excess objects only because of the greater ease of making sensitive observations with optical rather than infrared detectors. These infrared excesses can arise from two underlying causes: The source that powers the nuclear activity may emit a nonthermal spectrum that is observed directly in the infrared, or the output of this source may be absorbed by the surrounding interstellar dust and reemitted thermally in the infrared. In the first case, the infrared region usually accounts for a significant fraction of the total source luminosity and therefore can reveal fundamental aspects of the nuclear activity. In the second case, an energetically important infrared component requires that the interstellar material has altered the output spectrum at the wavelengths where the energy has been absorbed, typically in the blue and ultraviolet; the infrared properties help reveal the the underlying characteristics of the nuclear source in these other spectral regions as well as the placement and characteristics of the interstellar medium around the nuclear source. Figure 1 shows OBSERVATIONAL BEHAVIOR The behavior in the infrared of typical representatives of various types of active nucleus and QSO. Although the characteristics of these broad classes are discussed extensively elsewhere, we give here brief definitions with emphasis on specific aspects of interest from the infrared perspective. 1. Type 1 Seyfert galaxies have very broad(-5000 km s-1), high-excitation emission lines that are generated in highly disturbed gas very close (within -alight-year)to their nuclei. Their blue and ultraviolet excesses are typically variable on time scales of a few weeks, and they have strong x-ray fluxes. Type 1. 5 Seyfert galaxies have additional line components of moderate(-1000km s-1)width that originate from gas farther from the nucleus. 2. Type 2 Seyfert galaxies have moderate width, high-excitation emission lines that are similar to the additional line components of type 1.5 Seyfert galaxies. Type 2 galaxies have weaker blue and UV excesses than type 1 galaxies, they are less variable, and their x-ray fluxes are weaker. Both type 1 and 2 Seyfert galaxies have total nuclear luminosities of 10x9 - 10x12 L*(where L* is the luminosity of the sun). 3. QSOs appear to be related to type 1 Seyfert galaxies; the most fundamental difference is their much greater luminosities, which are typically 10x12-l0x14 L*. A minority of QSOs have very strong radio emission ("radio loud"). 4. Blazars are defined by extremely short variability time scales, of the order of a day. They can be strongly polarized and have optical spectra that approximate power laws with indices of about -1; for example, their spectra go as Sv(W m-2 Hz)=Cva, where C is a constant and &~-1. They are frequently strong x-ray sources and, so far as is currently known, always strong high-frequency radio sources. They have apparent luminosities of 10x11-10x14 L*, although a variety of arguments suggest that the emission from these sources is beamed strongly and we see only those where we fall in the beam. As a result, the luminosity estimates are strongly dependent on the detailed model of the source. As indicated by Fig. 1, these four classes of active source also tend to have distinctive infrared behavior. For example, virtually all type 1 Seyfert galaxies have a strong near-infrared excess that seems to fall very steeply toward the optical, becoming weak or undetectable at 1uM and shorter wavelengths. This component is variable, but with smaller amplitude than the ultraviolet variability. These galaxies have modest far-infrared emission, unless they are embedded in a galaxy that has strong far-infrared emission from the disk probably associated with star formation in the outer parts of the galaxy. Virtually all Seyfert 2 galaxies have a strong infrared excess rising steeply toward longer wavelengths from about 3uM, with few if any confirmed observations of variability. Far infrared emission from the galaxy disk is relatively common, and there is evidence that the far-infrared luminosities of the host galaxies tend to be larger than those from a similar set of "normal" galaxies. This suggested association between galaxy-wide properties and the presence of an active nucleus is not well understood. QSOs frequently have a spectrum that falls steeply toward 1uM, where there is an inflection with a much flatter spectrum extending to the blue and ultraviolet. In most cases, the variability damps out dramatically from the blue and visible to the near infrared. QSO far-infrared spectra often show only weak excesses above a vx-0. 7-vx-1 power law extending from the near infrared (such power laws connect smoothly between the near IR and radio for radio-loud QSOs). Out to the longest wavelength where there are abundant observations, 120uM, differences between radio-loud and radio-quiet QSOs are not apparent. Blazars generally are characterized by very nearly power law spectra extending from the optical all the way to the submillimeter or millimeter-wave regions, where the infrared joins onto the strong high-frequency radio emission. The strong polarization and rapid variability also seem to extend from the optical into the infrared with little attenuation. There are two important infrared-discovered variants to the patterns of behavior described previously. First, a census of the most luminous objects in the local part of the Universe shows them to be roughly equally divided into traditional QSOs and sources such as 13349+2438 that are relatively unremarkable optically, but have very large and powerful far-infrared emission. Current research is examining whether these latter objects are QSOs so heavily dust embedded that virtually virtually all their power is absorbed and reradiated thermally. Second, although about 10% of the high-frequency extragalactic radio sources have invisible or extremely faint visible counterparts, nearly all these objects are detectable in the near infrared and appear(from their rapid variability and strong polarization)to be a form of blazar with a very steep spectral cutoff from the near infrared to the visible. Because only a fraction of the total sample of high-frequency radio sources are blazars, the infrared variant is relatively common among this source source type, accounting perhaps 20-30% of them. INTERPRETATIONS The infrared emission of blazars is the most straightforward to interpret. Because it follows the rapid variations and strong polarization of the optical light very closely, it is clearly an extension of the optical synchrotron emission and, at least in the near infrared, is probably generated by the same population of relativistic electrons and in the same source region and magnetic field as the optical emission. The continuity of the power law spectra from the optical through the infrared provides further support for this view. The sharp spectral cutoffs in a significant portion of this source type suggest that the electron acceleration mechanism is operative only up to a very sharply defined upper energy limit. For the type 2 Seyfert galaxies, it is generally agreed that the infrared emission is dominated by thermal reradiation of the nuclear emission by circumnuclear dust. In the case of the archetype Seyfert 2, NGC 1068, the diameter of the nuclear source has been measured at 2-10uM and is in direct support of this conclusion. For other Seyfert 2 galaxies, the evidence is circumstantial, but convincing. For example, the spectral energy distributions are accurately reproduced by thermal radiation models and there is evidence in the optical emission line ratios for reddening by dust. The type of circumstantial evidence found for Seyfert 2 galaxies is weaker for Seyfert 1 and QSOs, and the nature of their infrared emission has been controversial for 20 years. Opinion is now leaning toward reradiation by dust as the dominant emission in most cases. The strongest argument is derived from the behavior during variability. The distance, d, of heated dust grains from the source of ultraviolet heating energy can be shown to be D2=eUV{LUV}, (1) =eIR{16*Tx4} where EIR is the infrared emission efficiency of the grains, eUV is the ultraviolet absorption efficiency, T is the grain temperature, * is the Stefan-Boltzmann constant, and LUV is the ultraviolet luminosity of the source. For any except highly contrived source geometries, this s distance must be small enough that the heated dust can respond in the observed variability time scale to changes in LUV propagating outward from the nucleus at the speed of light, or D