6.3. The X-ray BE

One of the distinguishing features of quasars and AGNs is that they emit over virtually the entire electromagnetic spectrum, from -rays to radio wavelengths, and their optical/UV radiation is only a fraction of the total emitted. Furthermore, radiation at other wavelengths almost certainly provides some of the most important keys for understanding both the BE and the nature of quasars themselves. Multi-wavelength observations are a vital part of quasar/AGN research.

The BE was discovered in the UV part of the spectrum, and observational constraints as well as atomic physics have dictated that studies of the luminosity dependence of broad-line emission are conducted almost exclusively in the rest-frame UV/optical bandpass. Improvements in the sensitivity and spectral resolution of X-ray telescopes have resulted in the detection of an important new line diagnostic of AGNs, in the form of the Fe K line at 6.4 keV. This feature is believed to form by fluorescence in high column-density material irradiated by an X-ray continuum. The observed K emission from Seyfert 1 nuclei can be very broad, and has been modeled as fluorescence in a relativistic accretion disk (Nandra et al. 1997b and references therein). Narrow components are also observed, however (e.g., Guainazzi et al. 1994), which may arise from structures such as a circumnuclear torus of high column density matter (Krolik et al. 1994; Ghisellini et al. 1994).

The recent availability of X-ray spectra of QSOs with decent spectral resolution has made it possible to investigate the luminosity dependence of Fe K. Iwasawa & Taniguchi (1993) employed GINGA spectra to argue that K equivalent widths were systematically weaker in more luminous systems, and drew attention to the parallel with the ultraviolet BE. While this finding was subsequently challenged (Nandra & Pounds 1994), a recent analysis based on ASCA spectra (Nandra et al. 1997a), reviewed by Paul Nandra at this conference, lends strong support to a BE for the Fe K line.

Several interesting comparisons can be made between the detailed phenomenology of the UV and X-ray BEs. The existence of an X-ray BE and its parallel in the UV lines initially led Iwasawa & Taniguchi to argue that this constituted evidence of a physical origin of the K emission in the BLR. However, the X-ray correlation appears to be strongly influenced by a luminosity dependence in the K wings, in contrast with the core-dominated UV trend. Under the existing interpretations of the K profile, this behavior can be taken as evidence that the X-ray BE is largely an accretion disk phenomenon. A connection between the K emission and the BLR may still be possible if the high-ionization broad lines are largely produced in a disk structure (e.g., Murray & Chiang 1998). While additional detailed physics is required to account for the different line profile behaviors of the UV features and K lines, a further interesting similarity exists in that the BE for K is strongest in the red wing, in agreement with the UV trend.

Fluorescence of Fe K in the AGN context is apparently part of a larger pattern of spectral ``reflection'' signatures, resulting from scattering and emission by X-ray-irradiated media with large column density, which also include a Compton-reflection hump peaking in flux density at ~ 30-50 keV and Fe K-edge absorption at ~ 7 keV; theoretical studies predict an additional reflection component of thermal emission that would contribute to the optical/UV ``Big Blue Bump'' (BBB; Guilbert & Rees 1988; Lightman & White 1988). Nandra et al. (1995) have pointed out an important difficulty in associating the BBB with reflection, in that luminous QSOs that typically exhibit a prominent BBB (as parametrized, for example, by a steep _ox) also exhibit weak reflection signatures in the X-ray bandpass, i.e. weak K and Compton reflection features. This result strongly suggests that processes other than reflection/reprocessing of the X-ray continuum dominate the generation of the optical/UV continuum in AGNs.

The X-ray BE shows some dependence on source radio properties, with radio-quiet QSOs featuring characteristically larger W for K than radio-loud sources show (Nandra et al. 1997a; Reeves et al. 1997). This result runs counter to the UV pattern. A possible means of reconciling these findings is to postulate a beamed component to the X-ray continuum in radio-loud sources, which is seen by us ⁽⁷⁾ and the broad-line clouds but not by the medium responsible for Fe K fluorescence. This component would give rise to greater heating of the BLR clouds and hence the observed enhancement in UV line strengths, while strengthening the observed X-ray continuum and thus diluting the K equivalent widths. An anisotropic component to the ionizing continuum might well be expected for radio-loud sources, which show evidence of relativistic outflows that naturally produce beamed radiation fields in the observer's frame. An anisotropic component to the X-ray emission has been suggested previously for other reasons (e.g., Browne & Murphy 1987), and an enhancement of this component in radio-loud sources would be consistent with the fact that radio-loud quasars are characteristically harder in _ox than radio-quiet systems (e.g., Zamorani et al. 1981).

AGNs show substantial variability in the X-ray continuum, and recent work has also revealed detections of variability in Fe K emission in individual sources (Iwasawa et al. 1996; Yaqoob et al. 1996; Nandra et al. 1997c; see also Iwasawa & Taniguchi 1993). The existing studies do not provide a clear basis for drawing an analogy between the behavior of K, and the intrinsic BE seen in UV lines; the degree to which K flux is correlated with the continuum in variable sources is ambiguous at this point, and may be a complicated function of velocity across the line profile. Improved measurements are of more general interest for study of AGN structure on very small scales. Iwasawa & Taniguchi (1993) have also discussed variability in relation to the global X-ray BE, and suggest that the range in luminosity over which the correlation is observed (~ 4 orders of magnitude) argues, as in the UV case, against variability as the source of the ensemble trend.