The Pennsylvania State University The Graduate School Department of Astronomy and Astrophysics UNVEILING THE ACCRETION DISKS THAT FUEL ACTIVE GALACTIC NUCLEI A Thesis in Astronomy and Astrophysics by Karen Theresa Lewis c 2005 Karen Theresa Lewis Submitted in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy December 2005 The thesis of Karen Theresa Lewis was read and approved1 by the following: Michael Eracleous Associate Professor of Astronomy and Astrophysics Thesis Adviser Chair of Committee Steinn Sigurdsson Associate Professor of Astronomy and Astrophysics W. Niel Brandt Professor of Astronomy and Astrophysics Donald Schneider Professor of Astronomy and Astrophysics L. Samuel Finn Professor of Physics Lawrence Ramsey Professor of Astronomy and Astrophysics Head of the Department of Astronomy and Astrophysics 1Signatures on file in the Graduate School. iii Abstract 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 transi- tion to a vertically extended, radiatively inefficient accretion flow at small radii. This scenario naturally explains the diminished wind in double-peaked emitters, but also of- fers a way to illuminate the outer accretion disk, which is necessary to produce the double-peaked emission lines. I critically analyze this hypothesis through robust esti- mates 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. Thus, alternative mechanisms for diminishing the disk wind should be sought; ideally these mechanisms should also offer a way to illuminate the outer accretion disk. Furthermore, robust estimates of the accretion rate should be determined for a much larger sample of double-peaked emitters in order to determine whether the distribution of accretion rates is continuous. 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; a few of the best observed objects have been modeled, with varying degrees of success, by invoking 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. To make further progress, it is important to continue monitoring these objects at least twice per year. Additionally, a few objects which showed significant variability should occasionally be monitored intensively (every few weeks) for several months at a time in order to probe variability taking place on the dynamical timescale. iv Table of Contents List of Tables ...................................... vi List of Figures ..................................... vii Acknowledgments ................................... viii Chapter 1. Introduction ................................ 1 1.1 The Accretion Disk Paradigm for Active Galactic Nuclei . ...... 1 1.2 A Brief History of Double-Peaked Emitters . 3 1.2.1 External Illumination by a Radiatively Inefficient Accretion Flow................................ 3 1.2.2 Connection Between Double-peaked Emitters and the General AGNPopulation ......................... 5 1.2.3 Challenges to the RIAF hypothesis . 6 1.2.4 Variability of the Double-Peaked Balmer Emission Lines . 7 1.3 Double-Peaked Emitters — Who Needs Them? . 8 1.4 TheGoalsofthisThesis ......................... 9 Chapter 2. Black Hole Masses in Double-Peaked Emitters ............. 11 2.1 Introduction................................ 11 2.2 Sample Selection, Observations, and Data Reduction . ...... 12 2.3 AnalysisandResults ........................... 17 2.3.1 FittingMethod .......................... 17 2.3.2 SourcesofSystematicError . 17 2.3.3 Notes on Individual Objects . 18 2.4 DiscussionandConclusions . 19 Chapter 3. XMM and RXTE Observation of 3C 111 ................ 23 3.1 Introduction................................ 23 3.2 Propertiesof3C111 ........................... 25 3.3 Observations and Data Reductions . 27 3.3.1 XMM-Newton ........................... 27 3.3.2 Rossi X-ray Timing Explorer .................. 29 3.4 TimingAnalysis ............................. 30 3.5 SpectralAnalysis ............................. 30 3.5.1 ContinuumModels ........................ 32 3.5.2 Models for the Fe Kα Line.................... 37 3.5.3 Combined Continuum and Fe Kα Emission Models . 42 3.5.3.1 Truncated Accretion Disk - Models #7a,b . 42 3.5.3.2 Highly Ionized Accretion Disk - Models #8a,b,c . 43 3.5.3.3 Partial Covering - Model #9 . 46 v 3.6 Discussion................................. 49 3.6.1 Origin of the Low Energy Component . 49 3.6.2 Interpretation of the Spectral Models . 50 3.7 Conclusions ................................ 52 Chapter 4. Long-Term Profile Variability in Double-Peaked Emitters ...... 54 4.1 Introduction................................ 54 4.2 Observations and Data Reductions . 57 4.3 ProfileAnlysis............................... 61 4.3.1 Relative Narrow and Broad Line Fluxes . 64 4.3.2 DifferenceSpectra ........................ 67 4.3.3 Variations in Profile Parameters . 75 4.3.4 Variations with Integrated Broad Hα Flux........... 79 4.4 Model Profile Characterization . 81 4.4.1 Physical Motivation . 81 4.4.2 Calculation of the Model Profiles . 83 4.4.3 Model Characterization . 84 4.5 Discussion and Interpretations . 89 4.6 Conclusions ................................ 91 Chapter 5. Conclusions and Suggestions for Future Work ............. 93 5.1 ViabilityoftheRIAFscenario. 93 5.2 Characterization of the Long-term Profile variability . ........ 94 5.3 Accretion Disk Winds in Double-Peaked Emitters . 95 Appendix A. Telluric Correction Method ...................... 96 Appendix B. Inclination Angle of the Disk in 3C 111 ............... 99 Appendix C. Total Galactic Hydrogen Column Density Towards 3C 111 ..... 100 Bibliography ...................................... 101 vi List of Tables 2.1 GalaxyProperties .............................. 13 2.2 TemplateStars ................................ 13 2.3 Velocity Dispersions and Derived Properties . ....... 21 3.1 ObservationLog ............................... 28 3.2 Best Fit Continuum Parameters . 33 3.3 Best Fit Parameters for Combined Continuum and Fe Kα Line Models . 47 4.1 GalaxyProperties .............................. 56 4.2 InstrumentalConfigurations . 58 4.3 LogofObservations ............................. 59 vii List of Figures 2.1 Observed Ca ii linesandBest-fitModels . 16 3.1 3C111X-rayLightcurve........................... 31 3.2 X-raySpectraof3C111 ........................... 34 3.3 Confidence Contours in the Column Density and Photon Index..... 35 3.4 Confidence Contour in the Folding Energy and Reflection Fraction . 36 3.5 Confidence Contours for Gaussian Fits to the Fe Kα Line ........ 39 3.6 Confidence Contours for the Disk-line Fits to the Fe Kα Line (Powerlaw +softGaussianContinuum) . 40 3.7 Confidence Contours for the Disk-line Fits to the Fe Kα Line (Powerlaw +Comptonreflectioncontinuum). 41 3.8 Ratio of the REFSCH and Power Law Models . 44 3.9 Confidence Contour in the Ionization Parameter and Reflection Fraction 45 3.10 Confidence Contour in the Column Density and Covering Fraction for thePartialCoveringModel . 49 3.11 Confidence Contours for the Gaussian Fit to the Fe Kα Line (Partial CoveringModel) ............................... 53 4.1 Double-Peaked Balmer Emission Line Profile . 61 4.2 Broad Hα LightCurves ........................... 66 4.3 DifferenceSpectra .............................. 68 4.4 Variations in Profile Properties as a Function of Time . ....... 77 4.5 Variations in Peak Separation with Broad Hα Flux............ 80 4.6 ModelProfiles................................. 85 A.1 Un-corrected Spectrum of IRAS 0236.6–3101 . 96 A.2 Example Fit of a Rapidly Rotating B-star and