Annu. Rev. Astron. Astrophys. 1997. 35: 445-502 Copyright © 1997 by Annual Reviews. All rights reserved

5.3. Other Issues Concerning the BLR

PARTIAL REDISTRIBUTION OF THE GAS IN THE BLR     Comparison of line profile variability during month-long campaigns separated by one or more years reveals that the line response is not stationary. This phenomenon, observed in the HIL and the LIL, is probably caused by changes in the distribution of the BLR gas in a few years (NGC 4151: Ulrich et al 1991, Perry et al 1994; NGC 3516: Wanders & Horne 1994; NGC 5548: Wanders & Peterson 1996). Although it is not possible at this time to offer a definitive interpretation of these changes, they suggest the presence of time variable inhomogeneities on the surface of the disk (possibly due to the magnetic field) which can (a) enhance the Balmer lines emission in some locations (e.g. bumps on the disk surface would intercept more continuum flux), thus altering the H; profile and (b) strengthen the gas extraction from the disk surface (at local enhancements of the magnetic field), thus lifting additional HIL clouds above the disk and producing shoulders and other features in the C IV line (Figure 5).

Figure 5. Examples of long- and short-term variations of the C IV line in NGC 4151. Variations within days are shown on each panel. One can appreciate the variations on time scales of years by comparing the two panels and considering that between 1985 and 1990 the C IV line was observed to be perfectly symmetrical, e.g. November 1988-January 1989. [From Ulrich et al (1991).] Ordinates in 10-14 erg s-1 cm-2 A-1.

That the line response evolves with time reduces the value of any single observation like the infall seen in NGC 4151 (November-December 1991) and in NGC 5548 (March-May 1993). Only if a certain behavior repeats itself over the years and for several AGN can it be considered to represent a standing feature of the velocity field or of the distribution of matter.

EMISSION LINES WITH DOUBLE PEAKS: ACCRETION DISKS, OUTF LOWS, OR BINARY BLACK HOLES?     A small number of AGN show double-peaked Balmer lines, suggestive of emission from a rotating disk. (Double peaks are present only in the LIL.) This profile appears preferentially among broad-line radio galaxies and galaxies with a very compact central radio source (Eracleous & Halpern 1994, Gaskell 1996b). A relativistic disk would have an extended red wing because of gravitational redshift, a stronger blue peak because of beaming, and a range of peak intensity ratios, B/R, that is well constrained. Several double-peak profiles have been successfully fitted with an axisymmetric accretion disk model at some epochs (Halpern & Filippenko 1988, Rokaki et al 1992, Eracleous & Halpern 1994), but later observations revealed changes of B/R inconsistent with the disk model (Miller & Peterson 1990). Elliptical disks, warps and spiral shocks have been proposed but with no definitive conclusions (Chakrabarti & Wiita 1994, Eracleous et al 1995b, Bao et al 1996).

The model where the BLR is a biconal inhomogeneous flow illuminated by a variable double beam appears unlikely because of the difficulties in accelerating the very dense matter that produces the LIL, and the model is little constrained (3C 390.3; Zheng et al 1991). Double peaks have also been interpreted as the signature of two orbiting black holes, each with its own BLR (Gaskell 1996a), resulting from the merger of two galaxies with central black holes. In 3C 390.3, the period is estimated to be ~ 300 years, corresponding to a total mass of 7 × 10⁹ M. A difficulty of this model is that no other drift has been well observed (except possibly in OQ 208; Marziani et al 1993) and also that one would expect the double peaks to appear in the HIL as well as in the LIL, in contrast to the available observations (Arp 102B: Halpern et al 1996; 3C 390.3: Wamsteker et al 1997).

Distinct double peak profiles form but one category of the complex profiles displayed by many AGN (Eracleous et al 1995a, Stirpe et al 1988). This suggests that the complex profiles may simply be caused by transient inhomogeneities and asymmetries in the emissivity and/or distribution of the emitting matter. This was also implied by the changes in the BLR gas distribution observed in NGC 4151 and NGC 5548 over several years (see Partial Redistribution of the Gas in the BLR). Symmetric inhomogeneities resulting in double peaks could be produced preferentially in the disk of radio sources through conditions related to radio jet formation.

EMISSION LINES DURING MINIMUM STATES; THE EFFECT OF THE LONG-TERM VARIATIONS     The long minimum of NGC 4151 in 1981-1988 (interrupted by short episodes at medium bright states, Figure 1b), has caused the regular decrease of the central part of the C IV, C III, and Mg II lines between 1978 and 1991 (Ulrich et al 1991). A similar variation of a medium width component of Ly has also been observed in 3C 390.3 (Clavel & Wamsteker 1987) and reveals, as in NGC 4151, the presence of an intermediate line region with a size of a few light-years and velocities of the order of 2000-3000 km s^-1.

During the deepest period of minimum NGC 4151 was the subject of two multi-month campaigns. The IUE spectra reveal the unexpected presence of two narrow emission lines, whose intensity varies by a factor of 2-3 in a few days apparently in phase with the small amplitude variations of the weak continuum (Ulrich et al 1985, Ulrich 1996). These lines, at (rest) 1518.5, 1594.4 Å and with FWHM less than 7 and 16 Å, respectively, are too narrow to be emitted by the entire BLR and must arise instead from two localized regions that have a special excitation mechanism.

These lines are best measured at minimum when the broad wings of C IV have faded but can also be seen, albeit with less contrast against the broad wings, at medium bright state (Clavel & Wamsteker 1987, Kriss et al 1992). Their origin is unclear: That they are C IV components (at -6100 km and +8500 km s^-1) emitted by a two-sided flow is an attractive possibility, but it is based on the uncertain assumption that the accretion disk is transparent.

THE APPEARANCE AND DISAPPEARANCE OF BROAD EMISSION COMPONENTS     Very large intensity changes in broad emission lines have been observed in different sets of circumstances, implying different origins for these large amplitude variations. First, some are in direct response to the large amplitude fluctuations of the ionizing continuum observed in a few AGN (see Sections 2.1 and 5.3). The nonzero delay between continuum and line variations rules out, in these cases, obscuration by a dust cloud. Second, unexpected appearances of broad components have been reported in three AGN that, while displaying definite signs of activity, had before the event rather narrow emission lines or somewhat broad lines with triangular profiles. Interestingly, in these three cases (Pictor A, NGC 1097, and M81), the prominent new broad component is double-peaked. The circumstances that produce these lines and the cause of their double-peaked profiles remain unclear, although the accretion of a star followed by the formation of an elliptic disk has been proposed (Halpern & Eracleous 1994, Sulentic et al 1995b, Bower et al 1996, Storchi-Bergmann et al 1995).

The third case of unexpected appearance of a broad emission component occurred in the nucleus of NGC 4552, an apparently normal elliptical. Between two HST observations separated by two years, the continuum brightened by a luminosity of 10⁶ L (Renzini et al 1995). Spectroscopic observations carried out after the brightening revealed the presence of a broad Mg II line. Was it accretion of a passing star or of an interstellar cloud by a dormant central black hole? A systematic search for such events in apparently quiescent galaxies could be a new way to discover black holes in the general galaxy population. Finally, as mentioned in Section 5.1, dust clouds crossing our line of sight to the BLR cause large reddening and flux variations of the continuum with simultaneous characteristic Balmer decrement variations.