AGN emit roughly equal amounts of energy throughout the major part of the electromagnetic spectrum, extending from > 100µm in the FIR to > 10 keV in the hard X-ray. Roughly 10% also emit strongly at radio wavelengths and a significant fraction of these are strong -ray sources. Thus, to obtain a complete picture of these powerful sources, observations must be made at all wavelengths. To fully observe a single AGN requires obtaining time and observing on many different telescopes and satellites, often using more than one instrument in each case. In many parts of the spectrum these objects are sufficiently faint to push the limits of current technology. Given that many of them are variable on fairly short timescales, obtaining a single snapshot of a AGN SED is a daunting task and one that has not been completed for many AGN to date.
The number of compiled AGN SEDs in the literature has increased rapidly in the past decade, including many for individual objects (eg. Treves et al. 1988, Kolman et al. 1991, Kuhn et al. 1995, Puchnarewicz et al. 1995) and larger compilations at both low-redshift (Kriss 1988, Sanders et al. 1989, Barvainis 1990, Masnou et al. 1992, Elvis et al. 1994 (EWM94), Fiore et al. 1995, Laor et al. 1997) and higher redshift (Bechtold et al. 1994, Tripp, Bechtold & Green 1994). While many of these incorporate data observed over many years and often by many different groups in order to produce an SED, the majority are partially contemporaneous and provide a reasonable approximation to an instantaneous SED for relatively non-variable (ts 1 year) objects in at least the near-IR-optical range. When multiple epochs are available, it is also possible to estimate the level of variability and thus the uncertainty in the final SED. Examples of typical SEDs in the literature are given in Figure 1.
Figure 1. Rest-frame radio-X-ray spectral energy distributions (SEDs) for low-redshift radio-loud (upper) and radio-quiet (lower) quasars (from EWM94, their Figure 1). These are typical of the quality of SEDs in the current literature. The display of L vs shows the energy output as well as highlighting the structure of the SED.
A small number of AGN known to be strongly variable have been the subjects of coordinated campaigns to monitor and relate their variability at a number of wavelengths simultaneously. The AGN which has the most complete, simultaneous coverage is the brightest nearby quasar: 3C273 (Tuerler et al. 1998, Courvoisier 1998) for which fairly complete SEDs are available at a number of epochs.
There remain parts of the SED which are not well-observed in more than a handful of objects. The FIR and sub-mm have very limited data which recent new capabilities, such as the photometer on the Infrared Space Observatory (ISO) and SCUBA (Submm Common-User Bolometer Array) on the James Clerk Maxwell Telescope (JCMT), should begin to rectify over the next few years. The extreme ultra-violet (EUV) region was observed by the wide field camera (WFC) on ROSAT and by the Extreme UV Explorer (EUVE) but only a handful of the brightest, low-redshift sources have been detected. In the -ray, the Gamma-Ray Observatory (GRO) has made an industry of observing core-dominated, radio-loud quasars (Fichtel et al. 1994). The lack of strong -ray emission from other classes of AGN implies that beaming is an important factor and results in a lack of knowledge on -ray emission from the general AGN population. In the high-energy -ray, only three BL Lac objects (e.g. MKN421, Zweerink et al. 1997, MKN501, Catanese et al. 1997) have been detected to date, two of which show flaring behaviour.
The radio region is also relatively unknown for radio-quiet AGN, which represent 90% of the class. This situation is rapidly improving as deeper surveys (FIRST: Gregg et al. 1996, NVSS: Condon et al. 1998) and directed studies (Kukula et al. 1998, Blundell & Beasley 1998) are completed.