Active Galaxies With Double-Peaked Emission Lines

The main part of my thesis project was a spectroscopic survey of moderate redshift, radio-loud active galactic nuclei (AGNs), in search of double-peaked H-alpha lines. The scientific goal was to find more examples of profiles that resemble the unusual, double-peaked H-alpha profile of Arp 102B. The main result was the discovery of eight new double-peaked H-alpha profiles that can be fitted with a simple, relativistic disk model. These profiles are the widest known in any AGNs. The hosts of these disk-like profiles (tentatively referred to as ``disk-like emitters'') stand out form the general AGN population because of a number of distinguishing characteristics that are reminiscent of LINERs. (1) average H-alpha full width at half maximum twice as large as the average radio-loud AGN, (2) an optical continuum around H-alpha which is dominated by starlight, (3) unusually large equivalent widths of low-ionization forbidden lines, and large [O I]/[O III] ratios compared to the other members of the sample. The properties of the disk-like emitters are most likely intrinsic, since they are not extensions of any of the known correlations which are thought to represent viewing angle effects in several unifying schemes for radio-loud AGNs.

Double-peaked emission lines provide the most direct dynamical evidence for the presence of accretion disks in radio-loud AGNs. A physical model in which the inner disk is an ion-supported torus which illuminates the outer disk with a hard ionizing continuum, goes a long way toward explaining the properties of disk-like emitters. Although alternative scenarios for their origin have been proposed (supermassive binary black holes, and jets and other radial flows), the data available today for at least two of the best-studied objects (3C 390.3 and Arp 102B) strongly favor the accretion disk over other possibilities. Thus, we are now concentrating on using double-peaked emission lines as tools for studying the dynamics of accretion disks in AGNs and more generally the structure of AGN broad-line emitting regions (these could be a combination of an accretion disk and its outflowing wind). My main collaborators in these projects are Jules Halpern, Alex Filippenko, Thaisa Storchi-Bergmann, Mario Livio, and Andrew Wilson. Our work in this area is now proceeding along two directions.

First, we are monitoring the variations of the profiles of double-peaked emission lines on time scales of years. Variability of the line profiles is a direct means of studying dynamical and thermal phenomena in AGN accretion disks, such as spiral waves, thermal instabilities, or the tidal influence of a second, unseen black hole. We have been monitoring about 20 objects for at least 5 years (some since the mid 1980s) by observing them from KPNO, CTIO, MDM, Lick, Keck, and the HET. We are finding clear patterns in the profile variability in some objects and by our continuing monitoring effort we hope to answer the following questions: (a) is this variability pattern universal? (b) do the variations repeat? (c) are the variations regular or even periodic? In parallel to the above observational program I have also undertaken theoretical work to refine and expand models of the origin of double-peaked emission lines and their variability and to apply them to available data. Included are calculations of emission-line profiles from relativistic elliptical disks which may result from the tidal disruption of star by a supermassive black hole. Using the elliptical disk model we were able to fit the profile of the H-alpha lines of the LINER galaxy NGC 1097. Now we are exploring whether the elliptical disk scenario can account for the variability of the double-peaked emission lines of NGC 1097 over the past decade. Another promising scenario that we have applied to 3C 390.3 and 3C 332 is a single-armed spiral superposed on an axisymmetric disk. The model line profiles match the observations fairly well and the observed variability time scale agrees with what the predictions dynamical models. A lot of the work of comparing models with data has been done by two graduate students at U.C. Berkeley, Andrea Gilbert and Jeff Newman.

Second, we are studying the relation between the profiles of the Balmer and UV emission lines of double-peaked emitters. A comparison of these line profiles constitutes a test of models for photoionized AGN accretion disks and winds. We have obtained some UV spectra with the HST and we are expecting a few more more. Our detailed study of the prototypical double-peaked emitter, Arp 102B, has showed that its Ly-alpha and H-alpha profiles are dramatically different, with the former much narrower than the latter. We interpreted this as a consequence of the extreme physical conditions in the parts of the disk where the H-alpha line is emitted. At those high densities and and column densities, the Ly-alpha photons are trapped by resonance scattering and destroyed by collisional de-excitation. Thus the Ly-alpha line that we do observe must come from a different region than the H-alpha line, perhaps a low-density, outflowing wind. With the larger sample of objects that we are now studying we will explore the relation between the Balmer and UV lines in AGNs with a very wide range of luminosities. Thus we hope to learn how the dominant line production site shifts from the accretion disk to the wind as the luminosity of the AGN approaches the Eddington limit.