Unveiling the accretion disks that fuel Active Galactic Nuclei

An increasing number of Active Galactic Nuclei (AGN) exhibit broad, double-peaked Balmer emission lines, reminiscent of those observed in Cataclysmic Variables; these double-peaked Balmer lines represent some of the best evidence for the existence of accretion disks in AGNs. There is considerable evidence to support the hypothesis that double-peaked emitters are "clean" systems in which the accretion disk is not veiled by a disk wind. This unobscured view affords the opportunity to study the underlying accretion disk which is believed to exist in all AGNs. In this thesis, I study two aspects of double-peaked emitters, namely the mechanism responsible for diminishing the accretion disk wind and the long-term profile variability of the double-peaked emission lines. It has been argued that double-peaked emitters have accretion flows that transition to a vertically extended, radiatively inefficient accretion flow at small radii. I critically analyze this hypothesis through robust estimates of the accretion rate in a few objects and also through an investigation of the X-ray spectra, which are sensitive to the structure of the inner accretion disk. I find that this hypothesis may be valid in some, but not all double-peaked emitters. A set of 20 double-peaked emitters has been monitored for nearly a decade in order to observe long-term profile variations in the double-peaked emission lines. Variations generally occur on timescales of years, and are attributed to physical changes in the accretion disk. The profile variability requires the use of non-axisymmetric accretion disk models such as, circular accretion disks with bright spots or spiral arms, or elliptical disks. I have characterized the variability of a group of seven double-peaked emitters in a model independent way and found that variability is caused primarily by the presence of one or more lumps of excess emission that change in amplitude, projected velocity, and shape over periods of several years. An elliptical accretion disk does not produce the correct variability patterns, and for those objects with a known black hole mass, the timescale for variability in this model is an order of magnitude longer than is observed. The spiral arm model produces variability on the correct timescale, but it is also unable to reproduce the observations. However, I suggest that with the simple modification of allowing the spiral arm to be clumpy, many of the observed variability patterns could be reproduced.